Handbook of Capillary and Microchip Electrophoresis and Associated Micro Techniques, 3rd Edition

Dedication
———————
This undertaking would not have been possible without the support of
my family and colleagues. Dedication is, first and foremost, to
family—first, my wife, Lianne, and my children, Miranda, Kate, and
Ben—who not only recognize my passion for these academic
endeavors but also fully encourage me to pursue them. I appreciate the
sacrifices they made en route to completion of this undertaking.
Second, this laboratory compendium would not exist without the
immeasurable effort from the scientists who authored the
chapters—they are leaders in their particular fields, both nationally
and internationally—I am grateful that they saw value in contributing
here. In recognition of their willingness to contribute to this handbook,
even in the midst of their already overextended commitments, this book
is also dedicated to them.
Contents
Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Part I Fundamentals and Methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Chapter 1 Introduction to Capillary Electrophoresis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
James P. Landers
Chapter 2 Protein Analysis by Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

James M. Hempe
Chapter 3 Micellar Electrokinetic Chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

Shigeru Terabe
Chapter 4 Capillary Electrophoresis for Pharmaceutical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

Eamon McEvoy, Alex Marsh, Kevin Altria, Sheila Donegan, and Joe Power
Chapter 5 Principles and Practice of Capillary Electrochromatography . . . . . . . . . . . . . . . . . . . . . 183

Myra T. Koesdjojo, Carlos F. Gonzalez, and Vincent T. Remcho
Chapter 6 Capillary Electrophoresis of Nucleic Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

Eszter Szántai and András Guttman
Chapter 7 Analysis of Carbohydrates by Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . 251

Julia Khandurina
Chapter 8 The Coupling of Capillary Electrophoresis and Mass Spectrometry in

Proteomics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
Haleem J. Issaq and Timothy D. Veenstra
Chapter 9 Light-Based Detection Methods for Capillary Electrophoresis. . . . . . . . . . . . . . . . . . . 305

Cory Scanlan, Theodore Lapainis, and Jonathan V. Sweedler
Chapter 10 Microfluidic Devices for Electrophoretic Separations: Fabrication and Use . . . . 335
Lindsay A. Legendre, Jerome P. Ferrance, and James P. Landers
Part IIA Capillary-Based Systems: Core Methods and Technologies . . . . . . . . . . . . . 359

Chapter 11 Kinetic Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
Maxim V. Berezovski and Sergey N. Krylov
vii
viii Contents
Chapter 12 DNA Sequencing and Genotyping by Free-Solution

Conjugate Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
Jennifer A. Coyne, Jennifer S. Lin, and Annelise E. Barron
Chapter 13 Online Sample Preconcentration for Capillary Electrophoresis . . . . . . . . . . . . . . . . . . 413

Dean S. Burgi and Braden C. Giordano
Chapter 14 Capillary Electrophoresis for the Analysis of Single Cells: Sampling,

Detection, and Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429
Imee G. Arcibal, Michael F. Santillo, and Andrew G. Ewing
Chapter 15 Ultrafast Electrophoretic Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445

Michael G. Roper, Christelle Guillo, and B. Jill Venton
Chapter 16 DNA Sequencing by Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467

David L. Yang, Rachel Sauvageot, and Stephen L. Pentoney, Jr.
Chapter 17 Dynamic Computer Simulation Software for Capillary Electrophoresis . . . . . . . . . 515

Michael C. Breadmore and Wolfgang Thormann
Chapter 18 Heat Production and Dissipation in Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . 545

Christopher J. Evenhuis, Rosanne M. Guijt, Miroslav Macka, Philip J. Marriott, and
Paul R. Haddad
Chapter 19 Isoelectric Focusing in Capillary Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 563

Jiaqi Wu, Tiemin Huang, and Janusz Pawliszyn
Part IIB Capillary-Based Systems: Specialized Methods and Technologies . . . . . 581

Chapter 20 Subcellular Analysis by Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 583
Bobby G. Poe and Edgar A. Arriaga
Chapter 21 Chemical Cytometry: Capillary Electrophoresis Analysis at the Level of the

Single Cell. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 611
Colin Whitmore, Kimia Sobhani, Ryan Bonn, Danqian Mao, Emily Turner, James Kraly,
David Michels, Monica Palcic, Ole Hindsgaul, and Norman J. Dovichi
Chapter 22 Glycoprotein Analysis by Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 631

Michel Girard, Izaskun Lacunza, Jose Carlos Diez-Masa, and Mercedes de Frutos
Chapter 23 Capillary Electrophoresis of Post-Translationally Modified Proteins

and Peptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 707
Bettina Sarg and Herbert H. Lindner
Chapter 24 Extreme Resolution in Capillary Electrophoresis: UHVCE,

FCCE, and SCCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 723
Wm. Hampton Henley and James W. Jorgenson
Chapter 25 Separation of DNA for Forensic Applications Using Capillary

Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 761
Lilliana I. Moreno and Bruce McCord
Chapter 26 Clinical Application of CE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 785

Zak K. Shihabi
Contents ix
Chapter 27 Solid-Phase Microextraction and Solid-Phase Extraction with Capillary

Electrophoresis and Related Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 811
Stephen G. Weber
Chapter 28 CE-SELEX: Isolating Aptamers Using Capillary Electrophoresis . . . . . . . . . . . . . . . 825

Renee K. Mosing and Michael T. Bowser
Chapter 29 Microfluidic Technology as a Platform to Investigate Microcirculation . . . . . . . . . 841

Dana M. Spence
Chapter 30 Capillary Electrophoresis Applications for Food Analysis. . . . . . . . . . . . . . . . . . . . . . . . 853

Belinda Vallejo-Cordoba and María Gabriela Vargas Martínez
Chapter 31 Separation Strategies for Environmental Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 913

Fernando G. Tonin and Marina F.M. Tavares
Part IIIA Microchip-Based: Core Methods and Technologies . . . . . . . . . . . . . . . . . . . . . . 979

Chapter 32 Cell Manipulation at the Micron Scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 981
Thomas M. Keenan and David J. Beebe
Chapter 33 Multidimensional Microfluidic Systems for Protein and Peptide

Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1001
Don L. DeVoe and Cheng S. Lee
Chapter 34 Microchip Immunoassays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1013

Kiichi Sato and Takehiko Kitamori
Chapter 35 Solvent Extraction on Chips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1021

Manabu Tokeshi and Takehiko Kitamori
Chapter 36 Electrophoretic Microdevices for Clinical Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . 1037

Jerome P. Ferrance
Chapter 37 Advances in Microfluidics: Development of a Forensic Integrated DNA

Microchip (IDChip). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1065
Katie M. Horsman and James P. Landers
Chapter 38 Taylor Dispersion in Sample Preconcentration Methods . . . . . . . . . . . . . . . . . . . . . . . . . 1085

Rajiv Bharadwaj, David E. Huber, Tarun Khurana, and Juan G. Santiago
Chapter 39 The Mechanical Behavior of Films and Interfaces in Microfluidic Devices:

Implications for Performance and Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1121
Matthew R. Begley and Jennifer Monahan
Chapter 40 Practical Fluid Control Strategies for Microfluidic Devices . . . . . . . . . . . . . . . . . . . . . . 1153

Christopher J. Easley and James P. Landers
Chapter 41 Low-Cost Technologies for Microfluidic Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . 1169

Wendell Karlos Tomazelli Coltro and Emanuel Carrilho
Chapter 42 Microfluidic Reactors for Small Molecule and Nanomaterial

Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1185
Andrew J. deMello, Christopher J. Cullen, Robin Fortt, and Robert C.R. Wootton
x Contents
Part IIIB Microchip-Based: Specialized Methods and Technologies . . . . . . . . . . . . . . 1205

Chapter 43 Sample Processing with Integrated Microfluidic Systems . . . . . . . . . . . . . . . . . . . . . . . . 1207
Joan M. Bienvenue and James P. Landers
Chapter 44 Cell and Particle Separation and Manipulation Using Acoustic Standing
Waves in Microfluidic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1229
Thomas Laurell and Johan Nilsson
Chapter 45 Optical Detection Systems for Microchips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1253

James M. Karlinsey and James P. Landers
Chapter 46 Microfabricated Electrophoresis Devices for High-Throughput Genetic

Analysis: Milestones and Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1277
Charles A. Emrich and Richard A. Mathies
Chapter 47 Macroporous Monoliths for Chromatographic Separations in
Microchannels. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1297
Frantisek Svec and Timothy B. Stachowiak
Chapter 48 Microdialysis and Microchip Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1327

Barbara A. Fogarty, Pradyot Nandi, and Susan M. Lunte
Chapter 49 Microfluidic Sample Preparation for Proteomics Analysis Using

MALDI-MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1341
Simon Ekström, Johan Nilsson, György Marko-Varga, and Thomas Laurell
Chapter 50 Implementing Sample Preconcentration in Microfluidic Devices . . . . . . . . . . . . . . . . 1375

Paul M. van Midwoud and Elisabeth Verpoorte
Chapter 51 Using Phase-Changing Sacrificial Materials to Fabricate Microdevices for

Chemical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1419
Hernan V. Fuentes and Adam T. Woolley
Chapter 52 Materials and Modification Strategies for Electrophoresis Microchips . . . . . . . . . . 1441
Charles S. Henry and Brian M. Dressen
Chapter 53 Microfluidic Devices with Mass Spectrometry Detection . . . . . . . . . . . . . . . . . . . . . . . . 1459
Iulia M. Lazar
Chapter 54 Nanoscale Self-Assembly of Stationary Phases for Capillary

Electrophoresis of DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1507
Kevin D. Dorfman and Jean-Louis Viovy
Chapter 55 Nanoscale DNA Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1527

Laili Mahmoudian, Mohamad Reza Mohamadi, Noritada Kaji, Manabu Tokeshi, and
Yoshinobu Baba
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1543
Foreword
The esteemed Chinese philosopher Confucius placed works such as this in succinct context when he
wrote “You cannot open a book without learning something.” Each chapter of this book describes
remarkable advancements in one of the most controversial and yet, powerful, analytical tools in
separation sciences. Even the most accomplished practitioners in separation science will extract new
information on methods and applications that can serve unmet needs in science and medicine.
Even after more than two-and-a-half decades of research and development, many scientists
debate whether capillary electrophoresis (CE) technology has provided the speed, resolving power,
peak capacity, sensitivity, robustness and cost reduction promised by the pioneers of CE during its
genesis in the 1980s and 1990s. Thousands of researchers, focusing on some of the most widespread
and practical analytical goals ever considered (that utilize microscale volumes of sample), have
demonstrated the utility of CE on a wide range of applications across many disciplines, such as
forensic science, medical diagnostics, pharmaceutical science, biotechnology, and environmental
science. The last 7 years have seen CE technology evolve quickly from the research laboratory into
practical applications in many fields, and the total body of international CE literature continues to
grow at an ever-increasing rate.
The first flush of excitement came with the key role that the CE technology played in completion
of the Human Genome Project. The project was successfully completed ahead of schedule and at a
fraction of the predicted cost. This monumental breakthrough allowed the placement of CE-based
DNA sequencers in the most prestigious institutions providing medical-legal testing and in numerous
research centers worldwide working on different DNA projects. In addition, an official compendium
for the analysis of erythropoietin is now part of a monograph of the European Pharmacopeia. Con-
sequently, CE is now part of the quality control that precedes the release of commercial batches
of this critically important drug. Other major biopharmaceutical companies also have official CE-
based quality control methods filed with the FDA. In fact, the growth of widely used biotechnology
compounds now rivals the growth of conventional pharmaceutical products and, thus, stricter regula-
tions have been applied to biomolecule development due to the complexity of their physicochemical
properties.
Each contributor to this book has been actively engaged in CE or microchip research and is a leader
in the field. Accordingly, each author contributes various aspects of their experience, specifically
the detailed descriptions and awareness of the practical problems inherent in their particular area
of research and application. Furthermore, contributors provide sound advice on how to overcome
practical problems, thus enabling readers to pursue novel applications equipped with much of the
practical knowledge and wisdom possessed by the leading researchers in the field.
The chapters in this book indicate the extraordinary breadth and scope of the work carried out
in capillary electrophoresis and microchip technology over the last few years. The core section on
“Fundamentals and Methodologies” contains 10 chapters focused on CE technology. These chapters
provide an introduction to CE in general, followed by in-depth descriptions of the most widely used
modes of CE and their associated detection methods. Several important applications related to protein,
nucleic acid, and carbohydrate research are presented, as well as an overview of the emerging field
of fabrication and use of microchips. The next 20 or so specialized chapters continue to describe
specific applications and methodologies that can be addressed with capillary systems, including
studies on the analysis of cellular and subcellular entities, food analysis, environmental science,
xi
xii Foreword
clinical and forensic applications, solid-phase microextraction/online sample preconcentration, and

extreme resolution CE.
This bandwidth of applications is illustrative of the exploding world of microchip systems
that perform chemical and biochemical analysis. It is difficult, if not impossible, to determine
whether microfabricated analytical devices were predominantly inspired by capillary systems, or
by visionaries who recognized the potential of microfabrication techniques (originally designed for
microelectronics) to revolutionize chemical, biomolecular, and cellular analysis. Clearly, inspiration
has flowed in both directions, and the revolution is upon us. Techniques to exploit the precision,
scalability, and unique microscale phenomena enabled by microchip technology now span much of
the routine practice of biochemistry and biology, including cell manipulation, protein separations,
forensic science, molecular and nanomaterial synthesis, etc. The second part of this book explores
core microsystem technologies, such as immunoassays, flow control, device fabrication and reli-
ability. These core technologies are widely used in many applications, such as those outlined in
the specialized chapters that conclude this book, and they span everything from on-chip sample
preparation to genetic analysis to devices coupled with mass spectrometry.
It would appear that the days of categorizing capillary and microchip electrophoresis as narrow
endeavors with limited impact are long gone. The confluence of chemistry, biology, and microengi-
neering undoubtedly will continue to produce new tools for scientific discovery, as well as for forensic
applications and clinical diagnostics. These latter applications are poised for a technology revolution
that will paradigm shift their respective fields. Microdevices that accept and analyze the compo-
nents of whole blood—ions, proteins, nucleic acids—using integrated sample preparation domains
fluidically interfaced with the analytic process are coming to fruition and will soon become avail-
able. Moreover, microfluidics offers the unique capability to control and manipulate (sub)nanoliter
volumes of liquids in unique, precise, and reproducible ways that were previously unimaginable
in capillary systems. This fact, together with innumerable new developments and insight on cell
manipulation and analysis, and recent progress in (inexpensive) polymer microfabrication combine
to create exciting opportunities for future analytical microsystems that can be applied as disposable
diagnostic tools for site-of-analysis testing. It is our hope that readers will find this book as helpful
as the last two editions have been. This work is an outstanding contribution to the development of
increasingly powerful analytical tools that will continue to have a positive impact on science and
society in the years ahead.
Norberto A. Guzman, Ph.D.

Senior Research Fellow
Bioanalysis, Drug Metabolism, and Drug Toxicity
Johnson & Johnson Pharmaceutical R&D
Raritan, New Jersey
Prof. Dr. Albert van den Berg

BIOS Lab-on-a-Chip group
MESA+ Institute for Nanotechnology
University of Twente
The Netherlands
Matthew R. Begley, Ph.D.

Associate Professor
Department of Mechanical Engineering
Department of Materials Science & Engineering
University of Virginia, Charlottesville, Virginia
Editor
James P. Landers is professor of chemistry and professor of mechanical engineering at the University
of Virginia, as well as an associate professor of pathology at the University of Virginia Health System.
Professor Landers received a Bachelor of Science degree in biochemistry with a minor in
biomedicine from the University of Guelph in Ontario, Canada, in 1984. He earned his Ph.D. in
biochemistry from the same department in 1988. After a one-year postdoctoral fellowship at the
Banting Institute at the University of Toronto School of Medicine, he was awarded a Canadian
Medical Research Council Fellowship to study cancer biology and diagnostics under Dr. Thomas
Spelsberg, a breast cancer biochemist at the Mayo Clinic. He launched and directed Mayo Clinic’s
Clinical Capillary Electrophoresis Facility in the Department of Laboratory Medicine and Pathology
developing clinical assays based on capillary electrophoretic technology—some are still on-board
at Mayo today. Beginning as an assistant professor of analytical chemistry at the University of
Pittsburgh in 1997, he forayed into analytical microfluidic systems with the goal of developing
the next-generation molecular diagnostics platform. These efforts were bolstered by a move to the
University of Virginia, where access to a dedicated class-100 cleanroom for microchip fabrication
allowed his group to rapidly prototype microdevices for separations, DNA purification, and DNA
amplification. In addition to editing the first two editions of this book, he has authored more than
175 papers and 25 book chapters on receptor biochemistry, capillary electrophoretic method devel-
opment, microchip fabrication, and integrated microfluidic systems for application in the clinical
and forensic arenas.
xiii
Contributors
Kevin Altria Maxim V. Berezovski

Research and Development Department of Chemistry
GlaxoSmithKline York University
Harlow, Essex, United Kingdom Toronto, Ontario, Canada
and
Imee G. Arcibal Campbell Family Institute for
Department of Chemistry Breast Cancer Research
Pennsylvania State University University of Toronto
University Park, Pennsylvania Toronto, Ontario, Canada
Rajiv Bharadwaj
Edgar A. Arriaga Microfluidics Group
Department of Chemistry Caliper Life Sciences
University of Minnesota Mountain View, California
Minneapolis, Minnesota
Joan M. Bienvenue
Yoshinobu Baba Department of Defense DNA Registry
Department of Applied Chemistry and Armed Forces DNA Identification Laboratory
MEXT Innovative Research Center for Armed Forces Institute of Pathology
Preventive Medical Engineering Rockville, Maryland
Nagoya University
Nagoya, Japan Ryan Bonn
and Department of Chemistry
Health Technology Research Center University of Washington
National Institute of Advanced Seattle, Washington
Industrial Science and Technology Michael T. Bowser
Takamatsu, Japan Department of Chemistry
University of Minnesota
Annelise E. Barron Minneapolis, Minnesota
Department of Bioengineering
Stanford University Michael C. Breadmore
Stanford, California Australian Centre for Research on Separation Science
School of Chemistry
David J. Beebe University of Tasmania
Department of Biomedical Engineering Hobart, Australia
University of Wisconsin
Dean S. Burgi
Madison, Wisconsin
Molecular Diagnostics
Affymetrix, Inc.
Matthew R. Begley Santa Clara, California
Department of Mechanical and Aerospace
Engineering Emanuel Carrilho
and Department of Chemistry and Molecular Physics
Department of Materials Science and Engineering Institute of Chemistry at São Carlos
University of Virginia University of São Paulo
Charlottesville, Virginia São Carlos, Brazil
xv
xvi Contributors
Wendell Karlos Tomazelli Coltro Christopher J. Easley

Department of Chemistry and Molecular Physics Department of Molecular Physiology and
Institute of Chemistry at São Carlos Biophysics
University of São Paulo Vanderbilt University Medical Center
São Carlos, Brazil Nashville, Tennesse
Jennifer A. Coyne Simon Ekström

Department of Chemical Engineering Department of Electrical Measurements
Stanford University Division of Nanobiotechnology
Stanford, California Lund University
Lund, Sweden
Christopher J. Cullen
Department of Chemistry Charles A. Emrich
Imperial College London Department of Chemistry and Biophysics
London, United Kingdom Graduate Group
University of California
Mercedes de Frutos Berkeley, California
Institute of Organic Chemistry
Madrid, Spain Christopher J. Evenhuis
Australian Centre for Research on Separation
Andrew J. deMello Science
Department of Chemistry School of Chemistry
Imperial College London University of Tasmania
London, United Kingdom Hobart, Australia
Don L. DeVoe
Andrew G. Ewing
Department of Mechanical Engineering
Department of Chemistry
University of Maryland
Pennsylvania State University
College Park, Maryland
University Park, Pennsylvania
Jose Carlos Diez-Masa
Jerome P. Ferrance
Institute of Organic Chemistry
Madrid, Spain
University of Virginia
Sheila Donegan Charlottesville, Virginia
Department of Chemical and Life Sciences
Barbara A. Fogarty
Waterford Institute of Technology
Tyndall National Institute
Waterford, Ireland
Lee Maltings
Kevin D. Dorfman Cork, Ireland
Department of Chemical Engineering and Material
Science Robin Fortt
University of Minnesota Department of Chemistry
Minneapolis, Minnesota Imperial College London
London, United Kingdom
Norman J. Dovichi
Department of Chemistry Hernan V. Fuentes
University of Washington Department of Chemistry and Biochemistry
Seattle, Washington Brigham Young University
Provo, Utah
Brian M. Dressen
Department of Chemistry Braden C. Giordano
Colorado State University Nova Research Inc.
Fort Collins, Colorado Alexandria, Virginia
Contributors xvii
Michel Girard Katie M. Horsman

Centre for Biologics Research Department of Chemistry
Health Canada University of Virginia
Ottawa, Ontario, Canada Charlottesville, Virginia
Carlos F. Gonzalez Tiemin Huang

Department of Chemistry Convergent Bioscience Ltd.
Oregon State University Toronto, Ontario, Canada
Corvallis, Oregon
David E. Huber
Rosanne M. Guijt Microfluidics Department
Australian Centre for Research on Separation Sandia National Laboratories
Science Livermore, California
School of Chemistry
University of Tasmania Haleem J. Issaq
Hobart, Australia Laboratory of Proteomics and Analytical
Technologies
Christelle Guillo SAIC Frederick Inc.
Department of Chemistry and Biochemistry Frederick, Maryland
Florida State University
Tallahassee, Florida James W. Jorgenson
András Guttman The University of North Carolina at
Horváth Laboratory of Bioseparation Sciences Chapel Hill
University of Innsbruck Chapel Hill, North Carolina
Innsbruck, Austria
Noritada Kaji
Paul R. Haddad Department of Applied Chemistry and
Australian Centre for Research on Separation MEXT Innovative Research Center for
Science Preventive Medical Engineering
School of Chemistry Nagoya University
University of Tasmania Nagoya, Japan
Hobart, Australia
James M. Karlinsey
James M. Hempe
Children’s Hospital Research Institute for Children
Penn State Berks
and Department of Pediatrics
The Pennsylvania State University
Louisiana State University Health Sciences Center
Reading, Pennsylvania
New Orleans, Louisiana
Thomas M. Keenan
Wm. Hampton Henley
Department of Biomedical Engineering
University of Wisconsin
The University of North Carolina at Chapel Hill
Madison, Wisconsin
Chapel Hill, North Carolina
Charles S. Henry Julia Khandurina

Department of Chemistry Anadys Pharmaceuticals
Colorado State University San Diego, California
Fort Collins, Colorado
Tarun Khurana
Ole Hindsgaul Mechanical Engineering
Carlsberg Laboratory Stanford University
Valby Copenhagen, Denmark Palo Alto, California
xviii Contributors
Takehiko Kitamori Lindsay A. Legendre

Department of Applied Chemistry Department of Chemistry
Nagoya University University of Virginia
Aichi, Japan Charlottesville, Virginia
and
Micro Chemistry Group Jennifer S. Lin
Kanagawa Academy of Science and Technology Department of Chemical and Biological
Kanagawa, Japan Engineering
Northwestern University
Myra T. Koesdjojo Evanston, Illinois
Herbert H. Lindner
Oregon State University
Division of Clinical Biochemistry
Corvallis, Oregon
Innsbruck Medical University
James Kraly Innsbruck, Austria
Department of Chemistry Susan M. Lunte
University of Washington Departments of Chemistry and Pharmaceutical
Seattle, Washington Chemistry
Ralph N. Adams Institute of Bioanalytical
Sergey N. Krylov
Chemistry
University of Kansas
York University
Lawrence, Kansas
Toronto, Ontario, Canada
Miroslav Macka
Izaskun Lacunza School of Chemical Sciences
Institute of Organic Chemistry Dublin City University
Madrid, Spain Dublin, Ireland
James P. Landers Laili Mahmoudian
Department of Chemistry Department of Applied Chemistry
University of Virginia Nagoya University
Charlottesville, Virginia Nagoya, Japan
Theodore Lapainis Danqian Mao

Department of Chemistry Department of Chemistry
University of Illinois at Urbana-Champaign University of Washington
Urbana, Illinois Seattle, Washington
Thomas Laurell György Marko-Varga

Department of Electrical Measurements Department of Analytical Chemistry
Division of Nanobiotechnology Lund University
Lund University Lund, Sweden
Lund, Sweden
Philip J. Marriott
Iulia M. Lazar Australian Centre for Research on
Virginia Bioinformatics Institute and Department of Separation Science
Biological Sciences School of Applied Sciences
Virginia Polytechnic Institute and State University RMIT University
Blacksburg, Virginia Melbourne, Australia
Cheng S. Lee Alex Marsh

Department of Chemistry and Biochemistry Research and Development
University of Maryland GlaxoSmithKline
College Park, Maryland Harlow, Essex, United Kingdom
Contributors xix
María Gabriela Vargas Martínez Monica Palcic

Universidad Nacional Autónama de México Carlsberg Laboratory
Depto. de Química Valby Copenhagen, Denmark
Cuautitlán, Edo. México, México
Janusz Pawliszyn
Richard A. Mathies Department of Chemistry
Chemistry Department University of Waterloo
University of California Waterloo, Ontario, Canada
Berkeley, California
Stephen L. Pentoney, Jr.
Bruce McCord Advanced Technology Center
International Forensic Research Institute Beckman Coulter, Inc.
Florida International University Fullerton, California
Miami, Florida
Bobby G. Poe
Eamon McEvoy
Chemistry Department
Department of Chemical and Life Sciences
Waterford Institute of Technology
Waterford, Ireland
David Michels Joe Power

Amgen Department of Chemical and Life Sciences
Seattle, Washington Waterford Institute of Technology
Waterford, Ireland
Mohamad Reza Mohamadi
Department of Applied Chemistry Vincent T. Remcho
Nagoya University Department of Chemistry
Nagoya, Japan Oregon State University
Corvallis, Oregon
Jennifer Monahan
Birck Nanotechnology Center Michael G. Roper
Purdue University Department of Chemistry and Biochemistry
Lafayette, Indiana Florida State University
Tallahassee, Florida
Lilliana I. Moreno
International Forensic Research Institute Juan G. Santiago
Florida International University Department of Mechanical Engineering
Miami, Florida Stanford University
Stanford, California
Renee K. Mosing
Michael F. Santillo
Pennsylvania State University
Pradyot Nandi University Park, Pennsylvania
Department of Pharmaceutical Chemistry
Ralph N. Adams Institute of Bioanalytical Chemistry Bettina Sarg
University of Kansas Division of Clinical Biochemistry
Lawrence, Kansas Innsbruck Medical University
Innsbruck, Austria
Johan Nilsson
Department of Electrical Measurements Kiichi Sato
Division of Nanobiotechnology Department of Applied Biological Chemistry
Lund University The University of Tokyo
Lund, Sweden Tokyo, Japan
xx Contributors
Rachel Sauvageot Wolfgang Thormann

Advanced Technology Center Department of Clinical Pharmacology
Beckman Coulter, Inc. University of Bern
Fullerton, California Bern, Switzerland
Cory Scanlan Manabu Tokeshi

Department of Chemistry Department of Applied Chemistry
University of Illinois at Urbana-Champaign Nagoya University
Urbana, Illinois Aichi, Japan
and
Zak K. Shihabi Micro Chemistry Group
Department of Pathology Kanagawa Academy of Science and Technology
Wake Forest University School of Medicine Kanagawa, Japan
Winston-Salem, North Carolina
Fernando G. Tonin
Kimia Sobhani
Institute of Chemistry
University of São Paulo
University of Washington
São Paulo, Brazil
Seattle, Washington
Dana M. Spence Emily Turner

Department of Chemistry Department of Chemistry
Michigan State University University of Washington
East Lansing, Michigan Seattle, Washington
Timothy B. Stachowiak Belinda Vallejo-Cordoba

Department of Chemical Engineering Centro de Investigation en Alimentacion y
University of California Desarrolla, A.C.
Berkeley, California Hermosillo, Sonora, Mexico
Frantisek Svec Paul M. van Midwoud

The Molecular Foundry Pharmaceutical Analysis Group
Lawrence Berkeley National Laboratory Groningen Research Institute of Pharmacy
Berkeley, California University of Groningen
Groningen, the Netherlands
Jonathan V. Sweedler
Department of Chemistry Timothy D. Veenstra
University of Illinois at Urbana-Champaign Laboratory of Proteomics and Analytical
Urbana, Illinois Technologies
SAIC Frederick Inc.
Eszter Szántai Frederick, Maryland
Institute of Medical Chemistry, Molecular Biology,
and Pathobiology
B. Jill Venton
Semmelweis University
Budapest, Hungary
University of Virginia
Charlottesville, Virginia
Marina F.M. Tavares
Institute of Chemistry
University of São Paulo Elisabeth Verpoorte
São Paulo, Brazil Pharmaceutical Analysis Group
Groningen Research Institute of
Shigeru Terabe Pharmacy
University of Hyogo University of Groningen
Hyogo, Japan Groningen, the Netherlands
Contributors xxi
Jean-Louis Viovy Adam T. Woolley

Laboratoire Physicochimie-Curie Department of Chemistry and Biochemistry
Section de Recherche, Institut Brigham Young University
Curie Provo, Utah
Paris, France
Robert C.R. Wootton
Department of Pharmacy and Chemistry
Stephen G. Weber Liverpool John Moores University
Department of Chemistry Liverpool, United Kingdom
University of Pittsburgh
Jiaqi Wu
Pittsburgh, Pennsylvania
Convergent Bioscience Ltd.
Toronto, Ontario, Canada
Colin Whitmore David L. Yang
Department of Chemistry Advanced Technology Center
University of Washington Beckman Coulter, Inc.
Seattle, Washington Fullerton, California
Part I
Fundamentals and Methodologies
1 Introduction to Capillary
Electrophoresis
James P. Landers
CONTENTS
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.2.1 Why Electrophoresis in a Capillary? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.2.2 The Family of CE Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2.3 Capillary Zone Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2.3.1 Instrumentation and CE Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2.3.2 Role of EOF in CE Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.2.3.3 A Description of the Electrophoretic Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2.3.4 The Capillary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.3 Method Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.3.1 Steps in Designing a Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.3.2 Sample Parameters to Consider . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.3.3 Separation Parameters to Consider . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.3.3.1 Electrode Polarity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.3.3.2 Applied Voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.3.3.3 Capillary Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
1.3.3.4 Capillary Dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.3.3.5 Buffers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.4 Introduction of Sample into the Capillary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
1.5 On-Capillary Sample Concentration Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
1.5.1 Sample Stacking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
1.5.2 Sample Focusing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
1.5.3 Isotachoporetic Sample Enrichment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
1.5.4 Online Concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
1.6 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Appendix 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
A.1 Mobility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
A.1.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
A.2 Corrected Peak Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
A.2.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
A.3 Quantity of Sample Introduced into the Capillary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
A.3.1 Hydrodynamic Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
A.3.1.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
A.3.2 Electrokinetic Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
A.3.2.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3
4 Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques
A.4 Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
A.4.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
A.5 Efficiency. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
A.5.1 Example 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
A.5.2 Example 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
A.6 Joule Heating. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
A.6.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Appendix 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Appendix 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
A.3.1 Ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
A.3.2 Small Molecules: Charged and Neutral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
A.3.3 Peptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
A.3.4 Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
A.3.5 Nucleic Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
1.1 INTRODUCTION
While the term “electrophoresis” was coined in 1909 by Michaelis,1 it was the pioneering experi-
ments of Tiselius in 19372 that first showed the separation of serum proteins—albumin, and α-, β-,
and γ -globulins—by “moving boundary electrophoresis”; this provided the first intimation of the
potential use of electrophoretic analysis for biologically active molecules. However, this approach
to electrophoresis was limited by the incomplete separation of proteins, the relatively large sample
volume needed, and the necessity of relatively low electrical fields due to the convection currents
generated by Joule heating, even in the presence of dense sucrose solutions. The historical events
that followed in electrophoretic method development were paradigm shifting—for historical detail,
the reader is referred to two reviews in the journal Electrophoresis, one by Hjertén3 in the late 1980s
and the other by Rilbe4 in the mid-1990s. The 1940s saw major efforts directed at improving anti-
convective media for zone electrophoresis, while paper electrophoresis was shown to be applicable
to the analysis of a wide variety of molecules.5,6 Gels formed from starch7 and, later, agarose8,9
were developed for the analysis of peptides, proteins, and oligonucleotides. In the 1950s, Kolin10
attempted isoelectric focusing (IEF), but the absence of “ampholytes” led to short pH gradients that
had poor stability over time. It was in this period that the pioneering work of Hjertén3 laid the
groundwork for the capillary electrophoresis (CE)analysis of diverse analytes, ranging from small
molecules (inorganic ions, nucleotides) to proteins and viruses. Much of this work was conducted
with a functional CE-like instrument, albeit with 3 mm tubes that Hjertén constructed as early as
1959 (see Figure 1.1).
It was not until the late 1960s that Vesterberg11 synthesized ampholines to create stable pH
gradients that allowed for effective IEF to be accomplished and take its place among the routine
molecular tools for biochemical and biomedical analysis. The functionality and acceptance of starch
gels was followed by the introduction of acrylamide as a sieving matrix in 1959 where, for the
first time, control over pore size and stability was possible.12 Ornstein13 and Davis14 independently
introduced disc gel electrophoresis in 1964, which was followed by the introduction of sodium
dodecyl sulfate (SDS) as a denaturing agent for protein separation in 1969. Building on 30 years
of development, the 1970s bought the concept of stacking gel together with the use of SDS for the
separation of the components in T4 phage,15 which laid the groundwork for the seminal work by
two-dimensional electrophoresis [combination of IEF separation followed by polyacrylamide gel
electrophoresis (PAGE)] that was described by Dale and Latner16 and Macko and Stegemann.17
Stegemann introduced IEF in polyacrylamide gels followed by SDS–PAGE,18 which was developed
Introduction to Capillary Electrophoresis 5
FIGURE 1.1 Photograph of the CE system designed by Hjertén. On the left are stacked the high voltage supply
and electronic components for the detector, topped off by a strip-chart recorder. In the center is the carriage
with the capillary and the electrode vessels above the immersion bath, while the cooling reservoir flanks it on
the right.
into a fine art by O’Farrell19 in 1975. Interestingly, around this time, Virtanen20 followed the work of
Hjertén a decade earlier with the use of smaller internal diameter (0.2 mm) tubes, which eliminated
convection problems and simplified instrumental design. In numerous ways, all of this work seeded
the later advances that included development of sequencing gels in 1977, agarose gels a short time
later, and then pulsed-field gel electrophoresis in 1983. It was in the very same year (1983) that
Jorgenson and Lukacs21 defined electrophoresis in micron-scale capillaries.
The concurrent development of advanced separation technology in the form of high-performance
liquid chromatography (HPLC) made speed, high resolution, quantitative results, and automation a
reality. This platform saw widespread adoption in industrial and clinical laboratory settings, filling the
need where electrophoresis could not to meet the demands of quantitative analysis, preparative iso-
lation, or automation. These advantages proved HPLC effective for the analysis of small molecules,
oligonucleotides, peptides, and small proteins, but with the disadvantage that larger analytes (e.g.,
structural proteins) were problematic. Moreover, more stringent regulations on waste resulted in
increased cost for waste disposal, leading to organic solvent waste generated by HPLC considered
as a major contributor to operating costs. Micro LC, in both open-tubular and packed formats,22−24
were shown to have promise,25 but technical difficulties in the design of high-pressure, low-volume
solvent delivery systems and operational challenges in handling the severe pressure drop in packed
column chromatography deterred commercial development.
It was with this historical backdrop of increasing demands for high resolution, quantitative
precision of biopharmaceuticals, and control of waste management costs that CE arrived on the
analytical scene. The pioneering work of Hjertén3 and Virtanen20 preceded the demonstration that
CE was shown to be a viable analytical technique by Mikkers et al.26 and Jorgenson and Lukacs.21
CE demonstrated the potential for producing high-resolution separations of biopolymers, as well as
smaller pharmaceutical agents, and used miniscule amounts of both sample and reagents. In 1989, a
decade after Mikkers et al.26 carried out their defining CE experiments, CE was ready for primetime
6000
5000
4000
3000
2000
1000
0
1990 1992 1994 1996 1998 2000 2002 2004 2006
Capillary
FIGURE 1.2 Growth of the CE and microchip literature. The ISI Web of Science database was searched with
the subject keywords of “CE” on a year-by-year basis beginning in 1983 through March 2007 and plotted as a
function of publication year.
as a result of improvements in the sensitivity of detectors, advances in automation technology, and,

most importantly, the widespread availability of high-quality narrow-bore capillary silica tubing.
Beckman Instruments introduced the scientific community to CE with the commercial launch of the
first fully automated CE instrument in the form of the P/ACE™ 2000. The result of sound engineering
is that some of these models can still be found in labs today (ours being one of them).
Within a few years, research groups throughout the world had expanded the horizons of CE.
Hjertén,3 Mazzeo and Krull,27 and Wehr28 were early in the pursuit of improvements in capillary
coatings to prevent analyte adsorption to the capillary surface while techniques, for casting poly-
acrylamide and agarose gels in capillaries were developed and applied to protein29,30 and DNA31−33
separations. These have now evolved into more sophisticated coatings involving phospholipids,34
lipoproteins,35 cholesterol,36 and polyelectrolytes.37 Terabe et al.38,39 pioneered micellar electroki-
netic chromatography (MEKC), with the use of micellar solutions to separate neutral and charged
species, while IEF was adapted to the capillary format,40−42 using chemical mobilization methods
to move the focused protein zones past the detector. In addition Zhu et al.,43 Bruin et al.,44,45 and
Ganzler et al.,46 developed non-gel sieving matrices for size-based separation of biopolymers, which
began a flurry of activity defining linear polymers for DNA and proteins analysis by CE.47 Theoret-
ical studies have led to understanding the problems associated with the technique, and the methods
required to overcome them. Productization of more effective and sophisticated instrumentation has
made room for the development of new detectors48−50 and mass spectrometer interfaces.51 Many
of these topics are covered in detail in the chapters that follow in this book.
Beginning with four publications in 1983, growth of the CE literature was exponential for roughly
the first decade (not surprising with such small numbers), linearized from 1993 to 1998, and then
slowed (but did not plateau) over the next 5 years (1999–2003) (Figure 1.2). While these types of plots
often seem superfluous, the point is this—despite the slowing of growth in the CE field around the
turn of the century, interest in CE has not stalled. In fact, there has been modest resurgence since then.
Method development still continues, but the application focus has intensified in the last half decade.
Diversity is evident by application of the technique to a wide spectrum of analyses in a variety
of disciplines ranging from the detection and quantitation of priority pollutants in environmental
samples, to the analysis of the components of a single cell, the screening for abnormal proteins or
DNA fragments indicative of disease or specific typing an individual.
Consistent with the theme of the handbook, this chapter has a practice-oriented focus, with the pre-
sentation of select theoretical aspects of CE limited to the basic principles needed for understanding
how molecules separate in an applied voltage, and the factors that affect the separation. Appendix 2
provides the reader with a guide to troubleshooting typical problems that may be encountered with CE.
1.2 CAPILLARY ELECTROPHORESIS

1.2.1 WHY ELECTROPHORESIS IN A CAPILLARY?
As introduced in Section 1.1, electrophoresis has been one of the most widely utilized techniques
for the separation and analysis of ionic substances. Almost all modes of electrophoresis utilize
some form of solid support to prevent convectional distortion of the analyte bands. This has been
in the form of paper (high voltage separation of amino acids and other small organic molecules)
or, more commonly, polyacrylamide or agarose gels, which have been used extensively for both
protein and deoxyribonucleic acid/ribonucleic acid (DNA/RNA) analysis. In light of this, there is
little doubt that gel electrophoresis has been an invaluable analytical tool for modern biochemical
research. However, despite the ability to resolve the components of complex systems, traditional
forms of electrophoresis suffer from several disadvantages, most of which scientists had endured
for lack of an effective alternative. Perhaps the most obvious disadvantage is the speed of separa-
tion, which is ultimately limited by Joule heating (the heating of a conducting medium as current
flows through it). The relatively poor dissipation of Joule heat in slab systems limits their use to
low potential electric fields. Moreover, from a methodological perspective, the entire process is a
series of cumbersome, time-consuming tasks, from the casting of the gel, preparation and loading
of samples, electrophoretic resolution of the ionic/molecular species, to the final stage where the
gel is stained, and the results obtained. Other problems include poor reproducibility, particularly
with two-dimensional analysis, analyte-dependent differences in staining, which makes quantita-
tive accuracy difficult to achieve (e.g., glycosylated proteins have different dye-binding properties
than their unglycosylated counterparts), and the cumbersome methodology involved in modern
gel electrophoresis that makes it virtually impossible for the entire electrophoretic procedure to be
automated.
The use of capillaries as an electromigration channel for separation of a diverse array of analytes,
not only presents a unique approach to separation, but is also associated with several advantages over
the standard solid supports used during the evolution of electrophoresis. In particular, the physical
characteristics of narrow-bore capillaries make them ideal for electrophoresis. Fused-silica capillaries
employed in CE typically have an internal diameter (i.d.) of 20–100 µm (375 µm outside diameter,
o.d.), lengths of 20–100 cm, and are externally coated with a polymeric substance, polyimide, which
imparts tremendous flexibility to a capillary that would otherwise be very fragile (Figure 1.3). The
high surface-to-volume ratio of capillaries with these dimensions allows for very efficient dissipation
of Joule heat generated from large applied fields. This is illustrated in Table 1.1, which compares the
surface-to-volume ratio of a standard analytical slab gel system with standard capillaries used in CE.
From this table, it is clear why the slab gel is limited to fields in the range of ≈15–40 V/cm whereas
up to 800 V/cm can be applied to a capillary containing the same type of gel matrix shown early on
by Dovichi and coworkers.52 As a result of this ability to dissipate heat, electrophoretic separations
can easily be performed at up to 30,000 V with the external capillary environmental thermostatted
at ambient temperature.
In addition to the ability to dissipate Joule heat efficiently, the use of capillaries for electrophoresis
is associated with many advantages. With typical electrophoresis capillaries, the small dimensions
yield total column volumes in the microliter range, thus requiring the use of only milliliter quantities
of buffer. Moreover, adhering to the chromatographic rule of thumb restricting sample volume to
1–5% of the total capillary volume, sample volumes introduced into the capillary are in the nanoliter
range (as low as 0.2 nL). As a result, as little as a 5 µL of sample will suffice for repetitive analysis
on some commercial instruments. Considering the small reagent (buffer) and sample requirements,
as well as the rapid analysis times associated with the application of high fields (30,000 V), it is
clear why CE has been applied to, and has provided the solution to, a diverse number of analytical
problems.
(a) (b)
375 µm (od)
20–100 µm (id)
~5 µm
polyimide coat
(c)
Static Mobile
Layer Layer
(Stern Layer) (Outer Helmholtz
(~0.1 nm) Plane)
FIGURE 1.3 Diagram of the capillary, its inner surface and the ionic layer. (a) Shows an end view of a capillary
with a 375 µm o.d. and a 50 µm i.d. (b) An illustration of the double-ionic layer formed in bare silica capillaries
critical in generating endoosmotic flow. (c) A scanning electron photomicrograph of a 375 µm (o.d.) × 75 µm
(i.d.) capillary enlarged 170 times. The polyimide coating which grants the capillary flexibility is visible and
the “etch” used to break the capillary is clearly seen on the left.
TABLE 1.1
A Comparison of Surface-to-Volume Ratios for an Analytical Slab Gel and a 57 cm
Capillary Having a Varied Internal Diameter
Surface Area (mm2 ) Volume (µL) Surface-to-Volume Ratio
Slab gel (14 × 11.5 × 0.15 cm) 32,200 24,150 1.3
20 mm i.d. capillary 35.81 0.179 200
50 mm i.d. capillary 89.53 1.119 80
75 mm i.d. capillary 134.3 2.518 53
100 mm i.d. capillary 179.1 4.477 40
200 mm i.d. capillary 258.1 17.907 20
1.2.2 THE FAMILY OF CE MODES

In much the same way that standard gel electrophoretic techniques diversified, so did CE. This has
resulted in a family of specialized modes that collectively constitute “CE.” The main modes of CE
that have been developed and are presently being exploited include capillary zone electrophoresis
(CZE), MEKC, capillary electrokinetic chromatography capillary (CEC), capillary isoelectric focus-
ing (CIEF), capillary gel electrophoresis (CGE), and capillary isotachophoresis (CITP). Over the last
decade or so, this latter mode has been primarily used as an on-capillary preconcentration technique.
TABLE 1.2
Modes Used for Analysis of Various Classes of Analytes
CZE MEKC CEC CIEF CGE
Ions Small molecules Small molecules Peptides Nucleic acids
Small molecules Peptides Peptides Proteins
Peptides Proteins
Proteins Carbohydrates
Carbohydrates
CZE is the most universal of the techniques, having been shown to be useful for the separation
of a diverse array of analytes varying in size and character. Although a brief description of CZE
is given in this chapter, various aspects of the technique will be discussed in a number of chapters
to follow in this handbook. The other CE modes are covered exclusively, and in detail, in separate
chapters dedicated specifically to the theoretical and practical aspects of those modes, along with
examples illustrating applications amenable to that particular mode. Table 1.2 provides a reference
table categorizing the modes used for the analysis of various classes of analytes.
1.2.3 CAPILLARY ZONE ELECTROPHORESIS

Capillary zone electrophoresis is not only the simplest form of CE, but also the most commonly
utilized. Discussion of this mode permits the presentation of a generic design for the instrumentation
for CE. The addition of specialized reagents to the separation buffer readily allows the same instru-
mentation to be used with the other modes mentioned in the previous section: addition of surfactants
with MEKC, ampholines for CIEF and a sieving matrix (linear polymers, entangled matrices) for
CGE. The discussion on CZE in the following subsections allows for analysis of some of the basic
principles governing analyte separation by this technique.
1.2.3.1 Instrumentation and CE Analysis

A diagrammatic representation of a CE instrument is presented in Figure 1.4. The basic components
include a high voltage power supply (0–30 kV; possibly 60 kV if you are in Jorgenson’s lab),
a polyimide-coated capillary with an internal diameter ≤200 µm, two buffer reservoirs that can
accommodate both the capillary and the electrodes connected to the power supply, and a detector.
As will be discussed later in this chapter, thermostatting of the capillary is critical to efficient and
reproducible separations and, hence, some type of capillary thermostatting system should be used.
To perform a capillary zone electrophoretic separation, the capillary is filled with an appropriate
separation buffer at the desired pH and sample is introduced at the inlet. Both ends of the capillary
and the electrodes from the high voltage power supply are placed into buffer reservoirs and up to
30,000 V applied to the system. The ionic species in the sample plug migrate with an electrophoretic
mobility (direction and velocity) determined by their charge and mass, and eventually pass a detector
where information is collected and stored by a data acquisition/analysis system.
Electrophoretic mobility (µ) of a charged molecular species can be approximated from the
Debye–Huckel–Henry theory
µ = q/6π ηr (1.1)
Net flow
Capillary length
(20–100 cm)
Lamp Optics Detector

HighVoltage
Power
(5–30 kV)
Inlet buffer/ OutletBuffer

sample injection Post-detection
Post-injection
FIGURE 1.4 General schematic of a CE instrument.
where q is the charge on the particle, η is the viscosity of the buffer, and r is the Stokes’ radius of the
particle. The mass of the particle may be related to the Stokes’ radius by M = (4/3)π r 3 V , where
V is the particle specific volume of the solute. Although one might infer the direct proportionality
of mass and radius of the particle, empirical data suggest modifications of Equation 1.1 to allow
for the nonspherical shape, the counterion effects, and nonideal behavior of proteins and biological
molecules.53
With CZE, the “normal” polarity is considered to be [inlet—(+), detector—(−) outlet] as shown
in Figures 1.4 and 1.5. As electrophoresis ensues, the analytes separate according to their individual
electrophoretic mobilities and pass the detector as “analyte zones” (hence, the term capillary zone
electrophoresis or CZE). The fact that, under appropriate conditions, all species (net positive, net
negative, or neutral) pass the detector indicates that a force other than electrophoretic mobility is
involved. If the applied field were the only force acting on the ions, net positively charged (cationic)
substances would pass the detector while neutral components would remain static (i.e., at the inlet)
and anionic components would be driven away from the detector. It is clear that, if this were the
case, CE would be of limited use. Fortuitously, there is another force, “electroosmotic flow” (EOF),
driving the movement of all components in the capillary towards the detector when under an applied
field (and a normal polarity). EOF plays a principle role in many of the modes of CE and most
certainly in CZE. This is discussed briefly in the next section.
1.2.3.2 Role of EOF in CE Analysis

Electroosmotic flow was first identified in the late 1800s when Helmoltz54 conducted experiments
involving the application of an electrical field to a horizontal glass tube containing an aqueous salt
solution. Curious about the ionic character of the inner wall and the movement of ions, he found
that the silica imparted a layer of negative charge to the inner surface of the tube, which under an
applied electric field, led to the net movement of fluid toward the cathode. More than a century later,
this phenomenon still plays the fundamental role in CE analysis. Moreover, the importance of the
control of EOF has been realized and has become the focus of several research groups.
As a continuation of the pioneering work of Helmholtz, the basic principles governing EOF have
been evaluated extensively. As shown by the expanded region of the inner wall of a capillary in
Figure 1.3, the ionized silanol groups (SIO) of the capillary wall attract cationic species from the
buffer. Obviously, the buffer pH will determine the fraction of the silanol groups that will be ionized;
understanding the amorphous nature of silica and the pKa range (4–6) associated with the various
types of silanol groups55 is key. The ionic layer that is formed has a positive charge density that
DETECTOR
(Apply voltage)
t=0 N– +
+N
–
N +
–
– N
+
Electro osmotic flow
t=1 –– N +
N
(+) – + (−)
N +
–
–
+N
t=2 – N +
– N+
(−)
(+) –
N +
N
– +
t=n
– +
N
– N +
(+) – N + (−)
– N +
+ N _
Intensity
Migration time = EOF
Time (min)
FIGURE 1.5 Mobility of charged and uncharged molecules in an applied field.
decreases exponentially as the distance from the wall increases. The double layer formed closest to
the surface is termed the “Inner Helmholtz” or Stern” layer and is essentially static. A more diffuse
layer formed distal to the Stern Layer is termed the “Outer Helmholtz Plane” (OHP). Under an applied
field, cations in the OHP migrate in the direction of the cathode carrying waters of hydration with
them. Because of the cohesive nature of the hydrogen bonding of the waters of hydration to the water
molecules of the bulk solution, the entire buffer solution is pulled toward the cathode. This EOF
or “bulk flow” acts as a pumping mechanism to propel all molecules (cationic, neutral and anionic)
toward the detector with separation ultimately being determined by differences in the electrophoretic
migration of the individual analytes. The importance of EOF in capillary electrophoretic analysis is
highlighted in Figure 1.5. The buffer entering the capillary inlet behind the sample plug is represented
by a “graded shading” for illustrative purposes, and is identical to the buffer preceding the sample. As
electrophoretic migration occurs, all analytes are swept towards the detector by bulk flow. Provided
that the EOF is adequate but not too strong, the respective electrophoretic mobilities of each of the
analytes leads to the formation of discrete zones by the time they pass the detector. If the EOF is low,
diffusion of the analyte zones could result in substantial band broadening and, under conditions of
very low EOF, some of the analytes may not reach the detector within a reasonable analysis time.
As previously mentioned, the EOF is pH dependent and can be quite strong. For example, in
20 mM borate buffer at pH 9.0, the EOF is ≈2 mm/s, which translates to a flow of ≈4 nL/s in a
50 µm i.d. capillary. The inclusion of EOF into the calculation of velocity is essential and results in
vi = µapp E = (µep µeo )E (1.2)
where µep is the mobility due to the applied electric potential and µeo is the mobility due to EOF.
From a practical perspective, EOF acts as an electric field-driven pump that may be considered
analogous to the mechanical pump used in HPLC. While the simplicity of this pumping system has
obvious advantages over mechanical pumps (e.g., no moving parts, no mechanical wear) one does
not have the same degree of control associated with mechanical systems.
It is important to have some idea of the magnitude of EOF under the conditions used for a
given separation for two reasons. First, under conditions where EOF is very fast, components of
the mixture may not have adequate “on-capillary time” for separation to occur. Second, it is useful
to know where neutral compounds migrate in the obtained electropherogram. Knowing this, infor-
mation about the charge character of the sample components is obtained on the basis of whether
they migrate faster (cationic) or slower (anionic) than EOF. The EOF marker is also useful as an
internal standard for calculating “relative” migration times for the components of the sample from
the apparent migration times. Some compounds that adequately serve as neutral markers include
dimethyl formamide (DMF), dimethyl sulfoxide (DMSO), and mesityl oxide, although, because
of rapid volatilization, the latter compound is of limited use with samples kept at ambient tem-
perature. Typically, a “marker” is a 0.1% solution in water; a 1-s hydrostatic (0.5 psi) sample
injection of the 0.1% solution provides an adequate signal with ultraviolet (UV) detection at either
214 or 200 nm. Fluorescent markers (e.g., BODIPY) can serve the same purpose for LIF detection
These markers are not of use with low pH separations where most analytes migrate faster than
the EOF, which is extremely low. If a peak is still required as an internal standard for correction
of migration time, any compound with a fast cathodic electrophoretic mobility will suffice as a
“frontal marker.” We have found that a synthetic peptide containing seven lysine residues and a
single tryptophan (K3 WK4 ) functions adequately for this purpose at pH 2.5. A frontal marker may
also be useful at higher pH (e.g., when the neutral marker comigrates with a species of interest),
where cationic species such as normetanephrine can suffice.
It is interesting that, since its inception, the holy grail in CZE has been a mechanism to control
EOF. The application of radial field in a manner that controls the magnitude of the “zeta potential”
(or the thickness of the double layer) on the capillary wall would, in theory, allow one to simply dial
up the desired magnitude of EOF. This mechanism has been illusive for the better part of 15 years,
still has not been determined, but clearly still is of interest.56
1.2.3.3 A Description of the Electrophoretic Process

In the early 1980s, Jorgenson and Lukacs21 discussed the theory and basis for electrophoretic sepa-
rations in capillaries. The following brief discussion develops, albeit at a basic level, the theoretical
concepts describing the movement of charged molecules in a buffer-filled capillary under an applied
electrical field and the shape of the zone during the electrophoretic process. The separation process
is discussed in terms of resolution (how adequately two components are separated) and efficiency
(how long the separation takes). Finally, we discuss factors that affect the resolution. For more
detailed evaluation of these processes, the reader is referred to several excellent discussions on this
subject.57−59
1.2.3.3.1 Mobility of an Analyte in a Capillary

A charged particle in solution will become mobile when placed in an electric field. The velocity,
vi , acquired by the solute under the influence of the applied voltage H, is the product of µapp , the
apparent solute mobility, and the applied field E (E = H/L, where L is the length of the field)21
vi = µapp E (1.3)
µapp is a property of the particle, and proportional to its charge, and inversely proportional to the
frictional forces acting upon it in solution. The electrical force can be given by
Fe1 = qE (1.4)
and the frictional force on a spherical ion is
Ffr = − 6π ηrvi . (1.5)
During the electrophoresis, a steady state is attained, where the two forces are equal, but in
opposite directions:
qEr = − 6π ηrvi (1.6)
Solving for velocity (or E) and substituting Equation 1.6 into Equation 1.3 and rearranging for
µ, we obtain
µapp = q/6π ηr (1.7)
From this equation, it is evident that the mobility of the analyte is a property of both the charge
and size; a small, highly charged particle will have a high mobility, while a large, minimally charged
species will have a low mobility. Refer to Appendix 1 for example on calculating mobility.
1.2.3.3.2 Shape of the Analyte Zone

Under the initial conditions of electrophoresis, the boundary between the buffer solution and the
solute mixture forms a zone of infinitesimal thinness at right angles to the direction of applied current
and migration. As migration proceeds, this initially sharp boundary will undergo a progressive
deterioration in shape. The most important influence on this process is diffusion, as the initial
conditions impart a severe concentration gradient across the zonal boundary. By applying Fick’s
second law of diffusion, Weber57 has shown that the variation of solute concentration in the direction
of migration is given by the equation
−(x−vi t)2
k
Cx,t = e 4Di t (1.8)
2(Di πt) 1/2
where C is the solute concentration at a distance x from the initial position after time t, vi is the
electrophoretic velocity of the solute, Di is the diffusion coefficient of the solute, and k is a constant.
By integration, initially with the boundary conditions at t = 0, at distant x = ∞, C = 0, and a second
time with t = 0, x = 0 and dC/dx = 0, we obtain a mathematical description of the concentration
profile. As the solute migrates a distance x from the origin after time t, the zone may be described
by a Gaussian curve.
The characteristics of a Gaussian profile describe the peak maximum and the peak width. The
maximum is dependent on initial concentration of solute. The width depends on the length of time
from initial conditions (i.e., application of sample in to the system) and the diffusion constant, Di .
The width may be given by the distance between the inflection points, in our case, 2/(2Di t)1/2 .
From analogy with probability calculations, the width of a Gaussian curve is termed the standard
deviation (σ )
σ = (2Di t)1/2 (1.9)
and the square of standard deviation is called the variance
σ 2 = 2Di t. (1.10)
Remembering that
t = Ld /vi = Ld Lt /µapp V = Ld /µapp E, (1.11)
where Ld is the distance from the origin to the detector, and Lt is the total length of the capillary (i.e.,
the length of the applied electric field), we may substitute into Equation 1.10, to obtain
σ 2 = 2Di Ld /(µapp E). (1.12)
One should remember that the discussion to this point has only dealt with diffusion as a dis-
persive phenomenon affecting peak shape. Other factors that contribute to what is observed on the
electropherogram as band broadening, the practical result of variance, will be discussed later.
1.2.3.3.3 Resolution and Efficiency

The simplest way to characterize the separation of two components is to divide the difference in
migration distance by the average peak width to obtain resolution (Res)
Res = 2(xi2 − xi1 )/(w1 + w2 ) (1.13)
where xi is the migration distance of the analyte i, and the subscript 2 denotes the slower moving
component, and w is the width of the peak at the baseline.57 We can readily see that the position
of a peak, xt , is determined by the electrophoretic mobility. The peak width, w, is determined by
diffusion and other dispersive phenomena. For two neighboring peaks, w1 = w2 , and
Res = (xi2 − xi1 )/w2 . (1.14)
From the equation describing a Gaussian curve, the two peaks touch at baseline when
xi = x2 = 4σ , and Res = 1, or
Res =
xi /4σ . (1.15)
Remembering that distance is equal to velocity multiplied by time (xi = vt t) and substituting for σ
from Equation 1.9 into Equation 1.15, we obtain
Res = (
µapp E)t/4(2Di,avg t)1/2 (1.16)
where
µapp is the difference in apparent electrophoretic mobility of the two solutes and Di,avg is
the average diffusion of the two solutes.
To obtain a measure of efficiency for the process, we use probability theory.58 For a random walk
process of length L, made of n steps, the variance is given by
σ = l(n)1/2 (1.17)
where l is the length of each step. If each step is independent of any other step, each contributes to
the total variance of the process59

σtot
2
= σi2 . (1.18)
Substituting from Equation 1.17 and rearranging
1/n = L 2 /σtot
2
= N. (1.19)
The number of steps in the random process, n, is inversely related to the number of theoretical
plates, N, a measure of efficiency for the process.
We may substitute for L and σ in Equation 1.19 to express
N = (µavg E)2 t 2 /(2Di t) (1.20)
where µavg is the average mobility of the two solutes. By comparing the expression for Res (Equation
1.16) with the definition of N (Equation 1.20), we obtain an expression relating resolution to the
number of theoretical plates
Res = (1/4)(
µapp /µavg )N 1/2 . (1.21)
The utility of Equation 1.21 is that it permits one to independently assess the two factors that affect
resolution, selectivity, and efficiency. The selectivity is reflected in the mobility of the analyte(s),
while the efficiency of the separation process is indicated by N.
If Res = 1, then
N = 16/(
µapp /µavg )2 . (1.22)
Another expression for N is derived from Equation 1.19, using the width at half-height of a
Gaussian peak
N = 5.54(L/w1/2 )2 , (1.23)
where 5.54 = 8 ln 2, and w1/2 is the peak width at half-height.60

At this point, it is important to note that it is, in fact, misleading to discuss theoretical plates
in electrophoresis. The concept is a carry-over from chromatographic theory, where a true partition
equilibrium between two phases is the physical basis of separation. In electrophoresis, separation of
the components of a mixture is determined by their relative mobilities in the applied electric field,
which is a function of their charge, mass and shape. The theoretical plate is merely a convenient
concept to describe the analyte peak shape, and to assess the factors that affect separation. Refer to
Appendix 1 for examples on calculating resolution and efficiency.
1.2.3.3.4 Source of Variance

While N is a useful concept to compare the efficiency of separation among columns, or between
laboratories, it is difficult to use to assess the factors that affect that efficiency. This is due to the fact
that it refers to the behavior of a single component during the separation process, and is unsuited to
describing the separation of two components or the resolving power of a capillary. A more useful
parameter is the height equivalent of a theoretical plate (HETP).58
HETP = L/N = σtot

2
/L. (1.24)
HETP might be thought of as the fraction of the capillary occupied by the analyte. It is more
practical to measure HETP as an index of separation efficiency, rather than N, as the individual
components that contribute to HETP may be individually evaluated and combined to determine an
overall value. The variance of multiple dispersive phenomena on the analyte may be summed:
σtot
2
= σdiff
2
+ σT2 + σint
2
+ σwall·
2
(1.25)
A consideration of all the factors influencing σtot

2 should include not only diffusion, but also differ-
ences in mobility or diffusion generated by Joule heating, the reality that the sample is not introduced
as a thin disk but as a plug of finite dimensions, and interaction of analytes with the capillary wall.
Each of these factors will be addressed separately in a simplistic manner. Theoretical derivation will
be given only to emphasize the importance of various parameters, and how they contribute to the
overall result. References shall direct those desiring more information to the appropriate literature.
Two studies have addressed the variance due to a variety of sources. Huang et al.61 investigated
small molecules; Jones et al.,62 studied proteins as well as amino acids and a neutral marker. The
two studies demonstrate that the band broadening observed in CE is in excess over calculated values
of diffusion and analyte interaction. Huang et al.61 and Jones et al.62 attribute the excess variance to
sample introductory practices.
1.2.3.3.4.1 Variance Caused by Temperature Temperature control is an issue, because it may

be effected by the efficiency with which the capillary is thermostatted, or by choice of buffer ionic
species or ionic strength. Thus, from a practical point of view, we discuss the variance caused by
temperature to impress upon the practitioner the importance of controlling temperature in obtaining
reproducible electropherograms.
Current passing through a conducting solution generates heat. A sample way of looking at
the problem is to compare the equivalent expressions for the applied potential, H, and heat
generation, W
H = i/κπ r 2 (1.26)
W = i2 /κ(π r 2 )2 (1.27)
where i is the current through the electrolyte solution, π r 2 the cross-sectional area of the capillary,
and κ is the specific conductance of the buffer. By combining Equations 1.26 and 1.27, to take the
ratio we obtain
H/W = (i/κπr 2 )(κ(πr 2 )2 /i2 ) = π r 2 /I (1.28)
From Equation 1.28, it can be readily seen that the reduction of heat production can be achieved
by reducing the current density in the capillary. This may be accomplished by increasing the cross-
sectional area of the capillary or by reducing the current. The latter is preferred, since increasing
the diameter of the capillary results in a reduction of the surface-to-volume ratio, and leads to less
efficient heat dissipation (see Table 1.1). The choices to reduce the sysetm current lie in carrying out
the separation at a lower voltage, or reducing the ionic strength of the separation buffer.
To develop the quantitative expression for thermal heating, we need to describe the electrophoretic
front, and its behavior under a thermal gradient. According to the Poiseuille equation, which describes
the parabolic flow due to pressure,
vz =
Pr 2 /8Lη. (1.29)
Joule heating of the electrolyte solution creates a similar flow profile, the equivalent expression
being
vz = v(1 + E 2 Cb Br 2 /4kb T 2 )[1 − (rx /r)2 ] (1.30)
where E 2 is the rate of heat generation per unit volume, Cb is the buffer concentration, kb is the
thermal conductivity of the electrolyte solution. is the equivalent conductanceof the electrolyte
solution, T is the absolute temperature, rx is the radial position (which varies from 0 at the center of
the capillary to r at the capillary wall), and B is a buffer-related viscosity constant.63 The variance
caused by dispersion in a parabolic velocity profile is given by
2
σT2 = 2Di t + r E
6 4
Cb2 B2 2 t 24Di 8kb T − E Cb Br
2 2 2
. (1.31)
For most capillary applications, where radius is small (25–50 µm), and E is kV, E 2 Cb Br 2 is
8kb T 2 and the second term reduces to
σT2 = r 6 E 4 Cb2 B2 2 t/1536Di kb2 T 4 . (1.32)
The strong dependence of σT2 on the radial dimension of the capillary (r 6 ) and the field strength
(E 4 )demonstrate the importance of performing high voltage electrophoretic separations in narrow-
bore capillaries. It also highlights the necessity of obtaining efficient capillary cooling, to prevent
thermal effects from not only affecting sample liability, but also to avoid affecting solute mobility.
For small molecules with relatively larger diffusion constants (on the order of 10−5 cm2 /s) Grushka
et al.63 have calculated a maximum radius of 65 µm with a 0.1 M buffer solution at 30 kV, and
130 mm with 10 µM buffer for less than 5% increase in dispersion. With larger molecules, having
diffusion constants on the order of 10−6 cm2 /s, these dimensions were halved. Jones et al.62 have
introduced a novel method to study Joule heating effects and nonideal plug flow contributions
to analyte dispersion, with polarity reversal at constant voltage. By recording the variance (band
broadening) of the peaks over time and plotting the slope of the obtained line versus applied potential,
these authors have developed a measure of nonideal behavior. A comparison of charged analyte
behavior with neutral solute variance (which should be unaffected by electrophoretic mobility effects
of thermal heating, but affected equally by the temperature-dependent effects on a diffusion and
viscosity) allows an estimate of the Joule heating effect. With adequate thermostatting, one need
not worry about thermal dispersion under normal operating conditions (see Reference 64), but one
needs to be aware of how temperature generated within the capillary can affect the separation and
resolution of the analytes.
1.2.3.3.4.2 Variance Caused by Finite Sample Introduction Volume In the mobility section
above, we assume the solute was initially present as a think disk that dispersed into a zone or band.
In reality, the sample is introduced into the capillary as a cylindrical plug. Sternberg60 derived the
commonly used equation describing variance due to sample introduction.
σint
2
= lint
2
/12 (1.33)
where lint is the length of the sample introduction plug. This formula assumes that the sample is
introduced as a rectangular plug, which is a close approximation for CE. As the plug has a finite
volume, it also has measurable length. Huang et al.61 and Jones et al.62 have investigated the relative
contribution of introduction plug length (by hydrodynamic or electrokinetic methods) to the observed
peak width, and concluded that lint /Ld of less than 3% does not lead to excessive band broadening.
Only at very small plug lengths (<200 µm) does one observe variance due to diffusion of the sample
plug; typical sample introduction plug lengths of 300–6000 µm (0.3–6 mm) obscure this effect.
Huang et al.61 and Jones et al.62 conclude that sample introduction volume, that is, plug length, is
the most significant factor in excessive band broadening observed in CE.
1.2.3.3.4.3 Variance Caused by Analyte–Wall Interactions The interactions of analyte and

capillary wall, or components within the sample solution are numerous, complex and sample specific.
The best approach to understanding the band broadening due to adsorption is the approach of
McManigill and Swedberg.64 The general equation of Giddings59 may be adapted to include an
adsorption term
2
2
HETP = l2 /12L + 2Di,T /u + k µ 1 + k r0 4Di,T + 2 kd , (1.34)
where l is the length of the injected plug, L is the total length of the capillary, u is the flow rate,
k the capacity factor, r0 is the radius of the capillary, Di,T is the diffusion coefficient in the des-
ignated solvent at the specific temperature, and kd in the first-order dissociation constant off the
surface.
Through carefully designated and executed experiments, a value of for k may be determined
at differing flow rates, and the effect on plate height calculated. McManigill and Swedberg64 have
shown that the value for k is not zero under any conditions, and that a value for k as low as 0.001
can affect the shape of the peak, and, consequently, the efficiency of separation.
To obtain estimates of the relative contribution to overall variance, Jones et al.62 used a voltage
interruption/polarity reversal method. Huang et al.61 attribute part of the excessive peak width
observed in their experiments to analyte–wall interaction. One may minimize such interactions by
the appropriate choice of buffer pH, ionic strength, or buffer additives (see Section 1.3).
The variance of multiple dispersive phenomena on the analyte may be summed as
σtot
2
= σdiff
2
+ σT2 + σint
2
+ σwall
2
. (1.35)
To enable the reader to visualize the magnitude of the contributions by each of these variances,
we include some typical calculations.
For a small molecule, Di = 10−5 cm2 /s, time of analysis might be 10 min (600 s), introduction
plug of 1.2 nL into a 50 µm by 47 cm capillary, 40 cm to the detector, results in a plug length
of 0.6 cm
σdiff
2
= 12 × 10−3 ,
σT2 = σdiff
2
= 12 × 10−3 ,
σint
2
= 3.0 × 10−2 ,
σtot
2
(exp . observed) = 5 × 10−2 .
σtot
2
(calculated) = 1.2 × 10−2 + 1.2 × 10−2 + 3.0 × 10−2 + σwall
2
= 5.4 × 10−2
These figures indicate σint

2 is, by far, the largest contributor to the band broadening with a small,
highly diffusible analyte.

For a protein, Di = 10−6 cm2 /s, time of analysis might be 10 min (600 s), introduction plug of
1.2 nL into a 50 µm by 47 cm capillary, 40 cm to the detector, results in a plug length of 0.6 cm
σdiff
2
= 12 × 10−4 ,
σT2 = σdiff
2
= 12 × 10−4 ,
σint
2
= 3.0 × 10−2 ,
σtot
2
(exp . observed) = 5 × 10−2 .
σtot
2
(calculated) = 1.2 × 10−3 + 1.2 × 10−3 + 3.0 × 10−2 + σwall
2
= 5.0 × 10−2
The injection plug is the largest contributor to the band broadening observed, but the wall
interaction has nearly equal contribution.
1.2.3.4 The Capillary

Having the discussed EOF and its dependence on the capillary wall surface, one can understand how
the capillary and its inner sufrace condition can have an impact on the efficiency and reproducibility
of CE analyses. Hence, the treatment of the capillary before, during, and following electrophoretic
separations is crucial. Microbore capillaries employed in CE can be purchased from a number of
companies in either bare silica or coated format. As shown in Figure 1.3, the standard fused-silica
capillary typically has an internal diameter of 50–100 µm, although a range of internal and external
diameters are commercially available (<1µm–>1 mm i.d.; 60 µm–>1 mm o.d.) in shapes that are
commonly cylindrical, but also rectangular. The following sections describe some pertinent points
regarding capillary preparation, conditioning before use, maintenance, and storage respectively.
1.2.3.4.1 Preparation for CE

For use in CE, the capillary is cut to the appropriate length with a ceramic knife, or an ampoule
file, so that both ends are square and flat. This is of particular importance with the inlet end of
the capillary where sample is to be introduced. Closer to the outlet, a “window” is created through
removal of the polyimide coating so that online detection is possible. The window should ideally be
approximately 0.3 cm in length (no longer than 1.0 cm) and can easily be made by burning off the
polyimide with a flame and wiping the surface with an ethanol-soaked lens tissue. Another, more
labor-intensive method is to scrape off the polyimide surface, being careful not to scratch the silica,
or break the capillary. Several mechanical devices have been devised to accomplish this task.65,66
Commercial instruments are available for carrying this out (e.g., Capital Analytical, UK) Caution
must be taken when handling the capillary once the window is created since the window area is
extremely fragile. It is important to note that with internally coated capillaries, detector windows
cannot be created by burning off the polyimide since the heat required will damage the internal
coating. Alternatively, dropwise addition of 103◦ C sulfuric acid33 or hot concentrated KOH67 from
a burette or glass pipette have been reported to remove the polyimide coating to create a window
without damaging the internal coating. We use an adaptation of the hot sulfuric acid method. A drop
of concentrated sulfuric acid is placed on the capillary where the window is desired, and the tip of a
hot soldering iron is briefly touched to the drop. This readily removes the polyimide coating without
overheating the interior surface, and prevent the necessity of heating a larger volume of acid than
would otherwise be needed. The excess acid must be washed off the exterior surface, especially if
the capillary is to be placed into a cartridge where the acid would damage the cartridge housing.
When the capillary is installed, it is useful to determine the “capillary fill time.” This information
will be useful for determining the length of time for each rinse step (NaOH, separation buffer or
rinse solution), as well as for diagnosing potential capillary problem, such as partial or complete
obstruction. One approach for accomplishing this (with ultraviolet, UV detection) is to pressure rinse
the capillary with water, zero the detector, and follow with a rise with 100 mM, NaOH. The time
required for a maximum change in absorbance to occur is the “fill time” to the detector. Factoring in
the length between the detector and the capillary outlet yields the “capillary fill time.”
Before using new bare fused-silica capillary for analysis, the capillary should be “preconditioned”
by rinsing with 5–10 column volumes of 100 mM NaOH followed by 5–10 column volumes of
water, before rinsing with 3–5 column volumes of separation buffer. If the capillary is coated,
preconditioning should be performed as per the protocols recommended by the manufacturer or as
deemed necessary by the surface coating.
1.2.3.4.2 Conditioning of the Capillary

When using a newly installed capillary or changing to a new separation buffer, the capillary must be
adequately equilibrated with the separation buffer, a process termed “conditioning.” Equilibration is
particularly important when a phosphate-containing buffer is involved. For acceptable reproducibil-
ity, a phosphate-containing buffer should be equilibrated in the capillary for a minimum of 4 h before
use.68 This process can be slightly accelerated by applying a separation-scale voltage to the system
during the equilibration period. When preparing to use a new separation buffer in a capillary that
has been equilibrated with a phosphate-containing buffer, extensive washing and equilibration of
the capillary is typically required. The capillary should be equilibrated with the desired separation
buffer for at least several hours before use and, if possible, conditioned overnight. It is for this reason
that “dedication” of capillaries to individual buffer systems may be highly recommended. In this
manner, no pre-equilibration time may be required, since the capillary can be stored in the buffer to
which it is dedicated. However, users often report reproducibility problem that are suspect of surface
chemistry deviation, particularly in the buffer ph range of 4–7. Watzig et al.69 describe preliminary
results using x-ray photoelectron spectroscopy to probe the Si-C landscape of fused-silica surfaces.
This led the authors to suggest that, in order to yield capillaries that will provide stable migration
times, especially in the pH range 4–7, preconditioned for longer than 1 h is required.
1.2.3.4.3 Regeneration of the Capillary Surface

Capillary maintenance plays a critical role in attaining reproducible results with CE. As with any
untreated silica surface, ionized silanol groups are ideal for interaction with charged analytes, particu-
larly peptides and proteins in neutral/basic pH buffers. Hence, following each separation, the capillary
surface must be “regenerated” or “reconditioned”, that is, cleansed of any wall-adsorpted material.
This is accomplished by following each run with a 3–5 column-volume rinse with 100 mM NaOH,
followed by flushing with 5–8 column volumes of fresh separation buffer. This, of course, should be
optimized for the particular buffer system used. However, the users should heed the results of Watzig
and coworkers.69 Alternative solutions for cleaning the capillary are 100 mM sodium tetraborate,
pH 11, or mM trisodium phosphate, particularly for use with phosphate containing buffers.
When using acidic buffers, rinsing with NaOH may be a disadvantage since the drastic changes
in pH may induce the requirement for extensive rinsing with the separation buffer for adequate pH
re-equilibration of the capillary. Under these conditions, it may be advisable to follow the NaOH
trines with a brief rinse with a concentrated separation buffer (10X or more), followed by the normal
rinse with the IX separation buffer. Alternatively, depending on the sample, the 10X separation
buffer rinse alone may be adequate for regenerating the capillary and, hence, the NaOH rinse may
be avoided completely. For peptide separations in phosphate buffer, pH 2.5, 1.0 M phosphoric acid
solutions may be used for capillary regeneration. Another good protein cleaning solutions is 1 M
HNO3 .
Occasionally, after extensive use with protein solutions, the capillary will become fouled and
will need to be cleaned to remove strongly adsorpted material. This is usually noticed with migration
times that are excessive, with peaks distinctly asymmetrical with extensive tailing, or filling times
become noticeably longer. A rinse with sodium ethoxide solution (blended from equal volumes of 1
M NaOH and 95% ethanol) for 5 min, followed by extensive (10–20 column volumes) water rinses
usually suffices. A partially plugged capillary may be unplugged by filling with sodium ethoxide and
allowing the solution to stand for 5 min before flushing with fresh solution, and following with an
extensive water rinse before buffer equilibration. It should be noted that the sodium ethoxide will
etch a new surface by removing the surface layer of silica, eventually making the capillary brittle.
While we have used this procedure in our laboratory, keeping some capillaries in use for up to one
year, this solution should NOT be used with coated capillaries, as it will remove the coating.
1.2.3.4.4 Storage
Because of the small dimensions of the capillary, plugging is always a potential problem. This may
occur as a result of solvent evaporation at the capillary ends, which leads to salt crystal formation.
This may be avoided in several ways. Dedicated capillaries to be stored for a short time (less than a
week) should be rinsed with 100 mM NaOH, re-equilibrated with separation buffer and stored with
the ends immersed in buffer/distilled water, or capped with silicone rubber stoppers. Optionally, the
capillary may also be stored dry using the procedure below. If the capillary is to be stored for longer
periods of time (i.e., >1 week), it should be washed with 100 mM NaOH, rinsed thoroughly with
distilled water, purged with nitrogen or air (by pressure) and stored dry. If a capillary has been used
with a detergent-/surfactant-containing solution, all traces of the additive must be removed before
storage. If this is not done, the capillary will require longer than normal re-equilibration time upon
reuse owing to the presence of the residual additive.
To place a stored capillary back into service, one should follow the same procedure outlined for
a new capillary. If the capillary has been stored dry, we have found it useful to rehydrate the lumenal
surface with a 5 min rinse of distilled water before rinsing with 100 mM NaOH and re-equilibration
with buffer.
Before storing coated capillaries, they should be thoroughly rinsed with the appropriate buffers
or solvents. For example, we have found that it is best to rinse hydrophobically coated capillaries
with one-column volume 100 mM NaOH, then extensively with water (5–10 column volumes),
and finally with methanol (5–10 column volumes), followed by drying and storage. Capillaries
used with physical gels should be rinsed, filled with fresh buffer, and stored with both ends tightly
capped. Capillary coating for particular CE applications can be of paramount importance because
of the affinity of some analytes, particularly proteins, for interaction with the wall. A number of
methodologies have been described in the literature for coating the internal surface of the capillary.70
The pH stability, as well as the effectiveness for preventing adsorption of analyte to the wall (i.e.,
separation efficiency), varies with the chemical nature of the coating. Some precoated capillaries are
commercially-available from several suppliers.
1.2.3.4.5 Basic Aspects of a Typical CE Method

As a means of collating the above information into practical form, the following provides what might
be considered to be a rudimentary CE method.
1. Mount capillary in the instrument playing close attention to alignment with the
aperature/detection optics. Once the new uncoated fused-silica capillary has been
conditioned as described above, the capillary is mounted in the instrument and properly
equilibrated.
2. Determine the capillary fill time. First fill the capillary with some low-absorbing solu-
tion (buffer or dH2 O) and then zero the detector. Using the rinse mode, fill the capillary
with a strongly absorbing solution (e.g., 0.1 M NaOH) noting the amount of time to reach
the detector. Knowing the ratio of the effective length (inlet → detector) and the total
length (inlet → outlet), calculate the time required to fill the entire capillary.
3. Conditioning the capillary as described in Section 1.2.3.4.2.
4. Make sure capillary is equilibrated. The capillary should be equilibrated preanalysis
with a minimum of a two capillary volume rinse with unelectrophoresed separation buffer,
that is, a separate vial of separation buffer that has not been electrophoresed. This assumes
that the capillary is well conditioned and equilibrated if new, or well re-equilibrated if it
is not the first run.
5. Inject sample. Sample is injected either by hydrostatic pressure or electrokinetically
depending on the nature of the analyte and the sample matrix. Ideally, the injection should
be followed by a second injection of separation buffer (equivalent of 1–2 s at 0.5 psi) to
avoid any loss of sample into the inlet vial during the first few seconds of voltage applica-
tion. This is particularly important with sample matrices that end to generate significant
Joule heat.
6. Apply voltage to the system. With most applications, it is recommended that a low
voltage is applied for a short period of time (e.g., 1 kV for 1 min) before application of
the higher separation voltage. If the desired resolution is not obtained, one may consider
examining voltage ramping.
7. Regenerate the capillary surface. The capillary should be regenerated with a mini-
mum two capillary volume rinse with regenerating solution before re-equilibration with
separation buffer. For neutral to high pH separations, 0.1–1.0 M NaOH is effective while
0.1–0.5 M phosphoric acid (or the acid equivalent of your separation buffer) is adequate
for low pH separations. This should be followed with a short (one capillary volume) rinse
with distilled water.
8. Re-equilibrate the capillary surface. The capillary should be re-equilibrated with
at least 8–10 capillary volumes of unelectrophoresed separation buffer. If this is not the
first run, this can be shortened to 5–6 capillary volumes since the first step in the method
(step 4 above) precedes injection with a 2–3 capillary volume rinse with separation buffer.
This is also an appropriate end point following the last run since the capillary remains in
separation buffer (and not 0.1 M NaOH) overnight.
It should be clear that these are basic guidelines for setting up a CE method. As emphasized
throughout this book, no method is universally applicable to all separations. Each method is likely
to require the incorporation of idiosyncratic steps specific to the mode, sample, type of capillary,
buffer, pH, etc. used.
1.3 METHOD DEVELOPMENT∗

As with HPLC, there will be no universal CE conditions that will be appropriate for the analysis of
all types of samples/analytes. In fact, it has been shown that one set of conditions is not sufficient for
the analysis of one class of proteins, for example, glycoproteins,71,72 or even variants of the same
proteins from different species.73 One of the first steps in designing a method is to determine, on
the basis of the type or “class” of analyte involved, which of the CE modes is best suited to the
sample (Table 1.2). Once the appropriate CE mode has been identified, analysis is carried out and
separation optimization initiated. Optimal separation of the components of any sample requires a
logical approach to sample solubilization and/or dealing with diverse sample matrices as well as
identification and utilization of the correct combination of CE operating parameters. Each of these
will have distinct effects on the resultant separation and resolution. There is a massive literature on
∗ Based on information given in the p/ACE 2000 Customer Training Manual. With permission.
CZE method development, not so much in the form of papers that address this per se, but more in
the papers that describe a novel separation of a previously unstudied analyte. In these it is likely
that fairly extensive details are provided on the developed method.74−80 In addition, several reviews
have been written on aspects of this (e.g., References 81 and 82).
1.3.1 STEPS IN DESIGNING A METHOD

A general strategy for CE analysis of a sample is diagrammed in Figure 1.6. The first step is an obvious
but important one: choosing the optimum wavelength for detection. This may not necessarily be the
most sensitive wavelength (i.e., 200 nm for proteins), but more importantly, one that specifically
enhances the detectability of the substance of interest and minimizes the absorbance of background
components. Choice of a buffer system is of the utmost importance. The buffer needs to be selected
with strict attention to not only the pH requirements of the sample from a stability perspective, but
also to the wavelength restrictions of the buffer (i.e., λmin for detection; see Table 1.3 in this chapter).
If the use of a specific high ionic strength (high conductivity) buffer system is required (e.g., for
preservation of bioactivity), the maximum working voltage should be determined by an Ohm’s Law
plot (discussed in Section 1.3.3.5 on buffers) to avoid buffer overheating problems. A review of
the literature will provide a general idea of the basic parameters for an initial test run. A quick trial
should establish the appropriateness of the sample introduction method, initial field strength, and
capillary length. Optimization of the separation and resolution should include modification of the
field strength, modification of the buffer by either altering pH or using additives to alter analyte
mobility or EOF, and changing the length of the capillary as necessary. If peak resolution is poor,
one might consider increasing the length of the capillary and/or the frequency of data collection.
The remainder of the steps rely largely on the experience and intuition of the operation based on
the characteristics of the sample/analyte. Once adequate separation of the sample components is
attained, sample introduction should be optimized for a maximum signal-to-noise ratio. This step
might include, where possible, altering the sample concentration or the buffer composition (ionic
strength, type, pH, etc.). Typical starting conditions for analysis of unknowns (e.g., proteins or
some drugs) in our laboratory are 50 µm i.d. × 30 or 40 cm in effective separation length capillary,
100 mM borate, pH 8.3, 25 kV, with a 3 s pressure injection at 0.5 psi. We find 200 nm to be a good
wavelength for UV detection of most organic compounds with most buffers; due to the absorbance
of conjugated bonds, there is a 5–10-fold enhancement of the signal over that observed at 280 nm
with most proteins and peptides. A reasonable starting concentration for protein samples is 0.1–1.0
mg/mL, usually in 10 mM borate, pH 8.3 or 10–20 mM Tris, pH 7.5.
1.3.2 SAMPLE PARAMETERS TO CONSIDER

Clearly, analysis of a sample requires that it be solubilized. Although it is not always possible to
know the physicochemical properties of the sample to be analyzed, it is important to be familiar
with at least some of the physical properties of the specific analyte of interest in the sample. Some
physicochemical properties may ultimately have to be determined empirically. Different questions
should be posed depending on whether the sample involved is lyophilized or already solubilized,
as in a biological fluid. In either case, it is important to have some knowledge of the detectability,
purity, and stability of the sample/component of interest. A number of questions need to be addressed.
How complex is the sample? Is the substance of interest a major or minor component? What other
substances in the sample might interfere with the detection or separation of the analyte of interest?
What is its λmax ? Is it thermally stable? If structural information is available, what are the pK values
of the ionizable groups? If the sample contains proteins or peptides, what pIs are involved? If the
sample is not in solution, there are additional questions that must be considered. Is it soluble in water
or low ionic strength buffer at mg/mL concentrations? Does it require extremes of pH for solubility
and, if so, is it stable at this pH? Is solubility enhanced by buffer additives such as urea, methanol,
24
ACTION SMALL MOLECULES PEPTIDES PROTEINS NUCLEIC ACIDS
UV - 254 or 260 nm
Do UV-Vis scan to determine nmax LIF - 488/520 nm (ex./em)
Determine Optimum Wavelength
Wavelengt
h • usually 200, 208, 214 or 260 nm Typically 200, 214 or 220 nm Typically 200, 214, 220 or 280 nm
• 260 nm if usingEtBr
• 488/514 nm wfluorescent intercalators
CZE MEKC For most peptides, low pH buffers Start with slightly alkaline buffers Use TBE or MES/Tris buffers
Select Appropriate Mode, Buffer Phosphate-pH 2-10 Phosphate/Borate in • phosphate, citrate,formate or acetate at with sievingmatrices
• 100 mM borate, pH 8.3-8.6 • HEC, HPMC, HPC, linear PA
and pH Acetate-pH 3-8 combination with pH 2.0 -4.5 • 50m M phosphate, pH 7.5
Borate-pH 6-10 surfactant-pH 7-10 • YoPro-1, TO,ToPro-1, etc.
Start with room temperature Start with roomtemperture Start with room temperature Start with 20 C-d
Select Capillary Temperature • increase if shorter migration times req’d
• increase if shorter migration times req’d • increase if shorter migration times req’d ecrease if current too high or
During Separation • decrease if current too high • decrease if current too high (>1-2 W/m) if resolution poor
• decrease if current too high
Determine Optimum Voltage

Do Ohm's Law Plot (OLP) Do Ohm's Law Plot (OLP) Do Ohm's Law Plot (OLP) Do Ohm's Law Plot (OLP)
(i.e., do Ohm's Law Plot)
A short capillary A short capillary and low voltage Amedium length capillary A short capillary and low voltage
Select Starting Parameters • start with a 20 cm (leff) x 50/75 µm cap. • 20 cm (leff) x 50 µm capillary to start • start with a 40-60 cm (leff) x 50/75 µm • 20 cm (leff) x 50 µm capillary to start
• applied voltage 25-30 kV (if OLP allows) • applied voltage of 10 kV • voltage of 25-30 kV (if OLP allows) • applied voltage of 8-10 kV
Adjust capillary length, voltage and T Adjust capillary length, voltage and T Adjust capillary length, voltage and T Adjust capillary length, and voltage
• use buffer additives • use buffer additives • use buffer additives • modify sieving agent concentration
Improve Resolution • organic solvents (MeOH,AcN) • organic solvents (MeOH,AcN) • organic solvents (MeOH,AcN) • optimize intercalator concentration by
• diamino alkanes • diamino alkanes
• ion pairing agents (HSA) • ion pairing agents (HSA) doing10X dilution series
Maximize sample solubility Use pressure injection Use electrokinetic injection

Use electrokinetic injection • keep injection <60 sec • start with 10-90 sec x 130 volts/cm
Optimize Sampling Volume • use additives such asMeOH orAcN
• start with 8-10 sec x 5-8 kV • if resolution lost -try EK injection • high salt samples -dilute with water
• try sample stacking or focusing
• try sample stacking or focusing or pre-inject water plug (1-2 sec)
FIGURE 1.6 Flow diagram for CE method development.

Handbook of Capillary and Microchip Electrophoresis and Associated Microtechniques
TABLE 1.3
Commonly Used CE Buffers and Their Associated Properties
Buffer Useful pH Range Minimum Useful λ (nm)
Phosphate 1.14–3.14 195
Formate 2.75–4.75 200
Acetate 3.76–5.76 200
Citrate 3.77–4.77 200
MES* 5.15–7.15 230
Citrate 5.40–7.40 200
PIPES* 5.80–7.80 215
Phosphate 6.20–8.20 195
HEPES* 6.55–8.55 230
Tricine* 7.15–9.15 230
Tris 7.30–9.30 220
Borate 8014–10.14 180
CAPS* 9.70–11.10 220
Phytic 1.9–9.5 200
∗ Zwitterionic buffers.
acetonitrile, hexane sulfonic acid (HSA), SDS, and so forth? It is these types of questions that will
allow for a sound methodological approach to methods development.
A final point worthy of note in this section is the importance of the sample matrix. As per the
discussion in Chapter 13 by Burgi and Giordano, the constituents of the sample matrix will play an
important role in attaining optimal separation of sample components. The conditions required for
solubilization may not provide and adequate sample matrix for introduction to analysis. Compromises
may have to be made between complete solubility and an ideal sample matrix. Moreover, it is crucial
to view the sample matrix in relation to the separation buffer. Ideally the sample matrix should be
approximately 10- to 200-fold less in total ionic strength than the separation buffer, in order that the
sample does not contribute to the EOF. This disparity will also enhance the possibility of adequate
detection, since the lower ionic strength sample matrix will lead to sample stacking and, hence,
on-capillary concentration.
Once solubility has been attained in a reasonable sample matrix, the sample should be filtered
through a 0.45 µm filter to remove unsolubilized particulates. A trial analysis is then conducted and
the efficiency of the separation optimized. This is accomplished by altering/adjusting a multitude
of variables (discussed below) in search of the combination that provide the best separation of the
nextsample components within a reasonable analysis time.
1.3.3 SEPARATION PARAMETERS TO CONSIDER

With the relatively large number of parameters to contend with in CE methods development, it is
beneficial to understand how a change in one parameter will affect others. Figure 1.7 provides a
quick reference chart for evaluation of the relationships between the variables discussed below. In
this table, the direction of the effect of altering effects on the parameters described on the vertical axis
on those in the horizontal is given (, signifies the effect is bidirectional). Parameters not covered
in Figure 1.7 include the composition of the sample matrix and the presence of buffer additives that
either interact with the sample components (e.g., surfactants) or affect the chemical nature of the
wall (through dynamic/covalent modification). It is clear that, with any particular sample, each of
these variables should be examined systematically for effective methods development.
Electrophoretic mobility
Analyte-wall interaction
Sensitivity/resolution
Affect
Electroosmotic flow
Capillary surface
these
Injection volume
System current
Analyte charge
Ionic strength
Temperature
Increasing
Viscosity
Diffusion
these
pH
pH
Temperature
Field strength
Viscosity
Molecular size/shape
Electroosmotic flow
Diffusion
Analyte-wall interaction
Current
Ionic Strength
Capillary diameter
Capillary length
Surface negative charge
Analyte charge
Electrophoretic mobility
FIGURE 1.7 A reference chart detailing the relationships between variables influencing performance in CE.
The following sections describe the qualitative effects on separation that result from alteration
of the parameters described above. For example illustrating the effects of many of these parameters
on CE analysis of mixtures, the reader is referred to an excellent review by McLaughlin et al.83
1.3.3.1 Electrode Polarity

Establishing the electrode polarity is of paramount importance in CZE and, obviously, one of the
first consideration before beginning analysis. As mentioned earlier, the normal polarity for CE is
to have the anode (+) at the inlet and cathode (−) at the outlet. In this format, EOF is toward the
cathode (detector/outlet). This is the standard polarity for most modes of CE utilizing a bare fused
silica capillary. If set in reverse polarity (cathode—inlet: anode—outlet) by mistake, the direction of
EOF is away from the detector and only negatively charged analytes with electrophoretic mobility
greater than EOF will pass the detector. This format is typically used with capillaries that are coated
with substances that reverse the net charge of the inner wall (and subsequently reverse EOF), or
when all analytes are net negatively charged (e.g., DNA or SDS–complexed proteins).
1.3.3.2 Applied Voltage

It is advisable to begin analysis with a mid-range voltage (10–20 kV). Increasing the voltage will
have a number of effects. While it will increase sample migration and EOF rate, as well as shorten
analysis time, it may increase the sharpness of the peaks and improve resolution. However, the
advantages associated with increasing the voltage may be lost if the sample matrix ionic strength
is much greater than the running buffer ionic strength such that the increased production of Joule
heat cannot be efficiently dissipated. Joule heating of the capillary results in a decreased solu-
tion viscosity. This leads to a further increase in EOF, ion mobility, and analyte diffusion, which
may ultimately result in band broadening. An excellent example of the effects of increased volt-
age on the separation of five vitamins has been provided by McLaughlin et al.83 and shown in
Figure 1.8. Decreasing the applied voltage from 30 to 10 kV has the expected effect of increas-
ing the migration time of all analytes in the mixture. At 10 kV the separation is inconveniently
long and peak 5 had not reached the detector after 33 min. However, adequate separation at higher
applied voltages is obtained at the expense of both resolution and efficiency (theoretical plates)
(Figure 1.8 inset).
1.3.3.3 Capillary Temperature

Separations should initially be attempted with the capillary thermostatted at close to ambient tem-
perature. The capillary temperature can be increased on most commercial CE units to as high as
60◦ C without substantially increasing current with most buffers. When this is done using the same
applied voltage, decreased buffer viscosity leads to an increase in analyte electrophoretic mobil-
ity, thus, decreasing separation times. Also, it is important to note that, when sample introduction
is hydrostatic (same pressure/vacuum and time), increased capillary temperature will lead to an
increase in the injected sample volume as a result of decreased buffer viscosity.2,84−89 Sensitiv-
ity may not necessarily be increased. However, Undesirable effects include concurrent changes in
buffer pH, band broadening due to increased diffusion, and possible thermal denaturation of the
sample.
TP (N; × 103)
Resolution (Rs)
4 500
N
3
2 20
1 300 Rs
5 30 kV 15
100
0 5 15 25
Absorbance (214 nm)
25 kV
20 kV
15 kV
2 3 4 5
1
10 kV
3 8 13 18 23 28 33
Time (min)
FIGURE 1.8 Effects of voltage on separation efficiency and resolution. A mixture of five vitamins
(1) niacinamide, (2) cyanocobalamine (vitamin B12 ), (3) pyridoxine (vitamin B6 ) niacin, and (5) thiamin
(vitamin B1 ). Separation conditions were as follows: capillary: 75 µm × 50 cm (effective length) fused-silica;
T : 25◦ C; buffer: 60 mM borate, 60 mM SDS, 15% methanol, pH 8.92; voltage varied as indicated to the right
of each electropherogram. (Modified from McLaughlin, G. M. et al., J. Liq. Chromatogr., 15, 961, 1992. With
permission.)
Increasing the capillary temperature can have both positive effects and negative effects on sep-
aration. In some cases, the effects of elevated capillary temperatures are not only advantageous
but necessary. A study by Guttman et al.89 showed that separation of a mixture of five proteins
in a physical gel was poorest at 20◦ C and optimized at 50◦ C. Apparently, higher temperatures
were needed to obtain the appropriate structural conformation of the physical gel required for siev-
ing of the proteins. Another positive effect of increasing capillary temperature is the substantial
decrease in analysis time (which may or may not be associated with an increase in resolution).
Figure 1.9 shows the effect of increasing temperature on the separation of 4-hydroxy-7-nitro-2,1,3-
benzoxadiazole (NBD-OH) from the derivatized valine (NBD-valine) at temperatures ranging over
20◦ C. The temperature shift from 25◦ C to 45◦ C was not only associated with the expected decrease
in analysis time (reduced by roughly half), but also with a change in the fluorescence signal inten-
sity for the peaks, lower fluorescence intensity observed at lower temperatures. Other effects can
include the negative effects of elevated capillary temperature on the stability of the sample. This
has been shown with the temperature-dependent CE analysis of α-lactalbumin87 where separa-
tions were optimal at 20◦ C and 50◦ C; however, the authors demonstrate that the faster migration
time is not solely due to increased EOF, but also to a temperature-induced conformational change
in the protein. Hence, while elevated capillary temperatures shorten analysis times, one should
be cognizant of the potential for adverse effects on analyte stability. More recently, Giordano et
al.90 defined a new approach for calculating intracapillary temperatures. In an interesting twist,
they define a potential positive outcome of Joule heating that overcomes the thermostatting capa-
bility of the instrument—the ability to exploit the elevated temperatures for fluorescent labeling
proteins.
In the first edition of this CE handbook, Nelson and Cooke91 addressed the developmental issues
associated with capillary temperature control in CE systems. This topic is revisited anew by Haddad
in Chapter 18.
(a) (b) (c) (d)
NBD-OH 100 NBD-OH NBD-OH

100 NBD-OH 100 100
Fluorescence intensity
NBD-valine
NBD-valine NBD-valine
NBD-valine
50 50 50 50
0 0 0 0
0.0 2.5 5.0 7.5 0.0 2.5 5.0 7.5 0.0 2.5 5.0 7.5 0.0 2.5 5.0
Migration time (min)
FIGURE 1.9 Effect of temperature on migration time and fluorescence intensity (RFU) with the separation
of 4-hydroxy-7-nitro-2,1,3-benzoxadiazole (NBD-OH) and the derivatized valine (NBD-valine). Temperatures
were (a) 25◦ C, (b) 30◦ C, (c) 35◦ C, and (d) 45◦ C. The conditions of the separation were15 kV applied voltage,
40/47 cm effective/separation length, 20 mM borate buffer pH 10.0, injection of 5.56 × 10−5 M valine for 4 s,
followed by absolute ethanol for 1 s, and then 5.24 × 10−2 M NBD-F for 2 s. (Modified from Zhang et al.
Anal. Bioanalyt Chem., 2006, 386, 1387–1394. With permission.)
1.3.3.4 Capillary Dimensions
1.3.3.4.1 Internal Diameter

The main advantage resulting from increasing the capillary i.d. is the enhancement in detection
sensitivity due to increasing the path length. However, a decrease in the surface–volume ratio
accompanies an increase in diameter (see Table 1.1). This may lead to less efficient dissipation of Joule
heat, which then results in a temperature gradient across the capillary and band broadening. While
a narrower capillary will accommodate a smaller sample volume (keeping Lint /Lt < 3%), owing
to the concentrating phenomenon during electrophoresis, the narrower diameter allows detection a
lower initial sample concentration.
1.3.3.4.2 Length
It is recommended that initial separations be performed on shorter capillaries (20–30 cm to detector),
which provide short analysis times for determining sample introduction mode and buffer choice, and
so forth. Once successful analysis has been achieved, the capillary length can then be increased
to improve resolution of closely migrating species since an increased on-capillary time allows for
subtle differences in analyte electrophoretic mobility to separate the components. Also, as the length
increases, there will be a concomitant decrease in the electrical field strength at constant voltage and,
hence, higher voltage may be used. There is a practical limit on capillary length, where the trade-off
with increased resolution is overcome by the decrease in sensitivity due to band broadening effects,
or the limits of the applied field. This practical limit appears to be about 100 cm.
1.3.3.5 Buffers
The choice of a buffer is critical to obtaining successful CE separation of the analytes. Once the
optimal wavelength for detection has been established, a buffer must be selected that does not
interfere with the ability to detect the analytes of interest, maintains solubility of the analytes,
maintain buffering capacity through the analysis, and produces the desired separation. These topics
are covered in more detail in the following sections, with examples highlighting the importance of
each parameter. Refer to Appendix 3 for some exemplary buffer systems and associated conditions
for separation of various analytes.
1.3.3.5.1 Selection and Preparation of a Separation Buffer

A wide variety of electrolytes can be used to prepare buffers for CE separations. When using
absorbance detection, a major requirement of any component used in the buffer system is a low UV
absorbance at the wavelength used for detection. This restriction substantially limits the choices to
a moderate number of non-UV absorbing electrolytes. However, a number of UV-absorbing organic
buffers have been used with success at low concentrations to minimize background absorbance. For
low pH buffers phosphate, acetate, formate, and citrate have commonly been used effectively. For
buffers in the basic pH range, Tris, Tricine, borate, and 3-(cyclohexylamino)propane-1-sulphonic
acid (CAPS) are acceptable electrolytes. For detection modes other than UV absorbance, a number
of other electrolytes can be utilized. In fact, for indirect detection, one might desire a buffer with
high background absorbance to enable detection limits of analytes. However, if one is considering
methods utilizing electrochemical detection, the buffer must not only be compatible with the ana-
lytes, but present a stable background conductivity against which the analytes can be detected. This
is covered by Lunte in Chapter 48 on electrochemical detection. Table 1.3 presents a list of useful
electrolytes for preparing separation buffers for CE. In addition, an indication of their useful pH
range and minimum functional wavelength (for absorbance detection) is given.
Perhaps one of the most important aspects of buffers used for CE is the requirement of pure
reagents. It is recommended that the water used for buffer preparation be of Milli-Q (or equivalent)
purified quality and that the reagents be highly purified or ultrapure (Gold Label) quality; the small
TABLE 1.4
Common Buffer Additives in CE and Their Effects
Additives Example Function
Inorganic salts NaCl, CaCl2 , K2 SO4 Modification of EOF; protein
conformational changes; protein
hydration
Organic solvents Methanol, acetonitrile, Modification of EOF; analyte
ethylene glycol solubilization; analyte salvation
Organic additives Urea Modification of EOF; protein
solubilization
Pyrenebutanoate Denaturation of oligonucleotides
dynamic modification of protein
Inorganic additives Borate Complex cis-diols; carbohydrate
or glycoprotein separations
Zwitterionic additives Z1-Methyl Reduce wall interaction; augment
EOF
Sulfonic acids hexane, heptane, octane, or Analyte ion-pairing; hydrophobic
nonane analogs interaction
Divalent amines diaminoalkanes; hexamethonium Modification of EOF; charge
bromide; decamethonium neutralization; analyte
bromide interaction
Cationic surfactants Dodecyltrimethylammonium Charge reversal on-capillary
bromide (DTAB); wall; hydrophobic interaction
cetyltrimethylammonium
bromide (CTAB);
tetradecyltrimethylammonium
chloride (TTAC)
Cellulose derivatives Hydroxyethyl cellulose; methyl Reduce EOF; provide sieving
cellulose; hydroxypropyl medium
methylcellulose
Miscellaneous polymers Polyethylene glycol Protein stability; reduce wall
interaction
Dextran Manipulate the electrophoretic
mobilities
Polyvinylpyrrolidone Enhance resolution of peptides
and amino acids
amounts required for CE do not entail great expense. Passage of buffers through an 0.22 or 0.45 µm
filter is recommended before use to remove any particulate matter. For buffers stored in 4◦ C, it is
imperative that they be brought to room temperature and thoroughly degassed before use.
Buffer additives may also be used to obtain enhanced or differential selectivity in a separation.
Several classes of additives have been identified as applicable to CE. Table 1.4 outlines some common
buffer additives and their mode of action. As can be seen from the table, the additives are multi-
functional, not only suppressing analyte–wall interactions, but also affecting analyte solubility, and
in some cases affecting selectivity. Some examples of the use of additives for enhancing/optimizing
CE separations are given later in this chapter.
1.3.3.5.1.1 pH Assuming that information about the sample is available, it is advisable to choose
a buffer pH that approximates the pK of the solute mixture. This choice is simple with pure or partially
purified preparations. With crude biological mixtures, the pKavg will typically be close to neutrality.
Proteins
pH 6.0 pH 10.0
20 25 30 35 40
pH 7.0
Peptides pH 8.61
0.02 AU
Absorbance (200 nm)
Shuffle 2,3,4
0.025 AU Native and reverse

pH 8.0
0 4 8 12
Native pH 11.64
pH 9.0 Shuffle 2,4
Reverse
Shuffle 3
20 25 30 35 40 2 4 8
Time (min)
FIGURE 1.10 Dependence of electrophoretic separation and resolution on buffer pH. Proteins: Ovalbumin
[1.0 mg/mL, in 100 mM borate, pH 8.3; 3 s pressure (0.5 psi) injection] was introduced into a pre-equilibrated
capillary containing 100 mM borate buffer at the appropriate pH, indicated above each trace. Analysis performed
on a Beckman P/ACE System 2050. Separation conditions: capillary: 50 µm × 80 cm (effective length), 87 cm
total length bare fused silica; T : 28◦ C; voltage: 25 kV; detection: 20 nm. (Modified from Landers, J. P. et al.,
Anal. Biochem., 205, 115, 1992. With permission.) Peptides: Five peptides containing the same 12 amino acids
but in varied sequence were analyzed under different pH conditions as labeled. Sequence of native, reverse, and
shuffled peptides are provided in the inset table. Separation conditions: capillary: 50 µm × 20 cm (effective
length), 27 cm total length bare fused silica; T : 20◦ C; voltage: 25 kV; detection: 20 nm. Separation buffer:
100 mM borate, plus 10 mM diaminopentane, adjusted to the appropriate pH with NaOH.
Remember that increasing the pH between 4 and 9 results in an increase in the EOF. Also remember
that the buffer pH may be altered in a secondary manner by other parameters such as temperature, ion
depletion of the buffer (caused by repetitive use of the same separation buffer) and organic additives.
Examples of the effect of buffer pH on CE separation are given in Figure 1.8 for both peptides and
proteins. With the separation of ovalbumin isoforms, the expected decrease in migration time with
increasing pH is observed and optimal separation is observed at pH 9.0. Separation of a mixture
of five peptide “isomers” over a wide range of pH is illustrated in the lower right of Figure 1.10.
At pH 2.5, all five peptides, which are identical in amino acid composition but vary in sequence,
comigrate as a broad peak (not shown). At pH 8.61, the peptides migrate as two groups while at
11.64, resolution of 4 of the 5 peptides is observed.
1.3.3.5.1.2 Ionic Species As discussed above, the choice of buffer for a particular CE separation
is important from the perspective of Joule heat generation. Obviously, the pH required for the
separation will limit the candidate buffer systems amenable for use. However, it is important to note
that, at a given pH, the type of buffer can have dramatic effects on resolution. Figure 1.11 (left)
Acetate (a) A Tricine

Dopamine
Tyrosine
Mesityl oxide Dopa

Absorbance (200 nm)
Acetate/ (b)
phosphate
0 2 4 6
B
Borate
Mesityl oxide/dopamine
Acetate/ (c)
sulfate Tyrosine
Dopa
0 5 10 15 20 0 1 2 3 4
Time (min)
FIGURE 1.11 Effect of buffer type (ionic species) on CE separation. (A) CE analysis of r-hEPO. Separation
conditions: capillary: 75 µm × 20 cm (effective length) fused silica; voltage: 10 kV; detection: 214 nm.
Buffers were of identical pH (4.0) and 100 mM concentration; (a) acetate buffer (30 µA); (b) acetate-phosphate
(120 µA); and (c) acetate–sulfate (200 µA). (Modified from Tran, A. D. et al., J. Chromatogr., 542, 469, 1991.
With permission.) (B) Separation of dopamine, tyrosine, dopa [0.5 mg/mL each, in water; 3 s pressure (0.5 psi)
injection] and mesityl oxide [neutral marker, 3 s pressure (0.5 psi) injection]. Analysis performed on a Beckman
P/ACE System 2050. Separation conditions: capillary: 50 µm × 50 cm (effective length), 57 cm (total length)
bare fused silica; T : 28◦ C; voltage: 25 kV; detection: 200 nm. (A) Separation in 100 mM tricine, pH 8.3. (B)
Separation in 100 mM borate, pH 8.3. (Modified from Landers, J. P. et al., Anal. Chem., 64, 2846, 1992. With
permission.)
illustrates the importance of buffer choice on the CE analysis of recombinant human erythropoietin
(r-hEPO).68 Although all buffers were of identical pH (4.0) and at a 100 mM concentration, resolution
of the glycoforms was only observed with acetate–phosphate buffer and not with the acetate alone
or acetate–sulfate solutions.
Other buffer systems that may have a dramatic effect on the separation are those capable of
interaction with specific analytes. The ability of borate to form stable complexes with cis-diols has
long as been known and this property exploited in chromatography.92 Several studies have shown
this property to be particularly advantageous in CE since, at a slightly basic pH, borate complexation
imparts an additional negative charge to the molecule (due to ionization of one of the free hydroxyl
groups; pKa = 9.14).93−95 Figure 1.11 (right) highlights the importance of borate for the separation
of cis-diol-containing compounds from those similar in structure, but lacking the functional group
in the separation of dopamine, tyrosine, dopa, and mesityl oxide (neutral marker).96 In tricine at
pH 8.3, dopamine is positively-charged and migrates faster than EOF (mesityl oxide) while both
tyrosine and dopa migrate slower than EOF (dopa more negatively-charged than tyrosine). In borate
buffer at the same pH, borate complexed with dopamine negates the positive charge of the amine,
while dopa becomes more negatively-charged. This example emphasizes the importance of buffer
selection for enhancing selectivity in CE separations.
1.3.3.5.2 Buffer Concentration: Ohm’s Law Plot

Useful buffer concentration is restricted by several parameters including the capillary length and
internal diameter, the applied electrical field strength and the efficiency of the capillary ther-
mostatting/cooling system. Generally, use of moderately high ionic strength buffers is desirable
for suppression of ion-exchange effects between the charged analyte ions and the ionized silanol
groups on the capillary wall. However, the current (Joule heat) associated with buffer concentrations
greater than 100 mM may overcome the capillary thermostatting capability of the system at higher
applied voltages. Excessive Joule heating can have desirable effects on both resolution and analyte
stability. Buffers that may be problematic in this respect include those containing high mobility
electrolytes such as chloride, citrate and sulfate. The excessive Joule heating associated with high
concentration buffers can be circumvented in two ways. The easiest solution is to lengthen the cap-
illary so that a tolerable current is maintained. However, this will increase the effective capillary
length and may compromise resolution. A reduction in the cross-sectional area of the capillary will
also reduce heating by decreasing the current density, and allows for more effective dissipation of
heat due to a greater surface-to-volume ratio. Alternatively, buffers that run at relatively low current
(and Joule heat) can be used. One such buffer in the pH 7–9 range is borate buffer that has been
shown to be an excellent CE buffer at concentrations as high as 500 mM.97 An added advantage of
using higher ionic strength separation buffer is the increased sample loading capacity that results
owing to the on-capillary stacking effect that will be described in a later section, and by Burgi and
Giordano in Chapter 13.
A simple method has been described by Nelson et al.,98 termed the “Ohm’s Law Plot,” that allows
for easy determination of the “functional” buffer concentration and the maximum voltage that can be
utilized with the particular buffer system (i.e., the functional limit for capillary thermostatting). After
filling the capillary with the desired buffer and allowing for adequate equilibration, voltage is applied
to the system for short intervals (e.g., 1 min) and the current recorded at each voltage. Linearity in a
plot of observed current vs. voltage is an indication that the capillary temperature is being adequately
maintained (i.e., the generated Joule heat is being effectively dissipated). At the point where linearity
is lost, the thermostatting capacity of the system has been exceeded. One should strive for heat
generation of <1 W/m (watts per meter) for optimum separation99 and should not exceed 5 W/m.
An example of the Ohm’s Law plot is given in Figure 1.12a for 100 mM concentrations of each
of three buffers: phosphate, pH 2.5; borate, pH 8.3; and CAPS, pH 11.0.100 At 20 kV, the current
is lowest with borate (≈10 µA) while CAPS (≈100 µA) and phosphate (150 µA) are dramatically
higher. The power associated with each buffer at 25 kV is 0.58, 10.07 and 5.88 W/m for borate,
phosphate and CAPS, respectively. Refer to Appendix 1 for example on calculating power (W/m).
The recommended limits have clearly been exceeded with the latter two buffers. More importantly,
linearity in the relationship is lost with the phosphate and CAPS buffer at relatively low applied
voltages in comparison with borate; the 100 mM phosphate and CAPS buffers should not be used
at voltages greater than 10 and 15 kV, respectively. In contrast, borate buffer (inset in Figure 1.12a)
exhibits a linear relationship between current and the applied voltage, even at 30 kV. Figure 1.12b
shows that this determination can be simplified if the software used for data acquisition and analysis
has the ability to directly monitor/plot current as a function of time. The figure presents a direct (on
screen) plot from System Gold Version 7.1 for 100 mM CAPS buffer where voltage is incremented
by 2.5 kV/min. This provides the same information without having to manually plot current versus
voltage.
1.3.3.5.2.1 Ionic Strength Increasing the ionic strength of the separation buffer increases the
thickness of the ionic double layer, and has the effect of decreasing EOF, hence, increasing the
analysis time. As mentioned earlier, increasing the ionic strength will also increase the current at
a constant voltage and, hence, adequate thermostatting of the capillary becomes a concern. An
advantage of increasing ionic strength, in addition to the obvious improvement in buffer capacity,
(a) (b)
20 Borate Phosphate Excessive
CAPS Joule
Borate Healing
10
150 Acceptable
0 10 20 30 Separation
Current (µA)
Voltage
200 100
100 50
0
0
0 10 20 30 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0 27.5 30.0
Applied voltage (kV)
FIGURE 1.12 The Ohm’s Law Plot. (a) Plot of observed current vs. applied voltage for each of three buffers.
The determinations were carried out on a Beckman P/ACE 2050. The voltage was incremented 2.5 kV/min,
and the current recorded through System Gold software, Version 7.1. Buffers were 100 mM phosphate, pH 2.5,
made by dilution of phosphoric acid and titration with NaOH; 100 mM borate, pH 8.3, made by titrating 25 mM
sodium tetraborate with 100 mM boric acid; and 100 mM CAPS, pH 11.0, made by titration of the appropriate
concentration dissolved in water with NaOH. The inset shows the borate data plotted on an expanded scale.
(b) Direct plot of current versus applied voltage for 100 mM CAPS, pH 11.0. A direct plot of observed current
versus applied voltage obtained through System Gold Version 7.1. Voltage is incremented by 2.5 kV/min. A
straight line drawn through the front edge of the plateau illustrates the ability of the cooling system to dissipate
the heat generated by the passage of current. The departure from linearity indicates the excessive increase in
current at the applied voltage, and is a reflection of the increase in the capillary temperature.
will be to decrease analyte–wall interactions.100 The net effect on the separation, therefore, will
be to increase resolution, provided that capillary thermosetting capability is not overcome and that
unwanted analyte dissociative processes (e.g., ligand-peptide/protein multimer dissociation) do not
occur. On the other hand, an increase in ionic strength might improve resolution in mixtures by
decreasing nonspecific analyte–analyte interaction. One example is with protein–DNA interactions.
Figure 1.13 illustrates the effect of increased ionic strength on oligonucleotides electrophoresed in
the presence of varying concentrations of urea. With urea, which maintains the oligonucleotides in
a denatured form at the thermostatted temperature, there is an obvious improvement in resolution
when increasing the urea concentration from 0 to 2 M. However, an optimum is reached and further
increase (from 2 to 4 M) results in decreased resolution and a concomitant decrease in the S/N ratio
with increasing urea. Other effects have been reported exploiting ionic strength in combination with
bovine serum albumin (BSA) to minimize nonspecific interaction with DNA restriction fragments.100
Addition of BSA to the sample obliterated resolution of the double-stranded DNA-stranded (dsDNA
fragments ranging from 72 to 310 base pairs. However, the simple addition of 50 mM NaCl to the
89 mM Tris-borate-ethylenediaminetetraacetic acid (Tris-borate-EDTA) separation buffer negated
these effects. While the voltage had to be decreased to accommodate the increased Joule heating
effects of the added ionic strength, there was a corresponding improvement in the detection of the
low molecular weight DNA fragments, most likely due to a stacking effect produced between the
lower ionic strength sample buffer and the higher ionic strength buffer.
Increasing the ionic strength is usually associated with increased Joule heating due to higher
current flux. In the above example provided by Figure 1.13, the voltage had to be decreased in the
separation that contained salt. One way to increase ionic strength without increasing Joule heating
is to use zwitterionic additives.101 Even at molar levels, zwitterionic agents, such as betaine or
16000 20
14000 (a) 0 M urea
26
Intensity (A.U.)
12000 22
10000
8000 18 28 30
24
6000 32
4000 14 16
2000
0
6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5
Time (min)
8000 (b) 2 M urea

Intensity (A.U.)
6000
4000
2000
7.5 8.0 8.5 9.0 9.5 10.0 10.5

Time (min)
7000
(c) 4 M urea
Intensity (A.U.)
6000
5000
4000
3000
2000
1000
7.0 7.5 8.0 8.5 9.0 9.5 10.0
Time (min)
FIGURE 1.13 The effect of increasing ionic strength on the CE separation of oligonucleotides. Dye was
mixed with Pluronic F127 solution and the capillary filled by syringe. 25% w/v F127 in 1XTBE buffer as a
separation medium at room temperature. Conditions: l (length to detector) = 7 cm, L (column length) = 13 cm,
E (field strength) = 200 V/cm, capillary i.d. = 50 mm, o.d. = 360 mm, Electrokinetic injection at 300 V/cm for
1 s. Peak assignments are as indicated in the figures. (Modified from Landers, J. P. et al., Biotechniques, 14, 98,
1993. With permission.)
sarcosine, do not contribute to the conductivity of the buffer, and therefore, do not raise the current
or induce Joule heating.
1.3.3.5.2.2 Organic Salts The addition of organic modifiers to the separation buffer will have
differing effects dependent on the nature of the additive. The positive effects on the separation
usually result from the alteration of the solvation properties of the buffer or the diminished EOF due
to changes in the thickness of the double-ionic layer (and hence the zeta potential) or the viscosity of
the buffer. One affect often results is a change in EOF. For example, the addition 1,4-diaminobutane to
the buffer has been proposed to enhance resolution by slowing EOF through a dynamic modification
of the capillary wall.71,72 Ion-pairing agents, such as the alkyl sulfonic acid salts, have been used with
success to resolve peptides.102 Figure 1.14 demonstrates the influence of HSA on the separation of
a series of peptides of identical amino acid composition but differing sequence. Here the resolution
appears to be due to the differential solvation and/or the microenvironment of nearest neighbor
effects on pK values of the charged amino acids in the peptides.
1.3.3.5.2.3 Organic Solvents Organic solvents such as methanol or acetonitrile have the effect
of decreasing both the conductivity of the buffer and EOF through their ability to disrupt the ordered
(a) 1,4 50 mM PO4

5
6 10 nM HSA
7
8
2Ala
(b) 1,4 50 mM PO4

5
3Ala 6 25 nM HSA
7
8
Absorbance (200 nm) 2Ala
(c) 1,4 50 mM PO4

6 50 nM HSA
3Ala
5
7 8
2Ala
(d) 1,4 50 mM PO4

100 nM HSA
3Ala
5 8
2Ala 7 6
(e) 1,4 100 mM HSA

only
3Ala
5 8
2Ala 7
6
10 15 20
Time (min)
FIGURE 1.14 Effect of organic salts on separation. The separation of eight peptides of which six peptides
contain the same 12 amino acids but a varied sequence, and two contain alanine substations for lysine residues.
Sequence of native peptide: KTNYCTKPQKSY. Separation conditions: capillary: 50 µm × 50 cm (effective
length), 57 cm total length, bare fused silica: T: 28◦ C; voltage: 15 kV; detection: 200 nm. Separation buffer: 50
mM phosphate, pH 2.0. (a) 50 mM phosphoric acid, pH 2.0, containing 10 mM HSA. (b) 50 mM phosphoric
acid, pH 2.0, containing 25 mM HSA. (c) 50 mM phosphoric acid, pH 2.0, containing 50 mM HSA. (d) 50 mM
phosphoric acid, pH 2.0, containing 100 mM HSA. (e) 100 mM HSA in distilled water, titrated to pH 2.0 with
sulfuric acid. (From Oda, R. P. et al., J. Chromatogr., 680, 341, 1994. With permission.)
structure of the water molecules. In such cases, the subsequent enhancement in resolution may result
from a combination of the decreased EOF (i.e., increased on-capillary time), decreased thermal
diffusion, and improved analyte solubility. Figure 1.15 shows the effect of acetonitrile concentration
on the solvation and resolution of the same series of peptides depicted in Figure 1.12.102 More
recently, Lookhart et al.103 showed that the addition of acetonitrile to a phosphate–glycine buffer
aided in the separation of storage proteins in maize. In another example illustrating the importance
of analyte solvation, Gordon et al.104 demonstrated that adding ethylene glycol to a borate buffer
enabled resolution of serum proteins. They found, however, that the presence of the ethylene glycol
in the sample, not in the buffer, led to greater sensitivity and reproducibility.
1.3.3.5.2.4 Modifiers of EOF Two approaches can be taken to modify EOF in an attempt
to enhance resolution: covalent coverage or dynamic modification of the capillary wall. Many
examples of both covalent105−107 and dynamic106−113 coatings are found in the literature, and
(a) 50 nM PO4
1,4,5,6
3Ala 7
8
2Ala
(b) 50 nM PO4
4
100 mM HSA
10% AcN
Absorbance (200 nm)
3Ala
5 6
7
8
2Ala 1
(c) 3Ala
4 50 nM PO4
100 mM HSA
6 20% AcN
6 8
7
2Ala 1
9 11
11 13
13 15
15 17
17 19
19
Time (min)
FIGURE 1.15 Effect of organic solvents on CE separation. The separation of the same peptides from Figure
1.12. Separation conditions: capillary: 50 µm × 50 cm (effective length), 57 cm total length, bare fused silica;
T: 28◦ C; voltage: 15 kV; detection: 200 nm. Separation buffer: 50 mM phosphate, pH 2.0. (a) Buffer without
any additive, (b) buffer containing 10% acetonitrile, and (c) buffer containing 20% acetonitrile. (From Oda, R. P.
et al., J. Chromatogr., 680, 341, 1994. With permission.)
while specific to capillary, surface coating is covered in more detail in Chapter 52 by Henry.
A study by Swedberg105 is an excellent example of the importance of capillary coating for pre-
venting protein–wall interactions during CE separations. A mixture containing seven proteins was
clearly resolved (with efficiency >500,000 theoretical plates per meter) in a capillary coated with
a pentafluoroaryl compound, whereas separation in bare fused-silica resulted in only a few broad
peaks barely greater than baseline. The EOF was not abolished by this coating, suggesting that not
all of the silanol groups were reacted, but that the bulky aryl groups prevented protein–wall interac-
tion. As would be expected when protein–wall interactions are diminished, the migration times are
markedly shortened. Figure 1.16 highlights the effects of capillary coating on the separation of DNA
fragments observed by Chu & coworkers. They describe two new methods for coating capillar-
ies with polyvinyl alcohol (PVA) on characterize the adsorption of PVA onto the wall using atomic
Coated capillary (PVAL-Control)
540 589
Relative fluorescence intensity
Coated capillary (PVAL-Si) 504

458
434
267
234
124 213
123 192
184
104
89
80
81
21 57
18 51
11
8
Coated capillary (Grignard/PVAL) 458

434 589
540
504
267
234
124 213
123 192
104 184
89
90
64
57
51
21
1118
8
6 10 14 18 22 26
FIGURE 1.16 The effect of capillary surface coating on the resolution of DNA fragments. Capillaries were
coated with polyvinyl alcohol and a pBR322 HaeIII digest used to evaluate resolution (fragment lengths from 8
to 589 bp[). Capillary: 13/10 cm total/effective length, 75 m i.d.; medium, 1 TTE solution of polyvinyl alcohol
(Mr 125 000; DH = 88 mol%; concentration = 5 wt% for a and c, 6 wt% for b); DNA: pBR322 HaeIII digest,
10 g/mL; buffer, 1 TTE, pH 8.3, including ethidium bromide at 3 g/mL; injection, 0.65 kV, 3 s; electrophoresis:
3.9 kV, 2.6 A, nominal 25C; detection, LIF using an Ar-ion laser beam at = 488 nm. (Modified from Moritani
et al. Electrophoresis, 2003, 24, 2764–2771. With permission.)
force nueroscopy. The authors describe performance for tens of runs but don’t comment on separation
reproducibility beyond that.
The other approach to minimizing analyte–wall interactions through modification of the cap-
illary wall is the use o additives that provide a dynamic (noncovalent) coating. Diaminoalkanes
are one class of buffer additives that appear to enhance the resolution of protein mixtures by coat-
ing the capillary wall,110,111 and modifying the zeta potential, although the mechanism is poorly
understood.71,72 Figure 1.17 shows the electropherograms resulting from the CE analysis of several
EOF (a) (b) Chorionic

Ovalbumin EOF
Gonadotropin
EOF
EOF
Absorbance (200 nm)
−DAB −DAB
+DAB
+DAB
5 15 25 35 45 0 10 20 30 40
(c) Pepsin (d) Chorionic

Abhydrase
−DAB
−DAB
+DAB +DAB
0 5 10 30 35 40 0 5 10 15 20
Time (min)
FIGURE 1.17 Effect of modifiers of EOF on separation. (a) The effect of 1 mM 1,4-diaminobutane (DAB)
on the resolution of ovalbumin glycoforms. Ovalbumin [1 mg/mL, in water; 3 s pressure (0.5 psi) injection].
Analysis was carried out on a Beckman P/ACE System 2050. Separation conditions: capillary: 50 µm × 80 cm
(effective length), 87 cm total length bare fused silica; T : 28◦ C; voltage: 25 kV; buffer: mM borate, pH 8.3;
detection: 200 nm. (b) The effect of 5 mM 1,3-diaminopropane on the resolution of hCG glycoforms. hCG [4
mg/mL, in water; 2 s vacuum (–127 mmHg) injection]. Analysis was carried out on a ABI 270A. Separation
conditions: capillary: 50 µm × 80 cm (effective length), 100 cm total length bare fused silica: T : 28◦ C; voltage:
25 kV; buffer: 25 mM borate, pH 8.8; detection: 200 nm. (Modified from Morbeck, D. E. et al., J. Chromatogr.,
680, 217, 1994. With permission.) (c) Partial resolution of pepsin glycoforms in the presence of 1 mM DAB.
Pepsin [0.5 mg/mL, in water; 3 s pressure (0.5 psi) injection]. Analysis was carried out on a Beckman P/ACE
System 2050. Separation conditions: capillary: 50 µm × 80 cm (effective length), 87 cm total length bare fused
silica; T : 28◦ C; voltage: 25 kV; buffer: 100 mM borate, pH 9.0; detection: 200 nm. (Modified from Landers, J. P.
et al., Anal. Biochem., 205, 115, 1992. With permission.) (d) Lack of effect of DAB on carbonic anhydrase
separation. Bovine carbonic anhydrase [1.0 mg/mL, in water; 3 s pressure (0.5 psi) injection]. Analysis was
carried out on a Beckman P/ACE System 2050. Separation conditions: capillary: 50 µm × 80 cm (effective
length), 87 cm total length bare fused silica; T : 28◦ C; voltage: 25 kV; buffer: 100 mM borate, pH 8.3; detection:
200 nm.
glycoproteins [ovalbumin, human chorionic gonadotropin (hCG), and pepsin] known to be micro-
heterogeneous with respect to glycan content, in the absence of any modifier, and in the presence
of a diaminoalkane.72,112 Ovalbumin is resolved into two groups of nine peaks on the basis of
protein phosphorylation states; the multitude of peaks observed in the presence of 1 mM diaminobu-
tane are presumed to represent the glycoforms of the protein. A similar effect has been observed by
Morbeck, et al.112 with the CE separation of urinary hCG, another glycoprotein known to be of micro-
heterogeneous character. The presence of 5 mM diaminopropane in the buffer results in a reduction in
EOF (as determined by reduced mobility of the neutral marker, DMF) and leads to resolution of eight
hCG glycoforms that cannot be completely resolved by any other method. Pepsin glycoforms were
only partially resolved in 1 mM diaminobutane. The nonglycosylated protein carbonic anhydrase

indicates no microheterogeneity in the presence of 1 mM diaminobutane.72
A second method of dynamic coating involves the use of materials to alter the viscosity of the
solution at the capillary wall. Hjertén3 suggest that EOF is inversely dependent on the viscosity of
the solution in the double layer. If the viscosity approached infinity, the EOF will fall to zero. Thus
polymers that adsorb to the wall should reduce or eliminate EOF, and prevent protein adsorption by
blocking the access to the reactive silanes. Various cellulose derivatives33,42−45 and polyhydroxy
compounds such as polyvinyl alcohols113,114 have been used with varying degrees of success for the
separation of peptides and proteins.
1.3.3.5.3 Buffer Depletion

Since the electrophoretic process occurring within a buffer-filled capillary under an applied elec-
trical field involves the movement of ions, the separation buffer can eventually be “depleted” if
electrophoresis continues for extensive periods of time. This is not surprising in light of the fact that
slab gels are used for electrophoresis only once and buffers up to two or three times. Use of “depleted”
buffer for CE analysis can result in deterioration of resolution and poor run-to-run reproducibility.116
This phenomenon is not isolated to free solution buffer systems, but also problematic in gel-based
systems. For this reason buffers should be replenished after a limited number of runs (inlet and outlet
every 5–8 runs; 10–15 min analysis; 4–5 mL buffer reservoirs). One method that may be useful
for extending the use of a buffer without adverse effects on reproducibility is a “limited replenish-
ment” system in which two inlet buffer reservoirs are used—one for electrophoresis and the other
for rinsing the capillary with buffer that has not undergone any changes due to electrophoresis. This
“limited replenishment” approach appears to substantially increase the number of runs (with the same
electrophoresis buffers) before the onset of buffer-associated reproducibility problems.117 Trushina
et al.118 have provided one example of pushing CE to the limit with respect to reproducibility under
extreme buffer depletion conditions. They used a phosphate buffer containing 10 mM HCl and 250
mM KCl to separate glutathione, nitrosylated glutathione, and glutathione disulfide solubilized in a
2 M HCl sample matrix. Using 11 kV for separation, the system was pushed to the maximum current
limit tolerated by the CE (249 µA) exceeding a total power of 7 W/m. Despite the dramatic change
in pH observed in the inlet and outlet reservoirs over several runs, reproducible separations of the
thiols (injection-to-injection) could be obtained provided that buffer depletion issues were addressed
with buffer replenishment after every run.
1.4 INTRODUCTION OF SAMPLE INTO THE CAPILLARY

As a result of the small dimensions of capillaries used in CE, the total capillary volume is typically
in the microliter range (see Table 1.1). Adhering to the chromatography “rule of thumb” restricting
sample volume to 1–5% of the total capillary volume, sample volumes must be in the low nanoliter
range if overloading is to be avoided.119 The technology for small sample volume introduction into a
capillary has converged at three accepted methods. Introduction, or as is sometimes referred, “injec-
tion” or “loading” of sample into the capillary can be accomplished hydrostatically, by siphoning,
or electrokinetically. With hydrostatic injection, sample is introduced by immersing the capillary
inlet into a vial containing the sample and either pressurizing the inlet vial containing the sample or
applying a vacuum to the outlet vial (opposite end). Sample introduction by gravity, which is more
commonly used with noncommercial systems, relies on the siphoning of sample into the capillary
by elevating the injection (inlet) end of the capillary relative to the outlet end. With electrokinetic
loading, the injection (inlet) end of the capillary is immersed in the sample, the outlet in the separa-
tion buffer and a low voltage (1–10 kV) applied for durations of 1–99 s, depending on the capillary
length and i.d.
With electrokinetic introduction, the quantity of sample introduced into the capillary is depen-
dent on a number of parameters, the most important of which are the electrophoretic mobilities of
the sample components and the EOF. Other parameters that can effect this mode of loading, there-
fore, may not be a quantitative representation of the sample components. In contrast, electrokinetic
loading may be advantageous if the analyte of interest is a small percentage of the sample, but has a
much higher electrophoretic mobility than other constituents. Under these conditions, electrokinetic
loading provides a positive sample loading bias and may enhance the detectability of the compo-
nent of interest. This is containly the case with smaller molecules weight DNA fragments. Another
situation where electrokinetic introduction is advantageous is where the capillary contains a poly-
merized or cross-linked matrix, where pressure injection cannot be used conveniently. Examples
of samples advantageously electrokinetically loaded are DNA and SDS–protein complexes. It is
important to bear in mid that sample matrices containing significant concentrations of electrolyte are
not efficiently electrokinetcally introduced.
This last point highlights a recurring theme in many of the chapters to follow, that is, the sample
matrix plays a critical role in obtaining efficient separations. The problems associated with high-salt
sample matrices and approaches to manipulating the matrix for adequate CE analysis, are addressed
by Shihabi in Chapter 26 and by Burgi and Gordano in chapter 13. In contrast, under the appropriate
conditions, the sample matrix can have a positive effect on a sample introduction, enhancing the
detectability of the components of a dilute sample by on-capillary preconcentration. This is particu-
larly important in light of the miniscule sample volume capacity inherent with CE and is described
in the next section.
1.5 ON-CAPILLARY SAMPLE CONCENTRATION TECHNIQUES

One of the main drawbacks of UV detection CE is the low sensitivity resulting from the inherently
small dimensions of the flow cell (i.e., the inner diameter of the capillary) and the sample volume
capacity. As mentioned above, introduction of sample volumes larger than 1%–5% of the total
capillary volume can be detrimental to resolution. Therefore, while CE is a good mass detector, it is
a relatively poor concentration detector. This limits the use of CE as an analytical technique to samples
having nominally high concentrations (10 µg/mL or greater). For this reason, the development of
several approaches for “on-capillary sample concentration” has been pivotal to the acceptance of
CE as a microanalytical technique. Since an entire chapter is dedicated to this subject (Chapter 13
by Burgi and Giordano), these techniques are discussed here in only the briefest of detail.
1.5.1 SAMPLE STACKING

One of the simplest methods for sample preconcentration is to induce “stacking” of the sample
components of this is easily accomplished by exploiting the ionic strength differences between the
sample matrix and separation buffer.120,121 Stacking results from the fact that sample ions have an
enhanced electrophoretic mobility in a lower conductivity environment. When voltage is applied
to the system, ions in the sample plug instantaneously accelerate toward the adjacent separation
buffer zone. On crossing the boundary, the higher conductivity environment induces a decrease in
electrophoretic velocity and subsequent “stacking” of the sample components into a smaller buffer
zone than the original sample plug. Within a short time, the ionic strength gradient dissipates and
the charged analyte molecules begin to move from the “stacked” sample zone toward the cathode.
Stacking can be utilized with either hydrostatic or electrokinetic injection and can typically yield
a 10-fold enhancement in a sample concentration and, thence, sensitivity. An example is given in
Figure 1.18, which shows separation of the standard peptide mixture in 50 mM phosphate separation
buffer, pH 2.5. There is an obvious enhancement in detectability when the sample matrix is 5 mM
phosphate in comparison with 50 mM separation buffer. The undesirable effects of high ionic strength
50 nM 5 nM 100 nM
4
Absorbance (200 nm)
0.001 AU 0.001 AU 0.001 AU

1
3
2
6
5 1 5
12 4 2 4
3 6 3 6
2 4 6 8 10 2 4 6 8 10 2 4 6 8 10
Time (min)
FIGURE 1.18 Sample “stacking.” Stacking of peptides in phosphate buffer. Separation of a standard peptide
mixture (peptide calibration kit, Bio-Rad, Richmond, CA). Peptides were (1) bradykinin, (2) angiotensin, (3)
a-melanin stimulating hormone, (4) thyrotropin-releasing hormone, (5) leutinizing hormone releasing hormone,
(6) leucine enkephalin, (7) bombesin, (8) methionine enkephalin, and (9) oxytocin, 50 µg/mL each. (a) Sample
was dissolved in 50 mM phosphate buffer, pH 2.5. (b) Sample diluted in 5 mM phosphate buffer, pH 2.5. (c)
Sample dissolved in 100 mM phosphate buffer, pH 2.5. Analyses were performed on a Beckman P/ACE System
Model 2100. Separation conditions were as follows: capillary: 50 µm × 20 cm (effective length), 27 cm total
length, bare fused silica; T: 20◦ C; voltage: 10 kV; separation buffer: 50 mM phosphate, pH 2.5; 5 s hydrostatic
injection; detection: 200 nm.
(100 mM phosphate buffer) sample matrix are clearly illustrated in the right panel. Detailed discussion
of sample matrix effects on separation can be found in Chapter 13 and Chapter 26.
When possible, dissolving the sample in dilute separation buffer may be more advisable than
in water since the dramatic differences in EOF between the sample plug and separation buffer may
cause laminar flow within the capillary and, hence, lead to peak broadening.63,84,85,122 Moreover,
excessive heat produced in the sample plug may denature sample components.84,85,88,89 Therefore,
it may be advisable to “stack” at lower applied voltages (e.g., 1-2 KV) or “ramp” to the separation
voltage over several minutes.
1.5.2 SAMPLE FOCUSING

An alternative approach to sample preconcentration by stacking in on-capillary “focusing” and is
based on pH differences between the sample plug and separation buffer. This has been shown to be
very useful for the analysis of peptides, mainly because of their relative stability over a wide pH
range.119 Focusing is easily accomplished by increasing the pH of the sample above that of the net
pI (isoelectric point) of the sample. The high pH sample plug is flanked between low pH separation
buffer zones (i.e., an equivalent volume of low pH separation buffer following introduction of the
sample plug) and, on applying a voltage, the negatively charged peptides in the initial sample zone
migrate toward the anode. Upon entering the lower pH separation buffer, a pH-induced changed
in their charge state causes a reversal in the direction of mobility resulting in a “focusing” on the
peptides at the interface of the sample (high pH) and low pH buffer plugs. After the pH gradient
dissipates, the peptides, again positively charged, migrate toward the cathode as a sharp zone. This
approach, limited to samples that can withstand the inherent changes in pH without substantial
denaturation, may yield as much as a fivefold enhancement in sample concentration and, hence sen-
sitivity. The practical and some theoretical aspects of these techniques are addressed in Chapter 26 by
Shihabi and Chapter 13 by Burgi and Gierdano.
1.5.3 ISOTACHOPORETIC SAMPLE ENRICHMENT

Isotachoporetic (ITP) is a specialized technique for sample application first described for sample
concentration by Foret et al.123 This technique has been covered in detail by Burgi and Chein124 in a
previous edition of this handbook. However, more recently Williams and colleagues125 utilized this
in a very clever way to exploit the original on-capillary sample concentration123 to enhance injection
of DNA into microchannels for higher sensitivity fluorescence detection.
1.5.4 ONLINE CONCENTRATION

The development of capillaries with a small “column” or plug of adsorptive resin at the inlet side of
the capillary has attracted much attention owing to its ability to allow the many-fold concentration
of dilute samples from large volumes. It is the simplicity of the solid-phase extraction (SPE) concept
that makes it attractive—load sample by flow through the phase where the analyte of interest is
adsorpted onto the resin, wash the resin to remove loosely bound or unwanted components, and
then desorb the analytes from the resin and separate by electrophoresis. The notion of online sample
concentrate before separation was pioneered by Guzman126,127 in the early 1990s coincident with
the effort of Fuchs et al. 128 A number of groups honed in on the potential utility of this approach for
solving sensitivity problems rooted in low concentration analytes in large volume biological sam-
ples, and shown the applicability of this method to the analysis of drugs,129,130 peptides,131,132 and
proteins.133,134 More recently, this has been applied to DNA extraction135−139 and immunoaffinity
capture.140−142 This technique is discussed in detail in a various chapters in this edition of the CE
handbook including Chapter 43 by Bienvenue.
1.6 CONCLUDING REMARKS

Capillary zone electrophoresis (CZE) is a powerful technique, magnetic in its analytical personality
as a result of the simple way diverse analytes can be resolved rapidly and with high efficiency.
The attraction is easy to understand—an electrophoretic technique with as much bandwidth as (and
complementary to) HPLC and multiple modes of separation available by simply changing the buffer
system. Yet within the simple instrumental framework that is, at its root, a power supply, a cap-
illary and a detector, lies the capability to analyze drugs, peptides, carbohydrates, and proteins in
sample matrices as simple as buffer or as complex as serum. That power is the magnet that draws
people in.
In this chapter, I hope to have imparted to the reader a taste of CE, and some practical suggestions
that will be useful in developing a method for the analysis of a specific compound(s). By using
the basic principles associated with electrophoretic separation as discussed here, and applying a
systematic approach to assess the factors that influence the resolution of the analyte(s) under study,
it is likely a successful analytical method can be developed for just about any analyte of interest.
In many of the core chapters that follow this chapter, specific methods are covered in detail with
an analyte-specific focus—however, method development of some kind will be found in the vast
majority one of the 55 chapters that constitute this handbook.
REFERENCES
1. Michaelis, L., Electric transport of enzymes, malt distaste and pepsin, Biochemische Zeitschrift, v 17,
1909, pp. 231–234.
2. Tiselius, A., A new apparatus for electrophoretic analysis of colloidal mixtures, Trans. Faraday Soc.,
33, 524, 1937.
3. Hjertén, S., High performance electrophoresis, Chromatogr. Rev., 9, 122, 1967.
4. Rilbe, H., Some reminiscences of the history of electrophoresis, Electrophoresis, 1995, 16, 1354–1359.
5. Consden, R., Gordon, A. H., and Martin, A. J. P., Quantitative analysis of proteins: partition
chromatographic method using paper, Biochem. J., 38, 224, 1944.
6. Wieland, T. and Fischer, E., Electrophoresis onto filter paper, Naturwissenschaften, 35, 29, 1948.
7. Smithies, O., Zone electrophoresis in starch gels: group variations in the serum proteins of normal
human adults, Biochem. J., 61, 629, 1955.
8. Graber, P. and Williams, C. A., Methode permittant l’etude conjugee des proprietes electrophoretiques
et immunichimiques d’un mélanges de proteins: application au serum sanguine, Biochem. Biophys.
Acta, 10, 193, 1953.
9. Hjerten, S., Agarose as an anticonvection agent in zone electrophoresis, Biochem. Biophys. Acta, 53,
514, 1961.
10. Kolin, A., Separation and concentration of proteins in a pH field combined with an electric field,
J. Chem. Phys., 22, 1628, 1954.
11. Vesterberg, O., Synthesis and isoelectric fractionation of carrier ampholytes, Acta Chem. Scand., 23,
2653, 1969.
12. Raymond, S. and Weintraub, L., Acrylamide gel as a supporting medium for zone electrophoresis,
Science, 130, 711, 1959.
13. Ornstein, L., Disc electrophoresis. I. Background and theory, Ann. NY Acad. Sci., 121, 321–349,
1964.
14. Davis, B. J., Disc electrophoresis. II. Method and application to human serum proteins, Ann NY Acad
Sci., 121, 404–427, 1964.
15. Laemmli, U. K., Cleavage of structural proteins during the assembly of the head of bacteriophage T4,
Nature, 227, 680–685, 1970.
16. Dale, G. and Latner, A. L., Isoelectric focusing of serum proteins in a acrylamide gels followed by
electrophoresis, Clin. Chem. Acta, 24, 61, 1969.
17. Macko, V. and Stegemann, H., Mapping of potato proteins by combined electro-focusing and
electrophoresis. Identification of varieties, Hoppe–Seyler’s Z. Physiol. Chem., 350, 917, 1969.
18. Stegemann, H., Protein mapping in polyacrylamide and its application to genetic analysis in plants,
Angew. Chem. (Internat. Ed.), 9, 643, 1970.
19. O’Farrell, P. H., High resolution two-dimensional electrophoresis of proteins, J. Biol. Chem., 250,
4007, 1975.
20. Virtanen, R., Zone electrophoresis in a narrow-bore tube employing potentionmetric detection, Acta
Polytechnica Scand., 123, 1, 1974.
21. Jorgenson, J. W. and Lukacs, K. D., Zone electrophoresis in open tubular glass capillaries, Anal.
Chem., 53, 1298, 1981.
22. Ishii, J., A study of high performance liquid chromatography. I. Development of technique for
miniaturization of high performance liquid chromatography, J. Chromatogr., 144, 157, 1978.
23. Hirata, Y. and Novotny, M., Techniques of capillary liquid chromatography, J. Chromatogr., 186, 521,
1979.
24. Kucera, P. and Guiochon, G., Use of open-tubular columns in liquid chromatography, J. Chromatogr.,
283, 1, 1984.
25. Schomburg, G., Problems and achievements in the instrumentation and column technology for
chromatography and capillary electrophoresis, Chromatographia, 30, 500, 1990.
26. Mikkers, F. E. P., Everaerts, F. M., and Verheggen, T. P. E. M., High performance zone electrophoresis,
J. Chromatogr., 169, 11, 1979.
27. Mazzeo, J. R. and Krull, I. S., Coated capillaries and additives for the separation of proteins by capillary
zone electrophoresis and capillary isoelectric focusing, Biotechniques, 10, 638, 1991.
28. Wehr, T., Recent advances in capillary electrophoresis columns, LC-GC Mag., 11, 14, 1993.
29. Cohen, A. S. and Karger, B. L., High performance sodium dodecyl sulfate polyacrylamide gel capillary
electrophoresis of peptides and proteins, J. Chromatogr., 397, 409, 1987.
30. Cohen, A. S., Paulus, A., and Karger, B. L., High performance capillary electrophoresis using open
tubes and gels, Chromatographia, 24, 15, 1987.
31. Cohen, A. S., Najarian, D. R., Paulus, A., Guttman, A., Smith, J. A., and Karger, B. L., Rapid
separation and purification of oligonucleotides by high-performance capillary gel electrophoresis,
Proc. Natl. Acad. Sci. USA, 85, 9660, 1988.
32. Drossman, H., Luckey, J. A., Kostichka, A., D’Cunha, J., and Smith, L. M., High-speed separation of
DNA sequencing reactions by capillary electrophoresis, Anal. Chem., 62, 900, 1990.
33. Bocek, P. and Chrambach, A., Capillary electrophoresis of DNA in agarose solutions at 40◦ C,
Electrophoresis, 12, 1059, 1991.
34. Mansfield, E., Ross, E. E., and Aspinwall, C. A., Preparation and characterization of
cross-linked phospholipid bilayer capillary coatings for protein separations, Anal. Chem., 79,
3135–3141, 2007.
35. Kuldvee, R., Wiedmer, S. K., Oorni, K., and Riekkola, M. L., Human low-density lipoprotein-coated
capillaries in electrochromatography, Anal. Chem., 15, 77, 3401–3405, 2005.
36. Linden, M. V., Holopainen, J. M., Laukkanen, A., Riekkola, M. L., and Wiedmer, S. K., Cholesterol-
rich membrane coatings for interaction studies in capillary electrophoresis: application to red blood
cell lipid extracts, Electrophoresis, 27, 3988–3998, 2006.
37. Danger, G., Ramonda, M., and Cottet, H., Control of the EOF in CE using polyelectrolytes of different
charge densities, Electrophoresis, 16, 28, 925–931, 2007.
38. Terabe, S., Otsuka, K., Ichikawa, K., Tsuchiya, A., and Ando, T., Electrokinetic separations with
micellar solution and open-tubular capillaries, Anal. Chem., 56, 111, 1984.
39. Terabe, S., Otsuka, K., and Ando, T., Electrokinetic chromatography with miceller solution and
open-tubular capillary, Anal. Chem., 57, 834, 1985.
40. Hjerten, S. and Zhu, M., Adaptation of the equipment for high-performance electrophoresis to
isoelectric focusing, J. Chromatogr., 346, 265, 1985.
41. Hjerten, S., Liao, J. L., and Yao, K., Theoretical and experimental study of high-performance
electrophoretic mobilization of isoelectrically focused protein zones, J. Chromatogr., 387, 127,
1987.
42. Hjerten, S., Elenbring, K., Kilar, F., Liao, J. L., Chen, A. J. C., Siebert, C. J., and Zhu, M., Carrier-
free zone electrophoresis, displacement electrophoresis and isoelectric focusing in a high-performance
electrophoresis apparatus, J. Chromatogr., 403, 47, 1987.
43. Zhu, M., Hansen, D. L., Burd, S., and Gannon, F., Factors affecting free zone electrophoresis and
isoelectric focusing in capillary electrophoresis, J. Chromatogr., 480, 311, 1989.
44. Bruin, G. J. M., Chang, J. P., Kuhlman, R. H., Zegers, K., Kraak, J. C., and Poppe, H., Capillary zone
electrophoretic separation of proteins in polyethylene glycol-modified capillaries, J. Chromatogr., 471,
429, 1989.
45. Bruin, G. J. M., Huisden, R., Kraak, J. C., and Poppe, H., Performance of carbohydrate-modified
fused-silica capillaries for the separation of proteins by zone electrophoresis, J. Chromatogr., 480,
339, 1989.
46. Ganzler, K., Greve, K. S., Cohen, A. S., Karger, B. L., Guttman, A., and Cooke, N. C., High perfor-
mance capillary electrophoresis of SDS–protein complexes using UV-transparent polymer networks,
Anal. Chem., 64, 2665, 1992.
47. Miksik, I., Sedlakova, P., Mikulikova, K., Eckhardt, A., Cserhati, T., and Horvath, T., Matrices
for capillary gel electrophoresis—a brief overview of uncommon gels, Biomed. Chromatogr., 20,
458–465, 2006.
48. Rech, I., Cova, S., Restelli, A., Ghioni, M., Chiari, M., and Cretich, M., Microchips and
single-photon avalanche diodes for DNA separation with high sensitivity, Electrophoresis, 27,
3797–3804, 2006.
49. Gras, R., Luong, J., Monagle, M., and Winniford, B., Gas chromatographic applications with the
dielectric barrier discharge detector, J. Chromatogr. Sci., 44, 101–107, 2006.
50. Castano-Alvarez, M., Fernandez-Abedul, M. T., and Costa-Garcia, A., Amperometric detector designs
for capillary electrophoresis microchips, J. Chromatogr. A, 24, 1109, 291–299, 2006.
51. Cappiello, A., Famiglini, G., Palma, P., and Siviero, A., Liquid chromatography-electron ionization
mass spectrometry: fields of application and evaluation of the performance of a Direct–EI interface,
Mass Spectrom. Rev., 24, 978–989, 2005.
52. Rocheleau, M. J. and Dovichi, N. J., Separation of DNA sequencing fragments at 53 bases per minute
by capillary gel electrophoresis, J. Microcol. Sep., 4, 449, 1992.
53. Rickard, E. C., Strohl, M. M., and Nielsen, R. G., Correlation of electrophoretic mobilities from
capillary electrophoresis with physicochemical properties of proteins and peptides, Anal. Biochem.,
197, 197, 1991.
54. Helmholtz, H. Z., About electrical interfaces (translated title), Ann. Phys. Chem., 7, 337, 1879.
55. Tian, H., Hühmer, A. F. R., and Landers, J. P., Evaluation of silica resins for the direct and efficient
extraction of DNA from complex biological matrices in a miniaturized format, Anal. Biochem., 283,
175–191, 2000.
56. Sazelova, P., Kasicka, V., Koval, D., Prusik, Z., Fanali, S., Aturki, Z., Control of EOF in CE by
different ways of application of radial electric field. Electrophoresis, 28, 756–766, 2007.
57. Weber, R., Concerning theories of electrophoretic separations in porous tubes, Helv. Chim. Acta., 36,
424, 1953.
58. Snyder, L. R. and Kirkland, J. J., Introduction to Modern Chromatography, 2nd ed., Wiley Interscience,
New York, 1979.
59. Giddings, C. J., Generation of variance, “theoretical plates”, resolution, and peak capacity in
electrophoresis and sedimentation, Sep. Sci., 4, 181, 1969.
60. Sternberg, J. C., Extracolumn contribution to chromatographic band broadening, Adv. Chromatogr., 2,
206, 1996.
61. Huang, X., Coleman, W. F., and Zare, R. N., Analysis of factors causing peak broadening in capillary
zone electrophoresis, J. Chromatogr., 480, 95, 1989.
62. Jones, H. K., Nguyen, N. T., and Smith, R. D., Variance contributions to band spreading in capillary
zone electrophoresis, J. Chromatogr., 504, 1, 1990.
63. Grushka, E., McCormick, R. M., and Kirkland, J. J., Effect of temperature gradients on the efficiency
of capillary zone electrophoresis separations, Anal. Chem., 61, 241, 1989.
64. McManigill, D. and Swedberg, S. A., Factors affecting plate height in high performance zonal cap-
illary electrophoresis (HPZCE), in Techniques in Protein Chemistry, Hugli, T., Ed., Academic Press,
San Diego, CA, p. 468, 1989.
65. Lux, J. A., Hausig, U., and Schomburg, G., Production of windows in fused silica capillaries for
in-column detection of UV-absorption or fluorescence in capillary electrophoresis or HPLC, J. High
Resol. Chromatogr., 13, 373, 1990.
66. McCormick, R. M. and Zagursky, R. J., Polymide stripping device for producing detection windows
on fused-silica tubing used in capillary electrophoresis, Anal. Chem., 63, 750, 1991.
67. Bruno, A. E., Gassmann, E., Pericles, N., and Anton, K., On-column capillary flow cell utilizing
optical waveguides for chromatographic application, Anal. Chem., 61, 876, 1989.
68. Tran, A. D., Park, S., Lisi, P. J., Huynh, O. T., Ryall, R. R., and Lane, P. A., Separation of
carbohydrate-mediated microheterogeneity of recombinant human erythropoetin by free solution
capillary electrophoresis, J. Chromatogr., 542, 459, 1991.
69. Kaupp, S., Bubert, H., Baur, L., Nelson, G., Watzig, H., Unexpected surface chemistry in capillaries
for electrophoresis, J. Chromatogr. A, 894, 73–77, 2000.
70. Dolnik, V. Wall coating for capillary electrophoresis on microchips, Electrophoresis, 25, 3589–3601,
2004.
71. Taverna, M., Baillet, A., Biou, D., Schluter, M., Werner, R., and Ferrier, D., Analysis of carbohydrate-
mediated heterogeneity and characterization of N-linked oligosaccharides of glycoproteins by high
performance capillary electrophoresis, Electrophoresis, 13, 359, 1992.
72. Landers, J. P., Oda, R. P., Madden, B. J., and Spelsberg, T. C., High-performance capillary electrophore-
sis of glycoproteins: the use of modifiers of electroosmotic flow for analysis of microheterogeneity,
Anal. Biochem., 205, 115, 1992.
73. Oda, R. P. and Landers, J. P., Effect of cationic buffer additives on capillary electrophoretic separation
of serum transferrin from different species, Electrophoresis, 17, 331, 1996.
74. Dabek-Zlotorzynska, E., Aranda-Rodriguez, R., and Buykx, S. E., Development and validation of
capillary electrophoresis for the determination of selected metal ions in airborne particulate matter
after sequential extraction. Anal. Bioanal. Chem., 372, 467–472, 2002.
75. Wang, M., Qu, F., Shan, X. Q., and Lin, J. M., Development and optimization of a method for the
analysis of low-molecular-mass organic acids in plants by capillary electrophoresis with indirect UV
detection, J. Chromatogr. A, 989, 285–292, 2003.
76. Jamali, B. and Lehmann, S., Development and validation of a high-resolution capillary electrophoresis
method for multi-analysis of ragaglitazar and arginine in active pharmaceutical ingredients and low-
dose tablets, J. Pharm. Biomed. Anal., 34, 463–472, 2004.
77. Sanger-van de Griend, C. E., Ek, A. G., Widahl-Nasman, M. E., and Andersson, E. K., Method
development for the enantiomeric purity determination of low concentrations of adrenaline in local
anaesthetic solutions by capillary electrophoresis, J. Pharm. Biomed. Anal., 11, 41, 77–83, 2006.
78. Kamoda, S., Ishikawa, R., and Kakehi, K., Capillary electrophoresis with laser-induced fluores-
cence detection for detailed studies on N-linked oligosaccharide profile of therapeutic recombinant
monoclonal antibodies, J. Chromatogr. A, 1133, 332–339, 2006.
79. Iqbal, J., Scapozza, L., Folkers, G., and Muller CE., Development and validation of a capillary electro-
phoresis method for the characterization of herpes simplex virus type 1 (HSV-1) thymidine kinase
substrates and inhibitors. J. Chromatogr. B–Anal. Technol. Biomed. Life Sci., 846, 281–290, 2007.
80. Wang, W., Tang, J., Wang, S., Zhou, L., and Hu, Z., Method development for the determination of
coumarin compounds by capillary electrophoresis with indirect laser-induced fluorescence detection,
J. Chromatogr. A, 2007. Vol 1148(1): 108–114.
81. Li, P., Li, S. P., and Wang, Y. T., Optimization of CZE for analysis of phytochemical bioactive
compounds, Electrophoresis, 27, 4808–4819, 2006.
82. Kamoda, S. and Kakehi, K., Capillary electrophoresis for the analysis of glycoprotein pharmaceuticals,
83. McLaughlin, G. M., Nolan, J. A., Lindahl, J. L., Palmieri, R. H., Anderson, K. W., Morris, S. C.,
Morrison, J. A., and Bronzert, T. J., Pharmaceutical drug separations by HPCE: practical guidelines,
J. Liq. Chromatogr., 15, 961, 1992.
84. Gobie, W.A. and Ivory, C. F., Thermal model of capillary electrophoresis and a method of counteracting
thermal band broadening, J. Chromatogr., 516, 191, 1990.
85. Davis, J. M., Influence of thermal variations of diffusion coefficient on non-equilibrium plate height
in capillary zone electrophoresis, J. Chromatogr., 517, 521, 1990.
86. Terabe, S., Otsuka, K., and Ando, T., Electrokinetic chromatography with micellar solution ad open-
tubular capillary, Anal. Chem., 57, 834, 1985.
87. Rush, R. S., Cohen, A. S., and Karger, B. L., Influence of column temperature on the electrophoretic
behavior of myoglobin and α-lactalbumin in high-performance capillary electrophoresis, Anal. Chem.,
63, 1346, 1991.
88. Kurosu, Y., Hibi, K., Sasaki, T., and Saito, M., Influence of temperature control in capillary
electrophoresis, J. High Resolut. Chromatogr., 14, 200, 1991.
89. Guttman, A., Horvath, J., and Cooke, N., Influence of temperature on the sieving effect of different
polymer matrices in capillary SDS gel electrophoresis of proteins, Anal. Chem., 65, 199, 193.
90. Giordano, B. C., Horsman, K. M., Burgi, D. S., Ferrance, J. P., and Landers, J. P., Method for
determining intracapillary solution temperatures: application to sample zone heating for enhanced
fluorescent labeling of proteins, Electrophoresis, 28, 714, 2007.
91. Nelson and Cooke, Handbook of Capillary Electrophorosis, 1st ed., CRC Press.
92. Moyer, T. P., Jiang, N.-S., Tyce, G. M., and Sheps, S. G., Analysis for urinary catecholamines by liquid
chromatography with amperometric detection: methodological and clinical interpretation of results,
Clin. Chem., 25, 256, 1979.
93. Wallingford, R. A. and Ewing, A. G., Separation of serotonin from catechols by capillary zone
electrophoresis with electrochemical detection, Anal. Chem., 61, 98, 1989.
94. Lui, J., Shirota, O., and Novotny, M., Capillary electrophoresis of amino sugars with laser-induced
fluorescence detection, Anal. Chem., 63, 413, 1991.
95. Honda, S., Makino, A., Suzuki, S., and Kakehi, K., Analysis of the oligosaccharides of ovalbumin by
high performance capillary electrophoresis, Anal. Biochem., 191, 228, 1990.
96. Landers, J. P., Oda, R. P., and Schuchard, M. D., Separation of boron-complexed diol compounds
using high-performance capillary electrophoresis, Anal. Chem., 64, 2846, 1992.
97. Chen, F. A., Kelly, L., Palmieri, R., Biehler, R., and Schwartz, H., Use of high ionic strength buffers
for the separation of proteins and peptides with capillary electrophoresis, J. Liq. Chromatogr., 15,
1143, 1992.
98. Nelson, R. J., Paulus, A., Cohen, A. S., Guttman, A., and Karger, B. L., Use of Peltier thermoelectric
devices to control column temperature in high-performance capillary electrophoresis, J. Chromatogr.,
480, 111, 1989.
99. Sepaniak, M. J. and Cole, R. O., Column efficiencies in micellar electrokinetic capillary chromatog-
raphy, Anal. Chem., 59, 472, 1987.
100. Landers, J. P., Oda, R. P., Spelsberg, T. C., Nolan, J. A., and Ulfelder, K. J., Capillary electrophoresis:
A powerful microanalytical technique for biologically active molecules, Biotechnique, 14, 98, 1993.
101. Bushey, M. and Jorgenson, J., Capillary electrophoresis of proteins in buffers containing high
concentrations of zwitterionic salts, J. Chromatogr., 480, 301, 1989.
102. Oda, R. P., Madden, B. J., Morris, J. C., Spelsberg, T. C., and Landers, J. P., Multiple-buffer-additive
strategies for enhanced electrophoretic separation of peptides, J. Chromatogr., 680, 341, 1994.
103. Bean, S. R., Lookhart, G. L., and Bietz, J. A., Acetonitrile as a buffer additive for free zone capillary
electrophoresis separation and characterization of maize (Zeamays L.) and sorghum (Sorghum bicolor
L. Moench) storage proteins. J. Agric. Food Chem., 48, 318–327, 2000.
104. Gordon, M. J., Lee, K.-J., Arias, A. A., and Zare, R. N., Protocol for resolving protein mixtures in
capillary zone electrophoresis, Anal. Chem., 63, 69, 1991.
105. Swedberg, S. A., Characterization of protein behavior in high-performance capillary electrophoresis
using a novel capillary system, Anal. Biochem., 185, 51, 1990.
106. Bolger, C. A., Zhu, M., Rodriguez, R., and Wehr, T., Performance of uncoated and coated capillar-
ies in free zone electrophoresis and isoelectric focusing of proteins, J. Liq. Chromatogr., 14, 895,
1991.
107. Wiktorowicz, J. E. and Colburn, J. C., Separation of cationic proteins via charge reversal in capillary
electrophoresis, Electrophoresis, 11, 769, 1990.
108. Towns, J. K. and Regnier, F. E., Capillary electrophoretic separation of proteins using nonionic
surfactant coatings, Anal. Chem., 63, 1126, 1991.
109. Bullock, J. A. and Yuen, L.-C., Free solution capillary electrophoresis of basic proteins in uncoated
fused silica capillary tubing, J. Microcol. Sep., 3, 241, 1991.
110. Song, L., Ou, Q., and Yu, W., Study on phosphate of ethylenediamine, 1,3-diaminopropane and 1,4-
diaminobutane as carrying electrolyte in open-tubular capillary electrophoresis, J. Liq. Chromatogr.,
17, 1953, 1994.
111. Oda, R. P., Madden, B. J., Spelsberg, T. C., and Landers, J. P., α, β-Bis-quarternary ammonium alka-
nes as effective buffer additives for enchanced capillary electrophoretic separation of glycoproteins,
J. Chromatogr., 680, 85, 1994.
112. Morbeck, D. E., Madden, B. J., Charlesworth, C. C., and McCormick, D. J., High resolution separation
of human choriogonadotropin (hCG) isoforms using capillary electrophoresis, J. Chromatogr., 680,
217, 1994.
113. Gilges, M., Husmann, H., Kleemiß, M. H., Motsch, S. R., and Schomberg, G., CZE separation of basic
proteins at low pH in fused silica capillaries with surfaces modified by silane derivatization and/or
adsorption of polar polymers, J. High Res. Chromatogr., 15, 452, 1992.
114. Schomberg, G., Belder, D., Gilges, M., Kleemiß, M. H., and Motsch, S. R., Wall chemistry in
capillary electrophoresis: the importance of maximal efficiency in analytical CE, Abstract 101, a
lecture presented at HPCE ’94, San Diego, CA, 1994.
115. Figeys, D., Renborg, A., and Dovichi, N. J., Spatial and temporal depletion of ions from noncrosslinked
denaturing polyacrylamide in capillary electrophoresis, Electrophoresis, 15, 1512–1517, 1994.
116. Strege, M. A. and Lagu, A. L., Studies of migration time reproducibility of CE proteins separations,
J. Liq. Chromatogr., 16, 51, 1993.
117. Landers, J. P., Oda, R. P., Madden, B., Sismelich, T. P., and Spelsberg, T. C., Reproducibility
of sample separation using liquid or forced air convection themostated high performance capillary
electrophoresis, J. High Res. Chromatogr., 15, 517, 1992.
118. Trushina, E. V., Oda, R. P., McMurray, C. T., and Landers, J. P., Effective and reproducible capillary
electrophoretic separation of thiols under conditions where exceptionally high current is generated.
Anal Chem., 71, 5569–5573, 1999.
119. Abersold, R. and Morrison, H. D., Analysis of dilute peptide samples by capillary zone electrophoresis,
J. Chromatogr., 516, 79, 1990.
120. Moring, S. E., Colburn, J. C., Grossman, P.D., and Lauer, H. H., Analytical aspects of an automated
capillary electrophoresis system, LC-GC, 8, 34, 1989.
121. Chien, R.-L. and Burgi, D. S., On-column sample concentration using filed amplification in CZE,
Anal. Chem., 64, 489A, 1992.
122. Chein, R.-L. and Helmer, J. C., Electroosmotic properties and peak broadening in field-amplified
capillary electrophoresis, Anal. Chem., 63, 1354, 1990.
123. Foret, F., Sustacek, V., and Bocek, B. J., On-line isotachophoretic sample preconcentration for
enhancement of zone detectability in capillary zone electrophoresis, J. Microcol. Sep., 2, 229, 1990.
124. Burgi, D. and Chien, R.-L., Handbook of Capillary Electrophoresis (2nd edition), Ed. J.P. Landers,
CRC Press, Boca Raton, FL, Chapter 16, 1996.
125. Kurnik, R. T., Boone, T. D., Nguyen, U., Ricco, A. J., and Williams, S. J., Use of floating electrodes
in transient isotachophoresis to increase the sensitivity of detection, Lab Chip, 3, 86–92, 2003.
126. Guzman, N. A., Trebilcok, M. A., and Advis, J. P., The use of a concentration step to collect urinary
components separated by capillary electrophoresis and further characterization of collected analytes
by mass spectrometry, J. Chromatogr., 14, 997, 1991.
127. Guzman, N. A., Automated capillary electrophoresis apparatus, U.S. Patent #5,202,010, 1993.
128. Fuchs, M. and Merion, M., Apparatus for effecting capillary electrophoresis, U.S. Patent #5,246,577,
1993.
129. Strausbauch, M. A., Xu, S. Z., Ferguson, J. E., Nunez, M., Machacek, D., Lawson, G. M., Wettstein,
P. J., and Landers, J. P., Concentration and analysis of hypoglycemic drugs using solid phase extraction-
capillary electrophoresis, J. Chromatogr., 717, 279–291, 1995.
130. Tomlinson, A. J., Benson, L. M., Braddock, W. D., Oda, R. P., and Naylor, S., On-line preconcentration-
capillary electrophoresis-mass spectrometry (PC-EC-MS), J. High Resolut. Chromatogr., 17, 729,
1994.
131. Strausbauch, M., Landers, J. P., and Wettstein, P. J., Mechanism of peptide separations by solid phase
extraction-capillary electrophoresis (SPE-CE) at low pH, Anal. Chem., 68, 306–314, 1996.
132. Strausbauch, M. A., Madden, B. J., Wettstein, P. J., and Landers, J. P., Sensitivity enhancement and
second dimensional information from the SPE-CE analysis of entire HPLC fractions, Electrophoresis,
16, 541, 1995.
133. Beattie, J. H., Self, R., and Richards, M. P., The use of solid phase concentrators for on-line preconcen-
tration of metallothionen prior to isoform separation by capillary zone electrophoresis, Electrophoresis,
16, 322, 1995.
134. Roche, M. E., Anderson, M. A., Oda, R. P., Riggs, L. B., Strausbauch, M. A., Okazaki, R.,
Wettstein, P. J., and Landers, J. P., Capillary electrophoresis of insulin-like growth factors: enhanced
UV detection using dynamically-coated capillaries and on-line solid phase extraction (SPE-CE), Anal.
Biochem., 258, 87–95, 1998.
135. Cao, W., Easley, C. J., Ferrance, J. P., and Landers, J. P., Chitosan as a novel polymer for pH-
induced DNA purification in a totally aqueous system: I. High density open-channel pattern for DNA
purification, Anal. Chem., 78, 7222–7228, 2006.
136. Wolfe, K. A., Breadmore, M. C., Ferrance, J. P., Power, M. E., Conroy, J. F., Norris, P. M., and
Landers, J. P., Toward a microchip-based solid-phase extraction method for isolation of nucleic acids,
Electrophoresis, 23, 727–733, 2002.
137. Breadmore, M. C., Wolfe, K. A., Arcibal, I. G., Leung, W. K., Dicks, D., Giordano, B. C., Power, M. E.
et al., Microchip-based purification of DNA from biological samples, Anal. Chem., 75, 1880–1886,
2003.
138. Wu, Q., Bienvenue, J. M., Giordano, B. C., Hassan, B. J., Kwok, Y. C., Norris, P. M., Landers, J. P.,
and Ferrance, J. P., A microchip-based macroporous silica sol-gel monolith for efficient isolation of
DNA from clinical samples, Anal. Chem., 78, 5704–5710, 2006.
139. Wen, J., Guillo, C., Ferrance, J. P., and Landers, J. P., DNA extraction using a tetramethyl orthosilicate-
grafted photopolymerized monolithic solid phase, Anal. Chem., 78, 1673–1681, 2006.
140. Guzman, N. A. and Phillips, T. M., Immunoaffinity CE for proteomics studies, Anal. Chem., 77,
61A–67A, 2005.
141. Guzman, N. A., Improved solid-phase microextraction device for use in on-line immunoaffinity
capillary electrophoresis, Electrophoresis, 24, 3718–3727, 2003.
142. Vizioli, N. M., Carducci, C. N., Pena, C., and Guzman, N. A., Monitoring the purity of a syn-
thetic peptide by capillary electrophoresis: utilization of an on-line preconcentration method for
improved separation and detection sensitivity, J. Capill. Electrophor. Microchip Technol., 6, 109–118,
1999.
APPENDIX 1
CALCULATIONS OF PRACTICAL USE
In this section, we utilize an exemplary electropherogram (Figure A1.1) and the corresponding data
obtained using the data management software (Figure A1.1, inset) to take the reader through a variety
of calculations of practical use.
A.1 MOBILITY
The velocity of an analyte, νi , is the distance traveled during the time of electric field application.
Velocity is related to mobility, µi , by field strength, E = H/L. Thus,
νi = µapp H/L = µapp E
Peak Migration Component Peak Peak Apparent

Number Time Area Height Mobility
1 1.723 Mesityl oxide 0.10762 0.00688 7.2735
2 2.642 ρ-OH φ acet. 0.10624 0.00463 4.7446
3 2.818 ρ-OH benz. 0.18887 0.00819 4.4477
4 2.924 benzoate 0.14680 0.00616 4.2868
0.54952 0.02565
4
0.015 A.U.
2
Abs. (200 nm)
0 60 120 180 240 300

Time (sec)
FIGURE A1.1 Simple separation of four small molecules for calculations of practical use. Separation of
mesityl oxide [an EOF (neutral) marker], p-hydroxyphenylacetic acid, p-hydroxybenzoic acid, and benzoic
acid [in borate buffer, pH 8.3; 3-s pressure (@0.5 psi) injection]. Analysis was carried out on a Beckman P/ACE
System 2050. Separation conditions: capillary, 50 µm × 40 cm (effective length), 47 cm total length bare fused
silica; T , 28◦ C; voltage, 25 kV; buffer, 100 mM borate, pH 8.3; detection, 200 nm. Inset: data derived from
System Gold (vs7.1) data management software.
The apparent mobility µapp measured from an electropherogram is the sum of the mobility of the
analyte and that due to electroosmotic flow (EOF):
µapp,i = µep + µeo

= Ld Lt /tapp,i H
= Ld Lt /H(1/tapp,i − 1/tref ) + µref
where Ld is length to the detector, Lt is total length of the capillary, and, if the reference peak is a
neutral marker,
µref = 0.
To calculate µref for a charged reference peak, the above equation may be used, substituting tapp,i
for the charged reference and tref for the neutral marker, with µref = 0.
A.1.1 EXAMPLE
From Figure A1.1, Ld = 40 cm, Lt = 47 cm, and t1 = 1.723 min:
ν = Ld /t1 = 0.4m/(1.723 min)(60s/min)

= 0.003869 m/s
= µapp E = µapp H/Lt
or apparent mobility = µapp,1 = νLt /H
= (0.003869 m/s)(0.47 m)/25 kV
= (0.001818/25) = 7.273 × 10−4 cm2 /(V · s)
Similarly for the other peaks,
µapp,i = (0.4)(0.47)/(60)(25 × 103 )ti = 12.533 × 10−4 /ti cm2 /(V · s)
µapp,2 = 12.533 × 10−4 /2.642 = 4.744 × 10−4 cm2 /(V · s)
µapp,3 = 12.533 × 10−4 /2.818 = 4.448 × 10−4 cm2 /(V · s)
µapp,4 = 12.533 × 10−4 /2.924 = 4.286 × 10−4 cm2 /(V · s)
Using
µapp,i = Ld Lt /H(1/tapp, i − 1/tref ) + µref
we many make the same calculation using µref = µeo = 7.273 × 10−4 cm2 /(V · s):
µapp, i = [(0.4)(0.47)/25 × 103 ](1/tapp, i − 1/1.723) + 7.273

µapp,2 = 12.533(1/2.642 − 1/1.723) + 7.273 × 10−4
= 12.533(0.3785 − 0.5804) + 7.273 × 10−4
= −2.530 + 7.273 × 10−4 cm2 /(V · s) = 4.743 × 10−4 cm2 /(V · s)
µapp,3 = 12.533 × 10−4 (0.3549 − 0.5804) + 7.273 × 10−4
= −2.826 × 10−4 + 7.273 × 10−4 = 4.448 × 10−4 cm2 /(V · s)
µapp,4 = 12.533 × 10−4 (0.3420 − 0.5804) + 7.273 × 10−4
= −2.988 + 7.273 × 10−4 = 4.285 × 10−4 cm2 /(V · s)
The electrophoretic mobility, µep , may be calculated from
µapp,i = µep + µeo
as peak 1 in Figure A1.1 is a neutral marker, µapp,1 = µeo = 7.273 × 10−4 cm2 /(V · s)
µep,2 = 4.744 − 7.273 = −2.529 × 10−4 cm2 /(V · s)

µep,3 = 4.448 − 7.273 = −2.825 × 10−4 cm2 /(V · s)
µep,4 = 4.286 − 7.273 = −2.987 × 10−4 cm2 /(V · s)
A.2 CORRECTED PEAK AREA

Integrators typically present peak area in dimensions of (response) (time). This calculation is a simple
transformation to obtain area in (response) (width).
Acorr = Ai (mAU × min)νi

= Ai (mAU × min)Ld (cm)/tm (min)
= Ai (mAU × cm)
This can usually be achieved through the data collection and software system.
A.2.1 EXAMPLE
Acorr,1 = (0.10762)(40)/1.723
= 2.498 mAU cm
Similarly,
Acorr,2 = (0.10624)(40)/2.642 = 1.608

Acorr,3 = (0.18887)(40)/2.818 = 2.681
Acorr,4 = (0.14680)(40)/2.924 = 2.008
Dividing the corrected area by peak height gives peak width, which approximates the peak width
at half-height, which may be used to calculate N (see calculations, Section A.5).
w1/2,1 = 2.498 mAU/0.00668 µAU = 374 µm

w1/2,2 = 1.608 mAU/0.00463 µAU = 347 µm
w1/2,3 = 2.681 mAU/0.00819 µAU = 327 µm
w1/2,4 = 2.008 mAU/0.00616 µAU = 326 µm
A.3 QUANTITY OF SAMPLE INTRODUCED INTO THE CAPILLARY

We assume
Q = (volume)int (concentration)
= πr 2 l [Ci ]
where r is the internal radius of the capillary, l is the length of the sample plug, and [Ci ] is the
concentration of the sample.
A.3.1 HYDRODYNAMIC INJECTION

Q = πr 2 [
Pr 2 tint /(8ηL)][Ci ]
where
P is the pressure difference, r is the capillary inner radius, tint is the introduction time, η is
the viscosity of the sample solution, and L is the length of the column.
A.3.1.1 Example
Assuming typical values for the constants,
Column dimensions : 50 µm i.d. × 47 cm, 40 cm to the detector

η = 0.9548 cP = 9.548 × 10−4 N/s/m2 at22◦ C

P = 0.5 psi = 3.435 × 103 N/m2
tint = 3 s
Then
2

−6 2 −6 −4
volint = (3.1416)(25 × 10 ) 3.435 × 10 25 × 10
3
3 8 9.548 × 10 0.47

= 19.635 × 10−10 17.94 × 10−4
= 3.523 × 10−12 m3 = 3.523 nL
which correlates with 1.2 nL/s injection.

A.3.2 ELECTROKINETIC INTRODUCTION
Q = [πr 2 (µapp )Vint tint /L][Ci ]
where µapp is the mobility of the analyte, µeo is the electroosmotic mobility, and Vint is the
introduction voltage.
A.3.2.1 Example
Again assuming typical values,
Column dimensions : 50µm i.d. × 47 cm, 40 cm to the detector

tint = 20 s
µapp = 4.744 × 10−8 m2 /V/s
µeo = 7.273 × 10−8 m2 /V/s
Vint = 1.5 kV
Then
2
volint = (3.1416) 25 × 10−6 4.744 × 10−8 1.5 × 103 20/0.47 m3

= 19.635 × 10−10 302.81 × 10−5 m3
= 5.945 × 10−15 m3 = 5.945 nL
However, if we calculate the volume of fluid injected, which is due solely to µeo ,

volint = (3.1416) (3.1416) (x)2 7.273 × 10−8 1.5 × 103 20/(0.47) m3

= 19.635 × 10−10 464.2 × 10−5 m3
= 9115 × 10−15 m3 = 9.115 nL
With electrokinetic introduction, the amount of sampled injected is proportional to the analyte
mobility. It is typically used when the sample has greater mobility than EOF. If one has a highly
mobile analyte in relatively low ionic strength sample buffer (1/x times separation buffer) one may
load an x-fold greater portion onto the column by electrokinetic introduction without adversely
affecting separation resolution. In the above example, if the sample were in 0.1 times separation
buffer, we could have loaded the equivalent of 59.5 nL of sample.
A.4 RESOLUTION
Using the equation for resolution, we obtain
Res = 2(xi2 − xi1 )/(w1 + w2 )

A.4.1 EXAMPLE
Res2−1 = 2(2.818 − 2.642)/(0.02295 + 0.02306)

= 0.352/0.04601
= 7.650
Res3−2 = 2(2.924 − 2.818)/(0.02306 + 0.02383)
= 0.212/0.04689
= 4.521
Using the criterion that peaks are resolved when Res = 1, we could state that all three peaks are
resolved.
A.5 EFFICIENCY
From the above example, we may calculate the efficiency (from Equation 1.17) of the separation of
the peaks in Figure A1.1.
N = 5.54(Ld /w1/2 )2
where Ld , the length to the detector, is 40 cm, and the peak width is (peak area/peak height), both
of which must be in the same units, either in time (min) or distance (cm). We shall calculate the
efficiency using both time and distance.
A.5.1 EXAMPLE 1
For time calculations,
Ni = 5.54(Ld /w1/2 )2 = 5.54[ti /(areai /peak heighti )]2
From the data in the inset of Figure A1.1, peak area is given in mAU × min, and peak height is in
AU. Thus,
N1 = 5.54 {(1.732)/[0.10762(0.00668) (1000)]}2
= 5.54(1.732/0.01611)2 = 5.54(106.95)2 = 63, 368
N2 = 5.54(2.642/0.02295)2 = 5.54(115.14)2 = 73, 445
N3 = 5.54(2.818/0.02306)2 = 5.54(122.20)2 = 82, 728
N4 = 5.54(2.942/0.02383)2 = 5.54(122.70)2 = 83, 406
A.5.2 EXAMPLE 2
For distance calculations,
N1 = 5.54(Ld /w1/2 )2 = 5.54[Ld /(areacon /peak height)]2

From calculations, Section A.2, corrected peak area above,
w1/2,1 = 2.498 mAU/0.00668 µAU = 374 µm
w1/2,2 = 1.608 mAU/0.00463 µAU = 347 µm
w1/2,3 = 2.681 mAU/0.00819 µAU = 327 µm
w1/2,4 = 2.008 mAU/0.00616 µAU = 326 µm
Therefore,
N1 = 5.54(40/0.374)2
= 5.54(106.95)2 = 63, 370
Likewise,
N2 = 5.54(40/0.347)2 = 73, 616

N3 = 5.54(40/0.327)2 = 82, 896
N4 = 5.54(40/0.326)2 = 83, 405
A.6 JOULE HEATING

To calculate the Joule heating of a buffer, one should run an Ohm’s law plot as outlined in Section
1.3.3.5.2, under Buffers. To calculate watts, we use the following:
Watts/m = (voltage)(amperage)/(column length, cm) 1000
A.6.1 EXAMPLE
For 100 mM borate, pH 8.3: (25 kV)(13.17 µA)/(57 cm) 1000 = 0.58 W/m
For 100 mM CAPS, pH 11.0: (25 kV)(134 µA)/(57 cm) 1000 = 5.88 W/m
For 20 mM CAPS, pH 11.0: (25)(30.03)/57,000 = 1.32 W/m
For 100 mM PO4 , pH 2.5: (25)(229.5)/57,000 = 10.06 W/m
For 50 mM PO4 , pH 2.5: (25)(116.9)/57,000 = 5.13 W/m
For 25 mM PO4 , pH 2.5: (25)(70.03)/57,000 = 3.07 W/m
APPENDIX 2
TROUBLESHOOTING
This table summarizes problems commonly encountered in capillary electrophoresis (CE) experi-
ments, and plausible solutions to these problems.
Problem Cause Remedy

I Peak-Associated Problems
A. No peak observed
1. Baseline on scale Inappropriate data collection • Reset scale for appropriate absorbance parameter
scale
Inappropriate detector range • Reset detector range
Inappropriate detector • Reset wavelength
wavelength
Separation time too short • Lengthen analysis time
• Lengthen capillary
Reasonable current • See Section 1.2.1
Flow through capillary? • See Section 1.5
Integrity of sample • Check sample level
• Check for air bubble in bottom of
sample vial
• Increase sample introductory time
• Check sample caps for leakage
• See Section 1.5
Using voltage introduction • See Section 1.2.1
2. Baseline off scale Offset baseline • Rezero detector
Inappropriate detector • Reset wavelength
wavelength
B. Peaks present
1. Too many peaks Random peaks may be caused • Warm buffer to room temperature
by microbubbles
• Reduce voltage
Solid contaminants in sample • Filter sample (0.45 µm pore size filter)
Solid contaminants in buffer • Filter buffer
Residue from previous analysis • Wash capillary
Sample degradation • Replace sample; check temperature of capillary
chamber
2. Too few peaks Proper wavelength used • Reset detector wavelength
Sufficient time • Lengthen analysis time
Unreasonable current • See Section 1.2.1
Unreasonable voltage • See Section 1.2.1
Inappropriate temperature • Reset temperature control
Analyte–wall interactions • Wash capillary
• Check running conditions
Fluid flow inadequate • See Section 1.5
Inadequate separation time • Lengthen analysis time
• Reduce electroosmotic flow (EOF)
Similar charge/mass ratio (not • Alter buffer pH
resolving components) • Reduce EOF
• Try alternative separation mode: micellar
electrokinetic chromatography (MEKC), capillary
isotachophoresis (CITP), capillary isoelectric focusing
(CIEF)
3. Distorted peak shape
a. Small peaks Improper wavelength • Reset wavelength
Inappropriate range on detector • Reset detector
Improper integration • Reset integrator parameters
Inappropriate sampling time • Longer sample introduction time
Continued

b. Flat-topped peaks Sample too concentrated • Reduce introduction time
• Dilute sample
c. Peak tailing Current too high • Ionic strength of separation buffer too high.
Check composition; dilute
• Reduce voltage
Sample buffer ionic strength too • Dilute sample buffer
high
Analyte–wall interactions Modify separation buffer:
• Increase ionic strength
• Add organic solvents
• Add cationic compounds
• Add zwitterions (phosphorylethanolamine)
• Use coated capillary
• Consider different CE mode (e.g., MEKC)
II Instrument-Based Problems
A. Current
1. Fluctuating current during Vial levels low • Replenish buffer
separation
Loose electrode connections • Tighten connections
Plugged capillary • Rinse/replace
Sample matrix effects • Check ionic strength of matrix
2. No current Safety interlock off • Reset interlock
Plugged capillary • Wash capillary
Broken capillary • Replace capillary
Buffer depletion • Replace buffers
Empty capillary • Fill reservoirs
3. Abrupt loss of current Short in system (buffer on • Dry cap
during separation reservoir cap)
4. High current High ionic strength separation • Decrease ionic strength
buffer
B. Baseline drift Contaminated capillary • Wash capillary
Contaminated aperture on • Clean aperture
detector
Bad capillary alignment • Check alignment; realign or replace
Detector instability • Give adequate warm-up time
• Replace lamp
Unstable capillary temperature • Check oven/bath temperature
• Replace thermostat
Unstable room temperature • Stabilize room temperature
• Deflect drafts away from instrument
• Move instrument
C. Data analysis
Peaks observed: not analyzed Inappropriate integration • Reset integrator attenuation
parameters
III Sample Introduction
A. Electrokinetic
No peaks Anodic sample • Revere polarity
• Change separation buffer pH
Sample ionic strength too high • Dilute sample buffer
Sample pH too high • Adjust pH
Sample ionic strength too low • Raise ionic strength
B. Hydrodynamic
1. No peaks Poor seal with sample vial • Replace seal
Pinched pressure/vacuum line • Replace line

Depleted pressure source • Replace
Anodic sample • Reverse polarity
• Change separation buffer pH
2. Irreproducible peak Poor seal • Replace cap
height/area
• Decrease percentage volatile
Sample matrix volatility • Decrease solvent concentration
IV Poor Quantitative Reproducibility
A. Migration time
1. Unstable temperature Ionic strength too high • Dilute buffer
Voltage too high • Decrease voltage
Temperature too high • Decrease thermostatted temperature (see Section 1.2.2)
Sample matrix ionic strength • Dilute sample
too high
2. Others Buffer depletion • Replenish with fresh buffer
Analyte–wall interactions • See Section 1.3.3.5
Buffer siphoning • Adjust level in inlet/outlet reservoirs1
Contaminated capillary • Rinse capillary extensively
Inadequate rinse steps • Increase rinse time with rinse solution/buffer
B. Peak height/area Analyte–wall interactions • Increase ionic strength
• Add dynamic coating agents (e.g., diaminobutane)
Current instability • See Section 1.2.1.1
Fouled capillary surface • See Section 1.2.3.4.3
V Capillary-Associated Problems2
A. No peaks—proper flow Anodic sample • Reverse polarity
through capillary
Sample buffer too viscous • Dilute sample
B. Reduced flow through Partially occluded capillary • Wash capillary. Check sample, buffers for
capillary particulates; filter if necessary
Depleted pressure source • Replace
C. No flow Broken capillary • Replace capillary
D. Peaks irreproducible Partially occluded capillary • Unblock chemically (e.g., sodium ethoxide) or
with pressure
Fouled capillary surface • Recondition capillary (e.g., NaOH or sodium ethoxide)
Poorly cut capillary inlet • Retrim inlet or replace capillary
1 This is unlikely to be significant with narrow diameter capillaries (i.e., <75 µm) since the hydrostatic head differential
must be in range of 5 cm of water in order for this phenomenon to contribute appreciably.

2 Be certain that there is adequate capillary flow. This can be accomplished as described in Chapter 1, Section 1.2.3.4.
Alternatively, one may check the flow by pressure rinsing and observing droplets forming at the outlet end. If this must be done
manually, a 6-cc syringe with a yellow Eppendorf pipette tip will produce enough pressure to create one drop every
20–30 s. with a 50 µm × 57 cm capillary.
APPENDIX 3
EFFECTIVE SEPARATION CONDITIONS FOR EACH CLASS OF
ANALYTES
The following tables provide examples of buffer systems, some containing additives or using surface-
modified capillaries, that have provided successful separation conditions for the select analytes.
Culled from the capillary electrophoresis (CE) literature, this list is not comprehensive, but is
intended to provide the practitioner with a guide for starting conditions for method development
for a particular type of analytes—no set of conditions is universally effective for any given class
of analytes. For convenience, individual tables are provided for the analyte classes—ions, small
molecules, peptides, proteins, and nucleic acids—in both bare silica and with coated surfaces,
and with the chromatographic hybrid, micellar electrokinetic chromatography. Full details on the
separation conditions can, of course, be found in the original reference.
A.3.1 Ions
Buffer Additives Analytes/Sample References
5 mM chromate or phthalate, 0.5 mM Nice-Pak OFM Organic/inorganic ions Romano et al. (1991)
pH 10.00 Anion-BT
5 mM chromate, pH 8.0 0.4 mM OFM Anion-BT Organic/inorganic ions Jones and Jandik (1992)
5 mM chromate, pH 11.0 None
(LiOH)
30 mM creatinine, pH 4.8 8 mM hydroxyisobutyric acid Metal ions Lu and Cassidy (1993)
(acetic acid)
100 mM borate 50 mM tetrabutyl ammonium Inorganic/organic ions Audalovic et al. (1993)
hydroxide
A.3.2 Small Molecules: Charged and Neutral
Buffer Additives Analytes/Sample Reference

CZE
50 mM tetraethylammonium 10 mM HCl in acetonitrile Organic bases Walbroehl (1984)
perchlorate
20–125 mM sodium None Cinnamic acid and analogs Fujiwara and Honda
phosphate, pH 2.5, 7.0–9.2 (1986)
50 mM sodium phosphate, 50% acetonitrile (v/v) Substituted benzoic acid Fujiwara (1987)
pH 7.0 isomers
50–100 mM sodium acetate, 0.1% hydroxypropyl Isotopic benzoic acids Terabe et al. (1988)
pH 3.9–4.5 cellulose (v/v)
16 mM sodium sulfate, 5 mM 30% methanol (v/v) Methotrexate analysis Roach et al. (1988)
MES, pH 6.7
20–125 mM sodium None Pharmaceuticals Altria and Simpson
phosphate, pH 2.5, 7.0–9.2 (1988)
25 mM tetrahexylammonium 50% acetonitrile (v/v) Neutral organic molecules Walbroehl and Jorger
perchlorate (e.g., polycyclic aromatic (1988)
hydrocarbons)
25 mM MES, pH 5.5–5.65 10–20% 2-propanol (v/v) Catecholamine analysis Wallingford and Ewing
(1989)
10 mM MES/His, pH 6.0 0.5 mM tetradecyltrimethyl- Carboxylic acids Huang et al. (1989)
ammonium
bromide
5 mM sodium borate, pH 9.0 2% SDS (w/v), 0–1% Polyamines Tsuda et al. (1990)
ethylene (v/v) diamine, and
5% ethylene glycol (v/v)
100 mM CAPS, pH 10.5 None Nucleotides Nguyen et al. (1990)
20 mM sodium citrate or 100 None Shellfish poisons Thibault et al. (1991)
mM acetic acid, pH 2.0
or 2.9

150 mM sodium dihydrogen Replacing water with D2 O, Aniline derivatives Okafo et al. (1991)
phosphate, pH 2.98 or 5.98 pD = 2.98 or 5.98
20 mM phosphoric acid/20% None Cimetidine in pharmaceutical Arrowood and Hoyt
KOH, pH 7.00 preparations (1991)
50 mM CAPSO and 12.5 mM None Tricyclic antidepressants Salomon and Romano
NAOH, pH 9.55 (1992)
10 mM sodium tetraborate Acetonitrile:water 50:50 cis/trans Isomers of Chadwick and Hsieh
and 50 mM boric acid or 40 butenedioic and retinoid (1991)
mM NaOH, pH 9.55 acids
20 mM sodium phosphate or 0.05% ethylene glycol (v/v) Organic/inorganic cations and Beck and Engelhardt
3–5 mM imidazole or 50 anions (1992)
mM acetic acid, pH 4.5–7.0
100 mM borate, pH 8.3 None Chloramphenicol acetyl Landers et al. (1992a)
transferase (CAT) enzyme
substrates and products
20 mM imidazolium acetate, None Sulfonamides Ackermans et al. (1992)
pH 7.0
20–50 mM phosphate, pH 6.8 0.02% Cefixime and its metabolites Honda et al. (1992)
hydroxypropylcellulose
(w/v) or 0.15% CPDAPS
(w/v) and 20% methanol
(v/v) or 0.5% polybrene
(w/v)
6 mM sorbate, pH 12.1 None Carbohydrates Vorndran et al. (1992)
100 mM borate, pH 8.4 None cis-diol containing Landers et al. (1992b)
compounds
100 mM tricine, pH 8.4 None Norepinephrine, dopamine
and metabolites
8 mM sodium carbonate, 10 None Simple carbohydrates O’Shea et al. (1993)
mM NaOH, pH 12.0
5 mM phthalate, and 50 mM 0.5 mM TTAB Organic acid counter ions (1: (1) Altria et al. (1997)
MES (1: pH 5.2; 2: pH 5.0) succinic and maleic; 2: (2) Little et al. (2007)
trifluoroacetic)
25 mM phosphate (pH 2.3) or Antifungal compounds Crego et al. (2001)
0.2 M formic acid (pH 2.15)
15 mM borax-sodium 0.1 mM TAR Uranium(VI) and other Evans and Collins (2001)
phosphate monobasic (pH transition metal ions
8.3)
50 mM borate (pH 8.15) 10 mM phytic acid Phycobiliproteins Viskari and Colyer (2002)
30 mM phosphate (pH 5.2) HVA and VMA (metabolites Li et al. (2002)
of catecholamines)
10 mM sulfated Amphetamine type stimulants Iwata et al. (2002)
γ -cyclodextrin w/ 50 mM
phosphate (pH 2.6)
30 mM phosphate (pH 9.8) Riboflavin, flavin Cataldi et al. (2003)
100 mM acetate (pH 4.5) Glycosaminoglycan Ruis-Calero et al. (2003)
monosaccharides
40 mM phosphate and 10 mM 40 mM SDS E. coli K4 polysaccharide Volpi (2004)
borate (pH 9.0)
10 mM glycine (pH 2.4) 5 mM QA-β-CD (quaternary Highly negative enatiomers Liu et al. (2004)
ammonium β-cyclodextrin)
10 mM borate (pH 9.5) 20 mM SDS, 30% Antiretroviral agents Pereira et al. (2005)
acetonitrile, 5% ethanol
Continued

0.1 M phosphate (pH 3.23) 15 mM β-cyclodextrin Metabolites of heroin Qi et al. (2005)
10 mM MES and MOPSO 0.25 mM TTAB EDTA, EDDS, IDS in Katata et al. (2006)
(pH 5.5) cosmetics
20 mM phosphate (pH 7.2) 60% acetone Amitriptyline, doxepin, and Li et al. (2006)
chlorpromazine in urine
10 mM DBA, 2 mM Cations in beverages Fung and Lau (2006)
18-crown-6, and 8 mM
lactic acid (pH 4.65)
15 mM chromate (pH 8.5) 1 mM CTAB Cyanoacrylate adhesives Whitaker et al. (2007)
32 mM borate (pH 9.2) 4.5 mM SDS Monohydroxyl stereoisomers Wang et al. (2007)
of flavanones
Borate Borate Corticosteroids Palmer et al. (1998)
MEKC MEKC separations
100 mM phosphate, pH 7.0 25 mM CM-β-cyclodextrin Cresol isomers Terabe et al. (1983)
25 mM tetraborate/50 mM 1 mM SDS Phenols Terabe et al. (1984)
phosphate, pH 7.0
50 mM phosphate/125 mM 50 mM SDS PTH-amino acids Otsuka et al. (1985)
tetraborate, pH 7.0
100 mM tris-HCL, pH 7.0 50 mM DTAB
20–50 mM phosphate, pH 10–50 mM SDS Pharmaceuticals Fujiwara and Honda
7.0–8.0 (1987)
10 mM disodium phosphate 50 mM SDS or 50 mM Phenols and polycyclic Burton et al. (1987)
and 6 mM borate, pH dodecyltrimethylammonium aromatic hydrocarbons
7.0–9.0 chloride or 50 mM STS
10 mM phosphate/6 mM 10 mM SDS Catechols and catecholamines Wallingford and Ewing
borate, pH 7.0 (1989)
20 mM phosphate and 20 mM 50 mM SDS and 20–60 mM Vitamins and pharmaceuticals Nishi et al. (1989)
borate, pH 9.0 tetralkylammonium salts
29 mM phosphate-borate, pH 50 mM sodium cholate Corticosteroids Nishi and Terabe (1990)
9.0
20 mM sodium dihydrogen 100 mM SDS and 10% Nucleosides and nucleotide- Lecoq et al. (1991)
phosphate and 20 mM acetonitrile (v/v) 3-monophosphates
borate, pH 9.2
50 mM phosphate, pH 6.0 50 mM SDS and 5% Creatinine and uric acid and Mikaye et al. (1991)
2-propanol (v/v) polycyclic aromatic
hydrocarbons
50 mM phosphate and 100 8 mM α-cyclodextrin, 1 mM Plant growth regulators Yeo et al. (1991)
mM borate, pH 8.09 β-cyclodextrin, and 1 mM
γ - cyclodextrin
2.5–5.0 mM borate, pH 10–50 mM SDS Organic gunshot and Northrop et al. (1991)
7.8–8.9 explosives
10 mM borate-phosphate, pH 50 mM SDS, 6 M urea, 20% Benzothiazole sulfenamides Nielsen and Mensink
8.7 methanol (v/v) (1991)
8.5 mM phosphate and 8.5 85 mM SDS and 15% (v/v) Acidic and neutral heroin Weinberger and Lurie
mM borate, pH 8.5 acetonitrile impurities (1991)
100 mM borate and 50 mM 30 mM SDS, 3% 2-propanol Vitamins Ong et al. (1991)
phosphate, pH 7.6 (v/v) or 3 mM
γ -cyclodextrin
10 mM borate, pH 9.5 75 mM SDS and 10% Hydroquinone and related Sakudinskaya et al.
methanol (v/v) molecules in skin toning (1992)
cream
50 mM phosphate and 100 3 mM β-cyclodextrin or 2 Sulfonamides and polycyclic Ng et al. (1992)
mM borate, pH 6.0 or 7.0 mM γ -cyclodextrin and 10 aromatic hydrocarbons
mM SDS

10 mM phosphate and 10 mM 100 mM SDS and 20% Enantiomers of amphetamine, Lurie (1992)
borate, pH 9.0 methanol (v/v) methamphetamine, and their
hydroxyphenethylamine
precursors
18 mM sodium tetraborate and 50 mM CTAB Glucosinolates and their Morin et al. (1992)
30 mM phosphate, pH 7.0 desulfo-derivatives
60 mM tris/phosphate, pH 2.5 20 mM cyclodextrin (α,β,γ ) Basic drugs Nielsen (1993)
10 mM sodium phosphate 10% acetonitrile Morphine-3-glucuronide Wernly and Thormann
(1993)
6 mM borate, pH 9.2 75 mM SDS
10 mM sodium phosphate 7 mM cyclodextrin (α,β,γ ), Mycotoxins Holland and Sepaniak
acetonitrile (0–15%) (1993)
6 mM sodium borate 50 mM SDS, 50 mM
deoxycholate
50 mM borate pH 8.5, 100 mM MeOH 15% Phthalates/soil Guo et al. (2005)
sodium cholate
20 mM phosphate pH 2.5, 50 ACN 15%, 5% THF Flavonoids Tonin et al. (2005)
mM SDS
25 mM borate pH 9.5, 30 mM Chloropropham aniline Orejuela et al. (2005)
SDS, 10 mM Triton X-100 metabolites / potatoes
20 mM borate, 20 mM Nitrofuran antibiotics Wickramanayake et al.
phosphate pH 9, 80 mM (2006)
sodium deoxycholate
50 mM borate pH 9.3, 50 mM γ −CD 20 mM Doxorubicin, doxorubicinol Eder et al. (2006)
SDS
20 mM borate, 150 mM CTAB, 31.5% H2 O Biphenyl nitrile derivatives Gong et al. (2006)
10% IPA in MeOH
15 mM borate pH 10.2, 40 mM Ibuprofen, tetrazepam/urine Nevado et al. (2006)
SDS
20 mM Tris pH 9, 300 mM SDS MeOH 18% Cefepime, Yang et.al. (2007)
vancomycin/plasma, CSF
100 mM borate pH 9, 60 mM IPA 2% Pesticides/wine Molina-Mayo et al.
SDS (2007)
20 mM borate pH 9.3, 50 mM β−CD 5 mM Amine metabolites/human Tseng et al. (2007)
sodium cholate, 20 mM Brij-35 biofluids
5 mM borate pH 9.3, 20 mM ACN 20% Sudan dyes/chili powder Mejia et al. (2007)
SDS
A.3.3 Peptides

150 mM phosphate, pH 3.0 None Angiotensin II octapeptides McCormick (1988)
10 mM tricine, pH 8.0–8.1 5.8–45 mM morpholine and LGH tryptic digest Neilsen et al. (1989a)
20 mM NaCl or KCl
100 mM phosphate, pH 2.5 30 mM ZnSO4 dl-His-dl-His Mosher (1990)
25 mM tris/25 mM phosphate, 50 mM HTAB Angiotensin analogs Liu et al. (1990)
pH 7.05
0.5 mol/L acetic acid, pH 2.6 None Di- and triglycine, synthetic Prusik et al. (1990)
growth hormone-releasing
peptide
Continued

250 mM borate, pH 7.0 1% ethylene glycol (v/v) and Proteinase-digested horse Tanaka et al. (1991)
7% acetonitrile (v/v) myoglobin
20 mM citric acid pH 2.5 None Motilin and synthetic Florance et al. (1991)
peptides
20 or 150 mM sodium Replacing water with D2 O, Simple peptides, tryptic Camilleri et al. (1991)
dihydrogen phosphate, pH 2.93, pD = 2.93, 7.93, or 7.95 digest of calcitonin, gluc-
7.93, or 7.95 agons, and cytochrome c
40–80 mM tris and tricine, pH None β−Casein tryptic and Krueger et al. (1991)
8.1–8.2 ACTH-endoproteinase Arg
C digests
25 mM phosphate, pH 2.2 (KOH) None Adrenocorticotropic hormone Van de Goor et al.
(ACTH) peptide fragments (1991)
20 mM formate, pH 3.8 Alanine
20 mM ε -aminocaproate, pH 4.4 None
(Acetic acid)
20 mM histidine, pH 6.2 MES
40 mM imidazole, pH 7.5 MOPS
100 mM borate, pH 8.3 (KOH) None
50 mM phosphate, pH 2.5 40% acetonitrile (v/v) Multiple antigen peptides Tanaka et al. (1991)
50 mM sodium dihydrogen 0.1% TFA (v/v) and 0.05% α and β species of CGRP Saria (1992)
phosphate, pH 3.93 hydroxymethylpropyl
cellulose (v/v)
20 mM citrate buffer, pH 2.50 None Peptide monomers and Landers et al. (1993)
dimmers
50 mM formic acid, pH 2.5 10 mM sodium chloride Basic peptides Gaus et al. (1993)
10 mM tris and 10 mM disodium, 50 mM dodecyltrimethyl Angiotensin analogs Novotny et al. (1990)
pH 7.05 ammonium bromide
50 mM borate and 20–25 mM tris 50 mM SDS or 20 mM Derivatized peptides and Liu and Novotny
and 10–25 mM disodium β-cyclodextrin and 1% THF angiotensin analogs (1990)
hydrogenphosphate, pH 9.50 or (v/v) or 15% methanol (v/v)
7.05 or additives such as 0.05 M
HTAB or 2–50 mM DTAB
10 mM sodium phosphate 7 mM cyclodextrin (α,β,γ ), Mycotoxins Holland and Sepaniak
acetonitrile (0–15%) (1993)
6 mM sodium borate 50 mM SDS, 50 mM
deoxycholate
200 mM borate buffer at pH 7.4 None Human serum proteins: Bossuyta et al. (1998)
albumin, 1-globulin,
2-globulin, β-globulin, and
gamma-globulin
5 mM ammonium acetate, pH 4 (Coupled to mass Phytochelatins in cell extracts Mounicou et al.
spectrometry) (2001)
15 mM sodium citrate (pH 2.1) 0.05% Tylose (methylhydrox- β−lactoglobulin and De Block et al. (2003)
yethylcellulose, 30,000 P, α-lactalbumin in milk
that is, viscosity of a 2% w/v product isolates
solution in H2 O at 20◦ C)
50 mM Tricine pH 8.0 20 mM NaCl and 2.5 mM Recombinant human Zhoua et al. (2004)
1,4-diaminobutane (DAB) granulocyte
colony-stimulating factor,
glycosylated and
nonglycosylated isoforms
10 mM Tricine, pH 5.5 0.01 M NaCl, 0.01 M sodium Human alpha-1-acid Lacunza et al. (2006)
acetate, 7 M urea, and 3.9 glycoprotein
mM 1,4-diaminobutane

40 mM sodium phosphate (pH None Cleavage products of peptide Purcell and
2.65) substrates for botulinum Hoard-Fruchey
neurotoxin (2007)
Peptides (Coated Capillaries)
150 mM phosphoric acid, pH 1.5 None Dipeptides McCormick (1988)
in polyvinylpyrrolidone-coated
capillary
1 M Acetate and 20% MeOH, MeOH Recombinant human Balaguer and Neusü
Polybrene coated capillary erythropoietin (rhEPO) (2006)
A.3.4 Proteins
20 mM borate, pH 8.25 None Model proteins Lauer and
McManigill (1986)
10 mM tricine, pH 8.22 20 mM KCl
10 mM tricine, pH 8.0 5.8 mM morpholine and 20 Biosynthetic human insulin Neilsen et al. (1989b)
mM NaCl derivatives and growth
hormone
50 mM glutamine-triethylamine, Vinyl-bound Model proteins Cobb et al. (1990)
pH 9.5 polyacrylamide-coated
capillaries
50 mM sodium phosphate, pH 7.0 Applied Biosystems coating Multiacetylated histone H4 Wiktorowicz and
reagent proteins Colburn (1990)
50 mM sodium borate None Human serum proteins Gordon et al. (1991)
decahydrate, pH 10.0
50 mM tricine, pH 8.0 30% methanol (v/v) Glycorproteins: recombinant Tran et al. (1991)
human
50 mM MES, pH 6.0 30% ethylene glycol (v/v) Erythropoietin
10 mM phosphate, pH 7.0 in None Basic proteins Towns and Regnier
capillaries coated with either (1991)
Tween 20/alkylsilane or Brij
10 mM citrate, pH 3.0 in 4 mM CHAPSO Standard proteins Swedberg (1993)
BSA-coated capillary
30 mM citrate, pH 3.0 in a None Standard proteins Kohr and Engelhard
polyacrylamide-coated capillary (1991)
(uncrosslinked)
50 mM sodium tetraborate, pH 25–50 mM LiCl Fluorescamine-labeled Guzman et al. (1992)
8.3 interferon and other proteins
100 mM borate, pH 9.0 1 mM diaminobutane Ovalbumin Landers et al. (1992c)
20 mM phosphate, pH 3.0 30 mM NaCl, 0.05% Basic proteins Gilges et al. (1992)
polyvinyl alcohol (w/w)
TES (NaOH), pH 7.0 0.2% SDS linear acrylamide Proteins (3–205 kDa) Werner et al. (1993)
(CGE)
SDS polyacrylamide gel Proteins Tsuji (1994a,b)
Shieh et al. (1994)
Iminodiacetic acid (IDA), 50 mM 10% Trifluoroethanol (TFE), Proteins and peptides Bossi and Righetti
0.5% hydroxyethylcellulose (1997)
(HEC); or 0.5% HEC and
6-8 M urea
Continued

50 mM aspartic acid 0.5% hydroxyethyl Alpha and beta-globin Capelli et al. (1997)
cellulose, 5% chains from human adult
trifluoroethanol, 1 Hb
zwitterionic detergent
(CHAPS)
Acetic acid Acetonitrile and hexane Heptathelical membrane Dong et al. (1997)
sulfonic acid proteins
Phytic acid Proteins Veraart et al. (1997)
Ammonium formate Trimethylamine (TEA) Proteins Lee and Desiderio (1997)
Phosphate buffer Guaran (neutral Basic proteins and drugs Liu et al. (1991)
polysaccharide)
50 mM phosphate with 11 11 mM sodium pentasulfate Protein content of a single Zhang et al. (2000)
mM sodium pentasulfate (SPS) cell
Acetic acid 10 mM hydroquinone Protein digests Moini et al. (1999)
Methanolic buffer Enzymic digest peptides Katayama et al. (2000)
(methanol/formic
acid/water = 60:20:20)
Glutamic acid 1 mM oligoamine Proteins/muscle Verzola et al. (2000)
(tetraethylene pentamine) acylphosphatase (AcP)
Iminodiacetic acid or TEPA Human alpha and beta Olivieri et al. (2000)
aspartic acid (tetraethylenepentamine) globin chains
Borate buffer SDS and PEO Proteins: Tseng et al. (2002)
microheterogeneites and
isoforms
Acetic acid OR ammonium Acetonitrile Human high density Deterding et al. (2002)
acetate OR ammonium apolipoproteins
bicarbonate OR sodium
bicarbonate/carbonate OR
phosphate OR tris OR
ammonium bicarbonate
Linear polyacrylamide Proteins Gomis et al. (2003)
(sieving)
Run buffer: Tris, 0.035 M
aspartic acid, 0.1% SDS,
4% acrylamide
1-Alkyl-3- Basic proteins Jiang et al. (2003)
methylimidazolium-based
ionic liquids
90-mM 1-ethyl-3-
methylimidazolium
tetrafluoroborate
(1E-3MI-TFB)
25 mM Borax pH 9.4 200 nM SSB (single-strand Basic proteins, separation Berezovski et al. (2003)
binding protein) for via Nonequilibrium
NECEEM Capillary Electrophoresis
of Equilibrium Mixtures
(NECEEM)
Acidic buffers Tetraalkylammonium and Peptides and proteins Quang et al. (2003)
tetrabutylammonium
cations
6 mM phosphate buffer Mexiletine (chiral CE) Human Serum Ablumin: Xu et al. (2004)
conformational change

SDS-pullulan (capillary Protein fingerprinting of Hu et al. (2004)
sieving electrophoresis) single mammal cells
followed by MECC
Formic acid (1 M, pH 1.78) Human metal binding Stutz et al. (2004)
for CZE followed by 125 proteins
mM ammonium formate
(pH 4) for tCITP
Phosphate buffer (50 mM) Heptanesulfonic acid (ion Proteins and peptides Miksik et al. (2004)
pairing agent)
Formic acid (pH 2.4, 50 Intact proteins Eriksson et al. (2004)
mM), sodium phosphate
buffer, ammonium
formate and ammonium
borate
20 mM phosphate buffer Pluronic F-127 Enzymic digests of insol. Miksik et al. (2004)
pH 2.5 Matrix proteins
Diethylenetriamine (DIEN) Basic proteins Corradini et al. (2005)
phosphate buffer
Ternary nonaqueous buffer: 12.5 mM ammonium Tryptic digests Assuncao et al. (2005)
60/30/10 v/v acetate
methanol/acetonitrile/acetic
acid
Triethanolamine and Microheterogeneity of Berkowitz and Zhong (2005)
phosphoric acid, pH 2.5 intact recombinant
glycoprotein
Replaceable cross-linked Lu et al. (2005)
polyacrylamide (rCPA)
MCRB (Moving Chem. Basic proteins Cao et al. (2005)
Reaction Boundary):
weak acidic run buffer
with alk. Sample buffer
Carrier ampholytes Tryptic digests Busnel et al. (2006)
Tris-phosphate 0.05% polyethyleneimine Proteins Sedlakova et al. (2006a)
0.1 M Phosphate pH 2.5 Pluronic F 127 Proteins, peptides, digests Sedlakova et al. (2006b)
Acid buffers: ammonium DDAB capillary coating Basic proteins Mohabbati and Westerlund
acetate, ammonium (2006)
hydroxyacetate,
phosphate
Borate pH 8.3 D-PEG Capsid proteins from Kremser et al. (2006)
dedecylpoly(ethylene rhinovirus
glycol ether)
Tris-HCl pH 8.2 “Smart aptamer” affinity MutS protein Drabovich and Krylov (2006)
probes
Tris-borate pH 10 SDS or PEO Microheterogeneous Huang et al. (2006)
proteins
50 mM Tris-HCl pH 8.5 Artificial gel antibodies Transferrin Takatsy et al. (2006)
30 mM ammonium acetate Mixtures of drugs, Fanali et al. (2006)
pH 5.5 peptides, tryptic digests of
proteins, biological fluids
MEKC run buffer: 20 mM Urinary CPIII and CPI Li and Huie (2006)
CAPS, 60 mM sodium
cholate, 20% v/v ACN
Continued

0.01 M Tricine, 0.01 M Intact forms of Alpha-1-acid Lacunza et al. (2006)
NaCl, 0.01 M sodium glycoprotein (AGP)
acetate, 7 M urea
20 mM Borax in heavy water Note: Online CE FTIR Model putrescine proteins Kulka et al. (2006)
detection
Barium tetraborate Tryptic digests Mendieta et al. (2006)
10 mM phosphate Gold nanoparticles Acidic and basic proteins Yu et al. (2006)
Poly Native proteins Zhang et al. (2006)
(N,N-dimethylacrylamide)-
grafted polyacrylamide
(self-coating sieving
polymer)
Tris-phosphate 0.05% (w/v) Proteins Sedlakova and Miksik
polyethlyeneimine (2006)
0.1 M Phosphate pH 2.5 Pluronic F 127 Proteins, peptides, digests Sedlakova et al. (2006)
1. Proteins and biogenic amines Chen et al. (2007)
TRIS-CHES-SDS-Dextran from mouse AtT-20 cell line
(sieving)
2. TRIS-CHES-SDS
(MECC)
10% acetic acid 0.5% chitosan (high Complex peptide mixtures Busnel et al. (2007)
molecular weight) with
0.1% acetic acid
100 mM borate pH 8.3 6 mM diaminobutate (DAB) Transferrin Bortolotti et al. (2007)
100 mM phosphate High concentration PDDAC: Cationic proteins Lin et al. (2007)
Poly(diallyldimethyl- Anionic proteins
ammonium chloride)
A.3.5 Nucleic Acids

50 mM phosphate, pH 7.0 5% ethylene glycol (v/v) Oligo dT ladder Kasper et al. (1988)
1 mM borate, pH 9.1 (CGE) 30% hydrolink (polymerized) Restriction fragments
100 mM trizma Cesium hydroxide Forensic DNA McCord et al. (1993)
100 mM boric acid, pH 8.7 0.1 mM EDTA
(CGE)
0.5% hydroxyethylcellulose
29 mM Tris, 68 mM HEPES (pH 0.5% HPMC (pretreatment); Single-stranded DNA Ren and Ueland
7.2); 15 mM Tris, 27 mM MES, 6% short-chain and linear (1999)
0.5 M EDTA (pH 6.4); 167 mM polyacrylamide (SLPA)
Tris, 33 mM boric acid (pH 9.0)
20 mM TAPS, pH 7.5 7 M urea; 4% HEC RNA analysis Saevels et al. (1999)
89 mM Tris-base [pH 8.3], 89 4 M urea; 1% PVP RNA analysis Khandurina et al.
mM boric acid, 2 mM EDTA (2002)
20 mM Tris, 10 mM phosphoric 4.5% HPC DNA analysis/determination Giovannoli et al.
acid, pH 7.3 of genetically modified (2004)
organisms
50 mM Tris-borate buffer 5% glycerol; 2% Single-strand conformation Endo et al. (2005)
methylcellulose polymorphism analysis
100 mM TB buffer (pH 8.0) 0.36 µg/mL CTAB; 0.5% Double-stranded DNA Lin and Chang (2006)
PEO

50 mM Tris-acetate (pH 8.2) 2 mM MgCl2 (gel free SNP analysis Drabovich and Krylov
separation) (2006)
40 mM Tris, 60 mM Mes, and 2 1.5% poly(N- DNA analysis Yu et al. (2006)
mM EDTA, pH 6.11 isopropylacrylamide)
(PNIPAM)
30 mM Trizma® Base, 100 mM 7 M urea; 8% PVP or 7% DNA sequencing Ekstrøm, and
TAPS, 1 M EDTA, pH 8.0 PVP, 1% PDMA Bjørheim (2006)
20 mM Tris-HCl, 2 mM KCl 0.03% Triton X-100 Nucleic acid aptamers Li et al. (2007)
50 mM sodium bicarbonate, pH 70 mM SDS DNA analysis Hua and Naganuma
9.0 (2007)
20 mM Tris, 9.5 mM 4% HEC DNA analysis/determination Sánchez et al. (2007)
orthophosphoric acid, 2 mM of genetically modified
EDTA, pH 7.3 (CGE) organisms
REFERENCES
Ackermans, M. T., Beckers, J. L., Everaerts, F. M., Koogland, H., and Tomassen, T. J. H. Determination of
sulphonamides in pork meat extracts by capillary zone electrophoresis, J. Chromatogr., 596, 101, 1992.
Altria, K. and Simpson, C. F. Analysis of some pharmaceuticals by high voltage capillary zone electrophoresis.
J. Pharm. Biochem. Med. Anal., 6, 801, 1988.
Altria, K. D., Assi, K. H., Bryant, S. M., and Clark, B. J. Determination of organic acid drug counter-ions by
capillary electrophoresis, Chromatographia, 44, 367, 1997.
Arrowood, S. and Hoyt, A. M. Determination of cimetidine in pharmaceutical preparanons by capillary zone
electrophoresis, J. Chromatogr., 586, 177, 1991.
Audalovic, N., Pohl, C. A., Rocklin, R. D., and Stillian, J. R. Determination of cations and anions by capillary
electrophoresis combined with suppressed conductivity detection, Anal. Chem., 65, 1470–1475, 1993.
Balaguer, E. and Neusü, C. Intact glycoform characterization of erythropoietin-alpha and erythropoietin-beta
by CZE-ESI-TOF-MS, Chromatographia, 64, 351, 2006.
Beck, W. and Engelhardt, H. Capillary electrophoresis of organic and inorganic cations with indirect UV
detection, Chromatographia, 33, 313, 1992.
Bossi, A. and Righetti, P. G. Generation of peptide maps by capillary zone electrophoresis in isoelectric
iminodiacetic acid. Electrophoresis, 18, 2012, 1997.
Bossuyta, X., Schiettekatte, G., Bogaerts, A., and Blanckaer, N. Serum protein electrophoresis by CZE 2000
clinical capillary electrophoresis system, Clin. Chem., 44: 749–759, 1998.
Burton, D. E., Sepaniak, M. J., and Maskarinec, M. P. Evaluation of the use of various surfactants in micellar
electrokinetic capillary chromatography, J. Chromatogr. Sci., 25, 514, 1987.
Camilleri, P., Okafo, G. N., Southan, C., and Brown, R.Analytical and micropreparative capillary electrophoresis
of peptides from calcitonin, Anal. Biochem., 198, 36, 1991.
Capelli, L., Stoyanov, A. V., Wajcman, H., and Righetti, P. G. Generation of tryptic maps of α-and β-globin
chains by capillary electrophoresis in isoelectric buffers. J. Chromatogr. A, 791, 313, 1997.
Cataldi, T. R. I., Nardiello, D., Carrara, V., Ciriello, R., and DeBenedetto, G. E. Assessment of riboflavin and
flavin content in common food samples by capillary electrophoresis with laser-induced fluorescence
detection, Food Chem., 82, 309, 2003.
Chadwick, R. R. and Hsieh, J. C. Separation of cis and trans double bond isomers using capillary zone
electrophoresis, Anal. Chem., 63, 2377, 1991.
Cobb, K.A., Dolnik, V., and Novotny, M. Electrophoretic separations of protein in capillaries with hydrolytically
stable surface structures, Anal. Chem., 62, 2478, 1990.
Cohen, A. S. High performance capillary electrophoresis of bases, nucleosides and oligonucleotides: retention
manipulation via micellar solutions and metal additives, Anal. Chem., 59, 1021, 1987.
Cohen, A. S., Najarian, D., Smith, J. A., and Karger, B. L. Rapid separation of DNA restriction fragments usiny
Crego, A. L., Marina, M. L., and Lavandera, J. L. Optimization of the separation of a group of antifungals by
capillary zone electrophoresis, J. Chromatogr. A, 917, 337, 2001.
De Block J. et al. Monitoring nutritional quality of milk powders: capillary electrophoresis of the whey protein
fraction compared with other methods. Intl. Dairy J., 13, 87, 2003.
Drabovich, A. P. and Krylov, S. N. Identification of base pairs in single-nucleotide polymorphisms by muts
protein-mediated capillary electrophoresis. Anal. Chem., 78, 2035, 2006.
Eder, A. R., Chen, J. S., and Arriaga, E. A. Separation of doxorubicin and doxorubicinol by cyclodextrin-
modified micellar electrokinetic capillary chromatography, Electrophoresis, 27, 3263, 2006.
Ekstrøm, P. O. and Bjørheim, J. Evaluation of sieving matrices used to separate alleles by cycling temperature
capillary electrophoresis, Electrophoresis, 27, 1878, 2006.
Endo, Y., Zhang, L., Katashima, R., Itakura, M., Doherty, E. A. S., Barron, A. E., and Baba, Y. Effect of
polymer matrix and glycerol on rapid single-strand conformation polymorphism analysis by capillary
and microchip electrophoresis for detection of mutations in K-ras gene. Electrophoresis, 26, 3380,
2005.
Evans III, L. and Collins, G. E. Separation of uranium (VI) and transition metal ions with 4-(2-thiazolylazo)
resorcinol by capillary electrophoresis, J. Chromatogr. A, 911, 127, 2001.
Florance, J. R., Konteatis, Z. D., Macielag, M. J., Lessor, R. A., and Galdes, A. Capillary zone electrophore-
sis studies of motilin peptides, effect of charge, hydrophobicity, secondary structure and length,
J. Chromatogr., 559, 391, 1991.
Fujiwara, S. and Honda, S. Determination of cinnamic acid and its analogues silica capillary tube. Anal. Chem.,
58, 1811, 1986.
Fujiwara, S. Effect of addition of organic solvent and the separation of positional isomers in high voltage
Fung, Y. S. and Lau, K. M. Separation and determination of cations in beverage products by capillary zone
electrophoresis, J. Chromatogr. A, 1118, 144, 2006.
Gaus, H.-J., Beck-Sickinger, A. G., and Bayer, E. Optimization of capillary electrophoresis of mixtures of basic
peptides and comparison with HPL, Anal. Chem., 65, 1399–1405, 1993.
Gilges, M., Hasmann, H., Kleemib, M. H., Motsch, S. R., and Schomburg, G. CZE separations of basic proteins
at low pH in fused silica capillaries with surfaces modified by silane derivitization and/or adsorption
of polar polymers, J. High Resolut. Chromatogr., 15, 452, 1992.
Giovannoli, C., Anfossi, L., Tozzi C., Giraudi, G., and Vanni, A. DNA separation by capillary electrophoresis
with hydrophilic substituted celluloses as coating and sieving polymers. Application to the analysis of
genetically modified meals, J. Sep. Sci., 27, 1551, 2004.
Gong, S. X., Liu, F. J., Li, W., Gao, F., Gao, C. J., Liao, Y. P., and Liu, H. W. Separation of hydropho-
bic solutes by organic-solvent-based micellar electrokinetic chromatography using cation surfactants,
J. Chromatogr. A, 1121, 274, 2006.
Gordon, M. J., Zare, R. N., Lee, K. J., and Arias, A. A. Protocol for resolving protein mixtures in capillary zone
Guo, B. Y., Wen, B., Shan, X. Q., Zhang, S. Z., and Lin, J. M. Separation and determination of phthalates by
micellar electrokinetic chromatography, J. Chromatogr. A, 1095, 189, 2005.
Guzman, N. A., Moschera, J., Bailey, C. A., Iqbal, K., and Malick, A. W. Assay of protein drug substances
present in solution mixtures by fluorescence denvitisation and capillary electrophoresis, J. Chromatogr,
598, 123, 1992.
Holland, R. D. and Sepaniak, M. J. Qualitative analysis of mycotoxins using micellar electrokinetic capillary
chromatography, Anal. Chem., 65, 1140–1146, 1993.
Honda, S., Taga, A., Kakehi, K., Koda, S., and Okamoto, Y. Determination of cefixime and its metabolites by
high performance capillary electrophoresis, J. Chromatogr., 590, 364, 1992.
Hua, N. and Naganuma T. Application of CE for determination of DNA base composition. Electrophoresis, 28,
366, 2007.
Huang, X., Luckey, J. A., Gordon, M. J., and Zare, R. N. Quantitative analysis of low molecular weight
carboxylic acids by capillary zone electrophoresis/conductivity detection, Anal. Chem., 61, 766, 1989.
Iwata, Y. T., Garica, A., Kanamori, T., Inoue, H., Kishi, T., and Lurie, I. S. The use of a highly sulfated
cyclodextrin for the simultaneous chiral separation of amphetamine-type stimulants by capillary
electrophoresis. Electrophoresis, 23, 1328, 2002.
Jones, W. R. and Jandik, P. Various approaches to analysis of difficult sample matrices of anions using capillary
ion electrophoresis, J. Chromatogr., 608, 385–393, 1992.
Kasper, T. J., Melera, M., Gozel, P., and Browniee, R. G. Separation and detection of DNA by capillary
Katata, L., Nagaraju, V., and Crouch, A. M. Determination of ethylenediaminetetraacetic acid, ethylenedi-
aminedisuccinic acid and iminodisuccinic acid in cosmetic products by capillary electrophoresis and
high performance liquid chromatography, Analytica Chimica Acta, 579, 177, 2006.
Khandurina, J., Chang, H., Wanders, B., and Guttman, A. Automated High-throughput RNA analysis by
capillary electrophoresis, Biotechniques, 32, 1226, 2002.
Kohr, H. and Englehardt, H. Capillary electrophoresis with surface coated capillaries, J. Microcolumn Sep., 3,
491, 1991.
Krueger, R. J., Hobbs, T. R., Mihal, K. A., Therani, J., and Zeece, M. G. Analysis of endoproteinase ArgC action
on adrenoconicotrophic hormone by capillary electrophoresis and reverse phase HPLC, J. Chromatogr.,
543, 451, 1991.
Lacunza I. et al. CZE of human alpha-1-acid glycoprotein for qualitative and quantitative comparison of samples
from different pathological conditions, Electrophoresis, 27, 4205, 2006.
Landers, J. P., Schuchard, M D., Subramaniam, S., Sismelich, T. and Spelsberg, T. C. High-performance
capillary electrophoresis analysis of chloramphenicol acetyl transferase activity, J. Chromatogr., 603,
247–257, 1992a.
Landers, J. P., Oda, R. P., and Schuchard, M. D. Separation of boron-complexed diol compounds using high-
performance capillary electrophoresis, Anal. Chem., 64, 2846–2851, 1992b.
Landers, J. P., Oda, R. P., Madden, B. J., and Spelsberg, T. C. High-performance capillary electrophoresis of
glycoproteins: the use of modifiers of electroosmotic flow for analysis of microheterogeneity, Anal.
Biochem., 205, 115–124, 1992c.
Landers, J. P., Oda, R. P., Liebenow, J. A., and Spelsberg, T. C. Utility of high performance capillary elec-
trophoresis for monitoring peptide homo- and hetero-dimer formation, J. Chromatogr., 652, 109–117,
1993.
Lauer, H. H. and McMangill, D. Capillary zone electrophoresis of proteins in untreated fused silica tubing,
Anal. Chem., 58, 166, 1986.
Lecoq, F., Leuratti, C., Marafante, E., and DiBase, S. Analysis of nucleic acid derivatives by micellar
electrokinetic capillary chromatography, J. High Resolut. Chromatogr., 14, 667, 1991.
Li, J., Zhao, F., and Ju, H. Simultaneous determination of psychotropic drugs in human urine by capillary
electrophoresis with electrochemiluminescence detection, Analytica Chimica Acta, 575, 57, 2006.
Li, T., Li, B., and Dong, S. Aptamer-based label-free method for hemin recognition and DNA assay by capil-
lary electrophoresis with chemiluminescence detection. Anal. Bioanal. Chem., [Epub ahead of print],
2007.
Li, X., Jin, W., and Weng, Q. Separation and determination of homovanillic acid and vanillylmandelic acid by
capillary electrophoresis with electrochemical detection, Analytica Chimica Acta, 461, 123, 2002.
Lin, Y. W. and Chang, H. T. Analysis of double-stranded DNA by capillary electrophoresis using poly(ethylene
oxide) in the presence of hexadecyltrimethylammonium bromide. J. Chromatogr. A, 1130, 206, 2006.
Little, M. J., Aubry, N., Beaudoin, M. E., Goudreau, N., and LaPlante, S. R. Quantifying trifluoroacetic acid as
a counterion in drug discovery by 19F NMR and capillary electrophoresis, J. Pharm. Biomed. Anal.,
43, 1324, 2007.
Liu, J. and Novotny, M. Capillary electrophoretic separation of peptides using micelleforming compounds and
cyclodextrins as additives, J. Chromatogr., 519, 189, 1990.
Liu, Q., Inoue, T., Kirchhoff, J. R., Huang, C., Tillekeratne, L. M. V., Olmstead, K., and Hudson, R. A. Chiral
separation of highly negatively charged enantiomers by capillary electrophoresis, J. Chromatogr. A,
1033, 349, 2004.
Liu, Y. and Chan, K. F. J. High performance capillary electrophoresis of gangliosides, Electrophoresis, 12, 402,
1991.
Lu, W. and Cassidy, R. M. Evaluation of ultramicroelectrodes for the detection of metal ions separated by
capillary electrophoresis, Anal. Chem., 65, 1649–1653, 1993.
Lurie, I. S. Micellar electrokinetic capillary chromatography of the enantiomers of amphetamine, metham-
phetamine and their hydroxyphenethylamine precursors, J. Chromatogr., 605, 269, 1992.
McCord, B. R., Jung, J. M., and Holleran, E. S. High resolution CE of forensic DNA using a non-gel sieving
buffer, J. Liquid Chromatogr., 16, 1963–1981, 1993.
McCormick, R. M. Capillary zone electrophoresis of peptides and proteins using low pH buffers in modified
silica capillaries, Anal. Chem., 60, 2322, 1988.
Mejia, E., Ding, Y. S., Mora, M. F., and Garcia, C. D. Determination of banned sudan dyes in chili powder by
capillary electrophoresis, Food Chem., 102, 1027, 2007.
Michaelis, L., Electric transport of enzymes, malt distaste and pepsin, Biochemische Zeitschrift, 17, 1909,
pp. 231–234.
Mikaye, M., Shibukawa, A., and Nakasgawa, T. Simultaneous determination of creatinine and uric acid in
human plasma and urine by micellar electrokinetic chromatography, J. High Resolut. Chromatogr., 14,
181, 1991.
Molina-Mayo, C., Hernandez-Borges, J., Borges-Miguel, T. M., and Rodriguez-Delgado, M. A. Determination
of pesticides in wine using micellar electrokinetic chromatography with UV detection and sample
stacking, J. Chromatogr. A, 1150, 348, 2007.
Morin, Ph., Villard, F., Quinsac, A., and Dreux, M. Micellar electrokinetic capillary chromatography of glu-
cosinolates and desulfoglucosinolates with a cationic surfactant, J. High Resolut. Chromatogr., 371,
1992.
Mosher, R. A. The use of metal ion-supplemented buffers to eliminate the resolution of peptides in capillary
zone electrophoresis, Electrophoresis, 11, 765, 1990.
Mounicou, S. et al. Determination of phytochelatins by capillary zone electrophoresis with electrospray tandem
mass spectrometry detection (CZE-ES MS/MS), Analyst, 126, 624, 2001.
Neilsen, R. G., Riggin, R. M., and Richards, E. C. Capillary zone electrophoresis of peptide fragments from
trypsin digestion of biosynthetic human growth hormone, J. Chromatogr., 480, 343, 1989a.
Neilsen, R. G., Sittampalam, G. S., and Richard, E. C. Capillary zone electrophoresis of insulin and growth
hormone, Anal. Biochem., 177, 20, 1989b.
Nevado, J. J. B., Flores, J. R., Penalvo, G. C., and Dorado, R. M. R. Determination of ibuprofen and tetrazepam
in human urine by micellar electrokinetic capillary chromatography, Anal. Bioanal. Chem., 384, 208,
2006.
Ng, C. L., Lee, N. H., and Li, S. F. Y. Systematic optimization of capillary electrophoretic separation of
sulphonamides, J. Chromatogr., 598, 133, 1992.
Nguyen, A. L., Luong, J. H. T., and Masson, G. Determination of nucleotides in fish tissue using capillary
Nielsen, M. W. F. Chiral separation of basic drugs using cyclodextrin-modified capillary zone electrophoresis,
Anal. Chem., 65, 885–893, 1993.
Nielsen, M. W. and Mensink, M. J. A. Separation of benzothiazole sultenamides using micellar electrokinetic
capillary chromatography, J. High Resolut. Chromatogr., 14, 417, 1991.
Nishi, H. and Terabe, S. Applications of micellar electrokinetic chromatography to pharmaceutical analysis,
Nishi, H., Tsumagari, N., and Terabe, S. Effects of tetraalkylammonium salts on micellar electrokinetic
chromatography of ionic substances, Anal. Chem., 61, 2434, 1989.
Northrop, D. M., Martine, D. E., and MacClehan, W. A. Separation and identification of organic gunshot and
explosive constituents by micellar electrokinetic capillary chromatography, Anal. Chem., 63, 10338,
1991.
Novotny, M. V., Cobb, K. A., and Liu, J. Recent advances in capillary electrophoresis of proteins, peptides and
amino acids, Electrophoresis, 11, 735, 1990.
O’Shea, T. J., Lunte, S. M., and LaCourse, W. R. Detection of carbohydrates by CE with pulsed amperometric
detection, Anal. Chem., 65, 948–951, 1993.
Okafo, G. N., Brown, R., and Camilleri, P. Some physico-chemical properties that make D.O-based buffer
solutions useful media for capillary electrophoresis, J. Chem. Soc. Chem. Commun., 864, 1991.
Ong, L. P., Ng, C. L., Lee, N. H., and Li, S. F. Y. Determination of antihistamines in pharmaceutical by capillary
Orejuela, E. and Silva, M. Rapid determination of aniline metabolites of chlorpropham in potatoes by micellar
electrokinetic chromatography using negative-charged mixed micelles and laser-induced fluorescence
detection, Electrophoresis, 26, 2991, 2005.
Otsuka, K., Terabe, S., and Ando, T. Electrokinetic chromatography with micellar solutions: separation of
phenylthiohydantoin amino acids, J. Chromatogr., 332, 219, 1985.
Palmer, J., Atkinson, S., Yoshida, W. Y., Stalcup, A. M. and Landers, J. P. Charged chelate-capillary
electrophoresis of endogenous corticosteroids, Electrophoresis, 19, 3045–3051, 1998.
Pereira, E. A., Micke, G. A., and Tavares, M. F. M. Determination of antiretroviral agents in human serum by
capillary electrophoresis, J. Chromatogr. A, 1091, 169, 2005.
Prusik, Z., Kasicka, V., Mudra, P., Stepanek, J., Smekal, O., and Hlavacek, J. Correlation of capillary zone
electrophoresis with continuous flow zone electrophoresis: application to the analysis and purification
of synthetic growth hormone releasing peptide, Electrophoresis, 932, 1990.
Purcell, A. L. and Hoard-Fruchey, H. M. A capillary electrophoresis method to assay catalytic activity of

botulinum neurotoxin serotypes: Implications for substrate specificity, Anal. Biochem., 366, 207, 2007.
Qi, X. H., Mi, J. Q., Zhang, X. X., and Chang, W. B. Design and preparation of novel antibody system and
application for the determination of heroin metabolites in urine by capillary electrophoresis, Analytica
Chimica Acta, 551, 115, 2005.
Ren, J. and Ueland, P. M. Temperature and pH effects on single-strand conformation polymorphism analysis
by capillary electrophoresis, Hum. Mutat., 13, 458, 1999.
Roach, M. C., Gozel, P., and Zare, R. N. Determination of methotrexate and its major metabolite, 7-
hydroxymethotreaxate using capillary zone electrophoresis and laser induced fluorescence detection,
J. Chromatograph., 426, 129, 1988.
Romano, J., Jandik, P., Jones, W. R., and Jackson, P. E. Optimisation of inorganic capillary electrophoresis for
the analysis of anionic solutes in real sample, J. Chromatogr., 545, 411, 1991.
Ruis-Calero, V., Puignou, L., and Galceran, M. T. Determination of glycosaminoglycan monosaccharides by
capillary electrophoresis using laser-induced fluorescence detection, J. Chromatogr. B, 791, 193, 2003.
Saevels, J., Van Schepdael, A., and Hoogmartens, J. Capillary Electrophoresis of RNA Oligonucleotides:
Catalytic Activity of a Hammerhead Ribozyme, Anal. BioChem., 266, 93, 1999.
Sakudinskaya, I. K., Desirodero, C., Nardi, A., and Fanali, S. Micellar electrokinetic chromatographic
study of hydroquinone and some of its ethers. Determination of hydroquinone in skin toning cream,
J. Chromatogr., 596, 95, 1992.
Salomon, D. R. and Romano, J. Applications of capillary ion electrophoresis in the pulp and paper industry,
J. Chromatogr., 602, 219, 1992.
Sánchez, L., González, R., Crego, A. L., and Cifuentes, A. A simple capillary gel electrophoresis approach for
efficient and reproducible DNA separations. Analysis of genetically modified soy and maize, J. Sep.
Sci., 30, 579, 2007.
Saria, A. Identification of alpha and beta series of calcitonin gene-related peptide in the rat amygdala after
separation with capillary zone electrophoresis. 573, 219, 1992.
Sedlakova, P., Svobodova, J., Miksik, I., and Tomas, H. Separation of poly(amidoamine) (PAMAM) dendrimer
generations by dynamic coating capillary electrophoresis. J. Chromatogr. B, 841, 135–139, 2006a.
Sedlakova, P., Svobodova, J., and Miksik, I. Capillary electrophoresis of peptides and proteins with plug of
Pluronic gel. J. Chromatogr. B, 839, 112–117, 2006b.
Shieh, P. C. H., Hoang, D., Guttman, A., and Cooke, N. Capillary sodium dodecyl sulfate gel electrophoresis
of proteins. I. Reproducibility and stability. J. Chromatogr. A, 676, 219, 1994.
Swedberg, S. The impact of column technology on protein separations in OT capillary electrophoresis: past
lessons, future promises, CRC Handbook of Capillary Electrophoresis: A Practical Approach, Chapter
19, 1993.
Tanaka, H. W., Kanabe, T., Yameda, Y., and Semba, T. Capillary electrophoretic monitoring for C-terminal
fragment identification, J. High Resolut. Chromatogr., 14, 491, 1991.
Terabe, S., Otsuka, K., Ichikawa, K., Tsuchiva, A., and Ando, T. Electrokinetic separations with micellar
solutions and tubular capillaries, Anal. Chem., 56, 113, 1984.
Terabe, S., Ozuki, H., Otsuka, K., and Ando, T. Electrokinetic chromatography with 2-0-carboxymethyl-
cyclodextrin as a moving stationary phase, J. Chromatogr., 332, 211, 1983.
Terabe, S., Yashima, T., Tanaka, M., and Araki, M. Separation of oxygen isotopic benzoic acids by capillary
zone electrophoresis based on isotope effects on the dissociation of the carboxyl group, Anal. Chem.,
60, 1673, 1988.
Thibault, P., Pleasance, S., and Laycock, M. V. Analysis of paralytic shellfish poisons by capillary
Tonin, F. G., Jager, A. V., Micke, G. A., Farah, J. R. S., and Tavares, M. F. M. Optimization of the separation of
flavonoids using solvent-modified micellar electrokinetic chromatography, Electrophoresis, 26, 3387,
2005.
Towns, J. K. and Regnier, P. E. Capillary electrophoretic separations of proteins using non-ionic surfactant
coatings, Anal. Chem., 63, 1126, 1991.
Tran, A. D., Pak, S., List, P. J., Huynh, O. T., Ryall, R. R., and Lane, P. A. Separation of carbohydrate-mediated
microheterogeneity of recombinant human erythropoietin by free solution capillary electrophoresis, J.
Chromatogr., 542, 459, 1991.
Tseng, H. M., Li, Y., and Barrett, D. A. Profiling of amine metabolites in human biofluids by micellar elec-
trokinetic chromatography with laser-induced fluorescence detection, Anal. Bioanal. Chem., 388, 433,
2007.
Tsuda, T., Kobayashi, Y., Hori, A., Matsumoto, T., and Suzuki, O. Separation of polyamines in rat tissue by
capital electrophoresis, J. Microcol. Sep., 2, 21, 1990.
Tsuji, K. Sodium dodecyl sulfate polyacrylamide gel- and replaceable polymer-filled capillary electrophoresis
for molecular mass determination of proteins of interest. J. Chromatogr. A, 666, 294, 1994a.
Tsuji, K. Factors affecting the performance of SDS gel-filled capillary electrophoresis. J. Chromatogr. A, 661,
257, 1994b.
Van de Goor, T. A. A. M., Janssen, P. S. L., Van Nispen, J. W., Van Zeeland, M. J. M., and Everaerts, F. M.
Capillary electrophoresis of peptides: analysis of adrenocorticotropic hormone-related fragments, 545,
379, 1991.
Viskari, P. J. and Colyer, C. L. Separation and quantitation of phycobiliproteins using phytic acid in capillary
electrophoresis with laser-induced fluorescence detection, J. Chromatogr. A, 972, 269, 2002.
Volpi, N. Application of high-performance capillary electrophoresis to the purification process of Escherichia
coli K4 polysaccharide, J. Chromatogr. B, 811, 253, 2004.
Vorndran, A. E., Oefner, P. J., Scherz, H., and Bonn, G. K. Indirect UV detection of carbohydrates in capillary
zone electrophoresis, Chromatographia, 33, 163, 1992.
Walbroehl, Y. and Jorgenson, J. W. Capillary zone electrophoresis of neutral molecules by solvophobic
association with tetraalkylammonium ions, Anal. Chem., 58, 479, 1988.
Walbroehl, Y. On-column UV absorption detector for open tubular capillary zone electrophoresis, J.
Chromatogr., 315, 135, 1984.
Wallingford, R. A. and Ewing, A. G. Separation of serotonin from catechols by capillary zone electrophoresis
using electrochemical detection. Anal. Chem., 61, 98, 1989.
Wang, S. P., Fu, M. D., and Wang, M. H. Separation mechanism and determination of flavanones with capillary
electrophoresis and high-performance liquid chromatography, J. Chromatogr. A, 1164, 306, 2007.
Weinberger, R. and Lurie, I. S. Micellar electrokinetic capillary chromatography of illicit drug substances, Anal.
Chem., 63, 827, 1991.
Werner, W. E., Demorgst, D. M., Stevens, J., and Wiktorewicz, J. E. Site dependent separation of proteins
denatured in SDS by CE using a replaceable sieving matrix, Anal. Biochem., 212, 253–258, 1993.
Wernly, P. and Thormann, W. Determination of morphin-3-glucuronide in human urine by capillary zone
electrophoresis and micellar electrokinetic capillary chromatography, J. Chromatogr., 616, 305–310,
1993.
Whitaker, G., Kincaid, B. J., VanHoof, N., Regan, F., Smyth, M. R., and Leonard, R. G. An investigation into the
sample preparation procedure and analysis of cyanoacrylate adhesives using capillary electrophoresis,
Intl. J. Adh. Adhes., 27, 604, 2007.
Wickramanayake, P. U., Tran, T. C., Hughes, J. G., Macka, M., Simpson, N., and Marriott, P. J. Simultaneous
separation of nitrofuran antibiotics and their metabolites by using micellar electrokinetic capillary
chromatography, Electrophoresis, 27, 4069, 2006.
Wiktorowicz, J. E. and Colburn, J. C. Separation of cationic proteins via charge reversal in capillary
Xu, H., Yu, X., and Chen, H. Enantiomeric separation of basic drugs with partially filled serum albumin as
chiral selector in capillary electrophoresis. Anal. Sci., 20, 1409–1413, 2004.
Yang, Y. H., Wu, W. Y., Yeh, H. H., and Chen, S. H. Simultaneous determination of cefepime and vancomycin
in plasma and cerebrospinal fluid by MEKC with direct sample injection and application for bacterial
meningitis, Electrophoresis, 28, 1788, 2007.
Yeo, S. K., Ong, C. P., and Li, S. F. Y. Optimisation of high performance capillary electrophoresis of plant
growth regulators usiny the overlapping resolution mapping scheme, Anal. Chem., 63, 2222, 1991.
Yu, S. B., Zhou, P., Feng, A. R., Shen, X. D., Xhang, Z., and Hu, J. M. pH effect on dynamic coating for
capillary electrophoresis of DNA. Anal. Bioanal. Chem., 385,730, 2006.
Zhang, Z., Krylov, S., Arriaga, E. A., Polakowski, R., and Dovichi, N. J. One-dimensional protein analysis of
an HT29 human colon adenocarcinoma cell. Anal. Chem., 72, 318–322, 2000.
Zhoua, G.-H. et al. Characterization of recombinant human granulocyte colony stimulating factor (rHuG-
CSF) by capillary zone electrophoresis, capillary isoelectric focusing electrophoresis and electrospray
ionization mass spectrometry. J. Pharm. Biomed. Anal., 35, 425, 2004002E.
2 Protein Analysis by Capillary
Electrophoresis
James M. Hempe
CONTENTS
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
2.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
2.2.1 Current Status of CE for Protein Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
2.2.2 The Challenges of Systems Biology and Proteomics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
2.3 Theoretical Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
2.3.1 Protein Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
2.3.2 CE Separation Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
2.3.3 Instruments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
2.3.4 Capillaries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
2.3.5 Separation Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
2.3.6 Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
2.3.7 Sample Preparation and Fraction Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
2.4 Practical Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
2.5 Analysis of Protein Properties by CZE and CSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
2.5.1 Identification of Beta Globin Isoforms by CIEF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
2.5.2 SP 1 Mobility Shift Assay by CAE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
2.5.3 Analysis of Free Prostate-Specific Antigen Isoforms by Capillary Zone
Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
2.5.4 Analysis of Histone H1.5 Isoforms by Offline
Multidimensional RPHPLC-CZE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
2.5.5 Online Multidimensional Protein Analysis by CSE-MEKC . . . . . . . . . . . . . . . . . . . . . . . 99
Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
2.1 INTRODUCTION
Almost any analytical protein application can be accomplished using capillary electrophoresis (CE).
CE can be used to characterize protein properties, identify specific protein isoforms, quantify expres-
sion levels, study protein interactions with other biomolecules or map the distribution of proteins
in one or more analytical dimension. Yet, protein analysis arguably remains the least widely used
major CE application. Compared to analysis of nucleotides or small organic compounds, protein
analysis by CE has faced greater technical challenges and more persistent competition from alterna-
tive separation and assay techniques. Although many protein applications are exquisitely or uniquely
accomplished with CE, others are still better achieved using conventional analytical methods (e.g., gel
electrophoresis, enzyme-linked immunosorbent assay (ELISA), or column chromatography) or other
75
advanced technologies (e.g., high-performance liquid chromatography (HPLC), mass spectrometry

(MS), or protein microarray). If an investigator has a small number of specific protein applications to
perform over and over again, CE’s automation and low reagent costs make it a very attractive alter-
native. If an investigator wants a versatile system that can be used to analyze more than just proteins,
opportunities abound on multimode open analytical CE platforms simply by changing the capillary,
separation medium, and/or detector. High resolution, versatility, automation, small sample volume
requirements, and minimal waste production are the hallmarks of CE for all applications. Poor repro-
ducibility and lack of sensitivity and specificity are traditional weaknesses of CE for protein analysis.
The purpose of this chapter is to provide an overview of how CE is used for protein analysis and how
it might be used in the future. CE instrument and method development will be strongly influenced by
systems biology and the search for biomarkers of complex diseases. Recent technological advances
that improve capillary performance and enhance detection of proteins suggest that CE will play a
major role in the evolving field of proteomics.
2.2 BACKGROUND
Capillary electromigration techniques can be effectively applied for protein analysis in all areas of
biology. Basic theoretical principles and applications of CE are discussed here in the context of
human biology and disease but are similar whether the analytical focus is plants or other animals.
Terminology used to describe the separation of proteins by CE will conform to the recommendations
of the Analytical Chemistry Division of the International Union of Pure and Applied Chemistry
(IUPAC) [1] except that CE will be used according to tradition as a collective term for all capillary
electromigration techniques. Recent progress in the use of CE for a wide variety of protein appli-
cations has been documented in a periodic series of reviews by Dolnik and colleagues [2–4]. Other
recent reviews specifically related to proteins [5], application of specific CE techniques to proteins
and other molecules [6–10], and the use of CE in systems biology and proteomics [11,12] are also
recommended.
2.2.1 CURRENT STATUS OF CE FOR PROTEIN ANALYSIS

Capillary electrophoresis has found widespread use for protein analysis in the pharmaceutical [13,14]
and food science [15–17] industries, primarily for process quality control in determining the purity
of natural and synthetic proteins, peptides, and monoclonal antibodies. Dedicated CE instruments
are commercially available for clinical diagnostics based on hemoglobin variants or serum and urine
proteins [18–22]. Future development of CE in the pharmaceutical, food science, and diagnostic
industries appears strongly linked to advances in microchip CE technology and its promise of high-
throughput, low-cost analyses [23–25]. CE is not a standard tool in most biology or biochemistry
research laboratories. One reason is that competing technologies were already entrenched when
CE was first developed. Another is that CE has low UV/Vis detection sensitivity and lacks speci-
ficity for analysis of proteins in complex biological matrices. CE is also consistently outperformed
by conventional techniques in applications that require fractionation and recovery of proteins for
further study [26].
Significant advances have been made in commercial CE instrumentation since the last edition of
this book was published. Open analytical systems are more reliable and dedicated microchip systems
are commercially available. High-throughput, multicapillary platforms with dedicated proteomic
applications are either on the market or nearing final development. Engineering advances include
better microfluidics, better optics and electronics, wider range of detection options, and improved
software for instrument control and for data processing, analysis and reporting. Newer coated cap-
illaries allow for better control of electroosmotic flow (EOF) and reduce protein–wall interactions
in ways that markedly improve the reproducibility and efficiency of CE analyses. The development
Protein Analysis by Capillary Electrophoresis 77
of multidimensional assay systems and advances in protein detection by MS and fluorescence has
improved the sensitivity and specificity of CE for analysis of proteins in the complex biological
matrices that are the targets of proteomic studies. More widespread use of CE for protein analysis
might be predicted as biomedical research turns more from the study of individual proteins to large-
scale assessment of the human proteome. What role CE will play in the evolving field of proteomics
depends on how effectively CE competes with or complements other advanced protein separation
technologies, and on how proteomics is applied to the study of complex biological processes.
Some CE techniques are better suited than others for analysis of proteins. A wide variety of appli-
cations have been developed for protein analysis by capillary zone electrophoresis (CZE), capillary
isoelectric focusing (CIEF), capillary sieving electrophoresis (CSE), and micellar electrokinetic chro-
matography (MEKC). All of these techniques are excellent for high-resolution analysis of specific
polypeptides in uncomplicated matrices, which makes CE ideal for use in process quality control
and characterization of purified proteins. CZE and CIEF are also excellent for the analysis of specific
proteins in complex biological matrices if the protein or proteins of interest are naturally abundant or
readily extracted from biological samples before analysis. Hemoglobin variants, for example, can be
directly analyzed in crude hemolysates using commercially available on-capillary detection systems
because globins are abundant and do not precipitate at the higher concentrations required for UV/Vis
detection. Analysis of less abundant or less soluble proteins often require sample prefractionation,
concentration and/or derivatization to enhance detection sensitivity. Like open analytical HPLC plat-
forms, a single multimode open analytical CE platform can be used to accomplish different protein
applications depending on the choice of capillary, separation medium and detection system. Cap-
illaries, reagents, and detectors can be readily interchanged, making it simple to switch from one
protein application to another. In general, CE offers superior separation efficiency compared to more
conventional protein separation techniques when using CZE, CIEF or MEKC. CSE performed on
an open analytical CE platform offers relatively fewer advantages over conventional size separation
methods beyond those generally characteristic of all CE techniques (i.e., automation, low sample
volume requirement, etc.). In contrast, automated microchip CSE represents a significant time- and
labor-saving improvement over conventional gel electrophoresis systems but the advantages come
at a price. High instrument start-up costs and the need for more advanced operator expertise are
obstacles to more widespread CE use as is the fact that CE methods are essentially nonpreparative
and fraction collection remains technically challenging.
A significant barrier to the use of CE for protein analysis in many biology and biochemistry
research laboratories is the lack of specificity and flexibility provided by conventional polyacry-
lamide gel electrophoresis, especially when combined with Western blotting. Commercially available
reagents and prepoured native, denaturing, isoelectric focusing and two-dimensional (2D) gels offer
a wide and usually adequate range of separation efficiencies at relatively low cost per sample, and it
requires little expense or expertise to introduce conventional gel electrophoresis into any laboratory.
Western blotting is generally as specific as the primary antibody employed, but that functional-
ity is conserved in immunoassays performed by capillary affinity electrophoresis (CAE). A main
difference between CAE and Western blotting is that Western blotting is usually performed non-
competitively with enzyme-linked secondary antibody detection, and the target protein can be easily
changed by switching from one primary antibody to another. Other than the relatively straightfor-
ward need to optimally titrate primary and secondary antibody concentrations, most other analytical
aspects of Western blotting are consistent regardless of the protein of interest. Moreover, antibodies
and reagent kits are widely and inexpensively available commercially. So are gel documentation
systems that feature sensitive chemiluminescence and/or phosphorescence detection and software
that makes data analysis simple. In contrast, most CAE protein immunoassays are performed non-
competitively using fluorescent-labeled primary antibodies or competitively using labeled peptides
as competing antigens [27], and it is not as easy to switch from one protein immunoassay to
another. Automated immunoassay of specific proteins in a microchip or multicapillary format offers
many advantages over traditional Western blotting in terms of separation efficiency and throughput
potential. However, ELISA and protein microarrays [28] competitively offer similar advantages for
many protein applications.
A PubMed search of research literature published over the last 10 years using CE as a search term
consistently identified approximately 1000 articles during each year of that period. The proportion of
articles identified by adding “protein” as a search term was consistently about one-fourth of that total,
suggesting sustained interest in CE for protein analysis. Continued evolution of CE immunoassays,
microchip platforms and multidimensional systems that incorporate CE with MS and other advanced
detection technologies can be expected to increase the use of CE for protein analysis as biomedical
research turns its attention from genomics to proteomics and demand for high-throughput protein
analysis increases.
2.2.2 THE CHALLENGES OF SYSTEMS BIOLOGY AND PROTEOMICS

To study proteins is to study biology, the focus of which is increasingly shifting from individual
molecules to systems. Systems biology uses top-down and bottom-up approaches to integrate life
science information and better understand the control and operation of biological processes in health
and disease [29]. Top-down systems biology is systemic-data driven and identifies molecular inter-
action networks on the basis of correlated molecular behavior by integrating data from the genome,
transcriptome, proteome, and metabolome. Bottom-up systems biology examines the mechanisms
through which functional molecular properties arise and constructs detailed mechanistic models of
biological processes that can be used to predict system properties. The systems biology approach is
more than just a neologistic exercise that combines “–omics” with existing words. It is a conceptual
framework that permits consideration of complex systems in a way that begins to encompass the
enormous amount of biological variation inherent to living organisms.
Proteomics is the systematic inventory of protein expression levels, posttranslational modifi-
cations and interactions with other biomolecules that along with genomics, transcriptomics and
metabolomics is an integral part of the systems biology approach. Unlike the relatively static genome,
the proteome is a dynamic entity that changes in different physiological states (e.g., health versus
disease) and biological compartments (e.g., fluids, cells, tissues, or organisms). The human genome
is estimated to contain 30,000–40,000 genes that by alternative splicing or other means code for the
production of around 100,000 different proteins [30]. These include enzymes, structural proteins,
peptide hormones, molecular transporters, signal transduction elements, transcription activators, and
a host of other functionally different proteins. More than two million polymorphic sites are predicted
in the human genome [31]. Consequently, many or most of the proteins produced by the human
genome exist in multiple primary sequence isoforms in human populations due to allelic polymor-
phism. Most, if not all, also exist in multiple chemically modified isoforms that are alternatively
glycated, phosphorylated, methylated, oxidized and/or otherwise enzymatically or nonenzymatically
posttranslationally modified at one or more sites on the protein. Since many of the 100,000 proteins
produced by the human genome exist as primary sequence isoforms that also differ by how and where
they are posttranslationally modified, the actual number of chemically distinct proteins in the human
proteome is undoubtedly in the order of millions. The sheer magnitude and diversity of the human
proteome assures that the probability of any two organisms of the same species having exactly iden-
tical proteomes is vanishingly small, as is the probability that any single individual’s proteome will
be identical in normal and diseased states. Identifying, characterizing, and quantifying chemically
distinct proteins in this vast human proteome is the monumental challenge that proteomics presents
to analytical science.
Why is it important to accept this challenge? A current and near future priority of biomedical
research is to understand and treat complex diseases like diabetes, heart disease, and cancer that
have only recently become prevalent sources of morbidity and mortality in human populations.
A major goal of systems biology and proteomics is to identify and quantify specific proteins for
use as predictive diagnostic and/or prognostic biomarkers of disease [32]. Biomarker discovery is
complicated, however, by the fact that individual response to complex diseases is heritable and
highly variable. Sickle cell disease, for example, is caused by a Glu→Val substitution at the sixth
position of beta globin in all patients with the disease [33]. But the clinical presentation of sickle
cell disease varies from mild to severe, depending on the genetic background of each individual
patient [34]. Heritable variation is also observed in human responses to other complex diseases
[35,36] and in resistance to drugs used to treat disease [37]. Indeed, remarkably few human traits
are truly Mendelian due, in part, to the activities of modifier genes that alter the expression of other
genes and obscure regular patterns of inheritance [38,39]. Instead, inheritance of most quantitative
traits, including disease risk, is polygenic and attributable to interactions between two or more genes
and the environment. As a result, genetic variation and genetic background strongly influence both
disease risk and the composition of each individual organism’s proteome. Protein expression lev-
els and interactions with other biomolecules are also influenced by environmental factors, like
pathogen exposure, nutritional status, or drug intake. Consequently, the types and concentrations
of chemically distinct proteins in the human proteome vary widely both between individuals and
within individuals over time. The specific challenge that systems biology presents to proteomics is
to identify biomarkers of disease within this vast framework of biological variation in the human
proteome.
The role CE will play in proteomics will depend in part on how systems biology is applied
to biomarker discovery and how biomarkers are applied in the diagnosis and treatment of disease.
Given the likelihood that millions of chemically distinct proteins exist in the human proteome, it is
clear that comprehensive characterization of an individual organism’s entire proteome is beyond the
reach of present and near future technology. Given the wide dynamic range of abundance of different
proteins in the proteome, and the practical demand for high-throughput, low-cost diagnostic testing,
comprehensive characterization of even a limited subset of the human proteome, like the plasma
proteome, represents a significant analytical challenge. The estimated dynamic concentration range
of proteins in the plasma proteome is 108 , ranging from 0.5 ng/mL to 50 mg/mL [40]. Even if it
were possible to cost-effectively assess all proteins in just the plasma proteome, organizing and
understanding the plethora of data produced represents a colossal biomedical and computational
challenge because it requires mathematical comparison of the clustering of patterns in proteomic
phenotypes (characterized on the basis of the types and concentrations of chemically distinct proteins)
with clustering of patterns in clinical phenotypes (characterized on the basis of observable traits
or symptoms associated with health and disease). Associating proteomic phenotypes with disease
phenotypes is further complicated by the fact that clinical phenotypes of disease can range from mild
to severe within and between individuals with the same disease. At present, biomarker discovery
appears to be in a sort of Catch-22 situation: biomedical science is technically unable to assess
complex disease risk because analytical science does not cost-effectively provide the information
needed to discriminate between high and low risk disease phenotypes; but analytical science does
not provide this information primarily because biomedical science does not yet know which of the
millions of genes, transcripts, proteins and metabolites in an organism are best analyzed in order to
characterize normal and disease phenotypes and subphenotypes.
The solution to this situation will require cooperation between clinical and basic scientists and
combined expertise in the areas of medicine, biology, chemistry, engineering, and bioinformat-
ics. National and international sample repositories are already being created and categorized to
make samples from case controls and subjects with known diseases available for analysis [11]. A
plausible top-down approach to biomarker discovery is to develop a general analytical foundation
for proteome characterization and later extract functional or diagnostic utility for various proteins
on the basis of the results of large-scale systematic data collection [40]. This work is well underway
and CE appears poised to be a major contributor, much as it was in the effort to characterize the
human genome. A corollary bottom-up approach to biomarker discovery is to comprehensively eval-
uate narrow proteomic subsets. Characterizing how the levels and functions of a limited number of
proteins change in specific cells or specific biochemical pathways associated with disease processes
could help differentiate “normal” phenotypes from mild or severe disease phenotypes.
As systems biology progresses toward these goals it is important to emphasize that a range of
“normal” human phenotypes exist in part because genotypic and phenotypic variation give human
populations the ability to adapt and survive under different environmental and physiological condi-
tions. Furthermore, hereditary risk for mild or severe susceptibility to complex disease is ultimately
a function of the interplay between “normal” genetic variation and environmental factors that cause
disease. Biochemical and physiological traits often tend to differ in a continuum between individuals
in normal populations and within individuals over time or as they progress from normal to diseased
states. The fact that phenotypic traits segue almost imperceptibly across a continuum is the ultimate
challenge that nature poses to systems biology as it attempts to parse out complex trait patterns asso-
ciated with normal and disease phenotypes. The advancement of systems biology and proteomics
promises to complicate clinical diagnostics due to the increased influx of complex information, but
it also holds promise for early detection of complex disease risk and more effective treatment of
disease on a case by case basis. The systems biology approach appears indispensable for understand-
ing the clinical heterogeneity of complex diseases, the role of heredity and genetic background in
susceptibility to disease, what proteins or metabolites might be effectively used as risk biomarkers
for specific diseases, and how biological processes can be manipulated to treat disease. What role
CE will play in proteome analysis will depend on how proteomics is applied as part of the systems
biology approach. It also depends on how CE instrumentation and method development evolve in
competition with other advanced separation technologies whose evolution will also be driven by the
demand for more and better information about very complex biological systems and their roles in
very complex human diseases.
2.3 THEORETICAL ASPECTS

Capillary electromigration techniques can be used to analyze proteins in a wide variety of ways, all
of which depend on principles of chemistry and physics as they apply to the movement of proteins in
narrow-bore capillaries. Different CE techniques separate proteins on the basis of different protein
properties like charge, mass, isoelectric point and/or hydrophobicity. Different CE techniques also
use different detector systems and variously employ hydrodynamic, electrophoretic and/or electroos-
motic forces for separation and mobilization. The efficiency and specificity of any CE separation is
a function of fluid dynamics, electrical field strength, analyte diffusion and other analytical system
characteristics that affect peak migration, resolution, and detection. The term analytical system is
used here to describe the specific CE technique employed and all properties pertaining to its use,
including the CE platform, detector type, capillary type and cooling system temperature, separation
media composition and properties (e.g., pH, ionic strength, and viscosity), voltage and other run
conditions like duration and capillary conditioning. All of these analytical parameters interactively
determine how CE can be effectively applied for different protein applications. The sheer number
and complexity of interactions make the theoretical aspects of CE for protein analysis difficult to
fathom but it is also the reason that proteins can be analyzed by CE in so many different ways, often
with distinct advantages over other separation techniques. Many theoretical aspects regarding the
electrophoretic separation of molecules by CE are the same for all analytes and are discussed in
greater detail elsewhere in this book. Theoretical aspects of CE most important for protein analysis
are discussed next.
2.3.1 PROTEIN PROPERTIES

Proteins are amino acid polymers (polypeptides) with distinctive charge and shape characteristics
that can change depending on their microenvironment. Selection of a specific CE analytical system
for a specific protein application depends on the objectives of the separation and how the protein or
proteins of interest behave in any particular analytical system. The specificity and selectivity of any
CE analytical system depends in part on the physical and chemical properties of the protein itself,
including concentration, amino acid composition, charge, mass, shape, hydrophobicity, absorption
and emission characteristics and interactions with other biomolecules. These properties determine
how a protein interacts with separation medium constituents and the capillary or other physical com-
ponents of the analytical system. The Debye–Hückel–Henry theory can be used to mathematically
predict the electrophoretic behavior of proteins [41] but often does not apply to predicting protein
mobility in CE due to protein interactions with charged constituents of the capillary wall [42]. Pre-
dicting protein mobility is also complicated by inter- and intramolecular chemical interactions that
influence a protein’s physical properties and electrophoretic migration.
How proteins behave in a particular CE analytical system is strongly influenced by the pH
(negative log of the hydrogen ion concentration) and ionic strength (concentration of all ions)
of the sample and electrolyte solutions. Both markedly influence the charge properties of ioniz-
able functional groups, especially the amino and carboxyl termini and the basic and acidic side
chains of amino acids like lysine or glutamate. In general, polypeptides become more negatively
charged as pH increases due to progressive deprotonation of acidic carboxyl and basic ammo-
nium groups. As pH increases and hydrogen ion concentrations decrease, carboxyl groups are
converted into carboxylate anions (R-COOH to R-COO− ) and ammonium groups into amino
groups (R-NH+ 3 to R-NH2 ). Decreasing pH (i.e., increasing hydrogen ion concentration) has the
opposite effect. The dissociation constants (pKa ) of the α-carboxyl groups of different amino
acids range from about 1.8 for histidine to 2.5 for tryptophan. The pKa of α-amino groups range
from about 8.8 for asparagine to 10.6 for proline. Consequently the carboxyl termini of pro-
teins tend to be negatively charged at pH greater than 2.5 and the amino termini of proteins
are positively charged at pH < 8.8. The pKa of carboxyl and amino side chains range from
about 3.9 for aspartic acid to 12.5 for arginine. The isoelectric point (pI) of a protein is the
pH at which the molecule has zero net surface charge, that is, the effective number of posi-
tive charges equals the effective number of negative charges. The absolute number of ionizable
groups in a specific polypeptide molecule is a function of its primary amino acid sequence.
The actual number is a function of (a) pH, (b) the pKa of the ionizable functional groups present,
(c) posttranslational chemical modification of ionizable functional groups, and (d) inter- and
intra-molecular protein interactions that determine which amino acids are present on the surface
of the molecule.
2.3.2 CE SEPARATION TECHNIQUES

Capillary electromigration techniques collectively represent a family of related microanalytical meth-
ods that effect molecular separations using narrow-bore capillaries and high electric field strengths.
Operational differences between different modes of separation are primarily a function of the com-
position of the electrolyte or buffer solutions and whether coated or uncoated capillaries are used.
Because different CE techniques variously employ hydrodynamic, electrophoretic and/or electroos-
motic forces to mobilize and separate proteins, peak profiles and separation efficiencies differ between
techniques based largely on the relative contributions of these forces. For most separations samples
are introduced into the capillary either hydrodynamically using pressure or vacuum, or electroki-
netically using voltage. The IUPAC subdivides capillary electromigration techniques into capillary
electrophoretic techniques and electrically driven capillary chromatographic techniques while recog-
nizing significant overlap between these two classifications [1]. The CE techniques most commonly
used for protein analysis include CZE, CIEF, CSE, and MEKC. CAE is an electrophoretic tech-
nique where separation is effected by interaction between an analyte of interest and a specific affinity
reagent, for example, a protein-specific antibody that changes the electrophoretic migration of a target
protein. Other CE techniques are less frequently used for protein analysis. Capillary isotachophoresis
(CITP) is used in conjunction with CZE to automate online sample concentration before analysis.
Capillary electrochromatography (CEC) is a special case of capillary liquid chromatography where
movement of a mobile phase through a capillary packed or coated with a stationary phase is achieved
by EOF. CEC is also most often used in conjunction with CZE for online analyte concentration or
prefractionation.
Smaller proteins move faster in an electric field than larger proteins with similar charge because
acceleration is inversely related to mass to charge ratio (m/z) at any applied force. Protein separation
based on differences in m/z in is the fundamental principle of CZE that uses both electrophoretic
and electroosmotic forces to mobilize and separate proteins of interest. Since these two forces can
act in opposite directions, directional movement of proteins through the capillary depends on the
relative direction and magnitude of each force. CZE is the simplest and easiest CE separation mode
to use and can be employed in a wide variety of applications. Anode and cathode electrolyte solutions
are usually identical and the efficiency and selectivity of protein separations can be modified simply
by adjusting pH and ionic strength. Since the pI of most proteins are between pH 3 and 10, most
proteins are positively charged in solutions at pH 3 or below and negatively charged at pH 10
or above. The relationship between the pI of a protein and the pH of the separation medium can
be exploited to adjust assay selectivity. CZE performed using electrolyte solutions at low or high pH
are nonselective because most proteins will be positively or negatively charged, respectively, and
migrate to the oppositely charged electrode when voltage is applied. One reason low pH (e.g., 2.5)
phosphate buffer is a popular electrolyte solution for exploratory studies is that essentially all proteins
introduced into the anode end of the capillary will be positively charged and migrate toward the
cathode in CZE. Another reason is that the internal surfaces of fused-silica capillaries bind less
protein at low pH. The down side of low pH buffers includes reduced EOF and longer run times.
Higher pH buffers increase EOF but also increase binding of proteins with higher pI to capillary walls
such that capillaries with fixed or dynamic coatings are often used for analysis of more basic proteins.
More selective CZE separations can be achieved, however, by using electrolyte solutions with pH
slightly above or below the pI of the protein or proteins of interest. Under appropriate separation
conditions proteins with charge opposite that of the protein of interest will be excluded from the
separation. The same principle applies to the pH of sample solutions and selective introduction of
proteins into the capillary by electrokinetic injection.
Smaller proteins also move faster than larger proteins when electrophoretically mobilized through
media containing a selective physical barrier like polyethylene oxide [43] or dextran [44]. Protein
separation based on frictional and other forces influenced by protein mass and hydrodynamic radius is
the fundamental principle of CSE. CSE uses fixed or replaceable sieving media to differentially retard
migration of proteins or protein complexes with different molecular weights. CSE can be effectively
used to determine molecular weight in single or multidimensional analyses, or in mobility shift assays
that separate a protein from its complex with an antibody or nucleotide. Capillary gel electrophoresis
(CGE) is a special case of CSE performed in capillaries containing cross-linked gels. The use of
CGE for protein analysis has declined with the advent of noncross-linked soluble polymers that are
replaced between runs and afford more reproducible and reliable CSE separations [45].
CIEF [8,46–52] separates proteins in a pH gradient on the basis of surface charge and is ideal
for identification and quantification of protein isoforms since different sequence or posttranslation-
ally modified isoforms of the same protein often have different pI. The pH gradients established by
including ampholytes in CIEF separation media progressively range from lower pH regions at the
anode to higher pH regions at the cathode when voltage is applied. When sample is introduced into
the anode end of the capillary, proteins with higher pI migrate further toward the cathode through
higher and higher pH regions. When balance is eventually achieved between positive and nega-
tive surface charges proteins become neutral and those with similar pI focus into stationary zones.
Proteins with lower pI migrate shorter distances and focus in zones closer to the anode. The pro-
tein bands remain focused because diffusional movement in either direction puts the protein into
a higher or lower pH region that adds positive or negative charge to the molecule. Addition of
charge induces electrophoretic repulsions and attractions that cause the protein to return to a pH
region where it either loses or gains protons and becomes neutral again. The focused zones may
or may not be fully immobilized, depending on the absence or presence of residual EOF. EOF is
often suppressed in CIEF separations by using neutral coated capillaries and/or including EOF-
suppressing compounds in the separation media. Proteins with higher pI have shorter migration
times because they have shorter distances to travel when mobilized past a fixed detector placed
near the cathode end of the capillary. Migration times are progressively longer for proteins with
progressively lower pI. Wide (e.g., pH 3–10) or narrow (e.g., pH 6.7–7.7) range ampholytes are
commercially available and can be used alone or in combination to modify selectivity and sep-
aration efficiency. Variations in CIEF operating parameters primarily include the use of pressure
or electromigration to mobilize focused protein zones past the detector when using fixed win-
dow on-capillary detection. Additives like methylcellulose are often included in CIEF separation
media to regulate viscosity and flow rates in separations that hydrodynamically mobilize focused
protein zones.
Analyte separation based on hydrophobicity is the fundamental principle of MEKC [53–
55]. MEKC is a special case of electrokinetic chromatography (EKC) that separates analytes on
the basis of a combination of electroosmosis, electrophoresis and interactions between analytes
and surfactants or other additives to the electrolyte solution [1]. At sufficiently high concentra-
tions, detergent molecules like sodium dodecyl sulfate (SDS) self-associate in pseudostationary
arrangements called micelles, which have hydrophobic inner cores and hydrophilic outer sur-
faces. Micelles constantly associate and dissociate and have electrophoretic mobility opposite that
of normal EOF. During separation, proteins with differing hydrophobicities distribute differently
between the hydrophobic inner core and the hydrophilic solution phase. Hydrophilic (polar) pro-
teins that are completely insoluble in the inner core of the micelle migrate at the velocity of EOF.
Hydrophobic (nonpolar) proteins that completely associate with the micelle migrate at the veloc-
ity of the micelle. Proteins with intermediate levels of hydrophobicity will spend varying amounts
of time between the hydrophobic and hydrophilic phases of the separation media and have migra-
tions times intermediate to those of the micelle and EOF. Variation in protein hydrophobicity is
a function of the relative content of nonpolar (e.g., alanine, glycine, leucine, valine, etc.) and
polar (e.g., lysine, arginine, asparagines, glutamine, etc.) amino acids. Because separations are
based on hydrophobicity, both charged and neutral proteins can be analyzed by MEKC, which
is increasingly used to map one dimension of protein distribution in multidimensional analyses
[12,56–59].
2.3.3 INSTRUMENTS
No single CE instrument is ideal for all protein applications. Multicapillary instruments have more
throughput capacity than single capillary instruments. Some CE instruments have interchangeable
detection systems while others do not and are limited to more specific applications. How different
CE instruments are used for protein analysis is primarily a function of what detector systems are
available for a specific platform, how many and what types of capillaries can be used, sample
throughput capacity and the functionality of the system’s software for automation of sampling and
data collection. Research laboratories at the forefront of CE instrument development use a wide
variety of in-house manufactured platforms and software that are not yet commercially available.
Most of these incorporate unusual on- or off-capillary detectors into the analytical system or are
designed to permit online pre- or postseparation sample treatment or fraction collection. Perhaps the
fastest growing segment of advanced instrument development with relevance to proteomics is that of
multidimensional systems that analyze proteins on the basis of more than one protein property, that
is, charge, mass and/or hydrophobicity [59]. Systems have been developed for protein separation
using CE in two dimensions, HPLC in one dimension and CE in the other, or CE followed by MS,
which is both a separation and detection technique.
The advent of microfluidic devices like the Agilent 2100 Bioanalyzer (Agilent Technologies,
Sanata Clara, CA) has diverse implications for CE-based protein analysis. This commercially avail-
able microchip CE platform is a dedicated instrument that uses manufacturer-supplied reagent
kits for the analysis of DNA, RNA and proteins. Analyses are quick (∼30 min) and simple with
digital control and output that makes it easy to manage and store data. Up to 10 samples can
be analyzed simultaneously and the results can be viewed as either electropherograms or gel-like
images. The system uses noncross-linked linear polymers, fluorescent protein dye and laser-induced
fluorescence (LIF) to separate and detect proteins on the basis of size, and is useful for protein
quantification or molecular weight determination. This dedicated analyzer is a feasible alterna-
tive to labor intensive conventional electrophoresis methods with advantages superior to that of
SDS–polyacrylamide gel electrophoresis (SDS–PAGE) in large part due to automation. One benefit
for many biology and biochemistry laboratories is that the system can be used for both proteins
and nucleic acids, essentially replacing multiple conventional electrophoresis applications in a sin-
gle platform. Disadvantages include the inability to recover fractionated proteins and high capital
equipment and reagent costs. How microchip systems will be used in proteomic applications will
depend on how these systems evolve and their ability to characterize diagnostic subsets of the human
proteome.
A relatively limited number of other commercial CE systems is available. Multimode open
analytical CE platforms, like those available from Beckman-Coulter (Fullerton, CA) and Agi-
lent Technologies, can perform a wide variety of protein applications because many different
separation modes and detection systems can be used. Sebia Electrophoresis (Norcross, GA) and
Beckman-Coulter manufacture dedicated platforms and reagent kits for automated clinical analysis
of hemoglobin and/or plasma and urine proteins. Convergent Biosciences (Toronto, Canada) mar-
kets a CIEF system for whole-column image detection (WCID) of proteins. The systems biology
demand for faster, more specific and more sensitive protein analysis has encouraged new companies
to enter the CE instrument market with systems specifically designed for proteomic applications.
Cell Biosciences (Palo Alto, CA) is developing a dedicated CIEF system specifically for immun-
odetection of proteins. The instrument is reportedly capable of analyzing 96 samples in 8–12 h and
could replace Western blotting for use in some protein applications, especially for analysis of low
abundance proteins in an automated format. CombiSep (Ames, IA) manufactures a high-throughput,
multicapillary instrument capable of determining the molecular mass of proteins in 96 samples in
30 min [60].
2.3.4 CAPILLARIES
Poor between-run analytical reproducibility has been a major impediment to the wider use of CE in
protein applications. Variable capillary performance continues to be the single most important source
of analytical variation. Uncoated fused-silica capillaries from different manufacturers or different
production lots from the same manufacturer can have significantly different chemical properties. The
internal surfaces of uncoated fused-silica capillaries chemically interact with proteins directly and
with components of electrolyte solutions in ways that markedly influence protein recovery, mobility
and separation efficiency. The need for consistent starting conditions for each analytical run is one
reason uncoated capillaries are typically rinsed with sodium hydroxide and/or other regenerating
solutions between analytical runs. Capillaries are also usually rinsed between runs with separation
media in an effort to establish a reproducible internal chemical equilibrium.
The single most important capillary property influencing protein analysis is the charge charac-
teristics of the internal surface which can be positively charged, neutral or negatively charged. The
charge properties of the capillary wall depend on the pH and chemical composition of the electrolyte
solution and whether the capillary is uncoated or chemically modified. The surface silanol groups of
uncoated fused-silica capillaries are negatively charged at pH > 2 and are progressively deprotonated
at higher pH. The negatively charged wall attracts a layer of hydrated cations from the electrolyte
solution that move toward the cathode when an electric field is applied. This generates the EOF char-
acteristic of many CE separations, which greatly influences the resolution and efficiency of protein
separations. When it exists, EOF is always in the direction of the electrode with the same charge
as the capillary surface. EOF in uncoated fused-silica capillaries is progressively higher at higher
pH and decreases with increasing ionic strength. Protein–wall interactions in uncoated fused-silica
capillaries also increase at higher pH, which can decrease protein recovery and degrade capillary
performance.
Because the surface of a fused-silica capillary is chemically reactive it can be chemically mod-
ified to better control, minimize, or even reverse surface charge. Many of the problems associated
with protein analysis in uncoated fused-silica capillaries have been largely overcome by the devel-
opment of methods to chemically modify the internal surface. Chemical modification produces
more consistent starting conditions between analytical runs and decreases protein–wall interactions.
Coated capillaries are particularly useful for the analysis of basic proteins, that is, proteins with
greater positive charge, because basic proteins bind more readily to the negatively charged surface
of uncoated fused-silica capillaries. Typical coating materials include polyacrylamide, polyvinyl
alcohol, polyethylene glycol, polyvinylpyrolidone, polyamines, and cationic detergents. Capillaries
with covalently-attached (fixed) coatings that are acidic, neutral, basic, hydrophobic, or hydrophilic
are commercially available. The performance of capillaries with fixed coatings can deteriorate over
time and the capillaries must be replaced periodically due to degradation of the coating material or
accumulation of proteins or other foreign matter that corrupts the quality of the separation. Problems
encountered with fixed coatings can be overcome somewhat by the use of “dynamic” coatings where
the coating agent is applied internally before a separation and then is stripped and replaced between
runs. Coating reagents like polyamines or cationic detergents bind to negatively charged fused-silica
capillaries and create a positively charged internal surface with reverse EOF (directed toward the
anode). MicroSolve Technology Corporation (Eatontown, NJ) markets a family of quality-controlled
CE reagents (CElixir) that includes multiple electrolyte solutions with a variety of pH that have been
optimized for use with different dynamic coating solutions.
Because low abundance proteins can be difficult to detect by absorption spectroscopy, light trans-
mittance and path length are important properties that also must be considered in capillary selection.
Although most capillaries used for protein analyses are constructed of fused-silica, borosilicate
glass, quartz, Teflon and other materials are also infrequently used (quartz transmits ultraviolet light
better than glass at wavelengths used for protein detection). The internal cavities of most CE cap-
illaries are round but rectangular capillaries have also been used [61]. Typical internal diameters
range from 2 to 100 µm. Optical path length and detection sensitivity both increase with diameter
but separation efficiency is usually better in narrower capillaries. Most capillaries have an opaque
polyimide coating on the outside that provides tensile strength and limits breakage. For analytical
systems using on-capillary spectrophotometric detection, the polyimide coating is usually removed
by heating or with acid to create a window through which light will pass. Transparent external
coatings that provide tensile strength and also optically transmit light have also been developed
but are not widely used. Modified capillaries with extended light paths (e.g., bubble- or z-shaped)
have been developed to increase detection sensitivity, which can be especially important for low
abundance proteins. The utility of capillaries with extended light paths for protein analysis has
diminished somewhat with the expanded use of protein-specific derivatizing agents and fluorescence
detection.
2.3.5 SEPARATION MEDIA

Individual electrons normally cannot pass through solutions from one electrode to another. Instead,
the electrons are carried between electrodes in solution by electrolytes that make protein electrophore-
sis possible. When a capillary is placed between two separated vials of electrolyte solution and
voltage is applied, chemical reactions occur that consume electrons and generate negative ions at
the cathode and also produce electrons and generate positive ions at the anode. As negative and
positive charge builds up at the cathode and anode, respectively, ions in the electrolyte solutions
and sample move through the capillary toward the oppositely charged electrode. This neutralizes
charge differences between the electrolyte vials and allows continued electron flow through the
electrodes and charged molecule flow through the capillary. Conducting multiple analytical runs
without replacing electrolyte solutions can lead to ion depletion, especially when using low ionic
strength electrolyte solutions and small volume reservoirs.
As previously noted, the pH and ionic strength of electrolyte solutions markedly influence the
charge and migration of proteins due to their effects on both the protein and the internal surface
of the capillary. EOF is positively related to pH and inversely related to ionic strength. Although
higher ionic strength electrolyte solutions can suppress ionic interactions between charged proteins
and ionized silanol groups, they can also generate excessive heat at high voltages that may surpass
the cooling capacity of the analytical system. Overheating can alter the stability of proteins, corrupt
capillary coatings and degrade the quality and reproducibility of CE separations. CZE often uses
the same simple electrolyte solution (e.g., phosphate or borate) at both the anode and cathode
ends of the capillary. Other separation techniques use more complex electrolyte solutions that are
buffered or contain additives that (a) selectively or nonselectively modify proteins to alter their
charge or conformation (e.g., affinity reagents or detergents), (b) selectively or nonselectively modify
proteins to increase signal strength (e.g., fluorescent antibodies or nonspecific derivatizing agents),
or (c) selectively impede the mobility of proteins through the capillary on the basis of their size,
shape, or hydrophobicity (e.g., linear polymers or micelles).
2.3.6 DETECTION
Various types of detectors are commonly or uncommonly used for protein analysis in CE analytical
systems. In all cases, detection is based on characteristic properties of the protein or proteins of interest
in their modified or unmodified state. Most commercial platforms use on-capillary absorption or
fluorescence spectrophotometers that measure light as it passes transversely through the lumen of the
capillary at a fixed narrow window through which analytes sequentially pass during electromigration.
In contrast, WCID systems monitor absorbance or fluorescence of all proteins simultaneously over
the entire length of the capillary [62,63]. WCID is ideal for use with CIEF because it eliminates
band distortions caused by mobilization of narrow focused protein zones. WCID systems are also
advantageous for real time on-capillary study of dynamic processes like biomolecular interactions
or how proteins migrate under different analytical conditions. In-house constructed CE platforms
use a variety of less common on- and off-capillary detection systems (Chapter 9). These include
UV/Vis or fluorescence detection using unusual optical systems, polarization, total internal reflection,
correlation spectroscopy, and multiphoton excitation. Electrochemical or amperometric CE detectors
measure analyte concentrations on the basis of change in current caused by oxidation or reduction
of chemical reactants or products at an electrode [64]. Kuijt et al. [65] demonstrated the use of
quenched phosphorescence for CE peptide detection using separation media containing 1-bromo-
4-naphthalenesulfonic acid whose phosphorescence and background signal is quenched by electron
transfer from peptide amino groups at higher pH.
Absorption spectroscopy using ultraviolet or visible light is the detection system most commonly
used in CE protein applications. Most commercially available UV/Vis CE detectors use narrow band
filters to limit transmittance to specific wavelengths but multiwavelength diode array detectors are
also available. Choice of wavelength for UV/Vis protein detection depends on the abundance of the
protein of interest and the absorption characteristics of the protein, capillary and electrolyte solution.
Nearly all proteins absorb light at 214 and 280 nm due to molecular absorption of electromagnetic
energy at these wavelengths by peptide bonds and aromatic amino acids, respectively. Absorbance
at any given protein concentration is usually greater at lower UV wavelengths but many separation
media constituents, like some ampholytes used in CIEF, also absorb low-frequency UV light and
can interfere with detection. The absorption characteristics of naturally-occurring protein prosthetic
groups can be used to increase detection specificity for some proteins. Hemoglobin variants, for
example, are usually detected by absorbance at 415 nm because the attached heme groups absorb light
energy at this wavelength and thus impart some degree of detection specificity. Proteins containing
aromatic amino acids like tryptophan, tyrosine, and phenylalanine absorb electromagnetic energy
when excited at wavelengths between 250 and 300 nm and naturally fluoresce at wavelengths between
300 and 400 nm. Native fluorescence detection can be used to enhance analytical specificity for
proteins with high aromatic amino acid content [66–68].
The obligatory use of narrow-bore capillaries in CE limits permissible optical path lengths. This
makes on-capillary UV/Vis absorption spectrophotometry relatively insensitive for analysis of low
abundance proteins or when conventional detection systems are used for native fluorescence appli-
cations. Optical path length is of less concern when using high-intensity fluorescence detection.
Fluorescence spectroscopy measures luminescence generated when molecular absorption of a pho-
ton induces the emission of another photon with a longer wavelength. The use of derivatizing agents
to enhance the sensitivity of fluorescence detection has greatly expanded the utility of CE for some
protein applications [27,69]. LIF detectors using argon ion (488 nm) laser sources have been available
for some time on commercial multimode open analytical CE platforms. Fluorescein isothiocyanate
(FITC) and other fluorescein derivatives are often used with argon ion lasers because fluorescein has
absorption and emission maxima at 490 and 514 nm, respectively. Other high-intensity, monochro-
matic excitation sources are also used for CE fluorescence detection, including other laser sources
and less expensive UV light-emitting diodes [66,67].
The need for greater detection sensitivity and the availability of a wider range of high-intensity
excitation wavelength options have combined to promote development and use of novel fluorophores
for derivatization of a variety of analytes [70]. Derivatizing agents are used to nonspecifically
label carboxyl, amine or sulfhydryl groups of proteins or peptides. Examples include FITC and
Cy3.5 [71], 3-(2-furoyl)quinoline-2-carboxaldehyde (FQ) [56,57,72], o-phthalic dicarboxaldehyde/
beta-mercaptoethanol [66], monobromobimane [73], dipyrrometheneboron difluoride (BODIPY)
[74], and MitoTracker Green [75]. Fluorophores are classified as either fluorescent or fluorogenic
depending on their emission properties [69]. Fluorescent derivatizing agents emit photons when
excited regardless of their binding state while fluorogenic agents only fluoresce when bound to their
target. Background fluorescence produced by fluorescent derivatizing agents can limit fluorescence
detection sensitivity.
The number of derivatization sites on a protein and the chemical nature of the selected derivatiz-
ing agent can influence protein speciation and quantification [69,76]. For example, lysine is a primary
target for many derivatizing agents and most proteins have one or more lysine residues. Modification
of lysine amino groups or other ionizable functional groups by uncharged derivatizing agents can
alter protein surface charge and electrophoretic mobility. Incomplete labeling of ionizable functional
groups can produce multiple fluorescent species of the same protein each with different charge prop-
erties. Incomplete labeling can also corrupt the relationship between fluorescence signal strength and
protein concentration thus jeopardizing accurate quantification. Cationic amine derivatizing agents
can be used to apply a fluorescence tag to a protein while maintaining its charge characteristics, but
all cationic fluorophores are fluorescent rather than fluorogenic and produce background signal that
can obscure signals of interest [69]. Proteins of interest that are not readily separated from contami-
nating proteins are difficult to specifically detect using nonspecific fluorescence labeling techniques
in one-dimensional CE analyses. Selectivity and sensitivity can both be enhanced by the use of
analyte-specific fluorescence labeling like that afforded using fluorescent antibodies [27,77–79] or
when using fluorescent proteins or peptides as competing antigens [80] in CAE applications.
MS is increasingly the detection method of choice for many proteomic applications because it
provides high sensitivity detection and information about protein structure (Chapter 8). Conventional
two-dimensional SDS–PAGE has traditionally been the workhorse off-line separation system used
in conjunction with MS for the study of proteins [12]. An investigator can run a 2D gel, use an
automated spot-cutting system to collect a sample separated in two property dimensions by pI and
molecular weight, then process and analyze the sample using a variety of single or tandem MS
techniques, or HPLC-MS. Samples can also be analyzed directly using reversed phase, ion exchange
or other HPLC-MS technique. Coupling CE with MS permits high efficiency protein separation
based on one or more protein properties in an automated format that eliminates post-separation
sample handling. CZE is most commonly combined with MS using soft ionization methods due to the
simplicity and compatibility of many CZE electrolyte solutions. However, the process of transferring
separated proteins from the capillary outlet to the mass spectrometer still presents obstacles to more
widespread CE-MS use due to issues related to solvent compatibility and the placement of the CE
electrode at the CE–MS interface. The amount of sample that CE can deliver to the MS also remains
a major limitation for CE-MS detection of low abundance proteins. Greater detail regarding the
advantages and disadvantages of CE-MS for protein analysis is provided in Chapter 8 and recent
reviews [6,59,81,82].
2.3.7 SAMPLE PREPARATION AND FRACTION COLLECTION

Proteins that are abundant in complex biological matrices or have unique spectral properties can
be readily detected and quantified with little or no sample preparation. Little sample preparation
may also be needed to analyze proteins in less complicated matrices or where the protein of inter-
est is effectively separated from all potential contaminants. However, many proteins are present
in biological samples at concentrations that are lower than the limit of detection of many CE
analytical systems. If fluorescence derivatization is not compatible with the desired application,
analysis of low abundance proteins may require sample pretreatment to concentrate the protein or
proteins of interest or remove interfering compounds. Pre-analytical sample concentration or frac-
tionation can be performed on- or off-column. If sample availability and volume is not severely
limited, the same off-line methods commonly used to concentrate or fractionate samples for HPLC
or gel electrophoresis can be used to treat samples for CE analysis. Dilute samples can be con-
centrated by ultrafiltration or reconstituted after lyophilization. Buffer exchange by ultrafiltration
or dialysis can be used to remove salts or other compounds that may interfere with a separation.
A variety of common off-column fractionation methods from affinity chromatography to acid pre-
cipitation can be used to select specific proteins or remove interfering proteins with similar charge,
size or hydrophobicity. Development of advanced robotic instruments for automated sample treat-
ment using a variety of different modalities makes off-column sample pretreatment an increasingly
feasible option for high-throughput CE protein analysis. CE techniques like CITP or CAC are
frequently used for on-column sample pretreatment to concentrate or fractionate proteins before
separation, which can be particularly useful in CE-MS for automated detection of low abundance
proteins.
Fraction collection by CE is possible but difficult to accomplish in yields sufficient for further
study of separated proteins even when using CE methods that are considered preparative. CE is
consistently outperformed by other protein separation techniques when it comes to preparative oper-
ation or fraction collection for post-separation analysis [26]. Because CE sample injection volumes
are very low, the concentrations of separated proteins eluting from the capillary are also very low.
Typical CE injection volumes are usually < 50 nL and only nanogram quantities of specific proteins
are isolated with each analytical run. Peaks from multiple CE runs can be pooled to increase sample
recovery but consistent fraction collection requires highly reproducible run to run separations and
accurate prediction of post-detector elution from the capillary. Fraction collection can be performed
by adsorption of analytes to moving blotting membranes as they exit the capillary [83]. Inclusion
of magnetic beads coated with immobilized antibodies represents an innovative new approach to
on-column fraction collection [84]. CE fraction collection is desirable because CE is often able to
separate analytes that cannot be resolved by other techniques and because CE separation in free solu-
tion often allows preservation of analytes in their native form. A key advantage to preparative CE
compared to preparative HPLC is that analyses can be performed using very small sample volumes
like those available from single cells.
2.4 PRACTICAL APPLICATIONS

This section details characteristic examples of protein analysis by CZE, CSE, CIEF, and MEKC.
The selected examples were chosen to demonstrate some of the more important theoretical aspects
of protein separation by each of these techniques while also highlighting the application of CE for
characterization of protein properties, isoform identification, use in mobility shift assays, and multidi-
mensional protein analysis. Table 2.1 lists recent references for specific proteins analyzed by different
CE methods and show how CE analysis of proteins can be used in a variety of biological applications.
2.5 ANALYSIS OF PROTEIN PROPERTIES BY CZE AND CSE

Information about a protein’s physical and chemical properties can help determine its biological
function and can be useful in a bottom-up approach to understanding a biological system. The
biological activities and binding properties of proteins are strongly influenced by the type and number
of ionizable functional groups that are available to participate in chemical reactions with other
molecules. Information about a protein’s size, shape, and hydrophobicity can also be used to help
deduce transport properties, compartmental distribution or other biological characteristics.
CZE separates proteins in narrow-bore capillaries under the combined influence of electrophoretic
and electroosmotic forces. Electrophoretic mobility is a function of charge and frictional forces and
thus distinguishes proteins with different mass to charge ratios (m/z). Analysis of protein ladders
by Sharma and Carbeck [85] elegantly demonstrates the fundamental theoretical principles that
permit the separation of proteins by CZE. In these experiments, net charge and hydrodynamic
radius ladders were prepared from purified lysozyme. Positively charged functional groups were
progressively chemically modified to produce a range of molecules with precisely defined charge and
shape characteristics. Figure 2.1 is an idealized schematic showing how charge and radius ladders
are synthesized and the net effects of progressive derivatization of one or more functional groups on
a protein’s charge, hydrodynamic radius and electrophoretic mobility.
To study purified lysozyme from chicken egg white (EC 3.2.1.17, 14.3 kDa, 6 lysine residues, pI
10.9) the authors produced a charge ladder by chemically derivatizing primary amines (RNH2 ) at pH
12 with acetic anhydride. Two radius ladders were produced by derivatizing the same amines with
two different sizes of polyethylene glycol N-hydroxysuccinimide ester (PEG-NHS, with 2 or 5 kDa
PEG chains). The products of multiple reactions were combined to produce complete ladders in a
single sample since single reactions did not always produce all derivatized forms. The high lysine
(pKa = 10.5) content of unmodified lysozyme makes it a very basic protein that is positively charged
at the hydrogen ion concentration of the separation media (pH 8.4). Because the protein would be
expected to bind to the negatively charged surface of an uncoated fused-silica capillary the internal
surfaces of the capillaries used in these experiments were coated with a cationic polymer to reverse
internal surface charge and reduce protein–wall interactions. Because the internal surface charge of
the capillary was the reverse of normal, EOF was also the reverse of normal, that is, toward the
anode instead of the cathode, and the separations were conducted in reverse polarity where samples
were introduced at the cathode end of the capillary with the detector oriented toward the anode.
Under the conditions of these experiments, electroosmosis is a more or less uniform direc-
tional force toward the anode on all proteins in the sample. In contrast, positively charged protein
molecules are electrophoretically attracted to the cathode and move in a direction opposite that
of EOF at rates inversely related to m/z. Electrophoretic mobility toward the cathode is inversely
related to m/z because increasing positive charge increases electrostatic attraction while increasing
mass increases frictional forces that retard velocity. When electroosmotic force is greater than the
opposing electrophoretic force, a protein will migrate in the direction of the detector. Migration times
will therefore be greater for proteins with higher electrophoretic mobility (lower m/z) because the
difference between electroosmotic and electrophoretic forces increases as electrophoretic mobility
increases.
90
TABLE 2.1
Recent Applications of CE for Protein Analysis
Protein Purpose CE Technique Detection References
Akt (protein kinase B) Enzyme activity CZE UV [99]

Albumin Heparin affinity Microchip CAE LIF [100]
Albumin; alpha(1)-acid glycoprotein Protein–drug interactions CZE UV [101]
Albumin; lysozyme; ribonuclease a Protein characterization CIEF UV-WCID [102]
Amyloid precursor protein Glycosaminoglycan affinity CZE UV [103]
Beta lactoglobulin Allergen characterization CZE LIF [104]
Botulinum neurotoxin Enzyme activity CZE LIF [105]
Erythropoietin Glycoform characterization CZE ESI-TOF-MS [106,107]
Erythropoietin Glycoform characterization CZE UV [108]
Hemoglobin variants, mouse Isoform identification and quantification CIEF UV [109]
Hemoglobin variants, human Isoform identification and quantification CIEF UV [87]
Hemoglobin, globin chains Isoform identification and quantification CZE UV/Vis [110]
HIV-1 reverse transcriptase DNA-binding affinity CZE LIF [111]
Insulin; trypsin Protein–protein interactions Multicapillary CZE CCD-LIF [112]
Metallothionein; superoxide dismutase Isoform characterization CZE ICP-MS [113]
Methionine-enkephalin Immunoassay CZE-CAE LIF [114]
Prion protein Immunoassay CZE LIF [115]
Prion protein Immunoassay CZE-CAE/CIEF-CAE LIF [116]
Serum proteins Immunotyping CZE UV [117]
Soybean proteins Protein profiling CZE UV [118]
Transferrin Isoform characterization CZE UV/ICP-MS [119]
Transferrin Isoform quantification CZE UV [120,121]
Whey proteins Protein profiling CZE LIF [72]
Charge ladder
O O
Reaction: Reaction:
O O O O
O
NH2 O N NH2 NH O
pH 12 H pH 12
Derivatives: Derivatives:
+ + + ++
+ + ++
+
0 Mobility Mobility
4 3 2 1 0 0 4 3 2 1 0
Number of modifications Number of modifications
FIGURE 2.1 Schematic illustration of the synthesis of protein charge and radius ladders and their charac-
terization by CZE. Ladders are synthesized by the partial modification of primary amino groups on a protein.
Charge ladders are produced using acetic anhydride and radius ladders are produced using PEG-NHS esters. The
rungs of a charge ladder contain proteins that differ incrementally only in net charge from those in neighboring
peaks and are thus equally spaced. The rungs of a radius ladder contain proteins that differ in both net charge
and hydrodynamic radius from those in neighboring peaks. (From Sharma U, Carbeck JD. Electrophoresis
2005;26:2086–91. With permission.)
Figure 2.2 shows the effect of progressive derivatization of lysozyme amino groups with acetic
anhydride or PEG-NHS on electrophoretic mobility. Mobility was empirically calculated for each
observed peak as µ = (LT LD /V )((1/tnm ) − (1/tx )), where LT is the total length of the capillary, LD
is the length of the capillary to the detector window, V is the applied voltage, tnm is the migration
time of a neutral marker, and tx is the peak migration time. The constant influence of electroosmotic
force on peak migration is accounted for by including the migration time of the neutral marker in the
equation. Since all variables besides tx are constants in each analytical run, 1/tx decreases as peak
migration time increases and the difference between 1/tnm and 1/tx increases. Consequently, peak
mobility is greater for molecules with lower m/z and longer migration times.
The precisely defined charge and shape characteristics of the lysozyme molecules in the charge
and radius ladders clearly demonstrate these principles. In both types of ladders, unmodified
lysozyme had the highest positive charge and the lowest m/z so it also had the longest migration
times and highest calculated peak mobility. In the charge ladder, progressive neutralization of charged
lysozyme amino groups by incremental addition of acetic anhydride progressively decreased charge
but had little influence on the mass of the molecule. Consequently, the magnitude of the increase
in m/z was approximately equal with each additional amino group modification as evidenced by
the even distribution of the differently charged peaks. In contrast, incremental addition of either
size PEG-NHS to lysozyme amino groups progressively increased hydrodynamic radius but also
decreased charge. Since both mass and charge were modified with each additional amino group
modification, the increase in m/z was disproportional as evidenced by the unequal peak spacing of
lysozyme molecules with different numbers of PEG-NHS-modified amino groups.
Separation of purified lysozyme molecules with varying but well-defined charge and mass prop-
erties clearly demonstrates the theoretical aspects of how proteins are separated on the basis of these
characteristics by CZE. It shows that electroosmotic force must be considered in all separations and
Number of modifications
Radius
ladder,
PEG 5-kDa
Radius
ladder,
PEG 2-kDa
6 5 4 3 2 1 0
Charge
ladder
0.00 0.05 0.10 0.15 0.20

Mobility (cm2/kV s)
FIGURE 2.2 Analysis of lysozyme charge and radius ladders by CZE. The analytical system included a
Beckman-Coulter P/ACE MDQ instrument operated in reverse polarity at 20 kV with UV detection at 214
nm and capillary cooling at 25◦ C. Derivatized lysozyme molecules with different hydrodynamic radii and/or
charge or were separated by CZE on the basis of m/z using a positively charged, 60 cm long (50 cm to detector)
by 50 µm ID fused-silica capillary (cationic polymer poly-(diallyldimethylammonium chloride) was applied
to suppress adsorption of lysozyme). Separation media contained 25 mM Tris and 192 mM glycine (pH 8.4,
ionic strength 7.9 mM). The bottom panel shows peak mobility for the charge ladder constituents while the
upper two panels show peak mobility for the two radius ladders. (From Sharma U, Carbeck JD. Electrophoresis
can be modified to tailor separations based on the pI of the protein of interest. It also shows how elec-
trophoretic mobility increases with protein charge and decreases with increased mass/hydrodynamic
radius. Besides aptly demonstrating the principles of CZE, protein charge and radius ladders can
be applied in a variety of practical ways. Sharma and Carbeck [85] used lysozyme radius ladders
to measure partitioning of different sized molecules in polymer hydrogels and also suggested that
radius ladders can be used to measure the effects of hydrodynamic size on the transport properties of a
protein. Ebersold and Zydney [86] used CZE and myoglobin charge ladders to quantify electrostatic
interactions between proteins and membranes during ultrafiltration.
Like CZE, CSE also separates proteins in narrow-bore capillaries under the combined influence
of electrophoresis and electroosmosis (if present) but markedly increases the influence of frictional
forces by including sieving polymers in the separation media. Since frictional force is largely a
function of protein mass and shape, CSE primarily separates proteins on the basis of molecular
weight if charge differences between proteins are normalized by addition of detergent. SDS–PAGE
is the analytical technique most commonly used to determine protein molecular weight in most
biology and biochemistry research laboratories. Size exclusion HPLC and MS can also accomplish
this task with varying degrees of sensitivity, specificity, and automation.
Betgovargez et al. [58] produced a multidimensional profile of the human serum proteome and
used CSE to map one dimension of protein distribution by molecular weight. Their approach is
characteristic of how CE techniques can be incorporated in multidimensional analyses for proteome
characterization—in this case, by first using an off-line 2D chromatographic system before analysis
of the resulting fractions by both CSE and CIEF. Figure 2.3 was selected for inclusion in this section
because it very clearly demonstrates the use of CSE for molecular weight determination. In this figure,
upper panel shows the electropherogram obtained from a sample containing purified transferrin
supplemented with a 10 kDa internal standard and lower panel shows the electropherogram obtained
(a) 10 kDa 98 kDa

Internal Human transferrin
standard standard
50
AU
AU
10 20
(b) 35 kDa
kDa kDa kDa 100
kDa 150
kDa 225
kDa
0 10 12 14 16 18 20 22 24 26 28 30
Minutes
FIGURE 2.3 Protein molecular weight determination by CSE. The analytical system included a Beckman-
Coulter ProteomeLab PA 800 operated in reverse polarity at 15 kV with photo diode array UV detection at
220 nm and capillary cooling at 25◦ C. SDS-treated proteins were separated by CSE on the basis of molecular
weight using a negatively charged, 30.2 cm long (20.2 cm to detector) by 50 µm ID uncoated fused-silica-
capillary. The separation media included a formulation of polymers optimized for separation of proteins over a
wide molecular weight range that is marketed in a kit as part of the system. Analysis of purified transferrin is
shown in the upper panel while molecular weight standards are shown the lower panel. (From Betgovargez E,
et al. J Biomol Tech 2005;16:306–10. With permission.)
from a molecular weight standard containing eight different proteins with known molecular weights
(10–225 kDa). Because all samples were treated with SDS, which was also included in the separation
media, all proteins were negatively charged and CSE was performed in reverse polarity with the
detector toward the anode. SDS also normalizes charge effects on proteins such that separation was a
function of molecular weight rather than native m/z and smaller proteins moved more rapidly through
the sieving media than larger proteins.
The molecular weights of unknown proteins in test samples can be determined using a calibration
curve calculated on the basis of the relationship between the log of the molecular weight standards and
1/mobility. CSE can be performed under denaturing or nondenaturing conditions using replaceable
soluble polymers or fixed cross-linked polymers. CSE can be used to determine molecular weight,
for direct quantification of abundant proteins in uncomplicated matrices, or for quantitative analysis
by immunoassay where bound and free antibodies are identified and quantified on the basis of
differences in molecular weight.
2.5.1 IDENTIFICATION OF BETA GLOBIN ISOFORMS BY CIEF

Allelic polymorphism in a specific gene can produce nearly identical proteins with primary sequences
that differ by one or more amino acids. Primary sequence isoforms can be readily separated by CIEF
on the basis of their pI if the amino acid substitution changes the surface charge of the protein.
Similarly, enzymatic or nonenzymatic posttranslational chemical modification of ionizable protein
functional groups can produce chemically distinct isoforms of the same protein with detectably
different surface charge and pI. Analysis of hemoglobin by CIEF is an excellent example of how this
technique can be applied for the identification and quantification of a family of related proteins that
represent a concise subset of the human proteome. Members of this family include many primary
sequence isoforms associated with monogenic diseases but also includes posttranslationally modified
isoforms whose chemical composition reflects the concentrations of metabolomic constituents and
disease status. The principles and approach used to evaluate primary sequence and posttranslationally
modified beta globin isoforms described below can be applied to the study of other proteins, only
not as easily since hemoglobin is unusually amenable to one-dimensional analysis by CIEF due to
its high abundance, high solubility, and the UV/Vis detection specificity afforded by the attached
heme groups.
Hemoglobin is a tetrameric oxygen transport protein composed of two mixed pairs of alpha and
beta globin subunits. Hemoglobin tetramers readily dissociate into αβ dimers that migrate as a unit
when analyzed by most charge-based separation techniques [33]. Polymorphism and mutation in
globin genes are responsible for the synthesis of hundreds of different primary sequence isoforms of
both alpha and beta globin, including the beta globin isoforms responsible for the anemia associated
with sickle cell disease and hemoglobin C disease. The most prevalent (normal) beta globin isoform
has a glutamic acid residue at position 6 (βGlu6), which imparts negative surface charge to the αβ
dimer at pH > 4.1 due to deprotonation of its carboxyl side chain. Sickle beta globin has a valine
residue at the same beta globin position (βVal6) which does not have an ionizable functional group
and thus imparts no additional negative or positive charge regardless of pH. The C beta globin
isoform has a lysine residue at the same beta globin position (βLys6), which imparts positive surface
charge to the αβ dimer at pH less than ∼10.5 due to protonation of its amino side chain.
Homozygous sickle cell disease and heterozygous S/C disease are hereditary hemoglobinopathies
prevalent among persons of African ancestry. Figure 2.4a demonstrates the application of CIEF
in my laboratory for diagnosis of congenital human hemoglobinopathies [87,88] using pH 6.7–7.7
ampholytes. The sample shown in the figure was prepared for use as a hemoglobin pI standard by pool-
ing erythrocyte hemolysates from multiple subjects that were homozygous or heterozygous for globin
alleles that produce normal, sickle and/or C beta globin isoforms. The figure shows that αβ dimers
containing the more positively charged C beta globin isoform (αβ Lys6+ , pI = 7.44) eluted first,
followed by αβ dimers containing the more neutral sickle beta globin isoform (αβ Val6 , pI = 7.21),
followed in turn by αβ dimers containing the more negatively charged normal beta globin isoform
(αβ Glu6− , pI = 6.97). The peak between αβ Val6 and αβ Glu6− contains fetal hemoglobin where
gamma globin replaces beta globin in the hemoglobin dimer (αγ , pI = 7.06). Multiple sequence
variations between gamma and beta globin are responsible for the pI difference between αγ and
αβ dimers. The proportions of different hemoglobin variants present in a sample can be determined
on the basis of peak area at 415 nm. Excellent peak resolution and capacity for automated peak
identification based on calculated pI make CIEF a highly effective analytical method for quantitative
analysis of primary sequence globin isoforms.
CIEF can also be used for quantification and identification of posttranslationally modified
hemoglobin isoforms. Virtually all human beta globin isoforms have a cysteine residue at posi-
tion 93 (βCys93) whose sulfhydryl group (pKa ∼ 8.4) participates in nitric oxide transport. βCys93
also nonenzymatically forms a mixed disulfide with glutathione in vivo under conditions of oxidative
stress; one physiological role of glutathione is to protect protein sulfhydryl groups from irreversible
oxidation. The mixed disulfide formed between βCys93 and glutathione is called S-glutathionyl
hemoglobin and has been recommended for use as a biomarker of oxidative stress [89–92]. The
sample shown in Figure 2.4b was prepared in vitro from erythrocytes obtained from a patient with
homozygous sickle cell disease, and consequently, only the sickle beta globin isoform and αβ Val6
dimers were present. A proportion of the dimers were converted into S-glutathionyl hemoglobin
by disulfide exchange in vitro with glutathione disulfide (oxidized glutathione). Glutathione is a
tripeptide (Glu–Cys–Gly) that is negatively charged at pH > 4.1 due to deprotonation of the carboxyl
side chain of its N-terminal glutamic acid. When the mixed disulfide is formed between βCys93
and the glutathione cysteine residue, the αβ Val6 dimer acquires negative charge that lowers its pI,
increases its migration time and produces a new peak (αβ Val6,Cys93−SG , ∼40% of total hemoglobin
based on peak area at 415 nm) that was not detected in the untreated hemolysate. Naturally occurring
0.06 (a) αβGlu6-

αβVal6
αγ
0.04
αβLys6+
0.02
0.00
Absorbance (415 nm)
0.06 (b)
αβVal6
0.04
αβVal6,Cys93-SG
0.02
0.00
αβGlu6-
0.06 (c)
0.04
0.02 αβGlu6-,Val1-Glucose
0.00
4 5 6 7 8 9 10 11 12
Minutes
FIGURE 2.4 Analysis of primary sequence and posttranslationally modified beta globin isoforms by CIEF. The
analytical system included a Beckman-Coulter P/ACE MDQ instrument operated in normal polarity at 30 kV
with UV/Vis detection at 415 nm and capillary cooling at 20◦ C. Hemoglobin αβ dimers with different beta
globin primary sequences or chemical modifications were separated by CIEF on the basis of pI using a neutral,
31 cm long (21 cm to detector) by 50 µm ID fused-silica capillary (DB-1, coated with dimethylpolysiloxane to
suppress EOF). Hemolysates were prepared by adding 10 µl of erythrocytes to 200 µl of hemolyzing reagent
(10 mM KCN, 5 mM EDTA) and pressure injected at the anode for 10 s at 1.0 psi. Separation media contained
0.375% methylcellulose and 2% pH 6.7–7.7 ampholytes in deionized water. Cathode solution was 80 mM borate,
pH 10.25. Anode solution was 100 mM phosphoric acid in 0.375% methylcellulose. Voltage was applied without
pressure for 3 min before focused zones were mobilized past the detector using pressure at 0.5 psi with constant
voltage. Samples were (a) mixed hemolysate containing Hbs C, S, F and A; (b) hemolysate containing HbS and
S-glutathionyl HbS; (c) hemolysate from a patient with diabetes and elevated glycated hemoglobin.
S-glutathionyl hemoglobin levels in vivo measured by MS or ion exchange HPLC are estimated to
be <3% of total hemoglobin in normal individuals and up to ∼9% in patients undergoing peritoneal
dialysis [89,91,92].
Enzymatic and nonenzymatic protein glycation are posttranslational biochemical processes
that also occur naturally in many cells. Attachment of sugars to specific serine, asparagine, and
hydroxylysine residues is under strict enzymatic control and serves a multitude of biological func-
tions, such as facilitating protein secretion or prolonging protein survival in the circulatory system
[33]. In contrast, nonenzymatic glycation is a concentration- and oxidation-dependent condensa-
tion reaction between protein amino groups and the carbonyl groups of aldehydes, ketones, and
reducing sugars, also known as the “browning” or Maillard reaction. Nonenzymatically glycated
proteins initially form Schiff bases that can chemically rearrange to form Amadori products that
can rearrange further to form advanced glycation end (AGE) products that disrupt normal pro-
tein function and play an important role in the etiology of vascular complications associated with
diabetes and other chronic diseases [93]. Glycated hemoglobin levels increase in proportion to
mean blood glucose that remains elevated in diabetes patients with poor metabolic control. Glycated
hemoglobin measurements (including HbA1c) are widely used to clinically monitor metabolic status
and therapeutic success in diabetes patients.
The N-terminal valine of beta globin (βVal1) is particularly susceptible to nonenzymatic
posttranslational chemical modification by glucose, glycolytic intermediates, and glucose oxi-
dation/fragmentation products. Chemical modification of the βVal1 α-amino group inhibits its
protonation and lowers both the pKa of the α-amino group and the pI of hemoglobin αβ dimers. Glu-
cose and other molecules react with other alpha and beta globin amino groups but glycation at these
sites apparently does not alter the surface charge of αβ dimers [33], probably because the charge on
these groups is not significantly altered by glycation at pH optimal for separation by charge-based
separation techniques. The sample shown in Figure 2.4c was prepared from erythrocytes obtained
from a patient with type 1 diabetes and elevated glycated hemoglobin levels. The larger of the two
major CIEF peaks (αβ Glu6− ) contains αβ dimers that are not posttranslationally modified in ways that
alter pI, for example, on βVal1. The smaller peak (αβ Glu6−,Val1-Glucose , ∼15% of total hemoglobin)
with longer migration time contains αβ dimers with lower pI that are posttranslationally modified
on β Val1 by glucose and other biomolecules.
These examples show how CIEF can be used to identify and quantify a family of sequence and
posttranslationally modified beta globin isoforms based on pI. They also show how oxidative stress
and hyperglycemia both influence the speciation of hemoglobin into a variety of chemically distinct
isoforms that differ by how and where they are posttranslationally modified. Because hemoglobin
is nonenzymatically modified by compounds like glutathione and glucose, isoform levels measured
by CIEF reflects the concentrations of these and other metabolites and can be used to quantitatively
assess a narrow subset of the human proteome that provides diagnostic information in a more or less
bottom-up approach to biomarker discovery. Comprehensive assessment of the hemoglobinome, if
you will, holds promise for identifying normal and disease risk phenotypes because beta globins
are posttranslationally modified in characteristic patterns associated with characteristic patterns in
disease phenotypes, that is, oxidative stress and diabetes. Many other proteins also undergo similar
nonenzymatic and/or enzymatic posttranslational modifications with glutathione, glucose and other
metabolites. Assessment of posttranslationally modified protein isoforms, especially long-lived pro-
teins like collagen where modifications accumulate irreversibly, could provide information about
both the proteome and the metabolome since the chemically distinct protein isoforms present are a
function metabolite exposure over time.
2.5.2 SP 1 MOBILITY SHIFT ASSAY BY CAE

Capillary affinity electrophoresis is often used to enhance selectivity or study binding interactions
between proteins and other biomolecules. CAE applications rely on the principle that the elec-
trophoretic mobilities of bound and unbound forms of a protein of interest differ and can be identified
and quantified by CE, most often using CZE or CSE. One example is electrophoretic mobility shift
assays, which are used to study protein–protein, protein–antibody, or protein–DNA interactions. A
study by Ronai et al. [94] demonstrated the advantages of a CE mobility shift assay (CEMSA) for
analysis of SP 1 binding capacity based on affinity between SP 1 and a fluorescent-labeled synthetic
oligonucleotide. SP 1 is a sequence-specific DNA-binding protein that plays an important role in the
transcriptional regulation of many genes. Figure 2.5 demonstrates the principle of the assay using
a fluorescent DNA probe and HeLa nuclear extracts that contained SP 1. In this example, DNA
probes were prepared from the 5 upstream regulatory region of the human dopamine D4 receptor
(DRD4) gene by polymerase chain reaction (PCR) amplification using fluorescein labeled primers.
The probes (at 40-, 100-, or 400-fold dilutions) were then incubated with or without HeLa nuclear
extract and analyzed by CE with fluorescence detection.
The m/z of the free DNA probe was lower than that of the DNA–protein complex and conse-
quently had greater electrophoretic mobility and shorter migration time when analyzed by CZE.
The analysis was performed in normal polarity but with the detector window closer to the inlet
1 RFU
400-fold dilution + HeLa
2.5 RFU
10 RFU
10 RFU
40-fold dilution
0 2 4 6 min
FIGURE 2.5 SP 1 capillary electrophoresis mobility shift assay (CEMSA). The analytical system included a
Beckman-Coulter P/ACE MDQ instrument operated in normal polarity at 200 V/cm with laser-induced fluo-
rescence detection (488 nm excitation, 520 nm emission) and capillary cooling at 20◦ C. Free DNA (∼4.2 min)
and DNA–protein complex (∼6 min) were separated by CZE based on m/z using a neutral coated, 50 cm long
(10 cm to detector) by 50 µm ID fused-silica capillary. Separation media (TBE buffer) contained 89 mM Tris
HCl, 89 mM boric acid, and 2 mM disodium EDTA, pH 8.3. Samples containing various dilutions of fluorescein
labeled DNA probe with or without a fixed amount of HeLa nuclear extract (containing SP 1) were pressure
injected for 5 s at 3 psi. (From Ronai Z, et al. Curr Med Chem 2004;11:1023–9. With permission.)
(10 cm) in order to shorten the migration time while still using a longer capillary (50 cm total
length) and high voltage. The proportion of DNA–protein complex present was a function of the
relative concentrations of the DNA probe and nuclear extract. In the absence of HeLa nuclear extract
(bottom electropherogram), a single large peak was observed at about 4.2 min representing the
unbound (free) DNA probe. Addition of HeLa nuclear extract with the same 40-fold dilution of
the same DNA probe (second electropherogram from the bottom) resulted in the appearance of a
second peak at approximately 6 min representing the SP 1-DNA probe complex. Decreasing the
concentration of the DNA probe while holding the concentration of HeLa extract constant (top two
electropherograms) resulted in a proportional increase of bound vs. free DNA probe. Competition
assays using nonfluorescent DNA probes with the consensus SP 1 binding sequence or a mutated
version of the consensus sequence showed that the effect of the HeLa extract was specific for SP 1.
The authors used this CEMSA to show that Sp 1 binding to the DRD4gene was influenced by the
number of SP 1 binding 120 base pair repeat sequences present in the highly polymorphic 5 upstream
regulatory region.
2.5.3 ANALYSIS OF FREE PROSTATE-SPECIFIC ANTIGEN ISOFORMS BY CAPILLARY

ZONE ELECTROPHORESIS
Prostate-specific antigen (PSA) is a single-chain glycoprotein that belongs to the kallikrein fam-
ily of serine proteases [95]. Overproduction of PSA, for example, in prostate hyperplasia, cancer
or prostitis, causes more of the protein to enter the circulatory system such that serum PSA can
be used as a biomarker for prostate-related diseases. PSA has one known glycosylation site that
is chemically modified by a diantennary N-linked oligosaccharide that itself can vary on the basis
of the presence or absence of fucose, one or two sialic acids, or the glycan content of fucose
and N-acetyl-d-galactosamine. Free PSA thus exists as chemically distinct posttranslational iso-
forms. Genetic variants of PSA have also been reported. In addition, PSA in serum binds to several
serine protease inhibitors, including α1-antichymotrypsin, α2-macroglobulin and α1-antitrypsin.
Consequently, PSA in serum exists as both free PSA and complexed PSA, primary sequence
Group I
Relative absorbance (A214)

(a) 2.0 (b) 3.0
Sample 1 Sample 2
free PSA free PSA
1.0 1.0
0 0
3 6 9 12 15 18 21 24 3 6 9 12 15 18 21 24
Time (min) Time (min)
(c) 2.0 (d) 2.0


Sample 3 Sample 4
free PSA enzymatically
active
1.0 1.0
0 0
3 6 9 12 15 18 21 24 3 6 9 12 15 18 21 24
FIGURE 2.6 Analysis of PSA heterogeneity by CZE. The analytical system included a Model 310 Thermo
Capillary Electrophoresis Crystal CE instrument operated in normal polarity at 25 kV with UV detection at 214
nm. Free PSA isoforms were separated on the basis of m/z using a 60 cm long by 50 µm ID fused-silica capillary.
Separation media contained 20 mM sodium borate, 5 mM 1,3-diaminopropane, pH 8.0. Samples were pressure
injected for 3 s at 200 mbar. Panels (a)–(d) show the separation of PSA isoforms in purified PSA obtained from
four different commercial vendors.
isoforms, and in posttranslationally modified isoforms. Free PSA isoforms were previously reported
to have pI between 5.7 and 7.2. Characterization of all structural variants/isoforms is important
since modification may differentially modulate PSA biological activity and serve as biomarkers
of disease.
Donohue et al. [95] developed a CZE method for the analysis of PSA and compared the profiles
of seven commercially available purified PSA products. Since different manufacturers use different
proprietary methods to produce purified PSA for research use, their goal was to evaluate the hetero-
geneity of commercially available PSA. The capillary used in these experiments was dynamically
coated by adding 1,3-diaminopropane to the buffer as an EOF modifier to enhance glycoprotein
separation and reproducibility. Figure 2.6 shows the heterogeneity observed in four of the seven
samples analyzed. Four to seven peaks were observed with one dominant peak in each case rep-
resenting 51%–70% of the total free PSA (complexed forms and genetic variants are not expected
in purified samples). The dominant peak in the sample shown in panel c was not the same as the
dominant peaks in panels a and b on the basis of peak retention time, which had intra-assay and
interassay relative standard deviations of 0.6% and 5.0%, respectively. The author suggested that
the dominant peak is most likely fucosylated disialylated diantennary PSA as demonstrated by other
investigators using LC-MS. The majority of the lower-abundance isoforms had longer migration
times indicating that these species are more electronegative and likely attributable to alteration of
the oligosaccharide structure. Since the profiles of all seven commercially available samples dif-
fered, the authors concluded that variation in purification methods strongly influence the isoform
composition and heterogeneity of purified PSA products.
2.5.4 ANALYSIS OF HISTONE H1.5 ISOFORMS BY OFFLINE

MULTIDIMENSIONAL RPHPLC-CZE
The fundamental subunit of chromatin is the nucleosome core, which consists of DNA wrapped
around an octamer of core histones [96]. H1 histones are linker histones associated with the core
histone–DNA complex and with the linker DNA between adjacent nucleosomes. Histone H1 phos-
phorylation is a function of the cell cycle, lowest in G1 phase, rising during S and G2 , and reaching
a maximum in M phase. H1 phosphorylation appears to be involved in chromatin decondensation,
destabilizing chromatin structure, and weakening binding to DNA. Individual H1 subtypes differ in
their degree of phosphorylation during the cell cycle.
To better understand H1 phosphorylation, Sarg et al. [96] incorporated CZE in a multidimensional
analysis to show that human lymphoblastic T-cells have unambiguous site specificity for histone
H1 phosphorylation. Human lymphoblastic T-cells were synchronized in culture and labeled with
32 P. Lyophilized perchloric acid cell extracts containing 32 P-labeled H1 histones (∼500 µg) were
first separated by reversed phase HPLC (RP-HPLC) on a Nucleosil 300-5 C4 column. Figure 2.7a
shows that H1.5, the main component histone and the most highly phosphorylated during inter-
phase and mitosis, was clearly separated from residual histones H1.2, H1.3, and H1.4. The extent
of H1.5 phosphorylation was then determined by analyzing the H1.5 fraction by CZE (Figure 2.7b),
which identified four peaks representing nonphosphorylated (p0), monophosphorylated (p1), diph-
sophorylated (p2), and triphosphorylated (p3) H1.5. A similar H1.5 sample pretreated with alkaline
phosphatase showed only two peaks, a large p0 peak and a smaller p1 peak but no p2 or p3, con-
firming the identity of the phosphorylated isoforms. Phosphate groups are negatively charged such
that their addition to H1.5 decreases surface charge and alters m/z.
2.5.5 ONLINE MULTIDIMENSIONAL PROTEIN ANALYSIS BY CSE-MEKC

Multidimensional separation of proteins based on more than one physical property can greatly expand
the number of proteins that can be qualitatively or quantitatively identified among the complement of
proteins present in any proteomic compartment. The process can be automated by coupling multiple
HPLC, CE and/or MS systems in a variety of hybrid/hyphenated combinations. Interfacing any
two analytical systems online, even two different CE techniques, requires the use of compatible
buffers/electrolytes and transfer of protein-containing fractions from the outlet of the first dimen-
sion to the inlet of the second dimension [59]. For complex biological samples, multidimensional
analyses can be quantitative if MS is used but is at best semi-quantitative if based on absorbance or
derivatized fluorescence intensity across a landscape of intersecting dimensional coordinates (e.g.,
pI vs. molecular weight). Compared to HPLC, CE provides exceptionally high separation efficiency
but is also uniquely suited for analysis of proteins in single cells due to the very low sample volume
requirement.
Harwood et al. [56] used CSE and MEKC to map protein distribution in single mouse embryos
in two dimensions based on size and hydrophobicity (Figure 2.8). Although CE can be used to
inject, lyse, and analyze whole cells completely online [68,69,77,97], mouse embryos are rela-
tively large (∼75 µm) and would require the use of greater diameter capillaries where Joule heating
might compromise separation efficiency. Embryo extracts were therefore prepared off-line and
protein and nonprotein primary amines were nonspecifically labeled with 3-(2-furoyl)quinoline-
2-carboxaldehyde (FQ) before analysis by CSE-MEKC. Over 100 mouse embryo components were
identified with fluorescence intensities over 10 times greater than the standard deviation of the
background fluorescence.
On average, mouse embryos contained about 27 ng of protein each and yielded approximately
1 µL of lysate. In this study, approximately 3 nL (90 pg of protein or 0.3% of total protein) was
injected such that analytical reproducibility could be tested on the basis of the results of multiple
analyses. Data collection and processing were performed using in-house developed software that
0.6
(a) RPC
H1.5
0.5
H1.2
0.4 H1.3
H1.4
A210
0.3
0.2
0.1
0.0
0 15 30 45
0.035
(b) p1
HPCE
H1.5p0
0.025
p2
A200
0.015
p3
0.005
–0.005
14 16 18
Time (min)
FIGURE 2.7 Isoform analysis of phosphorylated H1 histone by multidimensional HPLC-CZE. Panel (a) shows
the separation of H1 histones by RP-HPLC on a Nucleosil 300-5 C4 column. Panel (b) shows the separation
of the H1.5 RP-HPLC fraction by CZE. The CE analytical system included a Beckman-Coulter P/ACE 2100
instrument operated in normal polarity at 12 kV with UV detection at 200 nm and capillary cooling at 25◦ C.
Nonphosphorylated (p0) and mono-, di-, and tri-phosphorylated (p1, p2, p3) H1.5 isoforms were separated
on the basis of m/z using a 57 cm long (50 cm to detector) by 75 µm ID untreated fused-silica capillary.
Separation media contained 0.1 M sodium phosphate, 0.02% hydroxypropylmethylcellulose, pH 2.0. Samples
were injected for 2 s. (From Sarg B, et al. J Biol Chem 2006;281:6573–80. With permission.)
produced gel-like images or landscape images like those shown in Figure 2.8. Each analytical
run consisted of approximately 200 cycles where fluorescence intensity was measured between
0 and 17 s as 200 different CSE fractions were analyzed by MEKC. The reproducibility of the
separations in replicate analyses were determined in the first and second dimensions by comparing
cycle number and seconds, respectively, for 50 sample constituents with the highest fluorescence
intensities. The average standard deviations were 4.7 cycles and 0.30 s in the CSE and MEKC
dimensions, respectively.
This example shows that CSE-MEKC can separate and detect over 100 different cellular compo-
nents in a single mouse embryo and could potentially be used to semiquantitatively compare protein
expression levels in individual cells from a heterogeneous cell population. Though nonspecific, the
use of fluorescence derivatization and detection greatly increased the sensitivity and dynamic range
15
15
18 18
MW 12 12
50
50
s) MW 6 e ( s )
(kD 6 e ( m
tion ti
0
on tim (kD
0
0
10
0 ra
migrati
10
a) a) m ig
M E K C MEKC
(a) (b)
FIGURE 2.8 2D CSE-MEKC separation of proteins in a single mouse embryo. The analytical system included
an in-house constructed multidimensional CE system operated in reverse polarity with postcolumn laser-induced
fluorescence detection in a sheath-flow cuvette (473 nm excitation, 580 nm emission). Both CSE and MEKC
separations used 20 cm long by 30 µm ID fused-silica capillaries dynamically coated with EoTrol LN (Target
Discoveries, Palo Alto, CA) to stabilize normal EOF. Samples were injected at the inlet of the CSE capillary
for 5 s at 11 kPa using negative pressure. CSE was performed at 1000 V/cm in separation media containing
100 mM Tris, 3.5 mM SDS and 5% dextran polymer (513 kDa) at pH 8.7. Analytes were transferred sequentially
into the second capillary where MEKC was performed at 900 V/cm in separation media containing 100 mM
CHES, 100 mM Tris and 15 mM SDS at pH 8.7. The net electric field across the first capillary was held
at 0 V/cm during analyte separation in the second capillary, a process that was repeated for approximately
200 cycles where each cycle was composed of a 1 s transfer and a 17 s separation in the second dimension. Data
collection and processing were performed using in-house developed software that can produce the landscape
images shown here at full scale (a) or ten-times expanded scale (b). (From Harwood MM, et al. J Chromatogr A
of protein detection. Multidimensional resolution was better than that attainable in a single CE
dimension, and was perhaps similar to that attainable using conventional 2D SDS–PAGE. However,
analyzing the protein content of embryos by conventional methods like 2D SDS–PAGE and tandem
MS requires pooling extracts from many cells in order to acquire enough sample for analysis. Such
results only reflect average protein expression levels rather than expression levels in individual cells
as described in the example presented here.
Proteomic analysis of individual cells or specific tissues has important implications for diagnosis
of disease since comparing normal and diseased cells can help identify proteomic phenotypes char-
acteristic of disease. For example, Kraly et al. [98] used CSE-MEKC to analyze tissue homogenates
from esophageal and stomach biopsy samples collected from patients with Barrett’s Syndrome. They
identified 18 features from the homogenate profiles as biogenic amines and amino acids that differed
in diseased and normal tissues. The results suggest that two-dimensional CE might be useful for
rapid characterization of endoscopic and surgical biopsies for this and other diseases.

The probability that all chemically distinct proteins in the vast human proteome will be identified
and characterized any time soon is vanishingly small. However, research and development toward
this goal can only improve the quality and quantity of proteomic data that is available and the ability
of biomedical science to use that information to improve the human condition. CE appears poised
to play a significant role in the collection of proteomic data for use in both top-down and bottom-up
systems biology approaches to biomarker discovery. Electrophoresis in narrow-bore capillaries is

capable of extremely high efficiency protein separations based on one or more protein properties.
Some of the application examples included in this chapter show how one-dimensional CE techniques
can be used to identify and quantify specific protein isoforms or study their physical properties. Other
examples show how CE can be used in multidimensional analyses to separate proteins on the basis of
multiple physical properties. The efficiency of multidimensional protein separations is approaching
the scale needed for relatively comprehensive analysis of the more abundant proteins present in
discrete proteomic compartments like the plasma proteome or individual cell proteomes. Many of
the reproducibility, sensitivity and specificity issues traditionally associated with analysis of proteins
in a capillary format have been overcome by technological advances in CE instrument and method
development. MS is and will likely remain the detection method of choice for most microanalytical
protein separation techniques due to its high degree of sensitivity and selectivity; characteristics that
one-dimensional CE techniques often lack when it comes to protein analysis.
Commercial availability of CE-MS analytical systems has opened the door for high efficiency,
multidimensional protein analysis to those who can afford it. Commercialization of multidimensional
systems that combine multiple CE techniques, or CE with HPLC, could lead to even greater ability to
separate, identify and quantify proteins in complex biological samples. However, multidimensional
map coordinates of the distribution of proteins based on charge, size, and/or hydrophobicity provide
relatively low quality data because complex biological samples simply contain too many proteins with
similar multidimensional properties. Proteomic phenotyping based on limited protein information
can only be used to divide individuals in a population into relatively indistinct groups and provides
little or no mechanistic information. Higher quality proteome data quantifies specific proteins, as
many as possible, and can be used to quantitatively group individuals on the basis of discrete patterns
in the proteomic distribution of chemically distinct proteins.
Understanding and classifying biological variation in the vast human proteome is a major under-
taking and a challenge that nature and systems biology have presented to biomedical and analytical
science. Capacity to cost-effectively and simultaneously profile the concentrations of many chem-
ically distinct proteins in selected proteomes is what biomedical science most urgently needs for
biomarker discovery and management of complex diseases. By design or by good fortune, analytical
systems for protein analysis by CE are well-positioned to advance in both bottom-up and top-down
approaches to the collection of proteomic data. High-throughput microchip and multicapillary CE
instruments, and high-resolution multidimensional systems than incorporate CE, can greatly expand
both the quantity and quality of protein data that can be acquired. What remains to be determined is
how well CE will compete with other advanced protein separation technologies or assay techniques
whose development will also be influenced by the emergence of systems biology and biomarker
discovery as driving forces in the evolution of analytical instrumentation.
ACKNOWLEDGMENT
The author would like to thank Dr. Mark Richards for his kind and considered review of the manuscript
for this chapter.
REFERENCES
1. Riekkola ML, Jonsson L, Smith RM. Terminology for analytical capillary electromigration techniques.
Pure Appl Chem 2004;76:443–51.
2. Dolnik V, Hutterer KM. Capillary electrophoresis of proteins 1999–2001. Electrophoresis 2001;
22:4163–78.
3. Hutterer K, Dolnik V. Capillary electrophoresis of proteins 2001–2003. Electrophoresis 2003;24:
3998–4012.
4. Dolnik V. Capillary electrophoresis of proteins 2003–2005. Electrophoresis 2006;27:126–41.

5. Huang YF, Huang CC, Hu CC, Chang HT. Capillary electrophoresis-based separation techniques for
the analysis of proteins. Electrophoresis 2006;27:3503–22.
6. Klampfl CW. Recent advances in the application of capillary electrophoresis with mass spectrometric
detection. Electrophoresis 2006;27:3–34.
7. Kasicka V. Recent advances in capillary electrophoresis and capillary electrochromatography of
peptides. Electrophoresis 2003;24:4013–46.
8. Kilar F. Recent applications of capillary isoelectric focusing. Electrophoresis 2003;24:3908–16.
9. Eeltink S, Kok WT. Recent applications in capillary electrochromatography. Electrophoresis
2006;27:84–96.
10. Schou C, Heegaard NH. Recent applications of affinity interactions in capillary electrophoresis.
Electrophoresis 2006;27:44–59.
11. Babu CV, Song EJ, Babar SM, Wi MH, Yoo YS. Capillary electrophoresis at the omics level: towards
systems biology. Electrophoresis 2006;27:97–110.
12. Cooper JW, Wang Y, Lee CS. Recent advances in capillary separations for proteomics. Electro-
phoresis 2004;25:3913–26.
13. Kamoda S, Kakehi K. Capillary electrophoresis for the analysis of glycoprotein pharmaceuticals.
14. Little MJ, Paquette DM, Roos PK. Electrophoresis of pharmaceutical proteins: status quo. Electro-
phoresis 2006;27:2477–85.
15. Simo C, Barbas C, Cifuentes A. Capillary electrophoresis-mass spectrometry in food analysis.
16. Lookhart GL, Bean SR. Capillary electrophoresis of cereal proteins: an overview. J Capillary
Electrophor 2004;9:23–30.
17. Frazier RA, Papadopoulou A. Recent advances in the application of capillary electrophoresis for food
analysis. Electrophoresis 2003;24:4095–105.
18. Jabeen R, Payne D, Wiktorowcz J, Mohammad A, Petersen J. Capillary electrophoresis and the clinical
laboratory. Electrophoresis 2006;27:2413–38.
19. Fliser D, Wittke S, Mischak H. Capillary electrophoresis coupled to mass spectrometry for clinical
diagnostic purposes. Electrophoresis 2005;26:2708–16.
20. Kolch W, Neususs C, Pelzing M, Mischak H. Capillary electrophoresis-mass spectrometry as
a powerful tool in clinical diagnosis and biomarker discovery. Mass Spectrom Rev 2005;24:
959–77.
21. Bossuyt X, Lissoir B, Marien G, Maisin D, Vunckx J, Blanckaert N, Wallemacq P. Automated serum
protein electrophoresis by Capillarys. Clin Chem Lab Med 2003;41:704–10.
22. Petersen JR, Okorodudu AO, Mohammad A, Payne DA. Capillary electrophoresis and its application
in the clinical laboratory. Clin Chim Acta 2003;330:1–30.
23. Landers JP. Molecular diagnostics on electrophoretic microchips. Anal Chem 2003;75:2919–27.
24. Dolnik V, Liu S. Applications of capillary electrophoresis on microchip. J Sep Sci 2005;
28:1994–2009.
25. Fortina P, Surrey S, Kricka LJ. Molecular diagnostics: hurdles for clinical implementation. Trends Mol
Med 2002;8:264–6.
26. McLaren DG, Chen DD. A quantitative study of continuous flow-counterbalanced capillary elec-
trophoresis for sample purification. Electrophoresis 2003;24:2887–95.
27. Lacroix M, Poinsot V, Fournier C, Couderc F. Laser-induced fluorescence detection schemes for the
analysis of proteins and peptides using capillary electrophoresis. Electrophoresis 2005;26:2608–21.
28. Sanchez-Carbayo M. Antibody arrays: technical considerations and clinical applications in cancer.
Clin Chem 2006;52:1651–9.
29. Bruggeman FJ, Westerhoff HV. The nature of systems biology. Trends Microbiol 2007;15:45–50.
30. Prentice H, Webster KA. Genomic and proteomic profiles of heart disease. Trends Cardiovasc Med
2004;14:282–8.
31. Bloom FE. What does it all mean to you? J Neurosci 2001;21:8304–5.
32. Muddiman DC. “Lewis and Clark” proteomics. J Proteome Res 2006;5:221–2.
33. Bunn HF, Forget BG. Hemoglobin: Molecular, Genetic and Clinical Aspects. Philadelphia, PA:
W.B. Saunders Company, 1986.
34. Patrinos GR, Kollia P, Papadakis MN. Molecular diagnosis of inherited disorders: Lessons from
hemoglobinopathies. Hum Mutat 2005;26:399–412.
35. Fu Y, Lee AS. Glucose regulated proteins in cancer progression, drug resistance and immunotherapy.
Cancer Biol Ther 2006;5:741–4.
36. Becker JC, Kirkwood JM, Agarwala SS, Dummer R, Schrama D, Hauschild A. Molecularly targeted
therapy for melanoma: current reality and future options. Cancer 2006;107:2317–27.
37. Zimmermann GR, Lehar J, Keith CT. Multi-target therapeutics: when the whole is greater than the
sum of the parts. Drug Discov Today 2007;12:34–42.
38. Nadeau JH. Genetics. Modifying the message. Science 2003;301:927–8.
39. Nadeau JH. Listening to genetic background noise. N Engl J Med 2005;352:1598–9.
40. Anderson NL, Anderson NG. The human plasma proteome: history, character, and diagnostic prospects.
Mol Cell Proteomics 2002;1:845–67.
41. Mosher RA, Dewey D, Thormann W, Saville DA, Bier M. Computer simulation and experimental
validation of the electrophoretic behavior of proteins. Anal Chem 1989;61:362–6.
42. Pritchett T, Robey FA. Capillary electrophoresis of proteins. In: Landers JP, ed. Handbook of Capillary
Electrophoresis. Boca Raton, FL: CRC Press, 1997:260–90.
43. He Y, Yeung ES. Rapid determination of protein molecular weight by the Ferguson method and
multiplexed capillary electrophoresis. J Proteome Res 2002;1:273–7.
44. Karim MR, Janson JC, Takagi T. Size-dependent separation of proteins in the presence of sodium
dodecyl sulfate and dextran in capillary electrophoresis: effect of molecular weight of dextran.
45. Guttman A. Gel and polymer-solution mediated separation of biopolymers by capillary electrophoresis.
J Chromatogr Sci 2003;41:449–59.
46. Liu Z, Lemma T, Pawliszyn J. Capillary isoelectric focusing coupled with dynamic imaging detec-
tion: A one-dimensional separation for two-dimensional protein characterization. J Proteome Res
2006;5:1246–51.
47. Kuroda Y, Yukinaga H, Kitano M, Noguchi T, Nemati M, Shibukawa A, Nakagawa T, Matsuzaki K. On-
line capillary isoelectric focusing-mass spectrometry for quantitative analysis of peptides and proteins.
J Pharm Biomed Anal 2005;37:423–8.
48. Graf M, Watzig H. Capillary isoelectric focusing—reproducibility and protein adsorption. Elec-
trophoresis 2004;25:2959–64.
49. Martinovic S, Pasa-Tolic L, Smith RD. Capillary isoelectric focusing—mass spectrometry of proteins
and protein complexes. Methods Mol Biol 2004;276:291–304.
50. Storms HF, van der HR, Tjaden UR, van der GJ. Capillary isoelectric focusing-mass spectrometry for
shotgun approach in proteomics. Electrophoresis 2004;25:3461–7.
51. Shimura K. Recent advances in capillary isoelectric focusing: 1997–2001. Electrophoresis
2002;23:3847–57.
52. Righetti PG, Gelfi C, Bossi A, Olivieri E, Castelletti L, Verzola B, Stoyanov AV. Capillary elec-
trophoresis of peptides and proteins in isoelectric buffers: an update. Electrophoresis 2000;21:
4046–53.
53. Holland LA, Leigh AM. Bilayered phospholipid micelles and capillary electrophoresis: a new additive
for electrokinetic chromatography. Electrophoresis 2003;24:2935–9.
54. Maichel B, Kenndler E. Recent innovation in capillary electrokinetic chromatography with replaceable
charged pseudostationary phases or additives. Electrophoresis 2000;21:3160–73.
55. Roman GT, McDaniel K, Culbertson CT. High efficiency micellar electrokinetic chromatography of
hydrophobic analytes on poly(dimethylsiloxane) microchips. Analyst 2006;131:194–201.
56. Harwood MM, Christians ES, Fazal MA, Dovichi NJ. Single-cell protein analysis of a sin-
gle mouse embryo by two-dimensional capillary electrophoresis. J Chromatogr A 2006;1130:
190–4.
57. Michels DA, Hu S, Dambrowitz KA, Eggertson MJ, Lauterbach K, Dovichi NJ. Capillary siev-
ing electrophoresis-micellar electrokinetic chromatography fully automated two-dimensional cap-
illary electrophoresis analysis of Deinococcus radiodurans protein homogenate. Electrophoresis
2004;25:3098–105.
58. Betgovargez E, Knudson V, Simonian MH. Characterization of proteins in the human serum proteome.
J Biomol Tech 2005;16:306–10.
59. Stroink T, Ortiz MC, Bult A, Lingeman H, de Jong GJ, Underberg WJ. On-line multidimensional
liquid chromatography and capillary electrophoresis systems for peptides and proteins. J Chromatogr
B Analyt Technol Biomed Life Sci 2005;817:49–66.
60. Luo S, Feng J, Pang HM. High-throughput protein analysis by multiplexed sodium dodecyl sulfate
capillary gel electrophoresis with UV absorption detection. J Chromatogr A 2004;1051:131–4.
61. Cifuentes A, Poppe H. Rectangular capillary electrophoresis: some theoretical considerations.
Chromatographia 1994;39:391–404.
62. Wu J, Wu XZ, Huang T, Pawliszyn J. Analysis of proteins by CE, CIEF, and microfluidic devices with
whole-column-imaging detection. Methods Mol Biol 2004;276:229–52.
63. Wu XZ, Pawliszyn J. Whole-column fluorescence-imaged capillary electrophoresis. Electrophoresis
2004;25:3820–4.
64. Pasas SA, Lacher NA, Davies MI, Lunte SM. Detection of homocysteine by conventional and
microchip capillary electrophoresis/electrochemistry. Electrophoresis 2002;23:759–66.
65. Kuijt J, van Teylingen R, Nijbacker T, Ariese F, Brinkman UA, Gooijer C. Detection of nonderivatized
peptides in capillary electrophoresis using quenched phosphorescence. Anal Chem 2001;73:5026–9.
66. Hapuarachchi S, Janaway GA, Aspinwall CA. Capillary electrophoresis with a UV light-emitting diode
source for chemical monitoring of native and derivatized fluorescent compounds. Electrophoresis
2006;27:4052–9.
67. Sluszny C, He Y, Yeung ES. Light-emitting diode-induced fluorescence detection of native proteins in
capillary electrophoresis. Electrophoresis 2005;26:4197–203.
68. Yeung ES. Study of single cells by using capillary electrophoresis and native fluorescence detection.
J Chromatogr A 1999;830:243–62.
69. Dovichi NJ, Hu S. Chemical cytometry. Curr Opin Chem Biol 2003;7:603–8.
70. Underberg WJ, Waterval JC. Derivatization trends in capillary electrophoresis: an update. Elec-
trophoresis 2002;23:3922–33.
71. Kremser L, Petsch M, Blaas D, Kenndler E. Capillary electrophoresis of affinity complexes
between subviral 80S particles of human rhinovirus and monoclonal antibody 2G2. Electrophoresis
2006;27:2630–7.
72. Veledo MT, de Frutos M, Diez-Masa JC. Development of a method for quantitative analysis of the
major whey proteins by capillary electrophoresis with on-capillary derivatization and laser-induced
fluorescence detection. J Sep Sci 2005;28:935–40.
73. Zinellu A, Sotgia S, Usai MF, Chessa R, Deiana L, Carru C. Thiol redox status evaluation in
red blood cells by capillary electrophoresis-laser induced fluorescence detection. Electrophoresis
2005;26:1963–8.
74. Welder F, McCorquodale EM, Colyer CL. Proteinase assay by capillary electrophoresis employing
fluorescence-quenched protein-dye conjugates. Electrophoresis 2002;23:1585–90.
75. Presley AD, Fuller KM, Arriaga EA. MitoTracker Green labeling of mitochondrial proteins and
their subsequent analysis by capillary electrophoresis with laser-induced fluorescence detection.
J Chromatogr B Analyt Technol Biomed Life Sci 2003;793:141–50.
76. Stoyanov AV, Ahmadzadeh H, Krylov SN. Heterogeneity of protein labeling with a fluorogenic
reagent, 3-(2-furoyl)quinoline-2-carboxaldehyde. J Chromatogr B Analyt Technol Biomed Life Sci
2002;780:283–7.
77. Zhang H, Jin W. Single-cell analysis by intracellular immuno-reaction and capillary electrophoresis
with laser-induced fluorescence detection. J Chromatogr A 2006;1104:346–51.
78. Sowell J, Salon J, Strekowski L, Patonay G. Covalent and noncovalent labeling schemes for near-
infrared dyes in capillary electrophoresis protein applications. Methods Mol Biol 2004;276:39–75.
79. Heegaard NH, Nilsson S, Guzman NA. Affinity capillary electrophoresis: important application areas
and some recent developments. J Chromatogr B Biomed Sci Appl 1998;715:29–54.
80. Schultz NM, Huang L, Kennedy RT. Capillary electrophoresis-based immunoassay to determine insulin
content and insulin secretion from single islets of Langerhans. Anal Chem 1995;67:924–9.
81. Simpson DC, Smith RD. Combining capillary electrophoresis with mass spectrometry for applications
in proteomics. Electrophoresis 2005;26:1291–305.
82. Stutz H. Advances in the analysis of proteins and peptides by capillary electrophoresis with matrix-
assisted laser desorption/ionization and electrospray-mass spectrometry detection. Electrophoresis
2005;26:1254–90.
83. Eriksson KO, Palm A, Hjerten S. Preparative capillary electrophoresis based on adsorption of the
solutes (proteins) onto a moving blotting membrane as they migrate out of the capillary. Anal Biochem
1992;201:211–5.
84. Kaneta T, Inoue J, Koizumi M, Imasaka T. On-column capture of a specific protein in capillary
electrophoresis using magnetic beads. Electrophoresis 2006;27:3218–23.
85. Sharma U, Carbeck JD. Hydrodynamic radius ladders of proteins. Electrophoresis 2005;26:2086–91.
86. Ebersold MF, Zydney AL. Use of protein charge ladders to study electrostatic interactions during
protein ultrafiltration. Biotechnol Bioeng 2004;85:166–76.
87. Hempe JM, Vargas A, Craver RD. Clinical analysis of structural hemoglobin variants and Hb A1c by
capillary isoelectric focusing. In: Petersen JR, Mohammad AA, eds. Clinical and Forensic Applications
of Capillary Electrophoresis. Totowa, NJ: Humana Press, 2001:145–63.
88. Hempe JM, Craver RD. Laboratory diagnosis of structural hemoglobinopathies and thalassemias by
capillary isoelectric focusing. In: Palfrey SM, ed. Clinical Applications of Capillary Electrophoresis.
Totowa, NJ: Humana Press, 1999:81–98.
89. Biroccio A, Urbani A, Massoud R, di Ilio C, Sacchetta P, Bernardini S, Cortese C, Federici G. A
quantitative method for the analysis of glycated and glutathionylated hemoglobin by matrix-assisted
laser desorption ionization-time of flight mass spectrometry. Anal Biochem 2005;336:279–88.
90. Pastore A, Francesca MA, Tozzi G, Maria GL, Federici G, Bertini E, Lo Russo A, Mannucci L,
Piemonte F. Determination of glutathionyl-hemoglobin in human erythrocytes by cation-exchange
high-performance liquid chromatography. Anal Biochem 2003;312:85–90.
91. Takayama F, Tsutsui S, Horie M, Shimokata K, Niwa T. Glutathionyl hemoglobin in uremic
patients undergoing hemodialysis and continuous ambulatory peritoneal dialysis. Kidney Int Suppl
2001;78:S155–8.
92. Niwa T, Naito C, Mawjood AH, Imai K. Increased glutathionyl hemoglobin in diabetes mellitus and
hyperlipidemia demonstrated by liquid chromatography/electrospray ionization-mass spectrometry.
Clin Chem 2000;46:82–8.
93. Jakus V, Rietbrock N. Advanced glycation end-products and the progress of diabetic vascular
complications. Physiol Res 2004;53:131–42.
94. Ronai Z, Guttman A, Keszler G, Sasvari-Szekely M. Capillary electrophoresis study on DNA–protein
complex formation in the polymorphic 5 upstream region of the dopamine D4 receptor (DRD4) gene.
Curr Med Chem 2004;11:1023–9.
95. Donohue MJ, Satterfield MB, Dalluge JJ, Welch MJ, Girard JE, Bunk DM. Capillary electrophoresis
for the investigation of prostate-specific antigen heterogeneity. Anal Biochem 2005;339:318–27.
96. Sarg B, Helliger W, Talasz H, Forg B, Lindner HH. Histone H1 phosphorylation occurs site-specifically
during interphase and mitosis: identification of a novel phosphorylation site on histone H1. J Biol Chem
2006;281:6573–80.
97. Woods LA, Roddy TP, Ewing AG. Capillary electrophoresis of single mammalian cells. Electrophoresis
2004;25:1181–7.
98. Kraly JR, Jones MR, Gomez DG, Dickerson JA, Harwood MM, Eggertson M, Paulson TG, et al.
Reproducible two-dimensional capillary electrophoresis analysis of Barrett’s esophagus tissues. Anal
Chem 2006;78:5977–86.
99. Suresh Babu CV, Cho SG, Yoo YS. Method development and measurements of endogenous ser-
ine/threonine Akt phosphorylation using capillary electrophoresis for systems biology. Electrophoresis
2005;26:3765–72.
100. Liu X, Liu X, Liang A, Shen Z, Zhang Y, Dai Z, Xiong B, Lin B. Studying protein–drug interaction by
microfluidic chip affinity capillary electrophoresis with indirect laser-induced fluorescence detection.
101. Martinez-Gomez MA, Sagrado S, Villanueva-Camanas RM, Medina-Hernandez MJ. Characteriza-
tion of basic drug-human serum protein interactions by capillary electrophoresis. Electrophoresis
2006;27:3410–9.
102. Wu XZ, Zhang LH, Onoda K. Isoelectric focusing sample injection for capillary electrophoresis of
proteins. Electrophoresis 2005;26:563–70.
103. McKeon J, Holland LA. Determination of dissociation constants for a heparin-binding domain of
amyloid precursor protein and heparins or heparan sulfate by affinity capillary electrophoresis.
104. Veledo MT, de Frutos M, Diez-Masa JC. Analysis of trace amounts of bovine beta-lactoglobulin
in infant formulas by capillary electrophoresis with on-capillary derivatization and laser-induced
fluorescence detection. J Sep Sci 2005;28:941–7.
105. Laing TD, Marenco AJ, Moore DM, Moore GJ, Mah DC, Lee WE. Capillary electrophoresis laser-
induced fluorescence for screening combinatorial peptide libraries in assays of botulinum neurotoxin
A. J Chromatogr B Analyt Technol Biomed Life Sci 2006;843:240–6.
106. Balaguer E, Demelbauer U, Pelzing M, Sanz-Nebot V, Barbosa J, Neususs C. Glycoform characteri-
zation of erythropoietin combining glycan and intact protein analysis by capillary electrophoresis—
electrospray—time-of-flight mass spectrometry. Electrophoresis 2006;27:2638–50.
107. Neususs C, Demelbauer U, Pelzing M. Glycoform characterization of intact erythropoietin by capillary
electrophoresis-electrospray-time of flight-mass spectrometry. Electrophoresis 2005;26:1442–50.
108. Kinoshita M, Murakami E, Oda Y, Funakubo T, Kawakami D, Kakehi K, Kawasaki N, Mori-
moto K, Hayakawa T. Comparative studies on the analysis of glycosylation heterogeneity of sialic
acid-containing glycoproteins using capillary electrophoresis. J Chromatogr A 2000;866:261–71.
109. Hempe JM, Ory-Ascani J, Hsia D. Genetic variation in mouse beta globin cysteine content modifies
glutathione metabolism: implications for the use of mouse models. Exp Biol Med 2007;232:437–44.
110. Shihabi ZK, Hinsdale ME. Simplified hemoglobin chain detection by capillary electrophoresis.
111. Fu H, Guthrie JW, Le XC. Study of binding stoichiometries of the human immunodeficiency virus
type 1 reverse transcriptase by capillary electrophoresis and laser-induced fluorescence polarization
using aptamers as probes. Electrophoresis 2006;27:433–41.
112. Shimura K, Waki T, Okada M, Toda T, Kimoto I, Kasai K. Analysis of protein–protein interactions
with a multi-capillary electrophoresis instrument. Electrophoresis 2006;27:1886–94.
113. Montes-Bayon M, Profrock D, Sanz-MedelA, PrangeA. Direct comparison of capillary electrophoresis
and capillary liquid chromatography hyphenated to collision-cell inductively coupled plasma mass
spectrometry for the investigation of Cd-, Cu- and Zn-containing metalloproteins. J Chromatogr A
2006;1114:138–44.
114. Babu CV, Chung BC, Lho DS, Yoo YS. Capillary electrophoretic competitive immunoassay with
laser-induced fluorescence detection for methionine-enkephalin. J Chromatogr A 2006;1111:133–8.
115. Jackman R, Everest DJ, Schmerr MJ, Khawaja M, Keep P, Docherty J. Evaluation of a preclinical
blood test for scrapie in sheep using immunocapillary electrophoresis. J AOAC Int 2006;89:720–7.
116. Li G, Zhou X, Wang Y, El Shafey A, Chiu NH, Krull IS. Capillary isoelectric focusing and affinity
capillary electrophoresis approaches for the determination of binding constants for antibodies to the
prion protein. J Chromatogr A 2004;1053:253–62.
117. Roudiere L, Boularan AM, Bonardet A, Vallat C, Cristol JP, Dupuy AM. Evaluation of a capillary zone
electrophoresis system versus a conventional agarose gel system for routine serum protein separation
and monoclonal component typing. Clin Lab 2006;52:19–27.
118. Garcia-Ruiz C, Garcia MA, Garcia MC, Marina ML. Development of a capillary electrophore-
sis method for the determination of soybean proteins in soybean-rice gluten-free dietary products.
119. Arizaga RS, Blanco GE, Alvarez LG, Montes-Bayon M, Sanz-Medel A. Detection of transferrin iso-
forms in human serum: comparison of UV and ICP-MS detection after CZE and HPLC separations.
Anal Bioanal Chem 2005;383:390–7.
120. Joneli J, Lanz C, Thormann W. Capillary zone electrophoresis determination of carbohydrate-
deficient transferrin using the new CEofix reagents under high-resolution conditions. J Chromatogr A
2006;1130:272–80.
121. Lanz C, Kuhn M, Deiss V, Thormann W. Improved capillary electrophoresis method for the
determination of carbohydrate-deficient transferrin in patient sera. Electrophoresis 2004;25:2309–18.
3 Micellar Electrokinetic
Chromatography
Shigeru Terabe
CONTENTS
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

3.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
3.2.1 Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
3.2.2 Mass Spectrometric Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
3.2.3 Chemicals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
3.2.3.1 Surfactants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
3.2.3.2 Additives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
3.2.3.3 The Electrophoresis (Separation or Running) Solution . . . . . . . . . . . . . . . . . 113
3.2.4 Operating Conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
3.3.1 Retention Factor and Resolution Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
3.3.1.1 Number of Theoretical Plates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3.3.1.2 Selectivity Factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3.3.1.3 Retention Factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3.3.1.4 Migration Time Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
3.3.2 Interaction Mechanism between the Micelle and the Analyte . . . . . . . . . . . . . . . . . . . . . 116
3.3.3 Online Sample Preconcentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
3.3.3.1 Field-Enhanced Sample Stacking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
3.3.3.2 Sweeping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
3.3.3.3 Combination of Different Preconcentration Methods . . . . . . . . . . . . . . . . . . . 120
3.4.1 Pharmaceutical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
3.4.2 Clinical and Body Fluid Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
3.4.3 Food Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
3.4.4 Environmental Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
3.5 Methods Development Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
3.1 INTRODUCTION
Electrophoresis is a separation technique based on the migration of the analyte in a solution under
the influence of electric field. In electrophoretic separation it is essential for the analyte to be charged
or ionic. The migration velocity of the analyte is primarily depends on the charge and the size
of the analyte under a homogeneous electric field but it can also be modified by several chemical
or physical interactions between the analyte and the electrophoretic media including additives or
109
Surfactant Electroosmotic flow

Solute Electrophoresis
FIGURE 3.1 Schematic illustration of the principle of MEKC.
polymer networks. Neutral molecules were not considered to be separated by electrophoresis until
micellar electrokinetic chromatography (MEKC) was developed by Terabe et al. in 1984 [1]. The idea
using ionic micelles in electrophoresis was suggested by Nakagawa in 1981 as described by Terabe
et al. [1]. The neutral analyte will gain an apparent electrophoretic mobility when it is incorporated
into the ionic micelle and will migrate at the same velocity as the micelle under electrophoretic
conditions. Since the distribution equilibrium of the analyte between the micelle and the surrounding
aqueous phase can be quickly established, the apparent electrophoretic mobility or migration velocity
is directly related to the distribution coefficient. The more the analyte is incorporated, the higher the
mobility. The analyte also migrates by electroosmotic flow (EOF) under capillary electrophoresis
(CE) conditions, although EOF does not contribute to the separation at all. Thus, the migration
time in MEKC is a function of electrophoretic velocity of the micelle, distribution ratio, and EOF
velocity. A schematic principle of MEKC is shown in Figure 3.1, where EOF is stronger than the
electrophoretic migration of the anionic micelle and hence the anionic micelle migrates toward
cathode at a retarded velocity. The neutral analyte migrates at the velocity between EOF velocity
and that of the micelle and the migration velocity depends on how much fraction of the analyte is
incorporated into the micelle. It should be noted that there is optimum distribution of the analyte to
the micelle for good separation, that is, totally incorporated analyte by the micelle or free from the
micelle cannot be separated by MEKC. MEKC introduced the chromatographic principle into CE
by adding ionic micelles into the electrophoretic solution. MEKC is a branch of CE and extended
significantly the applicability of CE to a wide range of analytes, particularly small molecules. Since
MEKC also belongs to a chromatographic technique, it is familiar for most chromatographers to
design separation conditions.
MEKC can be easily performed just by adding ionic micelles to the electrophoretic solution
or background electrolyte solution in capillary zone electrophoresis (CZE), which is the simplest
separation mode of CE, without any instrumental modification. The ionic micelle added to the
background solution (BGS) is called pseudostationary phase because it plays a role as stationary
phase in chromatography, although it migrates inside the capillary. It should be noted that micelles are
classified as a pseudophase because the micelles exist only in equilibrium with surfactant molecules
in solution and cannot be isolated. The idea of separation of neutral analytes by electrophoresis was
extended to the use of other pseudostationary phases such as microemulsions, liposomes, charged
cyclodextrins (CDs), dendrimers, charged polymers, and so forth, and the technique can be generally
named as EKC [2].
Although MEKC has been widely accepted as a CE technique for the separation of neutral
analytes, it is also a powerful technique in improving selectivity of charged analytes in CE. MEKC
is almost 25 years old and popular among separation scientists as pointed out by a large number of
papers published, more than 2600 by 2005 [3], and coverage of the technique in the first edition of
this handbook. Therefore, this chapter does not describe much about the fundamental characteristics
of MEKC but the main target is to give the guiding principles for full availability of the technique.
Micellar Electrokinetic Chromatography 111
There are many books, book chapters, or review articles on fundamental MEKC [4–11]. In particular,
online sample preconcentration techniques are becoming popular in CZE to improve concentration
detection sensitivity and so are in MEKC, which will be introduced rather in detail. Although it is
generally stated that CE requires minimal amounts of sample (certainly true), it should be mentioned
that the volume of sample solution required is much greater than that actually introduced into the
capillary because the injection end of the capillary must be dipped into the sample solution—this is
typically at least a few microliters.
3.2 BACKGROUND
3.2.1 INSTRUMENTATION
No special modification of conventional CE instrumentation is required to perform MEKC. However,
the micellar solutions, prepared by dissolving a surfactant into a buffer solution, have higher electric
conductivity than the conventional CZE buffers and the current is usually high in MEKC. Therefore,
care must be taken to apply voltages that avoid excessive current (e.g., preferably less than 50 µA).
Otherwise, the separation efficiency will be deteriorated due to high Joule heating. Temperature of
capillary should be controlled higher than the Krafft point of the surfactant, which is the temperature
where the surfactant solubility increases rapidly with an increase in temperature. Krafft points change
depending on the solution, for example, the Krafft point of sodium dodecyl sulfate (SDS) is 15◦ C in
pure water, but it will be lower in buffer solutions.
Native fused-silica capillary with an inside diameter (ID) of 50 or 75 µm is generally used in
MEKC. To suppress EOF acidic conditions are employed. The use of surface-treated capillaries, such
as polyacrylamide or neutral coating, is effective to suppress EOF. Charged surface coating such as
successive multiple ionic layer (SMIL) coating [12] increases EOF even under acidic conditions or
reverses the direction of EOF. For example, the capillary is first coated with polybrene after rinsing
the surface followed by the coating with dextran sulfate, producing stable negatively charged surface
even at low pH. An additional third coating of polybrene produces stable positively charged surface.
Photometric detectors are the most popular in CE instruments including diode array detectors.
Laser-induced fluorescence (LIF) detection and electric conductivity detectors are also popular.
LIF is particularly sensitive and powerful for detecting low concentration analytes. However, most
analytes are not natively fluorescent and some derivatizations are necessary. Conductivity detector
is useful for the detection of non-ultraviolet (non-UV) absorbing analytes such as inorganic ions or
fatty acids. Both LIF detection and conductivity detectors are commercially available and easy to
interface with conventional CE instruments. Electrochemical detectors are also useful for selective
high-sensitivity detection. Several techniques have been developed to circumvent the problem of
strong effects of electrophoretic field on electrochemical detection, but despite this, commercial
electrochemical detectors are not used extensively.
3.2.2 MASS SPECTROMETRIC DETECTION

Mass spectrometry (MS) is becoming an indispensable detection method in every separation analy-
sis. High-performance liquid chromatography (HPLC)-MS is now a practical and robust technique
with electrospray ionization (ESI) or atmospheric pressure chemical ionization (APCI) interface to
generate analyte ions directly from the liquid phase. The same instrument can be utilized for CE-MS
with minor modifications. Since the liquid flow rate is significantly different between HPLC and CE,
some modifications are necessary; to keep the supply of the liquid to ESI interface a sheath liquid is
added through the coaxial ESI spray nozzle, which is also helpful to keep electrical connection for
CE run. A mixture of methanol and water-containing acetic acid or formic acid is usually used as a
sheath liquid at a flow rate around few µL/min in the positive ESI mode. The ESI voltage and CE
voltage are simultaneously applied at the spray nozzle but the CE current and ESI current are not
the same and the mismatch may cause some troubles interrupting current due to bubble formation
or unstable CE current. It is preferable to maintain the BGS at a low conductivity to minimize the
current mismatch. In some instruments, the spray nozzle can be grounded, thus solving the problem.
Since the presence of a high concentration of salts adversely affect the ionization efficiency in ESI,
volatile electrolytes such as ammonium formate or acetate are widely employed in CE buffer, which
minimizes contamination of the ion source. However, addition of the surfactant to BGS is essential
in MEKC, and most surfactants are nonvolatile contaminating the interface, and strong electrolytes
reduce the ionization efficiency in ESI. Several techniques have been developed to circumvent this
problem: partial-filling technique, reverse migration of the micelle, or the use of volatile surfactant.
The partial-filling technique is very effective for avoiding entrance of nonvolatile surfactant into the
ion source [13–17], but it is not a mature technique for MEKC-MS as of yet. Other techniques are less
popular and will not be introduced here. The more effective procedure for MEKC-MS is to perform
conventional MEKC with nonvolatile surfactant and electrolytes, noting that concentrations of SDS
and inorganic electrolytes are to be kept low (20 and 10 mM, respectively) [18,19]. Atmospheric
pressure photoionization (APPI) interface is a promising interface for MEKC, where up to 50 mM
SDS can be used [20].
3.2.3 CHEMICALS
3.2.3.1 Surfactants
Although a large number of surfactants are commercially available, limited numbers of surfactants
are widely utilized in MEKC. SDS is the most popular because it has many advantages over other
surfactants—specifically, it has high stability, relatively low Krafft point, low UV absorbance, high
solubilizing capability, and high quality reagent is readily available. Among several different sur-
factants frequently used in MEKC, bile salts such as sodium cholate and sodium deoxycholate, and
cetyltrimethylammonium bromide or cetyltrimethylammonium chloride (CTAB or CTAC) are also
useful to obtain different selectivity from that obtained with SDS. The effect of surfactant structures
on selectivity will be discussed later. The concentration of the surfactant must be higher than critical
micelle concentration (CMC), above which the surfactant forms micelles. The CMC of SDS is 8 mM
in water, but it will be lower (e.g., down to 3 mM) in buffer solutions. In most applications, SDS
concentrations are between 10 and 50 mM, but even 100 mM can be employed if the applied voltage
does not cause excessive current. Different surfactants will provide different selectivity for separa-
tion and, hence, surfactant choice is a means of selectivity manipulation. The mixed micelle, which
consists of two or more surfactants, can provide different selectivity in separation; in particular, the
mixed micelle between an ionic and a nonionic surfactant is considered of value because selectivity
is closely related to the surface structure of the micelle. Some surfactants such as bile salts and amino
acid derived surfactants are chiral and can be used for enantiomers separations.
3.2.3.2 Additives
As will be discussed later, selectivity can be manipulated by changing several parameters in addition
to using different surfactants. The choice of additives to modify distribution equilibrium can follow
the analogy to reversed-phase HPLC. Water miscible organic solvents such as methanol, acetonitrile,
2-propanol, and so forth, used as mobile phase modifiers in reversed-phase HPLC, are also effective
in MEKC. For hydrophobic or less water-soluble analytes, the organic solvent addition is useful to
reduce the distribution of the analyte to the micelle. The content of the organic solvents are usually
less than 30% to avoid the dissociation of the micelle. However, much higher concentration can be
used for separation, although it is not certain that the micelle is still stable. It should be noted that the
addition of the organic solvent to BGS affects viscosity; in particular, methanol or 2-propanol addition
increases viscosity significantly, and EOF is accordingly decreased because EOF mobility is inversely
proportional to viscosity. The addition of hydrophobic solvent may change the micellar structure
even if the concentration is low; for example, 1-butanol or 1-propanol may partly be incorporated
into the micelle as a cosurfactant, which causes a change in selectivity. Another advantage of the use
of an organic solvent is that the adsorption of the analyte on the capillary wall is minimized.
Cyclodextrins (CDs) are useful additives in MEKC to reduce the distribution of the analyte to
the micelle, improving the separation of highly hydrophobic analytes. CD addition can also change
selectivity, particularly for aromatic isomers. Since CDs are chiral, CD addition makes enantiomer
separation possible, and this is particularly useful for the separation of neutral enantiomers. In most
applications, β-CD or γ-CD is used at relatively high concentrations (e.g., 10–40 mM). The surfactant
molecule may be included into the cavity of the CD and, therefore, the inclusion complex formation
constant between the analyte and CD may be different from that observed in the absence of the
surfactant. Ion-pair reagents are useful in MEKC in a manner analogous to that in reversed-phase
HPLC for the separation of ionic analytes. However, the mechanism is slightly different between
MEKC and HPLC, because the micelle is charged [21]. A high concentration of urea [22] or glucose
[23] can be used to improve the separation of highly hydrophobic analytes or to change selectivity.
3.2.3.3 The Electrophoresis (Separation or Running) Solution

The micellar solution prepared by dissolving an ionic surfactant into a buffer solution is used as a
“running” solution in MEKC. Popular buffer solutions often used to prepare the micellar solution
are phosphate, borate, or tris(hydroxymethyl)aminomethane (Tris). It is well known the pH of the
buffer significantly affects the EOF; that is, EOF is almost completely suppressed below pH 2, low
at ∼pH 5 and strong (and relatively constant) above pH 7 when a native capillary is employed.
The concentration of buffer electrolytes is usually 20–50 mM, and while higher concentrations give
higher buffer capacity, this generally generates high current in addition to the contribution of the
ionic surfactant to conductivity. It should be noted that the counter ion of the ionic surfactant is
exchanged by the counter ion of the buffer electrolyte and, consequently, the character of the micelle
may be changed. The use of potassium-containing electrolytes must be avoided when using SDS
because potassium dodecyl sulfate has a higher Krafft point at room temperature, and the surfactant
will be precipitated out.
3.2.4 OPERATING CONDITIONS

No special care is needed to perform MEKC compared with the CZE operation described in Chapter 1.
The difference between the two techniques is only the electrophoretic solution or running solution.
The current tends to be higher in MEKC, requiring a more frequent replenishment of the running
solution to avoid the pH change in running solution associated with buffer depletion. When a cationic
surfactant such as CTAB is employed, the direction of EOF is reversed and regeneration of the surface
needs to be accomplished through sufficient rinsing of the capillary. It should be noted that CTAB
in the anodic vial will generate bromine by electrolysis which ultimately contaminates the running
solution.

3.3.1 RETENTION FACTOR AND RESOLUTION EQUATION
The migration behaviors of the analyte, the micelle marker, and the EOF marker are shown in
Figure 3.2a, where the corresponding electropherogram is also shown (Figure 3.2b). The EOF
marker migrates at the velocity of EOF and the micelle at the velocity of difference between EOF
and the electrophoretic velocity of the micelle, which is in the opposite direction of EOF. The neutral
analyte migrates at the velocity between the two extremes: the analyte is assumed to be equally
distributed between the micelle and the aqueous phase in Figure 3.2. EOF mobility is assumed to be
(a)
Micelle Solute Water
Injection Capillary detection
(b)
Water Solute Micelle
0 t0 tR t mc
FIGURE 3.2 Schematic illustration of the migration behavior in MEKC. (From Terabe, S. et al., Anal. Chem.,
57, 834, 1985. With permission.)
4/3 times larger than electrophoretic mobility of the micelle in absolute values or the migration time
of the micelle, tmc , is four times longer than that of the EOF marker, t0 . The EOF marker is free from
the micelle throughout the progress and methanol is often used as a marker. Although methanol does
not absorb UV, it can be detected due to the fluctuation of the reflective index inside the capillary.
Sudan III or IV is often used as a micelle marker, because it is assumed to be totally incorporated
into the micelle. Some cationic hydrophobic compounds such as quinine sulfate can be employed
as an SDS marker, because quinine interacts strongly with the SDS micelle due to electrostatic and
hydrophobic interaction.
The retention factor, k, can be defined as
nmc
k= , (3.1)
naq
where naq and nmc are the amount of the analyte in the aqueous phase and that incorporated into the
micelle. It can be calculated from the migration times tR of the analyte, t0 , and tmc as follows [1]:
tR − t0
k= . (3.2)
t0 1 − tR/tmc
The resolution equation in MEKC can be given by [24]

√
N α−1 k2 1 − t0 tmc
Rs = , (3.3)
4 α 1 + k2 1 + t0 tmc k1
where N is the number of theoretical plates and α is the selectivity factor equal to k2 /k1 , which
is assumed to be larger than 1, where k2 and k1 are retention factors of analyte 1 and analyte 2,
respectively. Equations 3.2 and 3.3 are similar to those of the conventional chromatography but
different due to the limited migration time window for the analyte between t0 and tmc . The last factor
(parameter) in right-hand side of Equation 3.3 is not part of the conventional resolution equation,
and is ascribed to the change in effective capillary length utilized for separation. In other words, the
micelle is not stationary in the capillary but migrates during the separation process, which makes
the length of the micelle interacting with the analyte in the capillary shorter than the physical length
by the distance the micelle migrates during tR . We may define this parameter as the migration time
window contribution, although the last factor becomes null when EOF is absent or the micelle
is stationary. The effect of each factor in Equation 3.3 is briefly discussed below for reasonable
optimization of MEKC separation.
3.3.1.1 Number of Theoretical Plates

Column efficiency, expressed as the number of theoretical plates (N: plate numbers), has been
shown to be as high as 100,000–200,000 in MEKC. The most important factor contributing to the
plate height is the longitudinal diffusion as in CZE—the shorter the migration time, the higher plate
number. Since the diffusion coefficient of the micelle is usually one order lower than that of the
small molecule, higher plate numbers can be expected for analytes having larger retention factors or
those more incorporated into the micelle. The next important factor to be considered is extracolumn
effects; in particular, the injection sample volume or sample plug length. In general, the sample plug
length should be restricted to less than 1% of the capillary length unless sample preconcentration
occurs. The use of a longer capillary alleviates the large volume injection problem, although the
plate number is not proportional to the capillary length, as in other CE modes. It should be noted that
if the initial sample zone length is minimal, extremely high efficiency can be expected. Temperature
effects are not serious if current is lower than 50 µA, and it should be noted that the adverse effects
on efficiency often observed are due to the sample adsorption on the capillary wall. Thorough rinsing
of the capillary or an addition of a small amount of organic solvent may solve the problem. If the
plate number is lower than 100,000, care must be taken to identify the reason of low efficiency;
overloading of the sample or adsorption of the analyte on the capillary wall should be considered.
3.3.1.2 Selectivity Factor

As in chromatography, selectivity factor is the most important parameter in optimizing the separa-
tion. Since the micelle corresponds to the stationary phase in chromatography, the selection of the
surfactant is of primary importance. SDS is a good initial choice and, if not successful, bile salts
or CTAB/CTAC should be the second choice. If other surfactants are available, they may be tested.
However, the number of different surfactants available for MEKC is rather limited and, hence, the
selection of additives to modify selectivity should be considered when expected separation is not
obtained by changing the surfactant. Some of the additives suggested above should be considered,
as well as the use of mixed micelles. In MEKC, the efficiency is much higher than that of HPLC,
the minimum α that provides successful separation may be as low as 1.02.
3.3.1.3 Retention Factor

The retention factor is contained in the third (retention factor) and fourth (migration time window)
parameter in Equation 3.3. Therefore, the optimum k value, kopt , that gives the maximum value for
the product of third and fourth parameters is a function of k, t0 , and tmc , and given by [25]

kopt = tmc t0 . (3.4)
When SDS is used under neutral or alkaline conditions in native or uncoated capillary, the value
tmc /t0 is usually 3–4 and the retention factor should be adjusted to about 1.7–2.0. Since the retention
factor is related to the surfactant concentration by [24]
KVmc ∼
k= = K v̄ (Csf − cmc ) , (3.5)
Vaq
where K is the distribution coefficient of the analyte between the micelle and the aqueous phase, Vmc
and Vaq are volume of the micelle and the aqueous phase, respectively, v̄ partial specific volume of the
micelle, and Csf is the surfactant concentration. Equation 3.5 suggests that k is nearly proportional
to the surfactant concentration; this means that the phase ratio Vmc /Vaq is adjusted by changing
the surfactant concentration. Therefore, the optimization of k is straightforward provided that the
specific volume and CMC are known; the v̄ value for an SDS micelle is about 0.85 mL/g. Even if these
values are not known, it is easy to adjust k by increasing or decreasing the surfactant concentration
considering the linear dependence of k on Csf .
3.3.1.4 Migration Time Window

The EOF velocity is difficult to control precisely and depends on the surface charge on the inside
wall of the capillary. If the wall is negatively charged, EOF is cathodic and vice versa. The factors
affecting EOF are described above. In general, if the EOF velocity is decreased, the migration time
window becomes wide and the resolution will be improved at the expense of migration time. The
detailed discussion on the effect of EOF on resolution is published by Zhang et al. [26].
3.3.2 INTERACTION MECHANISM BETWEEN THE MICELLE AND THE ANALYTE

Three types of interaction mechanisms are known between the micelle and the analyte as shown
in Figure 3.3: (1) incorporation of the analyte into the hydrophobic core, (2) adsorption of the
analyte on the surface or on the palisade layer, and (3) incorporation of the analyte as a cosurfactant.
Highly hydrophobic and nonpolar analytes such as aromatic hydrocarbons will be incorporated
into the core of the micelle. The selectivity may not be very different among long alkyl-chain
surfactants for this class of analyte but the distribution coefficient will be increased with longer
alkyl-chain surfactants. Thus, selectivity will not be altered significantly for nonpolar hydrophobic
analytes, even when different surfactants are used. However, bile salts may provide substantially
different selectivity in comparison with long-alkyl chain surfactants, even for nonpolar hydrophobic
analytes.
Most analytes are considered to be incorporated by the micelle on the surface or on the palisade
layer [Figure 3.3a (1)] or partly incorporated into the core at the hydrophobic group and on the surface
at the polar group [Figure 3.3a (2)]. Therefore, the polar group of the surfactant significantly affects
selectivity. The palisade layer will be much different in mixed micelles consisting of an ionic and a
nonionic surfactant and, hence, selectivity will be significantly different from the single surfactant
micelle (Figure 3.3b).
(a) (2) (b)

(4)
(3)
(1)
FIGURE 3.3 Schematic illustration of micellar solubilization. (a) Ionic micelle and (b) mixed micelle of ionic
and nonionic surfactants interacting (1) with the hydrophobic core, (2) on the surface, (3) as a cosurfactant, and
(4) with nonionic surface. (From Terabe, S., Anal. Chem., 76, 240A, 2004. With permission.)
The interaction mechanism between the micelle and the analyte or chemical selectivity in MEKC
has been investigated by several groups. Linear solvation energy selectivity relationship (LSER) can
be utilized, where structural descriptors such as size, dipolarity, and hydrogen-bonding capabilities
are employed to evaluate their contribution to interaction (retention factor) using many different
types of solutes and different surfactants [27]. As expected, the molecular size or hydrophobicity
causes the strongest positive interaction due to unfavorable energy term for the formation of properly
sized cavity in the solvent system for solute accommodation. The second most important term in
most MEKC systems is the solute hydrogen bond acceptor capability, but this interaction negatively
affects the interaction. Since hydrophobicity affects the interaction almost equally among different
MEKC systems, fine tuning in selectivity is performed by the hydrogen bond acceptor capability or
basicity of the solute or acidity of the surfactant. Similar conclusions have been obtained by other
groups [28–30].
3.3.3 ONLINE SAMPLE PRECONCENTRATION

Poor concentration detection sensitivity in CE is a serious problem in almost all CE modes. The main
reasons are due to the small amount of sample introduced into the capillary and a short optical path
length for photometric detectors. To solve the problem, three major solutions are available: sample
pretreatment such as solid-phase extraction (SPE), the use of high sensitivity detectors such as LIF,
and the implementation of online sample preconcentration techniques. In addition, extended path
length cells such as bubble cell or Z-type cell are commercially available for photometric detection,
enabling a 3- to 10-fold enhancement in sensitivity. Most online sample preconcentration techniques
are designed not to require any modification of the instrument and to be compatible with any detection
methods because the concentration of the sample is increased prior to separation. Several techniques
have been developed: transient isotachophoresis (t-ITP), field-enhanced (amplified) sample stacking,
sweeping, and dynamic pH junction. The sample zone injected as a much longer plug than normal
can be focused to the length less than that of the normal injection. These techniques are based on the
migration velocity change of the analytes, between the sample solution zone and BGS, and the analyte
must have an electrophoretic mobility even if it is apparent mobility. Among these methods, t-ITP and
dynamic pH junction are applicable only to ionic or ionizable analytes. Therefore, field-enhanced
sample stacking and sweeping are the main techniques described.
3.3.3.1 Field-Enhanced Sample Stacking

The principle of the technique is simple and is based on the fact that the electrophoretic migration
velocity is proportional to the field strength. Thus, the sample solution is prepared in a lower conduc-
tivity solution than that of the BGS; then the analyte ion migrates rapidly in the sample zone when an
electric field is applied. The analyte ion quickly migrates until it reaches the boundary between the
sample zone and the BGS zone, where the migration velocity decreases upon entering the BGS zone
due to low electric field as shown in Figure 3.4. The larger the difference in electric fields between
the sample zone and BGS, the higher the focusing efficiency. However, it is noteworthy that EOF
velocity is also proportional to the field strength, but the liquid flow by EOF must be homogeneous
throughout the capillary. That is, if the EOF velocity in the two zones is mismatched, mixing of the
two zones at the boundary must occur, which can seriously deteriorate the concentration efficiency.
Therefore, it is desirable to suppress EOF when online sample stacking is performed. There are
several techniques based on the same principle introduced by Chien and Burgi [31].
In principle, field-enhanced stacking is a technique for the concentration of charged analytes
using discontinuous solutions in the capillary. Neutral analytes are not concentrated by this method
because they do not have electrophoretic mobility. However, by utilizing charged micelles, an
apparent electrophoretic mobility can be imparted to a neutral analyte according to the principles
of MEKC. As in conventional field-enhanced sample stacking, sample solutions are prepared in a
Stacking
boundary (SB)
Detector
BGS zone Sample zone BGS zone
High-
Low-conductivity
conductivity
SB
E Electric field strength
FIGURE 3.4 Schematic illustration of the principle of field-enhanced sample stacking.
low electrical conductivity solution, with or without adding micelles, and a conventional micellar
solution is used as the BGS. When the sample solution is prepared without the micelle, the analyte
is incorporated by the micelle that enters the sample zone from the BGS zone when voltage is
applied. The analyte having a high k that will induce extensive incorporation into the micelle can
be efficiently concentrated by field-enhanced stacking because it is apparently highly charged. It
should be noted that the micellar concentration is low in the sample zone even if the sample solution
contains the micelle, because the migration velocity of the micelle in the sample zone is faster than
that in the BGS. As described in Chapter 1, there are two sample injection techniques amenable to
field-enhanced stacking: pressurized/hydrostatic injection and electrokinetic injection. With either
technique, an EOF suppressed condition will yield a higher concentration efficiency. Larger amounts
of sample can, therefore, be injected by electrokinetic injection, while with pressurized injection the
sample, volume is limited to that substantially less than the capillary volume. In summary, either
injection method will yield a maximum concentration efficiency of less than 100-fold under favorable
conditions [32,33].
3.3.3.2 Sweeping
Sweeping is an online concentration technique developed for MEKC where the neutral analyte is
efficiently concentrated [34,35]. The principle of sweeping is shown in Figure 3.5 under suppressed
EOF conditions. Sweeping seems similar to field-enhanced sample stacking but the difference is
that no field enhancement in the sample zone is required in sweeping with a field strength equal
to or lower [36] than that of BGS. The sample solution is prepared devoid of the micelle but at a
conductivity similar to that of BGS (acidic to suppress EOF) and injected hydrodynamically as a
long plug (Figure 3.5a). When the voltage (negative for anionic micelle such as SDS) is applied,
the micelle enters the sample zone and collects analytes at the front end of the entering micellar
zone (Figure 3.5b) until the front end reaches the original boundary between the sample zone and
the BGS (Figure 3.5c). The original sample matrix migrates at the velocity of the micelle and the
micelle vacancy zone migrates ahead of the swept sample zone (Figure 3.5c). The length of the swept
analyte zone is given by [34]

1
lsweep = linj , (3.6)
1+k
To the
detector
Injection
Sample BGS
(a) [micelle = 0] [micelles]
Micelles Micelles
_
(b) +
Concentrated [Micelle] = 0
zone
(c) _ +
FIGURE 3.5 Schematic illustration of the principle of sweeping under suppressed EOF conditions. The sample
solution is prepared devoid of the micelle and BGS is the micellar solution. (a) Sample injection as a long plug.
(b) Start of sweeping by application of a voltage. (c) End of sweeping and start of MEKC separation.
where lsweep and linj are lengths of the sample zone after and before the end of sweeping, respectively.
According to Equation 3.6, to obtain high concentration efficiency, k should be maximized. It should
be noted that k in Equation 3.6 is the value in the sample matrix; that is, it is different from that in
BGS. Therefore, the additive to BGS such as an organic solvent or CD used to improve resolution
should not be added to the sample matrix. The zone length of the swept zone of a hydrophobic analyte
was extremely narrow as observed when the sweeping process was traced with the peak shape of the
sample zone at different positions in the microchannel [37]. Under favorable conditions, 64 cm of
70 cm of effective capillary length was filled with a dilute sample solution and 6 cm of 70 cm was
used for the MEKC separation, yet good separation and high efficiency was obtained because the
sample zone length after sweeping was very narrow and the time spent for the separation was short.
EOF is not related to the sweeping mechanism, but higher concentration efficiency was obtained
under suppressed EOF conditions [35]. Consequently, sweeping gives high concentration efficiency
with improvements in detection sensitivity on the order of 1000-fold greater than the conventional
injection/separation.
Although sweeping is an online sample preconcentration technique for neutral analytes, charged
analytes, or ions can be more efficiently concentrated if the micelle employed has a polarity opposite
to that of the analyte, because the retention factor will be large due to a strong electrostatic interaction.
Therefore, the SDS micelle is convenient for concentrating cationic analytes, while cationic micelles
(e.g., CTAB) are suitable for anionic analytes such as aromatic sulfonates or carboxylates [38].
Sweeping is effective in other EKC modes that use pseudostationary phases such as charged CDs,
microemulsions, and so forth. In addition, the sweeping principle is applicable to online concentration
of metal ions and sugars by utilizing an in-capillary complexation reaction, sweeping of metal ions
by ethylenediaminetetraacetic acid (EDTA) [39], or sweeping of sugars (cis-diols) using borate
ions [40].
As mentioned before, when pressure-based sample injection is employed, the maximum injection
volume must be less than the effective length of the capillary. At least 10% of the effective capillary
length must remain available for separation. To inject a larger sample volume, electrokinetic injection
must be employed. In most online sample preconcentration techniques, the maximum amount that
can be injected without loss of separation efficiency is certainly less than the capillary volume.
However, with a large volume sample injection under cathodic EOF conditions in SDS MEKC, a
sample devoid of the micelle can be continuously electrokinetically injected for a volume equivalent
to seven times the capillary volume without significant loss of separation efficiency under favorable
conditions. Here, the analyte in the sample is continuously concentrated when it reaches the swept
zone [41].
3.3.3.3 Combination of Different Preconcentration Methods

A combination of the different online sample preconcentration techniques enables much higher con-
centration efficiency or simultaneous concentration of different types of analytes. First, cationic (or
anionic) analytes can be electrokinetically injected for a extensive periods of time to cause over-
loading. Second, the long preconcentrated sample zone resulting from the injection can be further
concentrated by sweeping using anionic (or cationic) micelle. This combination of field-enhanced
sample injection (FESI) and sweeping is termed cation-selective (or anion-selective) exhaustive
injection-sweeping (CSEI/ASEI-sweeping) and can generate extraordinarily high concentration effi-
ciency with close to a million-fold increase in sensitivity, enabling the analysis of samples with
analytes present in as low as parts per trillion (ppt) levels [42,43].
Dynamic pH junction is a technique suitable for online sample preconcentration of weakly acidic
or basic compounds, with the principle based on the change in electrophoretic velocity between that
of the ionized and neutral state of the analyte [44]. The sample solution is prepared in an acidic
medium for the acidic analyte with the BGS a basic buffer having a pH higher than the pKa of
the analyte. When electrophoretic process begins, the acidic sample zone is titrated by the alkaline
BGS and the pH of the sample zone increases, causing deprotonation of the analyte. The ionized
analyte migrates toward the anode against the EOF and focuses at the boundary of pH change. The
technique is very effective for the selective concentration of acidic or basic analytes having a range of
pKa values. When the sample contains both neutral and weak acidic compounds, the combination of
sweeping and dynamic pH junction is a useful online sample preconcentration method. For example,
with urine or plasma containing a combination of riboflavin, flavin mononucleotide (FMN), and
flavin adenine dinucleotide (FAD), an enhancement in detection sensitivity of more than 1000-fold
was observed for these analytes by dynamic pH junction-sweeping with the sample diluted in a
75 mM phosphate buffer pH 6.0 (and BGS of 140 mM borate) because almost neutral riboflavin
was concentrated by sweeping and other FMN and FAD were concentrated by dynamic pH junction
simultaneously [45].

More than 2000 articles involving “MEKC” have been published since 1990 according to Zare
[3]. Therefore, it is impossible to survey all papers to determine which provide the best practical
applications of MEKC—consequently, only few practical applications have been arbitrarily selected
by the author for inclusion here. For more detailed examples, it is recommended that the reader consult
more comprehensive reviews [7,46–48] or relevant chapters of a recently published book [6].
3.4.1 PHARMACEUTICAL ANALYSIS

MEKC is probably most widely accepted in pharmaceutical analysis and several review articles have
detailed these applications [49–53]; therefore, only a few recent publications are introduced here. In
addition, the reader may refer to Chapter 4 in this book contributed by Altria and colleagues.
As mentioned above, MS is becoming an indispensable detection technique, in particular, for the
detection and identification of impurities in drugs. An ESI has been successfully interfaced between
MEKC and a single quadrupole or ion-trap MS instrument using a running solution containing 20 mM
SDS and 10 mM sodium phosphate buffer (pH 7.5) without any devices to avoid SDS and inorganic
electrolytes from entering into the MS ion source [18]. In this report, it was demonstrated that
sub-µg/mL levels of mebeverine and related compounds could be detected in full-scan mode, while
5.0 × 106
m/z 184
0.0
8.0 × 106
Intensity (a.u)
m/z 332
0.0
1.3 × 106
m/z 314
0.0
0 5 10 15
Time (min)
FIGURE 3.6 Extracted-ion chromatograms obtained during MEKC–ESI-MS of a heat-treated solution of

ipratropium (1 mg/mL) using a BGE containing 10 mM sodium phosphate (pH 7.5), 20 mM SDS, and 12.5%
acetonitrile. Capillary length, 57 cm. Peaks: m/z 332, ipratropium; m/z 184 and m/z 314, degradation products.
(From Mol, R., et al., J. Chromatogr. B, 843, 283, 2006. With permission.)
detection limits are in the 10–50 ng/mL range when selected ion monitoring was applied. It was also
shown that 0.1% (w/w) levels of potential impurities in mebeverine could be detected in full-scan
MS. The technique was also successfully applied to the analysis of a galantamine sample containing a
number of related impurities [19]. Two degradation products were detected and identified by MS/MS
scan in a heat-stressed ipratropium sample as shown in Figure 3.6.
In the pharmaceutical industry, rapid developments and validations of suitable methods for
assessment of purity, potency, and stability of new drug substances and drug products are critical
to providing appropriate data for early project development decisions. For example, a simple, fast,
and selective MEKC method for the simultaneous assay of ketorolac tromethamine and its identified
related impurities in both the drug substance and coated tablets was developed by Orlandini et al.
[54]. To optimize separation conditions, a response surface study was performed. Flufenamic acid
(FL) and tolmetin (TL) were chosen as internal standards to quantify ketorolac tromethamine and
the impurities, respectively. The optimized BGS consisted of a mixture of 13 mM boric acid and
phosphoric acid (equal mixture), adjusted to pH 9.1 with 1 M sodium hydroxide, containing 73 mM
SDS. Optimal temperature and voltage were 30◦ C and 27 kV. Under these conditions, all compounds
were resolved in about 6 min as shown in Figure 3.7. The related substances could be quantified
up to the 0.1% (w/w) level. Method validation was performed by evaluating selectivity, robustness,
linearity and range, precision, accuracy, detection sensitivity, and quantification limits, as well as
system suitability for drug substances/drug product and a satisfactory method performance was
obtained.
mAU
KT TL FL
8
6
KK
4
HK DK
1 2 3 4 5 min
FIGURE 3.7 MEKC electropherogram of a real sample of Lixidol tablets using 73 mM SDS in 13 mM borate-
phosphate (equal molar) buffer (pH 9.1). FL, flufenamic acid; KT, 1:1 mixture of ketorolac tromethamine
and 2-amino-2-(hydroxymethyl)-1,3-propanediol; DK, decarboxylated ketorolac; HK, 1-hydroxy analog of
ketorolac; KK, 1-keto analog of ketorolac; TL, tolmetin. Capillary, 50 µm ID × 48 cm (39.5 cm to the detector);
applied voltage, 27 kV; temperature, 30◦ C; hydrodynamic injection: 50 mbar, 5 s; detection wavelength, 323 nm.
(From Orlandini, S., et al., J. Chromatogr. A, 1032, 253, 2004. With permission.)
3.4.2 CLINICAL AND BODY FLUID ANALYSIS

Development of the analytical separation methods for biomarker discovery and metabolite research
is an important domain for the early detection of certain diseases, as well as in pharmaceutical
investigations. Recent applications of MEKC and CZE in the field of urinary biomarkers have been
summarized by Iadarola et al. [55]. Most target analytes are present in body fluids at low concentra-
tions and, consequently, high sensitivity analysis is essential. The use of high sensitivity detectors or
implementation of online sample concentration techniques is essential. Sample pretreatment is also
critical in the analysis of body fluids for cleaning up and concentrating the target analytes, and for
this several techniques are available.
Direct injection of plasma samples in MEKC was first reported by Nakagawa et al. [56]. Deter-
mination of cefotaxime (C) and its deacetylated metabolite (DA) in human plasma were investigated
by direct injection of plasma in both CZE and MEKC [57]. MEKC was shown to be superior
with regard to simplicity, rapidity, precision, and sensitivity relative to CZE. In MEKC, plasma
samples spiked with C, DA, and theobromine (as internal standard) were directly injected after
dilution with water, and analyzed using a phosphate buffer (pH 8.0) containing 165 mM SDS as a
running solution. The interday precision (n = 4 days) was 1.49% in RSD when theobromine was
used as internal standard. A satisfactory interday precision between slopes was also obtained with
MEKC, even without the use of an internal standard (RSD = 4.38%). Detection limits (S/N = 3)
were about 1 mg/L in plasma for C and DA. The advantage of MEKC is due to the fact that
SDS solubilizes the proteins, and this liberates the drugs from protein-binding sites. The SDS–
protein complexes migrate with the micelles and, thus, do not interfere with the detection of the
analytes.
An MEKC-sweeping technique for the simultaneous determination of flunitrazepam and its major
metabolites, 7-aminoflunitrazepam and N-desmethylflunitrazepam, was reported by Huang et al.
[58]. The optimized conditions involved a running solution of 25 mM borate (pH 9.5) containing 50
mM CTAB and 30% methanol (v/v) with a 60 cm (50 cm to the detector) × 50 µm ID capillary and
sample injected as a 151-mm plug. The limits of detection were 13.4, 5.6, and 12.0 ng/mL for flu-
nitrazepam, 7-aminoflunitrazepam, and N-desmethylflunitrazepam, respectively, and the sensitivity
enhancement for each compound was within the range of 110- to 200-fold. The method was applied
1 mAU
Absorbance 1 3
0 2 4 6 8 10 12
FIGURE 3.8 Sweeping MEKC electropherogram of a spiked urine sample. Peak identification: 1, 7-
aminoflunitrazepam; 2, flunitrazepam; 3, N-desmethylflunitrazepam. Analyte concentration: 0.3 µg spiked
in a 3-mL urine sample before SPE extraction. Conditions: capillary, 60 cm long (50 cm to detector), 50 µm ID;
running solution: 25 mM borate buffer (pH 9.5) containing 50 mM CTAB and 30% methanol (v/v), conductivity
7.28 mS/cm; sample matrix: 25 mM borate buffer (pH 9.5); applied voltage: -25 kV; UV detection at 240 nm.
(From Huang, C.-W., et al., J. Chromatogr. A, 1110, 240, 2006. With permission.)
to a spiked urine sample in combination with SPE — an example of electropherogram is shown in

Figure 3.8.
Ewing’s group has developed ultramicro separation technique using CE with an electrochemical
detector and a narrow bore capillary of 13 µm ID. They applied the technique to the separation of
amine metabolite in the fruit fly, Drosophila melanogaster, and 14 biogenic amines and metabolites
were successfully resolved by MEKC using 25 mM borate buffer (pH 9.5) containing 50 mM SDS and
2% 1-propanol [59]. Flies were snap-frozen in a liquid nitrogen bath, the separated fly heads treated
to yield a 3 µL sample solution that could be analyzed for biogenic amines as shown in Figure 3.9.
Quantitative results were, for example, 37.2, 311.0, 142.0, 79.5, 732.8 fmol (1 × 10−15 mol) per
fly head of a wild-type Drosophila, for tyramine, octopamine, N-acetyloctopamine, dopamine, and
N-acetyldopamine, respectively. MEKC with electrochemical detection gave limits of detection as
low as 3.4 amol for octopamine, and made the technique useful for volume-limited sample analysis.
Kennedy’s group used microdialysis sampling to monitor extracellular dopamine concentration
in the brains of rats. The dialysate, mixed online with 6 mM naphthalene-2,3-dicarboxaldehye and
10 mM potassium cyanide in a reaction capillary, was periodically analyzed by MEKC at 90-s
intervals [60]. The MEKC system consisted of a 10 µm ID, 369 µm OD, and 16 cm (14.5 cm
to the detector) capillary, 30 mM phosphate buffer (pH 7.4) containing 6.5 mM SDS and 2 mM
2-hydoxypropyl-β-CD, with LIF detection using the 413-nm line of a 14-mW diode-pumped laser,
and an electric field of 850 V/cm. The detection limit for dopamine was 2 nM when sampling by
microdialysis. The separation capability of the developed method is illustrated in Figure 3.10 where
(a)
10 pA
3
5 6
7
1
5.5 9
9
(b)
20 pA
11
9 11
(c)
5 pA
12
11 15
(d) 11
10 pA
12
wt
iav
10 Time (min) 15
FIGURE 3.9 MEKC-EC separation of a Drosophila head homogenate. (a) Enlargement of 5.5–9 min high-
lighting dopamine (1), N-acetyloctopamine (3), octopamine (5), N-acetylserotonin (6), and N-acetyldopamine
(7). (b) Enlargement of 9–11 min emphasizing peaks for 4-dihydroxyphenylalanine (9) and catechol (11). (c)
Enlargement of 11–15 min showing tyramine (12). (d) Comparison of Canton-S (wt) and iav1 Drosophila head
homogenates highlighting the internal standard catechol (11) and tyramine (12). Conditions: running solution,
25 mM borate buffer (pH 9.5) containing 50 mM SDS and 2% 1-propanol; capillary, a fused silica capillary of
148 µm OD and 13 µm ID, 50 cm long; Field strength, 333 V/cm. The working electrode was held at +750 mV
vs. a Ag/AgCl reference electrode. (From Paxon, T.L., et al., Anal. Chem., 77, 5349, 2005. With permission.)
the basal dopamine concentration in dialysates collected from the striatum of anesthetized rats was
18 ± 3 nM (n = 12).
3.4.3 FOOD ANALYSIS

Applications of CE techniques to food analysis is rapidly increasing and recent progresses have been
reviewed [61,62]—a detailed discussion of this topic can be found in Chapter 30 by Vargas and
Cordoba. The analysis of food components or nutrients does not require high sensitivity but does
(a) 100 Gln

Ser Tau
80 Arg
60 Ile
Leu
Gly
40
20 Glu
His Asp
Cit
0
35 45 55 65 75 85
Migration time (s)
27
(b) 5 8 21 25 28 30 33 49
1.0 57
29 43 46 58
0.8 32 62
61
20 23 31 59
0.6 6 38
4 37
3 17 36 39 54
9 44 53
0.4 1 45 48 52 56 60
42 51 55
7
0.2 18 22 24 26
19
2
0.0
35 45 55 65 75 85
Migration time (s)
FIGURE 3.10 Sample electropherogram collected from the striatum of an anesthetized rat illustrating peak
capacity of the method. (a) Full scale plot with amino acids identified by migration time matching marked.
(b) Expanded scale for electropherogram collected from control samples (derivatization reagents sampling
artificial cerebral spinal fluid in vitro, lower trace) and in vivo samples (upper trace) at 10-fold higher gain
illustrating smaller peaks detected. Fluorescence units are equivalent in the plots. Peaks that were consistently
observed in vivo and not observed in control samples were counted. Numbers for off-scale peaks are not shown.
Peak 55 is dopamine. (From Shou, M., et al., Anal. Chem., 78, 6717, 2006. With permission.)
require good resolution, in particular for the analysis of plant nutrients; however, for the analysis
of contaminants or agrochemical residues, high sensitivity is essential in combination with sample
pretreatments.
The catechin content of tea is often analyzed by CE techniques. Matcha is a special powdered
green tea used in the Japanese tea ceremony and catechins contained in it have been quantified by
MEKC [63]. MEKC conditions are conventional; a 50 µm ID bubble capillary of 77 cm effective
length, 25 mM phosphate buffer (pH 7.0) containing 20 mM SDS, applied voltage of 27 kV, and
detection at 200 nm. Water and methanol extracts of matcha were compared with water extracts of a
popular green tea. An electropherogram of matcha extract is shown in Figure 3.11. The concentration
of epigallocatechin gallate (EGCG) available from drinking matcha is 137-times greater than the
amount of EGCG available from China Green Tips green tea, and at least three times higher than
the largest value reported in the literature for other green teas.
The residue analysis of pesticides or herbicides in food is an important social concern. The
contents of these agrochemicals are very low and the sample preparations are time consuming and
labor intensive. Therefore, development of simple, inexpensive, and efficient total analysis system
is urgently needed. The analysis of pesticides in fruits and vegetables was described by Juan-García
140 1
120
100
80
mAU
60
40
5 7
3 4
20 6
4 5 6 7 8 9
Time (min)
FIGURE 3.11 Electropherogram of a matcha tea sample prepared in deionized water using the traditional
Japanese method. Conditions: capillary, 80.5 cm (77.0 cm effective length), 50 µm ID with bubble (three times
extended path length); running solution, 25 mM phosphate buffer (pH 7.0) containing 20 mM SDS; applied
voltage, 27 kV; detection, at 200 nm. Peak identification: (1) caffeine, (2) catechin, (3) epigallocatechin, (4)
epigallocatechin gallate (EGCG), (5) epicatechin, (6) epicatechin gallate, and (7) internal standard (4-amino-
2-hydroxybenzoic acid). (From Weiss, D.J. and Anderton, C.R., J. Chromatogr. A, 1011, 173, 2003. With
permission.)
et al. [64] using SPE and stir-bar sorptive extraction (SBSE), in combination with MEKC. The
recoveries obtained by SPE ranged from 40% to 106% with RSD from 10% to 19%, whereas the
SBSE method was associated with recoveries in the 12–47% range with RSDs of 3–17%. It is
noteworthy that the limits of quantification were much better by SPE (0.2–0.5 mg/kg depending on
the processed sample amount) than those obtained by SBSE (1 mg/kg for each compound). MEKC
conditions using 75 mM sodium cholate in 6 mM tetraborate buffer (pH 9.2) was employed and
exemplary electropherograms are shown in Figure 3.12, where pesticide residues in lettuce were
analyzed.
3.4.4 ENVIRONMENTAL ANALYSIS

Most target analytes in environmental analysis are low in concentration, although the amount of
sample is not very limited. As a result, CE is not an ideal technique for environmental analysis.
A combination of a sample pretreatment such as SPE or liquid–liquid extraction, and online sample
preconcentration is clearly necessary here. There are several review articles published on the topic
[65–67] and a detailed discussion of this topic can be found in Chapter 31 by Tonin and Tavares.
Determination of pesticides in water has been performed by automatic online SPE in combination
with MEKC [68]. A C18 solid-phase minicolumn was used for the preconcentration, allowing a 12-
fold enrichment (as an average value) of the pesticides from fortified water samples. MEKC was
performed with 10 mM phosphate buffer (pH 9.5) containing 60 mM SDS and 8% acetonitrile.
Pesticides mixtures spiked in water were detected down to a concentration of 50 µg/L in less than
13 min as shown in Figure 3.13.
0.002 8
0.000 9
1 7
–0.002
4
2 3 5
nm
–0.004 (b) 6
–0.006
–0.008
(a) 9
–0.010
10.0 12.5 15.0 17.5
Minutes
FIGURE 3.12 Electropherograms of SPE extracts from 15 g sample of (a) lettuce that contains pyriproxyfen
at 0.2 mg/kg sample and (b) lettuce sample spiked with the pesticides at 0.5 mg/kg levels. Peak identification:
(1) flutriafol, (2) cyproconazole I, (3) cyproconazole II, (4) myclobutanil, (5) tebuconazole, (6) acrinathrin, (7)
bitertanol, (8) fludioxonil, and (9) pyriproxyfen. MEKC conditions: capillary, 75 µm ID × 57 cm (50 cm to the
detector); running solution, 6 mM sodium tetraborate (pH 9.2) containing 75 mM sodium cholate; detection,
214 nm. (From Juan-García, A., et al., J. Chromatogr. A, 1073, 229, 2005. With permission.)
0.004
1 3
2 5 6
4
7
0.002 8
AU
0.000
2 4 6 8 10 12
Time (min)
FIGURE 3.13 Electropherograms of an unspiked and spiked river water sample (0.25 µg/mL). (1) EOF,
(2) fenuron, (3) simazine, (4) atrazine, (5) carbaryl, (6) ametryn, (7) prometryn, and (8) terbutryn. MEKC
conditions: capillary, 75 µm ID × 47 cm; running solution, 10 mM phosphate buffer (pH 9.5) containing
60 mM SDS and 8% acetonitrile; detection, 226 nm. (From Hinsmann, P., et al., J. Chromatogr. A, 866, 137,
3.5 METHODS DEVELOPMENT GUIDELINES

Introductory conditions for initial MEKC separations of neutral or weakly ionizable analytes are
provided in Table 3.1 [4]. The choice of the running solution is rather arbitrary, but a weakly alkaline
borate buffer is recommended because the electrophoretic mobility of the borate ion is rather low
and, thus, the current can be kept low. The sample solution can be prepared in any solvents provided
TABLE 3.1
Suggested Standard Operating Conditions
Capillary 50–75 µm ID × 20–50 cm (to the detector) uncoated fused silica capillary
Running solution 50 mM SDS in 50 mM borate buffer (pH 8.5–9.0) or 50 mM phosphate buffer (pH 7.0) (use
sodium salts)
Applied voltage 10–25 kV (keep current below 50 µA)
Temperature 25◦ C or ambient
Sample solvent Water or methanol (other water miscible organic solvents are usable)
Sample concentration 0.1–1 mg/mL (lower concentrations are acceptable if detectable)
Injection method Hydrodynamic or siphoning at anodic end
Injection volume Less than 1% of the capillary length
Detection Absorbance at 200–210 nm (higher wavelengths are possible if detectable)
Source: Modified from Table 3 of Terabe, S., Micellar Electrokinetic Chromatography, Beckman, Fullerton, 1992.
they are miscible with water. However, water-rich solution is advisable if the analyte is water
soluble. When the sample contains high concentrations of an organic solvent, each peak may split
due to incomplete mixing of the sample solution with the running solution. When online sample
preconcentration is to be utilized, the sample solution must be optimized for each concentration
technique. For a preliminary run, a relatively high concentration of the analyte will be favorable
for easy detection. The separation time depends on the EOF velocity and effective capillary length
(length to the detector) but under the conditions listed in Table 3.1 the migration time of the micelle
may be less than 20 min.
When a preliminary run shows insufficient separation, several parameters may need to be adjusted
in order to obtain optimized separation [9]. First, estimate the retention factors of analytes by
Equation 3.1 using measured t0 and tmc . Should measuring the tmc prove to be nontrivial due to
the difficulty observing the micelle marker peak, simply assume tmc is four times longer than t0 .
If k values are between 0.5 and 10, k values should be optimized—alternatively, adjust to about 2
by changing the micelle (or surfactant) concentration. If k values are too high (k > 10), several
options are available; decrease the micelle concentration by adding an organic solvent (methanol or
acetonitrile), add a CD, or use another surfactant (e.g., a bile salt). If k values are too small (k < 0.5),
limited options are available; increase surfactant concentration if the analyte is neutral, add an ion-
pairing reagent (a tetraalkylammonium salt for anionic analytes and an alkanesulfonate for cationic
analytes) or use cationic surfactant such as CTAB instead of SDS if the analytes are anionic.
If optimum k values do not provide acceptable resolution, further tuning may be required; the use
of additives such as organic solvents or CDs may be needed, or the use of alternate surfactants may
prove effective. The selection of additives is extensive, but it is recommended that addition of a low
concentration of an organic solvent be trialed initially. If the unresolved analytes have closely related
structures, addition of a CD derivative may prove effective. There are many different CD derivatives
available, but simply β-CD or γ-CD can be tested initially. Further tuning can be performed by
selecting other CD derivatives (e.g., sulfated) if necessary. Another choice is the modification of the
micelle by using mixed micelles, in particular, addition of a nonionic surfactant such as Brij-35 or
Tween 60, adding cosurfactants such as 1-propanol or 1-butanol, or by adding an organic counter
ion such as tetraalkylammonium salts for SDS.
The above-mentioned guideline describes practical procedures based on the migration behavior
[9]. Computer-assisted modeling, predictions, and multifactor optimization strategy are proposed
based on physicochemical models describing the migration behavior for ionizable analytes [69–71].
Other than the factors mentioned above, several experimental parameters that include temperature,
applied voltage, buffer concentration, pH, and so forth affect resolution and many parameters can
be altered in a manner that provides a fine tuning via statistical approaches, using fitting procedures
of polynomial equations (chemometrics). There are several techniques reported to optimize MEKC
separation conditions. It is beyond the scope of this chapter to describe chemometric techniques in
detail, and only some references are mentioned [72–76]. Although the approaches based on statistical
designs are helpful to find the optimum values for many parameters, they should not be used as a
black box system [77]. Analysts need to play a crucial role in optimizing separation conditions by
carefully considering separation models.

Most fundamental characteristics on MEKC were investigated by mid-1990s and, hence, this chapter
is not much different from that in second edition of the book [78] in fundamental theory and major
procedures. New parts of the chapter are mainly MS detection, online sample preconcentration, and
new applications. Unlike a decade ago, MEKC is now a mature technique in CE arsenal, and while
a few new, innovative MEKC-related techniques may not arise, applications of MEKC is likely
to be widely expanded—this is certainly true in the bioanalytical fields and in the ever-expanding
microscale separations including those in microfabricated devices. Although MEKC is considered to
be suitable for the separation of small molecules, it is also useful for the separation of large molecules
such as peptides and proteins, and applications to large molecules will certainly be explored more
extensively in the future. The latest progress in MEKC has been recently reviewed [79].
One of the issues that still remains unsolved in MEKC is the improvement in reproducibility
in quantitative analysis, including migration time and peak height or peak area—a problem that is
not characteristic of MEKC only, but general to every CE technique. To verify the validity of the
MEKC method, intercompany cross-validation exercises have been performed and the results have
shown acceptable method performance in terms of relative migration time precision [80]. While
reproducible EOF is an issue, it is so in every CE separation modes (except for gel electrophoresis);
however, MEKC has more parameters to consider to maintain acceptable reproducibility. First, some
characteristic issues in MEKC are discussed to improve quantitative results. CMC and distribution
coefficient are both temperature sensitive: an increase in temperature causes an increase in CMC
and, hence, a decrease in the micellar concentration, which reduces the retention factor. On the other
hand, an increase in temperature causes a decrease in distribution coefficient and, hence, the retention
factor. Therefore, the temperature increase will significantly decrease the retention factor from the
viewpoint of CMC and distribution coefficient. It is well known that an increase in temperature
reduces viscosity of the running solution in any CE mode, which increases both electrophoretic and
EOF velocities, reducing the migration time. Thus, temperature and migration time have an inverse
relationship where a decrease in migration time accompanies an increase in temperature, and the
effect will be larger in MEKC than in CZE. Consequently, temperature must be controlled with
diligence in MEKC, relative to other CE modes, in order to obtain high reproducibility. To maintain
a low current is also a good strategy, particularly since MEKC usually produces a higher current than
CZE due to the addition of the surfactant. A compromise is, therefore, necessary between high Joule
heating and fast separation time. It should be noted that an added bonus with the use of micelles
is that they may prevent analyte adsorption on the capillary wall, which may contribute to good
reproducibility.
Minimizing the temperature effects discussed above could be obtained with the use of polymer
micelles or polymer surfactants [81–83], whose CMC is zero, and even in nonaqueous solvent, the
micelle is stable. Although several polymer surfactants are commercially available, no such sur-
factant is widely accepted, probably because SDS, CTAB, or CTAC, and bile salts are superior to
polymer surfactants as the pseudostationary phase in MEKC. Although microemulsion electroki-
netic chromatography (MEEKC) is not discussed in this chapter but covered in Chapter 4 by Altria
and colleagues, a similar optimization strategy to that in MEKC applies to MEEKC [84–86]. Since
microemulsion (oil-in-water or o/w) consists of four components, surfactant, oil (organic solvent),
cosurfactant, and buffer, available parameters for optimizing selectivity are more extensive than
that in MEKC. For most analytes, the choice of the surfactant is most important because the sur-
face of the microemulsion affects significantly selectivity as in the case of micelles. For nonpolar
hydrophobic analytes, the choice of the organic solvent is effective but conditions that confer stabil-
ity to the microemulsion limit the range of organic solvent composition. An advantage of MEEKC
over MEKC is the ability to vary the migration time window by changing the concentration of SDS
in microemulsion [84]. It should be noted that the composition of microemulsion tends to change
with evaporation of the organic solvent, particularly noteworthy if a volatile organic solvent such as
hexane is used as the core oil.
Finally, the author would like to emphasize that, although MEKC gives similar separation selec-
tivity to that observed with reversed-phase HPLC, the choice of selectivity manipulations are more
versatile and separation efficiency is much higher than that of HPLC. Detection sensitivity, in terms
of concentration, has been considered inferior to that of HPLC, but sensitivity is not an issue if high
sensitivity detectors are implemented or if online sample preconcentration is employed. One more
important feature of MEKC, as a mode of CE, is the ability to use small amounts of sample and
almost no need of organic solvents. Therefore, the basis for the apparent lower popularity of MEKC
in industrial routine analysis is, at least partially, because CE techniques are not widely accepted
among those who are engaged in industrial analysis based largely on the popular HPLC platform.
However, it is strongly encouraged that the instrument companies explore the development new CE
instruments that are more user-friendly and easy to operate.
REFERENCES
1. Terabe, S. et al., Electrokinetic separations with micellar solutions and open-tubular capillaries, Anal.
Chem., 56, 111, 1984.
2. Terabe, S., Electrokinetic chromatography: an interface between electrophoresis and chromatography,
Trends Anal. Chem., 8, 129, 1989.
3. Zare, R.N., Tribute to Professor S. Terabe, J. Chromatogr. A, 1106, 4, 2006.
4. Terabe, S., Micellar Electrokinetic Chromatography, Beckman, Fullerton, 1992.
5. Vindevogel, J. and Sandra, P., Introduction to Micellar Electrokinetic Chromatography, Hüthig,
Heidelberg, 1992.
6. Pyell, U., Ed., Electrokinetic Chromatography, John Wiley, Chichester, 2006.
7. Terabe, S., Chen, N., and Otsuka, K., Micellar electrokinetic chromatography, Adv. Electrophoresis,
7, 87, 1994.
8. Khaledi, M.G., Micellar electrokinetic chromatography, in High Performance Capillary Electrophore-
sis, Khaledi, M.G., Ed., Wiley-Interscience, New York, 1998, Chapter 3.
9. Terabe, S., Selectivity manipulation on micellar electrokinetic chromatography, J. Pharm. Biomed.
Anal., 10, 705, 1992.
10. Otsuka, K. and Terabe, S., Micellar electrokinetic chromatography, Bull. Chem. Soc. Jpn., 71, 2465,
1989.
11. Terabe, S., Micellar electrokinetic chromatography, Anal. Chem., 76, 240A, 2004.
12. Katayama, H., Ishihama, Y., and Asakawa, N., Stable capillary coating with successive multiple ionic
polymer layers, Anal. Chem., 70, 2254, 1998.
13. Nelson, W.M. et al., On-line partial filling micellar electrokinetic chromatography-electrospray
ionization mass spectrometry, J. Chromatogr. A, 749, 219, 1996.
14. Koezuka, K. et al., Separation and detection of closely related peptides by micellar electrokinetic chro-
matography coupled with electrospray ionization mass spectrometry using the partial filling technique,
J. Chromatogr. B, 689, 3, 1997.
15. Muijselaar, P.G., Otsuka, K., and Terabe, S., On-line coupling of partial-filling micellar electrokinetic
chromatography with mass spectrometry, J. Chromatogr. A, 802, 3, 1998.
16. Suomi, J. et al., Determination of iridoid glycosides by micellar electrokinetic capillary

chromatography-mass spectrometry with use of the partial filling technique, Electrophoresis, 22,
2580, 2001.
17. Stubberud, K. et al., Partial filling micellar electrokinetic chromatography optimization studies of
ibuprofen, codeine and degradation products, and coupling to mass spectrometry, Electrophoresis, 23,
572, 2002.
18. Somsen, G.W., Mol, R., and de Jong, G.J., On-line micellar electrokinetic chromatography–mass
spectrometry: feasibility of direct introduction of non-volatile buffer and surfactant into the electrospray
interface, J. Chromatogr. A, 1000, 953, 2003.
19. Mol, R. et al., Micellar electrokinetic chromatography–electrospray ionization mass spectrometry for
the identification of drug impurities, J. Chromatogr. B, 843, 283, 2006.
20. Mol, R., de Jong, G.J., and Somsen, G.W., Atmospheric pressure photoionization for enhanced compat-
ibility in on-line micellar electrokinetic chromatography-mass spectrometry, Anal. Chem., 77, 5277,
2005.
21. Nishi, N., Tsumagari, N., and Terabe, S., Effect of tetraalkylammonium salts on micellar electrokinetic
chromatography of ionic substances, Anal. Chem., 61, 2434, 1989.
22. Terabe, S. et al., Effect of urea addition in micellar electrokinetic chromatography, J. Chromatogr.,
545, 359, 1991.
23. Kaneta, T. et al., Effect of addition of glucose on micellar electrokinetic capillary chromatography
with sodium dodecyl sulphate, J. Chromatogr., 609, 369, 1992.
24. Terabe, S., Otsuka, K., and Ando, T., Electrokinetic chromatography with micellar solution and
open-tubular capillary, Anal. Chem., 57, 834, 1985.
25. Foley, J. P., Optimization of micellar electrokinetic chromatography, Anal. Chem., 62, 1302, 1990.
26. Zhang, C.-X., Sun, Z.-P., and Ling, D.-K., Micellar electrokinetic capillary chromatography theory
based on conventional chromatography, J. Chromatogr. A, 655, 309, 1993.
27. Yang, S. and Khaledi, M.G., Chemical selectivity in micellar electrokinetic chromatography: char-
acterization of solute-micelle interactions for classification of surfactants, Anal. Chem., 67, 499,
1995.
28. Chen, N. et al., Effect of physico-chemical properties and molecular structure on the micelle-water
partition coefficient in micellar electrokinetic chromatography, J. Chromatogr. A, 678, 327, 1994.
29. Poole, C.F., Poole, S.K., and Abraham, M.H., Recommendations for the determination of selectivity
in micellar electrokinetic chromatography, J. Chromatogr. A, 798, 207, 1998.
30. Muijselaar, P.G., Claessens, H.A., and Cramers, C.A., Characterization of pseudostationary phases in
micellar electrokinetic chromatography by applying linear solvation energy relationships and retention
indexes, Anal. Chem., 69, 1184, 1997.
31. Chien, R.L. and Burgi, D.S., Optimization in sample stacking for high-performance capillary
32. Quirino, J.P. and Terabe, S., On-line concentration of neutral analytes for micellar electrokinetic
chromatography. 3. Stacking with reverse migrating micelles, Anal. Chem., 70,149, 1998.
33. Quirino, J.P. and Terabe, S., On-line concentration of neutral analytes for micellar electrokinetic
chromatography. 5. Field-enhanced sample injection with reverse migrating micelles, Anal. Chem.,
70, 1893, 1998.
34. Quirino, J.P. and Terabe, S., Exceeding 5000-fold concentration of dilute analytes in micellar
electrokinetic chromatography, Science, 282, 465, 1998.
35. Quirino, J.P. and Terabe, S., Sweeping of analyte zone in electrokinetic chromatography, Anal. Chem.,
71, 1638, 1999.
36. Palmer, J., Munro, N.J., and Landers, J.P., A universal concept for stacking neutral analytes in micellar
37. Sera, Y. et al., Sweeping on a microchip: concentration profiles of the focused zone in micellar
electrokinetic chromatography, Electrophoresis, 22, 3509, 2001.
38. Kim, J.-B. et al., On-line sample concentration in MEKC using cationic surfactants, J. Chromatogr.
A, 916, 123, 2001.
39. Isoo, K. and Terabe, S., Analysis of metal ions by sweeping via dynamic complexation and cation-
selective exhaustive injection in capillary electrophoresis, Anal. Chem., 75, 6789, 2003.
40. Quirino, J.P. and Terabe, S., Sweeping of neutral analytes via complexation with borate in capillary
zone electrophoresis, Chromatographia, 53, 285, 2001.
41. Palmer, J., Burgi, D.S., and Landers, J.P., Electrokinetic stacking injection of neutral analytes under
continuous conductivity conditions, Anal. Chem., 74, 632, 2002.
42. Quirino, J.P. and Terabe, S., Approaching a million-fold sensitivity increase in capillary electrophoresis
with direct ultraviolet detection: cation-selective exhaustive injection and sweeping, Anal. Chem., 72,
1023, 2000.
43. Kim, J.-B., Otsuka, K., and Terabe, S., Anion selective exhaustive injection-sweep–micellar
electrokinetic chromatography, J. Chromatogr. A, 932, 129, 2001.
44. Britz-McKibbin, P. and Chen, D.D.Y., Selective focusing of catecholamines and weakly acidic
compounds by capillary electrophoresis using a dynamic pH junction, Anal. Chem., 72, 1242, 2000.
45. Britz-McKibbin, P., Otsuka, K., and Terabe, S., On-line focusing of flavin derivatives using dynamic
pH junction-sweeping capillary electrophoresis with laser-induced fluorescence detection, Anal.
Chem., 74, 3736, 2000.
46. Molina, M. and Silva, M., Micellar electrokinetic chromatography: current developments and future,
47. Pappas, T.J., Gayton-Ely, M., and Holland, L.A., Recent advances in micellar electrokinetic
chromatography, Electrophoresis, 26, 719, 2005.
48. Huie, C.W., Recent applications of microemulsion electrokinetic chromatography, Electrophoresis,
27,60, 2006.
49. Nishi, H. and Terabe, S., Micellar electrokinetic chromatography: perspectives in drug analysis,
J. Chromatogr. A, 735, 3, 1996.
50. Wätzig, H., Degenhardt, M., and Kunkel, A., Strategies for capillary electrophoresis: method devel-
opment and validation for pharmaceutical and biological applications, Electrophoresis, 19, 2695,
1998.
51. Garcia-Ruiz, C. and Marina, M.L., Recent advances in the analysis of antibiotics by capillary
52. Sung, W.-C. and Chen, S.-H., Pharmacokinetic applications of capillary electrophoresis: a review on
recent progress, Electrophoresis, 27, 257, 2006.
53. Ha, P.T.T., Hoogmartens, J., and Van Schepdael, A., Recent advances in pharmaceutical applications
of chiral capillary electrophoresis, J. Pharm. Biomed. Anal., 41, 1, 2006.
54. Orlandini, S. et al., Micellar electrokinetic chromatography for the simultaneous determination of
ketorolac tromethamine and its impurities multivariate optimization and validation, J. Chromatogr. A,
1032, 253, 2004.
55. Iadarola, P. et al., Micellar electrokinetic chromatographic and capillary zone electrophoretic meth-
ods for screening urinary biomarkers of human disorders: a critical review of the state-of-the-art,
56. Nakagawa, T. et al., Electrokinetic chromatography for drug analysis. Separation and determination
of cefpiramide in human plasma, Chem. Pharm. Bull., 37, 707, 1989.
57. Penalvo, G.C. et al., Evaluation of capillary zone electrophoresis and micellar electrokinetic capillary
chromatography with direct injection of plasma for the determination of cefotaxime and its metabolite,
Anal. Chem., 69, 1364, 1997.
58. Huang, C.-W. et al., Sweeping technique combined with micellar electrokinetic chromatography for
the simultaneous determination of flunitrazepam and its major metabolites, J. Chromatogr. A, 1110,
240, 2006.
59. Paxon, T.L. et al., Microcolumn separation of amine metabolites in the fruit fly, Anal. Chem., 77, 5349,
2005.
60. Shou, M. et al., Monitoring dopamine in vivo by microdialysis sampling and on-line CE-laser-induced
fluorescence, Anal. Chem., 78, 6717, 2006.
61. Frazier, R.A. and Papadopoulou, A., Recent advances in the application of capillary electrophoresis
for food analysis, Electrophoresis, 24, 4095, 2003.
62. Cifuentes, A., Recent advances in the application of capillary electromigration methods for food
analysis, Electrophoresis, 27, 283, 2006.
63. Weiss, D.J. and Anderton, C.R., Determination of catechins in matcha green tea by micellar
electrokinetic chromatography, J. Chromatogr. A, 1011, 173, 2003.
64. Juan-García, A., Picó, Y., and Font, G., Capillary electrophoresis for analyzing pesticides in fruits and
vegetables using solid-phase extraction and stir-bar sorptive extraction, J. Chromatogr. A, 1073, 229,
2005.
65. Dabek-Zlotorzynska, E., Capillary electrophoresis in the determination of pollutants, Electrophoresis,
18, 2453, 1997.
66. Menzinger, F. et al., Analysis of agrochemicals by capillary electrophoresis, J. Chromatogr. A, 891,
45, 2000.
67. Dabek-Zlotorzynska, E. and Celo, V., Recent advances in capillary electrophoresis and capillary
electrochromatography of pollutants, Electrophoresis, 27, 304, 2006.
68. Hinsmann, P. et al., Determination of pesticides in waters by automatic on-line solid-phase extraction–
69. Khaledi, M.G., Smith, S.C., and Strasters, J.K., Micellar electrokinetic capillary chromatography of
acidic solutes: migration behavior and optimization strategies, Anal. Chem., 63, 1820, 1991.
70. Strasters, J.K. and Khaledi, M.G., Migration behavior of cationic solutes in micellar electrokinetic
capillary chromatography, Anal. Chem., 63, 2503, 1991.
71. Quang, C., Wasters, J.K., and Khaledi, M.G., Computer-assisted modeling, prediction, and multifactor
optimization in micellar electrokinetic chromatography of ionizable compounds, Anal. Chem., 66,
1646, 1994.
72. Vindevogel, J. and Sandra, P., Resolution optimization in micellar electrokinetic chromatography: use
of Plackett-Burman statistical design for the analysis of testosterone esters, Anal. Chem., 63, 1530,
1991.
73. He, Y. and Lee, H.K., Orthogonal array design experiments for optimizing the separation of various
pesticides by cyclodextrin-modified micellar electrokinetic chromatography, J. Chromatogr. A, 793,
331, 1998.
74. Persson-Stubberud, K. and Åström, O., Separation of ibuprofen, codeine phosphate, their degradation
products and impurities by capillary electrophoresis I. Method development and optimization with
fractional factorial design, J. Chromatogr. A, 798, 307, 1998.
75. Mikaeli, S., Thorsén, G., and Karlberg, B., Optimisation of resolution in micellar electrokinetic
chromatography by multivariate evaluation of electrolytes, J. Chromatogr. A, 907, 267, 2001.
76. Wan, H., Öhman, M., and Blomberg, R.G., Chemometric modeling of neurotransmitter amino acid
separation in normal and reversed migration micellar electrokinetic chromatography, J. Chromatogr. A,
916, 255, 2001.
77. Corstjens, H. et al., Optimisation of selectivity in capillary electrophoresis with emphasis on micellar
electrokinetic capillary chromatography, J. Chromatogr. A, 715, 1, 1995.
78. Mazzeo, J.R., Micellar electrokinetic chromatography, in Handbook of Capillary Electrophoresis, 2nd
ed., Landers, J.P., Ed., CRC Press, Boca Raton, FL, 1996, Chapter 2.
79. Silva, M., MEKC: An update focusing on practical aspects, Electrophoresis, 28, 174, 2007.
80. Altria, K.D. et al., An inter-company cross-validation exercise on capillary electrophoresis testing of
dose uniformity of paracetamol content in formulations, Chromatographia, 39, 180, 1994.
81. Palmer, C.P. and Terabe, S., Micelle polymers as pseudostationary phases in MEKC: chromatographic
performance and chemical selectivity, Anal. Chem., 69, 1852, 1997.
82. Palmer, C.P., Recent progress in the development, characterization and application of polymeric
pseudophases for electrokinetic chromatography, Electrophoresis, 23, 3993, 2002.
83. Palmer, C.P. and McCarney, J.P., Recent progress in the use of soluble ionic polymers as
pseudostationary phases for electrokinetic chromatography, Electrophoresis, 25, 4086, 2004.
84. Terabe, S. et al., Microemulsion electrokinetic chromatography: comparison with micellar electroki-
netic chromatography, J. Chromatogr., 608, 23, 1992.
85. Altria, K.D., Highly efficient and selective separations of a wide range of analytes obtained by an
optimised microemulsion electrokinetic chromatography method, Chromatographia, 49, 457, 1999.
86. Altria, K.D., Clark, B.J., and Mahuzier, P.-E., The effect of operating variables in microemulsion
electrokinetic capillary chromatography, Chromatographia, 52, 758, 2000.
4 Capillary Electrophoresis for
Pharmaceutical Analysis
Eamon McEvoy, Alex Marsh, Kevin Altria, Sheila Donegan,
and Joe Power
CONTENTS
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

4.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
4.3 Theoretical Aspects (Electrophoretic Modes) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
4.3.1 Free Solution Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
4.3.2 Nonaqueous Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
4.3.3 Micellar Electrokinetic Chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
4.3.4 Microemulsion Electrokinetic Chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
4.3.5 Capillary Electrophoresis-Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
4.3.6 Multiplexed Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
4.4 Methods of Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
4.4.1 UV/Vis Absorbance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
4.4.2 Laser-Induced Fluorescence Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
4.4.3 Mass Spectrometric Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
4.4.4 Electrochemical Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
4.4.5 Selecting the Most Suitable Detector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
4.5 Selecting the Most Suitable Mode of CE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
4.6 Pharmaceutical Applications of Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
4.6.1 Chiral Pharmaceutical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
4.6.1.1 Cyclodextrins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
4.6.1.2 Crown Ethers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
4.6.1.3 Macrocyclic Antibiotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
4.6.1.4 Oligo- and Polysaccharides. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
4.6.1.5 Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
4.6.1.6 Chiral MEKC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
4.6.1.7 Chiral MEEKC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
4.6.1.8 Pharmaceutical Applications of Chiral CE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
4.6.1.9 Examples of Chiral CE and Validated Methods . . . . . . . . . . . . . . . . . . . . . . . . . 151
4.6.2 Pharmaceutical Assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
4.6.3 Impurity Profiling of Pharmaceuticals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
4.6.4 Physicochemical Profiling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
4.6.4.1 pKa Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
4.6.4.2 Log Pow Measurements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
4.6.5 Analysis of Small Molecules and Ions (Pharmaceutical) . . . . . . . . . . . . . . . . . . . . . . . . . . 161
4.7 Method Development and Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
4.7.1 Method Optimization Using Experimental Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
4.7.2 Method Validation Guidelines for Pharmaceutical Applications . . . . . . . . . . . . . . . . . . 166
135
4.8 Comparison of HPLC, GC, and CE for Pharmaceutical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 167

4.8.1 Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
4.8.2 Sample Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
4.8.3 Sample Volume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
4.8.4 Sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
4.8.5 Precision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
4.8.6 Analysis Times . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
4.8.7 Reagents and Consumables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
4.9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
4.10 Dedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
4.1 INTRODUCTION
The use of capillary electrophoresis (CE) as an analytical technique has become more widespread
and popular in recent years and has established itself as a method of choice for many applications.
The ability to analyze small ions and organic molecules has made it the mainstay in many industries
such as the food and beverage, biotechnology, and pharmaceutical industries, many of which have
incorporated CE into their research as well as their quality assurance (QA) and quality control (QC)
departments.
Pharmaceutical analysis is dominated by high-performance liquid chromatography (HPLC) with
thin layer chromatography (TLC) and gas chromatography (GC) also used to a lesser extent, as
their quantitative abilities are not widespread as HPLC. CE, however, is becoming a more widely
used technique and is recognized by several regulatory authorities as a reliable routine analytical
technique. Although CE was initially heralded for its high speed and low sample volume, resolving
power, and versatility, the technique is also valuable because it is quantitative, can be automated,
and can separate compounds that have been traditionally difficult to handle by HPLC. Qualitative
analysis is also possible by capillary electrophoresis-mass spectrometry (CE-MS).
The small sample volumes required by CE can be an advantage over HPLC if sample supply is
limited, for example, in drug discovery, however, with short path lengths using traditional ultraviolet
(UV) detection methods, the technique can suffer from low sensitivity. A number of solutions to
this problem have been introduced including “Bubble Cell” by Agilent, which increases the path
length at the point of detection. A “High Sensitivity Detection Cell” also by Agilent improves
detection sensitivity 10-fold over normal capillaries. Other detection methods such as “Laser-Induced
Fluorescence Detection” (LIFD) and CE-MS also give better sensitivity and have led to CE being
on a par with HPLC for many pharmaceutical applications and indeed CE is superior to HPLC for
many separations.
The advantages of CE for pharmaceutical analysis include its speed and cost of analysis, reduc-
tions in solvent consumption, and disposal and the possibility of rapid method development. CE also
offers the possibility of using a single set of operating conditions for a wide range of separations
that can improve laboratory efficiency. CE instruments can be coupled to a variety of detector types
including mass spectrometers, indirect UV detectors, LIFD’s, and low UV wavelength detectors.
Separation efficiencies achieved by CE can be an order of magnitude better than competing
HPLC methods, which highlights the potential resolving power of CE for complex and difficult
separations, which is a valuable advantage in pharmaceutical analysis. The use of an open tubular
capillary improves resolution relative to that of packed HPLC columns by eliminating the multiple
path term (A) in the following Van Deemter equation. CE reduces plate height further by also
eliminating the mass transfer term (Cux ) that comes from the finite time needed for solutes to
equilibrate between the mobile and stationary phases in liquid chromatography (LC). This stationary
phase is absent in CE [1] [with the exception of micellar electrokinetic chromatography (MEKC) and
Capillary Electrophoresis for Pharmaceutical Analysis 137
microemulsion electrokinetic chromatography (MEEKC) where a “pseudo-stationary phase” forms

part of the separation mechanism, these CE modes will be discussed later in the chapter].
B C
van Deemter equation for plate height H =A+ + ,
ux ux
where A is the multiple path term which is zero for CE capillaries, B/ux is the longitudinal diffusion
term and is the only contributor to band broadening in CE, and C/ux is the equilibration time (mass
transfer) term also zero for CE.
High-speed separations are possible with CE, sometimes as quick as a matter of seconds. Because
of this it is an ideal technique for “high-throughput” analyses such as forensic testing, DNA sequenc-
ing, and, in particular, screening of candidate drug compounds. Multiplexed CE instruments can be
used to perform multiple parallel separations, dramatically increasing sample throughput. For exam-
ple, the “CombiSep cePRO 9600™” utilizes multiple capillaries and can perform 96 separations in
parallel.
An important aspect of CE for pharmaceutical analysis is its ability to separate all molecules of
pharmaceutical interest such as small and large synthetic and natural drugs, proteins and peptides,
oligonucleotides, carbohydrates and polysaccharides, inorganic ions, and so forth.
This chapter will look at the use of CE for pharmaceutical analysis and will include descrip-
tions of the various modes of CE and their suitability for quantitative and qualitative analysis of
pharmaceutical compounds. Practical applications of CE for the analysis of pharmaceuticals will be
covered, these applications include drug assay, impurity determination, physicochemical measure-
ments, chiral separations, and the analysis of small molecules. A section covering the approach to
CE method development for pharmaceutical analysis will include guidelines to selecting the best
mode of CE for an intended separation. Extensive data will be provided on successful pharmaceu-
tical separations with references to extra source material for the interested reader. This chapter will
provide a comprehensive and up to date view of the role and importance of CE for the analysis of
pharmaceuticals and will provide the reader with practical information and real data that will help
them to decide if CE is suitable for an intended separation.
4.2 BACKGROUND
In the 1930s, Tiselius developed the first electrophoretic apparatus to perform separations based on
differential migration rates in an electrical field. Up until the early 1980s, most of the research carried
out using this technique was by biochemists and molecular biologists on the separation, isolation,
and analysis of proteins and other biological macromolecules. Since its introduction in the 1980s
[2–5] and with the appearance of commercial instruments for performing analytical electrophoresis
on a microscale in capillary columns, CE has become established as a fast, reliable, and highly
efficient alternative and complementary method to LC. While CE can perform separations that may
be difficult by traditional LC methods and provides different migration patterns, combined use of
these two techniques can be a powerful analytical tool for the pharmaceutical analyst. Over the
past number of years since the early 1990s, there have been many review papers [6–12] and books
[13,14] published concerning CE for pharmaceutical analysis. Since the first applications of CE to
pharmaceutical analysis by Altria and Simpson [15] in 1987, the number of reports of CE for a range
of pharmaceutical applications now is in hundreds and the number of pharmaceutical compounds
analyzed by CE is far greater [13]. There have also been a number of specialized reviews of CE for
pharmaceutical analysis including chiral CE [16–20], CE with electrochemical detection [21] and
conductivity detection [22], LIFD [23], CE coupled with electrospray-ionization MS (CE-ESI-MS)
[24], and CE determination of acid dissociation constants [25].
When deciding on the optimum CE conditions for an intended separation, the pharmaceutical
analyst must take a different approach to that of the analyst using HPLC techniques. The vast range
of HPLC methods routinely used for pharmaceutical analysis would lead the liquid chromatographer
to use literature databases and pharmacopoeia methods along with his/her own experience to decide
on initial HPLC conditions. Properties such as sample solubility, log P values, sample polarity,
stationary phase characteristics, and mobile phase compositions, among others, all are important to
the LC analyst when deciding on the initial chromatographic conditions and can be easily obtained
for most applications. The CE analyst, however, must consider the problem separation in a different
manner. The depth of knowledge and literature associated with CE method development is in its
infancy when compared to LC. As CE is an electro-driven separation technique, the approach to
method development must look at factors such as compound pKa data, functional groups, electrolyte
pH, and applied voltage. Compound solubility will also determine which mode of CE is applicable
to the required separation. CE method development guidelines will be given more consideration in
a later section of this chapter.
4.3 THEORETICAL ASPECTS (ELECTROPHORETIC MODES)

There are several electrophoretic modes that can and have been be used to analyze pharmaceuticals.
Each mode offers different possibilities to the CE analyst and the choice of CE mode is the first step
when developing a separation method. The mode of CE that best suits the sample to be analyzed
must be decided. Some questions that the analyst must ask include are all of the components to
be separated soluble in an aqueous electrolyte; are the components charged, neutral, or both; are
physicochemical measurements available for the analytes, that is, pKa and/or log P values; what
are the properties of the sample matrix; and so forth. This section will provide the reader with a
brief introduction to the various modes of CE, which have been utilized for pharmaceutical analysis
and give tips on deciding the mode and detection method for an intended separation/analysis. The
detailed theoretical aspects of each CE mode are too broad to be covered in this chapter so the reader
will be directed to other sources for more in depth theory.
4.3.1 FREE SOLUTION CAPILLARY ELECTROPHORESIS

Free solution CE (FSCE) or capillary zone electrophoresis (CZE), as it is sometimes called, is the
simplest and one of the most widely used forms of CE and involves the use of buffered aqueous
solutions as the carrier electrolytes. FSCE is principally used to separate charged, water-soluble
analytes and separations rely principally on the pH-controlled dissociation of acidic groups or the
protonation of basic functions on the solute. These ionic species are separated based on differences in
their charge-to-mass ratios. For example, basic drugs are separated at low pH as cations while acidic
drugs are separated as anions at high pH. In FSCE all neutral compounds are swept, unresolved,
through the detector together. Under the influence of an applied electric field, sample ions will
move toward their appropriate electrode. Cations migrate toward the cathode and anions toward the
anode. The speed of their movement toward the electrode is governed by their size and number of
appropriate charges. Smaller molecules with a large number of charges will move more quickly than
larger or less charged compounds. The speed of movement, known as the electrophoretic mobility,
is characteristic of each solute. The mobility of a species can also be changed by complexing the ion
as it migrates along the capillary. For example, additives such as cyclodextrins (CDs) can be used
to complex with drug enantiomers to achieve chiral separations. The presence of an electroosmotic
flow (EOF) allows the separation and detection of both cations and anions within a single analysis, as
above pH 7 the EOF is sufficiently strong to sweep anions to the cathode regardless of their charge.
There are a number of variables in FSCE that can be used in the optimization of FSCE methods.
These include the operating pH, electrolyte type and concentration, capillary dimensions, tempera-
ture, and injection volume. Electrolyte additives such as ion-pair reagents and chiral substances can
also be employed in order to manipulate selectivity.
As the majority of pharmaceutical compounds are basic, the highly polar nature of these com-
pounds can make separating them complex by LC methods. Ion-pairing reagents and column
regeneration is often necessary to reduce nonspecific ionic interactions in the LC column. With
CE, these highly functional groups can be exploited to enable separation. These basic drugs and
their impurities can be separated using FSCE in two ways. First, at low pH, the capillary surface is
essentially neutral resulting in a suppression of the EOF in the capillary and the protonated analytes
migrate toward the detector with little or no influence from the EOF. The second and less desirable
method is to use an amine inner capillary surface to repel cationic interactions within the capillary
wall, which allows separation over a wide pH range. Using a low pH phosphate buffer, Hudson and
coworkers [26] reported the use of FSCE to analyze over 500 basic pharmaceutical compounds.
To separate mixtures of acidic and basic drugs (or any cationic and anionic species), FSCE at
high pH can be used [27]. At pH 7 or greater, the EOF generated by the applied current is sufficiently
strong to sweep anions to the detector. Analyte migration time is dependent on solute charge type
and density, strongly cationic species migrate first while small highly charged anions attempt to
migrate against the EOF and are detected last. Neutrals will not be resolved and migrate with the
EOF. Readers who require a more detailed description of the theoretical aspects of FSCE will find a
number of books dealing with the subject [28–31]. A review paper by Smith and Evans [32] outlines
the use of FSCE in pharmaceutical and biomedical analysis.
4.3.2 NONAQUEOUS CAPILLARY ELECTROPHORESIS

Nonaqueous CE (NACE) is used for the separation of water-insoluble or sparingly soluble phar-
maceuticals. NACE employs electrolytes composed of organic solvents and has been successfully
utilized [33–38]. NACE is also useful for the resolution of water-soluble charged solutes as the selec-
tivity obtained can be different to aqueous-based separations. The viscosity and dielectric constants
of organic solvents can have an effect on both sample ion mobility and the level of EOF. Reso-
lution, efficiency, and migration times are critically affected by the nature of the organic solvent,
the electrolyte composition, its concentration, and temperature. Because of their different physi-
cal and chemical properties (viscosity, dielectric constant, polarity, autoprotolysis constant, etc.),
methanol and acetonitrile are the most frequently used solvents for NACE. In particular, methanol–
acetonitrile mixture containing 25 mM ammonium acetate and 1 M acetic acid is considered to be
the appropriate electrolyte solution for the separation of a large variety of basic drugs [39]. The
low currents present in NACE not only allow the use of higher electrolyte salt concentrations and
higher electric field strengths but also the sample load can be scaled-up by employing capillaries
with wider inside diameter. In addition, an effective sample introduction to a mass spectrometer,
in terms of volatility, surface tension, flow rate, and ionization can be expected to further extend
the use of NACE [40]. NACE exploits the vastly different physicochemical properties of organic
solvents to control EOF and analyte migration. The ability of organic solvents to accept protons from
the silanol groups of the capillary wall appears to play an important role in the development of an
EOF. Although an EOF may not be required or may be completely undesired in a few electromi-
gration capillary techniques, it plays a significant role in separations in free solution CE. However,
the purity of the solvents may affect the EOF as they may contain foreign ionic species. The veloc-
ity of EOF is higher in pure solvents than in electrolyte solutions, which may be beneficial when
fast liquid transport is desired. Salt-free solvents may also be advantageous in mass spectrometric
detection [40].
One of the most attractive features of organic solvents is that their physical and chemical prop-
erties differ widely, both from each other and from water. Accordingly, simply changing the organic
solvent or varying the proportions of two solvents can achieve selectivity manipulation in NACE.
pKa values in organic solvents can be significantly different from those in water allowing separations
that are difficult to achieve in aqueous electrolytes. It has been shown that all solvents in which a
measurable EOF is developed in the capillary are either amphiprotic or aprotic solvents. In addition
to self-dissociation, amphiprotic solvents act as proton donors or acceptors if there are other proton
donors or acceptors in the separation system. However, there are differences in their proton donor
and acceptor capabilities. Methanol, similar to water, has an equal tendency to donate and accept pro-
tons, but basic amide-type solvents are worse proton donors than proton acceptors. Aprotic solvents,
such as acetonitrile, can only accept protons. Inert solvents are capable of neither autoprotolysis nor
donation–acceptance of protons to a considerable extent, which makes them less suitable for NACE
[40].
A number of review papers have been published which look at detection methods [41], selectivity
manipulation [42], and pH of the background electrolyte [43] in NACE. This mode of CE has been
reported to be used for a range of pharmaceutical separations; separation of a number of opium
alkaloids [44], a mixture of cationic drugs [45], a range of tropane alkaloids [46], a range of beta-
blockers [47], tricyclic antidepressants [37], and different basic drugs [48]. NACE has also been
used to separate polar acidic and basic drugs [49] and to perform chiral separation of pharmaceutical
amines [50].
4.3.3 MICELLAR ELECTROKINETIC CHROMATOGRAPHY

Micellar electrokinetic chromatography (MEKC) was developed by Terabe et al. [4], and combines
the separation mechanism of chromatography with the electrophoretic and electroosmotic movement
of solutes and solutions for the separation of constituents in a sample. This mode of CE is covered
in detail in Chapter 3 by Terabe, and utilizes surfactant micelles as a “pseudo-stationary phase”
in the carrier electrolyte. Surfactants are molecules with detergent properties, which are composed
of a hydrophilic water-soluble head group and a hydrophobic water-insoluble hydrocarbon chain
group [e.g., sodium dodecyl sulfate (SDS), CH3 –(CH2 )11 –O–SO− +
3 Na ]. When anionic surfactant
molecules are present in a solution at a concentration above their critical micelle concentration
(CMC) they aggregate to form negatively charged micelles. These are generally spherical in shape
with the hydrophilic head groups oriented in the aqueous buffer and the hydrocarbon chains in the
centre. Figure 4.1 shows the arrangement of the molecules to form a micelle.
MEKC uses the same instrument set-up as in conventional CE and uses uncoated capillaries.
When a water-insoluble hydrophobic compound is added to a micellar solution, it partitions into
the core of the micelle and is solubilized. Conversely, if a water-soluble hydrophilic compound is
added to the solution it will solubilize in the aqueous phase and will not partition into the micelle.
Neutral compounds of intermediate water solubility will partition between the aqueous phase and the
micelle core and the extent of the partitioning depends on the hydrophobicity of the compound. When
high pH micellar solutions are used as the carrier electrolyte in CE, the negatively charged micelles
migrate in the opposite direction to the EOF but the more powerful EOF carries them to the cathode,
Hydrophilic Hydrophilic
head tail
FIGURE 4.1 The arrangement of surfactant molecules when a micelle is formed. Depending on the surfactant,
surfactants will form micelles with different numbers of molecules (Aggregation Number, AN). In the case of
SDS its AN is 62 and its CMC is 8.27 mM.
through the detector. Very water-soluble solutes will remain in the aqueous phase and will separate
due to differences in their electrophoretic mobilities. Solutes, which are completely solubilized by
the micelle, will migrate and reach the detector at the same time as the micelle. The main advantage in
employing MEKC is that uncharged neutral compounds will partition between the aqueous phase of
the electrolyte and the micelle and will separate based on the difference in micelle/water partitioning
and migration times are proportional to their micelle/water partition coefficients, log Pmw .
The composition of the micellar buffer solution can be changed in many ways in order to optimize
the separations. The nature of the surfactant, that is, charge and concentration, the use of additives
such as organic solvents, urea, and CDs can be altered in order to manipulate the separation of an
analyte mixture. Books by Pyell [51] and Baker [31] dedicate chapters to more detailed discussions
of MEKC theory and practice. Review papers by Pappas et al. [52], Molina and Silva [53], and Pyell
[54] also cover many MEKC applications and development options.
4.3.4 MICROEMULSION ELECTROKINETIC CHROMATOGRAPHY

Microemulsion electrokinetic chromatography (MEEKC), first presented in 1991 by Watari [55], is
an extension of MEKC and is probably the most versatile mode of CE that offers the possibility of
highly efficient separations of both charged and neutral solutes with a wide range of water solubilities.
The range of pharmaceutical applications of MEEKC has grown rapidly in recent years with chiral
analysis, pharmaceutical assay, impurity determination, and log P measurements among the most
common. MEKC and MEEKC differ in the type of pseudostationary phase used, as discussed in
the previous section, MEKC electrolytes consist of micellar solutions while MEEKC electrolytes
are composed of microemulsions, with solutes partitioning between the microemulsion droplets and
electrolyte buffer.
Microemulsions are optically clear, thermodynamically stable suspensions of <10 nm diameter
droplets of an immiscible liquid dispersed in another liquid, usually oil droplets in an aqueous contin-
uous phase. These are referred to as oil-in-water (O/W) microemulsions and are the most commonly
used in MEEKC applications; however, water-in-oil (W/O) microemulsions have also been used
for MEEKC analysis of pharmaceuticals [56,57]. The microemulsion droplets are stabilized by the
presence of a surfactant, for example, SDS, which was described in Section 4.3.3. The hydrocarbon
chain of the surfactant resides in the oil droplet while the hydrophilic head remains in the aqueous
phase, which reduces the interfacial tension at the oil–water interface. By incorporating a cosurfac-
tant (usually a short chain alcohol), which bridges the oil–water interface in a manner similar to the
surfactant, the interfacial tension is reduced further to almost zero resulting in a thermodynamically
stable system. This arrangement is shown in Figure 4.2. SDS, an anionic surfactant, is the most
commonly used surfactant in MEEKC and results in a negatively charged droplet. The aqueous
continuous phase of the microemulsion usually contains additives, for example, pH buffers and/or
ion-pair reagents, CDs, or organic modifiers to provide optimum separation conditions.
The separation mechanism in MEEKC is similar to that in MEKC where solutes are separated
by both electrophoretic mechanisms in the capillary and by chromatographic interaction between
the microemulsion droplets and the aqueous phase. Similar to MEKC, hydrophobic solutes favor
inclusion in the oil droplet while hydrophilic solutes will favor the aqueous phase, neutral solutes
will be separated according to differences in their oil/water partitioning coefficients (log Pow ) and
migration times can be directly related to a neutral solutes log P value. Generally, high pH buffers
such as phosphate or borate are used in MEEKC and when a voltage is applied across the capillary,
these buffers generate a high EOF toward the cathode end of the capillary. The negatively charged oil
droplets attempt to migrate through the capillary to the anode but the EOF sweeps them through the
capillary to the cathode. If charged solutes interact with the charged droplet this will also influence
their migration time. Separation, therefore, is dependent on solute size, charge, and hydrophobicity,
which allows the separation of neutral components.
Aqueous phase Na+

SO–3
Na+ Aqueous phase
OH OH
– Na+
SO 3
OH
SO–3
OH
Oil OH
droplet
OH
–
+ SO 3
Na
SO–3
OH
OH OH Na+
Aqueous phase
SO–3
+
Na Aqueous phase
FIGURE 4.2 Schematic representation of an o/w microemulsion droplet.
MEEKC is the most versatile mode of CE for pharmaceutical analysis and offers many advan-
tages over MEKC. It can be applied to a wider range of solutes than MEKC as solutes can penetrate
the surface of a microemulsion droplet more easily than a more rigid micelle and offers greater sep-
aration capability for water-insoluble compounds [58]. MEEKC has been found to provide superior
separation efficiency than MEKC, probably due to improved mass transfer between the microemul-
sion droplet and the aqueous phase [59]. MEEKC also provides a larger separation window, the
size of which can be controlled and therefore offers greater separation capability for hydrophobic
compounds [58,60]. Owing to the unique ability of microemulsions to solubilize water-soluble and
water-insoluble, charged and neutral compounds, MEEKC is particularly suited to performing com-
plex pharmaceutical separations. The many variables and operating parameters involved in MEEKC
provide a large number of method development options for difficult pharmaceutical separations.
Other advantages of utilizing MEEKC include the potential elimination or reduction in sample prepa-
ration steps resulting in rapid analysis times for difficult water-insoluble samples. This is due to the
solubilizing power of microemulsions for water-soluble and water-insoluble compounds. Samples
can be simply dissolved in the microemulsion and injected directly onto the capillary with little or
no pretreatment. This was demonstrated by Broderick et al. [56] and Altria et al. [57] when utilizing
W/O MEEKC for the analysis of a number of pharmaceutical preparations and sunscreen lotions. As
the microemulsion droplets are less than 10 nm in diameter and optically transparent, they are very
well suited to the use of low wavelength detection. This can be advantageous when detecting analytes
with weak chromophores or when the sample matrix interferes with detection at higher wavelengths.
This was demonstrated by McEvoy et al. [61] when detecting ibuprofen at 190 nm using O/W
microemulsion liquid chromatography (MELC). Many reports have been published detailing the use
of MEEKC for pharmaceutical applications and there have been a number of review papers [62–
66] written, which describe further examples of MEEKC applications and advancements in method
development with descriptions of operating parameter effects. An application-based research paper
by Altria [58] highlights the suitability of MEEKC for a broad range of pharmaceutical applica-
tions. Readers can also find detailed theoretical aspects of MEEKC in a paper by Watari [66] and
Electrokinetic Chromatography by Pyell [51].
4.3.5 CAPILLARY ELECTROPHORESIS-MASS SPECTROMETRY

Capillary electrophoresis-mass spectrometry combines the short analysis time and high separation
efficiency of CE with the molecular weight and structural information from the MS to provide a
powerful tool that can be utilized to quantify unknown compounds and impurities. Currently, ESI
serves as the most common interface between CE and MS, as it can produce ions directly from
liquids at atmospheric pressure, and with high sensitivity and selectivity for a wide range of analytes
of pharmaceutical significance. It can also be applied to the detection of a wide range of analytes
without derivatization and gives the information necessary to determine the structural formula of the
analytes of interest.
The most common detection technique in CE is on-column UV absorbance detection, which
can be applied to most pharmaceutical applications as most organic compounds display some UV
absorbance. A drawback of UV detection is its relatively low sensitivity due to the short optical
path length of the capillary detection window. The marriage of CE and MS instrumental analytical
techniques results in an extremely powerful and highly sensitive tool for the separation, identification,
and characterization of a wide range of molecules, especially pharmaceuticals. Smith et al. [67]
first introduced CE-MS in 1988 and since then there has been an increased utility of CE-MS in
pharmaceutical analysis. A chapter covering the coupling of CE to MS [51] describes in more detail
the interfacing of the two methods, in particular electrokinetic chromatography-MS (EKC-MS). A
number of CE-MS reviews have also been published [67–74], which cover the many applications of
this technique and deal with operational, development issues, and quantification [69].
4.3.6 MULTIPLEXED CAPILLARY ELECTROPHORESIS

The demand for high-throughput analytical methods to support the evolution of parallel synthetic
technologies in drug discovery applications has led to the development of multiplexed CE to charac-
terize libraries of compounds and alleviate backlogs in the discovery process. Multiplexed or parallel
analysis is achieved in CE through the use of instruments containing bundles (arrays) of capillar-
ies. The capability to use the same instrument to analyze multiple samples simultaneously, and the
diverse separation conditions that are possible, allows rapid turnaround times. In the past 10 years,
multiplexed CE instruments have become commercially available with a number of companies such
as CombiSep, Beckman Coulter, and Genteon offering multiplexed systems with up to 384 capil-
laries coupled to laser-induced fluoroscence (LIF) and UV detectors. Applications of multiplexed
CE in pharmaceutical analysis include physicochemical profiling [76–78], chiral analysis [79–81],
and organic reaction monitoring [82]. In a recent paper, Marsh and Altria [83] reported the use of
multiplexed CE for the measurement of log P and pKa values, tablet assay, and impurity determina-
tion. While still in its infancy, the use of multiplexed CE in pharmaceutical applications shows clear
advantages over traditional screening techniques. The ability to analyze multiple samples or vary
experimental conditions in parallel offers a unique and powerful tool for drug discovery. Multiplexed
CE has a broad range of high-throughput applications spanning the pharmaceutical, fine chemical,
agrochemical, and biotechnology industries. The multiplexed CE format provides the flexibility
to simultaneously vary separation conditions to speed method development processes. As this CE
approach is relatively new, there are few sources of literature dedicated to the topic. The papers pre-
viously mentioned [76–84] cover the use of multiplexed CE for various pharmaceutical applications
and give brief discussions of method development and instrumentation. A review paper by Pang et al.
[85] gives more detailed information about the instrumentation and a wider range of applications.
4.4 METHODS OF DETECTION

Detection methods for CE analysis are as diverse as those used for HPLC. The most widely used
detection methods used in CE include ultraviolet/visible (UV/Vis) absorbance, LIF, mass spectrom-
etry, conductivity, amperometry, radiometric, and refractive index. When deciding which detection
method is best suited to an intended application, the analyst must first know if the compounds to be
separated can be detected using a certain type of detector, that is, does the analyte have a chromophore,
does it fluoresce, and so forth. If selective detection is desirable then choosing a detector that will only
detect the compound of interest or classes of compound may be required, especially if components
reach the detector simultaneously. Should quantitative and detailed qualitative data be required,
then mass spectrometric detection would be the method of choice. The same criteria for detec-
tors in HPLC apply in CE including sensitivity, selectivity, linear range, and the ability to provide
qualitative data.
This section will only discuss the most common detection methods used for CE analysis of phar-
maceuticals, that is, UV/Vis absorbance, LIFD, mass spectrometry, and electrochemical detection.
In the tables of applications, the method of detection is indicated for each application. This should
aid the interested reader and provide information on detector suitability for intended applications.
4.4.1 UV/VIS ABSORBANCE

UV/Vis absorbance is by far the most common detection mode utilized in CE analysis of pharma-
ceuticals. This is mainly due to the fact that most pharmaceutical compounds are organic in nature
and contain one or more chromophores, which leads the analyst to choose this detection method
first. This reason along with the fact that UV/Vis detectors are the most common detectors in the
majority of laboratories and are easily coupled to the CE instrumentation has led to UV/Vis detection
being the usual method of choice in CE analysis of pharmaceuticals. The most advantageous type of
UV/Vis detectors are diode array detectors (DADs), which utilize multiple wavelength detection to
acquire spectral data that allows purity determination and some peak identification. A drawback of
UV/Vis detection for CE analysis is the reduction on concentration sensitivity compared to its use
in HPLC. With absorbance detection, minimum detectable concentrations in CE are 10–100 times
higher than in HPLC, this is due to the lower path length of the cell. With CE the path length is the
inner diameter of the capillary, which is typically 50–100 µm, compared to path lengths of 10 mm
for HPLC detectors.
Sensitivity can be enhanced through the use of larger inner diameter capillaries; however, larger
capillary diameters produce a higher current, which leads to joule heating and separation efficiency
decreases dramatically when the capillary diameter is increased above 100 µm. There have been a
number of approaches to enhance sensitivity by increasing the path length without increasing the
inner diameter of the capillary. Moring et al. [85] used a “z-cell” with a 3 mm capillary section for
path length, which gave a 10-fold increase in sensitivity. Tsuda et al. [86] used a 1000 µm path length
“rectangular capillary” to increase sensitivity, however, their application is limited by commercial
availability and cost. Wang et al. [87] used a “multireflection cell” to effectively increase the path
length more than 40-fold and reported a 40-fold increase in peak height compared to a single pass
cell. A “bubble cell” can also be used to increase the path length by only increasing the diameter
of the capillary where the light beam goes through it. This was demonstrated by Heiger [88] who
reported a threefold increase in peak height when using a 50 µm ID capillary with a 150 µm “bubble.”
Some manufacturers supply specialized detection cells such as Agilents high sensitivity detection
cell, which provides a 10-fold increase in sensitivity and an extended linear range.
4.4.2 LASER-INDUCED FLUORESCENCE DETECTION

Laser-induced fluorescence is the optical emission from molecules that have been excited to higher
energy levels by absorption of electromagnetic radiation. Laser-induced fluorescence detection
(LIFD) can be regarded as the most sensitive optical detection technique to be coupled to CE with
sensitivity proportional to the intensity of emitted light. The main advantage of fluorescence detection
compared to absorption measurements is the greater sensitivity achievable because the fluorescence
signal has a very low background. Intense light is available from the laser and this light can be
efficiently focused into the narrow channel of the capillary in the detection window. LIFD is not a
universal detection method and solutes must possess native fluorescence or must be able to be
derivatized to generate a fluorophore. The pharmaceutical analyst, when deciding on the suitability
of LIFD to their separation must consider this. There are several commercially available LIF detec-
tors designed for use with CE. Beckman Coulter offers a dual wavelength LIF detector that works
with a 488 nm (Ar+ laser) and a 633 nm (HeNe laser). The Picometrics ZETALIF detector is a single
wavelength excitation LIF detector which is a modular design and can be used externally to the CE
instrument. Picometrics also offer another commercial detector, which can be used inside a cassette
of any CE instrument.
A number of publications are available that describe the theoretical aspects of LIFD [51,89],
which will provide the reader with a more detailed coverage of this topic. A number of reviews have
also been published which cover the use of CE-LIF detection for the analysis of a range of compounds
including pharmaceuticals [90–94]. Zhang et al. [94] also show the use of wavelength-resolved LIFD
for quantitative and semi-qualitative analysis of target solutes.
4.4.3 MASS SPECTROMETRIC DETECTION

Section 3.5 covered briefly the use of CE-MS for pharmaceutical applications. The suitability of
CE-MS to pharmaceutical applications is due to its appeal as a universal and selective detector,
which provides very good sensitivity. Significantly, CE-MS can provide quantitative data and where
required, conclusive qualitative information.
4.4.4 ELECTROCHEMICAL DETECTION

The two types of electrochemical detection used most often in CE are amperometry and conductivity.
Although not as widely used in pharmaceutical CE as UV/Vis or LIFD, these methods are used when
alternative detection methods are not suitable, that is, solutes lack chromophores, fluorophores or
if sensitivity is poor using optical detection methods. Amperometric detection measures the current
that results from oxidation or reduction of electroactive solutes at a working electrode surface.
Oxidation occurs when an electron is transferred from a solute molecule to the working electrode in
the amperometric cell, during oxidation the charge on the solute increases. Reduction occurs when
electron transfer is in the opposite direction, from the working electrode to the solute molecule, and
the charge on the solute is reduced. Oxidation and reduction of the solutes is caused by a potential
being applied across a supporting electrolyte between a working and reference electrode. The current
that flows through the working electrode is proportional to the number of electron transfers taking
place, and therefore to the concentration of the solute. A review by Lui et al. [95] gives a more
detailed account of amperometric detection in CE. A number of recent applications of amperometric
detection for CE analysis of pharmaceuticals include determination of the hydrolysis rate constants
and activation energy of aesculin [96] and analysis of acyclovir [97]. The basis of conductivity
detection is the change in electrical conductivity of a solution when an ionic solute is introduced into
it. A conductivity detector has two electrodes in the cell and a high frequency alternating current is
applied to the electrodes. The buffer from the capillary flows between the electrodes and when an ionic
solute comes into the cell it decreases the electrical resistance of the solution, thereby increasing
the electrical conductivity. The measured conductivity is proportional to the concentration of the
solute. As only ionic solutes cause a change in conductivity, this method cannot be used with MEKC
or MEEKC for the detection of separated neutral solutes. As with amperometry, there have been
relatively few applications of conductivity detection for pharmaceutical analysis by CE, compared
to optical detection methods. Bowman et al. [98] used this technique for the analysis of small amines
in pharmaceuticals. Readers are encouraged to consult reviews by de Silva [99], Baldwin [100],
Holland and Leigh [101], and Wang and Fang [102] who discuss in more detail electrochemical
detection in CE analysis.
TABLE 4.1
Features and Limits of Detection of CE Detection Methods
Approximate
Detection Detection On-column LOD
Method Requirements Selective Universal Qualitative Detection (Molarity)
UV/Vis • Chromophore Yes No Some, Yes 10−5 –10−7
absorbance groups using
• Absorb UV above DADs
190 nm
LIF • Fluorophore Yes No No Yes 10−13 –10−16
groups
• Derivatized to
contain
fluorophores
MS • No special Yes Yes Yes No 10−8 –10−10
requirements
Amperometry • No special Yes No No No 10−7 –10−10
requirements
Conductivity • Solutes need to be No Yes No No 10−7 –10−9
charged
4.4.5 SELECTING THE MOST SUITABLE DETECTOR

When deciding which detection method is most suitable to an application, the pharmaceutical analyst
needs some fundamental information about their sample and also the analytical requirements. The
main points to be considered when choosing a detection method are sensitivity and selectivity
requirements and the ability to provide qualitative information. If these requirements are known
then the list of available detectors is reduced. Knowledge of the chemical structure of target analytes
is indispensable when choosing an appropriate detector as information about chromophore and
fluorophore groups, solute charge, molecular weight, and so forth narrows the options available
to the analyst. Table 4.1 lists the most common detectors as discussed in this section and shows
the limits of detection, detection requirements, and so forth for each type that will aid in detector
selection. Readers are encouraged to consult reviews by Xu et al. [92] and Swinney and Bornhop
[103], which cover CE detectors in more detail.
4.5 SELECTING THE MOST SUITABLE MODE OF CE

As previously discussed in Sections 3.1 through 3.5, the various modes of CE offer various possib-
ilities to the pharmaceutical analyst. Depending on the complexity of the sample, the nature of its
components, and the intended application, the nature of pharmaceutical analytes varies enormously
from highly water-soluble compounds to species only soluble in pure organic solvents. Each of these
modes of CE will provide various advantages and disadvantages for the separation and detection of
different classes of compounds. FSCE is well reported for many pharmaceutical assays and impurity
determinations, mainly for water-soluble and some water-insoluble acidic and basic compounds.
NACE is suited to the separation of highly water-insoluble charged compounds and the volatile nature
of the solvents used makes it ideal for coupling to a MS. NACE also offers different selectivity to
FSCE for the separation of pharmaceuticals [45,46]. For complex separations, MEEKC is the method
of choice as it gives the analyst more method development options, it has a wider application range
than MEKC and unlike FSCE and NACE, this mode of CE will separate both charged and neutral
compounds. The use of W/O MEEKC has shown applicability of this method for the analysis of
highly water-insoluble compounds in hydrophobic matrices without the need for lengthy sample
pretreatment or expensive organic solvents [56,57]. For rapid screening, enantiomeric separations
and physicochemical measurements of pharmaceuticals, multiplexed CE has been used [77–85]. The
main pharmaceutical application areas of CE are chiral separations, pharmaceutical assay, impurity
determinations, physicochemical measurements, and the analysis of small molecules and ions. Each
of these areas will be discussed in the next section and a selected number of applications will
be detailed in tabular form showing the mode of CE used. As the numbers of applications are too
numerous to be included in this chapter, sources of further applications will be cited for the interested
reader.
4.6 PHARMACEUTICAL APPLICATIONS OF CAPILLARY

ELECTROPHORESIS
Since the first application of CE to pharmaceutical analysis in 1988 [15], the number of reports
describing the analysis of numerous different pharmaceutical compounds and formulations by CE
has expanded to be in the hundreds. Lunn [13] has covered in detail the CE separation methods
for over 700 pharmaceutical compounds taken from hundreds of publications. The objective of this
applications section is to provide the reader with an introduction to the main pharmaceutical appli-
cation areas with examples of successful separations including sources of comprehensive reviews
for each area. Within each application area there may be more specialized applications (e.g., chiral
MEEKC) for enantiomeric separations and some aspects may be cross-linked with other applications
or detection methods to comprehensively cover the topic of this chapter. Sources of extra reading
material in the form of review papers will be cited where relevant.
4.6.1 CHIRAL PHARMACEUTICAL ANALYSIS

Chiral analysis has become one of the most studied areas of CE, as it is a powerful analytical tool
for separating chiral compounds, which is of major importance in pharmaceutical applications. The
two most important analytical techniques used in chiral separations still are LC and CE, followed
by GC and capillary electrochromatography (CEC). Compared to other techniques, CE has several
advantages, the high resolving power, and low consumption of sample, solvent, and chiral selec-
tor. In method development, the chiral selector is added to the background electrolyte instead of
using a range of expensive chiral LC columns. Chiral CE provides high flexibility in choosing and
changing types of selectors. The possibility of low wavelength UV detection for CE also allows
the separation and detection of analytes with poor chromophores, which are difficult to detect by
UV/high-performance liquid chromatography (UV/HPLC).
This section will look at recently developed chiral selectors and chiral CE separating techniques.
Readers are referred to papers by Rizzi [104] and Gubitz and Schmid [105], which deal with the
fundamental aspects and principles of chiral CE. A book dedicated to CE for chiral analysis by
Chankvetadze [106] covers a number of aspects of chiral CE including the use of chiral metal
complexes, macrocyclic antibiotics, crown ethers, and chiral MEKC. Review papers by Gubitz
and Schmid [107], Hoogmartens and coworkers [16], and Amini [108] cover recent applications of
chiral CE. The following paragraph gives a very brief introduction to the basic theory of chiral CE
separations.
At low pH, basic drugs are positively charged and their migration toward the cathode can be
retarded by a chirally selective complexing agent, resulting in the separation of enantiomers of
differing affinity for the agent. This principle has been demonstrated for the resolution of chiral basic
drugs using CDs as a chiral selector [109]. At high pH, chiral acidic drugs are negatively charged
and migrate against the EOF toward the anode. Neutral chiral selector agents are swept along the
capillary with the EOF toward the detector; thus, complexation reduces the migration time of the
drug and results in enantioseparation [110].
There are many types of chiral selectors that have been applied to the separation of enantiomers
by CE, but the most common are native and derivatized CDs. Other chiral selectors, which have been
applied to CE separations, include natural and synthetic chiral micelles, crown ethers, chiral ligands,
proteins, peptides, carbohydrates, and macrocyclic antibiotics [105,111–114]. A review by Blanco
and Valverde [114] describes the separation capabilities of various chiral selectors and provides
criteria for their choice in terms of molecular size, charge, and the presence of specific functional
groups or substructures in the analytes.
Chiral CE using mass spectrometric detection has also been utilized for detailed qualitative
analysis of pharmaceuticals, a review by Shamsi [115] covers the various modes of CE-MS including
the use of neutral CDs for FSCE-MS and MEKC-MS.
4.6.1.1 Cyclodextrins
Native and derivatized CDs are employed routinely for enantiomeric separations. These naturally
occurring carbohydrates have a bucket-like shape, which allows analytes to become included into
the CD cavity by complexation. The migration time of the analytes is dependent on their mobilities
and degree of interaction with the CD. Enantiomeric separations are brought about by the difference
in the stabilities of the complexes formed between each enantiomer and the CD molecule. The
chiral hydroxyl groups around the rim of the CD can interact enantioselectively with chiral analytes,
which can fit into the CD cavity, leading to the separation of enantiomers with differing binding
constants. There are three types of CDs, α, β, and γ, each differing in the number of glucose units
they are composed of. β-CDs are the least soluble in water but their solubility can be improved
by the addition of urea to the background electrolyte. Because enantioselection is based on the
formation of inclusion complexes between the CD host and the chiral solute, the type of CD chosen
is a major factor for achieving efficient resolution of enantiomers. Recent reviews [20,117] have
covered the background to the use of CDs in chiral CE. Systematic approaches to the development
of chiral CE methods using sulfated CDs for the separation of acidic, basic, neutral, and zwitterionic
species have also been reported [118,119]. Systematic method development approaches for several
selected compounds were performed by modifying method parameters, such as the concentration
of the chiral selectors, buffer pH, type of organic modifiers, buffer type, temperature, and applied
voltage. Many practical aspects were also discussed through several specific examples in order
to demonstrate how to develop and validate a precise, sensitive, accurate, and rugged separation.
Perrin et al. [75] studied the robustness of enantiomeric separations of acidic, basic, and neutral
compounds using highly sulfated CDs in a low pH phosphate buffer with short-end injection. An
eight-factor experimental design was used to study the robustness of chiral separations of propanolol
(basic), praziquantel (neutral), and warfarin (acidic). These factors included electrolyte pH, CD and
electrolyte concentration, capillary temperature, and applied voltage. Results showed that control of
pH is critical, especially for the separation of acidic enantiomers while CD concentration changes
did not adversely affect the robustness of the separations.
4.6.1.2 Crown Ethers

Crown ethers have been developed and synthesized for use in Chiral CE. Kuhn et al. [119] first
utilized chiral crown ethers for the enantiomeric separation of drugs and amino acids in 1992. Since
then there have been a number of applications of these chiral selectors in CE analysis, 18-crown-6
tetracarboxylic acid [18C6H(4)] is to date, the only chiral crown ether that has been used for chiral
CE separations and forms complexes with chiral analytes in a way similar to CDs, based on dif-
ferences in complex formation energies. Two new 18-crown-6 diaza derivatives were investigated
as chiral selectors [120]. These derivatives did not show any chiral selectivity toward the investi-
gated analytes, they could serve well as an additive to improve the chiral resolution in combination
with CDs.
4.6.1.3 Macrocyclic Antibiotics

Although CDs and their derivatives are present in the majority of chiral CE applications, the use
of macrocyclic antibiotics can be observed with increasing frequency [121,122]. The number of
macrocyclic antibiotics utilized for chiral CE has now exceeded 10 and includes four main groups:
glycopeptides, polypeptides, ansamycins, and aminoglycosides (although this group is not always
considered as macrocyclic). Thanks to their macrocyclic structure and the diversity in chemical
groups, they exhibit a variety of interactions (inclusion, electrostatic, hydrogen bond, hydrophilic-
lipophilic, or other Van der Waals bond type), which enables them to achieve high chiral resolution
with a wider range of analytes (acidic or basic, with large or small molecular sizes, etc.) [124].
Macrocyclic antibiotics are enantioselective for positively charged solutes using ansamycins and
enantioselective for anionic compounds using the glycopeptides. Within a given class of antibiotics
such as the glycopeptides, enantioselectivity may also be altered by use of micelles, uncoated vs.
coated capillaries, or manipulation of operating parameters such as pH or organic modifiers. In a
review paper, Aboul-Enein and Ali [123] describe the chemistry of these antibiotics, the effect of
chromatographic conditions on enantioselectivity, the mechanism of resolution, the applications and
limitations of the compounds in LC and CE.
4.6.1.4 Oligo- and Polysaccharides

Apart from CDs, many other linear and cyclic oligo- and polysaccharides have been used as chiral
selectors [127] (e.g., monosaccharides as d-glucose, d-mannose, or polysaccharides as dextrins,
dextrans, and many others). Park et al. [125] used highly sulfated cyclosophoraoses, the sul-
fated derivatives of chiral unbranched cyclic β-(1→2)-d-glucans, and successfully used them to
separate five basic chiral drugs. These derivatives exhibited higher resolution than the original
cyclosophoraoses. Another new charged polysaccharide, N-(3-sulfo, 3-carboxy)-propionylchitosan,
was studied by Budanova et al. [126]. This appeared to have a different chiral recognition mechanism
from the charged polysaccharides. The use of polysaccharides combined with CDs has also been
reported [128].
4.6.1.5 Proteins
A protein or glycoprotein consists of amino acids or amino acids and sugars, both of which are
chiral. Therefore, proteins have the possibility to discriminate between the enantiomers of a chiral
molecule. CE methods using proteins as the immobilized ligands or running buffer additives are
attractive for the separation of enantiomeric mixtures. Although separation efficiencies by CE are
generally somewhat higher than those obtained with HPLC, chiral CE methods based on proteins
have a disadvantage of low efficiencies in addition to low loadability. Proteins such as bovine serum
albumin (BSA) and human serum albumin (HSA) and several additional proteins have been used as
chiral selectors in CE. Reviews by Haginaka [111] and Millot [129] provide a more in-depth coverage
of recent chiral CE separations using proteins as chiral selectors. Although proteins display very good
chiral discrimination for pharmaceutical enantiomers, their high background UV absorbance limits
their usefulness in chiral CE. More can be gleaned on protein analysis by CE from Chapter 2 by
Hempe.
4.6.1.6 Chiral MEKC

Enantiomeric separation by MEKC involves the addition of a chiral surfactant or a chiral selective
agent to the background electrolyte. A number of selectors have been used in chiral MEKC including
crown ethers and CDs. These selectors are usually used in combination with chiral or achiral micelles,
for example, CD modified MEKC (CD-MEKC) [130]. MEKC usually utilizes negatively charged
micelles formed from anionic surfactants such as SDS, which constitutes the pseudostationary phase.
The separation is achieved by differential partitioning of analytes between the pseudostationary phase
and the bulk aqueous phase. Chiral surfactants used for chiral MEKC separations include natural
surfactants (bile salts, amino acids, and glucose), monomeric synthetic surfactants, and polymeric
surfactants [131–133]. Chiral separation in MEKC is affected by the affinity of the enantiomers
toward the micelles and the concentration of the micellar phase, which depends on the aggregation
properties if the chiral surfactants or chirally modified surfactants. Dobashi et al. [131] first reported
the use of chiral mixed micelles to obtain optical separation of enantiomers in 1989 and since then,
there have been many reports on enantiomer separations by MEKC. Otsuka and Terabe [134] has
reviewed the use of natural and synthetic chiral surfactants in MEKC, in a review by Ha et al. [16],
recent applications, and aspects of chiral MEKC are covered in more detail.
4.6.1.7 Chiral MEEKC

Chiral separation by MEEKC was first demonstrated in 1993 [135] using (2R,3R)-di-n-butyl tartrate
as a water immiscible chiral selector in the microemulsion electrolyte that successfully separated
ephedrine enantiomers. Chiral MEEKC offers increased method development flexibility, the ability
to custom tune chiral resolution through the increased method development options, and unique solu-
bility capabilities for both analytes and additives when compared to typical capillary electrophoretic
techniques. Chiral MEEKC offers the possibility to simultaneously determine drug enantiomers of
more than one compound along with chemical impurities/additives. Importantly, the greatest advan-
tage of this technique is the ability to separate more hydrophobic racemic components for which CE
is not currently a preferred methodology. Only in recent years, more research into this area of chiral
separations has been carried out by Pascoe and Foley [136], Mertzman and Foley [137,138], and
Zheng et al. [139] among others. Chiral separations in MEEKC can be achieved by utilizing a number
of chiral agents such as a chiral surfactant dodecoxycarbonylvaline (DDCV) [136], chiral alcohols
as cosurfactants [139], a combination of both of these chiral components [140], and the use of CDs
in CD modified MEEKC as the only chiral agent and in combination with both SDS and DDCV
as surfactants [137]. Although no literature in the form of reviews and few books are dedicated to
chiral electrokinetic chromatography (MEKC and MEEKC), Pyell [51] includes a chapter dealing
with enantiomer separations by electrokinetic chromatography. There are a number of reviews by
Ha et al. [16], Marsh et al. [63], and Huie [64], which cover these subjects in some detail and give
examples of recent applications.
4.6.1.8 Pharmaceutical Applications of Chiral CE

Chiral selectors are routinely used in industry for a wide variety of applications and there have been
many reports published describing their use in CE for the enantiomeric separations of pharmaceutical
compounds. A quick search of scientific databases will yield a far greater number of publications
dealing with chiral CE than is possible to cover in this chapter; however, Table 4.2 contains a
selection of such applications and details of carrier electrolytes used. Readers are referred to review
papers [16,63,64,116,107,141–144], which comprehensively cover recent chiral applications of CE
and provide more detailed theory about chiral separations and detection methods.
TABLE 4.2
Selected Chiral Separations of Pharmaceuticals
Application Chiral Selector Mode References
Enantiomeric purity methods for A range of α-, β-, and γ-cyclodextrins, FSCE [153]
three pharmaceutical compounds highly and singly sulfated
Evaluation of chiral purity of Sulfobutyl ether β-cyclodextrin FSCE [154]
frovatriptan
Chiral separation of four Hydroxypropyl-β-cyclodextrin FSCE [155]
fluoroquinolone compounds
Chiral separation of bupivacaine Human serum albumin FSCE [156]
enantiomers
Chiral separations of a range of Sulfated β-cyclodextrin (β-CD-(SO−
4 )4 NACE [157]
pharmaceuticals
Dopa enantiomers (+)-(18-crown-6)-2,3,11,12-tetracarboxylic FSCE [147]
acid
Chiral separation of Hydroxypropyl-β-cyclodextrin FSCE [145]
N -nitrosonornicotine in tobacco
Enantioseparation of Heptakis-(2,3, 6-tri-O-methyl)-β-CD FSCE [158]
erythro-mefloquine and its Heptakis-(2,3-di-O-methyl-6-sulfo)-β-CD
analogues Randomly sulfated β-CD
Enantioseparation of nine racemic Highly sulfated β-cyclodextrin FSCE [159]
arylglycine amides
Enantioseparation of basic Sulfated cyclodextrins FSCE [160]
pharmaceutical compounds
Enantiomeric separation of (+)-(18-crown-6)-2,3,11,12-tetracarboxylic CE-MS [161]
compounds containing primary acid
amine groups
Separation of 4 diastereomers of an Highly sulfated α, β, and γ-CDs FSCE [162]
antiviral agent with 2 chiral centers
Separation of (R-(−) and S-(+) Carboxymethyl-γ-cyclodrextrin (CM-γ-CD) FSCE [205]
citalopram) in tablets
Lisuride bulk substance Acidic electrolyte with the addition of FSCE [206]
gamma-cyclodextrin (γ-CD)
Identification and quantification of Heptakis-(2,3, CE-MS [207]
cis-ketoconazole impurity in tablets 6-tri-O-methyl)-beta-cyclodextrin
and gel
Separation of omeprazole 40 mM phosphate, pH 2.2, 30 mM β-CD, FSCE [208]
enantiomers and 5 mM sodium disulfide
4.6.1.9 Examples of Chiral CE and Validated Methods

N -nitrosonornicotine (NNN) is a nitrosamine found in cured tobacco and is believed to be one of
the main carcinogenic constituents of tobacco. Racemic NNN is present as (R) and (S) enantiomers
and it is known that (S)-NNN undergoes more 2 hydroxylation than (R)-NNN. For rapid and effi-
cient enantiomeric separation of this nitrosamine, McCorquodale et al. [145] developed a simple
CE method, employing a citric acid buffer with hydroxypropyl-β-CD (HP-β-CD). Separation was
achieved in 4 min.
A nonaqueous chiral CE method was developed and validated by Olsson et al. [146] for the
enantiomeric separation of omeprazole an antiulcer drug and its metabolite 5-hydroxyomeprazole.
Heptakis-(2,3-di-O-methyl-6-O-sulfo)-β-CD (HDMS-β-CD) was chosen as the chiral selector in an
5-OH-omeprazole
Omeprazole
mAU
50
40 (b)
30
(a)
20
10
5 10 15 20 25 30
Time (min)
FIGURE 4.3 Enantiomeric separation of (a) omeprazole and 5-hydroxyomeprazole at a concentration of 1 mM

and (b) omeprazole and 5-hydroxyomeprazole at a concentration of 1 and 3 mM, respectively, using 30 mM
HDMS-β-CD, 1 M formic acid, 30 mM ammonium acetate, 25 kV, 5 s injection at 50 mbar, capillary and tray
temperature 16◦ C. (From Olsson J., et al., J. Chromatogr. A, 1129, 291–295, 2006.)
ammonium acetate buffer acidified with formic acid in methanol. The separation is illustrated in
Figure 4.3.
The addition of a chiral crown ether [(+)-(18-crown-6)-2,3,11,12-tetracarboxylic acid] to the
background electrolyte allowed Blanco and Valverde [147] to separate the enantiomers of benser-
azide and determine the enantiomeric purity of Dopa (3,4-dihydroxyphenyl-alanine). Levodopa is
the main active ingredient in the pharmaceutical formulation Madopar® , which is used to treat
Parkinson’s disease, while dextradopa causes unwanted side effects. The dextradopa impurity was
clearly resolved from the main peak and determined to be 0.5%.
The validation criteria for chiral CE methods are similar to those employed for the validation
of chiral HPLC methods and include limits of detection, detector linearity, recovery, precision, and
robustness. A number of validated chiral CE methods have been reported, as described previously
Olssen et al. [146] developed and validated a chiral NACE method for the enantiomeric determination
of an antiulcer drug and its metabolite. Enantiomeric determination of a local anesthetic, ropivacaine
in a pharmaceutical formulation showed the required limit of quantification of 0.1% of the impurity
[148]. Jimidar et al. [149] reported a validated CE method for the chiral separation of an Alzheimer’s
treatment, galantamine hydrobromide, using an α-CD chiral selector and the method was successfully
included in a New Drug Application. Song et al. [150] achieved a LOD of 0.05% for the undesired
enantiomer of an M3 agonist using a highly sulfated γ-CD in a low pH buffer. Other validated chiral
CE methods are also cited [151,152].
4.6.2 PHARMACEUTICAL ASSAY

Assay of pharmaceutical substances and formulated products is one of the most important and
regulated activities in the pharmaceutical analysts laboratory. Regulatory authorities require strict
validation standards to show that analytical assay methods are robust, accurate, repeatable, and
suitable for their intended purpose. In the pharmaceutical industry, HPLC has dominated most ana-
lytical assay determinations and it is well established as the method of choice in most laboratories.
In addition, pharmacopoeia monographs specify HPLC and titrimetric methods for the majority of
pharmaceutical assays. To date, CE is not used extensively in QC work despite displaying excellent
efficiencies, resolution, asymmetry factors, and signal-to-noise ratio. This is mainly due to the fact
that CE can suffer from insufficient sensitivity and repeatability to control impurities in pharmaceu-
tical substances at the levels required. These issues have been addressed somewhat with sensitivity
improvements being reported through the use of alternative detection techniques such as mass spec-
trometry and LIFD and the use of high sensitivity detection cells for UV detection. When conducting
CE analyses, the electrophoretic conditions inside the capillary can sometimes vary slightly between
injections, which can lead to greater variability in peak migration times. Poor injection precision due
to the very small volumes injected can also cause variations in peak areas. A number of approaches
can be made to overcome these problems. Migration times and peak areas can be calculated relative
to an internal standard, which leads to improvements in repeatability [164]. Greater migration time
reproducibility can also be achieved by applying the separation voltage across the capillary for a
very short time prior to injection and separation [165]. The factors affecting CE reproducibility and
efforts to address the problem have been covered in more detail by Shihabi and Hinsdale [166],
Schaeper and Sepaniak [167], and Mayer [168]. Many factors are involved in reproducibility. Some
of these, such as temperature control, voltage control, and sample injection precision, are inherent in
the design of the instrument. Other factors, such as the quality of the reagents used and the manner
in which the instrument is programmed and operated, are completely in the hands of the user. These
two factors, the quality of the instrument and the quality of the operation, are both required in order
to achieve reproducible results. A commercially available capillary treatment system of proprietary
buffers and rinse solutions has been shown to improve CE repeatability as the capillary is coated
with a bilayer of surfactants ensuring that the surface coverage and the EOF is consistent between
injections and between capillaries [169]. Figure 4.4 highlights the consistency of the EOF when
using the buffer coating system compared to a standard phosphate buffer when using a capillary
composed of 19 different channels. In Figure 4.4a, the peaks in the channels have different speeds
and the separation is poor while Figure 4.4b shows a single peak due to a consistent EOF in each
channel, which makes the peaks move at the same speed.
There is a time lag before new analytical methods such as CE are accepted by regulatory author-
ities in the submission dossiers for new medicines and pharmaceutical substances. Regulatory
authorities are, however, beginning to recognize CE with general monographs appearing in the
BP, EP, and Japanese Pharmacopoeia using CE for identification tests. For example, a test for related
substances of levocabastine hydrochloride by MEKC has been included in the British Pharmacopoeia
2005 [163].
One of the major advantages of using CE as an alternative to HPLC methods for pharmaceutical
assay is the relatively small solvent consumption, that is, milliliters of CE electrolyte compared
to liters of HPLC mobile phase, increased efficiency, reduced analysis time, and the possibility of
fewer sample pretreatment steps. The ability to quantify a range of sample types using a single set
of CE conditions is another strong feature and can contribute to considerable savings in analysis and
system setup times. This is particularly true when using multiplexed CE systems where high sample
throughput is possible. The merits of using CE compared to HPLC for pharmaceutical analysis will
be discussed in more detail in a later section.
Some recent examples of pharmaceutical assay by CE include a MEEKC method for the deter-
mination of folic acid in tablets giving a precision of <1.2% relative standard deviation (RSD) and
recovery of 99.8 ± 1.8% at three concentration levels [170]. Lehmann and Bergholdt [171] devel-
oped and validated a high-precision CE method for the main component assay of ragaglitazar, which
met the acceptance criteria that are set for HPLC main component assays. In a separate study, Jamali
and Lehmann [172] used a FSCE method to analyze ragaglitazar and its counterion arginine in active
pharmaceutical ingredients (APIs) and low dose tablets. The method was suitable for the assay and
identification of ragaglitazar and arginine, chiral purity of ragaglitazar, and the purity of ragaglitazar,
with percentage recovery found to be 101–106% for ragaglitazar and 101–125%. Pajchel et al. [173]
used a phosphate buffer supplemented with SDS to develop a selective and precise assay method for
quantitative determination of benzylpenicillin, procaine, benzathine, and clemizole. The separation
is shown in Figure 4.5.
A number of validated CE methods for pharmaceutical assay have been reported; a selection of
recent methods along with the mode of CE used and the run buffer are shown in Table 4.3.
(a)
22.00
20.00
18.00
16.00
14.00
0.00 2.00 4.00 6.00 8.00

Time (min)
(b)
35.00
30.00
25.00
20.00
15.00
0.00 2.00 4.00 6.00 8.00

Time (min)
FIGURE 4.4 Separations on a multibore capillary using phosphate buffer or Celixir buffer. (a) Separation
using phosphate buffer: 50 mM phosphate 2.5, multibore capillary 19 × 25 µm channels, 27 cm long, 130 µA,
+5 kV, 30◦ C, detection at 200 nm, sample salbutamol 1 mg/mL, 1 s injection. (b) Separation using Celixir
buffer: 50 mM phosphate 2.5, multibore capillary 19 × 25 µm channels, 27 cm long, 90 µA, +5 kV, 30◦ C,
detection at 200 nm, sample salbutamol 1 mg/mL, 1 s injection. Multibore capillary 19 × 25 µm channels, 27
cm long, +5 kV, 30◦ C, Elixir buffer pH 2.5, 90 µA, 200 nm. (From Altria K.D., J. Pharm. Biomed. Anal., 31,
447–453, 2003.)
4.6.3 IMPURITY PROFILING OF PHARMACEUTICALS

Impurities in pharmaceuticals are the unwanted chemicals that remain with the APIs, or develop
during formulation, or upon aging of both the API and formulated pharmaceutical product. The pres-
ence of these unwanted chemicals even in small amounts may influence the efficacy and safety of the
pharmaceutical products. Impurity profiling (i.e., the identity as well as the quantity of impurity in
the pharmaceuticals) is another important aspect of the pharmaceutical analysts work and must meet
strict regulatory requirements. The different pharmacopoeias, such as the British Pharmacopoeia
(BP) and the United States Pharmacopoeia (USP) have tests for related substances incorporated
into most monographs for pharmaceutical compounds and formulations. In addition, The Inter-
national Conference on Harmonization (ICH) has published guidelines on impurities in new drug
substances [190]. Impurity profiling is generally carried out by HPLC and cross-correlated with other
28.00 BP
26.00
24.00
22.00 C
20.00
18.00
16.00
14.00 B
12.00
10.00
8.00
P
6.00
4.00 MEOH
2.00
0.00
–2.00
–4.00
0.00 5.00 10.00 15.00 20.00 25.00 30.00

Minutes
FIGURE 4.5 Electropherogram of benzylpenicillin and procaine, benzathine, and clemizole. Buffer
phosphate–borate (pH 8.7), supplemented with 14.4 g/L SDS. Detection wavelength was 214 nm. Separations
were performed in 60 cm (52 cm effective length) × 75 µm ID fused-silica capillary coated with polyimide
(AccuSep capillaries, Waters) thermo regulated at 25˚C, with voltage of 18 kV applied (current about 140 µA).
Hydrodynamic injection by gravity-driven siphoning 10 s. P, procaine; B, benzathine; C, clemizole; BP, ben-
zylpenicillin; MEOH, methanol (EOF marker). (From Pajchel G., et al., J. Chromatogr. A, 1032, 265–272,
2004.)
chromatographic methods such as TLC or an alternative HPLC method. The overriding requirements
in impurity determination methods are that all likely synthetic and degradative impurities are resolved
from the main compound and these can be quantified at 0.1% and lower levels. This quantification is
possible using commercial CE instrumentation with standard capillaries. The structural impurities of
a drug will often possess similar structural properties of the main component, which makes achieving
resolution of the compounds challenging. The high separation efficiencies and different separation
mechanism of CE often allows easier resolution of the main component and related substances than
when using HPLC or TLC. The ability to easily alter separation parameters and the applicability of
one CE method to a range of compounds can make CE a cheaper and more rapid complementary
method to LC methods. Because LC and CE utilize different separation mechanisms, the analyst can
have a high degree of confidence in the cross-correlation of agreeable results.
Various modes of CE have been used for impurity determination with CE. Hansen described
the comparison of CZE (FSCE), MEKC, MEEKC, and NACE for the determination of impurities
in bromazepam [193] and found that NACE provided the best technique for the poorly water-
soluble compounds with impurities determined to be 0.05%. Chiral CE methods can be used to
determine enantiomeric impurities [191]. Readers are referred to a paper by Sokoliess and Koller
[192] describing method development for chiral purity testing in CE. FSCE is the most widely used
as most drugs and impurities are acidic or basic. Low pH buffers are used for basic drug impurities
and high pH buffers for acidic compounds.
TABLE 4.3
Selected Validated CE Methods for Pharmaceutical Assay
Application Buffer Mode References
Determination of sertaconazole in 20 mM phosphate + 40% acetonitrile FSCE [174]
pharmaceutical preparations
Development and validation of a 20 mM acetate (pH 4.5) FSCE [175]
quantitative assay for raloxifene
Analysis of atorvastatin calcium 25 mM sodium acetate (pH 6) FSCE [176]
using capillary electrophoresis and
microchip electrophoresis
Determination of ibuprofen and 20 mM FSCE [177]
flurbiprofen N-(2-acetamido)-2-aminoethanesulfonic
acid (ACES) with 20 mM imidazole and
10 mM alpha-cyclodextrin of pH 7.3
Separation and quantification of 1/15 M phosphate buffer (pH 10) FSCE [178]
fibrate-type antihyperlipidemic
drugs
Simultaneous determination of six 60 mM phosphate (pH 2.5) FSCE [179,180]
angiotensin-II-receptor antagonists. 55 mM phosphate 15 mM SDS (pH 6.5) MEKC
Simultaneous analysis of four 80 mM sodium phosphate buffer (pH 3.5) FSCE [181]
atypical antipsychotics
Determination of cyclizine 50 mM phosphate (pH 2.3) FSCE [182]
hydrochloride in tablets and
suppositories
Simultaneous determination of six Borate buffer containing sodium cholate and MEKC [183]
corticosteroids in commercial sodium deoxycholate
pharmaceuticals
Simultaneous determination of 0.08 M borate (pH 8) FSCE [184]
penicillin G, procaine, and
dihydrostreptomycin in veterinary
drugs
Quantitative analysis of quinolizidine 1% acetic acid, 50 mM ammonium acetate, NACE [185]
alkaloids in Chinese herbs 20% acetonitrile in methanol
Simultaneous determination of four 50 mM ammonium acetate, 1 M acetic acid NACE [37]
tricyclic antidepressants in acetonitrile
Analysis of benzodiazepines CeofixTM Buffer system, using dynamically CE-MS [187]
coated capillaries CE-MS2
CE-DAD
Determination of fluoxetine and its 7:3 methanol–acetonitrile containing 15 mM NACE [186]
main metabolite norfluoxetine in ammonium acetate
human urine
Determination of heroin, basic CelixirTM Buffer capillary coating system FSCE [188]
impurities, and adulterants with plus added cyclodextrins
capillary electrophoresis
Determination of various Phosphate–borate buffer supplemented with MEKC [173]
benzylpenicillin salts SDS 14.4 g/L
Simultaneous determination of 0.81% pentane, 6.61% 1-butanol, 2% MEEKC [189]
ingredients in a cold medicine by 2-propanol, 4.47% SDS, 86.11% 10 mM
CD-MEEKC sodium tetraborate, 3 mM
2,6-di-O-methyl-beta-CD
Folic acid in tablets 0.5% ethyl acetate, 1.2% butanol, 0.6% MEEKC [170]
SDS, 15% propanol, 82.7% 10 mM
tetraborate (pH 9.2)
A method was developed and validated using a low pH buffer for the determination of ranitidine
and potential related impurities in bulk drug and formulations [194]. This method gave detection
limits of 0.03% diamine, 0.04% oxime, 0.1% Bis, and 0.24% nitroacetamide but importantly, it
detected a number of peaks, which were not resolved by either HPLC or TLC.
High pH buffers such as phosphate or borate are employed in the analysis of acidic components.
At high pH, acidic components migrate against the EOF, thus maximizing mobility differences.
A high pH borate buffer of 9.2 was developed and validated for the determination of homotaurine
as an impurity in calcium acamprosate by FSCE, detection limits of between 0.01% and 0.15%
homotaurine were reported [195]. A high pH CE method was used to quantitatively profile the
chloromethylated, monomethylated, and hydroxylated impurities of a new compound (LAS 35917).
The CE method allowed the quantification of the impurities at levels of 0.04–0.08% of the parent
drug, while HPLC failed to resolve the impurities [196].
MEKC and MEEKC have the unique capability to separate both charged and neutral substances
and have been employed for the quantification of a number of pharmaceutical compounds and related
impurities. A MEKC method was developed and validated for the determination of didanosine, an
anti-HIV treatment, which was separated from 13 of its potential impurities [197]. A simple, fast, and
selective MEKC method was used for the determination of ketorolac tromethamine and its known
impurities in 6 min and gave a quantification limit of 0.1% for impurities [198]. MEEKC is similar
to MEKC in its ability to separate and quantify charged and neutral compounds (see Sections 3.3
and 3.4) and has been used to a lesser extent for impurity studies than MEKC, mainly due to the
relatively recent introduction of this mode of CE. Future impurity profiling should see an increased
utilization of MEEKC methodology. Further to the determination of ketorolac and its impurities
using MEKC, Furlanetto et al. [199] used a mixture design in the optimization of a MEEKC method
for the same analysis, complete resolution of the analytes was obtained in 3 min. Wen et al. [200]
utilized MEEKC to simultaneously separate 17 species of heroin, amphetamine, and their basic
impurities and adulterants in under 10 min. Readers are referred to a review by Hilhorst et al.
[204] covering impurity profiling of drugs by CE. Some application-based papers will also provide
the reader with more information on profiling by CE-MS [201,202] and CE-LIFD [203] (see also
Table 4.4).
4.6.4 PHYSICOCHEMICAL PROFILING

During the early phase and later phase of drug development, knowledge of physiochemical properties
of pharmaceutical compounds is important in order to predict their bioavailability and blood-brain
barrier distribution to help in formulation design and drug delivery. Physicochemical properties
such as acid dissociation constant (pKa ), octanol–water partition-coefficient (log Pow ), solubility,
permeability, and protein binding are closely related to drug absorption, distribution, metabolism,
and excretion. About 30% of drug candidate molecules are rejected due to pharmacokinetic-related
failures [220]. When poor pharmaceutical properties are discovered in development, the costs of
bringing a potent but poorly absorbable molecule to a product stage can become very high. Fast and
reliable in vitro prediction strategies are needed to filter out problematic molecules at the earliest
stages of discovery. In drug discovery, there can be a vast number of compounds requiring physic-
ochemical screening mainly because of the high volumes of syntheses that can be carried out by
combinatorial chemistry. Consequently, there is a need for rapid and reliable methods of physico-
chemical profiling to maintain high throughput and efficiency. CE is a simple, versatile, automated,
and powerful separation technique and widely applied in physicochemical profiling for pharmaceu-
ticals. It has advantages over traditional potentiometric, spectrophotometric, chromatographic, and
other methods, as CE requires very small amounts of sample and can measure compounds with
impurities and low aqueous solubility. The advent of multiplexed CE instruments for physicochemi-
cal measurements can allow for the simultaneous analysis of up to 96 separate compounds using just
TABLE 4.4
Pharmaceutical Impurity Applications
Heroin and its basic impurities 100 mM DM-β-CD in Celixir reagent B FSCE [188]
(pH 2.5);
100 mM HP-β-CD in Celixir reagent B
(pH 2.5);
103.2 mM SDS, 50 mM phosphate–borate
(pH 6.5)
Separation of penicillin V and its SDS 20.0 g/L (69 mM) and pentanesulfonic acid MEKC [209]
impurities sodium salt 2.2 g/L adjusted to pH 6.3
Ranitidine hydrochloride and related 190 mM trisodium citrate (pH 2.6) FSCE [194]
substances
Aminopyridines and related 50 mM phosphate buffer (pH 2.5) FSCE [210]
substances
Homotaurine as an impurity in 40 mM borate (pH 9.2) FSCE [195]
calcium acamprosate
New substance (LAS 35917) 60 mM tetraborate (pH 9.2) FSCE [196]
5-Aminosalicylic acid and its major 120 mM CAPS buffer (pH 10.2), 65 mM SDS, MEKC [211]
impurities 55 mM TBAB, 5% MeOH
Ketorolac tromethamine and its 13 mM boric acid, phosphoric acid (pH 9.1) MEKC [198,199]
known related impurities with 1 M sodium hydroxide, 73 mM (SDS), MEEKC
90.0% 10 mM borate, 2.0% n-heptane, 8.0%
SDS/n-butanol in 1:2 ratio
Vancomycin and related impurities 120 mM Tris-phosphate buffer (pH 5.2) MEKC [212]
containing 50 mM CTAC
Ximelagatran thrombin inhibitor and Phosphate buffer (pH 1.9), 22% (v/v) MeCN, FSCE [213]
related substances in drug substance 11 mM hydroxypropyl β-CD
and tablet formulation
Penicillin and related impurities 20 mM ammonium acetate (pH 6.5) FSCE and [214]
In aqueous and nonaqueous electrolyte NACE
N-acetylcysteine and its impurities 100 mM borate (pH 8.40) FSCE [215]
Ciprofloxacin and its impurities Phosphate buffer (pH 6.0), 0.075 M FSCE [216]
pentane–1-sulfonic acid Na salt
Metacycline and its related 160 mM sodium carbonate + 1 mM EDTA FSCE [217]
substances (pH 10.35), 13% v/v MeOH
Loratadine and related impurities 100 mM H3 PO4 (pH 2.5), 10% acetonitrile FSCE [218]
Rofecoxib and photodegradation 25 mM borate, 15 mM SDS, 10% acetonitrile MEKC [219]
impurities
Galantamine impurities by CE-MS 50:25:25 (v/v/v) 100 mM ammonium CE-MS [202]
acetate/acetonitrile/methanol
Abbreviations: CAPS, 3-(Cyclohexylamino)-1-propanesulfonic acid; CTAC, cetyltrimethylammonium chloride; EDTA,

ethylenediaminetetraacetic acid; TBAB, tetrabutylammonium bromide.
one instrument [76–79], which can offer huge savings in time costs compared to the more traditional
methods mentioned previously. In a review paper by Jia [221], the principles and applications of CE
in profiling various physicochemical properties are covered.
Of the physicochemical properties mentioned previously in this section, only capillary elec-
trophoretic measurements of pKa and log Pow values will be covered here as they have received the
most attention in physicochemical profiling by CE.
Me × 10–4 (cm2/V.s)
2
0
4 6 8 10 12 14
pH
FIGURE 4.6 Dependence of the effective mobilities of a monovalent compound (LY334370) on pH. Arrow
indicates the pH equal to the pKa . (From Caliaro G.A., Herbots C.A., J. Pharm. Biomed. Anal., 26, 427–434,
2001.)
4.6.4.1 pKa Measurements

The majority of pharmaceutical compounds are either acidic or basic and are therefore ionizable.
Other physicochemical properties, such as lipophilicity and solubility, are pKa dependent; therefore,
pKa is one of the fundamental parameters of a drug molecule. The pKa determination of acids and
bases by CE is based on measuring the electrophoretic mobility of charged species associated with
the acid–base equilibria as a function of pH. The detailed theory of pKa and log Pow measurements
is beyond the scope of this chapter and readers are referred to a review by Jia [221], which describes
in more detail the theoretical aspects of physicochemical profiling by CE methods including FSCE,
NACE, MEKC, and MEEKC.
The pKa values of pharmaceutical compounds can be determined from migration time data
obtained by running the compound with free solution CE electrolytes at a range of pH values. The
mobility of the solute at each pH can be calculated from its migration time and the EOF (measured
against a neutral marker such as methanol), and a plot of mobility versus pH can be constructed (see
Figure 4.6). The pKa value can be calculated mathematically or obtained from the plot.
The pKa values of water-insoluble and sparingly soluble compounds can be determined by NACE
using methanol as the background electrolyte [48], or a 50% methanol/water electrolyte as used by
de Nogales et al. [226] for the measurements of acid dissociation constants of several hydrophobic
drugs. The most common method for pKa measurements, however, is FSCE (see Table 4.5).
CE instruments are highly automated and can be used for high-throughput applications, partic-
ularly with the use of capillary array instruments. CE compares favorably to other methods of pKa
measurements [48] and unlike titration methods, precise information of sample concentration is not
required as only analyte mobilities are used to calculate dissociation constants by CE. Sparingly
soluble compounds are easily analyzed and only very small amounts of material are required, which
is useful for screening of newly synthesized compounds where only small quantities may exist or
where the molecules are present in only very small amounts. By using a “sample stacking” tech-
nique along with pressure assistance and short-end injection for rapid analysis, Wan et al. [224] were
able to successfully measure the pKa values of pharmaceutical compounds with concentrations as
low as 2 µM. Techniques commonly used for pKa measurement such as potentiometric or UV-Vis
spectroscopy do not differentiate between the analyte of interest and any other degradant or impurity
present, which can cause problems if the analyte is not highly pure or unstable in solution. Because
CE is a separating technique, it can measure the pKa values of impure or unstable compounds where
electrolyte purity is not essential. This was demonstrated by Ornskov et al. [225] where a CE method
of pKa measurement was used successfully for a set of labile drug compounds that were unsta-
ble in solution. Determining the EOF intensity during pKa measurements may be time consuming,
TABLE 4.5
Selected CE Applications for pK a Measurements
Application Buffer References
pKa * values of 21 basic drugs with aliphatic or pH* range 4.9–9.7 [48]
aromatic amino groups Methanol and sodium acetate buffer + acetic acid
Medium-throughput pKa of 48 pharmaceutical pH range 2.5–11.0 [227]
compounds—acidic, basic, and multivalent Electrolytes of 0.1 M ionic strength composed of
phosphate, acetate, and borate buffers
adjusted to required pH with phosphoric,
acetic, or boric acid or NaOH
Pressure-assisted CE at 2 psi
Rapid pKa screening of 26 acidic, basic, and pH range 2.5–11.0 [224]
multivalent pharmaceuticals Electrolytes of 0.05 M ionic strength composed
of 0.5 M phosphate and 1 M acetate buffers
mixed to obtain required pH
PACE at 25 mbar, “short-end” injection
pKa determination of labile drug compounds PH range 2.0–12.0 [225]
Electrolytes of 0.05 M ionic strength composed
of 1 M phosphate, 0.1 M borate, and 1 M
acetate buffers mixed and adjusted with
phosphoric acid, acetic acid, and NaOH to
obtain required pH
“Short-end” injection
2-Amino-2-oxazolines (antihypertensive agents) pH 4.77–9.69 [228]
Cephalosporins (antibiotics) pH 2.0–9.0 [229]
pKa determination of 99m Technetium pH range 1.3–6.6 [230]
radiopharmaceuticals 50 mbar PACE
Citric acid adjusted with NaOH
Dissociation constants of anthrocyclines Phosphoric acid, disodium phosphate, and [231]
(antibiotics) monosodium phosphate (pH 4.20–8.20)
Dissociation constants of cytokinins Phosphate and acetate buffers of ionic strength [232]
(phytohormones) 0.015 M (pH 1.5–6.0)
Determination of pKa values Dihydrofolate pH 2.1–4.5 [233]
reductase inhibitors 50 mM phosphate buffer adjusted with
phosphoric acid/NaOH
Quinolones (antibacterial) pH 2.0–11.0 [234]
Dissociation constants of amino- and pH 3.50–11.25, ionic strength (25 mM) [235]
guanidinopurine nucleotide analogs and related
compounds by CZE
Determination of pKa values with dynamically Dynamic capillary coating [236]
coated capillaries
pKa of organic bases in aq. acetonitrile pH range 4.76–9.5 [237]
(0–70%, v/v) Tris, ethanolamine and acetate buffers
pKa of N-imidazole derivative aromatase pH range 3.88–9.16 [238]
inhibitors 25 mM phosphate buffer adjusted with
triethylamine
especially at a low pH. Geiser et al. [236] overcame this drawback by using a dynamic capillary
coating procedure to increase the EOF and thus to reduce the analysis time. In addition, this coating
procedure enhanced migration time stability. Table 4.5 lists selected recent applications of CE to the
measurement of dissociation constants.
4.6.4.2 Log Pow Measurements

Hydrophobic interaction (or liquid–liquid partitioning) of pharmaceutical compounds in the body
plays a significant role in partitioning of drugs into lipid bilayers of biomembranes, bioavailabil-
ity, and pharmacokinetics. As with pKa values, liquid–liquid partition coefficient measurements are
extremely important during drug discovery, screening, and formulation processes. Solute hydropho-
bicity is usually expressed by the octanol–water partition coefficient (log Pow ) that is defined as the
ratio of the concentrations of a species in the two phases at equilibrium. A number of methods to
measure log Pow are available including the shake flask method, potentiometric titration, and liquid
chromatographic separation methods [241] and extensive data collections of log Pow values can
be found in literature [239,240]. CE techniques using pseudostationary phases in the background
electrolyte, that is, MEKC and MEEKC, allow the measurement of log Pow values because of the
partitioning of solutes between the MEKC micelle and microemulsion droplet (MEEKC). Early work
using MEKC showed that the extent of solutes partitioning with micelles was related to the solutes
solubility [242,243], the rise in popularity of MEEKC, however, demonstrated its applicability to a
wider range of solutes and this technique has been used extensively in log Pow measurements. Exam-
ples of using MEEKC for log Pow measurements are cited [78,244,245]. MEEKC review papers by
Hansen [62], Marsh et al. [63], and Huie [64] comprehensively cover the many applications in this
area of physicochemical profiling.
Using MEEKC, the compounds solubility is assessed by bracketing it with neutral marker com-
pounds of known log Pow values, which are used to create a calibration graph of log Pow against time
or log k (log of the retention factor) (see Figure 4.7). The log Pow of the analyte of interest can be
calculated by its migration time or retention factor using the graph. The higher the compounds log
Pow value, the more it partitions into the microemulsion droplet and the longer it takes to migrate.
Table 4.6 contains a number of selected CE applications for log Pow measurements, the interested
reader should consult recent reviews [62–64], which contain a comprehensive collection of all recent
physicochemical measurements by CE.
4.6.5 ANALYSIS OF SMALL MOLECULES AND IONS (PHARMACEUTICAL)

The separation and detection of small organic and inorganic ions is an important activity in the
pharmaceutical industry. CE is routinely used for ion analysis in the pharmaceutical industry for a
number of applications, these include counterion determination, stoichiometry, salts, and excipients
in drug formulations. Most pharmaceutical molecules are charged and are commonly manufactured
3
log P
0
–1 0 1 2 3
log k
FIGURE 4.7 Plot of log P against the retention factor (log k) from MEEKC on a dynamically coated capillary
column for compounds of varied structure. (From Poole S.K., et al., J. Chromatogr. B, 793, 265–274, 2003.)
TABLE 4.6
Selected Log P Measurements Using MEKC and MEEKC
Neutral and weakly acidic compounds SDS (1.4% w/v), n-butanol (8% v/v) and MEEKC [246]
using dynamically coated capillary n-heptane (1.2% v/v) in 90 mL acidic buffer
columns
Determination of octanol–water partition 2.16% SDS (w/w), 0.82% heptane (w/w), MEEKC [245]
coefficients of 80 pesticides 6.49% 1-butanol (w/w) in 0.05 M sodium
phosphate–0.1 M borate at pH 7
Estimation of octanol–water partition Heptane/SDS/butanol in phosphate buffer at MEEKC [247]
coefficients of anticancer pH 7.4
platinum(II)-complexes
Log Pow measurement using 6.61% (w/v) 1-butanol, 0.81% (w/v) MEEKC [248]
multiplexed MEEKC n-heptane, 3.31% (w/v) SDS, 800 mL of 68
mM CAPS buffer (pH 10.3)
Neutral and basic compounds by 2% w/v SDS, 1.2% v/v n-heptane, 8% v/v MEEKC [249]
microchip MEEKC 1-butanol, 50 mM CAPS (pH 10.4)
Neutral pharmaceuticals 25 mM borate buffer (pH 8.5), 20–150 mM MEKC [242]
SDS
Neutral pharmaceuticals and steroids 0.05 M phosphate, pH 7 (pH 9 for steroids), MEKC [250]
80 mM sodium cholate, 100 mM SDS or
100 mM CTAB
Determination of drug in phospholipids 50 mM phosphate buffer (pH 7.4) containing MEKC [251]
different concentrations of sodium cholate
C (50, 60, 80, 125 mM)
with a counterion, commonly a metal cation for acidic drugs and an ionic salt or small organic acid for
basic drugs. During development of a new drug, a range of different counterions may be synthesized
to compare pharmaceutical properties such as solubility, stability, and crystallinity of the different
salts. The ratio of the drug to counterion is known as the drug stoichiometry and this needs to be
characterized analytically. The typical stoichiometry is a 1:1 drug–counterion mixture; however,
frequently 2:1 and 1:2 compositions are manufactured depending on the ionic nature of the drug
and/or counterion. There is a clear analytical need to quantify drug counterion levels to demonstrate
that the correct salt version has been manufactured and that the required stoichiometry can be reliably
achieved batch-to-batch when the final drug salt has been selected. The counterion of basic drugs
includes inorganics such as sulfate and chloride or organic acids such as maleate, fumarate, acetate,
or succinate. Cations analyzed involve a range of metal ions including Na+ , K+ , Mg2+ , Ca2+ ,
and simple low molecular weight amines. These analytes possess little or no chromophore, which
generally necessitates use of indirect UV detection. However, some larger anionic counterions such
as benzoates and simple organic acids can possess sufficient UV activity to allow direct UV detection.
Alternatively, metal ions may be complexed “on capillary” to form metal chelates that can then be
detected by direct UV measurement. Alternative detection methods such as conductivity detection
have also been used [268] to detect potassium counterion and other inorganic cationic impurities in
pharmaceutical drug substances.
Popular techniques for the analysis of small ions include ion-exchange chromatography and
flame atomic absorption spectrometry but CE is becoming more popular for such applications due to
its simplicity and speed of method development and analysis, elimination of the need for specialized
columns, high resolving power, and simple sample treatment steps (typically, the sample just needs
to be diluted in the background electrolyte and injected onto the capillary). Commercial ion analysis
3 11,12
2
13
Absorbance (a.u.) 6.8

16 19
4 5 18
1
9 71410
17
15
2 4 6 8 10 12 14 16 18 20
Min
FIGURE 4.8 Electropherogram of 20 common amino acids. 50 mM Ethanesulfonic acid, pH 2.8; applied
voltage, 30 kV; injection time, 10 s. Peaks: 1 = Lys; 2 = Arg; 3 = His; 4 = Gly; 5 = Ala; 6 = Val; 7 = Ser;
8 = Ile; 9 = Leu; 10 = Thr; 11 = Asn; 12 = Met; 13 = Gln; 14 = Trp; 15 = Glu; 16 = Phe; 17 = Pro; 18 =
Tyr; 19 = Cys. (From Fritz J.S., J. Chromatogr. A, 884, 261–275, 2000.)
kits are available that contain predefined method conditions, reagents and buffers that simplify
method development and allow analysis times of 2–10 min, comparing favorably to ion-exchange
chromatography methods.
Pharmaceutical excipients such as SDS or alginic acid can be analyzed as raw materials or when
present in formulations [263]. Kelly et al. [252] developed a reliable quantitative CE method for the
determination of SDS in a cefuroxime axetil pharmaceutical preparation. Separation of amino acids
can be troublesome and complicated using HPLC methods as they first need to be derivatized to
provide a chromophore for detection. Performing these separations using CE and low pH electrolytes,
however, can be relatively simplistic as lower detection wavelengths of 185 nm can be used to detect
zwitterionic amino acids, which become cations at low pH. A background electrolyte of 50 mM
ethanesulfonic acid (pH 2.8) was used to resolve a number of amino acids, which were detected at
185 nm without any sample pretreatment [253], as shown in Figure 4.8.
Simple organic acids often possess chromophores, which makes direct UV detection possible,
even at low wavelengths. CE can measure the presence of small ion contaminant impurities in
drug substances. For example, an NACE method with indirect UV detection was used to monitor
ammonium ion contaminant in pharmaceutical preparations with a limit of detection of 50 ppb
[254]. A number of reviews have been published that will provide the reader with a comprehensive
coverage of the status of CE for the analysis of ions and small molecules, including detection
methods, quantification, and stoichiometric determinations [8,9,11,253,255,256]. Table 4.7 shows
some recent analysis applications of CE for ion analysis.
4.7 METHOD DEVELOPMENT AND VALIDATION

As with any other analytical technique developing a CE method involves some basic steps. As a
general guide to method development, the following items should be considered: the objectives
of the intended separation, instrumental requirements, sample characteristics, sample pretreatment,
data handling, and reporting. The objectives may be straightforward and require the detection and
quantification of a single charged compound of which there is much information available and
separation can be achieved using simple aqueous buffers. They may be more complex and require
the separation, identification, and quantification of a number of charged and neutral components,
TABLE 4.7
Selected CE Applications for the Analysis of Small Molecules and Ions
(Pharmaceutical)
Application Electrolyte References
Anions—indirect detection
SDS levels in tablets 3–10 mM barbital buffer, pH ∼9 [252]
Determination of Br, Cl, and SO4 as 10 mM potassium chromate + 1 mM borate, [257]
impurities in calcium acamprosate pH 9.15
Acetate counterion in an antifungal 4.0 mM 4-hydroxybenzoic acid, [258]
lipopeptide myristyltrimethylammonium bromide, as
an EOF modifier
Sodium acetate in antisense oligonucleotides Phthalate, myristyltrimethylammonium [259]
bromide type EOF modifier
Drug inorganic counter ion determination Chromate, [260]
tetradecyltrimethylammoniumbromide
(TTAB)
Drug organic acid counter ion determination Phthalate, 4-morpholineethanesulfonic acid [261]
(MES),
tetradecyltrimethylammoniumbromide
(TTAB)
Anion screening for drugs and intermediates Chromate, myristyltrimethylammonium [262]
bromide type EOF modifier
Anions—direct detection
Quantitative determination of alginic acid in 12.5 mM borate/boric acid buffer (pH 8.3) [263]
pharmaceutical formulations
Determination of residual Br in excess of 60:40 MeCN:methanesulfonic acid buffer [264]
chloride for local anaesthetic analysis (pH 1.3)
Arginine counterion determination for 10:90 ACN:(2% SB-β-CD and 0.7% [172]
ragaglitazone DM-β-CD in 25 mM phosphate buffer,
pH 8.0)
Cations—indirect detection
Ammonium ion impurities NACE method, imidazole [254]
Ca, Li, K, and Na counterions of 4-Aminopyridine buffer (pH 9) [265]
glycosaminoglycans
Ca in calcium acamprosate drug substance 10 mM imidazole containing 1 mM [266]
tetrabutylammonium sulfate (pH 4.5)
Sodium in drug substance Imidazole, formic acid [267]
Quarternary amine residues in drug Quinine, THF [267]
substance
Piperazine in pharmaceutical drug 50 mM benzylamine (pH 8.7) [269]
substances
Cations—direct detection
K counterion and inorganic cationic 30 mM creatinine, 30 mM acetic acid, [268]
impurities of pharmaceutical drug 4.5 mM 18-Crown-6
substances by conductivity detection
which may be only possible using MEKC or MEEKC. The compound of interest may be present in
a number of isomeric forms, which requires chiral selectors added to the background electrolyte or
may be present in trace quantities as degradation products or impurities, which require separation
and quantification with high sensitivity. Other items to be considered when deciding the objectives
are as follows: do all of the sample components need to be separated/quantified/identified and what
is the sample matrix. Identifying these goals helps to determine the instrumental requirements.
For example, if qualitative analysis is required then a DAD or MS detector will be required. To
achieve lower limits of detection, LIFD or UV sensitivity enhancement techniques may be used.
The characteristics of the sample and sample matrix are helpful in selecting the mode of CE. If
the sample and/or sample matrix is water insoluble, then NACE, MEEKC may be the CE mode of
choice. Data handling depends on the type of the intended analysis, quantitative or qualitative and
whether internal or external standardization is used for quantification. If the analyst possesses enough
knowledge and experience of CE separations, he/she may be able to decide on initial CE conditions
and optimize the separation from there. This, however, is not always the case particularly as the in-
depth knowledge and experience of pharmaceutical analysis by CE is not as widely established as it is
for other separation methods such as HPLC, TLC, and GC. Once the analyst identifies the separation
objectives, a literature search of published reports to find sources of developed and/or validated
methods for the same or similar separations should be carried out. A variety of literature sources
are available. Sources of reviews and publications for each application area, CE mode and detection
method described in this chapter are cited toward the end of each individual section to direct the
reader to relevant literature sources. A number of tables of selected applications are also presented.
Lunn [13] provides detailed method information for several hundred applications. In addition to
searching published application literature, a number of commercial CE instrument and consumable
manufacturers and suppliers offer generic and sample methods for a range of applications. Beckman
Coulter, for example, has published a generic method development strategy for chiral CE, previously
published in LCGC Europe [283].
Developing a method involves selecting and optimizing a number of variables
• Mode of CE: as discussed previously the mode of CE must suit the separation.
• Capillary: fused-silica capillaries are the standard type used. Capillary length will affect
the speed of separations. Small internal diameter capillaries allow higher electric field
strengths to be applied resulting in more efficient separations. Typically, capillaries of
50 µm inner diameter and 0.5–1 m in length are used.
• Capillary conditions: these are affected by the capillary pretreatment reagents (e.g.,
NaOH) or dynamic capillary coating systems to achieve better reproducibility.
• Applied voltage and operating current.
• Injection mode and volume.
• Capillary temperature.
• Detector: as discussed previously, the detector type is governed by quantitative/qualitative
and suitability requirements.
• Buffer composition: this is one of the most important variables to consider, as the type of
compounds to be analyzed will dictate the pH, polarity, concentration, additives, and so
on of the buffer.
A number of books and reviews provide a more detailed insight and coverage of the operating
parameters of CE and their effects in method development, such as those mentioned above. This
chapter does not have the scope to cover these in detail and readers are encouraged to consult the
publications [13,14,34–37].
If there are no previous published methods for a particular pharmaceutical compound or class
of compound then evaluation of some of the properties and functionalities must be considered to
determine the initial method conditions. If pKa data are available for the solute of interest and other
species that need to be resolved then an appropriate electrolyte pH can be selected. If the compound
is present as a racemic mixture then selection of an appropriate chiral selector can be made based
on the compounds functional groups and other chiral separation applications. For basic compounds,
a 50 mM phosphate (Na2 HPO4 ) buffer adjusted to pH 2.5 with phosphoric acid, in a standard
capillary with an applied voltage of +20 kV and UV detection at 200 nm is a good starting point.
Samples should be dissolved in an appropriate solvent at a concentration of ∼1 mg ml−1 . This set

of operating conditions is appropriate for a wide range of basic compounds. Additional optimization
can be achieved using additives such as CDs, ion-pairing reagents, and organic solvents.
For the separation of acidic compounds, the buffer can be replaced with a 15 mM borate solution
with a natural pH ∼9.3. The same operating parameters as those used for the separation of basic
compounds can be used to separate a wide range of acidic drugs and buffer additives similar to those
used for basic separations can be used.
For the separation of neutral compounds using MEKC, 15 mM Na2 HPO4 containing 50 mM
SDS as the micellar electrolyte can be employed. Using different surfactants such as bile salts can
help if the solutes are particularly water insoluble. Optimization can be carried out by adjusting
the surfactant concentration and type, adding organic solvents, ion-pairing reagents, CDs, and so
forth. Reviews by Pappas et al. [52], Molina and Silva [53], and Pyell [54] discuss MEKC operating
parameters and applications in detail.
MEEKC offers more selectivity options than MEKC and can be used to separate a wider range of
solutes. Owing to the composition of the microemulsion electrolytes, MEEKC can be used for more
complex separations of water-soluble, water-insoluble, charged and neutral compounds, particularly
where water-insoluble matrices are involved. A standard O/W microemulsion composed of 3.3%
SDS, 0.8% octane, 6.6% butanol in 10 mM borate with fused-silica capillaries are used for most
applications and this is a good starting point for any intended MEEKC separation. With a range
of buffer additives available such as organic modifiers, CDs, ion-pairing reagents along with the
various microemulsion parameters, MEEKC offers a wider selection of optimization choices than
any other form of CE for pharmaceutical analysis. The background theory and operating parameters
of MEEKC have been reviewed by Altria [58,270,271] and Klampfl [272], which detail the effects
of each microemulsion and instrument variable on separation selectivity.
4.7.1 METHOD OPTIMIZATION USING EXPERIMENTAL DESIGN

Although the approach to CE method development is similar to HPLC methods, the number of vari-
ables involved can sometimes make the optimization of separation conditions a difficult task. This is
particularly true when dealing with MEKC and MEEKC where pH, nature, and concentration of the
surfactant, buffer, and organic modifiers combined with instrument variables such as applied voltage,
temperature, and injection volume all have a great influence on solute separation. Ideally, a method
should be optimized to find the best separation conditions in as short a time as possible with the
minimum number of experiments. Statistical experimental design is a powerful tool to quantify the
effect of one or more variables on a set of measured responses. It provides a methodological frame-
work for changing operating parameters simultaneously by the help of experimental designs. These
approaches involve the smallest possible number of useful experiments and provide maximum infor-
mation. The use of experimental design strategies for CE methods involving univariate (studying one
factor at a time) and multivariate (studying multiple factors simultaneously) techniques has evolved
as a rapid method development and optimization option. Multivariate techniques are probably the
least time consuming and are recommended. Multivariate (chemometric) experimental designs have
been reviewed by Altria et al. [273] and Sentellas and Saurina [274,275]. Veuthey and Rudaz [276]
have also covered the use of statistical and chemometric tools for pharmaceutical analysis. The use
of artificial neural networks (ANNs) in separation science has also been reviewed [277]. Pyell [51]
includes a detailed chapter containing many examples of experimental work (MEKC and MEEKC)
using experimental design.
4.7.2 METHOD VALIDATION GUIDELINES FOR PHARMACEUTICAL APPLICATIONS

Validation of pharmaceutical CE methods requires similar considerations to those examined during
HPLC validation procedures. A useful guide for method validation in analytical chemistry is the
Eurachem guide [285], which discusses when, why, and how methods should be validated. However,
for the pharmaceutical industry, the main reference source is the ICH Guidelines [286], which
provides recommendations on the various characteristics to be tested for the most common types
of analytical procedures developed in a pharmaceutical laboratory. The main characteristics of any
analytical method to be tested are specificity, linearity, accuracy, precision, solution stability, limits of
detection and quantification, and robustness. Specific aspects should be considered for a CE method
including method transfer between instrument manufacturers, reagent purity and source, electrolyte
stability, capillary treatment and variations in new capillaries, and buffer depletion. Fabre and Altria
[284] discuss CE method validation in more detail and include a number of examples of validated CE
methods for pharmaceutical analysis. Included in Table 4.3 are a number of validated pharmaceutical
assay methods.
4.8 COMPARISON OF HPLC, GC, AND CE FOR PHARMACEUTICAL

ANALYSIS
As HPLC is the most widely used pharmaceutical analysis technique among current chromatographic
methods, a short comparison of HPLC and CE should be very informative to the reader. Some
comparisons will also be made to GC. This section will illustrate the advantages, disadvantages, and
complementary natures of HPLC and CE.
The range of CE applications in pharmaceutical analysis is at least as extensive as that of HPLC.
Adopting CE testing provides several distinct advantages, including faster analysis and method
development, lower consumable expenses, and easier operation. All of these factors are important in
pharmaceutical separations where high throughput is becoming ever important. The disadvantages
of CE are not to be forgotten; they include poorer injection precision (hence the need to incorporate
internal standards) and the limited number of staff members who are trained and experienced in using
CE compared with the number who are competent with HPLC. The highly complementary nature of
the two techniques is a valuable tool for the pharmaceutical analyst where “cross-validation” using
the two methods can give a high degree of confidence in results.
4.8.1 EFFICIENCY
As solute peaks in HPLC and GC chromatograms are integrated and measured in the same way as
peaks in CE electropherograms, direct comparisons of peak efficiency for each method is possible.
Typical CE capillary dimensions are 30–100 cm long (although when using “short-end injection”
the effective length can be reduced further) with internal diameters of between 50 and 100 µm and
can generate ∼400,000 theoretical plates, compared to ∼150,000 plates for a typical wall coated
open tubular GC column 0.18 mm internal diameter and 20 m long. HPLC, however, can generate
at best 30,000 plates for a typical 150 mm C18 column. The main reason for the higher efficiency
possible in CE compared to HPLC is that the solutes move through the column with a “plug” or
flat flow profile, while they move through HPLC columns with a “laminar flow” profile. The other
reasons for higher plate numbers in CE are the absence of multipath and longitudinal diffusion and
slow equilibration of the solute in the capillary. Turbulent flow, often associated with GC columns,
is also absent from CE capillaries due to the electrodriven nature of the EOF.
4.8.2 SAMPLE TYPES

Similar to HPLC, CE is well suited for the analysis of both polar and nonpolar compounds but unlike
HPLC columns, the same capillary can be used for most CE applications, that is, chiral separations,
small ions, and so forth. Sample preparation for CE analyses is straightforward and usually requires
sample dissolution in the buffer or water, sometimes with the addition of a small volume of organic
solvent to help dissolve the solute. CE buffers can be used over a wider pH range than HPLC mobile
phases, as most HPLC columns can only be used within a specified pH limits without damaging the
column.
4.8.3 SAMPLE VOLUME

Relatively small sample volumes are required for CE analyses compared to HPLC and GC, only a
few nanoliters are injected compared to microliters used in GC and HPLC. This is due to the small
volumes of the capillaries, typically 0.98 µl for a 50 cm × 50 µm ID capillary. This is advantageous
when only limited amounts of the sample is available.
4.8.4 SENSITIVITY
As mentioned in Section 4, detection limits in CE are generally poorer than HPLC or GC due to
the narrow detection path length for UV/Vis detection and the nanoliter injection volumes. Various
approaches to improve CE sensitivity such as the use of LIFD detectors, low wavelength detection,
and so forth have been covered in Section 4.4. Some comparisons of CE and HPLC sensitivity for
pharmaceutical analysis are illustrated in Table 4.8.
4.8.5 PRECISION
Because of the small injection volumes used in CE, peak area reproducibility is poorer than HPLC.
The use of high sample concentrations and internal standards makes CE precision values more com-
parable to HPLC. The precision of solute migration times can also suffer because of variations within
the capillary between injections, affecting the EOF. This can be alleviated by careful preparation of
buffers and samples and the use of dynamic capillary coating systems.
TABLE 4.8
Comparisons of Recent HPLC and CE Applications
Analysis Precision
Application Method Time LOD RSD% References
Determination of four HPLC 21 min 0.02–0.05 0.41–1.14 [279]
parabens in cosmetics CE 16 min µg ml−1 0.77–2.17
0.16–0.21
µg ml−1
Determination of HPLC 7 min n/a 0.9–1.7 [287]
glycyrrhizinic acid in CE 15.5 min n/a 3.2–4.0
pharmaceutical
preparations
Stereoselective analysis HPLC 27 min 1 ng ml−1 n/a [280]
of carvedilol in serum CE 15 min 50 ng ml−1 n/a
Determination of HPLC 3 min HPLC = 2.5 times 1.91 [281]
ketoconazole in drug CE 5 min less than CE 2.95
formulations
Determination of HPLC 4.9 min 1.44 µg ml−1 0.98 [282]
diazepam in CE 4.1 min 4.24 µg ml−1 1.62
pharmaceutical tablets
4.8.6 ANALYSIS TIMES

Typical analysis times for both HPLC and CE are between 5 and 30 min. With CE, however, a number
of options are available to shorten analysis times; pressure-assisted CE where an external pressure is
applied to the capillary in conjunction with the EOF helps speed up separations. Short-end injection,
where the sample is introduced into the detector end of the capillary can also reduce analysis times,
these methods were demonstrated by Mahuzier et al. [278]. A shorter capillary can also be used to
achieve shorter analysis times. As discussed in Section 4.3.6, the use of multiplexed or capillary array
CE instruments can provide the pharmaceutical laboratory with the capability to conduct multiple
analyses simultaneously, offering a significant advantage in terms of time and cost savings. This is
particularly advantageous where the screening of large numbers of new pharmaceutical compounds
is required due to combinatorial organic synthesis in drug discovery.
4.8.7 REAGENTS AND CONSUMABLES

This area provides CE with many advantages over HPLC methods for a number of reasons. Owing
to the applicability of the same CE instrumentation and capillaries to a wide range of applications
only the buffer composition needs to be changed, whereas a variety of normal and reversed-phase,
chiral and ion-exchange, and so forth HPLC columns are available with a wide range of packing
material for different classes of compounds and separations. CE capillaries are also only a fraction of
the cost of HPLC columns, particularly chiral columns, which can cost in excess of 1200, a standard
fused-silica capillary that costs a few euro can be used with milligrams of a chiral selector to perform
the same separation. Relatively small amounts of buffers and reagents are used in CE, milliliters
compared to liters for HPLC mobile phases. This offers significant cost savings for both the purchase
and disposal of expensive chemicals.
As discussed in Section 4.3.6, the use of multiplexed CE instruments can provide the pharmaceu-
tical laboratory with the capability to conduct multiple analyses simultaneously, offering a significant
advantage in terms of time and cost savings.
4.9 CONCLUSIONS
Capillary electrophoresis has been shown as a suitable analytical method for a wide range of
pharmaceutical applications, indeed for principal applications such as pharmaceutical assay, physic-
ochemical measurements, and chiral analysis CE can be superior to HPLC in terms of speed and
range of method development options, cost efficiency, speed of analysis, ease of use, selectivity,
peak efficiency, and the possibility of implementation of a single set of method conditions for the
analysis of several different samples. The various advantages of CE have been highlighted which
include low cost of consumables, speed of analysis, the use of standard instruments and capillaries for
most applications, high-throughput analysis using capillary array instruments, superior efficiency,
and low wavelength detection. The main disadvantages of CE are its poor sensitivity due to the nar-
row detection path length using UV/Vis detection methods and poor injection precision. Measures
to improve detection limits include more sensitive detection methods such as LIFD, while CE-MS
offers the possibility of highly sensitive qualitative analysis. Quantitative methods such as the use
of internal standards serve to improve injection precision to levels approaching that of HPLC.
The various modes of CE available for pharmaceutical analysis means that effectively all types of
drug compounds can be separated using standard instrumentation, that is, charged, neutral, polar and
nonpolar, large and small molecules, and counterions. MEKC and MEEKC have been shown to be
particularly suitable for the separation of complex mixtures of both charged and neutral compounds,
in particular MEEKC, which can be applied to a wider range of solutes of a broadly varying water
solubility using both W/O and O/W microemulsions as the carrier electrolyte.
The range of pharmaceutical applications of CE is very extensive with hundreds of reports pub-
lished and this chapter only serves to give a brief introduction to each application area with tabulated
examples of proven methods. Many validated methods are cited which indicate the usefulness of
CE in the analytical laboratory, indeed the use of CE in pharmaceutical laboratories is now very
widespread and regulatory authorities have begun to accept CE for a number of new drug product
submissions. CE identification tests in general monographs for drug substances have been included
in the British and European Pharmacopoeia.
CE is regularly used to provide complementary information to HPLC or other methods as well
as being a technique of choice for certain applications. As further research is carried out and the
use of CE becomes more widespread in pharmaceutical laboratories, the use of CE is certain to
become more popular and possibly take over from traditional analytical methods such as HPLC and
titrimetric techniques for routine testing of pharmaceuticals.
4.10 DEDICATION
To the memory of Johnny McEvoy, a great man who left us with many great memories.
REFERENCES
1. Harris D.C., Quantitative Chemical Analysis, 4th ed. (W. H. Freeman and Company, New York, 1995),
pp. 643, 713.
2. Jorgenson J.W., Lukacs K.D., Zone electrophoresis in open tubular glass capillaries. Anal. Chem., 53,
1289–1302 (1981).
3. Mikkers F.E.P., Everaerts F.M., Verheggen T.P.E.M., High-performance zone electrophoresis.
J. Chromatogr., 169, 11–20 (1979).
4. Terabe S., Otsuka K., Ando T., Electrokinetic chromatography with micellar solution and open tubular
capillary. Anal. Chem., 57, 834–841 (1985).
5. Terabe S., Electrokinetic chromatography: an interface between electrophoresis and chromatography.
Trends Anal. Chem., 8, 129–134 (1989).
6. Altria K.D., Quantitative aspects of the application of capillary electrophoresis to the analysis of
pharmaceuticals and drug related impurities. J. Chromatogr., 646, 245–257 (1993).
7. Holland L.A., Chetwyn N.P., Perkins M.D., Lunte S.M., Capillary electrophoresis in pharmaceutical
analysis. Pharm. Res., 14, 372–387 (1997).
8. Altria K.D., Kelly M.A., Clark B.J., Current applications in the analysis of pharmaceuticals by capillary
electrophoresis. I. Trends Analyt. Chem., 17, 204–214 (1998).
9. Altria K.D., Kelly M.A., Clark B.J., Current applications in the analysis of pharmaceuticals by capillary
electrophoresis. II. Trends Analyt. Chem., 17, 214–226 (1998).
10. Altria K.D., Overview of capillary electrophoresis and capillary electrochromatography. J. Chro-
matogr. A, 856, 443–463 (1999).
11. Altria K.D., Elder D., Overview of the status and applications of capillary electrophoresis to the
analysis of small molecules. J. Chromatogr. A, 1023, 1–14 (2004).
12. Natishan T.K., Recent progress in the analysis of pharmaceuticals by capillary electrophoresis. J. Liq.
Chromatogr. Relat. Technol., 28, 1115–1160 (2005).
13. Lunn G., Capillary Electrophoresis Methods for Pharmaceutical Analysis (John Wiley & Sons, New
York, 2000).
14. Landers J.P. (Ed.), Handbook of Capillary Electrophoresis, 2nd ed. (CRC Press, FL, 1997).
15. Altria K.D., Simpson C.F., Analysis of some pharmaceuticals by high voltage capillary zone
electrophoresis. J. Pharm. Biomed. Anal., 6, 801–807 (1988).
16. Ha P.T.T., Hoogmartens J., Van Schepdael A., Recent advances in pharmaceutical applications of chiral
capillary electrophoresis. J. Pharm. Biomed. Anal., 41, 1–11 (2006).
17. Scriba G.K.E., Selected fundamental aspects of chiral electromigration techniques and their application
to pharmaceutical and biomedical analysis. J. Pharm. Biomed. Anal., 27, 373–399 (2002).
18. Scriba G.K.E., Pharmaceutical and biomedical applications of chiral capillary electrophoresis and
capillary electrochromatography: an update. Electrophoresis, 24, 2409–2421 (2003).
19. Amini A., Recent developments in chiral capillary electrophoresis and applications of this technique
to pharmaceutical and biomedical analysis. Electrophoresis, 22, 3107–3130 (2001).
20. de Boer T., de Zeeuw R.A., de Jong G.J., Ensing K., Recent innovations in the use of charged cyclodex-
trins in capillary electrophoresis for chiral separations in pharmaceutical analysis. Electrophoresis, 21,
3220–3239 (2000).
21. Wang A.B., Fang Y.Z., Applications of capillary electrophoresis with electrochemical detection in
pharmaceutical and biomedical analyses. Electrophoresis, 21, 1281–1290 (2000).
22. Zemann A.J., Conductivity detection in capillary electrophoresis. Trends Analyt. Chem., 20, 346–354
(2001).
23. Lacroix M., Poinsot V., Fournier C., Couderc F., Laser-induced fluorescence detection schemes for
the analysis of proteins and peptides using capillary electrophoresis. Electrophoresis, 26, 2608–2601
(2005).
24. Smyth W.F., Recent applications of capillary electrophoresis-electrospray ionisation-mass spectrom-
etry in drug analysis. Electrophoresis, 27, 2051–2062 (2006).
25. Poole S.K., Patel S., Dehring K., Workman H., Poole C.F., Determination of acid dissociation constants
by capillary electrophoresis. J. Chromatogr. A, 1037, 445–454 (2004).
26. Hudson J., Golin M., Malcolm M., Whiting C., Capillary zone electrophoresis in a comprehensive
screen for drugs of forensic interest in whole blood. .Can. Soc. Forens. Sci. J., 31, 1–19 (1998).
27. Altria K.D., Bryant S.M., Hadgett T., Validated CE method for the assay of a range of acidic drugs
and excipients. J. Pharm. Biomed. Anal., 15, 1091–1101 (1997).
28. Khaledi Morteza G. (Ed.), High-Performance Capillary Electrophoresis: Theory, Techniques, and
Applications. (John Wiley & Sons, New York 1998).
29. Foret F., Bocek P., Krivankova L. Capillary Zone Electrophoresis. (John Wiley and Sons, New York,
1993).
30. Altria K.D. (Ed.), Capillary Electrophoresis Guidebook: Principles, Operations and Applications.
(Humana Pr. 1996).
31. Baker Dale R., Capillary Electrophoresis. (John Wiley & Sons, New York, 1995).
32. Smith N.W., Evans M.B., Capillary zone electrophoresis in pharmaceutical and biomedical analysis.
J. Pharm. Biomed. Anal., 12, 579–611 (1994).
33. Esaka Y., Okumura N., Uno B., Goto M., Non-aqueous capillary electrophoresis of p-quinone anion
radicals. Anal. Sci.., 17, 99–102 (2001).
34. Yuqin L., Shuya C., Yuqiao C., Xingguo C., Zhide H., Application of nonaqueous capillary elec-
trophoresis for quantitative analysis of quinolizidine alkaloids in Chinese herbs. Anal. Chim. Acta.
2004, 508, 17–22.
35. Leung G. N. W., Tang, H. P. O., Tso, T. S. C., Wan, T.S. M., Separation of basic drugs with non-aqueous
capillary electrophoresis. J. Chromatogr. A, 738, 141–154 (1996).
36. Cherkaoui S., Varesio E., Christen P., Veuthey J.L., Selectivity manipulation using nonaqueous capil-
lary electrophoresis. Application to tropane alkaloids and amphetamine derivatives. Electrophoresis,
19, 2900–2906 (1998).
37. Cantú M. D., Hillebrand S., Costa Queiroz M. E., Lanças F. M., Carrilho E., Validation of non-
aqueous capillary electrophoresis for simultaneous determination of four tricyclic antidepressants in
pharmaceutical formulations and plasma samples. J. Chromatogr. B, 799, 127–132 (2004).
38. Anurukvorakun O., Suntornsuk W., Suntornsuk L., Factorial design applied to a non-aqueous capillary
electrophoresis method for the separation of β-agonists. J. Chromatogr. A, 1134, 326–332 (2006).
39. Cherkaoui S., Veuthey J.L., Separation of selected anesthetic drugs by non-aqueous capillary
electrophoresis. Analysis, 27, 765–771 (1999).
40. Riekkola M.L., Jisilli M., Porras S.P., Valko I.E., Non Aqueous capillary electrophoresis. J. Chrom. A,
892, 155–170 (2000).
41. Matysik F.M., Special aspects of detection methodology in nonaqueous capillary electrophoresis.
Electrophoresis, 23, 400–407 (2002).
42. Fillet M., Servais A.C., Crommen J., Effects of background electrolyte composition and addition
of selectors on separation selectivity in nonaqueous capillary electrophoresis. Electrophoresis, 24,
1499–1507 (2003).
43. Porras S.P., Kenndler E., Capillary zone electrophoresis in non-aqueous solutions: pH of the
background electrolyte. J. Chromatogr. A, 1037, 455–465 (2004).
44. Bjornsdottir I., Hansen S.H., Determination of opium alkaloids in pure opium using non-aqueous
capillary electrophoresis. J. Pharm. Biomed. Anal., 13, 1473–1481 (1995).
45. Bjornsdottir I., Hansen S.H., Comparison of the separation selectivity in aqueous and non-aqueous
CE. J. Chromatogr. A, 711, 313–319 (1995).
46. Cherkaoui S., Varesio E., Christen P., Veuthey J.L., Selectivity manipulation using non-aqueous CE.
Application to tropane alkaloids and amphetamine derivatives. Electrophoresis, 19, 2900–2906 (1998).
47. Servais A.C., Fillet M., Chiap P., Abushoffa A.M., Hubert P., Crommen J., Optimisation of the
separation of beta-blockers by ion-pair CE in non-aqueous media using univariate and multivariate
approaches. J. Sep. Sci., 25, 1087–1095 (2002).
48. Porras S.P., Reikkola M.-L., Capillary zone electrophoresis of basic analytes in methanol as non-
aqueous solvent: Mobility and ionisation constant. J. Chromatogr. A, 905, 259–268 (2001).
49. Siren H., Hiissa T., Min Y., Aqueous and non-aqueous capillary electrophoresis of polar drugs. Analyst,
9, 1561–1568 (2000).
50. Hedeland Y., Hedeland M., Bondesson U., Petterson C., Chiral separations of amines with
N-benzoxycarbonylglycil-l-proline as selector in non-aqueous CE using methanol and 1,2-
dichloroethane in the background electrolyte. J. Chromatogr. A, 984, 261–271 (2003).
51. Pyell U. (Ed.), Electrokinetic Chromatography: Theory, Instrumentation and Applications (John Wiley
& Sons, England, 2006).
52. Pappas T.J., Gayton-Ely M., Holland L.A., Recent advances in micellar electrokinetic chromatography.
53. Molina M., Silva M., Micellar electrokinetic chromatography: Current developments and future.
54. Pyell U., Micellar electrokinetic chromatography—From theoretical concepts to real samples
(Review). Fresenius J. Anal. Chem., 371, 691–703 (2001).
55. Watari H., Microemulsion capillary electrophoresis. Chem. Lett., 3, 391–394 (1991).
56. Broderick M., Donegan S., Power J., Altria K.D., Optimisation of water-in-oil MEEKC in
pharmaceutical analysis. J. Pharm. Biomed. Anal., 37, 877–884 (2005).
57. Altria K.D., Broderick M.F., Donegan S., Power J., The use of novel water-in-oil microemulsions in
microemulsion electrokinetic chromatography. Electrophoresis, 25, 645–652 (2004).
58. Altria K.D., Application of microemulsion electrokinetic chromatography to the analysis of a wide
range of pharmaceuticals and excipients. J. Chromatogr. A, 844, 371–386 (1999).
59. Hansen S.H., Gabel-Jenson C., Mohammed El-Sherbiny D., Pedersen-Bjergaard S., Microemulsion
electrokinetic chromatography—or solvent modified micellar electrokinetic chromatography. Trends
Anal. Chem., 20, 614–619 (2001).
60. Cherkaoui S., Veuthey J.-L., Micellar and microemulsion electrokinetic chromatography of selected
anaesthetic drugs. J. Sep. Sci., 25, 1073–1078 (2002).
61. Wilson I.D., Poole C.F., (Eds). Liquid chromatography: Microemulsion based HPLC methods.
Encyclopedia of Separation Science—Online Update ISBN: 978-0-12-226770-3 (Elsevier, 2007).
62. Hansen S.H., Recent applications of microemulsion electrokinetic chromatography. Electrophoresis,
24, 3900–3907 (2003).
63. Marsh A., Broderick M., Altria K.D., Clark B., Power J., Donegan S., Recent advances in
microemulsion electrokinetic chromatography. Electrophoresis, 25, 3970–3980 (2004).
64. Huie C.W., Recent applications of microemulsion electrokinetic chromatography. Electrophoresis, 27,
60–75 (2006).
65. McEvoy E., Marsh A., Altria K.D., Donegan S., Power J., Recent advances in the development and
applications of microemulsion electrokinetic chromatography. Electrophoresis, 28, 193–207 (2007).
66. Watari H., Microemulsions in separation sciences. J. Chromatogr. A, 780, 93–102 (1997).
67. Smith R.D., Barinaga C.J., Udseth H.R., Improved electrospray ionisation interface for capillary zone
electrophoresis-mass spectrometry. Anal. Chem., 60, 1948–1952 (1988).
69. Ohnesorge J., Neususs C., Watsig H., Quantitation in capillary electrophoresis-mass spectrometry.
70. Schmitt-Kopplin P., Englmann M., Capillary electrophoresis-mass spectrometry: Survey on develop-
ments and applications 2003–2004. Electrophoresis, 26, 1209–1220 (2005).
72. Shamsi S.A., Miller B.E., Capillary electrophoresis-mass spectrometry: Recent advances to the
analysis of small achiral and chiral solutes. Electrophoresis, 25, 3927–3961 (2005).
73. Issaq H.J., Janini G.M., Chan K.C., Veenstra T.D., Sheathless electrospray ionization interfaces for
capillary electrophoresis-mass spectrometric detection—Advantages and limitations. J. Chromatogr.
A, 1053, 37–42 (2004).
74. Schmitt-Kopplin P., Frommberger M., Capillary electrophoresis-mass spectrometry: 15 years of
developments and applications. Electrophoresis, 24, 3737–3867 (2003).
75. Tu J., Halsall H.B., Seliskar C.J., Limbach P.A., Wehmeyer K.R., Heineman W.R., Estimation of
log Pow values for neutral and basic compounds by microchip microemulsion electrokinetic chro-
matography with indirect fluorimetric detection (mu MEEKC-IFD). J. Pharm. Biomed. Anal., 38, 1–7
(2005).
76. Zhou C.X., Yin Y.K., Kenseth J.R., Stella M., Wehmeyer K.R., Heineman W.R., Rapid pKa estima-
tion using vacuum-assisted multiplexed capillary electrophoresis (VAMCE) with ultraviolet detection.
77. Wong K.S., Kenseth J., Strasburg R., Validation and long-term assessment of an approach for the high
throughput determination of lipophilicity (log P-OW) values using multiplexed, absorbance-based
capillary electrophoresis. J. Pharm. Sci., 93, 916–931 (2004).
78. Wehmeyer K.R., Tu J., Yin Y.K., King S., Stella M., Stanton D.T., Kenseth J., Wong K.S.,
The application of multiplexed microemulsion electrokinetic chromatography for the rapid deter-
mination of log Pow values for neutral and basic compounds. LC GC North America 90–95
Suppl. S, (2004).
79. Kenseth J., Bastin A., High-throughput chiral separations using the cePRO 9600 (TM) multiplexed
capillary electrophoresis system. LC-GC North America, 28–29 Suppl. S (2004).
80. Reetz M.T., Kuhling K.M., Deege A., Hinrichs H., Belder D., Super-high-throughput screening of
enantioselective catalysts by using capillary array electrophoresis. Angew. Chem. Int. Ed., 39, 3891–
3893 (2000).
81. Zhong W., Yeung E.S., Combinatorial enantiomeric separation of diverse compounds using capillary
array electrophoresis. Electrophoresis, 23, 2996–3005 (2002).
82. Zhang Y.H., Gong X.Y., Zhang H.M., Larock R.C., Yeung E.S., Combinatorial screening of homoge-
neous catalysis and reaction optimization based on multiplexed capillary electrophoresis. J. Combinat.
Chem. 2(5), 450–452 (2000).
83. Marsh A., Altria K.D., Use of multiplexed capillary electrophoresis for pharmaceutical analysis.
Chromatographia, 64, 327–333 (2006).
84. Pang H.M., Kenseth J., Coldiron S., High-throughput multiplexed capillary electrophoresis in drug
discovery. Drug Discov. Today, 9, 1072–1080 (2004).
85. Moring S.E., Reel R.T., Van Soest R.E.J., Optical improvements of a Z shaped cell for high sensitivity
UV absorbency detection in capillary electrophoresis. Anal. Chem., 62, 3454–3459 (1993).
86. Tsuda T., Sweedler J.V., Zare R.N., Rectangular capillaries for capillary zone electrophoresis. Anal.
Chem., 62, 2149–2152 (1990).
87. Wang T.S., Aiken J.H., Huie C.W., Hartwick R.A., Nanoliter-scale multireflection cell for absorption
detection in capillary electrophoresis. Anal. Chem., 63, 1372–1376 (1991).
88. Heiger D.N., High Performance Capillary Electrophoresis—An Introduction. (Hewlett-Packard Co.,
France), 1992.
89. Mathies R., Peck K., Optimisation of high sensitivity fluorescence detection. Anal. Chem., 62, 1786–
1791 (1990).
90. Dittrich P.S., Manz A., Single-molecule fluorescence detection in microfluidic channels—The Holy
Grail in mu TAS? Anal. Bioanal. Chem., 382, 1771–1782 (2005).
91. Johnson M.E., Landers J.P., Fundamentals and practice for ultrasensitive laser-induced fluorescence
detection in microanalytical systems. Electrophoresis, 25, 3513–3527 (2004).
92. Xu C.L., Li B.X., Zhang Z.J., Recent developments in detectors for micro total analysis systems. Chin.
J. Anal Chem., 31, 1520–1526 (2003).
93. McWhorter S., Soper S.A., Near-infrared laser-induced fluorescence detection in capillary elec-
trophoresis. Electrophoresis, 21, 1267–1280 (2000).
94. Zhang X., Stuart J.N., Sweedler J.V., Capillary electrophoresis with wavelength-resolved laser-induced
fluorescence detection. Anal. Bioanal. Chem., 373, 332–343 (2002).
95. Lui J.F., Yang X.R., Wang E.K., Progress in amperometric detection for capillary electrophoresis.
Chin. J. Anal Chem., 30, 748–753 (2002).
96. Zang L., Tong P., Chen G.N., Determination of the hydrolysis rate constants and activation energy of
aesculin with capillary electrophoresis end-column amperometric detection. J. Chromatogr. A, 1098,
194–198 (2005).
97. Zhang S.S., Yuan Z.B., Liu H.X., Zou H., Xiong H., Wu Y.J., Analysis of acyclovir by high performance
capillary electrophoresis with on-column amperometric detection. Electrophoresis, 21, 2995–2998
(2000).
98. Bowman J., Tang L.Y., Silverman C.E., Analysis of small amines in pharmaceuticals by capillary ion
electrophoresis with conductivity detection. J. Pharm. Biomed. Anal., 23, 663–669 (2000).
99. de Silva J.A.F., Electrochemical detection in capillary electrophoresis. Quimica Nova., 26, 56–64
(2003).
100. Baldwin R.P., Recent advances in electrochemical detection in capillary electrophoresis. Electrophore-
sis, 21, 4017–4028 (2000).
101. Holland L.A., Leigh A.M., Amperometric and voltametric detection for capillary electrophoresis.
102. Wang A.B., Fang Y.Z., Applications of capillary electrophoresis with electrochemical detection in
pharmaceutical and biomedical analyses. Electrophoresis, 21, 1281–1290 (2000).
103. Swinney K., Bornhop D.J., Detection in capillary electrophoresis. Electrophoresis, 21, 1239–1250
(2000).
104. Rizzi A., Fundamental aspects of chiral capillary electrophoresis. Electrophoresis, 22, 3079–3106
(2001).
105. Gubitz G., Schmid M.G., Chiral separation principles in chromatographic and electromigration
techniques. Mol. Biotechnol., 32, 159–179 (2006).
106. Chankvetadze, B. Capillary Electrophoresis in Chiral Analysis. (John Wiley & Sons, New York, 1997).
107. Gubitz G., Schmid M.G., Recent advances in chiral separation principles in capillary electrophoresis
and capillary electrochromatography. Electrophoresis, 25, 3981–3996 (2004).
108. Amini A., Recent developments in chiral capillary electrophoresis and applications of this technique
to pharmaceutical and biomedical analysis. Electrophoresis, 22, 3107–3130 (2001).
109. Denola N.L., Quiming N.S., Catabay A.P., Saito Y., Jinno K., Optimization of capillary electrophoretic
enantioseparation for basic drugs with native beta-CD as a chiral selector. Electrophoresis, 27, 2367–
2375 (2006).
110. Koppenhoefer B., Zhu X., Jakob A., Wuerthner S., Lin B., Separation of drug enantiomers by capillary
electrophoresis in the presence of neutral cyclodextrins. J. Chrom. A, 875, 135–161 (2000).
111. Haginaka J., Enantiomer separation of drugs by capillary electrophoresis using proteins as chiral
selectors. J. Chromatogr. A, 875, 235–254 (2000).
112. Chiou C.-S., Shih J.-S., Application of crown ethers as modifiers for the separation of amines by
capillary electrophoresis. Anal. Chim. Acta, 360, 69–76 (1998).
113. Fanali S., Aturki Z., Desiderio C., Righetti P.G., Use of MDL 63 246 (Hepta-Tyr) antibiotic
in capillary zone electrophoresis: II. Chiral resolution of α-hydroxy acids. J. Chromatogr. A,
838, 223–235 (1999).
114. Blanco M., Valverde I., Choice of chiral selector for enantioseparation by capillary electrophoresis.
Trends Anal. Chem., 22, 428–439 (2003).
115. Shamsi S.A., Chiral capillary electrophoresis-mass spectrometry: Modes and applications. Elec-
trophoresis, 23, 4036–4051 (2003).
116. Fillet M., Hubert P., Crommen J., Method development strategies for the enantioseparation of drugs
by capillary electrophoresis using cyclodextrins as chiral additives. Electrophoresis, 19, 2834–2840
(1998).
117. Evans C.E., Stalcup A.M., Comprehensive strategy for chiral separations using sulfated cyclodextrins
in capillary electrophoresis. Chirality, 15, 709–723 (2003).
118. Zhou L., Thompson R., Song S., Ellison D., Wyvratt J.M., A strategic approach to the develop-
ment of capillary electrophoresis chiral methods for pharmaceutical basic compounds using sulfated
cyclodextrins. J. Pharm. Biomed. Anal., 27, 541–553 (2002).
119. Kuhn R., Stoecklin F., Erni F., Chiral separations by host-guest complexation with cyclodextrin and
crown ether in capillary zone electrophoresis. Chromatographia, 33, 32–36 (1992).
120. Iványi T., Pal K., Lázár I., Massart D.L., Vander Heyden Y., Application of tetraoxadiaza-crown ether
derivatives as chiral selector modifiers in capillary electrophoresis. J. Chromatogr. A, 1028, 325–332
(2004).
121. Fanali S., Cartoni C., Desiderio C., Chiral separation of newly synthesized arylpropionic acids
by capillary electrophoresis using cyclodextrins or a glycopeptide antibiotic as chiral selectors.
Chromatographia, 54, 87–92 (2001).
122. Desederio C., Fanali S., Chiral analysis by capillary electrophoresis using antibiotics as chiral selector.
J. Chromatogr. A, 807, 37–56 (1998).
123. Aboul-Enein H.Y., Ali I., Macrocyclic antibiotics as effective chiral selectors for enantiomeric res-
olution by liquid chromatography and capillary electrophoresis. Chromatographia, 52, 679–691
(2000).
124. Ha P.T.H., Van Schepdael A., Roets E., Hoogmartens J., Investigating the potential of erythromycin
and derivatives as chiral selector in capillary electrophoresis. J. Pharm. Biomed. Anal., 34, 861–870
(2004).
125. Park H., Lee S., Kang S., Jung Y., Jung S., Enantioseparation using sulfated cyclosophoraoses as a
novel chiral additive in capillary electrophoresis. Electrophoresis, 25, 2671–2674 (2004).
126. Budanova N., Shapovalova E., Lopatin S., Varlamov V., Shpigun O., N-(3-sulfo, 3-carboxy)-
propionylchitosan as new chiral selector for enantioresolution of basic drugs by capillary electrophore-
sis. Chromatographia, 59, 709–713 (2004).
127. Bortocan R., Bonato P.S., Enantioselective analysis of primaquine and its metabolite carboxypri-
maquine by capillary electrophoresis. Electrophoresis, 25, 2848–2853 (2004).
128. Quan Z., Song Y., Feng Y., LeBlanc M.H., Liu Y.M., Detection of d-serine in neural samples by
saccharide enhanced chiral capillary electrophoresis. Anal. Chim. Acta, 528, 101–106 (2005).
129. Millot M.C., Separation of drug enantiomers by liquid chromatography and capillary electrophoresis,
using immobilized proteins as chiral selectors. J. Chromatogr. B, 797, 131–159 (2003).
130. Tesarova E., Sevcik J., Gag B., Armstrong D.W., Effects of partial/asymmetrical filling of micelles
and chiral selectors on capillary electrophoresis enantiomeric separation: Generation of a gradient.
131. Dobashi A., Ono T., Hara S., Yamaguchi J., Optical resolution of enantiomers with chiral mixed
micelles by electrokinetic chromatography. Anal. Chem., 61, 1984–1986 (1989).
132. Terabe S., Shibata M., Miyashita Y.J., Chiral separation by electrokinetic chromatography with bile
salt micelles. J. Chromatogr. A, 480, 403–411 (1989).
133. Tickle D.C., Okafo G.N., Camilleri P., Jones R.F.D., Kirby J., Glucopyranoside-based surfactants as
pseudostationary phases for chiral separations in capillary electrophoresis. Anal. Chem., 66, 4121–
4126 (1994).
134. Otsuka K., Terabe S., Enantiomer separation of drugs by micellar electrokinetic chromatography using
chiral surfactants. J. Chromatogr. A, 875, 163–178 (2000).
135. Aiken J.H., Huie C.W., Use of a microemulsion system to incorporate a lipophilic chiral selector in
electrokinetic capillary chromatography. Chromatographia, 35, 448–450 (1993).
136. Pascoe R., Foley J., Rapid separation of pharmaceutical enantiomers using electrokinetic chromatog-
raphy with a novel chiral microemulsion. Analyst, 127, 710–714 (2002).
137. Mertzman M.D., Foley J.P., Chiral cyclodextrin modified microemulsion electrokinetic chromatogra-
phy. Electrophoresis, 25, 1188–1200 (2004).
138. Mertzman M.D., Foley J.P., Effect of surfactant concentration and buffer selection on chromatographic
figures of merit in chiral microemulsion electrokinetic chromatography. Electrophoresis, 25, 3247–
3256 (2004).
139. Zheng Z.X., Lin J.M., Chan W.H., Lee A.W.M., Huie C.W., Separation of enantiomers in microemul-
sion electrokinetic chromatography using chiral alcohols as cosurfactants. Electrophoresis, 25,
3263–3269 (2004).
140. Kahle K.A., Foley J.P., Chiral microemulsion electrokinetic chromatography with two chiral com-
ponents: Improved separations via synergies between a chiral surfactant and a chiral cosurfactant.
141. Garcia-Ruiz C., Marina M.L., Sensitive chiral analysis by capillary electrophoresis. Electrophoresis,
27, 195–212 (2006).
142. Shamsi S.A., Miller B.E., Capillary electrophoresis-mass spectrometry: Recent advances to the
analysis of small achiral and chiral solutes. Electrophoresis, 25, 3927–3961 (2004).
143. Gubitz G., Schmid M.G., Recent advances in chiral separation principles in capillary electrophoresis
and capillary electrochromatography. Electrophoresis, 25, 3981–3996 (2004).
144. Gubitz G., Schmid M.G., Chiral separation by chromatographic and electromigration techniques. A
review. Biopharm. Drug Dispos. 22, 291–336 (2001).
145. McCorquodale M.E., Boutrid H., Coyler C.L., Enantiomeric separation of N -nitrosonornictine by
capillary electrophoresis. Anal. Chim. Acta, 496, 177–184 (2003).
146. Olsson J., Stegander F., Marlin N., Wan H., Blomberg L.G., Enantiomeric separation of omeprazole
and its metabolite 5-hydroxyomeprazole using non-aqueous capillary electrophoresis. J. Chromatogr.
A, 1129, 291–295 (2006).
147. Blanco M., Valverde I., Chiral and non chiral determination of Dopa by capillary electrophoresis.
148. Sänger-van de Griend C.E., Wahlström H., Gröningsson K., Widahl-Näsman M., A chiral capillary
electrophoresis method for ropivacaine hydrochloride in pharmaceutical formulations: Validation and
comparison with chiral liquid chromatography. J. Pharm. Biomed. Anal., 15, 1051–1061 (1997).
149. Jimidar M., Van Ael W., De Smet M., Cockaerts P., Method validation and robustness testing of an
enantioselective CE method for chemical quality control. LCGC Europe, 15, 230–243 (2002).
150. Song S., Zhou L., Thompson R., Yang M., Ellison D., Wyvratt J.M., Comparison of capillary elec-
trophoresis and reversed-phase liquid chromatography for determination of the enantiomeric purity of
an M3 antagonist. J. Chromatogr. A, 959, 299–308 (2002).
151. Kuo C.Y., Wu H.L., Wu S.M., Enantiomeric analysis of methotrexate in pharmaceuticals by
cyclodextrin-modified capillary electrophoresis. Anal. Chim. Acta, 471, 211–217 (2002).
152. Zhang B., Krull I.S., CohenA., Smisek D.L., KlossA., Wang B., BourqueA.J., Separation of quaternary
ammonium diastereomeric oligomers by capillary electrophoresis. J. Chromatogr. A, 1034, 213–220
(2004).
153. Ramstad T., Enantiomeric purity methods for three pharmaceutical compounds by electrokinetic cap-
illary chromatography utilizing highly sulfated-γ -cyclodextrin as the chiral selector. J. Chromatogr.
A, 1127, 286–294 (2006).
154. Khan M., Viswanathan B., Sreenivas R.D., Sreenivas R.G., A validated chiral CE method for
Frovatriptan, using cyclodextrin as chiral selector. J. Pharm. Biomed. Anal., 41, 1447–1452 (2006).
155. Zhou S., Ouyang J., Baeyens W.R.G., Zhao H., Yang Y., Chiral separation of four fluoroquinolone
compounds using capillary electrophoresis with hydroxypropyl-β-cyclodextrin as chiral selector.
J. Chromatogr. A, 1130, 296–301 (2006).
156. Martínez-Pla J.J., Martín-Biosca Y., Sagrado S., Villanueva-Camañas R.M., Medina-Hernández M.J.,
Chiral separation of bupivacaine enantiomers by capillary electrophoresis partial-filling technique with
human serum albumin as chiral selector. J. Chromatogr. A, 1048, 111–118 (2004).
157. Wang F., Khaledi M.G., Non-aqueous capillary electrophoresis chiral separations with sulfated
β-cyclodextrin. J. Chromatogr. B: Biomed. Sci. Appl., 731, 187–197 (1999).
158. Chankvetadze B., Burjanadze N., Blaschke G., Enantioseparation of erythro-mefloquine and its
analogues in capillary electrophoresis. J. Pharm. Biomed. Anal., 32, 41–49 (2003).
159. Guo L., Lin S.J., Yang Y.F., Qi L., Wang M.X., Chen Y., Fast enantioseparation of arylglycine amides
by capillary electrophoresis with highly sulfated-β-cyclodextrin as a chiral selector. J. Chromatogr. A,
998, 221–228 (2003).
160. Gómez-Gomar A., Ortega E., Calvet C., Andaluz B., Mercé R., Frigola J., Enantioseparation of basic
pharmaceutical compounds by capillary electrophoresis using sulfated cyclodextrins: Application to
E-6006, a novel antidepressant. J. Chromatogr. A, 990, 91–98 (2003).
161. Tanaka Y., Otsuka K., Terabe S., Separation of enantiomers by capillary electrophoresis-mass spec-
trometry employing a partial filling technique with a chiral crown ether. J. Chromatogr. A, 875, 323–330
(2000).
162. Lipka E., Selouane A., Postel D., Len C., Vaccher M.P., Bonte J.P., Vaccher C., Enantioseparation
of four cis and trans diastereomers of 2 ,3 -didehydro-2 ,3 -dideoxythymidine analogs, by high-
performance liquid chromatography and capillary electrophoresis. J. Chromatogr. A, 1034, 161–167
(2004).
163. British Pharmacopoeia 2005, Volume 2, p. 1177.
164. Dose E.V., Guiochon G.A., Internal standardisation technique for capillary zone electrophoresis. Anal.
Chem., 63, 1154–1158 (1991).
165. Ross G.A., Voltage pre-conditioning technique for optimisation of migration time reproducibility in
166. Shihabi Z.K., Hinsdale M.E., Some variables affecting reproducibility in capillary electrophoresis.
167. Schaeper J.P., Sepaniak M.J., Parameters affecting reproducibility in capillary electrophoresis.
168. Mayer B.X., How to increase precision in capillary electrophoresis. J. Chromatogr. A, 907, 21–37
(2001).
169. Altria K.D., Enhanced pharmaceutical analysis by CE using dynamic surface coating system. J. Pharm.
Biomed. Anal., 31, 447–453 (2003).
170. Aurora-Prado M.S., Silva C.A., Tavares M.F.M., Altria K.D., Determination of folic acid in tablets by
microemulsion electrokinetic chromatography. J. Chromatogr. A, 1051, 291–296 (2004).
171. Lehmann S.V., Bergholdt A.B., Development and validation of a high-precision capillary elec-
trophoresis method for main component assay of ragaglitazar. J. Chromatogr. A, 1011, 203–211
(2003).
172. Jamali B., Lehmann S.V., Development and validation of a high-resolution CE method for multi-
analysis of ragaglitazar and arginine active pharmaceutical ingredients and low-dose tablets. J. Pharm.
Biomed. Anal., 34, 463–472 (2004).
173. Pajchel G., Michalska K., Tyski S., Application of capillary electrophoresis to the determination of
various benzylpenicillin salts. J. Chromatogr. A, 1032, 265–272 (2004).
174. Nemutlu E., Yardimci C., Ozaltin N., Determination of sertaconazole in pharmaceutical preparations
by capillary electrophoresis. Anal. Chim. Acta, 547, 83–88 (2005).
175. Perez-Ruiz T., Martinez-Lozano C., Sanz A., Bravo E., Development and validation of a quantitative
assay for raloxifene by capillary electrophoresis. J. Pharm. Biomed. Anal., 34, 891–897 (2004).
176. Guihen E., Sisk G.D., Scully N.M., Glennon J.D., Rapid analysis of atorvastatin calcium using capillary
electrophoresis and microchip electrophoresis. Electrophoresis, 27, 2338–2347 (2006).
177. Hamoudova R., Pospisilova M., Determination of ibuprofen and flurbiprofen in pharmaceuticals by
capillary zone electrophoresis. J. Pharm. Biomed. Anal., 41, 1463–1467 (2006).
178. Komsta L., Misztal G., Majchrzak E., Hauzer A., Separation of fibrate-type antihyperlipidemic drugs
by capillary electrophoresis and their quantitation in pharmaceuticals. J. Pharm. Biomed. Anal., 41,
408–414 (2006).
179. Hillaert S., Van den Bossche W., Simultaneous determination of hydrochlorothiazide and several
angiotensin-II-receptor antagonists by capillary electrophoresis. J. Pharm. Biomed. Anal., 31, 329–339
(2006).
180. Hillaert S., De Beer T.R., De Beer J.O., Van den Bossche W., Optimisation and validation of a micellar
electrokinetic chromatographic method for the analysis of several angiotensin-II-receptor antagonists.
J. Chromatogr. A, 984, 135–146 (2003).
181. Hillaert S., Snoeck L., Van den Bossche W., Optimization and validation of a capillary zone elec-
trophoretic method for the simultaneous analysis of four atypical antipsychotics. J. Chromatogr. A,
1033, 357–362 (2004).
182. Mohammadi A., Kanfer I., Walker R.B., A capillary zone electrophoresis (CZE) method for the
determination of cyclizine hydrochloride in tablets and suppositories. J. Pharm. Biomed. Anal., 35,
233–239 (2004).
183. Kou C.Y., Wu S.M., Micellar electrokinetic chromatography for simultaneous determination of six
corticosteroids in commercial pharmaceuticals. J. Sep. Sci., 28, 144–148.
184. Michalska K., Pajchel G., Tyski S., Capillary electrophoresis method for simultaneous determination
of penicillin G, procaine and dihydrostreptomycin in veterinary drugs. J. Chromatogr. B, 800, 203–209
(2004).
185. Li Y., Cui S., Cheng Y., Chen X., Hu Z., Application of nonaqueous capillary electrophoresis for
quantitative analysis of quinolizidine alkaloids in Chinese herbs. Anal. Chim. Acta, 508, 17–22 (2004).
186. Flores J.R., Nevado J.J.B., Penalvo G.C., Diez N.M., Development and validation method for deter-
mination of fluoxetine and its main metabolite norfluoxetine by nonaqueous capillary electrophoresis
in human urine. Talanta, 65, 163–171 (2005).
187. Vanhoenacker G., de l’Escaille F., De Keukeleire D., Sandra P., Analysis of benzodiazepines in dynam-
ically coated capillaries by CE-DAD, CE-MS and CE-MS2 . J. Pharm. Biomed. Anal., 34, 595–606
(2004).
188. Lurie I.S., Hays P.A., Garcia A.E., Panicker S., Use of dynamically coated capillaries for the deter-
mination of heroin, basic impurities and adulterants with capillary electrophoresis. J. Chromatogr. A,
1034, 227–235 (2004).
189. Okamoto H., Nakajima T., Ito Y., Aketo T., Shimada K., Yamato S., Simultaneous determi-
nation of ingredients in a cold medicine by cyclodextrin-modified microemulsion electrokinetic
chromatography. J. Pharm. Biomed. Anal., 37, 517–528 (2005).
190. International Conferences on Harmonization, Draft revised guidance on impurities in new drug
substances. Q3A(R). Fed. Regist., 65, 45085–45090 (2000).
191. Cruz L.A., Hall R., Enantiomeric purity assay of moxifloxacin hydrochloride by capillary electrophore-
sis. J. Pharm. Biomed. Anal., 38, 8–13 (2005).
192. Sokoliess T., Koller G., Approach to method development and validation in capillary electrophoresis for
enantiomeric purity testing of active basic pharmaceutical ingredients. Electrophoresis, 26, 2330–2341
(2005).
193. Hansen S.H., Sheribah Z.A., Comparison of CZE, MEKC, MEEKC and non-aqueous capillary elec-
trophoresis for the determination of impurities in bromazepam. J. Pharm. Biomed. Anal., 39, 322–327
(2005).
194. Kelly M.A., Altria K.D., Grace C., Clark B.J., Optimisation, validation and application of a capil-
lary electrophoresis method for the determination of ranitidine hydrochloride and related substances.
J. Chromatogr. A, 789, 297–306 (1998).
195. Fabre H., Perrin C., Bosc N., Determination of homotaurine as impurity in calcium acamprosate by
capillary zone electrophoresis. J. Chromatogr. A, 853, 421–430 (1999).
196. Toro I., Dulsat J.F., Fábregas J.L., Claramunt J., Development and validation of a capillary elec-
trophoresis method with ultraviolet detection for the determination of the related substances in a
pharmaceutical compound. J. Chromatogr. A, 1043, 303–315 (2004).
197. Mallampati S., Leonard S., De Vulder S., Hoogmartens J., Van Schepdael A., Method development
and validation for the analysis of didanosine using micellar electrokinetic capillary chromatography.
198. Orlandini S., Fanali S., Furlanetto S., Marras A.M., Pinzauti S., Micellar electrokinetic chromatog-
raphy for the simultaneous determination of ketorolac tromethamine and its impurities—Multivariate
optimization and validation. J. Chromatogr. A, 1032, 253–263 (2004).
199. Furlanetto S., Orlandini S., Marras A.M., Mura P., Pinzauti S., Mixture design in the optimization
of a microemulsion system for the electrokinetic chromatographic determination of ketorolac and its
impurities: Method development and validation. Electrophoresis, 27, 805–818 (2006).
200. Wen T., Zhao X., Luo G.A., Wang J., Wang Y.M., Li P., Zhu J., Yu Z.S., Simultaneous determination
of heroin, amphetamine and their basic impurities and adulterants using microemulsion electrokinetic,
chromatography. Chin. Chem. Lett., 16, 1499–1502 (2005).
201. Vassort A., Barrett D.A., Shaw P.N., Ferguson P.D., Szucs R., A generic approach to the impurity
profiling of drugs using standardised and independent capillary zone electrophoresis methods coupled
to electrospray ionisation mass spectrometry. Electrophoresis, 26, 1712–1723 (2005).
202. Visky D., Jimidar I., Van Ael W., Vennekens T., Redlich D., De Smet M., Capillary electrophoresis-
mass spectrometry in impurity profiling of pharmaceutical products. Electrophoresis, 26, 1541–1549
(2005).
203. Wan H., Schmidt S., Carlsson L., Blomberg L.G., Some factors affecting enantiomeric impurity
determination by capillary electrophoresis using ultraviolet and laser-induced fluorescence detection.
204. Hilhorst M.J., Somsen G.W., de Jong G.J., Capillary electrokinetic separation techniques for profiling
of drugs and related products. Electrophoresis, 22, 2542–2564 (2001).
205. Berzas Nevado J.J., Cabanillas C.G., Villaseñor Llerena M.J., Rodríguez R.V., Enantiomeric deter-
mination, validation and robustness studies of racemic citalopram in pharmaceutical formulations by
206. Kvasnièka F., Bíba B., Cvak L., Capillary zone electrophoresis separation of enantiomers of lisuride.
J. Chromatogr. A, 1066, 255–258 (2005).
207. Castro-Puyana M., Garcia-Ruiz C., Cifuentes A., Crego A.L., Marina M.L., Identification and
quantitation of cis-ketoconazole impurity by capillary zone electrophoresis-mass spectrometry.
208. Berzas Nevado J.J., Castañeda Peñalvo G., Rodríguez Dorado R.M., Method development and valida-
tion for the separation and determination of omeprazole enantiomers in pharmaceutical preparations
by capillary electrophoresis. Anal. Chim. Acta, 533, 127–133 (2005).
209. Pajchel G., Michalska K., Tyski S., Analysis of phenoxymethylpenicillin potassium by capillary
electrophoresis. J. Chromatogr. A, 1087, 197–202 (2005).
210. Sabbah S., Scriba G.K.E., Development and validation of a capillary electrophoresis assay for the deter-
mination of 3,4-diaminopyridine and 4-aminopyridine including related substances. J. Chromatogr. A,
907, 321–328 (2001).
211. Gotti R., Pomponio R., Bertucci C., Cavrini V., Determination of 5-aminosalicylic acid related impu-
rities by micellar electrokinetic chromatography with an ion-pair reagent. J. Chromatogr. A, 916,
175–183 (2001).
212. Kang J.W., Van Schepdael A., Roets E., Hoogmartens J., Analysis of vancomycin and related
impurities by micellar electrokinetic capillary chromatography. Method development and validation.
213. Owens P.K., Wikström H., Någård S., Karlsson L., Development and validation of a capillary elec-
trophoresis method for ximelagatran assay and related substance determination in drug substance and
tablets. J. Pharm. Biomed. Anal., 27, 587–598 (2002).
214. Hilder E.F., Klampfl C.W., Buchberger W., Haddad P.R., Comparison of aqueous and non-aqueous
carrier electrolytes for the separation of penicillin V and related substances by capillary electrophoresis
with UV and mass spectrometric detection. Electrophoresis, 23, 413–420 (2002).
215. Jaworska M., Szulinska G., Wilk M., Tautt J., Capillary electrophoretic separation of N-acetylcysteine
and its impurities as a method for quality control of pharmaceuticals. J. Chromatogr. A, 853, 479–485
(1999).
216. Michalska K., Pajchel G., Tyski S., Determination of ciprofloxacin and its impurities by capillary zone
217. Gil E.C., Dehouck P., Van SchepdaelA., Roets E., Hoogmartens J., Analysis of metacycline by capillary
electrophoresis. Electrophoresis, 22, 497–502 (2001).
218. Fernández H., Rupérez F.J., Barbas C., Capillary electrophoresis determination of loratadine and
related impurities. J. Pharm. Biomed. Anal., 31, 499–506 (2003).
219. Nemutlu E., Ozaltin N., Altinoz S., Determination of rofecoxib, in the presence of its photodegradation
product, in pharmaceutical preparations by micellar electrokinetic capillary chromatography. Anal.
Bioanal. Chem., 378, 504–509 (2004).
220. Avdeef A., Physicochemical profiling (solubility, permeability and charge state). Curr. Top. Med.
Chem., 1, 277–351 (2001).
221. Jia Z.J., Physicochemical profiling by capillary electrophoresis. Curr. Pharm. Anal., 1, 41–56 (2005).
222. Caliaro G.A., Herbots C.A., Determination of pKa values of basic new drug substances by CE. J. Pharm.
Biomed. Anal., 26, 427–434 (2001).
223. Poole S.K., Patel S., Dehring K., Workman H., Poole C.F., Determination of acid dissociation constants
by capillary electrophoresis. J. Chromatogr. A, 1037, 445–454 (2004).
224. Wan H., Holmén A., Någård M., Lindberg W., Rapid screening of pKa values of pharmaceuticals by
pressure-assisted capillary electrophoresis combined with short-end injection. J. Chromatogr. A, 979,
369–377 (2002).
225. Ornskov E., Linusson A., Folestad S., Determination of dissociation constants of labile drug
compounds by capillary electrophoresis. J. Pharm. Biomed. Anal., 33, 379–391 (2003).
226. de Nogales V., Ruiz R., Rosés M., Ràfols C., Bosch E., Background electrolytes in 50% methanol/water
for the determination of acidity constants of basic drugs by capillary zone electrophoresis. J. Chro-
matogr. A, 1123, 113–120 (2006).
227. Jia Z.J., Ramstad T., Zhong M., Medium-throughput pKa screening of pharmaceuticals by pressure-
assisted capillary electrophoresis. Electrophoresis, 22, 1112–1118 (2001).
228. Matoga M., Laborde-Kummer E., Langlois M.H., Dallet P., Bosc J.J., Jarry C., Dubost J.P., Deter-
mination of pKa values of 2-amino-2-oxazolines by capillary electrophoresis. J. Chromatogr. A, 984,
253–260 (2003).
229. Mrestani Y., Neubert R., MunkA., Wiese M., Determination of dissociation constants of cephalosporins
by capillary zone electrophoresis. J. Chromatogr. A, 803, 273–278 (1998).
230. Jankowsky R., Friebe M., Noll B., Johannsen B., Determination of dissociation constants of
99m Technetium radiopharmaceuticals by capillary electrophoresis. J. Chromatogr. A, 833, 83–96
(1999).
231. Hu Q., Hu G., Zhou T., Fang Y., Determination of dissociation constants of anthrocycline by capillary
zone electrophoresis with amperometric detection. J. Pharm. Biomed. Anal., 31, 679–684 (2003).
232. Barták P., Bednár P., Stránský Z., Bocek P., Vespalec R., Determination of dissociation constants of
cytokinins by capillary zone electrophoresis. J. Chromatogr. A, 878, 249–259 (2000).
233. Cao J., Cross R.F., The separation of dihydrofolate reductase inhibitors and the determination of pKa
values by capillary zone electrophoresis. J. Chromatogr. A, 695, 297–308 (1995).
234. Jiménez-Lozano E., Marqués I., Barrón D., Beltrán J.L., Barbosa J., Determination of pKa values of
quinolones from mobility and spectroscopic data obtained by capillary electrophoresis and a diode
array detector. Anal. Chim. Acta, 464, 37–45 (2002).
235. Solinova V., Kasicka V., Koval D., Cesnek M., Holy A., Determination of acid–base dissociation
constants of amino- and guanidinopurine nucleotide analogs and related compounds by capillary zone
electrophoresis. Electrophoresis, 27, 1006–1019 (2006).
236. Geiser L., Henchoz Y., GallandA., Carrupt P.A., Veuthey J.L., Determination of pKa values by capillary
zone electrophoresis with a dynamic coating procedure. J. Sep. Sci., 28, 2374–2380 (2005).
237. Buckenmaier S.M.C., McCalley D.V., Euerby M.R., Determination of pKa values of organic bases in
aqueous acetonitrile solutions using capillary electrophoresis. J. Chromatogr. A, 1004, 71–79 (2003).
238. Foulon C., Danel C., Vaccher C., Yous S., Bonte J.-P., Goossens J.-F., Determination of ionization con-
stants of N-imidazole derivatives, aromatase inhibitors, using capillary electrophoresis and influence
of substituents on pKa shifts. J. Chromatogr. A, 1035, 131–136 (2004).
239. Hansch C., Leo A., Substituent Constants for Correlation Analysis in Chemistry and Biology. (John
Wiley and Sons, New York, 1979).
240. Sangster J., Octanol-water partition coefficients for simple organic compounds. J. Phys. Chem. Ref.
Data, 18, 1111–1229 (1989).
241. Berthod A., Carda-Broch S., Determination of liquid–liquid partition coefficients by separation
methods. J. Chromatogr. A, 1037, 3–14 (2004).
242. Chen N., Zhang Y., Terabe S., Nakagawa T., Effect of physico-chemical properties and molecu-
lar structure on the micelle–water partition coefficient in micellar electrokinetic chromatography.
J. Chromatogr. A, 678, 327–332 (1994).
243. Herbert B.J., Dorsey J.G., n-Octanol water partition-coefficient estimation by micellar electrokinetic
capillary chromatography. Anal. Chem., 67, 744–749 (1995).
244. Lucangioli S.E., Carducci C.N., Scioscia S.L., Carlucci A., Bregni C., Kenndler E., Comparison
of the retention characteristics of different pseudostationary phases for microemulsion and micel-
lar electrokinetic chromatography of betamethasone and derivatives. Electrophoresis, 24, 984–991
(2003).
245. Klotz W.L., Schure M.R., Foley J.P., Determination of octanol–water partition coefficients of pesticides
by microemulsion electrokinetic chromatography. J. Chromatogr. A, 930, 145–154 (2001).
246. Poole S.K., Patel S., Dehring K., Workman H., Dong J., Estimation of octanol–water partition coef-
ficients for neutral and weakly acidic compounds by microemulsion electrokinetic chromatography
using dynamically coated capillary columns. J. Chromatogr. B, 793, 265–274 (2003).
247. Rappel C., Galanski M., Yasemi A., Habala L., Keppler B.K., Analysis of anticancer platinum(II)-
complexes by microemulsion electrokinetic chromatography: Separation of diastereomers and
estimation of octanol–water partition coefficients. Electrophoresis, 26, 878–884 (2005).
248. Wong K.S., Kenseth J., Strasburg R., Validation and long-term assessment of an approach for the high
throughput determination of lipophilicity (log P-OW) values using multiplexed, absorbance-based
capillary electrophoresis. J. Pharm. Sci., 93, 916–931 (2004).
249. Tu J., Halsall H.B., Seliskar C.J., Limbach P.A., Arias F., Wehmeyer K.R., Heineman W.R., Estima-
tion of log Pow values for neutral and basic compounds by microchip microemulsion electrokinetic
chromatography with indirect fluorimetric detection (µMEEKC-IFD). J. Pharm. Biomed. Anal., 38,
1–7 (2005).
250. Yang S., Bumgarner J.G., Kruk L.F.R., Khaledi M.G., Quantitative structure–activity relationships
studies with micellar electrokinetic chromatography influence of surfactant type and mixed micelles
on estimation of hydrophobicity and bioavailability. J. Chromatogr. A, 721, 323–335 (1996).
251. Quaglia M.G., Fanali S., Barbato F., Donati E., Micellar electrokinetic chromatography for
determination of drug partition in phospholipids. Il Farmaco, 60, 77–83 (2005).
252. Kelly M.A., Altria K.D., Clark B.J., Quantitative analysis of sodium dodecyl sulphate by capillary
electrophoresis. J.Chromatogr. A, 781, 67–71 (1997).
253. Fritz J.S., Recent developments in the separation of inorganic and small organic ions by capillary
254. Gong X.Y., Shen Y., Mao B., Quantitation of ammonium cations in pharmaceutical samples by non-
aqueous capillary electrophoresis with indirect UV detection. J. Liq. Chrom. Rel. Tech., 27, 661–675
(2004).
255. Padarauskas A., CE determination of small ions: Methods and techniques. Anal. Bioanal. Chem., 384,
132–144 (2006).
256. Timerbaev A.R., Capillary electrophoresis of inorganic ions: An update. Electrophoresis, 25, 4008–
4031 (2004).
257. Fabre H., Blanchin M.D., Bosc N., Capillary electrophoresis for the determination of bromide, chloride
and sulfate as impurities in calcium acamprosate. Anal. Chim. Acta, 381, 29–37 (1999).
258. Zhou L., Dovletoglou A., Practical capillary electrophoresis method for the quantitation of the acetate
counter-ion in a novel antifungal lipopeptide. J. Chromatogr. A, 763, 279–284 (1997).
259. Chen D.H., Klopchin P., Parsons J., Srivatsa G.S., Determination of sodium acetate in antisense
oligonucleotides by capillary zone electrophoresis. J. Liq. Chrom. Rel. Tech., 20, 1185–1195
(1997).
260. Altria K.D., Goodall D.M., Rogan M.M., Quantitative determination of drug counter-ion stoichiometry
by capillary electrophoresis. Chromatographia, 38, 637–642 (1994).
261. Altria K.D., Assi K.H., Bryant S.M., Clark B.J., Determination of organic acid drug counter-ions by
capillary electrophoresis, Chromatographia, 44, 367–371 (1997).
262. Nair J.B., Izzo C.J., Anion screening for drugs and intermediates by capillary ion electrophoresis.
J. Chromatogr. A, 640, 445–461 (1993).
263. Moore D.E., Miao W.G., Benikos C., Quantitative determination of alginic acid in pharmaceutical
formulations using capillary electrophoresis. J. Pharm. Biomed. Anal., 34, 233–238 (2004).
264. Stålberg O., Sander K., Sänger-van de Griend C., The determination of bromide in a local anaesthetic
hydrochloride by capillary electrophoresis using direct UV detection. J. Chromatogr. A, 977, 265–275
(2002).
265. Malsch R., Harenberg J., Purity of glycosaminoglycan-related compounds using capillary elec-
trophoresis. Electrophoresis, 17, 401–405 (1996).
266. Fabre H., Blanchin M.D., Julien E., Segonds C., Mandrou B., Bosc N., Validation of a capillary
electrophoresis procedure for the determination of calcium in calcium acamprosate. J. Chromatogr. A,
772, 265–269 (1997).
267. Johnson B.D., Grinberg N., Bicker G., Ellison D., The quantitation of a residual quaternary amine in
bulk drug and process streams using capillary electrophoresis. J. Liq. Chrom. Rel. Tech., 20, 257–272
(1997).
268. Williams R.C., Boucher R.J., Analysis of potassium counter ion and inorganic cation impurities in
pharmaceutical drug substance by capillary electrophoresis with conductivity detection. J. Pharm.
Biomed. Anal., 22, 115–122 (2000).
269. Denis C.M., Baryla N.E., Determination of piperazine in pharmaceutical drug substances using
capillary electrophoresis with indirect UV detection. J. Chromatogr. A, 1110, 268–271 (2006).
270. Altria K.D., Background theory and applications of microemulsion electrokinetic chromatography.
J. Chromatogr. A, 892, 171–186 (2000).
271. Altria K.D., Mahuzier P.E., Clark B.J., Background and operating parameters in microemulsion
electrokinetic chromatography. Electrophoresis, 24, 315–324 (2003).
272. Klampfl C.W., Solvent effects in microemulsion electrokinetic chromatography. Electrophoresis, 24,
1537–1543 (2003).
273. Altria K.D., Clark B.J., Filbey S.D., Kelly M.A., Rudd D.R., Application of chemometric experimental
designs in capillary electrophoresis—A review. Electrophoresis, 16, 2413–2148 (1995).
274. Sentellas S., Saurina J., Chemometrics in capillary electrophoresis. Part A: Methods for optimization.
J. Sep. Sci., 26, 875–885 (2003).
275. Sentellas S., Saurina J., Chemometrics in capillary electrophoresis. Part B: Methods for data analysis.
J. Sep. Sci., 26, 1395–1402 (2003).
276. Veuthey J.L., Rudaz S., Statistical and chemometric tools applied to pharmaceutical analysis. Chimia,
59, 326–330 (2005).
277. Marengo E., Robotti E., Bobba M., Liparota M.C., Artificial neural networks applications in the field
of separation science optimisation. Curr. Anal. Chem., 2, 181–194 (2006).
278. Mahuzier P.E., Clark B.J., Bryant S.M., Altria K.D., High-speed microemulsion electrokinetic
chromatography. Electrophoresis, 22, 3819–3823 (2001).
279. Labat L., Kummer E., Dallet P., Dubost J.P., Comparison of high-performance liquid chromatography
and capillary zone electrophoresis for the determination of parabens in a cosmetic product. J. Pharm.
Biomed. Anal., 23, 763–769 (2000).
280. Clohs L., McErlane K.M., Comparison between capillary electrophoresis and high-performance liquid
chromatography for the stereoselective analysis of carvedilol in serum. J. Pharm. Biomed. Anal., 31,
407–412 (2003).
281. Velikinac I., Cudina I., Jankovic I., Agbaba D., Vladimirov S., Comparison of capillary zone elec-
trophoresis and high performance liquid chromatography methods for quantitative determination of
ketoconazole in drug formulations. Il Farmaco, 59, 419–424 (2004).
282. Aurora-Prado M.S., Steppe M., Tavares M.F.M, Kedor-Hackmann E.R.M, Santoro M.I.R.M., Com-
parison of capillary electrophoresis and reversed-phase liquid chromatography methodologies for
determination of diazepam in pharmaceutical tablets. J. Pharm. Biomed. Anal., 37, 273–279 (2005).
283. Chapman J., Chen F.T.A., Implementing a Generic Methods Development Strategy for Enantiomer
Analysis. LCGC 19, 427–431 (2001).
284. Fabre H., Altria K.D., Validating CE methods for pharmaceutical analysis. LCGC 14, 302–310 (2001).
285. Eurachem Guide, ‘The Fitness for Purpose of Analytical Methods’ (1998).
286. Text on validation of analytical procedures: methodology. International Conference on Harmonization
of Technical Requirements for the Registration of Pharmaceuticals for Human Use (ICH), IFPMA Ed.,
(Geneva, Switzerland, 1996).
287. Hansen H.K., Hansen S.H., Kraunsøe M., Petersen G.M., Comparison of high-performance liquid
chromatography and capillary electrophoresis methods for quantitative determination of glycyrrhizinic
acid in pharmaceutical preparations. Eur. J. Pharm. Sci., 9, 41–46 (1999).
5 Principles and
Practice of Capillary
Electrochromatography
Myra T. Koesdjojo, Carlos F. Gonzalez, and
Vincent T. Remcho
CONTENTS
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

5.2 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
5.2.1 Electroosmotic Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
5.2.2 Zone Broadening: CEC versus HPLC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
5.3 Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
5.3.1 Open Tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
5.3.2 Packed Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
5.3.3 Monolithic Columns. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
5.3.3.1 Molecularly Imprinted Polymer Sorbents for CEC . . . . . . . . . . . . . . . . . . . . . 193
5.3.3.2 Particle Entrapment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
5.4 Developing a CEC Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
5.4.1 Sorbents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
5.4.2 Separation Buffer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
5.4.2.1 pH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
5.4.2.2 Ionic Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
5.4.3 Organic Modifier Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
5.4.4 Influence of Temperature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
5.4.5 Voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
5.4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
5.5 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
5.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
5.1 INTRODUCTION
Capillary electrochromatography (CEC) is a hybrid of capillary electrophoresis (CE) and capillary
high-performance liquid chromatography (HPLC). As such, it has been characterized by some as
exhibiting the best of both technologies and by others as a composite of the worst attributes of
both—the reality lies somewhere between the extremes. Shifting fortunes and interest in CEC are
well represented in Figure 5.1, which illustrates a best estimate of the number of publications on the
topic over the past decade. What is evident is (1) a boom attributable to the promise of highly efficient
liquid phase chromatographic separations and (2) a decline in interest due to practical issues such as
the irreproducibility of electroosmotic flow (EOF) (also a factor in the limited market impact of CE)
183
CEC Publications over the lastdecade

250
200
Number of publications
150
100
50
0
96 97 98 999 000 001 002 003 004 005 006
19 19 19 1 2 2 2 2 2 2 2
Year
FIGURE 5.1 A bar chart showing the number of publications on the topic of CEC per year for the last decade.
and a rediscovery of ultra high pressure liquid chromatography (UPLC). It is possible that there will
be a resurgence of interest arising at least in part from the growth of microchip-based separations
technology. CEC is a good fit on the microfluidic platform owing to its compatibility with small-
pore-size monolithic media and fine particulate packings (due to the lack of backpressure limitations
in the absence of pressure-driven bulk flow), its relative independence from mechanical valves and
pumps, the inherent ease of miniaturizing electrical and electronic components, and its facility to
be interfaced to other chip-based unit operations such as solid phase extraction, filtration, optical
spectroscopic detection, and electrochemistry. This chapter describes the basic theoretical concepts
and operational principles of CEC. The discussion is intended to provide sufficient background on
CEC instrumentation, the principles of CEC operation, and on the range of applications of CEC to
equip the reader to employ CEC in the laboratory.
5.2 THEORY
Capillary electrochromatography is a hybrid separation method that couples CE with HPLC. It
combines the high separation efficiency that capillary zone electrophoresis (CZE) offers with the
wide range of parameters that can be manipulated in HPLC, particularly the wide range of stationary
phases from which to choose. Subsequently, CEC has become a powerful technique that has gained
interest in the last few years.
5.2.1 ELECTROOSMOTIC FLOW

Capillary electrochromatography utilizes an electric field rather than a pressure gradient to propel
the mobile phase through a packed bed. High efficiencies can be achieved in CEC mainly because
of the use of EOF. The flow profile of EOF is pluglike, unlike the parabolic flow of pressure-driven
HPLC, making for less dispersion and therefore higher efficiency than pressure-driven separations
under otherwise identical conditions. Another benefit of EOF is that it generates no pressure drop,
thus, small diameter packings (0.5–2 µm) can be used, thereby further enhancing efficiency. In CEC,
separations are generally carried out using aqueous buffers with organic modifiers in fused-silica
capillaries of 25–100 µm internal diameter (i.d.) packed with small particles (<5 µm) in an applied
electric field.
Principles and Practice of Capillary Electrochromatography 185
EOF Laminar flow
FIGURE 5.2 Flow profiles of EOF and laminar flow.
O– + + –
– –
+
O– +
+
+
Fused silica capillary walls
+
O –
– + –
+
+ +
O– + + –
– +
O– + –
+ + +
–
+ –
O– +
+ + –
O– + –
+ + –
O– + +
Bulk solution
Stern layer Gouy layer
FIGURE 5.3 Schematic representation of the electrical double layer at a negatively charged capillary wall.
CEC was first introduced more than 30 years ago (1974) when Pretorius et al. [1] first
demonstrated the use of EOF to drive a mobile phase through a liquid chromatography (LC) col-
umn. Pretorius used 75–125 µm particles in a 1-mm i.d. glass tube and was able to show that
band broadening with EOF was considerably smaller than with pressure-driven flow. In 1981,
Jorgenson and Lukacs [2] further proved the principle of electrically driven chromatography using
small-diameter particles in fused-silica capillaries. Although the efficiencies were relatively low
(∼60,000 plates m−1 ), they were able to separate 9-methylanthracene from perylene on a 170-µm
i.d. capillary packed with 10-µm reverse-phase packing material. The work that finally resurrected
the interest in CEC, however, was pioneered by Knox and Grant [3] in 1987 and 1991 with their
theoretical and practical approach that demonstrated the practical feasibility of CEC.
The stark contrast of EOF to the parabolic laminar flow profile generated by an external pump
used in HPLC is shown in Figure 5.2. The flat profile of EOF originates in the evenly distributed
charge on the capillary wall, or packing material surface, at which an electrochemical double layer
is formed. On application of a longitudinal potential, a shear plane develops and ions, their solvation
layer, and adjacent species slip toward the outlet with a uniform velocity cross section. This contrasts
with pressure-driven flow, as in HPLC, in which frictional forces at the column walls lead to laminar
flow profile. That yields broad peaks due to the dispersive nature of the broad range in flow velocities.
With a flat profile, zone broadening is minimized, leading to high separation efficiencies in CEC.
The uniform EOF velocity (in channels ∼0.1 to ∼150 µm in diameter) is independent of the channel
width.
To understand the generation of EOF inside the capillary, Stern’s model, as presented in
Figure 5.3, is discussed. The inner wall of a fused-silica capillary possesses ionizable silanol groups
on the surface. In a buffer filled capillary, these silanol groups dissociate to give the capillary wall
a negative charge (SiO− ). This is centered by cations from the buffer solution, giving rise to an
electrical double layer. The layer of ions closest to the surface is called the “Stern layer” and is
essentially rigid or immobile. A more diffuse layer is formed distal to the Stern layer and is known
as the “Gouy layer.” This mobile layer is rich in cations and decays into the bulk liquid. The double
layer is formed at the capillary wall and the surface of the particulate packing material.
The formation of this electrical double layer gives rise to a potential that falls off exponentially
as a function of distance from the capillary surface. This zeta potential (ζ ) has values ranging from 0
to 100 mV. The distance between the Stern layer and a point in the bulk liquid at which the potential
is 0.37 times the potential at the interface between Stern and diffuse layer is defined as the thickness
of the double layer (δ).
The equation describing δ (the thickness of double layer) g [3] is
δ = [εr ε0 RT /2cF 2 ]1/2 , (5.1)
where δ is the thickness of double layer, εr the dielectric constant or relative permittivity of the
eluent, ε0 the permittivity of a vacuum, R the universal gas constant, T the absolute temperature,
c the molar concentration, and F the Faraday constant.
Solving Equation 5.1, for a monovalent electrolyte at a concentration of 0.001 M in water
(εr = 80), the thickness of the electrical double layer would be 10 nm, and at a concentration of
0.1 M it would be 1 nm for the eluent/sorbent and eluent/capillary wall interfaces.
The zeta potential (at the shear plane) is dependent on δ and the charge density σ [4]
δσ
ζ = . (5.2)
ε0 ε r
The relationship between the EOF linear velocity (ueo ) and ζ is given by the Smoluchowski
equation
ε0 εr ζ E
ueo = , (5.3)
η
where E is the electric field strength and η is the viscosity of the solvent.
It is now apparent that the EOF depends upon the surface charge density, the field strength, the
thickness of the electrical double layer, and the viscosity of the separation medium, which in turn is
dependent upon the temperature. In a packed capillary, the vast majority of the total surface area is
contributed by the packing material, thus it is reasonable to approximate the surface charge density
of the system (both the capillary wall and the packing) using the value for the sorbent.
When a voltage is applied longitudinally along the capillary, cations in the diffuse (Gouy) layer
are free to migrate toward the cathode, carrying the bulk solution with them. The result is a net flow
in the direction of the cathode.
A main factor affecting electroosmotic mobility is buffer pH. EOF will be significantly greater at
high pH compared to low pH. At high pH (pH > 9), silanols on the capillary surface are completely
ionized and thus, EOF mobility is at its greatest. At low pH (pH < 4), however, the ionization of
silanols is low and the EOF mobility is nearly insignificant. The ionic strength of the buffer will also
affect mobility.
A solute’s apparent or effective mobility (µx ) takes into account both its individual elec-
trophoretic mobility (µep ) and EOF mobility (µeo ):
µx = µep + µeo . (5.4)
Under “normal” polarity with a fused-silica capillary, samples are introduced at the anode and
EOF migrates toward the cathode. In this case, cations have positive µep , neutrals have zero µep , and
(a)
-
+
+ Uep Ueo –
Electrical double layer
(b)
Ueo
+ –
Each particle bears its

own electrical double
layer
FIGURE 5.4 Schematic representation of (a) CZE and (b) CEC inside a capillary column. (From Knox, J.H.
and Grant, I.H., Chromatographia, 26, 329, 1988. With permission.)
anions have negative µep . In other words, cations migrate faster than the EOF and anions migrate
more slowly than the EOF, whereas neutrals migrate with the same velocity as the EOF. Thus, to
determine EOF velocity experimentally, a neutral and chromatographically unretained marker can
be injected into the capillary under a given set of conditions. Often, acetone is used as a marker in
reversed-phase CEC. By measuring the time that it takes for the neutral unretained marker to reach
the detector, EOF can be determined on the basis of the distance traveled.
As discussed above, the origin of EOF is the electrical double layer that exists at the interface
of a charged surface in contact with an electrolyte solution, as in the case of CZE (Figure 5.4a). In
a capillary packed with silica particles, the surface of both the capillary wall and of the particles
are negatively charged owing to the dissociation of silanol groups (Figure 5.4b). Again, the surface
areas of silica-based packing materials are much greater than that of the capillary wall, thus most of
the EOF is generated by the surface silanol groups of the packing material. When an axial electrical
field is applied to the column, ions in the diffuse section of the double layer migrate towards the
cathode moving the bulk solution by viscous drag.
Looking back at Equation 5.3, in an open tubular (OT) capillary with thin double layer and in the
absence of significant polarization, ζ is the zeta potential of the wall ζw , and is defined as potential
on a hypothetical surface of shear close to the tube wall. For porous/nonporous packing particles
that are nonconducting, ζ is zeta potential at the particle surface ζp .
For OT fused-silica capillaries, the ζ potential has been determined to be between 20 and
120 mV depending on the pH. The potential ζ in capillaries made of polymeric materials
(i.e., polytetrafluorethylene, polyethylene, or polyvinylchloride) ranges from 0 to 60 mV at pH 6.0 and
10, respectively. The potential ζ of C18 derivatized particles can be assumed to be in the same range
as that of polymeric materials [5]. Figure 5.5 shows the EOF velocities versus the electric field
strength measured for a C18 packing material (CEC Hypersil, C18, 3 µm). Depending on the sur-
face charge of the packing materials and the pH of the buffer, linear flow rates of as much as 2.5 mm
s−1 can be achieved in a 35-cm long column.
5.2.2 ZONE BROADENING: CEC VERSUS HPLC

Electroosmotic flow-driven chromatography yields higher separation efficiencies than HPLC because
of the use of small particles and reduction of plate heights as a result of the plug-flow profile.
2
1.8
EOF velocity (mm/a)

1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
0 200 400 600 800 1000
Field strength (V/cm)
FIGURE 5.5 Plot of EOF velocity vs. the applied electric strength. Conditions: 250 (335) mM × 0.1 mM CEC
Hypersil C18, 2.5 µm, 80% ACN/20% MES 25 mM pH 6, 10 bar pressure applied to both ends of capillary,
20◦ C. (Reprinted from Landers, J.P, Handbook of Capillary Electrophoresis, 2nd ed., CRC Press, FL, 1997,
Chapter 5. With permission.)
The linear velocity in a pressure-driven system is given by the following equation described by
Knox and Grant [3]:
dp2 p
u= , (5.5)
ηL
where u is the linear velocity, dp the particle diameter, the flow resistance parameter for packed
columns, p the pressure drop across the column, L the column length, and η the viscosity of the
solvent.
On the other hand, the linear velocity through a packed capillary under an applied electric field
is given by the Smoluchowski equation, as discussed above in Equation 5.3.
By comparing the two Equations 5.3 and 5.5, it can be seen that the linear velocity u is proportional
to dp2 in a pressure-driven system while it is independent of the particle diameter in an electrically
driven system. Since plate height values are generally lowered as a result of using small diameter
particles, it is possible in electrically driven systems to use very small diameter packing materials
and still maintain high linear velocities to yield rapid and very efficient separations.
Since electromatography uses a stationary phase just like in conventional LC, the principles of
band broadening in CEC and LC are similar. The plate height can be expressed using a modified Van
Deemter equation:
1/3 2
udp B Cudp
H = Adp + + , (5.6)
Dm u Dm
where A, B, and C are constants, and Dm is the diffusion coefficient.

The A term refers to eddy diffusion that arises from the different flow paths the solute molecules
may traverse through the packed bed. Band broadening occurs because molecules in different flow
streams are moving at different velocities. Solute molecules move faster (on average) in wider paths
than they do in narrow paths. This term is much more significant in a pressure-driven system, where
the flow rate varies from one channel to another, while in an electrically driven system like CEC, the
contribution of the A term to zone broadening is significantly lower, because of the closely matched
velocities between the channels (in the absence of electrochemical double layer overlap). The C term
represents the contribution to the plate height resulting from resistance to mass transfer between the
mobile phase and the stationary phase. The effect is greater as the mobile phase velocity increases
(a)
Particle
Channel
Linear velocity
(b)
Particle
Channel
Linear velocity
FIGURE 5.6 Schematic representation of (a) electroosmotic and (b) pressure-driven flow in a CEC packed
column. (From Knox, J.H. and Grant, I.H., Chromatographia, 26, 329, 1988. With permission.)
owing to the diminished equilibration time. Just as the use of small diameter particles reduces the
A term contribution to zone broadening, the C term contribution can be minimized by using smaller
packing materials.
In a recent study by Horváth and coworkers [6], band broadening between electrically driven
and pressure-driven flows in several packed capillaries were examined. Their results suggested that
detection and injection contributions to band spreading are negligible. The experimental data were
examined and fitted to the simplified Van Deemter equation to evaluate the eddy diffusion (A) and
the mass transfer (C) parameters of each mode of flow:
B
H =A+ + Cu. (5.7)
u
As expected, their column studies showed that the A term was much smaller in CEC than in HPLC.
This was attributed to the plug-like EOF flow profile in CEC as shown in Figure 5.6. This results in
reduced multipath dispersion effects by a factor 2–4. The C term in HPLC was shown to be greater
than for CEC using packing materials with pore sizes ≥300 Å, but was not observed in packing
materials with average pore size of 80 Å. This was attributed to the absence of EOF transport
through the pores of the particles because of double layer overlap. The intraparticle resistance to
mass transfer contribution to plate height, Hi , was evaluated using

θ(k0 + k + k0 k)2 d Deff µd
Hi = , (5.8)
30 k0 (1 + k0 )2 (1 + k) Dapp Dm
where θ is the tortuosity of the support, d is the diameter of particle, Dm is the diffusivity of the
solute in the mobile phase, µ is the interstitial mobile phase velocity, k is the retention factor, Deff
is the effective molecular diffusivity in the pores, and Dapp is the apparent diffusivity for transport
in the porous particle by diffusion and by intraparticle convective transport. The retention factor, k,
(a) (b) (c)

20
15 µ-HPLC
Parameter C (mS)
µ-HPLC
10
µ-HPLC
5
CEC CEC
CEC
0
0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25
Buffer concentration (mM)
FIGURE 5.7 C term vs. the concentration of electrolyte in solution evaluated for HPLC and CEC. (Reprinted
from Wen, E., et al., J. Chromatogr. A, 855, 349–366, 1999. With permission from Elsevier.)
is given by
εi (1 − εi )
k= , (5.9)
εθ
where εi is the intraparticle porosity and εθ is interstitial porosity. Figure 5.7 shows how the C term
can be reduced by intraparticle EOF for large particles. It was also shown that the contribution of
the C term is reduced as the concentration of electrolytes in solution is increased.
In an additional study on the differences of zone broadening in CEC versus HPLC, Dittmann
and coworkers modified the Horvath model for HPLC to account for the effects of EOF [3,7].
H = Hdisp + He,diff + Hi,diff + Ht,diff + Hkin , (5.10)
where Hdisp is the plate height increment due to the axial dispersion of the solute in interstitial space,
He,diff is the plate height increment resulting from film resistance at the particle boundary, Hi,diff is
the plate height contribution from interparticle diffusion, Ht,diff is the plate height contribution of
transchannel mass transfer, and Hkin is the plate height contribution from the interaction between
the solute and the stationary phase. The sources of band broadening from He,diff , Hi,diff , and Hkin ,
are shown in Figure 5.8 [8]. In Equation 5.10, height equivalent to a theoretical plate (HETP) is
expressed as the sum of independent terms that contribute to zone broadening. In a packed column,
dispersion in the axial direction is expressed as
Hdisp = Ha,diff + Heddy,diff , (5.11)
where Ha,diff is the plate height contribution arising from static diffusion in the axial direction, and
Heddy,diff is the plate height contribution due to the differences in axial transport velocity (eddy
diffusion).
In Equation 5.10, it is assumed that the contributions of all individual terms are independent of
flow profile except for the Heddy,diff and Ht,diff terms. Because of the plug-like flow velocity profile,
the contribution of transchannel diffusion is much smaller in an EOF-driven system (Figure 5.6).
The contribution of transchannel diffusion to the total HETP in a packed column is very small owing
to the small channel diameter; approximately one-sixth of the particle diameter. Flow velocity in
the channels between particles determines the magnitude of the eddy diffusion term contribution. In
Intraparticle diffusion (Hiam)
Film diffusion (Hiam)
Interaction
kinetics (Hiam)
FIGURE 5.8 Sources of dispersion in liquid chromatography. (Reprinted from Bartle, K.D. and Myers, P.,
J. Chromatogr. A, 916, 3–23, 2001. With permission from Elsevier.)
a pressure-driven system velocity varies with the diameter of the channels, while in an electrically
driven system, flow velocity is largely independent of the width of the channel. Therefore, solutes
exchanging between one channel and another do not usually experience a difference in flow velocity
as they would in a pressure-driven system, resulting in much lower dispersion.
5.3 COLUMNS
The packing in a CEC column plays the key role of providing sites for the sorptive interactions (as in
HPLC) as well as supporting EOF. The development of both particulate packings having properties
tuned for CEC, and alternative column technologies such as monoliths, is emerging rapidly.
CEC columns are generally made of fused-silica tubing, usually packed with the appropriate
stationary phase. Today, the most commonly used CEC columns have i.d. of 100 µm or less, with
50 and 75 µm i.d. being the most popular. The stationary phase is retained in the column by two frits.
Column designs can be categorized into two major types: OT columns and packed structures, which
include packed columns, monolithic columns, and microfabricated structures (open or continuous
beds). Packed capillary columns are most commonly used, as has been demonstrated in numerous
papers [9–11]. They can be subdivided into three different categories: columns packed with particles,
columns with continuous beds fabricated in situ creating a rod-like monolithic structure, and columns
with immobilized or entrapped particulate materials.
5.3.1 OPEN TUBES

Open tubular CEC (OTCEC) is similar to CE, with the exception of having a stationary phase attached
on the wall of the fused-silica capillary. OT columns offer some advantages over packed capillaries
[9]. OT columns with inner diameters of 10 µm or less can generate plate height lower than those
for packed columns with the same i.d. The limited band broadening in OT columns results from
the absence of packing material and end frits. The small diameter OT column also allows a faster
separation with higher field strength without causing significant Joule heating. The drawbacks of
Fabrication of an end-frit
Packing at a high pressure, 9000 psi
Formation of a retaining frit, 9000 psi
Excess of stationary phase is flushed out
Detection window formation
FIGURE 5.9 Schematic of capillary packing procedure.
OTCEC columns include difficulties associated with sample injection volume (due to the extremely
small loading capacity) and detection. Because of the low nL to pL sample size range, and the small
i.d. and therefore path length, optical detection can be difficult. This has restricted the growth of
application of OTCEC.
5.3.2 PACKED COLUMNS

Capillary electrochromatography capillaries are usually packed with 3–5 µm C18 or C8 packing
materials. A packed column typically consists of two parts, a packed and an OT section as illustrated
in Figure 5.9. The preparation of these columns includes the fabrication of retaining frits within a
capillary and the packing of small diameter particles into narrow-bore capillaries. The retaining frits
are typically made by filling a small section of the column with silica, then, sintering to produce two
porous plugs. After this temporary frit is formed, stationary phase is slurry packed into the column
at high pressure. Once packed, a second retaining frit is sintered in place with a heating element,
and the excess stationary phase is flushed out by applying pressure in the reverse direction. Packed
columns provide higher column capacity than the OT columns. Most columns for CEC have been
packed with reversed-phase materials.
Both these steps, fritting and packing, present technical difficulties and therefore reproducibility
remains problematic. The use of retaining frits in fabricating a packed column presents several issues.
The fabrication of frits causes the removal of the protective polyimide coating and causes the column
to be fragile at the frit. Nonreproducibility in manufacturing the frits is also another drawback of this
approach. Moreover, the heat applied in the process of making the frits changes the characteristics
of the packing materials at the frits. The difference in surface chemistry leads to bubble formation
at the interface between the frits and the unpacked segments of the capillary. These limitations have
led to the development of various alternative approaches.
5.3.3 MONOLITHIC COLUMNS

Fritless monolithic columns, which emerged during the last decade [4,12–19], have proven to
be a viable alternative to packed capillaries for CEC. They are prepared by in situ polymeriza-
tion to form a continuous rod-like porous bed, thereby, eliminating the difficulties encountered
with packed columns. The porous monoliths can be either in rigid structures [12,16,19–31] or soft
gels [4,13,24,32–38]. Monolithic columns have special characteristics and present several advan-
tages, mainly in the ease of their preparation [39–41]. These advantages include the following:
(1) the polymerization process is simple and can be performed directly within the confines of
a capillary or a microfluidic chip, thus avoiding the problems related to both frit formation and
packing; (2) columns of virtually any length and shape can be fabricated; (3) the polymerization
mixture can be prepared using a wide variety of monomers, allowing a nearly unlimited choice of
1) FILLING 2) POLYMERIZATION 3) WASHING
Capillary
Washing solvent
1) Attach
1) Seal to pump
2) Heat 2) Wash
Polymerization Thermostated
Syringe pump
mixture bath
FIGURE 5.10 Schematic representation of the preparation of a monolithic column. (Reprinted from Svec, F.,
et al., J. Chromatogr. A, 887, 3–29, 2000. With permission from Elsevier.)
both matrix and surface chemistries; and (4) the polymerization process can be easily controlled,
which enables optimization of the porous properties of the monolith, and consequently the flow
rate and chromatographic efficiency of the system. Studies conducted by Luo and Andrade [42]
suggest that continuous polymeric beds with submicron channels may be the ideal packing struc-
ture for CEC. With greatly reduced geometric tortuosity and more favorable EOF, this approach
should simplify column technology and alleviate other problems associated with CEC (i.e., bubble
formation) [43].
Polymeric monolithic columns have been widely developed and successfully applied for CEC. As
can be seen in Figure 5.10 [44], the preparation of a monolithic porous polymer sorbent is a simple
and straightforward process. The steps consist of (1) modification of the capillary wall in order
to provide functional groups on the surface that will participate in the subsequent polymerization
process; (2) filling the capillary with the homogeneous polymerization mixture consisting of the
monomers, initiator, and porogenic solvent; (3) initiation of polymerization thermally or by exposure
to ultraviolet (UV) radiation to obtain a rigid monolithic porous polymer; and (4) the removal of
unreacted components, such as the porogenic solvents and any other soluble compounds that remain
in the polymer.
As mentioned before, a large number of available monomers with a wide variety of functionalities
are available giving unlimited choices of surface chemistry. They may be used directly for the
preparation of polymer monoliths and some are shown in Figure 5.11.
5.3.3.1 Molecularly Imprinted Polymer Sorbents for CEC

Molecularly imprinted polymers (MIPs) are tailor-made materials able to selectively recognize an
analyte. These stationary phases have been successfully employed in CEC, particularly when a
high degree of selectivity is required. The first applications of MIPs were as stationary phases in
affinity chromatography [45,46]. Extensive peak broadening and tailing, however, were observed,
especially for the more retained compound (usually the template). It was reported that this was due to
the heterogeneity of binding sites, in terms of both affinity and accessibility, and different association
and dissociation kinetics.
The use of MIPs for CEC gained considerable interest, most likely owing to the need for a highly
effective liquid-based separation system with unique selectivity for predetermined molecular species
in a miniaturized format. This is especially attractive since CEC is known to have more efficiency
than conventional HPLC, and thus CEC-mode separations should lead to improved performance of
imprinted polymers compared to that achieved in conventional LC.
(a)
NH2 H3C NH2

H3C O
CH3
O O O
Methyl methacrylate (MMA) Styrene Acrylamide Methacrylamide
(b)
CH3 O H H
N N
O
O
O O
O CH3
Ethylene glycol dimethacrylate (EGDMA) p-divinylbenzene (DVB) N,N-methylene bisacrylamide (MDAA)
(c)
CH3 O
OH OH
H3C OH S
H
N O
O O CH3
O
Methacrylic acid (MAA) Acrylic acid 2-acrylamido-2-methyl-1-propane sulforic acid (AMPS)
(d)
OMe O
CN CH3
H3C N N CH3
OMe
CH3 CN
Azobisisobutyronitrile (AIBN) 2,2-dimethoxy-2-phenyl-actophenone
FIGURE 5.11 Selection of (a) monomers with functional groups, (b) cross-linkers, (c) charged monomers
for EOF generation, and (d) chemical initiators used for the preparation of polymer monolithic columns.
The first requirement for MIP–CEC is therefore the adaptation of the molecular imprinting
technology to the capillary format. This is not a simple task and the straightforward approach taken
from HPLC experiments by packing irregular MIP particles to prepare MIP sorbents presented some
technical difficulties in the capillary format. Thus, most approaches have reported the synthesis
of MIP stationary phases within the confines of the capillary column. This monolithic format is
attractive but deceptively challenging. While laborious packing procedures, such as frit making as
well as bulk polymerization of MIP with subsequent crushing and sieving steps, are avoided, other
challenges abound.
The first attempt to adapt MIPs to capillary columns was reported using an in situ dispersion
polymerization, which involved the formation of about 10 µm sized agglomerates of 0.5–4 µm
particles [47]. These MIP capillaries were imprinted against l-phenylanaline anilide, pentamidine,
and benzamidine. The results showed low selectivity of these columns and only pentamidine showed
(a)
Imprinted polymer
H
N O Selective
O Extraction recognition
O Polymerization
Initiator, crosslinker, Hydrolyze, Analyte

monomer (s) N reflux HN
O Polymer with shape and OH
O
functional memory
Imprinted polymer Polymer with shape and

(b)
functional memory
Polymerization Extraction
HN
HN
H
O H H
O
O O
FIGURE 5.12 Schematic illustration of (a) covalent and (b) noncovalent imprinting procedures.
a pH-dependent retention that might originate from imprinting. Despite the poor performance of these
columns, this work is of high importance in indicating the feasibility of synthesizing MIPs in situ.
As shown in Figure 5.12, the preparation of MIPs includes the polymerization of suitable
monomers around a template in the presence of an appropriate cross-linker and solvent (porogen).
Following polymerization, the template molecule is removed to leave cavities complementary in
size, shape, and chemical functionality to the analyte. Accordingly, the polymer should be able to
selectively rebind the target analyte.
Two main approaches are used to produce MIPs: the noncovalent [48] and the covalent [49]
approach. In the covalent approach (Figure 5.12a), the functional monomer is covalently bonded to
the template molecule before polymerization. When polymerization is complete, the covalent bonds
between the template molecule and the polymer are cleaved and the template molecule is extracted.
The resulting imprint is then able to recognize and rebind the imprinted analyte via reversible covalent
bonds. However, this technique suffers from lack of generality owing to the difficulties of finding
suitable monomers.
A more general technique is the noncovalent approach. So far, only the noncovalent imprinting
technique has been used for the preparation of MIP phases for CEC. In the noncovalent approach,
monomers interact with the template molecule via hydrogen bonding, ionic bonding, Van der Waals
forces, hydrophobic effects, and so forth. The noncovalent imprinting technique has been dominant
because of its generality and simplicity. A broad range of species have been imprinted during the past
several years [50]. Figure 5.12b shows the procedure for preparation of an MIP using the noncovalent
approach.
With the noncovalent approach, the selectivity of the MIP can be altered by optimizing the
types and quantities of functional monomers, cross-linking monomers and solvents used. Careful
selection of a porogenic solvent is necessary in order to produce a polymer with enough porosity to
assure good permeability. The degree of cross-linking is also important in achievement of stability
of the template–monomer complex during polymerization, and will also affect polymer porosity.
Figure 5.11 shows some of the most common monomers, cross-linkers, and initiators used for the
preparation of monolithic MIPs.
5.3.3.2 Particle Entrapment

In the entrapped monolithic column approach, particles are immobilized inside the capillary column
in a sol–gel matrix [14,16,17,19,51] or by sintering the particles with heat [25] instead of using
retaining frits. The entrapment approach eliminates the problems associated with retaining frits
in conventional packed columns. Moreover, column longevity improves significantly because the
problems caused by the failure of frits can be avoided.
Columns fabricated via the sintering method have reproducible separations with
∼125,000 plates m−1 . The column preparation involves packing the capillary with silica parti-
cles (premodified with the desired stationary phase), a washing step, drying with nitrogen gas, and a
two-step heat treatment at different temperatures (120◦ C and 360◦ C). To reintroduce stationary phase
that might have been damaged during the process, in situ resilanization is required. The multistep
fabrication and stationary phase reattachment process can be time consuming.
There are two means of immobilizing particles in sol–gel matrix. One is to prepare a sol–gel
solution (typically a mixture containing alkoxysilanes, ethanol, and hydrochloric acid), followed by
addition of particulate materials containing the stationary phase to form a suspension. The suspension
is introduced into the column by vacuum or pressure, and then dried to result in an immobilized packed
bed. Studies have shown that columns prepared using this method generate ∼80,000 plates m−1 .
Fabrication of these columns is problematic owing to irreproducibilities in the finished products.
The second approach to particle entrapment is to introduce the solution sol gelled after the column
has been packed. Chirica and Remcho [16] used high pressure to slurry pack the column. Tang et al.
[20,27,51] used supercritical CO2 to pack the column. After packing, the sol–gel solution was forced
into the column and dried. If temporary retaining frits are used, they are eliminated after drying.
These methods generated columns with efficiencies of ∼125,000 plates m−1 . In comparison to the
“one-step” method, less entrapment matrix is required and higher reproducibility and homogeneity
were noted.
5.4 DEVELOPING A CEC METHOD

Developing a CEC method can seem like a daunting task, but by focusing on some key items and
understanding how they are interrelated, the process can become reasonably straightforward. The
easiest path is of course to find and use a preexisting method. These may be found by either conducting
a literature search or contacting a column manufacture. If a method cannot be located, then the analyst
will need to take into consideration: the sorbent, separation buffer, organic modifiers, temperature,
and voltage. Table 5.1 is a compilation of CEC applications by compound class, suitable stationary
phase, particle size, and reference where the method can be found, and may serve as a good starting
point.
5.4.1 SORBENTS
One item to consider when developing a new method is that there are many capillary columns
available to the consumer, and even though the stationary phases may be similar there will be
variations in EOF and selectivity between them. This effect was studied by Dittmann and Rozing
[201] for five reverse-phase sorbents anchored onto silica particles. Figure 5.13 shows the EOF
variation for five C18 stationary phases from the work conducted by Dittmann and Rozing. Since
variations in EOF exist between different columns with similar stationary phases, it is important to
be aware that methods may need to be modified when using a column that is different in even a
seemingly minor way than the one used in creating the original method.
TABLE 5.1
Applications of CEC Classified on the Type of Compounds Separated
Compound Class Suitable Stationary Phase Particle Size References
Pollutants
PAHs C18 and silica 3 and 1 [52]
µm
PAHs Nonporous C18 1.5 µm [53]
PAHs Methacrylate monolith — [54,55]
PAHs Entrapped C18 5 µm [16,56,57]
PAHs 4000 Å Nucleosil C18 7 µm [58]
Triazine herbicides Hypersil C18 3 µm for [59]
Hypersil C8 all three
Spherisorb C6/SCX
Cinosulfuron and byproducts C18 3 µm [60]
Carbonyl 2,4 dinitrophenylhydrazones Spherisorb ODS 1 3 µm [61]
Carbamate insecticides ODS 3 µm [62]
Insecticidal pyrethrin esters Hypersil C 18 3 µm [63]
Pyrethroid insecticides Nucleosil C18 5 µm for [64]
Zorbax C8 all three
J.T. Baker C18
Carbamate and pyrethroid insecticides Nucleosil C18 5 µm for [65]
Zorbax C8 all three
J.T. Baker C18
Pyrethroid pesticides Hypersil ODS 5 µm [66]
Pirimicarb and azoxystrobin pesticides Hypersil C18 3 µm [67]
Phenols C18 3 µm [68]
Mono- and dichlorophenols Silica with methylated β-CDs — [69]
Pentachlorophenol Spherisorb ODS 1 3 µm [70]
Phenols in tobacco smoke Hypersil C 18 3 µm [71]
Nitroaromatic and nitramine explosives Nonporous C18 1.5 µm [72]
Nitroaromatic and nitramine explosives Hypersil C 18 3 µm [73]
Nonporous C18 1.5 µm
Amino acids, peptide, proteins
PTH amino acids Hypersil ODS 3 µm [74]
PTH amino acids Zorbax C18 3.5 µm [75]
PTH amino acids Zorbax C18 sintered 6 µm [25]
Dansylated amino acids ECTFE 10 µm [76]
Tryptophan, tyrosine OT CEC DNA aptamers — [77]
NDA amino acids Methacrylate monolith — [78]
PTH amino acids
Peptides (2–7 amino acids)
Trp-Arg, Arg-Trp OT CEC DNA aptamers — [79]
Dipeptides Methacylate monolith — [80]
Tripeptides OT CEC porphyrin — [81]
Peptides (2–3 amino acids) Methacrylate monolith — [82]
Peptides (2–4 amino acids) Sperisorb ODS 3 µm [83]
SAX/C18 5 µm
Peptides (2–5 amino acids) Methacrylate monolith — [30,84]
Peptides (5 amino acids) Hypersil C18 3 µm [85,86]
Peptides (4–5 amino acids) Gigaporous polymeric SCX 8 µm [87]
Continued
TABLE 5.1
(Continued)
Peptides (4 amino acids) Polyacrylamide poly(ethyleneglycol) — [13]
monolith
Synthetic peptides Hypersil C18 3 µm [88]
Synthetic peptides dipeptides Spherisorb SCX 5 µm [89]
Basic peptides Spherisorb Si 5 µm [90]
Angiotensins Acrylate monolith — [91]
Angiotensins PS/DVB monolith — [15]
Lysozymes, angiotensins OT CEC C18 — [ 92]
Cytochrome c digest Vydac C18 3 µm [93]

Chicken albumin digest, angiotensins
l-threonine, l-tyrosine (IBMA-EDMA-AMPS) monolith — [94]
Myoglobin (horse heart), transferrin C4 monolith — [95]
(human), α-lactalbumin (bovine milk)
β-Lactoglobulin Vydac C18 3 µm [96]
Tryptic digest of transferrin OT CEC Si — [97]
Tryptic digest of ovalbumin Monolith — [98]
Map of cytochrome c Gromsil ODS2 1.5 µm [99]
Tryptic digest of cytochrome c Spherisorb SAX/C6 3 µm [88]
Carbonic anhydrase, α-lactalbumin, Spherisorb S5- W SAX 5 µm [100]
trypsin inhibitor, ovalbumn,
conalbumin, haemoglobin variants
Ribonuclease, insulin, α-lactalbumin Acrylate monolith — [91]
Lysozyme, α-chymotrypsinogen, C18 monolith — [37]
ribonuclease A, cytochrome c
Cytochrome c mixture (horse, tuna, OT CEC cholesteryl — [101]
chicken and bovine)
Cytochrome c mixture (horse, tuna, OT CEC C18 — [102]
chicken and bovine)
Trypsinogen, α-chymotrypsinogen, OT CEC polyaspartic acid — [103]
Lysozyme, α-chymotrypsinogen, OT CEC DVB — [104]
Carbohydrates
Sucrose, saccharin C18 3 µm [105]
Sucralose and related carbohydrates C18 3 µm [106]
Glucose-maltohexose Polyacrylamide poly(ethyleneglycol) — [13]
monolith
p-Nitrophenyl labeled glucopyranosides, Zorbax C18 5 µm [107]
maltooligosaccharides
α- and β-anomers of glucopyranoside
Phenyl-methyl-pyrazolone labeled Hypersil ODS I 5 µm [108]
monosaccharides
Phenyl-methyl-pyrazolone labeled mono Hypersil ODS I 5 µm [109]
and disaccharides
Phenyl-methyl-pyrazolone labeled Aminopropylated Si — [110]
aldopentose and monosaccharides octadecylammonium Si
Nucleotides
AMP, ADP and ATP Nucleosil C18 5 µm [111]
TABLE 5.1
(Continued)
AMP, CMP, GMP, and UMP OT: 4, 8, 12, 18, 22, 26-hexaaza-1, — [112]
15-dioxacyclooctaeicosane
([28]ane-N6 O2 ) derivatized
Adenine, cytosine, guanine, thymine 9-ethyladenine MIP — [113]
Adenosine, cytidine, uridine, Hypersil phenyl 3 µm [114]
guanosine, thymidine
Adenosine, cytidine, inosine, uridine, CEC Hypersil C18 3 µm [115]
guanosine, thymidine
Thymine, cytosine, adenine, guanine, OT-cholesteryl-undecanoate and — [116]
adenosine cyanopentoxybiphenyl derivatized
Nucleic acids, mono-, di- and ODSS 10 µm [117]
triphosphonucleotides ODSS 5 µm
Dinucleotides Nonporous ODSS 2 µm
t RNAs
Purine and pyrimidine bases and ODSS 10 µm [118,119]
their nucleotides
Primicarb and related pyrimidines Hypersil C18 3 µm [120]
Synthetic nucleoside Spherisorb ODS1 3 µm [121]
PAH-DNA adduct products of in ODS 3 µm [122,123]
vitro reactions
Miscellaneous
Vitamin E in vegetable oils C18 3 µm [124]
Flavonoids (hespederin, hesperetin) Hypersil C8 3 µm [125]
Flavonoids (hop acids) Hypersil C18 3 µm [126]
Antraquionones in Rhubarb Hypersil C18 3 µm [127]
Antraquionones in Rhubarb C18 5 µm [128]
Triglycerides of vegetable and fish oil CEC Hypersil C18 3 µm [129–131]
Fatty acids of vegetable and fish oil CEC Hypersil C18 3 µm [130]
Unsaturated fatty acids methyl esters GROM-SIL ODS 3 µm [132]
Glycosphingolipids Porous ODSS 5 µm [133]
Cannabinnoids Hypersil C18 3 µm [134]
Hypersil C8 3 µm
Retinyl esters Nucleosil C18 5 µm [135]
Retinyl esters 7 µm Nucleosil C18 7 µm [136]
Retinyl esters C30 5 µm [ 137]
Carotenoid isomers C30 3 µm [138]

N-Nitrosodiethanolamine in OT-CEC C18 — [ 139]
cosmetics
Aloins, and related constituents of C18 3 µm [140]
aloe
Food colorants and aromatic Nucleosil C18 5 µm [141]
glucoronides
Azo and antraquinone textile dyes Hypersil C18 3 µm [142]
Alkaloids Nucleosil C18 7 µm [20]
Fullurenes C60 and C70 Vydac C18 3 µm [143]
Polystyrene standards Methacrylate monolith — [144]
Continued
TABLE 5.1
(Continued)
Celluloses Nucleosil silica 5 µm [145]
Pharmaceuticals
Steroids
Fluticasone propionate and synthesis Hypersil C18 3 µm [146]
impurities Spherisorb ODS-I 3 µm [147]
Triamcinolone, hydrocortisone, Hypersil C18 3 µm [71,148]
prednisolone, cortisone,
methylprednisolone,
betamethasone, dexamethasone,
adrenosterone, fluocortolone,
triamcinolone acetonide
Tipredane and related substances Spherisorb ODS-I 3 µm [121]
Corticosteroids
Corticosterone, testosterone, Zorbax ODS 6 µm [75]
androsten-3, 17-dione, androstan-3,
17-dione, pregnan-3, 20-dione
Aldosterone, hydrocortisone, Hypersil C18 3 µm [149]
testosterone
Digoxigenin, gitoxigenin, Spherisorb ODS-I 3 µm [150]
cinobufatalin, digitoxigenin,
cinobufagin, bufalin
Hydrocortisone, testosterone, Nonporous Chromspher-ODS 1.5 µm [151]
17-α-methyltestosterone,
progesterone
Estriol, hydrocortisone, estradiol, Zorbax ODS 1.8 µm [152]
estrone, testosterone,
17-α-methyltestosterone,
4-pregnen-20α-ol-3-one,
progesterone
Deesterified steroid, budesonide, Spherisorb ODS-I 3 µm [153]
steroid A
Hydrocortisone, prednisolone, Poly (AMPS-co-IPPAm) hydrogel — [38]
hydrocortisone 21-acetate,
testosterone
Hydrocortisone, prednisolone, Hypersil C18 3 µm [154]
betamethasone, betamethasone
dipropionate, clobetasol butyrate,
fluticasone propionate, clobetase
butyrate,
betamethasone-17-valerate
Aldosterone, dexamethasone, Hypersil C18 3 µm [155]
β-estradiol, testosterone NPS ODS 2 1.5 µm
Desogestrel and analogs, tibolon and Hypersil C18 3 µm [156]
analogs
Dexamethasone, betamethasone Hypersil C18/SCX 3 µm [157]
valerate, fluticasone propionate
Neutral and conjugated steroids Macroporous monolith, C12 — [158]
TABLE 5.1
(Continued)
Corticosteroids and esters Spherisorb ODS-I 3 µm [159]
(hydrocortisone, hydrocortisone
17-butyrate, hydrocortisone
21-acetate, hydrocortisone
17-valerate, hydrocortisone
21-caprylate, hydrocortisone
21-cypionate, hydrocortisone
21-hemisuccinate)
Cholesterol and ester derivatives Hypersil C18 3 µm [160]
Corticosteroids (ouabain, Spherisorb small pore ODS/SCX, sol-gel 3 µm [20]
strophantidin, 4-pregnene- bonded
6b,11b,21-triol-3,20-dione)
Bile acids and conjugates Macroporous monolith, amino (normal — [161]
phase)
Macroporous monolith, C12 (reverse
phase)
Levonorgestrel and racemic Cellulose 5 µm [162]
norgestrel tris(3,5-dichlorophenylcarbamate
coated onto spherical Daisogel
Estrogens (diethylstilbestrol, ODS 3 µm [163,164]
hexestrol, dienestrol)
Estrogens (estriol, estradiol, equiline, GROM-SIL ODS-0 AB 3 µm [132]
estrone)
Benzodiazepines
Nitrazepam and diazepam Hypersil C18 3 µm [148]
Nitrazepam, diazepam, triazolam 3-(1,8-Naphthalimido)propyl-modified — [165]
silyl silica gel
Oxazepam, lorzepam, temazepam, phenyl 3 µm [166]
diazepam, tofizepam
Nitrazepam, nimetazepam, Cholesteryl bonded silica OT, cholesteryl 6.5 µm [166]
estazolam, brotizolam, clonazepam, modified capillary wall
axazolam, haloxazolam,
cloxazolam, medazepam
Temazepam, oxazepam, clonazepam, OT fused silica etched with NH4 HF2 , — [116]
diazepam, nitrazepam chemically bonded
cholesteryl-10-undecanoate
Flunitrazepam, temazepam, PEM-coated capillary — [168]
diazepam, oxazepam, lorazepam,
clonazepam, nitrazepam
Flunitrazepam, temazepam, Reliasil C18 3 µm [169]
diazepam, oxazepam, lorazepam,
clonazepam, nitrazepam
Nonsteroidal anti-inflammatory
drugs
Fenoprofen, ibuprofen, indoprofen, LiChrospher 100 RP-18 5 µm [170]
ketoprofen, naproxen, suprofen
Continued
TABLE 5.1
(Continued)
Indoprofen, ketoprofen, naproxen, OT histidine coated capillary — [171]
ibuprofen, fenoprofen, flurbiprofen,
suprofen
Indoprofen, suprofen, tiaprofen, LiChrospher 100 RP-18 5 µm [172]
ketoprofen, naproxen, fenoprofen,
carprofen, flurbiprofen, cicloprofen,
ibuprofen
Etodolac and five metabolites LiChrospher 100 RP-18 5 µm [173]
Acetaminophen, caffeine Nucleosil C18, sol-gel bonded 5 mm [57]
Ibuprofen, indoprofen, fenoprofen, Acrylamide-based monoliths — [174]
ketoprofen, suprofen, diclofenac,
metenamic acid
Ibuprofen, naproxen, ketoprofen, Methacrylate-based macroporous SAX — [175]
suprofen monolith
Tricyclic antidepressants
Nortriptyline and amitriptyline Nucleosil 50 5 µm for [176]
Nucleosil 100 all
Nucleosil 300
Nucleosil 1000
Nucleosil 4000
Bendroflumethiazide, nortriptyline, Spherisorb ODS-I 3 µm [147]
chlomipramine, methdilazine, Spherisorb SCX 3 µm
imipramine, desipramine
Nortriptyline, Spherisorb ODS-I 3 µm [177]
N-methyl-amitriptyline,
amitriptyline, imipramine,
clomipramine,
N,N-dipropyl-protriptyline
Nortriptyline, amitriptyline, Nucleosil SCX 5 µm [141]
N-methyl-amitriptyline, desipramine, Zorbax SCX 5 µm
imipramine, chlomipramine,
N, N-dimethyl-protriptyline,
N,N-dipropyl-protriptyline
Nortriptyline, imipramine, Spherisorb SCX 3 µm [142]
amitriptyline, clomipramine
Nortriptyline, Continuous-bedpolyacrylamide with — [178]
N-methyl-amitriptyline, various contents of isopropyl and
amitriptyline sulfonate ligands
Nortriptyline, doxepin, imipramine, Nortriptyline MIP — [179]
amitriptyline, trimipramine,
clomipramine
Various pharmaceuticals
Prostaglandins and relates impurities Spherisorb ODS-I 3 µm [146]
Zorbax SBC8 1.8 µm
TABLE 5.1
(Continued)
Neutral related S-oxidation Hypersil C18 3 µm [180]
compounds
2-Phenylmethyl-1-naphthol Hypersil C18 3 µm [181]
p-Hydroxybenzoic acid, bumetanide, Hypersil C18 3 µm [121]
flurbiprofen
Thomapyrin: containing GROM-SIL 100 ODS-0 AB 3 µm [182]
acetaminophen, caffeine and
acetylsalicylic acid
Antiviral drug suramin Nucleosil 100 C18 5 µm [111]
Amino group containing drugs: Micra bare silica 3 µm [183]
codeine phosphate, ephedrine
hydrochroride, thebaine, berberine,
hydrochloride, jatrorrizine
hydrochloride, cocaine
hydrochloride
Isradepin and by-products Hypersil C18 3 µm [184]
Morphine alkaloids Nucleosil 100 C18 5 µm [111]
Barbiturates (barbital, phenobarbital, 3-(1,8-Naphthalimido)propyl-modified — [165]
secobarbital, thiopental) silyl silica gel
Antiepileptic drugs: ethosuccinimide, Spherisorb ODS-I 3 µm [125]
primidon, CBZ-10, 11 diol,
CBZ-10, 11-epoxid, phenytoin,
carbamazepine (CBZ)
Macrocyclic lactone, S541 Factor B CEC Hypersil 3 µm [128]
from Streptamyces S541
Sulfanilamide, sulfaflurazol, Nucleosil 100-5C8 5 µm [185]
sulfadicramide
Cardiac glycosides: digoxigenin, NPS ODS II 1.5 µm [155]
digoxin, digitoxigenin
Tetracyclines Etched and modified C18 OT — [186]
column
2-Phenylethylamine derivatives: Nucleosil 5C8 5 µm [125]
epinephrine, Dopa, 2-amino-3-
hydroxy-3-phenyl-propanol,
ephedrine
Vitamin D2 and D3 GROM-SIL 100 ODS-0 AB 3 µm [132]
Doxorubicin Luna C18 3 µm [187]
Thalidomide and hydroxylated Aminopropyl coated with amylose and 5 µm [188]
metabolites cellulose derivatives
Thalidomide and metabolites LiChrospher 100 RP-18 5 µm [189]
Methylamphetamine and impurities NPS ODS II 1.5 µm [190]
Related opiate compounds NPS ODS II 1.5 µm [191]
(morphine, hydromorphone,
nalorphine, codeine, oxycodone,
diacetylmorphine)
Continued
TABLE 5.1
(Continued)
Theophylline, caffeine, sulfanilamide C18 3 µm [105]
Theophylline, caffeine, Silica 3 µm [192]
aminophylline, theobromine,
β-hydrocyethyltheophylline,
phenylbutazone,
hydrochlorothiazide,
acetominophen
Fluvoxamine and possible isomers Spherisorb ODS-I 3 µm [193]
Hydroquinone and ethers LiChrospher 100 RP-18 5 µm [194]
Serotonin and metabolites OT, fused silica etched with NH4 HF2 , — [116]
chemically bonded
4-cyano-4’-pentoxybiphenyl
Neurotransmitters OT, chemically-modified wall with — [195]
(5-hydroxytryptamine, dopamine, macrocyclic dioxopolyamine
norepinephrine, epinephrine, (dioxo[13]aneN4 )
DOPA)
Clenbuterol, salbutamol, methadone Hypersil MOS 3 µm [196]
Salbutamol, salmeterol Spherisorb C18 3 µm [142]
Spherisorb C6/SCX 3 µm
Spherisorb SCX 3 µm
Carbovir, ranitidine, ondansetron, Symmetry Shield RP-8 3.5 µm [197]
imipramine, amitriptyline,
clomipramine
Benzylamine, nortriptyline, Hypersil C18 3 µm [198]
diphen-hydramine, terbutalin,
procainamide
Benzylamine, nortriptyline, Hypersil silica 3 µm [199]
diphen-hydramine, procainamide Hypersil BDS silica 3 µm
Hypersil silica 3 µm
Basic and acidic drugs Spherisorb ODS-I, Hypersil C18, 3 µm [200]
(metoclopramide, timolol, procain, Hypersil C8
ambroxol, antipyrine, naproxen)
Amphetamine, methamphetamine, Hypersil C8 3 µm [134]
procaine, cocaine, quinine, heroin,
noscapine, Phenobarbital,
methaqualone, diazepam,
testosterone, testosterone
propionate, CBN, d9-THC, d9-THC
acid-A
5.4.2 SEPARATION BUFFER

Proper selection of a separation buffer is crucial because it will help resist changes in pH, is used to
dissolve the sample, will mitigate the interactions between the sample and bonded phase, affect the
EOF, and is the electrical connection between the cathode and anode when voltage is applied during
a separation. There are some important features of the separation buffer that must be considered: pH,
ionic strength, and organic modifiers. Of these, pH and ionic strength will be covered in this section,
and organic modifiers will be discussed separately.
2 3
1
80 BDS-ODS-Hypersil
2 3
70 1
ODS Hypersil
60
2 3
50 1
Spherisorb ODS2
mAU
40
2 3
1
30
Spherisorb ODS1
20
2
10 3
1
CEC Hypersil C18
0
0 2.5 5 7.5 10 12.5 15 17.5 20

min
FIGURE 5.13 Separation of PAHs on five reversed-phase C18 stationary phases. Column 250 (335) mm ×
0.1 mm, 3 µm, mobile phase 80% ACN 20% 50 mM Tris-HCl, pH 8, 20 kV, temperature 20◦ C, 10 bar pressure
applied to both ends of capillary, 20◦ C. Samples were not identical but all contained (1) thiourea, (2) napthalene,
and (3) fluoranthene. (Reprinted from Dittmann, M.M. and Rozing, G.P., J. Microcol. Sep., 9, 399, 1997. With
permission.)
The separation buffer in the column and inlet/outlet vials should also be replaced—ideally after
each run—to prevent buffer depletion. Buffer depletion occurs as the cation and anion components of
the buffer migrate toward cathode and anode respectively during a run. Over time, the concentration
of anions at the cathode and cations at the anode will be reduced. This reduction of ions will also
occur within the buffer in the column itself. When this occurs to an extreme, major performance
degradation is the result.
5.4.2.1 pH
In selecting a separation buffer, the sample to be dissolved must be considered first. It is important
to consider the pKa s of the components of the sample as the pH will determine whether the sample
components are charged or neutral. Euerby et al. [121] showed that by changing buffer pH, and using
a 25-cm effective length 3-µm particle size Hypersil phenyl stationary phase, the migration order
of some of the barbiturates in their mixture changed when the pH was increased from 6.1 to 7.8,
owing to a change in analyte ionization state. For biological samples the pH is not only important in
dictating the charge of the analyte, but also critical in preventing sample denaturation.
EOF is also pH dependent. For example, in columns having silica support particles, ionizable
silanols will cause these columns to exhibit a more pronounced dependency of EOF on pH than
do other columns. The silanols begin to ionize at pH 2 and will continue to ionize progressively
(based on the regiospecificity of a given silanols’ pKa ) until approximately pH 8–9, at which point
the preponderance of silanols are fully ionized. The zeta potential will therefore increase, and it
is for this reason that the electroosmotic velocity increases with increasingly alkaline solutions.
1.6
Linear velocity (mm/sec)

1.4
1.2
0.8
0.6
2 3 4 5 6 7 8
pH
FIGURE 5.14 Effect of mobile phase pH on the linear velocity of the EOF. Comparison of CEC Hypersil
C18 (), C8 (•), and phenyl () packing materials. (Reprinted from Euerby, M.R., et al., J. Microcol. Sep.,
Euerby et al. [121] demonstrated that the linear velocities for three different commercially available
columns increased with rising pH, as shown in Figure 5.14. These Agilent columns were packed
with Hypersil C18, Hypersil C8, and Hypersil phenyl.
A possible solution to the low electroosmotic velocity encountered with standard reverse-phase
columns is to use a mixed-bed column. A mixed-bed column contains both normal-phase (NP) and
reverse-phase (RP) components, and when working at lower pH regions the NP component will
yield EOF from the surface charge associated with the anionic or cationic NP sorbent. Tang and Lee
[20] compared electroosmotic velocity over the pH range of 2–9 for reverse-phase and mixed-bed
columns; both columns were sol–gel bonded continuous-bed columns. The reverse-phase column
contained an ODS1 stationary phase whereas the mixed-bed column was an ODS/SCX mixture. They
noted that the mixed-bed column exhibited a nearly constant EOF for the entire pH range studied,
unlike the reverse-phase column that continually increased as the pH climbed. These findings are
shown in Figure 5.15.
5.4.2.2 Ionic Strength

Another variable to be optimized is the ionic strength of the separation buffer. The ionic strength
not only determines the buffering capacity of the buffer being used, but also affects EOF and Joule
heating.
From Equation 5.3, it seems that there is no relation between EOF velocity and double layer
thickness, but there is a direct relation to the zeta potential. To find the relation between EOF and
ionic strength, Equation 5.1 must be substituted into Equation 5.2 to obtain
[(ε0 εr RT /2F 2 cz2 )1/2 ]σ

ζ = . (5.12)
ε0 εr
Thus the zeta potential will have an inverse relation to the square root of the electrolyte concentration.
Likewise, it can be seen that EOF velocity will be dependent on the inverse square root of the ionic
strength of the electrolyte.
3.0
2.5
EOF velocity (mm s–1)

2.0
1.5
1.0
0.5
0.0
0 1 2 3 4 5 6 7 8 9 10
Buffer pH
FIGURE 5.15 Plots of EOF velocity vs. mobile phase pH for sol–gel bonded continuous-bed columns.
Conditions: 25/34 cm × 75 µm i.d. continuous-bed columns containing sol-gel bonded () 3 µm, 80 Å ODS1,
and () 3 µm, 80 Å ODS/SCX; 70: 25: 5 (v/v/v) ACN/H2 O/50 mM phosphate buffer; 5 kV × 2 s injection;
30 kV applied voltage; 0.3 mM thiourea used as EOF marker. (Reprinted from Tang, Q. and Lee, M.L., J. High
Resol. Chromatogr., 23, 73, 2000. With permission.)
Heating is also a concern when working with high ionic strength buffer solutions. As the ionic
strength of the solution is increased, the current will likewise increase. Utilizing the following
equation for estimating Joule heating (Q)
Q = E 2 c, (5.13)
where E 2 is the voltage gradient, is molar conductivity of the electrolyte, and c is the electrolyte
molar concentration, it can be determined that heating will increase by raising the molar conductivity
of the electrolyte. The changes brought on by heating will be discussed in more detail in Section 5.4.4.
5.4.3 ORGANIC MODIFIER EFFECTS

The effect of organic modifiers such as acetonitrile (ACN), ethanol, isopropanol, methanol, and
tetrahydrofuran on CEC separations is quite complicated. Organic modifiers added to an aqueous
buffer can increase the solubility of neutral and nonpolar compounds, but at the same time yield
some EOF effects in concert with changes in retention factors.
Addition of a modifier will alter the viscosity of the run buffer depending on its interactions with
the aqueous buffer. According to Equation 5.3, there is an inverse relation between electroosmotic
velocity and solution viscosity, and therefore, the electroosmotic velocity will change accordingly
when a modifier is used. In addition, the dielectric constant of the buffer solution will change with
increasing organic solvent concentration, which will affect the zeta potential inversely thus altering
EOF.
An example of addition of organic modifiers decreasing EOF was presented by Kanitsar et al.
[202] for the effect of various organic additives on the separation of three carboxylic acids. Separation
buffers consisting of 15% v/v solvent to pH 6.0, 25 mM KH2 PO4 and 0.75 mM TTAB buffer were
used in conjunction with benzyl alcohol, as the neutral marker, for EOF determination. For the
20
19
µeo(×105 cm2 V1,8–1)

18
17
16
15
20 40 60 80
%ACN
FIGURE 5.16 Effect of the ACN content on EOF mobility in CEC. Electropak phenyl column, EP-75-40-3-
PH [40 cm packed (47 cm total) 75 µm i.d.]; electrolyte: Tris HCl (pH 8)–ACN mixture (ionic strength 5 mM);
temperature: 25◦ C; applied voltage: +20 kV; UV detection: 220 nm; electrokinetic injection: 4 s (+10 kV); EOF
marker: thiourea. (Reprinted from Cahours, X., et al., J. Chromatogr. A, 845, 203–216, 1999. With permission
from Elsevier.)
solvent systems studied, the respective EOF values were determined to be 4.03 × 10−4 cm2 V−1 s−1
for water, 2.71 × 10−4 cm2 V−1 s−1 for ethanol, 2.56 × 10−4 cm2 V−1 s−1 for 1-propanol, 2.29 ×
10−4 cm2 V−1 s−1 for ACN, 2.24 × 10−4 cm2 V−1 s−1 for methanol, and 0.69 × 10−4 cm2 V−1 s−1
for ethylene glycol. An example of increasing EOF was given by Cahours et al. [166] for addition of
ACN to a 5 mM, pH 8 Tris–HCl buffer. They noticed a 17% increase in EOF velocity when the ACN
percentage was increased from 30% to 80%, and a plot of their findings is shown in Figure 5.16.
As with HPLC, organic modifiers will of course affect interactions between analytes and the
stationary phase. With reverse-phase materials, addition of an organic modifier to a buffer will lower
the partition coefficient of an analyte. At an appropriate concentration, it is possible to optimize the
partition function for a given analyte and tune the retention factor (k ). Zhang et al. [91] showed that
by increasing the percentage of ACN in their mobile phase that the retention times of four proteins
separated by RP CEC were not altered significantly. What they did notice, however, is that as the ACN
percentage was increased a reduction in retention attributed to chromatographic interactions occurred,
and at a sufficiently high ACN percentage the separation was largely electrophoretic in nature.
Electrochromatograms for this separation showing ACN percentages in the range of 20%–50% are
given in Figure 5.17.
The organic modifier/buffer species and organic modifier/surface interactions are not fully under-
stood, and as such there is no easy way to summarize the effect of organic modifiers on EOF. These
effects are specific to a given stationary phase/buffer species/buffer concentration combination, and
therefore experimentation and optimization are required.
5.4.4 INFLUENCE OF TEMPERATURE

Joule heating leads to zone spreading and both qualitative and quantitative irreproducibility; to
attenuate Joule heating the column is housed in a temperature-controlled environment. Unregulated
6 10 3
20% ACN 40% ACN 2
2
1
Absorbance at 214 nm [mAU] 1 4
3,4
0 0
0 5 10 0 5 10
6 2 10 2,3
30% ACN 3 50% ACN
1
1
4 4
0 0
0 5 10 0 5 10
Minutes
FIGURE 5.17 Effect of ACN concentration in the eluent on the separation of four proteins. Column, 39 cm
(effective length 29 cm) × 50 µm i.d., fused-silica capillary with porous methacrylic monolith having tertiary
amino functions; mobile phase, ACN (%, v/v) in 60 mM aqueous sodium phosphate, pH 2.5; applied voltage,
−25 kV; detection, 214 nm; sample: (1) ribonuclease A, (2) insulin, (3) α-lactalbumin, and (4) myoglobin.
(Reprinted from Zhang, S., et al., J. Chromatogr. A, 887, 465–477, 2000. With permission from Elsevier.)
Joule heating leads to a cycle of increased resistivity and increased heating. In the worst case scenario,
the buffer will heat until ultimately out-gassing occurs. Most of the CE instrumentation available
today is equipped with heating or cooling capability, but when dealing with a homebuilt system
temperature control should be considered.
EOF velocity will also be changed by fluctuations in temperature. With an increase in temperature
the EOF velocity will increase, owing to a decrease in the viscosity of the buffer system in accordance
with Equation 5.3. Zhang et al. [203] showed that an increase in temperature from 25◦ C to 55◦ C
increased the velocity of the EOF as shown in Figure 5.18. Also, the kinetics of partitioning of
analytes between the stationary and mobile phases will be more rapid which can aid in minimizing
band broadening effects.
5.4.5 VOLTAGE
As with any other separation method, the ultimate goal of developing a method is to have the
shortest analysis time that affords the desired resolution between sample components. In CEC,
reduced analysis times can be gained by working at higher voltages because of the linear relation
between applied voltage and both electrophoretic and electroosmotic velocity. One concern about
working with elevated voltages is that there is increased possibility of excessive Joule heating. Li and
Remcho [204] demonstrated the linear relationship between field strength and the EOF velocity using
acetone as an unretained neutral marker in a packed Nucleosil C18 column (shown in Figure 5.19)
for 3, 5, and 7 µm particles.
5.4.6 SUMMARY
There are numerous variables that can be altered in developing a CEC method, and the vast major-
ity of them are interrelated. This contributed greatly to the decline in interest in CEC in the late
1990s. For example a change in buffer pH may change EOF, the charge state of an analyte, and
3.5
(10–8 m2 V–1 s–1)

EOF mobility
3
2.5
2
3 3.1 3.2 3.3 3.4
1000/T (K–1)
FIGURE 5.18 Plots of the EOF mobility, measured with DMSO as the unretained neutral marker, against the
reciprocal absolute temperature. Mobile phase () 10% ACN, () 20% ACN, () 30% ACN in 60 mM sodium
phosphate buffer, pH 2.5; temperature, 25◦ C; applied voltage, −30 kV; detection, 214 nm. Column, 40 cm
(effective length 30 cm) × 75 µm, fused silica with styrenic monolith having quaternary ammonium functions.
(Reprinted from Zhang, S., et al., J. Chromatogr. A, 914, 189–200, 2001. With permission from Elsevier.)
0.14
0.12 5 µm
3 µm
Linear velocity (mm/sec)
7 µm
0.1
0.08
0.06
0.04
0.02
0
0 100 200 300 400 500 600
Field strength (V/cm)
FIGURE 5.19 Linear velocity as a function of field strength for 3-, 5-, and 7-µm-diameter Nucleosil C18
particles. (Reprinted from Li, D.M. and Remcho, V.T., J. Microcol. Sep., 9, 389, 1997. With permission.)
its partitioning behavior. The ionic strength of a mobile phase can also change the EOF, and cur-
rent will increase as salts are added to the system. If the current is sufficiently high (owing to the
mobile phase ionic strength) heating effects may be encountered, which can lower the partition
coefficients of analytes and the mobile phase viscosity, and may ultimately lead to bubble formation.
Therefore, it is important to consider how a change in one component of the method can alter other
components.
5.5 APPLICATIONS
Capillary electrochromatography can be an effective separation tool for a wide range of sample
types, as evidenced in Table 5.1. In this section, examples of separations utilizing CEC will be
highlighted.
(a) mAU
2
100
80 1
mAU
60
5
250 40
3 4 6
20
200 0
15 20 25 30 35 40 min
150
100
50
0
10 20 30 40 50 min
(b)
mAU 2
7
80
60
40
1
20
5
0
3 4 6
–20
10 20 30 40 50 60 min
FIGURE 5.20 CEC of (a) pyrethrin dip and (b) flea and tick mist. Conditions: MeCN-25 mM Tris–THF
(55:35:10, pH 9). (1) Cinerin II, (2) pyrethrin II, (3) jasmolin II, (4) cinerin I, (5) pyrethrin I, (6) jasmolin I,
and (7) piperonyl butoxide. (Reprinted from Henry, C.W., et al., J. Chromatogr. A, 905, 319–327, 2001. With
permission from Elsevier.)
A CEC separation of insecticidal pyrethrin esters was accomplished by Henry et al. [63]. Here it
was shown that it is possible to identify pyrethrin esters (cinerin I, pyrethrin I, jasmolin I, cinerin II,
pyrethrin II, and jasmolin II) in a commercially available pyrethrin dip and flea and tick mist
(Figure 5.20). The column inner and outer diameters were 100 and 350 µm, respectively, and
the total and effective lengths were 33 and 25 cm, respectively. The column was packed with 3 µm
particle size Hypersil C18. The mobile phase was composed of 55:35:10, pH 9, ACN-25 mM Tris-
tetrahydrofuran. Other parameters were an applied voltage of 30 kV, column temperature of 25◦ C,
and UV detection at 254 nm.
3
100 1 2 4 (a)
50
0
Relative intensity (%)
100 3 (b)
1
2 4
50
100 3
1 (c)
2 4
50
0
0 10 20 30
Time (min)
FIGURE 5.21 Separation of estrogenic compounds: (1) estriol, (2) estradiol, (3) equiline, and (4) estrone.
(a) With pCEC-CIS-MS: 80 bar and +15 kV applied, (b) pCEC-CIS-MS: 80 bar and +5 kV applied, and
(c) CHPLC-CIS-MS: 80 bar applied. (Reprinted from Rentel, C., et al., Electrophoresis, 20, 2329, 1999.
With permission.)
An example of a separation of estrogens (estriol, estradiol, equiline, and estrone) with pressurized
CEC (pCEC) was shown by Rentel et al. [132]. Capillaries with inner diameters of 100 µm and outer
diameters of 360 µm were used. The GROM-SIL ODS-0 AB packing material used had a particle size
of 3 µm. Detection of the estrogenic analytes was accomplished with coordination ion spray mass
spectrometry (CIS-MS). Other experimental parameters were: pH 9, 4 mM ammonium acetate in
5:95 water–ACN mobile phase, applied pressure of 80 bar, and applied voltage of 15 (Figure 5.21a)
and 5 kV (Figure 5.21b).
Tricyclic antidepressants were analyzed by Vallano et al. [179] via an MIP-CEC separation.
In this separation, doxepin, imipramine, amitriptyline, trimipramine, and clomipramine were sep-
arated from the template molecule nortriptyline (Figure 5.22). The capillary inner diameter and
total length were 100 µm and 33 cm, respectively, and the total length of the monolithic bed was
22.5 cm. The eluent was 92:2 ACN: 10 mM sodium acetate pH 3.0 to which 0.02% trifluoracetic
acid and 0.015% triethylamine (TEA) (v/v) was added. A constant applied voltage of 30 kV was
utilized.
CEC has also been used to successfully separate nucleotides. For example, Helboe and Hansen
[115] separated cytidine, uridine, inosine, guanosine, thymidine, adenosine, and thiourea, with a
CEC Hypersil C18 column. Column specifics are as follows: particle size of 3 µm, inner diameter
of 100 µm, and bed length of 25 cm. The mobile phase was 98% 5 mM acetic acid, 3 mM TEA, pH
5: 2% ACN. Figure 5.23, shows the optimization of temperature and voltage, 25◦ C and 20 kV for
5.4A, 25◦ C and 25 kV for 5.4B, and 20◦ C and 25 kV for 5.4C. Nucleotides were also examined by
Cahours et al. [114]. Figure 5.24 illustrates the effect ionic strength has on the migration order of
adenosine (A), cytidine (C), guanosine (G), uridine (U), and thymidine (T), and thiourea, an EOF
marker. For this separation, the column dimensions were 27 cm total length × 75 µm inner diameter.
(1) Doxepin (4) Trmipramine
(2) Imipramine (5) Clomipramine
(3) Amirriptyline
(6) Nortriptyline
(TEMPLATE)
1.5
0 1 2 3
min
FIGURE 5.22 MIP-CEC separation of a simulated combinatorial library consisting of several tricyclic antide-
pressants. Conditions: capillary i.d. 100 µm; Ltot: 33 cm; Lbed: 22.5 cm; eluent:ACN: 10 mM Na acetate pH 3.0
(98:2) with 0.02% trifluoracetic acid and 0.015% TEA (v/v); voltage +30 kV constant; injection: +2 kV, 2 s;
column temperature: 50◦ C. (Reprinted from Vallano, P.T. and Remcho, V.T., J. Chromatogr. A, 887, 125-135,
2000, with permission from Elsevier.)
The effective length of the column was 20 cm and it was packed with 3 µm Phenyl Hypersil. The
separation temperature and voltage were 20◦ C and 20 kV, respectively, and the detection wavelength
was 254 nm. The electrolyte was composed of pH 5 acetic acid/ammonia-ACN (95/5), and the ionic
strengths used were 10 and 5 mM, respectively.
Pharmaceuticals have also been separated by CEC. Taylor and Teale [148] conducted gradi-
ent CEC separations of drug mixtures. In their study, a mixture of corticosteroids consisting of
triamcinolone, hydrocortisone, prednisolone, cortisone, methylprednisolone, betamethasone, dex-
amethasone, adrenosterone, fluocortolone, and tramcinolone acetonide was examined. This mixture
was separated using a 3-µm Hypersil ODS. The packed length, length to detection window, and total
length of this column were 30, 30.1, and 42 cm, respectively. The applied voltage was 30 kV, and
the detection wavelength used was 240 nm. Gradient elution was used as follows: initially, 5 mM
ammonium acetate in ACN–water (17:83) for the first 3 min, then for 15 min; the ACN concentration
was ramped to 38% and kept constant for the remainder of the analysis. The flow rate for this analysis
was 10 µL min−1 for the first 3 min, and was then elevated to 100 µL min−1 . This separation is
shown in Figure 5.25.
mAU (a)
7 2
80
60 1 3 5
40 4 6
20
0
2 4 6 8 10 12 14 16 18 min
mAU (b)
80
60
40
20
0
2 4 6 8 10 12 14 16 18 min
mAU (c)
80
60
40
20
0
2 4 6 8 10 12 14 16 18 min
FIGURE 5.23 Final optimization of temperature and voltage. Order of peaks: (1) cytidine, (7) thiourea, (2)
uridine, (3) inosine, (4) guanosine, (5) thymidine, and (6) adenosine. Conditions: injection, 10 kV for 3 s;
25 cm × 100 µm column (CEC Hypersil C18 , 3 µm); mobile phase (5 mM acetic acid, 3 mM TEA, pH
5)–ACN (92:8, v/v). (a) 25◦ C and 20 kV; (b) 25◦ C and 25 kV; and (c) 20◦ C and 25 kV. (Reprinted from
Helboe, T. and Hansen, S.H., J. Chromatogr. A, 836, J. Chromatogr. A, 315–324, 1999. With permission from
Elsevier.)
An example of a CEC analysis of an illicit drug was shown by Lurie and collaborators [134].
In this study, a standard mixture of seven cannabinoids was separated, and this test mix-
ture was then compared with concentrated hashish and marijuana extracts (Figure 5.26). The
cannabinoid standards used were cannabigerol (CBG), cannabidiol (CBD), cannabinol (CBN),
-9-tetrahydrocannabinol (d9-THC), -8-tetrahydrocannabinol (d8-THC), cannabichromene
(CDB), and -9-tetrahydrocannabinolic acid (d9-THCA-A), with dimethyl sulfoxide (DMSO) as
the neutral marker. A Hypersil C18 column (3 µm dp ) was used. The column inner diameter was
100 µm and had a total length of 49 cm, and the effective length was 40 cm. UV detection was
conducted at 210 nm. The run buffer was 75% ACN and 25% 25 mM phosphate buffer pH 2.57. The
applied voltage was 30 kV, and the column temperature was kept at 20◦ C.
Carbohydrate separations too have been accomplished using CEC. Such an example was pre-
sented by Zhao and Johnson [106], where they studied sucralose and related carbohydrates, as shown
in Figure 5.27. The 100 µm inner diameter column was packed with 3 µm ODS particles, had a
total bed length of 25 cm, and a total length of 33 cm. During each run, an external pressure of six
bars was applied to the inlet and outlet vials to eliminate bubble formation. The compounds were
detected at 195 nm for this experiment. Different mobile phase compositions were studied, and for
Figure 5.28 the eluent was: 35:65 ACN/4 mM borate (Figure 5.28a), 30:70 ACN/4 mM borate, and
25:75 ACN/4 mM borate. For the three runs shown in Figure 5.28, the applied voltage was 15 kV.
Note that sucralose was successfully separated from related compounds A and B, a system peak, and
ethyl acetate (EA) traces from compound A.
Clearly, CEC is a flexible and efficient separation technique that, with careful attention in
method development, is capable of addressing analytical needs across a broad spectrum of sample
types.
U
C
I = 10 mM
Absorbance 0.07 EOF G
A
T
0.003
–0.001
0 5 10 15
Time (min)
C,U
I = 5 mM
0.07 EOF
G
Absorbance
A
T
0.003
–0.001
0 5 10 15
Time (min)
FIGURE 5.24 Effect of the ionic strength on the migration order of nucleosides in CEC. Column: 27 cm ×
75 µm i.d., bed length 20 cm, packing: 3 µm Phenyl Hypersil; electrolyte: acetic acid/ammonia (pH 5)/ACN
(95/5); temperature: 20◦ C; applied voltage: +20 kV; UV detection: 254 nm; electrokinetic injection: 4 s
(+10 kV); nucleoside concentration: 50 ppm. (Reprinted from Cahours, X., et al., J. High Resol. Chromatogr.,
5.6 CONCLUSIONS
What does the future hold for CEC? It is impossible to predict the path the technique will follow. It is
of course unlikely that it will ever enjoy the great breadth of application of HPLC and GC. That said,
it is entirely possible that the technique will find one or more niches in which it will thrive. Likely
areas of application include: field-portable, low-cost, chip-based analytical instruments; disposable
components in medical diagnostic/assay devices; and consumables for clinical diagnostic tools.
48 mAU
2
36
3
1
5 8
24 4 6
7 9
12
10 20 30 40 50
Time (min)
FIGURE 5.25 CEC-UV chromatogram (240 nm) of a mixture of 10 corticosteroids (100 µg mL−1 ) using a
linear gradient elution program. Voltage = 30 kV, HPLC injection volume = 10 µL, flowrate = 10 µL min−1 for
3 min then increased to 100 µL min−1 . Gradient program = initial: ammonium acetate, 5 mM, in ACN–water
(17:83), held for 3 min then ramped to 38% acetronitrile at 15 min and maintained to end of run. Column =
Hypersil ODS, 3 µm, 42 cm total length, 30 cm packed length, 30.1 cm to window, (l) triamcinolone,
(2) hydrocortisone and prednisolone coeluting, (3) cortisone, (4) methylprednisolone, (5) betamethasone,
(6) dexamethasone, (7) adrenosterone, (8) fluocortolone, and (9) triamcinotone acetonide. (Reprinted from
Taylor, M.R. and Teale, P., J. Chromatogr. A, 768, 89–95, 1997. With permission from Elsevier.)
mAU (a) CBDA

UV 210 nm
12.5
10
7.5
5
2.5
0
0 5 10 15 20 25 30 35 40
CBN
mAU (b) d9-THCA-A
CBG d9-THC
12.5 CBD
10 CBE
DMSO d8-THC
7.5
5
2.5
0
0 5 10 15 20 25 30 35 40
(c)
mAU
12.5
10
7.5
5
2.5
0
0 5 10 15 20 25 30 35 40
Time (min)
FIGURE 5.26 CEC of (a) concentrated hashish extract, (b) standard mixture of cannabinoids, and
(c) concentrated marijuana extract. Conditions: 75% ACN/25 mM phosphate buffer pH 2.57 with voltage
30 kV and temperature 20◦ C. A Hypersil C18, 3-µm [100 µm × 49 cm (40-cm length to detector)] column is
used. Electrokinetic injections of 8.0 s at 5.0 kV are used. (Reprinted from Lurie, I.S., et al., Anal. Chem., 70,
3255, 1998. With permission.)
CI
OH
HO O
OH
OH
CI
O HO
O
CI
Sucralose
CI CI
OAc OH
HO O O
HO
OH OH
OH OH
CI CI
O O
HO CI
O O
CI CI
(a) (b)
FIGURE 5.27 Structures of sucralose and related carbohydrate compounds (a) and (b). (Copyright 2000 from
CEC: analysis of sucralose and related carbohydrate compounds by Zhao, R.R. and Johnson, B.P. Reproduced
by permission of Taylor & Francis Group, LLC., http://www.taylorandfrancis.com)
Sucralose
(a) A
B
EA
System peak
Sucralose
(b)
EA B
A
System peak
A
Sucralose
B
(c)
EA
0 4 8 12 16 20
Time (min)
FIGURE 5.28 Electrochromatograms of sucralose and related compounds in different mobile phases at 15 kV
run: (a) 35/65 ACN/4 mM borate, (b) 30/70 ACN/4 mM borate, and (c) 25/75 ACN/4 mM borate. (Copyright
2000 from CEC: analysis of sucralose and related carbohydrate compounds by Zhao, R.R. and Johnson, B.P.
Reproduced by permission of Taylor & Francis Group, LLC., http://www.taylorandfrancis.com)
CEC is a technique that must compete in an increasingly crowded field of technologies, yet it has
a sufficient quantity of positive attributes—detailed in this chapter—to be worthy of consideration.
It will be intriguing to see where this road leads.
REFERENCES
1. Pretorius, V., Hopkins, B.J., and Schieke, J.D., Electro-osmosis: A new concept for high-speed liquid
chromatography, J. Chromatogr. A, 99, 23, 1974.
2. Jorgenson, J.W. and Lukacs, K.D., Zone electrophoresis in open-tubular glass capillaries: Preliminary
data on performance, J. High Res. Chrom., 4, 230, 1981.
3. Knox, J.H. and Grant, I.H., Thermal effects and band spreading in capillary electro-separation,
Chromatographia, 26, 329, 1988.
4. Schweitz, L., Andersson, L.I., and Nilsson, S., Capillary electrochromatography with predetermined
selectivity obtained through molecular imprinting, Anal. Chem., 69, 1179, 1997.
5. Landers, J.P, Handbook of Capillary Electrophoresis, 2nd ed., CRC Press, FL, 1997, Chapter 5.
6. Wen, E., Asiaie, R., and Horváth, C., Dynamics of capillary electrochromatography: II. Comparison
of column efficiency parameters in microscale high-performance liquid chromatography and capillary
electrochromatography, J. Chromatogr. A, 855, 349, 1999.
7. Yan, C., Schaufelberger, D., and Erni, F., Electrochromatography and micro high-performance liquid
chromatography with 320 µm I.D. packed columns, J. Chromatogr. A, 670, 15, 1994.
8. Bartle, K.D. and Myers, P., Theory of capillary electrochromatography, J. Chromatogr. A, 916, 3,
2001.
9. Colon, L.A., et al., Recent progress in capillary electrochromatography, Electrophoresis, 21, 3965,
2000.
10. Vanhoenacker, G., et al., Recent applications of capillary electrochromatography, Electrophoresis, 22,
4064, 2001.
11. Cikalo, M.G., et al., Capillary electrochromatography. Tutorial Review, Analyst, 123, 87R, 1998.
12. Peters, E.C., et al., Molded rigid polymer monoliths as separation media for capillary electrochro-
matography, Anal. Chem., 69, 3646, 1997.
13. Palm, A. and Novotny, M.V., Macroporous polyacrylamide/poly(ethylene glycol) matrixes as
stationary phases in capillary electrochromatography, Anal. Chem., 69, 4499, 1997.
14. Dulay, M.T., Kulkarni, R.P., and Zare, R.N., Preparation and characterization of monolithic porous
capillary columns loaded with chromatographic particles, Anal. Chem., 70, 5103, 1998.
15. Gusev, I., Huang, X., and Horváth, C., Capillary columns with in situ formed porous monolithic
packing for micro high-performance liquid chromatography and capillary electrochromatography,
J. Chromatogr. A, 855, 273, 1999.
16. Chirica, G. and Remcho, V.T., Silicate entrapped columns—New columns designed for capillary
electrochromatography, Electrophoresis, 20, 50, 1999.
17. Tang, Q.L., Xin, B.M., and Lee, M.L., Monolithic columns containing sol–gel bonded octadecylsilica
for capillary electrochromatography, J. Chromatogr. A, 837, 35, 1999.
18. Tanaka, N. et al., Monolithic silica columns for HPLC, micro-HPLC, and CEC, J. High Resol.
Chromatogr., 23, 111, 2000.
19. Ratnayake, C.K., Oh, C.S., and Henry, M.P., Particle loaded monolithic sol–gel columns for capillary
electrochromatography: A new dimension for high performance liquid chromatography, J. High Resol.
Chromatogr., 23, 81, 2000.
20. Tang, Q. and Lee, M.L., Continuous-bed columns containing sol-gel bonded packing materials for
capillary electrochromatography, J. High Resol. Chromatogr., 23, 73, 2000.
matography. 1. Fine control of porous properties and surface chemistry, Anal. Chem., 70, 2288,
1998.
matography. 2. Effect of chromatographic conditions on the separation, Anal. Chem., 70, 2296,
1998.
23. Ishizuka, N., et al., Performance of a monolithic silica column in a capillary under pressure-driven
and electrodriven conditions, Anal. Chem., 72, 1275, 2000.
24. Fujimoto, C., Preparation of fritless packed silica columns for capillary electrochromatography, J. High
25. Asiaie, R., et al., Sintered octadecylsilica as monolithic column packing in capillary electrochro-
matography and micro high-performance liquid chromatography, J. Chromatogr. A, 806, 251,
1998.
26. Dulay, M.T., et al., Preparation and characterization of monolithic porous capillary columns loaded
with chromatographic particles, Anal. Chem., 70, 5103, 1998.
27. Tang, Q., Wu, N., and Lee, M.L., Continuous bed columns containing sol–gel bonded large-pore
octadecylsilica for capillary electrochromatography, J. Microcol. Sep., 11, 550, 1999.
28. Schweitz, L., Andersson, L.I., and Nilsson, S., Molecular imprint-based stationary phases for capillary
29. Xiong, B. et al, Capillary electrochromatography with monolithic poly(styrene-co-divinylbenzene-co-
methacrylic acid) as the stationary phase, J. High Resol. Chromatogr., 23, 67, 2000.
30. Yu, C., Svec, F., and Frechet, J.M., Towards stationary phases for chromatography on a microchip:
Molded porous polymer monoliths prepared in capillaries by photoinitiated in situ polymerization as
separation media for electrochromatography, Electrophoresis, 21, 120, 2000.
31. Remcho, V.T. and Tan, Z.J., MIPs as chromatographic stationary phases for molecular recognition.
Anal. Chem., 71, 248A, 1999.
32. Liao, J.L., et al. Preparation of continuous beds derivatized with one-step alkyl and sulfonate groups
for capillary electrochromatography, Anal. Chem., 68, 3468, 1996.
33. Schure, M.R., et al., High-performance capillary gel electrochromatography with replaceable media,
Anal. Chem., 70, 4985, 1998.
34. Fujimoto, C., Charged polyacrylamide gels for capillary electrochromatographic separations of
uncharged, low molecular weight compounds, Anal. Chem., 67, 2050, 1995.
35. Ericson, C., et al., Preparation of continuous beds for electrochromatography and reversed-phase liquid
chromatography of low-molecular-mass compounds, J. Chromatogr. A, 767, 33, 1997.
36. Fujimoto, C., Sakurai, M., and Muranaka, Y., PEEK columns for open-tubular liquid chromatography
with electroosmotic flow, J. Microcol. Sep., 11, 693, 1999.
37. Ericson, C. and Hjerten, S., Reversed-phase electrochromatography of proteins on modified continuous
beds using normal-flow and counterflow gradients. Theoretical and practical considerations, Anal.
Chem., 71, 1621, 1999.
38. Fujimoto, C., Fujise, Y., and Matsuzawa, E., Fritless packed columns for capillary electrochro-
matography: Separation of uncharged compounds on hydrophobic hydrogels, Anal. Chem., 68, 2753,
1996.
39. Svec, F., Tennikova, T.B., and Deyl, Z., Monolithic materials: Preparation, properties, and applications,
Anal. Bio. Chem., 379, 8, 2004.
40. Svec, F., Preparation and HPLC applications of rigid macroporous organic polymer monoliths,
J. Sep. Sci., 27, 747, 2004.
41. Allen, D. and El Rassi, Z., Silica-based monoliths for capillary electrochromatography:
Methods of fabrication and their applications in analytical separations, Electrophoresis, 24,
3962, 2003.
42. Luo, Q.L. and Andrade, J.D., Electrokinetic flow through packed capillary columns, J. Microcol., 11,
682, 1999.
43. Svec, F., Monolithic stationary phases for capillary electrochromatography based on synthetic
polymers: Designs and applications, J. High Resol. Chromatogr., 23, 3, 2000.
44. Svec, F., et al., Design of the monolithic polymers used in capillary electrochromatography columns,
J. Chromatogr. A, 887, 3, 2000.
45. Nishi, H. and Terabe, S., Optical resolution of drugs by capillary electrophoretic techniques,
J. Chromatogr. A, 694, 276, 1994.
46. Sellergen, B., Imprinted chiral stationary phases in high-performance liquid chromatography,
J. Chromatogr. A, 906, 227, 2001.
47. Nilsson, K., et al., Imprinted polymers as antibody mimetics and new affinity gels for selective
separations in capillary electrophoresis, J. Chromatogr. A, 680, 57, 1994.
48. Arshady, R. and Mosbach, K., Synthesis of substrate-selective polymers by host–guest polymerization,
Macromol. Chem., 182, 687, 1981.
49. Wulff, G. and Sarhan, A., Über die Anwendung von enzymanalog gebauten Polymeren zur
Racemattrennung, Angew. Chem., 84, 364, 1972.
50. Sellergren, B., Molecularly Imprinted Polymers—Man Made Mimics of Antibodies and Their
Applications in Analytical Chemistry, Elsevier, Amsterdam, 2001, pp. 113–184.
51. Tang, Q., Xin, B., and Lee, M.L., Monolithic columns containing sol–gel bonded octadecylsilica for
capillary electrochromatography, J. Chromatogr. A, 837, 35, 1999.
52. Yan, C., et al., Capillary electrochromatography: Analysis of polycyclic aromatic hydrocarbons, Anal.
Chem., 67, 2026, 1995.
53. Dadoo, R. et al., Advances in capillary electrochromatography: Rapid and high-efficiency separations
of PAHs, Anal. Chem., 70, 4787, 1998.
54. Ngola, S.M., et al., Conduct-as-cast polymer monoliths as separation media for capillary electrochro-
matography, Anal. Chem., 73, 849, 2001.
55. Fintschenko, Y., et al., Chip electrochromatography of polycyclic aromatic hydrocarbons on an
acrylate-based UV-initiated porous polymer monolith, Fresenius J. Anal. Chem., 371, 174, 2001.
56. Chirica, G.S. and Remcho, V.T., A simple procedure for the preparation of fritless columns by entrap-
ping conventional high performance liquid chromatography sorbents, Electrophoresis, 21, 3093,
2000.
57. Chirica, G.S. and Remcho, V.T., Fritless capillary columns for HPLC and CEC prepared by
immobilizing the stationary phase in an organic polymer matrix, Anal. Chem., 72, 3605, 2000.
58. Vallano, P.T. and Remcho, V. T., Modeling interparticle and intraparticle (perfusive) electroosmotic
flow in capillary electrochromatography, Anal. Chem., 72, 4255, 2000.
59. Dittman, M.M. and Rozing, R., Capillary electrochromatography: Investigation of the influence of
mobile phase and stationary phase properties on electroosmotic velocity, retention, and selectivity,
J. Microcol. Sep., 9, 399, 1997.
60. Mayer, M., et al., Fritless capillary electrochromatography, Electrophoresis, 20, 43, 1999.
61. Engelhardt, H. and Hafner, F.T., Porous and non-porous stationary phases for capillary electrochro-
matography under conditions of reversed phase, Chromatography, 52, 769–776, 2000.
62. Wu, X., et al., Rapid separation and determination of carbamate insecticides using isocratic elution
pressurized capillary electrochromatography, Electrophoresis, 27, 768, 2006.
63. Henry, C.W., McCarroll, M.E., and Warner, I.M., Separation of the insecticidal pyrethrin esters by
64. Tegeler, T. and El Rassi, Z., Surfactant-mediated capillary electrochromatography with octadecyl-
silica- packed capillary columns for the separation of nonpolar compounds. Case of pyrethroid
insecticides, Electrophoresis, 23, 1217, 2002.
65. Tegeler, T. and El Rassi, Z., On-column trace enrichment by sequential frontal and elution electrochro-
matography: II. Enhancement of sensitivity by segmented capillaries with z-cell configuration—
Application to the detection of dilute samples of moderately polar and nonpolar pesticides,
J. Chromatogr. A, 945, 267, 2002.
66. O’Mahony, T., et al., Monitoring the supercritical fluid extraction of pyrethroid pesticides using
capillary electrochromatography. Int. J. Environ. An. Ch., 83, 681, 2003.
67. Cooper, P.A., Jessop, K.M., and Moffatt, F., Capillary electrochromatography for pesticide analysis:
Effects of environmental matrices, Electrophoresis, 21, 1574, 2000.
68. Fung, Y.S. and Long, Y.H., Determination of phenols in soil by supercritical fluid extraction–capillary
69. Araki, T., Chiba, M., Tsunoi, S., and Tanaka, M., Separation of chlorophenols by HPLC and capillary
electrochromatography using β-cyclodextrin-bonded stationary phases, Anal. Sci., 16, 412, 2000.
70. Cassells, N.P., et al., Microtox® testing of pentachlorophenol in soil extracts and quantification
by capillary electrochromatography (CEC)—A rapid screening approach for contaminated land,
Chemosphere, 40, 609, 2000.
71. Saeed, M., et al., Application of capillary electrochromatography (CEC) for the analysis of phenols in
mainstream and sidestream tobacco smoke, Chromatographia, 49, 391, 1999.
72. Bailey, C.G. and Yan, C., Separation of explosives using capillary electrochromatography, Anal. Chem.,
70, 3275, 1998.
73. Hilmi, A. and Luong, J.H.T., In-line coupling capillary electrochromatography with amperometric
detection for analysis of explosive compounds, Electrophoresis, 21, 1395, 2000.
74. Mueller, T., Posch, U.C., and Lindner, H.H., Separation of phenylthiohydantoin amino acids by
capillary electrochromatography, Electrophoresis, 25, 578, 2004.
75. Huber, C., Choudhary, G., and Horvath, C., Capillary electrochromatography with gradient elution,
Anal. Chem., 69, 4429, 1997.
76. Alicea-Maldonado, R. and Colon, L., Capillary electrochromatography using a fluoropolymer as the
chromatographic support material, Electrophoresis, 20, 37, 1999.
77. Kotia, R.B., Li, L., and McGown, L.B., Separation of nontarget compounds by DNA aptamers, Anal.
Chem., 72, 827, 2000.
78. Shediac, R., et al., Reversed-phase electrochromatography of amino acids and peptides using porous
polymer monoliths, J. Chromatogr. A, 925, 251, 2001.
79. Charles, J.A.M. and McGown, L.B., Separation of Trp–Arg and Arg–Trp using G-quartet-
forming DNA oligonucleotides in open-tubular capillary electrochromatography, Electrophoresis, 23,
1599, 2002.
80. Wu, R., et al., Separation of peptides on mixed mode of reversed-phase and ion-exchange capillary
electrochromatography with a monolithic column, Electrophoresis, 23, 1239, 2002.
81. Charvatova, J., Kral, V. and Deyl, Z., Capillary electrochromatographic separation of aromatic amino
acids possessing peptides using porphyrin derivatives as the inner wall modifiers, J. Chromatogr. B,
770, 155, 2002.
82. Wu, R., et al., Capillary electrochromatography for separation of peptides driven with electrophoretic
mobility on monolithic column, Anal. Chem., 73, 4918, 2001.
83. Steiner, F. and Scherer, B., Separation of small peptides by electrochromatography on silica-based
reversed phases and hydrophobic anion exchange phases, Electrophoresis, 26, 1996, 2005.
84. Svec, F., et al., Monolithic stationary phases for capillary electrochromatography based on synthetic
polymers: Designs and applications, J. High Resol. Chromatogr., 23, 3, 2000.
85. Walhagen, K., Unger K.K., and Hearn, M.T.W., Capillary electrochromatography analysis of hormonal
cyclic and linear peptides, Anal. Chem., 73, 4924, 2001.
86. Gucek, M., et al., Capillary electrochromatography/nanoelectrospray mass spectrometry for attomole
characterization of peptides, Rapid Commun. Mass Spectrom., 14, 1448, 2000.
87. Ludtke, S., Adam, T., and Unger, K.K., Application of 0.5-µm porous silanized silica beads in
88. Adam, T. and Unger, K.K., Comparative study of capillary electroendosmotic chromatography and
electrically assisted gradient nano-liquid chromatography for the separation of peptides, J. Chromatogr.
A, 894, 241, 2000.
89. Ye, M.L., et al., Separation of peptides by strong cation-exchange capillary electrochromatography,
J. Chromatogr. A, 869, 385, 2000.
90. Ye, M.L., et al., Capillary electrochromatography with a silica column with a dynamically modified
cationic surfactant, J. Chromatogr. A, 855, 137, 1999.
91. Zhang, S., et al., Capillary electrochromatography of proteins and peptides with a cationic acrylic
monolith, J. Chromatogr. A, 887, 465, 2000.
92. Pesek, J.J., et al., Synthesis and characterization of alkyl bonded phases from a silica hydride via
hydrosilation with free radical initiation, J. Chromatogr. A, 786, 219, 1997.
93. Wu, J.T., et al., Protein digest analysis by pressurized capillary electrochromatography using an ion
trap storage/reflectron time-of-flight mass detector, Anal. Chem., 69, 2908, 1997.
94. Lin, Z., et al., On-column coaxial flow chemiluminescence detection for underivatized amino acids
by pressurized capillary electrochromatography using a monolithic column, Anal. Chem., 78, 5322,
2006.
95. Bandilla, D. and Skinner, C.D., Protein separation by monolithic capillary electrochromatography,
J. Chromatogr. A, 1004, 167, 2003.
96. Huang, P., Wu, J.T., and Lubman, D.M., Separation of tryptic digests using a modified buffer in
pressurized capillary electrochromatography with an ion trap storage/reflectron time-of-flight mass
spectrometer, Anal. Chem., 70, 3003, 1998.
97. Pesek, J., et al., Protein and peptide separations on high surface area capillaries, Electrophoresis, 20,
2343, 1999.
98. He, B., Ji, J., and Regnier, F., Capillary electrochromatography of peptides in a microfabricated system,
J. Chromatogr. A, 853, 257, 1999.
99. Behnke, B. and Metzger, J. W., Tryptic digest mapping by gradient capillary electrochromatography,
100. Zhang, J., et al., Capillary electrochromatography of proteins on an anion-exchanger column, Anal.
Chem., 72, 3022, 2000.
101. Matyska, M.T., et al., Electrochromatographic characterization of etched chemically-modified
capillaries with small synthetic peptides, J. Chromatogr. A, 924, 211, 2001.
102. Pesek. J.J., Matyska, M.T., and Cho, J., Open tubular capillary electrochromatography in etched,
chemically modified 20 µm I.D. capillaries, J. Chromatogr. A, 845, 237, 1999.
103. Xu, W. and Regnier, F., Electrokinetically-driven cation-exchange chromatography of proteins and
its comparison with pressure-driven high-performance liquid chromatography, J. Chromatogr. A, 853,
243, 1999.
104. Huang, X., Zhang, J., and Horvath, C., Capillary electrochromatography of proteins and peptides with
porous-layer open-tubular columns, J. Chromatogr. A, 858, 91, 1999.
105. Guo, W., Koropchak, J.A., and Yan, C., Sensitive, universal detection for capillary electrochro-
matography using condensation nucleation light scattering detection, J. Chromatogr. A, 849, 587,
1999.
106. Zhao, R.R. and Johnson, B.P., Capillary electrochromatography: analysis of sucralose and related
carbohydrate compounds, J. Liq. Chromatogr. Rel. Technol., 23, 1851, 2000.
107. Yang, C. and El Rassi, Z., Capillary electrochromatography of derivatized mono- and oligosaccharides,
108. Suzuki, S., et al., Separation of 1-phenyl-3-methyl-5-pyrazolone derivatives of monosaccharides by
109. Gucek, M. and Pilar, B., Capillary electrochromatography of 1-phenyl-3-methyl-5-pyrazolone
derivatives of some mono- and disaccharides, Chromatographia, 51, S139, 2000.
110. Suzuki, S., et al., Preparation of various silica-based columns for capillary electrochromatography by
in-column derivatization, J. Chromatogr. A, 873, 247, 2000.
111. Verheij, E.R., et al., Pseudo-electrochromatography–mass spectrometry: A new alternative,
J. Chromatogr. A, 554, 339, 1991.
112. Lin, S.Y. and Liu, C.Y., An insight into the phenomena involved in a multiple-function stationary phase
for the capillary electrochromatographic separation of 2 -, 3 -, and 5 -monophosphorylated nucleoside
isomers, Electrophoresis, 24, 2973, 2003.
113. Huang, Y.C., Lin, C.C., and Liu, C.Y., Preparation and evaluation of molecularly imprinted polymers
based on 9-ethyladenine for the recognition of nucleotide bases in capillary electrochromatography,
114. Cahours, X., et al., Fast separation of nucleosides by capillary electrochromatography on non-
endcapped phenyl-bonded silica phase using short-end injection method, J. High Resol. Chromatogr.,
23, 138, 2000.
115. Helboe, T. and Hansen, S.H., Separation of nucleosides using capillary electrochromatography,
J. Chromatogr. A, 836, 315, 1999.
116. Matyska, M.T., Pesek, J.J., and Katrekar, A., Open tubular capillary electrochromatography using
etched fused-silica tubing modified with chemically bonded liquid crystals, Anal. Chem., 71, 5508,
1999.
117. Zhang, M., Yang, C., and El Rassi, Z., Capillary electrochromatography with novel stationary phases.
3. Retention behavior of small and large nucleic acids on octadecyl-sulfonated-silica, Anal. Chem.,
71, 3277, 1999.
118. Zhang, M. and El Rassi, Z., Capillary electrochromatography with novel stationary
phases. I. Preparation and characterization of octadecyl-sulfonated silica, Electrophoresis, 19,
2068, 1998.
119. Zhang, M. and El Rassi, Z., Capillary electrochromatography with novel stationary phases: II. Studies
of the retention behavior of nucleosides and bases on capillaries packed with octadecyl-sulfonated-
silica microparticles, Electrophoresis, 20, 31, 1999.
120. Moffatt, F., Cooper, P.A., and Jessop, K.M., Comparison of capillary electrochromatography with
high-performance liquid chromatography for the analysis of pirimicarb and related compounds,
J. Chromatogr. A, 855, 215, 1999.
121. Euerby, M.R., et al., Solvent and stationary phase selectivity in capillary electrochromatography
method development: Comparison of C18, C8, and phenyl-bonded phases for the separation of a
series of substituted barbiturates, J. Microcol. Sep., 11, 305, 1999.
122. Ding, J. and Vorous, P., Capillary electrochromatography–mass spectrometry for the separation and
identification of isomeric polyaromatic hydrocarbon DNA adducts derived from in vitro reactions,
J. Chromatogr. A, 887, 103, 2000.
123. Ding, J. and Vourous, P., Capillary electrochromatography and capillary electrochromatography–mass
spectrometry for the analysis of dna adduct mixtures, Anal. Chem., 69, 379, 1997.
124. Aturki, Z., D’Orazio, G., and Fanali, S., Rapid assay of vitamin E in vegetable oils by reversed-phase
125. Eimer, T., Unger, K.K., and Van der Greef, J., Selectivity tuning in pressurized-flow electrochromatog-
raphy, Trends Anal. Chem., 15, 463, 1996.
126. Vanhoenacker, G., et al., Single-run capillary electrochromatographic analysis of hop acids and
prenylated hop flavonoids, J. Sep. Sci., 24, 55, 2001.
127. Li, Y., et al., Optimized separation of pharmacologically active anthraquinones in Rhubarb by capillary
electrochromatography, Electrophoresis, 21, 3109, 2000.
128. Lane, S. J., et al., Evaluation of a new capillary electrochromatography/mass spectrometry interface
using short columns and high field strengths for rapid and rfficient analyses, Rapid Commun. Mass
Spectrom., 10, 733, 1996.
129. Sandra, P., et al., Analysis of triglycerides by capillary electrochromatography, J. Microcol. Sep., 9,
409, 1997.
130. Dermaux, A., et al., Analysis of the triglycerides and the free and derivatized fatty acids in fish oil by
capillary electrochromatography, J. High Resol. Chromatogr., 1998, 21, 545.
131. Dermaux, A., et al., Elucidation of the triglycerides in fish oil by packed-column supercritical fluid
chromatography fractionation followed by capillary electrochromatography and electrospray mass
spectrometry, J. Microcol. Sep., 11, 451, 1999.
132. Rentel, C., Gfroerer, P., and Bayer, E., Coupling of capillary electrochromatography to coordination
ion spray mass spectrometry, a novel detection method, Electrophoresis, 20, 2329, 1999.
133. Zhang, M.Q., Ostrander, G.K., and El Rassi, Z., Capillary electrochromatography with novel stationary
phases: IV. Retention behavior of glycosphingolipids on porous and non-porous octadecyl sulfonated
silica, J. Chromatogr. A, 887, 287, 2000.
134. Lurie, I.S., Meyers, R.P., and Conver, T., Capillary electrochromatography of cannabinoids, Anal.
Chem., 70, 3255, 1998.
135. Roed, L., Lundanes, E., and Greigrokk, T., Nonaqueous electrochromatography on continuous bed
columns of sol-gel bonded large-pore C30 material: Separation of retinyl esters, J. Microcol. Sep., 12,
561, 2000.
136. Roed, L., Lundanes, E., and Greibrokk, T., Nonaqueous electrochromatography on continuous bed
columns of sol-gel bonded large-pore C18 material: separation of retinyl esters, J. Chromatogr. A, 890,
347, 2000.
137. Roed, L., Lundanes, E., and Greigrokk, T., Nonaqueous electrochromatography on C30 columns:
Separation of retinyl esters, Electrophoresis, 20, 2373, 1999.
138. Sander, L.C., et al., Separation of carotenoid isomers by capillary electrochromatography with C30
stationary phases, Anal. Chem., 71, 3477, 1999.
139. Matyska, M., Pesek, J., and Yang, L., Screening method for determining the presence of N-
nitrosodiethanolamine in cosmetics by open-tubular capillary electrochromatography, J. Chromatogr.
A, 887, 497, 2000.
140. Girelli, A.M., et al., Reversed-phase capillary electrochromatography of aloins and related constituents
of aloe, Chromatographia, 53, S284, 2001.
141. Enlund, A.M., Isaksson, R., and Westerlund, D., Capillary electrochromatography of tricyclic
antidepressants on strong cation exchangers with different pore sizes, J. Chromatogr. A, 918, 211,
2001.
142. Spikmans, V., et al., Automated capillary electrochromatography tandem mass spectrometry using
mixed mode reversed-phase ion-exchange chromatography columns, Rapid Commun. Mass Spectrom.,
13, 141, 1999.
143. Whitaker, K.W. and Sepaniak, M.J., Nonaqueous packed capillary electrokinetic chromatographic
separations of large polycyclic aromatic hydrocarbons and fullerenes, Electrophoresis, 15, 1341,
1994.
matography. 2. Effect of chromatographic conditions on the separation, Anal. Chem., 70, 2296,
1998.
145. Stol, R., et al., Application of size exclusion electrochromatography to the microanalytical determina-
tion of the molecular mass distribution of celluloses from objects of cultural and historical value, Anal
Chem., 74, 2314, 2002.
146. Smith, N.W. and Evans, M.B., The analysis of pharmaceutical compounds using electrochromatogra-
phy, Chromatographia, 38, 649, 1994.
147. Smith, N.W. and Evans, M.B., The efficient analysis of neutral and highly polar pharmaceutical
compounds using reversed-phase and ion-exchange electrochromatography, Chromatographia, 41,
197, 1995.
148. Taylor, M.R. and Teale, P., Gradient capillary electrochromatography of drug mixtures with UV and
electrospray ionisation mass spectrometric detection, J. Chromatogr. A, 768, 89, 1997.
149. Gordon, D.B., Lord, G.A., and Jones, D.S., Development of packed capillary column electrochro-
matography/mass spectrometry, Rapid Commun. Mass Spectrom., 8, 544, 1994.
150. Lord, G.A., et al., Tapers and restrictors for capillary electrochromatography and capillary
electrochromatography–mass spectrometry, J. Chromatogr. A, 768, 9, 1997.
151. Seilar, R.M., et al., Capillary electrochromatography with 1.5 µm ODS-modified non-porous silica
spheres, Chromatographia, 46, 131, 1997.
152. Seilar, R.M., et al., Capillary electrochromatography with 1.8-µm ODS-modified porous silica
particles, J. Chromatogr. A, 808, 71, 1998.
153. Euerby, M.R., et al., “Short-end injection” rapid analysis capillary electrochromatography, Chro-
matographia, 47, 135, 1998.
154. Frame, L.A., Robinson, M.L., and Lough, W.J., Simplification of capillary electrochromatography
procedures, J. Chromatogr. A, 798, 243, 1998.
155. Mayer, M., et al., Fritless capillary electrochromatography, Electrophoresis, 20, 43, 1999.
156. Hilhorst, M.J., Somsen, G.W., and de Jong, G.H., Sensitivity enhancement in capillary electrochro-
matography by on-column preconcentration, Chromatographia, 53, 190, 2001.
157. Spikmans, V., Lane, S.J., and Smith, N.W., Capillary electrochromatography of complex plasma
matrix on a C18 /SCX column using UV-vis and mass spectrometric detection, Chromatographia, 51,
18, 2001.
158. Que, A.H., et al., Steroid profiles determined by capillary electrochromatography, laser-induced
fluorescence detection and electrospray–mass spectrometry, J. Chromatogr. A, 887, 379, 2000.
159. Djordjevic, N.M., et al., High temperature and temperature programming in capillary electrochro-
matography, J. Chromatogr. A, 887, 245, 2000.
160. Thiam, S., et al., Capillary electrochromatography of cholesterol and its ester derivatives, Anal. Chem.,
72, 2541, 2000.
161. Que, A.H., et al., Analysis of bile acids and their conjugates by capillary electrochromatography/
electrospray ion trap mass spectrometry, Anal. Chem., 72, 2703, 2000.
162. Chankvetadze, B., et al., Comparative study on the application of capillary liquid chromatography and
capillary electrochromatography for investigation of enantiomeric purity of the contraceptive drug
levonorgestrel, J. Pharm. Biomed. Anal., 30, 1897, 2003.
163. Liu, S., et al., Separation of structurally related estrogens using isocratic elution pressurized capillary
164. Liu, S., et al., On-line coupling of pressurized capillary electrochromatography with end-column
amperometric detection for analysis of estrogens, Electrophoresis, 26, 2342, 2005.
165. Ohyama, K., et al., Rapid separation of barbiturates and benzodiazepines by capillary electrochro-
matography with 3-(1,8-naphthalimido)propyl-modified silyl silica gel, Biomed. Chromatogr., 18,
396, 2004.
166. Cahours, X., Morin, P., and Dreux, M., Influence of ionic strength and organic modifier on performance
in capillary electrochromatography on phenyl silica stationary phase, J. Chromatogr. A, 845, 203, 1999.
167. Jinno, K., et al., Comparison of separation behavior of benzodiazepines in packed capillary elec-
trochromatography and open-tubular capillary electrochromatography, J. Chromatogr. A, 887, 479,
2000.
168. Kapnissi, C., et al., Analytical separations using molecular micelles in open-tubular capillary
electrochromatography, Anal. Chem., 74, 2328, 2002.
169. Kapnissi, C.P. and Warner, I.M., Separation of benzodiazepines using capillary electrochromatography,
J. Chromatogr. Sci., 42, 238, 2004.
170. De Rossi, A. and Desiderio, C., Separation of negatively charged nonsteroidal anti-inflammatory drugs
by reversed-phase capillary electrochromatography, J. Chromatogr. A, 984, 283, 2003.
171. Pai, Y.F., Lin, C.C., and Liu, C.Y., Optimization of sample stacking for the simultaneous determination
of nonsteroidal anti-inflammatory drugs with a wall-coated histidine capillary column, Electrophoresis,
25, 569, 2004.
172. Desiderio, C. and Fanali, S., Capillary electrochromatography and capillary electrochromatography–
electrospray mass spectrometry for the separation of non-steroidal anti-inflammatory drugs, J. Chro-
matogr. A, 895, 123, 2000.
173. Strickmann, D.B. and Blaschke, G., Capillary electrochromatography–electrospray ionization mass
spectrometry for the qualitative investigation of the drug etodolac and its metabolites in biological
samples, J. Chromatogr. B, 748, 213, 2000.
174. Hoegger, D. and Freitag, R., Acrylamide-based monoliths as robust stationary phases for capillary
175. Lammerhofer, M., et al., Capillary electrochromatography in anion-exchange and normal-phase mode
using monolithic stationary phases, J. Chromatogr. A, 925, 265, 2001.
176. Steiner, F. and Lobert, T., Capillary electrochromatography with bare silicas of different pore sizes as
stationary phases, J. Sep. Sci., 26, 1589, 2003.
177. Endund, A.M. and Westerlund, D., Effects of aliphatic amines on capillary electrochromatographic
performance of tricyclic antidepressants on octadecylsilica, J. Chromatogr. A, 895, 17, 2000.
178. Enlund, A.M., et al., Capillary electrochromatography of hydrophobic amines on continuous beds,
179. Vallano, P.T. and Remcho, V.T., Highly selective separations by capillary electrochromatography:
Molecular imprint polymer sorbents, J. Chromatogr. A, 887, 125, 2000.
180. Miyawa, J.H., Alasandro, M.S., and Riley, C.M., Application of a modified central composite
design to optimize the capillary electrochromatographic separation of related S-oxidation compounds,
J. Chromatogr. A, 769, 145, 1997.
181. Miyawa, J.H., Lloyd, K.D., and Alasandro, M.S., Capillary electrochromatography as a method
development tool for the liquid chromatographic separation of DUP 654 and related substances, J. High
182. Gfrorer, P., et al., Gradient elution capillary electrochromatography and hyphenation with nuclear
magnetic resonance, Electrophoresis, 20, 3, 1999.
183. Wei, W., et al., Capillary electrochromatographic separation of basic compounds with bare silica as
stationary phase, J. Chromatogr. A, 817, 65, 1998.
184. Yamamoto, H., Baumann, J., and Erni, F., Electrokinetic reversed-phase chromatography with packed
capillaries, J. Chromatogr. A, 593, 313, 1992.
185. Dekkers, S.E.G., Tjaden, U.R., and Van der Greef, J., Development of an instrumental configuration
for pseudo-electrochromatography-electrospray mass spectrometry, J. Chromatogr. A, 712, 201, 1995.
186. Pesek, J.J. and Matyska, M.T., Separation of tetracyclines by high-performance capillary electrophore-
sis and capillary electrochromatography, J. Chromatogr. A, 736, 313, 1996.
187. Nagaraj, S. and Karnes, H.T., Visible diode laser induced fluorescence detection of doxorubicin in
plasma using pressurized capillary electrochromatography, Biomed. Chromatogr., 14, 234, 2000.
188. Meyring, M., Chankvetadze, B., and Blaschke, G., Simultaneous separation and enantioseparation
of thalidomide and its hydroxylated metabolites using high-performance liquid chromatography in
common-size columns, capillary liquid chromatography and nonaqueous capillary electrochromato-
graphy, J. Chromatogr. A, 876, 157, 2000.
189. Meyring, M., et al., Investigation of the in vitro biotransformation of R-(+)-thalidomide by HPLC,
nano-HPLC, CEC and HPLC–APCI-MS, J. Chromatogr. B, 723, 255, 1999.
190. Lurie, I.S., et al., Profiling of impurities in illicit methamphetamine by high-performance liquid
chromatography and capillary electrochromatography, J. Chromatogr. A, 870, 53, 2000.
191. Lim, J.T., et al., Separation of related opiate compounds using capillary electrochromatography,
192. Lai, E.P.C. and Dabek-Zlotorzynska, E., Separation of theophylline, caffeine and related drugs by
normal-phase capillary electrochromatography, Electrophoresis, 20, 2366, 1999.
193. Hilhorst, M.J., Somsen, G.W., and de Jong, G.H., Capillary electrochromatography of basic compounds
using octadecyl-silica stationary phases with an amine-containing mobile phase, J. Chromatogr. A, 872,
315, 2000.
194. Desiderio, C., Ossicini, L., and Fanali, S., Analysis of hydroquinone and some of its ethers by using
195. Guan, N., et al., Open tubular capillary electrochromatography in fused-silica capillaries chemically
bonded with macrocyclic dioxopolyamine, Anal. Chim. Acta, 418, 145, 2000.
196. Stol, R., et al., Pseudo-electrokinetic packing of high efficiency columns for capillary electrochro-
matography, J. Chromatogr. A, 873, 293, 2000.
197. Smith, N.W., Comparison of aqueous and non-aqueous capillary electrochromatography for the
separation of basic solutes, J. Chromatogr. A, 887, 233, 2000.
198. Gillot, N.C., et al., The analysis of pharmaceutical bases on a silica stationary phase by capillary
electrochromatography using aqueous mobile phases, Chromatographia, 51, 167, 2000.
199. McKeown, A.P., et al., An evaluation of unbonded silica stationary phases for the separation of basic
analytes using capillary electrochromatography, Chromatographia, 52, 777, 2000.
200. Dittmann, M.M., Masuch, K., and Rozing, G.P., Separation of basic solutes by reversed-phase capillary
201. Dittmann, M.M. and Rozing, G.P., Capillary electrochromatography: Investigation of the influence
of mobile phase and stationary phase properties on electroosmotic velocity, retention, and selectivity,
J. Microcol. Sep., 9, 399, 1997.
202. Kanitsar, K., et al., Influence of organic modifiers on the separation of carboxylic acids using co-EOF
capillary electrophoresis, J. Liquid Chromatogr., 26, 455, 2003.
203. Zhang, S., Zhang, J., and Horvath, C., Rapid separation of peptides and proteins by isocratic capillary
electrochromatography at elevated temperature, J. Chromatogr. A, 914, 189, 2001.
204. Li, D.M. and Remcho, V.T., Perfusive electroosmotic transport in packed capillary electrochromatog-
raphy: Mechanism and utility, J. Microcol. Sep., 9, 389, 1997.
6 Capillary Electrophoresis of
Nucleic Acids
Eszter Szántai and András Guttman
CONTENTS
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

6.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
6.3.1 Electrophoretic Migration of Nucleic Acids. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
6.3.2 Efficiency and Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
6.4 Practical Applications of Capillary Electrophoresis Analysis of DNA Molecules. . . . . . . . 230
6.4.1 Polymerase Chain Reaction Product Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
6.4.1.1 Mutation Detection and Polymorphism Studies . . . . . . . . . . . . . . . . . . . . . . . . . 230
6.4.1.2 Forensic Application/Identity Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
6.4.1.3 Diagnosis of Infectious Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
6.4.1.4 Molecular Karyotyping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
6.4.1.5 Quantification of Cellular mRNA/Expression Analysis. . . . . . . . . . . . . . . . . 238
6.4.2 DNA Sequencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
6.4.3 Purity Control of Synthetic Nucleotides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
6.4.4 Oligonucleotides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
6.4.4.1 Antisense DNA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
6.4.5 Separation of Large DNA Molecules (>2 kb) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
6.4.6 Analysis of Mononucleotides and Nucleosides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
6.4.6.1 Nucleotides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
6.4.6.2 Nucleoside Analogs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
6.4.6.3 DNA Adducts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
6.5 Future Prospective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
6.1 INTRODUCTION
Since Francis Crick and James Watson discovered the double helical structure of deoxyribonucleic
acid in 1953 [1], which finding—with the understanding of the basic rules of inheritance suggested
by Gregor Mendel in the nineteenth century—became the theoretical basis of contemporary genetic
studies, there has been an urgent need for rapid, precise, sensitive, and high-throughput analysis
of nucleic acids. In the mid-1960s, electrophoresis-based methods employing polyacrylamide and
agarose gels were rapidly developed; however, these were fairly time-consuming and labor-intensive
techniques. The use of gels as sieving matrices for DNA electrophoresis was necessary to provide
and appropriate anticonvective media and to make electric field mediated separation of nucleic acid
227
molecules size based. Please note that the constant linear charge density of DNA chains results in
practically equal charge-to-mass ratio for the different lengths oligo- and polynucleotide molecules.
Slab gels are still widely used in DNA analysis in spite of the fact that they are not quantitative, they
lack online detection option and the analysis usually takes a long time. The next generation of DNA
separation techniques transpired in the late 1980s by the application of capillaries to electrophoresis.
This resulted in higher separation speed, as the large surface area-to-volume ratio allowed effective
Joule heat dissipation (i.e., higher voltages could be applied) [2]. In the past two decades, capillary
electrophoresis (CE) became a powerful separation technique featuring automation, ease of use, and
high efficiency. Novel detection methods, such as laser-induced fluorescence (LIF), increased the
sensitivity of the technique, thus advancing low-level DNA analysis from small amount of samples.
The next big step toward large-scale and automated DNA analysis was the introduction of low
viscosity so-called replaceable sieving polymer networks, which provided much higher flexibility
compared with high viscosity cross-linked gels [3].
Capillary gel electrophoresis (CGE) also shed light to the possibility of miniaturization (lab-on-
a-chip), as microfluidic devices can be manufactured similar to that of semiconductor chips, and
existing CE methods can be readily transferred from the capillary to microchip electrophoresis [4].
Thanks to the rapid technical developments of CE in the past 15 years [5], the identification of the
three billion base pair of the human genome was completed well ahead of schedule [6,7] enabling
new high-throughput genomics based approaches to be manifested in biomedicine, especially in
clinical diagnostics.
6.2 BACKGROUND
Slab gel electrophoresis is a well established and still very popular separation method for nucleic
acid analysis. Polyacrylamide and agarose are the most frequently used matrices for horizontal and
vertical slab gel electrophoresis [8]. DNA fragments can be detected after the separation either by
the application of a fluorescent dye, such as ethidium bromide, or by autoradiographic techniques.
Besides the traditionally used slab gel electrophoresis, analysis of DNA molecules in the beginning
of the twenty-first century is also based on high-performance liquid chromatography (HPLC) and
CE. Liquid chromatography partitions analytes between a very hydrophobic stationary phase and
a relevant mobile phase. As the diffusion characteristic of the analyte is an important issue, elec-
trophoresis is advantageous over HPLC when large molecules, such as longer DNA fragments, are
separated. The mass transport of DNA molecules is slow in the HPLC mobile phase, resulting in
increased band broadening. In CGE this does not represent a problem. CE instrumentation is also
less complex and capillaries are cheaper than HPLC columns. A good critical comparison of the two
methods is described in Reference 9.
Slab gels feature the benefit of separating multiple samples simultaneously in multilane for-
mat; however, the emergence of multicapillary array instruments prevailed over this advantage
[10]. Today, mostly 96 capillary units (ABI, Molecular Dynamics) but even 384-capillary array
electrophoresis instruments (Genteon) are commercially available, allowing high-throughput anal-
ysis primarily for industrial sequencing and genotyping laboratories. As these instruments are quite
expensive and the change of a single damaged capillary represents a major issue (i.e., in most
instances replacement of the entire array), smaller molecular diagnostic laboratories cannot afford
to use them. Microchip-based units (Agilent, Biorad) or small cartridge-based multicapillary sys-
tems (eGene) with fluorescence-based nucleic acid detection represent a good alternative to slab gel
electrophoresis for smaller molecular biology or clinical laboratories. In the past few years several
reviews have been published on DNA separation by CE discussing theoretical issues, separation
matrices and detection modes [11–16], and genetic diagnostic [17–23] and forensic applications
[24,25].
Capillary Electrophoresis of Nucleic Acids 229

6.3.1 ELECTROPHORETIC MIGRATION OF NUCLEIC ACIDS
Applying a uniform electric field (E) to an oligo- or polynucleotide molecule with a net charge of
Q, the electrical force (Fe ) is defined as
Fe = QE. (6.1)
In a gel or polymer network solution, a frictional force (Ff ) acts in the opposite direction

dx
Ff = f , (6.2)
dt
where f is the translational friction coefficient and dx and dt are the distance and time increments,
respectively. Differences in shape, size, and overall charge of the solute molecules result in variances
in electrophoretic mobilities providing the basis of the electrophoretic separation. Under steady-state
conditions, Fe and Ff are counterbalanced, thus the solute migrates with a steady-state velocity of v
dx EQ
v= = . (6.3)
dt f
The electrophoretic mobility (µ) is defined as the velocity per unit field strength
v
µ= . (6.4)
E
Retardation of the solute molecules in gel-filled or polymer-filled capillaries is a function of the
separation matrix concentration (P) and its physical interactions with the molecules subject to
electromigration is defined by the retardation coefficient (KR )
µapp = µ0 exp(−KR P), (6.5)
where µapp is the apparent electrophoretic mobility and µ0 is the free solution mobility of the analyte
(i.e., with no sieving matrix) [8].
When the average pore size of the matrix is in the same size range as the hydrodynamic radius
of the migrating analyte molecule, the classical sieving theory applies (Ogston regime) [26,27], that
is, at constant polymer concentration the retardation coefficient (KR ) is an apparent logarifunction
of the molecular weight (MW) of the migrating analyte [28]:
µ ∼ exp(−MW). (6.6)
In this instance, the so called Ferguson plots [29] cross each other at zero gel concentration.
The Ogston theory assumes that the migrating solute behaves as an unperturbed spherical object
with comparable size to the pores of the gel. However, DNA molecules can migrate through polymer
networks with pores significantly smaller than their size [30] by the phenomenon referred to as
reptation, suggesting “snakelike” motion for large biopolymers through the much smaller gel pores
[31–33]. The reptation model implies an inverse relationship between the size (MW) and the mobility
of the analyte molecules as shown in the following equation:
1
µ∼ . (6.7)
MW
At extremely high electric field strengths, reptation turns into biased reptation and the resulting
mobility of the analyte is described by

1
µ∼ + bE ,
a
(6.8)
MW
where b is a function of the mesh size of the sieving matrix, the charge, and segment length of the
migrating DNA molecules, and 1 < (a) < 2.
6.3.2 EFFICIENCY AND RESOLUTION

In CGE one of the major contributors to band broadening, besides the injection and detection extra
column effects, is the longitudinal diffusion of the solute molecules in the capillary tube [34]. The
theoretical plate number (N) is characteristic of column efficiency
E·
N =µ , (6.9)
2D
where µ is the electrophoretic mobility, D is the diffusion coefficient of the solute in the separation
gel-buffer system, and is the effective column length.
Resolution (Rs ) between two peaks can be calculated from the differences of their electrophoretic
mobilities (µ) [35]

E·
Rs = 0.18 · µ , (6.10)
D·µ
where µ̄ is the mean mobility of the sample components of interest. As one can see, Equa-
tions 6.9 and 6.10 suggest that higher applied electric field and lower solute diffusion coefficient
would result in higher separation efficiency (N) and concomitantly higher resolution (Rs ). One of
the limiting factors is the so-called Joule heat (Qj ), generated by the applied power (P = V × I) [2]
P
Qj = , (6.11)
r2 · I · L
where I is the current, L is the total column (electrode to electrode) length, and r is the inside radius of
the capillary. Owing to the temperature dependence of electrophoretic mobility, efficient temperature
control during CE separation is important in order to attain good reproducibility. Modern, automated
CE instruments are equipped with effective liquid- or air-cooling systems to address temperature
change-related problems.
6.4 PRACTICAL APPLICATIONS OF CAPILLARY ELECTROPHORESIS

ANALYSIS OF DNA MOLECULES
6.4.1 POLYMERASE CHAIN REACTION PRODUCT ANALYSIS
6.4.1.1 Mutation Detection and Polymorphism Studies
Polymerase chain reaction (PCR) is an in vitro DNA replication and amplification technique that
revolutionized nucleic acid analysis [36]. It enables small amount of nucleic acid molecules to be
exponentially amplified (i.e., to generate enough material for their analysis and sequencing). PCR is
a commonly used technique in biomedical, molecular biology, and clinical diagnostics laboratories
accommodating a variety of tasks, such as detection of hereditary diseases, identification of genetic

fingerprints, diagnosis of infectious diseases, cloning of genes, and paternity testing. Electrophoresis
is a crucial part of this method (except in case of real-time PCR) and CE—especially combined with
LIF—more and more replaces classical electrophoretic techniques in many fields.
The first research groups to investigate the utility of CE for PCR product and in general for DNA
analysis were Brownlee’s and Karger’s in the late 1980s [37–39]. In the years, novel high-resolution
polymer networks and capillary coatings were developed and CE conditions were optimized for better
DNA separation performance [12,40–42]. At the present time, many DNA-related applications are
still being reported by using similar conditions and capillaries. In this section, we list a few examples
from several important reports on PCR-based methods that greatly helped the fields of molecular
genetics. Chromosome 18q allelic loss has been reported to have prognostic significance in stage II
colorectal carcinoma. Erill et al. [43] have developed a robust and reliable fluorescent and multiplex
PCR assay to analyze five microsatellite markers for the allelic loss at the long arm of chromosome
18. Amplicon detection and evaluation was accomplished by an ABI 310 Genetic Analyzer (Applied
Biosystems).
The most frequently used methods for DNA length polymorphism analysis usually consists of two
major steps. First is the amplification of the genome region of interest followed by electrophoresis-
based separation of the resulting fragments. Hyytia-Trees et al. [44] developed a subtyping method
for Shiga toxin-producing Escherichia coli O157 strains, which was based on the analysis of sev-
eral variable number of tandem repeats in the bacterial genome. PCR products were sized using a
multicapillary electrophoresis-based sequencing system (CEQ™8000; Beckman Coulter) with the
following conditions: injection at 2.0 kV after 15 s and separation at 6.0 kV after 60 min.
One of the largest groups of genetic polymorphisms is referred to as single nucleotide polymor-
phisms (SNPs), which were conventionally detected by restriction fragment analysis. This method is
based on the cleavage of molecules by sequence-specific endonucleases. Interrogation of the size of
the resulting DNA fragments is usually electrophoresis separation based. The C677T mutation of the
methylenetetrahydrofolate (MTHFR) gene is a nutrient-oriented, “eco” genetic mutation, which is
associated with elevated levels of homocysteine and an increased risk for coronary heart disease. Sell
and Lugemwa [45] reported on an automated assay by means of the Hinf I RFLP to detect this muta-
tion. The resulting restriction fragments were analyzed on an ABI PRISM 310 single capillary-based
Genetic Analyzer.
Allele-specific PCR is another very popular genotyping method and is also referred to as ampli-
fication refractory mutation system. The technique is based on the utilization of an allele-specific
primer as its 3 end hybridizes to the SNP site followed by amplification using a DNA polymerase that
lacks 3 exonuclease activity. Amplification in this case occurs only if the primer sequence perfectly
matches with the template sequence. Carrera et al. [46] describe refractory mutation system analysis
of point mutations by CE. The first application of a multiplex multicolor assay for the simultaneous
detection of three of the most frequent mutations related to hereditary haemochromatosis was pre-
sented by Gomez-Llorente et al. [47]. One of the described methods was allele-specific PCR and
CE analysis of the amplified products that enabled easy, rapid, unambiguous, and high-resolution
discrimination between wild-type and mutant alleles. Fluorescently labeled products were analyzed
on an ABI PRISM 310 Genetic Analyzer using POP-4 polymer and fused-silica capillaries of 47-cm
length and 50-µm diameter [47].
Ligase detection reaction (LDR) developed by Barany [48] provides an elegant technique for
multiplexed typing of SNPs, deletions, and insertions. It utilizes the ability of the DNA ligase to
preferentially seal adjacent oligonucleotides hybridized to a target DNA with perfect complementa-
tion at a nick junction. In their paper, Thomas et al. [49] reported on the use of CE and microchip
electrophoresis format for detecting single base mutations in selected gene fragments with high
diagnostic value in colorectal cancer using LDR. The electrophoretic separations were carried out
for the single-stranded DNA products generated by allele-specific ligation assay to screen for a sin-
gle base mutation in the K-ras oncogene. Various separation matrices were investigated in CGE.
LDR products (44 and 51 bp) were analyzed by a cross-linked polyacrylamide gel (5%T/5%C) in
approximately 45 min having a 1000-fold molar excess of the LDR primers (25 bp). Interestingly,
when linear polyacrylamide gels were used, these same fragments could not be detected because of
significant electrokinetic bias during the electrokinetic injection process.
Another frequently used genotyping technique is based on primer extension. In this method, the
3 end of the applied primer anneals just before the SNP site as the first step of a quasi-minisequencing
reaction, which is carried out using dideoxy-ribonucleotide triphosphate terminators. For detection
purposes, either the primers or the dideoxy nucleotide terminators are labeled by an appropriate
fluorophor. Another version of this technique is allele-specific primer extension, which also repre-
sents a good method of choice for efficient genotyping. This technique is based on the sequence
specific extension of two allele-specific primers that differ at their 3 end defining the allele. Sin-
gle base extension (SBE) assay—next to allele-specific PCR—was used for the detection of three
mutations related to hereditary haemochromatosis [47]. ABI PRISM 310 Genetic Analyzer using
POP-4 polymer and fused-silica capillaries of 47-cm length and 50-µm diameter was applied for
separation.
Brazill and Kuhr [50] presented a model system to show possible advantages in combining
SBE technique with CGE and electrochemical detection (Figure 6.1). An electrochemically labeled
primer, with ferrocene acetate covalently attached to its 5 end, was used. The complementary
dideoxynucleotide (ddNTP) extended the primer by a single nucleotide and the reaction mixture
Unextended SBE primer
Extended SBE
product
10 nA
23 24 25 26 27 28
Time (min)
FIGURE 6.1 CGE coupled to sinusoidal voltammetric detection of SBE product. Separation capillary:
25-µm ID/25-cm length; Coating and separation medium: POP-4. Separation conditions: injection: 6 s at
–4 kV; Separation voltage: –4 kV. The SV detection employed a 21 Hz sine wave scanning from –200 mV
to 800 mV versus Ag/AgCl. Both of the time courses shown in this figure are from the second harmonic or
42 Hz. The solid trace is from the injection of the successful SBE utilizing ddATP terminator. The first peak
corresponds to migration of the unextended primer, whereas the late-eluting peak corresponds to the extended
product. The dotted trace represents the control reaction utilizing ddCTP. (From Brazill, S.A. and Kuhr, W.G.,
Anal Chem, 74, 3421, 2002. With permission.)
was subsequently separated by CGE in a 25-cm-long fused-silica capillary filled with POP-4 sieving
medium. The ferrocene-tagged fragments were detected with sinusoidal voltammetry.
Heteroduplex analysis is another very useful approach to analyze known or unknown genetic vari-
ations. The method is based on the differential migration of heteroduplex DNA molecules (formed
between a wild-type gene segment and the corresponding homologous segment containing an induced
mutation or a naturally occurring SNP), compared to the corresponding homoduplex in nondena-
turing polyacrylamide gel. Heteroduplexes migrate slower than their corresponding homoduplexes
due to their quasi open configuration surrounding the mismatched bases. Heteroduplex analysis is
usually carried out by slab gel electrophoresis, CE, or denaturing HPLC (dHPLC); however, it can
also be performed by microchip electrophoresis. Kozlowski et al. [51] demonstrated the influence
of a number of parameters on the electrophoretic properties of DNA duplexes such as temperature,
presence of glycerol, capillary length, as well as polymer (GeneScan) concentration, and evaluated
their contribution to the overall analysis time. Their study was carried out on an ABI 310 appara-
tus equipped with argon-ion laser. The time required for the detection of two typical BRCA1 gene
deleterious mutations by heteroduplex analysis could be significantly decreased through careful opti-
mization of analysis conditions. The effect of different silanizing reagents, polymeric coatings, and
polymer networks were investigated for detecting DNA mutations in the BRCA1 gene via heterodu-
plex analysis by Landers’ group [52]. Figure 6.2 shows the results of analyzing one heterozygous
BRCA1 mutant using capillaries coated with different silanizing reagents and polymers and how
they compare with commercial coated capillaries. In general, the electroosmotic flow (EOF) was
found to be 19- to 76-fold lower than that of with bare fused-silica capillary. Optimal performance
was observed using the chlorodimethyloctylsilane (OCT)- and poly(vinylpyrrolidone) (PVP)-coated
capillary and hydroxyethylcellulose (HEC) as the polymer network. More recently, Weber et al. [53]
proposed a novel method for the detection of unknown mutations called enhanced mismatch mutation
analysis, which is also based on electrophoretic heteroduplex analysis. Their experimental results
showed that the combination of high-resolution block-copolymer sieving matrix poly(acrylamide-
γ -polydimethylacrylamide), and nucleosides as additives in the electrophoretic medium increased
the resolution significantly between the homoduplex and heteroduplex peaks. The enhanced mis-
match mutation analysis method was compared to denaturing HPLC in a large-scale mutation study
of the breast cancer-associated gene BRCA2 and the success rate of detection of both methods was
comparable (94%).
Single-strand conformation polymorphism (SSCP) is a simple and versatile method combining
PCR amplification, denaturation of DNA molecules, and the analysis of denatured fragments by elec-
trophoresis. This technique, originally developed by Orita et al. [54], is based on subtle sequence
differences (often just a single base pair) that can result in a different three-dimensional conformation
and concomitantly measurable differences in their electrophoresis mobility. Hofman-Bang et al. [55]
developed a multiplex CE-based SSCP screening protocol on an automated genotyping platform for
mutation detection in the SCN5A gene coding for the alpha-subunit of the cardiac Na+ ion channel.
The separation was carried out by using a commercial polymer at 18◦ C and 30◦ C. Disease-causing
mutations were scattered over the DNA sequence, making it difficult to screen for specific mutations
that could cause several diseases, such as long QT syndrome, Brugada syndrome, idiopathic ventric-
ular fibrillation, sick sinus node syndrome, progressive conduction disease, dilated cardiomyopathy,
and atrial standstill. These diseases exhibited variable expressivity and identification of gene carriers
was clinically important, particularly in sudden infant and adult death syndromes. The method was
highly efficient with a false positive rate of 0.5% of the analyzed amplicons. Holmila and Husgafvel-
Pursiainen [56] compared CE-SSCP, denaturant gradient gel electrophoresis (DGGE), and direct
sequencing to investigate the benefits and sensitivity of each of the methods for the detection of
unknown TP53 mutations in human lung cancer. Their study revealed that direct sequencing per-
formed less well in finding mutations than the other two methods, and CE-SSCP was found to be
a fast and highly reproducible method, also considerably less laborious compared to DGGE, for
screening of unknown TP53 mutations. The CE-SSCP analysis was performed using ABI PRISM
(a) MET-PVP
10 RFU
(b) ALLYL-PVP
10 RFU
(c) BUTYL-PVP
Relative fluorescent units (RFU)
10 RFU
(d) OCTE-PVP
50 RFU
(e) MET-PA
10 RFU
(f) FC capillary
10 RFU
(g) OCT-PVP
10 RFU
4 5 6 7 8 9
FIGURE 6.2 Effect of different coatings on CE-based heteroduplex analysis. The PCR products were ampli-
fied from a heterozygous individual containing 1294del40 mutation in the BRCA1 gene and analyzed under
the following conditions: Injection: 20 s at 370 V/cm. Separation: 370 V/cm (current 27.0–27.6 µA), using
reversed polarity. EOF of the bare fused-silica capillary was 3.7 × 10−4 cm2 /V·s under the conditions used.
The different silanizing reagents and polymers used for modifying the surface of the capillary were as follows:
(a) CH2 = C(CH3 )COO(CH2 )3 Si(OCH3 )3 (MET), PVP, EOF is 8.5 × 10−6 cm2 /V · s; 24.8 µA. (b) CH2 =
C(CH2 )Si(CH3 )2 (Cl) (ALLYL), PVP, EOF is 8.9 × 10−6 cm2 /V · s; 25.0 µA. (c) Cl(CH2 )4 Si(CH3 )2 (Cl)
(BUTYL), PVP, EOF is 19.7 × 10−6 cm2 /V · s; 25.5 µA. (d) CH2 = CH − (CH2 )6 Si(OCH3 )3 (OCTE),
PVP, EOF is 7.3 × 10−6 cm2 /V · s; 25.4 µA. (e) CH2 = C(CH3 )COO(CH2 )3 Si(OCH3 )3 (MET), PA,
EOF is 4.9 × 10−6 cm2 /V · s; 25.3 µA. (f) FC capillary (J&W), EOF is 6.7 × 10−6 cm2 /V · s; 24.9 µA.
(g) CH3 (CH2 )7 Si(CH3 )2 (Cl) (OCT), PVP, EOF is 8.1 × 10−6 cm2 /V · s; 24.7 µA. (From Tian, H., et al., Anal
Chem, 72, 5483, 2000. With permission.)
310 capillary sequencer at 30◦ C, and with ABI PRISM 3100 Avant capillary sequencer at five dif-
ferent analysis temperatures (18◦ C, 25◦ C, 30◦ C, 35◦ C, and 40◦ C). Electrophoresis time was 30 min
with the applied voltage of 13 kV. The separation media were 5% GeneScan Polymer in 1× Tris-
borate- EDTA (TBE) buffer containing 10% glycerol. Endo et al. [57] studied SSCP parameters
that might affect the mutation analysis in the K-ras gene—such as electric field strength, separation
temperature, polymer, and additive in the sieving matrix (high concentration methyl cellulose)—by
CE and microchip electrophoresis. Figure 6.3 shows the effect of glycerol concentration in the buffer
(a) 2400000
Wild type
1500000 1.5% MC
RFU
600000
–900000
0 2 4 6 8 10
5000000 PSN 1
1.5% MC
3000000
RFU
1000000
–1000000
0 2 4 6 8 10
Time (min)
(b)
2000000 Wild type

1.5% MC + 5G
RFU
1200000
4000000
–4000000
0 2 4 6 8 10 12
400000
300000 PSN 1
1.5% MC + 5G
200000
RFU
100000
0
–100000
0 2 4 6 8 10 12
Time (min)
(c)
210000 Wild type
150000 1.5% MC + 10G
RFU
90000
30000
–30000
0 2 4 6 8 10 12 14
1600000 PSN 1
1.5% MC + 10G
1000000
RFU
400000
–200000
0 2 4 6 8 10 12 14
Time (min)
FIGURE 6.3 Effects of glycerol concentration in the buffer on CE-SSCP analysis of PSN1 and the wild-type
of K-ras gene. Experimental conditions: 1.5% MC in 50 mM Trisborate buffer; (a) without glycerol; (b) with
5% glycerol; (c) with 10% glycerol; electric field: 300 V/cm; samples were labeled with FAM. RFU represents
relative fluorescence unit. (From Endo, Y., et al., Electrophoresis, 26, 3380, 2005. With permission.)
on CE-SSCP analysis of PSN1 and the wild-type K-ras gene. Better resolution of single-stranded
DNAs was obtained with the addition of 5% glycerol to the buffer system. However, as it can be seen
in Figure 6.3c, the separation time was increased with increasing glycerol concentration. Analysis of
seven mutants of the K-ras gene was accomplished in 10 min by CE and in 1 min by microchip elec-
trophoresis under optimized separation conditions. Culiat et al. [58] identified point mutations in the
mouse genome induced by N-ethyl-N-nitrosourea using a new high-throughput mutation-scanning
technique: temperature-gradient CE (TGCE). TGCE detects the presence of partially denatured het-
eroduplex molecules that are resolved from homoduplexes through their differential mobilities during
CE conducted with a closely controlled temperature gradient. All TGCE analyses were conducted
using the SCE9610 Genetic Analysis System (Spectrumedix). The default running condition was a
broad 60–68o C temperature ramp with 5 kV injection voltage and 60 s injection time.
6.4.1.2 Forensic Application/Identity Testing

Individual DNA fingerprints can be provided by microsatellite- or SNP analysis and can be used in
forensic testing, paternity testing, and identification of suitable recipients for organ transplantation,
just to mention a few. A recent review summarizes the most relevant examples of analytical applica-
tions of CE techniques in the forensic field from 2001 till 2004 [25]. In an earlier review, Butler et al.
[24] detailed the characteristics of DNA typing with short tandem repeat (STR) markers widely used
for a variety of applications including human identification. Y-chromosome STRs (Y-STRs) pro-
vided valuable information in cases of rape and questioned paternity. This test also allowed genetic
identification of male lineages. In a study of Johnson et al. [59], a Y-STR 10-plex on a commercially
available automated genetic analyzer was validated for use in forensic and paternity laboratories. In
a comparison study, Dixon et al. [60] used artificially degraded samples and demonstrably obtained
a profile using SNPs instead of STRs for certain sample types with greater likelihood. A method for
simultaneous analysis of SNPs has been developed and validated to analyze highly degraded and
low copy number DNA templates (i.e., <100 pg) for scenarios including mass disaster identification.
The multiplex assay interrogated 20 autosomal noncoding loci—not in linkage disequilibrium—and
Amelogenin for sex determination. The samples were amplified in a single-tube PCR and analyzed
by an automated CE-based sequencer. Inagaki et al. [61] developed a novel 39-plex typing system
for SNP study. Thirty-seven genomic DNA fragments containing a total of 38 SNPs and one sex-
discriminating site were amplified in a multiplex PCR. Following the reaction, single nucleotide
primer extension reaction was performed by dividing these SNP loci into five groups. The SNP type
of each of the 39 loci was determined in the same run by an ABI PRISM™ 310 Genetic Analyzer
using POP-4 polymer in 47 cm × 0.050 mm ID capillaries applying a newly designed multi-injection
method. In this novel technique, the collection time was set to 1 or 2 min for the first four groups and
20 min for the last group. The method was applied in forensic cases of paternity testing and personal
identification.
An additional interesting forensic application of CE is the identification of differences to deter-
mine the species of botanical evidence found at a crime site or to associate a sample with a source.
Amplified fragment length polymorphism (AFLP) analysis of botanical forensic evidence provided
a means of obtaining a reproducible DNA profile in a relatively short period of time for species
with no available sequence information. Bless et al. [62] obtained AFLP profiles of 40 Acer rubrum
trees using an automated DNA fragment analyzer. This information could be used to link a piece
of evidence with a particular location or a suspect. Another important application for investi-
gating an accident or a crime is dog DNA profiling, as dogs are intensely integrated in human
social life. Several STR markers were analyzed for the individualization of dogs using an ABI
PRISM 3100 Genetic Analyzer with POP-4 gel in 36-cm capillary arrays using default instrument
settings [63].
Graft rejection is one of the most severe complications after allogenic transplants in leukemia.
Detection of increasing quantities of lymphoid and myeloid host cells might be predictive of graft
rejection and leukemia relapse, respectively. A sensitive and automatic method of quantifying the
degree of mixed chimerism after allogenic stem cell transplantation was based on PCR assays of
polymorphic STRs. CE allowed quantification of the recipient’s cells relative to the donor’s cells by
calculating ratios between STR alleles [64].
6.4.1.3 Diagnosis of Infectious Diseases

Identification of bacteria, viruses, and fungi is crucial for diagnostic and therapeutic purposes;
however, microbiologists often have problems with microorganisms that are difficult to culture.
Molecular methods on the other hand enable detection of these infectious agents without requir-
ing the presence of viable organisms to allow appropriate identification. PCR can be used to target
conserved stretches of DNA, similar to the region encoding 16S ribosomal RNA. These conserved
sequences enable robust applications for the identification of unknown or known infectious agents.
Several groups reported case histories in instances where infectious agents were identified based on
the sequence analysis of 16S rRNA gene (rDNA). Sequencing was carried out using the ABI Big
Dye cycle sequencing reaction kit with AmpliTaq FS DNA polymerase and the electrophoresis was
performed on an ABI 310 capillary analyzer. Analysis of the sequences and clustering was performed
by GeneCompar, version 2.0 (Applied Maths, Kortrijk, Belgium) [65,66]. In some cases, 16S rDNA
sequence determination does not seem to be sufficient for species identification just like in the case
of some Helicobacter species commonly found in dogs and cats. Phylogenetically, these species are
highly related to each other, thus their 16S rRNA gene sequences show >99% similarity. A multiplex
PCR method was described based on the tRNA intergenic spacers and in the urease gene, combined
with CE in the study. The PCR products were separated using an automated genetic analyzer and
the tRNA intergenic spacers amplified in tDNA-PCR were sequenced by using the BigDye Termi-
nator cycle sequencing kit (Applied Biosystems). The sequencing products were electrophoretically
separated by a commercially available multicapillary genetic analyzer. Their procedure was shown
to be very useful in determining the species identity of “Helicobacter heilmannii”-like organisms
observed in human stomachs and would facilitate research concerning their possible zoonotic impor-
tance [67]. CE also offers great versatility in studying viral systems. Krylova et al. [68] demonstrated
the utility of a P/ACE MDQ instrument for monitoring DNA release from virus particles in a run-
ning buffer of 25 mM sodium tetraborate containing 10 mM SDS. Drug development targeting viral
propagation requires fast and sensitive methods for in vitro monitoring of viral DNA release. A T5
bacteriophage/Escherichia coli K-12 model was reported to study molecular mechanisms of viral
infections and to evaluate antiviral drug candidates. The identification of clinically significant fungi
species was demonstrated by the amplification of noncoding internal transcribed spacer regions of
the rRNA gene and sequencing. All PCR/sequencing identifications from positive broths were in
agreement with the final species identification of the isolates grown from subculture. Early identifi-
cation of fungi by PCR/sequencing method may facilitate prompt and more appropriate antifungal
therapy [69].
An earlier review by Righetti and Gelfi [70] describes a number of applications using capillary
zone electrophoresis with sieving liquid polymers (mostly linear polyacrylamides and celluloses)
for the analysis of PCR products of clinically relevant, diagnostic DNA samples. The fields of
microbiology and virology were also covered, just like human genetics, quantitative gene dosage,
forensic medicine, and therapeutic DNA.
It has to be taken into account that microorganisms can also be separated, identified, and charac-
terized by CE without the necessity of using PCR. A comprehensive review of Desai and Armstrong
[71] gives details about these possibilities, such as the recent paper of Gao et al. [72] who detected
Staphylococcus aureus by a combination of monoclonal antibody-coated latex and CE. CZE separa-
tions were performed on a Beckman P/ACE MDQ System equipped with a photodiode array detector.
A 27-cm-long capillary column was used for the separation (20 cm to the detection window) applying
215 kV at a constant temperature of 25◦ C.
Control
1000
500
1p22 3q12 5q12 8q21 11q13 2pter 5pter 8pter 11pter 2qter 5qter 8qter 11qter
2q14 4q25 7q21 9q21 Xq11 3pter 6pter 9pter Xpter 3qter 6qter 9qter Xqter
6p21 10q22 1pter 4pter 7pter 10pter 1qter 4qter 7qter 10qter
Patient
2000
1000
1p22 3q12 5q12 8q21 11q13 2pter 5pter 8pter 11pter 2qter 5qter 8qter 11qter
2q14 4q25 7q21 9q21 Xq11 3pter 6pter 9pter Xpter 3qter 6qter 9qter Xqter
6p21 10q22 1pter 4pter 7pter 10pter 1qter 4qter 7qter 10qter
FIGURE 6.4 CE trace showing the deletion of the 3p telomere. The peak area of the 3p telomere is reduced
in the patient sample when compared to the control sample. (From Rooms, L., et al., Hum Mutat, 23, 17, 2004.
With permission.)
6.4.1.4 Molecular Karyotyping

Conventional karyotyping (chromosome banding) is one of the most widely used techniques in
routine cytogenetics and has been invaluable in the search for chromosomal aberrations causally
related to, for example, congenital mental retardation and malformation syndromes. It requires
culturing of cells from viable tissues but many clinical specimens may fail to provide adequate cel-
lularity for karyotyping and it also has a rather limited resolution (5–10 million bp). To solve these
problems, molecular karyotyping assays have been developed that allow sensitive and specific detec-
tion of single copy number changes of submicroscopic chromosomal regions throughout the entire
human genome. Among such methods are array-based comparative genomic hybridization (array
CGH), fluorescent in situ hybridization (FISH), and multiplex ligation-dependent probe amplifica-
tion (MLPA). MLPA is a recently developed technique, which is based on PCR amplification of
ligated probes hybridized to chromosome regions in question using unique primers attached to the
probes, making multiplexing easier. Only probes hybridized to a target sequence will be ligated and
subsequently amplified in the PCR. After separation by CGE, the peak area of each amplification
product reflects the relative copy number of that target sequence. Sequences that are deleted or
duplicated can also be easily identified, such as the deletion of the 3p telomere in Figure 6.4 [73,74].
However, it should be noted that balanced translocations could not be detected by CE methods at
this time.
6.4.1.5 Quantification of Cellular mRNA/Expression Analysis

Quantitative transcript analysis is usually carried out by microarray techniques, real-time PCR,
or even with serial analysis of gene expression (SAGE) [75–78]. An alternative of these tech-
niques is competitive template reverse transcriptase method described by Gilliland et al. [79] in
which the PCR mix contains a known amount of competitor DNA that differs from the cDNA
of interest only by having, for example, a mutated internal restriction enzyme site. There-
fore, the competitor and the unknown amount of target DNA compete for the same primers.
The ratio of products remains constant through the coamplification and can be readily quanti-
fied by CE and LIF [80,81]. In the paper of Warner et al. [82] bronchogenic carcinoma was
diagnosed with the help of standardized reverse transcriptase PCR (StaRT-PCR)—which is a
modification of competitive template reverse transcriptase method—since morphological analysis

of cytologic samples obtained by fine-needle aspirate or bronchoscopy was not sensitive enough
(65–80%). They measured the c-myc × E2F-1/p21 index in cDNA samples and found that
the index may augment cytomorphologic diagnosis of bronchogenic carcinoma biopsy sam-
ples, particularly those considered nondiagnostic by cytomorphologic criteria. After amplification,
each PCR product was analyzed by microchip electrophoresis on an Agilent 2100 Bioana-
lyzer instrument [82]. Relative quantification of mutated and normal mRNAs was carried out
by amplification refractory mutation sequencing PCR assay using 5-carboxyfluorescein dye-
labeled mutation-specific primers and a commercial DNA analyzer with standard protocols [83].
A method was designed and validated to detect a point mutation in the Janus tyrosine kinase
2 gene in patients with chronic myeloproliferative disorders. The proposed method might com-
plement current technologies based on genomic DNA analysis. Spyres et al. [84] described a
semiautomated PCR-based technology, called quantitative rapid analysis of gene expression (Q-
RAGE), which provided fast measurements of mRNA abundance with extremely high sensitivity
using fluorescent detection of the specific products. An array of sixteen 36-cm-long fused-
silica capillaries were used in an ABI 3100 instrument employing POP-4 separation polymer.
They claimed that their method was more sensitive compared to SAGE and its throughput was
higher than that of real-time PCR or StaRT-PCR. The flexibility of Q-RAGE makes it eminently
well suited for large-scale validation of microarray results and for more directed quantitative
studies of medium-sized sets of related genes, especially when low-abundance transcripts are
of interest.
6.4.2 DNA SEQUENCING

DNA sequencing enables to yield the greatest amount of information by identifying the order of
each deoxynucleotide base in a particular DNA molecule. With this knowledge, for example, one
can locate regulatory elements and gene sequences, make comparisons between homologous genes
across species, and identify mutations. In the 1970s, two DNA-sequencing methods were inde-
pendently developed. Maxam and Gilbert [85] used a “chemical cleavage protocol,” while Sanger
[86] designed a procedure similar to the natural process of DNA replication. Even though both
teams shared the 1980 Nobel Prize, the Sanger method became the industry standard because of its
practicality. Sanger’s method is based on dideoxy chain termination using ddNTPs in addition to
the normal nucleotides (dNTPs) in appropriate concentrations. When these ddNTPs are integrated
into the newly built-up nucleotide chain, they prohibit the addition of further nucleotides due to
the lack of their phosphodiester bond forming OH group, thus the DNA chain is terminated. When
different fluorescent labels are attached to each of the four reaction products (primer or terminator),
DNA sequencing can be carried out in one test tube as multispectral imaging can readily deter-
mine which fragment ends with what base by simple fluorescent signal based differentiation. In
the beginning of the various Genome Projects (Human, Mice, Rice, etc.), most laboratories did
DNA sequencing by means of very labor-intensive and time-consuming slab gel electrophoresis,
also facing difficulties with automation. The first paper on high-resolution single base separation
of DNA fragments using CGE was published in 1988 with the promise that the method was appli-
cable for automated DNA sequencing [87]. In this early work, cross-linked polyacrylamide gels
were utilized within the capillary, in analogy to slab gel electrophoresis [38]. Later, it was found
that the cross-linked matrix could not withstand osmotic shock occurring when salt plugs migrated
through the gel-filled capillary. Therefore, it became clear that non-cross-linked (so-called physical
gels) had much better characteristics in this respect. Initially, covalent coatings were utilized to
minimize EOF [88]; more recently, dynamic (i.e., adsorbed) coatings prevailed [89]. It was clear
that if CE were to be selected as the technology to sequence the Human Genome, researchers had
to find the means for automatic replenishment of the polymer matrix, as no one was interested
in reusing a polymer matrix due to fear of cross-contamination. Changing columns after each run
4 595–7 (a)
2
Florescence intensity (× 103)
80 85 90 95 100 105 110 115 120

40 595–7
(b)
697.9
30
768.9
20
10
70 75 80 85 90 95 100 105
697.9
4 595–7 768.9 (c)
869.70
1001.2
3
961.3
2
1
50 55 60 65 70 75 80
FIGURE 6.5 Separation of single-color DNA sequencing of larger than 590 bases by using (a) POP-6, (b)
MegaBACE matrix, and (c) quasi-IPN. Temperature: 60◦ C. Capillary, 40 cm effective length, 75 mm ID. DNA
injection, 41 V/cm and 30 s for POP-6, 50 V/cm and 40 s for MegaBACE matrix, 75 V/cm and 8 s for quasi-IPN;
running electric field strength, 200 V/cm for POP-6, 150 V/cm for MegaBACE matrix and quasi-IPN. (From
Wang, Y., et al., Electrophoresis, 26, 126, 2005. With permission.)
was also not a workable strategy. In 1990, Guttman [3] demonstrated that, through the use of
solutions of linear polyacrylamide and high pressure, it was possible to automatically replace the
polymer after each run and still maintain the high separation efficiency. This was a key step for-
ward, because it proved the potential of CE to become the basis of an automated DNA-sequencing
instrument. The technology was further developed for long read-length sequencing by Karger and
coworkers [90] in the late 1990s. It was also recognized that with longer DNA chains, reptation-
like migration became more significant and this stretching phenomenon reduced chain length-based
electrophoretic mobility differences to the point where separation was unsuccessful. After numerous
optimization and improvement steps by increasing the molecular weight of the sieving polymer,
1000 base read lengths were achieved in 1996 at high column temperatures (50◦ C) [91]. Today,
high-molecular-weight linear polymers are used routinely for DNA sequencing and further develop-
ments on DNA separation media are carried out by the use of copolymers, possessing high sieving
ability, low viscosity, and dynamic coating ability [92]. As it was difficult for homopolymers to
possess all the properties that can meet the different challenges, copolymers were used to com-
bine the desirable properties of the different monomers. On the basis of this principle, a range of
DNA separation matrices were developed featuring dynamic coating ability, low viscosity, and/or
adjustable viscosity covering a wide range of read length. The latest results are summarized in the
paper of Wang et al. [93] and Figure 6.5 compares the separation efficiency of two polymers and a
copolymer.
Another significant improvement in capillary-based DNA sequencing was multiplexing (i.e.,
using arrays of capillaries) [10]. This topic has been extensively reviewed by a number of groups,
just to list a few [5,94–97]; however, it is important to highlight the work of Kheterpal and Mathies
[98], Li and Yeung [99], Dovichi and Zhang [100], and Kamahori and Kambara [101] for their
contributions on automated multicapillary instrumentation.
6.4.3 PURITY CONTROL OF SYNTHETIC NUCLEOTIDES

6.4.4 OLIGONUCLEOTIDES
Oligonucleotides are commonly used for DNA amplification in PCR mixtures as primers, for in
situ hybridization techniques as probes for detecting PCR products, and for antiviral and anticancer
treatment as antisense oligonucleotide therapeutic agents, just to list the most important applica-
tions. In all of these instances, the oligonucleotides have to be of high purity, so their quality control
after synthesis is very important. All synthetic oligonucleotides should be tested for a defect in
length or sequence. The study of Willems et al. [102] showed that the combination of capillary zone
electrophoresis and electrospray ionization quadrupole-time of flight mass spectrometry allowed
the identification of oligonucleotides differing in length by only one nucleotide, and also any mis-
incorporated nucleotide by measuring tiny mass differences. Another approach for such tests was
published by Olson et al. [103] who modified an automated multicapillary DNA sequencer and
applied noncovalent fluorescent labeling.
6.4.4.1 Antisense DNA

Antisense therapeutics are synthetic oligonucleotides that have a complementary base sequence to a
target messenger RNA which encodes for disease-causing proteins or to the double-stranded DNA
which the mRNA is transcribed from. These molecules inactivate the genetic message, thus inhibiting
gene expression. Since oligonucleotides with a phosphodiester backbone are very susceptible to
nuclease degradation, DNA analogs with phosphorus-modified backbone (e.g., phosphorothioates
and methylphosphonates) are of high interest as they exhibit increased resistance against degradation.
Another type of antisense DNA is called as peptide nucleic acid. In these molecules, the deoxyribose-
phosphate backbone is substituted by a pseudopeptide backbone. Synthetic peptide nucleic acid
oligomers have been used in recent years in molecular biology procedures, diagnostic assays and
therapies. Some of these applications typically require stringent purity criteria for the antisense DNA
agents.
A paper by Chen and Gallo [104] on the instrumental and technological aspects of analyzing
antisense DNA in biological samples recommends the use of CGE to reveal the pharmacokinet-
ics of these molecules. An earlier review by DeDionisio and Lloyd [105] described a CGE-based
method for the purity analysis of antisense oligonucleotides after their chemical synthesis and for
a range of applications in the field of antisense-based technology. In addition, a CGE-based anal-
ysis method using eCAP™ ssDNA 100-R gel in a Beckman PACE 5510 instrument was presented
by the same group for the analysis of a N3 → P5 phosphoramidate/phosphorothioate chimera,
a second generation antisense oligonucleotide, prior and after purification [106]. More recently,
Malek and Khaledi [107] used CE with LIF (CE-LIF) detection to evaluate the effectiveness of
delivery and fate of a model 25mer DNA-based phosphorothioate antisense drug in HeLa cells.
Belenky et al. [108] described a CE method for sequence determination of antisense DNA analogs
of unknown base composition, while Froim et al. [109] developed a method of phosphorothioate
antisense DNA-sequencing analysis using UV detection (i.e., with no fluorescent or other labeling
requirement). Vilenchik et al. [110] applied high-performance CE for low level and rapid monitor-
ing of phosphorothioate DNA that modulated in vitro gene expression of human immunodeficiency
virus (HIV). The method was based on Watson-Crick hybridization between phosphodiester and
the target DNA. Various techniques of staining or labeling were investigated to improve detection
limits (0.1 ng/mL).
6.4.5 SEPARATION OF LARGE DNA MOLECULES (>2 kb)

Initial efforts in large DNA fragment analysis tended to apply constant electric fields. The size limit
of high resolution was demonstrably about 20,000 bp [111,112], while in case of low-resolution
requirement it was as high as 48.5 kb using linear polyacrylamide in CE [113]. DNA fragments
(72 bp to 23.1 kb) were separated in ultradilute HEC polymer solutions (<0.002% w/w) using high
voltage. On the basis of their results, Barron et al. [114] developed a transient entanglement coupling
mechanism theory for DNA separation, which suggested that there was no a priori upper size limit of
DNA separation by CE at a constant electric field. More recently, Chiu and Chang [115] demonstrated
that HEC solution at concentrations higher than its entanglement threshold provided good separation
for large DNA fragments in the presence of EOF at high pH. Separation of DNA fragments ranging
in size from 5 to 40 kb was completed in 6 min using 1.5% HEC, prepared in 20 mM methylamine-
borate (pH 12.0). A novel separation mechanism was demonstrated by Zheng and Yeung [116],
which was based on radial migration in CE with additionally applied hydrodynamic flow. With-
out the need for gel/polymer or complex instrumentation, this separation technique—while also
applicable for proteins, cells, and so on—was complementary to CGE and field-flow fractionation
techniques.
Plasmid analysis is often used in recombinant DNA technology. A new separation matrix, consist-
ing of poly(N-isopropylacrylamide) (PNIPAM) polymer and mannitol as small molecule additive,
was used for CE-based plasmid DNA separation. Supercoiled, linear, and nicked conformers of
lambda plasmid were separated in 1% PNIPAM + 6% mannitol. The effect of the applied mannitol
concentration on the separation quality is shown in Figure 6.6 [117].
Application of pulsed electric fields in conjunction with dilute polymer solutions are one way to
extend the size limit of DNA fragment analysis in CE. In the early 1990s, Heiger et al. [40] demon-
strated the use of pulsed-field electrophoresis to enhance the separation of large DNA molecules using
linear polyacrylamide matrix. Later, Kim and Morris [118] separated long-chain double-stranded
DNA (<10 kb and >1.5 Mb) in less than 4 min in buffers containing ultradilute polymer solutions.
They concluded that field inversion with higher peak amplitude in the forward direction than in the
reverse, but with equal pulse durations, provided the best resolution if running time was an issue.
Kabatek et al. [119] examined the behavior of large DNA fragments in pulsed-field CE under various
temperatures in short capillaries. Tseng et al. [120] reported the analysis of long DNA molecules by
nanoparticle-filled CE under the influences of hydrodynamic and electrokinetic forces. The polymer
composite prepared from gold nanoparticles and poly(ethyleneoxide) was filled into the capillary
column as separation matrix. Separations of lambda-DNA (0.12–23.1 kb) and high-molecular-weight
DNA markers (8.27–48.5 kb) by nanoparticle-filled CE, under the electric field strength of 140 V/cm
with the additional hydrodynamic flow velocity of 554 µm/s, were accomplished within 5 min. Elec-
trophoresis of very long DNA molecules (T4 DNA: 166 kb; S. pombe chromosomal DNA: 3–6 Mb)
in linear polyacrylamide solutions was investigated by fluorescence microscopy and CE [121]. It was
found that at higher polymer concentrations, the shape of the migrating DNA changed from U shape
to linear shape, and it was possible to separate the DNA molecules under this linear shape motion.
Actually, Mb-sized DNAs were well separated within 5 min in the region of the linear shape motion
by means of CE with a d.c. field. Considering that it usually takes 20 h to separate Mb-sized DNAs
by standard pulsed-field gel electrophoresis, this result proved extremely useful for the separation
of giant DNA molecules.
6.4.6 ANALYSIS OF MONONUCLEOTIDES AND NUCLEOSIDES

Deoxyribonucleic acid and ribonucleic acid, as well as their nucleotides, nucleosides, and base
constituents play an important role in many vital biochemical processes of medical interest. To
better understand these processes, fundamental investigations into the structure, occurrence, search
for modifications, and biochemical impact of structural variation are required. Thus, reliable high-
resolution analytical methods for the separation and identification of the nucleic acid constituents
(often at extremely low concentration levels) had to be developed. Chromatography (including
reversed-phase liquid chromatography), ion exchange chromatography, dHPLC, and electrophoresis
(a) 2%
15 20 25 30 35 40
(b) 4%
15 15 20 25 30 35 40
(c) 6%
15 20 25 30
(d) 8%
15 20 25 30 1353
1078
872
603
(e) 10%
310
281
271
234
194
118
75
15 20 25 30
FIGURE 6.6 Separation of _X174/HaeIII DNA by CE with: (a) 1.5% + PNIPAM + 2% mannitol; (b)
1.5% PNIPAM4% mannitol; (c) 1.5% PNIPAM + 6% mannitol; (d) 1.5% PNIPAM + 8% mannitol; (e) 1.5%
PNIPAM + 10% mannitol. Capillary, 75/365 µm i.d./o.d., and 32/40 cm efficient/total length; inject, 5 s at
−8 kV; separation electric field strength 220 V/cm. (From Zhou, P., et al., J Chromatogr A, 1083, 173, 2005.
With permission.)
(slab gel and other modes) were the major techniques that have been attempted for the analysis of
nucleic acids, nucleotides, nucleosides, and bases.
6.4.6.1 Nucleotides
Nucleotides are among the most important metabolites in a cell. Besides being the monomeric units of
the major nucleic acids of RNA and DNA, nucleotides are also required for numerous other important
functions within cells. These functions include (1) acting as energy carriers in phosphate transfer
reactions [adenosine 5 -triphosphate (ATP), guanosine 5 -triphosphate (GTP)]; (2) supplying the
basic structures for important compounds, such as cofactors (NAD+ , NADP+ , FAD, and coenzyme
A); (3) serving as mediators of several important cellular processes such as secondary messen-
gers in signal transduction events [cyclic adenosine 5 -monophosphate (cAMP), cyclic guanosine
5 -monophosphate (cGMP)]; (4) controlling many enzymatic reactions through allosteric effects
on enzyme activity; and (5) acting as activated intermediates in numerous biosynthetic reactions
(S-adenosylmethionine, also referred to as S-AdoMet). Nucleotide analyses by CE were reportedly
carried out in biochemical, biomedical, and pharmacological studies, as well as in the food industry
[122–128]. A new method based on pressure-assisted CE coupled to electrospray ionization mass
spectrometry was recently developed by Harada et al. [124] for anionic metabolome analysis, includ-
ing nucleotides. The key step of the method was that CE polarity was inverted from conventional
CE analysis for anions with the use of uncoated fused-silica capillary.
6.4.6.2 Nucleoside Analogs

Several nucleoside analogs are chemically synthesized and used as therapeutics, for example, to
inhibit specific enzymatic activities. Such analogs can be cytostatics (e.g., 6-mercaptopurine, 5-
fluorouracil, 5-iodo-2’-deoxyuridine, and 6-thioguanine), because they interfere with DNAsynthesis,
thus preferentially kill rapidly dividing cells like tumor cells. Another family of nucleoside analogs
has been used as therapeutics for viral infection treatment, such as for HIV. Azidothymidine (AZT)
was the first of such analogs, but followed by many others: ddI (dideoxyinosine), ddC (dideoxycyti-
dine), d4T (didehydro-deoxythymidine), and so forth. Nucleoside analogs are also used to suppress
the immune system after organ transplantation and to reduce the likelihood of transplant rejection by
the host. CZE has proved to be a suitable and useful method for the determination of acid–base dissoci-
ation constants (pKa ) of new synthetic compounds of amino- and (amino)guanidinopurine nucleoside
analogs, such as acyclic nucleoside phosphonate, acyclic nucleoside phosphonate diesters, and
other related compounds [129]. Determination of this characteristic seemed to be very impor-
tant in the optimization stage of new drug development projects to understand their passage and
metabolism.
In an earlier study, Balayiannis et al. [130] developed a capillary zone electrophoresis method for
the analysis of a series of novel synthetic dideoxynucleoside analogs with potential anti-HIV activity.
The method was readily applied for purity testing as well as to resolve cis and trans diastereomers.
The purity and separation of diastereomers of such analogs are of great importance for further testing
of their biological and pharmacological activity.
6.4.6.3 DNA Adducts

The initial step in chemical carcinogenesis is believed to be the covalent attachment of a chemical
to DNA to form DNA adducts, like the extensively studied 8-hydroxyguanine. DNA adducts are
most often used as molecular dosimeters of exposure. If not repaired correctly, these modifications
may lead to mutations and eventually to cancer, in particular if the adduct is located in an oncogene
or in a tumor suppressor gene. Therefore, significant efforts are being made to understand how the
diversity in DNA adduct conformations can affect cellular responses to DNA damage. Analysis of
DNA adducts can be carried out on the nucleotide or nucleoside level after enzymatic digestion.
Analytical methods used for the detection of DNA adducts must have excellent sensitivity, since
only 1 in 106 to 1 in 1012 bases are usually modified that represents a great challenge. Next to
immunoassays and fluorescence-based assays, the most common method is 32 P-postlabeling, but
CE-MS holds the potential to replace it in the near future. Markushin et al. [131] recently reported
that catechol estrogen quinines-derived DNA adducts are present in urine samples from subjects with
prostate cancer. Their finding suggests that if it can be repeated in larger sample sizes, the presence
of depurinating adducts in human urine samples may be used as prostate cancer biomarker.
6.5 FUTURE PROSPECTIVE

Recent technological developments in nucleic acid analysis by electric field mediated separation
techniques are tending toward miniaturization with the goal to provide high-speed and low-cost
clinical analysis tools. CE, and more recently microchip electrophoresis, features the use of small
sample volumes, low reagent consumption, and equally important in a high-throughput clinical
environment, almost negligible waste. Another essential impact of miniaturization in capillary and
microchip-based techniques is the potential for system integration of several processing steps, such
as sample preparation, desalting, PCR amplification, restriction digestion, separation, fraction col-
lection, and so forth. The speed of CE and microchip electrophoresis is significantly higher than
that of any other conventional electrophoresis methods, and its cost is also favorable, especially
in a disposable plastic chip format. Multicapillary and microchip-based bioanalytical methods will
probably play a significant role in the future of biomedical and clinical applications [4].
ACKNOWLEDGMENTS
This work was supported by the European Community Marie Curie Chair 006733. Eszter Szantai
gratefully acknowledges the grant of the Austrian Ministry of Arts, Science and Education’s GenAu
program.
REFERENCES
1. Watson, J.D. and Crick, F.H., Molecular structure of nucleic acids: a structure for deoxyribose nucleic
acid, Nature, 171, 737, 1953.
2. Nelson, R.J., et al., Use of peltier thermoelectric devices to control column temperature in high-
performance capillary electrophoresis, J Chromatogr A, 480, 111, 1989.
3. Guttman, A., US patent #5,332,481, Capillary Electrophoresis Using Replaceable Gels, 1994.
4. Heller, M. and Guttman, A. (eds), Integrated Microfabricated BioDevices: Advanced Technologies for
Genomics, Drug Discovery, Bioanalysis, and Clinical Diagnostics, Marcel Dekker, New York, 2001.
5. Karger, B.L. and Guttman, A., Capillary electrophoresis and the human genome project,
Genomic/Proteomic Technol., 3, 14, 2003.
6. Lander, E.S., et al., Initial sequencing and analysis of the human genome, Nature, 409, 860, 2001.
7. Venter, J.C., et al., The sequence of the human genome, Science, 291, 1304, 2001.
8. Andrews, A.T., Electrophoresis, Claredon Press, Oxford, 1986.
9. Gelfi, C., et al., Analysis of antisense oligonucleotides by capillary electrophoresis, gel-slab
electrophoresis, and HPLC: a comparison, Antisense Nucleic Acid Drug Dev, 6, 47, 1996.
10. Mathies, R.A. and Huang, X.C., Capillary array electrophoresis: an approach to high-speed, high-
throughput DNA sequencing, Nature, 359, 167, 1992.
11. Barbier, V. and Viovy, J.L., Advanced polymers for DNA separation, Curr Opin Biotechnol, 14, 51,
2003.
12. Guttman, A., Gel and polymer-solution mediated separation of biopolymers by capillary electrophore-
sis, J Chromatogr Sci, 41, 449, 2003.
13. Lin, Y.W., Chiu, T.C., and Chang, H.T., Laser-induced fluorescence technique for DNA and proteins
separated by capillary electrophoresis, J Chromatogr B Analyt Technol Biomed Life Sci, 793, 37, 2003.
14. Quigley, W.W. and Dovichi, N.J., Capillary electrophoresis for the analysis of biopolymers, Anal
Chem, 76, 4645, 2004.
15. Slater, G.W., et al., The theory of DNA separation by capillary electrophoresis, Curr Opin Biotechnol,
14, 58, 2003.
16. Xu, F. and Baba, Y., Polymer solutions and entropic-based systems for double-stranded DNA capillary
electrophoresis and microchip electrophoresis, Electrophoresis, 25, 2332, 2004.
17. Andersen, P.S., et al., Capillary electrophoresis-based single strand DNA conformation analysis in
high-throughput mutation screening, Hum Mutat, 21, 455, 2003.
18. Babu, C.V.S., et al., Capillary electrophoresis at the omics level: towards systems biology,
19. Godde, R., et al., Electrophoresis of DNA in human genetic diagnostics—state-of-the-art, alternatives
and future prospects, Electrophoresis, 27, 939, 2006.
20. Hempel, G., Biomedical applications of capillary electrophoresis, Clin Chem Lab Med, 41, 720, 2003.
21. Jabeen, R., et al., Capillary electrophoresis and the clinical laboratory, Electrophoresis, 27, 2413,
2006.
22. Landers, J.P., Clinical capillary electrophoresis, Clin Chem, 41, 495, 1995.
23. Mitchelson, K.R., The use of capillary electrophoresis for DNA polymorphism analysis, Mol
Biotechnol, 24, 41, 2003.
24. Butler, J.M., et al., Forensic DNA typing by capillary electrophoresis using the ABI Prism 310 and
3100 genetic analyzers for STR analysis, Electrophoresis, 25, 1397, 2004.
25. Tagliaro, F. and Bortolotti, F., Recent advances in the applications of CE to forensic sciences (2001–
2004), Electrophoresis, 27, 231, 2006.
26. Guttman, A., On the separation mechanism of capillary SDS gel electrophoresis of proteins,
27. Ogston, A.G., The spaces in a uniform random suspension of fibres, Trans Faraday Soc, 54, 1754,
1958.
28. Grossman, P.D., Menchen, S., and Hershey, D., Quantitative analysis of DNA-sequencing
electrophoresis, GATA, 9, 9, 1992.
29. Ferguson, K.A., Starch-gel electrophoresis: application to the classification of pituitary proteins and
polypeptides, Metab Clin Exp, 13, 985, 1964.
30. Lumpkin, O.J., Dejardin, P., and Zimm, B.H., Theory of gel electrophoresis of DNA, Biopolymers,
24, 1573, 1985.
31. De Gennes, P.G., Scaling Concept in Polymer Physics, Cornell University Press, Ithaca, NY, 1979.
32. Slater, G.W. and Noolandi, J., The biased reptation model of DNA gel electrophoresis: mobility vs.
molecular size and gel concentration, Biopolymers, 28, 1781, 1989.
33. Viovy, J.L. and Duke, T., DNA electrophoresis in polymer solutions: Ogston sieving, reptation and
constraint release, Electrophoresis, 14, 322, 1993.
34. Terabe, S., Otsuka, K., and Ando, T., Band broadening in electrokinetic chromatography with micellar
solutions and open-tubular capillaries, Anal Chem, 61, 2434, 1989.
35. Karger, B.L., Cohen, A.S., and Guttman, A., High performance capillary electrophoresis in the
biological sciences, J Chromatogr, 492, 585, 1989.
36. Mullis, K.B., Target amplification for DNA analysis by the polymerase chain reaction, Ann Biol Clin
(Paris), 48, 579, 1990.
37. Cohen, A.S., et al., Rapid separation of DNA restriction fragments using capillary electrophoresis,
J Chromatogr, 458, 323, 1988.
38. Cohen, A.S., et al., Rapid separation and purification of oligonucleotides by high-performance capillary
gel electrophoresis, Proc Natl Acad Sci USA, 85, 9660, 1988.
39. Kasper, T.J., et al., Separation and detection of DNA by capillary electrophoresis, J Chromatogr, 458,
303, 1988.
40. Heiger, D.N., Cohen, A.S., and Karger, B.L., Separation of DNA restriction fragments by high perfor-
mance capillary electrophoresis with low and zero crosslinked polyacrylamide using continuous and
pulsed electric fields, J Chromatogr, 516, 33, 1990.
41. Schwartz, H.E., et al., Analysis of DNA restriction fragments and polymerase chain reaction products
towards detection of the AIDS (HIV-1) virus in blood, J Chromatogr, 559, 267, 1991.
42. Zhu, M., et al., Factors affecting free zone electrophoresis and isoelectric focusing in capillary
electrophoresis, J Chromatogr A, 480, 311, 1989.
43. Erill, N., et al., A novel multiplexing, polymerase chain reaction-based assay for the analysis of
chromosome 18q status in colorectal cancer, J Mol Diagn, 7, 478, 2005.
44. Hyytia-Trees, E., et al., Second generation subtyping: a proposed PulseNet protocol for multiple-
locus variable-number tandem repeat analysis of Shiga toxin-producing Escherichia coli O157 (STEC
O157), Foodborne Pathog Dis, 3, 118, 2006.
45. Sell, S.M. and Lugemwa, P.R., Development of a highly accurate, rapid PCR-RFLP genotyping assay
for the methylenetetrahydrofolate reductase gene, Genet Test, 3, 287, 1999.
46. Carrera, P., et al., Amplification refractory mutation system analysis of point mutations by capillary
electrophoresis, Methods Mol Biol, 163, 95, 2001.
47. Gomez-Llorente, C., et al., Multiplex analysis of the most common mutations related to hereditary
haemochromatosis: two methods combining specific amplification with capillary electrophoresis, Eur
J Haematol, 72, 121, 2004.
48. Barany, F., The ligase chain reaction in a PCR world, PCR Methods Appl, 1, 5, 1991.
49. Thomas, G., et al., Capillary and microelectrophoretic separations of ligase detection reaction products
produced from low-abundant point mutations in genomic DNA, Electrophoresis, 25, 1668, 2004.
50. Brazill, S.A. and Kuhr, W.G., A single base extension technique for the analysis of known mutations
utilizing capillary gel electrophoresis with electrochemical detection, Anal Chem, 74, 3421, 2002.
51. Kozlowski, P., Olejniczak, M., and Krzyzosiak, W.J., Rapid heteroduplex analysis by capillary
electrophoresis, Clin Chim Acta, 353, 209, 2005.
52. Tian, H., et al., Effective capillary electrophoresis-based heteroduplex analysis through optimization
of surface coating and polymer networks, Anal Chem, 72, 5483, 2000.
53. Weber, J., et al., Improving sensitivity of electrophoretic heteroduplex analysis using nucleosides as
additives: application to the breast cancer predisposition gene BRCA2, Electrophoresis, 27, 1444,
2006.
54. Orita, M., et al., Detection of polymorphisms of human DNA by gel electrophoresis as single-strand
conformation polymorphisms, Proc Natl Acad Sci USA, 86, 2766, 1989.
55. Hofman-Bang, J., et al., High-efficiency multiplex capillary electrophoresis single strand conformation
polymorphism (multi-CE-SSCP) mutation screening of SCN5A: a rapid genetic approach to cardiac
arrhythmia, Clin Genet, 69, 504, 2006.
56. Holmila, R. and Husgafvel-Pursiainen, K., Analysis of TP53 gene mutations in human lung can-
cer: comparison of capillary electrophoresis single strand conformation polymorphism assay with
denaturing gradient gel electrophoresis and direct sequencing, Cancer Detect Prev, 30, 1, 2006.
57. Endo, Y., et al., Effect of polymer matrix and glycerol on rapid single-strand conformation polymor-
phism analysis by capillary and microchip electrophoresis for detection of mutations in K-ras gene,
58. Culiat, C.T., et al., Identification of mutations from phenotype-driven ENU mutagenesis in mouse
chromosome 7, Mamm Genome, 16, 555, 2005.
59. Johnson, C.L., et al., Validation and uses of a Y-chromosome STR 10-plex for forensic and paternity
laboratories, J Forensic Sci, 48, 1260, 2003.
60. Dixon, L.A., et al., Validation of a 21-locus autosomal SNP multiplex for forensic identification
purposes, Forensic Sci Int, 154, 62, 2005.
61. Inagaki, S., et al., A new 39-plex analysis method for SNPs including 15 blood group loci, Forensic
Sci Int, 144, 45, 2004.
62. Bless, C., Palmeter, H., and Wallace, M.M., Identification of Acer rubrum using amplified fragment
length polymorphism, J Forensic Sci, 51, 31, 2006.
63. Eichmann, C., et al., Estimating the probability of identity in a random dog population using 15 highly
polymorphic canine STR markers, Forensic Sci Int, 151, 37, 2005.
64. Fernandez-Aviles, F., et al., Serial quantification of lymphoid and myeloid mixed chimerism using
multiplex PCR amplification of short tandem repeat-markers predicts graft rejection and relapse,
respectively, after allogeneic transplantation of CD34+ selected cells from peripheral blood, Leukemia,
17, 613, 2003.
65. De Baere, T., et al., Bacteremic infection with Pantoea ananatis, J Clin Microbiol, 42, 4393, 2004.
66. Vaneechoutte, M., et al., Isolation of Moraxella canis from an ulcerated metastatic lymph node, J Clin
Microbiol, 38, 3870, 2000.
67. Baele, M., et al., Multiplex PCR assay for differentiation of Helicobacter felis, H. bizzozeronii, and
H. salomonis, J Clin Microbiol, 42, 1115, 2004.
68. Krylova, S.M., et al., Monitoring viral DNA release with capillary electrophoresis, Analyst,
129, 1234, 2004.
69. Pryce, T.M., et al., Rapid identification of fungal pathogens in BacT/ALERT, BACTEC, and BBL
MGIT media using polymerase chain reaction and DNA sequencing of the internal transcribed spacer
regions, Diagn Microbiol Infect Dis, 54, 289, 2006.
70. Righetti, P.G. and Gelfi, C., Capillary electrophoresis of DNA for molecular diagnostics, Electrophore-
sis, 18, 1709, 1997.
71. Desai, M.J. and Armstrong, D.W., Separation, identification, and characterization of microorganisms
by capillary electrophoresis, Microbiol Mol Biol Rev, 67, 38, 2003.
72. Gao, P., et al., Rapid detection of Staphylococcus aureus by a combination of monoclonal antibody-
coated latex and capillary electrophoresis, Electrophoresis, 27, 1784, 2006.
73. Rooms, L., et al., Subtelomeric deletions detected in patients with idiopathic mental retardation using
multiplex ligation-dependent probe amplification (MLPA), Hum Mutat, 23, 17, 2004.
74. Slater, H.R., et al., Rapid, high throughput prenatal detection of aneuploidy using a novel quantitative
method (MLPA), J Med Genet, 40, 907, 2003.
75. Cai, Y., et al., Abnormal expression of Smurf2 during the process of rat liver fibrosis, Chin J Dig Dis,
7, 237, 2006.
76. Lee, S., et al., Gene expression profiles in acute myeloid leukemia with common translocations using
SAGE, Proc Natl Acad Sci USA, 103, 1030, 2006.
77. Nakamura, N., et al., Isolation and expression profiling of genes upregulated in bone marrow-derived
mononuclear cells of rheumatoid arthritis patients, DNA Res, 13, 169, 2006.
78. Sanada, Y., et al., Down-regulation of the claudin-18 gene, identified through serial analysis of gene
expression data analysis, in gastric cancer with an intestinal phenotype, J Pathol, 208, 633, 2006.
79. Gilliland, G., et al., Analysis of cytokine mRNA and DNA: detection and quantitation by competitive
polymerase chain reaction, Proc Natl Acad Sci USA, 87, 2725, 1990.
80. Borson, N.D., et al., Direct quantitation of RNA transcripts by competitive single-tube RT–PCR and
capillary electrophoresis, Biotechniques, 25, 130, 1998.
81. Williams, S.J. and Williams, P.M., Quantitation of mRNA by competitive PCR using capillary
electrophoresis, Methods Mol Biol, 163, 243, 2001.
82. Warner, K.A., et al., The c-myc × E2F-1/p21 interactive gene expression index augments cyto-
morphologic diagnosis of lung cancer in fine-needle aspirate specimens, J Mol Diagn, 5, 176,
2003.
83. Vannucchi, A.M., et al., A quantitative assay for JAK2(V617F) mutation in myeloproliferative
disorders by ARMS-PCR and capillary electrophoresis, Leukemia, 20, 1055, 2006.
84. Spyres, L., et al., Quantitative high-throughput measurement of gene expression with sub-zeptomole
sensitivity by capillary electrophoresis, Anal Biochem, 345, 284, 2005.
85. Maxam, A.M. and Gilbert, W., A new method for sequencing DNA, Proc Natl Acad Sci USA, 74, 560,
1977.
86. Sanger, F. and Coulson, A.R., A rapid method for determining sequences in DNA by primed synthesis
with DNA polymerase, J Mol Biol, 94, 441, 1975.
87. Guttman, A., et al., Capillary Gel Electrophoresis: High Resolution and MicropreparativeApplications,
in Electrophoresis’88, Schafer-Nielsen, C. (Ed.), VCH, Weinheim, Germany, 1988, p. 151.
88. Hjerten, S., High-performance electrophoresis: elimination of electroendosmosis and solute adsorp-
tion, J Chromatogr, 347, 191, 1985.
89. Madabhushi, R.S., Separation of 4-color DNA sequencing extension products in noncovalently coated
capillaries using low viscosity polymer solutions, Electrophoresis, 19, 224, 1998.
90. Goetzinger, W., et al., Characterization of high molecular mass linear polyacrylamide powder pre-
pared by emulsion polymerization as a replaceable polymer matrix for DNA sequencing by capillary
91. Carrilho, E., et al., Rapid DNA sequencing of more than 1000 bases per run by capillary electrophoresis
using replaceable linear polyacrylamide solutions, Anal Chem, 68, 3305, 1996.
92. Barbier, V., et al., Comb-like copolymers as self-coating, low-viscosity and high-resolution matrices
for DNA sequencing, Electrophoresis, 23, 1441, 2002.
93. Wang, Y., et al., Quasi-interpenetrating network formed by polyacrylamide and poly(N,N-
dimethylacrylamide) used in high-performance DNA sequencing analysis by capillary electrophoresis,
94. Carrilho, E., DNA sequencing by capillary array electrophoresis and microfabricated array systems,
95. Dovichi, N.J., DNA sequencing by capillary electrophoresis, Electrophoresis, 18, 2393, 1997.
96. Dovichi, N.J. and Zhang, J.-Z., Methods in Molecular Biology, Totowa, N.J. (Ed.), United States,
2001, p. 225.
97. Yeung, E.S. and Li, Q., Chemical Analysis, New York, 1998, 767.
98. Kheterpal, I. and Mathies, R.A., Capillary array electrophoresis DNA sequencing, Anal Chem, 71,
31A, 1999.
99. Li, Q. and Yeung, E.W., Simple two-color base-calling schemes for DNA sequencing based on standard
four-label Sanger chemistry, Appl Spectrosc, 49, 1528, 1995.
100. Dovichi, N.J. and Zhang, J., How capillary electrophoresis sequenced the human genome this essay
is based on a lecture given at the Analytica 2000 conference in Munich (Germany) on the occasion of
the Heinrich-Emanuel-Merck prize presentation, Angew Chem Int Ed Engl, 39, 4463, 2000.
101. Kamahori, M. and Kambara, H., Capillary array electrophoresis analyzer, Methods Mol Biol, 163,
271, 2001.
102. Willems, A.V., et al., Development of a quality control method for the characterization of oligonu-
cleotides by capillary zone electrophoresis-electrospray ionization-quadrupole time of flight-mass
spectrometry, Electrophoresis, 26, 1412, 2005.
103. Olson, N.A., Khandurina, J., and Guttman, A., DNA profiling by capillary array electrophoresis with
non-covalent fluorescent labeling, J Chromatogr A, 1051, 155, 2004.
104. Chen, S.H. and Gallo, J.M., Use of capillary electrophoresis methods to characterize the pharmacoki-
netics of antisense drugs, Electrophoresis, 19, 2861, 1998.
105. DeDionisio, L.A. and Lloyd, D.H., Progress in HPLC-HPCE, VSP, Hayward, 1997, p. 293.
106. DeDionisio, L.A., Raible, A.M., and Nelson, J.S., Analysis of an oligonucleotide N3 →P5 phos-
phoramidate/phosphorothioate chimera with capillary gel electrophoresis, Electrophoresis, 19, 2935,
1998.
107. Malek, A.H. and Khaledi, M.G., Monitoring liposome-mediated delivery and fate of an antisense
drug in cell extracts and in single cells by capillary electrophoresis with laser-induced fluorescence,
108. Belenky, A., Smisek, D.L., and Cohen, A.S., Sequencing of antisense DNA analogues by capillary gel
electrophoresis with laser-induced fluorescence detection, J Chromatogr A, 700, 137, 1995.
109. Froim, D., et al., Method for phosphorothioate antisense DNA sequencing by capillary electrophoresis
with UV detection, Nucleic Acids Res, 25, 4219, 1997.
110. Vilenchik, M., Belenky, A., and Cohen, A.S., Monitoring and analysis of antisense DNA by high-
performance capillary gel electrophoresis, J Chromatogr A, 663, 105, 1994.
111. Pariat, Y.F., et al., Separation of DNA fragments by capillary electrophoresis using replaceable linear
polyacrylamide matrices, J Chromatogr A, 652, 57, 1993.
112. Strege, M. and Lagu, A., Separation of DNA restriction fragments by capillary electrophoresis using
coated fused silica capillaries, Anal Chem, 63, 1233, 1991.
113. Guszczynski, T., et al., Capillary zone electrophoresis of large DNA, Electrophoresis, 14, 523, 1993.
114. Barron, A.E., Blanch, H.W., and Soane, D.S., A transient entanglement coupling mechanism for DNA
separation by capillary electrophoresis in ultradilute polymer solutions, Electrophoresis, 15, 597, 1994.
115. Chiu, T.C. and Chang, H.T., Comparison of the separation of large DNA fragments in the presence
and absence of electroosmotic flow at high pH, J Chromatogr A, 979, 299, 2002.
116. Zheng, J. and Yeung, E.S., Mechanism for the separation of large molecules based on radial migration
in capillary electrophoresis, Anal Chem, 75, 3675, 2003.
117. Zhou, P., et al., Electrophoretic separation of DNA using a new matrix in uncoated capillaries,
J Chromatogr A, 1083, 173, 2005.
118. Kim, Y. and Morris, M.D., Ultrafast high resolution separation of large DNA fragments by pulsed-field
119. Kabatek, Z., Kleparnik, K., and Gas, B., Effect of temperature on the separation of long DNA fragments
in polymer solution, J Chromatogr A, 916, 305, 2001.
120. Tseng, W.L., et al., Nanoparticle-filled capillary electrophoresis for the separation of long DNA
molecules in the presence of hydrodynamic and electrokinetic forces, Electrophoresis, 26, 3069,
2005.
121. Ueda, M., et al., Electrophoresis of long DNA molecules in linear polyacrylamide solutions, Biophys
Chem, 71, 113, 1998.
122. Geldart, S.E. and Brown, P.R., Analysis of nucleotides by capillary electrophoresis, J Chromatogr A,
828, 317, 1998.
123. Grob, M.K., et al., Optimization of cellular nucleotide extraction and sample preparation for nucleotide
pool analyses using capillary electrophoresis, J Chromatogr B Analyt Technol Biomed Life Sci, 788,
103, 2003.
124. Harada, K., Fukusaki, E., and Kobayashi, A., Pressure-assisted capillary electrophoresis mass spec-
trometry using combination of polarity reversion and electroosmotic flow for metabolomics anion
analysis, J Biosci Bioeng, 101, 403, 2006.
125. Huang, M., et al., High resolution separation and quantitation of ribonucleotides using capillary
electrophoresis, Anal Biochem, 207, 231, 1992.
126. Nguyen, A.L., Luong, J.H., and Masson, C., Determination of nucleotides in fish tissues using capillary
electrophoresis, Anal Chem, 62, 2490, 1990.
127. Soga, T., et al., Quantitative metabolome analysis using capillary electrophoresis mass spectrometry,
J Proteome Res, 2, 488, 2003.
128. Soga, T., et al., Simultaneous determination of anionic intermediates for Bacillus subtilis metabolic
pathways by capillary electrophoresis electrospray ionization mass spectrometry, Anal Chem, 74,
2233, 2002.
129. Solinova, V., et al., Determination of acid-base dissociation constants of amino- and guanidinopurine
nucleotide analogs and related compounds by capillary zone electrophoresis, Electrophoresis, 27,
1006, 2006.
130. Balayiannis, G., Papaioannou, D., and Karamanos, N.K., Application of capillary electrophoresis in
the analysis of novel synthetic dideoxynucleoside analogues with potential anti-HIV activity, Biomed
Chromatogr, 15, 271, 2001.
131. Markushin, Y., et al., Potential biomarker for early risk assessment of prostate cancer, Prostate, 2006.
7 Analysis of Carbohydrates by
Capillary Electrophoresis
Julia Khandurina
CONTENTS
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

7.2 Background: High-Performance CE Separations of Carbohydrates. . . . . . . . . . . . . . . . . . . . . . . 252
7.3 Theoretical and Practical Aspects of Modern Carbohydrate Analysis: CE-MS. . . . . . . . . . . 255
7.3.1 Basic Underlying Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
7.3.2 CE-MS Characterization of Simple Sugars and Glycoconjugates . . . . . . . . . . . . . . . . 257
7.4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
7.4.1 Complex Polysaccharides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
7.4.2 Carbohydrates in Pharmaceuticals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
7.4.2.1 Glycoproteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
7.4.2.2 Analysis of Complex Polysaccharides in Drugs . . . . . . . . . . . . . . . . . . . . . . . . . 268
7.4.2.3 Glucuronidation of Drugs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
7.4.2.4 Characterization of Oligosaccharides in Pathogenic Bacteria . . . . . . . . . . 271
7.4.2.5 Sugar Analysis in Traditional Chinese Drugs. . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
7.4.3 Carbohydrates in Foods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
7.4.4 Bioindustrial Applications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
7.4.5 Miniaturization in Carbohydrate Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278
7.5 Method Development Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
7.6 Conclusions and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
7.1 INTRODUCTION
Carbohydrates are widely distributed in nature. Generally speaking, they represent a class of poly-
hydroxylated compounds, which contain a carbonyl functional group, aldehyde or ketone, often
as semiacetal or hemiketal. The abundance of hydroxyl groups and a carbonyl functionality result
in the formation of diverse oligomers and macromolecules. Carbohydrate biopolymers, oligo- and
polysaccharides, are often very structurally complex comprising monosaccharides linked by gly-
cosidic bonds. The most common examples are α- and β-1,4-linked polyglucans, for example,
cellulose, the main constituent of plant cell walls, or starch, which is present in large quantities in
plant seeds and roots. Many fruits contain high concentration of oligosaccharides based on glucose,
fructose, and raffinose building blocks. On the other hand, bacterial cell walls are built of complex
glycans containing glucosamine, muramic acid, and oligosaccharide derivatives of various monosac-
charides. Various kinds of carbohydrates play important roles in animal and human tissues supporting
their living activities. For example, glucose is a major constituent of blood that generates energy via
TCA metabolic cycle (tricarboxylic acid cycle, also known as citric acid cycle, or the Krebs cycle).
Glycogen, a highly branched glucose polymer, is the principal energy storage in animal and human
251
cells. Abnormal increase of glucose in blood is a symptom of diabetes; therefore, rapid estimation of
its level is an important task in clinical applications. Polymerized glucosamine constructs hard shell
animal and insect substances. Besides, carbohydrates exist in living organisms in various forms of
glycoconjugates. Glycan moieties in glycoconjugates, where oligosaccharide chains are covalently
linked to an aglycon (e.g., protein or lipid), are very structurally and functionally diverse. They are
responsible for key functions in protein folding and stability, biological recognition, and cell–cell
interaction processes [1–3]. Glycoconjugates act as receptors on the cell surface. Structures of glyco-
moieties range from relatively simple, as in glycosphingolipids (GSLs) and some glycoproteins, to
extraordinary complex mucins and proteoglycans (PGs). Saccharides have an enormous potential to
carry information, far exceeding that of proteins and nucleic acids [4], due to diversity of monomer
building blocks and linkages between them, as well as substitute groups. Recent advances in the
research of biological roles of glycans in glycoconjugates call for the development of more sensitive
and reproducible analytical methods to determine their compositions and structures.
In spite of the apparent significance, carbohydrate separation and analysis technology has not been
as universal in life sciences and biotechnology as nucleic acid and protein analysis. Heterogeneity of
oligo- and polysaccharides in molecular weight, primary sequence and branching, linkage, variety
of sugar structural isoforms, and lack of chromophore/fluorophore or active functional groups have
presented a challenge. A number of analytical methods have been applied to carbohydrate analysis,
including paper and thin-layer chromatography [5,6], gas chromatography [8], high-performance
liquid chromatography (HPLC) [7], high-pH anion-exchange chromatography (HPAEC) [9–12],
supercritical fluid chromatography [13], affinity chromatography [14], polyacrylamide slab gel elec-
trophoresis and capillary electrophoresis (CE) as separation, and mass spectrometry and nuclear
magnetic resonance spectroscopy as structural characterization techniques [15]. CE features the
highest separation efficiency in carbohydrate analysis compared with the other reported methods.
A number of excellent reviews on carbohydrate analysis by CE have been published in the past
[16–21], which described the method in detail.
This chapter is intended to summarize most significant advances in carbohydrate analysis by
CE-based techniques during the past decade. Clearly, it is no longer possible to write a compre-
hensive review of the massive literature published recently, which includes a variety of different
aspects and applications involving carbohydrate analysis. Therefore, the focus has been made on
the most significant and interesting, in the author’s opinion, developments in high separation effi-
ciency of CE, hyphenated techniques, in particular, CE-mass spectrometry (CE-MS) and analysis
of complex polysaccharides and glycoconjugates. In addition, special attention has been paid to
specific applications, recently emerging in pharmaceutical and bioindustrial arenas. These include
analysis of polysaccharides and glycoconjugates in drugs, analysis of traditional Chinese medicines,
and approaches to high-throughput carbohydrate profiling at the large industrial scale, particularly,
problems associated with the processes of biomass conversion to fermentable sugars. Recent trends
toward miniaturization of CE carbohydrate analysis have also been reviewed.
7.2 BACKGROUND: HIGH-PERFORMANCE CE SEPARATIONS OF

CARBOHYDRATES
Numerous research papers and reviews on carbohydrate separations by CE have been written for the
past several years. Researches have successfully addressed problems, such as tremendous diversity
and complexity of this class of compounds, polar and neutral nature of most carbohydrates, their
low ultraviolet (UV) extinction coefficients, and lack of functional groups. In the previous edition
of this book, Olechno and Nolan [16] published a comprehensive overview of the CE separation
techniques, attempted and developed for intact and derivatized carbohydrates, charged and neutral,
as well as detection approaches by UV, indirect fluorescence, electrochemical (e.g., amperometric)
detection, refractive index, and laser-induced fluorescence (LIF). A variety of buffer systems were
Analysis of Carbohydrates by Capillary Electrophoresis 253
introduced, depending on the nature of mono- and oligosaccharides of interest: acidic and basic,
using normal or reversed polarity, either with electroosmotic flow (EOF) present or suppressed.
For more details and specific examples, the reader is directed to Reference 16. In brief, the most
popular buffers utilized have been phosphates (pH 3 and 9), borates (pH 8–9), mixed phosphate-
borate systems, acetates, and so forth [16]. Addition of surfactants, such as sodium dodecyl sulfate
(SDS) or cetyltrimethylammonium bromide (CTAB), has often improved the peak efficiency due
to micellar electrokinetic chromatography (MEKC) effect. Other additives have been exploited
as well, for example, thriethylamine to improve run-to-run reproducibility, or polyethyleneoxide
(PEO) to enhance resolution, most likely, by imposing additional sugar–polymer interactions based
on hydrogen bonding. Owing to weak acidic properties, neutral sugars become ionized at very high-
pH values (pH 12–13) and can be separated according to their pKa differences under strongly alkaline
conditions (analogous to HPAEC). Borate-assisted CE has been widely used for the analysis of both
native and derivatized sugars. Borate ions are known to form anionic complexes with polyols (e.g.,
sugars) and more stable adducts are typically formed with cis-oriented hydroxyl groups. Therefore,
in the presence of borate anions the resolving power of CE can be significantly enhanced, and very
structurally similar carbohydrate molecules can be separated.
In most cases, carbohydrate molecules must be derivatized to render charge, prior to or dur-
ing electrophoretic separation, as majority of carbohydrates are neutral. This can be achieved by
interaction with oxoacid [22] or metal ions [23], or introduction of ionic tags [24,25]. On the other
hand, electrokinetic chromatography approach using hydrophilic monolithic columns, developed by
Novotny and coworkers [26], extended CE applicability to neutral sugar molecules. Certain surfac-
tants, such as SDS, for example, form negatively charged micelles in the running buffer and serve
as pseduostationary phase that incorporates carbohydrate molecules with a hydrophobic tag [e.g.,
1-phenyl-3-methyl-5-pyrazolone (PMP)] [27] to a various extent, resulting in differential migration
and separation.
Absence of chromophore or fluorophore groups and very low extinction coefficients of most
carbohydrate molecules called for application of alternative (to UV and fluorescence) detection
methodologies. Electrochemical detection and refractometry allow direct measurements although
the sensitivity is relatively low. Carbohydrates are very weak acids, and their hydroxyl groups get
ionized at very high pH values (>12–13), enabling CE analysis under such strong alkaline conditions
using gold or transition metal electrodes [28,29] in electrochemical detection settings.
With respect to UV or fluorescent detection, a number of methods have been developed to
introduce UV absorbing and/or fluorescent tags into carbohydrate molecules. Derivatization of car-
bohydrates, to render chromophore/fluorophore group and/or charge, was first demonstrated in CE
by Honda [22] who used 2-aminoacridone (first introduced by Hase for HPLC analysis [30]). Since
then, different derivatization strategies have been explored. Most typical of these are reductive
amination [e.g., 2-aminoacridone, 2-aminopyridine, 4-aminobenzoate, p-aminobenzonitrile, APTS,
8-amino-naphtalene-1,3,6-trisulfonic acid (ANTS), etc.]; reaction of amino sugars with 1,2-dioxo
aromatic compounds to form isoindoles [e.g., o-phthaldehyde, 3-(4-carboxybenzoil)-2-quinoline-
carboxaldehyde]; reaction of aldoses with pyrazalones (e.g., PMP); reaction of dioxo sugars with
1,2-diamonoaromatic compounds; amidation of sugar carboxylic acids with aromatic amines through
carbodiimide effect [16]. PMP tag allows quantitative labeling under mild conditions, in both pre-
column mode and in migratio [31]. The most effective fluorescent derivatization agents have been
APTS and methylglycamine-4-nitro-2,1,3-benzoxadiazole (MG-NBD) [17,32]. These methods are
based on reduction amination, which results in quantitative and uniform labeling of different car-
bohydrates, and utilize most common Ar-ion laser for excitation (488 nm) [33–36]. The details of
these and other derivatization approaches are well summarized in several articles [18–19,32].
High-performance and resolving power of CE enable distinguishing between very similar struc-
tural isomers of short oligosaccharides, which have different anomeric and positional configurations
of sugar monomers. Typical separations of various APTS-labeled homooligosaccharide series are
exemplified in Figure 7.1 [37]. Traces A and B correspond to oligomers, composed of glucose
1 2 3
RFU C
1 3 5 6
2 4
2 3 4 6 7
5 8
1
A
4 6 8 10
Time (min)
FIGURE 7.1 High-resolution CE separation of oligosaccharide structural isomers. Traces from the bottom: (A)
maltooligosaccharide oligomers (DP 1–8); (B) cellooligosaccharide oligomers (DP 1–6); (C) xylooligosaccha-
rides (DP 1–3); and (D) xylopentaose. Numbers above the peaks correspond to the DP. Conditions: fused-silica
capillary, 60 cm × 50 µm ID, effective separation length 50 cm; running buffer 25 mM lithium acetate (pH
5); dynamic coating, 1% PEO; injection 5 s at 0.5 psi (3447 Pa); applied electric field 500 V/cm; separation
temperature 20◦ C. (Adapted from Khandurina, J. and Guttman, A. Chromatographia, 62, S37, 2005. With
permission.)
molecules linked via α-1,4 (maltooligosaccharides) and β-1,4(cellooligosaccharides) bonds, respec-

tively. Numbers above the peaks denote the degree of polymerization (DP). As one can see, species
with the same DP but different anomeric configuration between the monomer units (α or β) have
different electrophoretic mobilities, and can be readily separated in CE, using 25 mM lithium acetate
(pH 5) buffer in combination with a dynamically coated column and reversed polarity. The shorter
cello-oligomers migrated faster than the corresponding malto-analogs (DP 2–5 peaks in traces A and
B, Figure 7.1). However, the migration order was reversed above DP 5, as the cellohexaose migrated
slower than maltohaxaose. Most likely, the differences in the spatial configuration, rigidity and
hydrodynamic radius of the short α- and β-glucopyranose oligosaccharides in the solution caused
this change in the migration order. Traces C and D in Figure 7.1 depict the separation of xylose series
(xylose, xylobiose, xylotriose, and xylopentaose), which possess higher electrophoretic mobility
compared to glucose series (traces A and B, Figure 7.1) with the same DP, due to more compact
structure of pentose-based sugars versus hexose ones. Very close structural isomers of di- and trisac-
charides can be resolved by CE, under suppressed EOF conditions, as demonstrated in Figure 7.2.
Here, maltose and isomaltose (α-1,4 and α-1,6 linkages) are baseline resolved (peaks 1 and 2). On
the other hand, trisaccharide analogs (maltotriose and panose, peaks 3 and 4) could be separated only
with the addition of the polymer additive, PEO. The PEO additive is capable of selectively inter-
acting with sugar hydroxyls via the formation of hydrogen bonds, thus, enhancing (or decreasing)
separation selectivity. As it is illustrated in Figure 7.2, the described CE approach can provide an
excellent tool for fine tuning of the separation selectivity of the carbohydrate molecules of interest.
CE profile of a complex mixture of 10 mono- and oligosaccharides, constituents of plant cell walls,
is presented in Figure 7.3. An excellent resolution of the three different galactobiose isoforms (β-1,4;
α-1,4; and α-1,3 linkages) was achieved in capillary zone electrophoresis (CZE) format [37].
12
1
C 2
3 4
8
1
RFU
B 2
3 4
4
1
A 2 3+4
0
4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0
Time (min)
FIGURE 7.2 CE analysis of close structural di- and trisaccharide isomers. Peaks: (1) maltose, (2) isomaltose,
(3) panose, and (4) maltotriose. Conditions: 0.5% PEO (B) and 1% PEO (C) was added to the running buffer (no
PEO additive in (A). Separation temperature 15˚C. Other conditions were the same as in Figure 7.1. (Adapted
from Khandurina, J. and Guttman, A. Chromatographia, 62, S37, 2005. With permission.)
RFU
3
2
9
5
40 4 6
1
8
10
20 7
0
3.0 4.0 5.0 6.0 7.0 8.0
Time (min)
FIGURE 7.3 High-resolution analysis of a combined mono- and oligosaccharide mixture representing frag-
ments of major cell wall components of bioindustrial interest. (1) Galacturonic acid, (2) digalacturonic acid, (3)
trigalacturonic acid, (4) rhamnose, (5) galactose, (6) galactobiose β-1,4 linkage, (7) galactobiose α-1,4 linkage,
(8) galactobiose α-1,3 linkage, (9) galactotriose α-1,3; β-1,4, and (10) galactotetraose α-1,3; β-1,4; α-1,3.
Conditions were the same as in Figure 7.1. (Adapted from Khandurina, J. and Guttman, A. Chromatographia,
62, S37, 2005. With permission.)
7.3 THEORETICAL AND PRACTICAL ASPECTS OF MODERN

CARBOHYDRATE ANALYSIS: CE-MS
7.3.1 BASIC UNDERLYING PRINCIPLES
The popularity of CE-MS hyphenation has been a fast growing analytical approach, since its intro-
duction in the late 1980s. Over the years the technique has advanced to the commercial ready-to-use
level. Still, there is a lot of development to be done toward optimization of CE-MS instrumenta-
tions, in particular, improving its robustness and efficiency. A variety of analytical applications in
different fields have been investigated [38]. With respect to carbohydrate analysis, MS detection
offers a way to overcome detection sensitivity problems associated with the lack of chromophore
groups, as well as structural elucidation of complex carbohydrate molecules. Structural charac-
terization of carbohydrates includes the determination of molecular weight, type, and number of
saccharide components, sequence and branching, glycosidic attachment sites, anomeric configura-
tions, conformation of glycosyl ring, and secondary structure. Electrospray ionization (ESI) and
matrix-assisted laser desorption ionization (MALDI) mass spectrometry, especially in the tandem
MSn mode, are the most popular approaches that offer high ionization efficiency of saccharides
and possibility of their sequencing. In particular, nano-ESI-MS/MS, in both positive and negative
mode, enable elaborate structural determination of carbohydrates from biological samples using
only sub-pmole amounts [39–46]. Basically, three major types of ESI interfaces have been used for
CE-MS coupling: coaxial sheath-flow, sheathless, and liquid junction, as schematically presented
in Figure 7.4 [47]. Different modern mass analyzers e.g., ion trap (IT), quadrupole time-of-flight
(QTOF), TOF–TOF, and even Fourier transform ion cyclotron resonance (FT-ICR) can provide
highly informative fragmentation spectra of oligosaccharides [26,48–53]. One of the challenges in
carbohydrate characterization in complex mixtures is associated with the presence of overlapping
isobaric peaks that complicates spectra interpretation. In addition, modified and substituted sugar
molecules are prone to in-source fragmentation due to labile nature of acetyl, phospho-, sulfo-, and
other functional groups, causing loss of structural information of intact molecules. Moreover, most
carbohydrates, unlike peptides and proteins, do not form multiply charged ions causing difficulties
analyzing oligo- and polysaccharides of high molecular weight, which often exceeds MS detection
capabilities [4].
MS analysis of complex biological mixtures of carbohydrates often follows a liquid phase sep-
aration method, such as chromatography (HPLC), CE, or capillary electrochromatography (CEC).
The advantages of mass spectrometry, combined with superior resolving power of CE, have boosted
glycoscreening development in biomedical research [4]. Very efficient and fast separations, automa-
tion and capabilities of miniaturization, already proven for small molecules, peptides, proteins, and
nucleic acids [54–57], represent a promising analytical alternative for all types of carbohydrates.
CE-MS combination can be carried out either off-line or online. The former approach was often real-
ized by CE fraction collection followed by ESI-MS characterization of individual fractions, which
contain a single carbohydrate species or mixtures of significantly reduced complexity. However,
fraction dilution associated with this approach (collection of nanoliter fractions into microliter total
volume) imposes additional detection sensitivity challenges, especially with limited amounts of gly-
coconjugate samples extracted from biological sources, and therefore, is of limited use. On the other
hand, online CE-MS interface has become truly widespread during the past decade. This method-
ology allows minimum sample handling and consequently maximum efficiency in MS detection.
A limitation though is the necessity to select ESI-MS compatible volatile buffers, which often do
not possess the highest resolving power characteristics for CE. Besides, high sensitivity and high
speed of data acquisition, especially for online CE-MS/MS, and certain restrictive conditions of ion
formation also exist [58]. In spite of these limitations, CE-MS technique is invaluable, providing
carbohydrate composition identification, quantification, and structural insights. To decipher molecu-
lar composition, MS/MS measurements are typically conducted using low-energy collision-induced
dissociation (CID) in ESI applications or MALDI postsource decay (PSD). These types of fragmen-
tation preferentially cause glycosidic bond cleavage. Sugar ring cut, on the other hand, resulting in
the formation of mostly C and Y type alcoholate ions [39], provide information on the branching
pattern. The latest developments in CE-MS analysis of carbohydrates are highlighted and summa-
rized in a number of comprehensive reviews (e.g., by Zamfir and Peter-Katalinic [4], Campa et al.
[59], and Kamoda and Kakehi [60]). Here, we briefly outline examples of CE-MS applications to
the analysis of carbohydrates, including complex glycoconjugate mixtures, neutral and negatively
(a)
ESI voltage
CE capillary Conductive coating Nebulizer gas

From CE to MS
(b)
Transfer capillary
Buffer reservoir Nebulizer gas
CE capillary
From CE to MS
Electrode
(c)
ESI voltage
Nebulizer gas
CE capillary Sheath liquid
From CE
Stainless-steel tubes
FIGURE 7.4 Interfaces for the direct coupling of CE to ESI-MS using (A) sheathless; (B) liquid-junction;
and (C) coaxial sheath-flow designs. (Adapted from Simo, C. et al. Electrophoresis, 26, 1306, 2005. With
permission.)
charged glycopeptides and glycoproteins, glycosaminoglycans (GAGs), lipopolysaccharides (LPS),

and gangliosides.
7.3.2 CE-MS CHARACTERIZATION OF SIMPLE SUGARS AND GLYCOCONJUGATES

CE-MS analysis of simple sugar mixtures is typically conducted using either derivatized neutral sug-
ars or underivatized carbohydrates, containing ionizable groups (e.g., carrageenan oligosaccharides)
[61]. Derivatization is typically accomplished through reductive amination coupling to negatively
charged tags, such as aminobenzoic acid (ABA) [62], 8-aminonaphtalene-1,3,6-trisulfonic acid
(ANTS) [59], APTS [63], and 7-amino-1,3-naphtalene-disulfonic acid (ANDSA) [64]. Positively
charged tags, for example, 3-(acetylamino)6-aminoacridone [65] or rhodamine 123 [59], are some-
times used as well. Labeling results in enhanced ionization efficiency and MS response, as well
as CE separation selectivity [48], for most carbohydrates regardless of their native charge. Simple
sugars are often used as standard mixtures to assess performance and optimize CE-MS conditions
prior to characterization of more complex samples.
Glycoproteins, a very important type of glycoconjugates, are formed during co- and posttrans-
lational modification in eukaryotic cells. Typically, N-glycans are linked to asparagines, while
O-glycan moieties to serine or threonine. Protein glycosylation is characterized by pronounced
microheterogeneity of glycosylation sites, which poses challenges in controlled manufacturing of
recombinant glycoprotein therapeutics [66]. CE-MS analysis has been successfully applied to charac-
terization of sugar moieties, native or derivatized, using direct or reversed polarity CE in combination
with ESI-MS or MALDI. Since typically only low concentrations of glycans are available from gly-
coconjugates, various derivatization approaches have been implemented to increase MS response
and facilitate interpretation of MS/MS spectra of glycans. Some research groups developed new
derivatizing agents with improved physicochemical properties and better suited for CE-MS applica-
tions [65,67]. Min et al. [68] reported a new interesting labeling approach based on enzymatic
reaction rather than chemical derivatization. In this work a fluorescent acceptor, naphthalene-
2,3-dicarboxaldehide-asparagine-N-acetylglucosamine, was used. Asn-linked glycan moieties were
enzymatically transferred from ovalbumin to the tag reagent. This strategy can help alleviate degrada-
tion problems sometimes observed in reductive amination, especially with unstable glycan residues
(e.g., sialylated carbohydrates). A number of research groups used APTS and ANTS labeling in
conjunction with CE-MS/MS analysis of N-glycans following their enzyme-assisted release from
glycoproteins [69–71]. CE-UV and CE-MS electropherograms, obtained for the same oligosaccha-
ride mixture, were compared under optimized analysis conditions in Reference 69. As expected,
separation efficiency and selectivity were lower in CE-MS, primarily due to sheath liquid dilution
and certain dead volume in the capillary outlet-MS interface. In addition, application of an MS-
friendly buffer (6-aminocapronic acid, pH 4.1), instead of 10 mM citric acid (pH 3) used in CE-UV
settings, could have contributed to the decrease of selectivity. Nevertheless, the use of powerful mass
spectrometers can counterbalance the loss of separation efficiency, as it was demonstrated by suc-
cessful structural characterization of a number of N-linked carbohydrates and other glycoconjugates
[4,72,73]. The MSn capacity of ion-trap spectrometer has been particularly useful in elucidation of
fine structural details.
Analysis of high-mannose-type oligosaccharides by a combination of micro-liquid chromatogra-
phy (µLC)-MS and CE was reported by Koller et al. [70]. The authors separated glycopeptides and
high-mannose-type oligosaccharides, derived and digested from a recombinant enzyme phospho-
lipase C expressed in Pichia pastoris yeast. The glycopeptides were subjected to µLC-ESI-MS
and µLC-ESI-MS/MS that revealed variation in high-mannose structures in the range between
Man7 GlcNAc2 and Man14 GlcNAc2 . Then, high-performance CE was applied to identify possi-
ble positional isomers of the high-mannose structures. One Man9 GlcNAc2 , two Man10 GlcNAc2 ,
three Man11 GlcNAc2 , Man12 GlcNAc2 , and Man13 GlcNAc2 , and two Man14 GlcNAc2 were iden-
tified, while no Man7 GlcNAc2 and Man8 GlcNAc2 were observed (Figure 7.5). It was found that
CE results provided complementary information to the µLC-ESI-MS/MS data, revealing the exact
number of positional isomers in the glycan pool.
Cell surface and extracellular proteins are typically O-glycosylated, and the most abundant type
of O-glycosylation is the attachment of N-acetylgalactosamine (GalNAc) to serine or threonine in the
protein sequence by a-glycosidic linkage [74]. Many eukaryotic cytoplasmic and nuclear proteins,
containing such carbohydrate moiety, exhibit reciprocal glycosylation and phosphorylation during
cell cycle, stimulation, and growth. In the group of O-GalNAc glycosylated proteins, musins are the
most ubiquitous. They are especially interesting with respect to cancer research, understanding of
cell adhesion, and metastasis processes. The distribution of O-glycans in human tissues and sera is
important for diagnostic and prognostic studies [74,75]. Mucin-like O-glycans are characterized by
high degree of structural variation, even at the core level, as well as in the chain elongation arising
from the core extension (Figure 7.6). The carbohydrate chains often contain repeating units of N-
acetyllactosamine, terminal sialic groups, and also noncarbohydrate substituents. Less investigated
types of O-glycosylation are O-fucosylation, O-mannosylation, and O-glucosylation. These types of
modifications are vital for physiological functions of proteins. O-glycans are traditionally detached
G3
G7
G5
G9
G11
G13
G15
B
RFU
M5 M6
M8
M7
M9
C
M9
M10
M13 D
M11
M12
M14
6 7 8 9 10 11 12 13 14 15
Time (min)
FIGURE 7.5 High-performance CE separation of the APTS-labeled high-mannose type oligosaccharides

released by PNGase F from bovine ribonuclease B (C) and PLC (D). (A) Electropherogram of the maltooligosac-
charide ladder; (B) derivatization background peaks. Conditions: 48 cm neutrally coated column (58 cm total
length), 50-µm ID. LIF detection: Ar-ion laser, 488 nm excitation, 520 nm emission. Separation buffer: 25
mM acetate (pH 5); 25◦ C capillary cartridge temperature; pressure injection 10 s at 0.5 (A) and 2 psi (B-D).
(Adapted from Koller, A. et al. Electrophoresis, 25, 2003, 2004. With permission.)
3 GalNAcaSer/Thr–R 3 GalNAcaSer/Thr–R
Galb1 Core 1 GalNAca1 Core 5
GlcNAcb1
6 GalNAcb1
GalNAcaSer/Thr–R
3
Core 2 6 GalNAcaSer/Thr–R
Galb1
Core 6
3 GalNAcaSer/Thr–R GalNAca1
GlcNAcb1 Core 3 6 GalNAcaSer/Thr–R
Core 7
GlcNAcb1
6 GalNAcaSer/Thr–R 6 GalNAcaSer/Thr–R
3
Gala1
GlcNAcb1 Core 4 Core 8
FIGURE 7.6 Eight types of O-glycan core structures. (Adapted from Peter-Katalinic, J. Methods Enzymol,
from the parent protein by β-elimination using strong bases (e.g., NaOH or hydrazine). Hydrozi-
nolysis does not always result in the selective cleavage of O-glycans, and often is accompanied by
N-glycan release, as well as modification of the reducing terminus [74] and references therein. β-
Elimination by NaOH under reductive conditions allows chemical derivatization of the reducing end
(e.g., with a fluorescent label). Nonreductive release of O-glycans by β-elimination is also possible
[74], as well as an enzymatic release, but the latter options are limited by high substrate specificity
of O-glycosidases available. Comprehensive profiling of complex glycan and glycopeptide mix-
tures requires sophisticated analysis methods to characterize glycosylation patterns and structures
of individual components, and the coupling of CE separation devices to ESI-MS or MALDI-MS is
a powerful approach. Off-line CE-MALDI analysis has been used for the characterization of musin
carbohydrate chains, comprising up to six GalNAc moieties [75]. CE-MS/MS interface enabled
deducting molecular structure of individual glycopeptide components based on specific fragmenta-
tion, especially when micro- and nano-ESI sources are used. O-glycopeptides in urine of patients,
suffering from a hereditary N-acetylhexosaminidase deficiency, were profiled by this approach, using
high-pH and MS-compatible buffer for CE separation, sheathless CE-ESI-MS interface, and QTOF-
MS equipped with automated high-speed MS-MS/MS switching [75]. The MS/MS fragmentation of
glycans results mainly in the formation of Y and C types of ions, which are indicative of the specific
structures of the parent ions. The sheathless CE-ESI/QTOF design proved to be one of the most
suited designs for glycomics, providing high coverage of detection and identification of CE-MS
data, MS-MS/MS switching, electrospray stability, and acquisition rate [74,75].
The glycan MS fragmentation nomenclature introduced by Costello, Domon, and Vath [76–
79] is a widely accepted system used by most of the MS and glycoconjugate community. The
simplest fragmentation of the carbohydrate moiety of glycoconjugates and glycosides results from
the cleavage of the glycosidic bond and therefore provides information on the sugar sequence. More
complex processes involving the fragmentation of the sugar ring can occur, particularly, in CID-
MS/MS spectra. These ions are more difficult to assign, although they contain important structural
information. Figure 7.7 schematically illustrates the nomenclature. Fragments are designated Ai , Bi ,
and Ci , where i represents the number of the glycosidic bond cleaved, counted from nonreducing
end. Ions containing aglycone, or the reducing sugar unit in case of oligosaccharides, are labeled as
Xj , Yj , and Zj , where j is the number of interglycosidic bond counted from aglycone, or reducing end.
The glycosidic bond linking to aglycone is 0. More complex fragmentations involve cleavage of C–C
bonds of the sugar ring. The product ions are designated as Ai and Xj . Since several ring cleavage
pathways are possible, two additional superscripts are used (e.g., k,1Aj ) to indicate the sugar ring
bonds that have been broken [78]. This system, described in more detail elsewhere [78,80], comprises
types of cleavages, observed in collision-induced decomposition MS/MS spectra of carbohydrate
moiety of glycoconjugates, and is applicable to both positive and negative ionization modes.
Glycolipids are carbohydrate-attached lipids (e.g., phospholipids) found on the outer exoplasmic
surface of all eukaryotic cell membranes. Their biological role is to provide energy and serve as
markers for cellular recognition. One subtype is LPSs, which are major components of the cell
membrane of Gram-negative bacteria, contributing greatly to the structural integrity, and protecting
the membrane from certain kinds of chemical attacks. LPS are endotoxins and induce a strong
response from normal animal immune systems. It comprises three parts: polysaccharide (O) side
chains, core polysaccharides, and lipid A. A lipid A contains certain fatty acids (e.g., hydroxy-
myristic acid), and is inserted into the outer membrane while the rest of an LPS projects from the
surface. A core polysaccharide contains unusual sugars, such as keto-deoxyoctulonate and heptulose.
It also contains two glucosamine sugar derivatives, each having three fatty acids with phosphate
or pyrophosphate attached. The core polysaccharide is attached to lipid A, which is also in part
responsible for the toxicity of Gram-negative bacteria. The polysaccharide side chain is referred as
the O-antigen of the bacteria. O side chain (O-antigen) is also a polysaccharide chain that extends from
the core polysaccharide. The composition of the O side chain varies between different Gram-negative
bacterial strains. O side chains are easily recognized by the host antibodies, however, the nature of
(a) Y2 Z2 1.5x 1 Y1 Z1 Y0 Z0
CH2OH CH2OH CH2OH

O O O
HO O R
O O
OH OH OH
OH OH OH
0.2A1 B1 C1 2.4A2 B2 C2 2.5A3 B3 C3
(b)
CH2OH
5 CH2OH 5
4 0 0 4 O 0 O R Ac2HN 4 O
COOH
O R
CHOH 0
3
OH 3 OH 3
1 1 1 0 R
HO
2 2 CHOH 2
OH OH OH
CH2OH
1 2 3
FIGURE 7.7 Types of carbohydrate fragmentation. (Adapted from Domon, B. and Costello, C. E.
Glycoconjugate J, 5, 397, 1988. With permission.)
the chain can easily be modified by the bacteria to avoid detection. LPS also increases the negative
charge of the cell wall and helps stabilize the overall membrane structure [81,82].
A comprehensive discussion on the application of CE-MS to the analysis of complex LPS can be
found in Reference 83 and references therein. Various research groups developed electrophoretic con-
ditions to conduct trace-level enrichment and separation of closely related glycoforms and isoforms.
Sensitive detection of glycolipids from as little as five bacterial colonies has been demonstrated.
Mixed MS scanning functions can assist in the identification of specific LPS functionalities [e.g.,
pyrophosphoethanolamine, phosphocholine, and N-acetylneuraminic acid (Neu5Ac)]. High resolv-
ing power of CE combined with sensitivity of tandem MS provides a unique analytical tool to probe
subtle structural differences and location of oligosaccharide isoforms of LPS. Correlation of structural
changes in bacterial strains and isogenic mutants allow establishing functional gene relationships,
using CE-MS screening capabilities and wide dynamic range.
Figure 7.8 shows structure of the oligosaccharide region of LPS from Neisseria meningitidis
and Haemophilus influenzae. LPS of H. influenzae contains a common l-glycero-d-manno-heptose
inner-core trisaccharide unit attached to the lipid A via a phosphorylated 2-keto-3-deoxyoctulosonic
acid (KDO) residue (Figure 7.8A). Similarly, N. meningitidis (Figure 7.8B) has a conserved inner
core structure with two KDOs, two heptoses (Hep), and one N-acetylglucosamine (GlcNAc). Further
carbohydrate extension can occur at each heptose. In addition, various noncarbohydrate substituents
can be incorporated, for example, acetyl (Ac), phosphate (P), phosphoethanolamine (PE), pyrophos-
phoethanolamine (PPE), phosphocholine (PC), as well as different aminoacids (Ala, Ser, Thr, Lys).
Details on typical LPS extraction and preparation methods are well described in Reference 83.
Hydrozinolysis can be utilized to release O-linked fatty acids, and, if necessary, enzyme sialidase to
remove Neu5Ac residues. Li et al. [83] described interfacing of a CE instrument with a triple quad
(a)
Outer core Inner core Lipid A
P PE P P
b1,4 41 41 21
Glc HepI KDO GlcN GlcN
Outer core extension a1,3
Gal-Glc Lactosyl HepII HE
Gal-Gal-Glc Pk antigen a1,2
GalNAc-Gal-Gal-Glc Globotetraose HepIII 3-OH C14:0 or C14:0
(b)
Outer core Inner core Lipid A
KDO P P
b1,4 5 61 41 21
Glc HepI KDO GlcN GlcN
Outer core extension a1,3
6
Gal-Glc Lactosyl HepII PE
Gal-Gal-Glc Pk antigen a1,2
Gal-GlcNAc-Gal-Glc Lacto-N-neotetraose 3 3-OH C14:0 or C14:0
OAc - GlcNAc
FIGURE 7.8 Structure of the conserved inner core oligosaccharide regions of LPS from N. meningitidis and
H. influenzae. (A) Haemophilus influenzae capsular and noncapsular strains. (B) Neisseria meningitidis strain.
(Adapted from Li, J. et al. Methods Enzymol, 405, 369, 2005. With permission.)
or QTOF mass spectrometers via a microionspray assembly, using isopropanol–methanol sheath

liquid, a bare fused-silica capillary, and morpholine–formic acid (or ammonium acetate) buffer (pH
9) with 5% methanol. To maintain high separation efficiency of CE (N > 150,000 plates/m), the
injection volume should be kept below 2% of the total column volume (e.g., 40 nL for 1 m × 50 µm
ID capillary) [84]. MS identification of bacterial glycolipids can be achieved using 100–300 pg on
column loadings, which typically translates into low µg/mL detection limit when using ESI. Differ-
ent techniques to improve LPS sample loading have been explored, including preconcentration via
transient isotachophoresis, hydrophobic adsorption membrane at the sample inlet, and field amplifi-
cation stacking using a microdialysis interface junction and polarity switching ([83] and references
therein).
CE-MS has been employed for the structural elucidation of an O-chain polysaccharide (LPS
component exposed on bacteria cell surface and mediating host–cell invasion and other cell–cell
interaction processes) from Aeromonas salmonicida strains [85,86]. The authors used in-source
fragmentation approach for the analysis of these complex macromolecules. As mentioned above,
saccharides typically form singly charged ions during the ESI process. Therefore, the masses of
these oligosaccharides often exceed the mass limit of common mass spectrometers (e.g., quadrupole
or ion-trap types). The in-source fragmentation technique produces fragments that can be readily
detected by these instruments.
LPS are often presented by a complex distribution of closely related glycoforms varying by
length and site of attachment. Structural characterization of these diverse populations is important
for the development of protein therapeutics—antibodies specifically targeted toward these immun-
odeterminate structures. Glycoforms in the different families of glycolipids are identified based on
the progressive extension of Hex residues on the core structure, and concurrent increase in molecular
weight and decrease in electrophoretic mobility. This approach is described in great detail by Li et al.
TIE (m/z 400–1500)

(a)
13.0 m/z 290 (OR 150 V)
13.6
2 4 6 8 10 12 14 16 18
3017.0 Hex3-Neu5Ac-PE-PPE-PC
(b)
3015 3020 3025
2 4 6 8 10 12 14 16 18
2725.9
Hex3-PE-PPE-PC
(c)
2720 2730
2 4 6 8 10 12 14 16 18
Time (min)
FIGURE 7.9 Analysis of sialylated O-deacylated LPS from H. influenzae 375. (A) Results for H. influenzae
375 wild type strain. (B) Strain grown in presence of CMP Neu5Ac. (C) Strain following incubation with
α-2,3 sialidase. The total ion electropherogram (solid line) is shown together with the fragment anion m/z 290
characteristic of Neu5Ac. The arrows indicate the migration of the glycolipid shown as an inset on the right in
B and C. (Adapted from Li, J. et al. Methods Enzymol, 405, 369, 2005. With permission.)
[83]. In addition to isoform analysis, monitoring of trace-level glycolipids present as a small subset
in a bacterial extract, is often required. To probe characteristic functional groups and residues, such
as P, PE, PPE, PC, and Neu5Ac, CE-ESI-MS has been successfully applied [83] using in-source
fragmentation and selected ion monitoring (SIM) mode. The ability to probe specific functionalities
can also be an important analytical tool in monitoring incorporation of monosaccharides by glycosyl-
transeferases or their removal by glycosidases [83], as illustrated in Figure 7.9 for Neu5Ac residue
in H. influezae 375 strain. The precise location of a modification or the assignment of branching
site on the glycolipid structure typically requires using tandem MS/MS. Higher order MS (MSn
with FT-ICR, for example) helps rationalize the observation of characteristic fragment ions or assess
fragmentation pathways.
Glycosphingolipids (GSLs) are a subtype of glycolipids, containing aminoalcohol sphingosine,
which include cerebrosides, gangliosides, and globosides. GSLs are ubiquitously distributed among
all eukaryotic species and some bacteria. Since GSLs are secondary metabolites, their direct and
comprehensive analyses must be considered an essential complement to genomic and proteomic
approaches and establishing the structural repertoire and functional roles of these compounds within
an organism. Gangliosides, for example, are composed of a GSL (ceramide and oligosaccharide)
with one or more sialic acids linked on the sugar chain. It is a component in the cell plasma mem-
brane, which modulates cell signal transduction events. They have recently been found to be of
great importance in immunology. Natural and semisynthetic gangliosides are considered possible
therapeutics for neurodegenerative disorders. A detailed review on GSLs and their structural anal-
ysis, mostly by various MS techniques, can be found in Reference 80. Several research groups
[87–89] applied CE-MS technique to ganglioside analysis. Online glycoscreening of gangliosides
by CE-nano-ESI-QTOF-MS [87] is feasible, provided suitable buffer systems, compatible with MS
and ensuring acceptable CE resolution, are developed. Ju et al. [90] demonstrated good sensitivity
and electrophoretic resolution of a number of gangliosides. MEKC has been coupled with ESI-MS
by Tseng et al. [91] who used nonvolatile buffer systems.
Bindila et al. [72] reported a method for the analysis of underivatized glycoconjugates by CE-
TOF-MS. The method was first optimized using simple monosaccharides followed by the application
to the studies of urine samples from patients with N-acetylglucosaminidase deficiency syndrome.
Identification of nearly 50 major and minor species, O-glycosylated sialylated amino acids and
peptides in patient’s urine, using a CE buffer of 6.4% formic acid in 60% methanol at pH 2.8
(adjusted with NH3 ), was demonstrated. Another research group applied CE-MS approach to the
analysis of APTS-labeled structural isomers of glucooligosaccharides [63]. The known properties
of borate ions to form complexes with sugars, changing their hydrodynamic characteristics and
therefore improving the resolving power of CE, were also exploited in the CE-MS format. Isomeric
LPS and their isomeric glycoforms from N. meningitidis were distinguished and characterized by Li
et al. [91], employing CE-MS interface in either positive or negative mode.
More examples on CE-MS analysis of various glycoconjugates, including PGs, LPSs, ganglio-
sides, highly glycosylated peptides, and so forth, can be found in References 4, 59,60, and 83.
7.4 APPLICATIONS
7.4.1 COMPLEX POLYSACCHARIDES
Complex polysaccharides, glucosaminoglycans (GAGs), are a class of ubiquitous macromolecules
exhibiting a wide range of biological functions. They exist as side chains of PGs in the cells and extra-
cellular matrix. The recent development of analytical tools for their study has spurred a virtual explo-
sion in the field of glycomics. A number of electrophoretic separation techniques, including agarose
gel, CE, and fluorophore-assisted carbohydrate electrophoresis (FACE), have been employed for the
structural analysis and quantification of hyaluronic acid (HA), chondroitin sulfate (CS), dermatan
sulfate (DS), keratan sulfate (KS), heparan sulfate (HS), heparin (Hep), and acidic bacterial polysac-
charides [93]. Moreover, certain diagnostic analytical applications have been developed to detect fine
structural and compositional changes of GAGs at various pathological conditions. These have been
accomplished by analyzing oligosaccharides derived from GAGs by enzymatic or chemical degrada-
tion. The major analytical advances in this field in the past decade have been reviewed in Reference 93.
GAGs are linear, complex, and polydisperse polysaccharides [94–99]. Most of them, except
for KS, consist of alternating copolymers of uronic acids and amino sugars, and their structures
are commonly represented by disaccharide sequences. KS, CS, DS, HS, and Hep are sulfated het-
eropolysaccharides with different degrees of density and positional distribution of sulfate groups.
These compounds are very heterogeneous not only by structure but also in biological and pharma-
cological activities. With the exception of HA, GAG chains are covalently attached at their reducing
end through an O-glycosidic linkage to a serine residue or N-linked to asparagine (e.g., KS) in a
core protein [100–103]. Some glycobiology studies suggest that PGs are not only structural com-
ponents but also regulate many cellular and physiological processes (e.g., cell proliferation and
differentiation) [101,104–106]. Most GAGs are obtained from animal sources by extraction and
purification processes [99]. Recently, these natural substances have also been chemically modified
to get synthetic analogues and GAGs-based drugs [107,108].
Figure 7.10 shows structures of some typical GAGs. For example, HA is a linear polysaccharide
composed of alternating residues of the monosaccharides d-glucuronic acid and N-acetyl-d-
glucosamine linked by β-1,3 bonds in repeating units; and these disaccharides are connected
via β-1,4 linkages [109,110]. On the other hand, the repeatable blocks in KS molecules are N-
acetyl-lactosamine fragments of -β-1,3-[d-galactose-β-1,4-N-acetyl-d-glucosamine]-β-1,3-, which
are typically sulfated at C6 position of acetylglucosamine and sometimes at C6 of galactose [111].
Structures of some other typical GAGs are schematically shown in Figure 7.10 and described in
more detail in Reference 93 and references therein.
CH2OH
COO– O
H O
O H
H H
H O H
O OH H OH
H H NHCOCH3
H OH
Hyaluronic acid
CH2OR CH2OR
COO– –O O
–O SO
O
3SO O H
3
O
O H O H
H H H H
H O H COO– O H
OH H H O OH H H
O
H H NHCOCH3 H H NHCOCH3
H OR H OR
Chondroitin sulfate Dermatan sulfate
CH2OSO3–
CH2OSO3–
COO– O
H H H
H O H O
O H H
H OH H H
H O O COO– H
OH H OR H O
OH H O O
O H H NHCOCH3/SO3–
H NHSO3–
H OR H OSO3–
Heparan sulfate Heparin
CH2/OSO3–
CH2OR
H O
O
O H
HO OH H
H O H
O H
H H NHCOCH3
H
H OH
Keratan sulfate R = H or SO3–
FIGURE 7.10 Structures of disaccharides forming GAGs. Major modifications for each structure are illus-
trated (R = H or SO−3 ) but minor variations are possible. (Adapted from Volpi, N. and Maccari, F. J Chromatogr
B Analyt Technol Biomed Life Sci, 834, 1, 2006. With permission.)
Separation and quantification of GAGs compounds in mixed samples have been conducted by
one- and two-dimensional electrophoresis on cellulose acetate strips, nitrocellulose membranes,
agarose gels, and polyacrylamide gels (PAGE), using staining dyes for detection (alcian blue, tolui-
dine blue, azure A, methylene blue, combined azure blue, and silver staining) [93]. CZE in uncoated
fused-silica capillaries, as well as MEKC, have been successfully exploited for GAGs analysis. In
the latter case, SDS or CTAB was used in the basic borate buffers [112–114]. In addition, microemul-
sion electrokinetic capillary chromatography mode was demonstrated, where GAGs species can be
separated based on partitioning into oil droplets moving in the running buffer [93]. Different modes
of CE were realized with either normal or reversed polarity, under basic or acidic (suppressed EOF)
conditions, respectively. Proven benefits of CE, such as ease of use, fast analysis, very small sample
amount requirements, high resolution and sensitivity, and reproducible quantification, worked very
K4d
0.0100 K4 0.0100
Absorbance
Absorbance
0.0050 0.0050
0.0000 0.0000
0.00 10.00 20.00 25.00
FIGURE 7.11 HPCE electropherograms of unsaturated disaccharide HexAFrc-GalNAc, forming polysac-

charide K4 (K4), and HexA-GalNAc of defructosylated K4 product (K4d). (Adapted from Volpi, N.
Electrophoresis, 24, 1063, 2003. With permission.)
well especially for nonsulfated GAGs, that is, HA and bacterial polysaccharides, oligosaccha-
rides, and disaccharides generated by enzymatic digestion reactions [112–114]. The possibility to
label GAGs polysaccharides or their depolymerized short fragments with chromophores or fluo-
rophores, to significantly boost sensitivity, is another attractive feature of CE analysis. FACE, based
on 2-aminoacridone (2-AMAC) derivatization, has been applied to the analysis of plant cell wall
polysaccharides [115], and HA and GAGs disaccharides [116]. As an example, Figure 7.11 presents
CE separation of two disaccharides, products of chondroitinase digestion of polysaccharides from
uropathogenic Escherichia coli K4 bacteria and its defructosylated product [117]. SDS MEKC mode
in uncoated fused-silica capillary under normal polarity and UV detection (230 nm) was used in this
work. However, this approach required enzymatic treatment of bacterial polysaccharides with lyases.
In the other study of the same group [118], direct CE analysis of native K4 and defructosylated K4
has been developed. The two polyanions were separated and detected within 30 min, as depicted in
Figure 7.12. Linear region of quantification was observed in the 30–210 ng range. Other bacteria
polysaccharides and membrane LPS have been analyzed by high-performance CE [119,120] and
the results were invaluable for characterization and quantification of these complex polymers in
biological and medical research.
A method for the determination of HA oligomers by CE-MS was reported by Kuhn et al.
[121]. Oligosaccharides of 4–16 DP were generated by enzymatic digestion of HA with bacterial
hyaluronidase, followed by CE-MS analysis. Structural information was obtained in MSn experi-
ments using an ion-trap instrument (IT-MS). Another research group [122] described a homemade
sheathless ESI interface for the highly sensitive analysis of GAGs by CE-MS: CS and dermatan
oligosaccharides of extended chain length and increased degree of sulfonation from decorin trans-
fected human embryonic kidney cells were detected. More examples of CE-MS characterization of
complex polysaccharides are presented in Sections 7.3.2 and 7.4.2, as well reviewed in References
4, 59, and 93.
10
K4
K4d
Absorbance at 200 nm (x 10–3) 5
0 10 20 30
Time (min)
FIGURE 7.12 HPCE electropherograms of the polysaccharide K4 (K4) and defructosylated K4 polysaccha-
ride (K4d). (Adapted from Volpi, N. Electrophoresis, 25, 692, 2004. With permission.)
7.4.2 CARBOHYDRATES IN PHARMACEUTICALS

In drug development and clinical and forensic applications, the popularity of CE and CE-MS tech-
nologies have grown due to high discrimination power of CE and possibility to directly analyze
complex biological matrices, such as serum, urine, or other fluids [123,124]. This section overviews
most typical examples related to carbohydrate analysis in biopharmaceutical research. More details
and references on the topic can be found in recent comprehensive reviews [38,59,60,123,125].
Overall, implementation of CE in pharmaceutical bioanalysis has tremendous potential by providing
additional detailed information on the structure of oligosaccharide fragments in glycoconjugates
of pharmaceutical interest and ultimately contributing to the development of more effective and
safe medicines.
7.4.2.1 Glycoproteins
Glycosylation is one of the most common co- and post-translational modifications of the proteins.
Carbohydrate moieties in glycoproteins have direct effect on their biological activity and cellular
functions, such as protein folding, recognition, signaling, immune response, differentiation, and so
forth [126–128]. Modern recombinant technologies enable production of protein pharmaceuticals
in living cells. However, inherent structural heterogeneity in glycosylation patterns affect their effi-
cacy, pharmacokinetic and pharmacodynamic properties [129,130], and therefore should be carefully
controlled. The structure and composition of carbohydrate chains depend on the expression condi-
tions in the host cells and purification steps involved in the production processes. Characterization
of complex heterogeneous carbohydrate chains is a challenging, yet important, task in production
optimization, regulatory submissions, and quality control. Glycosylation patterns and heterogene-
ity are not directly controlled by genes, therefore cannot be fully predicted by protein expression
studies. Stand-alone CE and CE-MSn have become methods of choice in analysis of glycoforms of
protein and antibody therapeutics in biopharmaceutical research and development. High sensitivity
and resolution of CE permit detection of fmol and amol amounts of oligosaccharides, when using an
appropriate fluorescent-labeling method and LIF detection. Remarkably, oligosaccharide analysis at
a single cell level has been reported [131].
Recombinant human erythropoietin (rhuEPO), one of the most successful recombinant protein
therapeutics on the market, has three N-linked and one O-linked glycosylation sites, and each
glycan can have variable number of sialic acid residues [132]. Isoelectrofocusing (IEF) in capillary
format has been successfully demonstrated to resolve rhuEPO glycoforms, replacing time- and labor-
consuming slab gels IEF [58,133–135]. Noticeably, this methodology is now included in European
Pharmacopoeia [136]. The separation was performed in a bare fused-silica capillary using a buffer
containing putresine and urea. To overcome irreproducibility issues [58,135,136], capillary surface
modification through covalent and/or dynamic coating, was utilized [137]. Monitoring changes in
glycoprotein heterogeneity by CE separation of glycoforms is now in use in many quality control
labs for release and stability testing.
Structural analysis of N-linked oligosaccharides is typically conducted by their enzymatic release
from the protein, derivatization with a suitable fluorescent tag, and CE separation. A variety of
oligosaccharide standards have been developed to facilitate identification of unknown species based
on migration time comparison [60,138]. For further structural characterization, glycans can be frac-
tionated and subjected to LC-ESI-MS or MALDI-TOF-MS analysis [139–141]. Alternatively, online
CE-MS analysis is feasible as described in Section 7.3.1.
Sequencing of oligosaccharides can be accomplished by their digestion with various exoglycosi-
dase enzymes, followed by CE analysis and peak shift monitoring [60,142]. Specific glycosidase
cleavage is useful to distinguish positional isomers and glycosidic linkages. For example, Galβ-
1,4GlcNAc and Galβ-1,3GlcNAc at nonreducing end can be discriminated using β-1,4 galactosidase
or β-1,3 galactosidase. In glycan profiling of therapeutic antibody rituximab, two positional isomers
of biantennary N-linked oligosaccharides (with one galactose at nonreducing terminal of different
arms) were separated in CE, as shown in Figure 7.13 (peaks 2 and 3) [143]. The oligosaccharides
can be identified using α-1,2 and α-1,3 mannosidase digestion following N-acetylhexosaminidase
treatment [144]. Such subtle structural differences could not be resolved by MS or MS/MS methods.
Capillary affinity electrophoresis (CAE) based on specific carbohydrate–protein interaction, is
a valuable method in carbohydrate characterization. A variety of carbohydrate-binding proteins
(e.g., lectins) specific to certain oligosaccharide structures are available and have been exploited by
different research groups to characterize glycoprotein pharmaceuticals [145,146].
In recombinant antibodies [147], oligosaccharide moieties are attached to the Fc region of anti-
bodies and often control their biological activity and stability. CE-LIF analysis of oligosaccharides
derived from recombinant monoclonal IgGs has been used by different researchers and exhibited
high degree of reproducibility and accuracy in the determination of composition of the released
glycan isoforms. A number of examples of application of CE-based technology to glycan profiling
of therapeutic antibodies can be found in References 59, 60, 144, and 142.
7.4.2.2 Analysis of Complex Polysaccharides in Drugs

Certain GAGs and other sulfated polysaccharides are known to possess anti-inflammatory properties.
Some of them are also used as anticlotting and anti-arthritis agents. Recent studies by Liang et al.
[148–150], using CZE, demonstrated that heparin, carrageenan, and dextran sulfate interact with
a hematopoietic growth factor, granulocyte colony-stimulating factor (G-CSF), and have potential
therapeutic effect on cancers through inhibition of the growth and induction of the differentiation of
the leukemia cells. Such phenomenon might be of great importance for the treatment of patients under
radio- and chemotherapy, since heparin, for instance, appears to protect G-CSF from degradation,
therefore, increase its circulation half-life. It was found that these interactions are dependent on
the polysaccharide chain length and sulfate content. For example, separation, identification, and
interaction of heparin with G-CSF were accomplished by CE, following enzymatic digestion of
heparin with haparinase to render smaller oligosaccharides [150]. The oligosaccharides were well
separated in CE using 50 mM phosphate buffer (pH 9). The smaller di- and trisaccharides were also
identified by CE-ESI-MS.
(A)
2
3
4
0 1 2 3 4 5 6 7
(B)
1
3
4
0 2 4 6 8 10 12
(C)
Fucα1\
GlcNAcβ1-2Manα1\ 6 6
1 Manβ1-4GlcNAcβ1-4GlcNAc
GlcNAcβ1-2Manα1/ 3
Fucα1
Galβ1-4GlcNAcβ1-2Manα1 6 6
GlcNAcβ1-2Manα1 3
Fucα1
GlcNAcβ1-2Manα1 6 6
Galβ1-4GlcNAcβ1-2Manα1 3
Fucα1
Galβ1-4GlcNAcβ1-2Manα1 6 6
Galβ1-4GlcNAcβ1-2Manα1 3
FIGURE 7.13 Oligosaccharide maps of trastuzumab by CE. (A) Oligosaccharides derivatized with APTS, (B)
oligosaccharides derivatized with 3-AA. Analytical conditions: (a) capillary DB-1 (50 µm ID, 20 cm effective
length, 30 cm total length); running buffer 50 mM Tris–acetate buffer (pH 7.0) containing 0.5% PEG70000;
injection 0.5 psi for 5 s; applied voltage, 18 kV at 25◦ C. (b) Capillary DB-1 (100 µm ID, 20 cm effective length,
30 cm total length); running buffer 100 mM Tris–borate buffer (pH 8.3) containing 10% PEG70000; injection
1.0 psi for 10 s; applied voltage 25 kV at 25◦ C. LIF detection 405 nm emission/325 nm ex (He-Cd laser). (C)
Structures of major oligosaccharides in rituximab, peaks 1–4. (Adapted from Kamoda, S. et al. J Chromatogr
A, 1050, 211, 2004. With permission.)
The presence of native ionizable groups in GAGs and some other polysaccharides make them
particularly suited for both CE and MS, without need for derivatization. A number of examples of
the CE separation and CE-MS characterization of GAGs oligosaccharides can be found in literature
[59,87–89,121,151,152]. Both positive and negative ionization modes, as well as normal and reversed
CE polarities, were tested for analysis of heparin oligomers [151], although negative ionization was
particularly useful for complex mixtures. In many cases, use of low pressure in conjunction with
CE voltage, especially in the reversed separation mode, helped boost electromigration, and electro-
spray stability [121,151]. MS/MS fragmentation enabled qualitative and quantitative identification
of coeluting species, for example, disaccharides IIS-IIIS and UAS(1→4)α-d-GlcNAc6s (IIA)-
UA2S(1→4)α-d-GlcNAc(IIIA) [152]. MS/MS fragmentation results in loss of sulfate groups in
di- and trisulfated disaccharides, and loss of water and ring cleavage in mono- and nonsulfated ones. It
was shown that limits of detection (LODs) were comparable (low-µM range) for CE-UV and CE-MS
[152]. A coated capillary was used to enhance and stabilize EOF and characterize oligosaccharides
of HA, up to 16mer, under positive CE polarity and negative ESI [152].
Degree of sulfation of GAGs, in particular oversulfation, affect their biological activity [87–
89]; therefore, structural studies are important in the analysis of these compounds. Zamfir et al.
[87–89] have extensively studied DS and CS using CE-ESI-MS, both off-line and online. Mild
MS/MS fragmentation conditions allowed preventing sulfate cleavage and more effectively detecting
multiply charge ions, while higher collision energy produced smaller fragments and was more useful
for structure elucidation. Introduction of sheathless CE-MS interface with nano-ESI probe resulted in
improved overall performance of the analysis and LOD in pmol range [89]. Structural characterization
of CS/DS derived from skin fibroblast secretion lit more light toward understanding of binding
activity to fibroblast growth factor-2 (FGF2) [88], which is related to epithelial cell proliferation and
growth of new blood vessels (i.e., angiogenesis). Overall, CE-MS/MS approach seems to represent
a powerful tool for structural characterization of sulfated GAGs.
7.4.2.3 Glucuronidation of Drugs

Glucuronidation and N-glucuronidation are major inactivating pathways for a vast variety of endoge-
nous and exogenous molecules, such as drugs. Therefore, analysis of glucuronidated drug metabolites
is of importance for drug development and clinical and forensic applications. Owing to ionic nature
of glucoronidated compounds, CE and MS are often the methods of choice for their identification
and characterization.
For example, detection of opoids in body fluids is needed in clinical and forensic toxicology
because of their widespread use for both therapeutic and illicit purposes. CE-MS analysis of urine
proved to be a very efficient aproach in the detection of opoids’ metabolites [153–155]. Urinary sam-
ples of codeine, dihydrocodein, morphine and their glucuronides, and oxycodone and its metabolites
were successfully analyzed by Wey et al. [153–155], utilizing CZE-based immunoassay, as well as
CE-UV and CE-ion trap MS. Theses researchers investigated effects of different pretreatments and
found that solid-phase extraction (SPE) and liquid–liquid extraction improved detection limits [154].
CE separation before MS detection versus direct MS has been beneficial, since it enabled resolution of
the compounds with identical fragmentation patterns (e.g., morphine-2-glucuronide and morphine-
6-glucuronide). Besides, direct MS analysis required significantly larger amount of samples.
CE-ESI-MSn has been used for the analysis of entacapone and tolcapone and their metabo-
lites [156]. These drugs are administered in combination with anti-Parkinson’s levodopa. The main
metabolites found were 3-O-glucuronide conjugates, and the analytical approach proved to be repro-
ducible and very sensitive. The latter should facilitate direct metabolite identification in complex
biological matrices.
CE and CE-ESI-MS were exploited for the analysis of lorazepam (3-hydroxy-1,4-
benzodiazepine) and its metabolites in urine [157]. This drug is commonly used for the treatment of
anxiety and as a sedative and hypnotic agent. The authors found that 75% of administered lorazepam
is excreted in urine as its 3-O-glucuronide. Interestingly, since glucuronidation occurs at the chi-
ral center of the molecule, two diastereoisomers can be formed, which was confirmed by MEKC
analysis of urine extracts and also in vitro, via incubation of the drug with human liver microsomes
and 5 -diphospho-glucuronic acid as coenzyme. The evidence of stereoselectivity of lorazepam
glucuronidation, with one diastereoisomer being formed preferentially than the other, was found
as well. Achiral analysis was performed with a running buffer composed of 6 mM Na2 B4 O7 ,
10 mM Na2 PO4 (pH 9.1), and 75 mM SDS. For enantioselective analysis, the same buffer was
diluted with isopropanol (2.5%), and 2-hydroxypropyl-cyclodextrin stereoselector was added. CE-
MS experiments were performed in negative ESI mode with ammonium acetate as a separation
buffer (pH 9 adjusted with concentrated ammonia) and a sheath liquid containing 50% methanol
with 0.1% concentrated ammonia to ensure the formation of negatively charged ions. CE-MS2 and
CE-MS3 experiments were conducted to confirm the presence of lorazepam and its 3-O-glucronide,
respectively.
7.4.2.4 Characterization of Oligosaccharides in Pathogenic Bacteria

Pathogenic bacteria contain LPS and capsular polysaccharides (CPS) in their outer membranes.
These compounds play important role in virulent properties of bacterial species, that is, invasion into
human immune system through molecular mimicry mechanisms [158]. LPS functions have been
under experimental research for several years due to their role in activating many transcriptional
factors, which become active after stimulation with LPS. LPS also produces many types of mediators
involved in septic shock. Structural characterization of the core oligosaccharide domain and O-
chain polysaccharide components help identify the infecting pathogens and classify new strains,
and therefore, contributes to the development of antibacterial drugs. For example, Pseudomonas
aeruginosa is a Gram-negative bacterium affecting individuals suffering from cystic fibrosis, certain
cancers, and immunodeficiency. The wild type of this pathogen expresses a heterogeneous population
of O-PSs, and LPSs obtained from cystic fibrosis patients have been found to be deficient of O-chain
extension [159].
To facilitate analysis of LPS/CPS, a partial hydrolysis is typically employed, either chemical
or enzymatic, which is specific for each type of polysaccharide. Hydrolysis is usually followed by
purification by a chromatographic means. Taking into consideration these time-consuming sample
preparation steps, online coupling of separation and structural characterization methods represent
advantage of potential minimization of sample pretreatment efforts. CE-MS have been exploited
in this type of applications by several research groups. Li et al. [86] used ESI-MS in-source
fragmentation technique, following CE separation, to break down CPSs and O-PS obtained from
Aeromonas pleuropneumoniae, instead of conventional chemical degradation pretreatment. This
approach helped to avoid extra sample preparation steps while keeping O-PS structure intact [85,86].
The latter also allowed acquiring information on O-acetylation of the molecules. In addition, online
preconcentration methods coupled with CE-MS have been attempted [159].
CE-MS was used for the characterization of LPSs at different structural levels, gaining more
understanding of mechanisms of action of the pathogens and, consequently, effecting drug design.
Moraxella catarrhalis is a pathogen affecting the respiratory tract and middle ear causing ottitis. CE-
MS/MS analysis of oligosaccharides derived from this bacterium revealed both basic (glucosamine)
and acidic groups (deoxy-manno-octulosonic acid) [158]. Mycobacteria contain lipoarabinoman-
nans (LAMs) as their capsular components, which comprise arabinimannan core and a hydrophobic
anchor, (e.g., phosphatidyl-myo-inositol). The immunological activity of LAMs depends on both
cap and anchor structure: arabinofuranosyl side chains can be substituted with either mannose
(ManLAMs) or phosphatidyl-myo-inositol (PI-LAMs) [59,160]. Nonvirulent mycobacteria have
been shown to have preferentially PI-LAMs, while ManLAMs were found in pathogenic Mycobac-
terium tuberculosis, Mycobacterium bovis, and Mycobacterium leprae. Mannooligosaccharides
can be hydrolytically released, derivatized by reductive amination with APTS, and undergo CE,
coupled online or off-line with MS, using either nano-ESI source or MALDI-TOF [160]. More
examples on characterization of oligosaccharides derived from bacterial sources can be found in
Reference 59.
7.4.2.5 Sugar Analysis in Traditional Chinese Drugs

There has been growing attention to traditional Chinese drugs (TCD) in the world, due to their
effectiveness in a number of disease therapies. Various sugars, including mono-, oligo-, and polysac-
charides, are significant constituents of traditional Chinese herbs, reaching in some cases up to
80–90% of the whole plant dry weight [125]. There are over 300 kinds of polysaccharides extracted
from natural plants. Among those, water-soluble polysaccharides typically possess most valuable
pharmaceutical properties. The mechanisms of pharmaceutical effects of bioactive sugars have been
extensively studied, and the number of their therapeutic applications have been increasing [161,162].
Most promising areas are anticancer activity [163–165], immune system stimulation [166–168]
through the production of interleukins and T and B cell regeneration, anti-HIV effects (polysac-
charides inhibiting binding of HIV to T cells) [169], blood sugar reduction (enhancing secretion of
insulin and carbohydrate metabolism) [125,170], and anti-inflammatory [171,172].
The analysis of pharmaceutical polysaccharides and other sugars in TCD typically involves
extraction, purification, and characterization steps. Water and weak basic solutions, as well as some
organic solvents, at different temperatures, are typically used to extract sugars [125]. Purification
can be accomplished by serial precipitation and chromatography steps (e.g., size exclusion, gel per-
meation, ion exchange, etc.) [125]. Electromigration separation methods for the analysis of sugar
constituents of TCD have been gaining popularity in the past decade, due to their high separa-
tion efficiency and compatibility with UV, LIF, refraction index (RI), and electrochemical detection
techniques, as well as hyphenation with MS and NMR. CE occupies a special niche in carbohy-
drate analysis, and TCD are not exception. In addition to high performance of CE, emphasized
throughout this chapter, simplicity of instrumentation and low sample and all the reagents con-
sumption are the main reasons for that. TCD sugars are usually separated and identified by CE as
mono- and oligosaccharides, from which the composition of polysaccharides can be derived. Thus,
complex polysaccharides are typically hydrolyzed to render simpler oligomers and monomers. Elec-
trochemical detection and LIF of underivatized and fluorescently labeled saccharides, respectively,
are highly sensitive. For example, when using amperometric detection approach µM sensitivity level
for monosaccharides can be achieved with copper or gold electrodes [173–175]. Different aqueous
separation buffers have been utilized, including highly basic ones (pH > 11–12), when most neu-
tral sugars become negatively charged due to their weak acidic properties. Sodium hydroxide based
running buffers, borate (pH 10), phosphate–borate buffers, used in combination with uncoated fused-
silica capillaries, have most often been mentioned. Certain buffer additives, for example, surfactants
(SDS, CTAB), THF, tryptophan, can often improve separation efficiency. Examples and references
to analysis of carbohydrates in traditional Chinese medicines can be found in the recent review article
by Wang and Fang [125].
One promising approach to be exploited more in the future with respect to TCD analysis is
microchip-based CE. It is expected to enable ultimate integrated and highly sensitive systems to be
built [176–178]. CE-MS coupling represents another area of fast development in carbohydrate analy-
sis of TCD, as it has advanced in other applications in research, biopharmaceutical, and bioindustrial
fields.
7.4.3 CARBOHYDRATES IN FOODS

Carbohydrate composition of foods and feeds allow characterization of the quality, ripeness, possible
adulterations, and so forth. CE-MS approach has been widely used in analysis of foods. With a few
exceptions [179], food carbohydrates are analyzed without any derivatization. Arabinose, ribose,
xylose, inositol, galactose, fructose, and mannitol were monitored by CE-ESI-MS, using highly
alkaline CE running buffer and negative ionization mode [180], reaching LOD 0.5–30 µg/mL.
Figure 7.14 exemplifies different carbohydrate patterns in red and white wines. The higher contents
for most sugars were found in white wines. CEC-ESI-MS approach was exploited for the analysis of
(a)
25000
TIC
1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 min
2000 m/z 133
2914
2 34
m/z 149
10000 11
m/z 179 7 8 10
5000 12
m/z 181
(b)
25000
TIC
3000 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 min
1
m/z 133
2000 2 3 4
m/z 149
10000 11 8 10
m/z 179 7
4000 12
m/z 181
FIGURE 7.14 CE-MS total ion current (TIC) and SIM electropherograms for (A) a red wine and (B) white wine
sample. CE conditions: fused-silica capillary (70 cm length to detection, 50 mm ID); running buffer 300 mM
diethylamine (DEA); running voltage 20 kV; injection at 50 mbar for 9 s. ESI-quadrupole MS conditions:
negative ion mode; capillary voltage 5 kV; sheath liquid 2-propanol:water (80:20 v/v) with 0.25% DEA at a
flow rate of 4 mL/min; drying gas N2 at a flow rate of 1.4 L/min and a temperature of 1507◦ C; nebulizer
pressure 0 psi. Peaks: 1: deoxyribose; 2: arabinose; 3: ribose; 4: xylose; 5: galactose; 6: glucose; 7: fructose;
8: inositol; 9: mannitol. (Adapted from Klampfl, C. W. and Buchberger, W. Electrophoresis, 22, 2737, 2001.
With permission.)
enzymatically digested GAGs [151] and other derivatized and underivatized glycans [69], complex
oligosaccharide mixtures using hydrophilic monolithic columns [26], and FT-MS to achieve high
mass resolution [53]. MEKC-ESI-MS was attempted to the analysis of iridoid glycosides in plant
samples from different species [181]; in this case, the partial filling technique was used to prevent
surfactants from entering the MS interface. Gellan gum, a complex sugar polymer food additive,
was detected based on the occurrence of a characteristic tetrasaccharide released in the process of
enzymatic digestion by a specific gellan degrading enzyme [182]. Low molecular weight sugar-based
acids (gluconic and isosaccharinic) present in ale were unambiguously identified by CE-MS, which
helped overcome the problem associated with comigration of the two acids by SIM application
during acquisition of MS spectra [183]. Glycoalkoloids of vegetable origin are also compounds of
interest due to their toxic properties and occasional presence in food products. CE-MS was used
to separate some typical glycoalkoloids, as well as related aglycones, in potatoes (α-chaconine,
α-solanine, solanidine) and tomatoes (α-tomatine, tomatidine) [184]. In this case, nonaqueous CE
coupled with SIM or SRM (selected reaction monitoring) MS mode resulted in excellent separation
and identification of all the investigated compounds. Moreover, nonaqueous buffer was found to
significantly improve ionization efficiency. Another interesting application of CE-MS was reported
by Bringmann et al. [185] who characterized glucosinolates from seeds, the compunds of anticar-
cinogenic effects. The authors employed IT analyzer for structural elucidation, while TOF MS was
used for information on elemental and isotopic composition. Higher analysis efficiency was achieved
compared to chromatographic methods, resulting in not only identification of known compounds but
also discovery of some new ones.
7.4.4 BIOINDUSTRIAL APPLICATIONS

Enzymatic saccharification of lignocellulosic biomass to fermentable sugars is of great importance
for a number of agroindustrial processes. These include conversion of cellulose and hemicelluloses
to fuels and other important chemicals, delignification of paper pulp, digestibility enhancement of
animal feed stock, clarification of juices, to name just a few. Chemical pretreatment methods in con-
junction with proper cost-effective enzyme combinations (e.g., hemicellulases or cellulases) tailored
to each specific biomass will enable production of fuel additives, such as ethanol, and other valu-
able chemicals by fermentation from this vast renewable resource [186–190]. Various agricultural
residues, such as corn fiber and stover, wheat and rice straw, and so forth, contain up to 40% hemicel-
lulose, the second most abundant polysaccharide in nature after cellulose. Cost-effective conversion
of these materials to valuable chemicals is a current issue for the industry [188]. Lignocellulosic
biomass structure is very complex [186,187]. Industrial corn fiber contains cellulose chains, assem-
bled into highly crystalline and aggregated structures, microfibrils and fibers. The cellulose fibers are
imbedded in a matrix of hemicellulose (highly branched heteropolysaccharides) and lignin. Cellulose
and hemicellulose degrading enzymes are found in nature either as enzyme complexes or free in solu-
tion [191]. Different enzyme classes are involved in the biological depolymerization of cellulose and
hemicellulose components: cellulases (endoglucanases, cellobiohydrolases, and β-glucosidases);
xylanases to degrade the hemicellulose backbone, as well as accessory enzymes, for example, arabi-
nofuranosidases, mannanases, galactosidases, glucuronidases, ferulic acid esterases and acetyl-xylan
esterases. It is desirable to find high-specific activity enzymes that effectively hydrolyze biomass fiber
in synergistic fashion [189]. Evaluation of different enzymes and their combinations, optimization
of the reaction conditions (temperature, pH, reagent concentration, reaction time, etc.) are critical.
Therefore, hundreds and thousands of samples are to be analyzed to screen the various enzyme
combinations, tune reaction parameters, and characterize novel expressed recombinant enzymes and
microorganisms.
Implementation of efficient and robust industrial scale bioconversion processes requires high-
throughput analysis of mono- and oligosaccharide products, released by enzymatic digestion of
various biomasses, to evaluate activity of enzymes and microorganisms. CE with LIF detection
offers excellent resolving power and high sensitivity, superior to many other analytical approaches
[192]. High-resolution separations of complex carbohydrate mixtures, simple sugars, glycoforms of
glycoproteins, and glycolipid oligosaccharides by CE have been demonstrated for the past decade
[19,48,193]. Using suitable fluorescent-labeling procedures, charged derivatives of various carbo-
hydrates can be formed enabling attomole detection level in CE [193]. The majority of carbohydrate
derivatization methods utilize tagging of the reducing termini with a fluorophore via reductive
amination [34]. APTS is one of the most popular of such derivatization agents, providing very bright
fluorescence and good spectral characteristics suitable for the commonly used Ar-ion laser [195].
There is a great demand for large, industrial-scale carbohydrate profiling methods to meet the
current growing needs to screen enzymatic activity of recombinant enzyme and microorganism
libraries. A few advances have been made to date to establish reliable high-throughput methodology
for automated oligosaccharide analysis. Morell et al. [196] adapted a polyacrylamide slab gel based
DNA-sequencing device to parallel multilane analysis of oligosaccharide products of enzymatic
digests of starches. Although, good resolution up to DP 75 and sensitivity (1 fmol per oligosaccharide)
were achieved, the technique had limitations with respect to reproducibility, precise tracking of gel
lanes, resolution of subtle differences in profiles, and labor consumption. Another research group
[197] utilized a similar DNA-sequencing instrument using 32 parallel lanes, in combination with
sample preparation, derivatization and cleanup steps, to accomplish ultrasensitive profiling and
sequencing of N-linked glycans. On the other hand, CE demonstrated superior performance enabling
fast and reproducible separation in the range of DP 1–100, providing at the same time flexibility
of separation formats for the development of various assays [48]. A multicapillary CE-based DNA
analyzer combined with 96-well-plate-based sample preparation method and a thermocycler were
used to perform high-throughput clinical analysis of N-linked oligosaccharides from serum protein
of liver disease patients [141].
Khandurina et al. [198] demonstrated the capabilities of a single CE instrument to provide auto-
mated high resolution and quantitative analysis of mono- and oligosaccharide products of enzymatic
digestion of cellohexaose, as a model substrate, and corn fiber biomass in bioindustrial settings.
The authors developed various protocols to accommodate different assays needs. Unattended batch
sample processing from 96-well plates after the one-step derivatization reaction enabled reliable
industrial-scale carbohydrate profiling. A single CE system (P/ACE MDQ; Beckman Coulter) was
employed, with the cathode on the inlet side and anode on the outlet side (reversed polarity), since the
negatively charged APTS-labeled oligosaccharides migrate toward the anode under applied electric
field. For the CE analysis, either bare fused-silica (Polymicro Technologies, Phoenix, AZ) or coated
capillary columns (eCAP™ N-CHO Capillary; Beckman Coulter) of 50 µm ID were used, with
the effective separation length either 48 cm to achieve high resolution or 10 cm in fast screening
experiments. 25 mM lithium acetate (pH 5) buffer was used as running buffer. Samples were injected
by vacuum for 5–10 s at 0.5 psi (3447 Pa) and separated at 500 V/cm electric field strength. In case
of uncoated capillaries, a dynamic coating by 1% polyethylene oxide (MW 600,000 Da) was utilized
to suppress EOF. Figure 7.15 shows high-resolution oligosaccharide profiling by CE using 48 cm
effective separation length. Traces A and B depict the electropherograms of a maltooligosaccharide
ladder and cello-oligomer standards, respectively. There are well-observable differences in migration
times of the oligosaccharides with the same DP but different linkages between glucose monomer
units (e.g., maltose vs. cellobiose, maltotriose vs. cellotriose, etc.). These differences result from the
changes in structure and, therefore, hydrodynamic radius of the molecules, having either α-1,4 or
β-1,4 glycosidic bonds. Profiles C–J in Figure 7.15 show CE analyses of the reaction products of
cellohexaose digestion (trace C) using different cellulases. Precise monitoring of the differences in
cellulase enzymatic activity, revealing a variety of combinations of glucose, cellobiose, cellotriose,
cellotetraose, and cellopentaose, was readily achieved by CE. The excellent migration time repro-
ducibility allowed easy identification of the cello-oligomer products of the cellohexaose enzymatic
hydrolysis.
Separation of monosaccharides is a challenging task, particularly due to very minor differ-
ences between most common sugar molecules, representing the building blocks of lignocellulosic
biomasses. Monosaccharide analysis by CE has been previously demonstrated using borate-based
buffer systems at high pH (>9), taking advantage of the resolution enhancement due to anionic com-
plex formation between the polyhydroxy compounds (sugars) and borate [199,200]. Good separation
performance was demonstrated using CZE [200] or MEKC [27] in bare fused-silica capillaries under
EOF conditions. However, in most cases an internal standard was required to account for possible
migration time shifts. The authors [198] have found that comparable separation performance of
some monosaccharides (major building blocks of hemicellulose) can be achieved using the same CE
3
2 4 5 6 7
10 RFU 1
A
1 3 4 5 6
2
B
6
C
1 2 3
D
1 2 3 4
E
1 2
F
2 3
G
2 3 4
H
1 2 3
I
1 2 3 4 6
5
J
3 4 5 6 7 8 9 10 11 12
Time (min)
FIGURE 7.15 High-resolution oligosaccharide profiling by CE. Traces: (A) Maltooligosaccharide ladder;
(B) cellooligosaccharide ladder; (C) model substrate cellohexaose; (D–J) products of enzymatic digestion of
cellohexaose by seven different enzymes with cellulase activity. Numbers above peaks correspond to DP of
oligosaccharides. Conditions: eCAP neutral coated capillary (50 µm ID); effective separation length 48 cm;
electric field strength 500 V/cm; running buffer 25 mM lithium acetate (pH 5); injection 5 s at 0.5 psi (3447
Pa). (Adapted from Khandurina, J. et al. Electrophoresis, 25, 2326, 2004. With permission.)
conditions, as described above for oligosaccharide profiling. Moreover, the absence of EOF typi-
cally renders more reproducible performance and elimination of additional rinsing steps, otherwise
required.
Trace A in Figure 7.16 shows a well-resolved separation of a mixture of the five monosaccharides
of interest in just 5 min, using 25 mM lithium acetate (pH 5) running buffer under apparently zero
EOF conditions. Glucuronic acid (1) is the fastest migrating sugar, due to its additional charge,
followed by the pentoses of xylose (2) and arabinose (3), and the hexoses of glucose (4) and galactose
(5), respectively. By this means, monosaccharide reaction products of corn fiber hydrolysis, either
chemical or enzymatic, can be easily identified, assuming very similar labeling efficiency for all
the released monosaccharides [34,196]. Traces B and C in Figure 7.3 depict separations of sugars
released from equal amounts of corn fiber by acidic hydrolysis, for 2 h at 100◦ C, with 1 M sulfuric
acid and 2 M tetrafluoroacetic acid (TFA), respectively. The glucose (4), most likely, comes from the
cellulose domains of corn fiber. To identify the peaks precisely, in ambiguous cases, the samples were
spiked with proper amounts of standards. Trace D in Figure 7.16 delineates decomposition products
of corn fiber under harsh hydrolysis conditions, started with 72% H2 SO4 for 1 h followed by dilution
to 4%, autoclaving at 120◦ C for 1 h, and storing for 4 months after the treatment. This procedure
apparently resulted in the release of significantly more monosaccharides, as one can observe from
the increased peak areas of this particular trace. In addition, the relative glucose concentration was
4
1 2 3
5
A
8
23
45
B
6 C
RFU
D
2 4
3
2
5
E
0
3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0
Time (min)
FIGURE 7.16 Monosaccharide analysis by CE under zero EOF condition. Traces: (A) five standard sugar mix
(1: glucuronic acid, 2: xylose, 3: arabinose, 4: glucose, and 5: galactose); (B) corn fiber acid hydrolysates—
1 M sulfuric acid; (C) 2 M TFA; (D) 72% sulfuric acid; (E) corn fiber enzymatic digestion. Conditions: same
as in previous figure, except a bare fused-silica capillary (50µm ID, 48 cm effective separation length) was
used with dynamic PEO coating. (Adapted from Khandurina, J. et al. Electrophoresis, 25, 2326, 2004. With
permission.)
substantially elevated (peak 4) compared with milder hydrolysis conditions. An illustrative example
of the enzymatic digestion of corn fiber by a fungal culture medium is shown in Trace E (Figure 7.16).
The lower relative content of the released xylose, compared with the acidic hydrolysates, was caused
by lack of xylanase activity in the enzyme system used in the experiment. Most of the other earlier
migrating peaks in the electropherogram are probably associated with background and matrix effects
originating from the labeling reaction and constituents of the enzyme medium.
The same research group introduced a 96-capillary array electrophoresis (CAE) approach for
large-scale mono- and oligosaccharide analysis and characterization [201,202]. In this work, a DNA
sequencer, MegaBACE1000 (Amersham Biosciences) was adapted for carbohydrate screening by
developing and optimizing operational protocols and data processing tools. This approach brings
high-performance carbohydrate analysis within reach of life sciences laboratories without the need
for additional costly equipment, and helps expedite the pace of discovery and development of biomass
conversion processes, as well as glycobiology studies. Carbohydrate samples (1 µL) were labeled
in 96-well plate format, through reductive amination by the addition of 2 µL of 0.2 M APTS in
15% acetic acid and 2 µL of 1 M NaBH3 CN in tetrahydrofuran [19,34,48,193,194]. The plates were
incubated for 1 h at 75◦ C, followed by the addition of 100 µL water to stop the reaction. Before CE
analysis, the APTS-labeled samples were further diluted in water (up to three orders of magnitude
depending on the initial concentration of carbohydrates), resulting in low nM concentration range
of each derivatized carbohydrate species. The authors found that uncoated fused-silica capillary
arrays (The Gel Company, San Francisco, CA) in combination with dynamic PEO coating ren-
der more consistent performance and significantly longer lifetime compared to original covalently
coated capillary arrays, designed for DNA-sequencing applications. The dynamic coating is easily
regenerated before each run by sequential washing with hydrochloric acid, water, PEO solution
in the running buffer, and finally lithium acetate running buffer. It is important to emphasize that
carbohydrate analysis was performed in CZE (open tube) mode and, therefore, more sensitive to
capillary surface quality and any residual EOF, unlike in DNA sequencing and genotyping, where
viscous sieving media can compensate for some coating nonuniformity and deterioration. To further
speed up the process, the dynamic coating polymer can be directly added to the running buffer.
One important problem to avoid in carbohydrate CAE is a possible siphoning effect caused by
uneven liquid levels in inlet and outlet buffer reservoirs [202]. Capillary-to-capillary and run-to-run
variation in migration time and signal intensity necessitated the development of data normalization
tools. Internal bracketing fluorescent standards were incorporated into the analysis enabling CAE
trace alignment. A four-color detection system allowed a choice of different fluorophores to spec-
trally separate the sample peaks and internal standards in the corresponding detection channels. An
excess of unreacted APTS, remaining in the samples after derivatization and migrating faster than
the labeled sugars, was used as leading bracketing standard. A fluorescent dye sulforhodamine B
(SRB), emitting in the red spectral region (λmax = 586 nm) and migrating with similar mobil-
ity as the maltooligosaccharide of DP 16, served as a terminating bracketing standard. A set of
programming tools were developed to process the data and properly normalize migration time for
each electropherograms based on internal bracketing standards. The procedure is described in detail
in Reference 201. Figure 7.17 presents a set of raw (left) and aligned (right) electropherograms,
utilizing the developed algorithm. APTS-derivatized maltoologosaccharide ladder mixed with the
terminating bracketing standard (SRB) was used in this experiment. As one can see, all correspond-
ing oligosaccharide peaks match very well in the aligned data set in the right panel. The lower panel
in Figure 7.17 delineates the migration time of APTS peak across the entire 96-CAE before and
after the normalization procedure. A mixture of four monosaccharides, xylose, arabinose, glucose,
and galactose were also successfully separated by CAE under low-pH conditions (Figure 7.18). The
data were normalized, similar to the maltooligosaccharide ladder, enabling an accurate peak identi-
fication of closely migrating monosaccharides. Figure 7.18 shows five representative traces of such
analysis before (upper panel) and after (lower panel) the data normalization. Albeit, the efficiency of
APTS-labeling reaction is very similar for most monosaccharide product and short oligosaccharide
product, precise quantification requires correction for possible small differences [201]. This can be
accomplished by introduction of another internal standard into the sample derivatization mix, for
example, oligosaccharide migrating with no overlapping with the sugars of interest. Relative label-
ing efficiency coefficients can be derived this way and used for accurate quantification of the sugars
of interest in the unknown samples [201]. The described large-scale qualitative and quantitative
carbohydrate profiling method based on 96-CAE can accommodate high-throughput demands of the
biotechnology industry. The areas of application for the developed methodology include biomass
conversion processes, which require massive screening of enzymatic reaction products to evalu-
ate large libraries of recombinant enzymes and microorganisms, as well as protein glycosylation
studies.
7.4.5 MINIATURIZATION IN CARBOHYDRATE ANALYSIS

The recent fast developing trend for miniaturization of analytical systems has influenced carbohydrate
analysis as well. Advances in microfabrication technologies, microfluidics-based devices, and micro-
chip electrophoresis in particular, in combination with available high-sensitivity detection methods,
stimulated a series of attempts to implement carbohydrate analysis on-chip. However, advancement
of on-chip glycomics has been somewhat limited compared to fast growth of microscale genomic
and proteomic applications. A number of limitations, such as lack of charge and functional groups
and detection issues, present a challenge. Suzuki and Honda [17] have briefly summarized the latest
RFU
Raw data Aligned data
200000
100000
0 5 10 2 15 0 5 10 15 20
Time (min)
4.5
Time (min)
4.0
3.5
3.0
0 40 80
Capillary number
FIGURE 7.17 Migration time normalization and CAE data alignment. Left and right panels represent raw
and aligned traces, respectively, for maltooligosaccharide ladder samples containing SRB as internal bracketing
standard. The lower panel shows migration time standard deviation forAPTS peak before and after normalization
procedure. See details of data processing in the text. Conditions: injection 10 s at 5 kV, separation 15 kV (240
V/cm), temperature 30◦ C, 25 mM lithium acetate (pH 5) separation buffer, uncoated fused-silica capillary arrays
dynamically coated with PEO (see details in the text). Arrows and numbers indicate DP of oligosaccharides,
for reference. (Adapted from Khandurina, J. et al. Electrophoresis, 25, 3122, 2004. With permission.)
developments in this area. Mechref and Novotny [32] described chip approaches in glycomics, both
CE and chromatography based, in their latest review.
Miniaturization of sample handling and processing steps should significantly reduce losses asso-
ciated with adsorption at the surfaces, sample transfers, as well as potential contamination. Integration
and concomitant minimization of the analysis steps on-chip, including purification, concentration,
separation, and detection, will ultimately enhance sensitivity and efficiency of the carbohydrate
analysis, and glycan screening in particular.
Several research groups implemented carbohydrate analysis on-chip with direct detection of
underivatized sugar molecules. Electrochemical detection is the most attractive approach, as it
offers reasonable sensitivity and selectivity, and it is ideally suited for microchip format. Schwarz
et al. [203] developed amperometric detection of sugars using microfabricated copper electrode.
They separated fructose, sucrose, and galactose in 70 s on a glass chip with 50-µm wide and 20-µm
deep microchannel and double tee injection geometry. The detection was based on Teflon-coated
platinum wire plated with copper and inserted in the end of the separation channel etched in a conical
shape. The detection limit down to 1 µM was achieved. Hebert and coworkers [204] reported an
RFU
200000
100000
0
5 7 9
Time (min)
FIGURE 7.18 CAE analysis and data alignment for monosaccharides. Only five traces representing wells #
60–65 of a sample plate are shown to better visualize the separation and data processing performance. Raw
and normalized data are in the upper and lower panels, respectively. The first migrating peak in each trace is
unreacted APTS followed by xylose, arabinose, glucose, and galactose, in order of migration. Electrophoretic
conditions (except for injection ay 1 kV for 5 s) and internal bracketing standards are the same as in previous
figure. (Adapted from Khandurina, J. et al. Electrophoresis, 25, 3122, 2004. With permission.)
interesting detection system based on sinusoidal voltammetry, where Fourier transformation was
used to convert current data, associated with carbohydrate molecules, into the frequency domain,
thus, effectively isolating the analyte signal from background current. A polydimethylsiloxane chip
with a planar copper electrode was fabricated, enabling detection limit of 200 amol and efficiency of
105 theoretical plates per meter. The signal-to-noise ratio was enhanced by utilizing a digital lock-in
method, which allowed selective picking and masking of carbohydrate signals. An on-chip-pulsed
amperometric detection system, described by Fanguy and Henry [205], accomplished analysis of
glucose, xylose, and maltose in 4 cm separation channel with a platinum electrode built in the end part
of the channel. The authors demonstrated a linear response in 20–500 µM concentration range, with
20 µM detection limit. A unique system, combining flow injection and short CE, was constructed by
Fu and Fang [206]. Some carbohydrates were monitored by amperometric detection on a horizontal
channel connected to a vertical plastic tubing. A solution of glucose and sucrose was continuously
introduced by droplets from the tubing into the junction with the channel followed by electrophoretic
separation. Sugars were separated in 60 s on a 5-µm ID and 5-cm-long capillary, with a dynamic
range of 10–1000 µM.
Refractometry is another approach for direct detection of carbohydrates. An interesting
holography-based refractive index detector was implemented by Burggraf et al. [207]. A cyclic
channel of 80 mm in circumference and 10-µm deep was fabricated. The detection scheme was con-
structed having a diode laser beam (670 nm) split into two beams, one passing through the separation
channel and the second through the adjacent reference channel. A photodiode array detected changes
in holographic optical image. Separation time (17 s) of raffinose, sucrose, and N-acetylglucosamine
was achieved, with the detection optimum at 10 mM level (600–900 fmol of injected material).
Although the analysis sensitivity was quite limited, the described coupling of electrophoresis
microchip and refractive index detector can be used for certain application when sensitivity is not
required (e.g., measurements of simple sugars in foods).
The most successful microchip separations were achieved with derivatized carbohydrates when
a UV absorbing or, better, fluorescent tag is introduced. On-chip UV detection suffered from low
sensitivity and allowed only mM detection level using PMP label and free zone electrophoresis in
borate buffer [208]. To alleviate poor detection limit, microchip Shimadzu instrument is equipped
with the special whole-channel detection system, where the signal is enhanced by simultaneous
accumulation and averaging of the absorption data from multiple points. Baba and coworkers [209]
analyzed APTS-labeled oligosaccharides on a Hitachi SV1100 microchip system using a blue light
emitting diode (LED) source and confocal fluorescent detection configuration. In this work, a plastic
polymethylmethacrylate (PMMA) chip was employed, and the detrimental adsorption of APTS to
PMMA material was alleviated by dynamic coating with methylcellulose to the running buffer. The
maltooligosaccharides were resolved up to DP 10 (maltodecaose) in 2 min. In another work from the
same group [210] protein glycosylation structures were analyzed: high-mannose-type glycans were
enzymatically released from bovine spleen ribonuclease B and separated on-chip, using asymmetric
pinched injection with field-amplified stacking resulting in 20-fold improved sensitivity. Derivatiza-
tion of sugar reducing ends by reductive amination has been undoubtedly one of the most successful
and universal approaches; however, some classes of carbohydrates contain functional groups that are
suitable for alternative chemical couplings. For example, hexosamines and their derivatives possess
an amino group at the C2 position, and therefore can be readily fluorescently labeled via reactions
developed for aminoacids. Suzuki et al. [211] implemented a condensation reaction of hexosamines
with 4-nitro-2,1,3-benzoxadiazole 7-fluoride (NBD-F), followed by a microchip electrophoresis on
a quartz chip using 33 mm separation distance and Ar-ion LIF detection. The authors separated
glucosamine, galactosamine, and their corresponding reduced forms, glucosaminitol and galac-
tosaminitol, in 1 min under free buffer conditions (phenylborate sugar complexes), reaching detection
limit of 2.5 µM (0.5 fmol injected amount). The same labeling methodology was successfully applied
to O-linked glycans enzymatically stripped from glycoproteins (e.g., bovine submaxillary mucin).
Another interesting example of miniaturized carbohydrate analysis is on-chip integration of
enzymatic reactions, conventionally used for sugar determination, with electrophoretic separation
of the reaction products. Wang and coworkers [174] demonstrated a simultaneous assay of glucose,
ascorbic acid, and acetaminophen on a microchip with cross-channels and a gold electrode assembly
in the separation channel. The sample containing glucose was mixed with glucose oxidase (GOx) in
the channel intersection, followed by the oxidation reaction and separation of the generated hydrogen
peroxide from ascorbic acid and acetaminophen. The latter two species migrated more slowly due to
their anionic nature. The gold-coated thick-film amperometric detector was positioned downstream
in the separation channel, and the glucose level was assessed based on comparison of response with
and without GOx (Figure 7.19) [174]. The same group reported an oxidase/dehydrogenase assay,
implemented on-chip, for the measurements of glucose and ethanol [175]. The developed assay was
applied to the analysis of wine samples.
Integration of microchip CE and MS is a new emerging technology, which is expected to
enhance high-throughput glycan mapping and sequencing. The combination of microfluidics,
automation/robotization, and software for assignment of MS data has already been implemented
by introducing Advion Nanomate dispenser and chip to glycoscreening by coupling it to QTOF and
FT-ICR mass analyzers [73]. The fully automated chip-based MS approach for complex carbohy-
drate system analysis was applied to urine analysis of patient suffering from hereditary diseases.
Addition of sample CE preseparation to the ESI-QTOF via Nanomate system has been demonstrated
Separation/reaction
Channel outlet
channel
Sample
Buffer with
GOx
Buffer 25 mm
80 mm
AA– Glu– H2O2

UA–
UA–
Glu AA– Glu H O
2 2
–
Glu
UA O2 Gla
E Glu
UA–
AA– H2O2 Glu
UA– Glu
AA– AA– H2O2
H2O2 Glu H2O2

UA– UA–
AA–
O2 Gla
Glu
Glu E
AA–
–
AA H2O2 Glu H2O2
AA–
UA– UA–
H2O2 O2 Gla
AA–
E H2O2
UA– AA– H2O2 H2O2 Glu
H2O2 H2O2
UA– AA–
UA– AA– Glu
H2O2
UA – AA– – H2O2
AA
H2O2 H2O2
UA– –
AA
UA– AA– H2O2
O2 Gla
UA– Uric acid Enzyme catlyzed
E
H2O2 Glu reaction
AA– Ascorbic acid
Hydrogen
H2O2
Glu Glucose peroxide
FIGURE 7.19 Layout of the separation/reaction microchip for bioassays of glucose, ascorbic acid, uric acid,
and acetaminophen (upper panel). Enzymatic and separation processes along the reaction/separation channel
of the CE biochip (lower panel). (Adapted from Wang, J. et al. Anal Chem, 72, 2514, 2000. With permission.)
to provide advantages compared to more conventional CE-MS interface, that is, higher ionization
yield, decreased in-source fragmentation, and stable spray [212].
Recently, a total serum protein N-glycosylation profiling was attempted on a CE chip by Ehrlich
and coworkers [213]. The authors employed a glass chip with a double-tee injector and a 4% linear
polyacrylamide sieving medium (analogous to nucleic acid separations). Profiling of serum sam-
ples from chronic hepatitis patients identified the differences in N-glycan composition in cirrhotic
and noncirrhotic cases, and demonstrated the potential of microchip approach for these types of
clinical studies.
Integrated microchips, packed with reversed-chromatographic phase and interfaced with TOF
[214] or IT [215] MS, for the analysis of oligosaccharides, O- and N-glycans of different sources
have also been reported.
Although only relatively simple carbohydrate systems have been transferred on-chip so far,
primarily due to resolution limitation in short separation distances and sensitivity challenges in
microchannels, rapid progress in this area is expected in the nearest future. Column efficiencies
observed in microchannels were similar or higher (the latter was due to shorter injection plugs
realized in microchip settings) compared to conventional CE format. Shorter separation paths utilized
in miniaturized analysis settings limit complexity of analytes, which can be successfully resolved. On
the other hand, microfabricated devices offer a number of other attractive analytical advantages. One
of them is possibility of integration, that is, combining sample prep and derivatization, separation,
and detection steps on a single integrated device [216,217]. Second, disposable microchips can be
inexpensively mass produced and utilized in testing analytical kits or point-of-care diagnostic tools
in food, environmental, and medical applications.
7.5 METHOD DEVELOPMENT GUIDELINES

As it has been emphasized throughout this chapter, CE in its various modes of separation and
detection is suitable for the analysis of a wide variety of carbohydrate molecules, that is, mono-,
oligo-, and polysaccharides, glycopeptides, glycoproteins, and glycolipids. This can be credited to
the progress made in the capillary column technology, introduction of novel electrolyte systems, as
well as the development of various detection approaches, such as indirect UV and LIF detection,
electrochemical detection, CE-MS interface, and precolumn labeling with suitable chromophores
and fluorophores. Precolumn derivatization proved to be one of the most elegant approaches for
the separation and detection of carbohydrates. Among the different reaction schemes, introduced
for the labeling of carbohydrates, multiply charged tags (ANDSA, ANTS, and APTS) have become
the most popular, not only because of their high detection sensitivity by either UV or LIF, but also
because they yield derivatives that are readily separated by CE.
Majority of CE separations are typically accomplished utilizing borate complexation, for both
derivatized and underivatized carbohydrates. However, noncomplexing electrolyte systems (phos-
phate, acetate based buffers, for example) are often successfully used as well. As it has been described
in Section 7.2, borate complexation enhances small structural differences of structurally similar
molecules, resulting into highly selective separations of multicomponent sugar mixtures. Under a
given set of conditions, various sugars, whether charged or neutral, undergo varying degrees of
complexation with borate leading to differences in the electrophoretic mobilities of the complexed
solutes and hence separation.
Highly alkaline electrolyte solutions, for example, lithium, potassium, or sodium hydroxide at
pH > 12 have been proved to be useful in the separation of underivatized saccharides by CZE.
The resolution of various saccharides is typically improved with pH increase (up to pH 13) due to
increasing ionization of the separated analytes. In addition, the nature of the alkali-metal influences
the resolution of the sugar analytes.
The separations at extremely high pH can only be performed in bare fused-silica capillaries,
since most coatings undergo hydrolytic degradation under the highly basic conditions. The use of
electrolyte systems containing alkaline-earth metals for the separation of neutral carbohydrates by
CZE is mainly based on differences in the extent of complexation of the divalent metals with the
carbohydrate solutes, as well as the size and shape of the molecule. These systems provided a
different selectivity from that of borate buffers, although the resolution is often inferior.
The electrolyte systems in CE can be easily modified and tailored to a specific separation task.
For example, most glycolipids are not compatible with aqueous electrolyte solutions but they can
be readily separated in their monomeric forms in hydroorganic buffers. Buffer additives, such as
surfactants and polymers, often enhance separation efficiency. Examples of various separation buffer
systems, suitable for different classes of carbohydrates, and approaches to their development can be
found in the above sections, dedicated to specific types of carbohydrate species and their sources,
and references therein.
Characterization of glycoconjugates are typically carried out by first cleaving the carbohydrate
moieties from glycoproteins, glycolipids, GAGs, or other complex molecules, often followed by
further reduction of the structure complexity through enzymatic digestion and/or depolymerization.
Owing to enormous diversity and complexity of carbohydrates and their conjugates with other
classes of biomolecules, it is impossible to briefly outline any definitive guidelines for analysis
method development. Each type of carbohydrate molecules should be approached separately, based
on its specific properties, such as molecular weight, linearity or branching, presence/absence of
the reducing end, charge, additional functional groups present, matrix complexity, and so forth.
Therefore, the reader is directed to the above sections to find basic modern strategies, as well as
specific protocols in the corresponding references, for sample preparation, derivatization, separation,
and detection in the analysis and structural characterization of simple and complex saccharides and
glycoconjugates in pharmaceuticals, foods, plants, bacteria, animal tissues, and other sources.
7.6 CONCLUSIONS AND OUTLOOK

Progress in carbohydrate research and concomitant advances of analytical methodologies during
the past decade have been undoubtedly remarkable. In particularly, the expansion of CE and its
various implementations supported the increasing demand for new level of information output, such
as integrated systems biology approach.
The future developments in CE-based analysis of carbohydrates are envisioned to bring forward
automation and miniaturization to ensure high-throughput and large-scale carbohydrate profiling to
keep up with the industrial needs. On the other hand, capabilities of elaborate structural insights
into complex carbohydrate molecules should be further developed as well. Hyphenated techniques,
such as CE-MS, offer possibility to identify composition of complex polysaccharides and glyco-
conjugates, as well as their sequence and fine structural details, that is, isomeric configurations,
branching, substitutions, linkages between monomer units, and so forth. Glycomics and glycopro-
teomics deserve special attention, as part of system biology efforts toward complete understanding
of the complexity of living organisms. Just like the developments in genomics, transcriptomics,
proteomics, and metabolomics, functional glycomics is strongly dependent on the current and future
advances in analytical methodologies and instrumentation. Quantitative measurements are essential
to better understanding of cellular and molecular interactions associated with different physiological
and pathological processes in living systems. CE-based methodologies directly combined with MS
are particularly useful in this respect. Electromigration techniques, for example, CZE, MEKC, and
CEC, exhibit most superior resolving power enabling separation and identification of close oligosac-
charide isomers. LIF is most preferred detection approach for high-sensitivity profiling of glycan
mixtures by CE. Coupling these techniques to MS is still a challenging task, although significant
progress has been made. Increasing number of manufacturers, offering ready-to-use CE-MS instru-
ments and a variety of available interfaces (especially nanospray sheathless approach), represent a
step forward toward a more routine use of carbohydrate analysis by CE-MS. Growing sophistication
of MS instrumentation fuels development of new CE-MS applications, involving IT, TOF, TOF-
TOF, and FT-ICR. CE-MSn offers powerful capabilities. However, optimization of this technology,
including instrumental solutions, sample preparation, and separation aspects, is still in progress, and
robust standardized systems and protocols are to be refined yet. Careful choice of separation buffers is
the key, as many salts and other buffer components adversely affect sensitivity of most MS detectors.
It is reasonable to expect that microfluidic and chip-based technology will become widely
applicable to glycomics, as it has already become in many aspects of genomics and pro-
teomics. Miniaturized CE-MS technology should develop into reliable task-oriented commercial
instrumentation, applicable to investigation of different biological systems under well-designed
conditions.
REFERENCES
1. Crocker, P. R. and Feizi, T. Carbohydrate recognition systems: functional triads in cell–cell interactions.
Curr Opin Struct Biol, 6, 679, 1996.
2. Helenius, A. and Aebi, M. Intracellular functions of N-linked glycans. Science, 291, 2364, 2001.
3. Karlsson, K. A. Meaning and therapeutic potential of microbial recognition of host glycoconjugates.
Mol Microbiol, 29, 1, 1998.
4. Zamfir, A. and Peter-Katalinic, J. Capillary electrophoresis-mass spectrometry for glycoscreening in
biomedical research. Electrophoresis, 25, 1949, 2004.
5. Gordon, H. T., Thornburg, W. W. and Werum, L. N. Rapid paper chromatographic fractionation of
complex mixtures of water-soluble substances. J Chromatogr, 9, 44, 1962.
6. Ghebregzabher, M., Rufini, S., Monaldi, B. and Lato, M. Thin-layer chromatography of carbohydrates.
J Chromatogr, 127, 133, 1976.
7. Honda, S. High-performance liquid chromatography of mono- and oligosaccharides. Anal Biochem,
140, 1, 1984.
8. Churms, S. C. Recent developments in the chromatographic analysis of carbohydrates. J Chromatogr,
500, 555, 1990.
9. Townsend, R. R., Hardy, M. R. and Lee, Y. C. Separation of oligosaccharides using high-performance
anion-exchange chromatography with pulsed amperometric detection. Methods Enzymol, 179, 65,
1989.
10. Hardy, M. R., Townsend, R. R. and Lee, Y. C. Monosaccharide analysis of glycoconjugates by anion
exchange chromatography with pulsed amperometric detection. Anal Biochem, 170, 54, 1988.
11. Hermentin, P. et al. The mapping by high-pH anion-exchange chromatography with pulsed ampero-
metric detection and capillary electrophoresis of the carbohydrate moieties of human plasma alpha
1-acid glycoprotein. Anal Biochem, 206, 419, 1992.
12. Hermentin, P. et al. A strategy for the mapping of N-glycans by high-pH anion-exchange chromatog-
raphy with pulsed amperometric detection. Anal Biochem, 203, 281, 1992.
13. Hanko, V. P. and Rohrer, J. S. Determination of carbohydrates, sugar alcohols, and glycols in cell
cultures and fermentation broths using high-performance anion-exchange chromatography with pulsed
amperometric detection. Anal Biochem, 283, 192, 2000.
14. Rohrer, J. S. Analyzing sialic acids using high-performance anion-exchange chromatography with
pulsed amperometric detection. Anal Biochem, 283, 3, 2000.
15. Townsend, R. R. Techniques in Glycobiology, Marcel Dekker, New York, 1997.
16. Olechno, J. D. and Nolan, J. A. Handbook of Capillary Electrophoresis, ed. Landers, J. p. 297, CRC
Press Inc., Boca Raton, FL, 1996.
17. Suzuki, S. and Honda, S. Miniaturization in carbohydrate analysis. Electrophoresis, 24, 3577, 2003.
18. Honda, S. Separation of neutral carbohydrates by capillary electrophoresis. J Chromatogr A, 720, 337,
1996.
19. Suzuki, S. and Honda, S. A tabulated review of capillary electrophoresis of carbohydrates.
20. Novotny, M. V. and Sudor, J. High-performance capillary electrophoresis of glycoconjugates.
21. El Rassi, Z. and Mechref, Y. Recent advances in capillary electrophoresis of carbohydrates.
22. Honda, S., Iwase, S., Makino, A. and Fujiwara, S. Simultaneous determination of reducing monosac-
charides by capillary zone electrophoresis as the borate complexes of N-2-pyridylglycamines. Anal
Biochem, 176, 72, 1989.
23. Honda, S., Yamamoto, K., Suzuki, S., Ueda, M. and Kakehi, K. High-performance capillary zone
electrophoresis of carbohydrates in the presence of alkaline earth metal ions. J Chromatogr A, 588,
327, 1991.
24. Grill, E., Huber, C., Oefner, P., Vorndran, A. and Bonn, G. Capillary zone electrophoresis
of p-aminobenzoic acid derivatives of aldoses, ketoses and uronic acids. Electrophoresis, 14,
1004, 1993.
25. Jackson, P. The analysis of fluorophore-labeled glycans by high-resolution polyacrylamide gel
electrophoresis. Anal Biochem, 216, 243, 1994.
26. Que, A. H. and Novotny, M. V. Separation of neutral saccharide mixtures with capillary electrochro-
matography using hydrophilic monolithic columns. Anal Chem, 74, 5184, 2002.
27. Chiesa, C., Oefner, P. J., Zieske, L. R. and O’Neill, R. A. Micellar electrokinetic chromatography
of monosaccharides derivatized with 1-phenyl-3-methyl-2-pyrazolin-5-one. J Capillary Electrophor,
2, 175, 1995.
28. O’Shea, T. J., Lunte, S. M. and LaCourse, W. R. Detection of carbohydrates by capillary electrophoresis
with pulsed amperometric detection. Anal Chem, 65, 948, 1993.
29. Colon, L. A., Dadoo, R. and Zare, R. N. Determination of carbohydrates by capillary
zone electrophoresis with amperometric detection at a copper microelectrode. Anal Chem, 65,
476, 1993.
30. Hase, S., Hra, S. and Matsushima, Y. Tagging of sugars with a fluorescent compound, 2-aminopyridine.
J Biochem, 85, 217, 1979.
31. Honda, S., Suzuki, S. and Taga, A. Analysis of carbohydrates as 1-phenyl-3-methyl-5-pyrazolone
derivatives by capillary/microchip electrophoresis and capillary electrochromatography. J Pharm
Biomed Anal, 30, 1689, 2003.
32. Mechref, Y. and Novotny, M. V. Miniaturized separation techniques in glycomic investigations.
J Chromatogr B Analyt Technol Biomed Life Sci, 841, 65, 2006.
33. Evangelista, R. A., Liu, M.-S. and Chen, F.-T. A. Characterization of 9-aminopyrene-1,4,6-trisulfonate
derivatized sugars by capillary electrophoresis with laser-induced fluorescence detection. Anal
Biochem, 67, 2239, 1995.
34. Guttman, A., Chen, F. T., Evangelista, R. A. and Cooke, N. High-resolution capillary gel electrophore-
sis of reducing oligosaccharides labeled with 1-aminopyrene-3,6,8-trisulfonate. Anal Biochem, 233,
234, 1996.
35. O’Neill, R. A., Hoff, L. B. and Khan, S. H. Method and apparatus for automated carbohydrate mapping
and sequencing. PCT Int Appl WO, 95, 989, 1995.
36. Guttman, A. and Herrick, S. Effect of the quantity and linkage position of mannose, alpha 1,2 residues
in capillary gel electrophoresis of high-mannose-type oligosaccharides. Anal Biochem, 235, 236, 1996.
37. Khandurina, J. and Guttman, A. High resolution capillary electrophoresis of oligosaccharide structural
isomers. Chromatographia, 62, S37, 2005.
38. Klampfl, C. W. Recent advances in the application of capillary electrophoresis with mass spectrometric
detection. Electrophoresis, 27, 3, 2006.
39. Peter-Katalinic, J. Analysis of glycoconjugates by fast atom bombardment mass spectrometry and
related MS techniques. Mass Spectrom Rev, 13, 77, 1994.
40. Sagi, D., Peter-Katalinic, J., Conradt, H. S. and Nimtz, M. Sequencing of tri- and tetra-antennary
N-glycans containing sialic acid by negative mode ESI QTOF tandem MS. J Am Soc Mass Spectrom,
13, 1138, 2002.
41. Beckedorf, A. I., Schaffer, C., Messner, P. and Peter-Katalinic, J. Mapping and sequencing
of cardiolipins from Geobacillus stearothermophilus NRS 2004/3a by positive and negative ion
nanoESI-QTOF-MS and MS/MS. J Mass Spectrom, 37, 1086, 2002.
42. Bahrke, S. et al. Sequence analysis of chitooligosaccharides by matrix-assisted laser desorption
ionization postsource decay mass spectrometry. Biomacromolecules, 3, 696, 2002.
43. Metelmann, W., Peter-Katalinic, J. and Muthing, J. Gangliosides from human granulocytes: a nano-
ESI QTOF mass spectrometry fucosylation study of low abundance species in complex mixtures. J Am
Soc Mass Spectrom, 12, 964, 2001.
44. Macek, B., Hofsteenge, J. and Peter-Katalinic, J. Direct determination of glycosylation sites in O-
fucosylated glycopeptides using nano-electrospray quadrupole time-of-flight mass spectrometry. Rapid
Commun Mass Spectrom, 15, 771, 2001.
45. Hanisch, F. G., Jovanovic, M. and Peter-Katalinic, J. Glycoprotein identification and localization
of O-glycosylation sites by mass spectrometric analysis of deglycosylated/alkylaminylated peptide
fragments. Anal Biochem, 290, 47, 2001.
46. Viseux, N., Costello, C. E. and Domon, B. Post-source decay mass spectrometry: optimized calibration
procedure and structural characterization of permethylated oligosaccharides. J Mass Spectrom, 34, 364,
1999.
47. Simo, C., Barbas, C. and Cifuentes, A. Capillary electrophoresis-mass spectrometry in food analysis.
48. El Rassi, Z. Carbohydrate Analysis by Modern Chromatography and Electrophoresis, ed. El Rassi,
Z., Elsevier, Amsterdam, 2002.
49. Zaia, J. Mass spectrometry of oligosaccharides. Mass Spectrom Rev, 23, 161, 2004.
50. Novotny, M. V. and Mechref, Y. New hyphenated methodologies in high-sensitivity glycoprotein
analysis. J Sep Sci, 28, 1956, 2005.
51. Tang, H., Mechref, Y. and Novotny, M. V. Automated interpretation of MS/MS spectra of
oligosaccharides. Bioinformatics, 21 (Suppl 1), i431, 2005.
52. Huang, L. and Riggin, R. M. Analysis of nonderivatized neutral and sialylated oligosaccharides by
electrospray mass spectrometry. Anal Chem, 72, 3539, 2000.
53. Que, A. H. et al. Coupling capillary electrochromatography with electrospray Fourier transform mass
spectrometry for characterizing complex oligosaccharide pools. Anal Chem, 75, 1684, 2003.
54. Shen, Y. and Smith, R. D. Proteomics based on high-efficiency capillary separations. Electrophoresis,
23, 3106, 2002.
55. Bergstrom, S. K., Samskog, J. and Markides, K. E. Development of a poly, dimethylsiloxane interface
for on-line capillary column liquid chromatography-capillary electrophoresis coupled to sheathless
electrospray ionization time-of-flight mass spectrometry. Anal Chem, 75, 5461, 2003.
56. Ivanov, A. R., Horvath, C. and Karger, B. L. High-efficiency peptide analysis on monolithic multi-
mode capillary columns: Pressure-assisted capillary electrochromatography/capillary electrophoresis
coupled to UV and electrospray ionization-mass spectrometry. Electrophoresis, 24, 3663, 2003.
57. Guetens, G. et al. Nanotechnology in bio/clinical analysis. J Chromatogr B Biomed Sci Appl, 739, 139,
2000.
58. Lopez-Soto-Yarritu, P., Diez-Masa, J. C., Cifuentes, A. and de Frutos, M. Improved capillary isoelectric
focusing method for recombinant erythropoietin analysis. J Chromatogr A, 968, 221, 2002.
59. Campa, C., Coslovi, A., Flamigni, A. and Rossi, M. Overview on advances in capillary electrophoresis-
mass spectrometry of carbohydrates: a tabulated review. Electrophoresis, 27, 2027, 2006.
60. Kamoda, S. and Kakehi, K. Capillary electrophoresis for the analysis of glycoprotein pharmaceuticals.
61. Hau, J. and Roberts, M. Advantages of pressurization in capillary electrophoresis/electrospray
ionization mass spectrometry. Anal Chem, 71, 3977, 1999.
62. Li, D. T., Sheen, J. F. and Her, G. R. Structural analysis of chromophore-labeled disaccharides by
capillary electrophoresis tandem mass spectrometry using ion trap mass spectrometry. J Am Soc Mass
Spectrom, 11, 292, 2000.
63. Joucla, G., Brando, T., Remaud-Simeon, M., Monsan, P. and Puzo, G. Capillary electrophoresis
analysis of glucooligosaccharide regioisomers. Electrophoresis, 25, 861, 2004.
64. Wolff, M. W., Bazin, H. G. and Lindhardt, R. J. Analysis of fluorescently labeled oligosaccharides
by capillary electrophoresis and electrospray ionization mass spectrometry. Biotechnol Tech, 13, 797,
1999.
65. Charlwood, J. et al. A probe for the versatile analysis and characterization of N-linked oligosaccharides.
Anal Chem, 72, 1453, 2000.
66. Muthing, J. et al. Effects of buffering conditions and culture pH on production rates and glycosylation
of clinical phase I anti-melanoma mouse IgG3 monoclonal antibody R24. Biotechnol Bioeng, 83, 321,
2003.
67. An, H. J., Franza, A. H. and Lebrilla, C. B. Improved capillary electrophoretic separation and mass
spectrometric detection of oligosaccharides. J Chromatogr, 1004, 121, 2003.
68. Min, J. Z., Toyo’oka, T., Kato, M. and Fukushima, T. Resolution of N-linked oligosaccharides in
glycoproteins based upon transglycosylation reaction by CE-TOF-MS. Chem Commun (Camb), 27,
3484, 2005.
69. Gennaro, L. A., Delaney, J., Vouros, P., Harvey, D. J. and Domon, B. Capillary electrophoresis/
electrospray ion trap mass spectrometry for the analysis of negatively charged derivatized and
underivatized glycans. Rapid Commun Mass Spectrom, 16, 192, 2002.
70. Koller, A. et al. Analysis of high-mannose-type oligosaccharides by microliquid chromatography-mass
spectrometry and capillary electrophoresis. Electrophoresis, 25, 2003, 2004.
71. Sandra, K. et al. Characterization of cellobiohydrolase I N-glycans and differentiation of their
phosphorylated isomers by capillary electrophoresis-Q-Trap mass spectrometry. Anal Chem, 76,
5878, 2004.
72. Bindila, L., Peter-Katalinic, J. and Zamfir, A. Sheathless reverse-polarity capillary electrophoresis-
electrospray-mass spectrometry for analysis of underivatized glycoconjugates. Electrophoresis, 26,
1488, 2005.
73. Zamfir, A., Konig, S., Althoff, J. and Peter-Katalinc, J. Capillary electrophoresis and off-line capil-
lary electrophoresis-electrospray ionization quadrupole time-of-flight tandem mass spectrometry of
carbohydrates. J Chromatogr A, 895, 291, 2000.
74. Peter-Katalinic, J. Methods in enzymology: O-glycosylation of proteins. Methods Enzymol, 405,
139, 2005.
75. Baldus, S. E. and Hanisch, F. G. Biochemistry and pathological importance of mucin-associated
antigens in gastrointestinal neoplasia. Adv Cancer Res, 79, 201, 2000.
76. Domon, B., Vath, J. E. and Costello, C. E. Analysis of derivatized ceramides and neutral
glycosphingolipids by high-performance tandem mass spectrometry. Anal Biochem, 184, 151, 1990.
77. Domon, B., Vath, J. E. and Costello, C. E. Analysis of derivatized ceramides and neutral
glycosphingolipids by high-performance tandem mass spectrometry. Anal Biochem, 184, 151, 1990.
78. Domon, B. and Costello, C. E. A systematic nomenclature for carbohydrate fragmentations in
FABMS/MS of glycoconjugates. Glycoconjugate J, 5, 397, 1988.
79. Costello, C. E. and Vath, J. E. Tandem mass spectrometry of glycolipids. Methods Enzymol, 193,
738, 1990.
80. Levery, S. B. Glycosphingolipid structural analysis and glycosphingolipidomics. Methods Enzymol,
405, 300, 2005.
81. Seydel, U., Oikawa, M., Fukase, K., Kusumoto, S. and Brandenburg, K. Intrinsic conformation of
lipid A is responsible for agonistic and antagonistic activity. Eur J Biochem, 267, 3032, 2000.
82. Netea, M. G., van Deuren, M., Kullberg, B. J., Cavaillon, J. M. and Van der Meer, J. W. Does the shape
of lipid A determine the interaction of LPS with Toll-like receptors? Trends Immunol, 23, 135, 2002.
83. Li, J. et al. Mapping bacterial glycolipid complexity using capillary electrophoresis and electrospray
mass spectrometry. Methods Enzymol, 405, 369, 2005.
84. Krylov, S. N. and Dovichi, N. J. Capillary electrophoresis for the analysis of biopolymers. Anal Chem,
72, 111R, 2000.
85. Wang, Z. et al. Structural and serological characterization of the O-chain polysaccharide of Aeromonas
salmonicida strains A449, 80204 and 80204-1. Carbohydr Res, 340, 693, 2005.
86. Li, J., Wang, Z. and Altman, E. In-source fragmentation and analysis of polysaccharides by capillary
electrophoresis/mass spectrometry. Rapid Commun Mass Spectrom, 19, 1305, 2005.
87. Zamfir, A., Vukelic, Z. and Peter-Katalinic, J. A capillary electrophoresis and off-line capillary elec-
trophoresis/electrospray ionization-quadrupole time of flight-tandem mass spectrometry approach for
ganglioside analysis. Electrophoresis, 23, 2894, 2002.
88. Zamfir, A., Seidler, D. G., Kresse, H. and Peter-Katalinic, J. Structural characterization of
chondroitin/dermatan sulfate oligosaccharides from bovine aorta by capillary electrophoresis and
electrospray ionization quadrupole time-of-flight tandem mass spectrometry. Rapid Commun Mass
Spectrom, 16, 2015, 2002.
89. Zamfir, A., Seidler, D. G., Kresse, H. and Peter-Katalinic, J. Structural investigation of chon-
droitin/dermatan sulfate oligosaccharides from human skin fibroblast decorin. Glycobiology, 13, 733,
2003.
90. Ju, D. D., Lai, C. C. and Her, G. R. Analysis of gangliosides by capillary zone electrophoresis and
capillary zone electrophoresis-electrospray mass spectrometry. J Chromatogr A, 779, 195, 1997.
91. Tseng, M. C., Chen, Y. R. and Her, G. R. A low-makeup beveled tip capillary electrophoresis
/electrospray ionization mass spectrometry interface for micellar electrokinetic chromatography and
nonvolatile buffer capillary electrophoresis. Anal Chem, 76, 6306, 2004.
92. Li, J., Cox, A. D., Hood, D., Moxon, E. R. and Richards, J. C. Application of capillary electrophoresis-
electrospray-mass spectrometry to the separation and characterization of isomeric lipopolysaccharides
of Neisseria meningitidis. Electrophoresis, 25, 2017, 2004.
93. Volpi, N. and Maccari, F. Electrophoretic approaches to the analysis of complex polysaccharides.
94. Ofosu, F. A., Danishefsky, I. and Hirsh, J. Heparin and Related Polysaccharides: Structure
and Activities, eds. Ofosu, F. A., Danishefsky, I. and Hirsh, J., N.Y. Acad. Sci., New York,
1989.
95. Mammen, E. F. Seminars in Trombosis and Hemostasis, ed. Mammen, E. F., Theme Medical Pub.
Inc., New York, 1991.
96. Lane, D. A. and Lindahl, U. Heparin: Chemical and Biological Properties, Clinical Applications, eds.
Lane, D. A. and Lindahl, U., Edward Arnold, London, Melbourne, Auckland, 1989.
97. Lane, D. A., Bjork, I. and Lindahl, U. Advances in Experimental Medicine and Biology, eds.
Lane, D. A., Bjork, I. and Lindahl, U., Plenum Press, New York, 1992.
98. Scott, J. E. Dermatan Sulfate Proteoglycans—Chemistry, Biology, Chemical Pathology, ed. Scott, J.E.,
Portland Press, University of Manchester, UK, 1993.
99. Crescenzi, V., Dea, I. C. M., Paoletti, S., Stivala, S.S. and Sutherland, I. W. Biomedical and Biotechno-
logical Advances in Industrial Polysaccharides, eds. Crescenzi, V., Dea, I. C. M., Paoletti, S., Stivala,
S.S. and Sutherland, I. W., Gordon and Breach Sc. Pub, New York, 1989.
100. Wight, T. N. and Mecham, R. P. Biology of Proteoglycans, eds. Wight, T. N. and Mecham, R. P.,
Academic Press, New York, 1987.
101. Kjellen, L. and Lindahl, U. Proteoglycans: structures and interactions. Annu Rev Biochem, 60,
443, 1991.
102. Ruoslahti, E. Structure and biology of proteoglycans. Annu Rev Cell Biol, 4, 229, 1988.
103. Heinegard, D. and Oldberg, A. Structure and biology of cartilage and bone matrix noncollagenous
macromolecules. FASEB J, 3, 2042, 1989.
104. Salmivirta, M. and Jalkanen, M. Syndecan family of cell surface proteoglycans: developmentally
regulated receptors for extracellular effector molecules. Experientia, 51, 863, 1995.
105. Iozzo, R. V. The biology of the small leucine-rich proteoglycans. Functional network of interactive
proteins. J Biol Chem, 274, 18843, 1999.
106. Ruoslahti, E. Proteoglycans in cell regulation. J Biol Chem, 264, 13369, 1989.
107. Sasisekharan, R., Shriver, Z., Venkataraman, G. and Narayanasami, U. Roles of heparan-sulphate
glycosaminoglycans in cancer. Nat Rev Cancer, 2, 521, 2002.
108. Capila, I. and Linhardt, R. J. Heparin–protein interactions. Angew Chem Int Ed Engl, 41,
391, 2002.
109. Fraser, J. R., Laurent, T. C. and Laurent, U. B. Hyaluronan: its nature, distribution, functions and
turnover. J Intern Med, 242, 27, 1997.
110. Laurent, T. C. and Fraser, J. R. Hyaluronan. FASEB J, 6, 2397, 1992.
111. Choi, H. U. and Meyer, K. The structure of keratan sulphates from various sources. Biochem J, 151,
543, 1975.
112. Koketsu, M. and Linhardt, R. J. Electrophoresis for the analysis of acidic oligosaccharides. Anal
Biochem, 283, 136, 2000.
113. Vynios, D. H., Karamanos, N. K. and Tsiganos, C. P. Advances in analysis of glycosaminogly-
cans: its application for the assessment of physiological and pathological states of connective tissues.
114. Mao, W., Thanawiroon, C. and Linhardt, R. J. Capillary electrophoresis for the analysis of
glycosaminoglycans and glycosaminoglycan-derived oligosaccharides. Biomed Chromatogr, 16,
77, 2002.
115. Hoffman, P., Hsu, D. and Mashburn, T. A., Jr. Acrylamide gel electrophoresis of acid glycosamino-
glycans for the study of molecular size distribution. Anal Biochem, 52, 382, 1973.
116. Calabro, A. et al. Fluorophore-assisted carbohydrate electrophoresis, FACE of glycosaminoglycans.
Osteoarthritis Cartilage, 9 (Suppl A), S16, 2001.
117. Volpi, N. Separation of capsular polysaccharide K4 and defructosylated K4 derived disaccha-
rides by high-performance capillary electrophoresis and high-performance liquid chromatography.
118. Volpi, N. Separation of capsular polysaccharide K4 and defructosylated K4 by high-performance

capillary electrophoresis. Electrophoresis, 25, 692, 2004.
119. Volpi, N. Separation of Escherichia coli 055:B5 lipopolysaccharide and detoxified lipopolysaccharide
by high-performance capillary electrophoresis. Electrophoresis, 24, 3097, 2003.
120. Volpi, N. Purification of the Escherichia coli K5 capsular polysaccharide and use of high-performance
capillary electrophoresis to qualitative and quantitative monitor the process. Electrophoresis, 25, 3307,
2004.
121. Kuhn, A. V., Ruttinger, H. H., Neubert, R. H. and Raith, K. Identification of hyaluronic acid oligosac-
charides by direct coupling of capillary electrophoresis with electrospray ion trap mass spectrometry.
Rapid Commun Mass Spectrom, 17, 576, 2003.
122. Zamfir, A., Seidler, D. G., Schonherr, E., Kresse, H. and Peter-Katalinic, J. On-line sheathless
capillary electrophoresis/nanoelectrospray ionization-tandem mass spectrometry for the analysis of
glycosaminoglycan oligosaccharides. Electrophoresis, 25, 2010, 2004.
123. Smyth, W. F. Recent applications of capillary electrophoresis-electrospray ionisation-mass spectrom-
etry in drug analysis. Electrophoresis, 26, 1334, 2005.
124. Schweda, E. K., Li, J., Moxon, E. R. and Richards, J. C. Structural analysis of lipopolysaccharide
oligosaccharide epitopes expressed by non-typeable Haemophilus influenzae strain 176. Carbohydr
Res, 337, 409, 2002.
125. Wang, Q. and Fang, Y. Analysis of sugars in traditional Chinese drugs. J Chromatogr B Analyt Technol
Biomed Life Sci, 812, 309, 2004.
126. Muramatsu, T. Carbohydrate signals in metastasis and prognosis of human carcinomas. Glycobiology,
3, 291, 1993.
127. Fukuda, M. Possible roles of tumor-associated carbohydrate antigens. Cancer Res, 56, 2237, 1996.
128. Zara, J. and Naz, R. K. The role of carbohydrates in mammalian sperm-egg interactions: how important
are carbohydrate epitopes? Front Biosci, 3, D1028, 1998.
129. Takeuchi, M. et al. Relationship between sugar chain structure and biological activity of recombinant
human erythropoietin produced in Chinese hamster ovary cells. Proc Natl Acad Sci USA, 86, 7819,
1989.
130. Egrie, J. C., Dwyer, E., Browne, J. K., Hitz, A. and Lykos, M. A. Darbepoetin alfa has a longer
circulating half-life and greater in vivo potency than recombinant human erythropoietin. Exp Hematol,
31, 290, 2003.
131. Krylov, S. N., Arriaga, E. A., Chan, N. W., Dovichi, N. J. and Palcic, M. M. Metabolic cytometry:
monitoring oligosaccharide biosynthesis in single cells by capillary electrophoresis. Anal Biochem,
283, 133, 2000.
132. Dordal, M. S., Wang, F. F. and Goldwasser, E. The role of carbohydrate in erythropoietin action.
Endocrinology, 116, 2293, 1985.
133. Bietlot, H. P. and Girard, M. Analysis of recombinant human erythropoietin in drug formulations by
high-performance capillary electrophoresis. J Chromatogr A, 759, 177, 1997.
134. Cifuentes, A., Moreno-Arribas, M. V., de Frutos, M. and Diez-Masa, J. C. Capillary isoelectric focusing
of erythropoietin glycoforms and its comparison with flat-bed isoelectric focusing and capillary zone
electrophoresis. J Chromatogr A, 830, 453, 1999.
135. Kinoshita, M. et al. Comparative studies on the analysis of glycosylation heterogeneity of sialic
acid-containing glycoproteins using capillary electrophoresis. J Chromatogr A, 866, 261, 2000.
136. Sanz-Nebot, V., Benavente, F., Valverdu, A., Guzman, N. and Barbossa, J. Separation of recombinant
human erythropoietin glycoforms by capillary electrophoresis using volatile electrolytes. Assessment
of mass spectrometry for the characterization of erythropoietin glycoforms. Eur Pharmacopoeia, 1316,
1123, 2002.
137. Sanz-Nebot, V., Benavente, F., Vallverdu, A., Guzman, N.A. and Barbosa, J. Separation of recombinant
human erythropoietin glycoforms by capillary electrophoresis using volatile electrolytes. Assessment
of mass spectrometry for the characterization of erythropoietin glycoforms. Anal Chem, 75, 5220,
2003.
138. Kamoda, S., Nakano, M., Ishikawa, R., Suzuki, S. and Kakehi, K. Rapid and sensitive screening of
N-glycans as 9-fluorenylmethyl derivatives by high-performance liquid chromatography: a method
which can recover free oligosaccharides after analysis. J Proteome Res, 4, 146, 2005.
139. Kawasaki, N., Ohta, M., Hyuga, S., Hyuga, M. and Hayakawa, T. Application of liquid chro-
matography/mass spectrometry and liquid chromatography with tandem mass spectrometry to the
analysis of the site-specific carbohydrate heterogeneity in erythropoietin. Anal Biochem, 285,
82, 2000.
140. Kawasaki, N. et al. Structural analysis of sulfated N-linked oligosaccharides in erythropoietin:
application of liquid chromatography/mass spectrometry and liquid chromatography with tandem
mass spectrometry to the analysis of the site-specific carbohydrate heterogeneity in erythropoietin.
Glycobiology, 11, 1043, 2001.
141. Okafo, G. et al. A coordinated high-performance liquid chromatographic, capillary electrophoretic,
and mass spectrometric approach for the analysis of oligosaccharide mixtures derivatized with 2-
aminoacridone. Anal Chem, 68, 4424, 1996.
142. Callewaert, N. et al. Noninvasive diagnosis of liver cirrhosis using DNA sequencer-based total serum
protein glycomics. Nat Med, 10, 429, 2004.
143. Kamoda, S. et al. Profiling analysis of oligosaccharides in antibody pharmaceuticals by capillary
electrophoresis. J Chromatogr A, 1050, 211, 2004.
144. Ma, S. and Nashabeh, W. Carbohydrate analysis of a chimeric recombinant monoclonal antibody by
capillary electrophoresis with laser-induced fluorescence detection. Anal Chem, 71, 5185, 1999.
145. Nakajima, K., Oda, Y., Kinoshita, M. and Kakehi, K. Capillary affinity electrophoresis for the screening
of post-translational modification of proteins with carbohydrates. J Proteome Res, 2, 81, 2003.
146. Nakajima, K. et al. Screening method of carbohydrate-binding proteins in biological sources by cap-
illary affinity electrophoresis and its application to determination of Tulipa gesneriana agglutinin in
tulip bulbs. Glycobiology, 14, 793, 2004.
147. Glennie, M. J. and van de Winkel, J. G. Renaissance of cancer therapeutic antibodies. Drug Discov
Today, 8, 503, 2003.
148. Liang, A. et al. Interactions of dextran sulfates with granulocyte colony-stimulating factor and their
effects on leukemia cells. Electrophoresis, 27, 3195, 2006.
149. Liang, A. et al. Structural features in carrageenan that interact with a heparin-binding hematopoietic
growth factor and modulate its biological activity. J Chromatogr B Analyt Technol Biomed Life Sci,
843, 114, 2006.
150. Liang, A. et al. Separation, identification, and interaction of heparin oligosaccharides with granulocyte-
colony stimulating factor using capillary electrophoresis and mass spectrometry. Electrophoresis, 26,
3460, 2005.
151. Duteil, S. et al. Identification of heparin oligosaccharides by direct coupling of capillary
electrophoresis/ionspray-mass spectrometry. Rapid Commun Mass Spectrom, 13, 1889, 1999.
152. Ruiz-Calero, V., Moyano, E., Puignou, L. and Galceran, M. T. Pressure-assisted capillary
electrophoresis-electrospray ion trap mass spectrometry for the analysis of heparin depolymerised
disaccharides. J Chromatogr A, 914, 277, 2001.
153. Wey, A. B., Caslavska, J. and Thormann, W. Analysis of codeine, dihydrocodeine and their glu-
curonides in human urine by electrokinetic capillary immunoassays and capillary electrophoresis-ion
trap mass spectrometry. J Chromatogr A, 895, 133, 2000.
154. Wey, A. B. and Thormann, W. Capillary electrophoresis-electrospray ionization ion trap mass spectrom-
etry for analysis and confirmation testing of morphine and related compounds in urine. J Chromatogr
A, 916, 225, 2001.
155. Wey, A. B. and Thormann, W. Capillary electrophoresis and capillary electrophoresis-ion trap
multiple-stage mass spectrometry for the differentiation and identification of oxycodone and its major
metabolites in human urine. J Chromatogr B Analyt Technol Biomed Life Sci, 770, 191, 2002.
156. Keski-Hynnila, H., Raana, K., Taskinen, J. and Kostiainen, R. Direct analysis of nitrocatechol-type glu-
curonides in urine by capillary electrophoresis-electrospray ionisation mass spectrometry and tandem
mass spectrometry. J Chromatogr B Biomed Sci Appl, 749, 253, 2000.
157. Baldacci, A. and Thormann, W. Analysis of lorazepam and its 30-glucuronide in human urine by
capillary electrophoresis: evidence for the formation of two distinct diastereoisomeric glucuronides.
J Sep Sci, 29, 153, 2006.
158. Kelly, J., Masoud, H., Perry, M. B., Richards, J. C. and Thibault, P. Separation and characterization
of O-deacylated lipooligosaccharides and glycans derived from Moraxella catarrhalis using capillary
electrophoresis-electrospray mass spectrometry and tandem mass spectrometry. Anal Biochem, 233,
15, 1996.
159. Auriola, S., Thibault, P., Sadovskaya, I. and Altman, E. Enhancement of sample loadings for the
analysis of oligosaccharides isolated from Pseudomonas aeruginosa using transient isotachophore-
sis and capillary zone electrophoresis—electrospray—mass spectrometry. Electrophoresis, 19,
2665, 1998.
160. Ludwiczak, P., Brando, T., Monsarrat, B. and Puzo, G. Structural characterization of Mycobacterium
tuberculosis lipoarabinomannans by the combination of capillary electrophoresis and matrix-assisted
laser desorption/ionization time-of-flight mass spectrometry. Anal Chem, 73, 2323, 2001.
161. Tzianabos, A., Wang, J. Y. and Kasper, D. L. Biological chemistry of immunomodulation by
zwitterionic polysaccharides. Carbohydr Res, 338, 2531, 2003.
162. Ishii, T. Structure and functions of feruloylated polysaccharides. Plant Sci, 127, 111, 1997.
163. Tang, Y.-J. and Zhong, J.-J. Exopolysaccharide biosynthesis and related enzyme activities of the
medicinal fungus, Ganoderma lucidum, grown on lactose in a bioreactor. Biotechnol Lett, 24,
1023, 2002.
164. Ho, C. Y., Lo, T. W., Leung, K. N., Fung, K. P. and Choy, Y. M. The immunostimulating activi-
ties of anti-tumor polysaccharide from K1 capsular, polysaccharide antigen isolated from Klebsiella
pneumoniae. Immunopharmacology, 46, 1, 2000.
165. Wong, C. K., Leung, K. N., Fung, K. P. and Choy, Y. M. The immunostimulating activities of anti-tumor
polysaccharides from Pseudostellaria heterophylla. Immunopharmacology, 28, 47, 1994.
166. Sonoda, Y. et al. Stimulation of interleukin-8 production by acidic polysaccharides from the root of
Panax ginseng. Immunopharmacology, 38, 287, 1998.
167. Lu, R. et al. Specific biological activities of Chinese lacquer polysaccharides. Carbohydr Polymers
43, 47, 2000.
168. Yang, J. and Du, Y. Chemical modification, characterization and bioactivity of Chinese lacquer polysac-
charides from lac tree Rhus vernicifera against leukopenia induced by cyclophosphamide. Carbohydr
Polymers, 52, 405, 2003.
169. Mitsuya, H., Yarchoan, R. and Broder, S. Molecular targets forAIDS therapy. Science, 249, 1533, 1990.
170. Yoshida, T. Synthesis of polysaccharides having specific biological activities. Prog Polymer Sci, 26,
379, 2001.
171. Xu, H.-X., Lee, S. H. S., Lee, S. F., White, R. L. and Blay, J. Isolation and characterization of an
anti-HSV polysaccharide from Prunella vulgaris. Antiviral Res, 44, 43, 1999.
172. Wang, Y. et al. Protective effect of Fructus Lycii polysaccharides against time and hyperthermia-
induced damage in cultured seminiferous epithelium. J Ethnopharmacol, 82, 169, 2002.
173. Hu, Q., Zhou, T., Hu, G. and Fang, Y. Determination of sugars in Chinese traditional drugs by CE with
amperometric detection. J Pharm Biomed Anal, 30, 1047, 2002.
174. Wang, Q., Ding, F., Zhu, N., He, P. and Fang, Y. Determination of the compositions of polysac-
charides from Chinese herbs by capillary zone electrophoresis with amperometric detection. Biomed
Chromatogr, 17, 483, 2003.
175. Wang, Q., Yu, H., Zong, J., He, P. and Fang, Y. Determination of the composition of Chinese ligus-
trum lucidum polysaccharide by capillary zone electrophoresis with amperometric detection. J Pharm
Biomed Anal, 31, 473, 2003.
176. Wang, J., Chatrathi, M. P., Tian, B. and Polsky, R. Microfabricated electrophoresis chips for simul-
taneous bioassays of glucose, uric acid, ascorbic acid, and acetaminophen. Anal Chem, 72, 2514,
2000.
177. Wang, J., Chatrathi, M. P. and Tian, B. Microseparation chips for performing multienzymatic dehy-
drogenase/oxidase assays: simultaneous electrochemical measurement of ethanol and glucose. Anal
Chem, 73, 1296, 2001.
178. Tseng, K., Liu, J., Lebrilla, C. B., Collins, S. D. and Smith, R. L. Fabrication and design of open
microchannels for capillary electrophoresis separations and matrix-assisted laser/desorption mass
spectroscopy analysis. Proc SPIE Int Soc Opt Eng, 3606, 137, 1999.
179. Larsson, M., Sundberg, R. and Folestad, S. On-line capillary electrophoresis with mass spectrometry
detection for the analysis of carbohydrates after derivatization with 8-aminonaphthalene-1,3,6-
trisulfonic acid. J Chromatogr A, 934, 75, 2001.
180. Klampfl, C. W. and Buchberger, W. Determination of carbohydrates by capillary electrophoresis with

electrospray-mass spectrometric detection. Electrophoresis, 22, 2737, 2001.
181. Suomi, J., Wiedmer, S. K., Jussila, M. and Riekkola, M. L. Determination of iridoid glycosides
by micellar electrokinetic capillary chromatography-mass spectrometry with use of the partial filling
technique. Electrophoresis, 22, 2580, 2001.
182. Craston, D. H. et al. Determination of gellan gum by capillary electrophoresis and CE-MS. Food Chem,
73, 103, 2001.
183. Hagberg, J. Analysis of low-molecular-mass organic acids using capillary zone electrophoresis-
electrospray ionization mass spectrometry. J Chromatogr A, 988, 127, 2003.
184. Bianco, G., Schmitt-Kopplin, P., De Benedetto, G., Kettrup, A. and Cataldi, T. R. Determination of gly-
coalkaloids and relative aglycones by nonaqueous capillary electrophoresis coupled with electrospray
ionization-ion trap mass spectrometry. Electrophoresis, 23, 2904, 2002.
185. Bringmann, G. et al. Analysis of the glucosinolate pattern of Arabidopsis thaliana seeds by capil-
lary zone electrophoresis coupled to electrospray ionization-mass spectrometry. Electrophoresis, 26,
1513, 2005.
186. Kirk, T. K. and Cullen, D. Environmentally Friendly Technologies for the Pulp and Paper Industry,
eds. Young, R. A. and Akhtar, M., John Wiley & Sons, New York, 1998.
187. Viikari, L. et al. Bioconversion of Forest and Agricultural Plant Residues, ed. Saddler, J. N., CAB,
Oxford, 1993.
188. Saha, B. C. Hemicellulose bioconversion. J Ind Microbiol Biotechnol, 30, 279, 2003.
189. Wyman, C. E. Potential synergies and challenges in refining cellulosic biomass to fuels, chemicals and
power. Biotechnol Prog, 19, 254, 2003.
190. Wyman, C. E. Ethanol from lignocellulosic biomass: technology economics, and opportunities.
Bioresour Technol, 50, 3, 1994.
191. Felix, C. R. and Ljungdahl, L. G. The cellulosome: the exocellular organelle of Clostridium. Annu Rev
Microbiol, 47, 791, 1993.
192. Paulus, A. and Klokow-Beck, A. Analysis of Carbohydrates by Capillary Electrophoresis, Vieweg
Publishers, Wiesbaden, 1999.
193. Mechref, Y. and Novotny, M. V. Structural investigations of glycoconjugates at high sensitivity. Chem
Rev, 102, 321, 2002.
194. Chiesa, C. and Horvath, C. Capillary zone electrophoresis of malto-oligosaccharides derivatized with
8-aminonaphthalene-1,3,6-trisulfonic acid. J Chromatogr, 645, 337, 1993.
195. Guttman, A. High-resolution carbohydrate profiling by capillary gel electrophoresis. Nature, 380, 461,
1996.
196. Morell, M. K., Samuel, M. S. and O’Shea, M. G. Analysis of starch structure using fluorophore-assisted
carbohydrate electrophoresis. Electrophoresis, 19, 2603, 1998.
197. Callewaert, N., Geysens, S., Molemans, F. and Contreras, R. Ultrasensitive profiling and sequenc-
ing of N-linked oligosaccharides using standard DNA-sequencing equipment. Glycobiology, 11,
275, 2001.
198. Khandurina, J., Blum, D. L., Stege, J. T. and Guttman, A. Automated carbohydrate profiling by
capillary electrophoresis: a bioindustrial approach. Electrophoresis, 25, 2326, 2004.
199. Stefansson, M. and Novotny, M. Electrophoretic resolution of monosaccharide enantiomers in borate-
oligosaccharide complexation media. J Am Chem Soc, 115, 11573, 1993.
200. Guttman, A. Analysis of monosaccharide composition by capillary electrophoresis. J Chromatogr A,
763, 271, 1997.
201. Khandurina, J., Anderson, A. A., Olson, N. A., Stege, J. T. and Guttman, A. Large-scale carbo-
hydrate analysis by capillary array electrophoresis: part 2. Data normalization and quantification.
202. Khandurina, J., Olson, N. A., Anderson, A. A., Gray, K. A. and Guttman, A. Large-scale carbohydrate
analysis by capillary array electrophoresis: part 1. Separation and scale-up. Electrophoresis, 25, 3117,
2004.
203. Schwarz, M. A., Galliker, B., Fluri, K., Kappes, T. and Hauser, P. C. A two-electrode configuration
for simplified amperometric detection in a microfabricated electrophoretic separation device. Analyst,
126, 147, 2001.
204. Hebert, N. E., Kuhr, W. G. and Brazill, S. A. Microchip capillary electrophoresis coupled to sinusoidal
voltammetry for the detection of native carbohydrates. Electrophoresis, 23, 3750, 2002.
205. Fanguy, J. C. and Henry, C. S. Pulsed amperometric detection of carbohydrates on an electrophoretic
microchip. Analyst, 127, 1021, 2002.
206. Fu, C.-G. and Fang, Z.-L. Combination of flow injection with capillary electrophoresis. Anal Chim
Acta 422, 71, 2000.
207. Burggraf, N., Krattiger, B., de Mello, A. J., de Rooij, N. F. and Manz, A. Holographic refractive index
detector for application in microchip-based separation systems. Analyst, 123, 1443, 1998.
208. Suzuki, S., Ishida, Y., Arai, A., Nakanishi, H. and Honda, S. High-speed electrophoretic analysis of
1-phenyl-3-methyl-5-pyrazolone derivatives of monosaccharides on a quartz microchip with whole-
channel UV detection. Electrophoresis, 24, 3828, 2003.
209. Dang, F., Zhang, L., Hagiwara, H., Mishina, Y. and Baba, Y. Ultrafast analysis of oligosaccharides on
microchip with light-emitting diode confocal fluorescence detection. Electrophoresis, 24, 714, 2003.
210. Dang, F., Zhang, L., Jabasini, M., Kaji, N. and Baba, Y. Characterization of electrophoretic behavior
of sugar isomers by microchip electrophoresis coupled with videomicroscopy. Anal Chem, 75, 2433,
2003.
211. Suzuki, S., Shimotsu, N., Honda, S., Arai, A. and Nakanishi, H. Rapid analysis of amino sugars by
microchip electrophoresis with laser-induced fluorescence detection. Electrophoresis, 22, 4023, 2001.
212. Bindila, L. et al. Off-line capillary electrphoresis/fully automated nanoelectrospray chip quadrupole
time-of-flight mass spectrometry and tandem mass spectrometry for glycoconjugate analysis. J Mass
Spectrom, 39, 1190, 204.
213. Callewaert, N. et al. Total serum protein N-glycome profiling on a capillary electrophoresis-
microfluidics platform. Electrophoresis, 25, 3128, 2004.
214. Ninonuevo, M. et al. Nanoliquid chromatography-mass spectrometry of oligosaccharides employing
graphitized carbon chromatography on microchip with a high-accuracy mass analyzer. Electrophoresis,
26, 3641, 2005.
215. Kang, P., Mechref, Y., Klouckova, I. and Novotny, M. V. Solid-phase permethylation of glycans for
mass spectrometric analysis. Rapid Commun Mass Spectrom, 19, 3421, 2005.
216. Khandurina, J. and Guttman, A. Microscale separation and analysis. Curr Opin Chem Biol, 7,
595, 2003.
217. Khandurina, J. and Guttman, A. Bioanalysis in microfluidic devices. J Chromatogr A, 943, 159, 2002.
8 The Coupling of Capillary
Electrophoresis and Mass
Spectrometry in Proteomics
Haleem J. Issaq and Timothy D. Veenstra
CONTENTS
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295

8.2 Online CE–ESI/MS System Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
8.3 Sheath versus Sheathless Interfaces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
8.4 Sheathless Design Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
8.5 Proteomic Analysis Using CE-MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302
8.1 INTRODUCTION
Modern science is firmly entrenched in the “omics” era. Spurred by the human genome project,
omics-thinking quickly spread to ribonucleic acid (RNA), proteins, and now metabolites. Although
different technologies are required for each specific omic endeavor, they all have similar goals of
attempting to garner as much information about a specific class of biomolecules as possible. Another
thing that each has in common that often goes overlooked is the need for separation. While very
complex mixtures of deoxyribonucleic acid (DNA), RNA, protein, or metabolites are introduced to
the appropriate analytical instrumentation, the technology still measures each basic unit (whether
it be a base, transcript, or peptide) individually. For example, even though entire transcriptomes
are introduced to array platform, the RNA transcripts must be separated via annealing to their
complementary strand for a quantitative signal to be deciphered for each. While proteomics, one of
the foci of this chapter, is arguably driven by the development of mass spectrometry (MS) technology,
separation technologies have been just as instrumental in its development. If proteomics was limited
to direct infusion techniques, we would still be floundering in the detection of only the highest
abundant proteins present within biological samples.
When integrating a separation technique into a MS-based proteomics study many different aspects
need to be considered. The first consideration is if an online or an off-line separation of the proteome
sample will be integrated into the research plan. There are a large number of separation techniques
available to do off-line.1 These include many different chromatography [e.g., strong/weak cation
exchange (S/WCX), strong/weak anion exchange (S/WAX), reversed-phase (RP), size exclusion,
etc.] and electrophoretic gel and capillary electrophoresis (CE) [e.g., capillary zone electrophore-
sis (CZE), isoelectric focusing (IEF), isotachophoresis (ITC), etc.] methods. Off-line separations
are simpler, easier to perform, and require no instrument modification. In addition, a simple sol-
vent exchange step can be administered to each fraction to introduce the correct conditions for
downstream MS analysis. RP columns packed with C18 particles are most commonly used to
295
remove salts and other MS incompatible materials from fractions collected via off-line separations.
The next consideration is the amount of material available for fractionation. This consideration
is where chromatographic methods have a decided advantage over gel-based electrophoretic tech-
niques. Proteins in a proteomic sample are fractionated off-line mainly to simplify the mixture prior to
high-performance liquid chromatography/mass spectrometry (HPLC/MS) analysis and are normally
done when a significant amount (50–200 µg) of protein is available. Although chromatography
columns have the capacity to handle such samples sizes, typical gel electrophoretic methods are
unsuitable for these amounts of materials.2 This aspect is the primary reason that SCX has become
the most popular choice of off-line liquid-based separations for MS-based proteomic studies. The
final design aspect that needs to be considered is the choice of online separation that will be used
with direct MS analysis. The most important consideration for this part of the study is making
the solvent conditions compatible with electrospray ionization (ESI).3 ESI requires the solvent to
be volatile, acidic, and contain only low levels of salts and buffers. ESI is almost intolerable to
detergents.
The most popular type of separation technique to which MS has been directly coupled is reversed-
phase liquid chromatography (RPLC). This chromatography has all the necessary attributes that
make it an excellent choice for online MS analysis. What is not optimal, however, is its speed
and resolution. RP columns require re-equilibration between analyses when gradient mobile phases
are used, leading to an almost 25% downtime when using this type of separation. While many
laboratories have implemented column-switching so that one column is being used for online MS
analysis while the other is re-equilibrating, this technology is not widely used and has issues related
to repositioning of the column within the ESI source. While the resolution of RPLC is quite good,
the duty cycles of tandem MS instruments is so short that even higher resolution separations could
provide a greater solution for maximizing proteome coverage. It is in these two areas that CE has
distinct advantages.
Capillary electrophoresis is an excellent microseparation technique that has been used for the
separation of a wide diversity of different molecules.4 Its separation capabilities extend to ions, small
molecules (such as amino acids), and large biomolecules (such as peptides, proteins, and nucleic
acids). Indeed the human genome project owes its success, in part, to the use of CE for the separation
of DNA bases. In the past, CE has been combined with detection devices such as ultraviolet (UV)
and laser-induced fluorescence (LIF) spectrophotometers. The detection of the separated analytes is
carried out on column by etching the capillary. Unfortunately, UV detection lacks sensitivity and
not every compound of interest will absorb in the UV region of the spectrum. Detection using LIF is
sensitive, however, the analytes of interest may require derivatization with a fluorescent tag or have
an aromatic amino acid in their structure (e.g., proteins and peptides). An advantage of MS detection
that neither UV nor LIF detection provides is the information necessary to directly determine the
structure of the detected analyte(s).
Mass spectrometry, however, has high sensitivity for the detection of a wide variety of analytes.
When operated in a multistage, MS mode can provide information to assist in the determination of
the structural formula of the molecules of interest. A marriage of CE and MS would potentially result
in an extremely powerful tool for the separation, identification, and characterization of a wide range
of molecules. Unfortunately, the online interfacing of CE with MS is not trivial and presents several
technological challenges not present with CE-UV or CE-LIF. Some major issues are the requirements
of both the CE continuous electrical circuit for electrophoretic separation and the mass spectrometer
electrical contact for efficient ESI. Running buffers normally used in CE, such as sodium phosphate
or borate that have low volatility, are not compatible with ESI-MS. These buffers need to be replaced
with volatile buffers such as ammonium formate or acetate. This substitution, however, may have a
detrimental impact on the quality of the separation.
The resolution afforded with CE is a good match for the duty cycle times of present day tandem MS
instruments, particularly in the area of proteomics. In the past, the speed at which mass spectrometers
selected, isolated, fragmented, and detected peptide ions did not warrant extremely high resolution
The Coupling of Capillary Electrophoresis and Mass Spectrometry in Proteomics 297
separations because some analytes would not be present long enough in the MS source to be surveyed
by the instrument. Nowadays, with the speed of mass spectrometers such as the linear-ion traps,
peptides that are present for only a few seconds (i.e., <10) still have a high probability of being
selected for tandem MS analysis. The biggest drawback in using CE online with MS is the amount
of material that can be loaded onto the capillary. Although MS is a sensitive detection technique,
it still has its limits. Sample sizes in the range of 1–10 µg are typically loaded onto a capillary
RP column for subsequent MS analysis. Unfortunately, CE capillaries have a very limited volume
and can only accommodate sample amount in the nanogram (ng) range. Fortunately, the use of
on-column preconcentration devices has enabled the loading of greater amounts of material onto CE
capillaries.5
8.2 ONLINE CE–ESI/MS SYSTEM REQUIREMENTS

An online CE–ESI/MS system is made up of a CE instrument, a mass spectrometer, an interface that
will allow the transfer of analytes from the CE column outlet to the MS source, a closed CE electrical
circuit for electrophoretic separation of the analyte mixture, and a closed electrical circuit for the
generation of continuous, stable and uniform fine spray stream that affords sensitive MS detection.
The challenge in coupling CE directly online with ESI-MS is achieving electrical continuity that
allows uninterrupted operation of both systems, without affecting the quality of the CE separation or
the ESI efficiency (which is directly related to the detection sensitivity). The coupling of CE to MS
has been a challenging problem because an ideal interface is one that is constructed in such a way
that the CE separation column and the spray tip form a single continuous unit in order to eliminate
any dead volume that may lead to diffusion and affect the quality of separation. In addition, the
design should preserve the electric circuits of both the CE system and the spray tip. Also, it would
be advantageous if no external solvent is added to the system (sheath liquid) that dilutes the analyte
concentration and affects the detection sensitivity.
The most popular choice for coupling CE with ESI/MS is the coaxial sheath flow interface. Coax-
ial sheath flow was introduced as the first CE-MS interface in 1987.6 This coupling device uses three
concentric capillaries in which the CE capillary (used for the actual analyte separation) is innermost
and protrudes into the ESI source region. A stainless steel capillary (the ESI needle) is placed around
the CE capillary and delivers a sheath liquid. The sheath liquid is required to complete the electrical
circuit necessary for both CE separation and ESI. The sheath liquid comes into contact with both
the CE buffer and ESI needle. The sheath liquid is usually a mixture of a volatile acid, H2 O, and an
organic modifier, such as methanol (CH3 OH) or acetonitrile (ACN). The sheath liquid may also con-
tain an electrolyte since it also acts as the outlet buffer for the CE separation process. The electrolyte
concentration must be carefully selected, as both separation and ionization efficiency must be con-
sidered. A high electrolyte concentration will provide good separation efficiency but will negatively
impact ESI efficiency. The sheath liquid is mixed with the CE buffer at the very tip of the capillary
within the ESI source region. It is critical to keep the mixing volume as small as possible to maxi-
mize ES stability and sensitivity. Therefore, the distance that the CE capillary extends past the sheath
liquid capillary is very important. While the correct distance may need to be determined through
experimentation, it is generally in the range of 0.1–0.5 mm. Both of these capillaries are placed
within a third concentric stainless steel tube that delivers a gas flow, which is necessary for effective
desorption of the electrosprayed ions as well as providing the necessary cooling of the CE capillary.
8.3 SHEATH VERSUS SHEATHLESS INTERFACES

There are many advantages of sheathless interfaces compared to those that require a sheath flow.7,8
The main difference between sheath and sheathless interface designs is that sheathless interface does
not require the external flow of a coaxial sheath liquid to establish electrical contact with the CE
effluent to facilitate the ESI process. Introduction of a sheath liquid dilutes the analytes and results
in poorer detection limits. Indeed, a comparable sheathless interface can give an order of magnitude
greater sensitivity than its counterpart that requires a sheath liquid. One of the major weaknesses
of CE is that although the mass detection levels are quite high the amount of sample that may be
injected onto the capillary is low. Dilution of the analytes by sheath liquid only serves to compound
this deficiency. The construction of an interface with sheath flow is also more complicated, requiring
coupling of different capillaries on connectors. In principle, the sheathless interface is the best design
for coupling CE online with ESI-MS, since the flow rates required for each technique are compatible
and no dilution of the analytes occurs during separation and elution, thereby maximizing the signal
recorded by the mass spectrometer.
8.4 SHEATHLESS DESIGN STRATEGIES

Several different types of sheathless designs have been introduced that satisfy the requirement of
closing the CE separation capillary circuit while simultaneously providing an electrical potential to
the spray tip for ESI. These include the use of a single capillary, two capillaries, and three capillaries.5
Many methods for interfacing CE with MS using a single-capillary format whereby uninterrupted
electrical contact required for CE and ESI was established by different means have been introduced.
An example of such is shown in Figure 8.1a. The first single capillary-sheathless interface was
introduced by Wahl et al.9 who coated the capillary outlet with a conductive metal (i.e., silver
deposition) at the cathode end. This coating served to define the CE field strength along the capillary
column and provide the necessary ESI potential. One of the best features of this design is that it did
not interfere with the CE separation process. Since this work, a number of studies reported using
similar strategies for coating the capillary’s cathodic end with metal, polymer, or carbon.10,11 The
biggest drawback to these coating methods was their lack of mechanical stability. The coatings would
slowly disintegrate and each CE capillary was only useful for a few hours to days. The stability of the
metal coating by precoating the capillary with materials such as silanes, silicon oxide, or chromium
improved the length of the capillary’s lifetime.
(a) Platinum wire
Epoxy
Capillary outlet
(b)
Buffer Copper wire
reservoir
Capillary outlet
Dialysis tubing
FIGURE 8.1 (a) Schematic diagram of a single capillary design for CE-MS. (From Cao, P. and Moini, M.,
J. Am. Soc. Mass Spectrom., 8, 561, 1997.) In this design, a small hole is drilled into the capillary through
which a platinum wire is inserted. (b) Schematic of a two capillary design for CE-MS in which two pieces of
fused-silica capillaries are connected via a section of dialysis tubing. (From Severs, J. C. and Smith, R. D.,
Anal. Chem., 69, 2154, 1997.) This junction is immersed into the CE buffer and into which a copper electrode
is inserted.
Other groups have taken different approaches that require directly inserting a metal wire into
the CE capillary. For example, Cao and Moini10 constructed a sheathless interface by inserting a
platinum wire (25 µm diameter by 0.2–0.3 cm in length), through a pinhole that was drilled into a
150-µm o.d. fused-silica capillary. The pinhole was positioned approximately 2 cm from the outlet
of the capillary. Epoxy was used to maintain the position of the wire as well as seal the capillary.
The platinum wire electrode serves as both the low voltage electrode of the CE electric circuit and
as a connection for the ESI voltage. This design proved to be stable over the maximum time studied
(i.e., 50 min), however, such capillaries are difficult to construct (requiring a dental drill to create a
small opening) and the presence of the wire may generate turbulent flow that affects the resolution of
the CE separation. The laboratories of Wahl and Smith12 developed a similar “micro-hole” method,
however, in this case an interface whereby the electrical connection was made was constructed by
drilling a small hole 2 cm from the end of the separation capillary. The micro-hole was then sealed
with conductive gold epoxy. While this method allows for the necessary electrical conductance, the
design is cumbersome and the proper capillaries are not easy to construct.
Our laboratory developed a simple sheathless interface for CE-MS in which the separation
column, an electrical porous junction, and the spray tip are integrated within a single piece of a
fused-silica capillary, as illustrated in Figure 8.2.13 The electrical potential to induce ESI is introduced
through a porous junction across a 3–4 mm length of fused silica. A 3–4 mm section of the polyimide
coating was removed from the capillary approximately 5 cm from end of the spray tip. Electrical
conductance between the solute within the capillary and the anode buffer was achieved by etching
the exposed section using hydrofluoric acid (HF). This etching reduced the outside diameter of the
capillary at this point without affecting the inner diameter. The HF etching requires approximately
6 h and reduces the capillary wall to about 15–20 µm. Since the porous junction created by HF
etching is fragile, the capillary must be secured inside the reservoir protecting it from breaking.
The reservoir contains the buffer for closing the CE circuit and providing the voltage necessary
for ESI. This design proved to be quite rugged as this group reported continuous operation of
the same column for over a 2-week period with no evidence of deterioration in separation or ESI
performance.
Fused
Extraction silica
disk sleeve
Epoxy ESI
glue Power supply Plexiglas reservoir
CE capillary
MS
Plexiglas slide
Spray tip
CE
power
supply
CE buffer
Etched segment
FIGURE 8.2 Schematic diagram of a CE-MS design that incorporates etching of the capillary wall to enable
electrical conductance and a solid-phase extraction disk for pre-concentration of the sample prior to CE-MS
analysis. (From Janini, G. M., et al., Anal. Chem., 76, 1615, 2003. With permission.)
The same group followed up on this initial design by adding the capability for on-column sample
enrichment into the sheathless interface.4 A small solid-phase extraction cartridge, made of RP
material, was attached to the CE capillary near the point of injection. Optimization of the proper
capillary diameters showed that they could achieve a mass limit of detection of 500 amol for CE-
MS/MS analysis of a standard peptide using a 20-µm i.d. capillary. This design brings to reality a
true zero dead-volume sheathless CE—MS interface with the ability to preconcentrate the sample
within the same capillary.
Another strategy for coupling CE with MS is a sheathless interface in which two pieces of
capillary are used. In this method the CE separation capillary is connected to a short spray tip (which
functions as the ESI needle) via a sleeve. The sleeve can be prepared from microdialysis tubing,
stainless steel tubing, or a micro-tee junction. Severs and Smith14 developed a separation capillary
that is connected to a 2-cm-long ESI emitter capillary via a 1.5 cm length of dialysis tubing. Epoxy
is then applied to maintain the connection between the tubing and capillary pieces, as illustrated
in Figure 8.1b. The manufactured capillary is then inserted through a 250 µL Eppendorf pipet tip
containing an electrolyte identical to that employed for the CE separation. Connection of the pipet tip
to an x-y-z stage allows the position of the spray tip relative to the MS orifice for maximum detection
sensitivity. A copper wire is inserted in the electrolyte reservoir and connected to a high-voltage
power supply to provide the necessary voltage for the CE separation and ESI.
The group of Tong et al.15 developed a sheathless liquid–metal junction interface for CE-MS
by removing approximately 2–3 mm length of polyimide coating from a coated capillary (5–10 cm
long), and exposing the fused-silica tip to 48–51% HF. This resulted in etching of the exposed fused
silica and the polyimide coating was further burned to expose approximately 3 mm of the tapered tip.
The butt ends of the tip and CE capillary were then carefully polished flat. A liquid–metal junction
was then created with a polyether ether ketone (PEEK) micro-tee. The side channel of the tee was
enlarged to accommodate a 2.5 cm length of 0.06 cm diameter gold wire. The electrospray tip and the
separation capillary were carefully butted together along the channel of the micro-tee and connected
via microfittings. The gold wire was inserted into the side channel of the micro-tee to supply the
electrical connection to supply the required voltage for ESI.
Other researchers have also reported on the design of three-piece capillary/spray tips in which
the electrical contact is established through a porous glass joint that is connected to the CE capillary
using polytetrafluoroethylene (PTFE) sleeves on one side and the spray tip on the other. A different
three-piece designed based on the one previously described, used a porous glass joint. In this design,
the polyimide layer was burned off a 2–3 mm piece of a 3-cm-long fused-silica capillary.16 This
exposed portion was etched with HF to reduce the thickness of the capillary wall to <20 µm. This
porous glass joint was inserted into a piece of PTFE, in which a small notch was cut through its
center. The diameter of this notch matched that of the outer diameter of the capillary. The porous
glass joint was coupled to the CE column with the aid of a PTFE sleeve, and the ESI emitter was
connected to the other side of the porous glass joint. To conduct online CE separations the porous
fused silica was immersed in 1% acetic acid and voltage was applied to the acidic solution to drive
both the CE separation and ESI.
8.5 PROTEOMIC ANALYSIS USING CE-MS

Currently there is no greater focus within proteomics than the discovery of biomarkers of various
disease states. The number of proteomic techniques and strategies that are being utilized in an attempt
to find novel biomarkers is a veritable “who’s who” list of available methods. The number of different
methods being attempted is related to a number of different factors including the enormity of the
experimental question, complexity of human biofluids, and the lack of success achieved using one
single design. While the primary separation technologies that have been combined with MS for
the discovery of biomarkers are two-dimensional polyacrylamide gel electrophoresis (2D-PAGE)
two-dimensional polyacrylamide gel electrophoresis, and RPLC, CE has also played a role in this
research area. In a study attempting to find indicators of graft vs. host disease (GvHD) in patients
that had received allogeneic hematopoietic stem cell transplants (HSCT), investigators examined the
CE-MS profiles of urine obtained from 40 HSCT patients and 5 patients with sepsis.17 They were
able to find 16 differentially urine-excreted peptides that indicated the presence of GvHD. In a cross
validation study using the pattern generated by these markers, they correctly diagnose GvHD with
a sensitivity and specificity of 82% and 100%, respectively.
While the above study utilized urine, probably no biofluid has attracted the attention of the
proteomic biomarker discovery field as serum and plasma. Although the analysis of serum by RPLC-
MS has almost become routine, such is not the case for CE-MS. In a proof of principle study to show
the efficacy of using CE-MS to find differentially abundant proteins in serum, the group of Sassi
et al.18 analyzed groups of sera that were spiked with different concentrations of known standard
peptides. The groups of sera were analyzed by CE-MS and the standard peptides were successfully
identified as being differentially abundant with a success rate of 95%.
These first promising data using CE-MS to detect biomarkers in serum/plasma was a prospec-
tive, randomized, open-label trial to assess the effect of vitamin C deficiency on oxidative stress and
analyze inflammatory markers in hemodialysis patients.19 The CE-MS analysis of plasma obtained
from dialysis patients and those with normal renal functions showed more than 30 polypeptides
that were significantly different in abundance between the two groups. This study showed that
oral vitamin C supplementation did indeed have an affect on the plasma proteome of dialysis
patients.
Another important biofluid that has been utilized in CE-MS biomarker discovery studies is
cerebrospinal fluid (CSF). Because of the blood–brain barrier, blood has limited diagnostic potential
for nervous system related diseases, such as Alzheimer’s disease (AD). CSF is a clear, colorless
liquid that provides mechanical protection for the brain. The proteomic content of CSF is similar to
that of blood, in that 70% of CSF’s protein content consists of isoforms of albumin, transferrin, and
immunoglobulins. A study conducted by Wetterhall et al.20 described a method for analyzing CSF, by
first digesting the proteome into tryptic peptides and then analyzing these using CE coupled directly
online with FT-ICR MS. In the first step, CSF was analyzed by CE-MS to establish the ability to
identify peptides/proteins within this biofluid. Using the mass measurement accuracy capabilities of
FTICR-MS, 30 proteins were identified with mass measurement errors of less than 5 parts per million
(ppm). In an alternative approach, Wittke et al.21 obviated the tryptic digestion step, however, they
depleted the high abundance proteins from CSF prior to its analysis by CE-MS. They were able to
detect 450 different proteins in the molecular mass range of 800–15,000 Da. They proceeded to use
this experimental setup to analyze protein patterns within CSF from 4 healthy donors, 8 patients
with AD and 7 with schizophrenia. They were able to find peaks that were capable of differentiating
AD and schizophrenic patients, but were unable to differentiate these two groups from the controls.
While the number of samples analyzed in this study was low, it did show the potential of direct
CE-MS as a diagnostic tool for brain-related disorders.
Current literature suggests CE-MS is potentially a powerful tool for the detection of biomarkers
in a variety of biofluids. It has the ability to reliably detect peptides within biological samples at a rate
far greater than conventional RPLC-MS. There are, however, several challenges that remain. While
the ability of CE-MS to provide information on hundreds of peptides in a single sample has been
established, and the diagnostic ability of many of patterns of these peptides has been shown, the link
to potential physiological function of these markers is still lacking. This deficiency is primarily due
to the inability to routinely identify the sequence and therefore the identity of these peptide peaks.
At this point, the diagnostic features are simply; just features and not known proteins. To identify
these putative biomarkers is extremely difficult when one considers that they must be isolated from
an extremely complex mixture (e.g., serum, plasma, etc.) and significant sequencing is required to
establish their identity. While CE has been established with MS/MS capable instrumentation, the
broad identification of proteins in complex biofluids is still not at the level of RPLC-MS/MS.
ACKNOWLEDGMENTS
This project has been funded in whole or in part with federal funds from the National Cancer Institute,
National Institutes of Health, under Contract NO1-CO-12400. The content of this publication does
not necessarily reflect the views or policies of the Department of Health and Human Services, nor
does the mention of trade names, commercial products, or organizations imply endorsement by the
United States Government.
REFERENCES
1. Edwards, E. and Thomas-Oates, J., Hyphenating liquid phase separation techniques with mass
spectrometry: on-line or off-line. Analyst, 130, 13, 2005.
2. Bonneil, E. and Waldron, K. C., On-line solid-phase preconcentration for sensitivity enhancement in
capillary electrophoresis. J. Capill. Electrophor. Microchip Technol., 6, 61, 1999.
3. Canas, B., Lopez-Ferrer, D., Ramos-Fernandez, A., Camafeita, E., and Calvo, E., Mass spectrometry
technologies for proteomics. Brief. Funct. Genomic. Proteomic., 4, 295, 2006.
4. Liu, Z. and Pawliszyn, J., Online coupling of solid-phase microextraction and capillary electrophoresis.
J. Chromatogr. Sci., 44, 366, 2006.
5. Janini, G. M., Zhou, M., Yu, L. R., Blonder, J., Gignac, M., Conrads, T. P., Issaq, H. J., and
Veenstra, T. D., On-column sample enrichment for capillary electrophoresis sheathless electrospray
ionization mass spectrometry: evaluation for peptide analysis and protein identification. Anal. Chem.,
75, 5984, 2003.
6. Olivares, J. A., Nguyen, N. T., Yonker, C. R., and Smith, R. D., On-line mass spectrometric detection
for capillary zone electrophoresis. Anal. Chem., 59, 1230, 1987.
7. Ding, J. and Vouros, P., Advances in CE/MS. Anal. Chem., 71, 378A, 1999.
8. Kelly, J. F., Ramaley, L., and Thibault, P., Development of electrophoretic conditions for the char-
acterization of protein glycoforms by capillary electrophoresis-electrospray mass spectrometry. Anal.
Chem., 69, 51, 1997.
9. Wahl, J. H., Gale, D. C., and Smith, R. D., Sheathless capillary electrophoresis-electrospray ion-
ization mass spectrometry using 10 µm I.D. capillaries: analysis of tryptic digests of cytochrome c.
J. Chromatogr. A, 659, 217, 1994.
10. Cao, P. and Moini, M., A novel sheathless interface for capillary electrophoresis/electrospray ionization
mass spectrometry using an in-capillary electrode. J. Am. Soc. Mass Spectrom., 8, 561, 1997.
11. Chang, Y. Z. and Her, G. R., Sheathless capillary electrophoresis/electrospray mass spectrometry using
a carbon-coated fused-silica capillary. Anal. Chem., 23, 626, 2000.
12. Wahl, J. H. and Smith, R. D., Comparison of buffer systems and interface designs for capillary
electrophoresis-mass spectrometry. J. Cap. Electrophoresis, 1, 62, 1994.
13. Janini, G. M., Conrads, T. P., Wilkens, K. L., Issaq, H. J., and Veenstra, T. D., A sheathless nanoflow
electrospray interface for on-line capillary electrophoresis mass spectrometry. Anal. Chem., 76, 1615,
2003.
14. Severs, J. C. and Smith, R. D., Characterization of the microdialysis junction interface for capillary
electrophoresis/microelectrospray ionization mass spectrometry. Anal. Chem., 69, 2154, 1997.
15. Tong, W., Link, A., Eng, J. K., and Yates, J. R., Identification of proteins in complexes by solid-phase
microextraction/multistep elution/capillary electrophoresis/tandem mass spectrometry. Anal. Chem.,
71, 2270–2278, 1999.
16. Settlage, R. E., Russo, P. S., Shabanowitz, J., and Hunt, D. F., A novel µ-ESI source for coupling
capillary electrophoresis and mass spectrometry: sequence determination of tumor peptides at the
attomole level. J. Microcolumn Sep., 10, 281–285, 1998.
17. Kaiser, T., Kamal, H., Rank, A., Kolb, H. J., Holler, E., Ganser, A., Hertenstein, B., Mischak,
H., and Weissinger, E. M., Proteomics applied to the clinical follow-up of patients after allogeneic
hematopoietic stem cell transplantation. Blood, 104, 340, 2004.
18. Sassi, A. P., Andel, F., III, Bitter, H. M., Brown, M. P., Chapman, R. G., Espiritu, J., Greenquist, A. C.,
et al.. An automated, sheathless capillary electrophoresis-mass spectrometry platform for discovery of
biomarkers in human serum. Electrophoresis, 26, 1500, 2005.
19. Weissinger, E. M., Nguyen-Khoa, T., Fumeron, C., Saltiel, C., Walden, M., Kaiser, T., Mischak, H.,
Drueke, T. B., Lacour, B., and Massy, Z. A., Effects of oral vitamin C supplementation in hemodialysis
patients: a proteomic assessment. Proteomics, 6, 993, 2006.
20. Wetterhall, M., Palmblad, M., Hakansson, P., Markides, K. E., and Bergquist, J., Rapid analysis of
tryptically digested cerebrospinal fluid using capillary electrophoresis-electrospray ionization-Fourier
transform ion cyclotron resonance-mass spectrometry. J. Proteome Res., 1, 361–366, 2002.
21. Wittke, S., Mischak, H., Walden, M., Kolch, W., Radler, T., and Wiedmann, K., Discovery of biomark-
ers in human urine and cerebrospinal fluid by capillary electrophoresis coupled to mass spectrometry:
towards new diagnostic and therapeutic approaches. Electrophoresis, 26, 1476–1487, 2005.
9 Light-Based Detection
Methods for Capillary
Electrophoresis
Cory Scanlan, Theodore Lapainis, and Jonathan V. Sweedler
CONTENTS
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306

9.2 General Considerations for Detectors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
9.2.1 The Detector Cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
9.2.2 Resolution and Response Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
9.2.3 Maximizing Separation Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
9.2.4 Figures of Merit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
9.2.5 Qualitative versus Quantitative Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
9.2.6 Indirect Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
9.3 Absorbance Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
9.3.1 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
9.3.2.1 Light Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
9.3.2.2 Focusing Optics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
9.3.2.3 The Detection Cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
9.3.3 Wavelength Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
9.4 Fluorescence Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
9.4.1 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
9.4.2.1 Light Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
9.4.2.2 The Detection Cell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
9.4.2.3 Collection Optics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
9.4.3 Multidimensional Fluorescence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
9.4.3.1 Wavelength-Resolved Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
9.4.3.2 Time-Resolved Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
9.4.3.3 Polarization Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
9.4.4 Detecting Biomolecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
9.4.4.1 Derivatization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
9.4.4.2 Native Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
9.5 Thermooptical Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
9.6 Chemiluminescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
9.7 Radionuclide Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
9.8 Other Light-Based Detection Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
9.9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326
305
9.1 INTRODUCTION
The small sample volumes used in capillary electrophoresis (CE) make the technique suitable for
myriad small-scale biochemical and materials science applications. These small volumes also require
detection methods that exhibit high sensitivity and suitable selectivity. Such rigorous demands on
detector performance can have important consequences with respect to other aspects of the CE
system. For example, it is well known that higher efficiency and faster separations are achieved
using smaller inner diameter (ID) capillaries. And yet 50-µm ID capillaries are more common than
5-µm ID capillaries in systems using light-based detection, primarily because the longer optical path
length of the larger capillary aids in detection. In this way, the performance of the detection system
can dictate the overall performance of a given CE system. Thus, knowledge of the various detection
modes available, as well as their respective figures of merit, aids in designing CE-based methods of
analysis (see Table 9.1).1
In addition to a working knowledge of the various detection schemes available for CE, the
sensitivity, selectivity, and detectability requirements of the particular analyte(s) and application
should be considered when choosing a suitable detector. In a setting where the CE system will be
devoted to a single application, the detector’s parameters can be tuned to optimize the detection of
the analyte(s) of interest for that application. If, however, the instrument is to be used for a wide
variety of applications, then versatility is of primary importance, which may come at the expense of
optimal performance for any individual analyte. If none of the available detection methods allows for
adequate detection of the analyte(s) of interest, it may be possible to chemically modify the analyte
by derivatization in order to render it detectable.2
This chapter describes detection methods for CE that involve probing a sample with electromag-
netic radiation. While the discussion will emphasize the two most common forms of light-based
TABLE 9.1
Light-Based Detection Modes for CE, and Corresponding Limits of Detection
Minimum Detectable Minimum Detectable Representative Detectable
Detection Principle Concentration (M) Quantity (mol) Molecules
Direct absorbance 10−5 to 10−7 10−13 to 10−16 Most
Indirect absorbance 10−4 to 10−6 10−12 to 10−15 Small ions, amino acids, fatty
acids, polymers
Laser-induced fluorescence 10−9 to 10−12 10−18 to 10−21 Dyes, amino acids, peptides,
proteins, carbohydrates,
nucleotides, vitamins,
pharmaceuticals
Indirect fluorescence 10−6 to 10−8 10−14 to 10−16 Small ions, pollutants, amino
acids, peptides
Chemiluminescence 10−7 to 10−9 10−14 to 10−16 Dyes, amino acids, peptides,
proteins, nucleotides
Thermooptical absorbance 10−5 to 10−7 10−15 to 10−18 Dyes, amino acids, proteins,
nucleotides, vitamins
Radioactivity 10−6 to 10−10 10−14 to 10−18 Small ions, amino acids,
peptides, proteins,
nucleotides, pharmaceuticals
Raman 10−3 to 10−5 10−12 to 10−15 Dyes, small ions, amino
acids, nucleotides
Source: Adapted from Khaledi, M. G., in Chemical Analysis: A Series of Monographs on Analytical Chemistry and Its
Applications, 1st ed., Winefordner, J. D. (ed.), Wiley-Interscience, Raleigh, 1998, pp. 308–9.
Light-Based Detection Methods for Capillary Electrophoresis 307
detection—ultraviolet-visible (UV-Vis) absorbance and fluorescence detection—other less common

methods, such as thermooptical detection, are discussed as well. Prior to discussion of the individ-
ual detection schemes, the basic considerations and requirements of optical detectors for CE are
presented.
There have been a number of important developments in light-based detection for CE since the
previous edition of this work was published. For example, multichannel detection has become more
common for absorbance as well as fluorescence, and Fourier transform infrared (FTIR) and nuclear
magnetic resonance spectroscopies are now viable detection modes. Thus, while the purpose of this
chapter is to describe the instrumentation and performance of more common optical detectors, newer
methods are presented throughout.
Finally, it is worth noting that limits of detection of a system mentioned throughout the chapter
depend on the separation performance, preconcentration steps, and mode of operation (i.e., zone
electrophoresis, electrochromatography, isotachophoresis, etc.) of the separation system. Thus, one
can often optimize a CE method to improve the overall system performance, given the sample
stacking and other approaches that may be appropriate for a particular sample type or application.3−6
9.2 GENERAL CONSIDERATIONS FOR DETECTORS

9.2.1 THE DETECTOR CELL
The capillaries typically used in CE are made up of fused silica, with IDs of 10–100 µm, outer
diameters (ODs) of 100–375 µm, and coated with a 10–30 µm protective layer of polyimide (PI).
There are several locations relative to this capillary where detection can take place. These detection
locations include on-column, end-column, whole-column, and post-column detection.
For on-column detection, a portion of the protective PI coating is removed, either by acid
application (hot HNO3 ) or by burning. This results in a cylindrical length of bare fused silica, which
is used as the detection window. The window is generally placed as close as practicable to the outlet
of the capillary in order to maximize the fraction of the capillary used for the separation.
End-column detection is done by coupling the outlet of the capillary to an external detection cell.
The properties of this detection cell (path length, material, etc.) can be tailored to allow for high
sensitivity and low analyte limits of detection. The careful selection of coupling and transfer fluidics,
such as zero dead-volume unions, ensures that the high resolution afforded by CE is maintained within
the end-column detector.
In the post-column detection configuration, detection is carried out at a location remote to the
outlet of the capillary. For example, in radionuclide detection, the capillary eluent is spotted onto a
membrane or collected as fractions, and subsequently exposed to a scintillator for detection.7,8 Post-
column detection allows for the separation and detection to be independently optimized. In the case
of radionuclide detection, the detector integration time can be maximized without a corresponding
loss in resolution (see Section 9.2.2).
Whole-column imaging can be used to view the entire separation process as it occurs. As an
example, this type of detection is often used in capillary isoelectric focusing in order to determine
when substances have become sufficiently concentrated for collection and further characterization.
To do this, the capillary can be held stationary while mobile scanning systems scan the length of
the capillary,9 or detectors with high spatial resolution such as charge-coupled devices (CCDs) can
image the entire capillary at once.10 These imaging detectors typically measure UV-Vis absorbance
or fluorescence.
9.2.2 RESOLUTION AND RESPONSE TIME

In general, the signal-to-noise ratio (SNR) of a detected peak increases with longer detector integra-
tion times. With the exception of post-column detectors, the integration time is limited by the velocity
of the migrating analyte bands. Typical flow rates are on the order of 10–100 nL/min, resulting in
analyte bands that migrate through the detector cell on the order of a second. In peak profile analysis,
six or more data points are desired over the width of the peak band in order to describe it adequately.
This means that the length of the detection cell should be restricted to one-sixth of the width of
the narrowest band; common cell lengths are 100–200 µm. There are several strategies that can
be used to increase detector response within these general spatial and temporal restrictions, includ-
ing reducing the operating voltage of the separation, as well as transient stopped- or reduced-flow
detection.11
9.2.3 MAXIMIZING SEPARATION EFFICIENCY

The magnitude of detector response for a given amount of analyte is maximized when the analyte is
confined within narrow bands. As described in Chapter 1, there are several important considerations
for maintaining high efficiency in CE. These include voltage, temperature, and sample introduction.
In general, the use of the highest practicable voltages maximizes efficiency. Also, since temperature
gradients decrease efficiency, narrow bore capillaries allow for more effective heat dissipation. Peak
dispersion can be minimized by ensuring that the concentration of any given analyte is more than
two orders of magnitude lower than the concentration of the background electrolyte.12
9.2.4 FIGURES OF MERIT

Selection of an appropriate detection method for CE is facilitated by a comparison of each method’s
respective figures of merit. The relevant figures of merit include the limit of detection (LOD),
sensitivity, and linear dynamic range.
The LOD is a measure of the smallest amount of analyte that can be detected. Typically, LODs
are defined as the concentration or absolute amount (mass or moles) of analyte that yields a signal
that is equal to three times the standard deviation of the baseline signal.
Sensitivity describes the ability of a detector to differentiate between different amounts of analyte.
When a plot of signal versus amount of analyte generally yields a straight line, the term “sensitivity”
refers to the slope of this line.
It is observed that the relationship between signal and concentration is linear over a finite range
of concentrations. At higher concentrations, various phenomena (scattering, nonlinear optical pro-
cesses, “inner filter” effect, etc.) lead to deviations from linearity. The range of concentrations where
a linear relationship between signal and concentration holds is known as the linear dynamic range
of the detector.
As mentioned in Section 9.1, these figures of merit depend not only on the detection system, but
also on other factors as well, including the nature of the analyte, the composition of the separation
medium, and the mode of operation of the separation system.
9.2.5 QUALITATIVE VERSUS QUANTITATIVE INFORMATION

The linear relationship between analyte concentration and detector response can be exploited to gain
quantitative information about an “unknown” sample. To do this, solutions of known concentrations
are used to generate a standard plot of intensity versus concentration, which is then used to determine
the concentration of the unknown.
In addition to this quantitative information, it is also possible to obtain qualitative information
about an analyte, such as the presence or absence of specific functional groups and their relative
locations or orientations. A subset of the available optical detection methods are capable of providing
this information, such as multiwavelength UV-Vis absorbance or fluorescence, nuclear magnetic
resonance, and Raman spectroscopies; these techniques will be discussed in more detail in this
chapter.
9.2.6 INDIRECT DETECTION

Detection methods are typically chosen based on the characteristics of the analyte of interest, for
example, absorbance detectors for highly absorbing species, or fluorescence detectors for intrinsic
fluorophores with sufficient quantum yields. There are, however, cases where the analyte lacks
the properties that enable it to be analyzed by any of the conventional detection methods. If this
is the case, and suitable derivatization chemistry is unavailable, it may be possible to detect the
analyte indirectly.13−17 To do this, a signal-generating substance is incorporated into the separation
medium, and the presence of the analyte is detected as a decrease in the measured signal. The
observed decrease is due to the displacement of the signal-generating species by the analyte; charge-
density conservation and solubility changes are mechanisms by which this displacement can occur.
Indirect detection has been used in conjunction with absorbance, electrochemical, and fluorescence
detection systems.
9.3 ABSORBANCE DETECTION

9.3.1 THEORY
The term absorbance refers to a phenomenon whereby a molecule is promoted to an excited state
by a photon. Absorbance is normally determined by measuring the transmission of light through a
sample. Transmittance, T , is the ratio of intensity of a light beam before and after it passes through
a solution, that is, T ≡ I/I0 , where I is the intensity after passing through the sample and I0 is the
initial intensity. Absorbance, A, is related to the transmission by A = − log(T ).
A solution’s absorbance can be related to its constituent molecular components by the Beer–
Lambert law: A = εlc, where ε is the molar absorptivity (in units of inverse length per M), l is the
path length through which the light passes, and c is the concentration of the absorbing species. The
molar absorptivity of a substance is dependent on various parameters, including the solution pH,
solvent polarity, and the wavelength of light. It can be seen from the Beer–Lambert law that, for a
fixed path length and molar absorptivity, the absorbance of a solution is linearly dependent on the
concentration of the absorbing species in solution. This relationship allows for the quantification
of an absorbing species, if the measured value is within the dynamic range of the instrument (see
Section 9.2.4).
There are several considerations with respect to measuring absorbance in CE. First, the decrease
in measured light intensity after passing through a solution is due to both stray light/scattering
processes and absorbance. Thus, it is important that the stray light/scattering be accounted for when
relating absorbance to analyte concentration. This is typically accomplished by making appropriate
“blank” measurements of a solution containing all of the pertinent buffers and solvents, but without
the analyte of interest. Second, the magnitude of the measured signal scales with the path length.
In CE, capillary IDs are kept small in order to minimize the deleterious effects of temperature
gradients on resolution. This trade-off between detector performance and separation resolution is
characteristic of light-based detection methods for CE and underscores the importance of a working
knowledge of both the principles of CE and of the various methods of detection.
Figure 9.1 shows the layout of a typical absorbance detector for CE. The light generated by the
source is collected and focused into a beam by the focusing optics. The spatial profile of this beam
can be defined by an aperture, which also serves to reduce background light. The beam passes through
the detection cell, where the analyte absorbs a portion of the incident light. The emerging beam is
then spectrally filtered by a prism or spectrograph to select the desired wavelength(s) and detected by
an appropriate signal transducer, such as a photodiode array or photomultiplier tube. The resulting
Detection cell
Focusing optics (capillary) Wavelength
selector
Light source Aperature
A D
Analog-to-digital Electronic filters Photodetector

Recording device converter
FIGURE 9.1 Flow-diagram of an absorbance detection system for CE.
signal is either displayed using an analogue device or it is converted to a digital signal and sent to a
PC for analysis. The following subsections describe some of these essential components in greater
detail.
9.3.2.1 Light Source

The primary considerations with respect to light sources for absorbance-based detectors are output
stability and spectral characteristics. Because analyte quantification is dependent on the measured
intensity of light after it emerges from the detection cell, it is important that the amount of light that
enters the cell does not fluctuate in time. Thus, light sources that produce a stable output of light
are integral to the satisfactory performance of an absorbance-based detector. In addition to output
stability, the spectral characteristics of the output are of importance. Not only must the wavelengths
provided match the absorbance bands of the analyte(s) of interest, but also the spectral profile
must be suitable as well. Some detectors monitor a single wavelength, so that narrow emission
lines are preferred, while wavelength-resolved instruments require a source with a broad spectral
output.
Light sources that are suitable for use in absorbance-based detectors include atomic emission
lamps, deuterium arc lamps, lasers, and light-emitting diodes (LEDs). In atomic emission lamps,
photons are generated when electronically excited atoms relax to their ground states. Atomic emission
lamps are available that use transition metals such as mercury, cadmium, and zinc, which generate
strong emission lines at 254, 229, and 214 nm, respectively. These line source lamps are a reasonable
choice for systems that require only a single (or relatively few) output wavelength(s). In addition,
there are a number of blackbody lamps available that emit broadband radiation, such as a tungsten
filament, which produces a range of wavelengths from 350 to 2500 nm. Another broadband source is
the deuterium arc lamp. In contrast to atomic emission lamps, the UV light provided by the deuterium
arc lamp is due to the radiant decay of molecular deuterium. The lamp produces a relatively stable
output in the range of 160–400 nm, and is the most commonly used light source for absorbance
detectors.
In addition to the various lamps, lasers can be used as light sources for absorbance detectors. The
highly directional, spatially coherent emission from lasers can be efficiently collected and focused
to a small spot within a capillary. These characteristics make lasers a viable choice for sources in
absorbance detection systems (especially for smaller-diameter capillaries), although their use is less
common than the various incoherent lamps. More recently, LEDs have been used as light sources
for optical detectors for CE, including absorbance detectors.18−20 The low cost and compact size of
LEDs make them well suited for field-deployable sensors and lab-on-a-chip applications.
As mentioned above, output stability is important. All sources exhibit some degree of output
fluctuation, and so many instruments include a means by which these fluctuations can be monitored.
The most common approach is to use a beam-splitter or displacer to generate a reference beam that
can be monitored by an auxiliary detector.21 The reference channel can serve to minimize the effects
of instability on detection limits, although various factors (including shot noise) place restrictions
on this ability.
9.3.2.2 Focusing Optics

The performance of an absorbance detector is dependent on the ability to effectively couple the
excitation source to the detection cell. This involves collecting and focusing as much source light
as practical into the detection zone, as well as minimizing the stray source light that reaches the
detector without interacting with the analyte.22
Various lenses and lens arrangements are available to collect and focus a large fraction of the
incident source light. Commonly used lenses include the plano-convex singlet, as well as the quartz
ball lens. The specific lens arrangement that best suits a given detector is dependent on the source
used. As an alternative to the use of lenses, fiber optics can be effective in coupling the source light
to the detection cell (as well as coupling the detection zone to the detector).23−25 The use of fiber
optics allows for flexibility in the placement of the source and detector and enables the illumination
of multiple capillaries or multiple locations on the same capillary, although fiber optics are more
difficult to interface for the smallest diameter capillaries.
The amount of stray light that reaches the detector can be minimized by the appropriate use of
apertures placed before/after the detection cell.22 A pre-cell aperture can be used to block extraneous
light and allow only a well-defined cross section to pass through the capillary. For applications
where high spatial resolution is critical, the lateral dimension of this area (along the capillary) can be
decreased. In addition, a second aperture, placed after the detection cell but in front of the detector,
can be used to prevent light that is scattered off the capillary walls from reaching the detector.
9.3.2.3 The Detection Cell

The most common illumination geometry used in absorbance detection cells is the transcolumn
configuration (Figure 9.2a), in which the light passes through the detection cell along a path that is
perpendicular to the length of the capillary. For on-column detection, this sets the maximum optical
path length at the ID of the capillary. As mentioned previously, the performance of a light-based
detector is directly proportional to this path length. Thus, it is not surprising that numerous approaches
to extending the effective path length of optical detectors have been reported in the literature.
Axial-capillary illumination (Figure 9.2b) is an approach to increasing optical path length that
involves focusing the source light into the capillary bore and directing it along the capillary axis.26
In this set-up, the tightly focused lamp light or laser beam is directed into the capillary inlet, and
the light exiting the outlet end of the capillary is detected by an appropriate signal transducer. Axial
illumination relies on partial internal reflection within the capillary, which, in turn, requires that the
refractive index of the separation buffer be greater than that of the capillary walls. The light exiting
the capillary contains both the wall-propagated and internally propagated light rays, and so spatial
filtering or increasing the refractive index of the separation medium is used to selectively measure
the internally propagated rays.
A Z-cell can be constructed by heating the capillary in several places, and bending it into the
shape of a Z, as shown in Figure 9.2c. The source light enters the cell at the first of the two bends,
propagates axially though the capillary, and exits the cell at the other bend. The use of a Z-shaped cell
(b)
(a) (c)
(d) (e)
FIGURE 9.2 Different detector cell types used to increase the path length for optical detection. (a) Trans-
capillary illumination, (b) axial-capillary illumination, (c) Z-cell capillary, (d) bubble cell capillary, and (e)
rectangular capillary trans-illumination.
has been reported to yield up to a 10-fold improvement in detection limits relative to conventional
transcolumn illumination.22
The bubble cell increases the optical path length by expanding the ID of a small length of
capillary that makes up the detection zone relative to the ID of the rest of the capillary, as shown
in Figure 9.2d. Path length increases of up to fivefold have been reported using the bubble cell, for
example, a 125-µm ID bubble cell formed on a 25-µm ID capillary.27
While the most common cross section for fused-silica capillaries is circular, square cross section
capillaries have been employed in CE (Figure 9.2e). Square capillaries minimize the scattering caused
by curved capillary surfaces, and increase the probed volume relative to a circular capillary of the
same width. Originally, these rectangular capillaries were made up of borosilicate glass, which
is opaque in the UV.28 More recently, square fused-silica capillaries have become commercially
available and are a viable option for increasing the path length in UV absorbance detectors.29,30
9.3.3 WAVELENGTH SELECTION

Measuring absorbance at 200–325 nm is often satisfactory for general-use absorbance detectors, as
most analytes (including many biomolecules) have some absorbance at these wavelengths.31 There
are, however, cases where the analyte(s) of interest do not absorb appreciably in this range.32 Thus,
it may be beneficial to determine a more suitable wavelength for a particular analyte, for example,
when dealing with low-abundance analytes or buffers that strongly absorb at these wavelengths. The
optimum wavelength is the one that gives the maximum SNR for the analyte. Often, this wavelength
is the one where the analyte exhibits maximum absorbance (λmax ). A spectrophotometer can be used
to generate background-subtracted absorbance spectra, which can in turn be used to identify λmax .
As an alternative to selecting and monitoring absorbance at a single wavelength, it is possible
to use a multiwavelength absorbance detection system. These systems involve either rapid scanning
gratings33 used in conjunction with a single channel detector or, more commonly, a fixed spectrograph
used with a photodiode array.34−36 Indeed, a majority of the available commercial CE instruments
use wavelength-resolved diode-array absorbance detection. There are a number of reasons for the
popularity of these instruments, which have been discussed in detail elsewhere.37,38 With these
instruments, it is not necessary to select an appropriate wavelength for detection and quantification
before analysis; the optimal wavelength for each analyte can be determined and used for subse-
quent data analysis after running the samples of interest. The spectral information provided by these
detectors can also be useful for analyte identification or characterization, by comparing the spectral
features of a peak of interest to those of standard solutions. Not only is this beneficial for analyzing
“unknown” samples, but also it facilitates method development, since differences in separation con-
ditions can alter migration times and even elution order. Moreover, completely unresolved peaks
can be identified as such, because the observed spectral characteristics will differ from those of any
pure analyte (i.e., peak purity confirmation).39 For partially resolved components, the multichannel
data can be used in conjunction with various computational approaches to enable quantification
of the constituent components, without having to change separation conditions and reanalyze
samples.39,40
9.4 FLUORESCENCE DETECTION

9.4.1 THEORY
Fluorescence is a process in which a photon is emitted by an electronically excited molecule as
it relaxes to its ground electronic state. Each of the relevant electronic states contains numerous
vibrational sublevels, and so fluorescence excitation and emission spectra are relatively broad. The
manifold of various vibronic states of a fluorophore results in unique and complex emission spectra,
which may enable differentiation between two closely related molecules. In addition, the observed
emission spectra are dependent on the local chemical environment of the fluorophore, and can differ
depending on factors such as solvent polarity and solution pH. Indeed, this exquisite dependence of
emission spectra on surroundings can be used to deduce the relative location of tryptophan residues
within a protein. An excellent overview of the qualitative aspects of fluorescence can be found in
Lacowicz’s text on the subject.41
The observed magnitude of fluorescence emission is given by If = cdkIl Qf , where If is the
intensity of fluorescence (photons cm−2 s−1 ), c is the concentration of the fluorophore (M), d is
the path length of excitation (cm), k is the base-ten logarithm of the molar absorptivity (ε, see
Section 9.3), Il is the intensity of the laser, and Qf is the fluorescence quantum yield (unitless).42 The
dependence of the fluorescence signal on the intensity of the excitation source suggests that high-
power light sources can be used. In practice, however, the optimal laser power is also a function of
other factors such as analyte photostability, and so should be determined empirically.43,44
Fluorescence detection affords the lowest reported detection limits of the various detector types
for CE (Table 9.1). The high performance of fluorescence-based detection systems results from
a number of characteristics. First, a single fluorophore molecule can cycle repeatedly between
excitation and emission, thus generating up to thousands of photons before exiting the detection
zone or undergoing photodestruction. Second, the observed emission spectrum is shifted relative to
the excitation wavelength(s), so that filters can be used to minimize the amount of scattered excitation
radiation that reaches the detector. As a result, single molecule detection is possible using LIF, and
LIF-based detection systems for CE are capable of providing LODs in the yoctomole range.45,46
The basic fluorescence detection system in CE consists of several components, including an excitation
source, a detection cell, collection optics, and a detector. Within these requirements, there are
a wide variety of possible optical configurations, and vastly differing performance specifications
between systems. The design constraints and general considerations relevant to the development of
a fluorescence detection system have been detailed in the literature.47−50 The following sections will
discuss the major components of a fluorescence detection system and recent innovations.
9.4.2.1 Light Source

There are a number of criteria relevant to the selection of an appropriate light source for a CE
fluorescence detection system. As with sources for absorbance detectors, the output wavelength and
spectral characteristics are important, as is the stability of the source output. In addition, the inten-
sity of a light source is important for fluorescence detectors, as the observed signal scales with the
excitation irradiance. The sources for fluorescence detection can be differentiated into two types,
coherent (lasers) and incoherent (lamps). Jorgenson and Lukacs51−53 built the first fluorescence
detector for CE in 1981, and it used a high-pressure mercury lamp for excitation. Such types of
incoherent lamps, combined with predetection cell monochromaters to filter out unwanted wave-
lengths from the emission source, provide flexibility in wavelength selection to match the wavelength
of maximum absorbance for an analyte. Stabilized arc lamps are becoming increasingly available
that, along with spatial filters and carefully constructed optical setups, allow for sensitive detection
for chromophore-labeled analytes, such as detection limits of 10 pM for fluorescein-labeled amino
acids.54,55
The use of LEDs as sources for fluorescence is becoming more commonplace.56−59 As mentioned
previously (in Section 9.3), the small footprint and low cost of these sources make them a good choice
for microchip (e.g., µTAS or lab-on-a-chip) devices. It is likely that these sources will become more
commonplace as the output power generated by these sources increases.
Lasers are another source of excitation radiation used in fluorescence detection systems. The
high-directional output of a laser maximizes the fraction of total output that can be easily focused
down to a spot size compatible with the dimensions of CE detection cells. The output of a laser is
also typically monochromatic, or a discrete set of spectrally narrow lines. This type of output makes
it relatively easy to filter out low-level incoherent plasma radiation and undesired emission lines
without greatly diminishing the overall output power. In addition, many lasers provide flexibility
in terms of pulse width and repetition rate, which allows one to optimize excitation with respect to
analyte photostability.
The original reported use of lasers in CE with fluorescence detection was described by Gassmann
et al.60 in which laser-induced fluorescence (LIF) was used to distinguish and measure the chiral
enantiomers in mixtures of dansylated amino acids. Since that time, lasers have become common-
place in CE systems with fluorescence detection. The most common types of lasers used have been
the He–Cd, Ar-ion, and He–Ne lasers, as they are relatively inexpensive and have emission lines
that match commonly used fluorescent reagents, such as o-phthalaldehyde (OPA) (325 nm) and
fluorescamine (354 nm).
The majority of LIF instruments use excitation wavelengths in the visible portion of the
spectrum.61 There is, however, interest in the development of LIF instruments that use near-infrared
(NIR) excitation.62 There are several reasons for this, including the availability of inexpensive NIR
diode lasers, as well as the decreased likelihood of photobleaching or significant Raman background
at longer wavelengths. The further development of suitable derivatization strategies will increase
the number of applications that can be addressed by NIR-LIF.63,64
At the opposite end of the spectrum, UV sources for CE-LIF are becoming increasingly popular.
UV radiation is capable of inducing fluorescence in many intrinsic fluorophores, including a number
of biologically relevant molecules. The frequency-doubled Ar-ion laser (257 nm) was one of the first
examples reported of UV-excitation for CE-LIF, and yielded improvements in LODs for a number
of substances, such as conalbumin (1.4 × 10−8 M).65 As another example, a frequency-doubled Kr
laser operating at 284 nm has been used for the analysis of neuropeptides and small biomolecules,
and exhibited LODs for tryptophan of ∼800 zmol.66 In addition to frequency-doubled ion lasers,
a number of relatively inexpensive pulsed lasers such as frequency quadrupled YAG (266 nm),67
KrF (248 nm),68 and hollow-cathode metal vapor lasers have appeared, which provide deep-UV
excitation (e.g., 224 and 248 nm).69−73
Multiphoton excitation is another strategy for exciting the same transitions that UV sources excite,
while using NIR or visible lasers and associated optical elements.74,75 In multiphoton excitation, two
or three photons are absorbed by a fluorophore, which is then promoted to a vibronic state with an
associated energy that is equal to the sum of the energies of the individual absorbed photons. Because
the photons must be absorbed simultaneously (e.g., within ∼1 fs), this technique requires high peak
n , where I
power. More specifically, the observed fluorescence emission is proportional to Ipeak peak is
the peak intensity of the excitation radiation and n is the number of photons involved in the transition.
The small probe volumes dictated by this nonlinear intensity dependence are compatible with the
small ID capillaries used in CE; this technique has been used to detect a variety of substances after
CE separation, including the aromatic amino acids, nucleotides, and monoamine neurotransmitters
such as serotonin.76,77
9.4.2.2 The Detection Cell

On-column capillary detection is the most widely employed configuration in CE with fluorescence
detection, as shown in Figure 9.3a. A small window is made on the capillary by burning off a section
(a)
24.3 Capillary
Filter PMT
Run
HV buffer
Collection optics
Laser Window
Focusing lens
Waste
(b)
Capillary
24.3
Run
HV buffer
Sheath Filter PMT

buffer
Collection optics
Laser
Focusing lens Quartz cuvette
Waste
FIGURE 9.3 Schematic diagrams of two CE-LIF detection systems. (a) On-column CE-LIF and (b) end-
column sheath-flow CE-LIF.
of the polyimide coating. In most cases, the source light is focused through this window, although
axial illumination (directed up the bore of the capillary) has been used as well.26 In either case,
the fluorescence emission is collected through the window. Collection can take place orthogonal to
the light source, or coaxial with it, as in the epi-illumination configuration.49,78 Immersion of the
cylindrical detection window in an index-matching solution can be helpful in maximizing excitation
and collection efficiency.79
End-column detection is another frequently used configuration for fluorescence-based CE detec-
tion systems. One example is the sheath-flow cuvette, originally used for flow cytometry adapted to
CE by Dovichi and coworkers (Figure 9.3b).80−82 In sheath-flow post-column detection, the end of
the capillary is inserted into a cuvette into which buffer is flowing. The linear flow rate of the sheath
buffer is set to match or slightly exceed the electroosmotic flow of the capillary, thus minimizing
turbulence and maintaining a narrow sample stream. The laser is focused to a spot immediately below
the outlet of the capillary; the use of a cuvette with a square cross section maximizes the amount
of excitation light that is transmitted into the probe volume. This type of detection cell is especially
useful for CE-LIF systems using UV-excitation, as metal ions present in fused-silica capillary walls
give rise to a broadband luminescent background. Impressive limits of detection have been obtained
using the sheath-flow cuvette.83,84 For example, detection limits of 10−13 M have been reported for
numerous rhodamine-based fluorophores.85,86
The post-column configuration has also been used in conjunction with fluorescence. Typically,
this involves depositing the eluent on a membrane, or collecting fractions, as is commonly done in
HLPC.87−89 The resulting spots or fractions can then be imaged with an appropriate fluorimeter,
fluorescence microscope, or other suitable fluorescence detector.
9.4.2.3 Collection Optics

The ideal collection optic would collect fluorescence from a 4πsr solid angle around the detection
zone with 100% efficiency, while being completely inefficient at collecting scattered laser light.
Obviously, there exists no such optic, but there are various approaches reported in the literature
taken to emulate this ideal. In most cases, either a high numerical aperture microscope objective or
a fiber-optic system is used to collect the fluorescence and transmit it to the signal transducer.
Microscope objectives used as collection optics have yielded the lowest reported LODs in CE
with fluorescence detection. They can be used alone or as part of a complete microscope setup. A
high-numerical-aperture objective collects a large quantity of light from a very small region and
focuses it onto the detector.
A number of fiber optic-based systems have been described for CE-LIF systems.90−92 These
systems boast a small footprint, and the terminal of the fiber optic can be placed exceedingly close to
the detection window, thus enabling high collection efficiency. Hence, fiber optics are an excellent
choice for systems where the overall footprint must be kept to a minimum, such as field-deployable
sensors or lab-on-a-chip applications.
Fluorescence-based CE detection systems generally also include a spatial filter and wavelength
filter located between the collection optics and the detector, in order to minimize the amount of
Rayleigh and Raman scatter that reaches the detector. Careful selection of the filter wavelength
range and rejection characteristics, based on the desired detected wavelength range and the spectral
output of the source, is important to achieving optimal detector performance. The use of line sources
such as lasers is advantageous in this respect, as a narrow pass filter can be used to block excitation
radiation while minimizing the amount of fluorescence that is blocked.
9.4.3 MULTIDIMENSIONAL FLUORESCENCE

9.4.3.1 Wavelength-Resolved Fluorescence
In wavelength-resolved fluorescence detection, a complete emission spectrum is generated for each
time point. The resulting data are plotted as a three-dimensional electropherogram, with axes
(a) Capillary
24.3 Mirror CCD

Slit
Run Filter
HV buffer
Collection optics
Grating
Focusing lens
Waste
(b)
600
4
W 85
av
el (s)
en Time
gt 2
h
(n 90
m 0
)
FIGURE 9.4 (a) Schematic diagram of a CCD-based wavelength-resolved CE-LIF instrument.

(b) Wavelength-resolved electropherogram of a mixture of seven peptides containing tryptophan and tyrosine
residues. (Adapted from Timperman, A. T., et al., Anal. Chem., 67, 3421, 1995. With permission.)
corresponding to time, wavelength, and fluorescence intensity. The first report of wavelength-
resolved fluorescence detection for microscale separations was by Swaile and Sepaniak in 1989,35
in which a spectrograph was used to disperse the fluorescence onto a photodiode array with 4-nm
spectral resolution. The advent of higher-sensitivity array detectors, such as CCDs,93 made this type
of detection feasible for low-abundance analytes (see Figure 9.4), as demonstrated by 100 molecule
detection limits for sulforhodamine94 and 3 nM LODs for the neurotransmitter serotonin.95
Multichannel fluorescence detection can also be carried out using a series of dichroic beamsplit-
ters and photomultiplier tubes (PMTs) in place of a spectrograph and CCD.96 Rather than generating
a high-resolution emission spectrum, this type of detector divides the collected emission into several
channels, each of which corresponds to a relatively broad range of wavelengths. Peaks are identified
based on a ratiometric “fingerprint” calculated from the relative intensities of the channels. This “fin-
gerprint” is capable of differentiating two closely related substances, such as various catecholamine
neurotransmitters.
Spectrally resolved fluorescence detection for CE has found a variety of applications and
is often used for the detection and identification of various biologically relevant intrinsic fluo-
rophores, including amino acids, proteins, neuropeptides, and small molecule neurotransmitters.66,95
For example, wavelength-resolved fluorescence was used to elucidate various aspects of sero-
tonin catabolism.97−99 In addition, multichannel fluorescence detection was used to differentiate
each of four fluorescently labeled dideoxy bases that terminate replication in a polymerase chain
reaction.48,100 This technique was instrumental in sequencing the human genome, as well as those
of other organisms such as Drosophila melanogaster and Caenorhabditis elegans.
9.4.3.2 Time-Resolved Fluorescence

While wavelength-resolved fluorescence is the most common form of multidimensional fluores-
cence detection in CE, other forms of multidimensional fluorescence have also been reported.
Time-resolved fluorescence involves measuring the temporal evolution of the fluorescence signal,
and using this information to differentiate various analytes.101 For example, high-resolution time-
domain data can be used to calculate fluorescence lifetime of an analyte, which in turn can assist in
peak identification by discriminating between short-lived and long-lived fluorophores.102 Moreover,
typical fluorescence lifetimes are orders of magnitude longer than scattering lifetimes (both Raleigh
and Raman). As a result, the signals resulting from scattering and fluorescence can be effectively
deconvolved.
9.4.3.3 Polarization Fluorescence

In addition to wavelength and time-resolved fluorescence techniques, polarization fluorescence can
yield important information about an analyte.41 This is especially true when differentiating between
chiral compounds. Combining, for example, a fluorescently tagged antibody immunoassay with
polarization detection allows for very sensitive detection limits of chiral enantiomers. Laser-induced
fluorescence polarization (LIFP) has been used to detect concentrations as low as 0.9 nM of an
antibody-bound cyclosporine A (an immunosuppressive drug) in human blood.103 A conventional
single channel fluorescence detector can be easily modified to perform such measurements, simply
by adding the appropriate polarization filters.
9.4.4 DETECTING BIOMOLECULES

9.4.4.1 Derivatization
As mentioned previously, fluorescence-based detection exhibits the lowest reported detection limits
of any detection mode for CE. And while this is true, the effective use of fluorescence detection is
contingent upon the ability to induce fluorescence in the analyte of interest with reasonable quantum
efficiency. While native fluorescence is appropriate for a number of analytes, many others (including
many of biological relevance) are either not intrinsically fluorescent or exhibit low fluorescence
quantum yields. As a result, a significant amount of research has been devoted to appropriate labeling
methods, whereby a fluorescent moiety is covalently bound to the analyte of interest in order to enable
its detection.104 While attaching fluorescent probes to molecules is not unique to CE, the small sample
volumes involved make the development of suitable derivatization reactions and procedures critically
important.
Derivatization can be carried out end-column, on-column, or precolumn, with the latter being
the most common. With precolumn derivatization, large starting sample volumes and a large molar
excess of derivatizing reagents can be used, and the reaction can be allowed to proceed indefi-
nitely. These steps minimize losses due to dilution or sample handling, and help to ensure that
the derivatization reaction proceeds to completion. Precolumn derivatization of lysine-containing
neuropeptides with fluorescamine has shown pmol detection limits,105 and 3-(p-carboxybenzoyl)
quinoline-2-carboxaldehyde was used to gain a detection limit of 9 zmol (10−21 ) for arginine.106
While using large volumes of the sample solution helps to minimize the effects of dilution by the
derivatizing reagents, there are times when this is not possible. This is the case in many biological
applications, such as single-cell analysis. For example, Oates and coworkers107,108 demonstrated
derivatization of proteins and peptides from a single cell in a 25 nL volume using naphthalene-
2,3-dicarboxaldehyde (NDA) as the derivatant. Derivatization of intracellular signaling compounds
can be done by using the cell itself as a reaction chamber. Selective derivatization with mono-
bromobimane of thiol-containing compounds within a single red blood cell,15 as well as selective
derivatization of glutathiones by NDA in single neuroblastoma cells were some of the first reports
published showing successful derivatization within a single cell.109
On-column analyte derivatization is another approach often used. For on-column labeling, it
is beneficial to use a reagent that is nonfluorescent in its unreacted form, such as NDA. Gilman
and Ewing110 were able to use on-column labeling with NDA to analyze a whole PC12 cell. To do
this, they injected the intact cell followed by a plug of derivatization mixture, and left the voltage
off for a brief time interval to facilitate cell lysis and derivatization prior to CE analysis. On-
column derivatization can also be carried out using low-volume connectors attached near the outlet
of the capillary to introduce derivatization reagent. Carefully constructed transfer fluidics maintain
separation efficiency upon introduction of the derivatant. This type of apparatus can be constructed
by laser-drilling holes into the separation capillary and introducing the reagent by gravity through a
T-connector. The introduction of OPA into the capillary using this arrangement allows for femtomole
amounts of amino acids to be detected.111
There are cases when it is desirable to carry out derivatization after the separation has taken
place. For example, small molecules can be difficult to separate after derivatization, because the size
of the label is often large relative to the analyte. In addition, some analytes have multiple reactive
sizes (such as amino groups) where derivatization can occur. This can result in a distribution of
fluorescence products from a single analyte, each giving rise to a peak on the electropherogram. To
address these and other complications, a number of methods have been developed to perform end-
column derivatization. Most end-column systems employ a similar set of components, including
a micropump to add derivatization reagent to the CE effluent and a secondary column where the
labeling reactions occur before the tagged analytes reach the detector. The simplest approach is to
use a sheath-flow system to introduce the reactants, such as the coaxial capillary reactor system
originally described by Rose and Jorgenson.112 In this setup, OPA was introduced at the outlet of
the capillary tip, allowing for attomole detection limits. Another approach is to use the sheath-flow
cuvette (discussed earlier) as a derivatization reactor; the fluorogenic reagent is simply added to the
sheath-flow buffer (Figure 9.5).113,114
Another end-column derivatization system involves a gap junction reactor, as described by Albin
et al.115 In the gap junction reactor, two capillaries are separated by a small gap in which the
fluorophore is introduced. Careful alignment of the two capillaries ensures that high separation
efficiency is maintained at the junction. The gap junction interface is an excellent choice when using
high-cost derivates, as smaller volumes can be used for reagent introduction.
For all of the various end-column derivatization schemes, it is important that labeling reac-
tions be chosen that exhibit fast kinetics. This ensures that derivatization is complete before the
analytes enter the detection zone, which in turn ensures maximum analyte detectability. Com-
mon derivates with suitable kinetics for end-column detection include fluorescamine, OPA, and
NDA.2,116,117 In addition to kinetics, the selection of an appropriate derivatization reaction can be
influenced by other factors, including the potential side reactions for a given set of reagents and
analytes, equilibrium and concentration considerations, and the pH at which the derivatization is
carried out.
9.4.4.2 Native Fluorescence

When possible, it can be advantageous to take advantage of an analyte’s intrinsic fluorescence.
Ideally, native fluorescence would exhibit a high quantum yield and could be excited with readily
available excitation sources. Unfortunately, few compounds exhibit fluorescence in the spectral range
of the most common lasers, that is, the visible portion of the spectrum. Nonetheless, there are some
compounds that fluoresce upon excitation with visible light. Examples include many pharmaceutical
compounds,118−120 porphyrins,121,122 and vitamin B6 metabolites.123 In contrast, excitation below
300 nm excites quite a number of biomolecules, including proteins and peptides containing Tyr, Trp,
or Phe residues, polycyclic aromatic hydrocarbons, as well as nucleosides and DNA.95,124−126 For
(a) Separation
capillary
PEEK Tee
Derivatization
regent
Ground
Reaction distance Laser

beam
Cuvette
Waste reservoir
(b)
(A)
1
Signal (arb)
5
2 3 4
(B)
(C)
5 6 7 8
Time (min)
FIGURE 9.5 (a) Schematic diagram of the sheath-flow post-column reactor. (b) Separation of protein standards
derivatized with NDA/β-mercaptoethanol at a height difference of (A) 0.5 cm, (B) 3 cm, and (C) 6 cm between
the derivatization solution reservoir and waste reservoir. (Adapted from Ye, M. L., J. Chromatogr. A, 1022, 201,
example, laser-induced native fluorescence can be used to detect epinephrine and norepinephrine
content of single bovine adrenal medullary cells (Figure 9.6).127
Most applications of native fluorescence detection involve the direct detection of a fluorescent
analyte of interest. However, native fluorescence can also be used indirectly to detect the presence of
nonfluorescent species. For example, the presence of a nonfluorescent enzyme can be detected based
on its conversion of substrate into natively fluorescent products. This is often used in conjunction
with electrophoretically mediated microanalysis, in which differences in buffer mobilities enable on-
column mixing and reaction of enzymes with substrates.128 One such study exploited the intrinsic
fluorescence of nicotinamide adenine dinucleotide (NADH) to investigate differences in the reactivity
of individual molecules of lactate dehydrogenase.129
The quantum efficiencies and photostabilities of most intrinsically fluorescent biomolecules are
modest relative to those of the available fluorescent labels. There are, however, several factors that
can affect these quantum yields. The ionic strength and pH of the separation conditions, as well
as the presence or absence of organic additives, can have a significant effect on the fluorescent
properties (i.e., quantum yields and Stokes’ shift) of these analytes.41,130 Thus, if sensitivity is of
Fluorescence
NE
1 2 3
4
0 2 4 6 8 10
Time (min)
FIGURE 9.6 Electrophoretic separation of intracellular components of a single bovine adrenal medullary cell
detected by laser-induced native fluorescence. NE = Norepinephrine; E = Epinephrine; Peaks 1–4 are unknown
components. (Reprinted from Chang, H. T. and Yeung, E. S., Anal. Chem., 67, 1079, 1995. With permission.)
utmost importance for a particular application, the separation buffer can be optimized in terms of
analyte fluorescence and separation efficiency.
9.5 THERMOOPTICAL DETECTION

Thermooptical detection exploits the deflection of light that occurs as a result of changes in the
refractive index of a light-absorbing solution. Although it is a transmittance-based detection method,
it is distinct from absorbance detection in that the measured decrease in transmittance is related to
changes in the index of refraction of the absorbing solution. This detection mode involves the use of
two laser beams that intersect in the detection zone of the capillary. One of the beams, the “probe”
beam, passes through the detection cell, and is detected by a photodiode or other appropriate signal
transducer. The second beam, the “pump” beam, is selected to match the absorbance profile of
the separation medium. This beam is temporally chopped, so that it passes through the detection
window only transiently. The index of refraction of the separation medium changes in time as a
result of heat-induced density changes in the solution, depending on whether or not the pump beam
is passing through the detection cell. As a consequence, the transmittance of the probe beam changes
in time, which can be measured by the photodetector. Furthermore, the change in refractive index
that occurs is dependent on the absorbing species in the separation medium. Hence, as different
substances migrate through the detection zone, the magnitude of the detected signal changes. The
electropherogram is generated by plotting the difference in transmittance between “pump on” and
“pump off” as a function of time.
Thermooptical detection has been combined with CE and used for protein and peptide Edman
degradation sequencing detection, native protein detection, as well as the analysis of derivatized
amino acid mixtures.131−134 Frequency-doubled argon-ion lasers have been employed to supply
an UV (257 nm) pump beam for the analysis of dansylated amino acids as well as the analysis of
etopside and etopside phosphate in human blood plasma.135,136 In addition, two-color thermoop-
tical absorbance detectors have been constructed, where each laser serves a dual role of “pump”
and “probe”; this system can be used to detect analytes that absorb in differing regions of the
electromagnetic spectrum.
The use of a near-field thermal lens system to increase pump beam absorption has recently been
reported by Pyell and coworkers.137 This innovation has resulted in a nearly 30-fold improvement in
detection limits for a multitude of nitroaromatic compounds, including a 0.15 mg/L detection limit
(compared to a previously reported LOD of 3.4 mg/L) for 2-nitrotoluene.
9.6 CHEMILUMINESCENCE
Chemiluminescence (CL) is a phenomenon whereby the electronically excited state product of a
chemical reaction generates optical radiation during relaxation to its ground state. There are two
general mechanisms of CL that are employed in the context of detection for CE.138 In direct CL,
the photon(s) are emitted by the excited state reaction product as it relaxes to the ground state.
In the second, the relaxation of the excited state takes place via energy transfer to a fluorophore,
which subsequently fluoresces; this is referred to as sensitized CL, because a fluorophore with a high
quantum yield can be used.
CL is an attractive detection method for CE in part due to its simple instrumentation. An external
radiation source is not required, reduces instrument cost, complexity, as well as the overall footprint
of the instrument. In addition, since there is no background from an excitation source, there is no need
to filter the collected radiation, which increases detection efficiency. In its simplest form, a CE-CL
instrument can consist of only a capillary, power supply, and a detector, such as a PMT or photodiode
detector. The small footprint of such instrumentation is particularly attractive for microscale total
analysis systems (µTAS) or lab-on-a-chip applications.139,140
CL detection can be used to probe intrinsic luminescence, such as that found in numerous
biological systems (bioluminescence), or analytes of interest can be derivatized by reaction with an
isothiocyanate or succinimidyl ester-linked chemiluminescent moiety. The first reported use of CL
for CE detection was by Hara et al.141−143 who used peroxyoxolate as the luminophore to detect a
protein–dye complex. Since then, CL has been used successfully to detect a variety of substances
after CE separation including metal ions,144 amino acids,145 peptides146 and proteins,147 and even
DNA.148
CL involves a chemical reaction between two or more species, and so the magnitude of the
detected signal is a function of each of these substances. To facilitate quantification of one of these,
the concentrations of the others must be held constant, or otherwise known. Several distinct strategies
have emerged to enable this, including coaxial flow on-column detection, off-column detection, and
end-column detection.149,150 One method that has gained considerable attention recently is to gen-
erate the luminophore in situ by an electron transfer reaction facilitated by an electrode; this specific
type of CL is referred to as electrogenerated chemiluminescence (ECL). For example, ruthenium
(II) tris(2,2 -bipyridine) can be oxidized to Ru(bpy)2+3 , which undergoes a chemiluminescent reac-
tion with various analytes. 151−156 Not only can this luminophore be used with biologically relevant
analytes such as amino acids and NADH without their prior derivatization, but the reactive state of
the luminophore can be regenerated by oxidation at an electrode in the detection zone.
Many developments in CL detection have taken place since the last edition of this book. There
are a number of reviews that highlight current research in the field.149,150 Given the relative low cost,
simplicity, and small footprint of CL detectors, as well as the persistent drive toward miniaturization
of analytical instruments, it is likely that research in CL detection for CE will continue to progress
at a rapid pace.
9.7 RADIONUCLIDE DETECTION

In radionuclide detection, the presence of a radiolabeled analyte is detected by the response of a
secondary substance to the radioactive decay of the radiolabel. The secondary substance is referred
to as a scintillator, and it responds to collision with a radioactive particle by the emission of pho-
tons. Since it is this emission of photons that is actually measured in radionuclide detection, it is
appropriate that this method of detection be included in a chapter covering light-based detection
methods. Radionuclide detection is both highly sensitive and selective. The high sensitivity of this
technique is due to the ease of detection of the energetic decay events, in conjunction with a very low
natural background. The high selectivity of this technique results from the fact that only the radiola-
beled sample yields a detectable response. As a consequence, it is possible, for example, to inject a
living system with a radiolabeled sample and investigate various metabolic pathways or biochemical
changes by following the movement of the radiolabeled nuclei, without interference from unlabeled
compounds.
The sensitivity and detection limits for radionuclide detection depend in large part on the particular
radioactive nuclei being used. More specifically, the isotope’s rate of decay and particle energy
determine the lowest attainable detection limit. In addition, the sensitivity is dependent on the
observation time of the isotope and the fraction of events detected, which can be expressed as
follows:
counts = (DPS) (time observed) (efficiency of detector)
where “counts” represents the number of detector counts recorded over an observed peak, “DPS”
(measured in Becquerels, or Bq) represents the number of radioactive decompositions per second
from the sample, “time observed” is the amount of time (in seconds) a radioactive molecule spends in
the detection volume, and “efficiency of detector” is the average fractional number of events sensed
by the detector per total number of events. DPS is directly proportional to the concentration of a
particular isotope in a given sample, which is determined by the isotope’s rate of decay. The rate
of decay for an isotope is described as the amount of time for half the nuclei in a sample to decay,
commonly expressed in days or years, also known as the half-life (t1/2 ). These rates can vary widely,
from more than 1022 years for very stable radionuclides, to 10−6 s for highly unstable ones. More
reasonable ranges for isotopes used in CE analysis are those of 32 P with a half-life of 14.2 days, and
14 C with a half-life of 5.73 × 103 years. Because peak times in CE are normally limited to a few
seconds, the “time observed” term in the preceding equation limits the counts observed, and hence
the overall detection sensitivity when measurements are performed online. In addition, it is evident
from the “DPS” term that the shorter the half-life of the isotope, the more likely the detector will see
enough events in a given amount of time to obtain a clear signal. Another consideration worth noting
is that high-energy decay events are more easily detected than low-energy events. This is due to the
decreased probability that an intervening atom will absorb a higher-energy particle before it reaches
the scintillator. Common particle energies for isotopes used in online radionuclide detection in CE
range from 1.71 MeV for 32 P to 0.156 MeV for 14 C. The higher-energy particles are generally able to
travel a longer distance before being absorbed. Besides the differing particle energies, there are also
different forms of decay for radioparticles, the most common in CE detection being the emission of
negative (β − ) particles, positive (β + ) particles, and neutral gamma rays (γ ).
Both on-column and post-column detection schemes have been developed for radionuclide detec-
tion for CE. The most common type used is an on-column configuration, which yields detection limits
in the 10−9 M range for isotopes such as 32 P. Isotachophoretic separations of 14 C were among the
first examples of online capillary radionuclide detection, performed by Kaniansky et al.157−159 The
associated instrument uses 300-µm ID fluorinated ethylene–propylene copolymer capillary tubing,
and the separation eluent flows directly into a plastic scintillator cell between two PMTs. The scintil-
lation events are detected coincidentally between the two PMTs, such that only if both PMTs receive
an input within a short time will they register the count as signal. This kind of coincidence detection
ensures that nonscintillation photons that come from outside the detection cell and only hit one PMT
are not counted. This system exhibits a detection limit of 16 Bq for 14 C analytes, with a detector
efficiency of 13–15%.
Gamma-emitting radioparticles are typically much higher in energy than beta particles, and
so their detection requires a more dense scintillator material. The CE separation and detection
of radiopharmaceuticals containing 99 Tc was first reported by Altria et al.160 who constructed an
instrument in which the capillary passes through a block of solid scintillator; the solid scintillator
emits photons in response to impinging γ -rays.
A similar design was employed by Pentoney et al.161 in the pursuit of detecting 32 P particles,
using a γ − and β − sensitive Cd-Te detector. The sensitivity of this instrument was augmented by the
use of flow-programming—reducing the voltage while an analyte is in the detection window in order
to increase residence (and hence integration) time. Similar systems were developed by Klunder et
al.162 to detect fissionable materials, in which the radioactivity detector consists of conical plastic
scintillating material, with the capillary passing through the center to provide a near 4π sr detection
geometry. This type of system provides an 80% detection efficiency and detects the decay of 152 Eu
and 137 Cs at the nCi level.
In biological studies, two important radionuclides are 3 H and 35 S, due to their common presence
in small molecules and proteins, respectively. Unfortunately, 3 H and 35 S lack sufficiently energetic
decay products to penetrate the capillary wall and so must be sampled outside the capillary tubing if
they are to be detected. Tracht et al.7,8 developed a post-column method to detect these low-energy
β − particle emitters. In this setup, the end of the capillary is painted with a conductive paint to
maintain an electrical circuit, and the capillary eluent is deposited directly onto a binding membrane
with the aid of a multiaxis translation stage. Scintillation compounds deposited onto the membrane
will luminesce from any radioactive decay present in the sample (Figure 9.7). The intensity can be
measured with a photodetector, or the plate can be read by a phosphorimager. This post-column CE
collection and exposure system has since been used for a number of nanoscale biological analyses,
from analysis of a single neuron163 to observation of synthase activity within a ganglion.164
A major advantage unique to the coupling CE with radionuclide experiments is the significant
reduction in the volume of hazardous waste generated as compared to what would be generated using
other separation methods such as high-pressure liquid chromatography (HPLC). Furthermore, the
abundance of potential applications in the biological sciences and environmental monitoring will
ensure the continued development of this method of analysis in years to come.
Nanovial with
Syringe pump neuron
(with MALDI matrix solution) Separation capillary
Nanovial
injection
Grounding
apparatus
wire
MALDI MS Target
1000 3000 5000 7000

m/z
Mass spectrometry detection Radionuclide detection
FIGURE 9.7 Post-column radionuclide detection system for CE. The spots can be analyzed by other detection
modes, such as mass spectrometry, to gain additional information. (Reprinted from Page, J. S., et al., Anal.
Chem., 74, 497, 2002. With permission.)
9.8 OTHER LIGHT-BASED DETECTION METHODS

The set of optical detection techniques presented thus far in this chapter is not comprehensive;
emphasis has been placed on those that are most common in CE, the most influential to the field,
or both. There are other optical detection modes in various stages of development, each one with
its own unique characteristics. Many of these alternative detection modes aim to provide structural
information about the analyte being detected. One such method uses FTIR detection. FTIR detection
is of interest because it is potentially a nearly universal detection method and provides a wealth of
structural information. The nearly universal nature of IR activity also presents challenges, as the
capillary itself as well as the separation buffer are often IR-absorbing. Thus, most attempts at FTIR
detection for CE have been off-line; the eluting bands are sprayed onto an IR-transparent (e.g., CaF2 )
disk using a glass or metal nebulizer for subsequent (off-line) FTIR analysis.165,166 The instrumen-
tation for online FTIR detection was developed for HPLC before its implementation in CE.167 The
first report of online CE-FTIR appeared some 4 years later, in which adenosine, guanosine, and
adenosine-5-monophosphate were separated, and the online-generated IR spectra were compared
with standard IR spectra.168 The most recent report of CE-FTIR uses mid-IR detection to distinguish
between different sugars in fruit juices; this is the first report of using CE-FTIR to differentiate
compounds in “real” samples.169 In these reports, LODs were in the sub- to low-millimolar (>0.2
mM) range; this technique holds great promise for elucidating structural information from samples,
and future work in this field will likely enhance the performance of this technique for commercial
application.
Raman spectroscopy (RS) is another detection method for CE that is potentially useful for
providing structural information about an eluting analyte. The use of special Raman modes, such
as resonance Raman spectroscopy (RRS) and surface-enhanced Raman spectroscopy (SERS), offer
substantial signal enhancements (102 –1015 ) compared to other Raman processes. RRS and SERS
have been coupled with CE both online and off-line (so-called at-line); an excellent review has been
published recently covering considerations for at-line and online coupling of Raman spectroscopies
to liquid separations, including CE.170 For “at-line” SERS, the CE eluent can be spotted on an
appropriate substrate, such as a thin film of gold, before being subjected to Raman spectroscopic
analysis.171,172 Online CE-SERS analysis has also been reported using a silver colloid solution
(with appropriate electrolytes) as a separation medium.173 RRS was coupled online to CE as early
as 1991. More recently, online UV CE-RRS was used to obtain spectral information from the
separated components of standard mixtures of aromatic sulfonic acids and nucleotides. Through the
enhancement factors of RRS and SERS, and recent work toward the adaptation of deep-UV and
tunable wavelength lasers to CE-RRS, RS has become an intriguing and useful detection method for
obtaining structural information from small-volume separations.
Radio frequency (RF) electromagnetic radiation can be used to induce nuclear spin transitions
in the presence of an external magnetic field, thus enabling detection via NMR. NMR can be used
to both detect and obtain structural information about analytes that are not electroactive and do not
possess a chromophore. Hence, research is ongoing with the aim of improving the online coupling
of NMR to microscale separations. The use of NMR on nanoliter volume samples was first enabled
by the development of the microcoil probe, which consists of a thin copper wire (rf coil) wrapped
around a small diameter (e.g., 50 µm) capillary.174−176 A variety of approaches have been used in
conjunction with the microcoil probe in order to enhance the detected signal. Sample concentration
techniques, such as capillary isotachophoresis, have been used to improve analyte detectability and
enable trace analysis.177 The generation of high-resolution NMR spectra can be accomplished using
periodic stopped-flow CE to increase the spectral acquisition time.11,178 Several methods have been
developed to enable continuous-flow separations and high resolution, including the dual microcoil
split-flow method179 and a “splitless” method.180 For example, in the dual microcoil method, multiple
capillaries are interconnected such that a separated band can be directed to a secondary capillary
and microcoiled, parked, and analyzed, while the separation continues in the primary capillary.179
Several relatively recent reviews have been published that discuss the application of NMR detection
for microscale separations (including CE) in more detail.181,182
Future research will determine whether the potential of these techniques for structure-based
analyte identification can be realized in the form of practicable and profitable detection methods
for CE.
9.9 CONCLUSIONS
As CE continues to evolve, it is not surprising to see a concomitant development of optical detection
methods. Detector innovations have been required to keep up with smaller diameter capillaries, faster
separations, parallel capillary systems, and microfluidic-based electrophoretic formats. In parallel
to these changes, more flexible and higher performance imaging detectors, UV light sources, and
clever applications of these advances continue to push the envelope in terms of light-based detection
for CE.
While this chapter summarizes the state-of-the-art in optical detection, it is perhaps worthwhile
to speculate as to the potential fruits of this continued development, that is, the future of light-based
detection for CE. While photodiode array-based absorbance detection will continue to dominate
CE detection, fluorescence detection will be used for an ever-increasing set of CE-applications.
For example, the widespread use of fluorescence microscopy in biological and biomedical research
ensures that new derivatization chemistries will be developed, allowing a wider range of analytes
to be characterized using fluorescence. Further development of low-cost deep-UV sources is likely
to lead to the proliferation of native fluorescence detection. This, in turn, will be accompanied by
a dramatic increase in the use of multichannel systems to provide information about the identity of
eluting analytes. And of course, solid-state light sources will replace lasers in many applications.
Outside of absorbance and fluorescence, microchip CL systems will become more common,
likely using ECL with in situ regenerated lumophores. Will the next decades finally witness the
commercial development of a capillary-scale Raman probe? Of course, the continued hyphen-
ation of UV-absorbance, mass spectrometry, and other detection schemes will push the envelope
of information-rich detection systems for small-volume separations.
ACKNOWLEDGMENTS
This material is based upon work supported by the National Science Foundation under Award No.
CHE-04-00768, the National Institutes of Health under Award No. R33 DK 070285, and the National
Institute on Drug Abuse under Award No. P30 DA 018310 to the UIUC Neuroproteomics Center.
REFERENCES
1. Khaledi, M. G., High-performance capillary electrophoresis: theory, techniques, and applications, in
Chemical Analysis: A Series of Monographs on Analytical Chemistry and its Applications, 1st ed.,
Winefordner, J. D. (ed.), Wiley-Interscience, Raleigh, 1998, pp. 308.
2. Szulc, M. E. and Krull, I. S., Improved detection and derivatization in capillary electrophoresis,
J. Chromatogr. A, 659, 231, 1994.
3. Burgi, D. S. and Chien, R. L., Optimization in sample stacking for high-performance capillary
4. Beckers, J. L. and Bocek, P., Sample stacking in capillary zone electrophoresis: Principles, advantages
and limitations, Electrophoresis, 21, 2747, 2000.
5. Osbourn, D. M., Weiss, D. J., and Lunte, C. E., On-line preconcentration methods for capillary
6. Urbanek, M., Krivankova, L., and Bocek, P., Stacking phenomena in electromigration: From basic
principles to practical procedures, Electrophoresis, 24, 466, 2003.
7. Tracht, S. E., Cruz, L., StobbaWiley, C. M., and Sweedler, J. V., Detection of radionuclides in capillary
electrophoresis using a phosphor-imaging detector, Anal. Chem., 68, 3922, 1996.
8. Tracht, S., Toma, V., and Sweedler, J. V., Postcolumn radionuclide detection of low-energy beta-
emitters in capillary electrophoresis, Anal. Chem., 66, 2382, 1994.
9. Beale, S. C. and Sudmeier, S. J., Spatial-scanning laser fluorescence detection for capillary
10. Wu, X. Z. and Pawliszyn, J., Whole-column fluorescence-imaged capillary electrophoresis, Elec-
trophoresis, 25, 3820, 2004.
11. Olson, D. L., Lacey, M. E., Webb, A. G., and Sweedler, J. V., Nanoliter-volume 1H NMR detection
using periodic stopped-flow capillary electrophoresis, Anal. Chem., 71, 3070, 1999.
12. Sustacek, V., Foret, F., and Bocek, P., Selection of the background electrolyte-composition with
respect to electromigration dispersion and detection of weakly absorbing substances in capillary zone
13. Foret, F., Fanali, S., Nardi, A., and Bocek, P., Capillary zone electrophoresis of rare earth metals with
indirect UV absorbance detection, Electrophoresis, 11, 780, 1990.
14. Mala, Z., Vespalec, R., and Bocek, P., Capillary zone electrophoresis with indirect photometric
detection in the visible range, Electrophoresis, 15, 1526, 1994.
15. Hogan, B. L. and Yeung, E. S., Determination of intracellular species at the level of a single erythrocyte
via capillary electrophoresis with direct and indirect fluorescence detection, Anal. Chem., 64, 2841,
1992.
16. Kuhr, W. G. and Yeung, E. S., Indirect fluorescence detection of native amino acids in capillary zone
17. Olefirowicz, T. M. and Ewing, A. G., Capillary electrophoresis with indirect amperometric detection,
J. Chromatogr., 499, 713, 1990.
18. Dasgupta, P. K., Eom, I. Y., Morris, K. J., and Li, J., Light emitting diode-based detectors: Absorbance,
fluorescence and spectroelectrochemical measurements in a planar flow-through cell, Anal. Chim. Acta,
500, 337, 2003.
19. Macka, M., Andersson, P., and Haddad, P. R., Linearity evaluation in absorbance detection: The use
of light-emitting diodes for on-capillary detection in capillary electrophoresis, Electrophoresis, 17,
1898, 1996.
20. Boring, C. B. and Dasgupta, P. K., An affordable high-performance optical absorbance detector for
capillary systems, Anal. Chim. Acta, 342, 123, 1997.
21. Xue, Y. J. and Yeung, E. S., On-column double-beam laser-absorption detection for capillary
22. Bruin, G. J. M., Stegeman, G., Vanasten, A. C., Xu, X., Kraak, J. C., and Poppe, H., Optimization and
evaluation of the performance of arrangements for UV detection in high-resolution separations using
fused-silica capillaries, J. Chromatogr., 559, 163, 1991.
23. Bruno, A. E., Gassmann, E., Pericles, N., and Anton, K., On-column capillary-flow cell utilizing
optical wave-guides for chromatographic applications, Anal. Chem., 61, 876, 1989.
24. Foret, F., Deml, M., Kahle, V., and Bocek, P., Online fiber optic UV detection cell and conductivity
cell for capillary zone electrophoresis, Electrophoresis, 7, 430, 1986.
25. Ludi, H., Gassmann, E., Grossenbacher, H., and Marki, W., Capillary zone electrophoresis for the
analysis of peptides synthesized by recombinant DNA technology, Anal. Chim. Acta, 213, 215,
1988.
26. Taylor, J. A. and Yeung, E. S., Axial-beam absorbency detection for capillary electrophoresis,
J. Chromatogr., 550, 831, 1991.
27. Heiger, D. N., Kaltenbach, P., and Sievert, H. J. P., Diode-array detection in capillary electrophoresis,
28. Tsuda, T., Sweedler, J. V., and Zare, R. N., Rectangular capillaries for capillary zone electrophoresis,
Anal. Chem., 62, 2149, 1990.
29. Dubber, M. J. and Kanfer, I., Application of reverse-flow micellar electrokinetic chromatography for
the simultaneous determination of flavonols and terpene trilactones in Ginkgo biloba dosage forms,
J. Chromatogr. A, 1122, 266, 2006.
30. Wu, J. Q. and Pawliszyn, J., Capillary isoelectric-focusing with imaging detection, J. Liquid
Chromatogr., 16, 1891, 1993.
31. Stamler, J. S. and Loscalzo, J., Capillary zone electrophoretic detection of biological thiols and their
S-nitrosated derivatives, Anal. Chem., 64, 779, 1992.
32. Oefner, P. J., Vorndran, A. E., Grill, E., Huber, C., and Bonn, G. K., Capillary zone electrophoretic
analysis of carbohydrates by direct and indirect UV detection, Chromatographia, 34, 308, 1992.
33. Jones, D. H. and Hinman, A. S., Thin-layer ultraviolet visible reflectance spectroelectrochemistry with
a spinning-grating monochromator, Can. J. Chem., 68, 2234, 1990.
34. Kobayashi, S., Ueda, T., and Kikumoto, M., Photodiode array detection in high-performance capillary
35. Sepaniak, M. J., Swaile, D. F., and Powell, A. C., Instrumental developments in micellar electrokinetic
capillary chromatography, J. Chromatogr., 480, 185, 1989.
36. Yeo, S. K., Lee, H. K., and Li, S. F. Y., Separation of antibiotics by high-performance capillary
electrophoresis with photodiode-array detection, J. Chromatogr., 585, 133, 1991.
37. Heiger, D. N., Kaltenbach, P., and Sievert, H. J. P., Diode array detection in capillary electrophoresis,
38. Cruz, L., Shippy, S. A., and Sweedler, J. V., Capillary electrophoretic detectors based on light, in High
Performance Capillary Electrophoresis, Khaledi, M. G. (ed.), Wiley, New York, 1998, p. 303.
39. Lilley, K. A. and Wheat, T. E., Drug identification in biological matrices using capillary electrophoresis
and chemometric software, J. Chromatogr. B: Biomed. Appl., 683, 67, 1996.
40. Li, H., Hou, J., Wang, K., and Zhang, F., Resolution of multicomponent overlapped peaks. A
comparison of several curve resolution methods, Talanta, 70, 336, 2006.
41. Lakowicz, J. R., Principles of Fluorescence Spectroscopy, 3rd ed. Springer, New York, 2006.
42. Larson, A. P., Ahlberg, H., and Folestad, S., Semiconductor laser-induced fluorescence detection in
picoliter volume flow cells, Appl. Opt., 32, 794, 1993.
43. Peck, K., Stryer, L., Glazer, A. N., and Mathies, R. A., Single-molecule fluorescence detection:
Autocorrelation criterion and experimental realization with phycoerythrin, Proc. Natl Acad. Sci. USA,
86, 4087, 1989.
44. Mathies, R. A., Peck, K., and Stryer, L., Optimization of high-sensitivity fluorescence detection, Anal.
Chem., 62, 1786, 1990.
45. Chen, D. Y. and Dovichi, N. J., Single-molecule detection in capillary electrophoresis: Molecular shot
noise as a fundamental limit to chemical analysis, Anal. Chem., 68, 690, 1996.
46. Soper, S. A., Davis, L. M., and Shera, E. B., Detection and identification of single molecules in
solution, J. Opt. Soc. Amer. B-Opt. Phys., 9, 1761, 1992.
47. Goodwin, P. M., Ambrose, W. P., and Keller, R. A., Single-molecule detection in liquids by laser-
induced fluorescence, Acc. Chem. Res., 29, 607, 1996.
48. Zhang, J. Z., Chen, D. Y., Wu, S., Harke, H. R., and Dovichi, N. J., High-sensitivity laser-induced
fluorescence detection for capillary electrophoresis, Clin. Chem., 37, 1492, 1991.
49. Johnson, M. E. and Landers, J. P., Fundamentals and practice for ultrasensitive laser-induced
fluorescence detection in microanalytical systems, Electrophoresis, 25, 3513, 2004.
50. Fu, J. L., Fang, Q., Zhang, T., Jin, X. H., and Fang, Z. L., Laser-induced fluorescence detection system
for microfluidic chips based on an orthogonal optical arrangement, Anal. Chem., 78, 3827, 2006.
51. Jorgenson, J. W. and Lukacs, K. D., High-resolution separations based on electrophoresis and
electroosmosis, J. Chromatogr., 218, 209, 1981.
52. Jorgenson, J. W. and Lukacs, K. D., Zone electrophoresis in open-tubular glass-capillaries, Anal.
Chem., 53, 1298, 1981.
53. Jorgenson, J. W. and Lukacs, K. D., Capillary zone electrophoresis, Science, 222, 266, 1983.
54. Green, J. S. and Jorgenson, J. W., Variable-wavelength on-column fluorescence detector for open-
tubular zone electrophoresis, J. Chromatogr., 352, 337, 1986.
55. Arriaga, E., Chen, D. Y., Cheng, X. L., and Dovichi, N. J., High-efficiency filter fluorometer for
capillary electrophoresis and its application to fluorescein thiocarbamyl amino-acids, J. Chromatogr.
A, 652, 347, 1993.
56. Chabinyc, M. L., Chiu, D. T., McDonald, J. C., Stroock, A. D., Christian, J. F., Karger, A. M.,
and Whitesides, G. M., An integrated fluorescence detection system in poly(dimethylsiloxane) for
microfluidic applications, Anal. Chem., 73, 4491, 2001.
57. Hillebrand, S., Schoffen, J. R., Mandaji, M., Termignoni, C., Henrik Grieneisen, H. P., and Ledur
Kist, T. B., Performance of an ultraviolet light-emitting diode-induced fluorescence detector in capillary
58. Kuo, J. S., Kuyper, C. L., Allen, P. B., Fiorini, G. S., and Chiu, D. T., High-power blue/UV light-emitting
diodes as excitation sources for sensitive detection, Electrophoresis, 25, 3796, 2004.
59. Zhao, S., Yuan, H., and Xiao, D., Optical fiber light-emitting diode-induced fluorescence for capillary
60. Gassmann, E., Kuo, J. E., and Zare, R. N., Electrokinetic separation of chiral compounds, Science,
230, 813, 1985.
61. Higashijima, T., Fuchigami, T., Imasaka, T., and Ishibashi, N., Determination of amino-acids by
capillary zone electrophoresis based on semiconductor-laser fluorescence detection, Anal. Chem., 64,
711, 1992.
62. Williams, D. C. and Soper, S. A., Ultrasensitive near-IR fluorescence detection for capillary gel-
electrophoresis and DNA-sequencing applications, Anal. Chem., 67, 3427, 1995.
63. Shealy, D. B., Lipowska, M., Lipowski, J., Narayanan, N., Sutter, S., Strekowski, L., and Patonay, G.,
Synthesis, chromatographic-separation, and characterization of near-infrared-labeled DNA oligomers
for use in DNA-sequencing, Anal. Chem., 67, 247, 1995.
64. Williams, R. J., Lipowska, M., Patonay, G., and Strekowski, L., Comparison of covalent and noncova-
lent labeling with near-infrared dyes for the high-performance liquid-chromatographic determination
of human serum-albumin, Anal. Chem., 65, 601, 1993.
65. Swaile, D. F. and Sepaniak, M. J., Laser-based fluorometric detection schemes for the analysis of
proteins by capillary zone electrophoresis, J. Liquid Chromatogr., 14, 869, 1991.
66. Timperman, A. T., Oldenburg, K. E., and Sweedler, J. V., Native fluorescence detection and spectral
differentiation of peptides containing tryptophan and tyrosine in capillary electrophoresis, Anal. Chem.,
67, 3421, 1995.
67. Chan, K. C., Muschik, G. M., and Issaq, H. J., Solid-state UV laser-induced fluorescence detection in
68. Chan, K. C., Janini, G. M., Muschik, G. M., and Issaq, H. J., Laser-induced fluorescence detection of
9-fluorenylmethyl chloroformate derivatized amino acids in capillary electrophoresis, J. Chromatogr.,
653, 93, 1993.
69. Miao, H., Rubakhin, S. S., and Sweedler, J. V., Analysis of serotonin release from single neuron soma
using capillary electrophoresis and laser-induced fluorescence with a pulsed deep-UV NeCu laser,
Anal. Bioanal. Chem., 377, 1007, 2003.
70. McNeil, J. R., Johnson, W. L., and Collins, G. J., Ultraviolet laser action in He–Ag and Ne–Ag
mixtures, Appl. Phys. Lett., 29, 172, 1976.
71. Sparrow, M. C., Jackovitz, J. F., Munro, C. H., Hug, W. F., and Asher, S. A., New 224 nm hollow
cathode laser-UV Raman spectrometer, Appl. Spectrosc., 55, 66, 2001.
72. Storrie-Lombardi, M. C., Hug, W. F., McDonald, G. D., Tsapin, A. I., and Nealson, K. H., Hollow cath-
ode ion lasers for deep ultraviolet Raman spectroscopy and fluorescence imaging, Rev. Sci. Instrum.,
72, 4452, 2001.
73. Warner, B. E., Persson, K. B., and Collins, G. J., Metal-vapor production by sputtering in a hollow-
cathode discharge: Theory and experiment, J. Appl. Phys., 50, 5694, 1979.
74. Xu, C., Zipfel, W., Shear, J. B., Williams, R. M., and Webb, W. W., Multiphoton fluorescence
excitation: New spectral windows for biological nonlinear microscopy, Proc. Natl. Acad. Sci. USA,
93, 10763, 1996.
75. Shear, J. B., Multiphoton-excited fluorescence, Anal. Chem., 71, 598A–605A, 1999.
76. Wise, D. D. and Shear, J. B., Quantitation of nicotinamide and serotonin derivatives and detection
of flavins in neuronal extracts using capillary electrophoresis with multiphoton-excited fluorescence,
J. Chromatogr. A, 1111, 153, 2006.
77. Gostkowski, M. L., McDoniel, J. B., Wei, J., Curey, T. E., and Shear, J. B., Characterizing
spectrally diverse biological chromophores using capillary electrophoresis with multiphoton-excited
fluorescence, J. Am. Chem. Soc., 120, 18, 1998.
78. Hernandez, L., Marquina, R., Escalona, J., and Guzman, N. A., Detection and quantifica-
tion of capillary electrophoresis zones by fluorescence microscopy, J. Chromatogr., 502, 247,
1990.
79. Abromson, D. and Bickel, W. S., Fluorescent angular scattering emissions from dye-filled fibers, Appl.
Opt., 30, 2980, 1991.
80. Cheng, Y. F. and Dovichi, N. J., Subattomole amino acid analysis by capillary zone electrophoresis
and laser-induced fluorescence, Science, 242, 562, 1988.
81. Swerdlow, H., Wu, S. L., Harke, H., and Dovichi, N. J., Capillary gel electrophoresis for DNA
sequencing. Laser-induced fluorescence detection with the sheath flow cuvette, J. Chromatogr., 516,
61, 1990.
82. Wu, S. and Dovichi, N. J., High-sensitivity fluorescence detection of fluorescein isothiocyanate
derivatives of amino acids separated by capillary zone electrophoresis, J. Chromatogr., 480, 141,
1989.
83. Chen, D. Y. and Dovichi, N. J., Yoctomole detection limit by laser-induced fluorescence in capillary
electrophoresis, J. Chromatogr. B, 657, 265, 1994.
84. Zhao, J. Y., Dovichi, N. J., Hindsgaul, O., Gosselin, S., and Palcic, M. M., Detection of 100 molecules
of product formed in a fucosyltransferase reaction, Glycobiology, 4, 239, 1994.
85. Chen, D. Y., Swerdlow, H. P., Harke, H. R., Zhang, J. Z., and Dovichi, N. J., Low-cost, high-
sensitivity laser-induced fluorescence detection for DNA sequencing by capillary gel electrophoresis,
J. Chromatogr., 559, 237, 1991.
86. Swerdlow, H., Zhang, J. Z., Chen, D. Y., Harke, H. R., Grey, R., Wu, S. L., Dovichi, N. J., and
Fuller, C., Three DNA sequencing methods using capillary gel electrophoresis and laser-induced
87. Eriksson, K. O., Palm, A., and Hjerten, S., Preparative capillary electrophoresis based on adsorption
of the solutes (proteins) onto a moving blotting membrane as they migrate out of the capillary, Anal.
Biochem., 201, 211, 1992.
88. Cheng, Y. F., Fuchs, M., Andrews, D., and Carson, W., Membrane-fraction collection for capillary
89. Huang, X. H. and Zare, R. N., Continuous sample collection in capillary zone electrophoresis by
coupling the outlet of a capillary to a moving surface, J. Chromatogr., 516, 185, 1990.
90. Mesaros, J. M., Luo, G., Roeraade, J., and Ewing, A. G., Continuous electrophoretic separations in
narrow channels coupled to small-bore capillaries, Anal. Chem., 65, 3313, 1993.
91. Quesada, M. A. and Zhang, S., Multiple capillary DNA sequencer that uses fiber-optic illumination
and detection, Electrophoresis, 17, 1841, 1996.
92. Stokes, D. L., Sepaniak, M. J., and Vo-Dinh, T., Development of a new capillary electrophoresis-based
fibre optic sensor, Biomed. Chromatogr., 11, 187, 1997.
93. Sweedler, J. V., Charge-transfer device detectors and their applications to chemical-analysis, Crit. Rev.
Anal. Chem., 24, 59, 1993.
94. Timperman, A. T., Khatib, K., and Sweedler, J. V., Wavelength-resolved fluorescence detection in
95. Fuller, R. R., Moroz, L. L., Gillette, R., and Sweedler, J. V., Single neuron analysis by capillary
electrophoresis with fluorescence spectroscopy, Neuron, 20, 173, 1998.
96. Lapainis, T., Scanlan, C., Rubakhin, S. S., and Sweedler, J. V., A multichannel native fluorescence
detection system for capillary electrophoretic analysis of neurotransmitters in single neurons, Anal.
Bioanal. Chem., 387, 97–105, 2007.
97. Stuart, J. N., Zhang, X., Jakubowski, J. A., Romanova, E. V., and Sweedler, J. V., Serotonin catabolism
depends upon location of release: characterization of sulfated and gamma-glutamylated serotonin
metabolites in Aplysia californica, J. Neurochem., 84, 1358, 2003.
98. Stuart, J. N., Ebaugh, J. D., Copes, A. L., Hatcher, N. G., Gillette, R., and Sweedler, J. V., Systemic
serotonin sulfate in opisthobranch mollusks, J. Neurochem., 90, 734, 2004.
99. Squires, L. N., Jakubowski, J. A., Stuart, J. N., Rubakhin, S. S., Hatcher, N. G., Kim, W. S., Chen,
K., Shih, J. C., Seif, I., and Sweedler, J. V., Serotonin catabolism and the formation and fate of
5-hydroxyindole thiazolidine carboxylic acid, J. Biol. Chem., 281, 13463, 2006.
100. Karger, A. E., Harris, J. M., and Gesteland, R. F., Multiwavelength fluorescence detection for DNA
sequencing using capillary electrophoresis, Nucleic Acids Res., 19, 4955, 1991.
101. Bright, F. V. and Munson, C. A., Time-resolved fluorescence spectroscopy for illuminating complex
systems, Anal. Chim. Acta, 500, 71, 2003.
102. Miller, K. J. and Lytle, F. E., Capillary zone electrophoresis with time-resolved fluorescence detection
using a diode-pumped solid-state laser, J. Chromatogr., 648, 245, 1993.
103. Ye, L. W., Le, X. C., Xing, J. Z., Ma, M. S., and Yatscoff, R., Competitive immunoassay for
cyclosporine using capillary electrophoresis with laser induced fluorescence polarization detection,
J. Chromatogr. B, 714, 59, 1998.
104. Haughland, R. P., The Handbook—A Guide to Fluorescent Probes and Labeling Technologies, 10th ed.
Molecular Probes, Eugene, 2005.
105. Shippy, S. A., Jankowski, J. A., and Sweedler, J. V., Analysis of trace-level peptides using capillary
electrophoresis with UV laser-induced fluorescence, Anal. Chim. Acta, 307, 163, 1995.
106. Arriaga, E. A., Zhang, Y. N., and Dovichi, N. J., Use of 3-(P-carboxybenzoyl)quinoline-
2-carboxaldehyde to label amino-acids for high-sensitivity fluorescence detection in capillary
electrophoresis, Anal. Chim. Acta, 299, 319, 1995.
107. Oates, M. D., Cooper, B. R., and Jorgenson, J. W., Quantitative amino acid analysis of individual snail
neurons by open tubular liquid chromatography, Anal. Chem., 62, 1573, 1990.
108. Oates, M. D. and Jorgenson, J. W., Quantitative amino acid analysis of subnanogram levels of protein
by open tubular liquid chromatography, Anal. Chem., 62, 1577, 1990.
109. Orwar, O., Fishman, H. A., Ziv, N. E., Scheller, R. H., and Zare, R. N., Use of 2,3-
naphthalenedicarboxaldehyde derivatization for single-cell analysis of glutathione by capillary
electrophoresis and histochemical-localization Ion by fluorescence microscopy, Anal. Chem., 67,
4261, 1995.
110. Gilman, S. D. and Ewing, A. G., Analysis of single cells by capillary electrophoresis with on column
derivatization and laser-induced fluorescence detection, Anal. Chem., 67, 58, 1995.
111. Pentoney, S. L., Huang, X. H., Burgi, D. S., and Zare, R. N., Online connector for microcolumns—
Application to the on-column ortho-phthaldialdehyde derivatization of amino-acids separated by
112. Rose, D. J. and Jorgenson, J. W., Post-capillary fluorescence detection in capillary zone electrophoresis
using ortho-phthaldialdehyde, J. Chromatogr., 447, 117, 1988.
113. Oldenburg, K. E., Xi, X., and Sweedler, J. V., Simple sheath flow reactor for post-column fluorescence
derivatization in capillary electrophoresis, Analyst, 122, 1581, 1997.
114. Ye, M. L., Hu, S., Quigley, W. W. C., and Dovichi, N. J., Post-column fluorescence derivatization
of proteins and peptides in capillary electrophoresis with a sheath flow reactor and 488 nm argon ion
laser excitation, J. Chromatogr. A, 1022, 201, 2004.
115. Albin, M., Weinberger, R., Sapp, E., and Moring, S., Fluorescence detection in capillary
electrophoresis—Evaluation of derivatizing reagents and techniques, Anal. Chem., 63, 417, 1991.
116. Gilman, S. D. and Ewing, A. G., Postcolumn derivatization for capillary electrophoresis using
naphthalene-2,3-dicarboxaldehyde and 2-dercaptoethanol, Anal. Method Instrum., 2, 133, 1995.
117. Zhu, R. and Kok, W. T., Postcolumn reaction system for fluorescence detection in capillary
118. Reinhoud, N. J., Tjaden, U. R., Irth, H., and Vandergreef, J., Bioanalysis of some anthracyclines in
human plasma by capillary electrophoresis with laser-induced fluorescence detection, J. Chromatogr.
B, 574, 327, 1992.
119. Roach, M. C., Gozel, P., and Zare, R. N., Determination of methotrexate and its major metabolite,
7-hydroxymethotrexate, using capillary zone electrophoresis and laser-induced fluorescence detection,
J. Chromatogr. B, 426, 129, 1988.
120. Soini, H., Novotny, M. V., and Riekkola, M. L., Determination of naproxen in serum by capillary
electrophoresis with ultraviolet absorbency and laser-induced fluorescence detection, J. Microcolumn.
Sep., 4, 313, 1992.
121. Quirino, J. P., Dulay, M. T., Fu, L., Mody, T. D., and Zare, R. N., Capillary electrophoresis separation
and native laser-induced fluorescence detection of metallotexaphrins, J. Sep. Sci., 25, 819, 2002.
122. Wu, N., Li, B. H., and Sweedler, J. V., Recent developments in porphyrin separations using capillary
electrophoresis with native fluorescence detection, J. Liquid Chromatogr., 17, 1917, 1994.
123. Burton, D. E., Sepaniak, M. J., and Maskarinec, M. P., Analysis of B6 vitamers by micellar electroki-
netic capillary chromatography with laser-excited fluorescence detection, J. Chromatogr. Sci., 24, 347,
1986.
124. Lee, T. T. and Yeung, E. S., High-sensitivity laser-induced fluorescence detection of native proteins in
125. Mcgregor, D. A. and Yeung, E. S., Detection of DNA fragments separated by capillary electrophoresis
based on their native fluorescence inside a sheath flow, J. Chromatogr. A, 680, 491, 1994.
126. Milofsky, R. E. and Yeung, E. S., Native fluorescence detection of nucleic-acids and DNA restriction
fragments in capillary electrophoresis, Anal. Chem., 65, 153, 1993.
127. Chang, H. T. and Yeung, E. S., Determination of catecholamines in single adrenal-medullary cells by
capillary electrophoresis and laser-induced native fluorescence, Anal. Chem., 67, 1079, 1995.
128. Regnier, F. E., Patterson, D. H., and Harmon, B. J., Electrophoretically-mediated microanalysis
(Emma), Trends Anal. Chem., 14, 177, 1995.
129. Xue, Q. and Yeung, E. S., Differences in the chemical reactivity of individual molecules of an enzyme,
Nature, 373, 681, 1995.
130. Park, Y. N., Zhang, X., Rubakhin, S. S., and Sweedler, J. V., Independent optimization of capillary elec-
trophoresis separation and native fluorescence detection conditions for indolamine and catecholamine
measurements, Anal. Chem., 71, 4997, 1999.
131. Yu, M. and Dovichi, N. J., Attomole amino acid determination by capillary zone electrophoresis with
thermooptical absorbance detection, Anal. Chem., 61, 37, 1989.
132. Saz, J. M., Krattiger, B., Bruno, A. E., Diezmasa, J. C., and Widmer, H. M., Thermooptical absorbency
detection of native proteins separated by capillary Electrophoresis in 10µM ID tubes, J. Chromatogr.
A, 699, 315, 1995.
133. Waldron, K. C. and Dovichi, N. J., Sub-femtomole determination of phenylthiohydantoin-amino acids:
Capillary electrophoresis and thermooptic detection, Anal. Chem., 64, 1396, 1992.
134. Yu, M. and Dovichi, N. J., Sub-femtomole determination of dabsyl-amino acids with capillary zone
electrophoresis separation and laser-induced thermooptical absorbance detection, Mikrochim. Acta, 3,
27, 1988.
135. Bruno, A. E., Paulus, A., and Bornhop, D. J., Thermooptic absorption detection in 25 micrometer
ID capillaries—Capillary electrophoresis of Dansyl-amino acids mixtures, Appl. Spectrosc., 45, 462,
1991.
136. Ragozina, N. Y., Putz, M., Heissler, S., Faubel, W., and Pyell, U., Quantification of etoposide and
etoposide phosphate in human plasma electrokinetic by micellar chromatography and near-field thermal
lens detection, Anal. Chem., 76, 3804, 2004.
137. Ragozina, N. Y., Putz, M., Heissler, S., Faubel, W., and Pyell, U., Quantification of etoposide and
etoposide phosphate in human plasma electrokinetic by micellar chromatography and near-field thermal
lens detection, Anal. Chem., 76, 3804, 2004.
138. Huang, X. and Ren, J., Chemiluminescence detection for capillary electrophoresis and microchip
capillary electrophoresis, Trends Anal. Chem., 25, 155, 2006.
139. Hashimoto, M., Tsukagoshi, K., Nakajima, R., Kondo, K., and Arai, A., Microchip capillary
electrophoresis using on-line chemiluminescence detection, J. Chromatogr. A, 867, 271, 2000.
140. Mangru, S. D. and Harrison, D. J., Chemiluminescence detection in integrated post-separation reactors
for microchip-based capillary electrophoresis and affinity electrophoresis, Electrophoresis, 19, 2301,
1998.
141. Hara, T., Nishida, H., Kayama, S., and Nakajima, R., Capillary zone electrophoretic separation of
protein labeled with rhodamine B isothiocyanate and its on-line chemiluminescence detection, Bull.
Chem. Soc. Japan, 67, 1193, 1994.
142. Hara, T., Okamura, S., Kato, S., Yokogi, J., and Nakajima, R., Separation and detection of protein as
a supramolecular complex using capillary electrophoresis, Anal. Sci., 7, 261, 1991.
143. Hara, T., Yokogi, J., Okamura, S., Kato, S., and Nakajima, R., On-line chemiluminescence detection
of proteins separated by capillary zone electrophoresis, J. Chromatogr., 652, 361, 1993.
144. Liu, Y. M., Liu, E. B., and Cheng, J. K., Ultrasensitive chemiluminescence detection of sub-fM level
Co(II) in capillary electrophoresis, J. Chromatogr. A, 939, 91, 2001.
145. Zhao, S., Xie, C., Lu, X., Song, Y., and Liu, Y. M., A facile and sensitive chemiluminescence detection
of amino acids in biological samples after capillary electrophoretic separation, Electrophoresis, 26,
1745, 2005.
146. Hendrickson, H. P., Anderson, P., Wang, X., Pittman, Z., and Bobbitt, D. R., Compositional analysis of
small peptides using capillary electrophoresis and Ru(bpy)3/3+-based chemiluminescence detection,
Microchem. J., 65, 189, 2000.
147. Tsukagoshi, K., Nakahama, K., and Nakajima, R., Direct detection of biomolecules in a capillary
electrophoresis-chemiluminescence detection system, Anal. Chem., 76, 4410, 2004.
148. Tsukagoshi, K., Shikata, Y., Nakajima, R., Murata, M., and Maeda, M., Analysis of a biopolymer by
capillary electrophoresis with a chemiluminescence detector using a polymer solution as the separation
medium, Anal. Sci., 18, 1195, 2002.
149. Huang, X. and Ren, J., Chemiluminescence detection for capillary electrophoresis and microchip
capillary electrophoresis, Trends Anal. Chem., 25, 155, 2006.
150. Huang, X. J. and Fang, Z. L., Chemiluminescence detection in capillary electrophoresis, Anal. Chim.
Acta, 414, 1, 2000.
151. Forbes, G. A., Nieman, T. A., and Sweedler, J. V., On-line electrogenerated [Ru(bpy)3 ]3+ chemilu-
minescent detection of beta-blockers separated with capillary electrophoresis, Anal. Chim. Acta, 347,
289, 1997.
152. Li, J., Yan, Q., Gao, Y., and Ju, H., Electrogenerated chemiluminescence detection of amino acids
based on precolumn derivatization coupled with capillary electrophoresis separation, Anal. Chem., 78,
2694, 2006.
153. Ding, S. N., Xu, J. J., and Chen, H. Y., Tris(2,2 -bipyridyl)ruthenium(II)-zirconia-Nafion com-
posite films applied as solid-state electrochemiluminescence detector for capillary electrophoresis,
154. Yin, X. B., Dong, S., and Wang, E., Analytical applications of the electrochemiluminescence of tris
(2,2-bipyridyl) ruthenium and its derivatives, Trends Anal. Chem., 23, 432, 2004.
155. Wang, X. and Bobbitt, D. R., In situ cell for electrochemically generated Ru(bpy)33+-based
chemiluminescence detection in capillary electrophoresis, Anal. Chim. Acta, 383, 213, 1999.
156. Chiang, M. T. and Whang, C. W., Tris(2,2-bipyridyl)ruthenium(III)-based electrochemiluminescence
detector with indium/tin oxide working electrode for capillary electrophoresis, J. Chromatogr. A, 934,
59, 2001.
157. Kaniansky, D., Marak, J., Rajec, P., Svec, A., Koval, M., Lucka, M., and Sabanos, G., On-column radio-
metric detector for capillary isotachophoresis and its use in the analysis of C-14-labeled constituents,
J. Chromatogr., 470, 139, 1989.
158. Kaniansky, D., Rajec, P., Svec, A., Havasi, P., and Macasek, F., Online radiometric detection in
capillary isotachophoresis. 1. Preliminary experiments, J. Chromatogr., 258, 238, 1983.
159. Kaniansky, D., Rajec, P., Svec, A., Marak, J., Koval, M., Lucka, M., Franko, S., and Sabanos, G.,
On-column radiometric detector for capillary isotachophoresis, J. Radioanal. Nucl. Chem., 129, 305,
1989.
160. Altria, K. D., Simpson, C. F., Bharij, A. K., and Theobald, A. E., A gamma-ray detector for capillary
zone electrophoresis and its use in the analysis of some radiopharmaceuticals, Electrophoresis, 11,
732, 1990.
161. Pentoney, S. L., Zare, R. N., and Quint, J. F., Online radioisotope detection for capillary electrophoresis,
Anal. Chem., 61, 1642, 1989.
162. Klunder, G. L., Andrews, J. E., Grant, P. M., Andresen, B. D., and Russo, R. E., Analysis of fission
products using capillary electrophoresis with on-line radioactivity detection, Anal. Chem., 69, 2988,
1997.
163. Page, J. S., Rubakhin, S. S., and Sweedler, J. V., Single-neuron analysis using CE combined with
MALDI MS and radionuclide detection, Anal. Chem., 74, 497, 2002.
164. Miao, H., Rubakhin, S. S., Scanlan, C. R., Wang, L. P., and Sweedler, J. V., D-aspartate as a
putative cell-cell signaling molecule in the Aplysia californica central nervous system, J. Neurochem.,
97, 595, 2006.
164. Jarman, J. L., Todebush, R. A., and De Haseth, J. A., Glass nebulizer interface for capillary
electrophoresis—Fourier transform infrared spectrometry, J. Chromatogr. A, 976, 19, 2002.
165. Todebush, R. A., He, L. T., and De Haseth, J. A., A metal nebulizer capillary electrophoresis/fourier
transform infrared spectrometric interface, Anal. Chem., 75, 1393, 2003.
166. Vonach, R., Lendl, B., and Kellner, R., High-performance liquid chromatography with real-time
Fourier-transform infrared detection for the determination of carbohydrates, alcohols and organic
acids in wines, J. Chromatogr. A, 824, 159, 1998.
167. Kolhed, M., Hinsmann, P., Svasek, P., Frank, J., Karlberg, B., and Lendl, B., On-line fourier transform
infrared detection in capillary electrophoresis, Anal. Chem., 74, 3843, 2002.
168. Kolhed, M. and Karlberg, B., Capillary electrophoretic separation of sugars in fruit juices using on-line
mid infrared Fourier transform detection, Analyst, 130, 772, 2005.
169. Dijkstra, R. J., Ariese, F., Gooijer, C., and Brinkman, U. A. T., Raman spectroscopy as a detection
method for liquid-separation techniques, Trends Anal. Chem., 24, 304, 2005.
170. Dijkstra, R. J., Gerssen, A., Efremov, E. V., Ariese, F., Brinkman, U. A. T., and Gooijer, C., Substrates
for the at-line coupling of capillary electrophoresis and surface-enhanced Raman spectroscopy, Anal.
Chim. Acta, 508, 127, 2004.
171. He, L., Natan, M. J., and Keating, C. D., Surface-enhanced Raman scattering: A structure-specific
detection method for capillary, Anal. Chem., 72, 5348, 2000.
172. Nirode, W. F., Devault, G. L., Sepaniak, M. J., and Cole, R. O., On-column surface-enhanced Raman
spectroscopy detection in capillary electrophoresis using running buffers containing silver colloidal
solutions, Anal. Chem., 72, 1866, 2000.
173. Olson, D. L., Peck, T. L., Webb, A. G., Magin, R. L., and Sweedler, J. V., High-resolution microcoil
1H-NMR for mass-limited, nanoliter-volume samples, Science, 270, 1967, 1995.
174. Peck, T. L., Magin, R. L., and Lauterbur, P. C., Design and analysis of microcoils for NMR microscopy,
J. Magn. Reson. B, 108, 114, 1995.
175. Wu, N., Peck, T. L., Webb, A. G., Magin, R. L., and Sweedler, J. V., 1H-NMR spectroscopy on the
nanoliter scale for static and on-line measurements, Anal. Chem., 66, 3849, 1994.
176. Wolters, A. M., Jayawickrama, D. A., Larive, C. K., and Sweedler, J. V., Capillary isotachophore-
sis/NMR: Extension to trace impurity analysis and improved instrumental coupling, Anal. Chem., 74,
2306, 2002.
177. Wolters, A. M., Jayawickrama, D. A., Webb, A. G., and Sweedler, J. V., NMR detection with multiple
solenoidal microcoils for continuous-flow capillary electrophoresis, Anal. Chem., 74, 5550, 2002.
178. Jayawickrama, D. A. and Sweedler, J. V., Dual microcoil NMR probe coupled to cyclic CE for
continuous separation and analyte isolation, Anal. Chem., 76, 4894, 2004.
179. Rapp, E., Jakob, A., Schefer, A. B., Bayer, E., and Albert, K., Splitless on-line coupling of capillary
high-performance liquid chromatography, capillary electrochromatography and pressurized capillary
electrochromatography with nuclear magnetic resonance spectroscopy, Anal. Bioanal. Chem., 376,
1053, 2003.
180. Jayawickrama, D. A. and Sweedler, J. V., Hyphenation of capillary separations with nuclear magnetic
resonance spectroscopy, J. Chromatogr. A, 1000, 819, 2003.
181. Webb, A. G., Nuclear magnetic resonance coupled microseparations, Mag. Reson. Chem., 43, 688,
2005.
10 Microfluidic Devices for
Electrophoretic Separations:
Fabrication and Use
Lindsay A. Legendre, Jerome P. Ferrance, and James
P. Landers
CONTENTS
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335

10.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
10.3 Glass Microchip Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
10.3.1 Photolithography and Wet Etching of Glass Microchips . . . . . . . . . . . . . . . . . . . . . . . . . . 338
10.3.2 Drilling and Dicing Glass Microchips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
10.3.3 Glass Microchip Bonding Techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
10.3.4 Postbonding Cleaning of Glass Microdevices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
10.4 Fabrication of Polymeric Microdevices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
10.4.1 Fabrication of the Master . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
10.4.2 PDMS Device Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
10.4.3 PDMS Bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
10.4.4 PDMS Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
10.5 Performing Microchip Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
10.5.1 Electrophoretic Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
10.5.2 Exemplary Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
10.1 INTRODUCTION
The first miniaturized analytical device reported was a gas chromatographic analyzer fabricated from
silicon in 1979.1 The work showed the separation of a simple mixture of compounds in only a few
seconds. While unperfected, this was an example unlike anything else reported in the literature at that
time, and while the scientific community did not immediately grasp the achievements of Terry and
coworkers, they did see a glimpse of the future provided in this work with respect to the value of minia-
turizing analytical methods. Multiple papers followed this initial work in miniaturization over the next
12 years2−11 ; however, it was not until the concept of a total analysis system (TAS) was first proposed
by Manz et al.12 in 1990 that the concept of miniaturization began to have a significant impact.
The TAS concept was originally motivated by a lack of adequate sensors for the detection
of specific species from a complex mixture. It was hypothesized in this seminal paper12 that, by
improving the sample treatment steps, an ultrasensitive sensor would not be required if interfering
chemical compounds were removed. A TAS, as proposed, entailed initial sampling, transport of
the sample, sample pretreatment steps, and the final detection of the analyte. Miniaturization of all
335
2000
Total
1600
Journal articles
Occurrences
1200 Patents
800
400
0
1990 1992 1994 1996 1998 2000 2002 2004 2006
Publication year
FIGURE 10.1 An increase in microfluidics publications. A search was performed using the term “microflu-
idics” and the number of publications was graphed as a function of publication year. The data shown represent
the total number of journal articles (- - -), patents (· · ·), and the sum of both (—).
of the processes required for total analysis, a micrototal analysis system (µTAS), would allow the
development of a TAS that performs these functions at the site of measurement. Additional benefits
are also derived in decreasing the size of a TAS, that include decreased volumes of required sample,
decreased reagent consumption and the potential for dramatically reduced analysis times.
From its inception, the driving force behind the miniaturization of electrophoretic separation
techniques was related to an enhancement in its analytical performance, rather than a reduction in
size.12 Because separation efficiency is a function of the applied voltage, microfluidic devices utilize
shorter separation lengths along with the application of high voltages typically used in capillary-
based analysis, resulting in fast, efficient separations. In addition, and perhaps most importantly,
microchips also offer the opportunity to integrate sample-processing steps onto the same device used
for separations, producing the µTAS initially described by Manz and coworkers. Process integration
results in a decrease in total analysis time as well as a reduction in potential contamination because
sample handling between processes is eliminated.
Microfluidic devices are becoming increasingly more prevalent in the scientific arena, not only
revealed by a rise in the number of reports in the literature but also by the release of commercial
products. A graph plotting publications versus time, as shown in Figure 10.1, highlights the increased
number of reports using these miniaturized devices since the inception of the original TAS concept.
The goal of this chapter is to educate the novice microchip user on fabrication methods and “tricks of
the trade” regarding glass and polymer [specifically poly(dimethylsiloxane) or PDMS] microdevices
with a view to application of these devices.
10.2 BACKGROUND
The initial experiments transitioning the capillary electrophoretic separation method to a microdevice
were performed by Manz et al.13 This paper, which provides details about the process of fabricating
the devices, describes the four different processes necessary for standard one-mask microfabrication;
these processes included film deposition, photolithography, etching, and bonding. These processes
were used to fabricate a fluidic layer in silicon, which was sealed with a glass cover plate, and then
the device was used for the electrophoretic separation of amino acids. The techniques outlined in
this original paper for microfabrication in glass and silicon are still used today (and detailed below),
along with additional advances required to generate the significantly more complex devices now
being reported in the literature. The goal of this background section is to cover the basic definitions
and concepts, all of which are addressed in more detail in subsequent sections.
Microfluidic Devices for Electrophoretic Separations: Fabrication and Use 337
Film deposition refers to the deposition of materials, such as photoresist and metals, onto the
wafer. These materials can serve as protective layers during etching, or can act as a master for
polymeric devices fabricated using molding. Metal layers are usually deposited using sputtering or
vacuum deposition, with chrome being used most frequently because of its avid bonding to both glass
and silicon. Photoresists are photoactive materials normally coated on wafers using spin coating to
deposit a material onto the surface at a desired thickness. An alternative to conventional spin coating
of photoresist has been developed using a constant-volume injection method,14 which provides an
increased pattern definition of the photoresist structures under the conditions investigated. All of
these methods require cleanroom conditions and extensive instrumentation to provide either the
protective surface that can be patterned for etching or for the development of a master used for
PDMS molding.
Once the photoresist has been deposited, the channel geometry required for the final device
is “imprinted” onto the photoresist. This is normally performed using a photomask containing the
microchip design, through which a ultraviolet (UV) light is passed, exposing the photoresist in
the appropriate areas, to transfer the pattern from the mask onto the photoresist. Photolithogra-
phy, a well-established method from the microelectronics industry, is widely utilized but requires
initial production of the photomask. Recent reports describe methods for patterning photoresist
without a mask, the most promising of which is laser direct-write systems in which the design is
transferred to the photoresist using a laser that travels along the desired pattern on the chip sur-
face creating features based on a schematic contained within the software file used to create the
design.15,16
With the traditional etching method for fabricating glass or silicon devices, the desired pattern
is transferred to the photoresist by exposure of the photoresist to UV light in the pattern designated
by the design. This UV-induced damage to the photoresist makes it susceptible to removal with a
developer, which is essentially a buffered alkaline solution, exposing the underlying metal layers
that cover the wafer. Once these are removed, exposure to hydrofluoric acid (HF) etches the glass
or silicon at a rate that can be predetermined based on the substrate used and the concentration of
HF used in the etch solution. It is noteworthy to mention that HF is a hazardous acid that must
be handled with extreme caution, and appropriate countermeasures must be taken upon exposure.
In another work, a combination of photoresist and sacrificial metal layers were used that have an
increased resistance to HF and, therefore, allowed even deeper glass etching (>300 µm).17 Wet
chemical etching with HF has also been performed without a mask, using patterned PDMS sealed
to the glass with HF flowed through the PDMS channels to etch corresponding channels in the glass
substrate.18 The same concept was also applied to a plastic substrate where an etchant, in place of
HF, was pumped through the PDMS channels.19
Because of the hazards of HF, alternatives to wet chemical etching have also been reported;
Belloy et al.20 used powder blasting to fabricate high-aspect ratio structures in glass. Their tech-
nology uses a beam of powder particles that erodes the protective layers patterned on a substrate;
this technique allowed for fabrication of free-standing monolithic glass microstructures. A similar
method called deep reactive ion etching (DRIE) was used by Ceriotti et al.21 to fabricate rect-
angular channels for UV-detection in fused-silica wafers. DRIE, as well as regular reactive ion
etching, involves bombarding the material to be etched with highly energetic chemically reac-
tive ions, typically fluoride, to collide with the material’s surface and remove surface atoms in
the process. All of these etching processes are used to generate channels with depths in the 1–
500 µm range. Pushing the limits of structure size, Hibara et al.22 fabricated channels with only
nanometer dimensions in fused silica by fast atom beam etching; this method uses accelerated
neutral particles to bombard the surface. An additional method for the generation of nanome-
ter channels was reported by Ionescu et al.,23 who used enzymes to create nanometer-sized
depressions in a protein surface for nanofluidic photolithography. Proteases were delivered with
scanning probe microscopy and were able to precisely control the dimensions of the resulting
features.24
Once channels/chambers have been etched in to the substrate, it is bonded to a flat cover plate to
enclose the channels and complete the fluidic architecture of the microchip. With glass microdevices,
thermal bonding is used most frequently, with heating to a temperature that approaches their glass
transition temperature while under pressure, allowing the glass plates to fuse. For silicon/silicon,
and for glass/silicon bonding, anodic bonding is often employed in which a high voltage is applied
between the top and bottom layers, causing the two to fuse. Both of these bonding procedures
utilize high temperatures, which may not be desirable if metals or other components are an integral
part of the device. Other microchip bonding methods have been reported including techniques
utilizing silicate solutions,25−27 UV-curable adhesive,28,29 and low-temperature methods,30−33 many
of which are carried out at low temperatures (<150◦ C), and, therefore, are ideal for surfaces that
require modification (i.e., metallization) prior to bonding.
10.3 GLASS MICROCHIP FABRICATION

Glass offers some distinct advantages over other substrates that are used to fabricate microdevices.
It is optically-transparent through much of the visible and UV spectrum, its surface chemistry is
reasonably well-characterized, it is compatible with most solvents, and it can withstand high temper-
ature applications due to its high glass transition temperature. One of the most significant drawbacks
associated with using glass as a substrate is the apparent requirement for a cleanroom-based fabrica-
tion facility. However, while cleanroom environments are useful if available, they are not a necessity
as glass devices can be fabricated outside a cleanroom. In line with this, the need to sputter metals
and spin-coat photoresist can be avoided with commercially available silicon and glass wafers that
are precoated with the necessary layers in a variety of sizes and material, in many ways eliminat-
ing the need for sophisticated facilities. For the photolithography step, a laminar flow hood can be
used to protect the surface and mask from dust during the exposure, ensuring exact replication of
the mask image in the photoresist. Yellow lighting is also required to prevent unwanted exposure
of the photoresist, and all the postexposure steps can be performed in an acid-resistant hood (e.g.,
polypropylene). More detailed methods for the fabrication steps, beginning with pre-coated wafers,
are presented in the following sections.
10.3.1 PHOTOLITHOGRAPHY AND WET ETCHING OF GLASS MICROCHIPS

Glass microchips are traditionally fabricated using standard photolithography techniques exploiting
either a film or metal masks (Photosciences, San Jose, CA) for design transfer, and wet chemical
etching for structure creation (Figure 10.2). For both types of masks, the chip image is created
using software (e.g., AutoCAD or Adobe Illustrator) that contains the design for the channels,
chambers, and reservoirs desired in the microfluidic chip. The computer image can be transferred
lithographically onto a film by most commercial printing companies using a high-resolution printer
(∼2500 dpi)—this makes the mask generation inexpensive (a few tens of dollars), allowing the user
to quickly and easily modify the design with successive iteration. While the resolution of the printer
restricts designs to features greater than about 20 µm (width), in most cases, microfluidic channels
range from 40 to 200 µm deep and, thus, film masks prove adequate for most applications. Creating
a metal mask is a more costly endeavor (hundreds of dollars per mask) and is typically done when
either higher resolution structures are required or the user has defined a final design that is tried and
true. This is created through the fabrication of the microfluidic architecture into a glass substrate,
followed by a metallization step to increase the robustness of the mask. While these metal-on-glass
masks are more expensive, they can be used repeatedly without degradation (unless dropped), and
allow for higher resolution structures (<20 µm initial channel width) to be fabricated.
Borosilicate glass has traditionally been used as substrate for glass microchips (e.g., Schott
1.1-mm-thick BOROFLOAT), but quartz34 and soda lime35 glass have also been used. Borosilicate
UV light source
Photomask
Photoresist
Chrome
Glass
Develop photoresist
Glass
Remove chrome
hard bake
Glass
HF etch
Glass
Remove photoresist
remove chrome
Glass
Bond to coverplate
FIGURE 10.2 A diagram illustrating the processes involved in microchip fabrication using standard
photolithography and wet chemical etching methods.
glass plates can be obtained commercially, precoated with sacrificial chromium (<1 µm thickness)
layered underneath a coating of photoresist. To transfer the image from the photomask to the pho-
toresist, a high power (e.g., 350 W) UV exposure system is normally used, with the exposure time
being dependent on the thickness and type of photoresist; a 5 second exposure works well with the
commercially coated plates described above. The UV exposure system can be used with a mask
aligner, but is only essential if exact alignment is required, or if multiple exposures are to be used;
extended low power UV sources can also be sufficient, but could require exposure times. For expo-
sure, the photomask is placed over the coated wafer and the photoresist irradiated with UV light
through the mask. If a film mask is used (flimsy in nature), a clean bare glass plate is placed over it
to obtain intimate contact with the surface of the photoresist (i.e., to hold it flat) and ensure accurate
transfer of the image. The exposed glass is placed in a developer solution and, in the case of a
positive photoresist, the photoresist that was exposed to the UV radiation is removed, revealing the
underlying chromium layer (see Figure 10.3). In contrast, negative photoresist becomes insoluble
where the UV radiation was incident,36 therefore, the photoresist not exposed to the UV would be
removed. In general, either a positive or a negative photoresist can be used as long as the photomask
is designed accordingly. For most photoresists, a “hard bake” is required after the photoresist has
been developed—this stabilizes it for further processing and eliminates additional light sensitivity.
The time and temperature for hard baking varies based on the choice of photoresist, but normally
these range around 110◦ C for 30 min and this can be carried out on a standard laboratory hot plate.
Photomask
Photoresist
Wafer
(a) Positive photoresist (b) Negative photoresist
Align photomask Align photomask

UV-expose UV-expose
Develop photoresist Develop photoresist
FIGURE 10.3 Positive and negative photoresists. An illustration demonstrating the difference between (a)
positive and (b) negative photoresists when used in photolithograpy.
TABLE 10.1
Empirically Determined
Glass Etch Rates Based on
the Amount of HF Used
%HF Etch Rate (µm/min)
49 10
45 8
25 3
10 0.6
The chromium layer exposed by removal of the photoresist is dissolved using a chromium etchant
solution to reveal the underlying glass; all of the solutions used in the photolithography process
(developer, chromium etchant, stripper) are commercially-available from a number of sources. Once
the chrome is removed, the glass plate is immersed in an HF solution, which will etch away the
exposed glass. The etching rate is dependent on the HF concentration of the solution and the type
of glass. Borosilicate glass etches very slowly, and higher concentrations of HF are normally used
with this glass; soda lime glass etches rapidly due to the high concentration of ions in the glass,
and buffered etch solutions utilizing ammonium fluoride are often used to control the etch rate.
The etch rate is also dependent on the final depth required for the features in the device, as long
exposures to HF solution begin to consume the photoresist and chrome protective surfaces, causing
pitting in the areas where etching was not desired. At the same time, etching at excessively rapid
each rates can cause irregular surfaces to be formed on the bottom and walls of the channel, and
other components are often included in the etch solution to ameliorate this effect. An exemplary
etch solution for borosilicate glass is 10% HF/10% HNO3 in water and provides an etch rate of
approximately 0.6 µm/min, which will decrease as the solution is used. Table 10.1 provides a
general guide for achieving different etch rates by varying the concentration of HF. Gentle rocking
during the process also provides a cleaner etched surface, as precipitate forms during etching which
can build up in the channel structures unless removed. For carrying out the etching, costly acid-
resistant containers are often specified (e.g., Nalgene), but better, lower-cost alternatives can be
reusable polypropylene food storage containers; the polypropylene is completely HF resistant, and
these containers have tight-fitting, spill-proof lids that make it easy to store the HF solutions when
not in use. Once the etching is complete, the protective layer of baked photoresist is removed using
a photoresist stripper, and the underlying chrome layer is removed using the chromium etchant.
One important aspect to bear in mind with this process is that glass is isotropically-etched by HF
meaning that the glass structure is etched laterally at the same time it is etched deep—this results in
hemispherical or U-shaped channels. As a result of this phenomenon, side etching of the glass quickly
undercuts the protective chrome layer that has been deposited on top of the substrate, resulting in a
top channel width that is slightly larger than what would be expected. This widening at the top occurs
in concert with depth etching and must be taken into account when determining the depth of the etch
and used accordingly to define the width of the channel designed into the mask. While this places
limitations on the possible dimensions of structures that can be attained, and does not allow for high
aspect ratio structures in finished glass devices, this can be advantageous in two important ways. The
first is that isotropic etching provides a simple and inexpensive method for determining the depth
of the etched structures. A microscope containing an eyepiece reticule can be used to measure the
channel width at the top of the structure (the distance between the chrome flanking the sides of the
structure). The etch depth can be calculated by
Depth = (top width of channel − original width)/2,
where the top width of the channel is measured using the microscope and the original width is
known from the mask design, or it can be measured prior to etching. This method provides an
approximate depth of the channel, and can be used for most applications. Exact measurements of
the etch depth can be obtain using a profilometer in cases where knowing if the depth is critical. A
traditional profilometer uses a stylus tip which is mobilized across the surface; its vertical deflection,
as it “profiles” the surface, provides measurements of surface features in the nano- to micrometer
range. More recently, optical profilometers have been developed that provide similar information
without the need for the stylus. In our hands, fabrication involves estimating an etch rate based
on the chemical ratio of the etch solution or etch rates reported in the literature for the use of etch
solutions of a particular composition. After a specific etch time, the glass plate is removed from the
etch solution, immersed into a large volume of water to terminate the etching, then rinsed thoroughly
with water, and the depth of the structure measured by profilometry. These measurements are used
to empirically determine the etch rate for better control of timing to achieve a specific channel
depth.
The second advantage of the isotropic etching is the ability to easily form weirs—an abrupt
narrowing of the channel—for retaining beads of the appropriate size. Within the channel design
itself, a line across the channel is left unexposed. As the exposed part of the channel etches, the
chrome line across it is undercut from both sides by the side etching of the HF. As long as the line is
less than twice the final depth, it will be completely undercut and removed during the etching process.
Careful predetermination of the width of the line allows the final height of the weir to be controlled
(i.e., when the etching from the two sides connect). A less reproducible method for generating weirs,
but useful for etching different depths on a device with a single exposure, is to cover parts of the
exposed glass with an HF resistant tape during the initial part of the etching process. The deeper
desired features can be etched part way and the tape removed for the remaining etch time.
10.3.2 DRILLING AND DICING GLASS MICROCHIPS

Once the etched layer containing the desired microchannel structures is completed, access holes
at the entrance and exit location of each channel must be fabricated in either the etched plate or a
cover plate. The simplest and most popular method involves drilling holes through the glass with
a 1.1 mm diamond-tipped drill bit using a simple Dremel tool and miniature drill press, both of
which are available at a local hardware store.37 Slight submersion of the glass plate in water while
drilling prevents heating of the drill bit, eliminating wear of the cutting surface and allowing the bits
to be reused multiple times (however appropriate precautions must be taken, as this is a power tool
connected to a high-voltage power source). Silicon carbide drill bits can also be used for drilling
glass and are also readily available in a variety of diameters. More complicated drilling can be
achieved using a computer numerical controlled (CNC) mill, which is a more expensive option for
drilling holes, but may be a good choice if accurate placement of the hole location is required or if a
large number of holes are required in a device.38 A simple program guides the CNC mill to each of
multiple locations on a glass substrate in the X–Y plane, with an accuracy in a simple system down
to ∼5 µm, as well as controlling the speed and drilled depth (the Z-axis). While less efficient on
simple chip designs, or initial prototypes because of the setup time involved (alignment in particular),
it is extremely useful for device designs that contain multiple access holes or designs that will be
utilized multiple times. We have utilized this system to drill holes in a wafer containing 8 chips each
with 20 holes.39 Alternative methods for generating holes in glass substrates are also available, such
as laser ablation,40 but these are mainly used with substrates that are are not amenable to standard
drilling or in situations where very small holes are required.
In addition to drilling, cutting the glass plates is often required. Glass substrates used to fabricate
microchips are commonly 4 –6 in diameter—those etched in our laboratory are normally 5 × 5
in size—but the final dimensions of the devices are significantly smaller. Consequently, each mask
will contain multiple devices to be fabricated in a single glass plate during the etching process.
Each of the individual devices, along with the corresponding cover plates, needs to be diced prior
to the bonding step. Scoring and snapping is the traditional method for cutting glass, which utilizes
a diamond-tipped scoring tool for scoring the glass on one side, followed by application of pressure
to the other side so as to snap it. While this method is simple and easy to employ, it is not foolproof
as the glass does not always break along the scored line, and requires both practice and confidence
to master. A more accurate and reliable way to dice the microchips utilizes a glass-cutting saw
with a thin diamond-edged blade. While wafer saws created for the microelectronics industry are
available for this purpose, a simple and inexpensive option is to use a commercial ceramic tile
saw (available at hardware stores). The blade on a tile saw is normally designed for grinding away
the tile during cutting, but this blade is easily replaced with a thin diamond-edged blade designed
for the more expensive wafer saws. This provides a simple method for cutting the glass wafers
into individual devices without worrying about cracking or breaking the etched plate at unwanted
locations.
10.3.3 GLASS MICROCHIP BONDING TECHNIQUES

The most common glass bonding technique is thermal bonding, where the etched glass chip and
cover plate are properly aligned, brought into intimate contact and placed into a programmable,
high temperature oven where heating to 640–690◦ C for roughly 6–8 hours is required for anneal-
ing the glass. While thermal bonding is simple in principle, it is the slowest and, arguably, the
most unpredictable step in microchip fabrication procedure. There are two main aspects that play
a role in achieving successful thermal bonding, cleanliness and contact. Any organic residue on
the glass plates will prevent them from annealing, thus the surfaces of the etched and cover plates
are often cleaned using piranha wash (3:1 mixture of concentrated H2 SO4 :H2 O2 ) before assembly.
An easier method involves simply wiping the plates with an ammonia-based window cleaner to
thoroughly clean the glass surface; however, use of this method may result in partially unbonded
devices that will never bond completely. Wiping the glass with lint-free wipes prior to assembly
is advantageous as any glass particles that might have adhered to the surface during the glass cut-
ting or drilling procedures will be removed. Care must be taken in wiping the glass plates in this
manner as the edges can be sharp. Once cleaned, the top and bottom must be aligned and pressed
together to achieve intimate contact between the two plates. This is easily detected by the formation
and disappearance of “Newton rings” (rainbow circles) that form when the distance between the
two plates approaches the wavelengths of visible light. Although the float glass is manufactured
to be optically flat, the processing alters the surface enough that significant pressure (under heat)
is required to achieve this intimate contact. This can be achieved with small individual weights or
a press, all the while, careful not to stress the glass to the point that it breaks. Simply squeezing
the glass plates between ones thumbs can often achieve some areas of intimate contact, and a drop
of water to wet the surface and keep the plates aligned during oven loading will also increase the
chances of success. Dust or glass particles between the plates are easily detected after assembly
(Newton rings), at which point, the plates must be disassembled and cleaned. Putting the plates
together in a cleanroom or laminar flow hood assists in avoiding some of these problems, but equal
success can be achieved directly in the laboratory environment. The assembled device (etched and
cover plates) are placed between two polished graphite sheets or ceramic plates in a high tempera-
ture furnace, with or without a weight (∼5 kg) placed on top to apply additional pressure during the
heating process. The bonding cycle is normally performed in four steps, a rapid heating up to about
550◦ C, a slow heating up to the annealing temperature (640–690◦ C), a hold at the annealing tem-
perature (3–10 h), and an unforced cool-down step that allows the glass to cool slowly back to room
temperature.
Multiple bonding cycles are often used to generate devices that are completely bonded. The
annealing temperature in the step program is normally raised 5◦ C with each subsequent bonding
step, as long as the melting temperature of the glass is not exceeded. Two issues to be aware of in the
high temperature bonding procedure are that the graphite or ceramic used to sandwich the fabricated
devices must be polished, otherwise imperfections or roughness in these may be transferred to the
outer surface of the glass devices, giving them a frosted appearance or creating unwanted features
in the surface that can interfere with optical detection and visualization of the channels inside. In
addition, above certain temperatures (e.g., ∼670◦ C with borofloat glass), the glass will fuse to the
ceramic plates during the annealing process. To prevent this, a liquid graphite coating is applied to
the ceramic plates, allowed to dry, then polished flat before the glass chips are placed between them.
In addition to high temperature bonding, a number of other bonding procedures have been
reported (described in the Section 10.2). These procedures are most often used when additional
functionality is to be incorporated into the chip. Most often this is metal sputtered onto the microchip
architecture for electrodes or electrochemical detection, but could also include modification to the
channel surface or incorporation of materials that cannot withstand the high bonding temperatures.
In these cases, alternative cleaning methods may also need to be employed and the protocol for
microchip bonding altered.41
10.3.4 POSTBONDING CLEANING OF GLASS MICRODEVICES

Owing to the high temperatures and long dwell times associated with the thermal bonding steps, the
surface of the glass within the channel becomes dehydrated, which will ultimately affect not only
the electroosmotic flow (EOF) in the channel, but also reactions on the glass surface important for
permanently or dynamically coating the channels. Rehydrating the glass surface involves increasing
the ability to wet the glass usually through the formation of hydroxyl groups in an isolated or
geminal formation; this results in an increase in reproducible electrophoretic separations and allows
for consistent coating of the channels. While many techniques have been suggested as effective for
cleaning glass microchannels, adequate rehydration is required after the devices are bonded, and
a simple rinse with NaOH is not sufficient for this purpose. Cras et al.42 evaluated nine different
popular techniques for glass cleaning and preparing the glass for silanization. We have had success
adapting to borofloat, a technique described by Cras et al. for soda lime glass: a 30-min rinse
with a 1:1 mixture of methanol:concentrated HCl, a thorough rinse with water, a 30-min rinse with
concentrated H2 SO4 , followed by a second rinse with water.
TABLE 10.2
Comparison of Polymers and Glass Properties
Glass PDMS PMMA PI
Polymer type NA Elastomeric Thermoplastic Thermoplastic
Glass temperature (◦ C) 525 −120 106 285
Useful temperature range (◦ C) ≤500 −40 to 50 −70 to 100 −73 to 240
Thermal conductivity (W/mK) 1.2 0.17 to 0.3 0.186 0.2
Visible transmittance (%) >90 91 92 87
Surface charge (native) Yes Weak Yes No
Chemical resistance
Acid Excellent Fair–good Good Fair–good
Solvent Excellent Poor Poor Fair
Alkalis Excellent Poor–fair Excellent Fair–good
The information for the polymer properties was adapted from Sun and Kwok43 and Becker and
Gartner.81 The polymers listed include poly(dimethylsiloxane) (PDMS), polymethyl methacry-
late (PMMA) and polyimide (PI). The information for the glass properties was obtained from
the Schott website, specifically BOROFLOAT® 33.
10.4 FABRICATION OF POLYMERIC MICRODEVICES

The glass fabrication technique detailed above is used extensively, but for many first time microde-
vice researchers, the rapid prototyping technique first described by Whitesides and coworkers49 using
PDMS provides an easy starting point. Moreover, glass is not the ideal substrate for many applica-
tions, and for this reason, a variety of polymers have been utilized for fabrication of microchips. In
general, polymers can be classified into three categories based on their properties: thermoplastics,
thermoset, and elastomers.43 Thermoplastics are a class of polymers that consist of unlinked or
weakly cross-linked chains, which soften and flow when heated above their glass transition temper-
ature (Tg). Examples of these polymers include polymethyl methacrylate (PMMA), polycarbonate
(PC), and polyimide (PI). Thermoset polymers44 contrast thermoplastics in that they are heavily
cross-linked and do not melt or flow upon heating; examples include phenol formaldehyde and vinyl
esters. The last category, elastomeric polymers, consists of weakly cross-linked polymer chains
that form structures, which remain flexible and elastic after polymerization; an increasingly popu-
lar example for microfluidic device fabrication is PDMS. A described list of polymers from each
class that have been used as having utility for microfluidic devices and their associated properties,
can be seen in Table 10.2. The choice of substrate is dictated by a range of variables, such as the
application of interest, the available materials and fabrication, and the solvents/reagents to be used,
to name a few. For example, PI does not have an inherent surface charge and channels in PI would
require surface treatment in order to perform separations requiring a substantial EOF. This section
will focus on PDMS because of its ease of use; chapters in this volume written by DeVoe and Lee
(Chapter 33), Carrilho (Chapter 41), Woolley (Chapter 51) and Henry (Chapter 52) provide details
on the fabrication of microdevices from other polymeric substrates.
Microdevices fabricated in PDMS have been used for electrophoresis45 as well as a variety of
other applications ranging from cell sorting46 to combinatorial screening.47 PDMS is a heterogeneous
polymer exhibiting methyl on the surface, which imparts significant hydrophobicity; its surface
structure is illustrated in Figure 10.4. It is flexible, inexpensive, and optically transparent (down to
280 nm) making it compatible with many optical detection methods.48 Owing to the flexibility of
PDMS, it easily forms reversible bonds with another flat surface through van der Waals interactions.
The seal can easily be broken by peeling the PDMS from the surface intact, thus yielding a single
piece reusable in this manner; irreversible bonds can also be formed between PDMS and a number
of other materials including glass and PDMS itself (discussed in detail below).48,49 Owing to the
CH2 O
O
Si CH2 Si CH3
H3C CH3
O
CH2 Si CH3
O
O CH2
Si CH2 Si CH3
H3C CH3
O
CH2 Si CH3
O
FIGURE 10.4 Chemical structure of PDMS at the surface. (Adapted from http://mrsec.wisc.edu/Edetc/back-
ground/PDMS/index.html)
popularity of PDMS for microfluidic devices, a number of reviews have been written43,48,50 detailing
the procedures for making and using these devices. However, one must be cognizant of the drawbacks
associated with the hydrophobic nature of the polymer.51
Numerous methods for device fabrication, based on the particular polymer selected, have been
reported.43 For fabrication of PDMS microdevices, rapid prototyping49 and replica molding are two
processes that are readily used. These methods allow for fabrication of microdevices in a matter of
hours, with the only drawback being the inability to make channels smaller than 20 µm. For rapid
prototyping, a photoresist is spin coated onto a wafer in the same way it is for as with the glass
fabrication; however, the photoresist for this application must be as thick as the desired depth of the
channels and a photoresist created from SU-8 is used most often for this purpose. Photolithography is
used to transfer the fluidic design from the mask to the photoresist, but in this case, the SU-8 remains
on the wafer where the channels and chambers are desired in the PDMS surface. With this approach
the SU-8 forms a master that can be used repeatedly for replica molding to fabricate devices. This
involves simply casting the polymer over the master to make the fluidic layer that can be sealed to
another piece of PDMS or a glass cover plate.
10.4.1 FABRICATION OF THE MASTER

A common method for fabricating PDMS microfluidic devices is to use a silicon wafer and SU-8
photoresist. To assist in adhesion of the photoresist to the Si wafer, a cleaning step and/or a surface
treatment with hexamethyldisilazane (HMDS) is performed prior to spinning on the photoresist.
Placing the wafer into a 10% (v/v) HF bath for roughly 1–2 min will remove surface silanols, ren-
dering the surface hydrophobic; piranha wash has also been used for this purpose.52 Treatment of
Si wafers with HMDS can be performed simply by rinsing the wafer with methanol, spinning a
small amount of HMDS onto the surface and baking at 105◦ C for 5 min before spinning on the
photoresist. Either method should provide a surface adequate for adhesion between the Si wafer and
SU-8. When spinning the photoresist onto a blank Si wafer, the viscosity of the SU-8 as well as the
speed and duration of the spin will dictate the thickness of the SU-8 layer (and therefore the depth
of the channels in the PDMS). SU-8 is available in a number of formulations designed specifically
for achieving particular photoresist thickness upon spin coating. A soft bake is then performed to
evaporate the solvent from the SU-8 layer and improve the adhesion of the resist to the Si wafer—this
step can be accomplished on a hot plate and is best performed at 95◦ C (the heating time is dependent
on the thickness of the photoresist). Standard photolithography methods are used to transfer the
image from the mask to the photoresist, where the exposed SU-8 is cross-linked by the UV light
to make it insoluble in the developer (propyleneglycolmethylether acetate-PGMEA). The master is
then placed into the developer to remove the uncrosslinked photoresist, leaving the design that is
desired in the PDMS as a raised structure on the surface. The final step involves a hard bake of the
master (done at 105◦ C for at least 15 minutes depending on the thickness of the photoresist) which
stabilizes the photoresist. To assist in removal of cured PDMS from the master, the SU-8 master can
be silanized through using vapor deposition by simply placing it into a chamber with a tridecafluoro-
1,1,2,2-tetrahydrooctyl-1-trichlorosilane solution, essentially making the surface a nonstick-type
coating.
For all steps that involve heating, there are several noteworthy details. Attention must be paid
to the flatness of the surface used for the soft and hard baking steps, particularly if a hot plate is
used. If the surface is not heated evenly, there could be an uneven distribution of the SU-8 layer,
causing unwanted variation in structure depth in the PDMS microchips fabricated from the master.
The transition rates associated with the heating and cooling are also important and should not occur
rapidly. The coefficients of thermal expansion for SU-8 and the Si wafer are very different, and if the
heating/cooling step is expedited, the SU-8 layer could delaminate from the wafer. In our experience
with using a hot plate for heating, it may be best to use a separate hot plate set to 65◦ C for an
intermediate heating step prior to bringing the master to 95◦ C on a second hot plate.52 The master
need only be exposed to 65◦ C for about 1–2 min before transfer to the higher temperature.
This method for fabrication of a master is the traditional method, but does require access to a
cleanroom and instrumentation for spin coating the photoresist to the desired thickness as well as
exposing the photoresist to UV light. A master for PDMS molding can be fabricated out of any
material; however, be aware that there are predefined limitations on the size of the features that
are set by the method used to generate the master. Micromachining capabilities can be used to
generate a master from metal substrates or other polymers (such as PMMA) that work just as well
as the traditional method, but removal of PDMS from the master after the casting process can be an
issue. We have found micromachined Teflon to be an excellent substrate for fabrication of masters,
particularly when it comes to ease of removal of the PDMS. A distinct advantages of this approach
is that the mold can be heated to more rapidly cure the PDMS without destroying or altering the
master in the process.
10.4.2 PDMS DEVICE FABRICATION

Assuming a master of sufficient quality has been created, fabrication of the PDMS microdevice can
be attempted. PDMS is purchased as a two-component system consisting of the elastomeric base
and a curing agent, both in liquid form at room temperature. Sylgard 184 (Dow Corning) is the
most widely-used PDMS formulation, which uses a base:curing agent ratio of 10:1 (by weight as
recommended by the manufacturer). Harder or stickier polymers can be created by varying the ratio
of the two as needed for a particular application. Once the two components are mixed thoroughly,
the mixture should be degassed to remove any air bubbles; this can be accomplished by placing
the PDMS into a vacuum chamber and applying adequate vacuum to gently degas the PDMS. The
master should be placed in a container to hold the liquid PDMS, which is then slowly poured over
the master (taking care to avoid introducing bubbles) and allowed to cure. At room temperature,
at least 18 hours is required for the PDMS to completely crosslink, but this time can be reduced
by increasing the temperature. For example, at 40◦ C the PDMS will cure in less than 4 h, and the
process will only consume 2 hours at 70◦ C. A PDMS cover plate for the device can be fabricated
simultaneously, by pouring additional PDMS over a blank Si wafer and allowing it to cure at the
same time. Once the PDMS has cured, the device can be easily cut to size with a razor blade, and
access holes introduced into the device using a simple hole punch.
10.4.3 PDMS BONDING

Bonding of PDMS devices, in general, is much simpler than the thermal bonding required for glass
devices. To obtain a reversible bond between PDMS and any number of complimentary substrates
(glass, silicon, or another piece of PDMS), a cleaning step must be performed to ensure that the seal
between the fluidic layer and the cover plate is water tight, at least to the extent that it can withstand
the pressures associated with flow through the device. Cleaning the surfaces to assure no dust or
particles are present on either surface is a must, followed by rinsing the PDMS thoroughly with
ethanol and allowing to dry. The layers are then gently pressed together manually, starting at one
corner and making sure no air bubbles are trapped between the two surfaces. Because the bonding is
reversible, if air bubbles do appear, or if particles remain on the surface, the layers can be separated
(by simply peeling them apart), recleaned, and rebonded. A more permanent seal can also be obtained,
both by activation of the PDMS surface using either a plasma oxidizer (plasma cleaner)48,49 or a
UV source that generates ozone near the surfaces to be activated. This activation of the surface is
immediately followed by manually pressing the layers together, allowing the two surfaces to react
and form a permanent bond; curing the devices at 80◦ C for 2.5 hours aids in improving the avid
nature of the bonding.
A clever alternative bonding procedure for sealing PDMS to itself was reported by Quake and
coworkers and involves altering the elastomer:base ratio in the two layers.53 They utilized a ratio of
20:1 for the fluidic layer and 5:1 for the cover layer, so that when the device was oven-bonded, the
two layers fused together as a result of the excess curing agent in one layer and excess base in the
other. This method of bonding completely eliminates the seam between the two layers, negating
the opportunity for fluid leakage between the layers.
10.4.4 PDMS SURFACE

After the PDMS device has been fabricated, the separation channel (or other structures) may require
surface modification to suit the needs of the chemistry desired. Unmodified PDMS is hydrophobic
with no charge,51 therefore yielding a very weak EOF. One approach to decreasing the hydrophobicity
in PDMS surface is to invoke the use of the plasma cleaning method prior to bonding. Through the
introduction of ionizable silanols created by the plasma, this method temporarily renders the surface
hydrophilic, allowing for direct use in this state or providing the option of further surface modification.
However, the user should be aware that this effect is only temporary, lasting several tens of minutes
to a few hours (depending on the system and the oxidation process)—consequently, expediency in
cleaning and assembling is advised.54 For more information about altering the surface of PDMS,
please refer to Chapter 52 by Henry and coworkers.
10.5 PERFORMING MICROCHIP ELECTROPHORESIS

From the earliest inception of the microchip as an analytical platform,1,55 rapid separations were
demonstrated. The major components required for microchip-based electrophoresis (ME) include a
detection system, a power supply, and a computer with programmable software for data collection
and control of the applied voltages and data collection/manipulation; a schematic diagram illustrating
these components can be seen in Figure 10.5. While there are a number of different detection systems
that can be utilized with microchips (see Chapter 45 by Karlinsey and coworkers), a fluorescence
detection system (a homemade fluorescence microscope at its core), is the most widely utilized. There
are several parameters to consider when choosing a power supply, including the number of voltage
outputs required and the range of electric fields that are desired. For a standard “cross-t” chip design,
a minimum of four voltage outputs are required to control the sample injection and separation. As
the complexity of the chip design increases, more outputs may be necessary. One of the benefits of
glass devices is the inherent ability to dissipate heat; consequently, reasonably large voltages can be
applied and dual-polarity, high-voltage supplies are typically used. In addition, the holes drilled into
the microfluidic device may not be large enough to accommodate a sufficient volume of buffer to
allow for electrical contact with the platinum electrodes during the separation. The simplest solution
Power
supply
Microscope
objective
Laser
Optics
PMT
FIGURE 10.5 A schematic diagram of an LIF system. The laser beam is reflected up by a dichroic filter
and focused into the separation channel by a microscope objective. The resultant excitation passes back down
through the dichroic and bandpass filters, and is directed onto the PMT. The computer is used to collect a voltage
output from the PMT proportional to the fluorescence detected. A power supply is used to apply the appropriate
voltages to the microchip reservoirs, to perform electrophoresis.
for obtaining expanded volume reservoirs involves cutting the large ends off of pipette tips and
gluing them onto the chip to act as reservoirs—more recently, commercially-developed reservoirs
for microchips have become available.
A detection system is required to track the progress of the separation, and with the increasing
popularity of microchips for separations, a variety of different detection options are now available. By
far, the most prevalent detection system described in the literature is laser-induced fluorescence (LIF).
Chapter 45 by Karlinsey and coworkers not only provides a guide to selection of the components
for LIF, but also easy-to-follow strategies for alignment of the optics in the LIF system (as well as
reviewing other detection modes used in microfluidics). The components used in an LIF system are
a light detection module, such as a photomultiplier tube or charge-coupled device (CCD) camera,
a laser source, optics for alignment (to guide excitation light in and collect emitted light out) for
efficient detection, a computer with reasonable processing speed for data collection, and a user-
friendly software for interfacing with the hardware. LabVIEW software (National Instruments) is
commonly used as the interface, and allows for easy application-specific computer programming
to provide software interfaces that provide exquisite control over data collection and control of the
detection system.
10.5.1 ELECTROPHORETIC SEPARATIONS

When performing electrophoretic separations, the cross-t design originally proposed by Verheggen
et al.5 for electrokinetic (EK) injection of sample is the most popular; Figure 10.6a shows a typical
cross-t design. Using the appropriate buffer for the separation of interest, the channel surface must
be conditioned in a manner not too dissimilar to those processes used with capillary systems. If the
separation to be performed requires EOF, the rigorous cleaning method suggested in Section 10.3.4
is recommended. All channels of the device are filled with buffer along with the reservoirs to ensure
electrical contact between the platinum electrodes and the solution. A larger volume of buffer in the
reservoirs minimizes problems with buffer depletion which can have adverse effects on the separation
(see Chapter 1 by Landers). Consistent with the scheme in Figure 10.6a, the sample is placed in
the reservoir SI and a voltage applied between SI and SO to allow for mobilization of sample from
reservoir SI to reservoir SO. After the voltage has been applied for a predetermined amount of time,
voltage is applied to the BO and BI reservoirs to electrophorese the sample (down) into the separation
channel, thus, completing the injection process. The detection system, in this case, the incoming
laser light and the photomultiplier tube (PMT), are positioned at a point in the separation channel
(a) (b)
Sample injection Sample injection
BI BI
+ − +
−
SI SO SI SO
Separation
Separation Separation
channel +
+
BO BO
− −
FIGURE 10.6 Schematic representation of microchip electrophoresis device designs (a) shows a traditional
cross-t design while (b) shows an offset cross-t. The reservoirs are labeled as follows: SI is the sample inlet, SO
is sample outlet, BI is buffer inlet, and BO is buffer outlet. A closer view of the cross-t can be seen in the dotted
boxes during sample injection and separation; the sample is in gray and the + and − represent the position and
what type of voltage is being applied.
commensurate with providing the necessary effective length for the separation—the sample analytes
passing that detection point are excited and the resultant fluorescence collected.
For sample injection on a microdevice, a few different modes of EK injection have been reported.
Ramsey and coworkers described a floating injection, in which the BI and BO reservoirs are both
floating during the sample injection; this can lead to a variable volume plug due to diffusion.56 A
“pinched” injection allows for a constant-volume injection by also pumping buffer from the separa-
tion channel along the sample channel, minimizing the length of the plug.56 Another injection mode
involves slightly modifying the device design to incorporate an offset cross-t.57 With this design,
the incoming and outgoing channels are offset relative to the separation channel, thus elongating the
sample plug length to be injected (Figure 10.6b). As with EK injections in capillary systems, there is a
bias toward higher mobility analytes when performing EK injections in a microdevice. For example,
when performing microdevice-based DNA separations with a sieving matrix, EOF is predominantly
suppressed and the applied voltage is used to attract the (anionic) DNA across the sample arm to the
sample outlet. However, long injection times (upwards of 60 seconds) are generally needed to obtain
a representative sample in the cross-t because the mobility of smaller DNA fragments dominates
during the initial phase of the injection process.
10.5.2 EXEMPLARY SEPARATIONS

In 1992, the first microchip-based electrophoretic separation was performed: two fluorescent dyes,
fluorescein and calcein, were separated in a glass/silicon device utilizing LIF as the detection
method.13,58 While the analysis time was not significantly shorter than separations that could be
accomplished using capillary electrophoresis (CE), there was a drastic reduction in the consumption
of sample and reagents (i.e., electrophoresis buffer). Shortly thereafter, faster chip-based sepa-
rations began to emerge in the literature. Seiler et al.59 were able to separate three fluorescein
5-isothiocyanate (FITC)-labeled amino acids—arginine, phenylalanine, and glutamine—in less than
120 seconds using a separation distance of 9.6 cm; the electropherogram and conditions are shown
in Figure 10.7. Ramsey and coworkers60 decreased the separation length to 0.9 mm and were able
to baseline resolve fluorescein and rhodamine B in a fraction of a second (150 milliseconds).
The early success of zone electrophoresis in a microchip gave way to DNAseparations in microde-
vices. Mathies and coworkers were among the pioneers in this section of the microfluidics field,
successfully resolving the restriction fragments of X174 HaeIII in 120 seconds, which ranged
from 70 to 1000 base-pairs in length.61 This technology has since been advanced tremendously, with
Arg Phe
Gin
0 30 60 90 120
Time (s)
FIGURE 10.7 Microchip electropherogram of three 10 µM FITC-labeled amino acids (Arg, Phe, and Gln
in a pH 9.2 carbonate buffer). Electrokinetic injection was performed by applying a voltage of 250 V for 60 s
(∼1-mm plug length). For the separation, a voltage of 10 kV was applied (6.3 kV between injector and detector).
The separation distance from injection to detection was 9.6 cm. (Reprinted from Seiler, K., et al., Analytical
Chemistry, 65, 1481, 1993. With permission.)
multiple groups having accomplished high-resolution separations (including DNA sequencing) and
contributed advancements to this arena.62,63 High-throughput DNA electrophoresis microchips have
also been demonstrated to be functional (shown in Figure 10.8), with Paegel et al.38 having multi-
plexed 96 channels on a radial capillary electrophoresis microchannel plate using a four-color rotary
confocal fluorescence scanner for detection. Tapered turns have been included in the channel design
to gain the effective separation length necessary in order to achieve the high-resolution required for
DNA sequencing separations. On the detection front, Karlinsey and Landers64 have shown the value
of multicolor detection using an acoustooptical tunable filter (AOTF), an approach easily capable of
detecting tens of different colors without any moving parts—they illustrate this with the detection of
biowarfare agents. The electropherogram shown in Figure 10.9 provides an exemplary separation of
a PCR-generated fragment amplified from Bacillus anthracis labeled with a TET dye and co-injected
with a ROX-labeled sizing standard. They also show the utility of the AOTF for rapid channel align-
ment, sample scanning to collect spectral information from the desired analytes, and displayed the
wide range of wavelength detection by performing simultaneous 19 wavelength detection. DNA
separations on microdevices have also been successful for decreasing the analysis time associated
with many post-amplification separations for molecular diagnostics, such as single-stranded con-
formation polymorphism (SSCP) for the detection of breast cancer,65 tandem SSCP/heteroduplex
analysis for detecting p53 mutations,66 and the detection of biowarfare agents using an AOTF to
detect multiple fluorescent signals.64 These topics are covered in more detail in individual chapters
in this book.
Electrophoretic separations involving proteins have also been successfully adapted to microflu-
idic devices utilizing electrophoretic techniques that go beyond capillary zone electrophoresis.
Giordano et al.68 separated partially purified human plasma using a microchip-based capillary
electrophoresis-SDS analysis. Figure 10.10 shows a comparison between the glass microchip
separation with an 11% acrylamide SDS-PAGE gel separation. Traditionally, this separation had
been performed using capillary electrophoresis and UV-detection and suffered from poor detection
limits. Through the use of a dynamic labeling dye, they were able to achieve improved sensitivity
without a loss of resolution. Liu et al.69 modified the separation channels in PMMA microchips using
an atom-transfer radical polymerization to graft polyethylene glycol to the surface. This passivation
Sample
(b)
(mm)
0
(a)
5
10
Cathode
Waste
(c)
FIGURE 10.8 Overall layout of the (a) 96-lane DNA sequencing microchannel plate. (b) expanded view of
the injector. Each doublet features two sample reservoirs and common cathod and waste reservoirs. The arm
from the sample to the separation channel is 85-µm wide, and the arm from the waste to the seapration channel
is 300-µm wide. The separation channel connecting the central anode and cathode is 200-µm wide (c) expanded
view of the hyperturn region. The turns are symmetrically tapered with a tapering length of 100 µm, a turn
channel width of 65 µm, and a radius of curvature of 250 µm. Channel widths and lengths are not drawn to
scale. (Reprinted from Paegel, B. M. et al., Proceedings of the National Academy of Science USA 99, 575, 2002.
With permission.)
7
Time (min)
2
0 250 500
Fluorescence (a.u.)
Size (bases)
3 4 5 6 7 8 9
Time (min)
FIGURE 10.9 Amplified TET-anthrax fragment coinjected electrokinetically with a ROX sizing standard and
separated under denaturing conditions. Data were obtained by the AOTF between emission wavelengths at
18 Hz, and the electropherograms are offset for clarity. Inset: sizing plot with ladder (diamond) and anthrax
(square) peaks. (Reprinted from Karlinsey J. M. and Landers, J. P., Analytical Chemistry 78, 5593, 2006. With
permission.)
step allowed for a separation with increased resolution of the components of bovine serum albumin
(BSA) when compared to an uncoated PMMA micro device; the resultant electropherogram can
be seen in Figure 10.11. Roman et al.70 used micellar-EK chromatography in a PDMS device to
provide high-resolution separations of standard proteins and hydrophobic molecules by adding SDS
as a dynamic channel coating.71 Isoelectric focusing has also been utilized successfully in the micro-
format for the separation of proteins with multiple detection techniques including LIF,72−75 UV
detection76 and whole-column detection.76,77 While most microchip IEF methods have used either
1.9
1.8
1.7
RFU
1.6
1.5
1.4
1.3
1.2
190 240 290 340
Seconds
FIGURE 10.10 Comparison of 11% acrylamide SDS–PAGE gel separation of partially purified human plasma
with microchip-based CE-SDS analysis. Separation conditions include 25 mM Tris-CHES, 1 mM DTT, 0.1%
SDS, 370 V/cm field strength, electrokinetic injection and 0.2% (v/v) NanoOrange (excitation 488 nm and
emission 590 nm). (Reprinted from Giordano, B. C. et al., Analytical Chemistry 76, 4711, 2004. With
permission.)
(a) (b)
0.12 1c 0.12
0.10 0.10
0.08
0.08
0.06 1c
0.06 1a 1b
0.04 1a 1b
0.04
0.02
0 20 40 60 80 0 20 40 60 80 100
Time (s) Time (s)
FIGURE 10.11 Microchip CE of FITC-BSA. (a) Polyethylene glycol (PEG)-grafted and (b) untreated chip.
Peaks 1a, 1b, and 1c are the three main components of the BSA sample. The PEG was grafted onto the PMMA
surface using an atom-transfer radical polymerization. (Reprinted from Liu, J. et al., Analytical Chemistry 76,
EOF- or chemically-induced mobilization of the focused zones past the detector, Guillo et al.75 illus-
trated the use of on-chip pumping, generated by elastomeric diaphragm pumps in a three-layer device
(one PDMS layer sandwiched by two glass layers), to mobilize the zones. The authors explored the
effect of mobilization flow rates on the separation of two amino acids, l-lysine and l-histidine; the
electropherograms can be seen in Figure 10.12.
Other biologically-relevant analytes, in addition to DNA and proteins, have been success-
fully separated on microchip. Figure 10.13 shows the successful electrophoretic separation of
8-aminopyrene-1,3,6-trisulfonate (APTES)-labeled N-linked oligosaccharides from ribonuclease B,
12
10 0.162 µL/min
8 0.097 µL/min
RFU
6
0.058 µL/min
4
0.043 µL/min
2
0.038 µL/min
0
5 10 15 20 25 30 35
Minutes
FIGURE 10.12 Effect of mobilization flow rates on the overall separation performance of l-lysine and l-
histidine. The glass fluidic layer was coated using a modified Hjerten coating. The samples were labeled with
FQ. The focusing step was performed at 7.6 kV for 6 min, and then the mobilization step was initiated. Valve
actuation conditions for mobilization: actuation vacuum: 60 kPa; actuation pressure ranging from 3 to 20 kPa.
(Reprinted from Guillo, C., et al., Lab on a Chip 7, 117, 2007. With permission.)
APTS G3
2500
M5
2000 M6
M7
a M8
1500 M9
INT
1000
500
b
0 10 20 30 40 50 60 70 80 90 100
Time (sec)
FIGURE 10.13 Microchip electropherograms of 8-aminopyrene-1,3,6-trisulfonate (APTS)-labeled N-linked

oligosaccharides from ribonuclease B (A) and APTS-labeled oligosaccharide ladder (B). Experimental condi-
tions: Esep = 300 V/cm, 0.5% methyl cellulose in 20 mM phosphate buffer (pH 6.66). (Reprinted from Dang, F.,
et al., Journal of Chromatography A 1109, 141, 2006. With permission.)
using microchips fabricated in PMMA.78 The authors achieved separation in a channel that had a 30-
mm effective separation length and have also extended this work to include fast profiling of N-linked
complex oligosaccharides released from three other glycoproteins. The separation of lipids, an area
of growing interest, has also been accomplished on microchips. Lin et al.79 reported on the separa-
tion of anionic phosphoinositides on a single-sipper microdevice (Figure 10.14). Once a method for
successful electrophoretic separations was established, the authors used their technology to monitor
the activity of lipid-modifying enzymes. They were able to show the applicability of the method to
high-throughput screening applications by demonstrating phospholipase A2 in a 384-well format;
this technology offers an attractive alternative to current methods that involve radioactive substrates.
In an elegant report, Vrouwe et al.80 showed the ability to detect lithium in blood samples down
A 14
Mixture PI
12
PI PIP
PIP
10 PIP2
PIP2
8
6
1000 1500 2000 2500 3000 3500 4000 4500 5000
Time (sec)
FIGURE 10.14 Separation of anionic phosphoinositides on a single-sipper microchip. Separation conditions:

sample injection time, 0.5 s; buffer time, 800 s; field strength: 530 V/cm; pressure, −1.65 psi (vacuum); lipid
concentration: 1.3 µM each and labeled with BODIPY FL on the acyl chain. (Reprinted from Lin, S., et al.,
Analytical Biochemistry 314, 101, 2003. With permission.)
3
d
c
b
Conductivity (A.U.)
2
a
Mg Li
Na + Ca
0
0 5 10 15 20 25 30
Time (s)
FIGURE 10.15 Electropherogram of an (a) aqueous calibration mixture containing 140 mmol/L sodium
and 2 mmol/L lithium, (b) citrated whole blood enriched with 2 mmol/L lithium, (c) whole blood without
anticoagulant, and (d) heparinized plasma from a patient on lithium therapy containing 0.62 mmol/L lithium.
(Reprinted from Vrouwe, E. X., et al., Clinical Chemistry 53, 121, 2007. With permission.)
to levels of 0.15 mmol/L. Figure 10.15 shows an exemplary separation from samples that included
whole blood, citrated whole blood, and heparinized whole blood while detecting sodium, calcium,
magnesium, and lithium. Through the incorporation of a plexiglass sample collector, the authors
have created a functional device capable of point-of-care analysis. To further challenge their device,
the plexiglass sample collector was tested with a blinded study where they tested five-patient serum
samples on the microchip, and were able to detect lithium concentrations similar to that determined
using ion-selective electrodes.

The purpose of this chapter has been two-fold. First, to highlight some of the basic methods and
substrates that have been successfully utilized to create microfluidic devices. We have attempted to
describe these methods, albeit at a rudimentary level, in a manner that allows for someone skilled in
CE to fabricate, functionalize and execute microchip-based electrophoresis. In addition to the basic
methodologies provided here, the second aim was to provide a few ‘tricks of the trade’ from our own
experience, ones that should enable a novice to create devices of standard cross-t design, or even
more complex and innovative designs tuned to their specific applications. It is noteworthy that, this
is not a comprehensive methods treatise on chip fabrication, and that every microchip fabricator,
regardless of the substrate used, has their own ‘tricks of the trade’—methodology preferences that
enhance their fabrication success with the substrate and design of choice. It is our goal that the
methods given here (and in other chapters in this book) enable the reader to enter into the microchip
fabrication arena where they can demonstrate the application of microfluidics in ways that, to-date,
have not been revealed. But perhaps more important than enhancing the availability of microfluidics
to researchers looking to adapt this analytical platform into their lab, we have enabled the novices
of today to become the microfluidics leaders of tomorrow.
REFERENCES
1. Terry, S. C., Herman, J. H., and Angel, J. B., A gas chromatographic air analyzer fabricated on a
silicon wafer, IEEE Trans. Elec. Dev. 26, 1880, 1979.
2. van der Schoot, B. and Bergveld, P., An ISFET-based microlitre titrator: integration of a chemical
sensor-actuator system, Sensors and Actuators 8, 11–22, 1985.
3. Shoji, S., Esashi, M., and Matsuo, T., Prototype miniature blood gas analyser fabricated on a silicon
wafer, Sensors and Actuators 14, 101–107, 1988.
4. van Lintel, H. T. G., van de Pol, F. C. M., and Bouwstra, S., A piezoelectric micropump based on
micromachining of silicon, Sensors and Actuators 15, 153–167, 1988.
5. Verheggen, T. P. E. M., Beckers, J. L., and Everaerts, F. M., Simple sampling device for capil-
lary isotachophoresis and capillary zone electrophoresis, Journal of Chromatography 452, 615–622,
1988.
6. Esashi, M., Shoji, S., and Nakano, A., Normally closed microvalve and micropump fabricated on a
silicon wafer, Sensors and Actuators 20, 163–169, 1989.
7. van de Pol, F. C. M., Wonnink, D. G. J., Elwenspoek, M., and Fluitman, J. H. J., A thermo-pneumatic
actuation principle for a microminiature pump and other micromechanical devices, Sensors and
Actuators 17, 139–143, 1989.
8. Esashi, M., Integrated micro flow control systems, Sensors and Actuators, A 21, 161–167, 1990.
9. Shoji, S., Nakagawa, S., and Esashi, M., Micropump and sample-injector for integrated chemical
analyzing systems, Sensors and Actuators, A 21, 189–192, 1990.
10. Smits, J. G., Piezoelectric micropump with three valves working peristaltically, Sensors and Actuators,
A 21, 203–206, 1990.
11. van de Pol, F. C. M., van Lintel, H. T. G., Elwenspoek, M., and Fluitman, J. H. J., A thermopneumatic
micropump based on micro-engineering techniques, Sensors and Actuators, A 21, 198–202, 1990.
12. Manz, A., Graber, N., and Widmer, H. M., Miniaturized total chemical-analysis systems—a novel
concept for chemical sensing, Sensors and Actuators B–Chemical 1, 244–248, 1990.
13. Manz, A., Harrison, D. J., Verpoorte, E. M. J., Fettinger, J. C., Paulus, A., Ludi, H., and Widmer, H. M.,
Planar chips technology for miniaturization and integration of separation techniques into monitoring
systems—capillary electrophoresis on a chip, Journal of Chromatography 593, 253–258, 1992.
14. Lin, C. H., Lee, G. B., Chang, B. W., and Chang, G. L., A new fabrication process for ultra-
thick microfluidic microstructures utilizing SU-8 photoresist, Journal of Micromechanics and
Microengineering 12, 590–597, 2002.
15. Paufler, J., Kuck, H., Seltmann, R., Doleschal, W., Gehner, A., and Zimmer, G., High-throughput
optical direct write lithography, Solid State Technology 40, 175–182, 1997.
16. Corbett, S., Strole, J., Johnston, K., Swenson, E. J., and Lu, W. X., Laser direct exposure of pho-
todefinable polymer masks using shaped-beam optics, IEEE Transactions on Electronics Packaging
Manufacturing 28, 312–321, 2005.
17. Bien, D. C. S., Rainey, P. V., Mitchell, S. J. N., and Gamble, H. S., Characterization of masking
materials for deep glass micromachining, Journal of Micromechanics and Microengineering 13, S34–
S40, 2003.
18. Rodriguez, I., Spicar-Mihalic, P., Kuyper, C. L., Fiorini, G. S., and Chiu, D. T., Rapid prototyping of
glass microchannels, Analytica Chimica Acta 496, 205–215, 2003.
19. Brister, P. C. and Weston, K. D., Patterned solvent delivery and etching for the fabrication of plastic
microfluidic devices, Analytical Chemistry 77, 7478–7482, 2005.
20. Belloy, E., Pawlowski, A. G., Sayah, A., and Gijs, M. A. M., Microfabrication of high-aspect ratio
and complex monolithic structures in glass, Journal of Microelectromechanical Systems 11, 521–527,
2002.
21. Ceriotti, L., Weible, K., de Rooij, N. F., and Verpoorte, E., Rectangular channels for lab-on-a-chip
applications, Microelectronic Engineering 67–68, 865–871, 2003.
22. Hibara, A., Saito, T., Kim, H. B., Tokeshi, M., Ooi, T., Nakao, M., and Kitamori, T., Nanochan-
nels on a fused-silica microchip and liquid properties investigation by time-resolved fluorescence
measurements, Analytical Chemistry 74, 6170–6176, 2002.
23. Ionescu, R. E., Marks, R. S., and Gheber, L. A., Nanolithography using protease etching of protein
surfaces, Nano Letters 3, 1639–1642, 2003.
24. Ionescu, R. E., Marks, R. S., and Gheber, L. A., Manufacturing of nanochannels with controlled
dimensions using protease nanolithography, Nano Letters 5, 821–827, 2005.
25. Ito, T., Sobue, K., and Ohya, S., Water glass bonding for micro-total analysis system, Sensors and
Actuators B–Chemical 81, 187–195, 2002.
26. Satoh, A., Water glass bonding, Sensors and Actuators A–Physical 72, 160–168, 1999.
27. Khandurina, J., McKnight, T. E., Jacobson, S. C., Waters, L. C., Foote, R. S., and Ramsey, J. M.,
Integrated system for rapid PCR-based DNA analysis in microfluidic devices, Analytical Chemistry
72, 2995–3000, 2000.
28. Satyanarayana, S., Karnik, R. N., and Majumdar, A., Stamp-and-stick room-temperature
bonding technique for microdevices, Journal of Microelectromechanical Systems 14, 392–399,
2005.
29. Huang, Z. L., Sanders, J. C., Dunsmor, C., Ahmadzadeh, H., and Landers, J. P., A method for
UV-bonding in the fabrication of glass electrophoretic microchips, Electrophoresis 22, 3924–3929,
2001.
30. Chiem, N., Lockyear-Shultz, L., Andersson, P., Skinner, C., and Harrison, D. J., Room temperature
bonding of micromachined glass devices for capillary electrophoresis, Sensors and Actuators B–
Chemical 63, 147–152, 2000.
31. Pigeon, F., Biasse, B., and Zussy, M., Low-temperature pyrex glass wafer direct bonding, Electronics
Letters 31, 792–793, 1995.
32. Jia, Z. J., Fang, Q., and Fang, Z. L., Bonding of glass microfluidic chips at room temperatures,
Analytical Chemistry 76, 5597–5602, 2004.
33. Howlader, M. M. R., Suehara, S., and Suga, T., Room temperature wafer level glass/glass bonding,
Sensors and Actuators A–Physical 127, 31–36, 2006.
34. Nakanishi, H., Nishimoto, T., Arai, A., Abe, H., Kanai, M., Fujiyama, Y., and Yoshida, T., Fabrication
of quartz microchips with optical slit and development of a linear imaging UV detector for microchip
electrophoresis systems, Electrophoresis 22, 230–234, 2001.
35. Lin, C. H., Lee, G. B., Fu, L. M., and Chen, S. H., Integrated optical-fiber capillary electrophoresis
microchips with novel spin-on-glass surface modification, Biosensors and Bioelectronics 20, 83–90,
2004.
36. Shaw, J. M., Gelorme, J. D., LaBianca, N. C., Conley, W. E., and Holmes, S. J., Negative photoresists
for optical lithography, IBM Journal of Research and Development 41, 81–94, 1997.
37. Ferrance, J. P., Wu, Q. R., Giordano, B., Hernandez, C., Kwok, Y., Snow, K., Thibodeau, S., and
Landers, J. P., Developments toward a complete micro-total analysis system for Duchenne muscular
dystrophy diagnosis, Analytica Chimica Acta 500, 223–236, 2003.
38. Paegel, B. M., Emrich, C. A., Weyemayer, G. J., Scherer, J. R., and Mathies, R. A., High through-
put DNA sequencing with a microfabricated 96-lane capillary array electrophoresis bioprocessor,
Proceedings of the National Academy of Sciences of the United States of America 99, 574–579, 2002.
39. Easley, C. J., Karlinsey, J. M., Bienvenue, J. M., Legendre, L. A., Roper, M. G., Feldman, S. H.,
Hughes, M. A., et al., A fully integrated microfluidic genetic analysis system with sample-in-answer-
out capability, Proceedings of the National Academy of Sciences of the United States of America 103,
19272–19277, 2006.
40. Rossier, J. S., Schwarz, A., Reymond, F., Ferrigno, R., Bianchi, F., and Girault, H. H., Microchannel
networks for electrophoretic separations, Electrophoresis 20, 727–731, 1999.
41. Simpson, P. C., Woolley, A. T., and Mathies, R. A., Microfabrication technology for the production
of capillary array electrophoresis chips, Biomedical Microdevices 1, 7–26, 1998.
42. Cras, J. J., Rowe-Taitt, C. A., Nivens, D. A., and Ligler, F. S., Comparison of chemical clean-
ing methods of glass in preparation for silanization, Biosensors and Bioelectronics 14, 683–688,
1999.
43. Sun, Y. and Kwok, Y. C., Polymeric microfluidic system for DNA analysis, Analytica Chimica Acta
556, 80–96, 2006.
44. Holland, A., Cher, C., Rosengarten, G., and Simon, D., Plasma etching of thermoset polymer films
for microchannels, Smart Materials and Structures 15, S104–S111, 2006.
45. Dou, Y. H., Bao, N., Xu, J. J., and Chen, H. Y., A dynamically modified microfluidic
poly(dimethylsiloxane) chip with electrochemical detection for biological analysis, Electrophoresis
23, 3558–3566, 2002.
46. Fu, A. Y., Spence, C., Scherer, A., Arnold, F. H., and Quake, S. R., A microfabricated fluorescence-
activated cell sorter, Nature Biotechnology 17, 1109–1111, 1999.
47. Hansen, C. L., Skordalakes, E., Berger, J. M., and Quake, S. R., A robust and scalable microfluidic
metering method that allows protein crystal growth by free interface diffusion, Proceedings of the
National Academy of Sciences of the United States of America 99, 16531–16536, 2002.
48. McDonald, J. C., Duffy, D. C., Anderson, J. R., Chiu, D. T., Wu, H., Schueller, O. J., and Whitesides,
G. M., Fabrication of microfluidic systems in poly(dimethylsiloxane), Electrophoresis 21, 27–40,
2000.
49. Duffy, D. C., McDonald, J. C., Schueller, O. J. A., and Whitesides, G. M., Rapid prototyping of
microfluidic systems in poly(dimethylsiloxane), Analytical Chemistry 70, 4974–4984, 1998.
50. Sia, S. K. and Whitesides, G. M., Microfluidic devices fabricated in poly(dimethylsiloxane) for
biological studies, Electrophoresis 24, 3563–3576, 2003.
51. Toepke, M. W. and Beebe, D. J., PDMS absorption of small molecules and consequences in
microfluidic applications, Lab on a Chip 6, 1484–1486, 2006.
52. Vickers, J. A. and Henry, C. S., Simplified current decoupler for microchip capillary electrophoresis
with electrochemical and pulsed amperometric detection, Electrophoresis 26, 4641–4647, 2005.
53. Liu, J., Enzelberger, M., and Quake, S., A nanoliter rotary device for polymerase chain reaction,
Electrophoresis 23, 1531–1536, 2002.
54. Vickers, J. A., Caulum, M. M., and Henry, C. S., Generation of hydrophilic poly(dimethylsiloxane)
for high-performance microchip electrophoresis, Analytical Chemistry 78, 7446–7452, 2006.
55. Harrison, D. J., Fluri, K., Seiler, K., Fan, Z. H., Effenhauser, C. S., and Manz, A., Science 261,
895–897, 1993.
56. Jacobson, S. C., Hergenroder, R., Koutny, L. B., Warmack, R. J., and Ramsey, J. M., Effects of
injection schemes and column geometry on the performance of microchip electrophoresis devices,
Analytical Chemistry 66, 1107–1113, 1994.
57. Effenhauser, C. S., Manz, A., and Widmer, H. M., Glass chips for high-speed capillary electrophoresis
separations with submicrometer plate heights, Analytical Chemistry 65, 2637–2642, 1993.
58. Harrison, D. J., Manz, A., Fan, Z. H., Ludi, H., and Widmer, H. M., Capillary electrophoresis and
sample injection systems integrated on a planar glass chip, Analytical Chemistry 64, 1926–1932,
1992.
59. Seiler, K., Harrison, D. J., and Manz, A., planar glass chips for capillary electrophoresis—repetitive
sample injection, quantitation, and separation efficiency, Analytical Chemistry 65, 1481–1488, 1993.
60. Jacobson, S. C., Hergenroder, R., Koutny, L. B., and Ramsey, J. M., High-speed separations on a
microchip, Analytical Chemistry 66, 1114–1118, 1994.
61. Woolley, A. T. and Mathies, R. A., Ultra-high-speed DNA fragment separations using microfabricated
capillary array electrophoresis chips, Proceedings of the National Academy of Sciences of the United
States of America 91, 11348–11352, 1994.
62. Woolley, A. T. and Mathies, R. A., Ultra-high-speed DNA-sequencing using capillary electrophoresis
chips, Analytical Chemistry 67, 3676–3680, 1995.
63. Liu, S. R., Ren, H. J., Gao, Q. F., Roach, D. J., Loder, R. T., Armstrong, T. M., Mao, Q. L., Blaga,
I., Barker, D. L., and Jovanovich, S. B., Automated parallel DNA sequencing on multiple channel
microchips, Proceedings of the National Academy of Sciences of the United States of America 97,
5369–5374, 2000.
64. Karlinsey, J. M. and Landers, J. P., Multicolor fluorescence detection on an electrophoretic microdevice
using an acoustooptic tunable filter, Analytical Chemistry 78, 5590–5596, 2006.
65. Tian, H. J., Jaquins-Gerstl, A., Munro, N., Trucco, M., Brody, L. C., and Landers, J. P., Single-
strand conformation polymorphism analysis by capillary and microchip electrophoresis: a fast, simple
method for detection of common mutations in BRCA1 and BRCA2, Genomics 63, 25–34, 2000.
66. Hestekin, C. N., Jakupciak, J. P., Chiesl, T. N., Kan, C. W., O’Connell, C. D., and Barron, A. E., An
optimized microchip electrophoresis system for mutation detection by tandem SSCP and heteroduplex
analysis for p53 gene exons 5–9, Electrophoresis 27, 3823–3835, 2006.
67. Giordano, B. C., Jin, L. J., Couch, A. J., Ferrance, J. P., and Landers, J. P., Microchip laser-induced
fluorescence detection of proteins at submicrogram per milliliter levels mediated by dynamic Labeling
under pseudonative conditions, Analytical Chemistry 76, 4705–4714, 2004.
68. Liu, J. K., Pan, T., Woolley, A. T., and Lee, M. L., Surface-modified poly(methyl methacrylate)
capillary electrophoresis microchips for protein and peptide analysis, Analytical Chemistry 76, 6948–
6955, 2004.
69. Roman, G. T., Carroll, S., McDaniel, K., and Culbertson, C. T., Micellar electrokinetic chromatog-
raphy of fluorescently labeled proteins on poly(dimethylsiloxane)-based microchips, Electrophoresis
27, 2933–2939, 2006.
70. Roman, G. T., McDaniel, K., and Culbertson, C. T., High efficiency micellar electrokinetic chro-
matography of hydrophobic analytes on poly(dimethylsiloxane) microchips, Analyst 131, 194–201,
2006.
71. Li, Y., DeVoe, D. L., and Lee, C. S., Dynamic analyte introduction and focusing in plastic microfluidic
devices for proteomic analysis, Electrophoresis 24, 193–199, 2003.
72. Tan, W., Fan, Z. H., Qiu, C. X., Ricco, A. J., and Gibbons, I., Miniaturized capillary isoelectric
focusing in plastic microfluidic devices, Electrophoresis 23, 3638–3645, 2002.
73. Han, J. and Singh, A. K., Rapid protein separations in ultra-short microchannels: microchip
sodium dodecyl sulfate-polyacrylamide gel electrophoresis and isoelectric focusing, Journal of
Chromatography A 1049, 205–209, 2004.
74. Guillo, C., Karlinsey, J. M., and Landers, J. P., On-chip pumping for pressure mobilization of the
focused zones following isoelectric focusing, Lab on a Chip 7, 112–118, 2006.
75. Mao, Q. L. and Pawliszyn, J., Demonstration of isoelectric focusing on an etched quartz chip with
UV absorption imaging detection, Analyst 124, 637–641, 1999.
76. Huang, T. M., Ertl, P., Wu, X. Z., Mikkelsen, S., and Pawliszyn, J., Microfabrication of microfluidic
cartridge for isoelectric focusing by screen printing, Sensors and Materials 14, 141–149, 2002.
77. Dang, F. Q., Kakehi, K., Nakajima, K., Shinohara, Y., Ishikawa, M., Kaji, N., Tokeshi, M., and
Baba, Y., Rapid analysis of oligosaccharides derived from glycoproteins by microchip electrophoresis,
Journal of Chromatography A 1109, 138–143, 2006.
78. Lin, S. S., Fischl, A. S., Bi, X. H., and Parce, W., Separation of phospholipids in microfluidic
chip device: application to high-throughput screening assays for lipid-modifying enzymes, Analytical
Biochemistry 314, 97–107, 2003.
79. Vrouwe, E. X., Luttge, R., Vermes, I., and van den Berg, A., Microchip capillary electrophoresis for
point-of-care analysis of lithium, Clinical Chemistry 53, 117–123, 2007.
80. Becker, H. and Gartner, C., Polymer microfabrication methods for microfluidic analytical applications,
Electrophoresis 21, 12–26, 2000.
Part IIA
Capillary-Based Systems: Core
Methods and Technologies
11 Kinetic Capillary
Electrophoresis
Maxim V. Berezovski and Sergey N. Krylov
CONTENTS
11.1 Introduction and Background of Affinity Methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361

11.2 Theory of Kinetic Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
11.3 Kinetic Capillary Electrophoresis Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
11.3.1 NECEEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
11.3.1.1 Affinity-Mediated NECEEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367
11.3.1.2 Temperature-Controlled NECEEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368
11.3.2 SweepCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
11.3.3 ppKCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
11.3.4 Other KCE Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
11.3.5 Multimethod KCE Toolbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
11.4 KCE for Aptamer Selection and Drug Discovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
11.4.1 KCE-Based Selection of Smart Aptamers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
11.4.2 NonSELEX Selection of Aptamers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
11.4.3 The Prospective of KCE in Drug Discovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
11.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
11.1 INTRODUCTION AND BACKGROUND OF AFFINITY

METHODS
Affinity methods play a crucial role in modern life sciences. In addition to affinity purification, their
applications include quantitative analyses of biomolecules, studies of biomolecular interactions,
and selection of affinity probes and drug candidates from complex mixtures, such as combinatorial
libraries. Conceptually, all affinity methods are based on noncovalent binding of a ligand (L) and a
target (T) with the formation of a ligand–target complex (C):
kon
L +T−
←−−
−−−−
−−−−
−−
→− C, (11.1)
koff
where kon and koff are rate constants of complex formation and dissociation, respectively. The
stability of the complex is typically described in terms of the equilibrium dissociation constant:
koff
Kd = (11.2)
kon
Methods for the measurement of equilibrium and rate constants can be classified into two
categories: mixture-based and separation-based (Figure 11.1). The first category includes light
361
Methods for Measurement of

Binding Parameters
Mixture-Based Separation-Based
Homogeneous Heterogeneous Homogeneous Heterogeneous

• UV, VIS, IR • Electrochemical • Ultracentrifugation • Chromatography
Spectroscopy Biosensors • Gel-free • Gel-electrophoresis
• Raman • Surface Plasmon electrophoresis • Ultrafiltration
Spectroscopy Resonance (SPR) • Dialysis
• NMR
• Mass
Spectroscopy
• Calorimetry
FIGURE 11.1 Classification of affinity methods.
absorption, fluorescence spectroscopy, nuclear magnetic resonance (NMR), Fourier transform

infrared spectroscopy (FTIR), mass spectrometry (MS), Raman spectroscopy, potentiometry, and
calorimetry. The separation-based methods include dialysis, ultrafiltration, ultracentrifugation, chro-
matography (liquid chromatography and thin-layer chromatography), and electrophoresis (planar and
capillary electrophoresis). Separation-based methods can detect individual interacting components
and/or complexes, thus avoiding the interference of other components.
Mixture-based and separation-based methods can be subdivided into two broad categories:
heterogeneous and homogeneous binding assays. In heterogeneous assays, T is affixed to a solid
substrate, while L is dissolved in a solution and can bind T affixed to the surface. In an advanced
heterogeneous mixture-based technique such as SPR, T is affixed to a sensor that can change its
optical or electrical signal upon L binding to T.1,2 In SPR, Kd can be found in a series of equilibrium
experiments. The concentration of L in the solution is varied and L and T are allowed to reach
equilibrium. The signal from the sensor versus the concentration of L has a characteristic sigmoidal
shape and Kd can be found from the curve by identifying the concentration of L at which the signal
is equal to half of its maximum amplitude. The koff value can be determined by SPR in a single
nonequilibrium experiment in which the equilibrium is disturbed by fast replacing the solution of L
with a buffer devoid of L. The complex on the surface dissociates in the absence of L in the solution,
and the complex dissociation generates an exponential signal on the sensor.
Heterogeneous binding assays have certain advantages and drawbacks. The most serious draw-
back is that affixing T to the surface changes the structure of T. The extent of such change depends
on the method of immobilization. The change in the structure can potentially affect binding of L to
T. This problem is especially severe for L that binds to T through interaction with a large part of T. In
addition, the immobilization of T on the surface may be time consuming and expensive. Moreover,
nonspecific interactions with the surface are always a concern.
In homogeneous binding assays, T and L are mixed and allowed to form a complex in solution;
neither of the molecules affixed to the surface. Complex formation is followed by either monitoring
the changing physical– chemical properties of L or T upon binding. Such properties can be optical
(absorption, fluorescence, polarization) or separation-related (chromatographic or electrophoretic
mobility). Equilibrium experiments with varying concentrations of L can be used similarly to het-
erogeneous analyses to find Kd . Nonequilibrium stopped-flow experiments, in which L and T are
mixed in a fast fashion and the change in spectral properties is monitored, can be used to find kon .
Separation-based affinity methods can also be classified as kinetic or nonkinetic. Kinetic methods
are those that do not assume equilibrium in reaction 1 and can thus be used for (1) quantitative affinity
Kinetic Capillary Electrophoresis 363
analyses with “weak” affinity probes (high koff ), (2) measuring kon and koff , and (3) selection of
binding ligands with predefined kon and koff . Nonkinetic methods, in contrast, assume equilibrium
and, thus, cannot serve for these tasks. The assumption of equilibrium in nonkinetic methods is
not conceptually required; moreover, equilibrium cannot be maintained in separation-based affinity
methods. Thus, all nonkinetic methods can be converted to kinetic methods by changing conditions
and approaches for data analysis.
In general, homogeneous methods are preferable due to their simplicity and kinetic methods are
preferable due to their enabling kinetic features. Until recently, the only mixture-based heteroge-
neous method with comprehensive kinetic capabilities was SPR. In this chapter, kinetic capillary
electrophoresis (KCE) is described as a conceptual platform for separation-based homogeneous
methods with comprehensive kinetic capabilities, which will find multiple applications in chemistry,
biology, medicine, and drug discovery.
11.2 THEORY OF KINETIC CAPILLARY ELECTROPHORESIS

Kinetic capillary electrophoresis is defined as capillary electrophoretic separation of species that
interact during electrophoresis.3–5 Thus, KCE involves two major processes: affinity interaction of
L and T, described by Equation 11.1, and separation of L, T, and C based on differences in their
electrophoretic velocities, vL , vT , and vC , respectively. These two processes are described by the
following general system of partial differential equations:
∂L(t, x) ∂L(t, x)
+ vL = −kon L(t, x)T (t, x) + koff C(t, x)
∂t ∂x
∂T (t, x) ∂T (t, x)
+ vT = −kon L(t, x)T (t, x) + koff C(t, x)
∂t ∂x
∂C(t, x) ∂C(t, x)
+ vC = −koff C(t, x) + kon L(t, x)T (t, x) (11.3)
∂t ∂x
where L, T , and C are concentrations of L, T, and C, respectively; t is time passed since the beginning
of separation; and x is the distance from the injection end of the capillary. System 11.3 describes
two basic processes, which are always present in KCE. Depending on species studied and a specific
analytical setup, other processes, such as binding with complex stoichiometry, diffusion, adsorption
to capillary walls, and so forth, can play significant roles in KCE. In such cases, mathematical terms,
describing additional processes, must be added to system 11.3. The solution of system 11.3 depends
on the initial and boundary conditions: initial distribution of L, T, and C along the capillary and the
way L, T, and C are introduced into the capillary and removed from the capillary during separation.
This solution can be found nonnumerically for specific sets of initial and boundary conditions and
specific assumptions.6–9 For KCE to be a generic approach, it is required that system 11.3 be solved
for any set of conditions; such solutions can be found only numerically.
11.3 KINETIC CAPILLARY ELECTROPHORESIS METHODS

Every set of qualitatively unique initial and boundary conditions for system 11.3 defines a unique
KCE method. Here we will describe six KCE methods: nonequilibrium capillary electrophoresis of
the equilibrium mixtures (NECEEM), sweeping CE (SweepCE), continuous NECEEM (cNECEEM),
short SweepCE (sSweepCE), plug-plug KCE (ppKCE), and short SweepCE of Equilibrium Mixture
(sSweepCEEM). Table 11.1 contains drawings that schematically illustrate initial and boundary
conditions and show the mathematical representation of initial and boundary conditions. It also
contains representative functions L(t), T (t), and C(t) for a fixed x for each method. The notion of
equilibrium mixture (EM) refers to the mixture of L, T, and C at equilibrium, typically prepared
TABLE 11.1
Summary of KCE Methods
KCE Schematic representation of initial Simulated

method and boundary conditions concentration profiles
Detection
Ligand
Inlet Outlet
Target
Capillary
Complex
C (t )
NECEEM EM L (t )
T (t)
ppKCE T L
SweepCE T L
cNECEEM EM
E
sSweepCE T L
sSweepCEEM T EM
Migration time to the detector
outside the capillary. The concentrations of the three components (T̃ , L̃, and C̃) in the EM are
interconnected through the equilibrium dissociation constant, Kd , as Kd = (T̃ L̃)/C̃.
11.3.1 NECEEM
Nonequilibrium capillary electrophoresis of the equilibrium mixtures (NECEEM) is a unique KCE
method that facilitates finding both Kd and koff from a single electropherogram.6 Mechanistically,
it is similar to previously introduced affinity probe capillary electrophoresis (APCE)10–12 and falls
into a broad category of capillary zone electrophoresis (CZE).13–18 APCE was introduced as being
applicable to systems having suitably stable complexes where the dissociation rate constants are small
in the scale of characteristic separation times. In other words, the characteristic time of dissociation
is much longer than the characteristic separation time. That is why, APCE has been applied only to
measuring equilibrium constants but not rate constants.
So far, NECEEM was used to study the interaction between several proteins and DNA such as an
Escherichia coli single-stranded DNA-binding protein (SSB) and a fluorescently labeled oligonu-
cleotide (ssDNA),6,7,19 Taq DNA polymerase and its aptamer,19 thrombin and its aptamer,20 Tau
protein and single-stranded and double-stranded DNA,21 protein farnesyltransferase (PFTase) and
its aptamer,22 MutS protein and its aptamer,23 h-Ras protein and its aptamer,24,25 and Mef2c pro-
tein and double-stranded DNA it naturally binds.26 It has also been used to study protein–peptide
interactions.12
In NECEEM, a short plug of the EM is injected into the inlet of the capillary, which is prefilled with
the run buffer. Separation is carried out with both inlet and outlet reservoirs containing the run buffer
only. C continuously dissociates during electrophoresis. If separation is efficient, re-association of T
and L can be neglected. The resulting NECEEM electropherograms contain three peaks of T, C, and
L and two exponential “smears” of L and T, which occur from the dissociation of C whose migration
times and areas are used to calculate koff and Kd (Figure 11.2). The important feature of NECEEM is
that a single electropherogram contains data sufficient for finding both Kd and koff . NECEEM starts
with the EM; therefore, it has a memory of the equilibrium necessary for finding Kd . Although a single
electropherogram is sufficient to find both Kd and koff , for accurate measurements of the constants
and their experimental errors several experiments have to be performed. The essential feature of
NECEEM electropherograms is that the areas of peaks and smears in them are proportional to the
amounts of corresponding species. A single NECEEM electropherogram can be used for finding four
measurable parameters required for the determination of Kd and koff (Figure 11.3). A1 is the area of
the peak corresponding to L, which was free in the EM. A3 is the area of the exponential smear left
by L dissociated from C during the separation. A2 is the area of the peak corresponding to C, which
remained intact by the time of passing the detector. Finally, tC is the migration time of the complex.
The values of Kd and koff can be calculated using the following algebraic formulas8 :
[T ]0 (1 + A1 /(A2 + A3 )) − [L]0
Kd = , (11.4)
1 + (A2 + A3 )/A1
ln ((A2 + A3 )/A2 )
koff = , (11.5)
tC
where [T ]0 and [L]0 are total concentrations of T and L in the EM, which include free components
and complexes. Advantageously, areas and migration time associated with a single species only
(L in our example) are required. This simplifies the use of fluorescence detection because finding
a strategy for labeling a single species is relatively easy. A major step in the method development
for NECEEM involves finding conditions for good quality separation of L from C. Figure 11.3
shows an experimental NECEEM electropherogram. In this example, interaction between ssDNA
and ssDNA-binding protein was studied. In the experimental electropherogram, the peaks have
Gaussian-type shapes rather than rectangular ones presented in a schematic electropherogram in
Figure 11.2. While measuring the areas, it is important to accurately define the boundary between
them. The boundary between A1 and A3 can be found by comparing the peak of free L in the presence
and absence of T. It was shown that the uncertainty in defining the boundaries between the areas
leads to experimental errors in the range of 10%. This is an acceptable level of experimental errors
for most applications. Alternatively, mathematical modeling of a NECEEM electropherogram can
be used to find both Kd and koff from the nonlinear regression analysis without the need to define
the areas. Typically, the area method is used as a simple, fast, and acceptably accurate approach.8
(a) EM = L + C + T
(b) Direction of the migration in electrophoresis
t0 EM
t1 L C T
t2 L L dissociated from C C T dissociated from C T
(c)
Concentration
L T dissociated
L dissociated T
from C from C
C
Position in the capillary at time t2
FIGURE 11.2 Conceptual representation of nonequilibrium capillary electrophoresis of equilibrium mixtures

(NECEEM). (a) Components of the equilibrium mixture (EM): free ligand (L), free target (T), and the ligand-
target complex (C). (b) NECEEM-based separation of L, T, and C. A short plug of the equilibrium mixture is
injected into a capillary at time t0 . High voltage is then applied. It is assumed that T migrates faster than L; C
typically has an intermediate mobility. Equilibrium fractions of free L and T migrate as individual zones, which
do not change in time. The equilibrium fraction of C continuously dissociates during separation (time t1 and t2 ),
leaving smears of L and T. By time t2 , only a fraction of C remains intact. (c) The graph shows concentrations
of the separated components as functions of the position in the capillary at time t2 .
0.4
0.80 0.3
Fluorescence signal
0.60 0.2 Exponential decay
0.1
0.40
200 240 280 320
0.20
100 200 300 400

tC Migration time (s) tL
FIGURE 11.3 Example of a NECEEM electropherogram for the interaction between fluorescently labeled
ssDNA (ligand) and single-stranded binding (SSB) protein (target). tC and tL are the migration time of the
complex and the ligand, respectively.
NECEEM-based determination of Kd and koff is fast and accurate, and it has a wide and adjustable
dynamic range. The upper limit of Kd values depends on the highest concentration of T available
and can be as high as millimolar. This allows for the measurement of Kd values for very low-bulk
affinities of naive combinatorial libraries.24 The lower limit of Kd depends on the concentration limit
of detection; for fluorescence detection, it can be as low as picomolar. The dynamic range of koff
values is defined by the migration time of the complex, which can be easily regulated by the length
of the capillary, electric field, or electro-osmotic velocity. The practically proven dynamic range of
koff spans from 10−4 to 1 s−1 .6,8,12,19–21,27–29 Although only one electropherogram is required for
finding both Kd and koff , the concentration of T (if L is used as a detectable species) should be within
an order of magnitude from the Kd value. Titration of T with 10-time increments in concentration is
recommended as the fastest way of finding suitable T. Furthermore, conducting several experiments
may be required to find the experimental deviation of the Kd value. The equilibrium is typically
established in the incubation buffer, whereas dissociation occurs in the electrophoresis run buffer.
The values of Kd and koff are, thus, measured for the incubation buffer and run buffer, respectively. If
the incubation buffer and the electrophoresis run buffer are identical, then Kd and koff are determined
under the same conditions, and kon can be calculated as kon = koff /Kd . It is typically possible to
make the incubation and run buffers identical. An example of when such matching is difficult
is when T is the protein, which requires a high salt concentration. CE cannot tolerate high salt
concentrations in the run buffer because of the high Joule heating, which can deteriorate the quality of
separation.
11.3.1.1 Affinity-Mediated NECEEM

To explain the rationale for affinity-mediated NECEEM, it needs to be emphasized that NECEEM
requires good separation of free ligand from the target–ligand complex. If the separation is poor,
the accuracy of the method with respect to the determination of rate constants and equilibrium
constants decreases. Affinity-mediated NECEEM is based on the insight that adding to the run
buffer a background affinity agent, which can bind free ligand but not the target–ligand complex,
can improve the separation by changing the mobility of free ligand while not affecting that of the
complex. Affinity-mediated NECEEM was demonstrated for interaction between thrombin and its
DNA aptamer by using the single-stranded binding (SSB) protein from E. coli as a background
affinity agent in the run buffer (Figure 11.4).20,30 Hypothetically, affinity-mediated NECEEM can
be also realized with a target-binding affinity agent, such as an antibody, instead of a ligand-binding
agent, provided that the agent binds the target but does not bind the target–ligand complex. A serious
assumption used in affinity-mediated NECEEM is that the affinity agent does not affect the interaction
between the target and a ligand.
Fluorescence signal (V)
0.4
A DNA + 11 C [Apt]eq A1
R? ?
Thrombin DNA [Apt Thr]eq A2
0.3
4 9
Fluorescent signal
[Thr]o=Kd/R + [Apt0/(1+R)
3 [SSB] = 0 7 0.2
LOD: 60 nM and 4 million molecules

2 B DNA 5 A1 A2
0.1
Decay of
3
1 Thrombin–DNA tApt 4 6 8 10
[SSB] = 100 n M
1 Migration time (min)
0.0
0 1 2 3 4 5 6 7 8 9 3’- GGTTGGTGTGGTTCGCGCCTCGCACCGTCC
Migration time (min) Fluorescein - GCGGAGCGTGGCAGG
FIGURE 11.4 Affinity-mediated NECEEM. (a) Separation of thrombin and its aptamer (DNA) in the absence
of SSB in the run buffer. (b) Separation of thrombin and DNAin the presence of SSB. (c) Quantitative detection of
thrombin without observing the peak of the complex and just uses a decay of the complex. The structure of
the aptamer (bottom of c) for human thrombin (boldface) with an additional 16-nucleotide sequence and a
fluorescein-labeled probe that is complementary to the additional 16mer sequence. Complementary strands are
given in italics.
11.3.1.2 Temperature-Controlled NECEEM

Temperature-controlled NECEEM was developed to study thermochemistry of target–ligand
interactions (protein–DNA interaction as one of the examples).19 Knowing how temperature influen-
ces kinetic and equilibrium binding parameters of noncovalent protein–DNA interactions is
important for understanding fundamental biological processes, such as gene expression and DNA
replication.31–35 It is also essential for developing analytical applications of DNA aptamers and
DNA-binding proteins in affinity and hybridization analyses and in optimizing the polymerase
chain reaction (PCR).30,36–38 Conventional methods for thermo-chemical studies of protein–DNA
interactions have limitations. Differential scanning calorimetry (DSC) and isothermal titration
calorimetry (ITC) are not applicable to finding kinetic parameters.39–41 SPR can serve to deter-
mine equilibrium and kinetic parameters but it is a heterogeneous method, which requires the
immobilization of either DNA or protein on the surface of a sensor.42,43 Being a homogeneous
kinetic method, temperature-controlled NECEEM uniquely allows finding temperature dependen-
cies of equilibrium and kinetic parameters of complex formation without the immobilization of the
interacting molecules on the surface of a solid substrate. Moreover, it requires much lower quan-
tities of the protein than DSC, ITC, and SPR. Two protein–DNA pairs: (1) Taq DNA polymerase
with its DNA aptamer and (2) E. coli single-stranded DNA-binding protein with a 20-base long
ssDNA were analyzed by temperature-controlled NECEEM (Figure 11.5).19 Temperature depen-
dencies of three parameters were determined: the equilibrium binding constant (Kb = 1/Kd ),
the rate constant of complex dissociation (koff ), and the rate constant of complex formation
(kon ). The Kb (T ) functions for both protein–DNA pairs had phase-transition-like points suggest-
ing temperature-dependent conformational changes in structures of the interacting macromolecules.
Temperature dependencies of kon and koff provided insights into how the conformational changes
affected two opposite processes: binding and dissociation. Finally, thermodynamic parameters,
H and S, for complex formation were found for different conformations. With its unique
features and potential applicability to other macromolecular interactions, temperature-controlled
NECEEM establishes a valuable addition to the arsenal of analytical methods used to study dynamic
molecular complexes.
Additionally, there is a bonus application of temperature-controlled NECEEM: a nonspectro-
scopic approach to determining temperature in CE (Figure 11.6).44 It is based on measuring a
temperature-dependent rate constant of complex dissociation by NECEEM. This work was demon-
strated on the dissociation of a protein–DNA complex to show that the new method allows
T, C
Binding parameters, arbitrary
60 50 40 30 20 10
kon 21
20
H = 1.0 kcal/mo
19 S = 43 cal/(mol K)
koff
b
LnK
18
Conformational
changes 17 H = -30 kcal/mo
Kd S = -58 cal/(mol K)
16
0.0030 0.0032 0.0034 0.0036
15 25 35 45 55
Temperature (°C) 1/T, K−1
FIGURE 11.5 NECEEM-measured temperature dependencies of equilibrium and kinetic binding parameters
and Van’t Hoff plots used for the determination of H and S for complex formation for Taq DNA polymerase
with its DNA aptamer.
6 0.03
Rate constant of complex dissociation (s )

−1
Intact complexes
reaching the
5 detector
Fluorescence intensity (a.u.)
4 0.02
288 K
3
293 K
2 298 K 0.01
303 K
1 308 K
313 K
0 0
1 3 5 7 9 285 290 295 300 305 310 315
Migration time to the detector (min) Temperature (K)
FIGURE 11.6 Temperature dependence of NECEEM electropherograms for SSB-ssDNA interaction (left
panel), and calibration curve “koff versus T” (right panel).
for temperature determination in CE with a precision of 2◦ C. With a number of advantages over

conventional spectroscopic approaches, NECEEM-based temperature determination will find practi-
cal applications in CE method development for temperature-sensitive analyses, such as hybridization
and affinity assays.
11.3.2 SWEEPCE
The monomolecular rate constant of complex dissociation, koff , can be determined by either SPR,45,46
or by NECEEM.6–8 As for the bimolecular rate constant of complex formation, kon , until now, the only
technique available for its direct measurements was stopped-flow spectroscopy.47–49 Stopped-flow
spectroscopy relies on the change of spectral properties of either protein or DNA during com-
plex formation. Such changes are often insignificant, which limits the applicability of stopped-flow
methods to studies of protein–DNA interactions. SweepCE was introduced for directly measur-
ing kon , and demonstrated for studying protein–DNA interactions.50 In contrast to stopped-flow
spectroscopy, SweepCE does not rely on spectral changes of the protein or DNA upon complex for-
mation. It requires only that electrophoretic mobilities of the protein and DNA be different, which is
always achievable.
The concept of SweepCE is based on the sweeping of slowly migrating L by a fast migrating
T during electrophoresis. The capillary is filled with L, while the inlet reservoir contains T and
the outlet reservoir contains a run buffer. During electrophoresis, T continuously moves through
L, causing continuous binding of T to L. Dissociation of C can also contribute to the resulting
concentration profiles that contain a single peak of C and plateaus of T and L. The value of kon
for complex formation can be determined from the time-profile of T concentration using a simple
mathematical model of the sweeping process. Mathematical analysis is an essential part of SweepCE.
The method was demonstrated on the phenomenon of DNA sweeping by a DNA-binding protein in
CE (Figure 11.7).
11.3.3 ppKCE
In ppKCE, the plugs of L and T are injected into the capillary prefilled with the run buffer. The
inlet and outlet reservoirs contain the run buffer as well. During electrophoresis T moves through
L causing the formation of C. When the zone of T passes L, C starts to dissociate. ppKCE can be
considered as a functional hybrid of NECEEM and SweepCE. The resulting concentration profiles
Sweeping of DNA byprotein

1 intensifies with increasing of kon
Concentration of DNA
kon(1) > kon(2) > kon(3) > kon(4) = 0
2
3
4
3
Complex
Sweeping
2 Sweeping
1
No sweeping
0
0 50 100 150 200
Migration time to the end of the capillary (s)
FIGURE 11.7 SweepCE is a first nonspectroscopic method for direct measuring of kon . It is based on the
sweeping of a slowly migrating ligand by a fast migrating target during electrophoresis. The value of kon for
complex formation can be determined from the time-profile of target concentration using a mathematical model
of the sweeping process.
10
Fluorescent signal (a.u.)
L free
8 free
2 C L dis
dis
0
2 4 6 8 10
FIGURE 11.8 ppKCE is a method for direct measurement of both rate constants, kon and koff . The resulting
concentration profile is very similar to NECEEM with a smaller peak of a complex (area = C), a decay of the
complex (area = Ldis ) and a free ligand (area = Lfree ).
resemble those of NECEEM with a smaller peak of C and “smears” of T and L (Figure 11.8).3,9
However, since ppKCE does not start with the EM of L and T, the resulting electropherogram
does not have a “memory” of Kd but rather has a memory of kon and koff . Both kon and koff can,
thus, be calculated from a single ppKCE electropherogram using areas of peaks and smears and
migration times of peaks. Calculation of rate constants of complex formation and dissociation from
one experimental electropherogram is possible due to the developed simple mathematical approach
without performing nonlinear regression analysis so that the method does not require expertise in
mathematical modeling. So far, there has not been any method that allowed direct measurement of
both rate constants. ppKCE is simple and robust. It requires only nanoliters volumes of reagents and
can be readily adjusted for different ranges of both constants. This method will be useful for screening
large libraries for drug candidates as well as the development of novel research and diagnostic
tools.
11.3.4 OTHER KCE METHODS

In cNECEEM, the inlet reservoir is filled with the EM while the capillary and the outlet reservoir
contain the run buffer. During electrophoresis, C is separated from T, which moves faster, and
from L, which moves slower. As a result, C continuously dissociates inside the capillary. Although
dissociation is a prevalent process in cNECEEM, re-association can also contribute to the resulting
concentration profiles, which are represented by smooth functions of T (t), L(t), and C(t) with no
pronounced peaks.
In sSweepCE, a short plug of T is injected into the capillary prefilled with L. Both inlet and
outlet reservoirs contain the run buffer. T moves through L during electrophoresis causing both
association of T and L and dissociation of resulting C to occur. The concentration profiles of T and
C are peak-like, while that of L is a smooth function.
In sSweepCEEM, a short plug of T is injected into the capillary prefilled with the EM. Both
inlet and outlet reservoirs contain the run buffer. During electrophoresis, an intricate interplay of
dissociation of C and association of T and L occur resulting in sophisticated concentration profiles,
which contain peaks and plateaus.
11.3.5 MULTIMETHOD KCE TOOLBOX

The degree of formation and dissociation of a complex differ in different KCE methods.
KCE methods, therefore, have different accuracies of determination of kon and koff . For exam-
ple, in NECEEM, complex dissociation prevails over complex formation, thus, making it more
“sensitive” to koff than kon . In SweepCE, in contrast, complex formation can prevail over complex
dissociation, making it more sensitive to kon than koff . The ppKCE method can be tuned to have
comparable accuracy of both kon and koff determination. Therefore, KCE methods, which involve
EMs (e.g., NECEEM, cNECEEM, and sSweepCEEM), are expected to be more accurate for the
determination of the equilibrium constant, Kd . The most accurate determination of all constants can
be achieved if multiple KCE methods are combined in a single kinetic tool. The approach can be
used for testing hypotheses about the mechanisms of interaction and finding kinetic parameters of the
interaction. Conceptually, experimental electropherograms are obtained by multiple KCE methods
first. A hypothetical model of interactions between L and T is suggested and the system of differential
equations (system 11.3) is built. The experimental KCE electropherograms are fitted with simulated
electropherograms simultaneously to obtain the best fits with one of the criteria used for nonlinear
regression analysis (e.g., minimum chi-square). If the quality of fitting is not satisfactory, a new
hypothesis is suggested for the interaction. The procedure is repeated until a satisfying hypothe-
sis is found. The best fits for the accepted hypothesis lead to the determination of stoichiometric
and kinetic parameters of the interaction. Using the general concept of KCE, other KCE methods
can be defined by simply selecting new sets of initial and boundary conditions. Importantly, this
approach requires no serendipity but, rather, a rational (or irrational) design of conditions, which
can be performed in an intuitive way schematically depicted in Table 11.1. The multimethod KCE
toolbox allowed, for example, the determination of kinetic parameters of specific and nonspecific
protein–DNA interactions.3
KCE establishes a new paradigm: separation methods can be used as comprehensive kinetic
tools. The majority of previous attempts to utilize chromatography and electrophoresis for studying
biomolecular interactions were limited to assuming equilibrium between interacting molecules.51
Not only does such an assumption limit applications to measuring equilibrium constants, but also this
assumption is conceptually mistaken since separation disturbs equilibrium. Kinetics must be appre-
ciated when separation methods are used for studies of noncovalent interactions. This appreciation
can dramatically enrich analytical capabilities of the methods.
11.4 KCE FOR APTAMER SELECTION AND DRUG DISCOVERY

DNA or RNA aptamers are single-stranded oligonucleotides that can bind proteins, small-molecule
compounds, and living cells with high affinity and specificity.52,53 Aptamers are very promising
affinity ligands with the potential to change the field of affinity probes and replace antibodies as
diagnostic, analytical, and therapeutic reagents.54–56 Aptamers have indisputable advantages over
antibodies due to the simplicity and low cost of production, simplicity of chemical modifications, and
simplicity of integration into different analytical schemes. The unique properties of aptamers have
led to their application in multiple areas of bioanalytical and biomedical sciences. They have been
successfully used in proteomics and development of bioanalytical assays,57 inhibition of enzymes and
receptors,7,8 development of artificial enzymes (ribozymes and aptazymes),9 target validation and
screening for drug candidates,58–60 cytometry and imaging of cellular organelles,61 and development
of biosensors.62 Aptamers are gaining a strengthening reputation as therapeutic reagents for the
treatment of different pathologies.63,64 Their potential medical applications also include gene therapy
and drug delivery to therapeutic targets.65 Despite a great promise and a significant effort in the
development of aptamers over a period of 15 years, they have only been obtained for approximately
100 protein targets.66 Such a slow progress is largely due to limitations of conventional technologies
used for aptamer development.
Aptamers are typically selected from large libraries of random DNA(RNA) sequences in a general
approach termed Systematic Evolution of Ligands by EXponential enrichment (SELEX).67,68 In
essence, SELEX involves repetitive rounds of two alternating processes: (1) partitioning of aptamers
from nonaptamers by separating target-bound DNA from free DNA and (2) amplification of aptamers
by the PCR (Figure 11.9a).
Noninstrumental methods of partitioning, such as filtration and gel electrophoresis, were initially
used for SELEX; they still dominate the area.69 Because of high background (the high level of target-
nonbound DNA collected along with target-bound DNA), SELEX based on conventional partitioning
methods requires a large number of rounds of selection, typically greater than 10. As a result, SELEX
based on conventional partitioning methods is a lengthy and resource-consuming process. It often
leads to DNA structures that bind to the surfaces of the filters or chromatographic support used in
partitioning rather than to the target. Counter selection is successfully employed to eliminate such
“surface aptamers”; however, it introduces additional rounds of selection, thus, making the procedure
even longer. Another disadvantage of too many rounds of selection is the very limited number of
unique aptamer sequences obtained at the output of conventional SELEX. This disadvantage is
(a) (b)
SELEX NonSELEX
N rounds
N steps
Partitioning
Analysis Partitioning Analysis
Amplification
FIGURE 11.9 (a and b) Schematic representation of SELEX vs. nonSELEX.

N rounds
• Partitioning • Cloning
Separation of a target-binding DNA
from a DNA library by KCE methods • KCE-based affinity
analysis of clones
• Sequencing of the best
• PCR Amplification aptamers
of the selected target-binding DNA to obtain
an enriched library • Chemical synthesis of
aptamers
• Modification and/or
• Strand Separation truncation of aptamers
• KCE-based affinity
analysis of synthetic
and modified aptamers
• Affinity Analysis
of the DNA pool to the target by NECEEM
FIGURE 11.10 Schematic representation of KCE-based selection of DNA aptamers. This figure shows a
simplified flowchart of in vitro selection of DNA aptamers. A random DNA pool containing 1011 –1015 unique
sequences is incubated with a target. The next steps are KCE-based measuring of bulk affinity of DNA library
to the target and KCE-based partitioning of a DNA-target complex from free DNA. The collected DNA is
then amplified using PCR with a fluorescein-labeled forward primer and a biotin-labeled reverse primer. DNA
strands are separated using streptavidin iron particles and a fluorescently labeled strand is collected to yield the
affinity-enriched DNA library. This lower diversity pool is incubated with new aliquot of the target to examine
its binding affinity using NECEEM. If the affinity is not high enough, a new selection round starts. When
desirable affinity is reached, pool of aptamers is cloned into bacteria, clones are screened, and the best aptamers
are sequenced. Aptamer sequences can be further modified to improve binding or other specific properties.
especially critical for aptamer-based drug development, which requires as many “lead molecules”
as possible. Finally, if the efficiency of partitioning is too low, SELEX can completely fail to
select aptamers.
Methods of KCE,3 which started with pioneering works of Heegaard and Whitesides on affin-
ity capillary electrophoresis (ACE),15,70 establish a new methodological platform for partitioning of
aptamers. Two distinct KCE methods have been used for the selection of aptamers: NECEEM24,25,29
and ECEEM.71,72 Bowser and co-authors73,74 were the first to use NECEEM in SELEX; they called
the approach CE-SELEX. The partitioning efficiency of KCE methods exceeds that of conven-
tional partitioning methods, such as filtration and column chromatography, by at least two orders
of magnitude.29 As a result, KCE methods decrease the number of rounds of SELEX from 10 or
more (required with conventional partitioning techniques) to 1–3 (Figure 11.10). In addition, KCE
methods can be equally used for the selection of aptamers and for measurements of all their bind-
ing parameters: Kd , kon , koff , H, and S.19 KCE methods have been demonstrated to facilitate
selection of “smart” aptamers—ligands with predefined binding parameters.
11.4.1 KCE-BASED SELECTION OF SMART APTAMERS

The designing of advanced aptamer-based diagnostics and therapeutics requires aptamers with
predefined kinetic and/or thermodynamic parameters of aptamer–target interaction. Technologi-
cal limitations of aptamer selection methods have so far precluded selection of such aptamers.
Two KCE methods, ECEEM71,72 and NECEEM,24,25,29 have been successfully used for the selec-
tion of aptamers with predefined Kd and koff of aptamer–target complexes. Conceptually, in ECEEM,
a mixture of a target with a DNA (RNA) library is prepared and equilibrated (Figure 11.11). A plug of
the EM is injected into a capillary prefilled with a run buffer containing the target at the concentration
identical to the target concentration in the EM. The components of the EM are separated by CE while
quasi-equilibrium is maintained between the target and aptamers inside the capillary. The unique
feature of ECEEM is that aptamers with different Kd values migrate with different and predictable
mobilities. Thus, collecting fractions with different mobilities results in aptamers with different and
predefined Kd values. As an example, ECEEM was used to select aptamers with predefined Kd for
binding MutS protein.71 Three rounds of ECEEM-based selection were sufficient to obtain aptamers
with Kd values approaching theoretically predicted ones (Figure 11.11).
In the NECEEM approach, smart aptamers can be selected with predefined Kd and koff as well.
A combinatorial DNA library is equilibrated with a protein target and the components of the EM are
separated under nonequilibrium conditions. The nonequilibrium conditions promote the dissociation
of the complex during separation. Fractions collected in a time window preceding the DNA library
yield pools of DNA sequences capable of binding the target and dissociating from the complex with
specific rates (Figure 11.12). Being a homogeneous method with comprehensive kinetic features,
NECEEM provides a means for the selection of DNA aptamers with predefined ranges of all binding
parameters of complex formation (Kd , koff , kon ). First, the selection can be with respect to Kd values
by varying the concentration of the protein target [T ] in the EM. The ratio between protein bound
and unbound ligands changes according to classical equilibrium: the ligands where Kd < [T ] are
preferentially bound to the protein and selected, while the ligands where Kd > [T ] are preferentially
unbound and not selected. Second, the selection can be implemented with respect to koff values
by varying time windows in which fractions are collected. Finally, selection with respect to kon
values can be carried out by varying the time of incubation of the library with the target. To select
for a single binding constant, the parameters that control the other two binding constants should be
kept unchanged.
NECEEM appears to be a more complicated method than ECEEM, because each collection
window may contain ligands with totally different koff values; however, ligands with koff defined
in Figure 11.12 are the most abundant and will be predominant after multiple rounds of selection.
The concept is different in ECEEM, in which every window theoretically contains only ligands with
calculated Kd values, and deviations occur only because of nonspecific interactions during separation,
nonzero width of the peaks, and disturbed equilibrium. Thus, NECEEM is a more “evolutionary”
method of iterative selection, which requires more than one round to select oligonucleotides with a
narrow range of koff values.
11.4.2 NONSELEX SELECTION OF APTAMERS

The outstanding partitioning capabilities of KCE methods have motivated the attempt to select
aptamers in a procedure that does not include intermediate amplifications steps; this approach is
called nonSELEX24,25 (Figure 11.9b). Excluding repetitive steps of PCR amplification accelerates
the procedure of aptamer selection without compromising its efficiency. Omitting repetitive steps of
PCR also excludes quantitative errors associated with the exponential nature of PCR amplification,
thus making nonSELEX a useful tool for studies of properties of DNA libraries with respect to
their interaction with targets. For example, nonSELEX can be used to accurately back-calculate the
number of aptamer molecules in the naive DNA library. Further, excluding repetitive steps of PCR
allows one to avoid the bias related to differences in PCR efficiency with respect to different oligonu-
cleotide sequences. Finally, nonSELEX can potentially provide a viable alternative to SELEX in the
commercial development of aptamers. It should be noted that the implementation of nonSELEX with
currently available commercial CE instrumentation has a limitation: only a fraction of the collected
ligands can be sampled for the next step of nonSELEX. This limitation requires that the fraction of
aptamers in the naive library be no lower than 5 × 10−10 for the parameters used in this protocol.
Usually, the abundance of aptamers in the naive library is greater than 5 × 10−10 , thus making
Separation scheme
Equilibrium mixture EM = protein-DNAcomplex + free Protein+ free DNA
[Protein]EM= [Protein]RunningBuffer
Injection EM Protein Protein
Separation
Constant flow of the protein in the running buffer
tDNAPAt-- tP•DNA
Kd(t)=[P]
tP•DNA tDNA-t
Kd = 0 K Id <KdCOL <KIId Kd = +
Collection free
More stable window Less stable D
complexes P DNA complexes P DNA N
A
tP DNA t1 t2 tDNA
Time of migration to the end of the capillary
Experimental affinities
Theoretical affinity
Round 0 Round 0 Round 0
2000
Round 1
1500 Round 1
Affinity,Kd(nM)
Round 2
1000
Round 2
500 Round 3
Round 1
Round 3
Round 2
0 Round 3
16 20 24 28 Migration time (min)

Aptamer collection regions: I II III
FIGURE 11.11 ECEEM-based selection of smart aptamers with predefined Kd values.
the limitation less important. As a real example, NECEEM-based nonSELEX for the selection of
DNA aptamers was shown for h-Ras protein.24 Three steps of NECEEM-based partitioning in the
NonSELEX approach were sufficient to improve the affinity of a DNA library to a target protein by
more than four orders of magnitude (Figure 11.13). The resulting affinity was higher than that of the
enriched library obtained in three rounds of NECEEM-based SELEX. Remarkably, NECEEM-based
NonSELEX selection took only 1 h to complete in contrast to several days or several weeks required
for a typical SELEX procedure by conventional partitioning methods. In addition, NECEEM-based
NonSELEX allowed to accurately measure the abundance of aptamers in the library. Not only does
(a) EM DNA DNA T

•T
Direction of the migration in electrophoresis

(b)
t0 EM
t1 DNA T
DNA
•T
t2 DNA DNA dissociated from DNAT

DNA
T
T dissociated from DNA T
•T
Concentration
D
N DNA dissociated T dissociated T
A from DNA•T from DNA T
DNA•T
Position in the capillary at time et2

(c) Negative selection
(Kd > [T])
Concentration
Positive selection
(Kd < [T])
Kd < [Protein]0
koff = F (t1,t2,tDNA,tDNA T)
D

T dissociated N
T DNA dissociated
from DNA•T from DNA •T
A tDNA–tDNA•T tDNA–t2
koff = In
tDNA•T(t1–t2) tDNA–t1
tDNA•T t2 t1 tDNA
Time of migration to the end of the capillary
FIGURE 11.12 NECEEM-based selection of DNA aptamers with predefined ranges of all binding parameters
of complex formation (Kd , koff , kon ).
Selection
Multiple NECEEM-based partitioning Free
DNA+Target Protein–DNA
DNA
7 Affinity after Step 1 complex
Fluorescence signal (a.u)
Step 1 Step 2 Step 3

Kd = 100 M
Target Target 5
Affinity after Step 2
Kd = 5M
3
Analysis Affinity after Step 3
1. PCR 1 Kd = 0.3 M
2. Strand separation
3. NECEEM binding analysis
4 5 6 7 8 9 10 11
Migration time to the capillary exit (min)
FIGURE 11.13 NonSELEX. Partitioning of aptamers from nonaptamers in nonSELEX by multiple steps of
NECEEM with no PCR amplification between the steps.
this work introduce an extremely fast and economical method for aptamer selection, but it also sug-
gests that aptamers may be much more abundant than they are thought to be. The step-by-step protocol
for nonSELEX has been published.25 Even more importantly, this work opens the opportunity for
the selection of drug candidates from large nonDNA libraries, which cannot be PCR-amplified and
are, thus, not approachable by SELEX.
11.4.3 THE PROSPECTIVE OF KCE IN DRUG DISCOVERY

The nonSELEX concept provides the opportunity for the selection of affinity ligands from DNA-
encoded libraries of small molecules and peptides.75,76 DNA tags in such libraries encode the
information, which allows one to identify structures of corresponding small molecules, when the
DNA tags are PCR amplified and sequenced. Owing to the small size of the molecules with respect to
that of the covalently attached DNA tag, such libraries are expected to have electrophoretic properties
identical to those of DNA libraries. SELEX is not applicable to such libraries since small molecules
and peptides cannot be amplified by PCR.
11.5 CONCLUSION
KCE is based on the major principle of separation science that a complex can be separated from
a target. In this case, the kinetics of molecular interactions must be appreciated. All methods of
KCE use a single instrumental platform and a single conceptual platform for solving multiple tasks
associated with biomolecular screening. A variety of different KCE methods can be designed by
defining different ways of interaction between molecules in CE. Because of their comprehensive
analytical capabilities, KCE methods have the potential to become a workhorse of biomolecular
screening. This makes KCE methods highly attractive for the pharmaceutical industry as a novel
approach to the selection and characterization of drug candidates. To conclude, it is clear that
KCE methods will find multiple applications in fundamental studies of biomolecular interactions,
designing clinical diagnostics, and the development of affinity probes and drug candidates. New
applications will emerge with further development of KCE.
REFERENCES
1. Englebienne, P., Van Hoonacker, A., and Verhas, M. Surface plasmon resonance: Principles, methods
and applications in biomedical sciences. Spectroscopy 17, 255–273 (2003).
2. Homola, J. Present and future of surface plasmon resonance biosensors. Anal. Bioanal. Chem. 377,
528–539 (2003).
3. Petrov, A., Okhonin, V., Berezovski, M., and Krylov, S. N. Kinetic capillary electrophoresis (KCE):
A conceptual platform for kinetic homogeneous affinity methods. J. Am. Chem. Soc. 127, 17104–
17110 (2005).
4. Berezovski, M., Okhonin, V., Petrov, A., and Krylov, S. N. Kinetic methods in capillary electrophoresis
and their applications. Proc. Int. Soc. Opt. Eng. 5969, 203–215 (2005).
5. Berezovski, M. Kinetic capillary electrophoresis and its applications. Ph.D. Dissertation, Department
of Chemistry, York University, Toronto (2005).
6. Berezovski, M., and Krylov, S. N. Nonequilibrium capillary electrophoresis of equilibrium mixtures—
A single experiment reveals equilibrium and kinetic parameters of protein–DNA interactions. J. Am.
Chem. Soc. 124, 13764–13765 (2002).
7. Krylov, S. N., and Berezovski, M. Non-equilibrium capillary electrophoresis of equilibrium mixtures—
Appreciation of kinetics in capillary electrophoresis. Analyst 128, 571–575 (2003).
8. Okhonin, V., Krylova, S. M., and Krylov, S. N. Nonequilibrium capillary electrophoresis of equilibrium
mixtures, mathematical model. Anal. Chem. 76, 1507–1512 (2004).
9. Okhonin, V., Petrov, A., Berezovski, M., and Krylov, S. N. Plug-plug kinetic capillary electrophoresis:
A method for direct determination of rate constants of complex formation and dissociation. Anal.
Chem. 78, 4803–4810 (2006).
10. Shimura, K., and Karger, B. L. Affinity probe capillary electrophoresis: Analysis of recombinant
human growth hormone with a fluorescent labeled antibody fragment. Anal. Chem. 66, 9–15 (1994).
11. German, I., Buchanan, D. D., and Kennedy, R. T. Aptamers as ligands in affinity probe capillary
electrophoresis. Anal. Chem. 70, 4540–4545 (1998).
12. Yang, P., Whelan, R. J., Jameson, E. E., Kurzer, J. H., Argetsinger, L. S., Carter-Su, C., Kabir, A.,
Malik, A., and Kennedy, R. T. Capillary electrophoresis and fluorescence anisotropy for quantitative
analysis of peptide–protein interactions using JAK2 and SH2-Bb as a model system. Anal. Chem. 77,
2482–2489 (2005).
13. Dolnik, V. Recent developments in capillary zone electrophoresis of proteins. Electrophoresis 20,
3106–3115 (1999).
14. Busch, M. H. A., Boelens, H. F. M., Kraak, J. C., Poppe, H., Meekel, A. A. P., and Resmini, M. Critical
evaluation of the applicability of capillary zone electrophoresis for the study of hapten–antibody
complex formation. J. Chromatogr. A 744, 195–203 (1996).
15. Heegaard, N. H. H., and Robey, F. A. Use of capillary zone electrophoresis to evaluate the binding of
anionic carbohydrates to synthetic peptides derived from human serum amyloid P component. Anal.
Chem. 64, 2479–2482 (1992).
16. Kraak, J. C., Busch, S., and Poppe, H. Study of protein–drug binding using capillary zone electro-
phoresis. J. Chromatogr. A 608, 257–264 (1992).
17. Wu, S., and Dovichi, N. J. High-sensitivity fluorescence detector for fluorescein isothiocyanate
derivatives of amino acids separated by capillary zone electrophoresis. J. Chromatogr. 480,
141–155 (1989).
18. Zhao, J. Y., Dovichi, N. J., Hindsgaul, O., Gosselin, S., and Palcic, M. M. Detection of 100 molecules
of product formed in a fucosyltransferase reaction. Glycobiology 4, 239–242 (1994).
19. Berezovski, M., and Krylov, S. N. Thermochemistry of protein–DNA interaction studied with
temperature-controlled nonequilibrium capillary electrophoresis of equilibrium mixtures. Anal. Chem.
77, 1526–1529 (2005).
20. Berezovski, M., Nutiu, R., Li, Y., and Krylov, S. N. Affinity analysis of a protein–aptamer
complex using nonequilibrium capillary electrophoresis of equilibrium mixtures. Anal. Chem. 75,
1382–1386 (2003).
21. Krylova, S. M., Musheev, M., Nutiu, R., Li, Y., Lee, G., and Krylov, S. N. Tau protein binds both double-
stranded and single-stranded DNA—The proof obtained with nonequilbrium capillary electrophoresis
of equilibrium mixtures. FEBS Lett. 579, 1371–1375 (2005).
22. Berezovski, M., Drabovich, A., Krylova, S. M., Musheev, M., Okhonin, V., Petrov, A., and
Krylov, S. N. Nonequilibrium Capillary Electrophoresis of Equilibrium Mixtures: A universal tool
for development of aptamers. J. Am. Chem. Soc. 127, 3165–3171 (2005).
23. Drabovich, A., Berezovski, M., and Krylov, S. N. Selection of smart aptamers by equilibrium capillary
electrophoresis of equilibrium mixtures (ECEEM). J. Am. Chem. Soc. 127, 11224–11225 (2005).
24. Berezovski, M., Musheev, M., Drabovich, A., and Krylov, S. N. Non-SELEX selection of aptamers.
J. Am. Chem. Soc. 128, 1410–1411 (2006).
25. Berezovski, M. V., Musheev, M. U., Drabovich, A. P., Jitkova, J., and Krylov, S. N. Non-SELEX:
Selection of aptamers without intermediate amplification of candidate oligonucleotides. Nat. Protoc.
1, 1359–1369 (2006).
26. Krylov, S., Krylova, S., and Berezovski, M. Non-equilibrium capillary electrophoresis of equilibrium
mixtures (NECEEM)-based methods for drug and diagnostic development. U.S. Pat. Appl. Publ.,
41 (2005).
27. Berezovski, M., and Krylov, S. N. Using nonequilibrium capillary electrophoresis of equilibrium
mixtures for the determination of temperature in capillary electrophoresis. Anal. Chem. 76, 7114–
7117 (2004).
28. Krylov, S. N. Nonequilibrium capillary electrophoresis of equilibrium mixtures (NECEEM): A novel
method for biomolecular screening. J. Biomol. Screen. 11, 115–122 (2006).
29. Berezovski, M., Drabovich, A., Krylova, S. M., Musheev, M., Okhonin, V., Petrov, A., and Krylov, S.
N. Nonequilibrium capillary electrophoresis of equilibrium mixtures: A universal tool for development
of aptamers. J. Am. Chem. Soc. 127, 3165–3171 (2005).
30. Berezovski, M., and Krylov, S. N. Using DNA-binding proteins as an analytical tool. J. Am. Chem.
Soc. 44, 13451–13454 (2003).
31. Ma, H., and Zou, Y. Thermodynamic characterization of the interaction of mutant UvrB protein with
damaged DNA. Biochemistry 43, 4206–4211 (2004).
32. Schubert, F., Zettl, H., Haefner, W., Krauss, G., and Krausch, G. Comparative thermodynamic
analysis of DNA–protein interactions using surface plasmon resonance and fluorescence correlation
spectroscopy. Biochemistry 42, 10288–10294 (2003).
33. Galletto, R., and Bujalowski, W. Kinetics of the E. coli replication factor DnaC protein–nucleotide
interactions. II. Fluorescence anisotropy and transient, dynamic quenching stopped-flow studies of the
reaction intermediates. Biochemistry 41, 8921–8934 (2002).
34. Mosig, G., Gewin, J., Luder, A., Colowick, N., and Vo, D. Two recombination-dependent DNA
replication pathways of bacteriophage T4, and their roles in mutagenesis and horizontal gene transfer.
Proc. Natl Acad. Sci. USA 98, 8306–8311 (2001).
35. Sutani, T., and Yanagida, M. DNA renaturation activity of the SMC complex implicated in chromosome
condensation. Nature 388, 798–801 (1997).
36. Savran, C. A., Knudsen, S. M., Ellington, A. D., and Manalis, S. R. Micromechanical detection of
proteins using aptamer-based receptor molecules. Anal. Chem. 76, 3194–3198 (2004).
37. Dang, C., and Jayasena, S. D. Oligonucleotide inhibitors of Taq DNA polymerase facilitate detection
of low copy number targets by PCR. J. Mol. Biol. 264, 268–278 (1996).
38. Yakimovich, O. Y., Alekseev, Ya. I., Maksimenko, A. V., Voronina, O. L., and Lunin, V. G. Influence of
DNA aptamer structure on the specificity of binding to Taq DNA polymerase. Biochemistry (Moscow)
68, 228–235 (2003).
39. Carra, J. H., and Privalov, P. L. Energetics of folding and DNA binding of the MAT alpha 2
homeodomain. Biochemistry 36, 526–535 (1997).
40. Knapp, S., Karshikoff, A., Berndt, K. D., Christova, P., Atanasov, B., and Ladenstein, R. Thermal
unfolding of the DNA-binding protein Sso7d from the hyperthermophile Sulfolobus solfataricus. J.
Mol. Biol. 264, 1132–1144 (1996).
41. Davis, K. G., Plyte, S. E., Robertson, S. R., Cooper, A., and Kneale, G. G. Comparison of Pf1 and
Fd gene 5 proteins and their single-stranded DNA complexes by NMR spectroscopy and differential
scanning calorimetry. Biochemistry 34, 148–154 (1995).
42. Shumaker-Parry, J. S., Aebersold, R., and Campbell, C. T. Parallel, quantitative measurement of protein
binding to a 120-element double-stranded DNA array in real time using surface plasmon resonance
microscopy. Anal. Chem. 76, 2071–2082 (2004).
43. Tsoi, P. Y., and Yang, M. Surface plasmon resonance study of the molecular recognition between
polymerase and DNA containing various mismatches and conformational changes of DNA–protein
complexes. Biosens. Bioelectron. 19, 1209–1218 (2004).
44. Berezovski, M., and Krylov, S. N. A non-spectroscopic method for the determination of temperature
in capillary electrophoresis. Anal. Chem. 76, 7114–7117 (2004).
45. Imanishi, M., and Sugiura, Y. Artificial DNA-bending six-zinc finger peptides with different charged
linkers: Distinct kinetic properties of DNA bindings. Biochemistry 41, 1328–1334 (2002).
46. Cheskis, B., and Freedman, L. P. Modulation of nuclear receptor interactions by ligands: Kinetic
analysis using surface plasmon resonance. Biochemistry 35, 3309–3318 (1996).
47. Kozlov, A. G., and Lohman, T. M. Kinetic mechanism of direct transfer of Escherichia coli SSB
tetramers between single-stranded DNA molecules. Biochemistry 41, 6032–6044 (2002).
48. Noel, A.-J., Wende, W., and Pingoud, A. J. DNA Recognition by the homing endonuclease PI-
SceI involves a divalent metal ion cofactor-induced conformational change. Biol. Chem. 279,
6794–6804 (2004).
49. Zeeb, M., and Balbach, J. Single-stranded DNA binding of the cold-shock protein CspB from Bacillus
subtilis: NMR mapping and mutational characterization. J. Protein Sci. 12, 112–123 (2003).
50. Okhonin, V., Berezovski, M., and Krylov, S. N. Sweeping capillary electrophoresis: A non-stopped-
flow method for measuring bimolecular rate constant of complex formation between protein and DNA.
J. Am. Chem. Soc. 126, 7166–7167 (2004).
51. Heegaard, N. H. H., and Kennedy, R. Identification, quantitation, and characterization of
biomolecules by capillary electrophoretic analysis of binding interactions. Electrophoresis 20,
3122–3133 (1999).
52. Wilson, D. S., and Szostak, J. W. In vitro selection of functional nucleic acids. Annu. Rev. Biochem.
68, 611–647 (1999).
53. Famulok, M., Mayer, G., and Blind, M. Nucleic acid aptamers—from selection in vitro to applications
in vivo. Acc Chem. Res. 33, 591–599 (2000).
54. Rimmele, M. Nucleic acid aptamers as tools and drugs: Recent developments. Chembiochem 4,
963–971 (2003).
55. Jayasena, S. D. Aptamers: An emerging class of molecules that rival antibodies in diagnostics. Clin.
Chem. 45, 1628–1650 (1999).
56. Ellington, A. D., and Conrad, R. Aptamers as potential nucleic acid pharmaceuticals. Biotechnol.
Annu. Rev. 1, 185–214 (1995).
57. Bock, C., Coleman, M., Collins, B., Davis, J., Foulds, G., Gold, L., Greef, C., et al. Photoaptamer
arrays applied to multiplexed proteomic analysis. Proteomics 4, 609–618 (2004).
58. Mayer, G., and Jenne, A. Aptamers in research and drug development. BioDrugs 18, 351–359 (2004).
59. Nutiu, R., Yu, J. M., and Li, Y. Signaling aptamers for monitoring enzymatic activity and for inhibitor
screening. Chembiochem 5, 1139–1144 (2004).
60. Famulok, M., and Mayer, G. Intramers and aptamers: Applications in protein-function analyses and
potential for drug screening. Chembiochem 6, 19–26 (2005).
61. Ulrich, H., Martins, A. H., and Pesquero, J. B. RNAand DNAaptamers in cytomics analysis. Cytometry
A 59, 220–231 (2004).
62. Tombelli, S., Minunni, M., and Mascini, M. Analytical applications of aptamers. Biosens. Bioelectron.
20, 2424–2434 (2005).
63. Nimjee, S. M., Rusconi, C. P., and Sullenger, B. A. Aptamers: An emerging class of therapeutics.
Annu. Rev. Med. 56, 555–583 (2005).
64. Lee, J. F., Stovall, G. M., and Ellington, A. D. Aptamer therapeutics advance. Curr. Opin. Chem. Biol.
10, 282–289 (2006).
65. Patil, S. D., Rhodes, D. G., and Burgess, D. J. DNA-based therapeutics and DNA delivery systems: A
comprehensive review. AAPS J. 7, 61–77 (2005).
66. Lee, J. F., Hesselberth, J. R., Meyers, L. A., and Ellington, A. D. Aptamer database. Nucleic Acids
Res. 32, D95–D100 (2004).
67. Ellington, A., and Szostak, J. In vitro selection of RNA molecules that bind specific ligands. Nature
346, 818–822 (1990).
68. Tuerk, C., and Gold, L. Systematic evolution of ligands by exponential enrichment: RNA ligands to
bacteriophage T4 DNA polymerase. Science 249, 505–510 (1990).
69. Gopinath, S. C. Methods developed for SELEX. Anal. Bioanal. Chem. 387, 171–182 (2006).
70. Chu, Y. H., Avila, L. Z., Biebuyck, H. A., and Whitesides, G. M. Use of affinity capillary electrophoresis
to measure binding constants of ligands to proteins. J. Med. Chem. 35, 2915–2907 (1992).
71. Drabovich, A., Berezovski, M., and Krylov, S. N. Selection of smart aptamers by equilibrium capillary
electrophoresis of equilibrium mixtures (ECEEM). J. Am. Chem. Soc. 127, 11224–11225 (2005).
72. Drabovich, A. P., Berezovski, M., Okhonin, V., and Krylov, S. N. Selection of smart aptamers by
methods of kinetic capillary electrophoresis. Anal. Chem. 78, 3171–3178 (2006).
73. Mendonsa, S. D., and Bowser, M. T. In vitro evolution of functional DNA using capillary
electrophoresis. J. Am. Chem. Soc. 126, 20–21 (2004).
74. Mendonsa, S. D., and Bowser, M. T. In vitro selection of high-affinity DNA ligands for human IgE
using capillary electrophoresis. Anal. Chem. 76, 5387–5392 (2004).
75. Kanna, M. W., Rozenman, M. M., Sakurai, K., Snyder, T. M., and Liu, D. R. Reaction discovery
enabled by DNA-templated synthesis and in vitro selection. Nature 431, 545–549 (2004).
76. Levin, A. M., and Weiss, G. A. Optimizing the affinity and specificity of proteins with molecular
display. Mol. Biosyst. 2, 49–57 (2006).
12 DNA Sequencing and
Genotyping by Free-Solution
Conjugate Electrophoresis
Jennifer A. Coyne, Jennifer S. Lin, and Annelise E. Barron
CONTENTS
12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381

12.2 Free-Solution Conjugate Electrophoresis of DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
12.2.1 Theory of FSCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385
12.2.2 Proof-of-Concept Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
12.2.2.1 Streptavidin. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
12.2.2.2 Oligosaccharides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 390
12.3 Drag-Tags for FSCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 390
12.3.1 Synthetic Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392
12.3.2 Polypeptoids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392
12.3.2.1 Linear Polypeptoids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
12.3.2.2 Branched Polypeptoids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
12.3.2.3 Analysis of Solid-Phase Synthesis Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
12.3.3 Genetically Engineered Protein Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
12.3.3.1 Designing the Drag-Tag Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
12.3.3.2 Analysis of Protein Polymers as Drag-Tags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 396
12.3.4 Double-Labeled DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
12.4 DNA Analysis by FSCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
12.4.1 DNA Genotyping: Single-Base Extension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
12.4.2 DNA Sequencing by FSCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
12.5 Methods development guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402
12.5.1 Drag-Tag Cloning and Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402
12.5.1.1 Standard Protocol for Cloning and Production of Protein Polymers . . . 402
12.5.1.2 Polypeptoid Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
12.5.2 DNA + Drag-Tag Conjugation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404
12.5.3 Thermal Cycling Protocols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405
12.5.4 Electrophoresis Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405
12.6 Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406
Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
12.1 INTRODUCTION
Most DNA sequencing today is done essentially by the same enzymatic method conceived by
Frederick Sanger in the 1970s—while DNA electrophoresis instruments and molecular labels have
improved (we now use fluorescent dyes, whereas Sanger used radioactive labeling), the basic
381
approach is the same. When we read sequence from samples produced by the Sanger method,
the problem is reduced to biomolecule separation and detection. DNA molecules, each carrying a
terminal fluorophore that encodes the identity of the terminal base (A, C, G, or T), must be frac-
tionated according to chain length with single-base resolution, and for the best process efficiency,
with the longest “read-length” possible. Ideally at least 600–650 bases of contiguous, high-accuracy
sequence is desired per read for the de novo sequencing of complex genomes such as the human
genome, which has a high content of repetitive DNA sequence that can be difficult to “read through”
and assemble correctly if only shorter reads are available.
Accordingly, since 1990 or so, a key aspect of pushing sequencing technology forward has been
perfecting the ability of electrophoresis instruments to separate DNA molecules according to size
with great efficiency. A major turning point, reached in 1999, was the widespread shift of sequenc-
ing from slab gel instruments, which required genome center technicians to carry out manual steps
(e.g., pouring very large, ultra-thin polyacrylamide slab gels, loading DNA samples), to automated
capillary array electrophoresis (CAE) systems, which separate fluorescently labeled DNA molecules
within 50- or 75-µm inner diameter glass capillaries, in fluid, uncrosslinked polyacrylamide solu-
tions rather than crosslinked gels. Although the basic technological approach to sequencing is still
essentially “gel electrophoresis,” the greater automation provided by CAE along with intensive logis-
tical and computational efforts in academia and private enterprise enabled a draft human genome
sequence to be obtained 2 years ahead of schedule, in 2003.1,2
The sequencing of that first composite human genome, a good deal of which was done with slab
gel systems, cost a total of $3 billion.1–6 Thanks to improvements to CAE and other technological
developments seeded in academic research laboratories with funds granted by the National Institutes
of Health (NIH)/National Human Genome Research Institute (NHGRI) as well as in industry, genome
centers now analyze genomes 24 h/day by automated CAE of DNA in solutions of entangled,
water-soluble polymers.7 The aforementioned viscous, entangled polymer solutions are at present
necessary, consumable, and expensive elements of CAE technology, as they provide the size-based
separation of DNA required to read sequence by the Sanger method.
In 2004, the NIH/NHGRI initiated a race to develop new and different technologies that could
decrease the cost for sequencing an individual human genome (which comprises 6 billion base pairs,
considering both sets of chromosomes) from the now-current (2007) price of $10–20 million to
$100K in the near future, and eventually to $1K.8 Changes to the separation technology could drop
the cost a lot, since about 60% of the present cost of sequencing a genome is due to CAE and its
required consumables (capillary arrays and polymer solutions). While CAE has transformed both
science and society and remains extremely important (as the only technology suitable for the de
novo sequencing of complex genomes, such as the human), less expensive, more efficient, and
more highly parallel technologies are still greatly needed. In 2006, the National Cancer Institute
(NCI) announced “The Cancer Genome Atlas” (TCGA), a new project that aims to map and analyze
genetic mutations occurring in different types of human cancers; this will require full sequencing
and assembly of ∼15,000 human genomes, an effort much greater in scale than the Human Genome
Project (HGP).9,10 A 100-fold reduction in sequencing cost to $100K per genome will be needed if the
Cancer Genome Atlas Project is to be accomplished within its presently planned budget.11 The second
100-fold cost decrease, to a $1K genome, would enable individual human genomes to be sequenced
as a new aspect of general medical care. In addition, the recently announced, privately backed
Archon X Prize for Genomics promises $10 million to the first group that sequences 100 human
genomes in less than 10 days,12,13 which would represent a giant leap beyond the 2003 completion
of the HGP. However, given current sequencing costs and the state of present technologies, old and
new, it certainly seems that we have a long way to go before sequencing is cheap enough to make
$10 million an attractive prize for sequencing 100 human genomes, so that the Archon Prize can be
paid out.14
Reducing sequencing costs requires both increased throughput and decreased cost per base.
In practical terms, capillary electrophoresis (CE) using polymer matrices for DNA separation is
DNA Sequencing and Genotyping by Free-Solution Conjugate Electrophoresis 383
limited to a maximum read-length (the number of contiguous DNA bases that can be sequenced with
high accuracy, i.e., with good single-base resolution) of about 800–900 bases, in part because of
limitations imposed by the physical mechanism of DNA separation that prevails in entangled polymer
solutions.14,15 A phenomenon known as “biased reptation” occurs as DNA molecules move through
a polymer sieving matrix in an electric field; the DNA molecules unstretch and move through the
polymer “mesh” head-first in a snake-like fashion, orienting themselves in the direction of the field.
The higher the electric field strength, the more DNA molecules—especially long DNA chains greater
than 300 bases in length—are likely to reptate with a strong field-bias, and the weaker the dependence
of DNA electrophoretic mobility on size for these reptating chains. This causes the size-dependence
of DNA mobility to decrease, thereby also decreasing read-length. Therefore, capillary in a polymer
matrix electrophoresis requires a balance to be struck between the increased separation speed possible
under high electric fields and the decreased read-length.16 DNA sequencing is also limited by the
number of capillaries that can be run in parallel; CAE instruments are currently available with up
to 384 capillaries, though instruments with 96 capillaries are much more typical. Polymer sieving
matrix costs are a significant percentage of the cost of DNA sequencing consumables, however, and
the cost of polymer increases linearly with the number of capillaries. Throughput may be increased
by highly parallel separations, but the cost per base is only minimally decreased.
Miniaturized microfluidic electrophoresis devices or “chips,” which are still primarily in the
research stage for the applications of DNA sequencing and genotyping [which often requires the
fractionation of small, single-stranded DNA (ssDNA) molecules] are a natural next step from CE
because they decrease the separation time and increase throughput, thus offering to reduce the
cost per base of DNA sequencing. DNA has been separated by electrophoresis on both glass17-21
and plastic22,23 microchips of varying lengths. However, polymer-sieving matrices that are easily
pressure-loaded into capillaries (with pressure limits > 1000 psi) are difficult to load into glass
and plastic microchips (which have pressure limits in the range of 50–200 psi). If DNA could
be separated by electrophoresis in free solution, that is, without a polymer sieving matrix, DNA
sequencing by microchip electrophoresis would be significantly easier. Eliminating the need for
polymer matrix in chip electrophoresis could decrease the cost per base even further, relative to
microchip electrophoresis using separation matrices.
Woolley and Mathies24 of U.C. Berkeley published the first four-color DNA sequencing results
obtained on microfluidic devices back in 1995. The Mathies group then showed sequencing of
500 bases in 20 min in 1999.19 Since then, the Jovanovich group at Molecular Dynamics has suc-
ceeded in sequencing 450 bases on a chip in 15 min.21 While these sequencing results are certainly
impressive, they are not yet fast enough, or cheap enough, to provide the “$100,000 genome” on their
own. Research groups are working to integrate several parts of the sequencing protocol (DNA extrac-
tion, purification, amplification, and detection) on a “lab-on-a-chip” microfluidic device17,25–41 that
includes polymer matrix-based CE as the final separation step.
In CE, polymer sieving matrices are necessary to separate DNA molecules according to size,
for the reason that DNA chains behave as “free-draining” polymers during electrophoresis and also
have a monotonic size-to-charge ratio, and hence elute as one peak in free-solution CE, regardless
of molecular size. In free-solution electrophoresis, the electric field pulls negatively charged DNA
to the cathode, as the positively charged buffer counterions that surround DNA migrate toward
the anode; solvated cations move around and through the DNA coil, giving it its “free-draining”
characteristic. The counterions effectively screen hydrodynamic interactions between different parts
of the DNA molecule, so the molecule adopts an open conformation, causing the electrophoretic
friction coefficient ζ to scale linearly with the number of monomers in the DNA chain M. The total
electrophoretic force F applied to the DNA molecule also scales linearly with the number of charges
along the DNA. Each of the four deoxynucleotides (dNTPs) carries the same amount of negative
charge from identical phosphate groups, causing the number of charges in a DNA molecule to scale
with M. The electrophoretic velocity of DNA in free solution is calculated by dividing total force
by the friction coefficient: v = F/ζ ∼ M/M ∼ M 0 , thus no size separation is achieved for DNA
in the absence of a gel. The sieving matrix used in capillary gel electrophoresis separates DNA
molecules by size as they collide with and interact with polymers and their movement is retarded.
Small DNA molecules move through the polymer matrix faster because they have fewer collisions,
so that DNA molecules to elute in order from smallest to largest. Interaction with polymers to effect
this size separation necessarily slows down the migration of DNA.
To accomplish free-solution separations of DNA would obviously require that DNA molecules
of different sizes are modified in such a way that they have size-dependent electrophoretic velocities,
which could be achieved either by changing the density of charge on each monomer or by changing
the dependence of the molecular friction coefficient on M. In free-solution conjugate electrophoresis
(FSCE), the linear scaling of DNA’s charge-to-friction ratio, described above, is abolished by attach-
ing another molecule (which we call a “drag-tag”) to each DNA chain, which perturbs the average
charge density per monomer. Noolandi first suggested this general idea in 199242 and mentioned
the possibility of using a natural protein as a drag-tag in 1993.43 The idea is potentially very power-
ful, since free-solution electrophoresis would be much easier to integrate into microfluidic devices
than gel electrophoresis. To try to make this concept a concrete reality, one has first to consider
what the structure and properties of such a “drag-tag” molecule ought to be. Should it be charged
or uncharged? If charged, should it be cationic or anionic? Experiments predict that negatively
charged drag-tags might only slightly modify the electrophoretic velocity of DNA while reducing
the separation efficiency, read-length, and effective drag, while a positively charged drag-tag could
interact electrostatically with the negatively charged DNA and/or the microchannel walls, increasing
band broadening. Thus, a neutral drag-tag seems to offer the best size-based separation of DNA;
the drag-tag should act as a sort of “parachute” behind the migrating DNA molecule, slowing down
the molecule. If we attach the exact same drag-tag to every DNA molecule in a mixture, the largest
molecule will migrate fastest because it experiences the smallest retardation from the additional
frictional drag. Hence, elution should occur in order from largest to smallest, which is the opposite
order of gel electrophoresis. This chapter concerns the present state-of-the-art of this bioconjugate
approach to DNA sequencing and genotyping.
Other nonelectrophoresis technologies are being developed in the search for the $100,000 and
$1000 genomes. Several of these technologies offer promising results and are expected to play a role
in the sequencing revolution currently at hand. Specifically, sequencing by synthesis (SBS) and nano-
pore detection methods are under development as potential nonelectrophoresis technologies.44−53
SBS is currently limited to average read-lengths of about 120 bases, is not as accurate as Sanger
sequencing,54 and requires difficult and demanding sample preparation, amplification, and genome
assembly.55 CE-MS is also being investigated as a DNA sequencing method, and is likely to have
relatively short read-lengths.56 While these nonelectrophoretic sequencing methods are being rapidly
developed, DNAsequencing by FSCE offers one potential avenue to the $100,000 genome, especially
if it is incorporated into an integrated “lab-on-a-chip”17,25–41 sequencing device. The separation time
for FSCE is faster than matrix-based electrophoresis, and the cost per base should be decreased by
eliminating the polymer matrix. This chapter will present an overview of the free-solution elec-
trophoresis of bioconjugates, including a brief summary of the theory, a description of the perturbing
entities, recently achieved free-solution DNA separations for genotyping and sequencing, and finally,
future directions for the development of this potentially powerful new technology.
12.2 FREE-SOLUTION CONJUGATE ELECTROPHORESIS OF DNA

Free-solution electrophoresis of DNA that has been conjugated to a friction-perturbing entity, which
we call a “drag-tag,” was not feasible until after the development of CE in the 1990s. Free-solution
electrophoresis generates a significant amount of heat, which must be dissipated. The narrow channels
used in CE have diameters in the order of 25–100 µm, which efficiently remove this “Joule heat.”
Free-solution conjugate electrophoresis (FSCE) was first examined quantitatively in 1994 by Mayer
et al.57 who developed the theory and named the method end-labeled free-solution electrophoresis
(ELFSE). However, we prefer the name FSCE (since the drag-tag is not a “label”) and will use that
name in this chapter. The original theory presented in 1994 has been further developed and was
recently reviewed.14 A brief summary of the general theory of FSCE will be presented here, along
with several “proof-of-concept” experiments.16,58–61
12.2.1 THEORY OF FSCE

If DNA is electrophoresed in an aqueous buffer only (i.e., in “free-solution”), the electrophoretic
mobility of an entire DNA molecule is essentially equal to the mobility of just one monomer unit
of DNA. Because DNA molecules of all lengths elute at exactly the same time by free-solution
electrophoresis, DNA is called a “free-draining” polymer15,58,62 (as if DNA molecules are in “free
fall”). The electrophoretic mobility of DNA µ is defined as the ratio between the electrophoretic
velocity v and electric field strength E. For a free-draining molecule, this ratio is equivalent to the
ratio of charge Q to friction ζ .57 As each of the four dNTP monomers has an identical negatively
charged phosphate group, the total charge Q of a molecule of DNA scales linearly with the number
of monomers in the DNA chain M. The electrophoretic friction coefficient ζ of a free-draining chain
also scales linearly with the number of monomers in the chain; hence, the electrophoretic mobility
of DNA in free-solution (µ0 ) is given by
v QE/ζ Q M
µ0 = = = ∝ . (12.1)
E E ζ M
Because both charge and friction scale linearly with the number of DNA monomers, µ(M) is
not length-dependent. However, when an uncharged drag-tag is attached to the DNA molecule, the
modified mobility is size-dependent in free-solution electrophoresis. In 1994, Mayer et al.57 theorized
that the mobility of this bioconjugate could be determined solely by dividing total electrical force by
total friction coefficient. Assuming that an uncharged drag-tag is used, they predicted the free-solution
mobility of the ssDNA conjugated to the drag-tag as
ρ(M) M
µ(M) ≈ ≈ µ0 , (12.2)
ζ (M + α) M +α
where µ0 = ρ/ζ is the mobility of untagged (free) ssDNA in free-solution electrophoresis and α is
the friction of the drag-tag in units of the friction of one base of DNA.
The original assumption made by Mayer et al.57 concerning how the velocity of the ssDNA +
drag-tag conjugate is calculated was found later to be true only in special cases, and to be determined
by the hydrodynamic conformation of the ssDNA + drag-tag conjugate.63 The theory was then further
developed to determine the probability that the experimental conditions coincided with the specific
hydrodynamic conformation of that special case. Luckily, that special case is actually the most
common when separating these charged-uncharged conjugates in free-solution electrophoresis.14,64
A brief explanation of the theory confirming this is presented below.
The ssDNA + drag-tag conjugate may adopt four basic types of hydrodynamic conformations,
depending on the relative sizes of the ssDNA and the drag-tag and the intensity of the electric field E
(Figure 12.1).64 The electrophoretic mobility of the DNA + drag-tag depends on the hydrodynamic
conformation of the conjugate when separated by free-solution electrophoresis. At equilibrium, the
neutral drag-tag is integrated into the random coil of the negatively charged DNA (Figure 12.1a).
When an electric field is applied during electrophoresis, the drag-tag itself does not have any intrinsic
mobility because it is not charged. The ssDNA moves through the electric field according to its
electrophoretic mobility and the retardation of the drag-tag. If a large enough electric field is applied
(E1 > E > E0 ), the DNA “engine” will begin to segregate itself away from the drag-tag as it is pulled
(a) Coil
Segregation without deformation

(b)
Stretched DNA
(c)
Steric segregation
(d)
FIGURE 12.1 Schematic representation of four potential conformations of DNA-label bioconjugates:

(a) label is part of the random coil of DNA, (b) label and DNA segregate without the DNA being deformed,
(c) label and DNA segregate, and DNA is stretched, (d) DNA is too small to form a coil around the label, forcing
steric segregation. (Reprinted with permission from Desruisseaux, C., et al. Macromolecules 34, 44–52, 2001.)
by the electric field (Figure 12.1b).64 When the applied electric field has a high intensity (E > E1 ),
the DNA and label are segregated, and the DNA becomes fully stretched in the electric field as it
is pulled through the capillary (Figure 12.1c).64 In extreme cases where the drag-tag is much larger
hydrodynamically than the DNA, the DNA is sterically segregated from the drag-tag because it is
unable to surround the large label completely (Figure 12.1d). As bioconjugate electrophoresis aims to
sequence large pieces of DNA, the theory focuses on determining the critical electric field intensity
(E0 ) required to move from the random coil hydrodynamic conformation to segregation without
deformation.
Before calculating the critical electric field strength E0 for the transition from the random coil to
the hydrodynamic segregation regime, the mobility of labeled DNA µ(M) in the random coil regime
must be determined. The uncharged drag-tag monomers cannot be assumed to have the same hydrody-
namic friction per unit as negatively charged DNA. Because each block of the DNA + drag-tag block
copolymer has different friction, the so-called blob theory is used to create “supermonomers” within
each block that have the same hydrodynamic properties and drag but different charge.14,65 Taking
into account the molecular weights of the “blobs” as determined by Kuhn lengths, the electrophoretic
mobility of the DNA-label conjugate simplifies to
Mc
µ = µ0 , (12.3)
Mc + α 1 M u
where Mc is the number of charged monomers of DNA, Mu is the number of neutral drag-tag
monomers, and α1 is the friction coefficient of each uncharged drag-tag monomer.14,64–69 Using
this equation, the value for E0 with typical experimental conditions was E0 ≈ 8 kV/cm,14 which is
much larger than typical electric fields used for CE (100–300 V/cm). The possibility of segregated
and stretched DNA + drag-tag conjugates is highly unlikely under typical electrophoresis conditions;
these conjugates are always in the random coil regime during electrophoresis.
The value of α = α1 Mu is critically important for determining the potential separation efficiency
of drag-tags for free-solution electrophoresis. Drag-tags are evaluated by their α-value, and higher
α-values correlate to a greater amount of drag imposed on each piece of DNA, indicating that larger
DNA sequencing fragments can be resolved by a particular drag-tag. The approximate elution time
of a DNA + drag-tag conjugate can be determined using the equation

L Mu
t(M) = 1 + α1 , (12.4)
µ0 E Mc
where L is the total length of the capillary channel. This equation clearly shows that longer pieces
of DNA (smaller Mu /Mc ratio) elute before shorter ones (larger Mu /Mc ratio and thus larger time),
which is the opposite order of traditional CGE. This equation also offers an easy way to calculate
α = α1 Mu .
The first iteration of FSCE theory predicted the potential sequencing of up to 2000 bases in less
than 1 h under perfect (diffusion-limited) conditions using a drag-tag with an α between 100 and
200.57 The initial theory, however, assumed that µ(M) and the diffusion coefficient D(M) could be
related using the Nernst–Einstein equation, which it turns out is not a valid assumption under free-
solution electrophoresis conditions.70 This assumption caused Mayer to overestimate the potential
performance of DNA sequencing by FSCE. Further development of the theory with a correction for
this error led to the following equation that estimates the maximum number of bases that can be
sequenced by a drag-tag with a specified α-value:
Mc (Mc + α1 Mu )5/4 ∼
1/2
= 215α1 Mu . (12.5)
The protein streptavidin was suggested as a potential drag-tag in the early ELFSE literature.43,57,71
Proteins are a natural choice for drag-tags because, in principle, they can be completely monodisperse
(homogeneous in size and structure) and are large enough to provide appropriate amount of drag.
Furthermore, streptavidin was chosen because it is close to electrostatically neutral under standard
experimental conditions14 and can be easily linked to DNA with biotin. The α-value of streptavidin is
approximately equivalent to the friction of 30 monomers of ssDNA.16 With an α of 30, the maximum
read-length of DNA sequencing with a streptavidin drag-tag is predicted by theory to be 129 bases.14
12.2.2 PROOF-OF-CONCEPT EXPERIMENTS

12.2.2.1 Streptavidin
The first confirmation of the potential capabilities of ELFSE was achieved in 1998 by Heller et al.58
when streptavidin was used as a drag-tag to separate double-stranded DNA (dsDNA).58 Using
biotinylated nucleotides and a polymerase enzyme, either one or both ends of the dsDNA were
biotinylated so the dsDNA could be conjugated, end-on, with either one or two streptavidins (one
on each end). Double-stranded DNA ladders (100 bp) with one or two terminal biotin moieties were
generated and conjugated to streptavidin. FSCE analysis of the conjugates was run in 1× TBE buffer
(89 mM Tris, pH 8.0, 89 mM boric acid, 2 mM EDTA), and the resulting electropherograms are
presented below (Figure 12.2).
Large α-values are necessary to achieve good resolution of DNA, as shown by the double-labeled
dsDNA ladder. When two drag-tags are attached to dsDNA, an α-value of 54 imparts enough drag
on the dsDNA so the 100 and 200 bp strands separate over approximately 2 min instead of less than
1 min with only one streptavidin.58 This first foray into FSCE also confirmed the prediction that
increased electric fields lead to higher resolution between peaks.58 The electrophoretic velocity is
36 (a)
0.5
32
0.4 2 streptavidins
28
µ/µ–1
slope = 54
24 0.3
20 0.2
16 0.1
1 streptavidin
12 slope = 23
0.0
8
54
4 3 2 1 0 2 4 6 8 10
I/M (1/kbp)
0
6.0 7.0 8.0 9.0 10.0 11.0
72 (b) 10x10–3
Absorbance (a.u.)
64 8
Fluorescence signal
56 6
48 4
40 2
32 0
24 8 3.5 4.0 4.5 5.0
9 76 5 Migration time (min)
16 10 4 3
2
8 1
0
6.0 7.0 8.0 9.0 10.0 11.0
FIGURE 12.2 Free-solution capillary electrophoresis analysis of 100 bp dsDNA ladder labeled on one
(a) or both (b) ends with streptavidin. The drag of two streptavidins clearly resolves the dsDNA fragments
more efficiently than one streptavidin. Inset of (a) is plot of µ0 /µ1 versus 1/M; the slopes of this plot are equal
to the α value of one or two streptavidin drag-tags. Plot inset of (b) shows the polydispersity of the strepta-
vidin used. The large unnumbered peak at the beginning of both electropherograms is the unlabeled dsDNA.
(Reprinted with permission from Heller, C., et al. Journal of Chromatography A 806, 113–121, 1998.)
also increased with higher electric fields without any loss of resolution, which is an advantage of
FSCE over matrix-based CE systems that must balance the increased speed of separations with loss
of resolution from increased electric fields.14,15
The polydispersity of the streptavidin used to separate the dsDNA (Figure 12.2b, inset) was
troublesome because it potentially causes each peak of DNA to split into two or more peaks. The
protein is fairly heterogeneous due to differing degrees of proteolysis and glycosylation as well
as the fact that it has not one, but four active sites to react with biotin. DNA sequencing requires
single-base resolution, which is clearly not possible with a polydisperse drag-tag that generates
two or more peaks for each length of ssDNA. When streptavidin was purified using a homebuilt,
preparative, nondenaturing polyacrylamide gel electrophoresis instrument, the polydispersity was
greatly decreased.16
To fully demonstrate the potential of ELFSE, sequencing fragments were generated with a
biotinylated primer. The purified streptavidin was attached to the fragments after cycle sequenc-
ing. Single-base resolution of the sequencing fragments in four colors was achieved for the first
100–110 bases in 18 min by ELFSE (Figure 12.3).16 The approximately monodisperse streptavidin
used here had an α-value of around 24. As mentioned in Section 12.2.1, single-base resolution of
129 bases of ssDNA is predicted for a drag-tag with an α of 30. Therefore, sequencing 100–110
bases with streptavidin is near the upper limit predicted by theory.
a a ag t g t g t c c t t tg t cg a t a c tgg t a c t a a t g c t t a ag c t
400
130 120 110 100 90
300
200
100
0
10.4 min 10.6 10.8 11.0 11.2
350
c g ag c c a t g g g c c c c t ag g ag a t c t c ag c t g g a
300
80 70 60
250
200
150
100
50
0
11.5 12.0 12.5 min
cg t ccg t a cg t t c g a a c c g t g a c c g g c a g c a a a a
200
50 40 30 23
150
100
50
0
13 14 15 16 17 18 min
FIGURE 12.3 DNA sequencing of M13mp18 in free-solution electrophoresis using purified streptavidin and
a biotinylated sequencing primer. The bases elute in reverse order compared to traditional CGE; the base
immediately after the primer elutes at 18 min. Electrophoresis conditions were 1X TAPS buffer (pH 8.5) with
7 M urea and 0.01% POP-6 as a dynamic wall coating, 34 cm effective length (45 cm total length) capillary
with 20 µm inner diameter. (Reprinted with permission from Ren, H., et al. Electrophoresis 20, 2501–2509,
1999.)
While reading 100–110 bases of DNA by free-solution CE is obviously not yet competitive with
current matrix-based and nonelectrophoresis DNA sequencing methods, it clearly shows ELFSE’s
potential. Based on the streptavidin drag-tag, a drag-tag with an α of 300 is predicted to resolve
ssDNA sequencing fragments up to 625 bases long.16 Separating sequencing fragments by microchip
electrophoresis may also be able to increase the read-length by applying a higher electric field than
is possible with traditional CE instruments.16,58 However, several issues must be resolved before
that many bases can be separated using ELFSE. First, the mobility of the fluorescent dye labels must
be determined to deconvolute the mobility shifts of the sequencing fragments induced by the four
dyes. Second, an effective capillary wall coating must be developed to further decrease interactions
between the walls and the bioconjugates.
Third, and most critical to long-read sequencing, is the development of a large, uncharged,
monodisperse drag-tag. While streptavidin was a natural choice of drag-tag for testing the first DNA
separation by FSCE, it will never be able to sequence more than 100–130 bases of DNA. Streptavidin,
like almost all natural proteins, folds into a fairly compact sphere under native (nondenaturing) con-
ditions. The frictional coefficient of spheres increases linearly with the radius; however, the radius
only increases with the one-third power of the molecular weight. Therefore, streptavidin would need
a molecular weight of approximately 30 million Da (a 600-fold increase in mass) to have an α-value
of 300.14 Also, as a natural protein, streptavidin has areas of positive and negative charge on its
surface that may adversely affect its ability to be an effective drag-tag. Natural, folded proteins like
streptavidin are thus not the best choice of drag-tags for long-read ssDNA sequencing. A monodis-
perse drag-tag that is uncharged and unstructured with a large α-value should theoretically provide
much better sequencing results. Denaturing conditions that unfold proteins cannot be used with strep-
tavidin without compromising the biotin-binding sites. If another method of end-on conjugation is
used, proteins of either random coil or α-helical secondary structure may be used. Two different
approaches have been taken toward developing appropriate drag-tags, and they will be discussed in
Section 12.3.
12.2.2.2 Oligosaccharides
While work on FSCE has mainly focused on separating DNA molecules, the Novotny group of
Indiana University has also successfully achieved free-solution separations of end-labeled, charged
oligosaccharides.59–61 Oligosaccharides are similar to DNA in that their charge-to-friction ratio
remains constant as length changes.59 Two different fluorescent labels were used as drag-tags
for carbohydrates: one to increase the charge-to-friction ratio of the oligosaccharides and one
to decrease it. The fluorescent label 8-aminonaphthalene-1,3,6-trisulfonic acid (ANTS) increased
the charge-to-friction ratio and separated the oligosaccharides by free-solution electrophoresis
with a migration order from smallest to largest. When the charge-to-friction ratio was decreased
by the attachment of 6-aminoquinoline, the migration order was reversed to largest to smallest
oligosaccharides, like ELFSE. Both types of labeled oligosaccharides were also run in a 1% linear
polyacrylamide (LPA) matrix to determine the impact of sieving matrix on separation. The matrix
increased the separation efficiency of the ANTS-labeled oligosaccharides but significantly decreased
separation of the oligosaccharides with a decreased charge-to-friction ratio. The LPA slowed the
larger oligosaccharides by a much larger percentage than the smaller ones, effectively cutting the
separation efficiency into half.59 The Novotny group has also applied FSCE to the separation of
heparin oligosaccharides61 and the study of how polygalacturonic acid interacts with metal ions.60
12.3 DRAG-TAGS FOR FSCE

Selection of an appropriate drag-tag is key to achieving successful DNA sequencing by ELFSE.
The ideal drag-tag for many DNA separations of interest would have a large value of α. With a large
hydrodynamic drag, higher resolution and performance can be achieved for the separation of large
sequencing fragments (greater than 200 bases). In addition to a high α-value, several other properties
are desired in the ideal drag-tag14
1. Complete monodispersity
2. Water solubility (DNA sequencing is performed under aqueous conditions)
3. Minimal to no electrostatic charge

4. Minimal adsorption to or nonspecific interaction with microchannel walls
5. Ability to be uniquely and stably attach to DNA (to ensure DNA is attached to exactly one
drag-tag)
The most important property of an ideal drag-tag is complete monodispersity, where every tag
is identical in charge and drag. When analyzed by free-solution CE, a monodisperse DNA + label
conjugate should correspond to exactly one peak on the electropherogram. A polydisperse DNA +
drag-tag conjugate, though, may show up as two or more ambiguous peaks on an electropherogram,
which is not conducive to single-base resolution of ssDNA sequencing fragments. Accurate DNA
sequencing would be challenging if not impossible with a polydisperse drag-tag.
Three different types of drag-tags have been examined in the search for appropriate drag-
tags for sequencing and genotyping: synthetic polymers such as poly(ethylene glycol) (PEG),69
poly-N-substituted glycines,68,72,73 and genetically engineered protein polymers.68,74–76 The dual
requirements for complete monodispersity and sufficiently large hydrodynamic drag eliminate many
commercially available synthetic polymers from consideration. Although completely monodisperse
PEGs are now available, their small sizes result in low hydrodynamic drag that is insufficient for
DNA separation. Larger PEG molecules with low polydispersity (PDI = 1.01) are also inadequate.
When an “engine” of fluorescently labeled monodisperse ssDNA is conjugated to such PEG sam-
ples, they separate into more than 100 peaks in free-solution electrophoresis, with single-monomer
resolution of the PEGs with different chain lengths.69 The development of poly-N-substituted
glycines and genetically engineered protein polymers for drag-tags will be discussed in the following
sections.
To characterize a drag-tag, a completely monodisperse, fluorescently labeled ssDNA molecule
is conjugated to a potentially polydisperse drag-tag. The DNA is conjugated to the drag-tag by
the reaction between a maleimide-terminated drag-tag and a 5 -thiol-terminated oligomer of DNA
(Figure 12.4). In electrophoresis, the monodisperse DNA acts like an “engine” that imparts the exact
same amount of force on each drag-tag. Each different drag-tag is represented by a peak on the
electropherogram of the free-solution separation; the polydispersity of the drag-tag can be evaluated
by the resulting number of peaks. If just one product peak is present, then the drag-tag is actually
O O
SH + Drag-tag
Dye-labeled DNA
O
O
S
Dye-labeled DNA Drag-tag
FIGURE 12.4 Conjugation of DNA to drag-tag by reacting dye-labeled DNA terminated with a 5 -thiol
group with a maleimide-terminated drag-tag. (Modified from Vreeland, W. N., et al. Analytical Chemistry 73,
1795–1803, 2001.)
monodisperse and can be used for DNA sequencing or genotyping. More than one product peak
present indicates that the drag-tag has an underlying polydispersity that may not have been identified
during a previous purification. Potential drag-tags are evaluated and their effective drag α (measured
in units equivalent to the drag from one base of DNA) is determined using Equation 12.3 and the
identity µ0 /µ ≈ t/t0 .14,68,69,73,75–78 .
12.3.1 SYNTHETIC POLYMERS

FSCE was first applied to the evaluation of synthetic polymers as potential drag-tags, specifically
the water-soluble polymer (PEG) whose antifouling properties should keep it from interacting with
the capillary walls. PEG with a polydispersity index (weight average molecular weight/number
average molecular weight) of 1.01 was tested, which is considered “monodisperse” by commer-
cial standards. The PEG-DNA bioconjugate was expected to show up on the electropherogram as
essentially one peak. The necessity for absolute monodispersity of a drag-tag was quickly realized
upon separating the PEG-DNA bioconjugates by free-solution CE (Figure 12.5). The polydispersity
of the “monodisperse” PEG immediately ruled out its potential application as a drag-tag for DNA
sequencing. Ever since this first experiment with PEG, FSCE has been the “gold standard” by which
the monodispersity of a drag-tag is determined.
12.3.2 POLYPEPTOIDS
Polypeptoids are similar to polypeptides except the side chains are attached to the amide nitrogen of
the peptide backbone instead of the α-carbon79–81 (Figure 12.6). Peptoids are synthesized by solid-
phase synthesis on a peptide synthesizer using a “submonomer” approach79,81 that proceeds by
adding each “monomer” in a two-step process. With this method, many different side chain groups
2.5
20-base DNA + nominal 5 kDa PEG
2.0
Fluorescence (520 nm)
1.5 Free
DNA
DNA + PEG
1.0 conjugates
0.5
0.0
15 20 25 30 35 40
Time (min)
FIGURE 12.5 Electropherogram of FSCE analysis of DNA–PEG bioconjugate. A 20-base ssDNA oligomer
was conjugated to a 5 kDa PEG. Separation was achieved in 1× TAPS buffer in a 100 cm (total
length) capillary with 25 µm diameter. Pressure injection was used (138 kPa·s), 300 V/cm running field
strength with 5.6 µA current. (Modified from Vreeland, W. N., et al. Analytical Chemistry 73, 1795–1803,
2001.)
H3C
N
H NH2
n
FIGURE 12.6 General structure of polypeptoid, with methoxyethyl as a glycine-like side chain. To attach
to ssDNA, a thiol group is part of the R group at the end of the peptoid. (Modified from Haynes, R. D., et al.
Bioconjugate Chemistry 16, 929–938, 2005; Zuckermann, R. N., et al. Journal of the American Chemical
Society 114, 10646–10647, 1992.)
can be incorporated into the growing peptoid, including side chain analogues of amino acids. In
particular, methoxyethyl has been used as a glycine-like side chain in peptoids for drag-tags because
an N-methoxyethylglycine (NMEG) drag-tag is water-soluble, uncharged, and shows minimal or no
interaction with the capillary walls.
12.3.2.1 Linear Polypeptoids

Peptoid drag-tags are advantageous because they can be purified to monodispersity by reversed-
phase high-performance liquid chromatography (RP-HPLC). However, the solid-phase method can
only synthesize linear polypeptoids up to 60 NMEG monomers before the yield drastically decreases
from < 100% coupling efficiency of solid-phase synthesis.69,78 The α-value of peptoid drag-tags
is α ≈ 0.2–0.25 per peptoid monomer.78 Therefore, a linear peptoid would have to be 100–125
monomers in length to have an α of 25 similar to streptavidin. Because solid-phase synthesis methods
are limited to 60 monomers, achieving a large α like 300 for long-read sequencing requires something
other than a linear peptoid drag-tag.
12.3.2.2 Branched Polypeptoids

Introducing oligopeptoid branches onto polypeptoid backbones is one approach to increase the effec-
tive drag of these synthetic drag-tags. At certain intervals in the synthesis of the peptoid backbone,
side chains with a terminal -amino group are incorporated. Branches with terminal glutamic acid
residues can then be attached to the backbone via amide bonds. Using a 30mer peptoid backbone
with five evenly spaced “grafting” sites, three differently sized branches were appended to the drag-
tag backbone to determine the impact of branch size on the effective drag. The smallest “branch”
was an acetyl group, and the other two branches were linear polypeptoid NMEGs with a length of
either 4 or 8 monomers. The α-value increased approximately linearly as the molar mass increased,
with the effective drag of the octamer-branched drag-tag more than doubling that of the drag-tag
with the acetyl groups as “branches.”77 The original FSCE theory did not predict this; branched
objects are more compact than if the branches were added serially to the end of a linear drag-tag
and theory did not predict that the branch length would make a dramatic impact on the drag. The
theory was recently re-examined, though, and now suggests several parameters to take into account
when designing branches to help achieve the highest α-value.14 The search for the most effective
drag-tag for long-read sequencing is now focused on finding the largest possible branch for a peptoid
backbone.
12.3.2.3 Analysis of Solid-Phase Synthesis Products

FSCE can also show impurities present in solid-phase synthesis products that are not clearly evident
by conventional purification methods. RP-HPLC with ultra violet (UV) detection is traditionally
used to analyze solid-phase synthesis products such as peptoids and peptides as well as to separate
them preparatively away from any deletion products. Since each synthetic monomer addition cycle
is less than 100% efficient,73 products that are one or two monomer units shorter than the desired
product must be purified out to ensure monodispersity of the final polypeptide or polypeptoid.
The separating power of FSCE was compared with analytical RP-HPLC traces after RP-HPLC
purification (Figure 12.7). The one peak in the RP-HPLC trace actually separates into one large peak
(a) 2.5
RP-HPLC: "Purified" 60-mer peptoid

2.0
UV Absorbance (214 nm)
1.5
1.0
0.5
0.0
5 10 15 20 25 30 35 40 45
Time (min)
(b)
FSCE: 20-base ssDNA + "Purified" 60-mer peptoid
1200
1000
RFU (488 nm ex/521 nm em)
800
600
400
200
0
10 12 14 16 18 20 22
Time (min)
FIGURE 12.7 Analysis of polydispersity of a 60-unit NMEG polypeptoid drag-tag post RP-HPLC purification.
RP-HPLC (a) shows one peak, but FSCE (b) shows a series of smaller peaks that may impact the drag-
tag’s read-length. (Reprinted from Vreeland, W. N., et al. Bioconjugate Chemistry 13, 663–670, 2002. With
permission.)
with several smaller peaks in FSCE. While this drag-tag might have an α appropriate for sequencing
and is practically monodisperse by RP-HPLC, the slight polydispersity might negatively impact the
read-length. Overall, FSCE is very useful for evaluating both polydispersity and the α-value of a
potential drag-tag and is recommended before any drag-tag is used to separate or sequence DNA.
12.3.3 GENETICALLY ENGINEERED PROTEIN POLYMERS

In order to meet all of the criteria for an ideal drag-tag and avoid the drawbacks of synthetic polymers
or natural proteins, the approach of nonnatural “protein polymers” has been employed. Protein
polymers are produced by genetic engineering and consist of a repeating amino acid sequence. In
general, this repetitive sequence can be a mimic of a natural sequence motif (e.g., elastin, silk) or
highly nonnatural and designed specifically for a particular structure (e.g., α-helix) or function.82–99
The properties of protein polymers are determined by the selected DNA sequence that codes for the
final protein and can be customized by arranging specific amino acids in a desired order. Unlike
conventional, synthetic polymerization techniques, protein-based materials produced in biological
systems, such as the bacterium Escherichia coli, offer much better control over the properties of the
final product.100 Protein engineering is not limited to just the 20 naturally occurring amino acids.
Several researchers have successfully incorporated a wide variety of unnatural amino acid analogs
into the protein production machinery of living cells.101–105
12.3.3.1 Designing the Drag-Tag Sequence

The first and most important step in producing a nonnatural protein polymer is to design the actual
sequence. In the genetic code, three DNA bases (a codon) code for either a particular amino
acid sequence or a stop codon. The genetic code is degenerate; several different codons can corre-
spond to the same amino acid. When designing the DNA sequence, it is important to use codons
preferred by the particular host species because expressing highly repetitive protein sequences can
rapidly deplete the available amino acid pool in the cell.
As mentioned previously, the ideal drag-tag characteristics are monodispersity, water solubility,
lack of charged residues, unique attachment to DNA, and minimal adsorption to the microchannel
walls. Based on these requirements, phenylalanine, isoleucine, tyrosine, and tryptophan were elim-
inated from consideration due to their strongly hydrophobic aromatic groups. Cysteine also was
excluded because its thiol side-chain is highly reactive and may oxidize to form disulfide bonds.
Several charged amino acids (lysine, arginine, histidine, aspartate, and glutamate) were similarly
eliminated from consideration.
The first-generation of six protein polymer drag-tags that were designed for FSCE sequencing
were designated PZ-1 through PZ-6. Glutamic acid was included, sparsely, in four of the sequences
(even with the negative charge) to address concerns about water solubility. Valine and leucine were
included in PZ-3 and PZ-4 to explore the effects of increasing hydrophobicity in drag-tags. The
sequences were designed to generate random-coil structured protein polymers so each residue is
exposed to the outside solvent and is not buried inside a sphere (such as in streptavidin). Unfortunately,
not all of the proteins expressed well in E. coli, providing an important lesson to the designers and
indicating the same persistence would be required. Of the proteins that could be expressed, none of
these first six sequence designs was suitable as drag-tags for FSCE sequencing for the various reasons
summarized in Table 12.1 and discussed in Reference 76. A key issue with the first generation drag-
tags was the observed polydispersity that was later attributed to structural degradation of the drag-tag
under acidic conditions, and more specifically, to deamidation of the glutamine into negatively
charged glutamic acid.75
Therefore, the second generation drag-tags (PZ-7 and PZ-8) replaced glutamine with either
serine or threonine, but of these only the PZ-8 sequence expressed well. Variants of this sequence
containing a few serine-to-arginine mutations at different positions were also produced. The first of
TABLE 12.1
Summary of Initially Studied (Unsuitable) Drag-Tag Sequences
Name Repeating Sequence Comments
PZ-1 Gly-Ser-Gly-Gln-Gly-Glu-Ser Good expressibility and water solubility; too hydrophilic for RP-HPLC
PZ-2 Gly-Ala-Gly-Gln-Gly-Glu-Ala purification; contains both glutamine and glutamic acid;
similar electrophoretic mobility to DNA makes it poor drag-tag
PZ-3 Gly-Val-Gly-Gln-Gly-Glu-Val Both: not well expressed by E. coli; could not be purified in good yield
PZ-4 Gly-Leu-Gly-Gln-Gly-Glu-Leu
PZ-5 Gly-Ala-Gly-Gln-Gly-Asn-Ala Only obtained in impure preparation; both asparagine and glutamine
included, which turned out to be nonideal; soluble in water only with addition
of 1 M urea
PZ-6 Gly-Ala-Gly-Gln-Gly-Ser-Ala Contains glutamine
PZ-7 Gly-Ala-Gly-Ser-Gly-Ser-Ala Not well expressed
PZ-8 Gly-Ala-Gly-Thr-Gly-Ser-Ala Well expressed and well behaved in aqueous solution
these variants has a mutation where one of every nine serines is replaced by a positively charged
arginine. This mutant sequence and the original PZ-8 sequence have been expressed in at least three
protein chain lengths (127, 253, and 505 amino acids). The average secondary structure of these
proteins was profiled by circular dichroism (CD) spectroscopy, and all had spectra characteristic of
random coil structures at 25◦ C in water. These results confirmed that the protein sequence had been
successfully designed to form random coil proteins upon expression and purification.
Much effort and discussion has focused on making linear protein polymers for drag-tags but
a parallel effort to create branched drag-tags is ongoing. These branched proteins have a sequence
similar to the PZ proteins with the one notable exception that they include multiple lysines at varying
intervals to be reactive sites for grafting on branches (either protein polymer or peptoid). A branched
protein polymer is expected to generate greater hydrodynamic drag than a linear protein polymer of
same length, based on the branched peptoid drag-tags.
12.3.3.2 Analysis of Protein Polymers as Drag-Tags

Unfortunately when these proteins are conjugated to DNA primers and analyzed in free-solution CE,
all but the two 127-amino-acid proteins show multiple peaks, despite the exclusion of glutamine
residues from the sequences. These additional peaks get more numerous as the protein polymer
length increases, rendering these proteins useless as drag-tags. The expected effective α-value can
still be determined by FSCE assuming the final peak in the electropherogram is the desired protein.
Figure 12.8 shows a plot of protein molecular weight versus α for each protein. A clear linear trend
exists for both series of proteins tested. The protein containing a few arginine mutations showed
higher “effective” α-values than their uncharged counterparts due to the positive charge “pulling”
the molecule in the opposite direction of the DNA. No significant detrimental interaction of the
positively charged drag-tags with the DNA or microchannel walls was observed. The inclusion of
even more positively charged residues may increase the hydrodynamic drag and avoid the need to
generate very long, uncharged, drag-tags to achieve high α-values.
Meeting the requirement of a completely monodisperse drag-tag has proven more challenging
than originally expected, and this work is still ongoing. Several theories have been explored and
eliminated as causes of the heterogeneity, including the possibility that cell lysate proteases degrade
the protein during purification. The process of histidine tag removal by formic acid with cyanogen
bromide was also eliminated as a possible cause by using an enterokinase cleavage site already
140
y = –12.79 + 0.0038578x R2 = 0.99725
y = 3.1328 + 0.0018438x R2 = 0.99653

120
100
80

60
40
20
PZ8 with Arg
PZ8
0
5000 1 × 104 1.5 × 104 2 × 104 2.5 × 104 3 × 104 3.5 × 104 4 × 104
Molecular weight (Da)
FIGURE 12.8 Chain length and charge of linear protein polymers (PZ-8 variants) versus effective α as
determined by FSCE.
coded into the expression vector to allow enzymatic removal of the histidine tag and by shortening
the cyanogen bromide reaction time from 24 to 4 h. Also, protein obtained from cells lysed by a
freeze/thaw process demonstrated the same polydispersity as those obtained from cells lysed by
ultrasonication. Even though MALDI-TOF mass spectrometry appears to only show a single peak, a
difference of one amino acid may be what generates additional peaks in the FSCE electropherogram.
12.3.4 DOUBLE-LABELED DNA

FSCE can also be used to test different labeling techniques. The dsDNA separations with streptavidin
drag-tags back in 1998 revealed an anomaly: the drag from a drag-tag on both ends of the dsDNA
was expected to be equivalent to double the drag of a drag-tag on only one end, but as can be seen in
Figure 12.2a (inset), the effective drag from dsDNA with drag-tags on both ends is 17% more than
double the drag of one drag-tag.16 Accordingly, the theory of ELFSE was re-examined to determine
the cause of this anomaly. The theory is thoroughly developed in Reference 68 to take these “end-
effects” into account. The monomer “blobs” at the ends of the bioconjugate molecule were found to
be weighted more than those in the middle of the bioconjugate in how they influence the effective
drag and electrophoretic mobility of the bioconjugate.68
To determine how the theory compared to experimental data, two different lengths of linear
polypeptoid NMEG drag-tags were used to test end-effects on ssDNA, and four different drag-tags
(one linear polypeptoid NMEG drag-tag, one branched NMEG peptoid drag-tag, and two different
protein polymer drag-tags) were used to test the effect of double-labeling dsDNA. One drag-tag on
both ends of ssDNA increased the effective drag α 6–9% over one drag-tag that was twice as long.
Double-labeling dsDNAincreased the α10–23% over one drag-tag twice as long.68 Figure 12.9 shows
4000
(a) 0
3
2000
1
2
Relative fluorescence
0
4 6 8 10 12
6000
0
(b)
3
3000 4
0
4 6 8 10 12
Time (min)
FIGURE 12.9 Free-solution capillary electrophoresis analysis of single- and double-labeled fluorescently
tagged ssDNA. (a) Is a separation with 20mer NMEG peptoid drag-tag, and (b) uses a 40mer. Peak 0 is
unlabeled, “free” DNA. Peak 1 is 40mer DNA oligomer with one drag-tag; peak 2 is 20mer DNA with one
drag-tag; peak 3 is 40mer DNA with two drag-tags; and peak 4 is 20mer DNA with two drag-tags. Separations
were achieved in 36 cm capillaries (47 cm total length; 50 µm inner diameter) with 22 V/cm electrokinetic
injection for 15 s and current of 15 µA. (Reprinted with permission from Meagher, R. J., et al. Electrophoresis
27, 1702–1712, 2006.)
the separation achieved with double-labeled ssDNA. Overall, the use of double-labeled DNA could
be a way to increase the effective drag imparted by a particular drag-tag without having to synthesize
another, much larger, drag-tag.
12.4 DNA ANALYSIS BY FSCE

12.4.1 DNA GENOTYPING: SINGLE-BASE EXTENSION
While DNA sequencing is often the “gold standard” in detecting disease-causing genetic mutations,
it is very expensive and time- and energy intensive. A majority of the genetic mutations responsible
for diseases such as sickle cell anemia, Alzheimer’s, and most cancers are single-nucleotide poly-
morphisms (SNPs) that are point mutations in a gene. Specifically, mutations in the p53 gene have
been implicated as the cause of many cancers; most are missense mutations that cause a change in
amino acid.106–111 Many different methods are currently used for detecting SNPs,112,113 including
single-base extension (SBE) assays. In SBE, the sample is amplified using primers that anneal
one base 5 of the known SNP location in a reaction mixture containing only fluorescently labeled
dideoxynucleotide terminators (ddNTPs) and no dNTPs. When the polymerase extends the chain
from the 3 end of the primer, the ddNTP incorporated corresponds to the complement of the SNP.114
Bioconjugate electrophoresis was first applied to SBE genotyping in 2002 to separate three SNPs
in the p53 gene in approximately 9 min, testing four different mutants.72 Free-solution electrophoresis
is important to SBE because it eliminates the polymer matrix cost. The degree of multiplexing can
also be increased much more easily with FSCE. The elution order and separation efficiency of peaks
for unique loci is determined by the length of drag-tag; a longer and more expensive primer is not
required. While the genotyping peaks were rather broad in these first separations, the advantages to
this method are clear: multiplexed separations are possible in free-solution CE and should be easily
achieved in microchip electrophoresis.
SBE-FSCE was further multiplexed in 2006 when 22 samples (21 mutants and 1 wild-type)
were tested for mutations at 16 loci.115 Thermal cycling times were decreased by 90%, and free-
solution electrophoresis of the multiplexed SBE-FSCE reaction was performed both in capillaries
(Figure 12.10a) and microfluidic devices (Figure 12.10b) with greatly improved separation resolution
over the 2002 study. Separations by microchip electrophoresis decreased the separation time from
20 min in capillaries to less than 70 s in an 8-cm-long chip. The SBE-FSCE technique was 96.3%
accurate in genotyping by this method; of 16 loci across 22 samples (total of 352 loci), 339 were
correctly genotyped, including 5 wild-type/mutant heterozygotes that were later confirmed by direct
sequencing.115
The “modular” approach of SBE-FSCE shows the promise of genotyping using free-solution
bioconjugate electrophoresis. Any drag-tag can be conjugated to any primer, allowing the design of
multiple different sets of primers to probe many loci. Because of the abbreviated thermal cycling
protocol and rapid separation on the microchip, this assay is a promising candidate to be incorporated
into an integrated microfluidic “lab-on-a-chip” device for genotyping.17,25–41,116
12.4.2 DNA SEQUENCING BY FSCE

Rapid, long-read DNA sequencing using capillary and/or microchannel electrophoresis is an ultimate
goal of FSCE. Free-solution electrophoresis in microchips, in particular, eliminates the difficulties of
loading viscous polymer matrix into the microchannels and ensures a higher run-to-run reproducibil-
ity. Because polypeptoids do not have sufficient drag capable of long-read sequencing, they have
not been tried as drag-tags for sequencing. The branched peptoid drag-tag with five 8mer NMEG
branches grafted to it had an α-value of approximately 17,77 and was used recently to sequence
almost 100 bases (unpublished results R. J. Meagher, R. D. Haynes). However, the 8mer NMEG
branches did not increase the effective drag of the branched polypeptoid drag-tag enough to make it
as effective as the original streptavidin drag-tag.
The recent successful cloning and purification of the PZ-8 protein polymer has allowed sequenc-
ing of end-labeled bioconjugates in free-solution to become a reality. The 127-amino-acid PZ-8
protein polymer has been purified almost to monodispersity, as mentioned previously. It has an α-
value of approximately 25, which is appropriate for ssDNA sequencing. Both PZ-8 variants have
been tested as sequencing drag-tags. The PZ-8 sequence variant with two serine-to-arginine muta-
tions has a higher effective drag than the unmutated PZ-8 (Figure 12.8) and thus was expected to have
a longer read-length. To create sequencing fragments, a 5 -thiol terminated M13 primer was first
conjugated to the pure, monodisperse 127-amino-acid PZ-8. A typical Sanger sequencing reaction
mixture, including the drag-tag-labeled M13 primer, M13 ssDNA template, dNTPs, fluorescent dye-
labeled ddNTPs, DNA polymerase, and a polymerase buffer is thermal cycled to create drag-tagged
ssDNA sequencing fragments.54 Notably, the sequencing reaction was carried out with the protein
polymer drag-tag already conjugated to the thiolated primer, allowing for easy post-amplification
cleanup and sample preparation.117 The fact that the 127-amino-acid protein polymer conjugated
to the 5 end of the thiolated 17mer M13 primer did not impact the DNA polymerase activity or
(a) 6000
G (221-3)
G
C
T
A
G (175-2)
G (173-1)
G (254-1)
A (330-2)
T (149-1)
A (328-1)
Fluorescence
3000
C (273-2)
C (198-1)
C (144-1)
C (196-2)
Residual
C (128-1)
primers C (249-3)
C (196-1)
T (173-2)
C (273-1)
8 10 12 14 16 18 20
Time (min)
(b) 1.6 1.6 C

A A
1.4
A G
1.4 1.2 T
1.0
Fluorescence
0.8
1.2
0.6 C
C 0.4
1.0
Fluorescence
0.2
0.0
0.8 35.5 36.0 36.5 37.0 37.5
Time (sec)
G
0.6 C
G
0.4 G
C
C T
A C C
0.2 G
C C T
0.0
34 36 38 40 42 44 44 48 50 52 54 56 58 60 62 64 66 68 70
Time (sec)
FIGURE 12.10 (See color insert following page 810.) Electropherograms of multiplexed SBE-FSCE sepa-
ration of 16 loci in a wild-type p53 gene. (a) CE separation in 36 cm capillary (47 cm total length, 50 µm inner
diameter) in 1× TTE (89 mM Tris, 89 mM TAPS, 2 mM EDTA), 7 M urea and 0.5% (v/v) POP-6 for dynamic
coating. (b) Separation on 8 cm microchip in 1× TTE (49 mM Tris, 49 mM TAPS, 2 mM EDTA) with 7 M urea
on chip coated with poly-N-hydroxyethylacrylamide. Both separations were performed at 55◦ C. (Reprinted
with permission from Meagher, R. J., et al. Analytical Chemistry 79, 1848–1854, 2007.)
primer–template hybridization is a positive step for the potential application of DNA sequencing by
FSCE in genome centers and on integrated microfluidic devices.
The first 120 bases of sequencing of the M13 ssDNA sequencing template were very clearly
resolved using the 127-amino-acid PZ-8 drag-tag with two serine-to-arginine mutations, and the
next 60 bases could be distinguished with prior knowledge of the sequence (Figure 12.11), giving
an effective read-length of nearly 180 bases.117 Comparing Figure 12.11 with Figure 12.3, the peaks
of the ssDNA sequencing analysis using the protein polymer drag-tag are significantly sharper. The
127-amino-acid protein polymer was also used to test the effect of capillary length, electrophoretic
velocity, and ionic strength of the buffer on resolution of the separation of sequencing peaks. The
longest capillaries allowed slightly narrower peaks with deeper valleys, at the cost of a fourfold
increase in separation time, while changing electrophoretic velocity did not drastically impact the
separations. Increased buffer concentrations slightly increased α, but higher buffer concentrations
caused more heat generation, which may degrade the quality of sequencing.117
MC = 220 190 170 150 130 120 110

GGAAAA GGG GA A A AACACA CC AACAC GCC TA G TT G GTG TCC TTTG TCGA TA
G TT TC AA CC C T C TT TC TT AAA T
C
T
A
8.0 8.5 9.0
MC = 100 90 80 70 60
CGGTA CTAATGC TTAAGCTGCAG C C A TGGG C C C C T A G G AG A T C T C AG C TG G A C G T
T
9.5 10.0 10.5 11.0
MC = 50 40 30 25 20 18
CCGT ACG T T C GA A C C G T G A C C G G C AG C A A A A T G T T G
12 14 16 18
Time (min)
FIGURE 12.11 (See color insert following page 810.) DNA sequencing analysis of M13mp18 ssDNA in
free-solution with genetically engineering protein polymer drag-tag in 36 cm capillary (47 cm total length,
50 µm inner diameter) with 312 V/cm field strength and electrokinetic injection of 1 kV for 20 s. Capillaries
were filled with 1× TTE (89 mM Tris, 89 mM TAPS, 2 mM EDTA) with 7 M urea and 0.5% (v/v) POP-6 to
dynamically coat the walls of the capillary. (Reprinted with permission from Meagher, R. J., et al. In preparation
2007.)
CTC TTC NNN NNN G CTC TTC NNN NNN

GAG AAG NNN NNN C GAG AAG NNN NNN
Eam1104 I Sap I
FIGURE 12.12 Eam1104 I and Sap I restriction enzyme recognition and cleavage sites are shown by the
arrows. The Sap I enzyme has the additional base requirement of G on the top strand and C on the bottom strand
as depicted here.

12.5.1 DRAG-TAG CLONING AND PRODUCTION
12.5.1.1 Standard Protocol for Cloning and Production of Protein Polymers
Protein polymers are synthesized using biological systems (often E. coli) and then purified from the
cell lysate by affinity chromatography. Chemical synthesis of peptides is limited to the production of
short chains (<50 amino acids), and it is similarly difficult to synthesize long DNA oligonucleotides.
The large repetitive gene necessary to express a protein polymer must be generated by joining several
DNA “monomers” with compatible end base-pair sequences using a ligase enzyme.
Several cloning methods have been developed over the years to create repetitive genes for protein
polymers.92,98,99,118–121 Nonetheless, obtaining large concatemer genes was still difficult because
of their limited yield after DNA ligation. Long concatemers could only be found by chance after
laborious screening of hundreds of colonies. In order to consistently create the long DNA con-
catemers needed for expression of protein polymer drag-tags, a modified version of the seamless
cloning method was developed called controlled cloning.74 This method allows for the construction
of large DNA concatemers without any specific sequence requirement using two Type II restric-
tion enzymes, Eam1104 I and Sap I, which cleave downstream of their nearly identical recognition
sites (Figure 12.12). Figure 12.13 illustrates the controlled cloning method for the example of
doubling a multimer.
Each “monomer” of the gene is a 100 bp synthetic oligonucleotide that encodes three tandem
repeats of a seven amino acid sequence. This oligomer is polymerase chain reaction (PCR)-amplified,
enzymatically digested by the Eam1104 I enzyme to generate the 63-bp monomer DNA fragments,
and then self-ligated to generate a ladder of different sized multimers. The gene is inserted into a
cloning vector, pUC18, and colonies obtained after transformation of the plasmid into E. coli cells
are screened for the desired insert sizes. Controlled cloning can then be used to combine one gene
of the desired size with another, which may be the same or different in sequence and size.
After DNA sequencing confirms that the gene is correct, it is then moved into an expression
plasmid, pET-19b, encoding an N-terminal histidine fusion tag, which is used to purify the desired
protein by immobilized metal affinity chromatography (IMAC) from other cellular proteins. The
protein is expressed in BLR(DE3) E. coli cells using either LB (Luria-Bertani) or TB (Terrific Broth)
media. Cells are grown under antibiotic selection at 37◦ C until they reach mid-log growth phase
with an OD600 value between 0.6 and 0.8. The cells are then induced to begin protein expression
with the addition of the lactose analog, isopropyl-β-D-thiogalactopyranoside (IPTG) at 1 mM, and
grown for 3–4 h before harvesting by centrifugation. Cells are resuspended in buffer then lysed
using multiple freeze/thaw cycles as well as ultrasonication. The clarified cell lysate is purified using
column chromatography under denaturing conditions on either nickel or cobalt-chelated resin, which
bind to the histidine tag. The protein with the affinity tag is eluted off the column with the addition
of imidazole. The collected fractions are analyzed using sodium dodecyl sulfate–polyacrylamide
gel electrophoresis (SDS–PAGE). Elution fractions are dialyzed against water to remove salts and
small molecular weight contaminants before being lyophilized into a dry powder. Typical protein
Sap I Eam 1104 I
PCR product of concatamer

Sap I digestion
Eam 1104 I digestion
P-
P-
CIP
OH-
T4 DNA ligation
OH-
Eam 1104 I digestion
OH-
T4 nucleotide kinase
P-
Re-ligation with pUC18 recipient vector
Sap I
Recombinant Concatemer gene

plasmid
Sap I
FIGURE 12.13 Controlled cloning method for generation of long concatemer genes. (Reprinted with
permission from Won, J. I. and Barron, A. E. Macromolecules 35, 8281–8287, 2002.)
polymer yields range from 20 to 30 mg/L of cell culture. The purified protein is then analyzed for
purity by RP-HPLC on a C4 column. Peaks are detected by UV absorbance at 220 nm. MALDI-TOF
analysis yields the molecular mass of the protein sample which can then be compared to the expected
mass. Removal of the affinity tag is achieved by reaction with cyanogen bromide in formic acid for
24 h.122 Cleavage of the affinity tag is essential as it contains multiple positively and negatively
charged amino acids that can adversely affect the drag-tag’s properties in free solution.76
12.5.1.2 Polypeptoid Synthesis

Polypeptoid synthesis reactions are carried out on automatic peptide synthesizers (ABI 433A;Applied
Biosystems, Foster City, CA). Fmoc-rink amide resin (0.25 mmol; Nova Biochem, San Diego, CA)
is deprotected by treating with 20% (v/v) piperdine (ABI) in dimethylformamide (DMF; Fisher
Scientific, Itasca, IL) in two 15-min treatments. To assemble the polypeptoid chain, alternating
cycles of bromoacetylation and amine displacement of the bromine from the N-terminal alkyl halide
are performed (Figure 12.14). To bromoacetylate, the resin is vortexed for 45 min with 4.3 mL of
1.2 M bromoacetic acid (Aldrich, Milwaukee, WI) in DMF and 1 mL of diisopropylcarbodiimide
(DIC; Aldrich). The liquid is drained from the resin, and the resin is washed four times with 7 mL
O
O
Br
NH2 Br NH
HO
Rink amide N C N
resin
O
CH3
N-Methylpyrrolidone
40x or 60x
O O
O O
O O
H HO
N N
NH NH
O O
20,40,60
O O
N C N
H3C N H3C
N
N
OH
In Dimethylformamide
FIGURE 12.14 Reaction scheme for synthesis of linear NMEG polypeptoids of length 20, 40, or 60 monomer
units and addition of maleimidopropionic acid. (Reprinted with permission from Vreeland, W. N., et al.
Bioconjugate Chemistry 13, 663–670, 2002.)
of DMF. After the DMF wash, 4 mL of 1M methoxyethylamine (Aldrich) in N-methylpyrrolidone

(NMP; Aldrich) is introduced to the reaction vessel and vortexed for 45 min to displace the amine
and introduce the methoxyethyl side-chain moiety to the growing chain. The 4× DMF wash is then
repeated.72,73,115
The peptoid chain is cleaved from the resin by treating it with 95% trifluoroacetic acid (TFA;
Aldrich) in water for 30 min. After the peptoid is cleaved from the resin and dissolved in water,
RP-HPLC is used to purify the peptoids (C18 packing, Vydac, 5 µm, 300 Å, 2.1 × 250 mm). A linear
gradient of 10–60% B in A is typically run at 60◦ C over 50 min at a flow rate of 0.1 mL/min, where
solvent A is 0.1% TFA in water and B is 0.1% TFA in acetonitrile. Polypeptoid concentrations are
detected by UV absorbance at 214 nm. To fully purify the peptoids, preparative HPLC (C18 packing,
Vydac, 15 µm, 300 Å, 22 × 250 mm) is run using the same solvents, gradient, time, and a flow rate
of 12 mL/min.72,73,115
12.5.2 DNA + DRAG-TAG CONJUGATION

Conjugation of DNA to a drag-tag can be used for one of two purposes. To test the monodis-
persity and effective drag of a newly produced drag-tag, a fluorescently labeled DNA oligomer
should be used so the bioconjugate can be tested by electrophoresis immediately postconjugation.
For a sequencing or genotyping reaction, however, DNA primers should not have a fluorescent
label if fluorescently labeled ddNTPs are to be used. DNA and drag-tags are conjugated through
the use of a maleimide linker. For DNA to be attached to a maleimide-activated drag-tag, it
must first be modified to have a thiol (-SH) functionality on its 5 terminus. Any DNA–DNA
disulfide bonds present must be reduced to ensure the highest degree of conjugation efficiency.
To reduce the DNA, 2 nmol of DNA primer are incubated with a 20:1 molar excess of Tris
(2-carboxyethyl)phosphine (TCEP; Acros Organics, Morris Plains, NJ) at 40◦ C for 90 min.115
Protein polymer drag-tags are activated with a maleimide by the addition of sulfosuccinimidyl
4(N-maleimidomethyl)cyclohexane-1-carboxylate (Sulfo-SMCC; Pierce Biotechnology, Rockford,
IL). A mixture of a 10:1 excess of Sulfo-SMCC to 1.2 mg of protein polymer in 80 µL of 100 mM
sodium phosphate buffer (pH 7.2) is vortexed for 1 h at room temperature. Excess Sulfo-SMCC
is removed from the activated protein polymer drag-tag by gel filtration with a Centri-Sep column
(Princeton Separations, Adelphia, NJ). The activated, purified protein polymer is frozen, lyophilized
and then resuspended to a concentration of 10 mg/mL.117 Polypeptoid drag-tags are activated with
a maleimide during the solid-phase synthesis on the peptide synthesizer before cleavage. To do this,
a mixture of 500 µL of 1.2 M maleimidopropionic acid (Fluka, Buchs, Switzerland) in DMF and
144 µL of DIC is added to the resin and mixed for 45 min, then washed three times with DMF
(Figure 12.14).73 After cleavage and purification by HPLC, the peptoid is frozen, lyophilized to a
golden, viscous oil and then dissolved in water to a concentration of 10 mg/mL.
To conjugate the activated drag-tag to the reduced DNA, 90 pmol of DNA is mixed with
2.5 nmol of drag-tag in a total volume of 10 µL of sodium phosphate buffer at pH 7.2 (Figure 12.4).
The mixture is then incubated at room temperature for 3–24 h.115,117 A large excess of drag-tag to
DNA is necessary to ensure nearly complete conjugation of drag-tags to each DNA molecule.
12.5.3 THERMAL CYCLING PROTOCOLS

Single-base extension (SBE) reactions are commonly used both for genotyping and also for testing
drag-tag + DNA primer conjugations. Because neither the drag-tag nor the DNA primer is fluo-
rescently labeled, one fluorescently labeled ddNTP is added to the DNA so the bioconjugate can
be detected by laser-induced fluorescence (LIF). SBE reactions require a sequencing polymerase,
polymerase buffer, dye-labeled ddNTPs, DNA primer, and DNA template. To this end, the SNaPshot
Multiplex Kit (ABI) has been frequently used for SBE reactions. The SNaPshot kit premix includes a
sequencing polymerase, reaction buffer, and dichlororhodamine (dRhodamine) dye-labeled ddNTPs.
SBE reactions are prepared by mixing 2.5 µL of the SNaPshot premix, 1 pmol of primer + drag-tag
conjugate, 0.025–0.10 pmol of template DNA, 0.5 µL of 125 mM HCl, and sufficient water for a total
volume of 5 µL. The mixture is then thermocycled for 5 cycles at 96◦ C for 2 s (denaturation), 51.5◦ C
for 5 s (annealing), and 60◦ C for 10 s (extension). Centri-Sep gel filtration columns are then used to
remove excess dye terminators and buffer salts from the reaction mixture before electrophoresis.72,115
DNA sequencing reactions differ from SBE reactions in that they also include dNTPs in the
reaction mixture to extend the chain further than just one base by Sanger sequencing. For ELFSE
sequencing, the following reaction mixture is used: 5 µL of SNaPshot premix, 8 nmol of dNTPs,
4.2 pmol of M13 primer (5 -X1 GTTTTCCCA-GTCACGAC, where X1 is a 5 -C6 thiol linker)
conjugated to a drag-tag, 0.16 µg of M13mp18 control DNA template (Amersham Biosciences,
Piscataway, NJ), and sufficient water for a total volume of 10 µL. The mixture is then thermocycled
for 26 cycles at 96◦ C for 5 s (denaturation), 50◦ C for 5 s (annealing), and 60◦ C for 30 s (extension).
For purification, Centri-Sep gel filtration columns are used to remove excess dye terminators, buffer
salts, and dNTPs before electrophoresis.117
12.5.4 ELECTROPHORESIS CONDITIONS

Capillary electrophoresis separations of DNA + drag-tag conjugates are performed in a denaturing
buffer: 1× TTE (89 mM Tris, 89 mM TAPS, 2 mM EDTA) with 7 M urea and a 0.5–1% (v/v) POP-5
or POP-6 polymer solution (ABI) for a dynamic wall coating agent to suppress electroosmotic flow
(EOF) and prevent adsorption to capillary walls. Filtering the buffer with a 0.45 µm filter before
loading it into the capillaries and buffer reservoirs reduces noise. In theory, any capillary array
instrument with LIF detection (preferably 4-color) can be used for ELFSE separations. Capillaries
with an effective length from inlet to detector of 36, 50, and 80 cm have been used to separate
both ELFSE sequencing and genotyping fragments. Before analysis by electrophoresis, samples
should be denatured at 95◦ C for 30–60 s and snap-cooled on ice for 2–5 min. Typical electrophoresis
conditions include electrokinetic injection with a potential of 1–2 kV applied for 5–30 s and running
voltage of 15 kV, all at 55◦ C.68,115,117
12.6 FUTURE DIRECTIONS

In order for DNA sequencing by FSCE to make an impact on the sequencing field, sequencing of at
least 400 bases at a time is necessary. At present, a monodisperse drag-tag with a significantly larger
effective drag α is the major obstacle blocking long-read sequencing by free-solution electrophoresis
of DNA + drag-tag conjugates. Long protein polymers of the PZ-8 family are currently the most
promising large drag-tags because the (GAGTGSA) sequence has proved amenable to this method.
Research is also underway to investigate the possibility of achieving drag-tags with appropriate
α-values for long-read sequencing by attaching very large branches to peptoid backbones.
As can be seen in Figure 12.11, the final peaks of FSCE sequencing are over-separated while the
beginning peaks are under-separated. For long-read sequencing to be efficient and competitive, the
smaller molecules (final peaks) need to be less resolved, and the large molecules (first peaks) need
better resolution. The over-separation of the last peaks does not limit the read-length of the sequencing
but does increase the separation time. The number of bases of sequencing read by FSCE is, however,
limited by the under-separation of the first peaks. FSCE separations have always been performed
under conditions to minimize EOF; however, strong EOF should cause rapid elution of small DNA
fragments with inherently low electroosmotic mobility. Therefore, performing FSCE separations in
the presence of strong EOF was predicted to reverse the migration order of DNA fragments so that
they elute in order from smallest to largest, like CGE. The possibility of performing FSCE with EOF
was examined theoretically to determine if it would also slow down the longer molecules and speed
up the smaller ones when the migration order was reversed.66
A theoretical study of FSCE with EOF showed that FSCE in the presence of EOF mobility exactly
equivalent to the free-solution mobility of DNA (scaled EOF mobility = 1) will more than double the
2 101
3.5 0 114
3
1.5 124
Size resolution factor (bases)
2.5
2 1.2 156
1.1 179
1.5
1 235
1
Scaled Read
EOF length
0.5 mobility (bases)
0
0 50 100 150 200 250
Number of ssDNA bases
FIGURE 12.15 Graph of size resolution factor as a function of the number of ssDNA bases that are sequenced.
When the size resolution factor is ≤1, single-base resolution of ssDNA fragments is possible. The scaled
EOF mobility is the EOF mobility divided by the free-solution mobility of DNA. Therefore, the read-length is
predicted to be the longest when the EOF mobility exactly equals the free-solution mobility of DNA. (Reprinted
with permission from McCormick, L. C. and Slater, G. W. Electrophoresis 27, 1693–1701, 2006.)
read-length of sequencing. The counterflow of EOF not only reverses the migration order of DNA
fragments but also speeds up small fragments while slowing down long fragments. The impact of
the scaled EOF mobility on read-length in bases of ssDNA is shown in Figure 12.15, which used the
effective drag of streptavidin as a comparison. According to these results, approximately 100 bases of
ssDNA that were sequenced in Figure 12.3 could actually have been 235 bases of ssDNA if the EOF
mobility was accurately tuned to be equal to the mobility of DNA in free-solution.66 According to this
theory, FSCE with a scaled EOF mobility of 1 could be a significant contributor in the search for long-
read free-solution sequencing. The combination of FSCE with EOF, a drag-tag with a significantly
large α-value, and electrophoretic separations on a microfluidic device offers the opportunity for
sequencing by FSCE to be competitive with current methods of DNA sequencing. Both glass and
plastic microfluidic devices can be easily functionalized with different coatings, allowing EOF to be
tuned for FSCE separations to test how experimental data compare to the theory of FSCE with EOF.
Once long-read sequencing by FSCE is achieved, FSCE has the potential to be seamlessly integrated
into a “lab-on-a-chip” integrated microfluidic device.25–37,40,41,123,124 If long-read sequencing in
free-solution can be incorporated into a fully integrated microfluidic device, it has the chance to be
competitive in the race for the $100,000 genome.
ACKNOWLEDGMENT
The authors would like to thank Louisa R. Carr for her help in the preparation of this manuscript.
REFERENCES
1. Lander, E. S. et al. Initial sequencing and analysis of the human genome. Nature 409, 860–921 (2001).
2. Venter, J. C. et al. The sequence of the human genome. Science 291, 1304–1351 (2001).
3. Kling, J. The search for a sequencing thoroughbred. Nature Biotechnology 23, 1333–1335 (2005).
4. Service, R. F. Gene sequencing—The race for the $1000 Genome. Science 311, 1544–1546 (2006).
5. Collins, F. S., Lander, E. S., Rogers, J., Waterston, R. H., and Conso, I. H. G. S. Finishing the
euchromatic sequence of the human genome. Nature 431, 931–945 (2004).
6. Collins, F. S., Green, E. D., Guttmacher, A. E., and Guyer, M. S. A vision for the future of genomics
research. Nature 422, 835–847 (2003).
7. Zhou, H. H. et al. DNA sequencing up to 1300 bases in two hours by capillary electrophoresis with
mixed replaceable linear polyacrylamide solutions. Analytical Chemistry 72, 1045–1052 (2000).
8. NIH News Release: NHGRI seeks next generation of sequencing technologies, http://www.genome.
gov/12513210 (2004).
9. Bonetta, L. Genome sequencing in the fast lane. Nature Methods 3, 141–147 (2006).
10. Project., R. f. a. H. C. G. Report of the Working Group on Biomedical Technology (2005).
11. Recommendation for a Human Cancer Genome Project: Report of the Working Group on Biomedical
Technology. http://www.genome.gov/15015123 (2005).
12. Ledford, H. Kudos, not cash, is the real X-factor. Nature 443, 733 (2006).
13. Pennisi, E. GENOMICS: On your mark. Get set. Sequence! Science 314, 232 (2006).
14. Meagher, R. J. et al. End-labeled free-solution electrophoresis of DNA. Electrophoresis 26, 331–350
(2005).
15. Viovy, J. L. Electrophoresis of DNA and other polyelectrolytes: Physical mechanisms. Reviews of
Modern Physics 72, 813–872 (2000).
16. Ren, H. et al. Separating DNA sequencing fragments without a sieving matrix. Electrophoresis 20,
2501–2509 (1999).
17. Paegel, B. M., Blazej, R. G., and Mathies, R. A. Microfluidic devices for DNA sequencing: Sample
preparation and electrophoretic analysis. Current Opinion in Biotechnology 14, 42–50 (2003).
18. Paegel, B. M., Emrich, C.A., Weyemayer, G. J., Scherer, J. R., and Mathies, R.A. High throughput DNA
sequencing with a microfabricated 96-lane capillary array electrophoresis bioprocessor. Proceedings
of the National Academy of Sciences of the United States of America 99, 574–579 (2002).
19. Liu, S. R., Shi, Y. N., Ja, W. W., and Mathies, R. A. Optimization of high-speed DNA sequencing on
microfabricated capillary electrophoresis channels. Analytical Chemistry 71, 566–573 (1999).
20. Dolnik, V., Liu, S. R., and Jovanovich, S. Capillary electrophoresis on microchip. Electrophoresis 21,
41–54 (2000).
21. Liu, S. et al. Automated parallel DNA sequencing on multiple channel microchips. Proceedings of the
National Academy of Sciences of the United States of America 97, 5369–5374 (2000).
22. Llopis, S. D., Stryjewski, W., and Soper, S. A. Near-infrared time-resolved fluorescence lifetime
determinations in poly(methylmethacrylate) microchip electrophoresis devices. Electrophoresis 25,
3810–3819 (2004).
23. Shi, Y. N. and Anderson, R. C. High-resolution single-stranded DNA analysis on 4.5 cm plastic
electrophoretic microchannels. Electrophoresis 24, 3371–3377 (2003).
24. Woolley, A. T. and Mathies, R. A. Ultra-high-speed DNA-sequencing using capillary electrophoresis
chips. Analytical Chemistry 67, 3676–3680 (1995).
25. Blazej, R. G., Kumaresan, P., and Mathies, R. A. Microfabricated bioprocessor for integrated nanoliter-
scale Sanger DNA sequencing. Proceedings of the National Academy of Sciences of the United States
of America 103, 7240–7245 (2006).
26. Burns, M. A. et al. An integrated nanoliter DNA analysis device. Science 282, 484–487 (1998).
27. Burns, M. A. et al. Microfabricated structures for integrated DNA analysis. Proceedings of the National
Academy of Sciences of the United States of America 93, 5556–5561 (1996).
28. Dunn, W. C. et al. PCR amplification and analysis of simple sequence length polymorphisms in mouse
DNA using a single microchip device. Analytical Biochemistry 277, 157–160 (2000).
29. Ferrance, J. P. et al. Developments toward a complete micro-total analysis system for Duchenne
muscular dystrophy diagnosis. Analytica Chimica Acta 500, 223–236 (2003).
30. Fredlake, C. P., Hert, D. H., Mardis, E. A., and Barron, A. E. What is the future of electrophoresis in
large-scale genome sequencing? Electrophoresis 27, 3689–3702 (2006).
31. Giordano, B. C., Ferrance, J., Swedberg, S., Huhmer, A. F. R., and Landers, J. P. Polymerase chain
reaction in polymeric microchips: DNAamplification in less than 240 seconds. Analytical Biochemistry
291, 124–132 (2001).
32. Grover, W. H., Skelley, A. M., Liu, C. N., Lagally, E. T., and Mathies, R. A. Monolithic membrane
valves and diaphragm pumps for practical large-scale integration into glass microfluidic devices.
Sensors and Actuators B-Chemical 89, 315–323 (2003).
33. Kan, C. W., Fredlake, C. P., Doherty, E. A. S., and Barron, A. E. DNA sequencing and genotyping in
miniaturized electrophoresis systems. Electrophoresis 25, 3564–3588 (2004).
34. Khandurina, J. et al. Integrated system for rapid PCR-based DNA analysis in microfluidic devices.
Analytical Chemistry 72, 2995–3000 (2000).
35. Koh, C. G., Tan, W., Zhao, M. Q., Ricco, A. J., and Fan, Z. H. Integrating polymerase chain reac-
tion, valving, and electrophoresis in a plastic device for bacterial detection. Analytical Chemistry 75,
4591–4598 (2003).
36. Lagally, E. T. and Mathies, R. A. Integrated genetic analysis microsystems. Journal of Physics
D-Applied Physics 37, R245–R261 (2004).
37. Lagally, E. T. et al. Integrated portable genetic analysis microsystem for pathogen/infectious disease
detection. Analytical Chemistry 76, 3162–3170 (2004).
38. Legendre, L. A., Bienvenue, J. M., Roper, M. G., Ferrance, J. P., and Landers, J. P. A simple, valveless
microfluidic sample preparation device for extraction and amplification of DNA from nanoliter-volume
samples. Analytical Chemistry 78, 1444–1451 (2006).
39. Paegel, B. M., Yeung, S. H. I., and Mathies, R. A. Microchip bioprocessor for integrated nanovolume
sample purification and DNA sequencing. Analytical Chemistry 74, 5092–5098 (2002).
40. Ugaz, V. M., Elms, R. D., Lo, R. C., Shaikh, F. A., and Burns, M. A. Microfabricated electrophoresis
systems for DNA sequencing and genotyping applications: current technology and future directions.
Philosophical Transactions of the Royal Society of London Series A-Mathematical Physical and
Engineering Sciences 362, 1105–1129 (2004).
41. Woolley, A. T. et al. Functional integration of PCR amplification and capillary electrophoresis in a
microfabricated DNA analysis device. Analytical Chemistry 68, 4081–4086 (1996).
42. Noolandi, J. A new concept for sequencing DNA by capillary electrophoresis. Electrophoresis 13,
394–395 (1992).
43. Noolandi, J. A new concept for separating nucleic-acids by electrophoresis in solution using hybrid
synthetic end labeled nucleic-acid molecules. Electrophoresis 14, 680–681 (1993).
44. Bai, X. P. et al. Photocleavable nucleotide analogues for DNA analysis. Biochemistry 42, 8597–8598
(2003).
45. Ju, J. Y. et al. Design and synthesis of fluorescence energy-transfer dye-labeled primers and
their application for DNA-sequencing and analysis. Analytical Biochemistry 231, 131–140
(1995).
46. Ju, J. Y., Ruan, C. C., Fuller, C. W., Glazer, A. N., and Mathies, R. A. Fluorescence energy-transfer dye-
labeled primers for DNA-sequencing and analysis. Proceedings of the National Academy of Sciences
of the United States of America 92, 4347–4351 (1995).
47. Li, Z. M. et al. A photocleavable fluorescent nucleotide for DNA sequencing and analysis. Proceedings
of the National Academy of Sciences of the United States of America 100, 414–419 (2003).
48. Seo, T. S. et al. Four-color DNA sequencing by synthesis on a chip using photocleavable fluorescent
nucleotides. Proceedings of the National Academy of Sciences of the United States of America 102,
5926–5931 (2005).
49. Seo, T. S. et al. Photocleavable fluorescent nucleotides for DNA sequencing on a chip constructed by
site-specific coupling chemistry. Proceedings of the National Academy of Sciences of the United States
of America 101, 5488–5493 (2004).
50. Margulies, M., Egholm, M., Altman, W. E., Attiya, S., Bader, J. S., Bemben, L. A., Berka, J.,
et al. Genome sequencing in microfabricated high-density picolitre reactors. Nature 437, 376–380
(2005).
51. Astier, Y., Braha, O., and Bayley, H. Toward single molecule DNA sequencing: Direct identification of
ribonucleoside and deoxyribonucleoside 5 -monophosphates by using an engineered protein nanopore
equipped with a molecular adapter. Journal of the American Chemical Society 128, 1705–1710
(2006).
52. Yan, H. and Xu, B. Q. Towards rapid DNA sequencing: Detecting single-stranded DNA with a solid-
state nanopore. Small 2, 310–312 (2006).
53. Shendure, J. et al. Accurate multiplex polony sequencing of an evolved bacterial genome. Science 309,
1728–1732 (2005).
54. Sanger, F., Nicklen, S., and Coulson, A. R. DNA sequencing with chain-terminating inhibitors.
Proceedings of the National Academy of Sciences of the United States of America 74, 5463–5467
(1977).
55. Rogers, Y. H. and Venter, J. C. Genomics—Massively parallel sequencing. Nature 437, 326–327
(2005).
56. Edwards, J. R., Ruparel, H., and Ju, J. Y. Mass-spectrometry DNA sequencing. Mutation Research-
Fundamental and Molecular Mechanisms of Mutagenesis 573, 3–12 (2005).
57. Mayer, P., Slater, G. W., and Drouin, G. Theory of DNA-sequencing using free-solution electrophoresis
of protein–DNA complexes. Analytical Chemistry 66, 1777–1780 (1994).
58. Heller, C. et al. Free-solution electrophoresis of DNA. Journal of Chromatography A 806, 113–121
(1998).
59. Sudor, J. and Novotny, M. V. End-label, free-solution capillary electrophoresis of highly-charged
oligosaccharides. Analytical Chemistry 67, 4205–4209 (1995).
60. Wiedmer, S. K., Cassely, A., Hong, M. F., Novotny, M. V., and Riekkola, M. L. Electrophoretic studies
of polygalacturonate oligomers and their interactions with metal ions. Electrophoresis 21, 3212–3219
(2000).
61. Sudor, J. and Novotny, M. V. End-label free-solution electrophoresis of the low molecular weight
heparins. Analytical Chemistry 69, 3199–3204 (1997).
62. Olivera, B. M., Baine, P., and Davidson, N. Electrophoresis of the nucleic acids. Biopolymers 2,
245–257 (1964).
63. Long, D., Viovy, J. L., and Ajdari, A. Simultaneous action of electric fields and nonelectric
forces on a polyelectrolyte: Motion and deformation. Physical Review Letters 76, 3858–3861
(1996).
64. Desruisseaux, C., Long, D., Drouin, G., and Slater, G. W. Electrophoresis of composite molecular
objects. 1. Relation between friction, charge, and ionic strength in free solution. Macromolecules 34,
44–52 (2001).
65. McCormick, L. C. et al. Desruisseaux, C., and Drouin, G. Capillary electrophoretic separation of
uncharged polymers using polyelectrolyte engines—Theoretical model. Journal of Chromatography
A 924, 43–52 (2001).
66. McCormick, L. C. and Slater, G. W. A theoretical study of the possible use of electroosmotic flow to
extend the read length of DNAsequencing by end-labeled free solution electrophoresis. Electrophoresis
27, 1693–1701 (2006).
67. McCormick, L. C. and Slater, G. W. The molecular end effect and its critical impact on the behavior
of charged-uncharged polymer conjugates during free-solution electrophoresis. Electrophoresis 26,
1659–1667 (2005).
68. Meagher, R. J. et al. Free-solution electrophoresis of DNA modified with drag-tags at both ends.
Electrophoresis 27, 1702–1712 (2006).
69. Vreeland, W. N. et al. Molar mass profiling of synthetic polymers by free-solution capillary
electrophoresis of DNA-polymer conjugates. Analytical Chemistry 73, 1795–1803 (2001).
70. Nkodo, A. E. et al. Diffusion coefficient of DNA molecules during free solution electrophoresis.
Electrophoresis 22, 2424–2432 (2001).
71. Ulanovsky, L., Drouin, G., and Gilbert, W. DNA trapping electrophoresis. Nature 343, 190–192 (1990).
72. Vreeland, W. N., Meagher, R. J., and Barron, A. E. Multiplexed, high-throughput genotyping by single-
base extension and end-labeled free-solution electrophoresis. Analytical Chemistry 74, 4328–4333
(2002).
73. Vreeland, W. N., Slater, G. W., and Barron, A. E. Profiling solid-phase synthesis products by free-
solution conjugate capillary electrophoresis. Bioconjugate Chemistry 13, 663–670 (2002).
74. Won, J. I. and Barron, A. E. A new cloning method for the preparation of long repetitive polypeptides
without a sequence requirement. Macromolecules 35, 8281–8287 (2002).
75. Won, J. I., Meagher, R. J., and Barron, A. E. Characterization of glutamine deamidation in a long,
repetitive protein polymer via bioconjugate capillary electrophoresis. Biomacromolecules 5, 618–627
(2004).
76. Won, J. I., Meagher, R. J., and Barron, A. E. Protein polymer drag-tags for DNA separations by
end-labeled free-solution electrophoresis. Electrophoresis 26, 2138–2148 (2005).
77. Haynes, R. D., Meagher, R. J., Won, J. I., Bogdan, F. M., and Barron, A. E. Comblike, monodisperse
polypeptoid drag-tags for DNA separations by end-labeled free-solution electrophoresis (ELFSE).
Bioconjugate Chemistry 16, 929–938 (2005).
78. Vreeland, W. N. and Barron, A. E. Free-solution capillary electrophoresis of polypeptoid-
oligonucleotide conjugates. Polymer Preprints 41, 1018–1019 (2000).
79. Kirshenbaum, K. et al. Sequence-specific polypeptoids: A diverse family of heteropolymers with stable
secondary structure. Proceedings of the National Academy of Sciences of the United States of America
95, 4303–4308 (1998).
80. Simon, R. J. et al. Peptoids—A modular approach to drug discovery. Proceedings of the National
81. Zuckermann, R. N., Kerr, J. M., Kent, S. B. H., and Moos, W. H. Efficient method for the prepara-
tion of peptoids [oligo(N-substituted glycines)] by submonomer solid-phase synthesis. Journal of the
American Chemical Society 114, 10646–10647 (1992).
82. Welsh, E. R. and Tirrell, D. A. Engineering the extracellular matrix: A novel approach to polymeric
biomaterials. I. Control of the physical properties of artificial protein matrices designed to support
adhesion of vascular endothelial cells. Biomacromolecules 1, 23–30 (2000).
83. Panitch, A., Yamaoka, T., Fournier, M. J., Mason, T. L., and Tirrell, D. A. Design and biosynthesis
of elastin-like artificial extracellular matrix proteins containing periodically spaced fibronectin CS5
domains. Macromolecules 32, 1701–1703 (1999).
84. Trabbic-Carlson, K., Setton, L. A., and Chilkoti, A. Swelling and mechanical behaviors of
chemically cross-linked hydrogels of elastin-like polypeptides. Biomacromolecules 4, 572–580
(2003).
85. Zhou, Y. T., Wu, S. X., and Conticello, V. P. Genetically directed synthesis and spectroscopic anal-
ysis of a protein polymer derived from a flagelliform silk sequence. Biomacromolecules 2, 111–125
(2001).
86. Asakura, T. et al. Synthesis and characterization of chimeric silkworm silk. Biomacromolecules 4,
815–820 (2003).
87. Fukushima, Y. Genetically engineered syntheses of tandem repetitive polypeptides consisting of

glycine-rich sequence of spider dragline silk. Biopolymers 45, 269–279 (1998).
88. Chilkoti, A., Dreher, M. R., and Meyer, D. E. Design of thermally responsive, recombinant polypeptide
carriers for targeted drug delivery. Advanced Drug Delivery Reviews 54, 1093–1111 (2002).
89. Chilkoti, A., Dreher, M. R., Meyer, D. E., and Raucher, D. Targeted drug delivery by thermally
responsive polymers. Advanced Drug Delivery Reviews 54, 613–630 (2002).
90. Lee, J., Macosko, C. W., and Urry, D. W. Elastomeric polypentapeptides cross-linked into matrixes
and fibers. Biomacromolecules 2, 170–179 (2001).
91. Di Zio, K. and Tirrell, D. A. Mechanical properties of artificial protein matrices engineered for control
of cell and tissue behavior. Macromolecules 36, 1553–1558 (2003).
92. McMillan, R. A. and Conticello, V. P. Synthesis and characterization of elastin-mimetic protein gels
derived from a well-defined polypeptide precursor. Macromolecules 33, 4809–4821 (2000).
93. Megeed, Z., Cappello, J., and Ghandehari, H. Genetically engineered silk-elastinlike protein polymers
for controlled drug delivery. Advanced Drug Delivery Reviews 54, 1075–1091 (2002).
94. Nagarsekar, A. et al. Genetic synthesis and characterization of pH- and temperature-sensitive
silk-elastinlike protein block copolymers. Journal of Biomedical Materials Research 62, 195–203
(2002).
95. Nagarsekar, A. et al. Genetic engineering of stimuli-sensitive silkelastin-like protein block copolymers.
Biomacromolecules 4, 602–607 (2003).
96. Farmer, R. S., Argust, L. M., Sharp, J. D., and Kiick, K. L. Conformational properties of helical protein
polymers with varying densities of chemically reactive groups. Macromolecules 39, 162–170 (2006).
97. Farmer, R. S. and Kiick, K. L. Conformational behavior of chemically reactive alanine-rich repetitive
protein polymers. Biomacromolecules 6, 1531–1539 (2005).
98. McGrath, K. P., Tirrell, D. A., Kawai, M., Mason, T. L., and Fournier, M. J. Chemical and biosynthetic
approaches to the production of novel polypeptide materials. Biotechnology Progress 6, 188–192
(1990).
99. McGrath, K. P., Fournier, M. J., Mason, T. L., and Tirrell, D. A. Genetically directed synthe-
ses of new polymeric materials—Expression of artificial genes encoding proteins with repeating
(Alagly)3proglugly elements. Journal of the American Chemical Society 114, 727–733 (1992).
100. Van Hest, J. C. M. and Tirrell, D. A. Protein-based materials, toward a new level of structural control.
Chemical Communications 1897–1904 (2001).
101. Kiick, K. L., Saxon, E., Tirrell, D. A., and Bertozzi, C. R. Incorporation of azides into recombinant
proteins for chemoselective modification by the Staudinger ligation. Proceedings of the National
102. Link, A. J., Mock, M. L., and Tirrell, D. A. Non-canonical amino acids in protein engineering. Current
Opinion in Biotechnology 14, 603–609 (2003).
103. Mock, M. L., Michon, T., van Hest, J. C. M., and Tirrell, D. A. Stereoselective incorporation of an
unsaturated isoleucine analogue into a protein expressed in E. coli. Chembiochem 7, 83–87 (2006).
104. Wang, C. C. Y., Seo, T. S., Li, Z. M., Ruparel, H., and Ju, J. Y. Site-specific fluorescent labeling of
DNA using Staudinger ligation. Bioconjugate Chemistry 14, 697–701 (2003).
105. Wang, J., Xie, J., and Schultz, P. G. A genetically encoded fluorescent amino acid. Journal of the
American Chemical Society 128, 8738–8739 (2006).
106. Birch, J. M. et al. Relative frequency and morphology of cancers in carriers of germline TP53 mutations.
Oncogene 20, 4621–4628 (2001).
107. Collins, F. S., Brooks, L. D., and Chakravarti, A. A DNA polymorphism discovery resource for research
on human genetic variation. Genome Research 8, 1229–1231 (1998).
108. Greenblatt, M. S., Bennett, W. P., Hollstein, M., and Harris, C. C. Mutations in the P53 tumor-
suppressor gene—Clues to cancer etiology and molecular pathogenesis. Cancer Research 54,
4855–4878 (1994).
109. Kirk, B. W., Feinsod, M., Favis, R., Kliman, R. M., and Barany, F. Single nucleotide polymorphism
seeking long term association with complex disease. Nucleic Acids Research 30, 3295–3311 (2002).
110. Soussi, T. and Beroud, C. Assessing TP53 status in human tumours to evaluate clinical outcome.
Nature Reviews Cancer 1, 233–240 (2001).
111. Soussi, T. and Lozano, G. p53 mutation heterogeneity in cancer. Biochemical and Biophysical Research
Communications 331, 834–842 (2005).
112. Chen, X. and Sullivan, P. F. Single nucleotide polymorphism genotyping: Biochemistry, protocol, cost
and throughput. Pharmacogenomics Journal 3, 77–96 (2003).
113. Landegren, U., Kaiser, R., Sanders, J., and Hood, L. A ligase-mediated gene detection technique.
Science 241, 1077–1080 (1988).
114. Pastinen, T., Partanen, J., and Syvanen, A. C. Multiplex, fluorescent, solid-phase minisequencing for
efficient screening of DNA sequence variation. Clinical Chemistry 42, 1391–1397 (1996).
115. Meagher, R. J. et al. Multiplexed p53 mutation detection by free-solution conjugate microchannel
electrophoresis with polyamide drag-tags. Analytical Chemistry 79, 1848–1854 (2007).
116. Lagally, E. T., Simpson, P. C., and Mathies, R.A. Monolithic integrated microfluidic DNAamplification
and capillary electrophoresis analysis system. Sensors and Actuators B-Chemical 63, 138–146 (2000).
117. Meagher, R. J., Won, J. I., and Barron, A. E. Sequencing of DNA by free-solution capillary
electrophoresis using a genetically engineered protein polymer drag-tag. Submitted (2006).
118. Mi, L. X. Molecular cloning of protein-based polymers. Biomacromolecules 7, 2099–2107 (2006).
119. Cappello, J. et al. Genetic-engineering of structural protein polymers. Biotechnology Progress 6,
198–202 (1990).
120. Meyer, D. E. and Chilkoti, A. Genetically encoded synthesis of protein-based polymers with precisely
specified molecular weight and sequence by recursive directional ligation: Examples from the elastin-
like polypeptide system. Biomacromolecules 3, 357–367 (2002).
121. Goeden-Wood, N. L., Conticello, V. P., Muller, S. J., and Keasling, J. D. Improved assembly of
multimeric genes for the biosynthetic production of protein polymers. Biomacromolecules 3, 874–879
(2002).
122. Gross, E. The cyanogen bromide reaction. Methods in Enzymology 11, 238–255 (1967).
123. Tan, H. D. and Yeung, E. S. Integrated on-line system for DNA sequencing by capillary electrophoresis:
From template to called bases. Analytical Chemistry 69, 664–674 (1997).
124. Tan, H. D. and Yeung, E. S. Automation and integration of multiplexed on-line sample preparation with
capillary electrophoresis for high-throughput DNA sequencing. Analytical Chemistry 70, 4044–4053
(1998).
13 Online Sample
Preconcentration for Capillary
Electrophoresis
Dean S. Burgi and Braden C. Giordano
CONTENTS
13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413

13.1.1 The Kohlrausch Regulation Function and Stacking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413
13.1.2 Three Solutions to the KRF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414
13.1.2.1 Continuous Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414
13.1.2.2 Discontinuous Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
13.2 Examples of Long Injection Stacking. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 416
13.3 Sample Stacking in Micellar Electrokinetic Chromatography. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
13.3.1 Modes of Sample Stacking in MEKC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 420
13.3.1.1 Field-Amplified Stacking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 420
13.3.1.2 Sweeping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
13.3.1.3 High-Salt Stacking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422
13.3.1.4 Electrokinetic Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
13.3.1.5 Field-Enhanced Sample Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
13.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
13.1 INTRODUCTION
The detection of material in a capillary column has proven to be interesting and dynamic over the
years.1 Ultimately, the ability to detect a given analyte comes down to the amount of material in
the cross-sectional area of the capillary at the detector location. For example in ultra violet (UV)
detection, the absorbance of light, A, is related to three terms in the Beer–Lambert law,2
A = εlc. (13.1)
The molar absorptivity, ε, is an intrinsic property of the analyte, the second term, l, is the path length
the light will travel through the medium; for capillary electrophoresis (CE), the second term is small
(25–100 µm) and cannot be easily change without impacting the resolution of the separation process.
Thus, manipulation of the third term, c, the concentration of the analyte of interest, has evolved into
a topic of significant interest.
13.1.1 THE KOHLRAUSCH REGULATION FUNCTION AND STACKING

While there are a number of methods by which analytes can be stacked or preconcentrated in CE,
ultimately, all stacking phenomena require a change in analyte velocity as the analyte transitions
from the sample zone to the background electrolyte (BGE).
413
The capillary can be viewed as a wire, which will carry current when a potential is placed across it.
The amount of potential drop is related to the resistivity of the cross-sectional area of the column. In
a wire, the free electrons move from one end of the wire to the other and their ability to move freely
is a measure of the resistance of the wire. In the liquid domain where CE resides, each ionic species
has the ability to carry a part of the current based on its charge, z, and concentration, c. Since each
ion has a unique mobility, the sum of all the ionic species in the cross-sectional area is a measure of
the resistance. Before the application of potential, the Kohlrausch regulation function (KRF)3 for a
given point in the capillary is defined by the following equation:
c i zi
ω= , (13.2)
|µi |
i
where ci is the concentration of a given ion, zi is the relative charge of strong electrolytes, and
µi is the mobility. For any given point in the capillary, the KRF value, ω, will remain constant.
With that in mind, upon application of voltage, species will move in and out of said cross-sectional
area and increase (stack) or decrease (destack) in concentration in order to satisfy this function. For
example, if the whole column was filled with a single mono electrolytic compound like sodium
chloride dissolved in water, Equation 13.2 would allow the calculation of the Kohlrausch value at
all points from one end of the capillary to the other like a wire. As different ions are added to the
system, the equation becomes a summation of all the ions and their mobilities as they move into
and out of the analysis point. It must be noted that each time the electric field is broken; the ionic
composition at the time of reconnection will generate a new Kohlrausch value. Thus, an injection
potential will have one Kohlrausch value and separation will have a second Kohlrausch value. The
delta in values between the two events can help further stack the samples into tighter bands.4
13.1.2 THREE SOLUTIONS TO THE KRF

13.1.2.1 Continuous Systems
There exist three types of states for the regulation function to operate. State 1 is where the BGE
is continuous throughout the capillary. It is assumed that in this state a low analyte concentration
is present in the sample plug and the presence of analyte does not affect the Kohlrausch value
anywhere throughout the capillary. Put simply, the sample plug has the same Kohlrausch value as
the BGE region surrounding it, thus, has no boundary. The flux of ions in and out of the cross-
sectional area is uniform. This state is known as zone electrophoresis. Upon application of voltage,
the sample compounds move out of the sample region as a function of their intrinsic mobilities and
more importantly, no velocity difference exists between analyte in the sample zone and analyte in
the BGE. Thus, the back end of the sample maintains equidistance from the front end of the sample.
That velocity, v, is a function of the ions mobility and is given in the following equation:
v = µE, (13.3)
where E is the electric field across the capillary. The width of the peak for each compound in
the sample would be as narrow as the first original injection plug barring the diffusion process.
Figure 13.1 shows several mobility markers from a short electrokinetic injection. The peak widths
are becoming wider as the amount of time spent in the column increase due simply to longitudinal
diffusion.
The second state is where the sample plug is higher or lower in BGE concentration than that of
the surrounding BGE. If the sample plug is lower in concentration, the back of the sample plug will
migrate rapidly toward the front of the plug because the Kohlrausch value at the boundary would
decrease the sample ion velocity whereas the rest of the sample is still under the original Kohlrausch
Online Sample Preconcentration for Capillary Electrophoresis 415
peak A_mobility 24.1

peak MF_mobility 28.5
peak E_mobility 18.1
peak H_mobility 16.0

peak B_mobility 21.9
peak C_mobility 20.7

200 RFU
peak M3_mobility 27.7
peak D_mobility 19.4
peak I_mobility 14.5

peak F_mobility 17.6
peak J_mobility 13.1

peak G_mobility 17.1
peak ML_mobility 11.6

875 2864
Seconds
FIGURE 13.1 An electropherogram of 13 mobility markers run on a MegaBACE 4000. Injection voltage
8 kV, 20 s, run voltage 8 kV, 60 min, and oven temperature is 44◦ C. Mobilities calculated using a 60 cm long
capillary, 43 cm to the detector. Concentration of each marker is 5 nM.
value. This difference in velocity is directly related to the difference in the concentrations as seen in
the following equation:
v = µE = µIR = µI/c. (13.4)
The resistance (R) is inversely proportional the concentration (c) of the electric field carrying ions.
Thus, sample is stacked into a small region and the concentration increases, which allows a lower
concentration sample to be detected. The front boundary (or the detector side of the sample plug) is
a stable stationary boundary and does not move; only the sample moves through it. This stacking
mechanism (see Figure 13.6) is commonly referred to as field-amplified sample stacking (FASS). If
the sample plug is higher in concentration than the surrounding BGE, the plug will spread out, as
the front of the plug has a higher velocity than the back end. The front boundary is a function of the
Kohlrausch value calculated at the time of electric application and will cause mixing of the sample
with the BGE to maintain the Kohlrausch value. This effect is known as electrodispersion and often
presents as analyte fronting, and is typically considered undesirable. In continuous systems, the
boundaries are stable and stationary and will not move except by electroosmotic flow (EOF).
13.1.2.2 Discontinuous Systems

The third state requires a multi-ion sample matrix (SM). In this system, there are several ions present
of varying mobilities at high enough concentrations to affect the electric fields. For example, the
column matrix can include an ion with a greater mobility than those in the sample. In effort to satisfy
the KRF, this ion with the greatest mobility will have the ion next largest mobility ion to stack up
behind it. The third ion and so further will fall in line; the ions are stacked into regions where the
concentration/mobility (KRF) equals the leading electrolytes value. Thus, low concentration ions
would stack into very narrow bands and higher concentration ions will be broader. This state is
commonly known as isotachophoresis (ITP). There are several review articles,5–7 which explain the
various states in more detail. In a discontinuous system, the boundaries between the ion species are
moving but stable through the system. They will move from one end of the column to the other and
are called system peaks.
Another version of a discontinuous system is called sample self-stacking.8,9 In this system, a
co-ion is placed into the sample matrix, which acts like the leading electrolyte in ITP but only during
the injection phase. After injection, the sample will continue to stack until the stacking ion leaves
the sample zone. In Reference 9, an equation is presented to calculate a stacking ratio known as acrit .
acrit = [(µb + µr )/(µa + µr )] × [(µa − µx )/(µx − µb )] (13.5)
where µx is the mobility of the last ion to be stacked, µa is the mobility of the major stacking co-ion,
µb is the mobility of the sample BGE (which may not be the same as the separation run buffer),
and µr is the mobility of the counter ion of the system. µa is the same as the leading electrolyte
in ITP and µb is the trailing electrolyte. Thus, a temporary ITP event is setup in the sample during
electrokinetic injection and during the first moments of the separation time.9 acrit is the theoretical
value at which the major co-ion in the sample system is proposed to self-stack the slower ions, µx ,
in a transitory ITP fashion.
In addition, in Reference 9, a value related to the major ion concentrations is give.
a = Cs /Cb (13.6)
where Cs is the concentration of the sample-stacking co-ion, Cb is the concentration of the sample
BGE. The value of “a” will indicate if the sample system is near or at a self-stacking domain. If the
“a” value is greater than or equal to acrit then self-stacking can occur. If not, then other stacking or
destacking conditions are the driving force for the analysis.
13.2 EXAMPLES OF LONG INJECTION STACKING

In general, sample stacking will occur when a boundary exists between the sample regions and the
surrounding electrolyte (BGE). This boundary must be stable for the length of time such that the
concentration of the sample of interest can be increased. When the boundary has passed on, stacking
will stop and normal zone electrophoresis occurs. Several standard methods for boundary formation
and use are described elsewhere1,4 like FASS, capillary ion electrophoresis,10 and ITP to name a few.
Several novel stable pseudostationary boundary techniques have appear, one technique dubbed
moving chemical reaction boundary11 forms a neutral zone in the column were the H+ ions combine
with the OH− ions to make a water zone. This zone is stable and leads to stacking across it. A second
method uses EOF balanced against back-pressure to hold the stacking boundary in a fixed location.12
The stability of the boundary allows for injections up to 3 h to be made.13 The most robust method
for long injects has arguably been electrokinetic injection out of a sample vial in which the column
has one Kohlrausch value and the vial a lower value. This method has shown to improve sample
injection by 1000 fold.14
When considering long electrokinetic injections, it becomes necessary to carefully consider
the ionic species included in the sample matrix and the BGE. Figure 13.2 illustrates the impor-
tance of matching the composition of the ions in the sample matrix to the BGE. In this case,
4-Morpholinepropanesulfonic acid was added to the sample matrix in order to achieve a desired pH
for purposes of a reaction; thus creating a discontinuity between the sample matrix and the BGE (100
mM TAPS). Consequently, a stable moving boundary (or system peak) results between the MOPS
in the sample matrix and N-[Tris(hydroxymethyl)methyl]-3-aminopropanesulfonic acid in the BGE
(noted in Figure 13.2 by arrows). With increasing injection time, this system peak becomes more
obvious, and more importantly, significant destacking of analytes that have to migrate through this
MF 2542.1
20 s inj
1773.1
1699.6
1252.8
1402
1.016.3
986.51
833.65
ML
641.17
715.02
545.75
537.44
469.88
449.54
703
213.12
306.13
269.6
150.33
250.6
133.3
6200.3
5175.6
40 s inj
3634.6
2900.5
MF
2320.9
2218.3
2136.1
2050.1
1594.6
1464.4
1356.1
ML
1204.1
1044.8
181.2
368.8
493.71
676.44
448.4
8673.9
6363.9
60 s inj
MF
3583.9
3345.9
2521.9
2092.3
ML
1816.9
1674.9
1378.5
1171.9
196.9
849.31
757.27
1433
526.02
MF
20110
11909
120 s inj
1039.3
ML
4965.4
3841.4
4605
2798.8
2714.3
2319.9
2372.5
1936.7
2383.2
1898.8
2286.2
1394.3
1519.2
386.15
882.82
952.22
0 45
Minutes
FIGURE 13.2 Injection of mobility markers prepared in a 100 mM MOPS buffer system (pH 7.9). The
injection time ranged from 20 s in the top image to 120 s in the bottom image. The separation was done on a
MegaBACE 1000, injection voltage was 15 kV, run voltage 15 kV, run time 45 min, oven is set to 44◦ C. The
MOPS system peak is located at about 28 min and indicated by an arrow.
boundary is observed. If the sample matrix composition is changed such that TAPS is in both the
sample matrix and BGE (Figure 13.3), the injection time can be increased to 120 s and the system
peak is located at the interface of the column and the sample vial.
To further push the limits of long injections a mobility difference between the major current
carrying ionic species is required, resulting in the formation of a stable moving boundary; sample
will stack in the plug during the time it takes to establish steady-state conditions. The leading or
terminating electrolyte serves to define the electric field that each of the analyte zones experience
such that band width will adjust itself to have the same concentration per cross section as the leading
or terminating ion (isotachophoresis). Using a self sample-stacking ion in the sample plug at the
appropriate “a” value can generate a transient ITP event to occur during the injection and the first
part of the separation time. The sample compounds isotachophoretically stack as the leading ion is
slowed down by the mismatch at the moving boundary between the sample plug and the BGE. After
the leading ion moves into the separation part of the column, the self-stacked stops and the sample
begins to separate as if under zone electrophoresis and the resolution are preserved.
Close proximity to acrit , as determined by the ratio of the concentration of the sample-stacking
co-ion to the sample BGE (Equation 13.6), will result in a steady-state ITP effect before the leading
ion is out of the sample zone. For the set of mobility markers shown in this work, a range of acrit can
20 s injection
2564.6
1703.1
1439.5
1256.5
582.66
712.71
490.91
257.29
285.67
579.13
712.94
328.33
177.76
167.17
992.2
849.4
1019
463.5
229.5
662
59039
20,000 RFU
21982
26973
120 s injection
21374
16971
16628
16121
15192
13414
12780
12817
11251
9584.1
5576.4
4021.3
3302.6
2804.3
2374.4
1412.3
1174.2
837.47
800 2300
Seconds
FIGURE 13.3 Comparison of a 20 s injection with a 120 s injection using 100 mM TAPS buffer (pH 8.3)
instead of 100 mM MOPS (pH 7.9) buffer in the sample.
be determined from Equation 13.5. On the basis of the use of chloride (from MgCl2 ) as the leading
electrolyte and TAPS as the trailing electrolyte; a range of acrit ’s from 1.23 to 34.24 can be calculated.
Figure 13.4 illustrates the difference between an “a” value (ratio of chloride concentration to TAPS
concentration) close to acrit versus not close to acrit . In the top trace, “a” is 0.5; it is apparent that the
mobility markers with acrit values closest to 0.5 are stacked more efficiently (peaks A and B with acrit
values of 1.23 and 1.43, respectively). Decreasing the TAPS concentration relative to the chloride
concentration to yield an “a” value of 4.0 drastically improves the efficiency of stacking to markers
with much higher acrit values, allowing injection times up to 150 s long with no loss in resolution or
distortion in peak shape. However, there is a functional limit to the amount of self-stacking that can
occur as determined by the time it takes the leading electrolyte to migrate out of the sample zone. If
electrokinetic injection continues beyond this point significant, destacking can occur as illustrated
by Figure 13.5.
Whether the stacking technique is called field-amplified stacking, transient isotachophoresis, or
sample self-stacking, the driving force in on-column sample preconcentration for charges species is
the need to satisfy the initial KRF conditions. Under separation and sample matrix conditions where
the KRF is difficult to determine, simply calculations such as acrit and “a” offer a rapid evaluation
of conditions for optimal transient ITP.
13.3 SAMPLE STACKING IN MICELLAR ELECTROKINETIC

CHROMATOGRAPHY
Micellar electrokinetic chromatography (MEKC) is a mode of CE primarily used in the separation of
neutral molecules. Since a mixture of neutral molecules will all migrate with EOF, it becomes neces-
sary to include a proxy within the BGE to afford a separation. This proxy is typically a micelle—an
MF
M3
a = 0.5
H
C
ML
F
I
G
MF
M3
a = 4.0
A
H
B
E
C
J
G
ML
FIGURE 13.4 Electropherogram of the mobility markers run under different sample self-stacking conditions
as defined by the concentration of chloride. The background buffer was 100 mM TAPS with 25 mM MgCl2 for
an “a” value = 0.5 and 12.5 mM TAPS BGE with 25 mM MgCl2 for an “a” value = 4.0. The x-axis is setup
as a relative scale in which the MF is set at zero and the ML peak is set as 1.0. All the other peak mobilities are
calculated relative to this unit scale. The data were generated on a MegaBACE 4000 under the same condition
as Figure 13.1.
MF
M3
H
E
B
Injection =150 s
D
C
J
G
ML
(A)
(B)
M3
Injection = 300 s
H
F
MF
ML
FIGURE 13.5 Electropherogram of the mobility markers at 150 s versus 300 s injection time. Data were
generated on a MegaBACE 4000 as per Figure 13.1.
aggregate of surfactant molecules. Neutral analytes migrate into and out of the micelle, and when in
the micelle, briefly take on the electrophoretic mobility of the micelle.
In the previous sections, several methods of sample preconcentration were described. Regard-
less of the mechanism, every example of sample stacking requires a change in the velocity of the
analyte as determined by the drive to maintain the KRF value. Unfortunately, neutral analytes do
not contribute to the KRF and a different perspective on sample stacking must be taken. There are a
number of stacking methods available when using MEKC. This section will describe the fundamental
mechanism and conditions by which sample preconcentration of neutral molecules occurs.
13.3.1 MODES OF SAMPLE STACKING IN MEKC

13.3.1.1 Field-Amplified Stacking
Field-amplified stacking is made possible by preparing sample in diluted surfactant solution such that
the conductivity of the sample matrix is lower than that of the BGE.15 This results in localized higher
electric field in the sample zone. When using normal polarity with a high-EOF velocity, this will
result in analyte and micelle in the sample zone to rapidly travel toward the negative electrode due
to the increased field strength in the sample zone. On entering the BGE, the analyte will experience
a lower electric field, resulting in stacking. This mechanism is illustrated in Figure 13.6.
Throughout this section, electropherograms illustrating the effectiveness of the various modes of
sample stacking will be presented. Figure 13.7 shows the electropherogram resulting from preparing
sample in a dilute micelle solution, thus under FASS-like conditions. The analyte set is a mix-
ture of neutral alkaloids. The bottom trace shows the separation resulting from preparing sample
in BGE, thus under normal zone electrophoresis conditions, the middle trace shows the resulting
electropherogram when sample is prepared in 0.1× BGE, and the top trace shows the resulting elec-
tropherogram when sample is prepared in 0.05× BGE. Experimental details are included in the figure
legend. It is clear from the figure that the peak height of several analytes increase, however, some
tailing is observed indicating that this method of preconcentration may be inefficient for this specific
analyte set.
Initial conditions
BGE SM BGE
On application of separation voltage
+ –
–
+ Low E High E Low E
FIGURE 13.6 Illustration of the mechanism of field-amplified stacking. In this case, the SM is dilute BGE
typically between 0.05× and 0.1× BGE.
2 mAU
3 5 6 8,9
10
4 7 1. Nicotine
2
2. Aconitine
3. Colchicoside
3 5 6 8,9 4. Strychnine
10 5. Thiocolchicoside
4
2 7 6. Colchicine
7. Benzoinmethyl ether
8. Yohimbine
8,9 9. Thiocolchicine
3 4 56 10
12 7 10. Emetine
0 1.5 3 4.5 6 7.5 9

Minutes
FIGURE 13.7 Separation of 10 alkaloids. Bottom trace: sample matrix is BGE. Middle trace: sample matrix
is 0.1× BGE. Top trace: sample matrix is 0.05× BGE. Experimental conditions: 50 µm internal diameter,
LT = 60 cm (50 cm effective length), linj = 10 mm (20 s × 0.5 psi), V = 25 kV (47 µA), UV absorbance at
214 nm. BGE: 10 mM Na2 B4 O7 , 80 mM sodium cholate.
13.3.1.2 Sweeping
Sweeping was originally defined as “the picking and accumulating of analytes by the pseudo sta-
tionary phase that fills the sample zone during application of voltage.”16 Sweeping was initially
implemented by preparing analytes in a sample matrix of equal conductivity to the BGE, but devoid
of surfactant sodium dodecyl sulfate.16,17 Terabe and coworkers18 later expanded the definition of
“sweeping” to include any situation in which the BGE contains a separation vector and the sample
matrix does not. The mechanism of stacking for an equal conductivity sample matrix is presented in
Figure 13.8 for a system with negatively charged micelles and normal EOF. Briefly, under a homoge-
nous electric field, neutral analytes are carried toward the negative electrode only to be swept up in
the micelles at the boundary of the sample region. The effectiveness of the stacking is dependent on
an analytes affinity for the micelle, the higher affinity of the micelle the larger the extent of stacking.
Terabe derived the following equation representing the extent of stacking as a function of an analytes
affinity for the micelle

1
lsweep = linj , (13.7)
1+k
where lsweep is the length of the swept analyte, linj is the length of the injected plug, and k is the
retention factor for a given analyte. Figure 13.9 shows the effectiveness of sweeping on the alkaloid
separation. Note that peaks 1 and 2 are not well stacked. This is expected as they have the lowest
affinity (k) for the cholate micelle.
Initial conditions
BGE SM BGE
+ –
+ –
Continuous field strength
FIGURE 13.8 Illustration of the mechanism of sweeping. In this case, the ionic strength of the sample matrix,
devoid of micelle, is adjusted such that a continuous field strength is achieved.
1.7 mAU
8,9 1. Nicotine
6 2. Aconitine
5 10 3. Colchicoside
3 4. Strychnine
4 7 5. Thiocolchicoside
2
6. Colchicine
1 8. Yohimbine
8,9 9. Thiocolchicine
3 4 56 10
12 7 10. Emetine
0 1.5 3 4.5 6 7.5 9

Minutes
FIGURE 13.9 Separation of 10 alkaloids. Bottom trace: sample matrix is BGE. Top trace: sample matrix is
30 mM Na2 B4 O7 . Experimental conditions: 50 µm internal diameter, LT = 60 cm (50 cm effective length),
linj = 10 mm (20 s × 0.5 psi), V = 25 kV (47 µA), UV absorbance at 214 nm. BGE: 10 mM Na2 B4 O7 , 80 mM
sodium cholate.
13.3.1.3 High-Salt Stacking

High-salt stacking is a mechanism of analyte preconcentration afforded by the stacking of micelles
at the negative electrode side of the sample plug.19 Analyte migrates into the stacked micelle region
experiencing a locally high retention factor due to the high micelle concentration. The nomenclature
“high-salt stacking” comes from the fact that in order to stack micelles the conductivity of the sample
Initial conditions
BGE SM BGE
+ –
+ High E Low E High E –
FIGURE 13.10 Illustration of the mechanism of high-salt stacking. In this case the ionic strength of the
sample matrix is adjusted such that it is 1.5–2.5 times the ionic strength of the BGE.
zone is typical 1.5–2.5× that of the BGE. High-salt stacking requires the following conditions be met:
µsample Esample < µev Eev (13.8)

µsample > µev , (13.9)
where µsample is the mobility of the sample stacking co-ion, Esample is the electric field in the sample
zone, µev is the mobility of the electrokinetic vector (micelle), and Eev is the field in the BGE.
The condition that the mobility of the sample stacking co-ion be greater than the mobility of the
micelle requires that the field in the BGE be greater than that in the sample zone, this is the need
to significantly increase the conductivity of the sample zone relative to the BGE. The stacking
mechanism is summarized in Figure 13.10. Note that the illustration shows micelles concentrating
at the detector side of the sample plug upon application of a separation voltage. The effectiveness of
high-salt stacking for the analysis of the alkaloid samples is shown in Figure 13.11. In this example,
peaks 1 and 2 are well stacked demonstrating the power of high-salt stacking for analytes with low k’s.
13.3.1.4 Electrokinetic Injection

Field-amplified stacking, sweeping, and high-salt stacking all require a discrete hydrodynamically
injected sample plug. Alternatively, electrokinetic injections can be utilized to introduce sample to the
column. Landers and coworkers20 presented electrokinetic stacking under continuous conductivity
conditions. Sample is injected with EOF and interacts with micelles included in the BGE. The
analytes concentrate at the sample zone/BGE interface. Provided a large enough deference between
the velocity of EOF and the net velocity of the micelle, it is possible to inject the equivalent of
multiple column volumes of analyte. This stacking scheme is illustrated in Figure 13.12, and the
corresponding example using this method for the alkaloid separation is shown in Figure 13.13.
13.3.1.5 Field-Enhanced Sample Injection

While the above method of electrokinetic injection requires only a single step, it is possible to couple
electrokinetic injections with a brief hydrodynamic injection of water.21 The column is initially full
of BGE, followed by a hydrodynamic injection of water. Analyte is then electrokinetically injected
until the measured current through the column is approximately 70–90% of the column when filled
1.7 mAU 10
8,9
6 1. Nicotine
5
4 2. Aconitine
3 3. Colchicoside
1
2 7
4. Strychnine
5. Thiocolchicoside
6. Colchicine
8,9 8. Yohimbine
3 4 56 10 9. Thiocolchicine
12 7
10. Emetine
0 1.5 3 4.5 6 7.5 9

Minutes
FIGURE 13.11 Separation of 10 alkaloids. Bottom trace: sample matrix is BGE. Top trace: sample matrix
is 150 mM NaCl. Experimental conditions: 50 µm internal diameter, LT = 60 cm (50 cm effective length),
linj = 10 mm (20 s × 0.5 psi) V = 25 kV (47 µA), UV absorbance at 214 nm. BGE: 10 mM Na2 B4 O7 , 80 mM
sodium cholate.
Initial conditions
SM BGE
+ –
–
+ Continuous field strength
FIGURE 13.12 Illustration of the mechanism of electrokinetic stacking injection. The sample matrix is
prepared devoid of micelles to an equal conductivity of the BGE.
exclusively with BGE. Upon application of voltage, analyte migrates into the water plug. As potential
continues to be applied across the capillary the water plug compresses—narrowing the analyte plug.
This scheme is shown in Figure 13.14. A separation of a mixture of alkylphenyl ketones using this
method is shown in Figure 13.15.
1.7 mAU
8,9
10
6
5 1. Nicotine
2. Aconitine
3
4 3. Colchicoside
7
4. Strychnine
2
5. Thiocolchicoside
6. Colchicine
1 7. Benzoinmethyl ether
8,9
8. Yohimbine
3 4 56 10
12 7 9. Thiocolchicine
10. Emetine
0 1.5 3 4.5 6 7.5 9

Minutes
FIGURE 13.13 Separation of 10 alkaloids. Bottom trace: sample matrix is BGE. Top trace: sample matrix
is 30 mM Na2 B4 O7 . Experimental conditions: 50 µm internal diameter, LT = 60 cm, linj = 20 s at 10 kV,
V = 25 kV (47µA), UV absorbance at 214 nm. BGE: 10 mM Na2 B4 O7 , 80 mM sodium cholate.
Initial conditions
SM H2O BGE
t=1 – +
t=2 – +
+
t=3 –
FIGURE 13.14 Illustration of the mechanism or field-enhanced sample injection.

65
55 I
45
(245 nm)
35
25
15
2 4 5
5 1 3 1. Heptanophenone
2. Valerophenone
–5
0 5 10 15 3. Butyrophenone
4. Propionophenone
2
65 5. Acetophenone
55 II
3
1
45
35
mAU
4
25
15 5
5
–5
0 5 10 15
Time (min)
FIGURE 13.15 Separation of five alkyl phenyl ketones. Sample matrix is 8.3 mM SDS, 16.7 mM phosphate
buffer (pH 2.5) (I). Water plug length of 8.7 cm for FESI (II). Experimental conditions: 50 µm internal diameter,
LT = 64.5 cm (56 cm effective length), linj = 2 s at 50 mbar, V = −20 kV, UV absorbance at 245 nm. BGE:
100 mM phosphate buffer (pH 2.25), 50 mM SDS. (Reprinted with permission from Quirino, J.P. and Terabe, S.,
Am. Chem. Soc., 70, 1893, 1998. Copyright 1998.)
13.4 CONCLUSIONS
The goal of this chapter was to present some of the basic concepts of online sample preconcentration
in capillary electrophoresis. The relationship between large volume injections of charged analytes
and the Kohlraush regulation function were discussed in Section 13.2. Irrespective of whether one
chooses a continuous or discontinuous buffer system, careful consideration of the background elec-
trolyte and sample matrix components and the formation of a long-term stable Kohlrausch boundary
are absolutely necessary for efficient large volume injections. Rapid evaluation of the potential
for sample stacking is made possible by calculations for “a” and acrit , precluding the sometimes
cumbersome KRF calculation. Stacking of neutral analytes, which do not contribute to the KRF,
was presented in Section 13.3. The mechanism and effect of several sample stacking modes were
presented. The underlying phenomenon common to each mode of sample stacking in MEKC is
that the analyte experiences a velocity change on partitioning into the micelle. In this regard,
preconcentration efficiency in MEKC is, at the very least, dependent on the analyte affinity for
the micelle and the mobility of the micelle. Ultimately, whether analytes are neutral or charged,
successful optimization of online preconcentration is made possible with a complete understand-
ing of the system components including the analytes of interest, sample matrix, and background
electrolyte.
REFERENCES
1. Landers, J.P., Ed., Handbook of Capillary Electrophoresis, 2nd ed., CRC Press, Boca Raton, FL, 1996.
2. Skoog, D.A. and West, D.M., Principles of Instrumental Analysis, 2nd ed., Chapter 6, p. 161, Saunders
College, Philadelphia, PA, 1980.
3. Kohlrausch, F., Ueber concentrations-verschiebungen durch electrolyse im inneren van Losungen und
losungegemischen, Ann. Phys. Chem., 62, 209, 1897.
4. Chien, R.L., Sample stacking revisited: A personal perspective, Electrophoresis, 24, 486–497, 2003.
5. Gerbauer, P. and Bocek, P., Theory of zone separation in isotachophoresis: A diffusional approach,
6. Beckers, J.L., Steady-state models in electrophoresis: From isotachophoresis to capillary zone
electrophoresis, Electrophoresis, 16, 1987–1998, 1995.
7. Urbanek, M., Krivankova, L., and Bocek, P., Stacking phenomena in electromigration: From basic
principles to practical procedures, Electrophoresis, 24, 466–485, 2003.
8. Gerbauer, P., Thormann, W., and Bocek, P., Sample self-stacking and sample stacking in zone elec-
trophoresis with major sample components of like charge: General model and scheme of possible
modes, Electrophoresis, 16, 2039–2050, 1995.
9. Gerbauer, P., Krivankova, L., Pantuckova, P., Bocek, P., and Thormann, W., Sample self-stacking in
capillary zone electrophoresis: Behavior of samples containing multiple major coionic components,
10. Landers, J.P., Ed., Handbook of Capillary Electrophoresis, 1st ed., Chapter 9, p. 209, CRC Press,
Boca Raton, FL, 1993.
11. Cao, C.X., Zhang, W., Qin, W.H., Li, S., Zhu, W., and Liu, W., Quantitative predictions to conditions
of zwitterionic stacking by transient moving chemical reaction boundary created with weak electrolyte
buffers in capillary electrophoresis, Anal. Chem., 77, 955, 2005.
12. Zhu, J. and Feng, Y.L., A powered online enhancement technique for the analysis of nucleotides using
capillary zone electrophoresis/mass spectrometry, Anal. Chem., 78, 6608, 2006.
13. Gong, M., Wehmeyer, K.R., Limbach, P.A., and Heimeman, W.R., Unlimited-volume electrokinetic
stacking injection in sweeping capillary electrophoresis using a cationic surfactant, Anal. Chem., 78,
6035, 2006.
14. Krivankova, L., Pantuckova, P., Gebauer, P., Bocek, P., Ccaslavska, J., and Thormann, W., Chloride
present in biological samples as a tool for enhancement of sensitivity in capillary zone electrophoretic
analysis of anionic trace analytes, Electrophoresis, 24, 505–517, 2003.
15. Liu, Z., Sam, P., Sirimanne, S.R., McClure, P.C., Grainger, J., and Patterson, D.G., Field-amplified
sample stacking in micellar electrokinetic chromatography for on-column sample concentration of
neutral molecules, J. Chromatogr. A, 673, 125, 1994.
16. Quirino, J.P. and Terabe, S., Exceeding 5000-fold concentration of dilute analytes in micellar
electrokinetic chromatography, Science, 282, 465, 1998.
17. Quirino, J.P. and Terabe, S., Sweeping of analyte zones in electrokinetic chromatography, Anal. Chem.,
71, 1638, 1999.
18. Quirino, J.P., Kim, J.B., and Terabe, S., Sweeping: Concentration mechanism and applications to
high-sensitivity analysis in capillary electrophoresis, J. Chromatogr. A, 965, 357, 2002.
19. Palmer, J., Munro, N.J., and Landers, J.P., A universal concept for stacking neutral analytes in micellar
electrokinetic chromatography, Anal. Chem., 71, 1679, 1999.
20. Palmer, J., Burgi D.S., and Landers, J.P., Electrokinetic stacking injection of neutral analytes under
continuous conductivity conditions, Anal Chem., 74, 632, 2002.
21. Quirino, J.P. and Terabe, S., Online concentration of neutral analytes for micellar electrokinetic chro-
matography. 5. Field-enhanced sample injection with reverse migrating micelles, Anal. Chem., 70,
1893, 1998.
14 Capillary Electrophoresis for
the Analysis of Single Cells:
Sampling, Detection, and
Applications
Imee G. Arcibal, Michael F. Santillo, and Andrew G. Ewing
CONTENTS
14.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429

14.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
14.2.1 Sampling Techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
14.2.1.1 Whole Cell Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
14.2.1.2 Release from Whole Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
14.2.1.3 Subcellular Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432
14.2.2 Detection Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
14.2.2.1 Electrochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
14.2.2.2 Laser-Induced Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 434
14.2.2.3 Mass Spectrometry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 434
14.2.2.4 Radiochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436
14.3 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436
14.3.1 Neuroscience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436
14.3.2 Immunology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437
14.3.3 Nucleic Acids and Gene Expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437
14.3.4 Enzymology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 439
14.4 Future Prospects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 440
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 441
14.1 INTRODUCTION
Conventional analysis of biological specimens typically involves sampling from populations of cells
(e.g., tissue homogenates). The heterogeneity of these populations, however, often yields information
averaged over various types of cells. Single-cell sampling, on the other hand, allows researchers to
study fundamental processes, which may not be observed in a heterogeneous cell population. To look
at these processes at the single-cell level, analytical techniques must have the capability to utilize
small volumes and investigate a variety of compounds concurrently. Obtaining both quantitative and
qualitative information from single cells in complex biological environments allows researchers to
learn more about basic cellular function as well as the mechanisms of drugs and toxins.
Capillary electrophoresis (CE) is an analytical technique ideally suited for studying the small
volume contents of single cells, sampling whole cell and subcellular volumes on the femtoliter scale.
429
In addition, high separation potentials can be applied when using small inner diameter capillaries
with CE since Joule heat can be rapidly dissipated. The use of high potentials results in fast, efficient
separations of over 106 theoretical plates. Furthermore, CE is compatible with a diverse array of
detectors including laser-induced fluorescence (LIF), ultraviolet (UV) absorbance, electrochemi-
cal, mass spectrometric, and radiochemical. This chapter builds upon the previous two editions of
this book discussing advances in sampling techniques and detectors used for analyzing individual
cells with CE. Furthermore, applications of single-cell CE analysis in the areas of neuroscience,
immunology, nucleic acids and gene expression, as well as enzymology are also presented. The use
of microfabricated devices for cellular analysis has become a popular area of research and is covered
elsewhere in this book.
14.2 METHODOLOGY
14.2.1 SAMPLING TECHNIQUES
14.2.1.1 Whole Cell Sampling
Sampling whole cells for CE involves either siphoning the cell by either applying a pressure differ-
ential or applying a potential across the capillary to electrokinetically inject the cell. In each case, a
drag force is produced by the fluid flow, driving the cell into the capillary for lysis and the separation
of its contents.1 Though these schemes are the most simplistic forms of injection, there have been
several recently developed complementary techniques for introduction of cells and their contents
into capillaries.
14.2.1.1.1 Continuous Cell Introduction

Individually selecting cells for siphoning or electrokinetic injection is a time-consuming process,
thereby limiting the number of samples that can be run in a given period of time. In an effort to
increase throughput, Chen and Lillard2 have created a two capillary scheme capable of continuously
introducing cells for injection (Figure 14.1a). The thinner first capillary is immersed in a dilute cell
suspension and cells are continuously pumped in single file by electroosmotic flow. Individual cells
flow toward the lysis junction connecting the two capillaries where they are lysed by mechanical
disruption. The lysate is then siphoned into the second capillary where separation takes place. This
system reduces analysis time from 6 to 4 min for each cell2 and is capable of repeatedly injecting
cells for over 40 min before many of the cells descend to the bottom of solution and out of reach of
the capillary.
14.2.1.1.2 Laser Cavitation and Electrical Lysis

Because removing a cell from its growth substrate alters many of the biological processes occurring in
the cell, its detachment increases the difficulty in analyzing the cell and the variation from its normal
function.3 Thus, in order to obtain the most accurate portrait of the cell’s internal environment, it must
be sampled quickly after removal from the substrate. Two innovative techniques, laser cavitation
and fast electrical lysis, allow for rapid sampling of substrate-bound cells.
Sims et al.4 have exploited the production of a shockwave created by a laser beam for their
laser cavitation scheme (Figure 14.1b). The shockwave generated by laser pulse is focused in close
proximity to the cell to be analyzed, forming plasma at the focal point. A cavitation bubble is
produced that subsequently collapses and causes a shockwave that ruptures the cell’s membrane.
Cellular contents can then be loaded rapidly after lysis, reducing the time available for deviation
from standard cell function to occur.
To circumvent the expense and complication of using a pulsed laser, a fast electrical lysis proce-
dure for single-cell analysis has also been developed (Figure 14.1c).3,4 This setup takes advantage
of gold conductivity, with both the capillary and the coverslip on which cells are grown coated
Capillary Electrophoresis for the Analysis of Single Cells 431
(a) ) (c)
m
5n
( 27
Lysing er
buffer Las
Teflon Ar
Capillary
Cell tubing
Lens Wire to pulse
generator
Capillary A Capillary B Conducting
Lysis
junction Objective epoxy
Insulating
Cell suspension Filters Buffer epoxy Metal-coated
buffer PMT PC capillary and
tapered tip
Metal-coated
+ –
Hig voltage coverslip
power supply
(b)
Side view Bottom view Microscope
Capillary objective
lumen
Metal-coated
Capillary wall Positive
Capillary capillary
voltage
wall Capillary
lumen
Cell
Cell
Metal-coated
Laser ‘divot’ coverslip
Cell Coverslip
Focused Nd:YAG laser
FIGURE 14.1 Recently developed whole cell sampling schemes. (a) Continuous cell introduction. (From
Chen, S. and Lillard, S.J., Anal. Chem., 73, 111, 2001. With permission.) (b) Laser cavitation. (From Sims, C.E.
et al., Anal. Chem., 70, 4570, 1998. With permission.) (c) Fast electrical lysis. (From Han, F., et al., Anal.
with gold prior to analysis. Once the tapered capillary is positioned above a chosen cell, a voltage
pulse is applied between the capillary and the coverslip. The applied pulse passes through the cell
concurrently lysing and injecting the contents into the capillary for separation.
14.2.1.2 Release from Whole Cells

In addition to quantifying the amount of a compound within a whole cell, exocytosis can be monitored
by the sampling, separation, and detection of compounds released following stimulation. Because
this release often results in modifications in the environment surrounding the cell by initiating changes
in adjacent cells, it is important to have the ability to sample releasate from single cells to further
understand the effects of different compounds in a complex cellular environment. Direct monitoring
of exocytosis was initially completed by Chen et al.5 from the giant dopamine cell of the Planorbis
corneus by manipulating the injection end of the separation capillary near the cell. After stimulation,
a plug of secreted compounds was injected into the capillary, which was then returned to the buffer
reservoir for separation. Since then, release from single mammalian cells has been achieved by
Tong and Yeung.6 In their setup, single bovine adrenal medullary cells were injected individually
and allowed to adhere to the capillary wall. Subsequently, a plug of stimulant was injected and the
separation run to quantify the amount of secreted catecholamines.
In order to maintain temporal resolution while obtaining chemical data, Liu et al.7 have used a
dynamic channel electrophoresis scheme to sample release from single cells. As shown in Figure 14.2,
the sampling capillary is scanned across the entrance of a separation channel such that the position of
the capillary at the channel inlet yields the temporal information of release. Single cells are placed in a
“nanoperfusion chamber” for analysis within the sampling capillary (inset of Figure 14.2). Following
Nanoperfusion chamber
PV tubing
Frit Cell
Flow
Channel
Syringe
Gas-tight
box
Reagent vials
Separation channel
Sampling capillary
Electrophoretic axis Time axis
Analyte derivatization zone
FIGURE 14.2 Dynamic channel electrophoresis instrumentation. (From Liu, Y.-M. et al., Anal. Chem., 71,
release of a plug of stimulant, secreted compounds flow through the frit into the separation channel,
minimizing diffusion and increasing time resolution.
14.2.1.3 Subcellular Sampling

Sampling the cellular contents has become more prevalent as capillaries have become smaller and eas-
ier to position. Subcellular sampling allows only the cellular contents of interest to be analyzed, and
offers the possibility to sample specific subcellular compartments, minimizing interference from the
extracellular environment. In this manner, the function or dysfunction of different compartments can
be investigated, simplifying the information being obtained and yielding a more basic understanding
of the mechanisms occurring during fundamental cellular processes.
14.2.1.3.1 Cytoplasmic Sampling

The synthesis, storage, and metabolism of different compounds are regulated by levels found in
cellular cytoplasm. Sampling the cytosol of an intact cell can thus yield about the use of various
chemical species. A microinjector having a tip tapering to 8–10 µm o.d. was previously created
to sample cytoplasm directly from single neurons in P. corneus.8 The size of this injector, though
suitable for large invertebrate cells, is not compatible for use with single mammalian cells, which
average 15 µm in diameter.9 To probe smaller rat adrenal cells, Woods et al.10 have fabricated
microinjectors from 770 nm i.d. capillaries with tips narrowing to 2.5 µm o.d. to use in conjunction
with electroporation. As transient pores are created in the cell by electroporation, the microinjector
can be inserted with minimal perturbation. The voltage pulse used to electroporate is also used to
electrokinetically inject cytoplasm into the capillary before the tip is removed and the separation is
carried out.
14.2.1.3.2 Optical Trapping

The utility of optical trapping for manipulating subcellular contents into a capillary was initially
demonstrated using single vesicles from the Aplysia californica, which were lysed and their contents
separated.11 Optical trapping has since also been employed to maneuver nuclei,12 mitochondria,13,14
and various other organelles15 into capillaries for separation. This noninvasive and sterile manner of
sampling subcellular contents utilizes a single beam laser diode focused through a high numerical
aperture objective onto the desired specimen.11,16 The “scattering force” produced from the scattering
of the light upon encountering the organelle is able to move the organelle along the direction of light
propagation, in this case into the capillary for lysis and separation.17
14.2.1.3.3 Organelle Separations

Much of the study of different organelles by single-cell CE has been completed by the Arriaga
group, particularly focusing on acidic organelles and nuclei. In both cases, the organelles of interest
were labeled with fluorescent tags for ease in detecting the desired species. During the analysis of
single acidic organelles, populations of cells were first incubated with fluorescent microspheres,
which were endocytosed and transported to these organelles. Cells were then lysed in a buffer
solution and the fluorescent “organelle fraction” obtained prior to injection and separation.18 For the
investigation of nuclei, individual nuclei were tagged with a fluorescent nuclear-targeted protein in
order to determine their presence with LIF detection. Analyzed cells were subsequently chosen based
on observed fluorescence with a microscope and injected individually into the capillary. Following
injection, the cell was lysed with digitonin and its cellular components separated.
14.2.2 DETECTION METHODS

CE systems can be coupled to many detectors, allowing a variety of different molecules to be sep-
arated and quantified. UV absorbance, LIF, electrochemistry, radiochemistry, mass spectrometry,
and refractive index are all detectors that have been successfully coupled to CE. However, electro-
chemical and LIF detection are the two most commonly used detectors for single-cell CE due to
their high sensitivity. There have also been reports of using mass spectrometry and radiochemistry
in single-cell CE analyses, though their use is not as widespread.
14.2.2.1 Electrochemistry
Electrochemical detection has found widespread use for single-cell studies with CE owing to both
its selectivity and sensitivity. Though once hampered by the presence of a high voltage electric
field, innovations such as the porous glass coupler for off-column detection and development of an
end-column conical detector have been used to isolate the microelectrode from the potential field
used for separation.19,20 Electrochemical detection is typically carried out in the amperometric mode
with the applied voltage held at a constant overpotential versus a reference electrode. Most often,
carbon fiber working electrodes are used since they can be easily coupled at the ends of capillaries.
As analytes encounter the working electrode surface following separation, they are either oxidized
or reduced to produce a measurable current.
Scanning electrochemical detection has also been employed for single-cell CE separations.21
In contrast to amperometry, scanning electrochemical detection varies the voltage of the working
electrode. As separated species pass the working electrode, a voltammogram is obtained for each
b c
a 1.0
0.8
0.6
0.4 Current (nA)
0.2
0.0
0.8 –0.2
0.6
E (volts) 0.4 10
7
0.2
4 Time (min)
–0.2 0.0 0.2 0.4 0.6 0.8 1.0

Current (nA)
FIGURE 14.3 Electropherogram of the separated contents from a single dopaminergic neuron obtained using
CE with scanning electrochemical detection and plotted in three dimensions. Peaks “a” and “b” correspond to
dopamine that was immediately released during lysis and dopamine located deeper in the cell, respectively.
Peak “c” represents a neutral electroactive species in cellular buffer. (From Swanek, F.D. et al., Anal. Chem.,
analyte and a three-dimensional surface is constructed showing current versus voltage and time
(Figure 14.3). This unique detection mode allows eluted substances to be identified according to their
characteristic voltammogram as well as retention time. Thus, scanning electrochemical detection
offers enhanced selectivity since co-eluting peaks can be further resolved by their unique oxidation
potentials. Scanning electrochemical detection has been utilized to detect dopamine at the femtomole
level from single P. corneus neurons.22
14.2.2.2 Laser-Induced Fluorescence

Along with amperometry, LIF is the most popular mode of detection in single-cell CE experiments
owing to its high sensitivity. Most CE-LIF systems employ visible argon ion lasers; however, UV
lasers have also been used for fluorescence exitation.23,24 One problem with LIF is that not all
species are natively fluorescent, so labeling is often required. Although LIF detection most often
involves a single emission wavelength, Sweedler has demonstrated simultaneous multiwavelength
detection (Figure 14.4).25 Over 30 compounds containing aromatic amino acids in an individual
neuron were separated and detected using native LIF with a UV laser, and a unique fluorescence
emission spectrum was obtained for each eluted species.
14.2.2.3 Mass Spectrometry

Mass spectrometry (MS) has perhaps the most potential as a detector since it can provide detailed
chemical information to identify separated species, particularly molecules with high molecular
weights like polypeptides. Unfortunately, mass spectrometers have not been widely used in single-
cell CE experiments as detectors due to their relatively poor sensitivity compared to LIF. Nonetheless,
2
3
FIGURE 14.4 (See color insert following page 810.) CE system with simultaneous multiwavelength fluo-
rescence detection. Cellular samples were electrokinetically drawn into a capillary (1) from a steel microvial
(2). The analytes were separated in the capillary, and eluted into a square-cuvette sheath-flow cell (3) where
fluorescence was induced with a laser beam (4). The fluorescent light was collected orthogonally and directed
into a spectrograph and CCD where a complete spectrum was obtained for each separated analyte (5). (From
Fuller, R.R. et al., Neuron, 20, 173, 1998. With permission.)
Separation capillary Syringe pump

(with MALDI matrix solution)
Grounding
CE inlet Silver epoxy
wire
apparatus (CE cathode)
Sample target
High-voltage
power supply
FIGURE 14.5 CE coupled to a MALDI-TOF-MS system for detecting neuropeptides. As the contents of a
single cell were separated in the capillary, the separated species were collected as fractions deposited onto a
MALDI-MS target. (From Page, J.S. et al., Analyst, 125, 555, 2000. With permission.)
there have been a few interesting papers using mass spectrometric detection for single-cell analysis.
One early application of CE-MS for single-cell analysis demonstrated the use of electrospray ioniza-
tion and Fourier transform ion cyclotron resonance mass spectrometry to detect the α and β chains
of hemoglobin in a single human erythrocyte.26
Sweedler and coworkers27 have developed a novel system coupling CE to matrix-assisted
laser desorption–ionization time-of-flight mass spectrometry (MALDI-TOF-MS) to study peptides
(Figure 14.5). In some initial work, the contents of a single cerebral ganglion cell from A. cali-
fornica were separated and mixed with MALDI matrix after exiting the capillary; every 30 s, a
new fraction was deposited onto the MALDI target plate for mass spectrometric analysis. In fol-
lowing work, using an instrumentation setup similar to Figure 14.5, MALDI-TOF-MS was coupled
with CE to identify neuropeptides and hormones released from individual A. californica cells.28
Although MALDI-TOF-MS is excellent for qualitative chemical identification of high molecular
weight species, it lacks the ability for quantification. Sweedler and coworkers,29 however, comple-
mented CE-MALDI-TOF-MS with radiochemical detection to achieve both chemical identification
and successful quantification of hormones and neuropeptides from individual cells.
14.2.2.4 Radiochemistry
Radiochemistry has not been extensively used for single-cell CE systems because of the cumbersome
nature of working with radioactive elements and lengthy time involved in collecting an image of radi-
olabeled eluents. A postcolumn radionucleotide detector has been developed in which radiolabeled
eluents are deposited onto a membrane with a scintillator, and then imaged with a charge-coupled
device (CCD). This unique system had limits of detection of 88 zmol for 32 P-labeled analytes, 17 amol
for 35 S, and 8 fmol for 3 H, which were found in peptides.30 An improvement of the previous system
was made by using a commercial phosphor plate and photomultiplier tube, which yielded even lower
limits of detection (4.9 zmol for 32 P and 0.13 amol for 35 S found in peptides), and was successfully
applied to detecting radiolabeled peptides single buccal ganglion cells in A. californica.31
14.3 APPLICATIONS
14.3.1 NEUROSCIENCE
Single-cell CE experiments in neurochemistry were first performed in the late 1980s with initial
cytoplasmic sampling from P. corneus completed in 198832 and whole cell injection of Helix aspersia
in 1989 to separate fluorescently labeled amino acids.33 Since these pioneering studies, many different
small molecule neurotransmitters and other related species have been detected in single cells by CE.
Serotonin is a monoamine neurotransmitter believed to regulate physiological functions such
as appetite and sleep. Most importantly, decreased serotonin transmission in the brain has been
implicated in playing a role in depression and other related mood disorders, making it an important
focus of research in the contemporary neuroscience community. In one study, individual Aplysia
metacerebral cells were electrically stimulated, followed by separation by CE and fluorescence
detection with a UV laser.23 Each cell soma contained approximately 450 fmol of serotonin with a
limit of detection for the experiment of 39 nM. A later report34 investigated serotonin metabolism
in Aplysia. Single cells from different locations of the Aplysian nervous system were incubated with
serotonin with two new serotonin catabolites discovered. Interestingly, it was recently shown that
electrokinetic injection of serotonin results in the formation of a serotonin dimer artifact, but when
hydrodynamic injection is used, the dimer is not present.35
Like serotonin, the amino acid d-aspartate is a small molecule neurotransmitter and is present in
the central nervous system of both invertebrates and vertebrates. While most proteins and polypep-
tides contain only amino acids of the l-enantiomer, mounting evidence suggests that the d-enantiomer
of aspartate plays a role in intercellular signaling. The addition of cyclodextrin to the separation
buffer in CE allows these two enantiomers to be separated and individually quantified.36–38 When
individual Aplysian cell processes and soma were analyzed separately by CE-LIF, it was shown
that d-aspartate was not only located in the nucleus but also in the processes and the amount of
d-aspartate in different processes of the same cell was similar.37 In a subsequent experiment, CE-
LIF was used to show that the amounts of d-aspartate varied among different clusters from single
A. californica neurons, d-aspartate is synthesized endogenously from its l-enantiomer, and it travels
long intracellular distances to its release sites.38 In addition to aspartate, the addition of cyclodextrin
has allowed fluorescently labeled d-glutamate and l-glutamate to be separated from an individual
cell of Aplysia.39 Neuropeptides may also contain d-amino acids with the importance of peptides
containing d-amino acids illustrated with the discovery of a neuropeptide containing d-tryptophan
in a single cell of A. californica.40
Nitric oxide (NO), a more recently discovered neurotransmitter and regulator of other physio-
logical processes, has been indirectly quantified by CE of its primary metabolites. Because of the
ephemeral nature of NO, nitrate and nitrite, two products of NO oxidation, have been assumed to
be good indicators of NO synthase activity. These anions in single Pleurobranchaea californica and
A. californica neurons have been separated by CE and quantified with UV absorbance detection.41
The limits of detection were less then 200 fmol and the concentrations of nitrite and nitrate were
typically 2 and 12 mM, respectively. A more recent report involved CE separations of arginine,
citrullene, nitrate, and nitrite with LIF and conductivity detection.42 Contrary to previous results, it
was determined that nitrates are not always reliable indicators of NO synthase activity. Similarly,
other metabolites associated with NO—arginine, citrullene, arginosuccinate, ornithine, and arginine
phosphate—were quantified in Pleurobranchaea and Aplysia with limits of detection ranging from
5 nM to 17 µM.43
In vivo detection of NO is commonly accomplished using the fluorophore 4,5-diaminofluorescien
(DAF-2). When the fluorophore reacts with N2 O3 (a species related to NO), it is converted to a
triazole derivative, DAF-2T. Dehydroascorbic acid present in cells also reacts with DAF-2 and
forms a complex with an emission spectrum similar to that of DAF-2T, thereby interfering with NO
detection. The separation of the interfering species with CE-LIF has allowed quantification of NO
in single cells,44 along with ascorbic acid and dehydroascorbic acid,44,45 which both play important
roles in cellular metabolism.
14.3.2 IMMUNOLOGY
For years, enzyme-linked immunosorbent assays (ELISA) have been the standard method for inves-
tigating antibody–antigen interactions. The time-consuming nature of ELISA, however, has led to
the development of single-cell CE for the rapid analysis of the reaction of antibodies and target anti-
gens. One such study has utilized single-cell CE-LIF to analyze human interferon-gamma (IFN-γ ),
which provides antiviral, anticell proliferation, and immunity adjustment effects as support for other
cells.46 Though initially employing on-capillary immunoreactions, where the contents of a lysed nat-
ural killer (NK) cell could interact with fluorescently tagged anti-IFN-γ monoclonal antibody (Ab*)
prior to the separation,46 Zhang and Jin47 further optimized sample preparation by loading Ab* into
cells via electroporation prior to entering the capillary. Single cells were subsequently washed and
lysed on-capillary following a short incubation time prior to analysis by CE-LIF. Labeled antibodies
for both IFN-γ I and II were detected in the cell electropherograms with increases in the relative flu-
orescence intensity observed for each form of the protein–antibody complex with the electroporated
loaded cells versus those that had undergone the on-capillary reaction due to a decrease in Ab* dilu-
tion during the intracellular immunoreaction (Figure 14.6). In addition, the average value determined
for total IFN-γ concentration in NK cells (98 pM) corresponded well with the value acquired from an
ELISA assay (99 pM),47 demonstrating the utility of the CE technique in performing immunoreaction
assays rapidly while maintaining accuracy.
In another investigation, the role of P-glycoprotein (PGP) in imparting multidrug resistance to
tumor cells by functioning as transmembrane drug pump was examined.48 Two groups of single
cells—those that were drug resistant and those drug sensitive—were incubated with primary (JSB-1,
a mouse raised antibody that binds to PGP) and secondary [goat antimouse IgG-fluorescein isothio-
cyanate (GAMIF), an antimouse antibody tagged with fluorescein isothiocyanate (FITC)] antibodies
prior to analysis with CE-LIF. Owing to the minimal nonspecific binding of JSB-1, its quantity in a
single-cell was taken as the total amount of PGP present on the cell since the antibody reacts with
the protein epitope on a 1:1 ratio. When comparing the amount of PGP expressed in each cell group,
a larger quantity was present on the drug resistant cell line, with 157,000 molecules of PGP present
versus 47,000 molecules present on the drug sensitive cell line.
14.3.3 NUCLEIC ACIDS AND GENE EXPRESSION

A popular technique used to observe gene expression is reverse transcriptase-polymerase chain reac-
tion (RT–PCR). RT–PCR works by reverse transcribing mRNA to a complementary DNA sequence
that corresponds to the gene of interest. Several years ago, Zabzdyr and Lillard49 showed that
CE could be used to separate and detect RT-PCR products corresponding to β-actin in individual
(a) 0.02
IFN-γ I-Ab*
Ab*
IFN-γ II-Ab*
RFU
0.01
2 3 4 5 6
Time (min)
(b) 0.02 IFN-γ I-Ab*

IFN-γ II-Ab*
0.01
RFU
Ab*
0
2 3 4 5 6
Time (min)
FIGURE 14.6 Electropherograms detailing IFN-γ −antibody complex (IFN-γ -Ab*) content in NK cells using
on-capillary immunoreaction (a) and intracellular immunoreaction (b) for both IFN-γ I and II. (From Zhang H.
and Jin, W., J. Chromatogr. A, 1104, 346, 2006. With permission.)
human prostate carcinoma cells, which was similar to an earlier study done by Li and Yeung50
who detected PCR products corresponding to β-actin from lymphoblast cells. The cells were lysed
with detergent or by a freeze-thaw method and a primer and other PCR reagents were added. The
resulting PCR products were complexed with ethidium bromide and were separated by CE-LIF.
The freeze-thaw lysing method yielded a better signal-to-noise ratio versus the use of detergent,
and limits of detection corresponded to 133 initial molecules of β-actin mRNA per nL. Yeung
and coworkers51 successfully developed a microthermocycler for integrated cell lysis and RT–PCR
in which the PCR products of β-actin were measured from a group of 16 cells, demonstrating
the possibility of using this novel system for single-cell analyses. Lillard and coworkers52 more
recently used CE-LIF to detect RT–PCR products corresponding to β-actin and the α estrogen recep-
tor simultaneously from single human breast carcinoma cells (Figure 14.7). Surprisingly, only a
single round of RT–PCR amplification was necessary, but additives were needed to enhance the
low signals for the analytes. It is hypothesized that the signals were low because two primers were
used simultaneously, and it is thought that they formed dimers thereby lowering the amplification
efficiency.
In addition to using CE for detecting single-cell RT–PCR products, CE has also been used to
detect RNA in individual cells without the need for extraction. Lillard and coworkers53 injected sin-
gle Chinese hamster ovary cells into a capillary, where the cells were lysed and the RNA complexed
with ethidium bromide for LIF detection. Only rRNA and tRNA were detected, which was verified
by monitoring changes in the peak patterns after adding RNase. mRNA was not detected because
it is more unstable than the other forms of RNA with the buffer conditions used in these experi-
ments. The total amount of RNA in individual cells was calculated to be approximately 10–20 pg.53
Since aging, carcinogens, and other toxins cause damage to nucleic acids, single-cell CE was also
used to observe changes in peak patterns of nucleic acids after single cells were incubated with
hydrogen peroxide.53 In addition to monitoring changes in nucleic acids following chemical dam-
age, single-cell CE-LIF has also been used to look at changes in RNA during cell division.54 Single
Chinese hamster ovary cells were taken from different phases of the cell cycle, injected individually
1000000
900000 β-actin
800000
Relative fluorescence
700000
600000
500000
ERα
400000
300000
200000
100000
0
12 13 14 15 16 17 18 19 20 21
FIGURE 14.7 Electropherogram of multiplex RT–PCR products corresponding to β-actin and the α estrogen
receptor from single human breast carcinoma cells. (From Zabzdyr, J.L. and Lillard, S.J., Electrophoresis, 26,
via suction into a capillary, followed by lysis and detection with LIF. The total RNA content
increased after each phase in the cycle, while the rates of RNA synthesis varied between the steps in
the cycle.54
14.3.4 ENZYMOLOGY
Understanding the mechanism behind the interaction between enzymes and substrates in a biological
system is inherently difficult because of the variation that exists in enzyme expression and function
across a population of cells. The use of single-cell CE, however, is particularly suited to enzymology
as the activity of individual cells can be assayed alone. In one study, for example, post-translational
enzymes have been investigated at potential targets for cancer therapy. The role of Ras proteins in cell
growth and differentiation, as well as the presence of their mutated forms in human cancers, has made
them a particularly interesting area of study for understanding how cancer proliferates and spreads. In
efforts to target possible mechanisms for thwarting the function of mutated Ras proteins, the activity
of three enzymes required to create functional proteins—farnesyltransferase, endoprotease, and
methyltransferase—was evaluated.55 Immortalized mouse cell lines were incubated with a peptide
analog of the Ras protein to evaluate the function of each enzyme. Three enzyme products, two
found in extracellular media and one intracellularly, were detected using CE-LIF upon lysis of single
incubated cells. Disappointingly, comparison of the experimental product migration times did not
correspond with those of the known products of the enzymes with the substrate, so no conclusion
could be made as to the identity of the experimental compounds. However, the protocol developed
was clearly shown to be compatible with single-cell CE.
While simply investigating the action of an enzyme in a cell at a single time point can provide
a great deal of information, the ability to monitor how a reaction proceeds over time would also be
advantageous. Shoemaker et al.56 have developed nanoliter reaction vessels to sample the contents
of a single lysed cell over time. Intracellular enzymes, such as α-galactosidase, from the lysed cell
are capable of interacting with a particular substrate in the vessel. In the case of α-galactosidase, a
disaccharide substrate, αGlc(1 → 3)αGlc-TMR, was utilized as the substrate and nanoliter aliquots
were sampled for analysis by CE-LIF over several hours to detect the presence of the enzymatic
product (Figure 14.8). A heterogeneity in enzyme function was observed with average conversion
1.0 DGTMR
(a)
Signal (V)
0.6
0.2 MGTMR
8.5 9.0 9.5 10.0

45 Migration time (min)
(b)
40
35
Conversion (%)
30
25
20
15
10
5
0
Co trol 1
Ce l 2
Cell 1
Cell 2
Cell 3
Cell 4
Cell 5
Cell 6
Cell 7
Cell 8
Ce ll 9
Cell 10
Cell 11
Cell 12
Cell 13
Cell 14
Cell 15
Cell 16
Cell 17
Cell 18
Cell 19
0
ll 2
ntro
n
Co
Sample
FIGURE 14.8 Electropherogram of an α-glucosidase II single-cell (Cell 7) reaction aliquot taken after
incubation (a). Percent conversion of the dissaccharide (DG-TMR) substrate by α-glucosidase II in single
cells and control assays (b). Percent conversion was calculated by comparing the DG-TMR peak to that of
the monosaccharide product MG-TMR. (From Shoemaker, G.K. et al., Anal. Chem., 77, 3132, 2005. With
permission.)
of 20.9% ± 9.4% of substrate to product and is attributable to differences in cell age, cell cycle,
and cell size.55,56
14.4 FUTURE PROSPECTS

Capillary electrophoresis of single cells has been employed in many scientific areas such as neu-
roscience, immunology, gene expression, and enzymology in the past 20 years. Though numerous
developments have been made, many avenues are still left to pursue. One area that can be devel-
oped further is sample acquisition. The Chiu group, for instance, utilizes nanosurgery to choose
which subcellular component (organelle) to analyze, using a combination of optical trapping and
electroporation15,57 to remove it from the cell. Other organelle types can be studied by pairing this
technique with CE separations to better elucidate their functions. In addition, although continuous
sampling from a single cell has already been achieved to study enzyme activity, it would be bene-
ficial to maintain an intact cell while sampling to gain a more accurate picture of what is occurring
in terms of cellular function. Further, sampling single cells within a population simultaneously will
give clues about intercellular interactions in tissues and organs. In this manner, one can determine
how a stimulus to one affects the others.
As always, separation conditions are constantly being optimized to gain the best resolution
between analytes. For instance, when analyzing single whole cells, both lipids and proteins are
capable of adsorbing to capillary walls, reducing peak resolution, and separation efficiency. Thus, it
is imperative to reduce adsorption to the capillary walls to maintain the advantages of using CE over
other techniques. Further, to lower detection limits, more sensitive detectors must be developed.
Also, since LIF and electrochemistry are most common detection methods, but most analytes are
not natively fluorescent or electroactive, derivatization of analytes must be optimized so that all
compounds present can be adequately measured.
While new innovations in sampling schemes and detectors will further increase its use for the
analysis of single cells, CE has shown great promise as a tool to study single cells. The key advantages
of CE are small volume capability combined with rapid and efficient separations. In addition, the
diversity of detectors that can be coupled to a CE system further increases its versatility as a variety of
analytes can be investigated. With continued use of CE to study individual cells, more information
can be learned about the basic function of cells as well as their interactions with each other and
different chemical species.
REFERENCES
1. Krylov, S.N. et al., Instrumentation for chemical cytometry, Anal. Chem., 72, 872, 2000.
2. Chen, S. and Lillard, S.J., Continuous cell introduction for the analysis of individual cells by capillary
3. Han, F. et al., Fast electrical lysis of cells for capillary electrophoresis, Anal. Chem., 75, 3688, 2003.
4. Sims, C.E. et al., Laser-micropipet combination for single-cell analysis, Anal. Chem., 70, 4570, 1998.
5. Chen, G. et al., Observation and quantitation of exocytosis from the cell body of a fully developed
neuron in Planorbis corneus, J. Neurosci., 15, 7747, 1995.
6. Tong, W. and Yeung, E.S., On-column monitoring of secretion of catecholamines from single bovine
adrenal chromaffin cells by capillary electrophoresis, J. Neurosci. Methods, 76, 193, 1997.
7. Liu, Y.M., Moroz, T., and Sweedler, J.V., Monitoring cellular release with dynamic channel
8. Olefirowicz, T.M. and Ewing, A.G., Capillary electrophoresis in 2 and 5 microns diameter capillaries:
Application to cytoplasmic analysis, Anal. Chem., 62, 1872, 1990.
9. Lu, X. et al., Recent developments in single-cell analysis, Anal. Chim. Acta, 510, 127, 2004.
10. Woods, L.A., Gandhi, P.U., and Ewing, A.G., Electrically assisted sampling across membranes with
electrophoresis in nanometer inner diameter capillaries, Anal. Chem., 77, 1819, 2005.
11. Chiu, D.T. et al., Probing single secretory vesicles with capillary electrophoresis, Science, 279,
1190, 1998.
12. Liang, H. et al., Giant cell formation in cells exposed to 740 nm and 760 nm optical traps, Lasers Surg.
Med., 21, 159, 1997.
13. Fuller, K.M. and Arriaga, E.A., Advances in the analysis of single mitochondria, Curr. Opin.
Biotechnol., 14, 35, 2003.
14. He, M. et al., Selective encapsulation of single cells and subcellular organelles into picoliter- and
femtoliter-volume droplets, Anal. Chem., 77, 1539, 2005.
15. Shelby, J.P., Edgar, J.S., and Chiu, D.T., Monitoring cell survival after extraction of a single subcellular
organelle using optical trapping and pulsed-nitrogen laser ablation, Photochem. Photobiol., 994, 2005.
16. Kuyper, C.L. and Chiu, D.T., Optical trapping: a versatile technique for biomanipulation, Appl.
Spectrosc., 56, 300A, 2002.
17. Block, S.M., Making light work with optical tweezers, Nature, 360, 493, 1992.
18. Fuller, K.M. and Arriaga, E.A., Analysis of individual acidic organelles by capillary electrophore-
sis with laser-induced fluorescence detection facilitated by the endocytosis of fluorescently labeled
microspheres, Anal. Chem., 75, 2123, 2003.
19. Wallingford, R.A. and Ewing, A.G., Capillary zone electrophoresis with electrochemical detection,
Anal. Chem., 59, 1762, 1987.
20. Sloss, S. and Ewing, A.G., Improved method for end-column amperometric detection for capillary
21. Ferris, S.S., Lou, G., and Ewing, A.G., Scanning electrochemical detection in capillary electrophoresis,
J. Microcol. Sep., 6, 263, 1994.
22. Swanek, F.D., Chen, G.Y., and Ewing, A.G., Identification of multiple compartments of dopamine in
a single cell by CE with scanning electrochemical detection, Anal. Chem., 68, 3912, 1996.
23. Miao, H., Rubakhin, S.S., and Sweedler, J.V., Analysis of serotonin release from single neuron soma
using capillary electrophoresis and laser-induced fluorescence with a pulsed deep-UV NeCu laser,
Anal. Bioanal. Chem., 377, 1007, 2003.
24. Zabzdyr, J.L. and Lillard, S.J., UV- and visible-excited fluorescence of nucleic acids separated by
25. Fuller, R.R. et al., Single neuron analysis by capillary electrophoresis with fluorescence spectroscopy,
Neuron, 20, 173, 1998.
26. Hofstadler, S.A. et al., Analysis of single cells with capillary electrophoresis electrospray ionization
Fourier transform ion cyclotron resonance mass spectrometry, Rapid Commun. Mass Spectrom., 10,
919, 1996.
27. Page, J.S., Rubakhin, S.S., and Sweedler, J.V., Direct cellular assays using off-line capillary elec-
trophoresis with matrix-assisted laser desorption/ionization time-of-flight mass spectrometry, Analyst,
125, 555, 2000.
28. Rubakhin, S.S. et al., Analysis of cellular release using capillary electrophoresis and matrix assisted
laser desorption/ionization-time of flight-mass spectrometry, Electrophoresis, 22, 3752, 2001.
29. Page, J.S., Rubakhin, S.S., and Sweedler, J.V., Single-neuron analysis using CE combined with
MALDI MS and radionuclide detection, Anal. Chem., 74, 497, 2002.
30. Tracht, S., Toma, V., and Sweedler, J.V., Postcolumn radionuclide detection of low-energy beta emitters
in capillary electrophoresis, Anal. Chem., 66, 2382, 1994.
31. Tracht, S.E. et al., Detection of radionuclides in capillary electrophoresis using a phosphor-imaging
detector, Anal. Chem., 68, 3922, 1996.
32. Wallingford, R.A. and Ewing, A.G., Capillary zone electrophoresis with electrochemical detection in
12.7 microns diameter columns, Anal. Chem., 60, 1972, 1988.
33. Kennedy, R.T. et al., Microcolumn separations and the analysis of single cells, Science, 246, 57, 1989.
34. Stuart, J.N. et al., Serotonin catabolism depends upon location of release: Characterization of sulfated
and gamma-glutamylated serotonin metabolites in Aplysia californica, J. Neurochem., 84, 1358, 2003.
35. Stuart, J.N. et al., Spurious serotonin dimer formation using electrokinetic injection in capillary
electrophoresis from small volume biological samples, Analyst, 130, 147, 2005.
36. Miao, H., Rubakhin, S.S., and Sweedler, J.V., Confirmation of peak assignments in capillary
electrophoresis using immunoprecipitation. Application to d-aspartate measurements in neurons,
J. Chromatogr. A, 1106, 56, 2006.
37. Miao, H., Rubakhin, S.S., and Sweedler, J.V., Subcellular analysis of d-aspartate, Anal. Chem., 77,
7190, 2005.
38. Miao, H. et al., d-Aspartate as a putative cell-cell signaling molecule in the Aplysia californica central
nervous system, J. Neurochem., 97, 595, 2006.
39. Quan, Z. and Liu, Y.M., Capillary electrophoretic separation of glutamate enantiomers in neural
samples, Electrophoresis, 24, 1092, 2003.
40. Sheeley, S.A. et al., Measuring d-amino acid-containing neuropeptides with capillary electrophoresis,
Analyst, 130, 1198, 2005.
41. Cruz, L. et al., Nitrite and nitrate levels in individual molluscan neurons: Single-cell capillary
electrophoresis analysis, J. Neurochem., 69, 110, 1997.
42. Moroz, L.L. et al., Direct single cell determination of nitric oxide synthase related metabolites in
identified nitrergic neurons, J. Inorg. Biochem., 99, 929, 2005.
43. Floyd, P.D. et al., Capillary electrophoresis analysis of nitric oxide synthase related metabolites in
single identified neurons, Anal. Chem., 70, 2243, 1998.
44. Kim, W.S. et al., Measuring nitric oxide in single neurons by capillary electrophoresis with laser-
induced fluorescence: Use of ascorbate oxidase in diaminofluorescein measurements, Anal. Chem.,
78, 1859, 2006.
45. Kim, W.S. et al., Ascorbic acid assays of individual neurons and neuronal tissues using capillary
electrophoresis with laser-induced fluorescence detection, Anal. Chem., 74, 5614, 2002.
46. Zhang, H. and Jin, W., Determination of different forms of human interferon-gamma in single nat-
ural killer cells by capillary electrophoresis with on-capillary immunoreaction and laser-induced
fluorescence detection, Electrophoresis, 25, 1090, 2004.
47. Zhang, H. and Jin, W., Single-cell analysis by intracellular immuno-reaction and capillary electrophore-
sis with laser-induced fluorescence detection, J. Chromatogr. A, 1104, 346, 2006.
48. Xiao, H. et al., Immunoassay of P-glycoprotein on single cell by capillary electrophoresis with laser
induced fluorescence detection, Anal. Chim. Acta, 556, 340, 2006.
49. Zabzdyr, J.L. and Lillard, S.J., Measurement of single-cell gene expression using capillary
50. Li, H. and Yeung, E.S., Selective genotyping of individual cells by capillary polymerase chain reaction,
51. Matsunaga, H., Anazawa, T., and Yeung, E.S., Integrated on-capillary instrumentation for gene
expression measurement directly from cells, Electrophoresis, 24, 458, 2003.
52. Zabzdyr, J.L. and Lillard, S.J., A qualitative look at multiplex gene expression of single cells using
53. Han, F. and Lillard, S.J., In-situ sampling and separation of RNA from individual mammalian cells,
Anal. Chem., 72, 4073, 2000.
54. Han, F. and Lillard, S.J., Monitoring differential synthesis of RNA in individual cells by capillary
electrophoresis, Anal. Biochem., 302, 136, 2002.
55. Arkhipov, S.N. et al., Chemical cytometry for monitoring metabolism of a Ras-mimicking substrate
in single cells, Cytometry A, 63, 41, 2005.
56. Shoemaker, G.K. et al., Multiple sampling in single-cell enzyme assays using CE-laser-induced
fluorescence to monitor reaction progress, Anal. Chem., 77, 3132, 2005.
57. Chiu, D.T., Micro- and nano-scale chemical analysis of individual sub-cellular compartments, Trends
Anal. Chem., 22, 528, 2003.
15 Ultrafast Electrophoretic
Separations
Michael G. Roper, Christelle Guillo, and B. Jill Venton
CONTENTS
15.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445

15.2 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 446
15.3 Methods Development. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
15.3.1 Applied Voltage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
15.3.2 Injection Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
15.3.4 Electronics and Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453
15.4.1 Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454
15.4.2 Microfluidic Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 457
15.5.2 Microfluidic Chips. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461
15.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462
15.1 INTRODUCTION
Many chemical and biological events are dynamic in nature and rapid analytical techniques are
required to monitor their occurrence. For example, reaction intermediates may be unstable and only
present for milliseconds. Similarly, receptor binding and affinity complex formation or dissociation
occurs on a rapid time scale. Neurotransmission and other cellular signaling processes are fast, with
rapid release and clearance of signaling compounds. In addition, high-throughput analyses such as
DNA analysis or drug screening require repetitive and rapid observations. Fast analytical techniques
are therefore needed to better understand the time course of these phenomena.
Separations have often been considered a slow step in analysis, unsuitable for monitoring rapid
processes. The first demonstration of an electrophoretic separation was achieved in hours1 and today,
conventional capillary electrophoresis separations are typically completed in around 10 min, depend-
ing on the sample. Other analytical methods, such as spectroscopy or electrochemistry, have therefore
been used to obtain dynamic information about chemical changes. However, separations provide
chemical information that may not be obtained with other techniques. Fast separations that allow
simultaneous collection of temporal and chemical information would be advantageous for monitoring
many of the real-world examples given above. Electrophoretic separations performed in capillaries
or planar substrates have advantages over high-performance liquid chromatography (HPLC) and gel
electrophoresis because rapid separations can be achieved without a loss of separation efficiency.
Advances in electrophoresis instrumentation have reduced separation times from the minute to the
445
second time scale, and even to microseconds in some applications. The improved temporal resolution
is leading to a better understanding of rapid chemical and biological processes. In this chapter, we
will outline the theory, technological advances, and practical aspects of performing ultrafast capillary
and microfluidic electrophoresis.
15.2 THEORY
When examining the theoretical basis of rapid separations, it is important to balance the consider-
ations of temporal resolution with well-resolved separations. The fundamental equations defining
electrophoresis show that it is possible to increase both the speed and efficiency of a separation by
increasing the voltage.2,3 The migration time (tmig ) is given by
L2
tmig = (15.1)
V × µep
where L is the separation length (and length to the detector), V is the separation voltage, and µep is
the electrophoretic mobility of the analyte. Because the migration time is inversely proportional to
voltage, higher voltages lead to shorter migration times. The equation for theoretical plates (N) is
L2 µep × V
N= = (15.2)
σtot
2 2D
where D is the diffusion coefficient and σtot 2 is the total zone variance due to band broadening.
If band broadening is limited to diffusion, then the equation simplifies to the term on the right,
which shows that N is proportional to voltage. This reveals that more efficient separations will result
from high applied voltages. Shorter separation lengths will result in faster separations, given the L 2
proportionality to migration time, but with no loss in separation efficiency since N is not dependent on
separation length. In contrast, in chromatography, shorter column lengths lead to faster separations
but also result in a lower number of theoretical plates. Therefore, the strategy for fast, efficient
electrophoretic separations is clearly high voltages and short separation distances.
The above analysis assumed that the only source of band broadening was diffusion, but other
sources of band broadening can occur in an experiment. The total variance is given by
σtot
2
= σdiff
2
+ σheat
2
+ σinj
2
+ σdet
2
+ σads
2
+ σED
2
(15.3)
where variances are due to diffusion, joule heating, injection, detection, adsorption, and electro-
migration dispersion, respectively. Joule heating increases with high voltages, which is problematic
for fast separations where electric field strengths are often greater than 1000 V cm−1 . Heat is
only dissipated through the surface of the separation path and radial temperature gradients between
the inside and outside of the separation path result. The temperature gradients can lead to band
broadening due to convection or differences in temperature-dependent variables that control the
migration rate, such as viscosity. The effect of Joule heating on plate height has been quantified
and the parameters can be predicted where heating will have a deleterious effect on efficiency.4 The
temperature difference between the wall and the middle of the separation path is proportional to
the radius (r) of the capillary inner diameter (i.d.) to the second power (i.e., r 2 ), so to reduce the
effects of Joule heating most fast electrophoretic studies employ small i.d. capillaries (<10 µm i.d.
compared to the standard 50–100 µm i.d. capillaries) or shallow microfluidic channels. Joule heating
can also be reduced by using low conductivity buffers.5
Small i.d. capillaries (or shallow microfluidic channels) present experimental challenges,
particularly for injection and detection techniques. Band broadening due to adsorption is also a
Ultrafast Electrophoretic Separations 447
prominent mode of dispersion in small i.d. capillaries due to high surface area to volume ratios, but
covalent or dynamic coatings can be used to reduce this effect.6,7 When small i.d. capillaries are
used, lower amounts of analyte must be injected to avoid band broadening due to sample overload-
ing. Therefore, injection instrumentation must reproducibly inject small plugs of sample; however, if
smaller quantities are being injected, then high-sensitivity detectors are needed to detect the analytes.
Typically, ultraviolet (UV) detectors are not sensitive enough, and so most fast separations rely on
high-sensitivity detection methods such as laser-induced fluorescence (LIF). Advances in injection
and detection methodologies for ultrafast separations will be described more in the Section 15.3.
Another parameter to consider in electrophoresis is electroosmotic flow (EOF), which is present
in most examples of rapid electrophoretic separations. Therefore, µep in Equations 15.1 and 15.2
must be replaced with the sum of electrophoretic mobility and the electroosmotic mobility (µEOF ):
L2
tmig = (15.4)
V × µep + µEOF

L2 µep + µEOF × V
N= 2 = . (15.5)
σtot 2D
At first glance, EOF appears to be favorable for both migration time and efficiency. However,
while EOF mobilizes the analytes off the column faster, allowing less time for diffusional band
broadening and an improvement in N, it is a nonselective mode of transport. As a result, if EOF is
in the same direction as the ion mobility, two analytes have less time to be separated and a loss of
resolution results. This fact is shown in the equation for resolution (Rs ) of two compounds:

V
Rs = 0.177 µep (15.6)
D (µ̄ + µEOF )
where µep is the difference in the electrophoretic mobilities of two analytes and µ̄ is the average
electrophoretic mobility of two analytes. Thus, the highest resolution is attained when µ̄ is approxi-
mately equal and opposite to µEOF ; however, this situation is at the cost of increased analysis time
(see Equation 15.4).2 Methods to reduce EOF, such as wall coatings and buffer additives, promote
better resolution while maintaining fast separations.8
15.3 METHODS DEVELOPMENT

Examining the theory for electrophoresis reveals that high voltages and small separation lengths are
two strategies primarily involved in ultrafast separations, for both capillary and microfluidic formats.
However, most commercial instruments are incompatible with short separation distances and are
not equipped with the fast injection devices and sensitive detectors needed for rapid separations.
Therefore, methods development for both ultrafast capillary and microfluidic separations has gone
hand-in-hand with instrument development. Instrument development has focused on methods to
maximize voltages, minimize injection volumes, detect small amounts of analyte, and automate
data analysis.
15.3.1 APPLIED VOLTAGE

Typical separations might employ electric fields from 100 to 300 V cm−1 while electric fields in the
1000–5000 V cm−1 range are common for fast separations. Ultrafast separations on the millisecond
or microsecond timescale can require fields up to MV cm−1 . These high voltages are a safety
risk and the user should be properly isolated from the voltage. Proper care must also be taken to
ensure that there are no additional pathways to ground for the applied voltage besides the capillary.
Buffer
Sample
Sample waste
Injection
valve
Separation
channel
2 mm
Waste
FIGURE 15.1 Schematic representation of a microchip used for high-speed electrophoretic separations. Nar-
row channels were etched in the injection and separation areas, while wide channels were fabricated for all
other sections. These differential channel widths ensured the majority of the potential was applied across the
narrow channels. Owing to the high separation field strengths (up to 6.1 V cm−1 per volt of applied potential),
subsecond separations were possible. (Reproduced from Jacobson, S. C. et al. Anal. Chem., 70, 3476, 1998.
With permission from American Chemical Society.)
Extremely high voltages (greater than 30 kV) can lead to breakdown of the material in the separation
path and research with ultrahigh voltages has shown that shielding, such as immersion in a weakly
conducting solution, will help prevent material breakdown.9 The Jorgenson lab has applied ultrahigh
voltages to increase the temporal response when analyzing tryptic digests of proteins by capillary
electrophoresis.9 High voltages yield faster, more efficient separations but a complicated separation
can still take over an hour, even with 120 kV applied. The technology could be applied to making
second or even subsecond separations in the future.
Amplified electric fields have been obtained by changing the geometry of a capillary as well.
Pulling a capillary to an hourglass shape gives amplified fields in a 50 µm region where the inner
diameter is thinnest. Fields of up to 0.15 MV cm−1 were obtained by applying 5–20 kV potentials.10
In the same vein, microfluidic devices have been fabricated that use differential channel widths
to apply a large percentage of the separation voltage across the appropriate channel. For example,
Ramsey and coworkers11 applied 8.7 kV across a narrow injection valve design (Figure 15.1) and
obtained a 53 kV cm−1 electric field in the separation channel, while 71 kV would have been required
if all channels had the same cross-sectional area.
15.3.2 INJECTION METHODS

Injection technology is also crucial for achieving efficient separations. A thin plug of analyte must
be injected because the band broadening caused by injection is proportional to the width of the
sample plug squared. Sample overloading, or electromigration dispersion, can also occur due to
differences in the electric field between the sample and the separation buffer if the plug is too wide.
Thin plug injections onto small i.d. capillaries or shallow microfluidic channels require that low
volumes of sample are loaded. Injection methods for microfluidic devices will be outlined at the
end of this section; however, the two most common injection methods for capillary electrophoresis
systems are hydrodynamic and electrokinetic injections. Modifications of hydrodynamic injections
and electrokinetic methods are necessary to obtain thin plugs.
Injection valves have been used to introduce samples into a capillary electrophoresis instrument.12
The injection system is similar to a six-port valve used in a conventional HPLC instrument, where
the sample can be loaded into a sample loop on the valve, then the valve actuated to inject the plug
onto the capillary. The size of the injection is determined by the plug size and flow rate. This system
is not the fastest or most reproducible, but it relies on fairly simple technology. Valved injections
have been used in conjunction with separation times down to 10 s.13
Flow-gated injection is also a common method of injection. Pioneered by Jorgenson, this method
involves a cross intersection between the reagent capillary and the separation capillary.14 During
the separation, a cross-flow buffer sweeps the continuously flowing sample out to waste. However,
to make an injection, this cross-flow buffer is stopped, the sample builds up in the gap at the mid-
dle of the cross, and then sample is injected electrokinetically. The disadvantages are the same as
normal electrokinetic injections, as sampling biases occur due to differences in analyte electromi-
gration rates. However, the reproducibility of injections is good, with relative standard deviations
of 4% between runs. A range of flow rates can be used and the injection voltage can be varied
to control the injection volume. Flow switching on a very rapid scale is difficult to achieve, so
flow-gated injection is not useful for the fastest microsecond separations. However, it has been rou-
tinely used for multidimensional separations of 2–3 s15 and coupling to in vivo monitoring with 10 s
separations.16,17
Optical gating is an extremely fast method of injection for CE. The sample is fluorescently tagged
and a separation voltage is applied to allow a continuous stream of sample to be electrokinetically
injected.18 An intense gating laser beam near the entrance to the capillary is applied to photobleach
the fluorophore and render the analyte undetectable with LIF at the probe beam (Figure 15.2).
Only when the gating beam is pulsed “off” is an injection made, allowing tagged molecules to be
separated and detected by LIF. The gating laser beam can be pulsed quickly, on the microsecond
Electroosmotic flow
Capillary
Capillary
support
Gating beam
Capillary
with
polyimide
removed
Probe beam
FIGURE 15.2 Instrumental diagram for optical gating. The gating laser beam photobleaches the sample most
of the time. To make an injection, the gating beam is pulsed off, and analytes can then be detected by LIF
at the probe beam. (Reproduced from Monnig, C. A. and Jorgenson, J. W. Anal. Chem. 63, 802, 1991. With
permission from American Chemical Society.)
level, making optical gating methods faster than fluid switching methods. However, effectively
photobleaching the fluorescent tags can be difficult and incomplete photobleaching can lead to
high background light levels.18 The separation voltage is continuously applied to the capillary, so
capacitive charging when applying a voltage does not limit the sample plug sizes. The sample
plug is limited by the dimensions of the laser focus, the time for switching and the velocity of
the sample migrating through the gate. A concern is that the sample is continuously being injected
onto the column. If the sample adsorbs to the capillary or contains high salt levels that alter EOF,
poor separations may result. Optical gating has the advantage of being amenable to fast, repeated
injections. Therefore, it has been used to apply pulse profiles for multiple injection methods such as
Hadamard transform.19
A twist on optical gating, which photodestroys most analytes, is optical generation of fluorescent
products. For example, multiphoton-excited photoreactions have been used to generate fluorescent
products from thermally labile, nonfluorescent reactants.20 As with optical gating, the laser beams
needed for the reaction can be pulsed, so the fluorescent products are only produced for a very short
period. Another approach is to generate a fluorescent analyte to photoactivate a caged fluorophore.21
Caged fluorescein can be photolyzed below 365 nm yielding a fluorescent species that can be detected
using the 488 nm line of an Ar ion laser. Therefore, a UV laser beam can be used as a gate to activate
a small plug of sample, making it fluorescent and detectable with LIF. This method has lower noise
than traditional optical gating because it does not suffer from the background currents caused by
incomplete photobleaching.
Injection methods for microfluidic experiments are different, but often based on similar principles.
For example, gated injections (either via a perpendicular cross-flow or via an optical gate) that are
similar to their CE counterparts are used for high-speed separations on microdevices. Often these
injection times are fast, less than 5 s is typical, and since the sample remains close to the injection
intersection, serial injections can be easily attained.
A tee intersection,22 double tee,23 or a tee intersection operated in a pinched format24 are the
most popular choices for injection schemes in microfluidic formats due to their simplicity and ability
to deliver time-independent concentrations of samples. These types of injection formats are unique
to microfluidic devices. While short and rapid injections are needed in CE experiments, long loading
times are regularly used with these injection systems even with high-speed separations since (ideally)
no sample enters the separation channel until analysis begins.
The downside to these methods is that sample often diffuses out of the loading channel dur-
ing the analysis, which leads to increased band broadening. One method to reduce this effect is
to apply “pullback” voltages. In this way, the analyte is driven away from the injection intersec-
tion after separation commences to ensure no diffusion of sample into the separation channel. The
disadvantage of the pullback method is that the sample is moved away from the injection intersec-
tion making it difficult to achieve rapid serial injections. To alleviate this problem, a double-tee
format has been used with one tee in the sample channel (allowing the sample to remain close
to the injection cross while the pullback voltage was applied), and a second tee in the separation
channel (which enabled loading of subsequent injections during the analysis).25 With this design,
repetitive injections up to 10 Hz could be achieved, although the method was not coupled to a
separation.
In gated injections the sample is shunted away from the separation channel using a perpendicular
flow.26 To perform an injection, the perpendicular flow is stopped for a specific amount of time
and the sample is allowed to fill the injection cross before reinitiating the perpendicular flow. If the
sample is being driven by application of an electric field, the injection is biased toward analytes
with the highest mobilities. This type of injection is similar to conventional flow-gate injections in
capillary systems. Often the injection times are fast, less than 5 s is typical, and since the sample
remains close to the injection intersection, highly reproducible serial injections can be easily attained
(Figure 15.3).27
16 2.0
1.6
12
Intensity (AU)
B/F
1.2
8
0.8
4
0.4
0 0.0
400 600 800 1000 1200
Time (s)
FIGURE 15.3 Series of 110 consecutive separations performed in a microchip to monitor insulin secretion
via a competitive immunoassay. Fast monitoring can be achieved by increasing the rate at which a reaction is
sampled (e.g., 1 s injections with a 5 s separation time). The gated injection scheme used in these separations
enabled rapid sampling, resulting in high temporal resolution of insulin secretion. The black line through the
peaks shows the bound/free ratios for each electropherogram; reproducibility ranged from 2% to 6% over the
course of 30 min. (Reproduced from Roper, M. G. et al. Anal. Chem., 75, 4711, 2003. With permission from
American Chemical Society.)

Absorbance detection is frequently utilized with capillary electrophoresis but is difficult to use in
conjunction with rapid separations. While most molecules absorb in the UV, small capillaries or
shallow channels entail small pathlengths, resulting in high limits of detection. Detection cells that
lengthen pathlengths, such as Z-shaped cells and bubble cells, are incompatible with fast separations
and can lead to band broadening. Signal averaging can boost signal-to-noise (S/N) ratios but is
inconsistent with rapid detection because it is time consuming.
The most common detection method for fast separations with CE or microfluidic platforms is
LIF. With this method, molecules are usually derivatized with a fluorescent moiety, then a laser is
used to induce fluorescence and the resultant light intensity measured. Lasers can be focused into
small beams, minimizing band broadening due to the length of the detector. On-column detection
is popular because it limits band broadening after the separation. A detection window is created on
the capillary by removing the protective polyimide coating; however, scattering of light from the
capillary can cause noise and decrease S/N ratios. Dovichi and coworkers28 designed a sheath-flow
cell for off-column detection that concentrates the analyte into a Taylor cone after the capillary.
The sheath-flow detection cuvette was square, eliminating the light scattering caused by the round
capillary. A schematic diagram of an LIF instrument is shown in Figure 15.4a with a close-up of
the sheath-flow cell for LIF detection in Figure 15.4b.29 Sheath-flow cells are used to increase S/N
ratios by a factor of 10 over on-column detection.
Similar to capillaries, the most common detection method for high-speed separations on microde-
vices is fluorescence, yet few methods have matched the limits of detection found in sheath-flow
formats routinely used in capillary systems. This is due in part to the microfluidic substrate being
in a planar format and therefore difficult to reproduce the three-dimensional Taylor cone found in
sheath-flow capillary systems. The most simple and common method for high-sensitivity detection
in microdevices is the use of a confocal detection scheme.30
Native-fluorescent LIF has been used to detect proteins with high tryptophan levels31 and
serotonin.32 However, most molecules are not fluorescent and must be derivatized with a fluorescent
moiety in order to be detected. Derivatization has been achieved both with covalent attachments
and affinity agents such as aptamers.33,34 The limitation for LIF as a detection technique for rapid
separations is usually the chemistry of the derivatization reaction. First, the kinetics of the tagging
reaction should be fast, especially for online analysis systems. Some fluorophores such as fluorescein
(a) 450 nm
bandpass
Ar* Laser
PMT Syringe
pump
OPA / -ME aCSF
Iris
Prism
Fused
silica lens
Dialysis probe
351 nm
bandpass
Alignment Flow-gate
optics sheath-flow cuvette
(b) Dialysate
Gating flow
Separation
capillary
Sheath-flow
Excitation beam Scatter
–20 kV Fluorescence
FIGURE 15.4 Instrumental diagram of a flow-gated injection CE instrument for monitoring neurochemi-
cals. (a) Schematic diagram of the optical setup for the detection of analytes after OPA derivatization online.
(b) Schematic diagram of the flow-gated interface and sheath-flow cuvette for detection. A cross-flow buffer
between the reaction and separation capillary normally sweeps the sample to waste. An injection is made by
stopping the cross-flow, allowing sample to build up in the gap, then applying a voltage for an electrokinetic
injection. The sheath-flow cuvette has a sheath buffer flowing around the capillary to concentrate the sample
into a Taylor cone and facilitate off-column detection. (Reproduced from Bowser, M. T. and Kennedy, R. T.
Electrophoresis, 22, 3668, 2001. With permission from WILEY-VCH.)
isothiocyanate (FITC) can take 6–12 h to react completely,35 making them incompatible with meth-
ods requiring fast analysis. However, a few tags for primary amines, such as o-phthaldialdehyde
(OPA)36 and naphthalene dicarboxylic acid (NDA)37 can react within a few minutes. The classes
of compounds that can be detected are limited by the chemical reactions. Methods for derivatizing
primary amines,33 thiols,38 and carboxylic acids39 have been developed, but secondary amines, for
example, have slower kinetics and are much more difficult to rapidly detect. Also problematic is
that derivatization of real-world samples in complex matrices with nM concentration is difficult.40
Despite the drawbacks of the chemical reaction, LIF is popular because the spectral properties of
the fluorescent tag can be tuned and high mass sensitivity detection can be obtained. Detection of
yoctomoles of analyte has been reported using LIF.41
Electrochemical detection has also been used for rapid electrophoretic separations.42,43 The
main challenge of electrochemical detection after electrophoresis is separating the electrode from
the high voltage applied for the separation. Noise at the electrode from the separation voltage must
be minimized for efficient, rapid separations. This is usually accomplished by detecting on-column
after a decoupler that grounds the separation voltage before the electrode or by end-column detection
after the capillary.44 Because there is no EOF after the column, care must be taken to avoid band
broadening during end-column detection. Electrochemistry can provide high sensitivity and low
mass detection limits. Electroactive molecules can be detected without derivatization, an advantage
over LIF. Microelectrodes are compatible with small i.d. capillaries and the small size of electrodes
makes them particularly attractive for interfacing with microchips where miniaturization is a goal.
Electrodes can even be fabricated into chips. Because mass limits of detection are not as good as
with LIF, electrochemical detection has been used with separations in the minute time frame45 but
it has not been used for ultrafast, subsecond separations.
Mass spectrometry (MS) is especially useful for monitoring changes in peptides and proteins.46
MS can be coupled to fast separations;47 however, there are multiple challenges including detection
of low mass samples inherent to rapid electrophoretic separations (see previous section) and coupling
the separation eluent to the mass spectrometer for detection. Electrospray detection has primarily
been used as the interface method, but the separation voltage and electrospray voltage are of different
magnitudes (kV and V, respectively). Coupling of nanospray and CE has been achieved to allow
faster separations.48 Ideal mass analyzers for rapid separations would make measurements several
times a second, making ion traps and time-of-flight analyzers preferred. Quadrupoles are more
sensitive but have slower scan speeds. MS coupled to CE has not been used for ultrafast microsecond
separations, but could be useful for proteomics and other applications that require separations of tens
of compounds. These separations could normally take hours, but could be reduced to the minute
time scale with rapid capillary electrophoresis procedures outlined here.
15.3.4 ELECTRONICS AND DATA ANALYSIS

Analysis of ultrafast separation data requires a consideration of instrumentation electronics and data
analysis. High collection frequencies are needed to sample sharp analyte peaks. In addition, filtering
cutoffs and rise times must be ample to allow fast signals to pass.49 Fast sampling rates, while
providing a better description of the peak shape, require larger bandwidths and can lead to greater
noise from environmental sources. For subsecond separations, sampling rates greater than 1000 Hz
and filter rise times less than 1 ms are needed. Often, high-speed separations are used for monitoring
purposes. In these cases, methods to analyze a large number of electropherograms in a reasonable
time are indispensable.50
A simple approach to increasing S/N ratios for samples, thereby lowering detection limits, is to
average multiple electropherograms from the same sample. The difficulty of this approach with rapid
separations is that performing multiple separations increases the analysis time and sample consumed.
Mathematical multiplexing approaches such as Hadamard transforms allow multiple separations to
be performed at once, increasing throughput and improving S/N. Aspinwall has designed a fast
Hadamard transform injection profile that when coupled with low volume optical gating injections
yields separations with ninefold enhancement in S/N in less than 10 s.19 This is the first multiplexing
method that is amenable to fast monitoring.

For ease in presenting literature examples of rapid electrophoretic separations, this section is broken
into examples using capillary and microfluidic electrophoresis systems.
15.4.1 CAPILLARY ELECTROPHORESIS

The fastest separation to date is the separation of 5-hydroxytryptamine (5-HT) and 5-hydroxy-
tryptophan (5-HTrp) in 20 µs (Figure 15.5).10 Multiphoton excitation was used to create the fluo-
rescent, highly unstable photoproducts.20 Short injection pulses were achieved by pulsing the laser
on for 200 µs to achieve a femtoliter reaction plug. Separation of these products could be achieved
in a 4.7 µm i.d. traditional capillary with a separation distance of 15 µm in under 10 ms. Different
reaction products were observed than when the separation products eluted 0.1–1.0 s after pulse,
indicating the reaction occurs in less than 0.1 s and the products were very unstable. To achieve
microsecond separations, higher electric fields were needed. This was accomplished by pulling
capillaries into an hourglass shape to create extremely high electric fields (MV cm−1 range) in
the pulled region.10 With a separation distance of 10 µm, separations could be performed in these
high fields in 20 µs. This experiment demonstrated that capillary electrophoresis could be used to
analyze short-lived reaction products. The intermediates have similar spectroscopic properties, but
different electrophoretic mobilities that are probed with CE. The technological advances of both the
geometry of the capillary and the width of the injection pulse will allow faster probing of a variety
of reactions.
The Kennedy lab has been at the forefront of using rapid capillary electrophoresis to measure neu-
rochemical changes. Optically gated injection of OPA-derivatized amino acids has been achieved
in less than 2 s.33 However, the high salt concentrations of physiological samples, such as cere-
bral spinal fluid, can reduce EOF, and make separations slower. Still, 10 s temporal resolution for
online monitoring of directly sampled and microdialysis samples were obtained. An instrument has
been developed in the Kennedy lab for online analysis of microdialysis samples after precolumn
800
5HT
product
5HTrp
Counts
product
600
400
0 5 10 15 20
Separation time (µs)
FIGURE 15.5 Microsecond separation of 5-HT and 5-HTrp. The capillary was pulled to an hourglass geometry
to achieve an electric field of 0.15 MV cm−1 over the 10-µm separation distance. (Reproduced from Plenert,
M. L. and Shear, J. B. Proc. Natl Acad. Sci. USA, 100, 3853, 2003. With permission.)
(a) 1
gln
7.5 GABA tau
ser
5.0 0
RFU
8 9 10
2.5 glu
gly
asp
0.0
0.0 2.5 5.0 7.5 10.0
Time (s)
(b)
300
Glutamate (%basal)
CE detection
HPLC equivalent
200
100
Fox odor
0
–10 0 10 20 30
Time (min)
FIGURE 15.6 Use of rapid CE for neurochemical monitoring during a behavioral experiment. (a) Example
electropherogram of a rapid separation of OPA-derivatized amino acids in a microdialysis sample. Overlapping
injections were used, so glutamate (glu) and aspartate (asp) are from the previous injection. Also separated were
taurine (tau), glutamine (gln), serine (ser), and glycine (gly). The enlarged inset shows a clear peak for GABA.
(b) Average peak height for glutamate changes during presentation of a predator fox odor to a rat (n = 14
animals, fox odor present during gray bar). The CE data (triangles) collected a point every 14 s and revealed
that changes in glutamate were fast and large. Averaging the data to show the equivalent 10 min temporal
resolution of HPLC experiments show that both the magnitude and the width of the peak are distorted with
slower temporal resolution techniques. (Adapted from Venton, B. J. et al. J. Neurochem., 96, 236, 2006. With
permission.)
derivatization with OPA or NDA (Figure 15.4).29,37 With flow-gated injection, short, small i.d. cap-
illaries, and electric field strengths of 2000 V cm−1 , separations of amino acids were achieved in
less than 15 s. Sheath-flow detection with LIF allowed low nM detection limits of real samples.
An example electropherogram from a behavioral experiment using this instrumentation is shown
in Figure 15.6a. With overlapping injections, six amino acids were detected every 10 s. Plotting
the peak current for glutamate revealed large, fast amino acid changes in response to a predator
fox odor (Figure 15.6b).16 Conventional detection by HPLC would allow only 10 min temporal
resolution, and grouping the rapid CE data in 10 min bins to mimic that temporal resolution reveals
that fast detection is needed to characterize both the time course and magnitude of the changes.
Capillary electrophoresis has also been used to demonstrate that glutamate and γ -aminobutyric acid
(GABA) transients correlate with behavioral changes during stereotypic mouse behaviors51 and fear
conditioning.52 The application of fast CE separations to neurochemical monitoring is allowing a
better picture of the dynamics of neurotransmission in the rat brain.
Others have also demonstrated fast separations that could be useful for neurochemical monitoring.
Shear’s laboratory has detected cyclization reactions of catecholamines to form trihydroxyindoles
that allowed them to be detected with two photon fluorescence.53 Their method allowed separation of
norepinephrine and epinephrine within 80 s and cofactors such as FADH and NADH within 4 min.
Bowser and coworkers54 have developed rapid chiral CE separations for d-serine in salamander
retinas and rat brain samples. Fast separations of nitrate and nitrite in rat brain perfusates have been
achieved using UV detection in under 1.5 min.55 Multiplexing rapid separation using Hadamard
transforms has improved the detection limits of amino acids.19 Monitoring of amino acid mixtures
at low nM concentrations indicates that this method may be beneficial for neurochemical monitoring
of compounds such as GABA and dopamine, which are usually present at low concentrations.
Conventional capillary electrophoresis is useful for separating small inorganic ions in a complex
mixture. CE with a contactless conductivity detector has been utilized to separate solutions of cations
or anions in under 20 s.13 Together, inorganic ions and cations were detected in under 2 min.56 The
conductivity detector allows detection of ions that are not UV active. Small anions and cations have
also been separated from the complex solutions blood serum and milk.57
Rapid separations of DNA fragments in conventional capillaries have been demonstrated. Com-
mercial instrumentation commonly used to sequence DNA is not equipped to use small diameter,
short length capillaries so extremely fast separations have not been extensively utilized. Boc̆ek and
coworkers58 have shown that CE-LIF detection of denatured DNA fragments can be achieved in
2.5 cm capillaries in less than 45 s. Comparable separations on a traditional CE-UV instrument with
50 cm capillaries required 23 min. Short tandem repeat polymorphisms could be detected because
baseline separation of DNA fragments differing by two nucleotides was possible in under a minute. In
another study, narrow electrophoretic injections were the key to minimizing dispersion and obtaining
fast, efficient separations of DNA. Here, short capillaries were used to demonstrate separations of
double-stranded fragments of DNA within 30 s.59 Fast mutation detection was achieved of a single
base pair mutation in double-stranded mitochondrial DNA in 72 s.
Separations of peptides and proteins are not routinely performed in under a minute. However, the
Jorgenson lab has used rapid CE to investigate isomerization between cis and trans peptide bonds.60
Isomerization of proline-containing peptide bonds from cis to trans can occur in a few seconds. The
multiple peaks detected with CE for peptides containing more than one proline are thought to be
due to differential migrations of the cis and trans conformations. Optical gating injections combined
with a 6 µm i.d. capillary was used to separate cis and trans isomers in 2 s, less than the rate of the
conformational shift. Longer peptides could be separated in 5 s. These studies show that rapid CE is
useful for studying fast isomerization rates for optical isomers.
Affinity probe capillary electrophoresis (APCE) uses a fluorescently labeled ligand to bind a
receptor. Separation of the bound and free affinity probe has been achieved with CE with LIF
detection. One advantage of the rapid CE method is that if the separation is rapid compared to the
dissociation time, affinity constants can be determined.61 Competition experiments can be performed
to observe the conditions of complex formation. For example, a fluorescent GTP analog was used as
an affinity probe and G protein receptor subunits as the ligand.62 The complex and free fluorescent
probe were separated within 20 s and Ras-like G protein complexes could be separated in less than
15 s. The rapid separations allowed kinetics of binding to be studied.
APCE has also been used to investigate the fast kinetics of peptide–protein interactions.63 A flu-
orescently labeled phosphorylated peptide was mixed with a Src homology 2-domain protein and
separations of complex and free probe could be achieved in less than 4 s. These rapid separations
were necessary, because no complex peak was observed with 8 s separations, presumably because
it had dissociated. Estimated dissociation constants for the complex were obtained from peak areas
and results were similar to constants from fluorescence anisotropy data. Aptamers have also been
used as affinity ligands for fast separations. Separations as short as 30 s were accomplished with a
fluorescently labeled aptamer against IgE.34 APCE has also been used to detect nM concentrations
of digoxin in less than 1 min.64
The rapid separations offered by capillary electrophoresis have made it amenable as a detector in
hyphenated techniques. For LC-CE, the total analysis time is usually governed by the LC separation,
which generally takes minutes. However, capillary electrophoresis detection adds more peak capacity
because of a second and orthogonal dimension for separation, and shorter separation conditions for
LC can often be tolerated. For example, a 2.5 min reversed-phase liquid chromatography gradient was
used in conjunction with 2.5 s CE separations for the detection of a tryptic digest of cytochrome c.65
The experiment used optical gating for injections, but flow-gated injections were actually developed
to interface the two separation techniques. For example, using a capillary LC column, flow-gated
injections of LC eluent onto the CE capillary were performed every 30–60 s.14 Fluorescently labeled
phenylalanine and glutamate could be separated in 35 s. A more complicated analysis of fluorescently
labeled urine required longer separations, up to 1 min. However, overlapping electrokinetic injections
were made before one run finished, allowing 30 s temporal resolution.
15.4.2 MICROFLUIDIC APPLICATIONS

Owing to the numerous examples of microfluidic separations in the literature, we have attempted to
limit the applications reviewed to those where the separation was performed in less than 30 s.
As an example of utilizing the gated methodology for rapid, sequential sample introduction,
Roper et al.27 made 1 s injections every 15 s over the course of 30 min to monitor secretion of a
peptide hormone from a single islet of Langerhans on a glass microfluidic chip. During the 15 s
separation, free FITC-insulin and FITC-insulin bound to an insulin antibody were separated and
used to quantify the amount of insulin released from the islet. The Kennedy and coworkers66 have
continued to expand on this initial work by improving the throughput using 0.5 s injections every
5.5 s. These rapid serial injections and separations were used to monitor insulin secretion from an
islet for over 2 h.
Gated injections have also been used in several applications where microdialysis sampling was
coupled to microfluidic devices for separation. For example, Lunte and coworkers67 coupled the
outlet of a microdialysis sampling probe to a microfluidic chip and used the inherent flow resistances
within the device to split a portion of the dialysate to the separation channel. Their device was used
to monitor the product of an enzyme reaction using 1 s injections every 30 s and could observe a
decrease in the substrate peak with a concomitant increase in the product peak. Sandlin et al.68 have
also coupled a microdialysis probe to a microfluidic chip and used this device to monitor changes in
neurotransmitter levels in the brains of anesthetized rats. In this report, the dialysate was mixed and
reacted on-chip with OPA, before a 0.1 s gated injection with a 25 s separation. With this protocol,
sequential injections were made every 130 s over the time course of 100 min while obtaining high
efficiency separations (N = 156,000 for glutamate).
Lapos and Ewing69 have performed 80 ms injections every 5 s for the separation of 4-choloro-7-
nitrobenzofurazan (NBD)-labeled amino acids. They have continued to improve upon this method
by utilizing optical gating to inject analytes in a parallel channel format70 and have achieved five
serial separations of four NBD-labeled amino acids in four parallel channels every 10 s. In addition,
they have performed three serial separations of five fluorescein-labeled oligonucleotides (10, 20, 30,
60, and 80 basepairs in length) within 20 s.71
Li and coworkers72 described the separation of FITC and FITC-labeled antihuman IgG by zone
electrophoresis in a glass microchip. The two compounds were electrokinetically injected into the
separation channel for 2 s and separated in less than 12 s. A comparison between the separation perfor-
mance in capillary and microchip showed that higher efficiency was achieved in the microchip format
due to the shorter separation length and the higher electric field that were applied in microfluidic
structures.
Belder and coworkers73 reported the fast chiral separation of acidic and basic compounds using
a commercially available quartz microchip with a simple cross-tee design and linear imaging UV
detection. Highly sulfated cyclodextrins were added to the separation medium to improve selectiv-
ity and samples were loaded for 60 s using the pinched injection mode. The chiral separation of
norephedrine was achieved in 2.5 s, and a mixture of three basic drugs was resolved in 11 s.
The same microchip electrophoresis system was used by the Takeda and coworkers74 to separate
phenolic compounds [bisphenol A, 4-nonylphenol, 4-(1,1,3,3-tetramethylbutyl)phenol and 4-tert-
butylphenol] by micellar electrokinetic chromatography (MEKC). The samples were loaded for 25 s
using a pinched injection scheme. β-Cyclodextrin was added to the MEKC buffer to further improve
the separation and the four phenolic compounds were resolved in 13 s in a 25 mm separation channel
using an electric field of 450 V cm−1 .
The requirement for low sample injection volumes in microchip electrophoresis often leads to
problems in detection sensitivity. In cases where the analytes of interest are present in low amounts,
sample concentration techniques are typically applied before separation. deMello and coworkers75
has reported performing in-column field-amplified sample stacking (FASS) for the separation and
detection of low concentrations of biogenic amines. A hybrid PDMS/glass device patterned with
a narrow sample channel tee-injector design allowed the sample to be directly injected into the
separation channel without leakage control. The separation length was 30 mm and the sample was
loaded into the separation channel for 10 s. The extent of the sample plug was defined by the chip
design, more specifically by the distance between the injection intersection and the inlet buffer
reservoir, which enabled reproducible injections. Stacking occurred on application of the separation
voltage, and separation of putrescine and tryptamine was performed in less than 20 s. Limits of
detection for putrescine and tryptamine using FASS were 20 and 25 pM, respectively.
Several research groups have investigated the use of polymeric materials to fabricate microfluidic
devices, which can help reduce production costs compared to glass or quartz materials. Henion and
coworkers76 reported the use of Zeonor (a polymer commonly used to manufacture CDs and DVDs)
as the analytical platform for the separation of carnitine derivatives by microchip electrophoresis-MS.
Using a double-tee junction design for sample injection (15 s loading time), carnitine, acetylcarnitine,
and butyrylcarnitine were separated in a 3.5 cm long separation channel in less than 10 s, with an
applied electric field of 2 kV cm−1 .
Nagata et al.77 reported the use of a polyethylene glycol-coated poly(methyl methacrylate)
(PMMA) microchip for microchip gel electrophoresis separations. Trypsin inhibitor, bovine serum
albumin (BSA) and β-galactosidase (labeled with Alexa Fluor 488) were electrokinetically loaded
for 60 s. The sodium dodecyl sulfate (SDS)–protein complexes were then separated and baseline
resolved in 8 s using a 5% linear polyacrylamide/0.1% (w/v) SDS-based separation buffer, 3 mm
separation channel, and 303 V cm−1 electric field.
Fast microchip electrophoresis in polymeric devices was also applied to multiplex enzyme assays,
which enabled high-throughput screening of one or several substrates against multiple enzymes
targets. One particular advantage of this technique is that the activity, specificity, and cross-reactivity
of enzymes for drug candidates may be monitored as well as protein–protein interactions in a single
fast separation. Gibbons and coworkers78 performed a fourplex protein kinase assay in a PMMA
chip patterned with a double-tee injection (150 µm offset) design. The separation was performed in
a buffer containing 1% polyethylene oxide (PEO) to suppress EOF. Three substrate peaks and four
product peaks were separated in less than 25 s.
The fabrication of separation platforms from less common materials has also recently been
described. Wirth and coworkers79 performed electrophoretic separations inside silica colloidal crys-
tals. The advantages of this separation medium include fast mass transport, high electric field
resistance, and chemically modifiable surfaces. The silica colloidal crystals were treated and reacted
with chloro-dimethyl-octadecyl silane to form a C18 stationary phase. The system differed from
traditional electrochromatography in that EOF was negligible. In such a system, separation occurred
through both adsorption and electrophoresis phenomena. A comparison between the Van Deemter
plots obtained from the silica colloidal system (obtained using an unretained dye since there was no
flow) and from a commercial monolithic column showed that the colloidal phase produced much
smaller plate heights due to smaller A (Eddy diffusion) and C (mass transfer) terms. Fast mass
transport was also beneficial to produce fast separations. Three hydrophilic peptides, labeled with
rhodamine and electrokinetically injected for 5 s, were separated in less than 10 s using an electric
field of 800 V cm−1 in a 6-mm-long separation channel.
More complex experimental designs have also been described. Hyphenation of microchip elec-
trophoresis with an orthogonal separation technique was reported by Zhang and coworkers.80
A two-dimensional HPLC/electrophoresis device (fabricated in glass) was used to separate

FITC-labeled peptides. To interface the two techniques, a microchip design similar to a cross-tee
was used, containing a hole in the sample-loading channel that connected the outlet end of the HPLC
capillary column to the microchip channels. The HPLC effluent was continuously introduced into
the sample-loading channels and was repeatedly injected (pinched injection for 3 s) into the sepa-
ration channel every 20 s. The peptides were separated in a 40-mm-long separation channel using
an electric field of 635 V cm−1 . The partial profile of a tryptic digest of BSA was obtained from 10
consecutive injections and separations of the HPLC effluent, with eight FITC-labeled peptides being
separated and detected in a total analysis time of 230 s. With these experimental conditions (i.e., sam-
pling and separation times), two peptides were detected per electrophoretic separation (20 s analysis
time).
Ramsey and coworkers11,81 were one of the first groups reporting subsecond separation in a
microfluidic device. A simple cross-T design, etched into a planar glass chip, was used to separate
rhodamine B and fluorescein in 150 ms, in a 0.9-mm-long separation channel.
Liu et al.82 also reported the microchip electrophoretic separation of three flavin metabolites
(riboflavin, flavin-adenine dinucleotide, flavin mononucleotide) in less than 1 s using a pinched
injection scheme. The experimental conditions were optimized to achieve ultrafast separation, by
shortening the separation channel length and increasing the electric field. The separation quality was
also improved by controlling the sample injection size by optimizing sample pinching dispensing
factors.
Subsecond chiral separations of DNS-labeled amino acids were also reported by Belder and
coworkers83 using quartz microchips. Highly sulfated γ -cyclodextrin was added to the separation
buffer as a chiral selector and the samples were loaded for 20 s using the pinched injection mode.
The separation length and electric field strength (1.5 mm and 2012 V cm−1 , respectively) were
optimized to achieve baseline resolution of DNS-norleucine enantiomers in 720 ms, and DNS-
tryptophan enantiomers in 800 ms (Figure 15.7). The authors also showed the baseline separation of
a more complex sample consisting of three amino acids (DNS-phenylalanine, DNS-norvaline, and
DNS-glutamic acid) in 3.3 s using a 7 mm separation length and 2012 V cm−1 electric field.
R = 1.6 720 ms
0.10
580 ms
0.09
0.08
Fluoresscence
0.07
0.06
0.05
0.04
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

Time (s)
FIGURE 15.7 Subsecond chiral separation of DNS-norleucine achieved using a simple cross-injector quartz
microchip. The separation was performed in a 1.5-mm-long channel, using an electric field of 2012 V cm−1
and 2% highly sulfated γ -cyclodextrin, 25 mM triethylammonium phosphate buffer, pH 2.5, as the separation
buffer. (Reproduced from Piehl, et al. Electrophoresis 25, 3848, 2004.)

In this section, it is our goal to provide a brief procedure that may be followed to produce successful
ultrafast separations.

Rapid capillary electrophoresis measurements in general require small i.d., short length capillaries
with fast injection and sensitive detection techniques. One example of a successful implementa-
tion of all these principles is a rapid capillary electrophoresis instrument developed by Bowser and
Kennedy29 to analyze online microdialysis samples for in vivo monitoring (Figure 15.4). This instru-
ment used small, 10 µm i.d. capillaries that were 10 cm long. Applied voltages were 20,000 V, or
2000 V cm−1 .
Analysis of amino acid neurotransmitters was achieved by online, precolumn derivatization with
OPA. The reaction of OPA with primary amines occurs in less than 2 min. An online reactor was
created by inserting a smaller capillary with the analyte into a larger capillary through, which the
OPA was pumped. After the end of the smaller capillary, the analyte would be mixed with OPA
and derivatized. A flow-gated interface was used for injection. The sample was swept to waste by a
cross-flow buffer except when an injection was made. Then analyte was allowed to accumulate in
the gap (30–50 µm) between the reaction and separation capillaries. A small voltage, less than the
separation voltage, was applied for only 200 ms to perform an electrokinetic injection. This allowed
very efficient separations with only 30 ms wide peaks.
Separations were performed with 40 mM borate buffer. One key is to filter all solutions so that
no dust particles are present that can clog the small capillaries. Hydroxypropyl-β-cyclodextrin was
added to the separation buffer to increase the selectivity and was useful for the separation of isomers
α- and β-aminobutyric acid from GABA. In addition, side chains of the amino acids differentially
interact with the hydroxyl groups on the cyclodextrin, allowing more separations that are efficient.
Overall, efficiencies as high as 500,000 theoretical plates were obtained with separation times less
than 20 s. Overlapping injections, where the separation is stopped (i.e., the voltage turned off)
and more sample injected before the first separation is finished, allowed separations every 10 s
(Figure 15.6a). Limits of detection for glutamate were less than 50 nM.
LIF in a sheath-flow cuvette allows detection of small amounts of analyte with high S/N ratios.
The sheath-flow cuvette improved S/N ratios 15 times over conventional on-capillary detection. The
end of the separation capillary was inserted into a square, quartz cuvette. Buffer was siphoned around
the outside of the capillary end, which had been ground to a point. This created a Taylor cone of
analyte. For OPA, excitation was with a 351 nm line of an Ar ion laser and emission at 450 nm was
collected at a 90 angle using an objective to focus it on a photomultiplier tube. Collection of light
at this angle avoids interference from the light of the incident beam. Software programs automated
data collection and allowed the analysis of hundreds of electropherograms simultaneously.
To summarize
1. High voltages are needed for rapid separations (electric fields greater than 1000 V cm−1 ).
2. Short separation distances are crucial for fast analyses.
3. Small i.d. capillaries reduce band broadening due to joule heating.
4. Precolumn derivatization with a fast reaction eliminates band broadening issues with
postcolumn derivatization.
5. Short injection times, less than a second, yield narrow peaks and efficient separations.
6. A flow-gated interface allows sampling and injection from a continuous flow stream.
7. LIF is a high-sensitivity detection scheme for low quantities of analyte.
8. A sheath-flow cuvette increases S/N ratios and decreases detection limits.
9. Rapid, batch analysis of hundreds of electropherograms enables fast data analysis.
15.5.2 MICROFLUIDIC CHIPS

As detailed in Section 15.4, there are multiple examples of fast separations performed in microfluidic
devices and the channel design and speed of the separation is ultimately dictated by the analytes to
be separated. In one example, Roper et al.27 detected insulin secretion from islets of Langerhans
using a combination of several factors designed for rapid separations. These factors included the use
of shallow channels, 1 s gated injections, 15 s separations, and high-speed data processing.
Use of shallow channels reduces current (and therefore, Joule heating) by increasing the overall
resistance of the device. Unfortunately, these type of channels were also easy to clog with dust
particulates, salt, or cellular debris so a thorough cleaning procedure before and after experiments
was outlined in a subsequent paper.67 This cleaning method included flushing the channels with
NaOH, followed by multiple buffer rinses. In addition, caps were placed on top of microfluidic
reservoirs to reduce dust and particulate entry into the device over a day of experimentation. With
these types of precautions in place, separations every 10 s over a period of several hours could be
performed.
Gated injections were used in this example since a pullback flow would perturb the insulin
concentration upstream that was being monitored. With 1 s injections, the absolute number of analytes
being detected was relatively low; however, the concentration being detected, approximately 100 nM
FITC-insulin was high for LIF detection. The separation buffer was also optimized for high-speed
separations by using a zwitterionic buffer. Zwitterionic buffers or separation buffers with added
organic modifiers allow the application of high separation voltages while maintaining a low current,
essential for rapid separations.7 Finally, a software program was written to analyze the large amount of
data being produced from these rapid separations.50 Without this specialized software, data analysis
was orders of magnitude longer than the separation time, limiting the utility of the method.
An analogous circuit diagram should be produced while the fluidic architecture is being designed.
With this diagram, the distribution of the separation voltage over the channel network can be tested
using relative resistance values84−86 before spending the time and money fabricating the photomask
or device. Through careful consideration of the relative resistances, the majority of the separation
voltage can be applied to specific areas within the channel network, such as across the separation
channel, leading to decreased separation times.
To summarize
1. Fabrication of analogous circuit diagram ensures proper channel lengths and widths (rel-
ative resistances) so that if only 3000 V is applied across the entire device, most of the
voltage is dropped across the separation channel.
2. Production of a glass microfluidic device with shallow channels. These two features allow
for high EOF rates while maintaining a low current.
3. Extensive cleaning of the channel surfaces before and after experiments with NaOH and
buffers.
4. Use a high concentration of zwitterionic buffer to allow application of a high voltage while
maintaining a low current.
5. One-second gated injections using a computer-controlled high voltage relay followed by
a 10–15 s separation allows rapid, serial monitoring.
6. LIF in conjunction with an epifluorescence microscope facilitates sensitive detection of
low mass, low volume injections.
7. High-throughput data program allows for efficient postprocessing of collected data.
15.6 CONCLUSIONS
Capillary electrophoresis has continuously demonstrated that short analysis times are compatible with
well-resolved separations. With the use of microfluidic devices, more integrated sample processing
and handling is possible while maintaining the benefits associated with rapid separations. The major
obstacles to performing rapid separations are nondiffusional sources of band broadening (i.e., large
injection amounts and finite detection zones). With the proper choice of instrumentation and ana-
lytes, separation time scales could be readily achieved below 1 s. The development of commercial
instrumentation is needed for widespread implementation of rapid electrophoresis methodologies.
It will be interesting to observe if technology development in the areas of injection and detection
techniques paves the way for new applications, or if the applications drive the development of new
technologies. Certainly now it appears that technology is at the forefront as many examples of high-
speed separations are performed on fluorescent dyes. When the reverse is true, the true power of
rapid separations may be observed. Thus, the research field of rapid electrophoretic separations is
expected to be active for quite some time.
REFERENCES
1. Tiselius, A. A new apparatus for electrophoretic analysis of colloidal mixtures. Trans. Faraday Soc.
33, 524, 1934.
2. Jorgenson, J. W. and Lukacs, K. D. Zone electrophoresis in open-tubular glass capillaries. Anal. Chem.
53, 1298, 1981.
3. Jorgenson, J. W. and Lukacs, K. D. Capillary zone electrophoresis. Science 222, 266, 1983.
4. Liu, J., Dolnik, V., Hsieh, Y.-Z., and Novotny, M. Experimental evaluation of the separation efficiency
in capillary electrophoresis using open tubular and gel-filled columns. Anal. Chem. 64, 1328, 1992.
5. Hjerten, S., Valtcheva, L., Elenbring, K., and Liao, J. L. Fast, high-resolution (capillary)
electrophoresis in buffers designed for high field strengths. Electrophoresis 16, 584, 1995.
6. Hjerten, S. and Kubo, K. A new type of pH- and detergent-stable coating for elimination of
electroendosmosis and adsorption in (capillary) electrophoresis. Electrophoresis 14, 390, 1993.
7. Bushey, M. M. and Jorgenson, J. W. Capillary electrophoresis of proteins in buffers containing high
concentrations of zwitterionic salts. J. Chromatogr. 480, 301, 1989.
8. Qiao, R. Control of electroosmotic flow by polymer coating: Effects of the electrical double layer.
Langmuir 22, 7096, 2006.
9. Hutterer, K. M. and Jorgenson, J. W. Ultrahigh-voltage capillary electrophoresis. Anal. Chem. 71,
1293, 1999.
10. Plenert, M. L. and Shear, J. B. Microsecond electrophoresis. Proc. Natl Acad. Sci. USA 100, 3853, 2003.
11. Jacobson, S. C., Culbertson, C. T., Daler, J. E., and Ramsey, J. M. Microchip structures for
submillisecond electrophoresis. Anal. Chem. 70, 3476, 1998.
12. Hogam, B. L., Lunte, S. M., Stobaugh, J. F., and Lunte, C. E. On-line coupling of in vivo microdialysis
sampling with capillary electrophoresis. Anal. Chem. 66, 596, 1994.
13. Wuersig, A., Kuban, P., Khaloo, S. S., and Hauser, P. C. Rapid electrophoretic separations in short
capillaries using contactless conductivity detection and sequential injection analysis manifold for
hydrodynamic sample loading. Analyst 131, 944, 2006.
14. Hooker, T. F. and Jorgenson, J. W. A transparent flow gating interface for the coupling of microcolumn
LC with CZE in a comprehensive two-dimensional system. Anal. Chem. 69, 4134, 1997.
15. Moore, A. W. and Jorgenson, J. W. Rapid comprehensive 2-dimensional separations of peptides via
RPLC optically gated capillary zone electrophoresis. Anal. Chem. 67, 3448, 1995.
16. Venton, B. J., Robinson, T. E., and Kennedy, R. T. Transient changes in nucleus accumbens amino
acid concentrations correlate with individual responsivity to the predator fox odor 2,5-dihydro-2,4,5-
trimethylthiazoline. J. Neurochem. 96, 236, 2006.
17. Ciriacks, C. M. and Bowser, M. T. Monitoring d-serine dynamics in the rat brain using online
microdialysis-capillary electrophoresis. Anal. Chem. 76, 6582, 2004.
18. Monnig, C. A. and Jorgenson, J. W. On-column sample gating for high-speed capillary zone
electrophoresis. Anal. Chem. 63, 802, 1991.
19. Braun, K. L., Hapuarachchi, S., Fernandez, F. M., and Aspinwall, C. A. Fast Hadamard transform
capillary electrophoresis for on-line, time-resolved chemical monitoring. Anal. Chem. 78, 1628, 2006.
20. Gordon, M. J., Okerberg, E., Gostkowski, M. L., and Shear, J. B. Electrophoretic characterization of
transient photochemical reaction products. J. Am. Chem. Soc. 123, 10780, 2001.
21. Hapuarachchi, S., Premeau, S. P., and Aspinwall, C. A. High-speed capillary zone electrophoresis with
online photolytic optical injection. Anal. Chem. 78, 3674, 2006.
22. Harrison, D. J. et al. Capillary electrophoresis and sample injection systems integrated on a planar
glass chip. Anal. Chem. 64, 1926, 1992.
23. Effenhauser, C. S., Manz, A., and Widmer, H. M. Glass chips for high-speed capillary electrophoresis
separations with submicrometer plate heights. Anal. Chem. 65, 2637, 1993.
24. Jacobson, S. C. et al. Effects of injection schemes and column geometry on the performance of
microchip electrophoresis devices. Anal. Chem. 66, 1107, 1994.
25. Thomas, C. D., Jacobson, S. C., and Ramsey, J. M. Strategy for repetitive pinched injections on a
microfluidic device. Anal. Chem. 76, 6053, 2004.
26. Jacobson, S. C., Hergenroder, R., Moore, A. W., and Ramsey, J. M. Precolumn reactions with
electrophoretic analysis integrated on a microchip. Anal. Chem. 66, 4127, 1994.
27. Roper, M. G., Shackman, J. G., Dahlgren, G. M., and Kennedy, R. T. Microfluidic chip for continu-
ous monitoring of hormone secretion from living cells using an electrophoresis-based immunoassay.
Anal. Chem. 75, 4711, 2003.
28. Ye, M., Hu, S., Quigley, W. W., and Dovichi, N. J. Post-column fluorescence derivatization of pro-
teins and peptides in capillary electrophoresis with a sheath flow reactor and 488 nm argon ion laser
excitation. J. Chromatogr. A 1022, 201, 2004.
29. Bowser, M. T. and Kennedy, R. T. In vivo monitoring of amine neurotransmitters using microdialysis
with on-line capillary electrophoresis. Electrophoresis 22, 3668, 2001.
30. Ocvirk, G., Tang, T., and Harrison, D. J. Optimization of confocal epifluorescence microscopy for
microchip-based miniaturized total analysis systems. Analyst 123, 1429, 1998.
31. Lee, T. T. and Yeung, E. S. High-sensitivity laser-induced fluorescence detection of native proteins in
capillary electrophoresis. J. Chromatogr. 595, 319, 1992.
32. Zhang, X., Stuart, J. N., and Sweedler, J. V. Capillary electrophoresis with wavelength-resolved
laser-induced fluorescence detection. Anal. Bioanal. Chem. 373, 332, 2002.
33. Tao, L., Thompson, J. E., and Kennedy, R. T. Optically gated capillary electrophoresis of
o-phthaldialdehyde/B-mercaptoethanol derivatized amino acids for chemical monitoring. Anal. Chem.
70, 4015, 1998.
34. German, I., Buchanan, D. D., and Kennedy, R. T. Aptamers as ligands in affinity probe capillary
35. Fukushima, T., Usui, N., Santa, T., and Imai, K. Recent progress in derivatization methods for LC and
CE analysis. J. Pharm. Biomed. Anal. 30, 1655, 2003.
36. Kutlan, D., Presits, P., and Molnar-Perl, I. Behavior and characteristics of amine derivatives obtained
with o-phthaldialdehyde/3-mercaptopropionic acid and with o-phthaldialdehyde/N-acetyl-l-cysteine
reagents. J. Chromatogr. A 949, 235, 2002.
37. Shou, M. et al. In vivo monitoring of amino acids by microdialysis sampling with on-line derivatization
by naphthalene-2,3-dicarboxyaldehyde and rapid micellar electrokinetic capillary chromatography.
J. Neurosci. Methods 138, 189, 2004.
38. Lada, M. W. and Kennedy, R. T. In vivo monitoring of glutathione and cysteine in rat caudate nucleus
using microdialysis on-line with capillary zone electrophoresis-laser induced fluorescence detection.
J. Neurosci. Methods 72, 153, 1997.
39. Toyo’oka, T. Use of derivatization to improve the chromatographic properties and detection selectivity
of physiologically important carboxylic acids. J. Chromatogr. B 671, 91, 1995.
40. de Bee, T., Velthorst, N. H., Brinkman, U. A. T., and Gooijer, C. Laser-based non-fluorescent detection
techniques for liquid separation systems. J. Chromatogr. B 971, 1, 2002.
41. Chen, D. Y. and Dovichi, N. J. Yoctomole detection limit by laser-induced fluorescence in capillary
electrophoresis. J. Chromatogr. B 657, 265, 1994.
42. Holland, L. A. and Leigh, A. M. Amperometric and voltammetric detection for capillary electrophore-
sis. Electrophoresis 23, 3649, 2002.
43. Vandaveer, W. R., Pasas, S. A., Martin, R. S., and Lunte, S. M. Recent developments in amperometric
detection for microchip capillary electrophoresis. Electrophoresis 23, 3667, 2002.
44. Wallenborg, S. R., Nyholm, L., and Lunte, C. E. End-column amperometric detection in capillary
electrophoresis: Influence of separation-related parameters on the observed half-wave potential for
dopamine and catechol. Anal. Chem. 71, 544, 1999.
45. Chen, G., Zhang, L. Y., and Wang, J. Miniaturized capillary electrophoresis system with a carbon
nanotube microelectrode for rapid separation and detection of thiols. Talanta 64, 1018, 2004.
46. Babu, C. V. et al. Capillary electrophoresis at the omics level: Towards systems biology. Electrophoresis
27, 97, 2006.
47. Servais, A. C., Crommen, J., and Fillet, M. Capillary electrophoresis-mass spectrometry, an attractive
tool for drug bioanalysis and biomarker discovery. Electrophoresis 27, 2616, 2006.
48. Hsieh, F., Baronas, E., Muir, C., and Martin, S. A. A novel nanospray capillary zone electro-
phoresis/mass spectrometry interface. Rapid Commun. Mass Spectrom. 13, 67, 1999.
49. Moore, A. W. and Jorgenson, J. W. Study of zone broadening in optically gated high-speed capillary
50. Shackman, J. G., Watson, C. J., and Kennedy, R. T. High-throughput automated post-processing of
separation data. J. Chromatogr. A 1040, 273, 2004.
51. Presti, M. F. et al. Behavior-related alterations of striatal neurochemistry in a mouse model of
stereotyped movement disorder. Pharmacol. Biochem. Behav. 77, 501, 2004.
52. Venton, B. J., Robinson, T. E., Kennedy, R. T., and Maren, S. Dynamic amino acid increases in
the basolateral amygdala during acquisition and expression of conditioned fear. Eur. J. Neurosci. 23,
3391, 2006.
53. Gostkowski, M. L., Wei, J., Okerberg, E., and Shear, J. B. Attomole electrophoretic analysis of
catecholamines using copper-catalyzed intramolecular cyclization. Anal. Biochem. 303, 199, 2002.
54. O’Brien, K. B. et al. A high-throughput on-line microdialysis-capillary electrophoresis assay for
d-serine. Electrophoresis 24, 1227, 2003.
55. Gao, L. et al. Determination of nitrate and nitrite in rat brain perfusates by capillary electrophoresis.
Electrophoresis 25, 1264, 2004.
56. Rainelli, A. and Hauser, P. C. Fast electrophoresis in conventional capillaries by employing a rapid
injection device and contactless conductivity detection. Anal. Bioanal. Chem. 382, 789, 2005.
57. Yang, W. and Zhang, Z. Fast and direct determination of small anions in proteinaceous samples using
low voltage-short tube capillary zone electrophoresis system. Anal. Lett. 36, 465, 2003.
58. Kleparnik, K. et al. Ultrafast detection of microsatellite repeat polymorphism in endothelin 1 gene by
electrophoresis in short capillaries. Electrophoresis 21, 238, 2000.
59. Muller, O., Minarik, M., and Foret, F. Ultrafast DNA analysis by capillary electrophoresis/laser-
induced fluorescence detection. Electrophoresis 19, 1436, 1998.
60. Moore, A. W., Jr. and Jorgenson, J. W. Resolution of cis and trans isomers of peptides containing
proline using capillary zone electrophoresis. Anal. Chem. 67, 3464, 1995.
61. Heegaard, N. H. and Kennedy, R. T. Antigen-antibody interactions in capillary electrophoresis.
J. Chromatogr. B 768, 93, 2002.
62. Jameson, E. E. et al. Detection of G proteins by affinity probe capillary electrophoresis using a
fluorescently labeled GTP analogue. Anal. Chem. 75, 4297, 2003.
63. Yang, P. et al. Capillary electrophoresis and fluorescence anisotropy for quantitative analysis of peptide-
protein interactions using JAK2 and SH2-beta as a model system. Anal. Chem. 77, 2482, 2005.
64. Hafner, F. T. et al. Noncompetitive immunoassay of small analytes at the femtomolar level by
affinity probe capillary electrophoresis: Direct analysis of digoxin using a uniform-labeled scFv
immunoreagent. Anal. Chem. 72, 5779, 2000.
65. Hutterer, K. M. and Jorgenson, J. W. Separation of hyaluronic acid by ultrahigh-voltage capillary gel
electrophoresis. Electrophoresis 26, 2027, 2005.
66. Shackman, J. G., Dahlgren, G. M., Peters, J. L., and Kennedy, R. T. Perfusion and chemical monitoring
of living cells on a microfluidic chip. Lab Chip 5, 56, 2005.
67. Huynh, B. H., Fogarty, B. A., Lunte, S. M., and Martin, R. S. On-line coupling of microdialysis
sampling with microchip-based capillary electrophoresis. Anal. Chem. 76, 6440, 2004.
68. Sandlin, Z. D., Shou, M., Shackman, J. G., and Kennedy, R. T. Microfluidic electrophoresis chip
coupled to microdialysis for in vivo monitoring of amino acid neurotransmitters. Anal. Chem. 77,
7702, 2005.
69. Lapos, J. A. and Ewing, A. G. Injection of fluorescently labeled analytes into microfabricated chips
using optically gated electrophoresis. Anal. Chem. 72, 4598, 2000.
70. Xu, H., Roddy, T. P., Lapos, J. A., and Ewing, A. G. Parallel analysis with optically gated sample
introduction on a multichannel microchip. Anal. Chem. 74, 5517, 2002.
71. Xu, H. et al. Parallel separations of oligonucleotides with optically gated sample introduction on
multichannel microchips. J. Sep. Sci. 27, 7, 2004.
72. Rodriguez, I., Zhang, Y., Lee, H. K., and Li, S. F. Y. Conventional capillary electrophoresis
in comparison with short-capillary capillary electrophoresis and microfabricated glass chip capil-
lary electrophoresis for the analysis of fluorescein isothiocyanate anti-human immunoglobulin G. J.
Chromatogr. A 781, 287, 1997.
73. Ludwig, M., Kohler, F., and Belder, D. High-speed chiral separations on a microchip with UV detection.
Electrophoresis 24, 3233, 2003.
74. Wakida, S-I. et al. On-chip micellar electrokinetic chromatographic separation of phenolic chemicals
in waters. J. Chromatogr. A 1109, 179, 2006.
75. Beard, N. P., Zhang, C-X., and deMello, A. J. In-column field-amplified sample stacking of biogenic
amines on microfabricated electrophoresis devices. Electrophoresis 24, 732, 2003.
76. Kameoka, J., Craighead, H. G., Zhang, H., and Henion, J. A polymeric microfluidic chip for CE/MS
determination of small molecules. Anal. Chem. 73, 1935, 2001.
77. Nagata, H., Tabuchi, M., Hirano, K., and Baba, Y. High-speed separation of proteins by microchip
electrophoresis using a polyethylene glycol-coated plastic chip with a sodium dodecyl sulfate-linear
polyacrylamide solution. Electrophoresis 26, 2687, 2005.
78. Xue, Q. F., Wainright, A., Gangakhedkar, S., and Gibbons, I. Multiplexed enzyme assays in capillary
electrophoretic single-use microfluidic devices. Electrophoresis 22, 4000, 2001.
79. Zheng, S., Ross, E., Legg, M. A., and Wirth, M. J. High-speed electroseparations inside silica colloidal
crystals. J. Am. Chem. Soc. 128, 9016, 2006.
80. Yang, X. et al. Comprehensive two-dimensional separations based on capillary high-performance
liquid chromatography and microchip electrophoresis. Electrophoresis 24, 1451, 2003.
81. Jacobson, S. C., Hergenroder, R., Koutny, L. B., and Ramsey, J. M. High-speed separations on a
microchip. Anal. Chem. 66, 1114, 1994.
82. Liu, B-F., Hisamoto, H., and Terabe, S. Subsecond separation of cellular flavin coenzymes by microchip
capillary electrophoresis with laser-induced fluorescence detection. J. Chromatogr. A 1021, 201, 2003.
83. Piehl, N., Ludwig, M., and Belder, D. Subsecond chiral separations on a microchip. Electrophoresis
25, 3848, 2004.
84. Seller, K., Fan, Z. H., Flurl, K., and Harrison, D. J. Electroosmotic pumping and valveless control of
fluid flow within a manifold of capillaries on a glass chip. Anal. Chem. 66, 3485, 1994.
85. Attiya, S. et al. Design of an interface to allow microfluidic electrophoresis chips to drink from the fire
hose of the external environment. Electrophoresis 22, 318, 2001.
86. Kim, D., Chesler, N. C., and Beebe, D. J.Amethod for dynamic system characterization using hydraulic
series resistance. Lab Chip 6, 639, 2006.
16 DNA Sequencing by Capillary
Electrophoresis
David L. Yang, Rachel Sauvageot, and
Stephen L. Pentoney, Jr.
CONTENTS
16.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467

16.2 Core Capillary Electrophoresis Sequencing System Technologies . . . . . . . . . . . . . . . . . . . . . . . . 469
16.2.1 DNA Sequencing Reaction Chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 469
16.2.1.1 Sequencing Chemistries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 469
16.2.1.2 Fluorescent Dyes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471
16.2.2 Surface Coating and Separation Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 472
16.2.2.1 Capillary Characteristics and EOF Suppression . . . . . . . . . . . . . . . . . . . . . . . . . 472
16.2.2.2 Role of the Separation Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473
16.2.2.3 Evolution of Separation Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 475
16.2.3 Excitation/Detection Methods: The Road to Four-Color Multicapillary
Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 480
16.2.3.1 Single-Capillary On-Column Detection Systems . . . . . . . . . . . . . . . . . . . . . . . 481
16.2.3.2 Multicapillary On-Column Detection Systems . . . . . . . . . . . . . . . . . . . . . . . . . . 484
16.2.3.3 Sheath-Flow Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 488
16.2.3.4 Other Detection Schemes Explored . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 492
16.2.3.5 Integrated Sample Processing and Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 494
16.2.4 Separation Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495
16.2.4.1 Sample Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495
16.2.4.2 Separation Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496
16.2.5 Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 498
16.2.5.1 Basic Sequence Calling Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 498
16.2.5.2 Quality Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
16.3 The Next Generation: Separations in Microfluidic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
16.4 Other Nucleic Acid Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505
16.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 506
16.1 INTRODUCTION
The cell is the basic functional unit of all living systems and the instruction set required to manage
all cellular activities is contained within the biopolymer known as deoxyribonucleic acid (DNA).
The sequential ordering of paired bases within the DNA polymer codes for the exact instructions
that define an organism and determine its ability to survive and reproduce. DNA sequencing is the
process of determining the unknown locations and ordering of bases within an organism’s genome.
Identifying and understanding sequence changes or “DNA variations” among individuals can lead
467
TABLE 16.1
Milestones in DNA Sequencing by Capillary Electrophoresis
Year References Description
1981 [204] Capillary zone electrophoresis revisited
1985 [31] High-sensitivity CE-LIF detection
1988 [64] Single base resolution using cross-linked polyacrylamide gel-filled capillary
1990 [15] Sequencing using a four-color CE-LIF system
1991–1995 [45–48] Replaceable gels introduced
1994 [172] Single base resolution demonstrated in a microfluidic system
1995 [90] Sequencing performed using noncovalent capillary coating
1995 [175] Four-color separation demonstrated in a microfluidic channel
1996 [13,123] Multicapillary four-color CE-LIF
1995–1998 * Three commercial capillary-based DNA sequencers introduced
1998 [76] Sequencing with low-viscosity polydimethylacrylamide performed using
noncovalent capillary coating
2001 and 2003 [10,11] Sequence and analysis of the Human Genome working draft published
*Applera Corporation, Beckman Coulter, Inc., GE Healthcare.
to significant improvements in the way physicians diagnose, treat, and take measures to prevent the
large number of disorders that affect us.
In the late 1980s and early 1990s, when laboratories involved in DNA sequence analysis were still
using labor-intensive, manually loaded slab gels, many people in the biotechnology field recognized
several potential process improvements offered by capillary electrophoresis (CE). The most highly
valued potential improvement areas include increased throughput, process simplification through
automation, and sequencing cost reduction achieved through minimizing the consumption of key
reagents. The promise of these significant improvements drove substantial R&D investment in CE
and as a result, rapid technological advances were made during the 1990s in the development of
replaceable gels, capillary surface chemistries, and sensitive optical detection methods (the reader
is directed to several informative references [1–9]). Table 16.1 provides an approximate time-line
for several of the advances made in the field of CE that would ultimately prove to be key elements
to improving the field of automated DNA sequencing.
At a very simple level, current state-of-the-art sequencing methodology may be viewed to involve
two fundamental steps. In the first step, an enzyme catalyzed reaction is utilized to create a compli-
cated mixture of molecules that are structurally very similar to one another. Then, in the second step,
these molecules must be rapidly size-sorted and accurately identified at low concentration levels.
None of the advances made in the field of CE were more important than the development of robust,
high-resolution separation matrices required to size-sort these mixtures and the degree of success
quickly realized in this area was quite impressive. Figure 16.1 illustrates rapid progress made during
the 1990s in our ability to resolve single-stranded DNA fragments of increasing size using CE. Over
roughly a 5 year period, reported read-lengths progressed from a “proof-of-principle” few hundred
bases to a much more impressive range of 600–900 bases and commercial efforts in CE ultimately
played a significant role in the early completion of the human genome sequencing initiative [10,11].
Today, automated DNA sequencing is the largest single commercial application of CE and CE has
effectively displaced slab gel based sequencing methodologies. Significant improvements have been
demonstrated in all three of the earlier mentioned process improvement areas, though one may argue
that significant opportunity remains for the reduction of expensive reagent consumption.
The key technology areas involved in the maturation of CE-based DNA sequencing can be broken
down into the following major categories:
• Sequencing reaction chemistry
• Separation matrix and surface coatings
DNA Sequencing by Capillary Electrophoresis 469
10000
73 18 24
64 15 30 205 17 19 21 23
Fragment length (nucleotides)
64 67 29 28 134 16 85 34 20 22 25
1000
100
10
1987 1989 1991 1993 1995 1997 1999 2001 2003
Year
FIGURE 16.1 Graph showing literature reports of single-stranded DNAfragment lengths resolved by capillary
gel electrophoresis versus publication year. The area between the two dashed lines identifies the current read-
length range typically achieved using commercial capillary-based DNAsequencing systems. Reference numbers
are provided above the data points.
• Sample introduction and separation methods

• Excitation/detection systems
• Algorithms
Advances were required in all of these areas in order to make reliable and competitive CE-based
DNA sequencing a reality and in the following sections of this chapter we describe fundamental
progress made in each of these areas as we survey the related literature.
16.2 CORE CAPILLARY ELECTROPHORESIS SEQUENCING

SYSTEM TECHNOLOGIES
16.2.1 DNA SEQUENCING REACTION CHEMISTRY
16.2.1.1 Sequencing Chemistries
All current CE-based DNA sequencing systems utilize chemistry based on the chain termination
method originally described by Sanger et al. [12]. In this approach, a DNA polymerase enzyme is
used, in the presence of chain-terminating dideoxynucleoside triphosphates, to catalyze the extension
of an oligonucleotide primer hybridized to target DNA, as illustrated in Figure 16.2. As chain
extension progresses, a subset of the original target pool is terminated at each base position thus
creating a nested set of single-stranded fragments in the denatured reaction product. These fragments
range in length from that of the primer (plus one nucleotide) to a few thousand bases, well beyond the
upper resolution limit of sequencing matrices developed to date (generally in the range of 1200 bases).
Fluorescent reporter molecules are integrated into the sequencing reaction at either the primer or
ddATP
Target DNA
ddTTP
ddCTP
Labeled, terminated fragments
ddGTP
Target DNA
PPi C
T
G
Primer strand
G A O
O O O O O
O O
O P O CH2
O– P O P O P O CH2 O P O CH2
O–
O– O– O– O–
Target DNA
Labeled, terminated fragment
ddNTP*
T C
dNTP
Primer strand
A polymerase
O
O Target DNA ddNTP*
OH
O P O CH2
C dNTP
O– T polymerase
G
Primer strand
O O O G
O A O
OH O O
O – P O P O P O CH2 O OH
O P O CH2
O – O– O– O P O CH2
– O–
PPi O
FIGURE 16.2 Graphical depiction of Sanger–Coulson chain terminator sequencing chemistry.
the dideoxy chain terminators. Once the extension reaction is complete, the mixture of fluorescently
labeled fragments is purified to remove unincorporated reagents, denatured to free-labeled fragments
from the template strand, and loaded into the inlet end of a gel-filled capillary. Under the influence of
an applied electric field, the carefully formulated sieving matrix contained within the capillary allows
smaller fragments to migrate more freely along the capillary length and so the migrating fragments
become size ordered by the time they reach the detection region located near the capillary outlet.
The identity of the 3 terminal base of each migrating fragment is associated with the corresponding
fluorescent reporter signal and this signal, combined with the ordering of migration, is used for
algorithm-based sequence calling.
During the 1990s, several variations of the chain terminator chemistry were utilized in CE-based
sequencing studies. The first of these variations is known as fluor-primer sequencing. Here, a single
fluorescently labeled primer is elongated during four separate extension reactions in which only
one of the four dideoxy chain terminators is present. The reaction products from each reaction are
separated electrophoretically in a capillary, yielding “one-color primer-labeled” sequencing data
[13]. The one-color primer sequencing chemistry was widely utilized in separation optimization and
gel matrix development studies but was not popular for actual CE-based sequencing of DNA because
four separate reactions and separations would have been required to read sequence. In an alternate
approach, four separate extension reactions are run, each employing one of the four dideoxy chain
terminators and a primer labeled with one of four different fluorophores. The reactions are then
combined and separated in a capillary, yielding “four-color primer sequencing” data [14–25]. This
approach was more acceptable as it required only one separation be performed, but it still suffered
from the requirement of running four separate reactions.
Afundamentally different variation of the dye-primer chain termination chemistry, originally sug-
gested by Tabor and Richardson [26], was explored by several CE groups in the 1990s. Here, relative
peak height, rather than the spectral identity of the fluorescent labels, was used to determine sequence.
This approach offered the hope of simplified instrumentation and chemistry since only one channel
of fluorescence needed to be monitored from a single separation of labeled fragments produced in a
single reaction. Tabor and Richardson [26] showed that incorporation of Mn2+ into sequencing reac-
tion buffers using T7 DNA polymerase nearly eliminated template sequence-dependent variability
in dideoxynucleotide triphosphate (ddNTP) incorporation and thereby resulted in vastly improved
peak height uniformity. They exploited this observation in a slab-based sequencing protocol using
a single fluorescently labeled primer and different concentrations of the four ddNTPs to determine
sequence using relative peak intensity. Ansorge et al. [27] also described using a similar strategy.
Using this approach, the advantages of single-lane sequencing were realized without the associated
use of four different fluors. Pentoney et al. [28] and Dovichi et al. [29,30] investigated the possibility
of adapting the Tabor–Richardson type approach for DNA sequence determination for capillary gel
electrophoresis (CGE) with laser-induced fluorescence (LIF) detection [31,32].
Pentoney et al. [28] demonstrated that Tabor–Richardson type sequencing could be effectively
utilized in combination with a relatively simple and inexpensive CE/LIF system. Here the authors
modified the Tabor–Richardson approach so that only three ddNTPs are included in the sequencing
chemistry. Each template was sequenced twice using two different sets of terminator concentra-
tions and both reactions were analyzed simultaneously. The position of the fourth nucleotide in
the sequence was determined from the location, size, and shape of gaps appearing in the electro-
pherogram in between peaks produced by the three ddNTP terminators present in the reaction. This
simplified the software task of discerning between the different intensity levels and more efficiently
utilized the available system dynamic range. The authors found that careful selection of the termi-
nator concentrations and algorithm-based comparison of the two complementary data sets resulted
in a robust sequence determination.
The current state-of-the art chemistry approach utilizes four dideoxynucleotides, each labeled
with a different fluorophore, to generate fragments from a single reaction [33], which are also ana-
lyzed in one capillary (four-color dye-terminator or “single tube” sequencing) [34]. This approach
eliminates the need to fluorescently label and purify different primer sets, the sequence of which
varies with each target. Note that slab gel based automated sequencers using one or more of these
chemistry approaches were originally developed by Pharmacia, Hitachi, EG&G, Applied Biosys-
tems, and Du Pont. CE-based automated DNA sequencers have now been commercialized by Applera
Corporation, Beckman Coulter, Inc., and GE Healthcare. All the CE-based systems currently utilize
the four-color dye-terminator chemistry in combination with nondiscriminating thermal stable DNA
polymerases that allow linear amplification of reaction product via a chemistry protocol known as
“cycle sequencing.” Cycle sequencing refers to a modified version of the above described Sanger–
Coulson sequencing chemistry in which a thermal stable DNA polymerase is used in order to
repetitively run the extension reaction in a cyclic fashion. The thermal stable DNA polymerases
used today can be repeatedly heated to 95◦ C and still retain their enzymatic activity. Through the use
of repeated temperature cycles (annealing, extension, denaturation), it becomes possible to repeat
the extension reaction over and over again in the same reaction vessel, each time generating more
reaction product from the same sample target. This cycle sequencing method results in an n-fold
increase in reaction product, where n is the number of times the cycle is repeated. Figure 16.3 depicts
today’s typical DNA sequencing chemistry process flow.
16.2.1.2 Fluorescent Dyes

Fluorescent dye sets used in capillary sequencing attached to either primers or terminators have been
developed for gas, solid state, and diode lasers, and they span the spectral range from roughly 488
to 800 nm. The basic requirements of a fluorescently labeled set of four terminators are demanding
and include high-extinction coefficients at wavelengths for which reasonably low-cost lasers exist,
high-quantum yields for fluorescence, sufficient separation of emission maxima for confident iden-
tification, reasonable aqueous buffer solubility, good photostability, minimal impact on enzyme
Basic DNA sequencing chemistry process flow

DNA extraction,
purification
DNA amplification
reaction
Cycle sequencing
reaction
Post reaction
clean-up
Product
suspension
FIGURE 16.3 High-level graphical depiction of the steps involved in running a typical Sanger–Coulson
sequencing chemistry.
incorporation efficiency, and uniform mobility effects upon labeled fragments generated in the
Sanger–Coulson chemistry. No truly perfect set of fluorophores has been developed for DNA
sequencing and algorithm-based data correction is utilized to compensate for some of the above
concerns.
Four-color DNAsequence determination involves the accurate recognition of one of four different
emission profiles at each base position. In addition, in some instances, such as the sequencing of
mixed viral strains, it is necessary to identify the presence of two or more bases at a single-base
location in the electropherogram. To minimize spectral cross talk, families of four dyes have generally
been selected to have emission maxima that differ by about 30 nm. Two different approaches have
been developed to efficiently excite these four-dye families. The first approach is to assign two pairs
of dyes to two different laser excitation sources. This allows the excitation wavelength that is used
to stimulate each dye pair to be located reasonably near the corresponding dye absorbance maxima.
The two laser excitation lines are usually alternately modulated in this scheme. A second approach
is to exploit the use of fluorescence energy transfer [16,35–37]. Here, a common absorber molecule
is used to efficiently couple each of the four labels with a single laser line and a large fraction of
the absorbed laser energy is then rapidly transferred to one of four different acceptor dyes. In this
manner, a single laser source can be used to sensitively address four different labels having very
different emission maxima.
16.2.2 SURFACE COATING AND SEPARATION MATRIX

16.2.2.1 Capillary Characteristics and EOF Suppression
Most CE-based DNA sequencing separations are performed using fused-silica capillaries and some
have been reported for channels etched into glass plates. Since DNA sequencing separations are
performed at alkaline pH (∼8.5) to provide a negatively charged DNA phosphate backbone, the
acidic capillary surface silanol groups will also be ionized. This leads to the formation of negatively
charged surface silanol groups and a charge imbalance near the capillary inner wall. Mobile, hydrated
cationic species attracted to this negatively charged zone define a region of net positive charge density
that decreases in magnitude with increasing distance from the capillary wall. Under the influence
n
H3CO OCH3
Si O O O
H3CO
O O O
R
+
OH
HO Binding of H3CO Si OCH3 Polymerization H CO
Si OH 3 Si OCH3
bifunctional O
HO O
compound Si OH HO Si OH
Glass surface
Glass surface Glass surface
FIGURE 16.4 Inner surface of a capillary derivatized using the chemical method described by Hjertén to
eliminate electroendosmosis and surface adsorption of solutes.
of an applied electric field, the hydrated cationic groups near the capillary wall will migrate toward
the cathode (negatively charged electrode) and will effectively drag bulk fluid along with them in an
interesting pumping phenomenon known as electroendoosmotic flow (EOF) [38]. Unless eliminated,
EOF will lead to gel movement and results in very poor separation quality.
Several methods exist to suppress EOF. The earliest method, described by Hjertén in 1985,
involved covalent attachment of a hydrophilic masking layer of noncross-linked polyacrylamide
to the inner capillary wall in order to prevent both EOF and solute adsorption in free zone and
isoelectric focusing separations of proteins. In this original work, the bifunctional compound γ -
methacryloxypropyltrimethoxysilane was covalently bonded to the inner surface of a glass capillary
and the other end of this bridging molecule was then reacted with acrylamide (AA) monomer under-
going linear polymerization, as illustrated in Figure 16.4 [39]. Several other combinations involving
bifunctional bridging groups and noncross-linked polymers may be used in a similar manner. Addi-
tional covalent attachment chemistries have since been developed to mask the surface silanol groups
[40–43].
A second simpler method that has been used to eliminate EOF is to mask the surface silanol groups
using an adsorbed polymer. This “self-coating” or “dynamic-coating” method will be discussed in
more depth in a later section of this chapter.
16.2.2.2 Role of the Separation Matrix

The separation matrix truly lies at the core of CE-based DNA sequencing. Since DNA fragments have
constant size to charge ratios (each additional base adds another phosphodiester linkage and hence
another unit of charge), both the force acting upon each fragment and the frictional drag experienced
by each fragment increase linearly with size during electrophoresis performed in free solution. This
effect causes the DNA mobility to be size-independent [44], and so a sieving matrix is required in
order to separate DNA fragments of differing lengths from one another. We should keep in mind that
the basic separation requirement in DNA sequencing is to rapidly resolve single-stranded DNA over
a range of approximately 20 bases to more than a thousand bases with single-base resolution. This
is a well-defined nucleic acid mixture, but a very demanding separation indeed.
One of the most important technical hurdles to overcome was the development of a practical
solution to the capillary separation matrix problem. Early capillary gel work involved the adaptation
of cross-linked polyacrylamide gels, used for years in slab-based DNA sequencing separations, to
the capillary format. In cross-linked gels, a DNA mixture is sieved through pores formed by a
network of polymer branches that are covalently fixed. The size selectivity of this network is tuned
through variation of both the monomer (%T) and cross-linker (%C) concentrations. Cross-linked
gels were introduced into capillaries by initiating polymerization in an external vessel and then
driving the flowable, prepolymerized solution into the capillary with a simple syringe. These “fixed”
capillary matrices were capable of resolving DNAfragments but suffered from frequent failure modes
including sharp current drops due to intermittent bubble formation. As a result, use-lives were often
as short as one or two runs [45,46] and fixed capillary gels proved to be an impractical solution for
automated DNA sequencing.
The development of replaceable noncross-linked or linear polyacrylamide gels for DNA sequenc-
ing was probably the single most significant advance made in the field of CE during the 1990s [45–48].
In replaceable gels, a dynamic network of entangled linear polymers forms the pores through which
the DNA mixture is sieved. The size selectivity of this noncross-linked network is tuned through
variation in the length and concentration of the polymers defining the replaceable matrix. Replace-
able gel matrices for CE were described in 1991 for the separation of oligodeoxycytidylic acids
(“poly-C”) 10–15 and 24–36 bases long [47] and analysis of double-stranded restriction digest
products [49]. These gel matrices were not only capable of resolving sequencing fragments with
single base resolution over a wide size range but also easily replaced, thereby eliminating the
need for frequent capillary removal or capillary inlet trimming. A number of reviews provide
detailed discussions of the theory of DNA separation in gels as well as a description of the var-
ious separation matrices that have been evaluated for DNA sequencing [2,44,50–52], especially in
References 53–56.
But what characteristics or physical properties of a replaceable sieving gel are most impor-
tant in terms of resolving single-stranded DNA over such a wide size range? And what metrics
have proven useful in the optimization of gel formulations and separation parameters? Quesada
and Menchen [55] have published a helpful theoretical discussion of the parameters affecting the
performance of semi-dilute replaceable polymer systems used in capillary-based DNA sequencing.
It turns out that for a given separation matrix polymer, the separation speed and resolving power
of the entangled gel matrix are generally dependent on polymer concentration. Below a critical or
“overlap” concentration c*, little or no interaction between matrix polymer strands exists. The over-
lap concentration is generally reached when the space volume 4π(RG 3 )/3 (where R is the radius
G
of gyration of the polymer) is occupied by exactly one polymer strand [55,57]. Above c*, the poly-
mers begin to impinge upon one another and the mesh size of the gel matrix becomes inversely
proportional to polymer concentration. Mesh size in an entangled polymer system may be thought
of as the average distance between polymer–polymer entanglements. For linear polyacrylamide
studied at different concentrations above c* (C > c∗ ), the mesh size was observed to follow the
relationship
ξ = 2.09C −0.76±0.03 , (16.1)
where ξ is the mesh size (in angstroms) and C is the polyacrylamide concentration (in gm/mL)
[2,58]. As an example, a 3.5% noncross-linked linear polyacrylamide (LPA) gel would correspond
to an estimated average pore size of ∼27Å using Equation 16.1.
The past 10 years of capillary gel research has indicated that optimum separation efficiency
and maximum read-length are achieved through the use of highly entangled systems of hydrophilic,
high-average molecular weight polymers. Typical replaceable gel systems utilize polymer molecular
weights on the order of 1–8 million Da and concentrations on the order of 2.5–6%. The polymer
molecular weight distribution, polymer concentration, column length, field strength, and separa-
tion temperature are typically settled upon empirically to yield an optimum compromise between
separation quality and analysis time.
Several models exist to describe the migration of DNAfragments through the separation matrix. In
the Ogston model [59], often applied to cross-linked gels, the DNA fragments are treated as migrating
rigid spheres traversing variously sized pores within the matrix; smaller “spheres” encounter more
pores with equal or greater diameters and migrate more quickly than larger “spheres.” The DNA
140
Migration time (minutes)

120
100
80
60
40
20
0
0 100 200 300 400 500 600 700 800 900 1000
Nucleotide length
FIGURE 16.5 Graph showing the proportional relationship between migration time and DNA fragment length
for a replaceable capillary gel separation of dye-terminator labeled sequencing fragments. (pUC 18 template
DNA, 1009 bases called to 98% accuracy, run on a CEQ analysis system from Beckman Coulter).
fragment mobility in the separation matrix is approximated by the following equation:
µ = µ0 exp(−Kr C) (16.2)
where µ is the mobility in the separation matrix, µ0 is the mobility in free solution, Kr is the
retardation factor and is proportional to both the radius of the matrix polymers and the radius of
gyration of the DNA fragment, and C is the polymer concentration. This model is valid only for
DNA fragments in the approximate size range of 50–300 bases in polyacrylamide gels [56,60].
A second model known as biased reptation with fluctuations (BRF) is an extension of the biased
reptation model (BRM), and treats the DNA fragment not as a solid sphere but as a long flexible tube
that snakes through the pores of the separation matrix [61,62]. The electrophoretic migration times
are predicted to be directly proportional to DNA fragment size, as observed experimentally [50] and
illustrated in Figure 16.5. A large number of reports have confirmed empirically that DNA fragment
mobility in well-formulated and properly run gels is inversely related to molecular weight over a
size range of roughly 20–1000 bases, after which mobility asymptotically approaches a minimum
and the remaining fragments are observed to comigrate. Interested readers are directed to Viovy [54]
for a detailed description of the physical mechanisms of DNA separations.
Another important feature of the separation matrix is its ability to relieve any secondary structure
that arises due to intramolecular base pairing in the single-stranded DNA molecules. The pres-
ence of such regions of secondary structure leads to increased electrophoretic mobility of DNA
fragments and hence congested regions in the electropherogram known as “compressions.” These
compressed regions make accurate base-calling difficult or impossible. To alleviate this problem
gels are typically formulated with the addition of urea, formamide, or other denaturing com-
pounds [25,63]. The use of elevated temperature also helps to disassociate the regions of secondary
structure.
16.2.2.3 Evolution of Separation Matrices
16.2.2.3.1 Cross-linked versus Linear Acrylamides

Capillary gel electrophoretic separation of DNA fragments at single base resolution was first demon-
strated by Karger and coworkers in 1988 [64] and this work was expanded in 1990 [65]. Extending
40
60
0 4 8
min
FIGURE 16.6 Capillary electrophoretic separation with single base resolution of poly-A 40–60 bases long.
(Reprinted with permission from Cohen, A.S., et al., Proc. Natl Acad. Sci. USA, 85, 9660, 1988.)
what had been learned from slab gel work, a cross-linked polyacrylamide gel (T = 7.5%, C = 3.3%)
was formed in a 75 µm i.d. column having an effective length of 13 cm. Using an absorbance-based
detector, the individual fragments in a sample containing polydeoxyadenylic acid homopolymers
(poly-A) 40–60 nt in length were resolved from one another with efficiencies equivalent to 5 × 106
plates per meter in less than 8 min (Figure 16.6). Because of this high-resolving power, the authors
accurately predicted that “capillary gel electrophoresis [may] be used as a tool for oligonucleotide
sequencing” as well as a tool for “rapid assessment of purity” [65]. Paulus et al. [66] demonstrated
single base resolution for samples containing polydeoxythymidilic acid 20–160 nt long using a cross-
linked T = 2.5%, C = 3.3% matrix. Subsequent work using fluorescence detection also demonstrated
DNA fragments with single base resolution [15,29,67–71]. These separation gels typically consisted
of a Tris–borate buffer at pH values >8.0 and with 7–8 M urea used as a denaturant.
However, the cross-linked nature of these separation matrices limited reuse since the polyacry-
lamide degraded, bubbles or voids tended to form in the gels and impurities injected with the sample
accumulated in the capillaries, all leading to instability and unacceptably short average use-life
[45,46]. In an attempt to improve capillary use-life, Righetti and coworkers [72] polymerized AA in
a capillary without a cross-linker (T = 10%, C = 0%) and then treated the resultant media with cys-
teine to scavenge any remaining monomer, since this monomer is both toxic and potentially reactive.
The walls of the capillary were pretreated such that during matrix polymerization, polyacrylamide
strands were also bonded to the inner wall thus eliminating EOF. The authors reported that these
matrix-filled capillaries were expected to have an improved use-life of approximately 2 weeks.
The use-life issue was resolved by replacing the medium within the capillaries after each use
using a non cross-linked gel. Bocek separated a sample of poly-C 10–15 and 24–36 bases long
in a replaceable gel created from a solution containing 10% AA [47] while Karger and coworkers
[45] used replaceable gels to analyze dye-primer labeled DNA sequencing reaction products. In the
latter work, Karger and coworkers [45,46] polymerized a degassed solution containing 6% w/v AA
monomer in a Tris–borate–EDTA(TBE) buffer containing 7 M urea with ammonium persulfate (APS)
and N, N, N , N -tetramethylethylenediamine (TEMED) in 100 µL syringes. Linear polyacrylamide
(LPA) viscosities, ranging from 160 to 3600 centipoise, as measured with a falling ball viscometer,
were found to be inversely proportional to the radical initiator concentration while migration times
were found to be proportional to viscosity, reflecting the size (or molecular weight) of the LPA
strands. LPA in this viscosity range required pressures of the order of 1200 psi to be forced through
capillaries 33 cm long with an inner diameter of 75 µm, well within the range of a syringe pump
[45], thus allowing the matrix to be replaced after every run. Using LPA with an average molecular
weight of approximately 1 × 106 Da (determined via light scatter) and fluorescently labeled primer
samples, fragments >370 bases in length could be identified using capillaries that were modified via
the covalent masking method of Hjertén [39].
Subsequent work using noncross-linked LPA matrices and run under more optimized conditions
led to incremental increases in read-length. Although the separation medium was not replaced after
each run, Dovichi and coworkers [73] reported reads of up to 570 bases in less than 140 min when
a primer-labeled sample was separated at room temperature using a matrix created from a 6% AA
solution polymerized directly in capillaries treated with a solution of [(γ -methacryloyloxy) propryl]
trimethoxysilane to bind a surface layer of polyacrylamide to the capillary walls, similar to the
method of Hjertén [39]. Read-lengths of up to 640 bases were achieved in 2 h run times at 60◦ C
using a capillary containing a matrix consisting of a 5% AA solution [17].
Karger and coworkers [18] using a replaceable separation matrix introduced into the capillary
before each run, performed separations at 50◦ C at a field strength of 150 V/cm using a 2% AA
solution that yielded >1000 bases for a set of dye primer-labeled sequencing reaction fragments.
Separations performed in the temperature range 50–60◦ C were reported to yield higher resolution
for large fragments, decreased separation times, and longer read-lengths [19]. A separation matrix
created by mixing LPA of two different molecular weight distributions, 0.5% (w/w) of 270 kDa and
2% (w/w) 17 MDa, operated at 70◦ C with a field of 125 V/cm yielded read-lengths of ∼1300 bases
in just over 2 h for primer-labeled sequencing reaction fragments [24].
Synthetic procedures incorporating inverse emulsion polymerization were later used to create
a fine powder of high-purity LPA with a molecular mass of 9 MDa [20]. Separations performed
using a matrix created from a combination of this powder and polymer having a molecular weight
of 50 kDa resulted in read-lengths of greater than 1000 bases in runs requiring less than 60 min [21].
More recently, Barron and coworkers [74,75] showed promising results using very sparsely cross-
linked “nanogel” matrices. The authors used very low concentrations of methylene bisacrylamide
(Bis) cross-linker (∼10−4 mol%) in an emulsion polymerization of AA, to produce polymers with
molecular weights of ∼10–11 MDa. The low level of cross-linking was carefully controlled to
retain polymer solution fluidity, but was apparently high enough to stabilize pores within the matrix,
thereby allowing separations of large DNA fragments to occur with higher resolution than is typically
observed using noncross-linked LPA [75]. The authors reported that the improvement in separation
quality for the larger single-stranded fragments led to an increase in read-length of approximately
19% [75].
To date, LPA remains a very popular separation matrix for DNA sequencing, and is used on the
family of sequencers and genetic analyzers available from both Beckman Coulter and GE Health-
care. Figure 16.7 illustrates a typical separation of dye terminator-labeled DNA sequencing reaction
products performed using the CEQTM system from Beckman Coulter and a linear polyacrylamide
based replaceable gel matrix. In this run, 1009 bases were called with 98% accuracy.
16.2.2.3.2 Other Matrix Materials

A gel matrix based upon the use of polydimethylacrylamide (PDMA) has also enjoyed great com-
mercial success [76,77]. Developed by Madabhushi and coworkers at Perkin Elmer Corporation
(now Applera Corporation) and used in their family of high-throughput DNA sequencers, PDMA
0 100 1000 1500 2000 2500 3000 3500 4000 4500 5000
5000 5500 6000 6500 7000 7500 8000 8500 9000 9500 10000 10500 11000
11000 11500 12000 12500 13000 13500 14000 14500 15000 15500 16000 16500 17000 17500 18000
18000 18500 19000 19500 20000 20500 21000 21500 22000 22500 23000 23500 24000 24500 25000 25500 2600
FIGURE 16.7 (See color insert following page 810.) Analyzed four-color dye-terminator sequencing from
a CEQ analysis system (Beckman Coulter): 1009 bases called to 98% accuracy.
polymers of various molecular weights were created by polymerization of N, N-dimethylacrylamide

(DMA) solutions with APS and TEMED at 50◦ C and purification by precipitation. The resulting
polymer had molecular weights in the range 20–200 kDa. A 6.5% PDMA (98 kDa) solution was
reported to have a low viscosity of 75 cP and yielded read-lengths of 600 bases. In addition, since
PDMA is reported to be more hydrophobic than LPA, the segmental adsorption energy of PDMA
onto fused silica is greater than that of LPA [76]. Consequently, PDMA will adsorb to capillary
walls and mask the surface silanol groups, significantly reducing EOF and eliminating the task of
having to chemically modify the capillary walls. Madabhushi [76] reported performing over 100
runs, replenishing the gel before each run, without ever having to recondition the capillary. Heller
[78,79] performed a detailed study of the utility of PDMA in DNA separations by creating polymers
with various molecular weights for use in optimizing polymer concentration, polymer chain length,
and electric field strength. Chu and coworkers [80] achieved read-lengths of up to 1000 bases in 96
min at room temperature using 2.5% w/v PDMA with a molecular weight of 5.2 MDa. Although the
added analysis time would be impractical, preconditioning of PDMA, by subjecting the gel medium
to an electric field strength of 162–320 V/cm for at least 6 h before separation, was reported to
improve peak separation and was predicted to extend the read-length [81].
Barron and coworkers [82] reported success using replaceable poly N-hydroxyethylacrylamide
(PHEA) as both a separation matrix for DNA sequencing and a dynamic coating that eliminated EOF.
Using bare fused-silica capillaries, 4–6% solutions of PHEA with a molecular weight of ∼5.2 MDa
and field strengths ranging from 40 V/cm to 117 V/cm, read-lengths of 445–750 bases were achieved
with 98.5% accuracy. Separation times of 1–5 h were reported in this study.
Chu and coworkers [83] combined LPA and PDMA in an attempt to obtain a gel with the
best properties of each polymer, namely high-quality separations using uncoated capillaries. They
observed that mixtures containing PDMA with high molecular mass or in high concentrations
adversely affected the quality of separations. A mixture of 2.5% (w/v) of LPA with mass 2.2 MDa
and 0.2% PDMA with mass 8 kDa resulted in a matrix that produced read-lengths of ∼700 bases
using bare (untreated) capillaries [83]. To more intimately entangle the LPA and PDMA poly-
mers, the group polymerized PDMA monomer in a homogeneous solution of LPA, thus creating
a “quasi-interpenetrating network” or quasi-IPN [84]. Using single-color dye-primer separations,
they reported sequencing up to 1200 bases in less than 60 min at 65◦ C using 2% quasi-IPN formed
by LPA with a molecular weight of 9.9 MDa and an AA to DMA ratio of ∼11 : 1.
Although AA-based matrices dominate the commercial field of DNA sequencing today, other
polymers have been tested with varied results. Bashkin et al. [85], Marsh et al. [86], and Dolnik
and Gurske [87] reported having some success using hydroxyethylcellulose (HEC), with Bashkin
et al. reporting base calls to 500 bases in less than 60 min using a matrix composed of 2% HEC in
the molecular weight range 90,000–105,000 using Hjertén-modified capillaries [85]. Kheterpal and
Mathies [88] compared the read-lengths achieved using HEC matrix and LPA matrix in sequencing
and found LPA read-lengths to be ∼1000 bases while that of HEC was limited to approximately
600 bases.
Yeung and coworkers [34,89–93] demonstrated success using poly(ethyleneoxide) (PEO) as a
sieving matrix. PEO is an attractive alternative to LPA in the sense that PEO is available commercially
over a broad and high-molecular weight range, eliminating the need to synthesize the long chain
polymer [89]. Similar to PDMA, PEO polymer adsorbs to the capillary walls and masks the surface
silanol groups. However, Yeung’s group [90,94] found that periodic capillary regeneration with HCl
or polyvinylpyrrolidone (PVP) [93] was required to maintain good performance. The authors reported
separations of >900 bases in less than 110 min using a 2.5% solution of PEO with a molecular weight
of 8 MDa and separation performed at 40◦ C using a capillary effective length of 40 cm, an inner
capillary diameter of 75 µm and a field strength of 160 V/cm [93]. Yeung and coworkers [95,96] also
reported using PVP as the separation matrix in uncoated capillaries and demonstrated read-lengths
approaching 350 bases. PVP-based separation media have the dual advantage of masking surface
silanol groups and possessing an extremely low viscosity (<30 cP) [76,93].
Dolnik and coworkers [97] used a novel, natural material to create a sieving matrix for capillary
sequencing. Guar gum, or guaran, is a polysaccharide originating from the endosperm portion of the
legume seed (Cyamopsis tetragonoloba) that grows mainly on the Indian subcontinent and in some
parts of Texas and Oklahoma and contains 75–85% galactomannan. In this sequencing application,
purified 2.1 MDa guaran was dissolved to a final concentration of 15 g/L in a urea/Tris/HEPES buffer,
pumped into a capillary array and run using the MegaBACE system available from GE Healthcare.
Separation of sequencing fragments using this matrix, as well as matrices using galactomannan from
tara gum and locust bean gum exhibited poorer efficiencies than did separations performed using
LPA; read-lengths were generally in the range of 600–700 bases (98.5% accuracy) [97].
Menchen et al. [98] created self-assembling polymers for sieving DNA by capping polyethylene
glycol (PEG) molecules with micelle forming fluorocarbon tails. In solution just above the critical
micelle concentration, these molecules are believed to form intramolecular micellular structures and,
at even higher concentrations, intermolecular micellular structures are believed to form. Reasonably
good resolution was obtained for fragment lengths up to ∼500 bases using a 6% solution of a 1:1
mixture of polymers created from PEG35000 end-capped with (C6 F13 )2 and PEG35000 end-capped
with (C8 F17 )2 [98].
Several new copolymer formulations have recently been created [99]. In the first type of copoly-
mer, AA monomer was copolymerized with a second monomer type, creating a polymer with units
of the two monomers randomly distributed. Chu and coworkers [100] created AA/DMA random
copolymers with molecular weights of ∼2.2 MDa, starting with 3:1, 2:1, and 1:1 molar ratios of
AA and DMA. Using a 2.5% w/v copolymer created with either 3:1 or 2:1 molar ratios, reasonably
105
Water
104 50mM Na. Tris
Viscosity (mPa.s)
50mM Na. Tris + 4M Urea
50mM Na.Tris + 7M Urea

103
102
101
10 20 30 40 50 60 70
Temperature (°C)
FIGURE 16.8 Viscosity change as a function of temperature for thermal associating gels with different
additives. (Reprinted with permission from Sudor, J., et al., Electrophoresis, 22, 720, 2001.)
good resolution up to 700 bases, with visual identification of ∼900 bases, was achieved in room
temperature separations of primer-labeled fragments using uncoated, bare fused-silica capillaries.
Barron and coworkers [101] copolymerized DMA with N, N-diethylacrylamide (DEA) at differ-
ent molar ratios. These copolymers generally performed poorly in comparison to either LPA or
PDMA homopolymers in terms of resolution, with performance suffering as the relative amount of
DEA increased, indicating the role that polymer hydrophobicity plays in separation quality [102].
Although performance suffered with increasing amounts of DEA, the viscosities also decreased,
possibly indicating a tradeoff in performance and ease-of-use [101].
A second class of copolymer is the “thermo-associating” or “thermo-gelling” copolymer [103–
107]. These are made up of two polymers, one of which exhibits a phase change across a critical tem-
perature known as the lower critical solution temperature (LCST). This phase change is characterized
by a change in turbidity and/or viscosity [103,105]. These thermo-associating matrix materials have
the interesting and useful property of switching from a low-viscosity state to a high-viscosity state
upon increasing temperature from ambient (Figure 16.8). This facilitates easy matrix replenishment
at ambient temperature while providing a good separation environment at higher temperatures, where
DNAsequencing separations are normally performed. Examples include the “comb-like” copolymers
consisting of poly (N-isopropylacrylamide) (PNIPAM) groups [106] or PDMA groups [107] grafted
onto an LPA backbone and copolymers of N-ethoxyethylacrylamide and N-methoxyethylacrylamide
[105]. The thermal-thickening behavior of the comb-like copolymers is primarily attributed to the
behavior of the LCST units: at lower temperatures, little interaction between or within the backbones
or grafts exists. As the temperature approaches the LCST, the grafted units begin to associate with
one another, forming the aggregates that provide sieving during separation (Figure 16.9). Presum-
ably due to interactions of the grafted PNIPAM or PDMA units with the capillary walls [107], the
comb-like polymers may be run on bare fused-silica capillaries. Separations made with the PDMA
grafts yielded resolutions and migration times comparable to those of LPA [107].
16.2.3 EXCITATION/DETECTION METHODS: THE ROAD TO FOUR-COLOR

MULTICAPILLARY SYSTEMS
The detection of DNA fragments in a capillary separation channel is a challenging proposition
[5,108]. Injected sample quantities for each (ultimately) resolved band are typically in the sub
nanomolar to picomolar range and detection volumes are on the order of a 100 pL. Peak efficiencies
Water-soluble backbone
LCST grafts
Thermal induced
microdomain
FIGURE 16.9 Simplified depiction of the mechanism for thermo-thickening. Upon heating above the LCST,
grafted side chains become less soluble and tend to aggregate with one another thereby forming a transiently
cross-linked structure. (Reprinted with permission from Sudor, J., et al., Electrophoresis, 22, 720, 2001.)
of ∼105 –106 are required for single base resolution. For a 30 cm capillary, these high-efficiency
numbers correspond to peak widths of approximately 500 µm at the detector window. LIF detection
[31] is the method of choice for capillary-based DNA sequencing because it extends to extremely
low limits of detection, is compatible with the narrow capillary bore confines, is easily extended to
multiwavelength monitoring, and is readily compatible with Sanger–Coulson sequencing chemistries
(both primer and terminator labeling). For example, the CEQ-family of four-color DNA analysis
systems (Beckman Coulter) utilizes diode lasers that are focused to an elliptical spot approximately
60 µm in length along the capillary axis. Within the corresponding 120 pL illuminated volume,
approximately 220 molecules of fluor-labeled DNA in 120 pL illuminated volume may be detected
at a signal-to-noise ratio (SNR) of 2 using LIF detection (unpublished data).
16.2.3.1 Single-Capillary On-Column Detection Systems

Several groups have designed on-column LIF detection systems for capillary-based sequencing.
Excitation and emission light is typically passed through a short portion of the capillary from which
BPF FL
Beam CAP
Argon laser
expander
CL
BPF = band pass filter

FL = focusing lens (f = 200 mm doublet) SF
CAP = capillary(50 µm ID)
CL = collection lens (40x, 0.65NA objective) BPF
SF = spatial filter (slits)
PMT = photomultiplier tube PMT
FIGURE 16.10 Single-color on-column CE/LIF detection system. (Reprinted with permission from Dross-
man, H., et al., Anal. Chem., 62, 900, 1990. Copyright 1990 American Chemical Society.)
the polyimide cladding has been removed, usually by burning or charring. The effective separation
length of the column is defined as the distance from inlet tip of the capillary to this detection window.
Zagursky and McCormick [109] provided the first report of DNA sequencing performed in a
capillary-like format. The authors modified a commercial automated slab-gel sequencing system
(Genesis 2000 DNA Analyzer from E. I. DuPont) to accommodate up to 12 polyacrylamide gel-
filled, 530 µm i.d. by 40 cm long (effective length) columns. The separation of dye-terminator
labeled sequencing reactions was performed using the Genesis 2000 system equipped with an argon-
ion laser and two-color LIF detection optics and sequencing calls to 500 bases with 96% accuracy
were performed in 9.5 h. This work was performed using low field strength but conversion to capillary
dimension tubing (i.e., <200 µm i.d.) was expected to allow the use of higher voltage with a resulting
decrease in separation time.
Smith and coworkers [67] designed one of the first on-column, single-capillary, single-
wavelength systems used to detect primer-labeled DNA sequencing reaction fragments
(Figure 16.10). The output beam of a 40 mW argon-ion laser was passed through a 488 nm bandpass
filter, expanded, and then focused into a 50 µm i.d. capillary to excite fluorescein-labeled DNA
sequencing fragments. Emission light was collected normal to the excitation direction using a 40×,
0.65 NA microscope objective. Slits placed at the objective plane reduced background interference
from excitation light scattered at the capillary walls. A bandpass filter placed just in front of a pho-
tomultiplier tube (PMT) further eliminated scattered light and selected the appropriate detection
wavelength range. The detection limit for this system, determined using vacuum injection of known
concentrations of fluorescein-labeled primers, was estimated to be 60,000 molecules (at a SNR of
2:1). Swerdlow and Gesteland [68] and Karger and coworkers [69], working in parallel with compa-
rable CE/LIF systems, reported detection limits of 10−11 M for a solution of fluorescein, at a SNR
of 3:1 and 5:1, respectively. Swerdlow and Gesteland [68] reported focusing the laser to a spot size
of approximately 20 µm diameter, corresponding to an illumination volume of 30 pL and a reported
detection limit of <200 molecules.
Pentoney et al. [28] used a 543 nm HeNe laser to excite primer-labeled sequencing reaction
fragments. The capillary was threaded though a simple aluminum parabolic reflector and the portion
of the capillary with the polyimide cladding removed to allow detection was located at the focal
point of the reflector (Figure 16.11). The collimated light exiting the reflector passed through several
optical filters to eliminate scattered light before detection using a PMT. A scattering block was placed
across the mouth of the parabolic reflector to reject the intense plane of scattered light that surrounds
the capillary. Detection limits in the low 10−11 M range for solutions of fluor-labeled primers were
reported [28].
Capillary
Parabola
Spring PMT
plunger
screws
Emission filters
Scatter block
FIGURE 16.11 A single-color CE/LIF detection system utilizing a parabolic reflector for capturing and
collimating fluorescence. (Reprinted with permission from Pentoney, S.L., Jr., et al., Electrophoresis, 13, 467,
1992.)
Recognizing the need to perform multicolor detection for truly robust DNA sequencing, a number
of groups devised methods to perform single-capillary four-color detection. Smith and coworkers
[15] expanded their single detector system by incorporating 50/50 mirrors in the emission path in
order to split the emission from four fluorophores and direct it toward four optically filtered PMTs
(Figure 16.12). The four fluorophores, FAM, JOE, TAMRA, and ROX, from Applied Biosystems
(now part of Applera) were coupled to primers and used in four separate chain extension reactions.
These four fluorophores can all be stimulated by a common argon-ion laser, allowing detection in all
four channels to occur simultaneously. However, since 50/50 mirrors were used in the emission path,
only 25% of the original emission intensity for any channel actually reached the detector. Dovichi
and coworkers [29] opted for an alternate time sharing approach in a similar system by placing a
spinning filterwheel, containing four bandpass filter segments, before the detector. Dovichi used two
lasers, an argon-ion laser at 488 nm and a HeNe laser at 543 nm for excitation of the same family of
fluorophores used by Luckey [15], because the 543 nm HeNe laser is a better spectral match with the
absorbance profile of TAMRA and ROX fluorophores. The two lasers were modulated with a two-
sector chopping wheel, with 488 nm light used to stimulate FAM- and JOE-labeled primers while the
543 nm laser was used to stimulate TAMRA- and ROX-labeled primers. Synchrony between the two
spinning wheels allowed assignment of the emission collected at the PMT to a particular spectral
channel and hence to a specific fluor-labeled primer. Since the single detector is now temporally
shared among the four dyes, sensitivity will depend on the integration time for each channel. This,
in turn, will depend on the data acquisition rate, typically set to 2 Hz or greater due to peak width
and sampling considerations [29].
Gesteland and coworkers [71] eliminated the need to split the fluorescence emission or time-share
the detector to obtain multicolor data by using a spectrometer to disperse the emission from the four
Applied Biosystems fluors before imaging onto a charge-coupled device (CCD) camera. Quantum
yield in the 500–600 nm range was 1.6–2.5× greater for the CCD camera than for the PMT which
made this imaging detection system quite attractive. However, the need to place multiple mirrors,
with a composite efficiency of 0.944 = 0.78, and the 0.70–0.75 efficiency of the grating employed
reduced the light levels directed toward the detector by 50%. Sensitivities using this CCD-based
optical system with a 500 ms integration time were reported to be comparable to PMT-based systems
[71]. Karger also eliminated the emission splitting/time sharing issue by using a 488 nm laser and a
543 nm laser focused to two different points on the capillary that were separated by approximately
Lens Capillary
Beam
Argon laser expander
Collection lens
50/50 Beamsplitters Spatial

filter
PMT 1
PMT 3
PMT 2
PMT 4
FIGURE 16.12 Four-color on-column detection system utilizing four dedicated PMTs. A 540 nm bandpass
filter was placed in front of PMT1; a 560 nm bandpass filter was placed in front of PMT2; a 580 nm bandpass
filter was placed in front of PMT3; and a 610 nm bandpass filter was placed in front of PMT4. All pass bands
are 10 nm. (Reprinted from Luckey, J.A., et al., Nucleic Acids Res., 18, 4417, 1990. With permission from
Oxford University Press.)
2 cm [110]. Two collection objectives were used in this study, one for each detection window, to
bring the emission to different regions of a diode array detector. A grating placed before the diode
array dispersed the emission from each of the detection windows, with a resolution of 1 nm per pixel
in the 555–700 nm range. Monitoring different portions of the array gave excitation wavelength-
associated spectral information about the emission and thus about the primer-labeled fragments as
they migrated through the capillary.
16.2.3.2 Multicapillary On-Column Detection Systems

The systems described above may all be regarded as important steps or feasibility demonstrations in
the development of CE-based optical systems for DNA sequencing. The use of capillaries, with their
small cross-sectional areas and high surface-to-volume ratios permitted the use of significantly higher
field strengths than had typically been used in slab gel separations but the corresponding increase
in separation speed was not so great as to give single-capillary sequencing systems a throughput
advantage over slab-based systems where many parallel lanes were run simultaneously. Therefore,
in order to significantly improve sequencing throughput over that already realized with slab-based
systems, many researchers developed fluorescence optical approaches that allowed simultaneous
monitoring of multiple capillaries run in a parallel format. Mathies and coworkers [13,112] and
Mathies et al. [111] developed a multicapillary DNA sequencing system in which a linear array of
capillaries, placed on a moving stage, is scanned under a stationary confocal microscope head, as
displayed in Figure 16.13. In their first report [13], 25 capillaries (100 µm i.d., 200 µm o.d.) were
bundled into a linear array. An array holder, mounted to a motor driven translation stage, held the
capillary windows in optical alignment. For excitation, the output beam of an argon-ion laser was
expanded, collimated, and reflected through a long pass dichroic mirror to a 32 × 0.4 NA microscope
objective, which focused the beam to a 9 µm spot at the plane of the arrayed capillaries. A portion
Preamplifier Computer
Phototube
Spectral filter Preamplifier
Confocal spatial filter
Dichroic beam splitter

Mirror
Focal zone detail

Laser input
Objective
Translation
Capillary stage
array
High voltage
power supply
FIGURE 16.13 Multicapillary detection system utilizing a planar capillary array arrangement. Although
two detectors are shown, one- and four-color detection systems have also been implemented. (Reprinted with
permission from Huang, X.C., et al., Anal. Chem., 64, 2149, 1992. Copyright 1992 American Chemical Society.)
of the fluorescence originating from the capillaries was captured by the same microscope objective
used to focus the laser (confocal arrangement) and transmitted back through the long pass dichroic
mirror to reduce background laser interference at the detector. The fluorescence emission was focused
though a 400 µm pinhole spatial filter, a 488 nm rejection filter, and a long pass filter, all to further
reduce laser scatter. The initial system contained a single PMT and monitored one or two colors, but
later systems reported by the same group used either two or four detectors for multicolor monitoring
[111–113]. In those systems, the emissions from different fluorophores were split and directed to the
appropriate PMTs using dichroic beam splitters. The translation stage, with the capillary array holder
mounted to it, was scanned beneath the microscope objective at 1 Hz. During a sweep, the velocity of
the translational stage was 20 mm/s while data were collected at 2000 Hz; thus, an image resolution
of 10 µm per data point was achieved [13]. Since 100 µm i.d. capillaries were used in this work, 10
points defined the inner bore of the capillary. These 10 measurements were co-added to yield a single
intensity measurement for that capillary and that scan. A detection limit of ∼2 × 10−12 M (SNR = 3)
was reported using solutions of fluorescein that were flowed through the system (so-called flowing
stream characterization measurements) [13].
Bashkin et al. [86,114] modified this system by placing the microscope objective on a trans-
lation stage: the position of the 48-capillary array in their system was fixed and the objective was
translated above the array. The capillary signals were sampled unidirectionally at a sampling rate
of 38 kHz for each of two PMTs, yielding a spatial resolution of 37.5 µm per pixel, and a temporal
resolution of 1 Hz. Flowing stream detection limits of 10−12 M were reported using solutions of
BODIPY505/515 (Molecular probes) prepared in methanol [114]. Automated replacement of the
hydroxyethyl cellulose (HEC) gel separation matrix used in their studies was accomplished via a
stainless steel pressure manifold/anode assembly [114]. In the upper part, four ports were available
for arrays of 16 capillaries to be inserted while one port introduced a common electrode and another
port allowed high-pressure nitrogen to be introduced. The lower part of the manifold assembly con-
tained a 25 mL beaker that contained either the separation matrix or water. Pressurization of up to
1000 psi allowed water or matrix to be pushed out the distal end of the capillary. This system was
extended to four PMTs and eventually commercialized by Molecular Dynamics (now part of GE
Healthcare) as the MegaBACE DNA sequencing system [114,56].
Mathies and coworkers [115] also developed a highly multiplexed system, with the potential
to monitor 1000 capillaries. In this “rotary capillary array electrophoresis scanner,” the capillaries
were supported in grooves on the outside of two coaxial ∼4 diameter cylinders (see Figure 16.14).
A microscope objective, designed to deliver excitation light and collect the resulting emission, was
placed between the two cylinders and mounted on a central shaft that was driven by a stepper motor
to rotate at 2–4 revolutions per second. Light traveled bidirectionally along the axis of the rotating
shaft between the dichroic beam splitter located just outside the top of the shaft and a diagonal
mirror located within the shaft. Simultaneous replacement of separation matrix in all capillaries was
achieved by placing matrix in an O-shaped well, where the capillary tips reside, and pressurizing
with helium. During a sequencing run, buffer and the high voltage electrode were placed in this same
well.
Yeung and coworkers [116–119] studied several on-column excitation and detection schemes and
developed several multicapillary systems. In one scheme [117], the windows of 96 capillaries, with an
i.d. of 75 µm, were aligned in a plane oriented at 45◦ with respect to a CCD camera detector. A sheet
of 514 nm excitation light, oriented perpendicular to the viewing axis of the CCD camera was focused
into the windows of the capillary array. Emission was collected through a holographic filter to reduce
background laser scatter. Two-color fluorescence detection was obtained by directing one image of
the array onto a portion of the CCD chip through a 630 nm longpass filter and directing a second image
of the array onto another region of the CCD chip through a quartz plate without additional optical
filtering. The 1.5 cm range of capillaries imaged covered 300 pixels on the CCD with each individual
capillary window encompassing three pixels in one image. An integration time of 300 ms was used to
acquire each frame at 1.75 frames per second and good uniformity in migration time was observed.
Cross talk in this sheet illumination approach, arising from emission in an adjacent capillary being
refracted by the capillary walls into the CCD camera and appearing in the electropherogram of the
second capillary, was estimated to be <10%. Automation of matrix replenishment was discussed in
this work but not implemented. Heller and coworkers [120] developed a similar system, but used a
grating to disperse the emission onto the CCD camera. Intercapillary cross talk was estimated to be
1–2% in their system. This system was coupled to a stacker that held up to forty 96-well mictotiter
plates, allowing up to 15,000 samples to be run without reloading. However, no automated matrix
replenishment system was described.
Yeung and coworkers [121,122] also developed a side-entry illumination system for exciting up to
24 capillaries. In this work, capillaries with either round or square bores were used. After windowing,
the capillaries were packed side-by-side with the windows aligned and fixed into place using epoxy.
Fused-silica plates were used to create a cell for the windowed region of the capillaries; this cell
was then filled with index-matching fluid to reduce laser light loss at the capillary interfaces. The
514 nm line of an argon-ion laser was focused through the side of the array and transferred through
the detection bores of all capillaries. Because of the refractive index-matching fluid, negligible light
loss was observed across the array and the authors predicted that greater than 34% of the incident
laser power would be available in the 500th capillary of a hypothetical 500-capillary array [122].
Detection was made above the plane of the capillary array with a CCD camera. A 514-nm notch filter
was used to reduce scattered light while an image splitter, consisting of a 610- or 630-nm long pass
filter and a quartz plate, allowed two-color imaging onto two regions of the CCD chip. Detection
limits of 90 pM were reported from long injections of fluorescein dissolved in run buffer, to eliminate
stacking. Two-color DNA sequencing was demonstrated in this work.
Shutter
Pin Acromat
hole lens Mirror
Photomul tiplier
Bandpass
filter Dichroic beam
splitter
Aperture Laser
Filter Quarter wave plate

amplifier

Computer
Trigger
Power supply
Motor
FIGURE 16.14 Illustration showing a 1000 capillary, four-color detection rotary capillary array electrophore-
sis scanner device. (From Mathies and Scherer, US Patent 7,090,758.)
Quesada and Zhang [123] and Quesada et al. [124] performed detailed ray tracing and also
demonstrated success using a side illumination scheme. A planar 12 capillary array was created and
the windowed region was placed in index-matching fluid. The output of a 488 nm laser was split into
two fiber-optic transmitters and delivered to either end of a planar array, thus reducing variation in
laser intensity across the array. Emission was collected with 12 optical fibers, each placed against one
capillary detection window, normal to the direction of laser excitation. The emission was dispersed
using a spectrograph and imaged onto a CCD camera, thus providing spatial and four-color spectral
information. Cross talk between adjacent capillaries was estimated to be ∼1–2% while a detection
limit of ∼3.7 pM at a SNR of 2, derived from flowing stream measurements using fluorescein, was
reported. Simultaneous four-color dye-terminator sequencing for all 12 capillaries was demonstrated
in this work [124]. Since the number of capillaries that could be used was generally limited by laser
attenuation due to beam divergence, glass rods were placed between capillaries by Anazawa et al.
[125]. This allowed them to achieve detection limits of <1 pM of flowing fluorescein solution in the
45th capillary, furthest away from the illumination source.
Recently, Roeraade and coworkers [126,127] developed a detection scheme in which each capil-
lary acted as a waveguide to direct the emitted fluorescence to the capillary tip, where it was collected
with a CCD camera. Strip illumination of up to 91 capillaries was made in a region where the cap-
illaries are arranged in a planar array; these capillaries were externally coated with a fluoropolymer
that did not require removal for use in LIF detection. Approximately 15 cm downstream from the
excitation region, the capillaries were cast into a two-dimensional hexagonal array with black epoxy
and placed into a container consisting of buffer and the electrode. Fluorescence emitted from the
ends of the capillaries transmitted through the glass wall of the container was collimated, passed
through two bandpass filters to eliminate 488 nm scattered light, and was then dispersed with a prism
onto a cooled-CCD camera. This arrangement yielded spatial and spectral information about the flu-
orescence emitted from the capillary outlets. Cross talk between capillaries, arising from emission
from one capillary being captured and transmitted by a second capillary, was found to be less than
0.4%. A detection limit of 2.7 pM fluorescein measured in flowing stream mode was reported, but
this sensitivity measurement was made without use of the prism to disperse the light [126].
The parabolic reflector concept developed by Pentoney et al. [28] was extended to a scanning
eight-capillary system, which was commercialized by Beckman Coulter in their family of four-color,
eight-capillary DNA analysis systems. In this work, two diode lasers, also used in References 128
and 129, were alternately modulated and the beams directed to a precision rotating mirror positioned
on a galvanometer. This mirror sequentially directed the focused laser beams into each capillary
window in the planar array. The windows were centered about the focus of a parabolic reflector,
which collimated and directed the emission though a four-quadrant spinning filter wheel, a dual-
notch filter (designed to reduce scatter from each laser) and then through a focusing lens and onto
a PMT (Figure 16.15). The mouth of the parabolic reflector again had a black bar placed across
it to block the intense plane of laser scatter. The spinning filter wheel was synchronized with the
modulation of the two diode lasers and the galvo scanner such that when the first laser was striking
the capillary, the first two filter quadrants passed in front of the detector. Similar interrogations of
the capillaries were made for laser 2 and filter wheel quadrants 3 and 4. To ensure high sensitivity,
the diode lasers were optically filtered to remove low level emission (“diode wings”) overlapping
those of the detection wavelengths. Detection limits below 10 pM at a SNR of 2 were determined
from flowing stream measurements using proprietary dye-labeled terminator solutions (unpublished
data, Beckman Coulter).
An interesting aspect of this step-scanned system is the automated optical alignment feature. To
maximize duty cycle, the galvo mirror rapidly advances from one capillary bore to the next and
then pauses during data acquisition. In order to predetermine the location of the capillary bores,
an automated scan was made before every run. In this prescan, which was completed in less than
2 min, the first laser was scanned across the entire region where the array was believed to reside. An
autoalignment photodiode, placed behind the array, facing the laser but off axis to the laser, detected
light refracted from the capillaries and an auto-alignment spectrum was obtained, such as the one
displayed in Figure 16.16. The peaks seen in this spectrum contain the information required to direct
the galvo to position the beams at the capillary bore centers. This process was repeated for the second
laser, yielding a second set of eight peaks corresponding to the galvo positions required to send the
second laser to each of the capillary bores. The 16 positions (eight-capillary positions for each of
the two lasers) were saved in a lookup table and used during the separation to place the first laser in
the first capillary, the second laser in the first capillary, the first laser in the second capillary, and so
on. Experience has shown that the positions of the capillary bores with respect to the laser positions
do not change during the course of the run.
16.2.3.3 Sheath-Flow Detection

Dovichi and coworkers [29,70,130] developed a high-sensitivity fluorescence detection system
for DNA sequencing based on a detection scheme borrowed from flow cytometry and previously
Dual laser module

Laser line filters
Galvo scan mirror

Diode lasers
(Modulated)
Plano convex lens
Dual block filter

Parabolic reflector
Eight capillary PMT
array
Filter wheel
Autoalignment photodiode
Biconvex lens
96-well Microtiter Plate
FIGURE 16.15 Illustration of the galvanometer scanner based, eight-capillary DNA sequencer optics
employed in the Beckman Coulter DNA analysis system.
0.5
Photodiode response
0.4
0.3
0.2
0.1
0
0 50 100 150 200 250
Galvo position
FIGURE 16.16 Capillary auto-alignment spectrum used to determine the capillary bore center positions.
developed by the same group for CE separations of labeled amino acids [131]. To minimize or
eliminate background scattered light arising from the excitation source striking the curved capillary
surface, the separated DNA fragments were carried away from the capillary outlet before detection.
This was accomplished by placing the end of a 1 m × 50 µm i.d. capillary in a flow cytometer cell
(Figure 16.17): the labeled sequencing fragments eluting from the end of the capillary were entrained
Electrophoresis
capillary Quartz
window
Sheath
inlet Laser
beam
Microscope
objective
Spectral
Quartz
filters
window
Pinhole
Laser
beam PMT
Sample
stream
Side view of flow chamber Top view of optical train
FIGURE 16.17 Postelectrophoresis sheath-flow detection of DNA fragments. (Reprinted with permission
from Zhang, J.Z., et al., Clin. Chem., 37/9, 1942, 1991.)
and carried down the flowcell by sheath fluid, flowing at 0.16 mL/h, into a square detection region.
The sample stream was illuminated with a focused laser and fluorescence from the labeled sequenc-
ing products was collected, filtered to eliminate potentially interfering Rayleigh and Raman scatter
and passed on to a PMT [17,29,30,70,73,130,132–134]. Two-color detection was achieved either
using a dichroic beam splitter to separate the emission from two fluors [29,30,73,134] or by detec-
tion on both sides of the sheath-flow cuvette [30], facilitating two-label peak-height encoded DNA
sequencing. A detection limit of less than 60 molecules of Rhodamine 6G injected into the system
was reported [133].
Four-color sequencing in a single capillary using this sheath-flow concept was also demonstrated
[17]. Two alternately modulated lasers, an argon-ion laser and a green helium-neon (HeNe) laser,
were combined using a dichroic filter. The collinear beams were focused onto the sheath flow entrain-
ing the separated four-color primer-labeled DNA fragments. To spectrally resolve the emission, a
four-sector filter wheel spinning at 2 Hz was placed between the flow cell and a microscope objective,
which in turn coupled the emission into an optical fiber mated to an avalanche-photodiode operated in
single-photon-counting mode. With the argon-ion laser on, the first two quadrants of the filter wheel,
containing bandpass filters centered at 540 nm and 560 nm, respectively, passed in front of the
microscope objective. Similarly, excitation with the green HeNe laser was synchronized to detection
through the 580 nm and 610 nm quadrants of the spinning filter wheel. Read-lengths of 640 bases in
2 h were reported for fluorescently labeled primers [17]. A multicolumn approach soon followed [22]
and with a planar array of five capillaries, the lasers were focused to illuminate across the outlets of
the capillary array, exciting all five sheath-entrained samples in the cuvette simultaneously approx-
imately 100 µm below the ends of the capillaries (Figure 16.18). Emission was collected normal to
the direction of the sheath flow and excitation laser direction using a microscope objective (20×,
ries
Capilla
Sheath
inlet
Quartz
cuvette
r
Lase
Sample be a m
streams
Microscope
objective
GRIN
e
lens
Waste
Fiber
des
optics odio
h e phot ounting
nc nc
Avala le photo s)
(sing module
FIGURE 16.18 A five-capillary sheath-flow CE/LIF detection scheme. (Reprinted from Zhang, J., et al.,
Nucleic Acid Res., 27, e36, 1999. With permission from Oxford University Press.)
0.5 NA) and was passed through a spinning filter wheel. At the image plane, five gradient refractive
index lenses, each coupled to a fiber optic, brought the emission from the five spots onto five dedi-
cated avalanche photodiodes [22]. The implementation of a micromachined sheath-flow cuvette for
capillary alignment extended multiplexing to 16 capillaries [135] while a two-dimensional capillary
array scheme using 32 capillaries, with illumination by a “sheet” of laser light and detection in the
sheath stream at the exit tips of the capillaries using a CCD camera, suggested feasible scaling to
“several thousand capillaries” [136].
Kambara and Takahashi also employed the sheath-flow technique for the interrogation of 20
capillaries using a CCD camera [137]. In that work, the 20 capillaries were aligned into a linear
array, with the outlet end of all capillaries engulfed by a sheathing fluid, in this case, a TBE buffer.
No flow cell was used in this work. A laser was directed through the sheath fluid carrying the
sample, normal to the direction of the sheath flow, allowing sample from all 20 capillaries to be
illuminated simultaneously (Figure 16.19). A cylindrical lens followed by a prism was used to
disperse the fluorescence from all 20 capillary bores onto a CCD camera. One axis of the CCD
image yielded spatial information about the interrogated capillaries and the other axis of the CCD
yielded spectral information. No moving parts was required in this approach. In addition, since
excitation and detection for all capillaries and fluors occur simultaneously, no time sharing was
required, improving duty cycle and greatly increasing the capability of this system for multiplexing.
PE Biosystems Inc. (now part of Applera) has commercialized a similar sheath-flow design on several
of their high-throughput DNA sequencing systems [6,56,127].
To buffer reservoir
6 mW, 594 nm He-Ne laser
Twenty 100 µm × 200 µm

gel-filled capillaries
Cylindrical lenses
CCD camera
Gel capillaries
Sample carried
by sheath fluid
To buffer reservoir
Laser
Open capillaries
FIGURE 16.19 A twenty-capillary sheath-flow CE/LIF system. Detection is accomplished in the gap region
shown in the inset. (Adapted by permission from Macmillan Publishers Ltd: Kambara, H. and Takahashi, S.,
Nature, 361, 565, 1993. Copyright 1993.)
16.2.3.4 Other Detection Schemes Explored

The detection methods thus far described focus on measuring the intensity of fluorophores at one
to four wavelengths. For multicolor work, an important requirement is that the fluorophores are
spectroscopically unique. To obtain spectroscopic information, the emission is either monitored
using one time-shared detector, split among several detectors, or dispersed with a grating or prism,
all resulting in reduced light levels and lower SNR.
One alternative method is to use differences in fluorescence lifetime as the metric to discrimi-
nate between the fluor-labels [138,139]. Several potential benefits of measuring lifetimes for DNA
sequencing have been described [140,141]. Lifetimes may be accurately determined in the absence
of spectral discrimination; thus, discrete lifetimes may be determined for mixtures of fluors coeluting
in a single peak. Lifetimes are also independent of intensity even at concentrations where relatively
poor SNR may lead to an erroneous fluorescence intensity-based call. In addition, since scattered
light has an effective lifetime of zero, scatter may be temporally rejected and is thus less of an issue.
However, this method requires a family of fluors in which the lifetimes are resolvable [141–143]
and the use of relatively complex electronics [141].
There are two methods that have been used to determine fluorescence lifetimes in DNA sequenc-
ing. In the first method, known as the frequency-domain or phase-modulation method, the excitation
beam is intensity modulated. The a.c. portion of the resulting emission is phase-shifted relative to the
laser modulation; this phase-shift contains information about the fluorescence lifetime, or lifetimes
if more than one fluor is present [140]. McGown and coworkers [144,145] used this method for
four-color sequencing. In that work, 488 nm or 514 nm laser light was electronically modulated
with a Pockels cell before being focused onto a capillary column. Detection, made normal to the
laser direction, was optically filtered to reduce laser scatter and was focused onto the detector of a
fluorescence lifetime instrument (Model 4850 MHF from Spectronics Instruments, Rochester, NY).
Fluorescence lifetimes of 1.7, 2.5, 2.9, or 3.5 ns were measured and assigned to primers labeled with
either CY3, the energy transfer dye fluorescein-dTMR, rhodamine green, or BODIPY-fluorescein,
respectively. Recent advances using acridone dyes with more resolvable lifetimes of 4, 6, 11, and 14
ns have been reported [143].
The second method of determining fluorescence lifetime is to modulate the excitation source
with a very short pulse width and measure the decaying fluorescent emission with high-temporal
resolution when the laser is off. Soper and Mattingly [146] constructed a near-infrared detection
system with picosecond resolution. In this study, an argon-ion pumped Ti:sapphire laser generated
210 femtosecond excitation light pulses at 76 MHz. They found a strong dependence of fluores-
cence lifetime on sieving matrix material. Detection of separated dye-primer labeled fragments was
demonstrated [147]. Wolfrum and coworkers [138] used a single 630 nm diode laser, pulsed at 22
MHz with 500 picosecond on durations, to excite a family of fluor-labeled primers. The lifetimes of
the four fluorophores studied were 3.7, 2.9, 2.4, and 1.6 ns. Emission, filtered with a dichroic lens
(required in their confocal illumination/detection setup) and a 675 nm bandpass filter, was moni-
tored using an avalanche photodiode. This work was expanded by placing up to 16 capillaries onto a
translation stage and scanning them under the confocal excitation/detection system [23]. In all cases,
relatively complex curve fitting of the fluorescence decay was required to extract the lifetime values.
Detection limits were not reported for any of the fluorescence lifetime-based capillary sequenc-
ing studies probably because that was an area of difficulty relative to conventional fluorescence
approaches.
Another detection scheme, studied by Gorfinkel and coworkers [148,149] with capillary-based
DNA sequencing in mind, was based on single-photon detection. In this approach, the output beams
from four lasers were combined in a single optical fiber and used to excite the labeled fragments
in the CEQ DNA test sample (Beckman Coulter) as they migrated through the detection window
of a capillary. Each laser, selected to preferentially excite one of the four BCI fluors at 635, 675,
750, and 810 nm, was modulated at a different frequency, in the range 1–2 Hz. Collection was
made via a fiber receiver connected to a single-photon PMT detector. Four stacked notch filters were
used to eliminate scatter from each of the lasers. Since the four dye-matched lasers were modulated
at different frequencies, the fluorescence signal arising from each fluor was also modulated at a
different frequency. A Fourier spectrum contained a peak or peaks corresponding to the fluorescence
frequency of the fluor(s) present. Cross talk, corresponding to the emission from a fluor generated
by excitation from a second laser, may be determined by running each fluor independently and
determining the relative Fourier amplitude for each of the lasers. Although demonstrated only for
sequencing in a single capillary, a multiplexed scheme to monitor separations in 32 capillaries was
also described [149].
Metzker and coworkers [150] developed a “color-blind fluorescence detection” system using
four fluorophores. In this system, each of the four fluorophores was stimulated primarily by only
one of four lasers. The four lasers, at 399, 488, 594, and 685 nm, were modulated using mechanical
shutters such that only one laser was striking the capillary at any time. These were used to excite
Alexa Fluor 405, BODIPY-FL, 6-ROX, and Cy5.5, respectively. Emission was collected normal to
the direction of excitation and was then transmitted through a long pass and several notch filters
to block light scattered from any of the lasers before being focused onto a PMT. Since each of the
four fluorophores is primarily stimulated by only one laser, fluorescence detected at the PMT was
assigned to the DNA fragment labeled with the fluorophore primarily excited by the laser currently
striking the capillary. In general, cross talk observed between lasers and fluorophores was in the range
of 2–4%. However, Cy5.5 excited by 594 nm yielded an emission that was 20% of the emission
generated upon excitation with 685 nm light. Mobility correction was required before good DNA
sequence could be called.
An interesting nonoptical detection scheme for CE-based DNA sequencing was proposed by
Brennan et al. [151]. In this approach to sequencing, capillary gel electrophoresis was coupled to
mass spectrometric detection. The idea was based on the use of four different sulfur isotopes, having
masses of 32, 33, 34, and 36 to establish the terminal base identity of Sanger–Coulson sequencing
fragments eluting from the capillary. The labels were incorporated in the form of thiophosphate
analogues. Described advantages included the possibility of internal labeling, which would produce
an increase in signal with increasing fragment length, and the elimination of spectral overlap issues
that exist with fluorescent labeling schemes. The novel proposition involved a CE/MS interface in
which capillary effluent droplets were created using a piezo-electric dispenser similar to those used
in ink-jet printers. The small, picoliter volume droplets were to be combusted and transferred to a
mass dedicated spectrophotometer tuned over the small mass range required to differentiate the four
sulfur isotopes.
16.2.3.5 Integrated Sample Processing and Detection

The systems described thus far addressed only a portion of the individual tasks required to elucidate
DNA sequence. A more efficient process would fully integrate the tasks of DNA extraction, purifi-
cation, template preparation, and amplification with sequencing reaction chemistry, post reaction
cleanup, separation, and detection [152]. Several groups have explored partially integrated systems
intended to streamline various steps in this process [152–156].
Swerdlow et al. [153] developed a system for DNA sequencing that integrated the cycle sequenc-
ing reaction, post reaction sample cleanup, sample injection, CE separation, and LIF detection. A
schematic representation of the system is displayed in Figure 16.20. Cycle sequencing reaction com-
ponents and sample were loaded by syringe into a temperature programmable rotary valve. Once the
cycle sequencing reaction was completed, the reaction product was directed to a gel filtration HPLC
column for cleanup. Their system allowed “heart cutting” of the product peak that eluted from the
gel filtration column and this purified aliquot of the product peak was then electrokinetically injected
into a gel-filled capillary using the “Tee” injection scheme depicted in Figure 16.20. After a 10 s
injection, pressure drove a stream of TBE buffer through valve 2 and into the “T” in order to chase
away any remaining sample. Once the “T” had been filled with TBE buffer, electrophoresis was
initiated. On-column detection was made using a single PMT to monitor one-color fluorescence
induced by 532 nm laser excitation. The entire process was completed in 90 min [153].
Yeung and coworkers devised a similar scheme to integrate thermal cycling and purification
followed by injection, separation, and detection, first in a single capillary [157] and then in eight
capillaries [154]. In the multicapillary work, multiplexed freeze/thaw values (MFTV) were used to
simplify the manipulation of samples during the thermal cycling process. In the MFTV, an array
of eight capillaries was threaded through two stainless steel tubes; the length of capillary between
the tubes was placed in a thermalcycler and was used as the reaction vessels. When liquid nitrogen
was flowed through the two stainless steel tubes, the liquid in the capillary portions within the
steel tubes froze, thus “closing” the valves and forming an immobile volume of liquid between
the two frozen plugs. Passage of warm air through the steel tubes melted the plugs, opening the
“valves.” Figure 16.21 displays the operation of the MFTV for a single capillary. A pump was used
to draw thermalcycling reagents from a well in a microtiter plate into the portion of the capillary
within the thermalcycler. With valves 1 and 2 in the closed position, the cycle sequencing reaction
was performed, thereby creating the labeled DNA fragments (Figure 16.21a). With 1 and 2 in
the open position, and valve 3 in the closed position, the pump drew the sample above the “T”
(Figure 16.21b). The sample was then pushed into the size exclusion column for sample purification
by closing valves 1 and 2 and opening valve 3 (Figure 16.21c). After this, sample purification and
capillary loading/separation occurred in much the same manner as described by Swerdlow et al. [153].
Detection of separated DNA fragments was made via side illumination of all eight capillaries via the
method described above [122]. This method was later expanded to include sequence determination
directly from single bacterial colonies [155].
Positive buffer
Pressure chamber and
pump
Valve 4
Low salt HPLC
buffer pump Fluorescence
Pressure detection
Valve 1
Valve 3
Polymer filled
Waste capillary
TBE
Sample chase
Valve 2
Thermal cycler
Waste
Gel filtration
HPLC column
Waste
Negative buffer
chamber
FIGURE 16.20 Integrated cycle sequencing reaction, sample cleanup, separation, and detection system.
(Reprinted with permission from Swerdlow, H., et al., Anal. Chem., 69, 848, 1997. Copyright 1997 American
Chemical Society.)
16.2.4 SEPARATION METHODS

16.2.4.1 Sample Injection
In terms of sample loading, capillary-based DNA sequencing systems have a fairly narrow window
of stable operation. This window ranges on the low injection end, from the minimum amount of
sample required to produce sufficient signal to reliably call bases throughout the sequencing run
(note that signal strength generally declines with increasing base number) to the high injection end
of the window, where sample overloading results in resolution losses and/or current instabilities.
Several complicating factors can further reduce the width of this stability window including residual
template DNA, protein contaminants in the sample, excess salts and unincorporated nucleotides, and
the unintentional introduction of air into the gel upon executing electrokinetic injection. The ideal
sample loading process, speaking strictly from a separations perspective, would introduce nothing
but labeled sequencing fragments onto the gel, and those fragments would be introduced over a
column length that contributed nothing to total peak variance at the detector. The fragments would
furthermore be introduced onto the gel at concentrations that were electrically transparent relative to
the conductive gel medium. Obviously, practical considerations having to do with post sequencing
reaction cleanup and low level fluorescence detection prevent us from completely realizing this ideal
injection goal.
In 1998, Karger and coworkers [158,159] published a two-part study on the role of sample
matrix components and post reaction sequencing product cleanup . In this study, the authors used
ultra-filtration to remove template DNA and gel filtration to reduce salt concentration in the sample
To pump
Valve 3
Buffer
Sample
To SEC
column
Valve 2
Thermal cycler
Valve 1
(a) (b) (c)
FIGURE 16.21 Multiplexed freeze/thaw valves used to isolate sample during thermalcycling. (Reprinted with
permission from Tan, H. and Yeung, E.S., Anal. Chem., 70, 4044, 1998. Copyright 1998 American Chemical
Society.)
loading matrix. The cleanup procedure was reported to be very reproducible, resulted in a 10- to
50-fold increase in the amount of sequencing fragments loaded onto the gel, and facilitated the
realization of read-lengths in excess of 1000 bases.
Manufacturers of commercial systems generally recommend a spin-column cleanup method or
an ethanol precipitation/washing cleanup protocol that must then be followed by suspension of the
lyophilized DNA pellet in a sample loading matrix designed to optimize or enhance sample loading.
Note that separating DNA sequencing fragments generated from polymerase chain reaction (PCR)
amplicons is generally less problematic than when larger template DNAis sequenced, thus supporting
the idea that the presence of template DNA in the sample often leads to poor performance.
16.2.4.2 Separation Parameters

Several groups have searched for optimum run parameter settings in an attempt to maximize sequenc-
ing rates and read-lengths in both fixed and replaceable gels [15,18,21,24,50,67,73,80,160]. Rather
than provide an exhaustive review of those studies here, we describe a few metrics that we have
found useful in the optimization of separation conditions. The first of these is the “cross-over plot”
in which peak width and adjacent peak spacing are plotted against base number [55]. Figure 16.22
shows cross-over plots for the separation of dye-terminator labeled sequencing fragments run at three
different field strengths using a CEQ system (Beckman Coulter) and a linear polyacrylamide gel. A
qualitative resolution limit is visible in these plots as the point where the two curves cross. At that
point in the separation, the peak-to-peak spacing has dropped to the level where it is equal to the peak
widths. Beyond this intersection point, discerning individual bands becomes increasingly difficult.
Note that the cross-over point in these figures is observed to occur at decreasing base number as the
field strength is increased and so any increase in speed realized by using higher voltage is offset by
a loss in read-length.
1.2
2.5 kV separation voltage
1 75.6 V/cm
Length (mm) 0.8
0.6
0.4
0.2
0
0 100 200 300 400 500 600 700 800 900
Nucleotide length (number of bases)
1.2
1 116 V/cm
0.8
Length (mm)
0.6
0.4
0.2
0
0 100 200 300 400 500 600 700 800 900
1.2
1 197 V/cm
0.8
Length (mm)
0.6
0.4
0.2
0
0 100 200 300 400 500 600 700 800 900
FIGURE 16.22 Peak-to-peak spacing (thick curves) and full-width at half maximum (FWHM) values (thin
curves) as a function of nucleotide length for several run voltages.
Another useful, and perhaps more significant, metric to monitor as run conditions are varied is
the “98% cutoff” or “length of read” number. This refers to the length of read achieved before the
cumulative errors reach a value equal to or greater than a prespecified value of 2%.
A third metric that has proven useful in the optimization of run parameters and gel formulations
is the “first base called” (FBC). This is the first position in the sequence that is correctly called and it
is located typically 5–15 bases from the 3 end of the primer used in the extension reaction. Factors
affecting the FBC include the gels ability to sieve over a wide range of sizes, including the smaller
extension products and congestion that often occurs early in the electropherogram due to coelution
of unincorporated terminators, salts, and unextended primer.
Basic CE-based sequencing process flow
Select system
run parameters
Load MTP with

suspended samples
Purge/refill
capillary array
Electrokinetic
sample injection
High voltage separation

& data acquisition
Base calling
FIGURE 16.23 High-level graphical depiction of the steps involved in CE/LIF-based DNA sequence
determination.
A fourth metric that we have found useful is to track resolution and/or base-calling accuracy over
short stretches of sequence “known to be difficult” (KD stretches). This term refers, for example, to
regions in electropherograms where anomalous migration occurs due to either secondary structure
or the presence of salt fronts.
Figure 16.23 illustrates the steps involved in executing a typical CE-based DNA sequencing
separation using one of today’s automated sequencers. Run parameters such as injection voltage and
duration, capillary temperature, denaturation time and temperature, separation voltage and duration,
and sample well assignments are pre-programmed using simple graphical user interfaces. After
cleanup, sequencing reaction samples are suspended in a suitable loading matrix and are transferred
to appropriate wells in a 384- or 96-well microtiter plate. Fresh gel is automatically introduced into the
capillaries of the array, optical alignment is verified, samples are denatured and then electrokinetically
loaded into the capillaries. Recommended injection parameters vary somewhat but generally involve
loading the samples onto the gel at low voltage (∼60 V/cm) for approximately 30 s. Separation is
initiated, often with the voltage ramped up over a prespecified time interval, to a final separation
voltage on the order of 100–200 V/cm. Higher field strengths will result in faster separations but often
at the cost of shorter read-lengths and/or reduced system stability. Multicolor detector output signals
are acquired at data rates suitable to accurately represent the corresponding peak profiles (generally
in the range of 2–4 Hz). Individual capillary current profiles, which are a useful diagnostic, are also
often collected throughout the injection and separation process.
16.2.5 ALGORITHMS
16.2.5.1 Basic Sequence Calling Steps
The importance and value of base-calling algorithms that have been developed for automated DNA
sequencing cannot be overstated. Figure 16.24 graphically depicts the major steps involved in the
analysis of raw detector output signals from today’s capillary-based sequencers. Raw LIF signals are
collected at a data acquisition rate sufficient to adequately represent the true concentration profiles
of the narrow bands migrating through the beam. The raw data is generally smoothed to reduce noise
and is baseline corrected to reduce drift. Color separation refers to dealing with the fact that fluo-
rescence emissions typically exhibit some degree of overlap in the spectral channels monitored for
Base calling algorithm process flow
Data smoothing
Baseline correction
and color separation
Normalize peak heights,

mobility-shift correction
Convert elution time to

approximate base number
Base calling
Determine quality
values
FIGURE 16.24 Graphical depiction of the steps involved in analyzing CE/LIF sequencing data for base
calling.
each of the four fluorescent labels. Manufacturers of these systems understand the degree of spectral
overlap that exists in these measurements and are able to mathematically eliminate this concern so
that an unambiguous base assignment is made at each position. Peak-height uniformity has seen
improvements over the years, owing both to improved sequencing reaction chemistry (e.g., the role
played by pyrophosphatase in References 26 and 161) and algorithm-based mathematical normaliza-
tion for differences in response across the dye set and for signal decay with increasing base number.
The four different dyes attached to the terminators can also have an impact on fragment mobility and
thus can complicate the clear size ordering assumed to be provided by the gel matrix. This issue has
also been minimized through both chemistry improvements (similarity of dye structures, common
degree of sulfonation, etc.) and mathematical algorithm corrections.
16.2.5.2 Quality Values

Phred is a base-calling algorithm, developed by Ewing and Green [162,163] that reads DNA sequence
data files in a variety of formats, makes base call assignments, and also associates a highly accurate,
base-specific quality score with each base call. The reported Phred quality scores are logarithmically
related to error probabilities as shown in Table 16.2. Manufacturers of commercial sequencing
systems also incorporate similar base-specific quality scores into their base-calling algorithms.
16.3 THE NEXT GENERATION: SEPARATIONS IN

MICROFLUIDIC SYSTEMS
With replaceable gel CE now having effectively displaced all slab-based DNAsequencing, many have
speculated that the next generation of fully automated, high-throughput DNA sequencing systems
will be based on the use of microfluidic technologies such as compact plates or discs that are channel
etched and micro-plumbed at high density [7,8,164–171]. Using technologies such as photo and
chemical etching, originally developed for the microelectronics industry, features much smaller than
those required for conventional CE may be created directly in fused silica or other materials. This
technology could allow for significant miniaturization of multicolumn systems, more control options
TABLE 16.2
Error Probability for Various Phred Quality Scores
Phred Quality Score Probability of Error at That Base Call Basecall Accuracy at That Base (%)
10 1 in 10 90.0000
20 1 in 100 99.0000
30 1 in 1000 99.9000
40 1 in 10,000 99.9900
50 1 in 100,000 99.9990
60 1 in 1,000,000 99.9999
for sample introduction, and simplification of fluid handling in even more highly paralleled systems.
The ability to reliably introduce much smaller sample plugs into the separation channels (reduced
injection variance) may one day allow even faster separations to be run using significantly shorter
separation channels. As demonstrated below, plug lengths as short as 30 µm may be introduced in
microfluidic channels. Low signal-to-noise issues anticipated with reduced sample loading may be
resolved using innovative solutions such as brighter fluorophores and increasing channel depths in
the detection region.
The first DNA separation with single base resolution performed in a microfluidic device was
reported by Effenhauser et al. [172]. Using standard photolithographic techniques channels were
etched into a polished glass plate over which a second plate containing micromachined holes was
thermally bonded (Figure 16.25). Both thin channels (50 µm wide and 12 µm deep) and broad
channels (1000 µm wide and 12 µm deep) were created in this study. Before use, all the channels
were filled with a 10% AA solution that was allowed to polymerize in situ. Pipette tips were glued to
the holes in the covering plate and acted as reservoirs. Referring to Figure 16.25, sample was injected
from reservoir 1 into the main separation channel by applying a field across 1 and 5. The loaded
sample plug was 150 µm long with a volume of 90 pL [164]. Separation of this plug was achieved
by switching the field to ports 3 and 7. LIF detection of the fluorescein-labeled fragments was made
3.8 cm downstream of the injection manifold and single base resolution of fragments 10–25 bases
long was achieved in less than 45 s [172].
Mathies and coworkers [173] developed a similar system to separate double-stranded DNA
restriction fragment digests. Standard photolithography techniques were also employed here [174].
Specifically, the channels were etched into a glass slide by first covering the slide with a photoresist
film and then transferring the channel pattern to the film by exposing UV light through a patterned
mask. The photoresist film exposed to the UV light dissolved while the remaining film hardened.
The exposed glass region was then chemically etched with solutions of NH4 F/HF to form the appro-
priate channels and the remaining photoresist film was removed with H2 SO4 /H2 O2 . A glass plate
with access holes drilled into it was then thermally bonded to the etched plate. Channel widths of
30–120 µm were created. To prepare the channels for electrophoresis and eliminate EOF, the surfaces
were treated using a modified version of the Hjertén protocol. A general scheme for sample injection
and separation is displayed in Figure 16.26. Low-viscosity HEC separation matrix was placed in
the separation channel by applying vacuum to opening 4. The cross channel regions were filled
with TBE buffer. Electrodes, buffers, and sample were placed in reservoirs numbered 1 through 4.
Two injection types were explored. For “stack” injection, sample was placed in 3 while a field was
applied across 3 and 4. For “plug” injection, sample was placed in 3 and a field was applied across
1 and 3 to move sample across the intersecting separation channel. In either case, once sample was
injected, separation occurred by applying the field across positions 2 and 4. Initially, single-color
detection was made using an optical arrangement similar to that described for multicapillary work,
as illustrated in Figures 16.13 and 16.14.
1 8
5 6 7
1 Sample
Buffer 3 7 Waste
5 Injection waste
FIGURE 16.25 Microfluidic design for sample injection from a twin-T channel geometry. (Reprinted with
permission from Effenhauser, C.S., et al., Anal. Chem., 66, 2949, 1994. Copyright 1994 American Chemical
Society.)
2 4
Separation region
3
Cross channel region
Injection Separation
“Stack” : + – +
Buffer –
+
Sample
“Plug”: – +
FIGURE 16.26 Schematic diagram describing two different modes of injection in a microfabricated cross
channel format. (Reprinted with permission from Woolley, A.T. and Mathies, R.A., Proc. Natl Acad. Sci. USA,
91, 11348, 1994. Copyright 1994 National Academy of Sciences, USA.)
The first reported DNA sequencing separation with this system utilized plug injection of primer-
labeled fragments onto an LPA matrix in a channel having an effective separation length of 3.5 cm and
resulted in a four-color basecall of ∼150 bases at 97% accuracy in 540 s [175]. Further work using
a replaceable 4% LPA matrix, delivered via a high-pressure loader [176], and the “twin-T” injection
scheme used earlier by Effenhauser [172], yielded read-lengths of 500 bases at 99.4% accuracy
in less than 20 min [177]. Microfluidic chips containing 12 [178], 16 [179], 96 [180–183], 188
[184], and 384 [185,186] separation channels have since been reported. Linear 96-channel systems
were constructed [179,180] as were novel radial microplates for 96- and 384-channel systems [181–
183,185,186], although no sequencing was demonstrated with the 384-channel system. Figure 16.27
displays a system with 384 channels etched into a 20 cm diameter wafer; lanes were ∼60 µm wide
and 30 µm deep, with an effective separation length of 8.0 cm. A multicapillary array loader was
created to transfer sample (via pressure) from a microtiter plate to the radial microplate [181]. For
detection, a rotary scanning confocal fluorescence detector was created [181,182]. Similar to the
linear scan head developed to interrogate a multicapillary array (see Figure 16.13), the objective is
placed on a rotating scan head (Figure 16.28) [181]. This system was capable of collecting data at
200 revolutions per second, with 5000 data points collected per revolution for each color. Given that
the scanning diameter was 2 cm, a spatial resolution of ∼12.6 µm was achieved. A detection limit
of ∼1 pM at a SNR of 2 was reported using this system [182]. Simultaneous four-color sequencing
of primer labeled DNA fragments yielded an average of 400 high-quality bases in each of the 95
channels (one channel failed due to a defect in the photolithography) [183].
Ehrlich and coworkers [187] performed a systematic study of LPA polymer concentration and
composition, device temperature, effective channel length, and electric field strength for microfluidic-
based DNA sequencing platforms. They determined that separations performed in an 11.5 cm channel
(the longest used in this work) at 50◦ C using a field strength of 125 V/cm and a separation matrix
consisting of 3.0% (w/w) 10 MDa plus 1.0% (w/w) 50 kDa LPA were most optimal, yielding a
read-length of 640 bases at 98.5% accuracy in ∼30 min for primer-labeled DNA fragments. Very
long read-lengths required an extension of the column length and separations in 40 cm channels
(40 µm deep and 90 µm wide at the top) yielded reads >800 bases at 98% accuracy in 80 min at
50◦ C [188]. Comparable results required 180 min of run time on the MegaBACE 1000 system (GE
Healthcare) [188,189].
In addition to the advantages of performing separations in microfluidic channels presented at the
beginning of this section, several others are now evident. For example, in all microfluidic systems
in which the sample is loaded through a gel-filled “T” region before separation, “preseparation” is
occurring. With judicious injection timing, unwanted materials contained in the sample solution may
be prevented from entering the separation zone, thereby increasing separation quality [190,191]. For
example, fragments contained in the sample that are larger than the upper resolving limit of the gel
normally coelute as a large unresolved “blob” at the end of the electropherogram. This signal can be
completely eliminated from the analysis by running sample through the T junction just long enough
to allow the largest resolvable fragments to transit the intersection. Once the high-voltage contacts
are switched from injection to separation positions, only the resolvable fraction of the sample will
be directed on for separation and analysis.
In addition, such injection methods have also been reported to enhance loading of larger fragments
relative to direct electrokinetic sample injection onto a gel-filled capillary [179,192]. Finally, because
there is no analogy to the polyimide cladding protecting capillaries in microfluidic channels, the entire
chip process may be monitored with a CCD camera, yielding information about sample injection
and providing flexibility in detection [193,194].
Ehrlich and coworkers [188] found that long channel lengths yielded longer read-lengths, with
a channel 40 cm long yielding reads of greater than 800 bases at 98% accuracy in 80 min. However,
long straight channels are difficult to create primarily because of the special equipment needed to
etch large plates [184,191]. To remedy this situation, some groups have created channels with several
turns (i.e., serpentine channels) to increase column length. However, dispersion effects degraded
(a)
Cathode
0
(b) (c)
S2 S3
8.0 cm 5
S1 S4
Detection point 10
(mm)
FIGURE 16.27 (a) Layout of 384-channel microfluidic electrophoresis device. (b) Expanded view of a sin-
gle quartet of channels with their injectors. (c) Expanded view of a quartet of individual sample reservoirs.
(Reprinted with permission from Emrich, C.A., et al., Anal. Chem., 74, 5076, 2002. Copyright 2002 American
Chemical Society.)
the individual bands because the path-length within the channel varied depending on whether the
molecules traveled along the interior or exterior radius [195,196]. New designs implemented with
unique channel geometries in the turn regions have minimized this problem [197,198].
To circumvent the drawbacks associated with the etching of long straight channels on large glass
plates, Liu [191] created a hybrid capillary/microchip design. In his work, twin-T injectors with round
channels as well as round capillary connection channels were created using very narrow line-width
isotropic etching (Figure 16.29). The chip was then diced and a 75/200 µm capillary inserted along
the chip edge into the connection channel, mating with the twin-T region (Figure 16.29b). Hybrid
16-capillary chips were created and used on a modified MegaBACE 1000 system (GE Healthcare)
for multicapillary detection [199]. Sample loaded in the twin-T was then separated in the 40 cm long
capillary, resulting in >800 bases read to 98.5% accuracy in 56 min. Comparable capillary-only runs
with electrokinetic injection yielded only ∼650 bases [191].
The use of microfluidics in a commercial DNA sequencing system has yet to be demonstrated.
This is likely due to the lack of a compelling driver. Capillaries are still relatively inexpensive
in comparison with etched plates and schemes have been engineered to allow fairly cost effective
replacement of subgroups of capillaries, even in the high-throughput systems employing hundreds of
Radial microplate
Scan head
Excitation/emission
light (bi-directional)
To stepper motor
(4 Hz rotation rate)
Confocal LIF excitation/

four-color detection
system
FIGURE 16.28 Rotary confocal fluorescence microplate detector. (Reprinted with permission from Shi, Y.,
et al., Anal. Chem., 71, 5354, 1999. Copyright 1999 American Chemical Society.)
(a)
Photomask
Photoresist/Cr/Au
Glass wafer
Isotropic etching
Semicircular
groove
Two chips thermally

bonded together
Twin-T injector
(b)
75 µm × 200 µm capillary inserted

into 200 µm channel
FIGURE 16.29 (a) Isotopic etching to create round channels. (b) Top view of capillary inserted into the
connection channel, mating with twin-T region. (Reprinted with permission from Liu, S., Electrophoresis, 24,
3755, 2003.)
separation channels. Future improvements in the field of microfluidics may result in a transition from
capillaries to etched substrates but competitive sequencing technologies that seem to have finally
eliminated the need for electrophoresis are now on the horizon [200–203] and progress in this area
will likely redefine the role of CE in DNA sequencing.
16.4 OTHER NUCLEIC ACID APPLICATIONS

Although the focus of this chapter has been on capillary-based DNA sequencing, it is important
to point out that the same systems originally developed for DNA sequencing are also capable of
supporting a wide variety of other important electrophoresis-based assays. Methods and kits have
been developed for applications such as gene expression, heterozygote detection, mutation analysis,
allele identification, single nucleotide polymorphism screening, microsatellite instability, and ampli-
fied fragment length polymorphism (AFLP) fingerprinting and these applications are now finding
widespread acceptance in a variety of fields of science including biomedical research, human and
animal diagnostics, and agriculture.
16.5 CONCLUSIONS
Clearly, a remarkable amount of progress has been made over the past two decades in the area of
automated DNA sequencing. Early sequencing methodologies were slow, tedious, labor–intensive,
and riddled with error-making opportunities. Until the late 1980s, slab gels were manually poured,
samples were loaded by hand, and images created using X-ray film were painstakingly read by
eye. Automated film readers were introduced to improve the manual reading process but their util-
ity was short lived as better technologies were soon developed. The late 1980s and early 1990s
saw important improvements as the manual sequencing process was partially automated. Radio
isotopic tags were displaced by the use of fluorescent labels and the process of reading slab gels
was automated using multilane fluorescence slab gel readers and powerful base-calling algorithms.
The initiative to sequence the human genome drove the development of faster, high-throughput
sequencing technologies, and CE emerged in the late 1990s as that decade’s winner. Slab gels
were displaced by capillary arrays filled with replaceable gels and sensitive fluorescence technolo-
gies were successfully adapted to the more challenging multicapillary format. Current large-scale
sequencing projects would be slowed unthinkably without automated capillary sequencing machines,
which became commercially available in the late 1990s and have now made DNA sequencing much
quicker and more reliable. Today, one automated capillary DNA sequencing system can produce
data at a rate of approximately 1 million bases of raw sequence per day with as little as 15 min of
user intervention required. Without question, the works cited in this chapter and the authors who
performed those studies have all contributed to a technology and field that has enjoyed impressive
progress.
New technologies involving nonelectrophoretic, massively paralleled sequencing strategies now
seem poised to deliver the next significant improvement in the area of lower cost, high-throughput
DNA sequencing. These systems are targeting single platform throughput levels of tens of millions
of bases per hour and may make whole genome sequencing more affordable.
ACKNOWLEDGMENTS
The authors wish to thank Cynthia Johnson, Clarence Lew, Feng Liu, Lucy Liu, Paul Kraght,
Veronica V. Colinayo, and Christopher Pentoney for their assistance in creating this chapter and
Dr James C. Osborne for his support.
REFERENCES
1. Dovichi, N.J., DNA sequencing by capillary electrophoresis, Electrophoresis, 18, 2393, 1997.
2. Quesada, M.A., Replaceable polymers in DNA sequencing by capillary electrophoresis, Curr. Opin.
Biotechnol., 8, 82, 1997.
3. Slater, G.W., Kist, T.B.L., Ren, H., and Drouin, G., Recent developments in DNA electrophoretic
separations, Electrophoresis, 19, 1525, 1998.
4. Dolnik, V., DNA sequencing by capillary electrophoresis (review), J. Biochem. Biophys. Methods, 41,
103, 1999.
5. Timperman, A.T. and Sweedler, J.V., Capillary electrophoresis with wavelength-resolved fluorescence
detection, Analyst, 121, 45R, 1996.
6. Schmalzing, D., Koutny, L., Salas-Solano, O., Adourian, A., Matsudaira, P., and Ehrlich, D., Recent
developments in DNA sequencing by capillary and microdevice electrophoresis, Electrophoresis, 20,
3066, 1999.
7. Carrilho, E., DNA sequencing by capillary array electrophoresis and microfabricated array systems,
8. Mitnik, L. Novotny, M., Felten, C., Buonocore, S., Koutny, L., and Schmalzing, D., Recent advances
in DNA sequencing by capillary and microdevice electrophoresis, Electrophoresis, 22, 4104, 2001.
9. Righetti, P.G., Gelfi, C., and D’Acunto, M.R., Recent progress in DNA analysis by capillary
10. Venter, J.C., et al., The sequence of the human genome, Science, 291, 1304, 2001.
11. The International Human Genome Mapping Consortium, Aphysical map of the human genome, Nature,
409, 934, 2001.
12. Sanger, F., Nicklen, S., and Coulson, A.R., DNA sequencing with chain-terminating inhibitors, Proc.
Natl Acad. Sci. USA, 74, 5463, 1977.
13. Huang, X.C., Quesada, M.A., and Mathies, R.A., DNA sequencing using capillary array electrophore-
sis, Anal. Chem., 64, 2149, 1992.
14. Smith, L.M., Sanders, J.Z., Kaiser, R.J., Hughes, P., Dodd, C., Connell, C.R., Heiner, C., Kent, S.B.H.,
and Hood, L.E., Fluorescence detection in automated DNA sequence analysis, Nature, 321, 674, 1986.
15. Luckey, J.A., Drossman, H., Kostichka, A.J., Mead, D.A., D’Cunha, J., Norris, T.B., and Smith, L.M.,
High speed DNA sequencing by capillary electrophoresis, Nucleic Acids Res., 18, 4417, 1990.
16. Ju, J., Kheterpal, I., Scherer, J.R., Ruan, C., Fuller, C.W., Glazer, A.N., and Mathies, R.A., Design and
synthesis of fluorescence energy transfer dye-labeled primers and their application for DNA sequencing
and analysis, Anal. Biochem., 231, 131, 1995.
17. Zhang, J.Z., Fang, Y., Hou, J.Y., Ren, H.J., Jiang, R., Roos, P., and Dovichi, N.J., Use of non-
cross linked polyacrylamide for four-color DNA sequencing by capillary electrophoresis separation of
fragments up to 640 bases in length in two hours, Anal. Chem., 67, 4589, 1995.
18. Carrilho, E., Ruiz-Martinez, M.C., Berka, J., Smirnov, I., Goetzinger, W., Miller, A.W., Brady, D., and
Karger, B.L., Rapid DNA sequencing of more than 1000 bases per run by capillary electrophoresis
using replaceable linear polyacrylamide solutions, Anal. Chem., 68, 3305, 1996.
19. Klepárnik, K., Foret, F., Berka, J., Goetzinger, W., Miller, A.W., and Karger, B.L., The use of
elevated column temperature to extend DNA sequencing read lengths in capillary electrophoresis with
replaceable polymer matrices, Electrophoresis, 17, 1860, 1996.
20. Goetzinger, W., Kotler, L., Carrilho, E., Ruiz-Martinez, M.C., Salas-Solano, O., and Karger,
B.L., Characterization of high molecular mass linear polyacrylamide powder prepared by emulsion
polymerization as a replaceable polymer matrix for DNA sequencing by capillary electrophoresis,
21. Salas-Solano, O., Carrilho, E., Kotler, L., Miller, A.W., Goetzinger, W., Sosic, Z., and Karger,
B.L., Routine DNA sequencing of 1000 bases in less than one hour by capillary electrophoresis with
replaceable linear polyacrylamide solutions, Anal. Chem., 70, 3996, 1998.
22. Zhang, J., Voss, K.O., Shaw, D.F., Roos, K.P., Lewis, D.F., Yan, J., Jiang, R., Ren, H., Hou, J.Y., Fang,
Y., Puyang, X., Ahmadzadeh, H., and Dovichi, N.J., A multiple-capillary electrophoresis system for
small-scale DNA sequencing and analysis, Nucleic Acids Res., 27, e36, 1999.
23. Neumann, M., Herten, D.-P., Dietrich, A., Wolfrum, J., and Sauer, M., Capillary array scanner for
time-resolved detection and identification of fluorescently labeled DNA fragments, J. Chromatogr. A,
871, 299, 2000.
24. Zhou, H., Miller A.W., Sosic, Z., Buchholz, B., Barron, A.E., Kotler, L., and Karger, B.L., DNA
sequencing up to 1300 bases in two hours by capillary electrophoresis with mixed replaceable linear
polyacrylamide solutions, Anal. Chem., 72, 1045, 2000.
25. Kotler, L., He, H., Miller, A.W., and Karger, B.L., DNA sequencing of close to 1000 bases in 40
minutes by capillary electrophoresis using dimethyl sulfoxide and urea as denaturants in replaceable
linear polyacrylamide solutions, Electrophoresis, 23, 3062, 2002.
26. Tabor, S. and Richardson, C.C., DNA sequence analysis with a modified bacteriophage T7 DNA
polymerase: Effect of pyrophosphorolysis and metal ions, J. Biol. Chem., 265, 8322, 1990.
27. Ansorge, W., Zimmermann, J., Schwager, C., Stegemann, J., Erfle, H., and Voss, H., One label,
one tube, Sanger DNA sequencing in one and two lanes on a gel, Nucleic Acids Res., 18, 3419,
1990.
28. Pentoney, S.L., Jr., Konrad, K.D., and Kaye, W., A single-fluor approach to DNA sequence
determination using high performance capillary electrophoresis, Electrophoresis, 13, 467,
1992.
29. Swerdlow, H., Zhang, J.Z., Chen, D.Y., Harke, H.R., Grey, R., Wu, S., and Dovichi, N.J., Three DNA
sequencing methods using capillary gel electrophoresis and laser-induced fluorescence, Anal. Chem.,
63, 2835, 1991.
30. Chen, D., Harke, H.R., and Dovichi, N.J., Two-label peak-height encoded DNA sequencing by
capillary gel electrophoresis: three examples, Nucleic Acids Res., 20, 4873, 1992.
31. Gassman, E., Kuo, J.E., and Zare, R.N., Electrokinetic separation of chiral compounds, Nature, 339,
813, 1985.
32. Gordon, M.J., Huang, X., Pentoney, S.L., Jr., and Zare, R.N., Capillary electrophoresis, Science, 242,
224, 1988.
33. Prober, J.M., Trainor, G.L., Dam, R.J., Hobbs, F.W., Robertson, C.W., Zagursky, R.J., Cocuzza,
A.J., Jensen, M.A., and Baumeister, K., A system for rapid DNA sequencing with fluorescent chain-
terminating dideoxynucleotides, Science, 238, 336, 1987.
34. Kim, Y. and Yeung, E.S., Separation of DNA sequencing fragments up to 1000 bases by using
poly(ethylene oxide)-filled capillary electrophoresis, J. Chromatogr. A, 781, 315, 1997.
35. Ju, J., Ruan, C., Fuller, C.W., Glazer, A.N., and Mathies, R.A., Fluorescence energy transfer dye-
labeled primers for DNA sequencing and analysis, Proc. Ntl. Acad. Sci. USA, 92, 4347, 1995.
36. Lee, L.G., Spurgeon, S.L., Heiner, C.R., Benson, S.C., Rosenblum, B.B., Menchen, S.M., Graham,
R.J., Constantinescu., A., Upadhya, K.G., and Cassel, J.M., New energy transfer dyes for DNA
sequencing, Nucleic Acids Res., 25, 2816, 1997.
37. Rosenblum, B.B., Lee, L.G., Spurgeon, S.L., Khan, S.H., Menchen, S.M., Heiner, C.R., and Chen,
S.M., New dye-labeled terminators for improved DNA sequencing patterns, Nucleic Acids Res., 25,
4500, 1997.
38. Lee, C.S., Improved capillary electrophoretic separations associated with controlling electroendoos-
motic flow, in Handbook of Capillary Electrophoresis, 2nd ed., Landers, J.P., ed., CRC Press, Boca
Raton, FL, 1996, chap. 25.
39. Hjertén, S., High-performance electrophoresis elimination of electroendosmosis and solute adsorption,
J. Chromatogr., 347, 191, 1985.
40. Cobb, K.A., Dolnik, V., and Novotny, M., Electrophoretic separations of proteins in capillaries with
hydrolytically stable surface structures, Anal. Chem., 62, 2478, 1990.
41. Schmalzing, D., Piggee, C.A., Foret, F., Carrilho, E., and Karger, B.L., Characterization and perfor-
mance of a neutral hydrophilic coating for the capillary electrophoretic separation of biopolymers, J.
Chromatogr. A, 652, 149, 1993.
42. Hjertén, S. and Kubo, K., A new type of pH- and detergent-stable coating for elimination of
electroendosmosis and adsorption in (capillary) electrophoresis, Electrophoresis, 14, 390, 1993.
43. Moritani, T., Yoon, K., Rafailovich, M., and Chu, B., DNA capillary electrophoresis using
poly(vinylalcohol). I. inner capillary coating, Electrophoresis, 24, 2764, 2003.
44. Slater, G.W., Kenward, M., McCromick, L.C., and Gauthier, M.G., The theory of DNA separation by
capillary electrophoresis, Curr. Opin. Biotechnol., 14, 58, 2003.
45. Ruiz-Martinez, M.C., Berka, J., Belenkii, A., Foret, F., Miller, A.W., and Karger, B.L., DNA sequenc-
ing by capillary electrophoresis with replaceable linear polyacrylamide and laser-induced fluorescence
detection, Anal. Chem., 65, 2851, 1993.
46. Pariat, Y.F., Berka, J., Heiger, D.N., Schmitt, T., Vilenchik, M., Cohen, A.S., Foret, F., and Karger, B.L.,
Separation of DNA fragments by capillary electrophoresis using replaceable linear polyacrylamide
matrices, J. Chromatogr. A, 652, 57, 1993.
47. Sudor, J., Foret, F., and Boček, P., Pressure refilled polyacrylamide columns for the separation of
oligonucleotides by capillary electrophoresis, Electrophoresis, 12, 1056, 1991.
48. Guttman, A., Capillary electrophoresis using replaceable gels, US Patents 5,332,481, 5,421,980,
RE37,606E, and RE37,941E.
49. Guttman, A. and Cooke N., Capillary gel affinity electrophoresis of DNA fragments, Anal. Chem., 63,
2038, 1991.
50. Grossman, P.D. and Soane, D.S., Experimental and theoretical studies of DNA separations by capillary
electrophoresis in entangled polymer solutions, Biopolymers, 31, 1221, 1991.
51. Karger, B.L., Foret, F., and Berka, J., Capillary electrophoresis with polymer matrices: DNA and
protein separation and analysis, Methods Enzymol., 271, 293, 1996.
52. Sunada, W.M. and Blanch, H.W., Polymeric separation media for capillary electrophoresis of nucleic
acids, Electrophoresis, 18, 2243, 1997.
53. Heller, C., Principles of DNA separation with capillary electrophoresis, Electrophoresis, 22, 629,
2001.
54. Viovy, J.-L., Electrophoresis of DNA and other polyelectrolytes: Physical mechanisms, Rev. Modern
Phys., 72, 813, 2000.
55. Quesada, M.A. and Menchen, S., Replaceable polymers for DNAsequencing by capillary electrophore-
sis, Methods Mol. Biol., 162, 139, 2001.
56. Albarghouthi, M.N. and Barron, A.E., Polymeric matrices for DNA sequencing by capillary
57. Ying, Q. and Chu, B., Overlap concentration of macromolecules in solution, Macromolecules, 20,
362, 1987.
58. Wu, C., Quesada, M.A., Schneider, D.K., Farinato, R., Studier, F.W., and Chu, B., Polyacrylamide
solutions for DNA sequencing by capillary electrophoresis: Mesh sizes, separation and dispersion,
59. Ogston, A.G., The spaces in a uniform random suspension of fibers, Trans. Faraday Soc., 54, 1754,
1958.
60. Slater, G.W., Mayer, P., and Drouin, G., Migration of DNA through gels, Methods Enzymol., 270, 272,
1996.
61. Slater, G.W. and Noolandi, J., On the reptation theory of gel electrophoresis, Biopolymers, 25, 431,
1986.
62. Duke, T., Viovy, J.-L., and Semenov, A.N., Electrophoretic mobility of DNA in gels. I. New biased
reptation theory including fluctuations, Biopolymers, 34, 239, 1994.
63. Konrad, K.D. and Pentoney, S.L., Jr., Contribution of secondary structure to DNA mobility in capillary
gels, Electrophoresis, 14, 502, 1993.
64. Cohen, A.S., Najarian, D.R., Paulus, A., Guttman, A., Smith, J.A., and Karger, B.L., Rapid separation
and purification of oligonucleotides by high-performance capillary gel electrophoresis, Proc. Natl
Acad. Sci. USA, 85, 9660, 1988.
65. Guttman, A., Cohen, A.S., Heiger, D.N., and Karger, B.L., Analytical and micropreparative ultrahigh
resolution of oligonucleotides by polyacrylamide gel high-performance capillary electrophoresis, Anal.
Chem., 62, 137, 1990.
66. Paulus, A., Gassmann, E., and Field, M.J., Calibration of polyacrylamide gel columns for the separation
of oligonucleotides by capillary electrophoresis, Electrophoresis, 11, 702, 1990.
67. Drossman, H., Luckey, A.J., Kostichka, J., D’Cunha, J., and Smith, L.M., High-speed separations of
DNA sequencing reactions by capillary electrophoresis, Anal. Chem., 62, 900, 1990.
68. Swerdlow, H. and Gesteland, R., Capillary gel electrophoresis for rapid, high resolution DNA
sequencing, Nucleic Acids Res., 18, 1415, 1990.
69. Cohen, A.S., Najarian, D.R., and Karger, B.L., Separation and analysis of DNA sequence reaction
products by capillary gel electrophoresis, J. Chromatogr., 516, 49, 1990.
70. Swerdlow, H., Wu, S., Harke, H., and Dovichi, N.J., Capillary gel electrophoresis for DNA sequenc-
ing laser-induced fluorescence detection with the sheath flow cuvette, J. Chromatogr., 516, 61,
1990.
71. Karger, A.E., Harris, J.M., and Gesteland, R.F., Multiwavelength fluorescence detection for DNA
sequencing using capillary electrophoresis, Nucleic Acids Res., 19, 4955, 1991.
72. Chiari, M., Nesi, M., Fazio, M., and Righetti, P.G., Capillary electrophoresis of macromolecules in
“syrupy” solutions: Facts and misfacts, Electrophoresis, 13, 690, 1992.
73. Best, N., Arriga, E., Chen, D.Y., and Dovichi, N.J., Separation of fragments up to 570 bases in length
by use of 6% T non-cross-linked polyacrylamide for DNA sequencing in capillary electrophoresis,
Anal. Chem., 66, 4063, 1994.
74. Doherty, E.A.S., Kan, C.-W., and Barron, A.E., Sparsely cross-linked “nanogels” for microchannel
DNA sequencing, Electrophoresis, 24, 4170, 2003.
75. Doherty, E.A.S., Kan, C.-W., Paegel, B.M., Yeung, S.H.I., Cao, S., Mathies, R.A., and Barron,
A.E., Sparsely cross-linked “nanogel” matrixes as fluid, mechanically stabilized polymer networks for
high-throughput microchannel DNA sequencing, Anal. Chem., 76, 5249, 2004.
76. Madabhushi, R.S., Separation of 4-color DNA sequencing extension products in noncovalently coated
capillaries using low-viscosity polymer solutions, Electrophoresis, 19, 224, 1998.
77. Madabhushi, R., DNA sequencing in noncovalently coated capillaries using low viscosity polymer
solutions, Methods Mol. Biol., 163, 309, 2001.
78. Heller, C., Finding a universal low viscosity polymer for DNA separation (II), Electrophoresis, 19,
3114, 1998.
79. Heller, C., Separation of double-stranded and single-stranded DNA in polymer solutions: I. mobility
and separation mechanism, Electrophoresis, 20, 1962, 1999.
80. Song, L., Liang, D., Fang, D., and Chu, B., Fast DNA sequencing up to 1000 bases by capillary
electrophoresis using poly(N, N-dimethylacrylamide) as a separation medium, Electrophoresis, 22,
1987, 2001.
81. Griess, G.A., Hardies, S.C., and Serwer, P., Matrix conditioning for lengthened capillary DNA
sequencing, Electrophoresis, 26, 102, 2005.
82. Albarghouthi, M.N., Buchholz, B.A., Huiberts, P.J., Stein, T.M., and Barron, A.E., Poly-N-
hydroxyethylacrylamide (polyduramide): A novel, hydrophilic, self-coating polymer matrix for DNA
sequencing by capillary electrophoresis, Electrophoresis, 23, 1429, 2002.
83. Song, L., Liang, D., Kielescawa, J., Liang, J., Tjoe, E., Fang, D., and Chu, B., DNA sequencing by cap-
illary electrophoresis using copolymers of acrylamide and N, N-dimethyl acrylamide, Electrophoresis,
22, 729, 2001.
84. Wang, Y., Liang, D., Ying, Q., and Chu, B., Quasi-interpenetrating network formed by polyacrylamide
and poly(N, N-dimethylacrylamide) used in high-performance DNA sequencing analysis by capillary
85. Bashkin, J., Marsh, M., Barker, D., and Johnston, R., DNA sequencing by capillary electrophoresis
with a hydroxyethylcellulose sieving buffer, Appl. Theor. Electrophor., 6, 23, 1996.
86. Marsh, M., Tu, O., Dolnik, V., Roach, D., Solomon, N., Bechtol, K., Smietana, P., Wang, L., Li, X.,
Cartwright, P., Marks, A., Barker, D., Harris, D., and Baskin, J., High-throughput DNA sequencing
on a capillary array electrophoresis system, J. Capillary Electrophor., 4, 83, 1997.
87. Dolnik, V. and Gurske, W.A., Capillary electrophoresis in sieving matrices: Selectivity per base,
mobility slope, and inflection slope, Electrophoresis, 20, 3373, 1999.
88. Kheterpal, I. and Mathies, R.A., Capillary array electrophoresis DNA sequencing, Anal. Chem., 71,
31A, 1999.
89. Chang, H.T. and Yeung, E.S., Poly(ethyleneoxide) for high-resolution and high-speed separation of
DNA by capillary electrophoresis, J. Chromatogr. B, 669, 113, 1995.
90. Fung, E.N. and Yeung, E.S., High-speed DNA sequencing by using mixed poly(ethylene oxide)
solutions in uncoated capillary columns, Anal. Chem., 67, 1913, 1995.
91. Tan, H. and Yeung, E.S., Characterization of dye-induced mobility shifts affecting DNA sequencing
in poly(ethylene oxide) sieving matrix, Electrophoresis, 18, 2893, 1997.
92. Fung, E.N., Pang, H.-M., and Yeung, E.S., Fast DNA separations using poly(ethylene oxide) in
non-denaturing medium with temperature programming, J. Chromatogr. A, 806, 157, 1998.
93. Wei, W. and Yeung, E.S., Improvements in DNA sequencing by capillary electrophoresis at elevated
temperature using poly(ethylene oxide) as a sieving matrix, J. Chromatogr. B, 745, 221, 2000.
94. Preisler, J. and Yeung, E.S., Characterization of nonbonded poly(ethyleneoxide) coating for capillary
electrophoresis via continuous monitoring of electroosmotic flow, Anal. Chem., 68, 2885, 1996.
95. Gao, Q. and Yeung, E.S., A matrix for DNA separation: Genotyping and sequencing using
poly(vinylpyrrolidone) solution in uncoated capillaries, Anal. Chem., 70, 1382, 1998.
96. Song, J.M. and Yeung, E.S., Optimization of DNA electrophoretic behavior in poly(vinyl pyrrolidone)
sieving matrix for DNA sequencing, Electrophoresis, 22, 748, 2001.
97. Dolnik, V., Gurske, W.A., and Padua, A., Galactomannans as a sieving matrix in capillary
98. Menchen, S., Johnson, B., Winnik, M.A., and Xu, B., Flowable networks as DNA sequencing media
in capillary columns, Electrophoresis, 17, 1451, 1996.
99. Chu, B. and Liang, D., Copolymer solutions as separation media for DNA capillary electrophoresis,
J. Chromatogr. A, 966, 1, 2002.
illary electrophoresis using copolymers of acrylamide and N, N-dimethylacrylamide, Electrophoresis,
22, 729, 2001.
101. Albarghouthi, M.N., Buchholz, B.A., Doherty, E.A.S., Bogdan, F.M., Zhou, H., and Barron, A.E.,
Impact of polymer hydrophobicity on the properties and performance of DNA sequencing matrices
for capillary electrophoresis, Electrophoresis, 22, 737, 2001.
102. He, H., Buchholz, B.A., Kotler, L., Miller, A.W., Barron, A.E., and Karger, B.L., DNA sequencing
with hydrophilic and hydrophobic polymers at elevated column temperatures, Electrophoresis, 23,
1421, 2002.
103. Buchholz, B.A., Doherty, E.A.S., Albarghouthi, M.N., Bogdan, F.M., Zahn, J.M., and Barron, A.E.,
Microchannel DNA sequencing matrices with a thermally controlled “viscosity switch,” Anal. Chem.,
73, 157, 2001.
104. Buchholz, B.A., Shi, W., and Barron, A.E., Microchannel DNA sequencing matrices with switchable
viscosities, Electrophoresis, 23, 1398, 2002.
105. Kan, C.-W., Doherty, E.A.S, and Barron, A.E., A novel thermogelling matrix for microchannel DNA
sequencing based on poly-N-alkoxyalkylacrylamide copolymers, Electrophoresis, 24, 4161, 2003.
106. Sudor, J., Barbier, V., Thirot, S., Godfrin, D., Hourdet, D., Millequant, M., Blanchard, J., and Viovy,
J.-L., New block-copolymer thermoassociating matrices for DNA sequencing: Effect of molecular
structure on rheology and resolution, Electrophoresis, 22, 720, 2001.
107. Barbier, V., Buchholz, B.A., Barron, A.E., and Viovy, J.-L., Comb-like copolymers as self-coating,
low-viscosity and high-resolution matrices for DNA sequencing, Electrophoresis, 23, 1441, 2002.
108. Pentoney, Jr., S.L. and Sweedler, J.V., in Handbook of Capillary Electrophoresis, 2nd ed., Landers,
J.P., ed., CRC Press, Boca Raton, FL, 1996, chap. 12.
109. Zagursky, R.J. and McCormick, R.M., DNA sequencing separations in capillary gels on a modified
commercial DNA sequencing instrument, BioTechniques, 9, 74, 1990.
110. Carson, S., Cohen, A.S., Belenkii, A., Ruiz-Martinez, M.C., Berka, J., and Karger, B.L., DNA sequenc-
ing by capillary electrophoresis: Use of a two-laser-two-window intensified diode array detection
system, Anal. Chem., 65, 3219, 1993.
111. Mathies, R.A. and Huang, X.C., Capillary array electrophoresis: An approach to high-speed, high-
throughput DNA sequencing, Nature, 359, 167, 1992.
112. Huang, X.C., Quesada, M.A., and Mathies, R.A., Capillary array electrophoresis using laser-excited
confocal fluorescence detection, Anal. Chem., 64, 967, 1992.
113. Kheterpal, I., Scherer, J.R., Clark, S.M., Radhakrishnan, A., Ju, J., Ginther, C.L., Sensabaugh, G.F.,
and Mathies, R.A., DNA sequencing using a four-color confocal fluorescence capillary array scanner,
114. Bashkin, J.S., Bartosiewicz, M., Roach, D., Leong, J., Barker, D., Johnston, R., Implementation of a
capillary array electrophoresis instrument, J. Capillary Electrophor., 3, 61, 1996.
115. Scherer, J.R., Kheterpal, I., Radhakrishnan, A., Ja, W.W., and Mathies, R.A., Ultra-high through-
put rotary capillary array electrophoresis scanner for fluorescent DNA sequencing and analysis,
116. Ueno, K. and Yeung, E.S., Simultaneous monitoring of DNA fragments separated by electrophoresis
in a multiplexed array of 100 capillaries, Anal. Chem., 66, 1424, 1994.
117. Pang, H., Pavski, V., and Yeung, E.S., DNA sequencing using 96-capillary array electrophoresis,
J. Biochem. Biophys. Methods, 41, 121, 1999.
118. Xue, G. and Yeung, E.S., Two-color excitation system for fluorescence detection in DNA sequencing
by capillary array electrophoresis, Electrophoresis, 23, 1490, 2002.
119. Zhong, W. and Yeung, E.S., Multiplexed capillary electrophoresis for DNA sequencing with ultra
violet absorption detection, J. Chromatogr. A, 960, 229, 2002.
120. Behr, S., Matzig, M., Levin, A., Eickhoff, H., and Heller, C., A fully automated multicapillary
electrophoresis device for DNA analysis, Electrophoresis, 20, 1492, 1999.
121. Lu, X. and Yeung, E.S., Optimization of excitation and detection geometry for multiplexed capillary
array electrophoresis of DNA fragments, Appl. Spectrosc., 49, 605, 1995.
122. Lu, S.X. and Yeung, E.S., Side-entry excitation and detection of square capillary array electrophoresis
for DNA sequencing, J. Chromatogr. A, 853, 359, 1999.
123. Quesada, M.A. and Zhang, S., Multiple capillary DNA sequencer that uses fiber-optic illumination
and detection, Electrophoresis, 17, 1841, 1996.
124. Quesada, M.A., Dhadwal, H.S., Fisk, D., and Studier, F.W., Multi-capillary optical waveguides for
DNA sequencing, Electrophoresis, 19, 1415, 1998.
125. Anazawa, T., Takahashi, S., and Kambara, H., Acapillary-array electrophoresis system using side-entry
on-column laser irradiation combined with glass rod lenses, Electrophoresis, 20, 539, 1999.
126. Hanning, A., Lindberg, P., Westberg, J., and Roeraade, J., Laser-induced fluorescence detection by
liquid core waveguiding applied to DNA sequencing by capillary electrophoresis, Anal. Chem., 72,
3423, 2000.
127. Hanning, A., Westberg, J., and Roeraade, J., A liquid core waveguide fluorescence detector for mul-
ticapillary electrophoresis applied to DNA sequencing in a 91-capillary array, Electrophoresis, 21,
3290, 2000.
128. Chen, F.-T., Tusak, A., Pentoney, S. Jr., Konrad, K., Lew, C., Koh, E., and Sternberg, J., Semiconductor
laser-induced fluorescence detection in capillary electrophoresis using a cyanine dye, J. Chromatogr.
A, 652, 355, 1993.
129. Mank, A.J.G. and Yeung, E.S., Diode laser-induced fluorescence detection in capillary electrophoresis
after pre-column derivatization of amino acids and small peptides, J. Chromatogr. A, 708, 309, 1995.
130. Zhang, J.Z., Chen D.Y., Wu, S., Harke, H.R., and Dovichi, N.J., High-sensitivity laser-induced
fluorescence detection for capillary electrophoresis, Clin. Chem., 37/9, 1942, 1991.
131. Wu, S. and Dovichi, N.J., High-sensitivity fluorescence detector for fluorescein isothiocyanate
derivatives of amino acids separated by capillary zone electrophoresis, J. Chromatogr., 480, 141,
1989.
132. Chen, D.Y., Swerdlow, H.P., Harke, H.R., Zhang, J.Z., and Dovichi, N.J., Low-cost, high-
sensitivity laser-induced fluorescence detection for DNA sequencing by capillary gel electrophoresis,
J. Chromatogr., 559, 237, 1991.
133. Chen, D.Y. and Dovichi, N.J., Yoctomole detection limit by laser-induced fluorescence in capillary
electrophoresis, J. Chromatogr. B, 657, 265, 1994.
134. Starke, H.R., Yan, J.Y., Zhang, J.Z., Mühlegger, K., Effgen, K., and Dovichi, N.J., Internal fluorescence
labeling with fluorescent deoxynucleotides in two-label peak-height encoded DNA sequencing by
capillary electrophoresis, Nucleic Acids Res., 22, 3997, 1994.
135. Crabtree, H.J., Bay, S.J., Lewis, D.F., Zhang, J., Coulson, L.D., Fitzpatrick, G.A., Delinger, S.L.,
Harrison, D.J., and Dovichi, N.J., Construction and evaluation of capillary array DNA sequencer
based on a micromachined sheath-flow cuvette, Electrophoresis, 21, 1329, 2000.
136. Zhang, J., Yang, M., Puyang, X., Fang, Y., Cook, L.M., and Dovichi, N.J., Two-dimensional direct-
reading fluorescence spectrograph for DNA sequencing by capillary array electrophoresis, Anal.
Chem., 73, 1234, 2001.
137. Kambara, H. and Takahashi, S., Multiple-sheathflow capillary array DNA analyzer, Nature, 361, 565,
1993.
138. Lieberwirth, U., Arden-Jacob, J., Drexhage, K.H., Herten, D.P., Müller, R., Neumann, M., Schulz,
A., Siebert, S., Sagner, G., Klingel, S., Sauer, M., and Wolfrum, J., Multiplex dye DNA sequencing in
capillary gel electrophoresis by diode laser-based time-resolved fluorescence detection, Anal. Chem.,
70, 4771, 1998.
139. Zhu, L., Stryjewski, W., Lassiter, S., and Soper, S.A., Fluorescence multiplexing with time-resolved
and spectral discrimination using a near-IR detector, Anal. Chem., 75, 2280, 2003.
140. Li, L.-C. and McGown, L.B., On-the-fly frequency-domain fluorescence lifetime detection in capillary
141. Flanagan, J.H., Jr., Owens, C.V., Romero, S.E., Waddell, E., Kahn, S.H., Hammer, R.P., and
Soper, S.A., Near-infrared heavy-atom-modified fluorescent dyes for base-calling in DNA-sequencing
applications using temporal discrimination, Anal. Chem., 70, 2676, 1998.
142. Nunnally, B.K., He, H., Li, L.-C., Tucker, S.A., and McGown, L.B., Characterization of visible dyes
for four-decay fluorescence detection in DNA sequencing, Anal. Chem., 69, 2392, 1997.
143. Mihindukulasuriya, S.H., Morcone, T.K., and McGown, L.B., Characterization of acridone dyes for
use in four-decay detection in DNA sequencing, Electrophoresis, 24, 20, 2003.
144. Li, L.-C., He, H., Nunnally, B.K., and McGown, L.B., On-the-fly fluorescence lifetime detection of
labeled DNA primers, J. Chromatogr. B, 695, 85, 1997.
145. He, H. and McGown, L.B., DNA sequencing by capillary electrophoresis with four-decay fluorescence
detection, Anal. Chem., 72, 5865, 2000.
146. Soper, S.A. and Mattingly, Q.L., Steady-state and picosecond laser fluorescence studies of nonradiative
pathways in tricarbocyanine dyes: Implications to the design of near-IR fluorochromes with high
fluorescence efficiencies, J. Am. Chem. Soc., 116, 3744, 1994.
147. Lassiter, S.J., Stryjewski, W., Owens, C.V., Flanagan, J.H., Jr., Hammer, R.P., Khan, S., and Soper,
S.A., Optimization of sequencing conditions using near-infrared lifetime identification methods in
capillary gel electrophoresis, Electrophoresis, 23, 1480, 2002.
148. Alaverdian, L., Alaverdian, S., Bilenko, O., Bogdanov, I., Filippova, E., Gavrilov, D., Gorbovitski, B.,
Gouzman, M., Gudkov, G., Domratchev, S., Kosobokova, O., Lifshitz, N., Luryi, S., Ruskovoloshin,
V., Stepoukhovitch, A., Tcherevishnick, M., Tyshko, G., and Gorfinkel, V., A family of novel DNA
sequencing instruments based on single-photon detection, Electrophoresis, 23, 2804, 2002.
149. Gavrilov, D.N., Gorbovitski, B., Gouzman, M., Gudkov, G., Stepoukhovitch, A., Ruskovoloshin, V.,
Tsuprik, A., Tyshko, G., Bilenko, O., Kosobokova, O., Luryi, S., and Gorfinkel, V., Dynamic range of
fluorescence detection and base-calling accuracy in DNA sequencer based on single-photon counting,
150. Lewis, E.K., Haaland, W.C., Nguyen, F., F., Heller, D.A., Allen, M.J., MacGregor, R.R., Berger, C.S.,
Willingham, B., Burns, L.A., Scott, G.B.I., Kittrell, C., Johnson, B.R., Curl, R.F., and Metzker, M.L.,
Color-blind fluorescence detection for four-color DNA sequencing, Proc. Natl Acad. Sci. USA, 102,
5346, 2005.
151. Brennan, T., Chakel, J., Bente, P., and Field, M., New methods to sequence DNA by mass spectrom-
etry, in Proceedings of SPIE—Volume 1206 New Technologies in Cytometry and Molecular Biology,
Salzman, G.C., ed., July 1990, pp. 60–77.
152. Soper, S.A., Williams, D.C., Xu, Y., Lassiter, S.J., Zhang, Y., Ford, S.M., Sanger DNA-sequencing
reactions performed in a solid-phase nanoreactor directly coupled to capillary gel electrophoresis,
Anal. Chem., 70, 4036, 1998.
153. Swerdlow, H., Jones, B.J., and Wittwer, C.T., Fully automated DNA reaction and analysis in a fluidic
capillary instrument, Anal. Chem., 69, 848, 1997.
154. Tan, H. and Yeung, E.S., Automation and integration of multiplexed on-line sample preparation with
capillary electrophoresis for high-throughput DNA sequencing, Anal. Chem., 70, 4044, 1998.
155. Zhang, Y., Tan, H., and Yeung, E.S., Multiplexed automated DNA sequencing directly from single
bacterial colonies, Anal. Chem., 71, 5018, 1999.
156. He, Y., Pang, H.-M., and Yeung, E.S., Integrated electroosmotically-driven on-line sample purification
system for nanoliter DNA sequencing by capillary electrophoresis, J. Chromatogr. A, 894, 179, 2000.
157. Tan, H. and Yeung, E.S., Integrated on-line system for DNA sequencing by capillary electrophoresis:
From template to called bases, Anal. Chem., 69, 664, 1997.
158. Ruiz-Martinez, M.C., Salas-Solano, O., Carrilho, E., Kotler, L., and Karger, B.L., A sample purifi-
cation method for rugged and high-performance DNA sequencing by capillary electrophoresis using
replaceable polymer solutions. A. Development of the cleanup protocol, Anal. Chem., 70, 1516,
1998.
159. Salas-Solano, O., Ruiz-Martinez, M.C., Carrilho, E., Kotler, L., and Karger, B.L., A sample purifi-
cation method for rugged and high-performance DNA sequencing by capillary electrophoresis using
replaceable polymer solutions. B. Quantitative determination of the role of sample matrix components
on sequencing analysis, Anal. Chem., 70, 1528, 1998.
160. Luckey, J.A. and Smith, L.M., Optimization of electric field strength for DNA sequencing in capillary
gel electrophoresis, Anal. Chem., 65, 2841, 1993.
161. Ruan, C.C., Samols, S.B., and Fuller, C.W., Role of pyrophosphorolysis in DNA sequencing, Editorial
Comments, Vol. 17, No. 2, p. 1, United States Biochemical Corporation, 1990.
162. Ewing, B., Hillier, L., Wendl, M.C., and Green, P., Base-calling of automated sequencer traces using
phred. I. Accuracy assessment, Genome Res., 8, 175, 1998.
163. Ewing, B. and Green, P., Base-calling of automated sequencer traces using phred. II. Error probabilities,
Genome Res., 8, 175, 1998.
164. Effenhauser, C.S., Bruin, G.J.M., and Paulus, A., Integrated chip-based capillary electrophoresis,
165. Dolnik, V., Liu, S., and Jovanovich, S., Capillary electrophoresis on microchip, Electrophoresis, 21,
41, 2000.
166. Gao, Q., Shi, Y., and Liu, S., Multiple-channel microchips for high-throughput DNA analysis by
capillary electrophoresis, Fresenius J. Anal. Chem., 371, 137, 2001.
167. Chen, L. and Ren, J., High-throughput DNA analysis by microchip electrophoresis, Comb. Chem.
High Throughput Screen., 7, 29, 2004.
168. Zhu, L., Stryjewski, W.J., and Soper, S.A., Multiplexed fluorescence detection in microfabri-
cated devices with both time-resolved and spectral-discrimination capabilities using near-infrared
fluorescence, Anal. Biochem., 330, 206, 2004.
169. Kan, C.-W., Fredlake, C.P., Doherty, E.A.S., and Barton, A.E., DNA sequencing and genotyping in
miniaturized electrophoresis systems, Electrophoresis, 25, 3564, 2004.
170. Ueberfeld, J., El-Difrawy, S.A., Ramdhanie, K., and Ehrlich, D.J., Solid-support sample loading for
DNA sequencing, Anal. Chem., 78, 3632, 2006.
171. Shi, Y., DNA sequencing and multiplex STR analysis on plastic microfluidic devices, Electrophoresis,
27, 3703, 2006.
172. Effenhauser, C.S., Paulus, A., Manz, A., and Widmer, H.M., High-speed separation of antisense
oligonucleotides on a micromachined capillary electrophoresis device, Anal. Chem., 66, 2949, 1994.
173. Woolley, A.T. and Mathies, R.A., Ultra-high-speed DNA fragment separations using microfabricated
capillary array electrophoresis chips, Proc. Natl Acad. Sci. USA, 91, 11348, 1994.
174. Simpson, P.C., Woolley, A.T., and Mathies, R.A., Microfabrication technology for the production of
capillary array electrophoresis chips, J. Biomed. Microdevices, 1, 7, 1998.
175. Woolley, A.T. and Mathies, R.A., Ultra-high-speed DNA sequencing using capillary electrophoresis
chips, Anal. Chem., 67, 3676, 1995.
176. Scherer, J.R., Paegel, B.M., Wedemayer, G.J., Emrich, C.A., Lo, J., Medintz, I.L., and Mathies, R.A.,
High-pressure gel loader for capillary array electrophoresis microchannel plates, Biotechniques, 31,
1150, 2001.
177. Liu, S., Shi, Y., Ja, W.W., and Mathies, R.A., Optimization of high-speed DNA sequencing on
microfabricated capillary electrophoresis channels, Anal. Chem., 71, 566, 1999.
178. Woolley, A.T., Sensabaugh, G.F., and Mathies, R.A., High-speed DNA genotyping using microfabri-
cated capillary array electrophoresis chips, Anal. Chem., 69, 2181, 1997.
179. Liu, S., Ren, H., Gao, Q., Roach, D.J., Loder, R.T., Jr., Armstrong, T.M., Mao, Q., Blaga, L., Barker,
D.L., and Jovanovich, S.B., Automated parallel DNA sequencing on multiple channel microchips,
Proc. Natl. Acad. Sci. USA, 97, 5369, 2000.
180. Simpson, P.C., Roach, D., Woolley, A.T., Thorsen, T., Johnston, R., Sensabaugh, G.F., and
Mathies, R.A., High-throughput genetic analysis using microfabricated 96-sample capillary array
electrophoresis microplates, Proc. Natl Acad. Sci. USA, 95, 2256, 1998.
181. Shi, Y., Simpson, P.C., Schere, J.R., Wexler, D., Skibola, C., Smith, M.T., and Mathies, R.A., Radial
capillary array electrophoresis microplate and scanner for high-performance nucleic acid analysis,
Anal. Chem., 71, 5354, 1999.
182. Medintz, I.L., Paegel, B.M., Blazej, R.G., Emrich, C.A., Berti, L., Scherer, J.R., and Mathies, R.A.,
High-performance genetic analysis using microfabricated capillary array electrophoresis microplates,
183. Paegel, B.M., Emrich, C.A., Wedemayer, G.J., Scherer, J.R., and Mathies, R.A., High throughput
DNA sequencing with a microfabricated 96-lane capillary array electrophoresis bioprocessor, Proc.
Natl. Acad. Sci. USA, 99, 574, 2002.
184. Backhouse, C., Caamano, M., Oaks, F., Nordman, E., Carrillo, A., Johnson, B., and Bay, S., DNA
sequencing in a monolithic microchannel device, Electrophoresis, 21, 150, 2000.
185. Emrich, C.A., Tian, H., Medintz, I.L., and Mathies, R.A., Microfabricated 384-lane capillary array
electrophoresis bioanalyzer for ultrahigh-throughput genetic analysis, Anal. Chem., 74, 5076, 2002.
186. Tian, H., Emrich, C.A., Scherer, J.R., Mathies, R.A., Andersen, P.S., Larsen, L.A., and Christiansen,
M., High-throughput single-strand conformation polymorphism analysis on a microfabricated capillary
array electrophoresis device, Electrophoresis, 26, 1834, 2005.
187. Salas-Solano, O., Schmalzing, D., Koutny, L., Buonocore, S., Adourian, A., Matsudaira, P.,
and Ehrlich, D., Optimization of high-performance DNA sequencing on short microfabricated
electrophoretic devices, Anal. Chem., 72, 3129, 2000.
188. Koutny, L., Schmalzing, D., Salas-Solano, O., El-Difrawy, S., Adourian, A., Buonocore, S., Abbey,
K., McEwan, P., Matsudaira, P., and Ehrlich, D., Eight hundred-base sequencing in a microfabricated
electrophoretic device, Anal. Chem., 72, 3388, 2000.
189. Swanson, D., The art of the state of nucleic acid sequencing, Scientist, 14, 23, 2000.
190. Schmalzing, D., Tsao, N., Koutny, L., Chisholm, D., Srivastava, A., Adourian, A., Linton, L., McEwan,
P., Matsudaira, P., and Ehrlich, D., Toward real-world sequencing by microdevice electrophoresis,
Genome Res., 9, 853, 1999.
191. Liu, S., A microfabricated hybrid device for DNA sequencing, Electrophoresis, 24, 3755, 2003.
192. Schmalzing, D., Adourian, A., Koutny, L., Ziaugra, L., Matsudaira, P., and Ehrlich, D., DNA
sequencing on microfabricated electrophoretic devices, Anal. Chem., 70, 2303, 1998.
193. Simpson, J.W., Ruiz-Martinez, M.C., Mulhern, G.T., Berka, J., Latimer, D.R., Ball, J.A., Rothberg,
J.M., and Went, G.T., A transmission imaging spectrograph and microfabricated channel system for
DNA analysis, Electrophoresis, 21, 135, 2000.
194. Haab, B.B. and Mathies, R.A., Single-molecule detection of DNA separations in microfabricated
capillary electrophoresis chips employing focused molecular streams, Anal. Chem., 71, 5137, 1999.
195. Jacobson, S.C., Hergenröder, R., Koutny, L.B., Warmack, R.J., and Ramsey, J.M., Effects of injection
schemes and column geometry on the performance of microchip electrophoresis devices, Anal. Chem.,
66, 1107, 1994.
196. Culbertson, C.T., Jacobson, S.C., and Ramsey, J.M., Dispersion sources for compact geometries on
microchips, Anal. Chem., 70, 3781, 1998.
197. Paegel, B.M., Hutt, L.D., Simpson, P.C., and Mathies, R.A., Turn geometry for minimizing band
broadening in microfabricated capillary electrophoresis channels, Anal. Chem., 72, 3030, 2000.
198. Molho, J.I., Herr, A.E., Mosier, B.P., Santiago, J.G., and Kenny, T.W., Optimization of turn geometries
for microchip electrophoresis, Anal. Chem., 73, 1350, 2001.
199. Liu, S., Elkin, C., and Kapur, H., Sequencing of real-world samples using a microfabricated hybrid
device having unconstrained straight separation channels, Electrophoresis, 24, 3762, 2003.
200. Shendure, J., Porreca, G.J., Reppas, N.B., Lin, X., McCutcheon, J.P., Rosenbaum, A.M., Wang, M.D.,
Zhang, K., Mitra, R.D., and Church, G.M., Accurate multiplex colony sequencing of an evolved
bacterial genome, Science, 309, 1728, 2005.
201. Margulies, M., et al., Genome sequencing in microfabricated high-density picolitre reactors, Nature,
437, 376, 2005.
202. Bonetta, L., Genome sequencing in the fast lane, Nat. Methods, 3, 141, 2006.
203. Pennisi, E., Cut-rate genomes on the horizon? Science, 309, 862, 2005.
204. Jorgenson, J.W. and Lukacs, K.D., Zone electrophoresis in open-tubular glass capillaries, Anal. Chem.,
53, 1298, 1981.
205. Manabe, T., Chen, N., Terabe, S., Yohda, M., and Endo, I., Effects of linear polyacrylamide concen-
trations and applied voltages on the separation of oligonucleotides and DNA sequencing fragments by
17 Dynamic Computer
Simulation Software for
Michael C. Breadmore and Wolfgang Thormann
CONTENTS
17.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515

17.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 516
17.3 Historical Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 519
17.4 Theoretical Aspects of SIMUL5 and GENTRANS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 520
17.4.1 Mathematical Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 520
17.4.2 Numerical Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521
17.4.3 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521
17.4.4 Execution of a Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 522
17.5.1 Isotachophoresis and Moving Boundary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . 525
17.5.2 Isoelectric Focusing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527
17.5.3 Zone Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 529
17.5.4 tITP Stacking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 530
17.5.5 pH Stacking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 532
17.5.6 Miscellaneous Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 540
17.1 INTRODUCTION
Since the early 1980s, the use of computers within society has increased at a significant rate. They
are now involved in almost every aspect of modern day life, and it is therefore not surprising that
this prevalence has transferred into the professional field of the modern day scientist. Of particular
interest in this chapter is the use in capillary electrophoresis (CE) to model or “simulate” the myriad
of chemical and physical processes that occur during an electrophoretic separation. This is achieved
through a series of model equations derived from the transport concepts in solution under the influence
of a d.c. electric field together with user-inputted experimental conditions, such as concentrations,
mobilities, field strengths or currents, capillary lengths, and so on, and a theoretical separation
is calculated. One of the potential advantages of a simulation is the possibility of determining
appropriate separation conditions well before any laboratory experiments are undertaken, making
method development a simpler task, although it must be noted that this can currently only be done
with limited capacity. Simulations can also provide detailed insight into the processes involved
515
in the separation allowing the researcher to understand the result of a particular combination of
experimental conditions, and with this information design new superior systems, and this has been
the major area in which simulation software has been used to date. In addition, simulations can be
employed for educational purposes.
The result sought by the researchers will quite often dictate the simulation approach to be used, and
in this regard it is necessary to distinguish the different types of computer models for electrophoresis.
The most important for CE are (1) the dynamic models providing complete sets of concentration,
conductivity, pH, and flow profiles as a function of time and location (e.g., at the point of detection
along the capillary) for zone electrophoresis (ZE), moving boundary electrophoresis (MBE), iso-
tachophoresis (ITP), and isoelectric focusing (IEF) [1,2]; and (2) the models for rapid assessment
of ZE buffer systems and analyte separability [3–6] with a comprehensive program available for
free (Peakmaster 5.1, http://www.natur.cuni.cz/Gaš). An overview of other models, including the
steady-state models for ITP and IEF, which can predict the shape but not the evolution of steady-
state electrophoretic boundaries, is given in [7], and a survey of training software for electrophoresis
is provided in [8]. Computer simulations using dynamic models are more powerful than the other
alternative as they are applicable to any buffer configuration and follow the entire separation at every
step from start to finish according to the underlying transport laws. This not only yields the end result
but also provides snapshots along the way, which can be used to monitor and understand the evolution
of a particular process. While this is a huge advantage, it has one main disadvantage, namely that the
algorithms are much more complicated and therefore requires more computer power. Furthermore,
in contrast to high-performance liquid chromatography (HPLC) [9–12], electrophoretic simulations
are much more computationally intense in three distinct ways: (1) electrophoresis usually has more
components to deal with, (2) the separation efficiency is much higher requiring that component
concentrations be updated more often and at more numerous column positions, and (3) the equations
required to describe the processes are more numerous. This is less of an issue today than it was
20 years ago as computer hardware technology has developed significantly since then. Simulations
that used to take several days can now be performed within several hours, allowing more complex
and realistic simulations to be undertaken. It has only been within the past 5 years, over 20 years
since the first reports on computer simulations, that it is possible to simulate simple CE separations
under conditions that approximate those routinely employed in the laboratory.
In this chapter, we will discuss the theory underlying the use of dynamic simulation software
and the benefits that the use of simulations can provide, illustrated through a number of simulated
examples and examples taken from the literature over the past two decades. Finally, the most common
computer simulation software will be contrasted and discussed from a practical perspective providing
some guidelines as to the best way to approach a simulation problem.
17.2 BACKGROUND
Simulators with dynamic models are based on the description of algebraic acid–base equations and
continuity equations, which are partial differential equations in time and space that can only be solved
numerically using computers. Such models calculate the transport of each component through the
electrophoretic space as a result of electromigration, diffusion, imposed and/or electrically driven
bulk flow, solution-based chemical reactions such as protolysis and, if incorporated, also interaction
of solutes with the capillary walls. An example is presented in Figure 17.1. A complete analysis of
the temporal behavior of an electrophoretic system is thereby obtained and such models are thus
often referred to as dynamic or transient state models. The numerical solutions are called dynamic
simulations or dynamic modeling. Many dynamic models of various degrees of complexity have
been described in the literature [13–46]. Only two of them are mostly used these days, namely the
generalized PC-based models developed by Gaš and coworkers [30] and Mosher and coworkers
[34–36], models that permit simulation of the transient states in all major electrophoretic modes
Dynamic Computer Simulation Software for Capillary Electrophoresis 517
with predictive values for MBE, ZE, ITP, and IEF and are referred to as SIMUL5 and GENTRANS,
respectively, in the remainder of this chapter. Furthermore, the construction of a three-dimensional
stochastic simulation model for electrophoretic separations has been reported [47]. In contrast to
the dynamic models referred to above, this approach is based on the modeling of the trajectories of
each individual molecule, requires extremely powerful computers in order to compute the motion
of a statistically significant number of molecules, and is not further considered in this chapter. The
(a) (b)
30 5.0
40 min 12.5 min FER MYO CYTC
15 2.5
3
H3PO4
0 0.0
0 1 2 3 0 1 2 3
30 5.0
20 min 10 min FER
5 MYO CYTC
4
15 3 2.5
H3PO4
Concentration (mM)
Concentration (mM)
0 0.0
0 1 2 3 0 1 2 3
30 5.0
10 min 9 7.5 min
5 6 78 10 NaOH
4 1112
15 FER CYTC
3 2.5
MYO
H3PO4
0 0.0
0 1 2 3 0 1 2 3
30 5.0
5 min 5 min
CYTC
15 NaOH FER
2.5
H3PO4 MYO
0 0.0
0 1 2 3 0 1 2 3
30 5.0
0 min 2.5 min
15 NaOH 2.5
FER MYO CYTC
Sample
0 0.0
0 1 2 3 0 1 2 3
Column length (cm) Column length (cm)
FIGURE 17.1 Computer predicted IEF of 10 carrier components and 3 proteins between 10 mM H3 PO4
(anolyte) and 20 mM NaOH (catholyte) in a bare fused-silica capillary using GENTRANS with (i) the
electroosmosis model that considers electroosmosis as function of pH and ionic strength according to exper-
imental data (electroosmosis model V of Reference 54), (ii) a 3 cm column divided into 600 segments
(x = 50 µm), (iii) boundary conditions with open column ends, (iv) a constant voltage of 10 V, and
(v) data smoothing. (a) Concentration profiles of carrier ampholytes at 0, 5, 10, 20, and 40 min of current
flow. (b) Dynamics of the three proteins at 2.5, 5, 7.5, 10, and 12.5 min (from bottom to top, respectively).
(c) Column properties at 10 min (solid lines) and 0 min (dotted lines) for component distributions, ionic strength,
pH, conductivity, and electroosmotic flow (from bottom to top, respectively). (d) Temporal behavior of current
density, pH at two column locations (70 and 100% of column length), detector profiles for the proteins at the
two detector locations, and detector profiles for all carriers and proteins for the two column locations (from top
to bottom, respectively). For constructing the detection traces, all analytes had a detector response of 1. The
cathode in panels a–c is to the right. The numbers in panel c refer to the pI values of the carrier ampholytes. S in
panel c denotes the initial sample and C, F and M in panel (d) refer to CYTC, FER, and MYO, respectively.
(a) (d)
200 50
EOF (µm/s)
EO (µm/s)
100 25
S
0 0
0 1 2 3 0 20 40
0.6 60
Cond.(S/m)
density (A/m2)
Current
0.3 30
S
0.0 0
0 1 2 3 0 20 40
14 14
S 100%
pH
7 7
pH
70%
0 0
0 1 2 3 Concentration 0 20 40
30 6
M F
IS (mM)
(mM)
MF C
15 S 3 C
70 % 100 %
0 0
0 1 2 3 0 20 40
Concentration (mM)
30 30
Concentration
NaOH
70 %
(mM)
H3PO4 100 %
15 15
S
0 0
0 1 2 3 0 20 40
Column length (cm) Time (min)
FIGURE 17.1 (Continued)
same is true for all multidimensional models that describe electrokinetically driven mass transport
in microfabricated chip devices, such as those of Ermakov et al. [48], Bianchi et al. [49], Chatterjee
[50], and Sounart and Baygents [51], as well as predict sample zone distortion in continuous flow
electrophoresis [52].
Dynamic simulation provides plentiful data of any given CE system. This is illustrated here with
the IEF separation example presented in Figure 17.1, which is performed in a fused-silica capillary
and comprises a sample composed of ten hypothetical biprotic carrier ampholytes and three proteins
[cytochrome c (CYTC, pI of 10.49), myoglobin (MYO, pI of 6.80), and ferritin (FER, pI of 4.40)]
bracketed between 10 mM phosphoric acid (anolyte) and 20 mM sodium hydroxide (catholyte).
A 3 cm separation space divided into 600 segments of equal length, 33% initial occupation of the
sample at the anodic capillary end (panel a, bottom graph), and a constant voltage of 10 V were used.
The pI values of the carrier ampholytes uniformly span the range 3–12 (pI = 1). For each ampholyte,
pK was 2, the ionic mobility was 3 × 10−8 m2 /Vs and the initial concentration was 2.0 mM. The
initial concentrations of the three proteins were 0.1 mM each. Physico-chemical input properties were
the same as used before [32,53] and the input data for the calculation of electroosmosis are those
for a fused-silica capillary and consideration of the impact of the ionic strength on electroosmosis
according to model V [54]. Simulation reveals that the carrier ampholytes are completely separated
within about 10 min of current flow (panel a) and the pH gradient is therefore established within
that time period (center graph of panel c). The data shown also reveal that carrier ampholytes at
the interfaces to anolyte and catholyte (nicely seen with the pI 3 and pI 12 carriers, respectively)
become gradually isotachophoretically adjusted and somewhat removed from the gradient. Owing
to electroosmosis, the liquid and thereby the whole zone structure is being transported toward the
cathode. With time, the catholyte and most carrier compounds are swept out at the cathodic capillary
end and the net cathodic transport is decreasing and eventually disappearing completely [53]. A
stationary steady-state zone configuration, which is characterized by an equilibrium between anodic
isotachophoretic zone transport and cathodic electroosmosis, is produced. Under the employed
conditions, the boundary between anolyte (phosphoric acid) and the most acidic carrier ampholyte
is predicted to become immobilized at 91% of column length (top graph in panel a). The graphs
presented in panel b depict the dynamics of the protein separation and are essentially those that are
seen in whole column imaging [36]. The data of panel c represent zone properties predicted for
the 10 min time point (solid line graphs) compared to the corresponding initial distributions (dotted
line graphs). Data of other time points are not displayed, as this would complicate the figure. The
electroosmotic data of each column segment presented in the top graph of panel c illustrate that the
IEF gradient functions as the driving fluid pump of the capillary. These data do not reflect a real
physical distribution of that property, but provide insight into the pumping activity of each zone.
The electroosmotic pumping activity within anolyte and catholyte are predicted to be significantly
smaller compared to that within the gradient system. Furthermore, computer-predicted temporal
data for the net electroosmotic flow (EOF), current density, and various detector profiles are given
in panel d. EOF is thereby shown to vary with time (even when the separation is executed under a
constant current density, as is shown with the broken line graph, which was obtained for a constant
20 A/m2 ). The same is true for the current density under the simulation conditions of constant voltage.
Detector profiles for pH, the proteins (e.g., absorbance at 280 nm), and the carriers and proteins (e.g.,
absorbance at 200 nm) are given for two column locations, namely 70% of column length (solid
line graphs) and the cathodic capillary end (100% of column length, dotted line graphs). The entire
IEF zone structure is detected at the first location (and this can be obtained with a detector placed
between about 70% and 90% of column length), whereas at the end of the column, the anolyte and
the most acidic part of the gradient is not detectable at this location. Information of this kind is useful
for designing experiments with detection at the capillary end, most notably with mass spectrometry
(MS). It should be appreciated that without the use of computer simulation software, it would be very
difficult to discern and delineate the different electrophoretic phenomena involved in the system in
Figure 17.1.
17.3 HISTORICAL CONTEXT

Shortly after computers became available to University-based research entities, separation scientists
at universities in the Czech Republic (Prague, supervision of J. Vacík), Switzerland (Bern, E. Schu-
macher), and Arizona, USA (Tucson, M. Bier) began to construct dynamic computer models for
electrophoresis with the goal of exploring the basics of electrokinetic separations in solution. The
early models of Moore [13], Ryser [14], Gaš [15], and Vacík and coworkers [16–18], although
restricted to strong electrolytes, allowed the simulation of ZE, ITP, and MBE configurations and
can be regarded as the first dynamic electrophoretic models. These efforts were associated with the
CE studies of the 1960s and 1970s, which led to ITP analyses in polymer capillaries. First dynamic
models predicting the behavior of strong and weak electrolyte systems were developed in the 1980s,
including those of Bier et al. [1,19,20], Radi and Schumacher [21,22], Roberts [23,24], and Schafer-
Nielsen [25]. Except for [21,22], they were applied to all basic electrophoretic modes, including
IEF. The model of Radi and Schumacher [21,22] is unique in using the kinetic constants of the asso-
ciation and dissociation reactions to describe a chemical equilibrium. The model of Bier et al. [1],
which led to a unified view of all basic electrophoretic modes, was extensively used to characterize
a large number of electrophoretic configurations (for overviews see [7] and the detailed monograph
of Mosher et al. [2]) and also simulates the behavior of proteins using effective and mean square
valences that are determined from titration data [26]. Furthermore, the first model with incorporated
EOF together with the transport based on electromigration and diffusion is that of Dose and Guiochon
[27]. This approach, in which strong electrolytes are considered only and EOF is treated as constant
plug flow, has been applied to the modeling of capillary ZE and ITP. The model of Roberts includes
two spatial dimensions, whereas all others are one-dimensional approaches [23,24]. It is important
to note that all these efforts were undertaken with hybrid or digital mainframe computers.
With the advent of PCs and the increasing popularity of capillary electrophoretic separations and
analyses in fused-silica capillaries, many other dynamic simulators were developed in the 1990s,
including those of Gaš et al. [28–30], Mosher and Thormann [31–36], Ermakov et al. [37–39], Shimao
[40,41], Schafer-Nielsen [42], Martens et al. [43], Beckers and Boček [44], Ikuta and Hirokawa [45],
and Sounart and Baygents [46]. Gaš et al. [28,29] extended their early model with incorporation of
weak electrolytes, an approach that recently resulted in a comprehensive package that can account
for any number of mono- and multivalent electrolytes and ampholytes (SIMUL5, [30]). The PC
adapted model of Mosher and Thormann, GENTRANS, which is based on the dynamic simulator
of Bier et al. [1,19,20] and was extended for a more realistic treatment of proteins [31], application
of plug flow [32] and in situ calculation of EOF using wall titration data as input [33,34], can
handle strong and weak electrolytes, simple univalent ampholytes, peptides and proteins (currently
limited to 150 components total), and be executed at voltage gradients that are typically employed in
experimental work [35,36]. It is interesting to note that the models in current use are again promoted
by researchers from the same regions from which the early dynamic models originated, namely
Prague, Czeck Republic (B. Gaš), Bern, Switzerland (W. Thormann), and Tucson, Arizona, USA
(R.A. Mosher).
17.4 THEORETICAL ASPECTS OF SIMUL5 AND GENTRANS

17.4.1 MATHEMATICAL MODEL
The generalized models SIMUL5 and GENTRANS are simulation programs that predict the impact
of current flow on a specified distribution of electrolytes. They are composed of a set of balance
laws governing the transport of components in electrophoretic separations as was originally devel-
oped in the 1980s by Bier et al. [1,19,20], later detailed in the monograph of Mosher et al. [2] and
recently reported by Hruška et al. [30]. They comprise a coupled set of nonlinear partial differen-
tial equations describing the appropriate balance laws and algebraic equations describing chemical
equilibria, which includes an unsteady electromigration–diffusion equation for each component, a
charge balance with inclusion of the diffusion current (for importance, see [55]), the electroneu-
trality approximation, expressions for dissociation–association equilibria of weak electrolytes and
amphoteric compounds, and a model for calculating protein mobilities as a function of pH and ionic
strength (GENTRANS only). Because dissociation–association reactions are fast compared with the
mass transport [22], ion concentrations are constrained by a coupled set of mass action relations,
namely the dissociation of water, and the dissociation–association equilibria of the components.
Flow (constant or time dependent) is incorporated by adding an additional term to the flux equation.
The assumed plug flow with cross-sectional uniformity does not contribute to any boundary and
zone dispersion. GENTRANS also includes in situ calculation of electroosmosis using wall titration
data as input [33,34]. For each column segment, electroosmosis is calculated with the use of a pH
and ionic strength dependent electroosmotic mobility and the voltage gradient. Then, in analogy to
Darcy’s equation, which is valid for pressure-driven flow, the bulk capillary flow is taken to be the
average of all of the segment flows.
Both models are one-dimensional and based on the principles of electroneutrality and con-
servation of mass and charge. Isothermal conditions are assumed and relationships between the
concentrations of the various species of a component are described by equilibrium constants. Com-
ponent fluxes are computed on the basis of electromigration, diffusion, and convection (imposed flow
and/or electroosmosis). In GENTRANS, electrophoretic mobilities of small molecular mass compo-
nents are considered to be independent of the ionic strength and temperature, but vary as a function
of pH. SIMUL5 has an optional feature to account for the influence of the ionic strength on ionic
mobilities and electrolyte activities. Furthermore, to save computational time, SIMUL5 calculates
only the part of the column where considerable changes from the initial values are expected.
17.4.2 NUMERICAL IMPLEMENTATION

The complete partial differential equations are solved numerically. The spatial region is overlaid with
a uniform set of grid points or segments of length x for which all properties are calculated for each
time step. The numerical treatment involves replacing the spatial derivatives in the partial differential
equations by finite difference approximations that yields a set of ordinary differential equations with
time as the independent variable. These equations are solved using sophisticated algorithms and
computer codes that enable one to use variable step sizes for the time discretization and thus control
error growth. SIMUL5 solves the equations using the Runge-Kutta and predictor-corrector method
by Hamming [30], whereas GENTRANS is based on the fifth order Runge-Kutta-Fehlberg time step
and second order central difference spatial discretization using DAREP simulation language [2,20],
an approach that was later referred to as CSD algorithm in the work of Sounart and Baygents [46].
In both models, the algebraic equations are treated with the Newton’s iteration method. Eventual
differences in performance of the two simulators have not been elucidated thus far although there
appears to be no major difference in the predictions between the two systems for a relatively simple
simulation of the capillary ZE separation of a number of inorganic and organic anions as shown
in Figure 17.2. It should be noted, however, that there are some practical differences, namely the
actual simulation time using GENTRANS was approximately 40 min, with another 20 min required
to process the data and import it to graphing software, while SIMUL5 required approximately 3 h to
perform the same simulation, although did not require postsimulation data processing to visualize
the results.
17.4.3 BOUNDARY CONDITIONS

For each simulation, the boundary conditions have to be specified. As the models rely on cross-
sectional uniformity, there is only one spatial dimension, the separation axis x to be considered.
For a constant voltage simulation, the boundary conditions on the potential φ (x,t) are φ (0,t) = V
and φ (L,t) = 0 where V is the potential and L the column length. Boundary conditions on the
component concentrations at the column ends, that is, the permeabilities for the components, vary
with electrophoretic technique. Both models feature the conditions for open column ends, which
allow mass transport into and out of the separation space. After each step and for each component, the
implemented algorithm sets the concentration at the first (last) mesh point equal to that of the second
(next-to-last) mesh point. This approach accounts for changing boundary conditions at the column
ends that occur when, for example, sample components are leaving the separation space. Simulations
with GENTRANS can also be executed with fixed concentrations at column ends. Experimentally,
this is equivalent to the presence of large electrolyte reservoirs at each column end. This choice is
generally used (1) for ZE, ITP, and MBE simulations in which no sample components reach the
column ends; (2) for ITP and MBE in which no buffer constituents from one electrode vessel reach
the opposite side of the separation space; and (3) for IEF simulations in the presence of anolyte and
catholyte whose concentrations remain unchanged at the column ends during the investigated time
interval. Conditions with column ends that are impermeable to any buffer and sample compounds
can also be selected. It represents a configuration having the column ends only permeable for OH−
and H+ at cathode and anode, respectively, and is equivalent to an IEF experimental system in which
0.010 0.010
Concentration (mM)
Concentration (mM)
1 (c) (f)
0.008 23 4 0.008
5 6 7 8 9 10 1 23 4 5 6 7 8 9 10
0.006 0.006
0.004 0.004
0.002 0.002
0.000 0.000
–0.002 –0.002
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
9 9
0.010 10 876 0.010 10 8 76
Concentration (mM)
Concentration (mM)
4 3 21 4 3 21
0.008 5 (b) 0.008
5 (e)
0.006 0.006
0.004 0.004
0.002 0.002
0.000 0.000
–0.002 –0.002
0.00 0.02 0.04 0.06 0.08 0.10 0.00 0.02 0.04 0.06 0.08 0.10
Column position (m) Column position (m)
0.010 1-10
0.010 1-10
Concentration (mM)
Concentration (mM)
0.008 (a) 0.008 (d)
0.006 0.006
0.004 0.004
0.002 0.002
0.000 0.000
–0.002 –0.002
0.00 0.02 0.04 0.06 0.08 0.10 0.00 0.02 0.04 0.06 0.08 0.10
Column position (m) Column position (m)
FIGURE 17.2 Simple capillary ZE separation of a mixture of inorganic anions and carboxylic acids in a
MES/HIS electrolyte using SIMUL5 (panels a–c) and GENTRANS (panels d–f) under identical conditions (no
EOF, same PC). Data were produced using a 10 cm length that was divided into 4000 segments (x = 25 µm)
with the simulation in GENTRANS featuring variable column ends and smoothing. A constant voltage of –2000
V (200 V/cm) was applied and the sample initially occupied 1% of the capillary length. Panels a and d show the
initial distribution of the inorganic and organic anions. Panels b and e show the column distributions 30 s after
application of the voltage. Panels c and f show the detector trace constructed with the detector placed 9.5 cm
from the injection end (segment 3800). For constructing the detection trace in GENTRANS, all analytes had
a detector response of 1. The electrolyte consisted of 20 mM MES (pKa = 6.095, µep = 28 × 10−9 m2 /Vs),
20 mM HIS (pKa = 6.04, µep = 44.7 × 10−9 m2 /Vs), pH 6.06. The sample consisted of 0.01 mM of the
following anions prepared as sodium salt (µep = 51.9 × 10−9 m2 /Vs) in BGE. Analytes are 1 = sulphate (µep
= 82.9 × 10−9 m2 /Vs), 2 = chloride (µep = 79.1 × 10−9 m2 /Vs), 3 = nitrate (µep = 75.1 × 10−9 m2 /Vs),
4 = formate (pKa = 3.752, µep = 56.6 × 10−9 m2 /Vs), 5 = iodate (µep = 42.0 × 10−9 m2 /Vs), 6 = acetate
(pKa = 4.756, µep = 42.4 × 10−9 m2 /Vs), 7 = propanoate (pKa = 4.874, µep = 37.1 × 10−9 m2 /Vs), 8 =
butyrate (pKa = 4.820, µep = 33.8 × 10−9 m2 /Vs), 9 = valerate (pKa = 4.842, µep = 31.6 × 10−9 m2 /Vs),
and 10 = caproate (pKa = 4.857, µep = 30.2 × 10−9 m2 /Vs).
electrodes are used to define the ends of an electrophoresis chamber. It is also quite similar to an
IEF arrangement in which (1) a cation exchange membrane is used to isolate the anode chamber
from the separation space, and the anolyte is an acid, and (2) an anion exchange membrane is used
to isolate the cathode chamber from the separation space, and the catholyte is a base. Furthermore,
GENTRANS also permits the specification of mixed boundary conditions. This selection allows the
user to specify a boundary condition, either no transport or free transport of each component at both
left and right boundaries.
17.4.4 EXECUTION OF A SIMULATION

Initial conditions that must be specified for a simulation include (1) the distribution of all components,
(2) the pK and mobility values of the buffer and sample constituents, (3) the diffusion coefficients and
charge tables of the proteins (GENTRANS only), (4) the electroosmotic input data [constant velocity
286/287/12MHz
386/387/40MHz
PIII/600MHz
486/33/MHz
PIV/1.9GHz
PIV/2.4GHz
PII/300MHz
486/50MHz
486/66MHz
P/120MHz
P/233MHz
1e+5
1e+4
Execution time (min)
1e+3
1e+2
1991
2002
1e+1
PC chip configuration
FIGURE 17.3 Execution time of GENTRANS under MS-DOS (MS Fortran version) as function of chip and
clock speed assessed with PCs that were equipped with Intel chips for a 40 min simulation of the IEF example
of Figure 17.1 under a constant current density of 10 A/m2 , a constant net EOF of 15 µm/s toward the cathode
and fixed concentrations at column ends. A 350-fold decrease in execution time was observed with the PCs
purchased during the 11-year time period. Using the FTN77 Fortran version on the Pentium PCs, execution
time was found to be about half compared to that with the MS Fortran version.
value (SIMUL5 and GENTRANS) or as wall titration data for in situ calculation of electroosmosis
(GENTRANS only)], (5) the magnitude of constant voltage or constant current density, (6) the
duration of power application and the number of time points (frequency) of data storage, (7) the
column length as well as its segmentation, and (8) the species permeabilities (boundary conditions)
at the ends of the separation space (GENTRANS only). The programs output concentration, pH,
conductivity, ionic strength, and flow (GENTRANS only) distributions, and allow the presentation
of these data either as profiles along the column at specified time intervals or as temporal data,
which would be produced by detectors at specified column locations, that is, segment numbers.
Furthermore, they output the current density and the net flow (GENTRANS only) as functions of
time. SIMUL5 also has the added option of viewing the results as an animation allowing simple
visualization of the progression of the electrophoretic system.
The programs are executed on Pentium computers typically running at 233–3000 MHz. Exe-
cution time, which can vary between a few seconds and a full month, mainly depends on the PC’s
clock speed (Figure 17.3), the complexity of the electrophoretic situation (mainly the number of
components), the applied power level, and the number of column segments. SIMUL5, which is
free and can be downloaded from the web (http://www.natur.cuni.cz/Gaš), features a comfortable
windows environment for data input, data evaluation, and visualization of the ongoing simulation,
as illustrated in Figure 17.4. Furthermore, SIMUL5 features an inbuilt database of mobilities and
pKa values of common components as well as visualization of a completed simulation in a movie
format. The GENTRANS program is somewhat older, runs in the DOS environment and exists as
two versions, a low-resolution version for up to 600 segments (written in MS FORTRAN), and a
high-resolution version for up to 100,000 segments (FTN77 FORTRAN). The use of GENTRANS
can be learned in our research laboratories at the Universities of Bern and Tasmania.
FIGURE 17.4 Screen capture of the windows environment of SIMUL5 performing exactly the same simu-
lation as displayed in Figure 17.11 with a current density of 2000 A/m2 . The main panel displays the current
simulation progress, and is currently only displaying the column distributions for potassium, sodium, and
lithium. The vertical line at a column position of 0.0075 m is the right-hand calculation boundary, with the
left-hand calculation boundary at a position of 0 m. The second vertical line at 0.0095 m is the position of
the detector. The top left corner features the simulation parameters, such as time, voltage, current, and so on.
The top right displays the physico-chemical properties of the buffer and sample components.

A considerable proportion of the published reports on the use of dynamic computer simulation
software have been on the development of the simulation software itself and the agreement of the
simulation with real experimental data. This is an important and essential part of the development
of any theoretical system for without agreement between what is obtained in the laboratory, the
knowledge gained from the simulation becomes meaningless. These applications will not be dis-
cussed here, instead, we will focus on the way in which computer simulations have been used to
increase our understanding of the system and the field of electrophoretic separations. The early work
is summarized in [7] and the book of Mosher et al. [2], which appeared in 1992. In examining the
literature published thereafter, it has been dominated by application into a number of key areas,
particularly in ITP and MBE, IEF and ZE, as well as in understanding various forms of stacking
in ZE, and will be discussed under these categories. It is interesting to note that the application of
simulation software seemed to peak in the mid 1990s with a number of groups working in this field,
with a noticeable decline after this, until a few years ago when there was renewed interest around
the world. While one may speculate about the causes, it is worthwhile noting that the past few years
have seen an increasing number of applications of computer simulations in electrophoresis for two
main reasons. First, with the advent of faster computers, it is possible to simulate more realistic
and complex separations thus improving the quantitative prediction ability of the software and the
ability to simulate a range of separations that were previously not possible. Second, the development
of simple to use windows based software, such as SIMUL5, means that computer simulations are
no longer restricted to highly trained researchers, but can be used by almost anyone with a basic
knowledge of electrophoresis. It is anticipated that this area will see considerable growth over the
next few years as new and more powerful computers are used in conjunction with more sophisticated
simulation software.
17.5.1 ISOTACHOPHORESIS AND MOVING BOUNDARY ELECTROPHORESIS

ITP is one of the oldest electrophoretic separation techniques, and as such there have been a number
of simulations on ITP and the behavior of the boundaries, although many of these have been used
for the verification of the simulation software and for educational purposes [2,7,8]. In one of the
first nontrivial studies on ITP boundaries, Mosher et al. [56] provided the first precise description
of enforced migration due to changes in pH between the leading and terminating electrolyte. A
boundary with self-sharpening properties was created by a conductivity decrease across the bound-
ary in the direction of migration due to the pH gradient, which could be used to force analytes to
migrate between the two electrolytes as a result in changes in ionization (and hence, mobility). They
also showed that experimentally observed “bumps” were not necessarily artifacts due to adsorp-
tion/desorption phenomena, but were due to changes in conductivity and hence a change in electric
field. A similar study was later undertaken by Gaš et al. [28] using an early version of the SIMUL5
software. They also simulated the occurrence of “bumps” due to the presence of H+ or OH− and the
migration inversion of analytes depending on the pH of the electrolytes used. They concluded that in
a well-buffered system, migration inversion could occur only for analytes that had a pKa difference
of approximately 0.3–1.0 pH units.
The impact of hydrodynamic flow upon the formation and stabilization of ITP zones has been
examined by Deshmukh and Bier [57]. They modified the GENTRANS software to account for the
addition of counter flow and found that it had no significant impact on the shape of the ITP boundary,
but did have an impact on the position and could be used to effectively immobilize a particular
ITP boundary. This, however, can be a problem when the counter flow is applied at the beginning
of the separation as forming zones that migrate with a velocity lower than the counter flow were
found to be pushed from the capillary, as illustrated in Figure 17.5. This explained experimentally
observed results for the ITP stacking of serum components in which some components were not
stacked appropriately. Subsequently, Mosher et al. [33] introduced equations to account for EOF,
and in contrast to other available software, to also account for the variation of EOF with pH by using
titration data. This was later used by Thormann et al. [58] to examine the effect of EOF on ITP and
illustrate the difficulties of performing anionic ITP in unmodified fused-silica capillaries. Particularly
interesting was the generation of a stationary anionic ITP boundary in which the boundary velocity
was perfectly balanced by the EOF. Subsequent work by Caslavska and Thormann [59] simulated
bidirectional ITP in which the EOF was used to mobilize the zones past the detector. The use
of high pH electrolytes allowed both cationic and anionic ITP zones to be detected, but using a
lower pH caused the formation of stationary anionic ITP zones that could not be detected without the
application of hydrodynamic flow. Importantly, simulation data agreed well with the results obtained
by experiment. The same was found to be true for the prediction of ITP in polymethylmethacrylate
(PMMA) capillaries [54]. Furthermore, very recent work by Caslavska and Thormann [60] revealed
that fused-silica capillaries double coated with Polybrene and poly(vinylsulfonate) feature a strong
cathodic EOF across the entire pH range such that bidirectional ITP zone patterns at acidic pH reach
the detector without the addition of imposed flow.
In a series of work, Mosher et al. [26,31,61] expanded GENTRANS to be applicable for proteins.
As proteins are complicated molecules with a large number of ionizable groups, the mobility changes
significantly as a function of pH and as such a table of pH titration data and diffusion constants is
necessary to calculate the mobility of the protein [26]. After their initial work demonstrating the
applicability to proteins, two algorithms to account for the influence of ionic strength on protein
mobility were evaluated [31] with the best system subsequently used to examine the feasibility of
using titration data generated from the amino acid composition and free amino acid pKa values
25
Concentration, (x10+3)M
2 t = 0 min
T L
20
3,4
15
*
Electro- 1
10 Counter flow
migration
5
(*scaled 10x) (a)
0
25
t = 50 min
20 2
L
15
*
1
10
3
5 4
(*scaled 10x) (b)
T
0
25
2 t = 150 min
T
20
3 * L
15 1
10
5 4
(*scaled 10x) (c)
0
0 1 2
Column length (cm)
FIGURE 17.5 Selective washout of sample components during ITP due to application of a counter flow. The
counter flow applied was sufficient to immobilize the leading front. For the first 50 min, the bands are pushed to
the left due to the counter flow, followed by a movement to the right due to ITP adjustment. Leader (L): 20 mM
cacodylic acid/Tris, pH 7.1; Terminator (T): 20 mM β-alanine/Tris, pH 9.1. The four sample components were
(1) albumin, (2) Gly-Gly, (3) Asn, and (4) Gln. Their initial concentrations were 1.5 mM, 25 mM, 20 mM, and 20
mM, respectively. Simulation was performed under constant current density of 20 A/m2 with a separation length
of 4 cm with 400 segments and an applied counter flow of 0.47 × 10−5 m/s. (Reproduced from Deshmukh,
R.R. and Bier, M., Electrophoresis, 14, 205, 1993. With permission.)
[61]. When using theoretical titration data, a different ITP order of several proteins was obtained
to that observed by experiment, although this could be rectified by shifting the pH data to correct
for differences between theoretical and experimental pIs. Thormann and Mosher [62] simulated
two different anionic ITP systems and one cationic ITP system for the separation of proteins. In all
cases, there was qualitative agreement with experimental data, although the plateau concentrations
were typically lower than those observed experimentally. This was attributed to additional protein
interactions that are not considered in the model. In subsequent work, Thormann et al. [63] performed
simulations of ITP separations of serum albumin and compared these with experimental capillary ITP,
continuous flow ITP, and recirculating flow ITP in various instruments. Distortion of protein zones
found in the collected fractions of continuous flow ITP, which was not observed for low molecular
mass compounds and was not observed in the simulations, was attributed to various instrument
properties, such as high density/viscosity gradient formed at the buffer interfaces as well as the
design of the outlet ports. This was not observed in recirculating flow ITP where faster fluid flows
appeared to alleviate this problem.
In a different application, Dubrovčáková et al. [64] performed simulations of ITP separations

in which a neutral complexing agent (a cyclodextrin in this case) was added to the electrolytes.
Mechanistic insight into the movement of the ITP boundaries with addition of the cyclodextrin to the
sample, leading and terminating electrolyte, and various combinations, was gained. Simulations also
provided insight into the optimum concentration of cyclodextrin required to separate two analytes
with particular interaction constants.
Assessment and characterization of the evolution of moving and stationary electrophoretic bound-
aries is another topic of interest in simulation [2,7,44]. As an interesting application, Foret et al. [65]
investigated with GENTRANS the formation of a moving boundary induced by the liquid sheath
in coupling of CE with electrospray ionization MS. When the buffer in the capillary and the liquid
sheath contain different counterions, a boundary migrating from the capillary tip into the capillary is
formed. Before reaching the MS interface, analytes will suddenly face another buffer composition
(difference in pH, conductivity, ionic strength, etc.) and thus a change in migration that can result
in delays, loss of resolution, or inversion of migration order. The authors showed that the migrating
boundary can be either diffuse or sharp, depending on whether the effective mobility of the sheath
liquid counterion is higher or lower than that of the buffer in the capillary. In the most recent appli-
cation, Gebauer and Boček [66] used SIMUL to evaluate a new type of boundary that shows both
steady-state (sharp) and nonsteady-state (diffuse) properties. After discussing the theory behind the
new boundary, simulation results of a simple system comprising picrate and acetate were used to
demonstrate the validity of their theory, with the new type of boundary shown in Figure 17.6b. Figure
17.6a depicts conditions in which a normal steady-state boundary is formed.
17.5.2 ISOELECTRIC FOCUSING

The unique and powerful nature of IEF, which comprises a complex discontinuous configuration,
has ensured that there have been a number of simulations of IEF, again with the majority of those
being used to validate the simulation software and to ensure that there was reasonable agreement with
experimental data. The majority of these simulations were performed with a current density typically
much lower than that employed in the laboratory, which provided qualitative information about the
system, but could not be used in any quantitative capacity. Nevertheless, these simulations provided
new insights into the separation, stabilizing, and destabilizing processes of IEF in closed columns
without electrolytes, in configurations with acid and base as anolyte and catholyte, respectively,
and in systems with immobilized pH gradients (for a review of that simulation work see Mosher
et al. [2]). Simulations also revealed the basics of one-step capillary IEF performed in the presence
of electroosmosis in fused-silica capillaries [32,53]. In 2000, Mao et al. [35] presented the first
results with a new version of GENSTRANS that could handle up to 150 components and simulate
IEF at experimental current densities. This allowed the simulation of a pH gradient spanning 7 pH
units from 3 to 10, with the pH gradient made from 140 ampholytes, and low molecular mass
amphoteric dyes as sample constituents. These simulation data were used in conjunction with whole
column capillary imaging to examine the mechanism and quantitative agreement between the two.
Both systems demonstrated the well-known transient “double-peak” approach to focusing with the
simulation software predicting complete focusing being achieved 5 min after current application,
while experimentally it was observed to be 6 min. This was remarkable agreement given that the input
data (pK and mobility values) of carrier ampholytes and the dyes used were not accurately known and
that the composition of the commercial ampholyte mixture is largely unknown. A subsequent study
by Mosher and Thormann [67] revisited the focusing and stabilization phases of IEF. The simulation
results performed at 300 V/cm suggested that focusing is complete within less than 10 min and is
followed by a lengthy stabilization phase (up to 7000 min), which is characterized by changes that
progress from the column ends toward neutrality and the formation of nonlinear pH gradients, as
illustrated in Figure 17.7. The presence of electrolytes at the column ends disrupts the stabilizing
phase, with the degree of disruption being dependent on the concentrations of the acid and base
(a) 12
10 Ac
Concentration (mM)
8 Pic
(b) 12
10 Ac
Concentration (mM)
8 Pic
0
50 60 70 80
Column length (mm)
FIGURE 17.6 Simulated picrate (solid line) and acetate (dotted line) concentration profiles evolved after 25 s
of electrophoresis from initial sharp boundaries between solutions in which the molar fraction of picrate in the
front and rear zone was (a) 0.1 and 0.8 and (b) 0.8 and 0.1, respectively. The simulation was performed with
6000 segments (x = 16 µm) at a current density of 4000 A/m2 . (Reproduced from Gebauer, P. and Boček, P.,
of anolyte and catholyte, respectively. The data suggest that it is impossible to reach a completely
stationary steady-state pH gradient when using carrier ampholyte solutions, and that the only way
to achieve this is through the use of immobilized pH gradients such as those of Immobiline gels.
As GENTRANS can also handle proteins, Thormann et al. [36] performed simulations of the
IEF focusing of hemoglobin variants and compared these with whole column imaging IEF (both
approaches at 300 and 600 V/cm). They simulated a broad range pH gradient from 3 to 10 with 20
ampholytes per pH unit (140 ampholytes over the entire gradient) and a short-range pH gradient
from 5 to 8 with 40 ampholytes per pH unit. In contrast to IEF simulations of small molecular mass
compounds, the number of segments had to be increased from 1000 to 8000 in order to get smooth
protein peaks, which was due to the sharper peaks produced by proteins. The authors also noted that
in order to establish a smooth stepless pH gradient at experimental field strengths, it was necessary to
use 40 ampholytes per pH unit to establish the focusing gradient, while previously most simulations
had used at most 20. Again, agreement between experimental results with whole column imaging
was obtained supporting the use of this software for studying IEF.
In their latest work, Thormann and Mosher [68] have used GENTRANS to examine IEF when
sodium chloride is added to the catholyte at the beginning of the separation. This approach is a
single-step approach to post focusing mobilization of the IEF gradient by the addition of salt to the
cathode. Results were shown to demonstrate that proteins still focus at a pH near their pI, but that the
anion flux from the addition of chloride caused the pH to be slightly lower giving all ampholytes a
6.0 9
Concentration (mM)
(a) 6.6 7.4 (b) 10
4.5 10,000 8
7 0 104
120 3.0
pH
4
1000 6 10
3.5 10
5
10
0.0
4
0 1 2 3 4 5 0 1 2 3 4 5
Concentration (mM)
90
10,000 min 6.6 7.4
60
6.6 7.4
1000 min
30
6.6 7.4
10 min
0 1 2 3 4 5
Column length (cm)
FIGURE 17.7 Computer-simulated distributions of the 140 carrier components and the three dyes for the pH
5–8.5 gradient system after 10, 1000, and 10,000 min of constant voltage application. The numbers refer to the
pI values of the dyes and the arrowheads mark their locations. Successive graphs are presented with a y-axis
offset of 30 mM. The insets a and b depict the concentration profiles of the pI 6.6 and 7.4 amphoteric dyes and
the pH profiles, respectively, at the indicated time points. Simulations were performed with x = 50 µm and
having column ends that are impermeable to any sample and carrier compounds. (Modified from Mosher, R.A.
and Thormann, W., Electrophoresis, 23, 1803, 2002. With permission.)
slight positive charge and hence movement toward the cathode. The rate of movement was found to
be dependent on the concentration of chloride added, with high concentrations causing more rapid
movement. This also caused a gradual pH loss at the anodic end, while the cathodic end (where the
salt was) remained unchanged causing a change in the profile of the pH gradient.
17.5.3 ZONE ELECTROPHORESIS

Simulations of capillary ZE where the sample is the only discontinuity present have appeared in the
literature, but the reports are fewer than those of ITP and IEF, and are dominated by confirming the
validity of the simulation model, showing the property changes across ZE zones and demonstrating
the effect of electrophoretic sample concentration when the sample is administered in a matrix of
reduced ionic strength [2,7,33,34,54]. Simulation of ZE is useful for designing configurations with
indirect detection, for the assessment of system zones and the investigation of unusual peak forms
or peak shape distortions that cannot be handled with Peakmaster [30,69–72]. Particular worthy of
discussion here is the work of Ermakov et al. [73–75] on developing algorithms to describe interac-
tion of analytes with the capillary wall and the effect that this has on peak shapes in capillary ZE.
Simulations were undertaken with different levels of adsorption and desorption at different concen-
trations of analytes and also as a function of the adsorption capacity of the surface. Peaks shapes
were then compared with experimental results to gauge the source of interaction of analytes. Small
monovalent cations were found to have minimal interaction with the wall (symmetrical peak shapes
and no baseline shift) while large polycations were found to have slow desorption and produced
a tailed peak with a shift in baseline after the peak, consistent with simulation data in which ana-
lytes showed very slow, almost irreversible, desorption. Independent from the work of Ermakov
et al., Gaš et al. [76] studied the dynamics of peak dispersion in capillary ZE with wall adsorp-
tion using a simulation model that accounts for linear and nonlinear adsorption isotherms together
with slow and fast sorption rates. Furthermore, options in electrolyte systems for online combined
capillary ITP and ZE were investigated theoretically and verified by computer simulation using
GENTRANS [77].
17.5.4 TITP STACKING

The combination of capillary ITP and ZE has long been recognized as an ideal approach to improve the
sensitivity of problematic samples, particularly those with significant ionic strength. For biological
and environmental samples such as serum, urine, and seawater, the presence of high concentrations of
chloride is perfect to function as a suitable leading electrolyte for the ITP step, and all that is required
is to ensure that an appropriate terminator is used and that there is sufficient time to complete the
ITP stage and then to destack and separate the components by capillary ZE. tITP systems with
samples naturally abundant in chloride have been the basis of several significant experimental and
computational studies, the most significant belonging to Gebauer et al. [78] for the analysis of
biological samples by tITP, or sample self stacking as referred to by these authors. Their first report
in 1992 presented a simple theory to provide explicit description of various zone parameters such
as detection time, variance, and resolution, which was confirmed by computer simulation using
the GENTRANS software. This was later continued to look at the concentrations of components
required to induce tITP stacking, with the concentration of the sample macrocomponent or stacking
ion required to be above a critical concentration [79]. This concentration is a function of the buffer
co-ion concentration in the sample and the mobilities of the buffer co-ion, sample macrocomponent,
sample minor component, and the counterion. Simulations were undertaken to show when stacking
does and does not occur as a function of this ratio. Results were also presented illustrating that
the closer the mobility of the minor component to the stacking macrocomponent, a higher critical
concentration of macrocomponent is required to induce efficient stacking. Later work focused on
simulating the effect of multiple macrocomponents upon stacking and conditions under which one of
them would function as a destacker [80]. There was again found to be a critical concentration of both
the stacking and destacking macrocomponents, with this ratio only a function of the mobilities of the
ions within the system. Above this ratio, the destacker is actually stacked also facilitating stacking
of the minor components. Using the concepts of this system, Křivánková et al. [81] performed
simulations to examine the conditions required to induce tITP stacking of minor components in
serum (citrate, maleate, and acetoacetate), at various ratios of the destackers (phosphate and lactate),
with simulation results for different amounts of salt in the sample shown in Figure 17.8. Simulation
data demonstrated that 50 mM Cl or more was sufficient to induce ITP stacking in all possible
phosphate and lactate concentrations found in human serum and could therefore be used for the
analysis of plasma and serum samples from patients with a number of disorders. Their latest work
used SIMUL to examine the properties of borate buffers and ways in which tITP stacking could be
induced for samples with a high concentration of NaCl [82]. They examined two approaches, first, the
addition of a suitable terminating ion (MOPS) to allow borate to function as the leader, and second,
the use of ammonium as the counterion that reduced the mobility of hydroxide, allowing chloride
to function as the leader, and borate to function as the terminator. Both approaches were able to
provide suitable tITP stacking of minor components and were found to agree well with experimental
approaches.
In the area of environmental analysis, Hirokawa et al. [83] used SIMUL to simulate the electroki-
netic injection of cations with and without the hydrodynamic injection of potassium acetate. Not
20 1.0 20 (b) 1.0

Concentration (mM)
Concentration (mM)
(a)
Concentration (mM)
Concentration (mM)
2 min. 0.8 4 min. 0.8
15 15
L 0.6 0.6
10 10 L
0.4 0.4
5 5 P CI
P CI 0.2 0.2
0 0.0 0 0.0
0 2 4 6 0 2 4 6
20 1.0 20 1.0
Concentration (mM)
Concentration (mM)
Concentration (mM)
Concentration (mM)
15 4 min. 0.8 6 min. 0.8
15 L
0.6 0.6
10 L 10
0.4 P 0.4
5 CI 0.2 5 CI
P 0.2
0 0.0 0 0.0
0 2 4 6 0 2 4 6
20 1.0 20 1.0
Concentration (mM)
Concentration (mM)
Concentration (mM)
Concentration (mM)
6 min. 0.8 8 min. 0.8
15 15
0.6 L 0.6
10 10
L 0.4 0.4
5 5 P CI
P CI 0.2 0.2
0 0.0 0 0.0
0 2 4 6 0 2 4 6
20 1.0 20 1.0
Concentration (mM)
Concentration (mM)
Concentration (mM)
Concentration (mM)
15 8 min. 0.8 10 min. 0.8
15
0.6 0.6
10 10
L 0.4 L 0.4
5 CI 5 CI
0.2 P 0.2
P
0 0.0 0 0.0
0 2 4 6 0 2 4 6
Column length (cm)
Column length (cm)
20 1.0 20 1.0
Concentration (mM)
Concentration (mM)
Concentration (mM)
Concentration (mM)
(c) (d) 15 min.
15 4 min. 0.8 0.8
15
0.6 0.6
10 L 10 L
0.4 CI 0.4
5 CI 0.2 5 0.2
P P
0 0.0 0 0.0
0 2 4 6 0 2 4 6
20 1.0 20 1.0
Concentration (mM)
Concentration (mM)
Concentration (mM)
Concentration (mM)
15 8 min. 0.8 20 min. 0.8
15
0.6 0.6
10 L 10 L
0.4 CI 0.4
5 CI 5
P 0.2 P 0.2
0 0.0 0 0.0
0 2 4 6 0 2 4 6
20 1.0 20 1.0
Concentration (mM)
Concentration (mM)
Concentration (mM)
Concentration (mM)
0.8 P 25 min. 0.8
15 12 min. 15
0.6 L 0.6
10 L 10
0.4 CI 0.4
P CI
5 0.2 5 0.2
0 0.0 0 0.0
0 2 4 6 0 2 4 6
20 1.0 20 1.0
Concentration (mM)
Concentration (mM)
Concentration (mM)
Concentration (mM)
16 min. 0.8 30 min. 0.8

15 15
0.6 0.6
10 10
0.4 CI 0.4
5 L CI 5 P
P 0.2 L 0.2
0 0.0 0 0.0
0 2 4 6 0 2 4 6
FIGURE 17.8 Simulated concentration profiles at the given four time points for the sample containing (a)
10 mM, (b) 20 mM, (c) 40 mM, and (d) 100 mM chloride. Left concentration axis relates to the macrocomponents
chloride (Cl, dotted line), phosphate and lactate (P and L, respectively, both thin lines). Right concentration
axis relates to the microcomponents citrate, malate, and acetoacetate (thick lines). Simulations were performed
with x = 62 µm at a field strength of 40.54 V/cm. (Reproduced from Křivánková, L., et al., Electrophoresis,
surprisingly, much better preconcentration was obtained for cations with a mobility faster than the
mobility of the electrolyte co-ion with the injection of potassium as this functioned as a leading ion
for tITP. Following this, they presented results for the simulation tITP stacking of iodide in seawater.
Simulation data showed that the sample of unbuffered seawater maintains a high pH during ITP and
therefore a suitable terminator with a low mobility at this pH was necessary, although it should be
noted that practically seawater is actually buffered slightly by the presence of carbonate and borate.
On the basis of these results, the authors used 500 mM MES at pH 6.0 as the terminating electrolyte
that showed excellent stacking of an analyte with a mobility of 30 × 10−9 m2 /Vs (approximating
the mobility of iodide after ion association with CTAC).
The only other report of note in the area of tITP is that by Schwer et al. [29] who used an early
version of SIMUL to examine tITP stacking with a discrete pH difference between the leading and
terminating electrolytes. Dissipation of the pH difference was used to increase the mobility of the
cationic terminator thus shifting the separation from ITP stacking to ZE separation. A simplified
system involving the injection of a high pH terminator before the sample using a low pH terminator
was developed and showed to be suitable for the stacking of a tryptic digest of α-caesin.
17.5.5 pH STACKING
An alternative approach to stacking components in highly saline samples is though the use of a pH
discontinuity between the sample and the electrolyte to stack ionizable analytes. This is similar to the
enforced migration method examined by Mosher et al. [56] discussed above, although in this case it
is transient in nature and is destroyed before migration by capillary ZE. Kim et al. [84] used SIMUL
to simulate the formation and movement of a pH boundary and its influence on stacking weak acids.
The pH boundary eventually dissipates allowing stacked analytes to be separated according to their
electrophoretic mobility. Simulations also explained the experimental observation of an improvement
in stacking of m-nitrophenol with increasing injection volume, which was related to the velocity of
the stacking pH boundary and the length of time before dissipation and hence, destacking of the
analytes.
Shim et al. [85] used their own program for the simulation of stacking weak acids (such as
fluorescein) in buffered and unbuffered saline samples. Buffered saline samples showed stacking via
tITP, while in unbuffered samples (NaCl in water) the rapid exit of the matrix ions from the sample
caused a localized pH change to maintain electroneutrality thus creating an additional pH boundary.
This pH boundary was restricted to the initial position of the sample injection and merged with
a migrating tITP boundary induced by the high concentration of Cl. Thus, analytes were initially
stacked on both sides of the sample zone and merged into a single focused zone in a fashion similar
to the transient “double peak” mechanism observed in IEF.
In the above two simulations, stacking was simulated for one analyte and the magnitude of the
pH boundary was only several pH units. Breadmore et al. [86] used GENTRANS to simulate the
stacking of 24 weak bases of different mobility and pKa in which the sample had a pH almost 7 units
higher than that of the electrolyte. The influence of the addition of a high concentration of sodium
chloride to the sample was also examined, with simulation results demonstrating that the movement
of the boundary can be influenced by the addition of sodium chloride to the sample. Upon the addition
of 100 mM NaCl, a significant reduction in mobility of the boundary was observed allowing bases
with a much lower mobility to be focused. However, one of the limitations of lowering the mobility
of the boundary in this fashion is that it causes prolonged migration of the boundary itself, which can
result in analytes migrating through the entire capillary as focused bands, as shown in Figure 17.9.
Care must therefore be taken to allow sufficient time for the boundary to dissipate and the analytes
to destack.
5000
0 mM NaCl
4000
2 2
3000
3
2000 1
1000 1.15 1.20 1.25 1.30 1.35
3
1 4
0
5000
4000 50 mM NaCl
Detector response
1,2
3000 3
1,2
2000
1000 1.40 1.45 1.50 1.55 1.60

3
4
0
5000
4000 100 mM NaCl
1,2,3 1,2,3
3000
2000
1000 1.65 1.70 1.75 1.80 1.85
4
0
0 1 2 3
Time (min)
FIGURE 17.9 Simulated detector response for the separation of four weak bases (pKa = 4.5) after stacking
with a transient pH boundary without and with the addition of NaCl to the sample zone. Buffer: 65.6 mM
formic acid, pH 2.85, with NaOH. Sample: 65.6 mM formic acid, pH 8.60 adjusted with NaOH, 0, 50, or 100
mM NaCl and the four bases (5 µM each). Insets show a close-up view of the peaks 1–3. Voltage 400 V/cm.
Detector location 5 cm. Peaks are (1) 25 × 10−9 m2 /Vs, (2) 20 × 10−9 m2 /Vs, (3) 15 × 10−9 m2 /Vs, and (4)
10 × 10−9 m2 /Vs. (Reproduced from Breadmore, M.C., et al., Anal. Chem., 78, 538, 2006. With permission.)
17.5.6 MISCELLANEOUS APPLICATIONS

Tesařová et al. [87] used a modified version of SIMUL, which they called SIMULMIC to simulate the
separation of neutral analytes in a system with a neutral cyclodextrin and anionic micelles. A number
of systems were examined in which various combinations of the inlet and outlet vials and the capillary
itself were filled with cyclodextrin. Simulation results were used to examine the micellar/cyclodextrin
boundary at various times and concentrations although no simulation results for a chiral separation
were reported. To the best of our knowledge, this is the only dynamic simulation of an electrokinetic
chromatography (EKC) separation to date.
Several reports have been presented to simulate capillary affinity electrophoresis (CAE) under
equilibrium (complex is formed with ligand in the electrolyte and migrates in an equilibrium state) and
nonequilibrium (ligand is not in the electrolyte allowing the complex to dissociate during separation)
conditions [88–90] using binding constants and association/dissociation rate constants. Because of
the added complexity of the additional equations, implementation often involves simplification by
assuming no changes in the electrolyte macrocomponents and hence no change in pH or conductivity,
thus only allowing more limited simulations to be performed. Nevertheless, these simulators can
successfully predict peak shapes in CAE, and as was demonstrated by Okhonin et al. [90] can be
used to determine the binding and rate constants by using nonlinear regression to fit the experimental
separation to that predicted with the CAE software.
SIMUL5 has been used to study oscillation patterns that occur in background electrolytes with
complex eigenmobilities when they are subjected to an electric field [91]. For a system composed
of 0.21 mM sebacic acid and 0.323 mM imidazole, simulation revealed the production of local
10−12 mM perturbations of the imidazole concentration, which exponentially grew as long as the
system was in the region of complex eigenmobilities. This unusual behavior could be experimentally
validated and is thus not a simulation artifact.

In performing a simulation, there are some points that the researcher must understand in order to
maximize the utility of their simulation. Selection of the simulation parameters, such as the number
of segments and the applied power level, is crucial for success. These aspects are visualized with
the following simple example, which represents the simulation of the ITP zone formation of sodium
between potassium (leader L) and lithium (terminator T) with acetate as the counterion (Figure 17.10).
This simulation was performed in a 1 cm column that was divided into 2000 segments of equal length
(x = 5 µm) and having a constant current of 200 A/m2 . All initial concentrations are 10 mM and
the sample occupies 5% of column length (between 5% and 10%, bottom graph of left panel of
Figure 17.10). Upon application of power, the sodium zone becomes adjusted within about 0.1 min
(right panel of Figure 17.10) to a concentration of 8.48 mM as it gradually penetrates into the space
originally occupied by the leader where it continues to migrate between potassium and lithium. The
same applies to lithium, the terminating constituent, which becomes adjusted to a concentration
of 7.59 mM. Thereafter, the entire zone structure is migrating at a constant velocity and without
change of the zone structure toward the cathode (left panel of Figure 17.10). The corresponding zone
distribution after 0.08 min (4.8 s) at 2000 A/m2 obtained with SIMUL5 is shown in Figure 17.4. The
migrating boundaries between the alkali metal ions are sharp, steady-state transitions whose widths
are dependent on the applied power level. At a low power, electrophoretic boundaries are rather
broad and easy to treat with a relatively low number of segments. Before the availability of fast
computers, this was the rule rather than the exception [2,7]. SIMUL5 and GENTRANS are capable
of handling high power levels. The example presented in Figure 17.11 illustrates the impact of current
density on the ITP zone boundaries in the configuration of Figure 17.10. Having 20 A/m2 (electric
field E = 1.79 V/cm) results in broad boundaries and simulation predicts that there is insufficient
sodium to produce an ITP zone with a plateau concentration (upper left graph of Figure 17.11). The
opposite is true with an increase of the current density to 200 A/m2 (E = 17.9 V/cm, upper right
graph), 2000 A/m2 (E = 179 V/cm, lower left graph), and 20,000 A/m2 (E = 1787 V/cm, lower
right graph), conditions with much sharper boundaries and 10-, 100- and 1000-fold, respectively,
faster solute transport. The latter two examples were simulated at power levels that are typically
used in experiments. ITP boundaries between the alkali metal ions of our example are predicted to
be in the order of 350, 35, and 3.5 µm, respectively (Figure 17.11, insets depict the sodium–lithium
transitions at elongated x-axis scales), that is, boundary width is indirectly proportional to the current
density which is in agreement with previous knowledge [92]. The transition between potassium and
20 20
0.8 min Sadjusted 0.10 min
10 10
T Tadjusted L Li+ Na+ K+
0 0
0 5 10 0 1 2 3
20 20
0.6 min 0.08 min
10 10
0 0
0 5 10 0 1 2 3
20 20
Concentration (mM)
Concentration (mM)
0.4 min 0.06 min
10 10
0 0
0 5 10 0 1 2 3
20 20
0.2 min 0.04 min
10 10
0 0
0 5 10 0 1 2 3
20 20
0 min 0.02 min
10 10
TS L Li+ Na+ K+
0 0
0 5 10 0 1 2 3
Column length (mm) Column length (mm)
FIGURE 17.10 Computer predicted ITP zone formation with sodium as sample (S) between potassium (leader
L) and lithium (terminator T) and having acetate as counter component and no electroosmosis. Data were
produced with GENTRANS using a column of 1 cm length that was divided into 2000 segments (x = 5 µm),
fixed concentrations at column ends and no data smoothing. A constant current density of 200 A/m2 was applied
and the sample initially occupied 5% of column length. Left-hand graphs depict concentration distributions at
0, 0.2, 0.4, 0.6, and 0.8 min after current application. Right-hand graphs depict the dynamics of the formation
of the migrating sodium ITP zone with concentration profiles after 0.02, 0.04, 0.06, 0.08, and 0.1 min. The
cathode is to the right. The mobilities of potassium, sodium, and lithium were taken as 7.91 × 10−8 m2 /Vs,
5.19 × 10−8 m2 /Vs, and 4.10 × 10−8 m2 /Vs, respectively. The pKa and mobility of acetic acid were 4.76 and
4.12 × 10−8 m2 /Vs, respectively.
sodium is somewhat sharper than that between sodium and lithium. The reason for this is the smaller
conductivity change across the latter boundary (Figure 17.12, right panel).
Using insufficient number of segments, there is a risk of numerical oscillations that are seen in
all profiles, including concentration, conductivity, and pH distributions. This is illustrated with the
data presented in Figure 17.12 that were generated with 2000 A/m2 . Strong oscillations are obtained
for a simulation with 1000 segments (x = 10 µm, left graph in Figure 17.12). With 2000 segments
(x = 5 µm, center graph in Figure 17.12), much smaller oscillations are predicted, whereas 4000
segments led to smooth transitions (x = 2.5 µm, right graphs in Figure 17.12). It is important to
realize, however, that this is at the expense of simulation time. For a total of 0.1 min electrophoresis
time, the simulations performed on a Pentium IV 2.4 GHz PC lasted 2.45, 7.20, and 20.68 min,
respectively, suggesting that doubling the grid requires 2.9-fold longer simulation time intervals.
20 20
12
20.0 A/m2 200 A/m2
t = 8.000 min t = 800 min
x = 10 µm 8
16 x = 10 µm 16
4
Concentration (mM)
12 12 0
6.600 7.300
acetate K+
8 K+ 8
Li+ Li+
4 4
Na+
Na+
0 0
0.0 2.5 5.0 7.5 10.0 0.0 2.5 5.0 7.5 10.0
20 20
12 12
2000 A/m2 20,000 A/m2
t = 0.080 min t = 0.008 min
8 x = 0.25 µm 8
16 x = 2.5 µm 16
4 4
Concentration (mM)
12 0 12 0
6.915 6.985 6.939 6.946
K+ K+
8 8
Li+ Li+
4 4
Na+ Na+
0 0
0.0 2.5 5.0 7.5 10.0 0.0 2.5 5.0 7.5 10.0
Column length (mm) Column length (mm)
FIGURE 17.11 ITP simulation of the system of Figure 17.10 at different constant current densities of 20 A/m2
(upper left graph), 200 A/m2 (upper right graph), 2000 A/m2 (lower left graph), and 20,000 A/m2 (lower right
graph) with data shown for 8 min, 0.8 min, 0.08 min, and 0.008 min, respectively. The insets depict the sodium-
lithium transitions on elongated x-axis scales. Data were produced with GENTRANS without smoothing using
1000, 1000, 4000, and 40,000 segments, respectively.
On the other hand, the simulation at 200 A/m2 and 1 min of electrophoresis time (Figure 17.10)
executed on a Pentium 233 MHz PC with 1000, 2000, and 4000 segments required 0.27, 2.0, and
16.5 h, respectively. For this case, execution time was found to increase about eightfold with doubling
of the mesh. This illustrates that simulation time intervals are also hardware dependent. It should
also be noted that the alternative to increasing the mesh size is to restrict the magnitude by which the
step size dt can change. This can be used successfully to overcome floating-point errors that result
from numerical oscillations, however, this also obviously results in an elongation in simulation time.
0.13 0.13 0.13

Cond. (S/m)
0.09 0.09 0.09
0.05 0.05 0.05
8.42 8.42 8.42

pH
8.27 8.27 8.27
8.12 8.12 8.12
12 12 12
∆x = 10 µm ∆x = 5 µm ∆x = 2.5 µm
Concentration (mM)
8 8 8
4 Li+ +
Na +
K 4 Li+ Na+ K+ 4 Li+ Na+ K+
0 0 0
6.6 7.0 7.4 7.8 6.6 7.0 7.4 7.8 6.6 7.0 7.4 7.8
Column length (mm) Column length (mm) Column length (mm)
FIGURE 17.12 ITP simulation data at 2000 A/m2 of the system of Figure 17.10 using 1000 (left graphs),
2000 (center graphs), and 4000 (right graphs) segments. Concentration, pH, and conductivity data shown are
for 0.08 min and were obtained with GENTRANS without application of data smoothing.
While no quantitative study has been undertaken, this approach can often be quicker than doubling
the number of segments.
Instead of an increased mesh, other algorithms have been tested in various laboratories as is
discussed in depth by Sounart and Baygents [46]. None of the proposed procedures, however, was
found to provide data without undesired boundary distortions in a much reduced time interval.
Instead, GENTRANS is having an optional smoothing algorithm that removes negative concen-
trations caused by numerical oscillations. Using smoothing will decrease accuracy and creates the
potential for nonphysical results to be produced. If smoothing is used and data are obtained that are
suspected to be nonphysical, that is, behavior not expected to occur in an experiment, the simulation
should be repeated without invoking smoothing, and the results compared to the suspect data. The
data presented in Figure 17.13 illustrate the impact of the smoothing option for the simulation of a
ZE separation of five dipeptides that are negatively charged in the CHES buffer at pH 9.5. Without
smoothing, peptides 2–4 show oscillations with negative concentrations at the rear ends of their peaks
(shown for 10 and 20 min of power application, left-hand graphs). These are not observed anymore
when smoothing is used (right-hand graphs). Smoothing is thereby shown that such a simulation can
be executed with a relatively large mesh size (x = 166.7 µm), that is, a low number of segments
that results in a rather fast simulation.
While there has been a significant improvement in computer hardware over the past decade, it
is still necessary to note that due to the intimate relationship between the length of the capillary,
number of segments, and current density or applied electric field strength, there are still some
laboratory experiments that are difficult to simulate. These are typically those that employ very high
current densities (approaching 45,000 A/m2 , which translates to a current of 88 µA in a 50 µm
ID capillary, or 200 µA in a 75 µm ID capillary) in which there is a self-sharpening discontinuity.
These are frequently encountered when trying to simulate various stacking methods, which are
difficult to simulate at realistic currents unless short capillaries, typically in the order of several cm,
with several thousands (typically 20,000) segments, are used. Longer capillaries or fewer segments
0.18 0.18
20 min 20 min
1 2 1 2
4 5 4 5
Concentration (mM) 3 3
Concentration (mM)
0.12 0.12
2,3 1 2,3
1 4 10 min 4 10 min
5 5
0.06 0.06
0 min 0 min
0.00 0.00
0.0 2.5 5.0 7.5 10.0 0.0 2.5 5.0 7.5 10.0
FIGURE 17.13 Simulation of the ZE separation of five dipeptides in a 20.0 mM CHES/9.457 mM NaOH
buffer at pH 9.5 and without electroosmosis using GENTRANS without (left-hand graphs) and with (right-hand
graphs) data smoothing. Data were produced using a column of 10 cm length that was divided into 600 segments
(x = 166.7 µm) and fixed concentrations at column ends. A constant voltage of –145 V (current density of
–100 A/m2 ) was applied and the sample was applied as Gaussian peak (0.1 mM peak height each, half width
of 10 segments) superimposed onto the buffer at 10% of column length. Sample distributions after 0, 10, and
20 min of power application are shown with a y-axis offset of 0.06 mM between graphs. The anode is to the
right. Key: 1, β-ala-his (pKa values: 2.73, 6.87, 9.73; mobility: 2.14 × 10−8 m2 /Vs); 2, lys-glu (pKa values:
2.93, 4.47, 7.75, 10.5; mobility: 1.96 × 10−8 m2 /Vs); 3, his-gly (pKa values: 2.40, 5.80, 7.82; mobility:
2.26 × 10−8 m2 /Vs); 4, asp-his (pKa values: 2.45, 3.02, 6.82, 7.98; mobility: 1.81 × 10−8 m2 /Vs); 5, asp-gly
(pKa values: 2.10, 4.53, 9.07; mobility: 2.46 × 10−8 m2 /Vs). The pKa and mobility values of the weak acid
CHES were 9.55 and 2.31 × 10−8 m2 /Vs, respectively. The mobility of sodium was 5.19 × 10−8 m2 /Vs.
cause undesirable oscillations, and while this can be avoided by lowering the current density, it again
deviates from the situation encountered in the laboratory. A version of GENTRANS has been created
that can handle up to 100,000 segments while SIMUL has no apparent restriction on the number
of segments, potentially allowing simulations to be performed with capillary lengths typically used
in the laboratory at experimentally used electric field strengths (up to 1000 V/cm), however as
noted above, doubling the number of segments increases the computational time by a factor of 4–8,
meaning current simulations with 20,000 segments taking 18 days, would stretch out to 72–144 days
(about 2.5–5 months!). While it is expected that the advent of newer computers will continue to push
the boundaries of simulation software, it will necessitate revision of the software itself, particularly
to make use of multiprocessor computers that are now becoming commonly available in desktop
computers.
Finally, it should be noted that there are a number of essential requirements that need to be
reported to allow others to repeat a particular simulation. Obviously, the software and the version
are important as are the input data used to describe the components (pKa values, mobilities, and
concentrations), but it is often the overlooked parameters that can mean the difference between
successfully repeating a simulation. The length of the capillary (L) and number of segments (either
as the number of segments, n, or the mesh size, x), the current density and the voltage, and the
initial step size (dt) and whether there are any restraints on these should always be given. Also
important is the position and distribution of the components, particularly the presence of boundaries,
which should be given with their position as well as their width. In principle, simulations can be
performed at constant voltage or constant current density. For simulations comprising boundaries
with a steady-state shape, such as ITP, simulations with constant current density rather than constant
voltage might be preferred as the width of a steady-state boundary is current dependent.

The state of dynamic computer simulation software has progressed significantly over the past two
decades. Software is available that will simulate electrophoretic systems including MBE, ZE, ITP, and
IEF under almost exactly the same conditions used in the laboratory, and this has been used to show the
detailed mechanisms of many of the fundamental phenomena that occur in a simple electrophoretic
separation. A detailed insight into a number of electrophoretic processes has been achieved and this
is having a significant impact on designing new and improved electrophoretic systems. Noticeably
absent is the inability of the current software to simulate EKC and electrochromatography (EC),
and while the underlying algorithms of the current software is applicable to these forms of CE, the
necessity of adding more equations and constants to describe the additional equilibria that necessitates
an increase in computation time, and is likely to be the reason this has not been implemented to date.
This will change over the coming years as improvements in computer hardware will result in a
decrease in total simulation times and will make it feasible to complicate the system by introducing
additional equilibria to account for other modes of electrophoresis making it possible to simulate
even more varied and complex CE separations. Furthermore, two- and three-dimensional dynamic
simulation models will have to be developed that permit simulations in capillaries and microchannels
of different shapes and geometries.
It is important to mention that dynamic electrophoretic simulations are relevant for separations
on any scale and instrumental format, including free-fluid preparative, gel, capillary, and chip elec-
trophoresis. Separations in the minute to subsecond time domains (Figure 17.11) can thereby be
assessed. Electrophoretic modeling is extremely flexible. In addition to the classical electrophoretic
modes (MBE, ITP, ZE, and IEF), it has been used to examine the behavior of systems containing
arrays of fixed charges such as immobilized pH gradients or the presence of ion exchange membranes
[1,2]. CE performed in capillaries and microchannels (chips) benefits from computer modeling as
simulation is well suited (1) to develop/modify buffer systems, including those employed for indirect
and conductivity detection (determination of the impact of the sample on background buffer composi-
tion); (2) to study the effects of sample composition on sample concentration/migration for different
injection procedures; (3) to determine the effect of a pH change on migration/separation; (4) to
investigate system peaks, unusual sample peaks, and buffer oscillations; (5) to explore combinations
of electrophoretic modes, including ITP/ZE separations and mobilization in IEF; and (6) to elucidate
the impact of EOF on separations in discontinuous buffer systems (ITP and IEF), to name a few.
Furthermore, it allows to predict the impact of various modifications on the electrophoretic behavior
of peptides and proteins, including genetic mutations/modifications, chemical modifications such as
deamidation during downstream processing, changes in amino acid composition, stripping of tightly
bound, charged small ion cofactors, and binding of metal ions. This and other areas of application,
including the prediction of enantioselective and other affinity separations, will be explored on a
broad basis by dynamic electrophoretic simulation in the not too distant future.
ACKNOWLEDGMENTS
The authors acknowledge the valuable contributions of Dr R. A. Mosher to the development of
GENTRANS and the discussions about SIMUL5 with Dr B. Gaš and V. Hruška. This work was
supported by the Swiss National Science Foundation and the Australian Research Council.
REFERENCES
1. Bier, M., Palusinski, O.A., Mosher, R.A., and Saville, D.A., Electrophoresis: Mathematical modeling
and computer simulation, Science, 219, 1281, 1983.
2. Mosher, R.A., Saville, D.A., and Thormann, W., The Dynamics of Electrophoresis, VCH Publishers,
Weinheim, 1992.
3. Reijenga, J.C. and Kenndler, E., Computational simulation of migration and dispersion in free capillary
zone electrophoresis. I. Description of the theoretical model, J. Chromatogr., 659, 403, 1994.
4. Reijenga, J.C. and Kenndler, E., Computational simulation of migration and dispersion in free capillary
zone electrophoresis. II. Results of simulation and comparison with measurements, J. Chromatogr.,
659, 417, 1994.
5. Jaroš, M., Hruška, V., Štědrý, M., Zuskova, I., and Gaš, B., Eigenmobilities in background electrolytes
for capillary zone electrophoresis: IV. Computer program PeakMaster, Electrophoresis, 25, 3080,
2004.
6. Gaš, B., Jaroš, M., Hruška, V., Zuskova, I., and Štědrý, M., PeakMaster—A freeware simulator of
capillary zone electrophoresis, LC-GC Eur, 18, 282, 2005.
7. Thormann, W. and Mosher, R.A., Theory of electrophoretic transport and separations: The study
of electrophoretic boundaries and fundamental separation principles by computer simulation, Adv.
Electrophor., 2, 45, 1988.
8. Reijenga, J.C., Martens, J.H.P.A., and Everaerts, F.M., Training software for electrophoresis,
9. Golshanshirazi, S. and Guiochon, G., Comparison between experimental and theoretical band profiles
in nonlinear liquid-chromatography with a binary mobile phase, Anal. Chem., 60, 2634, 1988.
10. Golshanshirazi, S., Ghodbane, S., and Guiochon, G., Comparison between experimental and theoretical
band profiles in nonlinear liquid-chromatography with a pure mobile phase, Anal. Chem., 60, 2630,
1988.
11. Guiochon, G., Golshanshirazi, S., and Jaulmes, A., Computer-simulation of the propagation of a
large-concentration band in liquid-chromatography, Anal. Chem., 60, 1856, 1988.
12. Guiochon, G. and Ghodbane, S., Computer-simulation of the separation of a 2-component mixture in
preparative scale liquid-chromatography, J. Phys. Chem., 92, 3682, 1988.
13. Moore, G.T., Theory of isotachophoresis development of concentration boundaries, J. Chromatogr.,
106, 1, 1975.
14. Ryser, P., Omegaphorese: Theorie und Anwendung einer Analysenmethode, Dissertation, University
of Bern, Bern, 1976.
15. Gaš, B., Modeling separation process in capillary isotachophoresis, Diploma Thesis, Charles
University, Prague, 1975.
16. Vacík, J. and Fidler, V., Dynamics of isotachophoretic separation, in Analytical Isotachophoresis,
Everaerts, F.M. (Ed.), Elsevier, Amsterdam, 1981, p. 19.
17. Fidler, V., Vacík, J., and Fidler, Z., Dynamics of isotachophoretic separation. 1. Computer-simulation,
J. Chromatogr., 320, 167, 1985.
18. Fidler, Z., Fidler, V., and Vacík, J., Dynamics of isotachophoretic separation. 2. Impurities in the
electrolytes and sample injection, J. Chromatogr., 320, 175, 1985.
19. Saville, D.A. and Palusinski, O.A., Theory of electrophoretic separations. Part I: formulation of a
mathematical model, AIChE J., 32, 207, 1986.
20. Palusinski, O.A., Graham, A., Mosher, R.A., Bier, M., and Saville, D.A., Theory of electrophoretic
separations. Part II: Construction of a numerical simulation scheme and its applications, AIChE J., 32,
215, 1986.
21. Radi, P., Numerische Integration der elektrophoretischen Transportgleichung, Diploma Thesis,
University of Bern, Bern, 1982.
22. Radi, P. and Schumacher, E., Numerical simulation of electrophoresis—The complete solution for
3 isotachophoretic systems, Electrophoresis, 6, 195, 1985.
23. Roberts, G.O., Electrophoresis modeling, USRA/RAI/ES73 Report, Marshall Space Flight Center,
Huntsville, 1984.
24. Roberts, G.O., Rhodes, P.S., and Synder, R.S., Dispersion effects in capillary zone electrophoresis,
J. Chromatogr., 480, 35, 1989.
25. Schafer-Nielsen, C., Steady-state gel electrophoresis systems, in Gel Electrophoresis of Proteins,
Dunn, M.A. (Ed.), Wright, Bristol, 1986, p. 1.
26. Mosher, R.A., Dewey, D., Thormann, W., Saville, D.A., and Bier, M., Computer simulation and
experimental validation of the electrophoretic behavior of proteins, Anal. Chem., 61, 362, 1989.
27. Dose, E.V. and Guiochon, G.A., High-resolution modeling of capillary zone electrophoresis and
isotachophoresis, Anal. Chem., 63, 1063, 1991.
28. Gaš, B., Vacík, J., and Zelensky, I., Computer-aided simulation of electromigration, J. Chromatogr.,
545, 225, 1991.
29. Schwer, C., Gaš, B., Lottspeich, F., and Kenndler, E., Computer simulation and experimental evalua-
tion of on-column sample preconcentration in capillary zone electrophoresis by discontinuous buffer
systems, Anal. Chem., 65, 2108, 1993.
30. Hruška, V., Jaroš, M., and Gaš, B., Simul 5—Free dynamic simulator of electrophoresis, Electrophore-
sis, 27, 984, 2006.
31. Mosher, R.A., Gebauer, P., Caslavska, J., and Thormann, W., Computer simulation and experimen-
tal validation of the electrophoretic behavior of proteins. 2. Model improvement and application to
isotachophoresis, Anal. Chem., 64, 2991, 1992.
32. Thormann, W., Molteni, S., Stoffel, E., Mosher, R.A., and Chmelík, J., Computer modeling and
experimental validation of the dynamics of capillary isoelectric focusing with electroosmotic zone
displacement, Anal. Methods Instrum., 1, 177, 1993.
33. Mosher, R.A., Zhang, C.-X., Caslavska, J., and Thormann, W., Dynamic simulator for capillary
electrophoresis with in situ calculation of electroosmosis, J. Chromatogr. A, 716, 17, 1995.
34. Thormann, W., Zhang, C.-X., Caslavska, J., Gebauer, P., and Mosher, R.A., Modeling of the impact of
ionic strength on the electroosmotic flow in capillary electrophoresis with uniform and discontinuous
buffer systems, Anal. Chem., 70, 549, 1998.
35. Mao, Q., Pawliszyn, J., and Thormann, W., Dynamics of capillary isoelectric focusing in the absence
of fluid flow: High-resolution computer simulation and experimental validation with whole column
optical imaging, Anal. Chem., 72, 5493, 2000.
36. Thormann, W., Huang, T., Pawliszyn, J., and Mosher, R.A., High-resolution computer simulation of
the dynamics of isoelectric focusing of proteins, Electrophoresis, 25, 324, 2004.
37. Ermakov, S.V., Finite-difference algorithm for convection diffusion equation applied to electrophoresis
problem, Informatica, 3, 173, 1992.
38. Ermakov, S.V., Mazhorova, O.S., and Zhukov, M.Y., Computer simulation of transient states in
capillary zone electrophoresis and isotachophoresis, Electrophoresis, 13, 838, 1992.
39. Ermakov, S.V., Bello, M.S., and Righetti, P.G., Numerical algorithms for capillary electrophoresis,
J. Chromatogr. A, 661, 265, 1994.
40. Shimao, K., Computer simulation of electrophoretic transport. II. Modified general procedure and
application to isoelectric focusing of proteins using immobilised pH gradient, Jap. J. Electrophor., 38,
221, 1994.
41. Shimao, K., Computer simulation of electrophoretic transport. V. Computer simulation of sodium
dodecyl sulfate-polyacrylamide gel electrophoresis, Jap. J. Electrophor., 43, 245, 1999.
42. Schafer-Nielsen, C., A computer model for time-based simulation of electrophoresis systems with
freely defined initial and boundary conditions, Electrophoresis, 16, 1369, 1995.
43. Martens, J.H.P.A., Reijenga, J.C., ten Thije Boonkkamp, J.H.M., Mattheij, R.M.M., and Ever-
aerts, F.M., Transient modeling of capillary electrophoresis. Isotachophoresis, J. Chromatogr. A, 772,
49, 1997.
44. Beckers, J.L. and Boček, P., Sample stacking in capillary zone electrophoresis: Principles, advantages
and limitations, Electrophoresis, 21, 2747, 2000.
45. Ikuta, N. and Hirokawa, T., Numerical simulation for capillary electrophoresis. I. Development of a
simulation program with high numerical stability, J. Chromatogr. A, 802, 49, 1998.
46. Sounart, T.L. and Baygents, J.C., Simulation of electrophoretic separations by the flux-corrected
transport method, J. Chromatogr. A, 890, 321, 2000.
47. Hopkins, D.L. and McGuffin, V.L., Three-dimensional molecular simulation of electrophoretic
separations, Anal. Chem., 70, 1066, 1998.
48. Ermakov, S.V., Jacobson, S.C., and Ramsey, J.M., Computer simulations of electrokinetic transport
in microfabricated channel structures, Anal. Chem., 70, 4494, 1998.
49. Bianchi, F., Ferrigno, A., and Girault, H.H., Finite element simulation of an electroosmotic-driven
flow division at a T-junction of microscale dimensions, Anal. Chem., 72, 1987, 2000.
50. Chatterjee, A., Generalized numerical formulations for multi-physics microfluidics-type applications,
J. Micromech. Microeng., 13, 758, 2003.
51. Sounart, T.L. and Baygents, J.C., Electrically-driven fluid motion in channels with streamwise
gradients of the electrical conductivity, Colloids Surf. A Physicochem. Eng. Asp., 195, 59, 2001.
52. Afonso, J.L. and Clifton, M.J., Optimization of protein separation by continuous-flow electrophoresis:
Influence of the operating conditions and the chamber thickness, Electrophoresis, 20, 2801, 1999.
53. Steinmann, L., Mosher, R.A., and Thormann, W., Characterization and impact of the temporal behavior
of the electroosmotic flow in capillary isoelectric focusing with electroosmotic zone displacement,
J. Chromatogr. A, 756, 219, 1996.
54. Caslavska, J. and Thormann, W., Electrophoretic separations in PMMA capillaries with uniform and
discontinuous buffers, J. Microcol. Sep., 13, 69, 2001.
55. Sounart, T.L. and Baygents, J.C., Simulation of electrophoretic separations: effect of numerical and
molecular diffusion on pH calculations in poorly buffered systems, Electrophoresis, 21, 2287, 2000.
56. Mosher, R.A., Thormann, W., and Bier, M., Computer aided analysis of electric field gradients within
isotachophoretic boundaries between weak electrolytes, J. Chromatogr., 320, 23, 1985.
57. Deshmukh, R.R. and Bier, M., Counterflow in isotachophoresis: Computer simulation and experimen-
tal studies, Electrophoresis, 14, 205, 1993.
58. Thormann, W., Caslavska, J., and Mosher, R.A., Impact of electroosmosis on isotachophoresis in open-
tubular fused-silica capillaries: Analysis of the evolution of a stationary steady-state zone structure by
computer simulation and experimental validation, Electrophoresis, 16, 2016, 1995.
59. Caslavska, J. and Thormann, W., Bidirectional isotachophoresis in open-tubular, untreated fused-silica
capillaries, J. Chromatogr. A, 772, 3, 1997.
60. Caslavska, J. and Thormann, W., Isotachophoresis in dynamically double coated fused-silica
capillaries, Electrophoresis, 27, 4618–4630, 2006.
61. Mosher, R.A., Gebauer, P., and Thormann, W., Computer simulation and experimental validation of
the electrophoretic behavior of proteins. III. Use of titration data predicted by the protein’s amino acid
composition, J. Chromatogr., 638, 155, 1993.
62. Thormann, W. and Mosher, R.A., Experimental and theoretical description of the isotachophoretic
behavior of serum albumin, Electrophoresis, 11, 292, 1990.
63. Thormann, W., Firestone, M.A., Sloan, J.E., Long, T.D., and Mosher, R.A., Isotachophoretic zone
formation of serum albumin in different free-fluid electrophoresis instruments, Electrophoresis, 11,
298, 1990.
64. Dubrovčáková, E., Gaš, B., Vacík, J., and Smolková-Keulemansová, E., Electromigration in systems
with additives in background electrolytes I. Addition of the neutral complexing agent, J. Chromatogr.,
623, 337, 1992.
65. Foret, F., Thompson, T.J., Vouros, P., Karger, B.L., Gebauer, P., and Boček, P., Liquid sheath effects
on the separation of proteins in capillary electrophoresis/electrospray mass spectrometry, Anal. Chem.,
66, 4450, 1994.
66. Gebauer, P. and Boček, P., A new type of migrating zone boundary in electrophoresis: 1.
General description of boundary behavior based on electromigration dispersion velocity profiles,
67. Mosher, R.A. and Thormann, W., High-resolution computer simulation of the dynamics of isoelectric
focusing using carrier ampholytes: The post-separation stabilizing phase revisited, Electrophoresis,
23, 1803, 2002.
68. Thormann, W. and Mosher, R.A., High-resolution computer simulation of the dynamics of isoelec-
tric focusing using carrier ampholytes: Focusing with concurrent electrophoretic mobilization is an
isotachophoretic process, Electrophoresis, 27, 968, 2006.
69. Ermakov, S.V., Zhukov, M.Y., Capelli, L., and Righetti, P.G., Artifactual peak splitting in capillary
electrophoresis. 2. Defocusing phenomena for ampholytes, Anal. Chem., 67, 2957, 1995.
70. Ermakov, S.V., Zhukov, M.Y., Capelli, L., and Righetti, P.G., Experimental and theoretical-study of
artifactual peak splitting in capillary electrophoresis, Anal. Chem., 66, 4034, 1994.
71. Williams, R.L., Childs, B., Dose, E.V., Guiochon, G., and Vigh, G., Peak shape distortions during
capillary electrophoretic separations of multicomponent samples in two co-ion buffers, J. Chromatogr.
A, 781, 107, 1997.
72. Williams, R.L., Childs, B., Dose, E.V., Guiochon, G., and Vigh, G., Peak shape distortions in the
capillary electrophoretic separations of strong electrolytes when the background electrolyte contains
two strong electrolyte co-ions, Anal. Chem., 69, 1347, 1997.
73. Ermakov, S.V. and Righetti, P.G., Computer simulation for capillary zone electrophoresis. A quantita-
tive approach, J. Chromatogr. A, 667, 257, 1994.
74. Ermakov, S.V., Zhukov, M.Y., Capelli, L., and Righetti, P.G., Wall adsorption in capillary
electrophoresis. Experimental study and computer simulation, J. Chromatogr. A, 699, 297,
1995.
75. Zhukov, M.Y., Ermakov, S.V., and Righetti, P.G., Simplified mathematical model of irreversible sample
adsorption in capillary zone electrophoresis, J. Chromatogr. A, 766, 171, 1997.
76. Gaš, B., Štědrý, M., Rizzi, A., and Kenndler, E., Dynamics of peak dispersion in capillary zone elec-
trophoresis including wall adsorption I. Theoretical model and results of simulation, Electrophoresis,
16, 958, 1995.
77. Křivánková, L., Gebauer, P., Thormann, W., Mosher, R.A., and Boček, P., Options in elec-
trolyte systems for online combined capillary isotachophoresis and capillary zone electrophoresis,
J. Chromatogr., 638, 119, 1993.
78. Gebauer, P., Thormann, W., and Boček, P., Sample self-stacking in zone electrophoresis. Theo-
retical description of the zone electrophoretic separation of minor compounds in the presence of
bulk amounts of a sample component with high mobility and like charge, J. Chromatogr., 608, 47,
1992.
79. Gebauer, P., Thormann, W., and Boček, P., Sample self-stacking and sample stacking in zone elec-
trophoresis with major sample components of like charge: General model and scheme of possible
modes, Electrophoresis, 16, 2039, 1995.
80. Gebauer, P., Křivánková, L., Pantůčková, P., Boček, P., and Thormann, W., Sample self-stacking in
capillary zone electrophoresis: Behavior of samples containing multiple major coionic components,
81. Křivánková, L., Pantůčková, P., Gebauer, P., Boček, P., Caslavska, J., and Thormann, W., Chloride
analysis of anionic trace analytes, Electrophoresis, 24, 505, 2003.
82. Křivánková, L., Brezkova, M., Gebauer, P., and Boček, P., Importance of the counterion in optimization
of a borate electrolyte system for analyses of anions in samples with complex matrices performed by
capillary zone electrophoresis, Electrophoresis, 25, 3406, 2004.
83. Hirokawa, T., Okamoto, H., and Gaš, B., High-sensitive capillary zone electrophoresis analysis by
electrokinetic injection with transient isotachophoretic preconcentration: Electrokinetic supercharging,
84. Kim, J.-B., Britz-McKibbin, P., Hirokawa, T., and Terabe, S., Mechanistic study on analyte focusing
by dynamic pH junction in capillary electrophoresis using computer simulation, Anal. Chem., 75,
3986, 2003.
85. Shim, S.-H., Riaz, A., Choi, K., and Chung, D.S., Dual stacking of unbuffered saline samples, transient
isotachophoresis plus induced pH junction focusing, Electrophoresis, 24, 1603, 2003.
86. Breadmore, M.C., Mosher, R.A., and Thormann, W., High-resolution computer simulations of stacking
of weak bases using a transient pH boundary in capillary electrophoresis. 1. Concept and impact of
sample ionic strength, Anal. Chem., 78, 538, 2006.
87. Tesařová, E., Ševčík, J., Gaš, B., and Armstrong, D.W., Effects of partial/asymmetrical filling of
micelles and chiral selectors on capillary electrophoresis enantiomeric separation: Generation of a
gradient, Electrophoresis, 25, 2693, 2004.
88. Andreev, V.P., Pliss, N.S., and Righetti, P.G., Computer simulation of affinity capillary electrophoresis,
89. Fang, N. and Chen, D.D.Y., General approach to high-efficiency simulation of affinity capillary
90. Okhonin, V., Krylova, S.M., and Krylov, S.N., Nonequilibrium capillary electrophoresis of equilibrium
mixtures, mathematical model, Anal. Chem., 76, 1507, 2004.
91. Hruška, V., Jaroš, M., and Gaš, B., Oscillating electrolytes, Electrophoresis, 27, 513, 2006.
92. Thormann, W. and Mosher, R.A., Theoretical and computer aided analysis of steady state moving
boundaries in electrophoresis: An analytical solution for the estimation of boundary widths between
weak electrolytes, Trans. Soc. Comput. Simul., 1, 83, 1984.
18 Heat Production and
Dissipation in Capillary
Electrophoresis
Christopher J. Evenhuis, Rosanne M. Guijt, Miroslav
Macka, Philip J. Marriott, and Paul R. Haddad
CONTENTS
18.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 545

18.2 Heat Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 546
18.2.1 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 546
18.2.2 Factors Affecting Heat Generation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 548
18.2.2.1 Buffer Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 549
18.2.2.2 Applied Voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 551
18.2.2.3 Capillary Length . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 551
18.2.2.4 Capillary Internal Diameter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 551
18.2.2.5 Buffer Viscosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 552
18.3 Heat Dissipation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 552
18.3.1 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 552
18.3.2 Factors Influencing Heat Dissipation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 552
18.3.2.1 Type of Cooling System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 552
18.3.2.2 Cooling System Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553
18.3.2.3 Capillary Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553
18.3.2.4 Capillary Outer Diameter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555
18.4 Determination of the Average Electrolyte Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555
18.4.1 Electroosmotic Flow Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555
18.4.2 Conductance Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557
18.5 Determining the Heat Transfer Coefficient (h) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 557
18.6 Estimating the Temperature Increase of the Electrolyte . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 559
18.7 Control of the Capillary Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 559
18.7.1 Reproducibility of Migration Times and Injection Volumes . . . . . . . . . . . . . . . . . . . . . . 559
18.7.2 Optimizing Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 560
18.7.3 Sample Decomposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 560
18.7.4 Band Broadening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 560
18.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 560
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 560
18.1 INTRODUCTION
Joule heating is an unavoidable phenomenon in capillary electrophoresis (CE) and results from
resistive heating that occurs when an electric current (I) flows through the electrolyte when a potential
difference is applied. The increase in conductivity with temperature results in a positive feedback
545
effect in which the current increases until a steady state is reached. This may even cause the elec-
trolyte to boil or superheat: this is known as autothermal runaway [1]. Temperature control is usually
employed in CE to aid heat dissipation and provide acceptable precision, but measurement of the
electrolyte temperature is often overlooked. In some older systems, only ambient temperature oper-
ation was available, with or without fan-forced airflow. The temperature of the electrolyte affects
its viscosity (η), its dielectric constant (εr ), and the zeta potential (ζ) [2], which affect the precision
of migration times through the effect of η, εr , and ζ on the electroosmotic mobility (µeof ) and the
electrophoretic mobility (µep ). Even small changes in temperature can cause significant deviations
in migration times [3]. The electrolyte temperature also has a major influence on peak broadening
[4–6], such that separation efficiency generally decreases with increasing temperature. Radial tem-
perature differences in the electrolyte result in viscosity differences across the capillary with analytes
traveling faster in the warmer, lower viscosity zone near the axis of the capillary than in the cooler
zones near the capillary wall [5]. Axial temperature differences that result from the layout of the
instrument [7], and differences caused by variations in conductivity as the sample migrates through
the electrolyte, also increase dispersion effects [6].
This chapter discusses the phenomena of Joule heating, heat dissipation, estimation of the elec-
trolyte temperature and its effects, and extends the discussion of Nelson and Burgi [3] provided in the
first edition of this book. The focus of this chapter is on the use of simple methods to determine the
temperature of the electrolyte, free from the influence of Joule heating. An improved understanding
of the radial temperature profile that exists during electrophoresis [8] allows a simple technique to
be introduced for evaluating the heat transfer coefficient for an instrument.
18.2 HEAT GENERATION

18.2.1 THEORY
A significant number of authors have modeled the theory of heat generation and heat dissipation in
CE [4,9–13]. The rate of Joule heating (P), which is the resistive heating power that occurs when
an electric current (I) flows through the electrolyte, can be quantified simply using the following
equation:
P = IV , (18.1)
where V is the applied voltage. If the current is measured in microamperes (µA) and the voltage is
measured in kilovolts (kV), P will be measured in milliWatts (mW). Instead of P, the power per unit
length (P/L) is generally used to take the length of the capillary into account.
Traditionally, it has been suggested that a graph of current versus voltage should be constructed
for an electrolyte to characterize the CE system. In Figure 18.1a, an Ohm’s Law plot is given for a
fused-silica (FS) capillary and for a fluorinated ethylene–propylene (FEP) capillary. The greater the
deviation of the curve from the Ohm’s Law plot, the less efficient is the cooling system. However,
assignment of the voltage at which the deviation of the curve from the Ohm’s Law plot becomes
excessive is somewhat arbitrary, although this is sometimes used as a guide for maximum operating
potential.
A more useful plot to characterize the CE system is a graph of conductance (G) versus the power
dissipated per unit length as this can provide information about the increase in the buffer temperature,
the internal radius of the capillary, and the cooling efficiency of the instrument. Conductance varies
linearly with temperature, making it very useful as a probe for the average electrolyte temperature
in the capillary. A detailed description is provided in Section 18.4.2.
Equation 18.2 shows that conductance is the reciprocal of resistance and can be found
by dividing the current by the voltage, parameters that can be easily monitored and recorded
Heat Production and Dissipation in Capillary Electrophoresis 547
(a) (b)
120 4.0 G(FEP) = 0.2416 P/L+ 1.7405

FEP R2 = 0.9976
100 3.5
Conductance (nS)
Current (µA)
80
3.0
60 FS
2.5 G(FS) = 0.1104 P/L + 1.5093
40 R2 = 0.9998
Ohm's Law 2.0
20
0 1.5
0 5 10 15 20 25 30 0 2 4 6 8 10
Voltage (kV) Power/Length (Wm–1)
FIGURE 18.1 Graphs of (a) I vs. V and (b) G vs. P/L for different capillaries of equal length. Conditions:
Ltot = 41.0 cm fused-silica (FS) di = 74.0 µm, do = 362.8 µm and fluorinated ethylene–propylene copolymer
(FEP) di 80 µm, do ≈ 370 µm. Electrolyte: 10 mM phosphate buffer at pH 7.21.
during the course of a separation.
I
G= . (18.2)
V
If the current is measured in µA and the voltage is in kV, the conductance is determined in
nanoSiemens (nS = 10−9 AV−1 = 10−9 −1 ). Plots of G versus P/L for a FS capillary and for a
fluorinated FEP capillary are given in Figure 18.1b.
The later onset of deviation from Ohm’s Law for the FS capillary as voltage increases in the
Ohm’s Law plot in Figure 18.1a demonstrates the superior heat dissipation of FS over FEP. This
is confirmed by the smaller slope of G vs. P/L for the FS capillary in Figure 18.1b. The ratio
of the gradient of the G vs. P/L graph to the intercept can be used to estimate the temperature
increase of the electrolyte. A more detailed description of temperature calculation will follow in
Section 18.4. In addition, the average internal diameter (di ) of the capillary can be derived from the
conductance at zero power (G0 ). The intercept, G0 , is proportional to the cross-sectional area of the
capillary. Provided that the lengths of the capillaries are equal, this allows one to use Equation 18.3
to calculate the internal diameter of an unknown capillary based on the internal diameter of another
capillary [14].
G0(FEP) (di FEP)2

= . (18.3)
G0(FS) (di FS )2
Figure 18.1b demonstrates that the FEP capillary has a larger internal diameter, as evident from
its larger conductance at zero power (G0 ). On the basis of the di of the FS capillary, the diameter of
the FEP capillary can be calculated as illustrated in the following equation:

G0(FEP) 1.7405
di FEP = · di Fs = · 74.0 µm = 79.5 µm. (18.4)
G0(Fs) 1.5093
The possibility of determining the electrolyte temperature and the capillary diameter make the
G versus P/L graph a superior method to the Ohm’s Law plot for characterizing Joule heating.
∆TAir
∆TRadial
TWall TAxis
TMean
TSet
∆TPI
∆TMean ∆TFS
FIGURE 18.2 Schematic diagram showing the locations at which the temperature determinations are made,
and definitions of temperature differences. (Reproduced with permission from Evenhuis, C. J., et al., Anal.
Chem. 2006, 78, 2684–2693. Copyright 2006 American Chemical Society.)
Grushka et al. [4] have provided a useful rule of thumb that the radial temperature difference in
the capillary (TRadial , see Figure 18.2) should not be allowed to exceed 1.5 ◦ C otherwise there is
an unacceptable loss of efficiency due to peak broadening.
It has been shown that the radial temperature difference is directly proportional to the power per
unit length and is independent of the cooling efficiency of the instrument. The following equation
applies for aqueous electrolytes:
1 P P
TRadial = · = 0.1315 ◦ C W−1 m · , (18.5)
4πλ L L
where λ is the thermal conductivity of the electrolyte (λwater = 0.605 Wm−1 K−1 ). To avoid excessive
peak broadening P/L should not be allowed to exceed 11 Wm−1 . For nonaqueous electrolytes, which
tend to have lower thermal conductivities, or electrolytes containing organic constituents, it is wise
to calculate or estimate the maximum value of P/L[15]. For example, for methanol (λMeOH = 0.202
Wm−1 K−1 ) the maximum power level would be 3.8 Wm−1 .
A number of other problems are associated with working at high values of P/L. McCormick
[16,17] observed that if the electrolyte heated up too quickly, part of the sample could actually be
expelled from the capillary by the rapid expansion of the buffer. The formation of bubbles or even
boiling of the electrolyte can also occur at high power levels, resulting in a sudden decrease of the
current. Microbubbles formed by outgassing of dissolved air allow the current to continue but can
often be detected as random sharp peaks (spikes). To avoid these spurious peaks, it is good practice
to degas the electrolyte in an ultrasonic bath before its use.
18.2.2 FACTORS AFFECTING HEAT GENERATION

The flowchart (see Figure 18.3) demonstrates that the capillary, electrolyte, and instrumental para-
meters all have an influence to a varying degree on the overall temperature rise of the electrolyte
(TMean , see Figure 18.2). An explanation of the terms used is shown in Table 18.1. Key parameters
are discussed in detail in the following sections.
∆TWall + 1/
1 ∆T
Radial + ∆TAir = ∆TMean
2
In(do / di) P 1 P 1 P
∆TWall = • ∆TRadial = • ∆TAir = •
2π λwall L 4πλ L πdoh L
P VI
=
L L
I = GV
κA
G =
L
πd 2i κ = ΛmC
A =
4
1
Λm = Λ0 – KC 2
iF vn
Λ0 = Σ
i Zi
λWall do di L c λ V VAir
Capillary Electrolyte Instrument
FIGURE 18.3 Flowchart showing the influences of various experimental variables on the electrolyte
temperature.
18.2.2.1 Buffer Composition

Figure 18.4 illustrates that TMean increases linearly with the power per unit length (P/L). The buffer
composition and concentration affect its conductivity and therefore the rate of heat generation during
a separation. Equation 18.6 illustrates the principle that the use of a lower conductivity buffer results
in a smaller increase in temperature for the same electrical field strength (E).
P κAV 2
= = κAE 2 . (18.6)
L L2
As conductivity depends on the mobility of the ions involved (see Equation 18.7), the use of
counterions of lower mobility can reduce the current.
µiF

0 = , (18.7)
zi
i
TABLE 18.1
Explanation of Terms
Abbreviation Quantity Unit
TAir Temperature difference across air layer surrounding capillary (see ◦C
Figure 18.2)
TWall Temperature difference across the capillary wall [sum of TFS and ◦C
TPI (see Figure 18.2)]

TFS Temperature difference across fused-silica wall (see Figure 18.2) ◦C
TMean Change in mean temperature of the electrolyte as a result of Joule ◦C
heating (see Figure 18.2)

TPI Temperature difference across poly(imide) coating (see ◦C
Figure 18.2)
TRadial Radial temperature difference across electrolyte (see Figure 18.2) ◦C
γ Thermal coefficient of electrical conductivity ◦ C−1
η Viscosity of electrolyte kg m−1 s−1

0 Limiting ionic conductivity Sm2 mol−1

m Molar conductivity Sm2 mol−1
ε Electrical permittivity of the electrolyte F m−1
ζ Zeta potential V
κ Electrical conductivity of electrolyte S m−1
λ Thermal conductivity of electrolyte Wm−1 K−1
λWall Thermal conductivity of wall Wm−1 K−1
µ Limiting ionic mobility m2 s−1 V−1
µi Limiting ionic mobility of i species m2 s−1 V−1
µep Electrophoretic mobility m2 s−1 V−1
µEOF Electroosmotic mobility m2 s−1 V−1
µEOF (0) Electroosmotic mobility at zero power m2 s−1 V−1
Constant relating TMean to P/L ◦ C W−1 m
A Cross-sectional area M2
c Molar concentration mol L−1
di Internal diameter of capillary M
do External diameter of capillary M
F Faraday’s constant 96,487 C mol−1
G Conductance S
G(0) Conductance at zero power S
h Heat transfer coefficient Wm−2 K−1
I Electric current A
K Kohlrausch constant Sm2 mol−3/2 L−1/2
L Total length of the capillary M
P Power W
P/L Power per unit length Wm−1
q Charge of ion C
rh Hydrodynamic radius M
V Applied voltage V
zi Valency of ionic species i Dimensionless
where
0 is the limiting ionic conductivity, µi is the limiting ionic mobility of species i, and zi is its
valency.
For example, Na+ should be used in preference to NH+ +
4 or K when counterions are being
considered. Large organic counterions such as histidine can significantly reduce the current but
one needs to be aware that separation selectivity can also be affected. Zwitterionic buffers such as
∆ T FEP = 4.816 P/L + 0.476

40
35 R2 = 0.9991
30 ∆ TPMMA= 4.507 P/L – 0.001

∆TMean (ºC)
2
R = 0.9995 ∆T PEEK = 4.094 P/L
25
R2 = 0.9995
20
15
∆ T FS = 2.715 P/L + 0.001
10 2
R = 0.9996
5
0
0 1 2 3 4 5 6 7 8
P/L (Wm–1)
FIGURE 18.4 Variation of mean increase in electrolyte temperature with power per unit length for capillaries
made from different materials •, FEP; , PMMA; , PEEK; and , FS. Conditions: h = 376 Wm−2 K−1 ,
Ltot = 32.2 cm, di ≈ 75 µm, do ≈ 365 µm, See Reference [14] for more details for each capillary. Electrolyte:
10 mM phosphate, pH = 7.21. (Reproduced with permission from Evenhuis, C. J., et al., Electrophoresis 2005,
26, 4333–4344. Copyright 2005 Wiley VCH.)
3-(N-morpholino)propanesulfonic acid (MOPS) and 2-(N-morpholino)ethanesulfonic acid (MES)

have high buffer capacities with low conductivities; they tend to absorb at low ultraviolet (UV)
wavelengths but may be used to advantage for UV detection above 210 nm [3].
18.2.2.2 Applied Voltage

The rate of Joule heating increases with the square of the applied voltage, so doubling the voltage
increases P/L by a factor of 4 (see Equation 18.6). The voltage is usually optimized for maximum
efficiency but a compromise between efficiency and heat generation can often lead to shorter analysis
times if there is sufficient resolution and the analytes are not influenced by elevated temperatures [3].
18.2.2.3 Capillary Length

The rate of Joule heating (P) is inversely proportional to the length of the capillary, so halving the
capillary length increases both P/L and TRadial by a factor of 4 (see Equation 18.6), but decreases
the separation time by a factor greater than 4 due to the increased mobility of the analytes. Resolution
can be improved by using a longer capillary if the same electrical field can be maintained [3].
18.2.2.4 Capillary Internal Diameter

As the conductance depends on the cross-sectional area of the capillary (see Equation 18.6), for a
constant applied voltage, P/L increases with the square of the internal diameter (di ).
P πκdi2 V 2
= . (18.8)
L 4L 2
Decreasing the internal diameter from 75 µm to 50 µm leads to a 56% reduction in P/L. It follows
that the use of smaller bore capillaries enables higher conductivity buffers to be used at similar field
strengths, which can significantly improve peak focusing. However, decreasing the internal diameter
of the capillary may result in a loss of sensitivity for concentration-sensitive detection techniques,
such as UV absorption, due to the decreased detection pathlength.
18.2.2.5 Buffer Viscosity

Increasing the buffer viscosity (η) affects the apparent mobility through the electrophoretic mobility
(µep , see Equation 18.9) and electroosmotic mobility (µEOF , see Equation 18.10) by increasing the
resistance to movement
q
µep = , (18.9)
6π rh η
where q is the charge on the solvated ion, rh is its hydrodynamic radius, ε is the electrical permittivity
of the electrolyte, and ζ is the zeta potential.
εζ
µEOF = − . (18.10)
η
To a good approximation, there is an inverse relationship between the viscosity and the resulting
current; doubling the buffer viscosity will halve the current and TRadial . In nonaqueous solvents,
electrical currents are generally lower than that for the same electrolyte dissolved in water, but
predicting TRadial becomes complex as one needs to take into account variations in the mobilities
of species, dissociation equilibria, dielectric constant, zeta potential, and thermal conductivity [18].
18.3 HEAT DISSIPATION

18.3.1 THEORY
It is not surprising that since heat energy is generated in the electrolyte and the heat is conducted
through the capillary walls, the electrolyte temperature is greater at the axis than at the inside wall.
A temperature gradient therefore exists from the central axis to the outside wall of the capillary. The
size of the radial temperature difference across the electrolyte (TRadial ) and across the capillary
wall(s) is dictated by the power generated per unit length (P/L) and the thermal conductivity of
each medium, but the overall increase in temperature of the electrolyte is determined by the cooling
efficiency, which is characterized by the heat transfer coefficient at the outer surface (h). A relatively
straightforward method of finding h for a CE instrument is described in Section 18.5. It has been
predicted that an approximately parabolic temperature profile exists within the electrolyte [9,12] and
this has been verified using noninvasive methods, such as Raman Spectroscopy [19]. In FS capillaries,
an exponential decrease in temperature occurs across the FS and poly(imide) layers. Interestingly, the
magnitude of the temperature difference across the thin polymer layer is comparable to that across
the FS. However, even when highly efficient cooling systems are in operation, the main temperature
drop occurs from the outside surface of the poly(imide) coating to the set temperature [12] (see
Figure 18.5).
18.3.2 FACTORS INFLUENCING HEAT DISSIPATION

18.3.2.1 Type of Cooling System
The thermal conductivity of the cooling medium and the speed at which the medium flows over
the capillary influence the cooling efficiency. For example, helium has been shown to be about six
times more effective than air for capillary cooling due to its greater thermal conductivity [20]. As
mentioned earlier, the elevation in electrolyte temperature is determined mainly by how efficiently
74.0 µm
320.8 µm
362.8 µm
523 µm
3.0
Relative temperature excess

100
Temperature excess (ºC)
Electrolyte
2.5 Fused silica
Poly(imide)
80
2.0
60
1.5 Air
40
1.0
0.5 20
0.0 0
–300 –200 –100 0 100 200 300
Radius (µm)
FIGURE 18.5 Calculated average radial temperature profile for a FS capillary. Conditions: P/L = 1.00 Wm−1 ,
h = 376 Wm−2 K−1 , Ltot = 32.2 cm, di = 74.0 µm, do = 362.8 µm, thickness of poly(imide) layer = 21.0 µm.
Electrolyte: 10 mM phosphate, pH = 7.21. (Reproduced with permission from Evenhuis, C. J., et al., Anal.
Chem. 2006, 78, 2684–2693. Copyright 2006 American Chemical Society.)
heat is conducted from the outer surface of the capillary, and this is characterized by the heat transfer
coefficient (h).
Typical values for h in CE instruments are approximately 50 Wm−2 K−1 for stagnant air [21],
300–700 Wm−2 K−1 for fan-forced air depending on the speed of flow and the fraction of the capillary
that is temperature controlled [8,21] and between 700 and 1200 Wm−2 K−1 for liquid cooling [5,22].
The most effective cooling system to be demonstrated for CE used a Peltier thermoelectric device to
cool an alumina block. The capillary was housed in a purpose built groove and thermal contact was
enhanced using ethylene glycol. A heat transfer coefficient of h = 2600 Wm−2 K−1 was reported
for this device [10].
Figure 18.6 illustrates that increasing the speed of the airflow becomes less effective at higher
airflow rates, but that there is a large change in heat transfer at low flows.
18.3.2.2 Cooling System Design

Axial temperature differences are to be avoided. Ideally, the temperature of the whole capillary should
be actively controlled to the same temperature. Unfortunately, the design of some commercial CE
instruments is such that there are substantial portions of the capillary without temperature control.
These portions of the capillary are surrounded either by stagnant air or by electrolyte in the vials.
Without redesigning the instrument, areas with stagnant air cannot be avoided, but many instruments
allow regulation of the temperature of the vials containing the sample and electrolyte vials, which
allows significant improvements in the reproducibility.
18.3.2.3 Capillary Material

Although poly(imide)-coated FS accounts for the vast majority of applications in CE, polymeric
materials have also been studied as materials for CE capillaries since these offer a variety of different
800
700
600
500
h (Wm–2 K–1)
400
300
200
100
0
0 10 10 15 20
vair (ms–1)
FIGURE 18.6 Theoretical variation of the heat transfer coefficient with the speed of the airflow for a
poly(imide) coated FS capillary with di = 75 µm and do = 375 µm.
TABLE 18.2
Thermal Conductivity of Various Substances
Thermal Conductivity (λ)
Material at 25 ◦ C (Wm−1 K−1 )a References
Air 0.025 [12]
Fused-silica (FS) 1.40 [12]
Borosilicate glass 1.0 [12]
Water 0.605 [12]
Poly(dimethylsiloxane) (PDMS) 0.18 [46]
Poly(etherether ketone) (PEEK) 0.252 [47]
Poly(tetrafluoroethene) (PTFE) 0.265 [48]
Copolymer of tetrafluoro-ethene and hexafluoropropene (FEP) 0.209 [49]
Poly(fluoroethene) 0.1 [50]
Poly(imide) 0.155 [4]
Poly(iminoadipoyliminohexane-1,6-diyl) (Nylon-6,6) 0.23 [51]
Poly(methyl methacrylate) (PMMA) 0.193 [52]
Poly(phenylethene) (Polystyrene) 0.11 [53]
Teflon AF 0.116 [54]
a To convert from Wm−1 K−1 to cal s−1 m−1 K−1 divide by a factor of 4.184.
surface physical properties. Examples include the following: poly(methylmethacrylate) (PMMA)

[23,24], FEP [25], poly(propene) [26], poly(butylene-terephthalate) [26], poly(tetrafluoroethene)
(PTFE) [27–30], and poly(etherether ketone) (PEEK) [31].
Generally speaking, the thermal conductivity of polymers is much less than FS (see Table 18.2)
so that the temperature difference across the capillary walls and the rise in temperature of the
electrolyte are both significantly greater for polymeric capillaries under the same conditions (see
Figure 18.5) [14].
18.3.2.4 Capillary Outer Diameter

It has been noted that increasing the external diameter of the capillary by using a thicker layer of FS
actually reduces the rise in temperature of the electrolyte [3,10]. Although the temperature gradient
across the capillary wall increases, heat is removed more effectively from the larger surface area at
the external surface. However, increasing the thickness of the polymer layer causes an increase in the
temperature but can lead to a reduction in the optical baseline noise [3]. These authors suggested that
the increased temperature of the electrolyte could be counterbalanced by reducing the temperature
of the cooling system [3].
18.4 DETERMINATION OF THE AVERAGE ELECTROLYTE

TEMPERATURE
A knowledge of the electrolyte temperature is important in CE as temperature changes in the elec-
trolyte influence precision, accuracy, separation efficiency, and method robustness [7,14,32]. During
the past two decades, a considerable amount of research has been conducted toward electrolyte tem-
perature measurements in CE [1,14,19,21,32–42]. Early methods have included using the variation
of electroosmotic mobility (µEOF ), electrophoretic mobility (µep ), and electrical conductivity (κ) to
measure temperature [38,39]. More recently, techniques such as external thermocouples [21], Raman
thermometry [19,40], NMR spectroscopy [32,35], thermochromic probes [41], and the variation in
fluorescence response [42] have been used to measure temperatures. Most of these methods require
the modification of the existing instrument and/or the purchase of additional equipment.
Two noninvasive methods of temperature measurement based on the electroosmotic mobility
and conductance are discussed in further detail below. The linear relationships between both the
electroosmotic mobility and the conductance versus P/L are illustrated in Figure 18.7.
18.4.1 ELECTROOSMOTIC FLOW METHOD

It should be noted that the electroosmotic mobility (µEOF ) is not a measure of the average elec-
trolyte temperature (TMean , see Figure 18.2) over the whole cross section; but instead reflects the
average temperature of the electrolyte near the inner wall of the capillary (TWall ). This is because
the electroosmotic flow is generated at the capillary wall. To determine the average temperature of
110 2.8
G
G = 0.1077P/L + 1.9351
R2 = 0.9996
2.6
µEOF(x10–9 m2s–1 V–1)
100 µEOF
G(x10–9 S)
2.4
90
2.2
µEOF = 4.2394P/L + 72.05
80 R2 = 0.9995
2
70 1.8
0 2 4 6 8
P/L (Wm–1)
FIGURE 18.7 Variation of electroosmotic mobility and conductance with power per unit length. Conditions
as in Figure 18.5.
the electrolyte (TMean ), a correction of 12 TRadial needs to be added to the wall temperature to take
into account the radial temperature profile of the electrolyte [8].
1 P
TMean = TWall + TRadial = TWall + 0.066 KW−1 m · (18.11)
2 L
The previous methods of using µEOF as a temperature probe have assumed that changes in
µEOF were due solely to variations in viscosity. More recently, it has been shown that changes
to the dielectric constant and zeta potential also influence µEOF [8]. To obtain the most accurate
results, an initial calibration curve for the variation of electroosmotic mobility with temperature is
required for a specific electrolyte and capillary material. To make this calibration plot, µEOF needs
to be determined at three or more different values of P/L for a range of set temperatures using
the instrument’s temperature control system. At each temperature, the graph of µEOF versus P/L is
extrapolated to P/L = 0 to find µEOF free of the effects of Joule heating, µEOF (0). These values of
µEOF (0) can be used to prepare the calibration curve for the variation of electroosmotic mobility
with temperature, as illustrated in Figure 18.8.
For the example illustrated in Figure 18.8, the electrolyte used was 10 mM phosphate buffer at
pH = 7.21. µEOF was found to increase at 2.22% per ◦ C, that is, the temperature coefficient (or slope
of µEOF versus T ) was 0.0222◦ C−1 . Once the temperature coefficient for electrophoretic mobility
is known, the electrolyte temperature can be easily determined for any value of P/Las long as the
same capillary and electrolyte are used.
µEOF − µEOF (0) P

TMean = T0 + + 0.066 KW−1 m · (18.12)
0.0222µEOF (0) L
where T0 is the temperature set for the instrument, µEOF is the electroosmotic mobility measured
for the electrophoretic separation, and µEOF (0) is the extrapolated value of µEOF to zero power at
ambient temperature.
µEOF (x10–9m2s–1V–1)
95 32.9 °C
1.20 29.3 °C
85
25.0 °C
75 23.7 °C
21.0 °C
65
1.10 55
µEOF(T)/(µEOF(25 °C)
0 2 4 6
P/L (Wm–1)
1.00
µEOF(T) /µEOF(25 ºC) = 0.0222T + 0.445
R2 = 0.9435
0.90
0.80
17.5 20.0 22.5 25.0 27.5 30.0 32.5
T (°C)
FIGURE 18.8 Calibration curve for µEOF versus Ambient temperature. Conditions as in Figure 18.5. (Repro-
duced with permission from Evenhuis, C. J., et al., Anal. Chem. 2006, 78, 2684–2693. Copyright 2006 American
Chemical Society.)
Example: If µEOF (0) = 7.209 × 10−8 m2 s−1 V−1 at 25.00 ◦ C and µEOF (P/L = 4.625 Wm−1 ) =
9.179 × 10−8 m2 s−1 V−1 , then
9.179 × 10−8 − 7.209 × 10−8

TMean = 25.00 ◦ C + + 0.066 ◦ CmW−1 · 4.625 Wm−1
0.0222 ◦ C−1 7.209 × 10−8
= 25.00 ◦ C + 12.31 ◦ C + 0.30 ◦ C = 37.61 ◦ C
The increase in mean temperature of the electrolyte (TMean ) is directly proportional to P/L.
P
TMean = , (18.13)
L
where is a constant relating TMean to the P.

For the example above, TMean = 12.61 ◦ C,
TMean L 12.61 ◦ C
= = = 2.726 ◦ C W−1 m.
P 4.625 Wm−1
To obtain the most accurate results, the calibration process described above would need to be
applied for each different electrolyte. This makes the method less attractive for routine use.
18.4.2 CONDUCTANCE METHOD

Conductance (G) can be calculated by simply dividing the current by the voltage (see Equation 18.2).
Conductance can be used as a probe for the mean electrolyte temperature (TMean ). For most elec-
trolytes, the thermal coefficient of electrical conductivity (γ ) varies from about 0.019 K−1 to 0.021
K−1 [43] at the concentrations used for CE. It is possible to calculate γ from first principles using
the Debye–Hückel–Onsager equation [8] but this is far from being a trivial process. As a rule of
thumb, using γ = 0.020 K−1 gives values of TMean to within 5%, but for more accurate values a
calibration process is necessary.
To measure TMean , the value of conductance at the ambient temperature is required without the
influence of Joule heating [G(0)]. This can be obtained by measuring G at 3 or more values of P/L
and extrapolating to P/L = 0 (see Figure 18.7).
G − G (0)
TMean = . (18.14)
γ G (0)
For example, 10 mM phosphate buffer at pH = 7.21 has γ = 0.0205 ◦ C−1 . Using the data from
Figure 18.7, G(0) = 1.935 nS and G(4.625 Wm−1 ) = 2.441 nS.
This value is within 0.15 ◦ C of the increase in temperature that was calculated above using µEOF .
The overall process is far more straightforward than the method based on µEOF and is less sensitive
to slight changes in wall chemistry that can occur over a period of time.
18.5 DETERMINING THE HEAT TRANSFER COEFFICIENT (h)

As illustrated in Figure 18.5, the difference between the outer wall temperature and the ambient
temperature (referred to as TAir , see Figure 18.2) can easily account for 80% of the temperature
increase of the electrolyte, even in an efficiently cooled instrument. The size of TAir is determined
by the size of the heat transfer coefficient, so it useful to be able to determine h experimentally rather
than to assume that the instrument is still functioning according to the manufacturer’s specifications.
It should be noted that it is not possible to calculate h for an instrument per se as the value calculated
depends on the outer diameter of the capillary used. However as most practitioners use capillaries
with do ≈ 375 µm, this value will be used in the example below for the purposes of illustration.
h is calculated from the difference in temperature between the outer wall and the ambient
temperature of the coolant (TAir )
1 P
h= · , (18.15)
πdo TAir L
where do is the outer diameter of the capillary and P/L is the power dissipated per unit length. TAir
can be found from the rise in the mean temperature of the electrolyte and the temperature differences
across the FS and poly(imide) layers (see Figure 18.2).

TAir = TMean − 1 2 TRadial − TFS − TPI . (18.16)
For both FS and poly(imide) capillaries, the temperature difference across the wall is found using

1 do P
TAcrossWall = ln · , (18.17)
2πλWall di L
where λWall refers to the thermal conductivity of the wall material. The thermal conductivities for
FS and poly(imide) are λFS = 1.40 Wm−1 K−1 and λPI = 0.155 Wm−1 K−1 , respectively [12,21].
For convenience, the following calculations use P/L = 1.00 Wm−1 but the calculations are
applicable for any value of P/L. For the purpose of the calculation, the values used are internal
diameter, di = 75.0 µm, diameter of FS dFS = 335 µm, and outer diameter do = 375 µm (the
thickness of the poly(imide) coating is 20 µm).

1 335 µm
TFS = ln · 1.00 Wm−1 = 0.170 K
2π × 1.40 Wm−1 K−1 75.0 µm

1 375 µm
TPI = ln · 1.00 Wm−1 = 0.116 K
2π × 0.155 Wm−1 K−1 335 µm
1 1 1 P
TRadial = × −1 −1
· = 0.066 KW−1 m · 1.00 Wm−1 = 0.066 K
2 2 4π × 0.605 Wm K L
Using the previously determined value of TMean = 2.726 K for P/L = 1.00 Wm−1 for a capillary
with a total length of 32.2 cm,
1
TAir = TMean − TRadial − TFS − TPI
2
= 2.726 K − 0.066 K − 0.170 K − 0.116 K = 2.374 K
1 P 1
∴h= · = −6
· 1.00 Wm−1 = 358 Wm−2 K−1 .
πdo TAir L π × 375 × 10 m × 2.374 K
It should be noted that the value of h calculated is an average over the whole length of the
capillary and that the size of h will depend on the fraction of the capillary that is under thermostat
control. Longer capillaries that have a larger fraction of the capillary under thermostat control will
tend to give larger values of h.
18 Passive
cooling
16
14
12
10
TMean
0
0 200 400 600 800 1000 1200
h (Wm–2 K–1)
FIGURE 18.9 Variation of the mean rise in electrolyte temperature with heat transfer coefficient for P/L =
1.00 Wm−1 .
18.6 ESTIMATING THE TEMPERATURE INCREASE OF THE

ELECTROLYTE
The temperature rise of the electrolyte can be estimated using a combination of the heat transfer
coefficient and the power per unit length using Equation 18.18. For the FS capillary described above
[di = 75.0 µm, do = 375 µm, thickness of poly(imide) = 20 µm]:
P
TMean = TAir + 1 2TRadial + TFS + TPI = TAir + 0.352 KmW−1 ·
L
(18.18)
1 P 849 P
= + 0.352 KmW−1 = + 0.352 KmW−1
πdo h L h L
A graph of Equation 18.18 is shown in Figure 18.9.

For capillaries with do = 375 µm, the error introduced by Equation 18.18 is less than 0.4◦ C for
di = 50–150 µm. For capillaries with narrower bores, the temperature rise is slightly greater than
predicted and for wider bore capillaries the predicted temperature rise is smaller.
18.7 CONTROL OF THE CAPILLARY TEMPERATURE

When selecting the desired temperature of the capillary, consideration should be given to repro-
ducibility of sample injection and migration times, optimization of resolution, the possibility of
sample decomposition, and the minimization of analysis times [3].
18.7.1 REPRODUCIBILITY OF MIGRATION TIMES AND INJECTION VOLUMES

Both the electroosmotic mobility and electrophoretic mobility are temperature-dependent so it is not
surprising that controlling the temperature of the capillary leads to improved intraday-, interday-, and
interlaboratory-reproducibility of migration times. Nevertheless, it is possible to achieve acceptable
run-to-run reproducibility using an instrument without temperature control if experiments are carried
out in a laboratory in which the ambient temperature is maintained within a narrow range.
Since the viscosity of the electrolyte decreases with increasing temperature at a rate of approxi-
mately 2% per ◦ C, the volume of sample injected using a particular pressure and time combination
will increase with the ambient temperature. As a rule of thumb, the product of pressure and time
should be reduced by 20% for each 10◦ C rise in the ambient temperature. Alternatively, peak areas
can be normalized by multiplying them by the ratio of the viscosities at the different temperatures
[3]. As mentioned earlier, where the facility exists, it is wise to control the temperature of the tray
housing the electrolyte vials as this improves reproducibility of both injections and migration times.
18.7.2 OPTIMIZING RESOLUTION

Nelson and Burgi [3] observed that the relative migration times could be affected differently by
changes to the ambient electrolyte temperature so that the order in which a range of anions is
detected can be changed considerably. Ions that comigrate at one temperature can be well resolved at
a different temperature. Although the mechanism for these changes is not well understood, changing
the temperature may be an effective strategy to improve resolution. However, it should be noted that
generally there is an increase in peak distortion that accompanies increases in temperature [5].
18.7.3 SAMPLE DECOMPOSITION

It is well known that biological samples may be sensitive to decomposition and/or conforma-
tional changes at elevated temperatures [10,44,45]. Such changes can be avoided by monitoring
the electrolyte temperatures using the procedures described earlier or by controlling the electrolyte
temperature at a lower value to counteract the effect of Joule heating.
18.7.4 BAND BROADENING

Joule heating is well established as a source of peak broadening. Petersen et al. [5] showed that a
10◦ C increase in temperature is associated with a 3.3% decrease in the number of theoretical plates
due to increased axial diffusion. Radial temperature differences tend to be more significant; Grushka
et al. [4] recommended that the radial temperature difference should not be allowed to exceed 1.5◦ C.
Gobie and Ivory [1] found that thermally induced parabolic distortion of the migration velocity can
be countered by applying an opposing laminar flow to the outlet end of the capillary. This method
was found to be particularly effective in wider bore capillaries.
18.8 CONCLUSIONS
Although Joule heating is inevitable during electrophoretic separations, its magnitude is simple to
quantify. Variation of the electrical conductance (G) with the power per unit length (P/L) is a simple
and effective method of measuring the temperature increase of the electrolyte. An experimental
method of determining the heat transfer coefficient (h) for a capillary has been outlined, along
with a simple equation to calculate the rise in the mean temperature of the electrolyte (TMean ).
TMean increases linearly with the power per unit length and can be minimized by using narrow bore
capillaries with low electrical conductivity electrolytes and by increasing the length of the capillary.
Finally, the ambient temperature of the capillary during a separation should always be considered as
a variable that must be optimized with respect to speed, efficiency, and resolution [3].
REFERENCES
1. Gobie, W. A. and Ivory, C. F., J. Chromatogr. 1990, 516, 191–210.
2. Evenhuis, C. J., Guijt, R. M., Macka, M., Marriott, P. J., and Haddad, P. R., Electrophoresis 2005, 27,
672–676.
3. Nelson, R. J. and Burgi, D. S., in Landers, J. P. (Ed.), Handbook of Capillary Electrophoresis, CRC
Press, Boca Raton, 1994, pp. 549–562.
4. Grushka, E., McCormick, J. R. M., and Kirkland, J. J., Anal. Chem 1989, 61, 241–246.
5. Petersen, N. J., Nikolajsen, R. P. H., Mogensen, K. B., and Kutter, J. P., Electrophoresis 2004, 25,
253–269.
6. Xuan, X. and Li, D., Electrophoresis 2005, 26, 166–175.
7. Rathore, A. S., J. Chromatogr. A 2004, 1037, 431–443.
8. Evenhuis, C. J., Guijt, R. M., Macka, M., Marriott, P. J., and Haddad, P. R., Anal. Chem. 2006, 78,
2684–2693.
9. Hjerten, S., Chromatog. Rev. 1967, 9, 122–219.
10. Nelson, R. J., Paulus, A., Cohen, A. S., Guttman, A., and Karger, B. L., J. Chromatogr. 1989, 480,
111–127.
11. Bello, M. S. and Righetti, P. G., J. Chromatogr. 1992, 606, 103–111.
12. Knox, J. H., Chromatographia 1988, 26, 329–337.
13. Knox, J. H. and McCormack, K. A., Chromatographia 1994, 38, 215–221.
14. Evenhuis, C. J., Guijt, R. M., Macka, M., Marriott, P. J., and Haddad, P. R., Electrophoresis 2005, 26,
4333–4344.
15. Porras, S. P., Marziali, E., Gas, B., and Kenndler, E., Electrophoresis 2003, 24, 1553–1564.
16. McCormick, R. M., Anal. Chem. 1988, 60, 2322–2328.
17. Knox, J. H. and McCormick, K. A., Chromatographia 1994, 38, 279–282.
18. Porras, S. P. and Kendler, E., Electrophoresis 2005, 26, 3203–3220.
19. Liu, K.-L. K., Davis, K. L., and Morris, M. D., Anal. Chem. 1994, 66, 3744–3750.
20. Cifuentes, A., Kok, W., and Poppe, H., J. Microcolumn Sep. 1995, 7, 365–374.
21. Nishikawa, T. and Kambara, H., Electrophoresis 1996, 17, 1115–1120.
23. Caslavska, J. and Thormann, W., J. Microcol. Sep. 2001, 13, 69–83.
24. Chen, S. and Lee, M. L., J. Microcol. Sep. 1997, 9 57–62.
25. Kvasnicka, F., Electrophoresis 2003, 24, 860–864.
26. Liu, P. Z., Malik, A., Kuchar, M. C., Vorkink, W. P., and Lee, M. L., J. Microcol. Sep. 1993, 5,
245–253.
27. Deng, B. and Chan, W.-T., Electrophoresis 2001, 22, 2186–2191.
28. Lukacs, K. D. and Jorgenson, J. W., HRC & CC 1985, 8, 407–411.
29. Macka, M., Yang, W.-C., Zakaria, P., Shitangkoon, A., Hilder, E. F., Andersson, P., Nesterenko, P.,
and Haddad, P. R., J. Chromatogr. A 2004, 1039, 193–199.
30. Mikkers, F. E. P., Everaerts, F. M., and Verheggen, T. P. E. M., J. Chromatogr. 1979, 169,
11–20.
31. Tanyanyiwa, J., Leuthardt, S., and Hauser, P. C., J. Chromatogr. A 2002, 978, 205–211.
32. Lacey, M. E., Webb, A. G., and Sweedler, J. V., Anal. Chem. 2002, 74, 4583–4587.
35. Lacey, M. E., Webb, A. G., and Sweedler, J. V., Anal. Chem. 2000, 72, 4991–4998.
36. Thormann, W., Zhang, C.-X., Caslavska, J., Gebauer, P., and Mosher, R. A., Anal. Chem. 1998, 70,
549–562.
37. Berezovski, M. and Krylov, S. N., Anal. Chem. 2004, 76, 7114–7117.
38. Burgi, D. S., Salomon, K., and Chien, R.-L., J.Liq. Chromatogr. 1991, 14, 847–867.
39. Knox, J. H. and McCormack, K. A., Chromatographia 1994, 38, 207–214.
40. Davis, K. L., Liu, K.-L. K., Lanan, M., and Morris, M. D., Anal. Chem 1993, 65, 293–298.
41. Wätzig, H., Chromatographia 1992, 33.
42. Ross, D., Gaitan, M., and Locascio, L. E., Anal. Chem. 2001, 73, 4117–4123.
43. Isono, T., J. Chem.Eng. Data 1984, 29, 45–52.
44. Kuldvee, R., Vunder, K., and Kaljurand, M., Proc. Estonian Acad. Sci. Chem. 1999, 48, 119–128.
45. Guttmann, A. and Cooke, N., J. Chromatogr. 1991, 559.
46. Dow-Corning, in 10-898E-01, F. N. (Ed.), http://www.dowcorning.com/DataFiles/090007b5802e201c
.pdf, 2005.
47. Boedeker Plastics, I., 2004.
48. Kerbow, D. L. and Sperati, C. A., in Brandrup, J., Immergut, E. H., and Grulke, E. A. (Eds.), Polymer
Handbook 4th Edition, J. Wiley & Sons, Hoboken, 1999, p. V/36.
Handbook 4th Edition, J. Wiley and Sons, Hoboken, 1999, p. V/44.
Handbook 4th Edition, J. Wiley and Sons, Hoboken, 1999, p. V/51.
51. Mehta, R. H., in Brandrup, J., Immergut, E. H., and Grulke, E. A. (Eds.), Polymer Handbook 4th
Edition, J. Wiley & Sons, Hoboken, 1999, p. V/93.
52. Wunderlich, W., in Brandrup, J., Immergut, E. H., and Grulke, E. A. (Eds.), Polymer Handbook 4th
Edition, J. Wiley & Sons, Hobken, 1999, p. V/88.
53. Schrader, D., in Brandrup, J., Immergut, E. H., and Grulke, E. A. (Eds.), Polymer Handbook 4th
Edition, J. Wiley & Sons, Hobken, 1999, p. V/93.
Handbook 4th Edition, J. Wiley & Sons, Hoboken 1999, p. V/56.
19 Isoelectric Focusing in
Capillary Systems
Jiaqi Wu, Tiemin Huang, and Janusz Pawliszyn
CONTENTS
19.1 Introduction of Capillary Isoelectric Focusing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 563

19.2 Review of cIEF Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 564
19.2.1 Conventional cIEF. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 564
19.2.2 Applications of the Conventional cIEF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 566
19.2.3 Difficulties Associated with Conventional cIEF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 566
19.2.4 Whole-Column Detection cIEF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 566
19.3 Theoretical Aspects of cIEF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 567
19.3.1 Nature pH Gradient and Steady-State IEF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 567
19.3.2 The Stability of pH Gradient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 568
19.3.3 Resolution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 568
19.4 Examples of Practical Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 568
19.4.1 Monoclonal Antibody 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 569
19.4.2 Recombinant Protein 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 569
19.4.3 Protein Conjugate 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 569
19.4.4 Deactivated Virus 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 570
19.4.5 Monoclonal Antibody 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 570
19.5.1 Initial Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 572
19.5.2 Method Development Flow Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575
19.5.3 pI Determination and Peak Identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575
19.6 Expanding cIEF Applications and Future Prospects of cIEF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 577
19.6.1 Online cIEF-Mass Spectrometer Coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 577
19.6.2 Multiple Dimension Separation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 577
19.6.3 Fluorescence Whole-Column Detection for cIEF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 577
19.6.4 New Applications of Whole-Column Detection cIEF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 578
19.6.4.1 Study of Protein Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 578
19.6.4.2 Estimation of Protein Molecular Weight Based on Their Diffusion
Coefficients Determined by Whole-Column Detection cIEF . . . . . . . . . . . 578
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 578
19.1 INTRODUCTION OF CAPILLARY ISOELECTRIC FOCUSING

Isoelectric focusing (IEF) [1] is a powerful electrophoretic method for characterizing proteins and
other biopolymers. IEF separates amphoteric substances based on their isoelectric point (pI) differ-
ences. In an electric field, a charged particle is subjected to the force, F (F = QE, Q is the charge of
the particle and E the strength of the electric field). At the pI, where the charge of the particle equals
zero, the mobility of the particle should be zero. IEF is a special electrophoresis performed in a pH
563
gradient. The result of an IEF separation is that amphoteric substances in a sample are separated and
focused at the positions along the pH gradient where their pI equals the pH.
IEF has been applied mainly to two areas in the biological sciences: the analysis of complex
protein mixtures and the characterization of purified proteins [2]. For the first application, IEF is
used as a separation dimension in multidimensional electrophoretical methods, such as 2-D poly-
acrylamide gel electrophoresis (2-D PAGE). The 2-D PAGE is a common tool in proteomics. For
the second application, IEF is one of the two most common methods (another one is ion exchange
chromatography) to characterize a protein’s charge heterogeneity. The charge heterogeneity of a pro-
tein is frequently caused by post-translational modifications. Thus, the IEF method is an important
tool in the study of these modifications. Pharmaceutical companies utilize IEF methods to monitor
product processing, study formulations, and perform quality control for their protein products.
As a major protein characterization tool in biotech laboratories, slab gel IEF is slow, labor
intensive, and semiquantitative. In addition, the quality of the results depends on the skill of the
analyst. It was recognized that if IEF could be performed in a column format, significant advantages
over slab gel IEF could be realized, in terms of automation, speed, and quantification. In 1985,
Hjertén and Zhu [3] first reported an IEF method performed in a capillary column format—capillary
isoelectric focusing (cIEF). In the early years of cIEF, it was performed on commercial capillary
electrophoresis (CE) instruments that were designed for multiple CE modes. cIEF was expected
to have both the high resolution of slab gel IEF and the advantages of a column-based separation
technology that include automation and quantification.
Despite the inherent appeal of cIEF, since it was first introduced more than 20 years ago,
widespread acceptance of cIEF by biotechnology laboratories and the pharmaceutical industry as a
substitute for slab gel IEF for protein characterization did not occur. The main factor for this slow
acceptance by the industry was related to the difficulty in performing cIEF using general purpose CE
instruments (conventional cIEF). Subsequently, the introduction of whole-column detection cIEF
[4] and its commercialization [5] revolutionized cIEF technology. It has been quickly adopted by
leading pharmaceutical companies and research laboratories. Many analysis methods based on the
whole-column detection cIEF have been validated in laboratories regulated by the U.S. Food and
Drug Administration (FDA) [6,7]. To our knowledge, the top 10 pharmaceutical companies in the
world are using the whole-column detection cIEF technique for protein characterization, for drug
discovery, product processing, including formulation development, and quality control (QC).
It has been almost 25 years since CE was first reported [8]. The late 1980s and early 1990s
were CE’s booming years. During that period, it was widely expected that CE would replace high-
performance liquid chromatography (HPLC) for most HPLC applications, although this did not occur.
After the 25 years, CE almost disappeared from industrial laboratories. Only two CE technologies
survived and were accepted by biotech pharmaceutical companies, including CE-sodium dodecyl
sulfate (SDS) and cIEF. CE-SDS could replace SDS–PAGE and cIEF could replace slab gel IEF for
many applications, especially in protein characterization and QC. The speed of the acceptance of
these two CE technologies in the biotech field has been accelerated in the past 5 years [9] due to the
improvements in these two CE technologies.
In this chapter, a review of cIEF applications for protein characterizations performed by biotech
and pharmaceutical laboratories will be presented.
19.2 REVIEW OF CIEF TECHNOLOGY

19.2.1 CONVENTIONAL cIEF
When a general purpose CE instrument is used to perform cIEF (conventional cIEF), as shown in
Figure 19.1a, the separation column, which is usually a 50–100 µm inner diameter, 25–60 cm long
capillary, is filled with a mixture of a protein sample and carrier ampholytes (CAs). After the sample
is injected, the two ends of the capillary column are dipped into the catholyte and the anolyte, as
Isoelectric Focusing in Capillary Systems 565
(a) Capillary column

High
pressure
Detection
point
Waste Sample
Sample injection
(b) – Separation voltage +
pH gradient Capillary column
Detection Sample zones

point
–
OH H+
Catholyte Anolyte
(c) Focusing
– separation voltage
+
pH gradient
Capillary column
Low
pressure
Detection Sample zones
point
– +
OH H
Catholyte Anolyte
Mobilization
FIGURE 19.1 Steps involved in cIEF using general purpose CE instruments (conventional cIEF). (a) Sample
injection, (b) focusing step, and (c) mobilization step.
shown in Figure 19.1b. A separation voltage (usually at 100–500 V/cm) is applied to the two ends of
the capillary column. Under the separation voltage, a pH gradient is established along the capillary
column from the anodic end of the column to the cathodic end of the column by isoelectric stacking
of components in the CAs [10,11]. In the pH gradient, driven by the separation voltage, proteins in
the sample migrate to the points within the capillary column where their isoelectric point (pI) values
are equivalent to the pH values and the migration then stops. The proteins are focused into very
narrow zones at their pI points.
There are three basic requirements of a separation column for cIEF; zero or substantially reduced
electroosmotic flow (EOF), hydrodynamic flow, and interaction between protein samples and the
column wall. Columns used in cIEF are usually coated on their inner wall with stable, neutral coatings
to control for EOF and prevent the protein from interaction with the wall [10]. During the focusing
process, the two ends of the separation column should be maintained at the same level to eliminate
the hydrodynamic flow. Under these conditions, at the end of the focusing process, all protein zones
are stationary or near stationary within the separation column [3].
Since general purpose CE instruments are only equipped with a single point on-column detector,
which is usually located close to one end of the capillary column as shown in Figure 19.1c, these
stationary protein zones within the column have to be mobilized past the detection point of the single
point detector. For cIEF performed on these CE instruments (conventional cIEF), a mobilization
process is necessary following the focusing process [10,11]. There are three ways of performing the
mobilization: pressure-driven mobilization [11], salt mobilization [3], and EOF-driven mobilization
[12]. These three methods can be used in any combination.
Pressure is often used for mobilization. As shown in Figure 19.1c, at the end of the focusing a
low pressure (usually about 0.5 psi) is applied to one end of the capillary column to mobilize the
entire pH gradient past the detection point. During the mobilization, the separation voltage remains
to reduce the sample zone broadening.
With salt mobilization, at the end of the focusing process salts are added into one of the elec-
trolytes. The addition of the salts creates a pH shift at this end of the column under the separation
voltage, and then the shift gradually progresses deeper into the column. This causes the whole pH
gradient within the column to shift toward this end of the column.
In EOF-driven mobilization, the EOF of the separation column is not totally eliminated, but is
controlled with additives, such as methylcelulose (M.C.). Upon application of the separation voltage,
the focusing process is initiated. Concurrently, the whole volume of solution within the column is
mobilized toward the cathodic end of the column, driven by the EOF. In this way, the mobilization
process is combined with the focusing process.
19.2.2 APPLICATIONS OF THE CONVENTIONAL cIEF

There are a range of applications of conventional cIEF to real-world analysis, including the char-
acterization of protein samples in laboratories of biotech pharmaceutical companies, and clinical
analysis. In the pharmaceutical industry, some methods have been developed based on this technol-
ogy [13,14] and validated [15–17] as the identity assays in regulated pharmaceutical laboratories
for the analysis of therapeutic monoclonal antibodies and glycoproteins. cIEF has also been used in
clinical analysis for human hemoglobin analysis [18,19] and the analysis of proteins in cerebrospinal
fluid [20]. There are several reviews on the applications of the conventional cIEF [10,21]. However,
the pace of acceptance of conventional cIEF for the analysis of real-world samples is slow.
19.2.3 DIFFICULTIES ASSOCIATED WITH CONVENTIONAL cIEF

The mobilization process in conventional cIEF inhibits its widespread use. The process introduces
many problems, such as poor resolution, poor reproducibility, and long sample analysis time (less than
2 samples/h). In conventional cIEF, the dynamic process of IEF within the separation column is not
monitored, which makes it difficult to optimize focusing time and mobilization parameters. Focusing
time is the most important parameter in cIEF. In addition, the real focusing time in conventional
cIEF is equal to the focusing time and the mobilization time, since the separation voltage is always
on during the mobilization process. This makes it almost impossible to optimize the focusing time
because the mobilization time of a sample component depends on its pI value. It is also difficult to
detect problems in the focusing process within the separation column, such as sample aggregation
and precipitation.
19.2.4 WHOLE-COLUMN DETECTION cIEF

The introduction of whole-column detection cIEF technology [22,23] solved the problems caused by
the mobilization process in conventional cIEF. Figure 19.2 illustrates the principle of whole-column
detection cIEF. The separation column is a short capillary (typically, 50 mm long, 100 µm inner
diameter silica capillary). The inner wall of the capillary is coated to eliminate EOF and prevent
Separation
voltage –
+
H+ Capillary OH–
Outlet column Inlet Sample injection
Column wash
Semipermeable
Sample zones tube
Whole-column detection
FIGURE 19.2 Principle of whole-column detection cIEF.
proteins from interacting with the wall surface. The whole-column detection system is an ultraviolet
(UV) absorption-imaging detector operated at 280 nm. During the operation of cIEF, the two tanks
on the cartridge are filled with the catholyte and anolyte, respectively, as shown in Figure 19.2.
A protein sample premixed with CAs is injected from the inlet of the column into the separation
column between the two semipermeable membrane tubes. After the column is filled with the sample,
the sample injection is stopped. A separation voltage, typically, 3 kV (600 V/cm), is applied across
the two tanks. IEF only occurs within the separation column between the two semipermeable tubes.
The structure of the separation column makes accurate sample injection unnecessary. As long as
the separation column between the two semipermeable tubes is filled with the sample solution,
the injection is quantitative. The IEF process within the entire length of the separation column is
monitored by the whole-column detection system. At the end of the focusing, after recording the last
image, the sample is washed out of the column by injecting a wash solution from the inlet, then, the
separation column is then ready for the next sample.
Since the whole-column detector monitors the IEF process within the separation column in
an online fashion, the focusing time can be determined in a single sample run. At the end of the
focusing process, all sample zones within the capillary column are simultaneously recorded by
the detector without disturbing the separation resolution. Any sample precipitation and aggregation
during focusing can be observed and distinguished while focusing is in process. Sample precipitation
and aggregation are the two most common problems in IEF faced while obtaining reproducible results.
Different additives may be used to improve peak pattern reproducibility. These features facilitate fast
method development. In turn, the commercialization of the whole-column detection cIEF technology
accelerated the acceptance of cIEF technology by pharmaceutical companies.
19.3 THEORETICAL ASPECTS OF CIEF

19.3.1 NATURE pH GRADIENT AND STEADY-STATE IEF
The separation of ampholytes in a pH gradient formed by CAs was proposed by Svensson in the
early 1960s [1,24–26]. It is known that the pH value of a pure ampholyte solution is approximately
its isoelectric point [24]. In a Svensson’s IEF system, the electrode reactions must be the electrolysis
of water ions, which ensures that protons are produced at the anode and hydroxide ions are produced
at the cathode. Before the electric field is applied, components of CA are uniformly distributed
throughout the separation channel between the anode and the cathode. The pH along the channel is
uniform, representing the average pH of the mixture of all of the components of the CAs. Upon the
application of electric field, influenced by the migration of protons from the anode and hydroxide
ions from the cathode of the separation channel, the components of the CAs begin to separate. The
components with pI values higher than the average pH are positively charged and migrate toward the
cathode whereas the components with pI values lower than the average pH are negatively charged
and migrate toward the anode. The components cease migrating when their net charge becomes zero,
at the region where the pH equals their pIs. Consequently, a smooth and stable pH gradient, which
is positive all the way from anode to cathode, can be “naturally” generated.
19.3.2 THE STABILITY OF pH GRADIENT

Svensson [24,25] regarded the natural pH gradient IEF as a true steady-state process and expected
its unlimited stability. However, as noted by Svensson, a true steady state was never reached in gel
IEF [27]. Instabilities of pH gradient, such as the plateau phenomenon, cathodic drift, anodic drift,
or symmetric drift, were commonly observed [1,27–34]. There have been hypotheses to explain
the instability of the pH gradient in IEF by isotachophoresis (ITP) mechanism [35,36] or stationary
neutralization reaction boundary equilibriums (SNRBEs) [37]. Hjerten et al. [38] suggested, as he
proposed the mechanism for chemical mobilization in cIEF, that pH instability is inherent in natural
pH gradient IEF due to the need for electroneutral conditions.
Nevertheless, it is commonly accepted that IEF is a steady-state technique, but this steady
state is limited and conditional. To exploit the high-resolution capability of IEF, it is necessary to
understand IEF’s uniqueness. IEF is a steady state or equilibrium technique in nature. Therefore,
factors that affect the steady-state IEF should be avoided. For example, as mentioned in Section
19.2, interferences that disturb the formation and stability of the pH gradient, such as the existence
of EOF, hydrodynamic flow within the separation channel, or impurities in the anolyte and catholyte
(even adsorption of CO2 ), should be controlled.
19.3.3 RESOLUTION
The resolution in IEF is defined by the resolving power on pI (pI) that depends on the diffusion
coefficient of the ampholyte, D, the electric field, E, the mobility slope (–du/dpH), and the pH
gradient (dpH/dx) [1]

D dpH dx
pI = 3 .
E −dµ dpH
The above equation shows that resolution can be improved by high E, low D, a narrow pH
gradient, and a high mobility slope.
Experimentally, electric field, diffusion coefficient, and dpH/dx can be manipulated. The rate
of change of mobility with a pH near the isoelectric point is intrinsic to the ampholytes and cannot
be changed. Performing IEF under higher applied voltages facilitates the development of narrow
focused protein zones. However, an increase in the electric field will intensify the Joule heating. As
a result, a moderate voltage, which can be determined experimentally, is applied in IEF.
19.4 EXAMPLES OF PRACTICAL APPLICATIONS

In this section, several examples of method development procedures for protein characterization
assays used for samples from biotech pharmaceutical companies are presented.
0.78
0.58 pI marker 9.6

Absorbance pI marker 7.5
0.38
0.18
–0.02
0 500 1000 1500 2000
Peak position (pixel)
FIGURE 19.3 (Trace 1) Monoclonal antibody 1 at 0.2 mg/mL in 4% Pharmalyte pH 3–10 and 0.35% MC.
Focusing time was 6 min at 600 V/cm with a 1 min prefocusing at 300 V/cm. (Traces 2 and 3) Two consecutive
runs of the same sample were performed in the same running buffer and under the same focusing conditions
as that of trace 1 Sample concentration was 0.6 mg/mL. Two pI markers, pI 7.5 and pI 9.6, were spiked for pI
calibration.
19.4.1 MONOCLONAL ANTIBODY 1

First, following the initial conditions, 4% Pharmalyte (pH 3–10), and 0.2 mg/mL sample concentra-
tion were tested for this sample. Under the conditions, as shown in the trace 1 of Figure 19.3, the
sample peak height is detected at 0.1 Abs. In the next run, the sample concentration was increased to
0.6 mg/mL. The focusing time was 6 min at 600 V/cm with a 1 min prefocusing at 300 V/cm. Two
pI markers were spiked into the sample for pI calibration. The results show good reproducibility in
peak pattern (traces 2 and 3 in Figure 19.3). Since the separation resolution under these conditions
was satisfactory, no further method development was pursued to enhance the resolution.
19.4.2 RECOMBINANT PROTEIN 1

This protein was expected to be very heterogeneous (multiple peaks). At the beginning, higher
than usual sample concentration (0.4 mg/mL) was tested. However, as shown in the trace 1 of
Figure 19.4, the peak height of the major peaks was only in 0.03 Abs range. In the next run, the
sample concentration was boosted to 2.5 mg/mL in order to increase the peak height to above the
0.1 Abs level. Since the sample was focused into about 15 peaks, higher resolution was required. To
enhance the resolution, Pharmalyte mixture (pH 3–10 and pH 5–8 at 1:3 ratio) was tested. As shown
in traces 2 and 3 of Figure 19.4, the resolution was improved under these conditions.
19.4.3 PROTEIN CONJUGATE 1

The purpose of developing an IEF assay for this sample was to monitor the pI values of different
production lots of the product. For this sample, it was very difficult to obtain reproducible peak
patterns that could match the peak pattern reproducibility of the previous two samples. As shown in
Figure 19.5, the sample started to aggregate and was focused into an unstable big spike when the
focusing time was longer than 4.5 min. Many additives and denatured conditions were tested, but
no solution was found. However, if the focusing time was limited to 4 min (before sample started
to aggregate), the peak pattern reproducibility was acceptable. The peak pattern reproducibility is
shown in Figure 19.6 for this protein conjugate. This cIEF method is therefore useful in detecting pI
changes in different lots.
0.27
pI marker 5.3
pI marker 7.9
Absorbance 0.17
0.07
–0.03
0 500 1000 1500 2000
FIGURE 19.4 (Trace 1) Recombinant protein 1 at 0.4 mg/mL in 4% Pharmalyte pH 3–10 and 0.35% MC.
Focusing time was 6 min at 600 V/cm with a 1 min prefocusing at 300 V/cm. (Traces 2 and 3) Two consecutive
runs of the sample were performed at a concentration of 2.5 mg/mL in 1% Pharmalyte pH 3–10, 3% Pharmalyte
pH 5–8, and 0.35% MC. Focusing time was 6 min at 600 V/cm with a 1 min prefocusing at 300 V/cm. Two pI
markers, pI 5.3 and pI 7.9, were spiked into the sample for pI calibration.
0.09
3.5 min
Absorbance
0.04 4.5 min
6 min
–0.01
0 500 1000 1500 2000
FIGURE 19.5 Effect of focusing time on sample protein conjugate 1’s peak pattern.
19.4.4 DEACTIVATED VIRUS 1

A second sample was also analyzed to determine pI values of different lots in production. As shown
in Figure 19.7, when focusing time was limited to 6 min, the peak pattern reproducibility was
satisfactory for the application. From our experiences with cIEF method development for virus
samples, cIEF may not achieve the peak pattern reproducibility that can match that of monoclonal
antibodies or other protein samples. However, in most cases, the reproducibility is satisfactory for
the desired applications [39].
19.4.5 MONOCLONAL ANTIBODY 2

The monocolonal antibody 2 protein started to aggregate before the completion of the focusing
process (Figure 19.8). In the early stage of the focusing process, it went smoothly. The protein was
focused into well-defined peaks. However, before the completion of the focusing process, a new peak
0.19 pI marker 6.5

pI marker 4.4
0.14
Absorbance
0.09
0.04
–0.01
0 500 1000 1500 2000
FIGURE 19.6 Four consecutive runs of protein conjugate 1 at 0.25 mg/mL in 4% Pharmalyte pH 3–10 and
0.35% MC. Focusing time was 4 min at 600 V/cm. Two pI markers, pI 4.4 and pI 6.5, were spiked into the
sample for pI calibration.
pI marker 4.4
0.28 pI marker 7.7
Absorbance
0.18
0.08
–0.02
0 500 1000 1500 2000
FIGURE 19.7 Three consecutive runs of deactivated virus 1 at 0.15 mg/mL in 4% Pharmalyte pH 3–10 and
0.35% MC. Focusing time was 6 min at 600 V/cm. Two pI markers, pI 4.4 and pI 7.7, were spiked into the
sample for pI calibration.
(indicated by an arrow in Figure 19.8) started to appear and developed quickly along the focusing
time. This is a typical symptom of sample aggregation created by the focusing process. In order to
eliminate the problem, 4 M urea was added to the sample, effectively eliminating the aggregation
problem (Figure 19.9), resulting in a reproducible peak pattern. This example illustrates the power
of the whole-column detection cIEF for facilitating fast method development.

In this section, the development of protein characterization methods for samples in typical pharma-
ceutical laboratories will be examined. All of the method development procedures discussed in this
chapter are based on the use of the whole-column detection cIEF instrument, since the IEF process
can be observed and the easily associated with optimizing the focusing time. However, the basic
conditions described in this chapter are also applicable to conventional cIEF method.
0 min
pH
1 min
2 min
3 min
0.95
4 min
5 min
Absorbance
6 min
7 min
8 min
0.45 9 min
10 min
11 min
12 min
13 min
–0.05
0 500 1000 1500 2000
FIGURE 19.8 Focusing process of sample monoclonal antibody 2. The sample concentration was 0.2 mg/mL.
Focusing conditions: 4% Pharmalyte pH 5–8 and 600 V/cm.
0.18
10 min
Absorbance
0.08 15 min
20 min
–0.02
0 500 1000 1500 2000
FIGURE 19.9 Focusing process of sample monoclonal antibody 2. The sample concentration was 0.2 mg/mL.
Focusing conditions: 4% Pharmalyte pH 5–8 and 600 V/cm; 4 M urea was added to the sample.
19.5.1 INITIAL CONDITIONS

At the beginning of the development of a cIEF method for a new sample, the initial conditions for the
first sample run must be determined. These conditions include brand name, pH range, pH gradient
linearity, and concentration of CAs, electrolytes, additives, sample concentration, focusing voltage,
and focusing time.
0.48
pI marker 4.22
0.38
Absorbance Ampholine 3.5–9.5

0.28
Biolyte 3–10
0.18
pI marker 9.46
Pharmalyte 3–10
0.08
Servalyt 2–11
–0.02
0 500 1000 1500 2000
FIGURE 19.10 Electropherograms of four commercially available CAs. Concentrations of the CAs were
8%. Focusing time was 6 min at 600 V/cm. Two internal pI markers, pI 4.22 and pI 9.46, were spiked into the
samples.
pI marker 4.22 pI marker 9.46

0.18
Absorbance
0.08
–0.02
0 500 1000 1500 2000
FIGURE 19.11 Three consecutive runs of Pharmalyte pH 3–10. Concentration was 8% and focusing time
was 6 min at 600 V/cm. Two internal pI markers, pI 4.22 and pI 9.46, were spiked into the sample.
The most important condition in cIEF is the CA. As mentioned in Section 19.3, CAs are “special
buffers” in cIEF. A CA may contain over 900 amphoteric components [2,40]. According to the
literature [2] and to our knowledge, there are four different kinds of commercial CAs: Ampholine,
Biolyte, Pharmalyte, and Servalyt. Experiments are needed to determine which CA results in the
best resolution for a given sample. Thus, the resolution cannot be the first criterion in CA selection
because it is unknown for a given protein sample. Usually, a wide pH range CA is the first choice for
a sample of unknown pI value. From our experience, the first criterion in selecting the CA for cIEF
should be the background UV absorption. All these CAs were designed for slab gel IEF in which
the CA’s UV absorption was not a concern [41]. Figure 19.10 shows the background absorption of
the four CAs at 280 nm along their entire pH gradient. In addition, as shown in Figure 19.11, the
background profiles of these CAs are reproducible from run to run and relatively stable for different
lots. From the point view of background absorption, Pharmalyte should be the first choice due to its
low background along its entire pH gradient, as illustrated in Figures 19.10 and 19.11.
Because of the UV absorption of the CAs, from background point of view, lower the concentration
of the CAs is in samples, the better the signal-to-noise ratio will be. However, the buffer capacity of
the CAs is determined by their concentration [2]. The sample peaks become broad and the relationship
between the peak height and the sample concentration becomes nonlinear when the buffer capacity
of the CAs is saturated. There should therefore be a trade-off between the buffer capacity and the
background. From the literature and our experiences, a 2–8% Pharmalyte solution (the stock solution
is considered to be 100%) results in acceptable background noise, while still maintaining enough
buffer capacity for most proteins at working concentrations. Thus, it is recommended that a 4%
Pharmalyte solution is a good starting point for method development.
As we will discuss later in this chapter, for many samples, separation resolution can be enhanced
with the use of narrow pH range CAs. In addition, if the pI value of a sample is known, a narrow
pH range CA can be selected for the initial conditions. It should be noted that there has been a
misunderstanding associated with the use of narrow pH range CAs. Traditionally, it is believed that
narrow pH range CAs must be mixed with one or more wide pH range CAs, or one or more narrow
pH range CAs of different pH ranges. From our experience, this is not necessary. A single narrow
pH range CA can be used alone for cIEF. Actually, the best resolution is usually achieved using a
single narrow pH range CA. For a given protein sample, as long as the pH range of a single narrow
pH range CA coves the pIs of the protein, this CA should be the first choice.
Literature suggests that weak acids and bases should be used as the electrolytes when narrow pH
range CAs are used [10,11]. In our experiments, we found that although weak acids and bases gave
similar results, if not better, to that when using strong acids and bases; they had to be frequently
refreshed in order to keep consistent results for a long batch of sample runs. This is probably due
to their weak buffer capacity. In our experiments, the same anolyte (0.08 M H3 PO4 ) and catholyte
(0.1 M NaOH) are used for all CAs of wide pH ranges and narrow pH ranges between 4% and 8%.
The advantage of this approach is the relative simplicity of subsequent method development. Rather,
different CAs can be tested without changing the electrolytes.
Focusing time is the second most important parameter for cIEF. Before the focusing time is
chosen, the focusing voltage must be determined. Although theoretically, as described in the Section
19.3.3, the separation resolution is higher when a higher voltage is applied, the detection noise also
increases along with the voltage [42]. The optimal voltage is around 500 V/cm [42].
Under the focusing voltage, focusing time of different samples can be easily optimized with
the use of the whole-column detection cIEF because it monitors the focusing process within the
separation column in an online fashion. There is a common misunderstanding related to focusing
time in cIEF. The literature suggests that one must determine the end of the focusing process by
the focusing current. It is believed that the focusing process is complete when the focusing current
stabilizes. However, in our experiments using the whole-column detection cIEF, we found that there
is no such relationship between the end of the focusing process and the current. The stabilization
of the focusing current only roughly reflects the end of focusing for the CAs. Proteins are usually
focused at a slower speed compared to CAs. Some proteins, such as PEGylated proteins, may require
an additional 20 min of focusing time after the focusing current stabilizes. Therefore, focusing time
has to be determined experimentally for each sample. In addition, the focusing time can be affected
by the concentration of salts in the samples. A higher salt concentration will result in a faster focusing
process, since the pH gradient created by the CAs is narrowed by the salts [43]. The focusing speed
becomes slower when the concentration of CAs is higher. In addition, the focusing speed is much
slower when narrow pH range CAs are used. From our experience, for most protein samples, when
4% pH 3–10 Pharmalyte is used, a 6 min focusing time is a good starting point.
For all samples, especially samples in a salt matrix, performing a prefocusing step at a voltage
lower than focusing voltage reduces detection noise and increases peak pattern reproducibility. The
prefocusing voltage is usually half of the focusing voltage. The step can be 1–2 min long depending
on the samples.
At the end of the cIEF process, proteins are focused into narrow zones within the separation
column and are concentrated hundreds of times. Confining proteins at their pI points (zero net
charge) and at high concentrations for a long time increases the likelihood that they will precipitate
or aggregate. Many chemicals can be used as the additives in cIEF [2] to stabilize proteins during
IEF, such as sugars, nonionic or zwitterionic surfactants, and urea, thus, effectively reducing the
likelihood of protein precipitation and aggregation occur during the focusing process.
For all cIEF methods, polymer solutions are added into the samples, including methyl cellulose
(MC) and (hydroxypropyl)methylcellulose (HPMC). These polymers modify the capillary surface
[10–12] and enhance separation resolution. The existence of the polymer in the sample solution
reduces diffusion coefficients of the proteins, thus, as described in Section 19.3.3, enhancing the
separation resolution. In all cIEF applications described in this chapter, 0.35–0.5% of MC are added
to the samples.
Because of the UV absorption of the CAs used in cIEF, all UV absorption detections in cIEF are
performed at 280 nm. At 280 nm, the absorption of the proteins is determined by the presence of
two amino acids in the proteins: tyrosine and tryptophan. Thus, the sensitivity of proteins at 280 nm
varies a great deal from protein to protein. It is not easy to determine the initial sample concentration.
However, we found that the sensitivity of monoclonal antibodies is relatively even and predictable.
For these samples and many other protein samples, 0.2 mg/mL is a good starting point for sample
concentration if the samples have only one major peak (the major peak is defined as a peak having
>50% of the total protein in a sample). If the samples can be separated into several major peaks (each
major peak should be >20% of the total protein in a sample), the starting point can be calculated as
follows:
Number of major peaks × 0.1mg/mL
In summary, for an unknown sample, the initial cIEF analysis conditions would include 4% pH
3–10 Pharmalyte, 0.5% MC as the additive, 0.08 M H3 PO4 as the anolyte, 0.1 M NaOH as the
catholyte, 0.2 mg/mL sample concentration, and 6 min focusing time at 500 V/cm voltage with a
1–2 min prefocusing step.
19.5.2 METHOD DEVELOPMENT FLOW CHART

Figure 19.12 illustrates a flow chart for the development of the cIEF method. At the beginning,
it is recommended that the initial conditions described in Section 19.5.1 are tested. The second
step should then involve adjusting the sample concentration. The third step is the critical step in
the method development because a high-resolution method is based on good reproducibility. All
methods should be attempted to ensure a reproducible peak pattern. Then, the separation resolution
should be enhanced with the use of narrow pH range CAs. The final step in the method development
is pI determination and sample peak identification.
19.5.3 pI DETERMINATION AND PEAK IDENTIFICATION

Determination of the pI of a protein using cIEF requires two basic conditions: pI markers with
accurate pI values and a linear pH gradient created by the chosen CAs. Using pI markers is the
only way to characterize a pH gradient within the capillary column created by CAs [44]. As for the
second condition, a pI value in a nonlinear pH gradient is difficult to characterize using a limited
number of pI markers since the number of pI markers in a given pH region is always limited. In most
cases, these pI markers do not distribute evenly within the pH region. Curve fitting in a nonlinear pH
gradient using these pI markers involves a large error that is difficult to estimate. The determination
of the pI value should be performed in a single CA with good linearity in its pH gradient.
Initial conditions
Adjust sample
concentration
If Abs < 0.1 or > 0.6
Yes
No
1. Additives
2. Different CAs
Peak pattern
No reproducible?
Yes
1. Narrow range CAs
2. CA mixtures
Need higher
Yes resolution?
No
Results
FIGURE 19.12 cIEF method development.
2000
Pharmalyte 3–10
1500
1000
r2 = 0.9883
500
0
3 5 7 9
pH
FIGURE 19.13 pH gradient profile of 8% Pharmalyte pH 3–10.
Profiling the pH gradient of different CAs is important for this purpose. Figure 19.13 illustrates
an example with a pH gradient for Pharmalyte pH 3–10. The overall linearity of the pH gradient
is good (r 2 is about .99). However, the gradient bends at two places, around pH 5 and 8. If pI
determination is performed around these two regions, multiple pI markers should be used to ensure
good accuracy. However, in other pH regions of this CA, the linearity in the pH gradient is well
above r 2 = .99. Two pI markers should therefore be adequate to calibrate the pI values.
In most practical applications, the determined pI values of a sample are only used to identify
component peaks. The true pI value of a peak is not relevant in the applications, as long as the
determined pI value of this peak (component) is reproducible in all samples run under identical
experimental conditions. It has been proven that peaks (components) in different samples can be
identified using the determined pI values within a standard deviation of 0.1 pH units (<2% RSD)
by using two internal pI markers, regardless of the linearity of the pH gradient created by the CAs
[43]. This method applies to salt concentrations up to 15 mM in final sample solutions.
19.6 EXPANDING CIEF APPLICATIONS AND FUTURE PROSPECTS

OF CIEF
In the above sections, the cIEF applications of protein characterization in biotech laboratories
and pharmaceutical companies were discussed. In this section, other applications of the cIEF are
presented, as well as new developments in cIEF.
19.6.1 ONLINE cIEF-MASS SPECTROMETER COUPLING

Coupling cIEF to mass spectrometer is a perfect match because cIEF separates proteins only based
on their charges while the results of other CE methods are more or less related to the molecular
weight of the samples. Early demonstrations of cIEF-MS work were published in 1996 [45,46]. For
this application, a mobilization process after the focusing process is necessary to move sample zones
into the MS detector. This technology has recently been applied to the field of proteomics [47–50].
One example is the analysis of Escherichia coli whole cell lysates using cIEF-MS for the study of
intact proteins [50]. A complete discussion of the applications of cIEF in proteomics is discussed in
a recent review [48].
The biggest difficulty in coupling cIEF to MS is that high concentration CAs [47] and most of
the additives used in cIEF are not compatible with MS. As we have seen in the previous sections,
many protein samples require these additives at high concentration in order to achieve reproducible
results. This method limitation must be addressed before cIEF-MS technology can be applied as a
tool for the analysis of real-world samples.
19.6.2 MULTIPLE DIMENSION SEPARATION

The high speed of separation and detection of the whole-column cIEF makes it attractive as a second
separation dimension in a two-dimension (2-D) separation scheme. For example, using a 1-cm-long
separation column, the IEF separation, and detection of the separated zones can be performed in
30 s [51]. As the second separation dimension, the whole-column detection cIEF has been success-
fully coupled to gel filtration chromatography [52]. The whole-column detection cIEF can also be
coupled with other CE separation dimensions [53].
19.6.3 FLUORESCENCE WHOLE-COLUMN DETECTION FOR cIEF

Using cIEF as the second separation dimension in a 2-D separation system may require high detection
sensitivity for cIEF because after the first dimension separation, the sample concentration can be
very low. In this case, UV absorption detection is apparently not enough. Fluorescence is considered
to be one of the most sensitive methods in optical detectors. A whole-column fluorescence detection
system has been used for cIEF and other CE modes [54]. Two schemes have been proposed for the
introduction of an excitation laser light beam into a narrow bore capillary column: side illumination
and axial illumination. Axial illumination has the advantages of simple optical alignment and low
background noise caused by scattering excitation light beam [55]. These advantages expanded the
application of the whole-column cIEF [56].
The difficulty of the fluorescence detector for the cIEF is the lack of a chromophore in most
proteins in the visible wavelength region. Labeling proteins with fluorescence dyes may change
their pI values. In addition, multiple labeling is unavoidable for high sensitive detection.
19.6.4 NEW APPLICATIONS OF WHOLE-COLUMN DETECTION cIEF

19.6.4.1 Study of Protein Interactions
Since the entire separation column is monitored in an online fashion, the whole-column detection
cIEF is an ideal tool for studies of protein–protein interactions and protein interactions with other
molecules, such as phospholipids [57,58]. The basic principle of applying whole-column detection
cIEF for such studies is based on the pI differences of the proteins and their complexes after interaction
with the molecules.
19.6.4.2 Estimation of Protein Molecular Weight Based on Their Diffusion

Coefficients Determined by Whole-Column Detection cIEF
At the end of focusing process, if the separation voltage is stopped, the protein zones focused within
the cIEF column will start to diffuse. This diffusion is almost one dimensional along the axis of the
capillary column because of its narrow inner diameter. The diffusion process can be well monitored
by the whole-column detection system. From this process, the protein’s diffusion coefficient can be
calculated. Once the diffusion coefficient is known, the protein’s molecular weight will be estimated
[59]. One potential application of this method is the determination of aggregation. If the aggregation
peaks of a protein can be separated from its monomer, the aggregation peaks should be identified
using the diffusion method since the difference of their molecular weight from that of the monomer
will be large.
REFERENCES
1. Righett, P. G. Isoelectric Focusing: Theory, Methodology and Application, Elsevier, Amsterdam, 1983.
2. Righetti, P. G., Stoyanov, A. V., Zhukov, M. Y. The Proteome Revisited, Elsevier, Amsterdam, 2001.
3. Hjertén, S., Zhu, M. J. Chromatogr., 1985, 346, 265–270.
4. Wu, J., Pawliszyn, J. Am. Lab., 1994, October, 48–52.
5. Wu, J., Watson, A. H., Torres, A. R. Biotech. Lab., 1999, 17, 24–26
6. Janini, G., Saptharishi, N., Waselus, M., Soman, G. Electrophoresis, 2002, 23, 1605–1611.
7. Li, N., Kessler, K., Bass, L., Zeng, D. J. Pharmaceut. Biomed. Anal., 2006, doi: 10.1016/
jpba.2006.09.024.
8. Jorgenson, J. W., Lukacs, K. D. Anal. Chem., 1981, 53, 1298–1302.
9. Annual conference of CEPharm, see website www.casss.org.
10. Kilár, F. Electrophoresis, 2003, 24, 3908–3916.
11. Rodriguez-Diaz, R., Wehr, T., Zhu, M. Electrophoresis, 1997, 18, 2134–2144.
12. Mazzeo, J. R., Krull, I. S. Anal. Chem., 1991, 63, 2852–2857.
13. Silverman, C., Komar, M., Shields, K., Diegnan, G., Adamovics, J. J. Liq. Chromatogr., 1992, 15,
207–219.
14. Costello, M. A., Woititz, C., De Feo, J., Stremlo, D., Wen, L.-F. L., Palling, D. J., Iqbal, K., Guzman,
N. A. J. Liq. Chromatogr., 1992, 15, 1081–1097.
15. Hunt, G., Hotaling, T., Chen, A. B. J. Chromatogr. A, 1998, 800, 355–367.
16. Lasdun, A. M., Kurumbail, R. R., Leimgruber, N. K., Rathore, A. S. J. Chromatogr. A, 2001, 917,
147–158.
17. Schmerr, M. J., Cutlip, R. C., Jenny, A. J. Chromatogr. A, 1998, 802, 135–141.
18. Hempe, J. M., Craver, R. D. Clin. Chem., 1994, 40, 2288–2295.
19. Mario, N., Baudin, B., Giboudeau, J. J. Chromatogr. B, 1998, 706, 123–129.
20. Hiraoka, A., Tominaga, I., Hori, K. J. Chromatogr. A, 2002, 961, 147–153.
21. Righetti, P. G., Gelfi, C., Conti, M. J. Chromatogr. B, 1997, 699, 91–104.
22. Wu, J., Pawliszyn, J. Anal. Chem., 1992, 15, 224–227.
23. Wu, J., Watson, A. J. Chromatogr. B, 1998, 714, 113–118.
24. Svensson, H. Acta. Chem. Scan., 1961, 15, 325–341.
25. Svensson, H. Acta. Chem. Scan., 1962, 16, 456–466.

26. Vesterberg, O. Acta. Chem. Scan., 1969, 23, 2653–2666.
27. Rilbe, H. In: Electrofocusing and Isotachophoresis, Radola, B. J., Graesslin, D., eds., Walter de
Gruyter, New York, 1977.
28. Chrambach, A., Doerr, P., Finlayson, G. R., Miles, L. E. M., Sherins, R., Rodbard, D. Ann. NY Acad.
Sci., 1973, 290, 44–64.
29. Catsimpoolas, N. Isoelectric Focusing, Chapter 1, Academic Press, New York, 1976.
30. Murel, A., Kirjanen, I., Kirret, O. J. Chromatography, 1979, 174, 1–11.
31. Finlayson, G. R., Chrambach, A. Anal. Biochem., 1971, 40, 292–311.
32. Miles, L. E. M., Simmons, J. E., Chrambach, A. Anal. Biochem., 1972, 49, 109–117.
33. Arosio, P., Gianazza, E., Gighetti, P. G. J. Chromatography, 1978, 166, 55–64.
34. Nguyen, N. Y., Chrambach, A. Anal. Biochem., 1976, 74, 145–153.
35. Mosher, R. A., Thormann, W., Bier, M. J. Chromatogr., 1986, 351, 31–38.
36. Mosher, R. A., Thormann, W. Electrophoresis, 1990, 11, 717–723.
37. Cao, C. X. J. Chromatogr., 1998, 813, 153–171.
38. Hjerten, S., Liao, J., Yao K. J. Chromatogr., 1987, 387, 127–138.
39. Goodridge, L., Goodridge, C., Wu, J., Griffiths, M., Pawliszyn, J. Anal. Chem., 2004, 76, 48–52.
40. Mao, Q., Pawliszyn, J., Thormann, W. Anal. Chem., 2000, 72, 5493–5502.
41. Chen, A. B., Rickel, A., Flanigan, G., Hunt, G., Moorhouse, K. G. J. Chromatogr. A, 1996, 744,
279–284.
43. Wu, J., Huang, T. Electrophoresis, 2006, 27, 3584–3590.
44. Righetti, P. G. J. Chromatogr. A, 2004, 1037, 491–499.
45. Yang, L., Tang, Q., Harrata, A. K., Lee, C. S. Anal. Biochem., 1996, 243, 140–149.
46. Kirby, D. P., Thorne, J. M., Göetzinger, W. K., Karger, B. L. Anal. Chem., 1996, 68, 4451–4458.
47. Storms, H. F., van der Heijden, R., Tjaden, U. R., van der Greef, J. Electrophoresis, 2004, 25,
3461–3467.
48. Simpson, D. C., Smith, R. D. Electrophoresis, 2005, 26, 1291–1305.
49. Wang, Y., Rudnick, P. A., Evans, E. L., Li, J., Zhuang, Z., Devoe, D. L., Lee, C. S., Balgley, B. M.
Anal. Chem., 2005, 77, 6549–6556.
50. Martinovic, S., Pasa-Tolic, L., Smith, R. D. Methods Mol. Biol., 2004, 276, 291–304.
52. Tragas, C., Pawliszyn, J. Electrophoresis, 2000, 21, 227–237.
53. Sheng, L., Pawliszyn, J. Analyst, 2002, 127, 1159–1163.
54. Wu, X.-Z., Huang, T., Liu, Z., Pawliszyn, J. Trends Anal. Chem., 2005, 24, 369–382.
55. Huang, T., Pawliszyn, J. Analyst, 2000, 125, 1231–1233.
56. Liu, Z., Pawliszyn, J. Anal. Biochem., 2005, 336, 94–101.
57. Bo, T., Pawliszyn, J. Anal. Biochem., 2006, 350, 91–98.
58. Bo, T., Pawliszyn, J. Anal. Chim. Acta., 2006, 559, 1–8.
59. Liu, Z., Lemma, T., Pawliszyn, J. J. Proteome Res., 2006, 5, 1246–1251.
Part IIB
Capillary-Based Systems: Specialized
20 Subcellular Analysis by
Bobby G. Poe and Edgar A. Arriaga
CONTENTS
20.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 583

20.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 584
20.2.1 Organelle Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 584
20.2.1.1 Cellular Fractionation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 584
20.2.1.2 Electrophoretic Preparation of Organelles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 585
20.4 Practical Application. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 587
20.4.1 Proteomic Analysis of Dissolved Subcellular Fractions . . . . . . . . . . . . . . . . . . . . . . . . . . . 587
20.4.2 Quantification of Hydroxychloroquine in the Rat Liver Microsomal Fraction . . . 590
20.4.3 Measurement of the pH in Individual Acidic Organelles . . . . . . . . . . . . . . . . . . . . . . . . . . 591
20.4.4 Direct Sampling of Dopamine from Mammalian Cell Cytoplasm . . . . . . . . . . . . . . . . 594
20.4.5 Detection of Individual Mitochondria Sampled from Muscle Tissue Cross
Sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 595
20.5.1 Isolation of Mitochondria from Mammalian Cell Culture . . . . . . . . . . . . . . . . . . . . . . . . . 597
20.5.1.1 Cell Lysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 598
20.5.1.2 Differential Centrifugation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 598
20.5.1.3 Density Gradient Centrifugation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 599
20.5.2 Capillary Modifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 599
20.5.3 Separation Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 600
20.5.4 Organelle Labeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 601
20.5.5.1 UV Absorbance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 602
20.5.5.2 Electrochemical Detection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 602
20.5.5.3 Laser-Induced Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 603
20.5.6 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 604
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 605
20.1 INTRODUCTION
Many cellular processes are localized within specific subcellular compartments; therefore, recent
bioanalytical research has been aimed at utilizing the natural organization found within cells to
reduce the complexity of biological samples. Subcellular compartments include (i) distinct, vesicular
organelles, such as nuclei, mitochondria, and acidic organelles; (ii) indistinct, continuous organelles
583
such as the endoplasmic reticulum, Golgi apparatus, plasma membrane, and cytoskeletal network;
and (iii) the cytoplasm.
In spite of the reduction in complexity that can be achieved by cellular fractionation, an analytical
separation is frequently required to separate one or more components from the cellular milieu. As
evidenced throughout this book, capillary electrophoresis (CE) provides high resolution and separa-
tion efficiency, both of which are necessary for subcellular analysis. In addition, CE is advantageous
because it requires very little sample volume, typically less than a nanoliter, and generally very
little sample preparation. Hence, capillaries have been used to directly sample subcellular compart-
ments within neurons,1 oocytes,2 and muscle tissue sections.3 They have also been used to analyze
individual organelles from a single cell following on-column lysis.4,5
Taking advantage of these attributes, CE has the potential to benefit diverse fields, that study
the vital cellular processes that are localized within specific subcellular compartments. In fact,
subcellular analysis by CE has already proven useful in different areas such as neurochemistry;1,6–14
cellular physiology;2,3,15–23 gene therapy;24 drug accumulation, metabolism and localization;25–32
disease diagnosis;33,34 and proteomics.35–38
20.2 BACKGROUND
20.2.1 ORGANELLE PREPARATION
In general, the aim of subcellular analysis is to quantify an analyte within a specific subcellular
compartment. Consequently, in most cases, an organelle fraction must be purified before analysis.
Cellular fractionation, that is, isolation and purification of organelles, has been indispensable in the
biochemical fields and, as evidenced in the literature, has been used pervasively. Since complete
reviews can be found in the biochemical literature, we will only briefly describe the principles of
cellular fractionation.
20.2.1.1 Cellular Fractionation

To purify an organelle fraction, the first step is cell homogenization, which can be accomplished using
diverse methods such as hypotonic shock or mechanical homogenization. Hard tissue samples such
as muscle require harsh homogenization techniques (e.g., using a Waring blender), but soft samples
such as cells from cultures can be homogenized with gentler techniques like nitrogen cavitation.
The choice of a homogenization technique must take into account the sample hardness, volume, and
desired organelle intactness.
Following cell homogenization, centrifugation techniques enjoy nearly universal use in the purifi-
cation of subcellular organelles. Centrifugation is relatively simple and can be performed on several
liters to submilliliter volumes of cell homogenate. As shown in Equation 20.1, the sedimentation
velocity dx/dt per unit centrifugal field (ω2 x) is dependent on the size of the particle (r 2 ), the density
difference of the particle compared to the medium (ρp − ρm ), and the viscosity of the medium (η):

dx/dt 2r 2 ρp − ρm
= (20.1)
ω2 x 9η
With advances in separation media, organelles can now be separated based on their density, size,
or both. This makes centrifugation the most versatile and practical cellular fractionation method.
Other techniques can be combined with centrifugation to achieve higher purity. These tech-
niques include immunoisolation and electrophoretic purification. Immunoisolation holds great
potential for isolating highly purified organelle fractions, since it relies on the molecular recog-
nition of surface antigens by antibodies. Prerequisites for immunoisolation are (i) an antigen that
is highly abundant on the surface of a specific organelle type; (ii) an antibody that recognizes the
Subcellular Analysis by Capillary Electrophoresis 585
antigen; and (iii) a monodisperse organelle suspension, which reduces contamination due to aggre-
gation. Frequently, the antibody is attached to magnetic beads to allow more efficient recovery of
the organelles.
20.2.1.2 Electrophoretic Preparation of Organelles

Although electrophoretic separations are chiefly preparatory, they provide much of the basis for
the separation of intact organelles by CE. Electrophoretic separations were originally developed for
proteins, but have been used since the 1970s as a purification technique for intact organelles.39–43 Bio-
logical particles, including organelles, carry a net negative charge at physiological pH, and therefore,
are mobilized in the presence of an electric field (see Section 20.3). Common separation modes used
in electrophoretic preparations are zone electrophoresis, isotachophoresis, and isoelectric focusing
modes. The separation scheme for free-flow electrophoresis (zone electrophoresis mode) is illus-
trated in Figure 20.1. A laminar fluid flow carries the organelles through the separation chamber,
while the electric field mobilizes the organelles perpendicular to the flow. Organelle types are then
separated based on their electrophoretic mobility differences. At the end of the chamber the sus-
pension is separated into many fractions and the material can be recovered for further analysis. As
such, organelles can be continually introduced into the chamber, allowing for the purification of
large quantities of material.
The earliest electrophoretic techniques concentrated mainly on the preparation of
mitochondria39,42,43 and lysosomes,40,41 but further research has made possible the purification
of secretory vesicles;44 clathrin-coated vesicles;45 endoplasmic reticulum;46 early, middle-, and
late-stage endosomes;46 peroxisomes;47 microsomes;48 and phagosomes.49 These preparatory tech-
niques have functioned as the proof-of-principle for analytical separations of intact organelles using
CE, by demonstrating that isolated organelles are amenable to electrophoretic separation techniques.
+ –
Electric field
Sample inlet
Laminar fluid flow
FIGURE 20.1 Free-flow electrophoresis separation. Laminar fluid flow carries the organelles perpendicularly
to the applied electric field. Organelles migrate in the opposite direction of the electric field according to their
electrophoretic mobility. Sample is introduced continually at the inlet and collected in many fractions following
separation.

The theories describing micellar electrokinetic chromatography (MEKC), capillary zone elec-
trophoresis (CZE), and capillary gel electrophoresis (CGE) separations of small molecules and
biopolymers are described in other chapters of this book and will not be discussed here. Here, we
will briefly touch upon the theoretical aspects of the electrophoretic mobility of organelles, foregoing
an in depth discussion of electrokinetic theory that can be found elsewhere in original publications
and comprehensive reviews.50–58 The electrophoretic mobility (µE ) is defined as
ν
µE = (20.2)
E
where ν is the velocity of the particle and E is the electric field in V/cm. The electrophoretic mobility
of small molecules is as given in the following equation:
q
µE = (20.3)
6π ηr
where q is the net charge, η is the viscosity of the separation medium, and r is the radius of
the molecule. As shown, electrophoretic mobility depends on the electrical force and Stokes
frictional drag.
On the other hand, the electrophoretic mobility of charged particles (e.g., organelles) is more
complex. Figure 20.2 illustrates the forces acting upon a charged particle in an electric field (E). While
electrophoretic mobility is still governed in part by the electrical force (F1 ) and the Stokes frictional
force (F2 ), it is also affected by electrophoretic retardation (F3 ) and the relaxation effect (F4 ). The
electrophoretic retardation force and the relaxation effect are both caused by the ionic atmosphere
(dashed circle) that surrounds the charged particle and are dependent on the zeta-potential (ζ ), which
is related to the surface charge. Electrophoretic retardation (F3 ) is caused by the force exerted upon
the solvent molecules by the oppositely charged ions in the ionic atmosphere, resulting in fluid flow
around the particle, which increases frictional drag. In addition, the ionic atmosphere lags slightly
behind the charged particle, since its formation requires a finite amount of time, which results in
a small electrical force in the opposite direction of the electric field; this is termed the relaxation
effect (F4 ).
The first solution for the electrophoretic mobility of a charged particle was reported by Von
Smoluchowski59 for the case of a thin double layer (i.e., κa 1); where κ is the Debye factor and
a is the particle radius. Accordingly, the electrophoretic mobility of a charged particle should follow
εζ
µE = (20.4)
η
F3
F1
F4
F2
FIGURE 20.2 Forces acting on a charged particle. The particle is negatively charged and surrounded by a
positively charged ionic atmosphere, indicated by the dashed circle. F1 is the electrical force, F2 is Stokes
frictional drag, F3 is electrophoretic retardation, and F4 is the relaxation effect.
in which ε and η are the dielectric constant and viscosity of the liquid surrounding the particle,
respectively, and ζ is defined as follows:
σa
ζ = (20.5)
ε(1 + κa)
where σ is the net electrokinetic charge density.

In the opposite case, κa 1, the electrophoretic mobility follows the Hückel equation:60
2εζ
µE = (20.6)
3η
Henry provided a solution61 (Equation 20.7) for various values of κa by taking into account the
deformation of the electric field lines near the particle:
2εζ
µE = f1 (κa) (20.7)
3η
When κa → 0, f1 = 1 and the Henry equation reduces to Hückel’s solution (Equation 20.6).
Conversely, when κa → ∞, f1 = 3/2 and the equation reduces to Smoluchowski’s solution (Equa-
tion 20.4). The function f1 was recently given as a single equation,62 as opposed to two power series
as reported by Henry.
Equation 20.7 can be applied when any of the following conditions are met: (i) κa 1, (ii)
κa 1, or (iii) ζ 25 mV. In these cases, the relaxation effect is negligible. More accurate theories
that take into account the relaxation effect and electrophoretic retardation have been developed by
Booth,52 Overbeek,53,54 and O’Brien.55–58
Regrettably, the electrokinetic theories described above do not adequately describe the elec-
trophoretic mobility of organelles. This is because the classical colloidal theories assumed the particle
was rigid and nonconducting with uniform charge, whereas in reality, organelles are nonspherical,
deformable, and the surface charges are mobile, which leads to polarization. This makes it difficult
to interpret electrophoretic mobility with the classical theories. However, Hayes’ group has aimed to
describe the electrophoretic mobility of deformable particles using liposomes as a model.63,64 It has
been demonstrated that models that incorporate the effects of particle deformation, mobile surface
charges, and polarizability predict the electrophoretic behavior of liposomes better than do traditional
electrokinetic models.64 Since liposomes more closely resemble biologically relevant particles, this
model provides a more complete description of organelle electrophoretic mobility.
20.4 PRACTICAL APPLICATION

Subcellular analysis by CE can be performed in three modes: (i) an organelle fraction can be dissolved
or lysed and the analytes found in the fraction analyzed; (ii) intact, isolated organelles can be separated
and detected; and (iii) analytes or organelles can be directly sampled from a single cell or tissue
section. Each mode of analysis will be illustrated below.
20.4.1 PROTEOMIC ANALYSIS OF DISSOLVED SUBCELLULAR FRACTIONS

Proteomics research has benefited greatly from subcellular fractionation, because reducing the com-
plexity of the entire proteome to smaller organelle proteomes makes it possible to separate and
detect low abundance proteins. Furthermore, since the goal of proteomics is not only to learn protein
sequences but also the localization and function of proteins, subcellular analysis is advantageous
because it provides the subcellular localization. Indeed, the benefits of cellular fractionation before
proteomic analysis have been proven with conventional two-dimensional gel electrophoresis and
multidimensional liquid chromatography-tandem mass spectrometry (LC-MS/MS).65–68
Although two-dimensional gel electrophoresis provides unmatched resolution (∼2000 protein
spots using conventional gels), it does have some drawbacks: these include the precipitation of
highly acidic or basic proteins in the first dimension (isoelectric focusing), limited sensitivity and
dynamic range, low throughput, and difficulty in automating time-consuming and tedious tasks. CE,
on the other hand, is easily automated, has a very low limit of detection (LOD) when combined with
off-column laser-induced fluorescence (LIF) detection, and has a large dynamic range. In addition,
isoelectric focusing does not need to be performed in the first dimension, thereby eliminating the
precipitation of acidic and basic proteins. Separations based on CE thus show potential for increasing
the throughput and sensitivity of proteomic analyses.
Working toward this goal, Dovichi and coworkers69–71 have developed highly sensitive two-
dimensional CE protein separations and automated the analysis. Since two-dimensional CE
separations do not resolve as many proteins as traditional two-dimensional gel electrophoresis,70
in order to increase the resolution and aid in the detection of low abundance proteins, a differential
detergent fractionation technique was used to reduce the complexity of cell homogenates before
separation.38,72 Differential detergent fractionation was developed by Ramsby et al.73 for isolated
hepatocyes and is based on the different solubilities of proteins in various buffers. A schematic rep-
resentation of the fractionation procedure is shown in Figure 20.3 where four fractions are obtained:
(i) cytosolic or soluble proteins, (ii) membrane-organelle proteins, (iii) nuclear membrane proteins,
and (iv) cytoskeletal-matrix proteins. The identity of each fraction was confirmed and its selectivity
measured with immunoblotting using over 20 antibodies.73
AtT-20 cells
(3 × 106cells)
PBS wash
(repeat four times)
Cytosolic Digitonin/EDTA
proteins extraction
Membrane/organelle Triton/EDTA
proteins extraction
Nuclear Tween/DOC
proteins extraction
Cytoskeletal-matrix SDS
proteins extraction
FIGURE 20.3 Schematic representation of the differential detergent fractionation procedure applied to
cultured cells. Four fractions are obtained: (1) cytosolic proteins, (2) membrane/organelle proteins, (3) nuclear-
membrane proteins, and (4) cytoskeletal proteins. (Modified from Fazal, M. A., et al., J. Chromatogr. A, 1130,
182–189, 2006. )
Homogenate
0
CS
100
E
fra
45 60
cti
200 0 15 30
on
tim e (s)
MECC migration
Homogenate × 10
0
CS
100
E
fra
45 60
30
cti
200 0 15
on
tio n tim e (s)

MECC migra
FIGURE 20.4 Two-dimensional electropherogram of whole cell homogenate. AtT-20, mouse adrenal gland
cells, were homogenized and the proteins were labeled with FQ. The electropherogram is shown as a land-
scape image, in which the fluorescence intensity is proportional to the peak height. Proteins were separated
first by capillary sieving electrophoresis and then by micellar electrokinetic capillary chromatography in the
second dimension. Approximately 150 fractions were transferred from the first capillary to the second capillary.
(Reprinted from Fazal, M. A., et al., J. Chromatogr. A, 1130, 182–189, 2006. Copyright 2006. With permission
from Elsevier.)
Ramsby’s procedure was used by Fazal et al.38 to fractionate the homogenate of AtT-20 cells
(mouse adrenal gland cells) before two-dimensional CE separation. Proteins were labeled with 3-
(2-furoyl) quinoline-1-carboxaldehyde (FQ) before separation and detected using LIF in a sheath
flow cuvette. Capillary sieving electrophoresis was performed in the first dimension to separate the
proteins based on size, using dextran in the sieving buffer; this was followed by micellar electrokinetic
capillary chromatography in the second dimension to separate proteins based on hydrophobicity.
The resulting two-dimensional electropherogram for the whole cell homogenate is displayed in
Figure 20.4. The separation detected 51 unique protein peaks, and the origin of the large mesa,
shown in the close up view, was determined using the differential detergent fractionation procedure.
Figure 20.5 shows all four detergent fractions and indicates that the large mesa in Figure 20.4 was
from proteins located in the nuclear protein extract. As anticipated, by performing the fractionation
technique before the two-dimensional separation, the number of resolved proteins increased from 51,
for the whole cell homogenate, to 231, for the detergent fractions. This increase is due to the detection
of low abundance proteins that had been previously masked by highly abundant proteins in the whole
(a) Digitonin fraction (b) Triton fraction
0 0
C
C
SE
SE
100 100
fra
fra
60 45 60
45
ct
200 0 15 30
ct
200 0 15 30
io
(s)
MECC migration time
io
(s)
MECC migration time
n
n
(c) Tween fraction (d) Detergent resistant fraction
0 0
C
C
SE
SE
100 100
fra
fra
60 60
ct
ct
200 0 15 30 45 200 0 15 30 45
io
io
(s)
MECC migration time
n
(s)
n
MECC migration time
FIGURE 20.5 Two-dimensional electropherograms of each differential detergent fraction. The digitonin frac-
tion (A) is the cytosolic proteins, the Triton fraction (B) is the membrane/organelle proteins, the Tween fraction
(C) is the nuclear-membrane proteins, and the detergent resistant fraction (D) is the cytoskeletal proteins.
(Reprinted from Fazal, M. A., et al., J. Chromatogr. A, 1130, 182–189, 2006. Copyright 2006. With permission
from Elsevier.)
cell homogenate. In addition, each detergent fraction had distinct electrophoretic features, indicating
that each fraction contained unique proteins.
20.4.2 QUANTIFICATION OF HYDROXYCHLOROQUINE IN THE RAT LIVER

MICROSOMAL FRACTION
A more conventional subcellular fractionation approach was taken for the quantification of hydrox-
ychloroquine (HCQ) in the microsomal fraction of mouse liver homogenates.27 Microsomes are
vesicles formed from the disruption of the endoplasmic reticulum and provide a convenient means
of studying metabolism by the cytochrome CYP450 enzyme superfamily, which metabolizes the
majority of drugs given to humans. HCQ has proven efficacious against rheumatoid arthritis, lupus
erythematosus, and has been used as an antimalarial drug. It is a chiral molecule and, as shown in
Figure 20.6, forms three chiral metabolites when incubated with rat liver microsomes. Since it is
well known that enantiomers show different activities toward drug targets, a method for quantifying
HCQ and its metabolites was developed.
Mouse livers were excised and homogenized with a Potter–Elvehjem homogenizer. The micro-
somal fraction was isolated by centrifuging the homogenate at 9000 × g for 15 min, to sediment
nuclei, mitochondria, and membrane debris. The supernatant was centrifuged at 100, 000 × g for
60 min to pellet microsomes. The microsomes were further purified by another centrifugation step
to remove soluble proteins (e.g., hemoglobin). HCQ was incubated with the microsomes for 2 h.
Then a liquid–liquid extraction technique was performed to recover HCQ and its metabolites. The
molecules were separated by a commercial CE system with ultraviolet (UV) absorbance detection.
Cl
Desethylhydroxychloroquine
HN
(DHCQ)
*
Hydroxychloroquine (HCQ) H3C
NH
Cl Cl OH
N N
HN HN Desethylchloroquine (DCQ)
* *
H3C CH3
H3C
N
NH
Cl CH3
HO
N
Bidesethylchloroquine
HN (BDCQ)
*
H3C
NH2
FIGURE 20.6 Illustration of HCQ and its three metabolites. The chiral center in each molecule is indicated
by an asterisk.
The resulting electropherogram is displayed in Figure 20.7, which shows the separation of all
eight molecules. The separation was performed in Tris buffer that contained sulfated-β-cyclodextrin
and hydroxypropyl-β-cyclodextrin, which provided the enantiomeric selectivity. Metabolites were
quantified from 0.41 to 0.89 µM with relative standard deviations of ∼10%. The subcellular analysis
shows the preferential production of (−)-(R)-metabolites (Table 20.1). Compared with a previously
reported high-performance liquid chromatography (HPLC) method,74 the CE method allowed these
researchers to quantify two more metabolites and required less time.
20.4.3 MEASUREMENT OF THE pH IN INDIVIDUAL ACIDIC ORGANELLES

An alternative approach to separating analytes that have been released from dissolved organelles
is to quantify the analytes within intact organelles. The major aims of this type of analysis are
to reveal the heterogeneity amongst organelles, determine the subcellular localization of analytes,
and quantify the analytes within individual organelles. Arriaga and coworkers successfully analyzed
intact nuclei,4,75,76 mitochondria,3,15,17,77–80 and acidic organelles16,28,81 by CE-LIF. The advantages
of individual organelle analysis are illustrated below using pH measurements in individual acidic
organelles as an example.
Multidrug resistance is a common problem with chemotherapeutic treatments in which cancer
cells become insensitive to drug treatment. It has been proposed that multidrug resistance may
be related to a larger than normal pH gradient between the cytoplasm and acidic organelles in
2
20
Absorbance (220 nm)

1
15
10
5
3 45 7
6 8
0
5 10
15 min
FIGURE 20.7 Chiral resolution of HCQ and its metabolites after incubation with mouse liver microsomes.
Peaks are numbered and correspond as follows: (1) (−)-(R)-HCQ, (2) (+)-(S)-HCQ, (3) (−)-(R)-DHCQ,
(4) (+)-(S)-DHCQ, (5) (+)-(S)-DCQ, (6) (−)-(R)-BDCQ, (7) (−)-(R)-DCQ, and (8) (+)-(S)-BDCQ. Elec-
trophoresis was performed in a 50 cm, 50 µm i.d. capillary with a separation voltage of 13 kV in 100 mM
Tris/phosphate, 1% sulfated-β-cyclodextrin, 30 mg/mL hydroxypropyl cyclodextrin, pH 9.0. The molecules
were detected with UV absorbance at 220 nm. (Reprinted from Cardoso, C. D., et al., Electrophoresis, 27,
TABLE 20.1
HCQ Metabolites Formed by Incubation with Mouse Liver Microsomes
(−)-(R)- (+)-(S)- (−)-(R)- (−)-(R)- (−)-(R)- (+)-(S)-
DHCQ DHCQ DCQ DCQ BDCQ BDCQ
Concentration (M) 8.9 × 10–7 6.5 × 10–7 7.8 × 10–7 6.3 × 10–7 4.9 × 10–7 4.1 × 10–7
RSD(%)n = 3 7.6 8.9 10.1 12.9 10.5 11
Source: Reprinted from Cardoso, C. D., et al. Electrophoresis, 27, 1248, 2006. With permission.
drug-resistant cells as compared to drug-sensitive cells.82 The larger pH gradient is hypothesized

to result in drug protonation and sequestration in acidic organelles and secretion from the cell
(Protonation, Sequestration, and Secretion, PSS model). When the drug is secreted from the cell, it
can no longer act to halt cell proliferation, which causes drug resistance. The average pH value of
acidic organelles can be determined with fluorescence microscopy and flow cytometry, but neither
technique has been used to measure the pH of individual acidic organelles or reveal acidic organelle
heterogeneity. To obtain this information, CE-LIF was used to measure, for the first time, the pH of
individual acidic organelles in drug-resistant and drug-sensitive cell lines.81
Fluorescein tetramethylrhodamine dextran (FRD) was used as a ratiometric probe, which
provides a pH-independent signal (tetramethylrhodamine), to compensate for different organelle vol-
umes, and a pH-dependent signal (fluorescein). The ratio of the pH-independent and pH-dependent
fluorescence was expected to provide a quantitative measure of the pH. Acidic organelles (lysosomes
and endosomes) were specifically labeled by FRD, because the cells endocytose small amounts of
the extracellular medium, which is accumulated in endosomes and eventually lysosomes. CE was
performed in a poly(N-acryloyl aminopropanol, AAP) modified capillary to reduce both adsorption
at the capillary surface and the electroosmotic flow. LIF detection was performed off-column using a
0.4
Fluorescence intensity ratio (R)

0.35
0.3
0.25
0.2
0.15
2 4 6 8
pH
FIGURE 20.8 Calibration of the pH measurement. R, the ratio of tetramethylrhodamine and fluorescein
fluorescence, versus the pH for liposomes constructed with internal pH values ranging from 3 to 7. The average
ratio and standard deviation is shown with a marker and bars at 7 pH values. The calibration curve represents
a quadratic model and was determined by multiple regression analysis. (Reprinted from Chen, Y. and Arriaga,
E. A., Anal. Chem., 78, 821, 2006. Copyright 2006. With permission from American Chemical Society.)
14 (a) 14 (b)
12 12
10 10
8 8
6 6
4 4
2 2
0 200 400 600 800 1000 600 620 640 660 680 700
Migration time (s) Migration time (s)
FIGURE 20.9 Electropherogram displaying CE-LIF detection of individual acidic organelles. The entire
electropherogram is shown in panel a, where the top trace is the fluorescein channel and the bottom trace is the
tetramethylrhodamine channel. Panel b shows an enlarged region of the electropherogram to illustrate detection
of acidic organelles simultaneously in each channel. Acidic organelles were injected for 1 s at 11 kPa and
separated at –300 V/cm in a poly(AAP) coated capillary. (Reprinted from Chen, Y. and Arriaga, E. A., Anal.
Chem., 78, 821, 2006. Copyright 2006. With permission American Chemical Society.)
single excitation source (488 nm argon-ion laser) and two emission wavelengths (λ1 -pH independent
and λ2 -pH dependent) that were monitored simultaneously.
To calibrate the CE-LIF detector, liposomes were constructed with internal pH values ranging
from 3.1 to 8.0. The liposomes were injected onto the capillary, electrophoretically separated and
detected individually following hydrodynamic focusing in a sheath flow cuvette. The calibration
curve obtained is shown in Figure 20.8. To test the PSS model, acidic organelles were isolated from
drug-resistant and drug-sensitive cell lines using a Dounce homogenizer and differential centrifuga-
tion. The acidic organelles were then analyzed in a fashion similar to the liposomes and the resulting
electropherogram is shown in Figure 20.9; the sharp spikes correspond to the detection of individual
acidic organelles. By tabulating the fluorescence in both channels and comparing the ratios with
3.0 CEM/C2
Corrected amount of FRD/acidic organelle

2.5
2.0
1.5
1.0
CCRF/CEM
0.5
0.0
1 2 3 4 5 6 7 8
pH
FIGURE 20.10 Amount of FRD per acidic organelle sorted from high to low pH. The y-axis is offset 1 unit
for the CEM/c2 cells (drug-resistant). (Reprinted from Chen, Y. and Arriaga, E. A., Anal. Chem., 78, 821.
Copyright 2006. With permission from American Chemical Society.)
the calibration curve, the pH for each individual acidic organelle was measured. The average pH
values were 5.1 ± 0.2 for drug-resistant cells and 6.1 ± 0.9 for drug-sensitive cells, respectively. The
average pH values correlated well with fluorescence microscopy values (5.0 ± 0.6 and 6.2 ± 0.7 for
drug resistant and drug sensitive, respectively).
Unlike fluorescence microscopy, which measures the average pH value of all the acidic organelles
in an entire cell, the individual pH measurements obtained with CE revealed a range of pH values
from 3 to 7 in both cell lines, which indicated that several types of acidic organelles were being
analyzed; specifically, early endosomes (pH 6–6.5), late endosomes (pH 5.5–6), and lysosomes (pH
4.5–5.5). This was confirmed by plotting the amount of FRD versus pH as shown in Figure 20.10.
The more acidic organelles contained more FRD, which is consistent with the maturation of the
acidic organelles as they transform from early endosomes to lysosomes, since they fuse with each
other to become larger during this process.
20.4.4 DIRECT SAMPLING OF DOPAMINE FROM MAMMALIAN CELL CYTOPLASM

The third approach to subcellular analysis involves using the separation capillary to sample sub-
cellular compartments within single cells or from tissue sections. Capillaries are uniquely suited to
provide not only high resolution and separation efficiency but also, due to their physical dimensions,
sufficient spatial resolution to sample subcellular compartments. Most commonly, the capillary is
used to sample the cytoplasm from single cells. However, the capillary can also be used to sample
intact organelles from tissue sections3,15 and analyze organelles released from single cells.5,76
In particular, neurochemistry has found CE especially useful for quantifying neurotransmitter
concentrations in single cells. Since important cellular processes such as neurotransmitter syn-
thesis, storage and release are affected by the cytoplasmic concentrations of neurotransmitters,
methods to sample and quantify cytoplasmic concentrations of neurotransmitters from single inver-
tebrate cells have been developed.6,7,9–11 Invertebrate neurons are particularly amenable for sampling
because they can be quite large, for example, the giant dopamine neuron of pond snails is ∼200 µm
1 2
3
4
FIGURE 20.11 Schematic representation of the end-column amperometric detection system: (1) the outlet
of the separation capillary was etched to 13 µm i.d., (2) Ag/AgCl reference electrode, (3) flame-etched carbon
fiber electrode, (4) buffer reservoir, and (5) separation capillary, 770 nm i.d. (Reprinted from Woods, L. A.,
et al., Electroanalysis, 17, 1192, 2005. With permission.)
in diameter. However, sampling from mammalian cells, which are over four times smaller, was
impracticable until recently.
To overcome these difficulties, a CE system utilizing a 770 nm i.d. capillary was developed
to sample and quantify dopamine from the cytoplasm of single mammalian cells.7 To sample the
cytoplasm, the inlet of the separation capillary was tapered to an outer diameter of 2.5 µm, which
is five times smaller than the diameter of the cells. The capillary was inserted through the plasma
membrane by applying a small voltage (e.g., 2 kV) to cause the electroporation of the membrane.
This voltage also electrokinetically injected the cytoplasm into the capillary, simultaneously. This
system allowed as little as 380 fL of cytoplasm to be injected from a single PC12 cell, which amounts
to ∼40% of the cell cytoplasm. The smallest percentage of cytoplasm that was sampled was 8%.
Such small injection volumes necessitate a very low LOD. This is achieved using an improved
end-column amperometric detection scheme,83 as shown in Figure 20.11; a 2.5 µm flame-etched
carbon fiber microelectrode is positioned inside the etched lumen of the separation capillary. The
conical tip of the flame-etched microelectrode allows for better positioning of the electrode into
the capillary lumen, which reduces the dead volume and hence decreases the LOD. The reported
LODs for dopamine and catechol using this detection scheme were 400 ± 100 and 410 ± 80 zmol
(10–21 moles), respectively. The separation and detection of dopamine sampled from a single cell is
displayed in the resulting electropherogram shown in Figure 20.12. The concentration of cytoplasmic
dopamine from three cells was 240 ± 60 µM, which indicated the presence of dense core vesicles
in the injection plug.
20.4.5 DETECTION OF INDIVIDUAL MITOCHONDRIA SAMPLED FROM MUSCLE

TISSUE CROSS SECTIONS
In a similar fashion, intact organelles can also be sampled directly from a cell sample. This was first
demonstrated by sampling and separating individual mitochondria from skeletal muscle tissue cross
sections.3 Skeletal muscle is a complex and heterogeneous tissue in which thousands of individ-
ual fibers extend approximately in parallel throughout the length of the tissue. Each fiber contains
multiple nuclei and thousands of mitochondria. Using conventional histochemical stains, the het-
erogeneity among muscle fibers can be visualized to investigate the effects of disease and aging.
However, an analytical method that could reveal heterogeneity among the organelles in a single fiber
as well as fiber-to-fiber heterogeneity was desirable.
(a) 200
150
Current (pA)
DA
100
50
0
5 10 15 20
Time (min)
(b)
200
CAT
DA
Current (pA)
150
100
50
0
5 10 15 20
Time (min)
FIGURE 20.12 Separation of dopamine (DA) and catechol (CAT) from a single mammalian cell. (a) Dopamine
was separated and detected in a 380 fL injection of PC12 cell cytoplasm and (b) a standard injection of DA
and CAT is used to verify the detection of dopamine. Both injections used a 3 kV injection voltage for 5 s and
a bare fused-silica capillary. Dopamine was separated at 25 kV in 50 mM TES, pH 7.2 with 2% 1-propanol.
(Reprinted from Woods, L. A., et al., Electroanalysis, 17, 1192, 2005. With permission.)
Semimembranous muscle tissues from Fisher rats were excised, frozen, and sectioned perpendic-
ular to the fiber length, with a cryostat. The ∼10 µm thick cross sections were mounted on microscope
slides and frozen before use. Mitochondria were labeled with the mitochondria-specific fluorescent
stain 10-N-nonyl acridine orange (NAO) by depositing a microdroplet on the tissue cross section and
using a capillary, micromanipulator, and a microscope to visualize deposition. Mitochondria were
sampled by positioning the separation capillary on top of the labeled fiber and applying suction. Since
the cross-sectional area of the 50 µm i.d. capillary was ∼2000 µm2 , which is slightly smaller than
the ∼3000 µm2 area of a single fiber, the capillary was able to sample mitochondria from a single
fiber. The labeling and sampling of mitochondria are shown in Figure 20.13, wherein the capillary
lumen can be easily seen above a single muscle fiber. After the mitochondria were introduced into the
capillary, the inlet was placed in separation buffer and the separation voltage applied. Mitochondria
were electrophoretically separated and detected individually by off-column LIF detection as they
migrated out of the capillary. Figure 20.14 displays an electropherogram showing the detection of
individual mitochondria sampled from a muscle cross section.
One complicating factor in the analysis of mitochondria from tissue cross sections is that mito-
chondria are found in two locations in muscle tissue: (i) subsarcolemmal mitochondria are found
directly beneath the sarcolemmal membrane and are easily extracted and (ii) interfibrillar mitochon-
dria are located between the myofibrils and require exposure to a protease to be efficiently released.
To release these latter mitochondria, a short treatment with trypsin was tested by depositing a micro-
droplet of trypsin on the fiber before sampling. Table 20.2 compares the analysis of mitochondria
from muscle fiber that had undergone trypsin treatment with the analysis of mitochondria from
(A) (B) (C)
50 µm
FIGURE 20.13 Sampling mitochondria from a muscle cross section. A series of fluorescence images show
(A) the tissue cross section labeled with a microdroplet of 5 µM NAO using a capillary and a micromanipulator,
(B) the separation capillary is positioned above the desired fiber and the mitochondria are sampled for 3 s with
negative pressure, and (C) the fluorescence image indicates that the sampling was successful since some of the
fluorescence in the region above the desired fiber has disappeared. (Reprinted from Ahmadzadeh, H., et al.,
Anal. Chem., 76, 315, 2004. Copyright 2006. With permission American Chemical Society.)
10
8
Signal intesity (V)
4
1
2 2
3
0
0 400 800 1200
Migration time (s)
FIGURE 20.14 Detection of individual mitochondria sampled from a muscle tissue cross section. CE-LIF
was performed on (1) cross sections that were labeled with 5 µM NAO, (2) an unlabeled cross section, and (3)
a solution of the free dye, NAO. The three traces show that mitochondria are detected when labeled with NAO
and there is relatively little interference from other cellular components. The small spikes seen in (2) are likely
caused by scattering of the excitation source or autofluorescence but only comprise 5% of the peaks detected
in (1). Mitochondria were separated at –200 V/cm in a poly(AAP) coated capillary in 250 mM sucrose, 10 mM
HEPES, pH 7.4. (Reprinted from Ahmadzadeh, H. et al., Anal. Chem., 76, 315, 2004. Copyright 2006. With
permission American Chemical Society.)
untreated muscle fiber. As can be seen, many more mitochondria, presumably interfibrillar, were
released from the fiber after exposure to trypsin.

20.5.1 ISOLATION OF MITOCHONDRIA FROM MAMMALIAN CELL CULTURE
Many complex cellular processes are best studied through the use of an appropriate cell model.
Cell cultures provide a relatively stable and virtually unlimited supply of organelles for subcellular
analysis. We will discuss organelle preparations that are commonly used in the author’s laboratory
TABLE 20.2
Effect of Trypsin on the Number of Mitochondrial
Peaks
Number of Peaksa
Cross Section Number Trypsin Treatment No Trypsin Treatment
1 1782 350
2 2339 256
3 1285 339
Mean 1802 315
RSD (%) 29 16
a Conditions for sampling and CE-LIF analysis are the same as for
Figure 20.14.
Source: Reprinted from Ahmadzadeh, H., et al., Anal. Chem., 76, 315, 2004.
Copyright 2006. With permission American Chemical Society.
for the isolation of mitochondria from human osteosarcoma 143B cells. With minor changes, the
procedure detailed below should be applicable to other mammalian cell lines.
20.5.1.1 Cell Lysis

Approximately 5 × 106 143B cells should be harvested and washed three times with an ice-cold, iso-
osmotic buffer, for example, 220 mM mannitol, 70 mM sucrose, 0.5 mM EGTA, 2 mM HEPES, pH
7.4 buffer (MSHE buffer). Whenever possible, the sample should be kept on ice and centrifugation
performed at 4◦ C to prevent organelle degradation. Cells are washed by centrifuging at ∼700 × g
for 10 min, followed by removal of the supernatant and resuspension in MSHE buffer. Approxi-
mately 3 mL of cells are then transferred into an ice-cold cell disruption chamber (Parr Instruments,
Moline, IL), which is then pressurized to 500 psi. After equilibration for 15 min, the pressure is
released as the cells are collected through the discharge port and kept on ice for 5 min. Disrup-
tion efficiency is generally >90% and the nitrogen pressure can be adjusted to obtain the desired
efficiency. A phase-contrast microscope or trypan blue staining can be used to visually confirm
cell disruption.
20.5.1.2 Differential Centrifugation

Following lysis, centrifugation is nearly universally implemented to separate organelle fractions
based on their size or density. Centrifugation is popular because it is very versatile; it can be used
to fractionate organelles from several liters of tissue homogenates or submilliliter volumes from
cell cultures. The simplest centrifugation procedure is differential centrifugation, which produces
crudely purified organelle fractions.
Differential centrifugation begins by pelleting nuclei at low centrifugal force (i.e., 600–1000 × g
for 10 min). The resulting nuclear fraction is always contaminated with organelles that have been
trapped in the cytoskeletal network, whole cells, and large membrane debris. The postnuclear super-
natant is then transferred to a new microcentrifuge tube and the light-organelle fraction can be
sedimented at 16,000 × g for 10 min, and the supernatant will contain the cytosolic components.
The light-organelle fraction is now enriched with mitochondria, but also contains significant amounts
of acidic organelles, microsomes, and peroxisomes. For some analyses that use organelle-specific
labels, no further purification is necessary. However, if purer organelle fractions are required, the
light-organelle fraction can be further fractionated as described below.
20.5.1.3 Density Gradient Centrifugation

The mitochondria in the light-organelle fraction can be further purified using common methods
such as continuous gradient and discontinuous gradient centrifugation. While continuous gradients
can be used (e.g., sucrose gradient), since the osmolarity changes quite drastically through the
gradient, discontinuous gradients are preferred to reduce mitochondrial disruption. Furthermore,
the mitochondrial band is harder to distinguish when continuous gradients are used, which reduces
the recovery.
Discontinuous density gradient centrifugation separates organelles based on their density and has
been described by Storrie and Madden.84 Unfortunately, the popular medium for density gradient
centrifugation, Metrizamide, is no longer produced, and therefore, the existing methods must be
redesigned with appropriate new media. We have found success using Histodenz (Sigma, St Louis,
MO) following the procedure of Okado-Matsumoto and Fridovich.85 Briefly, a 50% (w/v) solution
of Histodenz is prepared by dissolving in 5 mM Tris–HCl, 1 mM EGTA at pH 7.4. Solutions of
34%, 30%, 25%, 23%, and 20% solutions (w/v) of Histodenz are prepared by diluting the stock
solution in 250 mM sucrose, 5 mM Tris−HCl, 1 mM EGTA at pH 7.4. The light-organelle fraction
is suspended in ∼10 mL of 25% Histodenz and the discontinuous gradient is prepared as shown in
Figure 20.15. The samples are then centrifuged at 52,000 × g for 90 min at 4◦ C. After centrifugation
is complete, the bands can be removed sequentially including the band with the purified mitochondria
located at the 25/30% interface. Finally, the purity and yield of the mitochondrial preparation can be
determined using enzymatic marker assays, such as cytochrome c oxidase (mitochondrial marker)
and β-galactosidase (lysosomal marker).
20.5.2 CAPILLARY MODIFICATIONS

Mitochondria interact strongly with the surface of bare fused-silica capillaries, preventing any accu-
rate measurement of electrophoretic mobility. Since mitochondria and the capillary surface both
carry a net negative charge, the adsorption is probably not due to any electrostatic interactions,
but rather to hydrophobic intermolecular interactions. To prevent the mitochondria from interact-
ing with the capillary surface, coatings can be applied to the bare silica. Among these coatings,
AAP has been reported to be more hydrophilic and hydrolytically stable in comparison with acry-
lamide, dimethyl acrylamide, and N-acryloylaminoethoxyethanol.86 We have found that capillaries
20%
20/23%
23%
23/25%
Light-organelle 25%
fraction Purified
25/30%
mitochondria
30%
30/34%
34%
FIGURE 20.15 Density gradient centrifugation. The postnuclear supernatant, suspended in 25% Histodenz
(Hz), is layered on top of the 30% and 34% Hz solutions. Then the tube is filled sequentially with the 23% and
20% solutions of Hz. Centrifugation is performed at 52,000×g for 90 min at 4◦ C. The mitochondria will form
a band at the 25/30% Hz interface.
coated with poly(AAP), based on Gelfi’s procedure,87 reduce organelle adsorption, and thereby allow
reproducible separations; this coating remains stable for several months.
Covalent surface modification with poly(AAP) consist of the following steps:
1. Flush 4–5 m of fused-silica capillary sequentially with 0.1 M KOH for 3 h, 0.1 M HCl for
1 h, MeOH for 1 h with nitrogen pressure (90 psi for all flushes unless stated otherwise)
2. Heat the capillary to 110◦ C and flush with nitrogen at 10 psi for 8 h
3. Place the capillary in a moisture-free, inert environment (glove box) and do not remove
until step (9)
4. Flush the capillary with thionyl chloride for 1 h at 65◦ C
5. Seal the capillary ends with a gas chromatography (GC) septum and heat for 8 h at 65◦ C
6. Flush the capillary with freshly distilled tetrahydrofuran (THF) for 30 min at 65◦ C
7. Prepare a 0.25 M vinylmagnesium bromide solution with freshly distilled THF, flush the
capillary with this solution for 1 h and heat for 8 h at 70◦ C
8. Flush the capillary with freshly distilled THF for 1 h and water for 30 min
9. Cut the capillary into 50 cm pieces
10. Add 5 µL of freshly prepared 10% ammonium persulfate and 2 µL tetramethylethylene-
diamine (TEMED) to a monomer solution and immediately flush this solution through
the capillaries for 5 min
11. Polymerize the capillaries at room temperature for 1 h
12. Flush the capillaries with water for 1 h
Freshly distilled THF must be used in each step to prevent the precipitation of vinylmagnesium
bromide in the capillary. To increase the robustness of the capillary modification, steps (4) and
(7) may be repeated. The AAP monomer can be synthesized in a single step reaction with >99%
yield and >99% purity.86 The polymerization time and concentrations of ammonium persulfate and
TEMED must be optimized for each batch of AAP. The success of the procedure can be checked
by measuring the electroosmotic flow, which should be <0.5 × 10–5 cm2 V–1 s–1 . The success rate
of the procedure described above is ∼90%, but decreases dramatically if freshly distilled THF or
anhydrous reagents are not used.
As described above, covalently linking AAP to the capillary surface is a tedious and time-
consuming task. Procedures that utilize silane chemistry are faster and easier to perform, but the
resulting Si–O–Si bonds are prone to nucleophilic cleavage and degrade relatively quickly. Simpler
procedures using dynamic coatings or unmodified capillaries would be beneficial and have also been
investigated. Under tightly controlled conditions, it was reported that mouse liver mitochondria
could be analyzed in unmodified fused-silica capillaries, but this method was not easily adapted to
other samples, including rat liver mitochondria, and requires strict attention to buffer composition.88
Poly(vinyl alcohol) was investigated as a dynamic capillary coating and reduced mitochondrial
adsorption to values that were similar to those for poly(AAP) coated capillaries. Although such
dynamic coatings show potential, poly(AAP) coated capillaries continue to be the gold standard for
organelle separations.
20.5.3 SEPARATION CONDITIONS

To separate analytes from a dissolved organelle fraction, a useful and popular mode of separation is
MEKC. As applied to subcellular analysis, the MEKC buffer is often used to dissolve the organelles
and acts to separate the components based on their hydrophobicity. A common buffer utilized for
this purpose is 10 mM borate, 10 mM sodium dodecyl sulfate (SDS) pH 9.5 (BS buffer), which has
been used to separate doxorubicin and its metabolites from nuclear, mitochondrial, and cytosolic-
enriched fractions.25,26 Here, separations were performed from each fraction by injecting a small
sample plug that had been dissolved in BS buffer and separating at 400 V/cm in a fused-silica
capillary while using BS buffer as the running buffer. To increase the resolution between structurally
similar molecules, additive can be used to improve the selectivity. For example, γ-cyclodextrin
has been added to BS buffer to resolve doxorubicin and a structurally similar chiral metabolite,
doxorubicinol.89 MEKC buffers that contain detergents (e.g., SDS) also reduce the interactions of
molecules with the capillary wall; consequently, capillary surface modifications are rarely needed
for an MEKC separation, greatly simplifying the analysis.
To separate intact organelles, iso-osmotic buffers must be used to reduce disruption caused by
osmotic shock. For instance, 220 mM mannitol, 70 mM sucrose, 0.5 mM ethylene glycol tetraacetic
acid (EGTA), 2 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), pH 7.4, form a
commonly used buffer for isolating mitochondria. The inclusion of mannitol as an osmotic support
reduces glycogen binding, whereas the sucrose reduces organelle aggregation. Since organelles
will spend several minutes in the running buffer during the separation, it must also be iso-osmotic.
Accordingly, a buffer containing 250 mM sucrose, 10 mM HEPES, pH 7.4, has been used in the
CE analysis of nuclei,76 mitochondria,17 and acidic organelles.81 For mitochondria, the pH should
be adjusted with KOH, as K+ ions are less permeant to the mitochondrial inner membrane than is
Na+ , thereby reducing mitochondrial degradation. Another consideration is that large electric fields
(e.g., 600 V/cm) can electrically disrupt organelle membranes.90 To minimize electric field-induced
organelle disruption, smaller electric fields (e.g., –200 V/cm) must then be used to separate organelles.
20.5.4 ORGANELLE LABELING

Instead of performing time-consuming and tedious organelle purification techniques, it is often
possible to use organelle-specific fluorescent labels to provide the necessary detection selectivity.
Such labels allow crudely purified organelle fractions, prepared from differential centrifugation, to
be analyzed. Many organelle-specific fluorescent labels are commercially available and Table 20.3
lists some of the most commonly used labels.
In the biological chemistry field, fusion proteins have recently shown more potential as organelle-
specific labels. For subcellular analysis, fusion proteins can be expressed that contain both a
fluorescent protein and a subcellular localization sequence. This combination results in organelle-
specific labeling that retains the excellent photochemical properties of fluorescent proteins (e.g.,
intense fluorescence and photostability). We commonly use a commercially available plasmid that
codes for a fusion protein of red fluorescent protein (DsRed2) and the mitochondrial targeting
sequence from subunit VIII of cytochrome c oxidase, which contains the neomycin/kanamycin
resistance gene, as described below.
TABLE 20.3
Organelle-Specific Fluorescent Labels
Organelle Fluorescent Label Target
Nuclei Ethidium bromide Double-stranded DNA

Mitochondria NAO Cardiolipin
Mitotracker Green FM Reacts with thiol groups in mitochondria
Rhodamine 123 Membrane potential dependent accumulation
Acidic organelles Fluorescein dextran conjugate Endocytosed into acidic compartments
LysoTracker Green Weak base is protonated and accumulated
Cytoskeletal network Fluorescein Phalloidin Actin
Oregon Green 488 paclitaxel Tubulin
Transfection of mammalian cells:
1. Suspend pDsRed2-Mito plasmid (Clontech, Mountain View, CA) in OPTI-MEM medium

(Invitrogen, Carlsbad, CA) and incubate with DMRIE-C (Invitrogen) at room temperature
for 30 min
2. Wash the 143B human osteosarcoma cells with OPTI-MEM
3. Layer the DMRIE-C-plasmid complex over the cells and incubate for 6 h
4. Add cell culture medium (MEM with 20% calf serum)
5. After 24 h, seed the cells onto a Petri dish at low density and culture in growth medium
that contains 500 µg/mL geneticin
6. After 5–6 days, select a single colony of cells that have been successfully transfected
7. Subculture the stably fluorescent transformants in medium [Dulbecco’s modified Eagle’s
medium (DMEM)] containing 250 µg/mL geneticin
In addition to mitochondria, commercial plasmids are also available that code for fusion proteins
that incorporate subcellular localization signals for nuclei, endoplasmic reticulum, Golgi apparatus,
peroxisomes, and actin filaments.

Choosing a detection scheme is an important component in designing any CE system, and as shown in
the preceding examples (Section 20.4), this can be done in many ways. Owing to the small injection
volumes used in CE, often the LOD must be submicromolar or lower; in these instances, achieving the
required LOD becomes our chief concern when designing a detection strategy. As illustrated above,
LIF, UV absorbance, and electrochemical detection may be employed for subcellular analysis by CE,
but other techniques such as mass spectrometry35 and biological cell sensors2 have also been used.
20.5.5.1 UV Absorbance
UV absorbance is the most common detection method for any CE system, primarily because nearly
all compounds absorb somewhere in the UV region of the spectrum, and also because fused-silica
capillaries are transparent to UV radiation. Given these advantages, UV absorbance detectors from
HPLC instruments were naturally adopted into CE systems, but a problem arises when these detectors
are used for subcellular analysis: subcellular analyses commonly quantify analytes that are below
the LOD for UV absorbance. According to Beer’s law, the sensitivity of a UV absorbance detector is
related to the pathlength of the light. Therefore, UV absorbance has limited applicability to traditional
CE separations due to the short path of the typical capillary diameters.
20.5.5.2 Electrochemical Detection

Electrochemical detection is less widely used than UV absorbance or LIF detection, but has been used
successfully by Ewing and coworkers to detect neurotransmitters within subcellular structures.6–11
Electrochemical detection is attractive because it provides a lower LOD than UV absorbance,
does not require sample derivatization, and in the case of amperometric detection, can be tuned
to specific classes of compounds. Furthermore, unlike UV absorbance in which the sensitivity
is dependent on the sample volume (i.e., the pathlength), the LOD of electrochemical detection
improves when applied to miniaturized CE systems, since in this case sensitivity is related to contact
between the analyte and the electrode surface. In fact, as described above, LODs in the zepto-
mole range for the subcellular quantification of dopamine have been reported,7 with the sensitivity
of these results making electrochemical detection the rival of many LIF detectors. However, the
application of electrochemical detection to subcellular analysis has been limited to the quantification
of neurotransmitters.
20.5.5.3 Laser-Induced Fluorescence

Lastly, the most popular detection technique for subcellular analysis is LIF due to the extremely low
LODs that can be attained. Mass LODs on the order of attomoles can be achieved with commercial
CE-LIF instruments; and custom-built instruments, in conjunction with sheath-flow cuvettes, can
reach down to the yoctomole (10–24 mole) scale.91 LIF detection thus allows the quantification of
minute amounts of analyte, and is the only detection method that has been used to detect individual
organelles. We will therefore describe off-column LIF detection in further detail.
Sheath-flow cuvettes were originally developed for flow cytometry, but were adapted for use
with CE by Dovichi.92–94 Figure 20.16 illustrates LIF detection in conjunction with a sheath-flow
cuvette. As displayed, the capillary is surrounded by sheath-flow buffer inside the cuvette. As
organelles migrate out of the capillary, they are hydrodynamically focused by the sheath flow into a
narrow stream. The excitation source is focused beneath the capillary outlet to excite the organelles
once they have been focused. Since the diameter of the sample stream is dependent on the difference
between the volumetric flow rates of the sample stream and the sheath flow, the sample stream can
be narrowed to the desired width by increasing the sheath volumetric flow.
At this point, the fluorescence from each organelle can be collected by a microscope objective. As
with all LIF detectors, care must be taken to choose an objective with high numerical aperture (NA)
that can maximize the collection efficiency. The collected fluorescence can then be spatially filtered
with a pinhole placed at the image plane to remove out-of-focus scatter from the buffer–cuvette
interfaces. The size of the pinhole should be matched with the magnification of the collection objective
Capillary
Sheath flow Sheath flow
Detector
Sheath Sheath
flow flow
Laser
Sheath-flow cuvette Hydrodynamic focusing
FIGURE 20.16 Diagram of the sheath-flow cuvette used for LIF detection. The capillary is inserted into
the sheath-flow cuvette and the laser beam is focused several micrometers away from the outlet. Sheath fluid
focuses the sample as it migrates out of the capillary into a narrow stream, in which it is excited by a focused
laser beam. The fluorescence is collected at a 90◦ angle from the excitation source by a high N.A. objective.
(Reprinted from Johnson, R. et al., Anal. Bioanal. Chem., In Press. Copyright 2006. With kind permission from
Springer Science and Business Media.)
and desired detection area. For instance, if the desired spot diameter is 50 µm and the fluorescence
is collected with a 60× objective, a 3-mm diameter pinhole is required. The collected fluorescence
must also be spectrally filtered to remove Raman and Raleigh scattering. In general, a long-pass
optical filter is used to block Raleigh scattering from the excitation source, and a bandpass filter that
encompasses the fluorophore emission to exclude Raman scattering. The Raman scattering bands
must be taken into consideration when choosing the emission wavelength range, since a lower LOD
may be achieved at wavelengths that are not at the maximum emission of the fluorophore if it coincides
with a Raman band. Finally, the fluorescence is detected using a photomultiplier tube or photodiode.
20.5.6 DATA ANALYSIS

Electropherograms that contain wide peaks, that is, peaks with baseline widths of several seconds
or more, are routinely analyzed. However, in the case of individual organelle detection, the peak
widths are much smaller, typically ∼80 ms, and the electropherograms nearly always contain wide
peaks, baseline drift, and the sharp spikes that correspond to the detection of individual organelles.
Special data treatment is required to discern these types of detected events. Toward this end, Jorgen-
son and coworkers95 described an important data analysis tool, wherein a moving median filter is
used to remove low-frequency background drift from a chromatogram. The median filter is shown
mathematically as follows:
yi = median (Ji )
Ji = {xi−r , xi−r+1 , xi−1 xi , xi+1 xi+r−1 , xi+r }
i = 0, 1, 2, n − 1
where Ji is a subset of the input array X (data to be analyzed) centered around the ith element of X,
yi is the ith element of the output array Y , n is the size of the input array, and r is the median filter
rank.95 Consequently, peaks that do not comprise over 50% of the subset are excluded from the data
set and only broad peaks remain.
For organelle analysis, median filtering is performed with a large rank (large filter window) so that
all the sharp spikes are removed from the electropherogram. Then, the filtered electropherogram,
which contains the wide peaks and background drift, is subtracted from the original to yield an
electropherogram with a flat baseline and sharp spikes that represent the detection of individual
organelles. As reported, the optimum rank value should generally be equal to, or greater than,
the baseline peak width of the widest peak of interest.95 However, in the analysis of individual
organelles, we find that the optimum rank value is much greater than the baseline peak width; that
is, typically greater than 10 times the average peak width. When the rank is too low, negative peaks
will be observed in the filtered data. Following the median filter procedure, organelle events can be
discriminated from noise based on their intensities. Frequently, five times the standard deviation in
the background is used as the threshold to identify organelles.

As shown throughout this chapter, CE is a very versatile and powerful separation technique that has
proven useful in subcellular analyses. As in all applications of CE, a trend for subcellular analysis is
increasing throughput by using microfluidic devices to reduce separation times. This was recently
demonstrated by analyzing fluorescently labeled mitochondria, for the first time, on a microfluidic
chip.96 The analysis allowed a fivefold reduction in analysis time, from 20 to 4 min. To realize the full
potential of CE, advances in electrophoretic models must also be made. Models that can accurately
predict electrophoretic mobility, as well as explain observed differences, will allow this information
to be used more effectively. Finally, as the information that can be gleaned from an electrophoretic
separation is limited, the development of complimentary techniques that can be combined with CE
will allow more information about subcellular compartments to be discovered.
REFERENCES
1. Miao, H., Rubakhin, S. S., and Sweedler, J. V., Subcellular analysis of d-aspartate, Anal. Chem., 77,
7190–7194, 2005.
2. Sims, C. E., Luzzi, V., and Allbritton, N. L., Localized sampling, electrophoresis, and biosensor
analysis of Xenopus laevis cytoplasm for subcellular biochemical assays, Methods Mol. Biol., 322,
413–424, 2006.
3. Ahmadzadeh, H., Johnson, R. D., Thompson, L., and Arriaga, E. A., Direct sampling from muscle
cross sections for electrophoretic analysis of individual mitochondria, Anal. Chem., 76, 315–321,
2004.
4. Gunasekera, N., Xiong, G., Musier-Forsyth, K., and Arriaga, E., A capillary electrophoretic method
for monitoring the presence of alpha-tubulin in nuclear preparations, Anal. Biochem., 330, 1–9, 2004.
5. Johnson, R. D., Navratil, M., Poe, B. G., Xiong, G., Olson, K. J., Ahmadzadeh, H., Andreyev, D.,
Duffy, C. F., and Arriaga, E. A., Analysis of mitochondria isolated from single cells, Anal. Bioanal.
Chem., 387, 107–118, 2006.
6. Woods, L. A., Gandhi, P. U., and Ewing, A. G., Electrically assisted sampling across membranes with
electrophoresis in nanometer inner diameter capillaries, Anal. Chem., 77, 1819–1823, 2005.
7. Woods, L. A., Powell, P. R., Paxon, T. L., and Ewing, A. G., Analysis of mammalian
cell cytoplasm with electrophoresis in nanometer inner diameter capillaries, Electroanalysis, 17,
1192–1197, 2005.
8. Kristensen, H. K., Lau, Y. Y., and Ewing, A. G., Capillary electrophoresis of single cells: Observation
of two compartments of neurotransmitter vesicles, J. Neurosci. Methods, 51, 183–188, 1994.
9. Chien, J. B., Wallingford, R. A., and Ewing, A. G., Estimation of free dopamine in the cytoplasm of the
giant dopamine cell of Planorbis corneus by voltammetry and capillary electrophoresis, J. Neurochem.,
54, 633–638, 1990.
10. Olefirowicz, T. M. and Ewing, A. G., Dopamine concentration in the cytoplasmic compartment of
single neurons determined by capillary electrophoresis, J. Neurosci. Methods, 34, 11–15, 1990.
11. Olefirowicz, T. M. and Ewing, A. G., Capillary electrophoresis in 2 and 5 microns diameter capillaries:
Application to cytoplasmic analysis, Anal. Chem., 62, 1872–1876, 1990.
12. Ho, A. M. and Yeung, E. S., Capillary electrophoretic study of individual exocytotic events in single
mast cells, J. Chromatogr. A, 817, 377–382, 1998.
13. Lillard, S. J., Chiu, D. T., Scheller, R. H., Zare, R. N., Rodriguez-Cruz, S. E., Williams, E. R., Orwar, O.,
Sandberg, M., and Lundqvist, J. A., Separation and characterization of amines from individual atrial
gland vesicles of Aplysia californica, Anal. Chem., 70, 3517–3524, 1998.
14. Chiu, D. T., Lillard, S. J., Scheller, R. H., Zare, R. N., Rodriguez-Cruz, S. E., Williams, E. R., Orwar, O.,
Sandberg, M., and Lundqvist, J. A., Probing single secretory vesicles with capillary electrophoresis,
Science, 279, 1190–1193, 1998.
15. Ahmadzadeh, H., Thompson, L. V., andArriaga, E.A., On-column labeling for capillary electrophoretic
analysis of individual mitochondria directly sampled from tissue cross sections, Anal. Bioanal. Chem.,
384, 169–174, 2006.
16. Fuller, K. M. and Arriaga, E. A., Analysis of individual acidic organelles by capillary electrophore-
sis with laser-induced fluorescence detection facilitated by the endocytosis of fluorescently labeled
microspheres, Anal. Chem., 75, 2123–2130, 2003.
17. Fuller, K. M., Duffy, C. F., and Arriaga, E. A., Determination of the cardiolipin content of individual
mitochondria by capillary electrophoresis with laser-induced fluorescence detection, Electrophoresis,
23, 1571–1576, 2002.
18. Gunasekera, N., Lee, S. W., Kim, S., Musier-Forsyth, K., and Arriaga, E., Nuclear localization of
aminoacyl-tRNA synthetases using single-cell capillary electrophoresis laser-induced fluorescence
analysis, Anal. Chem., 76, 4741–4746, 2004.
19. Li, H., Sims, C. E., Wu, H. Y., and Allbritton, N. L., Spatial control of cellular measurements with the
laser micropipet, Anal. Chem., 73, 4625–4631, 2001.
20. Lochmann, H., Bazzanella, A., and Bachmann, K., Analysis of solutes and metabolites in single plant
cell vacuoles by capillary electrophoresis, J. Chromatogr. A, 817, 337–343, 1998.
21. Luzzi, V., Lee, C. L., and Allbritton, N. L., Localized sampling of cytoplasm from Xenopus oocytes
for capillary electrophoresis, Anal. Chem., 69, 4761–4767, 1997.
22. Muscari, C., Pappagallo, M., Ferrari, D., Giordano, E., Capanni, C., Caldarera, C. M., and
Guarnieri, C., Simultaneous detection of reduced and oxidized glutathione in tissues and mitochondria
by capillary electrophoresis, J. Chromatogr. B Biomed. Sci. Appl., 707, 301–307, 1998.
23. Yoon, S., Ban, E., and Yoo, Y. S., Direct monitoring of the expression of the green fluorescent protein-
extracellular signal-regulated kinase 2 fusion protein in transfected cells using capillary electrophoresis
with laser-induced fluorescence detection, J. Chromatogr. A, 976, 87–93, 2002.
24. McKeon, J. and Khaledi, M. G., Quantitative nuclear and cytoplasmic localization of antisense oligonu-
cleotides by capillary electrophoresis with laser-induced fluorescence detection, Electrophoresis, 22,
3765–3770, 2001.
25. Anderson, A. B. and Arriaga, E. A., Subcellular metabolite profiles of the parent ccrf-cem and the
derived cem/c2 cell lines after treatment with doxorubicin, J. Chromatogr. B Analyt. Technol. Biomed.
Life Sci., 808, 295–302, 2004.
26. Anderson, A. B., Ciriacks, C. M., Fuller, K. M., and Arriaga, E. A., Distribution of zeptomole-
abundant doxorubicin metabolites in subcellular fractions by capillary electrophoresis with laser-
induced fluorescence detection, Anal. Chem., 75, 8–15, 2003.
27. Cardoso, C. D., Jabor, V. A. P., and Bonato, P. S., Capillary electrophoretic chiral separation of hydrox-
ychloroquine and its metabolites in the microsomal fraction of liver homogenates, Electrophoresis,
27, 1248–1254, 2006.
28. Chen, Y., Walsh, R. J., and Arriaga, E. A., Selective determination of the doxorubicin con-
tent of individual acidic organelles in impure subcellular fractions, Anal. Chem., 77, 2281–2287,
2005.
29. Clohs, L. and Wong, J., Validation of a capillary electrophoresis assay for assessing the metabolic
stability of verapamil in human liver microsomes, J. Capillary Electrophor., 7, 113–117, 2002.
30. Liu, H. C., Wang, N., Yu, Y., and Hou, Y. N., Stereoselectivity in trans-tramadol metabolism
and trans-O-demethyltramadol formation in rat liver microsomes, Acta Pharmacol. Sin., 24,
85–90, 2003.
31. Sakai-Kato, K., Kato, M., and Toyo’oka, T., On-line drug-metabolism system using microsomes
encapsulated in a capillary by the sol-gel method and integrated into capillary electrophoresis, Anal.
Biochem., 308, 278–284, 2002.
32. Ward, E. M., Smyth, M. R., O’Kennedy, R., and Lunte, C. E., Application of capillary electrophoresis
with pH-mediated sample stacking to analysis of coumarin metabolites in microsomal incubations,
J. Pharm. Biomed. Anal., 32, 813–822, 2003.
33. Lin, C., Cotton, F., Boutique, C., Dhermy, D., Vertongen, F., and Gulbis, B., Capillary gel elec-
trophoresis: Separation of major erythrocyte membrane proteins, J. Chromatogr. B Biomed. Sci. Appl.,
742, 411– 419, 2000.
34. Prange, A., Schaumloffel, D., Bratter, P., Richarz, A. N., and Wolf, C., Species analysis of met-
allothionein isoforms in human brain cytosols by use of capillary electrophoresis hyphenated to
inductively coupled plasma-sector field mass spectrometry, Fresenius J. Anal. Chem., 371, 764–774,
2001.
35. Moini, M. and Huang, H., Application of capillary electrophoresis/electrospray ionization-mass
spectrometry to subcellular proteomics of Escherichia coli ribosomal proteins, Electrophoresis, 25,
1981–1987, 2004.
36. Presley, A. D., Fuller, K. M., and Arriaga, E. A., Mitotracker green labeling of mitochondrial proteins
and their subsequent analysis by capillary electrophoresis with laser-induced fluorescence detection,
J. Chromatogr. B Analyt. Technol. Biomed. Life Sci., 793, 141–150, 2003.
37. Kustos, T., Kustos, I., Gonda, E., Kocsis, B., Szabo, G., and Kilar, F., Capillary electrophoresis study
of outer membrane proteins of pseudomonas strains upon antibiotic treatment, J. Chromatogr. A, 979,
277–284, 2002.
38. Fazal, M. A., Palmer, V. R., and Dovichi, N. J., Analysis of differential detergent fractions of an
AtT-20 cellular homogenate using one- and two-dimensional capillary electrophoresis, J. Chromatogr.
A, 1130, 182–189, 2006.
39. Heidrich, H. G., Stahn, R., and Hannig, K., The surface charge of rat liver mitochondria and their
membranes. Clarification of some controversies concerning mitochondrial structure, J. Cell Biol., 46,
137–150, 1970.
40. Henning, R. and Heidrich, H. G., Membrane lipids of rat liver lysosomes prepared by free-flow
electrophoresis, Biochim. Biophys. Acta, 345, 326–335, 1974.
41. Stahn, R., Maier, K. P., and Hannig, K., A new method for the preparation of rat liver lysosomes.
Separation of cell organelles of rat liver by carrier-free continuous electrophoresis, J. Cell Biol., 46,
576–591, 1970.
42. Valdivia, E., Pease, B., Gabel, C., and Chan, V., Electrophoresis of isolated mitochondria, Anal.
Biochem., 51, 146–151, 1973.
43. Kamo, N., Muratsugu, M., Kurihara, K., and Kobatake, Y., Change in surface charge density and
membrane potential of intact mitochondria during energization, FEBS Lett., 72, 247–250, 1976.
44. Sengelov, H., Nielsen, M. H., and Borregaard, N., Separation of human neutrophil plasma membrane
from intracellular vesicles containing alkaline phosphatase and NADPH oxidase activity by free flow
electrophoresis, J. Biol. Chem., 267, 14912–14917, 1992.
45. Gottlieb, M. H., Steer, C. J., Steven, A. C., and Chrambach, A., Applicability of agarose gel elec-
trophoresis to the physical characterization of clathrin-coated vesicles, Anal. Biochem., 147, 353–363,
1985.
46. Tulp, A., Verwoerd, D., Fernandez-Borja, M., Neefjes, J., and Hart, A. A., High resolution density
gradient electrophoresis of cellular organelles, Electrophoresis, 17, 173–178, 1996.
47. Volkl, A., Mohr, H., Weber, G., and Fahimi, H. D., Isolation of rat hepatic peroxisomes by means of
immune free flow electrophoresis, Electrophoresis, 18, 774–780, 1997.
48. Weber, G. and Bocek, P., Optimized continuous flow electrophoresis, Electrophoresis, 17, 1906–1910,
1996.
49. Hasan, Z. and Pieters, J., Subcellular fractionation by organelle electrophoresis: Separation of
phagosomes containing heat-killed yeast particles, Electrophoresis, 19, 1179–1184, 1998.
50. Hunter, R. J., Foundations of Colloid Science, 2nd ed. Oxford University Press, Oxford, 2001.
51. Overbeek, J. T. G. and Wiersema, P. H., The Interpretation of Electrophoretic Mobilities, Academic
Press, New York, 1967.
52. Booth, F., Theory of electrokinetic effects, Nature, 161, 83–86, 1948.
53. Wiersema, P. H., Loeb, A. L., and Overbeek, J. T. G., Calculation of the electrophoretic mobility of a
spherical colloid particle, J. Coll. Interf. Sci., 22, 78–99, 1966.
54. Overbeek, J. T. G., Quantitative interpretation of the electrophoretic velocity of colloids, Adv. Coll.
Sci., 3, 97–134, 1950.
55. O’Brien, R. W., The solution of the electrokinetic equations for colloidal particles with thin double
layers, J. Coll. Interf. Sci., 92, 204–216, 1983.
56. O’Brien, R. W. and Ward, D. N., The electrophoresis of a spheroid with a thin double layer, J. Coll.
Interf. Sci., 121, 402–413, 1988.
57. O’Brien, R. W. and Hunter, R. J., The electrophoretic mobility of large colloidal particles, Can. J.
Chem., 59, 1878–1887, 1981.
58. O’Brien, R. W. and White, L. R., Electrophoretic mobility of a spherical colloidal particle, J. Chem.
Soc., Faraday Trans., 77, 1607–1626, 1978.
59. Von Smoluchowski, M., Bull. Acad. Sci. Cracovie, 182, 182–200, 1903.
60. Hückel, E., Die Kataphorese der Kugel, Physik Z., 25, 204, 1924.
61. Henry, D. C., The cataphoresis of suspended particles. I. The equation of cataphoresis, Proc. Roy. Soc.
Lond., A133, 106, 1931.
62. Ohshima, H., A simple expression of Henry’s function for the retardation effect in electrophoresis of
spherical colloidal particles, J. Coll. Interf. Sci., 168, 269–271, 1994.
63. Pysher, M. D. and Hayes, M. A., Examination of the electrophoretic behavior of liposomes, Langmuir,
20, 4369–4375, 2004.
64. Pysher, M. D. and Hayes, M. A., Effects of deformability, uneven surface charge distributions, and
multipole moments on biocolloid electrophoretic migration, Langmuir, 21, 3572–3577, 2005.
65. Gaucher, S. P., Taylor, S. W., Fahy, E., Zhang, B., Warnock, D. E., Ghosh, S. S., and Gib-
son, B. W., Expanded coverage of the human heart mitochondrial proteome using multidimensional
liquid chromatography coupled with tandem mass spectrometry, J. Proteome Res., 3, 495–505, 2004.
66. Lescuyer, P., Strub, J. M., Luche, S., Diemer, H., Martinez, P., Van Dorsselaer, A., Lunardi, J., and
Rabilloud, T., Progress in the definition of a reference human mitochondrial proteome, Proteomics, 3,
157–167, 2003.
67. Sickmann, A., Reinders, J., Wagner, Y., Joppich, C., Zahedi, R., Meyer, H. E., Schonfisch, B.,
et al., The proteome of Saccharomyces cerevisiae mitochondria, Proc. Natl. Acad. Sci. USA, 100,
13207–13212, 2003.
68. Strong, R., Nakanishi, T., Ross, D., and Fenselau, C., Alterations in the mitochondrial proteome of
adriamycin resistant mcf-7 breast cancer cells, J. Proteome Res., 5, 2389–2395, 2006.
69. Hu, S., Michels, D. A., Fazal, M. A., Ratisoontorn, C., Cunningham, M. L., and Dovichi, N. J., Cap-
illary sieving electrophoresis/micellar electrokinetic capillary chromatography for two-dimensional
protein fingerprinting of single mammalian cells, Anal. Chem., 76, 4044–4049, 2004.
70. Michels, D. A., Hu, S., Dambrowitz, K. A., Eggertson, M. J., Lauterbach, K., and Dovichi, N. J., Capil-
lary sieving electrophoresis-micellar electrokinetic chromatography fully automated two-dimensional
capillary electrophoresis analysis of Deinococcus radiodurans protein homogenate, Electrophoresis,
25, 3098–3105, 2004.
71. Michels, D. A., Hu, S., Schoenherr, R. M., Eggertson, M. J., and Dovichi, N. J., Fully automated
two-dimensional capillary electrophoresis for high sensitivity protein analysis, Mol. Cell. Proteomics,
1, 69–74, 2002.
72. Hu, S., Jiang, J., Cook, L. M., Richards, D. P., Horlick, L., Wong, B., and Dovichi, N. J., Capillary
sodium dodecyl sulfate-dalt electrophoresis with laser-induced fluorescence detection for size-based
analysis of proteins in human colon cancer cells, Electrophoresis, 23, 3136–3142, 2002.
73. Ramsby, M. L., Makowski, G. S., and Khairallah, E. A., Differential detergent fractionation of isolated
hepatocytes: Biochemical, immunochemical and two-dimensional gel electrophoresis characterization
of cytoskeletal and noncytoskeletal compartments, Electrophoresis, 15, 265–277, 1994.
74. Cardoso, C. D. and Bonato, P. S., Enantioselective analysis of the metabolites of hydroxychloroquine
and application to an in vitro metabolic study, J. Pharm. Biomed. Anal., 37, 703–708, 2005.
75. Gunasekera, N., Musier-Forsyth, K., and Arriaga, E., Electrophoretic behavior of individual
nuclear species as determined by capillary electrophoresis with laser-induced fluorescence detection,
76. Gunasekera, N., Olson, K. J., Musier-Forsyth, K., and Arriaga, E. A., Capillary electrophoretic
separation of nuclei released from single cells, Anal. Chem., 76, 655–662, 2004.
77. Anderson, A. B., Xiong, G., and Arriaga, E. A., Doxorubicin accumulation in individually
electrophoresed organelles, J. Am. Chem. Soc., 126, 9168–9169, 2004.
78. Fuller, K. M. and Arriaga, E. A., Capillary electrophoresis monitors changes in the electrophoretic
behavior of mitochondrial preparations, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci., 806,
151–159, 2004.
79. Meany, D. L., Poe, B. G., Navratil, M., Moraes, C. T., and Arriaga, E. A., Superoxide released into the
mitochondrial matrix, Free Radic. Biol. Med., 41, 950–959, 2006.
80. Strack, A., Duffy, C. F., Malvey, M., and Arriaga, E. A., Individual mitochondrion characterization:
A comparison of classical assays to capillary electrophoresis with laser-induced fluorescence detection,
Anal. Biochem., 294, 141–147, 2001.
81. Chen, Y. and Arriaga, E. A., Individual acidic organelle pH measurements by capillary electrophoresis,
Anal. Chem., 78, 820–826, 2006.
82. Schindler, M., Grabski, S., Hoff, E., and Simon, S. M., Defective pH regulation of acidic compart-
ments in human breast cancer cells (mcf-7) is normalized in adriamycin-resistant cells (mcf-7adr),
Biochemistry, 35, 2811–2817, 1996.
83. Woods, L. A. and Ewing, A. G., Etched electrochemical detection for electrophoresis in nanometer
inner diameter capillaries, Chemphyschem, 4, 207–211, 2003.
84. Storrie, B. and Madden, E. A., Isolation of subcellular organelles, Methods Enzymol., 182, 203–225,
1990.
85. Okado-Matsumoto, A. and Fridovich, I., Subcellular distribution of superoxide dismutases (SOD) in
rat liver: Cu, Zn-SOD in mitochondria, J. Biol. Chem., 276, 38388–38393, 2001.
86. Simo-Alfonso, E., Gelfi, C., Sebastiano, R., Citterio, A., and Righetti, P. G., Novel acrylamido
monomers with higher hydrophilicity and improved hydrolytic stability: I. Synthetic route and product
characterization, Electrophoresis, 17, 723–731, 1996.
87. Gelfi, C., Curcio, M., Righetti, P. G., Sebastiano, R., Citterio, A., Ahmadzadeh, H., and Dovichi, N. J.,
Surface modification based on Si-O and Si-C sublayers and a series of n-substituted acrylamide top-
layers for capillary electrophoresis, Electrophoresis, 19, 1677–1682, 1998.
88. Whiting, C. and Arriaga, E. A., CE-LIF analysis of mitochondria using uncoated and dynamically
coated capillaries, Electrophoresis, 27, 4523–4531, 2006.
89. Eder, A. R., Chen, J. S., and Arriaga, E. A., Separation of doxorubicin and doxorubicinol
by cyclodextrin-modified micellar electrokinetic capillary chromatography, Electrophoresis, 27,
3263–3270, 2006.
90. Zimmerman, U. and Neil, G., Electromanipulation of Cells, CRC Press, New York, 1996.
91. Chen, D. Y. and Dovichi, N. J., Yoctomole detection limit by laser-induced fluorescence in capillary
electrophoresis, J. Chromatogr. B Biomed. Appl., 657, 265–269, 1994.
92. Dovichi, N. J., Martin, J. C., Jett, J. H., and Keller, R. A., Attogram detection limit for aqueous dye
samples by laser-induced fluorescence, Science, 219, 845–847, 1983.
93. Cheng, Y. F. and Dovichi, N. J., Subattomole amino acid analysis by capillary zone electrophoresis
and laser-induced fluorescence, Science, 242, 562–564, 1988.
94. Cheng, Y. F., Wu, S., Chen, D. Y., and Dovichi, N. J., Interaction of capillary zone electrophoresis
with a sheath flow cuvette detector, Anal. Chem., 62, 496–503, 1990.
95. Moore, A. W., Jr. and Jorgenson, J. W., Median filtering for removal of low-frequency background
drift, Anal. Chem., 65, 188–191, 1993.
96. Duffy, C. F., MacCraith, B., Diamond, D., O’Kennedy, R., and Arriaga, E. A., Fast electrophoretic
analysis of individual mitochondria using microchip capillary electrophoresis with laser induced
fluorescence detection, Lab. Chip, 6, 1007–1011, 2006.
21 Chemical Cytometry: Capillary
Electrophoresis Analysis at the
Level of the Single Cell
Colin Whitmore, Kimia Sobhani, Ryan Bonn, Danqian
Mao, Emily Turner, James Kraly, David Michels, Monica
Palcic, Ole Hindsgaul, and Norman J. Dovichi
CONTENTS
21.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 611

21.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 612
21.2.1 The Challenge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 612
21.2.2 Classic Cytometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 613
21.2.3 Chemical Cytometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 613
21.2.4 Chemical Cytometry of Proteins, Biogenic Amines, and Metabolic
Cascades by One- and Two-Dimensional Capillary Electrophoresis . . . . . . . . . . . . . 614
21.2.4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 614
21.2.4.2 High-Sensitivity Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 614
21.2.4.3 Proteins and Biogenic Amines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 614
21.2.4.4 Fluorescence Labeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 616
21.3.1 Chemical Cytometry of Biogenic Amines and Proteins Using One-Dimensional
Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 617
21.3.2 Chemical Cytometry of Biogenic Amines and Proteins Using
Two-Dimensional Capillary Electrophoresis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 619
21.3.3 Metabolic Cytometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 620
21.4.1 Injection Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 624
21.4.3 Fluorescence Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 627
21.4.4 Data Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 627
21.5 Concluding Remarks and Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 627
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 628
21.1 INTRODUCTION
The cell is the organizing unit of life, and characterization of the composition of cells is of value in
both fundamental research and in clinical applications. Classic analytical methods are often used to
characterize cellular homogenates produced from thousands to billions of cells. These methods are
611
FIGURE 21.1 Comparison of two cellular populations, one homogeneous, with a key component present at
intermediate levels, and the other heterogeneous, with some cells lacking the compound while others have that
compound in high abundance.
unable to distinguish between a homogeneous population of intermediate composition and a hetero-

geneous population made up of a mixture of high- and low-expression cells. Figure 21.1 presents
two samples that have the same average composition. In one case, the population is homogeneous
and in the other is highly heterogeneous; the responses of these populations to a stimulus are likely
to be quite different.
In many cases, it is clearly useful to study single cells, rather than an average of the population.
Classically, flow and image cytometry are used to measure physical properties and chemical com-
position of single cells, and are very important in both clinical and fundamental research. In these
methods, affinity reagents are usually used to determine the distribution of a handful of components
in a cellular population. While extremely powerful, these methods are limited in the number of com-
ponents resolved. More importantly, classic cytometry methods can only detect those components
for which affinity reagents are available; the unexpected is undetectable.
In contrast, chemical cytometry employs powerful ultrasensitive analytical methods to charac-
terize the composition of single cells. In principle, chemical cytometry can resolve hundreds of
components from a single cell.
Our work in this field is motivated by three broad applications. First, this technology will be of
value whenever characterizing complex tissues, such as neurons and cells of the immune system,
where there is a tremendous cell-to-cell heterogeneity in composition. For example, one can imag-
ine providing a detailed cell-by-cell map of selected brain structures or monitoring the evolution
in a T-cell population in response to an infection. Second, this technology will be to be of value
in characterizing the molecular changes associated with the maturation of stem and precursor cells
into their differentiated progeny; in an ideal experiment, an organism such as Caenorhabditis ele-
gans would be microdissected so that each cell can be analyzed as the organism develops from a
single-cell embryo to an adult. Finally, we believe that chemical cytometry may prove of value in
improving the accuracy of cancer prognosis and diagnosis; just as cell-to-cell variation in ploidy can
have prognostic value, determination of cell-to-cell heterogeneity in other cellular components may
improve prognostic accuracy.
21.2 BACKGROUND
21.2.1 THE CHALLENGE
Chemical cytometry employs modern analytical tools to characterize single cells. Analysis of the
content of a single cell requires extremely sensitive instrumentation. Consider a typical mammalian
cell that has a 10-µm diameter, 500-fL volume, and 500-pg mass. If that cell is 10% protein by
weight, then the cell contains 50 pg of protein. Assuming an average molar mass of 30,000 g/mol, a
single cell contains about 2 fmol of protein. The fraction of the proteome expressed by a single cell
is unknown; perhaps 10,000 different proteins are present in each cell, with an average of 200 zmol
Chemical Cytometry 613
(1 zmol = 10−21 mol = 600 copies) per protein. As we show below, there is a large distribution in
protein expression within a cell. In addition, the cell contains ∼15 fmol of amino acids and other
biogenic amines. Chemical cytometry requires exquisite detection sensitivity.
21.2.2 CLASSIC CYTOMETRY

Cytometry is the measurement of physical properties and chemical composition of single cells,
and a wonderful review of the field is given by Howard Shapiro.1 Cytometry was pioneered by
Caspersson,2,3 who developed microspectrophotomers that were used to measure ultraviolet (UV)
absorbance of single cells at 260 nm to characterize nucleic acids and at 280 nm to characterize
protein content. This work was particularly important because it led to the discovery that DNA
content of cells doubled during the cell cycle, strengthening the hypothesis that DNA contained
genetic material.
Image cytometry had its genesis in Caspersson’s instrumentation, where a scanning stage was
used to translate a sample through the focal volume of the microspectrophotometer, recording trans-
mission data from a large number of cells. Since then, many technological improvements have been
made to measure transmission, scatter, and fluorescence from cells and tissues.
Flow cytometers record optical and electrical signals from single cells flowing in a stream.
The earliest instruments were based on light scatter and electrical impedance and were used for
cell counting, replacing the hemocytometer. More sophisticated instruments record fluorescence
intensity in one or more spectral bands and can characterize perhaps a half-dozen components in a
single cell. The most sophisticated instruments are cell sorters that employ an electrical signal to
deflect cells of interest from the flowing stream into a receiving reservoir for subsequent culture or
analysis. Modern instruments can measure the signal from 200,000 cells/s and sort 50,000 cells/s.4
Classic cytometry typically employs fluorescence to characterize a few components per cell; by use
of several excitation wavelengths and multiple detection channels, instruments can characterize up
to a dozen parameters per cell.5
Fluorescence-based assays employ reagents to characterize a wide range of cellular components,
including nucleic acids, proteins, biogenic amines, carbohydrates, and lipids. These assays are often
based on affinity probes, such as oligonucleotides, antibodies, aptamers, and lectins, which are used
to characterize a specific component within the cell. In addition, fluorogenic substrates are used to
characterize enzyme activity in single cells.
21.2.3 CHEMICAL CYTOMETRY

Chemical cytometry employs modern analytical tools to characterize the chemical composition of
single cells. As an example, mass spectrometry has been used to monitor hemoglobin in single ery-
throcytes and neuropeptides in giant neurons. The information content of the mass spectrometer’s
signal is invaluable in component identification. Caprioli and coworkers has reported the use of
matrix-assisted laser desorption/ionization time of flight (MALDI-TOF) to analyze tissue samples.6
Over 800 components are detected from a tissue section.7 Li and coworkers has reported the detec-
tion of hemoglobin from a single erythrocyte by MALDI-TOF mass spectrometry(MS).8 Each cell
contains about 450 amol of hemoglobin, and only extremely highly expressed proteins are detected
and identified in a single cell. Sweedler and coworkers9 has reported the use of MALDI-TOF to
analyze neuropeptides in single cells.
Several separation methods have been developed for chemical cytometry. In an early example,
silk fibers were used for the electrophoretic determination of 100 pg of RNA contained within a
single cell, employing a UV microscope to detect the components.10 Kennedy et al.11 inaugurated
the modern era of chemical cytometry by using open tubular capillary chromatography for the
analysis of biogenic amines in a single giant neuron from a snail.
21.2.4 CHEMICAL CYTOMETRY OF PROTEINS, BIOGENIC AMINES, AND METABOLIC

CASCADES BY ONE- AND TWO-DIMENSIONAL CAPILLARY ELECTROPHORESIS
21.2.4.1 Introduction
A rich literature has developed for the use of capillary electrophoresis (CE) for chemical cytometry.
Wallingford and Ewing12 reported the use of a capillary to sample the internal contents of a single
giant neuron for electrophoresis of biogenic amines. Hogan and Yeung13 reported the use of a
specific label to derivatize thiols in individual erythrocytes; once the derivatization reaction was
completed, a single cell was injected into a capillary, lysed, and the contents separated by CE.
Gilman and Ewing14 reported the on-column labeling of amines from a cell that had been injected
into and lysed within a capillary; the capillary was used for electrophoretic separation of amines.
Allbritton and coworkers15 reported the use of a pulsed laser to lyse a cell before analysis of kinase
activities by CE. Han and Lillard16 reported the use of cell synchronization based on the shakeoff
method for the characterization of RNA synthesis in single cells as a function of cell cycle. Zare and
coworkers17 reported a microfabricated device to automate cell lysis, labeling, and separation.
21.2.4.2 High-Sensitivity Electrophoresis

Capillary electrophoresis employing a postcolumn sheath-flow cuvette has produced outstanding
detection limits, approaching single molecule levels in favorable cases.18,19 We have used chemical
cytometry based on CE to study proteins and biogenic amines in a wide range of cell types, including
the HT29 colon cancer cell line; the MCF7 breast cancer cell line; the AtT20 adrenal gland cancer
cell line; the MC3T3 osteoprecursor cell line; the hTERT telomerase-expressing Barrett’s esophagus
cell line; the A549 lung cancer cell line; the SupT1 T-cell line; RAW 264.7 macrophage cells;
Deinococcus radiodurans cells; primary neurons, macrophages, and T-cells; and single-cell C.
elegans and mouse embryos.
21.2.4.3 Proteins and Biogenic Amines

In chemical cytometry of protein and biogenic amines, cells are lysed and their constituents are
labeled with a fluorescent or fluorogenic reagent before separation by CE and detection by laser-
induced fluorescence. The use of fluorescent reagents to label proteins will target the most highly
abundant proteins in the cell. The best data on the protein expression in a cellular homogenate were
generated by Ghaemmaghami and coworkers,20 who created a tick anticoagulant peptide (TAP)
fusion of each open reading frame (ORF) in yeast; these cells were homogenized and assayed by
Western blotting. Figure 21.2 presents a cumulative sum of protein abundance, starting from the most
abundant protein. A few proteins are present at very high levels, and dominate the total population,
while a large fraction (25%) of ORFs generated no detectable signal. The most abundant protein
accounts for 3.4% of the total, and over 50% of the total protein content is due to the 100 most
abundant proteins in a yeast cell. Chemical cytometry methods that target all proteins in a cell will
be dominated by a relatively small number of highly abundant proteins.
We have also downloaded the ORF data for yeast;21 this database includes the predicted sequence
of 6057 ORFs that are 100 residues or longer. Note that these data are for the full-length ORF. As
result, all ORFs contain a methionine residue, which corresponds to the start codon. Post-translational
proteolysis will remove the leader sequence from the N-terminus of the mature protein. Figure 21.3
presents a histogram of the abundances of each amino acid in those ORFs. The figure also reports
the number of ORFs that do not contain that amino acid (n0 ) and the percentage of all amino acids.
For example, only 9 out of 6057 ORFs contain no lysine residues. The three most abundant amino
acids in yeast are leucine, serine, and lysine, which account for 9.6%, 9.0%, and 7.3% of all amino
acids, whereas the three least abundant amino acids are tryptophan, cysteine, and methionine, which
account for 1.0%, 1.3%, and 2.1% of all amino acids; the methionine abundance in mature proteins
1.0
Cumulative protein abundance

0.8
0.6
0.4
0.2
0.0
0 500 1000 1500 2000 2500 3000 3500 4000
Protein number
FIGURE 21.2 Cumulative abundance of proteins in yeast, starting from the most abundant protein. The
100 most abundant proteins represent 50% of the total protein content in this organism. Data obtained from
Reference 27.
Asn 150 600

150 Ala 200 Arg Asp Cys
150 n 0 = 13
n 0 = 17 n 0 = 32 n 0 = 42 400 n 0 = 458
100 4.5% 100 6.1% 100 5.8%
5.5% 1.3%
100
50 50 50 200
0 0 0 0 0
0 50 100 150 0 50 100 150 0 50 100 150 0 50 100 150 0 50 100 150
150 Glu Gly 400 His Ile

200 Gln 150 150
n 0 = 29 n 0 = 23 n 0 = 153 n0 = 4
n 0 = 68 100
6.5% 100 5.0% 2.1% 100 6.6%
3.9% 200
100 50 50
Number of ORFs
50
0 0 0 0 0
0 50 100 150 0 50 100 150 0 50 100 150 0 50 100 150 0 50 100 150
Leu Lys Met 200 Phe Pro

100 400 200
n0 = 0 100 n0 = 9 n0 = 0 n 0 = 11 n 0 = 15
9.6% 7.3% 2.1% 4.5% 4.3%
50 50 200 100 100
0 0 0 0 0
0 50 100 150 0 50 100 150 0 50 100 150 0 50 100 150 0 50 100 150
150 800 300
Ser Thr Trp Tyr Val
150 150
100 n0 = 2 n0 = 3 600 n 0 = 624 200 n 0 = 52 n0 = 3
9.0% 100 5.8% 400 1.0% 3.4% 100 5.6%
50 50 100 50
200
0 0 0 0 0
0 50 100 150 0 50 100 150 0 50 100 150 0 50 100 150 0 50 100 150
Amino acids per ORF
FIGURE 21.3 Amino acid distribution in yeast ORFs. A total of 6057 ORFs coding for 100 or more amino
acids were analyzed. The histogram of amino acid abundance in each ORF is plotted. The number of ORFs
lacking that amino acid (n0 ) is listed. Finally, the percentage of each amino acid in the yeast ORFs is listed as
the bottom line.
will be lower than predicted from these data because of the loss of the start codon during cleavage
of the signal sequence from the precursor.
Most labeling chemistry targets primary amines, which are the ε-amine of lysine residues and
unblocked N-termini. Primary amines are also found on amino acids, aminolipids, aminosugars,
neurotransmitters, and other biogenic amines, and understanding their role is of interest in physiology,
molecular biology, and cell biology. These compounds make up 5–10% of the content of a cell, and
are important mediators of a wide range of cellular processes. Their study is extremely difficult by
classic cytometry methods, such as flow cytometry, because of the lack of specific affinity reagents.
21.2.4.4 Fluorescence Labeling

While it is possible to detect proteins based on the native fluorescence of tryptophan, that amino acid
is the least abundant in yeast, representing 1.0% of the total amino acid content of yeast. Excitation of
tryptophan requires expensive and temperamental UV lasers. Instead, it is usually more convenient
to label the protein with a fluorophore that is excited by inexpensive and reliable lasers that operate
in the visible portion of the spectrum.
21.2.4.4.1 Chemical Derivatization

As noted above, lysine is one of the most abundant amino acids, and most chemical cytometry exper-
iments rely on reagents that target the primary amine found on lysine residues and the N-terminus.
Classic fluorescent reagents, such as fluorescein, rhodamines, Cy-dyes, and bodipy, are not useful
in chemical cytometry. While they create highly fluorescent products, unreacted reagent and trace
levels of fluorescent impurities create a sea of background signals that can swamp the fluorescence
signal from analyte. For example, if 1 nL of a 10−4 M labeling reagent is used in an experiment,
then there will be 100 amol of impurities present at the 0.1% level, and impurities present at the
part-per-million level will contribute 100 zmol, which corresponds to the average protein level in a
cell. These trace impurities are inevitably present, many are highly fluorescent, and they create a sea
of background peaks that obscures target compounds.
We instead employ fluorogenic reagents, such as 3-(2-furoyl) quinoline-2-carboxaldehyde (FQ),
in our experiments. These compounds are nonfluorescent until they react with a primary amine in the
presence of a nucleophile. Fluorogenic reagents have very few fluorescent impurities and generate
a very low background signal that does not interfere in the chemical cytometry experiment.
There is another subtle issue in fluorescent labeling of proteins. Heroic efforts are required
to completely label all lysine residues in a protein, and these efforts are impractical in single-cell
analysis.22 Instead, the reaction targets the most labile residues, and usually creates a complex mixture
of products; there are 2N − 1 fluorescent products generated from a protein with N primary amines.23
For example, ovalbumin has 20 lysine residues. Incomplete labeling can result in the production of
220 − 1 = 1048575 different fluorescent products and each product will have a different mobility
during electrophoresis, creating a complex electropherogram.23–25 Both this group and Whitesides
and coworkers26–28 have noted that addition of an anionic surfactant, such as sodium dodecyl sulfate
(SDS), causes the complex set of peaks to collapse into a single peak, which can have extraordinarily
high separation efficiency that exceeds 105 plates. However, this phenomenon only appears to occur
when acylation produces a neutral product and when an anionic surfactant is present.
An alternative strategy is to use very mild reaction conditions, so that few proteins have more
than one fluorescent label. However, these conditions lead, by necessity, to inefficient labeling.
21.2.4.4.2 Labeling by Genetic Engineering

Chemical labeling is best used to characterize the most abundant proteins in a cell. Low abundance
proteins tend to be buried under the signal generated by high abundance components. Fluores-
cently labeled antibodies are commonly used to label specific components in flow cytometry, but
cellular autofluorescence limits their use to characterize the lowest abundance proteins. In addition,
immunoassay seldom provides information on post-translational modifications on the target protein.
Genetic engineering can be used to introduce a label with high specificity. In this case, the label
is a fluorescent protein, such as green fluorescent protein (GFP). Labeling is performed by fusing
the gene for GFP with the gene of the target protein. Transcription and translation of the target gene
leads to the production of a fusion protein that is labeled with GFP.
GFP is a relatively highly fluorescent protein that can be detected with high sensitivity.29 The
study of fusion proteins by microscopy and flow cytometry is common. However, there are three
issues associated with study of GFP. First, autofluorescence from endogenous components provides
a background fluorescence signal that limits the detection of trace level proteins. Second, GFP
fluorescence reflects the structure of GFP itself, and provides no information on post-translational
modifications to the target protein, such as phosphorylation or glycosylation. Third, GFP is a rel-
atively proteolysis-resistant protein. As a result, the target protein may have undergone proteolytic
destruction while GFP remains intact. Detection of fluorescence from GFP does not necessarily mean
that the target protein is present; it simply means that the target protein had been expressed.
The use of chemical cytometry for the study of GFP fusion proteins deals with these issues.29
Electrophoretic conditions can be manipulated to minimize overlap of GFP and autofluorescent
components, so that GFP can be detected on a very low background, which improves detection
limits. The migration time of the fusion will reflect any post-translational modifications to the protein,
including proteolysis of the target protein.

21.3.1 CHEMICAL CYTOMETRY OF BIOGENIC AMINES AND PROTEINS USING
ONE-DIMENSIONAL CAPILLARY ELECTROPHORESIS
Figure 21.4 presents a one-dimensional capillary sieving electrophoresis (CSE) analysis of the pro-
teins from a human primary T-cell. In this separation, the cell was injected into the capillary, lysed,
and primary amines were labeled with FQ. Components were separated by CSE. The electrophero-
gram consists of a set of ∼25 peaks, some of which are extraordinarily sharp; for example, the
component that migrates at 14.2 min generates a peak with 2.8 million theoretical plates (7.7 million
plates/m), whereas most peaks generate a few hundred thousand theoretical plates.
1.0
0.8
Intensity (MHz)
0.6
6 8 10 12 14 16
0.4
0.2
0.0
6 8 10 12 14 16
FIGURE 21.4 One-dimensional CSE analysis of a single human T-cell. The inset expands the fluorescent
signal to highlight lower amplitude components.
1.5
Intensity (MHz)
1.0
Cell 5
Cell 4
0.5 Cell 3
Cell 2
Cell 1
0.0
6 8 10 12 14 16
FIGURE 21.5 CSE analysis of five human T-cells. Curves are offset for clarity.
120
GFP-Gal4
fusion protein Autofluorescence
110
Intensity (kHz)
100
90
80
0 2 4 6 8 10 12 14 16
FIGURE 21.6 Chemical cytometry of a single yeast cell expressing the GFP–Gal4 fusion protein. The aut-
ofluorescence signal is similar in amplitude to the signal due to the fluorescent protein; without electrophoresis,
accurate quantification of fluorescence due to the fusion protein would be difficult.
Figure 21.5 presents the electropherograms generated from five cells. While some cells, such as
cells 1 and 2, generate similar electropherograms, there is a fairly wide cell-to-cell heterogeneity in
expression patterns. This pattern likely reflects the inherent heterogeneity in this complex primary
cell type.
Krylov and coworkers and this group29,30 have also employed chemical cytometry to character-
ize GFP in single cells (Figure 21.6). This electropherogram shows the analysis of a single yeast
cell that has been engineered to express the Gal4–GFP fusion protein. Figure 21.7 presents the elec-
tropherogram generated from a single cell of D. radiodurans expressing GFP. This organism is a
prokaryote, which is ∼1000-fold smaller in mass than typical mammalian cells.
20
15
Intensity (kHz)
10
0 0.5 1 1.5 2 2.5

FIGURE 21.7 Chemical cytometry of a single D. radiodurans cell expressing GFP. The set of peaks appears
to be associated with proteolytic fragments of GFP.
21.3.2 CHEMICAL CYTOMETRY OF BIOGENIC AMINES AND PROTEINS USING

TWO-DIMENSIONAL CAPILLARY ELECTROPHORESIS
We have developed two-dimensional CE systems for the characterization of proteins and biogenic
amines.27,31,32 The use of this technology for chemical cytometry is similar to the use of one-
dimensional electrophoresis: a cell is aspirated into the column, lysed, and its components labeled
with FQ. For two-dimensional electrophoresis, components are separated based on CSE in the first-
dimension capillary. Fractions are then transferred across an interface to a second capillary, where
they undergo additional separation based on micellar electrokinetic chromatography (MECC) before
detection by fluorescence. The voltage drop across the first capillary is set to zero during the second
dimension separation, holding components stationary. In a typical experiment, ∼300 fractions are
transferred between capillaries under computer control.
Figure 21.8 presents the raw fluorescence signal generated from the two-dimensional elec-
trophoresis analysis of a single RAW 264.7 macrophage cell. The fluorescence signal consists of
a long time series that contains a large number of peaks, and these peaks tend to repeat with 17 s
periods, which was the time between transfers of fractions from the first to the second capillary.
The data can be considered a spiral wrapped around a cylinder (Figure 21.9), where the axial
distance corresponds to the first dimension separation, the angle corresponds to the second dimension
separation, and the density of the image is related to the fluorescence intensity. To generate an image,
the cylinder is slit and flattened, creating a gel image of the separation.
Figure 21.10 presents such an image that resembles a silver-stained two-dimensional gel. The
image consists of a set of spots, distributed across the surface. The positions of these spots are highly
reproducible from cell to cell, with typical precision in spot position being on the order of the size
of the spot itself.
The gel image is convenient for comparing spot position. However, the image, particularly when
overexposed, is less useful in comparing the amplitude of spots or characterizing the dynamic range
of the experiment. Instead, the data can be plotted as a surface in the form of a landscape, where
height is proportional to the fluorescence intensity (Figure 21.11).
These data have a very large dynamic range, approaching 104 in favorable cases. This dynamic
range is suggested by expanding the image by a factor of 10 (Figure 21.12). These separations tend
Intensity
2
11.4 11.6 11.8 12 12.2 12.4
0 10 20 30 40 50 60 70 80 90 100
Time (min)
FIGURE 21.8 Raw data generated from a single macrophage cell. Each fraction transfer from the first to the
second capillary generates a set of peaks as the first capillary’s components are separated in the second capillary.
The inset shows a close up of a typical region.
Second
dim
e
nsi o
n
First dimension separation
FIGURE 21.9 The raw electrophoresis data of Figure 21.8 are wrapped as a spiral around a cylinder, where
the optical density is now related to the fluorescence intensity.
to give excellent peak capacity in the CSE dimension, but poorer capacity in the MECC dimension.
The overall spot capacity of this two-dimensional separation is typically between 250 and 500 spots.
This form of chemical cytometry has the potential of providing a very rich data set with which to
characterize a cellular population.
21.3.3 METABOLIC CYTOMETRY

This research group coined the terms chemical cytometry and metabolic cytometry in 1999.33
Metabolic cytometry is a form of chemical cytometry that monitors biosynthetic and biodegradation
enzymatic cascades in a single cell.
50
CSE fraction number

100
150
200
250
300
350
0 2 4 6 8 10 12 14 16
MECC migration time (s)
FIGURE 21.10 The cylinder is slit and flattened, creating a gel image, where the optical density is related to
fluorescence intensity. The image is overexposed to highlight low-amplitude components.
mi
gr MEC
ati C
on
tim
e CSE fraction number
FIGURE 21.11 Landscape view, where height is proportional to fluorescence intensity.
In metabolic cytometry, cells are incubated with a substrate that is tagged with a highly fluores-
cent dye; we prefer tetramethylrhodamine because of its excellent spectroscopic properties and its
compatibility with the frequency-doubled neodymium YAG laser. This substrate is prepared at high
concentration and undergoes chromatographic purification to eliminate fluorescent impurities.
As long as the dye is preserved during the transformations, any enzymatic transformation can
be monitored with exquisite sensitivity. Figure 21.13 presents an illustrative cartoon, where a sub-
strate, tagged with a fluorescent label (star), undergoes biosynthesis (left path) or biodegradation
(right path).
In our first report, we measured DNA ploidy, the uptake of a fluorescent carbohydrate, and that
of carbohydrate’s biosynthetic and biodegradation products in a single HT29 colon adenocarcinoma
cell.33 We also monitored similar transformations in a single yeast cell.34
More recently, we have been studying glycosphingolipid metabolism in single AtT20 cells.
Glycosphingolipids are among the most common molecules on neural cell surfaces. These
mi
gr MEC
ati C
on
tim
e
CSE fraction number
FIGURE 21.12 The image of Figure 21.11 has been expanded by a factor of 10 to highlight low-amplitude
components.
FIGURE 21.13 Metabolic cytometry. A substrate (top) is fluorescently labeled (star). This substrate can
undergo biosynthesis (left) or biodegradation (right). As long as the fluorescent tag remains intact, the metabolic
products can be monitored by chemical cytometry.
compounds consist of several components. Ceramide is the nonpolar tail that consists of a long-
chain amino alcohol linked to a long-chain fatty acid. A chain of saccharide groups attached to the
ceramide form the constituent glycosphingolipids, and the presence of one to four sialic acids turns
these glycosphingolipids into gangliosides. In the common naming of gangliosides, the subscript
letter indicates the quantity of sialic acid groups: M for monosialo, D for disialo, and so forth; A
refers to asialo that are glycosphingolipids without sialic acids. The number refers to the order of
peak appearance in thin layer chromatography and thus the length of the sugar chain; the sequential
removal of zero, one, and two terminal saccharides increases retention to give GM1 , GM2 , and GM3 ,
respectively.
These glycolipids form a large fraction of neuronal cell membranes, and defects in their
metabolism leads to a series of devastating diseases, the best known of which is Tay-Sachs. The
compounds are synthesized from the lipid ceramide by successive addition of monosaccharides
(Figure 21.14), where the numbers in boxes correspond to the E.C. number for the enzyme responsible
2.4.1.80 2.4.1.– 2.4.99.9 2.4.99.8 Siat8a/e

Cer Glc-Cer Lac-Cer GM3 GD3 GT3
2.4.1.92 2.4.1.92 2.4.1.92 2.4.1.92
GA2 G M2 GD2 G T2
2.4.1.62 2.4.1.62 2.4.1.62 2.4.1.62
GA1 GM1 GD1b GT1c
2.4.1.62 2.4.99.4 2.4.99.4 2.4.99.–
Key:
GM1b GD1a GT1b GQ1c
Ceramide
Glucos e
Siat8e Siat8e Siat8e 2.4.99.–
Galactose
GlcNAc
GalNA c GD1c GT1a GQ1b GP1c
Sialic acid
FIGURE 21.14 Partial metabolic pathway associated with sphingolipids. The number in the boxes corresponds
to the EC name for the biosynthetic enzyme responsible for the transformation.
O
HO OH HN
OH O OH OH
HO O O O HO
O O
NHAc O O
HO OH O OH
O OH HO
HO
HO HO O
HO N
AcHN OH TMR-labeled
compounds
O
O
NH
HN O N
O O
R
O
HO
FIGURE 21.15 Structure of GM1 . A fluorescently labeled compound has been prepared where the ceramide
group (box) is replaced with the TMR-label (bottom box).
for the addition of the corresponding sugar. Biodegradation is not shown; defect in the degradation
of GM2 to GM3 leads to accumulation of the former, and is the underlying cause of Tay-Sachs disease.
We have synthesized the fluorescently labeled GM1 (Figure 21.15).35 In the native compound,
top, ceramide is shown in the box. In our labeled compound, the fatty acid has been replaced with
tetramethylrhodamine.
We have also synthesized the biodegradation products for this compound, including the asialo
GA1 , and GA2 , which are used as standards to tentatively identify metabolic cytometry products
based on comigration.
GM1
GA2 Cer
2.0
GM2
1.6
Signal (MHz)
1.2
GA1 LacCer
GlcCer
0.8 GM3?
0.4
4.5 5 5.5 6
FIGURE 21.16 Metabolic cytometry analysis of three cells. AtT20 cells were incubated with the substrate of
Figure 21.15, aspirated into a capillary, lysed, and analyzed by CE.
Figure 21.16 presents metabolic cytometry data generated from single AtT20 cells. In these
experiments, a cell is incubated with the substrate. The substrate is taken up, which can be confirmed
by fluorescence microscopy. Enzymatic transformations convert the substrate to products. To analyze
the products, the cell is aspirated into a capillary, lysed, components are separated by CE, and products
detected by laser-induced fluorescence.
These data show remarkable cell-to-cell variations in metabolism. The substrate, GM1 , is present
in all cells, and all cells show a modest amount of GA1 , which requires complex transformation of
the substrate. Some cells show very active biodegradation, completely removing the carbohydrates
and leaving only the labeled ceramide. Other cells show much less biodegradation and no detectable
ceramide.

Instrumentation for chemical cytometry consists of three parts. The first is an injection module that
facilitates aspiration of a cell into the capillary. The second is instrumentation to perform one- or
two-dimensional CE. The third is a high sensitivity laser-induced fluorescence detector.
21.4.1 INJECTION BLOCK

Krylov et al.36 designed and constructed an injection block that greatly facilitates single-cell analysis.
This block consists of three Plexiglas pieces (Figure 21.17). The top piece holds both the capillary
and the high-voltage electrode. The second piece contains a hole to connect to pure nitrogen, which
is used to pressurize the chamber in order to flush buffer through the capillary. The bottom piece
holds a microcentrifuge vial that contains the running buffer.
This injection block is attached to micromanipulators. In operation, the capillary is first flushed
with buffer to prepare it for separation. The bottom piece is then removed, and micromanipulators
are used to center the capillary over a cell of interest. The cell is aspirated into the capillary, the vial
holder is replaced, and high voltage is applied to drive the separation.
High
voltage
Capillary
electrode
Purge
vent
Vial
holder
FIGURE 21.17 Injection block. A capillary and electrode are threaded through the top two blocks. The third
block forms a gas-tight seal and is used to pressurize a reservoir that pumps reagents through the capillary. The
bottom block is removed for injection of a cell.
Aspiration of a single cell typically requires a pulse of ∼11 kPa vacuum to the distal end of
the capillary for 1 or 2 s. Control of such modest pressure is not trivial. Krylov36 developed an
elegant method for generating an extremely reproducible vacuum. The distal end of the capillary is
connected to a sheath-flow cuvette, and the waste stream of the cuvette is connected to a three-way
valve. In normal operation, the valve directs the waste to a receiving reservoir. To inject a sample,
the valve directs flow to a solenoid valve, which is connected to a water-filled piece of tubing that
terminates in a beaker placed 1 m below the level of the injection block. A timer circuit is used to
open the solenoid valve for a precisely timed period, during which time the injection end of the
capillary is connected to a water column of 1 m high, which applies vacuum necessary to aspirate
the cell within the capillary.

Both one- and two-dimensional CE have been used for chemical cytometry. One-dimensional elec-
trophoresis is similar to conventional experiments, albeit with the issues of sample loading, lysis,
and on-column labeling.
Two-dimensional CE instrumentation is more complicated, with two power supplies, one for each
capillary (Figure 21.18).37,38 The sample undergoes separation by CSE. Fractions are transferred
between capillaries by manipulation of the voltages applied across the two capillaries.
The interface is manufactured using standard macromachining technology. Two capillaries are
aligned to be coaxial with a ∼50-µm gap in a buffer-filled chamber. Analyte is transferred between
capillaries by creating a voltage drop across the interface. Figure 21.19 presents a micrograph of
fluorescein being transferred between capillaries in an interface. The fluorescence is centered on the
two capillaries. A stagnant zone appears to form an annulus about 10 µm in radius about the inner
lumen of the first dimension capillary. However, there is no evidence that analyte leaks from the
interface or is lost during a transfer.
Capillary 1 Capillary 2
Sheath flow
cuvette
Power Power
supply 1 supply 2
Running Running Waste

buffer 1 buffer 2
FIGURE 21.18 Instrumentation for two-dimensional CE. Two capillaries are joined through an interface, that
allows fractions to be transferred by manipulation of the voltage applied to the capillaries. In our system, the
sheath-flow cuvette detector is held at ground potential.
FIGURE 21.19 Transfer of fluorescein between capillaries in the interface used in two-dimensional CE.
This interface differs from Jorgenson and coworkers39 flow-gated interface between a liquid
chromatographic column and a CE column. In that design, analyte continually flows from the
chromatographic column; cross-flow is used to flush analyte to waste except during fraction transfer.
In contrast, analyte flow is easily stopped in two-dimensional CE by applying equal voltage at
both ends of the capillary; as a result, there is no need for crossed flow with its accompanying
sample loss.
A typical voltage program is shown in Figure 21.20. After lysis, a preliminary CSE separation
is performed so that the fastest moving components approach the exit of the first capillary. Then, a
series of fraction transfers and MECC separations is performed. There is a voltage drop across the
first capillary during the transfer steps, which continues the CSE separation and which drives analyte
to the second dimension capillary. There is also a voltage drop across the second dimension capillary
during the transfer, so that separation also continues in that dimension during fraction transfer. There
is no voltage drop across the first dimension capillary during the MECC separation step, and analyte
does not migrate in that capillary.
Power supply 1
20 kV Transfer Transfer
and CSE and CSE
separation separation
(1 s) (1 s)
Power supply 2
10 kV
CSE MECC MECC MECC
preseparation separation separation separation
(120 s) (17 s) (17 s) (17 s)
Time
FIGURE 21.20 Typical timing diagram used in two-dimensional CE for chemical cytometry. The distal end
of capillary 2 is held in the sheath-flow cuvette, which is at ground potential.
21.4.3 FLUORESCENCE DETECTION

High-sensitivity fluorescence detection is required for chemical cytometry. We use a postcolumn
sheath-flow cuvette for detection. This device is based on a cuvette used in the venerable Ortho
cytofluorograph, a flow cytometer developed in the late 1970s. A stream of flowing buffer ensheaths
analyte flowing from the separation capillary. Fluorescence is detected in a flow chamber with high
optical quality windows, which decreases light scatter and improves the detection limit compared to
using the capillary itself as a detection cuvette.
21.4.4 DATA PROCESSING

There are a couple useful steps in data processing. First, a morphological opening procedure can be
used to remove baseline drift. This process is available in the image processing toolbox in Matlab.
Like many other filtering processes, baseline correction is subject to artifact. These artifacts are most
obvious when performing the morphological opening procedure on too small an area; in this case,
wide peaks and ridges in the data can disappear.
Second, filtering is useful. A number of filters are available for two-dimensional data. The most
important is the median filter that effectively removes noise spikes caused by bubbles or particles
passing through the laser beam. Next, it is often useful to smooth the data; convolution with a two-
dimensional Gaussian filter is particularly useful if the shape of the Gaussian filter is matched to
those of typical peaks.
21.5 CONCLUDING REMARKS AND FUTURE WORK

The field of chemical cytometry is in its infancy, and it is clear that there is much work to be
done before the technology is widely used. First, instrumentation throughput must be increased.
Flow cytometry today can process a 100,000 cells/s. While chemical cytometry will not produce
similar throughput, it is reasonable to expect the technology to process 10,000 cells/day in one-
dimensional electrophoresis and perhaps 1000 cells/day in two-dimensional electrophoresis. Such
instrumentation will likely resemble the multiple-CE systems that have become ubiquitous in DNA
sequencing.37,38,40
Second, separation efficiency must be improved. Current two-dimensional electrophoresis sys-
tems are generating a spot capacity of a few hundred components; an order of magnitude improvement
is desirable. Spot capacity is given by the product of the peak capacity in each dimension, so that a
threefold improvement in separation efficiency in each dimension will generate the desired improve-
ment. MECC is not an ideal separation mode for proteins, and would be replaced with a technique
such as isoelectric focusing, which can provide exquisite resolution. A major hurdle deals with pro-
tein labeling. FQ, and presumably other dyes that produce an anionic or neutral reaction product,
generate very complex isoelectric focusing electropherograms due to multiple labeling. Anionic
surfactants are not compatible with IEF, and it will be necessary to employ a labeling reagent that
produces a cationic product. The recently developed chameleon dyes are interesting examples of
fluorogenic reagents that produce cationic products.41,42
Third, component identification must be improved. Current methods rely on comigration with
standards. While practical for metabolic cytometry and biogenic amines, due to the relatively small
number of possible compounds, spiking is extremely tedious for protein analysis. Ideally, an online
mass spectrometer can be used to monitor compounds as they migrate from the capillary. Although
the current generation of mass spectrometers does not have sufficient sensitivity to monitor the
minute amounts of compounds present in a single cells, mass spectrometry could be used to study
concentrated homogenates prepared from large numbers of cells. Based on excellent run-to-run
reproducibility of the two-dimensional electrophoresis system, component identification should be
practical in fluorescence-based detection of single cells. As a complication, it would be necessary
to use fluorescently labeled proteins, which would need to be considered during interpretation of
mass spectra.
ACKNOWLEDGMENTS
This work was funded by grants from National Institutes of Health (NIH) and Department of
Energy (DOE).
REFERENCES
1. Shapiro, H.M. Practical Flow Cytometry, 4th ed., Hoboken, NJ, Wiley, 2003.
2. Caspersson, T.O. Cell Growth and Cell Function, New York, Norton, 1950.
3. Caspersson, T.O. History of the development of cytophotometry from 1935 to the present. Anal. Quant.
Cytol. Histol. 1987; 9, 2–6.
4. Ibrahim, S.F. and van den Engh, G. High-speed cell sorting: Fundamentals and recent advances. Curr.
Opin. Biotech. 2003; 14, 5.
5. Baumgarth, N. and Roederer, M. A practical approach to multicolor flow cytometry for immunophe-
notyping. J. Immunol. Methods 2000; 243, 77–97.
6. Xu, B.J., Caprioli, R.M., Sanders, M.E. and Jensen, R.A. Direct analysis of laser capture microdissected
cells by MALDI mass spectrometry. J. Am. Soc. Mass Spectrom. 2002; 13: 1292–1297.
7. Laurent, C., Levinson, D.F., Schwartz, S.A., Harrington, P.B., Markey, S.P., Caprioli, R.M. and
Levitt, P. Direct profiling of the cerebellum by matrix-assisted laser desorption/ionization time-of-
flight mass spectrometry: A methodological study in postnatal and adult mouse. J. Neurosci. Res.
2005; 81: 613–621.
8. Whittal, R.M., Keller, B.O. and Li, L. Nanoliter chemistry combined with mass spectrometry for
peptide mapping of proteins from single mammalian cell lysates. Anal. Chem. 1998; 70: 5344–5347.
9. Li, L., Garden, R.W. and Sweedler, J.V. Single-cell MALDI: A new tool for direct peptide profiling.
Trends Biotechnol. 2000; 18: 151–160.
10. Edstrom, J.E. Nucleotide analysis on the cyto-scale. Nature 1953; 172: 809.
11. Kennedy, R.T., Oates, M.D., Cooper, B.R., Nickerson, B. and Jorgenson, J.W. Microcolumn
separations and the analysis of single cells. Science 1989; 246: 57–63.
12. Wallingford, R. A. and Ewing, A. G. Capillary zone electrophoresis with electrochemical detection in
12.7 microns diameter columns. Anal. Chem. 1988; 60: 1972–1975.
13. Hogan, B.L. and Yeung, E.S. Determination of intracellular species at the level of a single erythrocyte
via capillary electrophoresis with direct and indirect fluorescence detection. Anal. Chem. 1992; 64:
2841–2845.
14. Gilman, S.D. and Ewing, A.G. Analysis of single cells by capillary electrophoresis with on-column
derivatization and laser-induced fluorescence detection. Anal. Chem. 1995; 67: 58–64.
15. Meredith, G.D., Sims, C.E., Soughayer, J.S. and Allbritton, N.L. Measurement of kinase activation in
single mammalian cells. Nat. Biotechnol. 2000; 18: 309–312.
16. Han, F. and Lillard, S.J. In-situ sampling and separation of RNA from individual mammalian cells.
Anal. Chem. 2000; 72: 4073–4079.
17. Wu, H., Wheeler, A. and Zare, R. N. Chemical cytometry on a picoliter-scale integrated microfluidic
chip. Proc. Natl Acad. Sci. USA 2004; 101: 12809–12813.
18. Cheng, Y.F. and Dovichi, N.J. Subattomole amino acid analysis by capillary zone electrophoresis and
laser-induced fluorescence. Science 1988; 242: 562–564
19. Chen, D.Y. and Dovichi, N.J. Single-molecule detection in capillary electrophoresis: Molecular shot
noise as a fundamental limit to chemical analysis. Anal. Chem. 1996; 68: 690–696.
20. Ghaemmaghami, S., Huh, W.K., Bower, K., Howson, R.W., Belle, A., Dephoure, N., O’Shea, E.K.
and Weissman, J.S. Global analysis of protein expression in yeast. Nature 2003; 425: 737–741.
21. Available at: http://yeastgtp.ucsf.edu/allOrfData.txt
22. Liu, H.J., Cho, B.Y., Krull, I.S. and Cohen, S.A. Homogeneous fluorescent derivatization of large
proteins. J. Chromatogr. A 2001; 927: 77–89.
23. Zhao, J.Y., Waldron, K.C., Miller, J., Zhang, J.Z., Harke, H. and Dovichi, N.J. Attachment of a single
fluorescent label to peptides for determination by capillary zone electrophoresis. J. Chromatogr. 1992;
608: 239–242.
24. Craig, D.B. and Dovichi, N.J. Multiple labeling of proteins. Anal. Chem. 1998; 70, 2493–2494.
25. Richards, D., Stathakis, C., Polakowski, R., Ahmadzadeh, H. and Dovichi, N.J. Labeling effects on
the isoelectric point of green fluorescent protein. J. Chromatogr. A 1999; 853: 21–25.
26. Pinto, D.M., Arriaga, E.A., Craig, D., Angelova, J., Sharma, N., Ahmadzadeh, H., Dovichi, N.J. and
Boulet, C.A. Picomolar assay of native proteins by capillary electrophoresis—Precolumn labeling,
sub-micellar separation and laser induced fluorescence detection. Anal. Chem. 1997; 69: 3015–3021.
27. Kraly, J.R., Jones, M.R., Gomez, D.G., Dickerson, J.A., Harwood, M.M., Eggertson, M., Paulson, T.G.,
et al. Rapid and reproducible two-dimensional capillary electrophoresis analysis of Barrett’s esophagus
tissues. Anal. Chem. 2006; 78: 5977–5986.
28. Gudiksen, K.L., Gitlin, I., Moustakas, D.T. and Whitesides, G.M. Increasing the net charge and
decreasing the hydrophobicity of bovine carbonic anhydrase decreases the rate of denaturation with
sodium dodecyl sulfate. Biophys. J. 2006; 91: 298–310.
29. Turner, E.H., Lauterbach, K., Pugsley, H., Palmer, V.R. and Dovichi, N.J. Ultrasensitive green
fluorescent protein detection in a single bacterium by capillary electrophoresis with laser-induced
fluorescence, Anal. Chem. 2007; 79: 778–781.
30. Hu, K., Ahmadzadeh, H. and Krylov, S.N. Asymmetry between sister cells in a cancer cell line revealed
by chemical cytometry. Anal. Chem. 2004; 76: 3864–3866.
31. Michels, D.A., Hu, S., Schoenherr, R.M., Eggertson, M.J. and Dovichi, N.J. Fully automated two-
dimensional capillary electrophoresis for high sensitivity protein analysis. Mol. Cell. Proteomics 2002;
1: 69–74.
32. Michels, D.A., Hu, S., Dambrowitz, K.A., Eggertson, M.J., Lauterbach, K. and Dovichi, N.J. Capil-
lary sieving electrophoresis-micellar electrokinetic chromatography fully automated two-dimensional
capillary electrophoresis analysis of Deinococcus radiodurans protein homogenate. Electrophoresis
2004; 25: 3098–3105.
33. Krylov, S.N., Zhang, Z., Chan, N.W.C., Arriaga, E., Palcic, M.M. and Dovichi, N.J. Correlating cell
cycle with metabolism in single cells: The combination of image and metabolic cytometry. Cytometry,
1999; 37: 15–20.
34. Le, X.C., Tan, W., Scaman, C.H., Szpacenko, A., Arriaga, E., Zhang, Y., Dovichi, N.J., Hinds-
gaul, O. and Palcic, M.M. Single cell studies of enzymatic hydrolysis of a tetramethylrhodamine
labeled trisaccharide in yeast. Glycobiology 1999; 9: 219–225.
35. Larsson, E.A., Olsson, U., Whitmore, C.D., Martins, R., Tettamanti, G., Schnaar, R.L., Dovichi, N.J.,
Palcic, M.M. and Hindsgaul, O. Synthesis of reference standards to enable single cell metabolomic
studies of tetramethylrhodamine-labeled ganglioside GM1. Carbohydr. Res. 2007; 342: 482–489.
36. Krylov, S.N., Starke, D.A., Arriaga, E.A., Zhang, Z., Chan, N.W., Palcic, M.M. and Dovichi, N.J.
Instrumentation for chemical cytometry. Anal. Chem. 2000; 72: 872–877.
37. Zhang, J., Voss, K.O., Shaw, D.F., Roos, K.P., Lewis, D.F., Yan, J., Jiang, R., et al. A multiple-
capillary electrophoresis system for small-scale DNA sequencing and analysis. Nucleic Acids Res.
1999; 27: e36.
38. Crabtree, H.J., Bay, S.J., Lewis, D.F., Zhang, J., Coulson, L.D., Fitzpatrick, G.A., Delinger, S.L.,
Harrison, D.J. and Dovichi, N.J. Construction and evaluation of a capillary array DNA sequencer
based on a micromachined sheath-flow cuvette. Electrophoresis 2000; 21: 1329–1335.
39. Opiteck, G.J., Lewis, K.C., Jorgenson, J.W. and Anderegg, R.J. Comprehensive on-line LC/LC/MS of
proteins. Anal. Chem. 1997; 69: 1518–1524.
40. Zhang, J., Yang, M., Puyang, X., Fang, Y., Cook, L.M. and Dovichi, N.J. Two-dimensional direct-
reading fluorescence spectrograph for DNA sequencing by capillary array electrophoresis. Anal. Chem.
2001; 73: 1234–1239.
41. Craig, D.B., Wetzl, B.K., Duerkop, A. and Wolfbeis, O.S. Determination of picomolar concentrations
of proteins using novel amino reactive chameleon labels and capillary electrophoresis laser-induced
fluorescence detection. Electrophoresis 2005; 26: 2208–2213.
42. Wetzl, B.K., Yarmoluk, S.M., Craig, D.B. and Wolfbeis, O.S. Chameleon labels for staining and
quantifying proteins. Angew. Chem. Int. Ed. Engl. 2004; 43: 5400–5402.
22 Glycoprotein Analysis by
Michel Girard, Izaskun Lacunza, Jose Carlos Diez-Masa,
and Mercedes de Frutos
CONTENTS
22.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 631

22.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 632
22.4.1 Sample Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 638
22.4.2 Separation of Glycoforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 641
22.4.2.1 Solutions to Protein Adsorption Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 641
22.4.2.2 Factors Affecting the Resolution of Glycoform Peaks . . . . . . . . . . . . . . . . . . 644
22.4.2.3 Reproducibility Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 655
22.4.3 Detection and Identification of Forms of Glycoproteins . . . . . . . . . . . . . . . . . . . . . . . . . . 656
22.4.3.1 UV Detection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 656
22.4.3.2 Laser-Induced Fluorescence Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 657
22.4.3.3 Mass Spectrometry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 659
22.4.3.4 Other Detection Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 662
22.5.1 Characterization and Quality Assessment of Biologicals and Biopharmaceuticals 662
22.5.1.1 General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 662
22.5.1.2 Product Characterization and Identity Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . 665
22.5.1.3 In-Process Monitoring and Product Consistency . . . . . . . . . . . . . . . . . . . . . . . . 671
22.5.1.4 Product Comparability and Analysis of Finished Products . . . . . . . . . . . . . 673
22.5.1.5 Determination of Biological Activity/Potency . . . . . . . . . . . . . . . . . . . . . . . . . . 677
22.5.2 CE of Isoforms of Intact Glycoproteins in the Clinical Field . . . . . . . . . . . . . . . . . . . . . 678
22.5.2.1 Transferrin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 679
22.5.2.2 Alpha-1-Acid Glycoprotein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 686
22.5.2.3 Other Glycoproteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 687
22.6 Concluding Remarks and Future Prospects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 687
Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 694
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 696
22.1 INTRODUCTION
Under the action of enzymes globally called glycosyltransferases, proteins undergo cotranslational
and post-translational modifications (PTMs) resulting in the covalent attachment of glycan chains to
the polypeptide core. This fundamental process is called glycosylation and results in the formation of
glycoproteins. Although many others exist, glycosylation is undoubtedly one of the most prevalent
631
PTM of eukaryotic proteins. For instance, in humans, glycoproteins account for more than 50%
of all proteins. Despite this high frequency of occurrence, no generalized role can be attributed to
glycosylation of proteins. In fact, the role of the glycan moiety can vary from one protein to another
and have direct involvement in the biological properties of proteins such as transport, cell surface
recognition, enzymatic processes, or the immune system. The large size of oligosaccharides may
allow them to cover functionally important regions of proteins, to modulate the interactions of glyco-
conjugates with other molecules, and to affect the rate of the processes that involve conformational
changes. Besides these important biological functions, glycosylation also confers physical properties
to proteins such as solubility, hydrophilicity, stability, and assists in protein folding.
Just as for the rest of proteins, other PTM events and the existence of genetic variants contribute
to the heterogeneity of glycoproteins. All of the heterogeneities are generically termed “forms” of
a protein (e.g., deamidated forms), while forms arising from the glycosidic part of the protein are
specifically called “glycoforms.” The term “sialoform” refers to the glycoforms that have the same
number of sialic acid residues. In the field of capillary electrophoresis (CE), “isoform” indicates an
electrophoretic peak or band, which can include one or more forms of a protein.
From an analytical point of view, there are several approaches to the study of glycoproteins.
The protein can be submitted to partial or total hydrolysis/proteolysis resulting in the formation of
glycopeptides, glycans, individual monosaccharides, peptides, or amino acids, which can then be
analyzed separately. Alternatively, the different forms of the intact glycoprotein can be separated
and studied using one of many suitable techniques providing high resolution of closely related pro-
teins such as sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS–PAGE), isoelectric
focusing (IEF), high-performance liquid chromatography (HPLC), mass spectrometry (MS), or CE.
Traditionally, conventional gel electrophoresis methods have been employed to study the intact forms
of glycoproteins and their hydrolysis/proteolysis products. However, these methods are laborious to
carry out, cannot be automated, and do not consistently provide quantitative results. In contrast, CE
analysis has been demonstrated to provide rapid, automated, and quantitative results. In addition, it is
now often used as a complementary technique to the more established HPLC in several glycoprotein
studies.
In this chapter, CE approaches to the study of intact forms of glycoproteins will be discussed with
particular emphasis on factors such as separation conditions and modes of detection. In addition, an
overview of practical applications of CE in different sectors of activities such as the pharmaceutical
industry and clinical research will be presented. Although examples illustrating various aspects will
be mentioned, the primary goal of this chapter is not to provide an extensive review of the literature
reported on the subject or details about all of the glycoproteins that have been studied. For in-
depth discussions several reviews covering this subject in part or in total, which are mentioned in
Section 22.2 of this chapter, can be read.
The general principles related to the different modes of CE and of protein separation by CE have
not been considered here as they are the subjects of other chapters in this book.
22.2 BACKGROUND
Glycoprotein formation is a complex process that involves multiple enzymatic reactions occurring
in the endoplasmic reticulum and the Golgi apparatus.
A glycoprotein is usually present in an organism as a population of glycoforms, the heterogene-
ity of which is cell/tissue or organism-dependent, as well as a function of the pathophysiological
status [1]. Changes in protein glycosylation occur in both congenital and acquired illnesses [2–4]
and the study of these changes is generating considerable interest in the field of biomarkers. The
production by recombinant DNA technology of glycoproteins for specific therapeutic uses is another
field that has attracted considerable research efforts. In fact, many of the biopharmaceuticals cur-
rently produced are glycoproteins [5,6]. The glycosylation of these proteins is crucial in their role
Glycoprotein Analysis by Capillary Electrophoresis 633
OH ~~ OH
(H,CH3)
~~
HO O HO O (H,CH3)
H H
H HO O H
+
OH H OH H
HN O O
H OH H O
HN O
H NHAc ~~ H NHAc
~~
α-GalNAc ~Ser / Thr ~ O-Linked glycoprotein
OH OH
~~ ~~
H O H O
OH HN2 NH
H H
OH H O OH H O
+
HO H O HN O HO H O HN O
H NHAc H NHAc
~~ ~~
β-GlcNAc ~Asn ~ N-Linked glycoprotein
FIGURE 22.1 Representation of O- and N-linked glycosylation from the reaction of α-N-acetylgalactosamine
with serine/threonine and β-N-acetylglucosamine with asparagine, respectively.
as biopharmaceuticals since it will affect, for instance, their biological activities, solubility, and
other properties [7,8]. In the case of recombinant glycoproteins, in addition to being cell-dependent,
glycosylation is affected by culture media conditions and the purification process [1]. These con-
siderations give an idea of the enormous variety of glycoforms that a glycoprotein can contain and,
therefore, the difficult task that the in-depth analysis of all of these structures involves, particularly
when the approach selected is the study of intact glycoprotein.
Glycans are attached to proteins through O- and N-linkages at specific amino acids (Figure 22.1).
O-linked glycosylation occurs at the hydroxyl functional group of the side chain of serine (Ser)
or threonine (Thr) residues and, to a lesser extent, as a secondary modification at hydroxylysine
(Hyl) and hydroxyproline (Hyp) residues (e.g., in collagen). N-linked glycosylation takes place at
the amide side chain of asparagine (Asn) residues. There are several monosaccharides involved in
the glycosylation of proteins: mannose (Man), galactose (Gal), fucose (Fuc), N-acetylglucosamine
(GlcNAc), N-acetylgalactosamine (GalNAc), and sialic acid. The latter is the generic term for
several derivatives of neuraminic acid that includes N-acetylneuraminic acid (NeuAc), one of the
most abundant sialic acids in mammals. All of these sugars are neutral at physiological pH except
for sialic acid, which is negatively charged due to the presence of a carboxylic acid functional group.
N-glycans are usually classified into the following three types: complex, high-mannose, and
hybrid [9]. Complex-type glycans have several antennae, with biantennary being the most abundant
ones, although tri- and tetra-antennary ones are also common (Figure 22.2a) [10], each of which may
feature high heterogeneity due to several possible permutations of the sugar residues. High-mannose-
type glycans are common in cell-surface proteins and have, in most cases, between 5 and 9 mannose
residues (Figure 22.2b). Hybrid glycans feature characteristics common to both high-mannose and
complex glycans (Figure 22.2c).
O-glycans are also composed of a variety of common monosaccharides that, in addition to
those found in N-glycans, include xylose (Xyl), glucose (Glu), and arabinose (Ara) (Figure 22.2d).
O-glycans are very common in mucines where long polypeptides with highly negatively charged
zones are present due to the presence of sialic acids.
The complexity of glycoproteins and the fact that CE compares favorably with conventional
electrophoretic techniques are among the reasons for which glycoprotein analysis has been a pre-
eminent feature of CE from its very early days. A monograph dedicated to CE of proteins and
peptides by Strege and Lagu [11] is available. Reviews covering the early development of analytical
CE methods for proteins, including glycoproteins, for years up to 1997 have been cited by Prichett
and Robey [12] in the previous edition of this handbook. Several recent reviews on the analysis of
glycoproteins by CE have discussed various aspects [13–26].

Several CE modes that will be considered in detail in Section 22.4 of this chapter have been used
in the separation of glycoproteins: capillary zone electrophoresis (CZE), capillary isoelectric focus-
ing (CIEF), capillary electrophoresis in SDS (CE-SDS), micellar electrokinetic chromatography
(MEKC), affinity capillary electrophoresis (ACE), and capillary isotachophoresis (CITP). It is out
(a) 1-4
1-6
1-6
1-4
1-2
2-6 1-4 1-4 1-
Basic tetraantennary
0-4 1-4
1-4 1-6
1-3
1-4 1-2 0-1
1-4
1-6
1-6
1-4
1-2
2-6 1-4 1-4 1-4
Basic triantennary
0-3
1-6
1-3
1-4 1-2
0-1
1-4 1-2
1-6
2-6 1-4 1-4 1-4

Basic biantennary
0-2
1-6
1-3
1-4 1-2
0-1
FIGURE 22.2 Typical structures of (a) complex-type N-glycans; (b) high-mannose N-glycans; (c) hybrid-
type N-glycans; and (d) core structures of O-glycans. Representation according to Nomenclature Committee of
the Consortium for Functional Glycomics (http://www.glycomics.scripps.edu/CFGnomenclature.pdf, accessed
on November 2006).
(b) 1-2
1-6
1-6
1-2 1-3
High-mannose (M9)
1-4 1-4 1-4
1-3
1-2 1-2
1-6
1-2 1-6
High-mannose (M6–M8) 1-3
1-4 1-4 1-4
0-2
1-3
1-2
1-6
1-6
1-3
High-mannose (M5) 1-4 1-4 1-4
1-3
(c)
1-6
1-3 1-6
Hybrid
1-4 1-4 1-4 1-4
2-6 1-4
1-4
1-3
0-1
1-2
1-2 1-2
1-6
Hybrid 1-4 1-4 1-4
1-6
1-3
1-2
0-1

(d)
1-6
1–3 1-
Core 1 1-
Core 2
1-3
1-6
1–3 1-
Core 3
Core 4 1-
1-3
1- 1-
Sialic acid
Core 5 1–3 Core 6 1-3
Mannose
Galactose
Galactosamine-NAc
Glucosamine-NAc
1–6 1- 1-3 1-
Core 7 Core 8 Fucose
of the scope of this chapter to review the theoretical aspects of each of these techniques for glyco-
proteins. In this section, only a few considerations about the separation of proteins with a special
emphasis on glycoforms separation by CZE will be presented.
In CZE, proteins can be considered from a theoretical point of view as charged particles. The
mobility, µ, can be described in a first approximation by [27]
q
µ= , (22.1)
f
where q is the charge of the particle and f is its friction coefficient in the separation buffer, which for
small molecules is proportional to their respective mass. For proteins, the friction coefficient is also
related to the shape of the particle. In fact, studies [12,28] have shown that the separation of proteins
in CZE is better explained in terms of hydrodynamic radius, because this parameter describes both the
mass and shape of the protein. In a similar way, studies have shown that the displacement of a protein
in a given separation buffer is better described by the net charge of the protein and, consequently, q
is better expressed as the net charge of the protein (for a review see Reference 29). The net charge
includes contributions from all of the electrostatic elements of the protein, such as charged amino
acids, metal ions, and cofactors bound to it. The net charge is screened to a certain degree by the
microenvironment of the charged group, by the association of the protein with the counterions in
solution, and by the characteristics of the protein that influence its attraction to counterions. In
conclusion, electrophoretic mobility of proteins can be interpreted in terms of their hydrodynamic
radius and net charge [12].
In CZE of glycoproteins the situation is more complicated than for nonglycosylated ones because
of additional differences originating from the glycosidic part. For the sake of simplicity, the following
will present only a limited discussion of the overall implications of the charge and composition of
the glycans on the migration of glycoproteins.
The simplest cases are those in which glycoforms differ only in electric charge. These differences
are very often due to the different content in sialic acid of the glycoforms (sialoforms). Sialic acid
moieties occur in most cases in the outer part of the protein, in a relatively free position to move
around the N- or O-linkage with the rest of the oligosaccharide. Therefore, differences in sialic
acid content of these glycoforms produce important variations in the net charge of the proteins
resulting in differences in their mobilities. One well-documented example of sialoforms separation
by CZE is presented by Balaguer et al. [30]. Using neutral coated capillaries, the authors carried
out the separation of recombinant human erythropoietin (rhEPO) isoforms (Figure 22.3).The main
peaks were identified using electrospray ionization (ESI) time-of-flight mass spectrometry (ESI-
TOF MS) coupled to CZE as several sialoforms (Figure 22.3a). Using the reconstructed extracted
ion electropherograms together with the deconvoluted mass of the MS isoform peaks, they observed
that, as predicted by the theory, only glycoforms differing in their sialic acid content migrated as
separate peaks and that their mobilities were directly related to the number of sialic acids present in
each sialoform.
More complex situations arise when different glycoforms have the same charge but the hydro-
dynamic radius increases with the size of the glycan chain. In the same study mentioned earlier
[30], the authors obtained the extracted ion electropherogram of rhEPO glycoforms containing the
same number of sialic acids but differing in the number of hexose and N-acetylhexosamine residues
(Figure 22.3b). An increase in the effective hydrodynamic radius of the protein causing a mobility
shift to lower migration times was observed.
Intens
13 SiA
× 104
21Hex19HexNAc
(a) LN coating (b) LN coating

12 SiA
13 SiA
1.25 13 SiA
23Hex20HexNAc
21Hex18HexNAc
25Hex22HexNAc
24Hex28HexNAc
1.00
11 SiA
14 SiA
26Hex23HexNAc
0.75
0.50
10 SiA
9 SiA
0.25
8 SiA
BPE
0.00
20 22 24 26 28 20 22 24 26 28 Time [min]
A Sialoforms B HexHexNAc isoforms
14 SiA. EIE 1776.2; 1887.2; 2012.9; 2156.7 (M:30178.6) 21Hex18HexNAc: EIE 1737.7; 1846.2; 1969.2; 2109.8 (M: 29522.5)
9 SiA. EIE 1662.6; 1773.6; 1999.2; (M:26588.6)
FIGURE 22.3 Electropherogram of rhEPO (BRP European Pharmacopeia, 2.5 µg/µL in water). Separation
performed in a capillary 80 cm × 50 µm coated with a neutral polymer (UltraTrol Pre-Coated LN), separation
buffer 2 M HAc, and separation voltage at +30 kV. (a) Clear trace represents the base–peak electropherogram
(BPE) and darker trace represents the electropherogram reconstructed from the extracted ions together with
the deconvoluted mass of the MS peaks of different sialic acid isoforms. (b) Extracted ion electropherogram
obtained for glycoforms with different HexHexNAc content and identical sialic acid number. (From Balaguer, E.
et al., Electrophoresis, 27, 2638, 2006. With permission.)
In some cases, however, a clear effect of either the charge or the size on the glycoform mobility
cannot be discerned. This is illustrated in the work of Yim et al. [31] on the study of recombinant
human bone morphogenic protein 2 (rhBMP-2). This basic protein (pI > 8.5) is a dimer containing
in each of the monomer an N-linked glycan of the high-mannose type. Using neutral hydrophilic
capillaries and phosphate buffer at pH 2.5 as background electrolyte (BGE), the 15 possible rhBMP-2
glycoforms were separated into nine peaks. The use of the glycan and peptide maps, together with fast
atom bombardment mass spectrum of the generated glycopeptides, enabled the authors to identify
these peaks as the nine glycoforms of the structure (rhBP-2)2 -(GlcNAc)4 -(Manz ), where z varies
from 10 to 18. As all of the glycoforms have the same charge, the authors concluded that the size of
a mannose residue was fundamental for the separation and that the increasing hydrodynamic radius
of the protein could control the separation. However, in rhBMP-2, the mannose glycans appear to be
large enough to cover part of the polypeptide backbone, a situation that would lead to the shielding
of the most charged region of the protein and to a reduction of its effective charge. An estimation of
the effect of the mass of a mannose on the mobility of rhBMP-2 in terms of differences in charge
and frictional coefficient led to the conclusion that it was not possible to assess whether the effect
of the mannose residue was from a reduction of the z potential, from an increase of the frictional
coefficient, or from a combination of the two.
From the above discussion, it can be concluded that both the effective charge and the hydro-
dynamic radius play a major role in the control of the electrophoretic mobility in the separation of
proteins by CZE. Since proteins are flexible molecules, the interplay of both factors has to be taken
into consideration to explain their mobility in electrophoresis. For glycoproteins, a full understanding
of the factors that control their separation in CZE is a complex task because the size and flexibility
of the glycans constitute an additional cause of structural variability.

22.4.1 SAMPLE PREPARATION
As in any other analytical procedure, sample preparation, in those cases where it is needed, is a
key step in the separation of glycoforms by CE. The main goals of the sample preparation step are
concentration and purification and, when designing the protocol, the peculiarities of each analyte and
of the sample matrix must be taken into account. For protein separations by CE, in addition to aspects
common to other types of analysis such as filtration, it is often necessary to eliminate compounds
that may interfere either by causing a high electrical current, by adsorbing to the capillary walls, or
by comigrating with the analytes.
Without sample preparation, the analysis of a dilute serum sample by CE with ultraviolet (UV)
detection leads to the separation and detection of the major serum proteins, γ-globulin, transferrin
(Tf), β-lipoproteins, haptoglobin, α2 -macroglobulin, α1 -antitrypsin (AAT), α1 -lipoproteins, human
serum albumin (HSA), pre-albumin, and complements [32], as shown in Figure 22.4. These proteins
often mask other important analytes that are usually more clinically relevant than the major proteins.
In fact, it has been widely mentioned that global approaches to proteome analysis detect less than
0.1% of the protein species present in a sample, which span a range of 10 orders of magnitude
between the most abundant and the less abundant proteins [33].
The following discussion will focus on an example of a glycoprotein to which several sample
preparation procedures have been reported, that is, erythropoietin (EPO) to which a lot of attention
has been devoted for its analysis by CE.
The elimination of low molecular weight (MW) compounds from a sample is widely carried
out by using microcentrifugal devices provided with membranes of different MW cutoffs. Clean-up
of the rhEPO Biological Reference Preparation (BRP) of the European Pharmacopoeia (EP) was
carried out by passing an aqueous solution through a microconcentrator cartridge with a 10 kDa
30
25
Miliabsorbance (200 nm)
20
9
15
10
5
2 3 5
4
2' 6 7
1 8
10
0
100 150 200 250
Time (s)
FIGURE 22.4 Electropherogram of a diluted normal control serum. Peaks: 1, neutral marker; 2, γ-globulin;
3, transferrin; 4, β-lipoproteins; 5, haptoglobin; 6, α2 -macroglobulin; 7, α1 -antitrypsin; 8, α1 -lipoproteins;
9, albumin; 10, prealbumin; 2 , complements. Analytical conditions: fused-silica capillary 25 cm × 25 µm i.d.;
buffer: Beckman proprietary buffer of pH 10.0; 10 kV; 22◦ C. (From Chen, F.-T.A., J. Chromatogr., 559, 445,
MW cutoff membrane and followed by two consecutive rinses with 0.2 mL water [34,35]. A more
detailed procedure detailing the time and the centrifugal force to be applied as well as an increase
in the number of washing steps was subsequently reported [36]. Ways to decrease the preparation
time by increasing the centrifugal force [37,38], and to increase the recovery by repeatedly washing
the membrane and inverting the filtering device were also described [38]. An intermediate procedure
led to an enhanced resolution of rhEPO glycoforms when compared with the analysis performed
under identical conditions but without sample pretreatment [30]. As it will be shown later, this type
of sample preparation step that eliminates excipients of low MW is especially important when the
separation mode is CIEF [36,39]. This same type of procedure for eliminating low MW excipients
was also useful to remove mannitol from a standard of human urinary EPO (uEPO) to be analyzed
by CZE [40].
When the sample matrix contains high MW components, such as in biological fluids or in phar-
maceutical products formulated with protein excipients, strategies other than MW-based sample
preparation are often required. Using again rhEPO as an example, glycoform analysis of products
containing HSA under the same CZE conditions as those used for samples containing low MW
components was unsuccessful. An approach at separation that did not require a sample prepara-
tion step was the development of a CZE method involving the selective binding of HSA. This was
accomplished on a polyamine-coated capillary by adding nickel ions to the BGE. The metal ions
selectively bonded HSA and decreased its mobility leading to separation of the otherwise comigrat-
ing rhEPO [41]. Polybrene (PB)-coated capillaries with a buffer containing acetic acid and methanol
also allowed the separation of rhEPO from HSA [42]. A different approach relied on performing the
immunochromatographic removal of HSA before separation. Figure 22.5 shows the electrophero-
grams of the unresolved peak obtained for an HSA-formulated rhEPO sample before HSA removal
and the separated rhEPO bands after the immunochromatographic HSA depletion [43]. Figure 22.6
(a)
0.03
Absorbance (U.A.)
0.02
0.01
0.00
0 10 20 30 40 50
Time (min)
(b) 0.02
0.01
Absorbance (U.A.)
0.00
-0.01
-0.02
10 15 20 25 30
Time (min)
FIGURE 22.5 CZE electropherograms of rhEPO alpha formulated with human serum albumin (HSA),
(a) without immunodepletion of HSA, (b) after HSA immunodepletion. Separation conditions: Uncoated fused-
silica capillary 87 cm × 50 µm i.d. Separation buffer: 0.01 M Tricine, 0.01 M NaCl, 0.01 M sodium acetate,
7 M urea, and 3.9 mM putrescine, pH 5.5; pressure injection 30 s at 0.5 psi; 25 kV; temperature 35◦ C; detection
at 214 nm. (Adapted from Lara-Quintanar, P. et al., J. Chromatogr. A, 1153, 227, 2007. With permission.)
shows the separation of a similar sample on polyamine-coated capillary and nickel ions added to the
BGE [41]. Both methods provide separation of glycoforms and while the one performed on coated
capillaries does not require a prior sample preparation step, the other method can be performed with
bare fused-silica capillaries and provides better glycoform resolution.
Sample preparation steps used for other samples are mainly a function of the matrix, the most
complicated ones being for the biological fluids. Specific methods used to prepare some of these
samples are detailed in Section 22.5.2.
(d)
Abs at 200 nm
(c)
(b)
(a)
0.00 10.00 20.00 30.00 40.00 45.00

Time (min)
FIGURE 22.6 CZE electropherograms of (a) human serum albumin, (b) rEPO, (c) HSA-formulated rEPO.
Inset (d): enlarged view showing rEPO glycoform separation in formulated product. Separation conditions:
Beckman eCAP amine capillary, 200 mM sodium phosphate, pH 4.0, 1 mM nickel chloride, –15 kV (75 µA),
UV detection at 200 nm. (From Bietlot, H.P. and Girard, M., J. Chromatogr. A, 759, 177, 1997. With permission.)
22.4.2 SEPARATION OF GLYCOFORMS

22.4.2.1 Solutions to Protein Adsorption Problems
Very high separation efficiency can be achieved when the interaction between a protein and the inner
wall of the capillary can be eliminated [44]. This is especially important for glycoproteins where
this type of unwanted interaction has to be prevented to achieve the high efficiency required for the
separation of glycoforms. Two basic strategies can be utilized to overcome this problem. The first
one consists in changing the buffer pH and the second one in modifying the inner capillary wall.
The first, and simplest, approach is to select a buffer pH at which interactions become small. At a
pH below 3, the silanol groups on the silica surface are fully protonated and bear no electric charge,
thus minimizing interactions since most proteins have pI values above 3. On the other hand, at pH
8 or higher, silanol groups are fully ionized and the silica surface becomes negatively charged. So,
at pH 8, a protein having a pI smaller than 8 will have a net negative charge and will be repelled
by the surface, while one having a pI higher than 8 (a basic protein) will be charged positively and
may be adsorbed on the capillary walls. In such a case, a buffer with a pH higher than the pI of the
protein could be used for preventing their adsorption. Although these strategies have been proposed
for the separation of proteins since the early days of CE [45,46], they limit the selection of buffer
pH as a freely adjustable parameter for optimization. In addition, too large differences between the
buffer pH and the protein pI can lead to structural changes resulting in band broadening and low
recovery. These limitations have led to other approaches for preventing protein adsorption, which
allow a wider selection of buffer pH.
Several methods employing capillary wall modifications have been developed and they can be
grouped into several categories: dynamic coatings by the addition of a cationic modifier (usually
an amine) to the BGE [47,48], permanent coatings by physical adsorption of a cationic modifier
(usually a polymer) [47,48], and permanent coating by covalent bonding of a hydrophilic polymeric
layer [49].
22.4.2.1.1 Cationic Buffer Additives

The addition to the buffer of cationic substrates that interact with the silanol groups at the surface
of the fused-silica capillary represents the simplest method to improve the separation efficiency in
glycoprotein analysis. Several effects have been attributed to these additives, which include decreas-
ing the interaction of the analyte with the inner capillary walls, controlling the electroosmotic flow
(EOF), increasing the protein solubility, and enhancing the selectivity of the separation. Although
additives of several types, including neutral and ionic substances, have been reported [47,48], amine
modifiers have been more frequently used for the separation of protein glycoforms. There are two
main groups of amines, one comprising diaminoalkanes and the other polyaminoalkanes.
In the first group, short alkyl chain diamines have been used at millimolar concentrations for the
effective separation of glycoproteins such as ovalbumin (OVA) [50,51], rhEPO [52], recombinant
factor VIIa [53], granulocyte colony-stimulating factor (rhGCSF) [54], Tf [55], and human chorionic
gonadotropin (hCG) [56]. The diaminoalkanes play a role in the separation by affecting both the
resolution and the migration time. The separation of hCG glycoforms in uncoated silica capillaries at
different concentrations of 1,3 diaminopropane (DAP) in borate buffer (Figure 22.7) shows that an
increase in the diamine concentration results in an increase in resolution with a concomitant increase
in migration times [56].
Replacing the terminal diamino groups of diaminoalkanes by quaternary ammonium moieties
gave rise to more efficient separations [57]. Although a similar resolution was achieved with
both types of additives, lower concentrations of quaternary ammonium salts were necessary and
shorter migration times were obtained, an indication of effectiveness of these bis-quaternary ammo-
nium compounds. Quaternary ammonium salts such as hexamethonium chloride and bromide and
decamethonium bromide (DcBr), having alkyl chains of 6 and 10 carbon atoms, respectively, have
been used for the separation of the isoforms of OVA and hCG [57], and Tf [55]. For quaternary
ammonium salts, the longer alkyl chains were shown to be more effective and chloride salts led to
shorter analysis time than bromide ones.
Polyaminealkanes, such as spermine and spermidine, have been used for the separation of OVA
glycoforms [51]. Interestingly, while the efficiency of a polyamine for preventing protein adsorption
to the silica walls was shown to increase with the number of amino groups in the chain, this number
needed to be higher than three for it to become an effective adsorption inhibitor [58].
22.4.2.1.2 Capillary Coatings by Physical Adsorption

Physical adsorption of polymeric cationic coatings on the inner surface can be easily accomplished
by equilibrating the capillary with a solution containing the polymeric additive. In some cases,
the polymer remains so tightly bonded to the surface that this does not lose its coating even after
performing rinses and, therefore, the polymer does not need to be included in the buffer.
Polybrene (PB) has been used to prepare physically adsorbed polymeric coatings [59,60].
PB is a polycationic polymer, composed of quaternary amines (N,N,N ,N -tetramethyl-1,3-
propylenediamine), which strongly adsorbs to the inner surface of the capillary and reverses the
(a)
DMF
(b) DMF
(c)
DMF
(d) DMF
0 10 20 30 40 50
Time post-injection
(min)
FIGURE 22.7 Effect of various concentrations of diaminopropane on the electrophoretic migration and sepa-
ration of hCG (4 mg/mL). Separation capillary: 100 cm × 50 µm fused silica. Separation buffer: 25 mM borate
(pH 8.8). Other separation conditions: 25 kV, 28◦ C. Detection: 200 nm. The diaminopropane concentrations
used were: 0 (a), 1.0 (b), 2.5 (c), and 5.0 mM (d). (From, Morbeck, D.E. et al., J. Chromatogr. A, 680, 217,
EOF in acidic buffers. In these buffers, most proteins are positively charged and are repelled from
the surface. This strategy has been used for the separation of rhEPO glycoforms [37], however, with
a lower resolution than that obtained with the diaminoalkane putrescine (1,4-diaminobutane, DAB).
Its main advantage resides in that this coating allows the use of volatile buffers compatible with
MS. Glycoforms of ribonuclease B (RNase B) and horseradish peroxidase (HRP) [61] have also
been separated using this strategy. Other polymers prepared from 1,3-propylenediamine and having
different charge density have also been reported [62].
PB has a limited stability and is degraded after a few separations. Better coating stability was
shown for multiple ionic polymer layers. They are prepared from a first layer made up of a cationic PB
coating that is then covered with another physically adsorbed layer from an anionic polymer such as
dextran sulfate [63] or poly(vinylsulfonate) [64]. In these coatings the EOF is essentially constant at
most pH values. However, because of the negative charge at the capillary surface, positively charged
proteins could adsorb to the coating and as such cannot be separated in this type of capillaries.
The use of a third layer composed of a cationic polymer, usually PB, has overcome this limitation
and has improved the stability of the coating [63,65]. Several groups [66–69] have used commercial
reagents that produce a double ionic layer (e.g., CEofix-CDT kit; Analisis, Namur, Belgium) for the
analysis of Tf isoforms in human serum as will be shown in detail in Section 22.5.2.
Recently, a new polyacrylamide (PAA) derivative that allows dynamic coating (UltraTrol LN-;
Target Discovery, Palo Alto, CA) has been reported for the efficient separation of glycoforms of
ATT [70]. A comparison of the ability of PB and UltraTrol LN for the separation of bovine alpha-
1-acid glycoprotein (AGP), bovine serum fetuin (BSF), and rhEPO sialoforms and their application
coupling with MS has been reported [30,71]. Better resolution of sialoforms was obtained with
capillaries coated with UltraTrol-LN in acidic volatile buffers used for coupling CE to ESI-TOF MS.
22.4.2.1.3 Capillary Coatings by Covalent Binding

The covalent binding of polymers to the inner capillary walls leads to permanent coatings with a
thin layer of polymeric material. In a way similar to physically adsorbed polymeric coatings, they
minimize protein adsorption by shielding the silanol groups at the surface. In addition, these types
of coatings eliminate or fix the EOF over a wide range of buffer pH allowing for a better control on
the separation. However, good permanent coatings are more difficult to prepare than dynamically
prepared coatings. Typically, preparation is carried out in a two-step process. First, a bifunctional
compound, usually a silane derivative, is reacted with the silanol groups at the capillary surface
through one of its functional groups, giving rise to a first layer and leaving the second functional
group available to be covalently bound to the second layer. The second layer is produced by reacting
a monomeric unit and polymerizing it. In some cases, an out-of-column synthesized polymer is
bonded directly to the first layer. Several hydrophilic polymers such as linear and cross-linked
PAA [72,73], methylcellulose and dextran [74], and polyvinyl alcohol (PVA) [75] to cite only a
few have been developed for the separation of proteins. Commercial hydrophilic-coated capillaries
have also been used for the separation of recombinant human tissue plasminogen activator (rhtPA)
glycoforms [76].
The importance of using coated capillaries for glycoform separations has been shown by Thorne
et al. [77] who demonstrated that rhtPA does not migrate in uncoated capillaries, even when ε-
aminocaproic acid was added to the separation buffer. However, when the same buffer was used
with capillaries having hydrophilic coatings, rhtPA could be separated into several isoforms. They
also obtained better separations using PVA-coated capillaries than with PAA-coated ones.
The importance of additives to the separation buffer, even for coated capillaries, should be
emphasized here. Both zwitterionic buffers and tensioactive additives have been shown to improve
the separation and recovery of rhtPA with PVA-coated capillaries [77]. On DB1 capillaries, a
dimethylpolysiloxane coating widely used in gas chromatography (GC), optimal separations of
fetuin, rhEPO, and AGP isoforms were obtained with acidic acetate buffers containing 0.4–0.5%
(w/v) hydroxypropylmethycellulose (HPMC). In this case, the addition of a neutral polymer to the
separation buffer makes the EOF almost negligible and protein interactions with the inner surface
of the capillary were prevented [78].
The importance of the elimination of protein adsorption to the surface to obtain good glycoform
resolution should also be emphasized. This is well illustrated in the separation of rhBMP-2 mentioned
earlier, where the use of a commercial hydrophilic-coated capillary with a phosphate buffer without
any additive led to efficiencies in excess of 7 × 105 plates/meter and up to nine isoforms differing
only in the number of mannose residues being separated (Figure 22.8) [31].
22.4.2.2 Factors Affecting the Resolution of Glycoform Peaks

Owing to the importance of the capillary on resolution, the previous section was devoted to its
coating. Besides preventing or diminishing interactions with the silanol groups on the silica surface,
the capillary coating modifies the EOF. In this way, the apparent electrophoretic mobilities of the
glycoform peaks are modified with respect to their effective mobilities, thus allowing for the mod-
ulation of the separation. In addition, some of the compounds used as dynamic coatings have been
3 4
1.0 2 5
Absorbance (200 nm)

0.8
6
0.6 1
7
0.4 8
N = 350 000/50 cm 9
0.2 R = 1.12
0.0
22.0 22.5 23.0 23.5 24.0 24.5 25.0

FIGURE 22.8 CZE separation of rhBMP-2 in a 50 cm × 50 µm i.d. Bio-Rad coated capillary. Sample was
electroinjected for 4–8 s at 6–12 kV in 0.1 M phosphoric buffer (pH 2.5). Detection: 200 nm. Separation
temperature 20◦ C. (From, Yim, K. et al., J. Chromatogr. A, 716, 401, 1995. With permission.)
shown to play a role in the separation of glycoforms through interactions with glycoproteins or their
complexes, and in this sense they will be considered in this section.
In principle, each CE mode should be adequate to perform the separation of intact glycoforms
as long as they differ in molecular properties that are distinguished in the particular mode used.
When taking into account that the change in mass between two glycoforms is often almost negligible
compared with the total mass of the protein and that relative changes in charge, when occurring, are
more noticeable, it is not surprising that most separations have been achieved by CZE and CIEF.
However, the other modes, MEKC, CGE, ACE, and CITP have also proven to be useful in some
instances. The respective advantages and drawbacks of different modes have been discussed for
some glycoproteins [36,76,77,79,80]. The following discussion will focus on aspects of the effect
of factors affecting glycoform resolution that can be helpful for methods development together with
providing examples to illustrate them. Although this section is organized according to separation
modes, it does not imply that in some instances more than one mechanism may be involved in the
separation achieved. Details can also be found in Section 22.5.
22.4.2.2.1 Capillary Zone Electrophoresis

CZE was used for the separations of forms of glycoproteins, namely iron-free Tf and RNase almost
two decades ago [79,81]. Different factors are known to influence the CZE separation of glycoforms
and, when possible, the influence of individual variables will be considered. However, as it will be
mentioned, in several instances an interconnection between the effects of two or more factors exists.
The resolution between two peaks can be described [45] as
µe
RS = N 1/4 , (22.2)
4 (µEOF + µe )
where µe is the difference in the electrophoretic mobility between two peaks, µEOF is the elec-
troosmotic mobility, and µe is the mean electrophoretic mobility of the two peaks. Consequently, a
decrease in the µEOF should result in an increase in resolution for other factors remaining constant.
An option to selectively modify µEOF is through manipulation of the zeta potential of the capillary
and ways of achieving low values of µEOF by using dynamically or covalently coated capillaries have
been described in the previous section of this chapter. Other options to increase resolution would
be to increase the capillary plate number, to decrease the mean electrophoretic mobility, to increase
the difference in the electrophoretic mobility of the two peaks, or to match the electroosmotic and
the mean electrophoretic mobilities of different signs so the vectorial result is close to zero. The
following discussion will briefly present the influence of some factors on these parameters. Buffer
additives, which can act as capillary coatings, will also be considered in this section with respect to
their effects on resolution by a joint action with other additives.
22.4.2.2.1.1 Buffer pH A variable that is typically changed to optimize differences in effective

mobilities between glycoforms is the buffer pH. This frequently used approach derives from the
potential improvement in resolution when using a buffer at a pH close to the (pI) of the glycoprotein
where both its charge and mobility are lower. In other words, the closer the pH and pI values the higher
the chances of increasing charge differences between two glycoforms. An example of this effect was
shown for rhtPA, a protein with pIs in the range 4.5–6.0, for which better resolution was obtained at
pH 3.6 than at pH 2. [82]. The enhancement of resolution is frequently accompanied by an increase
in the analysis time as a consequence of the decrease in apparent mobility. For example, enhanced
resolution at the cost of doubling the migration time was obtained by decreasing the pH from 5.5 to
4.5 when analyzing the novel erythropoiesis-stimulating protein (NESP), a protein known to be more
acidic than rhEPO [83]. The same effect was observed for AGP, a very acidic protein (pI 1.8–3.8),
when the buffer pH was changed from 4.5 to 3.5. In this case, the increase in analysis time was
compensated by using a shorter capillary, while maintaining baseline resolution (Figure 22.9) [84].
While the previous examples were carried out on bare fused-silica capillaries, a similar phenomenon
has been observed on coated capillaries. Resolution of fetuin (pI 3.2–3.8) glycoforms on DB-1 coated
capillaries gradually increased for the pH buffer going from 9.0 to 5.0 [78].
In some instances, the choice of the most appropriate buffer pH precludes the need for buffer
additives. This is well illustrated in a study of rhGCSF that showed that although additives, mainly
DAB, could be effective in enhancing the separation of glycoforms, the pH control in borate buffer
0.03 5 6
4
(c) 3 7
2 8
1 5
0.02 4 6
3 7
(b) 1 2 8
Absorbance at 214 nm
0.01
(a)
0
10 20 30 40 50
0
–0.01 Time (min)
–0.02
EOF marker
–0.03
FIGURE 22.9 CZE electropherograms of standard human AGP. Conditions: (a) uncoated capillary, 87 cm
× 50 µm; separation buffer: 0.01 M Tricine, 0.01 M NaCl, 0.01 M sodium acetate, 7 M urea and 3.9 mM
putrescine, pH 5.5; injection 30 s at 0.5 psi; 25 kV; 35◦ C; detection at 214 nm; pH of separation buffer 4.5,
other conditions as in (a); (c) capillary length 77 cm, other conditions as in (b). (From Lacunza, I. et al.,
was enough to obtain baseline resolution of the two glycosylated, the nonglycosylated, and the
desialylated forms [54].
22.4.2.2.1.2 Buffer Ionic Strength The ionic strength of the separation buffer is another factor
to be taken into account when trying to optimize glycoform resolution. Generally, an increase in ionic
strength increases resolution by decreasing the EOF and the analyte apparent mobility. However,
anomalous results have been observed. This is the case for formic acid used as the separation buffer in
the CZE-MS analysis of RNase B on PB-coated capillaries. Increased resolution and decreased EOF
were observed when the formic acid concentration was increased from 0.1 M to 2.0 M. Interestingly,
at 0.5 M formic acid, resolution was worse than for any other concentration although the EOF was
lower than at 0.1 M formic acid. This behavior was interpreted as resulting from conformational
changes of the protein [61].
22.4.2.2.1.3 Borate Buffer In view of its ability to resolve peaks not solely based on charge
differential borate has become a substance of major importance in the resolution of glycoforms. The
formation of complexes between hydroxyl groups and borate ions is at the basis of the separation
mechanism, making it possible to resolve peaks of glycoforms with the same charge. Borate com-
plexation is usually favored with cis 1,2-diols over trans 1,2-diols, but some carbohydrates do not
follow this rule [85]. For a constant amount of carbohydrate, the complex concentration increases
with increasing borate concentration and pH as a result of higher alkaline borate ion concentration,
which is known to be the species that complexes with diols [86]. Asialo fetuin analyzed in DB-1
coated capillaries with Tris-borate buffer at pH 8.5 and containing HPMC is an example. The addition
of the cellulose polymer, which shields the capillary surface from interactions with the protein, was
necessary to achieve resolution; however, excess HPMC hampered the separation, probably due to
higher viscosity [78]. Depending on the polymer concentration a sieving effect could also be in play.
A comparison of the separation buffers Tris-glycine, ammonium formate, and borate for the analysis
of hCG glycoforms also showed an enhancement in resolution when using the borate buffer [56].
However, borate is not necessarily always the buffer of choice and depending on the glycoproteins
and the specific separation conditions, other buffers provide better resolution [87].
22.4.2.2.1.4 Borate Buffer and Diamino Additives The combined action of borate and diamino
additives in the separation buffer reported in 1992 by the groups of Taverna and Landers [50,82] has
been widely used to achieve separation of glycoforms. In addition to those reported by the two groups,
other types of bifunctional alkyl amines or bis-quarternary amines have been reported in separations
involving borate-based buffers [56,57]. The necessity for the diamine to exist as a divalent cation
was stated in these studies and a mechanism in which the borate-complexed protein interacts with the
amino group of the capillary-bound additive was suggested. Such a partition mechanism that would
involve CZE and electrochromatography components has been accepted by other authors [88].
22.4.2.2.1.5 Other Components of the Separation Buffer The use of zwitterions as buffer
additives has the advantage of not affecting the conductivity of the separation buffer. These additives
are thought to enhance glycoform resolution by suppressing protein–wall interactions, by forming
ion pairs with the glycoproteins, and by preventing individual protein molecules from interacting
with lysine residues on adjacent molecules resulting in precipitation. For example, EACA [76] and
other similar ω-amino acid buffers have proven effective in the separation of rhtPA forms [77]. In
this instance, the addition of a nonionic detergent such as Tween 80 increased the protein recovery.
Modified celluloses added to a borate buffer in DB-17 coated capillaries allowed the separation of Tf
sialoforms. Although the separation mechanism is not clear, it is thought that both charge differences
and gel sieving are contributing factors [89].
22.4.2.2.1.6 CZE Separation of EPO Glycoforms To illustrate the many factors that can be
modified to optimize glycoforms resolution it is interesting to follow the different approaches reported
for the CZE separation of rhEPO, starting with the method reported by Watson and Yao [52]. On a
bare fused-silica capillary, no separation was possible for a commercial rhEPO sample at pH values
between 5 and 10 using buffers of different ionic strengths. However, at pH 6.2, the use of additives
that decreased the EOF allowed some resolution, with the best results achieved with 2.5 mM DAB.
Further improvement was achieved by adding 7 M urea to the buffer and the optimized conditions
consisted of a separation buffer of 10 mM tricine, 10 mM NaCl, 2.5 mM DAB, 7 M urea, pH 6.2,
in a 50 cm bare fused-silica capillary and applying 10 kV. Baseline resolution of six peaks was
obtained in 35 min. A modified version of this method was developed and used for an international
collaborative study to assess the BRP for rhEPO of the EP, which consisted in a mixture of alpha-
and beta-epoetin. [34]. The modified method consisted in the addition of 10 mM sodium acetate to
the separation buffer, adjusting the pH to 5.5 with 2 M acetic acid, and performing the separation at
30◦ C in a 107 cm capillary. In the initial study, large discrepancies in migration times were found
among participants (between 32 and 120 min). These results were possibly attributable to differences
of the capillary walls since capillaries from different manufacturers and even from different batches
from the same manufacturer have been shown to have very different EOFs. This method was later
implemented as an official method for the identification of rhEPO (see Section 22.5.1.1) [35]. A
reduction of migration times from 70 to 36 min while keeping enough resolution between peaks and
good repeatability has been achieved by decreasing the DAB concentration to 0.025 mM [36]. In
contrast, the increased analysis time at higher DAB concentrations could be reduced by applying
a higher voltage, from 15.4 to 30 kV [90]. In this instance, baseline resolution of glycoforms have
been performed in less than 30 min, and this method could also be applied, although without baseline
resolution, to the NESP [38,43].
The development of several CZE methods for the same glycoprotein, using buffers and capillaries
other than those previously mentioned, clearly shows the versatility of CE. As another example, the
early work by Tran et al. [91] on the influence of factors such as buffer pH and type, and the addition
of organic modifiers showed the possibility of obtaining partial resolution of multiple rhEPO peaks
with a 100 mM acetate–phosphate buffer at pH 4 in uncoated fused-silica capillaries. A marked effect
on resolution was observed when the pre-equilibration time before using the capillary was increased
from 4 to 10 h. The shorter pre-equilibration time and the same separation conditions have been
used by others [92]. On the other hand, a different method for rhEPO analysis has been performed
on C8-coated capillaries using a phosphate buffer [93]. A 300 µm i.d. tube made up of fluorinated
ethylene–propylene (FEP) copolymer was used in a hydrodynamically closed separation system,
which allowed enhancing the loadability. A buffer consisting on N-(2-hydroxyethyl)piperazin-N -
2-(hydroxypropane sulfonic acid) (HEPPSO), 1,3-bis[tris(hydroxymethyl)-methylamino] propane
(BTP), and methylhydroxyethylcellulose (MHEC) 30,000, at pH 7.25 allowed to obtain partial
resolution of seven bands for rhEPO [94]. CZE under similar conditions with a borate buffer
at pH 8.8 containing MHEC was used to monitor the fractionation of rhEPO by preparative
CITP [95].
The coupling of the CZE step to detection systems other than UV has required the development
of separation conditions compatible to the detection system used. For instance, the presence of
primary amines, such as DAB, in buffers needed to be avoided for compatibility with laser-induced
fluorescence (LIF) of compounds derivatized with fluorogenic substrates through their amino groups
[90]. Baseline resolution of eight peaks in approximately the same time was achieved by substituting
DAB by morpholine and tricine by boric acid (to avoid potential traces of primary amines in the
tricine buffer) and by adjusting the concentration of other buffer components to compensate for
the increase in electrical current. In the same work, modifications were also required to achieve
compatibility with MS detection where nonvolatile salts, urea, and amines should be usually avoided.
A physically adsorbed polyethylenimine-coated capillary was used to overcome protein adsorption
to the capillary walls in the absence of cationic additives and the use of an acetate buffer at pH 5.05
allowed the partial resolution of at least five bands of rhEPO. Other types of coated capillaries have
been used for the analysis of EPO by CE-MS as detailed in Section 22.4.3.3 [30,37,42,62,96].
22.4.2.2.2 Capillary Isoelectric Focusing

Almost two decades ago, CIEF was described by Kilar and Hjerten [79,97] for the separation of Tf
glycoforms and it has, in principle, some advantages over other CE modes, namely, (1) the large
volume of sample introduced in the capillary would result in an increase in sensitivity compared with
nanoliters amounts in other modes; (2) the measurement of an important molecular property, the
isoelectric point (pI), is determined; and (3) a large number of experimental factors can be modified
to improve resolution. With regard to the first two advantages, experimental considerations made
it that their importance is not what it should be. With respect to the sensitivity issue, it should be
stated that commercial ampholytes are generally not designed to work in a capillary format with UV
detection and they show high absorbance at 214 nm. In fact, they are manufactured for classical IEF
slab gels with detection at 280 nm, a wavelength at which the extinction coefficient of proteins is
lower than at 214 nm. So, in practice, the sensitivity of CIEF has been shown in some instances to be
similar to that obtained for the same glycoprotein by CZE [36]. With respect to the pI determination,
the value obtained for a given analyte should be considered cautiously since, in many instances, urea
or other additives that affect the pI of a protein are used in addition to the fact that the pH gradient
is not always linear [98].
CIEF can be performed in two general ways: a one-step or a two-step process. In the one-
step process, the capillary retains a residual EOF that allows protein bands to migrate through the
detection window, while in the two-step process the absence of EOF is desirable for better focusing
of the protein bands being mobilized to pass the detection point in the next step. However, even in
the one-step mode a controlled EOF is necessary to achieve enough focusing before reaching the
detector. Both modes have advantages and drawbacks as was shown for the effect of different factors
in the separation of AGP [99].
In CIEF, the total time elapsed between sample introduction and detection depends on the focusing
time and on the time for a band to reach the detection window during mobilization. These two effects
can take place either simultaneously or consecutively depending on whether the process is a one-step
or two-steps, respectively. So, the term “migration time” is not strictly correct in CIEF but it is used
in this chapter for simplicity. It is important to note that the need for mobilization through the detector
window can be eliminated in systems where whole-capillary imaging detection is used [100].
Samples are introduced in the capillary either separated or mixed from ampholytes, the latter
being the most frequently used. In addition to the sample and ampholytes, additives can also be
introduced in the capillary in the so-called sample mixture, as will be shown in later text.
The different types of capillary coatings providing reduced or no EOF have been described in a
previous section. A comparison of different capillary coatings for the resolution of rhtPA glycoforms
has been performed and, in some cases, capillaries with different degree of EOF reduction have
shown similar resolution, a situation that was probably due to the presence of additives controlling
the EOF [101]. While similar resolution has been achieved with different coatings, the stability of
the coatings in terms of the number of runs performed was evaluated in the analysis of recombinant
human antithrombin III (rhATIII) and the recombinant human immunodeficiency virus envelope
glycoprotein rgp 160s. Higher stability was observed for dextran-coated [102] and PVA-coated
[103] capillaries than for PAA-coated capillaries. Besides the capillary coating, other variables that
can be controlled for improving resolution of glycoforms are described in the following text.
22.4.2.2.2.1 Additives for Reducing EOF Besides the use of coated capillaries for controlling
the EOF, the addition of substances aimed toward reducing the EOF through an increase in the vis-
cosity is frequently used. This is the case for poly(ethylene oxide) [36,39,77] and HPMC [101,103].
The latter has proven to be effective even when using bare fused-silica capillaries, where it acts as a
dynamic coating [104].
22.4.2.2.2.2 Ampholytes This is one of the variables to which a lot of attention has been paid.
Both the nature and pH of ampholytes as well as the combination of several pH ranges have been
Absorbance (280 nm)

0.025
(b)
Absorbance (280 nm)
0.16 (a) 0.02
0.12 0.015
0.08 0.01
0.04 0.005
0 0
10 15 20 25 30 35 40 10 15 20 25 30 35 40
–0.04 –0.005
Absorbance (280 nm)

Absorbance (280 nm)
0.025 (c) 0.025

(d)
0.02 0.02
0.015 0.015
0.01 0.01
0.005 0.005
0 0
10 20 30 40
–0.005 –0.005 10 15 20 25 30 35 40
FIGURE 22.10 Effect of types of ampholytes on separation of AGP forms. Total ampholyte concentra-
tion in the sample mixture: 6.3% (v/v). Types of ampholytes, specified by their pH range (and ratios (v/v)):
(a) range 3–10; (b) mixture of ranges 3–10 and 2.5–5 (1:1); (c) mixture of ranges 3–10, 2.5–5, and 2–4 (1:1:1),
(d) mixture of ranges 3–10, 2.5–5, 2–4, and 3–5 (1:1:1:1). Other analytical conditions: PVA-coated capillary,
27 cm (effective length 20 cm). Sample mixture in CIEF gel: 1 mg/mL AGP, 5.6 M urea, 20 mM NaCl, and
ampholytes. Catholyte: 20 mM NaOH titrated with H3 PO4 to pH 11.85. Anolyte: 91 mM H3 PO4 in CIEF gel.
Focusing: 10 min at 20 kV. Mobilization step: 20 kV and 0.5 p.s.i. N2 pressure. Temperature: 20◦ C. Detection:
280 nm. (From Lacunza, I. et al., Electrophoresis, 28, 1204, 2007. With permission.)
considered. In general, better resolution is attained with mixtures of ampholytes. The influence of
the commercial source of ampholytes on resolution has been shown, even when only a wide-range
ampholyte solution is used [77]. Similarly, for a given total concentration of ampholytes mixed in
constant proportions, the resolution was shown to depend on the nature of the ampholytes [105].
Besides providing different resolutions of rhtPA glycoforms, commercial ampholytes have different
optical properties that have an impact on detection [106]. Ampholytes from different sources have
been shown to be different in nature and in focusing properties [107,108]. As such, mixtures from
different manufacturers are generally preferred, since the different nature of the compounds from the
different sources may help to achieve better focused zones by reinforcing the pI range of interest. This
effect was clearly seen for the separation of rhEPO [36] andAGP glycoforms [109] (see Figure 22.10).
Apart from the nature of ampholytes, their proportion and total concentration are other factors
to be considered. Usually, higher proportions of an ampholyte or narrow-range ampholytes close
to the pI of the glycoprotein provide better resolution [103]. For the amount of total ampholyte, a
compromise is generally achieved since higher percentages not only provide better resolution but
also lead to higher background absorption and, consequently, decreased sensitivity [39,103,109].
22.4.2.2.2.3 Agents to Avoid Precipitation Protein precipitation is favored during the focus-
ing process where a protein is concentrated in a narrow zone close to its pI, and associated salts
are separated from it. The addition of nonionic surfactants, organic modifiers, or chaotropes to the
sample can be used to decrease chances of precipitation [110]. The presence of urea in the sep-
aration buffer has been shown to markedly improve the resolution of AGP an EPO glycoforms
[36,100,109], being also used to obtain adequate separation of rhtPA glycoforms [77,104]. Triton
X-100 was found to be unnecessary when urea was used to separate rhtPA glycoforms [101]. This
detergent was used to avoid precipitation for the analysis of rhATIII [102]. However, it may be
necessary to resort to a combination of agents. This was the case for recombinant human immun-
odeficiency virus envelope glycoprotein where several additives including a chaotrope (urea), a
0.02 Absorbance
(a)
–0.015
0.01
0.005
–0.005
0 2 4 6 8 10 12 14 16 18 20
0.005 6
(b) 5
0.004 7
4
0.003
0.002
0.001 3
2
0 1
–0.001
–0.002
0 2 4 6 8 10 12 14 16 18
Time (min)
FIGURE 22.11 Effect of urea on CIEF separation of rhEPO. (a) Without urea in the sample mixture, and
(b) with 7 M urea. Analytical conditions: polyacrylamide-coated capillary 27 cm (effective length 20 cm) ×
50 µm i.d. Focusing: 6 min at 25 kV. Sample mixed with CIEF gel and ampholytes. Ampholyte mixture (1:2,
v/v) of pH 3–10 and 2.5–5 ranges. Detection: 280 nm. (From Cifuentes, A. et al., J. Chromatogr. A, 830, 453,
zwitterion [3-(cyclohexylamino)-1-propanesulfonic acid, CAPS], and a sugar (saccharose) provided

the best results [103].
22.4.2.2.2.4 pH Extenders In order to avoid the focusing of glycoforms past the detection win-
dow, a situation that would prevent their detection, it is necessary to add a substance with an
appropriate pI to block that zone of the capillary. It is referred to as a pH extender and the most
frequently used is N,N,N ,N -tetramethylenediamine (TEMED). It is usually added to the sample
mixture at a concentration proportional to the length of capillary to be blocked. The larger the pH
extender concentration is the farther away a glycoform will focus from the detection point and,
consequently, the greater the migration time will be, as was observed for rhtPA without any marked
influence on resolution [105]. The same authors also reported that for TEMED concentrations about
10 times higher, in the range of 3.75–7.5% (v/v), the effect on migration time was not noticeable
[101]. Caution should be taken when using TEMED as it has been shown to damage the capillary
coating [99,103]. Depending on the pI of the glycoprotein, other less basic extenders, such as alanine,
can be used [99].
22.4.2.2.2.5 Focusing Time and Voltage For the one-step CIEF separation of rhtPAglycoforms,
an increase in voltage led, as expected, to a decrease in migration times without any influence on
resolution [105]. For the two-step process, the two factors involved on the focusing step (i.e., voltage
and time) have an impact on resolution. According to the theory [111], increasing voltages up to
25 kV provided slightly better resolution of rhEPO glycoforms; however, loss of resolution was
observed at higher voltage, probably due to an excess of Joule heating [36].
It is usually considered that focusing is finished when the current decreases below 10% of its
initial value and stabilizes. However, in practice, focusing times as short as 0.5 min and corresponding
to 69% of the initial current have been shown to lead to the same resolution of EPO isoforms as
for longer focusing times. This can be understood taking into account that during the pressure-
mobilization step, the voltage is applied to avoid band broadening resulting in a continuation of
focusing [36]. In contrast, focusing times longer than needed may have a detrimental effect not only
on the analysis time but also on the quality of the separation. The total depletion of salts due to a
long focusing time led to precipitation of monoclonal anti-alpha-1 antitrypsin [100]. In addition,
focusing times that are too long may lead to greater anodic drift and can affect the isoform separation
of highly acidic proteins such as AGP [109]. Incomplete focusing due to times shorter than needed,
however, may lead to double peaks. These examples show that careful optimization of the focusing
time is required.
22.4.2.2.2.6 Presence of Salts in the Sample Mixture The presence of salts in the sample
mixture remains a controversial point. Some authors have claimed that the ionic strength of the
sample should be as low as possible while others have found it to help improve resolution. For the
analysis of anti-alpha-1 antitrypsin [100], AGP [109], or rhEPO [36,39], the presence of salt was
necessary for the separation of isoforms. This effect is likely related to prevent protein precipitation
and denaturation, although other effects such as improved focusing or masking of the silanols of the
capillary walls cannot be excluded.
22.4.2.2.2.7 Mobilization Step In the two-step CIEF mode the mobilization step is performed
after focusing. Hydrodynamic (more frequently performed by applying pressure rather than by
vacuum) and chemical mobilization are the two general ways of driving the focused zones to the
detection point.
As indicated earlier, when the mobilization is performed by pressure, a voltage is applied simul-
taneously, usually at the same strength as the voltage used during the focusing step. However, a
higher voltage may be needed in some instances as shown for a monoclonal antibody (mAb) [112].
Chemical mobilization with a strong electrolyte has been shown to lead to better resolution than by
using a weak one [97].
Performing hydrodynamic and chemical mobilizations together has been shown to be beneficial
for the separation of rhEPO [39] and even necessary for the separation of glycoforms of acidic
proteins such as AGP [109] (see Figure 22.12). The addition of phosphoric acid to the catholyte,
which consists of sodium hydroxide, at the very beginning of the focusing step promotes chemical
mobilization that continues when the pressure is applied at the hydrodynamic mobilization step.
22.4.2.2.3 Micellar Electrokinetic Chromatography

There are many factors involved in the MEKC separation mode that contribute to the complexity of
its mechanism. On the other hand, they provide a large number of variables that can be modulated to
optimize the separation. MEKC has been used on a few occasions for the separation of glycoproteins,
0,05
0,04
Absorbance (280 nm)
0,03
(b)
0,02
0,01
(a)
0
0 5 10 15 20 25 30
–0,01 Time (min)
FIGURE 22.12 One-step CIEF of AGP. PAA coated-capillary 27 cm (effective length 7 cm) × 50 µm i.d.
Voltage: –20 kV. Temperature: 20◦ C, Detection 280 nm. Sample mixture: 5.6 M urea, 1.7% (v/v) TEMED,
9.7 % (v/v) ampholytes in the following distribution of pH ranges: 3–10, 2.5–5, 2–4, 3–5 (2:2:3:3) and 1 mg/mL
AGP in CIEF gel. (a) anolyte: 91 mM H3 PO4 in CIEF gel, catholyte: 20 mM NaOH, (b) anolyte: 91 mM H3 PO4
in CIEF gel, catholyte: 20 mM NaOH titrated to pH 11.85 with H3 PO4 .
usually under conditions involving the use of borate or phosphate salts and SDS as the micellar
agent. Factors, such as borate/phosphate and SDS concentrations, and the nature and pH of the
buffer can be modified to optimize resolution. RNase B was separated into five peaks, each one of
them corresponding to one of the Man5 to Man9 oligomannose structures, in a fused-silica capillary
with a buffer consisting of sodium phosphate, SDS, and sodium tetraborate [113]. For recombinant
human interferon-γ (rhIFN-γ), over 30 species were resolved giving rise to three groups of peaks,
PG1, PG2, and PG3, as shown in Figure 22.13 [114]. The increase in the concentrations of the
borate buffer and SDS led to a concomitant increase in separation efficiency and migration times.
The comparison between phosphate buffers at pH 6.5 and 7.5 and borate buffers at pH 8.5, at a
constant SDS concentration and at buffer concentrations chosen to generate similar currents, showed
that the best resolution was obtained with borate buffers. At a higher pH (9.5) and a lower borate
concentration than that used at pH 8.5, higher current, slower separation, and lower efficiency were
observed. The better resolution obtained with the borate buffer pH 8.5 was likely due to the effect
of the complexation between borate anions and diols mentioned earlier. It has been hypothesized
that the early migrating glycoforms would be those with the larger glycan structures and that this
should be taken into account when designing a separation method or when assigning peaks to specific
glycoforms. The reasoning behind this hypothesis is that at high borate concentrations, sugar residues
would be extensively complexed and that the resulting negative charge, when added to that from sialic
acid residues, would cause repulsion between SDS micelles and the borate-complexed glycoprotein.
According to the authors, the repulsion intensity would depend on the glycan size. The same study
reported the possibility of separating the glycoforms of RNAse B and HRP, while the separation of
BSF led to the separation of only one major peak and three minor, broad ones, and no separation
at all of AGP glycoforms. According to the authors, the lack of separation for proteins glycosylated
at multiple sites such as BSF and AGP could be due to the fact that the heterogeneity of individual
oligosaccharides does not give rise to measurable differences in carbohydrate contents.
22.4.2.2.4 Capillary Electrophoresis in SDS

In conventional polyacrylamide gel electrophoresis (PAGE) as well as in its capillary counterpart
CE-SDS, proteins form stable complexes with SDS and are separated on the basis of their size. The
separation of glycoproteins is known, however, to deviate from the linear relationship of migration
versus MW observed with nonglycosylated proteins. This anomalous behavior arises from the fact
that carbohydrate moieties bind with much lower amounts of SDS than expected for a corresponding
(a)
4.5 5.0 5.5 6.0 6.5
(b)
7 8 9 10
(c)
10 12 14 16 18 20
(d) PG1
PG2
PG3
15 20 25 30 35 40
FIGURE 22.13 Optimization of the separation of IFN-γ glycoforms by MEKC. Borate/SDS buffer at the
following concentrations: (a) 40 mM borate, 10 mM SDS; (b) 40 mM borate, 100 mM SDS; (c) 400 mM borate,
10 mM SDS; (d) 400 mM borate, 100 mM SDS. Fused-silica column 57 cm × 50 µm i.d. Injection 5 s of
1 mg/mL protein in 50 mM borate, 50 mM SDS, pH 8.5. Voltage: 22 kV. In (d), the main groups are designated
PG1, PG2, and PG3 in order of migration. (From James, D.C. et al., Anal. Biochem., 222, 315, 1994. With
permission.)
peptide of similar MW. Advantages were taken from this fact in order to differentiate glycosylated
fetuin from the nonglycosylated protein produced by enzymatic hydrolysis [115]. The release of the
carbohydrate moiety gives rise to a decrease in size that translates into an increase in the charge-
to-mass ratio since the charge imparted by SDS to the carbohydrate moiety is low compared with
that imparted by the interaction of the surfactant with the peptide core. The resulting differences in
mobility make it possible to monitor the deglycosylation reaction. Similarly, rhATIIIα was separated
from rhATIIIβ by CE-SDS due to the lack of glycosylation at Asp 135 in the latter [116].
2,3,4 3,4
(c)
(a) (b)
3,4
(d)
1
2 34 (e)
1 2 2
1 2 1 4
1 3
14 17 14 17 14 17 14 17 14 17
FIGURE 22.14 Changes in the migration times of OVA glycoforms with LCA concentration. Capillary: linear
PAA-coated 58 cm × 50 µm i.d; Buffer: 100 mM phosphate pH 6.8, (A) not containing LCA, (b–e) containing
0.4, 0.8, 1.4, and 2.0 mg/mL LCA, respectively. Voltage: 20 kV; Detection: 214 nm. (From Uegaki, K. et al.,
Anal. Biochem., 309, 269, 2002. With permission.)
22.4.2.2.5 Affinity Capillary Electrophoresis

Glycoform separations by ACE have been carried out by taking advantage of the known interactions
between lectins and carbohydrates. The “partial filling technique” in which only a portion of the
capillary is filled with an affinity ligand present in the separation buffer before injecting the sample
was shown to be useful for separating AGP glycoforms. The separation of two peaks based on the
dependence of the strength of the AGP-concanavalin A (Con A) interaction as a function of the content
in biantennary glycans was achieved [117]. Differential interactions between lectins and glycans are
useful not only to perform glycoform separations but also to estimate the values of the apparent
association constant between the lectin and each of the separated glycoform. PB-coated and PAA-
coated capillaries completely filled with a buffer containing the lectin Lens culinaris agglutinin
(LCA) were shown to be suitable for separating glycoforms of RNase B and OVA, respectively,
and for calculating the affinity constants without a prior separation of the glycoforms outside the
capillary. Figure 22.14 shows the separation obtained for OVA glycoforms and the influence of the
LCA concentration [118].
22.4.2.2.6 Capillary Isotachophoresis

Few studies on the use of the CITP electrophoretic mode have been reported for the separation
of protein glycoforms. The separation of AGP isoforms by CIEF mentioned previously [109] can
be considered to involve an isotachophoretic mechanism taking into account the concurrence of
focusing and chemical mobilization [119]. Preparative CITP has been used to fractionate rhEPO
with a leading electrolyte containing chloride as anion and BTP as counter ion, and a terminating
electrolyte containing glycine as anion and BTP as counter ion. Under these conditions, the fractions
contained mixtures of glycoforms enriched in the predominant glycoform of each fraction. The
process was monitored by CZE [95].
22.4.2.3 Reproducibility Problems

With respect to the issue of reproducibility, the separation of protein glycoforms by CE is faced
with the same kind of problems than for any other protein. However, the consequences of a lack of
adequate precision can have a detrimental impact due to the fact that differences in the charge/mass
ratio or pI among isoforms are so small that the different peaks usually migrate very close to one
another, making it sometimes impossible to compare samples. For example, peaks of rhBMP-2
separated by CZE on the basis of a single mannose residue between adjacent peaks differ in migration
times by only 0.15–0.22 min or 0.01 × 10−4 cm2 V−1 s−1 in their respective mobility [31]. This can
give rise to unreliable comparisons of samples since the difference in migration time between two
consecutive peaks can be in the same range as the migration time dispersion for a given peak [38]. For
instance, this problem made it impossible to correctly compare a reference hCG sample with a sample
obtained from the urine of a patient with metastatic choriocarcinoma [56]. Solutions to this type of
problems require a two-pronged approach. First, experimental factors affecting reproducibility must
be carefully controlled. Once this is achieved, reliable migration parameters must be used for band
assignment.
Some experimental factors such as buffer preparation are usually well controlled. However, there
are other variables to which not enough attention is paid. One of these, which is of key importance for
achieving good intra- and interday repeatability, is capillary conditioning. All of the steps performed
for the conditioning of brand-new capillaries as well as the washing steps carried out between runs
and before the storage of capillaries are decisive in achieving good precision. This may help in
explaining the low reproducibility reported for some methods [17], while the same methods when
applied to the same analytes showed highly reproducible results in other studies [36,38,90].
Another factor that has a large influence on migration time reproducibility, especially for coated
capillaries, is capillary aging. Buffer components that can alter the coating need to be identified and
avoided. However, even when using the appropriate buffers, capillaries gradually degrade and they
should be discarded if the resolution achieved is no longer within easily verifiable limits or if peak
assignments cannot be reliably performed.
Once experimental factors are controlled as much as possible there are several ways to minimize
the problem of migration time reproducibility through the use of migration parameters. One of the
more common approaches is the use of internal standards from which it is possible to calculate the
migration time of each isoform relative to that of the internal standard. This approach has been
used in the analysis of several glycoprotein isoforms [38,39,109,120]. For example, in the CIEF
separation of rhEPO glycoforms, the intraday migration time reproducibility was improved from an
RSD value of around 1.8–0.3% when using an internal standard [39]. In another approach and using
the CZE analysis of rhEPO as a model, a statistical program was developed in order to choose, among
others, the best migration parameter to correctly assign peaks and reliably compare different samples.
The program estimated the probability of correct assignment for each migration parameter [38] and
was used for the comparison of rhEPO from different samples [38,43] and for the comparison of
electrophoretic profiles of AGP purified from healthy donors and cancer patients [84]. In the latter
publication, it was shown that the electrophoretic mobility of each AGP peak and the migration
time of each peak relative to the migration time of the EOF marker were more effective migration
parameters than the migration time of each peak.
22.4.3 DETECTION AND IDENTIFICATION OF FORMS OF GLYCOPROTEINS

22.4.3.1 UV Detection
The UV absorbance detector is one of the most widely used systems in CE. Detection takes place
in the same separation capillary, avoiding band broadening of the separated compounds. Since the
internal diameter of the capillary is very small (25–100 µm), the detection pathlength is very short
and the concentration sensitivity is typically 10–20 times poorer than that obtained with the same
type of detector in other separation techniques (e.g., HPLC).
Glycoproteins, similar to nonglycosylated ones, can be monitored at 280 nm at which wavelength
the aromatic residues of the polypeptide chain, tryptophan, tyrosine, and, to a lesser extent, pheny-
lalanine give relatively good absorption. However, the detection limit at this wavelength remains
in the low mg/mL range and instead, detection is usually accomplished at 200 nm where proteins
present 50- to 100-fold greater absorptivity. At this wavelength, sensitivities in the µg/mL range can
be obtained.
While this level of sensitivity may be enough for the quality control of glycoproteins, it is, in
most cases, insufficient for the analysis of these proteins in clinical samples. Strategies based on
sample concentration or protein derivatization can sometimes be used in these cases and some aspects
have already been discussed in other parts of this chapter and will only be considered briefly in this
section.
22.4.3.2 Laser-Induced Fluorescence Detection

For glycoproteins present at low sample concentrations (generally lower than 10−6 M), such as in
clinical samples, fluorescence detection will likely be one of the detection methods to be improved in
the future for analysis of these proteins by CE. It should be recalled that, due to the small size of the
detection volume in on-column monitoring (usually 5–10 nL), high optical power (>1 mW) from the
light source should reach the detection window of the capillary for intense fluorescence excitation
[121]. This is more efficiently achieved using a laser as a source of light than a ultraviolet-visible
(UV-VIS) lamp and confirmed by practical experience of the fluorescence detection of proteins and
glycoproteins by CE.
LIF detection of glycoproteins is not an easy task. Most analytes do not produce fluorescence at
the wavelength of the emission of the most frequently used lasers (those with light emission in the
400–600 nm range) and, therefore, native fluorescence cannot be obtained with these light sources.
Derivatization with a fluorescent reagent is a useful strategy for LIF detection when native fluores-
cence is unavailable. Although several fluorogenic compounds with good chemical and spectroscopic
characteristics have been described in the literature [122], this approach suffers from several prob-
lems when the analyte is a glycoprotein, including the formation of multiple reaction products [123]
and the difficulties associated with derivatization of proteins at low concentration (below 10−8 M). In
the following discussion, a few approaches reported in the literature to make LIF detection feasible
for glycoproteins are presented.
Glycoproteins can be directly detected by native fluorescence owing to the fluorescence emission
of the tryptophan and tyrosine residues; however, this requires the use of laser with emission in the UV
region of the electromagnetic spectrum. This approach has been scarcely used [124] for glycoproteins,
probably due to the high cost of such lasers and the expertise required for their maintenance. In the
past few years, however, pulsed diode-pumped solid-state (DPSS) lasers with emission in the UV
region (266 nm) have become affordable and this has opened the way to a more universal native LIF
detection of glycoproteins.
The covalent derivatization of a glycoprotein with a fluorescent agent has been shown to be useful
in some cases. As mentioned earlier, the main difficulty with this type of approach is the formation
of multiple products due to the fact that each lysine residue in the peptide chain can react to a
different extent with the derivatizing agent. As a result, protein molecules having a variable number
of covalently bonded fluorescent tags are produced, resulting in a variable number of species with
similar MW (the fluorescent moiety has a small MW in relation to that of the protein) but with
different electric charges. In these cases, analysis by CZE or CIEF may give rise to broad or even
multiple peaks for the same protein. On the other hand, when the separation mode is CE-SDS, narrow
peaks with good efficiency are obtained and make LIF a more suitable technique for this separation
mode.
Using this approach, a CE-SDS method with LIF detection has been developed and vali-
dated as a quality control procedure for the purity determination of a recombinant mAb [125].
5-Carboxytetramethylrhodamine succinimidyl ester was used as fluorescent reagent and the optimal
conditions for derivatization of the nonreduced and reduced (reacted with dithiothreitol) mAb and
their respective degradation products were determined. It is worth mentioning that for reduced sam-
ples, the proposed method was able to separate the fraction corresponding to the nonglycosylated
fragment of the mAb from the glycosylated one. The commercial LIF instrumentation (argon ion
laser, λexc 488 nm, λem 560 nm) provided excellent limits of detection (around 10 ng/mL) for the
intact antibody and its degradation products.
One approach to overcome the formation of multiple derivatives and the difficulties associ-
ated with derivatization of proteins at low concentrations was proposed by the group of Dovichi
[126,127]. They developed a derivatization procedure for proteins with the fluorogenic reagent,
5 -furoylquinoline-3-carboxaldehyde (FQ), that provided good sensitivity and minimized the band
broadening caused by the variable labeling from the reaction of FQ with lysine residues. The deriva-
tization could be carried out before the sample injection or inside the separation column. For the
latter, two successive short plugs, one containing the sample and the other containing the reagent,
were injected into the column; both plugs were mixed inside the capillary (due to their differences
in electrophoretic mobility) and allowed to react. After derivatization, either inside or outside the
capillary, the protein separation took place in a buffer containing an-alkylsulfate tensioactive (e.g.,
sodium dodecylsulfate or sodium pentylsulfate) at submicellar concentrations, which masked the
charge differences introduced in the protein by the multiple labeling. The method provided good
results in terms of sensitivity, with an assay detection limit of 10−10 M for OVA (the authors define
the assay detection limit as the minimum amount of protein that can be successively derivatized and
detected), and band broadening (around 20,000 theoretical plates) (Figure 22.15). Limitations of
the technique included variable protein sensitivity due to the different number of lysine residues and
the possibility of masking differences in mobility of the protein glycoforms from the added tensioac-
tive [128]. It should be noted that the reported sensitivity was achieved with postcolumn sheath-flow
LIF detection; sensitivity using a commercial on-column LIF detector was about 100 times poorer.
Migration time (sec)

140 150 160 170 180 190 200
6 70 x 10–3
60
5
Fluorescence signal (v )
50
4
Absorbance (au)
40
3
(c)
30
2
(b) (d) 20
1
10
(a)
0 0
180 200 220 240 260 280
Migration time (sec)
FIGURE 22.15 Comparison of ovalbumin analyzed under various conditions. Running buffer: (a–c) 25 mM
tricine, pH 8.0; (d) 25 mM tricine +5 mM SDS. (a) FQ-labelled ovalbumin, 15 s reaction, LIF detection;
(b) unlabelled ovalbumin, UV adbsorbance detection; (c) FQ-labelled ovalbumin, 10 min reaction. LIF detec-
tion. Note that electropherogram B is plotted versus top x-axis and right y-axis. (From Pinto, D.M. et al., Anal.
Affinity binding is another good alternative for fluorescent detection of glycoproteins. This kind of
binding involves several types of interactions between the glycoproteins and the fluorescent reagent
(excluding the covalent one) and include electrostatic, van der Waals, hydrophobic, and hydrogen-
bonding forces. A large choice of fluorescently labeled biomolecules can be used in this approach.
The preparation of a homogeneously labeled biomolecule is in some cases a labor-intensive task
involving chromatographic purification of the probe. Bornemann et al. [129] have used affinity
derivatization with LIF monitoring to improve the selective detection of rhEPO. They first prepared
a monomeric antigen-binding fragment (Fab) from the mAb 5F12 that bonded to a conformationally
independent epitope of the N-terminal region of human EPO. This fragment was then labeled with the
fluorescent dye, Alexa Fluor 488, yielding a mixture of labeled products (doubly and singly labeled
and unlabeled) from which a homogeneous fraction was prepared by HPLC. This fraction was used
as affinity reagent in CIEF with LIF detection (argon ion laser λexc 488 nm, λem 520 nm) for the
detection of less than 100 pmol of rhEPO. Although this detection limit is far from the sensitivity
necessary for the determination of this hormone in serum or urine, the authors believe that further
refinement of the labeled Fab fragment and decrease of sample volume should permit the detection
of smaller amounts of EPO.
Postcolumn derivatization with a fluorescent reagent is another alternative. However, this
approach requires the selection of a fast derivatization reaction to achieve good sensitivity and a
careful design of the postcolumn reactor that minimizes the loss in resolution achieved in the separa-
tion capillary. To this end, affinity interactions of some biomolecules, such as antibodies, with other
glycoproteins are especially useful. That has been the case reported by Kelly and Lee [130] who
have developed a CE method with online postcapillary affinity LIF detection for the monitoring and
quantification of microheterogeneities of a mouse antihuman follicle-stimulating hormone mAb in
samples containing culture medium without further purification. The fluorogenic reagent, consisting
in fragment B (BF) of protein A conjugated with fluorescein, was mixed with the effluent of the
CE column using a homemade postcolumn capillary reactor without causing large band broaden-
ing. Affinity binding between the antibody variants and the BF–fluorescein occurred in the reaction
capillary and the complexes were monitored by the LIF detector (argon ion laser, λexc 488 nm, λem
515 nm) placed in the same capillary. By binding the BF–fluorescein fragment to the antibody, the
emission of the fluorescein moiety was enhanced, so that the mAb peaks were detected with a low
fluorescence background. The detection of up to five peaks was reported and the authors speculated
that this could be due to the presence of different isoforms of the antibodies differing in the sialic
acid content of their oligosaccharides.
22.4.3.3 Mass Spectrometry

Recent progress made in the development of MS ionization techniques has enabled its applica-
tion to the analysis of intact glycoproteins. The most widely used ionization modes are ESI and
matrix-assisted laser desorption ionization (MALDI). When considering the gains in sensitivity
when compared with conventional UV detection, the structural information that can be derived from
its use and the possibility of increasing the selectivity in the selected ion monitoring (SIM) mode,
it is not surprising that MS has become a widely used detection system for coupling to CE. In this
section, only those systems where the online coupling of CE to mass spectrometry (CE-MS), and
for which MS is used as the detection (and information) system for the CE separation of glycoforms
will be considered.
Owing to the high heterogeneity and structural complexity of glycoproteins, a prior separation
step is generally necessary to obtain qualitatively and quantitatively useful results. Otherwise, only
broad signals are obtained. Despite the compatibility of CE with ESI-MS and the relatively easy cou-
pling of the two techniques, appropriate separation buffers, usually consisting of volatile substances,
must be employed. This is one of the more difficult issues to the coupling of CE to MS where buffer
electrolytes compatibility with MS involves choosing buffers that do not inhibit ionization and that,
at the same time, provide reasonable glycoform separation [30]. Formic acid was used as the separa-
tion buffer in a pioneering work aimed at the characterization of glycoforms [61]. Analysis required
constructing a CE–MS interface that could be adapted to a commercial CE cartridge. Under these
separation conditions, the CE-MS system provided analysis of intact RNase B and HRP, a more
complex glycoprotein in which the microheterogeneity arises from glycosylation at two residues
and variability in the carbohydrate composition at each site. Figure 22.16 shows the CE-UV and the
CE-MS electropherograms obtained for RNase B.
Although volatile buffers are usually employed, successful analyses of intact glycoproteins have
also been achieved by using low concentrations of nonvolatile compounds in separation buffers with
low ionic strength [131]. Using PAA-coated capillaries and a separation buffer consisting of 50 mM
β-alanine adjusted to pH 3.5 with acetic acid, two high mannose-containing proteins, RNAse B
and rhBMP-2, were analyzed by CE-MS employing a home-built CE instrument. Reporter ions for
carbohydrates generated by in-source fragmentation for the glycoforms separated by CE were used
to provide information about the degree of glycosylation of the different isoforms.
The presence of urea in the separation buffer was found to be needed to accomplish the separation
of antithrombin III (ATIII) glycoforms from a commercial sample purified from plasma [132]. In
order to make the CZE separation step compatible with the ESI-MS ion trap, special precautions
about the decoupling of the needle were needed to minimize the amount of urea introduced into the
spray chamber. Although the resolution was lower than that obtained by CZE-UV, it was enough to
allow deconvolution of the multiply charged peaks, even for minor isoforms. It was also possible to
distinguish isoforms differing in fucosylation for which small differences in mass for species having
similar pIs would not make it possible to be differentiated on the basis of electrophoretic mobilities.
The precision of the molecular mass determination was better than 1%. Assignment of particular
glycan structures to isoforms was aided by the known canonical structures of the main isoforms of
alpha- and beta-ATIII.
Successful characterization of protein glycoforms has been reported using MS-friendly condi-
tions on capillaries with dynamic coatings that generate low-EOF [70]. An acidic BGE containing
an organic additive was used (1 M acetic acid/20% methanol at pH 2.4). Under these condi-
tions, partial separation of glycoforms was attained for several glycoproteins including AGP and
BSF [71].
By using ionene-coated capillaries and an acetate buffer at pH 4.8, baseline resolution of
the three main isoforms of rhEPO and a comparison of rhEPO and uEPO were obtained by
CE-ESI-MS [62].
The development of CZE-ESI-MS methods using an orthogonal accelerated time-of-flight (TOF)
mass spectrometer has allowed distinguishing glycosylation differences between a reference prepa-
ration of rhEPO and a commercially available product sold for research purposes [42]. Membrane
concentrators were used to obtain sample concentrations at the required level followed by separa-
tion on PB-coated capillaries (the amine coating effectively reverses the EOF). The concentration
step provided the additional advantage of removing small MW excipients (e.g., salts, polysorbate),
which are known to interfere in the ionization step. Despite the lack of complete resolution in the
CE separation step, the MS dimension provided additional resolution and enabled the identification
of intact glycoforms varying in the number of HexHexNAc residues, sialic acids, and even the iden-
tification of fucosylated, acetylated, and oxidized glycoforms. A total of about 135 isoforms were
distinguished in the rhEPO BRP of the EP. A possible composition of the recombinant protein was
speculated taking into account the information provided by the CE-MS system and that obtained by
other methods. The same method allowed the characterization of the two pharmaceutical products,
epoetin-α and epoetin-β [96]. Ultimately, the two products could be distinguished on the basis of
the presence of two additional basic (i.e., glycoforms containing less sialic acids) glycoforms in
epoetin-β.
RNase B glycoforms
A
0.05
Man5
0.04
0.03 Man6
0.02 Man7
Man8
RNase A
0.01 Man9
0.00
10 15 20 25
B (a) 100 Man

Man5
75
Rel. int. (%)
Man6 Asn34
50
Man7 GlcNAC
25 Man8
Man9 RNase A
0
25 30 35 40 45
Time (min)
(b) 100 9+
Mr:14901
75
Rel. int. (%)
8+
50 10+
25 11+
(c) 100
9+ Mr:15063
75
Rel. int. (%)
50
10+
25 8+
11+
(d) 100 9+ Mr:13680

75
Rel. int. (%)
8+
50
10+
25
0
1400 1600 1800 2000
m/z
FIGURE 22.16 CZE of RNase B. (A) Fused-silica capillary 80 cm × 50 µm i.d. coated with a solution
of polybrene and ethylenglycol; Buffer containing 2M formic acid; injection of 6 pmol of RNase B; UV
detection. (B) Fused-silica capillary 110 cm × 50 µm i.d. coated with a solution of polybrene and ethyleng-
lycol; Buffer containing 2 M formic acid; injection of 6 pmol of RNase B; ESMS detection. Sheath flow
5 µL/min of an aqueous solution of 0.2% formic acid and methanol. (a) Total ion electropherogram for the
full mass scan acquisition (m/z 1300–2000). (b) Extracted mass spectra for peaks migrating at 35.2 min,
(c) at 35.6 min, and (d) at 40.8 min. The calculated molecular mass is shown on the right corner of each
spectrum. (From Kelly, J.F. et al., J. Chromatogr. A, 720, 409, 1996. With permission.)
A comparison of rhEPO by CZE-ESI-MS using capillaries that provided reversal of the EOF
(PB coating) or suppressed EOF (UltraTrol Dynamic PreCoat LN) showed, as indicated in a previous
section of this chapter, that increased resolution was obtained with the suppressed EOF capillaries
(see Figure 22.3) [30]. The improved separation allowed detection of low-level glycoforms. The
observed mass differences were attributed to differences in the content of sialic acids, hexoses, and
N-acetylhexosamines. In addition, the high mass resolution of the system allowed the detection of
comigrating substances differing by 42 and 16 Da corresponding to acetylated and oxidized variants,
respectively. Similarly to the results obtained for rhEPO, better glycoform resolution was obtained
for bovine fetuin and bovine alpha 1-acid glycoprotein standards on capillaries with suppressed
EOF [71].
Despite a level of sensitivity about 100 times lower than that obtained by CE-UV, inductively cou-
pled plasma mass spectrometry (ICP-MS) coupled to CZE has been shown to separate and distinguish
between isoforms containing 3, 4, and 5 sialic acid residues in standard Tf samples [133].
For more complete details on the use of CE-MS for glycoproteins some recent reviews have also
appeared [134–138].
22.4.3.4 Other Detection Modes

Conductivity detection has been used in a CITP instrument for the fractionation of glycoforms of
rhEPO mentioned previously in this chapter [95]. The preparative CITP instrument is composed of
two columns made out of a fluorinated ethylene copolymer, one acting as a separation column and the
other one as a trapping column. Both columns are provided with their respective conductivity detec-
tors. That of the first one served, in addition to monitoring the CITP separation, as a proper timing in
the switching of the columns. The other one in the second column provided the isotacopherograms
used in the control of the fraction collector placed at the end of this second column, and equipped
with a collection valve. Although the sensitivity of conductivity detectors was poor for the moni-
toring of proteins, in this case sensitivity was not an issue since 100 µg of rhEPO were injected in
each run.

22.5.1 CHARACTERIZATION AND QUALITY ASSESSMENT OF BIOLOGICALS AND
BIOPHARMACEUTICALS
22.5.1.1 General Considerations
The development of physicochemical methods of analysis based on high-resolution techniques such
as CE has had a significant impact on fields of activities dealing with complex biomolecules such
as glycoproteins. While undoubtedly variable in magnitude, this impact has been considerable in
the development of foods, pharmaceuticals, vaccines, blood and blood products, diagnostics, and
for regulatory authorities. In many of these areas, the establishment of more sensitive, precise, and
selective methods of analysis has particular relevance since manufacturers are required to provide
scientific evidence with regard to major tenets that guide the development of regulatory requirements
in most jurisdictions: that products entering the market must be safe and of high quality. More
specifically, the assessment of a product through the verification of its identity, the quantitative
evaluation of its purity, its stability, or the consistency of its manufacturing is of prime importance. For
example, considerations regarding the decision to market a pharmaceutical protein resides not only
on the demonstration of its biological activity but also on the demonstration that the manufacturing
process can provide a product consistently, at the same level of quality and purity as that used
for clinical trials. While the application of these specifications does not affect all areas universally
(e.g., purity evaluation may not be a significant factor for a diagnostic product), they nevertheless
are now generally well adhered to an industry and applied at various stages of the manufacturing
process. Data generated are then incorporated in the preparation of documentation to be submitted
to regulatory authorities for evaluation. In certain instances, harmonized guidelines relating to test
procedures have been developed. This is the case for biotechnological and biological products where
an international initiative, the International Conference on Harmonization (ICH), has enshrined these
principles in a series of guidelines of test procedures for the quality evaluation (Q5A–Q5E) and
specifications (Q6B) of these products [139].
Traditionally, complex biomolecules such as glycoproteins have been assessed using conven-
tional gel electrophoresis methods. However, these methods are laborious to carry out, cannot be
automated, and do not consistently provide quantitative results. In contrast, CE has been demon-
strated to provide rapid, automated, and quantitative results. In addition, it is now often used as a
complementary technique to the more established HPLC in several fields of activities. This progress
has led to the recognition by pharmacopeial authorities in Japan, the United States, and Europe of CE
for the generation of data relating to product identification, assay or tests for related protein impu-
rities, and to its use by industry in various other aspects of the development process including drug
substance and drug product characterization, in-process monitoring, and stability determination. It is
also one of the recommended physicochemical techniques of the ICH Q6B guideline on specifications
for biotechnological/biological products mentioned earlier. Underlying this recognition is the neces-
sity to devise validated methods based on widely accepted analytical parameters. Typically, method
validation involves the determination of one or more analytical parameters (e.g., accuracy, precision,
linearity range, limit of detection, limit of quantitation, robustness, specificity) to demonstrate that
the method performs according to its intended purpose. The degree with which a method requires
validation varies according to the procedure being implemented as shown in Table 22.1 derived from
the ICH guidelines Q2(R1). For example, while an identification test requires only that specificity
be demonstrated, the quantitative evaluation of impurities necessitates that all characteristics be
determined with the exception of the limit of detection.
The coming of age of CE as a recognized analytical method has been marked by its introduction
into pharmacopeial monographs. Given its improved separation and quantification capabilities, a
CZE method for rhEPO glycoform separation was found suitable to replace conventional IEF as
an identification test and, as previously mentioned, has been incorporated into the EP monograph
for rhEPO concentrated solution (i.e., the bulk substance solution before formulation) [35]. Using
TABLE 22.1
ICH Q2(R1) Recommendations for Method Validation of Analytical Test
Procedures
Type of Analytical Procedure Identification Testing for Impurities Assay (Content/Potency)
Characteristics Quantitation/Limit
Accuracy − +/− +
Precision
Repeatability − +/− +
Interm. precision − +/− +
Specificity + +/+ +
Detection limit − −/+ −
Quantitation limit − +/− −
Linearity − +/− +
Range − +/− +
Source: http://www.ich.org/LOB/media/MEDIA417.pdf, Accessed November 2006.

conditions adapted from the method of Watson and Yao [52], rhEPO glycoforms are separated on a
bare fused silica with a BGE consisting of 0.01 M tricine, 0.01 M sodium chloride, 0.01 M sodium
acetate, 7 M urea, and 2.5 mM putrescine at pH 5.55. The method system suitability specifies several
required criteria including the separation of the BRP into a number of well-resolved peaks [34] as
shown in Figure 22.17. For this particular preparation, the rhEPO BRP batch 1, eight glycoforms
are separated with nearly baseline resolution and the order of migration follows the increasing
total number of terminal sialic acid residues present on the carbohydrate chains. In other words,
glycoforms with the lowest total number of sialic acids migrate first and those with the highest
number migrate last. It should be noted here that the BRP was a mixture prepared from known
quantities of two commercialized rhEPO forms, rhEPO-α and rhEPO-β. Characterization of these
two products had shown that they had different glycoform profiles and distribution, with rhEPO-β
containing higher amounts of the more basic glycoforms (i.e., those with lower total amounts of sialic
acid residues). The use of a mixture of products for the BRP allowed the preparation of a unified
monograph appropriate to analyze both forms of rhEPO that have essentially identical biological
activities. Specifications on the distribution of glycoforms as a percentage of the total content in the
test solution are provided and constitute an integral part of the identity test (Table 22.2). (At present,
a new standard, rhEPO BRP batch 2, has been established with a slightly different electropherogram
to that shown in Figure 22.17).
Somatropin, the recombinant version of the 22 kDa form of human growth hormone, is another
important biopharmaceutical for which a CZE method has recently been adopted for inclusion
into the EP somatropin monographs. Although not a glycoprotein, it is mentioned here to exem-
plify the increased use of CE for the specifications of proteins. In this case, the test is for the
quantitative assessment of charged variants [140]. It is expected that CE will become a regu-
larly used technique in pharmacopeial monographs of biopharmaceuticals in view of its usefulness
for the characterization of these structurally complex molecules. In addition, the patent expira-
tion date for several biopharmaceuticals has either passed or is fast approaching and this will
undoubtedly lead to several manufacturers wishing to enter this growing market to produce similar
products.
5
6
2
1 8
0 20 40 60 min
1. isoform 1 3. isoform 3 5. isoform 5 7. isoform 7
2. isoform 2 4. isoform 4 6. isoform 6 8. isoform 8
FIGURE 22.17 Electropherogram of the European Pharmacopoeia erythropoietin biological reference

preparation batch 1. (From Bristow, A. and Charton, E., Pharmaeuropa, 11, 290, 1999. With permission.)
TABLE 22.2
Ranges of Glycoforms Content Specified
in the EP Monograph for Erythropoietin
Isoform Content (Percent)a
1 0–15
2 0–15
3 5–20
4 10–35
5 15–40
6 10–35
7 0–20
8 0–15
a Ranges for the current EP standard, rhEPO BRP batch
2 have been modified (see Behr-Gross, M.-E., Daas, A.

and Bristow, A., Pharmeuropa Bio, 2004, 1, 23, 2004).
Source: Erythropoietin concentrated solution,
01/2005:1316, European Pharmacopoeia 5th edition,
published by EDQM, June 2004.
From a strategic perspective, product characterization is paramount to the development of iden-

tity, purity, or assay procedures since it involves the determination of many of the structural and
physicochemical properties of the product under study. Consequently, product characterization is, by
necessity, first in the series of test procedures to be developed. For glycoproteins, in addition to the
more basic structural properties inherent to proteins such as amino acid sequence and higher-order
structural elements (i.e., secondary, tertiary, and quaternary structures), in-depth characterization of
carbohydrate-mediated properties are required.
In the following sections, an overview of the applications of CE with respect to the guiding prin-
ciples and analytical parameters described earlier will be presented. For more in-depth discussions
on specific aspects or individual glycoproteins, excellent reviews cited in the Section 22.2 can be
read.
22.5.1.2 Product Characterization and Identity Testing

As previously mentioned, glycoproteins exist as mixtures of closely related species that differ in their
glycosylation patterns. These differences are often the result of both compositional and sequence
variations of the glycan chains. Moreover, the biological activity of glycoproteins is frequently
linked to the presence of these carbohydrates and, consequently, the characterization of glycoprotein
microheterogeneity represents one of the more challenging tasks for the characterization and identity
testing. In both cases, high specificity is required since it is expected that the method will differentiate
between the active ingredient, variants, impurities, and contaminants. Identity tests typically involve
the determination of one or more intrinsic properties of a molecule, whether physicochemical,
biological, and/or immunological and are often carried out by comparison with a reference standard.
Apart from performing a simple identity test through coanalysis of the substrate and the reference
standard, qualitative and quantitative information derived from CE experiments with respect to
specific structural characteristics such as carbohydrate-mediated glycoform profiles, pI, or MW
determination can be useful for product characterization and identity testing.
22.5.1.2.1 Recombinant Sialoglycoproteins

Several important biopharmaceuticals that received market authorization around the world in the past
decades are recombinant glycoproteins that include such successful products as EPOs (rhEPO-α and
-β), darbepoetin-α (also called NESP), rhtPA, interferons (rhIFN-β1, -γ), and colony-stimulating
factors (rhGCSF, recombinant human granulocyte macrophage colony-stimulating factor, rhGM-
CSF). They are used for the management of diseases not otherwise well treated with small molecule
pharmaceuticals. Unequivocally, it is advances in molecular biology and recombinant DNA tech-
nology that are largely responsible for the availability of these molecules in large quantities, thus
allowing biopharmaceutical companies to introduce them in the markets. Despite these advances,
protein glycosylation remains a crucial factor that requires a thorough assessment especially when
considering that the glycosylation of recombinant proteins is inherently heterogeneous and rarely
identical to that found in humans, a situation that has the potential to result in problems linked to
immunogenicity.
The ubiquitous occurrence of sialic acid residues on glycan chains of recombinant glycoproteins
has provided an ideal tool for analysis by CE by taking advantage of the presence of the ionizable
carboxylic acid group. As indicated in previous sections, both CZE and CIEF have been shown to
be particularly sensitive to differences in net charge and, as such, have been widely applicable to
characterize sialic acid-mediated glycoprotein profiles. However, as it will become evident in the
remainder of this section, the biggest limitation to the application of CE methods is undoubtedly the
complex nature of sialoglycoproteins, especially when more than one glycosylation site is present
in the molecule. In such cases, the number of possible combinations and permutations of individual
carbohydrate residues make it difficult to even anticipate separating all glycoforms.
Studies of relatively simple sialoglycoproteins, that is, containing a single glycosylation site,
were among the first examples of the use of CE for characterization of intact glycoproteins. RhGCSF
stimulates progenitor cell proliferation, differentiation, and activation of neutrophils and features a
single glycosylation site at Thr-133. It was separated into two glycoforms under basic (pH 7–9) [54] or
acidic (pH 2.5) conditions [141]. In the latter case, rhGCSF was analyzed in a preparation containing
HSA and the addition of HPMC prevented adsorption to the capillary walls and facilitated detection at
low pH. In both cases, the order of migration occurred in a predictable fashion according to increasing
number of sialic acid residues. Both methods could distinguish between the nonglycosylated and
glycosylated products.
A highly efficient CZE method capable of resolving in excess of 30 glycoforms of a sim-
ple sialoglycoprotein, a recombinant 24 kDa glycoprotein containing a single glycosylation
site, has been reported by Berkowitz et al. [142]. Using a bare fused-silica capillary with tri-
ethanolamine/phosphoric acid, pH 2.5 as BGE, the 30 glycoforms were separated on a 27 cm capillary
in less than 15 min (Figure 22.18). Lengthening the capillary to 77 cm resolved over 60 peaks in less
than 30 min. Partial validation of the method was carried out by examining day-to-day reproducibil-
ity and sample matrix effect. For the latter, it became apparent that increased resolution was obtained
for samples dissolved in low ionic buffer and low pH. While the effect of low sample ionic strength
compared with the BGE is well known and is due to stacking effects, the low sample pH effect was
explained in terms of the initial movement of the protein, which if injected from a sample at high pH
was toward the anode and, consequently, against the stacking effect since the protein has a relatively
high pI (6.5 < pI < 8.0). Using a kinetic study of the enzymatic removal of terminal sialic acids
in conjunction with fractionation of the glycoprotein by anion exchange, they were able to identify
several glycoforms according to their antennarity and sialic acid content as shown by the shaded
areas in Figure 22.18. Although not all of the peaks were identified, this approach allowed to demon-
strate that parts of the observed microheterogeneity were due to the presence of large amounts of
deamidated variants (i.e., >30%). In addition to the qualitative assessment of the product, the results
when taken collectively allowed to derive quantitative expressions for determination of the content
in deamidated forms, nonglycosylated forms and monoglycosylated forms, biantennary forms, and
(a) (b)
BiNA2 (Area #3)

0.14 0.14 0.14 0.14
0.12 0.12 0.12 0.12
0.10 0.10 0.10 0.10
0.08 0.08
A.U.
A.U.
0.08 0.08
0.06 0.06 0.06 0.06

BiNA1 (Area #2)
BiNA0 (Area #1)
0.04 0.04 0.04 0.04

Non-and monosialylated
Glycans (Area #4)
0.02 0.02 0.02 0.02
0.00 0.00 0.00 0.00

7.5 10.0 7.5 10.0
Min Min
FIGURE 22.18 CZE separation of 24 kDa sialoglycoproteins with peak region assignments (shaded areas). In
(a): BiNA0, biantennary oligosaccharides with no terminal sialic acid residues; BiNA1, biantennary oligosac-
charides with one sialic acid terminal residue; and BiNA2, biantennary oligosaccharides with two terminal sialic
acid residues. In (b): nonglycosylated protein and monosialylated glycoforms (area #4). Separation conditions:
fused-silica capillary, 50 µm × 27 cm capillary, 100 mM triethanolamine–phosphoric acid, pH 2.5, 10 kV,
25 ◦ C and UV detection at 214 nm. (From Berkowitz, S.A. et al., J. Chromatogr. A, 1079, 254, 2005. With
permission.)
finally, sialylation level of biantennary forms. The method was also adequate to monitor the effects
of changes in culture conditions such as host strains or growth conditions.
The analysis of complex sialoglycoproteins, that is, containing multiple glycosylation sites has
been the subject of many reports. One such example is rhtPA, an important biopharmaceutical used
in the treatment of myocardial infarction, which contains three N-glycosylated sites. Among several
articles published on CE separation conditions, Thorne et al. [77] reported the separation of rhtPA
into no less than eight glycoforms by CIEF. The method showed high reproducibility and precision
for migration times and good total recovery of rhtPA from the capillary was demonstrated. CZE
conditions using amines as additives to the BGE were also used to separate numerous major and
minor glycoforms. In addition, they demonstrated the use of CE-SDS as a potential purity test for
rhtPA Type I and Type II, which differ in the level of glycosylation site-occupancy.
RhEPO is another complex sialoglycoprotein, with three N-linked and one O-linked glycosyla-
tion sites, that has been extensively characterized by CZE and cIEF as described in previous sections.
The comparison of rhEPO and uEPO has been reported on uncoated capillaries [40] and on capillaries
dynamically coated with ionene [62]. Although increases in resolution and sensitivity are desirable,
and the samples of uEPO corresponded to purified standards, these works showed the potential of
discriminating between exogenous and endogenous EPO.
CE was also used to fractionate rhEPO glycoforms as briefly discussed previously. Madajova
et al. [95] developed a preparative CITP method, operating in a discontinuous mode, which enabled
the separation of 100 µg of rhEPO into six fractions in 30 min. The number of collected fractions
corresponded approximately to the number of glycoforms detected in the original product by CZE.
Subsequent analysis of the collected fractions by CZE showed that several glycoforms had been
substantially enriched by 10- to 100-fold in some cases. Such an approach may provide a desirable
way to access specific glycoform levels in relation to their biological activity or for the evaluation
of low glycoprotein levels in biological matrices.
Recombinant human deoxyribonuclease (rhDNAse) is a complex sialoglycoprotein whose het-

erogeneity is due in part to the presence of phosphorylated mannose residues. CE studies of this
acidic (3.5 < pI < 4.5) phosphosialoglycoprotein that contains two glycosylation sites have been
reported [143,144]. CZE methods were developed under both acidic (pH 4.8) and basic (pH 8.0)
conditions. The addition of calcium ions significantly improved resolution at both pH as a conse-
quence of interactions with calcium-binding sites that stabilize the protein and make it more resistant
to proteases. The pH-dependent calcium binding provided a means to distinguish between acidic and
neutral glycoforms that were identified by a two-dimensional (2D) investigation of neuraminidase
and alkaline phosphatase–protein digestions. The acidic pH resolved acidic charge heterogeneity
and the basic pH discriminated neutral heterogeneity.
22.5.1.2.2 Naturally Occurring Sialoglycoproteins

The need for characterization and identity determination of naturally occurring glycoproteins has been
one of the driving forces behind the development of CE methods. Naturally occurring glycoproteins
are highly heterogeneous and sometimes present a formidable challenge. Only a few examples of
approaches to their characterization will be discussed in this section.
ATIII is a glycoprotein that plays a key role in the inhibition of blood coagulation. Despite
the availability of a recombinant version, plasma-derived ATIII has continued to be the subject of
recent characterization studies by CE. Two isoform groups isolated from plasma-derived ATIII that
differ in the number of glycosylation sites, ATIII-α (glycosylated at four asparagines) and ATIII-β
(glycosylated at three asparagines), were studied by a combination of techniques that included gel
IEF, 2D-gel electrophoresis, CIEF, and CZE [145]. The two isoform groups were separated by CZE
as shown in Figure 22.19. With a lower glycosylation site-occupancy and, consequently less sialic
acid residues, the major component of ATIII-β (peak A in Figure 22.19a and d) predictably migrated
before the ATIII-α components. The quantitative analysis indicated a content of about 70% ATIII-α
main isoform and about 6.6% of ATIII-β in good agreement with published data. The pI values of
ATIII determined by CIEF using an internal calibration were in fair agreement with values of the
main isoforms measured by 2D-gel electrophoresis. The CZE method provided a definite advantage
in terms of quantification and accuracy. In another report, the same group further characterized the
separated isoforms by online coupling with MS [132].
The commercially available glycoprotein, OVA, has been extensively used as a model protein
for approaches to glycoform separations by CE since the very early years of the technique. Recent
studies on OVA have focused on novel separation conditions that included the effect of new additives
to the BGE and new dynamic coatings [87,146,147] as well as conditions suitable for online coupling
to MS [148]. As indicated in a previous section, an innovative separation approach based on the free-
solution (i.e., without immobilization) selective affinity of OVA glycoforms to a common protein,
LCA, has been reported [118]. The separation mechanism was explained in terms of the increased
affinity of glycoforms having high-mannose type N-glycans for LCAand, consequently, these species
were more strongly retarded than other glycoforms. A similar type of approach was recently used to
separate AGP glycoforms without biantennary glycans from those having one or more biantennary
glycans [117]. Once again, the separation mechanism appeared to involve greater affinity of the
ligand for high-mannose species.
22.5.1.2.3 Glycoproteins without Sialic Acids

CE separation of glycoforms of proteins where no sialic acid residues are present has been reported.
RhBMP-2 is a protein with high-mannose glycan chains that has been unexpectedly separated under
CZE conditions [31,131]. Glycoforms were separated on the basis of the number of mannoses a
given rhBMP-2 molecule possesses (see Figure 22.8). This type of approach may prove useful
for the qualitative and quantitative monitoring of fermentation processes for glycoproteins without
substantial charge differences.
(a) 214 nm (b) 214 nm

C
15 C
ATIII
15
10 10
mAU D
A mAU B D
E E
B
5
5
0
0
24 26 28 30 32 34 36 24 26 28 30
(c) 214 nm (d) 214 nm
A
20 C
15 A
15
mAU 10
10 mAU
5
5
0 0
21 22 23 24 25 26 27 20 22 24 26 28
FIGURE 22.19 CZE of (a) ATIII, (b) ATIII-α, (c) ATIII-β, and (d) ATIII-(α + β). Separation conditions:
PVA-coated capillary 50 µm i.d. × 64.5 cm (56 cm effective length), 1 M acetic acid containing 4 M urea,
0.1 M sodium chloride added to the aqueous sample solution, 120 kV, UV at 214 nm. (From Kremser, L., et al.,
Cellobiohydrolase I (CBH-I) is a phosphoglycoprotein that contains a catalytic domain with four

N-glycosylation sites and where phosphate groups form phosphodiester links between the anomeric
carbon and the C6 of two consecutive Man residues (i.e., Man1-P-6Man). CIEF was used to separate
the phospho-isoforms of the catalytic domain of CBH-I into four glycoform species according to the
number of phosphate groups, that is, 0, 1, 2, and 3 groups [149]. Since glycosylation of CBH-I is
dependent on the strain and growth medium, the method may be useful for the analysis of cellulase
glycosylation from different strains, different organisms, mutated organisms, or cellulases expressed
in other organisms.
22.5.1.2.4 Monoclonal Antibodies

A number of therapeutic mAbs produced by recombinant DNA technology have been marketed
in recent years. Most mAbs exhibit some heterogeneity due to PTM and degradation occurring
during the manufacturing process and shelf life of the product. As such, they have been the
subject of several studies using CE-based methods and reviewed [150]. Charge heterogeneity is
a parameter commonly monitored for mAbs. Common sources of charge heterogeneity include
sialylation, deamidation, and C-terminal lysine cleavage. Sialylation can vary widely between
mAbs, from being insignificant in some cases to being the major source of charge heterogene-
ity in other cases. It follows that the sialic acid-mediated heterogeneity is commonly determined
from the characterization of the charge heterogeneity before and after enzymatic treatment with
sialidase.
Both CIEF and CZE-based methods have been used to study mAbs. A combination of cation
exchange chromatography and CIEF was used to study human tumor necrosis factor (hTNF) mAb
[151]. Of the four peaks separated by cIEF, three corresponded to C-terminal lysine variants and
the remaining peak to sialic acid-mediated charge differences. Similarly, CIEF was used to deter-
mine the sialic acidmediated heterogeneity of a IgG1 mAb that contained N-linked glycosylation
in the Fab region at Asn56, with approximately half of the biantennary glycans being sialylated
[152]. Using an approach based on the well-known formation of complexes between borate and
diols, the carbohydrate heterogeneity of a mAb was demonstrated by CZE [153]. Three partially
resolved peaks were observed after optimization of the conditions to 150 mM borate, pH 9.4, on a
fused-silica capillary. The method was further used to monitor batch-to-batch consistency and for
stability testing.
In addition to providing molecular mass information, CE-SDS has been used for the detection
and quantification of low levels of unglycosylated heavy chain (HC) in mAbs. The optimization
and validation of generic and quantitative CE-SDS procedures with LIF detection was reported
[125,154]. An alkylation step was incorporated to decrease thermally induced fragmentation of
nonreduced, labeled mAb samples. The unglycosylated variant was detected under reducing
conditions where light chain (LC) and HC are separated (Figure 22.20). A collaborative study
report involving multiple laboratories in different organizations has demonstrated the robustness
of CE-SDS without labeling and using UV detection for assessing mAbs and the unglycosylated
variant [155].
HC
100 LC
90
80
NGHC
70
RFU
Free dye
60
Incomplete
50 reduction
With IAM
40
30
With IAA
20
10
No alkylation
0
0 2 4 6 8 10 12 14
FIGURE 22.20 CE-SDS separations of reduced labeled mAb samples in the presence of different alkylating
agents: NGHC, nonglycosylated heavy chain; LC, light chain; HC, heavy chain; IAA, iodoacetic acid; IAM,
iodoacetamide. Separation conditions: uncoated fused-silica capillary 50 µm i.d. × 30 cm (effective length
19.4 cm), both anode and cathode buffers were the Bio-Rad SDS running buffer, samples were injected at a
constant electric field of 417 V/cm for 15 s and electrophoresed at 625 V/cm (21.2 µA) at 20◦ C, detection
was performed with LIF using a 3.5-mW argon ion laser, 488 nm excitation, 560 ± 20 nm emission. (From
Salas-Solano, O. et al., Anal. Chem., 78, 6583, 2006. With permission.)
22.5.1.3 In-Process Monitoring and Product Consistency

Studies of recombinant DNA technology processes have clearly demonstrated that PTM such as
glycosylation are not constant throughout culture and that many factors must be examined to obtain
and maintain the required glycosylation. They include choice of host cell, genetic engineering of
glycan processing, or control of bioprocess parameters such as culture environment, method of
cell culture, and culture time [156]. Furthermore, even after a process has been well established,
changes may be required from time to time, a situation that may also have an impact on PTM.
Examination of pre- and postchange products is then required to ascertain the absence of unwanted
changes to the product quality [157]. It follows that monitoring of glycosylation at all stages of a
bioprocess has become a necessity to ensure product quality and batch-to-batch product consistency.
Practical considerations to the application of an analytical technique for in-process monitoring and
product consistency assessment range from the necessity of the method to have high accuracy and
sensitivity, to provide high reproducibility and be relatively simple in its application, to be robust
enough to sustain minor changes in sample matrix, to be specific and quantitative, and to permit
fast development and analysis time. CE has become widely applicable at various stages of the
manufacturing process due to its already mentioned advantages regarding the flexibility provided by
the various separation modes that can be applied, the low sample volume required, and its capability
for automation and fast analysis.
22.5.1.3.1 In-Process Monitoring

RhIFN-γ is, as mentioned in a previous section of this chapter, a sialoglycoprotein with two
N-linked glycosylation sites at Asn25 and Asn97, which has been shown to be heterogeneous in
the number of species resulting from variable glycosylation site-occupancy (nonglycosylated: 0N;
monoglycosylated at Asn25: 1N; and diglycosylated at Asn25 and Asn97: 2N) as well as from vari-
ability in the number of terminal sialic acids [114,134,158]. In combination with CIEF, a method
based on MEKC conditions discussed previously (see Section 4.2.2.3 and Figure 22.13) was used
to monitor rhIFN-γ glycosylation during perfused fluidized-bed production [159]. Glycosylation
site-occupancy was monitored by MEKC. A plot of the relative amounts of each of 0N, 1N, and 2N
provided evidence of constant levels of glycosylation site-occupancy over the course of the process.
In addition, the sialic acid mediated isoform content was monitored by CIEF using a commercial kit
over a pI range of 3–10. No less than 11 major variants were detected between pI values of 3.4 and 6.4
(Figure 22.21b). After enzymatic desialylation of this mixture, a decrease in the number of variants
and, as expected, a major shift to higher pI values was obtained (Figure 22.21a), indicating that most
of the observed heterogeneity resulted from sialic acid variability. The remaining heterogeneity was
ascribed to C-terminal truncated variants. The quantitative analysis of sialylated variants indicated
a sharp decrease in the mean pI after 200 h of perfusion culture (Figure 22.21c). This suggested
that sialylation is increased during the perfusion process to reach a constant level afterward. On the
other hand, examination by MEKC of glycosylation site-occupancy for a stirred-tank batch culture
showed that the process did not provide constant levels, but gradually declined throughout the fer-
mentation process. They also demonstrated that this decline was at the expense of the 2N variant and
that there were generally higher proportions of the doubly glycosylated (2N) rhIFN-γ glycoform and
lower proportions of the nonglycosylated (0N) variant in the perfused fluidized-bed process com-
pared with stirred-tank culture, that is, the rhIFN-γ protein was more heavily glycosylated during
perfusion culture.
Similar MEKC conditions were used to monitor rhEPO glycosylation during continuous culture
of Chinese hamster ovary (CHO) cells in a fluidized-bed bioreactor [160]. While individual glyco-
forms were not separated under these conditions, clear differences in both peak shapes and migration
times were observed when compared with unglycosylated EPO, suggesting that adequate glycosy-
lation was obtained. Analysis of the product by 2D-electrophoresis provided further evidence that
the required isoforms had been obtained.
0.018
(a) (b)
9,6 2,9
4,2
9,5
0.012
8,2
9,3
0.006 3,8 5,7

4,5
3,3 3,9
6,4 3,73,5
3,4
0.000
10 15 20 25 30 35 40
5.2 (c)
5.1
Weighted mean pI
5.0
4.9
4.8
4.7
4.6
100 200 300 400 500
Culture time (h)
FIGURE 22.21 Monitoring of rh IFN-γ microheterogeneity by capillary isoelectric focusing. (a) rhIFN-γ
desialylated with neuraminidase from A. ureafaciens. (b) Typical electropherogram of immunoaffinity-purified
rhIFN-γ from perfusion culture of CHO cells. (c) The weighted mean pI of rhIFN-γ variants secreted by CHO
cells during perfusion culture. (From Goldman, M. H. et al., Biotech. Bioeng. 60, 596, 1998. With permission.)
The influence of changing conditions for the fermentation process of rhATIII using a combination
of liquid chromatography and CZE has been reported [161]. While the CZE method only partially
resolved glycoforms, it provided quantification within 2 min. The use of a 2D–HPLC/CZE design
provided additional sensitivity and resolution and quantitative results in 5 min.
Although not developed specifically for glycoproteins, a rapid CE-SDS method for the in-process
monitoring of fermentation, hydridoma cell cultivation, and purification has been reported [162]. It
was carried out in a total analysis time of less than 5 min.
22.5.1.3.2 Product Consistency

The ability to detect differences of product quality is a key component of the assessment of any
manufacturing process. As seen in the preceding section, biological processes are complex and
many factors influence the quality of the resulting product. In addition, manufacturing processes
are rarely static and more often than not involve continual refinement to improve yields, to obtain
higher quality products, or to implement new technologies. In these cases, manufacturing changes are
implemented and the significance of these changes on the product consistency needs to be assessed.
In addition to examples already described in previous sections, a CIEF method was used to
compare the isoform profiles of IgG2a obtained under different conditions of osmolality and CO2
partial pressure (pCO2 ) in growth media [163]. Significant increases in the pIs of the major peaks as
well as in the number of peaks were observed with increases in osmolality and pCO2 (Figure 22.22).
The change in pI was not a consequence of differences in sialic acid content (this particular mAb
435 mOsm/kg
375 mOsm/kg
320 mOsm/kg
pI std
8.6 6.4
FIGURE 22.22 Representative CIEF separations showing the effect of osmolality on the pI distribution of
IgG2a produced in serum-containing medium. The pI standard curve is shown at the bottom of the figure for
comparison. A shift in the main pI peaks toward higher pI values with increasing osmolality, as well as an
increase in the number of peaks, especially at higher pI values are observed. (From Schmelzer, A. E. and Miller,
W.M., Biotechnol. Prog., 18, 346, 2002. With permission.)
was not sialylated) and was due to an increase in the level of galactose incorporation, a situation that
indicated that hyperglycosylation was occurring, possibly arising from galactosylation at another
Asn site.
22.5.1.4 Product Comparability and Analysis of Finished Products

The ability to compare products derived from different manufacturing processes or sources is an
important aspect of the regulatory process, for the setting of standards, the evaluation of poten-
tial problems related to one specific product or the assessment of differences linked to product
characteristics such as biological activity. For example, currently throughout the world, there are
biopharmaceuticals for which more than one manufacturer have received marketing authorization
and this situation is likely to expand as the expiration dates of patents for several products are fast
approaching. In addition, some biopharmaceuticals are being used in situations other than those
for which they were marketed, such as their use as performance-enhancing substances by athletes.
In many instances, products can be obtained through a number of sources other than conventional
pharmacies and may, in some instances, be produced by unauthorized manufacturers. In this context,
it has become essential to develop generic methods for the evaluation of multisource products.
22.5.1.4.1 Product Comparability

There are several sources of commercially available OVA, and CZE can be used to distinguish their
glycoform profiles [146]. Two batches of turkey egg ovalbumin (tOVA) showed qualitatively similar
glycoform profiles with some differences in the levels of a number of glycoforms (Figure 22.23a
and b). On the other hand, the glycoform profile of chicken egg ovalbumin (cOVA) (Figure 22.23c)
6
(a)
4
A200 nm × 1000
6
(b)
4
A200 nm × 1000
6
(c)
4
A200 nm × 1000
10 15 20 25 30 35 40 45 10
Time (min)
FIGURE 22.23 Electropherograms of (a) and (b): two batches of turkey ovalbumin and (c): chicken ovalbu-
min. Separation conditions: fused-silica capillary 72 cm (50 cm effective length), 100 mM H3 BO3 /NaOH and
1.8 mM putrescine, pH 8.6, 20 kV, detection at 200 nm, temperature controlled at 30 ◦ C. (From Che, F.Y. et al.,
J. Chromatogr. A, 849, 599, 1999. With permission.)
revealed marked qualitative differences when compared with tOVA with respect to migration times.
These differences were a result of both polypeptide and glycan chains structural variations.
Caseinoglycomacropeptide (CGMP) is a small glycoprotein (MW of approx. 7000 Da) derived
from bovine κ-casein that has been shown to have a variety of biological activities such as inhibition
of gastric secretion, depression of platelet aggregation, growth promoting effect on bifidobacteria,
inhibition of oral Actinomyces adhesion to red blood cell membranes, inhibition of adhesion of
oral Streptococci to saliva-coated hydroxyapatite beads, inhibition of adhesion of Streptococcus
sanguis to human buccal epithelial cells, and inhibition of cholera toxin binding to its receptor. A
CZE method that had been previously reported for the separation of CGMP glycoforms [164] and
for stability monitoring in cosmetic lotions [165] was also used for the assessment of sialylation
levels of commercial batches of CGMP [166]. Results indicated that the quantitative assessment
using CZE were in good agreement with those obtained by enzymatic release of sialic acids or
colorimetric methods. The CZE method was fully validated for the usual analytical criteria (e.g.,
linearity, precision, accuracy, LOD, and LOQ). This approach provided a fast, reliable, precise, and
cost-effective way to compare products.
Another example of product comparability was reported for prostate-specific antigen (PSA), a
single-chain glycoprotein that is used as a biomarker for prostate-related diseases [167]. PSA has
one known PTM, a sialylated biantennary N-linked glycan chain attached to Asn45. It is commer-
cially available from different sources. The glycoform profiles of seven free PSA (fPSA) samples
from several manufacturers, of which two were specialized, enzymatically active PSA (EA-PSA)
and noncomplexing PSA (NC-PSA), were assessed by CZE. Results indicated that PSA samples
could be classified into three distinct groups according to glycoform profiles. Figure 22.24 shows
(a)
2.0 3.0 (b)

Sample 1 Sample 2
Free PSA Free PSA
1.0 1.0
0 0
3 6 9 12 15 18 21 24 3 6 9 12 15 18 21 24
2.0 (c) 2.0 (d)

Sample 3 Sample 4
Free PSA Enzymatically
active
1.0 1.0
0 0
3 6 9 12 15 18 21 24 3 6 9 12 15 18 21 24
FIGURE 22.24 Electropherograms of protein specific antigen (PSA) group I. (a) Sample 1 (f PSA); (b) Sample
2 (f PSA); (c) Sample 3 (f PSA); (d) Sample 4 (EA-PSA). Separation conditions: 50 µm i.d. × 60 cm fused-silica
capillary, 20 mM sodium borate (pH 8.0)/5 mM diaminopropane or 20 mM sodium phosphate (pH 7.0 or 8.0),
25 kV, detection at 214 nm. (From Donohue, M.J. et al., Anal. Biochem., 339, 318, 2005. With permission.)
electropherograms of products from group I that contained 4–7 peaks with one peak dominating and
representing 51–70% of the total fPSA in the sample. This study demonstrated that proprietary purifi-
cation procedures have a major influence on the composition and yield of the number of structural
forms observed in a PSA sample.
22.5.1.4.2 Glycoprotein Drug Products

Methods capable of analyzing the active ingredient in a formulated (i.e., finished) drug product are
useful to measure product integrity and/or its stability as well as for comparison to other products.
However, the approach to their analysis entails careful consideration of a number of specific issues
to drug products such as the low amounts of active ingredient, the presence of large amounts of
excipients, or the potential of modifying the product through manipulation.
(Glyco)protein drug products generally contain low amounts of the active ingredient since the
therapeutic effect can usually be achieved at low concentrations. For example, it is not unusual
to find active ingredient at low microgram quantities in a dosage form. In turn, formulations at
low protein content usually require the addition of large amounts of excipients to enhance product
stability and prevent nonspecific adsorption to the container. Commonly used excipients vary widely
in terms of their chemical nature and include inorganic salts, amino acids, sugars, and surfactants
such as polysorbate or other proteins such as HSA. Typically, isotonic salt preparations are produced
as most of these products are injectables and, consequently, high salt concentrations are present.
Furthermore, many of these excipients may be present simultaneously, leading to complex mixtures.
They have also been shown in some cases to interfere with traditional assay methodologies such as
UV or high-pressure liquid chromatography (HPLC). The choice of CE as a separation technique to
be applied to the analysis of protein drug products resides in its ability to provide sufficient selectivity
to separate the active ingredient from interfering species.
Studies aiming the direct analysis of finished products with no or minimal prefractionation or
manipulations in order to minimize potential loss of product or the generation of product artifacts
have been reported. As mentioned previously, a CZE method using an amine-coated capillary to sep-
arate rhEPO glycoforms in drug products containing large amounts of HSA and inorganic salts has
been reported [41]. The addition of nickel ions to the high ionic strength BGE (200 mM phosphate,
pH 4.0) selectively altered the mobility of HSA and allowed the separation of the otherwise comi-
grating proteins (Figure 22.6). Dosage forms containing HSA/rhEPO at a ratio of 250:1 w/w could
be analyzed under these conditions. The method was validated and provided satisfactory results for
active ingredient assay, quantification of glycoforms, and comparison of products from different
manufacturers. The usefulness of the method was further extended to assess manufacturing changes
and to compare rhEPO products formulated with different excipients as shown in Figure 22.25 for
different rhEPO-α and rhEPO-β products formulated with HSA or polysorbate [168]. As described
in a previous section, a similar comparison of finished rhEPO products was carried out following
the immunochromatographic removal of HSA from formulations (see Section 4.1) and enabled the
comparison of glycoform profiles of products from different manufacturers and from the BRP of the
EP formulated with excipients of low-MW and with HSA [43].
The biophysical properties of two commercially available rhEPO-α products, Epogen and Eprex,
both of which are produced under similar conditions by different manufacturers, were studied [169].
Among many techniques used in this study, the analysis by CZE showed that the products had
similar glycoform profiles in terms of isoform type and distribution, an indication that the products
had similar sialic acid contents.
CZE was also used for the analysis of two preparations of rhGCSF, Granocyte, a glycosylated
product formulated with large amounts of HSA, and Gran300, a nonglycosylated product [170].
The two products were compared with a bulk rhGCSF preparation that had been obtained from
Escherichia coli (nonglycosylated). Analysis with 50 mM Tricine, 20 mM NaCl, 2.5 mM DAB, pH
8.0, as BGE showed the separation of two glycoforms in Granocyte that corresponded to monosia-
lylated and disialylated glycoforms. As expected, the nonglycosylated product, Gran300, showed
0.005
0.004 0.004
0.003 0.003
0.002 0.002
A.U.
A.U.
0.001 0.001
0.000 0.000
−0.001 −0.001
−0.002 −0.002
−0.003 −0.003
8 10 12 14 16 18 20 22 24 26 28 30
Min
FIGURE 22.25 Typical electropherograms of rhEPO drug products (from top to bottom): rhEPO-α formulated
with HSA (product I), rhEPO-α formulated with polysorbate 80 (product II), rhEPO-α formulated with HSA
(product III), and rhEPO-β formulated with polysorbate 20 (product IV). Separation conditions: Beckman eCAP
amine capillary, 200 mM sodium phosphate/1 mM nickel chloride pH 4.0, –15 kV (75 µA), UV detection at
200 nm. (From Girard, M. et al., Presented at the 7th International symposium on capillary electrophoresis in
the biotechnology and pharmaceutical industries, Montreal, August 2005.)
a single peak with a shorter migration time than that of the two glycosylated counterparts and that
corresponded to that of the bulk rhGCSF from E. coli.
22.5.1.5 Determination of Biological Activity/Potency

The determination of the biological activity of biopharmaceuticals by physicochemical methods
has long been a desirable goal of manufacturers, regulatory authorities, and policy makers alike,
driven in part by the ability to replace costly and often imprecise biological assay testing using
animals. There are currently a few examples of biopharmaceuticals for which this has occurred.
Recombinant insulin and recombinant somatropin (human growth hormone) are two such products,
both of which can be described as relatively simple, well-characterized proteins. For both of these
two nonglycosylated proteins, physicochemical tests have replaced animal-based bioassays for batch
release of bulk products (for somatropin see [140]). In-depth physicochemical characterizations as
well as the establishment of a direct correlation with the in vivo bioassay were carried out.
For more complex products such as glycoproteins, difficulties in obtaining in-depth characteri-
zation of the glycosylation-mediated microheterogeneity have limited the development of this type
of approach. However, in a few cases, reports have shown that there exists a correlation between
glycoform profiles and biological activity. This is the case for recombinant follicle stimulating hor-
mone (rhFSH) where the CZE-derived glycoform profile has been shown to correlate with the in vivo
bioassay in a study directed at predicting the biological potency of several preparations [171]. Using
CZE conditions consisting of a fused-silica capillary and 100 mM borate, 5 mM DAP at pH 8.9 [172],
where rhFSH isoforms migrated as a broad peak between two internal standards, sucrose and the
dipeptide Lys-Asp, four highly purified rhFSH preparations differing in their isoform composition
and biological potencies were analyzed (Figure 22.26a). The inclusion of internal standards enabled
the accurate determination of the median migration time (tm) of the rhFSH peak in each of the four
(a)
Neutral marker
FSH FSH
A 200 nm (AU)
A Iys-asp C
FSH
Sucrose
0 20 40 60 0 20 40 60
FSH FSH
A 200 nm (AU)
B D
0 20 40 60 0 20 40 60
Migration time (min) Migration time (min)
(b) Anode
3.50
4.15
4.55
5.20
pl FSH FSH pl FSH FSH pl Cathode

markers D C markers B A markers
2.5–6.5 3–10 2.5–6.5
FIGURE 22.26 (a) Electropherograms of rhFSH preparations A–D with different isoform compositions and
biological potencies. Separation conditions: fused-silica capillary i.d.: 20 µm × 77 cm (effective length 70.6
cm), 100 mM borate, 5 mM diaminopropane (DAP) at pH 8.9, 20 kV (approx. 3.5 µA), 28 ◦ C, detection at 200
nm. (b) Isoelectric focusing electropherograms of preparations A–D. (From Storring, P.L. et al., Biologicals,
preparations; tm correlated directly with the biological potency on the basis that an increase in the
content of highly sialylated (the more acidic) isoforms led to a corresponding increase in biological
activity. The same four preparations were also analyzed by gel IEF (Figure 22.26b) and the predicted
biological potencies were derived and compared with those obtained by CZE. In both cases it was
found that the methods were sufficiently accurate, precise, and robust to predict the bioactivity of
batches of rhFSH when used in conjunction with a standard preparation.
22.5.2 CE OF ISOFORMS OF INTACT GLYCOPROTEINS IN THE CLINICAL FIELD

As mentioned in the introduction to this chapter, changes in the glycosylation of proteins have
been widely related to pathophysiological changes in an individual [1–4]. For instance, a common
phenotypic alteration in malignant cells is the transformation of their glycosylation [173]. CE is an

attractive technique for the study of glycoproteins with a clinical interest, due to its well-known
quantitative results and speed of analysis [21]. However, not many works dealing with this subject
have been published. There are many challenges in the application of CE to glycoproteins of clinical
interest. One of the problems is that, as for many other important molecules, these glycoproteins
often exist at low concentrations in biological matrices where other proteins are usually present in
large amounts and interfere in the analysis. Thus, a sample preparation step, sometimes including
concentration, is required. Moreover, increasing the detection sensitivity of CE for the analysis of
glycoproteins is one of the challenges reviewed in this chapter. Another pitfall of the technique is
that it is not easy to develop a single method for the analysis of most of the proteins as it is the case
for other analytical techniques such as SDS–PAGE. On the other hand, CE presents the advantage of
its high resolving power that enables the study of slight modifications in protein composition such
as for glycoproteins. In addition, its small sample volume requirements make it appropriate for the
analysis of clinical samples. This section is devoted to a discussion of CE reports on the separation
of isoforms of glycoproteins with a clinical interest.
22.5.2.1 Transferrin
The isoforms of Tf, one of the major proteins in serum, have been widely analyzed by CE as alcohol
abuse markers. Tf is a glycoprotein synthesized mainly in the liver, which consists, in a single
polypeptide chain, of 679 amino acid residues and two N-linked complex-type oligosaccharide
chains. It has an important role in iron metabolism and it also acts as a growth factor [174,175].
Human Tf presents several heterogeneities due to its iron content, its genetic polymorphism, and its
carbohydrate moieties. With respect to iron content, a Tf molecule can be apotransferrin, monoferric
transferrin, or diferric transferrin [174,175]. The influence of this heterogeneity on CE separations
of glycoforms can be easily overcome by iron saturation of the sample. On the other hand, more than
38 genetic Tf variants, attributable to substitutions at one or more amino acids, have been described,
although only four of them show a prevalence of ≥1% [176]. The common type of Tf is called C-type
and most Caucasian individuals express this allele [177]. With respect to carbohydrate heterogeneity,
each carbohydrate chain can be bi-, tri-, or tetra-antennary and each antenna possesses a terminal
sialic acid residue. Therefore, nine sialoforms (from zero to eight sialic acid residues per molecule)
may be present in serum, resulting in pI variations of 0.1 pH unit/residue. The most abundant Tf-
form in normal sera is tetrasialo-Tf and it has a pI of 5.4. The other forms have higher and lower
pI values [178].
Variations in the heterogeneity of Tf are induced by several pathological and physiological condi-
tions, such as rheumatoid arthritis (RA), congenital disorders of glycosylation (CDG), or pregnancy
[175,179], even though variations in the Tf pattern due to alcohol abuse has attracted the highest
interest [176,178,180]. The first to report abnormal Tf heterogeneity in cerebrospinal fluid (CSF)
associated with alcoholic cerebellar degeneration were Stibler and Kjellin [181] and further studies
determined the Tf abnormality in serum of alcohol abusers [182,183]. The variation in Tf het-
erogeneity in sera from alcohol abusers corresponds to increased amounts of asialo-, monosialo-,
and disialotransferrrin (generically called carbohydrate-deficient transferrin or CDT). At first, the
increased CDT was only attributed to changes in the number of sialic acid residues. However, it
was later learnt that the difference was more complex and included changes in some of the neu-
tral carbohydrates [178]. CDT has become the most specific marker for chronic alcohol abuse
[66,177,178,184].
The first studies dealing with the quantitative determination of CDT in sera from alcohol abusers
were performed by conventional IEF with immunological detection. This method can easily detect
genetic variants and has the power to resolve individual isoforms. However, it is laborious and time
consuming [178,184]. Since then, a number of CDT methods of analysis, such as anion-exchange
chromatography, chromatofocusing, HPLC, and CZE have been developed [176,184].
At the present time, there are several commercial kits for the quantification of CDT based on
the fractionation of CDT and non-CDT serum variants by anion-exchange chromatography with
immunochemical detection [176]. The main characteristic of almost all of these micro columns is
that they do not separate CDT sialoforms and they are quantified as a whole. Therefore, the CDT
forms may contain an undetermined amount of trisialo-Tf, which is under debate as to whether
it should be considered a CDT form, and that introduces an error in the determination [185,186].
Another drawback of the anion-exchange chromatographic methods is that they fail to detect genetic
variants, which may give rise to incorrect determination of CDT or even false-positive results when
sera from individuals with genetic Tf variants are analyzed [177,186]. Some authors have claimed
that the separation of individual Tf sialoforms is a more accurate procedure than CDT quantification
as a whole [68,69,185,186]. One of these commercial tests was compared with HPLC and CZE
methods to conclude that the data obtained needed to be systematically confirmed by HPLC or
CZE [187]. In spite of these drawbacks, these kits are still convenient for routine analysis due to
their speed.
HPLC and CZE methods have been developed to resolve all of the Tf sialoforms present in
serum. An HPLC method based on anion-exchange chromatography with direct detection at 420 nm
was developed by Jeppsson et al. [188] to individually separate Tf sialoforms in about 16 min. The
serum required to be saturated with iron and lipoproteins to be precipitated. Several authors followed
the method with minor changes and improvements [186,189,190].
22.5.2.1.1 CZE of Carbohydrate-Deficient Transferrin (CDT)

In general, the CZE identification of Tf sialoforms in the electropherograms has been performed
in three ways: first, using anti-Tf to perform immunosubtraction and then comparing the electro-
pherograms of the Tf sample before and after being immunosubtracted; second, the amounts of the
different sialoforms can be monitored by CZE analysis after progressive desialylation of Tf with
neuraminidase treatment of the sample; and third, by identification from the migration times of each
sialoform. With very few exceptions, Tf is iron-saturated before performing CE to eliminate the
iron-binding-mediated heterogeneity.
22.5.2.1.1.1 Sample Preparation Commercially available Tf standards were first separated by

CIEF and CZE by Kilar and Hjerten [79,97] and Oda and Landers [55]. Iourin et al. [120] purified Tf
from sera of one donor and of two CDG type I syndrome by consecutive precipitation of sera with
rivanol and ammonium sulfate for their analysis by CZE. On the other hand, immunopurification
of Tf from sera was used as sample preparation before CE analysis [89,191]. While these were
pioneering attempts at the use of CZE to study CDT as a disease marker, the laborious and time-
consuming sample preparation steps required were not conducive to their use in routine settings.
Sample preparation of Tf from serum became easier after the development of a procedure that
involved only iron saturation and dilution of the serum [192]. This sample preparation method has
since then been adopted by most authors with only slight modifications [66–68,177,185,193–201].
In another approach, interfering proteins, such as immunoglobulins, can be eliminated from serum
with protein A to enhance CDT detection [202]. Another interesting approach to sample preparation
is the direct injection of serum into the capillary, carrying out complexation of Tf with iron during
the electrophoretic separation [203].
22.5.2.1.1.2 CZE Separation With respect to CZE performance, many approaches have been
published for the separation of Tf sialoforms. Two groups of methods can be distinguished.
In the first one, studies were conducted with a commercial buffer system (CEofix-CDT kit;
Analisis, Namur, Belgium), as already mentioned in a previous section, widely accepted and
used, sometimes with slight modifications to the protocol proposed by the manufacturer [66–
69,177,185,195,197–199,203]. This commercial reagent kit offers the advantage of interlabora-
tory standardization. In the second group of studies, methods developed in-house were used
[55,79,89,97,120,185,191,192,196,200–202]. To the best of our knowledge, all of the reported meth-

ods were carried out at basic pH (around 8.5) so that Tf is negatively charged to prevent protein
adsorption to the capillary wall.
The commercial buffer system is based on the dynamic double coating of an uncoated, bare
fused-silica capillary. A first buffer (the so-called initiator) is used to coat the capillary wall with a
polycation and a second buffer containing a polyanion adds the second layer. The separation buffer is
usually borate-based. Between injections, the capillary is rinsed with NaOH so that the coating layers
are eliminated after each analysis. This procedure ensures a constant EOF and increases the negative
charges on the capillary wall for preventing protein adsorption and increasing the speed of analysis
(usually between 6 and 13 min, depending on the modifications of the method) [185,203]. Although
in one of the first studies with this commercial system, disialo- and trisialo-Tf could not be resolved
[203], modifications of the initial method, which included among others offline iron-saturation and
dilution of serum, increased capillary length, and injection of an SDS solution before the injection
of the sample, led to baseline resolution of tri, tetra, penta, and hexasialo-Tf in nonalcoholic donors
and asialo, di, tri, tetra, penta, and hexasialo-Tf in alcoholic patients [177]. It appears that the
SDS plug before sample injection is performed to keep β-lipoprotein peaks out of the area of the
electropherogram of interest. SDS is now incorporated in the FeCl3 solution used for iron saturation
of the commercial reagent set, so it is not necessary to add it separately if the complete reagent kit
is used [69].
In the second group of CZE separations, many different approaches have been published. The
main goal has been to avoid Tf adsorption to the capillary walls. This problem has been overcome by
working with covalently coated capillaries, in some cases with hydroxyethylcellulose in the running
buffer [79,89,191,202], or silica capillaries dynamically coated with DAB [185,196,200,202], sper-
mine [185], DcBr [55], or diethylentriamine (DETA) [201]. In general, the methods performed in
uncoated capillaries with dynamic coatings are slower than the ones performed with the commercial
buffers. On the other hand, the methods using covalently coated capillaries are as fast (usually in
less than 10 min of analysis time) as the methods developed with the commercial buffers.
Lanz et al. [185] compared two different methods in fused-silica capillaries with alkaline borate
buffers and different dynamic coatings (DAB and spermine, respectively) and with the double-
coating method based on commercially available buffers. The latter method provided faster and
more reproducible results. In addition, Tf isoforms peaks were more intense when analyzed with the
latter method.
The same group [66] studied the precision of a method based on the double-coating buffer kit
over a 20-day period. They found out that the method is highly reproducible both in migration times
and peak areas of Tf isoforms. Martello et al. [198] ran an interlaboratory comparison of a method
based on the dynamic double coating of the capillary wall performed with commercially available
reagents. Results from both laboratories showed high correlation.
22.5.2.1.2 Clinical Applications

In general, the study by CZE of CDT as an illness marker is performed by qualitative or quantitative
comparison of CDT between two different groups of participants. To the best of our knowledge,
CDT reference limits for direct classification of samples are not yet routinely used, although some
attempts have been performed. In that sense, not even a common CDT calculation is performed. For
example, area of disialo-Tf as a percentage of tetrasialo-Tf [185,192,201,204], area % of disialo-Tf
in relation to the sum of all Tf-isoforms [191,202], and CDT isoforms (asialo-, monosialo-, and
disialo-Tf) as area % of total Tf isoforms [177,194,203] have been used as CDT measurements.
Lanz and Thormann [205] worked on the establishment of normal CDT reference limits using a CZE
method based on a dynamic double coating of the capillary with commercially available reagent kits.
They found that the reference CDT intervals (in the case of CDT, only the upper reference limit is
important) for a group of 54 individuals with no or moderate alcohol consumption was dependent
on the complete or incomplete separation of disialo- and trisialo-Tf, on the integration approach
and on the applied voltage. The upper reference limit obtained by Lanz et al. [197] was used in a
subsequent study where more than 600 samples were analyzed with a CZE method established as
a routine method. Later on, the reference intervals found by those authors [205] were revised for
a modified version of the commercially available reagent kits [67]. The same samples from the 54
individuals were re-analyzed with the new reagents and the CDT reference limits were found to be
comparable, although slightly smaller, to the old ones so that the same upper reference limit was
valid for the study. The establishment of reliable reference CDT intervals is critical in the sensitivity
(number of false negatives) and specificity (number of false positives) of the method.
22.5.2.1.2.1 Alcohol Abuse Alcohol abuse is an important public health problem, and self-
reporting of alcoholism or alcohol abuse is not reliable. Furthermore, diagnosis on the basis of clinical
symptoms is not easy [197]. Therefore, reliable alcohol markers are required for the diagnosis of
alcoholism and CDT is one of the better markers of alcohol abuse.
Prasad et al. [191] based their CDT calculation of Tf separated by CZE on the ratio of the area
of disialo-Tf (they did not detect asialo-Tf) to the area of total Tf in the sample. On the basis of
this index and with a control population of social drinkers, they established a CDT cutoff value.
This cutoff value was exceeded by the majority of alcohol abusers studied. This CDT measurement
was compared with other alcohol abuse markers (e.g., aspartate amino transferase, alanine amino
transferase, and others) and it was found to be the most specific one.
In another study, Tagliaro et al. [192] found that by comparing the Tf profile of a control group
of 30 participants against a group of 13 alcoholics, the latter showed a significant increase in disialo-
and trisialo-Tf (expressed as percentages of the tetrasialo-Tf peak). They did not detect asialo-Tf in
alcoholic samples. These results are controversial since several studies have claimed that not only
trisialo-Tf is not an alcohol abuse marker [68,195] but also that it interferes with the correct CDT
quantification [204]. On the other hand, there are also favorable opinions to include trisialo-Tf in the
calculation of CDT or, at least, to establish a debate [177,206,207]. Crivellente et al. [193] improved
on their published method [192] and tested it with real samples. Although the presence of asialo-Tf
was suspected to be present in the sera of alcoholic individuals, no confirmation was performed. An
increase in disialo- and trisialo-Tf (expressed as percentage of the tetrasialo-Tf peak) compared with
healthy participants happened in alcoholic individuals.
Another study using CZE for the analysis of CDT as a marker of alcohol abuse was carried
out by Giordano et al. [202]. They did not detect asialo-Tf in sera from alcoholic people, but
they detected a significant increase in the disialo-Tf (related to total Tf) in those samples in com-
parison with control (nonalcoholic subjects) samples. In the same direction, and by establishing
the Tf index as % area of disialo-Tf in relation to tetrasialo-Tf, Lanz et al. [185] found that a
good classification of sera could be performed on the basis of this index increasing in sera of
alcoholic individuals.
Interesting CZE studies on CDT isoforms were reported by Legros et al. [68,69]. They found that
asialo-Tf was missing in teetotalers and was present in 92% of alcohol abusers, and that disialo-Tf was
increased in alcohol abusers. Figure 22.27 shows the comparison of the Tf profiles of an alcoholic and
a teetotaler. Under these conditions, C-reactive protein (CRP) comigrated with monosialo-Tf in the
alcoholics electropherogram. Samples before and after Tf immunosubtraction are also compared in
the figure. They proposed to use the presence of asialo-Tf as alcohol intake marker because it showed
the highest sensitivity and specificity when compared with other CDT isoforms or combination of
CDT isoforms. The study focused on clearly different groups (teetotalers and alcohol abusers). In
their subsequent work [69], they included moderate drinkers in another group. They found that
asialo-Tf was able to discriminate between moderate drinkers and alcohol abusers better than other
CDT measurements tested.
22.5.2.1.2.2 Congenital Disorders of Glycosylation Congenital disorders of glycosyla-

tion (CDG) (previously known as carbohydrate-deficient glycoprotein syndromes) are inherited
P4
P5
P2
P3 P6
P0 1/crp AA
Anti-Tf
P5
P3 P6
P2
TT
Anti-Tf
5.5 6.0
Time (min)
FIGURE 22.27 Comparison between CZE Tf electropherograms of an alcohol abuser (AA) and a teetotaler
(TT). In both cases, anti-Tf antibody was added after the first electrophoretic run. P0: asialo-Tf. 1/CRP:
comigration of monosialo-Tf and CRP; P2: disialo-Tf; P3: trisialo-Tf; P4: tetrasialo-Tf; P5: pentasialo-Tf; P6:
hexasialo-Tf. Analytical conditions: fused-silica capillary 57 cm × 50 µm i.d; buffer: reagents from the CEofix
CDT kit. Capillary was first rinsed with a solution of polycation dissolved in malic acid, pH 4.8; then rinsed
again with a polyanion dissolved in Tris-borate, pH 8.5. Capillary was rinsed with the same buffer for 0.5 min
under low pressure. After 3 s low-pressure injection of SDS, the iron-saturated sample was injected by low
pressure for 2 s; Voltage: 28 kV; Temperature: 40◦ C; Detection: 214 nm. (From Legros, F.J. et al., Clin. Chem.,
conditions that are usually recognized from glycosylation changes of serum proteins [2]. The
hypoglycosylation of different proteins and sometimes of other glycoconjugates leads to several
symptoms, such as mental retardation. IEF measurement of Tf is the main diagnostic test used for
the detection of CDG because these disorders appear to have an influence on its sialylation (and con-
sequently on its pI) [199,208]. There are several studies that have been reported in which the CZE
of Tf has been used to study CDG. An increase in disialo- and asialo-Tf in CDG patients has been
reported by several authors [89,120,196,203]. Carchon et al. [199] demonstrated that patients with
abnormal IEF results and with confirmed CDG could be identified by CZE. However, they warned
about the possibility of finding compounds migrating in the Tf-region and when this is suspected in
a sample, CZE analysis with immunosubtraction should be performed.
22.5.2.1.2.3 Genetic Variants In addition to the common C-type Tf, other variants such as B
and D have been reported. As stated earlier, genetic polymorphism is limited to Caucasians. Most
individuals present CC phenotypes (homozygous for Tf with the C gene variant) and, rarely, CB
phenotypes. West African, African American, and indigenous American populations, for exam-
ple, present higher frequency of the D allele [177]. Certain genetic variants are not detected
by anion-exchange chromatography-based methods and may result in false positives for alcohol
abuse [177,178]. In that sense, CZE is a promising tool for the determination of genetic vari-
ants due to its high resolving power. Several workers have published Tf CZE patterns of different
genetic variants [66,67,177,197,203]. In general, these patterns are characterized by having two
major peaks instead of one (corresponding to the tetrasialo-Tf of both genetic variants). Wuyts
et al. [177] found that their CZE method was able to separate (in an alcoholic carrying a CD
phenotype) D-asialo-, D-disialo-, D-trisialo-, D-tetrasialo-, and D-pentasialo-Tf in addition to C-
tetrasialo-, C-pentasialo-, and C-hexasialo-Tf. In spite of this high resolution power, some of
the D-Tf peaks comigrated with C-CDT forms. To solve this problem for the CDT calculation
for alcohol consumption diagnostic purposes, the authors proposed a CDT calculation adjusted
to CD-phenotypes in order to avoid false positives. Figure 22.28 shows the comparison of Tf
patterns from a nonalcohol consumer homozygous (CC), an alcoholic homozygous (CC), a non-
alcohol consumer heterozygous (CD), and an alcoholic heterozygous (CD). On the other hand,
Lanz et al. [67] analyzed a CD-phenotype alcoholic individual. They detected two peaks for
asialo-Tf and two peaks for disialo-Tf. Since C-disialo-Tf was not baseline resolved, the CDT
calculation was done by taking twice the area of D-disialo-Tf because both peaks had the same
height. The CDT value clearly exceeded the upper CDT reference limit established for normal
nonalcoholic-CC phenotypes.
22.5.2.1.2.4 Transferrin and Cancer A study has been recently performed in which the Tf CZE
profile of a group of cancer patients who consumed alcohol moderately and a group of cancer
patients who were alcoholics has been studied [195]. Their profiles were compared with a group of
participants without cancer and that were teetotalers, moderate drinkers, and alcohol abusers. While
asialo-Tf was present in 95% of alcohol abusers (with and without cancer), trisialo-Tf was higher in
cancer patients. The CZE peak corresponding to this trisialo-Tf was not completely eliminated by Tf
immunosubtraction. Some tests on the remaining peak suggested that it might be a polysaccharide.
22.5.2.1.2.5 Interferences in CDT Determination Usually, Tf analysis by CZE is performed

after the injection of diluted, iron-saturated serum, which is an extremely complex media. There-
fore, possible interferences in the determination of Tf should be taken into account. Generally,
interferences are found after Tf immunosubtraction of the sample and re-analysis. Sometimes,
immunosubtraction of the interference is also performed. Legros et al. [68] detected comigration of
CRP with monosialo-Tf, so this peak was not taken into account in the CDT calculations. CRP was
also identified and migrated before disialo-Tf in some samples analyzed by Lanz et al. [66,197].
In some patients with hepatic disorders, a broad interference, assumed to be due to high levels of
immunoglobulin A, was found under the Tf peaks [197]. Wuyts et al. [203] added antihemoglobin
and anti-C3c (degradation products of complement-C) antibodies to the running buffer to avoid inter-
ferences from these compounds. As described before, SDS was injected into the capillary before
the sample for methods performed with commercial reagent buffers to avoid β-lipoprotein peaks
comigrating with Tf [68,69,177,195]. This modification was later adopted by the manufacturer [69].
Moreover, a compound comigrating with trisialo-Tf (probably, a polysaccharide) in cancer patients
was described [195]. Recently, the commercial reagents for the double-coating CZE method have
been reformulated to avoid undesired interferences, and CRP comigration is no longer interfering,
neither are some paraproteins that interfered in some samples [67].
CC
nonalcoholic
6
3
CC
alcoholic
5
3 6
2
0
D4 C4
CC
nonalcoholic
D5
C5
D3 C6
D4
CC C4
alcoholic
D5
D3 C5
D0 D2 C6
8.0 8.4 8.8 9.2

FIGURE 22.28 Comparison between CZE Tf electropherograms of a healthy nonconsuming (CC nonalco-
holic) and an alcohol-consuming (CC alcoholic) carrier of the homozygote CC-Tf, and a nonconsuming (CD
nonalcoholic) and an alcohol-consuming (CD alcoholic) carrier of the heterozygous CD-Tf. C0 : asialo-C, C2 :
disialo-C, C3 : trisialo-C, C4 : tetrasialo-C, C5 : pentasialo-C, C6 : hexasialo-C-Tf. D2 : disialo-D, D3 : trisialo-D,
D4 : tetrasialo-D, D5 : pentasialo-D-Tf. Analytical conditions: fused-silica capillary 67 cm × 50 µm i.d; buffer:
reagents from the CEofix CDT kit. Capillary was first rinsed with the initiator solution, then rinsed again with
the buffer solution containing a polyanion. A plug of SDS was injected during 5 s before the sample. The
iron-saturated sample was injected for 1 s. Voltage: 28 kV; detection: 214 nm. (From Wutys, B. et al., Clin.
Chem. Lab. Med., 39, 937, 2001. With permission.)
22.5.2.1.3 Concluding Remarks

The CZE analysis of Tf seems to be an alternative to CDT measurements kits, which are prone
to inaccurate diagnostics. Normal CDT ranges are method-dependent and, consequently, results
must be interpreted by taking into account method-specific cutoff values [176,184]. In any case,
several authors support the application of CZE for CDT determination in the clinical laboratory
[191] and claim that it represents an alternative to HPLC and that it should be taken into account
as a reference method for CDT [67]. In comparison to HPLC, CZE methods have the advantages
of easier sample preparation, faster analysis times, higher isoforms resolution, and faster column
reconditioning [197].
22.5.2.2 Alpha-1-Acid Glycoprotein

AGP or orosomucoid is synthesized mainly in the liver. It is an acute-phase protein with an unclear
function, although it is usually accepted as a natural immunomodulatory and anti-inflammatory
agent. It is a 41–43 kDa protein with a pI of 2.8–3.8 with 5 N-complex type glycans. Terminal
sialic acid residues are usually present and contribute to the acidity of the protein [209]. Since it is
expressed by various alleles at two loci, it exhibits genetic polymorphism [210].
Changes in the glycosylation of AGP, as well as in the distribution of genetic AGP forms, have
been related to several pathophysiological states, such as cancer [210–213], RA [214,215], and other
types of inflammation [216–218].
There are several studies that have described the separation of AGP forms by CE
[78,84,87,100,109,117,219,220]. However, only a few authors have carried out CE analysis of AGP
samples from patients in order to study if the technique is suitable for distinguishing the changes
described earlier.
Kinoshita et al. [78] analyzed AGP purified from rat sera from normal and inflammation states.
They separated electrokinetically injected AGP into several nonbaseline resolved peaks in a DB-1
capillary with an acetate buffer (pH 4.1) containing HPMC. They found qualitative differences
between the two samples. The sample from rat sera in the inflammation state had more AGP bands
moving slower than the AGP sample from sera in normal state. The authors speculated about those
bands having decreased sialic acid content. Later on, Kakehi et al. [219] proposed an interesting
approach to the analysis of AGP by CZE. They performed electrokinetic injection of desalted serum
in a DB-1 coated capillary with a running buffer at pH 4.5. They assumed that almost all serum
proteins have pIs above that pH except for AGP, which is negatively charged at that pH. Therefore,
a selective electro-injection of AGP was performed. They obtained nonbaseline resolution of 10
AGP forms. They analyzed two serum samples from patients who had acquired methicillin-resistant
Staphylococcus aureus during hospitalization and one serum sample from a healthy volunteer. How-
ever, they did not find a relationship between the abundance of each glycoform and the clinical data
of patients.
A completely different approach, based on lectin-affinity CE was used by Bergstrom et al. [117]
to separate AGP fractions. As mentioned in a previous section, using Con A as affinity ligand, they
separated AGP into two peaks according to the biantennary content of the glycoforms. They applied
this methodology to two AGP samples from patients with severe RA and one sample from a healthy
donor. They found that the RA samples showed a decrease in the relative peak area of the biantennary
peak compared with normal AGP.
In another study, a CZE method capable of baseline separation of up to 11 AGP bands was
developed [84]. This method, together with a statistical program that allowed to correctly compare
AGP bands from different samples, was used to compare three AGP samples from one healthy donor
and two pools of sera from cancer patients. Results showed that samples from cancer patients had
one extra AGP band with a higher charge to mass ratio, while the AGP profile from the healthy
donor presented one extra band with lower charge to mass ratio. A quantitative study of the samples
demonstrated that there were peak area differences for the AGP forms. The cancer patient samples
presented higher proportion of forms with higher charge to mass ratio and lower proportion of forms
with lower charge to mass ratio when compared with AGP from the healthy donor (Figure 22.29).
The same authors also developed a CIEF method to separate AGP isoforms and applied it to the
analysis of AGP from sera from ovary cancer patients [109].
The studies discussed earlier represent the beginning of research efforts directed toward the
utility of CE analysis of AGP for clinical uses. In all cases, AGP samples from people and animals
with different pathophysiological conditions were analyzed as a proof of concept for the methods
developed. Studies of large populations are needed to make clinical conclusions.
22.5.2.3 Other Glycoproteins

hCG is a glycohormone of 38,000 Da that consists of two noncovalently bound subunits. Both
subunits (α and β, respectively) contain two N-glycans and the β-subunit also contains four
O-glycosylated chains. Its role is to avoid the disintegration of the corpus luteum in the ovary.
The more acidic hCG glycoforms have increased biological activity. In addition, the glycosylation
pattern of this hormone seems to change in patients with trophoblastic disease [56]. As indicated
previously, this glycoprotein (in its native heterodimer form) was separated into eight peaks in an
uncoated capillary with a borate buffer containing DAP [56]. The method was used to analyze a
reference hCG, two samples purified from crude urinary hCG, and one hCG sample from the urine of
a patient with metastatic choriocarcinoma. The four hCG electropherograms had the same number
of forms, but the relative concentration of each one appeared to vary. However, the hCG sample
coming from the cancer patient migrated in a way that did not allow an accurate comparison with the
other samples. Those peaks could be aligned with either the previous or subsequent peak, resulting
in a different quantitative result.
Hiroaka et al. [221] developed a one-step CIEF method for the analysis of CSF proteins with
molecular masses between 10,000 and 50,000 Da. They found differences in the distribution of
lipocalin-type prostaglandin D synthase isoforms (a sialic acid-containing glycoprotein) in the CSF of
patients with certain neurological disorders. It was speculated that the four different peaks attributable
to this protein were due to different numbers of sialic acid residues.
PSA is a glycoprotein that is used as a biomarker for prostate cancer. Its release into blood is
increased during the development of prostate cancer. It has a MW of 28,430 Da and contains one
N-glycosylation site. A fraction of circulating PSA is bound to plasma proteins. The measurement
of unbound PSA increases the diagnosis specificity of total blood PSA [222]. However, these mea-
surements do not take into account the glycosylation pattern of PSA. It was recently stated that when
comparing PSA oligosaccharides from healthy donors with that from a prostate tumor cell line, the
PSA glycosylation pattern may be different [223]. In that case, the use of CZE for the comparative
analysis of PSA glycosylation patterns in samples from healthy persons and prostate cancer patients
would contribute to the study of PSA glycosylation as prostate marker. Donohue et al. [167] obtained
a separation of PSA forms from several manufacturers in 12 min in an uncoated capillary with a
borate buffer containing DAP using UV detection. The commercial PSA samples had been purified
from human seminal fluid. As mentioned earlier in this chapter, an interesting finding of this work
was that the purification procedure influences the PSA pattern.
22.6 CONCLUDING REMARKS AND FUTURE PROSPECTS

As demonstrated in this chapter, CE has clearly become a widely applicable tool for the analysis
of glycoforms of glycoproteins, without which the determination of some significant qualitative
and quantitative information would not have been possible. This is aptly exemplified from some
of the earliest work in the field on Tf, work which has culminated in the use of CE in clinical
settings, to a more recent application of CE-MS from which over 130 glycoforms of rhEPO were
Absorbance (214 nm)

0.04 (a)
F H
E I
D
C J
AB K
15 20 25
-0.01
t (min)
0.04
Absorbance (214 nm)
(b)
H
G I
F
E J
D
BC K L
15 20 25
-0.01
t (min)
Absorbance (214 nm)
0.04 H
(c) G I
F J
E
D K
BC L
-0.01 15 20 25
t (min)
30 (d)
25
Area percentage of each AGP band
20
AGP pool from ovary
carcinoma patients
15
AGP pool from lymphoma
patients
10 AGP from a healthy donor
0
A B C D E F G H I J K L AGP bands
Lower Charge/mass ratio Higher
FIGURE 22.29 (a, b, and c): Electropherograms of AGP purified from serum of a healthy donor, a pool
of sera from ovary cancer patients, and a pool of sera from lymphoma patients, respectively. A statisti-
cal program was used to compare the three electropherograms, showing the different AGP profile between
the healthy donor and the cancer patients. Analytical conditions: fused-silica capillary 77 cm × 50 µm i.d;
buffer: 0.01 M Tricine, 0.01 M NaCl, 0.01 M sodium acetate, 7 M urea and 3.9 mM putrescine, pH 4.5;
voltage: 25 kV, 35◦ C. (d) Mean percentage areas and standard deviation (indicated as “I” and calculated with
standard AGP, n = 6) of AGP peaks in real samples (2 injections each). Names of the AGP peaks accord-
ing to electropherograms (a), (b), and (c). Discontinuous lines are used to link the real values for the sake
of clarity of the representation. (Adapted from Lacunza, I. et al., Electrophoresis, 27, 4205, 2006. With
permission.)
identified. Undoubtedly, efforts at understanding the separation mechanism of glycoforms have made
it possible. The applicability of CE at most stages of biopharmaceutical manufacturing processes
as well as its usefulness for comparing products from different sources, whether of recombinant
or natural origins, has led most manufacturers to endorse this technique. CE methods have also
appeared in pharmacopeial monographs to replace the more laborious, conventional gel methods.
A list of several proteins for which glycoforms have been separated and the CE method employed
in each case can be seen in Table 22.3.
One particular aspect of CE analysis that requires improvement to make its practical applica-
tion wider is sample preparation. There is little sense in developing simple and fast CE methods if
TABLE 22.3
Separation of Glycoprotein Forms by Capillary Electrophoresis
Protein Separation Mode Reference
24 kDa glycoprotein, CZE Berkowitz, S.A. et al., J. Chromatogr. A, 1079, 254, 2005
recombinant
Alpha-1-acid glycoprotein, ACE Bergstrom, M. et al., J. Chromatogr. B, 809, 323, 2004
human (AGP) (orosomucoid) CIEF Wu, J. et al., J. Chromatogr. A, 817, 163, 1998
Lacunza, I. and de Frutos, M., PACE Setter, 10, 5, 2006
Lacunza, I. et al., Electrophoresis, 28, 1204, 2007
MEKC James, D.C. et al., Anal. Biochem., 222, 315, 1994
CZE Kubo, K., J. Chromatogr. B, 697, 217, 1997
Kinoshita, M. et al., J. Chromatogr. A, 866, 261, 2000
Pacakova, V. et al., Electrophoresis, 22, 459, 2001
Kakehi, K. et al., Anal. Chem., 73, 2640, 2001
Sei, K. et al., J. Chromatogr. A, 958, 273, 2002
Alpha-1-acid glycoprotein, CZE Che, F.-Y. et al., Electrophoresis, 20, 2930, 1999
bovine (bovine AGP)
Balaguer, E. and Neususs, C., Anal. Chem., 78, 5384, 2006
Alpha-1-acid glycoprotein, rat CZE Kinoshita, M. et al., J. Chromatogr. A, 866, 261, 2000
(rat AGP)
Alpha-1-acid glycoprotein, sheep CZE Kinoshita, M. et al., J. Chromatogr. A, 866, 261, 2000
(sheep AGP)
Alpha-1-antitrypsin (AAT) CZE Kubo, K., J. Chromatogr. B, 697, 217, 1997
Chang, W.W.P. et al., Electrophoresis, 26, 2179, 2005
Antithrombin III, human (ATIII) CIEF Kremser, L. et al., Electrophoresis, 24, 4282, 2003
CZE Demelbauer, U.M. et al., Electrophoresis, 25, 2026, 2004
Kremser, L. et al., Electrophoresis, 24, 4282, 2003
Antithrombin III, recombinant CZE Reif, O.-W. and Freitag, R., J. Chromatogr. A, 680, 383, 1994
human (rhATIII) CIEF Reif, O.-W. and Freitag, R., J. Chromatogr. A, 680, 383, 1994
Buchacher, A. et al, J. Chromatogr. A, 802, 355, 1998
Antithrombin III β CE-SDS Buchacher, A. et al., J. Chromatogr. A, 802, 355, 1998
Avidin CZE Bateman, K.P. et al., Methods Mol. Biol., 213, 219, 2003
Beta-trace proteins (BTP) CIEF Hiraoka, A. et al., Electrophoresis, 22, 3433, 2001
Bone morphogenic protein-2, CZE Yim, K. et al., J. Chromatogr. A, 716, 401, 1995
recombinant human (rhBMP-2) Yeung, B. et al., Anal. Chem., 69, 2510, 1997
Caseinoglycomacropeptide CZE Cherkaoui, S. et al., J. Chromatogr. A, 790, 195, 1997
Cherkaoui, S. et al, Chromatographia, 50, 311, 1999
Daali, Y. et al., J. Pharm. Biomed. Anal., 24, 849, 2001
CD4, recombinant (rCD4) CZE Wu, S.-L. et al., J. Chromatogr., 516, 115, 1990
TABLE 22.3
(Continued)
Cellobiohydrolase I (CBH-I) CIEF Sandra, K. et al., J. Chromatogr. A, 1058, 263, 2004

Cell adhesion molecule (CAM) CITP Josic, D. et al., J. Chromatogr., 516, 89, 1990
Chorionic gonadotropin, human CZE Morbeck, D.E. et al., J. Chromatogr. A, 680, 217, 1994
(hCG) Oda, R.P. et al, J. Chromatogr. A, 680, 85, 1994
Darbepoetin-α (novel CZE Lacunza, I. et al., Electrophoresis, 25, 1569, 2004
erythropoiesis stimulating Sanz-Nebot, V. et al., Electrophoresis, 26, 1451, 2005
protein, NESP)
Desmodus salivary plasminogen CZE Apfel, A. et al., J. Chromatogr. A, 717, 41, 1995
activator (DSPAα1) Chakel, J.A. et al., J. Chromatogr. B, 689, 215, 1997
DNAse, recombinant human CZE Felten, C. et al., J. Chromatogr. A, 853, 295, 1999
(rhDNAse) Quan, C.P. et al., Chromatographia Suppl., 53, S39, 2001
Erythropoietin, recombinant CZE Tran, A.D. et al., J. Chromatogr., 542, 459, 1991
human (rhEPO) Watson, E. and Yao, F., Anal. Biochem., 210, 389, 1993
Nieto, O. et al., Anal. Commun., 33, 425, 1996
Bietlot, H.P. and Girard, M., J. Chromatogr. A, 759, 177, 1997
Kaniansky, D. et al., J. Chromatogr. A, 772, 103, 1997
Zhou, G.-H. et al., Electrophoresis, 19, 2348, 1998
Bristow, A. and Charton, E., Pharmaeuropa, 11, 290, 1999
Che, F.-Y. et al., Electrophoresis, 20, 2930, 1999
Lopez-Soto-Yarritu, P. et al., J. Sep. Sci., 25, 1112, 2002
Sanz-Nebot, V. et al., Anal. Chem. 75, 5220, 2003
European Pharmacopoeia 5th edition, published by EDQM.
June 2004
Neususs, C. et al., Electrophoresis, 26, 1442, 2005
Madajova, V. et al., Electrophoresis, 26, 2664, 2005
Yu, B. et al., J. Sep. Sci., 28, 2390, 2005
Balaguer, E. and Neususs, C., Chromatographia, 64, 351, 2006
Balaguer, E. et al., Electrophoresis, 27, 2638, 2006
Kamoda, S. and Kakehi, K., Electrophoresis, 27, 2495, 2006
Lara-Quintanar, P. et al., J. Chromatogr. A, 1153, 227, 2007
CIEF Kubach, J. and Grimm, R., J. Chromatogr. A, 737, 281, 1996
Cifuentes, A. et al., J. Chromatogr. A, 830, 453, 1999
Lopez-Soto-Yarritu, P. et al., J. Chromatogr. A, 968, 221, 2002
CITP Madajova, V. et al., Electrophoresis, 26, 2664, 2005
Erythropoietin, urinary (uEPO) CZE de Frutos, M. et al., Electrophoresis, 24, 678, 2003
Yu, B. et al., J. Sep. Sci., 28, 2390, 2005
Factor VIIa, recombinant human CZE Klausen, N.K. and Kornfelt, T., J. Chromatogr. A, 718, 195,
1995
Factor IX, human CIEF Buchacher, A. et al., J. Chromatogr. A, 802, 355, 1998
Fetuin, bovine CZE Kinoshita, M. et al., J. Chromatogr. A, 866, 261, 2000
Fetuin CE-SDS Werner, W.E. et al., Anal. Biochem., 212, 253, 1993
MEKC James, D.C. et al., Anal. Biochem. 222, 315, 1994
Continued
TABLE 22.3
(Continued)
FG basic chimeric glycoprotein, CZE Tsuji, K. and Little, R.J., J. Chromatogr., 594, 317, 1992
recombinant
Follicle stimulating hormone, CZE Mulders, J.W.M. et al., 3rd WCBP meeting, Washington, DC,
recombinant human (rhFSH) 1999
Storring, P.L. et al., Biologicals, 30, 217, 2002
Granulocyte colony-stimulating CZE Watson, E. and Yao, F., J. Chromatogr., 630, 442, 1993
factor, recombinant human Somerville, L. E. et al., J. Chromatogr. B, 732, 81, 1999
(rhG-CSF) Zhou, G.-H. et al., J. Pharm. Biomed. Anal., 35, 425, 2004
Hepatitis C virus highly MEKC Kundu, S. et al., J. Cap. Electrophor., 3, 301, 1996
glycosylated protein,
recombinant
Hirudin, novel O-glycosylated P6 CZE Steiner, V. et al., Biochemistry, 31, 2294, 1992
(leech)
HIV gp120, recombinant MEKC Jones, D.H. et al., Vaccine, 13, 991, 1995
CIEF Tran, N.T. et al., J. Chromatogr. A, 866, 121, 2000
Horseradish peroxidase (HRP) MEKC James, D.C. et al., Anal. Biochem., 222, 315, 1994
CZE Kelly, J.F. et al., J. Chromatogr. A, 720, 409, 1996
Interferon-γ, recombinant human MEKC James, D.C. et al., Anal. Biochem. 222, 315, 1994
(rhIFN-γ) James, D. C. et al., Prot. Sci., 5, 331, 1996
Goldman, M.H. et al., Biotech. Bioeng., 60, 597, 1998
Hooker, A.D. and James, D.C., Mol. Biotech., 14, 241, 2000
Interferon-ω MEKC Kopp, K. et al., Arzneim. Forsch./Drug Res., 46, 1191, 1996
Interleukin-2 (IL-2) CZE Knuver-Hopf, J. and Mohr, H., J. Chromatogr. A, 717, 71,
1995
Lipocalin-type prostaglandin D CIEF Hiroaka, A. et al., Electrophoresis, 22, 3433, 2001

synthase
Monoclonal anti-alpha-1 CIEF Wu, J. et al., J. Chromatogr. A, 817, 163, 1998
antitrypsin
Monoclonal antibody HER2, CIEF Hunt, G. et al., J. Chromatogr. A, 744, 295, 1996
recombinant humanized Harris, R.J., J. Chromatogr. A, 705, 129, 1995
(rhuMAbHER2) CE-SDS Hunt, G. et al., J. Chromatogr. A, 744, 295, 1996
Monoclonal antibody antihuman CZE Kelly, J.A. and Lee, C.S., J. Chromatogr. A, 790, 207, 1997
follicle stimulating hormone,
mouse
Monoclonal antibodies anti-HIV CZE Wenisch, E., et al., J. Chromatogr., 516, 13, 1990
gp-41
Monoclonal antibody, CZE Compton, B.J. J. Chromatogr., 559, 357, 1991
unspecified (mAb)
Monoclonal antibody, CIEF Kubach, J. and Grimm, R., J. Chromatogr. A, 737, 281, 1996
unspecified (mAb)
Monoclonal antibody, CIEF Schwer, C., Electrophoresis, 16, 2121, 1995
unspecified (mAb)
Monoclonal antibody, CE-SDS Salas-Solano, O. et al., Anal.Chem., 78, 6583, 2006
unspecified (mAb)
Continued
TABLE 22.3
(Continued)
Ovalbumin (OVA) CZE Bullock, J.A. and Yuan, L.C., J. Microcol. Sep., 3, 241, 1991
Landers, J.P. et al., Anal. Biochem., 205, 115, 1992
Taverna, M. et al., Electrophoresis, 13, 359, 1992
Bullock, J., J. Chromatogr., 633, 235, 1993
Kelly, J.F. et al., Beckman discovery Series 2, 1993
Thibault, P. et al., HPCE’94, San Diego, CA 1994, Abstract
P-323, p. 126
Oda, R.P. et al, J. Chromatogr. A, 680, 85, 1994
Legaz, M.E. and Pedrosa, M.M., J. Chromatogr. A, 719, 159,
1996
Oda, R.P. and Landers, J.P., Mol. Biotechnol. 5, 165, 1996
Chakel, J.A. et al., J. Chromatogr. B, 689, 215, 1997
Chen, Y., J. Chromatogr. A, 768, 39, 1997
Pinto, D.M. et al., Anal. Chem., 69, 3015, 1997
Che, F.Y., et al., J. Chromatogr. A, 849, 599, 1999
Kubo K. and Hattori, A., Electrophoresis, 22, 3389, 2001
Pacakova, V. et al., Electrophoresis, 22, 459, 2001
Catai, J.R. et al., J. Chromatogr. A, 1083, 185, 2005
ACE Uegaki, K. et al., Anal. Biochem., 309, 269, 2002
Ovalbumin (turkey) CZE Che, F.Y., et al., J. Chromatogr. A, 849, 599, 1999
Pepsin CZE Landers, J.P. et al, Anal. Biochem., 205, 115, 1992
Placental alkaline phosphatase CZE Eriksson, H.J.C. et al., J. Chromatogr. B, 755, 311, 2001
Pollen allergens CZE Pacakova, V. et al., Electrophoresis, 22, 459, 2001
Prostate-specific antigen (PSA) CZE Donohue, M. J. et al., Anal. Biochem., 339, 318, 2005
Proteinase A (S. cerevisiae) CZE Pedersen, J. and Biedermann, K., Biotechnol. Appl. Biochem.
18, 377, 1993
Ribonuclease B (RNase B) MEKC Rudd, P.M. et al., Glycoconj. J., 9, 86, 1992
James, D.C. et al., Anal. Biochem., 222, 315, 1994
Rudd, P.M. et al., Biochemistry, 33, 17, 1994
CZE Grossman, P.D. et al., Anal. Chem., 61, 1186, 1989
Kelly, J.F. et al., J. Chromatogr. A, 720, 409, 1996
Yeung, B. et al., Anal. Chem., 69, 2510, 1997
Bateman, K.P. et al., Meth. Mol. Biol., 213, 219, 2003
ACE Uegaki, K. et al., Anal. Biochem., 309, 269, 2002
Serum proteins (albumin, CZE Kim, J.W. et al., Clin. Chem., 39, 689, 1993
globulins, IgG), human
Somatropin, recombinant bovine CGE Tsuji, K., J. Chromatogr. A, 652, 139, 1993
(rbSt)
Superoxide dismutase, CZE Wenisch, E. et al., J. Chromatogr., 516, 13, 1990
recombinant
Tissue plasminogen activator, CZE Wu, S.-L. et al., J. Chromatogr., 516, 115, 1990
recombinant human (rhtPA) Yim, K., J. Chromatogr., 559, 401, 1991
Thorne, J. M. et al., J. Chromatogr. A, 744, 155, 1996
CIEF Yim, K., J. Chromatogr., 559, 401, 1991
Moorhouse, K.G. et al., J. Chromatogr. A, 717, 61, 1995
Moorhouse, K.G. et al., Electrophoresis, 17, 423, 1996
Chen, A.B. et al., J. Chromatogr. A, 744, 279, 1996
TABLE 22.3
(Continued)
Thorne, J.M. et al., J. Chromatogr. A, 744, 155, 1996

Kubach, J. and Grimm, R., J. Chromatogr. A, 737, 281, 1996
CGE-SDS Thorne, J.M et al., J. Chromatogr. A, 744, 155, 1996
Transferrin, human (Tf) CZE Kilar, F. and Hjerten, S., J. Chromatogr., 480, 351, 1989
Bergmann, J. et al., Pharmazie, 51, 644, 1996
Oda, R.P. and Landers, J.P., Electrophoresis, 17, 431, 1996
Iourin, O. et al., Glycoconj. J., 13, 1031, 1996
Kubo, K., J. Chromatogr. B, 697, 217, 1997
Prasad, R. et al., Electrophoresis, 18, 1814, 1997
Oda, R.P. et al., Electrophoresis, 18, 1819, 1997
Tagliaro, F. et al., Electrophoresis, 19, 3033, 1998
Tagliaro, F. et al., J. Cap. Electrophor Microchip Tech., 5, 137,
1999
Beisler, A.T. et al., Anal. Biochem., 285, 143, 2000
Giordano, B.C. et al., J. Chromatogr. B, 742, 79, 2000
Trout, A.L. et al., Electrophoresis, 21, 2376, 2000
Crivellente, F. et al., J. Chromatogr. B, 739, 81, 2000
Wuyts, B. et al., Clin. Chem. Lab. Med., 39, 937, 2001
Wuyts, B. et al., Clin. Chem., 47, 247, 2001
Lanz, C. et al., J. Chromatogr. A, 979, 43, 2002
Legros, F.J. et al., Clin. Chem., 48, 2177, 2002
Ramdani, B. et al., Clin. Chem., 49, 1854, 2003
Sanz-Nebot, V. et al., J. Chromatogr. B, 798, 1, 2003
Legros, F.J. et al., Clin. Chem., 49, 440, 2003
Wuyts, B. and Delanghe, J.R., Clin. Chem. Lab. Med., 41, 739,
2003
Lanz, C. et al., J. Chromatogr. A, 1013, 131, 2003
Lanz, C. and Thormann, W., Electrophoresis, 24, 4272, 2003
Ramdani, B. et al., Clin. Chem., 49, 1854, 2003
Fermo, I. et al., Electrophoresis, 25, 469, 2004
Lanz, C. et al., Electrophoresis, 25, 2309, 2004
Martello, S. et al., Forensic Sci. Internat., 141, 153, 2004
Carchon, H.A. et al., Clin. Chem., 50, 101, 2004
Bortolotti, F. et al., Clin. Chem., 51, 2368, 2005
Chang, W.W.P. et al., Electrophoresis, 26, 2179, 2005
Joneli, J. et al., J. Chromatogr. A, 1130, 272, 2006
CIEF Kilar, F. and Hjerten, S., Electrophoresis, 10, 23, 1989
Kilar, F. and Hjerten, S., J. Chromatogr., 480, 351, 1989
Molteni, S. and Thormann, W., J. Chromatogr., 638, 187, 1993
Trypsin inhibitor, human urinary CZE Che, F.-Y. et al., Electrophoresis 20, 2930, 1999
(UTI)
Tumor necrosis factor receptor CIEF Jochheim, C. et al., Chromatographia Suppl., 53, S59, 2001
fusion protein, recombinant
human (rhTNFR:Fc)
they first require time-consuming and labor-intensive steps for obtaining adequate samples. Addi-
tional efforts along the lines of work described in this chapter for developing CE methods that do
not require a prior purification step, or developing fast and simple purification methods will be
required. In combination to more general purification methods, specific ones could be included
to reach this goal. The use of affinity-based systems using lectins, antibodies, or aptamers may
provide elegant solutions in this regard. The online combination of an affinity purification step
with the CE separation has been demonstrated recently and appears promising even in cases where
high throughput is achieved by using multicapillary systems, such as the one described for other
analytes [224].
One of the major challenges to be faced in future years will be that of subsequent entry biologics
(Canada)/follow-on biologics (USA)/biosimilar medicines (EU), and the ability to establish parame-
ters that will allow the comparison of complex glycoproteins whether it is at the glycoform or isoform
levels. Work in this area has been reported in recent years but it is anticipated that with the expiration
of patents for several biopharmaceuticals currently on the market, many of them will be produced
under a wide variety of manufacturing conditions. This will be even more challenging as it has been
shown in this chapter that many parameters can affect the glycosylation profiles. Consequently, the
refinement of CE methods will be required in order to improve peak resolution as well as to detect
minor glycoforms.
The development of multidimensional, multimodal CE, that is, the online coupling of two or
more CE separation modes has been reported several years ago. However, despite its potential for
the analysis of complex mixtures such as glycoproteins, its application to practical problems has
been limited and it is expected that more applications will be reported.
The field of microchip CE (MCE) has progressed tremendously over the past few years and
its application to the analysis of glycoproteins has only been recently reported [225] and will
undoubtedly continue to find applications in the coming years.
The majority of the work on CE of glycoforms carried out up to now is related to methods devel-
opment and the successful demonstration of proof of concept for specific applications. This should
lead as a logical next step to its application to a larger number of practical situations, especially in
the clinical and diagnostic fields. This will require the methods to be applied to large, well-controlled
populations to achieve well-defined conclusions about their validity. However, the application of
CE techniques (including MCE) to the clinical field for specific applications such as the analysis
of glycoforms as disease markers may be somewhat restrained (if not completely prevented) if
more sensitive detection methods are not available. In this sense, newer detection schemes, perhaps
involving the development of new fluorescent labeling agents for LIF detection of glycoproteins are
required.
ACKNOWLEDGMENTS
Financial support from Spanish Ministry of Education and Science (Projects TIC2003-01906,
HH04-33, CTQ2006-05214, DEP2006-56207-CO3-01), Fundacion Ramon Areces (Project Biosen-
sors), Fundacion Domingo Martinez (project Microorganisms), and Comunidad de Madrid (Project
S2006/GEN-0247) is acknowledged.
ABBREVIATIONS
ACE Affinity capillary electrophoresis
AGP Alpha-1-acid glycoprotein, orosomucoid
Ara Arabinose
Asn Asparagine
AAT Alpha-1-antitrypsin
ATIII Antithrombin III
BGE Background electrolyte
BRP Biological Reference Preparation

BSF Bovine serum fetuin
BTP 1,3-Bis[tris(hydroxymethyl)-methylamino] propane
CAPS 3-(Cyclohexylamino)-1-propanesulfonic acid
CBH-I Cellobiohydrolase I
CDG Congenital disorders of glycosylation
CDT Carbohydrate-deficient transferrin
CE Capillary electrophoresis
CE-MS Online coupling of CE to mass spectrometry
CE-SDS Capillary electrophoresis in sodium dodecylsulfate
CGMP Caseinoglycomacropeptide
CHO Chinese hamster ovary
CIEF Capillary isoelectric focusing
CITP Capillary isotacophoresis
Con A Concanavalin A
CRP C-reactive protein
CSF Cerebrospinal fluid
CZE Capillary zone electrophoresis
DAB 1,4-Diaminobutane, putrescine
DAP 1,3-Diaminopropane
DcBr Decamethonium bromide
EACA ε-Aminocaproic acid
EOF Electroosmotic flow
EP European Pharmacopoeia
EPO Erythropoietin
ESI Electrospray ionization
ESI-TOF MS Electrospray ionization time of flight mass spectrometry
Fuc Fucose
Gal Galactose
GalNAc N-Acetylgalactosamine
GlcNAc N-Acetylglucosamine
Glu Glucose
HC Heavy chain
hCG Human chorionic gonadotropin
HEPPSO N-(2-hydroxyethyl)piperazin-N -2-(hydroxypropane sulfonic acid)
HPLC High performance liquid chromatography
HPMC Hydroxypropylmethylcellulose
HRP Horseradish peroxidase
HSA Human serum albumin
hTNF Human tumour necrosis factor
Hyl Hydroxylysine
Hyp Hydroxyproline
IEF Isoelectric focusing
LC Light chain
LCA Lens culinaris agglutinin
LIF Laser-induced fluorescence
mAb Monoclonal antibody
MALDI Matrix-assisted laser desorption ionization
Man Mannose
MEKC Micellar electrokinetic chromatography

MHEC Methylhydroxyethylcellulose
MS Mass spectrometry
NESP Novel erythropoiesis-stimulating protein
NeuAc N-Acetylneuraminic acid
OVA Ovalbumin
PAA Polyacrylamide
PAGE Polyacrylamide gel electrophoresis
PB Polybrene
pI Isoelectric point
PSA Prostate specific antigen
PTM Postranslational modification
PVA Polyvinyl alcohol
RA Rheumatoid arthritis
rhATIII Recombinant human anti-thrombin III
rhBMP-2 Recombinant human bone morphogenic protein 2
rhDNAse Recombinant human deoxyribonuclease
rhEPO Recombinant human erythropoietin
rhFSH Recombinant human follicle-stimulating hormone
rhGCSF Recombinant human granulocyte colony-stimulating factor
rhGMCSF Recombinant human granulocyte macrophage colony-stimulating factor
rhIFN-γ Recombinant human interferon-γ
rhtPA Recombinant human tissue plasminogen activator
RNase Ribonuclease
SDS Sodium dodecylsulfate
SDS–PAGE Sodium dodecylsulfate polyacrylamide gel electrophoresis
Ser Serine
SIM Selected ion monitoring
TEMED N,N,N ,N -Tetramethylene diamine
Tf Transferrin
Thr Threonine
TOF Time-of-flight
uEPO Human urinary EPO
UV Ultraviolet
Xyl Xylose
REFERENCES
1. Dwek, R.A., Glycobiology: Toward understanding the function of sugars, Chem. Rev., 96, 683, 1996.
2. Taylor, M.E. and Drickamer, K., Introduction to Glycobiolgy, Oxford University Press, New York,
2004, chap. 12.
3. Brockhausen, I. and Kuhns, W., Glycoproteins and Human Disease, Springer-Verlag, Heidelberg,
Germany, 1997, chap. 2.
4. Durand, G. and Seta, N., Protein glycosylation and diseases: Blood and urinary oligosaccharides as
markers for diagnosis and therapeutic monitoring, Clin. Chem., 46, 795, 2000.
5. Walsh, G., Pharmaceutical biotechnology products approved within the European Union, Eur. J. Pharm.
Biopharm., 55, 3, 2003.
6. Walsh, G., Second-generation biopharmaceuticals, Eur. J. Pharm. Biopharm., 58, 185, 2004.
7. Jenkins, N. and Curling, M.A., Glycosylation of recombinant proteins: Problems and prospects,
Enzyme Microb. Technol., 16, 354, 1994.
8. Jung, E. and Williams, K.L., The production of recombinant glycoproteins with special refer-
ence to simple eukaryotes including Dictyostelium discoideum, Biotechnol. Appl. Biochem., 25, 3,
1997.
9. Varki, A. et al., Essentials of Glycobiology, Cold Spring Harbor Laboratory Press, New York, 1999,
chap. 2.
10. Taylor, M.E. and Drickamer, K., Introduction to Glycobiolgy, Oxford University Press, New York,
2004, chap. 2.
11. Strege, M.A. and Lagu, A.L., Capillary Electrophoresis of Proteins and Peptides, Meth. Mol. Biol.,
Humana Press, Inc., NJ, 2004, Vol. 276.
12. Prichett, T. and Robey, F.A., Capillary electrophoresis of proteins, in Handbook of Capillary
Electrophoresis, Landers, J.P., Ed., CRC Press, Boca Raton, FL, 1997, chap. 9.
13. Kakehi, K., Kinoshita, M., and Nakano, M., Analysis of glycoproteins and the oligosaccha-
rides thereof by high-performance capillary electrophoresis—Significance in regulatory studies on
biopharmaceutical products, Biomed. Chromatogr., 16, 103, 2002.
14. Pantazaki, A., Taverna, M., and Vidal-Madjar, C., Recent advances in the capillary electrophoresis of
recombinant glycoproteins, Anal. Chim. Acta, 383, 137, 1999.
15. Taverna, M. et al., Electrophoretic methods for process monitoring and the quality assessment of
recombinant glycopropteins, Electrophoresis, 19, 2572, 1999.
16. Taverna, M., Tran, N.T., and Ferrier, D., Separation of protein glycoforms by capillary electrophoresis,
Meth. Mol. Biol. 213, 163, 2003.
17. Kamoda, S. and Kakehi, K., Capillary electrophoresis for the analysis of glycoprotein pharmaceuticals,
18. Watzig, H., Degenhardt, M., and Kunkel, A., Strategies for capillary electrophoresis: Method
development and validation for pharmaceutical and biological applications, Electrophoresis, 19,
2695, 1998.
19. Chen, A.B. and Canova-Davis, E., Capillary electrophoresis in the development of recombinant protein
biopharmaceuticals, Chromatographia Suppl., 53, S7, 2001.
20. Patrick, J.S. and Lagu, A.L., Review applications of capillary electrophoresis to the analysis of
biotechnology-derived therapeutic proteins, Electrophoresis, 22, 4179, 2001.
21. Mechref, Y. and Novotny, M.V., Structural investigations of glycoconjugates at high sensitivity, Chem.
Rev., 102, 321, 2002.
22. Little, M.J., Paquette, D.M., and Roos, P.K., Electrophoresis of pharmaceutical proteins: Status quo,
23. Dolnik, V., Capillary electrophoresis of proteins 2003–2005, Electrophoresis, 27, 126, 2006.
24. Huck, C.W., Bakry, R., and Bonn, G.K., Progress in capillary electrophoresis of biomarkers and
metabolites between 2002 and 2005, Electrophoresis, 27, 111, 2006.
25. Kishino, S. and Miyazaki, K., Separation methods for glycoprotein analysis and preparation,
J. Chromatogr. A, 699, 371, 1997.
26. El Rassi, Z., Recent developments in capillary electrophoresis of carbohydrate species, Electrophore-
sis, 18, 2400, 1997.
27. Grossman, P.D., Free-solution capillary electrophoresis, in Capillary Electrophoresis. Theory and
Practice, Grossman, P.D. and Colburn J.C., Eds., Academic Press, New York, 1992, chap. 2.
28. Kalman, F. et al., Charge and size effect in capillary zone electrophoresis of nuclease A and variants,
29. Gitlin, I., Carbeck, J.D., and Whitesides, G.M., Why are proteins charged? Network of charge-charge
interaction in proteins measured by charge ladders and capillary electrophoresis, Angew. Chem. Int.
Ed., 45, 3022, 2006.
30. Balaguer, E. et al., Glycoform characterization of erythropoietin combining glycan and intact protein
analysis by capillary electrophoresis-electrospray-time-of-flight mass spectrometry, Electrophoresis,
27, 2638, 2006.
31. Yim, K., Abrams, J., and Hsu, A., Capillary zone electrophoretic resolution of recombinant human bone
morphogenic protein 2 glycoforms. An investigation into the separation mechanisms for an exquisite
separation, J. Chromatogr. A, 716, 401, 1995.
32. Chen, F.-T.A., Rapid protein analysis by capillary electrophoresis, J. Chromatogr., 559, 445, 1991.
33. van der Greef, J., Stroobant, P., and van der Heijden, R., The role of analytical sciences in medical
systems biology, Curr. Opin. Chem. Biol., 8, 559, 2004.
34. Bristow, A. and Charton, E., Assessment of the suitability of a capillary zone electrophoresis method
for determining isoform distribution of erythropoietin Pharmaeuropa, 11, 290, 1999.
35. Erythropoietin concentrated solution 01/2005:1316, European Pharmacopoeia 5th edition, published
by EDQM. June 2004.
36. Cifuentes, A. et al., Capillary isoelectric focusing of erythropoietin glycoforms and its comparison
with flat-bed isoelectric focusing and capillary zone electrophoresis, J. Chromatogr. A, 830, 453, 1999.
37. San-Nebot, V. et al., Separation of recombinant human erythropoietin glycoforms by capillary elec-
trophoresis using volatile electrolytes. Assessment of mass spectrometry for the characterization of
erythropoietin glycoforms, Anal. Chem., 75, 5220, 2003.
38. Lacunza, I. et al., Selection of migration parameters for a highly reliable assignment of bands of
isoforms of erythropoietin separated by capillary electrophoresis, Electrophoresis, 25, 1569, 2004.
39. Lopez-Soto-Yarritu, P. et al., Improved capillary isoelectric focusing method for recombinant
erythropoietin analysis, J. Chromatogr. A, 968, 221, 2002.
40. de Frutos, M., Cifuentes, A., and Diez-Masa, J.C., Differences in capillary electrophoresis profiles of
urinary and recombinant erythropoietin, Electrophoresis, 24, 678, 2003.
41. Bietlot, H.P. and Girard, M., Analysis of recombinant human erythropoietin in drug formulations by
high performance capillary electrophoresis, J. Chromatogr. A, 759, 177, 1997.
42. Neususs, C., Demelbauer, U., and Pelzing, M., Glycoform characterization of intact erythropoietin
by capillary electrophoresis-electrospray-time of flight-mass spectrometry, Electrophoresis, 26, 1442,
2005.
43. Lara-Quintanar, P. et al., Immunochromatographic removal of albumin in erythropoietin biopharma-
ceutical formulations for its analysis by capillary electrophoresis, J. Chromatogr. A, 1153, 227, 2007.
44. Hjerten, S., Zone broadening in electrophoresis with special reference to the high-performance
electrophoresis in capillaries: An interplay between theory and practice, Electrophoresis, 11, 665, 1990.
45. Lauer, H.H. and McManigill, D., Capillary zone electrophoresis of proteins in untreated fused silica
tubing, Anal. Chem., 58, 166, 1986.
46. McCormick, R.A., Capillary zone electrophoretic separation of peptides and proteins using low pH
buffers in modified silica capillaries, Anal. Chem., 6, 2322, 1986.
47. Corradini, D., Buffer additives other than the surfactant sodium dodecyl sulfate for protein separation
by capillary electrophoresis, J. Chromatogr. B, 699, 21, 1997.
48. Righetti, P.G. et al., The state of the art of dynamic coatings, Electrophoresis, 22, 603, 2001.
49. Ahmadzadeh, H., Dovichi, N.J., and Krylov, S. Capillary coating for protein separation on Si− −O and
Si−−C covalent bond formation for capillary electrophoresis with laser-induced fluorescence detection,
in Capillary Electrophoresis of Proteins and Peptides, Strege, M.A. and Lagu, A.L., Eds., Humana
Press, Totowa, NJ, 2004, chap. 2.
50. Landers, J.P. et al., High performance capillary electrophoresis of glycoproteins: The use of modifiers
of electroosmotic flow for analysis of microheterogeneity, Anal. Biochem., 205, 115, 1992.
51. Legaz, M.E. and Pedrosa, M.M., Effect of polyamines on the separation of ovalbumin glycoforms by
52. Watson, E. and Yao, F., Capillary electrophoretic separation of human recombinant erythropoietin
(r-HuEPO) glycoforms, Anal. Biochem., 210, 389, 1993.
53. Klausen, N.K. and Kornfelt, T., Analysis of the glycoforms of human recombinant factor VIIa by cap-
illary electrophoresis and high-performance liquid chromatography, J. Chromatogr. A, 718, 195, 1995.
54. Watson, E. and Yao, F., Capillary electrophoretic separation of recombinant granulocyte-colony-
stimulating factor glycoforms, J. Chromatogr., 630, 442, 1993.
55. Oda, R.P. and Landers, J.P., Effect of cationic buffer additives on the capillary electrophoretic
separation of serum transferrin from different species, Electrophoresis, 17, 431, 1996.
56. Morbeck, D.E., Madden, B.J., and McCormick, D.J., Analysis of the microheterogeneity of the gly-
coprotein chorionic gonadotropin with high-performance capillary electrophoresis, J. Chromatogr. A,
680, 217, 1994.
57. Oda, R.P. et al., α, ω-Bis quaternary ammonium alkanes as effective buffer additives for enhanced
capillary electrophoretic separation of glycoproteins, J. Chromatogr. A, 680, 85, 1994.
58. Verzola, B., Gelfi, C., and Righetti, P.G., Protein adsorption to the bare silica wall in capillary
electrophoresis. Quantitative study on the chemical composition of the background electrolyte for
minimizing the phenomenon, J. Chromatogr. A, 868, 85, 2000.
59. Wicktorowicz, J.E. and Colburn, J.E., Separation of cationic proteins via charge reversal in capillary
60. Cordova, E., Gao, J., and Whitesides G.M., Noncovalent polycationic coating for capillaries in capillary
electrophoresis of proteins, Anal. Chem., 69, 1370, 1997.
61. Kelly, J.F. et al., Development of electrophoretic conditions for the characterization of protein
glycoforms by capillary electrophoresis-electrospray mass spectrometry, J. Chromatogr. A, 720,
409, 1996.
62. Yu, B. et al., Ionene-dynamically coated capillary for analysis of urinary and recombinant human
erythropoietin by capillary electrophoresis and online electrospray ionization mass spectrometry, J.
Sep. Sci., 28, 2390, 2005.
63. Katayama, H., Ishihama, Y., and Asakawa, N., Stable capillary coating with successive multiple ionic
polymer layers, Anal. Chem., 70, 2254, 1998.
64. Bendahl, L., Hansen, S.H., and Gammelgaard, B., Capillaries modified by noncovalently anionic poly-
mer adsorption for capillary zone electrophoresis, micellar electrokinetic capillary chromatography
and capillary electrophoresis, Electrophoresis, 22, 2565, 2001.
65. Graul, T.W. and Schlenoff, J.B., Capillaries modified by polyelectrolyte multilayers for electrophoretic
separations, Anal. Chem., 71, 4007, 1999.
66. Lanz, C., Marti, U., and Thormann, W., Capillary zone electrophoresis with a dynamic double coating
for analysis of carbohydrate-deficient transferrin in human serum. Precision performance and pattern
recognition, J. Chromatogr. A, 1013, 131, 2003.
67. Joneli, J., Lanz, C., and Thormann, W., Capillary zone electrophoresis determination of carbohydrate-
deficient transferrin using the new Ceofix reagents under high-resolution conditions, J. Chromatogr.
A, 1130, 272, 2006.
68. Legros, F.J. et al., Carbohydrate-deficient transferrin isoforms measured by capillary zone elec-
trophoresis for detection of alcohol abuse, Clin. Chem., 48, 2177, 2002.
69. Legros, F.J. et al., Use of capillary zone electrophoresis for differentiating excessive from moderate
alcohol consumption, Clin. Chem., 49, 440, 2003.
70. Chang, W.W.P. et al., Rapid separation of protein isoforms by capillary zone electrophoresis with new
dynamic coating, Electrophoresis, 26, 2179, 2005.
71. Balaguer, E. and Neususs, C., Glycoprotein characterization combining intact protein and glycan anal-
ysis by capillary electrophoresis-electrospray ionization-mass spectrometry, Anal. Chem., 78, 5384,
2006.
72. Hjerten, S., High-performance electrophoresis. Elimination of the electroendosmosis and solute
adsorption, J. Chromatogr., 347, 191, 1985.
73. Schmalzing, D. et al., Characterization and performance of a neutral hydrophilic coating for the
capillary electrophoretic separation of biopolymers, J. Chromatogr. A., 652, 149, 1993.
74. Hjerten, S. and Kubo, K., A new type of pH- and detergent stable coating for elimination of
electroendosmosis and adsorption in (capillary) electrophoresis, Electrophoresis, 14, 390, 1993.
75. Gilges, M., Kleemiss, M.H., and Schomburg, G., Capillary zone electrophoresis separation of basic
and acidic proteins using poly(vinyl alcohol) coatings in fused silica capillaries, Anal. Chem., 66,
2038, 1994.
76. Yim, K., Fractionation of the human recombinant tissue plasminogen activator (rtPA) glycoforms by
high-performance capillary zone electrophoresis and capillary isoelectric focusing, J. Chromatogr.,
559, 401, 1991.
77. Thorne, J.M. et al., Examination of capillary zone electrophoresis, capillary isoelectric focusing and
sodium dodecyl sulphate capillary electrophoresis for the analysis of recombinant tissue plasminogen
activator, J. Chromatogr. A, 744, 155, 1996.
78. Kinoshita, M. et al., Comparative studies on the analysis of glycosylation heterogeneity of sialic
acid-containing glycoproteins using capillary electrophoresis, J. Chromatogr. A, 866, 261, 2000.
79. Kilar, F. and Hjerten, S., Separation of the human transferrin isoforms by carrier-free high-performance
zone electrophoresis and isoelectric focusing, J. Chromatogr. A, 480, 351, 1989.
80. Hunt, G., Moorhouse, K.G., and Chen, A.B., Capillary isoelectric focusing and sodium dodecyl sulfate-
capillary gel electrophoresis of recombinant humanized monoclonal antibody HER2, J. Chromatogr.
A, 744, 295, 1996.
81. Grossman, P.D. et al., Application of free-solution capillary electrophoresis to the analytical scale
separation of proteins and peptides, Anal. Chem., 61, 1186, 1989.
82. Taverna, M. et al., Analysis of carbohydrate-mediated heterogeneity and characterization of N-linked
oligosaccharies of glycoproteins by high performance capillary electrophoresis, Electrophoresis, 13,
359, 1992.
83. Sanz-Nebot, V. et al., Capillary electrophoresis and matrix-assisted laser desorption/ionization-time
of flight-mass spectrometry for analysis of the novel erythropoiesis-stimulating protein (NESP),
84. Lacunza, I. et al., CZE of human alpha-1-acid glycoprotein for qualitative and quantitative comparison
of samples from different pathological conditions, Electrophoresis, 27, 4205, 2006.
85. Honda, S. et al., Simultaneous determination of reducing monosaccharides by capillary zone
electrophoresis as the borate complexes of N-2 pyridylglycamines, Anal. Biochem., 176, 72, 1989.
86. Hoffsteter-Kuhn, S. et al., Influence of borate complexation on the electrophoretic behavior of
carbohydrates in capillary electrophoresis, Anal. Chem., 63, 1541, 1991.
87. Pacakova, V. et al., Effects of electrolyte modification and capillary coating on separation of
glycoproteins isoforms by capillary electrophoresis, Electrophoresis, 22, 459, 2001.
88. Chen, Y., Critical conditions for separating the microheterogeneous components of glycoproteins by
89. Oda, R.P. et al., Capillary electrophoresis-based separation of transferrin sialoforms in patients with
carbohydrate-deficient glycoprotein syndrome, Electrophoresis, 18, 1819, 1997.
90. Lopez-Soto-Yarritu, P. et al., Comparison of different capillary electrophoresis methods for analysis
of recombinant erythropoietin glycoforms, J. Sep. Sci., 25, 1112, 2002.
91. Tran, A.D. et al., Separation of carbohydrate-mediated microheterogeneity of recombinant human
erythropoietin by free solution capillary electrophoresis, J. Chromatogr., 542, 459, 1991.
92. Zhou, G.-H. et al., Application of capillary electrophoresis, liquid chromatography, electrospray-mass
spectrometry and matrix-assisted laser desorption/ionization-time of flight-mass spectrometry to the
characterization of recombinant human erythropoietin, Electrophoresis, 19, 2348, 1998.
93. Nieto, O., Hernandez, P., and Hernandez, L., Capillary zone electrophoresis of human recombinant
erythropoietin using C8 coated columns without additives in the running buffer, Anal. Commun., 33,
425, 1996.
94. Kaniansky, D. et al., Capillary zone electrophoresis in a hydrodynamically closed separation system
with enhanced sample loadability, J. Chromatogr. A, 772, 103, 1997.
95. Madajova, V. et al., Fractionation of glycoforms of recombinant human erythropoietin by preparative
capillary isotachophoresis, Electrophoresis, 26, 2664, 2005.
96. Balaguer, E. and Neususs, C., Intact glycoform characterization of erythropoietin-α and
erythropoietin-β by CZE-ESI-TOF-MS, Chromatographia, 64, 351, 2006.
97. Kilar, F. and Hjerten, S., Fast and high resolution analysis of human serum transferrin by high
performance isoelectric focusing in capillaries, Electrophoresis, 10, 23, 1989.
98. Righetti, P.G., Determination of the isoelectric point of proteins by capillary isoelectric focusing,
J. Chromatogr. A, 1037, 491, 2004.
99. Lacunza, I. and de Frutos, M., Different capillary isoelectric focusing approaches for analysis of
α-1-acid glycoprotein, PACE Setter, 10, 5, 2006.
100. Wu, J., Li, S.-C., and Watson, A., Optimizing separation conditions for proteins and peptides using
imaged capillary isoelectric focusing, J. Chromatogr. A, 817, 163, 1998.
101. Moorhouse, K.G. et al., Rapid one-step capillary isoelectric focusing method to monitor charged
glycoforms of recombinant human tissue-type plasminogen activator, J. Chromatogr. A, 717, 61, 1995.
102. Reif, O.-W. and Freitag, R., Control of the cultivation process of antithrombin III and its
characterization by capillary electrophoresis, J. Chromatogr. A, 680, 383, 1994.
103. Tran, N.T. et al., One-step capillary isoelectric focusing for the separation of the recombinant human
immunodeficiency virus envelope glycoprotein glycoforms, J. Chromatogr. A, 866, 121, 2000.
104. Kubach, J. and Grimm, R., Non-native capillary isoelectric focusing for the analysis of the
microheterogeneity of glycoproteins, J. Chromatogr. A, 737, 281, 1996.
105. Moorhouse, K.G., Rickel, C.A., and Chen, A.B., Electrophoretic separation of recombinant tissue-
type plasminogen activator glycoforms: Validation issues for capillary isoelectric focusing methods,
106. Chen, A.B. et al., Comparison of ampholytes used for slab gel and capillary isoelectric focusing of
recombinant tissue-type plasminogen activator glycoforms, J. Chromatogr. A, 744, 279, 1996.
107. Sebastiano, R. et al., Mass distribution and focusing properties of carrier ampholytes for isoelectric
focusing: I. Novel and unexpected results, Electrophoresis, 27, 3919, 2006.
108. Simo, C. et al., Mass distribution, polydispersity and focusing properties of carrier ampholytes for IEF
II: pH 4-6 intervals, Electrophoresis, 27, 4849, 2006.
109. Lacunza, I. et al., CIEF with hydrodynamic and chemical mobilization for the separation of forms of
alpha-1-acid glycoprotein, Electrophoresis, 28, 1204, 2007.
110. Wehr, T., Rodriguez-Diaz, R., and Zhu, M., Recent advances in capillary isoelectric focusing,
Chromatographia Suppl., 53, S45, 2001.
111. Righetti, P.G., Isoelectric Focusing: Theory, Methodology, and Applications, Elsevier Biomedical,
Amsterdam, 1989, 1.
112. Schwer, C., Capillary isoelectric focusing: A routine method for protein analysis?, Electrophoresis,
16, 2121, 1995.
113. Rudd, P.M. et al., Separation and analysis of the glycoform populations of ribonuclease B using
capillary electrophoresis, Glycoconj. J., 9, 86, 1992.
114. James, D.C. et al., High-resolution separation of recombinant human interferon-γ glycoforms by
micellar electrokinetic capillary chromatography, Anal. Biochem., 222, 315, 1994.
115. Werner, W.E. et al., Size-dependent separation of proteins denatured in SDS by capillary electrophoresis
using a replaceable sieving matrix, Anal. Biochem., 212, 253, 1993.
116. Buchacher, A. et al., High-performance capillary electrophoresis for in-process control in
the production of antithrombin III and human clotting factor IX, J. Chromatogr. A, 802,
355, 1998.
117. Bergstrom, M. et al., Lectin affinity capillary electrophoresis in glycoform analysis applying the partial
filling technique, J. Chromatogr. B, 809, 323, 2004.
118. Uegaki, K. et al., Simultaneous estimation of the association constants of glycoprotein glycoforms to
a common protein by capillary electrophoresis, Anal. Biochem., 309, 269, 2002.
119. Thormann, W. and Mosher, R.A., High-resolution computer simulation of the dynamics of isoelec-
tric focusing using carrier ampholytes: Focusing with concurrent electrophoretic mobilation is an
isotachophoretic process, Electrophoresis, 27, 968, 2006.
120. Iourin, O. et al., The identification of abnormal glycoforms of serum transferrin in carbohydrate
deficient glycoprotein syndrome type I by capillary zone electrophoresis, Glycoconj. J., 13, 1031,
1996.
121. Thibault, P. and Dovichi, N., General instrumentation and detection systems including mass spectrom-
etry, in Capillary Electrophoresis. Theory and Practice, Camilleri, P., Ed., CRC Press, Boca Raton,
FL, 1998, chap. 2.
122. http://probes.invitrogen.com/handbook/sections/0900.html Accessed January 2007.
123. Craig, D.B. and Dovichi, N.J., Multiple labeling of proteins, Anal. Chem., 70, 2493, 1998.
124. Tseng, W.-L., Lin, Y.-W., and Chang, H.-T., Improved separation of microheterogeinities and isoforms
of proteins by capillary electrophoresis using segmental filling with SDS and PEO in the background
electrolyte, Anal. Chem., 74, 4828, 2002.
125. Salas-Solano, O. et al., Optimization and validation of a quantitative capillary electrophoresis
sodium dodecyl sulfate method for quality control and stability monitoring of monoclonal antibodies,
Anal.Chem.,78, 6583, 2006.
126. Pinto, D.M. et al., Picomolar assay of native proteins by capillary electrophoresis precol-
umn. Submicellar separation, and laser-induced fluorescence detection, Anal. Chem., 69, 3015,
1997.
127. Lee, I.H. et al., Picomolar analysis of proteins using electrophoretically mediated microanalysis
and capillary electrophoresis with laser-induced fluorescence detection, Anal. Chem., 70, 4546,
1998.
128. Wehr, T., Rodriguez-Diaz, R., and Zhu, M., Capillary Electrophoresis of Proteins, Marcell Dekker,
New York, 1999, chap. 3.
129. Bornemann, C. et al., Fluorescence-labelled antigen-binding fragments (Fab) from monoclonal anti-
body 5F12 detect human erythropoietin in immunoaffinity capillary electrophoresis, Anal. Bioanal.
Chem., 376, 1074, 2003.
130. Kelly, J.A. and Lee, C.S., Online postcapillary affinity detection of immunoglobulin G subclass and
monoclonal antibodies variants for capillary electrophoresis, J. Chromatogr. A, 790, 207, 1997.
131. Yeung, B., Porter, T.J., and Vath, J.E., Direct isoform analysis of high-mannose-containing glycopro-
teins by online capillary electrophoresis electrospray mass spectrometry, Anal. Chem., 69, 2510, 1997.
132. Demelbauer, U.M. et al., Characterization of glyco isoforms in plasma-derived human antithrombin by
online capillary zone electrophoresis-electrospray ionization-quadrupole ion trap-mass spectrometry
of the intact glycoproteins, Electrophoresis, 25, 2026, 2004.
133. Arizaga-Rodriguez, S. et al., Detection of transferrin isoforms in human serum: Comparison of UV
and ICP-MS detection after CZE and HPLC separations, Anal. Bioanal. Chem., 383, 390, 2005.
134. Hooker, A.D. and James, D.C., Analysis of glycoprotein heterogeneity by capillary electrophoresis
and mass spectrometry, Mol. Biotech., 14, 241, 2000.
135. Morelle, W. and Michalski, J.C., The mass spectrometric analysis of glycoproteins and their glycan
structures, Curr. Anal. Chem., 1, 29, 2005.
136. Bateman, K.P. et al., Glycoprotein analysis by capillary zone electrophoresis–electrospray mass
spectrometry, Meth. Mol. Biol., 213, 219, 2003.
137. Monton, M.R.N. and Terabe, S., Recent developments in capillary electrophoresis—Mass spectrom-
etry of proteins and peptides, Anal. Sci., 21, 5, 2005.
138. Servais, A.-C., Crommen, J., and Fillet, M., Capillary electrophoresis-mass spectrometry, an attractive
tool for drug bioanalysis and biomarker discovery, Electrophoresis, 27, 2616, 2006.
139. ICH Guidelines, http://www.ich.org/cache/compo/276-254-1.html, accessed November 2006.
140. Somatropin, 07/2007:0951, European Pharmacopoeia 5th edition, published by EDQM, April 2007.
141. Somerville, L.E., Douglas, A.J., and Irvine, A.E., Discrimination of granulocyte colony-stimulating
factor isoforms by high-performance capillary electrophoresis, J. Chromatogr. B, 732, 81, 1999.
142. Berkowitz, S.A. et al., Rapid quantitative capillary zone electrophoresis method for monitoring the
micro-heterogeneity of an intact recombinant glycoprotein, J. Chromatogr. A, 1079, 254, 2005.
143. Felten, C. et al., Use of acidic and basic pH and calcium ion addition in the capillary zone electrophoretic
characterization of recombinant human deoxyribonuclease, a complex phosphoglycoprotein, J. Chro-
matogr. A, 853, 295, 1999.
144. Quan, C.P., Canova-Davis, E., and Chen, A.B., Effects of the solution environment on the resolution of
recombinant human deoxyribonuclease variants in capillary zone electrophoresis, Chromatographia
Suppl., 53, S39, 2001.
145. Kremser, L. et al., Characterization of antithrombin III from human plasma by two-dimensional gel
electrophoresis and capillary electrophoretic methods. Electrophoresis, 24, 4282, 2003.
146. Che, F.Y. et al., Comparative study on the distribution of ovalbumin glycoforms by capillary
147. Kubo, K. and Hattori, A., Assessment of the capillary zone electrophoretic behaviour of pro-
teins in the presence of electroosmotic modifiers: Protein-polyamine interaction studied using a
polyacrylamide-coated capillary, Electrophoresis, 22, 3389, 2001.
148. Catai, J.R. et al., Noncovalently bilayer-coated capillaries for efficient and reproducible analysis of
proteins by capillary electrophoresis, J. Chromatogr. A, 1083, 185, 2005.
149. Sandra, K. et al., Combining gel and capillary electrophoresis, nano-LC and mass spectrometry
for the elucidation of posttranslational modifications of Trichoderma reesei cellobiohydrolase I.,
J. Chromatogr. A, 1058, 263, 2004.
150. Krull, I.S. et al., HPCE methods for the identification and quantitation of antibodies, their conjugates
and complexes, J. Pharm. Biomed. Anal., 16, 377, 1997.
151. Santora, L.C., Krull, I.S., and Grant, K., Characterization of recombinant human monoclonal tis-
sue necrosis factor-α antibody using cation-exchange HPLC and capillary isoelectric focusing, Anal.
Biochem., 275, 98, 1999.
152. Harmon, B.J., Imaged capillary isoelectric focusing for the characterization and monitoring of charge
heterogeneity of recombinant monoclonal antibodies, in Proceedings of the 5th Symposium on CE in
the Biotechnology and Pharmaceutical Industries, San Francisco, CA, 2003.
153. Hoffstetter-Kuhn, S., Alt, G., and Kuhn, R. Profiling of oligosaccharide-mediated microheterogeneity
of a monoclonal antibody by capillary electrophoresis, Electrophoresis, 17, 418, 1996.
154. Ma, S. in State of the Art Analytical Methods for the Characterization of Biological Products and
Assessment of Comparability, Mire-Sluis, A.R., Ed., Karger, Basel, 2005, pp. 49–68.
155. Nunnally, B. et al., Aseries of collaborations between various pharmaceutical companies and regulatory
authorities concerning the analysis of biomolecules using capillary electrophoresis, Chromatographia,
64, 359, 2006.
156. Jenkins, N., Parekh, R.B., and James, D.C., Getting the glycosylation right—Implications for the
biotechnology industry, Nat. Biotechnol., 14, 975, 1996.
157. Chirino, A.J. and Mire-Sluis, A., Characterizing biological products and assessing comparability
following manufacturing changes, Nat. Biotechnol., 22, 1383, 2004.
158. James, D.C. et al., Posttranslational processing of recombinant human interferon-γ in animal
expression systems, Prot. Sci., 5, 331, 1996.
159. Goldman, M.H. et al., Monitoring recombinant human interferon-gamma N-glycosylation during
perfused fluidized-bed and stirred-tank batch culture of CHO cells, Biotech. Bioeng., 60, 596, 1998.
160. Wang, M.D. et al., Erythropoietin production from CHO cells grown by continuous culture in a
fluidized-bed bioreactor, Biotech. Bioeng., 77, 194, 2002.
161. Freitag, R., et al., Production of recombinant h-AT III with mammalian cell cultures using capillary
electrophoresis for product monitoring, Cytotechnology, 21, 205, 1996.
162. Klyushnichenko, V., Capillary electrophoresis in the analysis and monitoring of biotechnological
processes, Meth. Mol. Biol., 276, 77, 2004.
163. Schmelzer, A.E. and Miller, W.M., Hyperosmotic stress and elevated pCO2 alter monoclonal antibody
charge distribution and monosaccharide content, Biotechnol. Prog., 18, 346, 2002.
164. Cherkaoui, S. et al., Development of a capillary zone electrophoresis method for caseinoglyco-
macropeptide determination, J. Chromatogr. A, 790, 195, 1997.
165. Cherkaoui, S. et al., Determination of the stability of caseinoglycomacropeptide in a cosmetic lotion
by use of capillary zone electrophoresis with a coated capillary, Chromatographia, 50, 311, 1999.
166. Daali, Y., Cherkaoui, S., and Veuthey, J.L., Capillary electrophoresis and high-performance anion
exchange chromatography for monitoring caseinoglycomacropeptide sialylation, J. Pharm. Biomed.
Anal., 24, 849, 2001.
167. Donohue, M.J. et al., Capillary electrophoresis for the investigation of prostate-specific antigen
heterogeneity, Anal. Biochem., 339, 318, 2005.
168. Girard, M., Boucher, S., and Kane, A., Analysis of erythropoietin preparations obtained from
various sources worldwide by capillary zone electrophoresis, in Proceedings of the 7th Interna-
tional symposium on capillary electrophoresis in the biotechnology and pharmaceutical industries,
Montreal, 2005.
169. Deechongkit, S. et al., Biophysical comparability of the same protein from different manufacturers: a
case study using epoetin alfa from Epogen® and Eprex®, J. Pharm. Sci., 95, 1931, 2006.
170. Zhou, G.-H. et al., Characterization of recombinant human granulocyte colony stimulating factor
(rHuG-CSF) by capillary zone electrophoresis, capillary isoelectric focusing electrophoresis and
electrospray ionization mass spectrometry, J. Pharm. Biomed. Anal., 35, 425, 2004.
171. Storring, P.L. et al., Physicochemical Methods for predicting the biological potency of recombinant
follicle stimulating hormone: An international collaborative study of isoelectric focusing and capillary
zone electrophoresis, Biologicals, 30, 217, 2002.
172. Mulders, J.W.M., Wijn, H., and Theunissen, F. Development and validation of a capillary zone
electrophoresis method which can be used to predict the in-vivo bioactivity of recombinant folli-
cle stimulating hormone, in Proceedings of the 3rd Symposium on the Analysis of Well Characterized
Biotechnology Pharmaceuticals (WCBP), Washington, DC, 1999.
173. Kim, Y.J. and Varki, A., Perspectives on the significance of altered glycosylation of glycoproteins in
cancer, Glycoconj. J., 14, 569, 1997.
174. de Jong, G. and van Eijk, H.G., Microheterogeneity of human serum transferrin: A biological phe-
nomenon studied by isoelectric focusing in immobilized pH gradients, Electrophoresis, 9, 589, 1998.
175. de Jong, G., van Dijk, J.P., and van Eijk, H.G., The biology of transferrin, Clin. Chim. Acta, 190,
1, 1990.
176. Arndt, T., Carbohydrate-deficient transferrin as a marker of chronic alcohol abuse: A critical review
of preanalysis, analysis, and interpretation, Clin. Chem., 47, 13, 2001.
177. Wuyts, B. et al., Carbohydrate-deficient transferrin and chronic alcohol ingestion in subjects with
transferrin CD-variants, Clin. Chem. Lab. Med., 39, 937, 2001.
178. Stibler, H., Carbohydrate-deficient transferrin in serum: A new marker of potentially harmful alcohol
consumption reviewed, Clin. Chem., 37, 2029, 1991.
179. Keir, G., Winchester, B.G., and Clayton, P., Carbohydrate-deficient glycoprotein syndromes: inborn
errors of protein glycosylation, Ann. Clin. Biochem., 36, 20, 1999.
180. Whitfield, J.B., Transferrin isoform analysis for the diagnosis and management of hazardous or
dependent drinking, Clin. Chem., 48, 2095, 2002.
181. Stibler, H. and Kjellin, KG., Isoelectric-focusing and electrophoresis of CSF proteins in tremor of dif-
ferent origins—Study of 38 cases with cerebellar syndrome of chronic-alcoholism, Parkinson’s-disease
and essential tremor, J. Neurol. Sci., 30, 269, 1976.
182. Stibler, H. et al., Abnormal microheterogeneity of transferrin in serum and cerebrospinal-fluid in
alcoholism, Acta Med. Scand., 204, 49, 1978.
183. Stibler, H., Borg, S., and Allgulander, C., Clinical significance of abnormal heterogeneity of transferrin
in relation to alcohol-consumption, Acta Med. Scand., 206, 275, 1979.
184. Wuyts, B. and Delanghe, J.R., The analysis of carbohydrate-deficient transferrin, marker of chronic
alcoholism, using capillary electrophoresis, Clin. Chem. Lab. Med., 41, 739, 2003.
185. Lanz, C. et al., Evaluation and optimization of capillary zone electrophoresis with different dynamic
capillary coatings for the determination of carbohydrate-deficient transferrin in human serum,
J. Chromatogr. A, 979, 43, 2002.
186. Helander, A. et al., Interference of transferrin isoform types with carbohydrate-deficient transferrin
quantification in the identification of alcohol abuse, Clin. Chem., 47, 1225, 2001.
187. Bortolotti, F. et al., Analysis of carbohydrate-deficient transferrin: Comparative evaluation of
turbidimetric immunoassay, capillary zone electrophoresis, and HPLC, Clin. Chem., 51, 2368, 2005.
188. Jeppsson, J.-O., Kristensson, H., and Fimiani, C., Carbohydrate-deficient transferrin quantified by
HPLC to determine heavy consumption of alcohol, Clin. Chem., 39, 2115, 1993.
189. Turpeinen, U. et al., Comparison of HPLC and small column (CDTect) methods for disialotransferrin,
Clin. Chem., 47, 1782, 2001.
190. Helander, A., Husa, A., and Jeppsson, J.-O., Improved HPLC method for carbohydrate-deficient
transferrin in serum, Clin. Chem., 49, 1881, 2003.
191. Prasad, R. et al., Analysis of carbohydrate deficient transferrin by capillary zone electrophoresis,
192. Tagliaro, F. et al., Optimized determination of carbohydrate-deficient transferrin isoforms in serum by
capillary zone electrophoresis, Electrophoresis, 19, 3033, 1998.
193. Crivellente, F. et al., Improved method for carbohydrate-deficient transferrin determination in human
serum by capillary zone electrophoresis, J. Chromatogr. B, 739, 81, 2000.
194. Trout, A.L. et al., Direct capillary electrophoretic detection of carbohydrate-deficient transferrin in
neat serum, Electrophoresis, 21, 2376, 2000.
195. Ramdani, B. et al., Analyte comigrating with trisialotransferrin during capillary zone electrophoresis
of sera from patients with cancer, Clin. Chem., 49, 1854, 2003.
196. Sanz-Nebot, V. et al., Characterization of human transferrin glycoforms by capillary electrophoresis
and electrospray ionization mass spectrometry, J. Chromatogr. B, 798, 1, 2003.
197. Lanz, C. et al., Improved capillary electrophoresis method for the determination of carbohydrate-
deficient transferrin in patient sera, Electrophoresis, 25, 2309, 2004.
198. Martello, S. et al., Determination of carbohydrate deficient transferrin (CDT) with capillary
electrophoresis: an inter laboratory comparison, Forensic Sci. Internat., 141, 153, 2004.
199. Carchon, H.A. et al., Diagnosis of congenital disorders of glycosylation by capillary zone electrophore-
sis of serum transferrin, Clin. Chem., 50, 101, 2004.
200. Tagliaro, F. et al., Carbohydrate-deficient transferrin determination revisited with capillary elec-
trophoresis: A new biochemical marker of chronic alcohol abuse, J. Cap. Electrophor. Microchip
Tech., 5, 137, 1999.
201. Fermo, I. et al., Capillary zone electrophoresis for determination of carbohydrate-deficient transferrin
in huan serum, Electrophoresis, 25, 469, 2004.
202. Giordano, B.C. et al., Dynamically-coated capillaries allow for capillary electrophoretic reso-
lution of transferrin sialoforms via direct analysis of human serum, J. Chromatogr. B, 742,
79, 2000.
203. Wuyts, B. et al., Determination of carbohydrate-deficient transferrin using capillary zone electrophore-
sis, Clin. Chem., 47, 247, 2001.
204. Tagliaro, F. et al., Caveats in carbohydrate-deficient transferrin determination, Clin. Chem., 48, 208,
2001.
205. Lanz, C. and Thormann, W., Capillary zone electrophoresis with a dynamic double coat-
ing for analysis of carbohydrate-deficient transferrin in human serum: Impact of res-
olution between disialo- and trisialotransferrin on reference limits, Electrophoresis, 24,
4272, 2003.
206. Vitala, K., Lähdesmäki, K., and Niemela, O., Comparison of the Axis % CDT TIA and the CDTect
method as laboratory tests of alcohol abuse, Clin. Chem., 44, 1209, 1998.
207. Delanghe, J.R., Wuyts, B., and De Buyzere, M.L., Caveats in carbohydrate-deficient transferring
determination–Response, Clin. Chem., 48, 209, 2002.
208. Marquardt, T. and Dencke, J., Congenital disorders of glycosylation: Review of their molecular bases,
clinical presentations and specific therapies, Eur. J. Pediatr., 162, 359, 2003.
209. Fournier, T., Medjoubi-N, N., and Porquet, D., Alpha-1-acid glycoprotein, Biochim. Biophys. Acta,
1482, 157, 2000.
210. Duche, J.-C. et al., Expression of the genetic variants of human alpha-1-acid glycoprotein in cancer,
Clin. Chem., 33, 197, 2000.
211. Mackiewicz, A. and Mackiewicz, K., Glycoforms of serum α1-acid glycoprotein as markers of
inflammation and cancer, Glycoconj. J., 12, 241, 1995.
212. Hashimoto, S. et al., α1 -Glycoprotein fucosylation as a marker of carcinoma progression and prognosis,
Cancer, 101, 2825, 2004.
213. Kremmer, T. et al., Liquid chromatographic and mass spectrometric analysis of human serum acid
alpha-1-glycoprotein, Biomed. Chromatogr., 18, 323, 2004.
214. Hrycaj, P. et al., Micoheterogeneity of alpha 1 acid glycoprotein in rheumatoid arthritis: Dependent
on disease duration?, Ann. Rheum. Dis., 52, 138, 1993.
215. Elliott, M.A. et al., Investigation into the Concavalin A reactivity, fucosylation and oligosaccharide
microheterogeneity of α1 -acid glycoprotein expressed in the sera of patients with rheumatoid arthritis,
J. Chromatogr. B, 688, 229, 1997.
216. Iijima, S. et al., Changes of α1 -glycoprotein microheterogeneity in acute inflammation stages
analyzed by isoelectric focusing using serum obtained postoperatively, Electrophoresis, 21,
753, 2000.
217. Higai, K. et al., Glycosylation of site-specific glycans of α1 -acid glycoprotein and alterations in acute
ad chronic inflammation, Biochim. Biophys. Acta, 1725, 128, 2005.
218. Higai, K. et al., Altered glycosylation of α1 -acid glycoprotein in patients with inflammation and
diabetes mellitus, Clin. Chim. Acta, 329, 117, 2003.
219. Kakehi, K. et al., Capillary electrophoresis of sialic acid-containing glycoprotein. Effect of the het-
erogeneity of carbohydrate chains on glycoform separation using an α1 -acid glycoprotein as a model,
Anal. Chem., 73, 2640, 2001.
220. Sei, K. et al., Collection of α1 -acid glycoprotein molecular species by capillary electrophoresis and
the analysis of their molecular masses and carbohydrate chains. Basic studies on the analysis of
glycoprotein glycoforms, J. Chromatogr. A, 958, 273, 2002.
221. Hiroaka, A. et al., Charge microheterogeneity of the β-trace proteins (lipocalin-type prostaglandin
D synthase) in the cerebrospinal fluid of patients with neurological disorders analyzed by capillary
isoelectric focusing, Electrophoresis, 22, 3433, 2001.
222. Lilja, H., Biology of prostate-specific antigen, Urology, 62, 27, 2003.
223. Peracaula, R. et al., Altered glycosylation pattern allows the distinction between prostate-specific
antigen (PSA) from normal and tumor origins, Glycobiology, 13, 457, 2003.
224. Guzman, N.A., Immunoaffinity capillary electrophoresis applications of clinical and pharmaceutical
relevance, Anal. Bioanal. Chem., 378, 37, 2004.
225. Mao, X., Chu, I.K., and Lin B., A sheath-flow nanoelectrospray interface of microchip electrophoresis
MS for glycoprotein and glycopeptide analysis, Electrophoresis, 27, 5059, 2006.
23 Capillary Electrophoresis of
Post-Translationally Modified
Proteins and Peptides
Bettina Sarg and Herbert H. Lindner
CONTENTS
23.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 707

23.3 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 711
23.3.1 Proteins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 711
23.3.1.1 Phosphorylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 711
23.3.1.2 Acetylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 713
23.3.1.3 Deamidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 714
23.3.1.4 Methylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 715
23.3.2 Peptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 716
23.3.2.1 Phosphorylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 716
23.3.2.2 Deamidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 717
23.3.2.3 Farnesylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 717
23.4.1 Buffers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 718
23.4.2 Capillary Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 718
23.4.3 Method Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 718
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 719
23.1 INTRODUCTION
A large number of chemical modifications can occur in individual amino acids that can fundamentally
affect their physicochemical and functional properties. Over 200 distinct covalent modifications have
been reported, with phosphorylation, glycosylation, acetylation, methylation, and ADP-ribosylation
being the most common. Some amino acids can be converted into other amino acids, for example,
asparagine in aspartic acid or glutamine in glutamic acid by deamidation. Knowledge of these mod-
ifications is extremely important because they may alter physical and chemical properties, folding,
conformation distribution, stability, activity, and consequently, function of the proteins.
Phosphorylation, principally on serine, threonine, or tyrosine residues, is one of the most
important and abundant post-translational modifications (PTMs), with more than 30% of pro-
teins being modified by the covalent attachment of one or more phosphate groups. It plays a
critical role in the regulation of various cellular processes including cell cycle, growth, apop-
tosis, and transmitting extracellular signals to the nucleus. In fact, protein phosphorylation is
707
probably the single most common intracellular signal transduction event. Owing to its biochemical
importance, various analytical techniques for the detection and analysis of protein phosphoryla-
tion have been described.1 A widely used approach for the detection of phosphorylation involves
metabolic labeling with [32 P]orthophosphate followed by one- or two-dimensional polyacrylamide
gel electrophoresis. Detection of phosphorylation sites is based on radiography combined with
enzymatic digestions of the separated proteins and two-dimensional phosphopeptide mapping. High-
performance capillary electrophoresis (HPCE) offers substantial method inherent advantages over
such protocols, for example, a more rapid and accurate analysis with less sample consumption, while
avoiding hazardous radioactive labeling.
In vivo methylation of the side chains of specific arginines, histidines, and lysines in proteins
is a common phenomenon in nature involving numerous classes of proteins in both prokary-
otic and eukaryotic cells.2,3 Methylation has been most well studied in histones, with distinct
lysine residues mono-, di-, or tri-methylated playing a major role in the regulation of gene
expression, DNA replication, and repair. Methylated amino acids have often been determined in
protein and tissue hydrolysates using amino acids analyzers and through cells radiolabeled with
[methyl-3 H]methionine.
Acetylation of proteins occurs in two different ways. The N-terminal acetylation is one of the
most common protein modification reactions in eukaryotes and it is estimated that up to about 85% of
all mammalian proteins are affected. It is an irreversible process occurring cotranslationally, unlike
the reversible side chain acetylation of internal lysine residues, most famously for histones and
transcription factors that affect selective gene transcription and chromatin structure. Despite many
hypotheses about the role of N-terminal acetylation, its biological significance is still unclear.
Protein glycosylation is a PTM of eukaryotic proteins and does not occur in prokaryotes. The
carbohydrate groups are highly variable with significant effects on protein folding, stability, and
activity. The importance of HPCE for the analysis of glycosylated proteins is described in Chapter 22.
Nonenzymatic deamidation of peptides or proteins represents an important degradation reaction
occurring in vitro in the course of isolation or storage and in vivo during development and/or aging
of cells.4,5 Deamidation is a hydrolytic reaction resulting not only in the introduction of negative
charges but also in a change in the primary structure of proteins or peptides. Deamidation is a
common PTM resulting in the conversion of an asparagine residue to a mixture of isoaspartate
and aspartate. Deamidation of glutamine residues can occur but does so at a much lower rate.
Detecting and separating deamidated forms from the parent molecules are still problematic aspects
of protein analysis.
HPCE continues to become more widely used for the detection, separation, and quantification
of peptides and proteins, since the use of capillaries greatly reduces sample volume and analysis
time compared to conventional gel electrophoresis. This offers perfect qualification for the analysis
of PTMs, but surprisingly only a small number of HPCE applications have been developed during
the past years. Since these modifications represent phenomena that occur in vivo often in very small
amounts, they can be easily missed. This may explain why in certain cases recombinantly expressed
proteins have an altered or absent activity compared with the naturally expressed proteins.
Taking this into account, more sensitive methods of sample detection have been developed,
for example, special detection cell constructions for ultraviolet (UV) adsorption and laser-induced
fluorescence (LIF), and particularly the introduction of mass spectrometry (MS) brought tremendous
progress in online and offline characterization not only of modified peptides but also of modified
proteins. In this chapter, we try to provide information about capillary electrophoresis (CE) methods
developed for the separation of proteins and peptides with various PTMs.

More or less all modification reactions occurring on proteins may alter their overall charge either
by introducing or eliminating charges, thereby changing to some extent the m/z ratio and the pI of
Capillary Electrophoresis of Post-Translationally Modified Proteins and Peptides 709
the parent proteins. For this reason, capillary zone electrophoresis (CZE) and isoelectric focusing
(IEF) are CE modes that are especially well suited to separate post-translationally modified forms.
However, the ease of use makes CZE the preferentially used method in the laboratory and only very
few applications employing other CE modes can be found in the literature.
Depending on the pI of the proteins and the pH of the buffers used for the separation in CZE
proteins are either positively or negatively charged (except when pI = pH, where the overall charge
of the proteins is zero). In the course of phosphorylation, for example, negatively charged phosphate
groups are bound to uncharged amino acids such as Ser, Thr, or Tyr. Also, glycosylation has an impact
on protein charge if one or more charged sialic acid molecules are attached (detailed in Chapter 22).
In both cases, modification causes either a decrease in the overall positive charge or at appropriate
buffer pH conditions an increase in the negative charge of the modified forms.
Acetylation can take place either on the N-terminal amino group of the first amino acid of a
protein or at the ε-amino group of lysines. As basic amino groups, forming cationic ammonium ions
under usual buffer conditions, become neutral due to amide formation one positive charge for each
acetyl group bound will be removed in course of this modification reaction. This charge difference
enables the separation of distinctly acetylated proteins from each other and from their unacetylated
form.
Another modification of lysine, its mono-, di-, and, tri-methylation, has been the focus of great
attention in histones due to its biological importance in gene regulation. As expected, the impact of
methylation on charge is not as pronounced as of acetylation; however, electron donor effects of the
alkyl group slightly increase the charge density of the nitrogen of the amino group. These differences
can be sufficiently high for a CE separation provided the molecular mass of the protein is low.
An essential and in terms of its biological importance as well as of its frequency often underes-
timated modification reaction of proteins and peptides occurring under both physiological and labo-
ratory conditions involves the nonenzymatic spontaneous deamidation of asparagine and glutamine
residues to aspartate and glutamate, respectively. The deamidation of Asn, which takes place much
more frequently than that of Gln, follows a rather complex mechanism via a cyclic imide intermediate.
End products of its hydrolytic cleavage are the formation of isoaspartic acid as main reaction product
and aspartic acid. In certain cases also truncation of the protein backbone can occur.4 Under buffer
conditions where proteins are cations (pH < pI) the presence of both isoaspartic and aspartic acid
reduces the overall positive charge of the proteins compared to the Asn containing protein. Moreover,
due to minor differences in the pKa values of iso-Asp and Asp (pKa isoAsp < pKa Asp) a separa-
tion of these two isomeric protein forms can be achieved if pH of the separation buffer is adjusted
thoroughly.
Certain protein classes are known to be multiply modified. Histone proteins are most probably
the best investigated representatives in this respect. Under certain biological conditions, they are
known to be acetylated, phosphorylated, and methylated, for example, in a definite manner at
the same time. Several examples of successful CE separations of even multiply modified proteins
are reported in the literature.6,7
Problems in CE occurring with the analysis of PTMs are usually not related to the presence of
the modifying residues. They are much more associated with problems generally occurring when
proteins are analyzed by capillary electrophoresis. Proteins stick to many different surfaces including
metals, plastics, and also glass, which are the most commonly used materials for capillaries in CE.
Depending on the extent of interaction between proteins and the glass surface of the capillary peak
tailing, loss of resolution, reduced sensitivity, or even total adsorption can occur. The primary
causes for these detrimental effects are ionic interactions between cationic protein regions and the
negatively charged silanol groups of the capillary surface. Moreover, also hydrophobic interactions
may contribute to these adsorption phenomena.
A variety of strategies were developed in the past to overcome these troublesome effects. Among
these, a few are very simple to accomplish, for example, avoiding very low buffer concentrations.
Increasing the buffer concentration decreases protein–wall interactions by reducing the effective
0.030 H1.5
A 200 0.020
0.010
0.000
20 22 24 26 28 30
Time (min)
FIGURE 23.1 CZE separation of phosphorylated histone H1.5 from human tumor cells (CCRF-CEM) in a
100 mM phosphate buffer (pH 2.0). Other conditions: voltage 12 kV, temperature 25◦ C, injection 2 s, detection
200 nm, untreated capillary (50 cm × 75 µm i.d.).
surface charge and, depending on buffer type also ion-pairing effects come into play. Another
approach to limit adsorption is simply by working at pH extremes: At low pH (∼ 2) silanol groups
are protonated, their dissociation is significantly reduced and in this way the interaction of the under
these conditions cationic proteins with the now more or less uncharged surface minimized. Con-
versely, at high pH (∼10) both the silica surface and the proteins will be deprotonated and negatively
charged. In this case, adsorption is remarkably reduced as a result of a charge repulsion effect. How-
ever, while these approaches may be very successful strategies for the separation of modified forms
of peptides and some small proteins (<5 kDa), their application is very limited to the separation
of “normal sized” proteins as in this case undesired wall interactions cannot sufficiently enough be
suppressed. A striking example is illustrated in Figure 23.1. It shows the separation of a mixture
of distinctly phosphorylated and nonphosphorylated human linker histone H1.5 using an uncoated
capillary and a sodium phosphate buffer with pH 2.0. Owing to their remarkably basic (pI > 10) and
also hydrophobic properties, histone proteins particularly strongly interfere with the inner surface of
the silica wall. From Figure 23.1, it is evident that even at this low pH essential electrostatic protein–
wall interactions still occur, which are responsible for the poor resolution, broad peaks, and low
sensitivity.
In contrast to separations performed at low pH, working at high pH is not always feasible for the
following reasons:
1. Artificial protein modification reactions can be induced, for example, deamidation of

proteins.
2. At very high pH (>11) dissolution of the silica becomes an issue.
For this reason, distinct coating procedures, often combined with low pH buffers, have been
found to be a prerequisite for the successful separation of proteins as well as of their PTMs.
Two basic strategies have been applied in CE to limit protein adsorption.
1. Permanent modification by covalently bonded or physically adhered phases. Basically,

hydrophilic polymers, polar functional groups, or even positively charged residues, for
example, amino groups are bound to the silica surface.8 In general, life time of the per-
manent coating and batch to batch reproducibility may still be a problem. Therefore, it is
not so surprising that most applications for the separation of protein modifications, which
can be found in the literature, are based on an alternative approach to suppress undesired
protein–surface interactions.
2. It is the concept of dynamically coating the inner wall of the capillary by the addition of
suitable additives to the running buffer. These additives themselves show high affinity to
the silica surface and act like a shield to prevent positively charged proteins from coming
into close contact with the capillary wall. Various compounds have been used for this
purpose; for instance, neutral and cationic hydrophilic polymers, cationic hydrophobic
polymers, diamines, and so on.8 Outstanding advantages of this approach are stability of
the coating and simplicity of handling. Since the coating agent is in the buffer, the coating
layer of the capillary surface is continuously regenerated and no permanent stability is
required. Moreover, cleaning of the capillary can be performed easily by rinsing with
0.1 M sodium hydroxide and water. However, potential disadvantages of the dynamic
coating approach are effects of the surfactants, for example, influencing protein structure,
incompatibility with MS analysis, range of pH applicable can be limited, time for equi-
librium needed to obtain reproducible surface coating, surface properties may influence
quality and reproducibility of the coating.
As a consequence, both methods have advantages and shortcomings and no method is clearly
superior. For this reason, depending on protein primary structure and type of modification method
development is still an issue; however, many promising recipes for their successful separation are
available in the meantime.
23.3 APPLICATIONS
Many proteins such as histones, ribonucleic acid (RNA) polymerase II, tubulin, myelin basic protein,
p53, and tyrosine kinases are post-translationally modified at multiple sites. Among them histones are
the most extensively studied group as their hydrophilic N- and C-terminal tail domains are subjected
to a great variety of PTMs including phosphorylation, acetylation, methylation, deamidation, ubiq-
uitination, and ADP-ribosylation. Distinct combinations of covalent histone modifications including
lysine acetylation, lysine and arginine methylation, and serine phosphorylation form the basis of
the histone code hypothesis.9–11 This hypothesis proposes that a pre-existing modification affects
subsequent modifications on histone tails and that these modifications generate unique surfaces for
the binding of various proteins or protein complexes responsible for higher-order chromatin organi-
zation and gene activation and inactivation. Owing to their biological importance, it is not surprising
that a variety of methods were developed particularly for the CE separation of histone modifications,
more than for any other protein family.
This chapter summarizes practical applications of HPCE on the analysis of various modified
histones, other proteins and, moreover, also some modified peptide separations are described.
23.3.1 PROTEINS
23.3.1.1 Phosphorylation
A set of different microscale techniques of CE exists for analyzing phosphorylated proteins and
peptides. The effectiveness of a buffer system containing hydroxypropylmethyl cellulose (HPMC)
as dynamic coating agent in combination with low pH preventing undesired interactions of positively
charged histone proteins with the silica surface could be first demonstrated by Lindner et al.12–14 A
complex mixture of rat testis H1 histones consisting of eight microsequence variants and, in addition,
0.040 H1.5 p2
p1
0.030
A200
0.020
p0
p5
p3 p4
0.010
0.000
16 17 18 19
Time (min)
FIGURE 23.2 CE separation of phosphorylated histone H1.5 from human tumor cells (CCRF-CEM) in a
100 mM phosphate buffer (pH 2.0) and in the presence of 0.02% HPMC. Other conditions: voltage 12 kV,
temperature 25◦ C, injection 2 s, detection 200 nm, untreated capillary (50 cm × 75 µm i.d.). Designations
p0–p5 = non-, mono-, di-, tri-, tetra-, and pentaphosphorylated histone H1.5.
various phosphorylated forms, was separated using an uncoated capillary and a sodium phosphate
buffer (pH 2.0) and 0.03% HPMC.Astriking example for the resolving power of this separation buffer
system is depicted in Figure 23.2, illustrating the CE separation of the same multiphosphorylated
histone H1.5 sample already shown in Figure 23.1. Applying the same CE conditions as described
in Figure 23.1, except that 0.02% HPMC was added to the phosphate buffer, all five phosphorylated
forms are now clearly separated from each other and from the unphosphorylated parent protein.
The nonphosphorylated protein migrates fastest, followed by the distinctly phosphorylated forms,
because binding of the negatively charged phosphate groups decrease the overall positive charge
of the histone molecule thus diminishing the electrophoretic mobility of the phosphorylated protein
species.
Recently, Yoon et al.15 developed a CE-LIF method to determine phosphorylation levels and to
follow the translocation of the green fluorescent protein-extracellular signal regulated protein kinase
2 (GFP-ERK2) from the cytoplasm to the nucleus. CE conditions applied were an untreated capillary
and a 100 mM CAPS buffer (pH 11.0) containing 2 M betaine. LIF detection was performed with a
5 mW air-cooled argon ion laser (excitation, 488 nm/emission, 520 nm). Phosphorylated and non-
phosphorylated GFP-ERK2 were not separated simultaneously, but in consecutive runs. Significant
differences in migration time allowed a clear assignment of the modified and unmodified protein.
Owing to separation conditions employed (high pH, high salt buffer), the GFP-ERK2 proteins are
negatively charged and, therefore, no dynamic coating agent was added. Phosphorylation causes
a further increase in the negative charge; thereby, phosphorylated GFP-ERK2 migrates slower.
Compared to conventional Western blotting, the CE method allows the analysis of sample vol-
umes as low as a few nanoliters and does not require a separate sample purification step and
radiolabeling.
On the basis of their differences in isoelectric point (pI), phosphorylated proteins can be separated
in capillary isoelectric focusing (CIEF). Wei et al.16 employed a CIEF method using a capillary
covalently coated with linear polyacrylamide and a pH gradient from 4 to 6.5 for the resolution of
mono- and diphosphorylated ovalbumins. Proteins were detected by their UV absorbance at 280 nm.
Additional ovalbumin variants within each of the mono- and diphosphoovalbumins, differing in their
amount of glycosylation, were further analyzed by online CIEF-electrospray ionization (ESI)-MS.
In a similar manner, non-, mono-, and diphosphorylated myosin light chain using CIEF, either
with UV or LIF detection, was separated by Shiraishi et al.17 Neutral coated capillaries (eCAP;
Beckman Coulter) and HPMC in the ampholyte solution reduced electroosmotic flow (EOF) and
protein adsorption. A detection limit of ∼1 pg fluorescently labeled myosin light chain/capillary was
achieved.
23.3.1.2 Acetylation
Histone H4 with a calculated pI of 11.9 is one of the most basic proteins known. Under certain
biological conditions, it can be reversibly acetylated in its N-terminal region at four lysine residues.
CE separation of these different acetylated forms was not possible applying chemically coated
capillaries. However, when HPMC was used as a dynamic coating agent, the histone H4 sample
isolated from whole histones by reversed phase chromatography (RPC) was clearly resolved into
the non-, mono-, di-, tri-, and tetra-acetylated forms within about 22 min.18 Applying a special
buffer system consisting of 500 mM formic acid/LiOH/10 mM urea (pH 2.0) with 0.02% HPMC, an
ultrafast CE separation (4 min) of the distinctly acetylated H4 proteins could be achieved (shown in
Figure 23.3). As acetylation of lysines decreases the positive charge of histone H4, the unacetylated
protein migrates fastest.
Using HPMC as buffer additive another core histone, subfraction H2A, which is a single peak in
RPC, could be further separated by CE into five peaks consisting of non-and monoacetylated H2A.2a
and H2A.2b, respectively, and even a third subtype H2A.3 was resolved.19
Wiktorowicz and Colburn20 separated core histone H4 into its different acetylated forms using
a commercially available cationic surfactant (MicroCoat, ABI). Coating of the silica surface was
performed by rinsing the capillary with the reagent followed by a wash with running buffer. Like
other cationic surfactants such as cetyl trimethylammonium bromide (CTAB), the coating reagent
is primarily bound to the silica surface by ionic interaction. In a second step, the surface-bound
neutralized surfactant binds additional reagent cations by virtue of hydrophobic interactions between
alkyl side groups. In this way, a remarkably stable bilayer is formed and the charge of the capillary
wall is reversed from negative to positive. Consequently, the positively charged histone molecules
0.008
H4
ac0
0.006
0.004
A200
0.002 ac1
ac2
0.000 ac3
ac4
–0.002
3.5 3.7 3.9 4.1 4.3 4.5
Time (min)
FIGURE 23.3 Electropherogram of multiacetylated histone H4 from mouse tumor cells (NIH) in 500 mM
formic acid/LiOH/10 mM urea buffer (pH 2.0) containing 0.02% HPMC. Conditions: voltage 30 kV, injection
1 s. Other conditions are as in Figure 23.2. Designations ac0–ac4 = non-, mono-, di-, tri-, and tetra-acetylated
histone H4.
are repelled from the capillary surface and under the influence of an electric field, a reversal of the
EOF takes place. It should be mentioned, however, that strictly speaking this particular approach of
charge reversal is a special case of the dynamic coating concept, as no coating reagent is included in
the running buffer. For this reason, periodic replenishment of the coating by flushing the capillary
with the surfactant solution is required.
Besides histones, very little is described about the CE analysis of other acetylated proteins.
Some articles report on the separation of N-terminally acetylated and unacetylated forms of met-
allothioneins (MTs). MTs are structurally unusual proteins due to their small size (6–7 kDa), high
cysteine and metal content, and remarkable kinetic lability. Purified rat MT-2 protein was separated
by MEKC using an electrolyte of 100 mM sodium borate buffer containing 75 mM SDS at pH 8.4 and
showed two peaks.21 Identification of both peaks was performed with online CE-MS using 100 mM
formic acid and 2% methanol. The two peaks differed by a mass equivalent to that of a single acetyl
group. Another group established a CE-ESI-time of flight (TOF)-MS method to characterize rabbit
liver MT isoforms and was successful in separating several N-acetylated and non-N-acetylated forms
using 100 mM acetic acid:100 mM formic acid (pH 2.3).22
CIEF has been applied to the separation and quantification of the three main hemoglobin compo-
nents of umbilical cord blood (fetal, acetylated fetal, and adult hemoglobins).23 CIEF was performed
with a poly (acryloylaminoethoxy-ethonal) [poly(AAEE)] coated capillary and a carrier ampholyte
consisting of 5% Ampholine pH 6–8, supplemented with 0.5% TEMED.
An interesting CE application has been the use of so-called charge ladders of proteins for measur-
ing the role of charge in protein stability, protein–ligand binding, and ultrafiltration.24–26 A protein
charge ladder is a collection of derivatives of a protein by converting its charged groups into elec-
trically neutral ones. Charge ladders can be easily generated in vitro by the treatment of a variety
of model proteins with acetic anhydride. CE has been shown to be an effective tool in the analysis
of such charge ladders, as it is used to separate the proteins that constitute the charge ladder into
individual “rungs,” each rung contains derivatives with the same number of modified groups.
Cordova et al.27 explored the CE behavior of protein charge ladders obtained by acetylation
of lysozyme and carbonic anhydrase II using noncovalent polycationic coated capillaries. Two of
them, polyethylenimine and Polybrene, were very effective in preventing the adsorption of positively
charged proteins. Conditions used were fused-silica capillary (50 µm i.d. × 38 cm) coated with the
polymer by flushing the capillary with a 7.5% (w/v) polymer solution, prepared in 25 mM Tris–
192 mM Gly buffer (pH 8.3), for 15 min. The running buffer was 25 mM Tris–192 mM Gly buffer
(pH 8.3) in the absence of polymer. Separations were obtained within 5 min.
Carbeck et al.28 showed CE-ESI MS to be a useful tool for the study of charge ladders of
lysozyme, carbonic anhydrase II, and bovine pancreatic trypsin inhibitor and for the examination of
the relationship between the properties of proteins in the solution phase and in the gas phase.
An example for a CE separation of such a charge ladder is shown in Figure 23.6 (Section 23.4.3).
Separation of multiply modified proteins place high demands on the method applied. An example
for a successful separation of a protein containing differently phosphorylated and acetylated forms
in a single run using the HPMC-based CZE method is shown in Figure 23.4.
23.3.1.3 Deamidation
The analysis of recombinant human growth factor (rhGH), one of the first biotechnologically pro-
duced proteins in Escherichia coli, by CE with UV absorbance and MS detection using bilayer-coated
capillaries was demonstrated very recently.29 The authors present an improved CE method using
capillaries noncovalently coated with polybrene and poly(vinyl sulfonic acid) and a background
electrolyte of 400 mM Tris phosphate (pH 8.5) to achieve efficient separations of intact rhGH
and degradation products like deamidated and oxidated forms.
0.050 H5
ac0p0
0.040
0.030
A200
0.020 ac1p0
ac0
p1 p2
0.010
ac1p1
0.000
20 21 22 23 24
Time (min)
FIGURE 23.4 Electropherogram of histone H5 from chicken erythrocytes. Conditions were the same as in
Figure 23.2. Designations ac0p0 = nonacetylated nonphosphorylated H5; ac0p1 and ac0p2 = nonacetylated
mono- and diphosphorylated H5; ac1p0 and ac1p1 = monoacetylated non-and monophosphorylated histone H5.
A novel method for the determination of glutamine deamidation in a long repetitive protein poly-
mer via bioconjugate capillary electrophoresis was published recently by Won et al.30 By conjugating
a monodisperse, fluorescently labeled DNA oligomer to long polydisperse nearly electrostatically
neutral protein polymers, protein polymers differing in degree of deamidation were successfully sep-
arated. For protein polymers with increasing extents of deamidation, the electromotive force of DNA
+ polypeptide conjugate molecules increases due to the introduced negative charge of deamidated
glutamic acid residues. CE analysis reveals increasing differences in the electrophoretic mobilities
of conjugate molecules, which qualitatively shows the degree of deamidation. CE of protein–DNA
conjugates was performed using a capillary coated with POP-5 polymer and a 50 mM Tris, 50 mM
TAPS, 2 mM EDTA buffer containing 7 M urea (pH 8.4) added with 3% POP-5 solution.
Histone H10 is known to be deamidated in the course of aging.5 Intact H10 contains asparagine in
position 3, while depending on the age of the organ deamidated forms of H10 containing either aspartic
acid or isoaspartic acid accumulate. The buffer system containing HPMC enables the separation of
intact histone H10 fragment (residues 1–52) as well as its deamidated forms (Figure 23.5).
23.3.1.4 Methylation
Unlike protein modification by acetylation or phosphorylation, methylation does not greatly influ-
ence the charge of individual amino acids, thus making the electrophoretic separation of distinct
methylated proteins from each other and from the unmethylated parent protein a problematic part.
Separation of acetylated histone H4 has already been described by Lindner et al.18 using CZE with
HPMC as dynamic coating reagent. The same method enables the simultaneous separation of mono-,
di-, and trimethylated H4, including their distinct acetylated forms.6 Changes in the modification
pattern of H4 from normal tissues, cancer cell lines, and primary tumors were found suggesting
that a global loss of monoacetylation and trimethylation of histone H4 is a common hallmark of
human cancer.
H10 (1–51) Asn
Asp
A200
isoAsp
17 19 21 23 25
Time (min)
FIGURE 23.5 CE separation of deamidated proteins. Histone H10 from rat liver was digested with chy-
motrypsin and the N-terminal fragment obtained by RPC analyzed by CE. Conditions were the same as
in Figure 23.2. Designation Asn, Asp, isoAsp = asparagine, aspartic acid, and isoaspartic acid containing
fragments.
23.3.2 PEPTIDES
23.3.2.1 Phosphorylation
A CZE procedure for detection and assay of protein kinase and phosphatase activities in complex
biological mixtures has been developed by Dawson et al.31 The phosphorylated and dephosphorylated
forms of several peptides were resolved using an uncoated capillary and a 150 mM phosphoric
acid buffer at pH 2.0 or pH 5.0. Furthermore, the CZE-based assay was capable of resolving a
peptide phosphorylated on different sites and permitted the quantification of each peptide. Since this
application, much research has been done on protein kinase assays based on CE.
For the same purpose, Gamble et al.32 tested both neutral and cationic coated and uncoated
capillaries and found that the best conditions for the separation of phosphopeptide isomers is formic
acid or phosphate buffer with pH ranging between 5.5 and 6.5 and a PVA-coated capillary. Synthetic
peptides were incubated with various protein kinases and phosphatases and the phosphorylation and
dephosphorylation was monitored using CZE. These methods enabled the assay of several protein
kinases and phosphatases and the determination of the sites of phosphorylation.
Improved CE-based methods to measure kinase activation in single cells have been described
recently.33 Phosphorylated and nonphosphorylated forms of peptide substrates for protein kinase C
and calcium-calmodulin activated kinase were separated by CZE using a polydimethylacrylamide
(PDMA) coated capillary and buffers containing different concentrations of betaine (0–1 M). The
separation system is compatible with a living cell and, therefore, adaptable to the laser micropipet
system, a strategy to measure the activation of enzymes in single mammalian cells.
Synthetic peptides containing phosphorylated tyrosine have been shown to be easily separated by
a CZE method using a linear polyacrylamide-coated capillary and a 300 mM borate buffer (pH 8.5).34
Another separation mode of HPCE, namely MEKC, has been used to separate a mixture of phos-
phopeptide isomers of the insulin receptor peptide.35 The resolution in coated and uncoated capillaries
was compared using a 50 mM phosphate buffer with 25 mM SDS (pH 6.1). An efficient separation
of diphosphorylated isomers could be obtained by using polyacrylamide-coated capillaries.
Mass spectrometry of phosphopeptides has become a powerful tool for phosphorylation site
identification. However, proteolytic digests examined by MS are often likely to fail to detect phos-
phopeptides because the ionization of phosphorylated peptides in positive ion mode MS is generally
less efficient compared with the ionization of their nonphosphorylated counterparts resulting in ion
suppression effects. A further problem is that phosphopeptides may not be retained by RP chro-
matography because they are too small and/or hydrophilic to bind to the C18 stationary phase.
Therefore, capillary electrophoresis coupled to MS is a powerful method to enhance the detection
of phosphoproteins and phosphopeptides due to their efficient separation by CE.
Sandra et al.36 developed a CE-MS technique to the characterization of the N-glycans from the
glycoprotein cellobiohydrolase I (CBH I) with special interest for the phosphorylated species. CBH I
shows phosphorylated glycans that are only present when the organism is grown under minimal
conditions and could be related to stress response. The glycans were labeled with the negatively
charged tag 8-aminopyrene-1,3,6-trisulfonate (APTS) by reductive amination and separated using
an untreated capillary and a 25 mM ammonium acetate buffer (pH 4.55), allowing the simultaneous
analysis of uncharged and charged glycans and, moreover, the differentiation of phosphorylated
isomers.
Detection of phosphorylated proteins and peptides can be enhanced by selectively isolating
these species. Online immobilized metal affinity chromatography (IMAC)-CE-ESI-MS is such a
powerful analytical tool. The IMAC resin retains and preconcentrates phosphorylated proteins and
peptides, CE separates the phosphorylated species and MS/MS identifies the components and their
phosphorylation sites. Cao and Stults37,38 applied this method to the analysis of phosphorylated
angiotensin II and tryptic digests of α- and β-casein (CE conditions: buffer, 0.1% acetic acid/10%
methanol; uncoated capillary). Beta-casein is a well-characterized protein with five phosphorylation
sites and is widely used as a standard for protein phosphorylation studies.
23.3.2.2 Deamidation
The effectiveness of CZE for the separation of deamidated peptides was shown by Ganzler et al.39
Excellent resolution of peptides containing all three forms, Asn, Asp, and isoAsp, could be achieved
by using a 20 mM sodium citrate buffer (pH 2.5).
23.3.2.3 Farnesylation
Protein farnesylation, catalyzed by protein farnesyltransferase, plays important roles in the membrane
association and protein–protein interaction of a number of eukaryotic proteins. The enzyme transfers
a 15-carbon farnesyl moiety from farnesyl diphosphate (FPP) to the sulfhydryl group of cysteine.
The activity of the enzyme was measured by CE with LIF detection, which is a powerful alternative
to classical methods involving radiolabeled FPP.40 LIF detection was performed with an argon ion
laser (excitation, 488 nm/emission, 520 nm). A fluorescently labeled pentapeptide that was used
as substrate was clearly separated from its farnesylated form under the four CE buffer conditions
investigated (e.g., 25 mM borax, 25 mM SDS, pH 9.3, uncoated capillary). The method will be of
great value in studies of inhibitors of protein farnesyltransferase in vitro.

Many factors may influence the quality of CE separations and contribute to the resolution of protein
modifications. Generally, optimization strategies, precautions, and CE conditions applied to peptides
are also applicable to the separation of modified proteins. However, optimum parameters for protein
separations must still be determined empirically to some extent. In addition to well-known factors
influencing resolution in CE like capillary surface and coating, buffer pH, also effects of different
buffer cations and anions, buffer concentration, additives, and chaotropic agents can be utilized to
enhance resolution of protein mixtures.
23.4.1 BUFFERS
The running buffer selection is substantial to the success of any CE separation. Factors being most
important are the UV absorbance at low wavelengths, good pH stability, and conductivity. The
highest sensitivity detection is achieved at low UV wavelengths (190–220 nm) where peptide bonds
have an absorbance maximum. Good sensitivity at these wavelengths requires buffer systems with
high-UV transparency. The most frequently used buffers are sodium phosphate, sodium citrate, Tris,
and sodium borate. Sodium formate shows optimal compatibility to mass spectrometric analysis and
is therefore often used in CE coupled online to MS.
The buffer pH should be stable during a series of analyses. For reproducible analysis times the
buffer should be exchanged or remixed every 5–10 runs. Cellulose-based buffers should be freshly
prepared at least once a week, depending on the number of runs, and should not(!) be stored in
the refrigerator. The quality and purity of all buffer components should be of the highest grade
and the buffers should be filtered through 0.45 µm pore size filters before use. A starting buffer,
which gives superior results in many cases, is a 100 mM sodium phosphate buffer (pH 2.5) with the
addition of 0.02% HPMC. Owing to the excellent UV transparency of the separation buffer, proteins
can be detected very sensitively at 200 nm. In case this system does not provide satisfying results,
more complex separation conditions must be taken into account according to specific properties of
the proteins (e.g., pI, hydrophobicity, etc.).
23.4.2 CAPILLARY TREATMENT

Using an untreated capillary every 10–15 injections the capillary should be rinsed with water, 0.1 M
NaOH, water, 0.5 M H2 SO4 , and finally with the running buffer. Washing should be done for 2 min
with each solvent. The same wash, but without running buffer, should be applied at the end of sample
analyses. After flushing the capillary with air, it can be stored.
There are many different static wall coatings described in the literature, recently reviewed by
Horvath and Dolnik.8 Static coated capillaries should be treated according to the manufacturer’s
guidelines. Solvents such as NaOH and H2 SO4 must not be used.
23.4.3 METHOD EVALUATION

The evaluation of the applicability of a certain CE method for the separation of modified proteins
can easily be performed by analyzing acetylated lysozyme generated by the treatment of the protein
with acetic anhydride (described in Section 23.3.1.2). An aqueous solution of lysozyme (0.1 mM,
0.5 mL) adjusted to pH 10 by using 0.1 M NaOH is incubated with 4.5 µL acetic anhydride (100 mM
in dioxane) for 5 min at room temperature according to Cordova et al.27 An example using the
suggested phosphate buffer with HPMC is shown in Figure 23.6.

The analysis of post-translationally modified proteins is due to their biological significance a field of
ever increasing importance. For many years, different kinds of PAGE have been the most commonly
used tool for studying modified proteins with all the shortcomings of this technique generally known.
The advent of capillary electrophoresis provided a promising new tool for their separation. HPCE has
made significant advances in recent years and is establishing itself as a rapid, reproducible, sensitive,
and highly resolving method. Hyphenated techniques, for example, the combination with MS, will
0.20
(a)
A200
0.10
0.00
0.02
(b)
A200
0.01
0.00
10 15 20 25 30
Time (min)
FIGURE 23.6 CE analysis of the lysozyme charge ladder. (a) Untreated lysozyme and (b) charge ladder from
acetylation of lysozyme. Conditions were the same as in Figure 23.2.
be able to provide detailed structural information like modification status and sites and will further
increase the range of potential CE applications in this field.
ACKNOWLEDGMENTS
This work, as part of the European Science Foundation EUROCORES Programme EuroDYNA,
was partly supported by funds from the Austrian Science Foundation (project I23-B03) and by the
Jubilee Fund of the Austrian National Bank, grant no. 9319.
REFERENCES
1. Yan, J. X., N. H. Packer, A. A. Gooley, and K. L. Williams, Protein phosphorylation: Technologies for
the identification of phosphoamino acids, J. Chromatogr. A, 808, 23–41, 1998.
2. Paik, W. K. and S. Kim, Protein Methylation, John Wiley & Sons, New York, 1980.
3. Wold, F., In vivo chemical modification of proteins (post-translational modification), Annu. Rev.
Biochem., 50, 783–814, 1981.
4. Lindner, H. and W. Helliger, Age-dependent deamidation of asparagine residues in proteins, Exp.
Gerontol., 36, 1551–1563, 2001.
5. Lindner, H., B. Sarg, B. Hoertnagl, and W. Helliger, The microheterogeneity of the mammalian H1(0)
histone. Evidence for an age-dependent deamidation, J. Biol. Chem., 273, 13324–13330, 1998.
6. Fraga, M. F., E. Ballestar, A. Villar-Garea, M. Boix-Chornet, J. Espada, G. Schotta, T. Bonaldi, et al.,
Loss of acetylation at Lys16 and trimethylation at Lys20 of histone H4 is a common hallmark of human
cancer, Nat. Genet., 37, 391–400, 2005.
7. Sarg, B., W. Helliger, B. Hoertnagl, B. Puschendorf, and H. Lindner, The N-terminally acetylated
form of mammalian histone H1(0), but not that of avian histone H5, increases with age, ABB, 372,
333–339, 1999.
8. Horvath, J. and V. Dolnik, Polymer wall coatings for capillary electrophoresis, Electrophoresis, 22,
644–655, 2001.
9. Strahl, B. D. and C. D. Allis, The language of covalent histone modifications, Nature, 403,
41–45, 2000.
10. Turner, B. M., Histone acetylation and an epigenetic code, Bioessays, 22, 836–845, 2000.
11. Jenuwein, T. and C. D. Allis, Translating the histone code, Science, 293, 1074–1080, 2001.
12. Lindner, H., W. Helliger, A. Dirschlmayer, H. Talasz, M. Wurm, B. Sarg, M. Jaquemar, and B. Puschen-
dorf, Separation of phosphorylated histone H1 variants by high- performance capillary electrophoresis,
J. Chromatogr., 608, 211–216, 1992.
13. Lindner, H., M. Wurm, A. Dirschlmayer, B. Sarg, and W. Helliger, Application of high-performance
capillary electrophoresis to the analysis of H1 histones, Electrophoresis, 14, 480–485, 1993.
14. Lindner, H., W. Helliger, B. Sarg, and C. Meraner, Effect of buffer composition on the migration order
and separation of histone H1 subtypes, Electrophoresis, 16, 604–610, 1995.
15. Yoon, S., K. Y. Han, H. S. Nam, L. V. Nga, and Y. S. Yoo, Determination of protein phosphorylation
and the translocation of green fluorescence protein-extracellular signal-regulated kinase 2 by capillary
electrophoresis using laser induced fluorescence detection, J. Chromatogr. A, 1056, 237–242, 2004.
16. Wei, J., L. Yang, A. K. Harrata, and C. S. Lee, High resolution analysis of protein phosphorylation
using capillary isoelectric focusing—electrospray ionization—mass spectrometry, Electrophoresis,
19, 2356–2360, 1998.
17. Shiraishi, M., R. D. Loutzenhiser, and M. P. Walsh, A highly sensitive method for quantification of
myosin light chain phosphorylation by capillary isoelectric focusing with laser-induced fluorescence
detection, Electrophoresis, 26, 571–580, 2005.
18. Lindner, H., W. Helliger, A. Dirschlmayer, M. Jaquemar, and B. Puschendorf, High-performance
capillary electrophoresis of core histones and their acetylated modified derivatives, Biochem. J., 283
(Pt 2), 467–471, 1992.
19. Lindner, H., B. Sarg, C. Meraner, and W. Helliger, Separation of acetylated core histones by
hydrophilic-interaction liquid chromatography, J. Chromatogr. A, 743, 137–144, 1996.
20. Wiktorowicz, J. E. and J. C. Colburn, Separation of cationic proteins via charge reversal in capillary
21. Beattie, J. H., A. M. Wood, and G. J. Duncan, Rat metallothionein-2 contains N(alpha)-acetylated and
unacetylated isoforms, Electrophoresis, 20, 1613–1618, 1999.
22. Andon, B., J. Barbosa, and V. Sanz-Nebot, Separation and characterization of rabbit liver apothioneins
by capillary electrophoresis coupled to electrospray ionization time-of-flight mass spectrometry,
23. Conti, M., C. Gelfi, and P. G. Righetti, Screening of umbilical cord blood hemoglobins by isoelectric
focusing in capillaries, Electrophoresis, 16, 1485–1491, 1995.
24. Negin, R. S. and J. D. Carbeck, Measurement of electrostatic interactions in protein folding with the
use of protein charge ladders, J. Am. Chem. Soc., 124, 2911–2916, 2002.
25. Gao, J., M. Mammen, and G. M. Whitesides, Evaluating electrostatic contributions to binding with the
use of protein charge ladders, Science, 272, 535–537, 1996.
26. Menon, M. K. and A. L. Zydney, Effect of ion binding on protein transport through ultrafiltration
membranes, Biotechnol. Bioeng., 63, 298–307, 1999.
27. Cordova, E., J. Gao, and G. M. Whitesides, Noncovalent polycationic coatings for capillaries in
capillary electrophoresis of proteins, Anal. Chem., 69, 1370–1379, 1997.
28. Carbeck, J. D., J. C. Severs, J. Gao, Q. Wu, R. D. Smith, and G. M. Whitesides, Correlation
between the charge of proteins in solution and in the gas phase investigated by protein charge lad-
ders, capillary electrophoresis, and electrospray ionization mass spectrometry, J. Phys. Chem. B, 102,
10596–10601, 1998.
29. Catai, J. R., T. J. Sastre, P. M. Jongen, G. J. de Jong, and G. W. Somsen, Analysis of recombinant
human growth hormone by capillary electrophoresis with bilayer-coated capillaries using UV and MS
detection, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci., 852, 160–166, 2007.
30. Won, J. I., R. J. Meagher, and A. E. Barron, Characterization of glutamine deamidation in a
long, repetitive protein polymer via bioconjugate capillary electrophoresis, Biomacromolecules, 5,
618–627, 2004.
31. Dawson, J. F., M. P. Boland, and C. F. Holmes, A capillary electrophoresis-based assay for protein
kinases and protein phosphatases using peptide substrates, Anal. Biochem., 220, 340–345, 1994.
32. Gamble, T. N., C. Ramachandran, and K. P. Bateman, Phosphopeptide isomer separation using
capillary zone electrophoresis for the study of protein kinases and phosphatases, Anal. Chem., 71,
3469–3476, 1999.
33. Li, H., H. Y. Wu, Y. Wang, C. E. Sims, and N. L. Allbritton, Improved capillary electrophoresis
conditions for the separation of kinase substrates by the laser micropipet system, J. Chromatogr. B
Biomed. Sci. Appl., 757, 79–88, 2001.
34. Bonewald, L. F., L. Bibbs, S. A. Kates, A. Khatri, J. S. McMurray, K. F. Medzihradszky, and S. T. Wein-
traub, Study on the synthesis and characterization of peptides containing phosphorylated tyrosine, J.
Pept. Res., 53, 161–169, 1999.
35. Tadey, T. and W. C. Purdy, Capillary electrophoretic resolution of phosphorylated peptide isomers
using micellar solutions and coated capillaries, Electrophoresis, 16, 574–579, 1995.
36. Sandra, K., J. Van Beeumen, I. Stals, P. Sandra, M. Claeyssens, and B. Devreese, Characterization
of cellobiohydrolase I N-glycans and differentiation of their phosphorylated isomers by capillary
electrophoresis-Q-Trap mass spectrometry, Anal. Chem., 76, 5878–5886, 2004.
37. Cao, P. and J. T. Stults, Mapping the phosphorylation sites of proteins using on-line immobilized metal
affinity chromatography/capillary electrophoresis/electrospray ionization multiple stage tandem mass
spectrometry, Rapid Commun. Mass Spectrom., 14, 1600–1606, 2000.
38. Cao, P. and J. T. Stults, Phosphopeptide analysis by on-line immobilized metal-ion affinity
chromatography-capillary electrophoresis-electrospray ionization mass spectrometry, J. Chromatogr.
A, 853, 225–235, 1999.
39. Ganzler, K., N. W. Warne, and W. S. Hancock, Analysis of rDNA-derived proteins and their posttrans-
lational modifications, in Capillary Electrophoresis in Analytical Biotechnology, P. G. Rhigetti (Ed.)
CRC press, Boca Raton, FL, pp. 185–238, 1996.
40. Berezovski, M., W. P. Li, C. D. Poulter, and S. N. Krylov, Measuring the activity of farnesyltrans-
ferase by capillary electrophoresis with laser-induced fluorescence detection, Electrophoresis, 23,
3398–3403, 2002.
24 Extreme Resolution in
Capillary Electrophoresis:
UHVCE, FCCE, and SCCE
Wm. Hampton Henley and James W. Jorgenson
CONTENTS
24.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 724

24.2 Ultrahigh Voltage Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 725
24.2.1 Background and Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 725
24.2.2 Practical Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 728
24.2.2.1 High Voltage Capillary Shielding System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 728
24.2.2.2 Power Supply and Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 728
24.2.2.3 HV Insulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 730
24.2.2.4 Modes of Operation and Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 730
24.2.2.5 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 731
24.2.2.6 Methods Development Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 732
24.2.2.7 Buffer Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 733
24.2.2.8 Capillary Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 735
24.2.2.9 Sample Separation Strategies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 735
24.3 Flow Counterbalanced Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 736
24.3.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 736
24.3.2 Theoretical Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 736
24.3.3.1 Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 738
24.3.3.2 Capillary Packing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 740
24.3.4 Methods Development Guidelines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 744
24.3.4.1 Handling Extended Separation Times. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 744
24.3.4.2 Band Broadening Prevention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 745
24.3.4.3 Determining the Best Separation Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 746
24.4 Synchronous Cyclic Capillary Electrophoresis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 746
24.4.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 746
24.4.1.1 Microchip SCCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 747
24.4.1.2 Capillary SCCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 747
24.4.2 Theoretical Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 748
24.4.2.1 Absolute versus “Effective” Voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 748
24.4.2.2 Band Broadening Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 748
24.4.3.1 Power Supply and Voltage Switching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 749
24.4.3.2 Connection of Fused-Silica Capillaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 749
723
24.4.3.4 Methods of Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 753

24.4.3.5 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 753
24.4.4 Methods Development Guidelines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 753
24.4.4.1 Handling Extended Separation Times. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 753
24.4.4.2 Band Broadening Prevention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 754
24.4.4.3 Sample Separation Strategies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 754
Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 755
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 755
24.1 INTRODUCTION
The majority of difficult separations can be divided into two main classes: complex mixtures and
challenging pairs (Figure 24.1). Complex mixtures, such as proteins and protein digests, demonstrate
“random” or broad spectrum behavior with respect to their properties, whereas “well-behaved”
analytes such as nucleic acids display predictable behavior determined mainly by the size of the
molecule. The analysis of these complex samples requires a separation system with high peak
capacity and the ability to cover a broad spectrum of physical properties.
Challenging pairs, such as enantiomers and isotopomers, display properties that are so similar
in nature that a separation technique must have extremely high efficiency in order to resolve the
analytes. The high efficiency can come at the expense of the ability to analyze a broad spectrum of
different analyte properties because the species of interest are confined to closely migrating bands.
Conventional approaches to increasing the resolution for difficult separations in capillary elec-
trophoresis (CE) usually require extensive development of buffers and/or capillary coatings particular
to an individual separation. The improvement in resolution is usually the result of exploiting a specific
intrinsic property of the analytes. A universal approach to achieving broad spectrum high resolution
separations can be achieved using CE with an extremely high applied potential. Separation efficiency
increases linearly with applied voltage, and resolution improves with the square root of applied volt-
age. Ultrahigh voltage CE (UHVCE) uses existing buffer systems and separation conditions but
gives the analytes more time to separate at a given electric field strength.
The separation of challenging pairs that are not resolved using traditional methods is often
achieved by developing novel pseudo-stationary phases for enantiomers or by utilizing mass spec-
trometry for analyzing isotopomers. Resolution of these tough pairs can also be achieved using
Difficult separations
Complex mixtures Challenging pairs
“Random” “Well behaved” Enantiomers Isotopomers
Peptides/digests Proteins Nucleic acids Molecules Small ions
FIGURE 24.1 Dividing difficult separations into several categories can help in the development of separation
strategies.
Extreme Resolution in Capillary Electrophoresis: UHVCE, FCCE, and SCCE 725
extended separation times under high electric field strength. Because the difficult pairs exist as
closely migrating analyte bands, techniques such as flow counterbalanced CE (FCCE) and syn-
chronous cyclic CE (SCCE) can be used. These methods use conventional separation voltages in
relatively short capillaries but use procedures to increase the duration in a high electric field.
24.2 ULTRAHIGH VOLTAGE CAPILLARY ELECTROPHORESIS

24.2.1 BACKGROUND AND THEORY
Early work with CE demonstrated how higher applied voltages offered vast improvements for elec-
trophoretic separations.1–5 Separation efficiency shows a linear dependence on applied voltage when
analytes are detected very close to the end of the capillary:
(µe + µEOF ) V
N∼
= , (24.1)
2D
where N is the efficiency in theoretical plates, µe is the electrophoretic mobility of the analyte, µEOF
is the mobility of the electroosmotic flow (EOF), V is the applied voltage, and D is the diffusion
coefficient of the analyte.6 Resolution between analytes with electrophoretic mobilities µ1 and µ2
show a dependence on the square root of the applied voltage:7
1/2
V
Rs = 0.177(µ1 − µ2 ) . (24.2)
D(µe + µEOF )
In addition, the speed of a separation can be greatly increased with higher applied voltages.
Migration time for an analyte (tm ) displays an inverse linear relationship with applied voltage for
the same length capillary (L) and separation distance (l):
lL
tm = . (24.3)
V (µe + µEOF )
The flat electrokinetic flow profile suggests that as long as Joule heating can be prevented, there is
no theoretical limit to the improvements to be gained by increasing the applied voltage. In practice,
several factors combined to limit the applied voltage used in early instrumentation to approximately
30 kV.
High electric field strengths can lead to resistive or Joule heating of the buffer solution within
the capillary. A detailed discussion of the effects of Joule heating with voltages up to 60 kV has been
described by Palonen et al.8 In the simplest of terms, the power (P) running through the capillary is
given by
V2
P = iV = , (24.4)
R
where R is the resistance of the capillary. Since it is desirable to have the greatest possible voltage,
the resistance of the capillary must be large enough to keep the power below about 1 W per meter
of capillary length.
Initial experiments conducted at voltages in excess of 30 kV resulted in broken capillaries. A
quick calculation of the radial electric field at the silica/buffer interface (Figure 24.2) using the
following equation explains why this occurs:9
Vi − V o
Eradial = , (24.5)
ln(b/rc )Krc
E
Longitudinal field Radial field
FIGURE 24.2 Longitudinal electric fields are used to separate analytes in CE. Radial fields also exist through
the capillary wall, and at high voltages, they can destroy the capillary via dielectric breakdown.
where Vi is the voltage inside the capillary, Vo is the voltage of a nearby conductor distance b
from the center of the capillary, rc is the length of the inner radius of the capillary, and K is the
dielectric constant of the capillary wall (∼3.9 for fused silica). For a capillary 50 µm in diameter,
1 m from the nearest grounded conductor, with 100 kV applied, the radial electric field strength is
about 100 million V/m. Radial electric fields greater than the dielectric strength of fused silica (∼15
million V/m) damage the fused silica wall and cause defects in the structure leading to fragmentation.
Early efforts using a van de Graff generator10 at ∼200 kV and a thick dielectric coating such as
a sleeve of Teflon® tubing around the capillary were unsuccessful. In another attempt, a plastic tube
surrounding the capillary was filled with isopropyl alcohol.11 In theory, both the capillary and the
tube filled with alcohol can be treated as long resistors, and therefore the potential will drop linearly
down the length of the capillary. The potential drop in the outer tube should match the potential drop
in the capillary and therefore reduce the radial field through the capillary wall. Unfortunately, the
capillaries still degraded in a very short time.
The first capillary shielding system that could prevent dielectric breakdown of the fused-silica
capillary was demonstrated by Hutterer and Jorgenson.12–15 The capillary was placed inside a series
of metal rings charged to potentials closely matching those inside the capillary (Figure 24.3). The
electric potential from the rings counteracts the radial field from the capillary and shields the capillary
from damage. The high voltage direct current (HVDC) was generated using a 26-stage Cockcroft–
Walton voltage multiplier16 with 2.4 nF HV capacitors. The voltages needed to charge the rings
were supplied by different stages of the multiplier. The entire instrument was submerged in a tank
containing transformer oil to provide electrical insulation.
This instrument showed the expected improvements in efficiency and resolution for samples of
peptides via CE,7,12 hyaluronic acid via capillary gel electrophoresis (CGE),12,14 and oligosaccharide
mixtures via micellular electrokinetic capillary chromatography.12,15 Application of up to 120 kV was
possible, and no degradation of the capillary was observed. Figure 24.4 shows electropherograms of
a cytochrome c digest separated on the same capillary using 28 kV and 120 kV.12 It is easy to see the
improvements in resolution and the decrease in separation time when higher voltages are applied.
The current state-of-the-art instrumentation uses larger capacitors (10 nF) to create a 100-fold
Cockcroft–Walton voltage multiplier capable of reaching potentials as high as 410 kV.6 A total of
25 shielding rings keep the potential difference between the capillary and shielding system at less
than 20 kV. The benchtop instrument runs in air, using plastic dielectric material instead of the tank
of transformer oil. The HV polarity can easily be reversed, allowing both positive and negative HV
separations. The temperature of the instrument can be controlled to within a few degrees over a range
of ambient temperature up to 60◦ C. It can be easily moved and coupled to different detection systems
including laser-induced fluorescence (LIF) detection, UV absorbance detection, and electrospray
ionization-mass spectrometry (ESI-MS). Separations of peptides, proteins, and nucleic acids using
the instrument with these detectors are described in Sections 24.2.2.4, 24.3.3.3, and 24.4.3.3.
80 kV
80 kV
60 kV
40 kV
20 kV
0 kV
FIGURE 24.3 Charged metal shielding rings can be used to reduce the radial electric field strength. Closely
matching the potential charge on the ring to that within the capillary prevents dielectric breakdown and allows
operation at very high voltages. (Reprinted from Henley, W. H., PhD Dissertation, University of North Carolina,
Chapel Hill, 2005. Copyright 2005. With permission.)
(a)
0.6
28 kV
0.4
mAbs
0.2
0.0
220 230 240 250 260 270 280 290 300 310 320 330
Time (min)
(b) 1.0
0.8
120 kV
0.6
mAbs
0.4
0.2
0.0
55 60 65 70 75 80 85 90 95
Time (min)
FIGURE 24.4 Separations of cytochrome c tryptic digest using the same capillary at conventional (top) and
UHV conditions. (Example: Reprinted from Hutterer, K. M., Doctorial Thesis. University of North Carolina,
24.2.2 PRACTICAL IMPLEMENTATION

24.2.2.1 High Voltage Capillary Shielding System
The first step in implementing UHVCE is to devise a capillary shielding system adequate for the
desired separation voltage. The radial field that causes dielectric breakdown of the capillary must be
dealt with in some manner. The dielectric breakdown strength of silica is approximately 15 MV/m.
Equation 24.4 can be used to show that small diameter capillaries have a large radial electric field
at relatively modest voltages. One approach is to simply use a large i.d. capillary with a thick wall.
This may work for relatively low voltages, but not without band broadening and heat dissipation
problems inherent to large bore capillaries. Another approach uses a conductive, but high resistance,
material to surround the capillary in a uniform coating. High voltage is applied to the capillary and
to the surrounding material. If the resistance of the material is equal along its length, the potential
drop should mirror that of the capillary and thus neutralize the radial electric field through the wall of
the capillary. Attempts using isopropyl alcohol in a tube surrounding the capillary resulted in failure
after a relatively short time.11 Nevertheless, such a system should be possible.
The most suitable method found so far is the use of charged metal rings to surround the capillary
and neutralize the radial field (Figure 24.3). Briefly, the capillary is surrounded by a series of metal
rings charged to a potential that roughly matches those inside the capillary. Long lengths of capillary
can be equally spaced within the rings by winding them onto a suitable spool. This method has been
successfully used without any noticeable degradation of the capillary.
24.2.2.2 Power Supply and Development

The development of the power supply for UHVCE requires several considerations. Voltages of the
magnitude used for UHVCE can easily arc through the air and “crawl” along insulating surfaces for
several feet. Safety is paramount, and a grounded metal shell or Faraday cage should be used to protect
the operator from electrocution and to protect sensitive nearby equipment from electromagnetic
pulses caused by stray electric arcs. The supply must generate high voltage with sufficient current
to drive the separation and charge the shielding system. Loss of charge from corona and stray
capacitance must be prevented if the full potential of the power supply is to be realized.
The generation of high voltage can be accomplished in many ways. Perhaps the simplest method
is the van de Graff generator.10 Briefly, charge is pumped from one roller to anther roller held inside
a metal sphere via a belt of dielectric material. The van de Graff generator is a constant current
source, and therefore the voltage is limited by the loss of charge to the surrounding air and resistive
load. Submerging the generator in transformer oil can dramatically reduce charge loss and improve
the efficiency of the generator. Since the voltage is generated in a single step, charging metal shield
rings to different potentials to shield the capillary may not be practical.
If a single commercial high voltage power supply is used, a string of high voltage resistors set
up to divide the voltage will be needed to bias the shielding rings. Such a system would be most
practical for voltages up to about 100 kV, and even then require careful insulation of the resistors
and high voltage wires.
The Cockcroft-Walton voltage multiplier has been used to generate the high voltage for the
two UHVCE systems developed thus far.6,7,12,14 The instrumentation has already been thoroughly
detailed elsewhere, and so will only briefly be discussed here. The multiplier is a convenient and
inexpensive solution to several problems surrounding UHVCE. It produces successively higher d.c.
voltages as the multiplier chain is extended, providing multiple voltage sources to bias the shielding
rings. The electronics are simple and can be easily contained within modestly sized shielding rings
so that dielectric stresses on the electronic components themselves can be minimized (Figure 24.5).
The efficiency of the multiplier decreases as the number of stages is increased. This can be mediated
somewhat by using large capacitors in the first stages and decreasing the capacitance of each succes-
sive stage. Using fast recovery diodes can prevent excessive losses from reverse current leakage. The
(a) Female
banana jack
connectors
(at bottom)
+ –
Capacitor C1 C2
Diode
Capillary
spool
Male banana
jack C3
C4
connectors
(sticking up and
out of top)
Aluminum
ring
(b)
(c)
FIGURE 24.5 (a) A schematic representation of the electronics in the individual modules that make up the
Cockcroft–Walton voltage multiplier and shielding rings. The capillary spool can be seen in the photo (b), as
well as the connections between modules (c). (Reprinted from Henley, W. H., PhD Dissertation, University of
North Carolina, Chapel Hill, 2005. Copyright 2005. With permission.)
a.c. input frequency should be carefully selected for maximum efficiency of the Cockcroft-Walton
generator.
24.2.2.3 HV Insulation
When the electric field at a metal surface becomes greater in magnitude than the dielectric strength
of the surrounding material, ionization of that material will occur. This phenomenon is called corona
when it occurs in air. It is undesirable because it is a source of potential voltage loss and can lead to
an electric arc. The strength of the electric field is dependent on the radius of curvature of the metal
shielding ring. For every 30 kV of potential, the radius of curvature must increase 1 cm to prevent
corona. Potentials of 500 kV require a radius of curvature of over 33 cm. Such large diameters can be
avoided if material of sufficient dielectric strength and thickness is used to surround a more modest
diameter. The first UHVCE instrument was surrounded by a tank of transformer oil. The oil proved
difficult to work in. It absorbed moisture from the atmosphere and became slightly conductive,
reducing the maximum attainable voltage.
The second UHVCE instrument6 used thick coatings of plastics for dielectric insulation
(Figure 24.5). All metal surfaces were painted with vinyl coatings. The electronic components
were potted in thermoplastic polyethylene commonly available as “hot melt” glue. This mate-
rial is inexpensive, has good dielectric properties, can be applied in thick coatings, and can
be easily reworked with a scalpel and heat. Insulation between the shielding rings was pro-
vided by polyethylene caps. The insulation from the surrounding air was provided by an acrylic
tube with many layers of polyester film. The polyester film had an exceptionally large dielectric
strength (∼3 × 106 V/cm). Every effort was taken to prevent an ionizable path to ground. The
high voltage end of the instrument housing was completely sealed with polyester/styrene casting
resin except for the tiny injection port. The injection port was plugged with a polyetheretherke-
tone (PEEK) injection rod (Figure 24.6) covered in silicone vacuum grease to exclude air from
the seal.
24.2.2.4 Modes of Operation and Examples

Open tube UHVCZE can be used for the separation of small molecules, peptides, proteins, and protein
digests.6,7,12 Increasing the time that the analytes migrate will improve resolution and efficiency. One
method of achieving this is to increase the length of the capillary, but when the maximum attainable
voltage is limited, the electric field strength will be relatively low in long capillaries. UHVCE
can improve these separations by maintaining a high field strength in a long capillary. Figure 24.7
shows two separations of a model peptide sample in a 566-cm-long capillary. The sample was first
separated at 28 kV (close to the maximum conventional CE applied voltage) resulting in an electric
field strength of only 50 V/cm. Leu-enk had an efficiency of 1 million theoretical plates and required
approximately 328 min to migrate to the detector. Using the UHVCE power supply at 330 kV, the
electric field strength was 580 V/cm. Leu-enk showed a 10-fold efficiency improvement at higher
voltage, resulting in an efficiency of 10 million plates. Migration time was reduced to 28 min.
Resolution between the analytes and impurity peaks can also been seen.
In open tube UHVCE, it is important to reduce analyte adsorption as much as possible. Analytes
that adsorb to the fused-silica surface do not tend to show much improvement as the length of capillary
and applied voltage are increased. Figure 24.8 shows a model protein separation that demonstrated a
small amount of adsorption in a 50 cm long separation. Using UHV in a 501 cm capillary, resolution
and efficiency are actually reduced, and migration time is proportionally longer.
Capillaries coated with aminopropyl species have been used to reduce the adsorption of peptides
and proteins. UHVCE separations of peptides and proteins using these coating have been reported
Terminal Capillary UHMWPE Alligator

(a) module injection clip
port
Buffer vial w/
electrode
Magnet
Sample PEEK injection

Electrode rod
connection
Acrylic tube Cast polyester and styrene External
wrapped in resin cap HVDC
polyester film
(b)
FIGURE 24.6 Schematic diagram showing the operation of the injection port for sampling at low voltage.
(Reprinted from Henley, W. H., PhD Dissertation, University of North Carolina, Chapel Hill, 2005. Copyright
with UV absorbance and ESI-MS.6 Large improvements were seen in resolution and efficiency using
UHVCE compared with shorter capillaries run at the same field strength at lower voltages.
If the capillary is filled with a suitable sieving matrix, polymers such as nucleic acids6 and
hyaluronic acid12,14 can be separated using UHV capillary gel electrophoresis (UHVCGE). Figure
24.9 shows four separations of MegaBASE™ 4 Color Sequencing Standard using different lengths
of 75µm i.d., 360µm o.d. acrylamide-coated capillary filled with MegaBACE™ Long Read Matrix
(Amersham Biosciences). The applied voltage ranges from −9 kV in the 60 cm capillary to –83 kV
in the 5.5 m capillary to maintain a constant field strength of 150 V/cm. The time has been rescaled
for the longer separations by a factor of X (given in the legend) so that the resolution improvements
at higher applied voltages can be easily seen.
Neutral analytes have been separated using UHV micellular electrokinetic chromatography
(UHVMEKC).12,15 This was demonstrated by the separation of oligosaccharide mixtures in a
surfactant-containing buffer. Higher resolution was achieved using UHV than standard running
voltages, and plate counts over 1 million plates were seen.
24.2.2.5 Limitations
Despite the advantages offered by UHVCE, several problems still limit its application. The lack of
available instrumentation and unfamiliarity with voltages in excess of 30 kV keep many potential
applications from being explored. The dangers involved with working at such high voltages are more
perceived than real, and safety can be assured by using a grounded Faraday cage with interlocks
designed to protect the operator from high voltage discharges. Also, the stored energy in most high
voltage systems is usually very small, much less than 50 J for the current UHVCE system.
(a)
GlyAsn Trp
1500
28 kV
1000 AlaGln
Leu-enk
LIF signal
~1 million plates
500 GlyLeuTyr
325 330 335 340

Min
3500 Trp
GlyLeuTyr GlyAsn
(b)
3000
Leu-enk
~10 million plates 330 kV
2500
LIF signal
AlaGln
2000
1500
1000
500
27.0 27.5 28.0 28.5 29.0 29.5 30.0

Min
FIGURE 24.7 Separations of a model peptide sample run on the same capillary at conventional and UHV. The
10-fold increase in electric field strength improves efficiency and reduces separation time 10-fold. Resolution
increase can be clearly seen by the closely migrating impurity peaks. (Reprinted from Henley, W. H., PhD
Dissertation, University of North Carolina, Chapel Hill, 2005. Copyright 2005. With permission.)
24.2.2.6 Methods Development Guidelines

Part of the advantage of UHVCE is that it allows the use of known separation systems that work well
for similar analytes but are unable to completely resolve the analytes of interest using conventional
voltages and capillary lengths. Several factors relating to the buffer system and capillary should be
taken into consideration before attempting to use a separation system for UHVCE.
2500 Ribonuclease A
Lysozyme 470 kplates
415 kplates
UVabs@200 nm x0.002 AU
2000 Trpsinogen
270 kplates
1500 L = 501 cm
253 kV
beta-lactoglobulin A and B
490, 240, 360 kplates
1000
500
4.4 4.6 4.8 5.0 5.2 5.4 5.6

Min
1000 Lysozyme
610 kplates Ribonuclease A
320 kplates
800
UVabs@200 nm x0.002 AU
Trpsinogen/
beta-lactoglobulin A
600 and B (unresolved)
220 kplates
L = 501 cm
400 253 kV
200
–200
60 65 70 75
Min
FIGURE 24.8 Separation improvements are not realized in longer capillaries at higher voltages if analyte
adsorption occurs. Shown here is a separation of proteins in short (a) and long (b) capillaries under the same
electric field strength. Efficiency and resolution are actually reduced at UHV conditions due to analyte adsorption
over the long capillary wall. (Reprinted from Henley, W. H., PhD Dissertation, University of North Carolina,
24.2.2.7 Buffer Selection

When selecting a buffer for UHVCE, the conductivity of the buffer must be as low as possible so
that the electric field strength can be high without unacceptable Joule heating. The maximum field
strength for a buffer/capillary system can easily be found by generating a current to field strength
Note resolution
improvements
5500
L = 60/ I = 50 cm (9kV) X = 100%

5000
LIF signal
L = 120/ I = 110 cm (18kV) X = 52%
4500
L = 184/ I = 174 cm (27.6kV) X = 31%
4000
L = 554/ I = 544 cm (83kV) X = 48%
40.0 40.5 41.0 41.5 42.0 42.5

Rescaled time (minutes*x)
FIGURE 24.9 Separations of DNAsequencing standard rescaled in the time domain by factor X for comparison
of resolution improvements at higher applied voltages. Longer lengths (L) of capillary were used to maintain
a constant 150 V/cm field strength. (Reprinted from Henley, W. H., PhD Dissertation, University of North
Carolina, Chapel Hill, 2005. Copyright 2005. With permission.)
plot using a conventional power supply in a short piece of capillary. The onset of appreciable Joule
heating can be found at the field strength where the plot begins to show nonlinear behavior.
High concentrations of salt are used in some buffers to prevent adsorption of analyte molecules
(peptides and proteins17 in particular) to the capillary wall. Zwitterionic molecules contain an equal
number of positive and negative charges at a pH equal to their isoelectric point. They can be used
to raise the ionic strength of the buffer without increasing the conductivity greatly. In UHVCE
applications, they can sometimes be substituted for more conductive buffers to allow the use of
higher field strengths while maintaining sufficient ionic strength to reduce analyte adsorption. It is
important to note that the pI of the zwitterionic molecule may be several pH units different from their
pKa values. In cases where this occurs, they cannot be relied upon to buffer the solution effectively.
High concentrations of ions with small radii of hydration such as H3 O+ , K + , and OH– should
be avoided. These small ions are rapidly transported within the linear electric field of the capillary,
resulting in high electric currents. For this reason, ions with lower mobility such as Li+ and CH3 COO–
should be substituted wherever possible.18
The buffer capacity can become important in separations of extended duration. For example, a
CGE separation requiring 30 min on a 30 cm capillary will require 300 minutes on a 300 cm capillary
if the same electric field strength is used. The 10-fold increase in analysis time results in 10-fold
the amount of electrolysis products that must be dealt with by the buffer. The moles of OH– at the
cathode can be calculated from the electrophoretic current (i):
1
it = molOH− , (24.6)
F
where t is time in seconds and F is Faraday’s constant. A significant pH shift can cause problems
with reproducibility. Large capacity buffer vials can be used if the instrumentation has enough
available space.
Buffer systems that do not adequately address problems of analyte adsorption will not be compat-
ible with UHVCE in long capillaries.6 Buffers that contain ions with significantly different mobilities
from the analyte ions will also result in electrodispersion that will be exacerbated by long separation
distances. Both of these problems can be avoided by examining the shape of the analyte bands in
short capillaries at lower voltages. Peaks that show fronting or tailing indicate potential problems,
whereas highly Gaussian peaks indicate compatibility with UHVCE.
24.2.2.8 Capillary Selection

The length of the capillary should be determined by the desired degree of improvement over a known,
lower voltage, separation. For example, if twice the resolution of a known system using 30 kV in a
50 cm capillary is desired, then 120 kV should be used in a capillary 2 m long. This separation will
also require fourfold the time.
Joule heating can be avoided by simply reducing the power consumption of the capillary. Reduc-
ing the capillary diameter by half decreases the current fourfold. For the separation described above,
a 50 cm length of capillary with half the diameter could be used at 120 kV without excessive Joule
heating. In this case, the separation would require fourfold less time.
The inner diameter of the capillary cannot be decreased too greatly if UV absorbance is used for
detection. Below ∼50 µm, the pathlength is too short for sensitive detection. ESI-MS also requires
a certain flow rate for optimal detection that may not be achievable in very narrow capillaries. LIF
detection is well suited for use with narrow capillaries, but not every analyte can be labeled with
fluorescent tags.
24.2.2.9 Sample Separation Strategies

Unlike techniques that are limited to a narrow range of analyte properties, UHVCE is well suited
for the analysis of complex samples containing analytes with a broad spectrum of properties (Figure
24.1). These samples include, among others, cell lysates, protein digests, or DNA sequencing prepa-
rations. The UHVCE analysis of samples should ideally start with a fairly concentrated sample
containing low amounts of salt. High sample ionic strength can result in electrodispersion of the
sample when used with low conductivity buffers. Dilute samples can be problematic if poor detec-
tion limits require excessively large sample injections.19 This may or may not be solved using sample
stacking techniques.
The best solution found thus far for the preventing peptide and protein samples from adsorbing
to the capillary wall in UHVCE is electrostatic repulsion. This can be accomplished in bare silica
capillaries by using very basic pH buffers where the deprotonated silanol groups and the majority of
peptides and proteins are negatively charged. This method is not directly compatible with positive
mode ESI-MS, but may work for UV absorbance detection or LIF.17
Various coatings that impart a positive charge to the surface can be used for low pH separations.
Covalently bound coatings such as aminopropyltrimethoxysilane have been used with good success
for UV detection and ESI-MS of peptides.20,21 Other coating schemes based on the same principles
have been used to varying degrees of success, and diisopropylaminopropylethyoxysilane has shown
remarkable stability to degradation from hydrolysis.6,22 Cationic polymers such as PolyE-32323,24
have been used to great effect for CE/ESI-MS for proteins and peptides. These coating are easily
applied to long capillaries and stable in buffers with organic modifiers.
The analysis or sequencing of DNA in CE and its advantages over standard slab gels has been
well documented.25–40 Preliminary study of UHVGCE for DNA separations indicates that it may be
best suited for high-resolution separations of relatively short sequences of DNA.6
Polysaccharides and other complex carbohydrates can be difficult to analyze for several reasons.
The lack of charged moieties requires the use of a sieving gel or pseudo-stationary phase. Sieving
gels can be quite useful for highly branched structures found in some of these analytes. UHVCGE
or UHVMEKC has been shown to be useful for high-resolution analysis of these samples.12,14,15
24.3 FLOW COUNTERBALANCED CAPILLARY ELECTROPHORESIS

24.3.1 BACKGROUND
The effectiveness of capillary electrophoretic separations can be greatly affected by the magnitude
and direction of the bulk fluid flow. As pointed out by Jorgenson and Lukacs,4 analyte bands that
migrate in the direction opposite to the EOF spend a greater amount of time in the electric field
and thus are more affected than analytes that migrate with the EOF. Direct manipulation of the
EOF or application of hydrodynamic (pressure driven) flow can be used to retard the migration of
analytes to improve efficiency and resolution. Resolution improvement by the direct control of the
EOF through the manipulation of radial electric fields has been shown at low pH and low ionic
strength.41 Cheng et al.42 noted the potential of a counter flow to improve the resolution of analytes
in a paper investigating the effect of a constant applied pressure from a sheath flow cuvette on sep-
arations of amino acids during CE separations. Pressure driven flows have also been used for the
preconcentration of samples for capillary isotachophoresis with long preconcentration times in short
capillaries.43–49 Reduction of the EOF using surfactants and polymers to increase the viscosity of
background electrolyte near the wall has also been demonstrated for resolution improvement.50–55
Lucy and McDonald51 used this technique to resolve the major isotopes of chlorine using CE. Several
preparative scale purification techniques using a counter flow have been reported for enantiomers
and other closely related species.56–63 Another study used a counter flow to prevent buffer compo-
nents that improve selectivity but increase the UV absorbance background signal from reaching the
detector.64
A practical system for analytical scale high-resolution separations was first demonstrated by
Culbertson and Jorgenson.65 Narrow diameter capillaries and LIF detection were used to resolve
analytes with extremely small differences in mobility such as peptide isomers. Misleveling of the
buffer reservoirs, using a siphoning action to provide a counter flow in large diameter capillaries,
was also reported with modest resolution improvements.66
Band broadening caused by the counter flow limited high-resolution FCCE to narrow bore capil-
laries and LIF detection for over a decade. Recently, the use of large diameter capillaries packed with
bare silica particles to flatten the parabolic flow profile of the counter flow has been demonstrated.67
Packed capillary FCCE should not be confused with capillary electrokinetic chromatography (CEC).
Packing material used in FCCE serves no chromatographic purpose and does not act as a station-
ary phase. The purpose of the packing material is simply to prevent band broadening in large bore
capillaries so that UV absorbance detection can be used to detect unlabeled analytes.
24.3.2 THEORETICAL ASPECTS

The power of FCCE lies in the ability to maintain analyte bands within a high electric field strength
for a long period of time. A counter flow is applied that pushes the bulk solution in the direction
opposite to the analyte’s migration, extending the duration of the separation. For a given length (L)
of capillary with applied voltage (V ), the electric field strength (E) is simply
V
E= . (24.7)
L
The time (tm ) required for an analyte band to migrate a distance equal to the length of the capillary
(L) can be determined from the following equation:
L
tm = , (24.8)
E(µe + µEOF )
where µe is the analyte’s electrophoretic mobility and µEOF is the mobility of the EOF. Suppose that
a counter flow is applied opposite to the direction of the analyte’s net migration so that the analyte
band now requires fourfold the time to transverse the same length of capillary. The electric field
strength, length of capillary, and mobilities of the analyte band and EOF remain unchanged, but
now the effective distance (Leff ) traveled by the analyte band is equal to 4L. Neglecting any band
broadening, an equivalent separation without a counter flow would require a capillary four times
as long. In order to obtain the same electric field strength, four times the voltage would have to be
applied. This can be thought of as the effective voltage (Veff ) that can be determined from the known
electric field strength:
V Veff
E= = . (24.9)
L Leff
Since the resolution between two analyte bands scales with the square root of the applied voltage,
fourfold the effective voltage would generate twofold the resolution. For example, consider two
analyte bands that transverse a 30 cm capillary in an average time of 30 s with 30 kV applied. If
the resolution between the bands is 0.5, then a capillary 120 cm long with 120 kV applied would be
needed to obtain a resolution of 1.0. The same resolution can be obtained using the 30 cm capillary
with 30 kV applied if a counter flow is applied so that the analytes transverse the same distance
in 120 s.
EOF is generated at the wall of the capillary, resulting in a flat flow profile that limits band
broadening
2 to that caused by longitudinal diffusion. This broadening can be expressed as a variance
σB using Einstein’s equation:
σB2 = 2Dt, (24.10)
where D is the diffusion coefficient of the analyte and t is time. The application pressure results in a
hydrodynamic flow. Flow velocity is greatest in the center of the capillary due to drag at the capillary
wall and a parabolic flow profile is the result. Analyte bands diffuse laterally across this flow profile
and broaden to a degree proportional to its magnitude (Taylor dispersion). Culbertson and Jorgenson
65 derived an equation for the degree of band broadening for FCCE in an open tube. In addition, the
maximum allowable diameter of the capillary was determined from the following expression:
dc2 νpa
2 t
σc2 = ≤ 2Dt = σB2 , (24.11)
96D
where the broadening of the analyte bands from hydrodynamic flow is expressed as a variance,
σC2 , which should be less than or equal to the band broadening caused by longitudinal diffusion
alone. The diameter of the open capillary (dC ) can be calculated from the average velocity of the
hydrodynamic flow (vpa ) required to counterbalance the electrokinetic migration of the analyte band
and the diffusion coefficient of the analytes (D). For the analytes investigated, open tubular capillaries
5–10 µm were deemed acceptable.65,68 Narrow bore capillaries have a limited optical pathlength
and LIF detection is usually required for sensitive detection.
Band broadening from hydrodynamic flow in packed capillaries is independent of capillary
diameter, allowing the use of capillaries with a pathlength large enough for UV absorption detection.
A similar equation for band broadening in packed capillary FCCE can be derived, as can a similar
method of calculating the required particle diameter (dp ) to minimize band broadening.6,67 Briefly,
the total broadening is expressed as a variance (σtotal
2 ), which is the sum of the contributions of
longitudinal diffusion and hydrodynamic flow:
dp2 νpa
2 t
b
σtotal
2
= σB2 + σC2 = 2Dt + , (24.12)
10D
where tb is the time during which the counter flow is applied. Using the same strategy as expression
11, the maximum particle diameter can be calculated. The efficiency for FCCE in packed capillaries
can be found using the following equation:
10D(tνek )2
N= , (24.13)
20D2 t + dp2 νpa
2 t
b
where vek is the average electrokinetic velocity for the analyte bands and N is the efficiency in
theoretical plates.6,67 High efficiencies can be most easily obtained by using small diameter particles
and long separation times.

24.3.3.1 Instrumentation
The instrumentation required to perform FCCE is relatively simple. The power supply can be any
commercially available HVDC power source, but an ideal power supply would have dual polarity,
0–30 kV at several 100 µA, with safety interlocks and computer control capabilities. The maximum
voltage needed is determined by the desired electric field strength and length of the capillary. In order
to allow for some margin of error in control of the counter flow, the minimum length of capillary
should be at least three to four times the physical width of the desired separation window. These
margins allow for the analyte bands to be maintained within the inner half of the capillary so that
they are not pushed out on either side of the capillary, with detection occurring in the center of the
capillary.
The source of the counter flow, or pumping system, is determined by the velocity of the analyte
band and the backpressure required to reverse or stop its forward migration. Where the analyte
bands migrate slowly in an open tube, the pumping system may be as simple as a misleveling of the
buffer vials. Rapidly migrating analyte bands in narrow open capillaries may require a compressed
gas driven system such as that shown in Figure 24.10.65,68 Packed capillaries require much greater
pressures and so an high-performance liquid chromatography (HPLC) pump or syringe pump capable
of generating the desired fluid velocity must be used (Figure 24.11).6,67 If an HPLC pump is used,
some consideration should be given regarding the volumetric flow rate needed to maintain a stable
pressure. Typically, flow rates for HPLC columns range in the mL or µL per minute range. The flow
rate for FCCE will be much smaller, typically nL/min. The use of a “splitter” capillary to divert the
majority of the flow volume back to the pump’s buffer reservoir can save many liters of running
buffer. This may be especially important for buffers containing expensive reagents. Syringe pumps
have limited capacity and the flow rate will therefore determine the maximum separation time. One
notable exception to this is the Haskel International Inc. (Burbank, CA) DSHF-300 Air Driven Fluid
Pump. This is a syringe-style hydraulic pump capable of producing up to 50,000 psi of liquid pressure
via a 300-fold amplification of applied gas pressure. The advantages of this pump include its stable
applied pressure over a volumetric flow rate ranging from no flow at all to a few mL per minute.
It has a low stoke volume of a few mL, but it automatically recycles and refills itself at the end of a
piston stoke, maintaining a fairly constant pressure with built in check valves.
High voltage power supply
LIF detection
system
Pressure
reservoir
Cutaway view
Interlock
box
He gas
pressure
Micro- regulator Electronically actuated
computer pneumatic valve
FIGURE 24.10 Open tube FCCE using a gas pressure driven pumping system. (Reprinted from Culbertson,
C. T. and Jorgenson, J. W., Analytical Chemistry 1994, 66, 955–962. Copyright 1994 American Chemical
Society. With permission.)
High voltage power supply Syringe

clean out
port
Buffer out
Manual (p = 300 X Gas pressure)
Gas in
valve (low pressure)
UV detector PEEK
tubing
Waste
Capillary Capillary
316SS
T-fitting
Pneumatically
actuated valve
PEEK Haskel DSHF-300
tubing coil air driven
fluid pump
100 k
current monitoring
resistor Waste
Ground
Buffer reservoir Buffer reservoir
FIGURE 24.11 Packed capillary FCCE using high pressure from a hydraulic pump controlled by a pneu-
matically actuated valve. (Reprinted from Henley, W. H., et al., Analytical Chemistry 2005, 77, 7024–7031.
Copyright 2005 American Chemical Society. With permission.)
Ideally, the velocity of the counter flow should be exactly equal and opposite to the velocity
of the analyte band so that band broadening from pressure driven flow is minimized. The practical
implementation of such a system requires a detector capable of imaging the entire width of the
separation window such as a linear diode array or charge-coupled device (CCD) camera.
Single point detectors using UV absorbance or LIF detection require that the analyte bands pass
though the detection point in order for data to be collected. Therefore, analytes are first allowed
to migrate past the detection point electrokinetically. Then a pressure great enough to push them
backwards past the detector is applied via a computer-controlled valve. Once the analytes have
been sufficiently pushed back, the pressure is released and the cycle is repeated. The electrophoretic
voltage is maintained during the entire course of the separation, so the analytes are continuously
separating electrokinetically, even while being pushed backwards by the counter flow. An alternative
scheme using constant pressure with intermittently applied voltage would only resolve the analytes
while the voltage was applied.
Valves for controlling the back pressure must actuate quickly, be rated to the desired pressure,
and ideally be computer controlled. As long as the pressure remains constant, dead volume is usually
not a major concern. If the valve is constructed from metal, it must be electronically isolated from the
electrophoretic system or else electrolysis processes may corrode and destroy the valve, in addition
to ruining the separation. Several suitable valves are manufactured by Valco Instrument Co. Inc.
(VICI) (Houston, TX).
When making the connection between the capillary and pumping system, it is very important
to consider the flow of electrophoretic current and the effect it may have on components of the
system. Figure 24.12 shows two ways of connecting a packed separation capillary to the high-
pressure pumping system. A stainless steel Swagelok® compression fitting forms a union between
the fused-silica capillary and a piece of PEEK tubing. The configuration in Figure 24.12a results
in electrolysis products entering the capillary and spoiling the separation. Even though the metal
fitting is electrically floating, it is not isolated from the electrophoretic current. Current flows as ions
through the capillary and through the PEEK tubing, but the highly conductive metal fitting offers a
low resistance path for electrons to travel. The result is that electrolytic processes occur at the surface
of the metal near the exit of the capillary and the entrance to the PEEK tubing. An easy way to avoid
this is to simply push the capillary into the PEEK tubing so that the metal fitting is truly electrically
isolated from the electrophoretic current (Figure 24.12b).
24.3.3.2 Capillary Packing

Packing the capillary with nonporous bare silica can greatly reduce the amount of band broadening
observed at similar counter flow linear velocities. Figure 24.13 shows the temporal variance of an
Electrolysis PEEK tubing

occursat connected to
both grounded T- PEEK tubing
metal/buffer fitting
Capillary connected to
junctions Capillary
grounded T-
fitting
Plastic Steel Plastic Steel

ferrule ferrule ferrule ferrule
Stainless steel Stainless steel

Swagelok ® Swagelok ®
1/16”union 1/16”union
FIGURE 24.12 In order to prevent electrolysis products from entering the capillary at the connection of the
capillary to high-pressure counter flow system, metal surfaces must be floating and isolated from the path of
the electrophoretic current. (Reprinted from Henley, W. H., PhD Dissertation, University of North Carolina,
100
Temporal variance (sec2)

50 µm OTC
80
50 µm PBC
60
40
20
0
0.0 0.1 0.2 0.3 0.4
Counter flow linear velocity (cm/sec)
FIGURE 24.13 A comparison between the band broadening seen in a pack capillary and open tubular capillary
under similar counter flow linear velocities shows much less broadening in the packed capillary. (Reprinted
from Henley, W. H., et al., Analytical Chemistry 2005, 77, 7024–7031. Copyright 2005 American Chemical
Society. With permission.)
R-(−)-epinephrine peak using a 50 µm open tubular capillary (OTC) and a 75 µm capillary packed
with 4.1 µm nonporous bare silica particles. The diameters of the capillaries were chosen so that
they would have approximately the same dead volume and electrophoretic current. It is easy to see
from the plot of temporal variance versus counter flow linear velocity that much less broadening is
seen in the packed capillary as compared with the large diameter open tube.
The preparation of the packed capillary for FCCE using UV absorbance detection is not diffi-
cult, but it differs in several aspects from the methods used to prepare capillary HPLC or UHPLC
columns.69 For chromatography, capillaries are typically run with mobile phases flowing in the
same direction as the capillary was packed. For FCCE, the alternative application and release of
pressure flowing opposite to the flow of the EOF causes large voids to develop within the packed
bed. Consolidation of the capillary can be achieved via sonication under hydrodynamic flow.70,71
Briefly, dry silica particles are tapped into one end of an OTC. Heat from an electric arc is used to
sinter them together and to the capillary wall. Pressure (up to 50,000 psi) from packing bomb is used
to force a ∼10 mg/mL slurry of packing material into the open end of the capillary. After the bed
is packed, an inlet frit is sintered at the head of the bed using a heated tungsten wire. The pressure
is released, and the capillary is cut to length. The capillary, except for the fritted ends, is placed
into a water bath and sonicated for ∼5 min. Pressure is then applied to the capillary in the direction
opposite to that in which it was originally packed. A ∼5–10% reduction in the length of the packed
bed should be observed. A new frit is then sintered using a heated tungsten wire while the capillary
is still under pressure. Sonication can be repeated with pressure applied from the opposite end if
further consolidation is desired. Figure 24.14 shows how the consolidation of the packing material
reduces band broadening caused by the counter flow. The same capillary was used for measurements
of an R-(–)-epinephrine peak using 3.1 µm nonporous bare silica packing.6,67

Open tube FCCE can be performed in capillaries of relatively large diameter if the analyte bands
migrate slowly and only a modest improvement in resolution is desired. Misleveling of buffer vials
can provide a simple way of improving the resolution for analyte bands that are almost, but not quite
resolved under standard conditions.66 Band broadening will occur but usually not to a great extent
over short durations.
350
300 Before sonication

After sonication one side of window
Temporal variance (sec2)

After sonication both side of window
250
200
150
100
50
0
0 5 10 15 20 25 30 35 40 45 50
FIGURE 24.14 Reducing the interparticle porosity of the capillary packing can decrease the amount of band
broadening caused by the counter flow. Data taken from migration of R-(−)-epinephrine using 3.1 µm nonporous
bare silica packing. (Reprinted from Henley, W. H., et al., Analytical Chemistry 2005, 77, 7024–7031. Copyright
2005 American Chemical Society. With permission.)
Open tube FCCE in narrow capillaries using zone electrophoresis is useful for resolving analytes
that can be tagged with fluorescent labels for LIF detection. Culbertson72 demonstrated resolution
of the peptide fragment YAGAVVNDL and YAGAVVNDI (Figure 24.15), which only have an
electrokinetic mobility difference of 1.0033.
Open tube FCCE using micellular electrokinetic chromatography (MEKC) can be used to obtain
high resolution for charged or neutral analytes. High-resolution separations of TRITC-labeled pheny-
lalanine with different numbers of hydrogen atoms replaced with deuterium atom showed resolution
between analytes with differences of a single neutron (0.16% mass difference). The relative levels
of deuterium substitution could also be determined from the peak areas. The ability of FCCE with
MEKC to separate these isotopomers is believed to be based on slight differences in hydrophobicity
caused by the influence of the heavier deuterium atom.68
Packed capillary FCCE can be used with larger diameter capillaries with sufficient pathlength
for sensitive UV/Vis absorption detection. UV detection allows analysis of nonfluorescent analytes
and sample impurities. Most organic and some inorganic compounds absorb UV or visible light to
some degree, and therefore packed capillary FCCE can be used for a much larger range of analytes
than open tubular FCCE.
Separations of racemic samples of several basic and acidic pharmaceuticals using packed capil-
lary FCCE have been reported previously.6,67 Figure 24.16 shows the separation of the enantiomers
of fenoprofen using β-cyclodextrin as the chiral pseudo-stationary phase. The top portion of the figure
(a) shows the entire separation. The analyte peak is allowed to electrokinetically migrate pass the
detector (forward pass) and then pushed backwards under hydrodynamic flow (reverse pass). Seven
cycles with forward passes (F1–F7) and six reverse passes (R1–R6) can be seen in Figure 24.16a.
Figure 24.16b shows the first electrokinetic pass in front of the detector. This part of the electro-
pherogram represents the resolution (0.34) and efficiency (21,000 plates) that would result from
traditional CE under these separation conditions. Figure 24.16c shows the final pass in front of the
detector, where the analyte bands have broadened by a small degree, but are approaching baseline
resolution with greatly increased separation efficiency (320,000 plates). Table 24.1 gives a list of
acidic and basic pharmaceutical compounds separated with packed capillary FCCE. Electrokinetic
only migration times, total number of passes, along with resolution and efficiency improvements are
given.
(a) (b)

(c) (d)

Velocity ratio = 1.0033
FIGURE 24.15 Open tube FCCE with LIF detection showing the resolution of YAGAVVNDL and
YAGAVVNDI. (Reprinted from Culbertson, C.T., p. 264, 1996, University of North Carolina, Chapel Hill,
NC. Copyright 1996. With permission.)
(a)
UVabs signal (AUFs x 106)
Time (min)
(b) (c)
FIGURE 24.16 Packed capillary FCCE electropherogram showing the separation of fenoprofen enantiomers
using β-cyclodextrin as a chiral selector. (Reprinted from Henley, W. H., et al., Analytical Chemistry 2005, 77,
7024–7031. Copyright 2005 American Chemical Society. With permission.)
TABLE 24.1
Racemic Mixtures Resolved Using FCCE in Packed Capillaries Using β-Cyclodextrin
Migration Time Degree of Resolution Efficiency of
Total Run Total No. of
Compound (min) for 1st at 1st Pass/at 1st Pass/Last Pass
Time (min) Forward Passes
Forward Pass Last Pass (103 Plates)
Epinephrinea 40.9 329 8 0.42/1.6 31/213
Norepinephrinea 66.7 275 7 0.36/1.02 14/77
Synephrinea 117 138 2 0.988/1.02 15/25
Norphenylephrine HCla 62.1 147 4 1.13/1.76 35/68
Phenylpropanolaminea 71 195 5 0.21/0.77 9/41
Octopaminea 61.07 403 10 ND/0.55 8/45
Chlorpheniramine maleatea 32.8 118 3 2.12/3.35 90/240
Doxylaminea 23.5 202 6 1.59/3.38 120/820
Ibuprofenb 34.1 259 9 0.496/1.08 44/550
Ketoprofenb 39.8 688 22 ND/0.88 10/340
Fenoprofenb 35.7 242 7 0.34/0.88 21/380
ND = none detected.
a Basic chiral analytes were separated in 50 mM phosphoric acid, 10 mM β-cyclodextrin 0.5% triethylamine.
bAcidic compounds were analyzed using a 100 mM MES, 10 mM β-cyclodextrin buffer system.
Source: Reprinted from Henley, W. H., et al., Analytical Chemistry 2005, 77, 7024–7031. Copyright 2005 American
Chemical Society. With permission.
Several isotopomer separations have been reported for packed capillary FCCE. Deuterium-
substituted phenylalanine showing no resolution during the first forward pass was almost baseline
resolved after an 11 h separation time.67 Figure 24.17 shows a packed capillary FCCE separation of
bromine-79 and bromine-81 using UV absorbance detection. After 6 h of run time (Figure 24.17b),
a distinct notch is seen in the bromide ion peak. The separation was maintained for over 60 forward
passes requiring over 1100 min, and clear separation of the isotopes can easily be seen.6,67
Before attempting MEKC in packed capillaries, an examination of the properties of the micelles
under high pressures should be considered. Unpublished work suggests that at the high pressures
required for reversing the analyte bands, the properties of the micelles may change, resulting in an
unstable system.
24.3.4 METHODS DEVELOPMENT GUIDELINES

24.3.4.1 Handling Extended Separation Times
The high resolution and efficiency obtained in FCCE separations comes at the expense of extended
separation time. Although the buffer within the capillary is replenished by the counter flow,60 care
should be paid to ensure sufficient buffer capacity for reservoirs with fixed volume. Measurements of
the electrophoretic current can be used to calculate the expected pH change in each buffer reservoir
for a given separation time. In addition, steps should be taken to prevent electrolysis products from
entering the capillary. Figure 24.11 shows a long coil of PEEK tubing connected between the capillary
and the grounded electrode. This coil has sufficient length and volume to prevent gas bubbles or UV
absorbing electrolysis products from entering the separation capillary for many hours under typical
counter flow rates.
Automation of the valve control using a computer program is not only useful for ensuring that
the analyte band stays within the capillary but also lets the operator leave the instrument running
(a) (b)
200
600
150
400 100
50
200
UV abs signal (AUFs x 106)
0
–50
6 7 8 9 10 360 362 364
(c) (d)
100
200
80
150 60
40
100
20
50 0
–20
0
–40
738 740 742 744 1112 1113 1114 1115 1116

Time (min)
FIGURE 24.17 A separation of bromine isotopes using packed capillary FCCE. (Reprinted from Henley,
W. H., et al., Analytical Chemistry 2005, 77, 7024–7031. Copyright 2005 American Chemical Society. With
permission.)
autonomously for hours at a time. There are many different ways to automate the control of the
system, but a few basic principles should be followed for reliable operation. After injection of the
sample, the electrokinetic migration velocity of the analyte band can be determined by the time it
takes to travel the distance to the detector. After allowing the band to travel some distance beyond the
detector, the counter flow can be applied (while maintaining the applied voltage). The time required
for the analyte band to be pushed back in front of the detector can be used to calculate the analyte
band velocity under hydrodynamic flow. These two parameters can be used to calculate the position
of the analytes at any given point in the FCCE separation cycle. The progress of the separation
should be checked every few hours to correct any timing errors. Adjustments will have to be made
for long separations due to variations in pump pressure, EOF velocity, and buffer viscosity caused
by temperature changes or buffer depletion.
24.3.4.2 Band Broadening Prevention

One of the main advantages of FCCE is that published buffer conditions that may not completely
resolve certain analytes in a traditional CE system can be used to fully resolve the analytes by giving
them more time to separate. This can save method development time when analyzing a variety of
compounds with similar characteristics. The main limitation on the timescale of an FCCE experiment
is the degree to which band broadening reduces the analyte signal strength, and therefore the counter
flow velocity should be kept as low as possible.
Adsorption of the analyte to the capillary wall or bare silica packing material must also be
prevented to avoid signal loss. Electrostatic repulsion works well for negatively charged species
in basic buffer with bare capillaries. Amine or other basic analytes can be analyzed at low pH in
bare capillaries if a competitive inhibitor such as triethylamine is added to the running buffer. The
use of dynamic coatings such as polymers or surfactants in the running buffer to reduce adsorption
of analytes is probably best suited to open tubular FCCE as increased viscosity may result in high
back pressures in packed capillaries. Permanent, covalently bound coating should be used only after
their long-term stability has been confirmed, and analyte band migration time should be closely
monitored.
Electrodispersion can be a major source of band broadening in FCCE due to the long separation
times. Close matching of the electrophoretic mobilities of the analyte bands and running buffer can
reduce electrodispersion greatly. Gaussian-shaped peaks usually indicate a compatible sample/buffer
whereas fronted or tailed peaks indicate potential problems. Furthermore, the ionic strength of the
running buffer should be closely matched to that of the sample.
In addition, in packed capillary FCCE, care must be taken to ensure that the packing material is
fully consolidated. Voids in the packed bed will dramatically increase the observed broadening as
seen in Figure 24.14.
24.3.4.3 Determining the Best Separation Strategy

The first thing to consider when designing an FCCE separation strategy for a particular sam-
ple is the desired method of detection. Both open tubular FCCE and packed capillary FCCE
have their advantages and drawbacks. Open tube FCCE uses a slightly less complicated instru-
mental setup (Figure 24.10) than packed capillary FCCE (Figure 24.11). While not difficult, the
equipment and skills needed to pack capillaries with bare silica particles can represent a consid-
erable investment. The high sensitivity of LIF detection makes open tube FCCE the method of
choice for analytes that can be easily labeled with fluorescent tags such as peptides and protein
fragments.
FCCE in packed capillaries can be used for many more types of analytes. Direct or indirect
UV/Vis absorbance detection can be used to detect everything from monoatomic ions like Br– to
basic and acidic pharmaceuticals.
Pseudo-stationary phases can be used to improve the resolving power of either type of FCCE.
MEKC works well in narrow diameter open tubes for separating neutral analytes based on their
hydrophobicity differences. Chiral selectors such as cyclodextrins can be used with either method.
Other methods of detection such as noncontact conductivity and modes such as CEC have yet to be
explored but may prove useful in the future.
24.4 SYNCHRONOUS CYCLIC CAPILLARY ELECTROPHORESIS

24.4.1 BACKGROUND
Synchronous cyclic CE, as the name implies, uses a series of electrodes placed at regular intervals
around a closed loop to “chase” analyte bands around a continuous separation channel. Voltage
switching is synchronized to maintain analytes of a particular mobility within the separation chan-
nel for a long period of time.73 A similar concept had been applied in chromatography using two
chromatographic columns and timed valve changes to continuously separate analyte bands.74 The
problems connecting separation channels on the extremely small scale of capillary electrophoretic
systems prohibited the implementation of SCCE until several years later.75
24.4.1.1 Microchip SCCE

Burggraf et al.75–77 reported the first practical SCCE system using microfluidic technology. Pho-
tolithographic techniques can be used to cheaply manufacture microfluidic structures with channels
one to several hundred microns in dimension. The interconnection of these channels can be used
to create networks with negligible dead volumes. The first pattern used for SCCE was four over-
lapping, 2-cm channels arranged in a square. An additional channel intersecting the middle of a
main channel was used to perform electrokinetic injections. Several other designs have since been
reported by Manz’s group.78,79 Figure 24.18 shows the layout for an SCCE that was used to separate
fluorescently label amino acids and human urine.79
24.4.1.2 Capillary SCCE

The use of polyimide coated fused-silica capillaries for SCCE with low dead volume connections
was first reported by Zhao et al. in 1999.80,81 Fused-silica capillaries offered several advantages
over the microchip-base SCCE systems. The capillaries provided longer separation distances for
the analyte bands. While the geometry of the connections on a microfluidic device cannot be easily
changed during an experiment, the use of adjustable gaps between the capillaries provides very low
dead volume. In addition, conventional UV-Vis absorption detection was feasible for SCCE for the
first time using capillaries 50 µm in diameter. A basic schematic representation of the instrumental
setup is shown in Figure 24.19.80
Another device has been reported, the “electrophoretron” that used two capillaries with different
polarity surface treatments. The capillaries were connected in a loop and a single power supply was
used to apply a fixed potential across the two capillaries. The different polarity surface treatments
cause the EOF to flow in a continuous loop around the capillaries. Injected analytes would contin-
uously travel around the loop under the influence of the EOF.82 Such a device may prove useful
for CEC separations of neutral analytes. However, any separation of charged species in one capil-
lary will be almost negated in the second capillary due to the change in the polarity of the electric
field.73
7 6
8 5
Sample
Volume defined
injection scheme
9 4
2 3
SW
2.0 cm
5.0 cm
FIGURE 24.18 Microchip-base SCCE system used with MEKC to separate FITC-labeled amino acids and
human urine components. (Reprinted from von Heeren, F., et al., Analytical Chemistry 1996, 68, 2044–2053.
Copyright 1996 American Chemical Society. With permission.)
UV power
supply
UV
detector
2
1
Capillary
Cross-flow cell
Polycarbonate boards
FIGURE 24.19 Schematic representation of capillary-based SCCE system used to resolve isotopomers and
enantiomers. (Reprinted from Zhao, J., et al., Journal of Microcolumn Separations 1999, 11, 431–437. Copyright
1999 John Wiley & Sons, Inc. With permission.)
24.4.2 THEORETICAL ASPECTS

24.4.2.1 Absolute versus “Effective” Voltage
The electric field strength is determined by the applied potential difference divided by the distance
over which it is applied. SCCE is similar to FCCE in that the effective separation voltage can be much
higher than the actual, physically applied voltages. For a single cycle, the effective voltage is twice
that of the applied voltage. For an analyte band transversing n cycles, the effective voltage (Veff ) is
Veff = 2nVapplied , (24.14)
where Vapplied is the applied voltage. Using Equation 24.2, an equation for the expected resolution
between two analyte bands for a given applied voltage can be derived:
1/2
2nVapplied
Rs = 0.177(µ1 − µ2 ) . (24.15)
D((µe + µEOF
24.4.2.2 Band Broadening Mechanisms

There are several sources of band broadening in SCCE to consider. Longitudinal diffusion occurs
in all CE separations. As seen in Equation 24.10, it is mainly governed by separation time and the
analyte’s diffusion coefficient.
Turns in the separation channel create another source of band broadening.83–87 As the analyte
bands migrate around a turn, the portion of the band on the inside of the turn travels a shorter distance
than the portion of the band on the outside of the turn. In addition, the electric field is concentrated
near the wall of the inside of the turn, increasing the rate of band broadening. Taylor dispersion then
smoothes out the distorted band shape. The dispersion created by a turn in the separation channel
can be described as a variance
(l)2 (2θw (1 − exp (−tD /tt )))2

σturn
2
= = . (24.16)
X X
where l is the difference in length traveled around the inside and outside of the turn, θ is the
angle subtended in radians, w is the width at the top of the separation channel, tD is transverse
diffusion equilibrium time, tt is the turn transit time, and X is a constant ∼
=12 when tD is large.83 The
smaller radius of curvature found in microchip-based channels exacerbate this phenomenon, which
is typically not a concern in capillary-based SCCE.
Broadening at the connections (corners on microchip-based SCCE) between the separation chan-
nels is another important source of band broadening and sample loss in SCCE. The total broadening
can be expressed as a variance, σtotal
2

σtotal
2
= σi2 + σj2 = σmigr.dist
2
+ nσcorner
2
, (24.17)
i=sides j=corners
where n is the number of corners the band travels through, σmigr.dist

2 is the variance from the total
migration distance, and σcorner is the broadening caused each time a band travels through a corner
2
or connection of channels.75

24.4.3.1 Power Supply and Voltage Switching
Power supplies for microchip-based SCCE and fused-silica capillary-based SCCE mainly differ in
the number of power supplies and in the magnitude of the applied voltage. The small size found in
most microchip-based SCCE systems limits the separation channel length and so relatively small
potential differences are needed to generate high electric field strengths (∼2 kV/cm). In contrast,
capillary-based SCCE use approximately 50 cm lengths of capillary, and so power supplies generating
up to 30 kV and a few hundred microamperes are typically used. The configuration of the separation
channels determines the number of HV electrodes needed. The actual number of HV supplies needed
can be reduced if relays are used to switch the potential between different electrodes. The fused-silica
capillary-based SCCE instrumentation reported used a single 30 kV HV supply with four HVDC
relays.80 The first chip-based SCCE system used four HV power supplies at 2.5 kV and eight HV
relays to control the voltage switching.75,76 Later work with chip-based SCCE used six 10 kV HV
supplies and thirteen HV relays.79 Different geometries with shorter capillary channels were also
studied. A triangular arrangement resulted in the fewest HV supplies needed.78
24.4.3.2 Connection of Fused-Silica Capillaries

While the connection of the separation channels used in microchip-based SCCE can be manufac-
tured with almost no dead volume using photolithographic techniques, connection of fused-silica
capillaries is challenging. Zhao et al.80,81 report the use of a controllable gap using a cross flow
connection and a tight fitting Teflon sleeve (Figure 24.20). The gap between the capillaries offers
several advantages over other methods. In order to allow the flow of electrophoretic current, the gap
is opened, which also allows fresh buffer to flow into the capillary network and old buffer to be
removed. While the analyte bands migrate through the gap, the gap is tightly closed, reducing the
dead volume dramatically. This system can be implemented with solenoid, piezoelectric, hydraulic,
or even manual actuation. Careful preparation of the capillaries is required to ensure a tight, low loss
to HV
Electrode
Buffer
PEEK tube
Teflon sleeve
Clamp
Capillary
Cross-flow cell
Lever
Pull solenoid
Pull spring
Micropositioners
Buffer
FIGURE 24.20 Schematic diagram detailing the connection of the capillaries and the controlled gap for
capillary-based SCCE. (Reprinted from Zhao, J., et al., Journal of Microcolumn Separations 1999, 11, 431–437.
Copyright 1999 John Wiley & Sons, Inc. With permission.)
fit. Capillaries were cut to precisely equal lengths so that E fields and flow velocities would be equal
in each length of the four capillaries. The ends of the capillaries were polished flat using a jeweler’s
lathe and fine grit sandpaper to ensure a perfectly flat mating surface.80

Capillary zone electrophoresis with microchip-based SCCE has been used to separate fluores-
cent dye from degradation products88 and to separate fluorescein isothiocyanate (FITC) labeled
amino acids.78 Capillary zone electrophoresis (CZE) in fused-silica capillaries has been used to
separate the racemic mixtures of (α-hydroxybenzyl)methyltrimethylammonium and (2-hydroxy-
1-phenyl)ethyltrimethylammonium with β-cyclodextrin as the chiral pseudo-stationary phase.81
l-Phenylalanine and l-phenylalanine-ring-D5 (the hydrogens in the aromatic ring were substituted
with deuterium)81 and another separation of the closely related amino acids phenylalanine and
tyrosine have been resolved using CZE with capillary-based SCCE.80
MEKC has been implemented in both microchip-based SCCE separations and capillary-based
SCCE separations. Figure 24.21a shows a microchip-based SCCE separation using MEKC of FITC-
labeled amino acids.79 The same chip was also used to resolve FITC-labeled components of human
urine using MEKC in addition to performing an MEKC-based immunoassay of serum theophylline
levels.79 MEKC using capillary-based SCCE has resulted in separations with as many as 100 million
theoretical plates in 15 h of run time. A separation of l-phenylalanine and l-phenylalanine-ring-D5 is
(a)
asn
gly
phe ser
arg gln gln
phe
asn
ser 1 1/4
Fluorescence
gly cycles
50 60 70 80
asn
phe ser
* asn ser
1/4 1 1/4 2 1/4 3 1/4

cycle cycles cycles cycles
0 40 80 120 160 200

Time (s)
(b)
gly ser asn phe
arg
gly
ser 1 1/4
asn cycles
*
phe
Fluorescence
60 65 70 75 80 85
ser phe
gly asn
asn
1/4 1 1/4 2 1/4 3 1/4

cycle cycles cycles cycles
0 40 80 120 160 200

Time (s)
FIGURE 24.21 (a) MEKC separation of FITC-labeled amino acids using microchip-based SCCE. (b) CGE
separation of FITC-labeled amino acids using the same microchip-based SCCE device. (Reprinted from von
Heeren, F., et al., Analytical Chemistry 1996, 68, 2044–2053. Copyright 1996 American Chemical Society.
With permission.)
5
(a) All cycles
4
abs., mAU 3
0
0 200 400 600 800
Time (min)
4
(b) Cycle #5
3
abs., mAU
120 121 122 123 124 125 126 127
Time (min)
1.4
1.2
(c) Cycle #26
1.0
abs., mAU
0.8
0.6
0.4
0.2
0.0
715 716 717 718 719 720 721 722
Time (min)
FIGURE 24.22 MEKC separation of phenylalanine and phenylalanine-ring-D5 using capillary-based SCCE.
(Reprinted from Zhao, J. and Jorgenson, J. W., Journal of Microcolumn Separations 1999, 11, 439–449.
Copyright 1999 John Wiley & Sons, Inc. With permission.)
shown in Figure 24.22. These two analytes have only a 0.4% mobility difference and were separated
in only 14 h.81
CGE separations have been reported using microchip-based SCCE. Figure 24.21b shows an
SCCE with CGE separation of FITC-labeled amino acids.79 SCCE with CGE has also been used
with another microchip design for the separation of a PhiX174/Hae III DNA ladder. Separations with
difference synchronization times were used to isolate different DNA fragments.78
24.4.3.4 Methods of Detection

Laser-induced fluorescence detection is a logical choice for high-resolution SCCE when performed
on microfluidic chip-based systems. High laser power can result in photobleaching, decreasing the
analyte signal every time it passes through the beam.75,76,79
UV absorbance can easily be done on capillary-based systems using capillaries at least 50–75
µm in diameter. It is a universal method of detection that can be used with high sensitivity with
most analytes. The short pathlength provide by the shallow channels make UV absorbance detec-
tion difficult to implement on microfluidic systems. In addition, microchip-based systems fabricated
from glass instead of fused silica absorb significant amounts of UV, further complicating its imple-
mentation. In future experiments, detection methods may include direct conductivity measurements
(microchip-based SCCE) and contactless conductivity methods (capillary-based SCCE).
24.4.3.5 Limitations
Theoretically, the ability of SCCE to separate analytes is limited only by longitudinal diffusion. In
practice, broadening from the dead volume connecting the separation capillaries and adsorption of
the analytes to the capillary wall ultimately limit the efficiency of the separation. Microchip-based
SCCE also loses efficiency from broadening around tight turns.
SCCE, similar to FCCE, is a method in which each cycle reduces the range of mobilities that
can be analyzed. The result is that the peak capacity of the separation will be reduced during each
subsequent cycle. This effect is most notable when short channels and low voltages are used.78
The switching of potentials in SCCE causes analytes of the selected mobility to continuously
travel around the separation loop in the same direction. Analytes that migrate at higher mobilities
are lost at the corners and flushed into the waste reservoirs. Analytes of lower mobility that do
not make it into the next separation channel before the potentials are switched will travel in the
reverse direction. This can cause data analysis problems for complex samples when these backward
migrating peaks pass in front of the detector again. This has been observed in several microchip-
based separations including the SCCE separation of fluorescein isothiocyanate and its degradation
products,88 the MEKC and CGE with SCCE separations of FITC-labeled amino acids and human
urine,79 and SCCE of double-stranded DNA in a sieving matrix.78 While this phenomenon is not
much of a problem with relatively simple samples, data analysis may be confusing and complicated
for complex samples such as cell lysates.
24.4.4 METHODS DEVELOPMENT GUIDELINES

24.4.4.1 Handling Extended Separation Times
The selection of the buffer components is very important in SCCE. Just as in FCCE and UHVCE,
careful attention must be paid toward the amount of analyte adsorption to the channel walls. At
extended separation times, losses from adsorption will add to reduce the signal to noise value
greatly. Although the separation channels are flushed with fresh buffer during the SCCE separation,
buffering capacity must be fairly high for long separation times if the reservoir size is small. This is
a bigger problem with microchip-based SCCE where typical reservoir volumes are less than 1 mL.
Capillary-based SCCE typically uses 20 mL buffer reservoirs and buffering capacity problems are
usually not observed.
Similar to FCCE, automation of SCCE voltage switching can greatly improve the reliability
and usability of the SCCE system. Zhao et al.80 report the use of software that determined the
rising and falling signal of the detected analyte above a predetermined threshold value during each
cycle and recalculated switching times to ensure that the analyte bands stayed within the separation
channels.
Longer separation channels have a few advantages over shorter channels when separation times
are long. Longer channels allow longer separation distances without as much broadening and sam-
ple loss associated with channel connections and corners. In addition, longer separation channels
maintain a higher peak capacity during extended run times, and the analyte bands can broaden
considerably without being cut off at the edges.
24.4.4.2 Band Broadening Prevention

Band broadening prevention strategies for SCCE are similar to those used in UHVCE and FCCE.
Known formulas can be used for buffer components if little peak tailing is seen under standard CE
separations. Electrostatic repulsion from channel walls that share the same electrical charge as the
analytes can reduce broadening greatly.
The electrophoretic mobilities of the buffer ions should closely match those of the sample ions
to prevent broadening from electrodispersion. Similarly, the ionic strength of the buffer should be
greater than or equal to that of the sample.
Several strategies for reducing dead volume at the channel connections have been reported. The
use of T-shaped junctions at the middle of the separation channels (Figure 24.19) has been employed
in an effort to reduce losses at the corner connections of the separation channels. Experimental
observation indicated that broadening was more significant than that seen in previous designs with
reservoir connections at the corners.79 Recently, Manz et al.78 have reported the use of side channels
that are much shallower (∼1/8th depth) than the separation channel in an effort to reduce losses at
the connections of the channels.
The use of narrower rounded turns for microchip-based SCCE is reported for reducing band
broadening at the corners.78 This approach has been shown to reduce the broadening seen for turns
in other microfluidic chip designs.83,86,87,89,90 Other, more exotic approaches to reducing dispersion
at the turns of channels on microfluidic devices have been reported. One technique using a pulsed
UV laser to modify the surface of plastic chips at the turn to increase the EOF by up to 4%. This
technique was shown to reduce band broadening at the turns.91
24.4.4.3 Sample Separation Strategies

Amino acids and peptides are easily labeled with fluorescent dyes. CZE, CGE, or MEKC in capillary
or microchip-based SCCE can typically resolve these analytes. Microchip-based SCCE offers higher
sensitivity for very small samples, and higher electric field strengths can be used without Joule heating
problems. Capillary-based SCCE offers a much higher peak capacity and the use of UV absorbance
detection of unlabeled analytes.
Isotopomers are best separated using capillary-based SCCE due to the extremely long separation
times required. The components of the buffer must be carefully selected so that their ionic strength
and ion mobility are very close to that of the analytes and adsorption to the capillary wall is prevented.
MEKC can be used in some cases to separate isotopomers more quickly than CZE, most likely due
to small changes in hydrophobicity caused by deuterium substitution.
Resolution of the enantiomers in a racemic mixture will require the use of a chiral selector such
as a cyclodextrin. Fluorescently labeled analytes can be detected at very low concentration using
microchip-based SCCE, and unlabeled analytes can be easily detected with UV absorption using
capillary-based SCCE. The long separation times available with capillary-based SCCE allows the
use of less than optimal chiral resolving agents, whereas a more selective chiral selector is needed
in microchip-based SCCE.

The three techniques presented in this chapter demonstrate different approaches to improving the
resolution and efficiency of CE separations. All three methods work by increasing the time and
distance that the analyte bands electrophoretically migrate under a high electric field strength, but
the different ways in which each method accomplishes this goal determine their useful application.
UHVCE uses the direct application of extremely high voltages on long capillaries, and it allows
the examination of analytes that span the entire range of mobilities found within the sample. Samples
containing complex mixtures of proteins, peptides, nucleic acids, and many others display a wide
range of mobilities and UHVCE is well suited to their analysis.
FCCE and SCCE use conventional voltages on relatively short capillaries, but they maintain
analytes with a narrow range of mobilities within the capillaries for very long periods of time. Because
high-resolution analysis is limited to such a narrow range of mobilities, FCCE and SCCE are best
suited for the analysis of difficult to resolve pairs such as closely related peptides, enantiomers, and
isotopomers. The ultimate utility of SCCE and FCCE depends on the ability of the experimenter to
eliminate sources of band broadening. The amount of broadening observed in SCCE depends mainly
on the design of the instrumentation. Band broadening in FCCE is inherent to the magnitude of the
counter flow required to reverse the migration of the analyte bands.
The analysis of extremely complex samples such as cell lysates can be simplified through a
combination of these methods. Preliminary high-efficiency, high-resolution analysis with UHVCE
can be used to elucidate areas of interest within the sample. After the mobilities of the analytes of
interest have been determined, FCCE or SCCE can then be used for much more detailed analysis
of particular bands. The analysis of complicated biological samples remains an area where the
implementation of these techniques can provide new insight.
ACKNOWLEDGMENT
National Science Foundation Predoctoral Fellowship Program.
REFERENCES
1. Jorgenson, J. W. and Lukacs, K. D., Capillary zone electrophoresis. Journal of Chromatography
Library 1985, 30 (Microcolumn Sep.), 121–131.
2. Jorgenson, J. W. and Lukacs, K. D., Capillary zone electrophoresis. Science (Washington, DC, United
States) 1983, 222, 266–272.
3. Jorgenson, J. W. and Lukacs, K. D., Free-zone electrophoresis in glass capillaries. Clinical Chemistry
(Washington, DC, United States) 1981, 27, 1551–1553.
4. Jorgenson, J. W. and Lukacs, K. D., Zone electrophoresis in open-tubular glass capillaries. Analytical
Chemistry 1981, 53, 1298–1302.
5. Lukacs, K. Theory, instrumentation, and applications of capillary zone electrophoresis. PhD
Dissertation, University of North Carolina, Chapel Hill, 1983.
6. Henley, W. H. Two distinct approaches for increasing the resolution and efficiency of capillary
electrophoretic separations. PhD Dissertation, University of North Carolina, Chapel Hill, 2005.
7. Hutterer, K. M. and Jorgenson, J. W., Ultrahigh-voltage capillary zone electrophoresis. Analytical
Chemistry 1999, 71, 1293–1297.
8. Palonen, S., Jussila, M., Porras, S. P., Hyotylainen, T., and Riekkola, M. L., Extremely high elec-
tric field strengths in non-aqueous capillary electrophoresis. Journal of Chromatography A 2001,
916, 89–99.
9. Halliday, D., Resnick. R., and Krane, K. S., Physics. John Wiley & Sons, Inc., New York, 1992.
10. Van de Graaff, R. J., Compton, K. T., and Van Atta, L. C., The electrostatic production of high voltage
for nuclear investigations. Physical Review 1933, 43, 149–157.
11. Kraft, E. M. S., Development of capillary zone electrophoresis at 100,000 to 150,000 volts. M.S.
Thesis, University of North Carolina, Chapel Hill, NC, 1996.
12. Hutterer, K. M., Ultra high voltage capillary electrophoresis. Doctorial Thesis, University of North
Carolina, Chapel Hill, NC, 2000.
13. Hutterer, K. M. and Dolnik, V., Capillary electrophoresis of proteins 2001–2003. Electrophoresis
2003, 24, 3998–4012.
14. Hutterer, K. M. and Jorgenson, J. W., Separation of hyaluronic acid by ultrahigh-voltage capillary gel
electrophoresis. Electrophoresis 2005, 26, 2027–2033.
15. Hutterer, K. M., Birrell, H., Camilleri, P., and Jorgenson, J. W., High resolution of oligosac-
charide mixtures by ultrahigh voltage micellar electrokinetic capillary chromatography. Journal of
Chromatography B 2000, 745, 365–372.
16. Verma, R., Shyam, A., and Nair, L., Development of battery powered 100 kV dc power supply. Review
of Scientific Instruments 2006, 77 (10, Pt. 1), 106104/1–106104/4.
17. Gordon, M. J., Lee, K. J., Arias, A. A., and Zare, R. N., Protocol for resolving protein mixtures in
capillary zone electrophoresis. Analytical Chemistry 1991, 63, 69–72.
18. Duso, A. B. and Chen, D. D. Y., Proton and hydroxide ion mobility in capillary electrophoresis.
Analytical Chemistry 2002, 74, 2938–2942.
19. Huang, X., Coleman, W. F., and Zare, R. N., Analysis of factors causing peak broadening in capillary
zone electrophoresis. Journal of Chromatography 1989, 480, 95–110.
20. Moseley, M. A., Deterding, L. J., Tomer, K. B., and Jorgenson, J. W., Determination of bioac-
tive peptides using capillary zone electrophoresis/mass spectrometry. Analytical Chemistry 1991, 63,
109–114.
21. Moseley, M. A., Jorgenson, J. W., Shabanowitz, J., Hunt, D. F., and Tomer, K. B., Optimization
of capillary zone electrophoresis/electrospray ionization parameters for the mass spectrometry and
tandem mass spectrometry analysis of peptides. Journal of the American Society for Mass Spectrometry
1992, 3, 289–300.
22. Kirkland, J. J., Glajch, J. L., and Farlee, R. D., Synthesis and characterization of highly stable bonded
phases for high-performance liquid chromatography column packings. Analytical Chemistry 1989,
61, 2–11.
23. Hardenborg, E., Zuberovic, A., Ullsten, S., Soderberg, L., Heldin, E., and Markides, K. E., Novel
polyamine coating providing non-covalent deactivation and reversed electroosmotic flow of fused-
silica capillaries for capillary electrophoresis. Journal of Chromatography A 2003, 1003, 217–221.
24. Ullsten, S., Zuberovic, A., Wetterhall, M., Hardenborg, E., Markides, K. E., and Bergguist, J., A
polyamine coating for enhanced capillary electrophoresis-electrospray ionization-mass spectrometry
of proteins and peptides. Electrophoresis 2004, 25, 2090–2099.
25. Albarghouthi, M. N. and Barron, A. E., Polymeric matrices for DNA sequencing by capillary
26. Albarghouthi, M. N., Buchholz, B. A., Doherty, E. A. S., Bogdan, F. M., Zhou, H., and Barron, A. E.,
Impact of polymer hydrophobicity on the properties and performance of DNA sequencing matrices
for capillary electrophoresis. Electrophoresis 2001, 22, 737–747.
27. Albarghouthi, M. N., Buchholz, B. A., Stein, T. M., and Barron, A. E., PolyDuramide: A
novel, hydrophilic, self-coating polymer matrix for DNA sequencing by capillary electrophoresis.
Abstracts of Papers, 222nd ACS National Meeting, Chicago, IL, United States, August 26–30, 2001,
ANYL-119.
28. Buchholz, B. A., Doherty, E. A. S., Albarghouthi, M. N., Bogdan, F. M., Zahn, J. M., and Barron, A. E.,
Microchannel DNA sequencing matrices with a thermally controlled “viscosity Switch”. Analytical
Chemistry 2001, 73, 157–164.
29. Carrilho, E., Ruiz-Martinez, M. C., Berka, J., Smirnov, I., Goetzinger, W., Miller, A. W., Brady, D.,
and Karger, B. L., Rapid DNA sequencing of more than 1000 bases per run by capillary electrophoresis
using replaceable linear polyacrylamide solutions. Analytical Chemistry 1996, 68, 3305–3313.
30. Dolnik, V., DNA sequencing by capillary electrophoresis (review). Journal of Biochemical and
Biophysical Methods 1999, 41, 103–119.
31. Dolnik, V., Gurske, W. A., and Padua, A., Galactomannans as a sieving matrix in capillary
32. Heller, C., Separation of double-stranded and single-stranded DNA in polymer solutions. Part 2.
Separation, peak width, and resolution. Electrophoresis 1999, 20, 1978–1986.
33. Salas-Solano, O., Carrilho, E., Kotler, L., Miller, A. W., Goetzinger, W., Sosic, Z., and
Karger, B. L., Routine DNA sequencing of 1000 bases in less than one hour by capillary
electrophoresis with replaceable linear polyacrylamide solutions. Analytical Chemistry 1998, 70,
3996–4003.
34. Slater, G. W., Desruisseaux, C., Hubert, S. J., Mercier, J.-F., Labrie, J., Boileau, J., Tessier, F., and
Pepin, M. P., Theory of DNA electrophoresis: A look at some current challenges. Electrophoresis
2000, 21, 3873–3887.
35. Song, L., Liang, D., Chen, Z., Fang, D., and Chu, B., DNA sequencing by capillary electrophoresis
using mixtures of polyacrylamide and poly(N,N-dimethylacrylamide). Journal of Chromatography A
2001, 915, 231–239.
illary electrophoresis using copolymers of acrylamide and N,N-dimethyl-acrylamide. Electrophoresis
2001, 22, 729–736.
37. Wei, W. and Yeung, E. S., Improvements in DNA sequencing by capillary electrophoresis at elevated
temperature using poly(ethylene oxide) as a sieving matrix. Journal of Chromatography B: Biomedical
Sciences and Applications 2000, 745, 221–230.
38. Zhou, H., Miller, A. W., Sosic, Z., Buchholz, B., Barron, A. E., Kotler, L., and Karger, B. L., DNA
Sequencing up to 1300 bases in two hours by capillary electrophoresis with mixed replaceable linear
polyacrylamide solutions. Analytical Chemistry 2000, 72, 1045–1052.
39. Huber, C. G. and Oberacher, H., Analysis of nucleic acids by on-line liquid chromatography-mass
spectrometry. Mass Spectrometry Reviews 2002, 20, 310–343.
40. Shibata, K., Itoh, M., Aizawa, K., Nagaoka, S., Sasaki, N., Carninci, P., Konno, H., et al., RIKEN
integrated sequence analysis (RISA) system—384-format sequencing pipeline with 384 multicapillary
sequencer. Genome Research 2000, 10, 1757–1771.
41. Culbertson, C. T. and Jorgenson, J. W., Increasing the resolving power of capillary electrophoresis
through electroosmotic flow control using radial fields. Journal of Microcolumn Separations 1999, 11,
167–174.
42. Cheng, Y. F., Wu, S., Chen, D. Y., and Dovichi, N. J., Interaction of capillary zone electrophoresis
with a sheath flow cuvette detector. Analytical Chemistry 1990, 62, 496–503.
43. Williams, B. A. and Vigh, G., The use of hydrodynamic counterflow to improve the resolution and
detection of the minor component in the capillary electrophoretic analysis of enantiomers. Enantiomer
1996, 1, 183–191.
44. Bier, M., Twitty, G. E., and Sloan, J. E., Recycling isoelectric focusing and isotachophoresis. Journal
of Chromatography 1989, 470, 369–376.
45. Everaerts, F. M., Vacik, J., Verheggen, T. P. E. M., and Zuska, J., Displacement elec-
trophoresis. Experiments with counterflow of electrolyte. Journal of Chromatography 1970, 49,
262–268.
46. Everaerts, F. M., Vacik, J., Verheggen, T. P. E. M., and Zuska, J., Isotachophoresis. Experiments with
electrolyte counterflow. Journal of Chromatography 1971, 60, 397–405.
47. Everaerts, F. M., Verheggen, T. P. E. M., and Van de Venne, J. L. M., Isotachophoretic experiments
with a counter flow of electrolyte. Journal of Chromatography 1976, 123, 139–148.
48. Reinhoud, N. J., Tjaden, U. R., and van der Greef, J., Automated isotachophoretic analyte focusing for
capillary zone electrophoresis in a single capillary using hydrodynamic back-pressure programming.
Journal of Chromatography 1993, 641, 155–162.
49. Reinhoud, N. J., Tjaden, U. R., and van der Greef, J., Strategy for setting up single-capillary
isotachophoresis-zone electrophoresis. Journal of Chromatography A 1993, 653, 303–312.
50. Clark, S. L. and Remcho, V. T., Electrochromatographic retention studies on a flavin-binding RNA
aptamer sorbent. Analytical Chemistry 2003, 75, 5692–5696.
51. Lucy, C. A. and McDonald, T. L., Separation of chloride isotopes by capillary electrophoresis based
on the isotope effect on ion mobility. Analytical Chemistry 1995, 67, 1074–1078.
52. Soga, T., Inoue, Y., and Ross, G. A., Analysis for halides, oxyhalides and metal oxoacids by capil-
lary electrophoresis with suppressed electroosmotic flow. Journal of Chromatography A 1995, 718,
421–428.
53. Terabe, S., Yashima, T., Tanaka, N., and Araki, M., Separation of oxygen isotopic benzoic acids
by capillary zone electrophoresis based on isotope effects on the dissociation of the carboxyl group.
54. Yeung, K. K. C. and Lucy, C. A., Isotopic separation of [14N]- and [15N]aniline by capillary elec-
trophoresis using surfactant-controlled reversed electroosmotic flow. Analytical Chemistry 1998, 70,
3286–3290.
55. Yeung, K. K. C. and Lucy, C. A., Ultrahigh-resolution capillary electrophoretic separation with indirect
ultraviolet detection: Isotopic separation of [14N]-and [15N]ammonium. Electrophoresis 1999, 20,
2554–2559.
56. Brewer, A. K., Madorsky, S. L., and Westhaver, J. W., The concentration of K39 and K41 by bal-
anced ion migration in a counterflowing electrolyte. Science (Washington, DC, United States) 1946,
104, 156–157.
57. Chankvetadze, B., Burjanadze, N., Bergenthal, D., and Blaschke, G., Potential of flow-counterbalanced
capillary electrophoresis for analytical and micropreparative separations. Electrophoresis 1999,
20, 2680–2685.
58. Ivory, C. F., Preparative free-flow electrofocusing in a vortex-stabilized annulus. Electrophoresis 2004,
25, 360–374.
59. McLaren, D. G. and Chen, D. D. Y., A quantitative study of continuous flow-counterbalanced capillary
electrophoresis for sample purification. Electrophoresis 2003, 24, 2887–2895.
60. McLaren, D. G. and Chen, D. D. Y., Continuous electrophoretic purification of individual analytes
from multicomponent mixtures. Analytical Chemistry 2004, 76, 2298–2305.
61. Thome, B. and Ivory, C. F., Continuous fractionation of enantiomer pairs in free solution using an
electrophoretic analog of simulated moving bed chromatography. Journal of Chromatography, A 2002,
953, 263–277.
62. Thome, B. M. and Ivory, C. F., Development of a segmented model for a continuous electrophoretic
moving bed enantiomer separation. Biotechnology Progress 2003, 19, 1703–1712.
63. Tracy, N. I. and Ivory, C. F., Preparative isoelectric focusing of proteins using binary buffers in a
vortex-stabilized, free-flow apparatus. Electrophoresis 2004, 25, 1748–1757.
64. Kutter, J. and Welsch, T., The effect of electroosmotic and hydrodynamic flow profile superposition
on band broadening in capillary electrophoresis. Journal of High Resolution Chromatography 1995,
18, 741–744.
65. Culbertson, C. T. and Jorgenson, J. W., Flow counterbalanced capillary electrophoresis. Analytical
Chemistry 1994, 66, 955–962.
66. Iwata, T. and Kuroshu, Y., Enhancement of the sensitivity and resolution by flow-counterbalanced
capillary electrophoresis. Analytical Sciences 1995, 11, 131–133.
67. Henley, W. H., Wilburn, R. T., Crouch, A. M., and Jorgenson, J. W., Flow counterbalanced cap-
illary electrophoresis using packed capillary columns: Resolution of enantiomers and isotopomers.
68. Culbertson, C. T. and Jorgenson, J. W., Separation of fluorescently derivatized deuterated isotopomers
of phenylalanine using micellar electrokinetic chromatography and flow counterbalanced micellar
electrokinetic chromatography. Journal of Microcolumn Separations 1999, 11, 175–183.
69. MacNair, J. E., Lewis, K. C., and Jorgenson, J. W., Ultrahigh-pressure reversed-phase liquid
chromatography in packed capillary columns. Analytical Chemistry 1997, 69, 983–989.
70. Shalliker, R. A., Broyles, B. S., and Guiochon, G., Evaluation of the secondary consolidation of
columns for liquid chromatography by ultrasonic irradiation. Journal of Chromatography, A 2000,
878, 153–163.
71. Tong, D., Bartle, K. D., Clifford, A. A., and Edge, A. M., Theoretical studies of the preparation
of packed capillary columns for chromatography. Journal of Microcolumn Separations 1995, 7,
265–278.
72. Culbertson, C. T. New methods for increasing the resolving power and lowering the UV absorbance
detection limits in capillary electrophoresis. PhD Thesis, University of North Carolina, Chapel Hill,
1996.
73. Eijkel, J. C. T., van den Berg, A., and Manz, A., Cyclic electrophoretic and chromatographic separation
methods. Electrophoresis 2004, 25, 243–252.
74. Ramsteiner, K. A., Systematic approach to column switching. Journal of Chromatography 1988,
456, 3–20.
75. Burggraf, N., Manz, A., De Rooij, N. F., and Widmer, H. M., Synchronized cyclic capillary elec-
trophoresis: A novel concept for high-performance separations using low voltages. Analytical Methods
& Instrumentation 1993, 1, 55–59.
76. Burggraf, N., Manz, A., Effenhauser, C. S., Verpoorte, E., de Rooij, N. F., and Widmer, H. M., Syn-
chronized cyclic capillary electrophoresis—A novel approach to ion separations in solution. Journal
of High Resolution Chromatography 1993, 16, 594–596.
77. Burggraf, N., Manz, A., Verpoorte, E., Effenhauser, C. S., Widmer, H. M., and de Rooij, N. F., A
novel approach to ion separations in solution: Synchronized cyclic capillary electrophoresis (SCCE).
Sensors and Actuators, B: Chemical 1994, 20, 103–110.
78. Manz, A., Bousse, L., Chow, A., Metha, T. B., Kopf-Sill, A., and Parce, J. W., Synchronized cyclic

capillary electrophoresis using channels arranged in a triangle and low voltages. Fresenius Journal of
79. von Heeren, F., Verpoorte, E., Manz, A., and Thormann, W., Micellar electrokinetic chromatography
separations and analyses of biological samples on a cyclic planar microstructure. Analytical Chemistry
1996, 68, 2044–2053.
80. Zhao, J., Hooker, T., and Jorgenson, J. W., Synchronous cyclic capillary electrophoresis using conven-
tional capillaries: System design and preliminary results. Journal of Microcolumn Separations 1999,
11, 431–437.
81. Zhao, J. and Jorgenson, J. W., Application of synchronous cyclic capillary electrophoresis: Isotopic
and chiral separations. Journal of Microcolumn Separations 1999, 11, 439–449.
82. Choi, J. G., Kim, M., Dadoo, R., and Zare, R. N., Electrophoretron: A new method for enhancing
resolution in electrokinetic separations. Journal of Chromatography, A 2001, 924, 53–58.
83. Culbertson, C. T., Jacobson, S. C., and Ramsey, J. M., Dispersion sources for compact geometries on
microchips. Analytical Chemistry 1998, 70, 3781–3789.
84. Fu, L.-M., Yang, R.-J., and Lee, G.-B., Analysis of geometry effects on band spreading of microchip
85. Griffiths, S. K. and Nilson, R. H., Design and analysis of folded channels for chip-based separations.
86. Wang, Y., Lin, Q., and Mukherjee, T., System-oriented dispersion models of general-shaped
electrophoresis microchannels. Lab on a Chip 2004, 4, 453–463.
87. Yao, Z.-H., Yoder, G. L., Culbertson, C. T., and Ramsey, J. M., Numerical simulation of dispersion
generated by a 180 Deg turn in a microchannel. Chinese Physics (Beijing, China) 2002, 11, 226–232.
88. Manz, A., Verpoorte, E., Effenhauser, C. S., Burggraf, N., Raymond, D. E., and Widmer, H. M., Planar
chip technology for capillary electrophoresis. Fresenius’ Journal of Analytical Chemistry 1994, 348,
567–571.
89. Culbertson, C. T., Jacobson, S. C., and Ramsey, J. M., Microchip devices for high-efficiency
separations. Analytical Chemistry 2000, 72, 5814–5819.
90. Ramsey, J. D., Jacobson, S. C., Culbertson, C. T., and Ramsey, J. M., High-efficiency, two-dimensional
separations of protein digests on microfluidic devices. Analytical Chemistry 2003, 75, 3758–3764.
91. Johnson, T. J., Ross, D., Gaitan, M., and Locascio, L. E., Laser modification of preformed polymer
microchannels: Application to reduce band broadening around turns subject to electrokinetic flow.
25 Separation of DNA for
Forensic Applications Using
Lilliana I. Moreno and Bruce McCord
CONTENTS
25.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 761

25.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762
25.3.1 The Capillary and the Sieving Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762
25.3.2 Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 763
25.3.3 Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 764
25.3.4 Size Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 764
25.4 Practical Applications in Forensic Biology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 764
25.4.1 STRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 765
25.4.2 Mini STRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 768
25.4.3 Mitochondrial DNA Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 770
25.4.4 Y-STRs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 771
25.4.5 Single Nucleotide Polymorphisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 773
25.4.6 Mutation Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 775
25.4.7 Nonhuman DNA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 775
25.4.7.1 Animal DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 775
25.4.7.2 Botanical DNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 776
25.4.7.3 Microbial DNA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 777
25.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 779
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 779
25.1 INTRODUCTION
Forensic science is defined as the application of science to the law. It is the goal of the forensic scientist
to identify and compare physical evidence and use the resulting observations to aid in solving criminal
or civil matters. From its inception, forensic scientists have recognized the potential of capillary
electrophoresis (CE) as a useful tool to assist in the analysis of a wide array of trace evidence samples
[1,2]. This is particularly true for applications in forensic biology. In this discipline, biological fluids
(blood, hair, semen) left behind at a crime scene are probed to establish the essential facts of the crime.
Sample analysis is performed via extraction and analysis of the genetic material within these samples.
The key to the procedure is targeting specific locations in the genome containing polymorphisms
(different allelic forms) that permit differentiation between individuals. The statistical probability
761
of an individual inheriting any given polymorphism can then be used to assess the evidence and
compare it with known samples from victims and suspects.
In the early years of DNA analysis, the separation and sizing of DNA was accomplished by slab
gel electrophoresis. However, since 1998 CE has gradually taken over slab gel techniques in this
field because of its ease of automation, minimal sample consumption (which is of utmost importance
in the field), and its high-throughput capabilities.
25.2 BACKGROUND
Although seldom used for genotyping in current forensic applications, slab gels set the precedent
for the development of CE. Slab gel systems had an advantage in forensic analysis in that multiple
samples could be run simultaneously, allowing easy comparisons and rapid analysis. The ability to
compare different samples on the same gel was particularly useful in early procedures involving
restriction fragment length polymorphisms (RFLPs) and multilocus probes. With the RFLP tech-
niques, DNA fragment sizes ranged up to 20,000 bases and results were based on the presence or
absence of bands within different size ranges. Separations took place on large format agarose gels
and genetic loci were detected using Southern blotting with radioactive probes. As a result, a com-
plete analysis could take several weeks because of the time involved in repetitive DNA transfer,
hybridization, and exposure to radioactive film [3].
With the advent of the polymerase chain reaction (PCR), there was a complete paradigm shift.
The PCR permits a few copies of a DNA sequence to be amplified to millions of copies [4], thus
increasing sensitivity and eliminating the need for radioactive probes. DNA fragment sizes were
reduced to a few hundred bases, and the total time for an analysis dropped to a single day. With
PCR, sample quantities were sufficient for detection via simple silver staining processes; how-
ever due to the smaller size of the alleles, higher resolution acrylamide gels were required. As the
exceptional capabilities of the PCR in DNA typing became apparent, law enforcement personnel
began to take advantage of the expanded speed and sensitivity that the new technique had to offer.
Consequently, an increasing number of evidence samples began to be submitted to laboratories for
testing.
Therein arose a new concern. How were laboratories going to be able to handle the bulk of
samples received in a timely manner? An analytical technique was needed that could provide as
good or better results than those obtained via slab gel systems in a standardized and automated
fashion. This technique was to be CE.
The principle governing both slab gel and CE techniques is the same: the separation of a series of
analytes based on size selective sieving through a gel or polymer network under the influence of an
electric field. The slab gel matrix permits a size-dependent separation of fragments since larger DNA
fragments contain an invariant charge-to-size ratio due to their sugar-phosphate backbone. However,
slab gels require manual operations such as the preparation of the gel and the loading of samples.
Automated CE systems eliminated these tedious tasks and substituted replaceable, entangled polymer
solutions for the rigid cross-linked acrylamide gels. Also, the improved heat dissipation of the fused-
silica capillary permits much higher voltages to be used, reducing the run time of the analysis
[5]. Laser-induced fluorescence detection also improved throughput by allowing multiwavelength
analysis [6]. Soon, the forensic science community became interested in this robust technology
and began to adopt it for a variety of analyses. In particular, CE systems can be used for DNA
quantification, genotyping, sequencing, and mutation detection.

25.3.1 THE CAPILLARY AND THE SIEVING MATRIX
In 1988, it was shown that DNA single nucleotide separation could be achieved using cross-linked
polyacrylamide gel filled capillaries coupled with UV/Vis detection [7]. These chemical gels provided
Separation of DNA for Forensic Applications Using Capillary Electrophoresis 763
size selective separation of DNA molecules based on their ability to migrate through transient pores
created in the polymer matrix. Unfortunately, cross-linked gel filled capillaries had a limited life-
time due to the ready formation of voids in the gel during electrophoresis [8]. There were also
concerns with carryover from one run to the next, an important point when precious evidentiary
samples were being run. The solution to this problem was the development of CE systems with
entangled polymer buffers such as polyacrylamide, hydroxyethylcellulose, and polyethylene oxide
[9]. With proper control of polymer molecular weight and concentration, these buffers provided
equivalent separation to cross-linked gels with the added advantage that they permitted refilling and
reuse of the capillary. This property allowed the same capillary to be used up to 100 times before
replacement.
An additional problem with these novel sieving polymers was the effect of the electroosmotic
flow (EOF) on migration time reproducibility. For optimum results, the polymer network must not
only separate the DNA fragments but also eliminate wall effects such as adsorption and osmotic flow.
This is an important issue as the primary function of the method is to precisely determine the size
of the DNA fragments. Variations in migration time due to slight differences in EOF can negatively
affect estimation of fragment size. For this reason, linear polydimethyl acrylamide “POP” became
the polymer of choice for these separations because of its relatively low viscosity and its capability
to eliminate EOF [10,11].
Another critical issue in the development of the separation was the choice of buffer composition.
The optimal buffer should produce low and stable currents and be able to separate DNA under both
denaturing—where the formation of secondary structure of the sample is halted by denaturing agents
such as heat, formamide, or urea [12,13]—or nondenaturing conditions—where secondary structure
formation is desired to study subtle differences in DNA sequences [14,15].
For the majority of genotyping applications such as sequencing and fragment length determina-
tion, denaturing conditions are preferred as resolution is superior and there is a predictable relation-
ship between fragment size and migration time. Buffers such as trishydroxymethylaminomethane
(Tris) and N-tris[hydroxymethyl]methyl-3-aminopropanesulfonic acid (TAPS) are commonly used
as they produce low currents and can buffer at physiological pH. A commonly used buffer for
sequencing and genotyping applications consists of 100 mM TAPS at pH 8.0 with 5% pyrolidinone
and 8 M urea as denaturants [11]. Four to six percent polydimethyl acrylamide at a controlled molec-
ular weight is added to sieve the DNA and eliminate EOF. Uncoated fused-silica capillaries of 50 µm
are typically used as this diameter provides a good compromise between resolution, sensitivity, and
resistance to clogging.
25.3.2 INJECTION
For most applications in forensic DNA analysis, electrokinetic injection is used. This injection mode
provides improved peak shape and intensity as a result of field amplified sample stacking. The stack-
ing process results when the DNA sample is prepared in low ionic strength solutions. The application
of an electric field accelerates the DNA to the sample/buffer interface where it concentrates because
of the drop in field strength at that point. Because this process is highly dependent on the ionic
strength of the sample solution, great care must be taken to keep salt concentrations low. Unfortu-
nately PCR amplifications occur at relatively high-salt concentrations (70–100 mM KCl) and various
buffers are often added to the stabilize DNA for long-term storage. Thus, the sample must be diluted
or dialyzed before injection as smaller buffer ions have higher electrophoretic mobility and interfere
with the injection of the DNA.
Sample injection is an important issue in forensic analysis as there should be a semi-quantitative
relationship between peak height and sample concentration. This relationship helps the analyst
assess the quality of the sample and its preparation. It also helps define the relative level of different
contributors to a mixture. The removal of salts through dialysis or the unintentional addition of
higher quantities of salt can produce peak intensities that are less representative of sample quantity
and affect the aforementioned relationship.
In a typical injection, 1–2 µL of a completed PCR is diluted in 10–20 µL of formamide. Water

can also be used but the formamide produces complete denaturation of the sample without further
processing. High purity, low conductance formamide must be used to avoid loss of peak intensity or
other injection artifacts.
25.3.3 DETECTION
Laser-induced fluorescence is the most common method of detection for DNA analyses due to
its high sensitivity and multiplex capabilities. Typically argon-ion lasers are used with excitation
wavelengths of 488 and 514 nm. Detection occurs using a charge-coupled device (CCD) that collects
the fluorescence emission produced by the various dyes bound to the DNA molecules. There are
three basic methods for detection of the DNA using these dyes: intercalation, amplification with
dye-labeled PCR primers, or incorporation of dye-labeled bases into the DNA sequence during
replication. For single channel detection of native DNA, intercalating dyes produce excellent results
[16]. These dyes may simply be added to the CE buffer and the DNA is labeled as it moves through
the gel toward the detector. For genotyping of denatured DNA, dye-labeled primers are used [17]
while dye-labeled bases are used in DNA sequencing [18].
The specific dyes used for genotyping and DNA sequencing are designed to simultaneously
absorb at a single laser wavelength but emit at a variety of different wavelengths. Rhodamine and
fluorescein derivatives are commonly used. Using these dyes, multiple loci can be amplified and
genotyped without interfering with each other by simply labeling each set of PCR reactants with a
different dye. Current commercial systems have the capability to detect as many as five different
dyes simultaneously on a single CE capillary. Special software is used to eliminate problems with
dye overlap by applying virtual filters and various calibration procedures [17].
25.3.4 SIZE ESTIMATION

As mentioned previously, the primary function of the separation in genetic analysis is the estimation
of fragment size. Multichannel fluorescence systems typically reserve one channel for use as an
internal standard. The internal standard consists of multiple peaks throughout the size range of the
analysis and is used by the computer to permit a precise estimate of unknown fragments in the other
dye channels. Computer algorithms can then be used to produce a size estimation of the unknown
DNA fragments in the other dye lanes. The samples can be further processed by comparison of these
data with an external standard that consists of all known mutations. Using both sets of standards, it
is possible to produce size estimates with a precision of better than 0.17 base pairs (bp) making the
system capable of distinguishing single base differences at sizes up to 350 bp [19]. It is important to
note at this point that the size estimates produced by these techniques can be influenced by temperature
or sequence effects, and thus careful control of temperature and denaturant concentration is important
in order to maintain precision [20,21].
25.4 PRACTICAL APPLICATIONS IN FORENSIC BIOLOGY

The major application of CE in forensic biology is in the detection and analysis of short tan-
dem repeats (STRs). STR markers are preferred because of the powerful statistical result that is
possible with these markers and the large databases that exist for convicted offenders’ profiles.
Other related applications include the analysis of haploid markers in the Y chromosome and in
mitochondrial DNA (mtDNA). Nonhuman DNA testing can also be performed depending on the cir-
cumstances of the case. The techniques involved include genotyping, DNA sequencing, and mutation
detection.
25.4.1 STRs
The analysis of STR loci in DNA is the most common method for the determination of human
identity and can indisputably distinguish between two or more unrelated individuals if sufficient loci
can be detected [22]. STR loci occur in noncoding regions of the human genome and consist of
short segments of DNA 2–7 bp in length such as AATG, which are repeated consecutively multiple
times. The number of repeats at a given locus can vary between individuals and there is a statistical
probability that a given individual will have a set number of repeats at a particular STR locus. It
has been estimated that over 100,000 STR loci exist in the human genome and research continues
in an effort to determine their exact function [23]. Among this large number of STRs, the forensic
community in the United States has established a set of 13 loci (Table 25.1) [24,25] that can be used
to develop a genetic profile for the identification of individuals in criminal casework. To process the
results from each analysis, large database known as Combined DNA Index System (CODIS) has
been set up. This database stores profiles from convicted offenders and unsolved casework. Similar
databases have been set up in Europe, Japan, and other countries. The information in these databases
can be used to detect and apprehend serial offenders by permitting rapid exchange of information
between crime laboratories [26–28].
STRs can be targeted for PCR amplification by preparing primer sequences that bind to more
conserved sequences flanking the variable STR regions. The chemistry and protocols necessary for
identifying each set of STRs are included in one or a combination of kits provided by companies such
as Applied Biosystems and Promega. These kits (Table 25.2) have been subjected to strict validation
processes to ensure the quality of the data [19,29–31]. The kits permit multiplex PCR amplification
of up to 16 loci (including the sex determination gene) simultaneously from a single sample (Figure
25.1). The different loci included in the kit contain multiple alleles and are separated by size and dye
label. The repeat motifs are 4–5 bases in length and the motif sequence can repeat itself up to 51 times
(Table 25.1). To help define the size and migration time of each known allele an external standard
known as an allelic ladder is run subsequent to each set of samples and used to more precisely define
the identity of each peak (Figure 25.2).
TABLE 25.1
Thirteen STR Markers Commonly Used for
Forensic DNA Analyses in the United States
Marker (Locus) Repeat Motif Allele Range
CSF1PO TAGA 6–16
FGA CTTT 15–51.2
THO1 TCAT 3–14
TPOX GAAT 6–13
vWA [TCTG][TCTA] 10–24
D3S1358 [TCTG][TCTA] 9–20
D5S818 AGAT 7–16
D7S820 GATA 6–15
D8S1179 [TCTA][TCTG] 8–19
D13S317 TATC 5–15
D16S539 GATA 5–15
D18S51 AGAA 7–27
D21S11 [TCTA][TCTG] 24–38
Source: Adapted from Butler, J.M., Forensic DNA Typing:

Biology, Technology, and Genetics of STR Markers, 2nd ed.,
Academic Press, San Diego, CA, 2005. With permission.
TABLE 25.2
Commercially Available Human STR Amplification Kits
Kit Name Target Loci Discrimination Power
(A) Applied Biosystems, Foster City, CA

AmpFLSTR® Profiler® D3S1358, vWA, FGA, THO1, TPOX, CSF1PO, D5S818, 1:3.6 × 109
D13S317, D7S820, and Amelogenin
AmpFL® SEFiler™ D2S1338, D3S1358, D8S1179, D16S539, D18S51,
D19S433, D21S11, SE-33, FGA, vWA, and Amelogenin
AmpFL® Cofiler® CSF1PO, D16S539, THO1, Amelogenin, TPOX, 1:8.4 × 105
D3S1358, and D7S820
AmpFL Profiler Plus D3S1358, D5S818, D7S820, D8S1179, D13S317, 1:9.6 × 1010
D18S51, D21S11, FGA, vWA, and Amelogenin
AmpFL® SGM Plus® D2S1338, D3S1358, D8S1179, D16S539, D18S51, 1:3.3 × 1012
D19S433, D21S11, THO1, FGA, vWA
AmpFL® Identifiler® CSF1PO, D3S1358, D5S818, D7S820, D8S1179, 1:2.1 × 1017
D13S317, D16S539, D18S51, D21S11, vWA, FGA,
THO1, TPOX, D2S1338, D19S433, and Amelogenin
AmpFL® Green I THO1, TPOX, Amelogenin, and CSF1PO 1:410
AmpFL® Blue D3S1358, vWA, FGA 1:5000
(B) Promega Corporation, Madison, WI
PowerPlex® 16 System Penta E, D18S51, D21S11, THO1, D3S1358, FGA, 1:1.8 × 1017
TPOX, D8S1179, vWA, Amelogenin, Penta D, CSF1PO,
D16S539, D7S820, D13S317, and D5S818
Source: Applied Biosystems, Foster City, CA (www.appliedbiosystems.com) and Promega Corporation, Madison,
WI (www.promega.com).
D8S1179 D7S820 CSF1PO

D21S11
D3S138 THO1
D13S317
D16S539 D2S1338
vWA TPOX D18S21

D19S20
AMEL
D5S818
FGA
FIGURE 25.1 AmpFlSTR Identifiler™16-plex amplification results. The electropherogram shows 15 STR
loci as well as the amelogenin (sex determining) marker. Individual STR loci are separated by amplicon size
and dye label. Each row represents a different dye marker: 1, the 6-FAM labeled loci; 2, VIC labeled loci;
3, NED labeled loci; 4, PET labeled loci; an internal standard labeled in rox is also run simultaneously but is
not shown. (Courtesy of Ada Nuñez, Florida International University.)
100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360
* Ladder.fsa 2 Blue ladder
D21S11
D8S1179 D7S820
3000
CSF1PO
2000
1000
* Ladder.fsa 2 Green Ladder

THO1 D13S317 D16S539 D2S1338
D3S138
2000
1000
* Ladder.fsa 2 Yellow ladder
D19S20 vWA TPOX D18S21 4000
3000
2000
1000
* Ladder.fsa 2 Red ladder

2000
AMEL D5S818 FGA 1500
1000
500
FIGURE 25.2 AmpFlSTR™Identifiler kit allelic ladder. A separation of the most common alleles for each
STR loci. The STR analysis software utilizes this ladder as an external standard to assign the correct alleles
to evidence samples such as is illustrated in Figure 25.1. Notice the presence of two base variant alleles (from
the normal four base STR repeat) at the D21, D18, D19, and FGA loci. There is also a one base variant allele
located in the THO1 locus. (Courtesy of Ada Nuñez, Florida International University.)
One of the major issues in the separation of these large multiplex sets by CE is the potential
presence of variant alleles that differ from the standard 4 base repeat unit by 1 or 2 bp. Single base
resolution and high precision are necessary over the range of fragment lengths up to 350 bp in order
to reliably detect these variant alleles and distinguish them from artifacts, spikes, and noise.
Another issue that is unique to this application is the role of the PCR in defining system sensitivity.
Amplification of DNA concentrations less than 100 pg (about 17 cells) can produce stochastic
intensity fluctuations leading to peak imbalance and occasional loss of signal (allele dropout). Low
level mixtures may also be present in the electropherogram further complicating the interpretation
of the data [32,33]. The ability to produce clear and unambiguous electropherograms is critical
in criminal casework since DNA evidence may be the only information tying the suspect to the
crime. Loss of peak intensity can complicate data interpretation. As a result, rigid rules have been
developed for interpretation of signals below an intensity threshold that can be defined by the ability
to reproducibly amplify DNA and detect it at a level above the system’s limit of quantification [34].
For mixtures such as those often encountered in sexual assault cases, special techniques have
been developed for isolation of male DNA (sperm heads) from female epithelial cells [35]. These
procedures aid the interpretation of the data by removing most of the female contribution to the
profile. This process, known as differential extraction, can be performed by selective digestion of
the epithelial cells followed by isolation of undigested sperm heads through centrifugation. The
sperm fraction is further digested and extracted using dithiothreitol (DTT). However in certain
situations such as the case with vasectomized males or multiple assailants, isolation of individual
profiles is more difficult. In these cases careful analysis of the results may be necessary to reach
FIGURE 25.3 Mixed male/female DNA Profile. The amplification of a 1:1 mixture of male and female DNA
using the Identifiler™STR multiplex. Note the presence of 2, 3, or 4 alleles at each locus indicating the presence
of 2-, 1-, or 0-shared alleles. Also, notice the 3:1 ratio of X to Y sex typing alleles in the third panel due to the
mixture of XX and XY alleles. Analysis performed on an ABI310 Genetic Analyzer. (Courtesy of Stephanie
King and George Duncan, Broward County Sheriff’s Office Crime Laboratory.)
a conclusion. Figure 25.3 illustrates the electropherogram of a mixed profile. Note that the sample
can be determined to be a mixture of male and female DNA based on the relative peak areas of the
Amelogenin sex marker. In this case there is a 3/1 ratio of X- to Y-chromosome, suggesting a 1/1
mixture of male (XY) to female (XX) DNA.
Once the separation is complete, the data are analyzed using specialized software that assigns an
allele repeat number to each peak based on the alleles identified by the previously run allelic ladder.
Statistical analyses can be performed using relevant population frequencies to determine the overall
probability of another individual having an identical DNA profile [36]. Because the different genetic
loci used in determining the profile are inherited independently, the individual probabilities of having
a particular set of alleles at one locus can be multiplied together for all loci producing a random match
probability for the 13 CODIS loci of 1×10−15 [24]. For all intents and purposes, the data from a single
source profile such as that shown in Figure 25.1 can be used to provide identification of a suspect
[37]. Analysis of more complex samples such as those containing multiple contributors (Figure 25.3)
or related individuals is less straightforward and requires in-depth statistical analysis [38–41].
25.4.2 MINI STRs

Large STR multiplex sets are used for most forensic evidence samples. However, biological fluids
or materials left behind in a crime scene may be degraded because of exposure to a variety of harsh
(a)
(b)
FIGURE 25.4 (See color insert following page 810.) A comparison between a 9947 control DNA sample (a)
and a degraded DNA sample (b) extracted from a recovered bone fragment. There is an evident loss of signal
at the larger size loci in sample b (arrow) due to sample degradation. The sample is extracted and amplified
DNA prepared using the Promega Powerplex STR multiplex kit and consists of 16 separate genetic loci labeled
with three different dyes and separated via capillary gel electrophoresis using the ABI 310 genetic analyzer.
(Courtesy of Kerry Opel, International Forensic Research Institute, Florida International University.)
conditions and/or due to the presence of various contaminants that have been mixed with the sample
[42,43]. In both cases, the result may be a partial DNA profile (Figure 25.4) in which some alleles
are missing from the profile or are present below a laboratory’s interpretational threshold [44,45].
Characteristically, the larger alleles lose intensity due to increasing decomposition of the template into
smaller fragments. The resulting partial electropherograms are far less definitive than a full profile
since fewer loci are available for statistical analysis. This leads to an increase in the probability of
finding a random unrelated individual in the general population with a matching profile.
In these situations, one approach has been to perform further testing using mtDNA sequencing.
Since there are multiple copies of mitochondrion in each cell, the likelihood of obtaining a result is
greatly increased when compared to nuclear DNA. However, mtDNA analysis involves difficult and
expensive analytical procedures, and because it is inherited solely through the female line, maternally
related individuals will all have the same profile and statistical results are much less conclusive.
In recent years, investigators driven by this issue have developed a viable alternative—the use
of mini STRs. Mini STRs are reduced size STR amplicons that can be obtained by redesigning the
amplification primers in such a way that they bind closer to the STR repeat regions [46–48]. To keep
the fragment size as short as possible only one or two STR loci can be used in each dye lane. These
shorter amplicons can be detected when the original template is too fragmented to properly amplify
(Figure 25.5).
Many of these STR markers targeted by modified primers are the same as those already established
for the CODIS database [46]. Concordance studies between the mini STR and the commercially
available STR primers have been performed and some discrepancies (0.2%) have been observed in
the allele calls. These, however, can be explained by the fact that the primers from the mini STRs
bind at different locations than those of the original STRs, thus insertions/deletions can affect primer
FIGURE 25.5 Mini STR amplification. The electropherogram illustrates the analysis of an extracted and
amplified bone sample. The electropherogram consists of four panels showing three STR loci and an internal
lane standard run simultaneously and detected with four different dyes. Unlike the larger multiplexes illustrated
in previous figures, only one locus appears in each dye lane in an effort to keep the amplified products as
short as possible. (Courtesy of Kerry Opel, International Forensic Research Institute, Florida International
University.)
binding resulting in an apparent discrepancy in allele size [49]. This problem is being addressed
through alternative primer designs, and researchers are continuing to identify new mini STRs to
increase the information content of STR assays [50].
25.4.3 MITOCHONDRIAL DNA ANALYSIS

In situations in which minimal DNA can be recovered from a sample due to severe degradation or
lack of recoverable DNA, mtDNA can be exploited due to its relatively small size and presence in
multiple copies within the cell [51]. The mitochondrion is a self-replicating organelle that is able
to synthesize its own DNA [52]. The DNA is circular and contains 16,569 bp [53]. It contains both
coding and noncoding regions, and it is in the noncoding hypervariable regions, HVI and HVII,
where information relevant to forensic information is found [18,54]. These hypervariable regions
have high-mutation rates that enable them to be a useful tool for human genetic analysis [55]. Owing
to its small size, mtDNA does not contain STRs or other repetitive elements and instead analysis
relies on variations in sequences [51,56]. Mutation detection takes place through the analysis of single
nucleotide polymorphisms (SNPs); deletions, additions, and substitutions of various nucleotides in
HVI and HVII. The treatment the samples receive before being loaded into the CE system is also
different from that of the STRs. Because these samples are generally present at low copy number
and are highly degraded, extensive use of control samples with strict isolation protocols is necessary
in order to avoid cross contamination and maintain data reliability.
The samples are first amplified and the resultant native, unlabeled PCR products are quantified
using CGE or microfluidic CGE with fluorescent intercalating dyes to determine input levels and
FIGURE 25.6 Quantitative analysis of PCR-amplified mtDNA by microfluidic CE. Determination of the
overall quality and quantity of the DNA before sequencing is important to assure high-quality results. The figure
illustrates the detection of an mtDNA amplicon from the HVI region located between two internal standards
used for sizing. The analysis is performed on an extracted blood stain by an Agilent 2100 bioanalyzer using
fluorescent detection with an intercalating dye. (Courtesy of the DNA Analysis Unit II at the FBI Laboratory,
Quantico, VA.)
sample quality for the subsequent sequencing reaction [16,57] (Figure 25.6). Sequencing reactions
are then performed using the PCR template, and products are separated using denaturing CGE. The
results are compared to a reference sequence to catalog the specific point mutations that are present
[58] (Figure 25.7). The statistical analyses of these data are not as definitive as that obtained from
STR typing since mitochondria are transmitted directly from mothers to their progeny and there is
no admixing or shuffling of genetic information such as occurs in meiosis. Thus mtDNA analysis
cannot produce the high-statistical certainty of identification produced in STR typing and instead is
most useful as a means for maternal lineage determination [59,60]. However, in situations in which
the only evidence available is badly degraded or where only a few cells containing DNA are present
(such as the situation in which a single shed hair is recovered), mtDNA may be the only way useful
genetic information can be obtained [51,61–63].
25.4.4 Y-STRs
As with mtDNA, which provides maternal lineage, the Y-chromosome can be used to provide paternal
lineage. The first Y-STR was discovered in 1992 [64] Since then, a number of additional Y-STRs have
been validated for use in the forensic field [65,66] and several commercial multiplex amplification
kits are now available (Table 25.3). Initially used for paternity testing and rape case scenarios [67,68],
Y-STRs are now being used to aid in missing persons investigations [69], genealogical research, and
evolutionary studies [70–72]. Y-chromosome markers are particularly useful in the detection of small
amounts of male DNA in the presence of an overwhelming abundance of female DNA. This type of
STR loci are most useful when trying to isolate the male fraction from a DNA mixture (Figure 25.8)
[73,74]. .
FIGURE 25.7 (See color insert following page 810.) Mitochondrial DNA profile of an HVII sequence from
a hair shaft. A repetitive analysis of the same hair sample is used to compare and align the sequences. The
nuclear DNA in hair is often badly degraded and difficult to amplify. mtDNA provides an alternative procedure
that can provide a DNA haplotype for forensic DNA profiling based on point mutations detected in the sequence
when compared to a reference sample. Analysis performed on an ABI 310 Genetic Analyzer. (Courtesy of the
DNA Analysis Unit II at the FBI Laboratory, Quantico, VA.)
TABLE 25.3
Validated Y-STR Loci for Forensic Casework
Locus Repeat Size
DYS393 AGAT 108–132

DYS392 TAT 236–263
DYS391 TCTA 275–295
DYS389I (TCTG)(TCTA) 239–263
DYS389II (TCTG)(TCTA) 353–385
Y-GATA A7.2 TAGA 174–190
DYS438 TTTTC 203–233
DYS385 GAAA 252–300
DYS19 TAGA 242–254
DYS425 TGT 104–110
DYS388 ATT 119–131
DYS390 (TCTA)(TCTG) 200–251
DYS439 AGAT 242–258
DYS434 ATCT 106–116
DYS437 TCTA 188–192
Y-GATA C.4 TATC 250–269
Y-GATA A7.1 ATAG 104–112
Y-GATA H.4 TAGA 130–143
Source: Adapted from Daniels, D.L., Hall, A.M., and Ballantyne J.,
J Forensic Sci, 43, 668, 2004. With permission.
sample file sample name
FIGURE 25.8 A Y-STR profile of the mixture in Figure 25.3 using the ABI Y Filer™Y chromosomal STR
multiplex. Note that the DY385 locus has two alleles due to a duplication of the sequence on the Y chromosome.
Analysis performed using an ABI 310 Genetic Analyzer. (Courtesy of Stephanie King and George Duncan,
Broward County Sheriff’s Office Crime Laboratory.)
An example of such a scenario might be a fingernail scraping from a female victim or the
detection of a mixed DNA profile from the handle of an automobile. The profile depicted in Figure
25.9a results from an extract taken from a knife handle that includes DNA from the male suspect
along with two female profiles. Use of the Y profile (Figure 25.9b) permits isolation of just male
DNA, although it should be noted that all of the suspect’s male relatives from his paternal lineage
would also have the same Y-haplotype. Statistical analysis of the result involves performing an
estimate of the frequency of a given profile in a database. Frequency estimates for Y profiles are
much less specific than those obtained with autosomal STRs as individual allele frequencies cannot
be multiplied together. Nevertheless, the data provide important information regarding the potential
placement of an individual at a crime scene.
25.4.5 SINGLE NUCLEOTIDE POLYMORPHISMS

The various types of point mutations detected via mtDNA sequencing can also be targeted in nuclear
DNA. SNPs are particularly useful in ethnicity testing and in the analysis of highly degraded or
compromised samples [57,75,76]. In the human genome SNPs occur every 1000–2000 bp, thereby
(a)
(b)
FIGURE 25.9 A comparison of (a) a mixed DNA profile (two females and one male contributor) followed
by (b) the Y-STR profile of the same sample. The sample was collected with a single surface swab (moistened
with DI water) from a knife handle (homicide case). Amplification was performed using PowerPlex®Y with
1.2 ng and Profiler Plus™with 0.8 ng of template DNA. “The male component DNA obtained from the knife
handle and the suspect has the same Y-haplotype; therefore, the suspect could not be eliminated as the source
of the male DNA in the mixture. The results also indicate the presence of only one male contributor in the DNA
extract. It should be noted that all of the suspect’s male relatives from his paternal lineage would also have the
same Y-haplotype.” (Courtesy of DNA Unit—Orange County Sheriff-Coroner Department, CA.)
accounting for about 90% of genetic variation [75,77,78]. Unlike tandem repeats, SNPs are found
both in the coding and noncoding regions of the DNA molecule. They can play an important role in
the field of forensics as they have, in contrast to STRs, lower mutation rates and may eventually be
linked to physical features such as hair color, stature, and skin shade [79,80]. Unfortunately, SNP
systems have limited numbers of alleles and therefore many more loci are required to perform a full
genetic profile than is the case with STRs. In fact, about 50 SNPs would require analysis to achieve
the same level of statistical uniqueness developed with 13 STRs [78]. In addition, SNPs cannot
easily be used in mixture studies as fragment sizes overlap making it difficult to isolate multiple
contributors to a profile. SNPs can be detected using a single base primer extension assay in which
a special primer is designed to target a known SNP location [81]. A polymerase is then added and
used to probe the site of the polymorphism (the next unincorporated base pair) through the addition
of a terminal fluorescently labeled dideoxynucleotide (ddNTP). The SNP assay can be multiplexed
by simultaneously labeling multiple numbers of these loci and using changes in primer length to
permit all locations to be detected simultaneously. Like mtDNA sequencing and mini STRs analysis,
SNPs can be valuable in the recovery of information from degraded DNA. Figure 25.10 illustrates
the comparison of results from the standard STR typing of a degraded DNA sample followed by
SNP typing of that same DNA extract. The SNP amplicons are much shorter and provide additional
genetic information to assist in the identification of the partial STR profile.
25.4.6 MUTATION DETECTION

Although more commonly used in oncology and detection of genetic diseases [82–84], mutation
detection techniques can provide useful information in forensic analysis. Sequence polymorphisms
resulting from one or more base pair changes in PCR-amplified DNA fragments can be exploited
through differences in melting points, heteroduplex formation, or conformational variations. A
variety of different CE procedures exist to detect these differences, including single-stranded confor-
mational polymorphism (SSCP), heteroduplex polymorphism, and constant denaturant CE. Using
these techniques, DNA fragments of the same or similar length that would otherwise comigrate can
be differentiated based on mobility differences resulting from the effect of temperature, denaturant
concentration, or rehybridization effects. These types of mobility assays have been used in paternity
disputes [85] and in ABO allele discrimination [86]. These techniques are relatively inexpensive to
perform and can be highly sensitive to slight differences in sequence [87] making them highly useful
in the detection of previously unknown mutational events.
25.4.7 NONHUMAN DNA

Nonhuman DNA can provide crucial information in a variety of crime scenarios [88]. Items found
at a crime scene, such as animal hair, plant material, fungal spores, and soils, may all contain
recoverable DNA. Genetic markers such as STRs have been isolated from these types of materials
and a number of different loci have been validated for use in criminal casework [89,90]. For other
items such as microbial DNA, soil and plant materials, sequence information may not exist, and
various exploratory procedures must be used to recover genetic information.
25.4.7.1 Animal DNA

Feline and canine STR analyses have been used in forensic casework [91–94] and specific kits
have been developed to determine the genetic fingerprint of both animals [95,96]. Canine genetic
material has also been analyzed via CE using mtDNA sequencing [97]. As with human samples,
canine DNA can help link a suspected animal to crimes such as bite attacks or mauling. Similarly,
it can also be used to exclude the animal from implication in attacks or unlawful deaths [98]. More
commonly, however, dog hair and other animal hair has been found associated with victims at crime
scenes [94,99]. Figure 25.11 illustrates an electropherogram of a suspect dog hair that might be used
to associate an animal with the victim of an abduction. Genetic analysis of this material provides
further information on the circumstances of the crime and can connect the victim to a particular
FIGURE 25.10 (See color insert following page 810.) The recovery of information from degraded DNA
using the Identifiler™STR multiplex and an SNP multiplex, both analyzed using the ABI 310 genetic ana-
lyzer. The top electropherogram depicts a degraded sample amplified via STRs. The bottom profile depicts a
SNP multiplex on the same sample. The STR profile is blank for many larger alleles such as D7 CSF, D16,
D2, TPOX, D18, and FGA. The smaller sized SNP fragments permit recovery of additional genetic infor-
mation from the sample. (From Vallone, P.M., Decker, A.E, and Butler, J.M., The evaluation of Autosomal
SNP assays, Chemical Science and Technology Laboratory, National Institute of Standards and Technology,
Div.831.)
location or suspect. Although more commonly used for lineage analysis, other genetic material such
as equine and bovine DNA could also be used in criminal casework [100].
25.4.7.2 Botanical DNA

Plant DNA analysis has been increasing over the past couple of years because of its ability to
pinpoint the origin of drug-related plant material [101,102]. Plant STRs are not yet well characterized;
however, other molecular methods can be combined with capillary gel electrophoresis to establish the
identity of plant species [103]. These techniques include amplified fragment length polymorphism
(AFLP), random amplification of polymorphic DNA (RAPD), and other similar random primer
annealing techniques used for mutation detection.
AFLP is a molecular tool in which the DNA is cut by a combination of restriction enzymes for
subsequent amplification utilizing relatively nonspecific primers that are tagged with a fluorescent
dye [104,105]. These primers will anneal only to the fraction of the cut fragments that contain a
short complimentary sequence, leaving behind those that do not. Since DNA sequences vary, the cut
FIGURE 25.11 Canine STR analysis. Ten canine markers were successfully amplified and analyzed after
extracting DNA from a single dog hair. (Courtesy of Lilliana Moreno, Forensic DNA Profiling Facility, Florida
International University.)
sites produced by the selected restriction enzymes will vary between the different plant materials
submitted for analysis. The result is a set of DNA fragments of different sizes and intensities that
can be related to a particular plant cultivar. The specific fragments can be isolated and detected
by capillary gel electrophoresis. Various computer software are then used to define and catalog the
differences between samples. AFLP has proven to be a particularly valuable tool for establishing
plant identification and for tracing plant material back to its original source [106]. Figure 25.12
illustrates the analysis of a seized marijuana sample using AFLP with multichannel fluorescent CE
detection.
25.4.7.3 Microbial DNA

Recent events involving the use of microbes as bioterrorists’ weapons have intensified research in
the developing field of microbial forensics. This new field couples law enforcement efforts with
existing procedures to identify patterns in disease outbreaks, determine the pathogen involved,
control its spread, and trace the microorganism back to its source [107]. A variety of techniques are
currently being used as tools for the classification and identification of microorganisms and these
procedures can also be used to assist investigators in the detection of criminal acts. While most of
the reported incidents to date involve hoaxes with nonpathogenic material, the consequences of such
an attack require that laboratories be properly prepared to meet the threat. In addition, these same
techniques can also be used to examine trace evidence such as soils left behind following a criminal
act [108,109]. A number of molecular techniques for microbial detection and analysis are listed
below.
Terminal restriction fragment length polymorphism (TRFLP) uses end-labeled primers that will
bind to specific primer sequences used when amplifying DNA by PCR [110,111]. After digestion of
the PCR products with restriction enzymes, the samples are loaded into the CE system for separation.
FIGURE 25.12 Genetic analysis of a marijuana sample using AFLP. The sample consists of amplified DNA
fragments characteristic of the plant overlaid on top of an internal lane standard used to size the individual
fragments. Analysis performed on an ABI 310 genetic analyzer. (Courtesy of Ira Lurie, Yin Shen, and Bruce
McCord.)
FIGURE 25.13 Soil microbial DNA profile. The 16S rRNA gene hypervariable region V1_V2 was amplified
using ALH technique. The different DNA fragments identified in the figure depict the different microbes present
in the soil. These data can then be used to aid in the identification of soils associated with a crime scene and to
help answer important questions such as has a body been moved from its current location. Analysis performed
using an ABI 310 genetic analyzer. (Courtesy of Lilliana Moreno, Forensic DNA Profiling Facility, Florida
International University.)
This method has been extensively used in conjunction with the small ribosomal subunit (16S
rRNA) to establish differences between microbial entities [112,113] and was the technique first used
in the analysis of microbial communities for forensic purposes [108]. Subsequent researchers in
microbial forensics examined a method known as amplicon length heterogeneity (ALH), which is
commonly used in microbial ecology. ALH bases its profiles on differences in length within select
hypervariable domains of the 16S rRNA genes. DNA extracts from different microbial communities
are amplified using universal primers that bind outside of the variable region and are capable of
hybridizing to the majority of target organisms. The output is a pattern of peaks that provide, just
like STRs, a unique pattern that is used to differentiate microbial communities (Figure 25.13). Unlike
TRFLP, this highly reproducible method does not require restriction endonucleases but is based on
the natural variation in sequence lengths of specific regions within the gene [114,115]. Because of
its ubiquitous nature, soil can be a valuable evidence in crime scene investigations; however, it is
seldom used because the physical analysis is complicated and requires experts in the field of geology.
A novel approach utilizing ALH-PCR to discriminate between soils for forensic purposes has been
developed and tested and provides a promising foundation for the future of the application [109].
25.5 CONCLUSIONS
Capillary electrophoresis technology has become an indispensable tool for forensic scientists in the
biology field since it is able to provide valuable information to aid in the process of law enforcement.
The primary application of the technique is in the qualitative analysis of STRs. Isolation of STR
mixtures is also possible using relative peak heights. Other applications of CE include quantitative
analysis of PCR products, mtDNA sequencing, and mutation detection for the analysis of plant and
bacterial DNA. Based on the performance of the methods illustrated above, it is reasonable to expect
future researchers and practitioners to continue working to exploit the capabilities of this robust
scientific technique and its application to criminal investigations.
ACKNOWLEDGMENTS
The authors would like to thank John Butler, Dee Mills, Alice Isenberg, Kate Theisen, Pete Vallone,
Ed Buse, Ada Nuñez, Kerry Opel, Ira Lurie, Stephanie King, Yin Shen, and George Duncan for their
contributions to this work. Major funding from the National Institute of Justice is also gratefully
acknowledged. Points of view in the document are those of the authors and do not necessarily
represent the official view of the U.S. Department of Justice.
REFERENCES
1. Tagliaro, F. and Smith, F., Forensic capillary electrophoresis, Trends Anal Chem, 15, 513, 1996.
2. Northrop, D., McCord, B., and Butler, J., Forensic applications of capillary electrophoresis, J Capillary
Electrophor, 1, 158, 1994.
3. Barron, A., Sunada, W., and Blanch, H., Capillary electrophoresis of DNA in uncrosslinked polymer
solutions: Evidence for a new mechanism of DNA separation, Biotechnol Bioeng, 52, 259, 1996.
4. Arnheim, N., Polymerase chain reaction strategy, Annu Rev Biochem, 61, 131, 1992.
5. Luckey, J. and Smith, L., Advances in DNA sequencing technology, Proc SPIE, 1891, 21, 1993.
6. Fregeau, C. and Fourney, R., DNA typing with fluorescently tagged short tandem repeats: A sensitive
and accurate approach to human identification, BioTechniques, 15, 100, 1993.
7. Cohen, A., Najarian, D., Paulus, A., et al., Rapid separation and purification of oligonucleotides by
high-performance capillary gel electrophoresis, Proc Natl Acad Sci USA, 85, 9660, 1988.
8. Kasper, T., Melera, M., Gozel, P., et al., Separation and detection of DNA by capillary electrophoresis,
J Chromatogr, 458, 303, 1988.
9. Albarghouthi, M., Buchholz, B., Huiberts, P., et al., Poly-N-hydroxyethylacrylamide (polyDuramide):

A novel, hydrophilic, self-coating polymer matrix for DNA sequencing by capillary electrophoresis,
10. Tseng, W. and Chang, H., A new strategy for optimizing sensitivity, speed and resolution in capillary
electrophoretic separation of DNA, Electrophoresis, 22, 763, 2001.
11. Mandabhushi, R., Separation of 4-color DNA sequencing extension products in noncovalently coated
capillaries using low viscosity polymer solutions, Electrophoresis, 19, 224, 1998.
12. Rocheleau, M., Grey, R., Chen, D., et al., Formamide modified polyacrylamide gels for DNA
sequencing by capillary gel electrophoresis, Electrophoresis, 13, 484, 1992.
13. Guttman, A. and Cooke, N., Denaturing capillary gel electrophoresis, Am Biotechnol Lab,
9, 10, 1991.
14. Brining, S., Chen, N., Yi, D., et al., Gel electrophoretic distinction between toxic and nontoxic forms
of beta-amyloid (1-40), Electrophoresis, 20, 1398, 1999.
15. Ozawa, S., Sugano, K., Sonehere, T., et al., High resolution for single-strand conformation
polymorphism analysis by capillary electrophoresis, Anal Chem, 76, 6122, 2004.
16. Butler, J., McCord, B., Jung, J., et al., Quantitation of polymerase chain reaction products by capillary
electrophoresis using laser fluorescence, J Chromatogr B, 658, 271, 1994.
17. Butler, J., Buel, E., Crivellente, F., et al., Forensic DNA typing by capillary electrophoresis using the
ABI Prism 310 and 3100 genetic analyzers for STR analysis, Electrophoresis, 25, 1397, 2004.
18. Wilson, M., Stoneking, M., Holland, M., et al., Guidelines for the use of mitochondrial DNA
sequencing in forensic science, Crime Lab Digest, 20, 68, 1993.
19. LaFountain, M., Schwartz, M., Svete, P., et al., TWGDAM validation of the AmpFlSTR profiler Plus
and AmpFlSTR Cofiler STR multiplex systems using capillary electrophoresis, J Forensic Sci, 46,
1191, 2001.
20. Hartzell, B., Graham, K., and McCord, B., Response of short tandem repeat systems to temperature
and sizing methods, Forensic Sci Int, 133, 228, 2003.
21. Nock, T., Dove, J., McCord, B., et al., Temperature and pH studies of short tandem repeat systems
using capillary electrophoresis at elevated pH, Electrophoresis, 22, 755, 2001.
22. Budowle, B., Chakraborty, R., Carmody, G., et al., Source attribution of a forensic DNA profile,
Forensic Sci Comm, 2, 1, 2000.
23. Moxon, E. and Wills, C., DNA microsatellites: Agents of evolution? Sci Am, 280, 94, 1999.
24. Butler, J., Forensic DNA Typing: Biology, Technology, and Genetics of STR Markers, 2nd edn.
Academic Press, San Diego, CA, 2005.
25. www.fbi.gov, accessed 12/2006.
26. Budowle, B., Moretti, T. R., Niezgoda, S. J., et al., CODIS and PCR-based short tandem repeat loci:
Law enforcement tools., Forensic Science Research and Training Center, FBI Academy, 73–77.
27. Ruitberg, C. M., Reeder, D. J., and Butler, J. M., STRBase: A short tandem repeat DNA database for
the human identity testing community, Nucleic Acids Res, 29, 320, 2001.
28. http://www.FBi.gov/hg/lab/codis/program.htm, accessed 8/21/07.
29. Cotton, E., Allsop, R., Guest, J., et al., Validation of the AMPFlSTR SGM plus system for use in
forensic casework, Forensic Sci Int, 112, 151, 2000.
30. Krenke, B., Tereba, A., Anderson, S., et al., Validation of a 16-locus fluorescent multiplex system,
J Forensic Sci, 47, 773, 2002.
31. Hobson, D., Smerick, J., Brown, A., et al., A summary of the validation of the STR analysis by
capillary electrophoresis: development of interpretation guidelines for the Profiler Plus and Cofiler
systems for use in forensic casework, in Proceedings from the 10th International Symposium on
Human Identification, Promega Corp, Coronoado Springs (1999).
32. Gill, P., Sparkes, R., Pinchin, R., et al., Interpreting simple STR mixtures using allele peak areas,
Forensic Sci Int, 91, 41, 1997.
33. Ladd, C., Lee, H., Yang, N., et al., Interpretation of complex forensic DNA mixtures, Croat Med J,
42, 244, 2001.
34. Gill, P., Whitaker, J., Flaxman, C., et al., An investigation of the rigor of interpretation rules for STRs
derived from less than 100 pg of DNA, Forensic Sci Int, 112, 17, 2000.
35. Gill, P., Jeffreys, A., and Werrett, D., Forensic application of DNA “fingerprints,” Nature, 318, 577,
1985.
36. Weir, B., The second national research council report on forensic DNA evidence, Am J Hum Genet,
59, 497, 1996.
37. Sajantila, A. and Budowle, B., Identification of individuals with DNA testing, Ann Med, 23, 637,
1991.
38. Fung, W. and Hu, Y., Interpreting DNA mixtures based on the NRC-II recommendation 4.1, Forensic
Sci Comm, 2, 1–11, 2000.
39. Weir, B., Triggs, C., Starling, L., et al., Interpreting DNA mixtures, J Forensic Sci, 42, 213, 1997.
40. Clayton, T., Whitaker, J., Sparkes, R., et al., Analysis and interpretation of mixed forensic stains using
DNA STR profiling, Forensic Sci Int, 91, 55, 1998.
41. Evett, I., Gill, P., and Lambert, J., Taking account of peak areas when interpreting mixed DNA profiles,
42. Ludes, B., Pfitzinger, H., and Mangin, P., DNA fingerprinting from tissues after variable postmortem
periods, J Forensic Sci, 38, 686, 1993.
43. Smith, S. and Morin, P., Optimal storage conditions for highly dilute DNA samples: A role for trehalose
as a preserving agent, J Forensic Sci, 50, 1101, 2005.
44. Whitaker, J., Clayton, T., Urquhart, A., et al., Short tandem repeat typing of bodies from a mass disaster:
High success rate and characteristic amplification patterns in highly degraded samples, BioTechniques,
18, 670, 1995.
45. Anderson, T., Ross, J., Roby, R., et al., A validation study for the extraction and analysis of DNA from
human nail material and its application to forensic casework, J Forensic Sci, 44, 1053, 1999.
46. Butler, J. M., Shen, Y., and McCord, B. R., The development of reduced size STR amplicons as tools
for analysis of degraded DNA, J Forensic Sci, 48, 1054, 2003.
47. Wiegand, P. and Kleiber, M., Less is more—Length reduction of STR amplicons using redesigned
primers, Int J Legal Med, 114, 285, 2001.
48. Tsukada, K., Takayanagi, K., Asamura, H., et al., Multiplex short tandem repeat typing in degraded
samples using newly designed primers for the TH01, TPOX, CSF1PO, and vWA loci, Legal Med, 4,
239, 2002.
49. Drabek, J., Chung, D., Butler, J., et al., Concordance study between miniplex assays and a commercial
STR typing kit, J Forensic Sci, 49, 859, 2004.
50. Coble, M. and Butler, J., Characterization of new miniSTR loci to aid analysis of degraded DNA,
51. Bender, K., Schneider, P., and Rittner, C., Application of mtDNA sequence analysis in forensic
casework for the identification of human remains, Forensic Sci Int, 113, 103, 2000.
52. Saccone, C., Gissi, C., Lanave, C., et al., Evolution of the mitochondrial genetic system: An overview,
Gene, 261, 153, 2000.
53. Anderson, S., Bankier, A., Barrell, B., et al., Sequence and organization of the human mitochondrial
genome, Nature, 290, 457, 1981.
54. Holland, M. and Parsons, T., Mitochondrial DNA sequence analysis—Validation and use for forensic
casework, Forensic Sci Rev, 11, 21, 1999.
55. Jazin, E., Soodyall, H., Jalonen, P., et al., Mitochondrial mutation rate revisited: Hot spots and
polymorphisms, Nat Genet, 18, 109, 1998.
56. Wilson, M., DiZinno, J., Polanskey, D., et al., Validation of mitochondrial DNAsequencing for forensic
casework analysis, Int J Legal Med, 108, 68, 1995.
57. Kline, M., Vallone, P., Redman, J., et al., Mitochondrial DNA typing screens with control region and
coding region SNPs, J Forensic Sci, 50, 377, 2005.
58. Isenberg, A. and Moore, J., Mitochondrial DNA analysis at the FBI laboratory, Forensic Sci
Comm, 1, 1999.
58. Cann, R., Stoneking, M., and Wilson, A., Mitochondrial DNA and human evolution, Nature, 325,
31–36 1987.
59. Baasner, A., Schafer, C., Juge, A., et al., Polymorphic sites in human mitochondrial DNA control
region sequences: Population data and maternal inheritance, Forensic Sci Int, 98, 169, 1998.
60. Budowle, B., Allard, M., Wilson, M., et al., Forensics and mitochondrial DNA: Applications, debates,
and foundations, Annu Rev Genom Hum G, 4, 119, 2003.
61. Stone, A., Starrs, J., and Stoneking, M., Mitochondrial DNA analysis of the presumptive remains of
Jesse James, J Forensic Sci, 46, 173, 2001.
62. vonWurmb-Schwark, N., Harbeck, M., Wiesbrock, U., et al., Extraction and amplification of nuclear
and mitochondrial DNA from ancient and artificially aged bones, Legal Med, 5, 169, 2003.
63. Roewer, L. and Epplen, J., Rapid and sensitive typing of forensic stains by PCR amplification of
polymorphic simple repeat sequences in case work, Forensic Sci Int, 53, 163, 1992.
64. Daniels, D., Hall, A., and Ballantyne, J., SWGDAM developmental validation of a 19-locus Y-STR
system for forensic casework, J Forensic Sci, 49, 1, 2004.
65. Hall, A. and Ballantyne, J., The development of an 18-locus Y-STR system for forensic casework,
Anal Bioanal Chem, 376, 1234, 2003.
66. Kayser, M. and Sajantila, A., Mutations at Y-STR loci: Implications for paternity testing and forensic
analysis, Forensic Sci Int, 118, 116, 2001.
67. Corach, D., Filgueira, R., Marino, M., et al., Routine Y-STR typing in forensic casework, Forensic Sci
Int, 118, 131, 2001.
68. Huffine, E., Crews, J., Kennedy, B., et al., Mass identification of persons missing from the break-up
of the former Yugoslavia: Structure, function, and role of the International Commission of Missing
Persons, Croat Med J, 42, 271, 2001.
69. Butler, J., Recent developments in Y-short tandem repeat and Y-single nucleotide polymorphism
analysis, Forensic Sci Rev, 15, 91, 2003.
70. Hedman, M., Pimenoff, V., Lukka, M., et al., Analysis of 16 Y STR loci in the Finnish population
reveals a local reduction in the diversity of male lineages, Forensic Sci Int, 142, 37, 2004.
71. Silva, D., Carvalho, E., Costa, G., et al., Y-chromosome genetic variation in Rio De Janeiro population,
Am J Hum Biol, 18, 829, 2006.
72. Cerri, N., Ricci, U., Sani, I., et al., Mixed stains from sexual assault cases: Autosomal or Y-chromosome
short tandem repeats? Croat Med J, 44, 289, 2003.
73. Hanson, E. and Ballantyne, J., A highly discriminating 21 locus Y-STR “megaplex” system designed
to augment the minimal haplotype loci for forensic casework, J Forensic Sci, 49, 40, 2004.
74. Budowle, B., SNP typing strategies, Forensic Sci Int, 146, 139, 2004.
75. Vallone, P., Decker, A., and Butler, J., Allele frequencies for 70 autosomal SNP loci with US Caucasian,
African American, and Hispanic samples, Forensic Sci Int, 149, 279, 2005.
76. Stoneking, M., Single nucleotide polymorphisms: From the evolutionary past, Nature, 409, 821, 2001.
77. Venter, J. C., Adams, M. D., Myers, E. W., et al., The sequence of the human genome, Science, 291,
1304, 2001.
78. Amorim, A. and Pereira, L., Pros and cons in the use of SNPs in forensic kinship investigation:
A comparative analysis with STRs, Forensic Sci Int, 150, 17, 2005.
79. Payseur, B. and Cutter, A., Integrating patterns of polymorphism at SNPs and STRs, Trends Genet,
22, 424, 2006.
80. Syvanen, A., Accessing genetic variation: Genotyping single nucleotide polymorphisms, Nat Rev
Genet, 2, 930, 2001.
81. Tian, H., Jaquins-Gerstl, A., Munro, N., et al., Single-strand conformation polymorphism analysis by
capillary and microchip electrophoresis: A fast, simple method for detection of common mutations in
BRCA1 and BRCA2, Genomics, 63, 25, 2000.
82. Atha, D., Kasprzak, W., O’Connell, C., et al., Prediction of DNA single-strand conformation polymor-
phism: Analysis by capillary electrophoresis and computerized DNA modeling, Nucleic Acids Res,
29, 4643, 2001.
83. Orita, M., Iwahana, H., Kanazawa, H., et al., Detection of polymorphisms of human DNA by gel
electrophoresis as single-strand conformation polymorphisms, Proc Natl Acad Sci USA, 86, 2766,
1989.
84. Fukuda, M. and Tamaki, Y., Subtyping of D11S488 STR alleles by single-strand conformation
polymorphism (SSCP) analysis in two cases of disputed parentage, Nihon Hoigaku Zasshi, 52, 42,
1998.
85. Tsai, L., Kao, L., Chang, J., et al., Rapid identification of the ABO genotypes by their single-strand
conformation polymorphism, Electrophoresis, 21, 537, 2000.
86. Sunnucks, P., Wilson, A., Beheregaray, L., et al., SSCP is not so difficult: The application and utility
of single-stranded conformation polymorphism in evolutionary biology and molecular ecology, Mol
Ecol, 9, 1699, 2000.
87. Day, A., Nonhuman DNA testing increases DNA’s power to identify and convict criminals, Silent
Witness, 6, 2001.
88. Hellmann, A., Rohleder, U., Eichmann, C., et al., A proposal for standardization in forensic canine
DNA typing: Allele nomenclature of six canine-specific STR loci, J Forensic Sci, 51, 274–281, 2006.
89. Eichmann, C., Berger, B., and Parson, W., Aproposed nomenclature for 15 canine-specific polymorphic
STR loci for forensic purposes, Int J Legal Med, 118, 249, 2004.
90. Sensabaugh, G. and Kaye, D., Non-human DNA evidence, Jurimetrics J, 38, 1, 1998.
91. Eichmann, C., Berger, B., Steinlechner, M., et al., Estimating the probability of identity in a random
dog population using 15 highly polymorphic canine STR markers, Forensic Sci Int, 151, 37, 2005.
92. Halverson, J. and Basten, C., A PCR multiplex and database for forensic DNA identification of dogs,
93. State of Washington v. Kenneth J. Leuluaialii and Geroge J. Tuilefano, in No. 43507-8-I consolidated
with No. 437-1-I, Court of Appeals of the State of Washington, 2003.
94. Butler, J., David, V., O’Brien, S., et al., The MeowPlex: A new DNA test using tetranucleotide STR
markers for the domestic cat, Profiles DNA, 5, 7, 2002.
95. www.appliedbiosystems.com, accessed 12/2006.
96. State of California v. David Alan Westerfield, in Case No. SCD165805, Superior Court of California,
County of San Diego Central Division, 2002.
97. Braunner, P., Reshef, A., and Gorski, A., DNA profiling of trace evidence—Mitigating evidence in a
dog biting case, J Forensic Sci, 46, 1232, 2001.
98. Menotti-Raymond, M., David, V., and O’Brien, S., Pet cat hair implicates murder suspect, Nature,
386, 774, 1997.
99. Giovambattista, G., Ripoli, M., Liron, J., et al., DNA typing in a cattle stealing case, J Forensic Sci,
46, 1484, 2001.
100. Gilmore, S., Peakall, R., and Robertson, J., Short tandem repeat (STR) DNA markers are hypervariable
and informative in Cannabis sativa: Implications for forensic investigations, Forensic Sci Int, 131, 65,
2003.
101. Hsieh, H., Hou, R., Tsai, L., et al., A highly polymorphic STR locus in Cannabis sativa, Forensic Sci
Int, 131, 53, 2003.
102. Yoon, C., Forensic science: Botanical witness for the prosecution, Science, 260, 894, 1993.
103. Vos, P., Hogers, R., Bleeker, M., et al., AFLP: A new technique for DNA fingerprinting, Nucleic Acids
Res, 23, 4407, 1995.
104. Coyle, H., Ladd, C., Palmbach, T., et al., The green revolution: Botanical contributions to forensics
and drug enforcement, Croat Med J, 42, 340, 2001.
105. Bless, C., Palmeter, H., and M.M., Identification of acer rubrum using amplified fragment length
polymorphism, J Forensic Sci, 51, 31, 2006.
106. Budowle, B., Genetics and attribution issues that confront the microbial forensics field, Forensic Sci
Int, 146, 185, 2004.
107. Horswell, J., Cordiner, S., Maas, E., et al., Forensic comparison of soils by bacterial community DNA
profiling, J Forensic Sci, 47, 350, 2002.
108. Moreno, L., Mills, D., Entry, J., et al., Microbial metagenome profiling using amplicon length
heterogeneity-polymerase chain reaction proves more effective than elemental analysis in discrim-
inating soil specimens, J Forensic Sci, 51, 2006.
109. Marsh, T., Terminal restriction fragment length polymorphism (T-RFLP): An emerging method
for characterizing diversity among homologous populations of amplification products, Curr Opin
Microbiol, 2, 323–327, 1999.
110. Clement, B., Kehl, L., DeBord, K., et al., Terminal restriction fragment patterns (TRFPs), a rapid,
PCR-based method for the comparison of complex bacterial communities, J Microbiol Meth, 31, 135,
1998.
111. Chan, O. C., Yang, X., Fu, Y., et al., 16S rRNA gene analyses of bacterial community struc-
tures in the soils of evergreen broad-leaved forests in south-west China, FEMS Microbiol Ecol, 58,
247, 2006.
112. Kvist, T., Ahring, B., Lasken, R., et al., Specific single-cell isolation and genomic amplification of
uncultured microorganisms, Appl Microbiol Biotechnol, 74, 926–935, 2006.
113. Mills, D., Fitzgerald, K., Litchfield, C., et al., A comparison of DNA profiling techniques
for monitoring nutrient impact on microbial community composition during bioremediation of
petroleum-contaminated soils, J Microbiol Meth, 54, 57, 2003.
114. Dunbar, J., Ticknor, L., and Kuske, C., Assessment of microbial diversity on four southwestern United
States soils by 16S rRNA gene terminal restriction fragment analysis, Appl Environ Microbiol, 66,
2943, 2000.
26 Clinical Application of CE
Zak K. Shihabi
CONTENTS
26.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 786

26.1.1 Special Aspects of CE in Clinical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 786
26.1.2 Advantages of CE in the Clinical Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 787
26.1.3 Limitations of CE in the Clinical Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 787
26.2 Practical Aspects of CE in Clinical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 787
26.2.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 787
26.2.2 Sample Matrix Effects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 787
26.2.3 High Salt Content in the Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 788
26.2.4 High Protein Content in the Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 788
26.2.5 Wide Variability in Concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 788
26.2.6 Precision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 789
26.2.7 Coated versus Noncoated Capillary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 789
26.2.8 Stacking of Compounds of Clinical Interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 789
26.2.9 Compounds Difficult to Analyze by CE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 790
26.2.10 Compounds Suited for Analysis by CE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 790
26.3 Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 791
26.3.1 Serum Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 791
26.3.2 Immunofixation (Immunosubtraction). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 791
26.3.3 Cryoglobulins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 792
26.3.4 Urinary Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 793
26.3.5 Cerebrospinal Fluid Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 793
26.4 Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 793
26.5 Hemoglobin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 797
26.5.1 Hemoglobin Variant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 797
26.5.2 Hemoglobin A1C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 797
26.5.3 Globin Chains of Hemoglobin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 799
26.6 Peptides and Polypeptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 799
26.7 CE and Proteomics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 799
26.7.1 Nucleic Acids (DNA). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 801
26.8 Small Molecule Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 801
26.8.1 Drug Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 801
26.8.2 Endogenous Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 801
26.8.3 Amino Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 801
26.8.4 Ion Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 802
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 805
785
26.1 INTRODUCTION
The movement of soil colloidal particles was the first description of electrophoresis as early as 1809.
However, Arne Tiselius (∼1937) was the first to construct a successful instrument useful for the
separation of serum protein by electrophoresis using the boundary separation principle. Because of
the clinical significance of this type of separation, many improvements and refinements followed,
such as utilizing paper, cellulose acetate, gel, and more recently capillaries in order to speed up and
better separate (into distinct zones) the different proteins. The electric current can be utilized in the
clinical applications to accomplish not just separation but other tasks:
1. It can be utilized to move fluids (through the electroosmotic flow, EOF). This feature is
very useful for analysis in the microchip and in capillary electrochromatography (CEC)
where the additions of pumps to move and mix fluids are not feasible.
2. It can be used to concentrate dilute compounds directly during the electrophoretic step
(through stacking) as it will be discussed later. This feature is very important for analysis
of clinical compounds present in biological fluids at very low concentration.
3. It can be used to separate, quantify, and identify different compounds in clinical samples
using different electrophoretic principles.
Capillary electrophoresis (CE) is a general analytical technique for separation and quantification
of a wide variety of molecules including those of clinical interest, utilizing narrow bore capillaries
under high voltage. It separates various compounds not only based on charge but also based on
size, hydrophobicity, and stereospecificity as discussed in Chapters 1, 2, 19. The flexibility of this
technique stems from its ability to incorporate easily in the separation buffer different additives, that
can interact with some of the analytes relative to others to alter their velocity so as to achieve the
desired separation. These additives give the CE great ability to separate numerous clinical compounds
in the same instrument using nonexpensive capillaries. Thus, most of the clinical tests can be adapted
to CE; however in practice, some are better suited than others for analysis by this method.
The clinical/biological field and pharmaceutical industry are the main areas that are benefiting
most from the application of the CE. Most of the clinical tests can be analyzed by CE as well by other
techniques, such as high-performance liquid chromatography (HPLC) or slab gel electrophoresis
(SGE). However, the CE offers certain advantages for clinical analyses notably a high plate number,
a characteristic that is useful when dealing with complex samples containing numerous compounds
such as serum or urine. The CE also offers rapid analysis time and a low cost per test with full
automation. These characteristics were the driving forces for adopting the CE for completing the
human genome sequence project.
However, the CE in clinical analysis requires more thoughtful planning to achieve a good sepa-
ration compared with other methods. For example, many steroids cannot be separated in free zone
electrophoresis. Many of these are neutral or weakly charged compounds and migrate with EOF.
They separate much better by the addition of a micellar compound, such as sodium dodecyl sulfate
(SDS) to achieve separation by micellar electrokinetic capillary chromatography (MEKC). The CE
has some challenges such as poor sensitivity of detection, problems with the sample matrix, and
adsorption to the capillary walls as it will be discussed later here and Chapter 13. However, with
careful planning these obstacles can be resolved. Many applications can fall under the umbrella
of clinical applications; however, tests that have routine applications in the clinical field will be
discussed more in detail in this chapter.
26.1.1 SPECIAL ASPECTS OF CE IN CLINICAL ANALYSIS

The clinical field is a very vast one encompassing molecules of different physical and chemical
characteristics. The sizes vary from small ions such as Na and K to very large ones such as protein
and DNA. It encompasses compounds that are very abundant in concentration such as albumin
Clinical Application of CE 787
(∼30,000 mg/L) to compounds very low in concentration such as prostate-specific antigen (∼4 µg/L).
It encompasses compounds with different solubility and hydrophobicity. Thus, the analysis of these
different compounds requires different strategies and represents different degrees of difficulties.
Some analyses work well and easily with capillary zone electrophoresis (CZE) (e.g., serum proteins),
size separation (e.g., DNA) while others are difficult to adapt to any form of CE. The lower the
concentration is the more difficult the analysis becomes.
26.1.2 ADVANTAGES OF CE IN THE CLINICAL FIELD

The main advantage of the CE in clinical analysis is the flexibility of the separation so that the same
instrument can be utilized to separate numerous biological compounds. It is easy to add different
additives to the separation buffer to induce a change in the velocity of some of the compounds leading
to a better separation. The principle of the separation can easily be changed from free solutions to
hydrophobicity separation (MEKC), size separation, chiral, isofocusing to suit a particular group
of compounds as described under Chapters 1, 2. The second important advantage is the high plate
number in CE generated in the capillary due to the flat profile, compared with the laminar flow in
HPLC, without dealing with the high pressure or the high cost of the HPLC packed column. A third
feature is the ability to perform separations without the need to use large volumes of organic solvents.
Organic solvents, which are used often in HPLC, are becoming more expensive to purchase and more
difficult, under many state laws, to store and dispose of. This eventually may compel shifting of
many separations from HPLC to the CE.
26.1.3 LIMITATIONS OF CE IN THE CLINICAL FIELD

On the other hand, CE suffers from a few problems. Unlike HPLC, CE is greatly affected by sample
matrix (i.e., salts and proteins) [1–4], which are very high in biological samples. Another major
problem in clinical analysis is the suboptimal detection sensitivity. The third problem is sample
interaction with the capillary walls. To utilize the CE successfully for practical separation of clinical
samples it is important to understand how these factors affect the CE and the different maneuvers
needed to overcome them.
26.2 PRACTICAL ASPECTS OF CE IN CLINICAL ANALYSIS

26.2.1 BACKGROUND
Many applications fall under the umbrella of clinical applications. For the sake of simplicity, these can
be divided into two general areas. One is the research, which is performed occasionally for gaining
basic scientific information. The other area is the routine analysis for patient care and diagnosis such
as detection of monoclonal gammopathies to detect specific tumors. This test is performed often. In
research, the need is for the flexibility and versatility with good separation with emphasis on detection
of compounds at low concentration. In routine work, the emphasis is more on good reproducibility
(precision) with minimum amount of analytical steps. The different forms of CE can fulfill both
goals. For example, the ability to add different additives to the buffer allows for great selectivity
and versatility, for example, from separation based on charge to separation based on hydrophobicity,
or to size. On the other hand, the MEKC form of the CE offers ability to simply inject the sample
directly on the capillary without any treatment.
26.2.2 SAMPLE MATRIX EFFECTS

Many compounds can be analyzed easily from pure standard solutions, tablets, or from samples with
clean matrix. However, the analysis of the same compounds from biological sources poses many
more difficulties. This is because of the effects of both the high salt and high protein content of the
biological samples on the separation in CE.
Although the sample, in most instances, constitutes a very small portion of the overall volume in
the capillary once injected (<1%), the matrix of the sample has profound effects in CE. This is due
to two main factors: contribution of the sample to the total current conductance and also due to its
interaction with the capillary walls (especially proteins). On the basis of how the sample is prepared
and how the separation buffer is selected, sample matrix effects can be either favorable or detrimental
to the analysis [5]. Peak shape, separation, quantification, and reproducibility are affected greatly by
sample matrix.
A simple, but limited, solution to overcome the problems of high salt and protein content is to
dilute the sample. However, this not only lowers the salt and the protein content but also decreases
the concentration of the analytes of interest too!
26.2.3 HIGH SALT CONTENT IN THE SAMPLE

Conductivity differences between the analyte zones and the pure background electrolyte (BGE)
zones can lead to local electric field strength differences that, in turn, can distort the shape of the
analyte bands yielding asymmetric peaks and result in reduced separation efficiencies [1,3,5]. Sample
desalting is not easy to perform. If it is necessary, this can be accomplished through special columns
or after sample extraction. However, samples with high salt content can be analyzed directly provided
the separation buffers have high ionic strength [3]. These aspects will be discussed more in detail
later on.
26.2.4 HIGH PROTEIN CONTENT IN THE SAMPLE

High protein content in the sample can mask the absorption of the compounds of interest especially
those with low concentration. Proteins can bind to many molecules and alter their migration. They
can also cause peaks asymmetry [1,5]. However, more importantly they adsorb to the capillary walls
and change the zeta potential and the characteristics of the capillary. Notably, the reproducibility
suffers greatly from protein adsorption. High protein content in the sample eventually ruins the
capillary because of their adsorption to the walls. Since proteins tend to adsorb more on the capillary
inlet, cutting off a few millimeters of an old capillary inlet sometimes can restore back the separation.
Excess protein such as in serum samples can be removed by deproteinization before the CE step. If
the molecule of interest is a small to medium in size (<5000 Da) acetonitrile (2 V acetonitrile: 1 V of
serum) is an effective means to remove the protein while providing ability to concentrate (stacking)
the analyte as described below.
If the molecule of interest is neutral or weakly charged then the sample can be analyzed directly
without removing the protein or any treatment using MEKC. Analysis by MEKC can tolerate proteins
since the surfactants in the MEKC solubilize them.
26.2.5 WIDE VARIABILITY IN CONCENTRATION

The recent interest in proteomics placed great emphasis on analyzing proteins and peptides that are
present in very low concentration. Serum and many biological fluids contain numerous different
proteins (∼10,000) the majority of these are in very low concentration. However, at the same time,
very few proteins (mainly albumin, globulins, and transferrin) are present in these fluids in very
large concentrations. If these abundant proteins are not removed first, they make the analysis very
difficult because they overload the capillary. Several methods can be used to remove these proteins
and enrich the ones of interest; some are commercially available as kits. In general, several separation
or binding steps are used to remove or enrich these proteins (e.g., size exclusion, anionic exchange
chromatography, hydrophobic chromatography, and solid-phase ligands binding) [6,7]. For analysis
of small molecules, sample deproteinization is necessary but easy to perform. This removes the
proteins but leaves the small molecules in the supernatant.
26.2.6 PRECISION
In general, but not in all cases, the HPLC tends to give slightly better reproducibility than the CE.
The reproducibility in CE depends greatly on the capillary surface and to some extent on the number
of steps for sample preparation. A thorough wash with diluted sodium hydroxide or phosphoric
acid removes the adsorbed proteins from the capillary surface and improves the reproducibility.
Furthermore, the reproducibility improves by employing internal standards, peak area rather than
peak height, effective mobility rather than migration time, and frequent calibration [8]. The MEKC
can decrease or eliminate the extra steps for sample preparation. It allows direct injection of serum
on the capillary thus indirectly improving precision.
26.2.7 COATED VERSUS NONCOATED CAPILLARY

A noncoated (untreated) capillary is sufficient for most routine CE analysis, especially when it is
used with high ionic strength buffers. It is less expensive and offers good precision. However, for the
analysis of proteins and peptides present in low concentrations, a coated capillary gives a much better
theoretical plate number for proteins [9]. This is true of the capillaries and also for microchips [10].
Such capillaries are also important for performing isofocusing and isotachophoresis (ITP) to elimi-
nate the EOF and are more suitable for the detecting microheterogenity. The coating can be covalently
(permanently) bound or hydrodynamically by adding the coating material to the buffer. The walls
can be modified to possess either positive or negative charges or neutral. Linear polyacrylamide
was the first polymer used for permanent coating but it was not very stable. Other coatings were
prepared with other polymers such as dextran, fluorinated aromatic hydrocarbons, and hexamethyl
disiloxane poly(dimethyl acrylamide). Some of the covalently coated capillaries have a very short
life. Several dynamic coatings have been described, for example, Polybrene-poly(vinyl sulfonate),
Polybrene, poly(methoxyethoxyethyl)-ethylenimine, poly(diallyldimethylammonium chloride), and
starch derivatives [11–13]; some of these are commercially available as kits. An interesting dynamic
coating (adsorbed surfactant) was described based on washing the capillary with special surfac-
tant such as dimethylditetradecylammonium bromide and didodecyldimethylammonium bromide.
The coating binds and remains tightly bound enough for next several samples eliminating analytes
adsorption [14].
26.2.8 STACKING OF COMPOUNDS OF CLINICAL INTEREST

Because of the need for better sensitivity of detection in CE, sample concentration is crucial for
the widespread and practical use of this technique. Sample concentration can be accomplished by
several physical means outside the capillary, such as by liquid- and solid-phase extraction; but more
easily in CE by concentration on the capillary directly (stacking). Stacking is a general term referring
to the contiguous concentrated zones resembling a “stack of coins.”
Several methods for staking including the theoretical aspects have been described [15–18]. Few of
these methods are general while others are specific and more suitable for certain types of compounds.
In most of the stacking techniques, discontinuous buffers of different kinds are employed as the basic
means for altering the charge of the analytes or the field strength of the sample zone [15]. Thus,
the same analyte molecules present at the different edges of the band move at different velocities
in such a way that leads to sample concentration. Buffer discontinuity can be brought about simply
by altering the sample conductivity or pH so as to be different from that of the separation buffer.
Sometimes this can occur unintentionally [16].
Two general types of stacking are used often in CE for clinical compounds. The first is the
high field strength stacking. In this method, the sample is diluted in the same separation buffer
but at 10 times dilution. Thus, the sample molecules are subjected to higher field strength than
that of the electrophoresis buffer. Thus, they accelerate in this region but slow at the interface
of the sample and the buffer. This method can be used for large and small molecules. The second
type is acetonitrile stacking. Mixing acetonitrile with the sample (2:1 v/v) is used mainly to remove
proteins. However, the presence of acetonitrile in the sample (not in the buffer), has several important
additional advantages. For example, (1) it counteracts the deleterious effects of salts, (2) it yields
better stacking for small molecules than that obtained in dilute buffers, and (3) it allows larger
volumes of sample (in some cases half of the capillary volume) to be injected. The overall effect is
an increased sensitivity of small and medium molecules (<5000 Da). The stacking occurs in both
the hydrodynamic and the electromigration injection. The sodium chloride present in serum at about
150 mmol/L together with the acetonitrile used in the deproteinization both bring about 10–30 times
sample concentration. The mechanism of the stacking is pseudotransient ITP [19]. Thus, biological
samples having high salt and protein content are well suited to CZE analysis after treatment with
acetonitrile.
ITP [15], especially transient ITP, are used for small and large molecules stacking. In this tech-
nique, the sample is mixed with appropriate leading and terminating ions [16] as described in
Chapter 13. This technique is used to selectively enrich low-abundance compounds (e.g., peptides
in the absence of any bioaffinity interactions) [20]. This method is more difficult in practice, but it
offers the benefits of speed and very high concentration.
Neutral molecules analyzed by MEKC require special and more difficult stacking methods [18].
They concentrate based on stacking either by having regions of difference in the electric field or by
sweeping, that is, picking and accumulating of analytes by the pseudostationary phase that penetrates
the sample [18].
26.2.9 COMPOUNDS DIFFICULT TO ANALYZE BY CE

In spite of few attempts, glycoproteins, carbohydrates, and lipids are difficult to analyze by CE.
Carbohydrates and glycoproteins are important in the immunological recognition, control, and attack
of pathogens and in protein folding. Carbohydrates do not have strong absorption in the UV and
most do not have strong charges. They are analyzed at very high pH and require labeling mostly
with fluorescent reagents. The analysis of carbohydrates by CE-MS has been reviewed [21].
Lipids are important since they are part of the cell membrane and the lipoproteins have important
diagnostic values in coronary heart disease. Lipids too do not have strong absorption and lack strong
charges; in addition to that, they are hydrophobic. Lipoproteins have been analyzed by MEKC
[22]. CZE was used to separate the isoforms of low-density lipoprotein (LDL) particles in human
serum based on charge/volume ratios of the particles. LDL, dense LDL, were analyzed by CE [23].
A CE-MS method incorporating SDS for separating very-low-density lipoprotein (VLDL), LDL,
high-density lipoprotein (HDL) has been described [24].
26.2.10 COMPOUNDS SUITED FOR ANALYSIS BY CE

Some of the clinical compounds are better suited than others for separation by this technique. In
general, polar compounds possessing a strong charge and present in high concentration are most
suited to analysis by CE (e.g., proteins, hemoglobins and their chains, some peptides, and drugs).
Enzymes and DNA are also suited for analysis by CE even though they are not present in high
concentration. The DNA is amplified first and then analyzed elegantly by CE using sensitive detection
fluorescent dyes while enzymes are analyzed through the accumulation of products (catalytic activity
on the substrates). Following sections discusses those clinical compounds that have been well studied
by CE and analyzed often for patient care.
26.3 PROTEINS
Proteins are important not only for their function but also for their diagnostic significance. Proteins
maintain structural integrity and perform different functions such as catalytic, hormonal, or receptors.
The function of many of the proteins of clinical interest remains not well understood nevertheless
these proteins remain to be important. The CE is useful for quantification, purity check, and detection
of microheterogeneity of proteins.
Proteins and peptides are composed of amino acids, which are zwitterions carrying both positive
and negative charges. Thus proteins possess different isoelectric points. These proteins acquire a
charge based on the buffer they are dissolved in and move toward the opposite electrode depending
on the net charge. Usually, slightly alkaline buffers (pH > 8.0) are chosen for separation of proteins.
Under these conditions, the majority of proteins are negatively charged moving toward the anode but
they are pulled toward the cathode by the EOF. They can be also separated at acidic buffer conditions
(pH < 3.0) (i.e., carrying a positive charge). However, because of the absence of the EOF at this
pH, the separation requires a much longer time. Different proteins tend to behave differently in CZE
depending on their net charge. For example, basic proteins tend to bind to the capillary wall and give
distorted peak shape especially when they are in low concentration. To improve their separation by
CE, different additives, high salts, or coated capillaries are used to decrease the binding to the walls.
The advantages of CE over agarose gel (AG) electrophoresis for analysis of proteins are the
speed, automation, small sample volume, and avoidance of many staining/destaining steps. This led
some companies to design instruments dedicated only to protein analysis by CE. Furthermore, other
manufacturers designed special CE instruments to perform capillary isoelectric focusing (CIEF)
with absorption imaging detectors in order to focus, concentrate, and separate better the different
proteins as discussed in Chapter 19. These instruments can detect protein microheterogeneity better
than the common CZE instruments. Here are some clinical applications of proteins and peptides
measurement by CE.
26.3.1 SERUM PROTEINS

The interest in serum proteins stems from their diagnostic significance. Few thousands of different
proteins are present in the serum arising from the different cells and tissues; however, mostly at very
low level—way below the CE detection limits. A few of these proteins are present at high enough
concentration to be detected directly by CE. These proteins separate into a few bands (∼5–12), which
are not pure or single proteins but a group of several proteins. Serum proteins are analyzed routinely
on a daily basis in most large hospitals to detect several disorders such as renal failure and infections;
but most importantly, monoclonal gammopathies.
In CZE, serum proteins have been separated using different buffers (e.g., Tris and Tricine), but
mostly borate, with pH of 8–11 [22,25–29]. Serum protein separation can be completed by CE in
about 2–10 min in contrast to 1–2 h for agarose electrophoresis (AG) (Figure 26.1). The correlation
coefficient between CE and AG for the separated bands is good [30–32]. Some commercial instru-
ments use multicapillaries of narrow diameter (25 µm) to increase the throughput of the analysis. The
narrow capillaries produce better resolution than the wider capillaries with a much shorter migration
time [28]. This is true for all CE separations.
26.3.2 IMMUNOFIXATION (IMMUNOSUBTRACTION)

Serum immunoglobulins are composed of heavy and light chains and classified based on their
reaction with specific antibodies into the classes of IgG, IgA, IgM, IgD, and IgE. These can be
secreted in high concentration as some of clones of cells become malignant. Multiple myeloma,
Waldenstrom’s disease, and light chain disease all represents different malignancies of plasma cells
and can cause increased levels of these proteins (paraproteins). The identification of the type of
(a)
0.005
I
A
Absorbance
C T a2 a1 p
N a2 a1
0
0 Min 10
(b)
G
C T a2 a1
FIGURE 26.1 Comparison of the CZE to agarose gel (AG) electrophoresis for a patient with a streptococcal
infection: (a) CZE and (b) AG. (I: internal standard; P, prealbumin; A, albumin; a1, α1 globulin; a2, α2
globulin; T, transferrin; C, complement; G, γ -globulins; N, neutral compounds). Arrows in the γ -region indicate
oligobands. (From Shihabi, Z., Electrophoresis, 17, 1607, 1996. With permission.)
these different paraproteins is important for the proper patient treatment. These paraproteins are
detected and classified with immunofixation, a laborious procedure performed in gel electrophoresis.
Immunofixation is an important part [32] of working up an unknown monoclonal band. CZE has
been adapted to perform the immunofixation method based on reacting serum proteins with specific
antibodies. The Antigen-antibody complex can be separated from the free antibody or antibodies
based on binding to a solid matrix or simple separation in the capillary. The sample is assayed before
and after the binding [30,32]. The difference between the two “immunosubtraction” represents the
specific type of the monoclonal abnormal serum protein. This simple method has been shown to
be reliable.
26.3.3 CRYOGLOBULINS
Cryoglobulins analysis is an important but unfortunately a neglected test in clinical practice because
of the difficulty of the test. Cryoglobulins are special type of immunoglobulins that reversibly pre-
cipitate from serum at cold temperatures [33]. They can be associated with skin lesions (purpura),
glomerulonephritis, and peripheral neuropathy, or malignancy [8]. These are divided into two general
categories. First, a small fraction (about 5–10%) that are pure monoclonal antibodies (type I) and
which representing malignancies. The second type, which is the majority ∼95%, are immune
complexes (mixed cryoglobulins) representing immune stimulation due to infection or presence
of autoantibodies. The latter group is further divided based on if the cryoprecipitate contains a mon-
oclonal rheumatoid factor (type II) or polyclonal rheumatoid factor (type III) immunoglobulins [33].
The mixed cryoglobulins can precipitate in the different tissues of the body such as the kidney and
the extremities causing vasculitis.
Cryoglobulins are detected, quantified, and phenotyped all simultaneously by precipitating an
aliquot of the serum at 4◦ C, centrifuging, and dissolving the precipitate in a buffer followed by elec-
trophoresis using the same conditions as those for serum proteins (Figure 26.2) [34]. Cryoglobulins
are well suited for analysis by CE. The main advantages are the higher sensitivity, the use of small
volumes of serum, speed (15 min vs. 2 h), and improved quantification compared to the AG method.
We have used the CE for routine cryoglobulins analysis for our patients for over a decade with good
success [21].
26.3.4 URINARY PROTEINS

Usually urine proteins are present at about 10–100 times lower concentrations compared with serum
but in the presence of numerous interfering UV-absorbing compounds. This renders urinary proteins
to be more difficult to measure directly by CE when compared with serum. The majority of urine
samples require concentration (based on the protein content) before the CE analysis. They are con-
centrated through special commercial “membrane concentrators” followed by washing with saline
solution or using chromatographic column [32] to decrease the interfering UV-absorbing materials.
The urine contains several proteins of clinical interest, especially Bence-Jones proteins, which are
important for detecting plasma cell malignancies and for diagnosis of nephropathy. Very few urine
samples that are very high in proteins can be analyzed directly without any preparation. The same
buffers and conditions for serum proteins are basically used for the analysis of urine protein [35].
26.3.5 CEREBROSPINAL FLUID PROTEINS

The main clinical significance of cerebrospinal fluid (CSF) protein electrophoresis is for the detection
of the oligoclonal bands, which are present in multiple sclerosis in the gamma region. Similar to
urine, proteins in the CSF are present fluid in very low concentration (100 times less than serum).
For the majority of the samples, a 10- to 20-fold concentration is preferred before analysis by CE
(by the same membrane concentrators used for urine). CSF protein separation can be accomplished
in less than 10 min with CE versus 2 h for AG with the ability to detect oligoclonal banding by this
technique [36].
26.4 ENZYMES
Since enzymes are essentially proteins, they can be measured in CE by direct light absorbency like
other proteins or by their enzymatic activity. The CE offers versatility for enzyme measurement:
the enzyme itself, the substrates, or the products (Figure 26.3) all can be measured in CE. Most
of the enzymes in biological fluids are present in very low concentration so that they cannot be
determined by direct light absorbency. The catalytic activity is much more sensitive and more
versatile. Catalytic activity is more suited for enzymes with low activity because the reaction product
can be amplified with time. The catalytic activity can be measured in several ways: (A) incubation in
the capillary that is used as a microreactor [37–39]; (B) online, postcapillary reaction (more difficult);
and (C) incubation outside the capillary that is more common. If a long incubation step is needed then
it is more convenient to perform the incubation outside the instrument. After the incubation step, the
reaction is stopped preferably by the addition of acetonitrile (rather than acid) (Figure 26.3), which
(a)
0.006
A
0
0 10
(b)
0.015
A
Absorbance
0 10
(c)
A
G M
Minutes
FIGURE 26.2 Serum electrophoresis of a patient with type II cryoglobulins (650 mg/L): (a) serum CZE,
(b) Cryoglobulins by CZE, and (c) cryoglobulins by AG (see legend of Figure 26.1). (From Shihabi, Z.,
(a)
10
0
mAUFS
1 2 3 4 5 6 7 8 9 10 11
(b)
10
1 2 3 4 5 6 7 8 9 10 11
Minutes
FIGURE 26.3 Glutathione transferase activity: rat heart tissue homogenate 25 µL were mixed with 50 µL
of 1-Cl-2-dinitrobenzene (0.2 mg/mL, pH 6.6) and 25 µL of reduced glutathione (2 mg/mL) and incubated for
10 min at 37◦ C. The reaction is stopped with 200 µL acetonitrile, mixed, and centrifuged. (a) At 0 min and
(b) at 10 min of incubation (P, product of the reaction; G, glutathione peak). (Separation buffer: 250 mM borate,
50 mM Tris, pH 8.0 at 10 kV, 214 nm, 20 s injection on a capillary 30 cm × 50 µm (ID).
can remove the proteins, decreasing the UV absorption, and concentrate the analyte (by stacking) at
the same time.
Proteolytic enzymes and those enzymes with low activity or with expensive substrates are
well suited for analysis by CE in this manner. The proteolytic enzymes can be detected simply
based on their absorption in the UV range in CE. Several enzymes have been analyzed by CE,
such as chloramphenicol acetyl transferase [40], glutathione peroxidase, glutathione transferase
0.02
(a)
0.02
I (b)
P
Absorbance
0.02
(c)
(d)
0.02
0
0 2 4 6
Min
FIGURE 26.4 Enzymatic activity of breast tumor homogenate activity for Cathepsin D (106 pmol/mg protein)
at different periods of incubation: (a) 0 min; (b) 10 min; (c) 20 min; and (d) at 20 min in the presence of pepstatin.
(P, split peptide; I, Iothalamic acid). (From Shihabi, Z.K. and Kute, T.E., J. Chromatogr. B, 683, 125, 1996.
With permission.)
(Figure 26.3), ornithine transcarbamylase [40], angiotension converting enzyme [41], Cathepsin D
(Figure 26.4) [42], and elastase [43].
The advantages of CE for analysis of enzymes are the use of small volumes, versatility, and ability
to avoid the extra steps of indicator reactions. We found in some instances, for example, analysis of
glutathione transferase it is easier to assign enzymatic activity units (IU) based on the CE because both
the substrate and the products can be monitored at the same wavelength. Unfortunately, researchers
did not take full advantage of the CE for enzyme analysis. In practice, kinetic spectrophotometric
methods remain to be most widely used for routine work while the CE is reserved for those difficult
and specialized tests.
26.5 HEMOGLOBIN
26.5.1 HEMOGLOBIN VARIANT
Hemoglobin (Hb) carries the vital oxygen to the different tissues. Its concentration is used clinically
as an indicator of the different types of anemia. Hemoglobin is a tetramer composed of 2α chains and
2β chains. The β chain is more susceptible than the α chain for amino acid substitution (mutations),
which results in different variants often present in special populations. Most of the variants are
harmless and of research interest. However, few such as Hb S are associated with severe anemia,
decreased capacity to carry oxygen, and altered red blood cell shape. The most encountered variants
of hemoglobin are A, F, S, C, and E. Hemoglobin electrophoresis is carried out for two purposes:
detection of Hb variants and detect the presence of thalassemias (decreased synthesis of one of the
Hb chains).
Because of the small charge difference of the isoelectric point (pI), Hb variants do not separate
well by CZE. For good separation of hemoglobin variants by CZE, a high buffer concentration, a
narrow capillary (20–30 µm) (i.d.), and coated (hydrodynamically treated) capillary are chosen. Tris,
Tricine, and arginine buffers at pH 8–8.4 give a good separation [27,44,45] (Figure 26.5), which
resembles very closely to that of the alkaline separation by AG.
Although CE instruments are not well designed for CIEF, many variants can be separated better
by this technique. In addition to the common variants, G Philadelphia, A2, and Bart’s can all be
separated by CIEF [46–48]. The separation by CIEF compares well with gel isoelectric focusing
and with HPLC. The variants have also been analyzed by both CE and CIEF equipped with special
absorption imaging detectors. These types of detection devices eliminate the extra steps needed
to move the peaks, after the focusing step, to the detector and can simultaneously detect several
capillaries with better precision and faster results than CE instruments. HB A2 , which is increased
in β-thal, is better quantified by CE compared with AG electrophoresis [44]. Hemoglobin analysis
by CE has been reviewed recently [49–51].
26.5.2 HEMOGLOBIN A1C

Hemoglobin HbA1C is very important clinically because it is used to follow the control of blood
glucose over extended periods of time in patients with diabetes. It measures the amount of glucose
bound to hemoglobin (glycated Hb). Unlike blood glucose, HbA1C is more stable and represents
the average glucose in the past 3 months (the average life of the red blood cell). Clinical labs
perform this test routinely on a daily basis because of the widespread of incidence of diabetes
and the need to modify the glucose level or modify the patient treatment. A dedicated HPLC
instrument solely for this test is sold commercially. However, the cost/test is relatively expen-
sive. For analysis of this hemoglobin by CE, a coated capillary either dynamically [52,53] or
permanently is required [54]. Both CZE [52,53] and isoelectric focusing [54,55] are used for its
analysis.
15
A
(a)
F
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
30
(b)
Absorbance (mA)
b
s
1
2
C
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
13
(c)
B
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Min
FIGURE 26.5 Separation of a mixture of the common Hb variants using the same diluted sample and same
CE conditions for the intact Hb molecules and the HB chains (4 s injection, 214 nm). (a) Intact Hb molecule;
(b) Hb chains under acidic conditions; and (c) Hb chains under basic conditions. Peak 1, albumin; peak 2,
unknown. (From Shihabi, Z.K. and Hinsdale, M.E., Electrophoresis, 26, 581, 2005. With permission.)
26.5.3 GLOBIN CHAINS OF HEMOGLOBIN

The globin chains are useful for investigating the presence of thalassemias that lead to decreased
globin synthesis (either α or β chain). This could be minor (without much clinical consequences) or
major (severe or life threatening) symptoms, depending on the number of genes involved. The devi-
ation of the ratio of α/β in the patient from that of the normal (∼1:1) indicates the presence of
thalassemia. The thalassemia represents a decrease in the synthesis of either α or β chains. The
α-thalassemia is common in people of the African origin while the β-thalassemia is common in
those of Mediterranean sea origin. These chains are analyzed by cellulose acetate, which is a labo-
rious technique requiring extraction with acetone, denaturation by high concentration of urea, and
strip staining. Few CE methods used similar protocol to the cellulose acetate methods with analysis
in phosphate buffer either at pH 11.8 or at pH 2.5–4.5 [56,57]. We simplified the CE by using direct
Hb hemolysates without extraction or denaturation simply by analyzing the diluted red blood cells
at high alkaline pH 12.6 [58] or low acidic pH 2.15 (Figure 26.5). These buffers with high or low pH
induce the chain denaturation and at the same time separate the heme and serum proteins from the
chains. Thus method is extremely simple and very rapid since it does not require sample treatment,
staining or destaining. Furthermore, the common variants of the β chains, such as βS, βC, and βE,
are also separated from each other. Thus, this method gives further confirmation of the Hb variants
separation by CZE.
26.6 PEPTIDES AND POLYPEPTIDES

Peptides and polypeptides are similar to protein in structure but smaller in size. Many of the peptides
are biologically active compounds with different functions, such as adrenocorticotropic hormone
(ACTH), angiotension, substance P, or glutathione. Most of these peptides are present at very low
concentration. However, peptides can also arise from the digestion of purified proteins in vitro
especially as a step during protein characterization and can be present in high concentration.
After sequencing the human genome, there is a great interest in peptide analysis as a means to
identify those proteins coded by the different genes discovered recently. Peptide analysis by CE can be
used for quality control or purity check in the pharmaceutical industry. In this respect, the CE is well
suited for this purpose. After protein hydrolysis, the different peptides can be separated by HPLC or
CE and analyzed by the mass spectra. On the other hand, peptides can be present naturally in different
biological fluids such as in serum or spinal fluid. In most of the biological fluids, they are usually
present in low concentration among high concentration of interfering substances. Thus, the analysis
becomes much more difficult. Sample concentration before or on the capillary becomes vital for the
analysis. We used acetonitrile to concentrate some peptides, for example, glutathione, insulin chains
(Figure 26.6), and enkaphlin [59,60]. However, for those peptides in very low concentration they
require much more concentration either by other types of stacking [61] or by chromatography [62].
For better analysis, coated capillaries are preferred [9,10]. Analysis of peptide has been reviewed
recently [63,64].
26.7 CE AND PROTEOMICS

Proteins are synthesized as a result of the transcription and translation steps of the information stored
in the DNA. However, because of post-translational modification, during development and disease,
proteins primary structure can be modified greatly and so the function. In cells or tissues, there are
usually a large number of proteins that vary widely in their physicochemical properties, including
molecular weight, pI, solubility, and folding. Many factors that affect the synthesis can result in
many subtle variations, for example, due to sugar binding or SH interaction, which can account for
the wide variations between different individuals.
(a)
0.004
0.002 E
0.000 I
A
C
–0.002
Absorbance
(b) 0.055
E
0.040
0.020
A
C
0.000
0.0 5.0 10.0

Min
FIGURE 26.6 Effect of stacking for some natural peptides on sensitivity of detection (A, angiotensin; I, insulin
B chain; C, impurity in the insulin B chain; and E, Leu-enkephalin; (a) at 1.5% loading of the capillary and
(b) at 30% (stacking in acetonitrile). (From Shihabi, Z.K., J. Chromatogr. A, 1996, 744, 231. With permission.)
At the present time, proteomics is an area far from the routine work; however, this probably
will change in the near future because of the importance and the accelerated research. We have
witnessed the rapid change from research to routine work for DNA analysis. DNA fragments now
are routinely analyzed for viral detection; and for leukemia’s typing. Finding proteins (biomarkers)
that play key roles in the development, malfunctions of cells or tissues is a helpful for early detection
and diagnosis of diseases as well as searching for new targets for production of new drugs. There is
a great interest in identification and quantification of multiple proteins, including those in very low
concentration that constitute or control a particular biological process. Thus, techniques providing
high sensitivity and high peak capacity are still greatly demanded for the analysis of biological
samples.
At the present time, most of the studies for proteomics employ 2D gel or HPLC-MS. The CE-MS
has some promise in this area because of the speed and high plate number [65]. However, Huang
et al. [66] have summarized some of the problems involved with separation of protein by CE as
part of the proteomics “poor reproducibility, low-sample loading capacity, and low throughput due
to ineffective interfaces between the separation and MS systems.” Size separation is often used
in protein characterization in gel electrophoresis. This separation works very well in CE for the
DNA strands; however, it is more difficult for serum proteins. It is based on adding polymers such
as different cellulose derivatives, dextrans, or linear polyacrylamide to the buffers to retard the
migration of the large peptides. The proteins have to be denatured and carry the same charge (in
SDS) [67]. Several companies offer special kits suitable for this purpose.
Several pure proteins have been purified and checked for their purity, microheterogeneity, or
diagnostic significance by CE. In this case, coated capillaries are preferred for this separation.
Many of the proteins can be analyzed by CE; however, sensitive detectors such as fluorescence are
necessary [68–70]. Examples of proteins studied by CE separately or in a profile are transferrin
isoforms, which are important as markers of alcoholism [71], α-1 antitrypsin [72], recombinant
human erythropoietin glycoforms that stimulates erythopiosis [73], plasminogen tissue activator
[74], prions [75], urothelial carcinoma proteins in urine [76,77], and numerous urinary proteins [78].
26.7.1 NUCLEIC ACIDS (DNA)

This is an area that represents the best application of CE in the clinical/biological field as discussed
in Chapters 12, 16 and 25.
26.8 SMALL MOLECULE ANALYSIS

26.8.1 DRUG ANALYSIS
Therapeutic drug monitoring is an expanding area for both the clinical and the pharmaceutical
industry as discussed in Chapter 4.
26.8.2 ENDOGENOUS COMPOUNDS

Many metabolites both ionizable and nonionizable can be measured by different forms of CE such
as CZE, MEKC, and chirality. Metabolites with strong UV absorption such as nucleotides, phenolic
amino acids, and their metabolites are easy to measure by CE. However, some of these require
concentration and clean up before the CE step [5]. This can be achieved by traditional concentrating
methods, such as solid phase and solvent extraction; or by concentration on the capillary (stacking).
Examples of small molecules that have been analyzed by CE are nucleotides [79,80], amino acids,
catecholamines [81–83], and sugars [84,85]. Below is a more detailed discussion of some of these
compounds.
26.8.3 AMINO ACIDS

Amino acid analysis can be performed for detecting all the amino acids or for few specific ones.
Regardless of the type of method, analysis of amino acids is a difficult task because the majority
of the amino acids lack a strong chromophore and they resemble each other in structure, which
makes the separation very difficult. Thus, they need extra reaction steps to be derivatized pre- or
postseparation. Furthermore, physiological fluids like serum contain many interfering compounds
such as peptides and the uncommon amino acids. Amino acid analysis can be requested for a wide
variety of purposes such as clinical (diagnostic), nutritional, or for basic structure determination.
1. Specific amino acids: Analysis of a few or specific amino acids is useful and easy to
perform. This often requested in the detection of inborn errors of metabolism (phenylke-
tonuria or maple syrup disease). In these cases, the abnormal amino acids are present
in high concentration that makes the analysis simple. Several amino acids have been
determined by CE for this purpose, such as tyrosine, proline, and phenylalanine.
2. Amino acids arising from protein hydrolysate: This analysis is used often in food industry
as well in basic biochemistry. It is important in determination of the structure of protein
and in the assessment of the nutritional value of different proteins. Here the analysis of
about 20 amino acids is sufficient. This type of analysis is relatively more difficult than
that for a single amino acid.
3. Quantification of free amino acids in serum, urine, or other physiological fluids: The
separation here is complicated by the presence of many interfering substances, such as
small peptides and uncommon amino acids. This analysis is requested often for screening
purposes.
Traditionally amino acid analysis is carried out by specialized column chromatography instru-
ments (amino acid analyzers). These instruments are quite expensive. More recently, general HPLC
with specialized columns have been adapted to run this analysis. Many workers have attempted
analysis of amino acids by CE using different pre- and postreactions to enhance their detection. Both
free CZE and MEKC have been used with different degrees of success similar to that of amino acids
detection by HPLC. The separation of amino acids is better achieved in coated capillaries and also
better with MEKC. Derivatization, especially with fluorescent agents, offers much better sensitivity.
Several additive agents such as urea, cyclodextrin, and tetrabutyl ammonium salts improve the sep-
aration. Chiral amino acid separation based on MEKC has been developed to analyze and quantify
both d- and l-amino, which can be useful for detecting bacterial contamination [86,87]. On-capillary
derivatization of amino acids in serum with laser-induced fluorescence (LIF) detection has also been
reported [88]. Amino acid by CE has been reviewed [89–92]. The analysis of amino acids from
biological fluids by CE without interference and with good reproducibility remains a challenge. A
dedicated CE instrument for amino acids separation might someday be commercially available.
26.8.4 ION ANALYSIS

Many organic and inorganic ions are components of the biological fluids and cells. Most of these
have clinical and physiological importance. A change in their normal level is associated with different
disorders. Inorganic ions in the serum are important for maintaining cell viability, muscle excitability,
osmotic pressure, and regulating the pH. Thus they prevent muscle, renal, and heart malfunctions.
Organic acids are intermediates in the metabolism of many compounds in the cell. Clinically they are
important for detection of inborn errors of metabolism, infection, and different metabolic disorders,
such as diabetes. Some of the ions are not easy to measure regardless of the method. These ions
resemble each other, lack strong absorbency and may be present in low concentration. Depending on
the type of the ion, they are analyzed by several techniques [e.g., HPLC, automated clinical analyzer,
gas chromatography (GC), atomic absorption, or mass spectra]—all technically very involved. Many
inorganic ions are measured routinely in clinical labs. In practice, the common inorganic ions, such
as Na, K, Ca, can be measured more conveniently and rapidly in the clinical labs by ion-selective
electrodes built into highly specialized instruments.
The CE is quite suited for analysis of these ions too. Because of their relative charge to the small
mass, they tend to migrate rapidly in this technique giving fast separation with very high theoretical
plate numbers at a low cost per test. Both cations and anions can be analyzed in the same run. The
separation can be based on simple free solution CE or based on a suitable chelating additive. In
general, cations are measured in a low pH electrolyte containing a UV active species (imidazole
or benzylamine), while anions are measured after reversing the EOF and also after adding a high
mobility UV active species.
Organic acids are more difficult to measure compared with inorganic ions. Usually, these are
measured by the GC or HPLC. However, CE offers speed, precision, and specificity over other meth-
ods. Many of these compounds have been measured by CE directly, or by indirect UV absorbency
after the addition of a UV-absorbing compound such as benzoate, naphthalene, sulfonate, imidazole,
or benzylamine. For example, oxalate and citrate (Figure 26.7), which are important in stone for-
mation, have been measured after urine dilution by both direct and indirect UV detection [93–95].
Lactate, pyruvate, ascorbate, and oxalate were measured by CE in serum and in cerebrospinal fluids
of patients in ∼10 min [96,97]. Methylmalonic acid, which is a sensitive measure of vitamin B12
deficiency, preceding any clinical symptoms or changes in the serum level, has been determined in
urine by CE after sample extraction and concentration [65,98]. Some of the uncommon ions such as
nitrite and nitrate can be measured easily with CE (Figure 26.8) [99]. Plasma NO2 and NO3 were
analyzed after sample dilution using absorbance at 214 nm [100] and after serum deproteinization
with acetonitrile (Figure 26.8). We found the CE is much faster and less expensive for nitrate analysis
when compared with the enzymatic methods. CE has been applied for the determination of arsenic
acid and its related compounds using indirect detection in ammonium formate buffer [101]. Few
32
16 Patient 214 nm
Standard 214 nm
(a)
M CS
MC
Absorbance (mA)
0 0
32 16
Standard 185 nm Patient 185 nm
M
(b) M C
0 0
0 5 10 0 5 10
Min
FIGURE 26.7 Effect of the wavelength on citrate detection by CE. (a) Standard of malonic acid (500 mg/L)
and standard of citric acid (1000 mg/L) with detection at 214 nm; and patient urine at 214 nm; (b) same as
top but detection at 185 nm (C, citric acid; S, Succinic acid; M, malonic acid. (From Shihabi, Z., et al., J. Liq.
Chromatogr., 24, 3197, 2001. With permission.)
10
(a) Br (b)
Ni Na
Br
Ni Na
10 (c) (d)
mAuFS
Na
Ni
Ni
Na
10
(e) (f)
Na
Ni
Na
Ni
0
3 13 3 13
Min
FIGURE 26.8 Nitrite (Ni) and nitrate (Na) (6 mg/L), analysis by CE, added to (a) water without acetonitrile,
(b) 0.9% NaCl without acetonitrile, (c) 66% acetonitrile in water, and (d) 66% acetonitrile in 0.9% saline.
(e) Serum from an individual with low levels of nitrite and nitrate (for spiking), deproteinized with acetonitrile
(66% acetonitrile final concentration), and (f) the same serum spiked with nitrite and nitrate (6 mg/L, 66%
acetonitrile final concentration). (From Friedberg, M. A., et al., J. Chromatogr. A, 781, 491, 1997.)
ions, such as Fe, Ni, Zn, citrate, and oxalate, have been determined in urine based on transient ITP
[102]. Hydrogen peroxide that is involved in many biological reactions has been measured after
acetonitrile stacking [103].

The CE can be adapted to the separation of numerous compounds of clinical interest with some
compounds being more suited than others. It offers different advantages in the clinical field notably,
speed, high plate number, automation, and low operating cost. It is a good complement to other
separation techniques. At the same time, the CE has few limitations, such as poor detection limits
and the effect of the sample matrix on the separation. Understanding the principles behind these
limitations allows better strategies to encounter them. Biological samples usually have a high content
of protein and salts with a very wide range of concentrations for many compounds. This requires
special attention to how the samples can be treated, how the separation buffer is chosen, and how the
capillary is treated. The practical aspects and the special conditions for the separation of biological
samples are described. The CE has matured enough that there are several companies putting on
the market specialized fully automated instruments dedicated to special functions or special clinical
compounds such as serum protein separation, DNA analysis, or isofocusing of proteins. Numerous
applications of CE in the clinical field are described especially in the area of proteins, hemoglobin,
enzymes, and ions. Since the operating costs are much less than that of the HPLC, and it does
not require any appreciable amounts of organic solvents, application of CE in the clinical field
is expected to expand. The CE-MS is being refined for applications in the area of proteomics.
There remains a need for simpler methods of isofocusing in capillaries and simpler methods for size
separation for proteins. Further researches concerning amino acid analysis, dyes that bind to proteins
to improve the detection limits and improve the specificity are needed. The analysis of proteins by
CE probably will expand in the future as the separation and detection of microheterogeneity are
improved.
REFERENCES
1. Shihabi, Z.K. Effect of sample composition on electrophoretic migration application to hemoglobin
analysis by capillary electrophoresis and agarose electrophoresis. J. Chromatogr. A, 1027, 179, 2004.
2. Foret, F., Klepárník, K., Gebauer, P., and Boèek, P. Ionic boundaries in biological capillary
electrophoresis. J. Chromatogr. A, 1053, 43, 2004.
3. Garcia, L.L., and Shihabi, Z.K. Sample matrix effects in capillary electrophoresis: I. Basic
considerations. J. Chromatogr. A, 652, 465, 1993.
4. Shihabi, Z.K. Sample matrix effects in capillary electrophoresis: II. Acetonitrile deproteinization.
J. Chromatogr. A, 652, 471, 1993.
5. Shihabi, Z.K. Sample Preparation and Stacking for Capillary Electrophoresis. Encyclopedia of
Chromatography, 2nd ed. (Cazes, J., ed.), Taylor & Francis, New York, NY, 2005, 1490–1498.
6. Fortis, F., Guerrier, L., Righetti, P.G., Antonioli, P., and Boschetti, E. A new approach for the
removal of protein impurities from purified biologicals using combinatorial solid-phase ligand libraries.
7. Schappler, J., Guillarme, D., Prat, J., Veuthey, J.-L., and Rudaz, S. Enhanced method performances
for conventional and chiral CE-ESI/MS analyses in plasma. Electrophoresis, 27, 1537, 2006.
8. Shihabi, Z.K. Capillary electrophoresis for the determination of drugs in biological fluids. Handbook
of Analytical Separation (Elsevier), 5, 77, 2004.
9. Dolnik, V., Capillary electrophoresis of proteins 2003–2005. Electrophoresis, 27, 126, 2006.
10. Liu, J., and Lee, M.L. Permanent surface modification of polymeric capillary electrophoresis
microchips for protein and peptide analysis. Electrophoresis, 27, 3533, 2006.
11. Cordova, E., Gao, J., and Whitesides, G.M. Noncovalent polycationic coatings for capillaries in
capillary electrophoresis of proteins. Anal. Chem., 69, 1370, 1997.
12. Catai, J.R., Tervahauta, H.A., de Jong G.J., and Somsen, G.W. Noncovalently bilayer-coated capillaries
for efficient and reproducible analysis of proteins by capillary electrophoresis. J. Chromatogr. A, 1083,
185, 2005.
13. Sakai-Kato, K., Kato, M., Nakajima, T., Toyo’oka, T., Imai, K., and Utsunomiya-Tate, N. Cationic
starch derivatives as dynamic coating additives for protein analysis in capillary electrophoresis.
J. Chromatogr. A, 1111, 127, 2006.
14. Yassine, M.M., and Lucy, C.A. Preparative capillary zone electrophoresis using a dynamic coated
wide-bore capillary. Electrophoresis, 27, 3066, 2006.
15. Shihabi, Z.K. Stacking and discontinuous buffers in capillary zone electrophoresis. Electrophoresis,
21, 2872, 2000.
16. Krivankova, L., Pantuckova, P., Gebauer, P., Bocek, P., Caslavska, J., and Thormann, W. Chloride
analysis of anionic trace analytes. Electrophoresis, 24, 505, 2003.
17. Shihabi, Z.K. Stacking in capillary zone electrophoresis. J. Chromatogr. A, 902,107, 2000.
18. Shihabi, Z.K. Transient pseudo-isotachophoresis for sample concentration in capillary electrophoresis.
19. Kim, J.B., and Terabe, S. On-line sample preconcentration techniques in micellar electrokinetic
chromatography. J. Pharm. Biomed. Anal., 30, 1625, 2003.
20. An, Y., Cooper, W.J., Balgley, M.B., and Lee, C.S. Selective enrichment and ultrasensitive iden-
tification of trace peptides in proteome analysis using transient capillary isotachophoresis/zone
electrophoresis coupled with nano-ESI-MS. Electrophoresis, 27, 3503, 2006.
21. Campa, C., Coslovi, A., Flamigni, A., and Rossi, M. Overview on advances in capil-
lary electrophoresisass spectrometry of carbohydrates: A tabulated review. Electrophoresis, 27,
2027, 2006.
22. Lehmann, R., Liebich, H., Grübler, G., and Voelter, W. Capillary electrophoresis of human serum
proteins and apolipoproteins. Electrophoresis, 16, 998, 1995.
23. Liu, M., McNeal, C.J., and Macfarlane, R.D. Charge density profiling of circulating human low-density
lipoprotein particles by capillary zone electrophoresis, Electrophoresis, 25, 2985, 2004.
24. Macfarlane, R.D., Bondarenko, P.V., Cockrill, S.L., Cruzado, I.D., Koss, W., McNeal, C.J., Spieker-
man, M., and Watkins, L.K. Development of a lipoprotein profile using capillary electrophoresis and
mass spectrometry. Electrophoresis, 18, 1796, 1997.
25. Gordon, M.J., Lee, K.J., Arias, A.A., and Zare, R.N. Protocol for resolving protein mixtures in capillary
zone electrophoresis. Anal. Chem., 63, 69, 1991.
26. Chen, F.T., Liu, C.M. Hsieh, Y.Z., and Sternberg, J.C. Capillary electrophoresis—A new clinical tool.
Clin. Chem., 37, 14, 1991.
27. Chen, F.T., and Sternberg, J. Characterization of proteins by capillary electrophoresis in fused-silica
columns: Review on serum protein analysis and application to immunoassays. Electrophoresis, 15,
13, 1994.
28. Chen, F.A. Rapid protein analysis by capillary electrophoresis. J. Chromatogr., 559, 445–453, 1991.
29. Dolnik, V. Capillary zone electrophoresis of serum proteins: Study of separation variables. J. Chro-
matogr., 709, 99–110, 1995.
30. Roudiere, L., Boularan, A.M., Bonardet, A., Vallat, C., Cristol, J.P., and Dupuy, A.M. Evaluation of
a capillary zone electrophoresis system versus a conventional agarose gel system for routine serum
protein separation and monoclonal component typing. Clin. Lab., 52, 19, 2006.
31. Louahabi, A., Philippe, M, Lali, S., Wallemacq, P., and Maisin, D. Evaluation of a new Sebia kit for
analysis of hemoglobin fractions and variants on the capillary system. Clin Chem. Lab. Med., 44, 340,
2006.
32. Mussap, M., Ponchia, S., Zaninotto, M., Varagnolo, M., and Plebani, M., Evaluation of a new capillary
zone electrophoresis system for the identification and typing of Bence Jones Protein. Clin. Biochem.,
39, 152, 2006.
33. Shihabi, Z. Cryoglobulins: An important but a neglected clinical test. Ann. Clin Lab Sci., 36, 395,
2006.
34. Shihabi, Z. Analysis and general classification of serum cryoglobulins by capillary zone electrophore-
sis. Electrophoresis, 17, 1607, 1996.
35. Friedberg, M.A. and Shihabi, Z.K. Urine protein analysis by capillary electrophoresis. Electrophoresis,
18, 1836, 1997.
36. Sanders, E., Katzmann, J.A., Clark, R., Oda, R.P., Shihabi, Z., and Landers, J.P. Development of
capillary electrophoresis as an alternative to high resolution agarose electrophoresis for the diagnosis
of multiple sclerosis. Clin. Chem. Lab Med., 37, 37,1999.
37. Bao, J., and Regnier, F.E. Ultramicro enzyme assays in a capillary electrophoretic system.
J. Chromatogr., 608, 217–224, 1992.
38. Wu, D., and Regnier, F.E. Native protein separations and enzyme microassays by capillary zone and
gel electrophoresis. Anal. Chem., 65, 2029–2035, 1992.
39. Miller, K.J., Leesong, I., Bao, J., Regnier, F.E., and Lytle, F.E. Electrophoretically mediated microanal-
ysis of leucine aminopeptidase in complex matrices using time-resolved laser-induced fluorescence
detection. Anal. Chem., 65, 3267, 1993.
40. Landers, J.P., Schuchard, M.D., Subramaniam, M., Sismelich, T.P., and Spelsberg, T.C. High-
performance capillary electrophoretic analysis of chloramphenicol acetyl transferase activity. J. Chro-
matogr., 603, 247, 1992.
41. Shihabi, Z.K.Analysis of angiotension-converting enzyme by capillary electrophoresis. J. Chromatogr.
A, 853, 185, 1999.
42. Shihabi, Z.K., and Kute, T.E. Analysis of cathepsin D from breast tissues by capillary electrophoresis.
Chromatogr. B, 683, 125, 1996.
43. Viglio, S., Zanaboni, G., Luistetti, M., Cetta, G., Guglielminetti, M., and Iadarola, P. Micel-
lar electrokinetic chromatography: A convenient alternative to colorimetric and high perfor-
mance liquid chromatographic detection to monitor protease activity. Electrophoreis, 19, 2083,
1998.
44. Shihabi, Z.K., Hinsdale, M.E., and Daugherty, H.K., Jr. Hemoglobin A2 quantification by capillary
zone electrophoresis. Electrophoresis, 21, 749, 2000.
45. Cotton, F., Lin, C., Fontaine, B., Gulbis, B., Janssens, J., and Vertongen, F. Evaluation of a cap-
illary electrophoresis method for routine determination of hemoglobins A2 and F. Clin. Chem., 45,
237, 1999.
46. Mario, N., Baudin, B., Aussel, C., and Giboudeau, J. Capillary isoelectric focusing and high-
performance cation-exchange chromatography compared for qualitative and quantitative analysis of
hemoglobin variants. Clin. Chem., 43, 2137, 1997.
47. Jenkins, M.A. and Ratnaike, S. Capillary isoelectric focusing of haemoglobin variants in the clinical
laboratory. Clin. Chim. Acta, 289, 121, 1999.
48. Hempe, J.M., and Craver, R.D. Quantification of hemoglobin variants by capillary isoelectric focusing.
Clin. Chem., 40, 2288, 1994.
49. Wang, J., Zhou, S., Huang, W., Liu, Y., Cheng, C., Lu, X., and Cheng, J. CE-based analysis of
hemoglobin and its applications in clinical analysis, Electrophoresis, 27, 3108, 2006.
50. Jenkins, M., and Ratnaike, S. Capillary electrophoresis of hemoglobin. Clin. Chem. Lab Med., 41,
747, 2003.
51. Jabeen, R., Payne, D., Wiktorowicz, J., Mohammad, A., and Petersen, J. Capillary electrophoresis and
the clinical laboratory. Electrophoresis, 27, 2413, 2006.
52. Gerritsma, J., Sinnige, D., Drieze, C., Sittrop, B., Houtsma, P., Hulshorst-Jansen, N., and Huisman, W.
Quantitative and qualitative analysis of haemoglobin variants using capillary zone electrophoresis.
Ann. Clin. Biochem., 37, 380, 2000.
53. Siren, H., Laitinen, P., Turpeinen, U., and Karppinen, P. Direct monitoring of glycohemoglobin A1c in
the blood samples of diabetic patients by capillary electrophoresis. Comparison with an immunoassay
method. J. Chromatogr. A, 979, 201, 2002.
54. Yao, X.-W., and Regnier, F.E. Polymer- and surfactant-coated capillaries for isoelectric focusing.
J. Chromatogr., 632, 185, 1993.
55. Vincenzi, J.A., and Franco, M.M. Novel approach for the analysis of glycated hemoglobin using
capillary isoelectric focusing with chemical mobilization. J. Chromatogr. B, 785, 285, 2003.
56. Ong, C.-N., Liau, L.S., and Ong, H.Y. Separation of globins using free zone capillary electrophoresis.
J. Chromatogr., 576, 346, 1992.
57. Ferranti, P., Malorni, A., Pucci, P., Fanali, S., Nardi, A., and Ossicini, L. Capillary zone electrophore-
sis and mass spectrometry for the characterization of genetic variants of human hemoglobin. Anal.
Biochem., 194, 1, 1991.
58. Shihabi, Z.K., and Hinsdale, M.E. Simplified hemoglobin chain detection by capillary electrophoresis.
59. Shihabi, Z.K. Peptide stacking by acetonitrile-salt mixtures for capillary zone electrophoresis.
J. Chromatogr. A, 744, 231, 1996.
60. Shihabi, Z.K., Hinsdale, M.E., and Cheng, C.P. Analysis of glutathione by capillary electrophoresis
based on sample stacking. Electrophoresis, 22, 2351, 2001.
61. Yang, Y., Boysen, R.I., and Hearn, M.T. Optimization of field-amplified sample injection for analysis
of peptides by capillary electrophoresis-mass spectrometry. Anal. Chem., 78, 4752, 2006.
62. Sandra, K., Lynen, F., Devreese, B., Van Beeumen, J., and Sandra, P. On-column sample enrich-
ment for the high-sensitivity sheath-flow CE-MS analysis of peptides. Anal. Bioanal. Chem., 385,
671, 2006.
63. Stutz, H. Advances in the analysis of proteins and peptides by capillary electrophoresis with matrix-
assisted laser desorption/ionization and electrospray-mass spectrometry detection. Electrophoresis,
26, 1254, 2005.
64. Kasicka, V. Recent developments in capillary electrophoresis and capillary electrochromatography of
peptides. Electrophoresis, 27, 142, 2006.
65. Fliser, D., Wittke, S., and Mischak, H. Capillary electrophoresis coupled to mass spectrometry for
clinical diagnostic purposes Electrophoresis, 26, 2708, 2005.
66. Huang, Y.-F., Huang, C.-C., Hu, C.-C., and Chang, H.-T. Capillary electrophoresis-based separation
techniques for the analysis of proteins. Electrophoresis, 27, 3503, 2006.
67. Manabe, T., Oota, H., and Mukai, J. Size separation of sodium dodecyl sulfate complexes of human
plasma proteins by capillary electrophoresis employing linear polyacrylamide as a sieving polymer.
68. Yan, W., Sloat, A.L., Yagi, S., Nakazumi, H., and Colyer, C.L. Protein labeling with red
squarylium dyes for analysis by capillary electrophoresis with laser-induced fluorescence detection.
69. Sluszny, C., He, Y., and Yeung, E.S. Light-emitting diode-induced fluorescence detection of native
proteins in capillary electrophoresis. Electrophoresis, 26, 4197, 2005.
70. Lacroix, M., Poinsot, V., Fournier, C., and Couderc, F. Laser-induced fluorescence detection schemes
for the analysis of proteins and peptides using capillary electrophoresis. Electrophoresis, 26, 2608,
2005.
71. Giordano, B.C., Muza, M., Trout, A., and Landers, J.P. Dynamically-coated capillaries allow for
capillary electrophoretic resolution of transferrin sialoforms via direct analysis of human serum.
J. Chromatogr. B, 742, 79, 2000.
72. Lupi, A., Viglio, S., Luisetti, M., Gorrini, M., Coni, P., Faa, G., Cetta, G., and Iadarola, P.
Alpha1-antitrypsin in serum determined by capillary isoelectric focusing. Electrophoresis, 21, 3318,
2000.
73. Sanz-Nebot, V., Benavente, F., Vallverdu, A., Guzman, N.A., and Barbosa, J. Separation of recom-
binant human erythropoietin glycoforms by capillary electrophoresis using volatile electrolytes.
Assessment of mass spectrometry for the characterization of erythropoietin glycoforms, Anal. Chem.,
75, 5220, 2003.
74. Thorne, J.M., Goetzinger, W.K., Chen, A.B., Moorhouse, K.G., and Karger, B.L. Examination of
capillary zone electrophoresis, capillary isoelectric focusing and sodium dodecyl sulfate capillary
electrophoresis for the analysis of recombinant tissue plasminogen activator. J. Chromatogr. A, 744,
155, 1996.
75. Schmerr, M.J., Jenny, A.L., Bulgin, M.S., Miller, J.M., Hamir, A.N., Cutlip, R.C., and Goodwin, K.R.
Prions have been studied by capillary electrophoresis [Use of capillary electrophoresis and fluorescent
labeled peptides to detect the abnormal prion protein in the blood of animals that are infected with a
transmissible spongiform encephalopathy]. J Chromatogr. A, 853, 207, 1999.
76. Theodorescu, D., Fliser, D., Wittke, S., Mischak, H., Krebs, R., Walden, M., Ross, M., Eltze, E.,
Bettendorf, O., Wulfing, C., and Semjonow, A. Pilot study of capillary electrophoresis coupled to
mass spectrometry as a tool to define potential prostate cancer biomarkers in urine. Electrophoresis,
26, 2797, 2005.
77. Theodorescum, D., Wittke, S., Ross, M.M., Walden, M., Conaway, M., Just, I., Mischak, H., and
Frierson, H.F. Urothelial carcinoma was identified by CE_MS [Discovery and validation of new
protein biomarkers for urothelial cancer: a prospective analysis]. Lancet Oncol., 7, 230, 2006.
78. Ru, Q.C., Katenhusen, R. A, Zhu, L.A., Silberman, J., Yang, S., Orchard, T.J., Brzeski, H., Lieb-
man, M., and Ellsworth, D.L. Proteomic profiling of human urine using multi-dimensional protein
identification technology. J. Chromatogr. A, 1111, 166, 2006.
79. Geldart, S.E. and Brown, P.R. Analysis of nucleotides by capillary electrophoresis. J. Chromatogr. A,
828, 317, 1998.
80. Liu, C.C., Huang, J.S., Tyrrell, D.L., and Dovichi, N.J. Capillary electrophoresis-electrospray-
mass spectrometry of nucleosides and nucleotides: Application to phosphorylation studies of
anti-human immunodeficiency virus nucleosides in a human hepatoma cell line. Electrophoresis, 26,
1424, 2006.
81. Arcibal, I.G., Santillo, M.F., and Ewing, A.G. Recent advances in capillary electrophoretic analysis of
individual cells. Anal. Bioanal. Chem., 387, 51, 2007.
82. Weng, Q., Xu, G., Yuan, K., and Tang, P. Determination of monoamines in urine by capillary elec-
trophoresis with field-amplified sample stacking and amperometric detection. J. Chromatogr. B, 835,
55, 2006.
83. Peterson, Z.D., Collins, D.C., Bowerbank, C.R., Lee, M.L., and Graves, S.W. Determination of
catecholamines and metanephrines in urine by capillary electrophoresis-electrospray ionization-time-
of-flight mass spectrometry. J. Chromatogr. B, 776, 221, 2002.
84. Chen, G., Zhang, L., and Zhu, Y. Determination of glycosides and sugars in Moutan Cor-
tex by capillary electrophoresis with electrochemical detection. J. Pharm. Biomed. Anal., 41,
129, 2006.
85. Yang, Y., Breadmore, M.C., and Thormann, W. Analysis of the disaccharides derived from hyaluronic
acid and chondroitin sulfate by capillary electrophoresis with sample stacking. J. Sep. Sci., 28,
2381, 2005.
86. Carlavilla, D., Moreno-Arribas, M.V., Fanali, S., and Cifuentes, A. Chiral MEKC-LIF of amino acids
in foods: Analysis of vinegars. Electrophoresis, 27, 2551, 2006.
87. Chen, F., Zhang, S., Qi, L., and Chen, Y. Chiral capillary electrophoretic separation of amino acids
derivatized with 9-fluorenylmethylchloroformate using mixed chiral selectors of β-cyclodextrin and
sodium taurodeoxycholate. Electrophoresis, 27, 2896, 2006.
88. Veledo, M.T., de Frutos, M., and Diez-Masa, J.C. On-capillary derivatization and analysis of
amino acids in human plasma by capillary electrophoresis with laser-induced fluorescence detection:
Application to diagnosis of aminoacidopathies. Electrophoresis, 27, 3101, 2006.
89. Poinsot, V., Lacroix, M., Maury, D., Chataigne, G., Feurer, B., and Couderc, F. Recent advances in
amino acid analysis by capillary electrophoresis. Electrophoresis, 27, 176, 2006.
90. Poinsot, V., Bayle, C., and Couderc, F. Recent advances in amino acid analysis by capillary
electrophoresis. Electrophoresis, 24, 4047, 2003.
91. Prata, C., Bonnafous, P., Fraysse, N., Treilhou, M., Poinsot, V., and Couderc, F. Recent advances in
amino acid analysis by capillary electrophoresis. Electrophoresis, 22, 4129, 2001.
92. Smith, J.T. Recent advancements in amino acid analysis using capillary electrophoresis. Electrophore-
sis, 20, 3078, 1999.
93. Holmes, R.P. Measurement of urinary oxalate and citrate by capillary electrophoresis and indirect
ultraviolet absorbance. Clin. Chem., 41, 1297, 1995.
94. Wildman, B.J., Jackson, P.E., Jones, W.R., and Alden, P.G. Analysis of anion constituents of urine by
inorganic capillary electrophoresis. J. Chromatogr., 546, 459, 1991.
95. Shihabi, Z., Holmes, R., and Hinsdale, M. Urinary citrate analysis by capillary electrophoresis, J. Liq.
Chromatogr., 24, 3197, 2001.
96. Dolnik, V. and Dolnikova, J. Capillary zone electrophoresis of organic acids in serum of critically ill
children. J. Chromatogr. A, 716, 269, 1995.
97. Hiraoka, A., Akai, J., Tominaga, I., Hattori, M., Sasaki, H., and Arato, T. Capillary zone electrophoretic
determination of organic acids in cerebrospinal fluid from patients with central nervous system diseases.
J. Chromatogr. A, 680, 243, 1994.
98. Marsh, D.B., and Nutall, K.L. Serum methylmalonic acid by capillary zone electrophoresis using
electrokinetic injection and indirect photometric detection. J. Capillary Electrophor., 2, 63, 1995.
99. Friedberg, M.A., Hinsdale, M.E., and Shihabi, Z.K. Analysis of nitrate in biological fluids by capillary
electrophoresis. J. Chromatogr. A, 781, 491, 1997.
100. Leone, A.M., Francis, P.L., Rhodes, P., and Moncada, S. A rapid and simple method for the measure-
ment of nitrite and nitrate in plasma by high performance capillary electrophoresis. Biochem. Biophys.
Res. Comm., 200, 951, 1994.
101. Kitagawa, F., Shiomi, K., and Otsuka, K. Analysis of arsenic compounds by capillary electrophoresis
using indirect UV and mass spectrometric detections. Electrophoresis, 27, 2233, 2006.
102. Timerbaev, A.R., and Hirokawa, T. Recent advances of transient isotachophoresis-capillary elec-
trophoresis in the analysis of small ions from high-conductivity matrices. Electrophoresis, 27, 323,
2006.
103. Shihabi, Z. Direct analysis of hydrogen peroxide by capillary electrophoresis. Electrophoresis, 27,
4215, 2006.
27 Solid-Phase Microextraction
and Solid Phase Extraction
with Capillary Electrophoresis
and Related Techniques
Stephen G. Weber
CONTENTS
27.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 811

27.2 Microextraction Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 812
27.3 Preparation of SPME Probes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 815
27.3.1 Commercial SPME Probes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 815
27.3.2 Films and Fibers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 815
27.3.3 Packed Beds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 816
27.4 Interfacing SP(M)E with CE and Related Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 817
27.4.1 General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 817
27.4.2 Offline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 817
27.4.3 Online . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 818
27.5 Range of Applicability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 820
27.6 Challenges for the Future . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 820
27.6.1 Selectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 820
27.6.2 Desorption/Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 821
27.6.3 Smaller Probes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 821
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 821
27.1 INTRODUCTION
Improving the concentration detection limit is the goal of much analytical research and development.
While miniaturization leading to lower mass detection limits is by no means trivial, it is at least
conceptually obvious how to proceed. On the other hand, approaches to better concentration detection
limits often require considerable creativity. In some cases, better detectors offer lower detection
limits. Because properties of molecules differ, it is difficult to be globally accurate in a statement of
relative merits of detectors, but experience shows that the inherent detection limits for compounds
that are well suited for each detector are fluorescence < electrospray ionization mass spectrometry
(ESI-MS) ∼ electrochemistry < optical absorbance. It is worth noting that low detection limit is
correlated with the selectivity of the detector. In fact, in real samples of sufficient complexity (are
there any real samples that are not complex?), Nagels1–4 has shown that detection limits correlate
811
with the selectivity of the detector. The reason that detection limits correlate with the selectivity of
the detector is that the limitation in real analyses is interferences from chemical constituents of the
sample. These may be perceptible peaks in an electropherogram, or they may be a “baseline” of
poorly resolved, multiple unknown components that vary from sample to sample. When this is the
case, attempts to improve the detection limit by increasing the concentration of all the components
in a sample will fail.
The high-resolution capability of capillary electrophoresis (CE) coupled with recent progress in
comprehensive two-dimensional separations give a glimmer of hope that some day all components
of a sample will be resolved, and then detection limits will be dictated by the inherent detector
properties. If this is the case, then the act of concentrating all of the components in a sample will lead
to better detection limits. A consideration of the peak capacity problem, however, tells us that we have
a long way to go to realize this goal. We must, therefore, consider ways to improve analytical methods
based on CE and related techniques for real samples knowing that there are more components in a
sample that our most well-designed separations system can resolve. What is required is a method
that concentrates selectively the desired analytes, and does not concentrate, or better yet, that dilutes
undesired components of a sample.
Microextractions have the capability to concentrate selectively compounds with particular chem-
ical properties, and to reject solutes with other particular chemical properties. Microextractions
function on a small volume scale. As the volumes involved in microextraction and CE can be
(but are not necessarily) the same, and given the capability of microextraction to concentrate ana-
lytes selectively, microextractions are a natural choice for improving virtually any CE (micellar
electrokinetic chromatography, MEKC; capillary electrochromatography, CEC; etc.) method.
27.2 MICROEXTRACTION APPROACHES

This chapter will describe two approaches to microextraction: solid-phase extraction (SPE) and solid-
phase microextraction (SPME). It is not a comprehensive review. Rather certain key elements of each
approach and how it interfaces to the separations step will be described. As SPE is more intuitively
understood, and SPME perhaps less so, there is a focus on SPME. SPME has been reviewed from
several different perspectives.5–17
SPME was first developed in Pawliszyn’s laboratory and applied to gas chromatography
(GC).18 The technique has been commercialized and is available through Supelco (now part of
the Sigma/Aldrich company in the United States). The technique has some marvelous attributes—it
is more than a different way to carry out a separation. Figure 27.1 shows in schematic form the basic
idea as originally conceived. An SPME probe is simply an inert rod or fiber (e.g., an optical fiber)
coated with any of a number of types of materials such as the highly polar poly(acrylic acid) or
the highly nonpolar poly(dimethylsiloxane). The probe tip can be immersed directly in the sample
or head space sampling may be preferred. The most common application is to gas chromatography
(GC). The probe is inserted into the injection port where the high temperature and gas flow desorb and
carry away the analytes. The commercial device incorporates the SPME probe into a syringe barrel.
This both protects it and makes the injection into a GC seamless. Recently, a commercial device, also
from Supelco, has become available for interfacing with high-performance liquid chromatography
(HPLC).
One property of SPME that distinguishes it from other types of extraction is the phase ratio
(organic or extracting phase volume/sample volume). The phase ratio in SPME can be very small
(e.g., 1/1000). When the phase ratio is small, then the number of moles of an analyte extracted can
be small in comparison to the total number of moles in the sample. Unlike in bulk extraction where
most of the analyte can be extracted from a sample, in SPME it is possible that a negligible fraction
of the analyte in a sample is extracted. Because of this, the quantity extracted will reflect the “free
Solid-Phase Extraction with Capillary Electrophoresis and Related Techniques 813
(a) (b) (c)
Septum
Extracting phase Liquid sample
FIGURE 27.1 Scheme of SPME. (a) An SPME probe with an extracting phase as a film on the end of a rod or
fiber, and a liquid sample in a container. (b) Extraction by immersion of the probe in a sample or by headspace
sampling. (c) Desorption in the inlet of a GC column.
concentration” of the analyte, not the “total concentration.” For example, it is possible to determine
total drug concentration from blood serum using a bulk extraction. On the other hand, an SPME
experiment can be configured so that the “free” drug concentration is measured. In a sense SPME is
to bulk extractions what pH measurements are compared to titrations with base.
Another remarkable property of SPME can be realized with volatile solutes. It turns out to be faster
(and more selective) to extract volatile solutes with headspace extraction rather than by immersing the
SPME probe into the sample. Gaining speed from headspace extraction seems counterintuitive —
after all, headspace analysis involves an extra step (sample—vapor phase—extracting phase vs.
sample—extracting phase). The reason that headspace can be faster (more rapid progress toward
equilibrium; more analyte extracted in a given time) is that the rate-limiting step in the extraction is
typically solute getting out of the sample. The flux of analyte in this case is related to a mass transport
coefficient (that depends on diffusion and fluid flow if present) for the analyte in the sample and also
to the surface area through which the analyte flux goes. The probe tip-sample surface area is the
area across which the rate-limiting flux occurs in the direct immersion case. In the case of headspace
analysis, the sample-vapor phase surface area is the area across which analyte flux occurs. The latter
area is far larger than the former, so headspace extraction proceeds faster than direct immersion.
Another advantage of SPME over bulk extractions especially and over SPE to a lesser degree is
the variety of extraction phases that can be made, and the ease with which phases can be developed
and used in research laboratories. In bulk extraction, there are certainly differences among organic
solvents in their ability to extract analytes. However, using one organic solvent in preference to
another is not a powerful method to control the extraction selectivity. In SPE, chromatographic
particles are used. Here, there is a variety of choices. However, preparation of a new phase is time
consuming at best, and difficult to do and difficult to analyze at worst. In contrast, making a film
of polymer or other type of material on a surface can be very easy. Thickness can be controlled by
simple means using dip coating.
SPME does have drawbacks. One potential drawback is the equilibration time. Probes typically
equilibrate with sample on the timescale of tens of minutes to an hour or more. In many applications,
it is not necessary to wait for equilibrium, thus the extraction step need not take a long time. However,
by making the extraction yield depend on kinetic as well as equilibrium phenomena, there will be
more matrix effects to consider. Things like viscosity, percent solids, and the presence of species
that can foul the probe surface have an influence on a kinetically limited extraction, whereas they do
not on an equilibrium extraction. SPME is ideally suited to GC. The fit with liquid-based techniques
is not as simple. Desorption must be into a solution or liquid film requiring some thinking about
the chemistry, while in GC, there is nothing really to think about except the details: elevating the
temperature in the gas phase will release the extracted analytes into the column. Desorption in CE
(and HPLC) often involves a solvent that can itself dissolve in the film, so desorption can be faster
than adsorption—on the minutes timescale. On the other hand, it takes some effort to mate a purely
organic extract with CE.
Finally, there are some secondary considerations in SPME that, if properly handled, can
improve analysis. SPME probes are reusable, which is great from a cost and environmental bur-
den perspective, but as a consequence carryover is a concern. Washing (or heating for volatile
solutes) is essential between applications. Modifying the sample, for example, by adding salt,
can influence analyte yield. Unlike in SPE or bulk extraction, it is not necessary to choose
between a polar and a nonpolar phase. Phases with both properties, e.g., Carboxen/PDMS
exists. Carboxen is a porous carbon that adsorbs polar and nonpolar solutes, and PDMS is quite
nonpolar.
SPE is conceptually simpler than SPME. Sample is passed through a packed bed. The bed strips
the analyte from the sample. A desorbing solvent is then passed through the bed and passed on for
analysis. Commonly used phases are C18 and polystyrene/divinylbenzene (PS/DVB), but others
more selective are also used as will be discussed below. It is difficult to generalize, but despite the
nominally similar size of the objects used for SPME and SPE (mm length scale, 100 µm diameter
scale) the phase ratio (extracting phase/sample) is larger in SPE than SPME because of the porosity
of most materials used in SPE. Consequently, it is not typically the case that the extraction reveals
the “free” concentration of an analyte. Also because of the porosity and thinness of the extracting
layer, for example, in a C18 material, the presence of the support (nominally the same in an optical
fiber-based SPME and a silica-based SPE) can be more evident in SPE. Some skill is required
to prepare CE- or capillary HPLC-scale packed beds. There are practical advantages of SPE, too.
Packed beds are a natural match to flowing fluid, so adsorption and desorption are easily carried out
by solvent switching. There is a considerable history of adsorbent selectivity that can be applied to
new problems. As the typical particles are small, adsorbent layers are thin and mass transport outside
of the particles is rapid because the fluid flow augments diffusion, equilibration times (equivalent
roughly to the C term in the van Deemter equation) are short.
The defining features of SPME are mostly film-based, and the film volume is markedly smaller
than the sample volume. SPE is typically particle-based with a larger phase ratio. However,
these terms are not precise, and they are used differently in the literature. Also, there are some
approaches that simply do not fall easily into one category or the other. This writer is not enthusiastic
about creating unique terms for even finer distinctions between techniques, so this will be avoided
here. The terms SPME and SPE will be used generally, though not exclusively, as the original authors
used them. When all else fails, or for general comments, the abbreviation SP(M)E will be used.
SPME has been applied to CE and related techniques, first by Li and Weber19 offline and then by
Nguyen and Luong20 with online back extraction both in 1997. The approach possesses a number of
potential advantages, not the least of which is the obvious applications to the chip format in which
film and gel (as in sol-gel) formation are possible. Another advantage is that desorption can often
be followed by stacking or a variant of it obviating the problem of bandspreading induced by the
desorption step. The size scale of CE and SPME are also similar, so the fit is natural from a conceptual
perspective. On the other hand, there are engineering difficulties that include among other things
how to manage the many fluids required (adsorption, desorption, separation, column rinsing, SPME
rinsing), and how to incorporate fluid flow and electrokinetic flow. There are several very useful
reviews cited above that are considerably broader than the current overview that provides examples
of SPME-CE and related techniques. The remainder of this chapter draws from the primary literature
in the hope of presenting the sort of detailed understanding that makes this approach to analytical
chemistry understandable, and allows the novice to anticipate problems and recognize opportunities.
27.3 PREPARATION OF SPME PROBES

27.3.1 COMMERCIAL SPME PROBES
Several workers use commercial SPME probes in conjunction with CE and related techniques. In all
cases, probes were conditioned both before first use and between uses. Poly acrylate (PA) probes were
conditioned in 50:50 methanol/water (30 min)21 or the desorption solvent, for example, acetonitrile.22
A carboxen/PDMS probe was cleverly conditioned by exposing it to GC inlet conditions (300◦ C
for 2 h).23 The solutes in this case were halophenols. Obviously, this treatment would not be useful
for ionic compounds, for example, cations in a PA phase (although if the cation were a low molecular
weight protonated nitrogen base, it may work). Thirty minutes in methanol for a PDMS/divinyl
benzene (DVB) probe proved suitable.24 Other workers25 used a 30 min initial conditioning for four
different types of probes, and used a 20 min conditioning between analyses. The theme is clear.
A 30 min exposure to a suitable solvent will work to condition SPME probes for use in CE both
before and between analyses.
27.3.2 FILMS AND FIBERS

Others interested in the development of SPME per se prepare their own probes. It is not necessary
to start from scratch. One group20 purchased a particular optical fiber from Polymicro and used
dimethylformamide (DMF) to srip off a layer of nylon. This revealed a layer of PDMS underneath.
Of course, PDMS is a well-studied and understood extraction phase, so this is a very clever route
to something home made, and thus suited to a particular experiment, but commercially available.
Another approach uses a differently coated optical fiber. The polyimide coating on an optical fiber
was removed using 98% sulfuric acid at 120◦ C for 15 h. (Note: Hot, concentrated sulfuric acid is
extremely corrosive and must be treated with utmost respect. Use the absolute minimum quantity
to carry out the task. Use an apparatus in which both the containers of sulfuric acid and the fiber
are held in place reliably and securely.) Following this treatment, the clean surface could react with
aminopropyltriethoxysilane (APTES) to form an amine (ammonium in neutral or acidic solutions)
surface. The surface could also be exposed to a solution of pyrrole and ammonium persulfate in
1:1 water/isopropanol for forming a poly(pyrrole) surface.26 When interfacing with CE, handling
the probe is a problem. Typically, CE does not involve a device for injection, rather the injection is
made directly into the separation capillary with no other mechanical device (even if a mechanical
device such as an autosampler is used, at most the mechanical assistance in the injection per se
is something to pierce the septum enclosing the sample). Thus, unlike HPLC and GC, interfacing
requires creativity. This topic will be taken up below. Here, though, it is appropriate to describe
an approach that permits both modification of a surface and adaptation of the probe to a device.
Whang and Pawliszyn27 developed a method to create a probe suitable for CE, and that fits into
the commercial, syringe-like holding apparatus (see Figure 27.2). The probe thus formed is placed
into the syringe-like apparatus that is commercially available after the probe with which it came is
removed. While the field of SPME has been dominated by a handful of phases spanning a wide range
of polarity, there are other phases with excellent properties. Poly vinylchloride (PVC), appropriately
plasticized, is the basis for ion-selective electrodes. In a sense, SPME experiments have been done on
this material for decades, but without that explicit purpose in mind. Valenta et al.28,29 has described
the fact that the polarity of plasticized PVC depends on the plasticizer. Zhang et al. has demonstrated
that the polarity also depends on the ratio of plasticizer to polymer.30 Li and Weber19 have used
this material in SPME-CE. The probes are based on a stainless steel rod. A primer coating of
poly(vinylchloride-co-vinyl alcohol-co-maleic acid) permits a coating of plasticized PVC (dip coat
from tetrahydrofuran). The probes are reusable. The polymer also plays a role in the selectivity with
of extractions based on molecular recognition. A barbiturate receptor in plasticized PVC extracts
barbiturates more selectively the lower the polarity of the plasticizer.
(a) (b) (c) (d) (e) (f)
FIGURE 27.2 Preparation of an SPME probe for CE.27 (a) A 245 µm diameter PA coated optical fiber
3 cm long. (b) Strip the PA from 2 cm. (c) Glue a 10 cm long piece of fused silica capillary (100 µm ID,
245 µm OD) to the distal 0.8 cm of the exposed 2 cm. (d) Cut at a point 1.0 cm from the end of the capillary.
(e) Etch with HF (Note: use in a hood. Breathing HF is harmful.) to about 40 µm diameter. (f) Dip coat X3 in
3% poly(acrylate).
Porous materials can be used for SPME. Carboxen particles have been mentioned above. A
monolithic porous polymer has been prepared for SPME. Wei et al.31 made a porous polymer
monolith with monomers methacrylic acid, ethylene glycol dimethacrylic acid, and a toluene
porogen. The material was prepared inside a 250 µm capillary, which had been previously deriva-
tized with a methacrylic acid-containing silane. Following polymerization at 60◦ C for 16 h, the
monolith was washed with methanol, a popular choice for conditioning. Application (which is
described more below) uses flow through the medium, so perhaps this is more properly called
micro-SPE. However, the authors call it SPME, and it is an interesting approach, so it is being
described here. Zhou et al.32 used molecular recognition to assist in the SPME. A conjugate
of the well-known small molecule host, calix-[4]-arene, with a silane yields a reactive ethoxy
silane that is functionalized with the calixarene. A sol of this and tetraethoxysilane with a poro-
gen (-OH terminated silicone oil) was prepared. A fiber was dipped into the sol several times, then
the sol was given time to react to form the gel. Control probes were made in which the calixarene-
modified silane was not present. These probes were used for head space extractions, then back
extraction and analysis with CE. The calixarene probe was more effective for propranalol than the
control.
27.3.3 PACKED BEDS

Packed beds, as mentioned above, are sometimes described as SPME, and indeed, it may be
accurate—that is the extracting phase may reach or approach equilibrium with a sample without
extracting a significant amount of analyte on average from the sample passed through the packed
bed. Typically, a small piece of a filter material is used as a “frit” to block particle flow at the boundary
between a capillary holding the particles and another capillary.33 If the particles are large enough
with respect to the downstream capillary diameter, then there is no need for a frit or other device to
contain the particles. In Reference 34, 10 µm particles were introduced into a 180 µm capillary as
a very dilute suspension at modest pressure. The bed was formed where the 180 µm capillary met a
restrictor that was 50 µm in diameter.
27.4 INTERFACING SP(M)E WITH CE AND RELATED TECHNIQUES

27.4.1 GENERAL CONSIDERATIONS
Interfacing SPME with CE is not as straightforward as interfacing SPME with GC or as interfacing
SPE with CE. Because of the difficulty of interfacing, a lot of work has been done with offline
“interfacing.” The general issues faced when considering interfacing (in general) are as follows:
Preconcentration: A good extraction/back extraction cycle not only results in a cleaner sample,
but analytes are concentrated in the process. Obviously, this argues for using the smallest
possible back extraction solvent volume. If this volume needs to be manipulated by hand,
that is, if it is not a volume resulting from a programmed mechanical device or by switch-
ing on a chip, then the minimum volume of the back extraction solvent will probably be
about 1 µL.
Kinetics: An effective extraction/back extraction cycle is fast. Slow extraction or back extrac-
tion requires that the designer of the system be cognizant of the details. Slow extraction can
obviously lead to poor yield. Operating in a kinetic regime means that many more parameters
influence the outcome in comparison to operating with the extraction at equilibrium. Slow
back extractions have the same problem. In addition, a slow back extraction online can add to
bandspreading.
Fluid management: Depending on the design of a system there may be five different fluids
needed for an analysis: sample, back extraction solution, running buffer/electrolyte, extraction
phase rinsing, and column rinsing/conditioning solution. These liquids need to be managed
appropriately for a successful outcome of a single analysis and for good reproducibility.
On balance, the design of a system is a trade-off between the complexity of mechaniza-

tion/automation versus the savings of time (and improvement in reproducibility) once the effort
to automate/mechanize has been made. Because the effort to create online systems is considerable,
there has been significant effort in offline systems. Both offline and online systems are covered
below. The section on “online” systems includes methods that have at least one online step.
27.4.2 OFFLINE
Both in-sample and headspace approaches have been used in conjunction with CE. The forward
extraction, as is typical for SPME, can take an hour or more to equilibrate. The yield of the extraction
is improved by the addition of salt (typically NaCl). As the probe is intended to reach equilibrium
with the sample and not extract a large fraction of the moles of analyte in the sample, the volume of
sample can be immaterial. The first SPME-CE paper used offline extraction and back extraction.19
Management of the back extraction solution was made easier by keeping it in a piece of Teflon® .
The general scheme is shown in Figure 27.3.
Using such a system, the back extraction solution can be kept to a few microliters. In most
instances, however, the back extraction is into a much larger volume of 20–200 µL.21–24,31,32 In
most cases, the back extraction medium is not highly ionic. This permits stacking, electric field-
induced concentration at the interface between the relatively high conductivity running buffer and
the relatively low conductivity back extraction solvent. The option to stack means that it is less
urgent to use a small volume of back extraction solvent.
In most of the articles cited above, the authors measured the desorption kinetics. These turn out
to be remarkably consistent—with about 90% recovery (from the probe) in minutes (∼3–10 min).
This is interesting, because at least two modes of back extraction are used. In one, a solvent more
“organic” than water is used (e.g., methanol or acetonitrile). On the other hand, the ionic nature of
(a) (b) (c) (d)
FIGURE 27.3 Off-line extraction and back extraction with back extraction solution management.19,21
(a) extraction with SPME (dark portion = extracting phase, diagonal hatching = sample). (b) Separately,
inject a small volume (few µL of back extraction solvent into a Teflon® tube sized to have a diameter just
larger than the diameter of the SPME probe. (c) Insert the probe into the back extraction solution-containing
probe. The solution will spread between the wall of the tube and the extracting phase. (d) Following the back
extraction, remove the SPME probe. Using an empty syringe, move the droplet, which re-forms when the probe
is removed, at will.
the analyte is switched by changing the pH of the back extraction solvent from the value existing in
the sample.
There is one report of an offline SPME (or perhaps more properly SPE) extraction system for
CE analysis in which the offline extraction is automated.31 The extraction medium is a monolith
placed in the loop position of a six-port loop injector. An injector loop upstream contains a large
loop with sample. The sample from the loop in the upstream injector is passed through the extraction
medium for a controlled time at a controlled flow rate. Following a rinse, the back extraction solvent
is then pumped through the monolith into a collection vial. This system has the advantage of the
reproducibility of an automated system, and simultaneously avoids the complexity of the interfacing
process.
27.4.3 ONLINE
When the separation medium is online, it may be within the separation path or not. Figure 27.4
shows the two most common arrangements. One pair of arms (horizontal or vertical) is for sample
management, while the other pair (vertical or horizontal) is for the separation. The tubing need not
be of the same diameter in each arm or branch. If the SP(M)E medium is in the center of the cross
arrangement (Figure 27.4a), then obviously the SP(M)E medium is “online” during the separation.
Elution from the SP(M)E material occurs in the separation column. This places severe demands
on the effectiveness of the back extraction and the volume of back extraction solvent. Of course,
stacking may be used to address problems of having an initial back extraction zone that is too large
in many cases. The back extraction must be complete in the arrangement, too. If it is not, the peaks
will tail, or worse, the separation will be impossible. If the SP(M)E medium is in one branch of
the cross, then the demands on the back extraction are not as severe. Incomplete back extraction is
permitted. A flow-gate type of injection is possible to assure narrow zones. On the other hand, the
back extraction solution can be directed onto the separation column if stacking, sweeping, or other
online reconcentration is possible. Variations on these themes exist as well.
This sort of cross arrangement has been used by several workers. Peptides were trapped, separated
by CE, and detected by ESI-MS using an arrangement as shown in Figure 27.4b.34 This paper
demonstrated the creation of the online SP(M)E in a PDMS chip. Once every 10 s a 1 s injection
was made from the back extraction solution. It appeared that the peptides eluted from the SP(M)E
(a) (b)
FIGURE 27.4 (a) A cross arrangement for online extraction and back extraction with the extraction medium
centered in the cross. (b) Reaction medium in one arm.
[very high surface area hypercrosslinked poly(styrene)] over about a 3 min period. In a similar way
(Figure 27.4b), a short capillary packed with small fibers (Zylon) was used for the extraction of
tricyclic antidepressants.35 Samples could be driven through the extraction medium at a high flow
rate (80 µL/min). Back extraction volumes varied, but were in the range of 2 µL, of which 4 nL
were injected onto the CE column. An alternative arrangement similar to Figure 27.4b has only
three arms. The “waste” line for the extraction and the running buffer source (positive electrode in
“normal” polarity) capillary are the same.36 This leads to interesting possibilities for flow control,
as pressure flow going through the SP(M)E medium splits—a predictable amount going to waste,
and a predictable amount going onto the separation column.
A cross like Figure 27.4a has been used for affinity capture by Guzman.37 In this case, each of
the four arms is outfitted with a valve as well, so the flow is completely controllable. As mentioned
above, however, it is important to desorb the analytes into a small volume and rapidly. Guzman uses
100 nL of desorption solution (pH 3.4 glycine buffer or acetonitrile) to dissociate the antibody–analyte
complexes. A cross with medium centered in it was fabricated on a chip.38 In this paper, the goal was
manufacturing, not applying, the SP(M)E. A small packed bed served as the hydrophobic medium.
Laser-induced fluorescence was used for detection of the “analyte,” fluorescein. Interestingly, the
beam was split so that the luminescence from the fluorescein in the extraction chamber could be
seen. Evidence of rapid and complete desorption was obtained in this way.
It is possible to construct a system in which there is only one pair of “arms”—that is with
an extraction medium and a separation capillary in series. In an example of this, Yates and
coworkers33 extracted proteins and then desorbed them into a zone of methanol introduced with
a 45 s electrokinetic injection.
The previous works have been based on beds of particles or fibers. Films, that is, with a construc-
tion similar to commercially available SPME probe tips, can be used online. The first publication in
this area was half online. The sample extraction was carried out offline, but desorption and analysis
was managed online.20 In this and other papers, the extraction fiber is constructed as in Figure 27.2.
A fiber with analytes adsorbed is inserted into the injection end of the capillary. The injection process
involves introducing not sample but a desorption solution. The flow may be stopped to allow for
desorption (recall that it can take several minutes), then restarted for separations. In a similar vein,
Whang and Pawliszyn27 created a system in which fibers with extracted analytes were moved to a
capillary for desorption/analysis. In this case, though, the desorption solution is injected into the
capillary before inserting the SPME fiber into the end of the capillary. Also in this case, the fiber
is attached to a syringe for ease of handling. While it is an advantage in some sense, the use of
a low (organic/aqueous) phase ratio does limit the number of moles extracted from a sample. As
this places limitations on the follow-up analysis (mass detection limit), there is some advantage to
keeping the SPME tip “long.” The “active” extracting tips are typically ∼1 cm in length. One may
wonder about the effect of creating an annular space filled with analyte on bandspreading in CE.
Stoyanov and Pawliszyn39 have shown that the problem is not too large. Because the cross-sectional
area of the annular space created by inserting the probe into the injection end of the capillary is small
compared with the full cross-sectional area of the capillary, the field is higher in the annular space
than the capillary. This requires that the velocity of the analytes slows down when they reach the
full capillary (having come from the annular space). The result is a compression of sorts that helps
keep the injection-induced bandspreading to a minimum.
Another way to fight the potential bandspreading from the desorption/injection process is to
perform isoelectric focusing rather than C(Z)E.26 The focusing effect of the separation technique
undoes any bandspreading resulting from the desorption injection process.
Despite clever approaches, the online desorption/injection from SPME is not a “natural” fit. An
approach that holds promise for macromolecular analytes is to use microdialysis as an intermediate
step. Liu and Pawliszyn40 achieved this by back extracting and moving analytes electrokinetically
into the lumen of a microdialysis probe, then inject from the lumen into the separation capillary. The
beauty of this is that the sample stays in the (small) microdialysis probe volume, but electrolyte and
current can pass through the walls of the microdialysis probe.
Some general practical comments are in order. Care must be taken to make good butt connections
when coupling capillaries. Band spreading can result from poor connections. Many applications
of SP(M)E-CE are partly offline, or involve physically working with capillaries. Also, solvents of
various sorts are used for desorption. Therefore, there is abundant opportunity for the formation of
bubbles. It is a good idea at some point in the process to use pressure to flush the system before using
electrokinetic forces for separations. Many publications discuss moving fluid where desired using
both pressure and electroosmotic flow. Because flow control generally needs to be quantitative, it is
often necessary to have valves or pressurized electrolyte containers to facilitate predictable fluid flow.
27.5 RANGE OF APPLICABILITY

The range of analytes accessible to SP(M)E/CE is wide. Drugs, pollutants (and very hydrophobic
ones like polyaromatic hydrocarbon, PAHs), proteins/peptides have all been determined using the
methods described herein. There does not seem to be any sacrifice in separation modes, either. CE,
capillary isoelectric focusing, MEKC, nonaqueous CE can all function with some form of SP(M)E.
Methods of detection are similarly broad, encompassing at least the three most common: optical
absorbance, laser-induced fluorescence, and MS. It is perhaps foolish to attempt to generalize, but it
does seem that an improvement in concentration detection limits of about 100-fold can be realized
through SPME in conjunction with CE.
27.6 CHALLENGES FOR THE FUTURE

27.6.1 SELECTIVITY
Many analyses are for a few similar molecules, or a class of molecules. In such cases, selectiv-
ity is highly desirable in the extraction. Selectivity in SPME has been reviewed.41 Information is
also available at the Supelco website (this author is not and has never been funded by Supelco,
and this author has no financial interest in Supelco) http://www.sigmaaldrich.com/Brands/
Supelco_Home/Spotlights/SPME_central.html.
Ultimately, molecular selectivity is desirable. Calix-[4]-arenes improved the extraction of pro-
pranolol. Antibodies can be used for affinity separations. Li et al.42 have demonstrated selectivity
for barbiturates using a specially designed receptor molecule. The need for more selectivity is made
clear by the following common observation. Recall that SPME extractions can take a long time to
reach equilibrium. Is it worthwhile to wait for equilibrium? It can be as mentioned above, it should be
more reproducible. But interestingly, for many samples extracted by direct immersion of the probe
into the sample there is no detection limit advantage to longer extractions because interferences
are extracted along with analytes. More research and development on selective phases is certainly
needed.
27.6.2 DESORPTION/INJECTION
There remains, despite some very clever approaches, a need for a seamless union of SPME and
CE. This may ultimately come with lab-on-a-chip developments, as the goal of much of this sort
of research is complete analysis. In other words, rather than the necessary voltages, pumps, valves,
and pressure sources being a bother (as it is to a worker using commercial CE, commercial SPME,
commercial valves, etc.) it is a research opportunity for the courageous chemists and engineers
designing chips. This prognosis, then, points out a need for the development of extraction phases
(with selectivity, see above) that can be emplaced in the microfabrication process. This is an important
area for future work.
27.6.3 SMALLER PROBES

There is room to make smaller, thinner probes to manage microanalysis. Interfacing them brings new
challenges, but also may obviate some problems. Smaller, thinner phases with faster equilibration
times may improve the desorption/injection process.

The combination of SP(M)E and CE is a good idea. CE suffers from poor concentration detection
limits; microextractions provide for preconcentration. There is apparently no inherent sacrifice that
must be made in separations mode or detection mode in order to make SP(M)E work with CE and
related techniques. The biggest advantage occurs when the extraction is fairly selective: then the
act of concentrating the analytes (and not concentrating interferences) improves the concentration
detection limit for the method. Examples of where this is the case would be in head space extractions,
and extractions based on selective or affinity phases. The engineering of the extraction and back
extraction online is still awkward, but improving.
ACKNOWLEDGMENTS
The author is grateful to the NSF Division of Chemistry for funding. Contribution from the
Department of Chemistry, University of Pittsburgh, Pittsburgh, PA.
REFERENCES
1. Creten, W. L. and Nagels, L. J., Computation of determination limits for multicomponent chro-
matograms, Analytical Chemistry, 59, 822, 1987.
2. Nagels, L. J., Creten, W. L., and Parmentier, F., Statistical model for organic chromatographic trace
analysis of complex samples. A case study: Plant extracts, International Journal of Environmental
Analytical Chemistry, 25, 173, 1986.
3. Nagels, L. J., Creten, W. L., and Van Haverbeke, L., Determination limits in high-performance liquid
chromatography of plant phenolic compounds with an ultraviolet detector, Analytica Chimica Acta,
173, 185, 1985.
4. Nagels, L. J. and Creten, W. L., Quantitative evaluation of chromatographic analysis of complex
mixtures by establishing limits of determination, Analytica Chimica Acta, 169, 299, 1985.
5. Andreu, V. and Pico, Y., Determination of pesticides and their degradation products in soil: Critical
review and comparison of methods, Trends in Analytical Chemistry, 23, 772, 2004.
6. Fritz, J. S. and Macka, M., Solid-phase trapping of solutes for further chromatographic or
electrophoretic analysis, Journal of Chromatography A, 902, 137, 2000.
7. Jinno, K., Sawada, H., and Han, Y., Drug analysis by capillary electrophoretic methods, Biomedical
Chromatography, 12, 126, 1998.
8. Liu, Z. and Pawliszyn, J., Online coupling of solid-phase microextraction and capillary electrophoresis,
Journal of Chromatographic Science, 44, 366, 2006.
9. Mile, B., Chemistry in court, Chromatographia, 62, 3, 2005.
10. Pawliszyn, J., New developments and applications of solvent-free sampling and sample preparation
technologies for the investigation of living systems, Australian Journal of Chemistry, 56, 155, 2003.
11. Pawliszyn, J., Applications of Solid Phase Microextraction, Royal Soc. Chem., Cambridge, 1999.
12. Pawliszyn, J., Comprehensive Analytical Chemistry, Volume 37. (Sampling and Sample Preparation
for Field and Laboratory: Fundamentals and New Directions in Sample Preparation.), Elsevier Science
B.V., Amsterdam, The Netherlands, 2002.
13. Pawliszyn, J., Special Issue: Microextraction, Part I. [In: Journal of Chromatography Science; 2006,
44(6)], Preston Publications, Niles, IL, 2006.
14. Pawliszyn, J. B., Solid Phase Microextraction: Theory and Practice, VCH, New York, NY, 1997.
15. Saito, Y. and Jinno, K., Miniaturized sample preparation combined with liquid phase separations,
Journal of Chromatography, A, 1000, 53, 2003.
16. Theodoridis, G., Koster, E. H. M., and de Jong, G. J., Solid-phase microextraction for the analysis
of biological samples, Journal of Chromatography, B: Biomedical Sciences and Applications, 745,
49, 2000.
17. Vas, G. and Vekey, K., Solid-phase microextraction: A powerful sample preparation tool prior to mass
spectrometric analysis, Journal of Mass Spectrometry, 39, 233, 2004.
18. Belardi, R. P. and Pawliszyn, J. B., The application of chemically modified fused silica fibers in the
extraction of organics from water matrix samples and their rapid transfer to capillary columns, Water
Pollution Research Journal of Canada, 24, 179, 1989.
19. Li, S. and Weber, S. G., Determination of barbiturates by solid-phase microextraction and capillary
electrophoresis, Analytical Chemistry, 69, 1217, 1997.
20. Nguyen, A.-L. and Luong, J. H. T., Separation and determination of polycyclic aromatic hydrocarbons
by solid phase microextraction/cyclodextrin-modified capillary electrophoresis, Analytical Chemistry,
69, 1726, 1997.
21. Fan, X. and Deng, Y., Separation and identification of aromatic acids in soil and the Everglades sediment
samples using solid-phase microextraction followed by capillary zone electrophoresis, Journal of
Chromatography, A, 979, 417, 2002.
22. Jinno, K., Han, Y., Sawada, H., and Taniguchi, M., Capillary electrophoretic separation of toxic drugs
using a polyacrylamide-coated capillary, Chromatographia, 46, 309, 1997.
23. Kannamkumarath, S. S., Wuilloud, R. G., Jayasinghe, S., and Caruso, J. A., Fast speciation analysis of
iodophenol compounds in river waters by capillary electrophoresis-inductively coupled plasma-mass
spectrometry with off-line solid-phase microextraction, Electrophoresis, 25, 1843, 2004.
24. Hernandez-Borges, J., Cifuentes, A., Garcia-Montelongo, F. J., and Rodriguez-Delgado, M. A., Com-
bining solid-phase microextraction and on-line preconcentration-capillary electrophoresis for sensitive
analysis of pesticides in foods, Electrophoresis, 26, 980, 2005.
25. Hernandez-Borges, J., Rodriguez-Delgado, M. A., Garcia-Montelongo, F. J., and Cifuentes, A., Highly
sensitive analysis of multiple pesticides in foods combining solid-phase microextraction, capillary
electrophoresis-mass spectrometry, and chemometrics, Electrophoresis, 25, 2065, 2004.
26. Liu, Z. and Pawliszyn, J., Coupling of solid-phase microextraction and capillary isoelectric focus-
ing with laser-induced fluorescence whole column imaging detection for protein analysis, Analytical
Chemistry, 77, 165, 2005.
27. Whang, C.-W. and Pawliszyn, J., Solid phase microextraction coupled to capillary electrophoresis,
Analytical Communications, 35, 353, 1998.
28. Valenta, J. N., Sun, L., Ren, Y., and Weber, S. G., Solvatochromic study of poly(vinyl chloride)
plasticizers and their solutions in chloroform: application to phenobarbital partitioning and molecular
recognition of phenobarbital, Analytical Chemistry, 69, 3490, 1997.
29. Valenta, J. N. and Weber, S. G., Molecular recognition of phenobarbital in plasticizers. Equilibrium
investigations on the solubility of the barbiturate artificial receptor and its binding to phenobarbital in
plasticizers, Journal of Chromatography, A, 722, 47, 1996.
30. Zhang, X., Zhao, H., Chen, Z., Nims, R., and Weber, S. G., Effect of polymer concentration on
partitioning and molecular recognition in plasticized poly(vinyl chloride), Analytical Chemistry, 75,
4257, 2003.
31. Wei, F., Fan, Y., Zhang, M., and Feng, Y.-Q., Poly(methacrylic acid-ethylene glycol dimethacrylate)
monolith in-tube solid-phase microextraction applied to simultaneous analysis of some amphetamine
derivatives in urine by capillary zone electrophoresis, Electrophoresis, 26, 3141, 2005.
32. Zhou, X., Li, X., and Zeng, Z., Solid-phase microextraction coupled with capillary electrophoresis
for the determination of propranolol enantiomers in urine using a sol-gel derived calix[4]arene fiber,
Journal of Chromatography, A, 1104, 359, 2006.
33. Tong, W., Link, A., Eng, J. K., and Yates, J. R., III, Identification of proteins in complexes by
solid-phase microextraction/multistep elution/capillary electrophoresis/tandem mass spectrometry,
Analytical Chemistry, 71, 2270, 1999.
34. Dahlin, A. P., Bergstroem, S. K., Andren, P. E., Markides, K. E., and Bergquist, J.,
Poly(dimethylsiloxane)-based microchip for two-dimensional solid-phase extraction-capillary elec-
trophoresis with an integrated electrospray emitter tip, Analytical Chemistry, 77, 5356, 2005.
35. Jinno, K., Kawazoe, M., Saito, Y., Takeichi, T., and Hayashida, M., Sample preparation with fiber-
in-tube solid-phase microextraction for capillary electrophoretic separation of tricyclic antidepressant
drugs in human urine, Electrophoresis, 22, 3785, 2001.
36. Tempels, F. W. A., Wiese, G., Underberg, W. J. M., Somsen, G. W., and de Jong, G. J., On-line coupling
of size exclusion chromatography and capillary electrophoresis via solid-phase extraction and a Tee-
split interface, Journal of Chromatography, B: Analytical Technologies in the Biomedical and Life
Sciences, 839, 30, 2006.
37. Guzman, N. A., Improved solid-phase microextraction device for use in on-line immunoaffinity
38. Tuomikoski, S., Virkkala, N., Rovio, S., Hokkanen, A., Siren, H., and Franssila, S., Design and fabrica-
tion of integrated solid-phase extraction-zone electrophoresis microchip, Journal of Chromatography,
A, 1111, 258, 2006.
39. Stoyanov, A. V. and Pawliszyn, J., CE in a nonuniform capillary modulated by a cylindrical insert, and
zone-narrowing effects during sample injection, Analytical Chemistry, 75, 3324, 2003.
40. Liu, Z. and Pawliszyn, J., Microdialysis hollow fiber as a macromolecule trap for on-line coupling
of solid phase microextraction and capillary electrophoresis, Analyst (Cambridge, United Kingdom),
131, 522, 2006.
41. Li, S. and Weber, S. G., in Applications of Solid Phase Microextraction, J. Pawliszyn, Ed., Royal
Society of Chemistry, Cambridge, 49, 1999.
42. Li, S., Sun, L., Chung, Y., and Weber, S. G., Artificial receptor-facilitated solid-phase microextraction
of barbiturates, Analytical Chemistry, 71, 2146, 1999.
28 CE-SELEX: Isolating Aptamers
Using Capillary Electrophoresis
Renee K. Mosing and Michael T. Bowser
CONTENTS
28.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 825

28.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 826
28.3 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 827
28.3.1 CE-SELEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 827
28.3.2 Library Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 828
28.3.3 Sequence Length . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 829
28.3.4 Modified Nucleic Acid Libraries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 829
28.3.5 Target Concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 830
28.3.6 Capillary Inner Diameter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 830
28.3.7 Capillary Length . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 830
28.3.8 Capillary Coatings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 830
28.3.9 Buffer Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 831
28.3.10 Negative Selections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 831
28.4 Practical Application. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 831
28.5.1 Identification of the Collection Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 832
28.5.2 Fraction Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 833
28.5.3 PCR Amplification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 833
28.5.4 Purification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 834
28.5.5 Dissociation Constant Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 834
28.5.6 Cloning and Sequencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 835
28.5.7 Aptamer Characterization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 835
28.6 Advantages of CE-SELEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 836
28.7 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 836
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 836
28.1 INTRODUCTION
Since their discovery in 1990, aptamers have gained increasing attention due to their high affinity
and specificity for a wide range of target molecules.1–5 While aptamer applications are growing
exponentially, the time and difficulty of the process used to isolate aptamers are limiting widespread
adoption. Recently, the high resolving power of capillary electrophoresis (CE) has been used to select
high-affinity aptamers.6,7 This offers a number of advantages over conventional selection protocols
due to the dramatic enrichment rate and selection stringency made possible by CE. These advantages
result in a significantly faster and simpler process allowing high-affinity aptamers to be obtained
825
in several days rather than the weeks to months typical of pre-existing procedures. This chapter
discusses the significance of CE-SELEX and describes details of the selection process.
28.2 BACKGROUND
The word aptamer comes from the Latin root “aptus” meaning “to fit.” This is appropriate since
aptamers are single-stranded deoxyribonucleic acid (ssDNA) or ribonucleic acid (RNA) sequences
that fold into unique structures allowing them to bind target molecules with high affinity and selec-
tivity. Aptamers are typically 70–120 bases long, containing primer sequences on both ends to
facilitate polymerase chain reaction (PCR) amplification. As shown in Figure 28.1, depending on
their sequence, single-stranded nucleic acids can fold into a variety of loops, hairpins, and bulges to
generate a wide range of structures. The plethora of structures available make nucleic acids an attrac-
tive combinatorial library since sequences can be isolated with affinity for virtually any molecular
target.
While the large number of structures available in a library of even relatively short nucleic acids
allow aptamers to be isolated for a wide range of targets, it also raises the question of how to identify
high-affinity sequences out of so many possibilities. This was addressed in 1990 with the introduction
of a process for isolating aptamers that has been referred to as SELEX (Systematic Evolution of
Ligands by EXponential enrichment),8 in vitro selection9 or in vitro evolution.10 The process has been
described in detail in a number of excellent reviews.1–5 As illustrated in Figure 28.2, a structurally
diverse library containing approximately 1013 –1015 ssDNA or RNA molecules is incubated with the
target molecule. The nucleic acids contain a 30–80 base random sequence region flanked by two
primer regions. Sequences with affinity for the target are separated from nonbinding sequences using
filtration, affinity chromatography, or panning. Binding sequences are PCR amplified, purified, and
made single stranded to generate a new nucleic acid pool suitable for further rounds of enrichment.
This process is continued until the pool converges on a collection of sequences with high affinity for
the target. Typically, 8–15 selection cycles are required to generate a pool containing a significant
abundance of high-affinity aptamers. Individual aptamers are then cloned and sequenced from this
final pool for further characterization.
Aptamers have been obtained for hundreds of targets.1–5 Low nanomolar to picomolar disso-
ciation constants are typical for large protein targets. Aptamers of therapeutic interest have been
isolated for a number of drug targets including HIV components,11,12 the influenza virus surface
FIGURE 28.1 Secondary structure of a DNA aptamer selected using CE-SELEX to have affinity for
neuropeptide Y. (From Mendonsa, S. D., Bowser, M. T., J. Am. Chem. Soc., 127, 9382–9383, 2005.)
CE-SELEX: Isolating Aptamers Using Capillary Electrophoresis 827
Filter or affinity
selection
1013 ssDNA Unbound

sequences sequences
Binding
sequences
Target
PCR
amplify
Purify
FIGURE 28.2 Schematic representation of the SELEX process. A random sequence ssDNA library is incu-
bated with the target molecule. Binding sequences are selected using a filter or affinity separation, PCR amplified,
purified, and made single stranded, generating a new pool suitable for further cycles of enrichment. High-affinity
aptamers are typically obtained after 8–15 rounds of selection.
glycoprotein hemagglutinin,13,14 and the amyloid conformer of the prion protein.15,16 Macugen, a
treatment for age-related macular degeneration (AMD), is the first aptamer-based drug to gain Food
and Drug Administration (FDA) approval and is now available to the public.17 There are a number
of other promising therapeutic aptamers in clinical trials.18
Aptamers have an even more diverse role in diagnostics.19,20 Highly specific aptamers have been
used as affinity probes in CE21–23 and proteomic microarrays.24–28 Aptamers have been incorporated
into fluorescent switches and beacons.29,30 They have been incorporated into chromatography31–33
and capillary electrochromatography34–40 stationary phases. They have even been used as the sensing
element of cantilever-based detectors.41 Clearly, the combination of high affinity and selectivity
provided by aptamers can be applied to a wide range of analytical techniques.
The high affinity and selectivity of aptamers draw obvious comparisons with antibodies.19
Although similar in many respects, aptamers hold several advantages. Once their sequence is known,
aptamers are cheaply synthesized with little variation between batches. Modifications such as fluo-
rescent tags and non-natural bases can be easily incorporated into the synthesis. Isolating aptamers
does not require animals. Aptamers can therefore be obtained for molecules that are toxic or do not
stimulate an immune response. Aptamers are not restricted to use in physiological conditions. The
structure and affinity of aptamers can be easily manipulated by changing buffer conditions, heat, pH,
and so forth. Aptamers will often refold into their original structure once returned to nondenaturing
conditions. The relatively small size of aptamers allows them to penetrate tissues more easily, which
is attractive in therapeutic applications. Furthermore, their smaller size allows tighter packing and
higher loading capacities when attached to stationary surfaces in chromatography or microarrays.
28.3 THEORY
28.3.1 CE-SELEX
Recently, the high resolution power of CE has been used to isolate high-affinity aptamers (see
Table 28.1 and references therein). This process has been referred to as CE-SELEX.6,7 The
CE-SELEX process is illustrated in Figure 28.3. The target molecule is incubated with a nucleic
TABLE 28.1
Comparison of Aptamers Selected Using CE-SELEX with Those
Selected Using Conventional SELEX Protocols
CE-SELEX SELEX
Target Kd (nM) Cycles Kd (nM) Cycles

IgE 23 ± 127 4 653 15
HIVRT 0.18 ± 0.0711 4 154 12
Neuropeptide Y 300 ± 20042 4 37055,∗ 12
PFTase 0.545 1 n/a
mutS 3.6 ± 0.556 3 n/a
Kinase C-δ 12257 9 n/a —
Ricin 58 ± 1944 4 105 ± 4144 9
∗ RNA aptamer.
CE
Incubate + separation
with target
Unbound
1013 ssDNA
sequences
sequences
–
Binding
sequences
Target
PCR
Purify amplify
FIGURE 28.3 Schematic representation of the CE-SELEX process. A random sequence ssDNA library is
incubated with the target molecule. Binding sequences are separated from nonbinding sequences using CE, PCR
amplified, purified, and made single stranded, generating a new pool suitable for further cycles of enrichment.
High-affinity aptamers are typically obtained after 2–4 rounds of selection.
acid library. Several nanoliters of this incubation mixture are injected onto a capillary and separated
using free zone CE. Nonbinding oligonucleotides migrate through the capillary with the same mobil-
ity, regardless of their length or sequence. Complexing the target changes the size and/or charge of
binding sequences causing them to migrate as a separate fraction. These binding sequences are col-
lected as a separate fraction at the outlet of the capillary. As in conventional SELEX selections, the
fraction containing binding sequences is amplified, purified, and made single stranded for further
rounds of enrichment. The process is repeated until no further improvement in affinity is observed
between rounds.
28.3.2 LIBRARY SIZE

In general, as many independent sequences as possible should be included in the initial library.
This maximizes the probability that high-affinity aptamers are present in the initial pool. Presum-
ably, sequences with the highest affinity for the target are characterized by more complex, and
0.75 (a)
Absorbance (arbitrary units)

0.50
(b)
0.25
(c)
0.00
0 1 2 3 4 5 6 7 8
Time (min)
FIGURE 28.4 Effect of injection size on the peak shape of a 2.5 mM ssDNA library separated using CE.
Injection parameters were (a) 1 psi, 5 s; (b) 1 psi, 3 s; and (c) 0.5 psi, 3 s. Using a 1 psi, 3 s injection, 1.8 × 1013
sequences were loaded onto the capillary. The arrow shows the position on the electropherogram where collection
of binding sequences was stopped. CE conditions: TGK buffer, pH 8.3, UV detection at 254 nm, 50.2 cm, 50 µm
i.d., 360 µm o.d. capillary, 30 kV separation voltage. (From Mosing, R. K. et al., Analytical Chemistry, 77,
6107–6112, 2005.)
therefore rare structures than moderate binders with lower selectivity. The number of possible inde-
pendent sequences in an oligonucleotide library is given by 4n , where n is the number of bases in
the random region. Total coverage of all potential sequences becomes quickly unmanageable. For
example, a moderate length library containing a 60 base random region contains over 1036 inde-
pendent sequences. In conventional selections, 1013 –1015 sequences are often used in the initial
round of selection as a compromise between an adequate coverage of sequence space and practical
constraints. CE selection puts further constraints on library size. Injection volume in CE is limited
to several nL. Considering this, the concentration of the library must be in the low mM range to
inject 1013 sequences onto the capillary. This high concentration of ssDNA introduces significant
band broadening, making separation of aptamers from nonbinding sequences more difficult (see
Figure 28.4). In some cases, the concentration of the library must be lowered to ensure that bound
sequences are well separated from nonbinding sequences. Therefore, choosing the optimum library
size in CE-SELEX is a trade-off between maximizing sequence diversity and achieving adequate
resolution to separate aptamers from nonbinding sequences.
28.3.3 SEQUENCE LENGTH

Nucleic acid libraries are typically made up of 70–120 base ssDNA or RNA molecules. These
single-stranded oligonucleotides contain a 30–80 base random sequence region flanked by two
primer regions. The random region should contain at least 30 bases so that it is long enough to form
common structural motifs such as hairpins, bulges, pseudoknots, and g-quartets. Longer sequences
add diversity to the pool by providing more positions along the sequence where these motifs can be
located. The upper limit to sequence length is often determined by synthetic limitations. Yield and
purity drop off dramatically during synthesis of oligonucleotides longer than 100 bases.
28.3.4 MODIFIED NUCLEIC ACID LIBRARIES

Fluorescently labeled libraries can be used to facilitate detection during selections. Laser-induced
fluorescence (LIF) detection is orders of magnitude more sensitive than UV-absorbance detection
making it much easier to reliably observe peaks for the unbound sequences during selections. Flu-
orescent labels are easily incorporated at the 5 end of ssDNA during initial synthesis. Fluorescent
labels can be incorporated into sequences generated during the selection process by using a 5 labeled
forward primer during PCR.
If the final application of the aptamer includes attaching a fluorescent label (e.g., an affinity
sensor), it is advantageous to incorporate the label during selection. Postselection modification of
the aptamer could modify its structure, and consequently affect affinity or selectivity.
28.3.5 TARGET CONCENTRATION

An advantage of CE-SELEX is the flexibility to manipulate the selection stringency by varying the
target concentration in the incubation mixture. Target concentrations as low as 1 pM have successfully
been used in CE-SELEX selections.6 Considering the injection volume this resulted in 1013 ssDNA
molecules competing for binding sites on 10,000 target molecules. This level of stringency is difficult
to obtain using conventional macroscale separations. More recent selections have used higher target
concentrations.7,11,42 There is some concern that extreme competition promotes the selection of
aptamers that bind target sites that do not optimally bind nucleic acids.
28.3.6 CAPILLARY INNER DIAMETER

One of the easiest ways to increase the injection volume, and therefore number of sequences, is to
increase the inner diameter of the CE separation capillary. This increase in loading capacity must
be balanced with the potential for Joule heating. Joule heating is generally avoided in CE due to
the additional band broadening it causes. This is not the major concern in CE-SELEX since peaks
are already broad due to destacking caused by the ionic strength of the nucleic acid library. In CE-
SELEX, the concern is that Joule heating can change or even eliminate the structures that allow
aptamers to bind their targets. Melting temperatures as low as 40◦ C are not uncommon for some
of the structural motifs found in aptamers. Therefore, strict temperature control and elimination of
Joule heating are critical for selections to be successful.
28.3.7 CAPILLARY LENGTH

Capillary length should be optimized in much the same way as in typical CE separations. An additional
factor to consider is that longer capillaries force the complex to stay intact for a longer period of
time before collection. Note that even if equilibrium is reached during incubation before injection, as
soon as the separation begins the target–aptamer complex migrates into buffer that does not contain
the uncomplexed target. Therefore, the off-rate must be on the order of minutes for the complex
to migrate through the capillary intact. This can be used to increase the stringency of the selection.
Longer capillaries will promote selection of aptamers with slow off rates, which generally have
higher affinity for the target.
28.3.8 CAPILLARY COATINGS

Coatings are typically used in CE to prevent analyte interactions with the capillary wall.43 These wall
interactions can give rise to severe peak tailing, degrading resolution. In CE-SELEX, interactions
between the oligonucleotide library and the capillary wall are minimal since both are negatively
charged. Interactions between the target and the wall are more common. These should be avoided
since it is unclear how adsorption onto the capillary surface will affect binding sites on the target.
Typically, a CE analysis of the target alone is performed before attempting a selection. If a sym-
metrical peak is observed, this indicates that interactions between the target and capillary wall are
minimal. Tailing would suggest that wall interactions are taking place and that buffer conditions or
the capillary surface should be modified.
28.3.9 BUFFER COMPOSITION

There are a number of important considerations when choosing the incubation and separation buffers.
The end use of the aptamer is most important when choosing the incubation buffer. The selection
protocol enriches sequences that bind best under these specific conditions. There is no guarantee
that the selected aptamers will bind the target under different conditions. Therefore, the incubation
buffer should mimic the conditions that the aptamer will be used in as closely as possible. The
separation buffer should be similar to the incubation buffer, taking constraints of the CE separation
into consideration. High ionic strength buffers give rise to Joule heating and should be avoided.
Divalent ions such as Ca2+ and Mg2+ interact strongly with the fused-silica walls of uncoated
capillaries, affecting electroosmotic flow (EOF). The incorporation of buffer additives to modify the
separation should also be avoided since this introduces the possibility of selecting aptamers with
affinity for the additive, not the target.
28.3.10 NEGATIVE SELECTIONS

Great care must be taken when performing conventional filter or chromatography-based SELEX to
avoid isolating aptamers with affinity for the stationary surfaces to which they are exposed to during
selection. Negative selections are often necessary to prevent this. Before incubation with the target
the library is passed through a filter or the raw stationary support to eliminate sequences that have
any background affinity for these stationary surfaces. While effective, this procedure decreases the
diversity of the nucleic acid pool before selection even begins.
Negative selections have not proven to be necessary in CE-SELEX selections. No aptamers have
been identified that exhibit affinity for the capillary surface. This has been equally true for selections
performed using uncoated or neutrally coated capillaries.6,7,11,42 This greatly simplifies the SELEX
procedure and removes one of the significant pitfalls of aptamer selection.
28.4 PRACTICAL APPLICATION

Mendonsa et al.6,7 were the first to isolate aptamers using CE-SELEX. They used CE to isolate
aptamers with low nM dissociation constants for immunoglobulin E (IgE) after as few as two rounds
of selection. This was a significant improvement over conventional selections, which typically take
8–15 rounds. Since then CE-SELEX has been used to isolate aptamers for a range of targets including
human immunodeficiency virus reverse transcriptase (HIV-RT),11 neuropeptide Y (NPY),42 ricin,44
protein famesyltransferase (PFTase),45 and MutS46 (see Table 28.1). Berezovski et al.45 were able
to isolate aptamers with subnanomolar dissociation constants for PFTase after a single round of
CE-SELEX selection, demonstrating the extreme enrichment provided by CE.
Aptamers selected using CE-SELEX exhibit similar if not better affinity for their targets than
aptamers selected using conventional methods (see Table 28.1). Mosing et al.11 used CE-SELEX
to isolate aptamers with dissociation constants for HIV reverse transcriptase as low as 180 pM,
fivefold better than aptamers selected using nitrocellulose filtration. Tang et al.44 directly compared
CE-SELEX with conventional selections by isolating aptamers for ricin using both methods. They
found that after four rounds of CE-SELEX 87.2% of the nucleic acid pool bound ricin while even
after nine rounds of conventional selection only 38.5% of the pool bound target.
Target size is a potential limitation to CE-SELEX. The target must be large enough to significantly
shift the mobility of the aptamer upon binding for fraction collection to be successful. To test this
limitation, Mendonsa and Bowser6 used CE-SELEX to isolate aptamers for NPY. NPY is a 36
amino acid peptide with a molecular weight of 4272 g/mol, significantly smaller than the ssDNA
in the nucleic acid library (∼25 kDa). Aptamers with 300 nM dissociation constants for NPY were
isolated, similar to RNA aptamers previously selected using conventional methods. These results
demonstrated for the first time that CE-SELEX can be used to select aptamers for targets smaller
than the library itself. It is important to note that the target did not need to be attached to a stationary
phase or a large affinity tag as is necessary to perform conventional SELEX selections against small
molecules.

28.5.1 IDENTIFICATION OF THE COLLECTION WINDOW
For CE-SELEX experiments to be successful, there must be adequate separation between the
aptamer–target complex and nonbinding sequences. Initial experiments should be performed to
assess the probability that resolution will be sufficient and determine the migration window of the
analyte–target complex. This is complicated by the low abundance of binding sequences in early
rounds of selection and the low concentration of target used in selections. Both these factors pre-
vent direct detection of aptamer–target complex during selections. Before attempting selections, the
nucleic acid library and target should be analyzed by CE separately to determine their migration
times. In almost all cases the aptamer–target complex will migrate at a time between the target and
unbound oligonucleotides. For large targets it can be assumed that the complex will migrate closer
to the free target peak. Note that in almost every case the aptamer complex will have a mobil-
ity that is less negative than the unbound sequences (see Figure 28.5). It is difficult to imagine
many targets with a more negative charge density than ssDNA or RNA that would give rise to an
aptamer–target complex with a more negative mobility than unbound oligonucleotides. If the target
and unbound sequences are not separated then it is unlikely that the aptamer–target complex will be
sufficiently resolved for successful fraction collection. In this case separation buffer, voltage, injec-
tion time, capillary coatings, capillary length, and/or inner diameter should be modified to achieve
resolution.
(a) (b)
Unbound
Collection Unbound ssDNA Collection
window ssDNA window
Intensity
Intensity
1-2 min
1–2 min
Detection Detection
Outlet window Inlet Outlet window Inlet
– + + –
ᐍD ᐍD
ᐍT ᐍT
FIGURE 28.5 Schematics illustrating the procedure for fraction collection in CE-SELEX. (a) When using an
uncoated capillary that generates EOF the aptamer–target complex will migrate earlier than the unbound ssDNA.
Equation 28.2 should be used to calculate the time required for an analyte that reached the detection window
1–2 min before the leading edge of the unbound ssDNA peak to reach the outlet of the separation capillary.
(b) When using a coated capillary that does not generate EOF the aptamer–target complex will migrate after
the unbound ssDNA. Equation 28.2 should be used to calculate the time required for an analyte that reached
the detection window 1–2 min after the trailing edge of the unbound ssDNA peak to reach the outlet of the
separation capillary.
28.5.2 FRACTION COLLECTION

Once an adequate separation is achieved, and the migration window of the aptamer–target complex
has been identified, CE-SELEX selections can begin. Before selection the nucleic acid library is
heated to 72◦ C for 2 min and allowed to cool to room temperature. This ensures that all sequences
fold into their stable room temperature conformations. The library is then incubated with the target
(50–500 pM) for 20 min at room temperature to allow binding to occur. It is important to keep the
target concentration lower than that of the nucleic acid pool to promote competition for binding sites.
An aliquot of the incubation mixture is injected onto the CE capillary and the separation voltage is
applied.
The strategy for fraction collection is shown in Figure 28.5. Figure 28.5a shows a typical selection
using an uncoated capillary in the presence of EOF. EOF pulls the negatively charged nucleic acid
library toward the cathode. The aptamer–target complex is generally less negative and therefore
will migrate earlier than the unbound sequences. All sequences that migrate before the unbound
nucleic acid peak should therefore be collected. Sequences are collected during the CE separation
directly into the outlet vial containing a small volume (∼50 µL) of separation buffer. The volume
of buffer in the collection vial should be small to facilitate PCR amplification after fractions have
been collected. To account for run-to-run variability, the collection window should be set to end a
certain period of time (e.g., 1–2 min) before the leading edge of the unbound sequences migrates
off the end of the capillary. This time can be adjusted based on the resolution anticipated between
the aptamer–target complex and the unbound sequences. As shown in Figure 28.5a, the on column
detection of CE facilitates this fraction collection strategy. As there is a length of capillary after the
detection window, it takes a certain period of time for analytes to reach the outlet of the capillary
after they have been detected. The time that an analyte will reach the outlet of the capillary (tout ) is
given by
T
tout = (tdet ) , (28.1)
D
where tdet is the time required for the analyte to reach the detector, T is the total length of the
capillary, and D is the capillary length to the detector. To account for variation in migration time
between CE separations the separation is monitored until the front of the peak corresponding to the
unbound sequences is observed. Equation 28.1 can then be used to calculate the time it would take
an analyte that reached the detector 1–2 min earlier to reach the outlet of the capillary. The separation
is stopped at this time. The vial at the capillary outlet is removed and any sequences remaining on
the capillary are washed to waste using a pressure rinse.
Figure 28.5b illustrates the fraction collection procedure when using a coated capillary that
eliminates EOF. The polarity of this separation has been reversed since in the absence of EOF
nucleic acids will migrate toward the anode. The aptamer–target complex will migrate after the
unbound oligonucleotides if it is less negative. In this case, the separation is monitored until the end
of the peak corresponding to the unbound sequences is observed at the detector. Equation 28.1 is used
to estimate the time it would take an analyte that reached the detector 1–2 min later than the trailing
end of this peak to reach the outlet of the separation capillary. At this time the separation voltage
is turned off and the vial at the capillary outlet is replaced with a collection vial containing a small
volume (∼10–50 µL) of PCR buffer. A pressure rinse is used to collect any sequences remaining on
the capillary into this vial.
28.5.3 PCR AMPLIFICATION

After fraction collection, binding sequences are PCR amplified using standard procedures. Care must
be taken during this amplification step due to the extremely low concentration of ssDNA collected.
The precise melting, annealing, and extension temperatures are determined by the lengths of the
primers and the sequences in the library. The reverse primer should be biotinylated at the 5 end
to facilitate the purification procedure (described in the following section). If the initial library was
fluorescently labeled, the forward primer should be similarly labeled at the 5 end.
Note that PCR facilitates release of even high-affinity aptamers from their targets. The first step of
the PCR process is a melting stage where the sequences are brought to ∼90◦ C. This melting eliminates
all secondary structure from the ssDNA, removing the aptamer’s affinity for the target. Once released,
the target is quickly diluted in the PCR buffer, making reformation of the aptamer–target complex
unlikely.
Successful PCR amplification can be confirmed using standard gel electrophoresis protocols
(2% agarose, ethidium bromide). A single solid band of the proper length with little smearing should
be observed. Careful controls should be carried out to ensure that the amplified DNA is from the
collection and not the result of contamination.
28.5.4 PURIFICATION
PCR amplification generates double-stranded DNA (dsDNA). ssDNA must be recovered for binding
sequences to fold into their proper binding structures. Complementary sequences are biotinylated
during PCR to facilitate ssDNA recovery. The dsDNA is passed through a streptavidin column.
After a short incubation period, the column is rinsed with streptavidin binding buffer (10 × 500 µL,
10 mM Tris, 50 mM NaCl, and 1 mM ethylenediamine tetraacetic acid, EDTA) to remove excess
PCR reagents. The column is then heated to 37◦ C in the presence of 0.15 M NaOH to separate the
dsDNA without disrupting the biotin–streptavidin complex. The single-stranded forward sequences
are eluted from the column while the complementary sequences remain attached to the stationary
phase through the biotin–streptavidin interaction.
The ssDNA is immediately neutralized (0.15 M acetic acid) to avoid degradation: 40 µL of 3 M
sodium acetate and 1 mL of 100% ice-cold ethanol are added to initiate ethanol precipitation to
further purify and concentrate the ssDNA. The solution is incubated at –80◦ C for 1 h to ensure
complete precipitation. The solution is centrifuged at 14,000 rpm for 15 min at 4◦ C to pellet the
ssDNA precipitate. The supernatant is discarded leaving approximately 50 µL at the bottom of the
vial. The following steps are repeated three times to wash the ssDNA pellet:
1. Add 1 mL of ice cold 70:30 ethanol/water

2. Centrifuge 14,000 rpm for 25 min at 4◦ C
3. Discard the supernatant leaving approximately 50 µL of solution remaining in the bottom
of the vial
The samples are then dried in a speed vac at 60◦ C for 25 min or until dry. The resulting DNA
pellet is resuspended in 30 µL of buffer, generating a new ssDNA pool suitable for further rounds
of enrichment. In the next selection cycle 10–15 µL is used. The remaining portion is used to assess
the affinity of the pool for the target. An aliquot is also stored for further characterization or cloning
and sequencing once the selection process is complete.
28.5.5 DISSOCIATION CONSTANT ESTIMATION

The affinity of the nucleic acid pool for the target is measured after every round of selection to
monitor the progress of the enrichment. Affinity of the library for the target should increase as the
selection progresses. Selection cycles should continue until no further improvement in affinity is
observed between rounds. Dissociation constants in the low nM to pM range are typically observed
at the completion of the CE-SELEX process.
Dissociation constants can be estimated using affinity capillary electrophoresis (ACE).47 In
these experiments, a constant concentration of the nucleic acid pool is titrated with increasing
concentrations of target. Use of fluorescently labeled sequences is recommended since LIF detection
allows detection of much lower concentrations. After incubation, each solution is analyzed using
CE. The intensity of the peak corresponding to the unbound sequences decreases as the target
concentration increases and the equilibrium is shifted toward the aptamer–target complex. If the
condition that the nucleic acid concentration is much lower than the target concentration is met, the
dissociation constant (Kd ) can be determined by fitting the peak intensities to the following equation:
I0 − I c
= , (28.2)
I0 Kd + [target]
where I0 is the height of the unbound nucleic acid peak in the absence of target, I is the height of the
unbound nucleic acid peak in the presence of target, c is a constant, and [target] is the concentration
of the target.
28.5.6 CLONING AND SEQUENCING

Once the affinity of the nucleic acid pool has plateaued, individual sequences can be cloned and
sequenced from the pool for further characterization. The final pool is PCR amplified for eight
cycles to ensure that all sequences are double stranded. Primers should not be 5 labeled with biotin
or a fluorescent tag since this will prevent ligation of the sequences into the cloning vector. dsDNA
is ligated into a pGEM vector and transfected into DH5 Escherichia coli and colonies are raised.7,11
Plasmids from approximately 30 (or more) colonies are chosen at random and the sequences of
individual clones are determined using the T7 promoter sequence.
28.5.7 APTAMER CHARACTERIZATION

Once the sequences of individual aptamers are obtained, it is relatively inexpensive to have sufficient
quantities for further characterization synthesized. In conventional selections, sequence patterns or
motifs are commonly observed in the aptamers recovered from cloning. This suggests that the
selection has converged on certain sequence elements that are important for binding the target.
Software programs such as ClustalW are readily available for identifying these motifs in large
numbers of sequences.48 Only the random region of the sequences should be entered when using
these programs. Otherwise, the homology of the primer regions will dominate the analysis. Contrary
to conventional selections, sequence motifs are rare in aptamers selected using CE-SELEX.6,7,11,42
This suggests that the heterogeneity of the pool remains high even after the selection has converged
on a collection of high-affinity aptamers.
Common structural elements can be identified using mFold, a program that predicts secondary
structure of ssDNA or RNA.49 For this analysis, it is necessary to include the primer regions
since they most likely contribute to the overall structure of the nucleic acid. This analysis is use-
ful in identifying common structural elements in aptamers even if they are generated by differing
sequences.
Sequencing and synthesis allow the affinities of individual aptamers to be determined. Dis-
sociation constants can be estimated using ACE as described above. Affinity of the aptamer
should be compared with that of the unselected library or a nonsensical sequence such as polyT
as a control experiment. Selectivity should be assessed by measuring the affinity of the aptamer
for molecules similar to the target. Affinity should be measured under a range of conditions
and in the presence of potential interferents depending on whether the aptamer is to be used
as a therapeutic or diagnostic agent. Activity assays should be performed if the aptamer was
designed to inhibit the function of the target to determine if the aptamer interferes with the
active site.
28.6 ADVANTAGES OF CE-SELEX

CE-SELEX provides a number of benefits over traditional SELEX methods. The high resolving
power and efficiency of CE generates high rates of enrichment allowing aptamers to be isolated after
2–4 rounds of CE-SELEX selection. Conventional selections typically require 8–15 selection cycles,
a process that can take weeks to months to complete. CE-SELEX performs selections in free solution,
dramatically reducing the potential for nonspecific interactions with stationary surfaces. CE-SELEX
generates a large number of independent aptamer sequences, producing many lead compounds for
therapeutic or diagnostic applications. The abundance of high-affinity sequences in the final nucleic
acid pool is nearly 100%. In comparison, it is not uncommon for half of the sequences selected
using conventional techniques to not demonstrate affinity for the target. Finally, CE-SELEX is much
more compatible with libraries that incorporate non-natural nucleic acids to improve aptamer sta-
bility. These non-natural nucleic acids often contain hydrophobic functional groups, which promote
nonspecific interactions with hydrophobic filter materials, such as nitrocellulose, commonly used in
conventional selections.
28.7 FUTURE WORK

CE-SELEX has been used successfully to isolate aptamers for a variety of targets (see Table 28.1).
While these initial proofs of concept experiments are promising, a more detailed study of the selection
process is now needed. CE-SELEX has made the selection process dramatically faster. This will
allow fundamental experiments assessing the effect of variables such as library concentration, target
concentration, and incubation time to be performed for the first time. The 4–6 weeks that are necessary
to complete a conventional selection have made these experiments unfeasible until now.
CE-SELEX opens the door to further advances in automating the selection process. Sev-
eral researchers have developed automated SELEX instruments based on conventional selection
protocols.50–52 These experiments have demonstrated that automation greatly improves the through-
put of aptamer selection. The electrophoretic selection of CE-SELEX suggests that development of
an automated microfluidic SELEX device may be possible. Considering the dramatic improve-
ments in the time required for separation, amplification, and purification, which can be obtained
by performing these processes on the microscale, it is not difficult to imagine a fully auto-
mated microfluidic device capable of performing the entire SELEX process in as little as several
hours.
REFERENCES
1. Bunka, D. H. J., Stockley, P. G., Aptamers come of age—At last, National Reviews Microbiology, 4,
588–596, 2006.
2. Klug, S. J., Famulok, M., All you wanted to know about SELEX, Molecular Biology Reports, 20,
97–107, 1994.
3. Famulok, M., Mayer, G., Blind, M., Nucleic acid aptamers from selection in vitro to applications in
vivo, Accounts of Chemical Research, 33, 591–599, 2000.
4. Osborne, S. E., Ellington, A. D., Nucleic acid selection and the challenge of combinatorial chemistry,
Chemical Reviews, 97, 349–370, 1997.
5. Forst, C. V., Molecular evolution: A theory approaches experiments, Journal of Biotechnology, 64,
101–118, 1998.
6. Mendonsa, S. D., Bowser, M. T., In vitro evolution of functional DNA using capillary electrophoresis,
Journal of American Chemical Society, 126, 20–21, 2004.
7. Mendonsa, S. D., Bowser, M. T., In vitro selection of high-affinity DNA ligands for human IgE using
capillary electrophoresis, Analytical Chemistry, 76, 5387–5392, 2004.
8. Tuerk, C., Gold, L., Systematic evolution of ligands by exponential enrichment: RNA ligands to
bacteriophage T4 DNA polymerase, Science, 249, 505–510, 1990.
9. Ellington, A. D., Szostak, J. W., In vitro selection of RNA molecules that bind specific ligands, Nature,
346, 818–822, 1990.
10. Joyce, G. F., Amplification, mutation, and selection of catalytic RNA, Gene, 82, 83–87, 1989.
11. Mosing, R. K., Mendonsa, S. D., Bowser, M. T., Capillary electrophoresis-SELEX selection of
aptamers with affinity for HIV-1 reverse transcriptase, Analytical Chemistry, 77, 6107–6112, 2005.
12. Dey, A. K., Khati, M., Tang, M., Wyatt, R., Lea, S. M., James, W., An aptamer that neutralizes R5
strains of human immunodeficiency virus type 1 blocks gp120-CCR5 interaction, Journal of Virology,
79, 13806–13810, 2005.
13. Jeon, S. H., Kayhan, B., Ben-Yedidia, T., Arnon, R., A DNA aptamer prevents influenza infection by
blocking the receptor binding region of the viral hemagglutinin, Journal of Biological Chemistry, 279,
48410–48419, 2004.
14. Misono, T. S., Kumar, P. K., Selection of RNA aptamers against human influenza virus hemagglutinin
using surface plasmon resonance, Analytical Biochemistry, 342, 312–317, 2005.
15. Proske, D., Blank, M., Buhmann, R., Ansgar, R., Aptamers—Basic research, drug development, and
clinical applications, Applied Microbiology and Biotechnology, 69, 367–374, 2005.
16. Rhie, A., Kirby, L., Sayer, N., Wellesley, R., Disterer, P., Sylvester, I., Gill, A., Hope, J., James, W.,
Tahiri-Alaoui, A., Characterization of 2 -fluoro-RNA aptamers that bind preferentially to disease
associated conformations of prion protein and inhibit conversation, Journal of Biological Chemistry,
278, 39697–39705, 2003.
17. Siddiqui, M. A. A., Keating, G. M., Pegaptanib: In exudative age-related macular degeneration, Drugs,
65, 1571–1577, 2005.
18. Lee, J. F., Stovall, G. M., Ellington, A. D., Aptamer therapeutics advance, Current Opinion in Chemical
Biology, 10, 282–289, 2006.
19. Jayasena, S. D., Aptamers: An emerging class of molecules that rival antibodies in diagnostics, Clinical
Chemistry, 45, 1628–1650, 1999.
20. Tombelli, S., Minunni, M., Mascini, M., Analytical applications of aptamers, Biosensors and
Bioelectronics, 20, 2424–2434, 2005.
21. German, I., Buchanan, D. D., Kennedy, R. T., Aptamers as ligands in affinity probe capillary
electrophoresis, Analytical Chemistry, 70, 4540–4545, 1998.
22. Haes, A. J., Giordano, B. C., Collins, G. E., Aptamer-based detection and quantitative analysis of ricin
using affinity probe capillary electrophoresis, Analytical Chemistry, 78, 3758–3764, 2006.
23. Pavski, V., Le, X. C., Detection of human immunodeficiency virus type 1 reverse transcriptase
using aptamers as probes in affinity capillary electrophoresis, Analytical Chemistry, 73, 6070–6076,
2001.
24. Collett, J. R., Cho, E. J., Ellington, A. D., Production and processing of aptamer microarrays, Methods,
37, 4–15, 2005.
25. Cho, E. J., Collett, J. R., Szafranska, A. E., Ellington, A. D., Optimization of aptamer microarray
technology for multiple protein targets, Analytica Chimica Acta, 564, 82–90, 2006.
26. Bock, C., Coleman, M., Collins, B., et al., Photoaptamer arrays applied to multiplexed proteomic
analysis, Proteomics, 4, 609–618, 2004.
27. Golden, M. C., Collins, B. D., Willis, M. C., Koch, T. H., Diagnostic potential of PhotoSELEX-evolved
ssDNA aptamers, Journal of Biotechnology, 81, 167–178, 2000.
28. Petach, H., Gold, L., Dimensionality is the issue: Use of photoaptamers in protein microarrays, Current
Opinion in Biotechnology, 13, 309–314, 2002.
29. Matthew Levy, S. F. C. A. D. E., Quantum-dot aptamer beacons for the detection of proteins,
ChemBioChem, 6, 2163–2166, 2005.
30. Cao, Z., Suljak, S. W., Tan, W., Molecular beacon aptamers for protein monitoring in real-time and in
homogeneous solutions, Current Proteomics, 2, 31–40, 2005.
31. Brumbt, A., Ravelet, C., Grosset, C., Ravel, A., Villet, A., Peyrin, E., Chiral stationary phase based
on a biostable l-RNA aptamer, Analytical Chemistry, 77, 1993–1998, 2005.
32. Ravelet, C., Grosset, C., Peyrin, E., Liquid chromatography, electrochromatography and capillary
electrophoresis applications of DNA and RNA aptamers, Journal of Chromatography A, 1117, 1–10,
2006.
33. Ravelet, C., Boulkedid, R., Ravel, A., Grosset, C., Villet, A., Fize, J., Peyrin, E., A l-RNA aptamer
chiral stationary phase for the resolution of target and related compounds, Journal of Chromatography
A, 1076, 62, 2005.
34. Kotia, R. B., Li, L., McGown, L. B., Separation of nontarget compounds by DNA aptamers, Analytical
Chemistry, 72, 827–831, 2000.
35. Vo, T. U., McGown, L. B., Selectivity of quadruplex DNA stationary phases toward amino acids in
homodipeptides and alanyl dipeptides, Electrophoresis, 25, 1230–1236, 2004.
36. Charles, J. A. M., McGown, L. B., Separation of Trp-Arg and Arg-Trp using G-quartet-forming DNA
oligonucleotides in open-tubular capillary electrochromatography, Electrophoresis, 23, 1599–1604,
2002.
37. Rehder, M. A., McGown, L. M., Open-tubular capillary electrophoresis electrochromatography of
bovine beta-lactoglobulin variants A and B using an aptamer stationary phase, Electrophoresis, 22,
3759–3764, 2001.
38. Rehder-Silinski, M. A., McGown, L. B., Capillary electrochromatographic separation of bovine milk
proteins using a G-quartet DNA stationary phase, Journal of Chromatography A, 1008, 233, 2003.
39. Vo, T. U., McGown, L. B., Effects of G-quartet DNA stationary phase destabilization on fibrinogen
resolution in capillary electrochromatography, Electrophoresis, 27, 749–756, 2006.
40. Deng, Q., German, I., Buchanan, D., Kennedy, R. T., Retention and separation of adenosine and
analogous by affinity chromatography with an aptamer stationary phase, Analytical Chemistry, 73,
5415–5421, 2001.
41. Navani, N. K., Li, Y., Nucleic acid aptamers and enzymes as sensors, Current Opinion in Chemical
Biology, 10, 272, 2006.
42. Mendonsa, S. D., Bowser, M. T., In vitro selection of aptamers with affinity for neuropeptide Y using
capillary electrophoresis, Journal of American Chemical Society, 127, 9382–9383, 2005.
43. Judit Horvath, V. D., Polymer wall coatings for capillary electrophoresis, Electrophoresis, 22, 644–655,
2001.
44. Tang, J., Xie, J., Shao, N., Yan, Y., The DNA aptamers that specifically recognize rixin toxin are
selected by two in vitro selection methods, Electrophoresis, 27, 1303–1311, 2006.
45. Berezovski, M., Drabovich, A., Krylova, S. M., Musheev, M., Okhonin, V., Petrov, A., Krylov, S. N.,
Nonequilibrium capillary electrophoresis of equilibrium mixtures: A universal tool for development
of aptamers, Journal of the American Chemical Society, 127, 3165–3171, 2005.
46. Drabovich, A. P., Berezovski, M., Okhonin, V., Krylov, S. N., Selection of smart aptamers by methods
of kinetic capillary electrophoresis, Analytical Chemistry, 78, 3171–3178, 2006.
47. Østergaard, J., Heegaard, N. H. H., Bioanalytical interaction studies executed by preincubation affinity
48. Chenna, R., Sugawara, H., Koike, T., Lopez, R., Gibson, T. J., Higgins, D. G., Thompson, J. D.,
Multiple sequence alignment with the clustal series of programs, Nucleic Acids Research, 31, 3497–
3500, 2003.
49. Zuker, M., Mfold web server for nucleic acid folding and hybridization prediction, Nucleic Acids
Research, 31, 3406–3415, 2003.
50. Cox, J. C., Hayhurst, A., Hesselberth, J., Bayer, T. S., Georgiou, G., Ellington, A. D., Automated
selection of aptamers against protein targets translated in vitro: From gene to aptamer, Nucleic Acids
Research, 30, e108, 2002.
51. Cox, J. C., Rudolph, P., Ellington, A. D., Automated RNA selection, Biotechnology Progress, 14,
845–850, 1998.
52. Hybarger, G., Bynum, J., Williams, R. F., Valdes, J. J., Chambers, J. P., A microfluidic SELEX
prototype, Analytical and Bioanalytical Chemistry, 384, 191–198, 2006.
53. Wiegand, T. W., Williams, P. B., Dreskin, S. C., Jouvin, M. H., Kinet, J. P., Tasset, D., High-affinity
oligonucleotide ligands to human IgE inhibit binding to Fce receptor I, Journal of Immunology, 157,
221–230, 1996.
54. Schneider, D. J., Feigon, J., Hostomsky, Z., Gold, L., High-affinity ssDNA inhibitors of the reverse
transcriptase of type 1 human immunodeficiency virus, Biochemistry, 34, 9599–9610, 1995.
55. Proske, D., Hofliger, M., Soll, R. M., Beck-Sickinger, A. G., Famulok, M., A Y2 receptor mimetic
aptamer directed against neuropeptide Y, Journal of Biological Chemistry, 277, 11416–11422, 2002.
56. Drabovich, A., Berezovski, M., Krylov, S. N., Selection of smart aptamers by equilibrium capillary
electrophoresis of equilibrium mixtures, Journal of the American Chemical Society, 127, 11224–11225,
2005.
57. Mallikaratchy, P., Stahelin, R. V., Cao, Z., Tan, W., Selection of DNA ligands for protein kinase
C-omega, Chemical Communications, 3229–3231, 2006.
29 Microfluidic Technology as a
Platform to Investigate
Microcirculation
Dana M. Spence
CONTENTS
29.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 841

29.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 841
29.2.1 Components of the Circulation Investigated with Microfluidic Techniques. . . . . . 842
29.2.1.1 Erythrocytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 842
29.2.1.2 Platelets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 843
29.2.1.3 Endothelial Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 844
29.3.1 Unique Flow Properties of the Bloodstream . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 844
29.4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 845
29.4.1 Disease Biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 845
29.4.2 High-Throughput Drug Efficacy Investigations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 846
29.4.3 Measuring Cell-to-Cell Communication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 846
29.5 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 847
29.5.1 Cell Harvesting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 847
29.5.3 Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 849
29.6 Future Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 849
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 850
29.1 INTRODUCTION
Of the numerous applications of capillary- or microflow-based analyses, especially those involving
biological assays, there perhaps is none more fitting than employing microflow techniques to study
the circulatory system. In fact, the dimensions of the microcirculation so closely match those of
microbore capillary tubing and the channels in microfluidic devices, it is rather amazing that such
analytical tools and devices have not been in use to study the events in the bloodstream to a greater
extent. Here, a brief introduction to the microcirculation is given such that the nonexpert is able
to understand better why microflow techniques are an excellent tool to study in vivo processes that
occur in the bloodstream. Next, real examples will be given in order to demonstrate how such devices
are effective as tools to increase our understanding of the processes occurring in the bloodstream.
29.2 BACKGROUND
One of the key features of employing microbore tubing in a microflow system or a microfluidic device
to study events occurring in the microcirculation is that the studies can be performed in a controlled
841
environment. That is, in vivo, the circulation is very dynamic. Not only do vessel diameters change
(due to vasodilation) but also do the linear rates of the bloodstream, the viscosity of the blood, and
hematocrit.1 There are also slight changes in pH,2 oxygen tensions,3 and concentrations of molecules
in the bloodstream. Finally, there are also changes that occur at the cellular level that can affect the
overall cellular function. Common examples are cell aging and cellular oxidant status that can affect
the cell’s membrane properties.4
The use of a microflow system enables quantitative determinations to be performed in an envi-
ronment where the many variables mentioned above can be controlled.5 An excellent example is the
linear rate of the blood flow. In vivo, it has been shown recently that red blood cells (RBCs) have the
ability to control vessel diameter due to their ability to release adenosine triphosphate (ATP) upon
shear-induced deformation.6–8 This ATP is a recognized stimulus of nitric oxide (NO) production in
certain cell types including the endothelium9 and platelets.10 NO is a potent vasodilator, resulting
in the increase of vessel diameter and an increase in blood flow.11–13 The ability to study the effect
of linear rate on the release of ATP from RBCs would be difficult in an animal model because, as
the ATP is released (resulting in subsequent NO-induced vasodilation), the vessel diameter would
increase, thus changing the overall linear rate of the blood flow through the vessel. By performing
such studies on a controlled in vitro platform (microbore tubing), the ATP can be measured rather
easily via chemiluminescence, but the linear rate will remain constant (because microbore tubing
does not dilate).
Another advantage of performing measurements involving the bloodstream using microflow tech-
nologies is that blood is a complex sample and, as such, often requires some type of preparation prior
to the measurement portion of the analysis. Examples of such preparation schemes using microfluidic
devices have been described for automated, high-speed genotyping from droplets of blood14 and
separation of plasma from a flowing blood sample.15 There also exists numerous examples of cell
sorting on microfluidic devices.
Finally, and what may prove in the immediate near future to be the most important aspect
of investigating the bloodstream with microflow techniques is the finding that many processes in
the bloodstream are actually activated due to the flow itself.8,16,17 It has long been accepted that
endothelium-derived NO is actually stimulated due to a shear stress placed upon the endothelium.11–13
However, it is also becoming increasingly clear that other events in the microcirculation are also
stimulated due to the forces placed upon certain cells when flowing through open tubes (vessels).
These same forces can be applied to cells in an in vitro format in a microflow system and the resultant
cellular response can be quantitatively measured or monitored.18 In order to better understand the
importance of flow for the induction of many cellular events in vivo, it is necessary to understand
the properties of blood flow in vivo.
29.2.1 COMPONENTS OF THE CIRCULATION INVESTIGATED WITH MICROFLUIDIC

TECHNIQUES
The microcirculation is generally considered to be the blood flow through vessels having diameters
that are less than 100 µm. Specifically, this set of vessels consists of capillaries, arterioles, and
venuoles. There do exist other classifications within this range of vessels (e.g., meta-arterioles are
often defined as those vessels whose diameters range between 10 and 25 µm). However, capillaries
can be defined to have diameters below 25 µm while arteriole diameters range between 25 and
100 µm.1 However, there exist many different cell types that are determinants of microcirculatory
maintenance. There traditional roles and potential new roles being that are being discovered through
microanalytical techniques follow.
29.2.1.1 Erythrocytes
Many researchers consider the microcirculation to be a very important component of the circulatory
system due to its 60,000 miles of architecture (vessels) that deliver the different cell types in the
Microfluidic Technology as a Platform to Investigate Microcirculation 843
Red blood cell Platelet
Stress ATP NO
Endothelial
cells
NO
Smooth muscle
cells
FIGURE 29.1 Proposed mechanism for the involvement of RBCs and platelets as determinants of vascular
caliber in the microcirculation.
bloodstream to the tissues and organs in the human body. Of these cell types, erythrocytes or red
blood cells (RBCs) are deemed important due to their classic task of delivering oxygen to the
various tissues and organs in vivo. RBCs are generally considered as one of the more nonexciting
cell types in the circulation. They do not possess a nucleus, mitochondria, or any other key type
of organelle. Recently, however, RBCs have gained some attention in the literature due to their
association with NO, both as a carrier of the molecule19,20 and as a potential determinant in NO
production.7,21
RBCs, when traversing microvascular beds such as in the pulmonary circulation, are subjected to
mechanical deformation. Previously, it was reported that, in the isolated perfused rabbit lung, RBCs
obtained from either rabbits or healthy humans were a required component of the perfusate in order
to demonstrate flow-induced endogenous NO synthesis in the pulmonary circulation. Importantly, it
was reported that the property of these RBCs that was responsible for the stimulation of NO synthesis
was their ability to release ATP in response to mechanical deformation. Moreover, it was reported
that the release of ATP from RBCs of rabbits and humans increased as the degree of deformation
increased.6,22,23
These reports suggested a novel mechanism for the control of vascular caliber (Figure 29.1).
In this construct, as the RBC is increasingly deformed by increments in the velocity of blood
flow through a vessel and/or by reductions in vascular diameter, it releases ATP that stimulates
endothelial synthesis of NO resulting in relaxation of vascular smooth muscle and, thereby, an
increase in vascular caliber. This vasodilation results in a decrease in vascular resistance as well
as a decrease in the stimulus for RBC deformation and ATP release. Indeed, it has been pro-
posed that RBC-derived ATP contributes to vascular resistance in both the pulmonary and systemic
circulation.
The finding that ATP is released from RBCs in response to mechanical deformation suggests that
variables such as vascular diameter and flow rate may be important determinants of ATP release as
these cells traverse the intact circulation. An understanding of the contribution of these parameters
individually and in combination to ATP release from RBCs in the microcirculation, and in particular
in resistance vessels, is of paramount importance for the comprehension of the role of the RBC as a
determinant of vascular caliber.
29.2.1.2 Platelets
Other formed elements in the bloodstream that are considered crucial are the platelets. Platelets
are non-nucleated cells in the bloodstream that are approximately 2 µm in diameter. The main
role of platelets involves hemostasis, the prevention of excessive blood loss following vascular
injury. The actions of platelets take place in a dynamic environment: flowing blood. As such, the
larger and more numerous RBCs occupy the axial portion of the vessel forcing platelets toward
the peripheral endothelial lining.24 Platelets normally circulate without adhering to undisturbed
vascular endothelium. Upon vascular insult, subendothelial collagen is exposed that acts as the
primary stimulus for platelet activation. However, several other endogenous agonists also exist
such as thrombin, ADP, thromboxane A2 , serotonin, and epinephrine, which also promote platelet
activation.25 Platelet activation initiates a change in the platelet shape, thus promoting adhesion to
the vascular walls and the subsequent recruitment of additional platelets. NO has been widely shown
to mediate this process by activating soluble guanylate cyclase (GC) which initiates a protein kinase
G (PKG) dependent pathway.26 In this construct, NO becomes an important determinant in platelet
adhesion in vivo.10 If uncontrolled, these adhered platelets become a major constituent of thrombus
formation and subsequent vessel blockage.
Recently, platelets have been reported to have unique properties in disease states other than
cardiovascular problems. For example, a procedure used in multiple sclerosis, plasma replacement
therapy, involves removing such formed elements as RBCs and leukocytes from the patient’s plasma
and then replenishing new plasma with the previously removed cells. The plasma contains the
platelets so, in addition to obtaining new plasma, the patient also receives new platelets. Interestingly,
the platelets of people with multiple sclerosis have been shown to be more susceptible to hyperactivity,
often aggregating more easily than the platelets of healthy people that do not have multiple sclerosis.27
The importance of this platelet activity is that platelets are known to create NO upon activation and
people with multiple sclerosis have been shown to have high levels of NO and NO metabolites in
their cerebral spinal fluid and urine.28,29 This trait of platelets (prone to activation and aggregation)
is also known to exist for patients that have cystic fibrosis.30 Finally, patients with diabetes also
have platelets that are known to be activated more readily than healthy, nondiabetic patients.31
Exemplifying the importance and complexity of the relationship between platelets and NO is that
NO has the ability to inhibit platelet activation.10
Interestingly, platelets have the ability to create bioavailable NO in the bloodstream in multiple
ways other than by activation. For example, NO can be synthesized inside of the platelet by one of
multiple isoforms of NOS32,33 or from the denitrosation of S-nitrosothiols (S-nitrosoglutathione or
S-nitrosoalbumin) enzymatically34 or by copper-containing proteins.35
29.2.1.3 Endothelial Cells

Endothelial cells form a single layer within the blood vessels comprising the circulatory system.
They separate the flowing blood within the lumen from the vessel wall. While they are present in
both arteries and capillaries they have significantly different properties depending on the vessel size.
Endothelial cells are involved in many biological roles some of which include vasoconstriction, blood
pressure, inflammation, and transfer of materials into and out of the circulatory system. Research
of vascular disease has significantly increased the culturing of endothelial cells that are dominantly
obtained from the bovine aorta, bovine adrenal capillaries, rat brain capillaries, human umbilical
veins as well as other sources. Once cultured, a simple protocol for cell dissociation involving trypsin
can be followed to subculture the cells, reseeding them for future divisions or integrating them into
microfluidic devices for in vivo mimic studies.

29.3.1 UNIQUE FLOW PROPERTIES OF THE BLOODSTREAM
Collectively, the formed elements that comprise blood in the microcirculation have the properties
of a non-Newtonian fluid. As such, flow properties encountered in open tubes are not necessarily
found when the fluid contains components such as RBCs or platelets. Many properties of blood
flow in a vessel in vivo are difficult to mimic in a section of microbore tubing or a channel in a
microfluidic device. For example, the overall pressure drop across an open tube can be described by
the expression shown in the following equation.
Pr 4 π
q=
8 µL
Here, the overall flow q is dependent on variables such as the pressure applied across the tube (P),
the radius (r), length (L) of the tube, and the viscosity (µ) of the solution. Unfortunately, this
equation must be modified somewhat because blood is a non-Newtonian fluid having some pseu-
doelasticity. Therefore, at similar applied pressures, blood will flow faster than a typical Newtonian
fluid. Furthermore, in vivo, the vessel diameter (due to dilation and constriction of the resistance
vessels) is dynamic and constantly changing. Moreover, the viscosity of blood is very difficult to
measure because the hematocrit of blood changes depending on the diameter of the vessel in which
it flows. Because of this, blood viscosity in vivo is often reported as an “apparent” viscosity. Finally,
in the bloodstream, blood flow is somewhat pulsatile, especially in the venous side of the circulation.
In sum, it is difficult to know the exact linear rate of the blood flow or a pressure drop at any one
point in the microcirculation.
Another aspect of blood flow in vivo, and one that plays a major role in such circulation compli-
cations such as atherosclerosis and stroke, is that cells will partition themselves into certain areas of
the flowing stream. For example, RBCs will typically travel in the middle of the open tube where the
convective forces are at a minimum. However, platelets, being smaller in size, then become forced
to flow more toward the wall. In fact, a close examination of the components of blood at the wall of
the vessel or tube will show that there is actually a cell-free layer (often called the skimming layer)
along the wall. This phenomenon was first noticed in the 1920s by Fahraeus and Lindqvist today
bears the name of their discoverers.1
Interestingly, it may also be a key to a potential new role for the RBC, namely, the ability to
participate in the control of vascular caliber in the intact circulation. Furthermore, it is these flow
properties that enable such formed elements as the platelets to participate in the repair of damaged
endothelium.
29.4 APPLICATIONS
29.4.1 DISEASE BIOMARKERS
Although tissue analysis is an outstanding point to begin for many experiments in search of biomark-
ers, the bloodstream represents one of the more accessible tissues from which to extract samples. An
example of microfluidic technology helping to determine such a biomarker was recently reported by
Carroll, where the amount of ATP released from RBCs that were exposed to shear-induced mechan-
ical deformation.4 Specifically, RBCs obtained from healthy control subjects were pumped through
microbore tubing having diameters that approximated those of resistance vessels in vivo. The ATP
released from these deformed RBCs was measured using the well-known luciferin chemilumines-
cence assay for ATP. When compared to RBCs obtained from patients having type II diabetes, the
control patients release approximately two times the amount of ATP (190 ± 10 nMvs. 90 ± 10 nM,
respectively) as the RBCs obtained from the patients with type II diabetes (Figure 29.2). Such a
screen could be an important marker for adult onset diabetes because one could screen the ATP
release, archiving the results for the patient, and subsequently monitor any decrements on perhaps
an annual basis.
0.25
0.20
ATP Release (µM) 0.15
0.10
0.05
0.00
Control Diabetic
FIGURE 29.2 Determination of ATP release from the RBCs of diabetic (n = 7) and nondiabetic (n = 7)
patients is shown. The RBCs were pushed through microbore tubing that had an inside diameter of 50 µm.
The error bars represent standard errors about the mean. The values shown in the figure differ significantly
(p < 0.001).
29.4.2 HIGH-THROUGHPUT DRUG EFFICACY INVESTIGATIONS

A microfluidic device has been utilized to explore the mechanism by which vasodilation occurs
and the effects of iloprost, a known vasodilator. Iloprost is a second generation prostacyclin ana-
log that has been shown to directly affect vasorelaxation in vivo by increasing the amount of ATP
released from RBCs transversing the circulatory system. ATP is then free to stimulate the produc-
tion of NO from endothelial cells resulting in the production of guanosine monophosphate, which
is known to relax smooth muscle cells surrounding resistance vessels. This study incorporates the
methodology of the previous research and examines the effect of iloprost on the release of ATP
from RBCs when forced through a microchannel to mimic the circulation system. An irreversibly
sealed poly(dimethyl)siloxane (PDMS) microfluidic device was fabricated from a 4 in. silicon wafer
master that was obtained using standard negative photoresist lithographic processes (Figure 29.3a).
The channel dimensions measured approximately 100 µm wide by 100 µm deep and were pat-
terned from a negative photomask. This T-channel chip was irreversibly sealed to another PDMS
chip absent of any channels after perforating sample introduction and waste ports. A displacement
syringe pump and Tygon tubing was used to introduce samples via hypodermic steel tubing inserted
into the channels. A 7% RBC solution treated and left untreated with iloprost was mixed with a
luciferin/luciferace solution to examine the iloprost stimulated, RBC-derived ATP. The chemilumi-
nescence for all samples and standards was measured in real time by a photomultiplier tube at a rate
of 10 measurements per second for 30 s using software written in-house. The results from this study
are shown in Figure 29.3b.
29.4.3 MEASURING CELL-TO-CELL COMMUNICATION

Another advantage of microflow techniques to investigate the events occurring in the microcircu-
lation is those involving multiple cell types in a controlled environment. For example, the ability
to place endothelial cells in a microfluidic device and monitor the ability of platelets to adhere to
these endothelial cells would be beneficial in studies pertaining to vessel blockage due to throm-
bus formation. Quite often, a thrombus is formed due to platelets, macrophages, and even RBCs
adhering to the endothelial wall under certain conditions. The ability to quantitatively monitor those
(a) (b)
FIGURE 29.3 (a, b) Determination of ATP release from RBCs incubated with iloprost, a stable analog of
prostacyclin, in a microfluidin format.
molecules involved in the thrombus formation or those molecules that prevent such adhesion events
would be useful to the biomedical community. Figure 29.4a shows an endothelium coating in the
channel of a microfluidic device.9 The dimensions of this channel approximate those of resistance
vessels in vivo. Figure 29.4b shows activated platelets flowing over an endothelium. Importantly,
a close observation reveals that, in certain areas, platelet recruitment is occurring; that is, upon
activation, platelets become somewhat sticky and adhere to each other in addition to the endothelial
wall. In such a scenario, these platelets become determinants in the mechanism of vessel blockage.
The ability to create a multicellular composition in a controlled environment will facilitate the inves-
tigations of thrombus formation, occurrence of stroke, and studies involved in discovering new
thrombolytic materials.
29.5 METHODOLOGY
29.5.1 CELL HARVESTING
One of the inherent challenges for performing studies involving the circulation is that, due to an
absence of a nucleus, RBCs and platelets cannot be cultured. Therefore, to work with a meaningful
sample of RBCs, platelets, plasma, and so forth, one has to gather these cells from a mammalian
subject prior to completing the analysis. The RBCs and platelets are both obtained from the whole
blood of the animal or human. This means that approval from either an animal investigation committee
or a human investigation committee is required before such harvesting can proceed. In addition, if
the harvesting will involve a nonsurvival procedure (with an animal subject), it is likely that certain
drugs will be employed that may be controlled substances or require a special license in order to be
used in the procedure. Of course, the appropriate training will also be required prior to any harvesting
procedures.
Once the necessary approvals are in place, the harvesting of the RBCs and platelets can proceed.
This procedure for removal of cells from rabbits has been outlined in many publications. Rabbits
can be anesthetized with ketamine (8 mL/kg, i.m.) and xylazine (1 mg/kg, i.m.) followed by pen-
tobarbital sodium (15 mg/kg i.v.). A cannula is then placed in the trachea so that the animals can
be ventilated with room air. A catheter is then placed into a carotid artery for administration of
heparin and for phlebotomy. After heparin (500 units, i.v.), animals are exsanguinated and the whole
blood collected in a 50 mL tube. Depending on the animal size, varying amounts of blood can be
collected.
(a)
ATP
Inverted microscope
Syringe pump
Fluid access ports
2.00∝
2.00 L/m
µL in
/min
PDMS chip Glass

substrate
Waste
Data analysis
(b)
FIGURE 29.4 (a) A PDMS chip was fabricated to contain 100 µm × 100 µm × 2 cm channels. bPAECs were
cultured into the microfluidic channels, and the chip was reversibly sealed to a glass substrate, making sure to
align the channel and the fluid access port. Solutions were perfused across the cells (2 µL min−1 ) using 75 µm
microbore tubing to connect the syringe pump to the microfluidic adaptor on the fluid access port. Fluorescence
measurements were acquired using an Olympus IX71M microscope equipped with an electrothermally cooled
CCD (Orca, Hamamatsu) and MicroSuite software (Olympus). (b) Platelets are shown adhering to an immobi-
lized endothelium, even in the presence of flow. Such devices may be used in the future to investigate thrombus
formation or other vaso-occlusive mechanisms.

Once the RBCs or platelets are harvested from the subject, there still exists some sample preparation
prior to the measurement portion of the analysis. For example, in order to purify platelets, the collected
blood is centrifuged at 500 g at 37◦ C for 10 min. The platelet-rich plasma (PRP) is decanted for the
subsequent isolation of platelets. Platelets are then isolated from the PRP by centrifugation (15 min
at 2000 g at 37◦ C) and washed three times in Tyrode-albumin solution (pH 7.4). The first wash
contains heparin (2 U/mL) and apyrase (1 U/mL); the second wash contains only apyrase (1 U/mL);
and the third wash contains Tyrode’s solution without apyrase and heparin.
In order to purify the RBCs, the plasma and buffy coat (white cells) are removed for further purifi-
cation (see above). RBCs are then resuspended and washed three times in a physiological salt solution
[PSS; in mM, 4.7 KCl, 2.0 CaCl2 , 140.5 NaCl, 12 MgSO4 , 21.0 tris(hydroxymethyl)aminomethane,
11.1 dextrose with 5% bovine serum albumin (final pH 7.4)]. Cells should generally be prepared and
studied on the day of use within 8 h of removal from animal or human subjects.
In addition to the cell purification, it is also important to remember that, in vivo, RBCs exist at
various hematocrits throughout the vasculature, depending on the size of the vessel. For example, it is
generally accepted that the average hematocrit of RBCs in vivo is approximately 40–45%. However,
in resistance vessels, this value can decrease to levels as small as single-digit percentiles.
Endothelial cells are, in some ways, easier to work with than the RBCs and platelets because
they can be grown in standard tissue culture flasks. However, even though these cells may be easier
to obtain (committee approval is not required), preparing them for a quantitative investigation in a
microflow system is challenging because one must immobilize these cells prior to investigation. For
example, Kotsis has demonstrated the ability to culture endothelial cells from bovine pulmonary
artery on the walls of a microbore tubing. Although these cells responded to an ATP bolus by
producing NO, the preparation of the cell microflow reactors was tedious and met with a low rate
of success. More recently, immobilizing these cells to the walls of a channel in a microfluidic
device has been reported by multiple groups. Immobilizing the cells in the microfluidic device is
advantageous because the chip device is planar (similar to standard tissue culture flasks), the channels
are accessible thus allowing media to flow over the cells, and the cells can be easily monitored using
standard microscopy (Figure 29.4a). There also exist different methods for cell immobilization.
Spence et al.36 employed a method by which endothelial cells were vacuum delivered into a
channel that was coated with fibronectin. Li et al.37 described the use of a “trough” method for
immobilizing a dopaminergic cell model. This method appears to be useful for cell types that have
a tendency to cluster. Other methods have been described as well. In most cases, hydrodynamic
pumping schemes have been employed for such immobilization determinations.38
29.5.3 DETECTION
Detection of the molecules produced, consumed, and secreted by the cells described here is chal-
lenging for two main reasons. First, the cell is dynamic and constantly tries to maintain balance.
As such, molecules concentrations or speciation are usually changing. Second, the matrix in which
these measurements are typically performed is very complex. Thus, the technique of choice needs
to have some built-in feature that enables the analyst to overcome the matrix. To date, a variety of
measurements have been employed to learn more about the roles of the cells in the microcirculation.
Specifically, fluorescence, chemiluminescence, and amperometry have all been used extensively. Not
surprisingly, all three of these detection schemes are readily employed in capillary electrophoresis-
based determinations. Therefore, many of the measurements employ technology from the CE field.
However, due to the cell matrix complexity, techniques are required to overcome potential interfer-
ents. For example, Kovarik et al.39 employed a Nafion coating over a micromolded ink electrode for
selectivity in detecting dopamine in the presence of an anion interferent (ascorbate). For similar rea-
sons, Ku40 employed the classic method of multiple standard additions to quantitatively determine
the amount of NO released from activated platelets in a flowing stream.
29.6 FUTURE ASPECTS

The concept of performing studies involving the circulation in an environment that closely mimics
blood vessels and its components is relatively new. Therefore, there has not been a great deal of
studies involving actual blood flow in microflow systems. However, as pointed out in a recent
review by Martin, Root, and Spence, the number of papers published over the past decade involving
biological cells and microfluidic devices is growing at a very rapid rate. One of the obstacles in
performing cellular studies with cells typically found in the circulation is that many of these cells
cannot be cultured; that is, they must be harvested. One of the initial steps in any analysis is obtaining
the sample and obtaining the sample is not easy in these studies. However, the potential range of
studies that still need to be investigated are immense.
One potential area where microflow studies of the circulation will be of major benefit involves
studies involving glycosylation products. In a recent report, it was stated that nearly 80% of all
proteins in the bloodstream were glycosylated. With such products being implicated heavily in such
diseases as diabetes and cancer, it would seem logical to employ the systems described here with
other separation schemes, detection schemes, or protocols in general to determine various biomarkers
in a system that more readily mimics the circulation in vivo.
As eluded to above, one of the key features that this technology offers is the ability to perform
important investigations early. For example, most biomarker detection is successful on plasma,
serum, or other tissues after the disease of interest has already taken hold of the patient. By monitoring
the bloodstream, one may be able to find certain biomarkers before the formation of tumors, vessel
blockages, organ damage, and so forth. Moreover, it may also be possible to perform biomarker
studies not only on patients already inflicted with a disease but also to perform such studies on
healthy patients, creating an archive of their health and subsequently monitoring the decline in
certain proteins, peptides, or metabolites.
REFERENCES
1. Enderle, J., Blanchard, S., Bronzino, J. Introduction to Biomedical Engineering; Academic Press,
San Diego, CA, 2000.
2. Bergfeld, G. R., Forrester, T. Release of ATP from human erythrocytes in response to a brief period of
hypoxia and hypercapnea, Cardiovascular Research, 1992, 26, 40–47.
3. Olearczyk, J. J., Stephenson, A. H., Lonigro, A. J., Sprague, R. S. Heterotrimeric G protein Gi is
involved in a signal transduction pathway for ATP release from erythrocytes, American Journal of
Physiology 2004, 286, H940–H945.
4. Carroll, J., Raththagala, M., Subasinghe, W., Baguzis, S., Oblak, T. D. A., Root, P., Spence, D. An
altered oxidant defense system in red blood cells affects their ability to release nitric oxide-stimulating
ATP, Molecular BioSystems, 2006, 2, 305–311.
5. Martin, R. S., Root, P. D., Spence, D. M. Microfluidic technologies as platforms for performing
quantitative cellular analyses in an in vitro environment, Analyst (Cambridge, United Kingdom),
2006, 131, 1197–1206.
6. Sprague, R. S., Olearczyk, J. J., Spence, D. M., Stephenson, A. H., Sprung, R. W., Lonigro, A. J.
Extracellular ATP signaling in the rabbit lung: Erythrocytes as determinants of vascular resistance,
American Journal of Physiology–Heart and Circulatory Physiology, 2003, 285, H693–H700.
7. Sprague, R. S., Ellsworth, M. L., Stephenson, A. H., Lonigro, A. J. ATP: The red blood cell link
to NO and local control of the pulmonary circulation, American Journal of Physiology, 1996, 271,
H2717–H2722.
8. Sprague, R. S., Ellsworth, M. L., Stephenson, A. H., Lonigro, A. J. Increases in flow rate stimulate
adenosine triphosphate release from red blood cells in isolated rabbit lungs, Experimental Clinical
Cardiology, 1998, 3, 73–77.
9. Oblak, T. D. A., Root, P., Spence, D. M. Fluorescence monitoring of ATP-stimulated, endothelium-
derived nitric oxide production in channels of a poly(dimethylsiloxane)-based microfluidic device,
Analytical Chemistry, 2006, 78, 3193–3197.
10. Freedman, J. E., Loscalzo, J., Barnard, M. R., Alpert, C., Keaney, J. F., Michelson, A. D. Nitric oxide
released from activated platelets inhibits platelet recruitment, The Journal of Clinical Investigation,
1997, 100, 350–356.
11. Ignarro, L. J., Buga, G., Dhaudhuri, G. EDRF generation and release from perfused bovine pulmonary
artery and vein, European Journal of Pharmacology, 1988, 149, 79–88.
12. Moncada, S., Palmer, R., Higgs, E. A. Biosynthesis of nitric oxide from l-arginine, Biochemical
Pharmacology, 1989, 38, 1709–1715.
13. Palmer, R., Ferrige, M. J., Moncada, S. Nitric oxide release accounts for the biological activity of
endothelium-derived relaxation factor, Nature, 1987, 327, 524–526.
Hughes, M. A., et al. A fully integrated microfluidic genetic analysis system with sample-in-answer-
out capability, Proceedings of the National Academy of Sciences of the United States of America, 2006,
103, 19272–19277.
15. Yang, S., Uendar, A., Zahn, J. D. A microfluidic device for continuous, real time blood plasma
separation. Lab on a Chip, 2006, 6, 871–880.
16. Sprague, R. S., Ellsworth, M. L., Stephenson, A. H., Kleinhenz, M. E., Lonigro, A. J. Deformation-
induced ATP release from red blood cells requires cystic fibrosis transmembrane conductance regulator
activity, American Journal of Physiology, 1998, 275, H1726–H1732.
17. Sprague, R. S., Stephenson, A. H., Ellsworth, M. L., Keller, C., Lonigro, A. J. Impaired release of
ATP from red blood cells of humans with primary pulmonary hypertension. Exp Biol Med, 2001, 226,
434–439.
18. Sprung, R. J., Sprague, R. S., Spence, D. M. Determination of ATP release from erythrocytes
using microbore tubing as a model of resistance vessels in vivo, Analytical Chemistry, 2002, 74,
2274–2278.
19. Gow, A. J., Stamler, J. S. Reactions between nitric oxide and haemoglobin under physiological
conditions, Nature, 1998, 391, 169–173.
20. Stamler, J. S., Jia, L., Eu, J. P., Mcmahon, T. J., Demchenko, I. T., Bonaventura, J., Gernert, K.,
Piantadosi, C. A. Blood flow regulation by S-nitrosohemoglobin: A dynamic activity of blood involved
in vascular control, Science, 1997, 276, 2034–2037.
21. Ellsworth, M. L., Forrester, T., Ellis, C. G., Dietrich, H. H. The erythrocyte as a regulator of vascular
tone, American Journal of Physiology, 1995, 269, H2155–H2161.
22. Sprague, R. S., Stephenson, A. H., Dimmitt, R. A., Weintraub, N. A., Branch, C. A., McMurdo, L.,
Lonigro, A. J. Effect of L-NAME on pressure-flow relationships in isolated rabbit lungs: Role of red
blood cells, American Journal of Physiology–Heart and Circulatory Physiology, 1995, 269, H1941–
H1948.
23. Fischer, D. J., Torrence, N. J., Sprung, R. J., Spence, D. M. Determination of erythrocyte deformability
and its correlation to cellular ATP release using microbore tubing with diameters that approximate
resistance vessels in vivo. Analyst, 2003, 128, 1163–1168.
24. Radomski, M. W., Radomski, A. S. In Vascular Endothelium in Human Physiology and Patho-
physiology, Vallance, P., Webb, D., Eds., Academic Publication, Dordrecht, The Netherlands,
1999.
25. Packham, M. A. Role of platelets in thrombosis and hemostasis, Canadian Journal of Physiology and
Pharmacology, 1994, 72, 278–284.
26. Morrell, C. N., Matsushita, K., Chiles, K., Scharpf, R. B., Yamakuchi, M., Mason, R. J., Bergmeier,
W., et al. Regulation of platelet granule exocytosis by S-nitrosylation, Proceedings of the National
Academy of Sciences of the United States of America, 2005, 102, 3782–3787.
27. Neu, I. S., Prosiegel, M., Pfaffenrath, V. Platelet aggregation and multiple sclerosis, Acta Neurologica
Scandinavica, 1982, 66, 497–504.
28. Giovannoni, G., Heales, S. J. R., Land, J. M., Thompson, E. J. The potential role of nitric oxide in
multiple sclerosis, Multiple Sclerosis, 1998, 4, 212–216.
29. Smith, K. J., Lassmann, H. The role of nitric oxide in multiple sclerosis, Lancet Neurology, 2002, 1,
232–241.
30. O’Sullivan, B. P., Linden, M. D., Frelinger, A. L., III, Barnard, M. R., Spencer-Manzon, M., Mor-
ris, J. E., Salem, R. O., Laposata, M., Michelson, A. D. Platelet activation in cystic fibrosis, Blood,
2005, 105, 4635–4641.
31. Coppola, L., Verrazzo, G., La Marca, C., Ziccardi, P., Grassia, A., Tirelli, A., Giugliano, D. Effect
of insulin on blood rheology in non-diabetic subjects and in patients with Type 2 diabetes mellitus,
Diabetic Medicine: A Journal of the British Diabetic Association, 1997, 14, 959–963.
32. Sase, K., Michel, T. Expression of constitutive endothelial nitric oxide synthase in human blood
platelets, Life Sciences, 1995, 57, 2049–2055.
33. Muruganandam, A., Mutus, B. Isolation of nitric oxide synthase from human platelets, Biochimica et
Biophysica Acta, 1994, 1200, 1–6.
34. Root, P., Sliskovic, I., Mutus, B. Platelet cell-surface protein disulfide-isomerase mediated
S-nitrosoglutathione consumption, Biochemical Journal, 2004, 382, 575–580.
35. Gordge, M. P., Hothersall, J. S., Neild, G. H., Dutra, A. A. N. Role of a copper(I)-dependent
enzyme in the anti-platelet action of S-nitrosoglutathione, British Journal of Pharmacology, 1996, 119,
533–538.
36. Spence, D. M., Torrence, N. J., Kovarik, M. L., Martin, R. S. Amperometric determination of nitric
oxide derived from pulmonary artery endothelial cells immobilized in a microchip channel, Analyst,
2004, 129, 995–1000.
37. Li, M., Spence, D. M., Martin, R. S. A microchip-based PC 12 cell reactor: Determination of immo-
bilization conditions and amperometric detection of catecholamines released upon stimulation with
calcium, Electroanalogy, 2005, 17, 1171–1180.
38. Khademhosseini, A., Langer, R., Borenstein, J., Vacanti, J. P. Microscale technologies for tissue
engineering and biology, Proceedings of the National Academy of Sciences of the United States of
America, 2006, 103, 2480–2487.
39. Kovarik, M. L., Torrence, N. J., Spence, D. M., Martin, R. S. Fabrication of carbon microelectrodes
with a micromolding technique and their use in microchip-based flow analyses, Analyst (Cambridge,
United Kingdom), 2004, 129, 400–405.
40. Ku, C. J. Analytical Chemistry, 2007, 79, in press.
30 Capillary Electrophoresis
Applications for Food Analysis
Belinda Vallejo-Cordoba and
María Gabriela Vargas Martínez
CONTENTS
30.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 853

30.2 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 854
30.2.1 Carbohydrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 854
30.2.2 Amino Acids and Biogenic Amines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 854
30.2.3 Vitamins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 863
30.2.4 Organic Acids and Inorganic Ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 864
30.2.5 Toxins, Contaminants, Pesticides, and Residues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 867
30.2.6 Analysis of Phenolic Compounds in Food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 876
30.2.7 Chiral Analysis of Food Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 876
30.2.8 Proteins and Peptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 886
30.2.9 DNA and Microchips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 891
30.2.10 Food Additives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 893
30.2.11 Food Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 895
30.3 Concluding Remarks and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 898
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 899
30.1 INTRODUCTION
The usefulness and importance of capillary electrophoresis (CE) in food analysis is now well recog-
nized and established, as shown with many applications in the study of substances of food concern,
ranging from naturally ocurring compounds to contaminants. This fact is corroborated by more than
650 papers published in the past 10 years,1 including several review papers2–4 and a handbook for
method development in food analysis.5 Also, the scope of applications is broad in terms of molecular
size of food components, ranging from the analysis of small molecules such as organic acids (OA) or
amino acids (AA) to the analysis of large biomolecules such as proteins, carbohydrates, or DNA.1 An
exhaustive review examining the latest developments in the application of capillary electromigration
methods for the analysis of food components, as well as for the investigation of food interactions and
food processing was recently published.2 In addition, the analysis of organic contaminants in foods
with emphasis on procedures for extracting and concentrating compounds such as pesticides, antibi-
otics, biological toxins, and pathogens was also reviewed.3 Another important area in food analysis
where CE offers good potential is food authentication. Several examples of food authentication were
summarized in two review papers.4,6 Consequently, the aim of this chapter is not a comprehensive
review of the literature as this has been done in other works, but to present selected applications of
CE that are most relevant to the analysis of different food components and food quality.
853
30.2 APPLICATIONS
30.2.1 CARBOHYDRATES
The analysis of carbohydrates by CE has been reviewed by several authors as scientific papers1,7–9
or chapters of books.5,10 In fact, Chapter 7 by Khandurina in this handbook deals specifically with
this topic. Carbohydrates have traditionally been classified by food researchers into sugars and
polysaccharides, although mixtures of them such as glucose syrups are also used, taking advantage
of their respective characteristics. Sugars are utilized for their sweetening power, preservative action
(osmotic pressure), and crystallinity in foodstuffs; polysaccharides provide foodstuffs with texture,
body, and colloidal properties.
Currently, the assay of carbohydrates is a very important criterion for the quality control of
drinks and foodstuffs, especially in dietary products. Assay of carbohydrates is also very important
in the fields of monitoring of food-labeling claims, analysis of sweeteners, bulking agents, and
fat substitutes as well as for establishing authenticity and in the fermentation monitoring in the
production of alcoholic beverages.
A representative selection of applications is shown in Table 30.1, which illustrates the wide
choice of food sample types that have been analyzed, with attention being paid to the CE mode,
its running conditions, the derivatization technique and its conditions, as well as the detection
method utilized.11–26 Since carbohydrates lack both a charge and a strong ultraviolet (UV) chro-
mophore, several derivatization techniques have been described during the development of the
CE methodologies.11,27 Wang et al.11 reported an excellent detection limit of 420 amol for a
standard mixture of monosaccharides previously derivatized with 3-aminophthalhydrazide and
an online chemiluminescence detector with hexacyanoferrate as the postcapillary chemilumines-
cence reagent. A more conventional approach is the derivatization with a UV chromophore,
although this does not give such good sensitivity as the chemiluminescence detection. Several
chromophores have been utilized to derivatize carbohydrates such as 2-aminobenzoic acid,12
6-aminoquinoline,13 8-aminonaphthyl-1,3,6-trisulfonic acid,14 9-aminopyrene-1,4,6-trisulfonic acid
(APTS),19,23 p-nitroaniline,22 and p-aminobenzoic acid.26 While these methods lead to improved
sensitivity and resolution, the complexity of derivatization limits its use.
Alternatively, methods for the analysis of underivatized carbohydrate have been developed.
These methods include the use of high alkaline electrolyte to ionize the carbohydrates and make
them suitable for indirect UV detection.15–17,20,24,28–30 These techniques enabled to analyze acidic,
neutral, and amino sugars without derivatization. Soga and Serwe18 reported the separation of the
standard mixtures of 28 carbohydrates (Figure 30.1)18 including monosaccharides, disaccharides,
acidic sugars, amino sugars, and sugar alcohols using indirect detection at high pH with 20 mM
2,6-pyridinedicarboxylic acid (PDC) for indirect UV detection and 0.5 mM cetyltrimethylammonium
hydroxide (CTAH) for reversing the direction of the electroosmotic flow (EOF). The detection limits
for fructose, glucose, and sucrose were in the range from 12 to 16 mg/L with pressure injection of
50 mbar for 6 s. The reader is referred to two additional reviews on carbohydrate analysis that have
recently been published.31,32
30.2.2 AMINO ACIDS AND BIOGENIC AMINES

An overwhelming majority of foods contain AA, either in the free form (e.g., fruit juice) or in
the form of protein (partially hydrolyzed or intact). Proteins are polymers of AA, and as such,
represent the principal source of dietary AA for humans when enzymatically digested to liberate their
constituents AA. Consequently, the determination of the AA content in food is important in a number
of applications that include food as the sole source of nutrition (e.g., infant formula), prescribed
fortified nutritional products (e.g., product enriched with glutamine), the verification of the absence
of an specific AA in certain inborn errors of metabolism (e.g., phenylalaline in phenylketoneurea),
as well as for regulatory concerns.
TABLE 30.1
Methods for the Analysis of Carbohydrate Samples by CE
Type of Food Detection Method Sample Derivatization CE Mode Run Buffer CE Conditions References
Standard mixture of Online chemilu- Precapillary derivatization with CZE 200 mM borate (pH 10), 107 cm × 75 µm i.d. fused-silica 11
monosaccharides minescence 3-aminophthal-hydrazine and 100 mM hydrogen peroxide capillary, 24 kV, 25◦ C
postcapillary reaction with
25 mM hexacyanoferrate in
3 M NaOH
Mono- and disaccharides UV absorbance Mono- and disaccharides CZE 50 mM sodium phosphate, pH 7 70 cm × 50 µm i.d. fused-silica 12
214 nm derivatized with capillary, 20 kV, 25◦ C
2-aminobenzoic acid
Peanut fungal pathogens UV absorbance N-Acetylglucosamine and CZE 100 mM sodium phosphate 80 cm × 20 µm i.d. fused-silica 13
and Baker’s yeast 254 nm glucose derivatized with monobasic (pH 7) capillary, 18 kV
6-aminoquinoline prior to
analysis
Cereals/corn UV absorbance Reducing sugars derivatized with CZE 250 mM phosphoric acid titrated 67 cm × 50 µm i.d. fused-silica 14
Capillary Electrophoresis Applications for Food Analysis
235 nm 8-aminonaphthyl- pH 4.5 with triethylamine capillary, 20 kV, 30◦ C

1,3,6-trisulfonic acid
Monosaccharides in Indirect UV at Underivatized CZE 70 mM NaOH, 2 mM naphthol 23 cm × 50 µm i.d. fused-silica 15
orange juice 570 nm blue-black, and 80 mM sodium capillary, 9 kV, 25◦ C
phosphate (pH 12.5)
Monosaccharides in soft Indirect Underivatized CZE 10 mM NaOH, 44 cm × 20 µm i.d. × 375 µm o.d. 16
drinks, isotonic contactless 4.5 mM Na2 HPO4 , fused-silica capillary 2 kV, 30◦ C
beverages, fruit juice, conductivity 200 µm CTAB
and sugarcane spirits detection
Carbohydrates in ESI-MS Underivatized CE-MS 300 mM diethylamine (DEA). 70 cm × 50µm i.d. fused-silica 17
alcoholic drinks/wine Sheath liquid: 2-propanol/water capillary, 20 kV, 25◦ C
(80/20% v/v) with 0.25% DEA
at 4 µL/min
Continued
855
856
TABLE 30.1
(Continued)
Mono- and disaccharides, Indirect detection Underivatized CZE 20 mM 2,6-PDC with 112.5 cm × 50 µm i.d. fused-silica 18
amino sugars, and sugar 350 nm 0.5 mM CTAH (pH 12.1) capillary, −25 kV, 20◦ C
alcohols in orange juice,
apricot, sake, yogurts
Oligo- and Capillary array Derivatized with 0.2 M CZE 25 mM Lithium acetate (pH 5.0) 62 cm × 75 µm i.d. fused-silica 19
monosaccharides in electrophoresis- 9-aminopyrene-1,4,6-trisulfonic capillary, 15 kV, 27 or 30◦ C
enzymatic LIF acid trisodium (APTS) in 15%
polysaccharide digestion acetic acid and 2 M NaBH3 CN
in tetrahydrofuran
Mix of carbohydrates Frequency based Underivatized CZE 0.1 M NaOH running electrolyte Poly(dimethylsiloxane)/glass 20
electrochemical microchip with a channel 20-µm
wide and 7-µm high. 50-µm wide
copper electrode vs. Ag/AgCl
Carbohydrates in plants UV detection Derivatization with 12.5 µL of MEKC 100 mM Sodium tetraborate 760 mm × 0.05 mm i.d. 21
(lupine sample) 220 nm 0.15 M tryptamine dissolved in decahydrate, 35 mM cholic fused-silica capillary, 30 kV, 30◦ C
10% propanol and heated at acid, and 2% 1-propanol, pH 9.7
90◦ C for 10 min + 4.5 µL
sodium cyanoborohydride
(0.3 g/mL), 90◦ C for 60 min
Carbohydrates in infant UV-visible Derivatization with 50 µL of CZE 0.17 M boric acid (pH 9.7) 72 cm × 50 µm i.d. fused-silica 22
and milk powder, rice on-column LED glacial acetic acid, 100 µL of capillary, −20 kV, 27 or 30◦ C
syrup, and cola drink 406 nm 0.036 M p-nitroaniline in
methanol and 75 µL of 0.080 M
NaBH3 CN
Polysaccharides, iota, LIF Derivatization with 2 µL of CZE 25 mM ammonium acetate 60/67 cm × 50 µm i.d. polyvinyl 23
kappa, and lambda 75% v/v acetic acid, 2 µL (pH 8.0) alcohol coated capillary,
carrageenans 9-aminopyrene-1,4,6-trisulfonic 30 kV, 50◦ C
acid (APTS) reagent (0.2 M in
15% v/v acetic acid) and 2 µL
of 1 M sodium
cyanoborohydride in THF
Polysaccharides Indirect DAD Underivatized CEC 1% DMSO in 20 mM Tris, Sulfonated polystyrene/DVB 24
210 nm pH 8/ACN stationary phase. 300 Å, 45 cm
(53 cm total) × 100 µm, 10 kV,
20◦ C
Derivatized with CEC THF with 2% water Mixed Bed X400-PLGel C (40:60), 24
phenyisocyanate 25(33) cm, 75 µm; 25 kV; 20◦ C
Carbohydrates in sake Indirect UV Underivatized CZE 20 mM PDC, 0.5 mM CTAH 112.5 cm × 50 µm i.d. fused-silica 25
(pH 12.1) capillary, 30 kV, 15◦ C
Glucose, maltose, and UV detection Precolumn derivatization 20 mg CZE 20 mM sodium tetraborate (pH 57 cm × 75 µm i.d. fused-silica 26
maltotriose in 280 nm of NaBH3 CN in 1 mL of 10.2) capillary, 20 kV, 25◦ C

nonalcoholic methanolic solution of 250 mM
drinks/beverages p-aminobenzoic acid (PABA)
and 20% AcOH
857
1 23 45 7
6 8 11 12
9 13
20 10
15
14 16
17
Absorbance (mAU)
15 20
19 21 22
18
23
10
24 25 26 27 28
5
0
10 15 20
Time (min)
FIGURE 30.1 Separation of 28 standard carbohydrates by CE. Peak assignment: (1) mannuronic acid, (2) glu-
curonic acid, (3) galacturonic acid, (4) N-glycolylneuraminic acid (NGNA), (5) N-acetylneuraminic acid
(NANA), (6) ribose, (7) mannose, (8) xylose, (9) glucosamine, (10) glucose, (11) galactosamine, (12) galac-
tose, (13) fucose, (14) mannitol, (15) sorbitol, (16) xylitol, (17) inositol, (18) fructose, (19) rhamnose, (20)
lactulose, (21) lactose, (22) sucrose, (23) galactitol, (24) N-acetylmannosamine, (25) N-acetylglucosamine,
(26) N-acetylgalactosamine, (27) arabinose, and (28) raffnose, 200 mg/mL each. Experimental conditions:
capillary, fused-silica 50 µm ID × 112.5 cm (104 cm effective length); background electrolyte, 20 mM PDC,
pH 12.1; injection, 6 s at 50 mbar; temperature, 20◦ C; indirect detection signal 350 nm, reference 275 nm.
(Reprinted from Soga, T., and Serwe, M., Food Chem., 69, 339–344, 2000. With permission from Elsevier
Science-NL.)
Biogenic amines (BA) are bacterial degradation products of AA found in virtually all foods. They
are also normal constituents in fermented foods, such as cheese, wine, and beer. Since excessive
dietary intake of BA can lead to adverse physiological effects, such as migraine headaches (due to
their toxicity), and because their presence is used as indicators of the degree of freshness or spoilage
of food, it is important to analyze BA content for food quality. Histamine, putrescine, cadaverine,
tyramine, tryptamine, β-phenylethylamine, spermine, and spermidine are considered to be the most
important BA occurring in foods. The reader is referred to several reviews1,7–10 that summarize
the CE methodologies that have been developed for the determination of BA33–39 and AA,40–42 as
well as some book chapters on the same topic.5,10,32 The most significant reports and its separation
conditions are given in Table 30.2.
BA is challenging in that it displays poor chromophore and fluorophore properties. A similar
situation exists with AA because very few of them have a chromophoric moiety that allows for
facile detection after separation; consequently, derivatization procedures to facilitate their detec-
tion are usually required. Fortunately, the presence of common functional groups (amine and
carboxylic acid) chemical handles for derivatization schemes that render the BA and AA amenable
to spectrophotometric and fluorimetric detection after separation.
TABLE.30.2
Methods for the Detection of AA and BA Separated by CE
AA growth medium LIF Derivatized with MEKC 6.25 mM borate (pH 9.66), 107 cm × 75 µm i.d. fused-silica 43
3-(4-carboxybenzoyl)- 150 mM SDS, capillary, 24 kV, 25◦ C
2-quinoline-carboxaldehyde 10 mM tetra-hydrofuran
before analysis
Standards of AA CCD spectrometer Derivatized with FITC and CZE 20 mM borate buffer at pH 10 Fused-silica capillary, 44
5-iodoacetamido fluorescein 50 mm i.d., 360 µm o.d., length,
(5-IAF) 50 cm inner surface modified
with γ-glycido-
xypropyltrimethoxysilane
(GOPTMS), aplied voltage
17 kV
AA in alcoholic Electrochemical Derivatized with naphthalene- CZE Borate (pH 9.48) 80 cm × 20 µm i.d. fused-silica 45
drinks/beer 2,3-dicarboxaldehyde before capillary, 18 kV
analysis
AA in beer and yeast Contactless Underivatized CZE 2.3 M acetic acid, pH 2.1, 0.1% 80 cm × 50 µm i.d. fused-silica 46
conductivity w/w hydroxyethylcellulose capillary, 30 kV, 25◦ C
Complex mixture of AA ESI-MS Underivatized CE-MS Background electrolyte of 130 cm × 20 µm i.d. fused-silica 47
and blood 1 M formic acid capillary, 30 kV
AA standard mixture ESI-MS Underivatized CE-MS 1 M Formic acid solution 100 cm × 50 µm i.d. fused-silica 48
capillary, 30 kV, 20◦ C
Standard solution of AA LIF Derivatized with 5 mM NDA and CZE 125 mM Borate bufffer (pH 8.7) 65 cm × 50 µm i.d. fused-silica 49
and catecholamine 43 mM KCN capillary, 10 kV/7 min,
20 kV/8 min
AA in soy sauce UV-VIS at 214 nm Underivatized CZE 50 mM Phosphate buffer (pH 2.5) Uncoated fused silica column, 50
i.d. 50 µm, 150 cm (effective
length 37.8 cm), 15 kV, 25◦ C
Continued
859
860
TABLE.30.2
(Continued)
BA (histamine, DAD Derivatization with CZE 40 mM aqueous sodium 67 cm × 75 µm i.d. × 375 µm o.d. 51
tryptamine, 1,2-naphthoquinone-4-sulfonate tetraborate solution, pH 10.5, fused-silica capillary, 30 kV, 25◦ C
phenylethylamine, 2-propanol (25%, v/v)
tyramine, agmatine,
ethanolamine, serotonin,
cadaverine, and
putrescine) in red wines
BA in red and white wines ESI-MS Underivatized CE-MS 25 mM citric acid (pH 2.0) 75 cm × 50 µm i.d. × 360 µm o.d. 52
fused-silica capillary,13 kV, 20◦ C
BA in fish and wine LIF Derivatized with OPA CZE 50 mM borate buffer (pH 9.0) 47 cm × 50 µm i.d. × 360 µm o.d. 53
20 mM MβCD, 25 mM SβCD fused-silica capillary, 25 kV, 25◦ C
10% ethanol
Standards mix of AA DAD at 205 nm Underivatized MEKC 20 mmol/L sodium borate buffer 60 cm × 75 µm i.d. × 365 µm o.d. 54
(pH 9.3) etched bare fused-silica capillary,
140 mmol/L SDS 20 kV, 25◦ C
Standards of AA UV detection at Underivatized CZE 100 mM sodium phosphate 57 cm × 50 µm i.d. fused-silica 55
200 nm (pH 2.0) capillary, 20 kV, 20◦ C
Standards of AA Fluorescence Derivatized with fluorescamine CZE 50 mM borate buffer (pH 9.5) 70 cm × 50 µm i.d. × 375 µm o.d. 56
2.5 mM fused-silica capillary, 20 kV
AA in infant food ESI-MS Underivatized CE-MS 300 mM formic acid or 70 cm × 50 µm i.d. fused-silica 57
100 mM triethylamine capillary, 20 kV
BA and AA in beer LIF Derivatized with 10 µL of MEKC 25 mM boric acid running buffer 57 cm × 75 µm i.d. uncoated 58
1 × 10–3 mol/L SAMF, 75 µm (pH 9.0), containing 24 mM fused-silica capillary, 22.5 kV,
H3 BO3 buffer, pH 8.0 SDS, 12.5% v/v AcN 25◦ C
Standards of BA Indirect Underivatized Microchip-CE 30 mM phosphate (pH 9.4) 3% Separation channel, 40 µm × 10 59
fluorescence CZE 2-propanol, 500 µM µm × 16 mm, 2000 V
Rhodamine 110, pH 2.2
AA in beer Conductivity Underivatized CZE 50 mM AMP/10 mM CAPS, pH 60 cm × 25 µm i.d. × 360 µm o.d. 60
10.8 fused-silica capillary, 25 kV
BA (histamine, DAD at 210, 214, Underivatized CZE 20 mM sodium citrate buffer, 56 cm × 50 µm i.d. fused-silica 61
putrescine, cadaverine, and 320 nm 25 mM phosphate buffer, both capillary, 10/25 kV, 25/35◦ C
tyramine, spermidine, BGE pH 2.5 and 6.5
spermine, agmatine) in
fish, cheese, meat
products, vegetarian
products
AA (cysteine, LIF Derivatized with 400 µM CZE 50 mM boric acid, pH 9.7–10.6 57 cm × 75 µm i.d. fused-silica 62
cysteinyl-glycine, 5-bromomethylfluorescein 400 µM 5-BMF, 1 mM EDTA capillary, 30 kV, 30◦ C
γ-glutamyl-cysteine, (5-BMF) 1 µM GSH
glutathione, and
(γ-glutamyl-cysteinyl)
2-glycine (PC2 ) in
Durum wheat seeds
Standards of aromatic Electrochemical Underivatized Microchip- 30 mM acetate buffer solution Chip 86 × 20 × 2 mm × 75-mm 63
CZE (pH 4.5) long separation channel, 5-mm
long injection channel, 1500 V
Stock solutions of AA UV detection at Derivatized with phthalic CZE 20 mM phosphate buffer, pH 10 40.5 cm × 50 µm i.d., fused-silica 64
200 nm anhydride 10 mM on-column capillary, 24 kV, 25◦ C
BA (trimethylamine, UV detection at Derivatized with 800 µL of a CZE 6 mM copper sulfate, 70 cm × 75 µm i.d. × 375 µm o.d. 65
putrescine, cadaverine, 214 nm 1% (w/v) Dns-Cl in acetone 6 mM 18-crown-6-ether, and capillary tubing, 20 kV, 30◦ C
spermine, tryptamine, solution 4 mM formic acid (pH 2.7)
spermidine,
phenylethylamine,
tyramine) in fish, meat,
and sausage
Mixture of BA standards Spectrophotometric 5 mM 4-(4-sulfo-phenylazo)- MEKC 10 mM phosphate buffer, pH 7.8, 48.5 cm × 50 µm i.d. fused-silica 66
(histamine, tyramine, at 420 nm 1-hydroxy-2-naphthaldehyde 70% v/v ethanol, 0.3 mM Brij capillary, 20 kV, 35◦ C
cadaverine, and sodium salt (AZO2 ) 35, 10 mM Na2 SO4
putrescine)
AA in plant seeds PDA Underivatized CZE 75 mM phosphoric acid (pH 1.85) 570 mm × 50 µm i.d., uncoated 67
MEKC 50 mM NaH2 PO4 100 mM SDS fused-silica capillary, 25 kV,
(pH 7.0) 15◦ C (CZE); 25 kV, 20◦ C
(MEKC)
Continued
861
862
TABLE.30.2
(Continued)
Standard mixtures of UV-visible In-column derivatization with CEC 100 mM phosphate buffer Proline-coated column 75 cm × 75 68
nine AA detection OPA (50 mM) in borate buffer (pH 4.0) µm i.d., –15 kV
(pH 9.5, 50 mM)
BA in a sample of Thai Fluorescence On-chip derivatization with Microchip CE 80 mM borate buffer (pH 9.2.) Window of 40 µm × 40 µm 69
fish sauce emission, 25 µM dichlorotriazine
488 nm fluorescein (DTAF)
excitation
Mixture of AA Chemiluminescence Underivatized pCEC MeOH/phosphate 20/80 (pH 8.0), In situ polymerization monolithic 70
(l-threonine and detection (CL) 5 mM Cu(II) + 1 × 10–5 mol/L stationary phase (30 cm);
l-tyrosine) standards with a on column luminol pump flow rate, capillary, 75 µm i.d. ×
coaxial flow 0.01 mL/min; CL reagents, 375 µm o.d., −4 kV; back
detection H2 O2 + NaOH media; CL pressure, 500 psi
interface reagent flow rates, 1.3 µL/min
BA (histamine, tyramine, UV spectropho- Derivatized with 10 mM OPA MEKC 25 mM sodium borate buffer Channel widths/lengths of reaction 71
putrescine, and tometer and 10 mM NAC and DNS-OH (pH 10.0); 25 mM sodium SDS; chambers 120 µm × 2.3 mm
tryptamine) in red wine 5% v/v methanol (PRSM 1), 200 µm × 1.8 mm
(PRSM 2), and 50 µm × 5.6 mm
(PRSM 3). Glass microchips of
soda lime silicate, 75 mm × 25
mm × 1 mm
AA in sake Indirect UV Underivatized CZE 20 mM 2,6-pyridinedicarboxylic 112.5 cm × 50 µm i.d. fused-silica 72
acid (PDC), 0.5 mM CTAH capillary, –30 kV, 15◦ C
(pH 12.1)
AA in beer UV at 210 and Underivatized MEKC 25 mM Na2 B4 O7 (pH 10.5) and 57 cm × 75 µm i.d. fused-silica 73
270 nm 110 mM SDS capillary
Histamine in tuna fish DAD Underivatized CZE 50 mmol/L phosphate (pH 2.5) 50 cm × 75 µm i.d. fused-silica 76
samples 0.05% hydroxyethylcellulose capillary, 23◦ C
Capillary Electrophoresis Applications for Food Analysis 863
120
FITC
100
80
RFU
60
40
Mango
20 pulp
Blank
0
4 5 6 7 8 9 10 11 12 13 14
Time (min)
FIGURE 30.2 Electropherograms of the polyamines standards spiked in Manila Mangos pulp and its blank.
Conditions: Pre-column derivatization of the sample with FITC, 20 mM borate buffer, pH 9.5, 50 mM SDS,
30 kV, 25◦ C. Put = putrescine, Spd = spermidine and Spm = spermine.
Many reported methods for BA and AA involve pre- or postcolumn (or capillary) derivatization
of these compounds before detection using 3-(4-carboxybenzoyl)-2-quinoline-carboxaldehyde,43 5-
(4,6-dichloro-s-triazin-2-ylamino) fluorescein,44 1,2-naphthoquinone-4-sulfonate,45 o-Phthalaldehyde
(OPA),53,68,71 fluorescamine,56 and many other procedures reported in Table 30.2. OPA is disadvan-
tageous in that it reacts only with primary amines, and the fluorescent derivatives are associated with
significant instability. Dabsyl- and dansylchloride65 are better in this respect as they react with both
primary and secondary amino groups, and provide stable derivatives. Indirect UV,72 indirect fluo-
rescence detection,59 conductivity,46,60 and electrochemical detection 45,63 have been utilized after
CE separation as well as mass spectrometry (MS).47,48,52,57 Kvasnicka et al.74 developed a direct,
sensitive, and rapid (<15 min) CE method with conductometric detection for the determination of
BAs in food products (salami, cheese, wine, and beer). Vargas et al.75 reported the use of fluorescein
isothiocyanate isomer 1 (FITC) for the derivatization of polyamines present in manila mangoes. The
separation by CE with detection by laser-induced fluorescence (LIF) was critical to studying the
levels of polyamines in Mexican manila mango and its relationship to storage at chilling temperature
(Figure 30.2).
Nineteen AA were detected in beer samples derivatized with napthalene 23-dicarboxaldehyde
(NDA) and cyanide (CN− ) by capillary zone electrophoresis (CZE) with electrochemical detection.45
Under the optimum conditions, the limits of detection (LODs) for individual AA were between 84
and 893 amol. An interesting method recently developed by Ruiz-Jiménez and Luque de Castro65
reported the determination of BA in nine solid food samples using a full-automated method based
on pervaporation coupled online with CE and indirect UV detection. The pervaporator allowed
leaching, formation of the volatile analytes, and their removal by evaporation and diffusion through
a membrane. The isolated analytes were injected online into the CE system while the solid matrix
remained in the pervaporator. With this approach, BA have been determined in fish, meat, and
sausage, with LODs ranging between 0.2 and 0.6 µg/mL.
30.2.3 VITAMINS
The fat-soluble vitamins include vitamins A, D, E, K, and the carotenoids, which are precursors
of vitamin A. The water-soluble vitamins are vitamins C, B1 , B2 , B3 , B6 , folacine, B12 , biotin,
and pantothenic acid. Vitamin assay in foods are carried out for a variety of purposes: (1) to assure
compliance with contract specifications and nutrient labeling regulation; (2) to provide quality assur-
ance for supplement products; (3) to provide data for food composition tables; (4) to study changes
in vitamin content attributable to food processing, packaging, and storage; and (5) to assess the
effects of geographical, environmental, and seasonal conditions. Several scientific reviews77–78 and
book chapters6,10,32 have been published for the analysis of vitamins by CE, and some separation
details and the associated references can be seen in Table 30.3. CE methods for the determination
of vitamins in food are limited to the determination of vitamin C in fruit,79,80 vegetables,80,81,88 and
beverages,82−89 niacin in the range of foods,90−93 and thiamine in the samples of meat94 and milk.95
As detailed in Chapter 3 by Terabe, micellar electrokinetic chromatography (MEKC) is a useful
technique in the retention analysis of water-soluble compounds. The separation and analysis of
lypophilic analytes, however, may be difficult in MEKC due to the strong affinity of lypophilic
compounds to the micelle resulting in long separation times and poor resolution. An interesting
approach for the simultaneous analysis of water- and fat-soluble vitamins by microemulsion elec-
trokinetic chromatography (MEEKC) was proposed by Sanchez.96 The separation of both water- and
fat-soluble vitamins (B1 , B2 , B3 , B6 , B12 , C, A palmitate, D, E acetate, and K) was obtained when
the microemulsion was prepared with sodium dodecyl sulfate (SDS) as the surfactant, octane as the
nonpolar modifier, butanol as the cosurfactant, and propanol as the second cosurfactant. Complete
separation of all vitamins was carried out within 55 min; however, this approach was tested only in
multivitamin formulation.
The electrophoretic profile of flavins in commercially available baker’s yeast was reported by
Cataldi et al.99 An assay for riboflavin in common natural products was developed using LIF
detection; benefiting from its intrinsic fluorescent nature, flavins could be selectivity detected at
very low concentrations using a middle basic running electrolyte (i.e., phosphate buffer at pH 9.8).
In addition, a high sensitivity methodology was obtained with a minimal sample preparation by
Aturki et al., who reported a rapid method for the separation and detection of tocopherols (TOHs)
in vegetable oils by capillary electrochromatography (CEC) (Figure 30.3).101 The method proposed
was faster than the conventional chromatographic separation, with complete resolution of four TOHs
achieved in less than 2.5 min with LODs of 1.25 mg/mL.
30.2.4 ORGANIC ACIDS AND INORGANIC IONS

Organic acids are among the most frequently assayed substances in food, occurring naturally in a
variety of plant and animal substrates. OA present in food originate from biochemical processes,
either from their addition or from the activity of some microorganisms (particularly yeasts and bacte-
ria), and are important contributors to the sensory properties of foods. According to the regulations set
forth by the Food and Drug Administration (FDA), OA can be used as acidulants (e.g., citric, fumaric,
malic, and sorbic acid), antimicrobial additives (e.g., propionic acid), and sequestrants (e.g., tartaric
acid).105 On the other hand, inorganic ions (IA) in food are important from a health-related view-
point. Cations such as Na+ and K+ are essential for the maintenance of a proper electrolyte balance,
while excessive Na+ levels is associated with high blood pressure. The divalent cations, Mg2+ and
Ca2+ , are important for bone growth and are regulated in infant formulas. Other IA are used in food
as additives, including nitrite as a preservative and color enhancer in meat, bromate to improve the
strength of flour and reduce fermentation time, and sulfite as antimicrobial growth factor in food.
A scan of recent literature on the topic clearly reveals that CE is increasingly used for the
analysis of OA and IA in food. These reports are summarized in Table 30.4, and excellent reviews
about this topic can be found in the literature.106,107 The detectors most frequently used for the
OA analysis by CE are direct UV at 206–220 nm, and conductivity; while for IA, indirect UV and
conductivity detection are the most often reported. Cortacero-Ramírez et al.109 applied direct UV
detection to analyze 19 OA in beer samples. The inner surface of the capillary was dynamically
coated to facilitate a fast anodic EOF with the addition of a polycation (hexadimetrine bromide,
TABLE.30.3
Methods for the Detection of Vitamins Separated by CE
Sample
Type of Food Detection Method Derivatization CE Mode Run Buffer CE Conditions References
Ascorbic acid in PDAD 254 nm Underivatized MEKC 0.05 M sodium deoxycholate, 56 cm × 50 µm i.d. light path 80
vegetables/blue berries 0.01 M Na2 B4 O7 , fused-silica column, 25 kV
0.01 M KH2 PO4 , pH 8.6
Niacin in cereals, meat, PDAD 254 nm Underivatized CZE 15% CH3 CN, 85% 0.01 M 56 cm × 50 µm i.d. light path 91
fruit, vegetables, and KH2 PO4 , 0.01 M Na2 HPO4 fused-silica column, 20 kV
selected foods (pH 7)
Ascorbic acid in spinach, DAD spectropho- Underivatized CZE 60 mM sodium dihydrogen 33.5 cm × 50 µm i.d. fused-silica 97
watermelon, potato, and tometric phosphate, 60 mM NaCl, capillary, −15 kV, 23◦ C
tomato 0.0001% HDM (pH 7)
l-Ascorbic acid and DAD 265 nm Underivatized CZE 0.2 M borate buffer (pH 9.0) 57 cm × 75 µm i.d. fused-silica 98
d-isoascorbic acid in capillary, 30 kV, 25◦ C

lemon, pineapple,
sunkist, spinach
Alcoholic drinks/beer UV absorbance, Underivatized MEKC 25 mM Na2 B4 O7 (pH 10.5) and 57 cm × 75 µm i.d. fused-silica 73
210 and 270 nm 110 mM SDS capillary
Riboflavins vitamers in LIF at 442 nm Underivatized CZE 30 mM phosphate buffer, pH 9.8 84 cm × 75 µm i.d. fused-silica 99
vegetables, wheat flours, capillary, 30 kV, 15◦ C
and tomatoes and
baker’s yeasts
Alcoholic drinks/wine LIF at 442 nm Underivatized CZE 30 mM phosphate buffer, pH 9.8 84 cm × 75 µm i.d. fused-silica 100
Continued
865
866
TABLE.30.3
(Continued)
Sample
Type of Food Detection Method Derivatization CE Mode Run Buffer CE Conditions References
Vegetables oils/virgin DAD at 205 nm Underivatized CEC Methanol and acetonitrile 33 cm × 75 µm i.d. × 375 µm 101
olive, hazelnut, (50/50 v/v) containing o.d. fused-silica capillary,
sunflower, and soybean 0.01% ammonium acetate partially packed with
ChromSpher C18 (3 µm),
−25 kV, 20◦ C
Ascorbic acid in soft Conductivity Underivatized CE 10 mM histidine/0.135 mM 60 cm × 50 µm i.d. × 360 µm 102
drinks detection tartaric acid, 0.1 mM CTAB, o.d. fused-silica capillary,
pH 6.5, 0.025% HP-β-CD −15 kV
Microchips- The same condition except, PMMA microchips 90 × 6 mm,
CE 0.06% HP-β-CD, with a pair of antiparallel
0.125 mM CTAB orientated electrodes (1 mm ×
1.4 cm), and with a gap of
0.5 mm
Riboflavin, flavin LIF Underivatized CZE 30 mM phosphate buffer, pH 9.8 92 cm × 75 µm i.d. fused-silica 103
mononucleotide in capillary, 30 kV, 15◦ C
wines, milk, yoghurt,
and raw eggs
Vitamin C in Lupinus UV detection Underivatized MEKC 1.08 g of deoxycholate in 50 mL 47 cm × 75 µm i.d. fused-silica 104
albus L. var. Multolupa of a 1:1 mixture of 0.02 M capillary, 18 kV, 28◦ C
borate/phosphate, pH 8.6
4
1
15 (a)
10
mAU
5
3
0
0 0.5 1 1.5 2 2.5 3 3.5 min
1
4
10
(b)
5
mAU
0
5 2 3
−10
0 0.5 1 1.5 2 2.5 3 3.5 min
3
25
1 (c)
20
mAU
15
10 2
5 4
0
0 0.5 1 1.5 2 2.5 3 3.5 min
FIGURE 30.3 CEC analysis of the TOHs content in different vegetable oils. (a) Virgin olive oil:
(b) Sunflower oil: (c) Soybean oil. 1, BHT; 2, δ-TOH; 3, γ-TOH; 4, α-TOH.: 1, BHT; 2, d-TOH; 3, g-TOH;
4, a-TOH. Oil samples were diluted 5 times (10 times for sunflower oil) with the mobile phase and injected
in the CEC system Butylated hydroxytoluene (BHT) was added as antioxidant to prevent the loss of TOHs.
(Reprinted from Aturki, Z., et al., Electrophoresis, 26, 798–803, 2005. With permission from WILEY-VCH
Verlag GmbH & Co.)
HDB) to the electrolyte. This resulting method was very useful because of the short time (22 min)
for the analysis of 19 compounds, which makes this method suitable for screening in the brewing
industry.
Recently, Suárez-Luque et al.111 developed a method for the analysis of chloride, nitrate, sulfate,
phosphate, and formic acid in honey samples by CE with indirect UV detection. The separation was
achieved using 2 mM potassium dichromate and 0.05 mM tetraethylenepentamine (TEPA), pH 4.00.
The detection limit was in the range between 0.03 and 20 mg/kg of the IA.
30.2.5 TOXINS, CONTAMINANTS, PESTICIDES, AND RESIDUES

Safety of food is a basic requirement of food quality. “Food safety” implies absence or acceptable
and safe levels of contaminants, adulterants, naturally occurring toxins, or any other substance that
may make food injurious to health on an acute or chronic basis. Most countries, therefore, have
established official tolerance levels for chemical additives, residues, toxins, and contaminants in
food products. Table 30.5 summarizes several CE reports that can be found in the literature that deal
with the analysis of toxins, contaminants, pesticides, and residues. A number of reviews appearing
868
TABLE.30.4
Methods for the Analysis of OA and IA by CE
Detection Sample
Application Method Derivatization CE Mode Run Buffer CE Conditions References
Tartrate in spiked sample Direct UV at Underivatized CZE 200 mM phosphate buffer (pH 7.5) Neutral polyacrylamide coated 108
matrices; cocoa, jam, lemonade, 214 nm capillary, 50-µm i.d., 66 cm total
sugar syrup, Madeira cake, and length (56-cm effective length) and,
digestive biscuit −14 kV, 22◦ C
OA in malt beer Direct UV at Underivatized CZE 50 mM sodium phosphate (pH 8), 57 cm × 75 µm i.d. fused-silica 109
210 nm 0.001% HDB, and 25% 2-propanol capillary, −15 kV, 23◦ C
OA in milk powder, Cheddar Indirect UV at Underivatized CZE 4.4 mM KHP and 0.27 mM CTAB 105 cm × 75 µm i.d. fused-silica 110
cheese, and plain liquid yogurt 200 nm (pH 11.2) capillary, −20 kV, 30◦ C
IA (chloride, nitrate, sulfate, Indirect UV at Underivatized CZE 2 mM potassium dichromate, 60 cm × 75 µm i.d. fused-silica 111
phosphate) and formic acid in 254 nm. 0.05 mM tetraethylene-pentamine capillary, −27 kV, 25◦ C
honey (TEPA), pH 4.0
Citric, isocitric, tartaric, and Direct UV Underivatized CZE 200 mM phosphate buffer (pH 7.50) 57 cm × 50 µm i.d. fused-silica 112
malic acids in natural and capillary, −14 kV, 25◦ C
commercial orange juices
OA in coffee Direct UV at Underivatized CZE 500 mM phosphate buffer (pH 6.25), 57 cm × 50 µm i.d. fused-silica 113
200 nm 0.5 mM cetyltrimethylammonium capillary, −10 kV, 25◦ C
bromide (CTAB)
OA (tartaric, malic, lactic, Indirect UV at Underivatized CZE 5 mM 2,6-PDC, 0.5 mM CTAB 78 cm × 75 µm i.d. fused-silica 114
succinic, and acetic acids) in 200 nm (pH 5.0) capillary, −10 kV, 18◦ C
white and red port wines
Acetic, citric, fumaric, lactic, Direct UV at Underivatized CZE 200 mM phosphate buffer (pH 7.50) 50 cm × 50 µm i.d. neutral coated 115
malic, oxalic, succinic, and 200 nm capillary, −14 kV, 20◦ C
tartaric acids, nitrate and sulfite
ions in white and red wines
OA (oftartaric, malic, acetic, Direct UV, Underivatized CZE 3 mM phosphate with 0.5 mM 60 cm × 75 µm i.d. fused-silica 116
succinic, and lactic) in wine 185 nm MTAB (pH 6.5) capillary, 15 kV, 25◦ C
Indirect UV, 7 mM phthalic acid, 2 mM MTAB,
254 nm 5% v/v methanol (pH 6.1)
IA and OA in sake Indirect UV at Underivatized CZE 20 mM 2,6-PDC, 0.5 mM CTAH, 112.5 cm × 50 µm i.d. fused-silica 72
200 nm pH 12.1 capillary, −30 kV, 15◦ C
Gibberellic acid in fermentation UV detection Underivatized MEKC 25 mM disodium tetraborate, pH 9.2, 57 cm × 50 µm i.d. uncoated 117
broth and commercial products at 214 nm 100 mM SDS fused-silica capillary, 30 kV, 25◦ C
IA (NH+ + 2+ + Conductivity Underivatized µ-chip CE 10.5 mM His, 50 mM acetic acid, Microchip 85 mm total length, 4 kV 118
4 , K , Ca , Na ,
2+ 2 mM 18-crown-6, pH 4.10
Mg2+ , Cl+ , NO+3 , SO4 ) in
alcoholic and nonalcoholic 20 mM MES/His, pH 6.0
beverages 10 mM His, 7 mM glutamic acid,
pH 5.75
Inorganic and organic acid anions Indirect UV at Underivatized CZE 3 mmol/L 1,3,5-benzenetricarboxylic 65 cm × 50 µm i.d. fused-silica 119
in beverage drinks 240 nm acid (BTA), capillary, −25 kV
15 mmol/L
tris(hydroxymethyl)aminomethane
1.5 mmol/L tetraethylenepentamine
(TEPA), pH 8.4
Citric and lactic acid in food and Conductivity Underivatized µ-chip-CE 10 mM MES/His, 2 mM 18-crown-6 Microchip 85 mm length, 3 kV 120
beverages (pH 6)
10 mM 2-(cyclohexylamino)
ethanesulfonic acid (CHES), 6 mM

arginine electrolyte, pH 9
10 mM 3-(cyclohexylamino)-
1-propanesulfonic acid (CAPS),
10 mM arginine buffer, pH 10
Malic acid in apple juice PDA Underivatized CZE 1 mM CuSO4 +1 mM L-tartaric 56 cm × 50 µm i.d. polyvinyl 121
acid, pH 5.1 alcohol-coated capillary, −20 kV,
30◦ C
869
870
TABLE.30.5
Methods for the Detection of Toxins, Contaminants, Pesticides, and Residues Separated by CE
Detection Sample
Type of Food Method Derivatization CE Mode Run Buffer CE Conditions References
Acidic pesticides UV Underivatized CE-UV 4 mM ammonium 60 cm × 75 µm i.d. fused-silica 127
(o-phenylphenol, ioxynil, ES-SM CE-MS formate-formic acid (pH 3.1), capillary, 25 kV, 25/15◦ C
haloxyfop, acifluorfen, 32 mM ammonium formate/acid
picloram) in apple, grapes, formic buffer (pH 3.1)
oranges, tomatoes
AA herbicides (glufosinate and LIF Derivatized with CZE (Cy5) 50 mM borate, 15% ACN 57 cm × 50 µm i.d. fused-silica 128
aminomethylphosphonic acid, sulfoindo-cyanine (pH 8.0) (NIR-641) 150 mM capillary, 25 kV, 25◦ C
the major metabolite of succinimidyl ester (Cy5) borate, 15% ACN (pH 9.5)
glyphosate) in agricultural soil and 1-ethyl-1-[5-(N-
samples succinimidyl-oxycarbonyl)
pentyl]-tetramethyl-indodi-
carbocyanine
(NIR-641)
Aniline metabolites LIF Derivatized with 150 µL of MEKC 20 mM boric acid, 30 mM SDS, 57 cm × 50 µm i.d. × 129
(3-chloroaniline,3-chloro- 5 mM 5-(4,6-dichloro- and 10 mM Triton X-100 375 µm o.d. fused-silica
4-hydroxyaniline and s-triazin-2-ylamino) (pH 9.5) capillary, 25 kV, 25◦ C
3-chloro-4-methoxyaniline) of fluorescein (DTAF)
chlorpropham in potato
Aflatoxins B1 , B2 , and G1 and MPE- Underivatized CZE 20 mM Tris, 10 mM 20.4 cm × 2.1 µm i.d. capillary 130
the cholera toxin stock Fluorescence carboxymethyl-β-cyclodextrin
solution buffer (pH 7.5)
Insecticide (pirimicarb) and DAD Underivatized CEC Acetonitrile/5 mM aqueous Tris 33 cm × 100 µm ID × 131
fungicide (azoxystrobin) in 40/60 v/v (pH 8.6) 375 µm OD polyimide coated
soil and tomato fused-silica capillary, 30 kV,
30◦ C
Substituted urea pesticides in DAD Underivatized MEKC 4 mM borate and 35 mM SDS 60 cm × 75 µm i.d. 132
orange and tomato (pH 9.0) polyimide-coated fused-silica
Pesticides (aldicarb, UV detection Underivatized MEKC 50 mM SDS, 10 mM borate 52 cm × 50 µm i.d. fused-silica 133
carbofuran, isoproturon, buffer, 15 mM β-CD, capillary, 25 kV
chlorotoluron, metolachlor, 22% acetonitrile, pH 9.6
mecoprop, dichlorprop,
MCPA, 2,4-D, methoxychlor,
TDE, DDT, dieldrin, and
DDE) in drinking water
Bacterial endotoxins UV absorption Underivatized CZE 0.05 M Tris–HCl buffer, pH 8.5 28 cm × 50 µm i.d. fused-silica 134
(lipopolysaccharides) in capillary, 25 kV, 10◦ C
protein complexation
Fungicide validamycin A in Indirect UV Underivatized CZE 10 mM aminopyrine, 60 cm × 75 µm i.d. fused-silica 135
rice plants detection 2 mM ethylenediaminete- capillary, 15 kV, 25◦ C
traacetic
acid (pH 5.2)
Herbicide, glyphosate ESI-MS Underivatized CE-MS Ammonium formate buffer 75 cm × 50 µm i.d. fused silica 136
(standard) (pH 2.5) capillary column, –15 kV, 25◦ C
Pesticides (pyrimethanil, UV and MS Underivatized CE-MS 0.3 M ammonium acetate/acetic 57 cm × 50 µm i.d. fused-silica 137
pyrifenox, cyprodinil, acid (pH 4.0) capillary, 25 kV, 25◦ C
cyromazine, and pirimicarb)

in grapes and orange juice
Pesticides (carbendazim , DAD at Underivatized MEKC 20 mmol/L phosphate buffer 58.3 cm × 75 µm i.d. × 375 µm 138
simazine, atrazine, propazine 200 nm (pH 2.5), 25 mmol/L SDS, o.d. fused-silica capillary,
and ametryn, diuron and 10% methanol –25 kV, 25◦ C
linuron, carbaryl and
propoxur), in drinking water
and carrot
Antibiotics [Quinolone residues ESI-MS Underivatized CE-MS 60 mM (NH4 )2 CO3 (pH 9.2) 75 cm × 75 µm i.d. × 139
(danofloxacin, enro-floxacin, 375 µm o.d. fused-silica
flumequine, ofloxacin, and capillary, 20 kV, 20◦ C
pipemidic acid)] in chicken
and fish
Continued
871
TABLE.30.5
872
(Continued)
Detection Sample
Pesticides (acrinathrin, DAD Underivatized MEKC 6 mM sodium tetraborate 57 cm × 75 µm i.d., uncoated 140
bitertanol, cyproconazole, decahydrate, 75 mM of cholic fused-silica capillary, 30 kV,
fludioxonil, flutriafol, acid sodium solution (pH 9.2) 25◦ C
myclobutanil, pyriproxyfen,
and tebuconazole) in lettuce,
tomato, grape, and strawberry
Pesticides (cyprodinil, UV Underivatized CZE 0.4 mM 57 cm × 50 µm ID, bare 141
cyromazine, pyrifenox, cetyltrimethylammonium fused-silica capillary, 22 kV,
pirimicarb, and pyrimethanil) chloride (CTAC), 0.4 M acetic 25◦ C
in water, apple, and orange acid (pH 4), 5% v/v 2-propanol
juice
Pesticide residues (dinoseb, DAD 214 nm Underivatized CZE-UV 0.3 M ammonium acetate 57 cm × 75 µm i.d. × 142
pirimicarb, procymidone, ESI-MS CZE-MS (pH 4.0), 10% methanol 375 µm o.d. fused-silica
pyrifenox, pyrimethanil, and capillary, 30 kV, 25◦ C
thiabendazole) in peaches and
nectarines
Fungicide residues ESI-MS Underivatized CE-MS 12 mM ammonium formate, 150 cm × 75 µm i.d. × 143
(procymidone and 20 mM formic acid (pH 3.5), 375 µm o.d. fused-silica
thiabendazole) in apples, 2% methanol capillary, 30 kV, 25◦ C
grapes, oranges, pears,
strawberries, and tomatoes
Pesticides and metabolites UV-visible Underivatized MEKC 30 mM ammonium 57 cm × 75 µm i.d., 375 µm o.d. 144
(naphthalene acetamide, chloride/ammonia buffer fused-silica capillary, 29 kV,
carbaryl, 1-naphthol, (pH 9.0), 15 mM SDS 25◦ C
thiabendazole, and
carbendazime) in cucumbers
Amino phosphonic acid LIF In-capillary derivatization MEKC 50 mM boric acid, 57 cm × 50 µm i.d. fused-silica 145
herbicides (model mixture) with 5-(4,6-dichloro- 30 mM Brij-35, adjusted to capillary, 5 kV, 35◦ C
s-triazin-2-ylamino) pH 9.5
fluorescein (DTAF)
Penicillin V and related UV Underivatized CZE 20 mM ammonium acetate 70 cm × 50 µm i.d. × 146
substances in mixture of a real ESI-MS (unadjusted, pH 6.5), 360 µm o.d. fused-silica
fermentation broth NACE-MS 20 mM ammonium acetate in capillary, −20 kV, 25◦ C
acetonitrile/MeOH (60/40 v/v)
Seven paralytic shellfish toxins UV at 200 nm Underivatized ITP-CE LE: morpholine buffer (pH 5.0) 57-cm-long fused-silica capillary 147
(PSTs), namely TE: 10 mM formic acid, (pH 2.7) with polyacrylamide (effective
decarbamoylsaxitoxin CZE: morpholine buffer (pH 5.0) length of 51.8 cm), 75 µm,
(dcSTX), saxitoxin (STX), 14 kV, 20◦ C
neosaxitoxin (NEO),
gonyautoxin-2 (GTX-2),
gonyautoxin-3 (GTX-3),
gonyautoxin-1 (GTX-1), and
gonyautoxin-4 (GTX-4)
Antibiotics residues UV Underivatized CZE Different concentration of 47, 57, 67, and 77 cm × 75 or 148
(amoxicillin, doxycycline borate + phosphate buffer 100 µm i.d. × 375 µm o.d.
hydrochloride, streptomycin (30 mM NaH2 PO4 ) fused-silica capillary, 25 kV,

sulfate, thiamphenicol, 25◦ C
florphenicol, nifursol,
enrofloxacin, ciprofloxacin,
norfloxacin) from poultry and
porcine tissues
Quinolones [enrofloxacin DAD Underivatized NACE 20 mM ammonium acetate 64.5 cm × 75 µm i.d. fused-silica 149
(ENR), ciprofloxacin (CPR), 0.004% polycation capillary, 30 kV, 25◦ C
danofloxacin (DAN), hexadimethrine bromide
difloxacin (DIF), (HDB), 4% acetic acid (pH 5.4)
marbofloxacin (MAR), in methanol/acetonitrile
flumequine (FLU), and (50:50 v/v)
oxolinic acid (OXA)] in pig
kidney tissue
Continued
873
874
TABLE.30.5
(Continued)
Detection Sample
E. coli in meat LIF Underivatized CE 4.5 mM Tris, 4.5 mM boric acid, 27 cm × 75 µm i.d. fused-silica 150
and 0.1 mM EDTA) capillary, 370 V/cm
polyduramide (0.1% w/v)
Six bacterial contamination in UV Underivatized CZE 25 mM phosphate buffer (pH 7.0) 47 cm × 75 µm i.d. fused-silica 151
corn flakes, milk, baby food, + 25 µM calcium chloride + capillary, 15 kV, 20◦ C
juice, and frankfurter 35 µM myoinositol
hexakisphosphate.
Glycoalkaloids in potato MS-MS Underivatized NACE-MS 90:10 v/v of MeCN-MeOH 80 cm × 50 µm i.d. × 152
containing 50 mM ammonium 375 µm o.d. fused-silica
acetate and 1.2 M acetic acid capillary, 25.5 kV, 20◦ C
Eight colorants in (orange, PAD Underivatized CE 20 mM NaOH, 15 mM disodium 50.2 cm × 50 µm i.d. fused-silica 153
apple, and grape) soft drinks tetraborate (borax) pH 10.0, capillary, 25 kV, 25◦ C
(grape, pineapple, and peach), 7 mM β-CD
jellies, and (apple) milk
beverages
Nine sulfonamides in meat UV visible Underivatized CZE 45 mM sodium phosphate, 64.5 cm × 75 µm i.d. fused-silica 154
(animal food) detection 10% methanol (pH 7.3) capillary, with bubble cell,
25 kV, 27◦ C
0.024
3
0.020
0.016
AU
14
0.012
10
7 12 13
0.008 4 5 6
9
1 11 15
0.004 8
10 15 20
Min
FIGURE 30.4 Electropherograms of a bovine milk sample spiked with 15 antibiotic standard mixture.
Conditions: 16 ppm antibiotics, 100 mM borate buffer pH 8.0, with the addition of 5 mM α-CD and
5 mM HPα-CD, 30 kV, 25◦ C, detection 200 nm. (1) Procaine, (2) trimethoprim, (3) spyramicin, (4) tylosin,
(5) enrofloxacin, (6) cloramphenicol, (7) tetraciclina, (8) oxytetracyclin, (9) clortetraciclin, (10) amoxicillin,
(11) cefalexine, (12) ampicillin, (13) dicloxacillin, (14) oxacillin, and (15) penicillin G.
in the literature over the past half decade provide an excellent encapsulation of publications on these
topics.122–126
Several drugs are frequently fed to domestic cattle because modern intensive animal breeding
demands permanent suppression of diseases caused by viruses, bacteria, protozoa, and/or fungi.
Therefore, residues of these drugs can be found in foods of animal origin such as milk, eggs, and
meat. Kowalski et al.148 reported the development and validation of CZE methodologies for eight
antibiotics and one coccidiostatic (nifursol) residues in poultry and porcine tissues. They proposed a
simplified clean-up procedure, including deproteinization by acetonitrile and liquid-liquid extraction
with ethyl acetate, for drug substances at concentrations below 20 mg/kg in a variety of food types.
Castillo and Vargas155 reported a CE method for the simultaneous analysis of different kinds of
antibiotic residues in bovine milk, which would allow for the routine monitoring of the presence of
these compounds for quality control (Figure 30.4).155
On the other hand, contaminants are substances that have not been intentionally added to food.
These substances may be present in food as a result of the various stages of their production, packag-
ing, transport, or holding, or might result from environmental contamination. In this area, Palenzuela
et al.151 proposed an excellent CE method for the detection of bacterial contamination. The method
was based on the interaction of ions with biocolloids that allows for their reliable separation of eight
different types of bacteria by CE in only 25 min—a dramatic reduction in the analysis time resulted
(7 h of enrichment vs. the 24–48 h typically required by culturing methods) compared with classical
microbiological analyses.
In the field of pesticide residues in food, CE continues to gain ground and garner significant
attention. However, one of the main drawbacks of CE is its low sensitivity in the monitoring of
pesticide residue. Therefore, the use of different preconcentration strategies before the separation
[e.g., solid-phase extraction (SPE), liquid-liquid extraction, etc.] or different stacking procedures in
the capillary column have been applied and recently reviewed.156,157 Da Silva et al.138 demonstrated
the usefulness of sample stacking techniques for pesticide analysis with the separation of nine
pesticides in complex matrices such as fruits and vegetables. Online preconcentration of the analytes
during MEKC separation was established, providing LODs of 2.5 mg/L for all analytes in carrots.
Moreover, the use of SPE, in combination with an online preconcentration strategy, allowed for the
determination of pesticides at the 0.1 mg/L level in drinking water. Other toxic compounds such
as mycotoxins130 and bacterial endotoxins,134 colorants,153 and sulfonamine residues154 have also
been analyzed and reported using CE methodologies.
30.2.6 ANALYSIS OF PHENOLIC COMPOUNDS IN FOOD

Phenolic compounds (PCs) are a group of naturally occurring secondary metabolites present in plants,
all of which possess one common structural feature—a phenol moiety, and behave as excellent
antioxidants due to the reactivity of this group. PCs are of interest due to their potential contribution
to the taste (astringency, bitterness, and sourness) and formation of off-flavor in food, including
tea, coffee, and various fruits juices, during storage. Currently, PCs are divided into two types:
simple phenols and polyphenols. Phenols are a widely found group of secondary plant metabolites
(e.g., benzoic and cinnamic acid derivatives), and polyphenols have at least two phenol subunits—
flavonoids are an example of a common polyphenol in plants. Some recent application of CE to the
analysis of PCs in tea, wines, and other foods are summarized in Table 30.6. It should be noted that
there is no need to derivatize samples of PCs because they are aromatic and, therefore, show intense
absorption in the UV region. Several reviews have been published that can be used as a guide for
the analysis of these compounds in food by CE.158−161 In a recent review, Li et al.160 discussed the
strategies that have been used during the optimization of CZE for analysis of phytochemical bioactive
compounds, and proposed the use of multivariate experimental designs to simplify this task. Blasco
et al.162 recently proposed a new and attractive route for fast, simultaneous detection of prominent
natural antioxidants, including phenolics and nonphenolic antioxidants, using a CE microchip with
electrochemical detection via a glassy carbon electrode. Separation of the pair of standard compounds
(+)-catechin and ascorbic acid, (+)-catechin and rutin, as well as arbutin and phlorizdin was carried
out in less than 400 s and later applied for the determination of these compounds in pear juice. An
interesting method for the simultaneous separation and determination of selected PCs in six different
collections of Hypericum perforatum (St. John’s wort) was proposed by Hamoudová et al.163 This
method used an online combination of CZE with capillary ITP (CE-ITP) allowing detection limit as
low as 60 ng/mL and total analysis time of less than 35 min.
Using a different mode of CE, MEEKC, and MEKC, Huang et al.181 separated 13 PCs. The
authors compared both methodologies and demonstrated that both MEEKC and MEKC methods
possess the ability to analyze PCs in different food samples. The same authors also reported a
method for the separation of the nine flavonoids most often found in grape wine, namely, apigenin,
baicalein, naringenin, luteolin, hesperetin, galangin, kaempferol, quercetin, and myricetine, using
a 35 mM borax, pH 8.9, 16.8 kV of applied voltage with an analysis time of within 16 min.182
A high-pressure liquid chromatography (HPLC) method was also developed for the same group of
compounds and later compared with the CZE. They concluded that both strategies were useful for
the analysis of PCs. An electropherogram of the flavonoid separation can be seen in Figure 30.5.182
30.2.7 CHIRAL ANALYSIS OF FOOD COMPOUNDS

Most of the major nutritional organic components of food and beverage are chiral, including proteins,
AA, carbohydrates, fats, and some vitamins. Many flavor and fragrance components of food and
beverage are chiral as well. The significance of the enantiomeric composition of food components
is that they can have a dramatic effect on aroma, taste, and nutritional value. Chiral separations
for food components are useful for evaluating age, treatment, and storage effects; for control or
TABLE.30.6
Methods for the Analysis of Phenolic Compounds by CE
Application Detection Method CE Mode Run Buffer CE Conditions References
Polyphenolic fraction of UV absorption CZE Buffer solution 45 mM of sodium 50 µm i.d., 47 cm of total length 162
extra-virgin olive oil tetraborate pH 9.3 (40 cm to the detector) with a
detection window of 100 × 200 µm
PCs in hypericum perforatum ITP conductivity detector CE-ITP LE of 10 mM HCl with Tris 0.2% ITP-fluorinated ethylene propylene 163
(St. John’s wort) CZE UV at 270 nm and 2-HEC, 20% v/v methanol pH* 7.20 copolymer (FEP), 9.0 cm ×
conductivity TE 50 mM H3 BO3 pH* 8.2, 20% v/v 60.8 mm i.d.
methanol CZE-FEP, 16 cm × 60.3 mm i.d.
CZE 50 mM Tris buffer + 25 mM
MOPSO + 65 mM boric pH* 8.3,
20% v/v methanol
PCs in olive oil ESI-MS CZE-ESI-MS 60 mM NH4 OAc at pH 9.5 with 5% of Fused-silica capillary 50 µm i.d. 164
UV at 200 nm 2-propanol (7 cm effective length for the
UV detector and 100 cm total length
for the MS detection

PCs in chess (Bromus UV at 254 nm CZE 20 mM sodium tetraborate (pH 9.2) 53.5 cm fused-silica capillary (45 cm 165
inermis L.) with 5% v/v methanol effective length), 75 µm i.d.,
365 µm o.d.
Rutin and quercetin in plants Electrochemical MEKC 20 mM borate (pH 8.8), 40 mM SDS, 50 cm × 50 µm i.d. fused-silica 166
10% acetonitrile capillary, 12 kV
Quercetin, rutin, kaempferol, UV at 270 nm CZE 100 mM borate (pH 10) 51 cm × 50 µm i.d. fused-silica 167
catechin, gallic acid in plants capillary, 15 kV, 32◦ C
Procyanidins after thiolysis UV absorption at 214 nm MEKC 50 mM phosphate (pH 7), 40 mM 47 cm × 50 µm i.d. fused-silica 168
sodium cholate, 10 mM SDS capillary, 15 kV, 25◦ C
Catechins in green tea UV absorption at 230 nm MEEKC Microemulsion of 50 mM phosphate 24 cm × 50 µm i.d. fused-silica 169, 170
(pH 2.5), SDS, n-heptane, 2-hexanol, capillary, 10 kV, 40◦ C
or cyclohexanol
Continued
877
878
TABLE.30.6
(Continued)
Catechins in green tea UV absorption at 200 nm MEKC 20 mM phosphate + 50 mM borate + 47 cm × 50 µm i.d. fused-silica 171, 172
200 mM SDS (3:1:2), (pH 7) capillary, 30 kV, 29◦ C
Theaflavin composition of UV at 380 nm NACE 71% v/v acetonitrile, 25% v/v 40 cm × 50 µm i.d. fused-silica 173, 174
black tea methanol, 0.1 M potassium capillary, 22.5 kV, 18.5◦ C
hydroxide, 4% v/v glacial acetic acid,
90 mM ammonium acetate
Resveratrol in wines, herbs, Electrochemical CZE 100 mM borate (pH 9.24) 65 cm × 25 µm i.d. fused silica 175
and health food capillary, 30 kV
Resveratrol and piceid in red DAD MEKC 20 mM sodium tetraborate, 57 cm × 75 µm i.d. fused-silica 176
wine 25 mM polyethylene glycol 400, capillary, 28 kV, 25◦ C
25 mM SDS, 10% methanol
Anthocyanins in wine Visible at 599 nm CZE 50 mM borate (pH 8.4), 15% methanol 46 cm × 75 µm i.d. fused-silica 177
Resveratrol, catechin, rutin, UV at 240 nm CZE 25 mM borate (pH 9.4) 75 cm × 50 µm i.d. fused-silica 178
quercetin, myricetin, caffeic capillary, 18 kV
acid, chlorogenic acid, gallic
acid in plants
Chlorogenic acid, ferulic acid, DAD CZE 50 mM borate (pH 9.5) 48.5 cm × 50 µm i.d. fused-silica 179
vanillic acid, caffeic acid, capillary, 20 kV, 25◦ C
catechol in coffee extracts
Phenolic acids (derivatives of Contactless conductivity CE 150 mM 2-amino-2-methylpropanol 48.5 cm × 50 µm i.d. × 50 µm o.d. 180
benzoic and cinnamic acids) detection (pH 11.6), 30 mM MES/His, fused silica capillary, 15 kV, 25◦ C
30 mM CTAB, 10% methanol (pH 6)
UV visible CEC AMPD, 300 mM Optical fibers 300 µm i.d., 330 µm
2-amino-2-methylpropanediol; AMP, clad diameter, 360 µm o.d.
150 mM 2-amino-2-methylpropanol;
MA, 50 mM methylamine
Phenolic compounds (syringic UV-Vis detector MEKC 2.89% (w/v) SDS, 1.36% (w/v) 48.5 cm × 50 µm i.d. fused-silica 181
acid, p-coumaric acid, vanillic heptane, 7.66% (w/v) cyclohexanol, capillary, −20 to −27 kV, 25-35◦ C
acid, caffeic acid, gallic acid, 2% (w/v) acetonitrile, and 86.1%
3,4-dihydroxybenzoic acid, (v/v) phosphate solution of pH 2.0
4-hydroxybenzoic acid, (25 mM)
(+)-catechin, MEKC 2.89% (w/v) SDS, 2% (w/v) methanol,
(−)-epigallocatechin, and 95.1% (v/v) phosphate solution of
(−)-epicatechin gallate, pH 2.0 (25 mM)
(−)-epigallocatechin gallate,
(−)-epicatechin, and
(−)-gallocatechin) in teas and
grapes
Flavonoids (apigenin, DAD CZE 35 mM Borax (pH 8.9) 70 cm × 75 µm i.d. fused-silica 182
baicalein, naringenin, luteolin, capillary
hesperetin, galangin,
kaempferol, quercetin, and
myricetine) in grape wine
Flavonoids (quercetin and UV-visible detection CZE-ITP LE 10 mM HCl, Tris, 20% CH3 OH Sampling, 800 V for 80 s; Stacking, 183
isorhamnetin) (pH 7.2) 3000 V (1000 V/cm) for 24 s;

TE 50 mM H3 BO3 , 20% CH3 OH Separation, 2700 V (675 V/cm)
(pH 8.2)
BGE 25 mM MOPS, 50 mM Tris,
55 mM H3 BO3 (pH 8.36)
Phenolic acids in olive oils UV detection CZE 25 mM sodium borate (pH 9.6) 57 cm × 75 µm i.d. × 375 µm o.d. 184
fused-silica capillary, 25 kV, 25◦ C
Flavonoids and phenolic Electrochemical CE 100 mmol/L borate buffer (pH 8.7) 75 cm × 25 µm i.d. × 360 µm o.d. 185
compounds (ferulic acid, fused-silica capillary, 18 kV, 20◦ C
apigenin, luteolin, rosmarinic
acid, and caffeic acid) in
Perilla frutescens L.
Continued
879
TABLE. 30.6
(Continued)
880

Flavonoids and phenolic acids Conductivity CZE-ITP LE: 10 mM HCl of pH* 7.2 with Tris, ITP-Pre-separation capillary 9 cm × 186
in red wine 20% v/v methanol. TE: 50 mM boric 0.8 µm ID, CZE-16 cm ×
acid of pH* 8.2, 20% v/v methanol. 0.3 µm ID, 25◦ C
CZE: 25mM
β-hydroxy-4-morpholino propanes
ulfonic acid (MOPSO), 50 mM Tris,
15mM boric acid and 5 mM β-CD of
pH* 8.5, 20% v/v methanol
Flavonoids and phenolic acids UV detection CZE 35 mM sodium tetraborate (pH 9.3) 55 cm × 50 µm ID fused-silica 187
(trans-resveratrol, cinnamic containing 5% v/v methanol capillary, 20 kV, 25◦ C
acid, ferulic acid, p-coumaric
acid, quercetin, and morin) in
berries
Catechins and theaflavins UV at 205 nm CE 800 mL of 500 mM boric acid Light path capillary 40 cm × 188
(+)-catechin, catechin gallate, (pH 7.2), 200 mL of 100 mM a 50 µm i.d. 25 kV, 30◦ C
(−)-epicatechin, KH2 PO4 (pH 4.5), 450 mL of 20 mM
epicatechin-3-gallate, β-CD and 550 mL of acetonitrile
epigallocatechin,
epigallocatechin-3-gallate,
theaflavin,
theaflavin-3-monogallate,
theaflavin-3 -monogallate and
theaflavin-3,3 -gallate in
green and black teas
Flavonoids and isoflavonoids in UV at 210 and 270 nm MEKC 25 mM Na2 B4 O7 (pH 10.5) and 57 cm × 75 µm i.d. fused-silica 73
beer 110 mM SDS capillary
Phenolic oligomers DAD detection MEKC 50 mM boric acid/sodium tetraborate, 27 cm × 50 µm i.d. fused-silica 189
[(+)-catechin and 100 mM SDS (pH 9 and 8.2) capillary, 10 kV, 25◦ C
(−)-epica-techin] in soaking
water from lentils, white beans
and black beans, and in food
by-products (almond peels)
Phenolic acids in red wine Amperometric Microchip CE 15 mM borate buffer (pH 9.5), 10% of Microchip 74 mm total length 2000 V 190
methanol
Luteolin
Caffeic acid
Baicalein
5
Quercetin
Galangin
Apigenin
Kaempferol
4
Naringenin
Myricetin
Hesperetin
3
mV (Response)
–1
–2
3 6 9 12 15
FIGURE 30.5 Electropherogram of the flavonoids separation. Conditions: 35 mM borax, pH 8.9, and applied
field strength of 240 V/cm on a fused capillary of 70 cm (effective length: 45 cm) × 75 µm. (Reprinted
from Wang, S. P. and Huang, K.-J., J. Chromatogr. A, 1032, 273–279, 2004. With permission from Elsevier
Science-NL.)
monitoring fermentation processes; to identify adulterated products; to understand and control flavors
and fragrances; and for fingerprint complex mixtures.
AA in food naturally occurs as the l-AA form. However, the d-enantiomers can be found in
food that have been exposed to microbial activity (fermentation, aging, etc.), highly processed
food (those that have been exposed to extremes of pH, heat, etc.), and food that naturally contains
d-AA. Excellent reviews have been published that deal with chiral separation by CE193–196,198 and
specifically those separations that are applied to food analysis.191,192,197 Table 30.7 summarizes the
most recently published scientific paper on chiral compounds present in food and beverages.
Nouadje et al.199 proposed a method some years ago using β-cyclodextrins (β-CDs) and SDS
to determine l- and d-AA previously derivatized with FITC. A number of AA, including l-proline,
l-aspartic acid, d-Asp, l-serine, l-asparagine, l-glutamic acid, d-Glu, l-alanine, l-arginine, d-Arg,
and the nonchiral γ-amino-n-butyric acid (GABA), were analyzed in orange juice. The buffer con-
sisted of 100 mM sodium tetraborate, 30 mM SDS, and 20 mM β-CD at pH 9.4, with an applied
voltage of 23 kV. It was shown that l-Arg, l-Asp, and GABA were the most important com-
pounds for the classification of commercial orange juices to provide useful information about quality
and the efficacy of processing. Simo et al.200 reported the use of stepwise discriminating analysis
for 26 standard samples of commercial orange juices (i.e., nectars, orange juices reconstituted from
TABLE. 30.7
882
Methods for the Detection Chiral Analysis of Food Compounds Separated by CE

Detection Sample
Chiral AA content in LIF Derivatized with 200 m MEKC 100 mM sodium tetraborate 57 cm × 50 µm i.d. fused-silica 200
orange juices 3.75 mM FITC 20 mM β-CD, 30 mM SDS capillary, 23 kV, 25◦ C
(pH 9.4)
Chiral AA in vinegars LIF Derivatized with 200 mL of MEKC 100 mM sodium tetraborate, 57 cm × 50 µm i.d. fused-silica 202
3.75 mM fluorescein 30 mM SDS, 20 mM β-CD capillary, 15 kV, 20◦ C
isothiocianate (FITC) (pH 9.7)
Dextrin 10, β-CD,
hydroxypropyl-β-CD, and
dimethyl-β-CD
Fungicide (Imazalil Direct detection at Underivatized CE 5 mM ammonium dihydrogen 64.5 cm × 75 µm i.d., 203
Residue) in orange 200 nm phosphate + 50 mM phosphate fused-silica bubble cell
buffer (pH 3.0), 4 mM capillary, 25 kV, 20◦ C
2-hydroxypropyl-β-cyclodextrin
Fungicide (vinclozolin) in Direct detection at Underivatized MEKC 20 mM phosphate, 5 mM borate 64.5 cm × 75 µm i.d. fused-silica 204
wine 203 nm buffer (pH 8.5), 50 mM γ-CD, capillary, 20 kV, 20◦ C
100 mM SDS
Biogenic dl-amino acids UV detection at Derivatized with 1 mL of CZE 20 mM borate buffer (pH 9.2) 73 cm × 75 µm i.d. open-tubular 205
(Ala, Phe, Tyr, Ser, Cys, 200 nm 2 mM N-fluorenylmethoxy (0, 4, 10 mM γ-CD) silica capillary, 15 kV
Met, Val, Leu, Ile, Thr, carbonyl-l-alanyl MEKC 20 mM borate buffer (pH 9.2),
His, Pro, Trp, Arg, Lys, N-carboxyanhydride (0, 10, 20, 80 mM) SDS
Glu, Gln, Asp and Asn) (FMOC-l-Ala-NCA) in
in solid-phase peptide acetone
synthesis (SPPS)
AA enantiomers UV absorption Derivatized with 9-fluoreny CCE 150 mM borate and 18% v/v 60 cm × 75 µm i.d. fused-silica 206
methyl chloroformate isopropanol (pH 8.0), 30 mM capillary, 15 kV, 25◦ C
(FMOC) β-CD and 30 mM STDC
β-CD and sodium taurodeoxy
cholate (STDC) as selectors
Chiral analysis of Direct UV at Underivatized CZE 60 mM phosphate buffer (pH 7.0) Fused-silica bubble cell capillary 207
pantothenic acid in a soft 200 nm + 60 mM 2-hydroxypropyl- of 56 cm, 375 mm i.d., 20 kV
drink β-cyclodextrin and 10% (v/v) at 15◦ C
methanol
Monosaccharides DAD-UV Derivatized with 1-phenyl- CZE 200 mM borate and 200 mM 80.5 cm × 50 µm i.d. fused-silica 208
(mannose, galactose, 3-methyl-5-pyrazolone (S)-3-amino-1,2-pro-panediol capillary, 20 kV, 10–50◦ C
fucose, glucose, xylose, (PMP) (SAP), pH 9.2
and arabinose)
Standards of UV detection Derivatives of MEKC 20 mM borate buffer (pH 9.0) + Fused-silica capillary, 75 µm i.d., 209
α-aminoisobutyric acid at 492 nm chloroethylnitro-sourea of 20 mM Brij 58 20 kV, 25◦ C
(AIB) and isovaline LIF 488 nm benzyloxycarbonyllysine 5 mM carbonate buffer (pH 10.3) UV: 64.5 cm total length
excitation, 520 nm (tert-butyl ester) + 20 mM Brij 58 (56.5 cm to detector)
emission (CENU-Lys) LIF: 60 cm total length (55 cm to
the detector)
Halogenated AA DAD Underivatized Ligand 20 mM HO-l-4-Hypro, 10 mM 38.5 cm × 50 µm i.d. fused-silica 210
l-4-hydroxyproline, exchange CE Cu(II) sulfate, in 5 mM capillary, 5–20 kV, 25◦ C
l-histidine (HP-l-4-Hypro) (LE-CE) phosphoric acid solution, pH 4.5,
and (HO-l-4-Hypro) as injection, 10 mbar 61 s; 10 kV
chiral selectors
Flavanone-7-O- DAD Underivatized CCE 0.2 M borate buffer (pH 10.0), 77 cm × 75 µm i.d. × 375 µm 211
glycosides (naringin, 5 mM γ-CD as chiral selector o.d. fused silica capillary, 30 kV,
prunin, narirutin, 25◦ C
hesperidin,
neohesperidin, and
eriocitrin) in lemon juice
dl-Lactic acid yoghurts, Direct detection at Underivatized CZE 90 mM phosphate buffer (pH 6.0), 50 cm × 50 µm i.d. 212
and beverages (wine, 200 nm 240 mM (PVA)-coated bubble cell
sake, beer, and a soft 2-hydroxypropyl-β-cyclodextrin capillary, −30 kV, 16◦ C
drink)
dl-tartaric acid in grape Direct detection at Underivatized LE-CE 1 mM copper(II) sulfate 48.5 cm × 50 µm i.d. 213
juices, wines, soft 250 nm 10 mM d-quinic acid (pH 5.0) (PVA)-coated bubble cell
drinks, sakes, cooking capillary, −15 kV, 30◦ C
sakes, jams, candies,
tablet candies, and
pickles
Continued
883
884
TABLE. 30.7
(Continued)
Detection Sample
Dansyl AA and phenoxy UV detection Underivatized CEC AA: 1.6 mM Sodium phosphate Capillary column: 27 cm × 214
acid herbicides composed of 20% v/v acetonitrile 100 µm i.d. × 360 µm i.d.
and 80% v/v aqueous phosphate fused-silica tubing, 15 kV
buffer (pH 5.5)
Herbicides: 2 mM sodium
phosphate composed of 60% v/v
acetonitrile and 40% v/v aqueous
phosphate buffer (pH 6.0)
AA and nonprotein AA LIF Derivatized with CEC P1: 5 µm silica particles modiofied Capillary column 30 cm × 75 µm 215
4-fluoro-7-nitro-2, 1, with (S)-N-3, 5-dinitrobenzoyl-1- i.d., 0.83 kV/cm 0.50 kV/cm
3-benzoxadiazole (NBD-F) naphthylglicine
P2: 5 µm silica particles modiofied
with (S)-N-3, 5-dinitrophenyl
aminocarbonyl-valine 5 mM
phosphate buffer (pH 2.5)
acetonitrile 30:70
Standards of AA LIF Derivatized with FITC MEKC 100 mM borate buffer (pH 9.5) 74.4/99 cm × 50 µm i.d. 216
30 mM SDS fused-silica capillary, 22 kV,
β- and γ-CD as chiral selectors 25◦ C
Standards of aliphatic Conductivity Underivatized On-chip CE 0.5 M acetic acid 60 cm × 50 µm i.d. × 375 µm 217
amines 5 mM of DM-β-CD o.d. fused-silica capillary,
5 mM of (+)-18C6 H4 (pH 2.45) 15 kV, 22◦ C (PMMA),
90 mm × 16 mm
Standards of AA UV-visible Underivatized CZE 20 mM ammonium acetate (pH 5) 35 cm × 50 µm i.d. × 375 µm 218
2.5 mM vancomycin o.d. fused-silica capillary,
–15 kV, 20◦ C
Aromatic AA and UV detection at Underivatized CZE 60 mM phosphoric acid buffer, 502 mm × 50 µm i.d. 219
l-mimosine in extracts 200 nm pH 2.5 containing 175 mM fused-silica capillary, 25.1 kV,
of plant seeds and HP-α-CD 60 mM phosphoric acid 20◦ C, 25.1 kV, 18◦ C
nonprotein AA buffer (pH 2.5) containing
2.3 mM HS-γ-CD
Aromatic AA DAD Underivatized CZE 20 mM phosphate (pH 2.0), 55 mm × 25 µm i.d. fused-silica 220
(phenylalanine, tyrosine, 0.5 mM S-β-CD, and 0.55% capillary, 30 kV, 25◦ C
and tryptophan) dextran sulfate
Flavonoids medicarpin UV visible Underivatized CZE 2 mM hidroxypropyl-β- 50 cm × 50 µm i.d. × 221
and vestitone from ciclodextrin 20 mM 375 µm o.d. fused-silica
transgenic legumes hidroxypropyl-γ-ciclodextrin capillary, 15 kV, 25◦ C
extracts 25 mM borate (pH 10.0), 10% v/v
methanol
Standards mixture of AA UV at 254 nm Underivatized CZE 100 mmol/L MES + 10 mmol/L Polyacrylamide coated capillary 222
80 cm total length (54 cm to the
His (pH 5.2) + 20 mmol/L

MA-β-CD + 2 mmol/L γ-CD detector) × 75 µm i.d., 24 kV
Chiral AA in orange LIF Derivatized with FITC MEKC 100 mM sodium tetraborate, 50 cm × 50 µm i.d. fused-silica 223
juices and orange 30 mM SDS (pH 9.4), 20 mM capillary, 20 kV, 15◦ C
concentrates β-CD
885
concentrates, and pasteurized orange juices not from concentrates) and selected l-Arg, l-Asp, and
GABA as the most important variable l-AA to differentiate the samples. With the use of β-CDs
as the chiral selector to separate the l-AA, it was possible to classify correctly the 100% of the
samples.
Carlavilla et al.202 recently published a chiral MEKC-LIF procedure that allows fast and sensitive
analysis of the 11 main l- and d-AAs typically found in vinegars. The separation of the 11 compounds
could be achieved in a relatively short time (less than 20 min) with a LOD down to 16.6 nM. It was
shown that using these profiles resulted in a very useful strategy for characterizing vinegars.
Gel-Moreto et al.211 reported, for the first time, the complete separation of the 2R- and
2S-diastereomer of six selected flavanone-7-O-glycosides (naringin, prunin, narirutin, hesperidin,
neohesperidin, and eriocitrin). CE measurements were done using 0.2 M borate buffer, 5 mM γ-CD,
pH 10.0, and later applied for the separation of 2S- and 2R-diastereomers of eriocitrin and hesperidin
in lemon juice. Another application of chiral analysis in food by CE was published by Liu et al.201
They developed a method for the determination of 1-aminocyclopropane-1-carboxylic acid (ACC) in
apple tissues based on the derivatization of ACC with 3-(2-furoyl) quinoline-2-carboxaldehyde (FQ)
and the use of CE-LIF (488 nm). Optimal conditions for the separation of ACC were obtained using
20 mM borate buffer, pH 9.35, with 40 mM SDS and 10 mM Brij 35, allowing for the analysis of
ACC in crude apple extracts to be done with an impressive LOD of 10 nM. Finally, Cinquina et al.76
reported the determination of histamine in tuna fish samples. Histamine is a degradation product of
the bacterial decarboxylation of the AA histidine, which is present at substantial concentration in
fish tissues. A CE method and an HPLC method with diode array detection were compared showing
that both techniques were associated with correlation coefficients that exceeded 0.999, and LODs of
1 and 2 mg/kg for HPLC, and 0.5 and 1 mg/kg for CE.
30.2.8 PROTEINS AND PEPTIDES

Proteins are one of the main components of foods and possibly one of the most challenging to
analyze, not only because of the heterogeneity of the fractions but also because of their tendency to
aggregate. Thus, it is not surprising that CE has evolved as an alternative technique to the classical gel
electrophoresis for the analysis of proteins and peptides with the advantage that it allows for online
detection and automation. In fact, the analysis of food proteins is one of the main applications of CE in
food analysis, with most focus on the analysis of dairy proteins, followed closely by cereal proteins.2
The third category includes other food proteins such as those from meat, eggs, and others.2 Table 30.8
summarizes some selected applications for the analysis of food proteins by CE. For a comprehensive
review, including applications up to 1999, the reader is referred to Frazier et al.5 and to Cifuentes2
for a more updated review including applications from June 2002 to June 2005. For specific reviews
on food proteins of animal origin, the review by Recio et al.224 describes CE methods of analysis for
milk, egg, meat, and fish proteins and peptides, with special emphasis on dairy products. In addition,
various hurdles associated with food analysis technology, such as the assessment of technological
processes, quality, and authenticity control of animal foods, were considered.224 Similarly, Bean and
Lookhart225 offer a specific review covering methods and applications for the separation of three
major groups of food proteins: meat, dairy, and cereal proteins.
The first prerequisite for the analysis of proteins is their solubilization from the food matrix.
Nonliquid food sources usually require homogenization before extraction to allow efficient recovery
of proteins from the material.226 For this purpose, milling, blending, homogenizing, and the use
of ultrasound (sonication) are common methods. The homogenization step may be combined with
extraction of the proteins using suitable extraction buffers, which may give a selective extraction of
specific groups of proteins, often combining precipitation, filtration, and dialysis steps.226 Defining
a strategy for isolation of a protein or a group of proteins from a specific food matrix will usually
encompass a choice of method chosen on the basis of known characteristics of the protein(s) that
TABLE. 30.8
Food Proteins Analyzed by CE
Detection
Application Method Sample Pretreatment CE Mode Run Buffer References
Caseins and whey UV at 214 nm Milk, diluted (1:5) with 10 mM phosphate CZE 50 mM phosphate buffer (4 M urea, 0.1% 230
proteins buffer (4.8 M urea, 0.2% Tween 20, Tween 20, pH 2.5), coated capillary
pH 2.5), heating 5 min at 40◦ C
UV at 214 Milk, diluted (1:5) with 5 mM sodium citrate CZE 10 mM phosphate or citrate buffer 231
(5 mM DTT, 6 M urea, pH 8.0) (0.05% MHEC, 6 M urea, pH 2.5),
coated capillary
UV at 214 Milk, diluted (1:5) with 167 mM Tris, CZE 20 mM sodium citrate buffer (0.05% 232
42 mM MOPS, 67 mM EDTA, 17 mM MHEC, 6 M urea, pH 3.0), coated
DTT, 10 M urea capillary
Caseins and whey UV at 214 Milk, cheese, 167 mM Tris, 42 mM MOPS, CZE 0.48 mM citric acid −13.6 mM 233
proteins, denatured 67 mM EDTA, 17 mM DTT, 8 M urea, trisodium citrate in 4.8 M urea (pH 2.3),
β-lactoglobulin, 0.5 g/L MHEC coated capillary
para-κ-casein
Caseins and whey UV at 214 Purified proteins, denatured by addition of MEEKC 3 mM sodium borate, 8.2 mM SDS 234
proteins 10% SDS and 7% DTT, heating 3 min at (pH 9.5), uncoated capillary
100◦ C
Whey proteins, separation UV at 214 nm Milk, whey proteins prepared by casein CZE 0.05 M borate buffer, Tween 20 0.1%, 235
of β-lactoglobulin precipitation diluted (1:4) with 8.25 mM pH 8.0, coated capillary
variants A and B borate buffer, 0.1% Tween 20, pH 8.0
Cereal storage proteins UV at 200 nm Various cereals, extraction with 50% CZE 50 mM iminodiacetic acid IDA, 20% 246
1-propanol, glutenins with 50% 1-propanol, AcN, 0.05% HPMC
1% DTT
Sarcoplasmic/myofibrillar UV at 214 Meat, sarcoplasmic proteins extracted in MEKC SDS-CE, nm, Bio-Rad SDS run buffer 247
meat proteins bidistilled deionized water, myofibrillar
extracted in 0.6 M NaCl/0.01 M phosphate
buffer, 0.5% polyphosphates, pH 6.0,
Biorad SDS sample buffer
Lysozyme from egg white MS Meat adulterated with chicken egg white CE-MS 75 mM ammonium acetate/acetic acid, 249
pH 5.5, coated capillary
887
include molecular weight or isoelectric point; however, some general strategies as outlined by
Sorensen et al. 226 may be useful as a starting point before performing a CE analysis.
Similarly, the separation of proteins by CE is highly dependent on the characteristics of the
proteins, especially the isoelectric points of the proteins and also information on the differences
between the proteins to be separatedare important.226 The most important point to consider in the
choice of buffer systems used in analyzing proteins is how to prevent or control the adsorption of
proteins to the capillary wall.226 Adsorption of the proteins to the silica surface of the capillary wall
is the main reason for the observed efficiency loss, poor reproducibility in migration times, and low
protein recovery.2 This problem may be solved by the use of coated capillaries.226,227,228A coating
polymer was demonstrated to provide good separations of acidic and basic proteins, including pro-
teins from whey.227 Coating regeneration was achieved by flushing the capillary between injections
with an aqueous solution of the polymer.227 Alternatively, the ionization of the silanol groups can be
suppressed by working with background electrolytes (BGE) with low pH values or with a buffer pH
above the pI of the sample protein and the use of dynamic coatings.226,228A comprehensive review of
different dynamic and static capillary coatings for reducing protein wall adsorptions offered different
alternatives for solving the problem.228
The separation of dairy proteins by CE has been generally carried out by CZE and has been
exhaustively covered in several review papers,2,224,225,229 thus Table 30.8 only presents the key
methodologies that offer the reader an overview of their most distinctive features. Basically, dairy
protein analysis has been performed in whole milk for the simultaneous determination of caseins
and whey proteins, or in fractions isolated from milk after casein precipitation. The first approach
being used when the quantitative determination of the major proteins is required for the calculation
of casein/whey protein ratios or for authentication purposes where an analysis of the whole protein
profile is required. In both cases, accurate quantitative data must be derived. However, few studies
have addressed the analysis of both groups of proteins in a single run by presenting quantitative data
based on calibration curves constructed with analytical standards and good recovery of all proteins
from milk samples.
The first method reported to provide quantitative recovery of major whey and casein proteins
in a single run was by Vallejo-Cordoba (Table 30.8).230 With this method,230 sample preparation
was critical for the quantitative recovery of all proteins in their native state; while in most other
methodologies, sample preparation included a strong denaturing reducing buffer.231,232,233 Sample
and run buffer contained the minimum concentration of urea (4 M) that allowed good casein solubility.
Keeping urea to a minimum concentration was essential to prevent β-lactoglobulin denaturation and
to minimize urea crystallization.230 In addition, the nonionic detergent, Tween, was also used to
help with maintaining caseins in solution. A typical electropherogram of whey proteins and caseins
in fresh milk separated in a coated capillary at pH of 2.5 is shown in Figure 30.6.
Several groups have applied the method of de Jong et al.231 with modifications to monitor milk
proteins; these conditions not only allowed the separation of the individual milk proteins but also
some of their genetic variants.7,232,233 Although the original method offered by de Jong et al.231 was
the first method described for simultaneously determining casein and whey proteins, quantitative
data were not presented. The original method of de Jong231 included a reducing denaturing sample
buffer at pH 8.5 and a running buffer at pH 2.5 or 3.0 containing a cellulose polymer. To minimize
protein and capillary wall interactions, separation was carried out in a coated capillary. However,
this method was later modified to minimize protein absorption by optimizing the sample and running
buffer (Table 30.8).232,233
Finally, a third method234 was based on MEKC, where proteins were separated after complete
denaturation with SDS and dl-dithiothreitol in uncoated capillaries at pH 9.5. Although the method
had the advantage of being very rapid (separation completed in less than 90 s), it was not quantitative
(Table 30.8). The second approach reported for the analysis of dairy proteins was the analysis of
the whey fraction after casein precipitation. Unlike the methods described earlier, separations were
carried out in uncoated capillaries using polymeric additives, a high ionic strength, and high pH
20.00
mAu
6
14.98
6
9.96
3
5
4.94 5
1
4
5 6
5
0.00
0.00 6.0 12.00 18.00 24.00 30.00
FIGURE 30.6 Typical electropherogram of proteins in fresh milk. (1) β-lactoglobulin, (2) bovine serum
albumin, (3) α-lactalbumin, (4) conalbumin (added internal standard), (5) α-casein and (6) β-casein (From
Vallejo-Cordoba, B. J. Cap. Elec., 4, 219, 1997. With permission.)
separation buffer, which collectively reduced adsorption of proteins to the capillary wall.229 Several
examples of this strategy were reviewed;224,229 however, few of these methods were demonstrated to
be quantitative.229 To this end, Olguin-Arredondo and Vallejo-Cordoba235 presented a method that
allowed the quantification of genetic variants of β-lactoglobulin (Table 30.8). Excellent resolution
for variants A and B was observed under the established conditions (Figure 30.7).
Most applications for the analysis of dairy proteins are based on the methods described earlier
(Table 30.8). However, since UV detection was used, the detection sensitivity of these methods is
usually limited to proteins in the µM range.236 Thus, to overcome this drawback, an on-capillary
derivatization method with the fluorogenic reagent 3-(2-furoyl)quinoline-2-carboxaldehyde (FQ)
and LIF was developed for the determination of dairy proteins.236 This method allowed the determi-
nation of trace amounts of β-lactoglobulin in hypoallergenic infant formulas.237 CE was used for the
determination of added rennet casein and caseinate to processed cheeses238 and for the simultane-
ous quantitative determination of bovine, ovine, and caprine casein fractions in Mexican unripened
cheese (Panela).239 CE was also useful for characterizing milk protein hydrolysates to investigate
the molecular basis for differences in bioactivity,240 for monitoring proteolysis of casein in packaged
pasteurized milk during refrigerated storage in relation to hygienic and microbiological character-
istics of starting raw milk,4 and for peptide mapping in the investigation of milk protein genetic
variants.242
Cereal proteins are important not only for their nutritional quality but also for their functional
role in foods. The analysis of cereal proteins by CE has been covered in several reviews.228,243–245
Cereal proteins are complex mixture of proteins that are often difficult to solubilize and separate.
Because of this, a wide range of analytical techniques including CZE or sodium SDS-CE have
mAU
35
30
25 β-lactoglobulin A
-lactoalbumin
20
α-lactoglobulin B
15
10
0
5.5 6 6.5 7 7.5 8 8.5 9 9.5 min
FIGURE 30.7 Separation of β-lactoglobulin variants A and B in cow’s milk. (From Olguin-Arredondo, H.
and Vallejo-Cordoba, B., J. Elec. Microchip Tech., 6, 145, 1999. With permission.)
been used.243 These reviews also present the applications of CE for cultivar identification, classifi-
cation, and prediction of quality.243–245 Since the objective of most cereal protein applications was to
qualitatively characterize protein profiles, most method development efforts were toward improving
resolution, reproducibility, and speed of the analysis. A good example was an improved method that
allowed rapid (2–8 min) separations of grain proteins for several cereals (wheat, oats, rice, barley,
and rye) with high resolution and reproducibility using isoelectric buffers (Table 30.8).246 These
buffers allow the use of extremely high voltages, which in turn produces very rapid separations
and resolution.246 Also, it was shown that acetonitrile worked as well as urea in solubilizing maize
and sorghum proteins, the most hydrophobic storage proteins. In addition, some cellulose polymer
modifier was added to the buffer to prevent protein–capillary wall interactions.246
CE has also been successfully applied to the study of muscle proteins, and some of these appli-
cations have been recently reviewed.6,224,244 Methods for the determination of muscle proteins were
based on CZE, SDS-CE, or isoelectric focusing (CEIF).6,224,244 Meat species identification was car-
ried out by analyzing sarcoplasmic or myofibrillar proteins by a replaceable polymer-filled SDS-CE
method (Table 30.8).247 However, the analysis of sarcoplasmic protein profiles allowed better differ-
entiation among beef, pork, and turkey meat (Figure 30.8).247 The importance of sample preparation
in the established method was highlighted since sarcoplasmic proteins extracted by simply homoge-
nizing meat with cold bidistilled water were most useful for meat species identification when protein
profiles were examined by linear discriminant analysis.248 On the other hand, myofibrillar proteins
extracted with 0.6 M NaCl/0.01 M phosphate buffer with 0.5% polyphosphates (pH 6.0) were not
useful for raw meat species identification, although they may be of importance in the identification
of heat-processed meats.248
The analyses of egg proteins by CZE, CIEF, or MEKC were reviewed by Recio et al.224 In
addition, an interesting application of CE coupled to mass spectrometry (CE-MS) was the detection
of lysozyme from chicken and turkey egg white.249 Since one of the problems of protein separations
was the possible interaction with the capillary wall, a polymer coating compatible with CE-MS was
developed, and the usefulness of this method was shown by the analysis of a minced meat extract
containing chicken egg white as adulterant (Table 30.8).249
5.00
Beef
1
4 9
3.72 10
1
7
2.44 6
5
12
1.16 3
2
–0.11
0.00 4.00 8.00 12.00 16.00 20.00
9.00
Pork
1
6.69
10
mAU
4.38
6
11
2.07
4 89
5 7 12
2
–0.24
0.00 4.00 8.00 12.00 16.00 20.00
5.00
Turkey
3.71 1
89
10
11
2.42
12
6
5
7
1.13
2
–0.16
0.00 4.00 8.00 12.00 16.00 20.00

Min
FIGURE 30.8 CE-SDS electropherograms of sarcoplasmic proteins extracted from beef, pork and turkey
meat. (From Cota-Rivas, M. and Vallejo-Cordoba, B., J. Cap. Elec., 4, 197, 1998. With permission.)
30.2.9 DNA AND MICROCHIPS

Unlike food proteins, nucleic acids have no nutritional value but are characteristic of the various
biological components in complex products. Thus, the analysis of nucleic acids in food allows
control laboratories to determine the presence or absence of certain ingredients in complex products
or the identification of specific single food components.250 These analyses were based on nucleic acid
probes, including the polymerase chain reaction (PCR), which made the detection of minute amounts
of nucleic acids and their sequence determination possible.250 Also, as DNA is more thermostable
than many proteins, analyses using nucleic acids are less liable to be disrupted by processing of
foods. Thus, nucleic-acid-based technologies are developing rapidly and the use of suitable methods
by food control laboratories has the potential to greatly simplify methods for food authentication.251
The use of molecular techniques and CE has tremendously facilitated analytical procedures since
this combination has the benefit of high specificity and sensitivity of molecular techniques coupled
with the high resolving power and automation of CE.2
DNA-based analysis methods are highly dependent on the DNA extraction and purification
techniques. In particular, the application of molecular methods to food samples requires stringent
extraction and purification strategies, which ensure efficient recovery of nucleic acid and removal
of the numerous compounds inhibiting PCR assay.252 A comparison of DNA extraction methods
for food analysis was recently reported that highlighted the efficiency of two different commer-
cial kits.252 Also, four different DNA extraction methods from maize flour in genetically modified
organisms (GMOs) in foods were compared and the SDS/proteinase K method was chosen as the
most convenient.253 Thus, these studies highlighted the need for suitable DNA extraction methods
to obtain highly purified nucleic acids without inhibitors before PCR-CE can be carried out.
PCR-based techniques combined with CE separations has resulted in powerful methods in two
specific areas of food analysis, namely, food authentication and microbiology. PCR-CE applications
for meat authentication were reviewed by Vallejo-Cordoba et al.,6 while applications including other
foods were reviewed by Cifuentes2 and Kvasnicka.4 An overview of the combined use of PCR and CE
for the detection of transgenic foods, meat species identification, and microbial analysis discussing
advantages and drawbacks of these combined techniques was presented.254 Specific applications of
food authentication consisted in the detection of GMOs255 and the identification of meat species256
or in beef sexing;257 while food microbiology applications included the detection of food-borne
pathogens,258 food-spoilage bacteria,259 and the characterization of toxigenic fungi in dry, cured
meat.260
Commercial use of transgenic plants and other GMOs has raised several ideological and ethical
issues in recent years. Therefore, this has created a demand for analytical methods that can detect and
quantify the amount of GMO in foods.254 A method that combines PCR and CGE-LIF was developed
for the quantification of genetically modified Bt maize in foods.255 The method developed was based
on the coamplification of specific DNA maize sequences with internal standards using QC-PCR.255
Different DNA isolation and amplification techniques, which are being used for detecting GMOs in
foods, addressing quantitative aspects were reviewed.261
In the past, species identification in muscle foods routinely involved the detection of species-
specific proteins when attempting to discern the origins of the material, but they were not without
problems. The processing of foods by heating can cause denaturation of the proteins under study and,
in addition, protein expression is usually tissue dependent. Thus, study in the past decade has seen
DNA replace protein in species identification owing to the protein’s instability at high temperatures,
and to the fact that the DNA’s structure is conserved within all tissues of an individual.251 To this end,
a PCR-RFLP method that used fluorescence sensor CE for DNA fingerprinting of pork, goat, and beef
generated by restriction enzyme digestion following a fluorescent-labeling PCR amplification was
reported.256 This method was based on the amplification of the mitochondrial 12S rRNA gene with
a unique primer pair and incorporated a fluorescent-labeling nucleotide with sufficient specificity
to detect the three animal species. The method could reliably identify pork, goat, and beef and
semiquantify any of these meats when they were present in meat mixtures at levels less than 1%.256
Although QC-PCR by CE for the determination of meat species has not been reported, this
methodology was useful for quantitatively analyzing GMOs in foods. An alternative method, real
time PCR (RT–PCR), is gaining popularity over QC-PCR for the quantification of GMOs in food
samples, although these methods are still under development for the simultaneous detection of several
transgenes. Also, the interlaboratory reproducibility of RT–PCR was very low since %RSD values of
40% were reported.255 In addition, although instrumentation required for RT−PCR or QC-PCR-CE
may be costly for most food laboratories, CE is a more versatile technique because a wider range of
analytes may be determined with the same instrumentation. Thus, QC-PCR-CE offers good potential
for the quantitative determination of meat species or soy in heat-processed products.
Several of these food DNA-based analyses were recently translated to microchips as shown from
the applications reviewed by Cifuentes.2 An exemplary application of microchips was an improved
PCR-RFLP method for fish species identification.262 The objective of the improved method was to
replace the gel-electrophoretic steps for fragment separation, detection, and analysis by employing a
chip-based CE system (Agilent 2100 Bioanalyser) to analyze PCR-RFLP fingerprints. The use of this
system allowed simultaneous postrestriction digestion analysis of 12 samples in under 40 min, which
is a considerable time reduction over conventional gel-based methods. In addition, repeatability was
less than 3%, allowing species identification without the need to run reference materials with every
sample. Using DNA admixtures, discrimination of 5% salmon DNA in trout DNA was detected.262
This improved lab-on-a-chip method was applied to the development of PCR-RFLP profiles on a
commercial microchip electrophoresis instrument that can be used to identify a range of white fish
species without the need for concurrent analysis of reference materials. The method was applied to
a range of products and subjected to an interlaboratory study carried out by five UK food control
laboratories. One hundred percent correct classification of single species samples and six of nine
admixture samples was achieved by all the laboratories. The results indicated that the fish species
identification could be carried out using a database of PCR-RFLP profiles without the need for
reference materials.263
30.2.10 FOOD ADDITIVES

Food additives are substances added to food with the sole purpose of preserving it or improving its
flavor and appearance. The extended use of additives has made necessary the development of new
analytical procedures that are able to characterize them to ensure that food manufacturers comply
with labeling regulations. A comprehensive review was reported by Boyce,264 focusing on antiox-
idants, preservatives, colorings, and sweeteners added to food. In addition, an exhaustive survey
of capillary electromigration methods to analyze natural antioxidants was presented together with
some discussion of the use of these substances as functional foods.161 Most recently, an updated
overview on this subject was reported by Cifuentes.2 In general, CZE and MEKC methods with UV
detection were used for the analysis of food additives. A very useful application that demonstrated
the versatility of MEKC for the simultaneous rapid separation of preservatives, sweeteners, preser-
vatives, and colors in soft drinks was reported by Frazier et al.265 Minimum sample preparation
was required since soft drinks were only subjected to degassing by sonication. Resolution between
all additives was achieved within a 15-min run using carbonate buffer containing an optimum SDS
concentration at the micellar phase. By using the UV-visible range (190–600 nm), the identity of
sample components was confirmed by spectral matching relative to standards.265 Food additives
that were used in this analysis included a range of seven synthetic food colors (quinoline yellow,
sunset yellow FCF, carnosine, ponceay 4R, brilliant blue FCF, green S, and black PN), three artifi-
cial sweeteners (acefulfame K, aspartame, and saccharin) and two preservatives (benzoic and sorbic
acid). Although this example required minimum sample preparation, sample matrix interference is
directly related to food complexity. The analysis of eight colorants in milk beverages required prior
sample pretreatment with a polyamide SPE column.266 Baseline resolution of the carminic acid, a
natural colorant, and seven synthetic dyes was achieved within 9 min by using a tetraborate buffer
contaning β-CD as additive. Also, the recoveries of the eight food colorants from milk beverages
were better than 85% with detection levels of less than 0.5 µg/mL.266
On the other hand, the development of a MEEKC allowed the determination of the same food
colorants with most instances not requiring sample pretreatment. The effects of SDS surfactant,
organic modifier (acetonitrile), cosurfactant, and oil were examined to optimize the separation. A
highly efficient MEEKC separation method, where the eight colorants were separated with baseline
resolution within 14 min, was achieved by using a microemulsion solution of pH 2.0 containing
3.31% SDS, 0.81% octane, 6.61% 1-butanol, and 10% acetonitrile.267 To this end, an electrolyte
composed of tetraborate (TBS), Brij 35, and acetonitrile (ACN) was optimized for the separation
of 11 dyes allowed in Brazil providing baseline separation in less than 9 min. The optimization
procedure that provided baseline separation in less than 9 min was aided by a factorial design.268
Other applications of CE to analyze food additives include the determination of vitamin C
and preservatives (benzoate and sorbate) by both conventional CE and microchip electrophoresis
with capacitively coupled contactless conductivity detection.269 The separation was optimized by
adjusting the pH value of the buffer and the use of hydroxypropyl-β-CD (HP-β-CD) and CTAB as
additives. For conventional CE, optimal separation conditions were achieved in a histidine/tartrate
buffer at pH 6.5, containing 0.025% HP-β-CD and 0.25 mM CTAB with a LOD ranging from 0.5 to
3 mg/L, whereas a histidine/tartrate buffer with 0.06% HP-β-CD and 0.25 mM CTAB gave a LOD
ranging from 3 to 10 mg/mL. By using a microchip electrophoresis format, a considerable reduction
of analysis time was accomplished.269
Another useful application of CE is the determination of the preservative nitrate and nitrite ions
in meat products.270 It is known that nitrite causes methaemoglobinaemia and, with secondary and
tertiary amines, yields the carcinogenic nitrosoamines. Owing to these toxic effects, it is important
to develop new analysis methods for the simultaneous determination of the two anions reducing
the matrix effect in meat samples. The method developed was based on the separation of the two
anions in a capillary coated with polyethyleneimine (PEI). Since PEI is a cationic polymer, the EOF
is reversed over a wide pH range and the fast separation of anions is achieved without the addition
of any electroosmotic modifier to the Tris separation buffer at pH 7.5. The LODs of the method for
nitrite and nitrate were 0.10 µg/mL and 0.09 µg/mL, respectively.270
Because of its aromatic properties, vanillin is the most widely used flavoring material in foods
and beverages.271 Vanillin is the main aromatic compound in natural vanilla, but it can also be syn-
thesized from low-cost materials such as 2-methoxyphenol, eugenol, and lignin. On the other hand,
ethylvanillin, which is commonly used as another synthetic compound, has much more flavoring
strength than vanillin and is used in formulation and imitation products. Thus, a CE method was
developed for the determination of vanillin, ethylvanillin, 2-methoxyphenol, and 2-ethoxyphenol,
simultaneously in cocoa drink samples by using a photodiode array detector.270 Separation was
carried out in phosphate buffer containing CTAH at pH 10 with 10% acetonitrile. Vanillin and
related compounds were determined in 7 min, with the LOD at 1.6 µg/mL. Mean recoveries were
96.3–103.8%. With this method, it was possible to determine vanillin and ethylvanillin, which were
originally contained as flavoring in cocoa drinks, but it was also found that vanillin and ethylvanillin
were metabolized, respectively, to 2-methoxyphenol and 2-ethoxyphenol, by growth of Bacillus
firmus causing an off-flavor.271
Artificial sweeteners are added to foods as a sugar substitute, particularly to low-calorie foods.
Sucralose is one of these artificial sweeteners that was determined by CE in different food matrices.272
A CE method was optimized chemometrically for the quantification of sucralose from different food
matrices. Separation from food matrix components was obtained in a dinitrobenzoic acid/sodium
hydroxide background electrolyte with a pH of 12.1 and detection was achieved at 238 nm by
indirect UV. The method allows the detection of sucralose at >30 mg/kg making it suitable for
implementation of the recently amended “Sweetners for use in foodstuffs” European Directive.272
Antioxidants are often added to food to retard oxidative processes that cause food off-flavors. The
most commonly permitted additives are butylated hydroxyanisole (BHA), butylated hydroxytoluene
(BHT), tert-butylhydroquinone (BHQ), and the esters of gallic acid. Some of the applications of CE
to synthetic antioxidants were reviewed by Boyce.264 A recently reported method was the analysis
of alkyl gallates and nordihydroguaguaiaretic acid (NDGA) by a microchip MEKC with pulsed
amperometric detection. Simultaneous separation of the antioxidants propyl gallate (PG), octyl
gallate (OG), lauryl gallate (LG), and NDGA was carried out in borate buffer, pH 9.7, containing
SDS. The measured detection limits were in the range of 2-6 fmol of the analyte.273
30.2.11 FOOD QUALITY

The versatility of CE makes it a very useful method for monitoring food quality from its different
aspects, namely, nutritional, sanitary, compositional (including authenticity), and sensorial. Also,
these technological processes such as heat treatment, roasting, freezing, drying, and so forth may
affect different aspects of food quality, starting with changes in compositional and nutritional quality.
While some recent applications of CE for evaluating food quality has been reviewed by Cifuentes,2
the CE analysis of dairy proteins to determine the effect of different processing conditions, to follow
proteolysis in different types of cheese, or to assess authenticity7 represent, by far, the largest number
of CE applications in food quality assessment. A comprehensive review of these applications has
been presented by Recio et al.224
Capillary electrophoresis has been very useful for studying the effect of heat treatment on dairy
proteins, particularly in the study of the interaction between carbonyl-containing compounds such
as reducing sugars with amino groups. These types of interactions, known as Maillard reaction,
have been found to be very important in food science and medicine. The Maillard reaction results
in the formation of large protein aggregates as well as low molecular weight compounds known
as Maillard reaction products (MRP) that impart various flavor, aroma, and color characteristics to
foods. Thus, the formation of MRP not only affects food compositional quality but also impacts
nutritional and sensorial aspects of quality. In this sense, CE methods were reported that allowed the
determination of a broad spectrum of compounds ranging from large molecular weight aggregates,
such as lactosylated milk proteins,274 to small molecular weight compounds, such as furosine.275
Deterioration due to Maillard reaction of milk powder upon storage was monitored by CE showing
a native and a modified fraction of β-lactoglobulin.274 The method used for the analysis of milk
proteins was the one previously reported by Recio and Olieman.232 In comparison to other methods,
the CE method proved to be rapid, easy to execute, and sensitive for monitoring early deterioration of
milk powders during storage due to Maillard reaction. The determination of the ratio of unmodified
whey proteins to total whey proteins by CE was considered a suitable method to evaluate the loss in
nutritional value of milk powders with reference to protein quality.274
A new CE method was established for the quantitative determination of furosine in dairy prod-
ucts. Sample preparation consisted of drying hydrolyzed samples, redissolving them in NaOH, and
purifying them by SPE. The electrophoretic separation was carried out in an uncoated capillary
using phosphate buffer containing the additive hexadecyl trimethylammonium bromide (HDTAB).
The LOD for furosine was 0.5 ppm, a concentration that corresponds to 4.5 mg/mL of protein in milk
samples. Electropherograms showing very well-defined peaks for furosine in UHT and evaporated
milks are depicted in Figure 30.9. As expected, higher furosine concentration was determined for
evaporated milk since it is the most severely heat-treated product. Thus, furosine determination by
CE method may be used to assess the extent of protein damage caused by heating and as a useful
indicator of the intensity of food processing conditions that cause the deterioration of the nutritive
value.275
Another heat-induced marker formed as a result of the Maillard reaction is hydroxymethylfurfural
(HMF). A MEKC method was developed for the determination of HMF in milk-based formulas by
using an uncoated capillary and phosphate buffer (pH 7.5) containing SDS.276 A similar compound,
furfural, is also formed as a result of Maillard reaction and could be detected in distilled agave
drinks such as Tequila and Bacanora by CE method. Furfural is formed during the cooking of agave
heads before juice extraction, fermentation, and distillation. A MEKC procedure was carried out
with phosphate buffer containing SDS at pH 7.5.277 Typical electropherograms showing the furfural
peak in Bacanora (Figure 30.10a) and Bacanora spiked with the analytical standard (Figure 30.10b)
are shown. In addition, some CE-MS and CE-MS/MS applications for the study of Maillard reaction
products were reviewed by Simo et al.249
Food authenticity is yet another food quality control parameter in addition to the parameters
described earlier. Food fraud may involve substitution or addition of ingredients of inferior value
Furosine
(a) (b) 0.0300
0.010
Absorbance (280 mn)
Absorbance (280 mn)

0.0200
Furosine
0.005
0.0100
0.000 0.0000
0.00 5.00 9.00 0.00 5.00 9.00

FIGURE 30.9 Typical electropherograms of milk showing the furosine peak. (a) UHT milk contanining
269 mg of furosine/100 g protein, (b) Evaporated milk containing 730 mg of furosine/100 g protein. (From
Vallejo-Cordoba, B., et al., J. Agric. Food Chem., 52, 5787, 2004. With permission.)
(a)
Absorbance (280 nm) (mAU)
Bacanora Furfural
4
0 1 2 3 4 5 6 7
(b)
Spiked sample
Absorbance (280 nm) (mAU)
7 Bacanora
6
5
4
3
2
1
0
0 1 2 3 4 5 6 7
FIGURE 30.10 Typical electropherograms of Bacanora (a), and Bacanora spiked with the analytical standard
(b). (From Vallejo-Cordoba, B. and Gonzalez-Cordova, A. F., in Hispanic Foods, Chemistry and Flavor, ACS
Series, 946, Tunick, M. H., and González de Mejia, E. Ed., 2007, p. 153. With permission.)
and/or quality required for normal or accepted processing or product formulation.6 A comprehen-
sive review of CE applications in food authenticity was offered by Kvasnicka.4 Authenticity issues
dealing with meat and meat products have been covered by Vallejo-Cordoba et al.,6 and applications
dealing with dairy products authenticity were reviewed by Recio et al.224 Detailed knowledge on
raw materials constituting foods is essential, and this is especially true for the diverse foods and
feeds made from cereal grains. Varietal identity and ability to predict dough properties of wheat
consignments are largely dependent on protein composition. Whereas DNA analysis indicates geno-
type, protein composition provides information on variety and likely processing parameters. In this
sense, the Australian industry has standard methods of protein and DNA analysis involving gel elec-
trophoresis, CE, and most recently, Lab-on-a-chip equipment for protein analysis and microarrays
for DNA composition.278 The treatment of foods can be another issue in food authenticity; for exam-
ple, meat may have been previously frozen and falsely represented as fresh, especially with fresh
or chilled meat associated with higher prices than their frozen counterparts. Thus, fraudulent mar-
keters may thaw frozen meat and sell it as chilled or fresh. Methods for the differentiation between
fresh and frozen/thawed meat were based on the release of enzymes. Thus, the application of elec-
trophoretically mediated microanalysis (EMMA) to determine β-hydroxyacyl CoA-dehydrogenase
(β-HADH) activity in a model system and in meat samples was reported. The enzymatic assay and
the separation of the reaction products were carried out in an uncoated capillary using a plug–plug
reaction mode at variable potential. The NAD+ produced by β-HADH activity in juice extracted
from previously frozen meat was at least three times the concentration of that produced in fresh meat
(Figure 30.11).279
The most important issue in food quality analysis is food safety. Therefore, the development of
rapid methods for the identification and quantification of bacterial contamination in foods is of utmost
importance. Thus, a CE method with UV detection was proposed for the identification and quantifi-
cation of bacterial contamination in food samples. The proposed method allowed for the effective
separation of eight different types of bacteria in only 25 min. Electrophoretic resolution was improved
by using cations in phosphate buffer (pH 7.0) that interacted with the bacterial surfaces changing
its electrical properties and electrophoretic mobility. The validity of the method was established
by comparison with the standard plate counting method,151 where bacterial cells were separated as
mAU
100
80 NAD+
Absorbance (260 nm)
(a)
60
40
20
(b)
2 4 6 8 10 12 14
Time (min)
FIGURE 30.11 Electropherogram showing β-HADH activity by EMMA in extracted juice from beef.
(a) Frozen thawed and (b) fresh. (From Vallejo-Cordoba, B., et al., J. Cap. Elec. Microchip Tech., 8, 81,
“biocolloids,” which under the influence of an electric field, move at a velocity that is proportional
to their electrophoretic mobility. This means that under appropriate experimental conditions, dif-
ferent microorganisms exhibit a differential electrophoretic mobility.151 Another approach for the
identification of foodborne pathogens or food-spoilage bacteria consisted in multiplex-PCR-CGE-
LIF procedures.258,259 The main advantage of these techniques over conventional microbiological
techniques is that it is possible to obtain fast and specific identification of microorganisms at the
molecular/genetic level. One of the PCR-CGE-LIF protocols allowed for the simultaneous detec-
tion and differentiation of the genera Leuconostoc and Carnobacterium,258 while a second protocol
allowed the detection of Staphylococcus aureus, Listeria monocytogenes, and Salmonella spp.259
The determination of food fatty acids is important from the standpoint of food quality com-
position since fatty acid composition is required in nutritional labeling. Also, the quantification of
trans-fatty acids, which are formed during oil hydrogenation, is particularly important since new
regulations require their labeling to allow the consumer to make informed choices on their diet.
Thus, a novel CE methodology using UV indirect detection at 224 nm for the analysis of trans-fatty
acids in hydrogenated oils was reported. The electrolyte consisted of phosphate buffer (pH 7.0)
containing sodium dodecylbenzenesulfonate, Brij 35, octanol, and acetonitrile, where, under opti-
mized conditions, 10 fatty acids including C:12, C:13 (internal standard), C:14, C:16, C:18, C18:1c,
C18:1t, C18:2cc, C18:2tt, and C18:3ccc were baseline separated in less than 12 min. The proposed
method was used to monitor the formation of trans-fatty acids during hydrogenation of Brazilnut
oil.280 Similarly, the determination of short-chain free fatty acids, which are associated with lipolized
flavors, is important for good sensory quality. However, the quantification of individual free fatty
acids (FFA) in dairy products is particularly complicated, since FFA represent less than 0.5% of
the total fat. FFA with chain lengths from 2 to 20 carbon atoms are present, and short-chain FFA
with fewer than eight carbon atoms are extremely volatile. Because of these problems, the methods
most frequently described for the analysis of FFA in milk and dairy products involve fat extraction,
separation of FFA from the bulk of the extracted lipid, and injection of the fraction containing FFA
directly in the gas chromatograph, or fat extraction, derivatization, and gas chromatography of the
methyl esters formed.282 To circumvent these problems, an attractive alternative was the use of CE,
particularly in a MEKC mode. To this end, the release of short-chain free fatty acids was monitored
from milk fat during hydrolysis with lipase by using MEKC.282 Sample buffer containing CD and
methanol allowed for the solubilization of FFA, followed by separation in an uncoated capillary with
Tris buffer containing p-anisate as chromophore and trimethyl β-CD at pH 8.0. Indirect UV detection
at 270 nm was used282 and excellent separation was observed for FFA C4 −C12 in a FFA standard
mixture (Figure 30.12a) or in milk fat lipolyzed with lipase (Figure 30.12b).
30.3 CONCLUDING REMARKS AND OUTLOOK

The versatility of CE in food analysis was clearly shown by the numerous, wide ranging methods
and applications presented. Moreover, the efficiency and rapidity of CE separations were the main
advantages that make this technique an attractive alternative in food analysis laboratories. CE is still
far from being as well established as chromatographic techniques, although it certainly has gained
popularity as an alternative to conventional gel electrophoresis for the analysis of food proteins, most
likely due to the quantitative nature of CE, high resolving power, and speed of analysis. However,
for the analysis of trace component determination for an organic contaminant, low sensitivity and
low reproducibility of quantitative analysis are issues that need to be addressed. Perhaps the greatest
potential of CE was found in food diagnostics for quality assurance and authenticity, with most
recent approaches involving Lab-on-chip equipment. Authenticity issues such as meat and fish
species, GMOs, and cereal variety identification and quantification are challenging problems where
DNA-based CE methods in microchip-based formats offer opportunities for development. Similarly,
DNA-based CE methods for microorganism identification is another niche that may be explored.
(a) (b)
15 C6 C4
C5
12.5
25 C6
10 C7
Absorbance (mAU)
Absorbance (mAU)
C4
7.5 C6 20
5 C12 C8
C11 C10 C8
2.5 15
0
10 C12 C10
−2.5
−5
5
5 6 7 8 9 10 5 6 7 8 9 10
FIGURE 30.12 Typical electropherograms of free fatty acids. (a) Standard mixture (100 ppm of each), (b)
milk fat after a 30 min hydrolysis with lipase. (From Vallejo-Cordoba, B., et al., J. Cap. Elec., 5, 11, 1998. With
permission.)
ACKNOWLEDGMENTS
M.G. Vargas would like to thank the National Autonomous University of Mexico (Universidad
Nacional Autónoma de México, UNAM, FES-Cuautitlán, Depto de Química) for time and resources
to complete this work. Author B. Vallejo-Cordoba would like to acknowledge the Research Center for
Food and Development (Centro de Investigación en Alimentación y Desarrollo, A.C., CIAD, A.C.)
for their financial assistance, Dr. James Landers for his kind invitation to make this contribution, and
Dr. Norberto Guzman for sharing his enthusiasm on CE technology.
REFERENCES
1. Castañeda, G., Rodríguez-Flores, J., and Ríos, A., Analytical approaches to expanding the use of
capillary electrophoresis in routine food analysis, J. Sep. Sci., 28, 915, 2005.
2. Cifuentes, A., Recent advances in the application of capillary electromigration methods for food
analysis, Electrophoresis, 27, 283, 2006.
3. Juan-García, A., Font, G., and Picó, Y., Determination of organic contaminants in food by capillary
electrophoresis, J. Sep. Sci., 28, 793, 2005.
4. Kvasnicka, F., Capillary electrophoresis in food authenticity, J. Sep. Sci., 28, 813, 2005.
5. Frazier, R. A., Ames, J. M., and Nursten, H. E., Capillary electrophoresis for food analysis: Method
development, The Royal Society of Chemistry, Cambridge, U.K., 2000, p. 67.
6. Vallejo-Córdoba, B., González-Córdoba, A.F., Mazorra-Manzano, M.A., and Rodríguez-Ramírez, R.,
Capillary electrophoresis for the analysis of meat authenticity, J. Sep. Sci., 28, 826, 2005.
7. Frazier, R. A., Ames, J. M., and Nursten, H. E., The development and application of capillary
electrophoresis methods for food analysis, Electrophoresis, 20, 3156, 1999.
8. Frazier, R. A., Recent advances in capillary electrophoresis methods for food analysis, Electrophoresis,
22, 4197, 2001.
9. Frazier, R. A., and Papadopoulou, A., Recent advances in the application of capillary electrophoresis
for food analysis, Electrophoresis, 24, 4095, 2003.
10. Sørensen, H., Sørensen, S., Bjergegaard, C., and Michaelsen, S., Chromatographic and Capillary
Electrophoresis in Food Analysis, The Royal Society of Chemistry, Cambridge, U.K., 1999.
11. Wang, X., Wang, Q., Chen, Y., and Han, H., Determination of carbohydrates as their
3-aminophthalhydrazide derivatives by capillary zone electrophoresis with on-line chemiluminescence
detection, J. Chromatogr. A, 992, 181, 2003.
12. He, L., Sato, K., Abo, M., Okubo, A., and Yamazaki, S., Separation of saccharides derivatized with
2-aminobenzoic acid by capillary electrophoresis and their structural consideration by nuclear magnetic
resonance, Anal. Biochem., 314, 128, 2003.
13. Zhang, M., Melouk, H. A., Chenault, K., and El Rassi, Z., Determination of cellular carbohydrates in
peanut fungal pathogens and Baker’s yeast by capillary electrophoresis and electrochromatography,
J. Agric. Food Chem., 49, 5265, 2001.
14. Doner, L. W., Johnston, D. B., and Singh, V., Analysis and properties of arabinoxylans from discrete
corn wet-milling fiber fractions, J. Agric. Food Chem., 49, 1266, 2001.
15. Cabálková, J., Žídková, J., Pribyla, L., and Chmelík, J., Determination of carbohydrates in juices
by capillary electrophoresis, high-performance liquid chromatography, and matrix-assisted laser
desorption/ionization-time of flight-mass spectrometry, Electrophoresis, 25, 487, 2004.
16. Carvalho, A. Z., da Silva, J. A. F., and do Lago, C. L., Determination of mono- and disaccha-
rides by capillary electrophoresis with contactless conductivity detection, Electrophoresis, 24, 2138,
2003.
17. Klampfl, C. W., and Buchberger, W., Determination of carbohydrates by capillary electrophoresis with
electrospray-mass spectrometric detection, Electrophoresis, 22, 2737, 2001.
18. Soga, T., and Serwe, M., Determination of carbohydrates in food samples by capillary electrophoresis
with indirect UV detection, Food Chem., 69, 339, 2000.
19. Khandurina, J., Anderson, A. A., Olson, N. A., Stegel, J. T., and Guttman, A., Large-scale carbo-
hydrate analysis by capillary array electrophoresis: Part 2. Data normalization and quantification,
20. Hebert, N. E., Kuhr, W. G., and Brazill, S. A., Microchip capillary electrophoresis coupled to sinusoidal
voltammetry for the detection of native carbohydrates, Electrophoresis, 23, 3750, 2002.
21. Andersen, K. E., Bjergegaard, C., and Sørensen, H., Analysis of reducing carbohydrates by reductive
tryptamine derivatization prior to micellar electrokinetic capillary chromatography, J. Agric. Food
Chem., 51, 7234, 2003.
22. Momenbeik, F., Johns, C., Breadmore, M. C., Hilder, E. F., Macka, M., and Haddad, P. R.,
Sensitive determination of carbohydrates labelled with p-nitroaniline by capillary electrophoresis with
photometric detection using a 406 nm light-emitting diode, Electrophoresis, 27, 4039, 2006.
23. Mangin, C. M., Goodall1, D. M., and Roberts, M. A., Separation of ι-, χ - and λ-carrageenans by
24. Mistry, K., Krull1, I., and Grinberg, N., Size-exclusion capillary electrochromatographic separation
of polysaccharides using polymeric stationary phases, Electrophoresis, 24, 1753, 2003.
25. Cortacero-Ramírez, S., Segura-Carretero, A., Cruces-Blanco, C., Hernáinz-Bermúdez de Castro, M.,
and Fernández-Gutiérrez, A., Analysis of carbohydrates in beverages by capillary electrophoresis with
precolumn derivatization and UV detection, Food Chem., 87, 471, 2004.
26. Campa, C., Coslovi1, A., Flamigni1, A., and Rossi1, M., Overview on advances in capillary
electrophoresis-mass spectrometry of carbohydrates: A tabulated review, Electrophoresis, 27, 2027,
2006.
27. Martínez-Montero, C., Rodríguez-Dodero, M. C., Guillen-Sánchez, D. A., and Barroso, C. G., Analysis
of low molecular weight carbohydrates in food and beverages: A review, Chromatographia, 59, 15,
2004.
28. Klockow, A., Paulus, A., Figueiredo, V., Amado, R., and Widmer, H. M., Determination of car-
bohydrates in fruit juices by capillary electrophoresis and high-performance liquid chromatography,
J. Chromatogr. A, 680, 187, 1994.
29. Vorndran, A. E., Oefner, P. J., Scherz, H., and Bonn, G. K., Indirect UV detection of carbohydrates in
capillary zone electrophoresis, Chromatographia, 33, 163, 1992.
30. Soga, T., and Heiger, D. N., Simultaneous determination of monosaccharides in glycoproteins by
capillary electrophoresis, Anal. Biochem., 261, 73, 1998.
31. Suzuki, S., and Honda S., Miniaturization in carbohydrate analysis, Electrophoresis, 24, 3577,
2003.
32. Legaz, M. E., Pedrosa, M.M., and Vicente, C., Carbohydrates, polymeric sugars and their constituents,
in Advanced Chromatographic and Electromigration Methods in Bioscience, Deyl, S., Milsik, I.,
Tagliaro, F., and Tesarova, E., Ed. J. of Chromatogr. 60, 264, 1998.
33. Chiu, T. C., Lin, Y. W., Huang, Y. F., and Chang, H. T., Analysis of biologically active amines by CE,
34. Pais, P., and Knize, M.G., Chromatographic and related techniques for the determination of aromatic
heterocyclic amines in foods, J. Chromatogr. B, 747, 139, 2000.
35. Önal, A., A review: Current analytical methods for the determination of biogenic amines in foods,
Food Chem., 108, 1475, 2007.
36. Oguri, S., Electromigration methods for amino acids, biogenic amines and aromatic amines,
J. Chromatogr. B, 747, 1, 2000.
37. Kovács, Á., Simon-Sarkadi, L., and Ganzler, K., Determination of biogenic amines by capillary
38. Liu, X., Yang, L. X., and Lu, Y. T., Determination of biogenic amines by 3-(2-furoyl)
quinoline-2-carboxaldehyde and capillary electrophoresis with laser-induced fluorescence detection,
J. Chromatogr.A, 998, 213, 2003.
39. Rodríguez, I., Lee, H. K., and Li, S. F. Y., Microchannel electrophoretic separation of biogenic amines
by micellar electrokinetic chromatography, Electrophoresis, 20, 118, 1999.
40. Molnar-Perl, I., Role of chromatography in the analysis of sugars, carboxylic acids and amino acids
in food, J. Chromatogr. A, 891, 1, 2000.
41. Poinsot, V., Lacroix, M., Maury, D., Chataigne, G., Feurer, B., and Couderc, F., Recent advances in
amino acid analysis by capillary electrophoresis, Electrophoresis, 27, 176, 2006.
42. Simó, C., Barbas, C., and Cifuentes, A., Capillary electrophoresis-mass spectrometry in food analysis,
43. Ummadi, M., and Weimer, B. C., Use of capillary electrophoresis and laser-induced fluorescence for
attomole detection of amino acids, J. Chromatogr. A, 964, 243, 2002.
44. Kostal, V., Zeisbergerova, M., Hrotekova, Z., Slais, K., and Kahle V., Miniaturized liquid core
waveguide-based fluorimetric detection cell for capillary separation methods: Application in CE of
amino acids, Electrophoresis, 27, 4658, 2006.
45. Dong, Q., Jin, W., and Shan, J., Analysis of amino acids by capillary zone electrophoresis with
electrochemical detection, Electrophoresis, 23, 559, 2002.
46. Coufal, P., Zuska, J., van de Goor, T., Smith, V., and Gaš, B., Separation of twenty underiva-
tized essential amino acids by capillary zone electrophoresis with contactless conductivity detection,
47. Schultz, C. L., and Moini, M., Analysis of underivatized amino acids and their d/l-enantiomers by
sheathless capillary electrophoresis/electrospray ionization-mass spectrometry, Anal. Chem., 75, 1508,
2003.
48. Soga, T., and Heiger, D.N., Amino acid analysis by capillary electrophoresis electrospray ionization
mass spectrometry, Anal. Chem., 72, 1236, 2000.
49. Siri, N., Lacroix, M., Garrigues, J. C., Poinsot, V., and Couderc, F., HPLC-fluorescence detection and
MEKC-LIF detection for the study of amino acids and catecholamines labelled with naphthalene-2,3-
dicarboxyaldehyde, Electrophoresis, 27, 4446, 2006.50.
50. Lioe, H. N., Apriyantono, A., Takara, K., Wada, K., Naoki, H., and Yasuda, M., Low molecular weight
compounds responsible for savory taste of indonesian soy sauce, J. Agric. Food Chem., 52, 5950,
2004.
51. García-Villar, N., Saurina, J., and Hernández-Cassou, S., Capillary electrophoresis determination of
biogenic amines by field-amplified sample stacking and in-capillary derivatization, Electrophoresis,
27, 474, 2006.
52. Santos, B., Simonet, B. M., Ríos, A., and Valcárcel, M., Direct automatic determination of biogenic
amines in wine by flow injection-capillary electrophoresis-mass spectrometry, Electrophoresis, 25,
342, 2004.
53. Male, K. B., and Luong, J. H. T., Derivatization, stabilization and detection of biogenic amines by
cyclodextrin-modified capillary electrophoresis–laser-induced fluorescence detection, J. Chromatogr.
A, 926, 309, 2001.
54. Qu, Q., Liu, Y., Tang, X., Wang, C., Yang, G., Hu1, X., and Yan, C., Etched bare fused-silica capillaries
for online preconcentration of amino acids in CE, Electrophoresis, 27, 4500, 2006.
55. Lee, J. H., Choi, O. K., Jung, H. S., Kim, K. R., and Chung, D. S., Capillary electrophoresis of
nonprotein and protein amino acids without derivatization, Electrophoresis, 21, 930, 2000.
56. Zhan, W., Wang, T., and Li, S. F. Y., Derivatization, extraction and concentration of amino acid and
peptides by using organic/aqueous phases in capillary electrophoresis with fluorescence detection,
57. Klampfl C. W., and Ahrer, W., Determination of free amino acids in infant food by capillary zone
electrophoresis with mass spectrometric detection, Electrophoresis, 22, 1579, 2001.
58. Cao, L., Wang, H., and Zhang, H., Analytical potential of 6-oxy-(N-succinimidyl acetate)-9-(2 -
methoxycarbonyl) fluorescein for the determination of amino compounds by capillary electrophoresis
with laser-induced fluorescence detection, Electrophoresis, 26, 1954, 2005.
59. Beard, N. P., and de Mello, A. J., A polydimethylsiloxane/glass capillary electrophoresis microchip
for the analysis of biogenic amines using indirect fluorescence detection, Electrophoresis, 23, 1722,
2002.
60. Tanyanyiwa, J., Schweizer, K., and Hauser, P. C., High-voltage contactless conductivity detection of
underivatized amino acids in capillary electrophoresis, Electrophoresis, 24, 2119, 2003.
61. Lange, J., Thomas, K., and Wittmann, C., Comparison of a capillary electrophoresis method with high
performance liquid chromatography for the determination of biogenic amines in various food samples,
J. Chromatogr. B, 779, 229, 2002.
62. Hart, J. J., Welch, R. M., Norvell, W. A., and Kochian, L.V., Measurement of thiol-containing amino
acids and phytochelatin (PC2) via capillary electrophoresis with laser-induced fluorescence detection,
63. Shin, D., Tryk, D. A., Fujishima, A., Muck Jr., A., Chen, G., and Wang, J., Microchip capillary
electrophoresis with a boron-doped diamond electrochemical detector for analysis of aromatic amines,
64. Zhang, Y., and Gomez, F. A., On-column derivatization and analysis of amino acids, peptides and
alkylamines by anhydrides using capillary electrophoresis, Electrophoresis, 21, 3305, 2000.
65. Ruiz-Jiménez, J., and Luque de Castro, M. D., Pervaporation as interface between solid samples and
capillary electrophoresis, Determination of biogenic amines in food, J. Chromatogr. A, 1110, 245,
2006.
66. Driouich, R., Takayanagi, T., Oshima, M., and Motomizu, S., Synthesis of water-soluble phenylazos-
alicylaldehyde analogues and their application to capillary electrophoretic determination of primary
amines, Electrophoresis, 27, 3460, 2006.
67. La, S., Kim, A., Kim, J. H., Choi, O. K., and Kim, K. R., Profiling and screening analysis of 27 aromatic
amino acids by capillary electrophoresis in dual modes, Electrophoresis, 23, 1080, 2002.
68. Lin, C. C., and Liu, C. Y., Proline-coated column for the capillary electrochromatographic separation
of amino acids by in-column derivatization, Electrophoresis, 25, 3216, 2004.
69. Beard, N. P., Edel, J. B., and de Mello, A. J., Integrated on-chip derivatization and electrophoresis for
the rapid analysis of biogenic amines, Electrophoresis, 25, 2363, 2004.
70. Lin, Z., Xie, Z., Lu, H., Lin, X., Wu, X., and Chen, G., On-column coaxial flow chemilumines-
cence detection for underivatized amino acids by pressurized capillary electrochromatography using
a monolithic column, Anal. Chem., 78, 5322, 2006.
71. Ro, K. W., Lim, K., Kim, H., and Hahn, J. H., Poly(dimethylsiloxane) microchip for precolumn
reaction and micellar electrokinetic chromatography of biogenic amines, Electrophoresis, 23, 1129,
2002.
72. Soga, T., and Imaizumi, M., Capillary electrophoresis method for the analysis of inorganic anions,
organic acids, amino acids, nucleotides, carbohydrates and other anionic compounds, Electrophoresis,
22, 3418, 2001.
73. Cortacero-Ramírez, S., Segura-Carretero, A., Cruces-Blanco, C., Hernáinz-Bermúdez de Castro, M.,
and Fernández-Gutiérrezet, A., Direct multicomponent analysis of beer samples constituents using
micellar electrokinetic capillary chromatography, Electrophoresis, 25, 1867, 2004.
74. Kvasnicka, F., and Voldrich, M., Determination of biogenic amines by capillary zone electrophoresis
with conductometric detection, J. Chromatogr. A, 1103, 145, 2006.
75. Vargas M. G., Muñoz Ramírez, E., and Trejo Márquez, A., Development of analytical methods for the
determination of polyamines by MECK using LIF detection, Presented at the 12th Latin-American
Symposium on Biotechnology, Biomedical, Biopharmaceutical and Industrial Applications of
Capillary Electrophoresis and Microchip Technology, Queretaro, Mexico, 2–5 December, 2006.
76. Cinquina, A. L., Longo, F., Cali, A., De Santis, L., Baccelliere, R., and Cozzani, R., Validation
and comparison of analytical methods for the determination of histamine in tuna fish samples,
J. Chromatogr. A, 1032, 79, 2004.
77. Ong, C. P., Ng, C. L., Lee, H. K., and Li, S. F. Y., Separation of water- and fat-soluble vitamins by
micellar electrokinetic chromatography, J. Chromatogr. A, 547, 419, 1991.
78. Craige Trenerry, V., The application of capillary electrophoresis to the analysis of vitamins in food
and beverages, Electrophoresis, 22, 1468, 2001.
79. Chiari, M., Nesi, M., Carrea, G., and Righetti, P. G., Determination of total vitamin C in fruits by
capillary zone electrophoresis, J. Chromatogr. A, 645, 197, 1993.
80. Thompson, C. O., and Trenerry, V. C., A rapid method for the determination of total l-ascorbic acid
in fruits and vegetables by micellar electrokinetic capillary chromatography, Food Chem., 53, 43,
1995.
81. Panak, K., Giorgieri, S. A., Ruiz, O., and Diaz, L. E., Postharvest analysis of vitamin C and inorganic
cations in lettuce by CZE, J. Capil. Electrophor., 5, 59, 1998.
82. Boyce, M. C., and Spickett, E. E., Separation of food grade antioxidants (synthetic and natural) using
mixed micellar electrokinetic capillary chromatography, J. Agric. Food. Chem., 47, 1970, 1999.
83. Buskov, S., Moller, P., Sorensen, H., Sorensen, J. C., and Sorensen, S., Determination of vitamins in
food based on supercritical fluid extraction prior to micellar electrokinetic capillary chromatographic
analyses of individual vitamins, J. Chromatogr. A, 802, 233, 1998.
84. Begona, B. M., and van de Werken, G., Determination of green and black tea composition by capillary
electrophoresis, J. High Resolut. Chromatogr., 22, 225, 1999.
85. Schiewe, J., Mrestani, Y., and Neubert, R., Application and optimization of capillary zone
electrophoresis in vitamin analysis, J. Chromatogr. A, 717, 255, 1995.
86. Koh, E. V., Bissell, M. G., and Ito, R. K., Measurement of vitamin C by capillary electrophoresis in
biological fluids and fruit beverages using a stereoisomer as an internal standard, J. Chromatogr. A,
633, 245, 1993.
87. Lin Ling, B., Baeyens, W. R. G., Van Ackers, P., and Dewaele, C., Determination of ascorbic acid and
isoascorbic acid by capillary zone electrophoresis: Application to fruit juices and to a pharmaceutical
formulation, J. Pharm. Biomed. Anal., 10, 717, 1992.
88. Choi, O., and Jo, J., Determination of l-ascorbic acid in foods by capillary zone electrophoresis,
J. Chromatogr. A, 781, 435, 1997.
89. Marshall, P. A., Trenerry, V. C., and Thompson, C. O., The determination of total ascorbic acid in
beers, wines, and fruit drinks by micellar electrokinetic capillary chromatography, J. Chromatogr. Sci.,
33, 426, 1995.
90. Ward, C. M., Trenerry, V. C., and Pant, I., The application of capillary electrophoresis to the
determination of total niacin in concentrated yeast, Food Chem., 58, 185, 1997.
91. Ward, C. M., and Trenerry, V. C., The determination of niacin in cereals, meat and selected foods
by capillary electrophoresis and high performance liquid chromatography, Food Chem., 60, 667,
1997.
92. Diaz-Pollan, C., and Vidal-Valverde, C., Niacin determination in legumes by capillary electrophore-
sis (CE). Comparison with high performance liquid chromatography (HPLC), J. High Resolut.
Chromatogr., 21, 81, 1998.
93. Windahl, K. W., Trenerry, V. C., and Ward, C. M., The determination of niacin in selected foods by
capillary electrophoresis and high performance liquid chromatography: Acid extraction, Food Chem.,
65, 263, 1998.
94. Vidal-Valverde, C., and Diaz-Pollan, C., Optimization analysis by capillary electrophoresis of thiamine
in meat: Comparison with high perfomance liquid chromatography, Eur. Food Res. Technol., 209, 355,
1999.
95. Vidal-Valverde, C., and Diaz-Pollan, C., Comparison of capillary electrophoretic and high performance
liquid chromatographic thiamin determination in milk, Milchwissenschaft, 55, 307, 2000.
96. Sanchez, J. M., and Salvado, V., Comparison of micellar and microemulsion electrokinetic chro-
matography for the analysis of water- and fat-soluble vitamins, J. Chromatogr. A, 950, 241,
2002.
97. Herrero-Martinez, J. M., Simo-Alfonso, E. F., Ramis-Ramos, G., Deltoro, V. I., Calatayud, A., and
Barreno, E., Simultaneous determination of l-ascorbic acid, glutathione, and their oxidized forms
in ozone-exposed vascular plants by capillary zone electrophoresis, Environ. Sci. Technol., 34, 1331,
2000.
98. Liao, T., Jiang, C. M., Wu, M. C., Hwang, J. Y., and Chang, H. M., Quantification of l-ascorbic acid
and total ascorbic acid in fruits and spinach by capillary zone electrophoresis, Electrophoresis, 22,
1484, 2001.
99. Cataldi, T. R. I., Nardiello, D., Carrara, V., Cirielloa, R., and De Benedetto G. E., Assessment of
riboflavin and flavin content in common food samples by capillary electrophoresis with laser-induced
fluorescence detection, Food Chem., 82, 309, 2003.
100. Cataldi, T. R. I., Nardiello, D., Scrano, L., and Scopa, A., Assay of riboflavin in sample wines by
capillary zone electrophoresis and laser-induced fluorescence detection, J. Agric. Food Chem., 50,
6643, 2002.
101. Aturki, Z., D’Orazio, G., and Fanali, S., Rapid assay of vitamin E in vegetable oils by reversed-phase
102. Law, W. S., Kubán̆, P., Hong Zhao, J., Li, S. F. Y., and Hauser, P. C., Determination of vitamin C and
preservatives in beverages by conventional capillary electrophoresis and microchip electrophoresis
with capacitively coupled contactless conductivity detection, Electrophoresis, 26, 4648, 2005.
103. Cataldi, T. R. I., Nardiello, D., De Benedetto, G. E., and Bufo, S. A., Optimizing separation conditions
for riboflavin, flavin mononucleotide and flavin adenine dinucleotide in capillary zone electrophoresis
with laser-induced fluorescence detection, J. Chromatogr. A, 968, 229, 2002.
104. Frias, J., Miranda, M. L., Doblado, R., and Vidal-Valverde, C., Effect of germination and fermentation
on the antioxidant vitamin content and antioxidant capacity of Lupinus albus L. var. Multolupa,
Food Chem., 92, 211, 2005.
105. Aurant, L., Woods, A., and Wells, M., Food laws and regulation. In Food Composition and Analysis,
Van Nostrand Reinhold, New York, 1987, p. 14.
106. Galli, V., Garcia, A., Saavedra, L., and Barbas, C., Capillary electrophoresis for short-chain organic
acids and inorganic anions in different samples, Electrophoresis, 24, 1951, 2003.
107. Cortacero-Ramirez, S., Hernainz-Bermudez de Castro, M., Segura-Carretero, A., and Cruces-
Blanco, C., Analysis of beer components by capillary electrophoretic methods, Trends Anal. Chem.,
22, 440, 2003.
108. Vickers, P. J., Braybrook, J., Lawrence, P., and Gray, K., Detecting tartrate additives in foods:
Evaluating the use of capillary electrophoresis, J. Food Composition Anal., 20, 252, 2007.
109. Cortacero-Ramírez, S., Segura-Carretero, A., Hernáinz-Bermúdez de Castro, M., and Fernández-
Gutiérrez, A., Determination of low-molecular-mass organic acids in any type of beer samples by
coelectroosmotic capillary electrophoresis, J. Chromatogr. A, 1064, 115, 2005.
110. Izco, J. M., Tormo, M., and Jimenez-Flores, R., Development of a CE method to analyze organic acids
in dairy products: Application to study the metabolism of heat-shocked spores, J. Agric. Food Chem.,
50, 1765, 2002.
111. Suárez-Luque, S., Mato, I., Huidobro, J. F., Simal-Lozano, J., and Sancho, M. T., Capillary zone
electrophoresis method for the determination of inorganic anions and formic acid in honey, J. Agric.
Food Chem., 54, 9292, 2006.
112. Saavedra, L., Rupérez, F. J., and Barbas, C., Capillary electrophoresis for evaluating orange juice
authenticity: A study on Spanish oranges, J. Agric. Food Chem., 49, 9, 2001.
113. Galli, V., and Barbas, C., Capillary electrophoresis for the analysis of short-chain organic acids in
coffee, J. Chromatogr. A, 1032, 299, 2004.
114. Esteves, V. I., Lima, S. S. F., Lima, D. L. D., and Duarte, A. C., Using capillary electrophoresis for
the determination of organic acids in Port wine, Anal. Chim. Acta, 513, 163, 2004.
115. Saavedra, L., and Barbas, C., Validated capillary electrophoresis method for small-anions measurement
in wines, Electrophoresis, 24, 2235, 2003.
116. Castiñeira, A., Peña, R. M., Herrero, C., and Garci⊥’ıa-Martín, S., Analysis of organic acids in wine
by capillary electrophoresis with direct UV detection, J. Food Comp. Anal., 15, 319, 2002.
117. Nhujak, T., Srisa-Art, M., Kalampakorn, K., Tolieng, V., and Petsom, A., Determination of gibberel-
lic acid in fermentation broth and commercial products by micellar electrokinetic chromatography,
J. Agric. Food Chem., 53, 1884, 2005.
118. Kubán̆, P., and Hauser, P. C., Application of an external contactless conductivity detector for the
analysis of beverages by microchip capillary electrophoresis, Electrophoresis, 26, 3169, 2005.
119. Fung, Y. S., and Lau, K. M., Analysis of organic acids and inorganic anions in beverage drinks by
120. Tanyanyiwa, J., and Hauser, P.C., High-voltage capacitively coupled contactless conductivity detection
for microchip capillary electrophoresis, Anal. Chem., 74, 6378, 2002.
121. Yamamoto, A., Akiba, N., Kodama, S., Matsunaga, A., Kato, K., and Nakazawa, H., Enantiomeric
purity determination of malic acid in apple juices by multibeam circular dichroism detection,
J. Chromatogr. A, 928, 139, 2001.
122. Menzinger, F., Schmitt-Kopplin, Ph., Freitag, D., and Kettrup, A., Analysis of agrochemicals by
123. Weinberger, R., Capillary electrophoresis of venoms and toxins, Electrophoresis, 22, 3639,
2001.
124. Hernández-Borges, J., Rodríguez-Delgado, M. Á., García-Montelongo, F. J., and Cifuentes A., Chiral
analysis of pollutants and their metabolites by capillary electromigration methods, Electrophoresis,
26, 3799, 2005.
125. Ahmed, F. E., Analyses of pesticides and their metabolites in foods and drinks, Trends Anal. Chem.,
20, 649, 2001.
126. Eash, D. T., and Bushway, R. J., Herbicide and plant growth regulator analysis by capillary
127. Rodríguez, R., Mañes, J., and Picó, Y., Off-line solid-phase microextraction and capillary elec-
trophoresis mass spectrometry to determine acidic pesticides in fruits, Anal. Chem., 75, 452,
2003.
128. Orejuela, E., and Silva, M., Rapid and sensitive determination of phosphorus-containing amino acid
herbicides in soil samples by capillary zone electrophoresis with diode laser-induced fluorescence
129. Orejuela, E., and Silva, M., Rapid determination of aniline metabolites of chlorpropham in potatoes
by micellar electrokinetic chromatography using negative-charged mixed micelles and laser-induced
fluorescence detection, Electrophoresis, 26, 2991, 2005.
130. Wei, J., Okerberg, E., Dunlap, J., Ly, C., and Shear, J. B., Determination of biological toxins using
capillary electrokinetic chromatography with multiphoton-excited fluorescence, Anal. Chem., 72,
1360, 2000.
131. Cooper, P. A., Jessop, K. M., and Moffatt, F., Capillary electrochromatography for pesticide analysis:
Effects of environmental matrices, Electrophoresis, 21, 1574, 2000.
132. Rodríguez, R., Picó, Y., Font, G., and Mañes, J., Determination of urea-derived pesticides in fruits and
vegetables by solid-phase preconcentration and capillary electrophoresis, Electrophoresis, 22, 2010,
2001.
133. Fung, Y. S., and Mak, J. L. L., Determination of pesticides in drinking water by micellar electrokinetic
capillary chromatography, Electrophoresis, 22, 2260, 2001.
134. Kilár, A., Kocsis, B., Kustos, I., Kilár, F., and Hjertén, S., CE to monitor endotoxins by protein
complexation, Electrophoresis, 27, 4188, 2006.
135. He, J., Chen, S. W., Ruan, L. F., Cao, L. L., Yao, J., and Yu, Z. N., Determination of the fungicide
validamycin A by capillary zone electrophoresis with indirect UV detection, J. Agric. Food Chem., 51,
7523, 2003.
136. Safarpour, H., and Asiaie, R., Determination of glyphosate as cross-contaminant in a commercial
herbicide by capillary electrophoresis-electrospray ionization-mass spectrometry, Electrophoresis, 26,
1562, 2005.
137. Hernández-Borges, J., Rodríguez-Delgado, M. Á., García-Montelongo, F. J., and Cifuentes
A., Highly sensitive analysis of multiple pesticides in foods combining solid-phase microextrac-
tion, capillary electrophoresis-mass spectrometry, and chemometrics, Electrophoresis, 25, 2065,
2004.
138. Da Silva C. L., de Lima, E. C., and Tavares, M. F. M., Investigation of preconcentration strate-
gies for the trace analysis of multiresidue pesticides in real samples by capillary electrophoresis,
J. Chromatogr. A, 1014, 109, 2003.
139. Juan-García, A., Font, G., and Picó, Y., Determination of quinolone residues in chicken and fish by
capillary electrophoresis-mass spectrometry, Electrophoresis, 27, 2240, 2006.
140. Juan-García A., Picó, Y., and Font, G., Capillary electrophoresis for analyzing pesticides in fruits and
vegetables using solid-phase extraction and stir-bar sorptive extraction, J. Chromatogr. A, 1073, 229,
2005.
141. Hernández-Borges, J., Cifuentes, A., García-Montelongo, F. J., and Rodríguez-Delgado, M. Á., Com-
bining solid-phase microextraction and online preconcentration-capillary electrophoresis for sensitive
analysis of pesticides in foods, Electrophoresis, 26, 980, 2005.
142. Juan-Garciá, A., Font, G., and Picó, Y., Quantitative analysis of six pesticides in fruits by capillary
electrophoresis-electrospray-mass spectrometry, Electrophoresis, 26, 1550, 2005.
143. Rodríguez, R., Pico, Y., Font, G., and Mañes, J., Analysis of thiabendazole and procymidone in fruits
and vegetables by capillary electrophoresis–electrospray mass spectrometry, J. Chromatogr. A, 949,
359, 2002.
144. Segura Carretero, A., Cruces-Blanco, C., Cortacero Ramírez, S., Carrasco Pancorbo, A., and Fernandez
Gutierrez, A., Application of micellar electrokinetic capillary chromatography to the analysis of
uncharged pesticides of environmental impact, J. Agric. Food Chem., 52, 5791, 2004.
145. Molina, M., and Silva, M., In-capillary derivatization and analysis of amino acids, amino phospho-
nic acid-herbicides and biogenic amines by capillary electrophoresis with laser-induced fluorescence
146. Hilder, E. F., Klampfl, C. W., Buchberger, W., and Haddad, P. R., Comparison of aqueous and
nonaqueous carrier electrolytes for the separation of penicillin V and related substances by capillary
electrophoresis with UV and mass spectrometric detection, Electrophoresis, 23, 414, 2002.
147. Wu, Y., Ho, A. Y. T., Qian, P-Y., Leung, K. S-Y., Cai, Z., and Lin, J.-M., Determination of par-
alytic shellfish toxins in dinoflagellate Alexandrium tamarense by using isotachophoresis/capillary
electrophoresis, J. Separation Sci., 29, 399, 2006.
148. Kowalski, P., Olȩdzka, I., and Lamparczyk, H., Capillary electrophoresis in analysis of veterinary
drugs, J. Pharm. Biomed. Anal., 32, 937, 2003.
149. Hernández, M., Borrull, F., and Calull, M., Using nonaqueous capillary electrophoresis to analyze
several quinolones in pig kidney samples, Electrophoresis, 23, 506, 2002.
150. Kourkine, I. V., Ristic-Petrovic, M., Davis, E., Ruffolo, C. G., Kapsalis, A., and Barron, A. E.,
Detection of Escherichia coli O157:H7 bacteria by a combination of immunofluorescent staining and
151. Palenzuela, B., Simonet, B. M., García, R. M., Ríos, A., and Valcárcel, M., Monitoring of bacterial
contamination in food samples using capillary zone electrophoresis, Anal. Chem., 76, 3012, 2004.
152. Bianco, G., Schmitt-Kopplin, P., De Benedetto, G., Kettrup, A., and Cataldi, T. R. I., Determina-
tion of glycoalkaloids and relative aglycones by nonaqueous capillary electrophoresis coupled with
electrospray ionization-ion trap mass spectrometry, Electrophoresis, 23, 2904, 2002.
153. Huang H. Y., Chiu, C. W., Sue, S. L., and Cheng, C. F., Analysis of food colorants by capillary
electrophoresis with large-volume sample stacking, J. Chromatogr. A, 995, 29, 2003.
154. Soto-Chinchilla, J. J., García-Campaña, A. M., Gámiz-Gracia, L., and Cruces-Blanco, C., Application
of capillary zone electrophoresis with large-volume sample stacking to the sensitive determination of
sulfonamides in meat and ground water, Electrophoresis, 27, 4060, 2006.
155. Castillo, M. A., and Vargas, M. G., Development of capillary electrophoretic methods for bovine
milk quality control, Presented at the 12th Latin-American Symposium on Biotechnology, Biomedical,
Biopharmaceutical and Industrial Applications of Capillary Electrophoresis and Microchip Technol-
ogy, Queretaro, Mexico, 2-5 December, 2006.
156. Vargas, M. G., Trejo-Márquez, M. A., and Rodríguez, F., Analysis of herbicides by capillary elec-
trophoresis, In Applications of Analytical Chemistry in Environmental Research, Special Review
Books, Research Singpost, Trivandrum, Kerala, India, 2005, ISBN: 81-303-0057-8.
157. Hernández-Borges, J., Frías-García, S., Cifuentes, A., and Rodríguez-Delgado, M. A., Pesticide
analysis by capillary electrophoresis, J. Sep. Sci., 27, 947, 2004.
158. Da Costa, C. T., Horton, D., and Margolis, S. A., Analysis of anthocyanins in foods by liquid chro-
matography, liquid chromatography–mass spectrometry and capillary electrophoresis, J. Chromatogr.
A, 881, 403, 2000.
159. Gu, X., Chu, Q., O’Dwyer, M., and Zeece, M., Analysis of resveratrol in wine by capillary
160. Li, P., Li, S. P., and Wang, Y. T., Optimization of CZE for analysis of phytochemical bioactive
compounds, Electrophoresis, 27, 4808, 2006.
161. Herrero, M., Ibáñez, E., and Cifuentes, A., Analysis of natural antioxidants by capillary electromigra-
tion methods, J. Sep. Sci., 28, 883, 2005.
162. Blasco, A. J., Barrigas, I., González, M. C., and Escarpa, A., Fast and simultaneous detection of
prominent natural antioxidants using analytical microsystems for capillary electrophoresis with a
glassy carbon electrode: A new gateway to food environments, Electrophoresis, 26, 4664, 2005.
163. Hamoudová, R., Pospíšilová, M., and Spilková, J., Analysis of selected constituents in methanolic
extracts of Hypericum perforatum collected in different localities by capillary ITP-CZE, Electrophore-
sis, 27, 4820, 2006.
164. Štĕrbová, D., Vc̆èek, J., and Kubán̆, V., Capillary zone electrophoretic determination of phenolic
compounds in chess (Bromus inermis L.) plant extracts, J. Sep. Sci., 29, 308, 2006.
165. Carrasco-Pancorbo, A., Gómez-Caravaca, A. M., Cerretani, L., Bendini, A., Segura-Carretero, A., and
Fernández-Gutiérrez, A., A simple and rapid electrophoretic method to characterize simple phenols,
ligands, complex phenols, phenolic acids, and flavonoids in extravirgin olive oil, J. Sep. Sci., 29, 2221,
2006.
166. Li, X., Zhang, Y., and Yuan, Z., Separation and determination of rutin and quercetin in the flowers of
Sophora japonica L. by capillary electrophoresis with electrochemical detection, Chromatographia,
55, 243, 2002.
167. Suntornsuk, L., Kasemsook, S., and Wongyai, S., Quantitative analysis of aglycone quercetin
in mulberry leaves (Morus alba L.) by capillary zone electrophoresis, Electrophoresis, 24, 1236,
2003.
168. Herrero-Martinez, J. M., Ràfols, C., Rosés, M., Bosch, E., Lozano, C., and Torres, J. L., Micellar
electrokinetic chromatography estimation of size and composition of procyanidins after thiolysis with
cysteine, Electrophoresis, 24, 1404, 2003.
169. Pompanio, R., Gotti, R., Santagati, N. A., and Cavrini, V., Analysis of catechins in extracts of Cistus
species by microemulsion electrokinetic chromatography, J. Chromatogr. A, 990, 215, 2003.
170. Pompanio, R., Gotti, R., Luppi, B., and Cavrini, V., Microemulsion electrokinetic chromatogra-
phy for the analysis of green tea catechins: Effect of the cosurfactant on the separation selectivity,
171. Bonoli, M., Pelillo, M., Toschi, T. G., and Lercker, G., Analysis of green tea catechins: Comparative
study between HPLC and HPCE, Food Chem., 81, 631, 2003.
172. Bonoli, M., Colabufalo, P., Pelillo, M., Toschi, T. G., and Lercker, G., Fast determination of catechins
and xanthines in tea beverages by micellar electrokinetic chromatography, J. Agric. Food Chem., 51,
1141, 2003.
173. Wright, L. P., Pieter Aucamp, J., and Apostolides, Z., Analysis of black tea theaflavins by nonaqueous
174. Wright, L. P., Mphangwe, N. I. K., Nyirenda, H. E., and Apostolides, Z., Analysis of the theaflavin
composition in black tea (Camellia sinensis) for predicting the quality of tea produced in Central and
Southern Africa, J. Sci. Food Agric., 82, 517, 2002.
175. Gao, L., Chu, Q., and Ye, J., Determination of trans-Resveratrol in wines, herbs and health food by
capillary electrophoresis with electrochemical detection, Food Chem., 78, 255, 2002.
176. Brandolini, V., Maietti, A., Tedeschi, P., Durini, E., Vertuani, and Manfredini, S., Capil-
lary electrophoresis determination, synthesis, and stability of resveratrol and related 3-O-β-D-
glucopyranosides, J. Agric. Food Chem., 50, 7407, 2002.
177. Saenz-Lopez, R., Fernandez-Zurbano, P., and Tena, M. T., Development and validation of a
capillary zone electrophoresis method for the quantitative determination of anthocyanins in wine,
J. Chromatogr. A, 990, 247, 2003.
178. Vaher, M., and Koel, M., Separation of polyphenolic compounds extracted from plant matrices using
179. Del Castillo, M. D., Ames, J. M., and Gordon, M. H., Effect of roasting on the antioxidant activity of
coffee brews, J. Agric. Food Chem., 50, 3698, 2002.
180. Kubán̆, P., Štĕrbová, D., and Kubán̆, V., Separation of phenolic acids by capillary electrophoresis with
indirect contactless conductometric detection, Electrophoresis, 27, 1368, 2006.
181. Huang, H.-Y., Lien, W-C., and Chiu, C.-W., Comparison of microemulsion electrokinetic chromatog-
raphy and micellar electrokinetic chromatography methods for the analysis of phenolic compounds,
J. Sep. Sci., 28, 973, 2005.
182. Wang, S. P., and Huang, K.-J., Determination of flavonoids by high-performance liquid chromatogra-
phy and capillary electrophoresis, J. Chromatogr. A, 1032, 273, 2004.
183. Ma, B., Zhou, X., Wang, G., Huang, H., Dai, Z., Qin, J., and Lin, B., Integrated isotachophoretic
preconcentration with zone electrophoresis separation on a quartz microchip for UV detection of
flavonoids, Electrophoresis, 27, 4904, 2006.
184. Carrasco Pancorbo, A., Cruces-Blanco, C., Segura Carretero, A., and Fernandez Gutierrez, A., Sensi-
tive determination of phenolic acids in extra-virgin olive oil by capillary zone electrophoresis, J. Agric.
Food Chem., 52, 6687, 2004.
185. Peng, Y., Ye, J., and Kong, J., Determination of phenolic compounds in Perilla frutescens L. by
capillary electrophoresis with electrochemical detection, J. Agric. Food Chem., 53, 8141, 2005.
186. Hamoudová, R., Urbánek, M., Pospíšilová, M., and Polášek. M., Assay of phenolic compounds in
red wine by on-line combination of capillary isotachophoresis with capillary zone electrophoresis,
J. Chromatogr. A, 1032, 281, 2004.
187. Ehala, S., Vaher, M., and Kaljurand, M., Characterization of phenolic profiles of Northern European
Berries by capillary electrophoresis and determination of their antioxidant activity, J. Agric. Food
Chem., 53, 6484, 2005.
188. Lee, B. L., and Ong, C. N., Comparative analysis of tea catechins and theaflavins by high performance
liquid chromatography and capillary electrophoresis, J. Chromatogr. A, 881, 439, 2000.
189. Cifuentes, A., Bartolomé, B., and Gómez-Cordovés, C., Fast determination of procyanidins and other
phenolic compounds in food samples by micellar electrokinetic chromatography using acidic buffers,
190. Scampicchio, M., Wang, J., Mannino, S., and Prakash Chatrathi, M., Microchip capillary electrophore-
sis with amperometric detection for rapid separation and detection of phenolic acids, J. Chromatogr.
A, 1049, 189, 2004.
191. Barbas, C., and Saavedra, L., Chiral analysis of aliphatic short chain organic acids by capillary
192. Simo, C., Rizzi, A., Barbas, C., and Cifuentes, A., Chiral capillary electrophoresis-mass spectrometry
of amino acids in foods, Electrophoresis, 26, 1432, 2005.
193. Ward, T. J., Chiral Separations, Anal. Chem., 78, 3947, 2006.
194. Van Eeckhaut, A., and Michotte, Y., Chiral separations by capillary electrophoresis: Recent
developments and applications, Electrophoresis, 27, 2880, 2006.
195. Vespalec R., and Bocĕk, P., Chiral separations in capillary electrophoresis, Chem. Rev., 100, 3715,
2000.
196. Gübitz, G., and Schmid, M. G., Advances in chiral separation using capillary electromigration
techniques, Electrophoresis, 27, 1, 2006.
197. Simó, C., Barbas, C., and Cifuentes, A., Chiral electromigration methods in food analysis,
198. Gübitz, G., and Schmid, M. G., Recent advances in chiral separation principles in capillary
electrophoresis and capillary electrochromatography, Electrophoresis, 23, 3981, 2004.
199. Nouadje, G., Nertz, M., and Courderc, F., Study of the racemization of serine by cyclodextrin-modified
micellar electrokinetic chromatography and laser-induced fluorescence detection, J. Chromatogr. A,
716, 331, 1995.
200. Simo, C., Martin-Alvarez, P. J., Barbas, C., and Cifuentes, A., Application of stepwise discriminant
analysis to classify commercial orange juices using chiral micellar electrokinetic chromatography-laser
induced fluorescence data of amino acids, Electrophoresis, 25, 2885, 2004.
201. Liu, X., Li, D. F., Wang, Y., and Lu, Y. T., Determination of 1-aminocyclopropane-1-carboxylic acid in
apple extracts by capillary electrophoresis with laser-induced fluorescence detection, J. Chromatogr. A,
1061, 99, 2004.
202. Carlavilla, D., Moreno-Arribas, M. V., Fanali, S., and Cifuentes, A., Chiral MEKC-LIF of amino acids
in foods: Analysis of vinegars, Electrophoresis, 27, 2551, 2006.
203. Kodama, S., Yamamoto, A., Ohura, T., Matsunaga, A., and Kanbe, T., Enantioseparation of imazalil
residue in orange by capillary electrophoresis with 2-hydroxypropyl-â-cyclodextrin as a chiral selector,
J. Agric. Food Chem., 51, 6128, 2003.
204. Kodama, S., Yamamoto, A., Saitoh, Y., Matsunaga, A., Okamura, K., Kizu, R., and Hayakawa, K.,
Enantioseparation of vinclozolin by β-cyclodextrin-modified micellar electrokinetic chromatography,
J. Agric. Food Chem., 50, 1312, 2002.
205. Pumera, M., Flegel, M., Lepša, L., and Jelínek, I., Chiral analysis of biogenic dl-amino acids deriva-
tized by urethane-protected α-amino acid N-carboxyanhydride using capillary zone electrophoresis
and micellar electrokinetic chromatography, Electrophoresis, 23, 2449, 2002.
206. Chen, F., Zhang, S., Qi, L., and Chen, Y., Chiral capillary electrophoretic separation of amino acids
derivatized with 9-fluorenylmethylchloroformate using mixed chiral selectors of β-cyclodextrin and
sodium taurodeoxycholate, Electrophoresis, 27, 2896, 2006.
207. Kodama, S., Yamamoto, A., and Matsunaga, A., Direct chiral resolution of pantothenic acid using
2-hydroxypropyl-β-cyclodextrin in capillary electrophoresis, J. Chromatogr. A, 811, 269, 1998.
208. Kodama, S., Aizawa, S.-I., Taga, A., Yamashita, T., and Yamamoto, A., Chiral resolution of monosac-
charides as 1-phenyl-3-methyl-5-pyrazolone derivatives by ligand-exchange CE using borate anion as
a central ion of the chiral selector, Electrophoresis, 27, 4730, 2006.
209. Vandenabeele-Trambouze, O., Albert, M., Bayle, C., Couderc, F., Commeyras, A., Despois, D.,
Dobrijevic, M., and Grenier Loustalot, M.-F., Chiral determination of amino acids by capillary elec-
trophoresis and laser-induced fluorescence at picomolar concentrations, J. Chromatogr. A, 894, 259,
2000.
210. Koidl, J., Hödl, H., Schmid, M. G., Pantcheva, S., Pajpanova, T., and Gübitz, G., Chiral separation
of halogenated amino acids by ligand-exchange capillary electrophoresis, Electrophoresis, 26, 3878,
2005.
211. Gel-Moreto, N., Streich, R., and Galensa, R., Chiral separation of six diastereomeric flavanone-7-
O-glycosides by capillary electrophoresis and analysis of lemon juice, J. Chromatogr. A, 925, 279,
2001.
212. Kodama, S., Yamamoto, A., Matsunaga, A., Soga, T., and Minoura, K., Direct chiral resolution of
lactic acid in food products by capillary electrophoresis, J. Chromatogr. A, 875, 371, 2000.
213. Kodama, S., Yamamoto, A., Matsunaga, A., and Hayakawa, K., Direct chiral resolution of tartaric
acid in food products by ligand exchange capillary electrophoresis using copper (II)–d-quinic acid as
a chiral selector, J. Chromatogr. A, 932, 139, 2001.
214. Zhang, M., and El Rassi, Z., Enantiomeric separation by capillary electrochromatography II. Chiral
separation of dansyl amino acids and phenoxy acid herbicides on sulfonated silica having surface-bound
hydroxypropyl-β-cyclodextrin, Electrophoresis, 21, 3135, 2000.
215. Kato, M., Dulay, M. T., Bennett, B., Chen, J.-R., and Zare, R. N., Enantiomeric separation of amino
acids and nonprotein amino acids using a particle-loaded monolithic column, Electrophoresis, 21,
3145, 2000.
216. Jin, L. J., Rodriguez, I., and Li, S. F. Y., Enantiomeric separation of amino acids derivatized with fluores-
ceine isothiocyanate isomer I by micellar electrokinetic chromatography using b- and g-cyclodextrins
as chiral selectors, Electrophoresis, 20, 1538, 1999.
217. Gong, X. Y., and Hauser, P. C., Enantiomeric separation of underivatized small amines in conventional
and on-chip capillary electrophoresis with contactless conductivity detection, Electrophoresis, 27,
4375, 2006.
218. Fanali, S., Crucianelli, M., De Angelis, F., and Presutti, C., Enantioseparation of amino acid derivatives
by capillary zone electrophoresis using vancomycin as chiral selector, Electrophoresis, 23, 3035, 2002.
219. La, S., Ahn, S., Kim, J.-H., Goto, J., Choi, O.-K., and Kim, K.-R., Enantioseparation of chiral
aromatic amino acids by capillary electrophoresis in neutral and charged cyclodextrin selector modes,
220. Zakaria, P., Macka, M., and Haddad, P. R., Optimisation of selectivity in the separation of aromatic
amino acid enantiomers using sulfated β-cyclodextrin and dextran sulfate as pseudostationary phases,
221. Allen, D. J., Gray, J. C., Payva, N. L., and Smith, J. T., An enantiomeric assay for the flavonoids
medicarpin and vestitone using capillary electrophoresis, Electrophoresis, 21, 2051, 2000.
222. Mikuŝ, P., Kaniansky, D., and Fanali, S., Separation of multicomponent mixtures of 2,4-dinitrophenyl
labelled amino acids and their enantiomers by capillary zone electrophoresis, Electrophoresis, 22, 470,
2001.
223. Simó, C., Barbas, C., and Cifuentes, A., Sensitive micellar electrokinetic chromatography-laser-
induced fluorescence method to analyze chiral amino acids in orange juices, J. Agric. Food Chem., 50,
5288, 2002.
224. Recio, I., Ramos, M., and López-Fandiño, R., Capillary electrophoresis for the analysis of food
proteins of animal origin, Electrophoresis, 22, 1489, 2001.
225. Bean, S. R., and Lookhart, G. L., High-performance capillary electrophoresis of meat, dairy and cereal
proteins, Electrophoresis, 22, 4207, 2001.
226. Sorensen, H., Sorensen, S., Bjergegaard, C., and Michaelsen, S., Protein purification and analysis, in
Chromatrography and Capillary Electrophoresis in Food Analysis, RSC Food Analysis Monographs,
Belton, P.S., Ed., Norwich, U.K., 1999, p. 315.
227. Gonzalez, N., Elvira, C., San Román, J., and Cifuentes, A., New physically coated adsorbed poly-
mer coating for reproducible separations of basic and acidic proteins by capillary electrophoresis,
J. Chromatogr. A, 1012, 95, 2003.
228. Dolnik, V., Capillary electrophoresis of proteins 2003-2005, Electrophoresis, 27, 126, 2006.
229. Recio, I., Amigo, L., Lopez-Fandiño, R., Assessment of the quality of dairy products by capillary
electrophoresis of milk proteins, J. Chromatogr. B, 697, 231, 1997.
230. Vallejo-Cordoba, B., Rapid separation and quantification of major caseins and whey proteins of bovine
milk by capillary electrophoresis, J. Cap. Elec., 4, 219, 1997.
231. de Jong, N., Visser, S., and Olieman, C. Determination of milk proteins by capillary electrophoresis,
J. Chromatogr. A, 652, 207, 1993.
232. Recio, I., and Olieman, C., Determination of denatured serum proteins in the casein
fraction of heat-treated milk by capillary zone electrophoresis, Electrophoresis, 17, 1228,
1996.
233. Miralles, B., Rothbauer, V., Manso, M., Amigo, l., Krause, I., and Ramos, M., Improved method
for the simultaneous determination of whey proteins, caseins, and para-k-casein in milk and dairy
products by capillary electrophoresis. J. Chromatogr. A, 915, 225, 2001.
234. Fairise, J. F., and Cayot, P., New ultrarapid method for the separation of milk proteins by capillary
electrophoresis, J. Agric. Food Chem., 46, 2628, 1998.
235. Olguin-Arredondo, H., and Vallejo-Cordoba, B., Separation and determination of β-lactoglobulin
variants A and B in cow’s milk by capillary free zone electrophoresis. J. Cap. Elec. and Microchip
Tech., 6, 145, 1999.
236. Veledo, M. T., Frutos, M., and Diez-Masa, J. C., Development of a method for quantitative analysis
of the major whey proteins by capillary electrophoresis with online capillary derivatization and laser-
induced fluorescence detection, J. Sep., Sci., 28, 935, 2005.
237. Veledo, M. T., Frutos, M., and Diez-Masa, J. C., Analysis of trace amounts of bovine β-lactoglobulin
in infant formulas by capillary electrophoresis with on-capillary derivatization and laser-induced
fluorescence detection, J. Sep., Sci., 28, 941, 2005.
238. Miralles, B., Krause, I., Ramos, M., and Amigo, L., Comparison of capillary electrophoresis and
isoelectric focusing for analysis of casein/caseinate addition in processed cheeses, Int. Dairy J., 16,
1448, 2006.
239. Rodriguez-Nogales, J. M., and Vazquez, F., Application of electrophoretic and chemometric analysis
to predict the bovine, ovine and caprine milk percentages in Panela cheese, an unripened cheese, Food
Control, 18, 580, 2007.
240. Otte, J., Shalaby, S. M., Zakora, M., Pripp, A. H., El-Shabrawy, S. A., Angiotensin-converting enzyme
inhibitory activity of milk protein hydrolysates: Effect of substrate, enzyme and time of hydrolysis,
Int. Dairy J., 17, 488, 2007.
241. De Noni, I., Pellegrino, L., Cattaneo, S., Resmini, P., HPLC of proteose peptones for evaluating ageing
of packaged pasteurized milk. Int. Dairy J., 17, 12, 2007.
242. Olguin-Arredondo, H., Vallejo-Cordoba, B., and Gonzalez-Cordova, A. F., Micropreparative sepa-
ration, fractionation, an peptide mapping of β-lactoglobulin A and B variants by capillary free zone
electrophoresis, J. Cap. Elec. Microchip Tech., 9, 2005.
243. Lookhart, G., Bean, L., and Scott, R., Capillary electrophoresis of cereal proteins: An overview, J. Cap.
Elec. Microchip Technol., 9, 23, 2004.
244. Lookhart, G., and Bean, L., High performance capillary electrophoresis of meat, dairy, and cereal
proteins, Electrophoresis, 22, 4207, 2001.
245. Bean, S. R., Bietz, J. A., and Lookhart, G., High performance capillary electrophoresis of cereal
proteins, J. Chromatogr. A, 814, 25, 1998.
246. Bean, S. R., Bietz, J. A., and Lookhart, G., Ultrafast capillary electrophoretic analysis of cereal storage
proteins and its applications to protein characterization and cultivar differentiation, J. Agric. Food
Chem., 48, 344, 2000.
247. Cota-Rivas, M., and Vallejo-Cordoba, B., Capillary electrophoresis for meat species differentiation,
J. Cap. Elec., 4, 195, 1998.
248. Vallejo-Cordoba, B., and Cota-Rivas, M., Meat species identification by linear discriminant analysis
of capillary electrophoresis protein profiles, J. Cap. Elec., 5, 171, 1998.
249. Simo, C., Elvira, C., Gonzalez, N., San Roman, J., Barbas, C., and Cifuentes, A., Capillary
electrophoresis-mass spectrometry of basic proteins using a new physically adsorbed polymer coating.
Some applications in food analysis, Electrophoresis, 25, 2056, 2004.
250. Meyer, R., and Candrian, U., PCR-based DNA analysis for the identification and characterization of
food components, Lebensm-Wiss u. Technol., 29, 1, 1996.
251. Lockley, A. K., and Bardsley, R. G., DNA-based methods for food authentication. Trends Food Sci.
Technol., 11, 67, 2000.
252. Di Pinto, A., Forte, V. T., Guastadisegni, M. C., Martino, C., Schena, F. P., and Tantillo, G. A.,
Comparison of DNA extraction methods for food analysis, Food Control, 18, 76, 2007.
253. Garcia-Cañas, V., González, R., and Cifuentes, A., Detection of genetically modified maize by
polymerase chain reaction and capillary gel electrophoresis with UV detection and laser-induced
fluorescence, J. Agric. Food Chem., 50, 1016, 2002.
254. Garcìa-Cañas, V., Gonzalez, R., and Cifuentes, A., The combined use of molecular techniques and
capillary electrophoresis in food analysis, Trends Anal. Chem., 23, 637, 2004.
255. Garcia-Cañas, V., Cifuentes, A., and Gonzalez, R., Quantitation of transgenic maize using double
quantitative competitive polymerase chain reaction and capillary gel electrophoresis laser induced
256. Sun, Y. L., and Lin, C. S., Establishment and application of a fluorescent polymerase chain reaction
restriction fragment length polymorphism (PCR-RFLP), method for identifying porcine, caprine, and
bovine meats, J. Agric. Food Chem., 51, 1771, 2003.
257. Zeleny, R., Bernreuther, A., Schimmel, H., and Pauwels, J., Evaluation of PCR-based beef sexing
methods, J. Agric. Food Chem., 50, 4169, 2002.
258. Alarcon, B., Garcia-Cañas, V., Cifuentes, A., González, R., and Aznar, R., Simultaneous and sensitive
detection of three foodborne pathogens by multiplex PCR, capillary gel electrophoresis, and laser-
induced fluorescence, J. Agric. Food Chem., 52, 5583, 2004.
259. Garcia-Cañas, V., Macian, M. C., Chenoll, E., Aznar, R., González, R., and Cifuentes, A., Detection
and differentiation of several food-spoilage lactic acid bacteria by multiplex polymerase chain reaction,
capillary gel electrophoresis, and laser-induced fluorescence, J. Agric. Food Chem., 52, 5583, 2004.
260. Martin, A., Jurado, M., Rodriguez, M., Nuñez, F., and Cordoba, J. J., Characterization of molds from
dry cured meat products and their metabolites by micellar electrokinetic capillary electrophoresis and
random amplified polymorphic DNA PCR, J. Food Prot., 67, 2234, 2004.
261. Garcìa-Cañas, V., Cifuentes, A., and Gonzalez, R., Detection of genetically modified organisms in
foods by DNA amplification techniques, Crit. Rev. Food Sci. Nutr., 44, 425, 2004.
262. Doodley, J. J., Sage, H. D., Brown, H. M., and Garrett, S. D., Improved fish species identification by
use of lab-on-a-chip technology, Food Control, 16, 601, 2005.
263. Doodley, J. J., Sage, H. D., Clarke, M. A. L., Brown, H. M., and Garrett, S. D., Fish species identi-
fication using PCR-RFLP analysis and lab-on-a-chip capillary electrophoresis. Application to detect
white fish species in food products and an interlaboratory study, J. Agric. Food Chem., 53, 3348, 2005.
264. Boyce, M. C., Determination of additives in food by capillary electrophoresis, Electrophoresis, 22,
1447, 2001.
265. Frazier, R. A., Inns, E. L., Dossi, N., Ames, J. M., and Nursten, H. E., Development of a capillary
electrophoresis method for the simultaneous analysis of artificial sweetners, preservatives and colors
in soft drinks. J. Chromatogr. A, 876, 213, 2000.
266. Huan, H. Y., Shih, Y. C., and Chen, Y. C., Determining eight colorants in milk beverages by capillary
267. Huan, H. Y., Chuang, C. L., Chiu, C. W., and Chung, M. C., Determination of food colorants by
microemulsion electrokinetic chromatography, Electrophoresis, 26, 867, 2005.
268. Jager, A. V., Tonin, F. G., and Tavares, M. F. M., Optimizing the separation of food dyes by capillary
269. Law, W. S., Kuba, P., Zhao, J. H., Li, S. F. Y., and Hauser, P. C., Determination of vitamin C and
preservatives in beverages by conventional capillary electrophoresis and microchip electrophoresis
with capacitively coupled contactless conductivity detection, Electrophoresis, 26, 4648, 2005.
270. Oztekin, N., Nutku, M. S., and Erim, F. B., Simultaneous determination of nitrite and nitrate in meat
products and vegetables by capillary electrophoresis, Food Chem., 76, 103, 2002.
271. Ohashi, M., Omae, H., Hashida, M., Sowa, Y., and Imai, S., Determination of vanillin and related
flavor compounds in cocoa drink by capillary electrophoresis, J. Chromatogr. A, 1138, 262, 2007.
272. McCourt, J., Stroka, J., and Anklam, E., Experimental design-based development and single laboratory
validation of a capillary zone electrophoresis method for the determination of the artificial sweetener
sucralose in food matrices, Anal. Bioanal. Chem., 382, 1269, 2005.
273. Ding, Y., Mora, M. F., and Garcia, C. D., Analysis of alkyl gallates and nordihydroguaiaretic acid
using plastic capillary electrophoresis—microchips, Analytic Chimica Acta, 561, 126, 2006.
274. De Block, J., Merchiers, M., Mortier, L., Brraekman, A., Ooghe, W., and Van Renterghem, R.,
Monitoring nutritional quality of milk powders: Capillary electrophoresis of the whey protein fraction
compared with other methods, International Dairy Journal, 13, 87, 2003.
275. Vallejo-Cordoba, B., Mazorra-Manzano, M. A., and Gonzalez-Cordova, A. F., New capillary elec-
trophoresis method for the determination of furosine in dairy products, J. Agric. Food Chem., 52,
5787, 2004.
276. Morales, F., and Jimenez-Perez, S., Hydroxymethylfurfural determination in infant milk-based
formulas by micellar electrokinetic capillary chromatography. Food Chem., 72, 525, 2001.
277. Vallejo-Cordoba, B., and Gonzalez-Cordova, A. F., Latest advances in the chemical characterization
of Mexican distilled beverages: Tequila, Mezcal, Bacanora and Sotol, In Hispanic Foods, Chemistry
and Flavor, ACS Series, 946, Tunick, M.H., and González de Mejia, E., Ed., 2007, 153.
278. Wrigley, C. W., Bate, I. L., Uthayakumaran, S., and Rathmell, W. G., Modern approaches to food
diagnostics for grain quality assurance. Food Australia, 58, 538, 2006.
279. Vallejo-Cordoba, B., Mazorra-Manzano, M. A., and Gonzalez-Cordova, A. F., Determination of
β-hydroxyacyl CoA-dehydrogenase activity in meat by electrophoretically mediated microanalysis,
J. Cap. Elec. Micro. Tech., 8, 81, 2003.
280. de Oliveira, M. A. L., Solis, V. E. S., Gioielli, L. A., Polakiewicz, B., and Tavares, M. F. M., Method
development for the analysis of transfatty acids in hydrogenated oils by capillary electrophoresis,
281. Vallejo-Cordoba, B., Mazorra-Manzano, M. A., and Gonzalez-Cordova, A. F., Determination of short-
chain free fatty acids in lipolyzed milk fat by capillary electrophoresis, J. Cap. Elec., 5, 11, 1998.
31 Separation Strategies for
Environmental Analysis
Fernando G. Tonin and Marina F. M. Tavares
CONTENTS
31.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 914

31.2 Separation Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 914
31.2.1 CZE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 914
31.2.2 EKC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 915
31.2.2.1 MEKC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 915
31.2.2.2 Other EKC Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 917
31.3 Sensitivity Enhancement Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 918
31.3.1 Preconcentration Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 918
31.3.2 Alternative Detection Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 919
31.3.2.1 Fluorescence Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 919
31.3.2.2 Chemiluminescence Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 920
31.3.2.3 Electrochemical Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 920
31.3.2.4 Mass Spectrometry Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 921
31.4 Representative Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 921
31.4.1 Pesticides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 921
31.4.2 PAH, PCB, and PCDD. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 928
31.4.3 Phenols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 928
31.4.4 Amines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 930
31.4.4.1 Aliphatic Amines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 933
31.4.4.2 Aromatic Amines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 934
31.4.5 Carbonyls. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 934
31.4.6 Small Ions and Organometallic Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 937
31.4.6.1 Inorganic Cations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 939
31.4.6.2 Inorganic Anions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 941
31.4.6.3 Simultaneous Detection of Cations and Anions . . . . . . . . . . . . . . . . . . . . . . 941
31.4.6.4 Speciation and Organometallic Compounds . . . . . . . . . . . . . . . . . . . . . . . . . 942
31.4.7 Explosives and Warfare Residues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 942
31.4.7.1 Explosives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 942
31.4.7.2 Warfare Residues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 942
31.4.8 Aromatic Sulfonates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 949
31.4.9 Surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 952
31.4.10 Dyes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 953
31.4.11 Endocrine Disruptors and Pharmaceuticals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 954
31.4.11.1 Phenolic Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 954
31.4.11.2 Phthalate Esters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 954
31.4.11.3 Pharmaceutical Residues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 954
31.4.12 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 959
913
31.4.12.1 Humic Substances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 959

31.4.12.2 Algal Toxins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 960
31.4.12.3 Other Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 960
31.5 A Method Development Guide for Environmental Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 961
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 963
31.1 INTRODUCTION
In the past 25 years, capillary electrophoresis (CE) has conquered a solid position in the scientific
community, supported by vast literature compiling the intricacy of its theoretical aspects and the
diversity of its applications (Chapter 1 by Landers brings an introductory panel on CE technology).
Several relevant characteristics of CE such as high resolution, high efficiency, and speed of analysis
in addition to excellent mass sensitivity, low consumption of reagents, and small sample volumes
are in perfect tuning with the demands of environmental analyses. Moreover, whenever limits of
detection (LOD) are favorable, the direct injection of complex samples is possible without being
detrimental to the column integrity, which in turn contributes to preclude time-consuming sample
preparation, typical of environmental procedures. Another particularity of the CE technique is the
variety of separation mechanisms that can be practiced in a single capillary column, allowing the
simultaneous assessment of distinct classes of pollutants in a given sample.
31.2 SEPARATION STRATEGIES

The CE determination of broad classes of atmospheric pollutants and water and soil contaminants
has been reviewed regularly.1–7 This chapter focuses on separation strategies for environmental
analysis involving the two most commonly used CE modes: capillary zone electrophoresis (CZE)
and electrokinetic chromatography (EKC).
31.2.1 CZE
CZE separations are based solely on the differences in the electrophoretic mobilities of charged
species, either in aqueous or nonaqueous media (this latter often referred to as nonaqueous capillary
electrophoresis, NACE). In CZE, the migration of a species within the capillary column is the net
result of mass transport phenomena and chemical equilibria. Two modes of migration are possible,
that is, under suppressed electroosmotic flow (EOF), achieved at low pH buffers or by the use of sur-
face modified capillaries, and in the presence of EOF; in the latter, two possibilities arise: separations
under co- and counter-EOF, depending on the relative mobility of the analyte and EOF itself. With
the proper control of electrolyte composition (buffer type regarding both co- and counterions, buffer
pH and concentration, as well as additives), the analyte mobility can be altered. Flow characteristics
are also dependable on the electrolyte composition as well as on the capillary surface condition.
Organic solvents are among the most commonly used additives in CZE. Several other additives
(complexing agents to discriminate metal cations, quaternary alkylammonium salts as flow reversers
in the separation of small anions, and neutral cyclodextrins [CDs] as chiral selectors) are listed in
the literature as modifiers in CZE separations. In contrast, separations with charged CDs or mixtures
of neutral and charged CDs are often classified as EKC methodologies (vide next section). Consid-
ering that low-molecular mass additives represent in fact borderline cases between CZE and EKC
systems, arbitrary assumptions will be made in this chapter with the purpose of classifying the liter-
ature methodologies. Cyclodextrins either neutral, charged, or mixtures of both will be considered
secondary phases in EKC separations whereas other complexing agents (ethylenediaminetetraacetic
acid [EDTA], o-phenanthroline, crown ethers, etc.) will be classified as additives of CZE electrolytes.
Separation Strategies for Environmental Analysis 915
Quaternary alkylammonium salts, used as flow reverser additives in the separation of small anions,
will fall technically into the EKC definition, since the surfactant micellization is anticipated due
to the electrolyte ionic strength in which the separation is conducted (vide section on small ions
analysis).
CZE is, therefore, suitable for the determination of explicitly charged small ions, whose mobil-
ities already differ by some extent or might be modified by additives, as well as ionizable organic
environmental pollutants (for instance, compounds containing carboxylic acid, phenol, or amine
functionalities). Cations are generally separated under co-EOF whereas anions with low to moder-
ately low mobility are separated counter-electroosmotically. Examples of the CZE analysis of small
ions in methanol (MeOH)-modified borate electrolytes include the speciation of mercury [inorganic
Hg(II), methyl-, ethyl-, and phenylmercury] in water and dogfish muscle8 and the direct ultra-
violet (UV) determination of anions (Br− , I− , NO− 2− 2− − −
2 , S2 O3 , CrO4 , NO3 , SCN , Fe(CN)6 ,
4−
MoO2− 2−
4 , and WO4 ) in effluents of a power plant. For ionizable compounds, the analyte acquires
9
an electrophoretic effective mobility (summation of ionic species mobilities weighted by the species
availability at a given pH). The selective inspection of 21 aromatic amines in groundwater and soil
sample using low pH phosphate buffer in capillaries modified by 1,3-aminopropane is a represen-
tative example of the CZE mode for ionizable compounds.10 Another example is the determination
of the priority phenols in spiked wastewater samples using ammonium acetate buffer modified by
N-methyl formamide/acetonitrile (ACN) mixtures.11
31.2.2 EKC
Electrokinetic chromatography separations are those put into practice with electrolytes containing
a secondary phase that migrate with a particular velocity distinct from that of the analyte. The
fundamental requirement for EKC separations is that either the analyte or the secondary phase
must be charged under the electrolyte medium pH. As pointed out previously, for charged analytes
forming adducts with small molecules, there is a blurred distinction between CZE and EKC separation
principles.
31.2.2.1 MEKC
Micellar electrokinetic chromatography (MEKC) is a particular EKC mode where the secondary
phase is composed by micellized surfactant (MEKC is discussed in detail in Chapter 3 by Terabe).
Solute differential retention occurs as a result of a partition mechanism between a dispersed phase
defined by the total volume of micelles and the remaining aqueous phase. MEKC modes of elution
comprise normal, restricted, and reversed MEKC, based on the relative migration of the analyte and
secondary phase apparent velocity.
The anionic surfactant sodium dodecylsulfate (SDS) is by far the most commonly used surfactant
in micellar separations. Examples of the use of simple buffered/SDS systems in environmental appli-
cations include the simultaneous analysis of 10 N-methylcarbamate pesticides and their hydrolytic
phenolic metabolites in river, well, and pond water (pH 8 phosphate/borate buffer/SDS),12 and
the analysis of insecticides (imidacloprid and its metabolite 6-chloronicotinic acid) in air sam-
ples collected from a greenhouse cropped with tomatoes (pH 8.5 ammonium chloride/ammonia
buffer/SDS).13
The use of SDS in MEKC poses two important shortcomings regarding the nature of the com-
pounds eligible for separation: hydrophilic compounds, especially anionic in character (electrostatic
repulsion), are poorly retained even at high SDS concentrations, whereas hydrophobic compounds
are highly retained even at low SDS concentrations and little selectivity is, therefore, provided.
The alternative use of cationic surfactants (quaternary alkylammonium salts) is recommended
to increase retention of highly hydrophilic compounds. Cationic surfactants reverses the EOF.
As the cationic micelles migrate electrophoretically against the anodic EOF, an extended migra-
tion window results. An example of MEKC separation using cationic surfactants is the analysis
of various s-triazines, including five chloro-, three methoxy-, and five alkylthio-s-triazines in
tetradecyltrimethylammonium bromide (TTAB) electrolytes.14
For the MEKC analysis of highly hydrophobic compounds, several strategies have been proposed:
the use of organic additives (short-chain alcohols, acetonitrile, tetrahydrofuran, dioxane, and
dimethylsulfoxide are among the most studied), and the addition of large amounts of urea and
glucose have been reported. A few examples of solvent-modified electrolytes within the envi-
ronmental context include (i) separation of seven polynuclear aromatic hydrocarbons (PAHs) in
borate/SDS/ACN for inspection of deliberately contaminated soils submitted to biological decon-
tamination process and spent machine oil;15 (ii) determination of 11 triazines in groundwater using
as optimized electrolyte, borate/phosphate/SDS/1-propanol;16 (iii) the analysis of 21 mono- and
dinaphthalenesulfonates, as well as their hydroxy- and amino-derivatives in river water, with
borate/SDS/ACN;17 and (iv) determination of homologues and isomers of linear alkylbenzene-
sulfonates (LAS) in household products and sewage sludge samples using borate/phosphate/ACN
electrolytes.18
Other strategies to cope with hydrophobic compounds include the addition of neutral surfactants
or neutral CDs to SDS micelle systems, use of bile salts, and by exploring solvophobic association
with tetrahexyl- or tetraheptylammonium ions (added as perchlorate salt [THxAP] or bromide salt
[THpAB]) or dioctylsulfosuccinate (DOSS).
In the separation of aromatic sulfonates, 10 electrolyte compositions were investigated with low
and high pH buffers containing SDS, Brij35, and octylamine. Twenty-one compounds could be
separated and the contents of aromatic sulfonates in an industrial effluent were determined.19 Other
examples of mixed micelle systems applied to compounds of environmental relevance include (i) the
analysis of eight azo dyes, mono-, and disulfonated compounds with Brij35/SDS electrolytes;20
(ii) the separation of 16 arylamine isomers with Tween80/SDS/cholate systems (lake water near
industrial area);21 (iii) the analysis of eight aromatic compounds (phenol derivatives and PAH) with
Tween20/SDS electrolytes;22 and (iv) the separation of six phenylureas and chlorsulfuron pesticides
with PEG400/SDS systems.23
Although the target function of neutral CDs is to exert chiral selection in the EKC separation of
optical isomers, they have often been used as auxiliary complex ligands as a means of improving reso-
lution of closely eluting achiral positional and structural related compounds or to reduce significantly
apparent retention factors. The separation of seven positional and structural naphthalenesulfonate
isomers (pH 3.0 phosphate buffer/β-CD)24 and five 2,4-dinitrophenylhydrazine (DNPH)-aldehyde
derivatives in vehicular emission (pH 9.0 borate buffer/SDS/β-CD)25 are examples of neutral
CD-mediated separations.
Studies involving bile salts in EKC separations include (i) determination of s-triazines and quats
in well water samples;26 (ii) separation of the 16 priority pollutant PAHs for inspection in ambient air
samples;27 (iii) characterization of the electrophoretic behavior of 56 aromatic compounds (phenoxy
acid herbicides, phenylalkanoic acids, aromatic carboxylic acids, aromatic sulfonic acids, azo and
other dyes, and nitrogen-containing aromatic acids) for further assessment of extraction and sample
clean-up procedures using spiked water and soils;28 and (iv) recovery of synthetic dyes from spiked
water and soil matrices.29 Another interesting application of MEKC with diverse bile salts is the
prediction of ecotoxicity of aromatic compounds.30
The use of carboxylic and dicarboxylic acids, SDS, bile salts, organic solvents, and alkylammo-
nium ions was explored to study the separation of LAS homologues and positional isomers,31,32 as
well as alkylether sulfate oligomers.33 The MEKC separation of mixtures of the surfactant classes
coconut diethanolamide, cocamido propyl betaine, and alkylbenzene sulfonate was studied in either
low pH phosphate or high pH borate or dipentylamine buffers containing as surfactants deoxy-
cholate or SDS, organic solvents (methanol, acetonitrile, n-propanol, and n-butanol), and anionic
solvophobic agents (DOSS, fatty acids).34
The use of DOSS/ACN mixtures in the separation of 9-fluoroenylmethyl chloroformate (FMOC)-
derivatized anilines was investigated for inspection of lake water.35 While the organic modifier
allowed the solubilization of the hydrophobic solutes and maintained the DOSS surfactant in its
monomeric form by inhibiting micellization, the DOSS surfactant associated with the FMOC ani-
lines to a varying degree led to their differential migration and separation. Seven conventional
chromatographic ion-pairing agents including tetraalkylammonium series and alkanesulfonic series
were tested comparatively in a recent study involving the separation of 13 PAH in methanolic/ACN
matrices.36
Alternative and more complex surfactants continue to be explored as a resourceful option for
MEKC separations. In situ generated micelles, which are anionic complexes formed by alkyl-
or steroidal-glycoside surfactants and borate ions, the use of sodium 10-undecylenate (SUA)
and sodium 10-undecylsulfate (SUS) oligomers as well as surfactants composed of two ionic
groups and two liphophilic chains, such as sodium 5,12-bis(dodecyloxymethyl)-4,7,10,13-(tetraoxa)-
1,16-hexadecanedisulfonate (DBTD), bilayered aggregates such as vesicles and liposomes, and
bilayer micelles are a few examples.
In situ generated micelles have been applied to the inspection of aniline pesticidic metabolites in
lake water.37 The separation of 16 PAH in SUA oligomer electrolytes was reported.38 Creosote-
contaminated soil samples were extracted by accelerated solvent extraction using methylene
chloride–acetone mixtures. The extracts were further fractioned by gel permeation chromatogra-
phy before analysis. The EKC chromatogram of a creosote-contaminated soil fraction shows the
resolution of at least 50 peaks. The separation of the 11 priority phenols in river and sea water39
was demonstrated in MEKC with DBTD surfactants, whereas examples of the use of liposomes as
carriers include the separation of benzene derivatives and phenols.40
High-molecular mass surfactants such as butyl acrylate-butyl methacrylate-methacrylic acid
copolymer sodium salts, starburst dendrimers, poly(amidoamines), and diaminobutane-based
poly(propyleneimine) as well as cationic polyelectrolytes (ionenes) had all been presented as suc-
cessful secondary phases for aromatic compounds. The determination of 10 nitrophenols in glycine
buffers modified by β-CD (0–10 mmol L−1 ) and polyvinylpyrrolidone (PVP) (0.5–2.5% w/v) is an
example of application of polymer-based electrolytes to rain, tap, and process water.41
31.2.2.2 Other EKC Modes
31.2.2.2.1 CD-EKC and CDCD-EKC

Single negatively and positively charged CDs (e.g., sulfobutyl ether β-CD, SB-β-CD, sulfated CD,
carboxymethylated β-CD, and methylamino substituted β-CD) or even mixtures of neutral and
charged CDs are frequently employed in EKC separations as secondary phases. The separation
mechanism is simply defined by host–guest interactions once micellized surfactants are absent. The
enantiomeric separation of polychlorinated biphenyls using mixtures of several neutral and charged
CD derivatives is a fine example of the EKC with charged CDs impact in environmental analysis.42
Other examples include the use of a mixture of SB-β-CD and the neutral methyl-β-CD in borate
buffer for the analysis of 16 PAH in contaminated soils extracted by supercritical CO2 43 and the use
of SB-β-CD in the analysis of 25 chlorinated and substituted phenolic compounds (including the
11 priority phenols).44 A complexation model was used for investigating the effect of pH and CD
concentration on the electrophoretic mobility. The latter method was applicable for quantifying the
level of pentachlorophenol in contaminated soil samples.
31.2.2.2.2 MEEKC
Microemulsion electrokinetic chromatography (MEEKC) is a relatively new technique that accom-
plishes electrokinetic separations using buffers containing surfactant coated oil droplets (Chapter 4
by Altria and coworkers presents further details on the MEEKC technique). The potential of MEEKC
for the separation of priority endocrine disrupting compounds in industrial and domestic wastewater
treatment effluents and sludges has been investigated.45 Using reverse migrating microemulsion, that
is, negative polarity at the electrode inlet and a pH 2.8 phosphate buffer containing octane, butanol,
and SDS, further modified by propanol, the separation of the breakdown products of alkylphenolic
detergents, a few synthetic estrogens, and the plastic monomer bisphenol A was demonstrated.
31.2.2.2.3 Calixarene EKC

Calixarenes are conical-shaped macrocyclic oligomers whose inner cavity can accommodate several
guest molecules and, therefore, can serve as a viable option to replace CDs in EKC separations.
Environmental examples of the use of calixarene as EKC secondary phases include the separation
of chlorinated phenols, benzenediols, and toluidines.46 Sulfonated calixarenes have also been used
as chromophores in the indirect detection of aliphatic amines.47
31.2.2.2.4 Suspension EKC

The use of suspended chromatographic silica-based particles in SDS buffers was first demonstrated
as secondary phases in the EKC separations of phenol derivatives.48 Applications of polymer and
silica-based particles to several other pollutants have been reviewed recently.49
31.2.2.2.5 Ion-exchange EKC

The use of polymer ion additives such as poly(diallyldimethylammonium chloride) (PDDAC) and
(diethylamino)ethyldextran (DEAE-dextran) has been successfully proposed to promote dynamic
interactions with certain moderately hydrophobic solutes such as naphthalenesulfonate and naph-
thalenedisulfonate isomers.50 The fundamentals and scope of ion-exchange EKC with ionic polymers
for the separation of small ions of potential importance in the environmental context have been
reviewed.51
31.3 SENSITIVITY ENHANCEMENT STRATEGIES

The small pathlength defined by the capillary internal diameter coupled with the need of a small
injection volume to preserve the high-resolution features of CE place a strong demand on the detec-
tion capability. This is especially problematic for absorbance detectors (LOD in the order of 10−5
to 10−6 mol L−1 ), considered unsuitable for many environmental applications, where trace level
occurrence or matrix complexity issues are generally of concern.
Several approaches to enhance sensitivity in CE have been devised: some are based on sam-
ple preconcentration procedures, others rely on special cell geometries and alternative detectors.
Both areas have been subjects of intense investigation as described in Chapters 9 by Sweedler and
coworkers, 13 by Burgi and Giordano, 14 by Ewing and coworkers, and 27 by Weber.
31.3.1 PRECONCENTRATION SCHEMES

The sensitivity enhancement strategies based on sample manipulation are further subdivided into
offline and online procedures (Chapter 13 by Burgi and Giordano). Liquid–liquid extraction (LLE)
with a variety of solvents of selected properties and solid-phase extraction (SPE) (Chapter 27
by Weber) with a large assortment of chromatographic stationary phases of distinct chemistries
are among the most often performed offline strategies for environmental applications as reviewed
recently.4,5 An interesting example of homogeneous LLE is the pH-induced phase separation of water
samples (rain, river, and spring water) treated with perfluorooctanic acetate (PFOA− ) surfactant in
different water-miscible solvents for the preconcentration of five PAH.52 At the optimal conditions
(above 99% recovery), a 40 mL sample aliquot-containing THF and PFOA− was treated with HCl
to separate 30 µL of sedimented phase. The sedimented liquid phase was then mixed with 30 µL
DMSO before analysis. Limits of quantification (LOQ) in the range 10−10 to 10−9 mol L−1 and
enrichment factors up to 125,000-fold were obtained. An example of SPE is the use of C18 -bonded
silica and polystyrene–divinylbenzene (PS–DVB) disks for the enrichment of four triazines and three
degradation products of atrazine in drinking and well water samples with recoveries better than 93%
and LOD in the range of 0.02–0.30 mg L−1 .53 Several other LLE and SPE procedures are compiled
in the application tables (Section 31.4).
Online preconcentration strategies involve the insertion of large sample volumes in the capillary
and can generally be classified into two categories. One category involves the criterious manipulation
of the analyte electrophoretic velocity. A collection of strategies grouped by the name of stacking
and transient isotachophoresis (tITP) belong to this category. Field-amplified sample stacking of
arsenic species in environmental reference materials54 and organonitrogen pesticides in drainage
water55 as well as the tITP preconcentration of iodide in seawater samples56 are fine examples of
these procedures.
The other category explores the analyte ability of interacting with secondary phases. Stacking
of micelles and sweeping are representative of this group (fundamental aspects of these strategies
were introduced in Chapter 3 by Burgi and Giordano). Stacking of micelles was employed to pre-
concentrate dioxin-related compounds57 and in the analysis of nonsteroidal anti-inflammatory drugs
in mineral water,58 whereas sweeping was employed for the online concentration of bisphenol A and
alkylphenols.59 Stacking of micelles and sweeping were contrasted in the analysis of phenylurea
herbicides in tap and pond water60 and in the multiclass pesticide analysis of drinking water.61
Combinations of online procedures are often employed in environmental applications. Examples
include the analysis of trace metal ions in factory wastewater by a combination of cation-selective
exhaustive injection (CSEI) (a form of stacking, CZE format) and sweeping.62 In this latter case,
sweeping was promoted by dynamic complexation with EDTA, used as carrier. The same CSEI prin-
ciple has also been applied to the analysis of quaternary ammonium herbicides63 and environmentally
relevant aromatic amines64 in water, both followed by sweeping with SDS.
Hyphenation of automatic continuous flow systems (such as SPE, dialysis, gas diffusion, evapo-
ration, direct leaching, etc.) to CE and the coupling of automatic sample preparation devices into
commercial CE equipments have been devised as a means to simplification and miniaturization
of analytical procedures. An automatic online SPE device for the multiresidue extraction of seven
pesticides has been described.65 Four river samples were spiked with the test mixture at three different
levels presenting recoveries from 90% to 114%.
An elegant example of SPE method based on ion-exchange retention was used for inline precon-
centration of inorganic anions.66 A single capillary containing a preconcentration zone (adsorbed
layer of cationic latex particles) and a separation zone (fused-silica modified by adsorption of a
cationic polymer) was used. Analytes were retained in the preconcentration zone and eluted isota-
chophoretically into the separation zone by means of an eluotropic gradient. This approach was used
to determine nitrate in Antarctic ice cores at the 2.2–11.6 µg L−1 level.
31.3.2 ALTERNATIVE DETECTION SCHEMES

For the absorbance-based detectors, alternative cell geometries have been designed to extend the
optical pathlength such as the Z-shaped cells, bubble cells, rectangular cells, and multireflection
capillaries.6 Although manufacturing and manipulation of these special cells can pose a few opera-
tional problems, sensitivity enhancement of tenths of fold have been reported. The determination of
the herbicide metribuzin and its major conversion products in soil is an example of the use of bubble
cells,67 whereas the UV-detection of derivatized carbonyl compounds in rain is an example of the
use of Z-shaped cells68 in environmental analysis.
Alternative detection schemes for environmental analysis such as those based on laser-induced
fluorescence (LIF), chemiluminescence, conductivity, amperometry, and mass spectrometry (MS)
have been reported.
31.3.2.1 Fluorescence Detection

Some pollutants are natively fluorescent such as PAH and metabolites as inspected in biota.69
Others must be derivatized to fluoresce such as aniline species metabolized from pesticides70 and
low-molecular mass amines labeled as fluorescein isothiocyanate (FITC) derivatives for further
inspection of atmospheric aerosol samples reaching 50 pg mL−1 (10−9 mol L−1 ) level.71
An interesting variant for labeling procedures is the so-called in-capillary derivatization reac-
tions, which explores differences in the electrophoretic mobilities to merge distinct zones of analyte
and labeling reagent under the effect of an electric field. Molina and Silva72 explored the possibili-
ties of LIF detection for in-capillary derivatization of nitrogen compounds (herbicides and biogenic
amines).
Indirect fluorescence detection is another possibility for nonfluorescent compounds. In this par-
ticular case, a fluorophore co-ion is part of the electrolyte system. The simultaneous determination
of cations (EDTA complexes) and anions in pond water by indirect fluorescence detection based on
the displacement of fluorescein is an example.73 The use of fluorescein at pH 9.5 as electrolyte was
also reported in the determination of anionic organophosphorus herbicides [glufosinate, glyphosate,
and its major metabolite aminomethylphosphonic acid (AMPA)] in groundwater,74 with LOD in the
µmol L−1 range.
31.3.2.2 Chemiluminescence Detection

Biogenic amines, labeled with N-(4-aminobutyl)-N-ethylisoluminol) (ABEI) in the presence of
N,N-disuccinimidyl carbonate (DSC) in methanolic medium, were determined in lake water sample
by an EKC method using online chemiluminescence detection.75 Parameters affecting the separation
process (pH, SDS concentration, applied voltage, and injection time) and chemiluminescence detec-
tion sensitivity [H2 O2 concentration in the buffer and the postcolumn reagents K3 Fe(CN)6 and NaOH
concentrations] were examined in detail. Detection limits in the 1.2 × 10−10 to 3.5 × 10−8 mol L−1
range were obtained.
31.3.2.3 Electrochemical Detection

Among the electrochemical detection schemes for CE, conductometric methods have gained great
visibility especially for the determination of the high-equivalent conductivity inorganic species.
Conductivity detection in CE is performed either by galvanic contact or in a contactless mode (CCD).
The galvanic contact mode is further subdivided into on-capillary detection without decoupling
and off-capillary, after grounding the separation applied voltage. Most contactless methods use
oscillometric techniques. Wet deposition in the 1983–2006 period and related atmospheric chemistry
in the metropolitan area of Sao Paulo, Brazil, using CZE-CCD methods has been investigated.76,77
Electrolytes composed of MES-His (pH 6.2) modified by 18-crown-6 ether for the cation analysis
(ammonium, magnesium, sodium, calcium, and potassium), and cetyltrimethylammonium bromide
(CTAB) for the anion analysis (sulfate, nitrate, chloride, acetate, and formate) were employed. Limits
of detection spanned from 10 to 100 µg L−1 . An interesting approach employing dual-opposite end
injection and CZE-CCD to determine cations and anions in the same run has been devised.78 A single
electrolyte containing all necessary modifiers was employed (MES/His containing 18-crown-6 ether
and CTAB) allowing simultaneous detection of 13 anions and cations in less than 6 min with detection
limits in the range of 7–30 µg L−1 (except fluoride and phosphate). The method was applied to the
inspection of ions in rain, surface, and drainage water and the results were contrasted with ion
chromatography showing excellent agreement.
Despite the high sensitivity of amperometric detection, its use in environmental applications
has not been exploited accordingly. Nevertheless, a few examples include the MEKC analysis of
herbicides in tap water79 and the CZE and MEKC analysis of aromatic amines in contaminated water
samples.80
Electrochemical detection of 19 chlorophenols in river samples was performed using a graphite-
epoxy working electrode at a potential of 800 mV versus Ag/AgCl.81 A palladium metal union was
used to decouple the electric field from the electrochemical cell and a compensating pressure was
applied during the EKC analysis. Detection limits from one to three orders of magnitude lower than
UV-detection were achieved (10 µg L−1 ).
31.3.2.4 Mass Spectrometry Detection

Mass spectrometry (MS) is one of the most powerful detection techniques employed in environ-
mental screening due to the intrinsic structural information and high sensitivity it provides. With
the advent of effective interfaces, soft ionization modes (electrospray ionization, ESI, and atmo-
spheric pressure chemical ionization, APCI) and increased affordability of CE-MS instruments, the
hyphenated technique has become widespread.
In CE-MS, the use of NACE or mixed aqueous-organic solvent systems is advantageous as non-
volatile buffers are usually avoided. CZE-MS with volatile buffers has been employed in a variety
of environmental applications such as the inspection of drug residues in river water (pH 5.1 ammo-
nium acetate),82 detection of explosives (nitroaromatic and cyclic nitramine compounds) in soil and
marine sediment samples (ammonium acetate at pH 6.9 modified by SB-β-CD),83 and the identifi-
cation of reactive vinylsulfone and chlorotriazine dyes in spent dyebaths and municipal wastewater
treatment plant receiving dyehouse effluents (ammonium acetate at pH 9 modified by 40% ACN).84
Other examples of volatile electrolyte systems include the use of a pH 3.0 formic acid/ammonium
formate system in 50% methanol for the CZE-ESI (sheath liquid: 9:1 MeOH:20 mmol L−1 acetic
acid)-MS (ion trap) analysis of quaternary ammonium herbicides in contaminated irrigation water
and spiked mineral water,85 and pH 9.1 ammonium acetate/ammonium hydroxide/MeOH or pH 11
ammonium hydroxide systems by CZE-ESI (50% isopropanol)-MS (ion trap) for the inspection of
methoxy phenols and aromatic acids in biomass burning aerosol samples.86 When concentrations
are rather high CE-MS can still be used as a means of confirming the pollutant presence. An example
of such approach includes the determination of LAS in wastewater from treatment plants and coastal
waters of Cadiz (Spain) receiving untreated domestic wastewaters by CZE-UV (ammonium acetate
at pH 5.6 containing 30% isopropanol) followed by confirmation of the [M-H]− ions by ESI (sheath
liquid: 80% isopropanol and 0.1% ammonia)-MS (quadrupole).87
Particularly in MEKC separations, the presence of high amounts of SDS in the buffer is detrimen-
tal causing low ionization efficiency. A number of approaches have been presented to overcome this
shortcoming such as the use of high-molecular mass surfactants and anodically migrating micelles
as well as partial filling techniques. Examples of this latter approach have been compiled within
the pesticide analysis context.88,89 More recently, technological improvements in the interfacing
systems were shown to be more tolerant to the usage of nonvolatile buffers as demonstrated by the
MEKC-ESI-MS separation of triazines90 and the MEKC-APCI-MS analysis of aromatic amines and
alkylphthalates.91
31.4 REPRESENTATIVE APPLICATIONS

The literature comprises a larger number of methodologies employing electromigration princi-
ples, diverse preconcentration strategies, and a variety of detection schemes for pollutant standards
combined as test mixtures and for real samples. Tables 31.1 through 31.9 compile representative
applications of CZE and EKC methods to the most important classes of pollutants in different
environmental compartments.
31.4.1 PESTICIDES
Pesticide is a generic term used to describe compounds employed in the control, prevention, and
elimination of plagues that attack crops and herds. Pesticides are associated with their persistence
and toxicity and due to their widespread use in agriculture, these compounds are important sources
of environmental contamination.
922
TABLE 31.1
Selected Applications of Electroseparation Methods for Pesticides in Environmental Samples
Preconcentration
Procedure Optimal Detection
Compounds CE Mode Matrix (Recovery) Electrolyte (LOD) References
Multiclass pesticides
Amitrole, atrazin-2-hydroxy in 8 min CZE WATER — 20 mmol L−1 phosphate UV, 200 nm 98
River water buffer (pH 3.2) 90–120 µg L−1
Drinking water
Acifluorfen, pentachlorophenol, CZE WATER Spiked sample 5 mmol L−1 NH4Ac, MS, single quadrupole 104
2,4-DB acid, dinoseb, 2,4,5-TP, Drinking water 91–124% 40% isopropanol (pH 10.0) 8–250 µg L−1
2,4,5-T acid, MCPP acid,
dichloroprop, MCPA acid,
bentazon, 2,4-D, picloram,
3,5-dichlorobenzoic acid,
chloramben, dicamba,
4-nitrophenol in 40 min
Atrazine, simazine, paraquat, diquat EKC WATER SPE 10 mmol L−1 tetraborate, 220, 254, 300 nm, 150 µm 26
in 3 min Well water C18 25 mmol L−1 SDS, extended optical path
80–95% 40 mmol L−1 perchlorate, capillary
15% ACN (pH 9.3) 0.6–1.9 µg L−1
Carbendazim, simazine, atrazine, EKC WATER SW, SRMM, SRW 20 mmol L−1 phosphate, UV, 220 nm 61
propazine, ametryn, diuron, linuron, Drinking water SPE 25 mmol L−1 SDS, 0.1 µg L−1
carbaryl, propoxur in 6 min (and carrots) C18 , NH2 10% MeOH (pH 2.5)
Fenuron, simazine, atrazine, EKC WATER online SPE 10 mmol L−1 phosphate, UV, 226 nm 65
carbaryl, ametryn, prometryn, River water C18 60 mmol L−1 SDS, 0.01–0.03 µg mL−1
terbutryn in 12 min 90–114% 8% ACN (pH 9.5)
Simazine, aziprotryne, hexazinone, EKC WATER SPE 20 mmol L−1 borate, UV, 215, 240 nm 124
diuron in 10 min River water DVN-VP 8.5 mmol L−1 SDS 0.02–0.17 ng L−1
Well water 41–109% (pH 8.30)
Pyrethroids
Pyrethrin esters (pyrethrin, cinerin EKC Pyrethrum extract — 25 mmol L−1 Tris, UV, 254 nm 109
and jasmolin, I and II) in 25 min 30 mmol L−1 SDS, 1.1–14.1 mg L−1
25% ACN (pH 9)

Triazines
Hydroxyatrazine, CZE WATER SPE 80 mmol L−1 acetate, UV, bubble cell, 205 nm 100
desisopropylhydroxyatrazine, Tap water Amberchrom 62 mmol L−1 phosphate
desethylhydroxyatrazine, ameline in resins 20–95% (pH 4.7)
ca. 40 min 4% for ammeline
Atrazine, simazine, propazine, CZE WATER SPE 7.5 mmol L−1 HClO4 , UV, 214 nm 105
ametryne, prometryne, terbutryne in River water Oasis HLB 17 mmol L−1 SDS in 50:50 2.1–3.4 µg L−1 (CE-UV)
ca. 9.3 min Drinking water 83–114% MeOH–ACN mixture 0.01–0.05 µg L−1
(nonaqueous) (SPE-CE-UV)
Desethylatrazin-2-hydroxy, simazine, EKC WATER SPE 24 mmol L−1 borate, UV, 214 nm 16
prometon, atrazine, simetryn, Ground water C18 18 mmol L−1 phosphate, 0.05 µg L−1
ametryn, propazine, prometryn, 52–87% 25 mmol L−1 SDS,
trietazine, terbutylazine, terbutryn 5% 1-propanol (pH 9.5)
in 30 min
Atrazine, simazine, propazine, EKC WATER SPE 10 mmol L−1 borate, 220 nm, 200 µm extended 53
prometryn, hydroxyatrazine, Drinking water 2 PS-DVB disks 60 mmol L−1 SDS, optical path capillary
deisopropylatrazine, deethylatrazine Well water 73.5–102.4% 20% MeOH (pH 9.2) 0.02–0.30 mg L−1
Separation Strategies for Environmental Analysis
in 7 min
Metribuzin, bromacil, terbacil, EKC WATER SPE 12.5 mmol L−1 borate, UV, 210 nm 55
hexazinone, triadimefon, DEET Drainage water C18 disks 50 mmol L−1 SDS (pH 9.0) 0.8 µg L−1
in 8 min from highway 85%
Hydroxyatrazine, EKC WATER SPE 30 mmol L−1 borate, UV, 210 nm 99
hydroxyterbutylazine, Tap water LiChrolut EN 30 mmol L−1 SDS (pH 9.3) 0.2–0.5 mg L−1 (CE-UV)
deethylhydroxyatrazine, Ground water 43.8–93.4% 0.1–0.25 µg L−1
deisopropylhydroxyatrazine ameline is not SPE-CE-UV
ameline in 9.5 min recovered
Atrazine, desethylatrazine, EKC HUMIC ACID — 15 mmol L−1 tetraborate, UV, 210 nm 119
desisopropylatrazine, Solutions 60 mol L−1 SDS, 2–4 mg L−1
hydroxyatrazine, chloro-, hydroxy- 10% MeOH (pH 9.3)
degradation products in 29 min
Hexazinone, metabolites C, A1, E, EKC WATER SPE 50 mmol L−1 SDS, 220, 225, 230, 247 nm 120
B, D in 30 min Ground water graphitized 12 mmol L−1 phosphate, <0.38 mg L−1
nonporous carbon 10 mmol L−1 borate,
30–120% 15% MeOH (pH 9.0)
Continued
923
TABLE 31.1
924
(Continued)
Preconcentration
Carbamates
Aminocarb, propoxur, carbofuran, EKC WATER SPE 45 mmol L−1 UV, 202, 214 nm 12
carbaryl, methiocarb, metabolites: River water PS-DVB borate/phosphate, 22–85 ng L−1
4-dimethylamino-3-methylphenol, Well water 82.2–104.8% 40 mmol L−1 SDS
2-isopropoxyphenol, 2,3-dihydro- Pond water (pH 8.0)
2,2-dimethyl-7-benzofuranol, 1-naphthol,
and 4-methylthio-3,5-xylenol in 20 min
Urea herbicides
Monuron, isoproturon, diuron in 16 min EKC WATER SW, SRMM, SRW 50 mmol L−1 SDS, UV, 244 nm 60
Tap water SPE 50 mmol L−1 tested z-shaped cell
Pond water C18 phosphoric acid, 1 µg L−1
15 mmol L−1 γ-CD
Chlorsulfuron, chlorimuron, metsulfuron EKC SOIL LLE-SPE 30 mmol L−1 borate, UV, 214 nm 114
in 20 min C18 80 mmol L−1 SDS, 10 µg L−1
>80% 14% MeOH,
20% isopropanol
(pH 7.0)
Monuron, linuron, diuron, isoproturon, EKC WATER SPE 4 mmol L−1 tetraborate, UV, 244 nm 121
monolinuron in 15 min Drinking water C18 12 mmol L−1 phosphate, 0.1 mg L−1
80.2–94.9% 30 mmol L−1 SDS
(pH 7)
Organophosphorus
Glufosinate, glyphosate, CZE WATER — 1 mmol L−1 fluorescein LIF, 488 nm 74
aminomethylphosphonic acid in ca. 8 min Ground water (pH 9.5 adjusted with 10 mW, argon-ion laser
sodium hydroxide) 0.6–1.7 µmol L−1
Glyphosate in 12 min CZE WATER — 5 mmol L−1 NH4Ac ESI-CNLSD 101
Lake water (pH 2.8 adjusted with 0.2 µg mL−1
acetic acid)
(CTAB rinsing before
analysis)
Glyphosate, glufosinate, CZE WATER Spiked sample 40 mmol L−1 NH4Ac ICP-MS 103
aminomethylphosphonic acid in 10 min River water 95.9–104.1% (pH 9.0) 0.11–0.19 mg L−1
Glyphosate, aminomethylphosphonic acid EKC WATER Strong 50 mmol L−1 phthalate, UV, indirect detection, 118
in ca. 4 min River water anion-exchange resin 0.5 mmol L−1 TTAB 254 nm
Tap water 50–99% (pH 7.5) 60–85 ng L−1
Quaternary ammonium herbicide

Paraquat, diquat, difenzoquat in ca. EKC WATER CSEI–SW 50 mmol L−1 phosphate, UV, 220 nm, 255 nm 63
17.5 min Tap water 80 mmol L−1 SDS, 0.075–1 µg L−1
20% ACN (pH 2.5)
Paraquat, diquat, difenzoquat, chlormequat CZE WATER — 200 mmol L−1 formic ESI-IT-MS 85
mepiquat in 17 min Contaminated acid–ammonium 0.5–2.5 mg L−1
Irrigation water formate buffer, (hydrodynamic
Mineral water 50% MeOH (pH 3.0) injection)
1–10 µg L−1
(electrokinetic injection)
Triazolopyrimidine sulfonanilide
Flumetsulam, florasulam, EKC SOIL SPE-FESI 11 mmol L−1 formic acid, UV, 205 nm 116
cloransulam-methyl, diclosulam, C18 16 mmol L−1 18–34 µg kg−1
metosulam in 6.5 min 50–84% (NH4 )2 CO3 ,
2.5 mmol L−1 α-CD,
0.00042% HDB (pH 7.6)
Chloroacetanilide herbicide
Metolachlor stereoisomers and two EKC WATER SPE 75 mmol L−1 borate, Diodo array detector 115
metabolites (ethane sulfonic acid, Ground water C18 5% (w/v) γ-CD, 5 mg L−1
oxanilic acid) in ca. 24.5 min >90% 20% MeOH (pH 9.2)
Triazinone herbicide
Metribuzin and its major conversion EKC SOIL Sonication 10 mmol L−1 NH4Ac UV, 220 nm, 260 nm 67
products, deaminometribuzin, SPE buffer, 19–23.4 µg kg−1
diketometribuzin, LiChrolut EN 100 mmol L−1 SDS
deaminodiketometribuzin in ca. 7 min 78.3–99.2% (pH 10)
Continued
925
926
TABLE 31.1
(Continued)
Preconcentration
Chlorophenoxycarboxylic acid
2,4-dichlorophenoxyoleic, EKC WATER SPE 10 mmol L−1 borate buffer, UV, 205 nm 117
2,4-dichlorophenoxypropionic, Tap water Diapak S16 2 mmol L−1 β-CD (pH 9.2) 0.5–1 µg L−1
2,4,5-trichlorophenoxyacetic, River water 12.6–84.7%
2,4-dichlorophenoxyacetic, Drinking water
phenoxyacetic, 2,4-dichlorophenol Well water
(product of their decomposition)
in ca. 11 min
Nicotinoids
Imidacloprid, metabolite EKC AIR SPE NH4 Cl/NH3 buffer at UV, 227, 270 nm 13
6-chloronicotinic acid in 6 min greenhouse air amberlite 15 mmol L−1 , 60 mmol L−1 0.71–1.18 ng L−1
85–92% SDS (pH 8.5)
FESI, field enhanced sample injection; SW, sweeping; SRMM, stacking with reverse migrating micelles; SRW, stacking with reverse migrating micelles and a water plug; SPE, solid-
phase extraction; C8 , octa-, C18 , octadecyl-, NH2 , amino-bonded silica; PS-DVB, poly(styrene-divinylbenzene); DVN-VP, poly(divinylbenzene-N-vinylpyrrolidone); SLM, supported liquid
membrane; LLE, liquid–liquid extraction; HDB, hexadimethrine bromide; CD, cyclodextrin; ESI-CNLSD, electrospray condensation nucleation light scattering detector; ICP-MS, inductively
coupled plasma-mass spectrometry; ESI-IT-MS, electrospray ionization ion trap-mass spectrometry; CSEI, cation-selective exhaustive injection; ASE, accelerated solvent extraction; TTAB,
tetradecylytrimethylammonium bromide; CTAB, cetyltrimethylammonium bromide; LIF, laser-induced fluorescence; MeOH, methanol; ACN, acetonitrile.
Pesticides encompass a large assortment of organic compounds, several of which are positional,
geometrical, and optical isomers, with differing degrees of ionization, polarity, and water solubility.
Many exhibit spectrochemical, electrochemical, or other functional properties suitable for detection.
Unless the pesticide class is explicitly charged, single- and multiclass pesticide analyses are usually
assessed by EKC modes. Several excellent review articles have covered the literature on pesticide
analysis by CE in the past 5 years.92–96
The analytical feasibility of electromigration methods for pesticides has been endorsed by funda-
mental studies with standards and analyses of spiked and real samples. An example of fundamental
study is the determinations of pKa (CZE: citrate/phosphate buffers and carbonate buffers covering
the pH range from 2.05 to 12.45) and pI (CIEF kit from Beckman Coulter) of 12 hydroxy-s-triazines
standards.97
The majority of CZE methodologies for pesticides employs simple buffers at varying
pH,70,98–100 volatile electrolytes101,102 (usually aiming at CZE-MS applications),103 solvent modi-
fied buffers,85,104,105 and electrolytes for special detector schemes (indirect fluorescence detection,
for instance).74 Table 31.1 compiles the details of a few CZE methodologies applied to pesticide
analyses, classifying the application by the pesticide class and the environmental compartment
assessed.
For the EKC mode, pesticide standards are either combined in multiresidue mixtures106–108 or
differentiated by classes: pyrethroids,109 s-triazines,14,90,99,110 carbamates,88,111 phenoxyacids,112
quaternary ammonium salts,63,113 phosphonic acids,72 and urea-derived pesticides.114 MEKC
methodologies include buffered systems containing SDS,99,107 cationic,14,110,111 or nonionic
surfactants72 and additives such as organic solvents63,106,113,114 and CDs for chiral separations.106
The use of poly-SUS for pyrethoids,109 DOSS for herbicide mixtures,108 and alkylglycoside chiral
surfactants for phenoxy acids112 have also been reported. Several examples of EKC methodologies
applied to single class12,13,16,53,55,60,63,67,109,114–121 and multiclass26,61,65,122–124 pesticide residues
in real samples are detailed in Table 31.1. Figure 31.1 illustrates the electromigration separation of
herbicides.
(a) (b)
12 4
AT
SM AT
SM
Absorbance (mAU)
8
Absorbance (mAU)
PQ 2 PQ
DQ
4
DQ 0
0
–2
0 2 4 6 0 1 2 3 4 5
Migration time (min) Migration time (min)
FIGURE 31.1 EKC separation of triazines and quats herbicides. (a) Standard solution and (b) spiked well
water. Electrolyte: 10 mmol L−1 sodium tetraborate, 25 mmol L−1 SDS, 15% ACN, and 40 mmol L−1 sodium
perchlorate. Other conditions: +30 kV; direct UV detection at 220 nm. Peak labels: simazine (SM), atrazine
(AT), paraquat (PQ), and diquat (DQ). (Modified from M.I. Acedo-Valenzuela et al., Anal. Chim. Acta, 519,
65–71, 2004. With permission.)
31.4.2 PAH, PCB, AND PCDD

Polynuclear aromatic hydrocarbons are listed by the U.S. Environmental Protection Agency (U.S.
EPA) and European Community as priority pollutants due to their mutagenicity and carcinogenicity.
PAHs can be formed during natural processes or emitted by anthropogenic activity.
PAHs are characterized by a multitude of structurally similar congeners with inherently neutral
and highly hydrophobic character. Therefore, they are not suitable to CZE analysis and as a result,
only EKC-based methodologies have been devised in the literature. PAHs have been used in test mix-
tures as model compounds to introduce a plethora of EKC approaches. MEKC separations include
diverse surfactant types (SDS,69,125 tetradecylsulfate,126 dodecylbenzenesulfonate,127 cationic
surfactants,128 bile salts129 ) as well as commonly used additives (urea,125,130 organic solvents,125–130
and neutral CDs).69,125 Other strategies invoke solvophobic association with ThxAP, THpAP,131 or
DOSS130 in buffered electrolytes as well as charge–transfer interactions with pyrylium salts in
nonaqueous media.132 The use of mixed CD systems (CDCD-EKC), containing both noncharged
and negatively charged CDs (SB-β-CD and carboxymethyl-β-CD), has also been demonstrated.133
The applicability of more elaborate secondary phases such as SUA134 and SUS135 oligomers, in
combination with neutral CDs,136 polymeric surfactants (11-acrylamidoundecanoate,137 sulfated
siloxanes138 ), starburst dendrimers,139 fullerenes,140 and silica-based particle suspensions141 have
all been demonstrated by PAH test mixtures.
Polychorinated biphenyl (PCB) mixtures were used extensively in the past as coolant fluids in
power transformers and capacitors. PCBs were widely used because of their higher stability, but
are persistent environmental contaminants due to careless disposal practices, leakage, or accidents.
Similar to PAHs, PCBs are neutral, highly hydrophobic, and present a large number of congeners,
thus demanding EKC methodologies.
The separation of PCB congeners has been addressed by SDS/neutral CD containing
electrolytes,142,143 modified by organic solvents144 or urea142,143 and mixtures of bile salts.145
CD-EKC and CDCD-EKC modes with a large assortment of ionic CDs and modifiers are often
employed for the chiral discrimination of PCB racemates.42,146 The use of polymeric surfactants such
as polysodium undecyl sulfate (poly-SUS), in acetonitrile147 and its valinate form (poly-D-SUV) in
combination with hydroxypropyl-γ-CD, methanol, and urea148 has also been reported.
The term “dioxin” comprises a group of 75 polychlorinated dibenzo-p-dioxin (PCDD), and
135 polychlorinated dibenzofuran (PCDF), whose effects on human health and the environment
include dermal toxicity, immunotoxicity, reproductive effects and teratogenicity, endocrine dis-
rupting effects, and carcinogenicity. PCDD and PCDF are produced as byproducts of a myriad of
processes, including bleaching of pulp, incineration of garbage, recycling of metals, and in the pro-
duction of common solvents. Fires of many kinds, including forest fires and those in incinerators,
also release dioxins into the environment. PCDD congeners have been successfully separated by
MEKC in SDS buffers,149 modified by urea and CDs.150,151
Table 31.2 compiles a few applications of EKC methodologies to PAH in real samples. For exam-
ple, soil,15,38,43 water,52,152 air,27 and other matrices.125 Figure 31.2 illustrates the electromigration
separation of PAH homologs. To date, no environmental applications of electromigration methods
for PCBs and PCDDs have been reported.
31.4.3 PHENOLS
Several phenols and their derivatives (chloro- and nitrophenols) are priority polluting substances
that are widespread in the environmental aquatic compartments. They originate as byproducts from
the coal and oil industry and also due to pesticide and drug decay.
Phenols are UV-absorbing compounds with weakly acidic OH groups (pKa ≈ 9), which, upon
dissociation at high pH electrolytes, generate anionic species, making them suitable for CZE anal-
ysis. However, when a large number of phenolic compounds and derivatives must be assessed
TABLE 31.2
Selected Applications of Electroseparation Methods for PAH in Environmental Samples
Preconcentration
7 PAH homologs in 10 min EKC SOIL LLE 8.5 mmol L−1 borate, UV, 280 nm 15
spiked cyclohexane 85 mmol L−1 SDS, 10 mg L−1
heath sand (48–90%) 50% ACN (pH 9.9)
16 PAH homologs in 35 min EKC AIR Air sample collected on 0.1 mol L−1 phosphate, UV, 214 nm 27
ambient air polyurethane foams 0.1 mol L−1 borate, 3–25 ng mL−1
50 mmol L−1 STDC,
30% acetone
13 PAH homologs in 15 min EKC WATER SPME 50 mmol L−1 NH4Ac, UV, 254 nm 36
aqueous standard Silica impregnated with 100 mmol L−1 THA+
mixture extracted polydimethylsiloxane in 100% MeOH
by SPME
16 PAH homologs in 21 min EKC SOIL ASE 20% THF, 0.00625 mol L−1 UV, 214 nm 38
contaminated 50:50 CH2 Cl2 -acetone, OSUA
with creosote 150◦ C, 2500 psi, 10 min
followed by HPGPC
11 PAH homologs in 14 min EKC SOIL SFE 25 mmol L−1 SB-β-CD, LIF, bubble cell, 325 nm, 43
wood preserving lot CO2 at 400 atm, 400 mL/min, 20 mmol L−1 M-β-CD, 2.5 mW HeCd laser
120◦ C, 20 min; sample 50 mmol L−1 borate 0.9–21.7 µg L−1
collected in CH2 Cl2 and
diluted in MeOH/water
5 PAH homologs in 30 min EKC WATER LLE 0.2 mol L−1 PFOS− , 50% UV, 280, 333 nm 52
rain water perfluoro surfactants in THF DMSO, 0.1 mol L−1 H3 PO4 10−10 to 10−9 mol L−1
river water (>99%)
spring water
5-Hydroxy-PAH homologs in 20 min EKC ISOPODS PROTEOLYSIS 30 mmol L−1 borate, LIF 125
hepatopancreas protease K in Tris buffer 60 mmol L−1 SDS, Nd-YAG, 266 nm
FLATFISH pH 9, 37◦ C, 18 h 12.5 mmol L−1 γ-CD
bile (pH 9.0)
ASE, accelerated solvent extraction; HPGPC, high-performance gel permeation chromatography; OSUA, oligomers of sodium undecylenic acid; SFE, supercritical fluid extraction; SDβCD,
929
sulfobutyl ether β-cyclodextrin; MβCD, methyl β-cyclodextrin; PFOS− , perfluorooctanic sulfate; DMSO, dimethyl sulfoxide; THA+ , tetrahexylammonium; SPME, solid-phase microextraction;
STDC, sodium taurodeoxycholate; LIF, laser-induced fluorescence; LLE, liquid–liquid extraction.
6
1
2 4 7 11
6 8 10 12 14 16 18 20 22 24
Retention time (min)
FIGURE 31.2 EKC separation of PAHs. Electrolyte: 10 mmol L−1 H3 PO4 and 70 mmol L−1 sodium
n-tetradecylsulfate in 75:25 methanol:H2 O. Other conditions: −20 kV, direct UV-detection at 254 nm. Peak
labels: benzo[a]perylene (1), perylene (2), benzo[a]anthracene (3), pyrene (4), 9-methylanthracene (5),
anthracene (6), fluorene (7), naphthalene (8), and benzophenone (9). (Modified from J. Li and J.S. Fritz,
Electrophoresis, 20, 84–91, 1999. With permission.)
simultaneously (chlorophenol congeners, for instance), EKC methodologies should be selected as

a means to increasing selectivity. Nevertheless, CZE has, by far, received more attention for real
sample applications.
Literature proposed CZE methods for phenols and derivatives using test mixtures based on
aqueous buffered systems (phosphate–borate153 and borate154 ), volatile electrolytes (ammonium
hydrogencarbonate,155 diethylmalonic acid/dimethylamine in isopropanol156 and l-cysteic acid,
3-amino-1-propanesulfonic acid, aminomethanesulfonic acid, and diethylmalonic acid157 ), and non-
aqueous media (ammonium acetate in ACN/acetic acid in MeOH;158 acetate, bromide, chloride, and
malonate in ACN; and diprotic acids/tetrabutylammonium hydroxide159 and maleate in MeOH160 ).
Environmental assessment using CZE methodologies includes pressurized hot water extrac-
tion of phenols from sea sand and soil,161 chlorophenols162 and priority phenols11 in wastewater,
nitrophenols163 and iodophenol164 in river waters, and methoxyphenols in biomass burning.86 Details
of these methodologies are compiled in Table 31.3.
The impact of different surfactants (SDS,22 DOSS,165 CTAB166 and hexadimethrine bromide,
HDB,167 bile salts29,30 ), nonionic168 and mixed micelles,22,169 and additives (neutral170 and anionic
CDs,44 tetraalkylammonium salts,171 organic solvents29,165 ) in EKC separations has been demon-
strated with phenol test mixtures. In addition, phenols have been chosen to introduce the applicability
of more exotic EKC secondary phases such as SDS modified by bovine serum albumin,172 water-
soluble calixarene,173 starburst dendrimers,174 cationic replaceable polymeric phases,175 ionenes,176
amphiphilic block copolymers,177 polyelectrolye complexes,178 and liposome-coated capillaries.179
The separation of phenols of environmental interest180 as well as the sources and transformations
of chlorophenols in the natural environment181 have been revised. Examples of the investigation of
phenols by EKC methodologies in aquatic systems,39,41,81 soil,44 and gas phase182 are compiled in
Table 31.3. Figure 31.3 illustrates the electromigration separation of phenols by both CZE and EKC
modes.
31.4.4 AMINES
Amines are organic bases that are usually present in biological materials (biogenic amines), processed
foods and beverages (of concern are the nitrosamines in fried bacon, smoked/cured meat and fish,
TABLE 31.3
Selected Applications of Electroseparation Methods for Phenols in Environmental Samples
Preconcentration
26 Priority phenols in 40 min CZE WATER SPE 150 mmol L−1 NH4Ac in 75:25 UV, 280 nm 11
Waste water PSDVB N-methylformamide-ACN 28–629 µg L−1
spiked sample mixture (non-aqueous)
63.3–113.4%
11 Priority phenols in 12 min EKC WATER SPE 50 mmol L−1 phosphate, UV, 214 nm 39
River water spiked sample 25 mmol L−1 28.1–215 nmol L−1
Sea water 74.2–106% tetraborate-phosphate,
7.5 mmol L−1 DBTHS (pH 7)
10 Nitrophenols in 20 min EKC WATER ITP 50 mmol L−1 glycine, UV, 254 nm 41
Rain water spiked sample 0.2% m-HEC, 2.5% PVP 19–80 µg L−1
Tap water (pH 9.1)
Process water
25 Phenols in 20 min EKC SOIL — 50 mmol L−1 phosphate, UV, 214 nm 44
Certified soil 1 mmol L−1 SB-β-CD (pH 7.5) 0.05–0.33 mg L−1
Reference
standard
19 Chlorophenols in ca. 35 min EKC WATER SPE 50 mmol L−1 ACES, ELECTROCHEMICAL 81
River water PSDVB 22 mmol L−1 SDS (pH 6.1) graphite-epoxy electrode
81–116% versus Ag/AgCl
0.07–0.2 µg L−1
12 Substituted methoxy phenols and CZE AEROSOL Berner type impactor (1) 20 mmol L−1 NH4Ac, 10% ESI-IT-MS 86
aromatic acids in 9.5 min Biomass burning filter (140–420 nm, MeOH (pH 9.1) (1) 0.1–1.0 µmol L−1
Aerosol 50% cut off) (2) 1 mol L−1 NH4 OH (pH 11) (2) 0.3–0.7 µmol L−1
Continued
931
932
TABLE 31.3
(Continued)
Preconcentration
Phenol, 3-methylphenol, CZE SOIL PHWE 30 mmol L−1 CHES (pH 9.7) UV, 220 nm 161
4-chloro-3-methylphenol, Soil and sea 77.7–105.2% 270–410 µg L−1
3,4-dichlorophenol in 6 min sand mixture (soil sample)
Sea sand 77.7–98.4%
(sea sand sample)
17 Chlorophenols in 15 min CZE WATER — 10 mmol L−1 phosphate buffer, UV, 214 nm 162
Waste water 40% acetone (pH 8.23) —
o-Nitrophenol, m-nitrophenol, CZE WATER Spiked sample 20 mmol L−1 borate-carbonate, UV, 191 nm 163
p-nitrophenol in 11 min River water 95.6–100.9% 10% MeOH (pH 9.4) 10.2–40.6 µmol L−1
2-Iodophenol, 4-iodophenol, CZE WATER SPME 20 mmol L−1 CAPS (pH 11.0) ICP-MS 164
2,4,6-triiodophenol in 6.6 min River water CAR-PDMS 2.4–3.9 µg L−1
ca. 100% (CE-ICP-MS)
0.03–0.04 µg L−1
(SPME-CE-ICP-MS)
12 Chloro and nitrophenols in 10 min EKC AIR Loop-supported liquid 25 mmol L−1 , borate UV, 205 nm 182
film 10 mmol L−1 phosphate, 3.5–17 µg L−1
10 mmol L−1 SDS (8.86)
DBTHS, disodium 5,13-bis(dodecyloxymethyl-4,7,11,14-tetraoxa-1,17-heptadecanedisulfonate; ITP, isotachophoresis; m-HEC, methyl-hydroxyethylcellulose; PVP, polyvinylpirrolidone;

PSDVB, polystyrene-divinylbenzene copolymer; SB-β-CD, sulfobutylether-β-cyclodextrin; CAR-PDMS, carboxen-poly(dimethylsiloxane); ESI-IT-MS, electrospray ionization ion trap-mass
spectrometry; CHES, 2-(N-cyclohexylamino)-ethanesulphonic acid; CAPS, 3-(cyclohexylamino)-1-propanesulfonic acid; PHWE, pressurized hot water extraction; ICP-MS, inductively coupled
plasma-mass spectrometry; SPME, solid-phase microextraction; SPE, solid-phase extraction.
(a)
5
3
5
4
9 7
1 6
3
mAU
8
2 4 2
EOF 10
11
1
0
2 3 4 5
(b)
1 3 5
2 4 6 9
8
7 10
11
5
6.0 7.0 8.0 9.0 10.0 11.0 12.0

Time (min)
FIGURE 31.3 (a) CZE and (b) EKC separation of EPA phenols. Electrolytes: (a) 20 mmol L−1 NaH2 BO3 at
pH 10.00; (b) 20 mmol L−1 phosphate, 8% 2-butanol, 0.001% HDB, pH 11.95. Other conditions: (a) +20 kV;
direct UV-detection at 254 nm; (b) +25 kV; direct UV-detection at 210 nm. Peak labels: (a) 2,4-dinitrophenol (1),
2-methyl-4,6-dinitrophenol (2), pentachlorophenol (3), 2,4,6-trichlorophenol (4), 4-nitrophenol (5),
2-nitrophenol (6), 2,4-dichlorophenol (7), 2-chlorophenol (8), 4-chloro-3-methylphenol (9), phenol (10), and
2,4-dimethylphenol (11). (b) 2-nitrophenol (1), 2-chlorophenol (2), 2,4,6-trichlorophenol (3), phenol (4),
4-nitrophenol (5), 2,4-dinitrophenol (6), 4-chloro-3-methylphenol (7), 2,4-dichlorophenol (8), 2-methyl-
4,6-dinitrophenol (9), 2,4-dimethylphenol (10), and pentachlorophenol (11). (Modified from (a) I. Canals et al.,
Anal. Chim. Acta, 458, 355–366, 2002 and (b) P. Kubáň et al., J. Chromatogr. A, 912, 163–170, 2001. With
permission.)
canned sausage, etc.), as well as environmental samples. Several amines are strongly toxic and
suspected carcinogens. Major sources of amines in the environment include effluents or byproducts
of several chemical industry sectors such as oil refining, synthetic polymers, adhesives, rubber tyre
manufacturing, leather tanning, pharmaceuticals, pesticide production, and explosives.
31.4.4.1 Aliphatic Amines

The analysis of small aliphatic amines by electromigration methods faces a few challenges. Although
readily protonated at low pH electrolytes, the lack of chromophore precludes direct UV-detection in
both CZE and EKC mode unless derivatizing schemes are devised or alternative detectors are selected.
A further complication of the EKC mode derives from the hydrophilicity and polar characteristic of
aliphatic amines, which usually impair strong interaction with commonly used micellized surfactants,
demanding alternative secondary phases or more elaborate electrolytes.
Indirect UV-detection is a feasible option and particularly interesting for the CZE determination
of primary to quarternary amines without derivatization. Imidazole, N-ethylbenzylamine, and ben-
zyltriethylammonium chloride have been proposed as electrolyte chromophoric constituents.183 The
indirect UV-detection of aliphatic amines in ambient air has been performed by CZE in imidazole-
based buffer modified by ethanol and EDTA.184 By replacing imidazole with ammonium, the
electrolyte became applicable to MS detection (these details and other CZE methodologies are
reported in Table 31.4).
EKC separations of aliphatic amines include electrolyte systems composed of several surfac-
tants (SDS,185 cholate,186 Brij 3572 ) modified by certain additives (urea,185 neutral CDs,185 organic
solvents75,186 ), mixed CDs,187 and more unusual secondary phases [resorcarene-octacarboxylic
acid,188 calixarene,47 poly(sodium 4-styrenesulfonate), PSSS189 ]. Derivatization is performed when
UV (o-phthaldialdehyde, OPA,185,187 as derivatizing agents) or LIF detection [3-(2-furoyl)quinoline-
2-carboxaldehyde,186 5-(4,6-dichloro-s-triazin-2-ylamino)fluorescein (DTAF)72 as labeling agents]
is designed.
Examples of the EKC determination of aliphatic amines in water compartments75 and
atmospheric aerosol samples71 are compiled in Table 31.4.
31.4.4.2 Aromatic Amines

Direct UV spectrophotometric detection of aromatic amines from azo dye reduction in textile industry
wastewaters190 and electrophoretic methods for biogenic and aromatic amines191 have been the
subject of recent reviews. In addition, environmental applications of mutagenic heterocyclic amines
have recently been revised.192
CZE analysis of aromatic amines include the use of simple buffers at varying pH (acetate,80
phosphate,10,193 and borate70 ) with direct UV,10 electrochemical,80 or fluorescence70,80,193 detection.
EKC methodologies for aromatic amines are based on electrolytes containing SDS,64,91,194,195
mixtures of SDS and nonionic surfactants,196 and bile salts29 as well as modifiers (tetraalkyl-
ammonium salts,195 urea,64 organic solvents,29,64,91,194,196 and CDs197 ). Combinations of anionic
soluble polymers198 or crown ether199 and CDs have also been reported. Despite being chromophoric,
alternative detectors for aromatic amines have been proposed (CE-ESI-MS,91 electrochemical,80 and
fluorescence80 detection).
Examples of EKC methodologies for aromatic amines in water21,35,37,80 are compiled in
Table 31.4. Figure 31.4 illustrates the electromigration separation of aliphatic and aromatic amines.
31.4.5 CARBONYLS
Low-molecular mass carbonyls are among the most abundant and ubiquitous volatile organic com-
pounds in the atmosphere. They are produced from industrial activity and incomplete combustion of
fossil fuels and biomass. Many aldehydes are also emitted indoors (plastic, foam insulation, lacquers,
etc.). As a source of free radicals, aldehydes play an important role in the ozone formation, in urban
smog events, as well as in the photochemistry of the unpolluted troposphere. Aldehydes are recog-
nized irritants of the eye and respiratory tract, and often, carcinogenic and mutagenic characteristics
are also attributed to them.
To implement CZE and EKC methodologies for the determination of carbonyl compounds,
structure-related issues must be addressed. Aldehydes and ketones of environmental importance
are essentially neutral molecules, interact poorly with micellar phases, and present no chromophoric
moieties. Two methodological approaches have been presented: either generation of charged adducts
or generation of large UV-absorbing or fluorescence derivatives. In the former, the charged adducts
can be detected directly (if aromatic) or indirectly (if aliphatic). The separation of anionic aldehyde-
bisulfite adducts is a fine example of the first approach.25 The second approach comprises the
derivatization of the molecules to generate either neutral or charged UV-absorbing or fluorescent
TABLE 31.4
Selected Applications of Electroseparation Methods for Amines in Environmental Samples
Preconcentration
Aliphatic amines
Methylamine, dimethylamine, EKC AIR FITC derivatization 20 mmol L−1 borate, LIF, 488 nm Ar laser, 520 nm 71
diethylamine, dipropylamine, piperidine, Particulate aerosol 20% acetone, 5 mmol L−1 emission 10−9 mol L−1
pyrrolidine, morpholine in 10 min DM-β-CD
Diaminopropane, putrescine, cadaverine, EKC WATER ABEI-DSC 10 mmol L−1 borate, Chemiluminescence, postcolumn 75
diaminohexane in 7.5 min Lake water derivatization 100 mmol L−1 H2 O2 , reagents: 3 mmol L−1
92.8–106.8% 80 mmol L−1 SDS (pH 9.3) K3 Fe(CN)6 , 0.8 mol L−1
NaOH.
(3.5–12) × 10−8 mol L−1
Methylamine, ethylamine, propylamine, CZE AIR Quartz filter — SPE 20 mmol L−1 imidazole, UV, indirect detection, 214 nm 184
butylamine, dimethylamine, diethylamine, Metal working poly(acrylate- 4 mmol L−1 EDTA, 2 µg mL−1
cadaverine, putrescin, spermidine in Fluid aerosols and methacrylate) 25% ethanol (pH 2.5 with
9.5 min ambient air copolymer acetic acid)
95–99%
Aromatic amines
21 Aromatic amines in 25 min CZE WATER LLE/SPE 50 mmol L−1 NaH2 PO4 , UV, 280 nm 10
SOIL methylene chloride/ 7 mmol L−1 0.06–1.8 mg L−1
20:80 (w/w) keto 1,3-diaminopropane,
derivatized– (pH 2.35 with H3 PO4 )
underivatized PSDVB
47–97%
2-Toluidine, 4-toluidine, 1-naphthylamine, EKC WATER SPE/LLE 50 mmol L−1 UV, 214 nm 21
2-naphthylamine, 2-aminobiphenyl, Waste water (cation exchange phosphate—adjusted to 0.263–9.525 µg mL−1
4-aminobiphenyl, resin; pH 8.5 with 100 mmol L−1
2-methoxy-5-methylaniline, tertiary butyl methyl borate, 300 mmol L−1 SDS,
4-methoxy-2-methylaniline, ether) 10 mmol L−1 cholic acid,
2-chloroaniline, 4-chloroaniline, 78.4–94.4% 10 mmol L−1 Tween 80
2,4,5-trimethylaniline,
2,4,6-trimethylaniline, 2-anisidine,
4-anisidine, 2,4-xylidine, 2,6-xylidine in
50 min
935
Continued
936
TABLE 31.4
(Continued)
Preconcentration
4-Chloroaniline, 4-bromoaniline, EKC WATER FMOC derivatization 35 mmol L−1 DOSS, UV, 214 nm 35
3,4-dichloroaniline, 3-chloroaniline, lake water (anilines) 8 mmol L−1 borate, 40% ACN LIF, 266 nm solid-state
3-chloro-4-methylaniline, (pH 8.5) UV laser, 310 nm emission
4-isopropylaniline, aniline, 5.7 × 10−8 to
3-methylaniline, acetophenone, 4.9 × 10−7 mol L−1
propiophenone, butyrophenone,
valerophenone, hexanophenone,
heptanophenone in 26 min
4-Chloroaniline, 4-bromoaniline, EKC WATER FITC derivatization 400 mmol L−1 borate, LIF, 488 nm Ar laser, 520 nm 37
3,4-dichloroaniline, 3-chloroaniline, Tap water 40 mmol L−1 OG (pH 9.0) emission 10−10 mol L−1
3-chloro-4-methylaniline, Lake water
4-isopropylaniline, 3-methylaniline,
aniline in 15 min
p-Toluidine, aniline, p-ethoxyaniline, CZE WATER FITC derivatization 15 mmol L−1 borate (pH 9.5) Multiphoton-excited 70
p-phenylenediamine in 4 min Lake water fluorescence detection, diode
laser, 808 nm
1.25–2.60 µmol L−1
1,3-Phenylenediamine, 2-methoxy EKC WATER Fluorescamine 5 mmol L−1 tetraborate, Fluorescence, mercury–xenon 80
aniline, 4-ethoxyaniline, Surface water derivatization 4.5 mmol L−1 boric acid, lamp, 495 nm emission
4,4 -diaminobiphenyl, near textile 20 mmol L−1 SDS (pH 9) 1 µg L−1
2-methylaniline, 2,4- dimethylaniline, and leather
2-ethylaniline, 2,6-dimethylaniline in industries
4 min
ABEI, N-(4-aminobutyl)-N-ethylisoluminol); DSC, N,N-disuccinimidyl carbonate; FITC, fluorescein isothiocyanate; OG, n-octylglucopyranoside; FMOC, 9-fluoroenylmethyl chloroformate;
DOSS, dioctyl sulfosuccinate; EDTA, ethylenediaminetetraacetic acid; PSDVB, polystyrene-divinylbenzene copolymer; DMβCD, di-methyl β-cyclodextrin; SPE, solid-phase extraction; LLE,
liquid–liquid extraction; LIF, laser-induced fluorescence.
(a)
4 5
6
78
12
1 2 9 10
3 11 14 1617 19
18
13 15
4 5 6 7 8 9 10
(b) 3 78
10
12 5 6 9
1112
4 13
25
18
14
16 20
mAU 15 17
19 21
–2.5
0 10 20 30
Time (min)
FIGURE 31.4 Separation of (a) aliphatic and (b) aromatic amines. Electrolytes: (a) 20 mmol L−1
imidazole-acetate at pH 3.5, 4 mmol L−1 EDTA and 25% ethanol; (b) 50 mmol L−1 phosphate buffer
at pH 2.35, 7 mmol L−1 1,3-diaminopropane. Other conditions: (a) +20 kV; indirect UV-detection at
214 nm; (b) +30 kV applied voltage; direct UV-detection at 280 nm. Peak labels: (a): methylamine (1),
dimethylamine (2), Na+ (3), ethylamine (4), cadaverine (5), putrescine (6), propylamine (7), methanol-
amine (8), morpholine (9), diethylamine (10), butylamine (11), isopropanolamine (12), piperazine (13),
2-(2-aminoethoxy)-ethanol (14), 2-amino-1-butanol (15), diethanolamine + methyldiethanolamine (16),
amino-2-methyl-1-propanol (17), amino-2-ethyl-1,3-propandiol (18), and spermidine (19); (b): pyri-
dine (1), p-phenylenediamine (2), benzidine (3), o-toluidine (4), aniline (5), N, N-dimethylaniline (6),
p-anisidine (7), p-chloroaniline (8), m-chloroaniline (9), ethylaniline (10), α-naphthylamine (11), diethy-
laniline (12), N-(1-naphthyl)ethylenediamine (13), 4-aminophenazone (14), o-chloroaniline (15), 3,4-
dichloroaniline (16), 3,3 -dichlorobenzidine (17), 2-methyl-3-nitroaniline (18), 2,4-dichloroaniline (19),
2,3-dichloroaniline (20), and 2,5-dichloroaniline (21). (Modified from (a) A. Fekete et al., Electrophoresis,
27, 1237–1247, 2006 and (b) A. Cavallaro et al., J. Chromatogr. A, 709, 361–366, 1995. With permission.)
derivatives, usually assessed by EKC methods. Several derivatizing reagents have been pro-
posed in environmental applications: DNPH,25,68,200,201 3-methyl-2-benzothiazoline hydrazinone
(MBTH),202 5-(dimethylamino)-naphthalene-1-sulfohydrazide (dansylhydrazine, DNSH),68,203 and
4-hydrazinobenzoic acid (HBA).204
Table 31.5 compiles the details of the electromigration methodologies applied to carbonyls in
the environment. Figure 31.5 illustrates the electromigration separation of carbonyl compounds as
bisulfite adducts and DNSH derivatives.
31.4.6 SMALL IONS AND ORGANOMETALLIC COMPOUNDS

Aspects of the determination of small ions and organometallic compounds by CE methodologies as
well as elemental speciation have been extensively covered in the literature by excellent reviews
TABLE 31.5
Selected Applications of Electroseparation Methods for Carbonyl Compounds in Environmental Samples
938
Preconcentration
Acetone, acetaldehyde, CZE WATER Dansylhydrazine 5mmol L−1 phosphate, UV, 218 nm 68
propionaldehyde, benzaldehyde, Rain water derivatization 10 mmol L−1 tetraborate, Z-shaped flow cell,
formaldehyde, methylglyoxal in 20% ACN (pH 8.0) 170–300 nmol L−1
10.5 min
9 DNSH derivatives including isomers CZE AIR C18 catridges 20 mmol L−1 phosphate buffer UV, 214 nm 203
and impurities in 9 min Indoors impregnated with (pH 7.02) 1.1–9.5 µg L−1
Outdoors dansylhydrazine/ LIF, He/Cd 325 nm excitation,
trichloroacetic 520 emission
acid/MeOH 0.29–5.3 µg L−1
4 HBA derivatives in 6 min CZE AIR C18 catridges 0.040 mol L−1 tetraborate UV, 290 nm 204
Indoors impregnated with (pH 9.3) 2.7–8.8 ng L−1
HBA
Formaldehyde, acetaldehyde, EKC AIR IMPINGER 10 mmol L−1 UV, indirect detection, 254 nm 25
propionaldehyde, acrolein, Vehicular emission bisulfite derivatization 3,5 dinitrobenzoic acid, 10–40 µg L−1
benzaldehyde 0.2 mmol L−1 CTAB (pH 4.5)
(bisulfite derivatives) in ca. 7.5 min
5 DNPH derivatives in 14 min EKC AIR IMPINGER 20 mmol L−1 borate, UV, 360 nm 25
Vehicular emission 0.05 g DNPH in 50 mL 50 mmol L−1 SDS, 0.2–2.0 mg L−1
2 mol L−1 HCl 15 mmol L−1 β-CD
LLE in chloroform
3 DNPH derivatives in 8 min EKC GAS IMPINGER 0.02 mol L−1 borate, UV, 214 nm 200
Stack gas from 1.56 g DNPH in 0.05 mol L−1 SDS (pH 9) 2.0 mg L−1
an organic plant 500 mL 2 mol L−1
HCl
LLE in CS2
4 DNPH derivatives in 20 min EKC WATER Spiked sample 0.02 mol L−1 borate-phosphate, UV, 360 nm 201
River water 97–102% 0.05 mol L−1 SDS (pH 9) 0.05 mg L−1 (formaldehyde)
0.08 mg L−1 (acetaldehyde)
5 MBTH derivatives in 10 min EKC AIR IMPINGER 20 mmol L−1 tetraborate, UV, 216 nm 202
Indoors 0.05% MBTH 50 mmol L−1 SDS (pH 9.3) 0.54–4.0 µg L−1
stacking with salt
Obs LODs refer to each single aldehyde, not the derivative; DNPH, 2,4-dinitrophenylhydrazine; DNSH, 5-dimethylaminonaphthalene-1-sulfohydrazide; HBA, 4-hydrazinobenzoic acid; MBTH,
3-methyl-2-benzothiazoline hydrazinone; LLE, liquid–liquid extraction; CTAB, cetyltrimethylammonium bromide.
(a) R (b)
1 #
0.030 R
S 1
Standards 2 34 Standards
A.U
0.015
53 4
2
6
5 7 89
4a
** 0.000
36 9
0.030 0.005 1
1 2
S 2
Vehicle Indoor air
emission sample
sample 0.
A.U
0.015 4.5 6.0 7.5
1
R * 2
0.000
0 2 4 6 8 10 3 6 9
FIGURE 31.5 Separation of carbonyl compounds as (a) bisulfite adducts and (b) DNSH derivatives.
Electrolytes: (a) 10 mmol L−1 3,5-dinitrobenzoic acid at pH 4.5 containing 0.2 mmol L−1 CTAB;
(b) 20 mmol L−1 phosphate buffer at pH 7.02. Other conditions: (a) −10 kV; indirect UV detection at 254 nm;
(b) +20 kV applied voltage; direct UV detection at 214 nm. Peak labels: (a): formaldehyde (1), acetaldehyde (2),
propionaldehyde (3), acrolein (4), and benzaldehyde (5), impurities (∗ ) system peak (S), excess reagent (R);
(b): formaldehyde (1), acetaldehyde (2 and 7), propionaldehyde (3 and 6), acrolein (4, 8, and 9), and acetone (5),
impurities (#), excess reagent (R). (Modified from (a) E.A. Pereira et al., J. AOAC Int., 82, 1562–1570, 1999
and (b) E.A. Pereira et al., J. Chromatogr. A, 979, 409–416, 2002. With permission.)
and key articles. A comprehensive review on applications of CE to the analysis of inorganic species
including organometallic compounds in diverse environmental matrices (drinking, mineral, surface,
and ground waters; rainwater; snow; seawater; brine and waste waters; aerosol; and others) has been
compiled.205 The determination of inorganic ions in environmental aquatic samples of high salinity,
particularly seawater, has been revised.206
31.4.6.1 Inorganic Cations

The CZE determination of alkali, alkaline earth, and transition metal cations may be conducted
under indirect UV-detection in electrolytes composed of cationic chromophores (protonated imida-
zole, aromatic amines, pyridine derivatives at moderately low pH electrolytes) and weak complexing
agents (α-hydroxyisobutyric acid, HIBA, or other mono- and dicarboxylic acids), eventually mod-
ified by solvents. The simultaneous separation of 16 metal ions for inspection of river samples is
an example of this approach.207 Direct UV-detection at low wavelengths (190 nm) using nonab-
sorbing electrolytes (pH 10 borate buffer) is a possibility for absorbing cations such as ammonium,
as determined in river and sewage samples.208 Alternatively, UV-absorbing chelating (EDTA)62
or ion-pair forming agents (o-phenanthroline)209–211 can be selected for the determination of alka-
line earth and transition metal ions. Figure 31.6 illustrates the electromigration separation of small
cations.
(a) (b)
0.05
940
0.04 Fe (II) Zn (II)

Cu (II)
Standards
14
0.03
16
12 15
13 0.02
Absorbance
9
11
1 Mn (II) Cd (II)
3 10
8 0.01
4 7
0
6
5
Standards
2
−0.01
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 7.0 7.5 8.0 8.5 9.0 9.5 10.0
0.35
Fe (II)
1
2 7
16 NIST SRM 1648
13 urban particulate
0.25
matter at three
stage sequential
extraction
Zn (II)
5 0.15
Absorbance
Cu (II) Mn (II) F3
River water F2
0.05
contaminated
by copper F1
industry 6
−0.05
10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 7.0 7.5 8.0 8.5 9.0 9.5 10.0
FIGURE 31.6 Separation of metal ions. Electrolytes: (a) 10 mmol L−1 4-aminopyridine and 6.5 mmol L−1 HIBA at pH 4.5; (b) 200 mmol L−1 ammonium acetate at
pH 5.5, 0.5 mmol L−1 1,10-phenanthroline, 10 mmol L−1 hydroxylamine, and 20% acetone. Other conditions: (a) +25 kV; indirect UV detection at 214 nm; (b) +14 kV
applied voltage; direct UV-detection at 226 nm. Peak labels: (a): potassium (1), sodium (2), barium (3), strontium (4), calcium (5), magnesium (6), manganese (7), lithium (8),
iron (9), cobalt (10), cadmium (11), niquel (12), zinc (13), lead (14), chromium (15), and copper (16). (Modified from (a) N. Shakulashvili et al., J. Chromatogr. A, 895,
205–212, 2000 and (b) E. Dabek-Zlotorzynska et al., Anal. Bioanal. Chem., 372, 467–472, 2002. With permission.)
31.4.6.2 Inorganic Anions

Similar to the inorganic cations, the determination of inorganic anions and those derived from low-
molecular mass carboxylic acids are commonly approached by indirect UV-detection. However, here
the electrolyte chromophore must be anionic (chromate,212,213 benzoate derivatives,214 naphthalene
derivatives215,216 among others217,218 ) and an EOF reverser (long-chain tetraalkylammonium qua-
ternary salts) is usually employed. It is worth mentioning that many methodologies for inorganic
anions and even metal and nonmetal elements determined as oxoanions or anionic complexes that
are reported in the literature as CZE separations are, in fact, EKC separations. This is due to the
electrolyte ionic strength that forces micellization of the cationic surfactant used as flow reverser.219
Direct UV-detection at low wavelengths using nonabsorbing electrolytes (phosphate, borate, chlo-
ride, formate/chloride buffers modified by solvents, and cationic surfactants) is also a possibility
for UV-absorbing anions such as iodide,56,220,221 iodate,222 bromide,223 bromate,224 nitrate,225 and
nitrite226 as well as mixtures of anions9,66,227 and aromatic acids.228 Figure 31.7 illustrates the
electromigration separation of small anions.
31.4.6.3 Simultaneous Detection of Cations and Anions

An interesting strategy based on dual-opposite end sample injection that accomplishes the simulta-
neous separation of small cations and anions within a single run was demonstrated in the analysis
(a) (b)
11
Standards
18 20 6 Standards
15000 13 1517 12 1 3 8
2 7
12 16 5 9
Absorbance (µV)
67 8 10
10000 9 4
3 4
2
8 10 14 21 23 24 25 26 27
5000 1 5 0
4 28
19 22
0 29 Rain water
12
−5000 8 6
4
2 3 4 5 6 7 8
t (min) 0
mAU
3
Size-classified 6 Slag plate
2 rain sample 12
14
20000 8
4
Absorbance (µV)
4 12
15000 0
16 Sludge
10000 6 ? extract
7 17 20
23 12 4
6
5000 1 8
3 8
4
0
0
2 3 4 5 6 7 8 1 2 3 4
FIGURE 31.7 Separation of inorganic and organic anions. Electrolytes: (a) 7.5 mmol L−1 salicylic acid,
15 mmol L−1 Tris, 400 µmol L−1 DoTAH, 1050 µmol L−1 Ca2+ , and 600 µmol L−1 Ba2+ ; (b) 50 mmol L−1
tetraborate at pH 9.3, containing 5% methanol. Other conditions: (a) −28 kV; indirect UV-detection at 232 nm;
(b) −23 kV applied voltage; direct UV-detection at 200 nm. Peak labels: (a): chloride (1), nitrate (2), sulfate (3),
formate (4), fumarate (5), malonate (6), succinate (7), maleate (8), glutarate (9), methanesulfonate (10),
carbonate/adipate (11), malate (12), pimelate (13), acetate (14), suberate (15), oxalate (16), azelate (17),
sebacate (18), glyoxylate (19), propionate (20), methacrylate (21), lactate (22), butyrate (23), hydroxy-
iso-butyrate (24), valerate (25), capronate (26), enanthate (27), caprylate (28), and pelargate (29); (b):
bromide (1), iodide (2), nitrite (3), thiosulfate (4), chromate (5), nitrate (6), Fe(CN)4− −
6 (7), SCN (8),
MoO2− 2−
4 (9), and WO4 (10). (Modified from (a) A. Mainka et al., Chromatographia, 45, 158–162, 1997
and (b) M.I. Turnes-Carou et al., J. Chromatogr. A, 918, 411–421, 2001. With permission.)
of environmental water samples with indirect fluorescence73 and contactless conductometric78

detection.
31.4.6.4 Speciation and Organometallic Compounds

Speciation of chromium (histidine/acetic acid;229 acetate buffer/EDTA230 ), mercury (borate/MeOH),8
tin (pyridine/CTAB),231 lead (SDS/β-CD),232 arsenic (borate,233 phosphate or phosphate/borate,54
phosphate/TTAB,234 phosphate and phosphate/TTAB235 ), sulfur (phosphate/TTAB/ACN),236 and
selenium (histidine/acetic acid,237 SDS/β-CD232 ) by CE methods have been reported. Other exam-
ples include the analysis of beryllium (as an acetylacetone complex) in digested airborne dust238 and
the determination of Fe(II)-, Cu(I)-, Ni(lI)-, Pd(II)-, and Pt(II)-cyano complexes and nitrate from
leaching solutions of automobile catalytic converters.239
Metals including As, Co, Fe, Mn, Ni, and V are found at concentrations of 1–1000 ppm in
crude oils. About 27–100% of the total metals in crude oil occur as organometallic porphyrin
complexes. These petroporphyrins, derived from heme or chlorophyll, usually contain Ni(II) or
V(IV)O and are distinguished by various ring types (Etio, DPEP, Rhodo analogs) and different
substituents (H, n-alkyl). Petroporphyrin separations are of interest in geochemical sciences, envi-
ronmental monitoring, and process control. Separation of petroporphyrin models [Ni(II) and V(II)O
Etio I and Octaethyl type porphyrins] was approached by MEKC in electrolytes of varied compo-
sition: 15.8–22.5 mmol L−1 borate buffer containing 15–42 mmol L−1 SDS and 0–30% acetone,
methylethylketone, acetonitrile, or methanol in a 6–30 min window.240
Several examples of the aforementioned methodologies and their application in environment
compartments are compiled in Table 31.6.
31.4.7 EXPLOSIVES AND WARFARE RESIDUES

31.4.7.1 Explosives
A SB-β-CD-assisted EKC method for the determination of cyclic nitramine explosives and related
degradation intermediates and the 14 EPA listed explosives (borate/SDS electrolyte) has been
described.241 A volatile electrolyte composed of SB-β-CD modified ammonium acetate buffer was
selected for the EKC-MS detection of nitroaromatic and cyclic nitramine compounds in soil and
marine sediment,83 as detailed in Table 31.7. The use of phosphate/SDS electrolytes was reported
in the separation of the 14 listed nitramine and nitroaromatic explosives for the analysis of extracts
of high explosives such as C-4, tetrytol, and detonating cord.242
The simultaneous detection of small cations (ammonium, sodium, potassium, calcium, mag-
nesium, and strontium) and anions (bromide, chloride, nitrite, nitrate, sulfate, perchlorate, thio-
cyanate, and chlorate) from low explosives in postblast residue using an elaborate electrolyte com-
posed of a cationic chromophore and modifiers (imidazole/HIBA/18-crown-6 ether/ACN), an anionic
chromophore (1,3,6-naphthalenesulfonic acid) and flow reversal agent (tetramethylammonium
hydroxide) has been presented.243
31.4.7.2 Warfare Residues

A CZE method (borate buffer at pH 9.5) with LIF detection (He-Cd laser, 325 nm excitation; 450 nm
emission) that simultaneously examines thiols (OPA derivatives) and cyanide (derivatized by OPA-
taurine) aiming at monitoring for the potential contamination of drinking water supplies by chemical
warfare (CW) nerve agents was described.244 Detection limits of 9.3 µg L−1 for cyanide and from
1.8 to 89 µg L−1 for the thiols were obtained.
Table 31.7 compiles the details of CZE245,247 and EKC247,248 methodologies for inspection of
contaminated water and soil for alkylphosphonic acids, the breakdown products of organophosphorus
TABLE 31.6
Selected Applications of Electroseparation Methods for Small Ions and Organometallic Compounds in Environmental Samples
Preconcentration
Small ions
2− 2− −
Br− , I− , NO−2 , S2 O3 , CrO4 , NO3 , CZE WATER — 50 mmol L−1 tetraborate UV, 200 nm 9
4− − 2− 2−
Fe(CN)6 , SCN , MoO4 , WO4 in rain water 5% MeOH (pH 9.3) 0.02–0.1 mg L−1
3.5 min river water
drinking water
industrial residual
water
Lactate, butyrate, salicylate, propionate, CZE WATER — 5 mmol L−1 borate buffer, Fluorescence, indirect 73
phosphate, formate, citrate, acetate, pond water 5 µmol L−1 fluorescein detection,
Ba2+ , Ca2+ , Mg2+ , Ni2+ , Cu2+ in ca. (pH 9.2) (LED-induced fluorescence)
7 min 3.7–14.6 µmol L−1
NH+ + 2+ + 2+ + CZE WATER — 20 mmol L−1 MES, Contactless conductometric 77
4 , K , Ca , Na , Mg , Li
(internal standard) in ca. 3.8 min rain water 20 mmol L−1 His, detection
0.2 mmol L−1 18-crown-6 9.4–24 µg L−1
(pH 6.2)
K+ , Na+ , Ba2+ , Sr2+ , Ca2+ , Mg2+ , CZE WATER — 10 mmol L−1 4-aminopyridine, UV, indirect detection, 214 nm 207
Mn2+ , Li+ , Fe2+ , Co2+ , Cd2+ , Ni2+ , river water 6.5 mmol L−1 HIBA 92–454 µg L−1
Zn2+ , Pb2+ , Cr3+ , Cu2+ in 24 min (pH 4.5 adjusted with
sulfuric acid)
Fe2+ , Cu2+ , Zn2+ , Mn2+ , Cd2+ CZE AIR Sequential extraction 200 mmol L−1 NH4Ac UV, 226 nm 210
in 9 min airborne particulate 78–87% (pH 5.5), 0.5 mmol L−1 29–142 µg L−1
1,10-phenanthroline,
10 mmol L−1 hydroxylamine
hydrochloride, acetone 20%
Continued
943
944
TABLE 31.6
(Continued)
Preconcentration
Zn2+ , Cu2+ , Co2+ , Fe2+ , Cd2+ in CZE AIR Andersen 30 mmol L−1 hydroxylamine UV, 265 nm 211
ca. 7 min Respirable, sampler, filters hydrochloride, 0.1 mmol L−1 0.5–3 µg L−1
fine, and sonication 1,10-phenanthroline,
Coarse air 1% MeOH (pH 3.7)
Particulate
C6–C12 perfluorinated carboxylic acids CZE WATER — 50 mmol L−1 TRIS, UV, indirect detection, 214
in 20 min Lake water 7 mmol L−1 2,4-DNBA, 270 nm
River water 50% MeOH (pH 9.0) 0.6–2.4 mg L−1
Tap water
Bromide in 6 min CZE WATER Spiked sample 5 mmol L−1 formic acid, UV, 200 nm 223
Ground water 61.6–102% 42 mmol L−1 NaCl (pH 3.5) 0.1 mg L−1
Surface water
Potassium in ca. 11 min CZE WATER Spiked sample 50 mmol L−1 18-crown-6 UV, indirect detection, 223
Ground water 73.4–97.5% 10 mmol L−1 imidazole 214 nm
Surface water (pH 4.5) 0.5 mg L−1
Bromate in 4 min CZE WATER Spiked sample 10 mmol L−1 phosphate, UV, 193 nm 224
Tap water 100–110% 10 mmol L−1 Na2 SO4 . 1 µg L−1
River water (pH 3.2)
Nitrite, nitrate in ca. 7 min CZE WATER tITP SE: artificial sea water with UV, 210 nm 226
Sea water pH adjusted to 3.0 using 2.7–3.0 µg L−1
phosphate buffer as nitrogen
TE: 600 mmol L−1 acetate
(pre-rinse with 0.1 mmol L−1
DDAB)
Ammonium in ca. 8 min CZE WATER — 20 mmol L−1 borate UV, 190 nm 228
River water (pH 10 adjusted with 0.24 mg L−1
SEWAGE 1 mol L−1 sodium hydroxide) (as nitrogen)

3,4,5,-Trimethoxybenzoic acid, CZE WATER SPE 13 mmol L−1 tetraborate UV, 214 nm 228
4-hydroxyphenylacetic acid, salicylic Lake water C18 (pH 9.7)
acid, ferulic acid, p-coumaric acid, 58–108%
vanillic acid, 4-hydroxybenzoic acid
in 11 min
Iodide in ca. 35 min EKC WATER Spiked sample SE: 0.5 mol L−1 NaCl, UV, 226 nm 56
Sea water tITP 25 mmol L−1 CTAC (pH 2.4) 0.6 µg L−1
99.6–100.4% TE: 500 mmol L−1 MES
(pH 6.0)
− 2−
NO− 3 , Cl , SO4 , formate, acetate EKC WATER — 20 mmol L−1 MES, Contactless 77
in ca. 3.7 min Rain water 20 mmol L−1 His, conductometric
0.2 mmol L−1 CTAB (pH 6.2) detection
48–115 µg L−1
(inorganic ions)
Ammonium, potassium, sodium, EKC WATER — 20 mmol L−1 MES/His Contactless 78
calcium, magnesium, manganese, Rain water 11.5 mmol L−1 18-crown-6, conductometric
lithium, chloride, nitrite, nitrate, Surface water 10 µmol L−1 CTAB (pH 6.0) detection
sulfate, fluoride, phosphate in 5.5 min Drainage water 10–250 µg L−1

2− − −
F− , Cl− , NO−2 , SO4 , NO3 , HCO3 EKC WATER Spiked sample 6.0 mmol L−1 chromate, UV, indirect 213
in ca. 0.7 min Snow 91–104% 2.5 mmol L−1 CTAB, detection, 254 nm
ACN 3.6%, (pH 9.5) 0.03–0.2 mg L−1
Oxalate, glycolate, malonate, pyruvate, EKC AEROSOL 47-mm PTFE filters 4 mmol L−1 NDC, UV, indirect 215
formate, suberate, malate, acetate, Atmospheric aerosol (aerosol) 14.5 mmol L−1 Bis–Tris, detection, 214 nm
succinate, glyoxylate, phthalate, GAS KOH-coated quartz 0.2 mmol L−1 TTAB (pH 6.2) 3–10 µg L−1
lactate, methanesulfonate, propionate, Vehicle emission fiber
glutarate, benzoate in 8.5 min (vehicle emission)
Chloride, nitrate, sulfate, formate, EKC WATER freezing in liquid 7.5 mmol L−1 salicylic acid, UV, 232 nm 217
fumarate, malonate, succinate, Size classified nitrogen 15 mmol L−1 Tris, Z-shaped flow cell,
maleate, glutarate, methanesulfonate, raindrops (Guttalgor method) 400 µmol L−1 DoTAH, 90–600 nmol L−1
carbonate/adipate, malate, pimelate, 1050 µmol L−1 Ca2+ ,
acetate, suberate, oxalate, azelate, 600 µmol L−1 Ba2+ (pH 6.2)
sebacate, glyoxylate, propionate,
methacrylate, lactate, butyrate,
hydroxy-iso-butyrate, valerate,
capronate, enanthate, caprylate,
pelargate in 9.5 min
Continued
945
TABLE 31.6
946
(Continued)
Preconcentration
Cl− , NO−
3, SO2−
4 , ClO−
4, F− in 2.5 min EKC WATER — 0.25mmol L−1 CTAC, MeOH UV, indirect detection, 218
Tap water 30%, 254 nm
River water 5 mmol L−1 chromate 0.09–0.23 mg L−1
(pH 8.0)
Iodide in ca. 33 min EKC WATER tITP SE: 0.5 mol L−1 NaCl, UV, 226 nm 221
Sea water 25 mmol L−1 CTAC 0.2 µg L−1
(pH 2.4 adjusted with HCl)
TE: 0.5 mol L−1 MES
(pH 6.0 adjusted
with NaOH)
Iodide, iodate in ca. 33 min EKC WATER — SE: 0.5 mol L−1 NaCl , UV, 210 nm, 226 nm 222
Sea water 12.5 mmol L−1 CTAC 0.23–10 µg L−1
(pH 2.4 adjusted with HCl) (tITP)
TE: 0.5 mol L−1 MES
(pH 6.5)
Nitrate in ca. 7.6 min EKC WATER — 0.1 mol L−1 sodium phosphate, UV, 210 nm 225
Sea water 0.15 mol L−1 DDAPS 35 µg L−1
(pH 6.2)
Nitrate, nitrite, bromide, iodide EKC WATER — (1) 10 mmol L−1 DDAPS, UV, 214 nm 227
in ca. 5.8 min Sea water 10 mmol L−1 phosphate 5–11 µmol L−1
(pH 8.0)
(2) 30 mmol L−1 DDAPS,
10 mmol L−1 phosphate
(pH 8.0)
Organometallic
Methylmercury, ethylmercury, CZE WATER Spiked sample 100 mmol L−1 boric acid, Atomic fluorescence 8
phenylmercury, mercury (II), as cysteine Lake water 86.6–111% 12% MeOH spectrometry
complexes in 14 min River water (pH 9.1) 6.8–16.5 µg L−1
(as Hg)
Arsenite, arsenate, dimethylarsinate, CZE Environmental LV-FASI 20 mmol L−1 phosphate, UV, 200 nm 54
monomethylarsonate in ca. 19 min reference material 10 mmol L−1 borate 24–93 µg L−1
(pH 9.28)
Cr(III), Cr(IV) in 3 min CZE WATER — 4.5 mmol L−1 His Contactless conductometric 229
Rinse water (pH 3.40 adjusted with detection
from the acetic acid) 10–39 µg L−1
galvanic
industry
Cr(III), Cr(IV) in ca. 4.5 min CZE WATER Spiked sample 20 mmol L−1 acetate buffer, Chemiluminescence 230
Tap water 98–103% 1 mmol L−1 EDTA (luminol)
Well water (pH 4.7) 1–8 pmol L−1
River water
Waste water
Se(IV), Se(VI) in ca. 8 min CZE SOIL Sonication with water 8.75 mmol L−1 His Contactless conductometric 237
90–101% (pH 4.0 adjusted with detection
acetic acid) 7.5–190 µg L−1
Trimethyllead, triethyllead, EKC WATER LLE 50 mmol L−1 SDS, UV, 210 nm 232
diphenylselenide, phenylselenyl Drainage water chloroform 5 mmol L−1 β-CD 26–67 mg L−1
in 20 min 80–104% (pH 6.0)
Arsenite, arsenate, dimethyl arsenic EKC WATER FASI 10 mmol L−1 phosphate UV, 185 nm 234
acid in 4 min Ground water buffer, 0.1–0.5 mg L−1
0.35 mmol L−1 TTAB

(pH 9.0)
2−
AsO2−2 , AsO4 , dimethylarsinic acid EKC WATER FASI 20 mmol L−1 phosphate, UV, 185 nm 235
in 2 min Ground water 0.75 mmol L−1 TTAB 1 µmol L−1
(pH 9.0)
Be, as diacetylacetonate-beryllium EKC AIR — 0.1 mol L−1 TRIS-nitrate, UV, 295 nm 238
in 15 min Digested airborne 50 mmol L−1 SDS 1 mg L−1
Dust (model (pH 7.8)
sample)
Fe(II)-, Cu(I)-, Ni(II)-, Pd(II)-, and EKC CATALYTIC alkaline NaCN 20 mmol L−1 phosphate, UV, 208 nm 239
Pt(II)-cyano complexes, nitrate CONVERTER solutions 100 mmol L−1 NaCI, 11–60 µg L−1
in 20 min RESIDUE 3 mmol L−1 NaCN,
leaching solutions 1.2 mmol L−1 TBAB,
40 µmol L−1 TTAB
(pH 11)
FASI, field-amplified sample injection; SPE, solid-phase extraction; LV-FASI, large-volume field amplified stacking injection; TBAB, tetrabutylammonium bromide; TTAB, tetradecylytrimethy-
lammonium bromide; CTAB, cetyltrimethylammonium bromide; NDC, 2,6-naphthalenedicarboxylic acid; Bis Tris, 2,2-Bis(hydroxymethyl)-2,29,20-nitrilotriethanol; DoTAH, dodecyltrimethy-
lammoniumhydroxide; CTAC, cetyltrimethylammonium chloride; 2,4-DNBA, 2,4-dinitrobenzoic acid; SE, separation electrolyte; TE, terminating electrolyte; tITP, transient isotachophoresis;
DDAB, dilauryldimethylammonium bromide; DDAPS, 3-(n,N-dimethyldodecylammonium)propane sulfonate; HIBA, α-hydroxyisobutyric acid; MES, 2-[morphine]ethanesulphonic acid;
947
LLE, liquid–liquid extraction.

TABLE 31.7
948
Selected Applications of Electroseparation Methods for Explosives and Warfare Residues in Environmental Samples
Preconcentration
Compounds CE Mode Matrix (Recovery) Electrolyte (LOD) Referencess
Methylphosphonic acid, CZE WATER SPE 50 mmol L−1 NH4Ac FPD, (P) mode, 245
ethylphosphonic acid, SOIL Bond-Elut SCX (pH 9.0 adjusted by ammonia 0.1–0.5 µg L−1
propylphosphonic acid, solution)
N -butylphosphonic acid,
phenylphosphonic acid,
ethyl methylphosphonic acid,
ethyl methylthiophosphonic acid,
isopropyl methylphosphonic acid,
pinacolyl methylphosphonic acid,
dimethyl phenylphosphonate in 16 min
Methylphosphonic acid, ethylphosphonic acid, CZE SOIL PHWE 15 mmol L−1 NH4Ac ESI-IT-MS 246
propylphosphonic acid, phenylphosphonic acid, 150 bar, 100◦ C (pH 8.8 adjusted by ammonia 5 µg L−1
isopropylphosphonic acid, SPE solution)
methyl ethylphosphonic acid, cation exchange cartridge
ethyl methylphosphonic acid,
ethyl ethylphosphonic acid,
methyl propylphosphonic acid,
propyl methylphosphonic acid in 15 min
2,4,6-Trinitrotoluene, 1,3,5-trinitrobenzene, EKC SOIL — 10 mmol L−1 SB-β-CD, ESI-IT-MS 83
hexahydro-1,3,5-trinitro-1,3,5-triazine, Contaminated soil 10 mmol L−1 NH4Ac 0.025–0.5 mg L−1
octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine, SEDIMENT (pH 6.9)
2,4,6,8,10,12-hexanitro-2,4,6,8,10,12- Marine sediment
hexaazaisowurtzitane in 16 min
Methylphosphonic acid and its monoacid/ EKC WATER — 30 mmol L−1 L-His, Conductivity 247
monoalkyl esters (ethyl, isopropyl, and Surface water 30 mmol L−1 MES,
pinacolyl) in 10 min Ground water 0.7 mmol L−1 TTAOH,
SOIL 0.03% wt Triton X-100
Methylphosphonic acid and its monoacid/ EKC WATER SPE 200 mmol L−1 boric acid, UV, indirect detection, 210 nm 248
monoalkyl esters (ethyl, isopropyl, Tap water On-Guard-Ba 10 mmol L−1 phenylphosphonic 1–2 µg L−1
and pinacolyl) in 3 min Ground water On-Guard-Ag acid, 0.03% wt Triton X-100, (water samples)
Artificial sea water On-Guard-H 0.35 mmol L−1 DDAOH 25–50 µg L−1
SOIL 90–110% (pH 4.0) (aqueous leachates of soil)
TTAOH, tetradecyltrimethylammonium hydroxide; ESI-IT-MS, electrospray ionization ion trap mass spectrometry; SB-β-CD, sulfobutylether-β-cyclodextrin; PHWE, pressurized hot water
extraction; FPD, flame photometric detector; SPE, solid-phase extraction; MES, 2-[morphine]ethanesulphonic acid; DDAOH, didodecyldimethylammonium hydroxide.
(a) (b)
EMPTA
DMPhP
nPrPA
PhPA
MPA
nBPA
13
EPA
EMPA
iPrMPA
10
PMPA
CH3CN
FPD response (mV)

8
2,3
6 Standards
14
4
4 6 Sudon III 1
7
1 8
9 2
Water sample
2
10
5
0 Blank solution
11,12
4 6 8 10 0 4 8 12 16
Time (min) Migration time (min)
FIGURE 31.8 Separation of explosives (a) and warfare residues (b). Electrolytes: (a) 12 mmol L−1
borate at pH 9 and 50 mmol L−1 SDS; (b) 50 mmol L−1 ammonium acetate at pH 9.0. Other conditions:
+30 kV, direct UV-detection; (b) +30 kV; make up: 0.5% formic acid; flame photometric detector. Peak
legends: (a): octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (1); hexahydro-1,3,5-trinitro-1,3,5-triazine (2);
1,3,5-trinitrobenzene (3); 1,3-dinitrobenzene (4); nitrobenzene (5); 2,4,6-dinitrotoluene (6); N-2,4,6-tetranitro-
N-methylaniline (7); 2,4-dinitrotoluene (8); 2,6-dinitrotoluene (9); 2-nitrotoluene (10); 3-nitrotoluene (11);
4-nitrotoluene (12); 2-amino-4,6-dinitrotoluene (13); and 4-amino-2,6-dinitrotoluene (14). (b): methylphos-
phonic acid (MPA), ethylphosphonic acid (EPA), n-propylphosphonic acid (nPrPA), n-butylphosphonic acid
(nBPA), phenylphosphonic acid (PhPA), ethylmethylphosphonic acid (EMPA), ehtylmethylthiophosphonic
acid (EMPTA), isopropylmethylphosphonic acid (iPrMPA), pinacolylmethylphosphonic acid (PMPA), and
dimethylphenylphosphonate (DMPhP); (1) and (2) are suspected contaminants of water sample. (Modified from
(a) C.A. Groom et al., J. Chromatogr. A, 999, 17–22, 2003 and (b) E.W.J. Hooijschuur et al., J. Chromatogr. A,
928, 187–199, 2001. With permission.)
nerve agents. The unusual detection system for CE (flame photometric detector)245 with an ability to
detect alkylphosphonic acids a fold below the required 1 µg mL−1 level is worth mentioning here.
Figure 31.8 illustrates the electromigration separation of EPA explosives and warfare residues.
31.4.8 AROMATIC SULFONATES

Aromatic sulfonates and their amino- and hydroxy-derivatives are produced on a large scale in the
chemical industry. Although the acute toxicity and the risk of bioaccumulation appear to be small,
they became persistent and widespread environmental pollutants due to their high mobility in the
aquatic compartments and limited biodegradability.
The anionic charge of sulfonates within the entire practical pH range of electromigration sep-
arations and the UV-absorbing capability of the aromatic moiety make CZE-based methodologies
of aromatic sulfonates feasible. However, improved selectivity and resolution of positional isomers
have been sought by means of interactions with CDs,249 nonionic,250 and cationic250 surfactants
as well as mixed micelles.251 An overview of the electrophoretic methods for the determination of
benzene- and naphthalenesulfonates in water samples has been presented in the literature.252
Details of a CZE methodology for the separation of 14 aromatic sulfonates in river water253 and
a CZE-MS evaluation of 22 compounds in influent and effluent samples of a wastewater treatment
plant254 are organized in Table 31.8. With the idea that phytoremediation could be used as an
alternative means of wastewater treatment of recalcitrant compounds, a simple borate buffer-based
950
TABLE 31.8
Selected Applications of Electroseparation Methods for Aromatic Sulfonates, Dyes, and Surfactants in Environmental Samples
Preconcentration
Aromatic sulfonates
14 Aromatic sulfonates of a wide range of CZE WATER SPE 25 mmol L−1 sodium borate UV, 210 nm 253
different structure in 9.5 min River water C18 (pH 9.3) fluorescence, Xe lamp
Contaminated LiChrolut EN 0.2 mg L−1 (CE)
seepage water 72–132% 0.1 µg L−1 (SPE-CE)
13 Aromatic sulfonates of a wide range of CZE WATER SPE 5 mmol L−1 NH4Ac (pH 10.5) ESI-MS 254
different structure in ca. 20 min Waste water treatment C18 0.1–0.4 mg L−1
LiChrolut EN (CE-MS)
>50% 0.1 µg L−1
(SPE-CE-MS)
Anthraquinone-1-sulphonic acid, CZE WATER Aqueous extraction 20 mmol L−1 tetraborate (pH 9.3) UV, 254 nm 255
anthraquinone-2-sulphonic acid, Hydroponic media (plant samples)
anthraquinone-1,5-disulphonic acid, PLANT EXTRACTS direct injection
anthraquinone-2,6-disulphonic acid, Plant cultivated (water samples)
anthraquinone-1,8-disulphonic acid in 7.5 min in spiked
hydroponic
media
21 Naphthalenesulfonate derivatives in 28 min EKC WATER SPE 50 mmol L−1 borate, UV, 230 nm 17
River water 100 mmol L−1 SDS (pH 8.7) 20 µg L−1
21 Naphthalenesulfonate derivatives in 30 min EKC WATER Sample was diluted 25 mmol L−1 tetraborate, UV, 220 nm 19
Industrial effluent 1/1000 and filtered 75 mmol L−1 Brij 35, 0.5–3.2 mg L−1
5 mmol L−1 octylamine (pH 9.0)
56 Aromatic organic acids were evaluated; EKC WATER Soxhlet (soil), SPE 50 mmol L−1 borate, UV, 214 nm 28
2,4-dichlorophenoxyacetic acid, Deionized water cation exchange 100 mmol L−1 cholate (pH 8.3)
2,4,5-trichlorophenoxyacetic acid, orange II, SOIL 38–101% (water)
trypan blue, 2,4,5-trichlorophenoxypropionic 26.5–94.4% (soil)
acid, 3-nitrobenzoic acid,
4-chlorobenzensulfonic acid (recovery studies)
in 20 min
Surfactants
4 LAS (C10, C11, C12, C13) in 45 min CZE WATER SPE (C18 and SAX) 50 mmol L−1 NH4Ac, UV, 214 nm 87
Coastal marine 94–98% (marine water) 30% isopropanol 1 µg L−1
water SPE (Isolute ENV+) (pH 5.6) (SPE-CE-UV)
Waste water 77–93% ESI-MS
Treatment plants (wastewater) (qualitative
confirmation)
19 LAS isomers (commercial formulation) EKC WATER LLE, MeOH/NaOH 10 mmol L−1 phosphate, UV, 200 nm 18
in 45 min Industrial (44.6–96.2%) 40 mmol L−1 SDS, 4 mg L−1
Waste water SPE, C18 and SAX 30% ACN (pH 6.8)
SEWAGE SLUDGE (33.1–75.7%)
Dyes
5 Vinylsulfone and chlorotriazine reactive CZE WATER SPE 5 mmol L−1 NH4Ac ESI-MS single 84
dyes in 34 min Waste water C18 (pH 9.0) quadrupole
81–121% 23–42 µg L−1
Eosin, fluorescein (internal standard) CZE WATER Charcoal adsorption 20 mmol L−1 borate LIF, 514.5 nm 259
in 5.3 min Ground water (pH 9.2) Ar/Kr ion laser

1 nmol L−1
Sulfonated azo dyes: Acid Blue 113, Acid EKC WATER SPE 9.5 mmol L−1 NH4Ac, UV, 214 nm 20
Red 73, Acid Red 13, Mordant Yellow 8, Waste water Isolute ENV 0.1% Brij 35 (pH 9) 19–230 mg L−1
Acid Red 1, Acid Red 14, Acid Red 9, 54–81%
Acid Yellow 23 in 16 min
Monosulfonated dyes: cresol red, acid EKC WATER SPE 100 mmol L−1 cholate, UV, 214 nm 29
blue, acid orange, tropaeolin, nuclear fast Spiked water ion-pair (water) 10% acetone (pH 8.35)
red, orange II, acid red in 27 min SOIL C18 (soil)
42.5–107% (water)
17.9–105% (soil)
SPE, solid-phase extraction; ESI- MS, electrospray ionization mass spectrometry; LIF, laser-induced fluorescence; LLE, liquid–liquid extraction; LAS, linear alkylbenzenesulfonates.
951
0.014
20+21
18 4
0.012
15
0.01
Absorption (relative intensity)
12+16+11 4
1 6+10
0.008 19
3
8 5 23
22
0.006 I
I
0.004
EOF
0.002
0
2 4 6 8 10 12 14
Time (min)
FIGURE 31.9 Separation of aromatic sulfonates. Electrolyte: 12 mmol L−1 ammonium acetate at pH 10.
Other conditions: +20 kV; direct UV detection at 214 nm. Peak legend: 2-amino-1,5-NDS (1); 1,3,6-NTS (2);
1,3-BDS (3); 1,5-NDS (4); 2,6-NDS (5); 1-OH-3,6-NDS (6); 1-amino-5-NS (7); BS (8); 1-amino-4-NS (9);
2-OH-3,6-NDS (10); 1-OH-6-amino-3-NS (11); 3-nitro-BS (12); 1-amino-6-NS (13); 4-methyl-BS (14); 1-OH-
4-NS (15); 4-chloro-BS (16); 2-amino-1-NS (17); 1-amino-7-NS (18); 4-chloro-3-nitro-BS (19); 1-NS (20);
2-NS (21); diphenylamine-4-sulfonate (22); and 8-NH2 -1-OH-3,6-NDS (23). Abbreviations: NS, naph-
thalenesulfonate; NDS, naphthalenedisulfonate; NTS, naphthalenetrisulfonate; BS, benzenesulfonate; BDS,
benzenedisulfonate; and OH, hydroxy. (Modified from R. Loos et al., J. Mass Spectrom., 35, 1197–1206, 2000.
With permission.)
CZE methodology was used to monitor sulfonated anthraquinones uptake in hydrophonic media, in
which plants, bacteria, and algae had grown for 6 weeks.255
EKC methodologies for aromatic sulfonates in industrial effluents,19 mono- and dinaphthale-
nesulfonates as well as hydroxy- and amino-derivatives in river water,17 and phenoxyacetic acids
in spiked water and soils28 are compiled in Table 31.8. Figure 31.9 illustrates the electromigration
separation of aromatic sulfonates.
31.4.9 SURFACTANTS
Anionic surfactants such as alkanesulfonates, alkyl sulfates, and LAS are water-soluble, surface-
active materials that are consumed in large quantities in industrial and commercial formulations.
Thus, their disposal may impact water reservoirs. LAS surfactants, which are used commercially,
are quite complex mixtures containing several homologs and positional isomers. The degradation
rates and toxicity of LAS depend on the alkyl chain length and the position of the phenyl ring.
Since LAS are chromophoric-charged compounds, they are amenable to CZE18,87 determination.
EKC mode is preferred when homologue discrimination or isomeric distribution is requested31,32 or
mixture of surfactant classes are screened.33,34
LAS were determined in influent and effluent samples of wastewater treatment plants and coastal
waters receiving untreated domestic effluents using a volatile buffer-based CZE method (UV and
15
10
Absorbance at 224 nm (mAU)
5 9
78
10 11
12 14 + 15
13
5 16
1 17
2 4
3 1819
20
4 8 12 16 20
Time (min)
FIGURE 31.10 Separation of LAS surfactants. Electrolyte: 10 mmol L−1 sodium phosphate at pH 6.8,
40 mmol L−1 SDS and 30% acetonitrile. Other conditions: +15 kV applied voltage. Peak legend: positional
isomers: 5-C10 (1), 4-C10 (2), 3-C10 (3), 2-C10 (4), 6-C11 (5), 5-C11 (6), 4-C11 (7), 3-C11 (8), 2-C11 (9),
6-C12 (10), 5-C12 (11), 4-C12 (12), 3-C12 (13), 2-C12 (14), 7-C13 (15), 6-C13 (16), 5-C13 (17), 4-C13 (18),
3-C13 (19), and 2-C13 (20); first number represents the attachment point of the phenyl ring to the carbon chain.
(From J.M. Herrero-Martínez et al., Electrophoresis, 24, 681–686, 2003. With permission.)
MS detection).87 Phosphate and borate buffers containing acetonitrile were used to discriminate LAS
homologues in sewage sludge, whereas unmodified buffers were applied to wastewater samples.18
Isomeric separation was only possible using electrolytes with high contents of SDS and acetonitrile as
organic modifier. Table 31.8 compiles the details of the aforementioned methodologies. Figure 31.10
illustrates the electromigration separation of LAS surfactants.
31.4.10 DYES
Synthetic dyes, including azo compounds, are widely used as coloring agents in a variety of products
such as textiles, paper, leather, gasoline, and foodstuffs. Synthetic dyes persist even after conven-
tional water treatment procedures due to their hydrophilic character and high solubility (they usually
bear carboxylic or sulfonic acid groups in their structure) and, therefore, can be distributed in the
environment from urban and industrial wastewater.
CZE separation of synthetic dyes has been approached by simple (borate and citrate)256 and
volatile buffers (ammonium acetate)257 modified by solvents as well as nonaqueous systems (ammo-
nium acetate/acetic acid in MeOH).258 Environmental applications of CZE methodologies include the
analysis of spent dyebaths and wastewater samples257 and the monitoring of groundwater migration,
where eosin was used as a fluorescent tracer (details in Table 31.8).259
EKC separations of synthetic dyes include buffered SDS systems260 and polymeric
electrolytes,261 modified by organic solvents.260 Examples of the determination of azo dyes, mono-
and disulfonated compounds in water samples,20 as well as synthetic dyes in spiked water and soil
matrices29 have been compiled in Table 31.8. Figure 31.11 illustrates the electromigration separations
of synthetic dyes.
31.4.11 ENDOCRINE DISRUPTORS AND PHARMACEUTICALS

Recently, it has been established that certain synthetic organic chemicals affect the reproductive
health of higher organisms by contributing to infertility in various ways and even increasing the rate of
cancer in reproductive organs. These chemicals have been termed “environmental estrogens” due to
their disrupting effects on the endocrine system of hormone production and transmission. In addition
to organochlorine insecticides, PCBs and dioxins (covered in Section 31.4.2), phenolic compounds
such as bisphenol A—a widely used substance that is polymerized industrially into polycarbonate
and nonylphenol—a product of the breakdown of large surfactants used in spermicides, some plastics
and plasticizers such as phthalic acid esters, as well as pharmaceuticals such as synthetic estrogens
belong to the list of chemicals suspected to have endocrine-disrupting effects.
31.4.11.1 Phenolic Compounds

The CZE separation of halogenated phenolic and bisphenolic compounds from 25 potentially inter-
fering phenolic derivatives for inspection in water, sludge, and sediments has been attempted262–264
using aqueous (borate buffer at pH 9.4) and nonaqueous (borate buffer in MeOH) systems (details
in Table 31.9).
MEKC separation and online concentration of bisphenolAand alkyl phenols has been approached
in a series of experiments using SDS and other alkyl chain anionic surfactants, bile salts, and TTAB
in organic solvent and CD-modified buffers.59,265,266
The EKC determination of several target endocrine disruptors in water267,268 and biosolids
(sewage sludge and treated sludge)269 has been detailed in Table 31.9. The potential of MEEKC
for the separation of priority endocrine-disrupting compounds in wastewater samples has been
investigated.45
31.4.11.2 Phthalate Esters

Excessive use of phthalate esters in industrial applications, mainly as plasticizers, is the main cause
of their persistent presence in consumer goods and in the environment. The economic and social
interest in the control of phthalate esters and the availability of analytical methodologies for areas
such as environmental and food analyses have been discussed.270
31.4.11.3 Pharmaceutical Residues

The occurrence and fate of pharmaceuticals from various prescription classes and related metabolites
and medicinal products for veterinary use in aquatic environments as well as their removal is one of the
emerging issues in environmental chemistry.271 More than 80 different prescribed pharmaceuticals,
metabolites, and veterinary drugs have been detected in aquatic environment (e.g., sewage influent
and effluent samples, surface and groundwater, and even drinking water), as reported by studies
carried out in several European countries, United States, Canada, and Brazil.271
Tetracyclines,272 nonsteroidal anti-inflammatory,58,273 antidepressants,274 and mixtures of acidic
drugs82,275 as well as veterinary drug276 residues have been determined by CZE (citric acid/citrate,
borate, borate in MeOH, phosphate/MeOH, ammonium formate/formic acid/ACN, ammonium
acetate, ammonium acetate/acetic acid/MeOH) and EKC (SDS/pH 2.5 phosphate buffer/ACN)
methodologies in water samples from influent and effluent of sewage treatment plants as well as
wastewater, river, surface and groundwater as compiled in Table 31.9. Estrogens have been receiv-
ing increased attention due to the already mentioned possible interference with the reproductive
(a)
100 Effluent
m/z 343.5
5 (OH)
%
0
(b) 100
Effluent
m/z 264.7
4 (OH/OH)
%
0
(c)
100
Effluent
m/z 270.7
% 4 (OH/CI)
0
(d) 100
Effluent
2 (OH) m/z 501
0
(e) 100 2 (OH) 3 (OH/OH) Ref. mixture
NSA 4 (VS/CI) RIC
1 (OH) 5 (OH)
5 (VS)
%
4 (OH/OH)
4 (OH/CI)
0
10 15 20 25 30 35 40
Time (min)
FIGURE 31.11 Identification of synthetic dyes in a sewage extract effluent (mass traces a–d) and a reference
mixture (e) by CE-MS. Electrolyte: 5 mmol L−1 ammonium acetate with 40% acetonitrile. Other conditions:
+30 kV applied voltage; sheath liquid: 80:20 isopropanol:water. Abbreviations: OH, hydroxy form; Cl, chloride
form; VS, vinylsulfone; NSA, naphthalene sulfonic acid; RIC, reconstructed ion chromatogram. (From T. Poiger
et al. J. Chromatogr. A, 886, 271–282, 2000. With permission.)
TABLE 31.9
Selected Applications of Electroseparation Methods for Endocrine Disruptors and Pharmaceuticals in Environmental Samples and
956
Miscellaneous
Preconcentration
Endocrine disruptors
2,4,6-Tribromophenol, CZE WATER SPE-LVSEP 20 mmol L−1 tetraborate in UV, 210 nm, 230 nm 262
pentabromophenol, River water Polystyrene– MeOH 0.4–1.7 µg L−1
tetrabromobisphenol A, Waste water divinylbenzene (non-aqueous)
tetrachlorobisphenol A, 95.5–105.8% (pH 9.4)
2,6-Dibromophenol in 28 min (river water)
73–94%
(Waste water)
4-Bromo-3-methylphenol, CZE WATER SPE 20 mmol L−1 sodium UV, 210 nm 263
2-bromo-4-methylphenol, Waste water PS–DVB tetraborate
pentabromophenol, (pH 9.6)
2,4,6-tribromophenol,
2,4-dibromophenol,
2-bromophenol,
2,6-dibromophenol,
tetrabromobisphenol A,
tetrachlorobisphenol A in ca. 8
min
2,6-Dibromophenol, CZE SEWAGE SEDIMENT MSPD-LVSEP 20 mmol L−1 tetraborate in UV, 210 nm, 230 nm 264
2,4,6-tribromophenol, River sediment Florisil MeOH 2.7–5.3 µg L−1
tetrabromobisphenol A, Marine sediment 75.7–106.4% (nonaqueous)
pentabromophenol, (pH 9.4)
tetrachlorobisphenol A in 28 min
Octylphenol, nonylphenol in 10 EKC WATER SPE 25 mmol L−1 phosphate, UV, 214 nm 45
min Waste water treatment C18 200 mmol L−1 SDS, 50 mg L−1 (tested)
effluents and sludges spiked sample 900 mmol L−1 butanol,
25.6–50.8% 80 mmol L−1 heptane,
20% propanol (pH 2)
17β-Estradiol, diethylstilbestrol, EKC WATER Spiked sample 100 mmol L−1 phosphate, UV, 214 nm 267
ethynylestradiol, octylphenol, River water analytes solubilized 12.5% ACN, 25 mmol L−1 low mg L−1 range
nonylphenol, bisphenol A in 15 in 10% ACN:90% SDS, 1 mmol L−1 HP-β-CD

min buffer (pH 1.8)
Estriol, phenol, trichlorophenol, EKC WATER Spiked sample 20 mmol L−1 CAPS, UV, 200 nm 268
bisphenol A, pentachlorophenol, River water 25 mmol L−1 SDS, 2.0–7.4 mg L−1
butylphenol, estrone, β-estradiol, 15% ACN (pH 11.5)
diethylstilbestrol, ethinylestradiol,
nonylphenol in 25 min
Estrone, β-estradiol, and EKC WATER SPE 30 mmol L−1 phosphoric UV, 214 nm 278
ethynylestradiol in 10 min Spiked water C18; sweeping acid, 80 mmol L−1 SDS, 0.16–0.30 nmol L−1
>96% 20% MeOH
Pharmaceutical residues
Clofibric acid, naproxen, bezafibrate, CZE WATER LLE/SPE 20 mmol L−1 NH4Ac ESI-MS 82
diclofenac, ibuprofen, mefenamic River water hexane/MTBE (pH 5.1 adjusted with acetic single quadrupole
acid in ca. 20 min (50:50)/Bondesil acid) 25–59 µg L−1
ODS
43.1–98%
Tetracycline, oxytetracycline, CZE WATER Online SPE 50 mmol L−1 citric acid UV, 260 nm 272
doxycycline in 17.5 min Surface water STRATA-X (pH 2.5 adjusted with HCl 1.6–2 µg L−1
94–106% 1 mol L−1 )
Acetylsalicylic acid, ibuprofen, CZE WATER SPE-LVSEP 30 mmol L−1 tetraborate, UV, 214 nm 273
fenoprofen, naproxen, diclofenac, Bottled water LiChrolut RP-18 70% MeOH 2–9 ng L−1
ketoprofen in 26 min 64–95% (pH 9.8)
Fluoxetine, venlafaxine, citalopram, CZE WATER SPE 1.5 mol L−1 formic acid, ESI-TOF-MS 274
sertraline, paroxetine, River water hydrophilic 50 mmol L−1 ammonium 13–53 µg L−1
clomipramine, trazodone in 17 min SEWAGE divinylbenzene formate, 15% ACN (SPE-CE-MS)
85–99%
Clofibric acid, naproxen, bezafibrate CZE WATER SPE 20 mmol L−1 NH4Ac, ESI-MS 275
in 15 min Surface water LiChrolut RP-18 60% MeOH single quadrupole
SEWAGE 70–89% (surface (pH 4.5 adjusted with acetic 100 ng L−1
water) acid)
LLE/SPE
hexane/MTBE
(50:50)/LiChrolut
RP-18
50–58% (sewage)
Continued
957
958
TABLE 31.9
(Continued)
Preconcentration
Sulfapyridine, sulfamethazine, sulfamerazine, CZE WATER SPE-LVSS 45 mmol L−1 sodium phosphate, UV, 265 nm 276
sulfamether, sulfadiazine, sulfadimethoxine, Ground water Oasis HLB 10% MeOH (pH 7.3) 2.59–22.95 µg L−1
sulfamethoxazole, sulfachlorpyridazine, 75.6–100.3%
sulfamethizole in 18 min
Diclofenac, ibuprofen, fenoprofen, naproxen, EKC WATER SPE-SRMM, 25 mmol L−1 phosphate, UV, 214 nm 58
ketoprofen in 35 min Mineral water SRMM-ASEI, 75 mmol L−1 SDS, 0.07–0.23 µg L−1
FESI-RMM 40% ACN (SRMM)
LiChrolut RP-18 (pH 2.5 adjusted with HCl) 0.050–0.195 µg L−1
70–100% (SRMM-ASEI)
0.7–1.6 µg L−1
(FESI-RMM)
Miscellaneous
Anatoxin-a, microcystin-LR, cylindrospermopsin CZE WATER Centrifugation, 25 mmol L−1 sodium tetraborate UV, 230, 240, 278 nm 286
in 3 min Water bloom freezing–thawing, (pH 9.3) 0.89–2.77 mg L−1
Crude extracts filtering
Anatoxin-a, microcystin-LR, cylindrospermopsin EKC WATER Centrifugation, 25 mmol L−1 sodium tetraborate, UV, 230, 240, 278 nm 286
in 9 min Water bloom freezing–thawing, 100 mmol L−1 SDS (pH 9.3) 0.73–2.81 mg L−1
Crude extracts filtering
Pollen allergens and organic pollutants EKC AIR LLE 20 mmol L−1 TRIS, 5 mmol L−1 UV, 206 nm 288
(nonspecified); 20 compounds in 13 min Airborne dust samples acetone followed by H3 PO4 , 50 mmol L−1 SDS
acidic and alkaline (pH 8.7)
aqueous solutions
HP-β-CD, hydroxypropyl-β-cyclodextrin; CAPS, cyclohexylamino-1-propanesulfonic acid; LVSS, large-volume sample stacking; SRMM, stacking with reverse migrating micelles; SRMM–
ASEI, stacking with reverse migrating micelles–anion selective exhaustive injection; FESI–RMM, field-enhanced sample injection with reverse migrating micelles; LVSEP, large-volume
sample stacking using the electroosmotic flow pump; ESI-TOF-MS, electrospray ionization time of flight mass spectrometry; ESI-MS, electrospray ionization mass spectrometry; MSPD,
matrix solid-phase dispersion; PS–DVB, polystyrene–divinylbenzene; LLE, liquid–liquid extraction; SPE, solid-phase extraction.
system of man and animals. A few applications have been reported (Table 31.9).277,278 Figure 31.12
illustrates the electromigration separation of endocrine disruptors and pharmaceuticals.
31.4.12 MISCELLANEOUS
31.4.12.1 Humic Substances
CZE of natural organic matter279 and separation methods for humic substances280 have been sub-
jects of recent reviews. CZE and EKC separations (borate/TRIS/EDTA modified by CDs/SDS) and
(a)
40
35
5
30
25 4
20 1 3
AU
15
10 2
5
River water
0
5 Blank
0
5 10 15 20 25 30
40
5
35
30
4
25 3
AU
1
20
15
Effluent 2
10 waste water
5
0
Blank
−5
35 5 10 15 20 25 30
30 5
25 4
20 3
AU
1
15
10 Influent 2
5 waste water
0
5 Blank
0
0 5 10 15 20 25 30
FIGURE 31.12 Separation of (a) endocrine disruptors and (b) veterinary drug residues. Electrolytes:
(a) 20 mmol L−1 sodium tetraborate at pH 9.4; (b) 45 mmol L−1 sodium phosphate at pH 7.3 with 10% methanol.
Other conditions: (a) −30 kV; direct UV-detection at 210 nm; (b) +25 kV; direct UV-detection at 265 nm. Peak
legends: (a): 2,4,6-tribromophenol (1), pentabromophenol (2), 2,6-dibromophenol (3), tetrabromobisphenol A
(4), and tetrachlorobisphenol A (5). (b): Sulfapyridine (SPD), sulfamethazine (SMZ), sulfamerazine (SMR),
sulfamether (SMT), sulfadiazine (SDZ), sulfadimethoxine (SDM), sulfamethoxazole (SMX), sulfachloropyri-
dazine (SCP), sulfamethizole (SMI), and p-aminobenzoic acid (PABA). (Modified from (a) E. Blanco et al.,
J. Chromatogr. A, 1071, 205–211, 2005 and (b) J.J. Soto-Chinchilla et al., Electrophoresis, 27, 4360–4368,
(b)
20 PABA
SDZ SMX
SDM SMI
17.5
SMR SMT
15 SCP
Intensity (mAU)
SMZ
12.5
10 SPD
7.5
5
Spiked
2.5 ground water
0
2 4 6 8 10 12 14 16 18
Time (min)
characterization (TOF-MS) of humic acids isolated from Antarctica soil281 as well as interactions of
organic matter and environmental pollutants282,283 were reported (details in Table 31.9).
31.4.12.2 Algal Toxins

Methods for determining microcystins and microcystin-producing cyanobacteria have been
reviewed.284 Details of CZE and EKC analyses of cyanobacterial toxins in environmental
samples285–287 are compiled in Table 31.9. Figure 31.13 presents the separation of cyanobacterial
toxins.
31.4.12.3 Other Applications

CZE and MEKC methods were used to monitor extraction procedures involving pollen allergens and
organic pollutants from dust samples collected before, during, and after pollen seasons at different
locations (car-traffic tunnel in Prague and a metro station in Paris) using air-filtration devices.288
Water and acetic acid extracts were analyzed by CZE using acetic acid as background electrolyte
whereas water and alkaline water–SDS-buffer extracts were analyzed by MEKC in Tris-phosphate–
SDS electrolytes. Significant differences were found in the profiles of dust extracts from different
origins.
Drinking water and humidity condensate samples collected from U.S. Space Shuttle and the
Russian Mir Space Stations are analyzed routinely at the NASA–Johnson Space Center as a
means of verifying water quality and monitoring the environment of the spacecraft.289 Anions and
cations were determined by ion chromatography whereas carboxylates and amines were determined
by CE (phthalate/TTAB for carboxylates and imidazole/HIBA for amines). Results showed that
Shuttle water is of distilled quality whereas Mir-recovered water contains various levels of min-
erals. Organic ions were rarely detected in potable water samples but were present in humidity
condensates.
Table 31.9 compiles the details of the previously cited methodologies.
(a) (b)
20
a YA
a
15
10 YR
YR YA LR
LR
mAU
10
7
mAU
b b
5 4
0 1
2 2.5 3 3.5 4 4.5 2 4 6 8 10 12
FIGURE 31.13 (a) CZE and (b) MEKC separation of cyanobacterial toxins. Electrolytes: (a) 25 mmol L−1
borate at pH 9.3; (b) 25 mmol L−1 borate at pH 9.3 containing 75 mmol L−1 SDS. Other conditions: (a) and
(b) +25 kV; direct UV-detection at 238. Peak legends: microcystin YA (YA), microcystin YR (YR), and
microcystin LR (LR); a: water bloom sample; and b: spiked sample. (Modified from G. Vasas et al. J. Biochem.
Biophys. Methods, 66, 87–97, 2006. With permission.)
31.5 A METHOD DEVELOPMENT GUIDE FOR

ENVIRONMENTAL ANALYSIS
Figure 31.14 depicts a simplified diagram that can be useful during the first stages of method devel-
opment for environmental analysis. In CE, method development relies strongly on the knowledge
of the analyte structural features. For neutral compounds (for instance, many classes of pesticides,
PAH, PCB, dioxins, derivatized carbonyls, etc.), the obvious choice is the selection of an EKC
protocol. As a first attempt, MEKC in borate/SDS electrolytes is usually recommended. Depending
on the result of the exploratory run, several additives can be further tested as a means to modulating
the solute–micelle interaction. Other EKC modes are also feasible options, especially when SDS
fails in promoting separation, which is the case of either highly hydrophilic or highly hydrophobic
compounds. Occasionally, EKC has been the technique of choice for certain classes of ionizable
pollutants, such as phenol derivatives and amines, simply because the separation of a large number
of positional isomers and conformers is attempted and extra selectivity is required.
Method development procedures for charged compounds, either organic pollutants with ionizable
functional groups within the CE operational pH range (phenols, amines, carboxylic acids, many
dyes, pharmaceuticals, etc.) or explicitly ionic compounds such as small ions, sulfonates, LAS
surfactants, and so forth, will follow two distinct routes in the diagram of Figure 31.14, depending
on the compound’s UV-absorbing characteristics (only UV-absorbance detectors were considered
here because they are practically a part of all commercially available CE equipments).
If the pollutant is charged and exhibits UV-absorbing properties, the CZE mode is readily rec-
ommended. For basic pollutants, moderately low to low pH buffers are indicated and the analyte
migrates coelectroosmotically as a cation (protonated species) whereas for acidic pollutants, high pH
buffers will promote the analyte dissociation and it migrates counterelectroosmotically as an anion.
In both cases, buffer pH and concentration are the variables to optimize before the addition of any
modifiers is considered.
If the pollutant is charged but does not present a chromophoric group, either a cationic (separation
of cation) or an anionic (separation of anions) chromophore must be added to the electrolyte and
Nature of pollutant
NEUTRAL CHARGED
CHIRAL
(ionic or ionizable)
EKC Chiral
selectors
Cyclodextrin UV absorbing UV non-absorbing
MEKC Crown ether
Other
CZE CATIONIC ANIONIC
EXPLORATORY RUN
(borate/SDS) Low High
pH pH
Hydrophobic cpds
EXPLORATORY RUN
Hydrophilic cpds BASIC ACID (proper chromophore)
Organic solvents
Additives
Urea
Anionic
Cationic
EXPLORATORY RUN
surfactants Neutral surfactants (buffered systems)
EOF
Bile salts Additives reverser
Cationic
Organic solvents
OTHER EKC Amines
MODES Complexing
Other agent
FIGURE 31.14 A method development guide for environmental analysis.
indirect detection is performed. For the cations, complexing agents will further improve resolution
of similar mobility species. In the case of nonabsorbing anions, a judicious choice of both co- and
counterions is recommended, the co-ion due to peak symmetry considerations while the counterion
for signal enhancement. Furthermore, an EOF modifier (alkyl quaternary ammonium salt) is usually
selected to expedite the separation.
Finally, in the electrolyte design during chiral pollutants method development, a chiral selector
must always be considered. Cyclodextrins are among the most commonly used chiral additives
and can be employed in both EKC and CZE methodologies, depending on the nature of the
pollutant.

On the basis of the diversity of applications presented in this chapter, it becomes clear that CE
has found a niche among separation techniques for environmental analysis. Future directions will
likely include refinement of preconcentration strategies and further detector improvements to achieve
the desirable low-concentration limits of detection. In addition, with the consolidation of CE-MS
technology, more robust methods are likely to emerge with enhanced sensitivity and superior selec-
tivity, improving further the acceptance of CE in the routine determination of pollutants in real
samples.
ACKNOWLEDGMENTS
The authors wish to acknowledge the Conselho Nacional de Desenvolvimento Científico
e Tecnológico (CNPq) and the Fundação de Amparo à Pesquisa do Estado de São Paulo
(Fapesp) of Brazil for financial support (Fapesp 04/08503-2; 04/08931-4) and fellowships (CNPq
306068/2003-6).
REFERENCES
1. E. Dabek-Zlotorzynska and V. Celo, Recent advances in capillary electrophoresis and capillary
electrochromatography of pollutants, Electrophoresis, 27, 304–322, 2006.
2. E. Dabek-Zlotorzynska, H. Chen and L. Ding, Recent advances in capillary electrophoresis and
capillary electrochromatography of pollutants, Electrophoresis, 24, 4128–4149, 2003.
3. I. Ali, V.K. Gupta and H.Y. Aboul-Enein, Chiral resolution of some environmental pollutants by
4. D. Martínez, M.J. Cugat, F. Borrul and M. Calull, Solid-phase extraction coupling to capillary
electrophoresis with emphasis on environmental analysis, J. Chromatogr. A, 902, 65–89, 2000.
5. M.C. Bruzzoniti, C. Sarzanini and E. Mentasti, Preconcentration of contaminants in water analysis,
J. Chromatogr. A, 902, 289–309, 2000.
6. D. Martínez, F. Borrull and M. Calull, Strategies for determining environmental pollutants at trace
levels in aqueous samples by capillary electrophoresis, Trends Anal. Chem., 18, 282–291, 1999.
7. G.W. Sovocool, W.C. Brumley and J.R. Donnelly, Capillary electrophoresis and capillary electrochro-
matography of organic pollutants, Electrophoresis, 20, 3297–3310, 1999.
8. X.-P. Yan, X.-B. Yin, D.-Q. Jiang and X.-W. He, Speciation of mercury by hydrostatically mod-
ified electroosmotic flow capillary electrophoresis coupled with volatile species generation atomic
fluorescence spectrometry, Anal. Chem., 75, 1726–1732, 2003.
9. M.I. Turnes-Carou, P. López-Mahía, S. Muniategui-Lorenzo, E. Fernández-Fernández and
D. Prada-Rodríguez, Capillary zone electrophoresis for the determination of light-absorbing anions in
environmental samples, J. Chromatogr. A, 918, 411–421, 2001.
10. A. Cavallaro, V. Piangerelli, F. Nerini, S. Cavalli and C. Reschiotto, Selective determination of aromatic
amines in water samples by capillary zone electrophoresis and solid-phase extraction, J. Chromatogr. A,
709, 361–366, 1995.
11. S. Morales and R. Cela, Highly selective and efficient determination of US Environmental Protection
Agency priority phenols employing solid-phase extraction and non-aqueous capillary electrophoresis,
J. Chromatogr. A, 896, 95–104, 2000.
12. M. Molina, D. Peréz-Bendito and M. Silva, Multi-residue analysis of N-methylcarbamate pesticides
and their hydrolytic metabolites in environmental waters by use of solid-phase extraction and micellar
electrokinetic chromatography, Electrophoresis, 20, 3439–3449, 1999.
13. A. Segura-Carretero, C. Cruces-Blanco, S. Pérez Duran and A. Fernández Gutiérrez, Determination
of imidacloprid and its metabolite 6-chloronicotinic acid in greenhouse air by application of micellar
electrokinetic capillary chromatography with solid-phase extraction, J. Chromatogr. A, 1003, 189–195,
2003.
14. C.-E. Lin, C.-C. Hsueh, T.-Z. Wang, T.-C. Chiu and Y.-C. Chen, Migration behavior and separa-
tion of s-triazines in micellar electrokinetic capillary chromatography using a cationic surfactant,
J. Chromatogr. A, 835, 197–207, 1999.
15. O. Brüggemann and R. Freitag, Determination of polycyclic aromatic hydrocarbons in soil samples by
micellar electrokinetic capillary chromatography with photodiode-array detection, J. Chromatogr. A,
717, 309–324, 1995.
16. S. Frías, M.J. Sánchez and M.A. Rodríguez, Determination of triazine compounds in ground water
samples by micellar electrokinetic capillary chromatography, Anal. Chim. Acta, 503, 271–278, 2004.
17. S.J. Kok, E.H.M. Koster, C. Gooijer, N.H. Velthorst and U.A.Th. Brinkman, Separation of twenty-one
naphthalene sulfonates by means of capillary electrophoresis, J. High Resol. Chromatogr., 19, 99–104,
1996.
18. K. Heinig, C. Vogt and G. Werner, Determination of linear alkylbenzenesulfonates in industrial and
environmental samples by capillary electrophoresis, Analyst, 123, 349–353, 1998.
19. S. Angelino, A. Bianco Prevot, M.G. Genaro and E. Pramauro, Ion-interaction high-performance chro-
matography and micellar electrokinetic capillary chromatography: Two complementary techniques for
the separation of aromatic sulfonated compounds, J. Chromatogr. A, 845, 257–271, 1999.
20. E.R. Cunha and M.F. Alpendurada, Comparison of HPLC and micellar electrokinetic chromatogra-
phy in determination of sulfonated azo dyes in waste water, J. Liq. Chromatogr. Rel. Technol., 25,
1835–1854, 2002.
21. R.J.G. Jeevan, M. Bhaskar, R. Chandrasekar and G. Radhakrishan, Analysis of arylamine isomers by
micellar electrokinetic chromatography, Electrophoresis, 23, 584–590, 2002.
22. M. Wang, D. Wu, Q. Yao and X. Shen, Separation and selectivity in micellar electrokinetic chromatog-
raphy using sodium dodecyl sulfate micelles or Tween 20-modified mixed micelles, Anal. Chim. Acta,
519, 73–78, 2004.
23. L. Song, Q. Ou, W. Yu and G. Li, Separation of 6 phenylureas and chlorsulfuron standards by micellar,
mixed micellar and microemulsion electrokinetic chromatography, J. Chromatogr. A, 699, 371–382,
1995.
24. M.H. Chen and W.H. Ding, Separation and migration behavior of positional and structural naph-
thalenesulfonate isomers by cyclodextrin-mediated capillary electrophoresis, J. Chromatogr. A, 1033,
167–172, 2004.
25. E.A. Pereira, M.F.M. Tavares and A.A. Cardoso, Alternative methodologies for the determination of
aldehydes by capillary electrophoresis, J. AOAC Int., 82, 1562–1570, 1999.
26. M.-I. Acedo-Valenzuela, T. Galeano-Días, N. Mora-Díez and A. Silva-Rodríguez, Determination of
neutral and cationic herbicides in water by micellar electrokinetic capillary chromatography, Anal.
Chim. Acta, 519, 65–71, 2004.
27. E. Dabek-Zlotorzynska and E.P.C. Lai, Separation of polynuclear aromatic hydrocarbons by micel-
lar electrokinetic capillary chromatography using sodium taurodeoxycholate modified with organic
solvents, J. Cap. Elec., 3, 31–35, 1996.
28. W.C. Brumley and C.M. Brownrigg, Electrophoretic behavior of aromatic-containing organic acids and
the determination of selected compounds in water and soil by capillary electrophoresis, J. Chromatogr.,
646, 377–389, 1993.
29. W.C. Brumley, C.M. Brownrigg and A.H. Grange, Capillary liquid-chromatography-mass spec-
trometry and micellar electrokinetic chromatography as complementary techniques in environmental
analysis, J. Chromatogr. A, 680, 635–644, 1994.
30. J.M. Bermúdez-Saldaña, M.A. García, M.J. Medina-Hernández and M.L. Marina, Micellar elec-
trokinetic chromatography with bile salts for predicting ecotoxicity of aromatic compounds,
J. Chromatogr. A, 1052, 171–180, 2004.
31. J.M. Herrero-Martínez, E.F. Simó-Alfonso and G. Ramis-Ramos, Separation of homologues and
isomers of linear alkylbenzenesulfonates by capillary electrophoresis with sodium dodecyl sulfate,
carboxylic acids and bile salts, Electrophoresis, 24, 681–686, 2003.
32. J.M. Herrero-Martínez, E.F. Simó-Alfonso and G. Ramis-Ramos, Separation and determination of
homologues of linear alkylbenzenesulfonates by nonaqueous capillary zone electrophoresis using
alkylammonium salts in ethanol, Electrophoresis, 22, 2017–2024, 2001.
33. V. Bernabé-Zafón, S. Ortega-Gadea, E.F. Simó-Alfonso and G. Ramis-Ramos, Characterization and
quantitation of mixtures of alkyl ether sulfates and carboxylic acids by capillary electrophoresis with
indirect photometric detection, Electrophoresis, 24, 2805–2813, 2003.
34. V. Bernabé-Zafón, J.R. Torres-Lapasió, S. Ortega-Gadea, E.F. Simó-Alfonso and G. Ramis-Ramos,
Capillary electrophoresis enhanced by automatic two-way background correction using cubic smooth-
ing splines and multivariate data analysis applied to the characterisation of mixtures of surfactants,
J. Chromatogr. A, 1065, 301–313, 2005.
35. W. Wall, K. Chan and Z. El Rassi, Electrically driven microseparation methods for pesticides and
metabolites. VI. Surfactant-mediated electrokinetic capillary chromatography of aniline pesticidic
metabolites derivatized with 9-fluoroenylmethyl chloroformate and their detection by laser-induced
fluorescence, Electrophoresis, 22, 2320–2326, 2001.
36. J.T. Koch, B. Beam, K.S. Phillips and J.F. Wheeler, Hydrophobic interaction electrokinetic chro-
matography for the separation of polycyclic aromatic hydrocarbons using non-aqueous matrices,
J. Chromatogr. A, 914, 223–231, 2001.
37. W. Wall and Z. El Rassi, Electrically driven microseparation methods for pesticides and metabolites.
V. Micellar electrokinetic capillary chromatography of aniline pesticidic metabolites derivatized with
fluorescein isothiocyanate and their detection in real water at low levels by laser-induced fluorescence,
38. T.W. Moy, P.L. Ferguson, A.H. Grange, W.H. Matchett, V.A. Kelliber, W.C. Brumley, J. Glassman and
J.Q. Farley, Development of separation systems for polynuclear aromatic hydrocarbon environmental
contaminants using micellar electrokinetic chromatography with molecular micelles and free zone
39. H. Harino, S. Tsunoi, T. Sato and M. Tanaka, Applicability of micellar electrokinetic chromatography
with a double-chain surfactant having two sulfonate groups to the determination of pollutant phenols
in water, Anal. Sci., 16, 1349–1351, 2000.
40. S.K. Wiedmer, J. Hautala, J.M. Holopainen, P.K.J. Kinnunen, and M.-L. Riekkola, Study on liposomes
by capillary electrophoresis, Electrophoresis, 22, 1305–1313, 2001.
41. D. Kaniansky, E. Krčmová, V. Madajová, M. Masár, J. Marák and F.I. Onuska, Determination of
nitrophenols by capillary zone electrophoresis in a hydrodynamically closed separation compartment,
J. Chromatogr. A, 772, 327–337, 1997.
42. C. García-Ruiz, Y. Martín-Biosca, A.L. Crego and M.L. Marina, Rapid enantiomeric separation of
polychlorinated biphenyls by electrokinetic chromatography using mixtures of neutral and charged
cyclodextrin derivatives, J. Chromatogr. A, 910, 157–164, 2001.
43. R.S. Brown, J.H.T. Luong, O.H.J. Szolar, A. Halasz and J. Hawari, Cyclodextrin-modified capil-
lary electrophoresis: Determination of polycyclic aromatic hydrocarbons in contaminated soils, Anal.
Chem., 68, 287–292, 1996.
44. C.A. Groom and J.H.T. Luong, Sulfobutylether-beta-cyclodextrin-mediated capillary electrophore-
sis for separation of chlorinated and substituted phenols, Electrophoresis, 18, 1166–1172,
1997.
45. B. Fogarty, E. Dempsey and F. Regan, Potential of microemulsion electrokinetic chromatography
for the separation of priority endocrine disrupting compounds, J. Chromatogr. A, 1014, 129–139,
2003.
46. D. Shobat and E. Grushka, Use of calixarenes to modify selectivities in capillary electrophoresis,
Anal. Chem., 66, 747–750, 1994.
47. L. Arce, A. Segura-Carretero, A. Rios, C. Cruces, A. Fernández and M. Valcárcel, Use of calixarene
compounds as selectivity modifiers in capillary electrophoresis separations, J. Chromatogr. A, 816,
243–249, 1998.
48. K. Bächmann, B. Göttlicher, I. Haag, K.I. Han, W. Hensel and A. Mainka, Capillary electrokinetic
chromatography with a suspension of chromatographic particles, J. Chromatogr. A, 688, 283–292,
1994.
49. B. Göttlicher and K. Bächmann, Application of particles as pseudo-stationary phases in electrokinetic
chromatography, J. Chromatogr. A, 780, 63–73, 1997.
50. S. Terabe and T. Isemura, Ion-exchange electrokinetic chromatography with polymer ions for the
separation of isomeric ions having identical electrophoretic mobilities, Anal. Chem., 62, 650–652,
1990.
51. J.S. Fritz, M.C. Breadmore, E.F. Hilder and P.R. Haddad, Use of ionic polymers as stationary
and pseudo-stationary phases in the separation of ions by capillary electrophoresis and capillary
electrochromatography, J. Chromatogr. A, 942, 11–32, 2002.
52. Y. Takagai and S. Igarashi, Homogeneous liquid-liquid extraction and micellar electrokinetic chro-
matography using sweeping effect concentration system for determination of trace amounts of several
polycyclic aromatic hydrocarbons, Anal. Bioanal. Chem., 373, 87–92, 2002.
53. E. Turiel, P. Fernández, C. Pérez-Conde and C. Cámara, Trace-level determination of triazines and
several degradation products in environmental waters by disk solid-phase extraction and micellar
electrokinetic chromatography, J. Chromatogr. A, 872, 299–307, 2000.
54. P. Zhang, G. Xu, J. Xiong, Y. Zheng, Q. Yang and F. Wei, Determination of arsenic species by capillary
electrophoresis with large-volume field-amplified stacking injection, Electrophoresis, 22, 3567–3572,
2001.
55. Y. He and H.-K. Lee, Combination of solid-phase extraction and field-amplified concentration for trace
analysis of organonitrogen pesticides by micellar electrokinetic chromatography, Electrophoresis, 18,
2036–2041, 1997.
56. T. Hirokawa, T. Ichihara, K. Ito and A.R. Timerbaev, Trace ion analysis of seawater by cap-
illary electrophoresis: Determination of iodide using transient isotachophoresis preconcentration,
57. K. Otsuka, H. Hayashibara, S. Yamauchi, J.P. Quirino and S. Terabe, Highly sensitive micellar
electrokinetic chromatographic analysis of dioxin-related compounds using on-line concentration,
J. Chromatogr. A, 853, 413–420, 1999.
58. A. Macià, F. Borrull, M. Calull and C. Aguilar, Different sample stacking strategies to analyze some
nonsteroidal anti-inflammatory drugs by micellar electrokinetic capillary chromatography in mineral
waters, J. Chromatogr. A, 1117, 234–245, 2006.
59. S. Takeda, A. Omura, K. Chayama, H. Tsuji, K. Fukushi, M. Yamane, S.-I. Wakida, S. Tsub-
ota and S. Terabe, Separation and on-line concentration of bisphenol A and alkylphenols by
micellar electrokinetic chromatography with anionic surfactant, J. Chromatogr. A, 1014, 103–107,
2003.
60. J.P. Quirino, N. Inoue and S. Terabe, Reversed migration micellar electrokinetic chromatography with
off-line and on-line concentration analysis of phenylurea herbicides, J. Chromatogr. A, 892, 187–194,
2000.
61. C.L. Silva, E.C. Lima and M.F.M. Tavares, Investigation of preconcentration strategies for the trace
analysis of multiresidue pesticides in real samples by capillary electrophoresis, J. Chromatogr. A, 1014,
109–116, 2003.
62. K. Isoo and S. Terabe, Analysis of metal ions by sweeping via dynamic complexation and cation-
selective exhaustive injection in capillary electrophoresis, Anal. Chem., 75, 6789–6798, 2003.
63. O. Núñez, J.B. Kim, E. Moyano, M.T. Galcerán and S. Terabe, Analysis of the herbicides paraquat,
diquat and dibenzoquat in drinking water by micellar electrokinetic chromatography using sweeping
and cation selective exhaustive injection, J. Chromatogr. A, 961, 65–75, 2002.
64. J.P. Quirino, Y. Iwai, K. Otsuka and S. Terabe, Determination of environmentally relevant aromatic
amines in the ppt levels by cation selective exhaustive injection-sweeping-micellar electrokinetic
chromatography, Electrophoresis, 21, 2899–2903, 2000.
65. P. Hinsmann, L. Arce, A. Ríos and M. Valcárcel, Determination of pesticides in water by automation
online solid-phase extraction-capillary electrophoresis, J. Chromatogr. A, 866, 137–146, 2000.
66. M.C. Breadmore, A.S. Palmer, M. Curran, M. Macka, N. Avdalovic and P.R. Haddad, On-column
ion-exchange of inorganic anions in open tubular capillary electrochromatography with elution using
transient-isotachophoresis gradients. 3. Implementation and method development, Anal. Chem., 74,
2112–2118, 2002.
67. J.F. Huertas-Pérez, M. del Olmo Iruela, A.M. García-Campaña, A. González-Casado and A. Sánchez-
Navarro, Determination of the herbicide metribuzin and its major conversion products in soil by
micellar electrokinetic chromatography, J. Chromatogr. A, 1102, 280–286, 2006.
68. A. Mainka and K. Bächmann, UV detection of derivatized carbonyl compounds in rain samples in
capillary electrophoresis using sample stacking and a Z-shaped flow cell, J. Chromatogr. A, 767,
241–247, 1997.
69. J. Kuijt, C. García-Ruiz, G.J. Stroomberg, M.L. Marina, F. Ariese, U.A.Th. Brinkman and C. Gooijer,
Laser-induced fluorescence detection at 266 nm in capillary electrophoresis. Polycyclic aromatic
hydrocarbon metabolites in biota, J. Chromatogr. A, 907, 291–299, 2001.
70. S. Chen, Y. Xu, Y. Bi, W. Du and B.-F. Liu, Analysis of environmental pollutants metabolized from
pesticides using capillary electrophoresis with multiphoton-excited fluorescence detection, Talanta,
70, 63–67, 2006.
71. E. Dabek-Zlotorzynska and W. Maruszak, Determination of dimethylamine and other low-
molecular mass amines using capillary electrophoresis with laser-induced fluorescence detection,
J. Chromatogr. B, 714, 77–85, 1998.
72. M. Molina and M. Silva, In-capillary derivatization and analysis of amino acids, amino phospho-
nic acid-herbicides and biogenic amines by capillary electrophoresis with laser-induced fluorescence
detection, Electrophoresis, 23, 2333–2340, 2002.
73. S.-J. Chen, M.-J. Chen and H.-T. Chang, Light-emitting diode-based indirect fluorescence detection
for simultaneous determination of anions and cations in capillary electrophoresis, J. Chromatogr. A,
1017, 215–224, 2003.
74. S.Y. Chang and C.-H. Liao, Analysis of glyphosate, glufosinate and aminomethylphosphonic acid
by capillary electrophoresis with indirect fluorescence detection, J. Chromatogr. A, 959, 309–315,
2002.
75. Y.-M. Liu and J.-K. Cheng, Separation of biogenic amines by micellar electrokinetic chromatography
with on-line chemiluminescence detection, J. Chromatogr. A, 1003, 211–216, 2003.
76. F.R. Rocha, J.A.F. da Silva, C.L. do Lago, A. Fornaro and I.G.R. Gutz, Wet deposition and related
atmospheric chemistry in the Sao Paulo metropolis, Brazil; Part 1. Major inorganic ions in rainwater
as evaluated by capillary electrophoresis with contactless conductivity detection, Atmos. Environ., 37,
105–115, 2003.
77. A. Fornaro and I.G.R. Gutz, Wet deposition and related atmospheric chemistry in the Sao Paulo
metropolis, Brazil; Part 2. Contribution of formic and acetic acids, Atmos. Environ., 37, 117–128,
2003.
78. P. Kubáň, B. Karlberg, P. Kubáň and V. Kubáň, Application of a contactless conductometric detector
for the simultaneous determination of small anions and cations by capillary electrophoresis with
dual-opposite end injection, J. Chromatogr. A, 964, 227–241, 2002.
79. M. Chicharro, A. Zapardiel, E. Bermejo and A. Sánchez, Simultaneous UV and electrochemical
determination of the herbicide asulam in tap water samples by micellar electrokinetic capillary
chromatography, Anal. Chim. Acta, 469, 243–252, 2002.
80. A. Asthana, D. Bose, A. Durgbanshi, S.K. Sanghi and W.Th. Kok, Determination of aromatic
amines in water samples by capillary electrophoresis with electrochemical and fluorescence detection,
J. Chromatogr. A, 895, 197–203, 2000.
81. M. van Bruijnsvoort, S.K. Sanghi, H. Poppe and W.Th. Kok, Determination of chlorophenols by micel-
lar electrokinetic chromatography with electrochemical detection, J. Chromatogr. A, 757, 203–223,
1997.
82. W. Ahrer, E. Scherwenk and W. Buchberger, Determination of drug residues in water by the com-
bination of liquid chromatography or capillary electrophoresis with electrospray mass spectrometry,
J. Chromatogr. A, 910, 69–78, 2001.
83. C.A. Groom, A. Halasz, L. Paquet, S. Thiboutot, G. Ampleman and J. Hawari, Detection of nitroaro-
matic and cyclic nitramine compounds by cyclodextrin assisted capillary electrophoresis quadrupole
ion trap mass spectrometry, J. Chromatogr. A, 1072, 73–82, 2005.
84. T. Poiger, S.D. Richardson and G.L. Baughman, Identification of reactive dyes in spent dyebaths and
wastewater by capillary-electrophoresis-mass spectrometry, J. Chromatogr. A, 886, 271–282, 2000.
85. O. Núñez, E. Moyano and M.T. Galceran, Capillary electrophoresis-mass spectrometry for the analysis
of quaternary ammonium herbicides, J. Chromatogr. A, 974, 243–255, 2002.
86. Y. Iinuma and H. Herrmann, Method development for the analysis of particle phase substituted
methoxy phenols and aromatic acids from biomass burning using capillary electrophoresis/electrospray
ionization mass spectrometry (CE/ESI/MS), J. Chromatogr. A, 1018, 105–115, 2003.
87. J. Riu, P. Eichhorn, J.A. Guerrero, Th.P. Knepper and D. Barceló, Determination of linear alkylben-
zenesulfonates in wastewater treatment plants and coastal waters by automated solid-phase extraction
followed by capillary electrophoresis-UV detection and confirmation by capillary electrophoresis-mass
spectrometry, J. Chromatogr. A, 889, 221–229, 2000.
88. M. Molina, S.K. Wiedmer, M. Jussila, M. Silva and M.L. Riekkola, Use of partial filling technique and
reverse migrating micelles in the study of N-methylcarbamate pesticides by micellar electrokinetic
chromatography-electrospray ionization mass spectrometry, J. Chromatogr. A, 24, 191–202, 2001.
89. F. Menzinger, P. Schmitt-Kopplin, M. Frommberger, D. Freitag and A. Kettrup, Partial-filling micellar
electrokinetic chromatography and nonaqueous capillary electrophoresis for the analysis of selected
agrochemicals, Fresenius J. Anal. Chem., 371, 25–34, 2001.
90. Y.-R. Chen, M.-C. Tseng, Y.-Z. Chang and G.-R. Her, A low-flow CE/electrospray ionization MS
interface for capillary zone electrophoresis, large-volume sample stacking, and micellar electrokinetic
chromatography, Anal. Chem., 75, 503–508, 2003.
91. K. Isoo, K. Otsuka and S. Terabe, Application of sweeping to micellar electrokinetic chromatography-
atmospheric pressure chemical ionization-mass spectrometric analysis of environmental pollutants,
92. O.H.J. Szolar, Environmental and pharmaceutical analysis of dithiocarbamates, Anal. Chim. Acta,
582, 191–200, 2007.
93. L. Gámiz-Gracia, A.M. García-Campaña, J.J. Soto-Chinchilla, J.F. Huertas-Pérez and A. González-
Casado, Analysis of pesticides by chemiluminescence detection in the liquid phase, Trends Anal.
Chem., 24, 927–942, 2005.
94. V. Andreu and Y. Picó, Determination of pesticides and their degradation products in soil: Critical
review and comparison of methods, Trends Anal. Chem., 23, 772–789, 2004.
95. J. Hernández-Borges, S. Frías-García, A. Cifuentes and M.A. Rodrígues-Delgado, Pesticide analysis
by capillary electrophoresis, J. Sep. Sci., 27, 947–963, 2004.
96. Y. Picó, R. Rodriguez and J. Mañes, Capillary electrophoresis for the determination of pesticide
residues, Trends Anal. Chem., 22, 133–151, 2003.
97. Ph. Schmitt, T. Poiger, R. Simon, D. Freitag, A. Kettrup and A.W. Garrison, Simultaneous deter-
mination of ionization constants and isoelectric points of 12 hydroxy-s-triazines by capillary zone
electrophoresis and capillary isoelectric focusing, Anal. Chem., 69, 2559–2566, 1997.
98. M. Chicharro, A. Zapardiel, E. Bermejo and M. Moreno, Determination of 3-amino-1,2,4-triazole
(amitrole) in environmental waters by capillary electrophoresis, Talanta, 59, 37–45, 2003.
99. R. Loos and R. Niessner, Analysis of atrazine, terbutylazine and their N-dealkylated chloro and
hydroxyl metabolites by solid-phase extraction and gas chromatography mass spectrometry and
capillary electrophoresis-ultraviolet detection, J. Chromatogr. A, 835, 217–229, 1999.
100. H. Stutz, K. Pittertschatscher and H. Malissa, Jr., Capillary zone electrophoresis determination of
hydroxymetabolites of atrazine in potable water using solid-phase extraction with amberchrom resins,
Mikrochim. Acta, 128, 107–117, 1998.
101. J. You, M. Kaljurand and J.A. Koropchak, Direct determination of glyphosate in environmental waters
using capillary electrophoresis with electrospray condensation nucleation light scattering detection,
Intern. J. Environ. Anal. Chem., 83, 797–806, 2003.
102. J. Hernández-Borges, F.J. García-Montelongo, A. Cifuentes and M.A. Rodríguez-Delgado, Determi-
nation of herbicides in mineral and stagnant waters at ng/L levels using capillary electrophoresis and
UV detection combined with solid-phase extraction and sample stacking, J. Chromatogr. A, 1070,
171–177, 2005.
103. R.G. Wuilloud, M. Shah, S.S. Kannamkumarath and J.C. Altamirano, The potential of inductively
coupled plasma-mass spectrometric detection for capillary electrophoretic analysis of pesticides,
104. X. Song and W.L. Budde, Determination of chlorinated acid herbicides and related compounds in
water by capillary electrophoresis-electrospray negative ion mass spectrometry, J. Chromatogr. A,
829, 327–340, 1998.
105. R. Carabías-Martínez, E. Rodríguez-Gonzalo, J. Dominguez-Álvarez and J. Hernández-Méndez,
Determination of triazine herbicides in natural waters by solid-phase extraction and non-aqueous
capillary zone electrophoresis, J. Chromatogr. A, 869, 451–461, 2000.
106. N.V. Komarova and L.A. Kartsova, Factors responsible for the electrophoretic behavior of car-
boxylic acid and triazine derivatives under conditions of capillary zone electrophoresis and micellar
electrokinetic chromatography, J. Anal. Chem., 59, 662–668, 2004.
107. M. Chicharro, A. Zapardiel, E. Bermejo and A. Sanchez, Multiresidue analysis of pesticides in envi-
ronmental waters by capillary electrophoresis using simultaneous UV and electrochemical detection,
Electroanalysis, 16, 311–318, 2004.
108. R. Carabías-Martínez, E. Rodríguez-Gonzalo, P. Revilla-Ruiz and J. Domínguez-Alvarez, Solid-phase
extraction and sample stacking-micellar electrokinetic capillary chromatography for the determination
of multiresidue of herbicides and metabolites, J. Chromatogr. A, 990, 291–302, 2003.
109. C.W. Henry III, S.A. Shamsi and I.M. Warner, Separation of natural pyrethrum extracts using micellar
110. C.-E. Lin, Y.-C. Liu, T.-Y. Yang, T.-Z. Wang and C.-C. Yang, On-line concentration of s-triazine
herbicides in micellar electrokinetic chromatography using a cationic surfactant, J. Chromatogr. A,
916, 239–245, 2001.
111. Y.S. Wu, H.K. Lee and S.F.Y. Li, A fluorescence detection scheme for capillary electrophoresis of
N-methylcarbamates with on-column thermal decomposition and derivatization, Anal. Chem., 72,
1441–1447, 2000.
112. A. Karcher and Z. El Rassi, Electrically driven microseparation methods for pesticides and metabolites:
IV. Effects of the nature of fluorescent labels on the enantioseparation of pesticides and their degra-
dation products by capillary zone electrophoresis with UV and laser-induced fluorescence detection,
113. O. Núñez, E. Moyano, L. Puignou and M.T. Galcerán, Sample stacking with matrix removal
for the determination of paraquat, diquat and difenzoquat in water by capillary electrophoresis,
J. Chromatogr. A, 912, 353–361, 2001.
114. G. Dinelli, A. Vicari and A. Bonetti, Separation of sulfonylurea metabolites in water by capillary
electrophoresis, J. Chromatogr. A, 700, 195–200, 1995.
115. C. Klein, R.J. Schneider, M.T. Meyer and D.S. Aga, Enantiomeric separation of metolachlor and its
metabolites using LC-MS and CZE, Chemosphere, 62, 1591–1599, 2006.
116. J. Hernández-Borges, F.J. García-Montelongo, A. Cifuentes and M.A. Rodríguez-Delgado, Analy-
sis of triazolopyrimidine herbicides in soils using field-enhanced sample injection-coelectroosmotic
capillary electrophoresis combined with solid-phase extraction, J. Chromatogr. A, 1100, 236–242,
2005.
117. N.V. Komarova and L.A. Kartsova, Determination of herbicides of the chlorophenoxycarboxylic acid
type in natural and drinking water by capillary zone electrophoresis, Russian J. Appl. Chem., 76,
238–243, 2003.
118. M. Corbera, M. Hidalgo, V. Salvado and P.P. Wieczorek, Determination of glyphosate and
aminomethylphosphonic acid in natural water using capillary electrophoresis combined with enrich-
ment step, Anal. Chim. Acta, 540, 3–7, 2005.
119. H. Prosen, M. Guček and L. Zupančič-Kralj, Optimization of liquid chromatography and micellar
electrokinetic chromatography for the determination of atrazine and its first degradation products in
humic waters without preparation, Chromatographia, 60(Suppl. 60), S107–S112, 2004.
120. D.T. Kubilius and R.J. Bushway, Determination of hexazinone and its metabolites in groundwater by
capillary electrophoresis, J. Chromatogr. A, 793, 349–355, 1998.
121. M.B. Barroso, L.N. Honda and G. Morovjan, Multiresidue analysis of phenylurea herbicides in envi-
ronmental and drinking water by capillary electrophoresis, J. High Resol. Chromatogr., 22, 171–176,
1999.
122. J.L. Jarman, W.J. Jones, L.A. Howell and A.W. Garrison, Application of capillary electrophoresis to
study the enantioselective transformation of five chiral pesticides in aerobic soil slurries, J. Agric.
Food Chem., 53, 6175–6182, 2005.
123. R. Rodríguez, Y. Picó, G. Font and J. Mañes, Analysis of post-harvest fungicides by micellar
124. M. Chicharro, A. Zapardiel, E. Bermejo, J.A. Perez and A. Moreno, Multiresidue analysis of
S-triazines herbicides in environmental waters by micellar electrokinetic capillary chromatography,
J. Liq. Chromatogr. Rel. Technol., 24, 461–478, 2001.
125. X. Xu and R.J. Hurtubise, Separation of polycyclic aromatic hydrocarbon metabolites by gamma-
cyclodextrin-modified micellar electrokinetic chromatography, J. Liq. Chromatogr. Rel. Technol., 23,
1657–1670, 2000.
126. J. Li and J.S. Fritz, Nonaqueous media for separation of nonionic organic compounds by capillary
127. G. Kavran and F.B. Erim, Separation of polycyclic aromatic hydrocarbons with sodium dodecylben-
zenesulfonate in electrokinetic chromatography, J. Chromatogr. A, 949, 301–305, 2002.
128. G.K. Belin, E.B. Erim and F. Gulacar, Capillary electrokinetic separation of polycyclic aromatic
hydrocarbons using cetylpyridinium bromide, Polycyclic Arom. Compds., 24, 343–352, 2004.
129. H. Ren, X. Li, M. Qi, C. Stathakis and N.J. Dovichi, Capillary electrophoretic separation of polynuclear
aromatic hydrocarbons using sodium cholate in mixed aqueous-organic buffers, J. Chromatogr. A, 817,
307–311, 1998.
130. J.H.T. Luong, The combined effect of acetonitrile and urea on the separation of polycyclic aromatic
hydrocarbons using dioctyl sulfosuccinate in electrokinetic chromatography, Electrophoresis, 19,
1461–1467, 1998.
131. Y. Shi and J.S. Fritz, Capillary zone electrophoresis of neutral organic molecules in organic–aqueous
solution, J. High Resol. Chromatogr., 17, 713–718, 1994.
132. J.L. Miller, M.G. Khaledi and D. Shea, Separation of hydrophobic solutes by nonaqueous capillary
electrophoresis through dipolar and charge-transfer interactions with pyrylium salts, J. Microcolumn
Sep., 10, 681–685, 1998.
133. A.-L. Nguyen and J.H.T. Luong, Separation and determination of polycyclic aromatic hydrocar-
bons by solid-phase microextraction/cyclodextrin-modified capillary electrophoresis, Anal. Chem.,
69, 1726–1731, 1997.
134. C.P. Palmer, M.Y. Khaledi and H.M. McNair, A monomolecular pseudostationary phase for micellar
electrokinetic capillary chromatography, J. High Resol. Chromatogr., 15, 756–762, 1992.
135. D. Norton and S.A. Shamsi, Separation of methylated isomers of benzo[a]pyrene using micellar
electrokinetic chromatography, Anal. Chim. Acta, 496, 165–176, 2003.
136. C. Akbay, S.A. Shamsi and I.M. Warner, Separation of monomethyl-benz[a]antracene isomers using
cyclodextrin-modified electrokinetic chromatography, J. Chromatogr. A, 910, 147–155, 2001.
137. C. Fujimoto, Y. Fujise and S. Kawaguchi, Macromolecular surfactant as a pseudo-stationary phase in
micellar electrokinetic capillary chromatography, J. Chromatogr. A, 871, 415–425, 2000.
138. D.S. Peterson and C.P. Palmer, An anionic siloxane polymer as a pseudostationary phase for
electrokinetic chromatography, Electrophoresis, 21, 3174–3180, 2000.
139. N. Tanaka, T. Fukutome, K. Hosoya, K. Kimata and T. Anaki, Polymer-supported pseudo-stationary
phase for electrokinetic chromatography. Electrokinetic chromatography in a full range of methanol–
water mixtures with alkylated starburst dendrimers, J. Chromatogr. A, 716, 57–67, 1995.
140. J.M. Treubig Jr. and P.R. Brown, Novel approach to the analysis and use of fullerenes in capillary
141. B. Göttlicher and K. Bachmann, Investigation of the separation efficiency of hydrophobic compounds
in suspension electrokinetic chromatography, J. Chromatogr. A, 768, 320–324, 1997.
142. I. Benito, J.M. Saz, M.L. Marina, J. Jiménez-Barbero, M.J. González and J.C. Díez-Masa,
Micellar electrokinetic capillary chromatographic separation of polychlorinated biphenyl congeners,
J. Chromatogr. A, 778, 77–85, 1997.
143. M.L. Marina, I. Benito, J.C. Díez-Masa and M.J. González, Separation of chiral polychlorinated
biphenyls by micellar electrokinetic chromatography using β- and γ-cyclodextrin mixtures in the
separation buffer, J. Chromatogr. A, 752, 265–270, 1996.
144. W.C. Lin, F.C. Chang and C.H. Kuei, Separation of atropisomeric polychlorinated biphenyls by
cyclodextrin modified micellar electrokinetic chromatography, J. Microcolumn Sep., 11, 231–238,
1999.
145. A.L. Crego, M.A. García and M.L. Marina, Enantiomeric separation of chiral polychlorinated biphenyls
by micellar electrokinetic chromatography using mixtures of bile salts and sodium dodecyl sulphate
with and without gamma-cyclodextrin in the separation buffer, J. Microcolumn. Sep., 12, 33–40, 2000.
146. C. García-Ruiz, A.L. Crego and M.L. Marina, Comparison of charged cyclodextrin derivatives
for the chiral separation of atropisomeric polychlorinated biphenyls by capillary electrophoresis,
147. S.H. Edwards and S.A. Shamsi, Micellar electrokinetic chromatography of polychlorinated biphenyl
congeners using a polymeric surfactant as the pseudostationary phase, J. Chromatogr. A, 903, 227–236,
2000.
148. S.H. Edwards and S.A. Shamsi, Chiral separation of polychlorinated biphenyls using a combina-
tion of hydroxypropyl-gamma-cyclodextrin and a polymeric chiral surfactant, Electrophoresis, 23,
1320–1327, 2002.
149. A. Hilmi, J.H.T. Luong and A.-L. Nguyen, Applicability of micellar electrokinetic chromatography to
kinetic studies of photocatalytic oxidation of dibenzo-p-dioxin, Chemosphere, 36, 3113–3117, 1998.
150. J. Grainger, P.C. McClure, A. Liu, et al., Isomer identification of chlorinated dibenzo-p-dioxins by
orthogonal spectroscopic and chromatographic techniques, Chemosphere, 32, 13–23, 1996.
151. S. Terabe, Y. Miyashita, O. Shibata, E.R. Barnhart, L.R. Alexander, D.G. Patterson, B.L. Karger,
K. Hosoya and N. Tanaka, Separation of highly hydrophobic compounds by cyclodextrin-modified
152. J.T. Koch, B. Beam, K.S. Phillips and J.F. Wheeler, Hydrophobic interaction electrokinetic chro-
matography for the separation of polycyclic aromatic hydrocarbons using nonaqueous matrices,
J. Chromatogr. A, 914, 223–231, 2001.
153. C.-E. Lin, C.-C. Hsueh, W.-C. Lin and C.-C. Chang, Migration behavior and separation of
trichlorophenols by capillary zone electrophoresis, J. Chromatogr. A, 746, 295–299, 1996.
154. I. Canals, E. Bosch and M. Roses, Prediction of the separation of phenols by capillary zone
electrophoresis, Anal. Chim. Acta, 458, 355–366, 2002.
155. S. Takeda, Y. Tanaka, M. Yamane, Z. Siroma, S.-I. Wakida, K. Otsuka and S. Terabe, Ionization of
dichlorophenols for their analysis by capillary electrophoresis-mass spectrometry, J. Chromatogr. A,
924, 415–420, 2001.
156. O. Jáuregui, E. Moyano and M.T. Galceran, Capillary electrophoresis-electrospray ion-trap mass
spectrometry for the separation of chlorophenols, J. Chromatogr. A, 896, 125–133, 2000.
157. O. Jáuregui, L. Puignou and M.T. Galceran, New carrier electrolytes for the separation of chlorophenols
by capillary electrophoresis, Electrophoresis, 21, 611–618, 2000.
158. S. Morales and R. Cela, Capillary electrophoresis and sample stacking in non-aqueous media for the
analysis of priority pollutant phenols, J. Chromatogr. A, 846, 401–411, 1999.
159. S.P. Porras, R. Kuldvee, S. Palonen and M.-L. Riekkola, Capillary electrophoresis of methyl-
substituted phenols in acetonitrile, J. Chromatogr. A, 990, 35–44, 2003.
160. S.P. Porras, R. Kuldvee, M. Jussila, S. Palonen and M.-L. Riekkola, Capillary zone electrophoresis of
phenol and methylphenols at high pH in methanol, J. Sep. Sci., 26, 857–862, 2003.
161. J. Kronholm, P. Revilla-Ruiz, S.P. Porras, K. Hartonen, R. Carabías-Martínez and M.-L. Riekkola,
Comparison of gas chromatography-mass spectrometry and capillary electrophoresis in analysis of
phenolic compounds extracted from solid matrices with pressurized hot water, J. Chromatogr. A,
1022, 9, 2004.
162. S. Fu, F. Li, S.G. Chu and X.B. Xu, The determination of chlorophenols in waste water by capillary
zone electrophoresis with an organic modifier, Chromatographia, 56, 69–72, 2002.
163. X. Guo, Z. Wang and S. Zhou, The separation and determination of nitrophenol isomers by high-
performance capillary zone electrophoresis, Talanta, 64, 135–139, 2004.
164. S.S. Kannamkumarath, R.G. Wuilloud, S. Jayasinghe and J.A. Caruso, Fast speciation analysis
of iodophenol compounds in river waters by capillary electrophoresis-inductively coupled plasma-
mass spectrometry with off-line solid-phase microextraction, Electrophoresis, 25, 1843–1851,
2004.
165. W. Wall and Z. El Rassi, Electrically driven microseparation methods for pesticides and metabo-
lites. Part VII: Capillary electrophoresis and electrochromatography of underivatized phenol pesticidic
metabolites. Preconcentration and laser-induced fluorescence detection of dilute samples, J. Sep. Sci.,
25, 1231–1244, 2002.
166. A. Zemann and D. Volgger, Separation of priority pollutant phenols with coelectroosmotic capillary
electrophoresis, Anal. Chem., 69, 3243–3250, 1997.
167. S.M. Masselter and A.J. Zemann, Influence of organic solvents in coelectroosmotic capillary
electrophoresis of phenols, Anal. Chem., 67, 1047–1053, 1995.
168. M.R.N. Monton, J.P. Quirino, K. Otsuka and S. Terabe, Separation and on-line preconcentra-
tion by sweeping of charged analytes in electrokinetic chromatography with nonionic micelles,
J. Chromatogr. A, 939, 99–108, 2001.
169. M.R.N. Monton, K. Otsuka and S. Terabe, On-line sample preconcentration in micellar electroki-
netic chromatography by sweeping with anionic-zwitterionic mixed micelles, J. Chromatogr. A, 985,
435–445, 2003.
170. G. Jeevan, M. Bhaskar, R. Chandrasekar and G. Radhakrishnan, Separation of harmful chlorophe-
nols by cyclodextrin-assisted capillary electrokinetic chromatography, J. Sep. Sci., 25, 1143–1146,
2002.
171. P. Kuban, M. Berg, C. García and B. Karlberg, On-line flow sample stacking in a flow injection
analysis-capillary electrophoresis system: 2000-fold enhancement of detection sensitivity for priority
phenol pollutants, J. Chromatogr. A, 912, 163–170, 2001.
172. L. Gaillon, S. Cozette, J. Lelievre and R. Gaboriaud, New pseudo-stationary phases for electrokinetic
capillary chromatography. Complexes between bovine serum albumin and sodium dodecyl sulfate,
J. Chromatogr. A, 876, 169–182, 2000.
173. T. Zhao, X.B. Hu, J.K. Cheng and X.R. Lu, P-sulfonic calix[4]arene as running buffer additive in
electrokinetic chromatography, J. Liq. Chromatogr. Rel. Technol., 21, 3111–3124, 1998.
174. A.L. Gray and J.T. Hsu, Novel sulfonic acid-modified Starburst dendrimer used as a pseudostationary
phase in electrokinetic chromatography, J. Chromatogr. A, 824, 119–124, 1998.
175. B. Maichel, B. Potocek, B. Gaš and E. Kenndler, Capillary electrokinetic chromatography with
polyethyleneimine as replaceable cationic pseudostationary phase. Influence of methanol and
acetonitrile on separation selectivity, J. Chromatogr. A, 853, 121–129, 1999.
176. A.V. Pirogov and O.A. Shpigun, Application of water-soluble polymers as modifiers in electrophoretic
analysis of phenols, Electrophoresis, 24, 2099–2105, 2003.
177. T. Nakamura, A. Ohki, M. Mishiro, O. Tsuyashima and S. Maeda, Aggregate of amphiphilic
block copolymer as a pseudo-stationary phase in capillary electrophoresis, Anal. Sci., 15, 879–883,
1999.
178. A.V. Pirogov, A.V. Shpak and O.A. Shpigun, Application of polyelectrolyte complexes as novel pseudo-
stationary phases in MEKC, Anal. Bioanal. Chem., 375, 1199–1203, 2003.
179. J.T. Hautala, S.K. Wiedmer and M.-L. Riekkola, Anionic liposomes in capillary electrophoresis: Effect
of calcium on 1-palmitoyl-2-oleyl-sn-glycero-3-phosphatidylcholine/phosphatidylserine-coating in
silica capillaries, Anal. Bioanal. Chem., 378, 1769–1776, 2004.
180. A.L. Crego and M.L. Marina, Capillary zone electrophoresis versus micellar electrokinetic chromatog-
raphy in the separation of phenols of environmental interest, J. Liq. Chrom. Rel. Technol., 20, 1–20,
1997.
181. M. Czaplicka, Sources and transformations of chlorophenols in the natural environment, Sci. Total
Environ., 322, 21–39, 2004.
182. S. Kar and P.K. Dasgupta, Measurement of phenols on a loop-supported liquid film by micellar
electrokinetic chromatography and direct UV detection, J. Chromatogr. A, 379, 379–387, 1996.
183. W.H. Matchett and W.C. Brumley, Preconcentration of aliphatic amines from water determined by
capillary electrophoresis with indirect UV detection, J. Liq. Chromatogr. Rel. Technol., 20, 79–100,
1997.
184. A. Fekete, M. Frommberger, G. Ping, M.R. Lahaniatis, J. Lintelman, J. Fekete, I. Gebefugi, A.K. Malik,
A. Kettrup and P. Schmitt-Kopplin, Development of a capillary electrophoretic method for the analysis
of low-molecular-weight amines from metal working fluid aerosols and ambient air, Electrophoresis,
27, 1237–1247, 2006.
185. W. Maruszak, Analysis of aliphatic amines by micellar electrokinetic chromatography, J. High Resol.
Chromatogr., 22, 126–128, 1999.
186. X. Liu, L.-X. Yang and Y.-T. Lu, Determination of biogenic amines by 3-(2-furoyl)quinoline-
2-carboxaldehyde and capillary electrophoresis with laser-induced fluorescence detection,
J. Chromatogr. A, 998, 213–219, 2003.
187. K.B. Male and J.H.T. Luong, Derivatization, stabilization and detection of biogenic
amines by cyclodextrin-modified capillary electrophoresis-laser-induced fluorescence detection,
J. Chromatogr. A, 926, 309–317, 2001.
188. A. Bazzanella, H. Mörbel, K. Bächmann, R. Milbradt, V. Böhmer and W. Vogt, Highly efficient
separation of amines by electrokinetic chromatography using resorcarene-octacarboxylic acids as
pseudostationary phases, J. Chromatogr. A, 792, 143–149, 1997.
189. B.A. Musial, M.N. Martin and N.D. Danielson, Effect of an anionic polymer on the separation of
cationic molecules by capillary electrophoresis with conductivity detection, J. Sep. Sci., 25, 311–318,
2002.
190. H.M. Pinheiro, E. Toraud and O. Thomas, Aromatic amines from azo dye reduction: Status review with
emphasis on direct UV spectrophotometric detection in textile industry wastewaters, Dyes Pigments,
61, 121–139, 2004.
191. S. Oguri, Electromigration methods for amino acids, biogenic amines and aromatic amines,
J. Chromatogr. B, 747, 1–19, 2000.
192. H. Kataoka, Methods for the determination of mutagenic heterocyclic amines and their applications
in environmental analysis, J. Chromatogr. A, 774, 121–142, 1997.
193. S.-A. Leung and A.J. de Mello, Electrophoretic analysis of amines using reversed-phase, reversed-
polarity, head-column field-amplified sample stacking and laser-induced fluorescence detection,
J. Chromatogr. A, 979, 171–178, 2002.
194. G. Taibi, M.R. Schiavo, P.C. Rindina, R. Muratore and C.M.A. Nicotra, Micellar electrokinetic
chromatography of polyamines and monoacetylpolyamines, J. Chromatogr. A, 921, 323–329, 2001.
195. C.M. Knapp and J.J. Breen, Effects of tetraalkylammonium salts on the micellar electrokinetic
chromatography of aniline and substituted anilines, J. Chromatogr. A, 799, 289–299, 1998.
196. C.-E. Lin, C.-C. Chen, H.-W. Chen, H.-C. Huang, C.-H. Lin and Y.-C. Liu, Optimization of separation
and migration behavior of chloropyridines in micellar electrokinetic chromatography, J. Chromatogr. A,
910, 165–171, 2001.
197. S.D. Mendonsa and R.J. Hurtubise, Capillary electrophoresis (CE) methods for the separation of
carcinogenic heterocyclic aromatic amines, J. Liq. Chrom. Rel. Technol., 22, 1027–1040, 1999.
198. P. Zakaria, M. Macka and P.R. Haddad, Mixed-mode electrokinetic chromatography of aromatic bases
with two pseudo-stationary phases and pH control, J. Chromatogr. A, 997, 207–218, 2003.
199. D.W. Armstrong, L.W. Chang and S.S.C. Chang, Mechanism of capillary electrophoresis enantiosep-
arations using a combination of an achiral crown ether plus cyclodextrins, J. Chromatogr. A, 793,
115–134, 1998.
200. S.-L. Zhao, T.-Y. Dai, Z. Liu, F.-S. Wei, H.-F. Zou and X.-B. Xu, Determination of lower aliphatic
carbonyl compounds in stack gas as their 2,4-dinitrophenylhydrazones by micellar electrokinetic
chromatography, Chemosphere, 35, 2131–2136, 1997.
201. S. Takeda, S. Wakida, M. Yamane and K. Higashi, Analysis of lower aliphatic-aldehydes in
water by micellar electrokinetic chromatography with derivatization to 2,4-dinitrophenylhydrazones,
202. E.A. Pereira, A.A. Cardoso and M.F.M. Tavares, Determination of low-aliphatic aldehydes indoors by
micellar electrokinetic chromatography using sample dissolution manipulation for signal enhancement,
203. E.A. Pereira, E. Carrilho and M.F.M. Tavares, Laser-induced fluorescence and UV detection of
derivatized aldehydes in air sample using capillary electrophoresis, J. Chromatogr. A, 979, 409–416,
2002.
204. E.A. Pereira, M.O.O. Rezende and M.F.M. Tavares, Analysis of low molecular weight aldehydes in
air samples by capillary electrophoresis after derivatization with 4-hydrazinobenzoic acid, J. Sep. Sci.,
27, 28–32, 2004.
205. K. Fukushi, S. Takeda, K. Chayama and S.-I. Wakida, Application of capillary electrophoresis to the
analysis of inorganic ions in environmental samples, J. Chromatogr. A, 834, 349–362, 1999.
206. A.R. Timerbaev and K. Fukushi, Analysis of seawater and different highly saline natural waters by
capillary zone electrophoresis, Marine Chem., 82, 221–238, 2003.
207. N. Shakulashvili, T. Faller and H. Engelhardt, Simultaneous determination of alkali, alkaline earth and
transition metal ions by capillary electrophoresis with indirect UV detection, J. Chromatogr. A, 895,
205–212, 2000.
208. K. Fukushi, H. Ito, K. Kimura, K. Yokota, K. Saito, K. Chayama, S. Takeda and S.-I. Wakida,
Determination of ammonium in river water and sewage samples by capillary zone electrophoresis
with direct UV detection, J. Chromatogr. A, 1106, 61–66, 2006.
209. E. Dabek-Zlotorzynska, M. Kelly, H. Chen and C.L. Chakrabarti, Application of capillary elec-
trophoresis combined with a modified BCR sequential extraction for estimating of distribution of
selected trace metals in PM2.5 fractions of urban air particulate matter, Chemosphere, 58, 1365–1376,
2005.
210. E. Dabek-Zlotorzynska and R. Aranda-Rodriguez, Development and validation of capillary elec-
trophoresis for the determination of selected metal ions in airborne particulate matter after sequential
extraction, Anal. Bioanal. Chem., 372, 467–472, 2002.
211. Y.-S. Fung and H.-S. Tung, Capillary electrophoresis for trace metal ion analysis on environmental
studies, Electrophoresis, 20, 1832–1841, 1999.
212. A. Gáspár and E. Dudás, Application of internal universal calibration for determination of fully dis-
sociated species in capillary electrophoresis using indirect UV detection and electrokinetic injection,
J. Chromatogr. A, 1110, 254–260, 2006.
213. E. Yang and Z. Zhang, Simultaneous and fast detection of anions in snow using short tube by capillary
zone electrophoresis, Intern. J. Environ. Anal. Chem., 82, 353–360, 2002.
214. L. Wójcik, K. Korczak, B. Szostek and M. Trojanowicz, Separation and determination of perflu-
orinated carboxylic acids using capillary zone electrophoresis with indirect photometric detection,
J. Chromatogr. A, 1128, 290–299, 2006.
215. E. Dabek-Zlotorzynska, M. Piechowski, M. McGrath and E.P.C. Lai, Determination of low-
molecular-mass carboxylic acids in atmospheric aerosol and vehicle emission samples by capillary
216. E. Dabek-Zlotorzynska, R. Aranda-Rodrigues and L. Graham, Capillary electrophoresis determinative
and GC-MS confirmatory method for water-soluble organic acids in airborne particulate matter and
vehicle emission, J. Sep. Sci., 28, 1520–1528, 2005.
217. A. Mainka, P. Ebert, M. Kibler, T. Prokop, B. Tenberken and K. Bächmann, Development of new
methods for the analysis of carboxylic acids and carbonyl compounds in size classified raindrops by
CE for application in modeling atmospheric processes, Chromatographia, 45, 158–162, 1997.
218. A.G. Diress and C.A. Lucy, Study of the selectivity of inorganic anions in hydro-organic solvents using
indirect capillary electrophoresis, J. Chromatogr. A, 1085, 155–163, 2005.
219. J.E. Melanson, N.E. Baryla and C.A. Lucy, Dynamic capillary coatings for electroosmotic flow control
in capillary electrophoresis, Trends Anal. Chem., 20, 365–374, 2001.
220. V. Martínez, N. García, I. Antigüedad, R.M. Alonso and R.M. Jiménez, Capillary electrophoresis as
a useful tool for the analysis of chemical tracers applied to hydrological systems, J. Chromatogr. A,
1032, 237–242, 2004.
221. K. Ito, T. Ichibara, H. Zhuo, K. Kumamoto, A.R. Timerbaev and T. Hirokawa, Determina-
tion of trace iodide in seawater by capillary electrophoresis following transient isotachophoretic
preconcentration. Comparison with ion chromatography, Anal. Chim. Acta, 497, 67–74,
2003.
222. Z. Huang, K. Ito, A.R. Timerbaev and T. Hirokawa, Speciation studies by capillary electrophoresis—
Simultaneous determination of iodide and iodate in seawater, Anal. Bioanal. Chem., 378, 1836–1841,
2004.
223. S. Rovio, M. Mäntynen and H. Sirén, Determination of bromide and potassium in saline groundwaters
by capillary electrophoresis without prior dilution, Appl. Geochem., 19, 1331–1337, 2004.
224. T. Takayanagi, M. Ishida, J. Mbuna, R. Driouich and S. Motomizu, Determination of bromate ion in
drinking water by capillary zone electrophoresis with direct photometric detection, J. Chromatogr. A,
1128, 298–302, 2006.
225. C. Tu and H.K. Lee, Determination of nitrate in seawater by capillary zone electrophoresis with
chloride-induced sample self-stacking, J. Chromatogr. A, 966, 205–212, 2002.
226. K. Fukushi, Y. Nakayama and J.-I. Tsujimoto, Highly sensitive capillary zone electrophoresis with
artificial seawater as the background electrolyte and transient isotachophoresis as the on-line concen-
tration procedure for simultaneous determination of nitrite and nitrate in seawater, J. Chromatogr. A,
1005, 197–205, 2003.
227. M.A. Woodland and C.A. Lucy, Altering the selectivity of inorganic anion separations using
electrostatic capillary electrophoresis, Analyst, 126, 28–32, 2001.
228. Y. Deng, X. Fan, A. Delgado, C. Nolan, K. Furton, Y. Zuo and R.D. Jones, Separation and deter-
mination of aromatic acids in natural water with preconcentration by capillary zone electrophoresis,
J. Chromatogr. A, 817, 145–152, 1998.
229. P. Kubáň, P. Kubáň and V. Kubáň, Speciation of chromium (III) and chromium (VI) by capillary elec-
trophoresis with contactless conductometric detection and dual opposite end injection, Electrophoresis,
24, 1397–1403, 2003.
230. W.-P. Yang, Z.-J. Zhang and W. Deng, Simultaneous, sensitive and selective online chemiluminescence
determination of Cr(III) and Cr(VI) by capillary electrophoresis, Anal. Chim. Acta, 485, 169–177,
2003.
231. F. Han, J.L. Fasching and P.R. Brown, Speciation of organotin compounds by capillary electrophoresis
using indirect ultraviolet absorbance detection, J. Chromatogr. A, 669, 103–112, 1995.
232. C.L. Ng, H.K. Lee and S.F.Y Li, Determination of organolead and organoselenium compounds by
233. G. Forte, M. D’Amato and S. Caroli, Capillary electrophoresis speciation analysis of various arsenical
compounds, Microchem. J., 79, 15–19, 2005.
234. K.F. Akter, Z. Chen, L. Smith, D. Davey and R. Naidu, Speciation of arsenic in ground water samples:
A comparative study of CE-UV, HG-AAS and LC-ICP-MS, Talanta, 68, 406–415, 2005.
235. Z.L. Chen, J.-M. Lin and R. Naidu, Separation of arsenic species by capillary electrophoresis with
sample-stacking techniques, Anal. Bioanal. Chem., 375, 679–684, 2003.
236. Z. Chen and R. Naidu, Separation of sulfur species in water by co-electroosmotic capillary elec-
trophoresis with direct and indirect UV detection, Intern. J. Environ. Anal. Chem., 83, 749–759,
2003.
237. P. Kubáň, P. Kubáň and V. Kubáň, Rapid speciation of Se(IV) and Se(VI) by flow injection-capillary
electrophoresis system with contactless conductivity detection, Anal. Bioanal. Chem., 378, 378–382,
2004.
238. M. Takaya, Development of an analytical method for beryllium in airborne dust by micellar
239. M. Aguillar, A. Farran and V. Martí, Capillary electrophoretic determination of cyanide leaching
solutions from automobile catalytic converters, J. Chromatogr. A, 778, 397–402, 1997.
240. J.M. Robert and C.D. Spinks, Separation of metallated petroporphyrin models using micel-
lar electrokinetic capillary chromatography, J. Liq. Chrom. Rel. Technol., 20, 2979–2995,
1997.
241. C.A. Groom, A. Halasz, L. Paquet, P. D’Cruz and J. Hawari, Cuclodextrin-assisted capillary elec-
trophoresis for determination of the cyclic nitramine explosives RDX, HMS and CL-20. Comparison
with high-performance liquid chromatography, J. Chromatogr. A, 999, 17–22, 2003.
242. S.A. Oehrle, Analysis of nitramine and nitroaromatic explosives by capillary electrophoresis,
J. Chromatogr. A, 745, 233–237, 1996.
243. K.G. Hopper, H. LeChair and B.R. McCord, A novel method for analysis of explosives residue by
simultaneous detection of anions and cations via capillary electrophoresis, Talanta, 67, 304–312,
2005.
244. C.L. Copper and G.E. Collins, Separation of thiol and cyanide hydrolysis products of chemical warfare
agents by capillary electrophoresis, Electrophoresis, 25, 897–902, 2004.
245. E.W.J. Hooijschuur, A.E. Kientz and U.A.Th. Brinkman, Application of microcolumn liquid chro-
matography and capillary electrophoresis with flame photometric detection for the screening of
degradation products of chemical warfare agents in water and soil, J. Chromatogr. A, 928, 187–199,
2001.
246. M. Lagarrigue, A. Bossée, A. Bégos, A. Varenne, P. Gareil and B. Bellier, Separation and identification
of isomeric acidic degradation products of organophosphorus chemical warfare agents by capillary
electrophoresis-ion trap mass spectrometry, J. Chromatogr. A, 1137, 110–118, 2006.
247. T.E. Rosso and P.C. Bossle, Capillary ion electrophoresis screening of nerve agent degradation products
in environmental samples using conductivity detection, J. Chromatogr. A, 814, 125–134, 1998.
248. A.-E.F. Nassar, S.V. Lucas and L.D. Hoffland, Determination of chemical warfare agent degradation
products at low-part-per-billion levels in aqueous samples and sub-part-per-million levels in soils using
capillary electrophoresis, Anal. Chem., 71, 1285–1292, 1999.
249. M.-H. Chen and W.-H. Ding, Separation and migration of positional isomers and structural naph-
thalenesulfonate isomers by cyclodextrin-mediated capillary electrophoresis, J. Chromatogr. A, 1033,
167–172, 2004.
250. M.J. Cugat, F. Borrull and M. Calull, Separation of aromatic sulfonate compounds by coelectroosmotic
micellar electrokinetic chromatography, Chromatographia, 50, 229–234, 1999.
251. M.J. Cugat, F. Borrull and M. Calull, Comparative study of capillary zone electrophoresis and
micellar electrokinetic chromatography for the separation of twelve aromatic sulphonate compounds,
Chromatographia, 46, 204–208, 1997.
252. M.J. Cugat, F. Borrull and M. Calull, An overview of electrophoretic methods for the determination
of benzene- and naphthalenesulfonates in water samples, Trends Anal. Chem., 20, 487–499, 2001.
253. R. Loos and R. Niessner, Analysis of aromatic sulfonates in water by solid-phase extraction and
capillary electrophoresis, J. Chromatogr. A, 822, 291–303, 1998.
254. R. Loos, J. Riu, M.C. Alonso and D. Barceló, Analysis of polar hydrophilic aromatic sulfonates in
waste water treatment plants by CE/MS and LC/MS, J. Mass Spectrom., 35, 1197–1206, 2000.
255. S. Aubert and J.-P. Schwitzguébel, Capillary electrophoretic separation of sulphonated anthraquinones
in a variety of matrices, Chromatographia, 56, 693–697, 2002.
256. L. Farry, S.A. Oxspring, W.F. Smyth and R. Marchant, A study of the effects of injection mode,
on-capillary stacking and off-line concentration on the capillary electrophoresis limits of detection for
four structural types of industrial dyes, Anal. Chim. Acta, 349, 221–229, 1997.
257. T. Poiger, S.D. Richardson and G.L. Baughman, Analysis of anionic metallized azo and formazan dyes
by capillary electrophoresis-mass spectrometry, J. Chromatogr. A, 886, 259–270, 2000.
258. A.R. Fakhari, M.C. Breadmore, M. Macka and P.R. Haddad, Non-aqueous capillary electrophoresis
with red light emitting diode absorbance detection for the analysis of basic dyes, Anal. Chim. Acta,
580, 188–193, 2006.
259. W. Brumley and J.F. Farley, Determining eosin as a groundwater migration tracer by capil-
lary electrophoresis/laser-induced fluorescence using a multiwavelength laser, Electrophoresis, 24,
2335–2339, 2003.
260. S.M. Burkinshaw, D. Hinks and D.M. Lewis, Capillary zone electrophoresis in the analysis of dyes
and other compounds employed in the dye-manufacturing and dye-using industries, J. Chromatogr. A,
640, 413–417, 1993.
261. P. Blatny, C.-H. Fisher, A. Rizzi and E. Kenndler, Linear polymers applied as pseudo-phases in
capillary zone electrophoresis of azo compounds used as textile dyes, J. Chromatogr. A, 717, 157–166,
1995.
262. E. Blanco, M.C. Casais, M.C. Mejuto and R. Cela, Analysis of tetrabromobisphenol A and other
phenolic compounds in water samples by non-aqueous capillary electrophoresis coupled to photodiode
array ultraviolet detection, J. Chromatogr. A, 1071, 205–211, 2005.
263. E. Blanco, M.C. Casais, M.C. Mejuto and R. Cela, Comparative study of aqueous and non-aqueous
capillary electrophoresis in the separation of halogenated phenolic and bisphenolic compounds in
water samples, J. Chromatogr. A, 1068, 189–199, 2005.
264. E. Blanco, M.C. Casais, M.C. Mejuto and R. Cela, Approaches for the simultaneous extraction of
tetrabromobisphenol A, tetrachlorobisphenol A and related phenolic compounds from sewage sludge
and sediment samples based on matrix solid-phase dispersion, J. Chromatogr. A, 78, 2772–2778, 2006.
265. S. Takeda, A. Omura, K. Chayama, H. Tsuji, K. Fukushi, M. Yamane, S.-I. Wakida, S. Tsubota
and S. Terabe, Separation and on-line concentration of bisphenol A and alkylphenols by micellar
electrokinetic chromatography with cationic surfactant, J. Chromatogr. A, 979, 425–429, 2002.
266. S. Takeda, S. Ilida, K. Chayama, H. Tsuji, K. Fukushi and S. Wakida, Separation of bisphenol A and
three alkylphenols by micellar electrokinetic chromatography, J. Chromatogr. A, 895, 213–218, 2000.
267. F. Regan, A. Moran, B. Fogarty and E. Dempsey, Novel modes of capillary electrophoresis for the
determination of endocrine disrupting chemicals, J. Chromatogr. A, 1014, 141–152, 2003.
268. B. Fogarty, F. Regan and E. Dempsey, Separation of two groups of oestrogen mimicking compounds
using micellar electrokinetic chromatography, J. Chromatogr. A, 895, 237–246, 2000.
269. F. Regan, A. Moran, B. Fogarty and E. Dempsey, Development of comparative methods using
gas chromatography-mass spectrometry and capillary electrophoresis for determination of endocrine
disrupting chemicals in bio-solids, J. Chromatogr. B, 770, 243–253, 2002.
270. A. Goméz-Hens and M.P. Aguilar-Caballos, Social and economic interest in the control of phthalic
acid esters, Trends Anal. Chem., 22, 847–857, 2003.
271. T. Heberer, Occurrence, fate and removal of pharmaceutical residues in the aquatic environment:
A review of recent research data, Toxicol. Lett., 131, 5–17, 2002.
272. L. Nozal, L. Arce, B.M. Simonet, A. Rios and M. Valcárcel, Rapid determination of trace levels of
tetracyclines in surface water using a continuous flow manifold coupled to a capillary electrophoresis
system, Anal. Chim. Acta, 517, 89–94, 2004.
273. A. Macià, F. Borrull, C. Aguilar and M. Calull, Improving sensitivity by large-volume sample stacking
using the electroosmotic flow pump to analyze some nonsteroidal anti-inflammatory drugs by capillary
electrophoresis in water samples, Electrophoresis, 24, 2779–2787, 2003.
274. M. Himmelsbach, W. Buchberger and C.W. Klampfl, Determination of antidepressants in surface
and waste water samples by capillary electrophoresis with electrospray ionization mass spectrometric
detection after preconcentration using off-line solid-phase extraction, Electrophoresis, 27, 1220–1226,
2006.
275. A. Macià, F. Borrull, M. Calull and C. Aguilar, Determination of some acidic drugs in surface and
sewage treatment plant waters by capillary electrophoresis-electrospray ionization-mass spectrometry,
276. J.J. Soto-Chincilla, A.M. García-Campaña, L. Gámiz-Gracia and C. Cruces-Blanco, Application of
capillary zone electrophoresis with large-volume sample stacking to the sensitive determination of
sulfonamides in meat and ground water, Electrophoresis, 27, 4360–4368, 2006.
277. J.-B. Kim, K. Otsuka and S. Terabe, On-line sample concentration in micellar electrokinetic
chromatography with cationic micelles in a coated capillary, J. Chromatogr. A, 912, 343–352, 2001.
278. H. Harino, S. Tsunoi, T. Sato and M. Tanaka, Applicability of micellar electrokinetic chromatography
to the analysis of estrogens in water, Fresenius J. Anal. Chem., 369, 546–547, 2001.
279. P. Schmitt-Kopplin and J. Junkers, Capillary zone electrophoresis of natural organic matter,
J. Chromatogr. A, 998, 1–20, 2003.
280. P. Janoš, Separation methods in the chemistry of humic substances, J. Chromatogr. A, 983, 1–18, 2003.
281. D. Gajdošová, K. Novotná, P. Prošek and J. Havel, Separation and characterization of humic acids from
Antarctica by capillary electrophoresis and matrix-assisted laser desorption ionization time-of-flight
mass spectrometry. Inclusion complexes of humic acids with cyclodextrins, J. Chromatogr. A, 1014,
117–127, 2003.
282. M.L. Pacheco, E.M. Pena-Méndez and J. Havel, Supramolecular interaction of humic acids with
organic and inorganic xenobiotics studied by capillary electrophoresis, Chemosphere, 51, 95–108,
2003.
283. H. Prosen, S. Fingler, L. Zupančič-Kralj and V. Drevenkar, Partitioning of selected environmen-

tal pollutants into organic matter as determined by solid-phase microextraction, Chemosphere, 66,
1580–1589, 2007.
284. L.N. Sangolkar, S.S. Maske and T. Chakrabarti, Methods for determining microcystins (peptide
hepatotoxins) and microcystin-producing cyanobacteria, Water Res., 40, 3485–3496, 2006.
285. G. Vasas, D. Szydlowska, A. Gáspár, M. Welker, M. Trojanowicz and G. Borbély, Determination of
microcystins in environmental samples using capillary electrophoresis, J. Biochem. Biophys. Methods,
66, 87–97, 2006.
286. G. Vasas, A. Gáspár, C. Páger, G. Surányi, C. Máthé, M.M. Hamvas and G. Borbely, Analysis of
cyanobacterial toxins (anatoxin-a, cylindrospermopsin, microcystin-LR) by capillary electrophoresis,
287. A. Gago-Martínez, J.M. Leão, N. Piñeiro, E. Carballal, E. Vaquero, M. Nogueiras and J.A. Rodríguez-
Váquez, An application of capillary electrophoresis for the analysis of algal toxins from the aquatic
environment, Intern. J. Environ. Anal. Chem., 83, 443–456, 2003.
288. P. Sázelová, V. Kašiška, S. Koval, A. Prusík and G. Peltre, Evaluation of the efficiency of extraction of
ultraviolet-absorbing pollen allergens and organic pollutants from airborne dust samples by capillary
electrophoresis, J. Chromatogr. B, 770, 303–311, 2002.
289. D. Horta, P.D. Mudgett, L. Ding, M. Drybread, J.R. Schultz and R.L. Sauer, Analysis of water
from the Space Shuttle and Mir Space Station by ion chromatography and capillary electrophoresis,
J. Chromatogr. A, 804, 295–304, 1998.
Part IIIA
Microchip-Based: Core Methods and
Technologies
32 Cell Manipulation at the
Micron Scale
Thomas M. Keenan and David J. Beebe
CONTENTS
32.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 981

32.2 Microfabrication Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 982
32.3 Controlling Cell Position in 3D Cultures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 982
32.3.1 Physical Entrapment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 983
32.3.2 Photopatterning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 983
32.3.3 Dielectrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 984
32.3.4 Inkjet Printing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 986
32.4 Engineering the Cellular Microenvironment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 987
32.4.1 Mechanical Microenvironment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 988
32.4.1.1 Static Manipulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 988
32.4.1.2 Dynamic Manipulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 989
32.4.2 Chemical Microenvironment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 991
32.4.2.1 Gradient Generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 992
32.4.2.2 Subcellular Chemical Compartmentalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 993
32.4.3 Electrical Microenvironment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 994
32.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 997
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 998
32.1 INTRODUCTION
The advent of microfabrication technology and its more recent application to the field of biology has
greatly enhanced our ability to engineer in vitro cell culture environments. Cells and their sensing
elements are several hundred nanometers to tens of micrometers (microns) in size, and thus interact
within a microscale environment, or “microenvironment.” Unlike conventional cell culture meth-
ods, microfabrication-based methods allow manipulation of the physical, chemical, and electrical
properties of the cellular microenvironment at the microscale. As a result of this enhanced func-
tionality, microengineered culture environments are becoming more widely used in the biological
research community for a variety of applications, including investigating fundamental questions in
biology, enriching specific cell types from mixed populations, and examining the response of cells
to novel chemical compounds. Microengineered culture environments impose unique requirements
on analysis tools for detecting and characterizing changes in cell behavior or physiology in response
to specific environmental perturbations. Optimizing existing analysis tools for use with microengi-
neered culture environments would greatly enhance the utility of microfabrication-based methods
for biological studies. In this chapter, we provide a broad overview of the most common and recent
microfabrication-based methods for controlling the position of cells within a culture environment,
and the properties of the cellular microenvironment.
981
Photolithography Soft lithography
Negative Photoresist
photoresist Silicon features
wafer
PDMS
prepolymer
UV light
Mask
Photoresist
features Microchannels
FIGURE 32.1 Photolithography and soft lithography. Photolithography selective polymerizes a photoactive
polymer coating using UV light filtered by a patterned mask to create microscale features of defined dimensions.
Soft lithography utilizes the photolithographically defined substrates to mold polymer replicas of the microscale
features. (Reprinted and adapted from Li, N., et al., Crit Rev Biomed Eng, 31, 423. Copyright 2003. With
permission from Begell House, Inc.)
32.2 MICROFABRICATION TOOLS

Many microfabrication tools have been developed over the past 50 years to create more intricate
and advanced microelectronics. The ability of these tools to precisely control the size and shape
of polymers, metals, and other materials at scales as small as 1 µm, make them extremely adept
at creating unique cell culture microenvironments. Two processes called “photolithography” and
“soft lithography” are the most commonly used for biological studies.1,2 Photolithography begins by
coating a substrate with a thin layer of photosensitive polymer called a “photoresist” (Figure 32.1).
The thickness of the photoresist layer is controlled by spinning the substrate at a defined speed
and length of time. Thinner layers are produced by spinning the wafer faster and longer. The coated
substrate is then placed in contact with a thin sheet of glass or plastic that has user-defined opaque and
transparent regions, called a “mask.” Ultraviolet (UV) light is then shone through the mask onto the
coated substrate causing exposed regions to degrade (positive photoresist) or polymerize (negative
photoresist). The exposed substrate is then placed in a chemical solution called “developer,” which
removes the degraded or unpolymerized photoresist leaving user-defined features patterned on the
substrate. Soft lithography uses these patterned substrates, called “masters,” to mold elastomeric
polymer replicas. The most commonly used elastomer is poly(dimethyl siloxane) (PDMS), due to
its biocompatibility, transparency, and high-fidelity replica molding properties.
32.3 CONTROLLING CELL POSITION IN 3D CULTURES

The ability to control where a cell is positioned relative to its physical environment, other cells, or
regions with particular chemical identities is useful for studying cell exploration and migration in
the presence of various physical and chemical cues, investigating specific cell–cell or cell–substrate
interactions, characterizing autocrine or paracrine signaling, or for automating cell analysis. There is
an extensive body of literature describing methods to control the position of cells on two-dimensional
(2D) substrates.1,3,4 For some applications 2D culture systems are ideal; however, for many oth-
ers they are not. Although invaluable in developing our current understanding of many biological
phenomena, 2D culture systems are artificial environments not representative of in vivo culture
Cell Manipulation at the Micron Scale 983
(a) (b)
2D Culture 3D Culture
FIGURE 32.2 Comparison of 2D vs. 3D culture architectures. (a) 2D cultures consist of patterned or random
cells (shaded spheres) on a rigid or semirigid substrate. 2D cultures can offer cells a richer environment with
cell organization in three dimensions and incorporation of extracellular matrix components (black lines) with
entrained growth factors and signaling molecules. (b) Scanning electron micrograph (SEM) of the deep-sea
gulper shark retina. (Photo courtesy of Jill Olin, Hofstra University. With permission.)
environments except in a few distinct cases (e.g., endothelial or epithelial cells). Nearly all cells
found in vivo are completely surrounded by an intricate and highly organized arrangement of extra-
cellular matrix proteins, other cells, and a rich milieu of soluble biomolecules distributed throughout
the extracellular fluid. Cell–cell contact and the juxtaposition of different cells or cell types rela-
tive to one another are essential components of normal biological function. Cells communicate in a
variety of ways including integrin and gap junction signaling, or by secreting signaling proteins into
the extracellular environment that act on the cell itself (i.e., autocrine signaling) or on neighboring
cells (i.e., paracrine signaling). Two-dimensional cell cultures cannot fully replicate these intricate
and complex environments (Figure 32.2) and as a result may be limited in the insight and informa-
tion they can provide about normal and abnormal cell physiology. Three-dimensional (3D) culture
architectures better simulate complex in vivo environments and may be able to provide more com-
prehensive and relevant information about cell responses to specific environmental factors. Because
of the thorough description of 2D patterning methods in existing literature and the advantages of
3D culture systems, we will focus our discussion on efforts to control cell position in 3D culture
architectures.
32.3.1 PHYSICAL ENTRAPMENT

The simplest method of patterning cells in a 3D culture system is to create voids on the surface
of a biocompatible gel (Figure 32.3). Voids can be created by degrading select regions of the gel,5
embossing the gel with patterns defined by a microfabricated stamp,6 or by molding the gel on a
microfabricated surface.7,8 Cell patterning is achieved by seeding cells at random and rinsing away
those that have not fallen into the voids created in the gel. The voids can subsequently be refilled
with new gel to provide full encapsulation. Physical entrapment methods can be used with virtually
any soft or moldable polymer; however, its reliance on conventional photolithography allows cell
patterning in only two dimensions.
32.3.2 PHOTOPATTERNING
Cells can also be patterned in 3D culture constructs using a process called “photopatterning.” Pho-
topatterning (Figure 32.4) utilizes photoactive hydrogels that become cross-linked or polymerized
when exposed to light.9 Cells are mixed with prepolymerized gel solution and exposed to UV light
through a mask, similar to photolithography. Regions exposed to UV light polymerize and trap the
cells contained within. The unpolymerized solution can be rinsed away, replaced with acellular gel
solution, and flood exposed with UV light to provide a blank background. Alternatively, solutions
containing other cell types can be photopatterned relative to the first pattern to create a wide variety of
coculture architectures. Photopatterning is a very effective means of patterning cells in 3D cultures,
Degradation Embossing Molding
Hydrogel
Substrate
Seed cells
Rinse
Encapsulation
FIGURE 32.3 Physical entrapment cell patterning. Cells can be patterned in 3D hydrogel cultures by creating
voids using laser (L) degradation, embossing, and molding. Cells are seeded in the voids and those that do not
can be rinsed away. Addition of more hydrogel allows full 3D encapsulation.
even though cell position can only be controlled in two dimensions. However, its use of UV light to
initiate polymerization may have adverse effects on cell viability and function. UV light can damage
deoxyribonucleic acid (DNA) and creates polymer or cross linker radicals that may directly damage
cells.
32.3.3 DIELECTROPHORESIS
A phenomenon called dielectrophoresis has been used to pattern a variety of cell types on 2D
substrates10,11 and more recently in 3D culture constructs.9 Unlike electrophoresis where charged
species move in an applied electric field due to Coulombic forces (F = qE), dielectrophoresis capi-
talizes on the ability of a cell to become polarized when placed in an electric field. Dielectrophoresis
is most often used in conjunction with alternating current (AC) electric fields since AC fields elimi-
nate electrophoretic movement, and have less physiological impact on cells than direct current (DC)
fields.11 When a cell is placed in an AC field, the magnitude and polarity of the induced dipole
depend on the frequency of the applied field and the conductivities of the cell and the surrounding
medium, described by the equation

εc − εm
p(r) = 4πεm R3 , (32.1)
εc + 2ε m
(a) (b) (c)
UV
(d) (e)
FIGURE 32.4 Photopatterning. (a) Cells mixed with prepolymerized hydrogel solutions can be selectively
illuminated with UV light through the use of an applied photomask (shown in black). (b) Regions exposed to
UV become polymerized and entrap the cells. (c) The unpolymerized hydrogel can be rinsed away, replaced
with acellular prepolymerized hydrogel solution, and flood exposed with UV to provide a blank background.
(d) Photopatterning has been used to pattern 3T3 fibroblasts in PEG hydrogels against a blank background.
(e) Aligned with other photopatterned cell types for coculture experiments. (Adapted from Albrecht, D. R.,
et al., Lab Chip, 5, 111, 2005. With permission from The Royal Society of Chemistry.)
where p(r) is the induced dipole, R is the radius of the cell, ε is the complex permittivity√(ε =
ε + σ/(jω)) of the cell (c) or medium (m), σ is the conductivity of the cell or medium, j = − 1,
ω is the angular frequency of the applied electric field, and E(r) is the applied electric field.
If cells are placed in a nonuniform AC field, the spatial field gradient imparts different amounts of
force on each half of the cell dipole (Figure 32.5a). When the complex permittivity of the cell is greater
than that of the surrounding medium, the cell forms a positive dipole that causes it to move toward
the electrode with higher field strength in what is known as “positive dielectrophoresis” (pDEP). If
cell permittivity is less than that of the surrounding medium, a negative dipole is formed resulting in
repulsion away from the electrode of higher field strength, or “negative dielectrophoresis” (nDEP).
By engineering the shape and position of the electrodes, spatial field gradients can be created that
localize cells to precise, user-defined locations (Figure 32.5b). The magnitude and direction of cell
movement can be tuned by adjusting the conductivity of the surrounding medium or the frequency
of the applied AC field.
Since dielectrophoresis relies on the geometry of the electrodes to create the nonuniform electric
fields, the method has been limited to the creation of 2D cell patterns within 3D culture constructs
(Figure 32.5c).9 True 3D cell patterning would require more complex electrode geometries than
those employed to date. Although dielectrophoresis has been used to levitate cells12–14 and could
be used for 3D patterning, the method has not been utilized solely or in combination with any other
methods to pattern cells in all three dimensions within a 3D culture construct.
Because many biological hydrogels (e.g., collagen, fibrin, Matrigel™) have conductivities sim-
ilar enough to cells to make dielectrophoretic cell patterning difficult, synthetic hydrogels such as
poly(ethylene glycol) (PEG) are most commonly used.9 PEG hydrogels are biocompatible, but are
devoid of the numerous ligands and other supportive signals that cells normally receive from the extra-
cellular environment. Although there are numerous chemistries that can be incorporated into PEG
(a)
V pDEP nDEP
(b) (c)
FIGURE 32.5 Dielectrophoresis. (a) Nonuniform magnetic fields place different amounts of force on cells
with induced dipoles, causing them to move toward (pDEP) or away from (nDEP) high field densities. (b) Dielec-
trophoresis has been used to pattern 3T3 fibroblasts by controlling the shape and position of the electrodes.
(c) Dielectrophoresis has also been used to pattern 3T3 fibroblasts in 3D PEG hydrogels. (Adapted fromAlbrecht,
D. R., et al., Lab Chip, 5, 111, 2005. With permission from The Royal Society of Chemistry.)
hydrogels,15 the inability to fully recapitulate the natural environment may confound experimental
observations. Other confounding variables may be contributed by localized heating from the applied
electric field, which could damage cells or activate intracellular signaling cascades. The electric field
can also perturb the electrical state of the cell membrane and the function of membrane-embedded
ion channels.
32.3.4 INKJET PRINTING

Inkjet printing of biological solutions has been used to pattern a variety of biological molecules
on 2D substrates16 and could prove to be an efficient method for patterning cells in engineered
3D culture architectures. Biological inkjet printers are adapted conventional inkjet printers17,18 that
deposit one or more biological solutions in user-defined patterns under the control of a computer.
The computer controls the location of fluid ejection by rastering the printer head over the substrate
or by moving the substrate relative to a stationary printer head. The solutions are ejected from the
microfluidic channels and nozzles that constitute the inkjet printer heads using either heat-based or
piezoelectric fluid displacement. Heat-based fluid displacement, also known as “bubble jet” printing,
utilizes a heating element to vaporize a portion of the ink and create a bubble that ejects fluid from
the nozzle and onto the substrate (Figure 32.6a). When the bubble collapses it creates a vacuum that
pulls more ink into the printer head. Piezoelectric fluid displacement uses a piezoelectric transducer
to physically displace the fluid within the microchannel when a voltage is applied, and eject the fluid
out of the nozzle and onto the substrate (Figure 32.6b). Addition of a z-stepper motor allows 3D
(a) (b)
Resistive Piezo
heater Transducer
Bubble
(c)
FIGURE 32.6 Inkjet printing. (a) Bubble jet printer heads use resistive heating elements to vaporize the ink
and create a gas bubble that displaces fluid out of the nozzle. (b) Piezoelectric printer heads use a piezoelectric
transducer to mechanically displace fluid out of the nozzle. (c) Bovine aortic endothelial cells were inkjet printed
on an alginate-coated scaffold to create tubes 4 mm in diameter. (Reprinted from Varghese, D., et al., J Thorac
Cardiovasc Surg, 129, 2, 470. Copyright 2005. With permission from American Association for Thoracic
Surgery.)
cell culture constructs to be created using a layer-by-layer process. This method has been shown to
effectively create 3D cultures of neuronal cell lines in fibrin gels19 and a variety of cells in alginate
gels18 (Figure 32.6c). Because of the use of microfluidic channels in inkjet printer heads, care must
be taken to choose microchannel dimensions that do not shear damage cells or other biological
molecules within the printing solution. Minimization of shear effects often comes at the expense of
drop size resolution, which limits how close regions of different composition can be placed relative
to one another. In addition, the heating of biological ink when using bubble jet biological inkjet
printers may damage cells or cause denaturation of proteins.
32.4 ENGINEERING THE CELLULAR MICROENVIRONMENT

The ability to control the physical, chemical, and electrical properties of the in vitro cell culture
environment will perhaps be the single greatest contribution of microfabrication methods to the field
of biology. Engineering the cellular microenvironment can be used to create more physiologically
relevant culture conditions, study biological phenomena (e.g., growth, development, response to
damage, death), or to direct the behavior of cells toward specific industrial outcomes (e.g., tissue
engineering, chemical manufacturing). There are three ways in which the cellular microenvironment
can be altered—mechanically, chemically, or electromagnetically. Although most methods intend to
alter the cellular microenvironment in only one of these ways, it is often difficult to avoid secondary
perturbations to the cellular environment via one or both of the other modalities. Here, we review
the major methods for altering the microenvironment using microfabrication technology.
32.4.1 MECHANICAL MICROENVIRONMENT

32.4.1.1 Static Manipulations
Mechanical manipulations of the cell microenvironment can be either static or dynamic. Static
manipulations consist of modifications to the physical environment with which the cell interacts.
Altering the topography of the cell culture substrate is useful for understanding how cell growth and
proliferation is influenced by topographical or morphological cues, and how cells explore or migrate
in specific physical environments. Many studies have shown that topographical cues alone can alter
cell behavior.20–22 The recent development of nanofabrication methods23–27 has enabled the creation
of substrate topographies similar in size to those cells encounter in vivo, such as the protein fibers
that comprise the extracellular matrix (Figure 32.7).
A variety of different microfabrication methods have been used to create complex topographies
on cell culture substrates. Many of these studies20 require expensive fabrication tools and complex
methods and will not be discussed further here. The most common method of adding topography to
cell culture substrates is by using photolithography to either add features to the substrate or mask
the substrate for selective chemical or plasma etching. Alternatively, cell culture substrates can be
molded28–30 or embossed31 directly from photolithographically defined substrates. Photolithography
and soft lithography have been used to generate a wide range of cell culture substrate topographies28,32
including replicas of cells33 (Figure 32.8).
More complex 3D topographies can be generated using 3D patterning techniques.34 The sim-
plest method of generating 3D topographies is employing multiple iterations of photolithography
(Figure 32.9a). Unfortunately, multilayer photolithography requires precise alignment of each sub-
sequent layer and may result in poor feature resolution and fidelity. Generating multiple level
features in a single photolithography step can be accomplished using gray-scale photolithography
(Figure 32.9b). Gray-scale photolithography uses positive photoresists and masks printed with dif-
ferent gray levels35 (i.e., transparency) or masks made of microfluidic channels filled with dye.36
Although gray-scale photolithography is effective at making features of different heights, it can-
not make more complex features such as overhangs, closed loops, or hollow objects. These types of
complex 3D structures require methods such as multiphoton polymerization.37–39 Multiphoton poly-
merization uses long wavelength lasers, where each photon has half the energy necessary to activate
the photoinitiator in a photoactive polymer. When two beams are timed so that two long wavelength
photons arrive at a single photoinitiator molecule at the same time, the photoinitiator absorbs both
photons and is excited to its active state causing localized polymerization. By moving the focal point
of the two incident laser beams, complex 3D structures can be constructed (Figure 32.9c). These
features can subsequently be combined with soft lithography to form more intricate 3D structures
than those that can be produced with standard photolithography (Figure 32.9d–e).39
(a) (b)
FIGURE 32.7 Natural versus synthetic substrates. (a) SEM of the basement membrane of the rhesus macaque
urothelium. (From Figure 2 in Abrams, G. A., et al., Urol Res, 31, 341, 2003. Copyright 2003. With permission
from Springer Science and Business Media.) (b) Electrospun poly (lactide-co-glycolide) (PLGA) fibers (scale
bar = 10 µm). (Adapted from Li et al., J Biomed Mat Res, 60, 4, 613, 2002. With permission.)
(a) (b)
(c) (d)
FIGURE 32.8 Photo-/soft lithography generated topographies. (a) Bone marrow derived connective tis-
sue progenitor cells growing on PDMS channels molded from photolithographically defined masters. (From
Figure 5b in Mata, A., et al., Biomed Microdev, 4, 4, 267, 2002. Copyright 2002. With permission from Springer
Science and Business Media.) (b) Neurons growing on PDMS ridges with neurite bridge (arrow) spanning the
gap. (Reprinted from Goldner, J. S., et al., Biomaterials, 27, 460. Copyright 2006. With permission from Else-
vier.) (c) Fixed rat Schwann cells were used to create (d) PDMS replicas that were subsequently found to guide
neurite extension when seeded with dorsal root ganglion neurons (not shown). (Reprinted from Bruder, J. M.,
et al., Langmuir, 22, 20, 8266. Copyright 2006. With permission from American Chemical Society.)
(a) (b)
Photoresist
(c) (d) (e)
FIGURE 32.9 3D topographies. (a) A three-level master made with multiple iterations of photolithography.
Three mesas, with the center mesa taller than the other two, are connected by ridges 1 × 1 µm in cross
section (inset). (b) Multiple level features fabricated with gray-scale photolithography and microfluidic masks.
(Adapted from Chen, C., et al. Proc Natl Acad Sci USA, 100, 1499, 2003. With permission.) (c) A 10-µm
long, 7-µm tall bull made with two-photon photopolymerization in urethane acrylate polymer. (Reprinted from
Kawata, S., et al., Nature, 412, 6848, 697, Copyright 2001. With permission from Macmillan Publishers Ltd.)
(d–e) 3D PDMS structures molded from masters fabricated with multiphoton photopolymerization. (Adapted
from LaFratta, C. N., et al., Proc Natl Acad Sci USA, 103, 23, 8589, 2006. With permission.)
32.4.1.2 Dynamic Manipulations

Dynamic manipulations of the mechanical microenvironment consist of active elements that impose
mechanical force on cells. Many cell types experience mechanical forces in vivo. Chondrocytes
and osteoblasts within skeletal and load bearing connective tissues endure large compressive forces
(a) (b)
(c)
(d)
FIGURE 32.10 Deformable substrates. (a) Force (black arrows) can be used to bend the cell culture substrate
and mechanically strain attached cells. (b) The cell culture substrate can also be stretched, but due to the Poisson
effect this results in accessory stresses and strains perpendicular to the direction of applied force (gray arrows),
preventing mechanical loads from being truly uniaxial. (c–d) Microfabrication technology enhances the ways in
which cells can be mechanically strained. Cells could be cultured on thin PDMS membranes that are patterned
with a variety of unique topographies, and then mechanically strained when the membrane is deflected by
pressurizing underlying microfluidic channels. (Adapted from Hoffman, J. M., et al., Adv Mater, 16, 23, 2201,
during locomotion. Osteoblasts and endothelial cells experience large shear forces generated by fluid
flow within the osteon and vasculature, respectively. The ability to culture cells in environments
where different forms of mechanical loadings can be imposed provides invaluable insight into cell
physiology.
A variety of methods have been developed to apply mechanical forces to cells.40 One very com-
mon method is to culture cells on a deformable substrate that can be bent (Figure 32.10a) or stretched
(Figure 32.10b). Transducers provide controlled loading with user-defined strains and strain rates, or
cyclic loading with user-controlled frequency and amplitude. A recent advance in microfabrication
technology now allows cells to be grown on microtopographies while being dynamically loaded.
The method developed by Folch and coworkers41 employs thin PDMS membranes less than 20 µm
thick that are patterned with complex substrate topographies (Figure 32.10c–d). The membranes are
bonded to microfluidic channels connected to pressure or vacuum sources, allowing the membranes
to be deflected dynamically. The resulting concave or convex surfaces mechanically strain cells
interacting with the unique substrate topography.
Fluidic shear forces are commonly generated in vitro using cone and plate viscometers.42 A cone
and plate viscometer (Figure 32.11a) consists of a cone that rotates in close proximity to a flat
substrate. Constant shear stress is created in the region between the cone and the substrate by
engineering the taper of the cone to offset the influence of the increasing tangential velocity that
(a) (b)
L
Po PL
z
z
x
y (Po–PL) 2
V = r = z/(tan ) V=
2L
( H4 –z 2 )
= –(dV/dz) = –(dVdz)
= –(/tan ) (PL–Po)
=– z
L
FIGURE 32.11 Generating shear forces. (a) Cone and plate viscometers create constant shear stress throughout
the volume between the cone and substrate. The shear stress (τ ) is independent of the radius (r) or distance from
the cell culture surface (z), and dependent on the angular velocity (ω) of the cone and the angle of the taper (θ).
(b) Microfluidic channels have at least one dimension less than 1 mm creating a laminar flow regime restriction
that creates steady velocity profiles and shear stresses along the direction of fluid flow. The well-characterized
flow regime allows mathematical calculation of both the velocity profile (V ) and the shear stress distribution
across the channel (white graph perpendicular to fluid flow direction) using only the microchannel dimensions
(L and H) and the applied pressure gradient (PO and PL ). The negative sign convention indicates that shear
forces push down on the cells.
accompanies increasing radius on the shear rate. Cells cultured on the flat substrate are exposed to
constant shear forces and can be examined for behavioral or physiological effects.
Microfabrication technology provides a complementary method to expose cells to specific shear
stress environments. Straight microfluidic channels (Figure 32.11b) with highly precise dimensions
and controlled fluid flow rates have been used to study the effects of shear stress on cells.43,44
Microfluidic channels have one or more dimensions less than 1 mm, which forces flowing liquids to
be confined to laminar flow regimes.45 Laminar fluid flow is characterized by a parabolic velocity
profile that does not change downstream of a short flow stabilization region near the entrance of
the microchannel. The steady-state parabolic velocity profile produces constant shear stresses in the
direction of fluid flow. The velocity profile can be calculated from the fluid channel dimensions and
the applied pressure gradient, and be used to calculate the shear stress throughout the microchannel.
Because the velocity profiles near the junction of two microchannel walls can be more complex
to predict, cells are usually cultured near the middle of microchannels that are much wider than
they are tall.
32.4.2 CHEMICAL MICROENVIRONMENT

Chemical manipulation of the cellular microenvironment encompasses chemical modification of the
physical components with which the cell comes into contact (i.e., substrate, extracellular matrix, etc.),
and the distribution of soluble chemical species in the culture medium in which the cells are grown.
Numerous methods have been developed to modify the chemistry of cell culture substrates.1,46 Here
we focus our discussion on different microfluidic methods for controlling the distribution of soluble
chemical species in the cellular microenvironment. For this application, microfluidic devices offer
unique advantages due to their restriction to laminar fluid flow.45 In addition to having a constant,
parabolic velocity profile, laminar fluid flow is also characterized by more linear streamlines that
follow the contours of the microchannel. The fluid does not flow in circuitous, disorganized paths
as it does in turbulent flow. The lack of eddies and convective mixing results in a system where
chemical species can only mix via diffusion. This limitation allows creation of predictable chemical
gradients, which is not possible in conventional cell culture systems. The lack of convective mixing
also allows tight control over the location of different flowing streams within the microchannel,
which has proven useful for exposing specific regions of cells to defined chemical environments.
Here, we will discuss both gradient generators and subcellular fluidic compartmentalization using
microfluidic channels.
32.4.2.1 Gradient Generators

Gradients of diffusible biological molecules play critical roles in many biological phenomena includ-
ing development, cancer, inflammation, and wound healing. A wide variety of methods have been
developed to expose cells to soluble chemical gradients.47–52 Unfortunately, all of these methods
offer little temporal or spatial control over the gradient, and in many cases provide no way to quantify
the gradient that each cell perceives.
A device we will refer to as the “Premixer Device” (Figure 32.12a)53 has been used to expose
a variety of cells to biochemical gradients.54–56 The device generates steady-state, user-controlled
gradients using constant fluid flow. The device splits and recombines inlet fluids in an upstream
microfluidic mixer to generate a variety of different gradient shapes (Figure 32.12b). The use of
transparent PDMS to fabricate the device and the closed channel geometry allows direct quantification
of the gradient to which cells are exposed using a conventional wide field fluorescence microscope.
The only disadvantage of the device is that it exposes cells to fluid flow and shear forces that can
destroy autocrine or paracrine signaling, activate intracellular signaling cascades, and alter cell
migration.57
(a)
Coverslip
PDMS
Inlets PDMS
Gradient
across channel
Flow
Cell
inlet
Outlet Neutrophils Coverslip
(b)
Normalized intensity (%)
100
80
60
40
20
0
0 100 200 300 0 100 200 300 400 500 0 100 200 300
Channel width (M)
FIGURE 32.12 Premixer device. (a) An upstream microfluidic mixer splits and recombines fluids to expose
cells to user-defined gradients under constant fluid flow. (b) The premixer can be designed to generate a wide
variety of gradient shapes. (Adapted from Jeon, N. L., et al., Nat Biotechnol, 20, 8, 826. Copyright 2002. With
permission from Macmillan Publishers Ltd.)
(a) (b)
(c) (d)
1
Normalized
intensity
0.5
0
0
450
Po 150 300 300
si
res tion a 150
acro
ss
erv cro 450 0 ition
Pos rvoir,
oir, ss m
m rese
FIGURE 32.13 Microfluidic multi-injector. (a) 3D schematic representation of the device shows two small-
diameter channels under the control of microfluidic valves that pneumatically eject fluid into a cell culture
reservoir to form soluble molecule gradients. (b) Top view of the device in operation forming gradients of
fluorescein isothiocyanate (FITC)-conjugated dextran. (c) 3D plot of the fluorescence intensity within the cell
culture reservoir. (Adapted from Chung, B. G., et al., Lab Chip, 6, 6, 764, 2006. Reproduced with permission
from The Royal Society of Chemistry.)
Another recently developed microfluidic device replicates a method developed by Gunderson

and Barrett49 in a microfluidic platform. In the Microfluidic Multi-injector58 (Figure 32.13a), fluids
are ejected pneumatically out of a pair of small orifices into a larger cell culture reservoir, under the
control of independent microfluidic valves. Once the fluid is ejected into the reservoir, it forms a
diffusive gradient (Figure 32.13b–c). Although the device generates gradients without appreciable
fluid flow in the cell culture reservoir, as evidenced by a symmetric gradient, it cannot generate
steady-state gradients or the complex gradients of the Premixer Device.
A microfluidic generator called the “Microjets Device” (Figure 32.14a) generates stable gradients
on open surfaces without exposing cells to appreciable fluid flow.59 The open architecture allows
efficient nutrient and gas exchange, and the lack of flow minimizes shear forces and facilitates
autocrine and paracrine signaling. Like the previously described device, the Microjets Device gen-
erates gradients by pneumatically ejecting fluids out of small orifices, which minimize convective
flow and allow a diffusive gradient to form in the cell culture area. However, the Microjets Device
uses two opposed arrays of small orifices delivering different concentrations of biochemical factor
to generate steady-state chemical gradients. In addition, dynamic changes to the slope and position
of the gradient can be accomplished by adjusting the air pressures driving the fluid out of each orifice
array (Figure 32.14b–c). The Microjets Device is not capable of generating the complex gradients
possible with the Premixer Device nor can the gradient be visualized with a conventional wide field
fluorescence microscope. The architecture of the device requires confocal microscopy to evaluate
the gradient to which the cells are exposed.
32.4.2.2 Subcellular Chemical Compartmentalization

The laminar flow restriction and small dimensions of microfluidic devices have been exploited to
expose different portions of a single cell to different chemical environments. A method developed
by Whitesides and coworkers60,61 utilized hydrodynamic focusing62 on a microfluidic platform to
selectively label portions of a cell with different dyes (Figure 32.15). Laminar fluid flow confined
the dye solution to particular regions of the channel despite being in contact with streams containing
(a) Microjet arrays Open-surface

gradient
C=0
P = PL C=0
P = PR
(b) 1.00 16 (c) 16

P 1.00 PL PR
C/Cº /m × 1000
C/Cº /m × 1000

0 psi 1.2 0.8
+0.4 psi
0.8 1.2
+0.7 psi
0.4 1.6
C/Cº
0.50 8 C/Cº
0.50 8
0.00 0 0.00 0
0 50 100 150 200 0 50 100 150 200
Position (µm) Position (µm)
FIGURE 32.14 Microjets device. (a) 3D schematic representation of the device showing opposed arrays
of small orifices (i.e., small microfluidic channels) creating gradients in an open cell culture area. (b) Plot
of the concentration profile (solid lines, left axis) and profile derivatives (dots, right axis) for three different
pressure conditions illustrating how increasing the driving pressure delivered to the respective orifice arrays
by the same magnitude causes an increase in gradient slope with little change in the gradient position.
(c) Equal magnitude driving pressure offsets cause the gradient to shift position within the cell culture reservoir
but does not affect the gradient shape. (Reprinted from Thomas M., et al., Appl Phys Lett, 89, 114103, 2006.
Copyright 2006. With permission from American Institute of Physics.)
no dye. Cells within the microchannel were thus labeled only in the specific subcellular regions
exposed to a particular dye. Microfluidic hydrodynamic focusing has since been used to examine the
dynamics of epidermal growth factor (EGF) receptor signaling63 and the effects of focal application
of agrin in stimulating in vitro neuromuscular junction formation.64,65
Microfluidic channels can also be used to physically compartmentalize portions of a cell. Jeon and
coworkers66 replicated a Campenot67 chamber using microfluidic channels to provide subcellular
isolation of central nervous system growth cones from their cell bodies. Neurons seeded in a central
microchannel were allowed to extend axons through narrow restrictions to another large microchannel
(Figure 32.16). The restrictions prevent appreciable fluid connections from being established between
the two microchannels allowing the growth cones to be exposed to a different chemical environment
than the cell bodies.
32.4.3 ELECTRICAL MICROENVIRONMENT

Several important cell types are sensitive to the electrical characteristics of the cellular micro-
environment. Neurons, cardiac myocytes, and retinal cells all generate and can be stimulated with
electrical impulses. Because of the importance of these cell types, various methods have been
devised to record or stimulate electrical activity within cells cultured in vitro. Traditionally, the
electrical activity of cells has been recorded or stimulated by simply placing electrodes in the
(a) (b)
Inlets
PDMS Coverglass
Outlet
(c)
(d)
Cell
FIGURE 32.15 Hydrodynamic focusing. (a) A 3D schematic representation of the device shows how a fluid
stream containing dye was confined by adjacent flow streams to expose only certain subcellular regions of
the cell to the dye. (b) A bovine capillary endothelial cell labeled on one side with Mitotracker Green, and
(c) on the other side with FM Mitotracker Red CM-H2 XRos. (d) Phase micrograph of the cell overlayed with
fluorescence images from (b) and (c). (Adapted from Takayama, S., et al., Nature, 411, 1016, Copyright 2001.
With permission from Macmillan Publishers Ltd.)
(a) (b)
Axonal side
Somal side
20 mm
(c)
3 mm
100 µm 3 µm
FIGURE 32.16 Physical compartmentalization. (a) 3D schematic representation and 2D cross section of the
device shows how neurons loaded in the somal side of the device (black) are unable to cross through the narrow
constriction, unlike their axons, which can extend into the axonal side (white). (b) Texas red dextran injected in
the axonal side (right) does not cross into the somal side showing fluidic isolation between the compartments.
(c) When green cell tracker dye is loaded into the axonal side it retrograde labels the axons and somas of the
neurons. (Adapted from Taylor, A. M., et al., Nat Methods, 2, 8, 599. Copyright 2005. With permission from
Macmillan Publishers Ltd.)
culture medium, or alternatively, using a process known as “patch clamping” (Figure 32.17a). Patch
clamping uses glass capillaries that are heated and then pulled at a defined speed to form a drawn
tip approximately 1 µm in diameter. The drawn capillary, or micropipette, is then filled with a
conductive electrolyte solution and a metal electrode, and brought to the surface of a cell using
a micromanipulator. In this configuration the micropipette can be used to record the extracellular
electrical activity or stimulate cells. To measure single cell transmembrane ion currents, a small hole
is ripped in the cell membrane by placing the micropipette tip in contact with the cell membrane and
(a) (b)
+ +
V1 V2
– –
FIGURE 32.17 Micropipette electrical recording/stimulation. (a) Pulled glass pipettes are used to measure
the electrical activity of cells across the cell membrane (V1 ) by accessing the intracellular fluidic environment,
or extracellularly (V2 ) by simply placing the micropipette near the cell. (b) A major limitation of micropipette-
based electrical recording/stimulation is the inability to record from more than a few cells at a time due to space
limitations of the micropipette set up. (Adapted Fitzsimonds, R. M., et al., Nature, 388, 439. Copyright 1997.
With permission from Macmillan Publishers Ltd.)
gently pulling suction in the capillary. The cell membrane is pulled into the micropipette forming
a high-resistance seal, and ruptured within the micropipette to provide fluidic, and thus electrical,
access to the intracellular environment.
Although the quality of the recordings or stimulations is excellent using micropipettes, the method
suffers significantly from its lack of scalability. It is very difficult to record or stimulate electrical
activity in more than just a few cells due to space limitations (Figure 32.17b). Trying to comprehen-
sively characterize the electrical activity of cells within a tissue section or large dissociated culture
is not possible using individually placed micropipettes. It is also difficult to record from or stimulate
cells for more than a few hours due to biofouling or clogging of the micropipette tip by proteins
and cell components, and the instability of any seal formed between the micropipette and the cell
membrane.
Microelectrode arrays (MEA) offer a solution to both the scalability and long-term recording
limitations of micropipette electrodes. MEAs (Figure 32.18) are made by depositing a conductive
metal on a cell culture substrate, patterning the surface with photolithography, and using the pho-
toresist pattern as a mask for subsequent metal etching. After selective metal etching, the photoresist
is stripped and the resulting metal electrodes are connected to a multichannel recorder or stimulator.
Dissociated cells or tissue sections are then seeded on the MEA and can be recorded from or stim-
ulated electrically. Micron-scale control over the size, shape, and position of the metal electrodes
allows many electrodes to be placed on the cell culture substrate in a variety of configurations.
Because MEAs are integrated into the cell culture substrate, they offer the ability to record from cell
cultures for long periods of time.
MEAs have been used primarily to record electrical activity in tissues and cultured cells,68
although they can also be used to electrically stimulate cells69,70 or other functions such as electro-
porate dissociated cell cultures with precise spatiotemporal control.71 MEAs are extremely useful
for providing a comprehensive electrical profile of an entire cell culture, or for increasing experi-
mental throughput by simultaneously recording from or stimulating multiple cells.68 Although the
quality of the electrical recordings of MEAs is not as good as patch clamping, it is comparable
with or better than other extracellular recording methods. Effective recording or stimulation of elec-
trically active cells using MEAs requires close contact between the cell and the electrode. Large
cell-electrode distances will make action potentials difficult to detect and electrical activity difficult
to stimulate due to decreasing field strength with distance from the cell or electrode, respectively. To
(a)
Tissue with signal Tissue/substrate
sources contact
MEA
Recording Stimulation
(b) (c)
200 µm
FIGURE 32.18 Microelectrode arrays. (a) MEAs can be used to record or stimulate electrical activity in
dissociated cell cultures, engineered 3D cultures, or whole tissue explants. (b) MEAs are constructed using
metal deposition and photolithography resulting in intricate, user-defined electrode configurations. (c) Brain
tissue explant cultured on a MEA after 7 days in vitro. (Adapted from Stett, A., et al., Anal Bioanal Chem, 377,
486, 2003. Copyright 2003. With permission from Springer Science and Business Media.)
encourage close contact between cells and electrodes, MEAs are often patterned with adhesive pro-
teins such as poly(lysine) and/or extracellular matrix proteins, or adhesion-promoting self-assembled
monolayers using microstamping or other patterning techniques.1
MEAs have an added advantage in that they are compatible with and can be incorporated into
many other methods discussed in this chapter to address more complex questions. For exam-
ple, microfluidic channels constructed around an MEA72 would allow the electrical activity of
cells or whole tissue sections to be recorded or stimulated while the cells are exposed to unique
chemical, mechanical, or thermal73 environments. Cells could be exposed to chemical gradients,
regulated shear forces, or disparate fluidic environments for many days while under constant electrical
monitoring or stimulation.
32.5 SUMMARY
Microfabrication technology has provided a plethora of tools and methods to engineer the position
and microenvironment of cells in vitro. The unprecedented level of control over the mechanical,
chemical, and electrical nature of the cellular microenvironment allows investigation of questions
not addressable with conventional tools and methods. The unique insight into normal and abnormal
cell behavior afforded by microfabricated tools and methods may one day lead to cures for injuries
and diseases, and the ability to direct cell growth and behavior for tissue engineering or industrial
applications.
Although microfabrication technology has greatly enhanced the type of complex environments
in which cells can be grown and studied, there is a large need for the development of complementary
analysis methods to detect and characterize changes in cell response. Existing analysis methods,
some of which are covered in the other chapters of this book, are often not adaptable to or optimized
for use with microfabricated cell culture tools or environments. The development of novel methods
that integrate engineered cell culture environments with complementary analysis tools would greatly
accelerate our understanding of normal and abnormal cell behavior.
REFERENCES
1. Folch, A. and Toner, M., Microengineering of cellular interactions, Annu Rev Biomed Eng, 2, 227, 2000.
2. Li, N., Tourovskaia, A., and Folch, A., Biology on a chip: Microfabrication for studying the behavior
of cultured cells, Crit Rev Biomed Eng, 31, 423, 2003.
3. Falconnet, D., Csucs, G., Grandin, H. M., and Textor, M., Surface engineering approaches to
micropattern surfaces for cell-based assays, Biomaterials, 27, 3044, 2006.
4. Yap, F. L. and Zhang, Y., Protein and cell micropatterning and its integration with micro/nanoparticles
assembly, Biosens Bioelectron, 22, 775, 2007.
5. Liu, Y., Sun, S., Singha, S., Cho, M. R., and Gordon, R. J., 3D femtosecond laser patterning of collagen
for directed cell attachment, Biomaterials, 26, 4597, 2005.
6. Suh, K. Y., Seong, J., Khademhosseini, A., Laibinis, P. E., and Langer, R., A simple soft litho-
graphic route to fabrication of poly(ethylene glycol) microstructures for protein and cell patterning,
Biomaterials, 25, 557, 2004.
7. Tsang, V. L. and Bhatia, S. N., Three-dimensional tissue fabrication, Adv Drug Deliv Rev, 56,
1635, 2004.
8. Zhu, A., Chen, R., and Chan-Park, M. B., Patterning of a random copolymer of poly[lactide-co-
glycotide-co-(epsilon-caprolactone)] by UV embossing for tissue engineering, Macromol Biosci, 6,
51, 2006.
9. Albrecht, D. R., Tsang, V. L., Sah, R. L., and Bhatia, S. N., Photo- and electropatterning of hydrogel-
encapsulated living cell arrays, Lab Chip, 5, 111, 2005.
10. Gray, D. S., Tan, J. L., Voldman, J., and Chen, C. S., Dielectrophoretic registration of living cells to a
microelectrode array, Biosens Bioelectron, 19, 1765, 2004.
11. Voldman, J., Electrical forces for microscale cell manipulation, Ann Rev Biomed Eng, 8, 425, 2006.
12. Kaler, K. V. I. S. and Jones, T. B., Dielectrophoretic spectra of single cells determined by feedback-
controlled levitation, Biophys J, 57, 173, 1990.
13. Fuhr, G., Arnold, W. M., Hagedorn, R., Muller, T., Benecke, W., Wagner, B., and Zimmermann, U.,
Levitation, holding, and rotation of cells within traps made by high-frequency fields, Biochim Biophys
Acta, 1108, 215, 1992.
14. Hakoda, M., Hachisu, T., Wakizaka, Y., Mii, S., and Kitajima, N., Development of a method to analyze
single cell activity by using dielectrophoretic levitation, Biotechnol Prog, 21, 1748, 2005.
15. Roberts, M. J., Bentley, M. D., and Harris, J. M., Chemistry for peptide and protein PEGylation, Adv
Drug Deliv Rev, 54, 459, 2002.
16. Boland, T., Xu, T., Damon, B., and Cui, X., Application of inkjet printing to tissue engineering,
Biotechnol J, 1, 910, 2006.
17. Pardo, L., Wilson, W. C., and Boland, T. J., Characterization of patterned self-assembled monolayers
and protein arrays generated by the ink-jet method, Langmuir, 19, 1462, 2003.
18. Varghese, D., Deshpande, M., Xu, T., Kesari, P., Ohri, S., and Boland, T., Advances in tissue
engineering: Cell printing, J Thorac Cardiovasc Surg, 129, 470, 2005.
19. Xu, T., Gregory, C. A., Molnar, P., Cui, X., Jalota, S., Bhaduri, S. B., and Boland, T., Viability and
electrophysiology of neural cell structures generated by the inkjet printing method, Biomaterials, 27,
3580, 2006.
20. Flemming, R. G., Murphy, C. J., Abrams, G. A., Goodman, S. L., and Nealey, P. F., Effects of synthetic
micro- and nano-structured surfaces on cell behavior, Biomaterials, 20, 573, 1999.
21. Stevens, M. M. and George, J. H., Exploring and engineering the cell surface interface, Science, 310,
1135, 2005.
22. Abrams, G. A., Murphy, C. J., Wang, Z. Y., Nealey, P. F., and Bjorling, D. E., Ultrastructural basement
membrane topography of the bladder epithelium, Urol Res, 31, 341, 2003.
23. Chen, Y. and Pepin, A., Nanofabrication: Conventional and nonconventional methods, Electrophoresis,
22, 187, 2001.
24. Murugan, R. and Ramakrishna, S., Nano-featured scaffolds for tissue engineering: Areview of spinning
methodologies, Tissue Eng, 12, 435, 2006.
25. Norman, J. and Desai, T., Methods for fabrication of nanoscale topography for tissue engineering
scaffolds, Ann Biomed Eng, 34, 89, 2006.
26. Pham, Q. P., Sharma, U., and Mikos, A. G., Electrospinning of polymeric nanofibers for tissue
engineering applications: A review, Tissue Eng, 12, 1197, 2006.
27. Li, W. J., Laurencin, C. T., Caterson, E. J., Tuan, R. S., and Ko, F. K., Electrospun nanofibrous
structure: A novel scaffold for tissue engineering, J Biomed Mat Res, 60, 613, 2002.
28. Mata, A., Boehm, C., Fleischman, A. J., Muschler, G., and Roy, S., Analysis of connective tissue pro-
genitor cell behavior on polydimethylsiloxane smooth and channel micro-textures, Biomed Microdev,
4, 267, 2002.
29. Martin, C. R. and Aksay, I. A., Microchannel molding: A soft lithography-inspired approach to
micrometer-scale patterning, J Mat Res, 20, 1995, 2005.
30. Li, N. Z. and Folch, A., Integration of topographical and biochemical cues by axons during growth on
microfabricated 3-D substrates, Exp Cell Res, 311, 307, 2005.
31. Narasimhan, J. and Papautsky, I., Polymer embossing tools for rapid prototyping of plastic microfluidic
devices, J Micromech Microeng, 14, 96, 2004.
32. Goldner, J. S., Bruder, J. M., Li, G., Gazzola, D., and Hoffman-Kim, D., Neurite bridging across
micropatterned grooves, Biomaterials, 27, 460, 2006.
33. Bruder, J. M., Monu, N. C., Harrison, M. W., and Hoffman-Kim, D., Fabrication of polymeric replicas
of cell surfaces with nanoscale resolution, Langmuir, 22, 8266, 2006.
34. Geissler, M. and Xia, Y. N., Patterning: Principles and some new developments, Adv Mat, 16,
1249, 2004.
35. Henke, W., Hoppe, W., Quenzer, H. J., Staudtfischbach, P., and Wagner, B., Simulation and process
design of gray-tone lithography for the fabrication of arbitrarily-shaped surfaces, Jpn J Appl Phys Part
1–Regular Papers Short Notes & Review Papers, 33, 6809, 1994.
36. Chen, C., Hirdes, D., and Folch, A., Gray-scale photolithography using microfluidic photomasks, Proc
Natl Acad Sci USA, 100, 1499, 2003.
37. Kawata, S., Sun, H. B., Tanaka, T., and Takada, K., Finer features for functional microdevices—
Micromachines can be created with higher resolution using two-photon absorption, Nature, 412,
697, 2001.
38. Baldacchini, T., LaFratta, C. N., Farrer, R. A., Teich, M. C., Saleh, B. E. A., Naughton, M. J.,
and Fourkas, J. T., Acrylic-based resin with favorable properties for three-dimensional two-photon
polymerization, J Appl Phys, 95, 6072, 2004.
39. LaFratta, C. N., Lit, L. J., and Fourkas, J. T., Soft-lithographic replication of 3D microstructures with
closed loops, Proc Natl Acad Sci USA, 103, 8589, 2006.
40. Brown, T. D., Techniques for mechanical stimulation of cells in vitro: A review, J Biomech, 33,
3, 2000.
41. Hoffman, J. M., Shao, J., Hsu, C. H., and Folch, A., Elastomeric Molds with tunable microtopography,
Adv Mat, 16, 2201, 2004.
42. Duerrschmidt, N., Stielow, C., Muller, G., Pagano, P. J., and Morawietz, H., NO-mediated regulation
of NAD(P)H oxidase by laminar shear stress in human endothelial cells, J Physiol, 576 (Pt 2), 557,
2006.
43. Frame, M. D., Chapman, G. B., Makino, Y., and Sarelius, I. H., Shear stress gradient over endothelial
cells in a curved microchannel system, Biorheology, 35, 245, 1998.
44. Bransky, A., Korin, N., Nemirovski, Y., and Dinnar, U., An automated cell analysis sensing system
based on a microfabricated rheoscope for the study of red blood cells physiology, Biosens Bioelectr,
22, 165, 2006.
45. Bird, R. B., Stewart, W. E., and Lightfoot, E. N., Transport Phenomena, John Wiley & Sons, Inc., 1960.
46. Shin, H., Jo, S., and Mikos, A. G., Biomimetic materials for tissue engineering, Biomaterials, 24,
4353, 2003.
47. Boyden, S., The chemotactic effect of mixtures of antibody and antigen on polymorphonuclear
leucocytes, J Exp Med, 115, 453, 1962.
48. Zigmond, S. H., Ability of polymorphonuclear leukocytes to orient in gradients of chemotactic factors,
J Cell Biol, 75 (Pt 1), 606, 1977.
49. Gunderson, R. W. and Barrett, J. N., Neuronal chemotaxis: Chick dorsal-root axons turn toward high
concentrations of nerve growth factor, Science, 206, 1079, 1979.
50. Zicha, D., Dunn, G. A., and Brown, A. F., A new direct-viewing chemotaxis chamber, J Cell Sci, 99
(Pt 4), 769, 1991.
51. Foxman, E. F., Campbell, J. J., and Butcher, E. C., Multistep navigation and the combinatorial control
of leukocyte chemotaxis, J Cell Biol, 139, 1349, 1997.
52. Chen, H., He, Z., Bagri, A., and Tessier-Lavigne, M., Semaphorin−neuropilin interactions underlying
sympathetic axon responses to class III semaphorins, Neuron, 21, 1283, 1998.
53. Jeon, N. L., Dertinger, S. K. W., Chiu, D. T., Choi, I. S., Stroock, A. D., and Whitesides, G. M.,
Generation of solution and surface gradients using microfluidic systems, Langmuir, 16, 8311, 2000.
54. Jeon, N. L., Baskaran, H., Dertinger, S. K. W., Whitesides, G. M., Van de Water, L., and Toner, M.,
Neutrophil chemotaxis in linear and complex gradients of interleukin-8 formed in a microfabricated
device, Nat Biotechnol, 20, 826, 2002.
55. Chung, B. G., Flanagan, L. A., Rhee, S. W., Schwartz, P. H., Lee, A. P., Monuki, E. S., and Jeon, N. L.,
Human neural stem cell growth and differentiation in a gradient-generating microfluidic device, Lab
Chip, 5, 401, 2005.
56. Saadi, W., Wang, S. J., Lin, F., and Jeon, N. L., Aparallel-gradient microfluidic chamber for quantitative
analysis of breast cancer cell chemotaxis, Biomed Microdev, 8, 109, 2006.
57. Walker, G. M., Sai, J., Richmond, A., Stremler, M., Chung, C. Y., and Wikswo, J. P., Effects of flow
and diffusion on chemotaxis studies in a microfabricated gradient generator, Lab Chip, 5, 611, 2005.
58. Chung, B. G., Lin, F., and Jeon, N. L., A microfluidic multi-injector for gradient generation, Lab Chip,
6, 764, 2006.
59. Keenan, T. M., Hsu, C. H., and Folch, A., Microfluidic “jets” for generating steady-state gradients of
soluble molecules on open surfaces, Appl Phys Lett, 89, 114103, 2006.
60. Takayama, S., Ostuni, E., LeDuc, P., Naruse, K., Ingber, D. E., and Whitesides, G. M., Subcellular
positioning of small molecules, Nature, 411, 1016, 2001.
61. Takayama, S., Ostuni, E., LeDuc, P., Naruse, K., Ingber, D. E., and Whitesides, G. M., Selective
chemical treatment of cellular microdomains using multiple laminar streams, Chem Biol, 10, 123, 2003.
62. Spielman, L. and Goren, S. L., Improving resolution in Coulter counting by hydrodynamic focusing,
J Colloid Interface Sci, 26, 175, 1968.
63. Sawano, A., Takayama, S., Matsuda, M., and Miyawaki, A., Lateral propagation of EGF signaling
after local stimulation is dependent on receptor density, Dev Cell, 3, 245, 2002.
64. Tourovskaia, A., Figueroa-Masot, X., and Folch, A., Differentiation-on-a-chip: Amicrofluidic platform
for long-term cell culture studies, Lab Chip, 5, 14, 2005.
65. Tourovskaia, A., Kosar, T. F., and Folch, A., Local induction of acetylcholine receptor clustering in
myotube cultures using microfluidic application of agrin, Biophys J, 90, 2192, 2006.
66. Taylor, A. M., Blurton-Jones, M., Rhee, S. W., Cribbs, D. H., Cotman, C. W., and Jeon, N. L.,
A microfluidic culture platform for CNS axonal injury, regeneration and transport, Nat Methods, 2,
599, 2005.
67. Campenot, R. B., Local control of neurite development by nerve growth factor, Proc Natl Acad Sci
USA, 74, 4516, 1977.
68. Stett, A., Egert, U., Guenther, E., Hofmann, F., Meyer, T., Nisch, W., and Haemmerle, H., Biological
application of microelectrode arrays in drug discovery and basic research, Anal Bioanal Chem, 377,
486, 2003.
69. Jimbo, Y., Kasai, N., Torimitsu, K., Tateno, T., and Robinson, H. P. C., A system for MEA-based
multisite stimulation, IEEE Trans Biomed Eng, 50, 241, 2003.
70. Shenai, M. B., Putchakayala, K. G., Hessler, J.A., Orr, B. G., Banaszak Holl, M. M., and Baker, J. R., Jr.,
A novel MEA/AFM platform for measurement of real-time, nanometric morphological alterations of
electrically stimulated neuroblastoma cells, IEEE Trans Nanobiosci, 3, 111, 2004.
71. Jain, T. and Muthuswamy, J., Microsystem for transfection of exogenous molecules with spatio-
temporal control into adherent cells, Biosens Bioelectron, 22, 863, 2007.
72. Pearce, T. M., Williams, J. J., Kruzel, S. P., Gidden, M. J., and Williams, J. C., Dynamic control of
extracellular environment in in vitro neural recording systems, IEEE Trans Neural Syst Rehabil Eng,
13, 207, 2005.
73. Pearce, T. M., Wilson, J. A., Oakes, S. G., Chiu, S. Y., and Williams, J. C., Integrated microelectrode
array and microfluidics for temperature clamp of sensory neurons in culture, Lab Chip, 5, 97, 2005.
33 Multidimensional Microfluidic
Systems for Protein and
Peptide Separations
Don L. DeVoe and Cheng S. Lee
CONTENTS
33.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1001

33.2 Microfluidic Platforms for Multidimensional Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1002
33.2.1 Time-Multiplexed Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1002
33.2.2 Spatially Multiplexed Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1003
33.2.3 Interfacing Multidimensional Separations with MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1005
33.3.1 Fabrication of Spatially Multiplexed Microfluidic Chips . . . . . . . . . . . . . . . . . . . . . . . . . 1008
33.3.2 Electrical and Hydrodynamic Crosstalk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1008
33.3.3 Sample Dispersion and Loss at Channel Intersections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1009
33.3.4 Inhibiting Bulk Flow in Multidimensional Microchannel
Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1010
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1012
33.1 INTRODUCTION
Since the concept of micro total analysis systems (µTAS) was first proposed,1 the field has advanced
rapidly, with ongoing developments promising to profoundly revolutionize modern bioanalytical
platforms and methodologies. Whether termed µTAS, lab-on-a-chip, or microfluidics, the technolo-
gies that define the field offer important innovations capable of transforming the ways in which
bioanalytical techniques are performed. For example, reduced size and power requirements can
enable improved portability and higher levels of integration, with lower per-unit cost for disposable
applications. Low volume fluid control enabled by microfluidics allows smaller dead volumes and
reduced sample consumption, while many pumping methods including capillary action and elec-
troosmotic flow scale favorably in these systems, enabling precise and valveless flow control at the
microscale. Similarly, thermal time constants tend to be extremely small due to large surface area
to volume ratios inherent in these systems, reducing the onset of significant Joule heating during
electrokinetic separations and thus allowing higher separation voltages and shorter analysis times
with equivalent or better separation resolution for complex mixtures in an integrated format.
A significant advantage of microfluidics is that lithographic and replication-based fabrication
techniques readily lend themselves to the formation of complex systems, providing a path for effec-
tive manufacturing of highly parallel analytical tools. Specifically, we consider here the promise that
microfluidics technology holds for realizing robust multidimensional separations of proteins and pep-
tides in an integrated platform. Assuming the separation techniques used in a given two-dimensional
1001
(2-D) system are fully orthogonal, that is, the two separation techniques are based on independent
physicochemical properties of analytes, the overall peak capacity is the product of the peak capaci-
ties of the individual one-dimensional (1-D) methods.2 As a result, multidimensional systems are of
great interest for the analysis of complex mixtures.3 Nowhere is the demand for effective analysis
of complex matrices greater than in the field of proteomics. A major challenge faced in proteomics
analysis is the vast number of proteins present in a typical biological sample, together with the large
variation of protein relative abundances (typically greater than 6 orders of magnitude,4 and higher
than 10 orders of magnitude for the case of blood plasma5 ), which ultimately dictate the need for
high resolution separations to be performed on protein or peptide samples before detection by mass
spectrometry (MS). Thus, multidimensional separation technologies capable of reducing dynamic
range and enhancing detection sensitivity without substantially sacrificing analytical throughput are
highly desirable for many proteomic applications.
While intact protein analysis via 2-D polyacrylamide gel electrophoresis (2-D PAGE)6
and bottom-up peptide analysis combining multidimensional chromatography with tandem MS
analysis7,8 remain the workhorse technologies for most modern proteomic studies, ongoing improve-
ments in microfluidics for sample preparation and separations in proteomics are beginning to reveal
their potential impact on the field. In this chapter, microfluidic platforms for protein and peptide
analysis employing multidimensional electrophoretic separations and combined electrophoretic and
chromatographic separations are discussed.
33.2 MICROFLUIDIC PLATFORMS FOR MULTIDIMENSIONAL

SEPARATIONS
33.2.1 TIME-MULTIPLEXED SEPARATIONS
The majority of reported multidimensional microfluidic systems employ serial separations that are
time multiplexed. In this approach, fractions from the first separation dimension are sampled sequen-
tially, with each fraction separated in series within the second dimension while the first-dimension
separation continues. In one of the first demonstrations of this approach, Rocklin et al.9 performed
2-D separations of peptide mixtures in a microfluidic device using micellar electrokinetic chro-
matography (MEKC) and capillary zone electrophoresis (CZE) as the first and second dimensions,
respectively. A schematic representation of their platform, fabricated in a glass substrate, is shown in
Figure 33.1a. The MEKC separation was performed within a 65-mm-long serpentine channel located
between the upper and lower channel intersection points shown in the figure, with a 10-mm-long
CZE separation performed past the lower channel intersection. With a time scale for CZE on the
order of a few seconds, substantially faster than the MEKC separation, large numbers of fractions
could be sampled from the first dimension and separated by CZE with an overall analysis time under
10 min. Because MEKC is a transient electrokinetic separation, only a small portion (∼10%) of the
analyte in the first dimension could be sampled into the CZE microchannel, since no fractions could
be sampled while the second dimension separation was being performed. Despite this sample loss
issue, a revised chip design was later reported for peptide separations with a peak capacity greater
than 4000, such as the comparative separation of tryptic digest from bovine and human hemoglobin
shown in Figure 33.1b revealing distinct peptide variations between these samples.10
Similar approaches based on time multiplexing have been employed by a number of researchers.
For example, Gottschlich et al.11 used a spiral shaped glass channel coated with a C18 stationary
phase for performing chromatographic separation of trypsin-digested peptides. By providing a cross
interface, peptides eluted from an MEKC9 or reversed-phase11 chromatography channel were sam-
pled and rapidly separated by CZE in a short glass microchannel. Herr et al.12 similarly coupled
isoelectric focusing (IEF) with serial CZE for 2-D separations of model proteins, a concept that
Multidimensional Microfluidic Systems for Protein and Peptide Separations 1003
(a) (b)
700 (A) Bovine hemoglobin
S W1 600
500
MEKC elution time (s)

400
B1
300
700 (B) Human hemoglobin

W2
600
500
400
B2 W
300
0.7 0.8 0.9 1.0 1.1 1.2

CE migration time (s)
FIGURE 33.1 Time-multiplexed MEKC/CZE peptide separation chip schematic (From Rocklin, R. D. et al.,
Anal. Chem., 2000, 72, 5244–5249.) and (b) pseudo-gel views for chip-based separations of human and bovine
hemoglobin. (From Ramsey, J. D. et al., Anal. Chem., 2003, 75, 3758–3764.)
was later implemented in a poly(dimethylsiloxane) (PDMS)-based microfluidic system using elas-

tomeric valving to isolate the separation dimensions.13 In addition, Shadpour and Soper14 developed
a microfluidic chip combining sodium dodecyl sulfate (SDS) gel electrophoresis with MEKC, again
with multiple fractions sampled serially from the gel electrophoresis separation for time-multiplexed
fractionation by MEKC in the second dimension.
In each of the preceding examples, the separations were time multiplexed, with rapid second
dimension separations chosen to enable repeated sampling of analyte eluted from a slower first dimen-
sion separation. Such time-multiplexed separations have also been employed in multidimensional
capillary separations, for example, through the combination of a first dimension chromatographic
separation together with a second dimension separation based on CZE,15 with low dead-volume
switching valves used to couple the separation dimensions. For many applications, multidimensional
microfluidic systems employing time-multiplexed separations may not offer compelling advantages
over their capillary analogs, but rather introduce additional complexities, costs, and performance
limitations, which may be avoided through the use of more traditional capillary technologies. On the
other hand, unique benefits exist for applications that can leverage the small footprints and integrated
nature of these microfluidic platforms, particularly when coupled with on-chip sample preparation.
33.2.2 SPATIALLY MULTIPLEXED SEPARATIONS

In contrast to time multiplexing, high-throughput multidimensional separations may also be per-
formed using spatial multiplexing, with an array of second dimension microchannels used to
simultaneously separate analyte sampled as multiple fractions from the entirety of a first dimen-
sion separation channel. Using this approach relaxes the requirement for the second dimension
separation to be substantially faster than the first dimension, opening the door to a wider range of
potential separation modes. Furthermore, spatial multiplexing enables complete sampling of the first
dimension separation, even for transient separation modes, thereby preventing sample loss due to
a duty cycle mismatch between the dimensions. The use of spatial multiplexing also leverages a
significant advantages of microfluidics, namely the ability to combine large numbers of channels in
a single chip without the need for complex fluidic interconnects between the channels.
The concept of spatially-multiplexed microfluidic separations was first embodied in a micro-
fabricated quartz device proposed by Becker et al.16 which contained a single channel for a first
dimension and an array of 500 parallel channels with submicron dimensions positioned orthogo-
nally to the first dimension. While no separations were actually performed in this device, the concept
was later extended and validated by several groups. For example, Chen et al.17 described a microflu-
idic 2-D capillary electrophoresis platform based on a reconfigurable system of six individual PDMS
layers. The chip consisted of a 2.5-cm-long microchannel for performing IEF, with an intersecting
array of parallel 6-cm-long microchannels for SDS–PAGE. This six-layer PDMS microfluidic device
required the alignment, bonding, removal, realignment, and rebonding of various combinations of
the six layers to perform a full 2-D protein separation.
A fully integrated system developed by Li et al.18 combined IEF with SDS gel electrophoresis of
intact proteins in a single polycarbonate chip containing 10 parallel second dimension microchannels
(Figure 33.2). By combining both separation dimensions in a single rigid polymer chip, dispersion of
analyte during the transfer between the dimensions was minimized. Furthermore, different separation
media were employed in each of the dimensions, allowing sample mobility to be independently
tailored for rapid free-solution IEF separations and high resolution SDS-capillary gel electrophoresis
(CGE) separations. By using a replaceable polyethylene oxide (PEO) gel in the SDS–CGE channels,
the interface between the separation media also provided for sample stacking during the transfer
of focused proteins bands into the second dimension. Using this system, a comprehensive 2-D
protein separation was completed in less than 10 min, with the majority of time consumed in the
required SDS–protein complexation reaction. A peak capacity of ∼170 in the second dimension of
size-based separation was estimated from a measured bandwidth of 150 µm over a 2.5 cm channel
length. Because the separation mechanisms in IEF and SDS–CGE were completely orthogonal, the
overall peak capacity was estimated to be 1700 (10 fractions from IEF × 170 from SDS–CGE).
Improvements in peak capacity were proposed through the use of longer channels during the size-
based separation, and by increasing the density of microchannels in the array to increasing the
number of IEF fractions analyzed during the size-based separation. Owing to the use of parallel
Reservoir D
Reservoir Reservoir
A B
Reservoir C
FIGURE 33.2 Schematic representation and photograph of a 2-D IEF/SDS–CGE microfluidic chip containing
10 spatially-multiplexed channels for SDS–CGE separation. (From Li, Y. et al., Anal. Chem., 2004, 76,
742–748.)
(a) (b)
FIGURE 33.3 (See color insert following page 810.) (a) Microfluidic PMMA chip for 2-D IEF/SDS–CGE
protein separations and (b) pseudo-gel image resulting from a yeast cell lysate separation.
separations in the second dimension, there is no accompanying increase in the analysis time. Further
improvements in separation resolution were later achieved through the use of individual reservoirs
for the injection of PEO gel and SDS solutions into the chip, and through electrical isolation of
each second dimension channel. A revised chip design is shown in Figure 33.3a, with an example
separation of intact proteins from yeast cell lysate depicted in Figure 33.3b.
Ivory and coworkers19 demonstrated a multistage separation platform based solely on multi-
ple IEF separations, with an initial separation performed in a straight channel using broad-range
ampholytes, followed by a second separation using a set of orthogonal microchannels that ter-
minated in reservoirs containing ampholytes covering more narrow pI ranges for higher focusing
resolution. By providing multiple sets of the orthogonal channels, progressively narrower pI ranges
were established within the main separation channel. Alternately, multiple parallel narrow-range
IEF separations could be performed using the apparatus. A three-stage transient IEF separation
was demonstrated in this system, and additional stages were proposed. Multidimensional microflu-
idic separation platforms with even higher dimensionality have been proposed, such as a concept
for a four-dimensional (4-D) platform combining IEF, isotachophoresis (ITP), SDS–PAGE, and
reversed-phase liquid chromatography (RPLC)20 (Figure 33.4). It is not yet clear whether sig-
nificant advantages can be realized by such high-dimensional fractionation, since analyte loss,
dispersion, and dilution may ultimately prevent effective downstream detection (e.g., by MS), in par-
ticular given the relatively low loading capacities typical of many microfluidic systems. Regardless,
this 4-D concept suggests an intriguing direction for new platforms and multidimensional separa-
tion modalities that can be enabled by the unique advantages offered by integrated microfluidics
technology.
33.2.3 INTERFACING MULTIDIMENSIONAL SEPARATIONS WITH MS

Mass spectrometry is an essential tool for the characterization of biomolecules, revealing the charge-
to-mass ratio of analyte molecules following ionization. There is a strong need for effective MS in
the analysis of proteins and peptides, where MS coupling is required to provide sufficient mass
resolution and sequence information for modern proteomic analyses. Unlike deoxyribonucleic acid
(DNA), no techniques exists for the direct amplification of proteins, and thus detection sensitivity
is paramount for the identification of low abundance species from limited samples. The matter is
further complicated by the enormous dynamic range of protein abundance in complex samples. For
IEF
ITP
SDS-PAGE
rp-HPLC
FIGURE 33.4 Conceptual layout of a spatially-multiplexed 4-D microfluidic system combining IEF, ITP,
SDS–PAGE, and RPLC. (From Ivory, C. F., Electrophoresis, 2007, 28, 15–25.)
multidimensional microfluidic technologies to substantially impact the field of proteomics, effective

methods for MS interfacing are a necessity.
The integration of microfluidic systems and MS received substantial attention following demon-
stration of the first microfluidic electrospray ionization (ESI)–MS interfaces in 1997 by the groups of
Karger and coworkers,21 Ramsey and Ramsey,22 and Aebersold and coworkers.23 Extensive studies
on ESI–MS interfacing for coupling microfluidics to MS have been reported, with online ESI–MS
offering a simple approach for directly interfacing microfluidic analyses with MS, while providing
good sensitivity and mass accuracy, and reproducible signals for effective analyte quantification.
A number of reviews have appeared in recent years, which illuminate this topic. Oleschuk and
Harrison24 offer an early summary of microfluidic ESI–MS interfacing, and Sung et al.25 provide a
more recent review addressing this area. A further review by Limbach and Meng26 in 2002 updates
the developments in microfluidics for ESI–MS, and reviews by Figeys and Pinto,27 Lion et al.,28 and
Marko-Varga et al.29 focus on microfabricated systems in proteomics, including summaries of rele-
vant microfluidic-to-MS interfaces. A more detailed discussion of microfluidic ESI–MS interfacing
is provided in Chapter 53 by Lazar.
While ESI–MS is well suited as an interface between serial time-multiplexed microfluidic
systems and MS, as reviewed by Foret and Kusý,30 interfaces based on matrix-assisted laser desorp-
tion/ionization (MALDI) offer several important advantages for spatially-multiplexed (i.e., parallel
analysis) microfluidics. Unlike online ESI–MS, MALDI–MS is an offline soft ionization method,
which is typically applied to the analysis of solid phase analyte co-crystallized with an energy-
absorbing matrix material on the surface of a supporting target plate. In comparison with ESI–MS,
MALDI–MS tends to be more tolerant of salts and other sample contaminants, offers excellent detec-
tion limits, and produces mass spectra which are relatively simple to interpret due to the absence of
multiple charge states. Although generally limited to the analysis of higher molecular weight species
due to interference from matrix components below ∼500 Da, MALDI–MS can be applied to a wide
range of sample mass, with good sensitivity above 300 kDa, compared with less than 100 kDa for
ESI–MS.
From the point of view of multidimensional microfluidic systems, the off-line nature of MALDI–
MS analysis represents a key advantage. First, MALDI–MS allows on-chip sample processing steps
to be decoupled from back-end MS analysis. This is necessary when, for example, the time scales
for biomolecular separations and online MS data acquisition are incompatible. More importantly for
the present discussion, off-line MALDI–MS analysis provides a method for coupling simultaneous
parallel on-chip analyses with MS detection, where ESI–MS from multiple microchannels is simply
not practical or even feasible. It should be noted that although MALDI–MS is a serial process once
the sample plate has been prepared, its high duty cycle enables high throughput for large numbers of
deposited samples. For reviews of work in this area, Foret and Preisler31 presented an early summary
of microfabricated systems for MALDI–MS interfacing, and DeVoe and Lee32 recently reviewed
the current state of microfluidic interfacing with MALDI–MS. Additional discussion of microfluidic
MALDI–MS interfacing may be found in Chapter 49 by Laurell.
For microfluidic systems that employ parallel chromatography channels, MALDI–MS interfac-
ing is particularly straightforward. The concept of interfacing multiplexed capillary chromatography
separations with MALDI–MS for high-throughput proteome analysis was first described by Aeber-
sold and coworkers33 using a system comprising four parallel capillary chromatography columns
coupled with MALDI–MS/MS. In their system, a network of capillaries connected together with
microjunctions was used to interface a single gradient pump to four parallel separation columns,
with eluent from each column deposited by direct mechanical spotting onto a MALDI target at
a flow rate of 1 µL/min using a commercial fraction collector. Flow rate variations of ±5–6%
were observed between the columns, presumably due to variations in hydrodynamic flow resistance
resulting from inhomogeneity of the packed silica beads. In a recent demonstration by Knapp and
coworkers,34 a multichannel cyclic olefin copolymer (COC) chip was developed for parallel nanoflow
RPLC analysis. Chip effluent was deposited onto a MALDI target through an electrically mediated
deposition technique,35 which was shown to be robust for the case of a two-channel chromato-
graphy chip.
While electrospray36 and pulsed electric field35 deposition has been demonstrated from mul-
tichannel chips, there are several disadvantages associated with the use of electric fields to assist
sample deposition. In addition to increased system complexity, the reliability and repeatability of
droplet deposition degrade as the number of parallel channels is increased. In contrast, the direct
mechanical spotting strategy described by Aebersold and coworkers33 in a capillary format is well
suited for multichannel microfluidic chips employing a chromatographic separation in the second
dimension. For example, the eight-channel COC chip shown in Figure 33.5 contains 100-µm wide
and 60-µm deep channels, with a single fluid port connected to all eight channels through a sym-
metric on-chip splitter. The end of the chip was cut with a high-speed semiconductor dicing saw to
expose the channel exits. Using a syringe pump, a panel of nine model peptides were deposited onto
64 target spots (8 channels × 8 spotting events) on a custom MALDI target, with a target deposition
volume of 150 nL per spot corresponding to sample amounts ranging from 25 to 75 fmol for each pep-
tide. Using a Kratos Amixa MALDI-time-of-flight (TOF) instrument, the resulting spectra showed
good uniformity, with an average relative standard deviation (RSD) of 44% across all peptides.
(a) (b)
FIGURE 33.5 (a) Photo of an eight-channel COC chip aligned to a MALDI target mounted on a three-axis
robotic positioner and (b) close up of chip exit showing droplet formation.
This compares well with an RSD of 50% in MALDI signal intensities reported for multiple peptide
spots deposited from a single capillary using an optimized matrix seed layer technique.37 Further
improvements in deposition volume uniformity were realized for an eight-channel chip containing
a methacrylate monolith developed for reversed-phase chromatography in the second separation
dimension, with negligible variations at flow rates down to 200 nL/min per channel.

33.3.1 FABRICATION OF SPATIALLY MULTIPLEXED MICROFLUIDIC CHIPS
There are a number of practical issues that must be addressed to realize functional multidimen-
sional microfluidic separation platforms. For devices based on spatially-multiplexed microfluidics,
challenges exist even at the fabrication level, since producing microfluidic substrates with large
numbers of parallel channels can be difficult. For example, when rapid prototyping methods based
on hot embossing of thermoplastics such as polymethyl methacrylate (PMMA) or COC are used
for substrate patterning, high frictional forces imposed during demolding of the plastic from dense
microchannel arrays can lead to distortion of channel geometry, substrate warping, template damage,
and in extreme cases prevent demolding from occurring entirely. It can also be difficult to achieve
sufficient adhesion strength when bonding chips containing closely-spaced microchannels due to the
reduced bond surface area. While these problems can be alleviated by employing low aspect ratio
and tapered channel geometries, limiting the overall chip real estate occupied by the channels, and by
keeping “reasonable” spacing between adjacent channels, such solutions can often be in conflict with
desired performance goals for the system. The appropriate channel spacing must be evaluated on the
basis of the particular fabrication methods, materials, and channel dimensions being employed, and
weighed against potential performance losses resulting from a compromised design.
33.3.2 ELECTRICAL AND HYDRODYNAMIC CROSSTALK

Another fundamental challenge results from electrical and hydrodynamic crosstalk between multiple
interconnected channels that are inherent in spatially-multiplexed systems. To illustrate this point,
consider the schematic representation of the 2-D IEF/SDS–CGE protein separation chip previously
depicted in Figure 33.2a. This chip contains 11 parallel channels above the horizontal IEF channel,
and 10 parallel channels below. As shown in the figure, each of these channels terminate in one
of two common reservoirs, which connect the distal ends of the upper and lower channel arrays.
In practice, this configuration results in a resistive network that shunts electrical current through
the reservoirs during IEF separation, thereby producing an uneven electric field distribution within
the IEF channel. More significantly, both analytes and ampholytes (for the creation of pH gradient
in the IEF channel) can be unintentionally injected into the upper and lower channel networks
by electrophoresis, resulting in potentially severe sample loss, and preventing establishment of
the desired pH gradient for the present example of IEF as the first separation dimension. Issues
with crosstalk can also manifest during hydrodynamic sample injection, since the resistive network
introduced by the interconnected SDS–CGE channels applies equally to hydrodynamic and electrical
current shunts.
One approach to solve this problem would be to integrate flow control elements such as elas-
tomeric valves into the system, but the additional complexity and inherent fabrication constraints
make this solution less than ideal for real-world applications. Another obvious approach for eliminat-
ing crosstalk would be to provide individual reservoirs for all channels within the second dimension
array, thereby isolating each channel both electrically and fluidically. As an example, this approach
was chosen for the IEF/SDS–CGE device shown in Figure 33.3. However, there are practical lim-
itations to this solution. As the number of reservoirs increases, so does the challenge of achieving
efficient interfacing between off-chip plumbing and on-chip channels. For applications requiring
(a) (b)
FIGURE 33.6 (a) Manifold interface for providing fluidic interconnects between on-chip microchannels and
off-chip pumps, valves, and electrodes and (b) manifold with integrated valves positioned above an IEF/SDS–
CGE chip during testing.
pressure injection of sample, buffer, or other solutions into the device, on-chip connections are
generally required at each reservoir (with the potential exception of waste reservoirs) to provide
interfacing to pumps, valves, and/or electrodes, leading to a complex overall interface. For applica-
tions in which simple open reservoirs are suitable, for example, when electrokinetic sample injection
is favored over pressure injection, the need for large numbers of interconnects to off-chip components
may be alleviated, but additional challenges are introduced by bulk flow induced by hydrodynamic
pressure gradients resulting from uneven fluid heights among large numbers of reservoirs in the
chip. Furthermore, each added reservoir represents an additional point for bubble formation and
entrapment, reducing the robustness of the microfluidic system.
There are several solutions that may be used to address this interface challenge. For the case of the
relatively simple IEF/SDS–CGE chip previously described in Figure 33.3a, 38 individual reservoirs
were ultimately required for effective operation of the system, with a manifold interface used to
simplify the fluidic and electrical interface. The manifold consisted of a machined PMMA block as
shown in Figure 33.6, with O-ring compression seals and integrated valves, which provided leak-
free hydrodynamic pressure control for chip preparation and sample introduction, and electrodes
for electrokinetic control. The use of manifolds for microfluidic interfacing can substantially reduce
the time and effort required for chip preparation and testing by eliminating the need for discrete
fluidic or electrical ports to be added to each on-chip reservoir. Another solution is to reduce the
required number of reservoirs by connecting limited numbers of channels to common fluid ports for
off-chip access. Consider the example of a chip design comprising 32 second dimension channels
shown in Figure 33.7a. Each set of eight adjacent channels within the upper and lower arrays are
connected to a single reservoir, substantially reducing the total number of required ports. Although
channels within each group remain subject to crosstalk during the first dimension separation, the
overall current uniformity is greatly improved by segregating each group from the others as revealed
in Figure 33.7b.
33.3.3 SAMPLE DISPERSION AND LOSS AT CHANNEL INTERSECTIONS

When spatial or time multiplexing is employed as a multidimensional separation strategy with a
first dimension based on an electrokinetic separation mode, an inherent difficulty arises from the
need for the first dimension separation channel to be intersected by one of more second dimension
channels. Regardless of whether the first dimension separation is operated in a transient (e.g., CZE)
or steady-state (e.g., IEF) mode, electric field lines extending into the intersecting channels result in
dispersion of sample out of the first dimension channel, and ultimately to sample loss as diffusion
(a) (b) 1.0
1 reservoir
Channel current (normalized)

0.9
0.8
4 reservoirs
0.7
2 reservoirs
0.6
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Channel position (cm)
FIGURE 33.7 (a) Schematic representation of 2-D separation chip containing 32 second dimension
microchannels, with groups of 8 second microchannels terminating at one of four common reservoirs and
(b) model results showing relative current uniformity within the first dimension channel when using 1, 2, and
4 reservoirs.
allows sample to travel beyond the influence of the radiating electric field lines. This is of particular
concern when analyte passing the intersections is in a concentrated band, since a substantial portion
of the analyte may be lost into the interconnecting channel, while the remaining portion suffers from
broadening after passing the intersection.
Li et al.18 addressed this issue by integrating multiple separation media, with relatively low
resistivity ampholyte and buffer in the first separation dimension for IEF and higher resistivity
PEO gel used for the second dimension for CGE. As a result, the electric field remained largely
constrained within the first dimension microchannel during IEF, and minimal sample dispersion was
observed. In addition to constraining the electric field, low diffusion in the gel allowed dispersed
analyte to be mobilized back into the first dimension channel without suffering a high degree of
loss. As discussed previously, another benefit of this approach is that the multiple separation media
further enabled effective sample stacking as focused analyte was electrokinetically transferred to the
second separation dimension. Cui et al.38 proposed a solution based on the integration of addressable
electrodes near each channel intersection to control the local current streamlines and thereby limit
sample dispersion and loss. This is a more general approach that may be applied to a range of
multidimensional systems. A potential disadvantage is that additional current is injected by each
electrode, leading to current and electric field gradients within the first dimension separation channel,
which can affect separation performance.
33.3.4 INHIBITING BULK FLOW IN MULTIDIMENSIONAL MICROCHANNEL

NETWORKS
An important practical issue that affects many microfluidic systems, and multidimensional microflu-
idics in particular, is unwanted hydrodynamic flow induced by on-chip pressure gradients. Owing
to the complexity of the microchannel networks and relatively large numbers of reservoirs or other
off-chip interconnects required to operate multidimensional separation chips, there is a large poten-
tial for pressure gradients to be induced by sources such as gravity (due to uneven fluid heights
in the reservoirs), variations in capillary forces at the reservoirs, or trapped air bubbles within the
reservoirs or microchannel networks. In the authors’ experience, fully eliminating these pressure
gradient sources in complex microfluidic devices is rarely feasible, even when employing auto-
mated manifold interfaces, taking great care in chip preparation, and allowing pressure gradients
to equilibrate over long periods of time. A more effective approach is to minimize hydrodynamic
flow through the use of media, which inhibit bulk fluid motion, either by increasing the effective
viscosity of fluids or by generating a physical barrier to flow. For example, Cui et al.19 added a 2.5%
solution of methylcellulose to all reservoirs in a multidimensional IEF chip to limit the intrusion
of reservoir solutions into the separation channels, while the replaceable PEO gel used as a second
dimension separation medium by Li et al.18 served a similar purpose. In general, multidimensional
systems employing gel-based separations are likely to suffer from less bulk flow than those using
free solutions, as are systems employing chromatographic media that require high-pressure gradients
to generate bulk flow.

The development of novel platforms for multidimensional separations of proteins and peptides
remains an area of great interest throughout the analytical community. Microfluidic technologies
offer real advantages over capillary-based systems in this area, particularly for the analysis of limited
samples. Despite the promise and numerous examples of multidimensional microfluidic platforms
for proteomic analysis, a number of challenges remain before these systems can become competitive
with modern capillary systems such as 2-D liquid chromatography system for the analysis of complex
substrates. The authors believe that as multidimensional microfluidic systems continue to evolve,
the benefits of the technology will become increasingly evident. Specific areas that require attention
include improved methods for fluidic interfacing and on-chip valving, which do not substantially
impact system complexity or cost. Another important area of research currently being addressed by a
number of groups is the development microfluidic systems with novel coupled separation dimensions
that may offer improvements in separation efficiency. Interfacing spatially-multiplexed systems with
MS also remains a significant challenge, and although a variety of solutions have been proposed and
explored more efforts are needed to achieve parity with serial capillary ESI–MS.
New development efforts in the area of time-multiplexed multidimensional separations may also
provide real-world benefits for proteomic analysis. One of the difficulties with time multiplexing
relates to sample loss resulting from the inability to fully sample all separated analyte bands within the
first dimension microchannel. When a transient separation is employed in the first dimension, sample
loss is inherent in this configuration due to the inability to collect fractions while the second dimension
separation is proceeding. The difficulties with sample loss imposed by the serial nature of the second
separation dimension could potentially be alleviated through the use of on-chip trap columns to store
fractions before injection into the second dimension channel, providing parity between the desired
rate of fraction acquisition and the duty cycle of the second dimension separation. Alternately, a
combined time- and spatial-multiplexing scheme could be envisioned, wherein multiple second
dimension separation channels sequentially sample the first dimension.
Clearly, there are research issues that remain to be addressed before multidimensional microflu-
idics is commercially competitive against more traditional multidimensional capillary systems.
However, with the demand for higher analytical throughput and smaller sample quantities in pro-
teomics continuing to increase, there is an open opportunity for novel multidimensional platforms
based on microfluidics to meet this demand. Beyond the concepts described in this chapter such
as rapid and parallel separations, zero dead-volume flow control, and effective multichannel MS
interfacing, further functionalities are also of great interest, for example, the combination of cell
lysis, protein preconcentration, and digestion before multidimensional separation. Such technolo-
gies, already demonstrated for related lab-on-a-chip applications, offer an intriguing glimpse of what
could ultimately be possible in a fully integrated system for protein or peptide analysis.
REFERENCES
1. Manz, A., Graber, N., Widmer, H. M., Sens. Actuat. 1990, B1, 244–248.
2. Giddings, J. C., J. High Res. Chromatogr. 1987, 10, 319–323.
3. Giddings, J. C., United Separation Science, Wiley, New York, 1991.
4. Corthals, G. L., Wasinger, V. C., Hochstrasser, D. F., Sanchez, J.-C., Electrophoresis 2000, 21,
1104–1115.
5. Jacobs, J. M., Adkins, J. N., Qian, W.-J., Liu, T., Shen, Y., Camp, D. G. I., Smith, R. D., J. Prot. Res.
2005, 4, 1073–1085.
6. Rabilloud, T., Anal. Chem. 2000, 72, 48A–55A.
7. Gygi, S. P., Rist, B., Griffin, T. J., Eng, J., Aebersold, R., J. Prot. Res. 2002, 1, 47–54.
8. Peng, J., Elias, J. E., Thoreen, C. C., Licklider, L. J., Gygi, S. P., J. Prot. Res. 2003, 2, 43–50.
9. Rocklin, R. D., Ramsey, R. S., Ramsey, J. M., Anal. Chem. 2000, 72, 5244–5249.
10. Ramsey, J. D., Jacobson, S. C., Culbertson, C. T., Ramsey, J. M., Anal. Chem. 2003, 75, 3758–3764.
11. Gottschlich, N., Jacobson, S. C., Culbertson, C. T., Ramsey, J. M., Anal. Chem. 2001, 73, 2669–2674.
12. Herr, A. E., Molho, J. I., Drouvalakis, K. A., Mikkelsen, J. C., Utz, P. J., Santiago, J. G., Kenny, T. W.,
Anal. Chem. 2003, 75, 1180–1187.
13. Wang, Y.-C., Choi, M. H., Han, J., Anal. Chem. 2004, 76, 4426–4431.
14. Shadpour, H., Soper, S. A., Anal. Chem. 2006, 78, 3519–3527.
15. Evans, C. R., Jorgenson, J. W., Anal. Bioanal. Chem. 2004, 378, 1952–1961.
16. Becker, H., Lowack, K., Manz, A., J. Micromech. Microeng. 1998, 8, 24–28.
17. Chen, X., Wu, H., Mao, C., Whitesides, G. M., Anal. Chem. 2002, 74, 1772–1778.
18. Li, Y., Buch, J. S., Rosenberger, F., DeVoe, D. L., Lee, C. S., Anal. Chem. 2004, 76, 742–748.
19. Cui, H., Horiuchi, K., Dutta, P., Ivory, C. F., Anal. Chem. 2005, 77, 7878–7886.
20. Ivory, C. F., Electrophoresis 2007, 28, 15–25.
21. Xue, Q., Foret, F., Dunayevskiy, Y. M., Zavracky, P. M., McGruer, N. E., Karger, B. L., Anal. Chem.
1997, 69, 426–430.
22. Ramsey, R. S., Ramsey, J. M., Anal. Chem. 1997, 69, 1174–1178.
23. Figeys, D., Ning, Y., Aebersold, R., Anal. Chem. 1997, 69, 3153–3160.
24. Oleschuk, R. D., Harrison, D. J., Trends Anal. Chem. 2000, 19, 379–388.
25. Sung, W.-C., Makamba, H., Chen, S.-H., Electrophoresis 2005, 26, 1783–1791.
26. Limbach, P. A., Meng, Z., Analyst 2002, 127, 693–700.
27. Figeys, D., Pinto, D., Electrophoresis 2001, 22, 208–216.
28. Lion, N., Rohner, T. C., Dayon, L., Arnaud, I. L., Damoc, E., Youhnovski, N., Wu, Z. Y., et al.,
Electrophoresis 2003, 24, 3533–3562.
29. Marko-Varga, G., Nilsson, J., Laurell, T., Electrophoresis 2004, 25, 3479–3491.
30. Foret, F., Kusý, P., Electrophoresis 2006, 27, 4877–4887.
31. Foret, F., Preisler, J., Proteomics 2002, 2, 360–372.
32. DeVoe, D. L., Lee, C. S., Electrophoresis 2006, 27, 3559–3568.
33. Lee, H., Criffin, T. J., Gygi, S. P., Rist, B., Aebersold, R., Anal. Chem. 2002, 74, 4353–4360.
34. Ro, K. W., Liu, J., Knapp, D. R., J. Chromatogr. A 2006, 1111, 40–47.
35. Ericson, C., Phung, Q. T., Horn, D. M., Peters, E. C., Fitchett, J. R., Ficarro, S. B., Salomon, A. R.,
Brill, L. M., Brock, A., Anal. Chem. 2003, 75, 2309–2315.
36. Wang, Y., Zhou, Y., Balgley, B., Cooper, J., Lee, C. S., DeVoe, D. L., Electrophoresis 2005, 26,
3631–3640.
37. Onnerfjord, P., Ekstrom, S., Bergquist, J., Nilsson, J., Laurell, T., Marko-Varga, G., Rapid Comm.
Mass Spectrom. 1999, 13, 315–322.
38. Cui, H., Huang, Z., Dutta, P., Ivory, C. F., Anal. Chem. 2007, 79, 1456–1465.
34 Microchip Immunoassays
Kiichi Sato and Takehiko Kitamori
CONTENTS
34.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1013

34.2 Basics of Immunoassay Microchips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1014
34.3 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1016
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1019
34.1 INTRODUCTION
It is predicted that an enormous number of analyses will be required in the future in our daily lives,
especially in the clinical and biochemical fields. Rapid progress in molecular biology and biochem-
istry has brought about a deluge of useful and important information. From these achievements, a
significant number of genetic mutations have been elucidated, and a number of biomarkers con-
cordant with disease have been established. These are not only useful for revealing the presence
and, potentially, staging the disease but also for the a priori prediction of patient predisposition to
diseases. By simple calculation, the number of required diagnostic analyses is thought to be the
product of the population size and the number of indices. To perform all of these analyses, it is clear
that significant investment of funds, time, and resources will be required, which is very difficult to
realize with conventional analytical methods. One form of evolving technology, microfluidic devices
or microchips can be a powerful solution to this problem.
Microchip devices for chemical and biochemical analyses have been greatly advanced owing
to the progress of microfabrication techniques. Microchemical systems using these devices have
attracted much attention, not only from scientists and engineers but also from the clinical, forensic,
and biomedical research communities. Thus far, most studies describing microchip-based analytical
systems have focused on deoxyribonucleic acid (DNA) analysis by microchip electrophoresis with
laser-induced fluorescence detection. These microchip-based electrophoretic systems have great
advantages in some applications, especially in the clinical diagnostics and molecular biological
fields. Other analytical methods must also be performed for select applications, which involve several
distinct biochemical processes. To realize process functionality in these complicated systems, it is
necessary to utilize the chemical properties and potential of molecules involved in an effective way.
Immunoassay, founded by Rosalyn Yalow [1] in the late 1950s in the form of the “radio-
immunoassay,” is still today one of the most powerful and important analytical methods used in
clinical diagnoses and biochemical studies, primarily because of its extremely high selectivity and
sensitivity. The antibody-based method used today has two generic classifications: homogeneous
and heterogeneous. In heterogeneous immunoassays, the antibody is immobilized on a solid surface,
while homogeneous assays take place in a solution phase. Integration of both types of immunoas-
say into microchips has been accomplished with reasonable success by numerous research groups.
Most of the homogeneous immunoassays carried out in the microchip format have been based on
a microchip electrophoresis [2,3], where the benefit of the chip-based integration was primarily the
1013
reduction in separation time. In heterogeneous microchip immunoassay, however, a number of other

advantages exist—and this represents the focus of this chapter.
Enzyme-linked immunosorbent assay (ELISA) or other immunosorbent assay systems, in which
antigen or antibodies are fixed on a solid surface, are applicable to many analytes with high sen-
sitivity. These are used practically in many fields including clinical diagnoses and life science
researches. The conventional heterogeneous immunoassay, however, requires a relatively long assay
time and involves troublesome liquid-handling procedures and large quantities of expensive anti-
body reagents. Moreover, realization of point-of-care (POC) testing is difficult with the conventional
immunoassay, since rather large devices are necessary for automated practical diagnosis systems.
To overcome these drawbacks, a microchip-based system is effective. The integration of analytical
systems into a microchip should bring about enhanced reaction efficiency, a reduced assay time,
simplified procedures, and a lowered consumption of samples, reagents, and energy.
The main reason of the long assay time in the conventional heterogeneous assay is that the reaction
efficiency is very poor. This is because the reactions occur only on the solid surface. Moreover, it
takes long time to complete the reaction due to the long molecular diffusion time. In a microchip,
since it is easy to reduce the diffusion distance and increase the surface area-to-volume ratio, the
reaction time can be reduced to several minutes rather than hours or days.
A major source of the loss of reproducibility in the conventional assay is human error, because
the assay procedure contains many troublesome manual operations. In the microchip format, all
procedures can be controlled by pumps and valves that are regulated by a computer. Automation is
very favorable without the need for any large-scale robotic equipment.
Antibodies, enzymes, and other reagents used for immunoassay are expensive, hence, when
used in the conventional method causes a large problem. Moreover, in many cases, samples are very
precious or only a very small amount of samples can be obtained. Therefore, any reduction of the
consumption in the microchip systems is welcomed.
34.2 BASICS OF IMMUNOASSAY MICROCHIPS

Microfluidic devices for heterogeneous immunoassays consist of microchannels for transporting
solutions, reaction solid phase, and detection area. There are several important factors to develop a
high performance system.
Microchip substrate is open to a variety of materials. Silicon wafer is a good material to build up
microstructures if fabrication facilities are available. Glass has good chemical and optical properties,
and some polymers are cost effective for mass production. The surface treatment of the microchannel
is critically important for all materials. This is because of the nonspecific adsorption of the analytes
and antibodies to the channel wall that will result in considerable analytical error. It is very important
to modify the surface with some blocking reagents or other materials to prevent protein adsorption
before experiments. Therefore, we must choose a material of which, surface chemistry is well
understood.
Reaction solid phase is the core of the immunoassay microchip where the primary antibody
(capture antibody) is immobilized on the surface. There are some formats for the immobilization
(Figure 34.1). The simplest method is a fixation of the antibody to the channel wall [4]. Several
strategies have been used to adsorb or fix antibodies on the surface (e.g., direct adsorption), cova-
lent bonding to react with function groups on the surface, and microcontact printing. While direct
adsorption is common on hydrophobic polymer surfaces, covalent bonding by silanization reagents
is used for silicon and glass surface. Since the amount and activity of the immobilized antibody is
strongly dependent on the immobilization method, study of the fixation method is important [5].
Another type of solid support for capture antibody is microbeads. Beads coated with antibodies
are packed in the microchannel with a dam or cage structure. Bead format is considerably superior to
Microchip Immunoassays 1015
(a) (b)
(c) (d)
FIGURE 34.1 Formats of the reaction solid phase for microchip immunoassay. (a) channel wall, (b) packed
microbeads, (c) monolithic structure, and (d) magnetic beads.
(a) (b)
6.5 mm
100 µm
Antigen
1.5 mm
Antibody
45 µm
FIGURE 34.2 Schematic illustrations of the antigen–antibody reaction. (a) Microtiter plate and (b) microchip.
the wall surface immobilization methods. Microbeads bring extremely large surface area-to-volume
ratio and decrease the size of the liquid phase (i.e., decreasing the diffusion distance). These effects
are useful to reduce the assay time dramatically. An example is shown in Figure 34.2.
The surface area-to-volume ratio of a 50-µL solution in a microtiter plate well (0.65 mm in
diameter) was estimated to be 13 cm–1 , whereas that of the microchannel (11 beads in 100 µm ×
100 µm × 200 µm channel space) was 480 cm–1 . Therefore, the surface area-to-volume ratio the
microchannel was 37 times larger than that of the microtiter plate, and the reaction rate may be
increased because of this larger reaction field.
In the case of the conventional microtiter plate assay, a 1.5-mm movement would be necessary
for the most distant located antibody molecule to react with an antigen fixed on the surface of the
well, since the liquid depth was 1.5 mm. On the other hand, the liquid phase of the microchannel
filled with polystyrene beads was much smaller. The maximum distance from an antibody molecule
to the reaction-solid surface will not exceed 20 µm. Because the diffusion time is proportional to
the squares of the diffusion distance, the diffusion time of the antibody molecule to the antigen in
the microchip would be more than 5600 times shorter than in the conventional method [6].
The monolithic porous structure has also the same properties and suitable for the reaction solid
phase. However, it is difficult to remove the monolith from the chip after an assay so the chip is
then disposed. Hence, this happens even though the change of the beads is very easy. Therefore, the
monolithic format is not favorable unless the cost of the chip is very low.
Magnetic fine particles are another choice for the solid support. These particles can be captured
in the channel easily with a magnet without any microstructures inside. In many cases, however,
the particles are gathered very tightly and reagents are unable to go through them. That limits the
number of particles contributed to the reaction.
There are several kinds of detection methods for microchip-based heterogeneous immunoas-
says. The most popular technique is a fluorescent detection. Fluorescent molecules are used as a
labeling material and detection is done at the reaction solid phase. A combination of a laser-induced
fluorescence microscope, a highly sensitive detector, and a good labeling material may bring very
sensitive determination, while the beads may scatter light. Nonfluorescent material, colloidal gold
nanoparticle, is also useful for detection with a laser-induced thermal lens microscope (TLM).
Instead of the optical labels, enzymes are also used as a labeling material [7]. ELISA is one
of the most popular techniques in the conventional immunoassay and suitable in the microchip
format. By using a fluorogenic reaction, the fluorescent signal can be taken at the downstream
microchannel without interruption by microbeads. In the case of chromogenic reaction, a TLM is
useful [8].
34.3 APPLICATIONS
An example of the application of the microchip-based immunosorbent assay is shown below. The
human carcinoembryonic antigen (CEA), one of the most widely used tumor markers for serodi-
agnosis of colon cancer, was assayed with a microchip-based system [9]. An ultratrace amount
of CEA dissolved in serum samples was successfully determined within a short time. Polystyrene
beads pre-coated with anti-CEA antibody were introduced into a microchannel. Then a serum sam-
ple containing CEA, the first antibody, and the second antibody conjugated with colloidal gold,
First reaction Washing Second reaction
Mouse anti-CEA CEA Rabbit anti-CEA
Detection
Washing Third reaction Washing
Anti-lgG
Colloidal gold
FIGURE 34.3 Schematic illustrations of the microchip-based immunosorbent assay using microbeads.
(a) (b)
8
Signal intensity (Arbitrary unit)
100
Microchip assay (ng/mL)

6
0
10
C 0.01 0.1 1 10 100 0 2 4 6 8
CEA in serum (ng/mL) Conventional ELISA (ng/mL)
FIGURE 34.4 Determination of a tumor marker, human CEA. (a) Calibration curve for CEA in human sera.
(b) Correlation between the conventional ELISA and the microchip-based immunoassay.
were reacted successively (Figure 34.3). The resulting antigen–antibody complex, fixed on the
bead surface, was detected using a TLM. A highly selective and sensitive determination of an
ultratrace amount of CEA in human sera was made possible by a sandwich immunoassay sys-
tem that requires three antibodies for an assay. A detection limit in the orders of magnitude lower
than the conventional ELISA was achieved (Figure 34.4). Moreover and pertinent to the clinical
application of this technology, the analysis of serum samples from 13 patients showed a high cor-
relation (r = .917) with the conventional ELISA. This integration reduced the time necessary for
the antigen–antibody reaction to ∼1%, thus shortening the overall analysis time from 45 h to 35
min. This microchip-based diagnostic system might be the first micrototal analysis system (µTAS)
to show a practical usefulness for clinical diagnoses with short analysis times, high sensitivity, and
easy procedures.
In numerous microchip systems, researchers have shown that multiplexed analysis can be
achieved by increasing the number of channels available for parallel processing of numerous samples
simultaneously for protein crystallization [10] and genetic [11] analysis. This type of higher integra-
tion has been achieved for multiplexed immunoassay analysis capable of processing several samples
simultaneously [12]. In this integrated system, the chip had branching multichannels connected to
four reaction and detection regions; thus, the system could process four samples at a time with
only a single fluid pumping unit (Figure 34.5). To show utility, interferon gamma was assayed by a
three-step sandwich immunoassay with the system coupled to a TLM as a detector. The biases of the
signal intensities obtained from each channel were within 10%, with percent coefficient of variance
(CVs) were almost the same level as the single straight channel assay. The total assay time for all
four samples was 50 min when compared with 35 min for one sample in the single channel assay;
hence, a higher throughput was realized with the branching structure chip. The simultaneous assay
of many samples may also be achieved by simply arraying many channels in parallel on a chip. This
approach, however, requires many pumps and capillary connections, and high integration seems to
be difficult. On the other hand, a microchip with branching microchannels seems to be suitable for
carrying out the simultaneous assay, since the numbers of pumps and capillary connections required
for the system can be minimized.
Sample injection
Suction
Reagent injection
Injection
Drain
FIGURE 34.5 Immunoassay chip with a branching microchannels for simultaneous assay.

It is clear that much research has been done in this area, with the general sense that the microfluidic
immunoassay can be superior to conventional immunoassay systems. It is interesting that, despite
the large potential market size, microchip-based immunoassay systems have not yet been commer-
cialized. One of the main reasons is the difficulty associated with creating a truly high-throughput
assay without interfacing with automated fluidic control systems. While one microchip immunoas-
say can be completed in 30 min with manual operation, 96 assays can be completed within several
hours in the conventional format. Therefore, it is clear that parallelization and automation will be
required to realize the full potential. However, this will require the development and perfection of
compact, cheap, precise, and automatic fluidic controller systems (including pumps, valves, and
mixers). Precise control of submicroliter solutions without dead volume is a technical challenge and
it is difficult to realize it only with inexpensive devices—however, such developments are underway
in numerous labs. In addition, assay cost presents a problem, largely due to the cost of a 96-well
microtiter plate at less than $2 dollars, while microchip costs could approach $100 dollars. Clearly,
developing a cost-effective mass production approach will be a necessity. A candidate design for
the automated and compact system might be a compact disk (CD) format. It is very easy to array
more than 96 microchannels radially on a CD and all solutions/reagents can be controlled by cen-
trifugal force without any pumping. The structured portable CD player or CD drive in a laptop PC
might be a good analyzer as it can control the rotation speed precisely and has optical readout and
could accommodate fluorescent detection. Ultimately, development of a functional, easy-to-use, and
cost-effective system is a key to the acceptance and dissemination of such technology.
REFERENCES
1. Yalow, R. S., Berson, S. A. J. Clin. Invest. 1960, 39, 1157–1175.
2. Koutny, L. B., Schmalzing, D., Taylor, T. A., Fuchs, M. Anal. Chem. 1996, 68, 18–22.
3. Chiem, N. H., Harrison, D. J. Clin. Chem. 1998, 44, 591–598.
4. Bernard, A., Michel, B., Delamarche, E. Anal. Chem. 2001, 73, 8–12.
5. Yakovleva, J., Davidsson, R., Lobanova, A., Bengtsson, M., Eremin, S., Laurell, T., Emneus, J. Anal.
Chem. 2002, 74, 2994–3004.
6. Sato, K., Tokeshi, M., Odake, T., Kimura, H., Ooi, T., Nakao, M., Kitamori, T. Anal. Chem. 2000, 72,
1144–1147.
7. Eteshola, E., Leckband, D. Sens. Actuators B-Chem. 2001, 72, 129–133.
8. Sato, K., Yamanaka, M., Hagino, T., Tokeshi, M., Kimura, H., Kitamori, T. Lab Chip 2004, 4, 570–575.
9. Sato, K., Tokeshi, M., Kimura, H., Kitamori, T. Anal. Chem. 2001, 73, 1213–1218.
10. Hansen, C.L., Skordalakes, E., Berger, J.M., Quake, S.R. Proc. Natl Acad. Sci. USA. 2002, 99,
16531–16536.
11. Yeung, S.H., Greenspoon, S.A., McGuckian, A., Crouse, C.A., Emrich, C.A., Ban, J., Mathies, R.A.
J. Forensic Sci. 2006, 51, 740–747.
12. Sato, K., Yamanaka, M., Takahashi, H., Tokeshi, M., Kimura, H., Kitamori, T. Electrophoresis 2002,
23, 734–739.
35 Solvent Extraction on Chips
Manabu Tokeshi and Takehiko Kitamori
CONTENTS
35.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1021

35.2 Cocurrent Solvent Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1021
35.2.1 Simple Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1021
35.2.2 Ion Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1025
35.2.3 Liquid Membrane. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1027
35.2.4 Continuous Flow Chemical Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1028
35.2.5 Sample Preparation for Gas Chromatography and Mass Spectroscopy . . . . . . . . . . . 1030
35.3 Counter-Current Solvent Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1032
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1033
35.1 INTRODUCTION
Solvent extraction is a method capable of separating compounds based on their solution preferences
for two different immiscible liquids. In other words, it is the extraction of a substance from one liquid
phase into another liquid phase. Solvent extraction is a fundamental technique in chemical/biological
laboratories and chemical industries, where it is carried out in the macroscale using separatory
funnels. The advantage of doing a solvent extraction in the microspace on a chip is the scale merit
of using small dimensions, that is, a large specific interface area (the interface-to-volume ratio)
and a short diffusion distance, which together results in a short diffusion time. High speed and
high performance solvent extraction systems are possible on chips, without the need for mechanical
stirring, mixing, or shaking, all of which are necessary for the conventional method.
This chapter focuses on the analytical aspects of solvent extraction on microchips. Although there
are several reports in the literature on the application of chemical reactions,1−7 the topics dealt with
here are limited only to application using parallel (side-by-side) flow schemes inside microchannels.
Other examples, such as using a porous membrane,8 metal mesh,9 or droplets10 lie outside the scope
of this chapter.
35.2 COCURRENT SOLVENT EXTRACTION

35.2.1 SIMPLE APPROACHES
In order to perform an on-chip solvent extraction, a chip with at least two inlets for two immiscible
liquids is required. However, the difficulty associated with its use, and the resultant extraction
efficiency depends on the shape and dimensions of the channel, as well as other variables. Glass
is the substrate usually used when performing solvent extraction using chips due to its chemical
inertness toward organic solvents. If a polymer is to be used as the substrate, chemical modification
of the channel is typically required.
1021
The first solvent extraction carried out using a two-phase parallel flow in a microchannel was
demonstrated by Tokeshi et al.11 using the typical extraction set up shown in Figure 35.1. An aqueous
solution of Fe-bathophenanthrolinedisulfonic acid complex and a chloroform solution of capriquat
were introduced into the microchannel, using syringe pumps at a constant flow rate to establish a
parallel two-phase flow in the microchannel (Figure 35.2). In the case of ordinary solvent extraction
using a separatory funnel, the two solutions in the separatory funnel are separated vertically by
the difference in their specific gravities. Generally, in a microchannel, the liquid–liquid interface
does not have this vertical arrangement, instead the flows are positioned parallel to the sidewalls
Capillary tube
Teflom screw
O-ring
Waste
13 mm 10 mm 13 mm
8 mm
8 mm
46 mm
φ 0.5
0.77 mm
66 mm
FIGURE 35.1 Schematic illustration of experimental setup. (From Tokeshi, M. et al., Anal. Chem., 72, 1711,
2000.)
Solvent Extraction on Chips 1023
Aqueous phase
Fe-complex
Organic phase
Microsyringe pump
Drain
Capillary tube
Capillary tube
FIGURE 35.2 Schematic illustration of microextraction system. (From Tokeshi, M. et al., Anal. Chem., 72,
1711, 2000.)
of the microchannel. This is due to the fact that the interfacial tension and friction force are much
stronger than the force of gravity in the microspace provided by the microchannel.12 However,
this also depends on the microchannel geometry—for example, the properties of the wall surface
(hydrophobicity, hydrophilicity, wettability), the properties of liquids (viscosity, interfacial tension),
and so forth.13 The time taken for the extraction in the 250 µm (w) × 100 µm (d) microchannel of
Figure 35.2 was determined to be 45 s—this roughly coincides with the molecular diffusion time. The
extraction time was also at least one order of magnitude shorter when compared with the conventional
extraction time using a separatory funnel and mechanical mixing (shaking). In a similar manner, the
Ni-dimethylglyoxime complex,14 Co-2-nitroso-5-dimethylaminophenol (DMAP) complex,15 and
Al-2,2 -dihydroxyazobenzen (DHAB) complex16 were all analyzed by performing an extraction
at a two-phase interface, water/chloroform, water/toluene, and water/1-butanol, respectively. The
determination of radioactive nuclides, such as uranium, was carried out by the extraction of U(IV)
from an HNO3 aqueous solution to 30% (70% n-dodecane) and 100% tri-n-butyl phosphate phase
(TBP).17 In the conventional method, 100% TBP was not used as an extraction solvent because it
is impossible to get a separation between these two phases due to the fact that the specific gravity
of the two liquids is almost identical. Alternatively, with the chip, the extraction of U(IV) can be
achieved by using a two-phase aqueous/100% TBP flow because specific gravity is not a dominant
variable. Moreover, using a three-phase flow, it is possible to obtain a rapid solvent extraction as
illustrated in the following example, in which the extraction solvent m-xylene, was sandwiched by
two aqueous solutions of Co-2-nitroso-DMAP complex to form a microchannel-based three-phase
flow (Figure 35.3).12 The concentration of Co-2-nitroso-DMAP complex in the m-xylene phase was
determined by using thermal lens microscopy (TLM)18 and corresponded well with the time taken
for the extraction process. The dependence of the TLM signal on the distance x from the junction is
shown in the graph of Figure 35.3. The extraction process reached equilibrium after only a distance
of 1.5 cm, indicating that the extraction equilibrium was attained within 3 s of contact. The extraction
efficiency can, however, be improved by using specially designed microchannels, for example, an
asymmetrical zigzag-side-walled microchannel19 or a microchannel, which has intermittent partition
walls in the center of the confluent part.20
Uniquely, Minagawa et al.21 demonstrated the integration of two chemical processes on a chip:
a chelating reaction and solvent extraction. The layout of the microchannels fabricated on the glass
plates is shown in Figure 35.4 and a schematic illustration of the molecular behavior demonstrated
(a) CoDMAP aq.
m-xylene
CoDMAP aq.
0 x
(b)
TLM signal (arbitrary units)
150
100
50
0
0 5 10 15 20 25 30 35 40
x (mm)
FIGURE 35.3 (a) Schematic illustration of multilayer extraction system. (b) Position dependence of thermal
lens signal intensity on distance x. (From Hibara, A. et al., Anal. Sci., 17, 89, 2001.)
16
8 = 0.5
30º
32 mm
8
8
8 15 15 5 15 8
66 mm
FIGURE 35.4 Layout and dimensions of microchip. Microchannel was 100-µm deep and 250-µm wide.
(From Minagawa, T. et al., Lab Chip, 1, 72, 2001.)
in the microchannels is given in Figure 35.5. First, the m-xylene and the 2-nitroso-1-naphthol (NN)
aqueous solutions were introduced into the two inlets of the first Y-shaped microchannel by the
first syringe pump at a constant flow rate. These two liquids met at the intersection point, and a
parallel two-phase flow, that is, an organic/aqueous interface, was formed in the microchannel.
In this region, the NN in the aqueous phase could not be extracted into the m-xylene, whether
or not the liquids were flowing. Next, the sample solution of cobalt ion [Co(II)] was introduced
into the third inlet of the second Y-shaped microchannel using a second syringe pump at the same
constant flow rate. Therefore, a three-phase parallel flow was formed upstream of the intersec-
tion point of the second Y-shaped structure. There was almost no interdiffusion in the three fluids
under the flow condition, although the lack of interdiffusion depended on the flow rate. How-
ever, under these experimental conditions, interdiffusion can never occur. Thus, the chelating
m-xylene
NN/NaOH
Co complex
NN
Co2+ solution(aq.) Co2+
FIGURE 35.5 Schematic illustration of the chelating reaction and solvent extraction in microchannel. (From
Minagawa, T. et al., Lab Chip, 1, 72, 2001.)
reaction of Co(II) with the NN and the extraction of Co chelates did not occur under flow.
When the flow was stopped, the NN solution and the Co(II) sample solution promptly mixed,
and Co(II) reacted with NN. Then, the reaction product, Co chelate in the aqueous phase was
extracted into the organic phase. By using on-chip detection with a thermal lens microscope, trou-
blesome operations such as the phase separation necessary for the conventional system can be
avoided.
35.2.2 ION SENSING

The liquid microspace provides a short diffusion distance, and large specific interfacial area, of the
liquid–liquid interface. As a result several novel and attractive analytical features arise including
extremely fast ion sensing, and a need for only an ultra small volume of reagent solution. In contrast
to the slow response time of a standard ion-selective optode, where the response time is basically
governed by the slow diffusion of ionic species in the viscous polymer membrane, that of the on-chip
ion-sensing system is clearly faster due to the short diffusion distance and low viscosity of organic
solution. The ion-pair extraction scheme is an established methodology employed for ion-selective
optodes, which provide highly selective optical ion determination for various kinds of ions, using a
single lipophilic pH indicator dye and a highly selective neutral ionophore. However, exploitation
of an ion-pair extraction reaction and chip technology provide attractive advantages that would not
be achieved by conventional ion sensors. The advantages of on-chip ion-sensing systems are as
follows:
1. Reaction time (response time) can be dramatically reduced by the fast molecular transport
achievable in a microspace.
2. Since the required reagent solution volume is extremely small (∼100 nL), a fresh organic
phase can be used in every measurement. Subsequently, response degradation caused
by the leaching of ion-sensing components, a typical problem of ion-selective optodes,
would not need to be taken into consideration. This merit is directly reflected in the
excellent reproducibility of the response during continuous measurements and the effective
reduction of the amount of expensive reagents consumed in one measurement.
3. Ion determination can essentially be carried out by detection of the protonation/
deprotonation process of a single lipophilic anionic dye in the organic phase. There-
fore, no special color-responsive chelating reagents are required for the measurement of
another kind of analyte ion. One only need to alter the neutral ionophore for different
ion-selective ionophores. From the viewpoint of optical instrumentation, this merit is
quite important; there is no need to change the excitation source to match the excitation
wavelength of the different chelating reagents.
4. Highly selective ionophores or carriers, for various kinds of ions developed for applica-
tion to ion-selective electrodes, are commercially available and can be used without any
chemical modification.
Figure 35.6 shows a schematic illustration of the experimental setup and ion-pair extraction
scheme.22 The organic solution containing a neutral ionophore, a lipophilic pH indicator dye, and
an aqueous solution containing sample ion (K+ or Na+ ) were independently introduced into the
microchannel to form an organic/aqueous interface. Then determination of the ion was done using
a thermal lens microscope positioned downstream, where the ion-pair extraction occurred in the
organic phase under continuous-flow conditions. The response time and minimal required volume
of reagent for the on-chip ion-sensing system were about 8 s and 125 nL, respectively.
In other work, a sequential multi-ion-sensing system using a single chip was successfully iden-
tified by expanding the concept of on-chip ion sensing.23 Figure 35.7 shows the basic concept
for this chip-based multi-ion sensor. Different organic phases containing the same lipophilic pH
indicator dyes, but different ionophores were introduced sequentially into the microchannel by
the on–off switching of syringe pumps. In this case, having each organic phase contain differ-
ent ionophores avoids contamination. The aqueous sample solution containing the different ions is
introduced from the other inlet, to form a parallel two-phase flow with the intermittently pumped
Organic phase containing Thermal lens microscope

lipophilic dye and ionophore detection
Syringe pump
Waste
mm
10 mm
30
Water phase
containing
sample ion 60 mm
Channel width 250 µm

Channel depth 10 µm
TLM
Protonated detection
lipophilic
dye
Ion pair
Ionophore
Org.
250 µm
Aq.
Ion-pair
extraction
Ion Proton
FIGURE 35.6 Schematic illustration of experimental setup and ion-pair extraction in microchannel. (From
Hisamoto, H. et al., Anal. Chem., 73, 1382, 2001.)
C + io
noph
ore so
lvent
B + io
nopho
re so
lvent A +
ionopho
re
(a)-(c) Organic phases
containing lipophilic dye
and different ionophores
(a)
(b)
(c)
(d)
(d) Organic phase
without ionophore
TLM
Defection
Aqueous phase
containing multiple ions
Waste
Recorder
FIGURE 35.7 Concept of sequential ion-sensing system using single microchip. (From Hisamoto, H. et al.,
Anal. Chem., 73, 5551, 2001.)
organic phases. The selective ion-pair extraction reaction proceeds during flow; thus, different ions
can be selectively extracted into different organic phases, depending on the selectivity of the neu-
tral ionophores contained in the respective organic phases. Downstream in the flow, the ion-pair
extraction reaction becomes equilibrated; thus, downstream detection of the color change of the
organic phase, allowed for sequential and selective multi-ion sensing in the single aqueous sample
solution containing multiple ions. In this case, valinomycin and 2,6,13,16,19-pentaoxapentacyclo-
[18.4.4.4.7,12 0.1,20 07,12 ]dotriacontane (DD16C5)—known to exhibit high selectivity when used in
conventional ion sensors—were used as highly selective potassium and sodium ionophores, respec-
tively. Three types of aqueous sample solutions were analyzed with the system: a buffer solution
containing 10−2 M K+ , 10−2 M Na+ , or both ions. When the aqueous phases containing a single type
of ion were used, selective extractions occurred in each case; that is, potassium ions were extracted
only for an organic phase segment containing valinomycin, and sodium ions were extracted only
for that containing DD16C5. When the aqueous phase containing both ions was examined, both
ions were independently extracted into different organic phases, depending on the nature of the
ionophores in the respective organic phase (Figure 35.8). The minimum volume of single organic
phase needed to obtain an equilibrium response without dilution by cross dispersion of two organic
phases was ∼500 nL in the system, indicating that the required amounts of expensive reagents in
one measurement could be reduced to a few nanograms.
35.2.3 LIQUID MEMBRANE

A liquid membrane, composed of an organic phase with two aqueous phases, has a wide vari-
ety of industrial and analytical applications, including separation, concentration, and removal of
(a) KCL aq.

K+ K+ K+
(b) NaCL aq.

TLM signal intensity (au)
Na+ Na+ Na+
(c) KCI aq. & NaCI aq.

K+ K+ K+
Na+ Na+ Na+
2 2 2 2 (min)
Org.
Aq.
Time
FIGURE 35.8 Response profiles for different aqueous solutions obtained by intermittent pumping of organic
phase: (a) Aqueous solution containing 10−2 M KCl, (b) aqueous solution containing 10−2 M NaCl, and (c)
aqueous solution containing both 10−2 M KCl and 10−2 M NaCl. (From Hisamoto, H. et al., Anal. Chem., 73,
5551, 2001.)
analytes from wastewater, environmental, and biomedical samples. Surmeian et al.24 demonstrated
molecular transport on a chip using a stable three-layer flow membrane system, water/organic and sol-
vent/water. Under continuous-flow conditions, the analyte (methyl red) was rapidly extracted across
the microchannel from the donor to the acceptor phase through the organic phase (cyclohexene).
The thickness of the organic phase, sandwiched by the two aqueous phases, was ∼64 µm, and it was
considered a thin liquid organic membrane. Permeability studies identified the effects of molecular
diffusion, layer thickness, and organic solvent-water partition coefficient on the molecular transport.
In the chip, complete equilibration was achieved in several seconds, in contrast to a conventional
apparatus, where it would require tens of minutes for the comparable extraction. Maruyama et al.25
showed another selective separation by using a three-layer flow membrane, water/n-heptane/water,
to separate yttrium ions within a few seconds. In studies such as these, the formation of a stable
liquid/liquid interface inside the microchannel under flow conditions is very important. The use of
the microchannel structures and surface modification are particularly effective for stabilization of
the liquid/liquid interface inside the microchannel.7,20,24−28
35.2.4 CONTINUOUS FLOW CHEMICAL PROCESSING

The integration of chemical processes on a chip by using continuous-flow chemical processing
(CFCP) in combination with microunit operations, such as solvent extraction, phase separation, and
Reaction and extraction

area
Sample
Co2+
NN/ NN
NaOH
Decomposition and removal

Co chelate area
m-xylene
HCL TLM detection
Reaction & extraction area Metal ions
Sample Waste
HCI
Metal chelates/
NN/NaOH
m-xylene
m-xylene
Waste
NaOH
Decomposition & removal area NN

Metal chelate
NaOH
FIGURE 35.9 Schematic illustration of Co(II) determination by using CFCP.
so forth, under continuous flow conditions is a strategy used for the construction of real micro-
total analysis systems.27−30 The integration of a Co(II) wet analysis provides a good example of
CFCP,27 and Figure 35.9 schematically illustrates this analysis. The microchip consists of two dif-
ferent areas: the former is the reaction and extraction area, and the latter is the washing, that is,
decomposition and removal, area. In the former area, a sample solution containing Co(II) ions, a
NN solution and m-xylene, are introduced at a constant flow rate through three inlets using syringe
pumps. The three liquids meet at the intersection point, and a parallel two-phase flow, consisting
of an organic/aqueous interface, forms in the microchannel. The chelating reaction of Co(II) and
NN and extraction of the resulting Co(II) chelate proceed as the reacting mixture flows along the
microchannel. Since the NN reacts with the coexisting metal ions [such as Cu(II), Ni(II), and Fe(II)],
these coexisting metal chelates are also extracted into m-xylene. Therefore, a postextraction washing
process is necessary for the decomposition and removal of the undesired coexisting metal chelates.
The coexisting metal chelates decompose when they make contact with hydrochloric acid, and the
metal ions are moved into an HCl solution (back extraction). The decomposed chelating reagent,
NN, is dissolved in a sodium hydroxide solution. In contrast to the coexisting metal chelates, the
Co chelate is stable in HCl and NaOH solutions, and remains. In the latter (washing) area, the
m-xylene phase containing Co chelates and the coexisting metal chelates from the former (reaction
and extraction) area is interposed between the other two inlets at a constant flow rate. Then, the
three-phase flow, HCl/m-xylene/NaOH, forms in the microchannel. The decomposition and removal
of the coexisting metal chelates proceed along the microchannel in the same manner as described
above. Finally, the target chelates in m-xylene are detected downstream. Cobalt in an admixture
sample was successfully determined with a rapid analysis time of less than 1 min. The advantages of
this approach, compared with a conventional method, are the simplified nature of the procedure and
avoiding troublesome operations. In the conventional method, the acid and alkali solutions cannot
be used simultaneously, and alternative washing procedures must be repeated several times. The
same effect can be obtained by using three-phase flow in the microchannel. In a subsequent paper,
Kikutani et al.31 expanded the concept of CFCP to integrate four parallel analyses of Co(II) and
Fe(II) on a chip in a three-dimensional microchannel network.
a b
e
LPS O2
(a) Cells NO NO2– + NO3–
Nitrate reductase
(b) NO3– NO2–
H+ N N+
(c) NO2– + H2NO2S NH2 H2NO2S
NH
(d) H2NO2S NH2 H NO S N N NH

N N+ 2 2
NH2
(e) Detection with a thermal lens microscope
FIGURE 35.10 Chemical processes carried out in microchip for bioassay of macrophage-stimulating agent.
(From Tokeshi, M. et al., Anal. Chem., 74, 1565, 2002.)
Goto et al.32 demonstrated the integration of a bioassay system that illustrated all processes
required for a bioassay, that is, cell culture, chemical stimulation of cells, chemical and enzymatic
reactions, and detection, on a chip using CFCP (Figure 35.10). By using the temperature control
device, spatial temperature control of the system was possible, with areas on the chip maintained at
different temperatures. Nitric oxide released from macrophage-like cells stimulated by lipopolysac-
charide was successfully monitored by this system. The total assay time was reduced from 24 to 4 h,
and the detection limit for nitric oxide was improved from 1 × 10−6 m to 7 × 10−8 M compared with
the conventional batch methods. Moreover, the system could monitor a time course of the release,
which is difficult to measure by conventional methods.
Smirnova et al.33 demonstrated the determination of the insecticide, carbaryl, using a two-chip
system. The first chip (for the hydrolysis of carbaryl) had a simple Y-shaped channel while the second
chip (for the diazo coupling reaction between hydrolyzed products and 2,4,6-trimethylaniline)—the
extraction required special channel shapes with a partial surface—modification obtained by using
capillary-restricted modification (CARM) (Figure 35.11).34 Determination of carbaryl pesticide in
water with sufficient sensitivity was carried out with an analysis time of 8 min. In a similar manner,
Honda et al.35 developed a combination of a tube-type enzyme-immobilized microreactor and a
microextractor with partial surface modification to produce optically pure amino acids.
35.2.5 SAMPLE PREPARATION FOR GAS CHROMATOGRAPHY AND

MASS SPECTROSCOPY
As the successful applications of a commercially available liquid chromatography (LC) chip for mass
spectroscopy (MS) have been made apparent, the development of sample preparation chips for other
= 532 mm
Toluene
TMA Chip 2 Coupling reaction,
extraction & detection
NaOH
Carbaryl Toluene Extraction
Chip 1 Carbaryl hydrolysis
ODS Interface Buffer
FIGURE 35.11 Concept of integration of carbaryl determination onto microchips. Target carbaryl was
hydrolyzed in chip 1 to the product 1-naphthol. In chip 2,1-naphthol was coupled with diazonium ion and
was extracted into the organic phase. (From Smirnova, A. et al., Anal. Chim. Acta, 558, 69, 2006.)
1/ether, Synthesis microchip

Capillary tubing
100 mM
Extraction microchip
TF AA/ether,
100 mM
Mass spectrometer
Capillary tubing
Syringe pumps
Phosphate buffer pH 6.8
FIGURE 35.12 Schematic illustration of synthesis microchip-extraction microchip-ESI-MS. (From

Takahashi, Y. et al., J. Mass Spectrom. Soc. Jpn., 54, 19, 2006.)
conventional analytical systems, such as gas chromatography (GC), are also strongly desirable. As
described above, microfluidic chips have great potential to integrate chemical processing such as
sample preparation (see Chapter 44 by Biemvenue). Recently, several groups have developed the
sample preparation chips for GC and GC–MS, and MS analyses.
Miyaguchi et al.36 developed a solvent extraction chip for the off-chip GC analysis of
amphetamine-type stimulants (a class of illegal drugs) in urine. Microchannels, partially fluoroalky-
lated by the CARM method,34 were employed for stabilizing the 1-chlorobutane/alkalize urine
interface and obtaining a GC-amenable fraction. As a practical demonstration, methoxyphenamine
hydrochloride was administered to three healthy volunteers, and the concentration of
methoxyphenamine in their urine determined. This showed the potential of chip-based sample
preparation to contribute to the rapid automated analysis in forensic toxicology. Similarly, Xiao
et al.37 carried out solvent extraction of ephedrine using a polydimethylsiloxane (PDMS)/glass
chip that had been obtained by surface modification of half of the glass wall, with octadecyl-
trichlorosilane. On-chip extraction of hydrocarbons from a North Sea oil for GC-MS analysis was also
reported.38
Figure 35.12 shows a schematic diagram of an online high-throughput detection system for a
reaction product synthesized by a microreactor.39 The system has a synthesis chip that a microreactor
for the synthesis of 2,2,2-trifluore-N-phenetyl acetamide (TPA), an extraction chip for the purification
of TPA from the reaction mixture, and an electrospray ionization mass spectrometer (ESI-MS) for
detection. In this system, the extraction chip is placed between the microreactor and the ESI-MS to
reduce the ionization hindrances for the reaction products. This placement makes the detection of
the reaction product possible under any reaction conditions. The system as described here may be
useful for the optimization of reaction conditions and the screening of valuable chemicals such as
precursors of pharmaceuticals.
35.3 COUNTER-CURRENT SOLVENT EXTRACTION

From the viewpoint of recovery efficiency, cocurrent solvent extraction as described above, can reach
a theoretical plate number of only unity. In contrast to cocurrent solvent extraction, a higher theo-
retical plate number is expected using counter-current solvent extraction. However, counter-current
flow in a microchannel cannot be easily formed because of interfacial tension and viscous force of
both (aqueous and organic) phases. In an ordinary microchannel, counter-current flow cannot occur
because the two phases collide, and high shear stress at the liquid/liquid interface causes breakup.
To form counter-current flow, the aqueous solution must flow along one side of the channel, and
the organic solution flow along the other, without breakup. Aota et al.40 successfully developed
a counter-current flow system, which was obtained by selectively modifying the lower half of a
microchannel with a hydrophobic group (Figure 35.13). Using this system, they realized that a theo-
retical plate number of 4.6 was needed for the extraction of a Co(II)-2-nitroso-dimethylaminopyridine
(DMAP) complex in an aqueous-toluene counter-current flow. Counter-current flow is expected
to be applicable to enrichment processes for various environmental analyses and biomolecular
separations.
(a) 70 mm
m
30 m
Hydrophilic Liquid/liquid
surface interface
Microchannel
Top plate
(hydrophilic)
Bottom plate org.

(hydrophobic) aq. Hydrophobic
surface
20 mm
(b) Butylacetate 300 µm Water

Water
Butylacetate
Water Butylacetate
20 mm
FIGURE 35.13 (a) Photograph of microchip and (b) schematic illustration of counter-current extraction
system. (From Aota, A. et al., Angew. Chem. Int. Ed., 46, 878, 2007.)

The development and application of chip-based solvent extraction is gradually expanding as shown
in this chapter. With methods for surface modification, the liquid–liquid interface in a microchannel
can be stabilized, and so special techniques for solvent extraction are not required. Therefore, since
solvent extraction is one of the basic chemical processes, it is expected that it will be used more
often in the future.
In the next 5 years, from a practical viewpoint, the combination of solvent extraction chips
with conventional analytical apparatuses, such as GC, LC, MS, and so forth, and the integration of
solvent extraction and other chemical processes will become more and more important. From the
viewpoint of the basic science of two-phase and multi-phase flows, the flow itself near the liquid–
liquid interface is very interesting. In fact, chaotic vortices and transient vortex-like flows have been
observed around the liquid–liquid interface in the counter-current flow.41
ACKNOWLEDGMENTS
We would like to thank our coworkers whose work has been referenced in this chapter. Kanagawa
Academy of Science and Technology and The Ministry of Education, Culture, Sports, Science and
Technology of Japan are acknowledged for financial support.
REFERENCES
1. Hisamoto, H., Saito, T., Tokeshi, M., Hibara, A. and Kitamori, T., Fast and high conversion
phase-transfer synthesis exploiting the liquid–liquid interface formed in a microchannel chip, Chem.
Commun., 2662, 2001.
2. Ueno, M., Hisamoto, H., Kitamori, T. and Kobayashi, S., Phase-transfer alkylation reactions using
microreactors, Chem. Commun., 936, 2003.
3. Shaw, J., Nudd, R., Naik, B., Turner, C., Rudge, D., Benson, M. and Garman, A., Liquid/liquid extrac-
tion systems using micro-contactor arrays, Proc. µTAS 2000 Symposium, Enschede, The Netherlands,
371, 2000.
4. Ehrfeld, W., Hessel, V. and Löwe, H., Microreactors, Wiley-VCH, Weinheim, 2000, Chapter 5.
5. Kikutani, Y., Horiuchi, T., Uchiyama, K., Hisamoto, H., Tokeshi, M. and Kitamori, T., Glass microchip
with three-dimensional microchannel network for 2 × 2 parallel synthesis, Lab Chip, 2, 188, 2002.
6. Kikutani, Y., Hibara, A., Uchiyama, K., Hisamoto, H., Tokeshi, M. and Kitamori, T., Pile-up glass
microreactors, Lab Chip, 2, 193, 2002.
7. Maruyama, T., Uchida, J., Ohkawa, T., Futami, T., Katayama, K., Nishizawa, K., Sotowa, K, Kob-
ota, F., Kamiya, N. and Goto, M., Enzymatic degradation of p-chlorophenol in a two-phase flow
microchannel system, Lab Chip, 3, 308, 2003.
8. Cai, Z.-X., Fang, Q., Chen, H.-W. and Fang, Z.-L., A microfluidic chip based liquid–liquid extraction
system with microporous membrane, Anal. Chim. Acta, 556, 151, 2006.
9. Wenn, D. A., Shaw, J. E. A. and Machenzie, B., A mesh microcontactor for 2-phase reactions, Lab
Chip, 3, 180, 2003.
10. Chen, H., Fang, Q., Yin, X.-F. and Fang, Z.-L., Microfluidic chip-based liquid-liquid extraction and
preconcentration using a subnanoliter-droplet trapping technique, Lab Chip, 5, 719, 2005.
11. Tokeshi, M., Minagawa, T. and Kitamori, T., Integration of a microextraction system on a glass
chip: Ion-pair solvent extraction on Fe(II) with 4,7-diphenyl-1,10-phenanthrolinedisulfonic acid and
tri-n-octylmethylammonium chloride, Anal. Chem., 72, 1711, 2000.
12. Hibara, A., Tokeshi, M., Uchiyama, K., Hisamoto, H. and Kitamori, T., Integrated multilayer flow
system on a microchip, Anal. Sci., 17, 89, 2001.
13. Kuban, P., Berg, J. and Dasgupta, P. K., Vertically stratified flows in microchannels. Computational
simulations and applications to solvent extraction and ion exchange, Anal. Chem., 75, 3549, 2003.
14. Sato, K., Tokeshi, M., Sawada, T. and Kitamori, T., Molecular transport between two phases in a
microchannel, Anal. Sci., 16, 455 2000.
15. Tokeshi, M., Minagawa, T. and Kitamori, T., Integration of a microextraction system: solvent extraction
of Co-2-nitroso-5-dimethylaminophenol complex on a microchip, J. Chromatogr. A, 894, 19, 2000.
16. Kim, H.-B., Ueno, K., Chiba, M., Kogi, O. and Kitamura, N., Spatially-resolved fluorescence spec-
troscopic study on liquid/liquid extraction processes in polymer microchannels, Anal. Sci., 16, 871,
2000.
17. Hotokezaka, H., Tokeshi, M., Harada, M., Kitamori, T. and Ikeda, Y., Development of the innovative
nuclide separation system for high-level radioactive waste using microchip-extraction behavior of
metal ions from aqueous to organic phase in microchannel, Prog. Nucl. Energy, 47, 439, 2005.
18. Kitamori, T., Tokeshi, M., Hibara, A. and Sato, K., Thermal lens microscope and microchip chemistry,
Anal. Chem., 76, 52A, 2004.
19. Ueno, K., Kim, H.-B. and Kitamura, N., Channel shape effects on the solution-flow characteristics
and liquid/liquid extraction efficiency in polymer microchannel chips, Anal. Sci., 19, 391, 2003.
20. Maruyama, T., Kaji, T., Ohkawa, T., Sotowa, K., Matsushita, H., Kubota, F., Kamiya, N., Kusakabe,
K. and Goto, M., Intermittent partition walls promote solvent extraction of metal ions in a microfluidic
device, Analyst, 129, 1008, 2004.
21. Minagawa, T., Tokeshi, M. and Kitamori, T., Integration of a wet analysis system on a glass chip: deter-
mination of Co(II) as 2-nitroso-1-naphthol chelates by solvent extraction and thermal lens microscopy,
Lab Chip, 1, 72, 2001.
22. Hisamoto, H., Horiuchi, T., Tokeshi, M., Hibara, A. and Kitamori, T., On-chip integration of neutral
ionophore-based ion pair extraction reaction, Anal. Chem., 73, 1382, 2001.
23. Hisamoto, H., Horiuchi, T., Uchiyama, K., Tokeshi, M., Hibara, A. and Kitamori, T., On-chip inte-
gration of sequential ion-sensing system based on intermittent reagent pumping and formation of
two-layer flow, Anal. Chem., 73, 5551, 2001.
24. Surmeian, M., Slyadnev, M. N., Hisamoto, H., Hibara, A. and Kitamori, T., Three-layer flow membrane
system on a microchip for investigation of molecular transport, Anal. Chem., 74, 2014, 2002.
25. Maruyama, T., Matsushita, H., Uchida, J., Kubota, F., Kamiya, N. and Goto, M., Liquid membrane
operations in a microfluidic device for selective separation of metal ions, Anal. Chem., 76, 4495,
2004.
26. Hibara, A., Nonaka, M., Hisamoto, H., Uchiyama, K., Kikutani, Y., Tokeshi, M. and Kitamori, T.,
Stabilization of liquid interface and control of two-phase confluence and separation in glass microchips
by utilizing octadecylsilane modification of microchannels, Anal. Chem., 74, 1724, 2002.
27. Tokeshi, M., Minagawa, T., Uchiyama, K., Hibara, A., Sato, K., Hisamoto, H. and Kitamori, T.,
Continuous-flow chemical processing on a microchip by combining microunit operations and a
multiphase flow network, Anal. Chem., 74, 1565, 2002.
28. Kikutani, Y., Ueno, M., Hisamoto, H., Tokeshi, M. and Kitamori, T., Continuous-flow chemical
processing in three-dimensional microchannel network for on-chip integration of multiple reaction in
a combinatorial mode, QSAR Comb. Sci., 24, 742, 2005.
29. Sato, K., Hibara, A., Tokeshi, M., Hisamoto, H. and Kitamori, T., Integration of chemical and
biochemical analysis systems into a glass microchip, Anal. Sci., 19, 15, 2003.
30. Tokeshi, M., Kikutani, Y., Hibara, A., Sato, K., Hisamoto, H. and Kitamori, T., Chemical process on
microchips for analysis, synthesis and bioassay, Electrophoresis, 23, 3583, 2003.
31. Kikutani, Y., Hisamoto, H., Tokeshi, M. and Kitamori, T., Micro wet analysis system using multi-phase
laminar flows in three-dimensional microchannel network, Lab Chip, 4, 328, 2004.
32. Goto, M., Sato, K., Murakami, A., Tokeshi, M. and Kitamori, T., Development of a microchip-based
bioassay system using cultured cells, Anal. Chem., 77, 2125, 2005.
33. Smirnova, A., Mawatari, K., Hibara, A., Proskurnin, M. A. and Kitamori, T., Micro-multiphase laminar
flow for the extraction and detection of carbaryl derivative, Anal. Chim. Acta, 558, 69, 2006.
34. Hibara, A., Iwayama, S., Matsuoka, S., Ueno, M., Kikutani, Y., Tokeshi, M. and Kitamori, T., Surface
modification method of microchannels for gas-liquid two phase flow in microchips, Anal Chem, 77,
943, 2005.
35. Honda, T., Miyazaki, M., Yamaguchi, Y., Nakamura, H. and Maeda, H., Integrated microreaction
system for optical resolution of racemic amino acids, Lab Chip, 7, 366, 2007.
36. Miyaguchi, H., Tokeshi, M., Kikutani, Y., Hibara, A., Inoue, H. and Kitamori, T., Microchip-based
liquid-liquid extraction for gas-chromatography analysis of amphetamine-type stimulants in urine,
J. Chromatogr. A, 1129, 105, 2006.
37. Xiao, H., Liang, D., Liu, G., Guo, M., Xing, W. and Cheng, J., Initial study of two-phase laminar flow
extraction chip for sample preparation for gas chromatography, Lab Chip, 6, 1067, 2006.
38. Bowden, S. A., Monaghan, P. B., Wilson, R., Parnell, J. and Cooper, J. M., The liquid–liquid diffusive
extraction of hydrocarbons from a North Sea oil using a microfluidic format, Lab Chip, 6, 740, 2006.
39. Takahashi, Y., Sakai, R., Sakamoto, K., Yoshida, Y., Kitaoka, M. and Kitamori, T., On-line high-
throughput ESIMS detection of a reaction product using synthesis and extraction microchips, J. Mass
Spectrom. Soc. Jpn., 54, 19, 2006.
40. Aota, A., Nonaka, M., Hibara, A. and Kitamori, T., Countercurrent laminar microflow for highly
efficient solvent extraction, Angew. Chem. Int. Ed., 46, 878, 2007.
41. Aota, A., Hibara, A. and Kitamori, T., Dependence of the number of theoretical plates of micro
counter-current extraction of flow rate, Proc. µTAS 2005 Symposium, Boston, USA, 118, 2005.
36 Electrophoretic Microdevices
for Clinical Diagnostics
Jerome P. Ferrance
CONTENTS
36.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1037

36.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1038
36.4.1 DNA Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1039
36.4.1.1 Genotyping and Mutation Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1039
36.4.1.2 Other Nucleic Acid Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1041
36.4.2 Protein Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1042
36.4.2.1 Protein Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1043
36.4.2.2 Microchip Immunoassays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1044
36.4.3 Additional Clinically Relevant Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1046
36.4.4.1 Enzymatic Reactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1049
36.4.4.2 Purification and/or Concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1050
36.4.5 Chip-Based Integration of Sample Preparation Steps. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1051
36.4.5.1 DNA Amplification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1052
36.4.5.2 Onboard Sample Preparation for Other Analytes. . . . . . . . . . . . . . . . . . . . . . . . 1054
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1057
36.1 INTRODUCTION
Microfluidic devices are changing the way analytical procedures are performed. The speed that they
bring to an analysis, the small sample size and limited reagent use, integration of multiple processing
steps into a single unit, as well as the possibility for on-site analysis, all point to changes in the way
analytical methods are to be used in the relatively near future. Nowhere is this going to be more
important than in the field of clinical analysis. Microfluidic devices in the doctor’s office, at the
patient’s bedside, or at any point-of-care will be as common and easy to use as the blood glucose
and home pregnancy testing devices currently available. In the operating room, where rapid analysis
while a patient is still in surgery can provide information to the surgeon regarding the extent of
damage, to the emergency room, where rapid feedback could identify agents that might cause an
epidemic, or are the result of a terrorist attack, the clinical analysis methods provided by microchips
will be invaluable. Developments in microfluidics methods directed toward the clinical laboratory
have been progressing over the past few years, and fully integrated devices are now being reported.
This chapter will review what has been achieved, and look to what is still needed to move this
technology into the mainstream of clinical diagnostics.
1037
36.2 BACKGROUND
Not long after the early development of microchips for electrophoretic separations of DNA, the
potential of these devices for clinical applications was realized.1−3 While microchip electrophoresis
remains a major area in clinical analysis to which microchip processing is applied, implementation
of additional processing and analysis steps on microchips over the past few years is also providing
new devices designed to replace existing methods utilized in the clinical laboratory.4−6 The goal
of many of these efforts is to eliminate the need for sample transportation to and processing in a
central laboratory, by bringing the microfluidic technology directly to the point of care.7 A report
sponsored by the National Cancer Institute discusses the development of point-of-care systems for
cancer diagnosis and prognosis, in which integrated microfluidic systems play an essential role.8
The purpose of this chapter is to explore clinical analyses that have been transferred to an
electrophoretic microchip, along with some of the microchip preparatory methods that are now
being integrated with an electrophoretic analysis step to provide lab-on-a-chip type devices. There
are a vast number of clinical analyses that are now being transferred to microdevices, but the scope of
this chapter is limited to only those that have an electrophoretic component in accord with the nature
of this volume. A number of clinical analysis methods simply take advantage of the microfluidic
architecture in microchips to decrease sample size and reaction times without requiring a separation
step. The most common example of these is microarrays, recently reviewed by Situma et al.,9 that
utilize oligonucleotide to detect the presence of specific DNA sequences.10−14 Similar types of
immunoassays for clinical detection of specific proteins have also been developed on microfluidic
devices by immobilizing antibodies within microchannels;15,16 while these are not included in this
chapter, a number of reviews17,18 as well as an earlier Chapter 34 in this book by Sato and Kitamori
cover these types of microchip immunoassays with applications to clinical analyses in more detail.
Protein immunoassays on microchips that incorporate a separation step are included in this chapter
along with analyses for specific ions,19 glucose,20 pH,21 enzymatic activity,21,22 and other types of
bioaffinity interactions.23
The use of mass spectroscopy in clinical analysis is increasing, but is not yet widely employed,
thus electrophoretic microchips designed for interfacing with MS detection are also not included
here; these devices are covered in Chapter 49 by Laurell for coupling with matrix-assisted laser
desorption/ionization-mass spectrometry (MALDI-MS) detection, and Chapter 53 by Lazar for
electrospray-MS. A number of microchip applications also relate to the trapping, growth, and anal-
ysis of cells on microdevices, but these are not methods utilized in clinical analysis and thus are not
included. The reader is directed to a recent review of these techniques by El-Ali et al.24 as well as
in Chapter 32 by Keenan and Beebe. Even with these limitations, the number of clinical methods
that have been transferred to microchips is impressive and a comprehensive review is not possible.
A number of reviews on microchip separations for clinical analyses have been written,25−29 the
most recent of which reported on microchip separations utilized for detection of cancer, cardiovas-
cular disease, renal disease, neurological disease, immune disorders, diabetes, hereditary disorders,
thyroid functioning, and infectious agents in body fluids.29

Microchip electrophoretic separations themselves and the theory behind them have been covered in
Chapters 33 by DeVoe and Lee and 55 by Mahmoudian et al. This is also true of sample processing
methods on microdevices, which have been covered in Chapter 43 by Bienvenue and Landers,
Chapter 50 by van Midwoud and Verpoorte. The theory and engineering behind integration of
multiple processes on a single device, in which both sample pretreatment and analysis steps are
performed without removal of the sample from the device, are also explored in Chapter 43 by
Bienvenue and Landers, as well as Chapter 40 by Easley and Landers describing valves to control
flow within and between processes. This chapter seeks to show how the methods and theory discussed
Electrophoretic Microdevices for Clinical Diagnostics 1039
in these other chapters are being applied to real problems in clinical analysis. In addition, some of
the issues involved in developing microdevices for utilization in the clinical laboratory not covered
elsewhere will be described.

36.4.1 DNA ANALYSES
The ability to use laser-induced fluorescence (LIF) for detection allowed DNA separations on
microchips to advance more rapidly than separations of other types of clinically relevant compounds.
Chapter 6 by Szántai and Guttman reviewed the use of capillary electrophoresis of DNA for clinical
diagnostic purposes, and any of those methods can easily be transferred to the microchip platform.
Most often, clinical microchip DNA diagnostic devices utilize separations of specific fragments of
DNA amplified using the polymerase chain reaction (PCR). Nucleic acid targets of clinical interest
for amplification include mutations associated with particular diseases, the presence of exogenous
DNA or RNA from pathogens, or over expressed DNA indicating the presence of a single clonal
variation. In the coming genetic medicine revolution, where treatment options are also based on
particular DNA sequences within a patient’s genome, the use of microchip DNA separations will
expand even further in the clinical laboratory.
36.4.1.1 Genotyping and Mutation Detection

One of the most common methods for detection of mutations in PCR-amplified DNA fragments is
single-stranded conformational polymorphism (SSCP) analysis. Tian et al.30 reported microchip
separations for the detection of common mutations in BRCA1 and BRCA2, two breast cancer
susceptibility genes specific for the Ashkenazi Jewish population. Separations were performed in
140 s, fourfold faster than the CE-based assay, with profiles from the wild-type and mutant alle-
les easily distinguishable. As was important in the development of all of these clinical microchip
methods, no loss of diagnostic ability was seen in moving to the microfluidic analysis method. This
method was further pursued by Kang et al.31 for rapid detection of a point mutation in the obesity
gene. In addition to the SSCP technique, heteroduplex analysis (HDA) is a second method utilized
for mutation analysis that has been transferred to microchips. Hestekin et al.32 used a combination
of SSCP and HDA separations on a microchip to evaluate mutations in specific exons in the p53
gene, a gene for a transcription factor important in regulating the cell cycle and preventing tumor
growth. Figure 36.1 shows microchip separations of fragments generated from samples of a wild
type and an exon 8 mutant of the p53 gene. Tian et al.33 also utilized HDA on a microchip, com-
bining it with allele-specific PCR to identify additional mutations in the BRCA genes as well as the
PTEN tumor suppressor gene. The clinical potential for all of these electrophoretic methods based
on migration differences due to the presence of mutations was recently reviewed by Hestekin and
Barron.34
The mutations for which the above methods are most applicable are simple deletion and insertion
mutations affecting only a few bases. Substitution mutations are difficult to monitor in this way,30
but represent a significant class of DNA mutations called single nucleotide polymorphisms (SNP),
which will play a significant role in the area of personalized medicine. Woolley et al.35 performed
SNP detection in HLA-H, a marker gene for hemochromatosis, on a 12-channel microchip using
selective restriction digestion of a PCR-amplified product followed by traditional DNA sizing elec-
trophoresis. Shi et al.36 utilized this same method on a 96-channel radial array microchip for SNP
detection in the gene for methylenetetrahydrofolate reductase, the protein product of which regulates
folate and methionine metabolism. This same 96-channel microchip design was used by Medintz
et al.37 to evaluate a polymorphism in the HFE gene resulting in hereditary hemochromatosis using
(a) p56 Exon 8 Wild-type (b) p53 Exon 8 Wild-type + Mutant

Temperature = ambient Codon 282, CGG to TGG
1.8 Rev Temperature = ambient Rev
dsDNA ssDNA 0.5 wt + mut
1.5 Home- ssDNA
0.4 duplexes Hetero-
1.2
duplexes Fwd
0.3
RFU
0.9 mut
RFU
Fwd ssDNA
0.6 0.2 wt
Fwd
ssDNA ssDNA
0.3 0.1
0 0
360 380 400 420 440 460 480 405 425 445 465 485 505 525
Migration time (s) Migration time (s)
FIGURE 36.1 Electropherograms showing analysis of p53 exon 8 amplicons for mutation detection by tandem
SSCP/HA. dsDNA peaks are identified by the overlap of the FAM (forward strand) and JOE (reverse strand)
dyes. ssDNA peaks are identified by the predominance of one fluorescent dye. Peaks due to the mutation are
indicated by the arrows. Separation conditions: ambient temperature, 0.1% PHEA dynamically coated channel,
350–450 V/cm applied electric field strengths. (Reprinted from Hestekin, C. N., et al., Electrophoresis, 27,
allele-specific primers to generate different sized products. Figure 36.2 shows the separation and
detection of both homozygote and heterozygote mutations in the S62C variants of the HFE gene on
a microchip. The authors extended this work to simultaneously detect three SNP mutations within
this same gene based on the sizes of the DNA fragments produced.38
Simple size-based electrophoretic separations in microchips can also be employed for inser-
tion and deletion mutations, which modify the size of the PCR-amplified DNA fragments. Sung
et al.39 utilized this method on a plastic microchip to evaluate fragile X alleles that have mul-
tiple repeat units of a trimeric CGG repeat. A tetranucleotide repeat in microsatellite alleles,
related to hypercholesterolemia in families, was investigated by Cantafora et al.40 using the Agi-
lent 2100 Bioanalyzer, a commercially available microchip electrophoresis instrument capable
of analyzing up to 12 samples in 30 min. This same instrumentation was used in the evalua-
tion of deletion and duplication mutations associated with Duchenne muscular dystropy in two
papers by Ferrance et al.,41,42 which looked at a total of 13 loci in the dystrophin gene using two
PCR amplifications and electrophoretic separations. A total of 50 samples were evaluated in the
initial work, with all 35 samples with mutations being identified.41 Figure 36.3 shows the tra-
ditional Southern blot analysis versus the microchip separation for the detection of a particular
deletion. The subsequent work went on to establish normal ranges for each exon product using 40
control samples, which were then used to correctly evaluate patient samples for the presence of
mutations.42
The Bioanalyzer was also used in two studies by Sohni et al.43,44 that looked at both
variable number of tandem repeat and deletion mutations in genes of clinical interest. The
first study investigated polymorphisms in the cytokine genes IL-1RN and CCR5, both of
which are important in immune system functioning.43 The second study investigated nucleotide
repeats in the promoter region of the inducible NO2 synthase gene and the endothelial NO2
gene, both associated with complications in diabetes, and a deletion/insertion mutation in the
angiotensin-converting enzyme gene associated with NO2 activity.44 In addition to the Agi-
lent instrument, Hitachi also developed a commercial microchip electrophoresis instrument (SV
1210) both of which were utilized by Zhang et al.45 to evaluate gene fragments generated
from 3 µL of blood in less than 20 min using a DNA extraction disk and a capillary PCR
instrument.
193A/A Homozygote sample 193A-> T Heterozygote sample
(a) A-allele specific reaction (d) A-allele specific reaction

136 400
211 136
268 211 400
268
(b) T-allele specific reaciton (e) T-allele specific reaction

136 201

136
281 281
400 400
(c) A- and T-allele reacitons (f) A- and T-allele specific reactions
136
136
211
211
Red channel Red channel
Blue channel Blue channel

201
211
136
268 281
400 136 211 268 281 400
60 70 80 90 70 80 90
60
Time (s) Time (s)
FIGURE 36.2 Detection of a heterozygous mutation (right panel) in the HFE gene using allele-specific PCR
with two-color (red/blue) detection. (a and d) Separation of the A-allele specific amplicon (211 bp) from the
136 bp and 400 bp internal standards (red channel); (b and e) separation of T-allele specific amplicon (201
bp) from the same samples (blue channel); (c and f) separation of both amplicons showing only the A-allele
fragment in the homozygous sample but both fragments in the heterozygous mutant sample. Numbers above
the peaks indicate fragment length (bp). (Reprinted from Medintz, I., et al., Electrophoresis, 21, 2352, 2000.
With permission.)
36.4.1.2 Other Nucleic Acid Analyses

Mutation detection and genotyping represent a significant portion of clinical DNA analyses,
but other DNA separations of clinical interest have also been transferred to microchips. In
lymphoproliferative disorders, a single clonal variation is often overexpressed in T- or B-
cells resulting in a loss of heterogeneity in the immune system. Munro et al.46 translated
slab gel separations of PCR-amplified fragments, traditionally used to detect these disorders,
to a microchip, effectively detecting clonal populations in the variable region of the T-cell
receptor-γ gene and the immunoglobulin heavy chain gene. Again, no loss of diagnostic
information was seen in the transition to the microchip separation system, in fact, a patient
diagnosed as not having a lymphoma using the gel-based method was re-evaluated based on
the microchip results and found to be in an early stage of the disease. Quantification of
mRNA species has also been proposed as a clinical method, and reverse transcriptase-PCR
(RT-PCR) products from renal biopsies have also been separated on microchips as a diagnostic
method.47
(a) 1 2 3 (b)
- Exon 47
- Exon 52
Microchip separation
Exon 47
Exon 6
Control
- Exon 48 200 Patient 22
Fluorescence
- Exon 50
- Exon 51
100
0
- Exon 49 50 60 70 80 90 100
Seconds
- Exon 48
FIGURE 36.3 (a) Southern blot results obtained using DNA digested with HindIII and probed with cDMD 8.
Lane 1: normal male control; lane 2: normal female control; lane 3: female patient with a deletion in exon 47.
(b) Microchip electropherograms of a multiplex amplification for a control and a female patient with a deletion
in exon 47.
The presence of exogenous DNA or RNA in a patient is also of clinical interest, as this can
be used to indicate the presence of a bacterial or viral pathogen. Höfgartner et al.48 used a single-
channel glass microchip for the detection of herpes simplex virus (HSV) in cerebrospinal fluid
(CSF) through separation of PCR-amplified products. The microchip-based electrophoresis method
allowed for rapid diagnosis of herpes simplex encephalitis (HSE), in less than 100 s, compared
to 18 h for identical results obtained by hybridization. Chen et al.49 were able to detect viral RNA
from the hepatitis C virus through RT-PCR amplification with product identification performed using
an electrophoretic separation on a plastic microdevice. This group went on to further quantify the
amount of viral DNA present using a competitive PCR procedure, with product analysis again on the
plastic microchips.50 The Agilent Bioanalyzer also provides an efficient method for the separation
of PCR products generated from exogenous DNA in CSF to show the presence of viral infections.
Separations showing the presence of infections from a number of viral pathogens are shown in
Figure 36.4. The Agilent instrument was also used to evaluate repetitive sequence PCR products
from Mycobacterium species by Goldberg and coworkers.51 This kind of genotypic analysis can be
used to identify particular strains, track tuberculosis transmission in outbreaks, and identify cross
contamination in the laboratory.
36.4.2 PROTEIN ANALYSIS

Electrophoretic analysis of proteins was one of the first methods transitioned to capillaries in clinical
laboratories—this is highlighted in Chapter 2 by Hempe. Unfortunately, protein analysis has not
made the same transition to microchips. This is due to the difficulties in implementing a ultraviolet
(UV) absorbance method on microchips, the method normally utilized for proteins detection on
CE. A number of fluorescent methods have been employed both for standard separations as well as
for immunoassay analyses that utilize separations. Two-dimensional separations have also been per-
formed on microchips,52 as reviewed in Chapter 33 by Lee, and these might have clinical significance
at some point in the future.
CMV samples
HSV samples 250
250
200 200
Fluorescence
Fluorescence
150 150
100 100
50 50
0 0
30 40 50 60 70 80 90 100 110 30 40 50 60 70 80 90 100 110
Time (s) Time (s)
Blank Control Patient 2 Patient 6 Blank Control Patient 9 Patient 20
FIGURE 36.4 Electropherograms of PCR-amplified cerebral spinal fluid samples from patients with suspected
infectious agents. (a) Samples from patients possibly infected with HSV; clinical diagnosis by traditional
means: patient 2—infected HSV type 1, patient 6—negative. (b) Samples from patients possibly infected with
cytomegalovirus; clinical diagnosis by traditional means: patient 9—infected, patient 20—negative.
36.4.2.1 Protein Separations

Human serum protein quantification is an important method for determining the relative abundances
of albumin, IgG, and other important serum proteins. Colyer et al.53 investigated serum proteins
through microchip separations of fluorescently labeled proteins, showing that quantification was
possible even after labeling. Real samples, however, could not be evaluated because of poor sen-
sitivity. Callewaert et al.54 also utilized a fluorescent labeling scheme to tag N-glycan components
in serum, but was able to separate these real clinical samples on a microchip; a typical separation
is shown in Figure 36.5, in which the difference between serum from a patient with cirrhosis and
one with chronic hepatitis is illustrated. Separation of lipoproteins in serum on a microchip was
reported by Weiller et al.55 who utilized NBD as a fluorescent label to allow detection of both high-
and low-density lipoproteins, important risk factors for cardiac disease. This group further refined
the work to allow analysis of the low-density lipoprotein (LDL) components to show the presence
of different forms of LDL in serum samples.56
A number of other covalent and dynamic fluorescent labeling schemes have now been pro-
posed for the detection of proteins on microchips, all of which could be applied for clinical protein
analysis.57−60 Bio-Rad has adopted an on-column dynamic labeling scheme in their commercial
microchip electrophoresis instrument, which Chan and Herold61 utilized to evaluate microalbumin-
uria, a marker for cardiovascular disease and nephropathy in diabetic patients. Through the use of
an SDS-based separation, these experiments can provide quantitative and qualitative information
about the proteins. However, the use of covalent and dynamic labeling of native proteins must be
evaluated for each system studied as fluorescent tags are not incorporated homogeneously across
proteins, and are known to be influenced by the hydrophobicity and secondary structure of the pro-
teins. This heterogeneity affects both the electrophoretic mobility and fluorescence response of each
protein differently, and thus must be accounted for in the results obtained from any clinical protein
separation. Giordano et al.60 developed a partially denaturing technique for fluorescent labeling, to
minimize preferential binding, in which SDS was added to the sample, but no heat denaturation
step was employed. Figure 36.6 shows the microchip protein profiling of human sera, using this
method for the detection of gammopathies. The separations show clear spikes in the protein profiles
of severely affected patient sera, but detection of more subtle changes was not evaluated.
UV detection on microchips has not been completely bypassed, but it does require quartz
microdevices, as the normal glass and plastic microchips absorb too much UV radiation in the
wavelengths of interest. Zhuang et al.62 have developed quartz microchips for protein detection, and
3 Chronic hepatitis
0.6
0.5
6
0.4 8 9
7
1
0.3
RFU
0.4 Cirrhosis
0.3
0.2
8 9 10 11 12 13
Time (min)
FIGURE 36.5 Profiling of two serum samples: one from a noncirrhotic chronic hepatitis patient and one from
a patient with cirrhosis. Log (peak 7/peak 8) is diagnostic for liver cirrhosis, and these two peaks are well
resolved in the microfluidics profiles. (Reprinted from Callewaert, N., et al., Electrophoresis, 25, 3128, 2004.
With permission.)
have utilized them for electrophoretic analysis of urinary proteins of clinical interest.63 This group
has also now applied quartz microchips to the analysis of serum lipoproteins64 without the need for
fluorescent tagging as discussed above.55 UV detection is also covered in a review by Wang et al.65
who detail the use of electrophoretic microchips for the clinical analysis of hemoglobin, with various
detection methods.
36.4.2.2 Microchip Immunoassays

A number of microchip immunoassays have been developed that utilize an on-chip reaction followed
by a separation component to separate free antigen from that bound to the antibody—if a labeled
antigen is used—or to separate free antibody from the antigen/antibody complex—if labeled anti-
bodies are used. Schmalzing et al.66 utilized the latter method to demonstrate an immunoassay for
T4 in serum for clinical evaluation of thyroid function on fused-silica microchips. The immunoas-
say used a competitive format with labeled antigens competing against antigens in the serum for
antibody-binding sites; the amount of free labeled antigen was used to quantify the amount of T4
in the samples. The same group also reported a microchip electrophoretic immunoassay for cortisol
in serum using a similar method,67 which could quantify cortisol in serum within the clinical range
without extraction from the sample or other preparation steps. Qiu and Harrison68 were able to
incorporate a calibration step into their immunoassay microchip that will be important for eventual
implementation in clinical analyses. Using a microchip with multiple inlets, they could control the
amount of antibody mixed with labeled antigen, in this case bovine serum albumin (BSA). While
this antigen is not clinically relevant, the ability to calibrate the microchip response, as shown in
Figure 36.7, will be important for protein quantification.
In addition to fluorescent-based assays, electrochemical detection has also been utilized for
microchip immunoassays. Wang et al.69 initially utilized an alkaline phosphatase-labeled antibody,
(a)
2.4
Albumin
2.2
2
RFU
1.8
β γ
1.6 α
1.4
1.2
1
200 250 300 350 400
(b) 2.4
2.3
2.1
1.9
β Spike
1.7
RFU
1.8
1.4
1.1
0.8
0.7
0.4
200 250 300 350 400
(c) 3.5
3
γ Spike
2.5
1.5
0.5
200 250 300 350 400
Seconds
FIGURE 36.6 Comparison of CZE analysis of partially denatured human sera using run buffer containing
0.04% NanoOrange dye. Sera were diluted 1:500 in 0.5% SDS, with 1% dye included in the sample buffer.
Separation conditions: 100 mM borate, 3 mM diaminobutane, pH 8.5. (Reprinted from Giordano, B. C., et al.,
Anal. Chem., 76, 4705, 2004. With permission.)
which was mixed with the antigen (mouse IgG), then separated on-chip to resolve the free
antibody from the complex. Using a post column reaction of the alkaline phosphatase to produce
4-aminophenol, they were able to monitor the products using electrochemical detection. This same
group also used ferrocene tagged antibodies, simply measuring the reduction of the ferrocene in
the free antibody and the complex using amperometric detection.70 Measurement of both IgG and
T3 antigens were shown in this format. More recently, Herr et al.71 showed fluorescent microchip
immunoassays for both tetanus antibody and tetanus toxin, utilizing fluorescently labeled tetanus
toxin C-fragment and a fluorescent marker as an internal standard for quantification. The microchip
separation provided a direct immunoassay for the presence of tetanus antibody, and a competitive
Ab Buffer
Buffer
Reactor
BSA* Waste
Sep. channel
500 AFU Waste Detector
Impurities
Ab (µg/ml)
46.1
27.7
19.3
Complex
BSA* 7.0
0 20 40 60 80
Time (s)
FIGURE 36.7 Traces from an on-chip calibration series. Intensity changes as a function of the amount of anti-
BSA delivered to the reaction, the complex peak increases in intensity while the labeled BSA peak decreases. At
the highest antibody concentration impurities can be seen in the BSA peak region. (Reprinted from Qiu, C. X.,
and Harrison, D. J., Electrophoresis, 22, 3949, 2001. With permission.)
assay for the presence of tetanus toxin could be used to generate a dose-response curve as shown in
Figure 36.8; the microchip design, however, did not include an integrated incubation step.
36.4.3 ADDITIONAL CLINICALLY RELEVANT ANALYSES

In addition to DNA and proteins, a number of other important clinical markers have been evalu-
ated using electrophoretic separations on microchips. Fanguy and Henry72 quantified uric acid in
urine using a microchip electrophoretic separation with electrochemical detection. This work was
extended to include three markers of renal function normally measured in clinical assays: uric acid,
creatinine, and creatine.73 Detection of glucose has also been performed directly from human plasma
on a microchip, with no need for sample pretreatment. Du et al.74 utilized a poly(dimethylsiloxane)
(PDMS) microchip containing a copper electrode for electrocatalytic oxidation of the glucose as the
detection method. Additional investigations involving the analysis of carbohydrates by microchip
electrophoresis, a few of which are of interest for clinical analyses, are covered in a review by
Suzuki and Honda.75 This review covers not only electrochemical detection methods but also
reports on the use of fluorescent and UV absorbing tags; the authors suggest, however, that the
need for a derivatization step eliminates the rapid separation advantage provided by the microchip
analysis.
A number of small molecules of interest in clinical analyses have also been analyzed using
microchip-based separations. Zhao et al.76 utilized a microchip separation for detection on the
antibiotic lincomycin directly from urine samples. This method, also employing electrochemical
detection, allowed lincomycin analysis in under 40 s. Crevillen et al.77 utilized a similar method for
Normalized detector response

4
3
Normalized peak height
5 2 (a)
1
(b)
4
(c)
3
0 40 80 120
2
Separation time (s)
0
0 0.025 0.05 0.075
[TTC]2 µM
FIGURE 36.8 Dose–response curve for the tetanus toxin C-fragment using a microchip competitive
immunoassay. Symbols show the normalized response (peak height of complex/peak height of free dye) with
error bars indicating the standard error in the measurements (n = 3–5). A correlation coefficient of .95 was
obtained for the linear fit through the points indicated. The inset shows electropherograms with TTC concentra-
tions of (a) 2.0, (b) 7.8, and (c) 15.6 nM, with a labeled TTC* concentrations of 13.0 nM and 6.0 nM anti-TTC.
Peaks 1 and 2 are due to free dye with peak 2 used for normalization; peak 3 is free TTC*, and peak 4 is the
complex. Separation conditions: 300 V/cm, total length 6.1 cm, and length to detector 6 mm. (Reprinted from
Herr, A. E., et al., Anal. Chem., 77, 585, 2005. With permission.)
separation and detection of hydroquinone and hydroquinone glycoside, important in clinical analyses
because of their presence in pharmaceuticals and natural remedies. They were able to separate and
detect these components using different injection methods, and also applied the microchip method
to the analysis of urine samples. Figure 36.9 shows application of this method to a variety of sample
types. Electrochemical detection was also utilized on microchips for the separation of six organic
acids (fumaric, citric, succinic, pyruvic, acetic, and lactic acids) and three cations (K+ , Na+ , and
Li+ ).78 This microchip used a four-electrode system with no direct contact between the electrodes
and the separation channel, but was not tested with real samples.
Munro et al.79 showed separation and detection of amino acids on microchips using an indirect
fluorescence detection method. Figure 36.10 shows application of this method to urine samples with
no pretreatment other than dilution in the appropriate separation buffer. Abnormal amounts of amino
acids can easily be detected in the two patient samples compared to the healthy control sample.
An absorbance detection based approach was utilized for the clinical analysis of calcium ion in
serum, which is important in the regulation of a number of physiological processes.80 Beads with an
immobilized calcium reactive dye were placed into the detection region, and the samples mobilized
past the beads using electrophoretic flow. While a true separation was not intended, the interference
of magnesium ions was significantly less than seen in pressure flow devices,19 since Mg2+ moves
through the channel faster than Ca2+ allowing less interaction time with the beads.

There are a large number of biological sample processing steps that have now been translated to
microchip methods and been shown to provide the same results, while utilizing smaller amounts
of sample and less reagents, thus reducing both cost and waste generated. At the same time, the
miniaturization of these processes has often produced a significant reduction in the time required,
(a) 2 (b)
10 nA
1 20 nA
1 2
b b
1
2
a
0 50 100 150 200 0 50 100 150 200

Time (s) Time (s)
(c) 1 (d)
1
2
20 nA
20 nA 2
b b
1
2
a
0 50 100 150 200 0 50 100 150 200

Time (s) Time (s)
FIGURE 36.9 Electropherograms of (a) the nutraceutical HMP, (b) urine, (c) a cosmetic formulation, and (d)
a wine sample. Peak labels correspond to (1) hydroquinone and (2) arbutin. The (a) traces are unspiked sample
(b and d) or spiked with 150 ppm of (a) hydroquinone and (c) arbutin. The (b) traces are spiked with 150 ppm
of hydroquinone and 125 ppm of arbutin. (Reprinted from Crevillen, A. G., et al., Anal. Chim. Acta, 562, 137,
(a) (b) (c)

Microchip
G
A.Q
A. M. Q
S
G
S
S
A.Q
G
H
N.T
H.V
T
I.L
T
I.L
3 RFU 3 RFU 5 RFU

50 70 90 110 50 70 90 110 50 70 90 110
Time (s)
FIGURE 36.10 Microchip electrophoretic separations of (a) normal and (b and c) abnormal urine samples.
Urine samples were diluted 1:10 in water with 0.5 mM fluorescein, 1 mM sodium carbonate and 0.2 mM CTAOH.
Separation conditions: 1.0 mM sodium carbonate, 0.5 mM fluorescein, and 0.2 mM CTAOH, pH 10.3; leff 5.5
cm, 15-s injection at 417 V/cm (reversed polarity), 183 V/cm separation voltage. Abbreviations: G, glycine; S,
serine; N, asparagine; T, threonine; A, alanine; M, methionine; Q, glutamine; H, histidine; V, valine; L, leucine;
I, isoleucine. (Reprinted from Munro, N. J., et al., Anal. Chem., 72, 2765, 2000. With permission.)
by reducing diffusional lengths, thermocycling times, and column volumes. This chapter will not
address all of these processes, but again will be limited to those processes directly related to clinical
analysis and those that would be utilized in conjunction with a subsequent electrophoretic separation
for analysis. Chapter 43 by Bienvenue and Landers, Chapter 50 by van Midwoud and Verpoorte in
this volume address additional methods and provide more in-depth coverage of the microchip sample
preparation literature. In addition, early work on sample preparation methods utilized for clinical
molecular diagnostics that have been translated to microfluidic formats is covered in a review by
Huang et al.81 These methods include cell sorting/selection, which can be used for clinical diagnostics
but are not covered here, as well as nucleic acid purifications, and nucleic acid amplifications methods.
36.4.4.1 Enzymatic Reactions

Enzymatic amplification of specific fragments of a nucleic acid template, particularly through the use
of the PCR, is one of the most widely translated methods to microchips associated with a subsequent
electrophoretic analysis. A number of microdevices and thermocycling methods have been described
and are summarized in a recent review by Zhang et al.82 covering the basic designs and practical
applications of microchip PCR. Not all of the microchip PCR developments showed direct clinical
utility, and thus are not included here, but most of the methods could theoretically be utilized for
clinical analysis. In addition to the traditional PCR-based DNA amplification, amplifications from
RNA using RT to first generate a cDNA are also of clinical interest, as well as other enzymatic
methods, which are not as widely utilized.
Most clinically relevant microchip PCR amplifications have been utilized for the detection of
infectious agents. Mikhailovich et al.83 showed allele-specific amplifications of mycobacterium
tuberculosis to evaluate rifampicin resistance as a method for monitoring the use of this drug for
treatment. They compared this method to a hybridization method, providing similar results but
greatly decreasing the analysis time. Yang et al.84 utilized a plastic microchip to test the sensitivity
and specificity of detection of a K12-specific fragment from Escherichia coli. They were able to
detect as low as 10 E. coli cells with this method in the presence of 2% blood, and could do multiplex
PCR to show specificity. For the detection of hepatitis B, a common infectious disease and blood
borne pathogen, Cho et al.85 conducted a large-scale study of a microchip PCR method for the virus
in real clinical samples. The authors utilized a real-time detection method, employing a miniature
fluorescence detection system using light emitting diodes and photodiode detectors, to monitor
amplification of sequences specific for the hepatitis B virus, thus a separation was not needed.
The study, however, evaluated specificity, sensitivity, cross reactivity, reproducibility, and limits of
detection in a large-scale clinical evaluation of the method (n = 563) to show the applicability of
microchip methods for routine clinical analysis.
Real-time detection was also used for an isothermal amplification procedure termed nucleic acid
sequence-based amplification, for the detection of artificial human papilloma virus sequences in SiHa
cells.86 This method was not applied to real samples, but demonstrates the clinical applicability of
additional microchip enzymatic amplification methods. Not all clinical microchip PCR amplifications
looked for infectious agents, however. Cheng et al.87 showed a microchip PCR method for multiplex
amplifications of dystrophin gene fragments for the detection of mutations associated with muscular
dystrophy.
Other types of enzymatic reactions have also been implemented on microchips. Using PCR-
amplified DNA as the template, Lou et al.88 utilized ligase chain reaction to identify a mutation
in the NOD2/CARD15 gene, a marker for inflammatory bowel disease, using DNA from blood of
healthy volunteers. The reaction used the ligase enzyme to join short oligonucleotides sequences
when the corresponding complementary wild-type or mutant DNA template was present. Digestion
of DNA on a microchip using restriction enzymes has also been reported, but at this time, this
has not yet been utilized for analyses relevant to clinical work.89 Likewise, digestions of proteins
have been shown on microchips, but have not been applied for clinical analyses.90,91 A number of
these enzymatic reactions and others have been transferred to microchips, but are covered within
Section 36.4.5 below since they have been directly coupled to the subsequent electrophoretic
separation on the same microchip.
36.4.4.2 Purification and/or Concentration

A number of diverse methods fall under this category with respect to clinical analysis. One important
process in clinical analyses is the isolation of plasma from whole blood, which has been performed on
a microchip using a transverse flow microfilter design.92 Utilizing capillary action for the separation,
both the plasma and the cell components of the blood were available on-chip for additional processing.
Processing of blood to separate the leukocytes from other components has also been shown on a
microchip for clinical applications requiring removal of erythrocytes.93 This microchip was designed
to mimic intrinsic blood flow processes, such as leukocyte margination, to separate the two cell types
and allow collection of the white cells. Another interesting technique applied to serum and urine
samples is laminar flow diffusion, which takes advantage of the laminar flow and short diffusion
lengths provided by microchips.94 By flowing the sample in parallel with a water stream, low
molecular weight components are extracted into the water phase, which can then be collected in a
separate stream from the original sample. While used only for spectroscopic detection in the original
work, this could be implemented for the removal of small molecular weight components for analysis,
such as urea and creatine from urine. Additional details on this process can be found in Chapter 35
on liquid/liquid extraction by Tokeshi and Kitamori.
One of the most widely investigated purification methods on microchips is DNA extraction,
which has been shown using a number of different phases and microchip designs, but only those
methods that have been shown to deal with real samples such as whole blood are of clinical interest.
Breadmore et al.95 capture DNA directly from blood utilizing silica beads immobilized in place with
a silica sol–gel matrix, showing that it was possible to perform PCR amplification on the eluted
genomic DNA. Chung et al.96 used beads with organic surface groups for capture of DNA on a
poly(methyl methacrylate) (PMMA) microchip. The beads were immobilized to the surface of the
PMMA through the organic surface groups, and the DNA containing solution passed back and forth
over the immobilized beads. DNA from E. coli cells could be extracted and amplified from serum
and whole blood, with the immobilized beads providing significantly better extraction efficiency
than free beads as the number of E. coli in the blood sample decreased.
Rather than packing beads into a microchannel, Wu et al.97 utilized the liquid precursor of the
silica sol-gel to form a monolithic matrix in a microchannel for DNA extraction. This solid phase
provided extraction efficiency with good reproducibility using standards, but extraction efficiency
fell off quickly with whole blood when the devices were used repeatedly. This is not necessarily a
problem, in that these devices for clinical applications would be designed for only a single use to
prevent cross contamination of samples. Figure 36.11 shows the protein and DNA profiles generated
on the sol-gel microchip during purification of DNA directly from lysed blood. While the sol-gel
matrix was easier to generate in the channel, the liquid precursor filled all of the channels. Wen
et al.98 solved this problem through the use of a photopolymerizable matrix, which they coated with
a sol-gel precursor to increase the DNA-binding capacity. For whole blood application, they found
that the proteins competed for binding sites, and significantly decreased the capacity of the monolith
for DNA. A dual phase microchip was developed using an initial C18 phase to capture the proteins
from the lysed blood as the DNA passed through for capture on the monolith. Volumes as large as
10 µL of blood could be processed on this device with good extraction efficiencies.
In addition to proteins competing for binding sites in the DNA extraction process, removal of
proteins from the solid phase is normally performed using an alcohol-based wash step to prevent
protein coelution with the DNA and possible interference with the subsequent processing steps. As
reported by Legendre et al.99 this wash step causes its own problems in that isopropanol inhibits
PCR, which is often the next process on an integrated device. Wen et al.98 were able to eliminate this
Load Wash Elution

3
65
Protein 2.6
55 Genomic DNA 2.2
45
1.8
Protein (µg)
DNA (ng)
35 1.4
25 1
15 0.6
5 0.2
–5 –0.2
0 20 40 60 80 100 120 140 160 180 200
Volume (µL)
FIGURE 36.11 Protein and DNA extraction profiles from a lysed human blood sample loaded onto a sol-gel
filled microchip. Protein (closed squares) is removed in the wash step and DNA (open circles) is released during
the elution step.
step using the dual phase method, but other approaches have also been investigated. Witek et al.100
utilized a photoactivated polycarbonate (PPC) microfluidic device for DNA extraction from E. coli,
which incorporated an ethanol wash step, but the device could be dried before elution of the DNA
to eliminate inhibition from the wash buffer. Larger elution volumes were required, however, and
this work did not address DNA purification from human whole blood. An amine-coated microchip
that had lower protein binding was used in work by Nakagawa et al.101 to extract DNA from whole
blood. No wash step was required, avoiding the use of PCR-inhibiting reagents, but the amine phase
showed only moderate efficiency. A competing approach was developed by Cao et al.102 using the
protonated amine groups on chitosan at a low pH to capture the negatively charged DNA. The DNA
was easily released at a high pH at which the amine groups became neutral. Very little protein was
found to bind to the chitosan phase, thus a wash step was also not required in this method. Using the
chitosan technology, a microchip capture device could be constructed simply by immobilizing the
chitosan directly on the surface of a bifurcated channel structure within a microchip.
While DNA represents a major focus of purification on microchips, capture of RNA has also
been explored. Kokoris et al.103 utilized a silica membrane immobilized in a microchip for RNA
purification. The “lab card” device provided RNA suitable for reverse transcriptase amplification
with a yield and quality better than that obtained using traditional RNA extraction methods. A number
of protein preconcentration or extraction microchip devices have also been developed, using either a
general hydrophobic capture phase to capture all proteins, or an antibody or ligand-based phase for
specific capture of a protein of interest. Most of these are designed for use with MS analysis of the
collected products, however, and are covered in Chapter 53 by Lazar. One exception is the use of a
phosphocholine ligand capture phase for selective capture of C-reactive protein (CRP), of interest as
a cardiac and inflammation marker.104 This work required subsequent electrophoretic separation to
quantify the CRP because of capture of additional phosphocholine-binding proteins in the extraction;
quantitative results generated from the microchip extraction process agreed well with current CRP
analysis method results.
36.4.5 CHIP-BASED INTEGRATION OF SAMPLE PREPARATION STEPS

The ultimate motivation for transfer of electrophoretic separations from a capillary to a microchip
was the ability to perform these additional processing steps in the microchip format, which was not
as easily achieved in a capillary system. This provides the microchip the advantage of sequential
integration of a number of processing steps directly with the electrophoretic analysis, and many
of the integrated devices that resulted are of interest in clinical analyses.105 As with microchip
electrophoresis itself, much of this work has focused on DNA for genetic analysis106 (see Chapter 43
by Bienvenue and Landers), but a number of integrated protein microchips have also been reported,
along with multiprocess assays for small molecules.
36.4.5.1 DNA Amplification

A significant number of groups have reported on integration of the PCR process directly with the sub-
sequent electrophoretic separation on the same microchip.87,107−115 As with the microchip PCR for
clinical applications, most of the integrated microchip work focused on detection of infectious agents,
starting in 1996 with Woolley et al.107 who showed amplification and electrophoresis of a fragment of
salmonella DNA. Figure 36.12 shows the microchip separation of the amplified salmonella fragment
along with a co-injection of the amplified product with a size standard for peak identification. Waters
et al.108,109 incorporated cell lysis, and amplified and separated fragments from multiple regions of
the E. coli genome in their integrated device. Bacterial detection of both E. coli 0157 and salmonella
was performed by Koh et al.113 in a plastic integrated device fabricated in cyclic olefin copolymer
(COC). Limits of detection as low as six copies were possible. Lagally et al.114 also evaluated an
integrated microchip for pathogen detection, amplifying genes from both E. coli and Staphylococcus
aureus. Limits of detection in this system were as low as two to three bacterial cells. Easley et al.115
developed an integrated microchip that utilized valves for pressure-based injections from the PCR
chamber into the separation channel for the detection of salmonella DNA. This allowed dilution
Primer-dimer
Salmonclla PCR product
1200
800
Fluorescence (arbitrary units)
400
0
50 60 70
Salmonclla PCR product
Primer-dimer
600
400
310
281
271
234
194
200
118
0
40 50 60 70 80
Time (s)
FIGURE 36.12 Integrated PCR-electrophoresis microdevice assay for the detection of salmonella DNA.
(a) Microchip separation of the PCR product was performed immediately following a 39-min amplification in
the integrated microdevice. Primer-dimer (light gray peak) appears at 51 s and the PCR product (dark gray peak)
appears at 61 s. (b) Sizing of the Salmonella product (1:100 dilution) in a separate microchip using a X174
HaeIII digest DNA standard (1 ng/µL). Total analysis time for the salmonella sample using the integrated
microdevice was less than 45 min. (Reprinted from Woolley, A. T., et al., Anal. Chem., 68, 4081, 1996. With
permission.)
of the PCR product with water to provide a stacking effect as it was co-injected with a DNA size
standard for accurate sizing.
For the detection of a severe acute respiratory syndrome coronavirus, Zhou et al.116 used an
integrated microchip to perform RT-PCR along with electrophoretic sizing. Nasopharyngeal swab
samples were utilized in the analysis with the microchip method showing infection in 17 of 18
patient samples; this was better than the conventional RT-PCR method that identified only 12 infected
patients. Pal et al.117 incorporated an additional reaction step in their integrated microchip for the
identification of hemagglutinin A subtypes of influenza virus. Identification was based on restriction
fragments, thus an on-chip enzymatic digestion was performed after the PCR. Separation of the
digested products allowed successful discrimination of two influenza strains. In more recent work,
Kaigala et al.118 used a PCR electrophoresis microchip to assess the risk of nephropathy in renal
transplant patients. The analysis could distinguish between low, medium, and high BK viral loads,
indicating patients at risk for complications; detection of as low as two viral copies was possible.
Easley et al.119 went further, incorporating a DNA extraction step with the PCR and separation steps
to show a fully-integrated analysis following off-chip cell lysis. They reported on analysis of mouse
blood for the detection of Bacillus anthracis infections showing detection in asymptomatic mice
from less than 1 µL of blood samples. A similar sized sample of nasal aspirate was used to show the
presence of Bordetella pertussis in a patient confirmed to have whooping cough. Figure 36.13a shows
a picture of the integrated microchip utilized in this work, Figure 36.13b shows the IR-mediated
thermocycling profile on the device, and Figure 36.13c shows consecutive injections and separations
of the amplified product for the B. anthracis analysis.
For detection of mutations, Cheng et al.87 used an integrated PCR-microchip electrophoresis
device for a multiplex amplification to identify deletions in the dystrophin gene, with successful
amplification and resolution of all of the products. Taylor et al.120 detected mutations in mitochondrial
DNA, which are indicated in a wide range of human disease states, utilizing a different set of on-chip
processes. DNA was digested using a restriction enzyme, then denatured and reannealed to perform
a HDA in the separation step for the identification of mutations.
(a) (b) SPE-PCR-ME anthrax 1 thermocycling profile
90
80
Temp (c)
70
60
50
0 2 4 6 8 10 12
Time (min)
(c) SPE-PCR-ME anthrax 1

7
Run 01_1
6
Run 01_2
5
RFU
4
3
2
60 120 180 240
Time (s)
FIGURE 36.13 (a) Integrated microchip used for the detection of B. anthracis infection in murine blood.
The right channels are the solid phase extraction region; the dark filled chambers are the PCR area; the left
channel is the separation channel, with the cross-t injection region covered by the pneumatic actuation lines
used for an on-chip pumped injection. (b) Thermocycling profile for the microchip using IR-mediated PCR.
(c) Electrophoretic analysis of the amplified product showing sequential pressure injections and separations on
the integrated device to confirm the presence of the amplified fragment.
36.4.5.2 Onboard Sample Preparation for Other Analytes

Integrated microdevices have also been developed for protein analysis, most often using MS detec-
tion, as in the discovery of new biomarkers useful for clinical analysis.121 Gottschlich et al.122 showed
early integration on a microchip, combining a protein digestion step directly with an electrophoretic
separation of the products, but often only a preconcentration step is integrated with the on-chip
separation. Membranes have been formed in a cross channel for concentrating proteins during the
injection step using both laser patterning123 and photopolymerization.124 While these methods might
be applicable to clinical analysis, no clinical sample was evaluated. Both concentration and selective
protein removal was shown on a microchip that integrated two electrophoretic channels with built-
in electrochemical detection.125 The first channel was used for sample preconcentration utilizing
an isotachophoresis method, followed by a normal CZE separation; a myoglobin sample could be
concentrated by greater than 60-fold in this system. Selective removal of proteins from a peptide
sample was also shown in conjunction with desalting of the sample for subsequent analysis of the
peptides.
In addition to proteins, small molecules have also been analyzed on integrated microchips.
Analyses of three markers of renal function, uric acid, creatinine, and creatine, along with
p-aminohippuric acid were investigated using integrated microchips by Wang and Chatrathi.126
This work utilized fluid control to mix in creatinase and creatininase enzymes with the samples for
on-chip enzymatic reactions before the electrophoresis. The unreacted acid species were separated
from the neutral hydrogen peroxide, generated in the two enzymatic reactions, with detection using
an amperometric method. To individually quantify the creatine and creatinine concentrations in a
sample, the creatininase was simply eliminated from the reaction. A similar method was employed
by Wang et al.127 for the measurement of uric acid integrated with measurements of ascorbic acid,
glucose, and acetaminophen on a single device. This work utilized the integrated microchip to mix
the sample with glucose oxidase for generating hydrogen peroxide from the glucose present. Again,
electrophoretic separation from the acidic species was possible with amperometric detection used for
quantification. Acetaminophen, another neutral species, which migrated with the hydrogen peroxide,
was detected by analyzing the sample with and without glucose oxidase to determine the difference
in detector response when the neutral components passed the detector.

Implementation of microchip methods for clinical analysis has thus far been almost exclusively
confined to the research laboratory. This is changing with the availability and ease of use of the
commercially available microchip electrophoresis instrumentation. These instruments provide both
quantification and size determination of DNA or protein samples, thus could be immediately imple-
mented into any clinical laboratory. Microdevices for sample preparation steps, at least for DNA
processing, are currently under commercial development, but have not yet been introduced. Sample
extractions and concentrations on microdevices are easily implemented in any laboratory, simply
by placing the appropriate solid phase matrix into a microchip rather than using a suspension or
large column, or centrifugal tube format. The only issue is fabrication or purchase of the microchips
that have the appropriate sized channel and a constriction or frit for retention of the solid phase
within the channel. Interfaces for the connection of syringe pumps to flow material through the
microchip column have been described in the literature,95 and microchip reservoirs/connectors are
also commercially available from IDEX® Corporation.
In-house development of a microchip PCR amplification system is a bit more challenging than the
DNA extraction system, but a noncontact IR heating design has been described in the literature.128
As an easier approach, simply putting the microfluidic device into a conventional thermocycler also
works, and Legendre et al.99 showed that it was possible to integrate microchip DNA extraction with
the PCR amplification using this technique. Integration of capture and separations, or enzymatic
or labeling reactions with the electrophoretic separation, are also possible in many laboratories,
again obtaining or fabricating the specifically designed microchips will be a significant part of the
processes. Development of fully-integrated devices, however, is at this point beyond the realm of
most laboratories that do not have experience with microfluidic design and utilization.
In addition to simply designing and fabricating the devices and instrumentation to use microchip
methods, a number of issues also have to be dealt with that are more specific for clinical appli-
cation of these devices. The use of microfluidic devices with biological samples provides some
important differences relative to the use of these devices for standards or “real samples” that have
been cleaned up or spiked to greatly increase the concentration of the compound of interest. The
proteins, lipids, and DNA mixture found in any biological sample have numerous charged species,
as well as both hydrophobic and hydrophilic compounds. Thus, surfaces in microchips formed
from almost any substrate will attract or interact with some component in the sample. In addition,
there are issues with the sheer mass of material, such as proteins that can compete nonspecifi-
cally with an assay, the particulate matter left by cell lysis that can clog the microstructures, and
the viscosity of biological samples that make it hard to pump material through the microdevice. The
development of microchips for clinical applications then must look beyond the initial testing of the
microchip assays using standards, and show that the assay, method, and microchip are rugged and
reliable enough for use with real samples at the required concentrations, which is not always the
case.2
Often, the primary issue when implementing on-chip clinical analyses is biofouling, where pro-
teins or lipids bind to the microchip surface interfering with the assay being performed on the device;
this is particularly an issue in microchips because of the large surface-to-volume ratio presented by
these devices. One of the biggest clinical applications in which this is important is PCR amplifica-
tion of DNA, in which binding of Taq polymerase and Mg2+ to the microchip surface can inhibit
the reaction.129,130 A review by Kricka and Wilding131 detailed the various methods that had been
utilized for passivating microchip surfaces to prevent PCR inhibition, and this group has shown
increased efficiency for PCR through coating selection.132 Preventing reaction components from
binding to the surface can also be accomplished by preventing the components in the aqueous solu-
tions from reaching the surfaces by surrounding the aqueous phase with an organic phase such as a
fluorocarbon oil, as reported by Roach et al.133
Of more interest is some recent work that looks at the modification of the microchip surfaces to
resist nonspecific adsorption in polymeric devices, since these represent the future of inexpensive,
disposable microchips that will be needed for clinical analyses. Bi et al.134 utilized an acrylate
copolymer to modify the surface of PMMA devices to exhibit poly(ethylene glycol) (PEG) on
the surface. PEG is known to prevent protein adsorption, and these microdevices were utilized
in electrophoretic separations of proteins, with good theoretical efficiencies. Application of these
devices directly to serum or plasma should be possible with no biofouling of the separation channel
surface. This same group was also able to generate a phospholipid coating on the surface of PMMA
microchip channels as a biomimetic surface to prevent protein binding.135 This coating was also
evaluated for the electrophoretic separation of proteins, showing stable electrophoretic mobility, with
little nonspecific adhesion of proteins or platelets when serum or plasma was utilized in the device.
For PDMS devices, Wu et al.136 grafted hydrophilic polymers onto the surface to create a surface
with a strongly suppressed electroosmotic flow (EOF), and significantly decreased adsorption of BSA
and lysozyme than the native PDMS surface. These microchips were also utilized for electrophoretic
separations of protein, both native and denatured, as well as peptides.
In addition to adsorption, issues to be addressed for the routine use of microchips for clinical anal-
ysis include the development of lower cost, compact instrumentation for detection on microchips.
Shrinivasan et al.137 detailed the development of a miniaturized LIF detector applicable for fluores-
cent detection of DNA on microchips, which could also be applied to proteins with the necessary
fluorescent tags. By replacing the argon-ion laser with a diode laser, the cost and power requirements
were significantly decreased, with only a small loss in sensitivity. Integration of fluid flow and mixing
components into the microchips must also be addressed. Kim et al.138 describe a microchip for blood
typing that incorporates flow splitting designs, micromixers, and microfilters directly fabricated into
an injection molded COC device.139 This group also developed a COC microchip for partial oxygen,
glucose, and lactate levels in blood that incorporates an on-chip power source for fluid manipula-
tion. Vestad et al.140 addressed the need to reduce the associated hardware for flow control through
microchips through the use of capillary pumping, allowing surface tension to pull fluid through their
devices. Flow path can be controlled through the use of elastomeric valves, similar to those utilized
in the integrated devices. Again, the goal is to eliminate the need for external power sources in the
instrumentation for fluid flow control, to simplify the instrumentation and make it cost effective and
compact enough for use in point-of-care applications.

There are a number of clinical areas in which microchip devices will be implemented over the next
decade. For DNA, the switch will actually be away from the devices described here to a large extent
as real time-PCR analysis becomes the predominant method for detecting and quantifying specific
DNA sequences. Real-time PCR takes advantage of the amplification method while eliminating the
separation step to reduce the complexity of the analysis and the microchip design. These devices
will be implemented both for mutation detection and infectious agent analysis. For proteins, the
separation will remain an integral part of the analysis, thus microchip separations will expand to
form a larger part of the clinical diagnostic arena. Most of these devices will utilize MS detection,
however, unless new breakthroughs in protein labeling or UV detection on microchips are seen.
Electrophoretic microchips for small molecule detection will play some part in clinical analysis, but
individualized instrumentation specific for a particular analyte or group of analytes may be the final
format.
For this to become reality, reductions in the costs of both the microdevices and the instrumentation
will be needed to move these methods into routine clinical laboratory use. On the instrumentation
side, replacing research designs with commercial grade instruments, with simple user interfaces
and integrated components for sample and reagent loading, alignment, and thermal processing must
be completed. Initial commercial microchip electrophoresis instruments show that this is possible,
but more demand from the clinical community may be needed to show the potential for additional
microchip instrumentation to commercial entities. The microchips themselves must also be cost
effective for the analyses being performed, with disposable devices required to prevent contamination
or false positive and false negative results. The cost of a PCR microchip, even a plastic and disposable
one, will never be as inexpensive as a PCR tube, thus additional advantages must be provided by the
microchip methods to justify their implementation. These can include reduced reagent consumption
in the microchip, or incorporation of additional processing steps such as the DNA extraction, but
some added benefit must be present in the microchips if they are to be accepted and used by the
clinical community.
In terms of integrating additional steps onto a microdevice, issues related to fluid flow control,
either through on-chip or off-chip valving or flow metering, still need to be more fully addressed.
Handling of the nanoliter volumes utilized in microchips presents additional challenges for instru-
mentation development, particularly with reagent addition and mixing with components already on a
microdevice. Further work is also needed in the area of on-chip storage of reagents, possibly already
lyophilized in the necessary compartments, or sealed off for storage but easily available on-chip
when needed. Development of integrated microchips will continue, however, and these will begin
to focus more on the actual end applications now that the concepts have been shown using standards
and simulants. More specific clinical analyses will be translated to microchips over the next few
years, and the complexity of the designs and instrumentation will determine how quickly they will
be transitioned into the clinical laboratory.
REFERENCES
1. Burtis, C. A., Converging technologies and their impact on the clinical laboratory, Clin. Chem., 42,
1735, 1996.
2. Colyer, C. L., Tang, T., Chiem, N., and Harrison, D. J. V., Clinical potential of microchip capillary
electrophoresis systems, Electrophoresis, 18, 1733, 1997.
3. Kricka, L. J., Miniaturization of analytical systems, Clin. Chem., 44, 2008, 1998.
4. Tudos, A. J., Besselink, G. A. J., and Schasfoort, R. B. M., Trends in miniaturized total analysis systems
for point-of-care testing in clinical chemistry, Lab Chip, 1, 83, 2001.
5. Schulte, T. H., Bardell, R. L., and Weigl, B. H., Microfluidic technologies in clinical diagnostics, Clin.
Chim. Acta, 321, 1, 2002.
6. Verpoorte, E., Microfluidic chips for clinical and forensic analysis, Electrophoresis, 23, 677, 2002.
7. Pugia, M. J., Blankenstein, G., Peters, R. P., Profitt, J. A., Kadel, K., Willms, T., Sommer, R.,
Kuo, H. H., and Schulman, L. S., Microfluidic tool box as technology platform for hand-held
diagnostics, Clin. Chem., 51, 1923, 2005.
8. Soper, S. A., Brown, K., Ellington, A., Frazier, B., Garcia-Manero, G., Gau, V., Gutman, S. I., et al.,
Point-of-care biosensor systems for cancer diagnostics/prognostics, Biosens. Bioelec., 21, 1932, 2006.
9. Situma, C., Hashimoto, M., and Soper, S. A., Merging microfluidics with microarray-based bioassays,
Biomol. Eng., 23, 213, 2006.
10. Aytur, T., Foley, J., Anwar, M., Boser, B., Harris, E., and Beatty, P. R., A novel magnetic bead bioassay
platform using a microchip-based sensor for infectious disease diagnosis, J. Immunol. Meth., 314, 21,
2006.
11. Soper, S. A., Hashimoto, M., Situma, C., Murphy, M. C., McCarley, R. L., Cheng, Y. W., and Barany,
F., Fabrication of DNA microarrays onto polymer substrates using UV modification protocols with
integration into microfluidic platforms for the sensing of low-abundant DNA point mutations, Methods,
37, 103, 2005.
12. Peytavi, R., Raymond, F. R., Gagne, D., Picard, F. J., Jia, G., Zoval, J., Madou, M., Boissinot, K.,
Boissinot, M., Bissonnette, L., Ouellette, M., and Bergeron, M. G., Microfluidic device for rapid
(<15 min) automated microarray hybridization, Clin. Chem., 51, 1836, 2005.
13. Rubina, A. Y., Pan’kov, S. V., Dementieva, E. I., Pen’kov, D. N., Butygin, A. V., Vasiliskov, V.
A., Chudinov, A. V., Mikheikin, A. L., Mikhailovich, V. M., and Mirzabekov, A. D., Hydrogel
drop microchips with immobilized DNA: Properties and methods for large-scale production, Anal.
Biochem., 325, 92, 2004.
14. Sohni, Y. R., Cerhan, J. R., and O’Kane, D., Microarray and microfluidic methodology for genotyping
cytokine gene polymorphisms, Human Immunol., 64, 990, 2003.
15. Honda, N., Lindberg, U., Andersson, P., Hoffman, S., and Takei, H., Simultaneous multiple immunoas-
says in a compact disc-shaped microfluidic device based on centrifugal force, Clin. Chem., 51, 1955,
2005.
16. Rossier, J. S., and Girault, H. H., Enzyme linked immunosorbent assay on a microchip with
electrochemical detection, Lab Chip, 1, 153, 2001.
17. Dupuy, A. M., Lehmann, S., and Cristol, J. P., Protein biochip systems for the clinical laboratory, Clin.
Chem. Lab. Med., 43, 1291, 2005.
18. Jia, H. X., Wu, Z. Y., and Fang, Z. L., Microfluidic chip-based immunoassay, Chin. J. Anal. Chem.,
33, 1489, 2005.
19. Malcik, N., Ferrance, J. P., Landers, J. P., and Caglar, P., The performance of a microchip-based fiber
optic detection technique for the determination of Ca2+ ions in urine, Sens. Act. B–Chem., 107, 24,
2005.
20. Srinivasan, V., Pamula, V. K., and Fair, R. B., An integrated digital microfluidic lab-on-a-chip for
clinical diagnostics on human physiological fluids, Lab Chip, 4, 310, 2004.
21. Morimoto, K., and Suzuki, H., Micro analysis system for pH and protease activities with an integrated
sample injection mechanism, Biosens. Bioelec., 22, 86, 2006.
22. Upadhyay, S., Ohgami, N., Kusakabe, H., and Suzuki, H., Electrochemical determination of gamma-
glutamyl transpeptidase activity and its application to a miniaturized analysis system Biosens. Bioelec.,
21, 1230, 2006.
23. Guijt, R. M., Baltussen, E., and van Dedem, G. W. K., Use of bioaffinity interactions in
electrokinetically controlled assays on microfabricated devices, Electrophoresis, 23, 823, 2002.
24. El-Ali, J., Sorger, P. K., and Jensen, K. F., Cells on chips, Nature, 442, 403, 2006.
25. Jin, L. J., Ferrance, J., and Landers, J. P., Miniaturized electrophoresis: An evolving role in laboratory
medicine, Biotechniques, 31, 1332, 2001.
26. McGlennen, R. C., Miniaturization technologies for molecular diagnostics, Clin. Chem., 47, 393,
2001.
27. Oin, J. H., Fung, Y. S., and Lin, B. C., DNA diagnosis by capillary electrophoresis and microfabricated
electrophoretic devices, Exp. Rev. Molec. Diag., 3, 387, 2003.
28. Mastrangelo, C. H., Burns, M. A., and Burke, D. T., Microfabricated devices for genetic diagnostics,
Proc. IEEE, 86, 1769, 1998.
29. Li, S. F. Y., and Kricka, L. J., Clinical analysis by microchip capillary electrophoresis, Clin. Chem.,
52, 37, 2006.
30. Tian, H. J., Jaquins-Gerstl, A., Munro, N., Trucco, M., Brody, L. C., and Landers, J. P., Single-strand
conformation polymorphism analysis by capillary and microchip electrophoresis: Afast, simple method
for detection of common mutations in BRCA1 and BRCA2, Genomics, 63, 25, 2000.
31. Kang, S. H., Jang, S., and Park, S. K., Single-strand conformation polymorphism analysis by microchip
electrophoresis for the rapid detection of point mutation in human obesity gene, Bull. Korean Chem.
Soc., 27, 1346, 2006.
32. Hestekin, C. N., Jakupciak, J. P., Chiesl, T. N., Kan, C. W., O’Connell, C. D., and Barron, A. E., An
optimized microchip electrophoresis system for mutation detection by tandem SSCP and heteroduplex
analysis for p53 gene exons 5–9, Electrophoresis, 27, 3823, 2006.
33. Tian, H. J., Brody, L. C., Fan, S. J., Huang, Z. L., and Landers, J. P., Capillary and microchip
electrophoresis for rapid detection of known mutations by combining allele-specific DNAamplification
with heteroduplex analysis, Clin. Chem., 47, 173, 2001.
34. Hestekin, C. N., and Barron, A. E., The potential of electrophoretic mobility shift assays for clinical
mutation detection, Electrophoresis, 27, 3805, 2006.
35. Woolley, A. T., Sensabaugh, G. F., and Mathies, R. A., High-speed DNA genotyping using
microfabricated capillary array electrophoresis chips, Anal. Chem., 69, 2181, 1997.
36. Shi, Y. N., Simpson, P. C., Scherer, J. R., Wexler, D., Skibola, C., Smith M. T., and Mathies, R.
A., Radial capillary array electrophoresis microplate and scanner for high-performance nucleic acid
analysis, Anal. Chem., 71, 5354, 1999.
37. Medintz, I., Wong, W. W., Sensabaugh, G., and Mathies, R. A., High speed single nucleotide poly-
morphism typing of a hereditary haemochromatosis mutation with capillary array electrophoresis
microplates, Electrophoresis, 21, 2352, 2000.
38. Medintz, I., Wong, W. W., Berti, L., Shiow, L., Tom, J., Scherer, J., Sensabaugh G., and Mathies, R. A.,
High performance multiplex SNP analysis of three hemochromatosis-related mutations with capillary
array electrophoresis microplates, Genome Res., 11, 413, 2001.
39. Sung, W. C., Lee, G. B., Tzeng, C. C., and Chen, S. H., Plastic microchip electrophoresis for
genetic screening: The analysis of polymerase chain reactions products of fragile X (CGG)n alleles,
40. Cantafora, A., Blotta, I., Bruzzese, N., Calandra, S., and Bertolini, S., Rapid sizing of microsatellite
alleles by gel electrophoresis on microfabricated channels: Application to the D19S394 tetranu-
cleotide repeat for cosegregation study of familial hypercholesterolemia, Electrophoresis, 22, 4012,
2001.
41. Ferrance, J., Snow, K., and Landers, J. P., Evaluation of microchip electrophoresis as a molecular
diagnostic method for Duchenne muscular dystrophy, Clin. Chem., 48, 380, 2002.
42. Ferrance, J. P., Wu, Q. R., Giordano, B., Hernandez, C., Kwok, Y., Snow, K., Thibodeau, S., and
Landers, J. P., Developments toward a complete micro-total analysis system for Duchenne muscular
dystrophy diagnosis, Anal. Chim. Acta, 500, 223, 2003.
43. Sohni, Y. R., Cerhan, J. R., and O’Kane, D., Microarray and microfluidic methodology for genotyping
cytokine gene polymorphisms, Human Immunol., 64, 990, 2003.
44. Sohni, Y. R., Burke, J. P., Dyck, P. J., and O’Kane, D. J., Microfluidic chip-based method for genotyping
microsatellites, VNTRs and insertion/deletion polymorphisms, Clin. Biochem., 36, 35, 2003.
45. Zhang, L. H., Dang, F. Q., Kaji, N., and Baba, Y., Fast extraction, amplification and analysis of genes
from human blood, J. Chrom. A, 1106, 175, 2006.
46. Munro, N. J., Snow, K., Kant, E. A., and Landers, J. P., Molecular diagnostics on microfabricated
electrophoretic devices: From slab gel- to capillary- to microchip-based assays for T- and B-cell
lymphoproliferative disorders, Clin. Chem., 45, 1906, 1999.
47. Cohen, C. D., Frach, K., Schlondorff, D., and Kretzler, M., Quantitative gene expression analysis in
renal biopsies: A novel protocol for a high-throughput multicenter application, Kidney Inter., 61, 133,
2002.
48. Höfgartner, W. T., Huhmer, A. F. R., Landers, J. P., and Kant, J. A., Rapid diagnosis of herpes simplex
encephalitis using microchip electrophoresis of PCR products, Clin. Chem., 45, 2120, 1999.
49. Chen, Y. H., Wang, W. C., Young, K. C., Chang, T. T., and Chen, S. H., Plastic microchip electrophoresis
for analysis of PCR products of hepatitis C virus, Clin. Chem., 45, 1938, 1999.
50. Young, K. C., Lien, H. M., Lin, C. C., Chang, T. T., Lee, G. B., and Chen, S. H., Microchip and
capillary electrophoresis for quantitative analysis of hepatitis C virus based on RT-competitive PCR,
Talanta, 56, 323, 2002.
51. Cangelosi, G. A., Freeman, R. J., Lewis, K. N., Livingston-Rosanoff, D., Shah, K. S., Milan, S. J.,
and Goldberg, S. V., Evaluation of a high-throughput repetitive-sequence-based PCR system for DNA
fingerprinting of Mycobacterium tuberculosis and Mycobacterium avium complex strains, J. Clin.
Microbiol., 42, 2685, 2004.
52. Ramsey, J. D., Jacobson, S. C., Culbertson, C. T., and Ramsey, J. M., High-efficiency, two-dimensional
separations of protein digests on microfluidic devices, Anal. Chem., 75, 3758, 2003.
53. Colyer, C. L., Mangru, S. D., and Harrison, D. J., Microchip-based capillary electrophoresis of human
serum proteins, J. Chromatogr. A, 781, 271, 1997.
54. Callewaert, N., Contreras, R., Mitnik-Gankin, L., Carey, L., Matsudaira, P., and Ehrlich, D.,
Total serum protein N-glycome profiling on a capillary electrophoresis-microfluidics platform,
55. Weiller, B. H., Ceriotti, L., Shibata, T., Rein, D., Roberts, M. A., Lichtenberg, J., German, J. B.,
de Rooij, N. F., and Verpoorte, E., Analysis of lipoproteins by capillary zone electrophoresis in
microfluidic devices: Assay development and surface roughness measurements, Anal. Chem., 74,
1702, 2002.
56. Ceriotti, L., Shibata, T., Folmer, B., Weiller, B. H., Roberts, M. A., de Rooij, N. F., and Verpoorte, E.,
Low-density lipoprotein analysis in microchip capillary electrophoresis systems, Electrophoresis, 23,
3615, 2002.
57. Yao, S., Anex, D. S., Caldwell, W. B., Arnold, D. W., Smith, K. B., and Schultz, P. G., SDS capillary gel
electrophoresis of proteins in microfabricated channels, Proc. Natl. Acad. Sci. USA, 96, 5372, 1999.
58. Liu, Y. J., Foote, R. S., Jabobson, S. C., Ramsey R. S., and Ramsey, J. M., Electrophoretic separation
of proteins on a microchip with noncovalent postcolumn labeling, Anal. Chem., 72, 4606, 2000.
59. Bousse, L., Mouradian, S., Minalla, A., Yee, H., Williams, K., and Dubrow, R., Protein sizing on a
microchip, Anal. Chem., 73, 1207, 2001.
60. Giordano, B. C., Jin, L. J., Couch, A. J., Ferrance, J. P., and Landers, J. P., Microchip laser-induced
fluorescence detection of proteins at submicrogram per milliliter levels mediated by dynamic labeling
under pseudonative conditions, Anal. Chem., 76, 4705, 2004.
61. Chan, O. T. M., and Herold, D. A., Chip electrophoresis as a method for quantifying total
microalbuminuria, Clin. Chem., 52, 2141, 2006.
62. Zhuang, G. S., Liu, J., Jia, C. P., Jin, Q. H., Zhao, J. L., and Wang, H. M., Manufacture of reversible
electrophoresis chip and its application to protein detection, Acta Chim. Sinica, 63, 1003, 2005.
63. Zhuang, G. S., Liu, J., Jia, C. P., Jin, Q. H., Wang, H. M., Yang, M. S., and Zhao, J. L., Capillary
electrophoresis microchip assay of clinical urinary proteins, Acta Chim. Sinica, 64, 229, 2006.
64. Zhuang, G. S., Jin, Q. H., Liu, J., Cong, H., Liu, K. D., Zhao, J. L., Yang, M. S., and Wang, H. M., A low
temperature bonding of quartz microfluidic chip for serum lipoproteins analysis, Biomed. Microdevice,
8, 255, 2006.
65. Wang, J., Zhou, S., Huang, W., Liu, Y., Cheng, C., Lu, X., and Cheng, J., CE-based analysis of
hemoglobin and its applications in clinical analysis, Electrophoresis, 27, 3108, 2006.
66. Schmalzing, D., Koutny, L. B., Taylor, T. A., Nashabeh, W., and Fuchs, M., Immunoassay for thyrox-
ine (T4) in serum using capillary electrophoresis and micromachined devices, J. Chromatogr. B, 697,
175, 1997.
67. Koutny, L. B., Schmalzing, D., Taylor, T. A., and Fuchs, M., Microchip electrophoretic immunoassay
for serum cortisol, Anal. Chem., 68, 18, 1996.
68. Qiu, C. X., and Harrison, D. J., Integrated self-calibration via electrokinetic solvent proportioning for
microfluidic immunoassays, Electrophoresis, 22, 3949, 2001.
69. Wang, J., Ibanez, A., Chatrathi, M. P., and Escarpa, A., Electrochemical enzyme immunoassays on
microchip platforms, Anal. Chem., 73, 5323, 2001.
70. Wang, J., Ibanez, A., and Chatrathi, M. P., Microchip-based amperometric immunoassays using redox
tracers, Electrophoresis, 23, 3744, 2002.
71. Herr, A. E., Throckmorton, D. J., Davenport, A. A., and Singh, A. K., On-chip native gel
electrophoresis-based immunoassays for tetanus antibody and toxin, Anal. Chem., 77, 585, 2005.
72. Fanguy, J. C., and Henry, C. S., The analysis of uric acid in urine using microchip capillary
electrophoresis with electrochemical detection, Electrophoresis, 23, 767, 2002.
73. Garcia, C. D., and Henry, C. S., Direct detection of renal function markers using microchip CE with
pulsed electrochemical detection, Analyst, 129, 579, 2004.
74. Du, Y., Yan, J. L., Zhou, W. Z., Yang, X. Y., and Wang, E. K., Direct electrochemical detection
of glucose in human plasma on capillary electrophoresis microchips, Electrophoresis, 25, 3853,
2004.
75. Suzuki, S., and Honda, S., Miniaturization in carbohydrate analysis, Electrophoresis, 24, 3577, 2003.
76. Zhao, X. C., You, T. Y., Qiu, H. B., Yan, J. L., Yang, X. R., and Wang, E. K., Electrochemilumi-
nescence detection with integrated indium tin oxide electrode on electrophoretic microchip for direct
bioanalysis of lincomycin in the urine, J. Chromatogr. B, 810, 137, 2004.
77. Crevillen, A. G., Barrigas, I., Blasco, A. J., Gonzalez, M. C., and Escarpa, A., Microchip-
electrochemistry route for rapid screening of hydroquinone and arbutin from miscellaneous samples:
Investigation of the robustness of a simple cross-injector system, Anal. Chim. Acta, 562, 137, 2006.
78. Laugere, F., Guijt, R. M., Bastemeijer, J., van der Steen, G., Berthold, A., Baltussen, E., Sarro, P., van
Dedem, G. W. K., Vellekoop, M., and Bossche, A., On-chip contactless four-electrode conductivity
detection for capillary electrophoresis devices, Anal. Chem., 75, 306, 2003.
79. Munro, N. J., Huang, Z., Finegold, D. N., and Landers, J. P., Indirect fluorescence detection of amino
acids on electrophoretic microchips, Anal. Chem., 72, 2765, 2000.
80. Caglar, P., Tuncel, S. A., Malcik, N., Landers, J. P., and Ferrance, J. P., A microchip sensor for calcium
determination, Anal. Bioanal. Chem., 386, 1303, 2006.
81. Huang, Y., Mather, E. L., Bell, J. L., and Madou, M., MEMS-based sample preparation for molecular
diagnostics, Anal. Bioanal. Chem., 372, 49, 2002.
82. Zhang, C. S., Xu, J. L., Ma, W. L., and Zheng, W. L., PCR microfluidic devices for DNA amplification,
Biotech. Adv., 24, 243, 2006.
83. Mikhailovich, V. M., Lapa, S. A., Gryadunov, D. A., Strizhkov, B. N., Sobolev, A. Y., Skotnikova, O. I.,
Irtuganova, O. A., et al., Detection of rifampicin-resistant Mycobacterium tuberculosis strains by
hybridization and polymerase chain reaction on a specialized TB-microchip, Bull. Exp. Biol. Med.,
131, 94, 2001.
84. Yang, J. N., Liu, Y. J., Rauch, C. B., Stevens, R. L., Liu, R. H., Lenigk, R., and Grodzinski, P., High
sensitivity PCR assay in plastic micro reactors, Lab Chip, 2, 179, 2002.
85. Cho, Y. K., Kim, J., Lee, Y., Kim, Y. A., Namkoong, K., Lim, H., Oh, K. W., et al., Clinical evaluation
of micro-scale chip-based PCR system for rapid detection of hepatitis B virus, Biosens. Bioelec., 21,
2161, 2006.
86. Gulliksen, A., Solli, L. A., Drese, K. S., Sorensen, O., Karlsen, F., Rogne, H., Hovig, E., and Sirevag,
R., Parallel nanoliter detection of cancer markers using polymer microchips, Lab Chip, 5, 416, 2005.
87. Cheng, J., Waters, L. C., Fortina, P., Hvichia, G., Jacobson, S. C., Ramsey, J. M., Kricka, L. J., and Wild-
ing, P., Degenerate oligonucleotide primed polymerase chain reaction and capillary electrophoretic
analysis of human DNA on microchip-based devices, Anal. Biochem., 257, 101, 1998.
88. Lou, X. J., Panaro, N. J., Wilding, P., Fortina, P., and Kricka, L. J., Mutation detection using ligase chain
reaction in passivated silicon-glass microchips and microchip capillary electrophoresis, Biotechniques,
37, 392, 2004.
89. Easley, C. J., Legendre, L. A., Roper, M. G., Wavering, T. A., Ferrance, J. P., and Landers, J. P., Extrin-
sic Fabry-Perot interferometry for noncontact temperature control of nanoliter-volume enzymatic
reactions in glass microchips, Anal. Chem., 77, 1038, 2005.
90. Peterson, D. S., Rohr, T., Svec, F., and Frechet, J. M. J., Enzymatic microreactor-on-a-chip: Protein
mapping using trypsin immobilized on porous polymer monoliths molded in channels of microfluidic
devices, Anal. Chem., 74, 4081, 2002.
91. Jin, L. J., Ferrance, J., Sanders, J. C., and Landers, J. P., A microchip-based proteolytic digestion
system driven by electroosmotic pumping, Lab Chip, 3, 11, 2003.
92. Crowley, T. A., and Pizziconi, V., Isolation of plasma from whole blood using planar microfilters for
lab-on-a-chip applications, Lab Chip, 5, 922, 2005.
93. Shevkoplyas, S. S., Yoshida, T., Munn, L. L., and Bitensky, M. W., Biomimetic autoseparation of
leukocytes from whole blood in a microfluidic device, Anal. Chem., 77, 933, 2005.
94. Mansfield, C. D., Man, A., Low-Ying, S., and Shaw, R. A., Laminar fluid diffusion interface precon-
ditioning of serum and urine for reagent-free infrared clinical analysis and diagnostics, Appl. Spec.,
59, 10, 2005.
95. Breadmore, M. C., Wolfe, K. A., Arcibal, I. G., Leung, W. K., Dickson, D., Giordano, B. C.,
Power, M. E., Ferrance, J. P., Feldman, S. H., Norris, P. M., and Landers, J. P., Microchip-based
purification of DNA from biological samples, Anal. Chem., 75, 1880, 2003.
96. Chung, Y. C., Jan, M. S., Lin, Y. C., Lin, J. H., Cheng, W. C., and Fan, C. Y., Microfluidic chip for
high efficiency DNA extraction, Lab Chip, 4, 141, 2004.
97. Wu, Q. R., Bienvenue, J. M., Hassan, B. J., Kwok, Y. C., Giordano, B. C., Norris, P. M., Landers, J. P.,
and Ferrance, J. P., Microchip-based macroporous silica sol-gel monolith for efficient isolation of DNA
from clinical samples, Anal. Chem., 78, 5704, 2006.
98. Wen, J., Guillo, C., Ferrance, J. P., and Landers, J. P., DNA extraction using a tetram-
ethyl orthosilicate-grafted photopolymerized monolithic solid phase, Anal. Chem., 78, 1673,
2006.
99. Legendre, L. A., Bienvenue, J. M., Roper, M. G., Ferrance, J. P., and Landers, J. P., A valveless
microfluidic sample preparation device for DNA extraction and amplification of DNA from nanoliter
volume samples using conventional instrumentation, Anal. Chem., 78, 1444, 2006.
100. Witek, M., Llopis, S. D., Wheatley, A., McCarley, R. L., and Soper, S. A., Purification and pre-
concentration of genomic DNA from whole cell lysates using photoactivated polycarbonate (PPC)
microfluidic chips, Nucleic Acids Res., 34, e74/1, 2006.
101. Nakagawa, T., Tanaka, T., Niwa, D., Osaka, T., Takeyama, H., and Matsunaga, T., Fabrication of
amino silane-coated microchip for DNA extraction from whole blood, J. Biotechnol., 116, 105, 2005.
102. Cao, W. D., Easley, C. J., Ferrance, J. P., and Landers, J. P., Chitosan as a polymer for pH-induced
DNA capture in a totally aqueous system, Anal. Chem., 78, 7222, 2006.
103. Kokoris, M., Nabavi, M., Lancaster, C., Clemmens, J., Maloney, P., Capadanno, J., Gerdes, J.,
and Battrell, C. F., Rare cancer cell analyzer for whole blood applications: Automated nucleic acid
purification in a microfluidic disposable card, Methods, 37, 114, 2005.
104. Roper, M. G., Frisk, M. L., Oberlander, J. P., Ferrance, J. P., McGrory, B. J., and Landers, J. P.,
Extraction of C-reactive protein from serum on a microfluidic chip, Anal. Chim. Acta, 569, 195, 2006.
105. Liu, Y., Garcia, C. D., and Henry, C. S., Recent progress in the development of µTAS for clinical
analysis, Analyst, 128, 1002, 2003.
106. Lagally, E. T., and Soh, H. T., Integrated genetic analysis microsystems, Crit. Rev. Solid State Mat.
Sci., 30, 207, 2005.
107. Woolley, A. T., Hadley, D., Landre, P., deMello, A. J., Mathies, R. A., and Northrup, M. A., Functional
integration of PCR amplification and capillary electrophoresis in a microfabricated DNA analysis
device, Anal. Chem., 68, 4081, 1996.
108. Waters, L. C., Jacobson, S. C., Kroutchinina, N., Khandurina, J., Foote, R. S., and Ramsey, J. M.,
Microchip device for cell lysis, multiplex PCR amplification, and electrophoretic sizing, Anal. Chem.,
70, 158, 1998.
109. Waters, L. C., Jacobson, S. C., Kroutchinina, N., Khandurina, J., Foote, R. S., and Ramsey, J. M.,
Multiple sample PCR amplification and electrophoretic analysis on a microchip, Anal. Chem., 70,
5172, 1998.
110. Cheng, J., Waters, L. C., Fortina, P., Hvichia, G., Jacobson, S. C., Ramsey, J. M., Kricka,
L. J., and Wilding, P., Degenerate oligonucleotide primed polymerase chain reaction and capillary
electrophoretic analysis of human DNA on microchip-based devices, Anal. Biochem., 257, 101,
1998.
111. Khandurina, J., McKnight, T. E., Jacobson, S. C., Waters, L. C., Foote, R. S., and Ramsey, J. M.,
Integrated system for rapid PCR-based DNA analysis in microfluidic devices, Anal. Chem., 72, 2995,
2000.
112. Lagally, E. T., Simpson, P. C., and Mathies, R. A., Monolithic integrated microfluidic DNA
amplification and capillary electrophoresis analysis system, Sens. Act. B–Chem., 63, 138, 2000.
113. Koh, C. G., Tan, W., Zhao, M. Q., Ricco, A. J., and Fan, Z. H., Integrating polymerase chain reaction,
valving, and electrophoresis in a plastic device for bacterial detection, Anal. Chem., 75, 4591, 2003.
114. Lagally, E. T., Scherer, J. R., Blazej, R. G., Toriello, N. M., Diep, B. A., Ramchandani, M., Sens-
abaugh, G. F., Riley, L. W., and Mathies, R. A., Integrated portable genetic analysis microsystem for
pathogen/infectious disease detection, Anal. Chem., 76, 3162, 2004.
115. Easley, C. J., Karlinsey, J. M., and Landers, J. P., On-chip pressure injection for integration of
infrared-mediated DNA amplification with electrophoretic separation, Lab Chip, 6, 601, 2006.
116. Zhou, Z. M., Liu, D. Y., Zhong, R. T., Dai, Z. P., Wu, D. P., Wang, H., Du, Y. G., Xia, Z. N., Zhang,
L. P., Mei, X. D., and Lin, B. C., Determination of SARS-coronavirus by a microfluidic chip system,
117. Pal, R., Yang, M., Lin, R., Johnson, B. N., Srivastava, N., Razzacki, S. Z., Chomistek, K. J., et al., An
integrated microfluidic device for influenza and other genetic analyses, Lab Chip, 5, 1024, 2005.
118. Kaigala, G. V., Huskins, R. J., Preiksaitis, J., Pang, X. L., Pilarski, L. M., and Backhouse, C. J.,
Automated screening using microfluidic chip-based PCR and product detection to assess risk of BK
virus-associated nephropathy in renal transplant recipients, Electrophoresis, 27, 3753, 2006.
Hughes, M. A., et al., A fully-integrated microfluidic genetic analysis system with sample in-answer
out capability, Proc. Natl. Acad. Sci. USA, 103, 19272, 2006.
120. Taylor, P., Manage, D. P., Helmle, K. E., Zheng, Y., Glerum, D. M., and Backhouse, C. J., Analysis of
mitochondrial DNA in microfluidic systems, J. Chromatogr. B, 822, 78, 2005.
121. Fortier, M. H., Bonneil, E., Goodley, P., and Thibault, P., Integrated microfluidic device for mass
spectrometry-based proteomics and its application to biomarker discovery programs, Anal. Chem., 77,
1631, 2005.
122. Gottschlich, N., Culbertson, C. T., McKnight, T. E., Jacobson, S. C., and Ramsey, J. M., Integrated
microchip-device for the digestion, separation and postcolumn labeling of proteins and peptides,
J. Chromatogr. B, 745, 243, 2000.
123. Song, S., Singh, A. K., and Kirby, B. J., Electrophoretic concentration of proteins at laser-patterned
nanoporous membranes in microchips, Anal. Chem., 76, 4589, 2004.
124. Hatch, A. V., Herr, A. E., Throckmorton, D. J., Brennan, J. S., and Singh, A. K., Integrated preconcen-
tration SDS-PAGE of proteins in microchips using photopatterned cross-linked polyacrylamide gels,
Anal. Chem., 78, 4976, 2006.
125. Silvertand, L. H. H., Machtejevas, E., Hendriks, R., Unger, K. K., van Bennekom, W .P., and de Jong,
G. J., Selective protein removal and desalting using microchip CE, J. Chromatogr. B, 839, 68, 2006.
126. Wang, J., and Chatrathi, M. P., Microfabricated electrophoresis chip for bioassay of renal markers,
Anal. Chem., 75, 525, 2003.
127. Wang, J., Chatrathi, M. P., Tian, B. M., and Polsky, R., Microfabricated electrophoresis chips for
simultaneous bioassays of glucose, uric acid, ascorbic acid, and acetaminophen, Anal. Chem., 72,
2514, 2000.
128. Easley, C. J., Legendre, L. A., Landers, J. P., and Ferrance, J. P., Rapid DNA amplification in
glass microdevices, in Microchip Capillary Electrophoresis: Methods and Protocols, C.S. Henry,
ed., Humana Press, Totowa, NJ, chap. 15, 2006.
129. Panaro, N. J., Lou, X. J., Wilding, P., Fortina, P., and Kricka, L. J., Surface effects on PCR reactions
in multichip microfluidic platforms, Biomed. Microdev., 6, 75, 2004.
130. Krishnan, M., Burke, D. T., and Burns, M., A polymerase chain reaction in high surface-to-volume
ratio SiO2 microstructures, Anal. Chem., 76, 6588, 2004.
131. Kricka, L. J., and Wilding, P., Microchip PCR, Anal. Bioanal. Chem., 377, 820, 2003.
132. Lou, X. J., Panaro, N. J., Wilding, P., Fortina, P., and Kricka, L. J., Increased amplification efficiency
of microchip-based PCR by dynamic surface passivation, Biotechniques, 36, 248, 2004.
133. Roach, L. S., Song, H., and Ismagilov, R. F., Controlling nonspecific protein adsorption in a plug-
based microfluidic system by controlling interfacial chemistry using fluorous-phase surfactants, Anal.
Chem., 77, 785, 2005.
134. Bi, H. Y., Meng, S., Li, Y., Guo, K., Chen, Y. P., Kong, J. L., Yang, P. Y., Zhong, W., and Liu, B. H.,
Deposition of PEG onto PMMA microchannel surface to minimize nonspecific adsorption, Lab Chip,
6, 769, 2006.
135. Bi, H., Zhong, W., Meng, S., Kong, J., Yang, P., and Liu, B., Construction of a biomimetic surface on
microfluidic chips for biofouling resistance, Anal. Chem., 78, 3399, 2006.
136. Wu, D. P., Zhao, B. X., Dai, Z. P., Qin, J. H., and Lin, B. C., Grafting epoxy-modified hydrophilic
polymers onto poly(dimethylsiloxane) microfluidic chip to resist nonspecific protein adsorption, Lab
Chip, 6, 942, 2006.
137. Shrinivasan, S., Norris P. M., Landers, J. P., and Ferrance, J. P., A low-cost, low-power consumption,
miniature laser-induced fluorescence system for DNA detection on a microfluidic device, Clin. Lab.
Med., 27, 173, 2007.
138. Kim, D. S., Lee, S. H., Ahn, C. H., Lee, J. Y., and Kwon, T. H., Disposable integrated microfluidic
biochip for blood typing by plastic microinjection moulding, Lab Chip, 6, 794, 2006.
139. Ahn, C. H., Choi, J. W., Beaucage, G., Nevin, J. H., Lee, J. B., Puntambekar, A., and Lee, J. Y.,
Disposable smart lab on a chip for point-of-care clinical diagnostics, Proc. IEEE, 92, 154, 2004.
140. Vestad, T., Marr, D. W. M., and Oakey, J., Flow control for capillary-pumped microfluidic systems,
J. Micromech. Microeng., 14, 1503, 2004.
37 Advances in Microfluidics:
Development of a Forensic
Integrated DNA Microchip
(IDChip)
Katie M. Horsman and James P. Landers
CONTENTS
37.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1065

37.2 Differential Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1066
37.3 DNA Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1068
37.4 DNA Quantification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1069
37.5 DNA Amplification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1070
37.6 Separation of PCR Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1074
37.7 Impact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1078
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1081
37.1 INTRODUCTION
The past half decade has seen deoxyribonucleic acid (DNA) analysis performed on microdevices
become more accepted and much more routine. Only recently, however, have analytical chemists
addressed the specific needs of forensic science to allow the forensics community to harness these
advances in microfluidics. The research focus of a number of groups has led to the development of
microfluidic systems specifically tailored to forensic applications. As a result, we are beginning to
witness the validation of microfluidic devices in forensic laboratories, and their widespread adoption
in the coming years can be anticipated. Microfluidics has the potential to advance forensic DNA
analysis in ways that were unforeseen at the time that restriction fragment length polymorphisms
(RFLPs) were adopted. It will introduce a rapid, automatable technology that will enable our severely
backlogged crime laboratories to process casework more efficiently.
In many respects, the forensics community has been awaiting the advent of microdevices.
Microfluidic chips have been proven to carry out processes more efficiently than their macroscale
counterparts (e.g., polymerase chain reaction or PCR and electrophoretic separations), thus, consider-
ably decreasing sample processing time.1,2 The current advances in integration of multiple processes
on a single device serve to further enhance the processing speed.3−5 Single-process devices, such as
those for DNA purification as well as multiprocess devices, will significantly reduce the opportunity
for laboratory sources of sample contamination by completing the sample preparation in a closed
system and, thus, eliminating the multiple tube transfers and additional handling. Forensics, perhaps
more than any other field, stands to benefit from this inherently closed-system design. The ease
1065
of automation of the microchip methods is advantageous to the forensics community in the same
way it will benefit the clinical community—by decreasing technician time and, thus, dramatically
decreasing costs associated with DNA analysis. This decrease in sample processing time and costs
could, in turn, have a dramatic impact on the DNA analysis backlog in crime laboratories. Although
yet to be fully realized, it is reasonable to expect a concomitant improvement in limit of detection
by the transfer to the microscale format from the macroscale counterpart. Therefore, we can expect
to see successful DNA typing in cases where limited or no profile was previously obtained with
conventional methods.
Conventional forensic DNA analysis involves DNA extraction and purification, quantification,
and amplification, followed by separation of PCR products, detection, and data analysis. In sex-
ual assault casework, differential extraction, a cell sorting process by which DNA is differentially
extracted from sperm and vaginal epithelial cells (or other nonsperm cell types), must be utilized
to obtain enriched fractions of male and female DNA. In this chapter, we will discuss the specific
development of these techniques on the microscale for application to forensic human DNA analysis.
37.2 DIFFERENTIAL EXTRACTION

Differential extraction is employed by forensic DNA analysts to obtain enriched fractions of male
and female DNA from sexual assault evidence. This evidence, typically obtained from a vaginal
swab during evidence collection at the emergency room, contains sperm cells from the perpetrator
often in an overwhelming number of vaginal epithelial cells from the victim. Although the victim’s
DNA profile can often be obtained by other means, a “clean” profile of the perpetrator is sought
for prosecution. As a result, Gill6 developed differential extraction, which exploits the differential
stability of the nuclei of sperm and vaginal epithelial cells to obtain enriched male and female
DNA fractions. The mixed cell sample is first extracted using mild conditions (Gill Buffer [10 mM
trizma, pH 8, 10 mM EDTA, 0.1 M NaCl, and 2% SDS] and 20 µg/mL proteinase K, incubate for
2+ h at 56◦ C), the swab removed, and sample centrifuged. The supernatant contains female DNA
(from lysed vaginal epithelial cells) while the pellet contains unlysed nuclei from sperm heads. After
extensive washing of the sperm cell pellet, incubation under harsh extraction conditions [Gill buffer
containing 40 mM dithiothreitol (DTT) and 20 µg/mL proteinase K] results in the lysis of the sperm
cell nuclei.
In the translation of this process to the microscale, several approaches have been developed. The
conventional method has not been directly translated to the chip, most likely due to the inherent
need for centrifugation. While centrifugation has been demonstrated on microdevices,7 it is not in
widespread use, requires specialized chip design, and the development of methods for washing the
sperm cell pellet in order to obtain results comparable to the macroscale method.
One microscale method developed for differential extraction involves a filter-based system and
lysis using acoustic energy.8 The sample is first infused over a filter (size and material not indicated)
in which the sperm cells (∼4–6 µm diameter) pass through unimpeded and the much larger epithelial
cells (∼50 µm diameter) are retained. The DNA is then extracted using ultrasonic disruption of the
cells. Although it is too early to gauge the success of this method, filtration has been explored on the
macroscale for this application without widespread success.
Another method for the analysis of sexual assault evidence exploits the differential physical
and chemical properties of sperm and epithelial cells to result in a cell separation, from which
DNA can be extracted from each population independently.9 In this simple method, the epithelial
cells settle to the bottom of the microchip inlet reservoir more rapidly than the sperm cells (as a
result of their size and density). By subsequently invoking flow in the microchannel, sperm cells
are swept from the inlet to the outlet reservoir, where they can be either collected for subsequent
analysis or packed up against a silica bead/sol-gel bed for DNA extraction (Figure 37.1a). Efficient
separation of a mixture of sperm and epithelial cells has been demonstrated via short tandem repeat
Advances in Microfluidics: Development of a Forensic Integrated DNA Microchip (IDChip) 1067
(a)
Epithelial cell/sperm Sperm cells in

mixture in inlet microchannel
(b)
Amel Cell separation product (sperm fraction)

D3S1358 THO1 TPOX
D16S539
CSF1PO
Amel Semen (positive control)

D3S1358 THO1 TPOX
D16S539
CSF1PO
FIGURE 37.1 Separation of sperm and vaginal epithelial cells on a glass microdevice. (a) Photographs of the
straight-channel microdevice (center), the inlet reservoir (containing a mixture of sperm and vaginal epithelial
cells), and the microchannel (containing sperm cells) during cell separation. In this method, the cells were
separated based on their physicochemical differences, resulting in differential sedimentation of the cell types.
Upon application of a negative pressure at the outlet reservoir, sperm cells were swept into the microchannel,
whereas the epithelial cells adsorbed to the glass substrate of the inlet reservoir. (b) Forensic STR profile
obtained from the material in the outlet reservoir following cell separation. This compares favorably with the
semen donor’s profile and indicates an efficient separation from the vaginal epithelial cell DNA. (Adapted
and reproduced from Horsman et al., Anal Chem 2005, 77, 742–749. Copyright 2005. With permission from
American Chemical Society.)
(STR) profiling (Figure 37.1b), and, in addition, cell separation with integrated DNA extraction has
also been demonstrated.10 A drawback associated with this method is the sample volume that can
be loaded into the chip. The proof-of-principle work demonstrated a sample load that was in the
10–50 µL range, where biological material from vaginal swabs are typically eluted in 500 µL or
more. Because the number of sperm cells on a given swab is unknown (and, thus, the amount of
DNA available for subsequent processing), the device must be capable of accepting the entire volume
from the evidentiary sample. In addition, this method represents a divergence from the conventional
macroscale method, assuming that both cell types exist on and are eluted from the swab intact.
Any epithelial cells that have lysed in storage (due to dehydration) will yield free female DNA that
co-migrates to the outlet reservoir with the sperm cells, thus contaminating the male fraction. While
this method has shown promise for use in forensic DNA analysis, it did not meet the requirements
for accommodating larger sample volumes.
An interesting approach reported for obtaining purified male and female fractions on microdevices
involves acoustic differential extraction (ADE)11 and exploits microacoustic transducer develop-
ments (see Chapter 44 by Laurell for details). This method involves the elution of biological
material from the vaginal swab under mild lysis conditions, resulting in a mixture of sperm cells
and epithelial cell lysate. The sample is then infused in the microdevice, and sperm cells trapped in
a monolayer above an ultrasonic transducer while free DNA (from the lysed epithelial cells) flows
through the system unretained. Flow of the epithelial cell lysate is directed to the outlet reservoir
using laminar flow valving12 while the sperm cells are retained by the transducer. Upon deactiva-
tion of the ultrasound, the sperm cells are released and directed to the sperm cell outlet. Highly
enriched fractions of male and female DNA have been shown to be obtained using this method. In
addition, this approach addresses the shortcoming of the aforementioned cell sorting method, in that
it is capable of readily processing large volume samples, although cell sorting integrated with DNA
extraction is under development but has not yet been demonstrated.
All of these methods are amenable to eventual integration with downstream sample preparation
steps, such as the solid-phase extraction of DNA, for automated processing of samples in a format that
is a closed, fully contained system, thereby diminishing contamination by outside sources. However,
none of the methods described here identify sperm cells specifically. Before DNA analysis, analysts
conduct a time-consuming “sperm search” using light microscopy to identify sperm morphologically
and gauge the relative abundance. While the “sperm search” is currently considered a separate entity,
any method that could incorporate the positive identification of sperm cells, would be attractive to
the forensic laboratories and eliminate another time-consuming analysis step. Although not on a
microchip, one method that has been detailed for this is laser microdissection, which has recently
been used to identify and capture sperm cells from a membrane-coated microscope slide.13,14 While
effective, particularly for evidence containing low cell numbers, this method is not likely to be used
for high-throughput casework.
While several microscale alternatives to differential extraction have been proposed and demon-
strated, one school of thought is that none are likely to significantly impact the forensic community
without direct integration with DNA extraction.
37.3 DNA EXTRACTION

DNA extraction and purification on microdevices has gained widespread use for clinical and forensic
applications, and has been reviewed in detail in Chapter 43 by Bienvenue. On the macroscale, DNA
purification is completed using organic extraction (historically), or, more commonly, solid phases
such as QIAamp and DNA IQ. The development of DNA extraction methods on the microscale
has largely focused on the use of silica-based solid phases,15−19 although purification has also been
recently demonstrated using an ion-exchange resin in a microdevice.20 Any microscale DNA extrac-
tion method must demonstrate high efficiency of extraction comparable to or better than the currently
available methods (e.g., ∼46% for whole blood21 using QIAamp). In addition, the extraction bed
must be of relatively high capacity to enable binding of sufficient DNA for downstream processing,
and the extraction must be capable of accepting large volumes (e.g., up to 500 µL) in a timely
manner.
Most solid-phase DNA extraction techniques involve loading the DNA onto the solid phase using
a chaotropic salt (e.g., Guanidine HCl), washing with ethanol to remove proteins, and eluting the
DNA with water or other solvent. A more detailed explanation of these methods can be found in
Chapter 43, entitled “Integrating Sample Preparation Into Multifluidic Devices.”
Christel et al.16 reported the first microdevice-based solid-phase extraction of DNA in 1999 using
an etched silica solid phase consisting of silicon pillars. This work reported extraction efficiencies
of approximately 50% from prepurified DNA. As a result, the extension of this method to forensic
samples, which are often low-copy number samples, seems unlikely. Higher extraction efficiencies
have been demonstrated using silica bead and sol-gel-based solid phases. These phases are added to
the microchannel as liquid suspensions, thus, circumventing the complex etching required for the
pillar design described previously. Silica beads and sol-gel have been demonstrated individually3,22
and in tandem17 where the sol-gel acts as a glue holding the beads in place. In both methods,
DNA purification has been demonstrated using blood and other biological fluids common to forensic
analyses with high extraction efficiencies (up to 60%) and high capacities (∼20 ng, 1.5 cm extraction
bed) comparable to the macroscale counterpart such as the QIAamp kit. In a clear demonstration
of the application of this work to forensic analyses, on-chip sperm cell lysis and DNA extraction
using a silica bead solid phase was reported by the addition of 40 mM DTT to the 6 M guanidine
lysis solution.18 The purified DNA was shown to be amenable to downstream (off-chip) processing
including STR amplification and separation of PCR products.
DNA purification on microdevices, as currently performed, appears to meet the needs of the
forensics community. That is, extraction efficiencies and capacities comparable to the conventional
methods have been demonstrated. Most notably, these methods typically result in elution volumes
on the order of 10 µL or less, negating the need for an additional concentration step (e.g., Microcon)
and, thus, further diminishing DNA loss and additional processing time. Because the commercially
available kits accept limited volumes of DNA, concentration of the eluted DNA is commonly required
in forensic samples in order to maximize the amount of template DNAadded to the STR amplification.
In addition, because the target amount of template DNA added to an STR reaction is important in
order to minimize possible PCR artifacts such as allelic dropout, off-scale alleles, and so forth, DNA
quantification following extraction is necessary.
37.4 DNA QUANTIFICATION

Unlike DNA diagnostics with clinical samples, where the amount of DNA in a volume of blood or
other biological fluid can be estimated with little consequence, postpurification DNA quantification is
of the utmost importance in forensic sample analysis. The target mass of DNA required for effective
use of the commercially available STR kits is ∼1 ng. When a reduced mass of DNA is added to
the PCR, stochastic effects that could include “allelic dropout,” may make interpretation difficult.
Alternatively, an excess of DNA in the PCR (>5 ng) can result in off-scale peaks, “pull-up” (color
bleeding usually due to off-scale peaks), increased levels of stutter (small peaks typically one repeat
unit shorter than the STR allele, believed to be due to slippage of the primer or template during
replication) or other artifacts in the subsequent electrophoretic separation.
Until recently, DNA quantification in forensic analyses has most commonly involved the use of
slot-blot methods. In the past few years, the forensics community has shifted focus to the use of quan-
titative PCR as the preferred method for quantification. Both methods allow for human-specific DNA
quantification, although qPCR has significantly lower limits of detection and gives an indication of
the PCR-amplifiability of the DNA as well. Recently, multiplex qPCR assays have been developed to
simultaneously determine the total amount of genomic and mitochondrial DNA,23−25 total genomic
and male DNA,26,27 or assess the extent of DNA degradation for specific forensic applications.28,29
Because of the utility of qPCR to the forensic community, we focus on the development of qPCR
on microdevices here.
Although real-time PCR has been demonstrated on microdevices, no qPCR has been reported to
date. That is, fluorescence detection during the accumulation of PCR products throughout cycling
has been demonstrated. However, no standard curve was simultaneously amplified to allow for DNA
quantification based on the signal generated. Real-time PCR has been demonstrated by Lin et al.30
in 25 µL reactions using fluorogenic probes and the SYBR Green I intercalating dye, respectively.
In both cases, multiplex PCRs that include generation of a standard curve for quantification have not
been described. In 2003, Quake and coworkers31 reported the development of picoliter volume PCR
chambers on a microfluidic device for parallel real-time PCR assays. Although not demonstrated,
the authors indicate that the eight flow channels can be used for the generation of a standard curve
and, thus, the chip can be used for quantitative PCR.
While DNA quantification remains the least-developed sample processing step on microchip
for forensic DNA analysis, it is anticipated that extension of the reported methods to qPCR
will be forthcoming. Any such qPCR method should be directly applicable to forensic DNA
quantification by use of the appropriate primers and probes described in the literature for the
macroscale counterpart. Following successful demonstration of qPCR on-chip, integration of this
method with DNA extraction and DNA amplification would be the next microfluidic challenge;
however, recent advances in this arena suggest this is possible.32−34 In addition, development of
valving or other means of metering the appropriate volume of purified DNA into the amplification
reaction must be carried out. This then results in the need for a storage mechanism on chip or
off chip for the unused extracted DNA, which is derived from the original evidence sample. This
material must be saved for subsequent reprocessing, permanent storage, or use by opposing legal
council.
37.5 DNA AMPLIFICATION

DNA amplification on microdevices has been demonstrated routinely, and Roper et al.35 have not
only provided an extensive review of the literature on this topic (also see Chapter 43 by Bienvenue)
but also provided a scoring method for device comparison. However, it is noteworthy that few
examples of the complex PCR associated with forensic DNA typing have been demonstrated on chip.
Forensic DNA typing most commonly involves the amplification of STRs via PCR using multiplex
PCRs involving up to 16 primer pairs for coamplification of STRs from multiple loci. Primers are
tagged with fluorescent tags (typically one of three or four fluorophores) for discrimination of loci.
Commercially available kits have been developed, are the industry standard, and are used by every
forensic DNA laboratory in the United States. Of primary interest to the forensics community in the
development of a microfluidic device for PCR is the limit of detection for the amplification fragments
(low-copy number amplifications), unbiased amplification, limiting the DNA template required for
effective amplification, total reaction time, the ability to perform multiplex amplifications (as in the
commercially available kits), and multiple chamber amplification (to amplify >1 sample at a time).
Of course, any microdevice method must show that the PCR amplification product is comparable
to that obtained on the macroscale, permitting any future results to be related to those previously
obtained and are contained in the national DNA database.
Microchip DNA amplification, to date, has already demonstrated the potential for low limits of
detection, with amplification demonstrated from single-copy sources.36 The useful limit of detection
(so as to prevent stochastic effects in forensic analyses) with microchip PCR is yet to be determined.
However, the reduced reaction volume inherent to the microchip supports the expectation of lower
limits of detection. Fast thermocycling has been demonstrated by a number of microchip PCR designs
and methods with amplification rates approaching the biological limit of the processivity of Taq poly-
merase, that is, with 20–25 cycles in under 5 min.1,2 This fast thermocycling has not yet been applied
to STR amplifications, although preliminary data have shown a reduction in total amplification time
by almost half.37 Therefore, it is not unreasonable to anticipate that the approximately 3-h conven-
tional thermocycling protocol will be reduced to 45 min or less. There are a few demonstrations
of STR amplifications in microdevices. Liu et al.4 have demonstrated the amplification of a mini
Y-STR multiplex, Bienvenue et al.5 and Legendre et al.37 have demonstrated amplification using
the AmpFlSTR COfiler and Profiler Plus kits, and Schmidt et al.40 demonstrated microchip ampli-
fication using the PowerPlex 16 STR kit. Recently, amplification of mitochondrial DNA has been
reported on a glass microchip.41 Following amplification, a subsequent mitochondrial sequencing
reaction was performed. These data indicated amplification with template amounts as low as 1 pg,
where product was not detected by the conventional method. Large masses of mitochondrial DNA
(500 pg–1 ng) resulted in some nonspecific product formation when amplified in the low volume
reaction. Table 37.1 summarizes the characteristics of these reported microchip forensic DNA ampli-
fications. In addition to multiproduct amplification in each chamber (corresponding to the multiple
loci), simultaneous amplification in multiple chambers will likely be needed to compete with the
conventional thermocyclers. This has been shown by Waters et al.42 (four chambers), Legendre
et al.37 (three reactions performed—although four chambers are fabricated in the device)
(Figure 37.2), Lutz-Bonengel et al.41 (60 reaction spots), and Schmidt et al.40 (23 reactions
TABLE 37.1
Forensic DNA Amplifications on Microdevices
Author (Citation) Reaction Volume Reaction Time Heating Method Amplification Demonstrated Integrated? Multichamber?
Liu et al.31 160 nL 64 min (35 cycles) Resistive Mini Y-STR multiplex incl. Yes, to microchip No
Amelogenin electrophoresis
Legendre et al.37 200 nL 100 min (28 cycles) Block (conventional) COfiler, Profiler Plus No Yes, 4
Schmidt et al.40 1 µL 210 min (32 cycles) Block (conventional) PowerPlex 16 No Yes, 60
Bienvenue et al.5 1.2 µL 180 mina (28 cycles) Block (conventional) COfiler, Profiler Plus Yes, to DNA purification Yes, 4
Lutz-Bonengel et al.41 1 µL 180 minb (32 cycles) Block (conventional) Mitochondrial No Yes, 60
a This reaction time is reported for the PCR amplification and the DNA extraction in total.
b This reaction time is reported for the PCR amplification only, not the sequencing step, which requires approximately an additional 220 min.
Advances in Microfluidics: Development of a Forensic Integrated DNA Microchip (IDChip)
1071
(b)
COfiler®
(a)
(c)
(d) Amel TH01 TPOX
D3S1358
D16S539
D7S820
CSF1PO
100 150 200 250 300 350

Fragment size (bp)
FIGURE 37.2 Multiple simultaneous STR amplifications on a single glass microdevice. (a) Photograph of
the microdevice for multiple chamber PCR microdevice. The chip contains five chambers: one for temperature
control (center) and four surrounding PCR chambers for parallel amplification. Amplification of the COfiler
multiplex STR loci was performed simultaneously in three of the four reaction chambers on this microdevice
using less than 1 ng template DNA. (Adapted and reproduced from Legendre, L. et al. Multiplex microchip
PCR for STR Analysis. Poster presented at The 15th International Symposium on Human Identification, 2004.
With permission.) The resulting STR profiles (following conventional DNA separation) are shown in panels
B–D, indicating amplification in each chamber with no apparent PCR artifacts that would complicate analysis
of the DNA profile. (Figures 37.2B through 37.2D are courtesy of L. Legendre.)
performed—although the chip contains 60 reaction spots) (Figure 37.3). Multiple PCR chambers
per device will enable high-throughput processing but also allow analysts to simultaneously process
the necessary controls (reagent blanks, positives, negatives).
The application of microfabricated devices to the amplification of STR fragments in foren-
sic DNA typing provides numerous advantages over the conventional methodology—specifically,
lower limits of detection,36 less reagent and sample consumption, enhanced processing speed,1 and
so forth. However, it also brings along issues that are not currently encountered with conventional
methodology. For example, the STR amplifications on microchip discussed above involve total
reaction volumes that range from 160 nL to 1.2 µL. Thus, the DNA added to the reaction must
be contained in a fraction of a microliter, rather than the common 10–200 µL obtained from a
Qiagen purification and then used for a conventional 25 µL amplification reaction. Consequently,
the DNA must be concentrated to a greater extent than needed for conventional amplification. As
a result, for microchip PCR to be utilized in forensic laboratories it will most likely require direct
integration with microchip DNA extraction, where the DNA is routinely eluted in microliter vol-
umes. This need to amplify the entire mass of DNA and, thus, concentrate the purified DNA to a
high degree, is most critical in cases where low starting copies of DNA template are available.
Bienvenue et al.5 have demonstrated DNA purification and PCR amplification on the same
microdevice, carrying out microliter-scale DNA extraction as reported previously in a glass
(a)
(b)
FIGURE 37.3 (a) Photograph of a glass microchip containing 60 reaction spots, each with a hydrophilic
surface surrounded by a hydrophobic ring. Each spot is loaded with 0.5 µL PCR mix and 0.5 µL of DNA
template and covered with 5 µL oil to decrease evaporation and contamination. Thus, the total reaction volume
on this device is 1 µL. The chip was thermocycled on conventional instrumentation using an in situ adapter.
(Photo courtesy of Schmidt, U.) (b) A complete STR profile generated by PowerPlex 16 amplification of 32 pg
genomic DNA on the chip shown in (a). Allele names are identified in the gray boxes above each set of peaks.
Repeat numbers for the corresponding alleles are indicated below each peak. (Adapted and reproduced from
Schmidt, U. et al., Int J Legal Med 2006, 120, 42–48. Copyright 2005. With permission from Springer-Verlag.)
microdevice.17,22,43 However, in a unique approach, the DNA was eluted from the solid phase
using PCR master mix and retained in a chamber. Subsequent placement of the device into a conven-
tional (block) thermocycler allowed for amplification of the STR fragments in the Cofiler and Profiler
Plus kits. These results are significant as they represent the first work demonstrating the potential
for interfacing microfluidic DNA extraction technology with conventional PCR instrumentation. Of
course, for application to forensic casework, the integration of a DNA quantification step will also be
necessary.
37.6 SEPARATION OF PCR PRODUCTS

As described in detail in Chapter 25 by McCord, forensic DNA profiling is routinely accomplished
by capillary electrophoresis (CE) in forensic laboratories. Recently, microchip electrophoresis is
being evaluated for use in casework by the Virginia Department of Forensic Science and the Forensic
Science Service (U.K.). While DNA separations on microdevices are becoming more commonplace,
separations for forensic DNA analysis have been less-frequently demonstrated. DNA separations
for forensic analyses, like DNA sequencing, require single base pair resolution and multicolor
fluorescence detection (see Chapters 6, 27, and 44). In addition, the STR separation must be rapid (less
than 30 min) to compete with the current capillary-based method. For nondisposable, single-process
devices, the sieving matrix must be of relatively low viscosity to enable loading and replacement of
the matrix after each separation (as in the capillary) to prevent carryover.
Table 37.2 provides some of the pertinent details associated with the STR separations carried out
on microdevices to date. All of these STR separations have utilized long-read polyacrylamide (LPA)
or polydimethylacrylamide (PDMA) [the main component of the commercially available polyolefin
plastomer polymers, known as “POP” polymers]; POP-4 (where the number dictates the %PDMA) is
the polymer used for capillary STR separations on ABI instruments. Acceptable polymers for these
separations must be capable of single base pair resolution at low temperatures (preferably, room
temperature) with short effective separation length (Leff – distance from injection point to detection
point) and of relatively low viscosity to enable replacement and ease of filling the microchannel.
Overall, it appears that LPA is capable of providing the requisite resolution with microchannels
having an Leff of 11.5 cm,44 although shorter distances are unlikely to be adequate with the current
methodology.45 In general, the separations require 25 or more minutes,46−48 which is comparable to
that obtained on the conventional capillary instrumentation. One embodiment46,48 , however, utilizes
a 96-microchannel device (Figure 37.4) and can accomplish 96 simultaneous separations, resulting
in greater throughput capabilities than the commonly encountered 4- and 16-capillary conventional
CE instruments. An example of data obtained on this microdevice is shown in Figure 37.5. POP-4
has also been used for the separation of STR fragments on microdevices, where the separation of
an Identifiler allelic ladder was accomplished with an Leff of only 8 cm and in under 12 min.49
This separation is distinct in that the optical requirements for the detection system are relatively
simple—containing only a laser and a single photomultiplier tube (PMT), because of the electronic
filtering capabilities of an inline acousto-optic tunable filter (AOTF) to select the desired emission
wavelengths sequentially (see Chapter 45 by Karlinsey). The value of the AOTF stems from the
ease with which four-color detection can be expanded to five or more colors, which may be opti-
mal for portable systems. One of the benefits of microchip electrophoresis reported by Goedecke
et al.47 (Figure 37.6) was the improved data quality and stability when compared to the commer-
cially available capillary array instruments. This advantage arises from the enhanced thermal and
mechanical stability of the separation channel, as well as the improved heat-sinking inherent to glass
microchips.
As mentioned earlier, devices similar to those described in Schmalzing et al.50 and Yeung et al.48
are currently undergoing evaluation in the United States and the United Kingdom. It is anticipated
TABLE 37.2
STR Separations on Microdevices
Separation
Channel Separation Resolution
Author Sample Microchip (Effective Sieving Channel Field STR Kit Separation of 9.3/10 (THO1) Detection
(Citation) Treatment Substrate Length, cm) Matrix Coating Strength Utilized Time (min) Demonstrated? Multichannel? (Color(s))
Schmalzing Diluted in Glass (fused 2.6 4% LPA in Modified 200 V/cm CTTv <2 No No 1
et al.52 buffer, silica) one time Hjerten
heat TBE/3.5 M
denatured, urea/30%
snap formamide
cooled
Schmalzing Diluted in Glass (fused 11.5 4% LPA in Modified 200 V/cm CTTv <10 Yes No 2
et al.50 H2 O, heat silica) one time Hjerten
denatured, TBE/3.5 M
snap urea/30%
cooled formamide,
50◦ C
Medintz Desalted Glass 5.5 Long-read Modified a (ET dyes) <8 Yes Yes, 96 4
et al.53 (Qiagen), (borofloat) LPA (Amer- Hjerten
resus- sham),
pended in 40◦ C
0.5 × TE,
diluted in
for-
mamide,
heat
denatured,
snap
cooled
Continued
1075
TABLE 37.2
1076
(Continued)
Separation
Channel Separation Resolution
Author Sample Microchip (Effective Sieving Channel Field STR Kit Separation of 9.3/10 (THO1) Detection
(Citation) Treatment Substrate Length, cm) Matrix Coating Strength Utilized Time (min) Demonstrated? Multichannel? (Color(s))
Mitnik Heat dena- Glass 11.5 4% LPA in Modified 180 V/cm PowerPlex <35 Yes No 4
et al.44 tured, (borosilicate) one time Hjerten 16
snap TTE/7 M
cooled urea, 50◦ C
Shi and Heat dena- Plastic 4.5 4% LPA in 2% 165 V/cm CTTv <10 Yes No 2
Anderson54 tured, (polyolefin) one time p(DMA/
snap TTE/7 M DEA)
Goedecke Combined, Glass 20 4% LPA in Modified 180 V/cm PowerPlex 40 Yes Yes, 16 4
et al.55 purified (borosilicate) one time Hjerten 16
(GFX TTE + 7 M
spin urea, 50◦ C
column),
resus-
pended/
diluted in
water,
heat
dena-
tured,
snap
cooled
Crouse Diluted in Glass 17–18 Long-read Modified b PowerPlex <25 Yes Yes, 96 4
et al.46 for- (borofloat) LPA Hjerten 16
mamide, (Amer-
heat sham)
dena-
tured,
snap
cooled
Shi56 Heat Plastic 10 3% LPA in 2% 150 V/cm CTTv and <18 Yes No 4
denatured, (polycycloolefin) one time p(DMA/ Profiler Plus
snap TTE/7 M DEA)
Yeung Diluted in Glass (borofloat) 15.9 Long-read Modified 150 V/cm PowerPlex <22 and Yes Yes, 96 4
et al.48 for- LPA, five Hjerten 16 and <17
mamide, times Profiler Plus
heat TTE, 67◦ C
denatured,
snap
cooled
Karlinsey Dilute in Glass (borofloat) 8 POP-4 Modified 250 V/cm Identifiler 18 No No 5
and for- (Applied Hjerten
Landers49 mamide, Biosys-
spin filter tems)
(Microcon
30), heat
denatured,
snap
cooled
a The separation channels are not straight and, therefore, the field strength is not uniform. Voltage was applied as follows: +1300 V at the anode reservoir, +200 V at the cathode reservoir, +325 V
at the sample reservoir, and +325 V at the waste reservoir.

b The separation channels are not straight and, therefore, the field strength is not uniform. Voltage was applied as follows: +2500 V at the anode reservoir, +0 V at the cathode reservoir, +180 V at
the sample reservoir, and +200 V at the waste reservoir.

1077
(a)
15 cm
(b)
Central
anode
Sample
Hyperturn
Common Common
waste cathode 0 5 mm
FIGURE 37.4 (a) Design for the 96-channel µCAE microdevice by Mathies and coworkers. The glass
microdevice is approximately the dimensions of a compact disk. The microchannel structure consists of 48
doublet structures, each containing two electrophoresis lanes that share common cathode and waste reservoirs.
(b) Magnified view of the doublet structure, highlighting the “hyperturn,” which prevents band-broadening dis-
persion as the DNA molecules traverse the turn, thereby enabling high-resolution DNA separations in a compact
device. (Adapted and reproduced from Yeung, S. H. et al., J Forensic Sci 2006, 51, 740–747. Copyright 2006.
With permission from Blackwell Publishing, Inc.)
that microdevices for the separation of STR products will appear in casework soon. Because little
advantage has been gained in processing time with these devices, as compared to the conventional
multiple-capillary CE instrumentation, there is little impetus for the forensic community to implement
this separation-only microdevice. While not used in forensic laboratories, there are commercial
96-capillary CE instruments capable of the same throughput as the 96-microchannel microcapillary
array electrophoresis (µCAE) device in a similar time frame. One would anticipate more widespread
adoption by forensic laboratories either following a reduction in the separation time or, better still,
when the integration of µPCR and DNA separation has been achieved in an effective and robust
manner. Liu et al.4 have demonstrated an instrument capable of µPCR of a mini-Y STR multiplex
and separation of the PCR product in under 64 min.
37.7 IMPACT
Advancements in forensic DNA analysis methodology have, historically, focused on alterations in
the basic science permitting individualization of biological material. In its origins, RFLP was used to
generate the DNA profile. However, the focus then shifted to the use of variable number of tandem
repeats (VNTRs) and, eventually, to STRs. It is the STRs that are currently being used to build
the national and international DNA databases (containing DNA of convicted offenders and from
unsolved cases) and are, therefore, unlikely to be supplanted in the near future. Microdevices do not
represent a shift in the basic science, but rather the technology platform on which the STR profiles
(a)
(b)
Time (min)
FIGURE 37.5 Electropherograms of (a) Promega PowerPlex 16 allelic ladder and (b) Applied Biosystems
AmpFlSTR Profiler Plus allelic ladder obtained on the µCAE microdevice shown in Figure 37.4. (Reproduced
from Yeung. S. H. et al., J Forensic Sci 2006, 51, 740–747. Copyright 2006. With permission from Blackwell
Publishing, Inc.)
are obtained much like the witnessed transitions from slab gels to the capillary format for the DNA
separations. While the effectiveness of this technology platform and its inherent ability to enhance
throughput might be considered revolutionary, it will not likely be viewed as a “paradigm shift” in the
same way that the introduction of PCR was to forensic DNA analysis. Instead, it may be viewed as a
necessary overhaul of the technology which, ultimately, will lead to better, more efficient processing
of DNA evidence. Because the underlying principles of the DNA analysis have not been changed,
merely the technological platform, its introduction, and acceptance by the courts should be relatively
straightforward. With similar validation studies as those completed when shifting from slab gels to
CE, their acceptance is not anticipated to be threatened.
As a result of the shift from the conventional methodology and technology to the microchip
format, it is possible that DNA testing in criminal laboratories will undergo a “make-over,” giving
it a very different look than it has today. While the criminalistics on the front-end (e.g., determining
which samples should be tested) will remain largely unchanged, there will be dramatically less analyst
Double-T
injector
Scan section Sample inlets
Anode
Sample waste Cathode
FIGURE 37.6 Microdevice for STR separations, capable of 16 simultaneous separations. The device is
constructed of two glass layers with interface boards at each end. The left interface board contains the anode
buffer reservoir (2 mL). The right interface board contains the cathode buffer reservoir (2 mL), 16 sample inlets
(70 µL), and 16 sample waste reservoirs (70 µL). The entire device measures approximately 5-cm wide and
25-cm long, and the separation channels have an effective length of 20 cm. (Adapted and reproduced from
Goedecke, N. et al., J Chromatogr A 2006, 1111, 206–213. Copyright 2006. With permission Elsevier.)
intervention in the analysis during the sample processing due to the automation. Thus, the job of the
analyst will necessarily shift from carrying out the entire analysis to only the criminalistics on the
front-end and the interpretation of the STR profile results produced by the instrument. While this
will increase the throughput of forensic DNA laboratories, it will, in most laboratories, not decrease
the number of analysts employed. Instead, the number of samples submitted to and processed by
laboratories is expected to increase. In this way, DNA testing in cases where it is not currently
routine (e.g., robbery) in all jurisdictions may become more commonplace. While a major advantage
of microfluidic devices in forensic analyses is the concomitant automation anticipated, automation
alone is not the only advantage of the microdevice, as robotic methods are also able to achieve
more automated analyses. Microdevices allow the distinct advantage of integrating multiple sample
processes (such as DNA extraction, quantification, PCR, and separation) on a single platform. This
allows for DNA analysis in a closed system and, therefore, less opportunity for contamination by the
analyst with tube transfers and sample handling. In addition, many microfluidic processes have been
shown to be more efficient than their macroscale counterparts such as DNA amplification, which
typically requires 2–3 h conventionally, but can be completed in an order-of-magnitude less time on
the microscale with concomitant decreases in sample and reagent consumption. Decreased sample
consumption in forensic casework, where there is often not abundant sample at the outset, is highly
valued. This might be expected to lead to the ability to obtain an STR profile where one cannot be
obtained with conventional methodology and will, likely, be a major driving force for the transition
to microfabricated technology in forensic laboratories.
In recent years, we have witnessed the incorporation of microdevices into the U.S. and U.K. crime
laboratories. Currently, only separation of PCR products is completed on these instruments, with
limited advantage over the current conventional methodology, CE. However, the major advantage of
microdevice technology is in the ability to incorporate multiple sample processing steps on a single
platform. Thus, the true value of the microchip in forensic laboratories will be realized when multiple
processes are integrated onto a single automated device, thus, decreasing analyst time as well as total
processing time and reagent consumption. Since microchip electrophoresis is the most mature of the
microchip processes described above, it is of no surprise that this would be the first process to be
introduced into the forensic laboratory. The more unseasoned microchip processes such as PCR and
DNA extraction are expected to follow suit in due time. The significant advantages of incorporating
these devices will, undoubtedly, drive their introduction into forensic laboratories.

The microfluidics community has successfully demonstrated microscale methods for cell sort-
ing/differential extraction, DNA extraction, PCR amplification, and electrophoretic separation of
STR products. Most of these individual microscale processes meet needs and desires of the forensics
community by demonstrating efficient extraction of DNA, rapid thermocycling, and single base pair
resolution (with multicolor detection) of STR fragments. However, quantitative PCR has yet to be
demonstrated on chip. Several examples of cell sorting (for use in sexual assault evidence) have
been demonstrated, although much work remains to show that efficiency and sensitivity of these
methods are comparable to the conventional differential extraction and, thus, merit implementation.
Most notably, many of these methods fall short addressing the macro-to-micro volume connection
where the macroscale sample size (mL) must be interfaced with the small volumes (nL to µL) of the
microscale environment. With forensic samples involving large volumes of solution that need to be
processed, volume reduction need to be invoked on chip in order to obtain sufficient DNA for STR
typing. Therefore, the front-end processes, such as cell sorting and DNA extraction, must be capable
of processing large sample volumes in order to compete with their macroscale counterparts. Devel-
opments like the ADE offer the potential to address this issue.51 In addition, with the maturity of the
microfluidics field, as with any other analytical development, one can anticipate further advances in
the automation of these processes. This automation, although not yet demonstrated for all processes,
provides a significant advantage over the conventional processes used in the financially strapped
forensic laboratories.
Also, the concept of single-process versus integrated (multiprocess) microdevices has been dis-
cussed. Single-process devices are designed with higher-throughput capabilities to process multiple
samples simultaneously. However, these devices are capable of only one step of the analysis pro-
cess. Integrated devices are typically lower-throughput but feature a totally integrated analysis in a
“sample-in-answer-out” fashion. For example, a microliter of blood could be introduced to the inlet
of the device and all sample processing steps occur on the device, resulting in a DNA profile detected
at the outlet. While both of these designs have promise in forensic DNA analysis, their adoption may
be highly dependent on the intricacies of the sample type and the application. For samples such as
those of convicted felons to be added to the national DNA databank, high-throughput single-process
devices may be the microchip of choice. These devices can be easily designed with throughput in
mind, such as the 96-channel microdevice for STR separations. Alternatively, casework samples
may be better suited for an integrated device that is all inclusive and would not require sample
handling after introduction of the sample to the inlet reservoir. Ultimately, the choice of microchip
format most appropriate in forensic laboratories will need to be driven by the forensics community,
as many caveats to DNA typing are not intuitive to the analytical chemists and engineers developing
microfluidic devices.
REFERENCES
1. Easley, C., Humphrey, J., Landers, J. P., J. Micromech. Microeng. 2007, 17, 1758–1766.
2. Hashimoto, M., Chen, P. C., Mitchell, M. W., Nikitopoulos, D. E., Soper, S. A., Murphy, M. C. Lab
Chip 2004, 4, 638–645.
Hughes, M. A., Hewlett, E. L., Merkel, T. J., Ferrance, J. P., Landers, J. P. Proc Natl Acad Sci USA
2006, 103, 19272–19277.
4. Liu, P., Seo, T. S., Beyor, N., Shin, K. J., Scherer, J. R., Mathies, R. A. Anal Chem 2007, 79, 1881–1889.
5. Bienvenue, J. M., Legendre, L. A., Ferrance, J. P., Landers, J. P. in preparation.
6. Gill, P., Jeffreys, A. J., Werrett, D. J. Nature 1985, 318, 577–579.

7. Duffy, D. C., Gillis, H. L., Lin, J., Sheppard, N. F., Kellogg, G. J. Anal Chem 1999, 71, 4669–4678.
8. Devitt, A. J., Aflatooni, N., Vinas, M., Loh, N., Pourahmadi, F., Yuan, R., and Northrup, M. A.
Proceedings of the International Symposium on Miniaturized Total Analysis Systems, Boston, MA,
USA, October 9, 2005, pp. 7–9.
9. Horsman, K. M., Barker, S. L., Ferrance, J. P., Forrest, K. A., Koen, K. A., Landers, J. P. Anal Chem
2005, 77, 742–749.
10. Horsman, K. M., Barker, S. L. R., Bienvenue, J. M., Voorhees, J. C., Blaiser, K., Schroeder, B., Koen,
K. A., Weingart, G., Landers, J. P., Ferrance, J. P. Special Publication—Royal Soc Chem 2004, 297,
288–290.
11. Horsman, K. M., Evander, M., Easley, C. J., Nilsson, J., Laurell, T., Landers, J. P. Micro Total Analysis
Systems 2006, Proceedings of the mTAS 2006 Symposium, 10th, Tokyo, Japan, 2006, 2, 1055–1057.
12. Blankenstein, G., Scampavia, L., Branebjerg, J., Larsen, U. D., Ruzicka, J. Proceedings of the 2nd
International Symposium on Miniaturized Total Analysis Systems µTAS96, Basel, Switzerland, 1996,
82–84.
13. Di Martino, D., Giuffre, G., Staiti, N., Simone, A., Le Donne, M., Saravo, L. Forensic Sci Int 2004,
146 Suppl, S151–S153.
14. Sanders, C. T., Sanchez, N., Ballantyne, J., Peterson, D. A. J Forensic Sci 2006, 51, 748–757.
15. Cady, N. C., Stelick, S., Batt, C. A. Biosens Bioelectron 2003, 19, 59–66.
16. Christel, L. A., Petersen, K., McMillan, W., Northrup, M. A. J Biomech Eng-Trans ASME 1999, 121,
22–27.
17. Breadmore, M. C., Wolfe, K. A., Arcibal, I. G., Leung, W. K., Dickson, D., Giordano, B. C., Power,
M. E., Ferrance, J. P., Feldman, S. H., Norris, P. M., Landers, J. P. Anal Chem 2003, 75, 1880–1886.
18. Bienvenue, J. M., Duncalf, N., Marchiarullo, D., Ferrance, J. P., Landers, J. P. J Forensic Sci 2006,
51, 266–273.
19. Chung, Y. C., Jan, M. S., Lin, Y. C., Lin, J. H., Cheng, W. C., Fan, C. Y. Lab Chip 2004, 4, 141–147.
20. Cao, W., Easley, C. J., Ferrance, J. P., Landers, J. P. Anal Chem 2006, 78, 7222–7228.
21. Wen, J., Guillo, C., Ferrance, J. P., Landers, J. P. Anal Chem 2006, 78, 1673–1681.
22. Wolfe, K. A., Breadmore, M. C., Ferrance, J. P., Power, M. E., Conroy, J. F., Norris, P. M., Landers,
J. P. Electrophoresis 2002, 23, 727–733.
23. Alonso, A., Martin, P., Albarran, C., Garcia, P., Garcia, O., de Simon, L. F., Garcia-Hirschfeld, J.,
Sancho, M., de La Rua, C., Fernandez-Piqueras, J. Forensic Sci Int 2004, 139, 141–149.
24. Andreasson, H., Gyllensten, U., Allen, M. Biotechniques 2002, 33, 402–404, 407–411.
25. Timken, M. D., Swango, K. L., Orrego, C., Buoncristiani, M. R. J Forensic Sci 2005, 50, 1044–1060.
26. Horsman, K. M., Hickey, J. A., Cotton, R. W., Landers, J. P., Maddox, L. O. J Forensic Sci 2006, 51,
758–765.
27. Nicklas, J. A., Buel, E. J Forensic Sci 2006, 51, 1005–1015.
28. Swango, K. L., Hudlow, W. R., Timken, M. D., Buoncristiani, M. R. Forensic Sci Int 2007, 170,
35–45.
29. Swango, K. L., Timken, M. D., Chong, M. D., Buoncristiani, M. R. Forensic Sci Int 2006, 158, 14–26.
30. Lin, Y. C., Li, M., Chung, M. T., Wu, C. Y., Young, K. C. Sensors Materials 2002, 14, 199–208.
31. Liu, J., Enzelberger, M., Quake, S. Electrophoresis 2002, 23, 1531–1536.
32. Unger, M. A., Chou, H. P., Thorsen, T., Scherer, A., Quake, S. R. Science 2000, 5463, 113–116.
33. Grover, W. H., Skelley, A. M., Liu, C. N., Lagally, E. T., Mathies, R. A. Sensors Actuat B—Chem
2003, 59, 315–323.
34. Karlinsey, J. M., Monahan, J., Marchiarullo, D. J., Ferrance, J. P., Landers, J. P. Anal Chem 2005, 77,
3637–3643.
35. Roper, M. G., Easley, C. J., Landers, J. P. Anal Chem 2005, 77, 3887–3893.
36. Lagally, E. T., Medintz, I., Mathies, R. A. Anal Chem 2001, 73, 565–570.
37. Legendre, L. L., Bienvenue, J. M., Ferrance, J. P., Landers, J. P. 15th International Symposium on
Human Identification, Phenix, AZ, USA, 2004.
38. Schmidt, U., Lutz-Bonengel, S., Weisser, H. J., Sanger, T., Pollak, S., Schon, U., Zacher, T., Mann,
W. Int J Legal Med 2006, 120, 42–48.
39. Lutz-Bonengel, S., Sanger, T., Heinrich, M., Schon, U., Schmidt, U. Int J Legal Med 2007, 121,
68–73.
40. Waters, L. C., Jacobson, S. C., Kroutchinina, N., Khandurina, J., Foote, R. S., Ramsey, J. M. Anal
Chem 1998, 70, 5172–5176.
41. Wu, Q., Bienvenue, J. M., Hassan, B. J., Kwok, Y. C., Giordano, B. C., Norris, P. M., Landers, J. P.,
Ferrance, J. P. Anal Chem 2006, 78, 5704–5710.
42. Mitnik, L., Carey, L., Burger, R., Desmarais, S., Koutny, L., Wernet, O., Matsudaira, P., Ehrlich, D.
Electrophoresis 2002, 23, 719–726.
43. Schmalzing, D., Adourian, A., Koutny, L., Ziaugra, L., Matsudaira, P., Ehrlich, D. Anal Chem 1998,
70, 2303–2310.
44. Crouse, C. A., Yeung, S., Greenspoon, S., McGuckian, A., Sikorsky, J., Ban, J., Mathies, R. Croat
Med J 2005, 46, 563–577.
45. Goedecke, N., McKenna, B., El-Difrawy, S., Gismondi, E., Swenson, A., Carey, L., Matsudaira, P.,
Ehrlich, D. J. J Chromatogr A 2006, 1111, 206–213.
46. Yeung, S. H., Greenspoon, S. A., McGuckian, A., Crouse, C. A., Emrich, C. A., Ban, J., Mathies, R. A.
J Forensic Sci 2006, 51, 740–747.
47. Karlinsey, J. M. Advances and Applications of Microfluidic Analysis Systems, University of Virginia,
Charlottesville, VA, 2007.
48. Schmalzing, D., Koutny, L., Chisholm, D., Adourian, A., Matsudaira, P., Ehrlich, D. Anal Biochem
1999, 270, 148–152.
49. Petersson, F., Nilsson, A., Holm, C., Jonsson, H., Laurell, T. Analyst 2004, 129, 938–943.
50. Schmalzing, D., Koutny, L., Adourian, A., Belgrader, P., Matsudaira, P., Ehrlich, D. Proc Natl Acad
Sci USA 1997, 94, 10273–10278.
51. Medintz, I. L., Berti, L., Emrich, C. A., Tom, J., Scherer, J. R., Mathies, R. A. Clin Chem 2001, 47,
1614–1621.
52. Shi, Y., Anderson, R. C. Electrophoresis 2003, 24, 3371–3377.
53. Goedecke, N., McKenna, B., El-Difrawy, S., Carey, L., Matsudaira, P., Ehrlich, D. Electrophoresis
2004, 25, 1678–1686.
54. Shi, Y. Electrophoresis 2006, 27, 3703–3711.
38 Taylor Dispersion in Sample
Preconcentration Methods
Rajiv Bharadwaj, David E. Huber, Tarun Khurana, and
Juan G. Santiago
CONTENTS
38.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1085

38.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1087
38.3.1 Taylor Dispersion Analysis via Area Averaging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1087
38.3.2 Scaling Relations to Estimate Dispersion Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1093
38.4.1 Field-Amplified Sample Stacking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1095
38.4.1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1095
38.4.1.2 Basic Theory and Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1096
38.4.1.3 Dispersion Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1097
38.4.1.4 Performance and Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1100
38.4.2 Temperature Gradient Focusing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1102
38.4.2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1102
38.4.3 Isotachophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1108
38.4.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1108
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1116
38.1 INTRODUCTION
Sample preconcentration methods, including stacking and focusing techniques, enable high-
sensitivity detection by increasing analyte ion concentration. Sample preconcentration also leads
to a decrease in the axial dimension of sample plugs (i.e., peak widths) and this can improve resolu-
tion in electrophoretic separations. In this chapter, we describe the basic principles and limitations
of electrokinetic sample stacking and focusing techniques. We highlight the importance of Tay-
lor dispersion in determining the efficiency of preconcentration methods. Sample preconcentration
techniques can be broadly classified into two types:
1. In an increase in sample ion concentration occurs due to a local decrease in the magnitude
of the drift velocity of an ion. For drift velocity vector field v̄d , stacking occurs when
1085
locally ∇ · v̄d < 0. These gradients are typically limited to a small region relative to
the length of a separation channel. Ions enter the region, are stacked, leave the region,
and subsequently disperse and decrease in concentration. After stacking, peak (spatial)
variance tends to grow linearly in time. Stacking includes field-amplified sample stacking
(FASS) and its derivative techniques.
2. In focusing, the condition ∇ · v̄d < 0 is typically satisfied throughout a large extent of
the channel and there is, additionally, a focal point, x̄foc (or focal region), for the solute.
There is a reference frame where the continuous vector field v̄d changes sign at x̄foc . In
this frame, the opposing signs of v̄d drive the initially widely distributed sample to x̄foc ,
where sample ions accumulate. For a finite amount of sample, sample peak width reaches
a steady-state value determined by the opposing effects of the v̄d gradient and dispersion.
Focusing includes isoelectric focusing (IEF) and thermal gradient focusing (TGF) where
x̄foc is fixed in space; and also includes isotachophoresis (ITP) where x̄foc propagates at a
wave speed set by the bulk flow and the ion mobilities in the system.
Leveraging heterogeneous buffer systems to effect changes in sample drift velocity is a com-
mon feature of electrophoretic preconcentration techniques. The sample and background electrolyte
(BGE) can differ in ionic strength and ionic makeup (e.g., ion mobility or valence). The theoretical
concentration enhancement of most of these techniques has been well known for many decades. For
example, the maximum concentration enhancement in FASS is predicted to be equal to the ratio
of the (relatively high) conductivity of the BGE to the (relatively low) conductivity of the sample
solution. This ratio determines the low-to-high analyte drift velocity ratio. In ITP, the theoretical drift
velocity ratio and associated prediction of concentration enhancement is a function of the leading
ion concentration and various ionic mobilities. Several numerical simulation tools that can predict
the concentration enhancement in various electrophoretic preconcentration techniques are available.
One fairly comprehensive and useful simulation tools is Simul, which is available free download
on the Web [1]. Simul solves one-dimensional multispecies electromigration–diffusion phenomena
(e.g., for electromigration in long thin tubes of constant area). The code includes ionization equi-
libria of weak electrolytes, calculation of local pH fields, and the dependence of ionic mobilities on
ionic strength and pH. Given initial conditions and ion information (e.g., fully ionized mobility, pKa
values, diffusivity), the code predicts the evolution of ionic species in the presence of an electric
field. Simul is, however, limited to one dimension and does not include flow-induced dispersion
caused by nonuniform electroosmosis or body forces due to conductivity gradients.
In practice, both molecular diffusion and advective dispersion (i.e., Taylor dispersion [2]) limit
the achievable concentration enhancement of preconcentration techniques. In some techniques, such
as TGF, an external-pressure-driven velocity field is integral to the focusing effect and dispersive
effects are inherent in the system. In other techniques such as FASS, external-pressure-driven flows
are minimal (e.g., due to unwanted hydrostatic pressure heads in end-channel reservoirs). How-
ever, despite efforts to suppress electroosmotic flow (EOF), electrophoresis in capillaries is typically
accompanied by at least trace EOF. The heterogenous nature of buffers creates gradients in EOF
mobility. These heterogeneities include ion density and ion mobilities and may also include tem-
perature gradients, valence (via chemical reactions), and perhaps permittivity and viscosity. More
importantly, variations in local ionic conductivity (and local permittivity) affect strong variations
in local electric field. Both these effects (gradients in EOF mobility and gradients in field) create
nonuniform EOF that leads to the generation of internal pressure gradients. Secondary flows gen-
erated by these pressure gradients tend to disperse analyte plugs and reduce the preconcentration
effect.
In this chapter, we will identify the role and the scaling of dispersive effects in determining
the ultimate limits of sample preconcentration techniques. Where possible, we will offer simple
quantitative theory for estimating the effects of dispersion, and in all cases we will provide scaling
arguments that can help guide empirical optimization of preconcentration techniques. First, we will
Taylor Dispersion in Sample Preconcentration Methods 1087
explain the basic physics of Taylor dispersion and provide reduced order models to qualitatively
and quantitatively describe it. Second, we will present simple relations to estimate the importance
of Taylor dispersion in relation to other peak broadening mechanisms in capillary electrophoresis
(CE) systems. These relations are useful in, for example, empirically identifying optima in applied
field and conductivity ratios. Finally, we will briefly describe three examples of preconcentration
techniques as instructive “case studies” in which dispersion plays an important role in determining
concentration enhancement.
38.2 BACKGROUND
Dispersion, the tendency for ordered molecules to decrease gradients and local concentration, is
caused by both molecular diffusion and nonuniform bulk liquid motion. High dispersion rates may be
advantageous for mixing and chemical reactions, but are undesirable in separation and purification
applications. For separations, minimizing dispersion improves resolution and sensitivity [3] and
yields improved dynamics for concentration and purification [4]. As a consequence, the physical
processes that lead to dispersion have been a subject of intense interest for more than a century. In
recent years, the development of the concept of the micro-total analysis system (µTAS) or “labs on
a chip” has motivated further exploration of dispersion in microchannel flows.
In the field of CE, researchers are most familiar with dispersion caused by molecular diffusion
and Taylor dispersion [2]. In 1953, G. I. Taylor demonstrated that, under certain conditions, the cross-
sectional average of the unsteady, three-dimensional concentration field within a channel evolves as a
one-dimensional convective–diffusion equation with a coefficient analogous to a modified diffusivity.
Here, the advective dispersion in the axial direction is balanced by spanwise and depthwise (e.g.,
radial in a cylindrical tube) diffusion, causing peak variance to increase linearly in time, with the
characteristic slope determined by an effective dispersion coefficient. Taylor further demonstrated
that given enough time, theoretically, all solute plugs flowing within a channel ultimately reach this
limit. Subsequently, Aris [5] proved that the (now named) Taylor regime could be unified with the pure
diffusive regime by using an effective dispersion coefficient, which was the sum of the molecular
diffusivity and the Taylor dispersion coefficient. The effective dispersion concept has proved to
be extremely useful and has been extended to other geometries [6], generalized using alternative
analyses [7,8], and extended to apply to EOF [9] and electrophoresis in nanochannels [10,11].

The nonuniform velocity distribution over the cross section of a capillary or a microchannel leads
to distortion of the sample plug. This velocity-induced distortion generates a radial concentration
gradient of the analyte. Molecular diffusion acts to reduce this radial concentration gradient. The two
effects in tandem determine the effective dispersion rate of the sample plug. The increase in variance
can be quantified in terms of an effective dispersion coefficient, Deff . In this section, we will describe
the details of calculating this effective diffusion coefficient for a sample velocity profile. This section
is targeted at readers who may want to understand the mathematical details behind calculation of
Deff . Practitioners who are not interested in these details may skip to the next section without loss
of continuity. The next section describes simple relations to estimate the role of dispersion in peak
broadening in CE.
38.3.1 TAYLOR DISPERSION ANALYSIS VIA AREA AVERAGING

We derive the Taylor dispersion equation using an area averaging approach as described by Stone
and Brenner [8]. We consider axisymmetric flow in a cylindrical capillary of radius “a,” but will
later discuss the application of these principles to other cross sections (such as typical wet etched
“D” profiles). We consider a nonuniform (in x) drift velocity to account for electrophoresis, and
the scaling of streamwise coordinates by a characteristic length Up tc , where Up is the magnitude of
the deviation of axial velocity from the area-averaged velocity and tc is the characteristic time for
stacking or focusing. This approach can be applied to stacking and focusing. The following equations
describe the convective dispersion of a charged species in a parabolic velocity field:

∂Ci ∂ 1 ∂ ∂Ci ∂ 2 Ci
+ (uti (x, r) Ci ) = Di r + ; i = 1 : N,
∂t ∂x r ∂r ∂r ∂x 2
∂Ci
=0 at r = a, 0, (38.1)
∂r
where Ci is the concentration of a dilute and, in general, charged solute of interest. In general
electrokinetic flow physics, applied electric fields couple with conductivity gradients and lead to the
generation of net charge in the bulk liquid. However, this charge density is negligible compared to
the total background ion concentration. We can, therefore, assume that the solution is approximately
electrically neutral everywhere outside of the electrical double layer, so that N i=1 zi Ci ≈ 0, where
zi is the valence. This approximation is discussed in detail elsewhere [12–14].
In Equation 38.1, we have made the typical “nearly parallel flow” approximation common in
lubrication analysis that the flow is nearly parallel to the axis in this long thin tube [15]. This approx-
imation holds as long as axial gradients in electroosmotic mobility and conductivity scale is 1/σ ,
where the characteristic plug width of the solute, σ , is such that σ a. This condition is commonly
met in sample stacking and focusing techniques, although it may not be met in, for example, high
field regimes of ITP in the so-called peak mode. This and other limitations are discussed later.
We consider the velocity field bounded by a cylindrical slip surface, which excludes a thin
electric double layer (EDL) [16]. The velocity scale ut includes the effects of local pressure gradients,
electroosmosis, and electrophoresis and can be expressed as follows:
uti (x, r, t) ∼
= up (x, r, t) + Ueph,i (x, t) + Ueof (x, t), (38.2)
where up (x, r, t) is the axial velocity component due to a local pressure gradient, ∂p(x, t)/∂x, and
equal to −a2 (∂p/∂x)(1 − r 2 /a2 )/4η, where η is dynamic viscosity.
Ueph,i (x, t) is an axial electrophoretic drift due to an approximately purely axial electric field,
E(x, t), and can be written as zi Fµeph,i E(x, t), where zi , µeph,i , and F are the valence, electrophoretic
mobility, and Faraday’s constant [12]. Ueof (x, t) is the (exactly axial) velocity component due to
electroosmosis, which can be expressed as zFµeof E(x, t), where µeof is electroosmotic mobility
and E(x, t) is the axial field component at the slip plane (near the wall). We also consider changes
slow enough such that fluid inertia and charge relaxation are negligible. See Storey et al. [17] and
Lin et al. [18] for more detailed discussion of quasi-steady electromigration, electroosmosis, and
pressure-driven flow in long thin channels.
We define Ue (x, t) = Ueph,i (x, t) + Ueof (x, t) as the sum of the electroosmotic and approximately
parallel electrophoretic velocities. Note that ∇ ·(Ueof + up ) = ∂(Ueof +up )/∂x = 0. The convective
diffusion equation then becomes

∂Ci ∂Ci ∂Ueph,i 1 ∂ ∂Ci ∂ 2 Ci
+ (Ue,i + up ) + Ci = Di r + . (38.3)
∂t ∂x ∂x r ∂r ∂r ∂x 2
The dependent variables are expressed in terms of cross-sectional averages and deviations:
Ci (x, r, t) =
Ci (x, t) + Ci (x, r, t),
(38.4)
u(x, r, t) =
up (x, t) +
Ue,i (x, t) + up (x, r, t) = Up + Ue,i + up .
a
The cross-sectional average is defined as
· · · = 1/π a2 0 2π r(· · · )dr. Note up is Up (1 − 2r 2 /a2 ),
where Up = −π a4 (∂p/∂x)/8η is the bulk (area-averaged) velocity due to pressure gradients. We
can interpret primed quantities as describing the concentration field in a frame moving with the
area-averaged velocity of the solute. Substituting these definitions into Equation 38.1:
∂
Ci ∂Ci ∂
Ci ∂C ∂
Ci ∂Ci
+ + (Ue,i +
up ) + (Ue,i +
up ) i + up + up
∂t ∂t ∂x ∂x ∂x ∂x

∂C
∂Ueph,i ∂Ueph,i 1 ∂ ∂
Ci ∂ Ci
2 2
+
Ci + Ci = Di r i + + . (38.5)
∂x ∂x r ∂r ∂r ∂x 2 ∂x 2
Subject to
∂Ci
= 0 at r = a, 0.
∂r

Note that
Ueph,i = Ueph,i and Ueph,i = 0. The boundary conditions (BC) reduce to this since
∂
Ci /∂r is exactly zero by definition,
Ci being only a function of x. Next, we take a cross-sectional
average of Equation 38.5:

∂
Ci ∂Ci ∂
Ci ∂Ci ∂
Ci
+ +(Ue,i +
up ) + (Ue,i +
up ) + up
∂t ∂t ∂x ∂x ∂x

= 0 by def = 0 by def = 0 by def
 


 1 ∂
2C 
∂Ci ∂Ueph,i ∂Ueph,i ∂C ∂ ∂
2
C
Ci = Di  ,
i
+ up +
Ci +  r i
+ + i
∂x ∂x ∂x  r ∂r ∂r ∂x 2 ∂x 
2


= 0 by def evaluate using BC = 0 by def
(38.6)
where the notes “= 0 by def” and “evaluate using BC” denote a quantity zero by definition and a
term that can be evaluated with the BCs. Evaluating the first term on the right-hand side:

1 ∂ ∂C 1 a ∂ ∂C 2 a ∂C ∂Ci ∂Ci
r i = 2π r i dr = 2 d r i =r −r = 0.
r ∂r ∂r πa2 0 ∂r ∂r a 0 ∂r ∂r a ∂r 0
(From BC, ∂Ci /∂r is zero at both r = 0 and r = a.) Collecting nonzero terms in Equation 38.6:

∂
Ci ∂
Ci ∂Ueph,i ∂ 2
Ci ∂Ci
+ Ue,i + up +
Ci = Di − up . (38.7)
∂t ∂x 2 ∂x ∂x 2 ∂x
The cross-correlation term is placed on the right-hand side as it acts as a source of dispersion (as does
diffusion). We now derive an expression for Ci . To do so, subtract Equation 38.7 from Equation 38.5:
∂Ci ∂
Ci ∂Ci ∂Ci ∂Ueph,i
+ up + Ue,i + up + up + Ci
∂t ∂x ∂x ∂x ∂x

∂Ci ∂ 2 Ci
1 ∂ ∂Ci
= Di r + + up . (38.8)
r ∂r ∂r ∂x 2 ∂x
Scale the variables as follows: r ∗ = r/a, x ∗ = x/σ , Ci∗ = Ci /Cio

,
C ∗ =
C /C , t ∗ = t/t , and
i i io c
∗ ∗ ∗
up =
up /Up = 1, up = up /Up , and
Ue,i =
Ue,i /(Ueph,i (x, t) + Ueof (x, t)) = 1. We choose
both Cio and C as we will use the ratio C /C as a smallness parameter in our scaling arguments.
io io io
We scale axial gradients using the characteristic width of dispersion σ . For techniques where
variance width of the sample plug increases approximately linearly in time, we can interpret σ as
the characteristic length scale of advective dispersion, Up tc . Here, tc is the time of observation or
time between injection and detection. In focusing techniques where solute dispersion length scales
reach a steady value, σ is an inherent length scale associated with the focusing stacking dynamics.
For example, in ITP stacking or TGF of finite injection volumes, sample widths eventually reach
a steady-state value. In these techniques, the axial concentration and velocity gradients of the peak
are determined by the interplay between electromigration (which provides the focusing fluxes) and
diffusion and dispersion (which limit the amount of stacking). The inherent length scale of dispersion,
σ , is then an internal length scale (i.e., determined by the specific condition). For such cases, we
will assume σ a with the understanding that the analysis is checked for self consistency and is
only valid if it indeed predicts long thin solute plugs.
Equation 38.8 can then be written as

Co,i ∂Ci∗ Up Co,i ∗ ∂
Ci
∗ Co,i ∗ ∂Ci
∗ Up Co,i ∂Ueph,i
+ up + (Up + Ue,i + Up up ) ∗ + Ci
tc ∂t ∗ σ ∂x ∗ σ ∂x σ ∂x ∗

∂C ∗
∂ 2C∗
∂C
∗
Di Co,i 1 ∂ D i C U p C
r∗ ∗ up∗ ∗ .
o,i o,i
= i
+ i
+ i
(38.9)
a2 r ∗ ∂r ∗ ∂r σ2 ∂x ∗2 σ ∂x
Next, multiply this equation by a2 /(Di Co,i ).

2
∗

a2 Co,i ∂Ci∗ a 2 Up ∗ ∂
Ci
∗ a Up Co,i Ue,i ∗ ∂Ci

a2 UpCo.i
+ up + 1+ + up +
Di tc Co,i ∂t ∗ σ Di ∂x ∗ σ Di Co,i Up ∂x ∗ σ Di Co,i

order ε2 order ε order ε2 order ε2

∗

∂Ueph,i Co,i 1 ∂ ∂Ci a2 Co,i ∂ 2 Ci∗ a2 U p Co,i ∂Ci∗
× Ci = r∗ + + up∗ .
∂x Co,i r ∗ ∂r ∗ ∂r ∗ σ Di Co,i ∂x ∗2 σ Di Co,i ∂x ∗

order ε order ε2 order ε2
(38.10)
The notes below the various terms denote the order of magnitude. Of interest are the long times
relative to streamwise transport so that tc a2 /Di . In other words, the plug is long compared to the
radius of the capillary such that σ a and the smallness parameter, ε, is
a 2 a2
ε∼ ∼ 1.
σ Di tc
For the third term on the left-hand side, we assume that the displacement time, σ/(Up + Ue,i (x)), is
much longer than the radial diffusion time, a2 /Di . For TGF, Up +Ue,i (x) = 0 and the term in question
is zero. For FASS and ITP, the solute plug velocity, Up + Ue,i , is finite and this assumption implies
the diffusion time, a2 /Di , must be smaller than the time required for the plug to move a characteristic
distance σ . This condition should hold for FASS where typically σ is significantly larger than a.
For ITP, we will restrict our analysis to the cases where this third term is negligible. In comparing
the second and third terms on the left-hand side, we see this assumption allows [Up + Ue,i ]/Up to
be somewhat larger than unity (or smaller), but not so large that [Up + Ue,i ]Co,i /(U C ) is order
p o,i
unity. This seems reasonable for ITP plugs with interface axial lengths, σ , of order a and larger.
(The challenges of analyzing Taylor dispersion in ITP are discussed further below.) The current
assumptions lead to the following:

∗ ∗
a 2 Up ∗ ∂
Ci 1 ∂ ∗ ∂Ci
Co,i
up ≈ r . (38.11)
σ Di ∂x ∗ Co,i r ∗ ∂r ∗ ∂r ∗
Or, in dimensional form:

∂
Ci 1 ∂ ∂Ci
up ≈ Di r . (38.12)
∂x r ∂r ∂r
Perturbations in the axial velocity must be balanced by radial diffusion. Substituting for up

2r 2 ∂
Ci 1 ∂ ∂C
Up 1− 2 ≈ Di r i .
a ∂x r ∂r ∂r
Multiply both sides by r and integrate in r

∂
Ci r2 r4 ∂C
Up − 2 ≈ Di r i + C1 .
∂x 2 2a ∂r
Apply BC at r = a: C1 = 0. Divide both sides by r (and D) and integrate again:

Up (x, t) a2 ∂
Ci r 2 r4
Ci (r, x, t) = Ci (0, x, t) + − 4 . (38.13)
4Di ∂x a 2 2a
where Ci (0, x, t) is the constant of integration obtained from the BC at r = 0. Now, we can evaluate
the cross term in Equation 38.7. Take the x-derivative of Equation 38.13,

∂Ci (r, x, t) ∂Ci (0, x, t) ∂ ∂
Ci a2 r 2 r4
= + Up − .
∂x ∂x ∂x ∂x 4Di a2 2a4
multiply this by u (from Equation 38.4), and integrate over the cross section:

2 2
∂Ci ∂ a Up ∂
Ci
u =− . (38.14)
∂x ∂x 48Di ∂x
Substitute this into Equation 38.7, which was the area average of the originally decomposed
convective diffusion equation, to yield
2
2 2
∂
Ci ∂
Ci ∂Ueph,i ∂
Ci ∂ a Up ∂
Ci
+ (Ue,i +
up ) +
Ci = Di + . (38.15)
∂t ∂x ∂x ∂x 2 ∂x 48Di ∂x
To review, the main assumptions are that the velocity and electric fields are nearly parallel and that
a2 /(Di tobs ) 1 and a/σ 1.
We rewrite Equation 38.15 as

∂
Ci ∂
Ci ∂Ueph,i ∂ ∂
Ci
+ (Ueph,i + Ueof + Up ) +
Ci = Deff ,i , (38.16)
∂t ∂x ∂x ∂x ∂x
where Deff ,i (x, t) = Di + a2 Up (x, t)2 /(48Di ) = Di (1 + Pe2i /48). Here Pei is a Peclet number
defined as Up a/Di . We do not know of analytical solutions to the nontrivial cases where Ueph,i ,
Ueof , and Up indeed vary in x (as in the case of nonuniform EOF velocities), and so this requires
numerical solutions. However, we argue that significant insight can be gleaned from this equation,
and this insight is valuable in optimizing sample preconcentration processes.
The previously mentioned equation shows that the area-averaged solute plugs will travel along
the capillary with a nonuniform wave velocity of the form Ueph,i + Ueof + Up , as expected. The
local dispersion coefficient, Deff ,i , is strictly a function of the pressure-driven flow component,
which has nonzero radial gradients, and has the same form as the well-known Taylor–Aris solution
[5], but with axial and temporal variations in the dispersion coefficient. The focusing effect of
electrophoresis is captured by the electrophoresis term
Ci ∂Ueph,i /∂x. This analysis captures the
interplay between dispersion and preconcentration forces in electrophoretic stacking methods. Deff ,i
acts to increase sample plug variance (in the absence of stacking fluxes, variance scales as Deff ,i t,
where t is time). This dispersion is countered by gradients of ion drif velocity that generate a
negative value of
Ci ∂Ueph,i /∂x. For a positive x-direction electric field, preconcentration (local
increase in
Ci ) occurs for a cation when Ueff ,i has negative axial gradient, ∂Ueph,i /∂x < 0 (and
where ∂Ueph,i /∂x > 0 for an anion). Opposite signs for ∂Ueph,i /∂x cause so-called electromigration
dispersion and associated reductions in
Ci [19]. As described earlier, focusing occurs when we have
a frame of reference in which ∂Ueph,i /∂x is accompanied by a change in sign of the electrophoretic
velocity, and solute is driven toward this focal point from large regions of the channel. [Ueof and
Up cannot cause preconcentration since ∇ · (Ueof + Up ) is identically zero.] The relative strength of
stacking and focusing versus dispersion is determined by the relative strength of ∂Ueph,i /∂x and Deff ,i ;
more simply, ∂Ueph,i /∂x acts to reduce σ while Deff ,i acts to increase it. This balance determines
the maximum achievable concentration increase, and can determine the resolution of simultaneous
preconcentration and separation. This concept is further developed in the next section.
As mentioned earlier, Taylor-type dispersion analyses have been extended to other geometries
including flow in rectangular channels of varying aspect ratio and channels with the characteristic
“D” shape of isotropic chemical etching [6,20]. In general, these analyses lead to an effective
dispersion coefficient of the form D(1 + κPe2 ), where κ takes into account geometric dependences.
Table 38.1 list sample κ values for several flow geometries. This elegant and general form of the
dispersion coefficient implies that the insights gained by the area-averaging are generally applicable
to dispersion in a wide range of geometries.
Often, microchannels are etched with finite (and sometimes high) width-to-depth aspect ratios,
w/h, with typical values ranging from 2 to 20. For such channels, the time scale for diffusion across
TABLE 38.1
Taylor Dispersion κ Values for Various Cross-sectional
Geometries
Cross-Sectional
Geometry κ
!
1 24 − 24e2 + 5e4
Ellipse ; e= 1 − d 2 /W 2 ;
192 24 − 12e2
where d and W are the minor and major axes, respectively

1 8.5W 2
Rectangle ;
210 d 2 + 2.4dW + W 2
where d and W are the channel depth and width, respectively
1
Cylinder
48
the channel depth,τh ∼ h2 /Di is short compared to across the channel width τw ∼ w2 /Di . Ajdari
et al. [20] points out that dispersion in shallow-channels with smooth spanwise height distributions is
controlled by the product κPe2w , where Pew = wU/Di . In wide, shallow channels dispersion, owing
to spanwise (width direction) velocity gradients, occurs at rates that are not negligible compared to
spanwise diffusion. Statistical sampling of solute molecules along the spanwise direction is, therefore,
less efficient and leads to increased dispersion over that of an idealized, infinitely wide channel with
the same depth. A key consequence of the analysis is that the largest cross-section dimension controls
the time scale to reach the Taylor dispersion limit. For arbitrarily shaped channels, the criteria for
achieving Taylor dispersion are then modified to tc w2 /Di (where w is the largest cross section
scale) and σ a, where again σ is the characteristic axial dimension of the solute (or interface
region) of interest.
38.3.2 SCALING RELATIONS TO ESTIMATE DISPERSION EFFECTS

In the presence of axial gradients in ionic strength (and conductivity) or pH, both electric field and
electroosmotic mobility will vary along the axis of the capillary or microchannel [21]. As discussed
earlier, this leads to the generation of internal pressure gradients to satisfy the continuity equation
(i.e., mass conservation). In the absence of an external, applied pressure difference, up is generated
strictly by gradients in EOF mobility Ueof = µeof (x)E(x). For this case, we can estimate the relative
importance of Taylor dispersion to molecular diffusion by performing a simple scaling analysis.
This exercise highlights the importance of judicious choices of system parameters to minimize
dispersion. These choices include separation voltage, EOF suppression strategy, and channel shape
and dimensions.
We present an example scaling analysis for dispersion and optimum electric field in a single-
interface FASS problem. (The scaling of pressure velocity magnitudes versus local electric field
and electroosmotic mobility will also hold for other single-interface preconcentration methods such
as the ITP method demonstrated by Jung et al. [22]). This simple case yields approximations as
with closed-form analytical expressions. Obtaining estimates for the electroosmotic velocity is the
first step. This can be achieved in at least two ways: direct experimental measurement (e.g., in
calibration experiments) and theoretical estimates. EOF velocities can be measured under various
conditions using the current monitoring method [23], micron-resolution particle image velocimetry
measurements [24], or neutral marker tracking [25]. If these measurements are difficult to obtain,
EOF velocity can be estimated using various models for the electrical double layer. For thin EDLs,
the Helmholtz–Smoluchowski model yields a simple relation [26]:
εζ E
uEOF = − , (38.17)
η
where ε is the permittivity constant and ζ is the zeta potential [27]. The zeta potential is a function of
ionic strength and pH of the electrolyte. Various models and correlations are available in the literature
to estimate the zeta potential for various substrates [28–30]. In the presence of axial gradients in
conductivity, electric field will also vary along the axis of the channel. Electric field variation can be
estimated by invoking conservation of current and Gauss’s law. For example, consider the two-zone
preconcentration problem in a cylindrical tube as depicted in Figure 38.1. After negligible charge
relaxation time scales [13], the following relations hold true [31]:
σe1 E1 ∼
= σe2 E2 ,
(38.18)
E1 xL + E2 (1 − x)L = V ,
where σ e refers to electrical conductivity and x is the fraction of channel filled with σe1 . These
equations can be solved simultaneously to obtain E1 and E2 . Upon deriving the estimates for the
1,1, E1 2,2, E2
x /L (1−x)L
FIGURE 38.1 Schematic representation of a capillary with axial gradient in conductivity, zeta potential, and
electric field. The interface is located at distance x/L from the end of the channel (i.e., the applied potential),
where x is the fraction of the length L occupied by the liquid 1. The potential drop across the capillary is V .
local electric fields and zeta potentials, the EOF mismatch-induced pressure-driven component of
the velocity field can be obtained. To do so, we invoke the continuity equation:
UEOF,1 + Up1 = UEOF,2 + Up2 , (38.19)
where Up1 and Up2 are related by following:
Up1 (1 − x)
=− . (38.20)
Up2 x
Finally, the maximum pressure-driven component of the velocity field is (assuming x < 0.5):
Up1 = (UEOF,2 − UEOF,1 )(1 − x). (38.21)
Burgi and Chien [31] have derived similar relations for estimating internal pressure gradients and
for predicting optimum conductivity ratio for FASS. Now, the dispersion coefficient is estimated:
2
a2 Up1 a2
Deff = D + =D+ (UEOF,2 − UEOF,1 )2 (1 − x)2 . (38.22)
48D 48D
Furthermore,
(ζ1 − γ ζ2 )
UEOF,2 − UEOF,1 = ε E1 , (38.23)
η
where γ = σ 1 /σ2 . Noting E1 = E0 /(x + γ (1 − x)), where E0 = V /L is the nominal electric field,
we have
a2 ε 2 E02 (1 − x)2
Deff = D + (ζ 1 − γ ζ 2 ) . (38.24)
48Dη2 x + γ (1 − x)
We now assume that we are concerned with a stacking technique such as FASS where variance
increases in time. For such a technique, the peak variance due to molecular diffusion and Taylor
dispersion scales is
σtotal
2
∼ 2Deff τobs . (38.25)
Note that Deff varies in space and time as per equation (38.24). However, for scaling purposes the
maximum dispersion coefficient can be used. The observation time, τobs , is a complex quantity to
determine, depending on the electrophoretic, electroosmotic, and induced pressure-driven velocities.
A closed form solution may be found in some simple scenarios, but in all cases the time scales linearly
with E0 , such that τobs ∼ L/µeff E0 , where µeff is an effective species mobility. Thus,

a2 ε 2 2 (1 − x)
2 L L 1
σtotal
2
∼ (ζ1 − γ ζ2 ) E0 + 2D . (38.26)
24Dη2 x + γ (1 − x) µeff µeff E0
In this relation, the first term on the right-hand side is the contribution of Taylor dispersion to the
overall band broadening. The second is due to molecular diffusion. A key point is that the contribution
to the variance due to Taylor dispersion increases with nominal electric field, E0 , whereas variance
due to molecular diffusion decreases with electric field, 1/E0 . This is a consequence of the Up2
and therefore, E02 scaling of Deff on the one hand, and the L/u and therefore, 1/E0 scaling of
time. Therefore, there exists an optimum electric field that will minimize the peak broadening.
This optimum electric field can be derived by minimizing the variance with respect to field to
obtain
"
48D2 η2 (x + γ (1 − x))
Eoptimum ∼ . (38.27)
a2 ε 2 (ζ1 − γ ζ2 )2 (1 − x)2
As we discuss later in this chapter, techniques achieving a steady-state dispersion-limited width

typically also have optimum electric fields that reduce dispersion. At negligible Taylor dispersion
conditions (i.e., low electric fields), the sample axial dimension is determined by a balance between
electrophoretic focusing fluxes and molecular diffusion. Plug axial dimension, therefore, scales as
D/E0 (see Equations 38.40, 38.41, 38.44, and associated discussions). At higher electric fields, Taylor
dispersion becomes important and axial dimensions scale as Deff /E0 . At sufficiently high fields, the
Taylor dispersion term, which scales as a2 ε 2 E02 /(48Dη2 ), dominates and then interface width again
scales as E0 .
In all cases, the optimum field for minimization of Taylor dispersion is a function of sample
(through D), channel geometry, solution viscosity, and zeta potential value for a given conductivity
ratio and solution permittivity. Finally, we note the earlier analysis neglects the effects of Joule
heating [32,33] on dispersion. Joule heating places additional constraints on the optimum field as
temperature rise is a function of both conductivity distribution and pore diameter. We, however, stress
the main purpose of the analysis is to argue that all preconcentration phenomena have optimum fields
and that they are often limited by the dispersion associated with even residual EOF.

38.4.1 FIELD-AMPLIFIED SAMPLE STACKING
Field-amplified sample stacking is a fairly widely applicable method of achieving increased sen-
sitivity for capillary and on-chip assays in a scheme that is easily integrated with electrophoretic
separation techniques [4,34–43]. FASS is typically used as a preconcentration step that occurs before
the electrophoretic separation of analyte ions.
The transport phenomena associated with FASS (as in all preconcentration methods) are, in
general, a complex coupling of convective–diffusion, electrostatics, and electrokinetics along with
the unsteady effects associated with the response of the electrical double layer to varying bulk ion
concentrations. The detailed understanding of the process dynamics is important for optimization of
high-sensitivity systems. The effects of EOF on preconcentration and separation are very important
to studies of FASS as even slight EOF couples with axial conductivity gradients to generate internal
(a) (b)
UEOF, BGE UEOF, S
t=0
–
- – –
-
–
- EBGE ES
+ –
- -– –
-–
High-conductivity Low-conductivity High-conductivity Low-conductivity

buffer sample buffer sample
t>0
-– –- -–
+ -– –- –
–
- –
-
Stacked analyte Favorable pressure Adverse pressure
FIGURE 38.2 (a) Schematic diagram showing FASS of anionic species in the absence of EOF. A gradient
in the BGE ion concentration is established. The sample is in a region of locally low conductivity. Upon
application of an electric field, the axial gradient in conductivity results in an electric field gradient. Since
area-averaged current density is uniform along the axis of the channel, the low conductivity section is a region
of high electric field, and the region of high conductivity contains relatively low electric field. As sample ions
exit the high field/high electrophoretic velocity region and enter the low velocity region, they locally accumulate
and increase in concentration. (b) Stacking in the presence of EOF. Gradients in conductivity generate axial
variation in electric field and electroosmotic mobility. The system generates internal pressure gradients, which
tend to disperse the sample.
pressure gradients. These internal pressure gradients disperse the sample and thereby, limit the
practically achievable concentration enhancement.
38.4.1.2 Basic Theory and Implementation

The principle behind FASS is shown schematically in Figure 38.2a. An axial gradient in ionic
conductivity (and therefore electric field) is achieved by preparing the sample in an electrolyte
solution of lower concentration than the BGE. Upon application of an axial potential gradient, the
sample region acts as a high electrical resistance zone in series with the rest of the channel and a
locally high electric field is generated within the sample zone. Under the influence of electric field,
sample ions migrate from the high to low drift velocity region. This leads to a local accumulation
or “stacking” of sample ions near the interface between high and low conductivity regions. This
stacking increases sample concentration and results in an increased signal. The process depicted in
Figure 38.2a is for an idealized case where diffusion and advection are neglected. The maximum
concentration enhancement is given by a conservation of species at the interface
CStacked ES
= = γ, (38.28)
CInitial EBGE
where ES and EBGE are the electric field in the sample and the BGE regions, respectively. In essence,
FASS relies strictly on electric field gradients generated by spatial variations in ion density. This
makes FASS unlike other preconcentration techniques (such as ITP and IEF), which require more
specific buffer chemistries and ion mobilities. For example, FASS can be implemented with the same
type of ions in the BGE and sample zones. As such, FASS provides flexibility in the choice of BGE
buffer chemistry and can nearly always be performed under well-controlled pH conditions.
Figure 38.2b shows a more realistic system where finite EOF is present. The gradient in the
electrolyte concentration required for stacking leads to a gradient in electric field and electroosmotic
mobility. This causes a mismatch of electroosmotic velocity and hence generation of a pressure
(a) (b)
N Waste Vacuum
Low-conductivity
Buffer sample
t=0 t = 0.38 s
50 µm
Buffer
V t = 0.76 s t = 1.14 s
50 100 150 200 250
FIGURE 38.3 Single interface stacking experiments. (a) Schematic diagram of a microchip. Width and
centerline depth of channels were 50 µm and 20 µm, respectively. (b) Epifluorescence CCD images showing
establishment of initial condition for conductivity gradient and subsequent stacking across interface. The sample
was anionic 17 µm bodipy dye and the buffer was HEPES at pH = 7.
gradient (consistent with the continuity constraint). The pressure gradient tends to disperse the
concentration fields and thereby lower the efficiency of stacking.
FASS has been applied in a variety of assay formats in both capillaries and microchips. The
configurations can be broadly classified as (1) single interface configuration and (2) finite-plug
configuration. The single interface configuration includes techniques such as field-amplified sample
injection (FASI) [36,37,44], large volume sample stacking (LVSS) [38], and Head-Column FASS
[45]. These techniques involve a single interface between the sample region and BGE. In these
techniques, the volume of sample loaded into the capillary or the microchannel often exceeds the
total volume of (initially low concentration) sample loaded into the capillary, or the microchannel
can exceed the total volume of the capillary. In contrast, the finite-plug technique involves sample
zones of fixed size. The sample zone size can be defined by duration of sample injection by either
electrokinetic flows or by hydrodynamics flow. In microchannel networks, the sample size can
be accurately controlled by chip geometry, for example, by using pinched-injection or staggered-
T injection [46–48].
Figure 38.3 shows schematic representation of microchip-based single-interface FASS system.
The interface between high and low conductivity buffer regions is generated by applying a vacuum at
the north reservoir. Once a buffer–buffer interface is established, the vacuum is released and an axial
electric field is applied. Upon application of an axial electric field along the west-to-east direction,
sample stacks at the interface between buffer streams. Figure 38.3b shows images of the stacking
process at selected times. In Figure 38.4, instantaneous images of stacking process are shown for
a case where EOF was not suppressed. Since the EOF velocity in all regions is greater than the
negative electrophoretic velocity of fluorescein dye, the stacked region moves in the direction of
EOF. The images clearly show the favorable pressure-gradient-induced curvature of the stacked ions
on the downstream (left-hand) side of the interface. As described earlier, these pressure gradients act
to disperse sample and reduce the efficiency of FASS. From Figures 38.3b and 38.4, the efficacy of
the EOF-suppression method is apparent as the conductivity interface is nearly stationary and there
is negligible pressure-induced curvature of stacked analyte. Figure 38.5 shows the temporal devel-
opment of the spatial concentration distribution of sample ions. The peak intensity increases roughly
exponentially at first and then saturates at a maximum achievable concentration enhancement of γ .
38.4.1.3 Dispersion Theory

In most electrophoretic experiments, the quantity of practical interest is the cross-sectional area-
averaged concentration distribution of sample ions. This quantity is, for example, proportional to the
50 µm
50 100 150 200 250
FIGURE 38.4 CCD images of on-chip FASS in an untreated channel with significant EOF mobility. The
images clearly show the development of stacked fluorescein in the favorable pressure gradient region (i.e., the
high conductivity region). The sample was 25 µm fluorescein dye dissolved in 5 mM Borate buffer. The BGE
for this particular experiment was 25 mM Borate buffer (pH = 9.2). The electric field in the sample region was
50 V/cm and γ = 4.
4
Normalized intensity
2
Increasing
time
1
0
–5 0 5
x
FIGURE 38.5 Measured axial intensity profiles for sample ions. The profiles were obtained by averaging
two-dimensional image intensity data along the width of the channel. The applied nominal electric field was
588 V/cm and γ = 4. Time between individual profiles was 0.15 s.
measured signal intensity of line-of-sight optical integrators such as pointwise fluorescence detectors,
transmitted-mode absorption detectors, and width-averaged electrophoregrams from charge-coupled
device (CCD) arrays. This signal determines the key detectability constraints of electrophoretic
separations [3]. As described earlier, Taylor dispersion analysis allows us to develop cross-sectional
area-averaged transport equations. Such models provide useful insight into the physics of the process
and lead to the identification of key parameters that can be used to develop optimization strategies
for FASS experiments.
The FASS model requires description of electromigration, diffusion, and advection of sample
ions as well as BGE ions. The general system of equations is highly coupled and nonlinear and,
therefore, difficult to solve. However, the concentration of sample ions is much smaller than the
buffer ions (typically µM sample ions concentration or less versus order 1 mM buffer ion concen-
trations). Therefore, we can decouple the buffer and sample ion concentration fields. Using this
approach, Bharadwaj and Santiago [4] have developed a dynamic model for FASS in a flat-plate
3.5
Model
3 Experiment
2.5
2
C
1.5
0.5
0
–2 0 2 4
x
FIGURE 38.6 Comparison of model prections and measured concentration profiles. γ = 4, E0 = 379 V/cm,
and time-between-frames was 76 ms. The model parameters are Pee = 55, α = 0.23, β = 0.28, δ = 1.27.
geometry. The analysis provides the following equation for the cross-sectional area-averaged sample
ion distribution, CS :

∂CS ∂CS ∂ 2 CS 1 8g(x, t) ∂ ∂CS ∂(CS E)
+ α
u =D 2 + α 2 β 2 Pee g(x, t) − z S µS . (38.29)
∂t ∂x ∂x Pee 105D ∂x ∂x ∂x
The parameters governing this system of equations are
E0 µeph FsS −ε0 εr ς /η µeof d sB

Pee = ; α= = ; β= ; and δ= .
DS µeph µeph sS sS
Pee is the electric Peclet number, expressed as the ratio of diffusion time to electromigration time; α is
the ratio of electroosmosis to electrophoretic mobility; β is the ratio of channel width to characteristic
length scale for the initial sample ion concentration distribution; and δ is the ratio of the length scale
of the initial BGE and sample ion concentration gradients. Here, d is the channel depth, sS and sB
are the initial sample and BGE ion concentration gradients, and E0 is the nominal electric field. The
function g(x, t) in the advective dispersion term accounts for the axial variation in pressure-driven
velocity profile (cf. Figure 38.2).
Figure 38.6 shows comparisons between dispersion model predictions and experimentally mea-
sured concentration profiles. Experiments are shown for γ = 4 and 9 and ES values of 379 and
588 V/cm. There is a quantitative agreement between measured area-averaged concentration pro-
files and the model prediction throughout the time of observation. The model predictions were
obtained by solving Equation 38.29. As shown in Figure 38.3b, for times approaching 1 s, the region
of high area-averaged concentration becomes two-dimensional as it enters the staggered-T injection
region of the system, which cannot be captured by one-dimensional model. The dispersion model is
able to capture important features such as the development of peak width and the temporal growth
of the maximum concentration. Also, the model describes convective–electromigration–diffusion
dynamics in a purely two-dimensional flow in a wide, shallow channel (neglecting the influence of
side walls). In reality, however, the microchannels in the experiment have a D-shape characteristic
of an isotropic etch with a width to maximum depth ratio of 2.5. This assumption may be improved
in future refinements of the dispersion model. For example, Dutta and Leighton [6] have investi-
gated the effect of isotropic-etched microchannel geometries on the dispersion coefficient for simple
pressure-driven flows. Their analysis shows, in the Taylor dispersion limit (ruled by the widest
channel dimension), the dispersion coefficients for the D-shaped channels can be three to four times
16
14
12
C 10
6 Pee = 50
Pee = 60
4 Pee = 70
100 200 300 400
FIGURE 38.7 Optimum value of γ for fixed stacking times of 1 s. The parameters for the dispersion model
are: α = 0.5, β = 2, and δ = 1. For a given analysis time and fixed values of Pee , α, β, and δ, there is a unique
value of γ that provides maximum concentration enhancement.
larger than those predicted by simple two-dimensional analysis. Such advective–diffusion effects
would be most important in flow for large γ values (associated with larger internally generated pres-
sure gradients) and the large Peclet numbers associated with high electric fields. Another possible
refinement of the model would include dispersive effects due to Joule heating [49], which should
also lead to reduction in the rate of concentration increase.
38.4.1.4 Performance and Guidelines

In this section, we summarize some parametric and optimization results from dispersion models.
First, we consider the effect of conductivity ratio, γ , on stacking efficiency. Figure 38.7 shows clearly
that there is an optimum value of γ for a given set of parameters and analysis time. In these plots, the
analysis time is fixed as we are interested in FASS as a preconcentration step before electrophoretic
separation. These fixed-time comparisons help to determine the time needed to achieve adequate
concentration enhancement and the initial condition of the subsequent uniform-conductivity sep-
aration process. These model predictions were generated with a parametric variation study using
the numerical model described earlier. This result is in contrast with the ideal concentration factor
described earlier, which shows that increasing γ always increases the concentration enhancement
(Equation 38.28). This in an important feature of the dispersion analysis as it gives experimentalists
a method of choosing values of γ to yield optimal signal detections.
The existence of an optimal value of γ can be better understood by considering the scaling of the
parameters of interest. Equation 38.28 shows that the maximum sample concentration is proportional
to γ :
CC,max ∝ γ . (38.30)
In contrast, the ratio of the EOF velocities in the low-conductivity region to the value in the
high-conductivity region scales is
UEOF,S ζS ES
∼ ∼ γ 1+n . (38.31)
UEOF,B ζ B EB
The parameter n refers to the ratio of the zeta potential in the sample and the BGE regions. Typical
values of this parameter range between 0.2 and 0.3 [50,51], so that the advective dispersion effects
of mismatched slip velocities is negligible for low γ but dominates at high γ . An analogous scaling
7 =5
= 10
6 = 20
C 5
2
50 100 150 200
0Pee
FIGURE 38.8 Optimum value of Pee for a fixed stacking time of 1 s. The parameters for the dispersion model
are α = 0.5, β = 2, and δ = 1. At low Pee , diffusive dispersion dominates and concentration enhancement
suffers. At high Pee , advective dispersion is dominant and again concentration enhancement suffers. For fixed
γ and analysis time, there is a unique, optimal Pee (e.g., an optimal electric field for a given channel system)
which results in maximum concentration increase.
observation was made by Burgi and Chien [31]. They discuss the existence of an optimum γ using
simple scaling arguments. They developed an algebraic model for the long-time behavior of a finite-
length sample plug variance as a function of γ , using a one-dimensional Taylor dispersion approxi-
mation. In contrast to their model, the dynamic model described here allows quantitative prediction
of both temporal and spatial development of the sample ion, BGE ions, and electric field profiles.
Another important parameter determining the convective dispersion and hence, rate of concen-
tration increase in FASS is α, or the ratio of electroosmotic and electrophoretic mobilities. For a
typical value of electrophoretic mobility (e.g., 3E−8 m2 V−1 s−1 ), α is approximately equal to 2
for glass microchips. Therefore, the dispersion dynamics of untreated glass chips are well in the
advection regime. However, at least an order of magnitude reduction in electroosmotic mobility is
possible by dynamic surface coatings using neutral water-soluble polymers [52]. It is, therefore,
interesting to experimentalist to quantify the importance of suppressing EOF in determining maxi-
mum achievable concentration increases. Bharadwaj and Santiago [4] show that, even for the case of
a 10-fold decrease in electroosmotic mobility (α = 0.2), there is significant convective dispersion.
The preconcentration time required to reach the maximum concentration enhancement can be as
much as 50% longer for α = 0.2 case as compared to a case where there is no EOF. This result has
important consequences in the design of microchip-based FASS systems because a slower rate of
concentration increase can adversely affect both the amount of sample required per separation and
the throughput of the device. To increase the rate of concentration enhancement, the electric field
and Peclet number should be increased.
Figure 38.8 describes the effect of Peclet number, Pee = E0 µeph FsS /DS , on maximum concen-
tration achievable for a given analysis time. Initially, increases in Pee are favorable for concentration
enhancement. This is due to a reduced contribution of molecular diffusion, which scales as Pe−1 e .
However, as Pee is increased further, the concentration increase slows down and, past a critical Pee ,
the achievable concentration begins to decrease. The latter effect is due to the aforementioned fact
that convective dispersion increases with increasing Pee . Equation 38.29 shows the dispersion term
scales as α 2 β 2 Pee . There is, therefore, an optimum value of Pee (e.g., an optimum electric field for
a given process and geometry) for a given analysis time and fixed values of γ , α, and β. Note that
the optimum value of Pee may be somewhat smaller in practice than that predicted by our dispersion
model since we do not account for the effects of Joule heating [49]. Joule heating is proportional
to the square of local electric field and is expected to be important for very high field strengths and
relatively large channels.
3
t=0
2.5 t = 3.2
t = 32
2
C 1.5 Separated species
0.5
0
–20 –15 –10 –5 0 5
x
FIGURE 38.9 Stacking and separation dynamics of three negatively charged sample species. The model
parameters are Pee = 40, γ = 5, β = 2, α = 0.05, δ = 1, and h = 6. The dimensionless electrophoretic
mobilities of the species are 1, 2, and 3, respectively. The initial dimensionless concentrations of the sample
species are 1/6, 1/2, and 1/3, respectively.
Finally, Figure 38.9 shows model results for sample stacking and separation dynamics of a finite
injection volume (initially approximately six channel depths wide) of three anionic sample ions.
Initially, there is rapid stacking (accumulation) of the sample ions as they exit the low conductivity
region and enter the high conductivity region. Once sample ions enter the high conductivity region,
sample stacking ends and ions are subsequently electrophoretically separated into three distinct
peaks. The dispersion model can be used to optimally design FASS-based electrophoretic separation
systems for the analysis of multiple sample species. For example, the model predictions can guide
the location of detector and width of initial sample plug to ensure adequate signal-to-noise ratio
(SNR) and resolution.
38.4.2 TEMPERATURE GRADIENT FOCUSING

So far, we have alluded to a number of preconcentration techniques that are of interest to the CE
community and, in the previous section, we examined FASS in detail. In this section, we consider
dispersion as it relates to focusing techniques. The quintessential focusing technique is IEF [53–
55]. In IEF, charged species migrate through a pH gradient under the influence of an electric field
until they reach their isoelectric, or pI, point. At the pI, a species becomes net neutral and ceases
to migrate. While this is the most common focusing technique, it is not the most general from the
perspective of dispersion analysis and scaling. Consequently, we choose microfluidic temperature
gradient focusing (TGF) as it represents an excellent case study in dispersion and provides a more
general model. The approach and discussions presented can be extended to the analysis of Taylor
dispersion in IEF.
TGF is a form of electric field gradient focusing, where a temperature-induced gradient in electric
field helps produce the gradient in electrophoretic velocity required for focusing (in contrast to
IEF, where the focusing effect is caused by the changing charge on an ampholyte). TGF was first
described in a seminal paper by Ross and Locascio [56], where they demonstrated the successful
focusing of charged fluorescent dyes, amino acids, green fluorescent protein, DNA, and polystyrene
particles, illustrating the general utility of TGF. TGF has subsequently been extended to DNA
hybridization assays and single nucleotide polymorphism detection [57], as well as the detection
of chiral enantiomers [58]. Focusing of neutral and ionic hydrophobic analytes (e.g., coumarin)
ueph – ubulk
ubulk ueph
T
FIGURE 38.10 Schematic of TGF process with advective dispersion. An electrophoretic velocity is countered
by an opposing liquid flow, composed of both pressure-driven and electroosmototic flow (top). A temperature
gradient is applied to a microchannel, inducing a gradient in the electrophoretic velocity of an analyte. The
analyte focuses where the electrophoretic and convective (“bulk”) fluxes sum to zero (middle). Both molecular
diffusion and advective dispersion broaden the band about the focus point. The bottom image shows Bodipy
focused in a 20 by 200 µm wide channel with an applied electric field and temperature gradient of 60 V/mm
and 10◦ C/mm, respectively.
has also been achieved [59,60] using TGF combined with micellar electrokinetic chromatography
(MEKC). The capabilities of TGF allow it to be used in analytical (i.e., detection and separation) and
preparative (i.e., concentration and purification) applications. In both applications, minimization of
dispersion is an important design goal.
We have already discussed the basics of dispersion, noting how decreased dispersion improves
resolution and sensitivity in separation applications [61], and also yields improved dynamics for
concentration and purification applications [4]. However, there are some key differences to consider
when comparing focusing techniques such as TGF with other techniques. We describe the basics of
TGF theory, implementation details, and the modifications to Taylor dispersion required for TGF.
Finally, we present tips for empirical optimization of TGF preconcentration factors and resolution.

Temperature gradient focusing focuses charged species by balancing an axially varying elec-
trophoretic flux with a counterflow, causing species to focus at locations where their net fluxes
sum to zero (Figure 38.10) [56].

ubulk + ueph (xfoc ) = 0, (38.32)
where xfoc is the focus location for the species in question and “bulk” refers to the net effect of both
pressure-driven flow and electroosmosis. Note that, unlike IEF, the electrophoretic velocity of the
charged species is always nonzero, so a counterflow, and its associated dispersive effects, is always
present and significant. In practice, the counterflow is provided by a combination of flow due to elec-
troosmosis and flow caused by an externally applied pressure difference. Variation in electrophoretic
flux is accomplished by applying a temperature gradient along the axis of the microchannel and
by using a buffer with a temperature-dependent ionic strength (e.g., due to temperature-dependent
dissociation of a weak electrolyte).
Within the channel, local electric field is inversely proportional to conductivity, which in turn is
a function of local viscosity and ion density. Using the convention of Ross et al. [56], we write the
conductivity as σ (T ) = µ0 σ0 /(µf ), where µ0 and σ0 are the buffer viscosity and conductivity at a
defined reference temperature, µ(T ) is the viscosity, and f (T ) is a nondimensional function incorpo-
rating any remaining conductivity dependencies (primarily the change in ionic strength). A similar
decomposition is applied to the electrophoretic mobility, yielding veph = v0,eph µ0 /(µfeph ), where
v0,eph is the analyte’s electrophoretic mobility at the reference temperature, and feph (T ) accounts for
any other temperature dependencies. The usefulness of this decomposition becomes apparent when
we assume uniform current, I, use Ohm’s law to determine the local electric field, and solve for the
electrophoretic velocity:
ueph = v0,eph E0 f /feph , (38.33)
where E0 = I/Aσ0 and A is the channel cross-sectional area. Most analyses assume feph (T ) remains
near unity, which is equivalent to assuming that the ion has a charge and drag coefficient independent
of temperature. In this case, we see the electrophoretic velocity is a function only of the temperature,
through f . In addition, we note the rate of (global) mass accumulation in the entire volume of a channel
due to electrophoretic focusing is proportional to the difference in f on the boundaries. However,
for a differential volume or for linear f (T (x)), the focusing is proportional to the gradient of f .
TGF has been implemented in a variety of microfluidic formats, including channels imprinted
in polymer substrates and embedded capillaries [56,62]. In most cases, the temperature gradient is
imposed by mounting the channel across temperature-regulated blocks, such that the temperature
gradient is established via thermal conduction in the gap between the blocks. Figure 38.11 shows
(a) Filter cube
Lamp CCD PC
Objective HV power
supply
Pressure TEC
controller controller
(b) (c)
FIGURE 38.11 Control schematic and images of TGF fixture and capillary assembly. The TGF assembly (a)
provides the thermal, fluidic, and electrical interface to the microchannel. A temperature gradient is established
across a gap between two copper plates, each heated or cooled by a thermoelectric (Peltier) device. Pressure
control is accomplished by adjusting the relative heights of two external reservoirs. The capillary assembly
(b) consists of a 20 by 200 µm rectangular glass borosilicate capillary that spans the distance between two
O-ring reservoirs and underneath an insulating PDMS block. In the assembly photo (c), the capillary assembly
is mounted on the TGF fixture. The encapsulating PDMS block is located directly below the objective, and the
fluidic manifolds are to the left and right of the block [62].
one such implementation. Alternatively, Joule heating has also been used to form the temperature
gradient independent of external fixtures [56,63].

To analyze dispersion in TGF, we must modify our previous derivation to account for changes due
to the temperature gradient. This includes an axially varying electrophoretic velocity and diffusivity.
In its most general form, the transport equation becomes
∂c
+ ubulk · ∇c + ∇ · ueph c = ∇ · ∇ (Dc) , (38.34)
∂t
where c is the concentration of the sample analyte and D is the analyte’s molecular diffusivity.
Note the placement of the electrophoretic velocity and diffusivity within the gradient operators. In
particular, the placement of the diffusivity within the second gradient operator reflects the use of
the Fokker–Planck diffusivity law, J = −∇(Dc). Although it is common practice to use Fick’s law
for the diffusive flux (yielding the traditional ∇ · D∇c), Fick’s law strictly applies only to diffusion
with homogeneous D [64]. To recover the more familiar diffusion representation, we differentiate,
yielding the terms, D∇c and c∇D. The latter term thus represents a flux due to a diffusivity-induced
velocity, ∇D, which is typically small in comparison to the electrophoretic velocity.
For a purely axial temperature gradient, if the cross section mean decomposition is again
performed, a new mean concentration transport equation is derived

∂
c ∂
c ∂ ∂ ∂
c
+
ubulk = (
ueph
c) = D −
ubulk c . (38.35)
∂t ∂x ∂x ∂x ∂x
The deviation transport equation again reduces to a form analogous to Equation 38.11 following our
scaling arguments, although with one exception, the characteristic time scale for TGF is no longer
the observation time, but is instead the focusing time for a nondiffusive particle approaching its
focal point, τfoc . (See solution for form of τfoc .) Note that if any of these assumptions are violated,
then new dispersion mechanisms arise. For example, if τfoc is of order a2 /D or less, then ballistic
dispersion [20] can become significant. The correlation term on the right-hand side of Equation 38.35
once again corresponds to the advective dispersion arising from transverse variations in the axial
velocity. To evaluate this term, we need an expression for u .
An analytical solution for u was derived by Huber and Santiago [62]. Using a decomposition
on the Helmholtz–Smolukowski equation similar to that performed by Ross for the electrophoretic
velocity, they determined the nonuniform electroosmotic slip velocity to be
ueo = veo,0 E0 f (T )g(T ), (38.36)
where veo is the electroosmotic mobility, g(T ) ≡ veo (T )µ(T )/veo,0 µ0 = ε(T )ζ (T )/ε0 ζ0 , ε is the
permittivity, ζ is the zeta potential, and the subscript zero indicates a value at the reference temper-
ature [16]. By extending the lubrication flow solution of Ghosal [65] to include variable viscosity
and the nonuniform electroosmotic slip velocity of Equation 38.3, Huber and Santiago derived an
analytical expression for the bulk velocity, from which the deviation velocity was determined:

1 3y2
ubulk = Up − , (38.37)
2 2a2
a2 P {µfg}
Up = + veo,0 E0 − veo,0 E0 fg, (38.38)
3{µ} Lch {µ}
where a is the channel half-height, Lch is the length of the channel, P the applied pressure difference,
and the curved brackets indicate an axial mean over the length of the channel. We see the flow field
is a superposition of a uniform electroosmotic component and a parabolic (in y) pressure-driven
component (for flow between parallel plates). The two are linked through continuity, so as one
decreases, the other must increase. Thus, Up , the pressure-driven flow component, contains both
the externally applied pressure gradient and the internally generated pressure gradient, which results
from the local slip velocity deviating from the axial average. Note that the axially averaged terms
are uniform and constant, and f and g are functions of T ; therefore, the velocity profile varies in x
in response to axial temperature changes.
We analyze a flat-plate geometry again for simplicity and as an approximation to the full focusing
problem. This implies, we are interested in characteristic focusing times much longer than that for
molecular diffusion along the depth but not for spanwise diffusion along the width of the channel.
This implies that our analysis is an estimate of the maximum concentrations achieved over these
relatively short focusing times (along the centerline of the channel) and does not account for the
finite amount of dispersion introduced by spanwise velocity gradients. Improvements on this model
are discussed in the following text.
Given u and c , the advective dispersion term may be determined, which yields

∂
c ∂
c ∂ ∂ ∂
c
+
ubulk · + (
ueph
c) = Deff , (38.39)
∂t ∂x ∂x ∂x ∂x

2 2
2 Up a
Deff = D 1 + . (38.40)
105 D2
This 1-D convection–diffusion equation has features of the Taylor–Aris dispersion equation.
However, in contrast to Taylor–Aris, here Deff is a function of temperature and the axial coordinate,
as it depends on both D and Up . When f is a linear function of the axial dimension, Equation 38.39
can be solved in closed form subject to the form of Deff . If Deff is uniform, the solution is a Gaussian
with peak variance σ 2 = 2Deff τfoc , where τfoc = 1/2E0 v0 |df /dx|. If Deff is also a linear function
of x, the solution is
# $
x/L − ln(1 + x/L)
c = c0 exp − , (38.41)
σ 2 /L 2
where c0 is the peak height and L = Deff (0)/(dDeff /dx) [62]. Although unusual in form, this solution
produces Gaussian-like peaks with a skew that grows with the slope of Deff . For more general cases,
the equation can be solved numerically.
Figure 38.12 shows sample full-field fluorescence images of focused Bodipy proprionic acid
in an applied temperature gradient of 10◦ C/mm and electric fields from 1 to 215 V/mm. The two
cases illustrate different dispersion regimes. The images on the left show focusing in the molecular
diffusion dominated regime. Note the lack of spanwise curvature in the peaks and the near inverse
square root dependence of peak width on electric field (dotted line). The images on the right-hand side
show focusing experiments where the Taylor dispersion criteria are violated about E0 = 40 V/mm
and there is enhanced advective (or ballistic) dispersion [20]. The solid line is a theoretical prediction
for the peak width based on an heuristic ballistic dispersion model [66]. In both cases the direction of
electrophoretic flux is left to right, while the bulk flow is right to left, driven by electroosmosis. The
experimental and theoretical peak widths agree closely with some deviation at low fields, where the
focused peak remains slightly “over-focused” as field was decreased, having had insufficient time
to diffuse out to full width.
(a) (d)
(b) (e)
(c) (f)
1 0.55
0.5
0.8
Peak width (mm)
Peak width (mm)

0.45
0.6
0.4
0.4
0.35
0.2 0.3
0 0.25
0 10 20 30 40 50 0 50 100 150 200
Reference field, E0 (V/mm) Reference field, E0 (V/mm)
FIGURE 38.12 Bodipy peak images and width as function of reference field. The images show focused
Bodipy dye within a 20 × 200 µm rectangular capillary, while the bottom figures plot the peak widths versus
electric field as determined by the fitting of Gaussian profiles to each peak. The applied temperature gradient
was 10◦ C/mm and the applied electric fields ranged from 1 to 215 V/mm. The left hand side shows results
for focusing in the molecular diffusion dominated regime. The current-normalized fields, E0 , were (a) −1.5,
(b) −15, and (c) −30 V/mm. The right hand gives results which violate the Taylor scaling arguments above
40 V/mm and thus feature enhanced ballistic dispersion. (Note the curvature in the images of the peaks.) The
fields were (d) −38, (e) −81, and (f) −212 V/mm [62,66].

As we have described, TGF has successfully demonstrated the simultaneous concentration and
separation of a wide range of species in a variety of implementations. Key figures of merit for
TGF as a separation modality are the concentration factor and peak capacity. Static TGF, where the
externally applied pressure remains constant, has demonstrated concentration factors in excess of
20,000 but is limited to peak capacities <10 [56]. As a result, there has been recent work to develop
a dynamic form of TGF called scanning TGF [67], where the externally applied pressure is varied
with time. This technique allows higher peak capacity and tunable resolution and concentration by
adjusting the rate at which the procedure scans through the applied pressures.
In implementation, the complexities of TGF extend beyond that due to advective dispersion and
encompass system design details such as device geometry, temperature control, and buffer selection.
In static TGF, larger temperature gradients produce faster focusing and sharper peaks but reduce
resolution [56]. The latter may be addressable using scanning TGF, but particular attention must
be paid to the temperatures within the system because analytes of interest (e.g., enzymes) may
be extremely sensitive to temperature and Joule heating may adversely influence the temperature
gradients. In order to produce desired temperature profiles, designers must consider conductive and
convective heat transfer within the channel, Joule heating due to the electric field, and heat transfer
from the channel to its fixture and the environment. While it is possible to model these effects, in
most cases, empirical studies will be required to validate models and refine optimizations.
Optimization of TGF preconcentration and (static) resolution is similar to other sample precon-
centration methods. As described earlier, the contribution of molecular diffusion to the total peak
variance scales as 1/E , while the Taylor dispersion coefficient scales as Up2 and therefore, E 2 . Sample
peak variance, therefore, scales as 1/E at low fields and as E at high fields. Thus, there will (again)
be an optimal field strength that minimizes the overall peak width. Finally, we note that reducing the
characteristic width of the channel simultaneously decreases the amount of Taylor dispersion and
Joule heating. However, we caution that most detection techniques rely on depthwise integration of
a signal (e.g., fluorescence), which then scales as the channel height, h. Thus, TGF system designers
must also optimize the channel width subject to their desired detection limits.
38.4.3 ISOTACHOPHORESIS
Isotachophoresis [68] is an electrophoretic preconcentration and separation technique that utilizes
a heterogenous buffer system of disparate electrophoretic mobilities. Typically, a plug containing
sample ions to be focused and separated is introduced between the leading and trailing electrolyte
(LE and TE) whose mobilities are respectively higher and lower than any of the mobilities of sample
ions. Under the influence of an electric field, the sample ions separate and redistribute themselves
in contiguous zones in order of reducing mobility starting from LE to TE (each is focused into its
respective, mobility-dependent focal point). At steady state, these focused sample zones migrate at
a same speed as the leading zone, hence the name “iso-tacho-phoresis.”
ITP, also known as displacement electrophoresis, was first performed in capillary tubes by Ever-
aerts et al. [69] for the separation of strong anions using a thermocouple detector. Since then,
ITP has been used for the analysis of various important chemical and biological species such as
amino acids [70], peptides [71], nucleotides [72], proteins [73,74], heavy metal ions [75], and other
organic/inorganic ions [76,77] on a variety of detection platforms such as UV absorbance, conduc-
tivity and fluorescence detection. Over the past 15 years, ITP has been used as a preconcentration
technique in conjunction with CE [78]. This mode of ITP, referred to as transient isotachophore-
sis (tITP), has been implemented on microchip platform in the recent years to achieve improved
sensitivity [22,79,80].

ITP leverages differences in the mobility of sample ions to create a segregation of species on the basis
of mobility. ITP separation results in a system of contiguous sample zones, sandwiched between
leading and trailing electrolytes, migrating at identical speeds. The inherent preconcentration effect
of ITP maintains sharp concentration boundaries between adjacent sample zones. In the absence of
dispersion due to (radially) nonuniform bulk flow, the thickness of these boundaries is governed
by the balance of electromigration and diffusion flux, as we shall discuss later. Away from this
diffused boundary region, the sample concentration is uniform and can be obtained in each zone using
species conservation equations and the electroneutrality condition [81]. Consider a simple model ITP
system, shown in Figure 38.13, consisting of a plug of sample ions (Xi ) injected between the leading
electrolyte (LE) and trailing electrolyte (TE), and common counterion (A) present everywhere. For
now, we assume EOF is fully suppressed to negligible levels. After sufficient focusing time, the
various zones in the system will develop locally uniform concentrations (concentration “plateaus”
around their respective focal points) where diffusive fluxes are locally insignificant. For this long-time
condition, the species conservation equations can be simplified to obtain the well-known Kohlrausch
regulating function (KRF) [82] given by

% Ci
= f (x), (38.42)
µeph,i
i x
(a) Sample LE (b) TE Sample LE

1 2 1 2
C L– C L–
X1– X2– X3– T– X1– X2– X3–
x x
(c) Sample zones (d) LE
TE LE LE + TE
1 E 2 1 E 2
X2– X3– X2– X1– L– +HV
+HV T–
C C
X3– X1–
T– L–
x x
FIGURE 38.13 A schematic of the steps involved in a typical anionic ITP experiment is shown. (a) First,
the capillary/microchannel is filled with the leading electrolyte using pressure driven flow. Sample electrolyte
is introduced in well 1 and the capillary/channel is partially filled with this sample electrolyte by applying
vacuum on well 2. (b) Well 1 is emptied and filled with the trailing electrolyte which is then drawn into the
capillary/channel. (c) Next, high voltage is applied across the capillary and an electric field is setup inside
the capillary/channel initiating separation of sample zones and achieving ITP condition. Sample ions X1 and
X3 are present in high initial concentration and hence their zones form plateau shaped peaks and achieve the
concentration as required by KRF. X2 appears as a peak in the diffused interface between X1 and X3 zones and
has a much lower concentration than predicted by KRF. (d) Finally, well 1 is emptied and is filled with LE. The
LE ions overspeed the TE and sample ions and “break” the ITP mode into CE mode. The sample peaks separate
in the CE mode and also disperse due to electromigration dispersion.
where Ci and µeph,i are the concentration and the mobility of species i, at axial location x in the
channel. The constant (in time) function f (x) is governed by the initial condition. This relation can
be stated as follows: The sum of concentration-to-mobility ratio of all species at a given location in
the channel (relative to the channel wall at negligible electroosmosis conditions) remains invariant
with time.
Using the KRF, we can arrive at the following expression to obtain the concentration adjustment
of an ITP zone assuming no dispersions and bulk fluid velocity [83]:
µX µA + |µL |
CX = CL , (38.43)
µL µA + |µX |
where the adjusted sample concentration, CX , is strictly a function of the LE concentration and
electrophoretic mobilities in the nondiffuse one-dimensional model. Note that KRF may be used
to estimate sample concentration in an ITP zone only if the species are strong ions (fully ionized).
Further, the species are present in sufficiently high initial concentration to form concentrated regions
with finite-width, but locally uniform concentrations (a plateau in analyte peaks, e.g., samples X1
and X3 in Figure 38.13c).
For low initial concentrations (or short preconcentration times), the ITP zone width can be on
the order of the diffusion width between zones (e.g., sample X2 in Figure 38.13c). In that case, as
pointed out by Svoboda [84], analyte width is a function of both initial concentration and injected
plug length (see also the quantitative data of Jung et al. [22]). This regime of ITP has been called the
spiked mode [85] and here, the analyte zone appears as a spike between adjacent zones rather than a
plateau. One often encounters this spiked mode ITP in trace analyte detection and separation. In this
regime, Taylor dispersion is a critical factor in determining the maximum concentration enhancement
achieved. As a result, the sample peak widths (in spike mode) and interfaces between ITP zones
can often be greater than peak widths predicted by diffusion alone. In our work, this factor is also
reflected in the rate of growth of sample peaks. Despite this, there has been very little work in the
analysis and modeling of Taylor dispersion in ITP.
(a) Sample
ions LE
TE ions
ions
50 µm
(b) Sample
ions
TE ions
LE ions
50 µm
FIGURE 38.14 (See color insert following page 810.) CCD camera images of on-chip sample peaks of
AlexaFluor 488 at the LE/TE interface in two different ITP experiments. In (a) there is finite (nonuniform) EOF
and the sample peak streamwise dimension is on the order of channel width or larger. In (b) EOF is suppressed
and the sample is concentrated in narrower zone (∼5 µm) at relatively high electric field. While Taylor dispersion
based analysis is probably applicable in the first case, more comprehensive modeling is required for case (b).
(a) (b)
C C
LE LE– FLT S1– S – LE–
S1–+S–2+TE– S1–+S–2+TE– S–2 1
x x
TE + LE GND +HV
sample Vacuum
(c) (d)
C C
S–2 S1–
GND LE– – – LE– GND LE–
– S2 S1
TE
x x
+HV FLT +HV
FIGURE 38.15 Schematic of ITP/CE assay protocol in a microchip. Configurations of co-ions are also shown
at each step. (a) The north and the south reservoirs are filled LE, and the west reservoir is filled with a mixture of
TE and sample. TE/LE boundary is formed by applying vacuum at the south reservoir. White arrows show the
direction of pressure-driven flows. (b) ITP preconcentration is initiated by applying high voltage and ground at
the east and west reservoirs, respectively. The black arrow denotes the direction of electric field. Sample anions
electromigrate toward the anode as EOF is suppressed. The early stage of ITP preconcentration results in a
partial separation (i.e., moving boundary electrophoresis). (c) The field is switched toward the north reservoir
to inject LE ions behind the sample and initiate CE. ITP preconcentration continues until LE ions overtake the
TE and sample ions. (d) Separation of samples occurs further downstream where sample ions electromigrate in
nearly homogeneous LE electrolyte (remnant of TE not shown).
Example visualizations of ITP-focused sample peaks in a 50 µm wide by 20 µm deep (isotrop-

ically etched) microchannel are shown in Figure 38.14. The top image shows a typical ITP sample
peak in a channel with finite EOF and Taylor dispersion. Interface lengths are on the order of or
larger than the characteristic channel cross section dimensions. In this experiment, the ITP stacking
occurs over a long duration (∼5 min) such that the peak has a substantial axial dimension compared
to the channel width and has clearly dispersed edges. The area-averaging dispersion analysis and
scaling presented earlier should apply here. Figure 38.14b shows early stage of an ITP plug created
on-chip with high-quality EOF suppression (using poly-N-hydroxyethylacrylamide [PHEA] coating
on a borosilicate glass wall [22]), high electric field (∼1000 V/cm in the TE), and high LE con-
centration (∼1 M). The injection protocol used here is shown in Figure 38.15a,b. Here, the sample
peak is a narrow concentration “shock wave” with extremely high electric field and concentration
gradients. Accurate prediction of the focusing dynamics of such a peak will probably require fairly
comprehensive two- and three-dimensional models.
The coupling of ITP with CE is shown in Figures 38.13d and 38.15c–d. This coupling is referred
to as transient ITP (tITP). tITP is typically achieved by replacing the TE ions with LE ions, which
interrupts the ITP process. After sufficient sample preconcentration via ITP (as in Figure 38.13c), LE
ions are introduced behind the contiguous sample zones. This is accomplished by either manually
exchanging the buffer in well 1 (e.g., with a hand pipette, as shown in Figure 38.13c) or by using
a side channel of a microchip to inject LE ions electrokinetically behind the train of sample zones
(Figure 38.15c,d) [22,86]. The LE ions enter the sample zones faster. Sample zone concentrations
reduce in order to satisfy the KRF-type regulation, and sample plug typically forms long tails as
shown. This form of dispersion is termed as electromigration dispersion (EMD), and has been
extensively investigated [87–89]. Once the ITP process is interrupted, the samples separate via
standard mobility-based CE separation.

The isotachophoretic boundary between two adjacent zones, under constant current condition and in
the absence of bulk flow, assumes a constant width governed by the balance of electromigration and
dispersion fluxes. For negligible electroosmosis (and negligible Taylor dispersion), the dispersion is
determined by diffusion alone. Analytical solution to the concentration of the species in this diffused
boundary, for a three-component fully ionized system, has been presented by Saville et al. [90]. The
characteristic length-scale, δ, of the ITP boundary in this case is given by

µ L µT C L kB T
δ= , (38.44)
µL − µ T j e
where µL and µT are electrophoretic mobilities of LE and TE, respectively, CL is the concentration of
leading ion, j is the current density, and kB T /e is the thermal voltage. For negligible Taylor dispersion
(e.g., low fields), peak axial dimension scales as D/E (since diffusivity is proportional to mobility
by the Nernst–Einstein relation). Typical concentration profiles of LE, TE, and counterion in an ITP
boundary (with molecular diffusion alone contributing to dispersion) are shown in Figure 38.16.
The case of nonsuppressed EOF conditions is much more complex. For even trace EOF mobilities,
there will be a mismatch in electroosmotic mobility and electric field in the adjacent zones. This results
in a mismatch in the electroosmotic velocity and hence generation of internal pressure gradients as
discussed earlier. These pressure gradients result in increased width of the interface due to dispersion
and have detrimental effects on preconcentration of trace samples via ITP.

Concentration
LE (L–)
TE (T–)
Counterion (A+)
FIGURE 38.16 Schematic of the distribution of ions in the diffused interface between leading and trailing
electrolyte in an ITP system. δ is the characteristic length scale of this diffused boundary, obtained from the
balance of electromigration and diffusion fluxes.
Saville [91] investigates the effect of electroosmosis on the interface between two ITP zones.
To our knowledge, this is the only published Taylor dispersion analysis for ITP. Saville’s anal-
ysis is for a simplified problem. He neglects the influence of electrolyte concentration and pH
on electroosmotic mobility. Similar to the analyses presented earlier, a purely axial flow veloc-
ity is split into the bulk area average velocity
u and a deviation velocity, up = u −
u. The
pressure-driven flow velocity component due to a mismatch in electroosmotic slip velocity (alone) is
up = 2(µeof E − {µeof E})(1 − 2r 2 /a2 ). While simplifying approximations, Saville assumes µeof is
uniform, and the local electric field governing EOF (but not the electric field governing electrophore-
sis) is approximately uniform in x. For example, assuming the TE voltage drop is much larger than
that of the LE; and assuming solute plug is approximately halfway along the length of the channel,
our expression reduces to µeof ETE (x)(1 − 2r 2 /a2 ). Saville writes this as
up = µeof E0 (1 − 2r 2 /a2 ), (38.45)
where E0 is assumed constant and uniform. Saville further expands µeof as −εζ E/η (where ε is the
permittivity, µ is the viscosity of the liquid, and ζ is the zeta potential) but we shall retain the more
compact form here. In a frame of reference moving at
ubulk = UEOF + Up , the species conservation
equation for any species i can be expressed as

∂Ci ∂Ci ∂ ∂Ci 1 ∂ ∂Ci
+ up = −µeph,i Ci Ex (x, r) + Di + r −µeph,i Ci Er (x, r) + Di ,
∂t ∂x ∂x ∂x r ∂r ∂r
(38.46)
where µeph,i is the electrophoretic mobility (m2 /V/s), E is the electric field (V/m), and Ci is
the concentration of the species i. The species relations are constrained by the electroneutrality
approximation expressed as
%
zi Ci = 0. (38.47)
i
Next, the conservation equations are transformed to the frame of reference moving with the
interface at the speed UITP (UITP = µL ELE , where ELE is the electric field strength in the LE region)
and subsequently nondimensionalized to obtain the following dimensionless equations:
∂C ∗
−Ci∗ + Pe(1 − 2r ∗2 ) ∗i
∂x

∂ ∂Ci∗ 2 1 ∂ ∗ ∂C ∗
= ∗ −µeph,i Ci Ex + µeph,i ∗ + λ ∗ ∗ r −µ∗eph,i Ci∗ Er∗ + µ∗eph,i ∗i
∗ ∗ ∗ ∗
(38.48)
∂x ∂x r ∂r ∂r
and
%
zi Ci∗ = 0. (38.49)
i
In arriving at the previous equation, the following scaling has been used:
Ci∗ = Ci /CA , µ∗eph,i = µeph,i /µeph,A , x ∗ = x/l, r ∗ = r/a, Ex∗ = Ex µeph,A l/DA
and
Er∗ = Er µeph,A a/DA .

Here, the following dimensionless parameters appear: Peclet number Pe = µeof E0 l/DA and the
aspect ratio λ = l2 /a2 , where l is the electrical length scale obtained by balancing diffusion and
electromigration flux, l = DA /UITP .
The typical value of Peclet number is ∼10 indicating that convection plays an important role in
determining the shape of the interface. By performing asymptotic expansion on the concentration
and potential terms in the species conservation equation and area-averaging the equation over the
cross section of the capillary, Saville obtains the following:

∂
Ci∗ ∂ ∗ ∗ ∗ ∗ (Pe) 2 ∂ 2
Ci∗
− = [−µ
C
E ] + µ + ∗ . (38.50)
∂x ∗ ∂x ∗ eph,i i x eph,i
48µeph,i λ2 ∂x ∗2
Reverting back to dimensional form, we write

∂
Ci ∂ (µeof E0 )2 a2 ∂ 2
Ci
−UITP = [−µeph,i
Ci
Ex ] + Di + . (38.51)
∂x ∂x 48Di ∂x 2
Hence, we arrive at a expression similar to the one derived in Taylor dispersion using area-averaging
section, with

(µeof E0 )2 a2
Deff ,i = Di + .
48Di
Note that Equation 38.49 follows from our more general Equation 38.16 (e.g., transform the equation
to the UITP frame of reference and assume a uniform constant value of Deff ). Saville presents
numerical solutions of his model for the case where λ = 1, µ∗eph,LE = 1, µ∗eph,TE = 0.5, and
Pe = 10. The model shows interface width (e.g., the interface between two adjacent species) as a
strong function of Pe (and E0 ). In the conditions Saville modeled, Taylor dispersion increases the
effective dispersion coefficient up to ∼18-fold over diffusion alone.
Despite the dearth of work on ITP Taylor dispersion, there is fairly clear experimental evidence
that Taylor dispersion often limits maximum achievable concentration increase. Next, we present
experimental studies of Taylor dispersion in ITP and present an empirical optimization of ITP that
minimizes dispersion to maximize preconcentration factor. We studied the effects of Taylor dispersion
on the width of single interface ITP systems shown schematically in Figure 38.15a-b, under constant
(in time) current conditions. In a typical ITP experiment, the LE contains high concentration of Cl−
(100 mM to 1 M), the TE contains HEPES− (5–100 mM) and the sample ions (Alexa Fluor 488) in
trace concentration (100 aM to 1 nM). We setup an interface between the LE and TE using pressure-
driven flow, and ITP is initiated by applying high voltage (∼3 kV) across this interface. Under the
influence of the applied electric field, sample ions overspeed TE ions in the TE zone and stack at
the LE/TE interface. Figure 38.17 shows a typical result for the width (along the axial direction) of
the focused sample plug as it propagates down the channel toward the anode. Data are shown for
four values of applied current in this 50 µm wide by 20 µm deep borosilicate glass microchannel
(Mycralyne Inc.). Here, EOF was suppressed by adding polyvinyl pyrrolidone (PVP) (0.2% w/w)
to the LE and TE. In the absence of Taylor dispersion (but including the effects of diffusion), we
expect the sample plug width to remain constant. However, the sample width clearly grows nearly
linearly in space. This behavior of sample plug width is observed even in ITP experiments with high
degree of EOF suppression. We hypothesize that this type of growth is due to a mismatch in the
EOF between the LE and TE region, which induces Taylor dispersion. This experimental evidence
suggests that the interface dimension in ITP is often controlled by Taylor dispersion.
Analytical and even numerical computations of ITP processes including diffusion and Taylor
dispersion are difficult. However, the analyses presented in this chapter can be used as a guide for
80
70
60
Width (µm)
50
40
5 µA
30 8 µA
20 10 µA
15 µA
10
0 1 2 3 4 5 6
Distance x (cm)
FIGURE 38.17 Plot of the width of the sample plug (10 nM Alexa Fluor 488) focused at the interface of LE
(750 mM Tris–Hcl) and TE (25 mM Tris-Hepes) at various locations downstream in a 50 µm wide (20 µm
deep) microchannel. Each data set represents five realizations of a constant current ITP experiment.
empirical optimizations of the ITP process. Key parameters influencing the sample preconcentration
achieved are the concentrations of LE, TE, and the initial sample concentration; the local applied
electric field; the initial conductivity gradient of the initial condition, which partly determines time
to reach steady state; and the degree to which EOF is suppressed.
Jung et al. [22] performed a systematic variation of LE, TE, and sample concentration and
measured the effect of these variations on concentration increase, CI (CI = Csample,final /Csample ).
In each case, we ensure that the ITP zones reached a fully preconcentrated state (i.e., maximum
focusing) by verifying the transients of the preconcentration procedure using full-field imaging at
low magnifications. In all cases, ITP zones reached a quasi-steady concentration, but we empha-
size this does not imply that the KRF analysis is valid (since plateau-shaped peaks were not
achieved [84]). We suppressed EOF (to minimize Taylor dispersion) by adding 0.1% (w/w) poly-N-
hydroxyethylacrylamide (PHEA) to all electrolytes to suppress EOF. We tried several suppression
strategies including poly-ethyPEO coating and PVP, but PHEA yielded the lowest electroosmotic
mobility (as verified by current monitoring measurements [23]). For convenience in working with
relatively high voltages (up to 3 kV), we used constant voltage control in this optimization of
preconcentration.
Effects of LE (NaCl) concentration, CLE , TE (HEPES) concentration, CTE and initial con-
centration of sample (Alexa Fluor 488), CS,initial on sample preconcentration are summarized in
Figure 38.18. CLE was varied from 10 mM to 1 M to study its effect on maximum focused sample
concentration, CS,final and concentration increase, CI (Figure 38.18a). The 5 mM HEPES TE solution
contained 1 nM Alexa Fluor 488 as a sample. The focused sample concentration is nearly directly
proportional to the concentration of LE, as expected from a one-dimensional nondispersive model
(i.e., KRF theory). However, the nondispersive model drastically underpredicts the proportionality
constant; the measured focused sample concentrations are 3500- to 7900-fold less than that predicted
by Equation 38.43 despite all cases reaching fully-focused state. This gross difference between KRF
theory and experiments is because the sample is in a “smeared” region of locally varying conductivity
and electric field, as dictated by the effects of diffusion and Taylor dispersion.
Figure 38.18a also shows the effect of CTE on ITP preconcentration, where CTE were varied
from 1 to 100 mM. The LE was fixed at 1 M NaCl, and CS,initial was fixed at 1 nM. The KRF
model suggests that focused sample concentration is not a function of CS,initial or CTE . However, the
measurements show that CS,final increased for lower CTE (i.e., as conductivity ratio increases). High
LE-to-TE conductivity ratios (associated with low TE concentrations) increase the electric fields
in the TE and focused sample zones. High electric field leads to fast-focusing dynamics and high
electric Peclet numbers (Ueph a/D, where a is the characteristic channel scale) and thereby, high
CS,final , as the preconcentration process is less susceptible to dispersion. Higher electric fields would
CTE (mM)
100 101 102 103
102
102
106
CS, final (µM)

CS, final (µM)
CI
101 101
105
100 100
100 101 102 103 100 101 102 103
(a) CLE (mM) (b) CS, initial (pM)
FIGURE 38.18 Parametric variations of initial concentration profile. The nominal applied field was 220 V/cm.
CCD viewing area was centered 30 mm downstream of the channel intersection. (a) Maximum focused sample
concentration, CS,final , versus LE concentration, CLE and TE concentration, CTE . For variation of CLE , sample
analyte and TE were respectively 1 nM alexa fluor 488 and 5 mM HEPES. The regression coefficient, R2 , is
0.95. For variation of CTE , sample analyte and LE were respectively 1 nM alexa fluor 488 and 1 M NaCl.
R2 = 0.97. (b) CI and CS,final versus initial sample concentration, CS,initial . TE and LE were 5 mM HEPES
and 1 M NaCl, respectively. The regression coefficients for CI and CS,final are respectively 0.97 and 0.98.
eventually lower maximum focusing due to the effects of dispersion (e.g., due to Taylor dispersion
or Joule heating or both).
Next, initial sample concentrations, CS,initial were varied from 1 pM to 1 nM as shown in
Figure 38.18b. The LE and TE were, respectively, fixed at 1 M NaCl and 5 mM HEPES. The LE-to-
TE conductivity ratio was kept constant at 1.3×103 in an effort to decouple the dependence of the TE
zone electric field on this conductivity ratio. The data shows CI increases as CS,initial decreases. This
trend is roughly consistent with the KRF model. However, the dependence of CI on CS,initial is weaker
than the inversely proportional dependence predicted by the simple model (CI changes just under
two orders of magnitude while CS,initial changes three orders of magnitude). This discrepancy is also
apparent in the measurements of the maximum concentration, CS,final . The experimental data show
CS,final as a linear function of (although not directly proportional to) CS,initial , which is not attributable
to changes in local field in the TE zone. We again hypothesize that this discrepancy between the
simple model and observations is due directly to the effects of dispersion. Dispersed interfaces of
finite width cause a final sample concentration to be a function of initial sample concentration.

These experimental parametric studies yield important insight into key ITP stacking parameters and
suggest strategies for optimizing ITP in practice. Suppression of EOF to minimize Taylor dispersion
is a key component. Also important are high LE concentration and low initial sample concentration
to maximize achievable concentration increase, and the implementation of a single-column ITP
configuration (where initially, there is a single interface between the LE and the TE/sample mixture) to
inject a large effective sample width. As shown in Figure 38.18b, these simple strategies derived from
earlier scaling arguments achieve ITP preconcentration with final-to-initial sample concentration
ratios exceeding one million.
An example implementation of these strategies is shown in Figure 38.19. These are results of a
tITP assay (again, ITP followed by injection of LE ions on the cathode side of the TE to interrupt
ITP and initiate CE), which uses the single interface ITP configuration described schematically in
Figure 38.15. The figure shows two example separations of 100 aM (100 attomolar) concentrations
30
20
CS (pM)
10
0
72 73 74 75 76 77 78
CS (pM) 30
20
10
0
72 73 74 75 76 77 78
Time (s)
FIGURE 38.19 Two 100 aM sample electropherograms of ITP/CE separation of Alexa Fluor 488 (the peaks
near 73.5 s) and Bodipy (peaks near 76.5 s). A glass microchip (microchannel cross-sectional dimensions are
50 µm wide and 20 µm deep) and 60× water immersion objective (N.A. = 0.9) were used. The detector was
located 30 mm downstream of the injection region.
each of Alexa Fluor 488 and Bodipy, detected 30 mm after injection with 40 s ITP preconcentration
under a nominal field of 220 V/cm. LE and TE were 600 mM NaCl and 5 mM HEPES, respectively.
The concentrations measured after injection, ITP preconcentration, and separation are, respectively,
21 and 16 pM for Alexa Fluor 488 and bodipy, as averaged across five realizations. This experiment
achieves a concentration increase of 2.1 × 105 fold relative to the initial sample concentration of
100 aM. This 100 aM sensitivity is, to our knowledge, the highest ever reported sensitivity for an
electrophoresis experiment.

Dispersion effects arising from nonuniform fluid motion have detrimental effects on CE sensitivity
and resolution. Quantification and minimization of dispersion effects is especially important for
sample stacking and focusing techniques. In preconcentration methods, EOF of heterogeneous buffer
systems give rise to internal pressure gradients and strong dispersive forces. Generalized Taylor
dispersion analysis is a powerful approach for quantifying these effects. More importantly, scaling,
analytical solutions, and numerical solutions of Taylor dispersion provide unique insight. This can
guide the choices of separation voltage, EOF suppression strategy, and channel shape, width, and
depth for a given target analyte and buffer chemistry. The goal of such models and understanding
is the optimization of preconcentration efficiency and resolution across a broad range of physical
regimes and techniques.
Further improvements in the predictive capability of the stacking and focusing models presented
here can be made by accounting for the 3D effects of typical D-shaped, wet-etched channel cross
sections. More sophisticated dispersion analyses of the unsteady three-dimensional velocity field
should also be carried out. This is particularly important for the high electric field regimes of TGF
and ITP, where the sample plug widths become on the order of, or smaller than, the channel width.
More comprehensive models should aid in the systematic design and optimization of CE separations
under a wide range of conditions.
REFERENCES
1. Vlastimil Hruska, M.J., Bohuslav Gas, Simul 5—Free dynamic simulator of electrophoresis.
Electrophoresis, 2006, 27: 984–991.
2. Taylor, G., Dispersion of soluble matter in solvent flowing slowly through a tube. Proc. Roy. Soc. A,
1953, 225: 186–203.
3. Bharadwaj, R., Santiago, J.G., and Mohammadi, B., Design and optimization of on-chip capillary
electrophoresis. Electrophoresis, 2002, 23: 2729–2744.
4. Bharadwaj, R. and Santiago, J.G., Dynamics of field-amplified sample stacking. J. Fluid Mech., 2005,
543: 57–92.
5. Aris, R., On the dispersion of a solute in a fluid flowing through a tube. Proc. Roy. Soc. London,
Ser. A, 1956, 235: 67–77.
6. Dutta, D. and Leighton, D.T., Dispersion reduction in pressure-driven flow through microetched
channels. Anal. Chem., 2001, 73: 504–513.
7. Brenner, H., A general theory of Taylor dispersion phenomena. PCH PhysicoChem. Hydrodyn., 1980,
1: 91–123.
8. Stone, H.A. and Brenner, H., Dispersion in flows with streamwise variations of mean velocity: Radial
flow. Ind. Eng. Chem. Res., 1999, 38: 851–854.
9. Griffiths, S.K. and Nilson, R.H., Hydrodynamic dispersion of a neutral nonreacting solute in
electroosmotic flow. Anal. Chem., 1999, 71: 5522–5529.
10. Pennathur, S. and Santiago, J.G., Electrokinetic transport in nanochannels. 1. Theory. Anal. Chem.,
2005, 77: 6772–6781.
11. Griffiths, S.K. and Nilson, R.H., Charged species transport, separation, and dispersion in nanoscale
channels: Autogenous electric field-flow fractionation. Anal. Chem., 2006, 78: 8134–8141.
12. Probstein, R.F., Physicochemical Hydrodynamics: An Introduction, 2nd ed. 1994, New York: John
Wiley & Sons, p. 400.
13. Lin, H., Storey, B.D., Oddy, M.H., Chen, C.-H., and Santiago, J.G., Instability of electrokinetic
microchannel flows with conductivity gradients. Phys. Fluids, 2003, 16: 1922–1935.
14. Chen, C.H., et al., Convective and absolute electrokinetic instability with conductivity gradients.
J. Fluid Mech., 2005, 524: 263–303.
15. Deen, W.M., Analysis of Transport Phenomena. 1998, New York: Oxford University Press.
16. Santiago, J., Electroosmotic flows in microchannels with finite inertial and pressure forces. Anal.
Chem., 2001, 73: 2353–2365.
17. Storey, B.D., et al., Electrokinetic instabilities in thin microchannels. Phys. Fluids, 2005, 17: 1922–
1935.
18. Lin, H., et al., Instability of electrokinetic microchannel flows with conductivity gradients. Phys.
Fluids, 2004, 16: 1922–1935.
19. Gas, B. and Kenndler, E., Dispersive phenomena in electromigration separation methods. Elec-
trophoresis, 2000, 21: 3888–3897.
20. Ajdari, A., Bontoux, N., and Stone, H.A., Hydrodynamic dispersion in shallow microchannels: The
effect of cross-sectional shape. Anal. Chem., 2006, 78: 387–392.
21. Anderson, J.L. and Idol, W.K., Electroosmosis through pores with nonuniformly charged walls. Chem.
Eng. Comm., 1985, 38: 93–106.
22. Jung, B., Bharadwaj, R., and Santiago, J.G., On-chip million fold sample stacking using single
interface isotachophoresis. Anal. Chem., 2006, 78: 2319–2327.
23. Huang, X.H., Gordon, M.J., and Zare, R.N., Current-monitoring method for measuring the
electroosmotic flow-rate in capillary zone electrophoresis. Anal. Chem., 1988, 60: 1837–1838.
24. Santiago, J.G., Wereley, S.T., Meinhart, C.D., Beebee, D.J., and Adrian, R.J., A particle image
velocimetry system for microfluidics. Exp. Fluids, 1998, 25: 316–319.
25. Devasenathipathy, S. and Santiago, J.G., Electrokinetic flow diagnostics, in Micro- and Nano-Scale
Diagnostic Techniques. 2002, Ed. K.S. Breuer, New York: Springer Verlag.
26. Probstein, R.F., Physicochemical Hydrodynamics. 1994, New York: John Wiley & Sons.
27. Hunter, R.J., Zeta Potential in Colloid Science. 1981, London: Academic Press.
28. Kirby, B.J. and Hasselbrink, E.F., Zeta potential of microfluidic substrates: 2. Data for polymers.
29. Kirby, B.J. and Hasselbrink, E.F., Zeta potential of microfluidic substrates: 1. Theory, experimental
techniques, and effects on separations. Electrophoresis, 2004, 25: 187–202.
30. Yao, S., et al., Porous glass electroosmotic pumps: Design and experiments. J. Coll. Interf. Sci., 2003,
268: 143–153.
31. Burgi, D.S. and Chien, R.L., Optimization of sample stacking for high performance capillary
electrophoresis. Anal. Chem., 1991, 63: 2042–2047.
32. Xuan, X.C. and Li, D.Q., Analytical study of Joule heating effects on electrokinetic transportation in
capillary electrophoresis. J. Chromatogr. A, 2005, 1064: 227–237.
33. Tang, G.Y., et al., Modeling of electroosmotic flow and capillary electrophoresis with the joule heat-
ing effect: The Nernst–Planck equation versus the Boltzmann distribution. Langmuir, 2003, 19:
10975–10984.
34. Burgi, D.S., Large-volume stacking of anions in capillary electrophoresis using an electroosmotic
flow modifier as a pump. Anal. Chem., 1993, 65: 3726–3729.
35. Burgi, D.S. and Chien, R.L., Optimization in sample stacking for high-performance capillary
36. Chien, R.L. and Burgi, D.S., Field amplified sample injection in high-performance capillary
electrophoresis. J. Chromatogr., 1991, 559: 141–152.
37. Chien, R.L. and Burgi, D.S., Field amplified polarity-switching sample injection in high-performance
capillary electrophoresis. J. Chromatogr., 1991, 559: 153–161.
38. Chien, R.L. and Burgi, D.S., Sample stacking of an extremely large injection volume in high-
performance capillary electrophoresis. Anal. Chem., 1992, 64: 1046–1050.
39. Chien, R.L., Sample stacking revisted: A personal perspective. Electrophoresis, 2003, 24: 486–497.
40. Chien, R.L. and Burgi, D.S., Field amplified sample injection in high-performance capillary
electrophoresis. J. Chromatogr., 1991, 559: 141–152.
41. Chien, R.L. and Burgi, D.S. On-column sample concentration using field amplification in CZE. Anal.
Chem., 1992, 64: A489–A496.
42. Jung, B., Bharadwaj, R., and Santiago, J.G., Thousand-fold signal increase using field amplified
sample stacking for on-chip electrophoresis. Electrophoresis, 2003, 24: 3476–3483.
43. Jacobson, S. and Ramsey, J. Microchip electrophoresis with sample stacking. Electrophoresis, 1995,
16: 481–486.
44. Chien, R.L., Mathematical modeling of field amplified sample injection in high performance capillary
45. Zhang, C.X. and Thormann, W., Head-column field-amplified sample stacking in binary system capil-
lary electrophoresis. 2. Optimization with a preinjection plug and application to micelllar electrokinetic
chromatography. Anal. Chem., 1998, 70: 540–548.
46. Jacobson, S.C., Ermakov, S.V., and Ramsey, J.M., Minimizing the number of voltage sources and
fluid reservoirs for electrokinetic valving in microfluidic devices. Anal. Chem., 1999, 71: 3273–3276.
47. Jacobson, S.C., et al., High-speed separations on a microchip. Anal. Chem., 1994, 66: 1114–1118.
48. Jacobson, S.C., et al., Effects of injection schemes and column geometry on the performance of
microchip electrophoresis devices. Anal. Chem., 1994, 66: 1107–1113.
49. Grushka, E., McCormick, R.M., and Kirkland, J.J., Effect of temperature-gradients on the efficiency
of capillary zone electrophoresis separations. Anal. Chem., 1989, 61: 241–246.
50. Kirby, B.J., Zeta potential of microfluidic substrates: 1. Theory, experimental techniques, and effects
on separations. Electrophoresis, 2004, 25: 187–202.
51. Yao, S., Hertzog, D.E., Zeng, S., Mikkelsen, J.C., and Santiago, J.G., Porous glass electroosmotic
pumps: Design and experiments. J. Coll. Interf. Sci., 2003, 268: 143–153.
52. Preisler, J. and Yeung, E.S., Characterization of nonbonded poly(ethylene oxide) coating for capillary
electrophoresis via continuous monitoring of electroosmotic flow. Anal. Chem., 1996, 68: 2885–2889.
53. Hjerten, S. and Zhu, M.D., Adaptation of the equipment for high-performance electrophoresis to
isoelectric-focusing. J. Chromatogr., 1985, 346: 265–270.
54. Righetti, P.G., Isoelectric Focusing: Theory, Methodology, and Applications, 1st ed. 1983, New York:
Amsterdam, xv, 386 pages.
55. Herr, A.E., et al., On-chip coupling of isoelectric focusing and free solution electrophoresis for
multidimensional separations. Anal. Chem., 2003, 75: 1180–1187.
56. Ross, D. and Locascio, L.E., Microfluidic temperature gradient focusing. Anal. Chem., 2002, 74:
2556–2564.
57. Balss, K.M., et al., DNA hybridization assays using temperature gradient focusing and peptide nucleic
acids. J. Am. Chem. Soc., 2004, 126: 13474–13479.
58. Balss, K.M., et al., Simultaneous concentration and separation of enantiomers with chiral temperature
gradient focusing. Anal. Chem., 2004, 76: 7243–7249.
59. Balss, K.M., et al., Micellar affinity gradient focusing: A new method for electrokinetic focusing.
J. Am. Chem. Soc., 2004, 126: 1936–1937.
60. Kamande, M.W., et al., Simultaneous concentration and separation of coumarins using a molecular
micelle in micellar affinity gradient focusing. Anal. Chem., 2007, 79: 1791–1796.
61. Ghosal, S., Electrokinetic flow and dispersion in capillary electrophoresis. Annu. Rev. Fluid Mech.,
2006, 38: 309–338.
62. Huber, D.E. and Santiago, J.G., Taylor-Aris dispersion in temperature gradient focusing. Electrophore-
sis, 2007, 28: 2333–2344.
63. Kim, S.M., et al., Low-power concentration and separation using temperature gradient focusing via
Joule heating. Anal. Chem., 2006, 78: 8028–8035.
64. Van Milligen, B.P., et al., On the applicability of Fick’s law to diffusion in inhomogeneous systems.
Eur. J. Phys., 2005, 26: 913–925.
65. Ghosal, S., Lubrication theory for electro-osmotic flow in a microfluidic channel of slowly varying
cross-section and wall charge. J. Fluid Mech., 2002, 459: 103–128.
66. Huber, D.E., Transport and Dispersion in Microfluidic Temperature Gradient Focusing, Ph.D Thesis,
2006, Stanford University.
67. Hoebel, S.J., et al., Scanning temperature gradient focusing. Anal. Chem., 2006, 78: 7186–7190.
68. Everaerts, F.M., Beckers, J.L., and Verheggen, T.P.E.M., Isotachophoresis: Theory, Instrumentation,
and Applications. 1976, New York: Amsterdam.
69. Martin, A.J.P. and Everaert, F.M., Displacement electrophoresis. Anal. Chim. Acta, 1967, 38: 233.
70. Everaerts, F.M. and Van der Put, A.J.M., Isotachophoresis—Separation of amino acids. J. Chromatogr.,
1970, 52: 415.
71. Holloway, C.J. and Pingoud, V., The analysis of amino acids and peptides by isotachophoresis.
72. Bruchelt, G.G., Niethammer, D.D., and Schmidt, K.K.H., Isotachophoresis of nucleic acid
constituents. J. Chromatogr., 1993, 618: 57–77.
73. Delmotte, P., Analysis of complex protein mixtures by capillary isotachophoresis. Sci. Tools, 1977,
24: 33.
74. Schmitz, G. and Möllers, C., Analysis of lipoproteins with analytical capillary isotachophoresis.
75. Everaerts, F.M., et al., Determination of heavy metals by isotachophoresis. J. Chromatogr. A, 1985,
320: 263–268.
76. Sollenberg, J., Analytical isotachophoresis in biological monitoring of exposure to industrial
chemicals. J. Chromatogr. A, 1991, 545: 369–374.
77. Gebauer, P. and Bocek, P., Recent progress in capillary isotachophoresis. Electrophoresis, 2000, 21:
3898–3904.
78. Krivankova, L., Gebauer, P., and Bocek, P., Some practical aspects of utilizing the online combination
of isotachophoresis and capillary zone electrophoresis. J. Chromatogr. A, 1995, 716: 35–48.
79. Wainright, A., et al., Sample preconcentration by isotachophoresis in microfluidic devices. J. Chro-
matogr. A, 2002, 979: 69–80.
80. Xu, Z.Q., et al., Performance of electrokinetic supercharging for high-sensitivity detection of DNA
fragments in chip gel electrophoresis. Electrophoresis, 2004, 25: 3875–3881.
81. Mosher, R.A., Saville, D.A., and Thormann, W., The Dynamics of Electrophoresis. 1992: VCH
Weinheim New York.
82. Kohlrausch, F., Ueber concentrations-Verschiebungen durch Electrolyse in Inneren Von Losungen
and losungsgemischen. Ann. Phys. (Leipzig), 1897, 62: 209–239.
83. Martin, A.J.P. and Everaert, F.M., Displacement electrophoresis. Proc. Roy. Soc. Lond. Ser. A—Math.
Phys. Sci., 1970, 316: 493.
84. Svoboda, M. and Vacik J., Capillary electrophoresis with ultra violet detection some quantitative
aspects, Journal of Chromatography A, 1976, 119: 539–547.
85. Nagyova, I. and Kaniansky, D., Discrete spacers for photometric characterization of humic acids
separated by capillary isotachophoresis. J. Chromatogr. A, 2001, 916: 191–200.
86. Wainright, A., et al., Preconcentration and separation of double-stranded DNA fragments by
electrophoresis in plastic microfluidic devices. Electrophoresis, 2003, 24: 3784–3792.
87. Thormann, W., Description and detection of moving sample zones in zone electrophoresis: Zone
spreading due to the sample as a necessary discontinuous element. Electrophoresis, 1983, 4: 383–390.
88. Beckers, J.L., Steady-state models in electrophoresis: From isotachophoresis to capillary zone
electrophoresis. Electrophoresis, 1995, 16: 1987–1998.
89. Hjertén, S.S., Zone broadening in electrophoresis with special reference to high-performance elec-
trophoresis in capillaries: An interplay between theory and practice. Electrophoresis, 1990, 11:
665–690.
90. Saville, D.A. and Palusinski, O.A., Theory of electrophoretic separations. 1. Formulation of a
mathematical-model. Aiche J., 1986, 32: 207–214.
91. Saville, D.A., The effects of electroosmosis on the structure of isotachophoresis boundaries.
39 The Mechanical Behavior of
Films and Interfaces in
Microfluidic Devices:
Implications for Performance
and Reliability
Matthew R. Begley and Jennifer Monahan
CONTENTS
39.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1122

39.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1123
39.2.1 The Mechanics of Interface Failure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1124
39.2.2 Overview of the Mechanics Describing Deformable Plates/Films . . . . . . . . . . . . . . . . 1126
39.3.1 Interface Fracture Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1127
39.3.2 Potential Energy and Pressure–Deflection Relationships . . . . . . . . . . . . . . . . . . . . . . . . . . 1129
39.3.2.1 Small Deformation Plate Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1129
39.3.2.2 Large Deformation Membrane Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1131
39.4 Practical Applications and Development Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1131
39.4.1 Debonding of Multilayers from Interface Flaws at Edges . . . . . . . . . . . . . . . . . . . . . . . . . 1131
39.4.2 Debonding of Pressurized Films Covering Long Channels . . . . . . . . . . . . . . . . . . . . . . . 1134
39.4.2.1 Plate Behavior: L >∼10 h and δmax < h . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1135
39.4.2.2 Membrane Behavior: L >∼10 h and δmax > h . . . . . . . . . . . . . . . . . . . . . . . . . . 1136
39.4.3 Pressure–Deflection Relationships for Valve Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1137
39.4.3.1 Long Capped Channels: Plane-Strain Deformation . . . . . . . . . . . . . . . . . . . . . 1137
39.4.3.2 Circular Films . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1139
39.4.4 Design of Check Valves for Specific Adhesion Energy and/or Dimensions . . . . . . 1139
39.4.5 Fluidic Capacitance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1141
39.4.6 Material Property Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1142
39.4.6.1 Indentation Modulus Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1142
39.4.6.2 Bulge Testing to Determine Modulus and Interface Toughness . . . . . . . . 1144
39.4.6.3 Peel Tests to Determine Interface Toughness . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1146
39.5 Future Prospects and Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1147
Nomenclature/Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1148
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1149
1121
39.1 INTRODUCTION
As electrophoretic separations move beyond traditional capillary systems to include microfabri-
cated “lab-on-a-chip” technology, researchers must consider the mechanical response of devices, as
dictated by the materials and methods involved in channel fabrication. Microfluidic chips created
using planar lithography (e.g., [1–4]) enable increasingly complex separations, such as multidimen-
sional chromatography [5–9] and on-chip sample preparation [10,11]. Unfortunately, lithographic
techniques generate open channels, which require a subsequent bond step to produce an enclosed
capillary network suitable for electrophoretic separations. In contrast to tubular capillaries comprised
of a single material piece, the bonded interface of a microchip separation channel introduces a new
failure mechanism, which limits electrophoretic or chromatographic methods that rely on pressure
(i.e., capillary electrokinetic chromatography (CEC) [12–15], microchip isoelectric focusing (mIEF)
[16], high ionic strength buffers [17], or polymer-based separations [18]).
Relatively recent additions of external and/or integrated features that control fluid motion (e.g.,
valves [19–21]) further increase the need to understand and quantify the mechanics controlling inter-
face strength. This is especially true as hybrid chips pairing different types of materials (e.g., glass
and polydimethylsiloxane (PDMS)) become increasingly prevalent; the bonding of surfaces with
elastic and thermal expansion mismatch can induce significant stresses, which influence the perfor-
mance and reliability of the final device. This chapter is designed to provide separation scientists with
an introduction to interface and film mechanics that can be used to promote successful microfluidic
design and fabrication.
Interface stability affects several aspects of lab-on-chip development. Successful device fab-
rication requires consistent bonding between different layers, often of different materials, and
this remains among the most challenging obstacles to device development. Figure 39.1 provides
schematic illustrations of several features/geometries that play increasingly important roles in
microfluidic separation devices. Figure 39.1a is a top view of a typical multilayer assembly; a
layer with lithographically patterned channels is “sealed” by bonding additional layers. The inter-
face formed by the layers is typically weaker than the layers themselves, such that interface failure
defines device reliability. Figure 39.1b is a side view of the interface failure shown in Figure 39.1a;
the corner formed by bonding a capping layer to the layer with etched channels serves as the initia-
tion site for failure. Device performance may further include deformable elements interacting with
adjacent elements (e.g., valving). Such dynamic elements may require controlled debonding (e.g.,
valve release [19–21]) or site-specific bonding for ideal performance (e.g., check valves [22,23]).
Figures 39.1c and 39.1d illustrate the side views of two common multilayer valve assemblies; the
clear layer is a deformable element that is pushed or pulled via external pressure actuation to allow
(Figure 39.1c) or prevent (Figure 39.1d) fluid motion in through the channels.
Thus, device reliability for many systems is dictated by interface stability, which in turn dic-
tates the maximum allowable conditions and pressures. The use of pressure in microfluidic chips
naturally introduces a competition between pressure limitations due to interface strength and that
required for operation. Consider two examples: (1) filling a microchip channel with a viscous sepa-
ration polymer (e.g., for DNA sequencing [24]) and (2) creating an integrated valve that is actuated
using pressure [19–21]. Both require elevated pressures, which must be balanced against the possi-
bility of breaking the chip apart. In designing features for these diverse devices, two fundamental
questions arise:
I. What are the physical parameters that affect interface stability?

II. How do geometry, material properties and device pressures govern performance and
reliability?
The principle motivation for this chapter is to establish the framework to predict device performance
and reliability a priori, by detailing the mechanics underlying the behavior of typical geometries.
The Mechanical Behavior of Films and Interfaces in Microfluidic Devices 1123
Reliability of Valve/diode
capped channels performance and reliability
Actuation pressure
−pa
Open
config.
(a) (c)
p1 p2
Delamination
Etched layer Top view crack fronts
Actuation pressure
Deformable
Capping layer
layer
Ef , νf pa
Pressure, p
(b) (d)
Closed p1 p2
config.
Es, νs
Side view
FIGURE 39.1 Schematic illustrations of common microfluidic features utilized in microchip separation
devices: (a) top view of a multilayer assembly of microchannels formed by capping a layer with etched channels,
(b) side view of micropatterned channel with capping film layer, (c) side view of a vacuum-driven three-layer
negative pressure valve, (d) side view of a positive pressure valve.
To complement this, we describe basic materials characterization experiments used to determine the
relevant parameters. This combination can be used to optimize proven device configurations and
develop new strategies to improve fabrication, performance and reliability.
39.2 BACKGROUND
Microfluidic separation devices have increased in complexity beyond the original cross-t design of
Manz and coworkers (e.g., [25,26]) to include a wide variety of designs with intricate intersecting
channels. The bulk of these devices are fabricated in glass or PDMS. Several reviews (and previous
chapters in this book) discuss the methods used in the fabrication of microfluidic devices [27,28].
Nearly all are based on planar lithography and “soft” lithographic techniques to produce channels
embedded in the surface of a substrate (e.g., [29,30]). To create an enclosed microcapillary network
suitable for electrophoretic separations, these channels must be bonded to a cover plate(s) as shown
in Figure 39.1a. The bonding of two different layers creates an interface that is typically weaker than
either layer; hence, the “seam” that runs along the edge of the channel acts as an initiation site for
chip failure. This is shown schematically in Figure 39.1a (top view) and Figure 39.1b (cross-section
view). Interface failure usually corresponds to device failure, as fluid from the channel is lost to the
interface “crack”; with sufficient pressure (or volume injected into the channel), the entire interface
may debond.
Here we will focus on two material systems common to chip based electrophoretic separations.
The first, glass-on-glass devices, are common for “cross-t” and gated electrokinetic injections. Such
devices depend on a high temperature anneal to facilitate bonding between the cover plate and
substrate with etched channels [27,28]. The successful high temperature anneal of two pieces of
glass requires flat, defect free surfaces and the uniform application of pressure during fabrication.
In many ways, successful bonding of glass-on-glass devices is considered a “black art”: while
strong bonds produce robust chips that are ideal for pressurized applications, fabrication yield can
be depressingly low, especially in the hands of a novice. The second class of chips, glass-on-PDMS
hybrid devices, are widely utilized because of their ease in fabrication/assembly and to produce
integrated elastomeric valves, such as those shown in Figures 39.1c and 39.1d. Depending on the
application, glass-PDMS devices are either weakly bonded using reversible van der Waals forces,
or more strongly bonded through oxidative modification of the material surfaces prior to bonding.
The physical demands placed on these devices can be quite different due to the rigid nature of
glass–glass interfaces and extreme deformation (arising from large pressures relative to the elastic
modulus) that may occur with PDMS hybrid chips. Regardless of the material(s) or fabrication tech-
niques, the performance and reliability of all microfluidic chips designed for CE separations depend
upon controlling the interface bond between neighboring substrates. Device operation (either inter-
face failure or desired debonding) is governed by (1) the physics of interface debonding and (2) the
relationship between loading (e.g., pressure and/or temperature) and chip deformation. The two are
strongly interrelated, because both are essentially governed by energy stored by elastic deformation.
39.2.1 THE MECHANICS OF INTERFACE FAILURE

The mechanics of interface failure, while often complex, has been firmly established in the context of
thin film multilayers commonly found in microelectronics and protective coatings. This large body
of work serves as the foundation for the present chapter; it is succinctly summarized by the review
article by Hutchinson and Suo [31] and covered in great detail in the book by Freund and Suresh
[32]. We present the most basic framework needed to quantify the performance of microfluidic
devices, and summarize results for geometries and materials common to such devices. The focus is
on prediction of debonding initiated from sharp changes in geometry, as illustrated schematically in
Figure 39.2.∗ For these systems, the act of debonding or delaminating under pressure is simplified
by considering the process as the formation and/or propagation of a crack at the interface.
As a first approximation for analyzing interfaces, an elementary “strength of materials” approach
is to determine the maximum stress in the component and then compare its value to the strength of
the material (or interface), as measured via a conventional uniaxial tension test. Unfortunately, this
approach is not applicable in the presence of material discontinuities, such as an interface that is
only partially bonded or a corner formed by bonding two materials (see Figure 39.2a). If one obtains
a complete solution for the stress distribution, the stress field diverges to infinity as the discontinuity
is approached: that is, the stress diverges to infinity as distance from the crack tip becomes smaller
and smaller. This is shown schematically in Figure 39.2b (refer to log-log scale on the right-hand
side). Hence, the discontinuity represents a singularity in the stress distribution. This nonphysical
result (i.e., an infinite stress at a crack tip) is a consequence of assuming elastic behavior applies at
Delamination log σ
Ef, νf crack front n
1
σ
Pressure, p log r
Es,νs Energy release rate, G
FIGURE 39.2 Schematic illustration of the stress distribution at the tip of a delamination crack: the presence
of a stress singularity necessitates an energy-based approach.
∗ “Delamination” is commonly used to describe the debonding of thin films or bonded stacks of thin films (i.e., multilayers);
hence, in the present context, debonding and delamination refer to the same phenomena.
an arbitrarily small length-scale. To get around these conceptual difficulties, linear elastic fracture
mechanics (LEFM) theory [33] is a very successful approach to predicting crack growth without
resorting to molecular simulations. That is, one can successfully predict failure using macroscopic
continuum solutions. It is interesting to note that for materials with limited mechanisms for nanoscale
deformation,∗ continuum elasticity solutions are applicable to impressively small scales, that is,
several nanometers. Hence, it is not surprising that they can serve as the basis for predicting failure
in predominantly elastic systems.
With the exception of an infinitely long material interface, any mathematically sharp geometry
change introduces a singularity, for example, right-angled corners, the edge of a debonded region, or
material interfaces that intersect a free edge. The form of this singularity depends on the angle of the
junction (e.g., the 90◦ corner junction shown in Figure 39.2a), as well as the elastic mismatch between
the materials comprising the junction (e.g., [34–36]). At the moment, the mechanics of crack stability
are considerably more established than that required to predict debond initiation from a bimaterial
corner junction. That is, the criteria needed to predict the advance of a sharp interface delamination
crack is more clearly defined than that needed to predict its initiation from a right-angled corner.
Debond initiation from blunt corners is a critical problem that requires continued study.† However,
as luck would have it, when the film and substrate have very large elastic mismatch—such as for
PDMS-glass systems—it is reasonable to assume that the substrate is “rigid.” In this limit, the
corner initiation problem reduces to that of an interface crack: the entire span shown in Figure 39.2a
corresponds to the “crack,” and well-established interface fracture mechanics is applicable. This
is because the substrate does not deform, such that it is immaterial whether or not the substrate
geometry forms a perfect 90◦ corner or is partially debonded as shown in Figure 39.2a.
In systems where the elastic moduli of the layers are comparable—critically, glass-glass or any
homogenous system—one must distinguish between the corner initiation geometry and the partially
debonded region shown in Figure 39.2a. This is because the deformation of the substrate lying
between the original corner and leading edge of the debond region (i.e., crack tip) may influence the
stress distribution at the edge of the crack tip. When the distance between the corner and the crack
tip is sufficiently large, the influence of geometry of the channel (and hence corner) can be neglected
and one can again treat the entire debonded length as a crack between geometrically uniform layers.
This limit is reached when the distance between the corner and debond edge is several times greater
than the film thickness (or thickness of the more compliant layer) (see [37] for the related study of
delamination from vertical edges of a two layer system). The fact that even short cracks (measured
from the edge of the corner to the debond crack tip) might lead to undesirable behavior—such as
fluid or pressure loss—highlights the need for additional mechanics analyses of the initiation and
short-crack propagation problems for elastically similar materials. Such analyses are beyond the
scope of this introduction, and hence, we must assume here that the debond crack tip is located at
least several film thicknesses from the corner.
The stress singularity implies that stress-based failure criteria are not feasible and one must adopt
energy frameworks to describe the conditions that lead to debonding.‡ The central concept that is
that the crack front advances when the energy available to drive the crack reaches a critical value,
often referred to as the interface toughness. The energy available to drive the crack, or energy release
rate (ERR), is dictated by the potential energy of the system. This is comprised of the strain energy
stored in the deformed structure and the work done by applied loads. The critical value of the ERR
∗ Glass is a classic example: atomic bonds can stretch (elasticity) or rupture, but typically do not reform. In metals, atomic
bonds can “slide” and reform, a phenomenon described by dislocation theory; this gives rise to energy dissipation, which
makes it much more difficult to propagate a crack in metals.
† Significant inroads into the corner initiation problem have been established (e.g., [34–36]; these studies, however, require
more extensive discussion of mechanics concepts that are beyond the scope of this work.
‡ Or, one can use equivalent stress-based frameworks that deal with the nature of the singularity rather than a “maximum
stress”: these approaches use the stress intensity factor, which is essentially the amplitude of the singularity (see Section 39.3.1.)
(a) h~α h (b) (c)

h
δ δ
L L δ L
Half-space: Plate behavior: Membrane behavior:

L << h, L > 8h, δ < h L > 8h, δ > h
δ α (p/E)L δ α (pL4/(Eh3) δ α (p/Eh)1/3L4/3
FIGURE 39.3 Schematic illustration of three types of film response: (a) elastic half-space behavior, dominated
by bulk compression of the film, (b) plate behavior, dominated by bending deformation of the film, and
(c) membrane behavior, dominated by axial stretching of the film.
(which has units of N/m or J/m2 ) is the energy per unit area needed to create new crack surface and
is fundamentally dictated by the bonding at the interface. As such, it is often referred as the adhesion
energy. This chapter outlines the mechanics needed to compute the ERR for various combinations
of load and geometry, as well as that needed to design and interpret experiments for extracting the
interface toughness.
39.2.2 OVERVIEW OF THE MECHANICS DESCRIBING DEFORMABLE PLATES/FILMS

The relationships between geometry, external loading (i.e., pressures) and deformation has been an
area of intense study in the solid mechanics community for more than century; hence, our goal is
to provide a condensed overview of the theory and solutions that apply specifically to microfluidic
applications. A number of classical mechanics results are relevant to microfluidic devices and the
trick is to determine which is applicable. The key consideration is the “aspect ratio” of the device
that relates the pressurized span, L, to the film thickness, h, as shown in Figure 39.3. Three different
types of deformation “modes” are amenable to closed-form analytical solutions. Figure 39.3a depicts
small aspect ratios L/h: very thick structures respond as half-spaces, in that the deformation does not
depend on the film thickness, which is taken to be semi-infinite. Figure 39.3b depicts moderate to large
aspect ratios L/h: moderately thin structures respond as plates, in that bending deformation of the film
dominates film stretching. Figure 39.3c depicts large aspect ratios L/h: very thin structures respond
as membranes, in that axial stretching deformation of the film dominates the effects of bending.∗
The scaling relationships shown in Figure 39.3 apply regardless of the out-of-plane geometry
(e.g., elliptical or rectangular), which only affects a constant prefactor. In this chapter, we review
the relevant analytical solutions that bracket the range of behaviors, as well as numerical solutions
that can be used for intermediate aspect ratios. The focus is on providing guidelines that enable chip
designers to identify appropriate solutions for a given geometry and pressure magnitude.

Since a truly comprehensive treatment of the mechanics relevant to microfluidic devices is beyond
the scope of a single chapter, we provide a basic introduction that adopts several simplifications. First
and foremost, we assume that materials are linearly elastic; while, this is obviously objectionable
for elastomers at large strains, it allows for closed-form analytical solutions that can be used to
guide initial device design and broadly assess reliability. It is important to note that even when
∗ It should be noted that many refer to flexible thin films (commonly elastomers) as “membranes,” regardless of their
dimensions, presumably because they are easily deformed by relatively low pressures (relative to atmospheric pressure).
Here, we adopt the classical mechanics definition of membrane, that is, a stretch-dominated structure (as opposed to bending-
dominated).
deformation is large (i.e., displacements are significant compared to film thickness), strains may still
be small. Moreover, the loss of accuracy due to linearization of material response is oftentimes not
significant for strains less than ∼30% (e.g., [38]). Secondly, we assume that interface debonding
occurs at an asymptotically sharp crack front, as opposed to a bimaterial corner defined by a finite
angle. This approximation has no impact on glass–PDMS interfaces due to the elastic mismatch;
for interfaces formed between similar materials, this approximation is valid when the debond crack
tip is sufficiently far from corners, other crack tips and so on. Finally, we assume that the interface
is sufficiently weak to ensure that delamination cracks propagate at the interface, that is, interface
delaminations will not “kink” and turn to run in the materials on either side of the interface.
39.3.1 INTERFACE FRACTURE MECHANICS

Interfacial stability is predicted using the ERR, which is defined as the mechanical energy that
is released during an incremental increase in crack length. Crack advance occurs when the ERR
reaches a critical value, often referred to as the interface toughness, denoted here as i . Thus, the
crack growth criterion is G ≥ i . The ERR is defined as [33]
∂U
G≡− , (39.1)
∂A
where A is the area of the new surface formed by crack advance (dictated by the geometry of the
crack front) and U is the mechanical energy of the system. U is a function of the current crack size,
loads such as applied pressure or residual stress and the elastic properties of the system. When the
crack propagates with a straight crack front, the ERR can be written interms of the crack length and
the associated length of the crack front. The ERR thus has units of N/m, or J/m2 and hence, the
interface toughness has units of energy per unit area; for this reason, the toughness is often referred
to as the critical adhesion energy, or simply the adhesion energy. The interface toughness is governed
by the nature of bonding at the interface and should be considered a material property to be measured
for each system and fabrication approach.
For bimaterial interfaces, the interface toughness is also a function of the mode-mixity, which
describes the relative amounts of crack tip opening, sliding and tearing deformation [31,32].
Figure 39.4 illustrates the definitions of the three fundamental modes of crack growth. For straight
crack fronts with no deformation gradients in the direction parallel to the crack front, there is no
out-of-plane tearing and mode III is identically zero. The stress intensity factors KI and KII relate to
the contributions of opening and sliding displacements, respectively and hence are outcomes of the
solution to the crack tip elasticity problem. In essence, the stress intensity factors describe the ampli-
tude of the stress singularity shown schematically in Figure 39.2b. As such, they are dependent on
the applied load, crack geometry and elastic properties on either side of the interface. It is important
to note that bending of thin films near crack tips can introduce a significant mode II component; thus,
even though pressurized crack faces may seem to be “pure mode I,” the deformation that involves
relative sliding of the material near the interface crack tip, which implies that mode II is present.
It has been demonstrated that any two-dimensional bimaterial problem can be uniquely described
in terms of the two Dundur’s parameters, defined as
Ēf − Ēs Ef (1 + vs ) (1 − 2vs ) − Es (1 + vf ) (1 − 2vf )

α= , β= , (39.2)
Ēf + Ēs Ef (1 + vs ) (1 − vs ) + Es (1 + vf ) (1 − vf )

where Ē = E/ 1 − v2 and the subscripts f and s refer to the film and substrate, respectively. That is,
the behavior of any interface crack with identical Dundur’s parameters is independent of individual
elastic properties of the film and substrate. Dundur’s parameters for several material systems relevant
to microfluidic systems are shown in Table 39.1.
Mode I Mode II Mode III
FIGURE 39.4 Modes of crack loading: for the scenarios considered here, mode III is assumed to be zero due
to the absence of deformation gradients parallel to the crack front.
TABLE 39.1
Dundur’s Parameters for Several Interfaces Relevant to Microfludic Devices
Materials Forming the Interface Dundur’s Parameters
Material #1 E, v Material #2 E, v α β
Elastomer 1 MPa, 0.5 Glass 70 GPa, 0.2 ∼ −1 1.4 × 10−5
Elastomer 1 MPa, 0.5 Polymer 3 GPa, 0.4 ∼ −1 1.2 × 10−4
Polymer 3 GPa, 0.4 Glass 70, GPa, 0.2 –0.91 –0.28
Identical — Identical — 0 0
Note: Interchanging materials #1 and #2 simply changes sign of Dundur’s parameters.
Here, we limit our attention to scenarios where β ∼ 0, which essentially eliminates combina-
tions of materials with comparable—but not identical—stiffness. Note that β is identically zero for
identical materials. When β ∼ 0 the crack tip √fields are analogous to those in a homogenous material
(i.e., the crack tip stresses scale as σij ∼ 1/ r) and KI and KII dictate the contributions of opening
and sliding displacements, respectively (or normal and shear stresses). If β = 0, the opening and
sliding modes are coupled and crack advance criteria require additional parameters [31,32].
The simplification that β = 0 implies that crack advance can be predicted in the context of a
mixed-mode interface toughness, that is,

−1 KII
i = f (ψ), where ψ ≡ tan , (39.3)
KI
where KI and KII are computed from the elasticity solution to the given crack problem.
Since interface toughness depends strongly on the amount of mode-mixity, the ERR alone is not
sufficient to predict delamination; one must also compute the phase angle from the relevant elasticity
solution and have measurements that describe i = f (ψ).
However, measurements on a variety of systems have illustrated that the minimum interface
toughness, imin , is obtained for pure mode I. That is, the measured toughness for cracks dominated
by mode II (ψ → π/2) is generally much higher. Hence, assuming ψ ∼ 0 and adopting the crack
advance criterion G ≥ i (ψ = 0) = imin is generally conservative, in that it underestimates crack
stability. Moreover, experiments on interfaces between glassy polymers (e.g., PMMA) and glass
suggest i ≈ i (ψ = 0) for −15◦ < ψ < 60◦ [31,32].
If the interface toughness is known, the prediction of debonding thus essentially boils down to
computing the potential energy of the system, U and computing its derivative with respect to crack
length. The potential energy of the system is computed from the solution to the solid mechanics
problem at hand; this is described in the next section. For simple geometries, this can be done in
closed-form, as illustrated in Section 39.4.2. For more complex geometries, this is typically done
with commercial finite element analysis (FEA) software and many codes have a built-in option for
doing so.
39.3.2 POTENTIAL ENERGY AND PRESSURE–DEFLECTION RELATIONSHIPS

The relationship between loads, geometry and deformation is governed by the solution of partial
differential equations (PDEs) that are derived using three fundamental relationships: (1) equilibrium
conditions, expressed as linear PDEs involving up to six stress components, (2) strain-displacement
compatibility conditions, expressed as differential equations involve up to six strain components and
three displacements and (3) a constitutive description, expressed as an algebraic equation relating
the six stress and six strain components. The governing equations are completely linear for linear
elastic materials that experience relatively small displacements. For large deflections, the strain-
displacement relationships are nonlinear, such that numerical solutions are typically needed. The
deformation of linearly elastic membranes is a notable exception, in that several important nonlinear
solutions (that account for large-displacements) can be derived in closed form.
The stresses and displacements at the boundaries of the deformable domain (applied or enforced)
dictate the boundary conditions, which effectively determine the particular solution to the PDEs. If
the problem is truly three-dimensional (because geometry and/or loads vary in all three directions)
then numerical solutions are needed for all but the most elementary example cases. FEA is the de
facto choice for such scenarios and commercial codes abound; ABAQUS and ANSYS (software
programs) are arguably the most common. An example of a simple geometry that is nevertheless
a truly three-dimensional problem is a film capping an elliptical hole, that is, an elliptical plate or
membrane.
If the geometry and loads do not vary in the third dimension, the derivatives with respect to
the third coordinate are identically zero and the problem is inherently two-dimensional. Important
examples include long, straight-walled channels, or axisymmetric structures such as circular plates:
one can obtain the relevant solution by analyzing a two-dimensional “slice” taken from the full
structure. For simple boundary conditions, such scenarios are highly amenable to analytical solutions.
This is the focus of the remainder of this section.
39.3.2.1 Small Deformation Plate Behavior

Consider a film to a pressure distribution shown in Figure 39.5a. It is assumed that the variation in
displacement in the z-direction is zero, that is, εz = 0: this is referred to as plane-strain deformation.
(Note that this does not imply σz = 0.) First, consider platelike behavior, wherein the bending strain
po(x)
(a) po(x) (b)
y N x (x)
y
(x)
x h M z (x)
dx
u(x)
dx
FIGURE 39.5 (a) Schematic illustration of film subjected to an arbitrary pressure distribution, (b) a differential
element of the film showing the variables used for strain–displacement and equilibrium equations.
in the film dominates axial stretching. For small deformations relative to the film thickness, the only
relevant strain-displacement relationship is:
εxx (x, y) = δ (x) · y, (39.4)
where δ (x) is the second derivative of the vertical deflection with respect to the x-coordinate (i.e.,
the local curvature) and y is the distance from the centerline of the film.
The net internal moment acting at any cross-section (about the z-axis) is given by
h/2
Ēh3
M (x) = σxx (x, y) · y · dy = δ (x) , (39.5)
12
−h/2
where σxx = Ēεxx is used as the constitutive law. Ē = E/(1 − v2 ) is the plane-strain modulus, where
v is the Poisson’s ratio of the film. This assumes that the in-plane stress σxx dominates the vertical
stress σyy through most of the film thickness; that is, the structure translates the vertical pressure
into a linear bending stress. Vertical equilibrium of the differential element shown in Figure 39.5b
dictates that the moment distribution, vertical displacements and external pressure related as follows:
Ēh3 IV
M (x) = δ (x) = −po (x), (39.6)
12
where po (x) is the resultant pressure acting on the film (i.e., the superposition of pressure on top and
bottom). Thus, by combining a kinematic relationship between strain and displacement, a constitutive
law and equilibrium, a differential equation is obtained for the displacements of the film in terms of
the applied load.
One can integrate this result for any pressure distribution: the four integration constants are
determined by the boundary conditions at either end. Force balance applied to the differential element

of
Figure 39.5b also dictates that the resultant vertical shear force is given by V (x) = M (x) =
Ēh3 /12 δ (x). This completes the possible boundary conditions that can be used to solve for
unknown constants: for example, a clamped or bonded film implies zero displacement and slope,
that is, δ (x = 0) = δ (x = 0) = 0 and δ (x = L) = δ (x = L) = 0. Or, for free ends (i.e.,
cantilevers), the net moment and shear force acting at a free end is zero, such that δ (x = L) =
δ (x = L) = 0.
For the case of a uniform applied pressure (i.e., po (x) = po ), the total potential energy of the
system (for a slice across of the channel of width b) is given by
L h/2 L
U=b 1/2σxx (x, y) εxx (x, y) − bpo δ (x) dx, (39.7)
0 −h/2 0
where L is the length of the span that can deform. The first term is the strain energy in the deflected
film and the second is the external work done by the applied pressure. Note that the last term is
equivalent to the pressure times the volume of the bulge: this observation becomes important in
calculation the ERR for bulge tests under fixed volume (see Section 39.4.2). Thus, for any uniform
film that is subjected to constant pressure and dominated by bending deformation, the potential
energy is given by
L 3
Ēh 2
U=b δ (x) − po δ (x) dx. (39.8)
24
0
This result can be used with Equation 39.1 to determine the ERR for scenarios involving small
deflections; Section 39.4.2.1 provides an illustration.
39.3.2.2 Large Deformation Membrane Behavior

Similar elementary results can be derived for deformation in the membrane limit, that is, when
the centerline stretch of the film dominates bending. In this scenario, the strain-displacement
relationship is
2
εxx (x) = u (x) + 1/2 δ (x) , (39.9)
where u(x) is the displacement in the x-direction (see Figure 39.5). One can illustrate via energy
minimization that the strain and hence the resultant stress σxx (x) = Ēεxx (x) = σo , is uniform
along the membrane.∗ This stress is an unknown that must be determined. Additionally, the energy
minimization implies the following equilibrium condition:
hσo δ (x) = −p (x). (39.10)
Again, the solution by direction integration involves two constants; these constants, as well as the
unknown membrane stress, are dictated by the displacements of the end points. Noting that the
membrane stress is spatially uniform and integrating Equation 39.9 yields
L
σo L 1 2
u (L) = + δ (x) dx. (39.11)
Ē 2
0
For uniform applied pressure, the total potential energy of the membrane system is given by
L L
hσo2 bhLσo2
U=b − po δ (x) dx = − bpo δ (x) dx. (39.12)
2Ē 2Ē
0 0
Once again, the last term represents the pressure times the volume of the bulge, which is relevant
to interface fracture toughness measurements conducted by injecting a fixed volume (see Section
39.4.2.2).
39.4 PRACTICAL APPLICATIONS AND DEVELOPMENT

GUIDELINES
39.4.1 DEBONDING OF MULTILAYERS FROM INTERFACE FLAWS AT EDGES
This section addresses interface failure between two elastic plates bonded together, as shown in
Figure 39.6. Stresses are generated when the plates have different initial curvatures and/or different
thermal expansion properties; the resulting stored elastic energy promotes debonding. It is assumed
that two plates with different curvatures bond together after they have been pressed completely flat
and their interfaces have been aligned. Put another way, the material on each side of the interface
experiences relative sliding prior to bonding. This is an important assumption, because it implies that
each layer will be stressed in the bonded state even when their initial curvature, elastic properties
∗ One can derive equivalent expressions using force balances: however, the finite angles that are retained in membrane theory
(e.g., cos β ≈ 1 − (1/2) β 2 ) make such derivations cumbersome.
Rs = 1/κos
R f = 1/κof R = 1/κ
L
d
Note: R = L2/(2d)
(a) Initial curvatures (+κ up) (c) Deformed state of bonded bilayer
(b) Press flat, then bond (d) Edge delamination relieves stress
(calculate steady-state ERR)
FIGURE 39.6 Geometry and variables used in the analysis of bilayer deformation and debonding: (a) initial
configuration of curved plates, (b) it is assumed the plates are pressed flat prior to bonding, (c) resulting curved
bilayer with continuous total strain distribution, and (d) schematic illustration of edge debonding and stress
state in bonded region.
and thermal expansion coefficients are identical. This is a consequence of the fact that “opposite”
faces are bonded: for example, if both plates have the same curvature, the compression side of the
top plate will be bonded to the tensile side of the bottom plate (see Figure 39.6b).
The mechanical strain distributions in the thickness direction in each plate can be expressed as∗
εxx
f
(y) = εo − κ · y + κof (y − hs − hf /2) − θf , (39.15a)
and
εxx
s
(y) = εo − κ · y + κos (y − hs /2) − θs . (39.15b)
Here, κoi is the initial curvature of the ith layer (i.e., the layer traces a circular arc with radius
Roi = 1/κoi ) and the subscripts f and s refer to the film and substrate, respectively. Positive curvature
is taken as concave up: if the curvature of the two layers opposes one another, the individual layer
curvatures are of opposite sign. y is the
distance from the bottom of the substrate. The thermal strain
in the ith layer is defined as: θi ≡ αi Ti − Toi , where αi is the coefficient of thermal expansion and
Ti is the current temperature. Toi is the reference temperature at which there is no thermal strain,
that is, the temperature at which the layer is fabricated. εo is the axial stretch along the bottom of
the substrate (i.e., y = 0) and κ is the curvature of the bilayer after bonding; hence, εo and κ are the
solution variables of interest.
∗ The total strain is simply ε

tot (x) = εo − κ (x) · y: one subtracts the thermal strains from the total strain to get the mechanical
strain, that is, the strain used to determine stresses via σ = Ēε.
Absent any external loading on the bonded bilayer, the net resultant axial force, N and net
resultant moment, M, are zero. This implies
hs s +hf
h
N= σxx
s
(y) dy + σxx
f
(y) dy = 0, (39.16)
0 hs
and
hs s +hf
h
M= σxx
s
(y) · y dy + σxx
f
(y) · y dy = 0 (39.17)
0 hs
Substituting σxx
i (y) = Ē ε i (y) and performing the integration yield two linear equations for
i xx
the unknowns εo and κ. The stored elastic energy of the bilayer can be calculated by integrating the
strain energy density per unit volume, 1/2 σxxi ε i , throughout the volume, using the results for ε and
xx o
κ with Equation (39.15); integration over thickness yields the strain energy per unit area in the plane
of the bilayer.
The ERR for an edge crack such as that shown in Figure 39.6d can be trivially calculated
using the elastic energy. Here, we assume that interface failure relieves all stresses in the layers,
that is, they return to their undeformed configuration for portions of the plate that are no longer
bonded.∗ When the length of the interface crack greatly exceeds either layer thickness, edge effects
are negligible; this case is referred to as the steady-state ERR, because the ERR is independent of
flaw size. The steady-state ERR is the maximum possible—that is, for shorter interface flaws the
ERR is smaller. The steady-state ERR is simply the change in strain energy per unit crack advance,
that is, G = (Ubonded − Ucracked ) (ba) = Ubonded (ba). This result can be written as†
G = Ēs hs fsf + Ēf hf ffs , (39.18a)
where
gij 2 2
2
fij = hj κi − 2κi κj + 4κj2 + 6hj κj hi κj + 2 (θs − θf ) + 3 hi κj + 2 (θs − θf ) (39.18b)
24
and
2 3
Ēi hi hj hj hj Ēj
gij = +4+6 +4 + . (39.18c)
Ēj hj hi hi hi Ēi
Note that the ERR is symmetric with respect to the film and substrate, that is, one can arbitrarily
label the top or bottom film as the substrate.
This broadly general result has interesting and important implications, even when the two plates
are these same material and thickness. For example, consider two identical plates with the same
initial curvature: the ERR is
Ēh3 Ēh3
G= (curvature opposing) or G= (curvature aligned) (39.19)
12Ro2 16Ro2
∗ In some instances, the plates in the debonded region remain subject to out-of-plane constraints, such that they are not entirely
stress-free: this requires slightly different algebra (see Reference 31).
† Note that the Einstein summation convention does NOT apply.
10 Substrate, (a) (b) (c) (d)
Energy release rate (J/m2)

Glass, Polymer, Glass, Glass,
5 Film polymer polymer elastomer glass
Es, αs, hs
70, 6, 1 4, 40, 2 70, 6, 10 70, 7, 2
(GPa, ppm/°C, mm)
(a) Ef, αf, hf
1 (GPa, ppm/°C, mm)
4, 40,1 4, 40, 2 0.001, 300, 5 70, 7, 2
0.5 same: temp 0, ∆Tf, temp, same: temp 0, ∆Tf, temp,
∆Tf, ∆Ts(°C) change after difference of change after difference of
bonding bonding bonding bonding
(d)
Rs,Rf (meters) 20, 20 20, −20 0, 0 20, −20
0.1 (b) (c)
Γi(Ψ=0°) ~2–5 J/m 2 ~2–20 J/m2 ~0.05–1 J/m2 ~0.1–5 J/m2
0.05
Γi(Ψ=90°) ~10–50 J/m2 ?? ~5–30 J/m2 ~5–30 J/m2
10 15 20 30 50 70 100
Temperature change (°C)
FIGURE 39.7 Examples of energy release rates for several different bilayer scenarios: (a) glassy polymer
film on glass, (b) two glassy polymers bonded at different temperatures, and (c) an elastomer film bonded to a
glass substrate, and (d) two glass slides bonded at different temperatures.
These somewhat surprising results imply that aligning initially curved plates may not be remark-
ably effective in suppressing delamination. It should be noted that if the initial curvatures are aligned,
the final curvature of the plate will be twice the initial curvature: the internal moments generated
by the internal stresses introduced during pressing are complementary. Conversely, as one would
expect, the final curvature of the bilayer is zero if the initial curvatures are opposed and identical.
As anyone in the microelectronics community can attest to, the true headaches begin when two
different materials with different thermal expansion coefficients are bonded together, often times at
different temperatures due to the fabrication process. In this case, the ERR arising from bonding
perfectly flat plates is given by

1 Ēf hf3
G= Ēf hf (θs − θf ) 1 +
2
. (39.20)
2gfs Ēs hs3
The implications of these results for several scenarios relevant to microfluidic devices are illustrated
in Figure 39.7, which depicts the ERR for bilayer debonding as a function of temperature change.
Approximate numbers for interface toughness are also listed, based on an attempt to roughly summa-
rize findings from a diverse range of interface studies. It is worth noting that the ERR for all cases is
in the ballpark of the pure mode I interface toughness; this indicates that many common fabrication
sequences inherently introduce stresses that are “large,” in that the device will be near the threshold
for debonding. This rather thumbnail sketch illustrates the difficulty in identifying highly repeatable
bonding processes. It must be noted, however, that many of the scenarios considered in Figure 39.7
will not necessarily delaminate; the presence of mode II deformation, for which the interface tough-
ness is considerably larger (see table in Figure 39.7), implies greater stability. Nevertheless, it is
clear that many common bilayer scenarios are not “far” from debonding, highlighting the need for
more precise and systematic characterization of such systems.
39.4.2 DEBONDING OF PRESSURIZED FILMS COVERING LONG CHANNELS

Figure 39.8a depicts film debonding due to pressurizing a sealed cavity; the width (span) of the cavity
is denoted as L. Recall from Section 39.2.1 that this length should be taken as the channel width plus
any debonded region next the interior edge of the film/substrate interface (see Figure 39.2b). For
PDMS-glass systems, this inclusion of a debonded region next to the channel involves no significant
approximation, because the substrate is effectively rigid and the geometry of the unbonded region
is not relevant. For glass–glass systems, the geometry of the unbonded substrate is relevant if the
distance from the edge of the channel (see Figure 39.2b) to the crack tip is greater than several
(a) Injected volume, Vi (b)

x
G2
∆
G po > 0 G
po > 0
G1
L L
FIGURE 39.8 Side views of geometries relevant to debonding along channels: (a) edge debonding due to
internal pressure, (b) debonding to release a “stuck” film.
times the film thickness. In the following, we assume that L refers to the total length of the debonded
region, with the understanding that this must extend beyond the edge of the channel when considering
glass-glass systems.
This section considers plate and membrane behavior, as opposed to scenarios where the film is
thick enough to be considered an elastic half-space (as in Figure 39.2a). Thus, the applications in
Section 39.4.2 thus correspond directly to the theory described in Section 39.3.2. The analysis of
two relatively thick films (compared to crack length) requires numerical solutions not discussed in
Section 39.3.2 and hence is discussed in Section 39.4.3.1 in conjunction with pressure-deflection
relationships.
39.4.2.1 Plate Behavior: L >∼10 h and δmax < h

Figure 39.8b illustrates a pressurized film that adheres to an adjacent layer: this is a more general
case of the simple span shown in Figure 39.8a. The film is subjected to a uniform pressure and is
deflected at one end (from its initially stress-free position). Solution to Equation 39.6 subject to the
boundary conditions δ (0) = δ (0) = δ (L) = 0 and δ (L) = − yield the following deflection
profile for the film
po L 4 x 2 x 2 x 2 x
δ (x) = 1 − − 3 − 2 , (39.21)
2Ēh3 L L L L
where po is the resultant pressure, that is, the superposition of that acting on the top and bottom of
the film. Plugging this into Equation 39.8, the ERR is
2
p2o L 4 Ēh3
G= 1±6 . (39.22)
24Ēh3 po L 4
Despite the fact that the pressure seems to act purely to “open the crack,” it should be noted that
there is a mode II component arising from bending of the film at the edge of the delamination: the phase
angle is ψ = −45◦ . This highlights the need for mixed-mode interface toughness measurements for
material systems relevant to microfluidic systems.
When utilizing Equation 39.22, one must determine the proper sign of the operator in parenthesis.
This operation depends on the stress condition at the interface. For instance, if the displacement of
the film places the interface in tension (e.g., G1 in Figure 39.8b where the film is debonding from the
bottom of the channel), then a positive sign should be used in Equation 39.22. In this scenario, there
is a driving force for debonding even when there is no internal pressure. Conversely, if one was to
consider debonding at the upper left edge of the channel—labeled G2 —the displacement of the film
acts to close an interface crack at the edge of the channel: for this scenario, the negative sign is used,
since the displacement decreases the driving force. One may use the same reasoning to determine
the sign of the contribution of the film displacement for interfaces created by the bonding of a third
layer on top of the film.
With = 0, the result corresponds to debonding from an interface crack along the edge of a
long, straight channel (Figure 39.8a). If the internal pressure is held constant, the ERR increases
as the crack extends (i.e., L increases). This corresponds to unstable crack growth: once debonding
initiates, G > i for all subsequent debond lengths L and the crack will not arrest. Conversely, the
crack extends in a stable manner if one injects a constant volume into a previously filled channel. The
injected volume will initially be accommodated by the deflection of the film, as shown in Figure 39.8a.
Integration of Equation 39.21 (i.e., the displaced profile of the film) yields the relationship between
injected volume, Vi and the resulting pressure. This result is then used with Equation 39.8. Since
the second term in Equation 39.8 refers to the pressure times the injected volume, it is constant with
respect to crack length and it does not enter into the ERR calculation. Thus, the derivative of the
potential energy is equal to the derivative with respect to the strain energy, that is, the first term only.
This yields
7Ēh3 (15Vi )2
G= , (39.23)
8bL 6
where b is the length of the channel (i.e., the distance in the z-direction over which the injected volume
is distributed). This result indicates that crack extension due to injected volume will be stable, in that
the ERR will decrease as the crack extends (L increases). This has important implications for both
design (i.e., devices that operated under fixed volume conditions), as well as bulge testing to infer
interface toughness (see Section 39.4.6.2).
39.4.2.2 Membrane Behavior: L >∼10 h and δmax > h

For films that experience deflections that are much greater than the film thickness, the membrane
models of Section 39.3.2.2 are appropriate. For a membrane covering a long channel of width L, the
solution is found by integrating Equations 39.10 and 39.11 and imposing the boundary conditions
δ (x = 0) = δ (x = L) = 0 and u (x = 0) = u (x = L) = 0. The results are
2/3
Ē po L
σo = , (membrane stress) (39.24)
2 · 31/3 Ēh
and
1/3
3po L xx
δ (x) = L 1− . (membrane deflection) (39.25)
Ēh L L
The ERR for constant pressure in the channel can be calculated from Equation 39.12: this yields
1/3
7po L po L
G= . (39.26)
8 9Ēh
Once again, one can solve for the ERR that arises due to constant volume injection, using
Equation 39.25 to calculate the volume for given pressure. In calculating the potential energy used
in the ERR calculation, the second term (pressure times injected volume) is dropped. The result is
4
Vi
G = 126Ēh . (39.27)
bL 2
Under constant volume injection, the ERR again falls as the crack extends (i.e., L increases), implying
stable crack extension under constant volume injection.
TABLE 39.2
Summary of Pressure–Deformation Relationships
Half-Space Plate Membrane
δmax Volume δmax Volume δmax Volume

pL π Lδmax Po L 4 8Lδmax L 3po L 1/ 3 2Lδmax
Channel
E 4 32Ēh3 15 4
Ēh 1/ 3
3
4po a 2π a2 δmax 3pa4 /2Ēh3 π abδmax 3 (1 − v) po a4 π a2 δmax
Circle or Ellipse h
π Ē 3 3 + 2 (a/b)2 + 3 (a/b)3 3 8Eh4 2
Ellipse: b ≥ a Ellipse: b ≥ a
39.4.3 PRESSURE–DEFLECTION RELATIONSHIPS FOR VALVE DESIGN

The relationship between pressure, deformation and geometry plays a critical role in microchip
design, as it essentially dictates the actuation pressure of valves and the fluidic capacitance introduced
by deformable channels. Table 39.2 summarizes several classical results for simple geometries.
39.4.3.1 Long Capped Channels: Plane-Strain Deformation

Figure 39.9 presents the film displacements resulting from pressurizing a cavity, for small pressures
(such that deformations are small compared to the film thickness) and small spans that are compara-
ble to the film thickness. Results are shown for an elastomer film bonded to a glass substrate: due to
the extreme compliance of the elastomer, the substrate is effectively rigid. The curves in Figure 39.9
were generated via numerical FEA of the two-dimensional, plane-strain geometry shown in the inset.
Such techniques are required when the deformation of the film is not dominated by bending, as is
the case for films that are of comparable thickness to the span.
For very thick films, the behavior is well described by results that assume the film behaves as
a semi-infinite half-space. In this limit, the geometry corresponds to that of a pressurized crack,
such that the inner surface of the film displaces according to the well-known analytical result
(e.g., [33])

2
2x
δ (x) = δ∞ 1 − , (39.28)
L
the maximum surface deflection is given by
pL
δ∞ = . (39.29)
Ē
Note that this analytical result is independent of the film thickness, which is a consequence of
invoking the assumption that the film behaves as a half-space. The maximum displacement of the
inner surface, δ∞ , is used to normalize the displacement results in Figure 39.9; hence, the curves
asymptote to unity for small aspect ratios (note a/h = L/2h).
Figure 39.9 illustrates that for shorter spans, the deflection of the top of the film will be consid-
erably smaller than that of the bottom; this has significant implications for the actuation pressure of
thick elastomer films used as valves. Monitoring the deflection of the top surface enables use of the
PDMS film, glass substrate

1000
Internal (bottom)
displ., δin/δα
Displacement, Eδ/(2poa)
External (top)
100 displ., δtop/δα
Capacitance, C/Cα
10
3 h x
1 1
2a
δ/δα α (a/h)3
0.1
0.1 1 10
Span to film thickness ratio, a/h
FIGURE 39.9 Film deformation as a function of span to thickness ratio for elastomer films bonded to glass
substrates, for long, straight-walled channels (plane-strain); results illustrate the transition from the half-space
solution (where bending is negligible) to the plate solution (dominated by bending).
film as a pressure sensor to quantify the internal cavity pressure, or direct measurement of the fluidic
capacitance of the channel. For larger spans where bending dominates, the deflection of the top
(or outer) surface is identical to the bottom surface. When bending dominates, the film deflection is
accurately described by the solution for a clamped plate. The deflections of the plate are given by
Equation 39.21 with = 0.
The capacitance of the deformable film (discussed more fully in Section 39.4.5) is also shown:
this is essentially the channel volume introduced by deformation of the film and is calculated by
integrating the film’s spatial displacement profile obtained from the numerical analysis. These results
are normalized by the volume obtained by integration of Equation 39.28. The capacitance results in
Figure 39.9 can be reinterpreted as the ERR for debonding, with the crack length defined according to
the convention discussed at the outset of Section 39.4.2. This reinterpretation involves replacing the
normalized capacitance with a normalized ERR, G/Go , where Go is the ERR for a pressurized crack
between two half-spaces: this quantity is given in Table 39.5. The coincidence of the normalized
capacitance and that of the normalized ERR is not exact (i.e., there is no obvious analytical reason
the two should coincide, at least not obvious to us), but rather an observation based on numerical
results. The error is less than ∼20% over the range shown in Figure 39.9.
For films that are sufficiently thin to allow for bending, a central question is whether or not
pressures are large enough to induce large deflections, in which case one must account for axial
stretching of the film (along its centerline). At sufficiently low pressure or small deflections, the
deflection varies linearly with pressure (i.e., plate behavior). At large pressures or large deflections,
the film deflections scale with p1/3 . This is referred to as membrane behavior. The transition between
plate and membrane behavior can be illustrated using a closed-form solution that implicitly defines
the deflection of the film for a given pressure. The solution that describes the applied pressure and
the deflection in terms of the axial tension developed in the membrane is [32]
√
po L 4 σ̂ 2 sinh σ̂
=λ= √ √ √
, (39.30)
32Ēh4 12 24 + 4σ̂ cosh 2 σ̂ − 18 σ̂ sinh 2 σ̂ − 24 − 16σ̂
√ √ 
σ̂ − 2tanh 2σ̂
δmax = 12hλ  , (39.31)
σ̂ 3/2
10 100
h 1
%Error in analytical asymptotes

χ δ 1
80
Membrane
L
result
1
Deflection, δ/h
60
1
3
40
0.1
20
0.01 0
0.01 0.1 1 10 100
poL4
Pressure,
32Eh4
FIGURE 39.10 Deflection–pressure relationship for films over long, rectangular channels (plane-strain defor-
mation) illustrating the transition from plate-to-membrane behavior. The shaded region represents combinations
of pressures and displacements for which simple analytical solutions have greater than 10% error.
where λ represents the normalized (dimensionless) pressure and is the normalized axial tension in
the film: σ̂ = 3NL 2 /Ēh3 where N is the axial resultant (force per unit width of the film). Clearly,
the deflection for a given applied pressure can found by numerical solution to Equation 39.5 and
substitution into Equation 39.6. However, it is often simpler to generate the load–displacement
relationship via the parametric plotting features of commercial codes such as Mathematica. The
outcome of this procedure is illustrated in Figure 39.10.
For sufficiently large pressure, one obtains the classical membrane solution, described by Equa-
tion 39.25. The shaded region in Figure 39.5 represents combinations of pressures and displacements
for which simple analytical solutions have greater than 10% error.
39.4.3.2 Circular Films

Similar behavior is observed for circular films, although a closed-form expression for the pressure–
deflection relationship in the transition region is not possible. Figure 39.11 illustrates the transition
from plate-to-membrane behavior as the pressure applied to a circular film is increased. It should
be noted that for circular (i.e., axisymmetric) films, the dependence on Poisson’s ratio is slightly
different from that of the plane-strain case. Strictly speaking, the pressure–deflection relationship
shown in Figure 39.11 depends on the Poisson’s ratio; however, the dependence is rather weak and
tabulated elsewhere [39]. The asymptotic limits at low and high pressure (i.e., plate and membrane,
respectively) are listed in Table 39.2; note that those listed inside Figure 39.11 are for v = 1/2.
39.4.4 DESIGN OF CHECK VALVES FOR SPECIFIC ADHESION ENERGY AND/OR

DIMENSIONS
The results of the previous sections can be used to design check valves that control fluid motion in
powerful ways. Consider the geometry shown in Figure 39.12 [40]: a wall of thickness w separates
two fluidic channels within a single layer. At low pressure, the check valve closes the channel by
seating on this wall. When a critical pressure is reached, the valve film debonds from its seat and
allows fluid to pass from one channel to the other. Assuming displacements are much smaller than
the film thickness (such that plate behavior applies) the critical pressure required to release the check
h
10 100
2a
80
% Error in analytical result

Plate
1 regime
Deflection, δ /h
Membrane 60
9poa4 regime
δ=
1/3
64Eh3 3poa4
δ=h
16Eh4 40
0.1
20
0.01
0.1 1 10 100 103
poa4
Pressure,
Eh3
FIGURE 39.11 Deflection–pressure relationship for circular films with v = 1/2, illustrating the transition
from plate-to-membrane behavior. The shaded region represents combinations of pressures and displacements
for which analytical solutions have greater than 10% error.
y
L2 p2
∆ w
x
p1
p1 G p2 L1
x
z
L1 w L2
Side view Top view
FIGURE 39.12 Schematic diagram of a fluidic check valve: by modulating the dimensions (and/or interface
adhesion energy), one can design valves that allow flow in only one direction, and achieve the desired deflection
after debonding at the design pressure.
valve and allow for fluid flow is dictated by Equation 39.22 with = 0: that is,

24Ēh3 i
pc1 = , (39.32)
L12
where i is the adhesion energy of the interface formed by the film and the valve seat. Clearly, one
may control the relative pressures to allow for flow from left-to-right (or vice versa) by suitable
choice of the channel width, L. The ratio of activation pressures for forward (1 to 2) and reverse (2
to 1) directions thus scales as
2
pc1 L2
= . (39.33)
pc2 L1
Above the critical pressure, the valve debonds, which connects the two fluid channels. The valve
displaces according to Equation 39.21 with L = L1 + L2 + w and = 0.
Suppose that in the “open” position, the desired film displacement is ; presumably this is
chosen to achieve a specific flow rate at the activation pressure of the valve. Alternatively, may
represent the maximum allowable displacement of the valve film to prevent adhesion with overlying
layers, as shown in Figure 39.12. Thus, one can identify the combinations of material properties and
dimensions that result in a valve that is either closed or open to a prescribed clearance by setting
pc1 = pmax (). This combination is
2 2
h
i L1 L1
= 8 . (39.34)
Ēh 1+ L2
+ w
L1 L1
In ideal world, one might modulate the adhesion energy of the valve seat to achieve the desired
performance for a specific set of dimensions. This would likely be difficult as the chemistry of the
fluidic environment, which obviously depends on the application, often affects interface toughness.
However, one can measure the interface toughness for a given application and then choose the
dimensions of the chambers to satisfy Equation 39.34: the result is a check valve that actuates to a
specific clearance for a given interface condition.
One can naturally derive similar design guidelines for chamber dimensions comparable to the film
thickness (i.e., the elastic half-space regime) or valve displacements greater than the film thickness
(i.e., the membrane regime).
39.4.5 FLUIDIC CAPACITANCE

Any feature of the fluidic network whose internal volume changes with pressure acts as a fluidic
capacitor, because those volume changes imply that additional mass will be “stored” inside the
feature. This behavior can be exploited to alter the characteristic timing of flow inside a fluidic
“circuit” (e.g., [41]). Assuming density changes are negligible, the mass flow rate can be written as

dm d dV
q̂ = =ρ Vo + d(p) , (39.35)
dt dt d (p)
where V0 is the initial fixed volume of the element, ρ is the mass density and p = pin − pout is
the difference between pressure inside and outside the element. Obviously, the volume flow rate (at
constant density) is simply q̂ = q/ρ. In terms of volume flow rate,
dV dp
q= · ≡ C ṗ, (39.36)
dp dt
where it should be noted dVo /dt = 0 by definition. Clearly, the flow rate is defined as positive
towards the capacitance element inside the chamber: when the pressure inside is greater, the film
deflects outward and the capacitor stores more fluid (positive flow rate). Hence, for any enclosed
volume, the fluidic capacitance is defined as the change in volume with respect to the difference
between internal and external pressure.
Table 39.2 lists the pressure-volume relationships for various geometries: the fluidic capacitance
is found simply by differentiating with respect to pressure. For small deformations, volume varies
linearly with applied pressure, such that the capacitance is not a function of the pressure: it merely
defines the proportionality between increases in pressure and increases in “stored” mass. For such
cases, the fluid circuit analysis is linear, because flow rate and pressure drops are related via linear
expressions. For large deformations (i.e., the membrane limit), the fluidic capacitance is a function of
the pressure: this implies that the fluidic circuit behavior will be nonlinear. Obviously, once the fluidic
resistance, capacitance and inductance have been identified via geometry (and material properties),
complicated networks can be analyzed using commercially available circuit analysis software such
as SPICE [42].
39.4.6 MATERIAL PROPERTY MEASUREMENTS

This section outlines several approaches to characterizing the elastic modulus of thin films and the
interface toughness when bonded to substrates. Naturally, there is an elastic enormous variety of test
configurations that can be used; attention here is limited to those with specimen fabrication that is
similar to that used for typical microfluidic devices. The text by Freund and Suresh [32] describes
additional test approaches, particularly those pertaining to quantifying the effects of mode-mixity
on interface toughness.
39.4.6.1 Indentation Modulus Measurements

As evident from the calculations in previous sections, the elastic modulus of a material must be known
if a researcher hopes to predict interface toughness for a fabricated device. An increasingly prevalent
method to characterize modulus is instrumented indentation, wherein an indenter tip of known
shape is pressedinto a surface: see Figure 39.13a. The plane-strain modulus of the material, that is,
Ē = E/ 1 − v2 , is extracted from the measured force–displacement relationship. The test method is
particularly attractive because it requires minimal sample preparation and mounting: one must merely
ensure that a nominally flat film can be attached to substrate that can be inserted into the apparatus.
This makes it ideal for testing thin films, as samples can be deposited via spin casting, chemical
vapor deposition and so on. Moreover, the dynamic ranges of force and displacement measurement
continue to expand, enabling tests on a broad range of film thickness. Conventional micro-indenters
typically apply forces in the Newton range and are capable of measuring displacements in the >
10 micron range. Nanoindentation systems can accurately apply forces beneath a milli-Netwon
and measure displacements with nanometer precision over a broad dynamic range spanning tens of
microns. Such systems are increasingly common and are becoming the de facto choice for thin film
modulus measurements, due to sophisticated and automated test control and interpretation.∗ The
book chapter by Hay and Pharr [43] provides an excellent, detailed overview of the instrumentation
and theory involved in indentation testing.
Typical load–displacement curves obtained from instrumented indentation experiments are
shown in Figure 39.13b. Results are shown for silica (E = 72 GPa, v = 0.2) and polystyrene
(PS, E = 2 GPa, v = 0.4). The load–displacement curve for PDMS (E = 1.5 MPa, v ∼ 0.5) involve
forces that are too low to appear on the scale used in Figure 39.13b, due to the fact its elastic modulus
is three orders of magnitude lower than polystyrene.
Depending on the material being tested, the loading path is a complicated function of elastic and
inelastic deformation (such as creep or plasticity). Conversely, the initial portion of the unloading
curve is dominated by purely elastic deformation; it is this portion of the measurement that is used
to extract the modulus. The initial slope of the unloading curve is often referred to as the contact
stiffness and is denoted here as S. The postprocessing method for determining S by curve-fitting the
unloading portion of the curve is described in detail in [43]. The basic relationship that is used is
EEi S
=√ , (39.37)
1 − v2 Ei + 1 − vi E
2 A (δc )
where Ei is the modulus of the indenter and S is the measured stiffness (i.e., slope of the initial
portion of the unloading curve). A is the area function of the indenter tip: this function describes the
relationship between contact depth, δc , and the projected contact area between the indenter tip and
surface (see Figure 39.13a).
∗ Indentation testing can also be performed with an atomic force microscope, which has pico-Newton force and angstrom
displacement resolution at the lower end; the draw back is that the dynamic displacement range is limited to several microns.
Moreover, on older systems, the control software, as well as that needed for test interpretation, may not be readily available.
(a) P
(Not to scale)
Incompressible mat'l
with strong adhesion: Aa(d)
Aa(d )>A(d )
d Classical elastic result
with no adhesion: A(d )
nm
0 200 400 600
100
(b) Silica (c)
30
10
Indenter load (mN)
Silica
Modulus (GPa)
S Polystyrene (PS)
20 1
1
0.1
10 PS
PDMS
0.01
0
0.001
0 1 2 3 100 1000
Indent depth (µm) Indent depth (µm)
FIGURE 39.13 Typical measurements obtained using a commercial nanoindenter: (a) schematic of the contact
area, (b) typical load–displacement measurements, and (c) inferred modulus as a function of penetration depth
for several materials.
The accuracy of the test method relies critically on the accurate determination of the area func-
tion, which varies with tip geometry. At the scale of many nanoindentation experiments, minute
differences in tip geometry—such as those caused by wear induced by repeated experiments—can
have a large effect on the area function. For this reason, the area function is typically measured by
indenting a known material to several depths and computing the function from Equation 39.37. This
experimentally determined tip function is then used to interpret further tests on unknown materials.
Fused silica (or quartz) is commonly used to “calibrate the tip,” that is, determine the area tip func-
tion, because its material response is nearly perfectly elastic and it does not exhibit adhesion with
diamond indenter tips.
Provided one is testing relatively stiff material such as glassy polymers, ceramics and metals, the
choice of indenter tip geometry is largely one of convenience; by far the most common is a diamond
that has been cleaved on crystallographic planes to generate a well-defined tip profile. The resulting
shape is that of a pyramid with 65.3◦ face angle and a three-sided base, known as a Berkovich tip.
Other common indenter tips include Vickers (a pyramid with a square base), spheres, cones and
even cube-corners (i.e., the corner of a cube that is lopped off and mounted such that the direction of
indentation aligns with the diagonal of the cube). It should be emphasized that the actual geometry
differs from these nominal shapes at small length scales; for example, a Berkovich tip will typically
have a rounded tip with a radius of ∼100–300 nm after hundreds of indents on stiff materials such
as ceramic or metals. This implies that careful tip calibration is critical for indents whose maximum
penetration is smaller than several microns.
Modern hardware and testing protocols are capable of superimposing a small oscillatory com-
ponent to the apply load. This effectively introduces a series of unloading cycles that can be used to
extract the modulus; this technique is commonly referred to as the dynamic contact method (DCM)
and is a standard option in commercial nanoindentation systems. The result is that one can measure
the modulus of the material as a function of indentation depth in one rapid test.∗ Examples of data
obtained from such nanoindentation testing are shown in Figure 39.13c for several materials: the
results represent the average of a dozen tests, with the error bars indicating standard deviation. The
results for the ceramic (silica) and glassy polymer (polystyrene) are typical of relative stiff materials
whose modulus is greater than 1 GPa: for penetration depths greater than one hundred nanometers,
the modulus is independent of probe depth. Moreover, for these stiff materials there is very little
scatter. This indicates that the tests successfully probed the “bulk” material properties and were not
unduly influenced by surface phenomena.
In contrast, the results in Figure 39.13c for the elastomer (PDMS) exhibit larger scatter and an
elastic modulus that depends on depth. The indentation testing of elastomers is obviously relevant
to microfluidic devices; unfortunately, such materials introduce a host of challenges that continue
to be addressed by the mechanics community. First and foremost, such materials exhibit strong
adhesion with indenter tips; adhesion obfuscates the inferred modulus by artificially increasing the
contact area for a given depth (see Figure 39.13a). This increases the slope of the unloading curve
and one infers a larger modulus than reality (see Figure 39.13c: the average modulus for strains less
than 3% is 1.5 MPa). Second, such materials exhibit nonlinear stress–strain behavior, which strictly
speaking, violates the models derived earlier. Even though the material is purely elastic (does not
experience permanent deformation), this leads to ambiguity regarding the imposed strain level at
which a modulus is extracted. The most desirable strategy that avoids these problems is to use a
flat-ended punch; in this case, provided the indenter is not unloaded too far, adhesion plays no role
during unloading because the contact area remains in compression and does not change.
Alternatively, one can indent a freestanding film (i.e., a blanket film covering a hole or microflu-
idic chamber) using an instrumented probe, as shown in Figure 39.14. With careful choice of
dimensions, the measured load-deflection response is dominated by plate bending or membrane
stretching, in which case analytical solutions are accurate (e.g., [38]). Adhesion is less of a factor,
because the load–deflection response is dominated by behavior outside the contact area. In many
regards, the instrumentation can be easier to set up; one can simultaneously measure load and dis-
placement with the same probe (as opposed to pressure loading in “bulge tests,” (Section 39.4.6)
where film deflection must be monitored separately from the applied pressured). This has been illus-
trated for macroscale elastomer films in the membrane regime (PDMS: ∼100 µm thick and 5 cm in
diameter [38,44]). Similar tests in the plate regime have been demonstrated for ultrathin glassy poly-
mers that cover microfabricated holes (∼750 nm thick PMMA films over holes 60 µm in diameter:
[45]), using a commercial nanoindenter as the instrumented probe.
39.4.6.2 Bulge Testing to Determine Modulus and Interface Toughness

The configurations shown in Figures 39.3, 39.10, and 39.11 can be used to probe mechanical prop-
erties of the system by monitoring film displacement while modulating the pressure: this is typically
referred to as a “bulge test” (e.g., [31,32,46–48]). If the delamination crack front can be monitored
while the pressure (or injected volume) is modulated, the interface toughness can also be deter-
mined. There are two design considerations for the test: (1) the range of strain imposed during the
test and (2) the nature of the load–deflection relationship as dictated by the film thickness-to-span
ratio. The former determines whether or not the strains are sufficiently small to allow for a linear
elastic approximation and to prevent film rupture. Table 39.3 summarizes the relationships between
applied pressure and the maximum strain in the film for various configurations. To address the sec-
ond consideration, the film thickness and span is ideally such the range of applied pressure enables
the accurate application of a simple closed-form analytical expression, that is, plate solutions or
membrane solutions.
∗ Or more importantly, one can extract an accurate area tip function by averaging the results of dozens of indents.
Clamped Mounting bracket

Load cell
film Guide rails
Positioning stage
Spherical indenter 8.8 mm total travel
Air bearing
Clamping plates
Screw-driven
clamping plate PDMS: Dow sylgard 184
0.2
Data
Guide holes Point load solution
Clamping plate Finite contact solution
Load (N)
PDMS film
E = 1.5 MPa
0.1
h
R/a = 0.5
h = 0.127 mm
a= 25.4 mm
a
0
0 2 4 6 8
Displacement (mm)
FIGURE 39.14 Schematic illustration of deflection test on freestanding PDMS film (Dow Sylgard 184 with
a 10:1 mixing ratio) [38].
TABLE 39.3
Estimate for Maximum Strain in the Film
Long Channel: Plane Strain Circular
Plate Membrane Plate Membrane

po L 2 1 3po L 2/3 3po a2 (1 − v)po a 2/3
4Ēh2 6 Ēh 4Ēh2 2Eh
Table 39.4 summarizes simple rules of thumb for designing microfluidic systems developed
from the pressure-strain solutions presented in Section 39.4. Based on the resolution of the available
instrumentation and anticipated modulus, these recommendations can be adjusted for film thickness
and span as required. The broad generalizations of Table 39.4 can be specified more quantitatively
using the results shown in Figures 39.9 through 39.11; that is, critical values of pressure and deflection
can be identified such that deflections are within a specified percentage of the analytical solutions.
The bulge test described above can be used to characterize the interface as well as the film itself.
Table 39.5 summarizes the ERR (G) for the half-space, plate and membrane limits, in terms of both
applied pressure and injected volume. The same limits identified in Table 39.4 can be used to establish
the limits of accuracy. The use of injected volume is preferred because the crack front advances in a
stable manner: this implies that a single film can be used to make multiple measurements. In order to
extract interface toughness, one must monitor the “crack length,” that is, the width of the debonded (or
bulging) region. For transparent films, the gap between the bulging film and substrate creates enough
contrast to monitor the position of the crack front using optical methods. Depending on the film
thickness, optical methods are often sufficient even for opaque films, as the edge of the bulging region
TABLE 39.4
Design Guidelines for Accurate Closed-Form Pressure-Deflection Solutions
Long Channel: Plane Strain Circular (v = 1/2 )
Plate Membrane Plate Membrane

δmax < h/3 δmin > 2h δmax < h/3 δmin > 2h
4 4 4 4
h h h h
pomax < 0.8Ē pomin > 30Ē pomax < 3Ē pomin > 40Ē
a a a a
TABLE 39.5
Summary of Energy Release Rates
Half-Space Plate Membrane
Pressure Pressure Volume Injection Pressure Volume Injection

π p2o L p2o L 4 3Ēh h 2 10Vi 2 7po L po L 1/3 Vi 4
Channel (*) 2
126Ēh 2
4Ē 24Ēh3 2 L bL 8
9Ēh 1/3 bL

2p2o a 3p2o a4 24Ēh h 2 Vi 2 5 3 (1 + v) p4o a4 112Ēh Vi 4
Circle (*)
π Ē 32Ēh3 π2 a a3 18 Ēh (1 + v) a3
(∗) These results assume the substrate is rigid compared to the film: if the film and substrate have identical elastic properties,
these results should be multiplied by two.
is clearly visible (particularly when using white light interferometry). Alternatively, if the volume of
injected material is known, the crack length can be inferred from the maximum deflection of the film.
The toughness can be determined by least squares fit to the measured L = f (Vi ) relationship, using
G = i as the unknown fitting parameter. Monitoring small deflections generally ensures that the
linearly elastic material assumption is valid; however, this also implies small injected volumes that
may be difficult to achieve. Thus, one should design the test—that is, adjust film thickness and span—
to lie in the appropriate regime of behavior based on the resolution limits of the volume injection and
displacement monitoring.
39.4.6.3 Peel Tests to Determine Interface Toughness

Perhaps the most straightforward test to determine interface toughness is the peel test, wherein
one measures the applied force required to peel a film from the substrate. The test is illustrated in
Figure 39.15 and has the advantage of measuring at least part of the dependence of the interface
toughness on mode-mixity. The ERR for this configuration is [32,49]
σa2 h
G = σa h (1 − cos θ ) + , (39.38)
2Ē
where σa is the stress applied to the film at the end. (Note that this simply the force applied to the film
divided by the cross sectional area of the strip being pulled.) The test involves mixed-mode loading
due to bending of the film near the edge of the debonded region, except for the case where θ = 0◦
that corresponds to pure mode II. For compliant films on much stiffer substrates (i.e., α ≈ −1) the
σa = P/bh −30
Mode-mixity phase angle, ψ

−40
0.001
L
b −50
Note:b>>L,h
for plane strain
−60 0.01
θ h ε0 = σa/E = 0.1
−70
−80 Dundur’s parameters:

α = -1, β = 0
0 25 50 75 100 125 150 175

Peel test angle, θ°
FIGURE 39.15 Schematic of the peel test, and associated mode-mixity phase angle for the case of a flexible
film on a rigid substrate; mode II arises from bending near the debond edge.
phase angle of the test ψ is given by

 
2(1−cos θ)
 cos θ − + sin θ  2
εo
ψ = − tan−1  , (39.39)
θ)
− cos θ + 2(1−cos
εo + sin 2
θ
where εo = σa /Ē is the level of strain in the film far from the debond edge. Care should be taken
to ensure that the material is appropriately modeled using linear elasticity for the strain level at
debonding.
The phase angle () is shown in Figure 39.15 as a function of peel test angle for several values
of applied strain. In all cases, the phase angle is negative due to a negative mode II component: thus,
the test is only capable of measuring mixed-mode interface toughness in this regime. When the film
is pulled vertically (i.e., θ = 90◦ ), the phase angle is ψ = −45◦ regardless of the applied strain
level: this is identical to the bulge test in the plate regime.
39.5 FUTURE PROSPECTS AND CONCLUDING REMARKS

The convergence of microfabrication technology developed for microelectronics, microelectro-
mechanical systems (MEMS) and lab-on-chip technology is steadily progressing, and creating rapidly
expanding opportunities for new types of miniaturized chemical analysis. It seems obvious that future
of microfluidic systems lies in hybrid/composite devices, which incorporate a wide range of materi-
als and geometric features and hence, fabrication techniques. Future chips will undoubtedly involve
integrated metallic electrodes, piezoelectric materials for actuation, micropatterned polymers for
chemical sensing and so on. The successful realization of such complex microfluidic devices relies
critically on the translation of thin film mechanics to those material systems that differ from typical
semiconductor and metallic devices. The theoretical and experimental approaches to predicting the
mechanical response of multilayer systems outlined here provide a well-established foundation that
should be applied to these material systems.
The increasing use of polymers in microfluidic devices, especially elastomers, is a significant
departure from microelectronic devices that tend to rely on comparatively stiff ceramics and metals.
Hence, there is a critical need for sustained and systematic experimental and theoretical characteri-
zation of polymer/ceramic and polymer/metal interfaces. The utility of such studies will be strongly
influenced by the successful partnering of chemists and engineers; it is clear that relevant material
systems must be defined by those with an understanding of chemical compatibility and device per-
formance requirements, while those familiar with mechanical integrity offer an efficient route to
characterization.
A critical area that needs analysis is the initiation of interface failure at the corners formed by
bonding capping layers over patterned channels. While there have been significant inroads to this
problem from the mechanics community, it has largely been motivated by microelectronic devices.
As such, the analysis of geometry and materials prevalent in microfluidic devices has received only
cursory treatment. Similarly, the behavior of PDMS/glass interfaces is particularly important, due
to their ubiquitous utilization in emerging devices. While there have been a number of studies in
this regard, the understanding of interface toughness as a function of both mode-mixity and surface
treatment requires significant additional study. Related to this, the role of nonlinear material behavior
and large strain deformation in mechanics of interface delamination is largely unexplored, yet is likely
to play an important role in successful device fabrication.
NOMENCLATURE/GLOSSARY
a Radius of a circular span, used in expressions to describe the behavior of a deformable film
bonded over a circular hole.
b The length of the crack front that experiences debonding, often taken as unity.
α, β Dundur’s parameters: dimensionless parameters that functions of the elastic properties of
two materials on either side of an interface, which play a critical role in interface mechanics.
αi Coefficient of thermal expansion of the ith layer, used to include the effects of thermal
strains introduced during deposition.
C Fluidic capacitance, that is, volume of fluid stored in a deformable element per unit pressure.
δ deflection of a film or half-space in the direction normal to the interface (or axis of the
film), may appear as a function of position along the film axis (i.e., δ(x)).
The maximum or requisite displacement of a valve film, which generally is dictated by the
fluidic channel depth.
ε Strain, usually refers to the direct strain (aligned with the major axis) of a deformable film.
εo The axial extensional strain resulting from bonding two layers.
Ei Elastic modulus (Young’s modulus) of layer “i”: if there is no subscript, then this refers to
the modulus of the film (as the substrate is assumed to be rigid).
Ēi The plane-strain elastic modulus of the layer “i”: Ē = E/(1 − r 2 )
G The energy release rate, or “crack driving force,” which is used to predict crack stability.
i The critical value of the energy release rate at which crack extension occurs, often referred
to as the fracture toughness, or adhesion energy: the subscript indicates that the quantity
refers to the interface toughness (as opposed to a bulk material).
hi Thickness of layer “i”: if there is no subscript, then this refers to the Poisson’s ratio of the
film (as the substrate is assumed to be rigid).
κ The curvature of a bilayer created by bonding together two layers.
L Total span (or crack length) across a long, rectangular channel.
M Resultant moment in a thin film.
N Axial force resultant in a thin film.
po The net pressure acting inside a cavity; that is, the sum of pressures acting on both sides of
the film: assumed to be spatially uniform.
P, S Indentation load and stiffness (i.e., the slope of the load–displacement curve).
Ti , Tio Temperature and reference temperatures, respectively, of the ith layer: the reference tem-
perature is defined as that at which there is no thermal strain, typically the fabrication
temperature
of the layer.
θi = αi Ti − Tio , the thermal strain in the ith layer.
σ Stress.
ψ Phase angle describing mode-mixity of crack driving force: describes the relative amounts of
mode I (opening) and mode II (sliding) deformation at a crack tip.
U Potential energy of the system, comprised of the strain energy of a deformed solid and the work
done by applied loads.
vi Poisson’s ratio of layer “i”: if there is no subscript, then this refers to the Poisson’s ratio of the
film (as the substrate is assumed to be rigid).
REFERENCES
1. Ruska, W.S. Microelectronic Processing: An Introduction to the Manufacture of Integrated Circuits,
McGraw-Hill: New York, 1987.
2. Morgan, D.V. and Board, K. An Introduction to Semicondutor Microtechnology, 2nd edn. John Wiley
& Sons: New York, 1990.
3. Madou, M. Fundamentals of Microfabrication, CRC Press: Boca Raton, 1997.
4. van Zant, P. Microchip Fabrication: A Practical Guide to Semiconductor Processing, McGraw-Hill:
New York, 2000.
5. Becker, H., Lowack, K., and Manz, A. “Planar quartz chips with submicron channels for two dimen-
sional capillary electrooresis applications,” Journal of Micromechanics and Microengineering 1998,
8, 24–28.
6. Rocklin, R.D., Ramsey, R.S., and Ramsey, J.M. “A microfabricated fluidic device for performing
two-dimensional liquid-phase separations,” Analytical Chemistry 2000,72, 5244–5249.
7. Gottschlich, N., Jacobson, S.C., Culbertson, C.T., and Ramsey, J.M. “Two dimensional electrochro-
matography/capillary electrophorsis on a microchip,” Analytical Chemistry 2001, 73, 2669–2674.
8. Ramsey, J.D., Jacobson, S.C., Culbertson, C.T., and Ramsey, J.M. “High-efficiency, two-
dimensional separations of protein digests on microfluidic devices,” Analytical Chemistry 2003, 75,
3758–3764.
9. Wang, Y.C., Choi, M.N., and Han, J.Y. “Two-dimensional protein separation with advanced sample
and buffer isolation using microfluidic valves,” Analytical Chemistry 2004, 76, 4426–4431.
10. Wu, H.K, Wheeler, A., and Zare, R.N. “Chemical cytometry on a picoliter-scale integrated microfluidic
chip,” Proceedings of the National Academy of Sciences 2004, 101351, 12809–12813.
11. Roper, M.G., Shackman, J.G., Dahlgren, M., and Kennedy, R.T., “Microfluidic chip for continuous
monitoring horome secretion from live cells using an electrophoresis-based immunoassy,” Analytical
Chemistry 2003, 75, 4711–4717.
12. Oleschuk, R.D., Shultz-Lockyear, L.L., Ning, Y.B., and Harrison, D.J. “Trapping of bead-based
reagents within microfluidic systems: On-Chip solid-phase extraction and electrochromatography,”
13. Jemere, A.B., Oleschuk, R.D., Ouchen, F., Fajuyigbe, F., and Harrison, D.J. “An integrated solid-phase
extraction system for sub-picomolar detection,” Electrophoresis 2002,23, 3537–3544.
14. Stachowiak, T.B., Svec, F., and Frechet, J.M.J. “Chip electrocrmatography,” Journal of Chromatog-
raphy A 2004, 1044, 97–111.
15. Peterson, D.S. “Solid supports for micro analytical systems,” Lab on a Chip 2005, 5, 132–139.
16. Guillo, C., Karlinsey, J.M., and Landers, J.P. “On-chip pumping for pressure mobilization of the
focused zones following microchip isoelectric focusing,” Lab on a Chip 2007, 7, 112–118.
17. Satow, T., Machida, A., Funakushi, I.C., and Palmieri, R. “Effects of the sample matrix on the
separation of peptides by high performance capillary electrophoresis,” Journal of High Resolution
Chromatography 1991, 14, 276.
18. Svec, F. “Preparation and HPLC applications of rigid macroporous organic polymer monoliths,”
Journal of Separation Science 2004, 27, 747–766.
19. Thorsen, T., Maerkl, S.J., and Quake, S.R. “Microfluidic large-scale integration,” Science 2002, 298,
580–584.
20. Fu, A.Y., Chou, H.P., Spence, C., Arnold, F.H., and Quake, S.R. “An integrated microfabricated cell
sorter,” Analytical Chemistry 2002, 74, 2451–2457.
21. Grover, W.H., Skelley, A.M., Liu, C.N., Lagally, E.T., and Mathies, R.A. “Monolithic membrane
valves and diaphragm pumps for practical large-scale integration into glass microfluidic devices.”
Sensors and Actuators B—Chemical, 2003, 89, 315–323.
22. Jeon, N.L., Chiu, D.T., Wargo, C.J., Wu, H., Choi, I.S., Anderson, J.R., and Whitesides, G.M. Design
and fabrication of integrated passive valves and pumps for flexible polymer 3-dimensional microfluidic
systems, Biomedical Microdevices, 2002, 4, 117–121.
23. Easley, C.J., Karlinsey, J.M., Leslie, D.C., Begley, M.R., and Landers, J.P. Directional and frequency-
dependent flow control in microfluidic circuits using passive elastomeric components, Proceedings
of the mTAS 2006 Conference, 10th International Conference on Miniaturized Systems for Chemistry
and the Life Sciences (Tokyo), 2006, 2, 1064–1068.
24. Buchholz, B.A., Doherty, E.A.S., Albarghouthi, M.N., Bogdan, F.M., Zahn, J.M., and Barron, A.E.,
“Microchannel DNA sequencing matrices with a thermally controlled “viscosity switch,” Analytical
Chemistry, 2001, v73, 157–164.
25. Manz, A., Harrison, D.J., Verpoorte, E., Fettinger, J.C., Paulus, A., Ludi, H., and Widmer, H.M., “Pla-
nar Chips Technology for Miniaturization and Integration of Seperation Techniques into Monitoring
Systems, CE on a chip” Journal of Chromatography, 1992, 593, 253–258.
26. Harrison, D.J., Manz, A., Fan, Z.H., Ludi, H., and Widmer, H.M., “Capillary Electrophoresis and
Sample Injection Systems Integrated on a Planar Glass Chip,” Analytical Chemistry, 1992, 64, 1926–
1932.
27. Geschke, O., Klank, H., and Telleman, P., Ed. Microsystem Engineering of Lab-on-a-Chip Devices,
Wiley VCH: Weinheim, 2004.
28. Li, Paul C.H. Microfluidic Lab-On-a-Chip for Chemical and Biological Analysis and Discovery, CRC
Press Taylor & Francis: Boca Raton, 2006.
29. Xia, Y. and Whitesides, G.M. “Soft lithography,” Angewandte Chemie, International Edition 1998,
37, 550–575.
30. Xia, Y., and Whitesides, G.M. “Soft lithography,” Annual Review of Materials Science 1998, 28,
153–184.
31. Hutchinson, J.W. and Suo, Z. “Mixed mode cracking in layered materials.” In Advances in Applied
Mechanics, Hutchinson, J.W., Wu, T.Y. Eds. Academic Press: London 1992, 29, 63–191.
32. Freund, L.B. and Suresh, S. Thin Film Materials: Stress, Defect Formation, and Surface Evolution,
Cambridge University Press: Cambridge, 2003.
33. Kanninen, M.F. and Popelar, C.H., Advanced Fracture Mechanics, Oxford University Press, New York,
1985.
34. Mohammed, H. and Leichti, K.M., “Cohesive zone modeling of crack nucleation at bimaterial corners,”
Journal of the Mechanics and Physics of Solids, 2000, 48, 735–764.
35. Labossiere, P.E.W., Dunn, M.L., and Cunningham, S.J., “Application of bimaterial interface corner
failure mechanics to silicon/glass anodic bonds,” Journal of the Mechanics and Physics of Solids 2002,
50, 405–433.
36. Reedy, Jr., E.D., “Strength of butt and sharp cornered joints,” Comprehensive Adhesion Science,
Elsevier Press, Amsterdam. 2001.
37. Ambrico, J.M. and Begley, M.R., “The role of flaw geometry in film delamination from two-
dimensional interface flaws along free edges,” Engineering Fracture Mechanics 2003, 70, 1721–1736.
38. Scott, O.N., Begley, M.R., Komaragiri, U., and Mackin, T.J. “Indentation of freestanding elastomer
films using spherical indenters,” Acta Materialia 2005, 52, 4877–4885.
39. Komaragiri, U., Begley, M.R., and Simmonds, J.G. The mechanical response of freestanding circular
elastic films under point and pressure loads, Journal of Applied Mechanics 2005, 72, 203–212.
40. Easley, C.J., Leslie, D.L., Landers, J.P., Utz, M., and Begley, M.R. “Design of microfluidic diodes and
check valves,” to be published, 2007.
41. Easley, C.J., Leslie, D.C., Karlinsey, J.M., Begley, M.R., and Landers, J.P. “Microfluidic waveform
shaping with passive elastomeric components,” to be published, 2007.
42. Roberts, G. and Sedra, A., SPICE, Oxford University Press, New York, 1996.
43. Hay, J.L. and Pharr, G.M. “Instrumented indentation testing,” chapter in ASM Handbook Volume 8:
Mechanical Testing and Evaluation, 10th edn., H. Kuhn and D. Medlin, Ed., ASM International,
Materials Park, OH, 2000, pp. 232–243.
44. Begley, M.R. and Mackin, T.J. “Spherical indentation in the membrane regime,” Journal of the
Mechanics and Physics of Solids 2004, 52, 2005–2023.
45. Maner, K.C., Begley, M.R., and Oliver, W.C. “Nanomechanical testing of circular freestanding polymer
films with sub-micron thickness,” Acta Materialia 2004, 52, 5451–5460.
46. Dannenberg, H. “Measurement of Adhesion by a blister method” Journal of Applied Polymer Science,
1961, 5, 125.
47. Malyshev, D.M. and Salganik, R.L., International Journal of Fracture Mechanics, 1965, 1, 11.
48. Cotterell, B. and Chen, Z. “The blister test—Transition from plate to membrane behavior for elastic
material,” International Journal of Fracture 1997, 86, 191–198.
49. Williams, J.G. “Energy release rates for the peeling of flexible membranes and the analysis of blister
tests,” International Journal of Fracture 1997, 87, 265–288.
50. Thouless, M.D. and Jensen, H.M. Elastic fracture mechanics of the peel-test geometry, Journal of
Adhesion 1992, 38, 185–197.
40 Practical Fluid Control
Strategies for Microfluidic
Devices
Christopher J. Easley and James P. Landers
CONTENTS
40.1 Microfluidic Flow Control Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1153

40.2 Active Microfluidic Valving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1154
40.2.1 Normally Open Valves Using Soft Lithography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1154
40.2.2 Normally Closed Valves Using Hybrid Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1155
40.2.3 Robust and Scalable Nature of Elastomeric Valving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1157
40.2.4 Manually Operated Torque Valves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1157
40.3 Passive Microfluidic Flow Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1158
40.3.1 Fluidic Resistors for Controlled Flow-Splitting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1158
40.3.2 Gravimetric Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1160
40.3.3 Capillarity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1160
40.3.4 Passive Structural Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1161
40.4 Applications of Active Valving Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1163
40.5 Valve Fabrication and Implementation Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1165
40.5.1 Normally Open Valves Using Soft Lithography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1165
40.5.2 Normally Closed Valves Using Hybrid Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1166
40.5.3 Instrumentation Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1166
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1167
40.1 MICROFLUIDIC FLOW CONTROL INTRODUCTION

Maintaining precise control over fluid within confined microfluidic networks has arguably become
one of the major hurdles in realizing the true potential of microfluidic science. The nearly 20-year-old
promise1 of integrating multiple processes into single devices—whether customized as miniature
factories for synthetic purposes or as microscale total analysis systems (µ-TAS) for sample processing
and analysis—had fallen short of expectations until the recent implementation of practical and robust
microfluidic valving systems.2,3 With the basic valving systems now in place, there is enormous
potential for many unique experiments relevant to a wide range of disciplines from fundamental
physics to applied biology. Furthermore, the development of passive components to complement
these active valves should prove to enhance control even further, while simultaneously reducing
peripheral instrumentation.
This chapter is not intended to be a comprehensive review of microfluidic valving. On the
contrary, we intend to provide a summary of the two most widely utilized valving strategies developed
in the past decade.2,3 Our approach is based on the practicality of implementing these types of valves.
1153
In particular, we propose that the fabrication simplicity and compatibility of active elastomeric
membrane valves2,3 are key factors in their robustness and usefulness, a notion that is supported by
the recent surge in related publications.4−15 The various microelectromechanical systems (MEMS)
approaches to microfluidic valving are typically limited by operational complexity and fabrication,5
thus they are not discussed here; interested readers are referred to a recent review by Oh and Ahn.16
Throughout this chapter, flow control techniques that require an external (off-chip) stimulus
are considered active, while those requiring no external stimulus are considered passive. The two
dominant active valving techniques are discussed in Section 40.2 along with a simplified manual
alternative, while some of the recent passive flow control methods are reviewed in Section 40.3.
Various applications of the active valving systems are discussed in Section 40.4, which is followed by
a section including guidelines that provide the reader with a “beginner’s knowledge” of implementing
these types of valves (Section 40.5). Although this chapter places more emphasis on active valving for
microfluidic flow control, we conclude (Section 40.6) by stating the importance of conducting further
research into practical passive flow control methods that should allow more widespread use of
microfluidic devices in general.
40.2 ACTIVE MICROFLUIDIC VALVING

Depending on the particular device, microfluidic volumes can range from picoliters to microliters,
with typical total solution volumes in the nanoliter range. Before the development of robust valves that
were simple to fabricate, first by the Quake and coworkers2 and then by the Mathies and coworkers,3 it
was nearly impossible to reliably control these small volumes of fluid using practical fabrication meth-
ods. The central theme of these two approaches is the use of a flexible poly(dimethylsiloxane) (PDMS)
membrane as a pneumatically actuated material to produce on-chip valves, or fluidic switches. While
it should be noted that PDMS is not the only elastomeric membrane amenable to such application, it
is the one that has been extensively used thus far. The use of PDMS in microfluidics was pioneered by
the Whitesides and coworkers,17,18 who developed the soft-lithography techniques used by Quake
and coworkers2 to fabricate their valves. The material is well suited for optical interrogation due to its
transparency in the visible spectral region,17 and it is also amenable to the integration of electrochem-
ical detection,19 making it well suited for analytical purposes. Included in this section are summaries
of the architectures and functionalities of the normally open Quake valves2 and the normally closed
Mathies valves,3 which are easily operated by computer control, followed by a brief description of
a low-power, manually operated alternative proposed by the Whitesides and coworkers.20
40.2.1 NORMALLY OPEN21 VALVES USING SOFT LITHOGRAPHY

Figure 40.1a and b illustrates the basic architecture of valves developed by the Quake and
coworkers.2,22 These normally open, pressure-actuated valves are fabricated in two configurations,
push-down2 or push-up22 valves. The push-down valves (Figure 40.1a) require approximately an
order of magnitude larger pneumatic actuation pressure to close than do the push-up valves (Figure
40.1b). Furthermore, the fluidic channels in the push-down configuration are restricted to a depth of
less than ∼20 µm, while the push-up valves will function with much deeper channels while simul-
taneously requiring lower pressures. A key feature of these types of valves is the essentially zero
dead volume, as can be inferred from the illustration of pressure-actuated closing of the fluid chan-
nel (Figure 40.1c). With this advantage, the valves can be arranged into high-density architectures
(≥30 valves mm−2 ),23 as shown in Figure 40.1d and e. With at least three valves coupled in series,
peristaltic pumps can be created,2 allowing the precise metering of solutions within the device at
flow rates up to 2.35 nL s−1 . Figure 40.1e shows a magnified view of Figure 40.1d, with labels for
the peristaltic pumps and cell growth chambers used in this bacterial chemostat study.23
Practical Fluid Control Strategies for Microfluidic Devices 1155
(a) PDMS (c) Pressure (d)
Fluidic Closed fluidic

channel
Pneumatic
(b)
(e) E
Glass
FIGURE 40.1 Basic architecture of Quake’s normally open valves. (a) Push-down configuration, which is
typically limited to fluidic channels <20 µm. (b) Push-up configuration, which requires approximately an order
of magnitude less pressure to actuate than the push-down valves and can be used with deeper channels. (Adapted
from Studer, V., et al. J. Appl. Phys., 2004, 95, 393–398.) (c) Illustration of pressure-actuated closing of the fluid
channel using a push-up valve. (d) High-density valve arrangements were shown in a microfluidic chemostat
for programmed population control of bacteria. (e) Magnified view of device in (d). (From Balagadde, F. K., et
al. Science, 2005, 309, 137–140. With permission from American Association for Advancement of Science.)
Since their development, these valves have been proven functional through their utilization in a
wide variety of applications such as microfluidic large-scale integration,4 protein crystallization,24
nucleic acid processing,6 and multistep radiolabel synthesis.9 These applications will be discussed
further in Section 40.4. One disadvantage of these valves is that, in order to actuate a large number of
individually addressable valves, there must be an equal number of external pneumatic controls (typ-
ically computer-controlled solenoid valves). In this situation, the controlling instrumentation begins
to overwhelm the microscale nature of the device. Although multiplexers have been developed,4
these pose restrictions on individual addressability of the valves. In other words, a continuous exter-
nal pressure is required on each valve to maintain actuation, thus a latchable fluidic control structure
has yet to be developed in this valve configuration.
40.2.2 NORMALLY CLOSED21 VALVES USING HYBRID DEVICES

Alternatively, the valves developed by the Mathies and coworkers3 are fabricated in a normally
closed architecture, as shown in Figure 40.2a, and are actuated open by vacuum, with the option
of pressure to promote closing. Rather than patterning the PDMS by soft-lithography, these valves
are defined by patterned glass layers (chemically etched), which sandwich a PDMS membrane.
The fact that the majority of the channel architecture can be made from glass is advantageous,
parleying the rigidity and chemical resistivity of the glass substrate with the system, while maintaining
valving capabilities. These valves can be fabricated in either three-layer (Figure 40.2a) or four-layer
(Figure 40.2b) architectures, in which the four-layer design minimizes contact with PDMS surfaces
that have a known incompatibility with certain solvents and analytes.25 Figure 40.2c illustrates the
actuation of the three-layer valves, where a vacuum is applied to the pneumatic control chamber to
open the fluid path below; this valving action is mechanistically equivalent to the four-layer valve
actuation. Either type of valve can be arranged into a diaphragm pumping configuration (at least three
valves), with reported flow rates up to 380 nL s−1 ,3 approximately two orders of magnitude larger
(a) Glass (c)

Vacuum
Pneumatic
Fluidic
PDMS (d) Slipper

(b) Fluidic
reservoir
Pneumatic 15
13
11
9
bc 7
a 5
3
1
Separation Microfluidic
channel bus
(e) Slipper
Drilled Fluidic
Fluidic
vias
PDMS
Manifold
Channel
FIGURE 40.2 Basic architecture of Mathies’ normally closed valves. (a) The three-layer configuration in
PDMS-glass hybrid devices and (b) the four-layer configuration (used to limit PDMS-solution contact) can be
(c) actuated open to fluid flow by applying a vacuum to the pneumatic chamber at the valve seat. (Adapted
from Grover, W. H., Skelley, A. M., Liu, C. N., Lagally, E. T., Mathies, R. A. Sens. Actuat. B, 2003, 89,
315–323.) (d) An example of the assembled four-layer device that was used to detect evidence of life in extreme
environments in the MOA. (e) The preassembled MOA device. (From Skelley, A. M., et al. Proc. Natl. Acad.
Sci. USA, 2005, 102, 1041–1046. Copyright 2005. With permission from National Academy of Science, USA.)
than the Quake pumps. An illustration of an exemplary four-layer device is shown in Figure 40.2d
(assembled) and Figure 40.2e (pre-assembly). This device was utilized for amino acid detection in
extreme environments with the long-term goal of searching for evidence of life on Mars.8,14 Since
these valves were developed three years after the Quake valves, they have been utilized in fewer
applications to date. However, recent reports describe the use of these valves for DNA computing,26
pressure injections for electrophoresis chips,10,27 DNA sequencing,28 serial dilution circuits,29 and
fully integrated genetic analysis devices.11 A disadvantage of these valves is that the dead volume
is not negligible, which is also related to the lower achievable valve density compared to the Quake
valves, although the valves can be designed for <10 nL dead volume3 and have shown capable of
injection of samples that are submicroliter in volume.10,27
An interesting modification of this valve configuration was recently made by the Mathies group,
in which the researchers showed that it was possible to fabricate “latching” structures that no longer
required a continuous external stimulus.12 Vacuum pulses of 120 ms were sufficient to latch the valves
for up to 2 min. This was an important advance in flow control technology, because it demonstrated
that 2(n−1) valves could be controlled with only n control lines. Although the authors showed just
five inputs to control 16 valves, it should be possible to extrapolate the concept, for example, to
control 1024 valves with only 11 external control lines, thereby greatly reducing the complexity of
the external hardware associated with valve control.
150 Including "valve" or "pump"
Citing Quake et al.
"Microfluidic" articles
Citing Mathies et al.
100
1000
100
50 10
1
0
1990 1994 1998 2002 2006
Publication year
FIGURE 40.3 The ISI Web of Science database was searched with the subject keywords of “microfluid∗
AND (valv∗ OR pump∗ )” through December 2006. These results were plotted as a function of publication
year (solid triangles), and the totals were further limited to those which cite the initial valve descriptions by
the Quake (open circles) or Mathies (open squares) group. The surge of publication since the introduction of
the Quake valves is evident in either the linear plot or the inset logarithmic plot. Ninety-seven percent of the
total “microfluidic valve or pump” publications in 2006 cited either the Quake or Mathies. Key reasons for the
success of these valves are simplicity of fabrication, valve actuation robustness, and design scalability.
40.2.3 ROBUST AND SCALABLE NATURE OF ELASTOMERIC VALVING

The Quake2 and Mathies3 valve configurations can arguably be considered as the most robust that
have been developed to date. This argument is supported by the surge in publications that have
referenced these valves. Using the ISI Web of Science database,30 a search was performed including
the terms “microfluid* AND (valv∗ OR pump∗ )” through December 2006. These results were plotted
as a function of publication year (Figure 40.3, solid triangles), and the resulting publications were
further limited to those that cite the initial valve descriptions by the Quake (open circles) or Mathies
(open squares) group. It is clear from Figure 40.3 that the number of publications including “valves”
or “pumps” began to surge after the introduction of Quake’s valves, a fact that is visually confirmed
in the inset logarithm-scale plot. Although the surge could have occurred due to general scientific
progress alone, this notion is refuted by the fact that an average of 76% of all the publications
referenced either Quake or Mathies since 2001, with 97% in 2006. While this representation of
the literature is not infallible, Figure 40.3 demonstrates the general importance of these elastomeric
membrane valves. The key reasons for the success of these types of valves are the simplicity with
which they can be fabricated, the valve actuation robustness, and the design scalability. The keys
to “fabrication simplicity” are that the fabrication methods necessary for either configuration2,3 are
essentially no more complex than typical microdevice fabrication,17,31 and that they have been shown
to be highly scalable, for example, with ≥30 valves mm−2 .4,5
40.2.4 MANUALLY OPERATED TORQUE VALVES

For certain applications, portability and power savings outweigh the need for highly scalable,
computer-controlled valving. It should be clear to those experienced in elastomeric valving on
nontrivial chip-based architectures that the external hardware also becomes nontrivial in both size
and complexity. A useful alternative was presented by Whitesides and coworkers,20 in which torque-
actuated valves were used to collapse PDMS channels, thereby closing the channels to fluid flow. The
mechanistic action of these valves was similar to the Quake valves,2 in that a pressure was applied
above a thin layer of PDMS to collapse a fluidic channel. However, since the torque-actuated valves
(coined as TWIST valves) relied on a miniature machine screw to collapse the channels, they did
not require power to maintain the valves in either the open (flow on) or closed (flow off) state indef-
initely, thus functioning as latchable fluidic switches. Moreover, these valves were also shown to be
amenable to settings between “on” and “off,” thereby functioning as variable fluidic resistors. The
machine screws were fixed into recessed regions of the PDMS devices through a simple photochem-
ical curing process that embedded them into polyurethane, which simultaneously bonded with the
PDMS. These devices were used to carry out sandwich immunoassays.20 The TWIST valves have
since been proven useful for the development of entirely portable microfluidic systems capable of
flow rate control, sample introduction, filtration, mixing, and bubble generation.32
40.3 PASSIVE MICROFLUIDIC FLOW CONTROL

Although active valves have clearly been proven as effective fluid control elements, the ultimate
usefulness of these valves is inherently limited by the pneumatic interfacing problem. As demands
for functionality increase, the number of valves must also increase; therefore, the controlling instru-
mentation often begins to overwhelm the microscale nature of the device. In this section, several
methods for passive fluid control are summarized. As before, we do not intend to provide a compre-
hensive review of passive flow control, only a sampling of the more popular methods. Although there
is limited review literature on this subject, the interested reader is referred to an informative focus
article on capillarity by Eijkel and van den Berg,33 who refer to passive flow control methods as “a set
of everyday forces that are ‘always there’ . . . gravity, suction, and the capillary force . . . .” Whenever
possible, these and other passive flow control strategies should be utilized as alternatives to active
valving, and active valves should be used only when necessary. Since passive components require no
external stimulus, careful management of active and passive methods—working in concert—should
offer the most optimal balance of instrumentation and space.
40.3.1 FLUIDIC RESISTORS FOR CONTROLLED FLOW-SPLITTING

A fundamental mode of design-driven passive flow control in microfluidic devices is to utilize
differential fluidic flow resistances to direct the majority of flow through the least resistant path. This
straightforward, yet powerful, technique is illustrated in Figure 40.4. When a pressure greater than
atmospheric pressure is applied to the inlet channel of a device, and this channel is split, by design,
into two exit channels, the flow rate of fluid through these channels depends on their relative flow
resistance values. The flow resistance, in turn, increases with an increase in length or a decrease
in cross-sectional area of the channel. In fact, Harrison and coworkers34 have outlined a simple
method to calculate flow resistance values, and they showed that highly accurate flow splitting could
be achieved simply by designing the channel layout accordingly. The researchers started with the
expression for the average linear flow rate, U, in a rectangular channel
wd P
U= F, (40.1)
η L
where w, d, and L are the half-width, half-depth, and length of the microchannel, respectively; η is
the viscosity of the fluid; and P is the pressure difference along the channel. The value F refers to
a geometric form factor that depends on the ratio d/w (when d ≤ w), which can be calculated by

∞ (2n+1)πd
w 64w2 tan h 2w
F= − 5 2 . (40.2)
3d π d (2n + 1) 5
n=0
(a) (b) (c) (d)

P = 1 atm P = 1 atm P = 1 atm V=0
(ground)
R1 R2
Flow
Rin
P > 1 atm P > 1 atm P > 1 atm V>0
FIGURE 40.4 Output fluid levels at an arbitrary time point illustrate the effects of length and cross-sectional
area of opposing exit channels. (a) Conceptual device designed for equal flow resistance in both exit paths,
forcing the output flow rates to be equal upon application of a pressure gradient. Other devices were designed
for a 10-fold larger flow resistance in the left exit path using (b) increased channel length or (c) decreased
cross-sectional area, thereby forcing the output flow rates in their right exit paths to be 10-fold larger than the
opposing (left) paths. (d) This behavior can be modeled using simple electrical circuit diagrams as analogs to
fluidic networks.
Equation 40.1 was then manipulated to give the pressure difference as follows:

Q ηL 4ηL
P = =Q , (40.3)
A wdF (wd)2 F
where Q is the volumetric flow rate and A is the cross-sectional area of the rectangular channel
(note that U = Q/A). The right-hand side of Equation 40.3 represents the Ohm’s law (V = IR)
equivalent for microfluidics, in which P is analogous to the voltage drop and Q is analogous to
current flow through a wire, meaning that the term in parentheses on the right in Equation 40.3
represents the fluidic resistance,
4ηL
R= . (40.4)
(wd)2 F
In the device of Figure 40.4a, fluidic flow rates through each exit channel are equal due to their
identical flow resistance values. However, in the devices of Figure 40.4b and c, the left exit channels
were designed to be 10-fold more resistant to flow than the opposing (right) exit channel. This was
accomplished in the Figure 40.4b device by simply designing a 10-fold longer channel in the left
exit path, while the Figure 40.4c device accomplished the identical goal by reducing the cross-
sectional area of the left exit path. Essentially independent of the magnitude of the input pressure,
the flow rates through these devices is 10-fold larger in the right exit path (by design), as illustrated
by the fluid levels at an arbitrary time point shown in Figure 40.4a–c. This flow behavior can be
accurately modeled using the analogy to Ohm’s law in electrical circuits, via Equation 40.3. The
electrical circuit equivalent to the devices in Figure 40.4 is shown in Figure 40.4d, where R1 = R2
in Figure 40.4a device, while R1 = 10 R2 in the Figure 40.4b and c device.
Unfortunately, even though control of relative flow resistances within channel networks is
straightforward, it appears to be somewhat overlooked by much of the microfluidic research commu-
nity. Through more judicious design constraints, researchers could benefit greatly from this simple
method. Harrison and coworkers34 utilized relative flow resistances to reproducibly sample small
volumes for electrophoresis injections from a much larger sample introduction channel. Landers
and coworkers also used this technique for microchip electrophoresis (ME) injections from on-chip
diaphragm pumps,10,11 where the injection plug size was minimized by carefully designed flow
splitting ratios at the injection cross-tee. Lam et al.35 combined valving with flow resistances, in this
case to produce a digitally variable fluidic resistor with 16 possible flow resistance values. A differ-
ent approach was employed by Whitesides and coworkers,36 who were able to show analog/parallel
computing capability using shortest-path determinations in microfluidic mazes. City maps were used
as mask patterns to fabricate channel networks, and the shortest path between two points in the “city”
were determined rapidly, suggesting that parallel computing with microfluidics is a viable alternative
for some computationally difficult problems.
40.3.2 GRAVIMETRIC CONTROL

When possible, flow control based on gravity alone can provide a simple, yet useful, alternative to
active valving. These siphoning effects can provide smooth pressure profiles based on Equation 40.5.
The pressure difference between input and output reservoirs, P, can be described by
P = ρgL (40.5)
with a working fluid of density ρ and height L, where g is the gravitational acceleration constant.
Although it is difficult to generate large pressures using this method, it can be sufficient for many
applications using channels with sufficiently large cross-sectional area (small fluidic resistance). For
example, Du et al.37 utilized gravity-driven flows for an automated flow injection analysis system for
on-chip core-waveguide spectrometric detection of the complexation of o-phenanthroline with Fe(II).
Highly precise flow control was achieved by Whitesides and coworkers,38 who utilized reservoirs
on motorized stages to adjust the L term of Equation 40.5, providing switching between forward
and reverse flow. This gravity-based flow was used to control droplet fluidics for encryption and
decryption of signals coded in the intervals between droplets. Disadvantages of this method include
pressure variability over time due to evaporation or source depletion, difficulty in flow switching,
and interference from feedback pulses based on device compliance.39
40.3.3 CAPILLARITY
In the aforementioned focus article by Eijkel and van den Berg,33 the authors highlight the fact
that capillary forces scale better with miniaturization when compared to other passive forces such
as gravity, suction, centrifugation, or evaporation. To this end, Walker and Beebe40 showed that
useful pumping work could be performed in microfluidic systems by simply controlling the shape
of droplets at fluidic input ports. They showed that the highest pressure attainable for a given port
radius is a hemispherical drop with a radius equal to that of the port. A flow rate of 1.25 µL s−1 was
demonstrated using only a 0.5 µL drop of water, and the method was shown to be strong enough
to pump against gravitational potential. In other words, based on the Young–Laplace equation, the
smaller droplet possessed a higher internal pressure than the larger drop, thus was capable of pumping
against a much larger droplet in an opposing reservoir simply based on the shape of the droplet at
the air–water interface40 (Figure 40.5). The change in volume with respect to time was described in
dV 1 2γ
= ρgL − , (40.6)
dt R r
where r is the radius of the small spherical “pumping” drop, R is the microchannel resistance (see
Equation 40.4), and γ is the surface free energy of the liquid. The pressure created by the larger
outlet reservoir drop of density ρ and height L is represented by the gravimetric term ρgL (see
Reservoir drop
Pumping drop
Flow direction
FIGURE 40.5 Capillarity was shown to be a more dominant force at the microscale by a passive pumping
method. The small “pumping” drop (right) was designed to provide enough capillary force to overcome the
gravimetric force from the large “reservoir” drop (left), thus the flow progressed from right to left in the figure
(see Equation 40.6). (Adapted from Walker, G. M., Beebe, D. J. Lab Chip, 2002, 2, 131–134.)
Equation 40.6), which opposes the capillary pressure but is typically much smaller in magnitude in
microscale systems. This technique was shown by Walker and Beebe40 to be useful for injecting a
plug within a microchannel for electrophoresis applications.
Eijkel and van den Berg33 also noted that the capillary pressure can be tuned by changing the
contact angle or the device geometry. When pumping in this a constant cross section, the meniscus
within the channel proceeds with the square root of time due to the linear increase in flow resistance as
the channel fills. Different geometries and cross sections can be used to control the relative wetting of
channels within a microfluidic network, providing more precise control over these autonomous sys-
tems. Kim et al.41 showed this method to be capable of passive control of two merging laminar streams
whose widths could be tuned based on flow resistance of downstream channels. Capillarity can also
be used to control sample and sieving polymer loading for DNA separations, as shown by Ono and
Fujii.42 This method allows for DNA injection and separation with a simple two-electrode system.
40.3.4 PASSIVE STRUCTURAL COMPONENTS

Instead of merely exploiting microscale physical phenomena using simple straight-channel patterns
with constant cross section, it is certainly feasible to make use of well-developed microfabrication
methods for introducing flow control functionality into individual components. Although various
MEMS approaches have been utilized for this purpose (as reviewed by Oh and Ahn16 ), many of
these techniques require complex fabrication and will not be discussed here. This section focuses on
more practical passive components that can be fabricated in PDMS or PDMS-glass hybrid devices,
requiring minimal fabrication complexity in addition to the typical microchannel assembly methods.
Passive flow rectifiers analogous to electrical diodes have been developed by several groups.43−45
In a single channel, these components allow larger flow rates in the “forward” direction over the
“reverse” direction. The flow rectifiers developed by the Landers and coworkers45,46 provide the most
facile fabrication of any reported to date (mask design shown in Figure 40.6a, with magnified view
of rectifier inset; fluidic layer in black, pneumatic in gray). These components can be fabricated
alongside the normally closed valves/pumps developed by the Mathies and coworkers,3 with no
additions to the fabrication steps other than the essentially negligible additions to the mask design
process. As shown in Figure 40.6b, the rectifiers45,46 were shown to eliminate negative flow pulses
that are inherent to the on-chip diaphragm pumps (flow rate of a diaphragm pump, gray trace; rectified
flow, black trace). This behavior was shown to mimic the behavior of an electrical half-wave rectifier
circuit (Figure 40.6c and d).
Easley et al.45,46 also provided proof-of-principle data for discrete components referred to as
“fluidic capacitors.” While the concept of fluidic capacitance is by no means a new concept, little
work has been done to exploit this behavior in microfluidic networks. To the knowledge of the authors,
only recently has there been a report that discusses the fluidic compliance of an entire microchip
made up of a flexible polymer, PDMS.39 In this work, which was concurrent with the work by Easley
(a) 8 Fluidic data (b)

4
2
Flow rate (nL s–1)

Resistor 0
Diaphragm
pump –4
–8
–12
Rectifier
2
–16
0.0 0.5 1.0 1.5 2.0
Time (s)
(c) (d)
Voltage Numerical simulation
source Resistor
Diode Voltage
Time (s)
FIGURE 40.6 Fluidic flow rectifiers that require no additional fabrication steps above the steps required to
construct the Mathies valves. (a) Mask design of a fluidic half-wave rectifier, including a diaphragm pump,
flow rectifier (with inset image), and a fluidic resistor for flow visualization. Fluidic layer is in black, while
the pneumatic layer is in gray. (b) Flow rate profiles from these rectifiers (black trace) revealed their suc-
cess in eliminating negative flow pulses that are inherent to the on-chip diaphragm pumps (gray trace). The
fluidic flow rate profiles matched qualitatively with the behavior of an analogous electrical circuit, (c) a half-
wave rectifier using a voltage source in series with a single diode, where the (d) voltage profiles of the pump
(gray trace) and the rectified flow (black trace) are shown here. (Adapted from Easley, C. J., et al. Micro-
Total-Analysis-Systems, 10th International Conference on Miniaturized Systems for Chemistry and Life
Sciences, Tokyo, Japan, November 2006.)
et al., Beebe and coworkers showed that the dynamic characteristics of the entire system were similar
to a lowpass filter in electrical circuits. However, there were no attempts to exploit these effects using
individual microfluidic components. Easley et al. proposed the use of discrete components to control
the fluidic capacitance of individual fluidic networks within the same device.45,46 This treatment was
based on the electrical analogy, where the volumetric flow rate, Q, through microchannels in the
presence of compliant membranes is dependent on the time derivative of the pressure, P, through
the channel as below:
dP
Q=C , (40.7)
dt
where C is the fluidic capacitance in units of mm3 kPa−1 , representing the volume stored in the
compliant membrane per applied pressure. This way, the frequency response of a first-order system
could be characterized based on similar equations. For example, a low-pass or high-pass filter could
be characterized by its cutoff frequency, f0 , at which the power output of a circuit is reduced to
one-half of its maximum value. Using the fluidic capacitance and resistance analogies presented
above, the fluidic cutoff frequency can be calculated by
1
f0 = . (40.8)
2π RC
250/500 µm
250/250 µm
50/300 µm
Normalized deflection
3.5 Hz
0.7 Hz 8.0 Hz
0.01 0.1 1 10 100

Actuation frequency (Hz)
FIGURE 40.7 Characteristic frequencies could be shifted by an order of magnitude by locally altering PDMS
membrane thicknesses (“fluidic capacitors”). Different curves represent different combinations of the thick-
nesses of two fluidic capacitor components in the same fluidic network. These fluidic bandpass filters provide
the proof-of-concept of a new paradigm in microfluidic flow control, where actuation frequency could be used
to passively control relative flow rates.
This equation was also deduced by Beebe and coworkers.39 In fact, Easley et al.45,46 have recently
shown that multiple fluidic capacitor components could be combined to produce actuation-frequency-
dependent control over flow rates through microfluidic networks. These fluidic bandpass filters with
tunable fundamental frequencies, which should provide a new paradigm in flow control, are currently
in development. As shown in Figure 40.7, the fundamental frequencies of the fluidic bandpass filters
could be shifted by an order of magnitude, from 0.7 to 8.0 Hz, by simply altering the thicknesses of the
deflectable PDMS layers (fluidic capacitors) at discrete locations within the network. Combinations
of membrane thicknesses between 50 and 500 µm were used in this work (see legend in Figure 40.7).
Finally, Kartalov et al.13 recently presented a straightforward method for creating microfluidic
vias, which allowed solutions to be passed between multilayers of channel networks. The fabrication
of these vias was consistent with the multilayer soft-lithography methods17,18 used by Quake and
coworkers.2 A novel, passive flow shaping component, coined an “autoregulator,” was enabled using
these vias. The autoregulator was shown to provide the fluidic equivalent to an electrical current
source, in which the output flow rate was essentially independent of input pressure above a certain
threshold pressure.13 These components were also shown to possess flow rectification character.
With further development of the types of passive components outlined in this section—ones that are
complimentary to the now well-developed valving techniques—the true potential of microfluidic
flow control could be realized.
40.4 APPLICATIONS OF ACTIVE VALVING SYSTEMS

As illustrated by Figure 40.3, microfluidic valving techniques have proven highly useful in recent
years. The Quake valves, in particular, have been put to use for a wide variety of functionali-
ties from synthesis to genetic screening. Hansen et al.24 developed a method for the screening of
conditions for protein crystallization based on free-interface diffusion. The researchers were able
to accurately meter solutions on the picoliter scale in a highly integrated and parallel fashion. A
spectrum of screening conditions, covered by 144 parallel reactions, was carried out with each
using only 10 nL of protein sample, to allow growth of diffraction-quality crystals. The device
was shown to outperform conventional techniques, while using a mass of protein sample that was
smaller by two orders of magnitude. More recent work has shown this technique to be useful for
crystallizing challenging proteins by using knowledge of phase behavior to generate more rational
screening tests.47 Balagadde et al.,23 with the device shown in Figure 40.1d and e, utilized the
Quake valves to provide long-term programmed population control of bacterial in a microchemo-
stat. A feedback mechanism, based on bacterial quorum sensing, was used to regulate the cell density
of extremely small cell populations with single cell resolution over hundreds of hours. Lee et al.9
fabricated a “miniature chemical factory” for multistep synthesis of a radiolabeled molecular imaging
probe, 2-deoxy-2-[18 F]fluoro-d-glucose, in an integrated microfluidic device. The device was able to
carry out five sequential processes: [18 F]fluoride concentration, water evaporation, radiofluorination,
solvent exchange, and hydrolytic deprotection. This automated method provided high radiochemical
yield and purity with shorter synthesis time relative to conventional automated synthesis, and the
products were used successfully for positron emission tomography (PET) imaging in mice.
The Mathies valves have also found practical use for a variety of applications. Indeed, the slight
chronological shift associated with the development of the Mathies valving system compared to the
Quake system (Figure 40.3) suggests that the Mathies valves will perhaps adopt a similar trend. Skel-
ley et al.8 have developed a Mars Organic Analyzer (MOA) (device schematic shown in Figure 40.2d
and e), which was capable of detecting a molecular signature of life (homochiral amino acids) in the
Atacama Desert, Chile, and in the Panoche Valley, CA—two of the driest locations on Earth. The inte-
grated and portable MOA system consisted of a microfabricated capillary electrophoresis instrument
for sensitive amino acid biomarker analysis, along with the necessary high voltage power supplies,
pneumatic controls, and fluorescence detection optics, and was capable of amino acid detection in
mineral deposits as low as 70 parts per trillion. This system has since been shown to be capable of
analyzing a wide variety of fluorescamine-labeled amine-containing biomarker compounds, includ-
ing amino acids, mono and diaminoalkanes, amino sugars, nucleobases, and nucleobase degradation
(a) (b) V2
PR
MR
V2
99.3%
V1
RPCR RPCR RPCR <<REW
TR
V5 <1%
BR
V4
EW SA EW H2O EW PCR
SW V3 V1 Gdn
mix
SI /IPA H2O
(c)
EW SA
60 x 85 mm
DNA (ng)/intensity (a.u.)
0.10 100
Temperature (°c)
SPE
FA SI PCR
ME
0.05 75
0.00 50
0 8 16 24 32
Total analysis time (min)
FIGURE 40.8 A clever combination of the Mathies valves and passive flow control methods allowed full
integration of sample cleanup and analysis steps directly from blood or nasal aspirates, resulting in (a) a
microfluidic genetic analysis (MGA) device. (b) The combination of valving, laminar flow, and differential
channel resistances was used to prevent fouling of PDMS valves with incompatible solvents when transitioning
from SPE to the PCR. (c) SPE and PCR were then coupled to ME using valve-based pumping to inject the
purified DNA sample. The presence of Bacillus anthracis could be detected directly from murine blood in less
than 24 min, and multiple injections were allowed using the valves to confirm the diagnoses. (From Easley, C. J.,
et al. Proc. Natl. Acad. Sci. USA, 2006, 103, 19272–19277. Copyright 2006. With permission from National
Academy of Science, USA.)
products.14 The Mathies group has also applied their valves to DNA sequencing on a microfluidic
format,28 integrating thermal cycling, sample purification, and capillary electrophoresis.
The Landers and coworkers27 first utilized the Mathies valves to show their utility for pressure
injection on a microfluidic format, then applied this technology to the integration of DNA ampli-
fication (polymerase chain reaction, PCR) with electrophoretic separation.10 As shown in Figure
40.8, this technology was combined with upstream sample cleanup (solid-phase extraction, SPE) of
crude clinical samples such as blood and nasal aspirates, providing fully integrated genetic analysis
in <30 min.11 The device architecture is shown in Figure 40.8a, flow control steps between SPE
and PCR are illustrated in Figure 40.8b, and the results of a <24 min detection of Bacillus anthracis
directly from the blood of an asymptomatic mouse are shown in Figure 40.8c. This work provided
true sample-in, answer-out analyses of complex samples, as predicted by Manz et al.1 over a decade
ago. The work was carried out on a relatively simple device design, for the authors utilized a creative
combination of active and passive flow control methods to interface the SPE and PCR steps (Figure
40.8c), where laminar flow and relative flow resistance values were exploited to prevent incom-
patible solvents from contacting the PDMS valves. Most recently, Guillo et al.15 used these valves
for pressure mobilization of focused protein zones following isoelectric focusing (IEF) to provide a
robust microchip alternative to conventional IEF.
40.5 VALVE FABRICATION AND IMPLEMENTATION GUIDELINES

This section provides brief summaries of fabrication methods and instrumentation requirements asso-
ciated with the microfluidic, pneumatic valving techniques outlined above. This section is meant
to provide a general guide for fabrication. For more detailed descriptions of chip fabrication meth-
ods, see Chapter 10 by Legendre and Landers. Readers interested in implementing these valving
techniques should refer to the original publications2,3,22 for details and/or clarity. The organization
of this section parallels that of Section 40.2 in that the normally open and normally closed valve
configurations are separately described.
40.5.1 NORMALLY OPEN VALVES USING SOFT LITHOGRAPHY

In order to achieve either the push-down2 or push-up22 versions of the Quake valves, as shown in
Figure 40.1a and b, standard multilayer soft-lithography equipment is necessary. For either valve
configuration, typically a layer of SU-8 (negative) or AZ (positive) photoresist is spun onto two
silicon wafers to the desired thicknesses. One wafer is then exposed to UV radiation using the fluidic
photomask, while the other is exposed using the pneumatic photomask. The fluidic layer master is
fabricated out of AZ photoresist, which can then be reflowed with heat to give a rounded channel
profile, and the pneumatic layer is fabricated out of SU-8 photoresist. The methods deviate at this
point, for the push-down valves (Figure 40.1a) require the fluidic layer to be below the pneumatic
layer, while the push-up valves (Figure 40.1b) require the opposite. A thin PDMS film (10–100 µm)
is then spun onto the lower layer’s master (whether fluidic or pneumatic), with the thickness of
the PDMS above the master determining the valving membrane thickness. A thick film of PDMS
(1–10 mm) is simply poured onto the opposing layer’s master, at which time the two layers are cured,
then the upper layer is peeled from the master, aligned with the lower layer, and mated. In order to
mate the two layers, it is often preferable to prepare the thin PDMS film with an excess of elastomer
base (∼20:1, base:curing agent), while the thick film is prepared with an excess of curing agent
(∼5:1, base:curing agent).2 This way, the two layers can be essentially fused into one by heating
postassembly at 60–80◦ C for 1–2 h. Finally, the PDMS device is peeled from the master, access
holes are punched, and the device is completed by either mating with another (unpatterned) layer of
PDMS or a glass slide.
Note that the push-down valves (Figure 40.1a) require approximately an order of magnitude
larger pneumatic actuation pressure to close than do the push-up valves (Figure 40.1b) due to the
preferential geometry of the push-up configuration.22 Furthermore, the fluidic channels in the push-
down configuration are restricted to a depth of less than ∼20 µm, while the push-up valves will
function with much deeper channels while simultaneously requiring lower pressures. However,
in cases where sample contact with a glass substrate (or some other material mated to PDMS) is
necessary, such as high-resolution confocal microscopy, the push-down valves must be used due to
the lower position of the fluidic layer. On the other hand, the recently developed microfluidic via
technology13 may provide a remedy to this problem in some cases.
40.5.2 NORMALLY CLOSED VALVES USING HYBRID DEVICES

When a normally open configuration is not sufficient, or when glass fluidic devices are preferred over
PDMS devices, the Mathies valve configuration provides a useful alternative.3 These valves also have
two configurations, the three-layer (Figure 40.2a) or four-layer (Figure 40.2b) mode. When PDMS
material compatibility is not an issue, the three-layer configuration should be adequate. However,
if certain reagents are incompatible, the four-layer configuration can be used instead to minimize
PDMS/solution contact. With either configuration, the photolithographic patterning is carried out in
a glass substrate rather than PDMS. The PDMS portion of these devices can simply be purchased in
thin sheets, and slices can be cut to the appropriate size. Glass slides can be purchased with prespun
photoresist, and fluidic/pneumatic layers can be etched into these slides using standard wet etching
methods with hydrofluoric acid (HF). (Warning: Exposure to HF can produce harmful health effects
that may not be immediately apparent, such as destruction of deep tissue and even bone. Extreme care
should be taken when working with these solutions.) For the three-layer configuration, the PDMS
layer (valving membrane) is then sandwiched between the aligned fluidic and pneumatic layers. With
the four-layer configuration, access holes must be drilled into another glass slide, in alignment with
the valves (two holes per valve), and this slide must be thermally bonded to the fluidic layer.
Although the fabrication of these valves is relatively simple, common problems associated with
these valves include delamination of the PDMS layer from the glass layers at higher fluidic or
pneumatic pressures, resulting in sample or solvent leakage. In addition, the dead volume of these
valves is not negligible, and care must be taken to avoid bubble trapping within the valve seats,
especially in the four-layer configuration (Figure 40.2b). Then again, there appears to be no evidence
of an upper-limit restriction on fluid channel depth, for the valve actuation dimensions are separate
from those of the fluid channel.
40.5.3 INSTRUMENTATION REQUIREMENTS

The common theme in these microfluidic valving systems is the requirement of a vacuum or pressure
pump/compressor that is used to actuate the pneumatic lines that are interfaced with the on-chip
valves. These pneumatic lines, in turn, are usually controlled individually using computer-operated
solenoid valves with the corresponding circuitry and manifolds. Various tubing can be used to
connect each solenoid valve to the corresponding microfluidic interface. Interfacing to the device
can be accomplished through a variety of means, which are not discussed here. Interested readers
should refer to the cited works that are particularly relevant to their own research.

In writing this chapter, the intent of the authors was to provide the reader with a rudimentary
knowledge of the practical aspects associated with elastomeric valving technology—one that could
be directly implemented in the laboratory. At this point, the reader should have a basic understanding
of the mechanics of actuation for the two featured valving methods,2,3 as well as an understanding
of how innovative design of passive components can considerably improve flow control with or
without valves in place. It should be noted that for many microfluidic applications, the exquisite
control of small volumes that is attainable with these valves and components is unparalleled. Thus,
researchers should not be hesitant to implement these techniques in their laboratories. In the years
to come, once microfluidic researchers have a set of tools comparable to the various active and
passive circuit components available to electrical engineers, many previously undiscovered avenues
of experimentation will undoubtedly be revealed.
REFERENCES
1. Manz, A., Graber, N., Widmer, H. M. Sens. Actuat. B, 1990, 1, 244–248.
2. Unger, M. A., Chou, H.-P., Thorsen, T., Scherer, A., Quake, S. R. Science, 2000, 288, 113–116.
3. Grover, W. H., Skelley, A. M., Liu, C. N., Lagally, E. T., Mathies, R. A. Sens. Actuat. B, 2003, 89,
315–323.
4. Thorsen, T., Maerkl, S. J., Quake, S. R. Science, 2002, 298, 580–584.
5. Hong, J. W., Quake, S. R. Nat. Biotechnol., 2003, 21, 1179–1183.
6. Hong, J. W., Studer, V., Hang, G., Anderson, W. F., Quake, S. R. Nature Biotechnology 2004, 22,
435–439.
7. Lagally, E. T., Scherer, J. R., Blazej, R. G., Toriello, N. M., Diep, B. A., Ramchandani, M., Sensabaugh,
G. F., Riley, L. W., Mathies, R. A. Anal. Chem., 2004, 76, 3162–3170.
8. Skelley, A. M., Scherer, J. R., Aubrey, A. D., Grover, W. H., Ivester, R. H., Ehrenfreund, P., Grunthaner,
F. J., Bada, J. L., Mathies, R. A. Proc. Natl. Acad. Sci. USA, 2005, 102, 1041–1046.
9. Lee, C. C., Sui, G., Elizarov, A., Shu, C. J., Shin, Y. S., Dooley, A. N., Huang, J., et al. Science, 2005,
310, 1793–1796.
10. Easley, C. J., Karlinsey, J. M., Landers, J. P. Lab Chip, 2006, 6, 601–610.
Hughes, M. A., et al. Proc. Natl. Acad. Sci. USA, 2006, 103, 19272–19277.
12. Grover, W. H., Ivester, R. H., Jensen, E. C., Mathies, R. A. Lab Chip, 2006, 6, 623–631.
13. Kartalov, E. P., Walker, C., Taylor, C. R., Anderson, W. F., Scherer, A. Proc. Natl. Acad. Sci. USA,
2006, 103, 12280–12284.
14. Skelley, A. M., Cleaves, H. J., Jayarajah, C. N., Bada, J. L., Mathies, R. A. Astrobiology, 2006, 6,
824–837.
15. Guillo, C., Karlinsey, J. M., Landers, J. P. Lab Chip, 2007, 7, 112–118.
16. Oh, K. W., Ahn, C. H. J. Micromech. Microeng., 2006, 16, R13–R39.
17. Duffy, D. C., McDonald, J. C., Schueller, O. J. A., Whitesides, G. M. Anal. Chem., 1998, 70, 4974–
4984.
18. McDonald, J. C., Duffy, D. C., Anderson, J. R., Chiu, D. T., Wu, H., Schueller, O. J., Whitesides, G. M.
Electrophoresis, 2000, 21, 27–40.
19. Garcia, C. D., Henry, C. S. Anal. Chem., 2003, 75, 4778–4783.
20. Weibel, D. B., Kruithof, M., Potenta, S., Sia, S. K., Lee, A., Whitesides, G. M. Anal. Chem., 2005, 77,
4726–4733.
21. Herein, “normally open valves” are defined as valves open to fluid flow in the nonactuated, resting state.
Conversely, “normally closed valves” are defined as valves closed to fluid flow in the nonactuated,
resting state. In both cases, the “resting state” refers to the state requiring no power to maintain.
22. Studer, V., Hang, G., Pandolfi, A., Ortiz, M., Anderson, W. F., Quake, S. R. J. Appl. Phys., 2004, 95,
393–398.
23. Balagadde, F. K., You, L., Hansen, C. L., Arnold, F. H., Quake, S. R. Science, 2005, 309, 137–140.
24. Hansen, C. L., Skordalakes, E., Berger, J. M., Quake, S. R. Proc. Natl. Acad. Sci. USA, 2002, 99,
16531–16536.
25. Lee, J. N., Park, C., Whitesides, G. M. Anal. Chem., 2003, 75, 6544–6554.
26. Grover, W. H., Mathies, R. A. Lab Chip, 2005, 5, 1033–1040.
27. Karlinsey, J. M., Monahan, J., Marchiarullo, D. J., Ferrance, J. P., Landers, J. P. Anal. Chem., 2005,
77, 3637–3643.
28. Blazej, R. G., Kumaresan, P., Mathies, R. A. Proc. Natl. Acad. Sci. USA, 2006, 103, 7240–7245.
29. Paegel, B. M., Grover, W. H., Skelley, A. M., Mathies, R. A., Joyce, G. F. Anal. Chem., 2006, 78,
7522–7527.
30. The Thomson Corporation. ISI Web of Science, 2007, Citation Database.
31. Manz, A., Harrison, D. J., Verpoorte, E. M. J., Fettinger, J. C., Paulus, A., Luedi, H., Widmer, H. M.
J. Chromatogr., 1992, 593, 253–258.
32. Garstecki, P., Fuerstman, M. J., Fischbach, M. A., Sia, S. K., Whitesides, G. M. Lab Chip, 2006, 6,
207–212.
33. Eijkel, J. C., van den Berg, A. Lab Chip, 2006, 6, 1405–1408.
34. Attiya, S., Jemere, A. B., Tang, T., Fitzpatrick, G., Seiler, K., Chiem, N., Harrison, D. J. Electrophoresis,
2001, 22, 318–327.
35. Lam, E. W., Cooksey, G. A., Finlayson, B. A., Folch, A. Appl. Phys. Lett., 2006, 89, 164105.
36. Fuerstman, M. J., Deschatelets, P., Kane, R., Schwartz, A., Kenis, P. J. A., Deutch, J. M., Whitesides,
G. M. Langmuir, 2003, 19, 4714–4722.
37. Du, W.-B., Fang, Q., He, Q.-H., Fang, Z.-L. Anal. Chem., 2005, 77, 1330–1337.
38. Fuerstman, M. J., Garstecki, P., Whitesides, G. M. Science, 2007, 315, 828–832.
39. Kim, D., Chesler, N. C., Beebe, D. J. Lab Chip, 2006, 6, 639–644.
40. Walker, G. M., Beebe, D. J. Lab Chip, 2002, 2, 131–134.
41. Kim, S. J., Lim, Y. T., Yang, H. S., Shin, Y. B., Kim, K., Lee, D. S., Park, S. H., Kim, Y. T. Anal.
Chem., 2005, 77, 6494–6499.
42. Ono, K., Fujii, T. Micro Total Analysis Systems, San Diego, CA, USA, 2005, p. 271.
43. Adams, M. L., Johnston, M. L., Scherer, A., Quake, S. R. J. Micromech. Microeng., 2005, 15, 1517.
44. Jeon, N. L., Chiu, D. T., Wargo, C. J., Wu, H., Choi, I. S., Anderson, J. R., Whitesides, G. M. Biomed.
Microdev., 2002, 4, 117–121.
45. Easley, C. J., Karlinsey, J. M., Leslie, D. C., Begley, M. R., Landers, J. P. Micro-Total-Analysis-
Systems, 10th International Conference on Miniaturized Systems for Chemistry and Life Sciences,
Tokyo, Japan, November 2006.
46. Easley, C. J. Ph.D. Dissertation, University of Virginia, Charlottesville, VA, USA, 2006.
47. Anderson, M. J., Hansen, C. L., Quake, S. R. Proc. Natl. Acad. Sci. USA, 2006, 103, 16746–16751.
41 Low-Cost Technologies for
Microfluidic Applications
Wendell Karlos Tomazelli Coltro and Emanuel Carrilho
CONTENTS
41.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1169

41.2 Fundamental Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1170
41.2.1 A Brief History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1170
41.2.2 The Working Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1171
41.2.3 Required Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1173
41.3 The Direct-Printing Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1173
41.3.1 PT Microchips Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1173
41.3.2 Microelectrode Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1175
41.3.3 Fabrication of Other Microanalytical Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1178
41.3.4 Fabrication of Glass-Toner Microchips and Toner-Mediated Lithography . . . . . . . 1179
41.3.5 Fabrication of PT Microchips with Integrated Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . 1180
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1183
41.1 INTRODUCTION
The amazing research field related to miniaturized systems has provided a true revolution in ana-
lytical and bioanalytical sciences in the past two decades. An exponential increase in the number
of publications has been observed since the concept of a micrototal analysis system (µTAS) was
first reported in the early 1990s. Short analysis time, reduced reagent consumption, low waste pro-
duction, and unprecedented device portability are key features of a µTAS. These, combined with
the integration of other analytical procedures, provide advantages that have stimulated the scientific
community to explore the value of executing analysis in the microworld.1−4 Since the first report
in literature,5,6 a large number of materials and microfabrication techniques have been theorized,
investigated, and developed in microfabrication sciences.1−4
Standard techniques based on photolithographic processes have been extensively used to fabri-
cate micromachined systems, and the details associated with these techniques can be found in Chapter
10 by Legendre. The initial development of µTAS was focused on glass or quartz substrates to create
microdevices. Both materials imparted the advantages associated with low electric conductivity and
high thermal conductivity, thus allowing the use of high voltages, providing good transparency for
optical detection and most importantly, well-developed surface chemistry.5,6 However, the conven-
tional microfabrication processes are laborious, time consuming, and present a relatively high cost
per device.
Furthermore, the standard processes can require some sophisticated instrumentation, particularly
if located in costly cleanroom environments, some of which are not always readily accessible to the
1169
researchers. For this reason, the quest for simple and low-cost procedures is a constant focus for all
the scientific community.
As a result, the introduction of new types of materials has offered new ways for the fast, low-
cost prototyping of microdevices of the disposable nature. In this light, the inexpensive nature of
certain polymeric materials presents unique opportunities for microchip fabrication, and there is a
significant interest in the development and use of these as substrates. One of the primary advantages
of most polymeric devices is that they can be fabricated outside of the clean room environment.2−4,7
The emerging use of polymers can be attributed largely to their enhanced biocompatibility, great
flexibility, reduced cost, and easy processing.7,8 Various polymeric materials have been used to
create microfluidic devices including poly(methyl methacrylate) (PMMA),8,9 poly(carbonate),10
photoresists (e.g., SU-8),11 Zeonor 102012 and, obviously, poly(dimethylsiloxane) (PDMS).13,14 A
variety of techniques have been utilized for the manufacture of polymeric microstructures including
laser ablation, X-ray lithography, injection molding, and imprinting from masters templates.2−4
Although these techniques are simpler, in some ways, than those used in glass and quartz µchip
fabrication, they still require sophisticated instrumentation for the micromachining and production
of the molding template.2−4
This chapter presents some detailed fundamentals of low-cost technologies that can be used to
produce microchannels, microelectrodes as well as integrated systems using the low-cost instru-
mentation commonly found chemical/biochemical laboratories. The main goal of this chapter is to
enable the reader with rapid prototype capabilities for lab-on-a-chip fabrication without a significant
investment in specific instrumentation, cleanroom facilities or time.
All fabrication techniques presented here are based on the use of toner from office laser printers,
which is used to print a layout on a transparency film or a wax paper sheet. The transparency
film is essentially used as a poly(ethylene terephthalate) (PET) substrate for polyester-toner (PT)
microchannels preparation. The wax paper is used as an initial surface onto which the pattern is
printed—the patterns are subsequently transferred either to a recordable compact disks (CD-R) or a
glass surface. In either case, this transference is carried out with heat and pressure using the type of
heat transfer machine that is used for T-shirt transfer machine or, in more sophisticated form, a tool
similar to that used for hot embossing.
41.2 FUNDAMENTAL ASPECTS

41.2.1 A BRIEF HISTORY
The potential for the use of toner in microfabrication technology has a short but promising history
that originated at the beginning of this century. Tan et al.,15 introduced a simple method for the
fabrication of PDMS microchips by using a photocopier to produce a high-relief master on polyester
film. This low-cost master was employed to produce channels in PDMS using a replica molding
process. This pioneering work, using toner-assisted microfabrication science, yielded channels in a
PDMS substrate in less than 1.5 h.
do Lago et al.16 proposed a direct-printing process for fast and direct production of microfluidic
devices at a very low cost. In this process, a laser printer is used to selectively deposit a toner layer
on a polyester film, which is subsequently laminated against a blank polyester film (single toner
layer, STL) or against a mirrored image of the layout (double toner layer, DTL). Compared to the
proposed technique by Tan et al.,15 the direct-printing process makes possible the fabrication of
microchips in a matter of minutes, not hours. Furthermore, tens of devices can be simultaneously
printed over a single transparency sheet. This simple and inexpensive technology has been used
to fabricate electrophoresis microchips coupled to end-channel amperometric17−19 and contactless
conductivity detection.16,20 Electrospray16 and mixing21 microdevices have also been proposed by
using do Lago’s process, as well as the production of printed masters for prototyping of PDMS
microchips.22
Low-Cost Technologies for Microfluidic Applications 1171
Daniel and Gutz23 have used this direct-printing process for quick production of both single
or multiple coplanar gold electrodes and microfluidic flow cells, as well as microfluidic cells with
interdigitated gold electrode array.24 In their work, the layouts were laser printed on a wax paper,
and thermally transferred onto the gold sputtered CDtrodes’ surface. This same technology was
used to fabricate disposable twin gold electrodes for amperometric detection in conventional cap-
illary electrophoresis25 (CE) and microchip electrophoresis.20 Lowinsohn et al.26 have recently
reported the fabrication and characterization of inexpensible gold-disk CDtrodes using toner masks,
while obtaining highly reproducible electrode area. The relative standard deviation (RSD) for
the voltammetric response of 10 different CDtrodes was below 1% for a 2 mmol L−1 solution of
K4 Fe(CN)6 .
Recently, do Lago et al.27 proposed a new process to fabricate glass-toner microdevices. In this
new approach, laser printers do not print the layout directly over the glass surface. Alternatively,
the layout is to be first printed on a wax paper and then transferred onto glass surface by heating
under pressure. In fact, following the printing step, each layout can be cut out and positioned
(like a decal) over a glass wafer for thermal transference. In a manner similar to the steps used
for preparing CDtrodes,17,23−26 heating under low pressure allows transfer of the toner from wax
paper to the glass surface. These glass-toner devices present a specific functional advantage in that
higher electroosmotic flow (EOF) is generated than in PT devices, primarily because of the greater
silanol content in the bottom and top surfaces of the channels. Using the same approach, free-flow
electrophoresis in glass-toner microchips have also been recently proposed by the same group.28
Liu et al.21 proposed the use of the direct-printing process to fabricate passive micromixers in
microfluidic devices. Micromixers were projected by applying a gray-scale coloring of the graphic
software. This gray-scale effect provides the presence of toner particles into channel that act as
obstacles to the flow stream for advection mixing, allowing these particles to be used as efficient
mixing elements. Vullev et al.29 used this concept of the direct-printing process to print positive-relief
masters on smooth substrates. In fact, although the authors have not referenced previously reported
works,16,17,22 in this “nonlithographic” process, the masters were printed over a polyester film just
as previously proposed for PT microdevices fabrication16,17 as well as for the masters production
for rapid prototyping of PDMS channels.22
41.2.2 THE WORKING PRINCIPLE

It is well known that UV photolithography is a standard protocol to transfer an image from a mask
to a planar surface of a substrate such as glass, silicon, or quartz (see Legendre Chapter 10 for
details). In this case, the layout drawn on a mask is exposed to UV-radiation and patterned on a
photosensitive layer (photoresist). After light patterning step, a wet chemical or dry etching step is
often employed to create channels in the substrate surface after removing the photoresist layer. In
this step, the resulting channel can have an isotropic or anisotropic profile, depending on both the
material and etch solution characteristics. The bonding of these etched channels is often performed by
a thermal bonding step, which requires elevated temperature. Even though this high-end technology
allows researchers to develop well-defined microstructures with desired aspect ratios, these steps are
laborious, time consuming, and require expensive equipment, thus driving the search for alternative
methods for fabrication with inherent simplicity and lower cost.
To the best of our knowledge, the direct-printing process proposed by do Lago et al.16 is the
simplest, easiest, fastest, and least expensive microfabrication process of all those described in the
literature. Basically, this process can be explained by the working theory of a laser printer. When
you design a layout by using graphic software, this layout is sent to a laser printer that selectively
deposits a toner layer over a surface (paper or transparency film). The graphic software should be
able to prepare the layouts on a 1:1 scale and for their printing without geometrical aberrations.
Toner is a complex powder deposited by a laser printer to form an image (in our case the channel
walls). Laser printers force toner powder to form the desired image on the polyester film and as
FIGURE 41.1 Typical EDX spectrum of (a) toner sample and (b) toner particle deposited by laser printer.
(Reprinted from Coltro, W.K.T. Fabrication and Evalution of Electrophoresis Microdevices with Electrochem-
istry Detection, Sao Carlos, 2004, 126 p. Dissertation (Master in Science). Institute of Chemistry at Sao Carlos,
University of Sao Paulo. Online available at: http://www.teses.usp.br/)
a final step, the toner image is melted onto the surface. In this process, the toner started off as a
powder, passed through a fluid, and ended up as a solid structure bonded to the transparency sheet
or to wax paper surface. At the chemical level, toner is constituted mainly of iron oxide (50–55%
w/w) and a polymeric resin (45–50% w/w). Other pigments or ingredients can be added according
to the manufacturer. Figure 41.1a shows a typical spectrum obtained by energy dispersive X-ray
(EDX) spectroscopy for a toner sample—the spectrum reveals the presence of iron and oxygen. The
gold that is present results from the metal deposition for a previous scanning electron microscopy
(SEM) analysis, presented in Figure 41.1b. In the EDX spectrum, it can also be seen that chlorine
is present, probably the result of residual FeCl3 . Interestingly, the polymeric resin is composed of
two components: PMMA (ca. 20%) and poly(styrene) (ca. 80%).16,17 The other material used in this
microchip preparation, the PET film, is coated with a thin layer of silica with a thickness that varies
depending on the manufacturer. X-ray photoelectronic spectroscopy (XPS) experiments showed the
presence of silicon (ca. 0.6% estimated as SiO2 ) over the polyester film.18 While the magnitude of
EOF on glass or PDMS chips basically correlates with the extent of deprotonation of the silanol
groups, this is not trivial on PT chips. Here, the parameters affecting EOF in this hybrid, highly
complex system are numerous and not well understood, since the bottom and the top of channel
are made up of silica doped PET and the channel walls are made up of toner. However, the low
abundance of silica in the channel explains the low EOF found in PT microchips, which is typically
10 times lower than glass and PDMS devices.17
As described previously,16,17 the limitation of this low-cost fabrication technology is related to
the laser-printer resolution. Theoretically, a laser printer with 600 dots per inch (dpi) resolution
could be used to print 50-µm width channels because such resolution (600 dpi) is about 42 µm.
However, toner layer is composed of smooth particles that are of irregular round or elliptical shape,
with dimensions around 6 × 8 µm (see Figure 41.1b). Experimentally, it is possible to produce
channels with 50-µm width, however, with limitations such in the channel definition and possible
appearance of edge regularity. Furthermore, the presence of some toner particles in narrow channels
at the mentioned dimension (50 µm) can result in partially or completely obstructed channels. The
RSD for channels with 50-µm width is around 25% with the high RSD inherently related to the
printing process. For channels having a width in the 150 µm to 1 mm range, this parameter is
decreased to 10% and 3%, respectively (n = 10).
While glass, silicon, or quartz channels are often thermally bonded at high temperatures (550–
650◦ C), the PT channels are bonded by a thermal lamination step in a standard office laminator.
The lamination should be carried out at temperature high enough to provide a desired seal between
the toner and the polyester (STL), or to other toner layer (DTL). For STL channels, the printed
channel is bonded against a blank transparency (i.e., a transparency cover without toner layer),
while for DTL channels, the printed layout should be firstly aligned with its mirror image, and
subsequently laminated together. During the lamination step, the toner layer binds by fusion on
the blank transparency or against another toner layer providing a stable sealing. After the bonding
process, a monolithic structure of polyester and toner results with channels and reservoirs produced
where no toner was deposited.
41.2.3 REQUIRED INSTRUMENTATION

The direct-printing process does not require any sophisticated instrumentation. Basically, two pieces
of equipment are required as fundamental tools: a computer equipped with graphic software (e.g.,
Corel Draw, AutoCad, Macromedia FreeHand, or Microsoft Power Point) and a laser printer
(although different brands provide different results as will be discussed later). Accessories such
as an office laminator, a heating press (T-shirt transfer machine), and a paper driller represent the
low-cost tools we have used. Finally, polyester films (transparency films or overhead projector films),
wax paper sheets, a commercially available toner cartridge, microscope slides (glass) and CDs are
the consumables, since they are used to produce microchips, microelectrodes, and integrated systems
at very low cost.
For the sake of comparison, the acquisition of this standard office equipment costs less than 3000
U.S. dollars. In contrast, the cost of a standard microfabrication laboratory including medium-size
cleanrooms, photo exposure devices, and spinners, for example, is close to one million U.S. dollars.
Even though a Class 100 laminar flow cleanhoods are substantially less expensive, the total amount
involved is still elevated. In addition to hardware costs, the comparative price of consumables for
the low-cost and standard techniques is staggering. While a toner cartridge, polyester films, CDs,
and wax paper can be obtained for a few hundred dollars, the purchase of photoresists, developer,
and remover solutions, as well as wafers (e.g., glass and silicon substrates) may amount to several
thousand dollars. Furthermore, the mask price used in the UV-exposure step can range from $20 to
$300, depending on the chosen technology.
41.3 THE DIRECT-PRINTING PROCESS

The direct-printing technology has been used for several applications including electrophoretic
microchips,16−20 CDtrodes23−26 for amperometric detection in conventional capillary25 and
microchip electrophoresis,17 microfluidic cells with interdigitated Au–CDtrodes,23,24 electrospray,16
and micromixers21 devices. Electrophoresis microchips with integrated electrodes for contactless
conductivity detection,20 glass-toner chips,27 microchip free-flow electrophoresis,28 as well as glass-
etched channels32 have recently been proposed using the direct-printing technique in one or more
stages during the device preparation.
41.3.1 PT MICROCHIPS FABRICATION

The PT microchip fabrication is depicted on Figure 41.2, where the production of both (A) STL
and (B) DTL channels are shown. For both types channel microfabrication, the first step is related
to the projection of the desired layout using any applicable graphic software such as Corel Draw,
AutoCAD, and Macromedia Freehand. The drawn layout is sent to the laser printer that prints the
image on a transparency film (I) carefully positioned on the printer tray. In this procedure, the white
color is used, during the drawing step, for the regions where microfluidic channels and reservoirs
(and consequently the channel width) are desired (II). On the other hand, the toner layer not only
defines the channel walls but, consequently, also the channel depth. The thickness of the toner
layer deposited on polyester surface depends of three main factors: (1) printing mode, (2) printer
a a
Channels Channels
II
III
IV
b b
(a) (b)
FIGURE 41.2 Step-by-step procedure of the direct-printing process for (a) STL and (b) DTL. (I) Polyester
film; (II) printed layout over the PET film; (III) printed layout with aligned cover [blank polyester in (a), or
printed mirror image in (b)] containing holes for accessing the microfluidic network; (IV) laminated STL and
DTL chips; and (V) microdevices with solution reservoirs. In this scheme, the toner layer and the reservoirs
glued on polyester film are depicted by A and B, respectively.
resolution, and (3) printer brand. Using a Hewlett Packard LaserJet printer and working at vectorial,
best quality printing mode, the thickness is around 7 ± 1 µm. Minor differences have been observed
in the channels defined by using laser printers with 600 or 1200 dpi resolution. The main observed
difference is that, for 1200 dpi, the toner layer is more compact than for 600 dpi, which will be
relevant only in toner-based lithography.32 Significant alterations will appear when using high-
resolution laser printers (≥3000 dpi) similar to photoplotters commonly employed to print masks
for photolithography.
For the simplest case shown at Figure 41.2, the STL channel, the printed channels are laminated
against a blank polyester film (containing no toner), where the STL channel is defined solely by
one toner layer. Consequently, the channel depth is close to toner layer thickness (i.e., 7 µm). The
DTL channels, require the use of two toner layers, which ultimately yields an improved aspect ratio.
However, since office laser printers were not projected to have to meet the needs of the analytical
microfabrication community, perfect alignment of both layers in repetitive printing steps is almost
impossible. The strategy found for do Lago et al.16 was to make these DTL channels by printing the
desired layout and its mirror image simultaneously on the same polyester film. Then, both structures
were aligned on top of each other for a DTL device.
Following the printing step, but before the lamination step, it is necessary to create access
holes to the microfluidic network on PT chips, these can simply be made by using an office paper
driller. This step is obviously much simpler (and faster) than preparing holes in glass channels with
specialized drilling tools (see Chapter 10 by Legendre). The bonding step of the PT channels is
provided by thermal lamination, but only after the printed channel and its cover (with holes) have
been carefully aligned (III). For DTL channels, special attention should be given to this step because
the alignment between the printed layout and its mirror image needs to be exact in order to avoid
channel obstruction during the lamination step. The use of alignment markers or guidelines on the
layout optimizes repeatability in chip-to-chip fabrication.
The lamination step, typically taking 40–50 s for each microdevice, however, the tempera-
ture at which this is carried out represents an important variable. The optimum temperature is
around the melting point of the toner (i.e., 120◦ C) if the working temperature is below this value,
repetitive lamination steps (two or three times) will be required in order to avoid fluid leakage.
Since the transparency film inherently is hydrophilic in nature, it is necessary to contain the buffer
solutions in reservoirs; otherwise, the fluid will spread out over the surface via capillarity. The
reservoirs can be prepared by gluing small cylinders (V), constructed from the base of 100 µL
pipette tips or poly vinyl chloride tubes onto the polyester surface using a bicomponent epoxy resin.
It is strongly recommended that the internal diameter (i.d.) of these reservoirs be larger than the
diameter of the access holes (access to the microfluidic network) in order to avoid blockage of
the channel entrance. The glue curing time for the preparation of reservoirs consumes the longest
portion of the fabrication time (ca. 10 min) in this low cost process. Likewise, a silicon or glass
wafer can be explored to micromachine several layouts simultaneously, a polyester sheet can also
be used to prepare many more devices in a same printing step. This advantage comes from the fact
that a letter or A4 size sheet has a total area four times larger than the area of a standard wafer
(25 in.2 ).
As described here, the direct-printing process is highly dependent on the laser-printer resolution.
Consequently, in this respect, the definition of the PT channels is worse when compared to glass
and PDMS chips produced by conventional photolithography. Figure 41.3 shows three micrographs
obtained by SEM depicting the surface of channels typically obtained in PT (Figure 41.3a), the edge
from channel to the bulk toner (Figure 41.3b), as well as a cross-sectional profile (Figure 41.3c)
of two channels. The presence of many toner particles inside the channel as well as a significant
roughness to toner wall is illustrated by Figure 41.3a and b. Figure 41.4 presents two examples of
electrophoresis microchips fabricated in PT using DTL channels format. For capacitively coupled
contactless conductivity detection (C4 D) measurements, the PT chips have been projected with a
double-T format injection channel,16,20 while for amperometric detection, a simple cross-channel
has provided satisfactory results.17−19
41.3.2 MICROELECTRODE PREPARATION

The toner material and the direct-printing approach may not only be used for PT microchannel
fabrication but also to produce a great variety of features, devices, or elements for analyti-
cal microsystems. For example, other essential elements for electrophoresis microchips are the
(a) (b) (c)
FIGURE 41.3 SEMs for an intersection of injection and separation 150-µm-wide channels (a), toner surface
at channel wall (b), and a transversal cut of a DTL channel (c). (Images (a) and (b) were reprinted from
Coltro, W.K.T. Fabrication and Evalution of Electrophoresis Microdevices with Electrochemistry Detection,
Sao Carlos, 2004, 126 p. Dissertation (Master in Science). Institute of Chemistry at Sao Carlos, University of
Sao Paulo. Online available at: http://www.teses.usp.br/) and image (c) with permission from He, F.-Y., et al.,
Anal. Bioanal. Chem., 382, 192, 2005 (c). With permission.)
b b
sw (B)
(A)
FIGURE 41.4 Examples of electrophoresis microchip fabricated in PT. (A) Microdevice with cupper tape
electrodes (a) positioned externally over the channel for C4 D measurements. Point b is a representation for the
solution reservoirs. (Reprinted from do Lago, C.L., Silva, H.D.T., Neves, C.A., Brito-Neto, J.G.A. and Fracassi
da Silva, J.A., Anal. Chem., 75, 3853, 2003. With permission.) (B) Microdevice layout for end-channel
amperometric detection. S, SW, and b represent, sample, sample waste, and buffer reservoirs, respectively.
(Reprinted from do Lago, C.L., et al., Anal. Chem., 75, 3853, 2003. and Coltro, W.K.T., et al., Electrophoresis,
electrodes. The preparation of microelectrodes can also benefit from this direct-printing process,
that is, electrodes for both electrokinetic control and electrochemical (EC) detection. Simulta-
neously to release of this technology, Daniel and Gutz23 described an innovative method to
manufacture gold microelectrodes using commercial recordable CD as metal source.31 In this
method, depicted in Figure 41.5, the direct-printing process is used to print the electrode design
over a wax paper (a), of the same type used as a support for commercially available adhe-
sive labels. Then, this printed image is thermally transferred onto the CD gold layer (b), from
which the protective film was previously removed with HNO3 .31 Commonly, the thermal trans-
ference is carried out in a thermal press at 120◦ C for 90 s under a pressure of approximately
0.1–0.5 MPa. Following the toner transference, the wax paper is easily removed from CD
surface, similarly to a decal step. The toner-free gold areas are etched away (c) by a short
exposure to an iodide/iodine solution. After this etching step, the toner is selectively removed
with the help of a cotton swab soaked with acetonitrile, exposing the underlying gold layer (d)
and, thus, producing the electrodes.23 An amplified image from a dual electrode is presented in
Figure 41.5e.
Richter et al.25 have used this process to produce electrodes (see Figure 41.5) for EC detection
in CE. These Au–CDtrodes were used first in a home-made CE system, in which an electrophoretic
separation of iodide, ascorbic acid, dipyrone, and acetaminophen was successfully performed.
Au–CDtrodes were later applied to an electrophoresis microchip fabricated with PT,17 where the
effectiveness of the proposed system was demonstrated with a separation of iodide and ascor-
bic acid. A series of 10 repetitive injections obtained in a conventional CE–EC system and one
electropherogram obtained in miniaturized system are shown in Figure 41.6.
The use of these Au–CDtrodes in both conventional capillary and microchip electrophoresis
presents some advantages. The first is that the low cost of electrode fabrication, considering that ca.
50 electrodes can be prepared from a single CD, aids in the quest of having cost-effective, disposable,
single-use devices. The second is that, for both systems, the electrode replacement can be made
quickly requiring only a new adjustment between the electrode and the capillary/channel extremity.
A third advantage involves the use of the dual electrodes (see Figure 41.5) to simultaneously detect
Au electrical
contacts
(a) (b) (c) (d) (e)
100 µm
Au PC Au-CDtrodes
FIGURE 41.5 Au–CDtrode microfabrication process. (a) Drawing and printing of electrode layout on a
wax paper; (b) thermal transference of the desired layout over a Au–CD surface; (c) Au etching step with
iodide/iodine solution; (d) dual microelectrodes obtained after toner removal; and (e) amplified image from a
dual Au–CDtrode with ∼100-µm wide and gap ∼100 µm. Black and gray colors indicate the apparent toner
and the Au surface, respectively. (Adapted and reprinted from Coltro, W.K.T., et al., Electrophoresis, 25, 3832,
2 4 80
(a) (b)
60 3 a
1
60
Current / nA
40
i/nA
40
b
20 20
0 0
1.0 1.5 2.0 2.5 3.0 0 20 40 60 80 100

Time (min) Time (s)
FIGURE 41.6 Electropherograms obtained for conventional capillary (A) and microchip electrophoresis (B)
coupled with end-column and end-channel amperometric detection, respectively. In (A) the sample is constituted
of iodide (1), ascorbic acid (2), dipyrone (3), and acetophen (4). In (B) the sample is composed of iodide (a) and
ascorbic acid (b). (Reprinted from Richter, E.M., et al., Electrophoresis, 25, 2965, 2004 and Coltro, W.K.T.,
et al., Electrophoresis, 25, 3832, 2004. With permission.)
both the oxidation and the reduction reaction at the electrodes. Since amperometry is a selective
method, the dual detection can provide additional information. In addition to these advantages, the
production of electrodes can be improved by using a high-resolution laser printer. With printer of
higher resolution, narrower electrodes can be prepared, thus decreasing the peak broadening related
to diffusion at electrode surface.
(a) (b)
FIGURE 41.7 Examples of electrospray devices fabricated in PT. (a) A drop of KCl solution at the channel
outlet and (b) electrospray tip with Taylor’s cone generated by electrostatic field. (Reprinted from do Lago,
C.L., et al., Anal. Chem., 75, 3853, 2003. With permission.)
41.3.3 FABRICATION OF OTHER MICROANALYTICAL DEVICES

The major contribution of this direct-printing process is that it is not restricted to the examples
described above. One possibility that has not completely been explored yet is the use of micro-
electrospray devices fabricated in PT. The generation of Taylor’s cone from a microchannel outlet
on a PT chips was shown by do Lago et al.16 Owing to the hydrophilic nature the substrate, the
edge of the device was dipped in a silicone solution to impart enough hydrophobic character to the
external surface to prevent the aqueous solution from spreading. The electrospray phenomenon on
the PT device was observed16 for a solution drop hanging at the wedge-cut tip (Figure 41.7), but the
coupling with mass spectrometry has not yet been shown and is currently under investigation in our
laboratory.
The concept of µTAS suggests that several analytical steps should be integrated on the same chip.
Different analytical operations have already been integrated using standard techniques and popular
substrates.33−37 For PT chips, this task is not trivial and requires creativity. As shown in Figure 41.3a,
the presence of toner particles in the channel is a limitation associated with the direct-printing process.
And while this may be viewed as disadvantageous, these toner particles can be exploited as obstacles
to create convective mixing, or to adsorb analytes/reagents. In other words, toner particles could act
as filter, preconcentrator, microreactor (the enhanced surface area is particularly enable with these
latter two points), or obstacles that induce mixing. It is well established that using photolithographic
techniques it is possible to create obstacles that have an ordered configuration inside the channel.
In the direct-printing process, toner particle clusters can be generated by using a gray-scale effect
from graphic software such as Corel Draw. This gray-scale effect allows the deposition of toner
particles inside the channel, potentially creating obstacles for flow or turbulent mixing—although it
remains to be seen whether this can be accomplished in an accurate way. Theoretically, it is possible
to apply gray-scale effects ranging from 0% to 100%, with the lowest percentage providing no
toner and 100% will generate a channel fully loaded with toner particles. Exemplary micrographs
for the gray-scale effect applied to PT microchannels are depicted in Figure 41.8. Three different
stages of this effect (0%, 50%, and 90%) are identified in a 150-µm wide channel in Figure 41.8a,
where it is possible to observe that the channel is almost totally obstructed at 90% gray. In previous
experiments,30,38 it has been estimated that beyond 80% of gray-scale effect, the channel is partially
obstructed. Figure 41.8b shows this effect applied to the 500-µm wide channel where Liu et al.21
were able to show that this effect could be applied to create passive micromixers in these low-cost
devices.
(a) (b)
FIGURE 41.8 SEM images showing a gray-scale effect applied to PT channels: (a) 150-µm wide chan-
nel with a gradient of gray-tone applied to different points over the channel and (b) 500-µm wide channel
with 80% of gray-tone applied to the entire channel. (Reprinted from Coltro, W.K.T. Fabrication and Evalu-
tion of Electrophoresis Microdevices with Electrochemistry Detection, Sao Carlos, 2004, 126 p. Dissertation
(Master in Science). Institute of Chemistry at Sao Carlos, University of Sao Paulo. Online available at:
http://www.teses.usp.br/)
(a) (b) E1 E2 E3 E4
E1 E3
3
1 4 0.5 mA
2
E2 E4
0 10 20 30 40 50 60
Time (min)
FIGURE 41.9 Example of electrodes produced by direct-printing process. (a) A schematic view of the
microfluidic device with four interdigitated Au–CDtrodes. (1) and (4) inlet and outlet reservoirs, respectively;
(2) and (3) reservoirs for reference electrodes; E1−4 , gold working electrodes. (b) amperometric response at
all four electrodes for a K4 Fe(CN)6 sample introduced into microchannel by FIA. (Reprinted from Daniel, D.
and Gutz, I.G.R., Talanta, 68, 429, 2005. With permission.)
Daniel and Gutz24 have also proposed the fabrication of microfluidic flow injection cells with
interdigitated array of gold electrodes. Figure 41.9 presents a schematic diagram of this microdevice
(a) as well as a sequence of amperometric response to a solution of K4 Fe(CN)6 exposed to all four
interdigitated gold electrodes in microfluidic flow cell (1.4 cm length × 0.1 cm width × 19 µm
depth) (b). The microfluidic channel was prepared by thermal transference of three toner layers over
the polycarbonate CD base, with the bonding provided by a second polycarbonate slice. The fluid
transport was carried out with a peristaltic pump for flow injection analysis (FIA) applications.
41.3.4 FABRICATION OF GLASS-TONER MICROCHIPS AND TONER-MEDIATED

LITHOGRAPHY
Polyester-toner devices present advantages such as low cost per device, as well as easy and speed
of fabrication. In addition, dozens of devices can be prepared in parallel using a single transparency
film.16,17,30 As mentioned earlier, the magnitude of EOF on PT devices is 10 times lower than that
in common microfabrication substrates such as glass, quartz, and PDMS.17,30 While the EOF in
these popular materials can be raised by chemical or physical modification of the surface (activation
by oxygen plasma or sodium hydroxide), similar treatment of PT chips does not yield the same
enhancement. The use of sodium hydroxide is not recommended because it attacks the toner surface
rapidly (in a few minutes), even at low concentration. However, the low EOF can certainly be an
advantage for the analysis of analytes with similar electrophoretic mobilities. One strategy to obtain
a simple microdevice but with greater EOF is the fabrication of glass-toner microchips.27
The toner layer cannot be printed over a glass surface since this substrate is not an acceptable
substrate for commercial laser printers. Alternatively, the layout can be printed on a wax paper
first, and then transferred to the glass surface using a thermal transfer press methodology already
described. In this case, one or more toner layers can be sequentially transferred onto glass surface
increasing the aspect ratio. The sealing of these channels is also provided by using thermal press, in
which a glass cover binds to upper toner layer under heating and pressure. Recently, this technology
has been used to fabricate microchips for free-flow electrophoresis.28
The problems related to toner porosity and weaker adherence of the toner to the glass surface,
compared to PT devices, reveals a drawback of these glass-toner devices. In recent studies,32 we
improved the adherence of the toner to the glass surface allowing the exploration of the toner layers
as layout masks for preparing glass channels by wet chemical etching (toner-based lithography).
In this case, glass channels are obtained without the use of any photolithographic step during the
process and the costs related to it. The toner possesses limited chemical resistance to hydrofluoric
acid solution, which can be overcome by multilayer stacking. However, the resistance is substantial
enough to create channels up to 40-µm deep in less than 10 min. Figure 41.10 presents two SEM
images, where it was possible to observe a cross-intersection (Figure 41.10a) and a transversal
section of the bonded channel (Figure 41.10b) but, as expected, the channel did not present a perfect
definition due to the limitation of the direct-printing process.
41.3.5 FABRICATION OF PT MICROCHIPS WITH INTEGRATED ELECTRODES

One of the most exciting applications using the direct-printing process is related to the fabrication of
PT microchips with integrated electrodes for C4 D.20 The C4 D system is an universal detection mode
that has several advantages when coupled with conventional CE or microchip electrophoresis: (i) the
electronic circuitry is simple, inexpensive, and decoupled from the high-voltage used for separation;
(ii) the formation of bubbles at the electrodes is avoided; and (iii) electrochemical modification
(a) (b)
100 µm 30 µm
FIGURE 41.10 SEM images showing the glass channel obtained by toner-mediated lithographic process.
(a) Intersection of injection and separation 200-µm-wide microchannels and (b) sectional profile of a bonded
glass–glass microchannel. (Reprinted from Coltro, W.K.T., Piccin, E., Fracassi da Silva, J.A., do Lago, C.L.
and Carrilho, E., Lab Chip, 7, 931, 2007. Reproduced by permission of The Royal Society of Chemistry.)
or degradation of the electrode surface is prevented, thereby, allowing a wide variety of electrode
materials and background electrolytes.39,40
The detection electrodes for C4 D are often prepared by manually gluing copper stripes over the
separation channel. This manual electrode positioning limits the repeatability on the chip-to-chip
preparation, in which the accurate width, gap, and position over the separation channel are essential
parameters. To minimize this problem, the integration of microfabricated electrodes on polyester
films was a successful alternative. The electrodes for C4 D were prepared using toner masks for metal
deposition via sputtering. Briefly, the electrode geometry was drawn and printed over a polyester
film, followed by the sputtering of a thin layer, of titanium/aluminum over the printed electrode
mask. The resulting electrodes were obtained by soaking the substrate in acetonitrile and dissolving
the underlying toner layer. As the toner layer was removed with the solvent, the sputtered Ti/Al
could be lifted off, and the electrode material remained anchored to polyester substrate. Note that
aluminum was chosen due to its lower cost when compared to gold and platinum. The integration
of these electrodes with PT channels was also carried by thermal lamination while maintaining the
electrodes carefully positioned upwards, that is, isolated from contact with the solution in the channel
by the thickness of the polyester film (100 ± 10 µm). The final C4 D-PT chip carries the detection
electrodes on the outside of the top polyester film and the microfluidic network is directly printed
on the bottom polyester film.
Figure 41.11a presents one layout of this integrated system as well as a typical electropherogram
obtained on this device (Figure 41.11b). In addition to enhanced chip-to-chip repeatability, the use
of integrated electrodes improved the signal-to-noise ratio. The use of toner masks for preparation
of electrodes can be also extended to other EC modes such as amperometry, voltammetry, and
potentiometry. Furthermore, the use of copper tapes as electrodes is limited to the classical format
(antiparallel design), while the use of toner masks allows one to investigate different electrode design
for C4 D.

As seen in this brief chapter, the direct-printing process, based on laser printing of layouts on
polyester films or wax paper, has the potential to become a powerful technology for the rapid
prototyping of microfluidic devices at very low cost, and even a source of low-cost production of
disposable devices. This is supported by the fact that the required instrumentation is commonly
found at offices and chemistry laboratories. Besides the typical injection and separation channels for
electrophoresis, this technology has shown that mixing, preconcentration, clean-up, reactor devices,
(a) (b)
1 0.2
K+
Function l 2
n ne
C4D Signal (V)
generator ha Na+
nc
tio 3
ec l
Inj ne 0.1 Li+
h an
nc
tio
p ara
Se
4 0.0
Detection
circuitry 0 30 60 90 120
Time (s)
FIGURE 41.11 Representation of a (a) PT microdevice with integrated electrodes for C4 D and (b) a typical
electropherogram obtained for a mixture of inorganic cations (25 µM each) in this device. (Reprinted from
Coltro, W.K.T., Fracassi da Silva, J.A. and Carrilho, E., unpublished material.)
1182
FIGURE 41.12 Examples of devices, elements, and features that can be achieved using the toner-based technology.
as well as electrodes can also be incorporated into PT channel design using graphic resources and
toner masks. Currently, our main goal is focused in the development of a working µTAS by exploring
just the direct-printing process. Furthermore, the versatility of the direct-printing process has been
demonstrated with microelectrodes for amperometric detection in conventional CE and microchip
electrophoresis, microfluidic cells with interdigitated electrodes, electrospray, micromixers, and
free-flow electrophoresis devices, as well as glass-toner and glass-etched microdevices that can be
fabricated and explored for specific applications.
Overall, as presented here in some detail, this technology presents the opportunity to all scientists
in any laboratory interested in lab-on-a-chip technology, requiring only common tools, easily found
in offices/general laboratories. To conclude this chapter, we present in Figure 41.12 a portfolio of
many devices prepared using toner technology.
ACKNOWLEDGMENTS
We would like to thank the scholarship granted from Fundação de Amparo à Pesquisa do Estado de
São Paulo (FAFESP—grant 04/01525-0) to W.K.T. Coltro and research fellowship granted from
Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) (grant 304100/2005-
6) to E. Carrilho. CNPq is also acknowledge for financial support (grants 477982/2003-5 and
478467/2006-0).
REFERENCES
1. Reyes, D.R., Iossifidis, D., Auroux, P.A. and Manz, A., Anal.Chem., 74, 2623, 2002.
2. Auroux, P.A., Iossifidis, D., Reyes, D.R. and Manz, A., Anal.Chem., 74, 2627, 2002.
3. Vilkner, T., Janasek, D. and Manz, A., Anal.Chem., 76, 3373, 2004.
4. Dittrich, P.S., Tachikawa, K. and Manz, A., Anal. Chem., 78, 3887, 2006.
5. Manz, A., Miyahara, Y., Miura, J., Watanabe, Y., Miyagi, H. and Sato, K., Sens. Actuators B, 1, 249,
1990.
6. Manz, A., Graber, N. and Widmer, H.M., Sens. Actuators B, 1, 244, 1990.
7. Becker, H. and Locascio, L.E., Talanta, 56, 267, 2002.
8. Kelly, R.T. and Woolley, A.T., Anal. Chem., 75, 1941, 2003.
9. Galloway, M., Stryjewski, W., Henry, A., Ford, S.M., Llopis, S., McCarley, R.L. and Soper, S.A.,
Anal. Chem., 74, 2407, 2002.
10. Liu, Y.J., Ganser, D., Schneider, A., Liu, R., Grodzinski, P. and Kroutchinina, N., Anal. Chem., 73,
4196, 2001.
11. Jackman, R.J., Floyd, T.M., Ghodssi, R., Schmidt, M.A. and Jensen, K.F., J. Micromech. Microeng.,
11, 263, 2001.
12. Kameoka, J., Craighead, H.G., Zhang, H.W. and Henion, J., Anal. Chem., 73, 1935, 2001.
13. Duffy, D.C., McDonald, J.C., Schuller, O.J.A. and Whitesides, G.M., Anal. Chem., 70, 4974, 1998.
14. McDonald, J.C. and Whitesides, G.M., Acc. Chem. Res., 35, 491, 2002.
15. Tan, A., Rodgers, K., Murrihy, J.P., O’Mathuna, C., Glennon, D., Lab Chip, 1, 7, 2001.
16. do Lago, C.L., Silva, H.D.T., Neves, C.A., Brito-Neto, J.G.A. and Fracassi da Silva, J.A., Anal. Chem.,
75, 3853, 2003.
17. Coltro, W.K.T., Fracassi da Silva, J.A., Silva, H.D.T., Richter, E.M., Furlan, R., Angnes, L., do Lago,
C.L., Mazo, L.H. and Carrilho, E., Electrophoresis, 25, 3832, 2004.
18. He, F.-Y., Liu, A.-L., Yuan, J.-H., Coltro, W.K.T., Carrilho, E. and Xia, X.-H., Anal. Bioanal. Chem.,
382, 192, 2005.
19. Liu, A.-L., He, F.-Y., Lu, Y. and Xia, X.-H., Talanta, 68, 1303, 2006.
20. Coltro, W.K.T., Fracassi da Silva, J.A. and Carrilho, E., unpublished results.
21. Liu, A.-L., He, F.-Y., Wang, K., Zhou, T., Lu, Y. and Xia, X.-H., Lab Chip, 5, 974, 2005.
22. Bao, N., Zhang, Q., Xu, J.-J. and Chen, H.-Y., J. Chromatogr. A, 1089, 270, 2005.
23. Daniel, D. and Gutz, I.G.R., Electrochem. Commun., 5, 782, 2003.
24. Daniel, D. and Gutz, I.G.R., Talanta, 68, 429, 2005.

25. Richter, E.M., Fracassi da Silva, J.A., Gutz, I.G.R., do Lago, C.L. and Angnes, L., Electrophoresis,
25, 2965, 2004.
26. Lowinsohn, D., Richter, E.M., Angnes, L. and Bertotti, M., Electroanalysis, 18, 89, 2006.
27. do Lago, C.L., Neves, C.A., de Jesus, D.P., Silva, H.D.T., Brito-Neto, J.G.A. and Fracassi da Silva,
J.A., Electrophoresis, 25, 3825, 2004.
28. de Jesus, D.P., Blanes, L. and do Lago, C.L., Electrophoresis, 27, 4935, 2006.
29. Vullev, V.I., Wan, J., Heinrich, V., Landsman, P., Bower, P.E., Xia, B., Millare, B. and Jones, G.,
J. Am. Chem. Soc., 128, 16062, 2006.
30. Coltro, W.K.T. Fabrication and Evalution of Electrophoresis Microdevices with Electrochemistry
Detection, Sao Carlos, 2004, 126 p. Dissertation (Master in Science). Institute of Chemistry at Sao
Carlos, University of Sao Paulo. Online available at: http://www.teses.usp.br/
31. Angnes, L., Richter, E.M., Augelli, M.A. and Kume, G.M., Anal. Chem., 72, 5503, 2000.
32. Coltro, W.K.T., Piccin, E., Fracassi da Silva, J.A., do Lago, C.L. and Carrilho, E., Lab Chip, 7, 931,
2007.
33. Easley, C.J., Karlinsey, J.M., Bienvenue, J.M., Legendre, L.A., Roper, M.G., Feldman, S.H., Hughes,
M.A., et al., Proc. Natl. Acad. Sci. USA, 103, 19272, 2006.
34. Skelley, A.M., Scherer, J.R., Aubrey, A.D., Grover, W.H., Ivester, R.H.C., Ehrenfreund, P.,
Grunthaner, F.J., Bada, J.L. and Mathies, R.A., Proc. Natl. Acad. Sci. USA, 102, 1041, 2005.
35. Burns, M.A., Johnson, B.N., Brahmasandra, S.N., Handique, K., Webster, J.R., Krishnan, M.,
Sammarco, T.S., et al., Science, 282, 484, 1998.
36. Liu, R.H., Yang, J., Lenigk, R., Bonanno, J. and Grodzinski, P., Anal. Chem., 76, 1824, 2004.
37. Unger, M.A., Chou, H.-P., Thorsen, T., Scherer, A. and Quake, S.R., Science, 288, 113, 2000.
38. Coltro, W.K.T. and Carrilho, E., Evaluation of Printed Patterns in Polyester Films as a Precon-
centration Microdevice for Separation Microchips, presented at the 9th Latin American Sympo-
sium on Biotechnology, Biomedical, Biopharmaceutical and Industrial Applications of Capillary
Electrophoresis and Microchip Technology, Mexico City, 2003.
39. Brito-Neto, J.G.A., Fracassi da Silva, J.A., Blanes, L. and do Lago, C.L., Electroanalysis, 17, 1198,
2005.
40. Brito-Neto, J.G.A., Fracassi da Silva, J.A., Blanes, L. and do Lago, C.L., Electroanalysis, 17, 1207,
2005.
42 Microfluidic Reactors for
Small Molecule and
Nanomaterial Synthesis
Andrew J. deMello, Christopher J. Cullen, Robin Fortt,
and Robert C.R. Wootton
CONTENTS
42.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1185

42.2 Theoretical Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1186
42.2.1 Fluid Flow on the Microscale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1186
42.2.2 Mixing and Chemical Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1188
42.3 Microfabricated Reaction System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1189
42.3.1 Fabrication of Microfluidic Reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1189
42.3.2 Fluid Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1189
42.3.3 Mixing Modalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1190
42.4 Application of Microfluidic Reactors in Synthetic Chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1191
42.4.1 Key Benefits of Microfluidic Systems in Synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1191
42.4.2 Fields of Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1192
42.4.2.1 High-Throughput Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1192
42.4.2.2 Multiscale Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1194
42.4.2.3 High-Selectivity Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1194
42.4.2.4 Catalytic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1194
42.5 Real World Applications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1195
42.5.1 Point of Use Manufacture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1195
42.5.2 Measurement of Chemical Phase Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1196
42.5.3 Microfluidic Systems for Nanomaterial Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1196
42.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1201
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1201
42.1 INTRODUCTION
In 1828, when attempting to prepare ammonium cyanate from silver cyanide and ammonium chloride,
Friedrich Wöhler accidentally synthesized urea [1]. Contemporary wisdom at the time held that
organic compounds could only be created from a “vital force,” which existed within living organisms.
In a letter to Jöns Jakob Berzelius, Wöhler’s excitement was apparent; “I can no longer, so to speak,
hold my chemical water and must tell you that I can make urea without needing a kidney, whether
of man or dog; the ammonium salt of cyanic acid is urea.” Although Wöhler’s discovery was not the
first synthesis of an organic compound it sparked huge interest in making organic compounds from
nonliving substances and marked the beginning of organic chemistry as an academic and industrial
discipline.
1185
Since that time, organic chemistry (the study of the structure, properties, composition, reactions,
and synthesis of carbon-containing compounds) has flourished and is of vital importance to the phar-
maceutical, chemical, cosmetic, petrochemical, and textile industries. For the vast majority of the
time, since (and for centuries before) Wöhler’s synthesis of urea, the chemist’s toolkit has predomi-
nantly consisted of macroscopic components fabricated from glass. Examples of such components
include round-bottomed flasks, test tubes, distillation columns, reflux condensers, Erlenmeyer flasks,
drying tubes, separation funnels, recrystallization tubes, and burettes. Despite the enormous advances
that have been made in experimental, mechanistic, and theoretical organic chemistry over the past
150 years it is noteworthy that the basic experimental techniques and associated laboratory equip-
ment remain largely unaffected. Perhaps this is unsurprising, since standard laboratory glassware
provides a fitting environment in which to perform the vast majority of synthetic transformations.
Glass as a generic material is robust, possessing good chemical inertness. It also exhibits high
thermal conductivity, low electrical conductivity, and good transparency within the visible region
of the electromagnetic spectrum. This means that standard glassware can be used to process most
chemical reagents under a wide range of temperatures and pressure. A more pragmatic reason for
the adoption of macroscale glassware in organic synthesis relates to the fact that chemists as indi-
viduals feel comfortable performing reactions in environments, which they can easily manipulate,
control, and observe. Nevertheless, although a chemist may prefer to perform a reaction within a
round bottom flask, at the molecular scale it makes little difference if a reaction is processed within
a volume of 100 mL or 100 pL. What is more important is that the “ideal” chemical reactor should
provide an environment in which chemical state functions are precisely controllable, allowing the
rapid synthesis of a desired product in high yield. To this end, over the past decade, the application
of micromachining techniques cultivated within the semiconductor and microelectronics industries
have allowed the creation of a new instrumental platform able to efficiently process and analyze
molecular reactions on the micron to nanometer scale. This chapter aims to provide an introduction
to the concept of reaction miniaturization. The theory of miniaturization is discussed and followed
by a practical assessment of the use of microfluidic systems in synthetic chemistry.
42.2 THEORETICAL BACKGROUND

42.2.1 FLUID FLOW ON THE MICROSCALE
The foremost reason why microfluidic systems provide attractive environments in which to perform
synthetic chemistry lies in the dependency of fluid flow on scale. Put simply, fluid handling within
microfluidic environments differs markedly from typical macroscale flow [2]. A variety of phenom-
ena manifest themselves upon moving from the macroscale to the microscale, which in turn have a
profound effect on both the efficiency and manner in which chemical reactions proceed.
A chemical reaction is initiated by bringing the required reactants into intimate contact. Efficient
execution of this process often defines the efficiency of the reaction and, therefore, requires consid-
ered examination when designing a reaction system. The physical transport of a component along a
concentration gradient by molecular diffusion and turbulent convection is known as mass transfer.
The transport of mass through an interface between two media or phases of the same medium is
extremely important, since chemical reactions are normally coupled to the mass transfer efficiency.
The mass transfer coefficient is used to quantify the efficiency of mixing in macrofluidic systems
and is given by
NA
Kc = , (42.1)
Cs − C B
where Kc is the mass transfer coefficient (m s−1 ), NA the molar flux (mol m−2 s−1 ), Cs the con-
centration at the phase boundary (kg m−3 ), and CB the concentration in bulk solution (kg m−3 ).
Microfluidic Reactors for Small Molecule and Nanomaterial Synthesis 1187
A conspicuous effect of reactor miniaturization is that fluid properties become increasingly controlled
by viscous forces rather than inertial forces as reaction volumes are reduced [3]. In macroscale envi-
ronments, reagents are brought into intimate contact by turbulent flow. In such a regime, fluid
elements (of varying sizes) travel randomly and create a mixture that is highly segregated with a
finely dispersed structure. These eddies are a product of areas of circulatory motion in the fluid cre-
ated by the inertial forces of flowing liquids. The mixture that is produced, although finely dispersed,
is still heterogeneous at the molecular level and mixing of these fine fluid elements is subsequently
achieved through the random, diffusive motion of the component molecules. The degree of turbu-
lence in a system depends on its scale and the density and viscosity of the fluid. For example, in
large systems such as fluidic jet streams, where turbulence is high, fluid elements can move large
distances and therefore mixing is much faster than mixing by molecular diffusion alone.
Fluid flow in the absence of turbulence is described as being laminar. In such a situation, viscous
forces dominate inertial forces and eddies are no longer present. Laminar fluid flow can be con-
ceptualized as consisting of concurrently flowing parallel streams with little or no mixing between
streams. Since the same viscous forces that restrain the fluid streams in laminae also dampen out
irregularities and streamlines, laminar flow is not disturbed by obstacles and instead maintains a
smooth path. With the removal of the highly segregated structure created by eddies, mixing must be
achieved by molecular diffusion alone.
The Reynolds’ number is a dimensionless number used in fluid dynamics to determine dynamic
similitude [4] and is generically defined as
vδ
Re = , (42.2)
U
where v defines the mean fluid velocity, δ is a characteristic dimension of the reactor and U defines
the kinematic viscosity of the fluid. Significantly, the Reynolds’ number provides a measure of the
ratio of inertial forces to viscous forces and is most commonly used to assess whether flow will be
laminar or turbulent. In simple terms, viscosity describes the resistance of a fluid to deformation
under shear stress and encourages fluid to flow in parallel layers, with no disruption between the
layers (laminar flow), while inertia describes the tendency of the fluid to resist acceleration and thus
counteract laminar flow. For microfluidic systems, such as capillaries or blood vessels in the body or
the microchannel networks common to microfluidic systems, Reynolds’ numbers are typically much
less than 102 . This represents a situation where the flow regime may be considered essentially laminar
and contrasts with macroscale pipes or reactors (exhibiting Reynolds’ numbers in excess of 104 ),
where flow regimes are almost always turbulent. This behavior has a direct consequence on mixing
within microfluidic systems. Before a reaction between two reagents can occur, intimate contact
between the component molecules must be realized through mixing. Mixing within microfluidic
systems in its simplest manifestation is achieved by bringing together pure fluid component streams
within channels having characteristic cross-sectional dimensions measured in tens of microns. Since
low Reynolds’ and Peclet numbers (which measure the relative importance of mass transport due
to convection with respect to diffusional mass transport) are typical of such environments, mixing
can only be accomplished through diffusion, rather than the fast convective processes that dominate
in turbulent systems. Thus under these conditions, two or more distinct fluid streams moving in the
same channel do not develop turbulence at the interface between them, or at the interface with the
channel walls and the only mechanism of mixing of their components is diffusion across the former
interface.
Diffusive mixing efficiencies can be assessed using the Fourier number, which is defined by
Dt
Fo = , (42.3)
l2
where D is the molecular diffusion coefficient, t is the contact time, and l is a characteristic length
over which diffusion occurs. Broadly speaking, adequate mixing of fluidic streams occurs for Fourier
numbers of 0.1, with complete mixing defined by Fourier numbers greater than 1 [5]. Equation 42.3
illustrates that diffusive-mediated mixing timescales increase with the characteristic dimension of
the reactor. Consequently, although mixing via diffusion is highly inefficient for reactors with char-
acteristic dimensions greater than 1 mm, when diffusion distances are reduced below 100 µm mixing
times can become small. For example, a small molecule such as fluorescein in water exhibits a mixing
time of approximately 90 ms across a diffusion distance of 10 µm.
42.2.2 MIXING AND CHEMICAL REACTIONS

The effect of mixing on the extent of a chemical reaction and the resulting product distribution is
crucial when designing microfluidic reactors. It is generally recognized that first-order irreversible
reactions are unaffected by local turbulent mixing, but are controlled by the residence time of the
reaction and the conversion of such systems can therefore be easily calculated. However, in the case
of fast reactions, where two or more reagents are initially present in separate streams, reaction rarely
occurs throughout the whole volume uniformly. As the mixture is initially highly segregated with
a heterogeneous and dispersed structure the reaction is stifled until mixing occurs. In this situation,
the rate of reaction is no longer related to the rate constant but instead is limited by the rate of
diffusion. In the case of fast reactions, where a single reaction product is produced, the yield can
be regarded as a direct measure of the degree of mixing in the reaction vessel. In the case of fast
reactions with two or more products, the product distribution has similar connotations. In this way,
the product distribution can be manipulated by increasing or decreasing the amount of segregation
or the efficiency of mixing within the reactor. The relationship between the rate of a reaction and
the rate of mixing in a reaction falls into one of three main categories. These are the chemical
regime, the diffusional regime and the mixed chemical/diffusional regime. The interaction between
chemical reactions and fluid dynamics is described by the Damkoehler number, which represents
the contribution of these two systems [6]. The Damkoehler number can be represented in either of
two forms, NDaI and NdaII, where NDaI is a ratio of the chemical reaction rate to the bulk mass
flow rate and NDaII is the ratio of the chemical reaction rate to the molecular diffusion rate. In the
chemical regime, the mixing time is fast in comparison to the reaction rate and is represented by a
Damkoehler number that is close to zero. In this situation, the chemical reaction is slow compared to
the mass transfer and the reaction system can be regarded as an unreactive fluid. That is to say that
mixing is complete before significant amounts of products are present and once mixed, the reaction
proceeds uniformly throughout the reaction volume. In the case of a competitive-consecutive reaction
as detailed in Scheme 1.1, the smallest amount of secondary products is formed. In the diffusion
regime (where NDa ≈ ∞), the chemical reaction is very fast and thus the reaction is limited
by the rate of mixing. In this condition, the area available for reaction is reduced from the vol-
ume of the entire reaction vessel at NDa ≈ 0 to a plane between reacting streams at which
each reagent has zero concentration. When a chemical reaction occurs, microscopic concentra-
tion gradients are introduced increasing segregation, which in turn leads to a decreased level of
mixedness. The reaction rate is then independent of the rate constant and the formation of sec-
ondary products in this situation is the greatest for the reaction system detailed in Scheme 1.1.
In a mixed chemical/diffusional regime (NDa = 1), the greatest interaction between chemical
reactions and fluid dynamics occurs and the product distribution depends on both the chemical
k1
A + B −→ R
k2
R + B −→ S
Scheme 1.1 The competitive reaction of the product R with the initial reagent B leads to the formation of
by-product S and a reduction in the overall yield of R.
factors, (e.g., rate constants) and also on the diffusional factors (e.g., efficiency of mixing). In
this situation, the amount of secondary product formed lies between the extremes of the chemical
and diffusional regimes. Some practical examples, where micromixing has a profound effect on the
product are fast consecutive competing reactions, polymerizations and precipitations. In competitive-
consecutive reactions, a reaction occurs between two reagents (A and B) to produce a product (R)
at rate k1 . The product in this situation is then able to react with one or more of the initial starting
reagents in a competitive manner at rate k2 to form a by-product (S), leading to a reduction in the
initial product (Scheme 1.1). The competitive reaction of the product R with the initial reagent B
leads to the formation of by-product S and a reduction in the overall yield of R In the case of fast
competitive-consecutive reactions k1 is much larger than k2 and k2 is still large. If a small amount of
B is mixed with a large excess of A, R is instantaneously formed. If the fluid is completely mixed,
R and B are dispersed throughout the reaction volume, B is consumed by the excess of A and R is
preserved. If a small amount of B is mixed with a large excess of A and the fluid is a segregated
mixture, R stays at the fluid interface of B. In the presence of B, R is immediately converted to S,
the concentration of which then may be used as a measurement of segregation or unmixedness in
the vessel.
42.3 MICROFABRICATED REACTION SYSTEM

42.3.1 FABRICATION OF MICROFLUIDIC REACTORS
Over the past decade a diversity of microfluidic systems have been designed for the rapid mixing
of pure fluid components. Microfluidic devices can be manufactured using a diversity of fabrication
techniques originally developed in the semiconductor and microelectronics industries. Since most
system features are relatively large (>1 µm), fabrication is straightforward and can be achieved
using well-established lithographic and microstructuring methods. Although, extensive details of
common fabrication methods can be found in excellent articles elsewhere [7], a brief description of
a common protocol for the fabrication of planar glass (or silicon) microfluidic reaction systems is
provided (Figure 42.1).
In the simplest case, a fluidic channel pattern is structured within a planar substrate using a
combination of photolithography, wet-etching, and bonding. Initially, a photosensitive polymer resist
is deposited on to a glass substrate and then exposed to ultraviolet (UV) radiation through a mask
(which defines the fluidic topography). Subsequent development in a solvent allows the removal of
portions of photoresist, which have been exposed (or unexposed) to radiation, leaving a polymerized
resist pattern with high chemical resistance in some areas and exposed substrate in others. Etching of
the substrate material then allows the two-dimensional resist pattern to be transferred to the substrate
material. Typically, wet etching protocols, involving the use of aqueous etchants, such as HF, HNO3 ,
KOH, and tetramethyl ammonium hydroxide, are used to create three-dimensional structures in the
substrate material. The final stage in the fabrication of a basic microfluidic reaction system involves
the assembly of the structured substrate with an unstructured coverplate (typically glass or silicon)
to form the enclosed chip structure. Anodic bonding is used to bond glass and silicon substrates,
while for glassy materials thermal bonding (450–900◦ C) provides the simplest way to assemble
substrates.
42.3.2 FLUID MOTIVATION

An important factor to consider when using microfluidic reactors to perform chemical synthesis is
the method by which reagents are moved through the microdevice. Normally, pressure-driven flow
using infusion pumps is used to force fluid through microchannels. The wide volumetric flow rate
ranges (pL/min to mL/min) and applicability to all solvent systems makes hydrodynamic pumping
the most versatile and popular method of flow motivation. In addition, electroosmotic pumps can be
used to generate fluid flow by application of a potential difference through a conductive solvent [8].
Photoresist
substrate
Align photomask and substrate

UV radiation
Photomask
Expose photoresist
Exposed
photoresist
Remove exposed photoresist
Etch exposed substrate
Remove unexposed photoresist
Bond coverplate to structured substrate
Coverplate
FIGURE 42.1 Schematic representation of component processes in the fabrication of a basic glass microfluidic
reactor using photolithography, wet-chemical etching, and thermal bonding.
The primary advantage of such an approach is that it creates a flat flow profile, and thus allows
definition of precise flow rates and narrow residence time distributions. Nevertheless, electroosmotic
flow (EOF) pumps are severely limited in their widespread application to molecular synthesis due
to a need for a conductive solvent and the fact that varying electrophoretic mobilities of reagents
and products leads to time-dependent concentration gradients within the reactor that can degrade
performance.
42.3.3 MIXING MODALITIES

A wide range of microfluidic systems have been designed to efficiently mix pure fluid components [9].
Broadly, all of these systems can be classified as being either passive or active in operation. Passive
mixers rely on the geometric properties of the channel shape and fluidic streams to achieve mixing,
whereas active mixers rely on time-dependent perturbations of fluid flow to achieve mixing. In
general, passive systems are simple to fabricate and integrate with other functional components, while
active systems require the integration of external actuators and involve more complex fabrication
techniques. Passive mixing is realized via a diversity of mechanisms including serial or parallel flow
lamination, chaotic advection, and injection of substreams of one component into a primary stream
of another component and microdroplet formation. Similarly, active mixers exploit actuation via
pressure, electrohydrodynamic, dielectrophoretic, acoustic, thermal, and electrokinetic mechanisms.
Detailed reviews of the structure, function, and mode of operation of microfluidic mixers are provided
elsewhere.
Of particular interest in relation to reactive microfluidic systems have been recent developments
in the use of chaotic advection to accelerate mixing. In many situations, practical limitations (such
as minimum feature dimensions) mean that basic flow lamination is inefficient at generating high
degrees of mixedness within short times. Rapid mixing with low reagent consumption is, however,
readily achievable using chaotic advection [10]. Chaotic advection may enhance mixing in laminar
flow systems by continuously “stretching” and “refolding” concentrated solute volumes, thereby
creating an exponential decrease in striation thickness. In microfluidic systems, this can be achieved
by introducing obstacles within channels or by modifying channel geometries. Chaotic advection
has been shown to generate efficient mixing within both high and low Reynolds’ number regimes.
For example, zigzag-shaped channels have been shown to cause chaotic advection at high Reynolds’
numbers by recirculation around turns [11], while generation of chaotic flow at low Reynolds’
numbers has been achieved through the use of grooves on channel surfaces [12].
The ability to achieve rapid mixing of fluids while minimizing the deleterious effects of dis-
persion is a significant challenge when creating synthetic systems. Put simply, dispersion acts
to increase residence time distributions within continuous-flow [13], which in turn causes sig-
nificant variation in the yield, efficiency and product distribution. Although chaotic flow can be
used to reduce dispersion in microfluidic systems, localization of reagents within discrete droplets
has been shown to be effective at entirely eliminating this phenomenon [14]. Indeed, a number
of recent studies have exploited the formation of droplets with microfluidic channels to perform
a variety of synthetic processes [15]. In these reports, droplets are made to spontaneously form
when multiple lamina streams of aqueous reagents are injected into an immiscible carrier fluid.
The formed picoliter droplets are isolated from channel surfaces and other droplets, with each
one acting as an individual reaction vessel. Importantly, the use of winding channel geometries is
highly efficient in generating chaotic mixing within droplets and allows rapid and dispersion free
mixing.
42.4 APPLICATION OF MICROFLUIDIC REACTORS IN SYNTHETIC

CHEMISTRY
42.4.1 KEY BENEFITS OF MICROFLUIDIC SYSTEMS IN SYNTHESIS
Small molecule synthesis forms the basis of the pharmaceutical and fine chemical industries. These
industries have to address sometimes competing concerns: the synthesis of high purity compounds
and the synthesis of large quantities of a compound. Microfluidic systems have been employed
to address these concerns and show many advantages over existing techniques. Many of these
advantages are case specific, but nevertheless certain key areas can be considered.
From a synthetic point of view, the essential features of the microscale are functions of mass and
energy transfer. The rapidity and efficiency of mixing is often one of the limiting factors on the rate
of reaction for chemical processes. The other commonly found limit is the rate at which heat can be
transferred out of (or less commonly into) the reaction vessel. As we have seen previously, mixing on
the microscale is a rapid and controlled diffusive process, limited only by the diffusion coefficients of
the solutes themselves. Given small enough diffusive distances, such as can be found in multilaminar
microfluidic mixers, homogeneity can be achieved faster than is possible in macroscale systems, and
the mixing is easily quantifiable [16]. Mass transfer between phases in microfluidic reactors is also
rapid, though more complex than the simple case for diffusive mixing. The flow regimes within
microreactors are normally strictly laminar and so readily quantifiable into flow modes. Even so,
the prediction of flow modes is complex and strongly dependent on local conditions such as solvent
viscosity and phase compressibility. Irrespective of this, work on microfluidic multiphase reactions
has shown that an acceleration of the rate of mass transfer of at least two orders of magnitude
is experienced in microfluidic systems compared to commonly quoted laboratory scale reactions
[17]. Studies on slug flow reaction systems have demonstrated convincing benefits to employing
microreactors for phase-transfer reactions due to the enhanced rate of transfer between phases in
these systems [18].
Heat transfer is limited by a “cube/square” relationship. Heat is generated throughout the reaction
volume, but can only be removed at the reactor surface. From this point of view, it can be argued that
the traditional spheroidal reaction vessel is perhaps the worst configuration imaginable for effective
heat transfer from the vessel to the surroundings. Given the very high surface area-to-volume ratios
present in microfluidic reactors these systems represent a very efficient heat transfer system indeed.
The limitation on the system in these devices is often found to be the intrinsic conduction properties
of the reactor materials themselves, rather than the heat transfer between the reaction medium and
the vessel wall [19]. The precise control of local thermal and concentration gradients possible within
microfluidic reactors enable much greater control of synthetic processes. Many systems have a
narrow optimal range of conditions for efficient and specific synthesis and benefit from the greater
control of these factors inherent in micron scale reactors. In general, it is fair to say that different areas
within the chemical and pharmaceutical industries require different approaches within the broader
theme of microfluidic flow systems. These requirements are briefly discussed below.
42.4.2 FIELDS OF APPLICATION

42.4.2.1 High-Throughput Synthesis
In high-throughput synthesis, the chief problems that have to be overcome are the time it takes to
perform individual syntheses and the purity of the products. The volume of product produced is
not of primary concern in most cases because the target compounds synthesized are destined for
lead screening, a process requiring milligram amounts or sometimes less. The chief aim within this
area is to produce as many pure compounds as possible within the shortest possible space of time
with high target purity. The compound libraries thus generated are then evaluated in a process often
completely divorced from the original synthesis. Traditional approaches toward library generation
have relied upon two divergent paradigms: solid phase synthesis and automation.
Solid phase chemistry has been the technique of choice for research into high-throughput syn-
thesis toward the generation of large compound libraries. In this approach, the synthetic targets
are attached by a linker group to insoluble functionalized, polymeric material, allowing them to be
readily separated from excess reagent, soluble reaction by-product or solvents. Such an approach
involves immobilization of substrates on a polymeric resin or beads using linker groups followed by
a synthetic sequence and finally a cleavage step to obtain the free product.
This approach suffers from problems related to the influence of the support on reaction chemistries
and the optimization of solid-supported chemistries, which may have completely different kinetics
from their solution-phase analogues. Difficulties exist in assessing product purity or structure of
compounds in the solid phase, often leading to expensive spectroscopic solutions. Working in the
solid phase also limits the range of chemistry, which can be applied to a synthetic problem, not least
because many solid phase approaches are so dependent upon amide bond formation. A further issue
arises from the need for the perfect linker: uncleaved in normal reaction conditions but perfectly
cleaved when desired, preferably without a footprint in the target molecule.
The above limitations with solid phase chemistries have necessitated the pursuit of solution-
phase chemistries for library generation. The advantages of solution-phase syntheses are legion, but
the major benefits include; unlimited numbers and types of reaction, lower requirements of solvents
and reagents compared to solid phase syntheses and that such chemistries can be developed and
monitored with relative ease. A conventional solution-phase chemical synthesis involves the use
of arrays of sub millimeter wells as discrete reaction vessels to which the reagents are delivered
using automated robotic systems. Although such systems have been successful, the batch nature of
(a)
(b)
A1 A2 A3 A4 A5 A6 A7
B1 C1 C2 C3 C4 C5 C6 C7
B2 C8 C9 C10 C11 C12 C13 C14
B3 C15 C16 C17 C18 C12 C19 C20
FIGURE 42.2 (a) Continuous-flow glass microfluidic reactor used for generating combinatorial libraries.
(b) The formation of a 7 × 3 pyrazole library was programmed and executed in the automated platform using
the device in part (a). (Taken from Garcia-Egido, E., et al., Lab on a Chip, 2003, 3: 73–76.)
the general approach is not ideal for efficient process optimization and high-throughput chemical
processing.
The use of microfluidic systems in high-throughput chemistry offers great advantages. Reactions
can be run in the solution phase, while still allowing rapid reaction throughput and with no detectable
cross-contamination. For example, work by Mitchell et al. [20] using an integrated continuous-flow
reaction/detection system (expressed as a logic gate) showed that the rapid generation of a small
library of compounds via a multicomponent reaction could be accomplished with ease. Using such
an approach, valuable information about reaction intermediates can be gathered in real time with
the entire process being readily controlled by simple automation protocols. Garcia-Egido et al. [21]
also showed that using slugs of reagents as separate reaction vessels small libraries of thiazoles
could be generated then purified and analysed by in-line LC-MS (Figure 42.2), with a noticeable
improvement in turnover compared to bulk systems and no detectable cross-contamination. Indeed
utilizing similar systems, the synthesis of libraries of cycloadducts [22], pyrazoles [21], nitrostilbene
esters [23], and thioethers [24] has been accomplished.
Although the bulk of these systems represent only one synthetic step, multistep combinatorial
processes have been undertaken, notably by Sakai et al. [25]. High-throughput synthesis on the
microfluidic scale has been proven to be advantageous in terms of the available chemistries, product
purity and, ease of separation. Table 42.1 lists most of the successfully advantageous examples
of synthesis performed within microfluidic reactors and recorded within the literature. One other
advantage that has been demonstrated is that the same chemistry used to produce the sample can be
used to produce large quantities of product. This multiscale nature of microfluidic reactors is one of
enormous potential value to the chemical and pharmaceutical sector.
42.4.2.2 Multiscale Synthesis

In traditional synthetic schemes, the transition from small scale production to large-scale production
is a troublesome one. Reactions will almost never perform in the same way on a production scale
as they do on a laboratory scale. Some reactions may be simply too dangerous to perform on a
large scale, especially if they involve a dangerous intermediate, or a strong exotherm. Scale-up of a
chemical process can, for these and other reasons, take years.
Continuous-flow processes allow greater access to laboratory reactions than batch processes.
This is because the instantaneous reactor size is smaller, better simulating laboratory conditions
and the assay of any dangerous intermediates generated can be kept small. Microfluidic reaction
systems offer one superlative advantage over other methodologies: that of scale-out. Scale-out (or
numbering up) is a simple proposition. If one microfluidic reactor produces a small quantity of
product, when more is desired the number of reactors and not the scale of the reactor, is increased.
Because microfluidic devices are easily mass produced, this system of parallel reactors is relatively
cheap. Furthermore, because of the continuous-flow nature of the reaction environment a surprisingly
large amount of product can be produced in a short space of time and the reaction conditions within
the production reactor exactly mimic those of the research reactor. In other words, no scale-up
development work is necessary. An excellent example of this approach is in the pioneering work
undertaken by R.D. Chambers and coworkers in the use of elemental fluorine for fluorinations.
Direct fluorination is a process that is generally viewed as unscaleable. The highly exothermic
nature of the system ensures that large-scale reactions are problematic and the use of fluorine gas in
large assay is in itself undesirable. Early work on microfluidic fluorinations showed that the direct
reaction was comparable in selectivity to the industrial Schiemann process [26] and that the system
could be applied to a wide range of compounds [27]. In a series of papers, Chambers et al. [28,29]
showed that by the simple numbering up of reactor channels a commercially viable reactor could be
constructed with no loss of reaction efficiency or selectivity.
The numbering up of microfluidic systems is a powerful tool in the rapid development of synthetic
processes from the laboratory to fine scale and beyond and is only beginning to show its versatility [30]
(Figure 42.3).
42.4.2.3 High-Selectivity Synthesis

The highly ordered mass and heat transfer processes within microreactors often produce very selective
reactions. As will be seen later this has implications for nanomaterial synthesis, but even within well-
known reactions this selectivity can be important. Burns and Ramshaw [31] showed that nitration
processes within microfluidic systems produce cleaner products than bulk scale systems. Similarly
high yields and low by-product formation has been reported in on-chip peptide formation [32]. This
is almost certainly attributable to the thermal flatness found within microfluidic channels and the
ordered and predictable mixing within the system. It appears that reactions within the microfluidic
regime are cleaner and often quicker, than their bulk equivalents.
42.4.2.4 Catalytic Systems

The use of microfluidic systems for catalytic reactions has several advantages. The high surface
area-to-volume ratios encountered on the microscale allow for good contact between heterogeneous
catalysts and the reaction medium, and the thermal flatness eliminates hotspot formation. This has
made microfluidic systems a valuable tool for process development in systems dependent upon a
limited range of catalyst temperatures for optimal operation. For example, Worz and coworkers [33]
at BASF showed that by using this approach, the efficiency of their catalytic process was increased
and the reliability of the catalyst improved. Worz [33] states that the use of this technology represented
a considerable saving in development time. For heterogeneous catalysis, many systems have been
F2/N2 gas
Substrate
solution
Product
Steel block Nickel or Stainless Transparent Top steel

with three steel steel PTFCE plate window
reservoirs plate channel plate
plate
FIGURE 42.3 A modular microreactor system for the direct fluorination of ethyl acetoacetate by fluorine gas.
The design allows for multiple channels to be supplied from single reservoir sources and a multichannel device
to be constructed in a facile manner from a disposable channel plate. (Taken from Chambers, R.D., et al., Lab
on a Chip, 2005, 5: 191–198. With permission.)
used from packed beds [17] through solid supports [34] to derivitized channel walls [35–37], each
with advantages for particular catalytic systems. Indeed the breadth of application for microfluidic
reaction systems in catalytic process development is hard to overestimate.
42.5 REAL WORLD APPLICATIONS

Real world synthetic problems are rarely solved in one step, but are more normally addressed by
a combination of factors leading to complex solutions. The “plug and play” nature of microfluidic
components makes them a flexible tool in problems of this nature, even when the whole process
in question cannot be transferred onto a single microfluidic device. The following section outlines
some illustrative examples of such applications.
42.5.1 POINT OF USE MANUFACTURE

Synthesis of short half-life radiotracers positron emission tomography (PET) is an emerging field
within radiography. The technique utilizes positron-emitting radiotracers to trace metabolic activity
directly within the body. The chief disadvantage of this methodology is the need for the tracers
themselves: the short half-lives dictate that the compounds must be synthesized close to the point of
use and the fiercely radioactive nature of the isotopes used entails prodigious shielding. The general
approach toward these syntheses is to use automated robotic synthesis stations, often custom built,
to perform the necessary synthesis and processing within a shielded compartment known as a hot
cell. Space within these cells is at a premium and the number of automated stations per cubic metre
is the major constraint on the procedure.
Recent work by Jeffery et al. [71] showed that many of the standard synthetic operations that
are necessary in the synthesis of [18] FDG (a major PET adiotracer) could be incorporated into
microfluidic devices. Moreover, Lu et al. [72] showed that the methodology was also applicable to
[11] C compounds. In an elegant recent study, Lee and coworkers [73] showed that all of the major
processes in [18] F radiolabeling tracer compounds could be integrated onto a single device. Such
advances are revolutionizing the field, with continuous-flow syntheses on microfluidic devices being
more efficient in terms of throughput, space and indeed energy requirements than their conventional
counterparts.
42.5.2 MEASUREMENT OF CHEMICAL PHASE SPACE

All chemical processes are optimized to some degree. This entire procedure can be thought of as
finding a set of localized optimal conditions within a reaction’s output. The most common way of
quantifying this form of relationship is by using phase space, an n-dimensional notional space, where
each input to or output from a reaction is dealt with as a separate dimension. Using this model allows
an intuitive handling of data. The main problem with applying the phase space model to reactions in
the real world is the lack of data. When each experiment takes 3 h to run the number of data points
generated will be quite low unless the optimization process has an unlimited budget. Automation of
kinetic and calorimetric data gathering has made some improvements and allows the use of Design
of Experiments algorithms, which can help to find minima quickly. However, using microfluidic
devices it is possible to speed up the data gathering process. Studies by Le Bars and coworkers [74]
has shown that calorimetric readings can be taken in continuous flow in a microfluidic device.
Furthermore, in-line monitoring using, inter alia, NMR spectroscopy allows real-time evaluation of
reaction conditions [75]. In addition, a useful demonstration of the benefits of microfluidic devices
for reaction optimisation was presented by Ratner et al. [68] who investigated the influence of
systematic variations in reaction time and temperature on glycosylation reactions (Figure 42.4).
In a series of studies our group demonstrated that not only could reactions be monitored on-
line [76] and these data used for very rapid optimization [77], but also that this information could
be used to plot local phase space in detail rarely attempted before. Several thousand data points
per minute can be generated using appropriate in-line sampling and control techniques, allowing
detailed phase space surfaces to be plotted [78]. This represents an increase in speed of over four
orders of magnitude when compared to conventional methods.
42.5.3 MICROFLUIDIC SYSTEMS FOR NANOMATERIAL PRODUCTION

As discussed, microfluidic systems are well suited for performing reagent mixing in a rapid and
controllable manner. This combined with control over other variables, such as reactor temperature,
concentration gradients and pressure dictate that continuous-flow processing on the microscale can
be used to synthesize species of specific yet variable characteristics. Perhaps the most interesting
demonstration of this feature is in the synthesis of nanomaterials (such as nanoparticles, nanotubes,
and nanorods). The physical characteristics of nanocrystallites are primarily determined by quantum
confinement effects with properties such as the optical band gap often differing considerably from
the bulk material. Consequently, they are seen as customized precursors for functional materials in
applications such as biological sensing and optoelectronics [79]. Since the electronic and optical
properties of these materials are ultimately controlled by the physical dimensions of the crystal-
lites, there is considerable interest in processing routes that yield nanoparticles of well-defined size
and shape. Synthetic approaches for nanomaterial synthesis involve particle growth on an atom-
by-atom or molecule-by-molecule basis until the desired size is achieved [80]. Such growth takes
place spontaneously in super-saturated solutions and has been successfully used to create crystal-
lites of well-defined size and shape. “Bottom-up” or synthetic approaches are attractive due to their
versatility and ease of use; however, standard syntheses rarely yield product size distributions bet-
ter than ±5%. This means that it is almost always necessary to use some form of post-treatment
(including electrophoresis, chromatography, precipitation and photocorrosion) to extract particles of
well-defined size. Although nanoparticles with extremely narrow size distributions can ultimately
be extracted, the starting point for all such methods is a polydisperse sample and thus product yields
and conversions are necessarily low.
(a) (b)
(c)
(d)
(e)
FIGURE 42.4 (a) Silicon microfluidic reactor for performing glycosylation reactions over a wide range of
experimental conditions. (b) Schematic representation of microfluidic circuit, which comprises three primary
inlets, a mixing and reaction zone, a secondary inlet for quenching reagent and an outlet for analysis/collection.
(c) Side view of microreactor. (d and e) Glycoside coupling reactions. (Taken from Ratner, D.M., et al., Chemical
Communications, 2005, 578–580. With permission.)
At a fundamental level monodisperse nanoparticle populations can be created by ensuring that

initial nucleation of solute molecules (to form “seed” particles) occurs on a timescale, which is
rapid compared with the growth process (in which the seed particles confine dissolved solutes) [81].
Moreover, nuclei formation and growth should occur within an environment in which state functions
are precisely controlled throughout. When these circumstances are not met, the size of nuclei and
the particle growth rates vary according to location and result in a wide distribution of particle sizes.
Unsurprisingly, a number of recent studies have demonstrated the effectiveness of continuous-flow
microfluidic reactors in performing controlled nanoparticle synthesis [82]. These systems generally
utilize simple flow regimes, whereby pure fluid component streams are mixed within low Reynolds’
number regimes. Variation in reaction residence times and reagent concentrations can be used control
average particle size, while product polydispersity is minimized through reduction in residence time
distributions and precise control of chemical state functions. For example, silver, cobalt, copper,
cadmium sulphide [83,84], cadmium selenide [14,85], gold [86,87], palladium [88], titania [89,90],
and CdSe-ZnS [91] core-shell have all been produced within a range of microfluidic reactors. In all
cases low-polydispersity nanoparticles of varying size could be synthesized with space-time yields
significantly higher than corresponding macroscale approaches.
More recent studies have addressed the issue of minimizing particle size distributions through
the development of segmented flow reactors. For example, Ismagilov and coworkers [92] have
reported multistep chemical production of CdS and CdS/CdSe core-shell nanoparticles within a
microdroplet-based reactor. Importantly, such an approach allowed for millisecond time control and
also enabled multiple reactions to be initiated by flowing additional reagent streams directly into
individual droplets. In addition, Chan et al. [93] have reported the use of microfluidic droplet reactors
for the high-temperature synthesis of CdSe nanoparticles, while Yen et al. [94] have a used gas–liquid
segmented flow reactor containing multiple temperature zones (Figure 42.5) for the synthesis of high
quality CdSe quantum dots. In all studies, enhanced mixing and reduced residence time distributions
provide the driving force for improvements in reaction yield and size distribution.
Precursor
ar solutions Deep trench
(a)
en
a
e
zo
zon
ng
ati
ting
he
b
hea
quench
zone
c
Outlet Thermal
insulation
(b)
gas Precursors
(c)
1 mm
FIGURE 42.5 Microfluidic reactor for nanoparticle production. (a) Schematic illustrating the key components
of the reactor. The reactor performs precursor mixing (section a), controlled particle growth (section b) and
reaction quenching (section c). Ahalo etch region allows localization of temperature zones for reaction (>260◦ C)
and quenching (<70◦ C). (b) Photograph of heated inlets and (c) photograph of main channel section. Red
segments show the reaction solution, with dark segments defining argon gas. (Taken from Brian, K.H., et al.,
Angewandte Chemie-International Edition in English, 2005, 44, 5447–5451. With permission)
TABLE 42.1
Diversity of small-molecule syntheses performed within microfluidic reactors.
Reaction Product(s) Conversion Comments Year/Reference
Bayer–Villiger oxidation Lactones <99% Increased reactivity and selectivity 2006/[38]
Baylis–Hillman reaction Baylis–Hillman adducts <95% C−C bond forming 2006/[39]
Autocatalytic nitration Nitrophenols 65–76% Autocatalysis 2005/[40]
Catalytic dehydrogenation Toluene 88% Gas-solid heterogenous catalysis 2005/[41]
Suzuki–Miyaura coupling Functionalised biphenols <76% Hydrodynamic pumping 2005/[42]
Grignard exchange reaction Pentafluorobenzene 92% Formation of Grignard reagents 2005/[43]
Friedel–Crafts aminoalkylation 1-(N-Butyl-N-methoxycarbonyl 96% Improved selectivity with highly reactive substrates 2005/[44]
aminomethyl)-2,4,6-trimethoxybenzene
Imine hydrogenation Amines Quantitative 2005/[45]
Stereoselective alkylation C–C bond 41% (ee 82%) Using chiral auxiliary 2004/[46]
Hydrogenations Alkenes Quantitative Triphase reaction system immobilized Palladium 2004/[47]
Michael addition 1,3-Diketones 95% Stopped flow technique 2002/[48]
Suzuki coupling Functionalized biphenols 70–100% Microwave assisted 2004/[49]
Heterocyclic synthesis 1,2-Azoles 98–100% Significant improvement on batch 2004/[50]
Fluorination Fluorinated diketones 70% Safety enhancement 1999/[27]
Esterification Benzoic acid esters, phenyl esters 100% EOF 2003/[51]
Photo-oxygenation Ascaridole 85% Safety enhancement 2002/[52]
Sandmeyer reaction Diazonium intermediates Various Control over unstable intermediate 2003/[53]
Microfluidic Reactors for Small Molecule and Nanomaterial Synthesis
Carbamate synthesis Methyl carbamates 91% Control over exothermic conditions 2004/[54]
Ring expansion (6–7 membered ring) N-tert-Butoxycarbonyl-5-ethoxycarbonyl-4- 89% Safety and control of reactive ethyl diazoacetate reagent 2004/[54]
perhydroazepinone
Nitration Nitrobenzene 65% 2001/[55]
Wittig reaction Alkene C=C bond formation 39–59% 2001/[23]
Peptide synthesis Peptides 100% Quantitative yield in 20 min 2001/[32]
Aldol reaction C−C bond via enolate 100% Enolate generated via silyl enol ethers 2001/[56]
Enamine formation Enamines 42% Formation using DCC 2001/[57]
Continued
1199
1200
TABLE 42.1
(Continued)
Reaction Product(s) Conversion Comments Year/Reference
Ugi 4 component coupling α-Dialkylacetamide — Detection of reaction products and intermediates 2001/[58]
Dehydration of alcohols Alkene 85–95% Sulfated Zirconia catalyst 2000/[59]
Kumada–Corriu coupling Coupling of aryl halide and Grignard reagent 60% Nickel chelated to immoblilsed salen ligand 2001/[60]
Knoevenagel condensation Condensation followed by hetero Diels–Alder reaction 50–68% 2 × 2 combinatorial array produced four products 2002/[61]
forming cycloadducts
Phase transfer diazo coupling ∼100% Increased surface area, rapid diffusive phase transfer 2001/[62]
Phase transfer alkylations Alkylated β-keto esters 71–96% Increased surface area, rapid diffusive phase transfer 2003/[63]
Hantzsch synthesis 2-Aminothiazoles 58–99% Yields improved over bulk—heated under EOF 2002/[64]
Knorr reaction Pyrazoles 52–99% Sequential synthesis of 7 × 3 library 2003/[21]
Acid catalyzed esterification Pyrenebutyric acid ethyl ester 83% Surface area related rate enhancement 2003/[65]
Swern oxidations Carbonyl compounds 81–100% Room temperature conversions—Batch Swern usually 2005/[66]
below −50◦ C
Photochlorination 1-Chloromethyl-2,4-diisocyanatobenzene 55% 80% selectivity (big incresase over batch) 2002/[67]
Glycosylations Glycosidic link for oligosaccharide assembly — Small quantity of glycosylating agent required 2005/[68]
Heck coupling reaction Ethyl cinnamate 99% Using a modular microreaction system 2005/[69]
Horner–Wadsworth–Emmons Olefins 91% Using a modular microreaction system 2005/[69]
Diels–Alder cycloaddition Cycloadduct 100% Using a modular microreaction system 2005/[69]
Henry reaction 2-Nitro-1-phenylethanol 76% Nitroaldol addition 2005/[69]
Reimer–Tiemann formylation Formylated β-napthol 10% Yields similar to bulk, allowed study of temperature 2005/[70]
effects
42.6 CONCLUSIONS
The nature of the microfluidic environment allows for a control over physical properties such as
energy transfer, mass transfer and mixing that is difficult to achieve using other approaches. The
consequences of this for the synthetic chemist are that processes, which are controlled in large part
by these properties, can be exploited to produce a range of products that are otherwise inaccessible.
The control on the microscale leads to faster, cleaner and more specific reactions. In addition to this,
the nature of the relationship between surface and bulk properties in the microfluidic environment
ensures that reactions involving rapid exchanges of energy become manageable in a way that is not
possible in larger reaction vessels.
Microfluidic approaches also show strong advantage in high-throughput systems, whether these
be for catalyst screening or for combinatorial synthesis. The rapidity of data generation, which
has already revolutionized the science of genomics, is also a major bonus in the application of
microfluidics to these synthetic problems.
REFERENCES
1. Wöhler F., Grundriß der Organischen Chemie. 1848, Berlin: Duncker und Humblot.
2. Purcell, E.M., Life at low Reynolds number. American Journal of Physics, 1977, 45: 3–11.
3. Atencia, J. and D.J. Beebe, Controlled microfluidic interfaces. Nature, 2005, 437: 648–655.
4. Rott, N., Note on the history of the Reynolds number. Annual Review of Fluid Mechanics, 1990,
22: 1–11.
5. Crank, J., The Mathematics of Diffusion. 2 edn. 1975, Oxford, England: Oxford University Press.
6. Chakraborty, S. and V. Balakotaiah, Low-dimensional models for describing mixing effects in laminar
flow tubular reactors. Chemical Engineering Science, 2002, 57: 2545–2564.
7. Kovacs, G.T.A., K. Petersen, and M. Albin, Silicon Micromachining. Analytical Chemistry, 1996, 68:
A407–A412.
8. Feng, X.Z., S.J. Haswell, and P. Watts, Organic synthesis in microreactors. Current Topics in Medicinal
Chemistry, 2004, 4: 707–727.
9. Nguyen, N.T. and Z. Wu, Micromixers—A review. Journal of Micromechanics and Microengineering,
2005, 15: R1–R16.
10. Ottino, M., The mixing of fluids. Scientific American, 1989, 260: 56–67.
11. Mengeaud, V., J. Josserand, and H.H. Girault, Mixing processes in a zigzag microchannel: Finite
element simulations and optical study. Analytical Chemistry, 2002, 74: 4279–4286.
12. Stroock, A.D., et al., Chaotic mixer for microchannels. Science, 2002, 295: 647–651.
13. Krishnadasan, S., et al., On-line analysis of CdSe nanoparticle formation in a continuous-flow
chipbased microreactor. Journal of Materials Chemistry, 2004, 14: 2655–2660.
14. Bringer, M.R., et al., Microfluidic systems for chemical kinetics that rely on chaotic mixing indroplet.
Philosophical Transactions of the Royal Society of London A, 2004, 362: 1087–1104.
15. Song, H., D.L. Chen, and R.F. Ismagilov, Reactions in droplets in microfluidic channels. Angewandte
Chemie-International Edition, 2006, 45: 7336–7356.
16. Ehrfeld, W., et al., Characterization of mixing in micromixers by a test reaction: Single mixing units
and mixer arrays. Industrial and Engineering Chemical Research, 1999, 38: 1075–1082.
17. Losey, M.W., M.A. Schmidt, and K.F. Jensen, Microfabricated multiphase packed-bed reactors: Char-
acterization of mass transfer and reactions. Industrial and Engineering Chemistry Research, 2001, 40:
2555–2562.
18. Burns, J.R. and C. Ramshaw, The intensification of rapid reactions in multiphase systems using slug
flow in capillaries. Lab on a Chip, 2001, 1: 10–15.
19. Rebrov, E., M.H.J.M. de Croon, and J.C. Schouten, Design of a microstructured reactor with integrated
heat-exchanger for optimum performance of a highly exothermic reaction. Catalysis Today, 2001, 69:
183–192.
20. Mitchell, M.C., V. Spikmans, and A. De Mello, Microchip-based synthesis and analysis: Control of
multicomponent reaction products and intermediates. Analyst, 2001, 126: 24.
21. Garcia-Egido, E., et al., Synthesis and analysis of combinatorial libraries performed in an automated
micro reactor system. Lab on a Chip, 2003, 3: 73–76.
22. Fernandez-Suarez, M., S.Y.F. Wong, and B.H. Warrington, Synthesis of a three-member array of
cycloadducts in a glass microchip under pressure driven flow. Lab on a Chip, 2002, 2: 170–174.
23. Skelton, V., et al., The preparation of a series of nitrostilbene ester compounds using micro reactor
technology. Analyst, 2001, 126: 7–10.
24. Jonsson, D., B.H. Warrington, and M. Ladlow, Automated flow-through synthesis of heterocyclic
thioethers. Journal of Combinatorial Chemistry, 2004, 6: 584–595.
25. Sakai, R., et al., On-line MS detection for a multi-step combinatorial synthesis system., in µTAS 2004,
Proceedings of the 8th International Conference on Miniaturised Systems for Chemistry and the Life
Sciences, T. Laurell, et al., Ed. 2004, Royal Society of Chemistry, Cambridge, UK, p. 96–98.
26. Jahnisch, K., et al., Direct fluorination of toluene using elemental fluorine in gas/liquid microreactors.
Journal of Fluorine Chemistry, 2000, 105: 117–128.
27. Chambers, R.D. and R.C.H. Spink, Microreactors for elemental fluorine. Chemical Communications,
1999, 883.
28. Chambers, R.D., et al., Gas–liquid thin film microreactors for selective direct fluorination. Lab on a
Chip, 2001, 1: 132–137.
29. Chambers, R.D., et al., Elemental fluorine—Part 16. Versatile thin-film gas–liquid multi-channel
microreactors for effective scale-out. Lab on a Chip, 2005, 5: 191–198.
30. Commenge, J.-M. and M. Matlosz, From process miniaturisation to structured multiscale design: the
innovative, high-performance chemical reactors of tomorrow. Chimia, 2002, 56: 654.
31. Burns, J.R. and C. Ramshaw, Development of a microreactor for chemical production. Transactions
of the Institute of Chemical Engineering, 1999, 77: 206–211.
32. Watts, P., et al., The synthesis of peptides using micro reactors. Chemical Communications, 2001,
990–991.
33. Worz, O., et al., Microreactors, a new efficient tool for optimum reactor design. Chemical Engineering
Science, 2001, 56: 1029–1033.
34. Wiles, C., P. Watts, and S.J. Haswell, The use of solid-supported reagents within EOF-based micro-
reactors., in µTAS 2004, Proceedings of the 8th International Conference on Miniaturised Systems for
Chemistry and the Life Sciences, T. Laurell, et al., Ed. 2004, Royal Society of Chemistry, Cambridge,
UK, p. 104–106.
35. Wootton, R.C.R. and A.J. deMello, A one-step protocol for the chemical derivatisation of glass
microfluidic devices. Lab on a Chip, 2006, 6: 471–473.
36. Kobayashi, J., Y. Mori, and S. Kobayashi, Triphase hydrogenation reactions utilizing palladiumimmo-
bilized capillary column reactors and a demonstration of suitability for large scale synthesis. Advanced
Synthesis and Catalysis, 2005, 347: 1889–1892.
37. Wan, Y.S.S., et al., TS-1 zeolite microengineered reactors for 1-pentene epoxidation. Chemical
Communications, 2002, 878–879.
38. Mikami, K., et al., Nanoflow microreactor for dramatic increase not only in reactivity but also ins-
electivity: Baeyer-Villiger oxidation by aqueous hydrogen peroxide using lowest concentration of a
fluorous lanthanide catalyst. Journal of Fluorine Chemistry, 2006, 127: 592–596.
39. Acke, D.R.J. and C.V. Stevens, Study of the Baylis-Hillman reaction in a microreactor environment:
First continuous production of Baylis-Hillman adducts. Organic Process Research and Development,
2006, 10: 417–422.
40. Ducry, L. and D.M. Roberge, Controlled autocatalytic nitration of phenol in a microreactor.
Angewandte Chemie-International Edition, 2005, 44: 7972–7975.
41. Roumanie, M., et al., Design and fabrication of a structured catalytic reactor at micrometer scale:
Example of methylcyclohexane dehydrogenation. Catalysis Today, 2005, 110: 164–170.
42. Phan, N.T.S., J. Khan, and P. Styring, Polymer-supported palladium catalysed Suzuki-Miyaura
reactions in batch and a mini-continuous flow reactor system. Tetrahedron, 2005, 61: 12065–12073.
43. Wakami, H. and J. Yoshida, Grignard exchange reaction using a microflow system: From bench to
pilot plant. Organic Process Research and Development, 2005, 9: 787–791.
44. Nagaki, A., et al., Control of extremely fast competitive consecutive reactions using micromixing.
Selective Friedel-Crafts aminoalkylation. Journal of the American Chemical Society, 2005, 127:
11666–11675.
45. Saaby, S., et al., The use of a continuous flow-reactor employing a mixed hydrogen-liquid flow stream
for the efficient reduction of imines to amines. Chemical Communications, 2005, 2909–2911.
46. Wiles, C., et al., Stereoselective alkylation of an Evans auxiliary derivative within a pressure-driven
micro reactor. Lab on a Chip, 2004, 4: 171–173.
47. Kobayashi, J., et al., A microfluidic device for conducting gas-liquid-solid hydrogenation reactions.
Science, 2004, 304: 1305–1308.
48. Wiles, C., et al., 1,4-Addition of enolates to a,b-unsaturated ketones within a micro reactor. Lab on a
Chip, 2002, 2: 62–64.
49. He, P., S.J. Haswell, and P.D.I. Fletcher, Microwave heating of heterogeneously catalysed Suzuki
reactions in a micro reactor. Lab on a Chip, 2004, 4: 38–41.
50. Wiles, C., et al., The application of microreactor technology for the synthesis of 1,2-azoles. Organic
Process Research and Development, 2004, 8: 28–32.
51. Wiles, C., et al., Solution phase synthesis of esters within a micro reactor. Tetrahedron, 2003, 59:
10173–10179.
52. Wootton, R.C.R., R. Fortt, and A.J. de Mello, A microfabricated nanoreactor for safe, continu-
ous generation and use of singlet oxygen. Organic Process Research and Development, 2002, 6:
187–189.
53. Fortt, R., R.C.R. Wootton, and A.J. de Mello, Continuous-flow generation of anhydrous diazonium
species: Monolithic microfluidic reactors for the chemistry of unstable intermediates. Organic Process
Research and Development, 2003, 7: 762–768.
54. Zhang, X.N., S. Stefanick, and F.J. Villani, Application of microreactor technology in process
development. Organic Process Research and Development, 2004, 8: 455–460.
55. Doku, G.N., et al., Electric field-induced mobilisation of multiphase solution systems based on the
nitration of benzene in a micro reactor. Analyst, 2001, 126: 14–20.
56. Wiles, C., et al., The aldol reaction of silyl enol ethers within a micro reactor. Lab on a Chip, 2001, 1:
100–101.
57. Sands, M., et al., The investigation of an equilibrium dependent reaction for the formation of enamines
in a microchemical system. Lab on a Chip, 2001, 1: 64–65.
58. Mitchell, M.C., V. Spikmans, and A.J. de Mello, Microchip-based synthesis and analysis: Control of
multicomponent reaction products and intermediates. Analyst, 2001, 126: 24–27.
59. Wilson, N.G. and T. McCreedy, On-chip catalysis using a lithographically fabricated glass
microreactor—the dehydration of alcohols using sulfated zirconia. Chemical Communications, 2000,
733–734.
60. Haswell, S.J., B. O’Sullivan, and P. Styring, Kumada–Corriu reactions in a pressure-driven microflow
reactor. Lab on a Chip, 2001, 1: 164–166.
61. Fernandez-Suarez, M., S.Y.F. Wong, and B.H. Warrington, Synthesis of a three-member array of
cycloadducts in a glass microchip under pressure driven flow. Lab on a Chip, 2002, 2: 170–174.
62. Hisamoto, H., et al., Fast and high conversion phase-transfer synthesis exploiting the liquid–liquid
interface formed in a microchannel chip. Chemical Communications, 2001, 2662–2663.
63. Ueno, M., et al., Phase-transfer alkylation reactions using microreactors. Chemical Communications,
2003, 936–937.
64. Garcia-Egido, E., S.Y.F. Wong, and B.H. Warrington, A Hantzsch synthesis of 2-aminothiazoles
performed in a heated microreactor system. Lab on a Chip, 2002, 2: 31–33.
65. Brivio, M., et al., Surface effects in the esterification of 9-pyrenebutyric acid within a glass micro
reactor. Chemical Communications, 2003, 1924–1925.
66. Kawaguchi, T., et al., Room-temperature swern oxidations by using a microscale flow system.
Angewandte Chemie-International Edition, 2005, 44: 2413–2416.
67. Ehrich, H., et al., Application of microstructured reactor technology for the photochemical chlorination
of alkylaromatics. Chimia, 2002, 56: 647–653.
68. Ratner, D.M., et al., Microreactor-based reaction optimization in organic chemistry glycosylation as a
challenge. Chemical Communications, 2005, 578–580.
69. Snyder, D.A., et al., Modular microreaction systems for homogeneously and heterogeneously catalyzed
chemical synthesis. Helvetica Chimica Acta, 2005, 88: 1–9.
70. Iles, A., R. Fortt, and A.J. de Mello, Thermal optimisation of the Reimer-Tiemann reaction using
thermochromic liquid crystals on a microfluidic reactor. Lab on a Chip, 2005, 5: 540–544.
71. Jeffery, N.T., et al., Radiochemistry on microfluidic devices: proof of principle. Supplement to the
Journal of Nuclear Medicine: Society of Nuclear Medicine, 2004, 145.
72. Lu, S.Y., et al., Syntheses of C-11- and F-18-labeled carboxylic esters within a hydrodynamicallydriven
micro-reactor. Lab on a Chip, 2004, 4: 523–525.
73. Lee, C.-C., et al., Multistep synthesis of a radiolabeled imaging probe using integrated microfluidics.
Science, 2005, 310: 1793–1796.
74. Le Bars, J., et al., A scale-transparent reaction calorimetric assay for rapid catalyst selection. Advanced
Synthetic Catalysis, 2001, 343: 207.
75. Wensink, H., et al., Real-time monitoring of chemical reactions inside a micro NMR chip, in µTAS
2004, Proceedings of the 8th International Conference on Miniaturised Systems for Chemistry and the
Life Sciences, T. Laurell, et al., Ed. 2004, Royal Society of Chemistry, Cambridge, UK, p. 192–194.
76. Fortt, R., R.C.R. Wootton, and A.J. deMello, Continuous-flow generation of anhydrous diazonium
species: monolithic microfluidic reactors for the chemistry of unstable intermediates. Organic Process
Research and Development, 2003, 7: 762–768.
77. Leung, S.-A., et al., A method for rapid reaction optimisation in continuous-flow microfluidic reactors
using online Raman spectroscopic detection. The Analyst, 2005, 130: 46–51.
78. Cullen, C.J., et al. Rapid phase space surface generation using an integrated microfabricated device
reaction detection system and automated control, in Micro Total Analytical Systems 2006: The 10th
International Conference on Miniaturized Systems for Chemistry and Life Sciences, 2006, Tokyo:
Society for Chemistry and Micro-Nano Systems.
79. Alivisatos, A.P., Semiconductor clusters, nanocrystals, and quantum dots 10.1126/ sci-
ence.271.5251.933. Science, 1996, 271: 933–937.
80. Murray, C.B., C.R. Kagan, and M.G. Bawendi, Synthesis and characterization of monodisperse
nanocrystals and close-packed nanocrystal assemblies. Annual Review of Materials Science, 2000,
30: 545–610.
81. LaMer, V.K. and R.H. Dinegar, Theory, production and mechanism of formation of monodispersed
hydrosols. Journal of the American Chemical Society, 1950, 72: 4847–4854.
82. Mello, J.D. and A.D. Mello, FocusMicroscale reactors: nanoscale products. Lab on a Chip, 2004, 4:
11N–15N.
83. Edel, J.B., et al., Microfluidic routes to the controlled production of nanoparticles. Chemical
Communications, 2002, 1136–1137.
84. Hung, L.-H., et al., Alternating droplet generation and controlled dynamic droplet fusion in microfluidic
device for CdS nanoparticle synthesis. Lab on a Chip, 2006, 6: 174–178.
85. Yen, B.K.H., N.E. Stott, K.F. Jensen, and M.G. Bawendi, A continuous-flow microcapillary reactor for
the preparation of a size series of CdSe nanocrystals. Advanced Materials, 2003, 15: 1858–1862.
86. Wagner, J., et al., Generation of metal nanoparticles in a microchannel reactor. Chemical Engineering
Journal, 2004, 101: 251–260.
87. Shalom, D., et al., Synthesis of thiol functionalised gold nanoparticles using a continuous flow
microfluidic reactor. Materials Letters, 2007, 61: 1146–1150.
88. Song, Y., C.S.S.R. Kumar, and J. Hormes, Synthesis of palladium nanoparticles using a continuous-
flow polymeric micro reactor. Journal of Nanoscience and Nanotechnology, 2004, 4: 788–793.
89. Wang, H., et al., Preparation of titania particles utilizing the insoluble phase interface in a microchannel
reactor. Chemical Communications, 2002, 1462–1463.
90. Cottam, B.F., et al., Accelerated synthesis of titanium oxide nanostructures using microfluidic chips.
Lab On a Chip, 2007, 7: 167–169.
91. Wang, H., et al., Continuous synthesis of CdSe-ZnS composite nanoparticles in a microfluidic reactor.
Chemical Communications, 2004, 48–49.
92. Shestopalov, I., J.D. Tice, and R.F. Ismagilov, Multi-step synthesis of nanoparticles performed on
millisecond time scale in a microfluidic droplet-based system. Lab on a Chip, 2004, 4: 316–321.
93. Chan, E.M., A.P. Alivisatos, and R.A. Mathies, High-temperature microfluidic synthesis of CdSe
nanocrystals in nanoliter droplets. Journal of the American Chemical Society, 2005, 127: 13854–13861.
94. Brian, K.H., A.G. Yen, M.A. Schmidt, K.F. Jensen, and M.G. Bawendi, A microfabricated gas–liquid
segmented flow reactor for high-temperature synthesis: the case of CdSe quantum dots. Angewandte
Chemie-International Edition in English, 2005, 44: 5447–5451.
Part IIIB
Microchip-Based: Specialized
43 Sample Processing with
Integrated Microfluidic
Systems
Joan M. Bienvenue and James P. Landers
CONTENTS
43.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1207

43.2 DNA Purification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1208
43.3 PCR Amplification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1213
43.4 Cell Sorting, Cell Lysis, and DNA Quantitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1215
43.5 Integration of Microfluidic Hardware: Design, Engineering, and Fluidic Control . . . . . . . 1216
43.6 Integrated Sample Processing without Online Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1216
43.7 Integrated-Sample Processing with Online Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1221
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1225
43.1 INTRODUCTION
As the genetic bioanalytical community continues to search for ways to improve DNA analysis for
human identification, pathogen detection, and disease diagnosis, affecting more rapid, efficient, and
timely results, the development of new analysis platforms becomes paramount.1,2 Microfluidic sys-
tems have become increasingly attractive analytical tools for applications in many fields. Polymerase
chain reaction (PCR)3−6 and high-resolution DNA separations7−13 are now readily carried out on-
chip, as well as microfluidic purification of DNA or a variety of applications, including those in
the clinical, biohazardous, and forensic sectors.14−19 With successful microchip adaptation of these
processes now commonplace, research focus has shifted toward the integration of these methods and
with other sample processing steps (cell lysis and sorting, DNA quantitation)—the first step toward
creation of a standalone device with full-genetic profiling capabilities. Due to the multistep nature
of the DNA analysis process, careful consideration of solution compatibility and chemistry, fluidic
interfacing and device engineering, as well as computer control and automation must be undertaken
for seamless integration of these sample processing technologies. That is, the firmware, hardware,
and “chemware” must all be carefully considered and optimized to create a multi-component, multi-
functional design that can accommodate the complex process of genetic analysis. As a result, much
attention is now being paid to device design and concept, interfacing diverse and complex chemi-
cal analyses, and computer-controlled automation, and, as a result, multi-component, microfluidic
sample processing is now becoming a reality.
Integrated microfluidic platforms offer unique solutions to many of the problems currently facing
genetic analysis for numerous applications. First, traditional genetic analyses necessitate multi-
ple, time-consuming sample processing steps, often requiring different instrumentation and sample
handling steps for each process, creating ample opportunity for contamination or loss of sample
1207
during transfer.1 Integrated microfluidic systems put forward the opportunity to have automated
sample handling, reducing the user intervention and sample manipulation, as well as instrumenta-
tion required to perform the analysis.1,2 In addition to reducing the time, cost, and handling required
to accomplish these processing steps, a reduction in the amount of sample required for analysis is
inherent to integrated microfluidic systems. By reducing sample and reagent consumption, a more
efficient analysis is achieved. Consequently, a shift to microfluidic technology for genetic analysis
will be impelled by the benefits of faster, more efficient, more automated and integrated sample
processing that produces timely and accurate results with the reduced possibility of contamination
and/or loss of sample.
Although microchips have been used for a variety of applications, this chapter is not a com-
prehensive review of this diverse field; rather, it will focus on the development of the “chemware”
required for microfluidic sample processing for genetic analysis. In particular, the discussion will
center on nucleic acid purification and PCR amplification, with highlights of efforts to develop other
on-chip sample processing techniques, such as sample cleanup, cell sorting and lysis, and DNA
quantitation, both alone and as components of integrated multi-process devices. Particular attention
will be paid to the development of solid-phase purification of DNA and RNA in microfluidic systems
due to the lack of review publications on this subject. The development of microfluidic hardware
will also be addressed; however, the preponderance of this chapter will devoted to the enhancement
and integration of sample processing from a chemware perspective. In addition, this chapter will
detail the development of integrated microfluidic processing devices for multiple genetic applica-
tions, as well as the utilization of microchips to enable fast, reproducible sample handling for human
identification, rapid disease diagnosis, and pathogen detection.
43.2 DNA PURIFICATION

As the microfluidic community looks to develop systems capable of performing a full genetic evalua-
tion, purified DNA or RNA is essential for most of these analyses. PCR-based methods for detection
of disease and infection, as well as for human identification and other specific genetic analyses, have
become the standard for many applications. In order for PCR to work efficiently and effectively,
however, the DNA or RNA template used in the reaction must be free of contaminants that will inhibit
the polymerase or interfere with other reagents necessary for the reaction to proceed. Consequently,
a DNA purification step is vital to ensure that the starting template for PCR is free of contaminants
and suitable for this enzymatic amplification. Thus, the inclusion of a DNA purification step in any
functional genetic analysis micrototal analysis system (µTAS) is imperative. DNA purification (as
opposed to cell isolation) prior to amplification provides many advantages, including, most impor-
tantly, the removal of sample PCR-inhibitors that can include cellular constituents (either proteins
or lipids) in cell lysates, such as hemoglobin,20 and an as yet unidentified factor in eosinophils that
specifically inhibits reverse transcription-PCR.21 In addition to these endogenous contaminants and
inhibitors, in forensic science, the removal of environmental exogenous impurities such as humic
acid from soil22 and pollen,23 is also vital, as the presence of these contaminants will also prevent
successful amplification. Effective removal of both of these external and sample-based inhibitors is
imperative to affect robust and efficient downstream PCR.
In addition, because many benefits associated with performing PCR in microdevices are achieved
only when the volume of the PCR chamber is reduced, the use of a concentrating method for DNA
isolation is not only advantageous, but also necessary, especially for samples that contain a low
number of DNA template starting copies. Solid phase extraction (SPE) methods, such as those
described herein, are not only effective at extracting highly purified DNA for subsequent analysis,
but can also function as a concentrating step prior to PCR amplification. This concentration effect is
not only beneficial for interfacing microfluidic techniques, but also could allow for a more successful
PCR amplification in cases where low starting copy numbers are present, by providing the sample
to be amplified in a smaller volume. Consequently, DNA purification, both as a means to remove
Sample Processing with Integrated Microfluidic Systems 1209
endogenous and exogenous contaminants and as a concentration step, is a critical component of

microfluidic sample processing.
Although purification of DNA from biological samples has been accomplished using a variety of
methods (see Rudi et al.24 for a more comprehensive review), more recently developed conventional
and commercially available techniques for DNA purification exploit silica SPE methods, these
methods reducing the time required for the extraction while maintaining recovery, sample purity,
and integrity. These techniques typically rely on a three-step bind/wash/elute protocol to purify
DNA from interfering proteins and cellular debris. In addition to enabling faster sample preparation,
these protocols are more easily translated into microdevice formats; as a consequence, solid phases
such as silica beads or sol-gels, which can be easily packed into microdevices to create a SPE bed or
column for DNA purification, have become some of the more common phases utilized for microscale
purification, as will be discussed in the next section.14,15,18,19,25−27 In addition, a variety of novel
silica solid phases can be created in microdevices during the fabrication process16,17 that are suitable
for sample purification. All of these phases will be highlighted in detail in the text.
Silica-based methods such as those referenced above typically rely on DNA adsorption to the
silica phase in the presence of a chaotropic agent, followed by elution of proteins with an isopropanol
or ethanol wash, and subsequent DNA elution in water or buffer. DNA, a strong polyelectrolyte
carrying two negative charges per base pair at most pHs, has a large negative surface charge. In
addition, the surface of silica is also negatively charged, due to weakly acidic silanol groups, with an
average pKa ranging from 5 to 7.28 This heterogenous, negatively charged silica surface makes the
net electrostatic repulsion of the fixed charges on the DNA and the silica surfaces strongly disfavor
adsorption at low ionic strength.28
In high ionic strength environments, however, the situation is markedly different. The dissolution
of the chaotropic agent, guanidine hydrochloride (GuHCl) in water, is an endothermic process with
a positive entropy change.28 One molecule of GuHCl will bind, on average, 4.5 molecules of water
(maximum six), which enables the effective dehydration of the DNA molecule and the surface of
the silica at high concentrations.29 Decreasing the water activity in solution through the addition of
a chaotropic salt results in a decrease in solvent-accessible surface area, a loss of water bound, and
an increase in entropy of the solution. In addition, the loss of water at high ionic strength conditions
greatly reduces the electrostatic penalty for placing the negatively charged DNA adjacent to the
negatively charged silica surface. Finally, when the pH of the solution is lowered from 8 to 5 (a pH
∼ 6 is typically utilized in many silica-based purifications), there is an increase surface hydroxyl
groups on the silica phase, thus increasing the ability of the silica to form hydrogen bonds with
the DNA in solution, and a decrease in free hydroxyls that diminishes solution competition with
DNA for these hydrogen-bonding sites on the silica surface.28 These combined effects are the major
driving force behind DNA adsorption to silica: an increase in the entropy of the water molecules
released from the DNA and silica surface and the reduction of the negative potential at the silica
surface by the pH of the solution. In combination, they allow for the binding of nucleic acids to
silica and the subsequent purification of nucleic acids from high ionic strength (6–8 M) chaotropic
solutions.
The first reference of a true microchip-based, silica solid phase extraction exploiting this purifica-
tion technique was published by Christel et al.17 in 1999. In this research, a microdevice containing
silica pillars with high surface-area-to-volume ratios to increase DNAadsorption was designed for the
purification and concentration of DNA for PCR amplification. This represented the first microchip-
based DNA purification accomplished; however, the fabrication of this device relied on a complex
reactive ion etching technique, which limits the potential utility of this format. In addition, capture
efficiencies reported with this device using prepurified stock lambda DNA were only 50%, consid-
erably lower than what would be expected (upward of 80% for other silica-based methods) for a
sample that did not contain proteins or other cellular debris (such as blood) limiting its utility for
low copy number samples where effective retention and release of DNA is imperative. This device,
however, provided the first example of a truly miniaturized DNA purification system and, conse-
quently, set the stage for other microfluidic, silica-based extraction methods. Cady et al.16 employed
(a) (b) (c)
FIGURE 43.1 Examples of solid phases used for DNA purification in microdevices. (a) Scanning electron
micrograph of silica pillars microfabricated into a DNA extraction microdevice. The spacing between pillars
was 10 µm and the depth and height of the pillars was adjusted between 20 and 50 µm. (Figure adapted and
reproduced from Cady, N. C., et al., Biosens. Bioelectron., 2003, 19, 59–66. Copyright 2003. With permission
from Elsevier.) (b) Micrograph (500 times) of a microchannel packed with silica particles immobilized with
sol-gel resulting in a high-surface area silica solid phase for DNA purification. (Figure adapted and reproduced
from Breadmore, M. C., et al., Anal. Chem., 2003, 75, 1880–1886. Copyright 2003. With permission from
American Chemical Society.) (c) Photomicrograph of a microchannel containing immobilized beads for DNA
extraction. (Figure adapted and reproduced from Chung, Y. C., et al., Lab Chip, 2004, 4, 141–147. Copyright
2004. With permission from Royal Society of Chemistry.)
a similar design concept (depicted in Figure 43.1a), fabricating a device containing high surface area
pillars, with the goal of providing a system with higher binding capacity (>200 ng) the capable of
handling large input volumes. It was, however, difficult to effectively evaluate this device design for
sample preparation as no extraction efficiencies were reported in this manuscript. In addition, while
effort toward the development of devices capable of extracting DNA from large volume samples is
certainly needed, any purification method that will be effective in a microfluidic system must also
yield concentrated DNA (in a small elution volume) amenable to downstream processes (e.g., PCR,
where typical reaction volumes range from nanoliters to picoliters). With DNA eluted from the solid
phase in an unusually large volume (250 µL) in this work, the likelihood that such a method could be
effectively integrated with other microchip processes is limited. Devices that contain pillars created
during the fabrication process do, however, represent one potential silica-based design that could
provide reproducible and robust microscale solid-phase purifications of nucleic acids.
In a more direct translation of current macroscale, silica-based, SPE protocols other, microchip-
based purification systems have focused on utilizing a packed silica-bead bed or silica sol-gel matrix
solid phase for purification. This type of extraction was first miniaturized in a capillary format, to
demonstrate the utility of the proposed method in the microscale, by Tian et al.,25 who utilized a 500
nL capillary-based chamber packed with silica particles to establish that PCR-amplifiable DNA (with
80–90% of proteins removed during the load and wash steps) could be obtained from white blood
cells with high extraction efficiencies (70%). This demonstrated the feasibility of incorporating such
silica-based column purification methods into microfabricated devices and the effectiveness of such
methods for the purification of DNA from a wide variety of biological species (white blood cells,
cultured cells, and whole blood).
This same microscale extraction technique was extrapolated to the microchip by Wolfe et al.,18
who evaluated a variety of silica and silica bead/sol-gel matrices for microchip DNA purification.
In this work, a potential problem associated with using silica beads or particles in a microdevice
was highlighted, that is, the tendency of these particles to pack tightly under flow during repetitive
use, thus affecting the reproducibility of flow and the repeatability of the extractions. Recently,
however, it has been suggested that if the devices are single use (which would be typical of devices
designed for many applications), packed silica-bead solid phases are acceptable purification phases
for DNA or RNA extraction, as demonstrated in Easley et al.,30 Bienvenue et al.,31 and Hagan
et al.,32 and highlighted elsewhere in this chapter. Alternatively, the use of sol-gels (liquid col-
loidal suspensions of silica-based materials that can be acid or base catalyzed to gel in place), as
described by Wolfe et al.,18 have been demonstrated as efficient, reusable solid phases. These solu-
tions are simply flowed into microchambers and, by controlling the catalyzed reaction, allowed to
gel and form a porous DNA extraction bed with enough surface area for the binding of DNA, as
demonstrated by Wu et al.,19 who utilized this phase to extract DNA from bacterial (anthrax), viral
(varicella zoster and herpes simplex), and human (blood) sources, with greater than 65% extraction
efficiency demonstrate from blood. In a more recent translation of this work, Wen et al.27 have
employed the use of a photopolymerizable sol-gel monolith to extract DNA in a capillary-based
system, which was further extrapolated to a microchip-based extraction. The photopolymerization
step allows for easy and precise formation of the solid phase within the microdevice, without the
use of retaining weirs or other microfabricated features, making it attractive for use in integrated
systems. In addition, this solid phase has recently been incorporated into a novel two-stage microde-
vice that was developed for DNA extraction from blood—a C18 reverse phase column for protein
capture (Stage 1) in series with a monolithic column for DNA extraction (Stage 2). This device has
a high capacity for DNA in blood (>240 ng) and was found to achieve ∼70% extraction efficiency,
with effective removal of contaminating proteins. Consequently, sol-gels, like column-based solid
phases, have been effectively employed as reproducible phases for DNA purification in microfluidic
systems.
In addition to being used as solid phase for DNA purification, the sol-gel extraction medium can
also be used as a “glue” to immobilize a silica-bead phase,33 thus maintaining a hybrid, reproducible
extraction column from run-to-run. This hybrid sol-gel/bead solid phase (depicted in Figure 43.1b)
was evaluated extensively by Breadmore et al.,15 who optimized flow rates and loading pH to effect
a sample purification in 15 min from bacterial sources (anthrax and salmonella) and whole blood,
with amplifiability of the resultant purified DNA demonstrated. In addition, intra- and interdevice
reproducibility was demonstrated, with extraction efficiencies as high as 79% reported. The utility
of this hybrid solid phase and extraction protocol was also demonstrated for DNA purification from
sperm cells, with a view to the analysis of sexual assault evidence,14 as described in Chapter 37. These
results demonstrated the potential of microchip-based extraction methods for forensic analysis. In
addition, this group has utilized silica beads alone as a purification phase, both for the purification of
DNA30,31 and RNA.32 The purification of RNA from biological samples remains a challenging and
often overlooked sample processing step. The work of Hagan et al.32 represents the first example of
a silica-based purification of RNA from crude samples, those of both clinical (cells lines of alveolar
rhabdomyosarcoma (ARS) tumors) and forensic (semen stains) interest. The only other instance of a
microfluidic purification of RNA present in the research to date was published by the Quake group34
and is discussed later in this section.
Another unique silica-based approach to microscale DNAextraction currently under development
utilizes a serpentine channel design, combined with an immobilized silica-bead solid phase and
fluidic oscillation. This method, developed by Chung et al.,35 relies on silica beads immobilized on
the plasma-oxidized surface of the polymethylmethacrylate (PMMA) channels, instead of a packed-
silica solid phase, as depicted in Figure 43.1c. Following bead immobilization, the solutions required
for DNA binding, purification, and release are flowed back-and-forth through the device. This fluidic
oscillation over the immobilized phase results in marked improvement of recovery and extraction
efficiency over the same extraction methods with free beads. This method represents yet another
variation of silica-based purifications that has been accomplished in microfluidic systems, exploiting
previously optimized chemistries. In summary, the development of macroscale, commercial, silica
SPE protocols has enabled the facile translation of DNA, and now RNA, extraction into microfluidic
systems for a variety of applications.
Although silica-based purifications of DNA and RNA are, by far, the most common microfluidic
methods for isolation of nucleic acids, there have also been other solid phases and methods explored
for DNA purification that deserve mention. Like more recently developed commercial technologies
for nucleic acid purifications, pH-induced methods for extraction have also been miniaturized. Cao
et al.26 have utilized chitosan-coated silica beads that extract DNA at pH 5 and, subsequently,
release the DNA into solution at pH 9. With this purification technique, the PCR inhibitory reagents
typically utilized in silica-based purifications (isopropanol, ethanol, GuHCl, etc.) are not necessary,
thus eliminating a major source of inhibitory contamination. Extraction efficiencies upward of 92%
from biological samples including blood were reported and the purified nucleic acids were suitable for
PCR amplification. Nakagawa et al.36 have also reported a novel purification method for DNA using
a microchannel coated with 3-aminopropyltriethoxysilane (APTES) or 3-[2-(2-aminoethylamino)-
ethylamino]-propyltrimethoxysilane (AEEA) to introduce amine groups on the surface. Relying on
surface electrostatic interactions between amine groups and DNA, this method also relies on a pH
change (from 7.5 to 10.6) to elute DNA from the device and these researchers were able to recover
27–40% of the DNA, which was PCR-amplifiable, from a whole blood sample.
Quake and coworkers34 have utilized an affinity column purification of both DNA and RNA in
microfluidic devices, packing a microfluidic channel with derivatized polymer magnetic beads. Cell
lysates can then be passed over the beads, the DNA or RNA retained on the column, and eluted
in a wash step. With this system and a moderately abundant target (zinc finger OZF), sensitivity
of detection on the order of 2 to 10 cells was obtained. In addition, the device was also utilized
for DNA purification, with PCR-amplifiable DNA successfully isolated from Escherichia coli (E.
coli) culture samples, demonstrating that the device design and method were versatile enough to
accommodate both nucleic acid types. Additionally, Witek et al.37 have utilized a photoactivated
polycarbonate (PPC) device for isolation of DNA from both cell lysates and whole blood. The
PPC chip was fabricated by exposing pristine polycarbonate surfaces to ultraviolet (UV) radiation
that resulted in the formation of surface carboxylate groups through a photo-oxidation reaction.
Subsequently, DNA can be precipitated onto the activated polycarbonate using a PEG/NaCl ethanol
buffer, as seen in Figure 43.2, the proteins rinsed away, and the purified DNA eluted with deionized
water. Extraction efficiencies of this device with prepurified DNA were upward of 85% and the
DNA purified from blood was also suitable for PCR amplification. Finally, in a slightly different
approach, Lee et al.38 as described in more detail in a latter section of this chapter, have developed a
way to use carboxyl-terminated beads to isolate DNA from lysed sample. By selectively binding and
removing the proteins in the sample using these magnetic beads, the resultant purified DNA is left
behind in solution and is PCR-amplifiable. This unique approach lends itself well to single-chamber
purification/amplification sample processing devices, as will be described. Each of the solid-phase
purification methods described here, although not as thoroughly characterized as their silica-based
counterparts, exploits unique chemistries to purify DNA from interfering species and each may be
appropriate in the development of diverse biological genetic analyses.
(a) (b)
FIGURE 43.2 Fluorescence microscopic images of the photoactivated surface of the channel incubated with-
out gDNA (a) and then, incubated with gDNA labeled with YOPRO-1 (b). Note the presence of fluorescently
labeled DNA bound to the channel. (Figure adapted and reproduced from Witek, M. A., et al., Nucleic Acids
Res., 2006, 34, 74. Copyright 2006. With permission from the author.)
43.3 PCR AMPLIFICATION

Polymerase chain reaction, an enzymatic process in which a specific region of DNA is replicated
repeatedly to yield many copies of a particular sequence, was first described by Mullis et al.39 in 1985.
PCR is, theoretically, an exponential amplification of a target DNA sequence, that is, for every cycle
completed, the number of target fragments doubles, such that after 30 cycles, approximately 30 billion
copies of the target region of the DNA template have been created.39 This amplification process
relies on the heating and cooling (thermocycling) of samples using precise thermal control, typically
through three temperatures, to allow for denaturing of the DNA (94◦ C), specific primer annealing
(∼60◦ C), and extension of the PCR fragment (∼72◦ C). Thermocycling for PCR amplification can
be accomplished in a number of ways. Conventional systems use thermal blocks that are heated and
cooled to, in turn, heat and cool the reaction tubes that contain the reaction solutions. This indirect
heating, combined with the large reaction volumes typically required (10–50 µL), limits cycling
rates and results in 1–2 h amplification times for many clinical analyses and over 3 h for standard
forensic amplifications.
Miniaturization of the reaction vessel provides one mechanism for decreasing the overall amplifi-
cation time; by increasing the surface area-to-volume ratio, as accomplished in microfluidic devices,
homogeneous solution temperatures can be achieved more rapidly, thus reducing the time required
for thermocycling (for a more thorough discussion of microfluidic PCR systems, see the following
reviews).40,41 In addition, microfluidic-based miniaturization allows for a number of different heat-
ing methods to be exploited. Many groups3,6,38,42−56 have chosen to utilize direct contact methods
for microchip thermocycling, similar to those used to accomplish conventional thermocycling. With
these methods, the microdevice is either in direct contact with a heating element or a heating element
is fabricated into the system. With noncontact methods, the solution in the chamber is heated directly,
independent of the chamber itself, to perform thermocycling and affect a more rapid analysis, as
pioneered by Landers and coworkers4,5,57 and demonstrated by Ahn and coworkers.58
Contact methods for PCR thermocycling are, by in large, the most common techniques used
for amplification in microfluidic systems. By utilizing microdevices and contact heating, a num-
ber of groups42,44−52 have significantly reduced the volume of sample and reagents required for
amplification (on the nanoliter to picoliter range) and the importance of accomplishing PCR ampli-
fication in reduced volumes cannot be underestimated. In addition to potentially reducing the time
required to perform amplification and the reduction in cost associated with the reduction in reagents
used, the volume of solution generated is amenable to upstream and downstream processing and
analysis components. To date, Quake and coworkers45 have accomplished the lowest volume PCR
amplifications, performing 450 pL RT–PCR on a 72-chamber chip. Although this paper has set the
lower limit of reaction volume, speed of analysis was not the major focus and thermocycling was
performed by placing the entire device in a conventional thermocycler. Consequently, even with a
reduced reaction volume ideally leading to extremely fast thermocycling times, over an hour was
still needed for RT–PCR amplification of a 240 bp fragment of human β-actin RNA. Although not all
of the nanoliter–picoliter amplification devices cited previously accomplished rapid thermocycling,
the reduction in volume permitted by these microprocessing systems will be an important component
of successfully integrated microfluidic systems.
In addition to reducing the sample and reagent volume necessary to perform PCR amplification,
multiple groups5,6,38,53−56,59 have focused on reducing the time required to carry out thermocy-
cling for PCR. Once again, using contact methods for heating in one of the fastest amplifications
accomplished on-chip to-date, Hashimoto et al.6 used a flow-through device consisting of discrete
temperature zones (depicted in Figure 43.3) produced by electrical resistance heaters to amplify a
500 bp region of λ-phage DNA in 1.7 min and a 997 bp fragment of the same target in 3.2 min. It
was determined, however, that the starting copy number of DNA will affect the speed at which PCR
amplification can be performed and with a linear velocity of 10 mm/s, the lowest DNA starting copy
to provide amplified signal in 20 cycles was 1 × 107 copies. Consequently, without refinement, low
copy number samples would be difficult to process with this system. Obeid et al.54 also described
Sample in
Denaturation
Extension
(95°C) (72°C)
Product out
Renaturation Extension
No active heating (72°C)
FIGURE 43.3 Schematic views of the flow-through PCR device layout and model. The device on the right
was used in the experiments described and had 50-mm wide channels separated by 250 mm. The device on
the left had 50-mm wide channels separated by 50 mm and was not used in the research presented, however,
it illustrates the versatility of the flow-through design. (Figure adapted and reproduced from Hashimoto, M.,
et al., Lab Chip, 2004, 4, 638–645. Copyright 2004. With permission from Royal Society of Chemistry.)
a flow-through amplification system that used four heating blocks, three for PCR amplification and
the fourth for reverse transcription, which was utilized to perform 30 cycles of PCR in 6 min. Once
again, however, the lower limit of starting copies required for this device was 6.25 × 106 , making
usage of this device problematic for situations where low template copy numbers are suspected. The
amplification was, however, accomplished in 700 nL, making this device not only capable of rapid
PCR but also lower volume amplification.
Finally, using infrared-mediated heating, a noncontact method has been developed to accomplish
rapid PCR amplification. With this method, near-infrared (IR) radiation is utilized to selectively heat
the reaction solution through excitation of vibrational modes of water molecules.4,5,57 By doing so,
the time required to heat the solution is dramatically decreased, resulting in a concomitant decrease in
amplification time. In addition, by etching away the glass surrounding the PCR chamber, the thermal
mass of the glass can be decreased, contributing to faster cycling rates. This modified glass device and
indirect heating method has been utilized to perform rapid amplification, accomplishing 30 cycles
in less than 5 min.60 This technique has also been utilized to perform RT–PCR in a device with two
IR-heating regions: one for the RT incubation and a second for PCR thermocycling.58 The further
development of this and other rapid thermocycling techniques presented herein represents a consid-
erable step forward in the development of rapid, microfluidic PCR amplification and, combined with
the reduced volume techniques described previously, provides a significant advancement of current
technology. In addition, by executing low-volume, rapid thermocycling techniques, microfluidic
PCR amplification will integrate more seamlessly with other microscale sample processing and
genetic analysis steps, helping to facilitate the ultimate goal of an integrated total analysis system.
43.4 CELL SORTING, CELL LYSIS, AND DNA QUANTITATION

The selective isolation of different cell populations can be an important step in sample processing
for a number of different genetic analyse. As described in Chapter 37, separation of sperm cells and
vaginal epithelial cells is an important step for the effective analysis of sexual assault evidence and
clinically, cell-sorting techniques have typically involved the isolation of specific cell types (i.e.,
cancer cells) from mixtures of other interfering species. A variety of methods can be exploited to
separate different cell types, including dielectrophoresis, separation on the basis of the different
physical and chemical properties of cell types (see Chapter 37), and acoustic differential separations.
A number of groups have utilized microscale dielectrophoresis, the manipulation of particles in
nonuniform electric fields, for separating different cell types (for comprehensive reviews on detailing
the use of dielectrophoresis for cell sorting, see References 61 and 62). Typically, for these methods,
electrodes are fabricated into the microdevice and differences in the dielectric makeup of the cells are
exploited to allow for separation when these particles are placed in an electric field. Consequently,
dielectrophoresis has become a noninvasive way to isolate different cell populations that is easily
miniaturized in microfluidic systems.61,62
In addition, as described in detail in Chapter 37, cell populations that have inherent physical
differences (such as sperm and vaginal epithelial cells) can be separated by simply exploiting these
innate dissimilarities.63 Also exploiting the size variation of different cell types to promote their
facile separation, Yuen et al.64 have utilized a weir-filter design to effectively isolate white blood
cells (12–15 µm) from red blood cells (7–8 µm), as will be described later in this chapter. In addition,
Quake and coworkers65,66 have developed a pressure-driven cell sorter that can selectively isolate
fluorescently-labeled cell populations. Austin and coworkers67 have translated magnetic-activated
cell sorting (MACS) to microscale using antibody-coated magnetic beads in a high magnetic field;
however, to date there has been no published data showing the successful sorting of cells with this
method. The same group,68,69 however, has utilized deterministic lateral displacement to isolate red
and white blood cells, and as well as platelets, from whole blood and for total cell removal from
whole blood for plasma preparation. Finally, the development of microfluidic acoustic technology
has enabled the separation of cells for a number of applications70−74 including its employment for
differential extraction (see Chapter 37). Selective cell sorting, although not the focus of this chapter,
is a technique that lends itself well to microfluidics and provides for efficient sample enrichment
prior to other sample processing techniques.
The lysis of cells on-chip (either after a cell-sorting step or from mixed sample sources) will,
undoubtedly, be an important step in the development of sample processing microfluidic devices
for total sample handling of crude biofluids. There are many examples in the literature that describe
the lysis of cells in microfluidic systems and that can be easily divided into two categories: (1)
those that utilize a chemical-based lysis and (2) those that employ a physical breakdown of the
cell. Chemical cell lysis has been demonstrated using a variety of buffers and reagents14,75−78 and
typically involves flowing the particular lysing reagent on-chip to mix with the cells of interest.
Physical breakdown of the cell can be caused by a number of different methods that disrupt the
cellular structure and release the nucleic acids into solution. Electroporation,79,80 electric field,81−84
thermal lysis,85−88 sonication,89 laser/vibrationally induced,38 and mechanical lysis with beads90 or
barbs91 are all methods that have been utilized to rupture cells on chip. Successful release of cellular
constituents through effective, complete, and rapid cell lysis is essential for downstream sample
processing steps, such as SPE and PCR. Consequently, although often overlooked, microscale cell
lysis is an indispensable component of any microfluidic sample processing systems.
Finally, on-chip DNA quantitation will undoubtedly be an important feature of microfluidic sam-
ple processing for a number of different applications where the amount of input DNA for downstream
genetic analysis is crucial. As discussed more extensively in Chapter 37, quantitative PCR (qPCR)
methods in microfluidic systems have not been demonstrated to date. However, there are a number
of publications describing real-time PCR in microdevices and it is likely only a matter of time before
these methods will be extended to include qPCR (for a more detailed discussion of real-time PCR,
see Chapter 37).
In addition to on-chip qPCR, another approach being developed to accomplish quantitation
involves a more direct translation of current macroscale approaches to the microscale. Using a
fluorescence-based assay, a valved device is used to generate a calibration curve, which is used to
quantitate an unknown sample with on-chip mixing of the DNA with an intercalating dye.92 This
method was designed with the goal of integrating this technology with microscale solid phase purifica-
tion techniques, for post-purification/pre-PCR quantitation, making it a potentially facile inclusion in
sample preparatory genetic analysis chips. Although less focus has been paid to developing microflu-
idic techniques for on-chip determination of DNA concentration, this sample processing step is no
less a necessity than the other more routinely miniaturized techniques and will be an important
inclusion in the development of genetic µTAS for a number of applications.
The discussion of on-chip sample processing presented here is by no means exhaustive, rather it
is meant to highlight the progress made in miniaturizing the main genetic processing steps required
to affect a total analysis. The methods described represent some of the first and most recent examples
of microchip-based sample preparation for downstream sample processing and a major step toward
the development of microfluidic systems capable of accepting crude samples for DNA analysis.
As advancement of these technologies has continued, the focus has shifted from basic method
development of these and other individual processing techniques to the integration of several sample
treatment steps in multi-process devices.
43.5 INTEGRATION OF MICROFLUIDIC HARDWARE: DESIGN,

ENGINEERING, AND FLUIDIC CONTROL
Although this chapter highlights sample processing in microfluidic systems in both stand-alone, sin-
gle process devices and integrated multi-process systems (i.e., the development of the “chemware”
to support miniaturized genetic analysis), it is of critical importance to note that the successful
integration of such diverse chemistries would not be functionally possible without the concomi-
tant advancement of microfluidic hardware. Advances at the device level have included innovative
engineering to allow for elegant chip fluidic control through valving48,93−98 and novel device
designs,30,34,99−102 as well as fluidic modeling of these systems where the surface-area-to-volume
ratios are radically different from the macroworld (see Chapter 40 by Easley). The complexity and
functionality of integrated microfluidic devices, as depicted in Figure 43.4, has made rapid sample
processing and analytics for genetic typing possible. Without these advancements, the integrated
“chemware” that will be detailed in the remainder of this chapter would simply not be feasible and
consequently, it is worth nothing these achievements and recognizing that it is these innovations that
have permitted the functional microfluidic integration of genetic analysis.
43.6 INTEGRATED SAMPLE PROCESSING WITHOUT ONLINE

DETECTION
Recent literature has provided descriptions of several devices that have integrated sample processing
steps, such as on-chip cell lysis, cell selection, DNA purification, and PCR amplification. In one of
FIGURE 43.4 (Continued) (c) An integrated microfluidic device designed by the Burns group. (Top)
Schematic representation of the device, containing multiple different liquid entry channels (“L,” sample, PCR
reagents and RD reagents), several metering channels, drop mixing intersections, a sealed PCR chamber, an
open RD chamber, individually controlled valves, and an electrophoresis channel. (Bottom) Photograph of an
assembled device (1.5 cm by 1.6 cm), which can control discrete liquid drops that are 100–240 nL, with fluidic
channel dimensions of 200–600-µm wide and 50-µm deep. (Figure adapted and reproduced from Pal, R., et al.,
Lab Chip, 2005, 5, 1024–1032. Copyright 2005. With permission from Royal Society of Chemistry.)
(a)
TC reactor
Cath-
Capture outlet RTD contact (c)
pads
ode Microvalve RD reagents PCR Test
e RD chamber reagents sample
Capture chamber
Waste
Capture inlet
d
b Electrophoresis PCR chamber
section
c
Sample Via hole
RTD
Pneumatic Pneumatic
lines ports
Capture-Inject
Separation
channel
Taper turn
Anode
(b)
Purge input
Sample out Sample in
Substrate
out Substrate
in
Sample Sample
collection colleciton
1 4
Multiplexor
control
Barrier 2 Sandwich Barrier 4
barrier Barrier 3
Mixer
Barrier 1
barrier
Sample
Multiplexor collection 5 mm
control 2 and 3
FIGURE 43.4 (a) The bioprocessor developed by the Mathies and coworkers. A photograph of the microde-
vice, showing one of two complete nucleic acid processing systems is shown (left), with sequencing reagent,
capture gel, separation gel, and pneumatic channels. (Scale bar, 5 mm). The pictures on the right show multiple
different components of the device, including (from top to bottom): a 250-nL thermal cycling reactor with RTDs
(Scale bar, 1 mm), a 5-nL displacement volume microvalve, a 500-µm-diameter via hole, a capture chamber
and cross injector, and a 65-µm-wide tapered turn (Scale bars, 300 µm). (Figure adapted and reproduced from
Blazej, R. G., et al., Proc. Natl. Acad. Sci. USA, 2006, 103, 7240–7245. Copyright 2006. With permission from
National Academy of Sciences.) (b) An optical micrograph of a nanofluidic system developed by the Quake
and coworkers that can be used for parallelized high-throughput screening of fluorescence-based single-cell
assays. The various inputs have been loaded with food dyes to show the channels and subelements of the fluidic
logic. This chip has 2056 valves, which are used to manipulate 256 compartments containing bacterial cells
expressing an enzyme of interest (or a library of mutants of that enzyme) that are combined on a pairwise
basis with 256 other compartments containing a fluorogenic substrate used to assay for a desired activity. Cells
that display a particularly interesting activity can be selected and recovered from the chip using valve-based
addressing of the compartments. (Figure adapted and reproduced from Hong, J. W., et al., Nat. Biotechnol.,
2003, 10, 1179–1183. Copyright 2003. With permission from Nature Publishing Group.) (Caption continued
opposite on page 1216.)
(a) (b) (c)

Pre-load Post-load
Empty Packed Empty Packed Packed sperm cells

channel beads channel beads
(d) (e) (f)
Initial DTT lysis t = 1 min t = 3 min
FIGURE 43.5 On-chip lysis of sperm cells integrated with microchip-based DNA extraction. All panels show
a section of the microchip extraction channel (∼1 mm) at the front edge of the silica bed by either light (Panels a
and b) or fluorescence (Panels c through f) microscopy. No inherent fluorescence was seen in the channel when
irradiated with 480–550 nm light (b), before cell loading. DEAD Red™-labeled sperm cells were flowed into
the channel and packed up against the beads, while the silica bed front was visualized (Panel c). After flow of
the dual-lysis/loading buffer was subsequently initiated (Panel d), the fluorescence decreased over the next few
minutes in (Panels e and f) and was essentially nonexistent after 5 min (data not shown). (Figure adapted and
reproduced from Bienvenue, J. M., et al., J. Forensic Sci., 2006, 51, 266–273. Copyright 2006. With permission
from Blackwell Publishing.)
the more simplistic integrations of two sample processing steps, cell lysis and nucleic acid extraction
were demonstrated in the same microdevice by Bienvenue et al.,14 who developed a method for
loading whole sperm cells into a device containing a hybrid silica-bead/sol-gel extraction column.
The cells packed against the leading edge of the silica-packed column, penetrating only slightly into
the solid phase, as depicted in Figure 43.5. Following the packing step (Figure 43.5a through c),
a dual cell lysis/DNA loading buffer was flowed over the packed cells and column (Figure 43.5d
through f), lysing them and releasing their contents into solution for subsequent binding to the silica
bed. Purification was then carried out as would typically be accomplished for silica-based methods,
with an isopropanol wash to remove interfering species, and finally elution with water. The resultant
eluate was subjected to forensic short tandem repeat (STR) amplification and the genetic profile
compared to that obtained using standard conventional silica methods. The microchip-obtained
profile was consistent with the profile generated using the kit-based purification, demonstrating
that not only was the dual-process device as least as functional as the open-system conventional
methods, but also that this closed microfluidic system have potential as a forensic sample processing
tool, where prevention of contamination is of paramount importance.
In 2001, Yuen et al.64 published a report describing an integrated system that performed cell
isolation and direct PCR amplification. Using weir-type filter constructed of silicon that spanned the
microfluidic channel to perform the isolation, blood (<3 µL) was flowed into the channel. Owing
to their larger size (12–15 µm in diameter as opposed to 7–8 µm in red blood cells), the white
blood cells were trapped between the top of the weir and the bottom of the microdevice cover
plate. Held there, the PCR master mix was flowed into the chamber and the device subjected to
thermocycling (and thermal lysis of the cells) using a copper block heater fabricated into a plexiglass
holder. With this system, successful cell isolation and PCR amplification of a 236 bp fragment of
the human coagulation Factor V was accomplished; however, the authors note that the presence
of too many white blood cells resulted in significant inhibition of amplification, most likely due to
the inhibitory factor in eosinophils.21 These results suggested that although the cell separation and
direct PCR method presented was a functional means to obtain sequence-specific genetic results,
the inclusion of a true DNA purification step will still be essential for further development of such a
technique.
In addition to this work, Hong et al.34 described a nanoliter-scale processor that performs cell
isolation, cell lysis, and nucleic acid purification on a single platform. In their work, a device was
designed that contained 100 µm wide fluidic channels, in which three holding chambers (lysis
buffer, cell, and bead) were created using valve isolation. The various reagents required for each
different step of the analysis were flowed into the chip through different inlets, with cells mobilized
in the device and retained in cell holding chambers. Following cell retention, the valve between the
cell chamber and the lysis buffer chamber was opened, allowing lysis buffer to mix with and lyse
the cells. The resultant corresponding cell lysate was then flowed over an on-chip mRNA affinity
column (as described previously in the extraction section of this chapter), during which the desired
mRNA was retained on the beads. The beads were then removed from the device and used directly
in the off-chip RT–PCR amplification. This device and method were utilized to isolate mRNA from
samples containing single cells, with two different targets (β-actin, a high-abundance transcript and
zinc finger OZF, a moderate-abundance transcript). This report also described the development of a
DNA purification chip, a device capable of cell lysis and affinity purification. Unique to this device
was the use of novel channel designs and flow patterns, as well as the use of rotary mixers that
provide comprehensive on-chip mixing of cells and lysate—a process that would require consume
hours to complete relying on diffusive mixing. Once again, the beads were removed from the
device subsequent to DNA capture, and the amplification of the gene encoding prelipin peptidase-
dependent protein (ppdD) was accomplished off-chip. Importantly, these two devices demonstrated
the feasibility of incorporating multiple sample processing steps in microfluidic systems for the
preparation of both DNA and RNA for genetic interrogation, successfully integrating cell lysis and
nucleic acid purification in devices with unique architectures and precision fluidic control elements.
Collectively, these reports highlight the importance of integrating multiple sample processing
steps into single multi-purpose microdevices. By incorporating on-chip cell lysis or cell sorting
with either purification or amplification, more efficient processing was accomplished in closed,
microfluidic systems, reducing not only the opportunity for contamination and the degree of user
intervention required, but also the volume of sample and reagents consumed. Although each of these
designs requires further refinement and inclusion of additional sample processing and/or analytical
steps before they can be considered complete, through these examples, great progress has been made
toward the development of a lab-on-a-chip, µTAS device capable of total sample preparation and
analysis.
The integration of DNA purification and PCR amplification remains a formidable and much-
overlooked undertaking. As argued earlier, purification of DNA before PCR amplification is
imperative for a variety of reasons—yet there are only a few reports of integrated DNA purifi-
cation and amplification devices in the literature to-date.31,38,103,104 The inherent incompatibility
of the reagents (chaotropic salts, organics, etc.) necessary to perform many types of nucleic acid
purifications (particularly silica-based nucleic acid isolation) with the polymerase chain reaction,
make fluidic coupling of these two dramatically and inherently different processes a challenging
task. Legendre et al.103 circumvent these problems with a valveless chip design that used the previ-
ously described silica-bead hybrid solid phase and combined with a conventional block thermocycler
for PCR . Using a dual syringe system, reagents from the load and wash step of this silica-based
purification were simply passed through the PCR domain and out of the device. DNA could then be
eluted from the SPE domain directly into the PCR chamber, while master mix containing the reagents
necessary for PCR was flowed in through a side channel for subsequent amplification, removing the
(a)
(d)
15
12 Primer
9 Anthrax
AU
6 Primer- product
dimer
3
0
2 3 4 5
Time (min)
(c) (b)
FIGURE 43.6 Microfluidic mixing in the integrated purification/amplification device. (a) Blue dye was flowed
through the side channel, while yellow food dye was flowed through the SPE bed. Note the laminar flow. (b)After
entrance into the PCR chamber, blue and yellow dyes were only mixed in the center of the channel, due to
diffusion. (c) Following the initial denaturation step of thermocycling, green color was observed in the PCR
chamber indicating adequate mixing of the two streams. (d) Electropherogram showing the results of application
of the integrated DNA purification/amplification device to biowarfare agent detection. Anthrax spores on a nasal
swab were eluted in lysis buffer and loaded onto the device for purification of DNA, integrated with IR-mediated
PCR. The capillary electropherogram depicts successful amplification of a 211 bp product peak of B. anthracis
and this total sample processing was accomplished in the total analysis time less than 23 min. (Figure adapted and
reproduced from Legendre, L. A., et al., Anal. Chem., 2006, 78, 1444–1451. Copyright 2006. With permission
from American Chemical Society.)
offending contaminating solvents from the device. Co-mobilization of the reagents delivered through
two separate syringes allowed for 1:1 mixing of the eluting DNA and the PCR reagents in the PCR
chamber (as depicted in Figure 43.6a through c), where the mixture was could be thermocycled for
DNA amplification using a conventional thermocycler. This dual-process microdevice was utilized
for the successful purification of DNA from a diversity of biofluids, including buccal cells, blood, and
semen followed by target-specific amplification.103 In addition, this microdevice was also utilized to
accomplish the microfluidically integrated purification and targeted DNA amplification of sequences
from Bacillus anthracis (B. anthracis) isolated from a mock nasal swab with a total analytical time of
less than 23 min,103 as depicted in Figure 43.6d, using the IR-mediated, noncontact thermocycling
method described earlier in this chapter. This work represented the initial efforts to integrate DNA
purification and PCR amplification, with a view toward the development of a sample processing
device capable of handling of a wide array of input samples for a wide variety of applications.
A similar device design was also utilized to explore the integration of DNA purification and
multiplexed amplifications for forensic analysis.31 In this work, device design and capacity were
further explored and methodologies to incorporate the commercially-available reagents commonly
utilized for forensic analyses were investigated. The challenges associated with microfluidic inte-
gration for forensic analysis are complicated by the need to develop systems that utilize established,
validated, and commercially available reagents/protocols, as described in Chapter 37. As a conse-
quence, microfluidic systems must be developed that interface seamlessly with existing, validated,
and court-tested reagents, protocols, and technology, providing a substantial decrease in reagents
and sample consumed by the analysis, as well as cost per sample, and significant improvements
in time to result and prevention of contamination. These changes will provide the motivation to
move beyond previously established techniques and promote the inclusion of microfluidic systems
in general casework analysis. The simple device described earlier was used to demonstrate the feasi-
bility of microfluidic sample processing in a closed system, utilizing common benchtop equipment
and commercially-available, kit-based reagents. It was utilized to purify DNA from semen (a com-
monly encountered biofluid in sexual assaults) and perform integrated, multiplexed amplification
using the standardized kit-based reagents employed for conventional forensic genetic interrogation,
as highlighted in Chapter 37.
Each of the devices described in this section has functionally integrated two or more sample
processing steps. Using novel chip designs, multiple sample processing steps (including cell sorting,
cell lysis, DNA purification, and PCR amplification) could be accomplished in an integrated system,
providing more automated and more efficient sample handling. The basic integration of these steps
has served as a catalyst to the development of fully integrated systems with both sample processing
and analytical capabilities. The functional development of these integrated chemware systems has
and will continue to impel, the evolution of µTAS for genetic analysis.
43.7 INTEGRATED-SAMPLE PROCESSING WITH ONLINE

DETECTION
It is important to distinguish integrated-sample processing devices from devices that have incorpo-
rated both sample processing and an endpoint analysis step. All of the devices highlighted in the
previous section are capable of sample manipulation, but do not have an online detection step in
their microfluidic architecture. The following section will detail µTAS that are capable of both sam-
ple handling and analytical detection for genetic analysis. Building in complexity, these devices
contain the hardware and “chemware” necessary to perform integrated total analysis, providing a
genotypic “answer” following sample handling, and will be discussed in detail in the next section.
In addition, although the focus of this chapter is on the development of sample processing, both
in stand-alone microfluidic devices and integrated processing systems, it is important to note that
the integration of PCR amplification and microchip electrophoresis has been accomplished by a
number of groups.42,99,100,105−116 Integration of PCR with electrophoretic analysis is an important
and essential step toward the development of µTAS devices for genetic interrogation; however, we
have chosen to focus this chapter on the integration of multiple sample processing steps, leaving
the integration of sample processing and analytical steps for future discussion, except in the rare
cases where an analytical step was incorporated with multi-process sample handling. Consequently,
detailed explanations of integrated PCR-(ME) devices will not be explored.
In one of the earliest examples of a device capable of performing multiple, integrated sample
processing steps followed by analysis, Waters et al.88 described a device in which that represents
one of the first examples of integrated sample processing combined with an analytical step to affect
a more complete analysis in a closed microfluidic system. In this particular device, thermal cell
lysis, followed by direct PCR amplification and electrophoretic sizing was accomplished. With this
method, whole E. coli cells were loaded into the device along with PCR master mix. Following
loading, the entire device was placed into a conventional benchtop thermocycler and subjected to a
typical PCR amplification. During the initial 4-min heating step (94◦ C), the cells were lysed, releasing
their DNA into solution, which could then be amplified during thermocycling, ultimately followed
by electrophoretic sizing using microchip electrophoresis. Although this method was successfully
employed to perform integrated cell lysis, amplification, and electrophoretic analysis, without the
inclusion of a DNA purification step in this method, this device would be limited to use with samples
that do not contain significant concentrations of PCR inhibitors. It does, however, represent one of
the first examples of integrated sample processing, combined with an analytical step, to affect a more
complete analysis in a closed microfluidic system.
As a more complex example of integrated sample processing, Lagally et al.117 developed a
device capable of dielectrophoretic cell sorting and on-chip cell lysis, followed by sequence-specific
(a) (b)
FIGURE 43.7 On-chip trapping of cells was accomplished using the integrated microfluidic system described
by Lagally et al. (a) An epifluorescence image taken after 10 min of cell flow over the interdigitated microelec-
trodes without DEP voltage applied. Nonspecific binding does not occur and no cells are visible. (b) The same
microelectrodes after 10 min of flow with 7 V applied at 1 kHz, as pictured, cells are captured by DEP onto the
electrodes where they can be further processed. (Figure adapted and reproduced from Lagally, E. T., et al., Lab
Chip, 2005, 5, 1053–1058. Copyright 2005. With permission from Royal Society of Chemistry.)
hybridization. This device was used to affect a 160-fold increase in cell concentration by trapping
and concentrating the cells in a continuous flowing stream (100 µL/h for 10 min) using positive
dielectrophoresis in a 100 nL plug and an on-chip valving system. Following cell trapping (as
depicted in Figure 43.7), the lysis buffer and molecular beacon are added to the cells and detection
of E. coli MC1061 cells was accomplished via the sequence-specific hybridization of the beacon to
rRNA of the bacteria with fluorescence measured by laser scanning confocal microscopy. By using
high concentrations of guanidine thiocyanate (4 M), not only was cell lysis promoted, but also a more
rapid hybridization of target DNA to the molecular beacon probe was achieved (immediate binding
versus upward of 40 min when placed in TE buffer), due to a decrease in the melting temperature of
the nucleic acid/molecular beacon probe hybrids. Detection of as few as 25 cells in less than 30 min
was accomplished in this manner, making this device a rapid total analysis tool capable of taking in
crude sample, performing integrated sample processing, followed by an analytical/detection step to
produce sequence-specific results.
Yeung et al.118 have described a device for integrated thermal cell lysis, specific target DNA
isolation, followed by PCR amplification and hybridization detection. In this work, intact cell were
broken down by thermal lysis (at 90◦ C) in the presence of biotinylated genome capture probes.
Following cell lysis, the temperature of the chamber was lowered to 50◦ C, where the probes bound
to their target sequences and, following the addition of avidin-coated magnetic particles, specific
genome sequences were isolated from cellular debris and interfering proteins by washing away
the unbound species. The captured DNA was then be amplified using asymmetric PCR and the
resultant amplicon hybridized to detection electrodes, labeled with gold nanoparticles, and detected
by electrocatalytic silver deposition and electrochemical silver dissolution. With this method, two
different bacterial species (E. coli and Bacillus subtilis (B. subtilis) could be detected in a mixed
sample, however, the device was only evaluated with cultured cells, so its performance with more
complex biofluids is unknown at this time.
In another example of integrated sample processing, Lee et al.38 accomplished cell lysis, DNA
purification, and PCR amplification for pathogen detection using a single-chamber device. In this
work, the authors developed a Laser-Irradiated Magnetic Bead System (LIMBS) for cell lysis and
DNA purification from pathogens. With this method, pathogen and carboxyl-terminated magnetic
beads were loaded into the microchip and placed into a chip guide module. A high-power laser beam
(808 nm) was then applied to the microdevice while it was simultaneously vibrated to facilitate lysis
of the cells. In addition to promoting cell lysis, when the laser was fired, the proteins were adsorbed
to the surface of the magnetic beads, while the DNA remained free in solution. Following removal
of the beads, the resultant DNA-containing solution was suitable for real-time PCR amplification.
Although extraction efficiencies with this method were not reported, this single-chamber protocol
was utilized to successfully lyse E. coli and Gram-positive bacterial cells, as well as hepatitis B virus
mixed with human serum, with a higher efficiency of DNA release using the LIMBS method than
that of other conventional cell lysis methods (e.g., boiling). In addition, successful lysis, purification,
and real-time PCR amplification was performed in this single-chamber device using cultured E. coli
cells, demonstrating the potential of this method for pathogen detection.
To date, only a few examples can be found in the literature describing a singular microfluidic chip
capable of accepting crude sample lysate (from complex biofluids), purifying and extracting the DNA,
accomplishing a target-specific amplification, and then performing online genetic analysis. In 2004,
Liu et al.85 described an integrated microfluidic system that performed multiple sample processing
steps and integrated detection of target amplicon. In this system, sample and immunomagnetic capture
beads were loaded into a chamber for incubation. Following successful isolation of the target cells, a
washing step was performed, followed by loading of the chamber with the necessary PCR reagents.
Subsequently, thermal lysis and PCR thermocycling was carried out, followed by incubation with and
acoustic mixing of the amplicon for a hybridization reaction. Following immobilization of the DNA,
ferrocene-labeled signaling probes hybridized with the target DNA, were bound to the immobilized
probes, and could be detected on-chip. The isolation and detection of E. coli from a mock E.
coli/rabbit blood mixture, as well as single-nucleotide polymorphism analysis directly from diluted
blood were demonstrated. This device is the earliest example of a fully functional microanalytical
system capable of accepting crude sample and producing a genotypic readout, establishing that
multifunctional, microfluidic devices could function as viable automated genetic analysis systems.
More recently, Cady et al.104 have described a method for integrating sample processing with
detection, by developing a device capable of silica-based purification using the silica-coated pillar
design described earlier in this chapter, integrated with online, real-time PCR. With this device,
a silica-based purification was achieved for the pathogenic bacterium Listeria monocytogenes—a
rod-shaped bacterium that is the causative agent in listeriosis, a serious infection caused by eating
contaminated food—followed by integrated real-time amplification using SYBR Green, in an average
of 45 min. In addition, this integrated system could detect as few as 104–107 cells, making this a
sensitive screening technique that combines both sample processing and analytical analysis in a
simple two-component device.
Finally, late in 2006, Easley et al.30 described a valved microdevice with discreet functional
domains: a silica-bead-based purification chamber, fluidically integrated to a PCR chamber suitable
for IR-mediated thermocycling and connected to an electrophoretic domain, where separation and
detection of amplicon was accomplished. This microgenetic analysis (MGA) device was the first of its
kind that was capable of accepting sample as crude as whole blood and performing complete sample
processing and genotypic analysis. As depicted in Figure 43.8, the device was evaluated with blood
drawn from a mouse infected with B. anthracis before prior to the onset of symptoms. Purification
of nucleic acids from less than 1 µL of this sample was accomplished in less than 10 min, followed
by target-specific amplification in 11 min. Utilizing the sophisticated valve system, amplicon could
then be repeatedly co-injected with sizing standard (both via pressure) for separation, detection, and
accurate sizing using microchip electrophoresis in less than 3 min, making the total time required
for this analysis less than 24 min.
In addition, the versatility of the same microdevice and method was demonstrated by application
to the complete genetic analysis of 1 µL of nasal aspirate from a patient symptomatic of whooping
cough. The presence of Bacillus pertussis (B. pertussis) was confirmed by the amplification of a
(a) (c)
1/ Time
MR PR
1.6 Log bp 3.0
PCR
BR TR 212 bp
SW
SPE
22 23
EW SA
ME
(b)
0.10 100
SPE
SI
DNA (ng)/intensity (a.u.)

FA
PCR
Temperature (°C)
CE
0.05 75
BW 10 mm 0.00 50
0 4 8 12 16 20 24 28 32
Total analysis time (min)
FIGURE 43.8 The integrated µTAS described by Easley et al. used for the integrated-sample processing and
detection of B. anthracis in murine blood. (a) The IgA device with dyes placed in the channels for visualization.
Domains for DNA extraction, PCR amplification, injection, and separation are connected through a network
of channels and vias. SPE reservoirs are labeled for sample inlet (SI), sidearm (SA), and extraction waste
(EW). Injection reservoirs are labeled for PCR reservoir (PR), marker reservoir (MR), and sample waste (SW).
Electrophoresis reservoirs are labeled for buffer reservoir (BR) and buffer waste (BW). Additional domains
patterned onto the device include the temperature reference (TR) chamber and fluorescence alignment (FA)
channel. The flow control region is outlined by a dashed box. Device dimensions are 30 × 63.5 mm, with a
total solution volume less than 10 mL. (a) Detector responses during all three stages of sample processing and
analysis are portrayed in terms of total analysis time. The SPE trace (green) was taken from an offline DNA
extraction of the same murine sample and is representative of the total DNA concentration observed in a typical
extraction. The temperature and fluorescence intensity represent online data, with a total analysis time of less
than 24 min. Three sequential injections and separations were carried out to ensure the presence of amplified
product. (c) Magnified view of the first separation shown in (b), with the product peak marked and sized. The
second and third runs are overlaid with the time axis cropped. The inset plot shows the sizing curve of inverse
migration time versus log (base pairs) with both the sizing standard peaks (open diamonds) and product (red
square) plotted for all three runs shown in (a) (error bars included). From this data, the product was sized as
212 ± 3 bp. (Figure adapted and reproduced from Easley, C. J., et al., Proc. Natl. Acad. Sci. USA, 2006, 103,
19272–19277. Copyright 2006. With permission from National Academy of Sciences.)
fragment of the IS481 repeated insertion sequence, as confirmed by sequencing of the amplicon.
With a total analysis time again of less than 24 min, this application further demonstrated the
versatility and robustness of the method and system, thus highlighting the potential of this type of
total analysis device for application to rapid, point-of-care testing for a variety of applications. Like
the work presented by Liu et al.,85 this device integrated multiple sample processing steps in a single
microfluidic device. However, improving upon this earlier work, Easley et al. were able to establish
a rapid analysis (<24 min versus 3.5 h) and utilize two crude input samples (infected murine blood
from asymptomatic mice and human nasal aspirate versus a mock rabbit blood/bacteria mixture).
This µTAS and the other integrated devices described here represent the first in what will surely be
a long line of increasingly more sophisticated integrated sample processing and analytical devices
for genetic interrogation and total genetic profiling.

The path to creating integrated microdevices for bioanalysis is not only riddled with cutting-edge
scientific challenges, it is ripe with cutting-edge engineering challenges as well. The µTAS field
today has reached a milestone, solving or circumventing many of these critical challenges, both
with hardware and “chemware,” to generate the integrated multi-functional microdevices described
herein. This has required clever design and novel approaches to integrating the necessary the hard-
ware, firmware, and “chemware” into the microdevice in a functional manner. Compartmentalization
for isolation of different chemistries in different functional domains, as described in the examples
presented herein represents a breakthrough in terms of functional bandwidth for microfluidic devices.
In addition, the development of precision fluidic control components and novel device designs has
enabled the advancement of integrated chemistries; providing the necessary hardware to support the
“chemware” as an essential component of integrated microfluidics.
The successful development of integrated microfluidic systems for genetic analysis represents
an important approach to the improvement of analytical technology for many applications (e.g.,
forensics, clinical analysis). The advancement of reproducible, optimized methods for microchip
DNA purification, PCR amplification, and other sample pretreatment techniques has enabled the
subsequent development of integrated sample processing systems; without consistent, reliable stan-
dalone microfluidic processing, successful and consistent integrated analyses are an impossibility.
The work presented here describes the development of these sample processing devices, highlights
the flexibility of microfluidic systems for a accomplishing a wide variety of genetic analysis appli-
cations, and the underscores importance of the inclusion of pretreatment steps in the development
of µTAS.
Future work with integrated sample processing microdevices will likely focus on the development
of systems capable of total sample manipulation, with the versatility to handle multiple input samples.
In addition, new generation devices will likely be able to perform sample-dependent processing, that
is, the device will have many different fluidic handling steps incorporated, and the steps applied in the
processing can be tailored to the type of input sample. Automation of these devices will be essential
and the integration of these sample processing techniques with microfluidic analytical technologies
will be imperative. The continued successful development of a genetic µTAS will hinge on sustained
effort directed toward the advancement of enhanced microfluidic sample processing and integrated
handling systems.
REFERENCES
1. Butler, J. M. Forensic DNA Typing, Biology and Technology Behind STR Markers, 2nd ed., Academic
Press, San Diego, CA, 2001.
2. Rathore, A., and Guttman, A., Eds. Electrokinetic Phenomena, Marcel Dekker, Inc., New York, 2004.
3. Kopp, M. U., Mello, A. J., and Manz, A. Science, 1998, 280, 1046–1048.
4. Huhmer, A. F., and Landers, J. P. Anal. Chem., 2000, 72, 5507–5512.
5. Giordano, B. C., Ferrance, J., Swedberg, S., Huhmer, A. F., and Landers, J. P. Anal. Biochem., 2001,
291, 124–132.
6. Hashimoto, M., Chen, P. C., Mitchell, M. W., Nikitopoulos, D. E., Soper, S. A., and Murphy, M. C.
Lab Chip, 2004, 4, 638–645.
7. Gawron, A. J., Martin, R. S., and Lunte, S. M. Eur. J. Pharm. Sci., 2001, 14, 1–12.
8. Jin, L. J., Ferrance, J., and Landers, J. P. Biotechniques, 2001, 31, 1332–1335, 1338–1340, 1342,
passim.
9. Zhang, L., Dang, F., and Baba, Y. J. Pharm. Biomed. Anal., 2003, 30, 1645–1654.
10. Khandurina, J., and Guttman, A. Curr. Opin. Chem. Biol., 2003, 7, 595–602.
11. Chen, L., and Ren, J. Comb. Chem. High Throughput Screen., 2004, 7, 29–43.
12. Dolnik, V., and Liu, S. J. Sep. Sci., 2005, 28, 1994–2009.
13. Li, S. F., and Kricka, L. J. Clin. Chem., 2006, 52, 37–45.
14. Bienvenue, J. M., Duncalf, N., Marchiarullo, D., Ferrance, J. P., and Landers, J. P. J. Forensic. Sci.,
2006, 51, 266–273.
15. Breadmore, M. C., Wolfe, K. A., Arcibal, I. G., Leung, W. K., Dickson, D., Giordano, B. C., and
Power, M. E., et al. Anal. Chem., 2003, 75, 1880–1886.
16. Cady, N. C., Stelick, S., and Batt, C. A. Biosens. Bioelectron., 2003, 19, 59–66.
17. Christel, L. A., Petersen, K., McMillan, W., and Northrup, M. A. J. Biomech. Eng-Trans ASME, 1999,
121, 22–27.
18. Wolfe, K. A., Breadmore, M. C., Ferrance, J. P., Power, M. E., Conroy, J. F., Norris, P. M., and
Landers, J. P. Electrophoresis, 2002, 23, 727–733.
19. Wu, Q., Bienvenue, J. M., Hassan, B. J., Kwok, Y. C., Giordano, B. C., Norris, P. M., Landers, J. P.,
and Ferrance, J. P. Anal. Chem., 2006, 78, 5704–5710.
20. Rolfs, A., Schuller, I., Finckh, U., and Weber-Rolfs, I. PCR: Clinical Diagnostics and Research,
Springer-Verlag, Berlin, 1992.
21. Hamalainen, M. M., Eskola, J. U., Hellman, J., and Pulkki, K. Clin. Chem., 1999, 45, 465–471.
22. Miller, D. N., Bryant, J. E., Madsen, E. L., and Ghiorse, W. C. Appl. Environ. Microbiol., 1999, 65,
4715–4724.
23. Wilson, I. G. Appl. Environ. Microbiol., 1997, 63, 3741–3751.
24. Rudi, K., and Jakobsen, K. S. Methods Mol. Biol., 2006, 345, 23–35.
25. Tian, H., Huhmer, A. F., and Landers, J. P. Anal. Biochem., 2000, 283, 175–191.
26. Cao, W., Easley, C. J., Ferrance, J. P., and Landers, J. P. Anal. Chem., 2006, 78, 7222–7228.
27. Wen, J., Guillo, C., Ferrance, J. P., and Landers, J. P. Anal. Chem., 2006, 78, 1673–1681.
28. Melzak, K., Sherwood, C., Turner, R., and Haynes, C. J. Coll. Interface Sci., 1996, 181, 635–644.
29. Mason, P. E., Neilson, G. W., Enderby, J. E., Saboungi, M. L., Dempsey, C. E., MacKerell, A. D., Jr.,
and Brady, J. W. J. Am. Chem. Soc., 2004, 126, 11462–11470.
30. Easley, C. J., Karlinsey, J. M., Bienvenue, J. M., Legendre, L. A., Roper, M. G., Feldman, S. H., and
Hughes, M. A., et al. Proc. Natl. Acad. Sci. USA, 2006, 103, 19272–19277.
31. Bienvenue, J. M., Legendre, L. A., Ferrance, J. P., and Landers, J. P. J. Forensic Sci., 2007,
Submitted.
32. Hagan, K., Bienvenue, J. M., Ferrance, J. P., and Landers, J. P. Anal. Chem., 2007, Submitted.
33. Kato, M., Dulay, M. T., Bennett, B. D., Quirino, J. P., and Zare, R. N. J. Chromatogr. A, 2001, 924,
187–195.
34. Hong, J. W., Studer, V., Hang, G., Anderson, W. F., and Quake, S. R. Nat. Biotechnol., 2004, 22,
435–439.
35. Chung, Y. C., Jan, M. S., Lin, Y. C., Lin, J. H., Cheng, W. C., and Fan, C. Y. Lab Chip, 2004, 4,
141–147.
36. Nakagawa, T., Tanaka, T., Niwa, D., Osaka, T., Takeyama, H., and Matsunaga, T. J. Biotechnol., 2005,
116, 105–111.
37. Witek, M. A., Llopis, S. D., Wheatley, A., McCarley, R. L., and Soper, S. A. Nucleic Acids Res., 2006,
34, e74.
38. Lee, J. G., Cheong, K. H., Huh, N., Kim, S., Choi, J. W., and Ko, C. Lab Chip, 2006, 6, 886–895.
39. Mullis, K., Faloona, F., Scharf, S., Saiki, R., Horn, G., and Erlich, H. Cold Spring Harb. Symp. Quant.
Biol., 1986, 51 (Pt 1), 263–273.
40. Zhang, C., Xu, J., Ma, W., and Zheng, W. Biotechnol. Adv., 2006, 24, 243–284.
41. Roper, M. G., Easley, C. J., and Landers, J. P. Anal. Chem., 2005, 77, 3887–3893.
42. Lagally, E. T., Scherer, J. R., Blazej, R. G., Toriello, N. M., Diep, B. A., Ramchandani, M.,
Sensabaugh, G. F., Riley, L. W., and Mathies, R. A. Anal. Chem., 2004, 76, 3162–3170.
43. Krishnan, M., Agrawal, N., Burns, M. A., and Ugaz, V. M. Anal. Chem., 2004, 76, 6254–6265.
44. Liu, J., Hansen, C., and Quake, S. R. Anal. Chem., 2003, 75, 4718–4723.
45. Marcus, J. S., Anderson, W. F., and Quake, S. R. Anal. Chem., 2006, 78, 956–958.
46. Lee, D. S., Park, S. H., Yang, H., Chung, K. H., Yoon, T. H., Kim, S. J., Kim, K., and Kim, Y. T.
Lab Chip, 2004, 4, 401–407.
47. Matsubara, Y., Kerman, K., Kobayashi, M., Yamamura, S., Morita, Y., Takamura, Y., and Tamiya, E.
Anal. Chem., 2004, 76, 6434–6439.
48. Pal, R., Yang, M., Johnson, B. N., Burke, D. T., and Burns, M. A. Anal. Chem., 2004, 76,
3740–3748.
49. Gulliksen, A., Solli, L., Karlsen, F., Rogne, H., Hovig, E., Nordstrom, T., and Sirevag, R. Anal. Chem.,
2004, 76, 9–14.
50. Guttenberg, Z., Muller, H., Habermuller, H., Geisbauer, A., Pipper, J., Felbel, J., Kielpinski, M.,
Scriba, J., and Wixforth, A. Lab Chip, 2005, 5, 308–317.
51. Liu, C. N., Toriello, N. M., and Mathies, R. A. Anal. Chem., 2006, 78, 5474–5479.
52. Koh, C. G., Tan, W., Zhao, M. Q., Ricco, A. J., and Fan, Z. H. Anal. Chem., 2003, 75, 4591–4598.
53. Lee, T. M., Carles, M. C., and Hsing, I. M. Lab Chip, 2003, 3, 100–105.
54. Obeid, P. J., Christopoulos, T. K., Crabtree, H. J., and Backhouse, C. J. Anal. Chem., 2003, 75,
288–295.
55. Woolley, A. T., Hadley, D., Landre, P., deMello, A. J., Mathies, R. A., and Northrup, M. A. Anal.
Chem., 1996, 68, 4081–4086.
56. Chen, J., Wabuyele, M., Chen, H., Patterson, D., Hupert, M., Shadpour, H., Nikitopoulos, D., and
Soper, S. A. Anal. Chem., 2005, 77, 658–666.
57. Oda, R. P., Strausbauch, M. A., Huhmer, A. F., Borson, N., Jurrens, S. R., Craighead, J., Wettstein, P. J.,
Eckloff, B., Kline, B., and Landers, J. P. Anal. Chem., 1998, 70, 4361–4368.
58. Lee, S., Kim, S.-W., and Ahn, C. H., µTAS, Tokyo, Japan 2006.
59. Krishnan, M., Ugaz, V. M., and Burns, M. A. Science, 2002, 298, 793.
60. Easley, C., Humphrey, J., and Landers, J. P. J. Micromech. Microeng., 2007, 17, 11–19.
61. Hughes, M. P. Electrophoresis, 2002, 23, 2569–2582.
62. Gonzalez, C. F., and Remcho, V. T. J. Chromatogr. A, 2005, 1079, 59–68.
63. Horsman, K. M., Barker, S. L., Ferrance, J. P., Forrest, K. A., Koen, K. A., and Landers, J. P. Anal.
Chem., 2005, 77, 742–749.
64. Yuen, P. K., Kricka, L. J., Fortina, P., Panaro, N. J., Sakazume, T., and Wilding, P. Genome Res., 2001,
11, 405–412.
65. Fu, A. Y., Chou, H. P., Spence, C., Arnold, F. H., and Quake, S. R. Anal. Chem., 2002, 74, 2451–2457.
66. Fu, A. Y., Spence, C., Scherer, A., Arnold, F. H., and Quake, S. R. Nat. Biotechnol., 1999, 17,
1109–1111.
67. Berger, M., Castelino, J., Huang, R., Shah, M., and Austin, R. H. Electrophoresis, 2001, 22, 3883–
3892.
68. Davis, J. A., Inglis, D. W., Morton, K. J., Lawrence, D. A., Huang, L. R., Chou, S. Y., Sturm, J. C.,
and Austin, R. H. Proc. Natl. Acad. Sci. USA, 2006, 103, 14779–14784.
69. Huang, L. R., Cox, E. C., Austin, R. H., and Sturm, J. C. Science, 2004, 304, 987–990.
70. Kumar, M., Feke, D. L., Belovich, J. M. Biotechnol. Bioeng., 2005, 89, 129–137.
71. Jonsson, H., Nilsson, A., Petersson, F., Allers, M., and Laurell, T. Perfusion, 2005, 20, 39–43.
72. Petersson, F., Nilsson, A., Holm, C., Jonsson, H., and Laurell, T. Analyst, 2004, 129, 938–943.
73. Petersson, F., Nilsson, A., Holm, C., Jonsson, H., and Laurell, T. Lab Chip, 2005, 5, 20–22.
74. Wang, Z., Grabenstetter, P., Feke, D. L., and Belovich, J. M. Biotechnol. Prog., 2004, 20, 384–387.
75. Irimia, D., Tompkins, R. G., and Toner, M. Anal. Chem., 2004, 76, 6137–6143.
76. Sethu, P., Anahtar, M., Moldawer, L. L., Tompkins, R. G., and Toner, M. Anal. Chem., 2004, 76,
6247–6253.
77. Di Carlo, D., Ionescu-Zanetti, C., Zhang, Y., Hung, P., and Lee, L. P. Lab Chip, 2005, 5, 171–178.
78. Heo, J., Thomas, K. J., Seong, G. H., and Crooks, R. M. Anal. Chem., 2003, 75, 22–26.
79. Lu, H., Schmidt, M. A., and Jensen, K. F. Lab Chip, 2005, 5, 23–29.
80. Lu, K. Y., Wo, A. M., Lo, Y. J., Chen, K. C., Lin, C. M., and Yang, C. R. Biosens. Bioelectron., 2006,
22, 568–574.
81. Culbertson, C. T. Methods Mol. Biol., 2006, 339, 203–216.
82. Wang, H. Y., Bhunia, A. K., and Lu, C. Biosens. Bioelectron., 2006, 22, 582–588.
83. McClain, M. A., Culbertson, C. T., Jacobson, S. C., Allbritton, N. L., Sims, C. E., and Ramsey, J. M.
Anal. Chem., 2003, 75, 5646–5655.
84. Gao, J., Yin, X. F., and Fang, Z. L. Lab Chip, 2004, 4, 47–52.
85. Liu, R. H., Yang, J., Lenigk, R., Bonanno, J., and Grodzinski, P. Anal. Chem., 2004, 76, 1824–1831.
86. He, Y., Zhang, Y. H., and Yeung, E. S. J. Chromatogr. A, 2001, 924, 271–284.
87. El-Ali, J., Gaudet, S., Gunther, A., Sorger, P. K., and Jensen, K. F. Anal. Chem., 2005, 77, 3629–3636.
88. Waters, L. C., Jacobson, S. C., Kroutchinina, N., Khandurina, J., Foote, R. S., and Ramsey, J. M. Anal.
Chem., 1998, 70, 158–162.
89. Belgrader, P., Hansford, D., Kovacs, G. T., Venkateswaran, K., Mariella, R., Jr., Milanovich, F.,
Nasarabadi, S., Okuzumi, M., Pourahmadi, F., and Northrup, M.A. Anal. Chem., 1999, 71, 4232–4236.
90. Kim, J., Hee Jang, S., Jia, G., Zoval, J. V., Da Silva, N. A., and Madou, M. J. Lab Chip, 2004, 4,
516–522.
91. Di Carlo, D., Jeong, K. H., and Lee, L. P. Lab Chip, 2003, 3, 287–291.
92. Shrinivasan, S., Norris, P. M., Landers, J. P., Ferrance, J. P. Clin. Lab. Med., 2007, 27(1) 173–181.
93. Grover, W. H., Ivester, R. H., Jensen, E. C., and Mathies, R. A. Lab Chip, 2006, 6, 623–631.
94. Grover, W. H., and Mathies, R. A. Lab Chip, 2005, 5, 1033–1040.
95. Skelley, A. M., Scherer, J. R., Aubrey, A. D., Grover, W. H., Ivester, R. H., Ehrenfreund, P., Grunthaner,
F. J., Bada, J. L., and Mathies, R. A. Proc. Natl. Acad. Sci. USA, 2005, 102, 1041–1046.
96. Unger, M. A., Chou, H. P., Thorsen, T., and Scherer, A., Quake, S. R. Science, 2000, 288, 113–116.
97. Melin, J., and Quake, S. R. Annu. Rev. Biophys. Biomol. Struct., 2007, 36, 213–231.
98. Thorsen, T., Maerkl, S. J., and Quake, S. R. Science, 2002, 298, 580–584.
99. Blazej, R. G., Kumaresan, P., and Mathies, R. A. Proc. Natl. Acad. Sci. USA, 2006, 103, 7240–7245.
100. Pal, R., Yang, M., Lin, R., Johnson, B. N., Srivastava, N., Razzacki, S. Z., Chomistek, K. J., et al. Lab
Chip, 2005, 5, 1024–1032.
101. Paegel, B. M., Emrich, C. A., Wedemayer, G. J., Scherer, J. R., and Mathies, R. A. Proc. Natl. Acad.
Sci. USA, 2002, 99, 574–579.
102. Hong, J. W., and Quake, S. R. Nat. Biotechnol., 2003, 21, 1179–1183.
103. Legendre, L. A., Bienvenue, J. M., Roper, M. G., Ferrance, J. P., and Landers, J. P. Anal. Chem., 2006,
78, 1444–1451.
104. Cady, N. C., Stelick, S., Kunnavakkam, M. V., and Batt, C. A. Sens. Actuat. B: Chem., 2005, B107,
332–341.
105. Burns, M. A., Johnson, B. N., Brahmasandra, S. N., Handique, K., Webster, J. R., Krishnan, M.,
Sammarco, T. S., et al. Science, 1998, 282, 484–487.
106. Dunn, W. C., Jacobson, S. C., Waters, L. C., Kroutchinina, N., Khandurina, J., Foote, R. S.,
Justice, M. J., Stubbs, L. J., and Ramsey, J. M. Anal. Biochem., 2000, 277, 157–160.
107. Hataoka, Y., Zhang, L., Mori, Y., Tomita, N., Notomi, T., and Baba, Y. Anal. Chem., 2004, 76,
3689–3693.
108. Hong, J. W., Fujii, T., Seki, M., Yamamoto, T., and Endo, I. Electrophoresis, 2001, 22, 328–333.
109. Khandurina, J., McKnight, T. E., Jacobson, S. C., Waters, L. C., Foote, R. S., and Ramsey, J. M. Anal.
Chem., 2000, 72, 2995–3000.
110. Lagally, E. T., Simpson, P. C., and Mathies, R. A. Sens. Actuat. B: Chem., 2000, B63, 138–146.
111. Lagally, E. T., Medintz, I., and Mathies, R. A. Anal. Chem., 2001, 73, 565–570.
112. Lagally, E. T., Emrich, C. A., and Mathies, R. A. Lab Chip, 2001, 1, 102–107.
113. Liu, C. N., Toriello, N. M., Maboudian, R., and Mathies, R. A. Special Publication—Royal Soc. Chem.,
2004, 297, 297–299.
114. Rodriguez, I., Lesaicherre, M., Tie, Y., Zou, Q., Yu, C., Singh, J., Meng, L. T., et al. Electrophoresis,
2003, 24, 172–178.
115. Waters, L. C., Jacobson, S. C., Kroutchinina, N., Khandurina, J., Foote, R. S., and Ramsey, J. M. Anal.
Chem., 1998, 70, 5172–5176.
116. Woolley, A. T., Hadley, D., Landre, P., deMello, A. J., Mathies, R. A., and Northrup, M. A. Anal.
Chem., 1996, 68, 4081–4086.
117. Lagally, E. T., Lee, S. H., and Soh, H. T. Lab Chip, 2005, 5, 1053–1058.
118. Yeung, S. W., Lee, T. M., Cai, H., and Hsing, I. M. Nucleic Acids Res., 2006, 34, e118.
44 Cell and Particle Separation
and Manipulation Using
Acoustic Standing Waves in
Microfluidic Systems
Thomas Laurell and Johan Nilsson
CONTENTS
44.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1230

44.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1230
44.3 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1231
44.3.1 Standing Wave Forces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1231
44.3.1.1 Primary Axial Radiation Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1231
44.3.1.2 Primary Lateral Acoustic Radiation Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1232
44.3.1.3 Secondary Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1232
44.3.2 Microscale Acoustic Standing Wave Separators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1232
44.4.1 Binary Modes of Acoustic Separation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1234
44.4.1.1 Acoustic Manipulation Based on Density Properties. . . . . . . . . . . . . . . . . . . . 1234
44.4.1.2 Carrier Media Density Manipulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1235
44.4.1.3 Acoustic Manipulation Based on Frequency Switching . . . . . . . . . . . . . . . . 1235
44.4.2 Applications of Ultrasonic Standing Wave Microresonators to Blood
Component Handling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1237
44.4.2.1 Lipid Microemboli Separation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1237
44.4.2.2 Blood Cell Washing by Carrier Media Switching . . . . . . . . . . . . . . . . . . . . . . . 1240
44.4.3 Guidelines to a Standard Lund-Separator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1241
44.4.4 Acoustic Trapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1242
44.4.4.1 Longitudinal Trapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1242
44.4.4.2 Lateral Trapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1243
44.4.4.3 Integrated Transducers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1243
44.4.4.4 Bioassays, Dynamic Arraying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1245
44.4.4.5 Cell Trapping and Culturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1246
44.4.4.6 Cell Enrichment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1247
44.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1249
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1249
1229
44.1 INTRODUCTION
There is an urgent need to find means to efficiently separate and manipulate cells and particles in
microfluidic chips because conventional mechanical separation concepts are not readily amenable
to downscaling as the channel dimensions and the particle sizes are approaching each other and
thus, physical restrictions in the flow paths tend to cause channel blockage. Therefore, noncontact
modes to induce forces on particles are of interest. Recently, ultrasonic standing waves have proven
beneficial in the quest for techniques that generate relatively large forces, which simultaneously are
compatible with microchip technologies and system integration.
44.2 BACKGROUND
Manipulation and spatial control of particulate matter in microspace are gaining interest within the
microtechnology community. This has become especially attractive when combining microfluidic
structures with physical concepts that induce controlled forces on particles inside microchannels.
Modern lithographic techniques enable the design and fabrication of static microstuctures that allow
mechanical separation and capturing of cells and particles in streaming microsystems. On the other
hand, concepts that allow noncontact cell separation and capturing offer an enhanced functionality in
terms of nonclogging continuous flow separation with low mechanical stress, that is, low shear rate
and low surface interaction. The literature is well provided with mechanical filter concepts, ranging
from conventional slits1,2 for size exclusion separation to advanced Patchinco-style single-cell cap-
ture devices.3 Recent developments by the Seki4 group demonstrate elegant microscale solutions
to laminar-flow-based size selection of particles and cells that, for example, enable clear discrim-
ination of erythrocytes from leukoytes.5 A similar approach was pioneered by the Austin group6,7
demonstrating the deterministic lateral displacement separation device that enables discrimination
of particles with size differences of tens of micrometers. Common to these concepts are all the uses
of sieving features or laminar flow splitting schemes in a continuous flow.
In contrast, techniques that induce an external force on a particle in a microfluidic environment,
not necessarily requiring a continuous flow mode for stable operation, are gaining attention. The
use of an optically induced force, optical tweezers, for spatial control of cells under microscope
surveillance has since long been a valuable research tool8,9 and is now available as commercial
instrumentation. Likewise, the use of the dielectrophoretic properties of cells and particles have been
widely researched and utilized in the development of microsystems for particle and cell separation
and trapping.10−14 More recent developments have taken dielectrophoresis electrodes to a single-
sided chip layout, thereby, vastly reducing complexity in dielectrophoretic chip designs.15,16 Further
work by Voldman17 has also demonstrated particle separation based on combined physical features
such as iso-dielectrophoresis where lateral ion gradients orthogonal to the induced dielectrophoretic
force serve as the basis for spatially defining a particle’s zero dielectrophoretic force position in a
channel.
With the advent of biofunctionalized magnetic microparticles, magnetophoretic separation has
also recently emerged as a potential mode of advanced particle separation. Pamme and Manz18 have
reported fundamental developments in microchip-integrated particle separation based on the net force
acting on the particles streaming in a flow orthogonal to the direction of the applied magnetic field.
This principle was later also employed to the magnetic bead-based extraction of mouse macrophages
and human ovarian cancer cells.19
An alternative to dielectrophoretic or magnetic force manipulation of microparticles in suspension
is the use of acoustic standing wave forces, which also have served as the basis for particle and cell
manipulation in microfluidic devices. Historical work by Kundt20 demonstrated early on the ability
to move cork particles in a standing wave pattern. The fundamental acousto-physical principles have
been well researched and described in the literature by King21 and Gorkov.22
Cell and Particle Separation and Manipulation Using Acoustic Standing Waves 1231
Normally, particles are trapped in the pressure node of the acoustic standing wave and stand-
ing wave patterns can thus be utilized to concentrate and aggregate particles in the pressure nodes.
Numerous applications have utilized the aggregation effect, for example, demonstrating yeast cells
concentration,23,24 blood plasma clarification,25,26 hybridoma cell aggregation,27 and cell reten-
tion in general.28 These are all examples of systems developed as macroscale devices in stagnant
systems where aggregated cells are collected at the bottom of a chamber. Efforts to realize contin-
uous flow separation systems based on acoustic focusing of particles and cells have, for example,
been reported by Yasuda et al.29 and Benes et al.30 The true benefits of continuous flow separa-
tion becomes evident as the scale is reduced and the flow becomes truly laminar at low Reynolds
numbers, reducing influence on separation from unsteady flow or turbulence. This fact clearly
speaks for acoustic separation to be most efficiently applied in combination with microfluidic-based
devices in a laminar flow mode, serving as a non-fluid-contact mode of controlling and separating
microparticles.
When reducing the dimensions, that is, going into the microscale domain with channel dimensions
smaller than 500 µm, inherent benefits in terms of an increased primary acoustic radiation force is
obtained due to the higher resonance frequency of the resonator.
With the progress within the lab-on-a-chip field, the need for new modes of controlling cells and
particles in a gentle but precise way is highly sought and acoustics combined with microfluidics is
now emerging as one strong candidate in this respect. This chapter will cover recent developments
within microchip-based cell and particle separation and manipulation. Applications in continuous-
flow blood component separation and blood purification will be outlined along with new modes of
cell trapping and in-chip manipulation.
44.3 THEORY
44.3.1 STANDING WAVE FORCES
44.3.1.1 Primary Axial Radiation Force
The primary axial radiation force (PRF) acting on a spherical particle in an acoustic standing wave
is defined by

−π · P02 · V · β0 4π · z 5ρp − 2ρ0 βp
FPRF = · sin · −
2λ λ 2ρp + ρ0 β0

−π · P02 · V · β0 4π · z
= · · sin (44.1)
2λ λ

5ρp − 2ρ0 βp
= − , (44.2)
2ρp + ρ0 β0
where P0 is the applied acoustic pressure amplitude, V defines the particle volume, βp is the com-
pressibility and ρp is the density of the particle in a fluid with compressibility β0 and density ρ0 ,
λ is the acoustic standing wavelength, and z is the position along the wave propagation axis. is
commonly denoted as the acoustic contrast factor. Force equation, Equation 44.1, was presented
by Gould and Coakley31 and is based on the theory on acoustic forces on small particles in a fluid
presented by Gorkov.22 Basic theory on the interaction between particles and acoustic pressure fields
assuming a rigid sphere and a friction-less liquid was presented by King.21 Later, Yosioka32 extended
the theory to also encompass compressible particles.
44.3.1.2 Primary Lateral Acoustic Radiation Force

The acoustic forces acting on the particles based on the theory by Gorkov are described in a three-
dimensional (3D) space using the negative gradient of a 3D potential function given by the standing
wave field. Equation 44.1 determines the axial force but there are also lateral/radial forces acting on
the particles from radial gradients in the standing wave field that will force the particles toward the
central axis. For particles with density and compressibility higher than the surrounding medium, the
PRF will be directed toward the nearest pressure node in the standing wave fields. When performing
acoustic focusing of particles in a continuous flow the particles will be retained in the pressure node,
that is, being trapped there as long as the drag force on the particle/cluster from the surrounding
medium is lower than the acoustic force. Normally, the axial force is larger than the radial force.
44.3.1.3 Secondary Forces

As particles gather and the interparticle distances become small, the influence of scattered sound
from neighbouring particles increases, which in turn gives rise to an interparticle acoustic force that
is commonly named secondary forces, or Bjerknes forces named after Bjerknes.33 He first described
the theory behind this phenomena, Equation 44.3, where a is the radius of the particle, d is the
distance between the particles, and θ is the angle between the centre line of the particles and the
direction of propagation of the incident acoustic wave.

(ρp − ρ0 )2 (3 cos2 θ − 1) 2 ω2 ρ0 (βp − β0 )2 2
FB (x) = 4πa 6
v (x) − p (x) (44.3)
6ρ0 d 4 9d 2
As an acoustic standing wave starts to act on a particle suspension, particles are driven toward the
standing wave nodal plane by the primary radiation force in an initial stage. As the particle density in
the nodal plane increases, the average interparticle distance becomes small and thus the influence of
the secondary force increases. This can be observed as particles moving toward one another, forming
clusters of densely packed particles. If monodispersed particles are used, nice periodic hexagonal
particles patterns are formed in the primary focal nodal plane, Figure 44.1.
44.3.2 MICROSCALE ACOUSTIC STANDING WAVE SEPARATORS

Acoustic standing wave separators are commonly designed as so-called layered resonators, where
the boundaries of a resonance chamber is defined at one side by a transducer (or a coupling layer
FIGURE 44.1 Fluorescently labeled polystyrene particles gathered in the pressure nodal plane (in plane
with the image focus) of an acoustic standing wave by the primary acoustic radiation force in a 70 µm deep
microchannel. The microbeads are also clustered in a dense hexagonal pattern by Bjerknes forces.
Reflecting layer n 3 /4, n odd

Fluid layer n 3 /2
Coupling layer n 3 /4, n odd
Transducer
FIGURE 44.2 Schematic of a layered resonator design.
Pyrex
Cavity
Silicon
In PZT Clean Dirty
FIGURE 44.3 A microfabricated layered resonator for continuous flow particle enrichment. (Reproduced
from Harris N.R., et al., Sensors and Actuators B: Chemical, 95, 425–434, 2003. With permission from
Elsevier.)
linked to a transducer) and a highly reflecting surface at the opposing side (Figure 44.2) generating
an acoustic standing wave pressure node in the centre of the fluid chamber.
The common design typically encompasses a quarter wavelength (n × λ/4; n = odd numbers)
coupling layer, a half wavelength fluid layer, and a quarter wavelength (n × λ/4; n = odd numbers)
reflecting layer. The reflecting layer can alternatively be defined by a second acoustic element. In
this case, the phase shift between the two transducers can be tuned to position the standing wave
pressure node at an arbitrary position along the direction of the wave propagation.
These types of layered separators have been researched by several groups and were applied to
various particle and cell separation applications. A microfluidic separation device was later devel-
oped by Hawkes and Coakley34 by electrodischarge machining of stainless steel films to fabricate
microchannels in a high Q-value material. The sidewalls were subsequently fine polished to obtain
a good acoustic reflecting surface and the component was assembled as a layered resonator device.
The separator was realized both in the form of an H-separator as presented by Benes et al.35 and
as a Y-shaped separator. This rather complicated procedure of fabricating a qualitative microscale
resonator clearly speaks for the transition into microfabrication technology for the acoustic separator
fabrication.
An early initiative to utilize microfabrication technology in this respect was presented by Harris
et al.36,37 who employed both anisotropically wet-etched silicon and isotropically etched glass that
were anodically bonded together, forming a separation device (Figure 44.3). An single inlet was
employed via the silicon chip and a flow-through cavity etched in the opposing acoustically reflecting
glass chip forming a resonator between the transducer surface and the glass recess and focused the
incoming particles into a sheet. The device had two outlets via the silicon chip, which allowed to
balance the flows at the two outlets such that the particle-enriched fraction was collected at one of
the outlets and the clear fraction at the other.
A common drawback with layered resonators is that visual inspection generally is not possible as
the reflecting glass layer serves as the observation window and thus, cells and particles are focused
in the plane of the observing window.
An optional mode of microscale resonator design, the Lund method, was proposed by Nils-
son et al.,38 describing a microchip with a rectangular-shaped acoustic focusing channel that was
anisotropically etched in standard 100 silicon. The end of the separation channel was provided
with a trifurcation flow splitting outlet. The chip was sealed by a anodic bonding of a glass lid.
Figure 44.4a shows a schematic cross-section of the separation chip and Figure 44.4b shows the
(a) (b)
Glass lid
Silicon chip
Gel Acoustic
PZT standing
wave
FIGURE 44.4 Schematic view of an acoustic particle concentrator according to the Lund method, enabling
online visual quality control of the particles focusing. (a) Cross-section of the Lund method acoustic focusing
chip. (b) Schematic top view of the Lund method acoustic focusing chip.
top view of the chip in operation. The almost perfect vertical silicon channel sidewalls served as
the acoustic reflecting surfaces, thus providing a standing wave nodal plane orthogonal to the visual
inspection window, which offers direct feedback on the performance of the separation.
As the chip is actuated from underneath and an acoustic standing wave is induced in the flow
channel since the excitation frequency is set to match the λ/2 resonance criterion of the channel width.
It should be noted that the obtained standing wave in the flow channel is orthogonal to the incident
wave propagation direction. This fact enables acoustic control of particle and cell displacement in
the lateral direction, in plane with the microchip, which in turns opens the route to simple integration
with other downstream acoustic separations steps or other microfluidic unit operations.
The following section of this chapter will outline several applications of using the Lund method
of ultrasonic standing wave particle/cell focusing, where the feature of visual control is of outmost
importance.

44.4.1 BINARY MODES OF ACOUSTIC SEPARATION
44.4.1.1 Acoustic Manipulation Based on Density Properties
Acoustic differentiation of cells and particles based on the primary acoustic radiation force equation,
Equation 44.1, inherently opens the route to design acoustic microparticle and cell separation systems.
The most straightforward system configuration is to make a binary separation (separation of mixed
particles into one or two populations) where the primary radiation force has different signs for, for
example, two population of particles. This situation is obtained when the acoustic contrast factor,
s in Equation 44.2 has different signs for the two particle populations to be separated. The decisive
factors for this are the density ratio and the compressibility ratio for the particle versus the carrier
solution. Most commonly, a sign shift in the acoustic contrast factor can be obtained by a proper
selection of carrier fluid density versus the density of the particle to be separated. A natural system
that nicely illustrates this can be seen in the focusing of lipid emulsions in milk into the pressure
antinodes along the sidewalls of the acoustic resonator rather than into the central pressure node.
Figure 44.5 shows milk streaming at the outlet of a Lund-separator without (44.5a) and with (44.5b)
(a) (b)
FIGURE 44.5 (a) Milk streaming through a Lund-separator without ultrasonic actuation. (b) Lipid emulsion
in milk focused in the acoustic standing wave pressure antinodes along the sidewalls of a Lund-separator.
the ultrasonic actuation at 2 MHz. Once the lipid emulsion is focused along the channel sidewalls,
the laminar flow directs the lipid emulsion to the side outlets at the channel trifurcation.
44.4.1.2 Carrier Media Density Manipulation

In situations where particles are difficult to separate owing to similar size and acoustic contrast
factors, it may still be possible to accomplish a good separation of the species by adapting the
density of the carrier media such that the acoustic contrast factor displays a sign shift for one of
the particles to separate. As tabular data on cell or particle density many a times are lacking, one
way to find the correct carrier media density for a qualitative separation is to perform a series of
centrifugation runs in which the carrier medium density is titrated. The buffer density at which the
particle populations are differentiated into the sedimentation pellet and the supernatant, respectively,
defines the buffer for a successful acoustic separation. Petersson et al.39 demonstrated this approach
when failing to separate 3 µm red coloured polystyrene particles (density 1.05 g/cm3 ) from 3 µm
white polymethyl methacrylate (PMMA) particles (density 1.22 g/cm3 ) suspended in distilled water.
By adding CsCl (0.22 mg/mL) to the aqueous carrier, the medium density was adjusted to 1.16 g/cm3 ,
which is between the densities of the two particle types to be separated. Centrifugation experiments
clearly confirmed this situation, collecting the white PMMA particles in the sedimentation pellet
and red particles were recovered in the supernatant. When running this buffer system in the acoustic
separator, 88% of the polystyrene particles were recovered in the side outlet and 96% of the PMMA
particles were collected from the centre outlet.
These findings clearly identify binary acoustic separations as a chip-integrated modality matching
density media centrifugations. Centrifugation is a major workhorse in all bio-oriented labs and the
vast variety of centrifugation kits using adapted density media for extracting cells or biological
particles can now be directly applied to acoustic separation systems.
44.4.1.3 Acoustic Manipulation Based on Frequency Switching

An optional mode of performing a binary particle separation utilizes the mobility of the particle
species to be separated. The motion of each particle is defined by its individual primary acoustic
radiation force and the opposing hydrodynamic drag force, Equation 44.4, where v is the fluid
velocity, µ is the fluid viscosity, and Dp is the particle diameter:
F = 3 · π · µ · Dp · v (44.4)
T1 T2
T3 T4
T5 T6
FIGURE 44.6 Particle separation by means of resonance frequency switching where separation is performed
based on the different migration velocities of differently sized particles. By accurate tuning of the frequency,
switching size discrimination to the first harmonic and the fundamental resonance nodes are obtained.
The net force acting on the particle defines a unique travelling speed in the acoustic standing
wave field for each particle type. If the actuation frequency is switched from the fundamental channel
resonance (λ/2) to a λ resonance criterion before the particles have reached their equilibrium position
in the channel centre, the particles will start to migrate to the pressure nodes defined at the λ/4 position
from each channel wall. By repeating the frequency switching, a situation can be accomplished where
particles of different sizes can be separated. This is accomplished due to the particle-specific primary
acoustic radiation force and the varying magnitude of the standing wave field with respect to the
spatial position in the channel. Figure 44.6 gives a schematic representation of the separation of
particles by means of acoustic resonance frequency switching. The schematic depicts six sequential
time points during the acoustic switching process.
At T1, the channel is actuated in a single wavelength (first harmonic of the resonator) standing
wave resonance at 4 MHz, driving particles to the pressure nodal planes λ/4 from each sidewall, T2.
When switching to the fundamental resonance (2 MHz), T3, the particles will migrate toward the
pressure nodal plane in the channel centre. Larger particles move faster due to a higher net force. If
the frequency is switched back at the right moment to a double node resonance mode, T4, the larger
particles will, at the switching occasion, be located close to the centre of the channel, where the axial
PRF is at its minimum in the first harmonic resonance. The smaller particles, on the other hand, will
have moved only a small distance from their original position and start to move back to the position
seen at time T2 under the influence of a higher radiation force. After switching frequency again, T5,
the larger particles will be closer to the centre than at T2 while the smaller will be closer to their
original double node position. If the switching continues, the larger particles will end up in the center
of the resonator channel and the smaller ones one-quarter of the channel width from each sidewall.
If this is performed in a continuous flow mode in a Lund-separator as presented by Siversson et al.,40
the switching system can be tuned such that the two particle types are aligned in laminar flow lanes
that can be diverted to different outlets at the end of the separation channel. Figure 44.7a shows the
outlet of a Lund separator operated in frequency switching mode. The particle suspension enters the
(a) (b)
3 and 8 µm 3 µm
particles particles
8 µm
particles
Off On
Flow Flow
FIGURE 44.7 (a) Microscope image of the situation when the ultrasound is turned off. Both 3 µm polystyrene
and 8 µm PMMA particles exit through the side outlets. (Reproduced from, Chemical Society Reviews, 36,
492–506, 2007. With permission from The Royal Society of Chemistry.) (b) Microscope image of the situation
when the ultrasound is turned on. The 8 µm particles are focused in the fundamental resonance pressure node
and exit through the centre outlet while the 3 µm particles are gathered in the first harmonic pressure nodes and
exit through the side outlets. The fundamental resonance and the first harmonic were typically active for 800 and
200 µs, respectively (total flow rate: 90 µL/min). (Reproduced from Chemical Society Reviews, 36, 492–506,
2007. With permission from The Royal Society of Chemistry.)
separation channel via two side inlets where the particles are seen laminated against the sidewalls.
In Figure 44.7b, the acoustic switching is in operation showing 8 µm PMMA beads being routed to
the center outlet and 3 µm polystyrene beads directed to the side outlets.
44.4.2 APPLICATIONS OF ULTRASONIC STANDING WAVE MICRORESONATORS TO

BLOOD COMPONENT HANDLING
44.4.2.1 Lipid Microemboli Separation
The Lund-method has been developed to address several clinical needs in blood component handling
where the direct visual access to the quality of separation has been instrumental for the platform
development. In thoracic surgery, a substantial loss of blood commonly occurs resulting in shed
blood being recovered from the chest cavity for retransfusing the blood to the patient. As the shed
blood is contaminated by tissue debris from the surgery as well as triglycerides, leaking from adi-
pose tissue undergoing surgery, the blood has to be purified before returning the blood to the patient.
Normally a mechanical filter with cutoff of about 40 µm serves as the main purification step. Cen-
trifugal instrumentation developed for this purpose is also available. In spite of these measures, a
substantial amount of triglycerides in the form of lipid microdroplets still prevail in the retransfused
blood. An increasing awareness now pinpoints lipid microemboli as an important source of cognitive
dysfunction in conjunction with major surgery where autotransfusion of shed blood is standard. The
lipid microemboli that cause cognitive dysfunction are in the size range of approximately 8 µm and
larger. As neither mechanical filters nor centrifugation eliminate lipid microemboli from shed blood
sufficiently well, new means of separation are needed. In view of these needs, acoustic separation
has recently emerged as a new potential and mild way of purifying shed blood. The components to
separate are erythrocytes and triglyceride microdroplets and as discussed earlier, a correctly tuned
acoustic standing wave binary separation mode should be feasible. When analyzing fundamental
acousto-physical data for the two components, it is evident that the two should be possible to sep-
arate in a standard acoustic resonator operated at its fundamental resonance criterion (Table 44.1).
The negative acoustic contrast factor, , for triglycerides and the corresponding positive factor
TABLE 44.1
Acousto-Physical Data of Relevant Blood Constituents, Showing a
Sign Shift in the Acoustic Contrast Factor, , for Erythrocytes versus
Triglycerides
Density (kg/m3 ) Compressibility (ms2 /kg) Velocity (m/s)
Triglycerids 913 5.34E−10 1435 −0.31
Erythrocytes 1100 3.48E−10 1616 0.29
Plasma 1010 4.40E−10 1500
(a) (b)
Erythrocyte
fraction
Lipid fraction Lipid fraction
Standing wave
FIGURE 44.8 Proof of principle of lipid particle separation from blood. Milk mixed with blood is shown in
the perfused nonactuated chip, left, and the same chip is shown when actuated at the fundamental resonance
frequency, 2 MHz, of the resonator channel. The erythrocyte fraction is clearly seen exiting via the center outlet
and the lipid fraction (white streaks along the sidewalls) in the side channels.
for erythrocytes shows that erythrocytes should be possible to collect at the center outlet of the a
Lund-separator and the lipid microemboli in the side channels.
A proof of concept of this hypothesis was performed on a mixture of blood and milk, where
the lipid emulsion in milk modeled the triglyceride suspension in real blood.41 Figure 44.8 shows a
binary separation of erythrocytes from the milk lipid fraction in a mixture of blood and milk. The
corresponding studies on blood and triglycerides were reported by Petersson et al.42 demonstrating
triglyceride microdroplet removal from blood at efficiencies as high as 95%. Figure 44.9 shows
the triglyceride elimination efficiency as measured on a radioactively labeled triglyceride emulsion
(0.3 µm droplet size) spiked at 1% level in blood. The separation was performed at a flow rate of
300 µL/min.
Clinical requirements on a separation system for autolog blood washing and transfusion request
throughputs of at least 1000 mL/h. In view of the limited throughput of the reported acoustic lipid
elimination microsystem, a clinial implementation of this separation strategy seems, at a first glance,
not feasible. However, when considering the potential of microfabrication to generate large arrays
of identical acoustic separation channels, this can very well be accomplished. A clear benefit of
the Lund-method is that multiple channels may be actuated simultaneously by the same trans-
ducer, which reduces the requirements on the actuation unit in an instrumental setup for clinical
purposes.
An effort to demonstrate the possibility of upscaling the throughput was presented by Jönsson
et al.43 Eight parallel separation channels, 350 µm wide and 125 µm deep each, were connected in
100
90
Separation efficiency, fat (%)

80
70
60
50
40
30
N- 3 3 3 3
2.5 5.0 7.5 10.0
Concentration (% erythrocytes)
FIGURE 44.9 Acoustic separation of lipid particles from blood at varying erythrocyte concentrations. Flow
rate = 300 µL/min. (Reproduced from Petersson F., et al., The Analyst, 129, 938–943, 2004. With permission
from Royal Society of Chemistry.)
Contaminated
blood inlet
Clean blood
outlet
Lipid outlets
FIGURE 44.10 To increase the throughput of the acoustic separation device, eight parallel channels were
connected in a bifurcation network. The image shows two eight channel separators on a 3 silicon wafer.
a microfluidic bifurcation structure such that the chip was supplied with a single inlet for blood, a
single outlet for purified blood, and a set of outlets at the chip backside for the triglyceride fraction.
The separator was operated at a flow rate of 60 mL/h. Triglyceride elimination was recorded to
range up to 90%. It should also be noted that the processing of the blood in the acoustic resonator
did not induce any measurable hemolysis above the background already present in the inlet blood
fraction. Figure 44.10 shows an early version of two eight-channel separator structures connected in
bifurcation structures on a 3 silicon wafer.
Particles in
buffer II
Particles in Particles free
buffer I buffer I
Particles in Particles in
buffer I buffer I
Buffer II Buffer II
Ultrasound actuation off Ultrasound actuation off
FIGURE 44.11 Acoustic switching of particles between two buffer systems.
44.4.2.2 Blood Cell Washing by Carrier Media Switching

The ability to spatially manipulate cells in a streaming medium as outlined in this chapter opens new
routes to perform online cell manipulation tasks. A fundamental unit operation is to be able to change
the buffer conditions for a particle or cell in suspension. This is commonly done by centrifugation,
removal of the supernatant, and finally resuspension in the new buffer. These steps can be directly
implemented in a streaming acoustic half wavelength resonator (Figure 44.11) by allowing particles
suspended in buffer I to enter a separation channel via two symmetrical side inlets to a main channel
that is provided with buffer II. The particle suspension is thus laminated in its original buffer fluid
along the sidewalls of the flow channel. Figure 44.11 schematically shows this setting and as the
ultrasonic standing wave is activated, the particles start to migrate to the pressure node in the centre
of the channel (Figure 44.11), which is occupied by the buffer II stream.
This mode of operation has been investigated using the Lund-separator to wash blood contam-
inated by inflammatory components and coagulation factors, which is the case for drainage blood
collected from patients in the intensive care units in the postsurgery state. Normally, this blood
is centrifuged and retransfused to the patients after removing the supernatant and resuspension in
plasma replacement fluid. Petersson et al.44 demonstrated an acoustic resonator microchip with a
channel configuration as outlined in Figure 44.11. Figure 44.12 shows a close up of the outlet from
such a separation chip where a model system with white polyamide particles, 5 µm, are suspended
in a buffer together with the colour compound, Evans blue. The nonactuated chip, Figure 44.12a,
shows the particles exiting the chip together with the original blue coloured buffer via the side outlets.
As the acoustic actuation is turned on, the particles are immediately transferred into the uncoloured
buffer stream in the centre of the channel, see Figure 44.12b.
Blood washing using the proposed particle switching strategy demonstrated wash efficiencies up
to 95%. To meet the needs for throughput, the same approach to parallel channel design as proposed
for the lipid microemboli removal chip can be implemented. Switching of particles between different
media is a powerful contact-free microfluidic unit operation for cell manipulation on chip where, for
example, protocols for compound testing and receptor activation ranges can be titrated.
(a) (b)
FIGURE 44.12 (a) White particles suspended in a blue colour compound, Evans blue, are seen to follow
their original buffer stream to the side outlets of the separation chip. As the chip is actuated at the resonance
frequency, 2 MHz, of the resonator channel, the particles are forced by the axial primary radiation force into
the central uncoloured buffer stream and exits the chip via the centre outlet (b).
(a) (b)
I II
B′ 90º
A–A′ C–C′
D D′ 125.7º
A
<110> B–B′ D–D′
I A′
C
II C′
FIGURE 44.13 Schematic of the layout for a separation chip fabricated in standard <110>- oriented silicon
with either x-shaped (I) or y-shaped (II) channel outlet. (a) Mask alignment. Reproduced from Nilsson A., et
al., Lab on a Chip, 4, 131–135, 2004. With permission from Royal Society of Chemistry. (b) Channel cross-
sections. Reproduced from Nilsson A., et al., Lab on a Chip, 4, 131–135, 2004. With permission from Royal
Society of Chemistry.
44.4.3 GUIDELINES TO A STANDARD LUND-SEPARATOR

A separation chip, according to the Lund-method, is most easily made by anisotropic etching of a
100-oriented silicon wafer. The channel mask should be aligned 45◦ offset to the 110-cut phase of
the wafer, Figure 44.13a, to accomplish the desired rectangular channel cross-section, Figure 44.13b.
After performing the photolithography and etching of the 1 µm surface oxide on the wafer in
buffered hydrofluoric acid (HF), the flow channel is etched in KOH (potassium hydroxide, Merck
KGaA, 64271 Darmstadt, Germany), 40 g in 100 mL deionised water to the desired channel width
and depth. Normal etch rates are approximately 1 µm/min. After terminating the anisotropic etching,
the remaining surface oxide is stripped in HF, washed and blown dry. Inlet and outlet holes should
be etched from the backside of the wafer to provide an inspection area at the channel side, free from
connecting tubings. The chip is sealed by anodic bonding of a Pyrex glass lid to the channel side of
the chip. Finally, silicone rubber tubings are glued to the backside of the wafer with standard silicone
rubber elastomer.
The ultrasonic transducer was a PZ26 piezoelectric ceramic disc (Ferroperm Piezoceramics A/S,
Kvistgard, Denmark). The tranducer should be designed to operate in the region of the resonance
frequency of the resonator channel.
The piezoelectric element is glued to the backside of the chip with epoxy (2 Ton Clear Epoxy,
ITWDevcon, Danvers, MA, USA). Alternatively, a small aliquot of ultrasonic coupling gel can be
clamped between the transducer and the resonator chip (Figure 44.4a). Normal actuation voltages
range between 5 and 20 V for performing a standard acoustic particles-focusing experiment. A power
amplifier may be required for the driving of the piezoelectric element. Flow rates are normally in
the range of 10–200 µL/min for a 350 × 125 µm channel (width × depth). Fluid is preferably
aspirated through the chip, that is, syringe pumps are connected to the chip outlets, drawing fluid
from reservoirs in which the inlet tubing is immersed.
44.4.4 ACOUSTIC TRAPPING

As have been mentioned earlier, trapping of particles will occur in an acoustic standing wave field if
the retaining forces from the acoustic field are larger than the drag forces exerted from a surrounding
fluid flow (Equations 44.1 and 44.4). The standing wave field may either be generated along the
fluid flow lines for longitudinal trapping or transversal to the fluid flow lines for lateral trapping.
According to the theory given earlier in this chapter, the trapping force is stronger in the longitudinal
case compared to the lateral for a given acoustic intensity.
44.4.4.1 Longitudinal Trapping

An example of a system using longitudinal trapping is shown in Figure 44.14. An 8.5 MHz focused
PZT-transducer, 20-mm in diameter, is used together with a molybdenum reflector to generate a
standing wave inside a quartz capillary.45 Figure 44.15 shows trapped 4.7 µm latex spheres inside
a 75 µm i.d. capillary. The standing wave pattern is clearly visible after the particles have been
trapped. The system was used to demonstrate size selective separation, where 4.7 µm latex particles
were retained in the capillary while the majority of 3.0 µm particles passed through the trap. This
method generates several trapping points along the capillary and it is quite difficult to selectively
trap particles in a certain position. The advantage is that the primary axial radiation force used will
generate a stronger trapping force for a given acoustic intensity, as discussed earlier. Alternatively,
the acoustic intensity can be decreased while maintaining trapping.
Transducer Reflector
Capillary Flow direction
Acoustic
wave
Trapping
Water cell region
FIGURE 44.14 Experimental setup for the longitudinal capillary ultrasonic trap. (Reproduced from
Wiklund M., et al., Ultrasonics, 41, 329, 2003. With permission.)
(a)
ID
(b)
/2 = 90 µm
FIGURE 44.15 Fluorescence images of 4.7 µm latex particles inside the 75 µm I.D. capillary without
(a) and with (b) activated transducer. (Reproduced from Wiklund M., et al., Ultrasonics, 41, 329, 2003.
With permission.)
The same group has also presented an alternative way to couple acoustic energy into a microfluidic
chip using a flat transducer and a refractive element placed on top of the chip.46 This setup is used
in combination with dielectrophoretic (DEP) forces with an outlook to combined manipulation of
bioparticles or individual cells using the two forces simultaneously.
44.4.4.2 Lateral Trapping

In the case of lateral trapping, the standing wave is applied perpendicular to the fluid flow through the
trapping device. A multilayer resonant structure is commonly used to generate the standing wave (cf.
Figure 44.2). Basically, two different transducer configurations have been presented in the literature.
The transducer is either connected via an intermediate coupling layer to the trapping cavity or is in
direct contact with it. The trapping cavity is normally sealed with a glass lid acting as both acoustic
reflector and inspection window. The trapping cavity height should be a multiple of λ/2, depending
on the number of desired trapping positions. The glass resonator should be an odd number of λ/4
in thickness and if an intermediate coupling layer is used between the transducer and the trapping
cavity, this is normally chosen to have a thickness of λ/4.
A transducer with an intermediate coupling layer was used in a setup by Bazou et al.
(Figure 44.16).47 A circular resonator was designed for cell manipulation experiments. A 1.5 MHz
disc transducer, 12 mm in diameter, was glued to a steel coupling layer (3λ/4 thick) forming the
bottom of the resonator. The resonator cavity (height: 0.5 mm) was sealed by a 1 mm quartz glass
lid and the device was driven at 1.57 MHz, corresponding to the resonance frequency of the res-
onator cavity. The cell suspension was filled into the resonator and the fluid flow was stopped as the
ultrasound was activated. The cells moved into the resonator nodal plane within 1 s (Figure 44.17).
Wiklund and coworkers48 used a similar setup with a glass-liquid-glass resonator activated by a
5 mm in diameter external transducer (Figure 44.18). The channel height (λ/2, 260 µm) was defined
by a polydimethylsiloxane (PDMS) spacer. The glass coupling layer and the glass reflector were
both chosen to have a thickness of λ/4, 550 µm. The system was run at a frequency close to 3 MHz
and was used for investigation of adherent COS-7 cell viability after being exposed to the ultrasound
in the trap under continuous perfusion with cell medium. No adverse effect was found for exposure
times up to 75 min.
44.4.4.3 Integrated Transducers

Trapping devices may also be designed without the coupling layer between the transducer and the
resonance chamber, that is, the transducer is in direct contact with the liquid in the chamber. This
has been demonstrated by, for example, Spengler et al.49 A 5 × 15 mm transducer was used in
contact with a 5 mm wide channel formed by a 0.8 mm polytetrafluoroethylene (PTFE) spacer. The
channel height corresponded to one wavelength, that is, two pressure node planes were generated
Reflector
Camera
Active area
Light
H.F. Drive source
Sample
In-/Outlet
Resonator
z-axis Sample in
Transducer SMA
connector
FIGURE 44.16 Acoustic resonator for cell trapping experiments. Its main components were a 1.5 MHz
disk transducer that was glue-attached to a steel acoustic coupling layer, a sample volume, and glass acoustic
reflector. The thicknesses of the different layers were selected to give a highly resonant system. (Reproduced
from Bazou D., et al., Ultrasound in Medicine and Biology, 31, 423, 2005. With permission.)
FIGURE 44.17 Neural cells suspended in the ultrasound trap 30 min after initiation of ultrasound. Scale
bar is 70 µm. (Reproduced from Bazou D., et al., Ultrasound in Medicine and Biology, 31, 423, 2005. With
permission.)
when driven at 1.93 MHz. The channel was closed by a 0.75 mm thick quartz glass lid acting as an
acoustic reflector.
A similar configuration using miniaturized transducers was demonstrated by Lilliehorn et al.50
Transducer elements measuring 0.8 × 0.8 mm were mounted on a printed circuit board (PCB) and
cast in epoxy that was polished down to the upper surface of the transducers (Figure 44.19). A cavity
was drilled in the PCB underneath the transducer to ensure air backing of the transducer, which is
important for the operation of the resonant system with a minimum off loss (high Q-value). The flow
channels (λ/2 height) were structured on a soda-lime glass plate using a photosensitive polymer,
SU-8, for defining the walls. The glass plate acted as a reflector and the thickness was 1.55 mm
that corresponded to an odd number of λ/4. The system was operated at around 10 MHz. A higher
frequency is advantageous to use since the forces acting on the particles will be larger for a given
acoustic intensity (Equation 44.1). When studying the resulting trapping behavior of the device, it was
(a) (b) PZT transducer

∅ = 5 mm
Tubing connecting f = 3 MHz
needle and syringe
Microchannel with cell medium
PDMS Glass λ /4
flap
PDMS λ /2
PDMS λ /4
spacer Cell injector Glass
PDMS
PZT transducer block
FIGURE 44.18 (a). Top view of the glass–PDMS–glass microfluidic chip with an integrated PDMS block
containing two separate inlets with tubing and needles. (b). Cross-section of the microfluidic chip showing
the circular 3-MHz PZT transducer and the three-layered structure. (Reproduced from Hultstrom J., et al.,
Ultrasound in Medicine and Biology, 33, 145, 2007. With permission.)
FPRF
(2n −1).λ
Glass reflector Flat
4
λ/2
λ/2 Epoxy Transducer Epoxy
Air backing
FIGURE 44.19 A side view schematic of the microfluidic trapping device using miniature transducers. The
acoustic forces trap the particles in clusters in the center of the channel as illustrated in the insert. (Reproduced
from Evander M., et al., Analytical Chemistry, 79, 2984–2991, 2007. With permission.)
seen that the 5 µm polyamide particles were trapped in several clusters above the transducer surface
in the centre of the channel, as indicated in Figure 44.19. This was theoretically and experimentally
shown to be a result of the nearfield pattern in the acoustic intensity generated by the transducer
(Figure 44.20). The acoustic field varies spatially across the transducer surface since the wavelength
of the acoustic signal is much shorter than the dimensions of the transducer, which results in an
interference pattern.
44.4.4.4 Bioassays, Dynamic Arraying

The intention of the device presented by Lilliehorn et al.50 was to utilize the miniaturized transducers
in combination with the laminar flow conditions in microfluidic systems to create a dynamic arraying
device holding several individually controlled transducers as envisioned in Figure 44.21.51 Here,
an array of 16 transducers are loaded with different functionalized beads in a first step. The beads
are acoustically trapped in the positions given by the transducers. In a second step, samples are
supplied through the orthogonal channels and the response is, for example, read by fluorescence.
The transducers are then deactivated and the beads are flushed out of the device and the whole
procedure is repeated again for a new set of beads and samples.
0
50
100
150
200
250
0
300
FIGURE 44.20 Spatial variation of the acoustic field across the transducer surface. The actual trapping
positions are given by the near field pressure distribution as shown in the 3D-image. Particles will be trapped
in clusters at the local pressure minima, indicated by the peaks in the figure. (Reproduced from Evander M.,
et al., Analytical Chemistry, 79, 2984–2991, 2007. With permission.)
(a) (b) (c)
FIGURE 44.21 Illustrations of the concept of dynamic arraying showing insertion of the solid phase of
different specificity (a) through inlets (A, B…X), trapping of bead clusters using the ultrasonic transducer array
(b) and perfusion of sample (c) through inlets (A, B…Y) followed by fluorescence read-out. (Reproduced from
Lilliehorn T. et al., Sensors and Actuators B, 106, 851, 2005. With permission.)
A simplified model system is shown in Figure 44.22 using three (550 × 550 µm) transducers.
A channel structure (width 600 µm, height 61 µm) incorporating hydrodynamic focusing of the
sample inlet was fabricated by wet etching of glass. Beads were trapped from a continuous flow and
moved between the different transducers by activating them in a sequence. A simple bioassay was
performed by trapping 6.7 µm biotinylated polystyrene beads and perfusing them with fluorescently
tagged avidin through the orthogonal channels (marked “Analyte inlet” in Figure 44.22).
44.4.4.5 Cell Trapping and Culturing

There is a clear trend today within the bioanalytical and biomedical fields toward more frequent use
of cell-based studies. The dimensions of microfluidic systems are well matched to meet the demand
on cell-based systems. Still, new methods are needed that can efficiently handle and manipulate cells
in those formats. Examples have already been given in this chapter where acoustic forces are used to
trap and manipulate cells. The device in Figure 44.22 has been further developed for use in cell-based
bioassays.52 The temperature characteristics of the device have been examined to be able to control the
temperature during the cell experiments. The major source of heat in the acoustic resonance systems
presented here is the power dissipation in the transducer itself. The power dissipation follows a
(a) Glas lid with fluidic connections (b)
(c) Buffer inlet (d)
Sample inlet
Trapping site
Analyte inlet
Analyte inlet
Analyte inlet
(d) Outlet
FIGURE 44.22 The microfluidic acoustic resonator is based on a PCB with three miniature PZT-transducers,
550 × 550 µm (a). The PCB provides fluidic and electric connections to the transducers. A glass lid with
microfluidic channels placed over the PCB defines the resonator cavity over each transducer (b), and the entire
assembly is fixed by a brass holder (c). A schematic of the channels with the transducers, that is, trapping sites,
marked with grey (d). (Reproduced from Evander M., et al., Analytical Chemistry, 79, 2984–2991, 2007. With
permission.)
2h 4h 6h
FIGURE 44.23 Growth of yeast cells trapped in the acoustic device while being perfused with cell medium.
The images show the increase of the number of cells in the cell cluster after 2, 4, and 6 h of cultivation.
(Reproduced from Evander M., et al., Analytical Chemistry, 79, 2984–2991, 2007. With permission.)
quadratic dependence on the drive voltage. Experiments have shown that it is possible to select a
drive voltage amplitude that gives both good trapping conditions and an appropriate temperature
profile for cell-based experiments. Figure 44.23 shows the results from growing yeast cells in the
trapping device. The cells were cultivated for 6 h under continuous perfusion of cell medium. Images
were recorded every hour to follow the growth of the cells.
44.4.4.6 Cell Enrichment

An interesting application of the acoustic trap is to use it for enrichment of particles/cells from
a dilute sample. The origin may be a rare event experiment or a sample where everything but
B B
S S
G1 G2 G1 G2
Piezo Piezo
on off
FIGURE 44.24 Channel structure with the sample inlet, S, and the hydrodynamically focusing buffer inlet, B,
together with with the valve-less switching inputs G1 and G2.
(a) (b) (c) (d)
FIGURE 44.25 Principle of cell enrichment using the acoustic trapping. (a) The tranducer is activated, (b) the
sample is supplied, (c) the particles/cells in the sample liquid are trapped at the transducer, and (d) the switching
flow is switched and the transducer is deactivated. The trapped particles/cells are released to the exit down left.
the cells/particles of interest have been lysed or dissolved. This has been demonstrated using the
transducer plate shown in Figure 44.22a together with new design of the glass channel system, shown
in Figure 44.24. Hydrodynamic focusing is used for the sample inlet (Inlet B and S) to ensure high
trapping efficiency.52 Two inlets have been added (G1 and G2) for valveless switching of the flow
between the two outlets below. If the flow in G1 is significantly higher than the flow in G2, the
sample/buffer flow will be directed to the right outlet and vice versa. The channel dimensions at the
transducer were 900 µm width and 70 µm height. The system was operated at 10.8 MHz.
The operation of the device is shown in Figure 44.25 using polystyrene beads in a dye as model
sample fluid. The transducer is activated and the hydrodynamically focused sample flow is supplied.
The switching flows (inputs G1 and G2) are set for guiding the sample flow to the right outlet after
passing the transducer. The particles/cells in the sample flow meeting the conditions for trapping
are retained at the transducer in a cluster (Figure 44.25c). After optional washing of the cluster, the
flows in inlets G1 and G2 are switched to address the left output and the transducer is deactivated.
The trapped cluster is released and supplied to the output for further processing.
This system has been used for sample preparation for enhanced analysis of sexual assault
evidence.53 A sample containing sperm cells and lysed female epithelial cells was infused into
the device during 5 min. The sperm cells were trapped at the transducer. After finished infusion, the
cells were washed with PBS for another 5 min, then released and collected at the outlet for qPCR
analysis. The original sample contained 5% male fraction and 95% female fraction. After processing
in the trapping device, the collected liquid at the female outlet (down right in Figure 44.25) contained
0.5% male fraction and 99.5% female, while the collected liquid at the male outlet (down left in
Figure 44.25) contained 85% male fraction and 15% female fraction, that is, a 17-fold enrichment
of the male fraction.
44.5 CONCLUSIONS
Acoustic standing wave manipulation of particles in microfluidic systems offers the implementation
of a wide range of particle/cell handling unit operations without introducing any moving parts, where
binary separations, buffer media switching, particles sizing, cell enrichment, and cell trapping can
be accomplished. Cell manipulation in microfluidic systems using ultrasonic standing waves have
proven to be a mild and yet robust technique well suited for performing chip-integrated cell handling
tasks. In contrast to dielectrophoretic and optical tweezer technology, higher throughputs are offered
in acoustically controlled systems. The low device complexity and the fact that changes in ionic
strength of the carrier buffer does not affect the acoustic forces induced on the cells to be handled
make acoustic separators an especially attractive platform for cell handling in their native body fluids
and in long-term experiments with, for example, gradient perfusion systems.
With the rapid development of cell-based microfluidic systems in the lab-on-a-chip field, ultra-
sonic standing wave manipulation and control of cells in microspace can be anticipated to play an
increasing role in future microfluidic systems.
REFERENCES
1. Kittisland G., Stemme G., and Nordén B., A submicro particle filter in silicon, Sensors and Actuators,
A21–A23, 904–907, 1990.
2. Brody J.P., Osborn T.D., Forster F.K., and Yager P., A planar microfabricated fluid filter, Sensors and
Actuators A, 54, 704–708, 1996.
3. Di Carlo D., Wu L.Y., and Lee L.P., Dynamic single cell culture array, Lab Chip, 6, 1445–1449, 2006.
4. Yamada M. and Seki M., Microfluidic particle sorter employing flow splitting and recombining,
Analytical Chemistry, 78, 1357–1362, 2006.
5. Yamada M. and Seki M., Microfluidic device for continuous and hydrodynamic separation of blood
cells, in Micro Total Analysis Systems, Proceedings of µTAS 2006 Conference. T. Kitamori, H. Fujita,
and S. Hasabe (Eds.), Society for Chemistry and Micro-Nano Systems, Tokyo, Japan, 1052–1054,
2006.
6. Huang L.R., Cox E.C., Austin R.H., and Sturm J.C., Continuous particle separation through
deterministic lateral displacement, Science, 304, 987–990, 2004.
7. Davis J.A., Inglis D.W., Morton K.J., Lawrence D.A., Huang L.R., Chou S.Y., Sturm J.C., and
Austin R.H., Deterministic hydrodynamics: Taking blood apart, Proceedings of the National Academy
of Sciences of the United States of America, 103, 14779–14784, 2006.
8. Ashkin A., Dziedzic J.M., Bjorkholm J.E., and Chu S., Observation of a single-beam gradient force
optical trap for dielectric particles, Optics Letters, 11, 288–290, 1986.
9. Block S.M., Making light work with optical tweezers, Nature, 360, 493–495, 1992.
10. Huang Y. and Pethig R., Electrode design for negative dielectrophoresis, Measurement Science and
Technology, 2, 1142–1146, 1991.
11. Markx G.H., Talary M.S., and Pethig R., Separation of viable and nonviable yeast using dielectro-
phoresis, Journal of Biotechnology, 32, 29–37, 1994.
12. Schnelle T., Hagedorn R., Fuhr G., Fiedler S., and Muller T., Three-dimensional electric field traps
for manipualtion of cells—calculation and experimental verification, Biochimica et Biophysica Acta,
1157, 127–140, 1993.
13. Fiedler S., Shirley S.G., Schnelle T., and Fuhr G., Dielectrophoretic sorting of particles and cells in a
microsystem, Analytical Chemistry, 70, 1909–1915, 1998.
14. Cummings E.B. and Singh A.K., Dielectrophoresis in microchips containing arrays of insulating posts:
Theoretical and experimental results, Analytical Chemistry, 75, 4724–4731, 2003.
15. Rosenthal A. and Voldman J., Dielectrophoretic traps for single-particle patterning, Biophysical
Journal, 88, 2193–2205, 2005.
16. Rosenthal A., Taff B.M., and Voldman J., Quantitative modeling of dielectrophoretic traps, Lab on a
Chip, 6, 508–515, 2006.
17. Vahey M.D. and Voldman J., Iso-dielectric separation: A new method for the continuous-flow screening
of cells, in Micro Total Analysis Systems, Proceedings of µTAS 2006 Conference. T. Kitamori, H. Fujita,
and S. Hasabe (Eds.), Society for Chemistry and Micro-Nano Systems, Tokyo, Japan, 1052–1054,
2006.
18. Pamme N. and Manz A., On-chip free-flow magnetophoresis: Continuous flow separation of magnetic
particles and agglomerates, Analytical Chemistry, 76, 7250–7256, 2004.
19. Pamme N. and Wilhelm C., Continuous sorting of magnetic cells via on-chip free-flow magnetophore-
sis, Lab on a Chip, 6, 974–980, 2006.
20. Kundt A. and Lehmann O., Ueber longitudinale Schwingungen und Klangfiguren in cylindrischen
Flüssigkeitssäulen, Annalen der Physik und Chemie, 153, 1, 1874.
21. King L.V., On the acoustic radiation pressure on spheres, Proceedings of the Royal Society of London,
A147, 212–240, 1934.
22. Gorkov L.P., On the forces acting on a small particle in an acoustical field in an ideal fluid, Soviet
Physics Doklady, 6, 773, 1962.
23. Limaye M.S., Hawkes J.J., and Coakley W.T., Ultrasonic standing wave removal of microorgansims
from suspension in small batch systems, Journal of Microbiological Methods, 27, 211, 1996.
24. Hawkes J.J. and Coakley W.T., A continuous flow ultrasonic cell-filtering method, Enzyme and
Microbial Technology, 19, 57, 1996.
25. Cousins C.M., Holownia P., Hawkes J.J., Limaye M.S., Price C.P., Keay P.J., and Coakley W.T., Plasma
preparation from whole blood using ultraound, Ultrasound in Medicine and Biology, 26, 881, 2000.
26. Peterson S., Perkins G., and Baker C., Development of an ultrasonic blood separator, Proceedings of
the IEEE Eighth Annual Conference of the Engineering in Medicine and Biology Society, 154–156,
1986.
27. Trampler F., Sonderhoff S.A., Pui P.W.S., Kilburn D.G., and Piret J.M., Acoustic cell filter for high-
denisty perfusion culture of hybridoma cells, Biotechnology, 12, 281–284, 1994.
28. Doblhoffdier O., Gaida T., Katinger H., Burger W., Groschl M., and Benes E., A novel ultra-
sonic resonance field device for the retention of animal cells, Biotechnology Progress, 10, 428–432,
1994.
29. Yasuda K., Umemura S., and Takeda K., Concentration and fractionation of small particles in liquid
by ultrasound, Japanese Journal of Applied Physics Part 1—Regular Papers Short Notes & Review
Papers, 34, 2715, 1995.
30. Benes E., Groschl M., Nowotny H., Trampler F., Keijzer T., Bohm H., Radel S., et al., Ultrasonic
separation of suspended particles, Ultrasonics Symposium, 2001 IEEE, 1, 649, 2001.
31. Gould R.K. and Coakley W.T., The effects of acoustic forces on small particles in suspension. In
Proceeding of the 1973 Symposium of Finite Amplitude Wave Effects in Fluids, Bjørno L. (Ed.), IPC,
Guilford, UK, pp. 252–257.
32. Yosioka K. and Kawasima Y., Acoustic radiation pressure on a compressible sphere, Acustica, 5,
167–173, 1955.
33. Bjerknes V.F.K., in “Die Kraftfelder,” F. Vieweg, (Ed.), Braunschweig, Germany, 1909.
34. Hawkes J.J. and Coakley W.T., Force field particle filter, combining ultrasound standing waves and
laminar flow, Sensors and Actuators B: Chemical, 75, 213, 2001.
35. Benes E., Groschl M., Nowotny H., Trampler F., Keijzer T., Bohm H., Radel S., et al., Ultrasonic
separation of suspended particles, Ultrasonics Symposium, 2001 IEEE, 1, 649–659, 2001.
36. Harris N., Hill M., Shen Y., Townsend R.J., Beeby S., and White N., A dual frequency, ultrasonic,
microengineered particle manipulator, Ultrasonics, 42, 139, 2004.
37. Harris N.R., Hill M., Beeby S., Shen Y., White N.M., Hawkes J.J., and Coakley W.T., A silicon
microfluidic ultrasonic separator, Sensors and Actuators B: Chemical, 95, 425, 2003.
38. Nilsson A., Petersson F., Jönsson H., and Laurell T., Acoustic control of suspended particles in micro
fluidic chips, Lab on a Chip, 4, 131–135, 2004.
39. Petersson F., Åberg L., and Laurell T., Acoustic separation of particles with similar acoustic properties
by means of medium density manipulation, in Micro Total Analysis Systems, Proceedings of µTAS
2006 Conference. T. Kitamori, H. Fujita, and S. Hasabe (Eds.), Society for Chemistry and Micro-Nano
Systems, Tokyo, Japan, 1052–1054, 2006.
40. Siversson C., Petersson F., Nilsson A., and Laurell T., Acoustic particle sizing in microchannels by
means of ultrasonic frequency switching, in Micro Total Analysis Systems, Proceedings of µTAS 2004.
T. Laurell, J. Nilsson, J. D., Harrison, K. Jensen, and J. P. Kutler (Eds.), The Royal Society for
Chemistry, Cambridge, UK, 330–332, 2004.
41. Petersson F., Nilsson A., Holm C., Jönsson H., and Laurell T., Separation of lipids from blood ultilizing
ultrasonic standing waves in microfluidic channels, The Analyst, 129, 938–943, 2004.
42. Petersson F., Nilsson A., Holm C., Jönsson H., and Laurell T., Continuous separation of lipid particles
from erythrocytes by means of laminar flow and acoustic standing wave forces, Lab on a Chip, 5, 20–22,
2005.
43. Jönsson H., Holm C., Nilsson A., Petersson F., and Laurell T., Separation by ultrasonic standing waves
can ameliorate brain damage after cardiac surgery, Annals of Thoracic Surgery, 78, 1572–1578, 2005.
44. Petersson F., Nilsson A., Holm C., Jönsson H., and Laurell T., Carrier medium exchange through
ultrasonic particle switching in microfluidic channels, Analytical Chemistry, 77, 1216–1221, 2005.
45. Wiklund M., Spégel P., Nilsson S., and Hertz H.M., Ultrasonic-trap-enhanced selectivity in capillary
electrophoresis, Ultrasonics, 41, 329, 2003.
46. Wiklund M., Günther C., Lemor R., Jäger M., Fuhr G., and Hertz H.M., Ultrasonic standing wave
manipulation technology integrated into a dielectrophoretic chip, Lab on a Chip, 6, 1537, 2006.
47. Bazou D., Kuznetsova L.A., and Coakley W.T. Physical environment of 2-D animal cell aggregates
formed in a short pathlength ultrasound standing wave trap, Ultrasound in Medicine and Biology,
31, 423, 2005.
48. Hultstrom J., Manneberg O., Dopf K., Hertz H.M., Brismar H., and Wiklund M., Proliferation and
viability of adherent cells manipulated by standing-wave ultrasound in a microfluidic chip. Ultrasound
in Medicine and Biology, 33, 145, 2007.
49. Spengler J.F., Jekel M., Christensen K.T., Adrian R.J., Hawkes J.J., and Coakley W.T., Observa-
tion of yeast cell movement and aggregation in a small-scale MHz-ultrasonic standing wave field,
Bioseparation, 9, 329, 2000.
50. Lilliehorn T., Simu U., Nilsson M., Almqvist M., Stepinski T., Laurell T., Nilsson J., and Johansson
S., Trapping of microparticles in the near field of an ultrasonic transducer. Ultrasonics, 43, 293, 2005.
51. Lilliehorn T., Nilsson M., Simu U., Johansson S., Almqvist M., Nilsson J., and Laurell T., Dynamic
arraying of microbeads for bioassays in microfluidic channels, Sensors and Actuators B, 106, 851,
2005.
52. Evander M., Johansson L., Lilliehorn T., Piskur J., Lindvall M., Johansson S., Almqvist M., Lau-
rell T., and Nilsson J., Noninvasive acoustic cell trapping in a microfluidic perfusion system for online
bioassays. Analytical Chemistry, 79, 2984–2991, 2007.
53. Evander, M., Using acoustic differential extraction to enhance analysis of sexual assault evidence on a
valveless glass microdevice, in Micro Total Analysis Systems, Proceedings of µTAS 2006 Conference.
T. Kitamori, H. Fujita, and S. Hasebe (Eds.), CHEMINAS, Tokyo, Japan, 1055, 2006.
45 Optical Detection Systems for
Microchips
James M. Karlinsey and James P. Landers
CONTENTS
45.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1253

45.2 Optical Detection Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1254
45.2.1 Laser-Induced Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1254
45.2.2 Absorbance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1255
45.2.3 Chemiluminescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1255
45.2.4 Electrochemiluminescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1256
45.2.5 Raman Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1257
45.2.6 Thermal Lens Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1257
45.2.7 Surface Plasmon Resonance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1258
45.2.8 Refractive Index Detection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1259
45.3 Fluorescence Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1259
45.3.1 Fluorescent Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1259
45.3.1.1 Native Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1259
45.3.1.2 Protein Labeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1261
45.3.1.3 DNA Labeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1262
45.3.2 LIF Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1263
45.3.2.1 Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1263
45.3.2.2 Detectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1264
45.3.3 Multidimensional Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1264
45.3.3.1 Multichannel Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1264
45.3.3.2 Multicolor Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1265
45.3.4 Single Molecule Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1266
45.3.5 Integrated Optics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1267
45.4 Practical Application of Lif . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1268
45.4.1 Experimental LIF Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1268
45.4.2 Alignment of the Optics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1269
45.4.3 Evaluation of the Detection System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1271
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1272
45.1 INTRODUCTION
Several detection techniques adapted for the capillary format have been applied to fluidic microchips,
including optical, electrochemical, and mass spectrometric detection. While the aim of this chapter
is not to review all of these various techniques, interested readers are directed to several recent
discussions presented in the literature.1–6 Instead, the focus will be on optical methods, which will be
1253
described briefly along with some examples—the reader is also referred to the complimentary chapter
on capillary-based systems by Sweedler in Chapter 9. Because of its broad application in microchip
electrophoresis, a significant amount of discussion will be paid to laser-induced fluorescence (LIF)
detection, which has played a key role in many of the microchip advancements achieved to date.
This discussion will include a practical guide to assembling and aligning an LIF detection setup.
Other optical techniques presented include ultraviolet (UV) absorption, chemiluminescence (CL) and
electrochemiluminescence (ECL), Raman spectroscopy, thermal lens microscopy (TLM), surface
plasmon resonance (SPR), and refractive index (RI) detection.
45.2 OPTICAL DETECTION SYSTEMS

The optical detection methods applied to microchip-based analysis are, for the most part, similar
to those employed in capillary electrophoresis (CE). The main differences, however, are that UV
absorbance is not as common and that the microchip format creates the possibility of incorporating
additional functionality during the fabrication steps (especially in plastic devices). Each of the optical
methods described here includes specific requirements, such as the nature of the analyte and the
microchip substrate, and these must be taken into consideration when choosing and implementing a
system.
45.2.1 LASER-INDUCED FLUORESCENCE

Despite the increase in alternative methods of optical detection, laser-induced fluorescence (LIF)
remains the most widely applicable detection method in microchip electrophoresis systems. There
are several reasons why LIF is ideally suited to the microchip format; these include (1) its high
sensitivity, (2) its compatibility with microchannel dimensions, (3) its fast response time, (4) its
noncontact interrogation, and (5) its addressable nature. When Harrison and coworkers7 initially
miniaturized a CE channel onto a planar glass chip in 1992, the fluorescent signals obtained for
fluorescein and calcein dyes were used to evaluate the feasibility of performing CE separations on
a microchip (Figure 45.1). Since then, LIF detection had been used in many different applications
on-chip including single molecule detection and DNA sequencing. A more thorough examination of
various aspects of LIF detection, including setup and alignment of a confocal system, can be found
later in this chapter.
pH 8.5
20 mM Fluorescein
20 mM Calcein
3000 V applied
0 2 3 4 5 6
Time (min)
FIGURE 45.1 First published microchip electropherogram using LIF. A mixture of fluorescein and calcein
(concentrations shown) were injected for 30 s across a separation channel and subsequently electrophoresed
past the detector. (Reprinted from Harrison, D.J., et al., Analytical Chemistry, 64, 1926, 1992. Copyright 1992.
With permission from American Chemical Society.)
Optical Detection Systems for Microchips 1255
45.2.2 ABSORBANCE
Although it is a universal detector in CE and high pressure liquid chromatography (HPLC),
absorbance detection has not found the same success in microchip applications. This is primar-
ily due to the short optical path lengths (i.e., channel depths) and the difficulty in coupling light into
and out of the microchip, resulting in detection limits several orders of magnitude higher than LIF.
One of the earliest examples of UV detection on-chip was the imaging of an isoelectric focusing
(IEF) channel with a charge-coupled device (CCD) camera. Sample was loaded onto a quartz chip
and excited at 280 nm using a Xe lamp. With the benefit of the concentrating effect of the IEF step,
the detection limit of myoglobin was found to be 30 µg·mL−1 .8 In most microchip applications,
however, there is no concentration step and the short optical path lengths present a challenge. An
initial attempt to increase the path length involved the patterning of a silicon flow cell that achieved
path lengths up to 5 mm using successive reflections in a channel 50 µm deep. However, there was
significant amount of loss at each surface.9 This was addressed by another group that fabricated
a multireflection absorbance cell in a glass microchip using a three-mask process to deposit alu-
minum mirrors above and below the flow channel. With entrance and exit apertures patterned into
the mirrors, optical path lengths of 50–272 µm were obtained in the multireflection absorbance cell
(compared to channel depths of 10–30 µm), increasing the sensitivity 5- to 10-fold over single-pass
devices.10 Another approach exploited total internal reflection (TIR) to form a liquid-core waveg-
uide by sandwiching Teflon tubing embedded in a polydimethylsiloxane (PDMS) block between two
silica plates containing fluidic channels. By increasing the optical path length to 5 mm while main-
taining a detection volume less than 1 µL, the limit of detection (LOD) for crystal violet was found to
be ∼1 µM.11
The different approaches for coupling light into the absorbance cell primarily involve optical
fibers and planar waveguides. Using fibers alone, a separation of peptides was shown by Jindal and
Cramer12 using on-chip electrochromatography with a UV absorbance detector. Optical fibers for
excitation and collection were positioned above and below a hybrid quartz/PDMS device, which
featured an optical path length of only 23 µm, yet resulted in a LOD of 167 µM for thiourea.
A similar approach involved embedding a ball lens in a layer of PDMS that sealed against a glass
microchannel. An optical fiber brought incident 668 nm light into the lens, which then collected
the reflected light from a layer of aluminum patterned on the other side of the channel. Calcium
levels were monitored through a reaction with arsenazo III, and a detection limit of 85 µM was
found for calcium before applying the assay to clinical urine samples.13 The alternative approach for
coupling light with the absorbance cell was taken by Kutter’s group, where fiber couplers and planar
waveguides were fabricated into the microchip along with the fluidic channels. Using deep-reactive
ion etching (DRIE) to form high-aspect ratio channels in silicon, an effective optical path length of
1.2 mm was obtained in a U-shaped channel (Figure 45.2). By eliminating propagation loss along
the waveguides, the detection limit for paracetamol at 254 nm was found to be 3 µg·mL−1 , at present
the most sensitive waveguide device in the literature, and a three-component separation was shown
on the device.14 The U-shaped channel design was initially presented by Liang and Harrison15 in
a microchip that featured a dual absorbance and fluorescence cell, with 488 nm light brought into
the device with an optical fiber. This work provided an excellent comparison between UV and LIF
detection on microchip, as detection limits were found to be 6 µM (for fluorescein isothiocyanate)
and 3 nM (for fluorescein), respectively.
45.2.3 CHEMILUMINESCENCE
Chemiluminescence refers to the release of light energy from a chemical reaction. This is based on
the type of reaction first apparent to scientists in the form of “cold” light from fireflies. The firefly’s
glow mechanism, which hinges on the oxidation of firefly luciferin, is incredibly efficient, as 80% of
reacting molecules present generate a photon of light. The inherent benefit of a CL system is that no
Separation
channel
Waste
channel
Waveguides
FIGURE 45.2 Scanning electron micrograph of the U-shaped detection region of a microchip designed for
absorbance detection using integrated waveguides. (Reprinted from Mogensen, K.B., et al., Electrophoresis,
excitation source is required, making it ideally suited to a miniaturized analysis system. Xu et al.16
used this feature, along with the benefit of the reduced reagent consumption in a microreactor chip,
to monitor chromium III from aqueous samples. A detection limit of 100 nM was obtained in the
continuous flow system, where the CL yield was observed to benefit from the short (<1 s) mixing
time. Liu et al.17 examined two model CL systems (the ion-catalyzed luminol-peroxide reaction and
the dansyl species conjugated peroxalate-peroxide reaction) using three different designs in PDMS
microchips, and sub-µM concentrations were detected. Importantly, the identification of three metal
cations in the first system and the chiral recognition of dansyl phenylalanine enantiomers in the second
was shown in <1 min, demonstrating the utility of CL detection coupled with a separation mode.
Some other analytes examined on microchip using CL detection include dopamine and catechol
(detection limits of 20 and 10 µM, respectively),18 glucose (detection limit of 10 µM),19 and the
cancer marker, immunosuppressive acidic protein (detection limit of 100 nM).20
45.2.4 ELECTROCHEMILUMINESCENCE
Electrochemiluminescence detection is similar to CL, except that the luminescent reaction is effected
by electrochemical stimulation. Along with the advantages of CL, the stimulation reaction allows
the time and position of the light-emitting reaction to be controlled. Ru(bpy)2+
3 is one of the most
efficient and thoroughly examined ECL molecules and, while the reaction scheme may change
depending on the analyte, the excited product is always Ru(bpy)2+
3 *, which emits light at a maximum
21
intensity at 620 nm when returning to its ground state. In 1998, a fully integrated probe was
designed that featured a gold interdigitated microelectrode array (IDA) and a photodiode detector on
a silicon microchip to perform ECL. The device was applied to the model system of Ru(bpy)2+ 3 and
tripropylamine, and a 500 nM detection limit was obtained for the ruthenium complex.22 L’Hostis
et al.23 fabricated an ECL detector using a hybrid silicon and SU-8 photoresist device. The detector
(with a platinum IDA) was initially evaluated using a reaction of codeine and Ru(bpy)2+ 3 (detection
limit of 100 µM), and it was then incorporated into an enzymatic microreactor (with a carbon IDA)
to assay glucose (detection limit of 50 µM). Applying an ECL detector to a micellar electrokinetic
chromatographic (MEKC) separation on a glass microchip, a U-shaped floating platinum electrode
was placed across the separation channel to perform both direct and indirect detection.24 Applying
ECL to clinical samples, the antibiotic lincomycin was analyzed using a microchip CE system with an
integrated indium tin oxide (ITO) working electrode (Figure 45.3). Without pretreatment, lincomycin
was analyzed in urine in <40 s with a detection limit of 9.0 µM.25 A benefit of using ECL is that
(a) (b)
4200
1
ITO electrode
Buffer reservoir 3900
ECL intensity (counts)

2
Electrode plate
3600
Separation channel
3300
PDMS layer 3000
2700
Detection reservoir
Sample reservoir
30 45 60 75 90 105
Time (s)
FIGURE 45.3 Microchip design and sample data using ECL detection. (a) The microchip was fabricated in
PDMS using a glass mold and the ITO electrodes were pattered with an etch step. (b) The separated mixture
contains 100 µM each of lincomycin (1) and clindamycin phosphate (2), which were detected using Ru(bpy)2+
3 .
(Reprinted from Zhao, X., et al., J Chromatography B, 810, 137, 2004. With permission.)
it can be combined with EC detection, as demonstrated by Qiu et al.26 in the determination of the
neurotransmitter dopamine and several pharmaceuticals.
45.2.5 RAMAN SPECTROSCOPY

Raman spectroscopy is similar to IR spectroscopy, but superior in microchip systems because water
is virtually Raman transparent. Similar to LIF, sample is typically excited with a laser and spectra
can be collected from a very small volume. A real benefit of Raman spectroscopy is the ability to
monitor structural changes or the progress of a chemical reaction on chip. Pan and Mathies24 used a
chip-based flow experiment to examine the chromophore structure of rhodopsin photointermediates
as well as changes in protein–chromophore interactions in a glass microchip at room temperature.
Fletcher et al.27 used a Raman microscopic spectrometer to observe the formation of ethyl acetate
from ethanol and acetic acid in a glass microreactor (Figure 45.4). Using an excitation wavelength
of 780 nm, it was shown that the spectral intensities of Raman bands specific to each species were
proportional to the concentration present in the microreactor at a given time, providing the time
dependence of the product formation. A similar system was used to optimize the catalytic oxidation
of isopropyl alcohol to acetone in a continuous flow microreactor.28 Ramser et al.29 utilized the
microchip environment to trap a single red blood cell and follow the dynamics of an oxygenation
cycle after environmental stimulation using Raman spectroscopy in real-time. Of the various optical
methods presented in this chapter, Raman spectroscopy possesses the greatest promise as a microchip
detection technique, especially considering the signal enhancement (up to 1014 )30 that can be obtained
performing surface-enhanced Raman spectroscopy (SERS) on a surface that has been coated with
silver or gold.
45.2.6 THERMAL LENS MICROSCOPY

The Kitamori group has pioneered the use of TLM in microchip systems, where both an excitation
and a probe beam are focused into a liquid sample—an example of its utility is provided with
immunoassay in Chapter 34. The energy of the excitation beam is absorbed by the sample species
and results in a localized temperature increase that affects the refractive index (RI) within the medium.
The probe beam, which is selected such that there is no absorption, is subject to the “thermal lens
(a) (b)
FIGURE 45.4 Three-dimensional plots of Raman intensity where two laminar flow streams come together.
The region was probed for specific bonds to identify (a) acetic acid at 893 cm−1 and (b) ethanol at 882 cm−1 .
(Reprinted from Fletcher, P.D.I., et al., Electrophoresis, 24, 3239, 2003. With permission.)
Objective lens
Pump beam
Sample Thermal lens

effect
Probe beam
Filter
aperture
Detector
FIGURE 45.5 Schematic illustration of the principle of the thermal lens effect in TLM. (Reprinted from
Tokeshi, M., et al., Electrophoresis, 24, 3583, 2003. With permission.)
effect” that results from the change in RI due to the increased temperature (Figure 45.5). Because
the probe beam is modulated by RI changes, the resulting change in intensity can be recorded with
a photodetector. The sensitivity of this detection method has been reported at subsingle-molecule
detection for a nonfluorescent species. To demonstrate this, a test sample of Pb(II) octaethylporphyrin
was prepared in benzene and analyzed in a 100 µm deep quartz microchannel. The concentrations
used corresponded to 0.4–3.4 molecules with a detection limit of 0.34 (and an average temperature
rise for a single molecule of 3.1 µK).31 Another impressive demonstration of the TLM technique
was in the two-dimensional imaging of a single cell to monitor the release of cytochrome c during
the apoptosis process with a spatial resolution of ∼1 µm.32 A review of the advances made by
the Kitamori group can be found in the literature,33 but it should be noted that another group has
recently developed a portable thermal lens spectrometer and shown a 30 nM limit of detection for
xylenecyanol in a fused silica microchip.34
45.2.7 SURFACE PLASMON RESONANCE

Surface plasmon resonance is a surface-sensitive method for chemical sensing based on RI changes
on the surface of a metal film. Variations in light intensity reflected from the back of the film are
detected, thus, no labeling is required. Because the SPR signal is strongest when measuring near
the sensor surface, it is often used to detect the surface binding characteristics of biomolecules.
Yager and coworkers35 developed a special microscope for performing SPR detection, and applied
it to characterize protein absorption on the microchannel surface of Mylar devices. An amplification
technique was then developed using a precipitating enzyme to increase the signal reflectivity by up to
70%.36 Furuki et al.37 sputtered gold onto glass before sandwiching a UV resin between glass layers
to define a microchannel. The gold surface was then chemically modified with a photobiotin layer to
promote the specific adsorption of avidin at a concentration of 25 µg·mL−1 , probing with a 670 nm
laser diode. PDMS microchips have also been used for SPR, and have been demonstrated for protein
binding to a gold surface in both flow-through microchip (using a commercial SPR detector)38 and
microarray39 formats. A review of SPR technology and applications can be found elsewhere.40
45.2.8 REFRACTIVE INDEX DETECTION

Refractive index detection is based on the factor by which light is slowed down (relative to vacuum)
as it travels through a medium. RI is sensitive to changes in temperature, pressure, and flow rate, and,
as previously discussed, plays a role in other detection techniques. Before those methods, however,
a holographic RI detector was applied to the separation of carbohydrates on chip.41 In the setup, an
incident laser beam passed through a holographic optical element that divided the beam into two. One
beam served as a reference and passed solely through the glass substrate while the probe beam passed
through the solution in the separation channel. A photodiode array was placed to collect the fringe
pattern from the interfering beams, and this pattern changed whenever an analyte passed through the
separation channel (Figure 45.6a). Using this system, a proof of principle separation of three sugars
was performed in less than 17 s, although their concentration was high (33 mM). An alternative
method was demonstrated by Costin and Synovec,42 who detected the angular deflection of a diode
laser beam incident on adjacent laminar flow streams. The beam entered the PDMS microchannel
orthogonal to the direction of flow and the diffusion gradient of the two streams, and the resulting
angle of the beam corresponded to a RI gradient (Figure 45.6b). The device was evaluated as a
molecular weight sensor for poly(ethylene glycol) solutions and was found be sensitive to changes
in RI of 4.5 × 10−6 .
45.3 FLUORESCENCE DETECTION

The main considerations when performing LIF detection include the target analyte, the instrumen-
tation, and the microchip, and these will be discussed in the following sections. Some advances in
microchip-based LIF will also be presented, including multichannel and multicolor detection, single
molecule detection, and integrated optics.
45.3.1 FLUORESCENT MOLECULES

While one of the limitations of LIF detection is that few target molecules exhibit native fluorescence,
especially biologically relevant proteins and DNA of clinical interest, several different labeling
approaches both on- and off-chip have been demonstrated. For the former, the microchip fabrication
steps provide the ability to incorporate additional structures into the design, in many cases without
adding more steps (or cost) to the fabrication process. Examples include the additional channels and
reaction chambers used to perform both precolumn43 and postcolumn44 labeling of amino acids.
45.3.1.1 Native Fluorescence

There are few examples in the microchip literature that feature analytes of interest exhibiting native
fluorescence, such as the phycobiliproteins present in cyanobacteria and some algae.45,46 Recently,
(a) (b)
Laser beam
Holographic
optical element Sample
inlet
Reference Probe beam
beam
Mobile
Microchip PSD
phase
inlet θ
Deflected
Channel
beam
Interference Laser X
pattern beam
Y Z
To outlet
Photodiode array
FIGURE 45.6 Different experimental approaches for RI detection. (a) Principle of holographic RI detection,
where a reference and probe beam interact to produce an interference pattern that changes when an analyte
passes through the channel. (From Burggraf, N., et al., Analyst, 123, 1443, 1998. Reproduced by permission of
The Royal Society of Chemistry.) (b) Illustration of angular detection as a gradient forms between two different
laminar flow streams to create an RI gradient. (Reprinted from Costin, C.D. and Synovec, R.E., Analytical
Chemistry, 74, 4558, 2002. Copyright 2002 American Chemical Society. With permission.)
however, deep UV fluorescence detection has been demonstrated on microchip using 266 nm exci-
tation (using a frequency quadrupled Nd:YAG laser). Hellmich et al.47 demonstrated UV-LIF of the
amino acid tryptophan (detection limit of 17 µM) and a protein mixture of 500 µM lysozyme C
and 125 µM avidin in hybrid PDMS/glass devices. With a similar system, Schulze et al.48 demon-
strated UV-LIF of several different proteins including a mixture of lysozyme (900 nM), trypsinogen
(500 nM), and chymotrypsinogen (500 nM) using fused silica devices.
Still, the majority of literature describing naturally fluorescent molecules analyzed on-chip deal
primarily with novel device, separation, and detection technologies. In these studies, fluorescein
and rhodamine dyes are most common because they can be readily obtained in various derivatized
forms.49 For example, our group used a combination of fluorescein and carboxy-X-rhodamine dyes
in an electrophoresis experiment to evaluate an on-chip pressure injection,50 and later added rho-
damine 6G to the test mix to evaluate a scanning multicolor detection platform.51 Also, fluorescent
dyes can be used to perform indirect fluorescent detection of nonfluorescent species, as demonstrated
by Munro et al.52 by adding fluorescein to the run buffer to analyze amino acids in urine. Sirichai
and de Mello53 also added fluorescein to the run buffer to perform quantitative analysis on a com-
mercial photographic developer solution and reported detection limits of 5 µg·mL−1 . Recently, deep
UV fluorescence detection has been demonstrated on microchip using 266 nm excitation (using a
frequency quadrupled Nd:YAG laser). Hellmich et al.47 demonstrated UV-LIF of the amino acid
tryptophan (detection limit of 17 µM) and a protein mixture of 500 µM lysozyme C and 125 µM
avidin in hybrid PDMS/glass devices. With a similar system, Schulze et al.48 demonstrated UV-LIF
of several different proteins (Figure 45.7) including a mixture of lysozyme (900 nM), trypsinogen
(500 nM), and chymotrypsinogen (500 nM) using fused silica devices. This technique is promis-
ing because it offers the benefits of LIF detection while enabling native detection of biomolecules,
HN
(a) HO (b)
1
NH2 O
(2) 2 3
1 4
OH
Fluorescence
(1) 2 4
3 6 9
OH
COOH
O OH
3 NH2
(3) N
H
(4)
3 6 9 12 Time (s)
FIGURE 45.7 Peaks collected using deep UV-LIF detection. The electropherogram in the foreground rep-
resents a mixture of (1) serotonin, (2) propranolol, (3) 3-phenoxy-1,2-propanediol, and (4) tryptophan, each
at a concentration of 40 µg·mL−1 . The same mixture is shown in the inset with concentrations closer to their
limits of detection. (Reprinted from Schulze, P., et al., Analytical Chemistry, 77, 1325, 2005. Copyright 2005
American Chemical Society. With permission.)
however, there are greater demands on the optics and microchip substrate to provide high UV
transmittance with low autofluorescence.
45.3.1.2 Protein Labeling

As mentioned earlier, the ability to integrate precolumn and postcolumn labeling of amino acids
brought an increased functionality to the microchip format that is not readily achieved on capillary
alone. The initial report of a postcolumn reactor contained a mixing tee after the separation col-
umn where an additional reagent reservoir, separated from the reservoirs used for the injection and
separation steps, electrokinetically introduced the fluorescent tag o-phthaldialdehyde (OPA). The
OPA then mixed with the partially separated amino acids by diffusion (Figure 45.8a) in the reac-
tion column as the separation continued.44 With the precolumn reactor, the analytes and OPA were
introduced into the same chamber prior from separate reservoirs before diffusing into the separation
channel in a gated injection scheme (Figure 45.8b).43 Both integrated techniques, performed in glass
devices, enabled the labeling and separation to occur in very small volumes (<1 nL) in a matter of
seconds, demonstrating the benefits of using microfluidic analysis systems. The same group later
used a postcolumn reactor to label proteins noncovalently with a NanoOrange dye, which offered
fast reaction kinetics and high fluorescence yield.54 This dye also proved effective for dynamic on-
column labeling, resulting in a detection limit of 500 ng·mL−1 for bovine serum albumin.55 Another
precolumn labeling reaction was demonstrated on a glass/PDMS device using dichlorotriazine fluo-
rescein (DTAF), a derivatizing agent for biogenic amines. In addition to fast reaction kinetics (<60 s
for on chip mixing and reaction), detection limits were reduced to 1 nM.56
(a) (b)
Analyte Reagent
Analyte
reservoir reservoir
reservoir
Buffer Analyte
reservoir waste
2.7 mm reservoir
Buffer Analyte
7.0 mm 6.6 mm reservoir waste
reservoir
Injection Reaction
Separation Injection
cross chamber
7.0 mm column cross
Mixing tee
11.6 mm
Reaction
Separation
10.8 mm column
column
Reagent
Cover Cover
reservoir
Waste slip slip
reservoir Waste
reservoir
10 mm Substrate 5 mm Substrate
FIGURE 45.8 Integrated reactors for amino acid labeling. (a) Schematic of the microchip layout for post-
column labeling, incorporating a reactor column after the separation. (Reprinted from Jacobson, S.C., et al.,
Analytical Chemistry, 66, 3472, 1994. Copyright 1994 American Chemical Society. With permission.) (b)
Schematic of the layout for precolumn labeling, incorporating a reactor chamber prior to injecting sample into
the separation channel for analysis. (Reprinted from Jacobson, S.C., et al., Analytical Chemistry, 66, 4127,
1994. Copyright 1994 American Chemical Society. With permission.)
45.3.1.3 DNA Labeling

Laser-induced fluorescence is the most common detection method in DNA sizing and sequencing
applications, on both capillary and microchip, owing to the different ways to incorporate fluo-
rophores. In double-stranded DNA (dsDNA) separations, intercalating dyes are typically added to
the separation matrix. Upon binding the DNA strand, the dye exhibits a significant increase in
fluorescence yield over unbound dye. While a wide variety of nucleic acid stains are available,
depending on the application,49 the protocol for dsDNA separations in our lab involves adding
the monomeric cyanine dye, YO-PRO® -1 iodide (Invitrogen), to the sieving matrix at a 0.1% v/v
concentration.57 For single-stranded DNA (ssDNA) separations, fluorescently tagged primers are
incorporated into the target DNA by the polymerase chain reaction (PCR) method. This allows
multiple fluorescent labels to be incorporated in a single application step and is commonly used
in four- and five-color sequencing and genotyping applications. Derivatization chemistries enable
conjugations of most dyes to oligonucleotides,49 and fluorescent labels can often be added (for
a fee) when ordering DNA primers. In many commercial multiplexed PCR amplification kits (i.e.,
Applied Biosystems AmpFlSTR® kits) proprietary dyes are used, but the Mathies group has designed
and synthesized fluorescent oligonucleotide primers using the principle of fluorescence resonance
energy transfer (FRET),58 and have used their dyes in sequencing59 and genotyping60 applications on
microchip.
APD
achr
Filter
PH
λ /2
LF
pol
DB
Ar + laser M
M
M
MO
CE
FIGURE 45.9 Schematic of a confocal LIF detection system with source, optics, and detector shown. The
optics include mirrors (M), laser line filter (LF), half-wave plate (λ/2), polarizer (pol), dichroic beamsplitter
(DB), microscope objective (MO), pinhole (ph), filter, and achromat lenses (achr). The source shown is an
argon ion laser, and the detector is an avalanche photodiode (APD). While the electrophoresis channel shown
here is in a capillary (CE), the system could be readily applied to a microchip. (Reprinted from Johnson, M.E.
and Landers, J.P., Electrophoresis, 25, 3515, 2004. With permission.)
45.3.2 LIF INSTRUMENTATION

A typical method for performing LIF detection on a microchip is to bring the excitation light into
the microchannel orthogonal to the device plane, which is relatively simple to align, and to then
focus the beam into the channel using a microscope objective or some combination of achromat
lenses. LIF emission is then collected with a separate objective or lens combination, spatially and
spectrally filtered, and recorded on a photodetector. Confocal detection systems are quite popular
in the literature, first demonstrated by Guzman and coworkers61 for LIF detection on a capillary
in 1991, where the focusing and collection optics are the same and the excitation and emission
wavelengths are separated with a dichroic beamsplitter (Figure 45.9). Variations in LIF detec-
tion setup exist, however, and He et al.62 present a detailed illustration of their laser-induced
detection system where the excitation is brought into the microchannel from an angle. A recent
review on the fundamentals and practice of LIF detection in microanalytical systems presents a
thorough examination of ultrasensitive systems (sub-pM concentration).63 While the goal of this dis-
cussion is not necessarily to achieve ultrasensitivity, the principles and considerations are very much
the same.
45.3.2.1 Sources
The most commonly used excitation sources for LIF detection on microchip are helium-neon and
argon ion gas lasers, both of which feature various emission lines. Typical laser powers range from
0.5 to 50 mW. Benefits of these lasers include their beam quality and long lifetimes, but they are
still costly and bulky relative to the microchip scale. There have been other reports of excitation
sources, including dye lasers (i.e., rhodamine 6G), metal-vapor lasers (i.e., HeCd), and solid-state
lasers (i.e., Nd:YAG), and semiconductor lasers (i.e., laser diodes), but these still present size and
cost issues. What is promising is the recent use of light-emitting diodes (LEDs) for LIF detection
because, in addition to being small and cheap, they can be incorporated into the devices themselves.
It is important to note that there are several probes and dye labels available that are ideally suited to
the different laser excitation wavelengths.49
45.3.2.2 Detectors
Photomultipliers tubes (PMTs) remain a common choice for detection. PMTs consist of a pho-
tocathode, a series of dynodes, and an anode. Incident photons strike the photocathode surface,
causing electrons to be emitted by the photoelectric effect. These electrons are accelerated toward
the first dynode, which is held at a positive potential, and collide to release additional electrons.
This process is repeated several times, with each dynode held at a higher potential, until the final
accumulation of cascading electrons (multiplying a signal by orders of magnitude) reach the anode.
This results in a sharp current pulse indicating that photons have been detected by the PMT. While
PMTs are quite common as photodetectors for microchip LIF, some of the other detectors that have
been used include avalanche photodiodes (APDs) and CCD arrays. While the former is most often
used to achieve ultrasensitive detection,63 the latter is typically used in whole channel imaging and
multidimensional applications, which will be described next.
45.3.3 MULTIDIMENSIONAL APPROACHES

45.3.3.1 Multichannel Detection
While several different glass types have been reported in the literature for microchip separations, they
are all compatible with LIF detection due to their optical transparency. This is significant because it
creates the opportunity to interrogate any region of the microchip, offering more flexibility than the
capillary format. Because the glass capillaries used in CE are coated with a polymer layer that must be
removed to expose the glass window for excitation, there is inherent limitation to where detection can
be performed. In addition to performing LIF detection at a single location, several techniques have
been shown for multipoint (within a single channel) or multichannel detection. The former includes
the incorporation of waveguide structures into the microchip design to create several excitation points
along a single channel, which was used to perform velocity measurements of particles flowing a
microchannel.64 Several approaches for multipoint detection have been demonstrated for microchip
IEF as an alternative to mobilizing resolved peaks past a fixed detection point.65 Similar to the
approach for detecting IEF by UV absorbance,8 Yao et al.66 collected the fluorescence signal from
an IEF separation using a CCD array for whole column imaging. Interestingly, the excitation source
was also an array, featuring organic LEDs. As opposed to imaging the entire IEF channel, Raisi
et al.67 designed a scanning fluorescence detector based on a computer-controlled translation stage
with a single PMT and were able to scan the entire separation channel every 9 s to collect temporal
and spatial measurements. It is ironic this approach parallels that devised by Hjerten68 and reviewed
more than half a century ago. Another scanning approach using a single PMT is presented using an
acoustooptic deflector (AOD) system (described later) to sweep the excitation beam across an IEF
channel during the focusing step.69
Considering the capabilities of current capillary instrumentation, which routinely handle bundles
of capillary at a time, the ability to increase the number of parallel analysis channels and, therefore,
sample throughput is arguably one of the biggest challenges (and successes) in the microchip field.
Modest increases in throughput have been achieved featuring both moving and nonmoving parts for
beam displacement to address multiple channels. Cheng et al.70 used a mirror attached to a galvano-
scanner to raster an excitation beam across a six-channel device designed to perform independent
immunoassays simultaneously (Figure 45.10). The galvano-scanner changed its position in an arc
in response to a change in current. Huang et al.71 used an AOD to change the diffraction angle of an
incident laser beam to address an eight-channel device designed for parallel DNA separations. The
AOD featured no moving parts, relying solely on the interaction of the laser beam and an acoustic
Photo
multiplier
tube
Emission
interference
filter
Optical
scanner
"Kick-down"
Microscope mirror
objective Galvo mirror
SPIDY device
Laser beam
FIGURE 45.10 Multichannel approach for LIF detection on a microchip. A mirror attached to a galvano-
scanner was used to position the excitation beam so that all eight channels on the device (named SPIDY) could
be addressed to monitor parallel immunoassays. (Reprinted from Cheng, S.B., et al., Analytical Chemistry, 73,
1472, 2001. Copyright 2001 American Chemical Society. With permission.)
wave propagating through the crystal to scan across the channels, and exhibited fast response times
(∼200 ns). Most impressive, however, are the advances made by the Mathies group with their radial
devices for DNA analysis, which have achieved parallel processing of 9672 and 38473 samples. Using
circular devices, with analysis channels extending from the center, the detection points can be kept
close together, and a circular scan path enables high scan rates and positional accuracy. A description
of the radial scanner, which includes a rhomb prism at the top of a hollow rotating shaft, can be
found in the literature.74
45.3.3.2 Multicolor Detection

In multicolor LIF experiments, the most common technique for sampling at multiple emission
wavelengths is to add additional photodetectors along with the appropriate dichroic beamsplitters
to separate the fluorescent signal. This is the approach taken by several groups to perform multi-
DNA sequencing75 and genotyping76 on microchip. Despite the existence of commercial capillary
instruments that utilize a spectrograph to disperse the emission wavelengths onto a CCD array,
this technique has seldom been reported in the literature for microchip LIF detection. However,
Backhouse et al.77 used a diffraction grating to disperse the LIF emission onto a CCD array for
four-color sequencing, and Simpson et al.78 used a transmission imaging spectrograph with a wide
imaging area to disperse LIF emission from a two-color genotyping experiment run in 48 parallel
channels onto a CCD array. In the CCD systems, the diffracted emission wavelengths are displaced
and then binned in rows or columns of CCDs. Recently, a system was developed to complement the
microchip platform by featuring a single acoustooptic tunable filter (AOTF) and a single PMT.51 The
filter behaves similar to the previously described AOD, except that when an acoustic wave traveling
through the crystal at a specific frequency interacts with the light, only select wavelength bands are
diffracted by the crystal and collected by the PMT. Capable of addressing several wavelengths with
only a single photodetector (19 are shown in Figure 45.11), an AOTF-based detection system has
20
18
16
14
FL
12
Time (s)
10
6
ROX
R6G
4
0
535 555 575 595 615
Wavelength (nm)
FIGURE 45.11 Three-dimensional multicolor data collected with an AOTF. A mixture of three dyes (FL:
fluorescein, R6G: rhodamine 6G, and ROX: carboxy-X-rhodamine) was injected and separated on chip while
scanning the AOTF through 19 wavelengths. (Adapted from Karlinsey, J.M. and Landers, J.P., Analytical
Chemistry, 78, 5590, 2006.)
been shown for four- and five-color genotyping applications on microchip.79 In all of these cases,
the appropriate optics are placed after the pinhole in the confocal detection setup.
45.3.4 SINGLE MOLECULE DETECTION

The earliest report of single-molecule detection on microchip was in 1998, with concentration detec-
tion limits of 1.7 pM for rhodamine 6G and 8.5 pM rhodamine B separated in 10 µm deep glass
channels in less than 35 s. The fluorescent molecules were excited with the 514 nm line from an
argon ion laser at 540 µW, a 100× objective was used in a confocal setup, and a single-photon
avalanche diode was used to collect the signal.80 Shortly afterward, single-molecule detection was
applied to DNA separations, focusing the sample through a 1 µm diameter focused laser beam.
Using the 488 nm line from an argon ion laser and a 100× objective in a confocal microscope,
thiazole orange intercalator was excited to detect single DNA molecules during the separation of a
100–1000 bp DNA sizing ladder.81 Foquet et al.82 patterned nanochannels in fused silica, creating
channel volumes ∼100 times smaller than the observation volumes obtained using conventional
confocal optics. This enabled single-fluorophore detection to be achieved at a higher concentration
(2 nM) than typically reported (pM). Although the examples presented here involve LIF detection,
single-molecule detection using TLM has also been demonstrated (see earlier).31
45.3.5 INTEGRATED OPTICS

Recently, several groups have integrated optical components into their devices, and some of the
more impressive examples are presented here. In some cases, no additional fabrication steps are
required, as the optics are patterned along with the fluidic features. After pioneering microchip
fabrication in PDMS, the Whitesides group assembled an integrated fluorescence detection system
including an optical fiber embedded in the patterned fluidic layer to couple excitation light into the
channel, a microavalanche photodiode (µAPD) embedded in a second PDMS layer, and a colored
polycarbonate filter placed between the two layers (Figure 45.12). The device was then used to
detect proteins and small molecules separated on the chip, and a detection limit of ∼25 nM was
obtained for fluorescein solution.83 A microoptical system was designed in glass that involved the
additional patterning of microlens arrays and aperture arrays deposited on both sides of the chip
to enable the detection of a 3.3 nM dye solution, comparing favorably with confocal systems.84
Polystyrene chips were fabricated by Maims et al.85 with an embossed diffractive element to obtain
fluorescence and absorption spectra using laser and LED sources. One research group integrated an
optical waveguide network with fluidic microchannels to perform particle sorting.86 Optical trapping
was performed by focusing a femtosecond pulsed laser into fused silica to form the waveguide, and the
(a)
PDMS
Detail in B.
microfluidic system
Blue LED
Optical fiber
µAPD in PDMS
Polymeric
optical filter
Personal Active quenching

HP counter circuit
computer
(b) Microfluidic channel

50 µm µAPD
Optical fiber
140 µm
50–100 µm
30 µm
FIGURE 45.12 Diagrams for a PDMS chip containing integrated optics. (a) The microchip consists of three
layers containing the fluidics, a filter, and an APD detector. Light from an LED source is coupled in the microchip
through an optical fiber embedded in the fluidic layer. (b) Top-down view of the microchip that shows the spatial
relationship between the channel, detector, and optical fiber. (Reprinted from Chabinyc, M.L., et al., Analytical
Chemistry, 73, 4491, 2001. Copyright 2001 American Chemical Society. With permission.)
trapped particles (exhibiting different fluorescence signals) were then sorted into different flow paths.
Miniaturized spectrometers have been assembled by combining molded elastomeric microchannels
with filtered silicon detector arrays, and a solution of Bromophenol Blue was detectable down to
the sub-µM range by monitoring its absorption at 588 nm.87 Thin-film photoelements have been
incorporated into PDMS devices to serve as both excitation sources and detectors. Using a confocal
detection setup, a thin-film polyfluorene LED was used to excite fluorescein and carboxyfluorescein
dyes in an electrophoresis experiment, with a detection limit of 1 µM.88 A thin-film organic diode
was implemented as a photodetector for a CL assay on chip, with a preliminary detection limit
of ∼1 mM.89
45.4 PRACTICAL APPLICATION OF LIF

The following sections are designed to assist someone interested in assembling a microchip LIF
detection setup, or merely to help reinforce some of the considerations for those already collecting
data. It is by no means the only way to perform microchip LIF, and it is probably not the best. However,
one should be able to routinely achieve 1–100 nM detection limits on the system described herein
using a fluorescein dye in a typical CE buffer system.
45.4.1 EXPERIMENTAL LIF SETUP

The confocal epifluorescent detection scheme we use is common among many who use this detec-
tion mode90 and features a cube-and-rail assembly system, specifically the microbench system from
LINOS Photonics (Milford, MA) on an optical breadboard to maintain proper alignment of compo-
nents (Figure 45.13). The excitation source is a multiline argon ion gas laser (model Reliant 150 m,
Laser Physics, West Jordan, UT) that features user-selectable wavelengths (457, 488, and 514 nm)
FIGURE 45.13 Confocal epifluorescent detection setup. Laser excitation (dark shading) enters the cube
assembly from the steering mirrors and is reflected off of the dichroic beamsplitter before being focused into a
microchannel. LIF emission (light shading) is then collected with the objective, passed through the beamsplitter
and off a mirror before focusing onto a pinhole. The light passing though the pinhole and subsequent filter is
finally collected by the photodetector. The microchip and microchip stage, which should feature translation in
the z-axis and at least one of the x or y-axes, have been left out of the diagram. Frontal views are provided
of some components when their features are not clear in the side-view. (1: mounting post, 2: steering mirror,
3: iris diaphragm, 4: cube assembly, 5: dichroic beamsplitter on 45◦ adapter (with three positioning knobs),
6: objective adapter, 7: microscope objective, 8: mirror on 45◦ adapter, 9: 90◦ adapter plate, 10: pinhole holder
with x-y positioning, 11: emission filter holder, 12: PMT detector, 13: rail, 14: rail mount.)
and can be operated in multiline mode up to 150 mW. The excitation wavelength enters normal to
one of the cubes and is incident on a dichroic beamsplitter tilted at a 45◦ angle. The dichroic is
selected with a particular cutoff wavelength, below which light is reflected and above which light is
transmitted. Reflected excitation light is directed into the back end of a microscope objective where
it is then focused into the microchannel. The fluorescence emission that is collected by the same
objective is then transmitted through the dichroic beamsplitter and focused onto a pinhole placed at
the appropriate distance from the objective. The pinhole serves as the spatial filter in the image plane
of the objective and defines the collection volume. An appropriate emission filter is placed after the
pinhole, and the filtered emission light is then incident on the fluorescence detector where signal is
collected for analysis. Both the dichroic beamsplitters and emission filters we use are obtained from
Omega Optical (Brattleboro, VT), and the pinhole is from National Aperture (Salem, NH).
In a true confocal system, a point source for excitation is imaged onto the sample and the resulting
fluorescence is imaged onto a spatial filter, although most microchip systems use a laser beam as the
source. Several discussions regarding the selection of optics, including the objective and pinhole,
can be found in the literature.63,90 The standard microchip LIF setup used in our lab features a 488
nm excitation beam diameter of 2 mm and a 16× magnification objective (model 04 OAS 012,
Melles Griot) with a focal length of 10.8 mm and a numerical aperture (NA) of 0.32, yielding a
theoretical spot size of 3.4 µm. Aberrations, however, such as those arising from RI mismatches at
the air/glass and glass/buffer interfaces, enlarge the effective spot diameter, but our channel widths
are typically 30 µm across at their narrowest point (due to the narrowest linewidths we can pattern
on our photomasks). Assuming a typical isotropic etch depth of 40 µm, a hemispherical channel is
formed with a bottom width of 30 µm (the initial line width on the photomask) and a top width of
110 µm (the initial line width in addition to the etch in each direction). Imaging the channel onto
a 1 mm pinhole using the 16× objective, therefore, probes a 62.5 µm spot, which is adequate for
sampling from the middle of the channel. When increased sensitivity is desired, a 40× objective
(model 440864, Zeiss) with an NA of 0.6 and a correction ring to adjust for cover glass thickness is
used, experimentally increasing the signal almost two-fold. By switching objectives, the excitation
spot size increases to 7 µm and the collection spot decreases to 25 µm, and so the channel is probed
more efficiently.
45.4.2 ALIGNMENT OF THE OPTICS

Alignment of the optical system is critical to the LIF detection, and a technique for setting up and/or
aligning a cube-and-rail system is presented here. While this technique is not the only approach, it is
relatively straightforward and easy to adapt for various applications. Before alignment, it is helpful
to assemble the cube system containing the dichroic beamsplitter and mirror cubes (both containing
adjustable 45◦ adapters), objective adapter, emission filter holder, 90◦ adapter plate, pinhole holder
with x-y positioning, and appropriate rails and rail mounts.
1. The excitation beam must first be aligned parallel to the optical table or breadboard. A way
to insure that the height is maintained along the table is to position an iris diaphragm at
an appropriate height (i.e., where the light exits the laser), and alternately move it along
the beam path from a point close to the laser output to a point as far away as possible,
while still on the table. This works best if the beam path follows the predrilled holes on the
optical table. Depending on the laser system, this may involve leveling a box enclosure
or adjusting the mounts on a tube enclosure.
2. Once the beam is parallel, a set of two steering mirrors are used in tandem to position the
beam at a height that matches the center of the top cube containing the dichroic beamsplit-
ter. The mirrors should be mounted at 45◦ angles on mirror mounts with adjustable pitch
and should be placed so that the first mirror reflects the excitation beam perpendicular to
the table and the second reflects the beam parallel to the table, with the beam incident at
the desired height. The steering mirrors are then adjusted to align the beam parallel to the
table, again using the iris diaphragm. The iris should be centered at the height of the center
of the top cube and moved alternately between two holes on the optical table, one close to
the steering mirrors and one further away. The pitch on the initial mirror mount should be
adjusted while the iris is located close to the steering mirrors, and the pitch on the second
mirror should be adjusted after moving the iris further away. This step is repeated until the
beam is parallel at the appropriate height to enter the cube. The iris diaphragm can then
be fixed in place before the entrance of the cube assembly with the iris diameter adjusted
slightly smaller than the beam.
3. The excitation beam should be normal to the cube assembly, and this is tested by placing a
microscope slide across the face of the cube. A mirror held against the slide should reflect
the excitation beam back through the iris. Once the cube assembly is positioned, it should
be fixed to the table. By maintaining an excitation path parallel to the holes in the optical
table, the cube assembly should already be centered using the appropriate rail mounts.
4. With the microchip objective removed, the dichroic beamsplitter should be adjusted so
that the excitation beam exits the assembly perpendicular to the table. This will put the
dichroic at the desired 45◦ . The position of the beam on the dichroic beamsplitter should
then be checked using a mirror above the objective adapter to direct the excitation beam
back out of the cube assembly. This is easily observed with the iris still in place. If the
beam is not centered, the dichroic mount can be moved back and forth along the beam
path. This may require the use of a washer between the cube and the mount to back it out
sufficiently.
5. The alignment of the dichroic and the microscope objective must then be checked. The
best way to check this is to place a card with a hole in it above the beam exiting the top
of the cube assembly, leaving enough space for the objective to be put in place. After
lining the hole up with the beam, the objective should be placed in the adapter. With
proper alignment, the beam exiting the objective will be centered on the hole. If the
beam does not line up, the dichroic adapter should be adjusted accordingly. After this is
done, step 4 should be repeated and the alignment of the microscope objective should be
rechecked.
6. Once the dichroic beamsplitter is positioned, the mirror adapter in the bottom cube should
be adjusted to direct the emission beam parallel to the table as it exits the cube assembly.
Because of the filtering behavior of the dichroic, this step is most easily performed with
a multiline laser (with a mirror above the objective adapter and the microscope objective
removed), but a concentrated solution of fluorophore can also be used to generate a
strong signal for alignment (with the objective in place and the beam incident on a filled
microchannel). Since the dichroic is aligned, the beam incident on the mirror is normal
to the table and so the mirror should be adjusted for a 45◦ . This is done by setting an iris
diaphragm to the height of the center of the bottom cube, and placing it alternately close
to the assembly and further away, while adjusting the adapter knobs until the beam is
parallel. Because the 45◦ adapter has three adjustable knobs, only two should be adjusted
at a time to ensure that all three are used to align the exit beam.
7. The x–y positioner with the pinhole should then be placed at the appropriate distance from
the shoulder of the microscope objective (the top of the objective adapter). This distance
is the tube length (typically 160 mm, specified on the objective housing) minus 10 mm
(that accounts for the placement of the eyepiece in finite-corrected light microscopes).
Once the pinhole is in position (in this case, 150 mm from the shoulder of the objective)
a fluorophore-filled microchannel should be placed above the objective with the focused
excitation beam in the channel. Adjusting the z-axis should increase or decrease the size
of the focused emission beam on the pinhole. The x- and y-axes should be adjusted on
the positioner to align the pinhole. The easiest way to do this is to adjust the z-axis until
the outer fringes of the focused emission beam are visible around the pinhole, allowing it
to be centered easily.
8. After all of the key components are aligned, the emission filter and detector, typically
a PMT, can be placed in the setup. Depending on the detector, additional optics, such
as focusing lenses or a spectrograph, may be placed after the pinhole. The setup should
then be covered to reduce background signal from stray light. The main source of back-
ground signal is scattered excitation light, which can be blocked with a notch filter. This
type of filter is very expensive, however, and should only be necessary for ultrasensitive
fluorescence detection experiments.
The benefit of using fluorescence detection is that it is path length independent and, therefore, the
microchannel dimensions (which commonly range from 5 to 200 µm) do not limit the effectiveness of
the detection. However, it is important that the probe volume is located within the channel. While the
lateral position of the focused beam is readily identified by eye, observing the sidewalls of the channel
that scatter the light and provide an overhead image like a contour map, the depth of the focus within
the channel is more difficult to discern. The most sensitive alignment in the z-axis is achieved by
observing a fluorescent signal while adjusting the z-position. In some applications, such as when using
a DNA intercalating dye that exhibits a weak fluorescent signal when unbound, there is enough signal
present to maximize the emission signal. However, in most cases, the separation channel is filled with
run buffer or polymer matrix that does not exhibit a fluorescence signal. Instead of optimizing the
z-position by adding a fluorophore to the analysis channel, and therefore contaminating the detection
region, it is recommended that a separate alignment channel, characterized by the same width and
etch depth as the separation channel, be included in the microchip design. Once the z-position of
the microchip stage is tuned using the alignment channel, the chip can then be repositioned with the
beam focused in the analysis channel.
45.4.3 EVALUATION OF THE DETECTION SYSTEM

One of the more problematic issues when working with glass microchips is the glass–glass bonding
step, usually performed by cleaning the bonding surfaces and thermally annealing at high temper-
atures (>600◦ C) between ceramic plates with some added weight to generate pressure. Not only is
this a time-consuming step (typically performed overnight), but it can also introduce optical defects
in the glass due to the combination of high temperature and pressure. It is not unusual to observe
“frosting” in the glass following one or more bonding cycles. In regions of the chip where the glass
is frosted or clouded, it is difficult to both focus the excitation beam and collect the fluorescence
emission. It is recommended, following thermal bonding, that the analysis channels in a microchip be
inspected using the LIF detection system before running any samples, in order to identify appropriate
windows for interrogation.
When evaluating a detection setup, it is most efficient to do so in the context of a given set
of experimental conditions (i.e., device dimensions, buffer system, etc.). These conditions can be
standardized within a research group or lab setting for comparison between users and instruments,
as well as to evaluate any changes to a system. There may be times, however, when a specific
application (i.e., DNA sequencing) merits its own evaluation of the LIF setup. In these cases,
it is important to maintain similar conditions in evaluating the detection as will be implemented
in the analysis. For example, an off-chip desalting of PCR product before being loaded into the
sample well on a microchip will have a significant effect on the amount of sample that gets injected
into the separation channel. In this case, the change is not made to the detection setup, but the
fluorescence emission collected from the sample will be very different compared to a test sample
that may not have been desalted.

Using primarily LIF detection, the microchip community has achieved the analysis speeds and
reagent reductions proposed some 15 years ago, and the process of miniaturizing and integrating
optical components onto and around the microchip has already begun. While a universal detector is
yet to be developed, researchers have increased the capability of LIF detection systems by integrating
additional complexity on the devices themselves, incorporating multiple reaction steps as well as
optics. At the same time, the number of microchip analysis techniques is steadily growing. According
to SciFinder (which offers access to the Chemical Abstracts database), the number of publications
per year regarding microchip detection techniques doubled in the years 2003–2004 and is currently
holding steady. Using the optical methods presented in this chapter along with various other forms
of chip-based analysis, including electrochemical detection, mass spectroscopy, and even nuclear
magnetic resonance, it will be the responsibility of the microchip community to find new applications
and elucidate more processes in the future as we increase our ability to control solution flow and
mass transfer on the micro- and nanoscale.
REFERENCES
1. Auroux, P.A., Iossifidis, D., Reyes, D.R., and Manz, A., Micro total analysis systems. 2. Analytical
standard operations and applications, Analytical Chemistry, 74, 2637–2652, 2002.
2. Uchiyama, K., Nakajima, H., and Hobo, T., Detection method for microchip separations, Analytical
and Bioanalytical Chemistry, 379, 375–382, 2004.
3. Vilkner, T., Janasek, D., and Manz, A., Micro total analysis systems. Recent developments, Analytical
Chemistry, 76, 3373–3385, 2004.
4. Mogensen, K.B., Klank, H., and Kutter, J.P., Recent developments in detection for microfluidic
systems, Electrophoresis, 25, 3498–3512, 2004.
5. Dittrich, P.S., Tachikawa, K., and Manz, A., Micro total analysis systems. Latest advancements and
trends, Analytical Chemistry, 78, 3887–3907, 2006.
6. Viskari, P.J. and Landers, J.P., Unconventional detection methods for microfluidic devices, Elec-
trophoresis, 27, 1797–1810, 2006.
7. Harrison, D.J., Manz, A., Fan, Z.H., Ludi, H., and Widmer, H.M., Capillary electrophoresis and
sample injection systems integrated on a planar glass chip, Analytical Chemistry, 64, 1926–1932,
1992.
8. Mao, Q.L. and Pawliszyn, J., Demonstration of isoelectric focusing on an etched quartz chip with UV
absorption imaging detection, Analyst, 124, 637–641, 1999.
9. Verpoorte, E., Manz, A., Ludi, H., Bruno, A.E., Maystre, F., Krattiger, B., Widmer, H.M., Van-
derschoot, B.H., and Derooij, N.F., A silicon flow cell for optical-detection in miniaturized total
chemical-analysis systems, Sensors and Actuators B—Chemical, 6, 66–70, 1992.
10. Salimi-Moosavi, H., Jiang, Y.T., Lester, L., McKinnon, G., and Harrison, D.J., A multireflection cell for
enhanced absorbance detection in microchip-based capillary electrophoresis devices, Electrophoresis,
21, 1291–1299, 2000.
11. Duggan, M.P., McCreedy, T., and Aylott, J.W., A noninvasive analysis method for on-chip spectropho-
tometric detection using liquid-core waveguiding within a 3D architecture, Analyst, 128, 1336–1340,
2003.
12. Jindal, R. and Cramer, S.M., On-chip electrochromatography using sol-gel immobilized stationary
phase with UV absorbance detection, Journal of Chromatography A, 1044, 277–285, 2004.
13. Malcik, N., Ferrance, J.P., Landers, J.P., and Caglar, P., The performance of a microchip-based
fiber optic detection technique for the determination of Ca2+ ions in urine, Sensors and Actuators
B—Chemical, 107, 24–31, 2005.
14. Mogensen, K.B., Eriksson, F., Gustafsson, O., Nikolajsen, R.P.H., and Kutter, J.P., Pure-silica optical
waveguides, fiber couplers, and high-aspect ratio submicrometer channels for electrokinetic separation
devices, Electrophoresis, 25, 3788–3795, 2004.
15. Liang, Z.H., Chiem, N., Ocvirk, G., Tang, T., Fluri, K., and Harrison, D.J., Microfabrication of a
planar absorbance and fluorescence cell for integrated capillary electrophoresis devices, Analytical
Chemistry, 68, 1040–1046, 1996.
16. Xu, Y., Bessoth, F.G., Eijkel, J.C.T., and Manz, A., Online monitoring of chromium(III) using a fast
micromachined mixer/reactor and chemiluminescence detection, Analyst, 125, 677–683, 2000.
17. Liu, B.F., Ozaki, M., Utsumi, Y., Hattori, T., and Terabe, S., Chemiluminescence detection
for a microchip capillary electrophoresis system fabricated in poly(dimethylsiloxane), Analytical
Chemistry, 75, 36–41, 2003.
18. Su, R.G., Lin, J.M., Qu, F., Chen, Z.F., Gao, Y.H., and Yamada, M., Capillary electrophoresis
microchip coupled with online chemiluminescence detection, Analytica Chimica Acta, 508, 11–15,
2004.
19. Xu, Z.R. and Fang, Z.L., Composite poly(dimethylsiloxane)/glass microfluidic system with an
immobilized enzymatic particle-bed reactor and sequential sample injection for chemiluminescence
determinations, Analytica Chimica Acta, 507, 129–135, 2004.
20. Tsukagoshi, K., Jinno, N., and Nakajima, R., Development of a micro total analysis system incorporat-
ing chemiluminescence detection and application to detection of cancer markers, Analytical Chemistry,
77, 1684–1688, 2005.
21. Gerardi, R.D., Barnett, N.W., and Lewis, S.W., Analytical applications of tris(2,2 -
bipyridyl)ruthenium(III) as a chemiluminescent reagent, Analytica Chimica Acta, 378, 1–41,
1999.
22. Fiaccabrino, G.C., de Rooij, N.F., and Koudelka-Hep, M., On-chip generation and detection of
electrochemiluminescence, Analytica Chimica Acta, 359, 263–267, 1998.
23. L’Hostis, E., Michel, P.E., Fiaccabrino, G.C., Strike, D.J., de Rooij, N.F., and Koudelka-Hep, M.,
Microreactor and electrochemical detectors fabricated using Si and EPON SU-8, Sensors and Actuators
B—Chemical, 64, 156–162, 2000.
24. Pan, D.H. and Mathies, R.A., Chromophore structure in lumirhodopsin and metarhodopsin I by time-
resolved resonance Raman microchip spectroscopy, Biochemistry, 40, 7929–7936, 2001.
25. Zhao, X.C., You, T.Y., Qiu, H.B., Yan, J.L., Yang, X.R., and Wang, E.K., Electrochemiluminescence
detection with integrated indium tin oxide electrode on electrophoretic microchip for direct bioanalysis
of lincomycin in the urine, Journal of Chromatography B—Analytical Technologies in the Biomedical
and Life Sciences, 810, 137–142, 2004.
26. Qiu, H.B., Yin, X.B., Yan, J.L., Zhao, X.C., Yang, X.R., and Wang, E.K., Simultaneous electrochemical
and electrochemiluminescence detection for microchip and conventional capillary electrophoresis,
27. Fletcher, P.D.I., Haswell, S.J., and Zhang, X.L., Monitoring of chemical reactions within
microreactors using an inverted Raman microscopic spectrometer, Electrophoresis, 24, 3239–3245,
2003.
28. Leung, S.A., Winkle, R.F., Wootton, R.C.R., and deMello, A.J., A method for rapid reaction optimi-
sation in continuous-flow microfluidic reactors using online Raman spectroscopic detection, Analyst,
130, 46–51, 2005.
29. Ramser, K., Enger, J., Goksor, M., Hanstorp, D., Logg, K., and Kall, M., A microfluidic system
enabling Raman measurements of the oxygenation cycle in single optically trapped red blood cells,
Lab on a Chip, 5, 431–436, 2005.
30. Moskovits, M., Surface-enhanced Raman spectroscopy: a brief retrospective, Journal of Raman
Spectroscopy, 36, 485–496, 2005.
31. Tokeshi, M., Uchida, M., Hibara, A., Sawada, T., and Kitamori, T., Determination of subyocto-
mole amounts of nonfluorescent molecules using a thermal lens microscope: Subsingle molecule
determination, Analytical Chemistry, 73, 2112–2116, 2001.
32. Tamaki, E., Sato, K., Tokeshi, M., Sato, K., Aihara, M., and Kitamori, T., Single-cell analysis by a
scanning thermal lens microscope with a microchip: Direct monitoring of cytochrome c distribution
during apoptosis process, Analytical Chemistry, 74, 1560–1564, 2002.
33. Tokeshi, M., Kikutani, Y., Hibara, A., Sato, K., Hisamoto, H., and Kitamori, T., Chemical processing
on microchips for analysis, synthesis, and bioassay, Electrophoresis, 24, 3583–3594, 2003.
34. Mawatari, K., Naganuma, Y., and Shimoide, K., Portable thermal lens spectrometer with focusing
system, Analytical Chemistry, 77, 687–692, 2005.
35. Munson, M.S., Hasenbank, M.S., Fu, E., and Yager, P., Suppression of nonspecific adsorption using
sheath flow, Lab on a Chip, 4, 438–445, 2004.
36. Hasenbank, M.S., Fu, E., and Yager, P., Lateral spread of an amplification signal using an enzymatic
system on a conductive surface, Langmuir, 22, 7451–7453, 2006.
37. Furuki, M., Kameoka, J., Craighead, H.G., and Isaacson, M.S., Surface plasmon resonance sensors
utilizing microfabricated channels, Sensors and Actuators B—Chemical, 79, 63–69, 2001.
38. Wheeler, A.R., Chah, S., Whelan, R.J., and Zare, R.N., Poly(dimethylsiloxane) microfluidic flow
cells for surface plasmon resonance spectroscopy, Sensors and Actuators B—Chemical, 98, 208–214,
2004.
39. Kanda, V., Kariuki, J.K., Harrison, D.J., and McDermott, M.T., Label-free reading of microarray-
based immunoassays with surface plasmon resonance imaging, Analytical Chemistry, 76, 7257–7262,
2004.
40. Homola, J., Yee, S.S., and Gauglitz, G., Surface plasmon resonance sensors: review, Sensors and
Actuators B—Chemical, 54, 3–15, 1999.
41. Burggraf, N., Krattiger, B., de Mello, A.J., de Rooij, N.F., and Manz, A., Holographic refractive index
detector for application in microchip-based separation systems, Analyst, 123, 1443–1447, 1998.
42. Costin, C.D. and Synovec, R.E., A microscale-molecular weight sensor: Probing molecular diffusion
between adjacent laminar flows by refractive index gradient detection, Analytical Chemistry, 74,
4558–4565, 2002.
43. Jacobson, S.C., Hergenroder, R., Moore, A.W., and Ramsey, J.M., Precolumn reactions with
electrophoretic analysis integrated on a microchip, Analytical Chemistry, 66, 4127–4132, 1994.
44. Jacobson, S.C., Koutny, L.B., Hergenroder, R., Moore, A.W., and Ramsey, J.M., Microchip capillary
electrophoresis with an integrated postcolumn reactor, Analytical Chemistry, 66, 3472–3476, 1994.
45. Colyer, C., Kinkade, C., Viskari, P., and Landers, J., Analysis of cyanobacterial pigments and proteins
by electrophoretic and chromatographic methods, Analytical and Bioanalytical Chemistry, 382, 559–
569, 2005.
46. Karlinsey, J.M., Phycobiliprotein Analysis on Microchip, Advances and Applications of Microfluidic
Analysis Systems, 2007, University of Virginia: Charlottesville, VA.
47. Hellmich, W., Pelargus, C., Leffhalm, K., Ros, A., andAnselmetti, D., Single cell manipulation, analyt-
ics, and label-free protein detection in microfluidic devices for systems nanobiology, Electrophoresis,
26, 3689–3696, 2005.
48. Schulze, P., Ludwig, M., Kohler, F., and Belder, D., Deep UV laser-induced fluorescence detection
of unlabeled drugs and proteins in microchip electrophoresis, Analytical Chemistry, 77, 1325–1329,
2005.
49. Haugland, R.P., The Handbook: A Guide to Fluorescent Probes and Labeling Technologies, 10th ed.
2005: Invitrogen Corp.
50. Karlinsey, J.M., Monahan, J., Marchiarullo, D.J., Ferrance, J.P., and Landers, J.P., Pressure injection
on a valved microdevice for electrophoretic analysis of submicroliter samples, Analytical Chemistry,
77, 3637–3643, 2005.
51. Karlinsey, J.M. and Landers, J.P., Multicolor fluorescence detection on an electrophoretic microdevice
using an acoustooptic tunable filter, Analytical Chemistry, 78, 5590–5596, 2006.
52. Munro, N.J., Huang, Z.L., Finegold, D.N., and Landers, J.P., Indirect fluorescence detection of amino
acids on electrophoretic microchips, Analytical Chemistry, 72, 2765–2773, 2000.
53. Sirichai, S. and de Mello, A.J., A capillary electrophoresis microchip for the analysis of photographic
developer solutions using indirect fluorescence detection, Analyst, 125, 133–137, 1999.
54. Liu, Y.J., Foote, R.S., Jacobson, S.C., Ramsey, R.S., and Ramsey, J.M., Electrophoretic separation of
proteins on a microchip with noncovalent, postcolumn labeling, Analytical Chemistry, 72, 4608–4613,
2000.
55. Giordano, B.C., Jin, L.J., Couch, A.J., Ferrance, J.P., and Landers, J.P., Microchip laser-induced
fluorescence detection of proteins at submicrogram per milliliter levels mediated by dynamic labeling
under pseudonative conditions, Analytical Chemistry, 76, 4705–4714, 2004.
56. Beard, N.P., Edel, J.B., and deMello, A.J., Integrated on-chip derivatization and electrophoresis for
the rapid analysis of biogenic amines, Electrophoresis, 25, 2363–2373, 2004.
57. Sanders, J.C., Breadmore, M.C., Kwok, Y.C., Horsman, K.M., and Landers, J.P., Hydroxypropyl cellu-
lose as an adsorptive coating sieving matrix for DNAseparations: Artificial neural network optimization
for microchip analysis, Analytical Chemistry, 75, 986–994, 2003.
58. Ju, J.Y., Kheterpal, I., Scherer, J.R., Ruan, C.C., Fuller, C.W., Glazer, A.N., and Mathies, R.A.,
Design and synthesis of fluorescence energy-transfer dye-labeled primers and their application for
DNA-sequencing and analysis, Analytical Biochemistry, 231, 131–140, 1995.
59. Berti, L., Xie, J., Medintz, I.L., Glazer, A.N., and Mathies, R.A., Energy transfer cassettes for facile
labeling of sequencing and PCR primers, Analytical Biochemistry, 292, 188–197, 2001.
60. Medintz, I.L., Berti, L., Emrich, C.A., Tom, J., Scherer, J.R., and Mathies, R.A., Genotyping
energy-transfer-cassette-labeled short-tandem-repeat amplicons with capillary array electrophoresis
microchannel plates, Clinical Chemistry, 47, 1614–1621, 2001.
61. Hernandez, L., Escalona, J., Joshi, N., and Guzman, N., Laser-induced fluorescence and fluorescence
microscopy for capillary electrophoresis zone detection, Journal of Chromatography, 559, 183–196,
1991.
62. He, B., Tait, N., and Regnier, F., Fabrication of nanocolumns for liquid chromatography, Analytical
Chemistry, 70, 3790–3797, 1998.
63. Johnson, M.E. and Landers, J.P., Fundamentals and practice for ultrasensitive laser-induced
fluorescence detection in microanalytical systems, Electrophoresis, 25, 3513–3527, 2004.
64. Mogensen, K.B., Kwok, Y.C., Eijkel, J.C.T., Petersen, N.J., Manz, A., and Kutter, J.P., A microfluldic
device with an integrated waveguide beam splitter for velocity measurements of flowing particles by
Fourier transformation, Analytical Chemistry, 75, 4931–4936, 2003.
65. Guillo, G., Karlinsey, J.M., and Landers, J.P., On-chip pumping for pressure mobilization of the
focused zones following microchip isoelectric focusing, Lab on a Chip, 7, 112–118, 2007.
66. Yao, B., Yang, H.H., Liang, Q.L., Luo, G., Wang, L.D., Ren, K.N., Gao, Y.D., Wang, Y.M., and
Qiu, Y., High-speed, whole-column fluorescence imaging detection for isoelectric focusing on a
microchip using an organic light emitting diode as light source, Analytical Chemistry, 78, 5845–5850,
2006.
67. Raisi, F., Belgrader, P., Borkholder, D.A., Herr, A.E., Kintz, G.J., Pourhamadi, F., Taylor, M.T.,
and Northrup, M.A., Microchip isoelectric focusing using a miniature scanning detection system,
68. Hjerten, S., Free zone electrophoresis, Chromatographic Reviews, 9, 122–219, 1967.
69. Sanders, J.C., Huang, Z.L., and Landers, J.P., Acousto-optical deflection-based whole channel scan-
ning for microchip isoelectric focusing with laser-induced fluorescence detection, Lab on a Chip, 1,
167–172, 2001.
70. Cheng, S.B., Skinner, C.D., Taylor, J., Attiya, S., Lee, W.E., Picelli, G., and Harrison, D.J., Develop-
ment of a multichannel microfluidic analysis system employing affinity capillary electrophoresis for
immunoassay, Analytical Chemistry, 73, 1472–1479, 2001.
71. Huang, Z., Munro, N., Huhmer, A.F.R., and Landers, J.P., Acoustooptical deflection-based laser beam
scanning for fluorescence detection on multichannel electrophoretic microchips, Analytical Chemistry,
71, 5309–5314, 1999.
72. Simpson, P.C., Roach, D., Woolley, A.T., Thorsen, T., Johnston, R., Sensabaugh, G.F., and Mathies,
R.A., High-throughput genetic analysis using microfabricated 96-sample capillary array electrophore-
sis microplates, Proceedings of the National Academy of Sciences of the United States of America, 95,
2256–2261, 1998.
73. Emrich, C.A., Tian, H.J., Medintz, I.L., and Mathies, R.A., Microfabricated 384-lane capillary array
electrophoresis bioanalyzer for ultrahigh-throughput genetic analysis, Analytical Chemistry, 74, 5076–
5083, 2002.
74. Shi, Y.N., Simpson, P.C., Scherer, J.R., Wexler, D., Skibola, C., Smith, M.T., and Mathies, R.A., Radial
capillary array electrophoresis microplate and scanner for high-performance nucleic acid analysis,
Analytical Chemistry, 71, 5354–5361, 1999.
75. Liu, S.R., Shi, Y.N., Ja, W.W., and Mathies, R.A., Optimization of high-speed DNA sequenc-
ing on microfabricated capillary electrophoresis channels, Analytical Chemistry, 71, 566–573,
1999.
76. Schmalzing, D., Koutny, L., Chisholm, D., Adourian, A., Matsudaira, P., and Ehrlich, D., Two-color
multiplexed analysis of eight short tandem repeat loci with an electrophoretic microdevice, Analytical
Biochemistry, 270, 148–152, 1999.
77. Backhouse, C., Caamano, M., Oaks, F., Nordman, E., Carrillo, A., Johnson, B., and Bay, S., DNA
sequencing in a monolithic microchannel device, Electrophoresis, 21, 150–156, 2000.
78. Simpson, J.W., Ruiz-Martinez, M.C., Mulhern, G.T., Berka, J., Latimer, D.R., Ball, J.A., Rothberg,
J.M., and Went, G.T., A transmission imaging spectrograph and microfabricated channel system for
DNA analysis, Electrophoresis, 21, 135–149, 2000.
79. Karlinsey, J.M., Multicolor Short Tandem Repeat (STR) Analysis on Microchip, Advances and
Applications of Microfluidic Analysis Systems, 2007, University of Virginia: Charlottesville, VA.
80. Fister, J.C., Jacobson, S.C., and Ramsey, J.M., Ultrasensitive cross correlation electrophoresis on
microchip devices, Analytical Chemistry, 71, 4460–4464, 1999.
81. Haab, B.B. and Mathies, R.A., Single-molecule detection of DNA separations in microfabricated
capillary electrophoresis chips employing focused molecular streams, Analytical Chemistry, 71, 5137–
5145, 1999.
82. Foquet, M., Korlach, J., Zipfel, W.R., Webb, W.W., and Craighead, H.G., Focal volume confinement
by submicrometer-sized fluidic channels, Analytical Chemistry, 76, 1618–1626, 2004.
83. Chabinyc, M.L., Chiu, D.T., McDonald, J.C., Stroock, A.D., Christian, J.F., Karger, A.M., and White-
sides, G.M., An integrated fluorescence detection system in poly(dimethylsiloxane) for microfluidic
applications, Analytical Chemistry, 73, 4491–4498, 2001.
84. Roulet, J.C., Volkel, R., Herzig, H.P., Verpoorte, E., de Rooij, N.F., and Dandliker, R., Performance
of an integrated microoptical system for fluorescence detection in microfluidic systems, Analytical
Chemistry, 74, 3400–3407, 2002.
85. Maims, C., Hulme, J., Fielden, P.R., and Goddard, N.J., Grating coupled leaky waveguide micro
channel sensor chips for optical analysis, Sensors and Actuators B—Chemical, 77, 671–678, 2001.
86. Applegate, R.W., Squier, J., Vestad, T., Oakey, J., Marr, D.W.M., Bado, P., Dugan, M.A., and Said,
A.A., Microfluidic sorting system based on optical waveguide integration and diode laser bar trapping,
Lab on a Chip, 6, 422–426, 2006.
87. Adams, M.L., Enzelberger, M., Quake, S., and Scherer, A., Microfluidic integration on detector arrays
for absorption and fluorescence microspectrometers, Sensors and Actuators A—Physical, 104, 25–31,
2003.
88. Edel, J.B., Beard, N.P., Hofmann, O., deMello, J.C., Bradley, D.D.C., and deMello, A.J., Thin-film
polymer light emitting diodes as integrated excitation sources for microscale capillary electrophoresis,
Lab on a Chip, 4, 136–140, 2004.
89. Hofmann, O., Miller, P., Sullivan, P., Jones, T.S., deMello, J.C., Bradley, D.D.C., and deMello,
A.J., Thin-film organic photodiodes as integrated detectors for microscale chemiluminescence assays,
Sensors and Actuators B—Chemical, 106, 878–884, 2005.
90. Ocvirk, G., Tang, T., and Harrison, D.J., Optimization of confocal epifluorescence microscopy for
microchip-based miniaturized total analysis systems, Analyst, 123, 1429–1434, 1998.
46 Microfabricated
Electrophoresis Devices for
High-Throughput Genetic
Analysis: Milestones and
Challenges
Charles A. Emrich and Richard A. Mathies
CONTENTS
46.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1277

46.2 Background and Evolution µCAE Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1279
46.2.1 Advancing to 96 Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1280
46.2.2 Transition to Radial Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1281
46.2.3 Optimizing Microchip DNA Sequencing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1283
46.2.4 To the Ultimate 384 Lane Microdevice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1287
46.2.5 Integration: The Final Frontier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1288
46.3 Conclusions and Prospects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1293
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1293
46.1 INTRODUCTION
The first microfabricated separation system introduced by Harrison et al. [1,2] highlighted the unique
marriage of photolithography with analytical chemistry to perform ultrafast, high-resolution separa-
tions of subnanoliter volume samples. The lab-on-a-chip concept that grew from these early papers
leveraged two important aspects of microfabrication. First, the ability to integrate complicated elec-
trical and mechanical structures directly with the separation channels enables prepurification sample
preparation, a concomitant reduction in manual sample handling steps and costly reagent volumes,
and the eventual miniaturization of the “lab” for point-of-analysis studies. The second advantage
that microfabrication adds is the ability to create arbitrarily complex arrays of channels on the same
substrate. The photolithographic techniques used are equivalent regardless of whether the desired
channel geometry is simple or complex, in sharp contrast to conventional capillary arrays where each
additional capillary exponentiates the complexity and decreases robustness. Thus, microfabricated
capillary electrophoresis (CE) systems are ideally suited to perform high-throughput analysis.
The first microfabricated capillary array electrophoresis (µCAE) device appeared in 1997 [3]
featuring 12 separation channels interrogated by a scanning laser-induced fluorescence (LIF) detector.
1277
50 mm 100 mm 100 mm 150 mm 150 mm 200 mm
1997 1998 1999 2000 2001 2002

Waste Sample Sample Sample Sample Cathode
Cathode Cathode Cathode
Sample Cathode
Waste Waste Waste Sample
Cathode
Wa
Anode
Anode Anode
Anode Anode Anode
FIGURE 46.1 The evolution of high-throughput µCAE devices. The earliest devices were fabricated on large
microscope slides and used rectilinear arrays of channels, which terminated at a common anode. More complex
rectilinear arrays were clearly not scalable and were replaced by modern radial arrays. Radial channel arrays
are simple to design and operate and are easily scalable by increasing the wafer diameter.
A race for extending and maximizing the density of channel arrays quickly followed. As Figure 46.1
illustrates, arrays grew in channel density first to 48, then 96, and ultimately to 384 channels [4].
To accommodate the growing number of channels, the “microchips” on which they were fabricated
also grew from the size of microscope slides to window-size plates half a meter in length [5]. Much
of the growth in channel number and areal density was facilitated by an early shift from rectilinear to
radial arrays of channels [6] and the simultaneous development of rotary confocal detection systems.
Despite the great increase in throughput provided by µCAE systems, they are only half way
to displacing conventional separation formats. While the resolution and sequencing read lengths of
µCAE systems are on par with their conventional counterparts and the time required for similar
separations is shorter, they are much closer to a capillaries-on-a-chip than they are to labs-on-a-
chip. The next step in the evolution of lab-on-a-chip devices for high-throughput analysis is the
integration of sample preparation components before separation. DNA amplification processes such
as the polymerase chain reaction (PCR) and Sanger cycle sequencing are the most attractive sample
preparation steps and have been integrated with CE exploiting the development of on-chip tem-
perature and valve control elements [7]. On-chip extraction techniques have also been recently
demonstrated for purification of DNA directly from clinical samples that often contain an array
of inhibitory background molecules [8]. Sample cleanup steps have also been developed to desalt
and concentrate Sanger extension fragments before electrokinetic injections, greatly increasing the
performance of high-resolution DNA-sequencing separations [9]. The integration of these elements
into multichannel arrays has advanced with the recent demonstration of four-channel PCR-CE sys-
tems [10,11]. Total integration of sample preparation techniques with on-chip CE analysis came
of age in 2006 with the demonstration of single-channel systems for DNA sequencing [12] and
Microfabricated Electrophoresis Devices for High-Throughput Genetic Analysis 1279
infectious disease typing [13]. Clearly, the successful implementation of integrated components is
the next step in µTAS lab-on-a-chip systems.
The demand for high-throughput electrophoretic analysis, which grew exponentially after the
birth of PCR in 1988 [14] and with the sequencing of the human genome [15] has only increased
in recent years. New metagenomic tests are being used to sample the genomes of environmental
specimens [16], mitochondrial barcodes are now used to catalog the world’s animal species [17],
and increasingly DNA evidence has become the gold standard for human identification in criminal
investigations. Taken further, human identification by electrophoresis can be used beyond the crime
lab in cases of parental determination and to identify victims of mass disasters [18].
DNA electrophoretic techniques are also becoming increasingly relevant not only in the under-
standing of cancer but also, enabled by recent breakthroughs in so-called targeted therapies, in
charting a course for its treatment. The Lazarus-like response of certain patients to the lung cancer
drug “gefitinib” was explained by mutations in the growth factor receptor EGFR, which were found
by DNA sequencing [19]. More recently, the identification of a metagene construct for evaluating
the severity of nonsmall cell lung cancers promises a more accurate determinant of individual patient
therapy than the traditional staging system [20]. Because of the genetic variability of both cancers
and patients, it is likely that the need for high-throughput genetic screens will only increase in the
future. Thus, it is imperative that electrophoretic techniques be pushed toward higher-throughput,
lower cost, and smaller size to truly enable the coming generation of personalized medicine. Such
developments in integrated µCE systems incorporating sample purification, amplification, and anal-
ysis with robust temperature and fluidic control have recently begun to emerge [21] and will provide
an effective foundation on which to build future bioanalysis devices.
46.2 BACKGROUND AND EVOLUTION µCAE SYSTEMS

The first high-throughput arrayed CE microdevice was demonstrated 5 years after the first single-
channel analysis system and consisted of 12 channels with linear scanning detection [3]. The design
of the 12-channel µCAE device presented in Figure 46.1 is a modification of the original cross-
injector design set out by Harrison et al. [22] and proven by Woolley and Mathies [23] for CE
separations of DNA. The injectors of the various channels are arrayed to minimize the dimensions of
the chip (maximizing channel density) while leaving sufficient space between reservoirs to maintain
electrical and sample isolation. Each of the 12 channels terminates at a common anode reservoir
located just below the detection point, for a total of 3N+1 drilled reservoirs where N is the number of
samples. Interrogation of the channels is performed by linearly translating the µCAE device below
a laser confocal fluorescence point detection system.
The 12-channel µCAE devices were fabricated on glass microscope slides by using photolithog-
raphy and wet chemical etching to 8-µm depths. The etch depth of the channels, in this case,
was limited because during photolithography the photoresist was patterned directly on the sur-
face of the wafers without the use of a thin-film “hard” mask layer. Photoresists typically exhibit
poor adhesion to hydrophilic surfaces such as glass, resulting in significant undercutting of the
photoresist during wet chemical etching steps. This undercutting introduces defects in the etched
surface, especially with long etches, and significantly reduces the resolution of the photolithographic
pattern.
The 12-channel µCAE device was successfully demonstrated by genotyping twelve individuals
for mutations in the human HLA-H gene in a restriction fragment length polymorphism (RFLP) assay.
Samples were noncovalently labeled with DNA intercalating dyes and run simultaneously with DNA
standards for accurate sizing. The 0.5% hydroxyethylcellulose (HEC) separation matrix has low
viscosity facilitating its loading into the shallow channels. Despite the differing channel routing in
the µCAE device, the electropherograms presented in Figure 46.2 display little channel-to-channel
variation, achieved resolution of better than 10 bp, and were complete in 160 s.
Lane number
1 2 3 4 5 6 7 8 9 10 11 12
Primer-dimer
110 845A
147 845G
160
180
201
pBR322 MspI Peaks (bp)
238+242
255 bp
HLA-H Peaks
307
404
527
622
1.1 mm
FIGURE 46.2 Electropherograph of a HLA-H genotyping separation using the 12-channel µCAE device.
Fluorescent labeling is achieved by the noncovalent intercalating dyes thiazole orange and butyl TOTIN. Sep-
arations were monitored by linearly translating the device under a laser confocal detection system focused
just upstream of the common anode shared by all channels. A pBR MspI digest is added as a sizing standard.
(Reproduced from Woolley, A.T. et al., Anal. Chem., 69, 2181, 1997. With permission.)
46.2.1 ADVANCING TO 96 SAMPLES

The next significant step in the evolution of µCAE devices required transitioning from glass
microscope slides to larger, 100-mm diameter wafers made from Borofloat glass, which is more
compatible with microfabrication processes because of its low sodium content and also exhibits
lower autofluorescence, improving detection limits. A 48-channel µCAE device built on the larger
wafers is presented in Figure 46.3 and features 96 sample reservoirs, pairs of which share the forty-
eight 10-cm long separation channels [24]. The 48-channel µCAE chips were microfabricated using
an amorphous silicon hard-mask deposited by plasma-enhanced chemical vapor deposition. The addi-
tion of this hard-mask step was critical to achieving high-fidelity microfabrication of 21-µm deep
channels by wet chemical etching. These deeper channels permit higher-viscosity sieving matrices
to be used, which enable higher resolution separations. Interrogation of the channels is performed
by a linear galvoscanner to successfully sample each channel at a rate of 40 Hz.
The 48-channel µCAE device employs 96 sample reservoirs, 6 cathode reservoirs, 24 waste
reservoirs, and a common anode reservoir. This design reduces the number of necessary reservoirs
to 5/4N+7, which is close to the theoretical minimum of N+3 for the case where the anode, cathode,
and waste reservoirs are all shared. This minimization of total reservoirs is important because the
reservoirs are the largest consumer of area on the chip surface. Not only are they 1.25-mm in diameter,
but positional error in drilling places limits on the reservoir density. The 48-channel µCAE chip was
S1 S2
Cathodes
W
Anode
S3 S4
Inject Run
Detector
scanning
path
100 mm
FIGURE 46.3 Design of the 48-channel µCAE chip. The unique cross-injector design allowed two sample
reservoirs (S) to utilize the same separation channel in succession. The total number of reservoirs necessary
for operation is greatly reduced by grouping waste (W) and cathode reservoirs. Each channel has a 10-cm
effective separation length terminating in a common anode. Detection is achieved using a galvoscanner along
the indicated linear path. (Adapted from Simpson, P.C. et al., Proc. Natl. Acad. Sci. USA, 95, 2256, 1998. With
permission.)
demonstrated by genotyping 96 individuals for the C282Y mutation in the human HFE gene using
a PCR-RFLP assay. Each separation channel in the 48-channel device is used to serially analyze
two separate samples. Figure 46.4 presents the results from such a double-injection run of 96 HFE
genotyping samples. Electrophoresis is complete within 8 min for both sets of samples but significant
variation in peak retention times is observed, possibly as a result of channel geometry or variation
of electrode position in the reservoirs.
The rectilinear layout of this chip enables quick sample loading via multichannel pipettors, but
this design paradigm scales poorly for higher channel density arrays. Similarly, the integration of
sample preparation components is clearly difficult for such designs because of the limited area around
each sample reservoir. Another concern is that while channel lengths are equalized in this design,
they still incorporate right-angle turns that were later revealed to be a significant source of dispersion
in on-chip electrophoretic separations [25].
46.2.2 TRANSITION TO RADIAL FORMATS

Further improvements in high-resolution, high-throughput µCAE necessitated a paradigm shift from
rectilinear to radial arrays of channels and a concomitant advancement in detection technologies.
The first radial µCAE device [6] arrayed 96 separation channels around a central common anode
on a 100-mm diameter glass substrate as presented in Figure 46.5. Pairs of channels are grouped
into doublet structures that share common cathode and waste reservoirs. The design of this radial
µCAE device offers significant advantages over previous rectilinear arrays. First, all the channels
have equivalent lengths and geometries, negating any variability in electrophoretic conditions that
might be present in irregular channels. Second, the radial array paradigm is perfectly complimentary
to the circular wafers used in microfabrication processes, maximizing utilization of existing wafer
(a) (b) First injection
3
1 12 24 36 48
Channel number
Minutes
4
(c) Second injection
7 1 12 24 36 48
Channel number
7.4 mm
FIGURE 46.4 Electrophoresis pseudo gel image from a genotyping separation performed on the 48-channel
µCAE device. The 96 samples are analyzed in two successive runs, successfully discriminating between full-
length 167-bp fragments and 111 or 140-bp digested fragments. (Reproduced from Simpson, P.C. et al., Proc.
Natl. Acad. Sci. USA, 95, 2256, 1998. With permission.)
area by positioning the drilled reservoirs at the edge of the wafer where the per-channel-area is
greatest. Positioning the reservoirs at the periphery of the wafer in a radially symmetric fashion
also allows simple construction of electrode arrays to address each reservoir. Third, the separation
channels in a radial array µCAE device can be straight between the injection point and the detection
zone. While longer separation channels generally yield higher resolution than shorter channels,
any turns in those channels can act as a source of dispersion in the bands. Fourth, radial arrays of
channels are clearly the best design for scaling to ultrahigh-throughput analysis devices, as wafer
sizes >100-mm diameter are readily available and the tools for processing of wafers up to 300-mm
diameter are well developed and commonplace in the semiconductor industry. Thus, increasing the
number of channels can be accomplished by increasing wafer size to accommodate the necessary
number of sample, cathode, and waste reservoirs. Further, integration of sample preparation steps
into a µCAE design becomes viable when the sample reservoirs are at the periphery of the analysis
device where there is more available surface area. Components of such an integrated device can then
be incorporated by placing them distal to the sample reservoirs, and the number of preseparation
steps will be limited only by the ultimate size of the substrate.
The shift to radial channel arrays also necessitated a shift in detection methodology from linear
scanning to rotary scanning as presented in Figure 46.6. The current generation rotary scanner directs
excitation from an Ar+ laser up through a hollow rotating shaft atop which sits a rhomb prism and
objective that respectively deflect the laser 1-cm off the axis of rotation and focus the beam into
the microchannels. Fluorescence is collected along the same path as excitation, passing through a
dichroic beamsplitter and into a four-color confocal photomultiplier tube (PMT) detector array. In
addition, the Berkeley rotary confocal fluorescence scanner is an inverted optical system, which puts
all the optical elements on the underside of the chip and all electrical connections on the opposite,
top side of the chip, thereby maximizing flexibility of the chip design.
The performance of the 96-channel, 100-mm diameter µCAE device was demonstrated by
simultaneously genotyping 96 individuals for the C677T mutation in the human MTHFR gene [6].
Electrophoresis was performed at 200 V/cm through a 1.0% w/v HEC sieving matrix and used dimeric
intercalating dyes for noncovalent fluorescent labeling. Separations on the 3.5-cm long channels were
complete within 120 s, corresponding to a sample throughput of approximately 0.5 samples/s. The
utility of this radial design was further advanced by increasing the diameter of the substrate from
100 to 150 mm. By increasing the wafer size, the effective separation length of each channel was
increased from 3.5 to 5.5 cm, which increased analysis time, but also increased the ultimate reso-
lution of the µCAE device. This increase in resolution enabled more demanding analyses such as
the short tandem repeat typing assays commonly used in forensic identification, which require near
single-base resolution. Medintz et al. [26] performed the first µCAE short tandem repeat (STR) sep-
arations using a covalent fluorescent multiplexing scheme, demonstrating extremely high-resolution
separations in under 8 min with <1% sizing variance. Results from a single multiplex STR genotyp-
ing separation are presented in Figure 46.7. The improvement in resolution over previous separations
can be attributed not only to the increase in effective separation length from 3.5 to 5.5 cm, but also
to the use of a high-performance, denaturing linear polyacrylamide (LPA) separation matrix. The
development of a simple and robust high-pressure gel loader and washer by Scherer et al. [27] facil-
itated the use of high-viscosity LPA in the 150-mm µCAE chip. By applying pressurized helium gas
to the sieving matrix, it can be forced into the channels of µCAE device through the common central
anode reservoir in under 10 min. After the completion of an electrophoretic run, sieving matrix can
then be washed from the radial µCAE chip with water and dried using the same apparatus.
46.2.3 OPTIMIZING MICROCHIP DNA SEQUENCING

The ultimate test of any electrophoretic separation system is DNA sequencing of Sanger extension
fragments. Sequencing separations not only require single-base resolution, but also require that
degree of resolution over 2–3 orders of magnitude of sample fragment size. In addition, the longer the
DNA sequencing read length, the more value it has during the final sequence assembly step. The read
lengths demonstrated by microfabricated DNA separation systems by 2000 had been significantly
less than the 500+ base reads from commercial capillary array systems [28,29]. Notable exceptions
include straight-channel microdevices with separation lengths from 11 to 40 cm [30,31], which
achieved DNA sequencing reads of up to 800 bases. However, the increased size and complexity
of fabrication and design of these longer “capillary-on-a-window” style separation devices negated
many of the advantages of miniaturization that are the hallmarks of microfabricated separation
systems. The obvious solution was to increase the effective separation length of channels on µCAE
chips without increasing substrate size by folding the channels back along themselves. However,
turns are effective sources of dispersion in µCE separations [25,32,33]. The dispersion is due to the
racetrack effect, in which species all traveling at the same velocity will pass through a turn more
quickly along the shorter inside track of the turn than along the longer outside track. This effect is
S1 S2
1 mm
100 mm
FIGURE 46.5 The first radial array µCAE chip featured 96 sample channels arrayed about a single, common
anode on a 100-mm diameter glass substrate. Utilization of space was maximized by placing reservoirs at the
periphery of the wafer and grouping pairs of channels into a doublet structure, which shared cathode (C) and
waste (W) reservoirs. Detection was performed along a circular path 1 cm from the anode with the scanner
described in Figure 46.6. (Adapted from Shi, Y.N. et al., Anal. Chem., 71, 5354, 1999. With permission.)
Electrode
array HV Power
supplies
µCAE Chip
60 X 0.7NA
Rhomb
Objective
prism
Hollow shaft 4-Color

stepper motor confocal
detector
Preamp/
10 kHz filter
d1
16-bit, 25 kHz
ADC data
Dichroic collection
beamsplitter
Ar + Laser
488 nm
FIGURE 46.6 The Berkeley rotary confocal fluorescence scanner developed by James R. Scherer. Laser
excitation is directed up through the hollow drive shaft of a stepper motor, deflected off axis by a rhomb prism,
and is then focused with a microscope objective, tracing a 1-cm radius circular detection path on the chip.
Fluorescence is collected along the same path and then spatially and spectrally filtered by a four-color PMT
array. The inverted nature of this system allows easy access to the upper surface of a µCAE device for electrical,
pneumatic, and sample interfacing.
Size (bp)
120
90
100
140
160
180
190
200
220
240
360
280
260
80
320
340
380
300
400
70
96
ber 72
um
nel n
48
Chan
24
1
2 3 4 5 6 7 8
Time (min)
FIGURE 46.7 (See color insert following page 810.) Electropherograms from a multiplexed short tandem
repeat sizing separation on a 150-mm diameter, 96-channel µCAE device. Separations are successful in all lanes
and sizing against DNA standards (blue and red) was accurate to within 1% of known values. (Reproduced
from Medintz, I.L. et al., Clin. Chem., 47, 1614, 2001. With permission.)
exacerbated in electrokinetic transport because the shorter inside path also has a lower resistance than
the outside path, and will thus experience a higher electric field strength that further increases the
difference in transit time between the inside and outside tracks. In an elegant series of experiments,
Paegel et al. [34] discovered that the racetrack dispersion of bands passing through turns could be
almost eliminated simply by narrowing the width of the channel for the duration of the turn.
The development of these hyperturns was a critical advance toward development of a µCAE
system for high-throughput DNA sequencing by enabling the longer separation lengths necessary for
high-resolution separations. By folding the separation channels back on themselves with hyperturns
it was possible to extend the channels on a 150-mm diameter wafer to a total effective separation
length of 15.9 cm (see Figure 46.8). The total combined separation length on the µCAE device
presented in Figure 46.8 is 15.25 m, with the channels taking up the majority of the surface area
of the wafer. Maximizing the amount of utilized space on the wafer surface was also found to
significantly improve the success rate of thermal bonding of µCAE devices for two reasons. The
smaller contact area between the wafers will experience greater pressure during bonding and the
presence of the channels provides convenient escape routes for gas trapped between the wafers
during bonding.
S1 S2
150 mm
FIGURE 46.8 Design of the 96-channel µCAE sequencer. Channels are grouped into doublets and folded
back on themselves twice for an effective separation length of 15.9-cm. Each turn uses the indicated hyperturn
geometry to minimize turn-related dispersion. The lengths and widths of channels in the 250-µm offset cross-
injector are balanced to provide equal fluidic resistance during gel loading. (Adapted from Paegel, B.M. et al.,
Proc. Natl. Acad. Sci. USA, 99, 574, 2002. With permission.)
The performance of the 96-channel µCAE sequencer was demonstrated by simultaneous elec-
trophoresis of 96 identical M13mp18 vector DNA sequencing standards as presented in Figure 46.9.
Electrophoresis was performed at 240 V/cm through LPA under denaturing conditions resulting in
continuous DNA sequencing reads of 430 bases with >99% accuracy in 24 min [35]. This level of
accuracy, corresponding to a Phred score of 20 is recognized as the standard metric for determining
read length. The longer time required for separations of DNA sequencing fragments is necessary
because of the increased length of the channels, the more efficient long-chain sieving matrix, and
the upper practical limit placed on the electric field due to the onset of biased repetition of the
larger fragments. The longer electrophoresis times necessary for high-resolution separation of DNA
sequencing fragments also necessitated an increase in the buffering capacity of the cathode and anode
reservoirs. This volume increase was made possible by affixing Plexiglas moats to the surface of
the chip, and filling the moats with 3 mL of Tris-TAPS-EDTA run buffer. The added buffer volume
was necessary to offset the hydrolytic production of hydroxyl ions at the cathode and protons at
the anode reservoirs. These hydrolytic products titrate and then acidify the anode buffer, creating a
local pH drop that reduces the net negative charge on the DNA and significantly reduces the overall
separation efficiency.
The 96-channel µCAE sequencer proved to be a robust high-resolution separation platform that
is easily extended to other applications. Through creative fluorescent multiplexing, Blazej et al. [36]
used the µCAE sequencer to mine two human mitochondrial genomes for mutations, analyzing an
entire genome versus a reference genome in a single electrophoretic run. More recently, Yeung et al.
[37] demonstrated the utility of the µCAE sequencer to high-resolution STR typing separations for
forensic DNA identification. The system performed separations with resolution and sensitivity equal
12
Run time (min)

16
20
24
1 24 48 72 96
Lane number
FIGURE 46.9 Results from a DNA sequencing separation of M13mp18 standards on the 96-channel µCAE
device. Average sequencing read lengths of 430 bases of ≥99% accuracy were obtained, and only one of
the 96 lanes failed. (Reproduced from Paegel, B.M. et al., Proc. Natl. Acad. Sci. USA, 99, 574, 2002. With
permission.)
to conventional capillary systems in less than half the time, and proved useful not only for standard
sizing ladders but also for actual (nonprobative) casework samples.
46.2.4 TO THE ULTIMATE 384 LANE MICRODEVICE

Increasing throughput beyond what was demonstrated for 96-channel µCAE devices require an
increase in the wafer diameter. Figure 46.10 presents the design of a 384-lane µCAE device for
ultrahigh throughput genotyping [4]. This channel density is close to the practical maximum for
radial arrays and the Berkeley rotary confocal fluorescence detection and required the replacement
of standard cross-injector with a direct injector. The direct injector used in the 384-lane µCAE
bioanalyzer provides significant savings of area by eliminating the waste reservoirs associated with
a cross-injector and also groups channels into quartets, which share a common cathode reservoir.
Thus, the total number of reservoirs is reduced to 5/4N+1, but the elimination of the cross-injector
comes at a cost of reduced resolution because of the larger direct-injected plug size. Much of
this lost resolution is made up by the increased 8.0-cm effective separation length afforded by the
larger diameter wafer, achieving theoretical plate numbers of up to 4 × 106 for separations of
DNA size standards. The performance of the 384-channel µCAE bioanalyzer was demonstrated by
simultaneously genotyping 384 individuals for the H63D mutation in the human HFE gene via an
RFLP assay. The electrophoretic separation was performed at 260 V/cm and was complete within
325 s for an overall throughput of >1 samples/s. Results of the separation presented in Figure 46.11
reveal excellent reproducibility between channels and the 98.7% success rate of the separations
stands as a testament to the robustness of both the radial array concept and the rotary confocal
fluorescence scanner.
The design, manufacture, and operation of the 384-lane µCAE bioanalyzer likely represents
a practical maximum for throughput in microfabricated electrophoresis systems. The fabrication
S2 S3
S1 S4
200 mm
FIGURE 46.10 Design of the 384-lane µCAE bioan

Handbook of Capillary and Microchip Electrophoresis and Associated Micro Techniques, 3rd Edition

Uploaded by

Copyright:

Available Formats

Handbook of Capillary and Microchip Electrophoresis and Associated Micro Techniques, 3rd Edition

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Handbook of Capillary and Microchip Electrophoresis and Associated Micro Techniques, 3rd Edition

Uploaded by

Copyright:

Available Formats

Dedication

Part I Fundamentals and Methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Chapter 2 Protein Analysis by Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

Chapter 3 Micellar Electrokinetic Chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

Chapter 4 Capillary Electrophoresis for Pharmaceutical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

Chapter 5 Principles and Practice of Capillary Electrochromatography . . . . . . . . . . . . . . . . . . . . . 183

Chapter 6 Capillary Electrophoresis of Nucleic Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

Chapter 7 Analysis of Carbohydrates by Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . 251

Chapter 8 The Coupling of Capillary Electrophoresis and Mass Spectrometry in

Chapter 9 Light-Based Detection Methods for Capillary Electrophoresis. . . . . . . . . . . . . . . . . . . 305

Part IIA Capillary-Based Systems: Core Methods and Technologies . . . . . . . . . . . . . 359

Chapter 12 DNA Sequencing and Genotyping by Free-Solution

Chapter 13 Online Sample Preconcentration for Capillary Electrophoresis . . . . . . . . . . . . . . . . . . 413

Chapter 14 Capillary Electrophoresis for the Analysis of Single Cells: Sampling,

Chapter 15 Ultrafast Electrophoretic Separations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445

Chapter 16 DNA Sequencing by Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467

Chapter 17 Dynamic Computer Simulation Software for Capillary Electrophoresis . . . . . . . . . 515

Chapter 18 Heat Production and Dissipation in Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . 545

Chapter 19 Isoelectric Focusing in Capillary Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 563

Part IIB Capillary-Based Systems: Specialized Methods and Technologies . . . . . 581

Chapter 21 Chemical Cytometry: Capillary Electrophoresis Analysis at the Level of the

Chapter 22 Glycoprotein Analysis by Capillary Electrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 631

Chapter 23 Capillary Electrophoresis of Post-Translationally Modified Proteins

Chapter 24 Extreme Resolution in Capillary Electrophoresis: UHVCE,

Chapter 25 Separation of DNA for Forensic Applications Using Capillary

Chapter 26 Clinical Application of CE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 785

Chapter 27 Solid-Phase Microextraction and Solid-Phase Extraction with Capillary

Chapter 28 CE-SELEX: Isolating Aptamers Using Capillary Electrophoresis . . . . . . . . . . . . . . . 825

Chapter 29 Microfluidic Technology as a Platform to Investigate Microcirculation . . . . . . . . . 841

Chapter 30 Capillary Electrophoresis Applications for Food Analysis. . . . . . . . . . . . . . . . . . . . . . . . 853

Chapter 31 Separation Strategies for Environmental Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 913

Part IIIA Microchip-Based: Core Methods and Technologies . . . . . . . . . . . . . . . . . . . . . . 979

Chapter 33 Multidimensional Microfluidic Systems for Protein and Peptide

Chapter 34 Microchip Immunoassays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1013

Chapter 35 Solvent Extraction on Chips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1021

Chapter 36 Electrophoretic Microdevices for Clinical Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . 1037

Chapter 37 Advances in Microfluidics: Development of a Forensic Integrated DNA

Chapter 38 Taylor Dispersion in Sample Preconcentration Methods . . . . . . . . . . . . . . . . . . . . . . . . . 1085

Chapter 39 The Mechanical Behavior of Films and Interfaces in Microfluidic Devices:

Chapter 40 Practical Fluid Control Strategies for Microfluidic Devices . . . . . . . . . . . . . . . . . . . . . . 1153

Chapter 41 Low-Cost Technologies for Microfluidic Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . 1169

Chapter 42 Microfluidic Reactors for Small Molecule and Nanomaterial

Part IIIB Microchip-Based: Specialized Methods and Technologies . . . . . . . . . . . . . . 1205

Chapter 45 Optical Detection Systems for Microchips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1253

Chapter 46 Microfabricated Electrophoresis Devices for High-Throughput Genetic

Chapter 48 Microdialysis and Microchip Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1327

Chapter 49 Microfluidic Sample Preparation for Proteomics Analysis Using

Chapter 50 Implementing Sample Preconcentration in Microfluidic Devices . . . . . . . . . . . . . . . . 1375

Chapter 51 Using Phase-Changing Sacrificial Materials to Fabricate Microdevices for

Chapter 54 Nanoscale Self-Assembly of Stationary Phases for Capillary

Chapter 55 Nanoscale DNA Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1527

clinical and forensic applications, solid-phase microextraction/online sample preconcentration, and

Norberto A. Guzman, Ph.D.

Prof. Dr. Albert van den Berg

Matthew R. Begley, Ph.D.

Kevin Altria Maxim V. Berezovski

Wendell Karlos Tomazelli Coltro Christopher J. Easley

Jennifer A. Coyne Simon Ekström

Michel Girard Katie M. Horsman

Carlos F. Gonzalez Tiemin Huang