LOAN DOCUMENT___
______
PHOTOGRAPH THIS SHEET
INVENTORY
LEVEL
DOCUMENTMIENTIFICTION
I~~
H
s'pr
r
Neeug
~Approved for puiblic
DiatrbutionUnhjn~tedD
L
DISTRIBTION STATEMENT
E
GAMI
RTS
vnrc
TRAC
UNANNOUNCED0
DI
JUSTIFICATION
___
__LECTE
___
__
i
NOV191990
DISTRIBUJTION/
AVAILk'!LITY CODES
AV~n.AB~ ff
ANDI
ACCE SSIO NEDH
MCIAL.DATE
A
R
E
DISTRIUTION STAMP
DATE RETURNED
REGISTERED OR CERTIIE
DATE RECEIVED INDTIC
NUMER
111010(APH THIS SHEET AND RETURN TO DTIC-JDAC
DTIO
7A
?C
oOaUMM
FSOCESWE
81=~
LOAN DOCUMNT
mou
hATNX
mm-Au%
p~..
.w
.
Vo
Jame
U
I
171
E. Mar
.
-w.---.--.---.
1
REPORT DOCUMENTATION PAGE
n m . sN g t g,
A W e 1i
Of t4 Om u l e
nqm e.vm sl@ow,
Officat
e t o asmae
,V U S R
utIS,
1. AGENCY US1 ONLY (Leeve
2 IMPORT DATE
, Spt mber 1990
4L TMIT ANO SUSM-LE
iS.
Polymer Based Molecular Composites
i9 ow ti m to t #
1 AMvod
ORS N. 070C401L
41 ,
WRf
UP
IJ
l~
OmOw]. t
0€U~
m
3. R-oRT TYPE ANDODATES COVERED
26 Nov 89 - 25 Nov 90
Final
FUNOING NUMBERS
2303
61102F
A3
4L AUTHOWS1
Dale W. Schaefer, and James E. Mark
7.PWO MI
CGANIZATION NAME(S) AND AOOIESS(1S)
Research Society
L PIRFOMN ORGANIZATION
REPORT NUMBERi
Materials
9800 McKnight Road, Suite 327
Pittsburgh, PA 15237-6005
. SPONSORM/IMONITORi
AGENCY NAME(S) AND AOORIISS(ES)
AFOSR/NC
Building 410
Bolling AFB, DC
11. SIJPPUEtwA
AFO
*'R-
90
1054
I.LSFONSOAINM/MONTOING
AFOSR-90-0089
20332-6448
NOTES
12. DISTRIUTIONI AVAILAIUJTY STATEMENT
j12
OISTRIU
CM
Approved for public relaese; distribution is unlimited
13. ABSTRACT (P
m.-n 200 b-avIW
A symposium entitled "Polymer Based Molecular Composites" was organized as
part of the Materials Research Society Fall Meeting Held November 27-30, 1989
A total of 57 papers were presented during the
in Boston, Massachusetts.
(1)
The papers were arranged in the following eight categories:
symposium.
(4)
(3) Rigid-Flexible Systems;
(2) Emulsions/Blocks;
Inorganics/Emulsions;
Properties;
Synthesis/Electrooptical
(6)
lonomers/Structure;
(5)
Blends/IPN's;
Conventional
Miscellaneous/
(8)
Properties;
Interfaces/Mechanical
(7)
with
organizers
symposium
the
by
recognized
were
papers
Two
Composites.
symposium
the
in
papers
other
Two
papers.
contributed
awards as outstanding
were recognized by the Materials Research Society with Graduate Student Awards
All papers appear in their entirety in the Materials
to their presenters.
Research Society Symposium Proceedings, Volume 171, edited by Dale W. Schaefer
Society,
Research
Materials
the
by
published
and
Mark,
E.
James
and
Pittsburgh, PA.
SUSIRTTim"1I.
17. SICUF- CLASS iCATION
18. SECURITY CLASSIFICATION
I REPOW
-T _1
OF THIS PAGE
NUMBER (W PAGIS
i1. SECURUtY O.ASSWICATION
I
Of ABSTRACT
0Uft'ATIN OF ABSTRA
Approvort
W OfFFCE OF SCITI~rFIC RUEAKW4AM)C
cpcnTnical repari USs been rPvltwed And il
,nftP' ublir %elp&S- TAW AF 194-J2
i ,,c!I fmitpv.
r'.Cllude -vender
4et, reChvn1CAI fnforimation Oiviriom
BEST
AVAILABLE COPY
MATERIALS RESEARCH SOCIETY SYMPOSIUM PROCEEDINGS
VOLUME 171
Polymer Based Molecular Composites
Symposium held November 27-30, 1989, Boston,
Massachusetts, U.S.A.
EDITORS:
Dale W. Schaefer
Sandia National Laboratories, Albuquerque, New Mexico, U.S.A.
James E. Mark
University or Cincinnati, Cincinnati, Ohio U.S.A.
MMATERIALS
RESEARCH SOCIETY
Pittsburgh, Pennsylvania
This work was supported by the Air Force Office of Scientific Research, Air Force
Systems Command, USAF, under Grant Number AFOSR 90-0089.
This work was supported in part by the U.S. Army Research Office under Grant
Number DAAL03-90-G-0014. The views, opinions, and/or findings contained in this
report are those of the authors and should not be construed as an official Department of
the Army position, policy, or decision unless so designated by other documentation.
Single article reprints from this publication are available through University Microfilms
Inc., 300 North Zeeb Road, Ann Arbor. Michigan 48106
CODEN:
MRSPDH
Copyright 1990 by Materials Research Society.
All rights reserved.
This book has been registered with Copyright Clearance Center, Inc. For further
information, please contact the Copyright Clearance Center, Salem, Massachusetts.
Published by:
Materials Research Society
9800 McKnight Road
Pittsburgh, Pennsylvania 15237
Telephone (412) 367-3003
Fax (412) 367-4373
Library of Congress Cataloging in Publication Data
Polymer based molecular composites : symposium held November 27-30, 1989,
Boston, Massachusetts, U.S.A. /editors, Dale W. Schaefer, James E. Mark.
p. cm. - (Materials Research Society symposium proceedings : ISSN 02729172; v. 171)
Includes bibliographical references.
ISBN 1-55899-059-3
11.Mark, James E.,
1. Polymeric composites-Congresses. I. Schaefer, Dale W.
.
Ill. Materials Research Society. IV. Series: Materials Research Society
1943symposium proceedings : v. 171.
TA455.P58P67
620.1'92-dc2O
1990
Manufactured in the United States of America
90-34129
CIP
Contents
PREFACE
xi
GRADUATE STUDENT AWARD WINNERS
xiii
ACKNOWLEDGMENTS
xv
MATERIALS RESEARCH SOCIETY SYMPOSIUM PROCEEDINGS
PART I:
xvii
INORGANICS/EMULSIONS
*INORGANIC-ORGANIC COMPOSITES BY SOL-GEL TECHNIQUES
Helmut Schmidt
3
*THE SYNTHESIS, STRUCTURE AND PROPERTY BEHAVIOR OF
INORGANIC-ORGANIC HYBRID NETWORK MATERIALS PREPARED
BY THE SOL GEL PROCESS
G.L. Wilkes, A.B. Brennan, Hao-Hsin Huang,
David Rodrigues, and Bing Wang
15
THE CATALYTIC SYNTHESIS OF INORGANIC POLYMERS FOR HIGH
TEMPERATURE APPLICATIONS AND AS CERAMIC PRECURSORS
31
Jeffrey A. Rahn, Richard M. Laine, and Zhi-Fan Zhang
CONDUCTING MOLECULAR MULTILAYERS:
INTERCALATION OF
CONJUGATED POLYMERS IN LAYERED MEDIA
V. Mehrotra and E.P. Giannelis
39
NYLON 6-CLAY HYBRID
Akane Okada, Masaya Kawasumi, Arimitsu Usuki,
Yoshitsugu Kojima, Toshio Kurauchi, and
Osami Kamigaito
45
REINFORCEMENT OF ELASTOMERS BY THE IN-SITU GENERATION
OF FILLER PARTICLES
James E. Mark and Dale W. Schaefer
51
STRUCTURE OF MICROPHASE-SEPARATED SILICA/SILOXANE
MOLECULAR COMPOSITES
Dale W. Schaefer, James E. Mark, David McCarthy,
Li Jian, C.-C. Sun, and Bela Farago
57
NMR IMAGING OF SILICA-SILICONE COMPOSITES
Leoncio Garrido, Jerome L. Ackerman, and
James E. Mark
65
*SYNTHETIC POLYMERS IN WATER-IN-OIL MICROEMULSIONS
Francoise Candau
71
POLYMER-DERIVED Si 3 N4 /BN COMPOSITES
Wayde R. Schmidt, William J. Hurley, Jr.,
Vijay Sukumar, Robert H. Doremus, and
Leonard V. Interrante
79
*Invited Papez
I
"
I
/
mum •
m
m
•
M
•
*
n
-
" '
,V
PART II:
EMULSIONS/BLOCKS
*STABILIZED NANOPARTICLES OBTAINED FROM SYNTHETIC
POLYMERIZABLE MICELLES AND VESICLES
Constantinos M. Paleos
THE PHYSICAL PROPERTIES OF MICROCELLULAR COMPOSITE FOAMS
Alice M. Nyitray, Joel M. Williams, David Onn, and
Adam Witek
*SYNTHESIS AND PROPERTIES OF COPOLYMERS OF DIPHENYLSILOXANE
WITH OTHER ORGANOSILOXANES
J. Ibemesi, N. Gvozdic, M. Kuemin, Y. Tarshiani, and
D.J. Meier
DYNAMIC IR STUDIES OF MICRODOMAIN INTERPHASES OF
1SO'OPE-LABELED BLOCK COPOLYMERS
I. Noda, S.D. Smith, A.E. Dowrey, J.T. Grothaus,
and C. Marcott
PART III:
87
99
105
117
RIGID-FLEXIBLE SYSTEMS
*LtGHT SCATTERING STUDIES OF THE STATE OF DISPERSION IN
MOLECULAR COMPOSITES
Benjamin S. Hsiao, Richard S. Stein,
Silvie Cohen Addad, Russell Guadiana,
and Norman Weeks
*RECENT ADVANCES IN THE MORPHOLOGY AND MECHANICAL
PROPERTIES OF RIGID-ROD MOLECULAR COMPOSITES
125
131
Stephen J. Krause and Wen-Fang Hwang
RHEOLOGY OF BLENDS OF A RODLIKE POLYMER (PBO) AND ITS
FLEXIBLE CHAIN ANALOG
V.J. Sullivan and G.C. Berry
PBZT MICROCOMPOSITES WITH ADVANCED THERMOPLASTIC MATRICES
W. Michael Sanford and Gerard M. Prilutski
141
147
PBZT/POIYAMIDE THERMOPLASTIC MICRO-COMPOSITES - AN
OUTGROWTH OF MOLECULAR COMPOSITES DEVELOPMENT
William C. Uy and E.R. Perusich
153
MORPHOLOGY AND FORMATION OF FIBRILLAR STRUCTURE IN
PRO FIBER
C.C. Chau, J.H. Blackson, H.E. Klassen, and
159
W.-F. Hwang
IN STTU COMPOSITES BASED ON THERMOTROPIC AND FLEXIBLE
P11[YMER i
cuido Cievecoeur and Gabtiel Groeninckx
r'101,I(vr.Ak COMPOSITES OF RODLIKE/FLEXIBLE POLYIMIDES
C.R. Rojstaczei, D.Y. Yoon, W. Volksen, and
!'.A. Smith
VI
165
171
EQUILIBRIUM AND NON-EQUILIBRIUM PHASES AND PHASE
DIAGRAMS IN BLENDS OF POLYMER LIQUID CRYSTALS WITH
ENGINEERING POLYMERS
Witold Brostow, Theodore S. Dziemianowicz,
Michael Hess, and Robert Kosfeld
177
STRUCTURE AND PROPERTIES OF BLENDS OF POLYCARBONATE
AND POLY (ETHYLENE-TEREPHTHALATE-CO-p-HYDROXYBENZOATE):
PHASE DIAG;RAM AND MECHANICAL BEHAVIOR
Robert Kosfeld, Frank Schubert, Michael Hess,
and Witold Brostow
183
CHARACTERIZATION OF POLYQUINOLINE BLENDS USING SMALL
ANGLE SCATTERING
Wen-Li Wu, John K. Stille, Joseph W. Tsang, and
Alex J. Parker
PART IV:
189
BLENDS/IPN'S
*MISCIBILITY IN BLENDS OF POLYBENZIMIDAZOLE AND FLUORINE
CONTAINING POLYIMIDES
Hiroaki Yamaoka, Norman E. Aubrey,
William J. MacKn~ght, and Frank E. Karasz
*SMALL ANGLE NEUTRON SCATTERING STUDIES OF BLENDS OF
PROTONATED LINEAR POLYSTYRENE WITH CROSSLINKED
DEUTERATED POLYSTYRENE
Robert M. Briber and Barry J. Bauer
DYNAMICS OF PHASE SEGREGATION IN POLY-P-PHENYLENE
TEREPHTHALAMIDE AND AMORPHOUS NYLON MOLECULAR COMPOSITES
Thein Kyu, Jan Chang Yang, and Tsuey Ing Chen
197
203
211
FACTORS INFLUENCING PRCPERTIES OF SAN/PMMA BLENDS
R. Subramanian, Y.S. Huang, J.F. Roach, and
D.R. Wiff
217
DIELECTRIC STUDIES OF POLYESTER/POLYCARBONATE BLENDS
James M. O'Reilly and Joseph S. Sedita
225
*INTERPENETRATING POLYMER NETWORKS
TOPOLOGICAL ISOMERS
Harry L. Frisch
PART V:
AND RELATED
231
IONOMERS/STRUCTURE
SMALL ANGLE X-RAY SCATTERING ON POLY(ETHYLENE-METHACRYLIC
ACID) LEAD AND LEAD SULFIDE IONOMERS
Benjamin Chu, Dan Q. Wu, and Walter Mahler
*EXCIMER AND EXCITON FUSION OF BLENDS AND MOLECULARLY
DOPED POLYMERS--A NEW MORPHOLOGICAL TOOL
Zhong-You Shi, Ching-Shan Li, and Raoul Kopelman
*Invited Paper
v'i
237
245
THE ORDERED BICONTINUOUS DOUBLE DIAMOND STRUCTURE IN
BINARY BLENDS OF DIBLOCK COPOLYMER AND HOMOPOLYMER
Karen I. Winey and Edwin L. Thomas
STUDIES ON THE EXCESS FREE ENERGY AND THE EARLY STATE OF
SPINODAL DECOMPOSITION OF THE BLEND d-PS/PVME AND THE
ISOTOPIC BLEND d-PS/PS WITH SMALL ANGLE NEUTRON SCATTERING
255
261
D. Schwahn, T. Springer, K. Hahn, and J. Streib
MECHANICAL PROPERTIES AND STRUCTURE OF MELAMINE
FCRMALDEHYDE/POLY(VINYL
ALCOHOL)
MOLECULAR COMPOSITES
267
Kecheng Gong and Xinghua Zhang
PART VI:
SYNTHESIS/ELECTRO-OPTICAL PROPERTIES
*MORPHOLOGICAL CONSEQUENCES OF CATALYTIC HYDROGENATION OF
POLYMERS IN THE BULK
Laura R. Gilliom, Dale W. Schaefer, and James E. Mark
275
SYNTHETIC CONTROL OF MOLECULAR STRUCTURE IN ORGANIC
AEROGELS
Richard W. Pekala
285
SYNTHETIC PROCEDURES FOR PREPARING CROSS-LINKABLE
ACRYLIC COMB-LIKE COPOLYMERS VIA MACROMONOMERS
Gang-Fung Chen and Frank N. Jones
293
SYNTHESIS AND CHARACTERIZATION OF SEGMENTED COPOLYMERS
OF A METHYLATED POLYAMIDE AND A THERMOTROPIC LIQUID
CRYSTALLINE POLYESTER
Gregory T. Pawlikowski, R.A. Weiss, and S.J. Huang
299
*AGGREGATION STRUCTURE AND ELECTRO-OPTICAL PROPERTIES OF
(LIQUID CRYSTALLINE POLYMER) / (LOW MOLECULAR WEIGHT
LIQUID CRYSTAL)
COMPOSITE SYSTEM
305
Tisato Kajiyama, Hirotsugu Kikuchi, Akira Miyamoto,
Satoru Moritomi, and Jenn-Chiu Hwang
PART VII:
INTERFACES/MECHANICAL PROPERTIES
*DIBLOCK COPOLYMERS AT SURFACES
Peter F. Green, Thomas M. Christensen,
Thomas P. Russell, and Spiros H. Anastasiadis
ON THE SCALE OF DIFFUSION LENGTHS OBSERVABLE BY NEUTRON
REFLECTION: APPLICATION TO POLYMERS
A. Karim, A. Mansour, G.P. Felcher, and T.P. Russell
329
NEUTRON REFLECTION STUDY OF SURFACE ENRICHMENT IN AN
ISOTOPIC POLYMER BLEND
R.A.[.. Jones, L.J. Norton, E.J. Kramer, R.J. Composto,
R.S. Stein, T.P. Russell, G.P. Felcher, A. Mansour,
335
and A.
*
317
Karim
td:tt Papet
Vil
X-RAY REFLECTIVITY AND FLUORESCENCE MEASUREMENTS FROM
POLYSTYRENE-CO-BROMOSTYRENE/POLYSTYRENE INTERFACES
J. Sokolov, M. Rafailovich, X. Zhao, W.B. Yun,
R.A.L. Jones, E.J. Kramer, R.J. Composto, R.S. Stein,
A. Bommannavar, and M. Engbretson
INTERFACIAL SEGREGATION EFFECTS IN MIXTURES OF
HOMOPOLYMERS WITH COPOLYMERS
Vijay S. Wakharkar, Thomas P. Russell, and
Vaughn R. Deline
A NEW VARIABLE ANGLE FT-IR ELLIPSOMETER
J.L. Stehle, O.T. Thomas, J.P. Piel, P. Evrard,
J.H. Lecat, and L.C. Hammond
337
343
349
POLYMER MOLECULES AT INTERFACES:
STUDIES BY SMALL-ANLE
NEUTRON SCATTERING
W.C. Foraman, B.E. Latshaw, and D.T. Wu
355
MECHANICALLY INDUCED SILICA-SILOXANE MIXTURES.
STRUCTURE OF THE ADSORBED LAYER AND PROPERTIES OF THE
NETWORK STRUCTURE
J.P. Cohen-Addad
365
THE EFFECT OF MASKED ISOCYANATES ON THE MECHANICAL
PROPERTIES OF MY720/DDS EPOXY RESIN
N. Rungsimuntakul, S.V. Lonikar, R.E. Fornes,
and R.D. Gilbert
A STUDY OF SHORT METAL FIBER REINFORCED COMPOSITE
MATERIALS
W.C. Chung
PART VIII:
371
379
MISCELLANEOUS/CONVENTIONAL COMPOSITES
*DEFORMATION BEHAVIOR OF POLYMER GELS IN ELECTRIC FIELD
Toshio Kurauchi, Tohru Shiga, Yoshiharu Hirose,
and Akane Okada
BIAXIAL EXTRUSION OF POLYIMIDE LARC-TPI AND LARC-TPI
BLENDS
R. Ross Haghighat, Lucy Elandjian, and
Richard W. Lusignea
STRUCTURAL STUDIES OF SEMIFLEXIBLE FLUOROCARBON CHAINS
CONTAINING AN AROMATIC CORE
A. Schulte, V.M. Hallmark, R. Twieg, K. Song,
and J.F. Rabolt
389
395
401
THE EFFECT OF LOW POWER AMMONIA AND NITROGEN PLASMAS ON
CARBON FIBRE SURFACES
C. Jones and E. Sammann
407
DETERMINATION OF PARTICLE SIZE OF A DISPERSED PHASE BY
SMALL-ANGLE X-RAY SCATTERING
Frank C. Wilson
413
*Invited Paper
ix
ifSUBJECT
SYNTHESIS AND CHARACTERIZATION OF A THERMOTROPIC
POLYALKANOATE OF 4,4-DIYDROXY-a.t'DIMETHYLBENZAiAZINE
H. Friutwala, A.L. Cimecioglu, and R.A. Weiss
419
AUTHOR INDEX
427
INDEX
MATERIALS RESEARCH SOCIETY SYMPOSIUM PROCEEDINGS4-
1x
429
Preface
The development of polymeric materials over the last 30
years is directly traceable to the chemist's ability to
manipulate structure and tailor the properties of organic
materials for specific applications. The phenomenal success
of polymer chemistry, of course, rests on a half century of
small-molecule chemical research.
In the current technological environment, demands for organic materials often
involve competing requirements which cannot be met by
single-component polymers. Although multicomponent systems
are often proposed to meet these competing demands, the
complex interrelation between chemical and physical factors
makes it difficult to synthesize these materials much less
optimize their properties. It has become clear, therefore,
that a research base is needed in multiphase polymeric
materials similar to that which underlies single-phase
organics. This symposium was organized to encourage the
development of such a research base.
When we first proposed this symposium, we were anticipating a one or two day starter symposium.
In fact we
received over 80 papers from all over the world indicating
the tremendous interest in molecular composites.
In
addition, the meeting supported another closely-related
symposium titled "Multifunctional Materials," which was
equally large.
The symposium was deliberately organized to include not
only "traditional molecular composites (miscible blends of
rod and coil molecules,)" but also to include a variety of
other alloy systems from blends to copolymers' to microphase-separated systems. To some extent this broad definition recognizes the fact that true molecularly miscible
rod/coil systems are exceedingly rare.
The idea of materials development by intimately mixing poiymers with
fundamentally different properties, however, remains viable
if phase separation can be restricted to microscopic scales.
We were delighted that two graduate student award winners
presented their papers in Symposium 0.
Karen Winey,
University of Massachusetts, received the award for her
paper titled "The Ordered Bicontinuous Double Diamond
Structure in Blends of Diblock Copolymer and Homopolymer,"
and Dan Q. Wu, State University of New York at Stony Brook,
received the award for his paper titled "Small Angle X-Ray
Scattering on Poly(Ethylene-Methacrylic Acid) Lead and Lead
Sulfide Ionomers." In addition, Dr. Wu chaired the session
on Synthesis/Electro-Optical Properties.
Two additional contributions were recognized by the
symposium organizers as outstanding contributed papers.
These papers were given a $500 grant to cover travel
expenses at the meeting. The two outstanding papers were
0-2.8, "Dynamic IR
Studies of Microdomain Interphases of
Isotope-Labeled Block Copolymers," by I. Noda, S.D. Smith,
Xi
p
A.E. Dowrey, J.T. Grothaus, and C. Marcott from the Procter
& Gamble Company, and 0-6.4, "Synthesis and Characterization
of Segmented Copolymers of a Methylated Polyanide and a
Thermotropic Liquid Crystalline Polyester," by Gregory T.
Pawlikowski, R.A. Weiss, and S.J. Huang of the University
of Connecticut.
December 1989
Dale W. Schaefer
James E. Mark
Ai
I'
GRADUATE STUDENT AWARD WINNERS
Fall 1989
Symposium 0
1(aren Winey
Dan Q. Wu
xiii
Acknowledgments
following
Symposium received generous support from the
institutions.
Air Force office of Scientific Research
Allied Signal, Inc.
Army Research Office
Dow Chemical Company
Dow Corning Corp.
E.I. duPont de Nemours & Co., Inc.
IBM Corp.
Proctor & Gamble Company
Rhone-Poulenc, Inc.
Union Carbide Corporation
xv
MATERIALS RESEARCH SOCIETY SYMPOSIUM PROCEEDINGS
Recent Materials Research Society Symposium Proceedings
Volume 145-I11-V Heterostructures for Electronic/Photonic Devices, C.W. Tu,
V.D. Mattera, A.C. Gossard, 1989, ISBN: 1-55899-018-6
Volume 146-Rapid Thermal AnnealinglChemical Vapor Deposition and Integrated
Processing, D. Hodul, J. Gelpey, M.L. Green, T.E. Seidel, 1989,
ISBN: 1-55899-019-4
Volume 147-on Beam Processing of Advanced Electronic Materials, N.W. Cheung,
A.D.Marwick, J.B. Roberto, 1989, ISBN: 1-55899-020-8
Volume 148-Chemistry and Defects in Semiconductor Heterostructures, M. Kawabe,
T.D. Sands, E.R. Weber, R.S. Williams, 1989, ISBN: 1-55899-021-6
Volume 149-Amorphous Silicon Technology-1989, A. Madan, M.I. Thompson,
P.C. Taylor, Y. Hamakawa, P.G. LeComber, 1989, ISBN: 1-55899-022-4
Volume 150-M.terials for Magneto-Optic Data Storage, C.J. Robinson, T. Suzuki,
C.M.Falco, 1989, ISBN: 1-55899-023-2
Volume 151-Growth, Characterization and Properties of Ultrathin Magnetic Films and
Multilayers, B.T. Jonker, J.P. Heremans, E.E. Marinero, 1989,
ISBN: 1-55899-024-0
Volume 152-Optical Materials: Processing and Science, D.B. Poker, C. Ortiz, 1989,
ISBN: 1-55899-025-9
Volume 153-Interfaces Between Polymers, Metals, and Ceramics, B.M. DeKoven,
A.J. Gellman, R. Rosenberg, 1989, ISBN: 1-55899-026-7
Volume 154-Electronic Packaging Materials Science IV, R. jaccodine, K.A. Jackson,
ED. Lillie, R.C. Sundahl,
1989, ISBN: 1-55899-0.7-5
Volume 155-Processing Science of Advanced Ceramics, I.A. Aksay, G.L. McVay,
D.R. Ulrich, 1989, ISBN: 1-55899-028-3
Volume 156-High Temperature Superconductors: Relationships Between Properties,
Structure, and Solid-State Chemistry, JR. Jorgensen, K. Kitazawa,
J.M. Tarascon, M.S. Thompson, J.B. Torrance, 1989, ISBN: 1-55899-029
Volume 157-Beam-Solid Interactions: Physical Phenomena, J.A. Knapp, P. Borgesen,
R.A. Zuhr, 1989, ISBN 1-55899-045-3
Volume 158-In-Situ Patterning: Selective Area Deposition and Etching, R. Rosenberg,
A.F. Bernhardt, J.G. Black, 1989, ISBN 1-55899-046-1
Volume 159-Atomic Scale Structure of Interfaces, R.D. Bringans, R.M. Feenstra,
J.M. Gibson, 1989, ISBN 1-55899-047-X
Volume 160-Layered Structures: Heteroepitaxy, Superlattices, Strain, and
Metastability, B.W. Dodson, L.J. Schowalter, J.E. Cunningham,
F.H. Pollak, 1989, ISBN 1-55899-048-8
Volume 161-Properties of Il-VI Semiconductors: Bulk Crystals, Epitaxial Films,
Quantum Well Structures and Dilute Magnetic Systems, J.F. Schetzina,
F.J.
Bartoli, Jr., H.F. Schaake, 1989, ISBN 1-55899-049-6
Volume 162-Diamond, Boron Nitride, Silicon Carbide and Related Wide Bandgap
Semiconductors, JT.Glass, R.F. Messier, N. Fujimori, 1989,
ISBN 1-55899-050-X
Volume 163-Impurities, Defects and Diffusion in Semiconductors: Bulk and Layered
Structures, J. Bernholc, E.E. Hailer, D.J. Wolford, 1989,
ISBN 1-55899-051-8
Volume 164-Materials Issues in Microcrystalline Semiconductors,
P.M. Fauchet, C.C. Tsai, K. Tanaka, 1989, ISBN 1-55899-052-b
Volume 165-Characterization of Plasma-Enhanced CVD Processes, G. Lucovsky.
D.E. Ibbotson, D.W. Hess, 1989, ISBN 1-55899-053-4
Volume 166-Neutron Scattering for Materials Science, S.M. Shapiro, S (
JD.Jorgensen, 1989, ISBN 1-55899-054-2
Mis%,
MATERIALS RESEARCH SOCIETY SYMPOSIUM PROCEEDINGS
Volume 167-Advanced Electronic Packaging Materials, A. Barfknecht, 1. Partridge,
C-Y. Li, C.J. Chen, 1989, ISBN 1-55899-055-0
Volume 168-Chemical Vapor Deposition of Refractory Metals and Ceramics,
T.M. Besmann, B.M. Gallois, 1989, ISBN 1-55899-056-9
Volume 169-High Temperature Superconductors: Fundamental Properties and Novel
Materials Processing, J. Narayan, C.W. Chu, L.F. Schneemeyer,
D.K. Christen, 1989, ISBN 1-55899-057-7
Volume 170-Tailored Interfaces in Composite Materials, C.G. Pantano, E.J.H. Chen,
1989, ISBN 1-55899-058-5
Volume 171-Polymer Based Molecular Composites, D.W. Schaefer, J.E. Mark, 1989,
ISBN 1-55899-059-3
Volume 172-Optical Fiber Materials and Processing, J.W. Fleming, G.H. Sigel,
S. Takahashi, P.W. France, 1989, ISBN 1-55899-060-7
Volume 173-Electrical, Optical and Magnetic Properties of Organic Solid-State
Materials, L.Y. Chiang, D.O. Cowan, P. Chaikin, 1989,
ISBN 1-55899-061-5
Volume 174-Materials Synthesis Utilizing Biological Processes, M. Alper, P.D. Calvert,
P.C. Rieke, 1989, ISBN 1-55899-062-3
Volume 175-Multi-Functional Materials, D.R. Ulrich, F.E. Karasz, A.J. Buckley,
G. Gallagher-Daggitt, 1989, ISBN 1-55899-063-1
Volume 176-Scientific Basis for Nuclear Waste Management XIII, V.M. Oversby,
P.W. Brown, 1989, ISBN 1-55899-064-X
Volume 177-Macromolecular Liquids, C.R. Safinya, S.A. Safran, P.A. Pincus, 1989,
ISBN 1-55899-065-8
Volume 178-Fly Ash and Coal Conversion By-Products: Characterization, Utilization
and Disposal VI, F.P. Glasser, R.L. Day, 1989, ISBN 1-55899-066-6
Volume 179-Specialty Cements with Advanced Properties, H. Jennings, A.G. Landers,
B.E. Scheetz, I. Odler, 1989, ISBN 1-55899-067-4
MATERIALS RESEARCH SOCIETY MONOGRAPH
Atom Probe Microanalysis: Principles and Applications to Materials Problems,
M.K. Miller, G.D.W. Smith, 1989; ISBN 0-931837-99-5
Earlier Materials Research Society Symposium Pro eedings li.ted in the bail
F-mmm
m
m
m
i
m
l
m,.
mmm
PART I
Inorganicsl/Emulsions
3
INORGANIC-ORGANIC COMPOSITES BY SOL-GEL TECHNIQUES
Helmut SCHMIDT
Fraunhofer-Institut fUr Silicatforschung, Neunerplatz 2, D-8700
WUrzburg, Federal Republic of Germany
ABSTRACT
The sol-gel process opens the possibility of combining inorganic and organic units to new hybrid polymers. Organic units
can be used for structural modification of the inorganic backbone, for creating new functions within an inorganic network
and for building up organic polymeric chains. The materials show
interesting perspectives with respect to structural (surface
hardness, strength) and functional properties (e. g. diffusion,
photocuring, incorporation of dyes, optical properties). A review over structural and functional properties of sol-gel derived inorganic-organic polymers (ORMOCERs = organically modified ceramics) is given.
INTRODUCTION
The sol-gel process is a synthesis route to inorganic nonmetallic materials. Reactive monomers, oligomers or colloids
can be used as starting materials. They have to be "activated"
in order to undergo a polycondensation step and to form polymeric networks. This can be achieved by various means but, in
general, by creating reactive sMeOH groups, which are able to
form aMe-O-Me- bonds during the condensation step. A convenient
route is the use of alkoxides as precursors which react with
water to hydroxides, condensing spontaneously to polymeric species (1). Similar reactions are well-known from inorganic salts,
forming hydroxides and precipitates (2) by pH change.
mMeOR
nMeOH
,
4
0
---.
+
+
H2 0
HMeX
----
*MeOH
+
+
.MeOH (irreversible step)
ROMew HOME -
(1)
wMeOMew + HOR
MeOMew + H 2 0
R = alkyl
.MeOH (reversible step, pH!)
XMe
HOMe
-
nMeOMem
MeOMew
+
+
HX
H2 0
(2)
X = anion (CI-, N03 ,..)
Another
routein is
the destabilization
colloidal
sols by
pH change
either
organic
solvents or inofwater.
All these
reactions lead to gels, containing solvent or air after drying.
Due to the (in general) high specific surface areas of these
gels, they contain adsorbed molecules from the processing steps
(water or organics) or, in the case of the alkoxide route, unhydrolysed alkoxy groups. Organics are oxidized or pyrolyzed
during heat treatments of gels, converting them into glasses or
ceramics. Heat treatments are necessary to enhance diffusion
Mat. Res. Soc. Symp. Proc. Vol. 171.
IM Materials Resarch Society
4
processes (for crystallization) or to decrease viscosity (ror
glass formation) in gels. The viscosity of inorganic gels is
high due to the three dime~sional crosslinking of the inorganic
leading to the typical
building units (e. g. Si0 4 -tetrahedron)
brittleness of ceramics.
Decreasing network connectivity by introducing organic
groupings can cause a remarkable decrease of viscosity of systems with pure inorganic backbones and leads to dense materials
at temperatures between 50 and 150 *C, as shown in (1-2]. An
analogue effect can be observed in inorganic glasses where the
introduction of inorganic network modifiers leads to lower viscosities compared to fused silica. The combination of organic
groupings with ceramics can take place via various routes, e. g.
mechanical mixing, infiltration and chemical synthesis as indicated in figure 1. Mechanical processes are the most common and
most economical processes, but the chemical routes offer new and
interesting perspectives. Linking organic units to inorganic
compaosie
polymer
fler
gas phas'e
and liquid
nm:
poru
ceramics, gels
fiber%
nm ao
orgni
monomers
swelinog of
polymer$
comSposite
morlerst
cuna
inorganic
_M_
Zrpep
__
_
01,20mer.,
colloids
compote@
tible
gets
Figure 1. Routes to inorganic organic composites.
ones provides the possibility of controlling the size scale of
inorganic and organic domaines in the final product but also to
keep them on the liquid-liquid mixing level (molecular or ohigomer). Therefore, crosslinking and phase s;eparation mechanisms
have to be controlled. The simplest way to prevent separation of
inorganic from organic units is to have them attached together
by chemical bonds, for example in precursors like substituted
alkoxy silanes, with appropriate functionial groupings, alkoxides
with complex formers or salts of organic acids (figure 2).
-Ne-CE
He = Si, Sn, Ge ...
a bond
BAe-O-Cm
me = Si,. .....
a bond
-me+_-OO-CS
He = Al, Zr, Ti, ..
ionic bond
=Me -L-C-
Me = AI,
coordination
Zr, Ti, Cu, Pt ...
bond
Figure 2. Chemical links between
organic and inorganic units.
Then, the inorganic crosslinking reaction can take place by
hydrolysis and condensation of the alkoxy group via a typical
sol-gel reaction (equation 3). The properties of the reaction
R
RMe(OR) 3 - RMe(OH) 3
-
R
R
-Ae-O-ge-O-ge-O-
(3)
product depend strongly on the type and the number of R per inorganic unit. In silicones, RSim and R2 Si= units are used [3].
The uSi-Cm bond can be used as a general link between the
inorganic and the organic side. For R corresponding to a nonreactive organic group, a pure organic network modification
results. This principle is used for spin-on glasses [4] which
can be densified at temperatures as low as 200 - 300 C, compared to 800 - 900 °C of sol-gel silica and applied in thicknesses of more than 20 om. This type of material represents the
group of organically network modified glasses and ceramics.
For R corresponding to a functional organic group, chemical
functions can be introduced into these networks, leading to special properties like adsorptive properties (5], reactive surfaces to special molecules (6] or sensors [7]. This group represents the type of organofunctional modified glasses or ceramics.
Functional groups which are able to be polymerized lead to the
additional organic chains within the inorganic network and to
hybrid oolvmer (inorganic-organic polymers), equation (4). This
principle can be varied by incorporation of organic monomers as
indicated in equation (5).
The construction principles shown above are all characterized by a chemical bond between the inorganic network and the
organic group. A lack of this bond leads to two independent or
interpenetrating networks. In opposition to the "chemically
linked" case, the possibility of separation of the "organic"
from the "inorganic" phase exists, leading to a composit type
of material, as indicated in figure 1. Examples, therefore, are
given by Mark and Wilkes [8-9] and [10]. These inorganic-organic
phase separated materials show a drastic change in mechanical
properties compared to the original organic polymer (e. g. increase of tensile strength and modulus of elasticity).
6
alene
ihethacrate
0
-C
I
-1
3H'C C-CH-Si-CH2
Ci4C1H-Si-Cfl
CDOCHs
\-coocH 's.
R-\
polymerisation
o
(5)
icmCHsCC
S
(4)
addition
R
0
-
/
0
o
CH2-CH-Si
A
R-\
CH2-CH2-CHj-S -CH 2
00
polyethylen
chain
II
R
Me)-
C
ee.g. COOMI
COOCH
2 cI4C-o
ASPECTS OF PROPERTIES
General asoects
As indicated above, the introduction of organic groupings
into an inorganic network leads to structural variations. On the
average a decrease of the network connectivity number decreases
Tg and increases the thermal expansion coefficient (TCE) a. Figure 3 shows Tg and a as a function of different "organic-inorganic" bonds aMe-R (x) per total number of inorganic groupings.
There is a clear tendency showing that for these properties x
is the dominating parameter.
200... /ZrO 2
$ 100G
0 Epoxyl TiO
0
0- 42SiOITiO2
0
x@S30
31 2S,0
x O iO12 SiO2 2
O
I
i
15
10
0.5
0
X
Figure 3. a and Tg
depending on x.
x
crosslinkilg grfOI4l$
like epouy-, methacroyl-
200-
-CIHS as
100
network
modifir
00
0.5
1,0
X
1.5
2.0
=e-x
*!Me-O-
7
The influence of the type of the organic group does not seem
to play the most important role, but a tendency to a steeper
slope of the Tg function with increasing concentration of crosslinking organic units (figure 3) is observed; for x = < 0.7 no
Tg values could be measured below decomposition, probably due to
the increasing immobilization of the organic group within the
inorganic network.
The densities (compared to the pure inorganic materials)
decrease drastically as a result of the organic modification.
As pointed out in [11], the computed densities (between 2 and
3 g/cm ) differ from the3 measured ones remarkably which are between 1.35 and 1.65 g/cm . This indicates the existence of a
free volume similar to organic polymers [12]. In [13] neutron
scattering experiments show that the glass transition can be
correlated with the chain movement, supporting the hypothesis of
a free volume based low density. The investigations were carried
out with amino modified SiO 2 condensates.
Homogeneity is an important property if optical applications
are of interest. In a single component system without intergranular interfaces optical transparency should be obtained easily.
Organic polymers, however, tend to form crystalline phases thus
distv Oing optical transparency. This can also be induced by
ageing, thermal or mechanical stresses and leads to the question of structural stability. Basically, a three-dimensional
inorganic network should be stable and no structural rearrangement should be possible after sufficient curing. For the preparation of sol-gel derived multicomponent systems one has to
take into consideration phase separation effects based on different rates of reactions of the precursors, e. g. Ti and Si
alkoxides. Adaption of rates is limited and cannot successfully
be used to enable a simple addition of water to a mixture of
Ti (OEt)4 and (CH0)3Si-CH -CH2-O-CH -NH-CH -OJ(Si-epoxy): If
water is added, iO precipitates. Controlled release systems
for water in order ?o avoid high concentration, accompanied by
heavy stirring leads to a partial hydrolysis of the fast reacting components and to a condensation via the alkoxide substitution (6).
x's H2 0
Ti(OEt) 4
........
Ti(OH) 4 - TiO 2 .H 2 0
slow release
Ti(OEt) 4
--------
.....
of water
wMeOR + HOTi(OR) 3
-
(RO) 3 TiOH
(6)
=MeOTi(OR) 3
This is in accordance with Livage's electronegativity driven
reaction scheme [14]. In a composition (molar ratio) of about
30 Ti(OEt) 4, 65 Si-epoxy and 5 Si(OEt)4 , the addition of only
1/16 of the total amount of water is necessary to be added by a
controlled release process (CCC according to [15] or H2 0 doped
silica) to prevent precipitation after addition of the full
amount. The reaction scheme is shown in figure 4. The hydrolysis
rates increase in the older Si(OEt) < Si-epoxy < Ti(OEt) 4 . The
preparation leads to a homogeneous liquid with a general structure proposed in figure 4 which can be used for coatings [16].
ba
The material does not show any phase separation down to 1 nm
(resolution of the TEM).
Si(OR),
Ti/Zr/AI (OR),
(RO)
+
3 Si(CHt30C12CHH-CH
2
R"
Ethanol,
optionol/
+ Acti,
Acid
-J
/
Silicagot
OR
OR
OR
-Ti-O-Si-O-Si"'
O
OR OR
0
0 potymem cham
Pelltrs
loaded wilt,
RO-St-OR
1420+ 0.1 N MC'
hydolyes0
OR
~bym~iso~ionRO
*
*
cnd~soten6R
-Si
-0 -Ti -0
6IR
R'SiOR)3+
r(OR
H20- O R Si(OR)
3 +1HOT(OR)
3
RWi(GR)3
+14KO OR)IOHIR RWSi(OR)
2
Figure 4.
kinetics
(IR):K. Ti(0R)A'
tSi(OR) 3 > KHVR
ing hard coatings.
O-(OR) 3
Reaction scheme for
the preparation of Ti contain-
Similar techniques can be used for combinations with other
alkoxides like Al(OR)3, Zr(OR)4 or Sn(OR) 4 . This demonstrates
that homogeneous multicomponent materials can be prepared by an
appropriate chemical processing.
Structural properties and aspects
One of the most interesting properties of these groups of
compositions is the abrasion resistance if used as coatings
[16). Haze measurements after taber abrader tests show values
close to those of float glass surfaces. The moduli of elasticity
are between 1 and 7 GPa, that means remarkably lower than those
of glass and ceramics. However, the low moduli are an important
prerequirement for coating organic polymer surfaces. First, the
lower modulus can help to overcome stresses resulting from TCE
differences between substrate and coating (high CTE of the coating materials according to figure 3 is helpful, too). Second,
for practical use, coating thicknesses for abrasion resistance
have to be in the range of ? 5 4m, a thickness which can be
achieved for densif led inorganic aol-gel coatings only by a
multiple coating process and which leads to cracks caused by
stresses due to the CTE mismatch dependent on the preparation
technique. As a consequence, there are no pure inorganic thick
protective coatings on soft plastics available. The comparison
of the behavior of ORMOCER coatings on ceramic coatings is
schematically illustrated in figure 5.
The modulus is very sensitive to the concentration of inorganic units and to the type of inorganic network formers and
increases in the order Zr < Ti < Al. The abrasion rqsistance
corresponds to this order. Based on these findings, various
scratch resistant systems have been developed, as described in
(16-173, for example for CR 39, polycarbonates, PC, PMMA, but
also for metals, like brass.
inorganic, brittle
coating
SE =50OGPa
5 An
organically modified
flexible coating
4 -7
E= 5 GPa
'D
C
polymer substrate
5 ILm
polymer substrate
crack formation by indentation
polymer substrate
polymer substrate
crack formation by CTE mismatch
based stresses
Figure 5. General behaviour of low and high
moduli coatings on plastics.
Other mechanical properties like bending strength
depend, as expected, very much on the degree of organic
crosslinking. In table I, the results of 3-point bending
experiments are given.
Table I. 3-point bending experiments
on various polymers.
Type
Curing
B(NPa)
A SiO 2 /Si-epoxy/
TiO 2
thermal
(130 "C)
B SiO 2 /Si-eoxy
Si methacryl/T10 2*
thermal (130 'C)
+ photo
4- 5
C SiO /Si-epoxy/Si-
thermal (130 'C)
8 - 9
methacryl/T1O 2
MKA*
D SiO /(C6 H ) SiO/
CH3fCH 2'CH )iO**
3 ZrO2 (methacrylic
acid (MA)
+ photo
thermal (200 'C)
+ photo + thermal
(130 "C)
photo + thermal
(100 - 150 "C)
Si-methacryl***
accordlnq to [131
accordlnq to (21
coslazation ofIr[o-prop)a vith mthacrylic acid (1:1) an sumqunt
bydrolysis aad ondeatlon by pbotocurinq VitIrcure 194(Ciba Giqy)
(191
*
*0*
3
30
80 - 90
i0
In composition A the epoxy is reacted to a glycol group that
means no organic crosslinking takes place. The materials are
brittle, they are cured in polypropylene moulds, the surface
quality of which is defining the number and size of surface
flaws (influence of flaws on strength has not been investigated). The number of organic crosslinking units increases from A
to E, but all materials are brittle with negligible plastic
deformation. According to this data, a clear connection between
strength and organic crosslinking exists. The first results
obtained from methacrylic acid condensates show bending strength
values, already suitable for structural applications, even if
the systems are not optimized so far.
Another example for property changes is the building up of
inorganic networks within already formed organic polymers, for
example ethylvinyl acetate copolymer (EVA) according to [10]
where the extremely low EVA modulus increases to about 5 GPa by
addition of 50 wt.% Ti(OEt) to the polymer solution in toluene.
Depending on the route of aadition of water, phase separated
(figure 6a; H20 is added rapidly via atmosphere) or non-phase
separated polymers (figure 6b, H20 is added via a controlled
chemical condensation process [16]; phase separation not visible
within the resolution of the instrument
2 nm).
Figure 6a and 6b. TEM of a TiO 2 /EVA copolymer.
6a: phase separation about 100 nm; 6b: no phase
separation or separation 5 2 nm; bar: 100 nm.
Functional properties
Despite the fact that the investigation of structural properties
of ORMOCERs (except abrasion resistance) is at its beginning,
the question arises whether they can be combined with functional
properties. Developments of special functional properties have
already been published elsewhere. Table II gives a survey over
various developments.
11
AS one can see from these examples, the introduction of functions into ORMOCERs can lead to interesting properties and
potentials for applications.
Table II. Development of functional ORMOCERs
and corresponding organofunctional groups.
Property
Application
Chemical Function
hydrophobicity
contact lenses
epoxy to glycol
reactive
surfaces
carrier for enzymes vNH
radioimmunoassay
adsorption
of CO 2
heat pump
SO 2
adsorption
S02 gas sensor
mild abrasion
acne ointment
2
,
-CH 3 ,
/tvCOOH
^ANH 2
Reference]
[18]
[20]
[5]
[7,
'AtNR2
(CH 3 )2 SiO/SiO 2
O "
21]
[23]
adhesion to
glass surfaces
protective coatings on medieval
glasses; sealing
agent
(CH ) 2 Si'
nSiO
OH
NIR absorption
solar protection
V4 +
H+ conductivity
solid state
electrolytes
/AANH
3 +X
low e, high
resistance
dielectric
materials
SiO , hydrocarons
[29]
low hydrocarbon diffusion
gas tanks
LSi- 8 :
-AY
(29]
micro patterns by photolithography
SMT technologies,
laser writing,
resists
wSi-methacryl
oSi-vinyl
aSi-epoxy
[30]
fluorescent
coatings
solar collectors
fluorescent organic
dyes
[31]
thermoplasticity
hot melts
(C6 H5 )2 SiO
[32]
refractive
index
optical
coatings
Ti, Zr, aryl
[33]
[24-25]
[26]
-
[27-28]
12
CONCLUSIONS
The combination of functional with structural properties in most
cases can be achieved and functions can be transferred to hard
coatings. Of special interest is the question, how far these
materials are able to play an important role for key technologies. Their ability of being patterned is important for microelectronics or microsystems, especially in combination with
other properties like sensor properties, low or high dielectric
constants, low H 0 diffusion or temperature stability. Therefore, the knowledge of structure to property relations has to be
improved. The high homogeneity shows an interesting potential
for optical applications, especially in combination with other
functions like dye incorporation. The properties of nanocomposites based on ORMOCERs are hardly investigated up to now despite the possibility to tailor such composites by controlled
phase separation.
REFERENCES
[1]
H. Schmidt, H. Scholze, G. Tunker: J. Non-Cryst. Solids 80
(1986) 557.
(2]
H. Schmidt, G. Philipp, H. Patzelt, H. Scholze: Collected
Papers, XIV. Intl. Congress on Glass, Vol. II, 1986, 429.
[3]
W. Noll: Chemie und Technologie der Silicone. 2. Aufl.,
Weinheim. Verlag Chemie 1968.
(4]
B. G. Bagley, W. E. Quinn, P. Barboux, S. A. Khan, J. M.
Tarascon: In: Proceedings Fifth International Workshop on
Glasses and Ceramics from Gels. August 1989, Rio/Brazil.
J. Non-Cryst. Solids (in print).
[5]
H. Schmidt, J. Strutz, H. G. Gerritsen, H. Mdhlmann: DP
35 18 738, 24.05.85.
[6]
H. Schmidt, 0. von Stetten, G. Kellermann, H. Patzelt, W.
Naegele:
IAEA-SM-259/67, Wien 1982, 111.
[7]
F. Hutter, K. H. Haas, H. Schmidt: In: Proceedings of the
Second International Meeting on Chemical Sensors. Eds.:
J.-L. Aucouturier et al., Bordeaux, July 1986, 443.
(8]
J. E. Mark: In: Ultrastructure Processing of Advanced Ceramics. Eds.: J. D. Mackenzie, D. R. Ulrich. John Wiley &
Sons, New York, 1988, 623.
(9]
G. L. Wilkes, B. Orler, H. H. Huang, Polym. Prepr. 26 (1985)
300.
(10] H. Schmidt: ACS Symposium Series No. 360 (1988) 333.
[11] H. Schmidt: J. Non-Cryst. Solids 112 (1989) 419.
[12] H. Scholze, P. Strehlow: Wise. Ztschr. Friedrich-SchillerUniv. Jena, Naturwiss. R. 36 (1987) 753.
13
(13] Y. Charbouillot: Ph.D. thesis. Institut National Polytechnique, Grenoble, 1987.
[14] J. Livage, M. Henry:
In: Ultrastructure Processing of Ad-
vanced Ceramics. Eds.: J. D. Mackenzie, D. R. Ulrich. John
Wiley & Sons, New York, 1988, 183.
[15) H. Schmidt, B. Seiferling: Mat. Res. Soc. Symp. Proc. 73
(1986) 739.
(16] H. Schmidt, B. Seiferling, G. Philipp, K. Deichmann: In:
Ultrastructure Processing of Advanced Ceramics; Eds.: J. D.
Mackenzie, D. R. Ulrich. John Wiley & Sons, New York 1988,
651.
(17] H. Schmidt, H. Wolter: In: Proceedings Fifth International
Workshop on Glasses and Ceramics from Gels, August 1989,
Rio/Brasilien. J. Non-Cryst. Solids (in print).
[18] G. Philipp, H. Schmidt: J. Non-Cryst. Solids 63 (1984) 283.
[19) E. Arpac, R. NaS, H. Schmidt: to be published separately.
[20] A. Kaiser, H. Schmidt, H. Hasenfratz-Schreier, K. D. Kulbe:
Chem.-Ing.-Tech. 56 (1984) 653.
[21) K. H. Haas, F. Hutter, H. Schmidt: In: Proceedings of the
International Conference on Materials with Exceptional Properties, EXPERMAT 87. Bordeaux, 1987.
(22] H. Schmidt, A. Kaiser, H. Patzelt, H. Scholze: J. Phys. 43
(1982), Coll. C9, Suppl. 12, 275.
[23] G. Trnker, H. Patzelt, H. Schmidt, H. Scholze: Glastech.
Ber. 59 (1986) 272.
[24] F. Hutter, H. Schmidt, H. Scholze: J. Non-Cryst. Solids 82
(1986) 373.
[25] D. Ravaine, A. Seminel, Y. Charbouillot, M. Vincens: J.
Non-Cryst. Solids 82 (1986) 210.
(26] F. Rousseau, M. Popall, H. Schmidt, C. Poinsignon, M. Armand: In: Proceedings Second International Symposium on
Polymer Electrolytes. Juni 1989, Siena/Italien. Elsevier
Applied Science Publishers Ltd., UK (in print).
[27] H. Schmidt, M. Popall, J. Schulz: In: Proceedings of the
Second Intcrnational Symposium on New Glass, Tokyo, November
1989 (in print).
[28] K.-H. Haas, H. Schmidt, H. Roggendorf: In: Proceedings of
the Topical Meeting on Glasses for Optoelectronics, Tokyo,
December 1989 (in print).
[29] B. Lintner, N. Arfaten, H. Dislich, H. Schmidt, G. Philipp,
B. Seiferling: J. Non-Cryst. Solids 100 (1988) 378.
15
THE SYNTHESIS, STRUCTURE AND PROPERTY BEHAVIOR OF INORGANIC-ORGANIC
HYBRID NETWORK MATERIALS PREPARED BY THE SOL GEL PROCESS
G.L. WILKES, A.B. BRENNAN, HAO-HSIN HUANG, DAVID RODRIGUES AND BING WANG
Dept. of Chemical Engineering. Polymer Materials & Interfaces Laboratory,
Virginia Polytechnic Institute & State University, Blacksburg, VA 24061
ABSTRACT
The synthesis, structure/property behavior of inorganic-organic
hybrid network materials prepared by the sol gel process have been
chronologically reviewed with emphasis on those systems prepared in the
authors laboratory. Specific features of reactions as well as the nature
of reactants are included. The morphological features of these "ceramer"
systems formed have many features in common even though the reactants may
be quite different. The mechanical properties of the final materials in
conjunction with SAXS have proven beneficial in establishing the basics of
their morphological texture. Finally, it is demonstrated how microwave
radiation in some cases, can serve as an efficient way in processing the
ceramer systems.
INTRODUCTION
In 1985 work was initiated within our laboratories with the objective
being to develop novel organic-inorganic hybrid network materials by
reacting metal alkoxides with functionalized polymeric/oligomeric species
by the sol gel process [1]. This paper addresses this topic with emphasis
on work from our laboratory but also will touch upon related work of
others.
Since the authors will principally address their research,
omissions to the other workers has not been by design but is due to space
limitations.
The classic sol gel reaction generally involves both the hydrolysis
and condensation of metal alkoxides. Generally, these two reactions occur
simultaneously once the hydrolysis step has been initiated. The reaction
rates of each are highly dependent upon such variables as the temperature,
pressure, local pH, nature of the metal alkoxide and the environment in
which the reaction takes place (e.g. concentration of water, alcohols and
other solvent media components utilized).
Of course, the condensation
step generally lead to loss of a water molecule while each hydrolysis step
leads to the production of an alcohol component - both of these features
contribute to the high shrinkage that occurs during the sol-gel process.
However, there is still much more information needed to understand the
classical sol gel reactions. Our goal, however, has not been to focus on
these features as much as to utilize the sol gel reactions of metal
alkoxides as a means of developing new materials by the incorporation of
appropriately functionalized orqanic olioomeric/oolvmeric components.
Restated, the goal has been to incorporate both inorganic and organic
based materials into a single network system that may possess properties
that are either unique and/or somewhat intermediate between those of the
pure organic and pure inorganic networks. They might also allow control
of a wide range of refractive index values and at the same time the
material may still retain flexibility. Such flexibility would be in
contrast to the more brittle nature common to the alkoxide networks.
These hybrids might allow new "window" materials to be developed for
specific regions of the electromagnetic spectrum.
Furthermore, their
mechanical and dielectric behavior may suggest their potential for
coatings, encapsulants, etc. for commercial applications. Indeed, within
this paper, controlling the optical property of refractive index will be
Mat. Reas.
Soc. Symp. Proc. Vol. 171.' 1990 Materials
Research Society
16
illustrated as a means of showing the novelty of our hybrid network
materials.
The majority of our efforts in this field has emphasized the
oligomeric/polymeric species as the dominant component within the final
network material. Hence, the materials tend to be flexible. We have
termed such materials "ceramers" to imply some combination of ceramic like
behavior with that of polymeric behavior. However, it is recognized the
firing of such materials would lead to the destruction of the organic
component and thus destroy the original network. Yet, high temperature
treatment of these materials could potentially lead to novel materials in
terms of controlled porosity. These materials might be suitable for
chromatographic supports, catalyst supports, etc. However, our goal has
been not to utilize high temperature treatments of the final system but
rather to investigate the initial hybrid network.
Other researchers, in particular Schmidt, have reported studies on
related materials. Schmidt and coworkers have developed hybrid systems
which they termed "ormosils" for organically Modified silicates [2].
Their studies have emphasized the use of tetraethylorthosilicate (TEOS)
and related silicon based alkoxides. The ormosils, in general, arc based
on the incorporation of much lower molecular weight, functionalized
organics for purposes of modifying the more alkoxide dominated systems.
This is in contrast with the work that will be reviewed principally in
this paper. Mark, et al. have also developed hybrid inorganic/organic
systems, but their approach in general has been the base catalyzed
formation of small particulate "in situ" inorganic fillers within a
polydimethylsiloxane network dominated matrix [3].
Some Structural Features of the Ceramer Systems
The initial ceramers that we prepared utilized silanol terminated
polydimethylsiloxane (POMS) oligomers which were of functionality two.
They were incorporated into a silicate network by condensation of their
terminal groups with the silicic acid species from hydrolyzed TEOS. These
reactions were acid catalyzed to promote more chainlike inorganic
structures as reported by Schaefer, et al. [4]. The effect of acid level,
acid type, water content and oligomer molecular weight were investigated
for their effect on the morphology and general structure property
behavior. As reported elsewhere [1,5,6], the final materials in multi-mil
film form (i.e. typically 5 to 20 mils thick) showed flexibility and
optical transparency.
However, brittleness occurred if only a small
amount of the PDMS component was utilized. While these materials did
demonstrate our ability to prepare hybrid networks, they did not display
high tensile strength or elongation which was attributed to some loose or
dangling chain ends. Dangling chain ends, which may result from lack of
incorporation of the silanol endcapped PDMS, would promote network
defects.
These network defects and the generally poorer mechanical
properties of PDMS would contribute to lower tensile strength. In order
to promote better mechanical properties in terms of strain to break and
tensile strength, we utilized hydroxyl terminated poly(tetramethylene
oxide) (PTMO), another flexible, functionalized oligomer. The PTMO was
further functionalized by reacting the terminal hydroxyl group with
isocyanatopropyltriethoxysilane which leads to a reactive oligomer of
functionality six. The molecular weight of these oligomers was varied
from 650 to 2900 g/mol and reacted with TEOS to produce transparent
flexible materials. The moduli (stiffness) of these materials was found
to be dependent on the oligomer molecular weight as well as the
composition ratio of alkoxide to PTMO. The high transparency, along with
recognition of the large difference in refractive index between the two
components clearly illustrated that no major macro phase separation
17
occurred. However, on a very localized scale, i.e.,
a., 10 nm, a
variation in composition was developed as noted by small angle x-ray
scattering (SAXS) [7,8,9,10]. The general morphological model that was
developed based on this work is schematically given in Fig. 1. The reader
should recognize that a variation in the level of ohase mixing can occur
H%,
dSi-
Fig. 1.
1
Schematic showing general morphology of a ceramer on a molecular
level - see text for details.
depending on the relative compatibility of the reacting components as well
as the specifics of the reaction conditions. The basis for the model in
Fig. 1 extends from the observation that the SAXS behavior typically
illustrated a single shoulder or peak with an estimated correlation
distance of ca.. 10 nm. However, this distance has varied somewhat
depending upon oligomer molecular weight, composition ratios, etc.
[7,8,9,11,12,13,14,18].
Furthermore, the high angular tail of the
scattering curve also generally suggested that some level of mixing of the
inorganic occurs within the oligomeric phase thereby broadening the
scattering profile. This same conclusion regarding partial phase mixing
was also based on the many dynamic mechanical spectra (DMS) that have been
gathered on the PTMO/TEOS hybrids as well as other ceramers that will be
discussed.
In these earlier studies, we noted that the level of reaction may be
influenced by diffusion control following vitrification which limits
mechanical properties. However, if the gels that had been cast under
ambient conditions were heated to a higher temperature, the extent of
reaction was increased. In some cases, this post gelation treatment put
elongation to break through a maximum while tensile strength tended to be
enhanced (19,201. Depending upon the general glass transition behavior of
the functionalized oligomeric species, the importance of diffusion control
reactivity was also more or less enhanced as would be anticipated. This
suggested that higher temperature glass transition oligomers would need to
be processed into ceramer networks above their respective glass transition
values accordingly. A good example of this latter situation has been
demonstrated by our studies of triethoxysilane capped poly(ether ketone)
18
(PEK) oligomers of ca. 3500 grams per mole [18]. These species were
compatible within a reacting mixture of TEOS up to reasonably high levels,
but gelation and vitrification occurred at low extents of reaction due to
the higher glass transition behavior of the PEK component (Tg ca. 180°C).
Thus, the final network solids needed to be post-cured above 200°C to make
strong transparent films.
Following the studies just described, our efforts began to focus on
the incorporation of other metal alkoxides in conjunction with TEOS.
Titanium (IV) tetraisopropoxide (Ti(OPr)) was a principal alkoxide since
this species promotes a higher refractive index network. However, due to
the much higher hydrolysis and condensation rates of the Ti(OPr) one
cannot add water directly to the reaction system. Thus, a new reaction
scheme was established by modifying Schmidt's procedure denoted as
controlled chemical condensation [21]. Our modification of this procedure
has been discussed elsewhere [8,20].
It was successful in promoting
transparent multi-mil films of the two mixed metal alkoxides with a
variety
of
functionalized
PTMO
or
triethoxysilane
capped
poly(dimethylsiloxane) oligomers. In all cases, an acidic environment was
utilized for purposes of minimizing any particulate growth of the
inorganic component.
It was also noted that any time the titanium
reactant was used, the modulus of a given system was higher following
network formation. This suggested that this reactant served to catalyze
the reaction to a higher level of conversion. Thus the final product
developed a higher network density and higher modulus accordingly as
denoted by DMS [8,20]. Further proof of this speculation came from the
fact that post curing caused less change in properties again indicating
that a higher extent of reaction had occurred.
The morphological features of these mixed alkoxide systems also
seemed to be rather similar to those discussed earlier. Their SAXS
behavior suggested that the general schematic model given previously in
Fig. 1 was still valid as a simplified description of these systems.
However, the inorganic regions would now likely contain both alkoxide
moieties.
Recent Ceramer Systems and Their Structure-Property Features
Turning to some of our more recent studies (hence more details will
be provided) we have been successful in directly incorporating aluminum
alkoxides or titanium alkoxide with functionalized oligomers but without
any TEOS. Our first success in this endeavor was to utilize a bidentate
chelating agent - that of ethyl acetoacetate (EACAC). EACAC promoted a
stable sol when reacted with the alkoxides of aluminum secondary butoxide
or Ti(OPr). This sol could be reacted later with additional alkoxide and
functionalized oligomer leading to transparent multi-mil films of these
EACAC modified ceramers. One disadvantage found was noted specifically
for the alumina based materials. They illustrated that on the order of 15
weight percent could be extracted from the final network. The extractable
by-products were speculated to be based on a cyclic species from the self
condensation of the EACAC component [23]. Yet, the Ti(OPr)/EACAC based
ceramers showed only a very small percentage (1-2 wt. %) of extractables.
SAXS analysis of the latter ceramer systems again illustrated that a
single correlation distance existed on the order of 10 nm. Therefore, the
morphological features were somewhat similar to what was illustrated
earlier in Fig. 1. OMS data and stress strain response provided general
support of this same model. [Later within this paper, another newer route
for preparing the more highly reacted metal alkoxides into ceramer systems
but without the chelating agent will be discussed.
On a somewhat different subject, we have recently demonstrated that a
polymeric catalyst can be effectively utilized to promote reactivity in
19
the ceramer systems. This approach is in contrast to the general classic
sol gel reactions that utilize low molecular weight acidic or basic
catalysts such as hydrochloric acid (HCl) or an organic amine respectively
[24]. In fact, our newer approach should also function for the classic
sol gel reactions where no additional oligomer is added. Basically, our
goal was to illustrate that the use of a higher molecular weight polymeric
catalyst, (e.g., polystyrene sulfonic acid (PSS)) would not greatly change
the final materials compared to those prepared by a lower molecular weight
catalyst as HCl. Also, with the polymeric species present, the viscosity
would be enhanced more quickly which therefore would influence the
rheological behavior of this system - a feature of importance in coating
or spinning operations. Indeed, shown in Fig. 2 are the DMS obtained on a
TEOS-PTMO system where the initial weight fraction of the TEOS component
and the 2000 g/mol molecular weight functionalized PTMO were each 50
weight percent [11,15]. One hundred percent of the stoichiometric amount
of water required for the hydrolysis of the alkoxide groups was utilized.
A constant equivalence of acid of both the PSS as well as the HCl was
10
.0.6
*
*
PS6 Cotolyzftd
M'CI Cotalyxftd
.0.5
0.4
000
E
07
J
0.1
0.0
E
-150
-100
-50
0
50
100
150
TEMPERATURE (C)
Fig. 2.
Storage modulus and tan s behavior (at a frequency of 11 Hz)
illustrating a higher extent of reaction in TEOS/PTMO ceramers
catalyzed with PSS vs. HCl.
maintained at 0.0143 eq. The PSS polymer was of 70,000 g/mol molecular
weight and was fully functionalized into the acid form as was determined
by end point titration. The resulting DMS of the ceramers in Fig. 2
indicated that there is some difference between the two catalyzed
materials. Above Tg, the modulus of the PSS system is higher relative to
the HCl catalyzed system. Certainly this suggests a higher level of
network formation. In addition, the tan 6 response is of a bimodal nature
for both materials with nearly equivalent breadth of the combined
dispersions.
As discussed elsewhere [15], the higher modulus in
conjunction with the somewhat lower tan 6 response for the PSS catalyzed
system has been attributed to the "tightening" of the network promoted
20
from the condensation of the TEOS that is partially mixed within the PTMO
phase. This restricts the PTMO mobility and decreases the tan 6 response
accordingly. However, since the thermal breadth of the dispersion for
both samples is nearly the same, i.e., -75"C to +1000C, the general
features of the morphological character were speculated to be the same
although the extent of curing is different. Further proof for this
extended from Fig. 3 where the relative SAXS intensity obtained is
provided and plotted against the scattering variable s where s is defined
as (2/1)sine with X being the wavelength utilized (1.54A) and e
IN.
ll-pI-C
4.0
U4.0
000
.000
.120
.240
a
Fig. 3.
AN
l
(i/nm)
SAXS profiles demonstrating a higher scattering contrast factor
for a TEOS/PTMO ceramer catalyzed with PSS versus HCl.
representing one half of the radial scattering angle. One clearly notes
from Fig. 3 that both the PSS and HC catalyzed materials display a
scattering peak with the estimated correlation distance between regions of
comparable electron density to be about 10 nm. (This latter result and
those prior to it are based on smeared SAXS scans. Desmearing procedures
however, lead to rather comparable results and therefore the general
correlation distance mentioned before is still approximately correct.)
One notes from these SAXS scans that the PSS material shows a higher
intensity at smaller s relative to that of the HCI catalyzed system yet
the high angular tail regions of each are nearly the same. Indeed, this
was expected since the PSS material was believed to have a higher extent
of reaction (recall Fig. 3). This in turn would also lead to a greater
level of condensation of the metal alkoxide which would raise its electron
density relative to PTMO. Thus, the scattered intensity would increase
accordingly (15]. The overall results therefore suggested that the same
morphological features exist and were similar for the two materials but
that the level of network density was higher in the PSS catalyzed system.
Certainly it must be recognized that the greater modulus in the PSS
catalyzed material was not a consequence of the polymeric nature of the
catalyst but more likely due to the difference in the reactivity features
promoted by the polymeric sulfonic acid moiety in contrast to HCI.
Further work
is being
undertaken
utilizing
para
toluene
21
sulfonic acid at an equivalent acid content to investigate this point.
These results will be reported in the near future.
As indicated earlier, one of our goals has been to develop higher
refractive index materials by the incorporation of the high refractive
index metal alkoxide Ti(OPr) but to do so without the use of a chelating
agent. To achieve this goal, a reaction scheme was developed to partially
react Ti(OPr) into a stable sol that could later be reacted with
functionalized PTMO components. A successful procedure for doing this is
shown in Scheme 1 [22].
This recipe incorporates a 2000 g/mol
functionalized PTMO using the reaction conditions of ambient temperature
PREPARATION OF HYBRID WITH Ti(OPr)
4
5gm Ti(OPr)
SLOW ADDITION OF
0.1 ml ION HCI
0.6 ml WATER
20 ml IPA
5omPTMO(2K)
+ 10ml THF
MIX FOR 5 TO 10 MIN.
CAST INTO MOLD, COVER
J
GEL 1 DAY, AMBIENT
I
DRY 2 to 3 DAYS
ANALYZE
Scheme 1.
Methodology for preparation of Ti(OPr)/PTMO ceramers.
ir a THF/HCI water isopropanol (IPA) media. The result of casting these
materials into multi-mil films usually led to optically transparent
materials. Sometimes a slight yellow color developed which is believed to
be caused by the difference in oxidation states for titanium. However, the
high transparency clearly indicated good dispersion of the two components.
The OMS data for two such compositions are shown in Fig. 4 where the
weight percentages of the Ti(OPr) are 49 and 58 percent respectively. One
notes that upon raising the temperature of these materials to over 150°C,
there was little sign of post curing.
Again this suggested Ti(OPr)
promoted a higher extent of reaction. Also noted from these data was that
the Tg dispersion of the PTMO was somewhat different from the bimodal
character displayed by the TEOS ceramers. In particular, while the -f
transition (-125°C) was common to all materials resulting from the
rotational features of the methylene groups in the PTMO chain [24], Tg of
the PTMO component was clearly more narrowed than before. This transition
was now initiated at approximately -80"C and passed through a minimum just
above 0°C with a smaller dispersion above at higher temperatures. Also,
the general temperature range over which the modulus decays to a lower
level was narrower for these materials. Both of these facts together
suggested less phase mixing between the two components such that the PTMG
oligomer now displayed more of a Tg as would be expected of its own pure
phase [19]. This is not to say that no mixing occurred for indeed if the
22
S3
0. 0
0 7
07 00
-150
-100
-50
0
50
100
150
TEMPERATURE (C)
Fig. 4. Storage modulus and tan 8 behavior (at a frequency of 11 Hz) for
two selected Ti(OPr)/PTMO ceramers at 49 and 58 wt.% fraction of
Ti(OPr).
PTMO were totally unrestricted and in pure form, one would anticipate the
tan 8 response to go to a minimum somewhere around -30°C at the frequency
which the DMS spectrum were determined (11 Hz).
Corresponding SAXS
profiles of these same materials along with two additional ones are shown
in Fig. 5. A single maximum was measured again whose intensity was in
proportion to the amount of the Ti(OPr) species. Also, the correlation
distance was the order of 10 nm. It was somewhat larger than the ceramers
discussed earlier that utilized identical functionalized PTMO. We have
speculated that this slight increase was due to less mixing of the
titanium based species with the PTMO thereby building slightly larger
inorganic domains and hence slightly increasing the correlation distance.
Furthermore, analysis of the high tail region of the SAXS curves given in
Fig. 5 showed there was a sharper interface in contrast to the previous
samples. Again this gave support to less phase mixing as was suggested
from the DMS results.
The general stress-strain response of these systems as determined at
ambient at an initial elongation rate of 2 mm/min is given in Table 1.
This table provides the modulus, stress at break as well as the elongation
at break for ceramers of four different Ti(OPr)/PTMO contents where again
the PTMO oligomer was of 2000 g/mol. As expected, the modulus and tensile
strength, at ambient, increased as the content of the inorganic species
was increased. The corresponding decrease in the elongation at break was
due to the lower content of the elastomeric PTMO oligomer. While the data
will not be shown here, we have also successfully incorporated zirconium
tetraisopropoxide in a similar manner to that of the titanium alkoxide.
DMS and stress strain results in conjunction with the SAXS behavior
suggested that these latter systems were also quite comparable to those
just discussed. Thus, there was a higher level of phase separation in
these systems in contrast to the earlier TEOS/PTMO or TEOS/Ti(OPr)/PTMO
comixed ceramers.
23
500.
-54S
400.
,66
1- 300.
F.-
Z
-200.
100.
.000
.000
.120
.240
.310
.
.0
Fig. 5. SAXS profiles indicating an increasing scattering intensity with
increasing weight fraction of Ti((OPr) (as shown in the figure) in
selected Ti(OPr)/PTMO ceramers.
Table 1) Elastic modulus, stress at break and elongation at break for
increasing weight percent (Ti(OPr) loadings in the Ti-PTMO ceramers.
Wt. Percent
[Ti(OPr) 4]
Elastic
Modulus (MPa)
Stress at
Break (MPa)
Elongation at
at Break (%)
XX=
XX=
101
185
244
411
19
20
23
34
58
51
25
10
49%
54%
58Y.
66%
As indicated above, we have been successful in directly incorporating
the titanium alkoxide with functionalized PTMO into transparent monolithic
systems. The materials not only displayed reasonable mechanical behavior
as presented above but also provided control of optical properties as
illustrated in Fig. 6. Here, the refractive index (sodium line) has been
plotted against the weight percent of what is denoted as TiO 2 of the
titania-PTMO material - see curve A. In this figure, the weight percent
of TiO 2 was calculated using the assumption that the titanium alkoxide
undergoes complete hydrolysis and condensation and therefore leads to
T10 2 . This assumption is not likely perfect but appears to be reasonable.
The extrapolated value of the refractive index at 100% Ti0 2 achieves a
value just below 1.8 as would be anticipated for the more amorphous form
of T10 2 [26]. A linear rise in the refractive index occurred as the
titania content increased. In fact, the initial value of about 1.45 for
pure PTMO can be raised substantially to 1.65 at about 60 weight percent
of the calculated T10 2 component. This is a significant increase of
refractive index in a flexible polymer based system which shows a very
strong dependence of refractive index on this particular alkoxide. No
such large rise in refractive index occurred with the zirconium based
alkoxide due to its lower refractive index properties (data not shown
4
24
I..
o
1 5
-
....
1.70*
-
I.0.1
...
..........
-
Q......
. ......
*1
1A0
45 I..
0
-
&
1
M
3D
40
M
70
OD
WtZ of TiO
80
0
I..
2
Fig. 6. Linear response of the refractive index as a function of
calculated wt. % Ti0 2 for selected ceramers. A) Ti(OPr)/PTMO and
B) Ti(OPr)/PSF.
here). Clearly these results are very positive in terms of controlling
this optical property. However, it would be of interest to start with a
higher refractive index oligomer for purposes of achieving even higher
refractive index materials that might serve as optical coatings, etc. In
an attempt to achieve this goal, functionalized polyethersulfone oligomers
were prepared in a similar manner to the other triethoxysilane capped
systems. In this case, amine capped oligomers of the poly(ethersulfone)
(PSF) were reacted with isocyanatopropyltriethoxy silane. Thus a urea
linkage was formed in contrast to the urethane linkages promoted by the
end capping reactions of the hydroxyl terminated PTMO discussed earlier.
We then developed a reaction scheme between the Ti(OPr) sol and that of
the end capped PSF species using a procedure similar to Scheme I except
for a post curing step at 200°C. This latter step was necessary in order
to extend the reaction to a higher network density that could not be
achieved at room temperature due to limitation by vitrification caused by
the high Tg oligomeric backbone (Tg ca. 200oC). Transparent amber colored
films were prepared using this procedure that maintained good flexibility
up to reasonably high levels of titanium. At higher levels, the materials
became more brittle as expected.
An indication of the general DMS obtained on one of these materials
is shown in Fig. 7. This material which contained 25 weight percent of
the Ti(OPr) component showed the typical glassy modulus expected up to the
range of 200°C where the Tg for the PSF component was initiated. The fact
that this dispersion was quite narrow and in the temperature region
expected for a pure PSF phase, clearly indicated that good phase occurred.
Indeed this was expected due to the fact that the titanium component was
utilized. However, enhanced phase separation likely occurred because the
oligomeric molecular weight was of the order of over 7000 g/mol greater
incompatibility with the alkoxide might well be anticipated. The SAXS
data obtained for this same material as well as materials of higher
.4
I
25
10
2.0
TiOPr (25) -- PSF (7. 2K)
VT200C
0
Tan d
1
8
•
i
A
1.6
7
0A
J0.4
a
a
25
75
0
10.04
175
TEMPERATURE
(C)
Fig. 7. Storage modulus and tan 6 behavior (at a frequency of 11 Hz) for
a selected Ti(OPr)/PTMO ceramer.
titanium content (data not shown here) clearly illustrated a sharper decay
in scattering intensity with angle. The tail region was flatter which
indicated that the phase separation was sharper. Of particular interest
is to recall Fig. 6 where the refractive index for this Ti(OPr)/PSF series
of materials is shown in curve B but now the line initiates at the
refractive index value of the pure PSF component which is about 1.6.
Indeed, PSF is one of the higher refractive index polymers known and hence
one of the reasons that we selected it as the basis for developing this
ceramer material. It is also of interest to note that the extrapolated
value to 100% TiO 2 was in agreement with that obtained from the PTMO based
systems. This suggested that the form of the TiO 2 domains were of an
equivalent nature.
In the last stage of this paper some time is spent pointing out a
potential way to process these systems more effectively than by the usual
ambient temperature casting procedure utilized in the classic sol gel
approach. As part of a much larger project in our laboratory, the use of
microwaves as a means of processing polymers has been under investigation.
For this reason we included our ceramer materials as one potential system
in this study. The details of the instrumentation for the microwave
cavity and its control in terms of power input, control, etc. are
presented elsewhere [271. Briefly, the basic instrument was a tunable,
cylindrical cavity equipped with a temperature monitoring fiber optics
probe in direct contact with the sample. The microwave frequency of 2.45
GHz was used. This allowed one to rapidly heat the material if its dipole
character responds to this microwave frequency.
Basically the power
absorbed scales with the second power of the applied voltage, the first
power of the frequency as well as the dissipation factor of the dielectric
constant (29].
While these laws are well known, the ability of the
microwave to potentially promote a higher rate of reaction due to specific
interactions with the dipoles of the appropriate functionalized reacting
groups may potentially lead to different reaction kinetics than would be
26
promoted at an equivalent bulk temperature of the system.
Indeed,
differences in reaction kinetics (but at the same temperature) have been
demonstrated in other network forming systems e.g., epoxies, as recently
reported [27]. We therefore anticipated that this same approach might be
effective in promoting the sol gel reaction and the development of the
ceramers in a much shorter time scale. One is limited initially in going
to high reaction temperatures due to the higher volatility of the alkoxide
and the solvent(s) in which the reaction is carried out. However, we found
that depending on the alkoxide and choice of reacting media, it was
possible to rapidly achieve gelation and almost final properties of the
ceramer systems in a matter of minutes (13]. This was in contrast to the
usual several days over which gelation occurs at ambient temperatures. As
an example, we have studied one microwave processed system where the
temperature of the reactants was ramped to 70*C in 400 seconds ,;dthr'n
held at that temperature for 12 min. This material, prepared in tha' 2
minute time period, was analyzed shortly thereafter.
Its general
stress-strain properties (modulus, stress at break, and elongation at
break) were compared with the same material that had been cast at room
temperature for six days. Also a third material, prepared by processing
in a simple conventional oven at 70°C for two hours, was included for
comparison. Table 2 shows a major difference in the results of these
Table 2) Stress strain analysis for a TEOS-PTMO system cured in a
microwave cavity (M.W.), in a conventional oven (O.C.) and at room
temperature (R.T.) [12].
Elastic
Modulus (MPA)
Stress at
Break (MPa)
250
40
21
153
29
29
4
1
50
M.W. 70°C
380
38
21
R.T. cure
380
40
21
O.C. 70*C
150
7
Elongation
at Break (%)
Fresh
M.W. 70°C
(20 min)
RT cure
(6 days)
O.C. 70*C
(2 hrs)
Aged (I Month)
7
three experiments. In particular it is noted that the sample that had
undergone the microwave treatment at 70*C (M.W. 70C) achie',ed a modulus
nearly equivalent to a glassy polymer with a significant sttess at break.
The corresponding oven cured sample (O.C. 70°C) that was heat treated for
two hours was a very soft material with a low stress at break and a much
higher elongation at break due to poor network formation.
The
conventional room temperature cured material that was also not completely
cured even at six days showed a slightly lower modulus and stress at break
than that of the microwave prepared material.
The properties of a
corresponding one month aged sample of each is shown in the lower part of
the table to illustrate that some additional level of curing did take
place even at room temperature following the microwave treatment. One can
see that the microwave treated material achieved a slightly higher modulus
and the room temperature cured material also achieved this same level
(given the extra time period). However, the oven cured 70*C sample never
27
In essence, then it was shown that
achieved the same properties.
microwave processing is effective in shortening the time period to reach
the solid state characteristics. Although, an additional post cure at a
higher temperature may be required for these systems to achieve their
ultimate properties. Investigations of this nature are being actively
pursued in our laboratory at this time.
Addressing the SAXS behavior of these three materials, it is of
interest to inspect the results given in Fig. 8. Here one notes
400.
-
......
M. W. 70C (20 W11)
ROOM TEMP 16 DAYS)
0. C. 70C ( HOUMS)
320.
240.
Cn
Z
w
/
150.
80.0
.000
.000
.120
.240
.360
.4
.600
S (i/nm)
Fig. 8. SAXS profiles for selected TEOS/PTMO system prepared by different
processing conditions - see text for details [13].
particularly sharp scattering peaks with a decrease to a flat tail region
over the same angular range utilized in the earlier plots. There were
some particular points that have been noted, the first being that the
level of intensity was highest for the microwave treated sample with the
room temperature material being second and the lowest level of peak
intensity being for the oven cured materials. This was exactly as
expected based on the level or extent of reaction being the least in the
oven cured and the highest for the microwave treated systems. In the
latter the electron density contrast would be higher as the extent of
reaction increased. Secondly, the flat tail region that extended from
using OMF as the liquid reacting media lead to a much higher level of
phase separation. The cause for this higher level of phase separation in
a TEOS/PTMO ceramer is unknown but we have further verified it by GMS in
conjunction with SAXS and TEM analysis. Further work is proceeding on the
influence of different reaction media and how it influences the microwave
In addition, we have prepared
versus conventional casting procedures.
many of the other ceramers more rapidly by microwave processing. There
were some differences between these systems and how they respond to
microwaves which will be the subjects of separate publications [29].
28
CONCLUSIONS
This paper has attempted to review the chronological development of
the ceramer materials as prepared under the direction of one of the
authors (GLW). Details regarding the earlier studies were presented in
brief form since numerous other publications provide many of the details.
More emphasis has been given to those studies that have led to the
incorporation of the more reactive metal alkoxides based on aluminum,
titanium and in more limited cases zirconium. Higher refractive index
flexible ceramers were made by incorporating titanium alkoxide into high
refractive index functionalized oligomers such as poly(ether sulfone) or
poly(ether ketone). Finally, it has been demonstrated that the use of
microwave processing of the sol gel ceramer hybrids can effectively
decrease the time to achieve final properties. However, it has been
shown that the choice of reaction media, as well as alkoxide and oligomer,
will influence the relative differences between microwave and conventional
processing.
Acknowledgement
The authors would like to acknowledge the financial support of the
Akzo Corporation, Johnson & Johnson Foundation, Eastman Kodak, DARPA, and
the Office of Naval Research.
References
1.
G. L. Wilkes, B. Orler and H. Huang, Polymer Prepr. 26(2), 300
(1985).
2. H. Schmidt, Kate. Res. Soc. Symp. Proc. 32, 327 (1984).
3. J. E. Mark, C. Y. Jiang and M. Y. Tang, Macromol., 17, 2613 (1984).
4. 0. W. Schaefer and K. D. Keefer, Mate. Res. Soc. Symp. Proc. 32, 1
(1984).
5. H. Huang, B. Orler and G. L. Wilkes, Polym. Bull. 14, 557 (fi85).
6. H. Huang, B. Orler and G. L. Wilkes, Macromolecules 2U(6), 1322
(1987).
7. H. Huang and G. L. Wilkes, Polym. Prep. 28(2), 244, (1987).
8. H. Huang, R. H. Glaser and G. L. Wilkes, ACS Symposium on "Inorganic
and Organometallic Polymers", 360, 354 (1987).
9. H. Huang and G. L. Wilkes, Polym. Bull. 18, 455-462 (1987).
10. H. Huang, R. H. Glaser, A. B. Brennan, 0. E. Rodrigues and G. L.
Wilkes, Symposium on Ultrastructure Processing of Materials Tucson,
1989 in press.
11. H. Huang and G. L. Wilkes, Polym. Prep. 30(2), 105-106 (1989).
12. B. Wang, H. H. Huang, A. B. Brennan and G. L. Wilkes, Polym. Prepr.
30(2), 146-147 (1989).
13. B. Wang, H. H. Huang, A. B. Brennan and G. L. Wilkes, Polym. Prepr.
30(2), 227-228 (1989).
14. B. Wang, A. B. Brennan, H. Huang and G. L. Wilkes, J. Appl. Polym.
Sci. submitted.
15. A. B. Brennan and G. L. Wilkes, In preparation.
16. M. Spinu, A. B. Brennan, G. L. Wilkes and J. E McGrath, Mat.
Res. Symp. Polym. Based Mol. Comps. (1989).
17. R. H. Glaser, G. L. Wilkes and C. E. Bronnimann, J. Non-Cryst.
Solids submitted.
18. John Lee W. Noell, G. L. Wilkes, D. K. Mohanty, J. E. McGrath,
J. Appl. Polym. Sc. submitted
19. H. Huang, Ph.D. Dissertation, Virginia Polytechnic Institute & State
University, 1987.
29
20.
R. H. Glaser, Ph.D. Dissertation, Virginia Polytechnic Institute &
State University, 1988.
21. H. Schmidt, H. Scholze, A. Kaiser, J. Non-Cryst. Solids 63, 1
(1984).
22. B. Wang, A. B. Brennan, H. Huang, D. Rodrigues and G. L. Wilkes,
manuscript in preparation.
23. B. Wang and G. L. Wilkes, manuscript in preparation.
24. K. D. Keefer, Mate. Res. Soc. Symp. Proc. 32, 15 (1984).
25. N. G. McCrum, B. E. Reed and G Williams, Anelastic and Dielectric
Effects in Polymer Solids, John Wiley S Sons, London (1967).
26. Handbook of Physics and Chemistry, CRC Press, 68, B140 (1987).
27. D. A. Lewis, T. C. Ward, J. D. Summers and J. E. McGrath, Polym.
Prepr. 29(1), 174 (1988).
28. A. C. Metaxas and R. J. Meredith, Industrial Microwave Heating,
Peter Peregrinus Ltd., London (1988).
29. D. Rodrigues and G. L. Wilkes, manuscript in preparation.
31
THE CATALYTIC SYNTHESIS OF INORGANIC POLYMERS FOR HIGH
TEMPERATURE APPLICATIONS AND AS CERAMIC PRECURSORS
Jeffrey A. Rahn, Richard M. Laine ° and Zhi-Fan Zhang
Contribution from the Department of Materials Science and Engineering
and the Polymeric Materials Program of the Washington Technology
Center, University of Washington, Seattle, WA 98195
ABSRACT:
Polysilsesquioxanes,-[RSi(O)1 .5 ]x-, exhibit many properties that are
potentially quite useful for industrial applications. These properties
include high temperature stability (- 6000C in 02); good adhesion and,
liquid crystal-like behavior for some derivatives. Moreover, [MeSi(O) 1 .5]x,
polymethylsilsesquioxane has been used successfully as a precursor for
the fabrication of carbon fiber/black glass' (SiO 2 /SiC/C) composites and
"black glass' fibers.
Current methods of preparation depend on hydrolysis of RSiCI3 or
RSi(OR) 3. Unfortunately, this approach leads to several products that are
difficult to purify because polysilsesquioxanes exhibit a great propensity
for forming gels. We describe here a simple catalytic approach to the
synthesis of polymethylsilsesquioxane copolymers of the type
-[MeRSiO].3[MeSi(O)i.r where R - H, OMe, OEt, OnPr and OnBu. The R - H
copolymer is produced by catalytic redistribution of -[MeHSiO]xoligomers using dimethyltitanocene, Cp2"iMe 2 as the catalyst precursor.
Following catalytic redistribution, the resulting copolymer,
-[MeHSiO]3[MeSi(O)I.5].7-, is reacted in situ with alcohols to produce
-[Me(R'O)SiO]j.[MeSi(O)s.5].7- (where R - Me, Et. nPr and nBu) which serve
as masked forms of the polymethylsilsesquioxane. These new copolymers
have been characterized by 1H, 13C and 29Si NMR TGA and DTA. The NMR
studies allow us to assign structures for the copolymer.
These new copolymers exhibit improved tractability. Their high
temperature properties are all quite similar; although, the MeO-, EtO- and
especially the nPrO- derivatives give much higher ceramic yields than
expected.
Met. Moo. Soc. Symp. Proc. Vol. 171.
1990 Materials Research Society
32
INTRODUCTION
Polysilsesquioxane polymers,1 -[RSi(O) 1.5] x-, represent a very poorly
exploited area of polysiloxane chemistry despite the fact that they exhibit
a variety of potentially useful properties including: high temperature
stability in air; 2 good adhesion to a wide variety of substrates3 and, in
some instances, liquid crystal-like behavior.1 Moreover, -[MeSi(O) 1 .5]x-.
polymethylsilsesquioxane has been used as a preceramic polymer for
fabrication of silicon carbide powders, 4 "black glass" (70% SiO2 /20%
SiC/1 0% C) composite matrices for carbon fibers5 and for the fabrication
of black glass fibers.6
The primary problems associated with using polysifsesquioxanes for
engineering applications are: (1) the lack of good, high yield synthetic
routes and, (2) the highly crosslinked nature of the polymers which limits
their tractability and ease of purification. Literature syntheses generally
rely on the hydrolysis of RSiCI 3 or RSi(OR') 3 :1
RSiCI 3 [RSi(OR') 3] + H20
ctalyst> HCI (R'OH) + [RSi(O)1.5], +
-[RSi(O)1.51x[RSi(OH)O]y- + -[RSi(O)1.5 ]x-
(1)
These reactions usually lead to the coincident formation of polyhedral
oligosilsesquioxanes, [RSi(O) 1 .s]x, where x - 8, 10, 12; polysilsesquioxanes
with partially condensed monomer units, -[RSi(O)I 5]x[RSi(OH)O]y-, and
polyhedral polysilsesquioxane itself. Because polysilsesquioxanes exhibit
a strong propensity to form intractable gels with organic solvents, there
are significant problems with purification which result in low yields.
Thus, this synthetic route is unattractive for the rapid, large-scale
preparations.
Harrod and coworkers have recently developed a novel titanium
catalyzed redistribution reaction, as illustrated in reaction (2), 7 that can
3MeHSi(OEt) 2 0.05 fole % Cg2.*Me2Q_> MeSiH 3 + 2MeSi(OEt) 3
(2)
also be used to prepare methylsilsesquioxane copolymers, reaction (3):2.8
-xMeHSiOe-
0+" moleeHa2TM~iO
1
]00C>)
0.33xMeSiH3 + "[MeHSiOjo,3[MeSi(O)1.5]o,7
(3)
33
The copolymer -[MeHSiO] 0.3[MeSi(O)1 S]Jo.7 forms as a gel if neat
oligomethylhydridosiloxane, Me 3SiO-[MeHSiOx-H (Mn - 2K D), is exposed
to the catalyst. Fortunately, gelation can be avoided if polymerization is
conducted in toluene with a greater than 5:1 toluene to -[MeHSiO]x- volume
ratio. The resulting polymer can be used to prepare coatings, fibers and
monolithic shapes.8 At lower volume ratios, gels form quite readily.
Although the 5:1 volume ratio solutions can be used for some
applications, the gel-like material that results on solvent removal limits
the copolymer's utility for some applications, e.g. as a matrix material for
composites. As such, we sought to modify the copolymer's physical
characteristics by taking advantage of the reactive Si-H bonds. Reactions
at these bonds should permit one to vary the side chains on the copolymer
backbone and thereby control some of its physical properties.
The long term objectives of the work described here are to prepare
tractable silsesquioxane copolymers that exhibit improved high
temperature performance and that are also useful for the fabrication of
polymer and ceramic, membranes and fibers. This report concerns
preliminary studies on the modification of the -[MeHSiO]x[MeSi(O)1.S1ycopolymer by alcoholysis of Si-H bonds.
RESULTS AND DISCUSSION
The copolymer produced in reaction (3) either neat or in toluene gives
the same results when characterized by magic angle spinning (MAS),
multinuclear NMR, solution NMR, diffuse reflectance infrared fourier
transform spectroscopy (DRIFTS), chemical analysis, TGA or DTA. These
extensive studies are described elsewhere.2. 7 For illustration purposes,
the TGA of -[MeHSiO] 0.3[MeSi(O) Ilo.7- is shown below (Figure 1). This
Figure also contains the elemental analyses (at selected temperatures) as
the polymer is heated in nitrogen at 50C/min to 9000 C.
At 4000C, the elemental analysis corresponds to pure -[MeSi(O)l.5]xwithout any of the starting monomer. This is confirmed by the MAS 29Si
NMR which shows a single peak at -65.7 8 relative to TMS.2 Consequently,
the 20% weight loss at temperatures below 4000C corresponds to the
depolymerization and volatilization of almost all of the -[MeHSiO]x- units.
34
Sttianq
oI.9omw
~
N (72 h)
-34.6
1 75OS.(CH
3 )3
1 04 SCH3
4.60 (3 43) MeHSPO
0.18(0 02) O$I(CM)3
0.10 (0.68) NSiC14
3
-33.5.,34 4.,35.9
-S72..5$.
1.60OSI(CH)
0.66 N5CH3
-3"2 08.0143
4.668() Si
0.18() 044C 3 )3
0.13 (00-R.
7 ) 145)013
0.10(Y)O05S13
49-000013
I1.54064(013)3
-3.2btvOSiCH3
-5.10br MSOSiCI3
4.70 (1.7) Sim
3.46(17)0054f3
0. 13 (49-K)001t64
0.10(X) M.0)C4 3
57.96OCt42
16.16CI43C*3
0.70 b ROSICS43
-3.25 br 061014
3.76(7
M oel4
1.20(10) 0142043
804
0.12
M)143
0042013
0(0142)20143
-36.7. 47.0.
46.3
64.0200145
25.550C1 2 C14
2
10.2701420543
i AS 06S(013)3
.3.12010143
-4.32 ROtION3
4.70(3) HSWH13
3.656(6.4) Wt14
1.55 (0)0054242
0.67 (6) C4C
0.12 (33-K)011053
0.10 (X)064053
0(0142)3CH3
-M6.5W -57.94.
-64.33 46.45
6210 OC52
34M50C4 3 C1 2
16.6 4C2C
1&.82C142C1
3
-&0 -p-lA 06403
366 (0.6) OCI4Q
I.s
10(a)
oc14sC)4
1.3
-W6) C01I3
OA (41) 0542H
0.12(0 CM
$43
ToweI. wM Specr lf +WS(0)wSIIitI(Oh.fr
uWn TUB w on Ww"e sgmider
spowa ww~ in COCs,
C 59&4%
0
11.1
la
aa
"A
U-US
750.
We6.
NG41
".0
a.s
1.,
W eS
N.0
RPMw 1.ThWMogrmvnsk Neyso oMjps6IO)j 7 .
as a FWUcio of TempWra.
The TGA heat Who" WaU
5'CfMi In
N2. OAiSaONye fr I*e and* e Wars huao Ina uiv manne.
0n
35
Copolymer Preparation and Characterization
Efforts to modify the -[MeHSiO]x[MeSi(O)1.5]y- copolymer began with
attempts to promote alcoholysis of the Si-H groups, reaction (4),9,10 in
situ, following completion ( 72 h) of reaction (3).
0
-[MeHSiO]o.3[MeSi(O) 1.5 ]o. + ROH 0.5 mole % Cp TiMeq/20 C >
H2 + -[Me(RO)SiO]o. 3 [MeSi(O)1.5]0.7r
The motivation for the alcoholysis experiments was to create a
polysilsesquioxane wherein some of the T groups, [MeSi(O) 1 .5], are masked
as the alkoxy derivative, [MeSi(O)OR]. In this way, the yield of T groups in
the polymer would increase significantly. Moreover, if R is a long chain
alkyl group it would also be possible to introduce more flexibility and
perhaps reduce or eliminate the elastomeric or gel character.
To our surprise, the addition of alcohols rapidly extinguishs the royal
blue Ti(lll) color of the original active catalyst system that forms in
reaction (3) leaving a yellow solution. If the alcohol is MeOH, then rapid,
almost violent H2 evolution ensues coincident with the color change. The
reaction can be somewhat exothermic depending on the initial catalyst
concentration.
Removal of the solvent and characterization by NMR indicates that
almost all of the Si-H bonds react with alcohol converting the remaining
silicons to T groups. NMR characterization, Table 1, confirms the 30:70
composition of the initial copolymer in that the integrated ratios of the
alkoxy groups to T groups in the product copolymer are nearly the same.
To date, we have made the derivatives R = Me, Et, Pr, nBu and
bis-1,4-(2-hydroxyethoxy)benzene (hydroquinone). These alcohols exhibit
reactivities with the copolymer strictly in accord with the size of the
alkyl group. The MeOH reaction is quite vigorous and is over in minutes to
hours while the nBuOH reaction requires two to three days. The
hydroquinone reaction results in extensive crosslinking that makes
further characterization impossible. The other copolymers are
moderately (MeO-) to completely (nBuO-) tractable following solvent
removal; however, it is expedient to redissolve the polymer in the
corresponding alcohol as these polymers still show a tendency to gel with
time (days to weeks).
The alkoxy copolymers have been characterized by 1H, 13C and in part
(4)
36
by 29Si as recorded in Table 1. The 13C shifts of the alkoxy carbon bound
directly to the oxygen are quite similar (except for MeO). Consequently,
characterization by 29Si was not deemed essential in all instances. The
proton and carbon spectra are all standard values for alcohols or alkoxy
substituents. However, the 29Si results are interesting because they
provide some understanding of the polymer backbone structure if we use
published 29Si peak assignments for standard siloxane monomer units. 11
The 29Si and 13C spectra for Me 3SiO[MeHSiO]xH, (Mn - 2K D), the
starting oligomer, are very simple. The MeHSiO 29Si peak appears at -34.6
8. The product, -[MeHSiO]o. 3[MeSi(O) 1 .5]0.7, obtained from Ti catalyzed
redistribution, shows several 29Si peaks in the same vicinity, -33 to -36
8. It also shows two peaks at -57.2 and -65.5 8. These results, when
coupled with the reproducible 30:70, [MeHSiO] : [MeSi(O) 1 .5] ratio, suggest
a polymer structure consisting of open cubes of T groups bridged by one or
two -MeHSiO- groups as depicted below. By visual inspection, the peak at
Me
Me
SS -
Mee
O
__O
1 4.6
0-M'
Me M
-
-6
Mell(
Me
/'
OR/
"IS' = -34.68
Sl z -57.2 8
29Sl = -64.38
-57.2 8 is much smaller than the peak at -65.5 8. Consequently, we assign
this peak to the open silicons in the cube and the -65.5 8 peak to the
remaining T group silicons in the cube. In the nBuO- derivative, the -34.6
8 peak is replaced by the appearence of a peak at ca. -64 8. This peak
appears to overlap with the T groups in the cube in the nPrO derivative.
We assign this peak to the alkoxy substituted silicons.
Hiah Temoerature Studies
TGA studies indicate that the high temperature stability of the
alkoxy derivatives is very similar to that of the starting copolymer. The
9000C ceramic yields for the set of copolymers are 76% (MeHSiO), 75%
37
(Me(MeO)SiO], 74% [Me(EtO)SiO], 78%/ [Me(nPrO)SiO] and 62%/6 for the
[Me(nBuO)SiO] derivative. As in Figure 1, most of the weight loss occurs
below 400-450 0C. If weight loss in the alkoxy derivatives occurs by a
mechanism similar to that found for the hydrido copolymer, then weight
loss must occur by depolymerization and volatilization of Me(RO)SiO
groups. One would expect increasing weight losses with increases in the
size of the R' group. Clearly this is not the case with the R' - Me, Et or nPr
derivatives. In these cases, the ceramic yields are comparable to that of
the original copolymer. Even in the nBuO- derivative, the mass of the
group increases from 60 D (MeHSiO) to 132 D [Me(OnBu)SiO]. If complete
loss of 60 D leads to a ceramic yield of 75% then complete loss of 132 D
should lead to a ceramic yield of <50%. These results inidcate that the
ceramic products from pyrolysis of these materials retain the carbons in
the alkoxy groups.
This is in contrast to studies by Fox et al. 4 on the pyrolysis of the
polyalkylsilsesquioxanes, -[RSi(O) 1. 5]x - , where increasing the size of R
from Me to Et to Pr resulted in drops in the 9000C ceramic yields from
86% (Me) to 470/ (Et) to 44% (nPr). In all instances, except for the Me
derivative, our ceramic yields are much higher.
It is likely that these differences arise because the bond dissociation
energy for Si-C bonds is approximately 85-90 kcal/mole whereas O-C
bond dissociation energies are typically around 100 kcal/mole. 12 Thus,
the decomposition mechanisms for the two types of polymers are quite
different. The important point to be made is that proportionately, the
EtO- and nPrO- derivatives incorporate more carbon in the ceramic
product than the EtSi and nPrSi derivatives, which should result in a
higher proportion of SiC in the final composite ceramic/black glass.
ACIN.EDGEMENTS
We would like to thank the Strategic Defense Sciences Office through
the Office of Naval Research for support of this work through ONR
contract No. NOOO14-88-K-0305. We would also like to thank the
Washington Technology Center for support of this work through equipment
purchases in the Advanced Materials Program. RML would like to thank IBM
for partial support of this work. RML would also like to acknowledge the
Department of Chemistry, the Technion, Haifa, Israel and Professor J.
Katriel for providing accomodations and facilities during the preparation
of this manuscript.
38
REFEECES
1.
2.
3.
4.
5.
a. M. G. Voronkov, V. 1.Lavrent'yev, Top. Curr. Chem. =0, 199 (1982). b.
C. L. Frye, J. M. Klosowski, J. Am. Chem. Soc., 3, 4599 (1971).
R. M. Lamne, J.A. Rahn, K. A.Youngdahl, F. Babonneau, J. F. Harrod
submitted to Chem. Mat.
See for example: a. J. R.January; U. S. Patent No. 4,472,510. b. S.
Uchimara, Eur. Pat. Appl. EP 312,280--CA 111:1 56040j. c. H. Adachi,
E. Adachi, 0. Hayashi, K. Okabashi, Jpn. Kok. Tokk. Koho JP 01 92,224.
a. D. A.White, S. M. Oleff, R. D. Boyer, P. A. Budringer, J. R. Fox, Adv.
Cer. Mat., 2, 45 (1987). b. D. A. White, S. M. Oleff, J. R. Fox, ibid p.53.
a. R. Baney in Ultrastructure Processing of Ceramics- Glasses, and
Comoasote, edited by L. L. Hench and D.R.Ulrich, (Wiley-lnterscience,
6.
7.
8.
9.
10.
11.
12.
1984) pp 245-255. b.H. Zhang and C. Pantano in "Proceedings of the
Fourth Internat. Confer. on Ultrastruct. of Ceramics, Glasses and
Composotes, edited by D. Uhlmann and D. R. Ulrich 1989, in press.
K. Kamiya, 0. Makoto, T. Yoko, J. Noncryst. Sol.; 83, 208 (1986).
a. J. F. Harrod, S. Xin, C. Aitken, Y. Mu, and E. Samuel, International
Conference on Silicon Chemistry, June,1986; St. Louis, Mo. b. J. F.
Harrod, S. Xin, C. Aitken, Y. Mu, E. Samuel, submitted to Can. J. Chem.
c. For a review on transition metal catalyzed synthesis of inorganic
polymers see: , R. M. Lamne in Aspecs of Homoageneous Catalyss
edited by R. Ugo (Kluwer pub., Dor2'echt, 1989) vol Z, in press.
K. A. Youngdahl, M. L. Hoppe, R. M. Lamne, J. A. Rahn, and J. F. Harrod,
"Proceedings of the 4 th Internat- Confer, an Ultrastruct. of Ceramics,
Glasses and Composites, edited by D. Uhlmann and D. R. Ulrich (Wileylnterscience, 1989) in press.
R. M. Lamne, Z.-F. Zhang and J. A. Rahn unpublished results. Detailed
experimental results will be reported elsewhere.
X-L. Luo and R.H. Crabtree, J. Am. Chem. Soc. =ii, 2527 (1989).
a. H. Marsmann and J. P. Kintzinger in Oxygen 17 and Silicon 29 NMR.
(Springer-Verlag, 1981) New York pp 74-239. b. E. A. Williams in IJ32
Chemistry of Organic Silicon Compgunds, edited by S. Patai and Z.
Rappoport (John-Wiley and Sons, 1989) pp 512-554.
a. Handbook of Chemistry and Physics. CRC CoL (Chemical Rubber Co.
6 4 th Ed.) p F-193. b. R.Walsh in The Chemistry of Organic Silicon
Compounds. edited by S. Patai and Z. Rappoport (John-Wiley and Sons,
1989) pp 371-392 and references therein.
CONDUCTING MOLECULAR MULTILAYERS:
INTERCALATION OF CONJUGATED POLYMERS IN LAYERED MEDIA
*
V. MEHROTRA AND E.P. GIANNELIS
Department of Materials Science and Engineering, Cornell University,
Ithaca, NY 14853.
ABSTRACT
Polyaniline has been synthesized in the galleries of fluorohectorite, a
two-dimensional mica-type layered silicate. Intercalation of aniline in the
intracrystalline region of Cu-exchanged fluorohectorite results in oxidative
polymerization to polyaniline (emeraldine base form) as demonstrated by
electronic, infrared and Raman spectroscopy and x-ray diffraction data.
The intercalated insulating form of polyaniline becomes conducting on
exposure to HCl.
In-plane electrical conductivity data measured in the
temperature range 274 to 573 K show a complex thermally activated behavior
"1
with room temperature conductivity 0.05 Ohm'icm
. The polyaniline/layered
silicate hybrids represent a new class of nanocomposites consisting of
synthetic conductors with molecular dimensions contained in a quasi twodimensional environment of a crystalline host.
INTRODUCTION
Research on electrically conducting polymers has recently focused on
the synthesis and characterization of environmentally robust materials with
high degree of processability and high conductivity [1].
An intriguing new
challenge in the area of conducting polymers is to isolate or confine them
in a well defined environment.
Such systems would not only provide
fundamental information on the electronic structure and the conduction
mechanism but they could also represent a new class of hybrid composites
with nanometer dimensions.
The properties of these new materials can be
tailored by inserting suitable guest molecules into the well defined
galleries of the host lattice.
Previous work in this area includes intercalation of polypyrrole and
polythiophene in layered FeOCI [2], intercalation of aniline in V 20 5 [3],
and encapsulation of chains of aniline, pyrrole and thiophene ini zeolites
[4,5].
Formation of conjugated polymers in the galleries of mica-type
silicates was first observed almost twenty years ago by Pinnavaia and
Mortland by a unique charge-transfer reaction (6].
Certain aromatic
molecules can be incorporated into the galleries of layered silicates whose
exchangeable cations have been replaced by transition metal ions like Cu,
Ru, etc.,
and form charge-transfer complexes.
For example in the case of
benzene, intercalation leads to the formation of poly-p-phenylene under
exhaustively dry conditions.
Previous studies focused mainly on spectroscopic studies offering no
information on their electronic properties or conduction mechanism (7,81.
In addition, intercalation of aniline in layered silicates has not been
studied. Polyaniline (PANI) is - novel conducting polymer, that in contrast
to other conducting polymers its electronic properties can be switched from
insulating to conducting by electrochemical oxidation and/or chemical
protonation.
Mat. Res. Sot. Syrp. Proc. Vol. 171.
190 Materials Research Society
40
We
report
here
the
first
successful
synthesis
of
in
situ
intercalation/polymerization of aniline in layered mica-type silicates. The
resulting materials constitute a pseudo two-dimensional composite with
nanometer architecture.
EXPERIMENTAL
Materials
Synthetic fluorohectorite was kindly provided by Corning Inc. Typical
particles have a platelike morphology with an average layer diameter of 5
microns. The cation exchange capacity for fluorohectorite is 190 meq/100 g.
The copper-exchange form was prepared by treating the host with 1 M solution
of Cu(N0 3 )2 , centrifuging and discarding the supernatant liquid.
Excess
Cu(N0 3 )2 was removed by washing with deionized water.
Thin films of fluorohectorite (0.5 - 25 microns) were prepared by
evaporation of a suspension on a polyethylene plate (for self-supporting
films) or on a glass substrate.
Aniline was intercalated in a P2 05
containing desiccator from the vapor phase until the reaction was completed
within 3-4 days. An alternative route employed dipping the fluorohectorite
film into aniline for a few days.
The conducting form of polyaniline was
formed by exposing the PANI/fluorohectorite films to HCl vapors for 2 to 3
0
hours followed by mild heating at 45 C to remove excess HCl.
Instrumentation and Methods
X-ray diffraction patterns of oriented film samples were obtained on a
Scintag x-ray diffractometer using Ni-filtered CuKa radiation.
Infrared
spectra of self-supporting films were recorded on a Perkin Elmer 1330
spectrometer.
Electronic absorption spectra were obtained with a Perkin
Elmer Lambda 4A spectrophotometer.
Raman experiments were performed with a
SPEX 1877 triplemate spectrometer, using a Coherent Nova 90-5 Ar+ laser with
a maximum power of 2 W in the 457.9 nm line.
Four-point "bulk"
conductivity measurements were performed at room temperature using a
Prometrix Versaprobe VP1O System.
In-plane conductivity of oriented films
was measured by a Hewlett-Packard 4145B semiconductor parameter analyzer.
Samples for conductivity measurements were prepared by depositing a film
onto HF cleaned glass slides.
Silver electrodes were deposited on both
edges of the film.
The current-voltage characteristics of the sample were
measured at each temperature in a shielded test enclosure.
Transient
currents were allowed to dissipate before each measurement.
The applied
voltages ranged from -10 to +10 Volts in steps of 0.5 V at each temperature.
0
0
Electrical properties were studied in the 1 C to 300 C range.
RESULTS AND DISCUSSION
Fluorohectorite is a synthetic mica-type silicate (MTS) with a layered
lattice structure in which two-dimensional multiple cross-linked planes of
atoms are separated by layers of hydrated cations.
Figure 1 schematically
illustrates the layered structure where two tetrahedral silicate sheets are
fused to a central octahedral sheet of magnesium hydroxide [9].
The
stacking of these layers to form crystals leads to the formation of
interlayers or "galleries" where the layers are held together by van der
Waals forces.
In fluorohectorite and other charged silicates the galleries
are occupied by hydrated cations in order to balance the charge deficiency
41
Figure 1. Idealized structure of a mica-type silicate.
that has been generated by the isomorphous substitution in the tetrahedral
or octahedral sheets. The strong intraplanar and weak interplanar binding
forces that arise from the two-dimensional structure allow the introduction
(intercalation) of guest species into the galleries of the host lattice.
Simple intercalative and/or ion exchange procedures permit a variety of
neutral molecules or cations of virtually any size to be accommodated in the
galleries of MTSs.
Principal themes that have emerged from various
spectroscopic techniques is that the properties of the gallery species are
generally preserved upon intercalation.
Intercalation of aniline in Cu-exchanged fluorohectorite from the
gaseous phase is relatively slow; however, when the film is dipped into neat
aniline a rapid color change occurs and the reaction is complete U.thin 24
hours.
In contrast, no polymerization occurs when Li+ ions are present in
the galleries.
Chemical oxidation of aniline yields the emeraldine base
form of polyaniline that consists of equal number of the amine [-(C 6 H4 NH)-]
and
quinoidal
[-(C6H 4 NH)-l
repeating
units.
In
contrast,
when
polymerization
is
carried
out
in
acidic
media,
the
protonated
poly(semiquinone), a radical cation structure is formed. This conducting
form of polyaniline is believed to consist of equal numbers of reduced ((C6 H4 NH)-] and oxidized, protonated quinone (-(C6 H4 NH+)-] repeat units 110].
2
Gallery Cu + ions, inti Juced by an intercalative ion exchange process,
serve as the oxidation centers for the oxidative polymerization of aniline
in the intracrystalline environment of the host structure. The reaction, by
analogy to
other conjugated polymers in layered
silicates, can be
represented by the following equation, where the horizontal lines identify
the layered structure.
Cu
2+
+
nC6 H 5NH 2
2
-
Cu +(PANI)
42
X-ray diffraction patterns
of oriented films
show
that
in situ
intercalation/polymerization
of
aniline
results
in
a
highly
ordered
composite with a quasi two dimensional structure (Figure 2).
Several (001)
harmonics are observed corresponding to a primary repeat unit (d spacing) of
14.9 A.
The difference of 5.3 A from the corresponding 9.6 X for the
silicate framework is in agreement with intercalation of single chains of
polyaniline.
The electronic absorption spectrum of the as prepared material shows an
absorption band at 510 rn (2.4 eV) and 330 (3.7 eV) which have been
attributed to the quinone diimine structure [-N-(C 6 H 4 )-N-] of the emeraldine
base (Figure 3).
The host exhibits a characteristic absorption at 4.5 eV.
However, when the sample is exposed to HCl vapors the band at 510 rum
gradually disappears and a new broad band appears centered at 760 run (1.6
eV) that
extends into the near ir.
The absorption at 3.4 eV and the broad
feature extending into
ir have been
associated with
radical cations
supporting polarons as charge carriers.
Further evidence for the presence of intercalated polyaniline comes
from ir and Raman spectroscopy.
Ir spectra of oriented films show typical
-1
vibrations of polyaniline at 1595, 1490, 1305 and 1245 cm .
The bands at
1595 and 1245 cmare
associated with the
reduced
repeat units
["1
(C 6 H 4 NH)-I while the 1490 and 1305 cm
are attributed to the quinone
"I
diimine and protonated oxidized repeat units respectively.
The 1305 cm
band is also characteristic of an electron-phonon interaction.
Raman
spectroscopy with 457.9 nm excitation shows absorptions at 1630, 1370, 1325
"1
and 1205 cm
in agreement with the above assignments.
2
The role
of Cu +
initiating
in
the polymerization
of
analogous
aromatic systems in the silicate galleries has been established in the past
by electron paramagnetic resonance and x-ray photoelectron spectroscopy [681.
Similar experiments for the polyaniline intercalate are underway in our
laboratory and the results will be communicated in the near future.
The as
prepared
polyaniline
intercalate
shows
excellent
thermal
0
stability up to 700 C.
After an initial weight loss of approximately 3%
presumably due to residual monomer trapped during the reaction and/or
adsorbed atmospheric moisture there is no further weight loss up to 700'C.
Electrical conductivity measurements were performed in the four-probe
and in-plane geometry using oriented film samples.
The room temperature
"
conductivity of HCI exposed samples is 0.05 Ohm-lcm
an increase by five
orders of magnitude with respect to Cu-exchanged host.
Figure 4 shows the
variable temperature in-plane conductivity of samples exposed to HCI.
The
current-voltage characteristics were ohmic throughout the temperature range.
Experimental points in the figure are connected as a guide to the eye.
As
the sample is heated from room temperature the conductivity increases with
increasing temperatures
as expected for
semiconducting behavior or in
different variable
range
hopping models.
The
conductivity
increases
reversibly
from
room
temperature
to
about
60
°C,
above
which
the
conductivity starts to decrease slowly.
After heating to temperatures
0
higher than 300
C and rapidly cooling to room temperature the sample
exhibits a conductivity substantially lower from that at room temperature
but the behavior is now metallic; i.e. an increase in temperature leads to a
decrease in conductivity.
The variable temperature conductivity data agree
very well with the behavior observed for bulk polyaniline [Il1.
Preliminary
conductivity measurements Rervendicular to the silicate layers affirm the
highly anisotroic nature of the material.
43
Figure 2. X-ray diffraction pattern for PANT-MTS multilayer
I ..
.
I
"
o
"I
/
I
0
.
0.5
200
400
600
,v'ovelenq
Figure 3. Electronic
800
(nni)
absorption
intercalated MTS (solid line);
line).
spectra of:
(a)
as
prepared PANI
(b) after exposure to HCI vapors (broken
The polyaniline/layered silicate hybrids represent a new class of
conducting nanocomposites.
These materials consist of conducting polymer
chains and insulating host layers alternately stacked to form a multilayered
structure with molecular dimensions and atomically sharp interfaces.
We
have
extended this
approach of
intercalative
polymerization
polymer/host
systems.
The
structural,
spectroscopic
and
properties of these systems will be reported in the future.
to other
transport
4,4
10
4
103
"0102
101
0
100
50
Figure 4. Resistivity of
function of temperature.
1;0
200
Temperature ('C)
intercalated MTS
PANI
3;0
2;0
exposed
350
to HC1
as a
ACKNOWLEDGMENT
Foundation
This work was sponsored by the National Science
8818558) through the Materials Science Center at Cornell.
(DMR-
REFERENCES
of Conductin g
Polymers, Vol.
1,
(Marcel
1.
T.A. Skotheim Ed. , Handbook
Dekker, New York, 1986).
2.
C.R. Kannewurf and T.J.
M.G. Kanatzidis, H.O. Marcy, W.J. McCarthy,
Marks, Solid State IonTcs, 32/33, 594 (1989).
3.
M.G. Kanatzidis, C.-G. Wu, H.O. Marcy and C.R. Kannewurf, J. Am. Chem.
Soc., Mi, 4139 (1989).
4.
P. Enzel and T. Bein, J. Phys. Chem., 93, 6270 (1989)
5.
P. Enzel and T. Bein, J. Chem. Soc., Chem. Commun., 1326 (1989)
6.
f.M. Mortland and T.J. Pinnavaia, Nature,
7.
J.P. Rupert, J. Phys. Chem., 2Z, 784 (1973).
8.
Y. Soma, M. Soma and I. Harada, J. Phys. Chem.,
9.
T.J. P
Science, 2
tnnavaia,
W2, 75 (1971)
365
ce, (1983).
Chiang and A.G. MacDiarmid, Synth. Met.,
J.C.k0.
2K.
11.i
3034 (1984).
11, 193 (1986).
. Uvdal, M.A. Hasan, J.O. Nilsson, W.R. Salaneck, . Lundstrom, A.G.
MacDarmid, A. Ray and A, Angelopoulos, in Electronic Pro5ert(es o
3 onugated Polymers, edited by H. Kuzmany, M. Mehrng and S. Roth
(Spr.ngrVeterlag, Berlin, 1987).
-
UvdlM.Hs,
nl
iK.
J.0. ilson
W..Sl
cI
udtoAG
45
NYLON 6-CLAY HYBRID
AKANE OKADA, MASAYA KAWASUMI, ARIMITSU USUKI,
YOSHITSUGU KOJIMA, TOSHIO KURAUCHI AND OSAMI KAMIGAITO
Toyota Central Research and Development Laboratories, Inc.,
Nagakute, Aichi, 480-11, Japan
ABSTRACT
t-Caprolactam was polymerized in the interlayer
spacing of montmorillonite, a clay mineral, yielding
a nylon 6-clay hybrid (NCH) ".
X-ray and TEM measurements revealed that each template of
the silicate, which is 10 A thick, was dispersed in the
nylon 6 matrix and that the repeat unit increased from 12 A
in unintercalated material to 214 A in the intercalated
material.
Thus NCH, is a "polymer based molecular composite" or
"nanometer composite". NCH, when injection-molded, shows
excellent properties as compared to nylon 6 in terms of
tensile strength, tensile modulus and heat resistance.
Heat distortion temperature increased from 65 r for nylon 6
to 152 t for NCH, containing 4 wtS (1.6 vol6) of clay
mineral.
INTRODUCTION
Nylon 6 (polycaploractam) has good mechanical
properties and is a commonly used engineering polymer. It
has been successfully reinforced by glass fiber or other
inorganic materials 2 . In these reinforced composites, the
polymer and additives are not homogeneously dispersed at the
microscopic level. If the dispersion could be achieved at
the microscopic level, the mechanical properties would be
expected to be further improved and/or new unexpected
features might appear. Clay mineral is a potential candidate
for the additive since it is composed of layered silicates,
10 A thick, and undergoes intercalation with organic
molecules. A conceptive picture is illustrated in Figure 1.
EXPERIMENTAL
Materials
Montmorillonite "Kunipia F" was supplied by Kunimine
Ind. Co., with a cation exchange capacity of 119 mili
equivalents/100 g. Montmorillonite is a fine sheet-like
particle with a dimension of about 0. 1 pm in length, 0. 1 pm
width and 10 A in thickness. Other inorganic and organic
materials were commercially available.
Mat. Res.Soc. Symp. Proc. Vol. 171.
1990Maerials
O
Research Society
48
0
0 0
0/0 00
00
0
000
00
/0
o00
0 0
o
0
0
0
O0 0
0 0
0 0000o0
0 00
0
00
monomer
polymerization
0\\\
0
0 0o 0
0
0
0
0
o oo0
0000
00
o
0
0 oO0
o
layered clay mineral
a layer of clay
Fig.1
polymer
A conceptive picture of polymerization
in the presence of clay
Preparation of 12-montmorillonite
In an electric mixer, 25 g of montmorillonite and
1. 75 1 of water were mixed.
The mixture was stirred at 5000
rpm for 5 min. , and then an aqueous suspension of clay was
obtained. 12. 8 g of 12-aminolauric acid and 6. 0 ml of
hydrochloric acid was then added to the suspension and the
mixture was further stirred at 5000 rpm for 5 min. The
product was filtered, washed with 1 1 of water, freeze-dried
and dried in vacuo at 100 t, which finally yielded
intercalated montmorillonite with 12-aminolauric acid
(termed as 12-montmorillonite).
Preparation of hybrid
A typical run is described. The reaction vessel was
composed of a 500 ml-three necked separable flask with
mechanical stirrer.
In the vessel, 113 g of i-caprolactam
and 5.97 g of 12-montmorillonite was placed. The mixture
was heated at 100 t in an oil bath with stirring for 30
min. The temperature was then elevated to 250 t and
maintained for 48 hr.
After cooling, the product was
mechanically crushed.
The finely divided particles were
washed with 2 1 of water at 80 r for I hr. Drying at 80 t
overnight yielded a nylon 6-clay hybrid (termed as NCH).
Nylon 6-clay composites (termed as NCC) were
prepared by blending commercial nylon-6 and montmollironite
in an extruder for comparison with NCH. These materials
were injection-molded for measurement.
47
Characterization
X-ray measurement was done with a Rigaku RU-3L X-ray
Diffractometer using Co Ka radiation.
Transmission electron micrographs were obtained with a
Jeol-200CX TEM using an acceleration voltage of 200 kV.
Viscoelastic measurement was done using Iwamoto
Seisakusho VES-F Viscoelastometer.
Tensile strength and other mechanical properties were
measured following ASTM.
RESULTS AND DISCUSSION
Polymerization
When E-caprolactam (mp 70 t) and 12-montmorillonite
were heated at 100 12under stirring, the mixture yielded a
viscous dispersion. Interlayer distance in the suspension
was 40 A as compared to 17 A for 12-montmorillonite. This
indicates that the monomer was intercalated into the
silicates. Polymerization was performed at 250 1.
Interlayer distance, D, could be directly obtained in
XRD diagrams.
Figure 2 compares the transmission electron micrographs
of sections of molded NCH. The dark lines are intersections
of sheet silicate of 10 A thickness and spacings between
the dark lines are interlayer distances.
Table I shows the interlayer distance, D, obtained by
XRD and TEM. The D values agree very well.
D was found to be inversely proportional to the montmorillonite content. A maximum D of 214 A was observed.
The thickness of a layer of silicate is about 10 A.
This is of the order of molecular-size and therefore can be
thought to be an "inorganic macromolecule" so that in NCH,
the polymer and montmorillonite are mixed at a molecular
level forming a "polymer based molecular composite".
On the other hand, D in the NCC was 12 A and therefore it
is not a "nanometer composite".
Properties
The dynamic elastic moduli (E') were obtained at 10 ft.
between -150 t and 250 t. The moduli of NCH-5 exceeded those
of NCC-5 and nylon 6 in all the region of temperature.
The modulus of NCH-5 was more than twice that of other
specimens around 120 t.
Mechanical properties of NCH-5 are shown in Table 2
together with nylon 6 and NCC-5. The tensile strength and
tensile modulus of NCH were superior to others. Impact
strength of NCH was comparable to nylon 6.
48
(a)NCH- 15
(b)NCH-30
1000 A
OO
Fig. 2 TEM of sections of NCHs
49
Table 1
Interlayer distances of NGHs
Specimen
Montmorillonite
Distance
Distance
(wt%)
(X-ray, A)
(TEM, A)
NCH-5
4.2
>150
214
NCH-10
9.0
121
115
NCH- 15
14.5
64
62
NCH-30
25.0
51
50
NCC-5
5.0
12
Table 2
Properties of NCH-5 (1)
Specimen
Tensile strength
Tensile modulus
Montmorillonite
(MPa)
(GPa)
Charpy impact
strength
(Wt%)
(KJ/mz)
NCH-5
(4.2)
107
2.1
6.1
NCC-5
(5.0)
61
1.0
5.9
nylon 6
(0)
69
1.1
6.2
Table 2 (continued)
Properties of NCH-5 (2)
Specimen
HOT
Rate of water
Montmorillonite
at 18.5kg/cm,
absorption
(wt%)
(C)
23*C,1 day
Coefficient of thermal expansion
flow
Perpendicular
direction
direction
(cm/cm*C)
M%
NCH-5
(4.2)
NCC-5
(5.0)
152
0.51
89
0.90
10.3
13.4
nylon 6
(0)
65
0.87
11.7
11.8
6.3x10'
13.lxI10
50
The most prominent effect was observed in heat distorHDT of NCH-5 containing only 4 wt%
tion temperature (HDT).
higher than
, which was 87
of montmorillonite was 152
This effect in NCH is a drastic
that of nylon 6.
improvement in the quality of nylon 6.
The rate of
Resistance to water was also improved.
water absorption in NCH was lowered by 40 % as compared to
nylon 6 and NCC.
The molded specimen was found to be anisotropic.
The coefficient of thermal expansion of NCH-5 in the flow
direction was lower than half of that in the perpendicular
direction. Nylon 6 was isotropic and NCC was intermediate.
Sheets of
silicate were parallel to the flow direction of
the mold.
The polymer chains also oriented in the same
It seems that anisotropy of thermal expansion
direction.
resulted from the orientations of silicate and polymer
chains.
Excellent properties in NCH can be considerd to have
origin in an enormous surface area and ionic bonds
the organic polymer and inorganic silicate sheet.
between
CONCLUSION
t-Caprolactam was polymerized in the interlayer
spacing of montmorillonite, a clay mineral, yielding
a nylon 6-clay hybrid (NCH).
XRD and TEM studies have revealed that this NCH is a real
"polymer based molecular composite" or "nanometer
composite".
NCH can be injection-molded and shows excellent
properties as compared to nylon 6 in terms of tensile
strength, tensile modulus and heat resistance.
NCH is now open to practical use.
We believe that this method has opened a new field and is
a novel processing technique exploiting the intercalation
properties of layered compounds.
REFERENCE
l.Presented in part at the 194th National Meetings of
American Chemical Society. New Orleans. La., August 1987.
Polymer Preprint 28, 447 (1987).
"Nylon Plastics". lnterscience, New York
2.M. l.Kohan Ed.,
1973.
51
REINFORCEMENT OF ELASTOMERS BY THE IN-SITU
GENERATION OF FILLER PARTICLES
JAMES E. MARK* AND DALE W. SCHAEFER**
*University of Cincinnati, Dept. of Chemistry, Cincinnati, OH 45221
"Sandia National Laboratories, Albuquerque, NM 87185
ABSTRACT
The goal of primary interest in these investigations was the development of novel
methods for filling elastomeric networks. The techniques developed employ the in-situ
generation of reinforcing fillers such as silica or a glassy polymer such as polystyrene either
after, during, or before network formation. The reaction involves decomposition of
organometallic compounds, using a variety of catalysts and precipitation conditions, or freeradical polymerization of a suitable monomer. The effectiveness of the technique is gauged by
stress-strain measurements carried out on these elastomeric composites to yield values of the
maximum extensibility, ultimate strength, and energy of rpture. Also of interest are calorimetric
studies of the networks, to determine their crystallizability. Information on the filler particles
themselves is obtained from density determinations, electron microscopy, and scattering
measurements.
INTRODUCTION
There are a number of disadvantages to reinforcing an elastomer by the usual technique
l1,2] of blending a finely divided filler (such as carbon black or silica) into a polymer before
cross linking it [3]. A number of alternative techniques are therefore also under development.
Examples of such techniques are presented here, with a strong emphasis on results beyond those
described in three recent reviews which are at least partly on the same subject 14-6]. They
include hydrolysis of organometallic compounds within a polymeric matrix to give ceramic
particles such as silica and titania. The semi-inorganic polymer, poly(dimethylsiloxane)
(PDMS) has been most studied in this regard. The case where the ceramic predominates and
becomes
the continuous
is also mentioned.
An alternative
approach
monomers
such as styrene
or methylphase
methacrylate
are polymerized
in-situ to give
glassywhere
polymers
is also
described, as are some related systems in which there are magnetic particles or zeolites.
4
Of primary interest here is the reinforcement provided by these fillers. It is easy to
switch the focus, however, so that the elastomer is viewed as only a matrix in which the ceramic
materials are being generated. In this "matrix isolation" approach (7], X-ray and neutron
scattering techniques, for example, can be used to obtain information that transcends these
particular systems. It should be useful in a variety of areas, including the new sol-gel technique
for preparing ceramics of carefully controlled ultrastructure 18-11].
IN-SITU PRECIPITATIONS
Typical Hydrolysis Reactions
The most important reaction in this area is the acid or base catalyzed hydrolysis of
tetraethoxysilane (TEOS) as described by the chemical equation 14-61,
Si(OCH
+2H 2
4 +)
- SiO 2 + 4C2 H50H
(1)
Analogous reactions [8-11 can be carried out, however, on titanates [12-141, aluminates [I5],
zirconates [161. In the sol-gel technique, the process first gives a (swollen) gel, which is then
dried, fimd, and densif'ed into a final, monolithic piece of silica 18-11). There have now been a
number of additional studies using essentially the same reactions, but in a very different context
14-61. Specifically, the hydrolysis reactions are carried out within a polymeric matrix, with the
Mat. Res. S"c. Symp. Ploc. Val. 171. '1990 Mateials ReSearch Society
52
ceramic frequently generated in the form of very small, well-dispersed particles. When the
matrix is an elastomer, these particles provide the same highly desirable reinforcing effects
obtained by the usual blending of a filler (such as carbon black) into polymers (such as natural
rubber) prior to their being cross-linked or cured into tough elastomers of commercial
importance [ 1,2].
These reactions can be carried out in three ways [4-6]. In the first, the polymer is cross
linked and then swelled with the organometallic, which is then hydrolyzed in-situ. In the
second, hydroxyl-terminated chains are blended with enough TEOS to both end link them and
provide silica by the hydrolysis reaction. Thus, curing and filling take place simultaneously, in
a one-step process. In the final technique, TEOS is blended into polymer having end groups
(e.g. vinyls) that are unreactive under hydrolysis conditions. The silica is then formed in the
usual manner [Eqn (1)], and the product dried. The resulting slurry of polymer and silica is
stable and can be cross linked later using any of the standard cross linking techniques, such as
vinyl-silane coupling [5,17,18].
The kinetics of this reaction are being studied [19] using air-pressure deformation
measurements [20] on the gels, in a manner similar to that used to characterize thernoreversible
polyethylene gels [21].
Comarisons Among Various Silica-Based Fillers
There are a variety of ways to generate silica-type fillers useful for reinforcing PDMS
networks. The extent to which such fillers provide reinforcement was characterized in a recent
study [22]. The materials and techniques employed were (i) incorporating a commercial silica
which had been treated with hexamethyldisilazane as a coupling agent, (ii) PDMS with silica
which had been precipitated from an aqueous dispersion, (iii) precipitating silica directly into
PDMS during its curing, (iv) precipitating silica directly into a swollen PDMS network after it
was cured, (v) incorporating silica prepared from tetraethoxysilane (TEOS) and containing some
PDMS, and (vi) incorporating silica prerared from partially hydrolyzed TEOS and also
containing some PDMS. The resulting filled elastomers showed the largest values of the
ultimate strength in the case of (iv) and (vi), and the largest value of the rupture energy for (iv).
Most of the studies to date have involved PDMS, primarily because of its great
miscibility with TEOS. Similar studies [23] on the related polymer, poly(methylphenylsiloxane), however, are of considerable importance. In this case the stereochemically irregular
structure of the polymer prevents strain-induced crystallization [5,24]. Good in-situ generated
reinforcement was also achieved in this polymer, suggesting that such crystallization is not
necessary for reinforcement [231.
The same techniques have also been shown to give good reinforcement in
polyisobutylene elastomers [25], and in poly(ethyl acrylate) 1261. In the latter case, it appears
that the precipitation can be carried out during an emulsion polymerization [26].
Other Ceramic-Type Fillers
Silica particles in PDMS elastomers can be a problem at high temperature, since the
silanol groups on their surfaces can cause degradation of the polymer 113,271. For this, and
other reasons, a variety of other fillers have been precipitated into this polymer. Included are
titania (TiO 2) [12-141, alumina (AI 2 03) [151, and zirconia (ZtO2) [16].
These non-silica fillers also provide good reinforcement. One interesting difference,
however, is the observation that the stress-strain isotherms in these cases frequently have much
better reversibility [141. Reversibility is presumably due to different interactions between the
surface groups present on these particles and the PDMS elastomeric matrix.
,IF
+
Aging Effects
Permitting precipitated silica particles to remain in contact with their aqueous catalyst
solution can permit them to "age' or "digest" [28,291. Electron microscopy results suggest that
some reorganization is occurring, with the particles becoming better defined, more uniform in
size, and possibly even less aggregated. There seem to be interesting parallels with "Ostwald
ripening" in the area of colloid science 130].
"3
Comparisons between the values of wt % filler obtained from density measurements and
the values obtained directly from weight increases can give very useful information on the filler
particles. For example, the fact that the former estimate is smaller than the latter in the case of
silica-filled PDMS elastomers 131] indicates that there are probably either voids or unreacted
organic groups in the filler particles.
Differential scanning calorimetry measurements at low temperatures were carried out on
PDMS elastomers containing in-situ precipitated silica 1321. The presence of the silica was
found to reduce both the extent of crystallization and the rate of crystallization when the
elastomers were in the unstretched state in contrast to similar studies of PDMS in the stretched
state, where the filler may facilitate the crystallization process 133).
Both tran'- nission [4-6,28] and scanning 1341 electron microscopy have been used to
characterize these novel composite materials. The information obtained in this way includes (i)
the nature of the precipitated phase (particulate or non-particulate), (ii) the average particle size,
if particulate, (iii) the distribution of particle sizes, (iv) the degre to which the particles are well
defined, and (v) the degree of agglomeration of the particles.
One interesting result of this type is the conclusion that basic catalysts generally yield
particles that are well defined, whereas acidic catalysts yield particles that ar rather "fuzzy"
[4,28]. This conclusion is in agreement with results obtained earlier in sol-gel ceramics
investigations 1351.
Scatering Studies
A number of X-ray and neutron scattering studies have been carried out on these filled
elastomers 14,5,36,371. Although the results are generally consistent with those obtained by
electron microscopy, there are some intriguing differences. Of particular interest is the
observation that some fillers which appear to be particulate in electron microscopy, appear to
consist of a continuously interpenetrating phase by scattering measurements 136,37). Additional
comparisons could certainly be very illuminating in this regard.
POLYMER-MODIFIED CERAMICS
The technique of hydrolyzing an organometallic substance such as an alkyl silicate can be
generalized to make the silica generated the continuous phase, with domains of PDMS dispersed
in it. Of course, relatively high concentrations of the silicate are necessary. By varying its
amount, composite materials can be obtained ranging from relatively soft elastomers, to tough
hybrid materials, to brittle ceramics 138-401. Important properties to be correlated with
composition would include impact resistance, ultimate strength, maximum extensibility, and
viscoclastic effects.
54
IN-SITU POLYMERIZATIONS
Isotropic System~s
It is also possible to obtain reinforcement by polymerizing a monomer such as styrene to
yield hard glassy domains within the elastomer [41]. In PDMS, low concentrations of styrene
give low molecular weight polymer that acts more like a plasticizer than a reinforcing filler. The
initial plasticization effect is revealed by the stress-strain results which show an initial decrease
in the energy of rupture. This conclusion is supported by the absence of evidence for
polystyrene (PS) particles at lower styrene concentrations. Polyisobutylene has also been
reinforced in this manner [42]. In both cases, the particles are roughly spherical and the system
isoiropic.
The glassy particles thus generated are relatively easy to extract from the elastomeric
matrix, which means that there is little effective bonding between the two phases. It is possible,
however, to get excellent bonding onto the filler particles. One way is to include some
R'Si(OC 2H5 ) 3 , where R' is an unsaturated group. The R' groups on the surfaces of the
particles then participate in the polymerization, thereby bonding the elastomer chains to the
reinforcing particles [43]. Alternatively, the R'Si(OC 2H5 ) 3 can be used as one of the endlinking agents, placing unsaturated groups at the cross links [44]. Their participation in the
polymerization would now tie the PS domains to the elastomer's network structure.
The PS domains have the disadvantage of having a relatively low glass transition
temperature Tg (-1000C) [45] and being totally amorphous. Above Tg they would therefore
soften and presumably lose their reinforcing capability. For this reason, similar studies have
been carried out using poly(diphenylsiloxane) as the reinforcing phase. For an elastomer, this
material has a relatively high Tg of 490 C [46], is crystalline, and has an extraordinarily high
melting point of 5500C [46].
Anisotropic Systems
It is possible to convert the essentially spherical PS domains described above to rod-like
ellipsoidal particles [341. First, the PS-elastomer composite is raised to a temperature well
above the Tg of PS. It is then stretched uniaxially, and cooled while in the stretched state. The
particles are deformed into prolate ellipsoids, and retain this slope when cooled. When the
deforming force is removed, the elastomer is observed to retract, but only part of the way back
to its original dimensions. The particles themselves were characterized using scanning and
transmission electron microscopy, and found to have axial ratios of approximately 2, and to
have their axes preferentially oriented in the direction of the high-temperature stretching. The
reinforcement they provided were characterized using stress-strain measurements in elongation
at room temperature. In these anisotropic materials, the moduli in the direction parallel to the
original stretching direction was found to be significantly higher than that of the untreated
(isotropic) PS-PDMS elastomer, whereas in the perpendicular direction it was significantly
lower.
It should also be possible to generate oblate ellipsoids by stretching such a PS-PDMS
elastomer biaxially, for example, by inflation of a sheet of the material. Such experiments are in
progress [341.
SOME RELATED SYSTEMS
Magn tic Paicles
Some filler particles can be manipulated with a magnetic field [47,481. For example,
magnetic ferrite particles dispersed in PDMS can be aligned in a magnetic field during the crosslinking process. In this way anisotropic mechanical properties can be obtained, even from
cssentially spherical particles. The reinforcement is found to be significantly larger in the
dirrction parallel to the magnetic lines of force.
55
This technique could be combined with the in-situ approach by generating metal or metal
oxide magnetic particles in a magnetic field [49,501 for example by the thermolysis or photolysis
of a metal carbonyl.
Zeolites
One of the problems with fillers used to reinforce elastomers, however introduced, is
their amorphous nature [ 1,2]. This lack of a well-defined structure makes them poor choices for
determining how the structure of a filler affects the reinforcement it provides [51].
The zeolites are a related group of silicate-based materials which (i) are crystalline, (ii)
have conveniently-sized holes or cavities, and (iii) have had their structures extensively
documented [51,52]. Two zeolites have been investigated as fillers and found to give good
reinforcement of PDMS elastomers [51]. The one that had cavities 3 A in diameter was not
nearly as effective as the one having 13 A diameter cavities. In the latter case, the cavities may
have been large enough for them to be invaded by the PDMS chains, which could explain the
enhanced reinforcement.
It would be particularly exciting if such sieve-like materials of known structures could be
prepared in-situ.
ACKNOWLEDGEMENTS
It is a pleasure to acknowledge the financial support provided by the Air Force Office of
Scientific Research through Grant AFOSR 83-0027 (Chemical Structures Program, Division of
Chemical Sciences), the Army Research Office through Grant DAAL03-86-K-0032 (Materials
Science Division), and the National Science Foundation through Grant DMR 84-15082
(Polymers Program, Division of Materials Research). Work performed at Sandia National
Laboratories was supported by the U. S. Department of Energy under contract DE-AC0476DP00789.
REFERENCES
I.
2.
3.
4.
5.
6.
7.
8.
9.
10.
II.
12.
13.
14.
15.
16.
17.
18.
B. B. Boonstra, Polyme, 20, 691 (1979).
Z. Rigbi, Adv. Polym. Sci., 36, 21 (1980).
1. E. Mark and S.-J. Pan, Makromol. Chem., Rapid Commun., 3, 681 (1982).
J. E. Mark, in Ultrastructure Processing of Advanced Ceramics, ed. by J. D. MacKenzie
and D. R. Ulrich, Wiley, New York, 1988. Reprinted in CHEMTECH, 19, 230 (1989).
J. E. Mark and B. Erman, Rubberlike Elasticity. A Molecular Primer, Wiley-Interscience,
New York (1988).
J. E. Mark, in Frontiers of Macromolecular Science, ed. by T. Saegusa, T. Higashimura,
and A. Abe, Blackwell, Oxford, 1989.
S. Craddock and A. Hinchliffe, Matrix Isolation, Cambridge University Press, New York,
1975.
L. L. Hench and D. R. Ulrich, Eds., Ultrastructure Processing of Ceramics. Glasses. and
Compsites, John Wiley & Sons, New York, 1984.
L. L. Hench and D. R. Ulrich, Eds., Science of Ceramic Chemical Processing, John Wiley
& Sons, New York, 1986.
Ultrastructure Processing of Advanced Ceramics, ed. by J. D. MacKenzie and D. R.
Ulrich, Wiley, New York, 1988.
D. R. Ulrich, CH E MTECH, 18, 242 (1988).
G. S. Sur and J. E. Mark, Eur. Polym. J., 21, 1051 (1985).
S.-B. Wang and J. E. Mark, Polym. Bulletin, 17, 271 (1987).
S. J. Clarson and J. E. Mark, Polym. Commun., 30, 275 (1989).
J. E. Mark and S.-B. Wang, Polym. Bulletin, 20, 443 (1988).
S.-B. Wang and J. E. Mark, ms. in preparation.
M. A. Llorente and J. E. Mark, Macromolecules, 13, 681 (1980).
K. 0. Meyers, M. L. Bye, and E. W. Merrill, Ma
gjoleu, 13, 1045 (1980).
56
19. J. Li and J. E. Mark, unpublished results.
20. P. R. Saunders and A. G. Ward, in Proceedings of Second International Congress of
R.eolgy, Butterworths, London, 1953.
21. Z. Li, J. E. Mark, E. K. M. Chan, and L. Mandelkern, Macromolecules, 22, 000 (1989).
22. C.-C. Sun and J. E. Mark, Plymr, 30, 104 (1989).
23. S. J. Clarson and J. E. Mark, Polym. Commun., 28, 249 (1987).
24. J. E. Mark, Polyrn. Eng. Sci., 19, 409 (1979).
25. C.-C. Sun and J. E. Mark, L Polym. Sci., Polym. Phys. Ed., 25, 1561 (1987).
26. W. Qu and J. E. Mark, unpublished results.
27. D.K. Thomas, Poixme 7, 99 (1966).
28. J. E. Mark, Y.-P. Ning, M.-Y. Tang, and W. C. Roth, Polymer, 26, 2069 (1985).
29. P. Xu, S. Wang, and .. E. Mark, ms. in preparation.
30.
31.
32.
33.
34.
35.
36.
37.
38,
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
W. Ostwald, Z. Physik. Chem., 34, 495 (1900).
J. E. Mark, C,-Y. Jiang, and M.-Y. Tang, MacromolCgles, 17, 2613 (1984).
S. J. Clarson, J. E. Mark, and K. Dodgson, Polym. Commun., 29, 208 (1988).
V. Y. Levin, G. L. Slonimskii, K. A. Andrianov, A. A. Zhdanov, Y. K. Godovskii, V. S.
Papkov, and Y. A. Lyubavskaya, Polym. Sci. USSR, 15, 256 (1973).
S. Wang and J. E. Mark, ms. in preparation.
D, W. Schaefer and K. D. Keefer, Phys. Rev. Lett., 53, 1383 (1984).
D. W. Schaefer, L. Jian, C.-C. Sun, D. McCarthy, C.-Y. Jiang, Y.-P. Ning, J. E. Mark,
and S. Spooner, submitted to "Proceedings of Fourth International Conference on
Ultrastructure Processing of Ceramics, Glasses, and Composites", ed. by D. R. Ulrich and
D. R. Uhlman, Wiley, New York, 1989.
D. W. Schaefer, J. E. Mark, L. Jian, C.-C. Sun, D. McCarthy, C.-Y. Jiang, and Y.-P.
Ning, in Polymer-Based Molecular Composites, ed. by D. W. Schaefer and J. E. Mark,
(Mat. Res. Soc. Symp. Proc, 171, Pittsburgh PA, 1989).
H. K. Schmidt, in Inorganic and Organometallic Polymers, ed. by M. Zeldin, K. J.
Wynne, and H. R. Allcock, Am. Chem. Soc., Washington, DC, 1988, and relevant
references cited therein.
H.-H. Huang, G. L. Wilkes, and J. G. Carlson, Polymer, 30, 2001 (1989), and relevant
references cited therein.
J. E. Mark and C.-C. Sun, Polym. Bulletin, 18, 259 (1987).
F.-S. Fu and J. E. Mark, J. Polym. Sci., Polym. Phys. Ed., 26, 2229 (1988).
F.-S. Fu and 1. E. Mark, J. Apol. Polym. Sci., 37, 2757 (1989).
G. S. Sur and J. E. Mark, Polym. Bulletin, 20, 131 (1988).
G. S. Sur and J. E. Mark, Eur. Poly. J., 24, 913 (1988).
Polymr Handbook 3rd Ed., ed. by J. Brandrup and E. H. Immergut, Wiley-lnterscience,
New York, 1989.
J. Ibemesi, N. Gvozdic, M. Keumin, M. J. Lynch, and D. J. Meier, Polym. Preprints,
26(2), 18 (1985).
Z. Rigbi and J. E. Mark, J. Polym. Sci., Polym. Phys. Ed., 23, 1267 (1985).
G. B. Sohoni and J. E. Mark, J. Apol. Polym. Sci., 34, 2853 (1987).
S. Liu and 1. E. Mark, Polym. Bulletin, 18, 33 (1987).
G. S. Sur and J. E. Mark, Polym. Bulletin, 18, 369 (1987).
A. M. S. Al-ghamdi and J. E. Mark, Polym. Bulletin, 20, 537 (1988).
Perspectives in Molecular Sieve Science, ed. by W. H. Flank and T. E. Whyte, Jr., Am.
Chem. Soc., Washington, DC, 1988.
II
U
57
STRUCTURE OF MICROPHASE-SEPARATED SILICA/SILOXANE
MOLECULAR COMPOSITES
DALE W. SCHAEFER*, JAMES E. MARK**, DAVID McCARTHY*, LI JIANt,
C. -C. SUN* AND BELA FARAGOt
*Sandia National Laboratories, Albuquerque, NM 87185, USA
fDepartment of Chemistry and the Polymer Research Center, The University of Cincinnati,
Cincinnati, OH 45221, USA
tnsititut Laue-Langevin, 38042 Grenoble, France
ABSTRACT
The structure of severzi classes of silica/siloxane molecular composites is investigated
using small-angle x-ray and neutron scattering. These filled elastomers can be prepared through
different synthethic protocols leading to a range of fillers including particulates with both rough
and smooth surfaces, particulates with dispersed interfaces, and polymeric networks. We also
find examples of bicontinuous filler phases that we attribute to phase separation via spinodal de-
composition. Jn-shu kinetic studies of particulate fillers show that the precipitate does not develop by conventional nucleation-and-growth. We see no evidence of growth by ripening
whereby large particles grow by consumption of small particles. Rather, there appears to be a
limiting size set by the elastomer network itself. Phase separation develops by continuous
nucleation of particles and subsequent growth to the limiting size. We also briefly report studies
of polymer-toughened glasses. In this case, we find no obvious correlation between organic
content and structure.
INTRODUCTION
Historically, the development of specialty polymers has proceeded largely through the
manipulation of polymer chain architecture. Glassy vs rubbery behavior, for example, can be
adjusted with backbone stiffness. Strain-induced crystallization can be enhanced via stereoregularity. Flame retardancy is augmented by incorporation of chlorinated moieties. Silicon-based
systems provide enhanced high-temperature stability. In all these cases, the enhancement of a
targeted property usually implies the sacrifice of another. If backbone stiffness is increased to
raise the glass transition temperature, for example, toughness is bound to suffer.
Multicomponent systems that are homogeneous on length scales exceeding lpm offer
new promise to meet the competing requirements of high-performance polymers. The hope is
that by appropriate manipulation of phase structure, it will he possible to simultaneously enhance
multiple properties. Unfortunately, successful techniques for achieving these so-called molecular composites (MCs) are limited. In the absence of systematic relationships between synthetic
protocol, structure and properties, it is difficult to optimize these materials. Even for conventional composites prepared by mixing, for example, the properties (ramification, stiffness, interfacial properties etc.) of the ideal filler are not well established.
Two factors have limited the understanding of the microstructure of complex phase-separated materials: the absence of unambiguous methods for characterizing structure, and the lack of
reasonable models to predict structure based on chemical and physical parameters. Substantial
progress, however, has been made in both these areas due to advances in instrumentation, advances in the interpretation of scattering from disordered systems, and advances in development
of simple models describing complex disorderly growth processes. Here we attempt to build on
these advances to enhance our understanding of silica/siloxane MC's.
We focus on the structure of microphase-separated silica-filled siloxane elastomers I].
We study materials ranging from the unfilled siloxane rubber to polymer-toughened silicate
glasses. Based on insights from kinetic growth models and known results for solution polymerization of silicon alkoxides[2], we formulate synthetic recipes designed to generate fillers of
varying degrees of ramification. We infer relevant aspects of the filler structure from small-angle
x-ray (SAXS) and small-angle neutron scattering (SANS). We establish that filler structure does
indeed depend on synthetic protocol. Elsewhere we discuss the relationship between structure
and mechanical properties[3].
Met. Res. Soc. Symp. Proc. Vol. 171. c'UO Materials Research Society
Because of the importance of "structure" to our study, we first review the nature of smallangle scattering of neutrons (SANS) and x-rays (SAXS) from multiphase materials. To understand the relationship between structure, scattering and synthetic technique it is useful to consider
the distinction between equilibrium, and kinetic factors that influence morphology, and therefore
scattering behavior.
STRUCrURE AND SMALL-ANGLE SCATTERING
Even a brief review of the history of crystalline materials reveals that Bragg's discovery
of sharp x-ray diffraction lines played a crucial role in the understanding of the structure and
properties of crystalline solids. No less important were the advances in the interpretation the
broad x-ray diffraction features from disordered systems such as liquids and glasses. In both
cases the knowledge extracted from diffraction pertains to atomic and molecular length scales.
Molecular composites are typically disordered with "structure" occuring on length scales
that are large compared to atomic dimensions. Often this large-scale morphology is of dominant
importance in determining material properties. Diffraction (actually diffuse scattering) from these
materials is found at small angles (<10) and is often rather featureless. Interpretation of these
featureless profiles is the key to structure/property relationships in microphase-separated materials.
Recently, fractal geometry[4, 51 has emerged as the pertinent description for countless
random physical phenomena and disordered natural forms ranging from branched polymers to
geographic coastlines. Fractal geometry not only delivers a quantitative measure of disorder but
also provides insight into the origin of that disorder in terms of growth models[6]. In addition,
the featureless scattering curves (often power-law in form) described above fall to simple interpretation using fractal geometry.
Disordered fractal forms are often observed in systems that develop far from equilibrium
via kinetic growth processes. Order, by contrast, is the signature of equilibrium and results because the system has the opportunity to test many configurations and find the lowest energy state,
albeit in accordance with Boltzmann's law. An example of an equilibrium structure is a crystal
grown from solution where depositing atoms or molecules can move on the surface to find an
optimum solidification front. Kinetic growth, on the other hand, occurs when bonds, once
formed, do not dissociate so the system develops far from equilibrium. By near-equilibrium
growth we envision an intermediate
process where thermodynamic factors, like
.system
.equilibrium
S,"
z
1,_
JU
M0U
Nuceai
*d
approaches the lowest-energy
state.
tegrow nish
csuctypical
Figure astoi
1compares
scattering
e
ore
hich:uf
patterns expected for equilibrium, nearQ .w
vector
equilibrium and kinetic systems. The
B
.
-4
scattered intensity I(Q) is parameterized by
or scattering
wave
transfer
is
scattering,
for elastic
Q, which,
ovectorthe momentum
simply proportional to sin(0/2) with 0
being the scattering angle measured from
.....
Kiet
LATE
ANT
Q
FIG. 1.Scattering profiles for ordered equilibrium systems compared to those expected
from threegrowthprocesses.
the transmitted beam,
Here g(r) is the autocorrelation function of
scattering centers (e.g., electrons in the
I(Q)
Qualitatively
of the
x rays).
case
measures
strength of
the spatial Fourier
component with wave vector Q in the
spectrum of density or concentration
fluctuations. Thus, the sharp lines in Fig.
Fourier
to discrete
I-A correspond
reflections) and
(Bragg
components
59
represent the pattern seen for ordered,
equilibrium systems. If the growth process
is accelerated, say by a quench, then new
broad features are found in the scattering
profiles particularly at small angles
0
0
SAXS
+
-
a
0
corresponding to large distances (actually t)
exceedingly large compared to atomic
scales) equivalent to Bragg spacings of 100
A or more. Figure I-B, for example,
shows the development of a peak in the Z
J
+ BASE 9.9% 4.
scattering profile for growth following a
o ACID 24%
spinodal quench.
Fig. 1-C is the Z
-A
corresponding profile for nucleation-andgrowth, a process characteristic of
quenches into the metastable regime. Here
1
0.1
0.01
0.001
the limiting slope is -4, the signature of
smooth sharp interfaces and the scattering
profile develops in a characteristic manner
as indicated in the figure. For these nearFIG. 2. Scattering profiles for thick samples
processes,
growth
equilibrium
polymerized under acidic and basic condithermodynamic concepts like surface
tions by the two-step procedure.
tension are valid and provide an essential
understanding of the observed structures
and their time development[7, 8, 91.
A qualitatively distinct scattering profile is observed in kinetic systems that evolve far
from equilibrium. In this case featureless power-law scattering is observed (Fig. I-D), indicating
a power-law continuum of Fourier components[2, 5, 61. From the slope S of log-log plots of I
vs Q, one can distinguish smooth surfaces (S = -4) from fractally rough surfaces (S = -3 to -4)
from a branched or linear polymer (S = -1 to -3). Slopes less (i.e. steeper) than -4 indicate a
gradient in the concentration of the two phases at the interface. Detailed analysis of the shape of
the scattering curve for these systems yields the interfacial profile[10, 11].
REINFORCED ELASTOMERS
Based on the above ideas, we studied the structure of a series of silica-filled siloxane
elastomers[l]. Typically the elastomer was prepared from hydroxy-terminated polydimethylsiloxane (PDMS) of number-average molecular weight 18.0 x 103 end-linked with TEOS
(tetraethylorthosilicate), using stannous-2-ethylhexanoate as a catalyst. The weight ratios of
PDMS:TEOS:CATALYST were 100:0.58:0.3 These networks were extracted using tetrahydrofuran and dried.
The above networks were then filled by first swelling in pure TEOS and subsequently
placing swollen strips into acidic and basic water solutions prepared from acetic acid and ammonia at pH = 2.5 and 11. We call this a two-step procedure. Room temperature reactions were
carried out for 6 to 12 hours. The amount of filler was calculated from the weight gain.
Fig. 2 shows the scattering profiles for acid and base-catalyzed filler precipitated in 3 mm
thick PDMS elastomers. In Fig, 2 as in Figs 4 and 5, a Q-independent background has been
subtracted from the data. The limiting slopes of -4 (acid) and -4.5 (base) in Fig. 2 imply compact
particulate fillers with sharp and diffuse interfaces respectively. In some cases, we observe
slopes of exactly -4.0 for basic systems at low filler content, although the rule is that the slopes
are steeper indicating a gradient in the two phases at the interface. The base-catalyzed data were
fit to a model[ 11] that assumes a sigmoidal interfacial gradient between SiO 2 and PDMS yielding
an interfacial layer of 10A. For the acid-catalyzed system, the interface is sharp (limiting slope =
-4). For both systems, the surface area (SA) was calculated from the invariant and the magnitude
of the scattering in the limiting high Q regime[1 2]. Assuming independent particulate fillers, these
surface areas give particle radii Rp listed in Table 1. These radii match reasonably with those
obtained from Guinier anaysis(log I vs Q2) in the regime near 0.02A- t . For comparison we also
:,elude the average chord Rc for lines passing through a random two-phase system[131. The
similarity of these lengths supports the assumption of independent particles.
60
TABLE I: SURFACE AREAS AND LENGTH SCALES (A)
Rc is the average chord calculated from die surface area (SA) R is the purticle madius, calcu-
lated fron SA, Rg is the Guinier radius, and Rh is the hard sphere radius from Rg.
CATALYSIS
BASE
ACID
ACID
%SILICA
10
24
17
SA (m2/gm)
14
39
....
R,
118
71
8
53
81
57
...
56
Rh
112
74
---
The upturn in the data below o.OiA-1 signals the presence of long-range structure. In
general slopes near -4 are observed in this regime, consistent sharp interfaces with no evidence of
a limiting length-scale bel9w 1000A. To further elucidate the relationship between this largescale structure and the 1OA particulates, we studied the kinetics of the glassy phase development
by in-situ SANS during precipitation.
Fig. 3 shows the time development of the glassy phase for an ammonia vapor-catalyzed
precipitation. As can be seen in data taken 5 minutes after catalyst introduction, large scale
structure is present in the swollen networks before there is any evidence of particulates. The particulate form factor develops later and is superimposed on the large-scale structure. This largescale structure is presumably due to incompatibility between TEOS polymerization product and
PDMS which leads to microscopic phase separation very early in the polymerization.
It is interesting to note that the
developing particulates in Fig. 3 show slopes
-4.
of -3.3 consistent with a fractally rough
C
interface[14].
Rough particulates have
previously been observed by Keefer and
tW
Schaefer for base-catalyzed solutionOpolymerized
TEOS[15]. When the sample is
dried, a limiting slope of -4.4 is found
DRY
consistent with a graded interface. It should be
l'0 230 MIN
noted that the data for dried1 samples in the
A
regime 0.005 A-15 Q5O.1A t were acquired
Zi
+ 60 MIN
2
W
6
by SAXS whereas all the rest of the data were
MIN
by SANS. With one exception (Fig. 7),
-nevertheless, we have never observed limiting
between
-3 and
-4 for any dry samples
0.001
0.01
0.1 slopes
using either
SAXS
or SANS.
O(A -1)
Note that the time progression in Fig. 3
does not follow the nucleation-and-growth patFig. 3. In-situ neutron scattering data for
tern in Fig. I-C. In normal surface-tensionammonia vapor catalyzed precipitation. Dry
driven growth, large particles grow at the
sample was studied 3 months after the
expense of small. For in-situ precipitation,
kinetic runs.
however, growth is limited to about 1OA.
Later growth occurs by nucleation of new
particles leading to the time-development seen
in Fig. 3. Similar behavior is reported by Gilliom, Schaefer and Mark[ 161 for MCs produced
by catalytic hydrogenation in bulk polybutadiene.
Because of the possibility of effects due to leaching of the TEOS from the sample during
polymerization, we also studied thin (.3mm) samples in addition to the 3mm samples described
above which have a visible skin. For base-catalyzed systems, we found particulate fillers in all
cases, whereas for acid-catalyzed systems we occasionally found polymer-like fillers, particularly
at low loading and in thin samples. Fig. 4, for example, shows the SAXS data for an acid-catalyzed system with 17% filler. The limiting slope of -2.8 is consistent with a branched polymeric
filler, probably an interpenetrating network. In solution, alkoxides are known to produce polymer-like clusters in acid solution due to a growth mechanism called reaction-limited cluster
aggregation[2]. By contrast, polymerization in basic systems is believed to proceed by reactionlimited monomer-cluster growth which leads to compact particles. Of course, one of the motivations for our work was to exploit these different classes of growth to control the morphology of
filler phases. Clearly, however, the essential controlling factors for filled systems are not yet totally clear.
61
.-
2.
0
*
I
.
SAXS
0'
LU
LU
SANS
SA
""
-3.7
Z
--
W
ACID
0.001
17%
0.01
16% NS102
0.1
1
0.001
0
0.01
a(A- 1 )
0.11
a(A"- )
Fig. 4.SAXS scattering profile for a thin
sample polyrmrized under acidic conditions.
Liming
slope of -2.8 indicated a
non-particulate filler morphology.
Fig. 5. Scattering profile for silica filled
PDMS showing a peak characteristic of
spinodal decomposition.
Inafew samples, we observed(3] a peak in the scattering profiles (Fig. 5) consistent with
those reported by Wilkes et all 17] for "ceramers." We interpret the peak to be the remnant of
phase separation by spinodal decomposition. Since phase incompatibility usually increases during polymerization[ 18], we believe that crosslinking of the glassy phase induces phase separation
in the fluid precursor. The resulting spinodal morphology is then locked in place by "gelation" of
the glassy phase. The shape of the curve in Fig. 5 is very close to that found in simulations of
spinodal decomposition reported by ChakrabanietalPJ7.
It is interesting to note that the material in Fig. 5 is quite brittle showing a maximum extensibility of only 1.2. We believe that the brittle nature of the material as well as the peak in the
scattering profile indicate that the glassy phase is continuous. The spinodal process is known to
produce bicontinuous morphologies(9, 19].
POLYMER-TOUGHENED GLASSES
We also studied glassy sol-gel
silicates toughened by the incorporation of
PDMS. The samples studied here are those
previously reported by Mark and Sun[20].
These materials were characterized via the
D-hardness method. Mark and Sun found
that the D-hardness decreased smoothly
with increasing organic content as shown
in Fig. 6. The materials with high
organic/Si ratio have mechanical properties
similar to the reinforced elastomers
discussed above whereas those with D 50 are leathery.
The SAXS profiles of the
toughened glasses in Fig. 7 reflect a variety
of structures. For low silica-content
glasses (D = 2 and 18), there is evidence of
the formation of a spinodal-like peak, but
the limiting slope does not approach -4
indicating that the short-scale structure
should be viewed as an interpenetrating
network, rather than as a distinctly separate
70
60
U
-,50
Z
0 40
<
7- 30 0 20
0
0
0
10
0
0.9
1
1.1
1.2
ORGANICISI
1.3
1.4
Fig. 6. Hardness as a function of organic
content for polymer toughened silica glasses.
62
phase. At higher glass content, however,
phase separation is distinct (slope - -4) at
the 100A level. The SAXS profile for the
D
0
-4.0
e
0
material with the highest glass content
- (D =
60
6
05
o
50
60) is flat in the region near 0. 1 I This
'
plateau implies composition or density
K
x 18
o
0
fluctuations on very short scales (<20A) 2.
2
o
- beyond the limits of SAXS.
0
_1All of the toughened samples show
0
)
evidence of structure on scales exceeding
0
.
at
500A. The limiting slopes of about -4
small Q in Fig. 7 are consistent with large=
scale phase separation similar to that
discussed in connection with reinforced
o
SAXS
.
the
of
elastomers. To establish the structure
large-scale domains, we studied the two
extreme materials in Fig. 7 using the high
Q(A.1)
resolution D-I 1 SANS camera at the Institut
Laue-Langevin in Grenoble, France. The
Fig. 7. SAXS data for the series of
data, presented in Fig. 8, demonstrate that
polymer-toughened glasses shown in Fig. 6.
the large-scale domains are similar for the
two materials. Both data sets are consistent
with uniform domains with a Guinier radius
of 1500A. The flat background at large Q in Fig. 8 is due to incoherent scattering so data in this
regime cannot be directly compared with the SAXS data in Fig. 7.
We conclude that hardness for toughened glasses is determined largely by organic content. No obvious correlations exist between the hardness and observed scattering patterns.
Large-scale structure is independent of filler
whereas distinct changes in intermediate
structure (100k) are found as a function of
0
organic content.
D
-
e050
CONCLUSION
The
morphological
0 2
tendencies
g-
observed in silica-filled siloxanes are clearly
rb
dependent on synthetic protocol. Although
particulate fillers are the rule, we can
generate both polymeric fillers and
particulates with varied interfacial properties
ranging from fractally rough surfaces to
compositionally graded interfaces.
In
general, the degree of filler ramification
follows that previously observed in the
solution polymerization of alkoxides.
Although correlations exist between
structure and mechanical properties for filled
elastomers[3], no clear pattern is obvious
for the rubber toughened glasses.
Z
W
/
a
0
1500A
z
-.
0
-
D11
I
0.001
0(A-1)
0.01
Fig. 8. High-resolution SANS data for two
polymer-toughened glasses. The Guinier
radius is independent of hardness.
ACKNOWLEDGEMENT
Work performed at Sandia National Laboratories was supported by the U. S.
Department of Energy under contract DE-ACD4-76DP00789. SANS data were taken at the
Manuel Lujan, Jr. Neutron Scattering Center at Los Alamos National Laboratory and the SAXS
data were taken at the National Center for Small-Angle Scattering Research at Oak Ridge
National Laboratory. We thank Steve Spooner, Phil Seeger and Rex Hjelm for important contributions in collection of the SAXS and SANS data.
63
REFERENCES
*Permanent address: Department of Chemistry and the Polymer Research Center, The
University of Cincinnati, Cincinnati, OH 45221, USA.
1.
2.
3.
4.
5.
6.
7.
J.E. Mark and D. W. Schaefer in Polymer-Based Molecular Cm sites edited by D. W.
Schaefer and J. E. Mark (Mat. Res. Soc. Symp. Proc. 11 Pittsburgh, PA 1990).
D. W. Schaefer, Science 24, 1023 (1989).
D. W. Schaefer, J. E. Mark, L. Ran, C.-C. Sun, D. McCarthy, C.-Y. Jiang, Y.-P. Ning
and S. Spooner, in Ultrasmrcture Processing of Ceramics. Glasses and Conmosites, edited
by D. R. Uhlman, D. R.Ulrich and S. H. Risbut (J. Wiley, N. Y., 1990).
B. B. Mandelbrot, The Fractal Geometry of SAM (Freeman, San Francisco, 1982).
J. E. Martin and A. J. Hurd, J. Appl. Cryst. 20. 61 (1987).
D. W. Schaefer, Bull. Mat. Res. Soc. 13 (2), 22 (1988).
A. Chakrabarti, A. Toral, J. D. Gunton and M. Muthukumar, Phys. Rev. Lett. 63, 2071
(1989).
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
W. Hailer and P. B. Macedo, Physics Chem, Glasses 2 153 (1968).
J. W. Cahn, J. Chem. Phys. 42. 93 (1965).
W. Ruland, J. Appl. Cryst. 4, 70 (1971).
J. T. Koberstein, B. Morra and R. S. Stein, J. Appl. Cryst. .1, 34 (1980).
A. J. Hurd, D. W. Schaefer and A. M. Glines, J. Appl. Cryst. 21, 864 (1988).
A. Guinier and G. Fournet, Small-Angle Scattering of X-Rays (Available from University
Microfilms International, Ann Arbor, MI, 1955).
H. D. Bale and P. W. Schmidt, Phys. Rev. Lett. 53, 596 (1984).
K. D. Keefer and D. W. Schaefer, Phys. Rev. Lett. 56, 2376 (1986).
L. R. Gilliom, D. W. Schaefer and J. E. Mark in Polymer-Based Molecular Comnmsites
edited by D. W. Schaefer and J. E. Mark (Mat. Res. Symp. Proc. J1.L Pittsburgh, PA
1990).
G. L. Wilkes, A. B. Brennan, H. Huang, D. Rodrigues and B. Wang in Polymer Ba
Molecular Comoosites, edited by D. W. Schaefer and J. E. Mark (Mat. Res. Soc. Symp.
Proc. 11L Pittsburg, PA 1990).
P. G. deGennes, Scaling Concents in Polymer Physics (Cornell University Press, Ithaca,
NY, 1979).
N. F. Berk, Phys. Rev. Lett. 2. 2718 (1987).
J. E. Mark and C.-C. Sun, Polym. Bull. 11, 259 (1987).
66
NMR IMAGING OF SILICA-SILICONE COMPOSITES
LEONCIO GARRIDO-, JEROME L. ACKERMAN- AND JAMES E. MARK"
* NMR Center, Massachusetts General Hospital, 149 13th St., Charlestown, MA 02129.
Department of Chemistry and the Polymer Research Center, University of Cincinnati,
Cincinnati, OH 45221.
ABSTRACT
Polydimethylsiloxane (PDMS) model networks reinforced by in aitu precipitated
SiO2, and polymer-modified silica glasses were obtained following the usual sol-gel methods. The conditions were chosen to increase the probability of observing inhomogeneities:
(i)bulky samples, and (ii) limited reaction times. These composites were characterized
by measuring bulk spin-lattice (T 1 ) and spin-spin (T 2 ) relaxation times and using 'H
NMR two-dimensional Fourier transform (2DFT) spin echo imaging techniques. The T1
and T 2 maps show clear and significant variations of NMR signal intensity throughout
the sample due to nonuniform hydrolysis of the tetraethylorthosilicate (TEOS) in the
specimens.
INTRODUCTION
The search for materials with optimal properties for specific applications, including
new pathways for their synthesis and processing, is a continuing process. Over the past
20 years, sol-gel processes have been extensively studied as alternatives to the existing
preparation methods for composites [1,2). The ability to manipulate the composite microstructure by the sol-gel reaction results in materials with improved and sometimes
unexpected physical and chemical properties. The understanding of a material's behavior requires knowledge of how the properties of interest depend upon its chemical composition and molecular structure. Nuclear magnetic resonance (NMR) spectroscopy is
very sensitive to both the chemical composition and the structure of a substance. These
material properties are reflected in the chemical shift spectrum as well as in the NMR
relaxation parameters. NMR imaging techniques [3-111, by producing visual pictures
of the spatial variation of selected NMR properties, offer the possibility of selectively
mapping the distribution of particular chemical species in a region of interest. Moreover,
NMR imaging can also provide spatial information about changes in NMR properties
that can be correlated with alterations in molecular structure and dynamics.
The aim of this work is to develop NMR imaging techniques for the characterization ,f
sol-gel prepared organic-inorganic composites by mapping the organic phase distribution
and the degree of alkoxide hydrolysis. We have obtained 'H NMR images of intentionally heterogeneous polydimethylsiloxane (PDMS) model networks reinforced by in situ
precipitated silica (SiO2) and polymer-modified silica glasses. In these images the variations in NMR signal intensity between different regions in the sample (image contrast)
are a function of proton density, spin-lattice (TI) and spin-spin (TI) relaxation times.
Such maps of NMR parameters provide a measure of the molecular mobility, which can
in turn be related to the spatial variation of the relationship between the organic and
inorganic phases throughout the specimen.
Mat. R@. lift. Symp. Proc. Vol. 171. *11110Materials Ressarch Soceoty
66
EXPERIMENTAL
Preparation of reinforced PDMS model networks.
PDMS model networks were prepared by end-linking reaction of dihydroxyl-terminated PDMS chains having a number average molecular weight, M., of 4,200 g mol'
with tetraethylorthosilicate (TEOS) in the usual manner (121. The networks obtained,
cyclindrical pieces 20 mm in diameter and 9 mm in height, were swollen at equilibrium
in TEOS (which correspond to a volume fraction of polymer of 0.70). Each swollen
sample was then immersed in aqueous solution of CF 3 COOH at 5 percent w/w for 15
to 120 min. The acidic catalyst was chosen because of its high efficiency to hydrolyze
TEOS [131. The samples were dried under vacuum to constant weight. The increase in
dry weight gave the amount of Si0 2 precipitated within the sample in the elapsed time
(see Table I). The large sample size and the short hydrolysis time assure inhomogeneous
specimens.
Preparation of polymer-modified silica glasses.
I
The polymer-modified silica glasses studied in this work were prepared as described
elsewhere [14].Briefly, the functionality of divinyl-terminated PDMS was greatly increased by a substitution reaction to give PDMS with triethosilyl chain ends [151. Samples having M. of 720 and 17,600 g mol', and mixtures thereof, were added to TEOS
or related silane. The functionalized PDMS/silane mixtures were hydrolyzed in aqueous
solutions of acetic acid following the usual sol-gel procedures.
Instrumentation and techniques.
All NMR measurements were performed in a Bruker MSL 400 spectrometer/imager
equipped with an Oxford 9.4 T (proton frequency at 400.13 MHz) 8.9 cm vertical bore
superconducting magnet. The RF coils used are saddle type with diameters ranging
from 10 to 30 mm and their longitudinal axis parallel to the static magnetic field. The
pulsed gradient amplitudes in the imaging experiments varied between 4 and 40 G cm'.
Bulk T, and T2 measurements were carried out using inversion recovery and spin
echo sequences, respectively. The results are shown in Table I. The inversion time in the
IR sequence was varied between 0 and 10 s. The echo time (TE, time between the 900
RF pulse and the center of the echo) in the spin echo sequence ranged from 0.2 to 300
ms. The repetition time, TR, was 10 s in both cases, more than five times T1.
'H NMR images of the reinforced networks and modified glasses were obtained using
two-dimensional Fourier transform (2DFT) spin echo techniques with TEs on the order
of 3 to 20 ms. The selective excitation of a slice throughout the sample 500 jm thick
was achieved with a - ms wide sinc-function amplitude modulated RF pulse. The pulse
sequence TR was typically 3s. The total imaging time varied between 25.6 and 76.8 min.
The digital resolution was 128X by 128Y pixels ranging from 65 to 200 ,m in both axes.
67
Table I: Amount of silica, T, and T 2 values of reinforced
PDMS model networks
Sample
ref.
1
2
3
4
5
6
Hydrolysis
SiO,
time (mmin)_ (%.w/w)
120
90
60
45
30
15
4.7
-
4.6
4.2
2.4
1.9
T,
T 2.*
(a)
(ms)
1.26
1.19
1.21
1.23
1.24
0.86
0.86
0.81
0.82
0.84
0.79
f.'
T2('
(ms)
0.968
0.965
0.970
0.966
0.967
0.962
106.3
68.6
74.4
51.6
113.7
220.7
Spin-spin relaxation time of the major component.
Fraction of major component contributing to the NMR signal (f. + f, = 1).
Spin-spin relaxation time of the minor component.
RESULTS AND DISCUSSION
The T, data was analyzed assuming the presence of only one component (a monoexponential function). The agreement found between the theoretical curves and the
experimental results was very good. As shown i.. Table I, T1 does not depend significantly on the time allowed for the hydrolysis of TEOS (increasing amount of Si02). The
NMR spin-lattice relaxation mechanism is apparently not affected by the presence of
silica and the T, values obtained are similar to those of unfilled PDMS model networks
with molecular weight between crosslinks, M,, between 3,700 and 18,000 g mol - 116).
The analysis of the T 2 data is more complex. An apparent two component model
gives the best fit to the experimental results. The values thus obtained are shown in
Table I. This may be interpreted ts the result of two contributions to the NMR signal, one
from a major fraction of material attributed to the PDMS chains forming the network
and the other, most likely, from the ethyl groups in partially hydrolyzed TEOS. The
T 2 values of the major component agree very well with the results obtained for unfilled
PDMS model networks 116). The T2 of the minor component (3 to 4 percent) varies
with the hydrolysis time of TEOS. It is known that TEOS tent to form polymeric chains
when the hydrolysis reaction is catalyzed by acids 1171. Therefore, it is possible to have
TEOS polymer with ethoxyl and hydroxyl side groups in various proportions. Their
relative contribution to the NMR signal is uncertain at this point and further ongoing
experiments might clarify it.
The complexity of the process is clearly manifested by the 'Il NNIR imaging experiments. Figure 1 shows two images, along its longitudinal axis, of a PDMS model
network (sample # 6 in Table 1) obtained with a two dimensional spin echo sequence
having TE of 3.3 (A) and 22.7 (B) ms. The rest of the experimental conditions were the
same for both images. The heterogeneity of the sample is readily apparent. Thl! dark
rim around the sample in Figure 1 A may indicate the reduced mobility of the network
chains in this region compared to that in the sample core, probably because of a high
concentration of SiO 2 . The availability of water and catalyst in the sample periphery
might be the reason for the observed difference. At long TE the PDMS chains do not
contribute to the echo 116) and only material with high molecular mobility, i.e. oligomers
68
Figure 1. 'H NMR images of the sample # 6 obtained with a two dimensional spin
echo sequence having a TE of 3.3 (A) and 22.7 (B) ms. TR was 3 s in both
cases. Selective excitation was used to define a 500 pm slice thickness in the
sample axial plane. The resolution is 128X by 128Y pixels of 180 and 200
pm, respectively. The time required to acquire tho data for each image was
25.6 min. The dark rim at the edge of the sample may indicate a reduced
molecular mobility of the network chains due to the presence of SiO 2.
from partial TEOS hydrolysis, is visible to the NMR experiment. A significant change
in the NMR images with short TE is not observed with increasing hydrolysis time.
However, T2rweighted images (long TE) show some variations in signal intensity. Figure
2 shows two images (A and B) of sample # 4 obtained with the same conditions as
Figures 1 A and B, respectively. As mentioned above, CF 3 COOH is a very efficient
catalyst in promoting the TEOS hydrolysis. Therefore, chanhes at the edges of the
sample will occur relatively fast while at the center those chan es are controlled
the
Figure 2. Axial 'H NMR images of sample # 3. Both images, Figures A and B,
were taken with the same experimental conditions as for Figures 1 A and B,
respectively. A significant variation of NMR signal intensity in Figure 2 B
(TI.weighted image) is observed when compared to that of Figure I B.
a
I.
69
Figure 3. On the left (A) is shown the 1H NMR image of a modified silica glass
which corresponds approximately to the cross-sectional photograph displayed
on the right (B). A pulse sequence similar to that described in the previous
figures was used. TE was changed to 3.0 ms. The resolution is 128X by 128Y
pixels of 110 #tm in both axes. The total imaging time was 76.8 mrn.
diffusion rates of the water and catalyst in the swollen rubber.
As a part of the characterization of silica-silicone composites by NMR imaging, we
are currently investigating polymer-modified silica glasses. Figure 3 B shows a photograph of the sample section of a modified glass which corresponds approximately to the
NMR image shown in Figure 3 A. There is a correlation between the visual appearence
and the NMR experiment. Heterogeneities are visible in the latter but, as in the case
of
reinforced PDMS rubbers, more detailed understanding requires further studies.
CONCLUSIONS
Our results demonstrate that proton magnetic resonance imaging is capable of revealing inhomogeneities in in situ precipitated Si0 2-filled PDMS networks and in polymermodified silica glasses, at least when relatively short reaction times and thick sections
are employed. The contrast between different regions is highlighted with T -weighted
2
imaging pulse sequences, suggesting that the underlying variations seen in the
images
may be closely related to variations in the mobility of the network segments. In the
case of the filled networks there is a clear, if nonmonotonic, progression of spin-spin
relaxation time values of the long T2 component with hydrolysis time; T, values are
constant. We suggest a variation in ethyl group concentration in partially hydrolyzed
TEOS as a possible source of this long T 2 signal.
Additional investigations into the molecular naturc of the spatial inhomogeneities
seen in the composite networks and glasses are required for a more complete understanding of these materials. Destructive analysis such as electron microscopy or local
measurements of specific gravity could be used to validate these NMR results, as well as
iid in further elucidation of the origin of the T2 variations. We are currently pursuing
.hese and other avenues. However, it is clear that NMR imaging will be extremely use-
lN•. I__ --m
A
Immi
m
immml...
m
mm
~
70
ful as a nondestructive tool for monitoring the uniformity of these and other synthetic
materials and for learning about the details of spatially varying chemistry.
ACKNOWLEDGEMENTS
This work was supported in part by the MGH NMR Center, NIH grant RR03264,
and NSF grant DMR 84-15082.
REFERENCES
1. Ultrastructure Processing of Advanced Ceramics, edited by J. D. Mackenzie and
D. R. Ulrich (John Wiley & Sons Inc., New York, 1988).
2. Better Ceramics Through Chemistry III, edited by C. J. Brinker, D. E. Clark and
D. R. Ulrich (Mater. Res. Soc. Proc., 121, Pittsburg, PA 1988).
3. P. Mansfield and P. K. Grannell, Phys. Rev. B 1-2, 3618 (1975).
4. A. N. Garroway, J. Baum, M. G. Munowitz and A. Pines, J. Magn. Reson. LO,
337 (1984).
5. N. M. Szeverenyi and G. Maciel, J. Magn. Reson. 60, 460 (1984).
6. F. De Luca and B. Maraviglia, J. Magn. Reson. 67, 169 (1986).
7. C. G. Chingas, J. B. Miller and A. N. Garroway, J. Magn. Reson. 66, 530 (1986).
8. W. A. Ellingson, J. L. Ackerman, J. D. Weyand, R. A. DiMilia and L. Garrido,
Ceram. Eng. Sci. Proc. a, 503 (1987).
9. L. Garrido, J. L. Ackerman, W. A. Ellingson and J. D. Weyand, Ceram. Eng. Sci.
Proc. 2, 1465 (1988).
10. D. G. Cory, J. C. de Boer and W. S. Veeman, Macromolecules 22, 1618 (1989).
11. J. B. Miller and A. N. Garroway, J. Magn. Reson. a2, 529 (1989).
12. J. E. Mark and J. L. Sullivan, J. Chem. Phys. 6,
1006 (1977).
13. C.-Y. Jiang and J. E. Mark, Makromol. Chem. 1U, 2609 (1984).
14. J. E. Mark and C.-C. Sun, Polym. Bull. 13, 259 (1987).
15. G. S. Sur and J. E. Mark, Eur. Polym. J. 21, 1051 (198).
16. L. Garrido, unpublished results.
17. K. D. Keefer, in Better Ceramics Through Chemistry, edited by C. J. Brinker, D.
E. Clark and D. R. Ulrich (Mater. Res. Soc. Proc., 32, Elsevier, New York, 1984),
p. 15-24.
71
SYNTHETIC POLYMERS IN WATER-IN-OIL MICROEMULSIONS
FRANCOISE CANDAU
Institut Charles
Sadron
Strasbourg Cedex, France
(CRM-EAHP),CNRS-ULP
6,
rue
Boussingault
67083
ABSTRACT
High molecular weight water-soluble polymers are usually supplied in the
form of water-in-oil emulsions which have advantages of low viscosity and easy
storage and dissolution. Most uses in water treatment, flocculation, paper
manufacture or mining fields require polymer latexes formed of finely dipersed
particles. Polymerization in reverse micelles or microemulsions appears to be
an attractive technique because it can lead, under appropriate formulations,
to
high molecular weight polymers entrapped within small-sized stable
particles. The main characteristics and properties of the latexes and polymers
formed by this process are described.
INTRODUCTION
Polymerization of water-soluble monomers in hydrocarbon fluids has
attracted a renewed interest over the past decade, owing to the suitability of
the process for producing high molecular weight polymers at high reaction
rates. In this process, a water-soluble monomer (usually in aqueous solution)
is emulsified in a continuous oil medium using a water-in-oil emulsifier.
Polymerization can be initiated with either oil -or -water-soluble initiators.
The product is a dispersion of fine particles of an aqueous high polymer
solution which can be easily inverted into water so that the water-swollen
polymer particles dissclve rapidly, contrary to solid-powder which forms gels
or aggregates when added to water. These high viscosity polymer solutions find
applications
in water-treatment, flocculation of colloidal suspensions,
tertiary oil
recovery as
pushing fluids,
fines retention
in paper
manufacturing etc [1].
Inverse emulsion polymerization has been far less
investigated than conventional (i.e. aqueous) emulsion polymerization. Apart
from a large number of patents, there have been few fundamental studies. The
pioneering work is due to Vanderhoff et al
[2] who studied in 1962 the
polymerization of sodium-p-vinyl benzene sulphonate in xylene. More recently,
polymerization processes
in different inverse
colloidal systems,
i.e.
suspension [31,
microsuspension [4,5], dispersion 16],
and emulsion J7-12]
have been described in the literature.
In general,
the problems of latex stability are more severe than for
aqueous latexes, due to the absence of electrostatic forces between particles,
and to the large difference in density between the organic continuous medium
and the polymer core. This led us in the past years to investigate the
possibilities
offered by
a polymerization
reaction proceeding
in a
thermodynamically stable microemulsion
rather
than
in
an
emulsion.
Microemulsions are water-oil colloidal
dispersions
stabilized
by
an
appropriate mixture of surface-active agents. While inverse emulsions are
unstable, turbid and consist basically of two populations with a broad
particle size distribution (fig.l ), microemulsions are thermodynamically
stable and can adopt a large variety of labile structural organizations. The
small size of the domains explains their optical transparency [13].
In the
oil-rich regions, they are formed of water-swollen spherical droplets of
uniform and small size (d - 6 nm) dispersed in the oil medium (fig.2a). When
the amount of the aqueous phase tends to be of the same order of magnitude as
that of the organic phase, the description generally given is that of a
bicontinuous structure formed of randomly interconnected oily and aqueous
domains [14,15], with the surfactant molecules located at the interface
(fig.2b). It should be noted that the formation of a microemulsion requires a
Mat. Poe. Soc. Symp. Proc. Vol. 171. e1SO Materials Research Society
4
72
Oit
monomer- potmer
partice
RR--2R"
0
C"°
,,0
M
-
-0
0
+
,,o H2 0
..
9,
0
d-(200-5000A)
micetle d-(0-100
Figure 1
Schematic representation of an inverse emulsion
Q)
b)
,H 0
Oiloil
iigure 2 : Schematic representation of a microemulsion
a) inverse globular structure
b) bicontinuous structure
minimum amount of surfactant of around 10%, imposed by the surface coverage of
the small domains while somewhat smaller amounts are sufficient for an
emulsion.
The microemulsion polymerization of various water-soluble monomers,
mostly used in industrial applications, has been investigated. In this paper,
we report the main characteristics and properties of the latexes and polymers
formed by this process.
STRUCTURE OF OIL/(WATER+MONOMER)/SURFACTANT MICROEMULSIONS
The choice of the system to be polymerized is of critical importance,
since it controls the properties of the resultant latexes. Water-soluble
monomers such as acrilamide (AM), were polymerized either inside water-swollen
micelles (d - 6.10'
pm) stabilized with an ionic surfactant, aerosol OT,
(sodium 1,4-bis
(2
ethylhexyl)sulfosuccinate)
(16]
or
in
nonionic
microemulsions
(17-19]. The latter systems have proven to be the most
effective, as they can incorporate up to 25% monomer still remaining stable.
The addition of monomer to the microemulsion produces a considerable
extension of the microemulsion domain in the phase diagram. This was
interpreted as a cosurfactant effect of the monomer which preferentially
73
locates at the w/o interface (16,18,191. The presence of monomer molecules
between the surfactant molecules increases both the flexibility and the
fluidity of the interface, resulting in a change of its curvature. Eventually,
when the radius of curvature becomes very large the globular configuration
converts into a bicontinuous structure. The latter structure is favoured by
addition of monomers producing a salting-out of the nonionic ethoxylated
surfactants. This is the case
of
sodium
acrylate
(NaA)
[18]
or
methacryloyloxyethyltrimethylammonium chloride (MADQUAT) [20].
In the case of AOT systems, there is no clear evidence of a globular
bi-continuous transition. However, under certain conditions, addition of
acrylamide induces sharp rises in conductivity, indicative of the formation of
transient conducting water channels from particle to particle which were
attributed to a percolation process (21].
A typcal polymerization recipe is the following : cyclohexane : 37.5%
MADQUAT : 25% ; water : 25% ; nonionic surfactants (Arlacel 83 + Tween 80
HLB - 12.9) : 12.5%.
POLYMERS PREPARED IN MICROEMULSIONS
The control of molecular weight of the final polymer is of paramount
importance since most applications require ultra-high molecular masses. The
polymer molecular weight has been determined from viscometry experiments in
aqueous solutions, using the Mark-Houwink relationship [16,22], or from static
light scattering [23]. Various parameters control the molecular weight. The
production of high molecular weights (106-107) is achieved by polymerization
at lower temperature and at high (monomer)/surfactant levels (fig.3) [22,23].
The properties of poly(acrylamide-co-acrylates) prepared in nonionic
microemulsions have been studied by means of several techniques (24,25]. The
free radical copolymerization of acrylamide with ionogenic monomers is
13
strongly influenced by the microenvironment. A
C NMR study performed on
poly(acrylamide-co-acrylates) prepared in
microemulsions
confirmed
the
influence of the reaction medium. The average copolymer composition was found
to be independent of the degree of conversion, also the sequence monomer
distribution analyzed
from triad proportions,
conforms to Bernouillan
statistics. The reactivity ratios of both monomers are therefore close to
unity, contrary to the literature values reported for copolymers prepared in
solution (rA - 0.3, r - 1) [26]. As the local monomer concentration is much
higher in a microemulsion (- 5 M) than in a solution (- I M), this could
produce an increased screening of the carboxylate groups by sodium ions. The
observed increase in the reactivity parameter of sodium acrylate in the
microemulsion (rA - 1) is consistent with this expectation.
The copolymers exhibit high intrinsic viscosities in aqueous solutions
3
(up to 3700 cm /g) with maximum around 40 mol% acrylate content due to a
superimposition of electrostatic effects and of intramolecular bonds.
STRUCTURAL PROPERTIES OF LATEXES
The
inverse
latexes
formed
highly
stable with no
apparent
after polymerization are clear, fluid and
settling.
A common feature, found in all
experiments, is that the final system always consists of an uniform dispersion
of spherical latex particles, regardless of the structure of the starting
system (globular or bicontinuous). This can be accounted for by the internal
dynamics of microsmulsions which are constantly rearranging on the time-scale
of microseconds. The size of the latex particles has been determined by quasielastic light scattering (QELS) and electron microscopy (EM) experiments. Both
techniques require an examination with dilute samples, which, in turn, implies
that the structure and the particle size are not affected by the dilution
process. One thousandfold dilution and the drying process used in EM for the
water-swollen polymer particles dispersed in oil can introduce artefacts. A
74
10-
10/
10
/
(MADQUAT)/( S)
I
I1
1.0
0
Figure 3
2.0
Variation of the molecular weight of polyMADQUAT with the MADQUAT
over surfactant ratio (wt/wt) in the initial microemulsion (from
re f.23).
dhnnm)
100
20 0
(MADOLUAT)/(S)
1.0
2.0
Figure 4 : Variation
of
the
hydrodynamic
diameter
of the latex particle
with
the
MADQUAT
over
surfactant
ratio
(wt/wt) In the initial
*icroemulsion (from ref.23).
75
flattening of the particles has indeed been observed in some cases (211.
In the case of AOT micellar systems, one observes a notable increase of
the particle size during the polymerization so that each final latex particle
is the result of the fusion of around a hundred initial micelles (161. The
particle size ranges typically between 30 nm and 120 nm depending upon the
experimental conditions;
it augments upon increasing monomer content or
decreasing surfactant concentration (fig.4) [22,231.
From the value of the average diameter of the dry polymer particle and
that of the polymer molecular weight, one can estimate the number of
macromolecules, Np, contained in a particle. Calculations give extremely low
values of N p, around I or 2, in contrast with what is usually found in
emulsion polymerization, where thousands of chains are commonly observed. This
result was supported by a thorough mechanistic study performed on AOT systems
121,27,281
where a continuous particle nucleation was shown to occur all
throughout the reaction and not only at the very early stages as is the case
in emulsion polymerization. The high molecular weight polymeric chain must be
strongly collapsed since it is entrapped within a water-swollen particle of
small size (d < 100 nm). For comparison, the radius of gyration of a 5.106
molecular weight polyMADQUAT is around 160 rm. The water in the particle core
acts here more as a plasticiser than as a solvent for the polymer.
These inverse microlatexes provide interesting models for rheological
studies [29].
They differ from the more conventional aqueous or nonaqueous
colloidal dispersions by a lower particle size and by a large swelling of the
particles (e.g. 50% water, 50% polymer). In addition, the process permits to
attain large volume fractions of the disperse phase (up to 60%) and thus to
accurately characterize the latex rheological behavior.
Figure 5 shows an example of the variation of the relative viscosity
with the shear stress for a typical latex. Qualitatively, one has the usual
behavior observed for colloidal dispersions, that is a Newtonian behavior at
low volume fractions and a shear-thinning effect for the largest volume
fractions investigated.
However, the shear-thinning effect occurs at volume fractions much higher
than those observed for conventional latexes (50-53% as against 25%). A
possible explanation is deformability of the microspheres due to the low
interfacial tension and to the high water content of the dispersed phase. The
full lines in Figure 5 represent the best fit of the well-known Williamson
) / (1
1
equation [301 (q - q . + (n
- qr+ o/ao) where 'r0 and Pr. are the low
shear and high sbear iimiting relative viscosities and a. the shear stress for
2
which r - (nr0 + nr.)/ )l to the experimental data.
We have determined the intrinsic viscosity [q] and the close-packing
volume fraction (p of a series of copolymer latexes containing various
contents of sodium acrylate, by using the Krieger-Dougherty equation [31]
nr " (1 -
/
)
A value of - 2.5 is found for (q] in good agreement with was is obtained for
hard spheres suspensions. The data shows that v decreases when increasing the
percentage of electrically charged sodium acrylate in the comonomer feed. For
instance, 4P is around 59% for a sodium acrylate concentration of 12%. This
value is significantly lower than that of 64% predicted
for a
random packing
of hard spheres. On the other hand, other experiments performed on neutral
polyacrylamide latexes have given values of V around 64%. The divergence of
viscosity
which
occurs
at
smaller
characteristic
volume
fractions
for
poly(acrylamide-co-acrylates)
latexes is probably due to the effect of
electrical charges in the system, which prevent particles from approaching too
closely to each other. Rheological studies on other ionogenic monomers are in
progress to check this hypothesis.
76
-- 4
so
40
30
F
20
0
10
£A
101
AM.
Al
MA
101
100
46
AA A A&A
104
103
102
Or (dynes/cm 2 )
Figure 5 : Variation of
the relative viscosity
with shear stress for a
poly(acrylamide-co-sodium acrylate) latex containing 9.10% sodium
acrylate (wt/wt) and with a particle size of 62 nm [29].
a : q-
0.35 ;
-
0.44 ;
: V-
0.48
; A
:
- 0.52
;
-
0.55.
APPLICATIONS OF MICROLATEXES AND POLYMERS
Polymerization
in inverse microemulsions overcomes
some
of the
problems classically encountered in emulsions and provides a novel technique
for the production of high molecular weight water-soluble polymers. The main
advantages are the low viscosity and high stability of the microlatex formed.
The small particle size prevents flocculation since gravity forces are
reduced. The large number of micelles contained in microemulsions compared to
that in
emulsions contributes to
the formation of
high degrees of
polymerization (DP a N). Also, the microlatex is self-inverting so that no
additional surfactant is needed to promote its inversion. With respect to the
economical aspect, the main drawback of the process is the rather expensive
formulation (high surfactant concentration). This is partially balanced by the
very high rate of polymerization due to the great number of micelles, loci of
the polymerization. Total conversions are usually achieved in a few minutes
compared with hours in the usual process.
Most of the applications described for water-soluble polymers prepared in
water-oil emulsions can in principle be extended to the inverse microemulsion
polymerization process and several patents have recently been issued [32-35].
For example, inverse microlatexes can be used after dilution to water to form
thickened solutions for improving the production of oil fields. They have
advantages with respect to conventional latexes, as a result of their lover
particle size, their lower degree of polydispersity and their great stability.
They result in a better scavenging of the oil formation and thus in a more
efficient oil recovery.
77
Most uses in the paper manufacture, water treatment and mining fields
are based on the ability of water-soluble polymers such as polyacrylamides to
flocculate solids in aqueous suspensions [1].
Small mineral or pigment
particles settle very slowly and are difficult to eliminate or recover.
Addition of charged polyacrylamides permits them to agglomerate. The ultrahigh molecular weight polymers produced in the microemulsion process can be
• ery effective in connecting together the small particles through bridging or
charge
neutralization. Moreover, classical polyacrylamide emulsions are
subjected to rapid changes in temperature (in winter time) which cause them to
rather than remain finely dispersed particles. This reduces
coagulate
drastically their usefulness as flocculants. Inverse microlatexes exhibit
excellent freeze-thaw properties and contain finely divided polymer particles
(- one polymer chain in a low size particle) which should insure a higher
activity.
Other applications include surface coatings, adhesives, photographic
emulsions, lubricating and cleaning drains, retention aid in paper making and
food processing. Finally, the low viscosity and good stability of microlatexes
can be useful for assembling glass fibers.
REFERENCES
1.
W.M. Thomas and D.W. Wang, in Encvclooedia of Polymer Science and
Enzineerina, edited by H. Mark,
N. Bikales, C.G. Overberger and G.
169
Menges, 2nd ed, New York, 1985, p.
.
2. J.W. Vanderhoff, H.L. Tarkowski, J.B. Shaffer, E.B. Bradford and R.M.
Wiley, Adv. Chem. Ser. )A, 32 (1962).
3. M.V. Dimonie, G.M. Boghina, N.N. Marinescu, M.M. Marinescu, C.I. Cincu
and C.G. Oprescu, Eur. Polym. J. 18, 639 (1982).
4. D. Hunkeler, A.E. Hamielec and W. Beade, Polymer 30, 127 (1989).
5. W. Beade, D. Hunkeler and A.E. Hamielec, Am. Chem. Soc. Div. PMSE
'reprints, 5., 850 (1987).
6. W. Baade and K.H. Reichert, Eur. Polym. J. 20, 505 (1984).
7. V.F. Kurenkov, T.M. Osipova, E.V., Kuznetsov and V.A. Myagchenkov,
Vysokomol. Soedmn. Ser. 20, 647 (1978).
8. C. Graillat, C. Pichot, A. Guyot and M.S. EI-Aasser, J. Polym. Sci.
Polym. Chem. Ed. 2A, 427 (1986).
9. V. Glukhikh, C. Craillat and C. Pichot, J. Polym. Sci. Polym. Chem. Ed.
25, 1127 (1987).
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
J.W. Vanderhoff, F.V. Distefano, M.S. El-Aasser, R. O'Leary, O.M. Shaffer
and D.L. Visioli, J. Disp. Sci. Tech. , (3&4) 323 (1984).
D.L. Visioli, PhD thesis, Lehigh University, 1984.
M.T. McKechnie, Proceedings of the Conference on Emulsion Polymers,
London, Psper 3/1 (1982).
See for example
: A.M. Bellocq, J. Biais, P. Bothorel, B. Clin, G.
Fourche, P. Lalanne, B. Lemaire, B. Lemanceau and D. Roux, Adv. Colloid
Interface Sci. 2, 167 (1984).
L.E. Scriven, Nature (London), 26, 123 (1976).
S. Friberg, I. Lapczynska and G. Gillberg, J. Colloid Interface Sci. 56,
19 (1976).
F. Candau, Y.S. Leong, G. Pouyet and S.J. Candau, J. Colloid Interface
Sci. 161, 167 (1984).
C. Holtzscherer and F. Candau, Colloids Surf. 29, 411 (1Q88).
F. Candau, Z. Zekhnini and J.P. Durand, J. Colloid Interface Sci. ]4,
398 (1986).
C. Holtzscherer and F. Candau, J. Colloid Interface Sci. 12,
No.l. 97
(1988).
P. Buchert and F. Candau, J. Colloid Interface Sci. (in press).
M.T. Carver, E. Hirsch, J.C. Wittmann, R.N. Fitch and F. Candau, J. Phys.
Chem. 21, 4867 (1989).
78
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
C. Holtzscherer, J.P. Durand and F. Candau, Colloid Polym. Sci. W,
1067
(1987).
P. Buchert, PhD thesis, Louis Pasteur University, Strasbourg, 1988.
F. Candau, Z. Zekhnini and F. Heatley, Macromolecules 19, 1895 (1986).
F. Candau, Z. Zekhnini, F. Heatley and E. Franta, Colloid Polym. Sci.
264, 676 (1986).
S. Ponratnam and S.L. Kapur, Kakromol. Chem. la, 1029 (1977).
M.T. Carver, U. Dreyer, R. Knoesel, F. Candau and R.M. Fitch, J. Polym.
Sci. Polym. Chem. Z, 2167 (1989).
M.T. Carver, F. Candau and R.M. Fitch, J. Polym. Sci. Polym. Chem. 2Z,
2179 (1989).
D. Collin, F. Kern and F. Candau, presented at the 3rd European Colloid
and Interface Society Conference, Basel (Switzerland), 1989 (to be
published).
R.V. Williamson, J. Rheol. 1, 283 (1930).
I.M. Krieger and T.J. Dougherty, Trans. Soc. Rheol. III 137 (1959).
F. Candau, Y.S. Leong, N. Kohler and F. Dawans, French Patent (to C.NRSIFP) No. 2 524 895 (1984).
J.P. Durand, D. Nicolas, N. Kohler, F. Dawans and F. Candau, French
Patent (to IFP) No 2 565 623 and 2 565 592 (1987).
J.P. Durand, D. Nicolas and F. Candau, French Patent (to IFP) No 2 567
525 (1987).
F. Candau and P. Buchert, French Patent (to Soc. Chim. Charb.) No 87
08925 (1987).
79
POLYMER-DERIVED Si3N4/BN COMPOSITES
Wayde R. Schmidtl . William J. Hurley, Jrl., Vijay Sukumarl , Robert H. Doremus 1,
2
Engineering
and
and Leonard V. lnterrante , Departments of I Materials
2
Chemistry, Rensselaer Polytechnic Institute, Troy, NY
12180-3590.
ABSTRACT
Partially crystalline silicon nitride, with a specific surface area greater
than 200 m2 /g. is obtained by the pyrolysis of an organometallic, polymeric
precursor under NH3 to 1000 *C. Additional heating to 1400 OC under N2 produces
alpha-Si3N4. The addition of up to 15% h-BN was found to affect the coarsening
characteristics of amorphous silicon nitride by promoting surface area reduction
and suppressing crystallinity.
By combining Si 3 N4 and BN molecular and
polymeric precursors prior to ceramic conversion, or incorporating Si. N. and B
into a single preceramic polymer, the relative proportion and crystallinity of the
ceramic phases can be controlled in the resulting Si3N4/BN composites.
INTRODUCTION
Silicon nitride has a unique combination of properties which makes it an
important material for high temperature and electronic applications.
In addition
to its resistance to thermal shock, creep, corrosion, and oxidation, dense Si3N4 has
high electrical resistivity, a low coefficient of thermal expansion, low density, and
good high temperature strength.
These attributes are exemplified by the use of
Si3N4 as crucibles. turbine blades, nozzles, tiles, bearings, cutting tools, and
electronic
components.
Boron nitride is another useful material
for high temperature
and
electronic applications due to its high melting point, low density, low coefficient
of thermal expansion, and resistance to oxidation up to 1000 *C.
BN is also
transparent over a large spectral range and is often used as a boron diffusion
source, a surface passivator, or a solid lubricant.
A composite consisting of homogeneously dispersed BN in a Si3N4 matrix
may exhibit beneficial properties of both components.
To date however, limited
research has been reported on such a material.
Past work with Si3N4/BN
composites has concentrated on their preparation by CVD and CVI techniques [I5].
Fukunaga, et. al. (1] prepared electrically conductive films of amorphous
Si3N4/BN composites.
They concluded that the 10 A voids within the amorphous
Si3N4/BN composite were occupied by turbostratic BN since Si3N4 and BN do not
constitute a solid solution even in the amorphous state.
They also claimed a
decreased free volume in the Si3N4/BN composite and an increased crystallization
temperature of the amorphous Si3N4 matrix in the composite.
Annealing at ca.
1600 OC led to precipitation of h-BN in the amorphous Si3N4.
Hirai and coworkers
12.31 produced opaque and transparent amorphous Si3N4/BN composite films with
up to 36% BN by weight using CVD.
Mazdiyasni and Ruh 16) found improved
electrical and thermal shock behavior of powder-based, hot-pressed Si3N4/BN
composites over Si3N4 alone.
The dispersion of the low modulus BN in the high
strength, high modulus Si3N4 matrix significantly improved the thermal stress
resistance, lowered the dielectric constant by ca. 20%. and
did not drastically
affect the mechanical integrity of the composite.
Fabrication of uniform Si3N4IBN material by more conventional ceramic
powder fabrication methods is virtually impossible due to the large differences in
the melting points of Si (1415 *C) and B (2177 °C) and the differences in N
diffusivities in both materials.
The use of organometallic precursors to ceramic
composites may provide several advantages, including 1) improved compositional
Mat. les. Sm. Spyp. Proc. Vol. 171.
11m110
Materials Resarch Soctoty
80
homogeneity due to atomic and molecular mixing of the elemental components, 2)
lower processing temperature, 3)
higher purity ceramic products with
controllable microstructures, and 4) processing flexibility to prepare thin films,
powders, binders, fibers, and monoliths.
Past researchers have employed polysilazanes as organometallic precursors
to Si3N4.
Unfortunately, the resulting ceramics typically consisted of mixed
alpha- and beta-Si3N4 phases, with considerable contamination by SiC, SiO2, and C
phases[7,S].
Polymeric prec-irsors to boron nitride are generally based on
substituted borazines [9-111, but were also obtained from decaborane- [12] and
pentaborane-based systems [131, or from substituted borane Lewis Base adducts
[14].
Our approach to fabricating Si3N4/BN composites from organometallics is to
mix a polymeric precursor to Si3N4 with a polymeric or molecular precursor to BN
prior to ceramic conversion and in addition, to prepare a single polymer which
contains Si, N, and B. The effects of incorporated BN on crystallinity, phase type
and distributicn, and surface area of the solid phases were examined.
EXPERIMENTAL METHODS
Glovebox and Schlenk techniques were employed whenever possible.
Dried
solvents, starting materials, and glassware were used throughout.
Precursor and
ceramic materials were characterized by a combination of FTIR. XRD. and BET.
A
vinylic
polysilane (VPS)
with the
approximate
composition
[(Me3Si)w(CH2=CHSiMe)x(HSiMC)y(SiMe2)z], where Me represents a methyl group,
was used as the precursor for Si3N4.
We have recently described the conversion
of this polymer to "amorphous" Si3N4 at 1000 OC in NH3, and the subsequent
crystallization of the ceramic to high purity alpha-Si3N4 with heating above 1400
OC in N2 [15].
The polymeric BN precursor was a poly(borazinylamine) (PBZA), which was
prepared according to the method described by Narula et. al. [11]. Briefly. 2 g (10.9
mmol) of trichloroborazine was added to 50 ml of dry diethyl ether. The solution
was stirred and cooled to -65 *C.
Hexamethyldisilazane, (HMDS) 2.63 g (16.32
mmol), was added at once and the mixture was slowly wanned to room temperature.
The reaction was carried out under N2.
Volatiles were removed under vacuum,
leaving a white, fluffy powder.
The powder was transferred into a N2-filled
glovebox whereby 25 ml of dry hexane was added to create a suspension. NH 3 was
bubbled through the suspension for 12 hours, after which volatiles and
remaining
solvent were removed under vacuum.
The resulting material was a
white powder.
4
VPS was mixed with the PBZA in 1:1 and 10:1 wt ratios. In both instances,
the powder was eveny dispersed in the VPS by preparing and mixing a slurry
with dry hexane, followed by removal of the solvent under vacuum. The resulting
white, gluey pastes were placed in a molybdenum pyrolysis boat and heated in
anhydrous. prepurified NH3 to 1000 *C according to the schedule described for
preparing the Si3N4 [15]. The resulting white solids were subsequently heated up
to 1600 °C under N2 for 4 hours.
In separate experiments, (CH3CH2)3B:NH3. an adduct prepared by
condensing NH3 over a pentane solution of (CH3CH2) 3 B at -50 *C followed by
removal of excess NH3 and solvent by vacuum, was mixed with VPS in a 1:3 wt ratio
and heated as above.
provide
A polymeric precursor which contains Si. N. and B was synthesized to
a homogeneous mixture of the elements on the atomic scale.
The
61
preparation and characterization of this polymer will be reported in a future
publication.
This poly(borosilazane) (PBS) was pyrolyzed in NH3 by heating from
25 OC to 1000 OC in 10 hours and holding at 1000 OC for an additional 10 hours. The
resulting solid was subsequently annealed under N2 up to 1600 °C for 4 hours.
EXPERIMENTAL RESULTS
The effect oi heating on the physical appearance of the various sample'
summarized in Table I.
is
Table I. Si3N4/BN Composite Appearance as Determined by Temperature
Sape1000
VPS
1:1 VPS:PBZA
10:1 VPS:PBZA
3:1 VPS:(CH3CH2)3B-NH3
PBS
Infrared
OC/NH,;
tan solid
white solid
tan solid
black-tan solid
orange-tan solid
1600 *C/N?
white
white
white
white
white
solid
solid
solid
solid
solid
Analysis
Typical FTIR spectra for the various composites are shown in Figures 1-2.
For the 1:1 and 10:1 VPS:PBZA samples, the presence of B-N stretches at 1390 cm - 1
and 801 cm -4 [16], and the Si-N stretch between 1100 cm - 1 and 800 cm - 1 [16].
confirm the formation of composite materials (Figre 1).
The Si-N stretches are
significantly more intense in the 10:1 samples than in the 1:1 samples, with
simultaneous reduction in the B-N stretches.
As the temperature is increased from
1000 *C to 1600 OC, fine structure in the spectra becomes more pronounced as
crystallization occurs, and the intensities of residual bands at 3430 cm -4 (N-H) and
2900 cm- I (C-H) decrease.
Mixtures of 3:1 VPS:(CH3CH2)3B-NH3 do not form detectable silicon
nitride/boron nitride composites upon pyrolysis. The FTIR of the 1000 *C sample
shows only a broad Si-N stretch centered near 900 cm - I, which is typical of
amorphous Si 3 N4 (Figure 2).
No peaks representing B-N bonding are observed
near 1400 cm "1 or 900 cm "1 . although these bands may be hidden beneath the
large Si-N band.
With heating to 1600 °C, the FTIR shows only the increased
crystallinity of the amorphous Si3N4, and confirms the absence of BN.
The PBS sample shows an extremely broad band ranging from 1600 cm " I to
500 cm"I (Figure 2) after heating to 1000 OC. Residual N-H and C-H bands are also
apparent.
With heating to 1600 *C, however, the skewed band near 1400 cm- I,
indicative of BN and the large structured band from 1100 cm " I to 800 cm - I,
attributed to the Si-N stretch, confirm the presence of both ceramics.
X-Ray
Diffraction
Powder diffraction patterns for the various composites are shown in
Figures 3-4. The 1:1 VPS:PBZA 1000 OC powder shows two broad peaks near 25 and
43 degrees (Figure0 3). which are due to the presence of turbostratic BN [171. Upon
heating at 1600 C, these peaks sharpen, due to increased crystallinity of the BN.
and several smaller peaks are now obvious, due to crystalline alpha-Si 3 N4.
The 10:1 composite provides very different patterns (Figure 3). At 1000 *C.
the sample primarily consists of amorphous Si3N4, noted by the very broad peaks
centered near 33 and 70 degrees (15]. By 1600 OC. the alpha-Si3N4 has extensively
crystallized, but the presence of turbostratic or crystalline BN is not obvious.
82
1000 OC
A 1000 0
A
1600 °C
1600jC
100O
100000
B
-V
B
16000cc10
4000
3000
2000 1600 1200 800
400
WAVENUMBER (CM-')
4000
3000
2000 1600 1200 800 400
Figure 2. M7rR spectra ofSi 3N4/BN composites
A) 3:1 V PSt
3 B-NIH,
B) PBS
WAVENUMBER (CM-)
0
Figure 1. FI'IR spectra of SigN 4 /DN com~posites60
A) 1:1 VPS:PBZA, B)10:1 VPS:PBZA
1600 OC
OC100000C
1750
A
I"
16000
-
100000C
5
25
35
45
55
26
FIgure 4. XRD pattema of PBS
d
75
'
25
35
*S
55
"
79
20
Figure 3. XRD pattemrns of various SiN
N eomposites
A)I: VPS:PBZA, B)1:1 VPS:PBZA
x
83
The patterns obtained for pyrolyzed VPS:(CH3CH2)3B-NH3 show only the
crystallization of the Si3N4, suggesting that the borane adduct does not serve as a
BN source in this context.
PBS also exhibits a unique series of diffractograms (Figure 4).
At 1000 OC
0
and 1600 C, the very broad peaks indicate that the product is amorphous [15].
Heating to 1750 °C for 2 hours, however, produces a mixed composition of both
alpha- and beta-Si3N4.
The amount of crystalline BN is below the detection limit of
XRD, however, BN is present since the B-N stretch is detected by FTIR.
Surface
Area
Measurements
Preliminary surface area measurements are listed in Table II.
Initial
results indicate that BN generated in-situ from polymeric precursors reduces the
surface area of the ceramic product, as observed for added h-BN powder [18,191.
This trend is clearly seen at 1000 *C. where the surface area decreases from 247
2
2
m /g to 50.2 m /g as the initial concentration f BN precursor is increased.
The
borane adduct, (CH3CH2)3B-NH3, is a poor BN precursor as shown by the relatively
unchanged surface areas relative to VPS alone.
Table If.
2
Specific Surface Area (m /g)
Precursor
VPS
10:1 VPS:PBZA
1:1 VPS:PBZA
3:1 VPS:(CH3CH2)3B-NH3
VPS+10% h-BN
Dependence
on
Temperature
lop010
247
226
50.2
260
220
12.3
19.7
7.5
11.6
2.2
DISCUSSION
Previous efforts in our
laboratory have shown that for temperatures
0
ranging from 1000 *C to 1600 C, the addition of up to 15 wt % of commercial BN
powder to precursor-derived Si3N4 reduces both the surface area and crystallinity
of the solid compared to precursor-derived Si3N4 alone [18,19].
Added BN does not
help sintering of Si3N4 since it suppresses crystallization and mass transport,
which are the first steps toward complete densification.
The effect is possibly due
to the lowering of oxygen activity in the presence of boron.
By combining polymeric precursors to both Si3N4 and BN prior to pyrolysis,
composites can be fabricated.
The relative proportion of the ceramic phases can
be controlled by varying the weight ratio of the starting polymers, as seen with
the 1:1 and 10:1 VPS:PBZA composites. The adduct (CH3CH2)3B-NH3 was not a useful
precursor to BN, probably because the adduct decomposes prior to reaction with
the VPS.
A polymer which contains Si, N, and B provided a single component,
homogeneous, molecularly mixed precursor to Si3N4/BN composites.
FTIR analyses indicate a mixture of both Si-N and B-N bonding
environments following pyrolysis of the VPS:PBZA and PBS precursors.
XRD
suggests that phase separation of the Si3N4 and BN occurs; this is particularly
evident for the 1600 *C 1:1 VPS:PBZA sample. The crystallinity of the Si3N4 in the
1600 °C 10:1 composite was somewhat lower than that obtained from the precursorderived Si3N4 alone, while that of the 1:1 composite was significantly less.
These
results indicate that molecularly mixed BN from a polymeric precursor also
suppresses crystallinity of amorphous Si3N4, as seen for added h-BN powder (18,
191.
OC,
The PBS polymeric precursor resulted in both alpha- and beta-Si3N4 at 1750
yet showed little crystallinity after 1600 OC treatment.
This observation
84
suggests that the BN has increased the crystallization temperature of the Si3N4. It
is possible that the presence of both phases is due to a better distribution of the BN species within the polymer or the formation of a high temperature liquid.
The reduction in surface area of Si3N4/BN composites relative to Si3N4
alone may be explained by the formation of B203 or borosilicate glass through
reaction with surface oxygen and B. added either in precursor form or as h-BN.
These reactions lead to a reduction of free volume in the Si3 N 4 and thus a
reduction in surface area.
Composites derived from polymeric precursors are expected to show
increased strength and toughness over the component ceramic phases while
retaining characteristic refractoriness and resistance to abrasion and corrosion.
This preparation method also shows potential for producing glass/ceramics or
amorphous material for high temperature applications in composites.
ACKNOWLEDGEMENT
This work was funded by the National Science Foundation under a Materials
Chemistry Initiative Grant.
REFERENCES
1. T. Fukunaga, T. Goto, M. Misawa, T. Hirai. and K. Suzuki, J. Non-Cryst. Solids,
95/96 (1987) 1119.
2. T. Hirai in Emergent Process Methods for Hieh-Technoloev Ceramics, Materials
Science Research Vol. 17, eds. R. F. Davis, H. Palmour, and R. L. Porter, Plenum
Press (1984) 329.
3. T. Hirai, T. Goto, and T. Sakai in Emereent Process Methods for High-Technoloay
Ceramics, Materials Science Research Vol. 17, eds. R. F. Davis, H. Palmour, and R. L.
Porter, Plenum Press (1984) 347.
4. T. Goto and T. Hirai, J. Mater. Sci. Letters. 7 (1988) 548.
5. K. Sugiyama and Y. Ohsawa, J. Mater. Sci. Letters, 7 (1988) 1221.
6. K. S. Mazdiyasni and R. Ruh, J. Am. Ceram. Soc., 64[7] (1981) 415.
7. R. M. Laine, Y. Blum, R. Hamlin. and A. Chow in Ultrastructure Processin of'
Advanced Ceramics, eds. J. D. Mackenzie and D. R. Ulrich, J. Wiley & Sons (1988)
761.
8. D. Seyferth, G. H. Wiseman, and C. Prud'homme, J. Am. Ceram. Soc., 66 (1983) C13.
9. K. J. L. Paciorek and R. H. Kratzer, Ceram. Eng. Sci. Proc., 9[7-8] (1988) 993.
10. K. J. L. Paciorek, D. H. Harris, and R. H. Kratzer, J. Polym. Sci. Polym. Chem., 24
(1986) 173.
i. C. K. Narula, R. Schaeffer, and R. T. Paine, J. Am. Chem. Soc., 109 (1987) 5556.
12. D. Seyferth and W. S. Rees, Jr., Mat. Res. Soc. Symp. Proc., Vol. 121, Materials
Research Society (1988).
13. M. Mirabelli and L. Sneddon, Inorg. Chem., 27 (1988) 3271.
14. J. Beck, C. Albani, A. McGhie, J. Rothman, and L. Sneddon, Chemistry of
Materials, I (1989) 433.
15. W. R. Schmidt, V. Sukumar, W. J. Hurley, Jr., R. Garcia, R. H. Doremus, L. V.
Interrante, and G. M. Renlund, submitted to J. Am. Ceram. Soc., Oct. 1989.
16. R. A. Nyquist and R. 0. Kagel, Infrared Snectra of Inoreanic Comnounds
Academic Press (1971) 114.
17. J. Thomas, Jr., N. E. Weston, and T. E. O'Connor, J. Am. Chem. Soc., 84[241 (1963)
4619.
18. V. Sukumar. Master's Thesis, Rensselaer Polytechnic Institute, December, 1989.
19. V. Sukumar, W. R. Schmidt, R. H. Doremus, and L. V. Interrante, submitted to
Mat Letters.. Dcc. 1989.
PART 11
Emulsions /Blocks
87
STABILIZED NANOARTICLES OBTAINED FROM SYNTHETIC
POLYMERIZABLE MICELLES AND VESICLES.
QCt4STANTII4OS M. E'ALHOS
NEC "Demokritos, Aghia P'eraskevi, 15310 Attiki, Greece.
ABSTRACT
The structural charateristics anid the formation of monomeric and
stabilized polymeric micelles and vesicles are reviewed. Characterization of
these nanoparticles
involved
stability studies,
molecular
weight
determination, permeability and fluorescence investigations, as well as
electron microscopy and DSC studies.
INTRO~DUCIONI
Spherical micelles of 3-6 na in diameter (1] are in a dynamic
equilibrium with their moomrs andican be stabilized either by the
solubilization within them of
appropriate monomers
arid
subsequent
polymerization,
or by the polymerization of mice lle-forming mmosr
surfactants [2]. Synthetic vesicles on the other hand with diameters from 30
to 300nm possess greater kinetic stability then micelles; [1]. In addition,
the dependence of the atructrure endi size of vesicles on the mode of their
formation [3,4] allows greater structural flexibility than micelles. However
further stabilization [5,6) of these particles is required and this was
achieved by addition polymerization or polycciviensat ion.
In the present study we will discuss the stuctural requirements for the
formation of micelle and vesicle forming polymerizable surfactants and the
diversified methods for the formation of the their polymerized counterparts.
The stability of nanoparticles was the main property to be investigated.
Furthermore their molecular weights were determined anid the structure of
micelles was investigated by fluorescence spectroscopy whereas vesicles were
further characterized by permeability studies, electron microscopy, DSC and
currently by video enhanced optical microscopy. Currant and prospected
applications of both monomeric anid polymierized particles include their use
as energy conversion system [7,83, drug-carriers [5,9] in medicine, and as
media for biomimetic reactions [103.
1. POLYMERIZED) NICELLE
-
FORMATION AND CHARACTERIZATION4
1 1- Synthesist of miml la-formning oolvmmr7Able mairfoantantot
Surfactants bearing one long alkyl-chain coupled with a hydrophilic
head associate in water above a critical concentration ((ClC) forming
sicelles. Introduction of a polymer izable group does not in general
MC=C4
MC=CH
H,,C-C-CH
CH Br -
(C k
I
C o
CHI),,
CHI
CH,
'CPBr
II
,:
HM
(CH,).
O~
CH, -N-CH,
Mat. No*. Soc. Symp. Proc. Vol. 171. 01990 Materials Research Society
85
drastically modify the basic molecular structure and the resulting monomers
also aggregate forming micelles. The functionalization of
surfactant
structure by polymerizable groups can be performed either near the heed or
at the lipophilic group, as in monomers I-V.
1
G2.neral onsisratinns of minellar nlvmerizatim
Studies on micellar polymerization have been limited to addition
polymerization
under conventional polymerization conditions. Certainly
polycaidensation cannot be ignored for the formation of
polymerized
aggregates.
These
latter nanoparticles,
in addition to their other
properties will be biodegradable.
The study here will be restricted to the polymerization of monomers
which are themselves surfactants. Several parameters may in principle affect
micellar polymerization and structure of nanoparticles [1i] such as:
a) The concentration which controls the size and shape of monomeric
micelles may very well affect polymerized counterparts. The type of
aggregates as a function of surfactant concentration is shown pictorially
below.
b) The location and the nature of the polymerizable group at the heed
or lipophilic moiety of the
surfactant
may
in principle
affect
polymerization and also the architecture and conformation of polymerized
micellar particles. Thus two types of polymerized micelles can be envisaged,
named H (for head) and T (tail) respectively. In type H spherical or
ellipsoidal polymerized micelles is very difficult to visualize because
their formation is prevented by formidable packing constraints. In this case
disk-like intermicellar aggregates or bilayer structures will be visualized.
Above considerations on polymerized micelles are however valid if the
polymerization
occurs
intramicellarly
under topochemical conditions.
Dynamics of micellization (12] fall into two categories: a) The millisecond
domain in which whole micelles dissolve and reform and b) the microsecond
domain when a single surfactant exchanges between the micelle and the
solution. In order for the polymerized particles to have more or less the
shape and size of the monomeric aggregates the polymerization must be fast
enough that is there is no chance for monomeric micelles to be altered.
89
1 3
Polmerized wicfllAR ots
g
ingi thru gh
elllar
oalpmarition_
In
reviewing polymerized aggregates originating from micelle-forming
monomeric surfactants a more or less chronological order is
followed.
Studies performed sofar are rather fragmentary and above mentioned general
aspects have only been partially taken into account in micelle formation.
a.
Heterocyclic polyquaternary ammonium salts.
The first
substituted
exaples in micellar polymerization involved
vinyl2yridinium
salts[13-16],
typical of micelle-forming surfactants and
certain
methyl
the structures of which are not
which
aggregate
at
abnormally
high concentrations. The investigations concerning these salts focused
primarily on the polymerization mechanism rather than on the effect of
organization-aggregation on polymerization and the structure of polymerized
particles. Thus we should not further discuss these polymers.
The polymerization of micelle-forming
3-n-dodecyl-l-vinylimidazolium
iodide (1) C17] was perforred comparatively in isotropic and micellar media
(18]. Although polymerization kinetics were not performed and the particles
it seems that micellization, at least as
obtained were not characterized
judged from the isolated polymeric products, does not affect polymerization
[18].
Thus polymers obtained by the two modes of polymerization cannot be
differentiated as far as viscosity, aicrostructure and solubility of the
polymers are concerned. Howeer, CKC obtained by electrical conductivity for
this salt is
zero bemause of its polysap structure.
b. Poly(sodium 10-undecenoat)
Sodium 1O-undsoenoate (IV)
has been studied rather
extensively.
Since
it
bears the double bond at the end of the aliphatic chain it can form
T-type polymerized micelles.
Polymerization was accomplished by y-rays
[19,20] or UV light [21] and the degree of polymerization was found equal to
equal to 10.
Thus
the aggregation number of the monomeric micelles i.e.
polymerization occured intramicellarly each polymer chain forming a micelle.
When polymerization was conducted below CC the rate was practically zero.
The type of organization of polymerized particles was somehow illustrated
when this monomer was polymerized in the liquid crystalline state
[22].
In
this case, structure changes from hexagonal closely packed cylinders to a
lamella structure. The degree of polymerization was 270.
The same mnomer form intermolecular micelles composed
of more than
one intramolecular micelles with a CHC = 10-2 H. The aggregational behavior
of monomeric end polymerized micelles has been studied by electrical
conductivity and confirmed by dye solubilization (19].
The intrinsic
viscosities of monomeric and polymerized micelles had shown that they both
have equal hydrated sizes while a larger hydrated size was found for the
intermolecular micelle [19].
By fluorescence probing
[23]
with pyrene and
eploying
the
Is/li
micropolarity index the internal structure of monomeric and polymerized
micelles were copred. Thus polymerized micelles %how a more compact
structure as comared to monomeric counterparts. An a consequence pyrene
does not penetrate inside polymerized micelles as deeply as in the monomeric
micelles which is attributed to the proximity of the aliphatic chains due to
their attachent on the backbone at the core of the polymerized micelles.
The dependence of polymerization and polymerized micelles on the
structure of monomeric aicelles was also exeeplified in the polymerization
of the same monomer with UV light [21].
90
c.Aliphatic polyquaternary salts
(24]
was
polymerized
Allyldimethyldodecyl ammonium bromide (II)
comparatively
under
micellar
and isotropic conditions with y-rays.
Isotropically obtained polymer was partially destructed from the radiolysis
products of water. It seems that micellization protects the aliphatic chains
from the same agents. For this monomer the aggregation number of monomeric
micelles was found equal to the degree of polymerization of polymerized
micelles i.e 33. This may be interpreted by the fact polymerization is fast
enough that polymerized micelles are formed before the surfactants of
As it was
monomeric micelles can exchange with the monomers in solution.
found by fluorescence probing olymerized micelles are more compact then the
monomeric ones.
Other examples of micelle-forming monomers are the long alkyl chain
derivatives of dimethylaminoethyl methacrylate (III) quaternary ammonium
salts (25,26]. For these monomers the rate of polymerization in aqueous
media increases as the alkyl chain length becomes longer. This behavior had
been interpreted as indicating that polymerization is taking place in the
micellar state. The same head methacrylate derivatives were polymerized by
S. Hamid and D. Sherrington [25)
and
determined
their
average
molecular
weights
which range from about 10,000 to 11,000.
It appears that
polymerization is facilitated in micellar media as compared to isotropic and
topochmical polymerization has been achieved. However since polymeric
solutions become increasingly opaque due to the presence of polymerized
species of higher molecular weights as polymerization proceeds, a skepticism
was expressed whether really polymerized micelles were formed through
topochemical
polymerization. According to their analysis on micellar
dynamics, topochemical polymerization was
rather
excluded
and
the
experimentally determined low molecular weight was fortuitous or attributed
to the facility of monomer transfer reaction. For these methacrylate
derivatives however micellization dynamics and polymerization rate data were
not a-wailable and therefore the parameters employed [25] were taken from the
literature for alkyl sulfates and various other monomers.
The structure of polymerized micelles may be affected by the position
of the polymerizable group, at the heed or at the tail (28] as shown for
monomers VI and VII.
The micellar parameters obtained by fluorescence
measurements for monomeric and polymerized micelles are summarized in Table
I.
c H3
NBr*
C
1.CH
iq
N
Bre
3
SI
,
O..C.C H=CH 2
611
VI
V11
According to these results pyrene senses the same micropolarity in both
monomeric VI and polymerized VI micelles while senses higher polarity in
polymerized micelles VII than VI. Thus in polymerized micelles originating
from the tail monomer it is easy for water to penetrate into the interior
thus enhancing its polarity. Concernig aggregation number of polymerized
micelle VI it is seen that it is almost half of of its monomeric counterpart
which very probably arizes from the structural constraints when the backbone
is formed at the interface. On the contrary it is twice as large for the
polymerized micelle of monomer VII. In this case the location of the main
91
chain at the core of the polymerized micelle
packing problems.
does
not
create
significant
Table I. Polarity Index (13I/)
and aggregation number N. of monomeric
and polymerized micelles of monomers VI end VII,
Surfactant
I/1I
N.
0.83
62
Monomeric
VI
Monomeric
VII
-
-
Polymerized
Polymerized
VI
VII
0.83
0.72
24
42
d. Polymerized micelles from non-ionic surfactants
The monomer l-O-3-(4-vinylphenyl)propyl-3-D-glycopyranose
(V) [29) a
stiff polymerizable surfactant polymerized with free radical initiators
above and below the CMC. 2,2
azoisobutyronitrile(AIBN),
dipotassium
peroxodisulfate
(K2S2Os)
and
2-(phenylazothio)naphthalene (ATE)
were
employed as initiators but only polymerization by ATE led to polymerization
with preservation of aggregation number. The polymerization with the other
catalysts resulted primarily in the formation of bigger aggregates. When ATE
catalyst was used it was speculated that the phenyl radicals might initiate
the polymerization mid that the naphtyl thio radicals might terminate
growing chains before diffusion effects become noticeable. Thus it was
possible to have topochemical polymerization conditions by changing the
employd catalyst.
2. POLYMERIZED VESICLES - SYNTHETIC APPROACHES
2 1 Formation nf polvmerized veginlsc
ventional addition polwn-rization.
frnm mnnmerie vesicles by con-
The usual structural feature of vesicle-forming surfactants is the
presence of two long alkyl chains in conjuction with a polar head group.
These structures when functionalized by the introduction polymerized groups
form vesicle-forming monomers C30,31] whose polymerization leads to the
formation of polymerized vesicles. The monomers VIII-XVII are indicative of
the diversity of molecular structures that form vesicles.
Utilization of vesicles for drug encapsulation has led to extensive
research on the formation of single and multi-compartment vesicles. For this
purpose the methods of eonication, reverse phase, evaporation, solvent
injection, extruction have been developed (3,4]. Depending on the method
and conditions, mltilamellar and unilamellar vesicles have been prepared.
Methods for the formation of
vesicles
have
been
critically
reviewed
by
Szoka-Papahadjopoulos (3] and Hope et al. Quite recently a method has been
developed (32) based on the swelling of phospholipid films deposited on
special supports in excess water. The method is simple and it is performed
under mild experimental conditions.
Polymerization is accomplished
by
conventional
methods.
Irradiation
polymerized groups in the exterior and interior of the vesicles. On the
contrary, polymerization may be limited to external double bonds when the
initiator ,added to sonicated vesicles, cannot permeate to the interior.
Depending on the position of the polymerizable group, polymerized
vesicles can either be linked at the polar head, at the middle,
or at the
end of the lipophilic chains as shown below. The location of the backbone
in polymerized vesicles certainly affects their properties.
Heed group
92
C'4j1i,C )C-COO-CMk,/ C
CNI1CH,,.N
C
U
C
N=
CN,
C
,,
CC
Nr,
.C
kI
C"
CC
-C
CC,,,
IX
C..
COOic -0.
C
'
XlV
C'.
Z.C-CNI
CNC"
. C1,
Xl
0
Ci
XV
0
.0cC-Ci.C.-C
CO Ol",,,
-4
CH,(CH,i..CN,
N 4
XI
xi
CI%OCO-iC-", iCOO -C(C-%)=Ob%
C.I~,CC-~C
"~CCi"i
nobility is preserved when the monomers are linked at the end of the long
aliphatic chains, whereas head group mobility is lost when the main chain is
located near the head grop. Indiacetylene monomers, polymerization leads
to extremely rigid structures
which
transitions.
Ci.OCO-(CNCi,,
C" do not exhibit phase(Cm,6sw
i.
0
aipthe
c cins,
whe ras
ftA.C
hed
Kog(cCHi.
H. . $%N (CH0
CNO
group obiity i
lwos
when theps
Man caiIs
6" IXX
2-2- Farmto
yeil~n
ofol~il
~l
rnm
nnonvantimn!lmnr~~
by radny reae-tnna.
The synteis of m
ers, XVII an IX(, form"' this type of vesicles
form
bilayer structures
woieh are polylerized
[33]vesile-forming
or rather "editched
is accomplished
by a funtionalization
on"
of typical
molecules
by
oxidation.
The
resgltin
polymer is reversibly depolymerized"eitohed
off,"
byreduction. It has also beew found that the macrocyclio analogue IM]
is polymerized
[34] in the vesicular ph e by ring openig
poly
riation
initiated with catalytic amount of dithiothreitol([Uff). Them vesicles are
promising candidates for mechanistic and practical applications due to their
biodegredabi lity
93
2.3 Fnrmtion of polvmeriwI veeiolis fromahiohilic aminnis
Functionalization of amino acids with long alkyl chains my, in
principle, lead to the formation of vesicle forming molecules whose
condensation results in the formation of polype9tides which can also form
vesicles. It has been found that the formation of stable polypeptide
vesicles is governed by proper balance of hydrophilic and hydrophobic
moieties in the monomers. Such monomers [35] fulfilling above condition are
the XX and XXI. The presence of the carboxylic group enhances the
hydrophilicity of the resulting polypeptide after polycondensation. The
vesicles are stable but susceptible to biodegradation. Polycondensation
occurs in the vesicular phase, in the presence of a water soluble
carbodiieide. It must be noted, however, that in general vesicle formation
is not a prerequisite for the polycondensation.
No
C,0oo-CH,
ON
C HCOO- M .,C4,
CH
4-
c , XXII
H
c" ¢'%"N-oCC-
N
H
H
C,
C,,-M,-COC - (CM, CM,O
o o- OC-C=C M,
'C...
CHI
C,.OOC -
e
-C
-CO-(C"M.0)M,-OC-C =C,
YXIV
C.,-WO-CM,
C.N--.,CH
51%
1
1
CMf-o-P-0(CH, CHMo).-oC- C=CN
XX
2 4
XI
XXV
ForMAtiM of vesinles from innene polvmers.
In this case a long-chain dibromide interacts with a ditertiary amine
of the same chain length, forming ionene polymers (36] leading to vesicles
after sonicatiun. lonene polymers cceposed of alkyl chain of different
lengths do not form membrane structures. Further work is required for the
synthesis and structure elucidation of this type of vesicle forming
polymers.
N--(C,*I N
rC,.m li *
II
I-
I
C1
--0
?ro
(I
CM" CH
CHI
o-
4
"--4CI4m
C,
C,
- o
CMO
10
o
/- ? --oP-o0
/--
-
C,
CH.
XXVI
2_5. Fo~mti
A
crucial
of olynerized vasioles from oreformed mlmers.
problem
of polymerized vesicles is whether their formation
94
requires formation of monoeric vesicles which subsequently polymerize. This
problem had been answered recently. Thus polymers prepared isotropically
formed polymerized vesicles under usual
from monomers [37) JOII-XXV,
is
conditions. A characteristic of these monomers
introduction
the
of
a
hydrophilic spacer between the polymerizable group and the amphiphilic
moiety. In this way an efficient decoupling of the motion of the polymer
backbone
and
the
groups
amphiphilic
is
achieved.
The
fluidity of the
polymerized vesicles, due to these spacers, is preserved and this results in
membrane structures simu'lating biological membranes. However, it has been
for
preserving
essential
not
are
reported recently that spacers
" omomer-like" packing behavior of polymeric surfactants. Monomer [38] XXVI
Sonication of the
was polymerized through oxidation of the thiol groups.
polymer dispersion afforded vesicles having diameters 200-1000 A.
3. KINETICS IN THE VESICULAR PHASE.
Understanding of the potential of polymerized vesicles necessitates a
detailed elucidation of the kinctics and mechanism of their formation as
they
vesicle structure. Such studies had been performed for styrene
affect
bearing quaternary ammonium surfactants [39-41] XXVII-XXIX.
Continuous
UV
irradiation of monomer XXVII, or irradiation by laser pulses, decreases
styrene absorbance in a first order process. Calculated rate constants were
independent of the vesicle concentration but increased linearly with
increasing intensity of the laser pulses. Rates were considerably slower in
isotropic ethanolic solution compared to vesicular medium, and in contrast
to vesicle polymerization the rate depended on monomer concentration. This
finding ,and the fact that the sizes of vesicles remained practically
occurs
polymerization
suggest
that
polymerization,
unchanged
on
intravesicularly on the surface with an apparent reduced dimensionality
which can be analysed on a per vesicle rather than per volume basis.
c,..,, COO (CM.
C,
C
C
C
p
C.C,
Cop)
-'CM,
'-.
o
VIII
M
0=
COC
0% \c,
.,J
c.,&o-4HI(-c./
CMOC>C'.
(C,),-M
C
CCHB
0=
CI,
C.M.
C,(CM,),, ,
o,
XXVll
CM,
C.,,
C,
Br
, CC=OCM,3
XXXIV
'o(o
,o
\..
CC
XXIX
4. CHARCERIZATION4 OF POLYMERIZED VESICLES.
4-1. Stability Studies
Polymerized vesicles in general retain the structure of their -momeric
radius Rh of the vesicles
counterparts. For instance, the hydrodynamic
upon
A
formed from monomer XXVII is 2500 A and changes to 2750
polymerization
[28).
Electron
microscopy also provides excellent evidence
for the retention of vesicular structure of the momeric vesicles upon
polymerization.
Evaluation of polymerized vesicle stabilities is performed by the
addition to their dispersions of surfactants such as sodium dodecyl sulfate
[42) or of increasing quantities of ethanol (43). In the latter case, UV
absorbance of solutions was found constant for polymerized vesicles upon
addition of 0-25% (V/V) ethanol. Monomeric counterparts showed a dramatic
decrease in turbidity. Thus, polymerized vesicles originating from monomers
(47] XXXI (48)
VIII [43] IVX [44, 45] XV [48) of Table I, as well as M
XXXII (48] showed enhanced stability as compared to their mononeric ones.
An interesting case in which excellent stabilization wee obtained
involves the polymerization of di(undecenyl)phosphate [48]. Fulymerized
vesicles resulting from this monomer were stable for years (own unpublished
results)
without
crosslinking
of
precipitating.
the adjacent
Their
stabilization
was
attributed
to
layers of the vesicles, (below on the left)
involving 502 polymerization of the vinyl groups.
The stabilization achieved through polymerization of monomeric vesicles
is
not always a staightforward process nor is it always predictable. For
instance, it has been found for allyl and diallyl vesicle-forming quaternary
ammonium salts [44,45] that only the polymerized vesicles derived from the
diallyl derivative (below on the right), incorporating a pyrrolidine moiety
and prepared by a mechanism involving alternate
in
the
backbone,
intramolecular and intermolecular growth steps exhibited good stability.
Polymerized vesicles obtained from the allyl derivative are unstable and
dissipate shortly after y-irrdiation. Although this behavior is not easily
rationalized,
it
seem
that the
insertion
backbone of diallyl polymerized vesicles is
of pyrrolidine moiety in the
probably
responsible
for
the
stabilization of the vesicles.
ucH.
{C.yjyu
Cu.
C.
C,..
C.,
C.U
C-,
C.,
C.,
Stabilization of synthetic vesicles was also treated in a manner
i.e. by
stabilizing biome branes,
analogous to that which nature uses in
coating their surface with polypeptides or polysaccharides. Polymers were
hydrophobic anchor
attached to vesicles surface by ionic interactions,
These ionic
groups, or polymerization of charged, water-soluble monomers.
monomers were attached to charged lipid molecules either via salt formation
or as 1ounterions [5,
493.
In
the latter case methasrylate was the
molecule
dioctedecyldimethylaimium
vesicle-forming
counterion
of
methacrylate [48]. Upon polymerization of the methacrylate counterions, the
monolayers.
vesicles were encased within two concentric poly(methacrylate)
Salt formation at the vesicle surfaces is demonstrated (51] by the use of
4-vinylpyridine (4VP) which polymerizes in organic solvents or water with
the addition of protic acids. In the present case, dicetyl phosphate is used
as vesicle-forming acid. Salt formation is achieved by the addition of 4VP
which is
followed by spontaneous polymerization. A polyaddition poly(1.4
the
obtained covering the surface of
pyridiniundimethylene salt) is
Retention of the liposmm structure after coating with the
vesicles.
it
should
polymer was confirmed by electron microscopy. In this connection
a well-known
be mentioned that cetyltrimethylammium bromide (CrAB),
micelle-forming surfactant, is forming bilayer membrane structures when the
bromide is replaced by a polyacrylate counterion. In this way, the nature of
the counterion controls (52] micellar or bilayer phase formation from a
single-chain surfactant.
2
PArmeaility Shiu
Controlled permeability in a significant property of both monomeric and
Vesicle permeability can be determined with any
polymerized vesicles.
water-soluble marker using methods based on fluorescence, enzymatic, redox
96
detection, electron paramagnetic resonance spectroscopy, and radiochemical
tecqniques.
[sH] glucose is a preferred compound because it does not
interact with cationic surfactants and can be used at low ionic strength.
This is required since vesicles formed from cationic surfactants tend to
aggregate at salt concentrations exceeding 20mM. Thus the permeability of
poly(XXXIII) and poly(XXXIV) vesicles is about half of those derived from
their respective monomers, XXXIII and XXXIV.
Copolymerization with a cross-linking agent further reduces the leakage
rate. In contrast, formation of peptide vesicles through the condensation of
long-chain amino acids [35] XX and XXI showed increased permeability. Thus,
permeability of vesicles from the homocysteine derivative is 0.2 to 0.4 of
that of cysteine vesicles. The difference in permeability behavior is likely
due to the higher lactam content, and hence, lower amounts of oligopeptides
in vesicles from XX than in vesicles from XXI.
4-3. Molecular Weight Determination.
Concerning molecular weights of polymerized vesicles it has been found
that many vesicles consist of several polymer fragments and not, of only two
polymer chains resulting from the polymerization of each layer of an
one-compartment vesicle. The structure of polymerized vesicles may be
related to or modeled as intermolecular micelles consisting of individual
polymeric chains (each being an intramolecular micelle) aggregated into
bigger assemblies. The fragments vary signigicantly depending on the
specific structure of vesicle-forming monomers and conditions of the
experiments. Thus, while for monomer XXVII (39] chain length is 20, the two
methacryloyl monomers [53],
(XOMIII)
and (XXXIV) which were polymerized
catalytically, exhibited chain lengths consisting of about 500 units. For
the polymerized vesicles derived from XO0III the calculated number of
monomers per vesicle was 104-3. 108 and therefore there must be 600 chains
per vesicle while for those from XXXIV, with a number of monomer units per
vesicle equal to 104-8.104, there must be only 20-60 polymeric chains. In
addition, preliminary experiments showed that the molecular weight of
poly(XXXIII) varies inversely with the time subjected to sonication before
polymerization. These results suggest that lower molecular weight polymers
are formed in smaller vesicles.
4-4 Miscellaneous Charanterization Studies.
The thermal behavior of
polymerized
vesicles
as
investigated
by
Differential Scanning Calorimetry depends on the type and position [54] of
the polymerizable group of their monomeric counterparts and the phase
transitions that are observed correspond to changes from the gel to liquid
crystalline phases (30].
Thus the phase transition of polymerized vesicles
is retained when the backbone is located in their hydrophilic surface while
is lacking when a rigid, fully conjugated main-chain is formed by the
polymerization of monomers bearing diacetylenic moiety at the middle of the
hydrocarbon chains. It seems that the crucial parameter for the exhibition
of phase transition by polymerized vesicles is a decoupling of mrsion of the
backbone from that of the amphiphilic side-chain. In addition, whi a spacer
is introduced between the polymerizable group and the amphiphilic moiety the
phase transition appears at higher temperatures, and is narrower as compared
to that of monomeric vesicles [37]. This behavior is in contrast with that
of polymerized vesicles without spacers (54,55] in which transitions are
broadened and shifted to lower temperatures.
Recently
introduced
video
microscopy
[56]
and
specifically
video-enhanced differential interference contrast microscopy [57,58) for
the study of molecular assemblies can effectively been used for the study of
polymerized aggregates. Specifically this technique allows an immediate and
97
rapid characterization of organized assemblies
and other colloidal
suspensions free from artifacts by direct visualization cn a television
screen. Particles with sizes down to 50nm, their dynamics, stability and
slow flocculation cain be directly pictured, recorded, analysed in real time.
CNCLUDING REMARKS
Elucidation
of micellar
polymerization and characterization of
polymerized micelles; has rather delayed, primarily due to the lack of
systematic studies. At the noment, what is needed, is LVe right choice of a
series of micelle
forming mooers,
each with minor
structural
differentiation from the other, and the investigation of the effect of these
modifications on polymerized micelles. In polymerized vesicles, on the other
hand, intensive and systematic work resulted in the clarification of
vesicular polymerization and of the problem associated with the structure
of these particles. The diversity of methods that have been developed for
their formation coupled with varying degrees of stability and permeability
render these aggregates appropriate for a wide variety of applications.
WEEPCES
1. J. H. Fendler, J. Phys. Chem.,fl, 2730 (1985).
2. V. V. Egorov and V. P. Zubov, Foss. Chain. Reviews, 1%, 1153 (1987).
3. F. Szoka, and D. Papahaijopoulos, Amn. Rev. Biophys. Biceng., a, 467
(1980)
4. M. J. Hope, M. B. Bally, L. D. Mayer, A. S. Janoff, Cullis., J. Chem.
and Phys. of Lipids, 40, 89 (1986).
5. H. Ringsdorf, B.Schlarb, J. Venzmer, Angew. Chemn. Int. Ed. Engl., _M2,
114 (1988)
6. C. M. Palaes, Chem. Soc. Rev., 45 (1985).
7. M. Gratzel, Ber. Bunsenges. Phys. Chem., ft, 981 (1980).
8. J. F. Rabek, Frog. Polym. Sci., 1, 83 (1988).
9. D. A. Tirell, L. G. Donamura, A. B. Turek, Eds., Macroeclecules as Drugs
and as Carriers for Biologically Active Materials, Annals of the New York
Academy of Sciences, Vol. 446, 1985.
10. J. H. Fendler, Membrane Mimetic Chemistry, Wiley-Interacience, New York,
1982.
11. C. H. Paleos and A. Malliaris, J. Macrom. Sci., Rev. Macros. Chaem. Phys.
C28 (364), 403, (1988).
12. J. Lang, T. Tondre R. Za, R. Bauer, H. Hoffumann, and W. Ulbright, J.
Phys. Chem.,72, 276, (1975).
13. V. A. Kabanov, Pure Appi. Chaem., 1L, 391 (1987).
14. J. C. Salamone, H. V. Majiud, A. C. Watterson, and A. P. Olson, J.
Polym. Sci., Polym. Chaem. Ed., 20, l1r3, (1982).
15. V. Martin, W. Sutter, and H. Ringsdorf, Makrom. Chem., MI, 89 (1976).
16. C. M. Palae and Dais, 3. Polyn. Sci., Polym. Chatn. Ed., 16, 1945
(1978).
17. 3. C. Salemone, S. C. Israel, P. Taylor, and B. Snider,
J. Polym. Sci., Poly.. Symp., 45, 65 (1974).
18. C. H. Palae, S. Voliotis, G. Margomenou-Leonldopoulou and P. Dais, 3.
Polyn. Sci. Chaem. Ed., J&f,
3484, (1980).
19. E. D. Sprague, D. C. Duecker, and C. E. Larrabee, J. Colloid Interface
Sci., &Z,416 (1983).
20. C. H. Peleon, C. I. Stasslnopoulou, and A. Malliaria, J.
Phys. Chemn., flZ,
251 (1983).
21. K. Arai, Y. Haseki, Y. 09iwara, Makrom. Chem., Rapid Comm., 6, 563,
(1987).
22. R. Thuindathil, J. 0. St ,ifer and S. E. Friberg, 3. Poly.. Sci., Polym.
Chem. Edit. ,ia, 2829, (1980).
A4
98
23. J. K. Thomas, Acc. Chem. Re., If, 133 (1977).
24. C. M. Paleos, P. Dais, and A. Malliaris, J. Polym. Sci., Polym. Chem.
Ed., 22, 3383 (1984).
25. K. Nagai, Y. Ohishi, H. Inaba, and S. Kudo, Ibid., 23, 1221 (1985).
26. K. Nagai and Y. Ohishi, Ibid., 2L, 1 (1987).
27. S. Hamid and D. Sherrington, J. Chem. Soc., Chem Commun., 936 (1986).
28. C. M. Paleos, G. Margomenou- Leonidopoulou and A. Malliaris, Mol. Cryst.
Liq. Cryst., 11, 385, (1988).
29. K. Nagai, H. G. Elias, Makromol. Chem., 108, 1095, (1987).
30. J. H. Fendler, P. Tundo, Ace. Chem. Res., 17, 17, (1984).
31. C. M. Paleos, J. Macrom. Sci., Rev. Macrom. Chem. Phys., In Press.
32. D. Lasic, J. Colloid and Interface Science, 124, 428 (1988).
33. S. L. Regan, K. Yamaguchi, N. K. P. Samuel, M. Singh, M., J. Am. Chem.
Soc., 1983, 105, 6354-6355.
34. S. L. Regen, N. K. P. Samuel, J. M. Khurana, J. Am. Chem. Soc., 1M7,
5804 (1985).
35. R. Neumann, H. Ringsdorf, E. V. Patton, D. F O'Brien, Biochim. Biophys.
Acta, 88, 338 (1987).
36. T. Kunitake, N. Nakashima, K. Takarabe, M. Nagai, A. Tsuge, H. Yanagi,
J. Am. Chem. Soc., 103, 5945 (1981).
37. R. Elbert, A. Laschewsky, H. Ringsdorf, J. Am. Ch
Soc.,
107, 4134
(1985)
38. B. A. Weber, N. Dodrer, S. L. Regen, J. Am. Chem. Soc., 109, 4419
(1987).
1.0,
39. W. Reed, L. Guterman, P. Tundo, J. H. Fendler, J. Am. Chem. Soc.,
1897 (1984).
J. Serrano, S. Murino, S. Millan, R. Reynoso, L. Fucugauchi, . W.
40.
Reed, F. Nome, P. Tundo, J. H. Fendler, Macromolecules, 1B, 1999 (1985).
41. F. Nome, W. Reed, M. Politi, P. Tundo, J. H. Fendler, J. Am. Chem.
Soc., 106, 8086 (1984).
42. A. Sadownik, J. Stefely, S. L. Regen, J. Am. Chem, Soc., 10I, 7789
(1986).
43. S. L. Regen, B. Czech M. Singh, . J. Am. Chem. Soc., 102, 6638 (1980).
44. D. Babilis, P. Dais, L. H. Margaritis, C. M. Paleos, J. Polym. Sci.
Chem. Edit., 2, 1089 (1985).
45.
D. Babilis, C. M. Paleos, P. Dais, J. Polym. Sci., Chem. Edit., 2a,
2141 (1988).
46. C. M. Paleos C. Christias, G. P. Evangelatos, P. Dais, J. Polym. Sci.
Chem. Ed. 20, 2565 (1982).
47. M. F. Poks, H. G. J. Visser, J. W. Zwikker, A. J. Verkley, R. J. M.
Nolte, J. Am. Chem. Soc. 10, 4507 (1983).
48. P. Tundo, D. J. Kippenberger, P. L. Klahn, N. E. Prieto, T. C. Jao, J.
H. Fendler, J. Am. Chem. Soc. 104, 456 (1982).
49. H. Ringsidorf, B. Schlarb, Makromol. Chem., 189, 299 (1988).
50. H. Fukuda, T. Diem, J. Stefely, F. J. Kezdy, S. L. Regen, J. Am. Chem.
Soc., 108, 2321, (1986).
51. K. V. Aliev, H. Ringsdorf, B. Scharb, K. H. Leister, Makromol. Chem.,
Rapid Comun., 5, 345 (1984).
52.
M. Wakita, K. A. Edwards, S. L. Regen, D. Turner, S. M. Gruner, , J.
Am. Chem. Soc., 1=0, 5221 (1988).
53. K. Dorn, E. V. Patton, R. T. Klingbiel, D. F. 0' Brien, H. Ringsdorf,
Macromol. Chem., Rapid Comaun., 4, 513 (1983).
54. R. Buschl, T. Folda, H. Ringsdorf, Makromol. Chem. Suppl., E, 245
(1984).
55. T. Kunitake, M. Nakashima, K. Takarabe, M. Nagai, A. Tsuge, H. Yamani,
J. Am. Chem. Soc., 10, 5945 (1981).
56. S. Inoue, Video Microscopy, Plenum Press, New York and London, 1986,
57. D. F. Evans, J. Brady, B. Kachar, and B. W. Ninham, J. Solution
Chemistry, 14, 141, 1985.
58. D. D. Miller, J. R. Bellare, D. F. Evans, Y. Talmon,and B. W. Ninham, J.
Phys. Chem., 91, 674, 1987.
99
THE PHYSICAL PROPERTIES OF MICROCELLULAR COMPOSITE FOAMS
Alice M. Nyitray and Joel M. Williams, Los Alamos National Laboratory, Los
Alamos, NM
David Onn and Adam Witek, Applied Thermal Physics Laboratory, University of
Delaware, Newark, DE
ABSTRACT
Recently we reported on a method of preparing microcellular composite
foams.
In this procedure an open-celled polystyrene foam is prepared by
the polymerization of
a high-internal-phase water-in-oil
emulsion
containing styrene, divinylbenzene, surfactant, free-radical initiator and
water. After drying, the cells of the polystyrene foam are then filled
with other materials such as aerogel or resoles. The physical properties
of these materials e.g., surface area, density, thermal conductivity, and
compressive strength will be presented.
INTRODUCTION
Physics experiments require very special materials.
For inertial
confinement fusion (ICF) experiments, foams have been suggested as ideal
material to hold uniformly a mixture of liquid deuterium and tritium in a
To do this, the foam must be a material of low itomic
fusion target.'
number with very small cells (< 0.1 um), low density (< 40 mg/cm ) but
handleable.
Other experiments requ re open-celled foams with specific
densities that can range from 7 mg/cm (5 X the density of air) to nearly
full density, and range in length from a millimeter to meteri. One type of
foam that we use extensively is a polystyrene emulsion foam prepared by a
technique introduced by Unilever.
The foam is prepared by mixing styrene,
divinylbenzene, sorbitan monooleate and free radical initiator with water
to form a water-in-oil emulsion that can be polymerized by heating at 600 C.
The solid mass is then oven-dgied to remove3 the water. In this way, foams
with densities from O.012g/cm to 0.20 g/cm and cell diameter of 3-100 Pm
can be prepared depending on the concentration of monomers, surfactant,
initiator and salt used to prepared them. (Fig. I)
lO
m
I
0.025 g/cm 3
Figure 1.
density.
0.133 g/cm 3
The size of the cell "windows" decreases with increasing foam
Met. Res. Soc. Symp. Proc. Vol. 171. 11990 Mterials Research Society
100
Composite foams were developed to reduce the cell size of the foam
matrix by backfilling
e cells of the polystyrene foam with materials such
as aerogel, or resole.T The preparation of the composite foams have been
reported on elsewhere.
The polystyrene foam can be filled completely
(Fig. 2,3), or the cell walls can be coated depending upon the
concentration of filler. This method has been extended to other filler
materials such as backfilling with styrene and divinylbenzene to prepare
foams of higher density.
In the remainder of this paper we will examine
some of the physical properties of the composite materials.
Fig.2
Fig. 3
Figure 2. Composite foams prepared by backfilling polystyrene emulsion
foam with (10%) Si aerogel.
Figure 3. Composite foams prepared by backfilling polystyrene emulsion
foam with (6%) Resorcinol-Formaldehyde (RF).
RESULTS AND DISCUSSION
Specific surface area measurements were made on a Monosorb
(Quantachrome Corporation) single point surface analyzer using 30 mole
percent nitrogen in helium. The single point method was used to obtain
specific surface areas; therefore, the data (Table 1) should be considered
as relative values only.
Compression analyses were determined on 0.75 cm right cylinders using
a Materials Testing System (MTS) with a 500 lb. load and a platen speed of
0.005 inch/minute. Results are reported in Table II. The composite foams
were found to have physical characteristics similar to those of the
polystyrene foam. Compression testing data show the filler increased the
density, but in the case of the Si aerogel filler, added nothing to the
modulus.
Conversely, the resole matrix increased the modulus.
Although
intrinsically weak, the resole appeared to stiffen the walls of the
polystyrene foam. This effect is very dramatic in the carbonized forms of
these moejials
which appear to be one of the best materials for ICF
targets. ,0
101
Table 1. Specific surface area measurements on foam composites
Sample
I.D.
C-1
141-3
141-2
141-1
146-3
146-2
146-1
14-5
14-3
Foam
Type
PS
PS+(2%)RF
PS+(4%)RF
PS+(6%)RF
PS+(2%)PF
PS+(3%)PF
PS+(7%)PF
PS+(5%)Si
aerogel
PS+(7.5%)
Si aerogel
Density
(g/cm )
Specific
Surf ce Area
(m /g)
SEM
Observations
0.105
0.116
0.118
0.116
0.121
0.126
0.158
0.119
3
35
27
28
30
14
20
127
empty cells
coating on cell walls
coating on cell walls
filled cells
coating on cell walls
filled cells
filled cells
partially filled
cells
coating on cell
walls
0.139
138
0.100
0.149
0.111
5
105
391
empty cells
filled cells
filled cells
0.090
538
filled cells
0.089
309
partially filled
cells
0.021
0.048
0.032
9
252
800
empty cells
filled cells
filled cells
...........................................................................
Control
118G
92-1
92-2
92-3
PS
PS+(4%)RF
PS+(10%)
Si aerogel
PS+(7.5%)
Si aerogel
PS+(5%)Si
aerogel
...........................................................................
C-2
28A
67-IB
PS
PS+(2%)RF
PS+(]%)RF
(carbonized)
Table II. Compression strength on foam composite
Foam
Type
Density
(g/cm )
Yield
Strength (psi)
Compression
at Yield (psi)
E-mod.
(psi)
PS
PS+(10%)Si
aerogel
PS+(7.5%)Si
aerogel
PS+(5%)Si
aerogel
PS+(2%)RF
PS+(4%)RF
PS+(6%)RF
0.105
0.139
174
183
4.7
6.4
5800
5800
0.126
181
6.4
5900
0.106
181
6.4
5900
0.111
0.122
0.131
248
294
367
6.2
6.8
6.4
7600
9500
10600
9
Since aerogels are known to be good insdlating materials, we have
attempted to make thermal conductivity measurements on the composite foams.
Thermal diffusivity (@) of the composite samples was measured using a
laser-flash thermal diffusivity system at the Applied Thermal Physics
Laboratory (ATPL), University of Delaware. ;his system and its associated
analytical software is described elsewhere.
The samples in the form of
discs 1 cm in diameter by 1.5 mm or 2.5 mm thick were pre-coated on both
sides with copper, followed by a thin graphite surface to ensure uniform
lateral heat dispersion on both surfaces.
102
Following energy deposition by the infra-red laser pulse on the front
face of the sample, the temperature rise of the second face of the sample
was monitored.
The temperature showed a time dependence distinctly
different from a homogeneous isotropic material.
The theoretical thermal
diffusivity values were calculated from the fractional temperature rise at
each 0.1 fraction of the trace rise-time (tv: V-0.1 to 0.9) between the
pulse and the peak temperature rise of the second face. Only the relative
values of thermal diffusivity (tl/ 2 ) are reported in Table Ill.
We noted that for values of V between 0.1 and 0.5 the calculated
values of @ decreased steadily with time. Thus the values for V-0.1 are
approximately three times the values reported in Table Il. These initial
high diffusivity measurements may be due to residual effects of rapid
radiant heat transfer through the composite materials since there is no
component in the composite which should give such a high initial
diffusivity. For values of V from 0.5 to 0.9 the average calculated value
@ was typically within 10% of the value at V-0.5. This suggests that
thermal transport in this time range is more typical of a homogeneous,
isotropic material and is due to a combination of skeletal conductivity and
thermal transport through the entrapped air.
Pending further transient analysis, we can gain a qualitative
understanding of the changes in heat transport mechanism in these composite
foams by comparing the t1 /2 values of @ listed in Table Ill.
We assume
that the contribution to thermal transport by radiation are small, while
contriubtions from skeletal diffusivity and entrapped air diffusitity are
comparable to each other. Using the blank foam value of 0.0018 cm /s as a
reference, we note that only for sample 146-1 is there a reduction of @
below the blank foam value.
We assume that in this material the
phloroglucinol-formaldehyde (PF) foam completely filled the polystyrene
foam thus reducing the cell-size for the entrapped air.
The resulting
reduction in gas molecule mean-free-path reduced the entrapped air
diffusivity while the skeletal diffusivity was not altered. At the other
extreme we note that sample 14-5 has the highest @ value despite the low
density.
This suggest that the silica filler (which as a thermal
conductivity about ten times that of the polystyrene) has enhancing the
skeletal conductivity without inhibiting the air diffusivity.
The other
two PF aerogels and the three RF aerogels give values equal to, or slightly
above, those of the blank foam. We propose that this relatively small
enhancement could be due to partial filling of the foam cells possibly
providing thermally conductive bridges across the cells. This effect may
be partially compensated by a reduced gas mfp due to some reduction in cell
size. Similar comments apply to the other Si airglass sample 14-3 though
in this case wall-coating giving enhanced skeletal diffusivity is more
likely than for the organic fillers. Further analysis and modeling of the
transient thermal response of the composite foams is in progress.
CONCLUSIONS
Although the materials needed for laser fusion targets or other
physics experiments are highly specialized, we envision that composite foam
technology can be extended to other foam combinations and that novel
filters, insulation, catalytic devices and chromatographic materials could
be made from these materials. At this time we are continuing to study a
way to optimize these materials.
103
V04.
.,1
>
-4
In
to
:j
44
41 44
u
a
f~-
toi
.
0
44
'-4
>
S
,q41
>
~
U1
'.Z4
1
4)41
4
00 U
0
4444
141
10
-A-41
0
SM
c.
0
4
E>.
4
0011
0G
(A
4.)
In
-4
440
41
r
m-o4o-:3. *l
0).44-IN
4.J
04w
44414
Q.
4).. 4
41
0
M'4 _
4104
ad-I
E-'0
EH
4
4
ton
410
4)
U0
.. ,
-44.V
*.4'
40
41
to41
4 u- t'W .
0Q : -14 U)
44-0 4J
414(a m
0
-w
0
4
44 o
41 CL
j 44
4.40
1.44to 0 4- CO
-4
0.
0:
0.0
t t
j
4
0 41>
.4)~ -4
0()
ON :% -4 >1
R10.
0
t 4)
'4 tr
0 O
-4
1-4r.'0
1
4 14 4)
-,-44
0
4J : 4J
44 4.J
1CO-4 '0N r
it1
044 . 0-A-4 to
-4 0
t
41-1
4
0-M
410u
1.0
I a
H-4I
.)to4 )
-4 w
$
)4
'-q4)
Q)
4r.O 4
MU
3:41O
0~ 3
N
0
$4
.H -w
0
to 0
fn
0E0
4.
t41
*41.
4..-
m>,.,I
V V44IO
w 4
4) >44
t0'--H 1-4
0 11
0 51
144M
fo1'0V
A0
0
414-4
41o
m 41 0
4-4
Ul
44
0 to i
V
14
VO>
V.)0
410 4.)>,.0
4.) V
w
W1 H
044-444
-q1
-444-4>
m
I4
4 0 C44
44
40 9
>
4)
O R1-.4
w r.0
41-4 r.4
c
-4 414
c
to>444441)-4
~
>q
0 04-)
---4
10.
c t
41 0 41
a)
.-q 4)
4
4
-444414JZ
> r. -4
4140O-A
'0 4 -44-4
410.a
41
04
0.0
C
0
144=4in
4
.of
W U) 410
N
'0
4
0
4~J
0
V1-4
1t
1-
(a
41
'-1
-a4l
--4
- 1 4-1
Ml
(n
U)I41MC
13
4
0
-4
4
V.)V.-4>
-4
V Q 44
M
44-I0
M11
ON r
-4
0
r.0'
A'4
-4
-4
4v
41)
41
~ to.441
V0
(W
0)I
*UNI
0
10'0
) r-4
04 w
.0
'0
41m1
0
N
0
E4 in0
.4
.-4
N
0
C;0
0
0
.4
N
000
00
0
Il
)0
0D
M
.4
in
Nn
N
-4'4
0
0
0
0
00
+
+
+.
14
.-4
-4
-4
41
o
41.
0O
CO-U
0
iv1.
+
+
+
(0
+
'4
'-4
.-I
1.4
to
EnCOCO
104
REFERENCES
1.
Sacks, R.A. and Darling, D.H., Nuclear Fusion, L., 44 (1987).
2.
Williams, J.M., Langmuir, 1, 44 (1988).
3.
4.
Unilever Research Laboratory, European Patent 60138, (Sept. 3, 1982).
Pekala, R.W. and Kong, F.M., UCRL-99846 (1988).
S.
Nyitray, A.M. and Willliams, J.M.,
(1989).
6.
Williams, J.M. and
Polymer (1989).
7.
Kong, F.M., Polymer Preprints, LO, 258 (1989).
8.
Nyitray, A.M.,
J. Cellular Plastics,
Wilkerson, M.H.,
J. Vac.
Sci.
Technol.,
accepted
5, 217
for publication
in
submitted for publication,
1989.
9.
Fricke, J., in Aeropels, (Springer-Verlag 1985), p. 94.
10.
Allitt, M., Whittaker, A., and Onn, D., to be published in Thermal
Conductivity (21), Plenum Press, NY.
105
SYNTHESIS AND PROPERTIES OF COPOLYMERS OF
DIPHENYLSILOXANE WITH OTHER ORGANOSILOXANES
a
J. Ibemesi , N. Gvozdicb, M. Kuemin', Y. Tarshianid and D. J. Meier*
Michigan Molecular Institute, 1910 W. St. Andrews Rd., Midland, MI 48640
ABSTRACT
We describe the synthesis, characterization and properties of various
types of siloxane polymers containing diphenylsiloxane (P) as a component.
The polymer types include di-and tri-block copolymers with dimethylsiloxane
(M) as the second component, and random and statistical copolymers with
dimethylsiloxane or methylphenylsiloxane (P/M) as the second component. Such
copolymers combine siloxane units whose polymers have very different
properties. The polydiphenylsiloxane chain is rigid and inflexible, and the
polymer is a highly crystalline solid with a liquid crystalline or condis
crystalline state and a very high melting (clearing) temperature. In
contrast, the polydimethylsiloxane or polymethylphenylsiloxane chains are
very flexible and the polymers have very low glass transition temperatures.
Polymers of controlled molecular composition, size and architecture
were prepared by anionic polymerization of the "cyclic trimers", using
lithium-based initiators.
The physical properties of the copolymers vary dramatically with
composition and architecture. Two types of "random" copolymers can be
prepared. In one type, siloxane units of a given type are randomly placed in
the chain in groups of three, i.e., the minimum sequence length of a given
siloxane type is three siloxane units. In the other type of random
copolymer, individual siloxane units are randomly distributed so that the
minimum sequence length is a single siloxane unit. The properties of the two
types are quite different, showing that subtle changes in sequence
distribution can have major effects on physical properties. At molar ratios
near 1/1 and with molecular weights of -105 , the first type of "random"
copolymer is an elastic solid with appreciable mechanical properties,
whereas the latter type is a sticky gum.
Diblock copolymer (P-M) with dimethylsiloxane as the major component
are paste-like, whereas the triblock (P-M-P) and star-block copolymers of
the same composition are tough elastomers. The block copolymers are
molecular composites, in which the polydiphenylsiloxane component separates
into crystalline microphases with very uniform fibrillar or lamellar
morphologies, and with widths or thicknesses comparable to the length of the
polydiphenylsiloxane block, i.e, typically of the order of 100 A.
INTRODUCTION
A program has been underway at MMI for several years to investigate
the synthesis and properties of homopolymer and cnpolymers of diphenylsiloxane with other organosiloxanes. A wide variety of homopolymers and
copolymers have been prepared by ring-opening polypmerization of the "cyclic
trimers" e.g., hexaphenylcyclotrisiloxane and hexamethylcyclotrisiloxane,
using lithium-based anionic polymerization techniques, as first described by
Bosticl. Polymers of controlled size and molecular architecture can be
prepared by these techniques since, when propertly used, lithium-based
Met. A*$. sec. Symp. Proc. Vol.
171. 019O Materials
Resewth Society
106
initiators do not "scramble" or randomize the sequence distributions of
resulting copolymers. Polydiphenylsiloxane is of interest since the polymer
is highly crystalline, with a liquid crystalline (or condis) transition near
250°C and a clearing temperature above 500'C. The polydiphenylsiloxane chain
is quite rigid in contrast to the highly-flexible polydimethylsiloxane or
polymethylphenylsiloxane chains.
A variety of block copolymers of the di-block (P-M), triblock (P-M-P)
and star-block, (P-M)n-x types have been made, where "P" and "'M"represent
polydiphenylsiloxane and polydimethylsiloxane blocks, respectively, and "x"
represents the common junction of n-arms of a star-block molecule. With
appropriate block compositions, the tri- and star-block copolymers are
highly elastic, with strength properties among the highest ever reported for
2
a siloxane elastomer .
Random and statistical copolymers are of interest since they
demonstrate in a striking way the influence that subtle differences in
molecular architecture can have on physical properties. The term
"statistical copolymer" is used here to designate polymers prepared by the
lithium-initiated copolymerization of the cyclic trimers. The resulting
copolymers are "random", but the randomness is for groups of three siloxane
units of a given type, i.e., the siloxane moities appear in the chain in
minimum sequence lengths of three. With the polymerization conditions used,
lithium-based initiators do not scramble the units of each cyclic trimer
molecule. In contrast, truly random copolymers can be prepared using other
5
types of initiators and monomer types, e.g., KOH with the cyclic tetramers .
With such initiators, the siloxane moities are completely scrambled and
hence appear in the chain in minimum sequence lengths of one. Although the
difference between minimum lengths of one unit vs. three units would appear
to be minor, the differences in the physical properties of the two types of
"random" copolymers are very large, as will be shown.
EXPERIMENTAL
As mentioned, the various siloxane polymers are prepared using
lithium-based anionic polymerization techniques to initiate the
polymerization of the cyclic trimers (obtained from HUls Petrarch Systems
Inc., Bristol, PA). The sensitivity of anionic polymerizations to
adventitious impurities such oxygen, moisture, polar substances, etc.
requires careful precautions to purify monomers and solvents and to exclude
air during the polymerizations. Hexaphenylcyclotrisiloxane (HPTS) was
purified by repeated recrystallization from toluene, hexamethylcyclotrisiloxane (HMTS) by vacuum sublimation at room temperature, and triphenyltrimethylcyclotrisiloxane (TPTMTS) by distillation. Trivinyltrimethylcyclotrisiloxane (TVTMTS) was used as received. Diphenylether (DPE) and
tetrahydrofuran (THF) used as solvents in the polymerization were purified
by distillation over n-butyllithium (DPE) or LiAlH4 (THF). The polymerizations were conducted under a positive-pressure of purified nitrogen.
Diblock P-M copolvmera
The diphenylsiloxane-dimethylsiloxane diblock copolymers are prepared
by polymerizing the diphenylsiloxane block first and then followed by
107
polymerization of the dimethylsiloxane block. This sequence, as shown in
Figure 1, allows the dimethylsiloxane block to be terminated with reactive
end-groups, e.g., vinyl groups, for post-polymerization coupling reactions
to produce star-block copolymers.
FIGURZ 1
diphenyi block
,0
/SI \
U p-J0
Si
Si
(25
I/
Hexaphenylcyclo-
X in DPE
0Sl Li*
HMTS/s-BuLl
1SO-160*C Initiator
trlsiloxno
(HPTS)
,0 ,'
Sc.S Sl"c
THF
(HIITS)
moth)_vny
metiil-vIny-I
"blockCH
0
-0.
-R,
n
divmethyl
'0'
c
'Si.
S
c
C
q1 L.3 H
block
Cc
"L
'". 'S'
" 3-25
10
I
0
.- n
I
L
-CH3M
C
-l..
[,l[,A1
C-
0
n0-500
SYNTHESIS
n ..
H3
-
m
C14
3
.
CH 3
3-25
S
0-5000
OF VINYL-TERMINATED
DIBLOCK P-N COPOLYMERS
Polymerization of the diphenysiloxane block is done at 150 - 180oC in DPE
solution at concentrations below 25% to avoid phase separation during the
polymerization. Lithium dimethylbutylsilanolate, which is formed by
reacting s-butyl lithium and HMTS in a 1/1 molar ratio (Li/Si), is used as
2
the initiator . The silanolate reaction product is required for initiation
of the polymerization of HPTS, since, in contrast to HMTS, s-butyl lithium
will not initiate the polymerization of HPTS. A small concentration (e.g.,
0.2-0.4 molar) of a polar additive such as tetrahydrofuran (THF) is
108
required as a polymerization "promoter" for polymerization of the
diphenylsiloxane block. The polymerization of the diphenylsiloxane block
is complete in approximately two hours, after which the solution is cooled
to 70*C and a solution of HMTS in THF (20 w/v %) is added. Polymerization
of the polydimethylsiloxane block also requires approximately two hours
for completion. The reaction is terminated by the addition of
vinyldimethylchlorosilane to provide a single terminal vinyl group, or,
alternatively, trivinyltrimethylcyclotrisiloxane can be added and
polymerized to give a short terminal methylvinylsiloxane "blcck" with
additional terminal vinyl groups. Typically only a few (3-20) methylvinylsiloxane units are added to each polymer chain The polymers are
recovered and impurities and unreacted monomers are eliminated by repeated
washing with acetone and methanol.
Triblock P-M-P Coolvmers
The triblock polymers P-M-P are prepared in the manner first described
by Bosticl. The polymerization sequence is shown in Figure 2.
FIGURE 2
LI
o, ,-S--o
3
CM
I
,,-u
3 CH'
ct.
Sl Sk
CH;
0,
rr.
L
j
Lr
150"C, 1.5 hrs.
In THF
CH,
'
O- CH-O
i-. - -o - Cli-O
' II °
~tI~
I
-
I
Hebamethllcuclo-
I
Qk .0
(HXrS)
CVc1ic trimor-)iI
0 0*C
1S
""
YI )a)-- 32hrs.
HPTS
CI
,<1
.6>l
r . 0t
eL
oJL
-"140-SI
SYNTHESIS
-.jI.-1t~rmination
* ..['J
0-Si-o-l--Si-
-
M
i-O
OF P-X-P TRIBLOCK COPOLYIURS
I
Li
109
A difunctional lithium silanolate initiator is used to polymerize the center
dimethylsiloxane block first, then followed by polymerization of the end
diphenylsiloxane blocks.
Statitical cnopolymers
4
Poly(diphenylsiloxane-stat-dimethylsiloxane) (P-s-M) and
poly(diphenyl-stat-phenylmethylsiloxane) (P-s-P/M) are prepared by
copolymerization of the cyclic trimers at 150-180*C, using 1/1 HMTS/s-BuLi
as initiator and THF or dimethylsulfoxide (DMSO) as promoters. The P-s-M
copolymers are prepared in DPE solution, while the P-s-P/M copolymers are
prepared in DPE solution or in bulk. The yield of polymer is typically 7590%. The reactivity of diphenyl cyclic trimer is slightly greater than that
of the dimethyl or methylphenyl analogues, as shown in Table 1 which
29
displays the feed and copolymer composition (determined by
Si NMR and
FTIR) of several poly(diphenylsiloxane-stat-dimethylsiloxane) copclymers.
It should be noted that the polymerization of the cyclic trimers to
form block and statistical polymers of controlled architecture and molecular
weights is the result of the polymerization being under kinetic control.
Under equilibrium conditions, the polymer is in equilibrium with cyclic
3
4
species and the polymer can be the minor component . Veith and Cohen have
Table I
STATISTICAL COPOLYMRS OF DIPHONLSILOXANE AND DIMXTHYLSILOXANZ
Sample
1
2
3
4
5
6
7
8
Promoter
THF
THF
THF
THF
THF
DMSO
DMSO
DMSO
Polymerization
Temperature (*C) Time
175
180
170
170
175
145
150
155
240
210
180
180
240
45
15
30
(min)
Diphenylsiloxane, mole %
Feed
Copolymer
27
57
72
80
88
20
30
40
30
60
76
86
93
33
36
46
Samples 1-5: 25-30% concentration in DPF, 0.2 molar THF promoter
Samples 6-8: 75% concentration in DPE, 0.2 molar DSMO promoter
'initiators
examined the kinetics of ring-opening siloxane polymerizations. Although the
initiation of polymerization of the cyclic trimers with lithium-based
minimizes the tendency towards equilibration, such reactions can
occur at higher temperatures when higher concentrations of promoters are
present. An example of the effect of higher concentrations of a promoter on
the polymerization of HPTS is shown in Figure 3. With a limited concentration of the THF promotor present, the yield of polydiphenylsiloxane
approached 80% and remained constant with additional reaction time. In
110
contrast, with an excess amount of THF present, the yield of polymer also
reached approximately 80%, but then rapidly declined as the system began
equilibrating, with conversion of the polymer to the cyclic tetramer as the
predominant product.
FIGURZ 3
Composition
"limited
"excess
THY"
THF"
HPTS 15 g
12.5 g
100
|DPE
limited THF
so8
THF
4S m.
50 Ml
2 ml
10 ml
60-
0
20
0
2
4
6
8
10
Reaction time, hrs.
KINETIC VS. EQUILIBRIUM POLYMERIZATION OF HEXAPI:ENYLCYCLOTRISILOXANE
Random Coonlymer
A random copolymers of diphenylsiloxane and phenylmethylsiloxane was
prepared by the ring-opening polymerization of the cyclic tetramers,
octaphenylcyclotetrasiloxane and tetraphenyltetramethycyclotetrasiloxane.
The polymerizations was conducted in a concentrated (85%) DPE solution at
1200C, using KOH as initiator. Since this polymerization is a equilibrium
polymerization, only a 35% yield of polymer (MW = 22,000) was obtained after
purification by repeated precipitation from toluene solution with MeOH and
heating in vacuum at high temperatures to remove the remaining cyclics by
sublimation.
CHARACTERIZATION
MMR and Ascuence
distrihution
Si MR data have been used to characterize the sequence distribution
29
of the random and statistical copolymers of 1i diphenylsiloxane and
phenylmethylsiloxane, and these data are shown in Figu:re 4, with assignments
5
of the various peaks as given by Babu, Christopher and Newmark and by
3
Ziemelis, Lee and Saam . These data show clearly that there are major
differences in the sequence distributions of the two polymer types. In
particular, the P12Pand DP triads, in which a central siloxane unit is
flanked on both sides by the opposite siloxane type, are present in the
random copolymer, but not in the statistical copolymer. These triads can be
formed only if the elements of the cyclic monomers are scrambled. Their
absence in the statistical copolymers then shows that the repeat unit in the
statistical copoymers has a minimum length of three siloxane units,
corresponding to the three units of the cyclic trimer.
FIGURE 4
diphenyl -phenyl/methyl siloxane
statistical copolymer
D0=PhMOSIO
2
h1
diphenyl-phenyllmethyl slloxane
random copolymer
b
=-PhMOSIO
b
b P=Ph 2 SIO
C
a=P2P
d=PEP
b=PQ2P
e=PED
aPl
=Q
b=DflO
=E
=E
d=DIEP
d
C
aa
35
30
PPM
40
45
50
-30
.3s
.40
.45
.50
1W"n
29S1 NNR OF RAN~DOMAMD STATISTICAL COPOLYTURS
OF
DIPHENYLS ILOXJISE AMD0 METHYLPHENYLS ILOX30OE
Molecular weights
As a result of the high crystallinty and high melting point of
polydiphenylsiloxane, polymers and block copolymers of it are soluble in
very few solvents and then only at high temperatures. This makes the
molecular characterization of such polymers quite difficult. However, we
found that it was possible to prepare solutions for analysis at moderat'o
temperatures hy pouring a dilute s~iution of the polymers in DPE (at 160"')
0
into chloroform at 50 C. Although the solutions are unstable and the r),)yrmer
112
will eventually crystallize and precipitate, the polymer will remain in
solution for approximately an hour, which is sufficient time for the
molecular weight and its distribution to be determined by gel permeation
chromatography (GPC). We find, in general, that the GPC molecular weights
(based on polystyrene standards) agree well with those expected from
stoichiometry, and that the molecular weight distribution can be quite
narrow, e.g., Mw/Mn - 1.1-1.2. Figure 5 shows GPC traces obtained during the
polymerization of a P-M diblock copolymer
FIGURE 5
precurser dlphonyl block
PDPS
phenyl/ethyl diblock
PM-DB
Uv
S~uv
chloroform solvent
551C
Mn = 83,000
MW= 111,000
Mn = 65,00
Mw 71,600
Mw/Mn = 1.2
Afw~ln= 1.1
RI
RI
Elutlon volume
Elutlon volume
GPC OF POLYDIPHENYLSILOXANE
AND
POLY (DIPHENYLSILOXANE-B-DIETHYLSILOXANE)
POLYMERS
Thermal propoerties
Thermal analysis of the block copolymers shows, as expected, the
characteristic transitions of the homopolymers, i.e., Tg at -225oC and Tm at
-40'C for the dimethylsiloxane block, and transitions near 260*C and 500'C
for the diphenylsiloxane block. We believe the 260'C transition is a
0
crystal/liquid crystal transition and the 500 C transition is the transition
to the isotropic liquid ("clearing temperature"). Liquid crystal formation
.
has also been reported for other polyorganosiloxanes 1. The glass
transition of polydiphenylsiloxane is not directly observable because of its
3 5
very high degree of crystallinity. Literature data
for the glass
transition behavior of random copolymers of diphenylsiloxane and dimethylsiloxane (prepared by ring-opening polymerization of the cyclic tetramers)
are not particularly informative either in establishing Tg for polydiphenylsiloxane, since depending on the extrapolation used, e.g., Tg or I/Tg vs
weight or mole %, the data can be extrapolated to give Tg of polydiphenylsiloxane between 13'C and 60'C.The variation between various sets of data
perhaps can be understood in terms of sequence distribution when it is noted
113
that the statistical copolymers of this investigation have Tg's approx-
imately 20*C above those of their random counterparts.
PHYSICAL PROPERTIES
Trihlock cooolvmers P-M-P
Tri-block copolymers of the P-M-P type (with polydimethylsiloxaneas
the major component) are thermoplastic elastomers which can be molded (with
difficulty) at very high temperatures (>300*C) . It proved to be more
practical to prepare samples for testing by casting films from DPE solution,
even though it was difficult to prevent flaws in the films from bubbles from
the high-temperature evaporation of the DPE solvent. Such films showed
tensile strength properties as high as 8 MPa and ultimate elongations above
600%. In addition, appreciable mechanical properties were retained at
0
temperatures as high as 150 C (at the highest temperature tested) . However,
in order to avoid the difficult processing of the P-M-P triblock copolymers,
2
we examined the end-to-end coupling of P-M-v diblock copolymers to form
tri- and star-block copolymers in situ. The diblock copolymers are soft
pastes which are easily processed, e.g., by compression molding or
extrusion, after which the reactive terminal vinyl groups on the
polydimethylsiloxane block allow the diblock copolymer to be coupled to give
the desired tri- and star-block copolymers.
Cuouled diblock copolvmers P-M-v
The effect of the number of terminal vinyl groups and of molecular
weight on the tensile properties of coupled P-M-v diblock copolymers is
shown in Table II.
Table I
TENSILE PROPERTIES OF COUPLED P-M-v COPOLYMERS
Sample
DB3
DB4
DB5
DB6
Vinyl
units
1
10
16
16
HPTS
mol %
20
20
27
27
MW
-3
x 10
102
70
72
150
Tensile strength
MPa
0.4
2.7
5.0
9.8
Elongation
%
195
90
70
195
The vinyl-specific peroxide 2,5 dimethyl-2,5 di(t-butylperoxy)hexane was
used for coupling through the terminal vinyl groups (3 phr, 15 min. at
170°C). The data in Table II show that the strength properties increase with
increasing numbers of terminal vinyl groups, most likely as the tesult of
increased coupling efficiency. The tensile strength of Sample DB6 approaches
the highest levels attainable with conventional silica-reinforced silicone
elastomers, in spite of the fact that these star-block copolymers are not
filled (in the conventional sense).
114
Since a silica filler is necessary in conventional silicone elastomers
to achieve appreciable strength properties, it was of interest to see if the
addition of a silica filler to our block copolymer systems would lead to a
further enhancement of properties. Various amounts of a hydrophobic silica
(Cab-O-Sil N70-TS) were added by mill mixing to a P-M-v block copoymer (70K1IOK-2.SK). After coupling (3.5 phr of 2,5 dimethyl-2,5 di(t-butylperoxy)hexane, 15 min. at 170*C), it was found that the mechanical properties were
improved, and the improvement maximized with the addition of about 20 phr of
silica. At this concentration the tensile strength increased from 4 MPa to
8.5 MPa and the ultimate elongation increased from 400% to 650%. The 100%
modulus remained more-or-less constant until the silica concentration was
above 20 phr. Additional experiments are required to determine if similar
improvements will occur when the polymer itself has inherent strength
properties comparable to the values achieved here with added filler.
SUMMARY
Table III sunnarizes our results on the effect of molecular
architecture on the character of siloxane homopolymers and copolymers.
Table III
SILOXANK POLYMERS
P = DIPHENYLSILOXANE
M = DIMETHYLSILOXANE
MW = 100,000
POLYMER
P homopolymer
H homopolymer
P-H random copolymer (30-70)
P-M statistical copolymer (30-70)
P-H diblock copolymer (30-70)
P-M-P triblock copolymer (15-70-15)
15 35
(P-H-)n starblock copolymer ( - -)n
CHARACTrR
rigid, crystalline solid
viscous liquid
sticky gum
flexible elastic solid
paste-like
"snappy" tough elastomer
"snappy" tough elastomer
ACKNOWLEDGEMENTS
We wish to thank Dr. J. C. Saam of the Dow Corning Corporation for his
helpful advice, and to thank the Lawrence Livermore National Laboratory and
the Dow Corning Corporation for their support.
PRESENT ADDRESSES
a. University of Nigeria, Nsukka, Nigeria
b. Dow Corning Corporation, Midland, MI 48686
c, Dow Chemical Europe, Norgen, Switzerland
d. Silar Optical, St. Petersburg, FL 33709
* To whom inquiries should be addressed
115
REFERENCES
1. E. E. Bostic, ACS Polymer Preprints, i,
877
(1965).
2. J. Ibemesi, N. Gvozdic, M. Kuemin, M. J. Lynch and D. J. Meier,
Polymer Preprints 2.f, 18
3. M.
Ziemelis, M. Lee and J. C. Saam, ACS Polymer Preprints, 31,
in press.
4. C. A. Veith and R. E. Cohen, J. Poly. Sci:
1241
ACS
(1985).
Part A:
No.1,
Polymer Chemistry, 22,
(1989).
5. B. N. Babu, S. S. Christopher and R. A. Newmark, Macromolecules 2Q,
1987).
2654
6. C. L. Beatty, J. M. Pochan, M. F. Froix and D. D. Hinman, Macromolecules
., 547 (1975).
7. Yu. K. Godovskii, N. N. Makarova and N. N. Kuz'min, Polymer Science
U.S.S.R., 30, 341 (1988).
117
DYNAMIC IR STUDIES OF MICRODOMAIN INTERPHASES
OF ISOTOPE-LABELED BLOCK COPOLYMERS
I. NODA, S. D. SMITH, A. E. DOWREY, J. T. GROTHAUS, and C. MARCOTT
The Procter & Gamble Company, Miami Valley Laboratories, P.O. Box 398707,
Cincinnati, OH 45239-8707
ABSTRACT
By probing the localized segmental motion of isotope-labeled block
copolymers, the physical nature of the interphase region between microphaseseparated domains of block polymers was examined. Dynamic infrared linear
dichroism (DIRLD) spectroscopy, which measures the reorientations of submolecular structures induced by a small-amplitude oscillatory strain, was combined with specific isotope-labeling using deuterium-substituted monomers.
The latter technique enabled us to differentiate the dynamic responses of
well-defined parts of block segments, e.g., near the segment junction, chain
end, or middle of the block. The degree of segmental interactions near the
interphase region of styrene-isoprene diblock copolymers were studied as a
function of the segment location and temperature.
The reorientational
motion of the polystyrene segment, especially near the block junction, was
monitored around the glass transition temperature of the polyisoprene
matrix.
From this result, the degree of segmental mixing in the interphase
region which leads to local plasticization of the polystyrene segment was
determined.
INTRODUCTION
Block copolymers owe their unique properties to the molecular architecture consisting of different polymer segments joined together by a covalent bond.
The repulsive interactions between dissimilar block segments
often results in microphase separation where the size of the phase domain is
restricted to a scale comparable to the block segment length. It has been
postulated for some time that the boundary between the adjacent microphase
domains is not a sharp two-dimensional layer but rather a region of finite
thickness characterized by a substantial intermixing of different block segments.
The existence of the microdomain interphase [1] due to the diffuse
concentration gradient across the boundary has been predicted by statistical
thermodynamic theories based on the mean-field approach [2,3). The results
of several experimental works (e.g., the systematic deviation of SAXS intensity profiles from the behavior of sharp-boundary systems described by
Porod's law [4,5] and the modeling of rheological behavior measured by dynamic mechanical analysis [6]) support the view of a segmentally mixed interphase.
Different models, such as a coarse interface with a sharp boundary
[7], may also account for some of the observed results.
Spectroscopic techniques which can provide direct information at a
molecular scale are especially suited for probing the presence of segmental
mixing at the interphase. Recent NMR studies [8] of isotope-labeled block
copolymers with relatively low molecular weights, for example, reveal the
existence of substantial segmental interactions at the domain boundary. We
used a recently developed analytical technique called dynamic infrared
linear dichroism (DIRLD) spectroscopy in conjunction with the selective
deuterium labelling of well-defined portions of the block segment to investigate the submolecular environment of microphase separated block copolymers.
Mat. As. Soc. Symp. Proc. Vol. 171. 11990 Materials Research Society
lie
BACKGROUND
DIRLD spectroscopy [9,101 is a rhea-optical polymer characterization
technique based on the combination of dynamic mechanical analysis and IR
dichroism
spectroscopy.
In this
technique, a polymer sample is deformed with a small-amplitude oscilOTh|S$
IR
latory strain, and strain-induced
dichroism (i.e., difference in absorAtlE
,
to the applied
bances
the directions
paral-A
perpendicular
and between
lel
leland
erpeentdiareton
heaprle
strain) is monitored with polarized
light as a function of time, temperature, and IR wavenumber (Fig. I).
Dynamic
dichroism induced by the
applied strain is directly related
to the local reorientational motions
of polymer molecules. Thus, DIRLD
spectroscopy can be used as a powerful tool to probe the detailed local
dynamics of polymer chains.
OSCU.I.UATW
IT
GAUGE
VT:AD
OMER
Fig. 1. Schematic representation of a DIRLD spectrometer.
For a small-amplitude sinusoidal strain, (t)= Z sin wt, the dynamic
variation of dich-oic difference AX(v, t) is given by
&(v, t) = AA'(v) sin wt + AA"(v) cos wt
(1)
where the wavenumber dependent terms, WA(v) and AA"(v), are referred to as
the in-phase and quadrature spectrum of dynamic dichroism. These spectra
represent, respectively, the components of dynamic dichroism proportional to
the extent and rate of applied strain. From a pair of such DIRLD spectra,
strain-induced reorientations of various functional groups contributing to
individual molecular vibrations can be determined. Figure 2 shows a typical
set of DIRLD spectra.
A positive peak indicates the dipole-transition
moment associated with the band is reorienting parallel to the strain direction, while negative peaks indicate perpendicular reorientations.
2 .5
4 . 6.
2.3-
2.6
4
In-phase
*..
x--
ma
----
Quadrature
--- ---
.
1.5
< - 2 .e _0.
Asymmetric•
-4.6
Phenyl
ring
..........
..
-6..
2356
CD2 stretching
Symmetric
2266
9.5
6
2s6
Wavenumber
Fig. 2. D[RLD spectra of deuterium-substituted atactic polystyrene
obtained under a 23-Hz oscillatory strain at room temperature.
119
EXPERIMENTAL SECTION
Polymer synthesis:
A series of styrene and isoprene block and homopolymers were synthesized by anionic polymerization. Cyclohexane (Burdick
and Jackson HPLC Grade) was degassed and passed through columns containing
activated alumina and molecular sieves. Dibutyl magnesium and s-butyl lithium were used as received from Lithium Corp. of America. Styrene (Aldrich)
was purified by titration with dibutyl magnesium then passed through an activated alumina column under an inert atmosphere to remove magnesium salts.
Deuterated styrene-d8-(98%) (Cambridge Isotopes Labs) was purified similarly
to styrene. Isoprene was provided by the Goodyear Tire & Rubber Company.
The reactions were carried out in Chemco reactors as described in
Ref. 11.
A reactor at 601C was charged with the appropriate amount of styrene and cyclohexane.
This was titrated slowly with s-butyl lithium till
the first yellow persistent color was obtained. The calculated charge of
butyl lithium was then added to obtain the desired molecular weight. The
second purified monomer was added after one hour to ensure complete conversion.
By these steps block copolymers were synthesized consisting of styrene, deutero styrene, and isoprene blocks. The copolymers were terminated
with degassed isopropanol and stabilized with Irganox 1010. The molecular
weights and molecular-weight distributions of polymers were analyzed by GPC
(Ultra Styragel columns ranging from 10' to 106 A porosities in THF). Compositions were determined using a combination of 'H and 13C NMR (GE QE300).
Films of these copolymers were prepared by dissolving an appropriate amount
of copolymer in toluene and allowing this solution to slowly evaporate from
a Teflon mold. After being allowed to air dry for 48 hours, the films were
vacuum-dried at room temperature for 12 hours then annealed under vacuum for
0
8 hours at 125 C.
The sample-film thickness was selected so that the IR
transmittance for the phenyl band at 2280 cm ' was approximately 10%.
DIRLD Analysis:
Dynamic dichroism data were obtained with the timeresolved IR spectrometer described in Ref. 10. The IR dichroism measurement
was carried out by mechanically perturbing the sample films at specified temperatures with an oscillatory tensile strain (ca. 0.1% amplitude and 23-Hz
frequency) and recording the time-dependent fluctuations of directional IR
absorbances induced by the perturbation at a spectral resolution of 8 cm -1 .
The effect of small variations in sample thickness and deformation amplitude
were corrected by normalizing the results to a condition of 0.1% strain amp- .
litude and 10%
IR transmittance for the phenyl ring band at 2280 cm '
All spectra were plotted on the same scale as Fig. 2 except for the room
temperature traces on Figs. 4 and 6 which were expanded by 20X.
RESULTS
The in-phase and quadrature DIRLD spectra and normal IR absorbance
spectrum of perdeuterated atactic polystyrene (PS) obtained under an oscillatory deformation have already been shown in Fig. 2. Peaks located at 2100
-I
and 2180 cm
are assigned, respectively, to the symmetric and asymmetric
CD 2 -stretching modes of the PS mainchain backbone. The peak near 2280 cm -'
is attributed to the CD-stretching vibrations of phenyl ring side groups.
The dipole-transition moments for the CD 2 modes are oriented perpendicular
to the backbone of the polymer chain; they are also oriented orthogonal to
each other.
The sign of the two DIRLD peaks for the CD2-stretching vibrations are both negative, indicating that the electric dipole-transition
moments associated with the peaks are both reorienting in the direction perpendicular to the sample-stretch direction. Thus, the molecular chain of PS
is reorienting parallel to the direction of the applied tensile strain.
P.
--
-~~~
120
Another dynamic dichroism peak, associated with the reorientation of
phenyl groups attached to the main chain, is observed near 2280 cm '. The
positive sign of this peak indicates that the reorientation of dipole-transition moments of phenyl ring CD-stretching vibrations of perdeuterated PS occur predominantly in the direction parallel to the applied strain. The complete identification of side group reorientation direction, however, is not
straightforward, since the peak consists of several overlapping bands corresponding to different types of CD-stretching vibrations of phenyl rings.
Significant changes are observed for DIRLD spectra when the temperature is raised above the glass-to-rubber transition (Tg) of PS. Figure 3
shows a comparison of DIRLD spectra for perdeuterated PS in the glassy (a)
and rubbery (b) state. In going from the glassy to the rubbery state, the
magnitude of dynamic dichroism is decreased for all peaks, and the sign of
the peak associated with the reorientation of phenyl rings near 2280 cm '
is inverted.
The dipole-transition
moments of phenyl groups in rubbery
PS are predominantly reorienting in
the direction perpendicular to the
PS 30C
applied strain.
Such shifts of the
local orientational motions of functional groups are believed to be a
consequence of a change in the sub(b)
PS 120
molecular
e.g., the
environment,
local free volume accessible to the
reorienting groups. Thus, the characteristic inversion of the dynamic
aSI
2296
2352
dichroism peak for the phenyl groups
Wav~enuber
above and below Tg may be used as a
molecular level indicator of the glass
Fig. 3. DIRLD spectra of pertransition phenomenon.
deuterated PS below and above Tg.
0
Figure 4 shows the CD-stretching region of DIRLD spectra at -70 C and
0
30 C for a 1:3 homopolymer blend of perdeuterated PS and hydrogenous polyisoprene (PI) and for a microphase-separated diblock copolymer made of a
perdeuterated PS segment and hydrogenous PI segment. The molecular weights
of the blend components are about 100,000 each, while those of PS and PI
blocks are 50,000 and 150,000 respectively. For both systems, macro- or
microphase-separated domains of PS components are dispersed in a continuous
PI matrix. Because of the selective deuterium substitution, the DIRLD spectra in this wavenumber region show only the dynamic reorientaLion of PS components.
PSIPI blend
(c)
0
(.1
(d)
(b)~
0
30 C
PSIPI blend -70 C
~
lc
~~
ooyu, II 1
23562266235
Havenumbue r
Sl Block
0
a
3
2206
Wvenumber
0
Fig. 4. DIRLD spectra at -70 C and 30'C for
a homopolymer blend and diblock copolymer.
o
|opolymer
300C
-_----
2056
121
At -701C, the basic feature of DIRLD spectra for both blend and block
copolymer is very similar to that of the spectra for pure PS at room temperature (Fig. 2).
DIRLD spectra for these heterogeneous systems all exhibit
characteristic glassy responses with positive peaks for phenyl groups. Such
results are expected since PS and PI phases at -701C are both in the glassy
state.
Consequently, the applied strain is evenly distributed within the
system as in pure glassy PS. At room temperature, on the other hand, deformation should occur predominantly in a rubbery PI phase. As long as the
dispersed PS phase remains rigid, no significant orientational motions of
segments in glassy PS domains should be expected. Interestingly, while the
DIRLD spectra of the PS/PI blend at room temperature (Fig. 4c) have no significant peaks to indicate the existence of PS segmental motion, spectra of
the SI block copolymer (Fig. 4d) clearly show peaks associated with dynamic
reorientation of the PS segments. The sign of the dynamic dichroism peak
for the phenyl ring in Fig. 4d is negative, which is a characteristic of the
response of rubbery PS, For pure PS, such a response can be observed only
at temperatures well above 1001C.
This result shows that, even at room
temperature, a portion of the PS segment apparently is in a plasticized
state where the chain can move around in a manner similar to those at temperatures above Tg.
The response of the PF segment of the SI block copolymer at room temperature could be ex"-aI ed by the possible existence of a diffuse interphase region between -.. jdomains where substantial mixing of PS and PI segments is believee t
take place. Such mixing can plasticize PS segments,
causing them to
(
.&ave like rubber. To test this hypothesis, two types of
selectively segment-labeled block copolymers consisting of PS and PI blocks
(block MW oi 50,000 and 150,000) were studied. The PS block is further
divided into two distinct subsegments: deuterium-substituted and hydrogenous
portions.
A deuterium-labeled subsegment (MW of 10,000) is placed either at
the tail end of the PS block (Fig. 5a) or near the junction with the PI
block (Fig. Sb).
While the part of the PS segment at the tail end of the
copolymer may concentrate at the interior of the microdomain (Fig. 5c), the
part near the block junction is expected to be located close to the domain
boundary region (Fig. 5d). Thus, if plasticization of the PS segment occurs
in the diffuse interphase region between microdomains, the molecular mobility of the segment near the junction is expected to be much higher than the
tail end.
Tail-labeled
(c)
Junction-labeled
(d)
Fig. 5. Tail-labeled (a) and junction-labeled (b) SI diblock copolymers
and their spatial distributions in the microphase-separated domains. The
deuterium-labeled portion of each PS block is indicated by (*").
122
The DIRLD spectra of the selectively segment-labeled SI diblock co0
polymers at -70 C (Figs. 6a and b) show typical glassy response of PS. For
the junction-labeled block (Fig. 6d), a rubber-like reorientation response
similar to that of the fully labeled block (Fig. 4d) is observed at room temperature.
Thus, PS segments near the junction appear to be plasticized,
most likely by intermixing with PI segments. DIRLD signals for the taillabeled block (Fig. 6c), on the other hand, virtually disappear at room temperature. The core of the PS microdomain, where the tail ends of the block
segments are concentrated, remains in the rigid glassy state. Unlike the
rubbery PI matrix or plasticized interphase, significant segmental reorientations do not take place in this part.
(c)
(.)
Tai-labeled
(d)
(b)
2358
__
TaIi-abeled
30'C
-700C
junction-labeed 30'C
Jubioelaald -100C
2269
Wavenumber
2SS.
2359
2206
2656
Wavenumbe.
Fig. 6. DIRLD spectra at -701C and 301C for
selectively segment-labeled SI diblock copolymers.
CONCLUSIONS
DIRLD analysis of PS above and below the Tg reveals the existence of
distinct reorientational molecular motions characteristic of rubbery and
glassy states.
Diblock copolymers consisting of PS and PI blocks show certain rubber-like DIRLD response of PS segments at a temperature well below
the Tg of normal PS homopolymer. By using selectively segment-labeled block
copolymers, it was found that the rubber-like behavior of the PS block at
room temperature takes place predominantly in the segment near the PS-PI
junction.
Meanwhile, the tail end of the PS block remains glassy. This
result is consistent with the model that the PS segment near the block
junction is located in a diffuse, segmentally mixed interphase region where
the PS segment is plasticized by PI, thus leading to rubber-like behavior.
14
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
D.F. Learry and M.C. Williams, J. Polym. Sci., Polym. Phys. Ed. 12, 265
(1974).
D. Meier, J. Polym. Sci., Part C 26, 81 (1969).
E. Helfand and Z.R. Wasserman, Macromolecules 2, 879 (1976).
W. Ruland, J. Appl. Crystallogr. 13, 34 (1971).
T. Hashimoto, M. Fujimura, and H. Kawai, Macromolecules 13, 1660 (1980).
G. Kraus and K.W. Rollman, J. Polym. Sci., Polym. Phys. Ed. 14, 1133
(1976).
W. Ruland, Macromolecules 20, 87 (1987).
W. Gronski and G. Stbppelmann, Polym. Prepr. 29 (1), 46 (1988).
1. Noda, A.E. Dowrey, and C. Marcott, in Fourier Transform Infrared
Characterization of Polymers, edited by H. Ishida (Plenum, New York,
1987), pp. 33-59.
I. Noda, A.E. Dowrey, and C. Marcott, Appl. Spectrosc. 42, 203 (1988).
J.M. Hoover and J.E. McGrath, Polym. Prepr. 27 (2), 150 (1986).
PART III
~Rigid-Flexible
I
I
5i
Systems
125
LIGHT SCATTERING STUDIES OF THE STATE OF DISPERSION IN
MOLECULAR COMPOSITES
Benjamin S. Hsiao*, Richard S. Stein and Silvie Cohen Addad + , Polymer Research
Institute, University of Massachusetts, Amherst, MA 01003.
Russell Gaudiana and Norman Weeks, Polaroid Corporation, Cambridge, MA 02139.
*Present address: Fibers Department, E. I. du Pont de Nemours Co. Inc., Wilmington,
DE 19880-0302.
+Present address: College de France, Paris, France.
ABSTRACT
Polymer solutions comprising stiff-chain polyester and flexible polysulfone were
examined via light scattering techniques. Results were analyzed using the SteinWilson extension of the Debye-Bueche theory, in which the correlation lengths due to
orientation fluctuations and mean-squared fluctuations of the molecular anisotropy
were obtained. For a molecular dispersion, the correlation length is small and a
function of concentration; as the anisotropy is attributed to the rod molecules.
Aggregation of rods is associated with an increase in the magnitude and size of the
density fluctuations, and a change in anisotropy fluctuations is dependent on the
degree of orientation correlation of the rods in the aggregate. Blends prepared by
solution casting were studied by a small-angle light scattering method. Results thus
far demonstrate that aggregates are present in most of the rod/coil composites
prepared.
INTRODUCTION
Molecular composites comprising stiff-chain molecules and flexible matrices
offer potential advantages over the conventional composites in terms of processing,
optical and certain mechanical properties. However, attempts to prepare molecular
composites with a true molecular dispersion have often been unsuccessful. This may
be explained by the theories of Onsager [1] and Flory [2], which stated that rod
molecules are not soluble in random coil polymers.
Preparation of molecular composites usually employs solutions of rods and
coils, where aggregation of rods may often be present prior to or during the making of
the composite. In this case, the light scattering technique can be used to characterize
the state of dispersion of the rod molecules in these solutions. Light scattering studies
of dilute solutions of isotropic molecules and of anisotropic molecules [3-5] are well
developed; however studies of concentrated solutions comprising rod and coil
molecules are rare. It is conceivable that one can use the Stein-Wilson theory [6]
extended from the Debye-Bueche approach [7] to determine the correlation functions
for both density and orientation fluctuations in these solutions.
In a binary solution of rod molecules, intensity measured from polarized light
scattering (Vv) is attributed to the density plus orientation fluctuations of the rods,
whereas intensity of the depolarized light scattering (Hv) is only due to orientation
fluctuations. For a solution comprising rod and coil molecules, intensity of Hv
scattering is mostly from the orientation fluctuations of rods, but not appreciably from
the isotropic coils, whereas intensity of Vv scattering is from the combination of density
and orientation fluctuations of both rods and coils. Therefore, the measurements of Hv
scattering reflect the state of molecular dispersion of the rods.
In many systems, the orientation correlation function f(r) can be represented by
the empirical exponential equation [8]
f(r)=exp(-r/a)
Mat. Res. Soc. Symp. Proc. Vol. 171. ' 1990 Materials Research Society
(1)
126
where "r" is the scattering distance and "a"is a correlation length. In the case of
randomly correlated assemblies of rods, intensities measured from Hv and Vv
scattering can be expressed as
IHV(q)=K<" 2 >(ao3/(1 +q2ao2)2 )
2
(2)
3
lv(q)=K[15<6 >(ar /(1 +q2ar2)2)+(4/3)<q 2>(ao 3/(1+q2ao2)2 )] (3)
where K is a constant, depending on the number of rod molecules, the scattered
volume and the wavelength of the light; <62> and <?12> are mean-squared fluctuations
of two average refractive indices and its anisotropy, respectively; ao and ar are
correlation lengths from the orientation and density fluctuations, respectively; and q is
the magnitude of the scattering vector. It is noted that a plot of I v(q) "1 /2 vs. q2 should
be linear with a (slope/intercept) equal to ao, and a plot of (Iw (q- (4/3)IHV(q))-1/ 2 vs. q2
should determine ar. This approach is referred to as the Debye-Bueche plot.
In this work, it is our intention to characterize the orientation correlation of rod
molecules in rod/coil solutions by using the depolarized light scattering technique. A
change in orientation correlation lengths will reveal information about the state of
dispersion of rods in these solutions. Additional work also includes the small-angle
light scattering s-udy of polymer films prepared from solutions with various rod/coil
ratios. The purpose of this study is to correlate the phase behavior of solid blends with
that of the concentrated solutions.
EXPERIMENTAL
Materials and Sample Preparation
The stiff-chain polymer is a wholly aromatic polyester containing substituent
groups on the aromatic rings. This material is synthesized by Polaroid Corp.,
ambridge MA [9], and has a chemical structure as shown:
C
CF3
+1
0F
0.
CF
3
1I
00
The substituting groups have improved the solubility in many solvents, including
tetrahydrofuran (THF) which was used in this work. This polyester is completely
amorphous with a glass transition temperature of 11 0OC, but exhibits thermotropic
behavior. No lyotropic behavior has been observed, even as the solution
concentration exceeds 50% (w/v), the maximum solubility in THF. Its weight-average
molecular weight Mw is 21800, determined by light scattering measurements on dilute
solutions. The flexible polymer is a polysulfone (from Aldrich Chemical), which has a
Mw of 30,000 and is also0 soluble in THF. The glass transition temperature of the
polysulfone is about 180 C. Solutions of various concentrations were prepared,
ranging from 0.1% to 30% (w/v), with different rod/coil ratios. Dry polymer films, a few
microns thick, were obtained by spin-coating of these solutions. These films were
annealed above the glass transition temperatures of both polymers to equilibrate their
phase behavior.
Instrumentation
A laser light scattering goniometer (Brookhaven Instrument Corp.,
Ronkonkoma, NY), was used to collect the scattering data in polymer solutions. This
127
instrument consists of a 15mW He-Ne laser, polarizer/analyzer and an automated
photomultiplier/goniometer, that is capable of measuring both polarized and
depolarized light scattering. Toluene, whose Rayleigh ratio and depolarization ratio
are known [10], was used as a calibrient. Absolute scattered intensity was calculated
after suitable corrections were made, which included the corrections for scattering
volume, reflection, refraction, absorption, multiple scattering and dark current
subtraction [11]. Multiple scattering was proven insignificant at the range of the
chosen concentrations. The small-angle light scattering device consisting of a twodimensional optical multichannel analyzer (OMA3, by EG&G PARC, Princeton, NJ), a
2mW He-Ne laser, and assemblies of optical components, was utilized to characterize
the polymer films. The detail of this device has been described before [12].
RESULTS AND DISCUSSION
The non-coplanar biphenyl unit coupled with the random configuration of
repeat unit dipoles in the chosen polyester has significantly increased the solubility in
THF. Though the solutions show evidence of the molecules being stiff (Mark-Howink
exponent is 1.1), no lyotropic behavior has been observed, even when solution
concentration approaches 50% (w/v). This may be attributed to the intermolecular
interactions which oppose the aggregation of anisotropic molecules and/or the
unusually large molecular diameter which may reduce the molecular aspect ratio below
the level requiring for the formation of an order phase. It is thought that this system
may offer a unique opportunity, that molecules are stiff enough to give reinforcement in
a blend, but permit solubility in a flexible coil matrix to form true molecular dispersion.
A phase diagram of the solutions containing stiff-chain polyester and flexible
polysulfone, determined by a laser cloud-point measurement, is shown in Figure 1. It
is noted that above the phase line, a miscible single-phase region is present, where
below this boundary, the immiscible multiple-phase systems occur. The phase
separation in high concentrations is expected, since the entropy of mixing is
unfavorable for mixtures of rods and coils. It is conceivable that although strong
repulsive interactions are present among the stiff molecules, the unfavorable entropy
of mixing prevents the molecular dispersion of these molecules at high concentrations.
THF
10
20
30
90
80
70
40
Polyester 10 20 30 40 50 60 70 80 90
60
PSF
Figure 1. Phase diagram of solutions of stiff-chain polyester and flexible polysulfone.
128
20
Polyester(3)/PSF( 1)
iS
,-
16,
~ ~~
Polyester(7)/PSF(l)
14
12
2
0
+
10-
+
"
Polyester
4
20
2
6
'a
1
Polyester Concentration (g/cc), .
Figure 2. Plot of the orientation correlation length vs. polyester
concentration at various rod/coil ratios.
Results from the Debye-Bueche plot of depolarized intensity (Hv) for the rod/coil
solutions inthe single-phase region are illustrated in Figure 2, where the orientation
correlation lengths are plotted against the polyester concentration at various rod/coil
ratios. These characteristic lengths represent the correlation of the orientation
fluctuations of the rod molecules, which reflect the state of anisotropic molecular
dispersion. An increase in the orientation correlation length indicates that the
molecular dispersion becomes unfavorable and the rods tend to aggregate. This is
seen in Figure 2, where the correlation length increases with polyester concentration
and also increases with the coil composition. The addition of coil molecules obviously
enhances the segregation of rods and coils; the entropy of mixing is so unfavorable
that it has hindered the molecular dispersion.
Similar results are found from the small-angle light scattering (SALS) study of
polymer films. Figure 3 shows photographs of the Hv scattering patterns at six
different rod/coil compositions. Inthis figure, the pure polyester exhibits a 0/90 cross
pattern, which has been reported before [13), whereas an increase in the coil content
changes this pattern to a 450 cloverleaf shape. The azimuthal dependence of the Hv
pattern, which may be due to the interactions between the anisotropic and isotropic
domains, indicates non-random orientation correlations which also suggests
aggregation. It was verified via Vv scattering, optical microscopy and DSC analysis
that phase separation occurred at the high coil content. Intensity obtained from Vv
scattering was always stronger than that from Hv (about 2 times), an indication of
phase separation, since aggregation of rods is associated with an increase inthe
magnitude and size of the density fluctuations. Such an observation was consistent
with the optical microscopic observation of isolated anisotropic domains. Results from
DSC showed two glass transition temperatures, again indicating two separate phases.
However a slight glass temperature shift suggested that a partial miscibility might be
present.
A particularly interesting case is the blend made of 7/1 rod/coil ratio. As seen in
Figure 3, the Hv scattering pattern of this blend remains a 0/90 cross shape, similar to
129
that of the polyester, which suggests that its orientation correlation resembles pure
polyester. In the polymer films, Hv scattering is angularly depdent, therefore the
previous equations derived for the random assemblies of r
no anger
applicable.
LCP
1/1
Nv
7/1
1/3
3/1
1/7
Figure 3. Photographs of Hv scattering of polymer films prepared
from various rod/coil ratio solutions.
1.0
0.90.8
S0.7-
0.6.
O0.5
0.4-
0.3
0.2
0.1
0.0
1.0
0.8
0.8
0.4
Polyester Composition
0.2
0.0
Figure 4. Depolarization ratio of total integrated intensity vs. polyester composition.
130
However, the state of molecular dispersion in these films can be estimated by the use
of total integrated intensity Q [14]. This is seen in Figure 4, where the depolarization
ratio (QHv/CvV) of total integrated intensity vs. the polyester concentration is
illustrated. It is noted that the depolarization ratio of the polyester and that of a blend
of a 7/1 rod/coil ratio are both about unity. This observation indicates that the
magnitude of orientation fluctuations dominate the scattered intensity and that of
density fluctuations is small, which implies phase separation may not occur in this
blend. Therefore, a blend of 7/1 rod/coil ratio may form a composite having a
molecular dispersion at the level of wavelength of the light.
CONCLUSION
In summary, solutions comprising stiff-chain polyester and flexible polysulfone
were characterized by light scattering techniques. The orientation correlation lengths
in these solutions were determined by the Stein-Wilson theory for depolarized light
scattering. It was found that orientation correlation lengths increased with
concentration as well as with addition of coil molecules. An increase in these lengths
indicates the molecular dispersion of rods is depressed and the molecular aggregation
is more favorable. The modification of the rod molecules with strong polarized
substituting groups has improved solubility in non-interactive solvents, indicating that
solubility in flexible polymer matrices may also be enhanced, though the effect of
unfavorable entropy of mixing is significantly greater. A polymer blend made of a 7/1
rod/coil ratio solution showed a possible molecular dispersion, at least at the level of
wavelength of the light, was determined by a small-angle light scattering method.
ACKNOWLEDGMENT
The authors would like to thank Polaroid Corporation for providing the materials.
Financial support of this work was provided in part by a grant from Polaroid
Corporation and Center of Excellence Corporation of the Commonwealth of
Massachusetts, and in part by the Division of Materials Research and the Engineering
Division of the National Science Foundation.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
L. Onsager, Ann. N.Y. Acad. Sci., 51, 627 (1949).
P. J. Flory, Macromolecules, 11, 1 19 (1978).
C. P. Wang, H. Ohnuma and'G. Berry, J. Polym. Sci. Polym. Symp., 65, 173
(1978).
0. Ying, B. Chu, R. Qian, J. Bao, J. Zhang and C. Xu, Polymer, 26, 1401
(1985).
0. Ying and B. Chu, Macromolecules, 20, 871 (1987).
R. S. Stein and P. R. Wilson, J. Appl. Phys., 33(6), 1914 (1962).
P. Debye and A. Bueche, J. Appl. Phys., 20, 518 (1940).
R. S. Stein and S. N. Stidham, J. Appl. Ph-s.,
1 42 (1964).
R. Sinta, R. A. Gaudiana, R. Minns and H. G. Rogers, Macromolecules, 20,
2374 (1987).
W. Kaye and J. B. McDaniel, Appl. Optics, 13. , 1934 (1974).
Brookhaven Instrum. Corp., Instruction Manual For Laser Light Scattering
Goniometer, 1985.
R. J. Taber, R. S. Stein and M. B. Long, J. Polym. Sci. Polym. Phys., 20, 2041
(1982).
S.Rojstaczer and R. S. Stein, Mol. Cryst. Liq. Cryst., 157, 293 (1988).
J. Koberstein, T. P. Russell and R. S. Stein, J. Polym.Sci. Polym. Phys., 17,
1719 (1979).
131
RECENT ADVANCES IN MORPHOLOGY AND MECHANICAL PROPERTIES OF
RIGID-ROD MOLECULAR COMPOSITES
STEPHEN J.KRAUSE* AND WEN-FANG HWANG**
*Dept. of Chemical, Bio, and Materials Engineering, Arizona State University, Tempe, AZ 85287
**Dow Chemical Co., Central Research Laboratories - Advanced Polymeric Systems, Midland M1 48674
ABSTRACT
Rigid-rod molecular composites are a new class of high performance structural polymers which have high
specific strength and modulus and also high thermal and environmental resistance. The concept of using a rigid-rod,
extended chain polymer to reinforce a ductile polymer matrix at the molecular level has been demonstrated with
morphological and mechanical property studies for aromatic heterocyclic systems, butnew naterials
systems and
processing techniques will be required to produce thermoplastic or thermoset molecular composites. Improved
characterization and modeling will also be required. In this regard, new results on modeling of mechanical
properties of molecular composites are presented and compared with experimental results. The Halpin-Tsai
equations from "shear-lag" theory of short fiber composites predict properties reasonably well when using the
theoretical modulus of rigid-rod molecules in aromatic heterocyclic systems, but newer matrix systems will require
consideration of matrix stiffness, desired rod aspect ratio, and rod orientation distribution. Application of traditional
and newer morphological characterization techniques are discussed. The newer techniques include: Raman light
scattering, high resolution and low voltage SEM, parallel EELS in TEM, synchrotron radiation in X-ray scattering,
and ultrasound for integrity studies. The properties of molecular composites and macroscopic composites are
compared and it is found that excellent potential exists for use of molecular composites in structural applications
including engineering plastics, composite matrix resins, and as direct substitutes ofr
fiber reinforced compo_,ilc;.
INTRODUCTION
Since the concept of a self-reinforcing molecular composite of rigid-rod and flexible coil polymer
components was first proposed by Helminiak et al. [1, 2],and first successfully applied by Hwang et al. 131,a
variety of candidate systems have been studied. The advantages of rigid-rod molecular composites over macroscopic
fiber reinforced composites are based upon the elimination of discrete fiber/matrix interfaces and upon the
intrinsically superior properties of the aromatic heterocyclic chemistry of the rod molecules. These advantages
include: higher specific mechanical properties; higher environmental and thermal resistance; and the potential for a
wider choice of processing options. Possible applications for molecular composites include: engineering plastics;
high performance fibers; composite matrix resins; and direct substitutes forfiber reinforced composites.
The choice of the reinforcing molecule for a molecular composite is critical, in order to maintain both high
aspect ratio (ratio of length to diameter) for efficient reinforcement and to have a molecule with inherently high
strength and stiffness. The rigid-rod aromatic-heterocyclics are a class of molecules which fulfill these criteria.
Table I list various high performance fibers. The rigid-rod molecule fibers, poly(p-phenylene benzobisthiazole)
(PPBT) (Fig. 1)and poly(p-phenylene benzobisoxazole) (PBO) (Fig. 1),
have uniaxial tensile strength and stiffness,
which significantly exceed the properties of other commercial fibers used for composites. The limiting factor of
application of organic fibers to composites, including PPBT. PBO, poly(p-phenylene terepthalate) (PPTA) Kevlar 149, and gel-spun polyethylene (PE) - Spectra 1000. is the low compressive strength, which is two orders
of magnitude less than graphite T300 or E-glass fibers. However, for reinforcement at a molecular level, PPTA and
PBO are superior since PE cannot maintain an extended chain conformation insolution and PPTA, with a relatiely
high persistence length, has intrinsic stiffness and strength that are much less than
the rigid-rod mtolccules.
After synthesis of the appropriate molecule, the critical factor inprocessitig a tiotecular coinpisitt (Cotmthe
solution to the solid state is that the rigid-rod reinforcing component be well dispersed and not phase separate from
the matrix component during any stage of solution processing, including blending, extrusion, and coagulation, and
during solid state consolidation. Phase separation of the components of a molecular composites can significantly
affect mechanical and thermal properties. A high aspect ratio (the ratio of length to diameter) of the reinforcing
phase must be maintained for efficient reinforcement, Phase separation will also limit thermal stability tothat of
the minimum of the individual homopolymer components. The mechanical properties of a molecular composite
are chiefly controlled by the composition, orientation, and dispersion of the reinforcing rod molecules and can be,
predicted by using appropriate models with a knowledge of the intrinsic properties of the homopolymer components
and their morphological arrangement in the material. Thus, a thorough description of morphology is necessar) for
assessing, understanding, and predicting ultimate properties of rigid-rod molecular composites.
Research Society
Mat. Res. Soc. Symp. Proc. Vol. 171. 1190 Materials
132
In this paper the study of molecular composite candidates with a variety of characterization techniques,
including optical microscopy (OM), electron microscopy (EM). and x-ray scattering will be discussed and new
techniques will be highlighted. Results will bediscussed with regard to models for traditional composite theory of
mechanical pirotiem, such as "shear lag- theory for short fiber composites, in order to compare experimentally
observed and theoretically predicted properties. Molecular composite properties will then be considered for
macroscopic systems t0 assess
the potential for molecular composites in engineering applications.
{
'CS>_
SD
&
PPBT
+~-(~1-PBO
+0~Ic-©-1-PDIAB
N
N
H
N--a
HlC
PPTA
OC
ABPBI
H
Figure 1.Chemical structures
of PPBT, PBO, PDIAB, PPTA, and ABPBI.
TABLE I
Mechanical Properties of High Performance Fibers
Material
Modulus
Predicted
Measured
GPa. (Msi)
GPa, (Msi)
Strength
Tensile
Compressive
GPa, (tsi)
GPa,(ksi)
Ductility Density
%etong.
(g/cc)
PPBT
PBO
30PPBT/70ABPBI
GraphitcT300
615 (88)
635 (90)
184 (26)
1100(157)
4.2
5.8
1.3
3.3
1.58
1.58
1.58
1.76
'TA
. Kcvlar 149
220 (32)
K - glass
I'F-Spectra 1000
328
363
120
237
(47)
(52)
(17)
(34)
186 (27)
76 (11)
300 (43)
174 (25)
(600)
(830)
(190)
(470)
0.027 (39)
0.021 (30)
.
2.9 (417)
1.6
1.6
1.4
1.0
3.5 (500)
0.033 (48)
1.2
1.47
2.0 (286)
2.0 (286)
2.5
2.6
3.0 (435)
0.017 (24)
0.97
133
MODELING OF MECHANICAL PROPERTIES OF MOLECULAR COMPOSITE
The ability of the Air Force Materials Laboratory (AFML), Polymer Branch to synthesize aromatic
heterocyclic molecules that were chemically similar, but configurationally different, led to thecreation of both the
inherent rigid-rod and the semi-flexible coil molecular architectures used in the first molecular composites 141.The
molecular composite concept was first tested by blending rigid-rod polyparaphenylene benzimidazole (PDIAB) and
semi-flexible coil poly 2,5(6)benzimidazole (ABPBI) (Fig. 1)in a dilute solution of methanesulfonic acid (MSA)
and casting thin films in a glass container placed in a vacuum evaporator [5f.The properties of this material are
listed in Table 2, along with those of other molecular composite systems. Although the properties of the
PDIAB/ABPBI films were somewhat enhanced over that of cast homopolymer ABPBI, especially when stretched
55%, they did not approach what could be expected from the intrinsic properties of the rigid-rod PDIAB, which
would probably be in the range of those for PPBT. as shown in Table 1.It was found from optical microscopy
(OM) and scanning electron microscopy (SEM) that the bright yellow opaque films showed that large scale phase
separation (I-5 jtm) had occurred during the casting process, resulting in a loss of potential for reinforcement at the
molecular level. This problem was addressed by developing an understanding of the phase behavior of rigid-rod
solutions and then applying this knowledge to the technology for processing for rigid-rod systems from the
solution to the solid state.
This work contained the earliest application of the "rule of mixtures" to molecular composites (5). The
"rule of mixtures" is a special case of the Halpin-Tsai equations, whi.h is a simplilication of"shear-lag" theory.
which models the modulus of short fiber reinforced composites 16. The 'rule of mixtures" predicts that the
modulus of a material with a uniaxially oriented reinforcing phase and is given by a parallel element model, which
states that the overall modulus is the sum of the volume fraction times themodulus of each component. In order
for the "rule of mixtures" to be applied, the Halpin-Tsai equations show that thereinforcing phase must have a high
aspect ratio (typically >100) to approach 100% efficiency, as given for the modulus of a continuous fiber reinforced
composite. Donaldson 171has applied short fiber composite theory (ie. shear-lag theory) to molecular composites
to predict the effect of the factors of rod length, concenuation. dispersion, and orientation variables on properties.
The phase behavior of rigid-rod homopolymer and blend systems was first elucidated by Hwang ctal. 131
by applying Flory's [8] theory of phase equilibria of rigid-rod polymers. This theory predicts that, when the total
concentration (C) of a rigid-rod polymer system in a solvent is above a critical concentration (Ccr), liquid
crystalline domains will phase separamte
from a homogeneous solution. Hwang et al. 13)calculated the ternary phase
diagram for a blend of rigid-rod poly(p-phenylene benzobisthiazole) (PPBT) and setni-filexible coil poly-2,5(6)
benzimidazole (ABPBI) as components in a dilute solution of MSA and used OM hot stage techniques to verify the
predictions. It was then necessary develop processing techniques for retaining the dispersion as the rod and coil
components in solution are processed into the solid state.
Molecular composite fibers were produced with the dry-jet/wet-spin technology used in processing of
poly-p-phenylene terepthalate fibers. Processing homogeneous of dilute solutions of PPBT / ABPBI "froze-in" the
dispersion of rod molecules in the matrix polymer to form molecular composite films and fibers (3). This process
of rapid immersion of the polymer acid solution into water was referred to as "quenching" or "coagulation". After
subsequent neutralization, drying, and heat treatment the 30 PPBT / 70 ABPBI films and fibers were optically
transparent and had excellent mechanical properties, as shown in Table 2. Any phase separation which occurred was
clearly below the resolution limit in OM,since the film was clear, but it was also 0ound
to be below the reslution
limit of the SEM of 20 nm. WAXS showed that any crystalline phase separation could be calculated to be <5nm),
Thus, morphology results demonstrated that the concept of a rigid-rod molecular comttposite had been achieved.
Application of the "rule of mixtures" toa 30 PPBT / 70 ABPBI spun fiber, based uAon the experimentally
measured modulus of PPBT fiber, also indicated that a molecular composite had been achieved (3). The problem in
extending the "shear-lag" or Halpin-Tsai theory from the macroscopic to the molecular level is the choice of the
value of the modulus of the reinforcing phase. The theoretically predicted value of the modulus of PPBT is 615
GPa, which is almost double the value of the highest Measured modulus of a macroscopic PPBT fiber 1241. It
seems that, for molecular composites, the basis for modulus predictions should be the theoretical modulus of the
rod molecules, rather than the experimental fiber modulus, since factors such as local misorientation, disorder, and
defects will degrade the macroscopic fiber modulus, but should not affect the modulus of rod molecules acting to
reinforce a material at the molecular level. If the jj3
lWtmodulus of PPBT is used with the "rule of mixtures"
then the modulus of the 30 PPBT / 70 ABPB1 spun fiber is about 60% of thevalue predicted theoretically (9).
This still approaches the value for an ideal molecular composite, although some reduction from the ideal could have
resulted frion
PPBT rod molecules which may nothave been entirely well enough oriented or dispersed. Improved
processing could improve the value.
134
TABLE 2
Mechanical Properties of Rigid-Rod Molecular Composite Candidate Systems
Material
Epedc
GPa, (Msi)
Emeasured
GPa, (Msi)
Tensile Str.
GPa, (ksi)
%elong.
30 PDIAB /70 ABPBI
*CastABPB
-
-
Castfilm(C>Ccr)
*Cast film (CCr)+ 55% Draw
-
-
30 PPBT / 70 ABPBI Blend
* PPBT fiber
* ABPBI fiber
* Spun fiber (C<Ccr)
* Spun fiber (C>Ccr)
*Extruded film (C<Ccr)
*Castfilm (C>Ccr)
30 PPBT / 70 ABPBT Blend
* ABPBTfiber
-
615 (88)
-
-
184 (26)
-1-
87(12)
-
-
-
-
1.0 (0.15)
3.1 (0.46)
9.7 (1.4)
328
36
120
1
88
1.1
(47)
(5.4)
(17)
(1.6)
(13)
(0.16)
0.080(12)
0.092(13)
0.24 (35)
15
3
3.3 (470)
1.1 (160)
1.3 (190)
0.31 (45)
0.92 (130)
0.035 (5)
1.6
5.2
1.4
13
2.4
5.6
-
36 (5.4)
1.1 (160)
5.2
184(26)
120 (17)
0.9 (130)
1.4
30 PPBT / 70 ABPBI Copolymer
* Spun fiber (C<Ccr)
* Cast film (C>Ccr)
184 (26)
- -
100 (15)
2.4 (0.35)
1.7 (250)
2.2 (32)
2.4
43
30 PPBT / 70 PPQ
* Spun fiber (C<Ccr)
184(26)
18 (2.7)
0.35 (50)
2.4
* Spun fiber (C<Ccr)
PPBT/ Nylon
*Nylon6
* 25 / 75 Spun fiber (C<Ccr)
* 30 / 70 Spun fiber (C<Ccr)
* 50/50Spun fiber (C<Ccr)
0 60/40Spun fiber (CcCcr)
30 PPBT / 70 PPOT-50
* Spun fiber (C<Ccr)
* Extruded film
5 PPTA / 95 Nylon
* Spun fiber (C<Ccr)
0.9
0.051 (7)
5.3
154(22)
185(26)
36 (5)
31 (5)
0.35 (50)
0.35 (50)
7.0
2.4
308(43)
370(52)
80 (12)
80 (12)
0.88 (116)
0.90 (130)
1.4
1.4
184(26)
87(12)
140(21)
103(14)
- 1.1 (100)
2.4
1.5
0.058 (250)
2.4
-
11 (.012)
1.7(0.26)
Molecular composite films of 30 PPBT/ 70 ABPBI had a morphology which showed that the rod
molecules were well dispersed in very small crystallites randomly oriented parallel to the surface of the film
suggesting that the film was a planar isotropic molecular composite. Mechanical property modeling used an
extension of "shear Lag" theory for planar isotropic orientation results in a value for in-plane modulus which is 3/8
of the value for uniaxial orientation (9). In Table 2 the properties of the 30 PPBT / 70 ABPBI film give a modulus
tol 88 GPa and tensile strength of 0.92 GPa. The modulus exceeds the theoretically predicted value by 30%. which
indicates that the film is a molecular composite, and also that improved modeling may be required to prcdict
properties on a molecular level. The film had a modulus closer to the theoretical than the fiber, probably because
randoxn rod orientation in the film improved the their dispersion compared to the uniaxially oriented fiber.
Until recently, it has not been possible to determine the efficiency of the reinforcing phase in a molecular
composite. Day et al. [101 have developed a technique using Raman spectroscopy of fibers in a stressed
135
macroscopic composite, from which it is possible to directly measure strain distribution along the length of the
fibers. They were then able to compare the results to those predicted by "shear lag" theory and found a good
correlation over a wide range of strains. Young [11] has extended the technique to molecules in a stressed 30
PPBT / 70 ABPBI molecular composite fim and were able to directly measure strain in rod molecules along the
macroscopic stress direction. This makes possible a comparison of macroscopic for a direct assessment of
reinforcement efficiency of the rod phase.
The discussion of mechanical property modeling of molecular composites with the Halpin-Tsai equations
has only dealt with the concept of requiring a high aspect ratio, eg. 100, for efficient reinforcement. However, the
required value of the aspect ratio for efficient reinforcement, according to "shear lag" theory for short fiber
composites, is strongly dependent on the ratio of the modulus of the fiber (Ef) to that of the matrix (Em) (7). The
term "reinforcement efficiency" refers to the ratio of the stress carried by chopped fibers in a short fiber composite
to the stress which could be carried by unbroken fibers in a continuous fiber composite. For example, for 90%
efficiency when Ef/Em = 5 the aspect ratio must be only about 30 compared to the case where, with Ef/Em = 100,
requires an aspect ratio of 400. This demonstrates that the problem of phase separation, which reduces the aspect
ratio, becomes more critical when a softer polymer is chosen for the matrix material in candidate systems.
OTHER MOLECULAR COMPOSITE CANDIDATE SYSTEMS
Although the PPBT/ABPBI system formed a molecular composite, it did not have a glass transition
temperature below the degradation temperature and could not be consolidated by traditional thermal processing
techniques. Because of the versatility in fabrication and use of a thermoplastic polymers, new blends and
copolymers have been synthesized and processed with the potential for achieving thermoplastic and thermoset
rigid-rod molecular composites. A variety of polymer systems have been examined as rigid-rod molecular
composite candidates. The properties of some of these systems are listed in Table 2. This section will provide
brief descriptions of some aromatic heterocyclic systems and also some of the thermoplastic and thermoset systems.
The earliest systems studied were all aromatic heterocyclic polymers consisting of components of rigid-rod
and semi-flexible coil molecules. As previously mentioned, the PDIAB / ABPBI blend was examined as a
molecular composite and, although properties of the matrix were increased somewhat (Table 2). they did not
approach the high values, expected of a molecular composite, due to phase separation [5]. The first successful
system, as previously discussed, was 30 PPBT / 70 ABPBI, first processed and characterized by Hwang et at. [31
and later studied by Krause et al. [9]. Another system was the blend of PPBT and poly-2,5(6) benzthiazole
(ABPBT) which achieved excellent property enhancement (Table 2) (121. Another system, a coil / rod / coil
triblock copolymer of ABPBI / PPBT/ ABPBI, as synthesized by Tsai et al. [131. was processed into a molecular
composite with correspondingly excellent properties (Table 2) [14]. This system demonstrated that excellent
properties could be obtained over a range of rod lengths, indicating that the minimum aspect ratio had been
achieved with the shortest rod components. Because these systems do not have a glass transition temperature below
the degradation temperature and cannot be consolidated by thermal processing techniques they are intractable and can
only be used in the form in which they have been processed. The need for improved processability could be
answered with thermoplastic matrix systems.
The first efforts to develop a thermoplastic molecular composite system at the AFML were reported by
Hwang et al. [15] for a nylon / PPBT blend which was also examined later by other researchers [16-18]. Table 2
shows that the properties of the neat matrix resin are enhanced, by a factor of 30 to 80, but are still only about a
quarter as much as the PPBT/ ABPBI systems. These lower-than-expected values are due to a number of factors
including the need for a higher aspect ratio because of the higher value of Ef/Em, reduced chemical compatibility
which would enhance phase separation (19), and some probable misorientation. PPBT / PPQ is another
thermoplastic blend examined by the AFML was, but, here also, only limited property enhancement was achieved
due to phase separation. Another system studied was a blend composed of PPBT / polyetheretherketone (PEEK),
and it was found that the molecular weight of the PEEK ws so low that it readily phase separated during
processing. Overseas, another block copolymer of PPBT / PPO was synthesized and processed into a material
which had achieved properties predicted for a molecular composite [15], but still was not thermally processable, so
the search for an appropriate system continues.
Other concepts to achieve a molecular composite system have also been examined. A thermoset blend
composed of PPBT and benzocyclobutnne has also been examined, but large scale phase separation occurred during
processing, resulting in correspondingly poor mechanical properties. Another system under study is the "in-situ"
rod molecular composite, in which a rod molecule is formed by a ring-closing reaction after solution processing.
136
It might be noted here that numerous researchers are exploring the possibility of a producing a molecular
composite with stiff and semi-flexible chain molecules as the reinforcing phase. In this sense, the term "molecular
composite" really refers to a fine dispersion of molecules which may, or may not, be in an extended chain
conformation. Although property enhancement of the matrix may occur, it will not be to the same extent as the
rigid-rod molecules, which have intrinsically higher properties and an inherent extended chain conformation,
Takayanagi [21 ] studied several stiff chain / flexible-coil polymer blends and observed that the finest dispersion was
15-30 nm diameter microfibrils of poly(p-phenyene terephthalamide) (PPTA) in a matrix of nylon 6 or nylon 66.
He also studied a block copolymer of an aramid and nylon 6 (Table 2) and suggested that a finer dispersion was
achieved but dimensions were not quantified.
PROCESSING OF MOLECULAR COMPOSITES
After synthesis, processing of the molecular composite candidate materials is the most important factor in
achieving a fine dispersion of rod molecules or segments in the matrix material. Phase separation has many
disadvantages for material performance, including the reduced aspect ratio of the reinforcing phase, reduced
interaction and entanglement of the matrix polymer with the reinforcing phase, an increasing amount of
unreinforced matrix, and development of discrete interfaces. To date, the primary technique for processing molecular
composites has been dry-jet / wet-spinning of particles, fibers, or films of a dilute homogeneous acid solution into
a coagulating water bath. Ideally, subsequent thermal consolidation would then be used for fabricating
thermoplastic molecular composites to the desired component geometry. To achieve the desired mechanical and
thermal properties of a molecular composite, large scale phase separation must not occur at any stage of processing,
including solution blending, extrusion, coagulation, and solid state consolidation.
MORPHOLOGICAL CHARACTERIZATION OF MOLECULAR COMPOSITES
A thorough description of morphology is necessary for assessing, understanding, and predicting the ulumatc
properties of rigid-rod molecular composites. A variety of characterization techniques, including optical and
electron microscopy and x-ray scattering, must be used to determine the composition, orientation, and dispersion of
the reinforcing rod molecules in a molecular composite. With this information, and a knowledge of the intrinsic
properties of the homopolymer components, it is possible to model mechanical properties, both to evaluate the
efficiency of reinforcement of the rod molecules and also to predict the ultimate properties possible for a rigid rod
molecular composite. In this section capabilities of major morphological characterization tools will be briefly
reviewed and the potential for application of new characterization techniques will be discussed.
Light-optical characterization techniques, which use the visible wavelengths of light, provide information
down to a scale of about 0.5 pn. The techniques include optical microscopy, light scattering, and Raman
spectroscopy which have proven to be valuable tools for assessing morphology, phase behavior, and
structure-property correlations. Transmission, reflection, and binocular OM has been used to assess the
morphology of fibers, films, and fracture surfaces of consolidated material. Large scale
phase separation (>0.5pm)
is easily detected, and enhanced by color differences, in transmission OM. Ductility, or conversely, brittleness, of
fracture surfaces and adhesion in consolidated molecular composites can be evaluated with reflection and binocular
OM. Hot stage OM to detect phase separation as a function of temperature and composition for PPBT / ABPBI
blend and copolymer systems to determine their ternary phase diagrams (3). More recently a hot stage light
scattering apparatus was used to determine the kinetics of solid state phase separation of a PPBT/nylon system (19).
Electron-optical characterization techniques of transmission electron microscopy (TEM) and SEM. which
use electrons accelerated from I to 200keV or more, have resolution which vary from about 0.2 nm to 2 Pm.
depending upon the operational mode. In SEM each mode is capable of analyzing a sample point-by-point as the
beam scans over the surface. Fracture surfaces of films and fibers can be examined for phase separation, ductility.
and orientation in molecular composite candidate systems with secondary electron imaging (SEI) (typical resolution
= 5 nm) backscattered electron imaging (BEI) (resolution = 20 nm), which reveals atomic nuntber differences. The
SEM is the simplest and most versatile tool for studying morphology down to 10 nm (9.14). Recent advances in
SEM equipment technology have introduced a field emission gun (FEG) SEM which is capable of operating with
excellent resolution at both higher and lower voltage, (0.7 nm at 30 keV and 4 nm at I keV) (23). Overall
resolution limits in the PEG SEM are significantly improved over those of the traditional tungsten hairpin gun
SEM. They ar about I nm in SEI, 5 nm in BEl, and 50 nm in EDS (18).
The FEG SEM has excellent
potential for analysis of moleculaIr composite candidates because of the capability for direct observation of uncoated
surfaces atlow voltages and because of the improved resolution limits in all operating modes (24).
137
TEM is capable of achieving the highest resolution (0.2 nm) of any morphological characterization
technique, but electron beam damage can cause significant alteration of the original structure of the sample, so skill
and patience are rquired to obtain useful results. Additionally, TEM sample preparation techniques are difficult and
may themselves induce artifacts into the original sample structure. Thus, caution must exercised in interpreting
TEM results. TEM dark field (DF) imaging, using diffraction contrast, is usually used to image the size, shape,
and orientation of phase separation of a crystalline component of a molecular composite candidate (9,14). Selected
area electron diffraction (SAED) may be used to determine the orientation and crystal structure of crystalline regions
in a sample. Phase separation of size, shape, and orientation of amorphous regions of a molecular composite
candidate can be studied with TEM bright field (BF) imaging, using mass-thickness contrast (18). An important
recent development in analytical electron microscopy is simultaneous parallel acquisition of electron energy loss
spectra (EELS) which has the potential for acquiring entire spectra before significant beam damage (25).
WAXS provides a rapid and very useful means for assessing important morphological features, such as
lattice parameters, crystallite size, and orientation, in molecular composites. One of the most valuable capabilities
of WAXS is its ability to assess orientation, which has a significant effect on material properties. In rigid-rou
molecular composite systems crystallite orientation of the rod-rich and the matrix-rich can be qualitatively assessed
by the degree of arcing of equatorial reflections which are usually perpendicular to the machine processing direction.
A more improved method of evaluating orientation is with the orientation function, which assigns a weighted
numerical value to orientation (9). A recent development in this field is the use to high intensity synchrotron
radiation, which will make possible real-time studies of dynamic phenomena such as phase separation kinetics.
Other morphological characterization techniques have been applied to molecular composites to a lesser
extent, mainly to determine the level of phase separation. Ultrasound has been used to examine phase separation
of a blend on a scale of 0.1 to 5 gim and should prove useful in examining the integrity of consolidated molecular
composite systems. There are other promising techniques which need to be explored and developed, such as
scanning tunneling microscopy, for improved characterization of morphological features, such as phase separated
material, in rigid-rod molecular composite candidate systems.
ENGINEERING APPLICATIONS OF MOLECULAR COMPOSITES
The driving force for development of molecular composites is their excellent potential for use in structural
applications. Although success has not been achieved in synthesizing and processing a "consolidatabl'"
thermoplastic molecular composite, it would be interesting and useful to see what would be the potential for such a
system. Table 3 lists, for a 30 PPBT / 70 ABPBI system, for fiber, film and bulk geometries for measured and
theoretically predicted properties. "Shear-lag" theory was used to calculate theoretical properties. It will be assumed
briefly, for the sake of argument, that this system can be thermally consolidated and compare the properties, tor the
appropriate geometry, for applications of: engineering plastics; high performance fibers; composite matrix resins,
and direct substitutes for fiber reinforced composites.
Many engineering resins have similar mechanical properties including PEEK, polyamideimide Clorlon),
polyetherimide (Torlon), polyimide (Avimid), polysulfone (Udel), and polyester (Xydar). The moduli are about
3.5 GPa (0.5 Mi) and tensile strengths about 0.11 GPa (I5ksi). The maximum operating temperatures of the
0
engineering resins vary from about 150 to 300 C. The bulk 30 PPBT/ 70 ABPBI exceeds mechanical properties
of other engineering resins by a factor of 5 to 10. Since it is expected that the Tg of a thermoplastic molecular
0
composite would probably range from 300 to 400 C the thermal stability would probably exceed that of
engineering resins by 100°C or more. There would be excellent potential for molecular composites for complex
shapes in parts for underhood applications in the auto industry, where thermal properties are critical, and in the
aerospace industry, where high specific properties art critical.
For high performance fibers Table I lists the properties of some of the commonly used materials. The
fiber 30 PPBT / 70 ABPBI does not compete particularly well with the highest performance fibers, T300. Kevlar,
and PBO. However, compared to E-glass the, 30 PPBT / 70 ABPBI has double the specific modulus and about the
same specific strength. Thus, it appears that applications of molecular composites as high performance fibers are
limited. unless properties can be increased with increased loading of the rigid-rod molecules.
Matrix resin applications for bulk 30 PPBT / 70 ABPBI can beconsidered for both continuous and chopped
fiber macroscopic composites, with predicted properties in Table 3. It can be seen that substituting the 30 PPBT 1
70 ABPBI for epoxy in a unidirectional composite would result in only small increases in properties since overall
properties are chiefly controlled by the fiber component. However, in substituting the 30 PPBT / 70 ABPBI for
epoxy in a planar isouopic composite would give a modest increase in modulus, but could result in up to a 3 fold
138
increase in tensile strength because lateral strength is strongly dependent on the matrix resin. Additionally, a major
limiting factor of matrix resins is thermal stability and it may be possible to produce molecular composites with
thermal stability of 100oC or so more than traditional resins.
TABLE 3
Mechanical Properties of Rigid-Rod Molecular Composites and Macroscopic Composites
Material
Modulus
Measured
GPa, (Msi) GPa, (Msi)
Prdicte
Molecular Compgsite
30 PPBT / 70 ABPBI blend
* Spun fiber (C<Ccr)
*Spun film (C<Ccr)
* Ideal 3-D bulk material
Continuous Fiber Macro-Comnosite
60 graphite / 40 epoxy
" Unidirectional
* "Planar" Isotropic (0/4510/90)
60 graph. /40 (30PPBTf/7OABPBI)
" Unidirectional
* "Planar" Isotropic (0/45/0/90)
184 (26)
87 (12)
69 (10)
120 (17)
88 (13)
-
Tensile Strength
Predicted
Measured
GPa, (ksi)
GPa, (ksi)
%el.
Density
1.3 (190)
0.92 (130)
1.4
2.4
-
1.58
1.58
1.58
1.0
1.0
1.90
1.90
1.8 (260)
0.87 (120)
0.69(100)
-
1.3 (185)
0.17 (24)
132 (19)
70 (10)
160 (23)
98 (14)
-
1.6 (225)
0.47 (64)
-
-
1.78
1.78
(35)
-
1.90
1.72
(30)
(36)
(39)
-
1.50
1.33
1.46
1.72
Chopped Fiber Macro-Composite
40 fiber / 60 matrix
" Glass/Epoxy
* Glass/30PPBT/7OABPBI
" Graphite/Epoxy
* Graphite/Nylon
* Graphite / PEEK
" Graphite / 3OPPBT/7OABPBI
30 (4.2)
92 (12)
-
-
51 (7.2)
24 (3.4)
23 (3.2)
99(14)
(.25
-
0.41
-
(60)
-
0.36 (52)
-
0.21
0.25
0.28
-
Substituting the 30 PPBT /70 ABPBI bulk for epoxy in a (40 fiber / 60 matrix) chopped fiber composite
would yield significant increases in modulu, by a factor of 2 to 5 in comparison with other thermoplastic and
thermoset resins. Modest increases in strength of roughly 50% or so might be possible. Once again, a major
advantage could be gained with the thermal stability of molecular composites.
The potential for direct substitution of 30 PPBT / 70 ABPBI material for continuous and chopped fiber
composites can be considered. Substituting a 30 PPBT / 70 ABPBI fiber for a 60 graphite / 40 epoxy
unidirectional composite would give similar specific properties, even though the level of loading of the molecular
composite is 1/2 of that of the macroscopic composite. The 30 PPBT / 7(1ABPIBI film has similar specilic
modulus compared to that of the 60 graphite / 40 epoxy planar isotropic composite, but its specific strength is
more than doubled. In comparing ideal bulk 30 PPBT / 70 ABPBI to chopped fiber composites it is seen that
specific strength is 1.5 to 3 times greater and specific modulus is up to 4 times greater, even though loading is
30% in the molecular composite compared to 40% in the chopped fiber. The potential for improvements in
thermal stability is good, as previously discussed. Although the potential for direct substitution of molecular
composites for macroscopic composites seems excellent, a comparison of other important properties, espciall>
compressive strength, compressive modulus, and toughness cannot be made until a thermoplastic or thermoset
molecular composite system is devised.
SUMMARY AND CONCLUSIONS
The molecular composite concept has been demonstrated with morphological and mechanical property
studies for aromatic heterocyclic systems, but new materials systems and processing techniques will be required to
produce thermoplastic or thermoset molecular composites. Improved characterization and modeling of these
systems will be required. In this regard, recent work in modeling of mechanical properties of molecular composites
139
E1
*
has been reviewed here. The Halpin-Tsai equations from "shear-lag" theory of short fiber composites predic,
properties reasonably well when using the theoretical modulus of rigid-rod molecules in all ar(mit '
o..tf lic
systems. However, modeling of newer matrix systems with the "shear-lag" theory will t.quire additional
consideration of matrix stiffness, desired rod aspect ratio, and rod orientation distribution. Traditional
morphological characterization techniques were discussed and new techniques were considered including: Raman
light scattering for in-situ morphology property-morphology correlation, high resolution and low voltage SEM,
parallel EELS for chemical analysis in TEM, synchrotron radiation in X-ray scattering for dynamic studies, and
ultrasound for integrity studies in consolidated material. Excellent potential exists for use ol molecular composites
in structural applications including engineering plastics, composite matrix resins, and as substitutes for fiber
reinforced composites. Some limitations may exist when compressive and toughness properties are evaluated, but
these factors cannot be considered until a thermoplastic or thernoset molecular composite is devised.
ACKNOWLEDGEMENTS
The author wishes to acknowledge partial support of this research by the Dow Chemical Co. Valuable discussions
are acknowledged from W.W. Adams, T.E. Helminiak, S. Kumar, H. Chuah and S. Donaldson of the AFML.
REFERENCES
j
I.
2.
3.
4.
5.
6.
7.
8.
9.
().
I .
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
4
4t
T.E. Helminiak, F.E. Arnold, and C.L. Benner, Am. Chem. Soc. Poly. Preprints, 6, 659 (1975).
T. E. Helminiak, C.L. Benner, F. Arnold, G. Husman, U.S. Pat. Appl. 902,525 (1978).
W-F. Hwang, D.Wiff, C.Benner, T.Helminiak, J. Macromol. Sci. Phys., B22 , 231 (1983).
J. Wolfe, "Polybenzthiazole and Polybenzoxazole Review" in Encyclopedia ol Polymer Science and
Engineering, 2nd Edition, J. Wiley & Sons. New York, 1988.
G. Husman, T.E. Helminiak, W.W. Adams, D. Wiff, and C.L. Benner, Am. Chem. Soc. S>mp. Ser., 132,
203 (1980).
R.M. Christensen, Mechanics of Comnosite Materials, Wiley, New York. 1979.
S. Donaldson, private communication.
PJ. Flory, Proc. Roy. Soc. London, A 234, 73 (1956).
S.I. Krause, T. Haddock, G.E. Price, P.G. Lenhert, iF. O'Brien,
ltchnlnijk, and W.N. Adam,
tl
J. Polymer Sci. - Polym. Physics Edition, 24. 1991 (1986).
R.J. Day, I.M. Robinson, M. Zakikhani. and R.J. Young, Polymer, 2 , 1833 \1988.
RJ. Young, private communication.
W-F. Hwang, D.Wiff. C.Verschoore, G. Price, T.Helminiak, and W.W. Adams, Poly. Eng. and Sci.. 23.
784 (1983).
T.T. Tsai, F.E. Arnold. and W.F. Hwang, Am. Chem. Soc. Poly. Preprints, 26, 144 (1985).
S.J. Krause, T.B. Haddock, P.G. Lnhert, W-F. Hwang, G. Price, T.E. Helminiak, J.F. O'Brien, and
W.W. Adams, Polymer, 22. 1353 (1988).
W.F. Hwang, D.R. Wiff, TE. Helminiak, and W.W. Adams, ACS Preprints. Org. Coat.
and Plast. Chem., 48. 922 (1983).
S.M. Wickliffe, M.F. Malone, and RJ. Fars, J. Appil. Polym. Sci.. 3.4,9l (1987I
0. Nehme. C. Gabriel, RJ. Farris, E.L. Thomas. and M. Malone. J. AppI. lol.tin. Set 1.. 1955 (19,81.
S.J. Krause and W.W. Adams, Elect. Mic. Soc. Am. Proc., 46, 748 (1OXX)
H.C. Chauh, T. Kyu, and T.E. Helminiak. Am. Chem. Soc. Poly Eng. Sci. Proc., 5 I.
I (R) L
1988
T. Nishihara, H. Mera, and K. Matsuda, Am. Chem. Soc. Poly. Eng. Set. Prock . , 821 1986).
H.C. Chauh, L.S. Tan, and F.E. Arnold, Poly. Eng. and Sci., 22, 107 (1989).
M. Takayanagi, T. Ogata, M Morikawa, T. Kai, 1. Macro. Sci. Phys.. 131, 519 (1981).
S.J. Krause, W.W. Adams, S. Kumar, T. Reilly, and T. Suzuki Elect. Mic. Soc. Am. Proc., 44. 66
(1987).
SJ. Krause. W.W. Adams, and D.C. Joy, Elect. Mic. Soc. Am. Proc., 47. 336 (1989)
O.L. Krivanek, Elect. Mic. Soc. Am. Proc., 46, 660 (1988).
141
RHEOLOGY OF BLENDS OF A RODLIKE POLYMER (PBO) AND ITS FLEXIBLE
CHAIN ANALOG
V. J. SULLIVAN AND G. C. BERRY
Carnegie Mellon University, Dept. of Chemistry, 4400 Fifth Avenue, Pittsburgh,
PA 15213
ABSTRACT
Rheological
properties of isotropic solutions of rodlike poly(p-phenylene
benzbisoxazole), PBO, flexible chain poly(2,5-benzoxazole), ABPBO, and their
miscible blends in solution are described.
Measurements include steady state
properties (the viscosity and recoverable compliance as functions of shear
transient
properties
(the
recoverable
rate),
compliance),
and
dynamic
mechanical properties (the loss and storage compliances as functions of
frequency).
The relaxation spectrum of the blends is broader than that for
the rodlike chain, and tends to occur at longer times, reflecting a viscosity
enhancement that occurs with the blends. The measured zero shear viscosities
for rod and blend solutions are compared with predictions based on the model
of Doi and Edwards.
INTRODUCTION
Rodlike polymers containing heterocyclic aromatic groups in the main
chain have shown promise as reinforcing agents in
rod-flexible chain
composites (1).
Above a critical composition for formation of a nematic phase,
rod-coil-solvent mixtures typically exhibit a brced biphasic regime, in which
the flexible chain is predicted to be essentially excluded from the nematic
phase [2].
In order to maintain an intimate dispersion of the rodlike chain in
the flexible chain matrix, fibers and films are processed from the iaotropic
phase.
It is therefore important to understand the dynamics of isotropic
mixtures of rods and flexible chains.
The model of Doi and Edwards for isotropic solutions of caged rods treats
the rod constraint release mechanism, for movement of a rod into a new cage,
as the translation of a rod along its length by a distance proportional to its
contour length 131. This treatment gives a result for the rotational diffusion of
a caged rod which leads to an expression for the zero shear viscosity 7o
given by
7
o/ ls
f
=
104i[]
3
(cL/ML)
(1 - B(c/c*)]
f
(I)
-2
where 'rs is the solvent viscosity, c, M, L and ([ql are the rod concentration,
molecular weight, length and intrinsic viscosity, respectively, ML = M/L and K
and B are parameters determined by fitting the equation to experimental data.
Here, f is a crowding factor accounting for the severe restriction of the rod
rotation which occurs at concentrations approaching (but less than) the
critical rod concentration cR
for the formation of an ordered phase 141.
In
5
practice, B is found to be close to unity (4].
In the following, eq. I will be
used to predict the viscosities of both rod and blend solutions.
Mat. Net. Soc. Symp. Proc. Vol. 171. 11990 Materials Research Society
142
EXPERIMENTAL
Dry polymers were received from the Dow Chemical company.
Static and
dynamic light scattering and intrinsic viscosity measurements give a weight
average chain length L w = 170 nm for PBO, and Lw = 320 nm for ABPBO.
Rheological experiments included steady state viscosity and recovery, transient
creep and recovery and dynamic mechanical measurements.
Steady state
viscosity and creep and recovery measurements were carried out using either
a wire suspension rheometer (5] or a gas bearing rheometer f6).
Dynamic
mechanical measurements were performed using the latter.
The linear transient shear creep compliance J(t) =7(t)/a is given by [7]
3(t) = R(t) + t/jo
(2)
where 7(t) is the strain at time t after imposition of a constant stress a, 'Io is
the linear steady state shear viscosity and R(t), the recoverable compliance.
The latter may be determined as 7r(t)/w, where 7r(t) is the recovered strain
at time t after cessation of an imposed streas a for a time long compared to
the longest relaxation time of the sample. At long times R(t) equals the linear
steady state recoverable compliance Re;
R(t) is often represented as a
discrete spectrum of retardation times Ai and weight factors Ri (7]:
Ro - R(t) =
N
E
i=1
R i exp-t/Xi )
(3)
The dynamic storage and loss compliances, J'(w) and J"(), respectively,
were determined by measurement of the in-phase and out-of-phase components
of the sinusoidal strain 7(t) resulting from imposition of a sinusoidal stress.
The complex viscosity
d()
[
1w
d(0)]l
and the dynamic viscosity ,I'()
2
J"(-)/-Jd(-) ], where Jd(w)
(J'(w)z + j-(, )211/2, will be discussed below.
RESULTS
The viscosities of isotropic PBO/MSA solutions are described by eq. I
-6
with the K = 8.6 x 10
and B = 0.96.
These are in accord with values
obtained with other rod solutions (4]. The viscosities of solutions of PBO and
1
ABPBO are plotted against cLw in Fig. 1. The enhancement of Vo on addition of
flexible chain to rod solutions is also shown in Fig. I, where the viscosities
for blend systems are plotted against total polymer concentration times the
rod length.
In Fig. I and in the following, the rod concentration in a rod or
blend system is denoted ci.
Data on J'(w) and J"(,) for two blend compositions are shown in Fig. 2,
along with data on J(t) and R(t) for these systems.
As has been reported
elsewhere (4,8,9], rheological functions for solutions of rod, flexible chain and
blend systems can be reduced using the parameter Tc = "ORo.
For example,
R(t) is
plotted
as R(t)/Ro versus t/rc in
Fig.
2 for different rod
concentrations.
For both rod and blend systems, the steady state viscosities
'vi, at shear rate x, is about equal to 'I'(w) for x = w, as expressed by the
Cox-Merz relation (71; '"d(w) tended to be larger than i for the larger w. In
addition, over the range in the time scale for which both dynamic and
transient date are available, the Markovits-Riande relation (71
143
.0
4
LOG
*m
3
2
U
-6.8
.8.6
-6.4
4.2
-6.0 -S.8
-5.4
-5.6
2
LOG cLw (g/cm )
solutions
FIGURE 1: Zero shear viscosities of rod ( a and flexible chain (
For comparison with rod viscosities,
).
equation
as a function of cL (ee
concentration
blend viscosities are plotted versus the product of total polymer
* denotes blends in which various amounts of
The symbol
and rod length.
denotes
0.028 g/ml; a
flexible chain were added to rod solutions at cR
blends with cR = 0.043 g/ml.
to.
0
,0'
"
z
m
Ida
1
Id
t/rc
or
10
l/(i'c,)
FIGURE 2:
Dynamic and transient rheological
compositions: ca = 0.043 a/mi, cF - 0.031 g/w| (with
cg = 0.015 gl
(without pipe). Symbois denote the
J'fw)/Ro o , J"M l)/o
a
, R(t)/R o
• , J(t)I11
Jd(w)/Ro for the blends; the dotted linoe is J'(w)/R
o
co = 0.043 C/=L
1
functions for two blend
pips)and cq
0.043 g/"d
fonwinK functic"J
The sold linogve
for a rod zoiution at
144
J(t)
at 'r
jd(-)
J
Z t
(4)
is a reasonable approximation for theme systems.
The addition of the flexible
chain to rod solutions broadens the relaxation spectrum, as can be seen by
the change in the shapes of rheological functions in the short time regime
(see Fig. 2).
DISCUSSION
The nonlinear viscoelastic behavior exhibited by the materials examined in
the present study is described by a factorized form of the BKZ constitutive
equation (10]:
(t)
tK*[(t
-
); 7(t)
-
Y(s)] [7(t)
7(s)] ds
-
(5)
where 7(t,s) = 7(t) - 7(s).
A form for K* which has proved successful for
both isotropic rodlike and flexible chain polymer solutions is [4,8]
aG(t - a)
K*((t -s);y(t,s))
=
-f(7(t,s))
a(t
s)
-
where
f(7(ts))
m
m
exP-1
1(fi
for 17(t,s)l
for 17(ts)I
0
I
7
<
b
]
(6)
'
-Y
-Y'
Here, G(t) is the linear shear relaxation modulus.
The latter may usually be
represented as a discrete relaxation spectra.
An experimental procedure to
determine 7' and 7" is described in reference 8.
This model, previously
successfully applied to the blend system PBT/nylon/MSA [9], provides good
fits to the nonlinear behavior of the PBO/ABPBO/MSA blend and the PBO/MSA
systems.
The narrow range of relaxation times available for the ABPBO/MSA
system prohibited adequate representation of its nonlinear behavior.
A phase diagram for the room temperature phase behavior of the
PBO/ABPBO/MSA system is shown in Fig. 3.
As can be seen in Fig. 3, the
addition of flexible chain causes a drop in the critical rod concentration
required for the onset of the ordered phase. For a given rod system, eq. 1
indicates that the critical rod concentration cR*
affects the zero shear
viscosity through the crowding factor f.
For blends in which the relaxation
mechanism of longest time scale is rod rotation, i.e. Tr
< ra J where TF is
the time scale for flexible chain disengagment and T-tR is the time scale for
rod rotational diffusion, we may c-,,sider the rod rotational diffusion to be the
sole determinant of no,s, 1 the blend zero shear viscosity. In this case, o,s is
given by eq. 1, with c 5
replaced by 8R, the critical rod concentration for
formation of an ordered phase in the blend. The ratio of 70,@ and vo for a
given rod concentration is then given by
[ 1
1'os
-i--=1
-
B(cR/cR*) 12
SJ(c 5 /e5 )
j(7)
US
0.02
0.10
0.14
PEO
ASIU
FIGURE 3: Phase diagram, at T
25 OC, for PO/ABPO/MSA: Open circles
denote the isotropic phase; filled circles denote the ordered ph.s.; the
coexistence line shown represents the prediction of the Flory model (21.
403020100,
0.01
0.02
0.03
0.04
0.65
cfr (glml)
FIGURE 4:
Ratios of blend to rod viscosities an a function of flexible chain
concentration CF.
Point. denote experimentally determined values, with
symbols denoting blend compositions an in FIGURE 1. Lines denote viscosity
ratios calculated by equation 7, aa described in text. Upper line; blond@ with
c*0.043 g/al. Lower line; blend@ with cl,
0.028 g/ml.
146
A
For the blends 6R will be estimated from the phase diagram shown in Fig. 3.
A line from the solvent (USA) apex through a given point represents all
compositions with rod to flexible chain ratios equal to that of the blend
represented by that point; el is estimated as the rod concentration for the
point at which this line intersects the isotropic-nematic coexistence line. With
PE values obtained in this way from Fig. 3, and assuming B' = B, eq. 7 can be
Calculated ,qo,g/,qo values, along with experimentally
used to calculate -o, /o.
determined values, are presented in Fig. 4.
For the two blends at the higher rod concentration (cR = 0.043 g/ml), the
calculated viscosity ratios are close to the measured values. For blends with
cp = 0.028 g/ml, the calculated viscosity ratios are substantially lower than the
Thus the above calculation underestimates the extent of
measured values.
viscosity enhancement that occurs upon adding flexible chain to rods at this
This discrepancy may result from neglect of the possibility of
concentration.
a contribution to the blend viscosity involving the disengagement time of the
This possibility must be considered for the present system,
flexible chain.
Indeed the
since the pure component viscosities are comparable (see Fig. 1).
broadening of the relaxation spectrum upon addition of flexible chain (Fig. 2)
suggests significant contributions from relaxation mechanisms involving the
T
R., sited above as a
TF
<
In this case, the relation
flexible chain.
neccessery condition for the use of eq. 7, would no longer be valid, and
Models which predict the viscosity of rodshould be replaced by TF o T"
coil blends in the latter regime are not yet available.
Acknowledgment.
This study was supported in part by DARPA and AFOSR.
REFERENCES
1)
S. J. Krause, T. Haddock, G. E. Price, P. G. Lenhert, J. F. O'Brien, T. E.
Helminiak and W. W. Adams, Polym. Sci.: Part B: Polym. Phys., 24, 1991
2)
3)
P. J. Flory, Macromol., 11, 1138 (1979)
M. Doi and S. F. Edwards, The Theory of Polymer Dynamics (Clarendon
Press, Oxford, 1986)
S. Venkatramen, G. C. Berry and Y. Einaga, J. Polym. Sci.: Polym. Phyr.
Ed., 23 , 1275 (1985)
G. C. Berry and C.-P. Wong, J. Polym. Sci.: Polym. Phys. Ed., 13, 1761
(1975)
(1986)
4)
5)
6)
7)
G. C. Berry, J. 0. Park# D. W. Meits, M. H. Birnboim and D. J. Plazek, J.
Polym. Sci.: Part B: Polym. Phys., 27, 273 (1989)
G. C. Berry and D. J. Plazek, in Glass: Science and Technology, Vol. 3, Ed.
by D. R. Uhlmann and N. J. Kreidl (Academic Press,
Chapter 6
8) K. Nakamura, C.-P. Wong and G. C. Berry, J. Polym.
22,1119 (1984)
9) C. S. Kim, Ph.D. Thesis, Carnegie Mellon University,
10) B. Bernstein, H. A. Kearsley and L. J. Zaps, Trans.
391 (1963)
New York, 1986),
Sci. Polym. Phys.,
Pittsburgh, PA (1988)
Soc. Rheol., 1,
147
PBZT MICROCOMPOSITES WITH ADVANCED THERMOPLASTIC MATRICES
W. MICHAEL SANFORD AND GERARD M. PRILUTSKI
E. I. du Pont de Nemours and Co., Inc., Experimental Station, P.O. Box 80302,
Wilmington, Delaware
19880-0302
Thermoplastic microcomposites offer the potential for better economics
and improvements in composite processing, and possibly performance, over
conventional "string-and-glue" composites. The early development of molecular composite technology focused on polyamide matrix polymers; however, for
many aerospace applications higher use temperatures and greater solvent resistance than that of conventional polyamide matrices will be required.
This paper describes work performed under contract to the U.S. Air Force to
develop PBZT (poly p-phenylene benzobisthiazole)/thermoplastic molecular
composites with high performance matrix resins into a viable technology.
A scaleable process has been defined based on a novel technology developed by Du Pont.
Advantages of this process include better economics,
superior processing performance, and improved MC fiber tensile properties
versus prior art.
Uring this process we have obtained rule-of-mixtures
properties in our microcomposite
fibers with matrix polymers offering use
temperatures from 330 to 600*F.
Consolidation of PBZT/PEKK fibrous preforms
into uniaxial panels up to 10" x 15" has been demonstrated and material propperty evaluation and data base development are in progress. Uniaxial property levels achieved to date for all systems compare favorably with conventional
"string-and-glue" PBZT/epoxy composites although as with other organic fiber
reinforcements, compressive and shear performance may be limiting factors in
MC applications.
INTRODUCTION
Conceptually, molecular composites are dispersions on a molecular scale
of rigid rod polymer molecules in a matrix of flexible coil polymers, formed
by the coagulation of a dilute isotropic solution containing these components
[1-31.
In the original concept, phase segregation of the rod and coil polymers into separate phases was to be avoided, because the rigid rod molecules
tended to aggregate to form domains with low aspect ratios ("footballs")
which led to ineffective reinforcement and hence low mechanical properties
[2].
To avoid phase segregation, very low concentrations (below the critical
concentration of 3.5%) were generally used.
We have found, however, that
under certain conditions, and using higher, more practical, solution concentrations, the rigid rod polymer aggregates to form a fibrillar microscopic
rather than molecular dispersion (rod domains of 1,000-10,000 vs. <30 angstroms), with high aspect ratio rigid rod domains.
These "microcomposites",
(MC's), exhibit higher tensile properties in fiber preforms than molecular
composites [4].
These thermoplastic microcomposites offer the potential for
better economics (no resin impregnation step) and performance over conventional "string-and-glue" composites.
MATRIX SELECTION CRITERIA
There are several important requirements for the matrix resins to be
used in PBZT-based molecular or microcomposites. Particularly, the matrix
must be soluble and stable in the strong acid solvents (e.g., methanesulfonic,
polyphosphoric) required for processing rigid rod polymers.
Additionally,
the polymer must be sufficiently thermally and thermo-oxidatively stable to
permit heat treatment at elevated temperatures. Lastly, the matrix must be
capable of forming a stiff, strong molecular composite preform which can be
Mel. Res. Soc. Symp. Proc. Vol. 171.
1990 Materials Research Soclety
148
consolidated by heat and pressure. Additionally, for the "new thermoplastic"
matrix resin, we sought to identify polymers which offered advantages over
the polyamides used in earlier studies of molecular composites, primarily
higher Tg and improved solvent resistance. Goal levels for the use temperature of the matrix were set at 250, 350, 450 and 600'F (121, 177, 232 and
315C). On the basis of processibility and potential use temperature, three
polymers were selected for in-depth study in this program, PEKK, and the
polyimides used in Du Pont's Avimid* K and N reinforced resin composite
structures (referred to as "K-polymer" and "N-polymer" respectively). The
work described in this paper focuses on these three matrix systems.
MICROCOMPOSITE PREPARATION
4the
Microcomposite preforms (e.g., fiber, film) are prepared via a process
similar to that described for forming molecular composite preforms [2,31.
First, a solution consisting of both rigid rod and flexible coil polymers in
an appropriate solvent is prepared and intimately mixed to effect dispersion.
After deaeration, the solution is then extruded to produce the desired preform, which in this work has primarily been in the form of continuous fibers.
The preforms are produced via air gap spinning, with orientation imparted to
the rigid rod component (PBZT) through shear in the capillary and spin draw
in the air gap. After spinning and solvent extraction, the fibrous microcomposite preforms are heat treated under tension to increase orientation,
thereby enhancing tensile strength and modulus.
The microcomposite fibrous preforms are converted into three dimensional
structures via a two-step process. First, the preform is wrapped around a
flat plate and pressed to form a coherent sheet with the fiber axes aligned
unidirectionally, analogous to a conventional composite prepreg. Secondly,
sheet is cut into plies of the desired size which can be stacked to form
a laminate which is then consolidated via application of heat and pressure.
PBZT/PEKK MC's
PEKK (poly(etherketoneketone)) is a semi-crystalline thermoplastic resin
offering intermediate use-temperature performance comparable to 3501-6 epoxy
or PEEK (poly(etheretherketone)) resins. It has a glass transition temperature of 156*C and offers excellent resistance to organic solvents [5,6].
Based on these properties as well as its excellent solubility and stability
in the strong acid solvents required to process PBZT, PEKK was selected as
a matrix for microcomposites evaluation.
Acid solutions of PBZT/PEKK (60/40 v/v rod/coil ratio) were air-gap spun
on a capillary rheometer over a range of conditions with excellent continuity
(no filament breaks) to form 10 filament fibrous
microcomposite precursor
yarns (nominal denier 45).
High spin stretch was achieved, which resulted
in a high degree of orientation of the rigid rod molecules in the as-spun
fibers, as evidenced by average orientation angles for the PBZT of 12-17
degrees measured by X-ray diffraction. The apparent crystal size was typically in the range 20-25 angstroms. As-spun tensile strengths were in the
range of 100-200 ksi, with elongations to break in the range 1-2.5% and
moduli of 9-14 Msi. Tension heat treatment of the as-spun yarns increased
the degree of orientation, resulting in average orientation angles of 5-7
degrees and apparent crystal sizes of 50-56 angstroms. Typical tensile properties of the heat treated PBZT/PEKK microcomposite fibrous preforms were
200-310 ksi strength, 19-25 Hsi modulus and 0.5-1.2% elongation to break,
(single filament breaks), with heat treated property levels varying with the
is-spun fiber properties and the temperature and tension used in heat treatment.
If these properties are normalized to account for the fraction of PBZT
Sii
Pont Rvgiatered Trademark.
149
(60% by volume), a strength and modulus of 500 ksi/38 Msi are obtained for
the PBZT reinforcement. These property levels are typical of neat PBZT fibers [7], indicating that "rule-of-mixtures" properties have been obtained
in the microcomposite fibrous preforms.
PBZT/PEKK fibrous preforms were molded into 6" x 112" unibars by the
two-step process described above. These bars had theoretical densities and
good C-scans (loss <2 dB at 5 MHz) indicating good consolidation. The mechanical properties of PBZT/PEKK (60/40 v/v) unibars are summarized in Table I.
The specific mechanical properties are compared to several other composite
materials in Figures 1-2. The tensile properties and flex and compressive
moduli are comparable to those obtained for conventional composites reinforced
with PBZT or a mid-range carbon fiber such as T300. In Figure 2, the specific
tensile strength reported for the PBZT/PEKK MC bar is somewhat lower than for
the PBZT/epoxy conventional composite; however if full translation of the MC
yarn strength to the bar can be achieved, the tensile strength of the PBZT/
PEKK should be approximately that of the PBZT/epoxy. The lower density of
the microcomposite versus carbon reinforced composites leads to an advantage
in specific tensile properties. Preliminary tensile tests of 9" microcomposite specimens have yielded somewhat higher moduli, up to 24 Msi. Based on
These high moduli, the microcomposites would be suitable for stiffness-critical
Ipplications.
TABLE I:
Mechanical Properties of PBZT/PEKK (60/40) Unibars
Tensile
Flex
Strength (ksi)
175
80
32
4.8
(Msi)
19
20
20
---
Modulus
Compressive
Short-Beam Shear
The flex and compressive strengths compare poorly with carbon fiber
reinforced materials. The properties are typical of organic fiber reinforced
composites and reflect the poor compressive strength of PBZT. This deficiency
in compressive and flexural strength will limit the utility of rigid rod MC's
fr primary structural applications. Note that no significant advantage is
fomnd for microcomposites versus conventional composites in mechanical properies, including compressive strength, based on our results to date.
600,
PBZTIPEKK
C
*
MC
500,
4
400.
AS4/PEKK
T300/5208
U•
2
Tensile Modulus
SFlex
Modulus
10
0
11
1
2
3
4
FIGURE 1: Specfic moduli of PBZT/PEKK MC compared to conventional composites
150
5.
PBZT/EPOXY PBZ/PEKK MC
fiber
AS4/PEKK
T30015208
•
4-
bar
3
3
•
Tensi
Strength
Flex
e Strength
Compressive Strength
2-
.u
Z
0
1t
0
1
2
3
4
FIGURE 2: Specific strengths of PBZT/PEKK MC compared to conventional composites
In addition to molding 6" x 1/2" unibars, several 6" x 3" panels were
also consolidated. These panels had theoretical densities and good C-scan
quality, similar to that of the smaller unibars. Tensile and flexural test
coupons cut from these panels had mechanical properties typical of the individually molded unibars. Panels up to 10" x 15" x 0.100" in size with theoretical densities and acceptable C-scans have been produced and shipped to
Boeing Advanced Systems Co. for evaluation.
PBZT/PI MC's
K-polymer is the matrix resin used in Du Pont's composites of Avimid® K.
This amorphous polyimide has a glass transition temperature of 250*C, good
thermo-oxidative stability and excellent resistance to organic solvents. Npolymer is the polyimide matrix resin used in Du Pont's composites of Avimi8d
N. This non-crystalline polymer has a glass transition temperature of 340370*C, excellent thermo-oxidative stability and good resistance to organic
solvents. Because of their high Tg's and stability, both K and N polymers
are of interest in high temperature applications.
Acid solutions of K polymer and PBZT were mixed and spun similarly to
the PBZT/PEKK microcomposite solutions to form 10-filament yarns of the fibrous microcomposite precursor. The PBZT/K yarn, however, did not sustain as
much spin stretch as the PBZT/PEKK. The effect of this lower attenuation was
evident in the as-spun orientation angles of 25 to 32 degrees for PBZT/K as
compared to the 12-17 degree orientation angles obtained for as-spun PBZT/
PEKK. Tensile properties of the as-spun PBZT/K yarns were in the range 100163 ksl strength, 2.5-7% elongation to break, 4-8 Msi modulus. Tension heat
treatment of the PBZT/K yarns decreased the orientation angle to 8-12 degrees.
The tensile properties of heat treated PBZT/K microcomposite preform yarns were
increased to 250 ksl/25 Msi strength and modulus, which represent approximately
"rule-of-mixtures" properties and are comparable to carbon fiber reinforced
composite tensile properties.
11eit treated PBZT/K yarns were consolidated Into 6" x 1/2" unibars by
the s.ime process used for PBZT/PEKK. The as-molded bars had good C-scans
( n dii loss at 5 MHz) and near theoretical densities. Tensile strength and
mdtiii ol the bars averaged 70 ksi and 15 Nsi, respectively.
The low strength
i heI
,vcd to he the result of tab failure rather than true tensile breaks.
151
Flex strength and modulus of the PBZT/K bars averaged 58 ksi and 15 Msi,
which are comparable to the values reported for PBZT/epoxy conventional composites [8].
Similarly to the PBZT/PEKK values, the low flexural strength
of the PBZT/K microcomposlte is attributed to the low compressive strength
of the PBZT reinforcement. The average short beam shear strength measured
for the PBZT/K unibars was only 2.5 ksi, approximately half that of the PBZT/
PEKK microcomposite and a PBZT/epoxy conventional composite. This indicates
that even though the PBZT/K microcomposite consolidated well, there is relatively poor adhesion within the laminate.
An acid solution of PBZT/N (60/40 v/v) was spun in a similar manner to
the PBZT/K. High spin stretch and moderate orientation (25 degrees) were
obtained in the as-spun fibrous preform which had tensile properties of 200
ksi strength and 9.4 Msi modulus. Tension heat treatment decreased the average orientation angle to 6.5 degrees with a corresponding increase in tensile
strength and modulus to 280-400 ksi and 23.8-29.5 Msi, respectively, depending on the heat treatment conditions. An initial attempt at consolidating
the PBZT/N yarns was only marginally successful; however, this experiment was
performed at relatively low pressures and further experimentation is expected
to lead to improved adhesion. Based on the temperature capabilities of PBZT/
N polymer microcomposites and the excellent tensile properties obtained, this
material appears quite promising if good consolidation can be obtained.
SUMMARY
PBZT-based microcomposites with advanced thermoplastic matrices have
been investigated. A scaleable spinning process has been developed to produce high yields of PBZT/PEKK microcomposite preform yarns having "rule-ofmixtures" tensile properties. PBZT/PEKK microcomposite unibars have moduli
comparable to T300 or AS-4 carbon fiber reinforced compositeF ,nd may be
suitable for stiffness-critical applications. Microcomposite fibrous pretorms of PBZT/K have been produced with high tensile properties and compression molded to form well-consolidated unibars. PBZT/N microcomposite fibrous
preforms have outstanding tensile properties and may be suitable for applications at temperatures >650*F, although consolidation of these preforms has
yet to be demonstrated. Further effort will be aimed at developing these
polyimide matrix microcomposites. Preliminary data indicates that microcomposites offer no distinct differences in static mechanical properties vs.
conventional composites with the same constituents. Structural applications
for rigid rod microcomposites are likely to be limited by the poor compressive strengths, which are typical of organic polymer reinforced composites.
ACKNOWLEDGMENTS
The authors gratefully acknowledge the U.S. Air Force Wright Research
and Development Center and the Defense Advanced Research Projects Agency for
their support of this research under contract number F33615-86-C-5069.
REFERENCES
1.
Thaddeus E. Helminiak et al., U.S. Patent No. 4,207,407 (1980).
2.
W.-F. Hwang, D.R. Wiff, C.L. Benner, T.E. Helminiak, J. Macromol. Sci.
Phys., B22 (2) 231-257 (1983).
3.
W.-F. Hwang, D.R. Wiff, C. Verschoore, G.E. Price, T.E. Helminiak,
W.W. Adams, Polym. Eng. Sci., 23 (14) 784-788 (1983).
4.
W.C. Uy, G.M. Prilutski, W.M. Sanford, WRDC-TR-89-4040 (1989).
153
PBZT/POLYAMIDE THERMOPLASTIC MICRO-COMPOSITES AN OUTGROWTH OF MOLECULAR COMPOSITES DEVELOPMENT
WILLIAM C. UY AND E. R. PERUSICH
E. I. du Pont de Nemours & Co., Inc., Experimental Station, P. 0. Box 80302,
Wilmington, DE 19880-0302
ABSTRACT
rMolecular
~applications.
composites are dispersions of rigid-rod polymer molecules in a
matrix of flexible coil polymers. formed by the coagulation of a solution containing these components,
Where there is aggregation of the rigid-rod
molecules, such composites are called micro-composites (MC's).
These composites offer the potential for better economics and improvements in composite
processing, and possibly performance, over conventional "string and glue"
This paper describes work performed under contract to the U. S. Air
composites.
Force to develop PBZT/thermoplastic molecular composites into a viable techA commercially viable MC spinning
and heat-treatment process has been
defined based on a novel mixed solvent/quaternary solution technology developed by Du Pont.
Advantages of this process include better economics, superior
processing performance, and improved MC fiber tensile properties versus prior
art.
PBZT/polyamide MC fibers with strength/modulus of 332 ksi/29 Msi have
been produced using this process.
Adhesion equivalent to that obtained in conventional composites has been demonstrated.
Uni-axial properties achieved to
date compare favorably with conventional "string and glue" PBZT/epoxy composites although compressive and shear strengths may be limiting factors in MC
INTRODUCTION
4of
This work was carried out at the Advanced Structural Materials Technology Center of the Fibers Department of E. 1. du Pont de Nemours & Co., Inc.
This presentation is on the development of general process technology
for PBZ-based molecular composites and, in particular, of PBZT-based products
in thermoplastic polyamide matrix system.
Du Pont involvement with PBZ* technology, the Air Force Materials
Laboratory and SRI International dates back to 1981, when research focused on
the neat PBZT polymer in developing spinning and heat-treatment processes
for this true rigid-rod polymer from its unusually viscous as-polymerized
polyphosphoric acid solution [I].
In the following year the process was refined
and scaled-up [2).
In 1985 and 1987, we expanded the experimental production
PBZT fiber for the Air Force and converted a total of 160 lbs. of the polymer to
heat-treated fiber in continuous 290-filament yam.
As-spun PBZT fiber has a
purplish brown color while optimally heat-treated fiber has a shiny, metallic
blue color.
The average tensile strength and modulus values of these
production yams exceed 400 ksi/40 Msi while values as high as 614 ksi/49 Msi
(Table I) have been obtained for lab-scale produced fibers.
PBZ polymers in general have outstanding thermal, oxidative, and hydrolytic stability [2).
Trans-PBZT and cis-PBO. in particular, are the most important and true rigid-rod PBZ polymers [3.41.
Therefore, they are ideal as reinforcing components in molecular composites.
PBZ is the generic terminology used to refer to the class of rigid-rod heterocyclic polymers, which have been under development with support of the Air
Force Materials Laboratory.
*
Ma. Res. Soc. Symp. Proe. Vol. 171. €1i"0 Matediat Rftoarch Society
154
The specific mechanical properties (Fig. 1) of various high performance
fibers and metals are compared in this chart.
PBZT. because of its unique
properties, occupies an area to itself.
MOLECULAR AND MICRO-COMPOSrTES
Dr. T. Helminiak (U. S. Air Force Wright Research & Development Center)
pioneered the concept of molecular composites (MC's) as early as 1978. A patent
[4] was issued to him and his colleages wherein the matrix is not a thermoplastic.
Their patent reads "Rod-like aromatic heterocyclic polymers are used as a
reinforcement in coil-like heterocyclic polymer matrices to provide composites
at the molecular level that are analogous to chopped fiber composites."
In the
strictest sense, in a true molecular composite the reinforcing component would
be dispersed on a molecular scale without aggregation.
However, theoretical
and experimental considerations have shown that true molecular dispersion is
attainable only at a few percent of rod content [5,6].
In practical applications.
where the rod content would more likely be more than a few percent (e.g., for
normal composite application, rod content is usually 60 volume %). aggregation
is unavoidable.
The scale of aggregation can range from a state where it is not
visible even at very high TEM magnification to a state where there are clearly
two phases.
If the dispersion of the rigid-rod molecules is not on a molecular
level but rather on a sub-micron scale, we call such composites micro-cornposites (MC's).
Therefore, realistically, this discussion will be on micro-composites
rather than molecular composites.
Why MC's?
Potential advantages of MC's over conventional "string-andglue" composites are better economics and potential superior performance.
Better economics, because with MC's, the pre-pregging step is not necessary
since the matrix is already built-in.
Superior performance, because there is no
macroscopic interface between the reinforcing and matrix components in MC's.
The interface is on a suo-micron level and, therefore, the very high interfacial
area can reduce stress concentration.
The chart compares the relative sizes of
the reinforcing component for conventional and micro-composites with two
degrees of dispersion.
The general objective of our current research is to develop thermoplastic
molecular composites based on PBZT into an industrially viable technology by
developing process techniques to produce and fabricate bulk material forms.
The program elements are essentially the same for the thermoplastic polyamide
and "new" thermoplastic matrix system with one exception, i.e., research on
basic processing technology and design/fabrication of equipment applicable
for both polyamide and new systems.
This presentation addresses primarily the
polyamide system.
Thermoplastic
Polyamide
System
Six thermoplastic polyamide resins (Table I1) were chosen and evaluated
for the effects of crystallinity versus amorphous character and glass transition
temperature.
Four criteria (Table I11) were used in screening the matrix candidates:
Stability in strong acid solvents, physical and chemical compatibility, MC
fiber properties and uni-directional bar properties.
Good compatibility between PBZT and the matrix resin (Table IV) was
found to be important in achieving good properties.
Compatibility was determined from calculated solubility parameters 17] and interfacial bond adhesion
[8].
As-spun fibers were heat-treated to develop full property potential and
properties were found to improve with increasing compatibility.
The solutions from which MC fibers are spun from can be isotropic or
anisotropic depending on the solution concentration [51 (Fig. 2).
When the solution is isotropic, viscosity increases with concentration in the normal man-
155
*
mtYLL
VICFIRERPROPERTIES
Poo-
g.70IQ4A4f
~
.. q.j
TVE At.PWo
P04PB~M'TADI 447*TWAL OI
POY
lFFCTIONCDITRIL
Pa'OITSZ~IL:ID:&TARI;FinUATRIX
'll
0 WOLFEOFSRI
PRZPOLYN43RROCESS
*IG M40*4W
NOUYATPADMIIVU0DWI.%DE AIRFIIaI) RDERED
*
L'.2DIRECTIONAL BAR PROPERTIES
*.
ramNER PROGRAM
a
44n
IIVEL.OPETIE PROES TECHNLOGY
*Du PONFT
FaOLLW BYFIRSTSCALELiP PRooufcnNs iI(41_
DENfltZR-FIAMAIT YARNFORTHlEAIR FORCE
* OTIS EARLYCC4ThRI!TIrRS CELISSE I OF
MASS.CARNECIE
MELLO'IISITR1E
*TE.N,%EMOI) . it I0 GPD~3 . 1% 4".
44
_
_
TAIIIF IV
PRZ7TVATRIX
CO%PATRILITV
ANDM(ICPPFTE
PZ
IPE
~
SPECIFICM&4RASIIAI,paPtaE44s
I
TA444.
UTP
TPA C
IR&IA
MATRIX
PIG.I
I.
n.*
W'0r~nn
3
ZT.0
AN WhiIc
YouRCA
60SI1.2
4551540
0A
*
*
SA
-
-3
'1~-
CCaaLUD
~~ ~
~
~
~
~
I-
~N
Z
-A -C1ELl
-l-a -A
S- -. PRPR
-.
'A
__________________________
____________________
11.
q
-
Of
_T_
1C4
is At
9
4.
W TYCPE
Me(,/txm~
KA
.0
xTO
T-MZ-%"POLMOCP-CLTUi
ONVSCST
I
lN.?E
ev,
SOLVENTS
5
156
Since PBZT is a liquid crystalline material, the increase in solution viscosner.
ity will reverse when it reaches the critical concentration of about 3.5% and the
The viscosity drops as sharply as the increase in the
solution becomes biphasic.
isotropic phase and then levels off at higher concentraiton.
State-of-the-art MC fibers (Table V) with tensile properties as good as
uni-axial conventional carbon/epoxy composites were obtained from below
Further
critical concentration using a Du Pont patented process technology (91.
achieved
from
higher
concentration
in
properties
was
improvement
The high meaanisotropic solutions also using the same Du Pont technology.
sured strength/modulus of 332 ksi/29 Msi (18.4/1,600 gpd) for the MC back calculated to 553 ksi/48 Msi (27.0/2,320 gpd) for neat PBZT indicates that rule-ofmixture has been achieved and, therefore, that we have a technologically sound
process.
Because of many technical problems in making MC's in prior art
a
proprietary mixed
II],
Du Pont developed
ternary
technology (10,
solvent/quaternary process which not only eliminated these problems but also
gained several process advantages, such as flexibility in accepting different
matrix resins.
Comparison between this novel process versus prior art is shown
in Table VI.
The quaternary process for MC fiber properties made from below
the critical concentration is between 1.5 to 2.5x higher than that from ternary.
Above the critical concentration, the quaternary process is 48 to 55x higher
than ternary.
Consolidated
Products
Except in applications such as ropes or soft armors, most high tech applications require the high performance fibers to be converted into composites,
which are typically articles composed of the fibers glued together with a matrix
resin at a typical 60/40 volume ratio of reinforcing to resin materials.
An "H"
mold was used to make direct-wound uni-directional composite bars.
Since MC
fibers already have built-in thermoplastic matrix, composites were made
without a pre-pregg-ag step, by simply applying heat and pressure to effect
consolidation.
Uni-directional MC composite bars have comparable flex,
compressive and short-beam-shear properties as conventional PBZT/epoxy
composites [2] and are just as deficient as PBZT in compressive strength as compared to carbon fibers [12].
As shown in Table VII, MC fiber precursors spun
from
below
critical
concentration
appears
to
yield
somewhat
higher
compres-
sive strength, but the compressive strength is still only 1/5th that of carbon
fiber-based composites.
MC film precursor offers no advantage in compressive
strength versus fiber precursor.
The tensile and compressive
significant
increase
over
other
strength
of PBZT,
low modulus
although
commercially
representing
available
a
organic
fibers [131, is in the range of other organic, high modulus fibers like Kevlar®.
Nevertheless, the compressive strength of 30-41 ksi is considered low for
primary structural applications.
The high tensile and compressive moduli of
PBZT. however, are as high as pan-based carbon fibers [14].
In addition, the fact
that PBZT is not an electrical conductor otfers an advantage over carbon fibers
in applications where high stiffness, non-conductive and corrosion resistance
are important(Table VIII).
An advantage of using thermoplastic matrix is reprocessability of composites.
Uni-directional MC bars were found to recover 100% of their flex and
SBSS properties when they were reconsolidated by remolding the failed bars
after test failure.
However, a test of recovery of tensile strength after test
failure might be a more sensitive measure of repairability.
A correlation between fiber modulus and x-ray orientation angle was developed for PBZT (2] and its MC's.
Since the polyamide matrices in the MC's do
not contribute to the orientation significantly, the orientation angle values are
the reflections of the PBZT molecular orientation.
And since orientation angle
is not a function of PBZT content while modulus is a function of PBZT content,
157
TAwLE
V,
TERNARXYj QUOfAIERNARY
TARNICE VI
sTATE.OF.THF-ARTMP
.ATRlIX
CON4ATII.
,PSOTMl3369. 60/
)
30.700)
wTs
6g0CONTDOT.
DI&
ST~im Sa.Nmi
6
66
BELOWC~r4$I,.
rAILS
3311 ,
3-032,10)
C-,66
ABDLO
SACK0
C fl3
ABOVE C 66jij..l
T6NEATM0PZT
M30
*X6M.C06VEPOI4NAL
CO&@0600203
CAROOK0EPO"
STREN0330K360
3301
WMiUS. 66S 06
9
* 13
'
'
2,
OF06Bt66E
RATIO0
US
'
S113
ITC6LN6.O.
P6366613-tO
A34
OPI
PONT
US PAMNT.
em -
PBZTTPICITE3-HT
vALU TONEAT
*S-CA,60.CT16.30
C6 ,toi,
TOBELOW
ABOVE
6
003
661240H
.~m..
nS.
1. 03.
-O
CONN0ST6ONA 600,010
*EXC06DEE34M-AXIAL
IN
FIG. 4. pBUTI.3346 MC 31136 TENStLE
PROPERTIES. TJ30EPOXYVMI.O6IPOSITE
*
OC456115fl6I,
F13fI~~lS
~LE
~
ANE
TABLE
NNInNRCTUINAKI
YIN
636POE4CS
UF
LEANNNI
PRT
MC BR
PROPERTIN
(A.
S
w
KO
I
,Po
933
in"aC
2"(.43 C-,
3
Ill
1
3-681
140
3I
336lx
6
),.6....
334
3SIN
m
POP
63Z
66
AToW-0C-1-.
"n...,,.,r
-0/03
''
-~-uT666016066
(R0
3IM.
4i
Isi
"
....
636I3463)0163W6
*~~~~~~~~~~~~~~~
QIII
K-))6 43364n3
01331)67)4)6
31413)6
1..P3
00
I
158
the modulus vs. orientation angle relationship for MC's naturally falls below the
relationship for neat PBZT (Fig. 3).
Status of Process Development
Spinning and heat-treatment processes were successfully scaled-up from
Numerlaboratory, intermediate and finally to "production" scales.
ous 60-lb "production" spin batches were made to complete a total of 25 lbs of
1,000 nominal denier/290-filament MC yam.
Therefore, the prognosis of the
MC fiber tendeveloped processes for commercial viability appears very good.
sile properties from "production" runs are as good as intermediate-scale fibers
after heat-treatment (Fig. 4).
MC Data Base Development
The "production" MC yarn will be used to produce consolidated plates with
variety of orientation for testings in collaboration with Boeing Advanced Systems in order to develop the first comprehensive MC data base. A wide variety
of mechanical, electrical and optical tests will be performed.
SUMMARY
A promising thermoplastic polyamide matrix candidate has been identified.
Excellent spinning and heat-treatment processes were scaled-up from
Commercial viability appears very good.
laboratory to "production" scale.
Consolidated MC fibers have exciting properies and are repairable with 100%
recovery in properties.
Consolidated uni-directional bars have comparable flex,
compressive and short-beam-shear properties as conventional PBZT/epoxy
composites and with the same deficiency in compressive strength.
Modulus,
however, is comparable to pan-based carbon/epoxy in tension/flex/compression.
REFERENCES
I.
2.
3.
4.
5.
6.
7.
8.
9.
10.
II.
12.
13.
14.
W.C. Uy. Air Force Report AFWAL-TR-82-4154, Part 1, 1982.
J.F. Maznmone and W.C. Uy, Air Force Report AFWAL-TR-82-4154, Part i. 1982.
J.F. Wolfe and B.H. Loo, U.S. Patent No. 4,225,700 (30 January 1980).
T.E. Helminiak, C.L. Benner, F.E. Arnold and G.E. Husman, U.S. Patent No.
4,207,407 (10 June 1980).
P.J. Flory, Macromolecules 11,IM38 No. 6 (1978).
W.F. Hwang, D.R. Wiff and C. Verschoore, Polymer Engineering and Science,
23 No. 14, 789 (1983).
P.A. Small, J.Applied Chem., 3, 71 (1953).
B. Miller, M.A. Tallent, K.P. Hewitt, K.L. Adams and G.A. Desio, Textile Research
Institute Report No. 6 (1 July 1985).
W.C. Uy. U.S. Patent No. 4,810,735 (7 March 1989).
W.F. Hwang, T.E. Helminiak and D.R. Wiff, U.S. Patent No. 4,631,318 (23
December 1986).
W.F. Hwang, 6th Industry/Government Review of Thermoplastic Matrix
Composites, Arlington, VA, 1989 (unpublished).
I.Y. Chang and .K. Lees, J.of Thermoplastic Composite Materials, 1, 277
(1988).
M. Katz, E.I.du Pont de Nemours & Co. (unpublished).
S. Kumar, SAMPE Quaterly, 3 (1989).
159
MORPHOLOGY AND FORMATION OF FIBRILLAR STRUCTURE IN PBO FIBER
C.C. CHAU, J.H. BLACKSON, H.E. KLASSEN AND W.-F. HWANG
The Dow Chemical Company, Midland, MI 48674
ABSTRACT
Electron microscopy studies showed that the porous structure of PBO fiber may contain
fractal geometries; i.e., the void spaces are self-similar with variations in magnification. At the
fiber surface, a dense skin which consists of fibrils was observed. In the matrix, the fibril size
is about 5 to 50 am with the voids distributed randomly among the fibrils. The fractal
dimension of voids in PBO fiber as determined by microscopy and image analysis was found to
be 2.44. The fibrillated fiber showed a continuous fibril size distribution with no evidence of
fibril size hierarchy. These observations suggest a nucleation and growth mechanism for the
formation of the fibrillar structure in PBO fiber.
INTRODUCTION
The structure of high performance fibers is an interesting subject and has been studied for
many years. Published data indicates that organic fibers are highly oriented in the fiber direction
II].Studies on Kevlar aromatic polyamide fibers showed that rod shaped crystallites are
oriented [2-4] along the fiber axis with layered stackings [5-7] held together by H-bonds, and
with weak van der Waals' attractions in the lateral direction. Observations on PBT [8] have
shown that these fibers could contain fibrils, with a fibril size 10 rim or smaller, oriented along
the fiber axis. The fibrillar nature seems to allow the fiber to be fibrillated easily by both tensile
and compressive deformation [3, 9] or by a simple peel test [10].
Interest is further revolving around the possibility of fibrillar hierarchy. Although still not
clear, fibrillar morphology has been recognized as one of the most important features of organic
fibers since mechanical properties, such as high tensile strength and modulus, and weak
compressive strength [11], are believed to be closely related to the fibrillar nature of the fiber. In
a continuing effort to understand the fibrillar morphology and the mechanism of fiber formation,
microtomed thin sections of PBO fibers were examined by transmission electron microscopy.
Details of the fibrillar morphology were observed and analyzed quantitatively. Some
considerations regarding the fibrillar formation mechanism are given.
EXPERIMENTAL
(1)
Electron Microscov: PBO fibers were prepared from dopes of a copolymer of
poly(phenylene benzobisoxazole) (PBO) in polyphosphoric acid (PPA). The spun filaments
were collected in water, dried, and subsequently heat treated at elevated temperatures under
tension. Fibers with a diameter of about 15 um were embedded in epoxy resin. Flat silicon
embedding molds were used with the fibers oriented parallel to the long dimension of the mold.
After appropriate trimming of the epoxy embedded sample, ultramicrotomy was performed at
room temperature with a Reichert Ultracut E microtome to produce sections ranging in thickness
from 40-70 nm. New areas of a diamond knife were used continuously to avoid damage from
the cutting edge. Sections were collected on carbon supported copper TEM grids. Complete
intact sections without folds were found difficult to obtain. Thin sectioned samples were
examined using a JEOL 100 CX ATEM at an accelerating voltage of 100 KV.
(2) Image Analysis: Image analysis was performed by using a Kontron SEM-IPS image
analyzer. Segments of images from TEM photomicrographs were transferred to the image
analyzer using a high resolution TV camera. The captured image was then enhanced by
expanding the gray level to the full 256 gray levels. The processed image was compared with
the original image to insure proper selection of the desired region of the image.
Mat. Res. Soc. Symp. Proc. Vol.171.
1990 Materials Research Society
160
(3) Fiber Peel Studies: Strands of fiber were sandwiched between two pieces of Scotch
tape. The tape was then peeled apart. The peeled tape with the attached fibers was examined in
an ISI-DS I30C scanning electron microscope.
RESULTS AND DISCUSSION
(1) The Morphology of Fiber after Fibrillation
Sawyer and Jaffe [ 10] have suggested that fibers of liquid crystalline polymers may possess
a hierarchical nature, e.g., macrofibrils, fibrils and microfibrils. In the present investigation
fibers were examined on a macro and micro level in an attempt to determine if such a hierarchy
exists in PBO. Examination of the peeled PBO fibers showed separated segments of various
sizes as shown in Fig-i. The fibrils form a reticulated network with the larger fibrils mostly
oriented along the fiber direction. The possibility of tearing prohibits determination of the basic
structural unit by a hand peel test. The number and size distribution of the fibrils is shown in
Fig-2. The distribution is continuous with the most probable size occurring at 35 nm or lower.
Fibrillar size smaller than 20 nm is not clear or measurable from the micrographs. No
hierarchical distribution could be identified within the area examined.
Fig-I Fibrillar network produced by hand peeling of a strand of fiber
(2) Skin-Core Structure and the Morphology of PBO Fiber
An ultramicrotomed thin section of a fiber prepared from a copolymer of PBO is shown in
Fig-3(a). The fiber cross-section consists of two regions of distinctly different
characteristics. A low contrast region is seen around the periphery; it is probable that this is a
skin. However, examinations of fibers from various spinning runs indicate that the skin
morphology is not prevalent in PBO fibers. The average thickness of the skin in this particular
sample is about 0.8 um as shown in Fig-3(b). The skin seems to consist of densely packed
fibrils.
24
20
4
000a
01
02.03 0406 081
234A
6681
2
3 4
66a10
Fig-2 The number and size plot of fibrils in Fig-1(a) and (b) showing a continuous
distribution with no apparent size hierartchy
4
Fig-3 (a) The morphology of amicrotomed thin section of a PBO copolymer fiber showing
skin-core structure (Mhe dark stripes are folds of sample produced in microtomning)
(b) An enlarged view of a portion of skin in Fig-3(a) showing densely packed
fibrils in the skin
In the core, irregularly shaped fibrils ame seen with boundaries separating them as shown in
Fig-4. The size of the individual fibril appears to range from about 5 to 50 nm and no size
hierarchy was observed. The size range is consistent with the order of 10 nm reported in PBT
(81. Individual fibrils with abnormally large or small sizes are not observed. These fibrils are
presumably oriented along the fiber axis. Another feature evident in Fig-4, is the contrast
162
variation, which could indicate the presence of voids. A separate study by small angle x-ray
diffraction concurred with such a possibility. These voids appear to be irregular in shape and are
distributed randomly among the fibrils. The voids, although low in volume content, are well
dispersed in the matrix with the size ranging from submicron to nanometer level. As is evident
in Fig-3(b), some void area is created in this fragile material during microtomy.
:
100 nm
Fig-4 A detailed view of the cross-sectional morphology of fibrils in the core of a
PBO fiber
Assuming that the relatively bright features are voids, the size and number of voids were
measured from TEM micrographs using an image analyzer. The average void size was
measured radially across the fiber at a magnification of 50 kx. The void size did not vary
systematically with distance from the fiber center and the void spacing was variable. The void
area percer.vige in the cross-section was measured at three different magnifications (10.8 kx, 97
kx, 216 kx) in order to detect voids of all sizes. The void area percentage was plotted as a
function of the average void size for each area on a log-log scale as shown in Fig-5. While
there is considerable scatter there is an inverse relationship between void area percentage and
void size, i.e., most of the void volume is contributed by the smaller voids.
If the void space is considered to possess fractal [ 12] geometry within the studied
magnificaL,)n range, the fractal dimension may be determined from Fig-5. For a fractal object.
the fractal a imension can be determined based on the number and size relationship: n - r-D
(0<D<3), where n is the number of pores, r is the size of the pore, and D is the fractal
dimension or Hausdorf dimension which relates number and size for a fractal object. For
relating the irea ratio of a species in a unit area to the size of the species, r2 n - r2-D (2<D<3),
The fractal dimension as calculated from the slope in Fig-5 is 2.44.
(3) Phase Separation and Fibril Formation
The fractal nature of voids implies a kinetic process of fibril formation resulting from phase
separation. According to Flory's consideration [ 13], the phase diagram for a ternary system
such as PBO, PPA and water, a nonsolvent, would consist of a narrow single phase region and
a much broader two phase region where the homogeneous PBO/PPA solution is separated into a
polymer-rich and a polymer-lean phase. It is therefore expected that voids will develop in fibers
spun from a dry jet-wet spinning process in which the PBO/PPA dope is precipitated in water.
Subsequent drying usually causes fiber to shrink. Complete densification, however, is not
likely to occur.
163
Some mechanistic understanding can be obtained from the void characteristics. Since the
average void size measurement across the fiber diameter shows no signs of spacial periodic
variation, spinodal decomposition is not likely. Consider a polymer solution which is in contact
with a precipitant, such as water, where precipitation takes place by heterogeneous nucleation,
as time goes by the nuclei grow in size while new nuclei are developing. The process continues
until the polymer composition is consumed. At that stage the polymer is prcipitated in the form
of a reticulated network. Examples showing such a precipitated polymer stcture are those of
asymrnetric microporous membranes [14-16]. When the precipitation process occurs in a dope
that is uniaxially oriented, the network is likely to develop in elongated fibrillar form. This
consideration suggests that all fibrils are interconnected with voids between the fibrils. It is
expected that the void density is proportional to fibril density. This relationship is assumed
unchanged during drying and shrinkage. The relationship between number and size of fibrils
can then be approximated by following a fractal growth model [17] proposed for rock
formation.
40
lII
'1
I I
PO Cowfyme, Ftber
3.0
2.0
0D
0.8
0
2
S
-
244
F
-5
0.3
02
10
20
30 40 50 60
60
100
200
300 400
600
00 000
Man Vo'dSze. A
Fig-5 A log-log plot of cross-sectional void area percentage vs void size for a PBO fiber
4
,
The rate of nucleation can be expressed as dN/dt = K1 Cn ,and the rate of growth is dRldt
K2Cr, where N is the number of nuclei, P. is the size of fibril, C is the local polymer
concentration, K, and K2 are the respective rate constants, and n and r are the orders of
nucleation and growth. For the coagulation of rigid rod polymers, the local polymer
concentration, C, is assumed to be close to constant throughout in the shaped dope during
precipitation.
This is likely to be true due to the low mobility of rigid and oriented polymer
chains,
Since the growth process depends upon local concentration to form an extended aggregate,
the rate of growth can be reasonably assumed to be first order. The final fibril size should then
be proportional to the rate of growth but inversely proportional to the rate of nucleation for a
given polymer concentration. Namely, the fibril size and number can not be both maximized at
any time t: Therefore, R is proportional to (dR/dt)/(dN/dt), or R - Ct-n. The total fibril
number N with a size R is expressed as N - C n .The relationship between the fibril number and
size can then be scaled as follows: N(R) - R -(/n-), where 0.(n/n-1)<3. By definition this
relationship shows that the number and size of fibrils are correlated by a fractal dimension:
164
n/n- 1. Based on fibril and void size proportionality, the fractal dimension gives the order of
nucleation, n=1.69. The kinetic equations governing the formation of fibrils may be expressed
as dN/dt = K1C' 69 , dR/dt = K2 C. These considerations suggest that the kinetics of fibril
formation may be indicated from a fractal analysis of the porous structure developed within the
fibrils.
SUMMARY AND CONCLUSIONS
(1) Electron microscopy showed that PBO copolymer fibers consisted only of fibrils with
varying sizes in the range of 5 to 50 nm. No hierarchy in the fibrillar size was observed.
(2) In an isolated case, one PBO fiber was found to have a 0.8 urn thick skin consisting of
closely packed fibrils.
(3) It was assumed, based on TEM image contrast, that a small void content (0. 1-I1 vol%)
existed within the PBO fiber. The small (2-100 rum) voids were randomly distributed
throughout the fiber.
(4) The void appeared to contain fractal properties. The void density and size gave a fractal
dimension of 2.44. Based on these observations, a nucleation and growth model is
suggested for the formation of fibrillar structure with a variation in the order of nucleation.
ACKNOWLEDGEMENTS
Discussions with professor JC.M. Li of the University of Rochester is gratefully
acknowledged. We wish to thank T. Helminiak of the Air Force (AFWAL) Research
Laboratories and D. McLemore of The Dow Chemical Company for supporting the work.
REFERENCES
I. W.W. Adams and R.K. Eby, MRS Bulletin, Nov 16/Dec 31, 22 (1987)
2. L.S. Li, L.F. Allard and W.C. Bigelow, J. Macromol. Sci.-Phys., B22(2), 269 (1983)
3. R.J. Morgan, C.O. Pruneda and WJ. Steele, J. Polym. Sci., Polym. Phys., 21, 1757
(1983)
4. M. Panar, P. Avakian, R.C. Blume, K.H. Gardner, T.D. Gierke and H.H. Yang, J.
Polym. Sci., Polym. Phys., 21, 1955 (1983)
5 M.G. Dobb, D.J. Johnson and B.P. Saville, J. Polym. Sci., Polym. Phys., 15, 2201
(1977)
6. S.C. Simmens and J.W.S. Hearle, J. Polym. Sci., Polym. Phys., 18, 871 (1980)
7. P.H. Young, Spectroscopy, 3, 9, 24 (1988)
8. Y. Cohen and E.L. Thomas, Polym. Eng. Sci., 25, 1093 (1985)
9. J.H. Greenwood, P.G. Rose, J. Mat. Sci., 9, 1809 (1974)
10. L.C. Sawyer and M. Jaffe, J. Mat. Sci., 21, 1897 (1986)
11. S. Kumar, IFJ/February, 4 (1989)
12. S.H. Liu, "Fractal and Their Applications in Condensed Matter Physics" in Solid State
Phys (Academic Pre.), 39,207 (1986)
13. P.J. Flory, "Principles of Polymer Chemistry", Chap. XIII, Corell Univ. Pre., 1953.
14. J.G. Wijmans, J. Kant, M.H.V. Mulder and C.A. Smolders, Polymer, 26, 1539
(1985)
15. D.M. Koenhen, M.H.V. Mulder and C.A. Smolders, J. Appi. Polym. Sci., 21, 199
(1977)
16. H. Strathmann, K. Kock, P. Amar and R.W, Baker, Desalination, 16, 179 (1975)
17. A.J. Katz and A.H. Thompson, Phys. Rev. Lett., 54, 1325 (1985)
165
IN SITU COMPOSITES BASED ON THERMOTROPIC AND FLEXIBLE POLYMERS
GUIDO CREVECOEUR AND GABRIEL GROENINCKX
Catholic University of Leuven, Laboratory for Macromolecular Structural
Chemistry, Celestijnenlaan 200 F, B 3030 Leuven, Belgium
ABSTRACT
Blends of a thermotropic liquid crystalline polymer (TLCP) and a thermoplastic matrix were compounded. Upon subsequent injection moulding and spinning,
the TLCP was deformed into fine fibrils in the matrix, giving in-situ reinforcement. Especially after spinning, the composite fibres contain fibrils
with very high aspect ratio, and exhibit mechanical properties in accordance
with simple composite models for modulus and strength.
INTRODUCTION
In-situ composites consisting of a thermotropic liquid crystalline
polymer (TLCP) and a thermoplastic matrix material can be regarded as an intermediate between conventional (short) fibre reinforced composites and
molecular composites. All three of these materials have in common that the
aspect ratio of the reinforcing species is of great importance with respect
to the final mechanical properties of the product. In short fibre composites
the length of the fibres is limited by the processing method used, e.g. extrusion or injection moulding. A high aspect ratio can be obtained in molecular composites, where one or a few stiff single macromolecules are dispersed on a molecular level. Due to their intrinsic immiscibility with flexible polymers, the rigid macromolecules, typically being iyotropic liquid
crystalline polymers, have to be dispersed in the thermoplastic matrix by
casting from dilute solutions. Indeed, dispersion on a nanometer scale and
outstanding mechanical properties can be obtained in such systems [1]. However, the solvent step provides an extra complication in industrial processing. Therefore, a reasonable alternative seems to be blending a TLCP with an
engineering polymer in the melt, so that dispersion, elongation and orientation of the TLCP take place in one processing step [2-6]. An additional advantage is that, because the TLCP acts as a lubricant, the melt viscosity of
the blend, contrary to those of short fibre reinforced composites, can be
lowered as much as one or two orders of magnitude [7-10].
The present paper is concerned with blends of a commercial thermotropic polyester-amide, Vectra B 950, in a matrix consisting of a
(miscible) mixture of polystyrene (PS) and poly-2,6-dimethyl-1,4-phenylene
ether (PPE).
EXPERIMENTAL
Vectra B 950, which is believed to consist of 58 mole-% hydroxy naphthoic acid, 21 mole-% terephthalic acid and 21 mole-% 4 aminophenol, was
purchased from Hoechst Celanese. Blends containing 70 weight-% PPE and
30 weight-% PS were kindly supplied by General Electric Plastics Europe.
Vectra B 950 melts at 280°C, where it shows liquid crystalline behaviour.
PPE and PS are known to be miscible over the entire composition range; the
as received PPE/PS blend exhibits a single T at 1770C.
Blending was performed on a Berstorf# corotating twin screw extruder
at 320 0 C. Strands with a draw ratio (DR) of approximately 2.5 were taken directly from the extruder, and their morphology was examined. From the granulated strands, ASTM tensile bars were injection moulded at 320°C, and fibres
were spun at approximately 310°C; the draw ratio was determined from the
Mat. Res. Soc. Symp. Proc. Vol. 11. ' 1990 Materials Research Society
1e
haul-off speed and the throughput. Although fibres with different draw ratios were prepared, the discussion in the present paper will be confined to
those having a draw ratio of approximately 30. The influence of draw ratio
on the morphology and properties of the blends is described elsewhere [11].
Complex viscosity of the blends was examined on a Rheometrics RMS 500
mechanical spectrometer at 300°C, in plate-plate geometry, at a strain of
I %, which is well in the linear region of Vectra. Dynamic modulus and loss
angle in the solid state as functions of temperature, were measured on a
Polymer Labs dynamic mechanical thermal analyser (DMTA) in bending mode.
Cryogenic fracture surfaces were investigated using scanning electron microscopy (SEM). Furthermore, the PPE/PS matrix from the strands and composite fibres was dissolved in toluene, to yield the insoluble TLCP particles or fibrils which were examined under an optical microscope. The mechanical properties of the blends were dqtermined on an Instron tensile tester,
applying P strain rate of I mm.min " for the injection moulded parts and
50 %.min- for the fibres. The fibres with low TLCP content have a high
strain to break, necking occurring over the entire sample; therefore the ultimate strength of these samples is corrected for the decrease in cross-sectional area, assuming incompressibility during necking. Molecular orientation of the TLCP in the fibres was measured using wide angle X-ray scattering (WAXS). Single fibres were placed in a Kiesig camera, and the orientation was characterized either by the peak width at half height (wh) or the
Hermans orientation parameter P2. Both parameters were obtained from an
azimuthal scan over the strong equatorial 110 reflection of the TLCP. Background scattering resulting from the amorphous matrix was subtracted from
the intensity profile prior to these calculations.
RESULTS AND DISCUSSION
The complex viscosity of the blends versus the TLCP content, for three
different angular frequencies, is presented in fig. 1. A substantial decrease in viscosity is observed for the higher TLCP contents. The fact that
the viscosity remains constant or even increases for the low contents, could
indicate the presence of a yield stress in the TLCP. Due to this yield
stress, the TLCP particles would act as rigid fillers at the low strain that
was applied during the measurements. Complex viscosity behaviour with intermediate extremes has also been reported for blends of flexible polymers
[12].
Storage modulus and loss tangent for a 50 %Vectra-50 %PPE/PS blend
(compression moulded sample) versus temperature are shown in fig. 2. Three
successive thermal transitions are observed:
the Vectra -transition at 70°C
[13], the Vectra glass-transition at 140 0 C, and the PPE/PS T at 1800 C.
Fig. 2 clearly demonstrates the single T of the miscible PPE/PS natrix.
Fig. 3 contains scanning electron nicrographs of an injection moulded
tensile bar, an extruded strand and a fibre. It is readily seen that fibril
formation ismuch more developed in the strand and the fibre than in the injection moulded part. Spinning and strand-extrusion involve extensional
flows, whereas during injection moulding merely shear flow (a weaker flow
due to its non-zero rotational component) takes place. When the PPE/PS matrix was dissolved in toluene, itappeared that the as spun fibres contained
microfibrils with nearly infinite aspect ratio, in contrast with the only
moderate aspect ratio of the fibrils in the injection moulded samples.
Because of these high fibril asp('-t ratios, the best mechanical properties
can be expected from the composite fibres. There is a tendency for the
fibril diameter to increase with TLCP content from 0.2-0.6 Ipmfor the
5 weight-% Vectra blend, to 0.4-2 pm for 25 %TLCP. The fibril thickness is
deteriniiied by the particle size prior to elongation, which is the result of
a dynamic equilibrium between dispersive mixing and coagulation in the twin
scr'w extruder.
187
10000
2
0
,
-
C
000
-2
10
0
10
20
30
40
Voklua % Vecira B950
Figure 1. Complex viscosity versus
volume fraction TLCP, for 101
+),
1.6 (A) and 0.10 (o) rad.s-
100
pm
50
a
so
tO
ISO
250
200
T rCl
Figure 2. Storage modulus and loss
tangent versus temperature,
50 1 Vectra-S0Z PPE/PS
for
blend, freq.-1O Hz
Figure 3. Scanning electron micrographs
of cryogenic fracture surfaces.
Injection moulded tensile bar,
10 Z TLCP (a), strand, 25 Z TLCP (1),
and fibre.
100 Im
10 Z TLCP (c)
10pm
C!
168
The tensile modulus of the injection moulded bars and the fibres, as a
function of TLCP content is plotted in fig. 4a. The only moderate aspect ratio of the TLCP fibrils in the injection moulded samples gives rise to a
modulus below the rule of mixtures [14]. Indeed, a curved plot is observed
in fig. 4a. The tensile modulus of the fibres is seen to increase linearly
with volume fraction TLCP. This is in full agreement with the morphological
observation of microfibrils of nearly infinite aspect ratio that, according
to composite theory [15], requires rule of mixtures for modulus:
Ec=VfEf+(1-Vf)Em
(1)
where Ec, Ef and E are the modulus of the composite fibre, TLCP and matrix
respectively, and V represents the volume fraction of TLCP. Moreover, the
modulus of 74 GPa for the pure Vectra, suggests a satisfactory molecular
orientation [16] for both the pure TLCP, and the TLCP dispersed in the composite fibres. Elongation to break for the fibres (plotted on a logarithmic
scale in fig. 4b) falls from approximately 100 % for the pure PPE/PS to I %
for the pure Vectra. Of course, this transition has great influence on the
ultimate strength (fig. 4c). A minimum occurs at 20 volume-A TLCP, which can
be understood qualitatively by indicating two limiting cases:
i) For high volume fractions of TLCP, the fracture behaviour will be dominated by the TLCP, i.e. the composite will break when the maximum strain of
the TLCP is reached, and the strength can be described by the following
equation [13]:
oc=Vfof+(1-Vf)ofEf/Em
(2)
where ac and of are strength at break of the composite and TLCP, respectively.
ii) For low TLCP contents, the TLCP cannot follow the plastic deformation of
the matrix (possessing very high strain to break), and may therefore be regarded as voids. In this case the composite strength is given by:
Oc=(l-Vf)- m
(3)
with a the strength of the pure matrix material. Equations 2 and 3 are
plotteT in fig. 4c. The minimum in the experimental data is seen to coincide
with the volume fraction at which equations 2 and 3 cross, but all experimental points are below the theoretical lines. However, during the derivation some assumptions are made that are only partially fulfilled. One of
these is that for the higher volume fractions, the TLCP is assumed to exhibit linear deformation behaviour up to break, which it does not. Furthermore, for the lower LCP contents, the strength according to equation 3 is
based on the strain to break for the pure matrix material, fig. 4b however,
clearly demonstrates that the presence of the TLCP actually constrains the
plastic deformation, thus also limiting the ultimate strength.
The Hermans orientation factor and peak width at half height versus
volume fraction TLCP are shown in fig. 5. Molecular orientation of the TLCP
as dispersed in the composite fibres is seen to improve with increasing TLCP
content. Similar results have been observed for blends of polycarbonate and
a thermotropic copolyester of poly(ethylene-terephthalate) and p-hydroxybenzoate [17]. A tentative explanation for this feature lies in the size of the
TLCP droplets in the melt, just before entering the spinneret. Since the
lower TLCP content blends contain smaller TLCP particles, these will have a
higher resistance against deformation, due to interfacial tension [18,19].
Thus, a smaller deformation is imposed by the macroscopic flow, leading to a
lower molecular orientation.
169
100
B0
70
*
60
~
50
10
50I
40
300
.3
L
20
10
I.
~~~~~0.11-___________
0
20
0
-
40
60
so
0
100
60
40
20
Vokirne % Vactra 6950
VOITI.
80
% Vectra B950
(b)
(a)
Figure 4. Young's modulus (a) and strain to break (b) versus
volume fraction TLCP. 0 injection moulded parts, A fibres
Boo
1.00
25
0.80
20
700
A600
~500
15
400.60
1300
100
A'
0
0.20
0
20
40
60
s0
VOW*j % Vectra 8950
of
10
0.40
200
Figure 4c. Strength for
composite fibres versus
TLCP content
100
0
20
40
60
80
5
100
Voklna % Vectra B950
Figure 5. Hermans orientation
factor (P ), and peak width at
half heig t (wig)versus volume
fraction Vectra
100
170
CONCLUDING REMARKS
Blends of a thermotropic liquid crystalline polymer and a matrix consisting
of flexible polymers were injection moulded and spun into fibres. In both
cases the formation of thin TLCP fibrils, oriented in the processing direction was observed. However, the average aspect ratio of these fibrils was
found to be moderate in the injection moulded samples, but nearly infinite
in the composite fibres. These observations clearly point out how much more
effective elongational flow (spinning) is than shear flow (injection moulding) in inducing elongated structures. The mechanical properties of the composite fibres could be described by composite models for transversely
isotropic geometry, i.e. rule of mixtures for modulus and an initial decrease, followed by an increase for strength versus TLCP content. Satisfactory molecular orientation of the dispersed TLCP fibrils in the composite
fibres was achieved.
ACKNOWLEDGEMENT
The authors are indebted to DSM Research: W. Bruls for compounding, R.
Keulers for spinning and C. van Halen for help with the X-ray work. Financial support to the laboratory by the Belgian National Foundation for Scientific Research (NFWO) is acknowledged.
REFERENCES
1. D.R. Wiff, T.E. Helminiak and W.F. Hwang, in High Modulus Polymers,
edited by A.E. Zachariades (Marcel Dekker, New York, 1988), p. 225
2. G. Kiss, Polym. Eng. Sci. 27, 410 (1987)
3. R.A. Weiss, W. Huh and L. Nicolais, Polym. Eng. Sci. 27, 684 (1987)
4. P.D. Frayer, Polym. Comp. 8, 379 (1987)
5. A.I. Isayev and M. Modic, Polym. Comp. 8, 158 (1987)
6. W. Brostow, T.S. Dziemianowicz, J. Romanski and W. Werber, Polym. Eng.
Sci. 28, 785 (1988)
7. K.G. Blizard and D.G. Baird, Polym. Eng. Sci. 27, 653 (1987)
8. B.L. Lee, Polym. Eng. Sci. 28, 1107 (1988)
9. A. Siegmann, A. Dagan and S. Kenig, Polym. 26, 1325 (1985)
10. M.R. Nobile, E. Amendola, L. Nicolais, D. Acierno and C. Carfagna,
Polym. Eng. Sci. 29, 244 (1989)
11. G. Crevecoeur and G. Gioeninckx, submitted Polym. Eng. Sci.
12. M.V. Tsebrenko, Intern. J. Polym. Mater. 10, 83 (1983)
13. T.S. Chung and P.E. McMahon, J. Appi. Polym. Sci. 31, 965 (1986)
14. J.C. Halpin and J.L. Kardos, Polym. Eng. Sci. 16, 344 (1986)
15. M.R. Piggott, Loadbearing Fibre Composites (Pergamon, Oxford, 1980),
p. 62
16. T.S. Chung, J. Polym. Sci. B26, 1549 (1988)
17. S.H. Jung and S.C. Kim, Polym. J. 20, 73 (1988)
18. G.V. Vinogradov, N.P. Krasnikova, V.E. Dreval, E.V. Kotova,
E.P. Plotnikova and Z. Pelzbauer, Intern. J. Polym. Mater. 9, 187 (1982)
19. P.H.M. Elemans, Ph. D. Thesis, Eindhoven University of Technology, 1989
171
MOLECULAR COMPOSITES OF RODLIKE/FLEXIBLE POLYIMIDES
S. R. ROJSTACZER, D.Y. YOON, W. VOLKSEN AND B.A, SMITH
IBM Research Division, Almaden Research Center, San Jose, CA 95120
ABSTRACT
Mixtures of a rodlike and a flexible polyimide were prepared
by solution-blending of the respective poly(amic alkyl eater) and
poly(amic acid), followed by solvent evaporation and thermal
imidization. The size scale of the phase separation, as measured
by light scattering, is primarily set during the demixing of the
precursor polymers, with no significant coarsening observed due
to the imidization performed at 400'C. The observed variation of
the domain size with parameter% such as composition, molecular
weight and film thickness is discussed in terms of the miscibility
of the precursor polymers as well as the thermal history to which
these were exposed.
INTRODUCTION
Mixtures of rodlike and flexible polyimides have the potential to achieve the desired properties for applications in microelectronic packaging by combining the required thermal
and
mechanical properties characteristic to the rodlike component
along with enhanced adhesion strength provided by the flexible
polyimide. The use of such mixtures as thin film coatings, however, requires that any heterogeneity on composition has to be
kept on a scale well below the micron level, raising the question
of the feasibility of polyimide based molecular composites.
Of
particular interest in such systems is the effect of the
imidization carried out at high temperatures on the morphology of
the precursor blends.
Yokota et al. [1 studied mixtures of polyimides prepared by
imidization of mixtures of different poly(amic acid) (PAA) precursors. Feger [2] and Ree et al. [3] showed, however, that mixtures of poly(amic acid)s can lead to the formation of copolymers
rather than physical blends of homopolymers.
The exchange reaction leading to the copolymer formation occurs due to the fact
that the poly(amic acid) coexists with a small amount of anhydride
and amine form through chain scission and recombination. The exchange reaction takes place either slowly at room temperature or
during a slow curing procedure.
Recently, Ree et al. [4) showed that polyimide/polyimide
blends can be formed from stable precursors mixtures if at least
one of the poly(amic acid)s is replaced by its alkyl ester derivative.
The poly(amic alkyl ester) (PAE) form provides a more
stable polyimide precursor system than its PAA counterpart by
eliminating the possibility for reequilibration reactions.
In the present paper, the polyimide blend system introduced
by Ree et al. [4) is used to study the relationship between the
miscibility of the precursor mixtures, the drying and curing conditions and the final morphology of the imidized mixtures.
Mal. Res. Soc. Symp. Proc. Vol. 171. '1990 Malerials
Research Society
172
EXPERIMENTAL
Fig.l shows the chemical structure of 6F-BDAF and PMDA-PDA
polymers synthesized for the present study. 6F-BDAF polyamic acids were prepared via classical solution polycondensation of sublimed hexafluoroisopropylidene diphthalic anhydride (6F) with
distilled
2,2-bis(4-aminophenoxy-p-phenylene)hexafluoropropane
(BDAF) in N-methylpyrrolidone (NMP). The degree of polymerization
(DP) was controlled by the monomer stoichiometric imbalance using
the diamine in excess. In this way, samples with estimated weight
average molecular weight of 10,000 and 48,000 were obtained (DP=l0
and DP=50, respectively).
Meta.PMDA/PDA poly(amic alkyl ester)
was prepared via low temperature solution polycondensation in NMP
of the diacyl chloride with freshly sublimed p-phenylene diamine
Meta-diethyl dihydrogen pyromellitate vwas prepared
(PDA) (6].
similar to the procedure of Bell and Jewell [7] except that after
most of the para-isomer had been selectively removed by slow concentration of the excess ethanol, the remaining ethanol was
stripped and the residue twice crystallized from n-butyl acetate.
This rendered the meta-isomer in greater than 90 % isomeric purity. The corresponding diacyl chloride was prepared by reaction
of the diacid with an excess of oxalyl chloride in ethyl acetate
at 60'C. The ethyl acetate was then stripped and the crystalline
residue twice crystallized from hexane. The molecular weight of
PMDA-PDA precursor measured by light scattering is 200,000.
Transparent common solutions of various PMDA-PDA/6F-BDAF ratios and a total polymer concentrations in the range 15-20% were
prepared by mixing of the homopolymer solutions.
0
o
PMDA-PDA Poilyamic Alkyl Eser
EtOH
1 -2
PMDA-PDA Polyimide
0
,L
CF
1I 3
C
O1
c\
CF 3
|
I
6F-SDAE Polyamic Acid
1 -2 H20
S 0
0
CF3
I
IO
CF 3
/
01
0
6F-BDAF Poilyimde
Fig.l. Chemical structure of PMDA-PDA and 6F-BDAF.
173
The ternary phase diagrams were determined by cloud point
measurements performed during isothermal solvent evaporation at
75'C using a 2 mw red He-Ne laser and a photodiode detector positioned at 90' from the incident beam.
The compositions corresponding to the binodal curve were calculated based on the weight
loss of NMP at the cloud point.
Films of PMDA-PDA/6F-BDAF precursors were prepared either by
doctorblading or by spinning, followed by a drying step performed
either at 75"C for 90 min. or at 150' C for 15 min. The respective
polyimide films were obtained by curing the dried precursor films
at 400'C for one hour, using a heating rate of 5°C /min,
Film
thickness were measured by an Alpha-Step 200 (Tencor Instruments).
The morphological observations were made on a Amplival Jena optical microscope operated in the phase contrast mode. The scattered light intensity from flat films was measured as a function
of angle at a wavelength of 632.8 nm. The source was a polarized
He-Ne laser with output power 1.8 mW, beam diameter 0.75 mm
(Ie 2 ), and beam divergence 1.1 milliradian.
The light electric
field was perpendicular to the plane of incidence and a film
polarizer oriented parallel to the incident polarization was used
in front of the PIN photodiode detector, to record the VV intensity.
The measured intensity was corrected for variation in
scattering volume and reflectance at the sample surfaces to
produce normalized scattering intensity values.
RESULTS AND DISCUSSION
Fig.2 shows the ternary phase diagram of the PMDA-PDA PAE /
6F-BDAF PAA / NMP system measured at 75'C, for two different molecular weights of 6F-BDAF.
The compositions are in units of
weight fraction, and the numbers on the right side of the ternary
diagram are the overall polymer concentration corresponding to the
compositions lying along the horizontal lines. Mixtures containing the low molecular weight 6F-BDAF in up to 15% of the solid
contents did not show a detectable cloud point.
Drying of these
cosolutions for up to 40 min. yielded a transparent solid film.
NMP
o 6F-BDAF M w = 48.000
* 6F-BDAF M w = 10,000
0.3
0Z
0.5
0.7
PMOA-POA
6F-BDAF
Fig.2. Ternary phase diagrAms meastired at 75'C (see text).
174
When dried at low temperatures, polyamic acids are known to
be able to solidify while retaining a considerable amount of solvent. This gelation process is characterized by a large increase
in the viscosity of the system and it affects some important
properties such as planarization (6].
The gelation of polyamic
acids in NMP has been attributed to the formation of PAA/NMP complexes stabilized by hydrogen bonding (7 - 9].
In order to investigate the possible effect of the gelation on the phase separation
of the ternary mixtures, the drying of the homopolymer solutions
was studied by isothermal thermogravimetric analysis. Fig.3 shows
a plot of the polymer concentration as a function of annealing
tims at 75'C. The gelation is characterized by a substantial decrease in the rate of solvent evaporation.
For the three solutl.ons studied, the gelation takes place at polymer concentrations
in the range of 62-65%.
100
80
PMDA-PDA
6F-BDAF (10K)
0
............
6F-BOAF (48K)
60
0
040
120
Time (min.)
Fig.3. Polymer concentration during drying at 75°C.
The concentrations corresponding to the gelation point are
represented by a dotted line in Fig.2.
If the binodal curve of
the low molecular weight 6F-BDAF mixtures is extrapolated to
PMDA-PDA concentrations in excess of 85% of the solid contents,
such a line would fall below the gelation line. Thus, the transparency of the mixtures in this composition range is believed to
be a consequence of the large viscosity increase occurring at the
gelation, which prevents phase separation, at least at long
scales.
Fig.4a shows the optical micrograph recorded from a 50/50
PMDA-PDA/6F-BDAF (48K) blend dried at 75'C.
The phase separated
structure is characterized by spherical domains of diameter in the
range of 1-3 pm. The light scattering curve measured from the same
film is shown in Fig.4b. The scattering is represented as relative intensity versus the amplitude of the scattering vector q,
defined as q
-sin-.The peak in the scattering intensity
indicates that the spatial concentration fluctuations have a
characteristic wavelength. The domain size of the phase separated
structures can be defined as d=2-/qmax, qmax being the scattering
vector corresponding to the maximum intensity. Fig.4b also shows
that upon imidization, the scattering intensity increases by an
order of magnitude while qmax remains unchanged (d-l.Oim).
In
other words, the imidization appears to cause an increase in
scattering contrast only, without significantly affecting the
phase separated morphology. This behavior was observed in all the
mixtures studied.
.4
175
(b)
4i
idized
0.1
dried
a
aC
0.01
-N
2
4
6
8
10
12
1
q (pm- )
20pm
Fig.4. Phase-contrast micrograph from a 50/50 (48K) film dried
at 75 'C (a), and its light scattering curves (b).
Apart
from
a
slight
composition dependence,
the
domain
size
measured from the mixtures containing the higher molecular weight
6F-BDAF was found to be insensitive to other parameters such as
drying conditions.
This is in contrast to the behavior exhibited
by the blends containing the low molecular weight fluoro-polymer,
which showed that their phase separated morphology can be affected
by the rate of solvent evaporation.
This is exemplified for the
most sensitive mixture, i.e., the 75/25 PMDA-PDA/6F-BDAF (10K) in
Fig.5, whicX shows the effect of film thickness and drying temperature on the shape of the scattering curves.
By changing the
thickness of the dried films from 70 to 15,m, the domain size drops
from about
1.0 to
0.61,m, while in the
81#m thick film, the disap-
pearance of the scattering peak accompanied by the decrease in the
scattering intensity indicates that if any phase separation occurs, it is limited to scales below the wavelength of light. The
effect of thickness is believed to be related to the mass transfer
involved in the solvent evaporation.
V'Ao0jm
(a)
Ib)
md.,zed
•15/m
l~m
•
02
4
- 01
dried
006
Bpm
004
'/
002
4
C
12
8
q (m
,
)
0
4
12
8
q
Im i
Fig.5. Scattering curves from ?5/25 PMDA-PDA/6F-BDAF (10K) films:
(a) dried at 75 IC; (b) a 2 51,m film dried at 150'C.
16
176
Transparent, homogeneous films were also obtained by drying
the 75/25 PMDA-PDA/6F-BDAF (10K) mixtures at 150'C into films of
thickness of up to 251,m, which is the limiting thickness for obtaining f4 ims at 150'C without the formation of voids. It should
be pointed out, however, that the phase diagram as well as the
gelation are expected to be a function of temperature, and the
effect of drying temperature can not be interpreted in terms of
an increased solvent evaporation rate only.
The effect of molecular weight, film thickness, and composition on the domain size can be rationalized as a kinetic effect
having to do with the difference in the polymer concentrations
corresponding to the onset of phase separation and gelation. The
fact that the morphology of the low molecular weight blends was
found to be sensitive to the film thickness indicates that the
time interval between the onset of the phase separation and the
gelation is comparable to the time scale of the structure
coarsening. This is in agreement with the observed trend that the
sensitivity to such parameters was more pronounced in mixtures for
which the binodal concentration is closer to the gelation point.
CONCLUSIONS
The morphology of the polyimide/polyimide mixtures studied
was found to be set at the gelation point of the precursor
cosolutions.
The domain size of the phase separated morphology
is affected by parameters such as film thickness and drying temperature, in particular in the more miscible mixtures. Transparent, fully imidized films of a rodlike polyimide containing up to
25% of a flexible component were obtained.
REFERENCES
1. R. Yokota, R.Horiuchi and M. Kochi, J. Polym. Sci.:Part C 26,
215 (1988).
2. C. Feger, in Polymeric Materials for Electronics Packaging and
I_terconnection, edited by J.H. Lupinski and R.S. Moore, ACS
Symposium Series, 407, Washington 114 (1989).
3. M. Ree, D.Y. Yoon and W. Volksen, to be published.
4. M. Ree, D.Y. Yoon and W. Volksen, to be published.
5. V.L. Bell and R.A. Jewell, J. Polym. Sci.:Part A-1, 5, 3043
(1967).
6. S. Nishizaki and T, Moriwaki, J. Chem. Soc. Japan, 71, 1559
(1967).
7. D.R. Day, D. Ridley, J. Mario and S.D. Senturia, in Polyimides.
Vol.2, edited by K.L. Mittal, Plenum Press, New York, 767
(1984).
8. P.D. Frayer, in EQivmi__
Vol.1, edited by K.L. Mittal, Plenum Press, New York, 273 (19S4).
9. M.J. Brekner and C. Feger, J. Polym. Sci.:Part C 25
2006
(1987).
10. M.J. Brekner and C. Feget, J. Polym. Sri.:Part C 25 , 2479
(1987).
177
EQUILIBRIUM AND NON-EQUILIBRILIM PHASES
AND PHASE DIAGRAMS INBLENDS OF POLYMER
LIQUID CRYSTALS WITH ENOINEERINO POLYMERS
WITOLD BROSTOW *, THEODORE S. DZIEMIANOWICZ *, MICHAEL HESS *'
and ROBERT KOSFELD -'*
* Center for Materials Characterization and Department of Chemistry, University of North Texas, Denton, TX 76203-5371
*
Himont U.S.A., Inc., 800 Greenbank Road, Wilmington, DE 19808
FB6-Physikalische Chemle, Universitit Duisburg, Postfach 10 16 29,
D-4100 Duisburg I, Federal Republic of Germany
ABSTRACT
This work represents a continuation of earlier studies of blends of polymer liquid crystals (PLC) with ordinary engineering polymers (EP).
We
now focus on connections between mechanical and other properties and phase
structures and phase diagrams. Pure PLC are already two-phase systems; In
each case addition of an EP complicates the situation further. In particular,
we are concerned with phases which we call quasi-liquids, at temperatures
between the glass transition and the melting point.
Quasi-liquids do not
have the mobility usually associated with liquids - because of the presence
of other constituents and also because of orlentational effects produced by
the mesogenic groups. In phase diagrams of PLC-containing systems one
should also take Into account non-equlibrium phases. We are trying to show
how such diagrams make possible Intelligent processing and a better control
of properties of the PLC + EP materials.
INTRODUCTION
This paper represents a continuation of earlier work 11-41 aimed at mixing polymer liquid crystals (PLC) with ordinary engineering polymers (EP) In
such a way that the valuable properties of PLC (reinforcement by mesogenic
groups, stability at hi temperatures, low Isobaric expanslvity, etc.) are
preserved while costs are lower than that of pure PLC. However, In the present study we deal with variation of LC content in a series of PLC copolymers, while blends of one of these PLCs with an EP are dealt with in a companion paper [5.
While In earlier papers we used a variety of techniques and covered a
wide range of properties, the present work Is focused on phase structures
and phase diagrams. Reasons for this are stated In the following Sectlon
Met. RMe.Soc. Symp. Proc. Vol. 171. %1 "0 Marals Rmerch Sockey
178
PHASE STRUCTURES, PHASE DIAGRAMS AND INTELLIGENT PROCESSING
Properties of materials, polymeric or otherwise, are of course determined by chemical composition as well as by structures produced during processing. For PLCs the situation is more difficult than for EPs; we know already from the first paper by Jackson and Kuhfuss [61 how easily acquire
PLCs orientation during processing; high anisotropy of properties ensues.
While In some cases the anisotropy is desirable, control of the resulting
properties is possible only if we have suff iccient knowledge of morphologies
and phase structures, and also if we can locate each structure In the correspor,ding region in the phase diagram. While much interesting and useful
work on processing of PLCs and their blends has already been done, our particular approach consists in the determination of the appropriate phase
diagram first, and defining processing conditions only afterwards; we call
this approach Intelligent Drocessing. In the present paper we define some
peculiarities of phase diagrams of PLC-containing systems, and also we
present one phase diagram. A striking example of how knowleoge of the
phase diagram makes intelligent processing possible is provided in the
companion paper [5).
It is customary to show in phase diagrams equilibrium phases only.
However, PLC-containing systems, apparently because of the presence of
rigid constituents, seem to show non-equilibrium phases with high longevity. Hence, just as for inorganic glasses, we have to take these phases
seriously Into account; sluggishness combined with easy orientation of
mesogens In PLCs affects processing and resulting properties to a high degree.
Consequently, we postulate that non-equilibrium phases should be
included in phase diagrams of PLCs and PLC-containing systems.
In the course of our work we have dealt often with a phase which we
believe deserves a name. auasi-liui.
This Is a the non-crystalline part of
a semi-crystalline polymer between Its glass transition temperature and the
melting temperature. Except for elastomers when one then talks about the
leathery state (see for instance 171), one calls such materials simply liquids.
However, In the case of PLCs in particular, that name Is not appropriate,
not only the material does not exhibit the ordinary liquid mobility, but we
are more than one phase transition away from the isotropic liquid arising
from the same component.
Further, the material Is sluggish because of
the simultaneous presence of the crystalline component, while this is also
true for polymer "liquids" In ordinary semi-crystalline polymers, in PLCs
there Is an additional contributions to sluggishness or to maintaining low
mobility because of rigidity of the mesogens.
Finally, It Is in the quasiliquid phase that the process of cold crystallization can occur.
179
CHOICE OF THE SYSTEM
A variety of otructure of PLCo i possibe, a3 cla5fied by one of u5
[8,91. The reason for the development of this classification is the fact that
properties of a given PLC are defined first of all by the class to which it
belongs. For this reason, in a number of cases It is not enough to talk about
main-chain PLCs, since these can be longitudinal, orthogonal, stars, soft
discs or rigid discs [8,91. In the present work we have chosen a longitudinal
PLC since at the present time these seem to be best known and understood.
In particular, we have chosen the copolymers of poly(ethylene terephthalate)
(PET) with l-hydroxybenzoic acid (PHB) studied already earlier by us [1-4]
as well as by a number of other authors from different points of view 16, 10-
181.
THE PHASE DIAGRAM
We have studied the phase diagram of PET/xPHB copolymers in function
of the mole fraction x of the liquid crystalline (that is PHB) component. In
the determination we have used a variety of techniques, Including differential scanning calorimetry (DSC),
wide-angle X-ray scattering (WAXS),
thermomechanical analysis (IMA), as well as dynamic mechnical testing
with a torsional pendulum. Some measurements on samples of the same
composition were performed at several locations; hence this international
collaboration. Because of the limited space, we do not provide here experimental details.
The resulting diagram is shown in Fig. I, Non-equilibrium transition
lines are Included. We believe that in general they have to be included, since
some non-equlibrium phases have fairly long lives; inorganic glasses come
to mind in this context.
The regions in the diagram marked with Roman numbers contain the
following phases:
I
- PET crystals
Isotropic glass
ii - PET crystals, PHB crystals
Isotropic glass, nematic glass
III- PET crystals, PHB crystals
nematic glass
quasi-liquid
IV - PET crystals
quasi-iquid
180
T
400
350
/ I
ii
II
300
v
1500
100
J3
150
"
A
J.
0.0
Fig. 1
0.
0.4
0.6
molar fraction PHB
0.8
1.0
181
V
- PET crystals, PHB crystals
quasi-liquid
VI = nematic
VII - nematic, liquid
VIII - liquid
The symbols in the diagram correspond to the following sources:
o - This work, independently of the technique used and also independently of the location where the experiments were performed. Hence
some points come from more than one technique and/or more than one laboratory
O - Meesiri et al.
(121
0 - Jezlorny [18]
o - Chouet al. [17]
* - Viney et al.
[1]
a - Benson et al. [15]
- Gedde et al. [14]
V - Hedmark [ 16]
* - Jackson et al. [6]
0 - Kricheldorf et a: [131
An analysis of connections between the phases present and their posi-
I
tion inthe diagram and properties will be presented insubsequent papers, In
part already Inthe companion paper (51.
References
1. R.Kosfeld, M.Hess and K. Friedrich, Mater. Chem. & Phys. 1. 93
(1987).
2. W.Brostow, T.S. Dziemianowlcz, J. Romanski and W.Werber, Polymer
Eng. &Scl. 2L 785 (1988).
3. K. Friedrich, M.Hess and R.Kosfeld, Makromol. Chem. Symp. 16. 251
(1988).
4. F.Schubert, K.Friedrich, M.Hess and R.Kosfeld, Molec. Cryst. Liq.
Cryst. l5. 477 (1988).
5. R.Kosfeld, F.Schubert, M.Hess and W.Brostow, the following paper In
the same MRS Symp. Proceedings volume.
6. W.J. Jackson Jr. and H.F. Kuhfuss, J. Polymer Scl. Chem. JA 2043
(1976).
7. W.Brostow and R.D. Cornellussen, In Failure of Plastics, edited by W.
Brostow and R.D. Cornellussen (Hanser Publishers, Munich - Vienna New York, 1986), Chapter 1.
182
8. W.Brostow, Kunststoffe 7.
411 (1988).
W.Brostow, Polymer aL In press (1990).
J Menczel and B, Wunderlich, J.Polymer 56. Phys. 1. 1433 (1980).
C. Viney and A. H. WIndle, J. Mater. Scl. 12, 2661 (1982).
W.Meesirl, J. Menczel, U. Gaur and B. WunderlIch, J. Polymer Sci. Phys.
2k,719(1982),
13. H.R.Krlcheldorf, G.Schwarz, Makromol. Chem. 1L. 475 (1983).
14. U.W. Gedde, D.Buerger and R. H.Boyd, Macromolecules 2%. 988 (1987).
15. R.S.Benson and D.N.Lewis, Polymer Commun. 2. 289 (1987).
16. P. Hedmark, Ph.D. thesis, The Royal Institute of Technology, Stockholm,
1988.
17. Ch. Chou and S. B. Clough, Polymer Eng. & Sci. 2& 65 (1988).
18. A.Jezlorny, Pollmery 34 210 (1989).
9.
10
11,
12.
4
183
*
3
STRUCTURE AND PROPERTIES OF BLENDS OF POLYCARBONATE
AND POLY(ETHYLENE-TEREPHTHALATE-CO-p-HYDROXYBENZOATE).
PHASE DIAGRAM AND MECHANICAL BEHAVIOR
ROBERT KOSFELD*, FRANK SCHUBERT*, MICHAEL HESS*, ** and WITOLD
BROSTOW**
* Department of Physical Chemistry, University of Duisburg,
D-4100 Duisburg, Federal Republic of Germany
**Center for Materials Characterization and Department of Chemistry,
University of North Texas, Denton TX 76203-5371
ABSTRACT
The Investlgatlon of the thermal behavior of polymer blenos leads to
phase diagrams which Involve important Information 'bout the system.
From these diagrams, equilibrium as well as non-equilibrium phases can
be deduced and ranges of miscibility or partial miscibility of the polymers
become obvious. Hence the diagrams are of a great value for processing of
advanced polymer blends, especially If a polyphasic polymers such as
a polymer liquid crystal Is one of the constituents of the system.
INTRODUCTION
-and
Polymer blending has become a promising method in the search
for polymeric materials with enhanced thermal and mechanical properties
(and sometimes also with lower prices). To achieve this, studies were
made on blends of ordinary engineering plastics (EP) with Polymer liquid
crystals (PLC) in order to preserve high-temperature stability, reinforcement by mesogenic groups , low thermal expansivities , etc.
The large
number of polymers available and the different methods of mixing open
the way to numerous products. With a detailed knowledge of interrelations between blending procedures, morphology of the resulting blend,
mechanical as well as thermal behavior of the product, new materials can be tailored for sspecific applications
inearlier studies [I-4) the copolyester of poly(ethylene terephthalate and p-hydroxybenzoate (COP) was investigated In blends with poly(carbonate) (PC) and poly(ethylene terephthalate) (PET). Transesterification, phase behavior In the ternary system with solvent and solid state
morphologies of the pure copolyester were studied. The thermal behavior
of different compositions of the copolyester is reported In 151.
Met. Re. Soc. Syrup. Proc. Vol. 171." 11,0 Materlflls Research Society
I
j
184
MATERIALS
The liquid crystalline material used In this study Is a COP containing 40 mole 9 PET and 60 mole X p-hydroxybenzoate (PHB). This material, with weight-average molecular mass M - 19,000, was first prepared
by Jackson and Kuhfuss [6]. As discussed in [2], the polymer shows a
two-phase morphology with Island structure; see also [5] and papers cited therein. The EP was a poly(blsphenol-A-carbonate) with M - 31,000.
Most of the blends were prepared by dissolving certain amounts of
the pure polymers In a chloroform + trifluoroacetic acid solvent system
and subsequent combined precipitation of both polymers from this solution
with acetone. The precipitate was dried under vacuum and then compression molded or extruded.
EPERIMENTAL TECHNIQUES
The blends were analysed with light- and electron- microscopes;
calorimetric experiments were carried out with a Perkin-Elmer DSC-2
thermomechanical analysis was done using a Perkin-Elmer TMA; and the
dynamic-mechanical analysis was performed with a torsion pendulum
working at a constant frequency of 1 Hz made by Myrenne, Roetgen, FRG.
WAX studies were executed with synchroton radiation at DESY in Hamburg.
RESULTS AND DISCUSSION
From light microscopy it became obvious that phase separation
occured In the whole concentration range and that application of pressure
resulted In a fine and very homogeneous distribution of the phases. The
Identification of the phases could be done as only the frozen-in nematic
structures of the PLC are bIrefringent.
Calorimetry and dynamic mechanical analysis showed that there
exists a very complex phase structure with several non-equilibrium
phases. T~te locallsation and the dimensions of these non-equilibrium
phases Is strongly infuenced by the degree of dispersion of the phases and
by the thermal history of the sample. The non-equilibrium phases can be
stabilized In a metastable state.
The most Important part of the phase diagram with respect to the
mechanical properties Is shown in Fig. 1. A more detailed discussion of
the whole diagram (broken vertical lines, the nature of various phases)
will be the subject of future papers.
185
o'g
100
o
60
I
I
I
I
I
I
80
low T~( oP)
-
0
100
-a- weight % PC
Fig. I A part of the phase diagram of COP containing 60 mole 9 PHB with
PC. Note the partial miscibility between 80 and go weight $ PC.
The experiment also showed that the presence of COP assists - in fact makes possible - the nucleation of PC crystals; see Fig. 2. in pure PC under
identical conditions the nucleation was n& observed. COP did crystallize
too; the PHB-rich phases of COP could be Identified as the regions where
the crystallization took place; see Figures 3 and 4.
Fig. 5 shows the influence of annealing on the dynamic-mechanical
behavior. The storage modulus is increased; the second glass transition
line, which can be attributed to PC, is less steep.
In Fig. I there is a region of partial miscibility between about 80
and 90 mole X PC. In this concentration range PC Is capable of incorpora-
186
% crystallinity
60
to0
160
IC60
.
Fig. 2 Change in crystallinity during a heating process which lasted 45
min. WAXS experiment with synchrotron radiation. 80 weight S PC.
"4°•
degrees
Fig. 3 Temperature-dependent WAXS. 80 weight X PC.
187
PHB
bl
COP
PET
20
10
20
degrees
Fig. 4 WAXS powder diagram at 250C.
a) Poly (PHB) with high crystallinity;
b)COP annealed f or 18B
hours at 4 MIa and 21 0-220*C;
0) PET annealed for 60 min at 4lPa and 210-220 0C.
MPQ
Mpa
101.v
\a)
1600
Fig. 5 Influence of sample preparation (crystallinity) on the storage
modulus G'.
a)blend with 67 weight 9 PC annealed for 1 mi.;
b)blend with 67 weight XPC annealed for 180 mi.
Conditions: 4MPa and 2l0-2200 C.
188
llexural modulus
MPo 20001500-
10
0
m
204m0
2o
8
0
1060
0
oi
-weight-%
PC
Fig. 6 Influence of crystallinity and composition on compression-molded
samples. Conditions: 4MPa and 210-2200 C.
ting certain amounts of COP, as demonstrated by the decrease of the
glass-transition temperature of PC. By contrast, COP does not Incorporate PC: there Is no altering of the glass transition temperature of COP at
- 500C. As a consequence of the partial miscibility In this narrow concentration range, COP fibres which may be formed during an extrusion
process are glued together by a thin layer of PC. This results In an Increase of the flexural modulus which begins just Inthe concentration range
where partial miscibility Is observed; see Fig. 6. Further results related
to the flexural modulus, shear behavior and the relationship of properties
to the phase diagram of the pure COP [5] are In preparation.
ACKNOWLEGEMENT
We are Indebted to the Deutsche Forschungsgemeinschaft, Bonn, for
financial support.
REFERENCES
1. R.Kosfeld, M.Hess and K.Friedrich, Mater. Chem. & Phys. 18 93
(1987).
2. W.Brostow, T,S, Dziemlanowicz, J, Romanski and W.Werber, Polymer
Eng. & Sc1. 2. 785 (1988).
3. K.Friedrich, M.Hess and R.Kosfeld, Makromol. Chem. Symp. 1. 251
(1988).
4 F.Schubert, K.Frledrlch, M.Hess and R.Kosfeld, Molec. Cryst. Liq.
Cryst. 15. 477 (1988),
5. W.Brostow, T.S. Dzlemlanowlcz, M.Hess and R.Kosfeld, preceding paper In the same MRS Proceedings volume,
6. W.J. Jackson and H.F. Kuhfuss, J. Polymer 51, Chem. 4. 2043 (1976).
CHARACTERIZATION OF POLYQUINOLINE BLENDS USING SMALL ANGLE SCATTERING
WN-LI WU*, JOHN K. STTLLE**, JOSEPH W. TSANG** AND ALEX J. PARKER"
*Polymers Division, Materials Science and Engineering Laboratory, NIST,
Gaithersburg, MD 20899
**Colorado State University, Fort Collins, CO
ABSTRACT
To determine the compatibility between the rigid rod and the
flexible chain polyquinolines, both small angle x-ray and neutron
scattering measurements were conducted on blends containing deuterated
flexible chains.
The scattering intensities from both x-ray and neutron
were reduced to their absolute scales in order to remove the scattering
contribution from microvoids which tended to overshadow the signal of
molecular
origin.
Quantitative
information
regarding
the
molecular
dispersion in a 50/50 rigid rod and flexible chain blend was obtained. The
result indicated that this material was partially segregated but not to
the point of single component phases.
INTRODUCTION
Polyquinoline[l] is a candidate for molecular composites.
In a
molecular composite, rigid rod molecules are dispersed in a matrix of a
flexible polymer,
and thus individual molecular rods act as the
reinforcing fibers.
Molecular composites have the potential to combine
the performance of a conventional composite with easy processing of
thermoplastics while retaining excellent properties at high temperature.
In an ultimate molecular composite, each molecular rod should be
completely surrounded by flexible polymer molecules.
There is, however, a
strong tendency for the rigid rod molecules to aggregate and separate on a
larger scale.
To overcome this problem, a number of modifications to
chemical compositions are being explored.
There is still a need to
develop test methods to characterize the degree of dispersion in these
materials such that progress can be quantified. It has been demonstrated
by the present authors that small angle scattering (SAS) is a viable
technique for the structure determination of molecular composites in
bulk[2].
The presence of microvoids has hampered the use of SAS for
quantitative measurements of rigid rod polymers and their blends.
To
circumvent this difficulty, both small angle neutron and x-ray (SANS and
SAXS) experiments were carried out on the same specimen, Because of the
differences in the scattering contrast factors of the voids between x-ray
and neutron scattering, the void contribution to the scattering intensity
can be estimated and then removed.
This SAS scheme has been applied successfully to a polyquinoline
multiblock copolymer composed of rigid rods and flexible blocks[2].
The
result suggested that this copolymer was strongly associated but not yet
segregated into phases of single components. The sample was prepared by a
solution casting method, and it took a significant amount of time for the
polymer to precipitate.
In the present work the film samples were
prepared by using the extrusion equipment at the Wright Patterson Air
Force laboratory.
The extrudate was rapidly coagulated in a nonsolvent
bath. It is conceivable that the extent of mixing can be improved through
this extrusion process.
Both copolymer and blend samples were prepared.
The SAS results for the copolymer, as will be shown later, support the
notion that extrusion provides improved mixing compared to the solution
Mat. Res. So. Symup. Prec. Vol. 171. c1990 Matedsle Resesrch Society
casting samples.
EXPERIMENTS
The maltiblock copolymer used in this work has a number average
molecular weight of 150,000, with each block having a value of about
17,000.
This value corresponds to a degree of polymerization (DP) of 30
for each block.
The blends were made of 50/50 rigid rod and flexible
chain polyquinolines and their DPs were 660 and 330 respectively.
The
details for the chemical structure and the synthesis of this material can
be found elsewhere[l).
The flexible chain component in both the block
copolymer ind the blends was partially deuterated to enhance the neutron
To prepare the SANS and SAXS samples a number of
scattering contrast.
layers of the extruded films were stacked to randomize the molecular
orientation within individual films.
A deuterated polyethylene (DPE) specimen was used to cross-calibrate
the SANS and the SAXS instruments.
Both the SANS and SAXS measurements
were conducted at the NIST facilities. A silica gel specimen was used as
the secondary absolute intensity standard for the SANS measurements.
RESULTS AND DISCUSSION
Based on the density values of the copolymer and the homopolymers,
one has the following contrast factors for SAXS and SANS[21 intensity,
I(q), expressed in their absolute scales.
IsAxs(q) = 9.23 V(q) + 3.37 x 10-ZS(q).
2
IsA~s(q) = 0.855V(q) + 9.46 x 10- S(q).
(1)
(2)
V(q) and S(q) are the structure factor of the voids and the polymer
respectively.
Both quantities are expressed in terms of v which is the
molar volume of the repeat unit of a rigid rod segment, and its value is
2
3
6.995 x 10-2 cm
corresponding to a bulk density of 1.32.
The material
studied in this work is composed of three components; they are microvoids,
rigid and flexible chains with each component having a different
scattering cross-section. Equations I and 2 are merely an approximation
of a three-component scattering theory[31 for the case in which the volume
fraction of the microvoid is rather small in comparison with the other two
components.
It is noteworthy that the volume fractions of the voids and
the polymers are included in these factors V(q) and S(q).
The calculated results for V(q) and S(q) for the block copolymer are
given in Figure 1. The height of the S(q) maximum is about 70, and it is
substantially less than 400 as observed in the solution cast film[21.
Based on this finding we conclude that the extruded film has a more
As was
homogeneous molecular structure than the solution cast one.
estimated in a previous publication[2], the peak height for the block
copolymers with an ideal mixing is only 1.43.
Apparently the extent of
the molecular dispersion in the extruded film is still far from ideal;
partial segregation between the rigid rod and the flexible chain
polyquinoline still prevails.
The SAXS and the SANS results for the homopolymer blend lead to the
The theoretical value of S(q) will be
results given in figure 2.
estimated for the ideal case in which all the chains are randomly
For this
dispersed regardless of their compositions or their rigidity.
case one has the following expression for S(q)(4]:
191
1
1
1
$("'-(q"
=6
+
=()
S~(3)
2 X
(3)
X is the Flory-Huggins interaction parameter and Sjx(q) is the single
chain correlation function for the flexible chain polymer. Sz2 (q) is then
the correlation function for the rigid rod. It is noteworthy that the
factor involving the volume fraction is included in the functions Sll(q)
and S22(q).
160
1
- 1
I
120Figure 1:
Form factors
of the microvoids and
z
0
the multiblock
polyquinoline. The
sample was prepared by a
extrusion and followed
by a rapid coagulation
lid
80 -
t
process.
40-
o
160
1
1
1
1
1
120
Figure 2: Form factors
of the microvoids and
the blend of high
weight
Zo
omolecular
80
poLyqr
40-
0.00
0.02
Q (A-')
0.04
0.06
olines.
192
Based on the DP values of 330 and 660 and the radii of gyration of 159 A
and 860 A for polymer 1 and 2 respectively the calculated result of
equation 3 is given in figure 3. The radius of gyration of the rigid rod
is obtained from a light scattering measurement [l and the value for the
flexible one is estimated from a freely joint chain model; the monomer
The
length in its fully extended state was chosen as the step length.
solid curve on figure 3 is for the case in which X equals zero and the
value
dashed line is for X equal to 0.004 which is close to the critical
for phase separation. By comparing the experimental result (solid circles
of
on the same figure) to the theoretical ones we conclude that the extent
segregation present in this blend is beyond that which can be described by
a mean field theory such as equation 3.
0
Beoo3: The
form factor
the blend using
equation 3; solid line
for x=O, dashed line for
IFigure
Itheoretical
Tof
04
. The
x=.0
experimental result of
points of figure 2
are also shown.
400 t
"
.data
....
o... ...
0.00
...
&.............
0.06
0.04
0.02
q (V)
For a partially segregated blend, the form factor of the cluster can
be approximated by a Debye type correlation function shown in equation 4
where C denotes the correlation length of the composition fluctuation.
7(r) = exp
(-i)
(4)
The corresponding form factor in Fourier space is given as
= v
(.)(6c)2
(+f
q2)
(5)
The
where 6c stands for the amplitude of the composition fluctuation.
value of 6c is unity for a completely segregated blend and zero for an
The term 0 denotes the volume fraction occupied by the
ideal solution.
rigid rod rich phase, hence 1-0 is for the flexible chain rich phase.
The best fit between the experimental results and equation 5 is
The upward curvature of the
given in figure 4 via a Zimm type plot.
explains why equation 2 fails to
experimental results in figure 4
accommodate the experimental result. The best fit value of f is i5A and
By assuming the value of 0 be 0.5, an ideal
the prefactor is 0.0015.
value for the case of a matched molecular weight between these two
polyquinolines, the value of bc can be calculated according to the best
This bc value
fit value of the prefactor and gives a result of 0.4.
suggests that the local compositional fluctuation of the rigid rod or the
For a completely
flexible chain is ;20% from the mean value of 0.5.
segregated blend the values of 0 and 6c are 0.5 and 1.0 respectively.
1A
193
This results in a prefactor of 0.009 which is significantly greater than
the best fit value of 0.0015.
It is also noteworthy that the value of
is much less than 890A the R, of the rigid rod molecule. One possible
e:.planatlon is that the rigid rod rich phase has the shape of a thin
cylinder, and the value of
reflects the average lateral dimension of
this cylindrical object.
Z0.04
Z 0Figure
4: The data
z
points for the observed
form factor of the blend
presented via a Zimn
type plot. The best fit
result using equation 5
given as the solid
line.
W
X
> 0.02
zis
0 .0 0
I
-
o0.0
0.04
0.08
q**2 (10 - 2 'r2)
II
0.12
In figure 4 some deviation of equation 5 from the experimental data
can be observed in the high q region, i.e. in the region of q >>1.0 .
This is expected since equation 5 is merely an approximation of the
scattering intensity for the low q region. In the high q region the
1
intensity should decrease as qor q-2 depending the local chain
conformation in this partially segregated blend.
CONCLUSION
The high molecular weight polyquinoline blend studied in this work
is partially segregated with a correlation length of 115& which is much
less than the R. of the rigid rod molecules.
Results for the block
copolymer indicate that the extrusion process enhances the mixing of this
material.
ACKNOWLEDGEMENT
The research was supported by a contract from DARPA and AFOSR.
REFERENCES
III
12)
[3]
[4)
J. K. Stille, Macromolecules 14_:870 (1981).
J. K. Stille, A.
Parker, J. Tsang, G. C. Berry, M. Peacherstone, D. R. Uhlmann, S.
Subramoney, V. L. Wu, Contemporary Topics in Polymer Science,
,
(1989).
W. L. Wu, J. K. Stille, J. Tsang, and A, Parker, Proceedings of MRS
meeting, Symposium J. (1988).
WL.Wu, Polymer 21: 1907 (1982).
L. Leibler, Macromolecules U: 1602 (1980).
PART IV
BlendslIPN's
_W
197
MISCIBIUTY IN BLENDS OF POLYBENZlMIDAZOLE AND FLUORINE CONTAINING
POLYIMIDES
Hiroaki Yamaoka', Norman E. Aubrey, William J. MacKnight* " and Frank E. Karasz*"
"Mitsubishi Monsanto Chemical Company, Yokkaichi Research
& Development Dept.
1 Toho-Cho Yokkaichi-City, Mie, JAPAN
.. Polymer Science & Engineering Dept., University of Massachusetts, Amherst, MA
01003
ABSTRACT
Blends of polybenzimidazole (PBI) with either of two fluorine-containing polyimides
were prepared by casting from solution and by precipitation. Dynamic mechanical
thermal analysis (DMTA) and differential scanning calorimetry (DSC) were used to study
miscibility in the two blend systems.
The blends of PBI with the first polyimide, the polysulfonimide PI-1, consisted of a
single phase when the blends contained less than 30 wt% PI-1; above 50 wt% PI-1,
phase separation occurred even at room temperature. The PBI blends containing the
second polyimide, PI-2, were immiscible.
INTRODUCTION
Miscible polymer blends consisting of high performance polymers, aromatic
polybenzimidazoles and polyimides, have been studied for several years [1-41.
Miscibility in these blend systems was confirmed by the presence of single, compositiondependent Tg's tying between those of the constitutent polymers, by well-defined
composition-dependent tan 8 peaks associated with the glass transition and by the
formation of clear films.
In this contribution, the phase behavior in blends of polybenzimidazole (PBI) with two
fluorine containing polyimides, PI-1 and PI-2, is presented.
EXPERIMENTAL
Materials and blend oreparatlon
Poty-2,2'-(m-phenylene)-5,5'-bibenzimidazole (PBI) (Hoechst-Celanese) was
used as the representative polybenzimidazole. The fluorine containing polyimides, PI-I
and Pl-2, with the structures shown below were prepared by NASA-Langley.
CF3
O
II
-
I
jN
0
i
il
CCF
3 ks4O
0
C
N
N
0
PI-1
Mat. Res. Soc. Symp, Proc. Vol. 171.
1990 Materials Research Society
198
0
CF3
0
II
N,
0
N
n i3
o
0
P1-2
Elemental analyses of PI-I and PI-2 were performed by the University of
Massachusetts Analytical Laboratories. mw and Mn were obtained by size exclusion
chromatography (SEC) using dimethylformamide as solvent at 60'C. Four columns of
pore sizes 106, 105, 104 and 500 A were used and were calibrated with poly(methyl
methacrylate) standards. The results of elemental analysis and SEC are listed in Table I.
Table I
Elemental Analysis and SEC results for PI-1 and PI-2
Elemental Anaylses
7lw
M,
w/jtn
Tg (C)
PI-t
Calc. C: 56.71% H: 2.13 N: 4.27 F: 17.37 S: 4.87
Found C: 54.88
H: 2.57 N: 4.08 F; 17.50 S: 4.96
17.2
3.1
x 105 x 105
PI-2
Calc. C: 61.18 H: 2.30 N: 4.61 F: 18.75
Found C: 61.22 H: 2.55 N: 4.65 F: 19.07
67.0
x 105
36.8
x 105
5.5
1.82
276
232
Blends were prepared by mixing 3% (w/v) solutions of PBI and polyimides in N,Ndimethylacetamide (DMAc) in the desired proportions. The polyimides dissolved readily
in DMAc under ambient conditions but even in a pressure vessel at 2250C, PBI left an
insoluble residue which was removed by filtration. Films of PBI/PI-1 and PBI/PI-2
blends were prepared by casting 3% (w/v) solutions on glass plates. The solvent, DMAc,
was evaporated under dry N2 by heating to 80C for 48 hours. The films were dried
further under vacuum with a gradual increase in temperature, from 100 to 2200C in
200C increments. Each temperature was held for two to three days; finally the films
were held at 2200C for 7 days, until thermogravimetric analysis (TGA) showed less than
0.2% weight loss in a heating cycle from 100 to 350C.
Precipitated blends were prepared by adding excess methanol to solutions containing
PBI and the appropriate polyimide. A fine powder was obtained which was washed with
water to remove residual DMAc and vacuum dried in the same way as the films. Residual
solvent was also assessed by TGA. Differential scanning calorimetry (DSC) measurements
were made on samples with residual solvent contents of less than 0.2 wt%; film and
powder samples gave identical results.
Dynamic mechanical analysis and differential scanning calorimetry
Dynamic mechanical analysis (DMTA) experiments were carried out using a
Polymer Laboratories DMTA equipped with a high temperature (5000C) head under dry
199
N2 . Films about 0.2 mm thick were used in the flexural mode, under constant strain, at 1
Hz. The scanning rate was 4°C/min in the temperature range from 150 to 450C.
Differential scanning calorimetry (DSC) experiments were made under N2 using film
and powder samples with a Perkin-Elmer DSC-7 differential scanning calorimeter
controlled by a 7500 PC. The heating rate was 20°C/min.
A Perkin-Elmer thermogravimetric analyzer was used tomeasure the residual solvent
content of film and powder samples. The heating rate was 20°C/min and the experiments
were performed under N2 .
RESULTS AND DISCUSSION
PBI/PI-1 blends
Itisimportant to note that the blend samples were not heated above 2200C prior to
determining Tg's. Thus the phase relationships observed in these systems can be regarded
0.8
04
2)
0.4
o
cc
.)
,,
0.4
"
4)
0.4
0
04-
0
200
300
400
T(°C)
Fig. 1
Dynamic mechanical analysis of PBI/PI-1 Blends
1) PI-1, 2) PBI/PI-1:25/75 wt%, 3) PBI/PI-1: 50/50 wt%
4) PBI/PI-l: 75/25/wt %, 5)PBI
200
as stable to 2200C (or at least unchanging within the time scale of the experiment).
Figure 1 shows single, composition-dependent tan 8 peaks indicating miscibility for
the PBVIPI-1 blend up to a blend composition of about 30 wt% PI-I. The observed
clarity and strength of the cast films Is consistent with these results; the 85/15, 75/25
and 70/30 wt % PBI/PI-1 blends are clear and strong as expected of miscible, one-phase
polymer systems.
When the PBI/PI-1 blend contained more than 50 wt% PI-, two phases are observed
for both film and powder samples. One phase, which appears to be a mixed phase has a Tg
of about 345°C. According to the Fox equation [5] this Tg corresponds roughly to a blend
composition of 50/50 wt% which appears to be the composition around which the limit of
miscibility in the PBI/PI-1 bend system is reached. The second phase, with a Tg of
2750C corresponds to pure PI-1(Fig. 1-(2)). At the blend compositions of 60/40 and
50/50 wt% a single, broad tan 8 peak is observed (Fig. 1(3)). The broadening may be
ascribed to phase separation which occurs as the sample is heated to temperatures above
its Tg and may result in the presence of higher Tg mixtures whose damping behavior is
then measured as the temperature is increased. The PBI/Ultem 1000 system [1,3]
shows the same behavior except that these two polymers are miscible over the entire
composition range.
The films cast from solutions of PBI and PI-1have a physical appearance which is
consistent with the phase behavior in these blends. The films become more opaque anc'
their strength decreases as the PI-1concentration in the blend is increased beyond th,
apparent miscibility limit of 50/50 wt%.
Phase behavior at high temperature
Figure 2 shows DMTA data for the PBI/PI-1 blends after heating to 4500C. A damping
A)
-
Fig. 2
TEMPERATURE
t"C)
Dynamic mechanical analysis of PBI/Pl-lblends; samples heated to 4500C.
A) PBtIPl-1: 75/25 wt%, B) PBI/PI-1: 50/50 wt%, C) PBI/PI-1: 25/75/wt%
201
peak clearly associated with the polysulfonimide component is observed only for the P1-rich samples (Fig. 2C). The damping associated with a PBI-rich phase or pure PBI is
apparently suppressed and shifted to higher temperatures as a result of crosslinking
reactions known to occur in PBI at or above T9 [6,71. To clarify the phase behavior for
PBI-rich blends, DSC experiments were conducted on mixtures with blend compositions
of 85115, 75125 and 70/30 wt% PBI/PI-1 before and after annealing at temperatures
between 300 and 450°C. No indication of phase separation was observed although when
the samples were annealed above 4000C, chemical changes apparently occurred. Thus the
phase boundary for PBIIPI-1 blends at these compositions appears to be above 4000C,
although the exact position of the boundary is obscured by the chemical changes which
take place at high temperatures.
The PBI/PI-2 blends
Figure 3 shows DMTA measurements for the PBI/PI-2 blend system. In contrast
to the results for the PI-1 containing blends, two separate transitions can be observed
over the entire range of blend compositions. The T g's of the two phases are slightly
displaced from those of the pure components which indicates the presence of minor mixed
phases. We were unable to obtain mixtures which clearly showed a single phase even at
the extremes of the composition range. Similar small shifts were observed in the T9 peak
corresponding to P1-2 which can be attributed to small amounts (s 10 wt%) of PBI in
this phase. This mixture appears to become phase separated after heating above 450°C.
04
0
04
-
4
3)
04
T °C)
1) PBl,
Dynamic mechanical analysis of PBI/PI-2 blends
2) PBI/PI.2: 75/25/wt%,3) PBI/Pl-2: 50/50 wt%,
4) PBI/PI-2: 25/75/wt%, 5) PI-2
202
CONCLUSI
PI-1 exhibits partial miscibility in blends with PBI; the approximate phase behavior is
shown schematically in Figure 4.
PBI/Pt-2 blends are essentially immiscible except perhaps at blend compositions of
10 wt% or less of PBI.
450
400
.,0
os
_
350
TI
ephm
300-
2501
0
Fig. 4
50
25
75
100
Wt % PI-1
Schematic phase diagram for PBI/PI-1. The dotted line is the calculated Tg;
the circles are the observed Tg.
ACKNOWLEDGEMvENT
This research was supported by the AFOSR through grant # 88-011 and by Mitsubishi
Monsanto Chemical Company and Mitsubishi Kasei Corporation. We thank Dr. T. L. St. Clair
(NASA-Langley) for providing the polyimides and DR. P. Das (Monsanto) for measuring the
molecular weights of the polyimides.
REFERENCES
1. L. Leung, D. J. Williams, F. E. Karasz and W. J. MacKnight, Polym. Bull. 16,457
(1986).
2. S. Choe, W. J. MacKnight and F. E. Karasz, in Polyimides Materials Chemistry and
Characterization ed. C. Feger, M. M. Khojasteh and J. E. McGrath (Elsevier, Amsterdam
1989).
3. S. Choe, W. J. MacKnight and F. E. Karasz, in Multinhase Macromolecular Systems ed. B.
M.
Culbertson (Plenum, New York, 1989).
4. P. Musto, F. E. Karasz and W. J. MacKnight, Polymer 30, 1012 (1989).
5. T. G. Fox, Bull. Am. Phys. Soc. 1, 123 (1956).
6. H. Vogel and C. S. Marvel, J. Polym. Sd. 50, 511 (1961).
7. J. K. Gilham, Science 495, 1257 (1963).
203
SKALL ANGLE NEUTRON SCATTERING STUDIES OF BLENDS OF PROTONATED) LINEAR
POLYSTYRENE WITH CROSSLZNMED DEUTERATED POLYSTYRENE
ROBERT M. BRIBER AND BARRY J. BAUER
National Institute of Standards and Technology, Polymers Division,
Gaithersburg, MD 20899, USA
ABSTRACT
Small angle neutron scattering (SANS) has been used to study the
scattering function and thermodynamics of blends of linear protonated
polystyrene (PSH) and crosslinked deuterated polystyrene (PSD).
Two series
of samples were synthesized, In both cases the samples were made by
dissolving the linear PSH in deuterated (de) styrene monomer containing a
small amount of divinyl benzene as a crosslinker which was then polymerized
to form the PSD network around the linear PSH chains. The samples were all
made at a concentration of 50/50 by weight PSD/PSH. A special effort was
made to keep the samples single phase so that SANS could be used to study
the thermodynamics of the system and compare with theory. This entailed
working at relatively low crosslink densities (<I mole % crosslink units).
Series I is a set of samples with the same crosslink density varying the
length of the linear chain. Series 2 is a set of samples containing the same
length linear chain varying the crosslink density systematically. By
extrapolating S(q) obtained from SANS to q=O the zero angle scattering,
S(O), was obtained. S(0) is inversely proportional to the second derivative
2
2
of the free energy with respect to composition, a (Af/kT)/a0 . Assuming
additivity of the free energies of mixing and elasticity, the portion of the
zero angle scattering due to elasticity is calculated.
.INTRODUCTION
Small angle neutron scattering has been used in recent years to study
the thermodynamics of phase separation in linear polymer blends with much
success 11-4].
In this work we wish to extend the use of neutron
scattering to study the phase separation transition in systems where one of
the components is crosslinked. The topic of linear polymer chains in
networks has been studied in the past under the topic of semiinterpenetrating polymer networks (semi-IPN) but a careful review of this
area shows that the systems studied are almost exclusively phase separated.
Generally, the phase separation occurs during the polymerization and the
systems exhibit a nonequilibrium, kinetically controlled two phase
morphology 15-7). The emphasis in the work presented here is on single
phase systems and the thermodynamics which control the miscibility of the
mixture.
THEORY
The classical theory for the free energy of a network swollen by a
solvent due to Flory and Rehner [8] (and later revised by Flory 191) can be
extended to a network containing linear chains [10,11].
Assuming additivitv
of the free energies described by Gaussian rubber elasticity and a FloryHuggins theory the free energy per unit volume of the system, Af, can be
written as
&f
3
-
-
kT
2N,
B0
(0,2/310
/
3
-
0) +
-
I4
ln(O/,)
nl4
+ xO(1-0)
+
2N,
(1)
Nb
where 0 is the volume fraction of the network, 0. is the volume fraction
Mat. Res. Soc. Syrp. Proc. Vol. 171. 11990 Muterial$ Research Society
204
where the network is relaxed (the network reference state; usually taken as
the composition where the network was formed), N. is the average number of
monomer units between crosslink points, Nb is the number of monomer units in
the linear chain (b component) and X is the Flory interaction parameter.
The specific volumes of the different monomers have been left out for
simplicity. The constant B has been generally taken as 2/f by Flory [9],
where f is the functionality of the crosslink points in the network, but
others have argued for different values (12-14].
The coexistence curve is
obtained when the chemical potential of the linear b chains inside and
outside the network are equal. The chemical potential of the b chains is
given by ,=(AF/kT)/anb where AF is the total free energy of the system and
n b is the number of moles of the linear b chains [Il].
B'
o.2/134s113
l (1A
2
--
pb=
N.
+
Ne
++
X
Nb
(2)
Nb
Because the swollen network is in equilibrium with a phase of pure b
chains, equation 2 can be used to calculate the coexistence curve directly
(upon setting pb=0) without resorting to solving a set of simultaneous
T
equations involving both j. and Ub as is necessary in polymer blends. For
2
2
the spinodal line extra care must be taken in calculating 8 (Af/kT)/aO
in
order to correctly account for the effect of fluctuations in composition on
the already swollen network (15,161.
If this is done the equation for the
spinodal (and the zero angle scattering) is given as
8
2
(Af/kT)
80 2
where k n
B
=
2NO
o.2/3
+
N 0
5/ 3
1
+
Nb ('-O)
-
2
x
kn
S(0)
(4)
is a contrast constant that depends on the type of radiation used
for the scattering experiment.
This equation also agrees with the criteria
for when an elastic body becomes unstable to fluctuations given by
-
s(o) '
(5)
X + (4/3)G
where x is the osmotic bulk modulus and G is the shear modulus 115-17).
Figure 1 shows a phase diagram calculated for network with N,=500. Nb=50,
8=1/2 and 0,=0.5.
The ordinate is in arbitrary temperature units based on
the observation that X generally has the form A + B/T [1,2,18,191.
The
addition of an elastic component to the free energy causes the phase diagram
to be asymmetric.
In addition, with the form of the free energy given in
equation I there is no critical point with the phase transition being first
order at all compositions.
205
N,=500 Nb=50
single phase
Figure 1:
Phase diagram
calculated for linear
chains in a network.
0
02
0.0
0.2
two phase
0.4
0.6
0.8
1.0
The scattering from the blends studied in this work is expected to
follow classical Ornstein-Zernike form with the total scattering given by
s(o)
s(q) =
(6)
1 +
( q)2
where E is the correlation length of the concentration fluctuations in the
system and q is the scattering vector (q = (4*/A)sin$).
If the data
follows equation 6 then a plot of S(q)-' versus q1 should be linear with
the correlation length being equal to the square root of the ratio of the
slope to the intercept.
EXPERIMENTAL
Linear protonated polystyrene (PSH) was purchased from the Pressure
Chemical company while the deuterated (d.) styrene monomer and divinyl
benzene was purchased from Aldrich 120].
The styrene was dried over
calcium hydride and distilled. Azobis(isobutyronitrile) (AIBN) was used as
the initiator.
Samples were prepared by dissolving the PSH in the
deuterated styrene monomer/divinyl benzene mixture.
When the PSH was
completely dissolved a few drops of a 10% AIBN solution in toluene was
added to give 0.1 wt % initiator.
The mixture was then sealed in a SANS
cell and placed in a oven at 70"C overnight. The temperature was then
increased to 130*C for an additional 12 hours. The samples were all made at
a concentration of 50/50 by weight PSD/PSH.
Series 1 is a set of samples
with the same network (N,=387) varying the length of the linear chain
91
(Nb=
, 308, 422, 981, 1711 and 3413). Series 2 is a set of samples
containing the same length linear chain (Nb= 4 2 2 ) varying the crosslink
density systematically (N.-, 960, 475, 345, 260, 158).
Table I gives the
details of the two sets of samples. The crosslink densities were calculated
based from the amounts of monomers charged to the reaction and a reported
divinyl benzene activity of 57%.
The neutron scattering was done at the NIST SANS facility.
The
wavelength of the incident neutron beam was 9A with a AA/X of 25% as
determined by a rotating velocity selector. The scattering was performed
206
above T, at 150°C. The data were collected by using a two-dimensional x-y
detector and were corrected for scattering from the empty cell, incoherent
background, and sample thickness and transmission. The scattering was
placed on an absolute scale using a calibrated secondary standard. Data
was then circularly averaged to obtain the S(q) versus q plots.
RESULTS AND DISCUSSION
The two series of samples provide a probe of different parts of the
second derivative of the free energy as given in equation 4. If the
assumption of additivity of free energies is valid, i.e.
Afro
Af±A,
=
t
(7)
Afo..
+
then series 1, where the crosslink density is held constant, while the
length of the linear chain is varied, allows examination of AEj.. Series
2 on the other hand, keeps the length of the linear chain constant, while
N. varies, thereby probing Afoj..
Figure 2 presents the SANS data for the series 1 samples in both S(q)
2
versus q and S(q) ' versus q forms. As the length of the linear chain
2
increases the scattering intensity also increases. In the S(q)Y' versus q
plot the data forms a series of straight lines which move roughly parallel
with progressively smaller intercepts as Nb increases. The sample with
3
3
Nb= 41 was phase separated as indicated by the negative intercept in the
2
S(q) 1 versus q plot. In analogy with linear polymer blends the slope of
these lines is proportional to the square of the statistical segment length
,p2 [1,18].
The behavior of the samples in series 1 is qualitatively similar
to that of a linear polymer blend as the temperature is changed.
K.-S6
a Mb-N
~
O
l
200-m
3000q0
OL~
.
400
~
Nb-3414
~ ~~~Nd1
O.1
1
0,00
o.4[
0.00
0.0
-
.1
.3)004
.5
.7
.
0.-
-
q
Figure 2:
A
SANS data for the series I samples. S(q) versus q and S(q)
versus
"I
q2.
Figure 3 presents the SANS data for the series 2 samples. The
scattered Intensity Increases as N. decreases. The samile with the highest
crosslink density (N.-158) was phase separated. The ph.%se separation of
this sample can be attributed to the stretching of the network necessary to
accommodate the linear chains at this crosslink density and its unfavorable
contribution to the free energy of the system. The small x between PSH and
207
PSD is not sufficient to explain the phase separation. Unlike the data for
2
series I the S(q) versus q plots show a systematic change in slope with
increasing crosslink density.
0 N.-960
0
* M.-345
•
0.00
• Ne-M
0.06
0.04
0.02
0.
0.04
q-' .1o*
0.0
q A
Figure 3:
0.12
0.00
A 2
SANS data for the series 2 samples. S(q) versus q and S(q)
2
versus q .
"
Rearranging equation 4 yields
[
k1
=
5(0)
-
Nb(1-#)
+
]
#.2/3
2N,#
+
- 2X
N.#
"s
(8)
51 3
indicating that a plot of k./S(0)
versus I/Nb(l-#) should give a straight
line with a slope of 1 if the Flory-Huggins combinatorial entropy
adequately accounts for the change in scattering brought about by the
increase in the length of the Nb chains. Equation 8 also assumes that the x
parameter between the PSH chains and the PSD network is independent of the
crosslink density. While recent studies h,
shown that X can be a
function N. the effect should be small at tw.e crosslink densities studied in
"
1
this work [21].
Figure 4 is a plot of S(0)
versus N* .
The straight
line is a linear least squares fit to the data and has a slope of 1.04±0.05.
The x intercept yields the value of Nb where S(0)"s=0 and the favorable free
energy of mixing is Just balanced by the elastic energy of the network and
phase separation occurs.
The value of Nb for phase separation from figure 4
is 12700.
This is larger than what is found for the sample which phase
"
separated (Nb=3413) but the extrapolation in figure 4 yields Nb
which will
"
give relatively large errors in Nb.
Assuming a value of X=l.5xl0
for
PSH/PSD [221 the elastic contribution to the second derivative of the free
energy can be calculated from the intercept of the line ii figure 4.
The
2
2
3
value of a (Af/kT)/la oL.. is 1.54 x 10
.
208
24 Nc=387
20
S16
S1 2
-
8
Figure 4: k0 /S(O) '
I/Nb(l-O) for
the series 1 samples.
/versus
0
4.
0
0
4
8
12
16
I/Nb(l-0)
X10
20
24
3
For the series 2 samples equation 4 can be arranged
1[B
k.
S(O)
=-
+
2#
N,
1
0%2/3]
-
+
2X
o5/3
(9)
Nb(l-)
-
and a plot of k,/S(O) l versus 1/N, should give a straight line. Figure 5
shows such a plot. The line is a linear least squares fit to the data and
the point where it crosses the x axis is the value of N, where phase
separation occurs S(0)'o=0. The value of N, obtained from the
extrapolation is 1,60. This compares favorably with the value of N,=158
where phase separation was observed to have occurred.
6
04
N
a2.
.X
Figure 5: k0 /S(O)- 2
versus N, ' for the
series 2 samples.
0
1
2
3
I/Ne
4
x10
°
3
5
6
2
7
If the slope of the S(q) 1 versus q- can be taken as proportional to
the square of the statistical segment length, 12, as in linear polymer
blends, then the systematic change in slope with increasing crosslink
density for the series 2 samples indicates an corresponding increase in 1.
The slope of the lines in figure i is proportional to an average of the
statistical segment lengths for the two components in the mixture. The
209
implication is that either the PSH or the PSD chains (or both) are
stretching as N, decreases. The data cannot distinguish between the two
possibilities, but one might speculate that the network chains are forced to
stretch in order to accommodate the linear PSH chains. As Nc decreases this
stretching increases until the free energy penalty caused by the deformation
drives the system to phase separate. If one does not worry about the
magnitude of 1Z but examines instead the ratio of the slope of the S(q) '
2
versus q plot at a given value of N c to the slope for the uncrosslinked
blend (N.-) an estimate of the amount of chain deformation, A, can be made.
The values of A obtained for the series 2 samples in this manner are given
in table I.
CONCLUSIONS
SANS has been used to study the thermodynamics of linear PSH chains
i.n a crosslinked matrix of PSD. It was found that the addition of an
elastic component to the free energy due the crosslinked PSD caused the
system to phase separate at relatively low crosslink densities (<l mole C
crosslinking agent). This indicates the importance of accounting for the
presence of a crosslinks on the compatibility of polymer blends, even in
miscible systems. Classical Flory-Huggins free energy of mixing combined
with Gaussian rubber elasticity has been used to calculate the zero angle
scattering, S(O) and compare with experiments. The combinatorial entropy
of mixing term adequately describes the data for the samples where N. is
held fixed and the "ength of the linear chain, Nb, is varied. In the
samples where Nb is fixed and N. varied, a systematic increase in the
average statistical segment length is observed, indicating that one or both
of the components is being stretched as N. decreases.
TABLE I
Series I
Nc=387
Nb
S(0)
A" 1
cm1
k/S(0) x10'
91
308
422
981
1711
19.2
26.3
215
54.0
51.1
76.2
95.6
62.6
43.0
18.7
93.6
220
1494
263
2.7
****
34l31
4 3
ISample with Nb=3 1 was phase separated.
SERIES 2
42
Nb2
N0
S(0)cm'
79.0
65.9
113
127
227
***
15 8
'Slopes are from
2x 101
3 x 103
'The sample with
960
475
345
260
-
A A
slope'
A
k./S(0)
51.9
60.0
66.3
93.6
111
2
3.37
--5.2
3.71
1.05
4.3
3.82
1.06
3.6
4.78
1.19
3.2
5.41
1.27
1.8
***
***
***
"2
2
d
the S(q)
versus q plot. x le
N.=158 was phase separated.
2
8'(&f/kT)/a0 elas
4.4
-61
-101
-243
3
210
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Shibayama, M.; et al.; Macromolecules, U, 2179 (1985)
Jelenic, J.; et al.; Makromol. Chem., 185, 129 (1984)
Briber, R.M. and Bauer, B.J.; Macromolecules, 21, 3296 (1988)
Bates, F.S.; et al.; Macromolecules, 12, 1938 (1986)
Sperling, L.H.; Interpenetratin2 Polymer Networks and Related
Materials, Plenum, New York, 1981
Coleman, M.M.; et al.; Macromolecules, 20, 226 (1987)
Frisch, H.L.; et al.; Macromolecules, 13, 1016 (1980)
Flory, P.J. and Rehner, J.; J. Chem. Phys.; 11, 455 (1943)
Flory, P.J.; Princioles of Polymer Chemistry, Cornell University
Press, Ithaca, New York, 1953
Binder, K. and Frisch, H.L.; J. Chem. Phys., 81, 2126 (1984)
Bauer. B.J.; Briber, R.M.; Han, C.C.; Macromolecules, 22, 940 (1989)
James, H.; Guth, E.; J. Chem. Phys., 14, 669 (1947)
Hermans, J.J.; J. Polym. Sci., 52, 197 (1962)
Kuhn, W.; J. Polym. Sci., 1, 183 (1946)
Onuki, A.; Phys. Rev. A., 38, 2192 (1988)
Olvera de Is Cruz, M.; Briber, R.M.; to be published
Landau, L.D. and Lifshitz, E.M.; Theory of Elasticity, Pergamom Press
1986
Han, C.C.; et al.; Polymer, 29, 2002 (1988)
Bates, F.S.; Macromolecules, 8, 525 (1985)
Certain equipment, instruments or materials are indentified in this
paper in order to adequately specify the experimental details. Such
identification does not imply recommendation by the National Institute
of Standards and Technology nor does it imply the materials are
necessarily the best available for the purpose.
McKenna, G.B.; et al.; Polymer Communications, 22, 272 (1988)
Bates, F.S.; et al.; Phys. Rev. Lett., 55, 2425 (1985)
211
I
"DYNAMICS OF PHASE SEGREGATION IN POLY-P- PHENYLENE
TEREPHTHALAMIDE AND AMORPHOUS NYLON MOLECULAR COMPOSITES"
THEIN KYU, JAN CHANG YANG AND TSUEY ING CHEN
Institute of Polymer Engineering, University of Akron, Akron,
Ohio 44325
ABSTRACT
Time-resolved light scattering has been employed to elucidate the
dynamics of phase segregation of poly-p-phenylene terephthalamide (PPTA)/
amorphous nylon (AN) molecular composites. Miscible PPTA/AN blends can be
prepared from sulfuric acid solution by rapidly coagulating the sclution in
distilled water. The composites, however, undergo phase segregation upon
thermal treatment and exhibit a miscibility window reminiscent of a lower
critical solution temperature (LCST). Several temperature-jump experiments
were undertaken from ambient to a two-phase temperature region of 240, 250
and 2600C. Time-evolution of scattering profiles are analyzed inaccordance
with non-linear and dynamical scaling t eories.
INTRDDI]TION
4
The field of molecular composites has gained considerable interest for
its potential in structural applications, such as high modulus and high
strength materials (1-8]. Conceptually, molecular composites are mixtures
of two dissimilar polymers with vastly different configurations, in which
the rigid component isdispersed in flexible matrix so that reinforcement
takes place at a microscopic level. The maximum performance of the
materials can be expected if reinforcement were to occur at a molecular
level. It is difficult to meet such expectation due to the low entropy of
mixing and high orientability of rigid component. The compatibility between
vastly dissimilar polymers is always a central issue in multicomponent
systems. The immense difference in the molecular topology, i.e. one being a
rigid-rod and the other component being a flexible coil, makes the mixtures
thermodynamically unstable [9]. However, such thermodynamic tendency for
phase decomposition may be overcome by rapidly coagulating from a
homogeneous ternary solution to give miscible systems [0.
These homogeneous mixtures processed from solution are generally in the
form of thin films, fibers or precipitates, thus are of limited use. For
structural applications, melt processing is required to consolidate intoa
desired shape. This consolidation step involves thermal cycle and flow
which occasionally lead to thermally induced phase separation [10-13].
Hence, the understanding of phase behavior and its kinetics of phase
decomposition during thermal consolidation is indispensable inorder to gain
control of the structure for improved materials properties [11].
The kinetics of thermally induced phase decomposition in
poly-p-phenylene benzobisoxazole (PBZT)/nylon 66 was first studied jointly
by the Air Force Materials Laboratory and us using time-resolved light
scattering [11]. Phase segregation occurs above the onset of the melting
temperature of nylon 66 and is dominated by the late stages of spinodal
decomposition. The kinetic behavior isnon-linear in character and follows
power-law kinetics with exponents of -1/3 and 1, inclose agreement with the
prediction of Binder and Stauffer [141. In the scaling analysis, the
structure function exhibits universality with time, thereby confirming
self-similarity of the system.
As a complementary study, we have examined the phase equilibrium of
poly-p-phenylene terephthalate (PPTA)/ amorphous nylon (AN) [12,13]. The
cross-hydrogen bonding was found to occur between amide groups of PPTA and
mat e.
Soc. Symp. Proc. VoL.171.
sm.
*
l0g9
Materials Rese.rch Society
212
AN. iowever, it is not sufficient to prevent thermally induced phase
separation. The mixture showed a phase behavior reminiscent of an LCST. In
the present study, we focus our attention on the kinetic aspects of phase
decomposition of PPTA/AN molecular composites. The advantage of the present
system is that amorphous nylon shows no crystallinity, thereby simplifying
the interpretation of dynamical results. Although PPTA is strictly not a
rigid-rod, it has a tendency to self-associate as a result of hydrogen
bonding and align with respect to each other due to its stiff extended
molecular structure. The presence of such a complex PPTA structure may
affect the dynamical behavior of phase decomposition. Temperature (T)-jump
experiments were undertaken from ambient to a two-phase temperature region.
Time-evolution of scattering curveq are then analyzed according to
non-linear theories [14-17] and dynamical scaling laws [18-21].
EXPERIMENTAL
Materials
Poly(p-phenylene terephthalamide) (PPTA, Mn , 20,000) was supplied by
Du Pont Co. in the form of quarter-inch fiber (Kevlar 29). Amorphous nylon
used in this study was Zytel 330 kindly supplied by Du Pont Co. Zytel 330
is a copolymer of (30/70) iso-/terephthalic acid and hexamethylene diamine
with a molecular weight Mn - 14,000 and Mw - 50,000. The solvent used was
967. sulfuric acid from Fisher Co. The preparation method of molecular
composites is thoroughly described in a previous paper [13].
Methods
Real time light scattering experiments were carried out using our
static light scattering set-up described elsewhere [221. The system
consists of a 2 mW le-Ne laser with a wavelength of 632.8 nm. The scattered
intensity was monitored using two-dimensional Vidicon detector (Model 1254
I, EG k G) inconjunction with a detector controller (Model 1216) and
Optical Multichannel Analyzer (OMA III, Model 1460). Data analysis, such as
background subtraction, smoothing, angle calibration, rescaling, etc., was
undertaken on an off-line computer (IBM-PC). Temperature (T)- jump
experiments were undertaken from ambient to 240, 250, and 2600C.
RESULTS AND DISCUSSION
In a previous paper [13], we have shown that PPTA/AN mixtures revealed
an L(ST with a minimum around 50/50 and at 2250C. When the 50/50 PPTA/AN
was annealed at 2400C for 10 min, a scattering halo developed in the V,
vertical polarizer with vertical analyzer) configuration. Interconnected
(1mains
were also discerned under a microscopic investigation, suggesting
tie possibility of spinodal decomposition (SO). If an alternative mechanism
of nucleat ion- growth (NG) were to occur, the scattered intensity should
oiiollto, icallv decrease from q = 0 without revealing a maximum. This was not
th, case here. We carried out several T-jump experiments by rapidly
rvansf'viring the specimen from ambient to two-phase temperatures.
Figure 1
shows a typical time-evolution of scattering curves, following a T-jump to
2501'1.
lh scattering maxima first occurs at relatively low scattering
s+avvnmmmbvrs q. suggestive of large periodic fluctuations. Here, q is
,i,'find as (Ir/A) sin #/2. where A and 0 are the wavelength of light and
.
eral
,imi :ng le neasured in the medium, respectively. TIme scattering peak
ll
.I' llfflv'diiItely to lower scattering angles with elapsed time associated
ih h1lbisv 'lOWili. There is mto
period at which the scattering peak is
213
So
r1l.(1)
no~
600l
4
Fig. 1. Time-evolution of scattering
curves for 50/50 PPTA/AN, following
a T-jump from ambient to 2500C.
400
300
200
100
0
0.0 0.2 0.4 0.6 0.8 1.0
q (1/0m)
1.2
invariant, indicating the lack of a linear SD regime [23].
The non-linear growth may be best explained in terms of the dynamical
scaling behavior in the late stages of SD [14-21]. This concept is implicit
in the cluster diffusion-reaction theory of Binder and Stauffer [141 who
proposed that the structure factor S(q,t) for an isotropic system with
dimensionality, d, obeys the scaling law
S(q,t) = (qm(t)
-d
s[q/qm(t)]
(1)
where s[q/qm(t)] is a universal scaled function. The characteristic
wavenumber qm(t), which is inversely proportional to the average cluster
size R(t), has a simple power law form,
qm(t)
"
t0
(2)
where subscript m stands for peak maxima. The kinetic exponent 9 has been
predicted to have various values. The classical evaporation-condensation
model of Lifshift and Slyozov [24] predicted that a droplet size grows
according to the power law with exponent 9 = 1/3, even though molecular
details were not considered. Binder and Stauffer [141 postulated V =
1/(d+3) or 1/(d+2) in the intermediate regime (depending on temperature
quenched depth into unstable region) and a value of 1/3 at late stages of
decomposition. Furukawa [18] also predicted the same formulae, V = 1/(d+3)
and 1/(d+2), for surface and bulk mobilities, respectively, in the
asymptotic behavior of the kinetic equation. On the basis of non-linear
statistical considerations, Langer, Bar-on and Miller [15] obtained V
0.21. The change of exponent from 0.2 to 0.28 was revealed in the
simulation work of Marro et al. [16]. According to Siggia [17],
hydrodynamic flow plays a crucial role in the coalescence of the growing
domains in which the materials have to be squeezed out through
interconnected channels. fieestimated a value of 1/3 in the intermediate
stage and I at late stages of SD. It appears that different mechanisms
operate at different stages of decomposition, thus the power law is not
expected to hold over the entire range of SD.
In Figure 2, are shown the plots of loq qm(t) versus log t for various
T-jumps at 240, 250, and 260oC. The slopes were close to -1, in ood
agreement with the prediction of Siggia[17] for the coarsening driven by
surface tension. As the T-jump temperatures were considerably high, the
214
100 31s.(U)
N£
A.
*
AA
£wavenumber
1time
IDA
10
101
40
f
tO4
10
,1 9
::"
:
30
10
Pi*
so.
10
Fig. 3. Plots of I.qm 3 versus q/qm
for 50/50 PPTA/AN at 2500C.
INto
0
0.0
103
Fig. 2. Log-log plots of maximum
q. versus phase separation
t for 50/50 PPTA/AN at various
temperatures.
0.5
r
1.0
1.5
30
.O
()
I COPintensity
Fig. 4. Log-loK plots of scattered
against scattering wavenumber
for 50/50 PPTA/AN at 2500C.
t0 L
0.3
1.0
4 iMm
1.5
215
liquid-liquid phase decomposition occurs very rapidly, thereby missing the
early and intermediate stages of SD. At a later time, the slope chan~es to
a smaller value due to the pinning effect. Wide-angle x-ray diffraction
studies reveal the presence of small amount of disordered PPTA crystals in
the blends. This mesophase structure may be anciwred during rapid
coagulation. If the amount of PPTA crystal were large, the diffusion
process during phase growth may be reminiscent of a liquid-solid type. For
such a case, Furukawa (21] predicted 9 = 1/2, in which solid-liquid
interphase mobility is dominant. Our experiment was not long enough in
duration to discriminate the exact pinning mechanism.
In order to test with the scaling theory [181 we calculated the scaled
structure function s(q/qm) for a three-dimensional case as follow;
s(q/q.) = {qm(t)}
3
S(q,t)
(3)
where the normalized structure factor S(q,t) isdefined as
S(q,t) - I(q,t)/J I(q.t)qidq
(4)
The denominator or the invariant function will be constant in the late
stages of SD where the mean-square fluctuations of refractive indices
reaches a limiting value.
In such cases, normalization is not necessary.
Figure 3 shows the plots of I(q,t)qmi(t) against q/qm for various times.
The good superposition of the scattering curves sugests that the structure
function isuniversal with time. The self-similarity behavior appears the
same for all T-jumps studied here.
The shape of the scaled function is important, as it is associated with
the correlation between clusters and their shapes. Furukawa [191 proposed
that the shape of structure function can be scaled with x = qR, i.e.,
s(x) - x2/[7/2 + x- 7 ]
(5)
where 7 isrelated to dimensionality d as follows;
7=
d+1 for off-critical mixtures
2d for a critical mixture
For
three-dimensional
growth, i.e., d=3, the asymptotic form of s(x) is
x2 at q<qm and x- 4 at q>qm for off-critical mixtures. In the case of
6
critical mixtures, an x- dependence has been suggested at q>qm[21]. As can
be seen in Figure 4, the slope of -3 was obtained at q>qm , which is lower
than that predicted for three-dimensional growth. We have to admit that our
results cover only limited q range. Moreover, the initial fluctuation size
is very large, i.e. q varies from about I to 0.3 pm' which corresponds to a
periodic domain size of approximately 6 to 15 pm. Since the film thickness
is about 10 am, it is conceivable, although by no means conclusive, that the
growth may be two-dimensional rather than three-dimensional, which would
give a slope of -3 for off-critical mixtures. The slope at q<qm is close to
the predicted value of 2. The effect of PPTA crystallization on the
scattering profiles during phase growth should not be ruled out in
explaining the behavior of asymptotic scaled structure function.
CONCLUSIONS
We have demonstrated that phase decomposition in PPTA AN molecular
composites appears to be dominated by late stages of SO. The surface
mobility of PPTA mesophase appears to play a crucial role in the kinetics of
216
phase growth. The structure function exhibits universality with time,
suggesting self-similarity. The asymptotic behavior of the scaled structure
function suggests that the growth process may be two-dimensional because of
thin specimens.
Acknowledgement: Support of this work by the U.S. Army Research Office
Grant number DAAC29-85-KO219 is gratefully acknowledged.
REFERENCES
I. ilelminiak, T.E., Benner, C.L. and Arnold, F.E., Polym. Prep. ACS 16(2),
659 (1975).
2. Hwang, W.F., Wiff, D.R., Benner, C.L. anid
Helminiak, T.E., J. Macromol.
Sci. Phys. B22 231 (1983).rcor
.
oym n.Si
3. Hlwang, W.F., Wiff, D.R., VrcorCPlm
n.Si
13 88
ibid. 23, 784,(1983).
(1983);
4. Wiff, D.R., Timms, S., Helminiak, T.E. and Hwang, W.F., Polym. Eng.
Sci. 27, 424 (1987).
5. Takayanagi, N. and Kajiyama, T., U.K. Patent No. 2,008,598 (1978).
6. Takayanagi, M., Ogata, T., Morikawa, M. and Kai, T., J. Nacromol. Sci.
Phys. B-1,591 (198).
7.* Takayanagi, M., Pure Appi. Chem. 55, 819 (1983).
8. Yamada, K., Uchida, M. and Takayanagi, N., J. Appl. Polym. Sci L2,
5231 (1986).
9.' Flory, P.J., Macromolecules L1, 1138 (1978).
10. Chuah, 11.1., Kyu, T. and Helminiak, T .E., Polymer -2a 2129 (1987).
11. Chuah, H.H., Kyu, T. and Helminiak, T.E., Polymer NQ, 1591 (1989).
12. Kyu, T.,,Chen, T.l., Park. H.S. and White, J.L., J. Appl. Polym. Sci.
37, 201 (1989).
13. Chen, T.I* and Kyu, T., Polym. Commun., submitted.
14. Binder, K. and Stauffer, D., Phys. Rev. Lett. 33, 1006 (1974).
15. Langer, J.S., Bar-on, N. and Nil ler, H.D., Phys. Rev. A, LjI,1417
(1975). J1.,
16. Narro,
Lebowitz, J.L. and Kalos, N.H., Phys. Rev. Lett. 43, 282
(1979).
17 * Siggia, E.D., Phys. Rev. A, 20, 595 (1979).
18. Furukawa, H., Phys. Rev. Lett. 43, 136 (1979); Phys. Rev. A, 23, 1535
1.Furukawa, H1.,
Physica A, 123, 497 (1984)
20. Binder, K., Phys. Rev. B, j, 4425 (1977).
21. Furukawa,' H., . 1.Ap.Cryst. 21, 805 (1988).
22. Kyu, T. and Saldanha, J.N., J. Polym. Sci. Polym. Lett. Ed. 26, 33
23
n JW and Hilliard, J.E., J1.
Chem. Phys. 22, 258 (1958).
2.Lifshitz, I.N. and Slyozov, V.V., J. Phys. Chem. Solids 12.,
35 (1961).
2.Hashimoto, T., Itakura, N. and Hasegawa, II.,
J. Chem. Phys. 85, 6118
(1986).
217
OF SAN/PMK4A BLENDS
FACTORS INFLUENCING PROPERTIES
R. SUBRAMANIAN, Y. S. HUANG, 3. F. ROACH AND D. R. MIFF
GenCorp Research, 2990 Gllchrist Road, Akron, Ohio 44305
ABSTRACT
The versatility of polymeric blends is reflected in the range of
usable end properties that can be achieved through alterations in the
composition and/or effective control of the morphologies of the mixtures.
Solvent cast and melt mixed blends of SAN and PMHA have been studied to
understand the influence of PMMA tacticity and the acrylonitrile content
of the SAN copolymer on their phase behavior. The results show a shifting
of the cloud point curves and the miscibility windows for the blends of
SAN with different stereo-regular PMMAs. This is interpreted as being due
to specific interactions between the acrylonitrile and methacrylate
groups. The 'goodness' of mixing in the melt mixed blends were determined
by FTIR ATR spectroscopy. The effects of the processing conditions on the
mixing characteristics and subsequent improvement in mechanical properties
are discussed. The limits of mixing that can be achieved using standard
mixing procedures and their role in affecting the end use properties such
as the flexural strength will also be discussed.
INTRODUCTION
Miscibility In multicomponent polymeric systems is not always
desirable; but gross phase separation invariably leads to inferior
As in block copolymers, controlled chemical structure and
properties.
microphase heterogeneity In polyblends often lead to superior mechanical
Such intricate phase
properties and Improved processability [1-71.
morphology can be obtained upon phase demixing (separation) of a
homogeneous blend via a spinodal mechanism. ihich would result in phase
domains approaching molecular dimensions (- 1OA*) and the phase separated
blend would behave like a self-assembled composite (8].
It has been known for quite sometime now that polymethylmethacrylate
(PMMA) forms miscible blends with the copolymer, styrene acrylonitrile
This
(SAN) over a limited range of acrylonitrile (AN) content (9].
'miscibility window' has been explained by Paul and Barlow [10] and
others
[11,12] as resulting from a dilution of the unfavorable
leading to a net negative
Interactions among the various segments
contribution to the free energy of mixing, AGM.
The SAN/PMMA system has been widely studied for phase behaviour by
many researchers [9-12]. However, the investigations have been limited to
blending SAN (with varying amounts of AN) with commercial, heterotactic
Pt#A (h-PMMA) and other higher order methacrylates (13]. The influence of
PMNA tacticity on the phase behavior of this system has not been
Investigated.
Isotactic, syndiotactic and heterotactic forms of PMMA
exhibit large differences in T and other physical properties (14,15].
Thus, It is reasonable to epect these different PMWAs to behave
Based on this consideration, a
differently when blended with SAN.
systematic study of the effects of PMMA tacticity on phase behavior,
processing and mechanical properties of the blends was undertaken.
Mat. Re. Soc. Symp. Proc. Vol. 171. I1F
I
Materials
Research Society
218
EXPERIMENTAL
a. Materials:
The polymers used In this study are described in Table 1. The SAN
and h-PMMA materials were obtained from Dow Chemical Company (Tyril') and
Rohm and Haas Company (Plexiglass') respectively. The isotactic PMIA
(i-PNMA) was produced in-house by anionic techniques and the syndiotactic
material (s-PMMA) was an experimental material obtained from Rohm and Haas.
Table I
POLYMIEtS
SANI
M INTHIS STUDY
% AN TJOC) On-xilY mi~ % Tacticity*
30
111-0
50.6
3-5
25
112.0
44
3-4
-
SAN3
20
108.0
49.3
3.6
-
SAW
10.4
103-8
47.4
3.5
-
h-PMMA
-
110.6
25.2
1-8 14
41 45
i-PMMA
-
59-0
41.0
-6-0 83 13 4I
s-PMMA
-
131-5
110.0
1.3
SAN2
Tnad Anaift by H NMR
3
26
71
0'GPCAJWY" UWg PMMA SWan<rds
b. Samle Preparation for Liaht Scattering:
The samples for determining the phase boundaries in the SAN/PMMA
systems were prepared by solution casting from 1,2 dichloroethane (OCE)
using procedures described in the literature [133.
Optical cloud points
were determined using a standard laser light scattering apparatus. The
morphologies in this system were identified using phase contrast optical
microscopy.
c. Melt Mixing:
Laboratory scale mixing of the two polymers was performed to assess
the effect of mixing on the end use properties of the binary blend.
Blends of SAN (30% AN) with h-PMA were prepared by four different
techniques. To facilitate blending, the materials were first ground to a
uniform particle size in a Condux grinder and dried overnight in a
convection air oven at 100C. Appropriate amounts of the samples were
then melt mixed using the following:
1. Haake Single Screw Extruder
2. An In-house built Fiber/Film Processor with Static Mixer.
3. Haake-Buhler Rheocord System 40
4. Physical mixing using a blender, followed by compression molding.
d. Mechanical Prooertes:
Tensile stress-strain and 3-point flexural measurements were carried
out on an Instron Model 1122 under a constant crosshead speed of O.05"/min
(1.27 m/mn). The flexural measurements were done as per specifications
given In ASTM D790.
219
e. Estimation of 'Goodness of Mixing':
-
f
The extent of mixing between SAN and PMMA upon blending by the above
methods was followed by optical phase contrast miscroscopy and comparison
of FTIR spectra obtained from different sections of the samples.
Generally speaking, a well mixed blend would show no streaks or islands
when viewed under the microscope, and the FTIR of different sections of
such a sample should be identical.
RESULTS AND DISCUSSION
a. Phase Behavior of SAN/PMiA Blends:
SAWi (Table I) formed miscible blends with hetero and syndlotactic
PMMA over the entire range of composition. The i-PMMA was immiscible with
the SAN and so solution cast blends were prepared using SAN2. The dried
films were all optically clear and exhibited a single Tg by DSC, The
results are shown in Figure 1. The calculated lines were obtained using
the Gordon-Taylor equation:
ThereTn AT A + K
TgB) / (H + K HS)
(1)
where%A
B are weigh fraction of component A and B, and the
value of K was taken as equal to 1 (161. As can be seen from the figure,
the experimental results fit quite satisfactorily to the theoretical line,
for the SAN/i-PMMA and SAN/h-PNMA systems. More scatter Is evident for
the SANIs-PMMA system.
Figure I. Glass transition
temperatures of melt mixed
blends obtained by DSC. The
calculated lines were obtained
using the Gordon-Taylor
equation (16].
100e
*
so PS
i CM
60
4O.
%ISAN
The phase boundaries in these SAN/PMMA systems were determined by
locating the cloud points using the light scattering setup.
The cloud
points were measured by the detection of a sudden jump In the Intensity of
220
the scattering profiles and the corresponding temperatures were plotted as
a function of composition, thus generating the cloud point curves (CPC).
The system forms a well modulated, two phase structure above the cloud
point, similar to those discussed In the literature for other polymer
pairs (17].
b. Effect of PMMA Tacticity:
All the three types of PMMA form miscible blends with SAN and exhibit
LCST type behavior. Figure 2 is a plot of CPCs obtained at a heating rate
of 20 C/min for the hetero, syndio and isotactic PMMA blended with SAN and
shows clearly the effects of tacticity on the phase behavior. The hetero
and syndiotactic blends behave similarly, whereas the isotactic blend
shows a minimum in its CPC at higher SAN compositions.
340
22o
A
Figure 2. Cloud point curves
of blends of SAN with
isotactic, syndiotactic and
heterotactic PMMAs. SANI was
used to blend with s-PMKA and
h-PMMA; SAN2 was used for
blending with i-PMMA.
2"
I3S AWI'4%MA
20
410
6
% SAN
Blends of SAN through SAN4 with the three types of PMMA were also
prepared and analyzed for their phase behaviour. The results, for 50/50
blends by weight of SAN/PMWA are given in Table II. It is clear that in
Table II
TACTICITY EFFECTS
Composition: 50150
CLOUD PoINT (.C)
% AN
SAEh-PWAA
30
192
25
>300 or
199
232
20
>300 or
300 or
240
10.4
Immscible
'ift
al 2oCVMn.
SAo-PMMA
Immlsci*
immiscible
SAMS-PUMA
210
> 300 w
221
this system, the miscibility window is the largest for s-PI*lA and is the
We have estimated the individual interaction
smal lest for I-PI4MA.
parameters for the SAN/i-PMMA and SAN/h-PMMA systems from the phase
The
boundary data using the procedures of Krammer and Kressler (18].
results indicate enhanced interaction between the AN and methacrylate
groups for the I-PMMA blends compared to the h-PMMA blends (XNJd.IA0.085 and 0.07 respectively). The reduced width of the miscibili ywindow
for the SAN/i-PMMA could be related to this enhanced interaction.
c. Processing. Morphology and Mechanical Progerties:
The pure components, h-PMMA, SAN (30% AN) and a 50/50 mixture of the
two polymers were extruded using a Haake Single Screw Extruder with a slit
die. The pure components gave optically clear (1cm x .1cm) strips. The
blends, however, all appeared to be turbid at the surface. Cross sections
of the extruded strips showed a clear inner core, with opacity Increasing
Figure 3 is a phase contrast optical
toward the outer regions.
photomicrograph of a microtomed cross-section of the extrudate and shows
areas of incomplete mixing. To further confirm this state of incomplete
mixing, FTIR spectra were obtained on various areas of these microtomed
(15pm) samples. The spectra obtained from an area near the core appeared
to be uniform. Analysis showed it to be essentially a 50/50 mixture of SAN
and PMMA. The spectra obtained from an area near the skin that appears to
be an inclusion showed PMMA to be the dominant component in this outer
area. Outer areas show various combinations of SAN and PMMA, indicating
Incomplete mixing.
Figure 3. Optical phase
contrast photomicrograph
of a 15 pm thick crosssection of 50/50
SAN1/h-PMA extrudate.
The uniform dark areas are
regions of good mixing.
222
To improve the mixing, two additional techniques were used. In the
first attempt, about 30g of the two polymers were mixed in the Haake
Buehler Rheocord System 40. The conditions that were varied included time
of mixing, and mixing speed.
The mixed samples were then molded at
various temperatures and then subjected to flexural measurements (Table
III). The flexural strength increased with molding temperature.
Table III
FLEXURAL STRENGTH
vS. MOLDING TEMPERATURE
30 'g
6MW.
Mb"Condkmu
MOLWS
40. 10¢.C
PUMP"
2W0
225
240
250
26
91.57
0&.33
Mal8
0&53
100.12
The polymers were also melt blended using the Fiber/Film Processor with
static mixer and a coat-hanger die. The extruded films were all optically
clear with no indication of phase separation.
T e samples were also
analyzed for compositional homogeneity by FTIR and 11 NMR. The results,
shown in Table IV,
Indicate that good mixing was obtained using this
n h)ecase of FTU R ATR the ratio of the absorbance of' t~e C O
teih u I
peak at 1150 cm(due to PMA) to that of the aromatic at 760 cm-" (due
to SAN)
Is a constant for various regions of the sample,
tndicatin§j
compositti onal1homogeneitty.
Table IV
COMPOSITIONAL ANALYSIS OF
EXTRUDED SAMPLES
- STATIC MIXER
SAMPLE
A1IIS/A750
AREA
1
.SAWpuMA
(50/50)
]wm~
mm
l- mi
i61240
(50/,50)
(30170)
(30070)
i
i
2
1
2
-WT%*
(FTIR)
1.62
1.62
0.52
0.67
906
PUMA
49, 2
53.2
26.6
30.0
SAN
60. 8
44L.8
71.4
70.0
•'HNMR
The samples were also directly compression molded in avacuum press.
The appropriately weighed mixtures were molded at 1800, reground and then
remolded In a vacuum press at various temperatures.
They were then
analyzed for their flexurai properties and compositional homogeneity.
The
results of the flexural measurements are shown in Figure 4.
The solid
line in the center of the figure is based on simple addittivty rule for
binary mixtures.
The figure
clearly shows
an
improvement
in the
properties when molded at 2500C.
At this temperature, an interconnected
223
morphology is visible under the microscope. The samples molded at 200 0C
showed insufficient mixing as determined by FTIR. This Is reflected in
the data points
falling quite close to the calculated line. The samples
molded at 275 0 C, show both phase separation and the onset of sample
degradation as evidenced by the appearance of slight brownish tinge in the
sample.
900
eA
A
so
Figure 4. Flexural strengths
(ASTM D790) of SANl/h-PMMA dry
blended and compression molded
at various points in the (T,O)
phase diagram. Measurements
were performed at 230C and 50.
humidi ty.
op
C_
mmA
to
40
SO
j aft
s0
%SAN
Table V compares the flexural strength of blends obtained by the
different techniques. The results indicate that the Rheocord 40 gave the
best combination of good mixing and improved flexural strength. But the
technique cannot be operated continuously like an extruder.
It is,
however, a very useful tool for exploratory mixing studies. The physical
blending did not yield extremely well mixed samples.
Nevertheless, the
improvement In the mechanical properties shows that extremely Intricate
level of mixing may not be necessary for reinforcement, if the
morphological contributions can be controlled [19].
Table V
PROCESSING TECHNIQUE
vs. FLEXURAL STRENGTH
Sample: 50/50 SAN/h-PMMA
Processing/Molding Temp.: 250C
FLEXURAL.
MIXING
PRCSSING TECHNIOUESTRENGTH
(MPa)
CHARACTERISTICS
Extder (Single Screw)
Static
Mixer
Rhacord 40
Compression Molding
90.72
87.01
96.53
90.12
Average
Good
Good
Poor
224
CONCLUSIONS
The PMMA tacticity influences the phase behavior of blends with SAN
by altering the windows of miscibility. The blends of SAN with s-PMMA has
the widest miscibility window while the SANII-PMMA has the narrowest for
the three Isomers. The minimum In the CPC for the i-PMMA blends occurs at
higher SAN compositions compared to the h-PMMA and s-PMMA blends.
FTIR ATR spectroscopy is a convenient way to determine the
effectiveness of mixing In binary blends. Blending using the Rheocord
System 40 followed by compression molding gives the best combination of
adequate degree of mixing and improved flexural strength. However, the
various techniques used for melt mixing left a complex history of shear,
elongation, compression and spatial distribution on the multiphase melt of
the polyblends. Hence the properties reported here could very well be a
'compromised' average.
ACKNOWLEDGMENT
He wish to thank GenCorp for giving us permission to present this
study. He also wish to thank L. F. Marker for his helpful suggestions.
We acknowledge the assistance of I. G. Hargis, P. G. Venoy, P. T. Suman,
S. H. Daroowalla, and P. Brookbank at various stages of this work.
REFERENCES
1. J. K. Gillham, L. C. Chan, A. J. Kinloch and S. J. Shaw, Intn'l.
Conf. Toughening of Plastics II, 1, (1985).
2. L. A. Utracki, Intn'l. Polym. Processing, 2, I (1987).
3. J. H. Barlow and D. R. Paul, Ann. Rev. Mat. Sci., 11, 229 (1981).
4. S. Krause, J. Macromol. Sc. - Rev. Macromol. Chem. C7(2), 251 (1972).
5. 0. Olabisi, L. M. Robeson and M. T. Shaw, "Polymer Polymer
Miscibility," Academic Press, NY (1979).
6. I. C. Sanchez
In "TIer___Qmpait1!ty
and__1UnfllnjtbI!ity
Prl_3_ and Practice," edited by K. Solc, MMI Press Symposium
Series, Vol. 2, Harwood Publishers, NY (1982).
7. D.
W. Fox and R. B. Allen, Encycl. Polym. Sci, and Engg., 3, 133
(1985).
4
8. J. D. Hoffmann and R. B. Miller in 'Advancing Materials Research,'
edited by P. Psaras and H. D. Longford, Nat'l. Acad. Press,
Hashington, D.C. (1985).
9. M. Suess, J. Kressler and H. H. Krammer, Polymer 28, 957 (1987).
10. D. R. Paul and J. W. Barlow, Polymer 25, 487 (1984).
11. J1.M. F. Cowie and D. Lath, Makromol. Chem. Macromol. Symp. 16, 103
(1988).
12.
13.
14.
15.
16.
17.
18.
19.
L. A. Utracki and B. D. Favis in Hb9JQojPo
.5Lt.h.
Vol. 4, edited by N. P. Chemisinoff, Marcell-Dekker, NY (1989).
H. E. Fowler, J. H. Barlow and D. R. Paul, Polymer 28, 1177 (1987).
R. Subramanian, R. D. Allen, J. E. McGrath and T. C. Hard, Polymer
Preprints 26(22, 238 (1985).
A. M. Halstrom, R. Subramanian, T. E. Long, J. E. M-Grath and T. C.
Hard, Polym. Preprints 27(2), 135 (1986).
K. Naito, G. E. Johnson, D. L. Atlana and T. K. Kweim, Macromol. Ll,
1260 (1978).
T. Inoue, T. Oulgizawa, 0. Yasuda and K. Mlyasaka, ibild, 18, 57
(1985).
H. H. Krammer and J. Kressler, Makromol. Chem. Macromol Symp., 18, 63
(1988).
P. Van Ghelue, B. D. Favis and 3. P. Challifoux, 3. Mat. Scl. 23,
3910 (1988).
225
Dielectric Studies of Polyester/Polyoarbonate Blends
James X. O'Reilly and Joseph S. Sedita, Eastman Kodak Company.
Corporate Research Laboratories, Rochester, NY 14650-2110
Abstract:
Dielectric and enthalpio relaxation times have been measured
as a function of composition and temperature. A fractional exponential (W-W) distribution of relaxation times fits both the
dielectric and enthalpio relaxation times with different values which are a function of composition. Free volume parameters
calculated using the WLF equation are not simple functions of
the composition. Concentration fluctuations are considered to
be important in these phenomena.
Introduction:
There is widespread interest in polymer blends which offer the
opportunity to obtain improved properties at lower cost. The
class of miscible and immiscible polyester/polyoarbonate blends
have been extensively studied (1) and reviewed (2). Copolyesters of ethylene glycol (EG) and cyclohexanedimethanol(CHDM)
and terphthalic acid (T) and bisphenol A polycarbonate(PC)
blends were studied by Paul and co-workers (3-6). These blends
were miscible under most conditions but because of high temperatures, 250C and higher required for melt blending, transesterification of the polymers may contribute to the miscibility. PET is not miscible with PC in the absence of transesterification. CHDMT is reported to be miscible with PC by Paul(7)
We
willcopolymer
report dielectric
and blended
thermal in
properties
of a20/80
withvarian extruder
of EG/CHDH/T
mole%
ous amounts of PC. The temperature dependence of the dielectric
and thermal relaxation times will be analyzed in termsof the
free volume models of relaxation.
Experimental:
Merlon X40 PC used in these studies has Mw-31,000 and Tg is
140C. The oopolyester EG/CHDM 20/80 terphthalate is a commercial product of Eastman Chemicals and has Xw-39,000 and Tg is
85C. The blends were melt extruded at 285C and films were mol
ded in compression at 295C. DSC measurements were made using a
DSC II. Dielectric measurements were made using a DETA apparatus from Polymer Labs over a temperature range of -100 to 200
C and from 1 to 100 kHz. Only tang measurements are reported
here because they show all the interesting behavior and are
the most accurate.
Dielectric Results:
A beta relaxation process is observed in both the polyester
(Tmax --10C) and polycarbonate homopolymers (Tmax - -BOC). The
beta process of the blends appears to be an average of the two
pure component losses and little additional information is
discernible from these measurements. In the Tg region (alpha
relaxation) a single narrow peak is observed for each blend
and indicates that the blends are miscible (Fig. 1.) For the
pure polyester, an additional peak is observed at high temperature which is attributed to reorystallization of the polyester
Mal. Res. Soc. Symp. Proc. Vol. 171. , 1990 Materials Research Society
226
and iads to an alpha' loss oharaoteristio of semi-orystalline
polymers. The intensity of the loss peak is not a linear funotion of the composition. This may be due to an anti-parallel
dipole-dipole interaotion leading to a lower effective dipole
moment.
These data can be presented as Tmax (temperature of maximum
loss for different frequencies) as a function oomposition in
Fig. 2. It is apparent that the Tmax of dielectric loss is a
different function of composition than is the Tg determined
from DSC measurements. This result was initially surprising
because free volume considerations would suggest linear behavior. Another representation of the frequency, temperature, and
composition dependence of the dielectric loss at Tg is shown
in Fig. 3. In this form it can be readily seen that the slopes
of the frequency vs temperature curves (i.e. activation energy
or WLF parameters vary systematically with composition. The
precision of the data is well represented in this curve. Reproducibility of the temperature is better than one degree, the
composition is known to within a few percent and the error in
the frequency is negligible compared to the other variables.
These data can be fitted with a WLF/VTF equation to obtain the
parameters Cl and C2 for different compositions.
log f(T) - log fCTg) -C(.T-Tg - A
C2+C--Tg)
--IT-To
(1)
The solid line in Fig. 3 is a non-linear regression fit of the
data and the calculated values of C1, C2, and log f(Tg) are
listed in Table I. Tg was taken as the onset temperature from
the DSC measurements at 20 deg/min.
%PC
0
20
50
60
80
100
Dielectric
C1
10.7
13.7
13.3
14.7
13.3
11.3
Table I
Relaxation Parameters
fg
C2
logfm(Tg)
.040
39
.35
.032
-1.5
27
.032
54
.20
.030
52
.32
1.34
.033
63
.038
48
2.4
afxlo
10
12
6.0
5.7
5.1
8.0
4
.59
.54
.51
.43
.51
.50
Many polymers and other materials can be described by the
Williams-Watts distribution function for the alpha and beta
relaxation. We used the W-W function to fit the alpha relaxation of the blends as shown in Fig. 4. In order to cover a wide
reduced frequency range, the temperature data are converted to
reduced frequency using eq. 1 {8}. Values of/I calculated for
a frequency of lkHz. are included in Table I. Data at higher
frequencies give slightly lower values ofj which indicates
that the distribution broadens with temperature.
DSC Results:
The thermal behavior of these polymer blends reflects the longer relaxation times associated with oonformational changes
and intermolecular energy changes at Tg. Tg changes with heating rate as shown in Fig. 5. and the activation energies, Ha,
calculated from the Arrhenius equation are listed in Table II.
The derived Ha show a maximum at high polycarbonate oonoentrations but this maximum should not be given much emphasis beoa-
227
.4
........
ajjou4u[4gDI00
228
use Ha of PC is low and the precision of the Ha is about ±20%.
The heating rate results for the 20% and 100% PC samples
appear to give low Ha's and the experiments were repeated for
the PC sample with similar results. Nevertheless Ha changes
with PC concentration. We have applied the widely used method
of enthalpy relaxation to these blends. Since the analysis is
fully described {9,10} only the essentials will be repeated
here. The parameters in addition to Ha, needed to characterize
the relaxation are the breadth of the distribution of relaxation times and the structure dependence of the relaxation times
which is proportional to {1-X). A characteristic relaxation
time, to, is also necessary. The eq.2.
(2)
71o + XHa+ {l- 1XKq_
RT
RTf
is used to analyze the data. Tf is the fictive temperature.
A characteristic of polymer blends is that the annealing peak
associated with aging is smaller or broaden with blending as
shown in Fig. 6. Only the short time aging behavior is reported here because a shortooming of this model is that the parmeters X and/, change with long annealing times {11). Values of
X and/,f derived from non-linear regression fits of the Cp curves of aged samples vary with composition and are listed in
TableII. An interesting feature of these data is the that the
distribution of relaxation times are narrower and more composition dependent for enthalpy relaxation than for dipolar relaxation (Fig. 7.). The changes in Ha with composition may
influence the calculated values ofA4 • Nevertheless the dielectrio4 's clearly show that Ha increases with PC concentration
and this effect direotlV or indirectly leads to the changes in
the enthalpio/A's.
Table II
Enthalpic Relaxation Parameters
X
Ha kJ/mol I
%PC
.74
.68
107
0
.83
.71
100
20
.84
.45
153
50
.74
.51
145
60
.80
.41
188
80
1.04
.74
107
100
ln1-
Discussion:
The dielectric and enthalpic-'s can be combined and analyzed
as shown in Fig.8. The dieleotri -'s are referenced to the Tg
at 20 deg./min and then the Tg's at other heating rates are
shifted relative to the 20 deg./min value. These data were then
re-analyzed using the WLF/VTF equation and the values were not
statistically different from the values in Table I except for
the 20 and 100% PC. In terms of the free volume models fg should be a constant at Tg. In these blends it appears that fg
decreases and shows a shallow minimum with composition. We expected af to show a monotonic behavior with composition but it
fluctuates with composition. Because of the limited frequency
range and experimental uncertainties we do not want to speculate about the fluctuations in af. However the changes in fg
are real and could be associated with negative volume changes
on mixing the blends {121. Although the density changes on
mixing are usually less than 1% the changes in fg are larger
than this and reflect larger change in relaxation volume on
mixing. In addition the dielectric relaxation time at Tg
229
00
~0
4
0-
64
sit
-u
OS
*
0s
P40IA
230
varies with composition and this could be related to the ohangee in fg. Since the distribution parameters are different and
have different dependences on composition, it is clear that the
concentration fluctuations on a molecular level affect
properties in different ways.
We have learned from this study and related research that the
relazational and time-dependent properties of polymer blends
are not simple monotonic functions of their compositions as
might be inferred from the glass transition behavior. These
phenomena are probably related to concentration fluctuations
in these blends which are intimately related to the thermodynamic interactions between blend components.
Acknowledgments: We thank Dr. Robert W. Seymour of Eastman
Chemicals Research Laboratories for the samples of these
blends and discussions on regarding them.
References
1. 0. Olabisi. L.M. Robeson, and M.T. Shaw, "Polymer-Polymer
Miscibility" Academic Press, NY (1979).
2. R.S. Porter, J.M. Jonza, M. Kimura, C.R. Desper, and E.R.
George, Polym. Soi. Eng. 29, 55, (1989).
3. C.A. Cruz, D.R. Paul. and J.W. Barlow, J Appl. Polym. Soi.
23, 589, (1979).
4. C.A. Cruz, D.R.Paul, J.W. Barlow. ibid. 24, 2101 (1979)
5. R. S. Barnum, J.W. Barlow. D. R. Paul. 27, 4065 (1982)
6. B.A. Joseph, M.D. Lorenz, J. W. Barlow, and D.W. Paul.
Polymer, 23, 112, (1982).
7. R.N. Mohn, D.R. Paul, J.W. Barlow and C.W. Cruz, J. Appl.
Polym. Soi. 23, 875, (1979).
8. N.G. MoCrum, B.E.Read, G.W. Williams, "Anelastio and
Dielectric Effects in Polymers", Chap. 4, John Wiley, 1967
9. X. Sasabe and C.T. Moynihan, J. Polym. Soi. Polym. Phys.
16, 1667, (1978)
10. I.M. Hodge and A.R. Berens, Macromolecules, 15, 762.
(1982), and ibid. 16, 371, (1983). 16, 898. (1983).
11. J.T. Tribone, J.M. O'Reilly, and J. Greener.
Macromolecules, 19. 1732 (1986).
12. M. Wolf and J.H. Wendorff, Polymer, 30, 1524, (1989)
231
Interpenetrating Polymer Networks and Related Topological Isomers
HARRY L. FRISCH
Department of Chemistry, State University of New York, Albany, NY 12222
ABSTRACT
We have studied the phase compatibility of simultaneous interpenetrating
polymer networks poly (2,6-dimethyl-l,4-phenylene oxide) with a number of
other polymer networks as a function of their solubility parameters. Even
when the blends of the corresponding linear chain polymers are immiscible
their interpenetrating polymer networks can be miscible in all proportions.
We have also prepared the corresponding pseudo (or semi) interpenetrating
polymer networks where one component is linear and in one system the
polymeric catenanes. All these systems generally show microphase separation.
1.
Introduction
Interpenetrating polymer networks (IPM's) are intimate mixtures of two
or more crosslinked polymers, held together predominantly by permanent
entanglements (1,21.
IPSI's can be prepared sequentially, i.e. by swelling
one formed crosslinked network by the monomers (or prepolymers) crosslinking
agents and initiators (or catalysts) of the other network (s) and
subsequently polymerizing in situ. They can also be prepared simultaneously
i.e. by mixing all the monomers (or prepolymers), crosslinking agents of the
different networks, initiators and catalysts and polymerizing all at the
same time. We will restrict ourselves here to simultaneous IPi's. Besides
their interest as macromolecular topological isomers these materials allow
one to produce more controlled and/or enhanced phase miscibility (3].
IPN's
are often the only way of producing miscible alloys of crosslinked
polymers. Besides depending on kinetic factors sueh as whether the
half-life of formation of permanent entanglements in the IPUs is shorter
than the half-life for phase separation due to "uphill diffusion" the extent
of incompatability is dependent on thermodynamic factors such as the free
volume change or the magnitude of the positive enthalpy of mixing found in
many systems. The latter can be crudely estimated by means of the
6
6
solubility parameter (31 (6) or parameters (31 ( d, &p, h )
associated with the corresponding linear polymers. To see the effect of the
latter we will focus on the properties of a number of recently synthesized
simultaneous IPN's [4-91 in which one network consists of crosslinked poly
(2,6-dimethyl-l.4-phenylene oxide) (PPO) while the chemical nature of the
other component will be varied.
Related to the IPN's are the pseudo (or semi) IPH's in which only one
component polymer of the mixture is crosslinked and the other component is
left as a linear polymer 11,21. Finally if one polymer is available as a
sufficiently large ring then such rings can be permanently trapped in a
different crosslinked polymer network forming a macromolecular catenane
[10-121.
We will briefly comment on these other macromolecular topological
isomers of the PPO-IPv's.
Mat. Re. Soc. Symp. Proc. Vol. 171. t1990 Materials Research Society
232
2.
PPO-IPN's
PPO can be conveniently crosslinked by first methyl brominating it
using N-bromosuccinimide and then condensing the resulting bromine
containing polymer with ethylene diamine. PPO was chosen because the
linear polymer is compatible at all compositions with linear polystyrene
(PS). It is not surprising therefore that the PPO-PS IPN's, in which the
PS is crosslinked by a free-radical reaction, which does not interfere with
brominated PPO condensation, using divinyl benzene and benzoyl peroxide,
There is
[4].
produces homogeneous single phase IPH's at 6all compositions
6
h ) of the
a perfect match of solubility parameters ( d,' &
linear polymers as can be seen from Table 1. Linear poly (methyl
methacrylate) (51 (PHKA), linear poly (urethane acrylate) 16] (PUA),
Table I
Solubility Parameters and Glass Transition Temperatures of the
Linear Polymers Corresponding to the Network Components of the
IPN's.
Linear Polymer
PPO
PMMA
PS
PUA
PB
PDMS
Solubility Parameter (s)
6
3
6
2 0
p= -0, h= .
d=8.
2
6
5
d=9.2,6p= .0, h=4.
60=8.6,6 =3-0,6b=2.0
6=9.1
6=8.38
6=7.3
6
Glass TrisiLtion
Temperature (*C)
211
105
100
-95
-127
linear polybutadiene (PB) and linear poly (dimethyl siloxane) [7] (PDNS)
are not miscible to any significant extent with linear PPO, as would be
expected from the solubility parameters listed in Table 1. Simultaneous
IPN's of PPO and PMMA, PUA, PB and PD0S can be produced employing free
radical crosslinking which appears not to interfere with the PPO
crosslinking mechanism. For experimental details we refer our readers to
the original papers 14-91 which also describe how the pseudo-IPN's of some
of these systems were prepared.
These IPM samples were studied by transmission electron microscopy.
The glass transition temperatures (Tg) were obtained from differential
scanning calorimetry and in some cases confirmed by rheovibron or other
thermal-dynamical spectroscopy. Instron stress-strain measurements
provided tensile stress and elongation to break data 14-9]. Measurements
were made generally on both unextracted and solvent extracted samples. In
a number of instances, on the homogeneous samples the average molecular
M
weight between crosslinks ( c ) deduced from stoichiometry was confirmed
M
by swelling studies. In general 14-91 these c values varied from about
several thousand to a maximum of seventeen thousand MA].
Binder and Frisch [131 had already theoretically suggested that weakly
interpenetrating simultaneous IPY's composed of immiscible linear chains
could form single phase IPW's over essentially the whole composition
range. Thus, the IPY's of PPO and PHMA, PPO and PUA and PPO and PB do
this. Some particulars are given in Table 2. These generally transparent
to slightly translucent materials show no microphase separation in their
transmission electron micrographs and exhibit a single TS intermediate in
value to the Tg's of the pure component crosslinked networks 14-6,9].
233
When the solubility parameter difference (cf. Table 1) becomes
sufficiently large as with PPO and PD0S we found microphase separation at
almost all compositions 17,81.
These samples showed two inwardly
displaced Tg's and exhibited phase domains ranging from a few tenths of
nanometers to one to two hundreds of nanometers 18].
Table II
Compatability of the PPO-IPM's end the weight percentage
of PPO at which the maximum value of the tensile strength to
break (ASTND 638) is found
IPN
Linear
Polymer
Compatability
IPM Compatibility
PPO-PS
Miscible at
all compositions
IPf's are single phase
and exhibit a single Tg
decreasing monotonically
with wt % PPO
Wt. % of PPO
Reference
75
14)
60
15]
PPO-PW. Linear
polymers
are immiscible
PPO-PUA
o80
PPO-PB
PPO-PDMS
"
IPY's exhibit two Tg's and
exhibit phase separation;
become most miscible around
90 wt. % PPO
[9]
90
[6]
90
(7,81
Under certain conditions [7,8] around 10 weight percent PDMS single phase
IPN's could be found.
While the Tg's of these IPN's were always intermediate to the Tg's of
the pure crosslinked component networks and varied monotonically with
composition.the tensile stress to break (T.S., AST638) of all these full
IPH's showed a maximum at an intermediate composition, which is noted in
Table 2. The exact reasons for this have not yet been experimentally,
fully confirmed but the prevailing speculation (1,2,4] is that at those
compositions the IPN exhibits the maximum extent of permanent
entanglements which also need to be broken to provide ultimate failure of
the sample.
3.
'1PUA.
I
Other Topological Isomers Involvinj Crosslinked PP0.
The pseudo IPV's of cross-linked PPO and linear chains of the other
polymer are generally opaque to translucent solids except for the PS and
The pseudo IPM's of PPO and PMMA, PB and PDOS exhibit two glass
transition temperatureSintermediate in value to the corresponding pure
polymers. The tensile strengths to break are usually monotone decreasing
functions with decreasing weight percentage of PPO.
& number of polymeric catenanes 110) of cyclic PD0S consisting of
rings from 33 to 122 siloxane units have been trapped in crosslinked PPO
234
(11,12]. As expected the larger the degree of polymerization of the
cyclic PDMS the larger is the fraction of trapped (unextractable) cyclic
in the PPO [141. These solid materials exhibit a melting point close to
that of the pure cyclic PDKS and a somewhat lower TS than the pure
crosslinked PPO. These polymeric catenanes reveal microphase separation
with domain sizes of 10-50 p [12]. The values of the tensile strength
to break of these catenanes is smaller or of the same magnitude (within
experimental error) as the pure crosslinked PPO.
4.
Acknowledgement
This work was supported by the National Science Foundation under Grant
DMR 8515519.
5.
References
1. D. Klempner and L.Berkowski, "Encyclopedia of Polymer Science and
Engineering", (John Wiley and Sons, New York, 1989), Vol 8, p. 282.
2. K.C. Frisch, Jr. and D. Klempner, editors, "Recent Developments in
Polyurethanes and Interpenetrating Polymer Networks", (Technomic
Publishing Co.. Inc., Lancaster, PA 1988).
3. 0. Olabisi, L.M. Robeson and M.T. Shaw, "Polymer-Polymer Miscibility",
(Academic Press, New York 1979).
4. H.L. Frisch, D. Klempner, H.K. Yoon and K.C. Frisch, Macromolecules,
13. 1016 (1980).
5. S. Singh, H.L. Frisch and H. Ghiradella, Macromolecules (in press).
6. H.L
Frisch and Y.H. Hua, Macromolecules, 22, 91 (1989).
7. H.L. Frisch, K. Gebreyes and K.C. Frisch, J. Polym. Sci. (Chem. Ed.)
26. 2589 (1988); K. Gebreyes and H.F. Frisch, ibid. 26, 3391 (1988).
8. W. Huang and H.L. Frisch, Makromol. Chem. Suppl. 15, 137 (1989).
9. P. Mengnjoh and H.L. Frisch, Polymer Letters 27, 285 (1989); J.
Polymer Sci. (Chem. Edit.) (in press).
10. D. Callahan, H.L. Frisch and D. Klempner, Polym. Eng. and Sci. 15, 70
(1975).
11. T.J. Fyvie, H.L. Frisch, J.A. Semlyen, S.J. Clarson and J.E. Mark, J.
Polym. Sci. (Chem. Edit.) 25, 2503 (1987).
12. W. Huang, H.L. Frisch, Y. Hua and J.A. Semlyen, J. Polym. Sci. (Chem.
Edit.) (in press).
13. K. Binder and H.L. Frisch, J. Chem. Phys., 81, 2126 (1984).
14. H.L. Frisch and 1. Wasserman, J. Am. Chem. Soc.,
83, 3789 (1961).
PART V
lonomers / Structure
237
SMALL ANGLE X-RAY SCA'TERING ON
POLY(ETHYLENE-METHACRYLIC ACID) LEAD AND
LEAD SULFIDE IONOMERS
BENJAMIN CHU*, DAN 0. WU
Department of Chemistry, State University ot New York at Stony Brook, Long Island, NY 11794-3400
and WALTER MAHLER
Central Research & Development Department, Experimental Station, E.. du Pont de Nemours & Co.,
Inc., Wilmington, DE 19880-0328
* author to whom all correspondence should be addressed.
ABSTRACT
The morphology of lonomers, e.g., poly(ethylene-methacryic acid) (EMA) lead salts (EMA/Pb)
and lead sulfide compounds (EMA/PbS), has been studied by using the techniques of small angle x-ray
scattering (SAXS), anomalous SAXS (ASAXS), wide angle x-ray scattering (WAXS), and differential
scanning calorimetry (DSC). EMA/Pb containing less than 5 wt% of lead exhibited two characteristic
SAXS peaks which corresponded to the lamellar structure of the partially crystalline polymer matrx and
the ionic structure of the lead aggregates that were present in the amorphous regions. The lead
aggregates were not distributed uniformly and increased in packing density with Increasing lead
content. Both DSC and WAXS showed that the crystalline phase was present for all EMA/Pb samples
and that the crystallinity decreased slightly with increasing lead content. ASAXS near the L3 absorption
edge of lead permitted the extraction of the scattered intensity of lead ions from the SAXS patterns of
the superimposed crystalline and ionic structures. Correlation function analysis revealed that the ionic
aggregates of the EMA/Pb containing 5 wt% of lead could be described by a liquid-like model with a
short range order of 2-4 nm. EMA/PbS samples were made by a reaction of EMA/Pb lonorners with
hydrogen sulfide. Instead of an Ionic peak as shown by EMA/Pb samples, the SAXS patterns of
EMA/PbS showed a broad diffraction peak located at the same q value as the lamellar peak of the EMA
in acid form. The (lamellar) peak could be attributed to the interference between the PbS crystallites in
the neighboring lamellae.
INTRODUCTION
Copolymer poly(ethylene-metharylic acid) (EMA, 85 wt% of ethylene) is partially crystalline
because of the presence of a large amount of ethylene segments. EMA ionomers are usually the metal
salts of EMA whose carboxyllc acid groups have been partially or completely neutralized. Sudyn, (a
trade mark of Du Pont) an EMA In sodium or zinc salt form, has been well-known for its superior
properties as packaging and coating materials, thanks to the ionic crosslinks formed by the aggregation
of metal ions. Although ethylene-carboxylate Ionomers have been studied extensively on their structureproperties relations [1-121 and several models have been proposed to describe its morphology [4-5, 7,
10-11[, controversies about which model to use still remain. This work aims at a better understanding of
the morphology of EMA/Pb lonomers by carefully designed small angle x-ray scattering (SAXS) and
anomalous SAXS (ASAXS) experiments and by utilizing the correlation function analysis, a
straightforward and effective method of selecting a correct model and of determining the structural
parameters from SAXS and ASAXS patterns covering a very broad q range.
Recently, Mahler treated EMA/Pb films with H2S to obtain a composite polymer material
(EMA/PbS) possessing semiconductor properties [13]. In EMA/PbS samples, PbS crystaliltes are
embodied in the semicrystalline polymer matrix. A morphological characterization of EMA/PbS is
desirable because the size and the spatial arrangement of PbS particles are the key factors that control
the semiconducting properties.
Anomalous small angle x-ray scattering is capable of varying the scattering power of a selected
element [14-20]. Its application to the EMA/Pb and EMA/PbS samples allows us to remove any nonlead related scatterings Including the reflection of the ethylene lamellae, possible Impurities and voids
and to study the structure and the spatial arrangement of lead Ions or PbS particles in the polymer
matrix. In this paper, the use of ASAXS technique in the EMA ionomer structural determination Is
exemplified by the EMA/Pb containing 5 wt% of lead.
EMA/Pb AND EMA/PbS IONOMERS
Poly(ethylene-methacryllc I7 cd) (EMA) copolymer containing 15 wt% (or 5.4 Mol%) of
methacrylic acid had Mw-8.6xl0 g/mol and MW/Mn-7. The EMA/Pb ionomer containing 20 wt% of
Mat. Res. Soc. Symp. Proc. Vol. 171. t1990 Materials Research Society
238
lead was obtained by neutralizing EMA with lead acetate. The neutralization was accomplished by
milling EMA with lead acetate at 1600C. The ionomer sample containing 20 wt% of lead was then
compression molded Into films. The samples with 10, 5.0, 2.0, 1.0, and 0.2 wt% of lead were
subsequently diluted from the 20 wt% sample by mixing with an appropriate additional amount of the
lead-free EMA. Fairly homogeneous films could be obtained by reprocessing the sample several times.
The ionomer films with a thickness ranging from 0.1 mm to 0.3 mm were essentially transparent. The
0
samples were annealed at a 60 C vacuum oven for 24 hours before SAXS and ASAXS measurements.
EMA lead sulfide compounds (EMA/PbS) containing 20, 10, 5, 2, 1, 0.2 wt% of lead were made
by exposure, respectively, of the corresponding EMA/Pb films to 1 atm 12S at room temperatures for at
least 2 hours. The change of color of the EMA/PbS films from light brown to black for 0.2 to 20 wt% of
lead suggested the formation of PbS particles In the EMA polymer matrix.
Differential scanning calorimetry (DSC) measurements (using - 5 mg sample and 10 degree/min
heating rate) of the ErMA/Pb samples showed a slight decrease in heat of fusion and melting
temperature with increasing lead content. Wide angle x-ray scattering (WAXS) showed the ethylene
(110) and (200) reflections. Both DSC and WAXS suggest the presence of the ethylene crystalline
phase, although the crystallinity decreases slightly with increasing lead content. WAXS of the EMA/PbS
showed reflections of PbS crystallites that were enhanced (became sharper) with increasing lead
content suggesting growth in the size of PbS particles in the semicrystalline polymer matrix.
Tables 1 and 2 list the DSC results, respectively, of the EMA/Pb and EMA/PbS samples.
Table 1. Physical Constants for EMA/Pba
Pb content (wt%)
lamellar peak position (1/nm)
ionic peak position (l/nm)
ionic peak height
heat of fusion (J/g)
T CC)
0
0.6
c
(4)
39
92
0.2
0.6
-3.6
-5
2
0.6
3.5
20
37
92
5
(0.6)
3.5
48
36
92
10
20
-
2.9
125
30
89
2.5
210
19
86
10
0.6
22
89
20
Table 2. Physical Constants for EMA/PbS4
Pb content (wt%)
lamellar peak position (1/nm)
heat of fusion (J/g)
T. CC)
0.2
0.6
37
91
1
-0.4
33
90
2
0.5
5
0.55
30
90
-
26
87
aData of heat of fusion and melting temperature T were collected from the endotherms of second heating circle
,o ensure the same crystallization hk*tory for all samtples.
CAppearance of lamellar peak was sensitive to sample annealing.
Ionic peak was hard to observe. The scattered intensity value at q- 3.6 nm was taken.
SMALL ANGLE X-RAY SCATTERING (SAXS)
All scattering experiments were conducted at the State University of New York (SUNY) X3A2
Beamline, National Synchrotron Ught Source 121]. A small angle x-ray diffractometer (SAXD) equipped
with a modified Kratky block collimator 1221and an x-ray photodiode array (PDA) detector 1231was
capable of making SAXS as well as ASAXS measurements with a q (=(47r/X)sl(0/2), with A and 0
being, respectively. x-ray wavelength and scattering angle) range of 003-13 nm . A Si (111) double
crystal monochramator with an energy resolution of ± 5 eV was used. A gold-coated quartz torroldal
mirror (60 cm long, 60 mm radius) used in SAXS at A-0.154 nm was removed in ASAXS measurements
near the L3 absorption edge (Eedge
= 13040 eV Or x -0.095 nm) of lead, because t could not reflect and
focus x-rays of wavelength shorter then 0.097 nm to the small angle x-ray diffractometer (SAXD) which
was -10 m away from the mirror. The x-ray beam had a cross-section of -0.2 mmx-2.0 mm at the
polymer sample. The jlstance between the sample and the detector was 220 mm. Desmeerdng was not
needed for q> 0.2 nm
239
SAXS measurements were performed at room temperature. The SAXS patterns were corrected
for detector sensitivity, parasitic scattering, detector dark counts, sample attenuation, Incident x-ray
intensity fluctuation, and normalized to I mm sample thickness and I sec experimental duration.
Figure 1 shows SAXS patterns for the EMA/Pb lonomers of, respectively, 20, 10, 5, 2. and 0 (in
acid form) wt% of lead. A diffraction peak at q - 0.6 nmI, indicated by a double arrow, could be
observed upto 2 wt% of lead (also In the sample with 5 wt% of lead after annealing the sample). The
(lamellar) peak became less sharp and weaker with Increasing Pb content and eventually disappeared.
Another broad diffraction peak, located at 2.5 < q < 3.6 nm", marked by an arrow, shifted to a smaller q
value and grew in peak height with increasing lead content.
The peak at q-0.6 nm" Is due to the lamellar structure of partially crystalline ethylene segments.
Little dependence of the lameilar peak position on the lead content was observed suggesting that the
lamellar structure was hardly affected by a small amount of lead aggregates (:s 2 wt%) in agreement with
the DSC results (Table 1). As DSC showed the presence of the ethylene crystalline phase even In the
EMA/Pb with 20 wt% of lead, diminishing lamellar peak with increasing lead content suggests that the
lead ions were distributed mainly In the amorphous region so that the electron density difference
between the crystalline and the amorphous regions was reduced with more lead Ions present In the
amorphous region. The electron densities of the crystalline and the amorphous phases were about to
be matched In the EMA/Pb containing 5 wt% of lead. Further increase in the lead content could
certainly affect the formation of lamellar stacks because the aggregates of Pb groups could form
crosslinks that would reduce the folding of ethylene segments and thereby the crystallinity. DSC also
showed a decrease In the heat of fusion and the melting temperature especially In the sample
containing 20 wt% of lead (Table 1), The disappearance of the lamellar reflection In the EMA/Pb
samples of higher than 5 wt% of lead Implied that (1) the lead ions were not distributed uniformly in the
amorphous region, otherwise a lamellar reflection should have reappeared and peaked at the same q
value as a consequence of "reversed-density" lamellae whose amorphous phase (containing ions)
would have a higher electron density than that of the crystalline phase and/or (2) the scattering from
the ionic aggregates oversnadowed that from the lamellae.
The peak at 2.5-3.6 nm' Is clearly the Ionic peak which originates from the lead Ion aggregate j
The magnitudes of the ionic peak as well as the small angle scattering (upturn) at q< - 0.6 nm,
another well-observed feature of Ionomers [6], are Increased with increasing lead content suggesting its
contribution from the ions 125,26]. Table 1 lists the positions and the heights of the two characteristic
peaks of the EMA/Pb samples.
Figure 2 shows SAXS patterns for the EMA/PbS samples containing, respectively, 20, 10, 5 2 1
and 0.2 wt% of lead. No appreciable Ionic peak was observed. A broad diffraction peak at q - 0.6 nm
was observed for all the EMA/PbS samples. The peak position was located near the lamellar peak
position of the ethylene crystalline matrix and could be related to the reflection of larnalla-Ilke structures
due to the spatial distribution of PbS particles. The above interpretation could be supported by the
following observations. At 0.2 wt% of lead, the number and the size of the PbS particles were so small
that the reflection was due to Inter-ethylene amellae. At 1 wt% of lead, however, the reflection peak
became less sharp with an Il-defined peak poeltior. Since the electron density matching between
crystalline and amorphous phases should occur near 5 wt% of lead as we have observed for the case of
peak could ory be attributed to the randomness In the
EMA/Pb with 5 wt% of lead, the Ile
concentration of the PbS particles In the periodic lamellar layers. This suggestion Is supported by the
EMA/PbS samples containing 2 of more wt% of lead. The increase in the reflection peak height and the
shift of the peak from q-0.4 nm at I wt% to q-0.6 nm" at 10 wt% illustrated that the PbS particles
grew In size and at the same time arranged themselves In a more ordered manner. A simlar larnellar
peak growth pattern was observed In the polyethylene crystallization [271. It Is thus reasonable to
conclude that In the EMA/PbS sample containing I or more wt% of lead, the PbS crystallites contribute
to the lamellar reflection at q-0.6 nm". The reflection peak of the EMA/PbS samples was broader than
that of the EMA and might be explained by the fact that the PbS particles with finite particle sizes would
have a particle form factor superimposed to the lamellar reflection. The EMA/PbS with 20 wt% of lead
showed very different behavior from the others. There, the PbS particles might no longer be confined
within the ethylene lamelae. The peak positions of the EMA/PbS samples are listed In Table 2.
The most straightforward Interpretation of the ionic peak as shown In FIgure 1 was made by
Longworth [4] who assumed that the Ionic aggregates occupied the lattice points and attributed the
Ionic peak to Bragg reflection of those Ionic aggregates. Thus the Inter-aggregates distance could be
estimated as ~ 2r/qpck. MacKnght at al. attributed the Ionic peak to core-she shaped aggregates
that had no Inter-aggregate Interference (7]. An alternative core-shell model was given by Hashimoto at
a/. (24]. Cooper et at., on the other hand, explained the ionic peak in temis of IlquklIke Interaggregates (spheres) Interference [111. The use of these morphological models, however, requires
avalabllity of the excess scattered Intensity from the lead ions. Furthermore the model fitling of SAXS
patterns measured In a limIted q-range often prevents a critical comparison Ol the theotical models
with experimental data. These difficulties could be resolved by using ASAXS covwng a very 'road q
240
E"
Soo
920
m4aOO
Wt
* 10 wt$
- 5 Wts
-
F5
SAXS patterns of the EMA/Pb
containing respectively, 20, 10, 5, 2, and
wt% of Pb. Double arrow indicates
6
form)
0 (acid
of the ethylene lamellae at q-0.
the refl
. ecion increasing Pb content, the peak
nma With
Figur1.
pm
b
"samples
becomes less sharp and disappears for EMA/Pb
wth more than 5 we% of lead. Arrowmarks the
3 6
2
is
peak, q- .5- . nm , which Pb
ionic diffraction
originated from the Pb ion aggregation. The
in the
aggregates are not distributed uniformly
region of the EMA polymer matrix.
- a Wts
- icid fi
w20
*iamorphous
"
2A
0
150E
A/PS
Fgrt Z. SAXS patterns of the EMA/PbS
5, 2, 1, 0
samples containing, respectively, 20, 10,ionic
peak
(acid form) wt% of Pb. No appreciableand tbroad
is observed. Instead, a strong 6
p. ,
-20
Wtz
-'10 wtU
0100C ,V. "."
%
5
\
•
%%
500
-Cid
amorphous region with certain ordering. Thc
diffraction peak is due to interference of shePbS
particles in neighboring lamellae. The PhS
grows in size with increasing Pb
content.
wt%
foit
icrystallites
10
2
j
0
*"
3
ca nm)
Figure3. ASAXS difference scattering pattern
(5 wt)"
" EM /Pb
(hollow squares) of the EMA/Pb sample
containing 5 w% of Pb, which was obtained by
measured at 7eV
of the SAXS curve
subtractionPb
L3 absorption edge (corrected for a
the
that measured at
from
background)
fluorescence
-31- eV below the edge. Solid line represents
fitting of the fiquid-like model. Dashed line is the
structure factor corresponding to a Debyetype inhomogeneity. The ASAXS pattern
"ueche
is well described by the two structures. The -nodcl
are shown in Figure 4.
S£xPertaenta1
fin
SA
n , For
EMA/PbS with higher than I wt% of Pb, PbS
particles, in crystalline form, are distributed in the
wit
Wt
-5
-2
-
"
6-4
diffraction peak is seen at q~O.
-LIquid Model
Sbelow
*
'* ..
._
0 ,parameters
h
0
0.08
2
EM/P
A
EMA/Pb (5
W
Vii
0
1
wtX)
IFgure4.
-,
squars
obtaind by
-&e
am (solid ofline).
two curves in the inset, represent a
subtraction
local structure that can be well described by the
liquid-like model (solid line). with R-nt=0-%
am, Rca=0.72 am, and Vp . 4 am3.
L 0
-0
I
lntheinset.thcttdline-thc
correlation function p(r)=r "Y(r) (Eq. 1) of the
ASAXS difference scattering pattern (Figure 3).
broad hump can be fitted by aDebye-Bueche
type equation fol random inhomogencity,
typetr)=exp.r/a+r/b'), with a-8&I am and b=3
,IIThe
0.04
'ii
6.1
r
tn )
o
241
range [19, 20, 281 and by applying correlation function analysis [29, 30] to the excess SAXS patterns
due to lead ions only as suggested by Williams (311 and demonstrated by an example In the following
section.
ANOMALOUS SMALL ANGLE X-RAY SCAITERING (ASAXS)
The ASAXS technique is one of the best methods for getting the pure scattering structure factor
of metal ions. Correlation function analysis of the scattering profiles is the most effective and leastbiased mean to demonstrate whether the ionic aggregates have inter-particle interferences [321. We are
here to present an example of using ASAXS to overcome the difficulties in determining an ionic
structure in the presence of polymer lamellar structure.
SAXS measurements on the EMA/Pb sample were conducted at, respectively, E= 13033 eV (7 eV
below the edge) and 12723 eV (317 eV below the edge). Each SAXS curve was accumulated for 200
sec. The energies were carefully chosen according to the anomalous scattering factors F and f "values
of lead in the ionomer samples which had been experimentally determined (281. The scattered Intensity
of lead ions could only be obtained by subtracting the SAXS pattern measureq at 13033 eV from that at
12723 eV after fluorescence correction (matching SAXS patterns at q> 10 nm" where the form factor of
lead ionic aggregates could be experimentally terminated). The experimental details will be described In
our forthcoming paper [28].
Figure 3 shows the ASAXS difference scattfring pattern for the EMA/Pb containing 5 wt% of lead
(hollow squares). A lamellar peak at q-0.6 nm" in the SAXS curves has been cancelled out by the
subtraction. The ASAXS difference pattern, possessing two typical features of ionomers: an ionic peak
and a small angle upturn, was due to lead ions only. One notes that the ionic peak terminates q> 10
nm
The electron density auto-correlation, -y(r) = <n (r )Y7(r )>,with ,7being the local electron density
fluctuations and r= r -r2 1, is related to the scattered intensitg 1(q) by [29, 30[,
2
j(r)= f q 1(q) sin qrdq/J q l(q)dq
a
qr
0
(1)
In the inset of Figure 4, the correlation function p(r) = r -(r) of the smoothed ASAXS difference
scattering pattern (Figyre 3) shows a broad hump which can be fitted with an empirical equation (solid
line): -y(r)
-exp(-r/a +r /b), with a =8.1 nm and b =36 nm. This equation is similar to well-known DebyeBueche equation for random inhomogeneties [29[, -y(r) - exp(-r/a), with a defined as the correlation
length. Fitting of the Debye-Bueche equation to the experimental correlation function yielded a =9.1 nm,
although the fitting was not as good as the empirical equation. The subtraction of the fitted (DebyeBueche like) correlation function from the experimental overall correlation function yielded a difference
curve as shown in Figure 4 (hollow squares). The damping oscillations seen in the difference correlation
function are typical in the correlation function of a system with inter-particle interference. The difference
correlation function could be well fitted to a liquid-like model [11]resulting in the followIng model
parameters: radius of the Ionic aggregates Rwt=0.36_+0.01 nm, ridius of closest approach
Rca= 0.72±0.02 nm,and volume per each ionic aggregate, Vp= 4.0+0.4 nm.
In Figure 3, the solid line and the dashed line are the structure factors corresponding to,
respectively, the correlation function of the liquid-like structure, as shown in Figure 4 and the correlation
function similar to Debye-Bueche's Inhomogeneity, as shown in the inset of Figure 4. The ASAXS
difference scattering pattern could be represented fairly well by the sum of the two types of structure
factors
Figures 3 and 4 reveal the following information about the lead ions. There are two structures in
the EMA/Pb with 5wt% of lead: a Debye-Bueche type Inhomogeneity with a correlation length of about
8-9 nm and a local structure that could be well described by the liquid-like model. The Ionic peak can be
attributed to the Inter-aggregate Interference while the small angle upturn Is due to the Inhomogeneous
spatial distribution of the lead Ions. The lead Ion aggregates are quite small In size with a mean diameter
of - 0.7 nm The most probable Inter-aggregate distance is -2.0 nm (second peak In Figure 4) slightly
greater than - 1.8 nm estimated from the Bragg spacing corresponding to the ionic peak position.
SUMMARY
By using the techniques of SAXS, ASAXS, WAXS, and DSC, the ionomer samples EMA/Pb and
EMA/PbS could be well characterized. The lead aggregates and PbS crystalline particles were present
in the amorphous region of EMA semcrystalline polymer matrix. The distribution of lead Ions was not
242
uniform while certain ordering was observed for PbS particles. The presence of the crosslinks formed
by the lead aggregates reduced the crystallinity of the polymer matrix. ASAXS was shown to be an
effective method of extracting the scattered Intensity of ions from complex scattering functions with
backgrounds including the crystalline matrix, possible impurity, air bubbles, and voids. The ASAXS
difference scattering pattern of the EMA/Pb sample containing 5 wt% of lead was analyzed by using the
correlation function scheme. The structure and the spatial arrangement of the Ionic aggregates could
thereby be described. The liquid-like model could be used to depict the Ionic peak with the local
structure of the ionic aggregates represented by the correlation function, while the Debye-Bueche type
long range inhomogeneity offered a good scheme to represent the small angle upturn.
ACKNOWLEDGEMENT
We are indebted to Dr. J.C. Phillips for his help in ASAXS experiments. B.C. gratefully
acknowledges support of this research by the U.S. Department of Energy (DEFG0286-ER45237A003).
The work was carried out at the SUNY Beamline supported by the U.S. Department of Energy
(DEFG0286-ER45231A003) at the National Synchrotron Light Source, BNL, which is sponsored by the
U.S. Department of Energy under contract (DE-AC02-76CH00016).
REFERENCES
[1]
RW. Rees and D.J. Vaughan, Polym. Prepr. Am. Chem. Soc. Div. Polym. Chem., §, 287 (1965).
(21
F.C. Wilson, R. Longworth and D.J. Vaughan, Poym. Prepr, Amer. Chem. Soc. Div. Polym.
Chem., 9. 505 (1968).
[3]
R. Longworth and D.J. Vaughan, Nature (London), 2M8, 85 (1968); and Polyn. Prepr., Amer.
Chem. Soc. Div. Polym. Chem., 2, 525 (1968).
[4]
R. Longworth, in Ionic Polymer, edited by L. Holliday (Halsted-Wiley, New York, 1975).
[5]
C.L. Max, D.F. Caulfield, and S.L Cooper, Macromolecules, §, 344 (1973).
[6]
T.R. Earnest, Jr. and W.J. MacKnight, Macomolecules, 10, 206 (1977) and J. Polym. Sci., Po/ym.
Phys. Ed., 16, 143 (1978)
[7]
W. J. MacKnight, W.P. Taggart, and AS. Stein, J. Polym. Sci., Polym. Syrup., 4, 113 (1974).
[81
J. Kao, A.S. Stein, W.J. MacKnight, W.P. Taggart, and G.S. Cargill, Macromolecules, 7, 95 (1974).
[9]
TR. Earnest, Jr., Ph.D Thesis, University of Massachusetts, (1978)
[10] W.J. MacKnight and T.R. Earnest. J. Polym. Sci: Macrom. Rev., 16§ 41(1961).
[111 D.J. Yarusso, S.L Cooper, Polymer, 26, 371 (1985) and Macromolecules, 16,1871 (1983).
[12] K. Tadano, E. Hirasawa, H. Yamamoto, and S. Yano, Macomolecules, 2.2, 226 (1989).
[131 W. Mahler, J. lnorg. Chem., ?7, 436 (1968).
[14] H. B. Stuhrmann, Adv. Polym. ScI., 7, 123 (1985).
[15] J.C. Phillips and K.O. Hodgson, In Synchrotron Radiation Research, edited by H. Winick and S.
Doniach, (Plenum, New York, 1980) and Acta Cryst., A&, 856 (1980).
[16] R,C. Mlake-Lye, S. Donlach, and K.O. Hodgson, Biophys. J., i. 287 (1983).
[17] P. Goudeau, A. Fontaine, A. Naudon, and C.E. Williams, J. Appl. Cryst., 1, 19 (1986).
[18] J.P. Simon, 0. Lyon, and D. de Fontaine, ,. Appi. Cryst., !a, 230 (1985).
[19] Y.S. Ding, S.R. Hubbard, K.O. Hodgson, R.A. Register, and S.L. Cooper, Macromolecules. 1,
1698 (1988).
[20] R.A. Register and S.L Cooper, to be published In Macromolecules.
[21] B. Chu, J.C. Phillips, D.Q. Wu, In Polymer Research at Synchrotron Radiation Sources, edited by
T.P. Aussel and A.N. Goland, report no. BNL51847, Brookhaven National Laboratory, Upton, NY (1986)
p.126 and J.C. Phillips, K.J. Baldwin, W.F. Lehnert, A.D. LeGrand, C.T. Prewiltt, Nucl. Instrum. and
Methods in Phys. Res., 6M, 182 (1986).
[221 B. Chu, D.O. Wu, and C. Wu, Rev. Sci. Instrum., (f, 1158 (1987).
"
3224 (1989).
[23] B. Chu, D.C. Wu, and R. Howard, Rev. Sci. Instrum.,
[24] M. Fujimura, T. Hashimoto, and H. Kawal, Macromolecules, 14, 1309 (1981) and a5, 136 (1982).
[25] B. Chu, D.O. Wu, W.J. MacKnight, C. Wu, J.C. Phillips, A. LeGrand, C.W. Lantman, and A.D.
Lundberg. Macromolecules, 1, 523 (1988).
[26] D.C. Wu, J.C. Phillips, A.D. Lundberg, W.J. MacKnight, and B. Chu, Macromolecules, 22, 992
(1989).
[27] H. Song, A.S. Stein, D.C. Wu, M. Re, J.C. Phillips, A. LeGrand, B. Chu, Macromolecules, 21,
1180 (1988) and H. Song, D.O. Wu, M. Satkowsk, A.S. Stein, and B. Chu, Macromolecule,, In press.
243
[281 D.O. Wu, B. Chu, and W. Mahler, to be submitted to Macromolecules.
[29] P. Debye and A.M. Bueche, J. App/. Phys., 20, 299 (1949) and H. Brumberger and P. Debye, J.
Phys. Chem., §1, 1623 (1957).
[30] G. Porod, in Small Angle X-ray Scattering, edited by 0. Glatter, 0. Kratky, (Academic Press,
London 1982)
[311 C.E. Williams, in Structure and Properties of lonomers, edited by M. Pined and A. Eisenberg, (D.
Reidel Publishing Co., Dordrecht 1987).
[32] D.O. Wu, B. Chu, R.D. Lundberg, and W.J. MacKnght, manuscript in preparation.
F
245
EXCIMER AND EXCITON FUSION OF BLENDS AND MOLECULARLY
DOPED POLYMERS--A NEW MORPHOLOGICAL TOOL.
ZHONG-YOU SHI, CHING-SHAN LI and RAOUL KOPELMAN,
Department of Chemistry, The University of Michigan, Ann Arbor, MI 48109-1055.
ABSTRACT
Exciton-exciton and exciton-excimer triplet fusion kinetics is monitored in
medium molecular weight P1VN/P1MA solvent cast films with concentrations from
0.005 to 100% (weight), at temperatures of 77 to 300 K, via time resolved
fluorescence and phosphorescence (10 nas to 10 sec). The heterogeneity exponent (h)
is 0.5 for isolated P1VN chains, zero (classical) for pure P1VN and "fractal-like"
throughout certain concentration regimes. However, h is not monotonic with blend
concentration but rather oscillates between zero and 0.5. Correlation is made with
morphology changes (phase separation, filamentation). 4s expected, the triplet
exciton kinetics is dominated by short-range hops (about 5 A) and thus monitors the
primary topology of the chains. At concentrations below 0.01%, the excitons are
constrained to a truly one-dimensional topology. At higher concentrations there is a
fractal-like topology. Similar studies were conducted on naphthalene-doped PMMA
(1-20% weight). The lower concentration samples are neither segregated nor
random solution phases.
INTRODUCTION
Can luminescence techniques establish themselves as important tools for
polymer morphology?
Industrial materials are becoming more complex,
structurally, as they are being optimized for specific performance criteria. This
complexity goes beyond the chemistry of copolymers and into the specific physical
arrangement of composite materials containing monomers, homopolymers, and
copolymers. This leads to heterogeneous mixtures containing microphase and
interphase domains. There is a general belief that the microscopic structure of the
material and its dynamic response are responsible for its desirable properties [11.
However, we know little about the morphology and dynamics at the 50 to 500 A
scale (typical polymer dimensions). Characterizing the structure and dynamics on
this scale might lead to : (1) a better understanding of materials performance, (2)
improved analytical methods for quality control, and 3) guidance for future
syntiesis of improved materials.
While spectroscopic methods have been around for a long time, the recent
advances in laser technology and in the theory of photophysical processes have led
to new approaches. Fluorescence and phosphorescence approaches, based largely
on energy transfer (ET) and quenching, have been of much recent interest[I-7.
Here we focus on a relatively new approach of monitoring longer-range energy
transport and energy fusion kinetics. This approach has been applied successfully
to simple organic crystals, mixed crystals, crystallites embedded in porous
membranes and glasses, and vapor-deposited organic films [8,9]. Most powerful has
been the utilization of triplet excitation transport and fusion, via delayed
fluorescence and phosphorescence. We apply this approach to the study of polymer
blend micro-morphology.
TRIPLET EXCITON TRANSPORT
Triplet excitation transfer (TET) is characterized by an extremely short geometrical
range, as short or shorter than electron transfer. The following table is adapted
from Tazuke and Winnik [2]:
Mat. Res. Soc. Symp. Proc. Vol. 171.
1990 Materials Research Society
246
Sinlet ET (dipole)
TET(electron exchange)
Electron Transfer
Exciplex Formation
Excimer Formation
Chem. Bond Formation
10 to 100
4 to 15
4 to 25
4 to 15
ca. 4
ca. 2 to 4
For chromophores such as naphthalene, TET is limited, effectively, to 8 A [8-10].
While excimer formation has a shorter range, its studies are usually contaminated
bY the much longer range of singlet ET, which leads to the formation of the excimer
[3,4]. Inaddition, the much longer lifetime of the excited triplet state (e.g., 3 sec. in
the naphthyl group, compared to 10-7 sec. for the excited singlet) makes it possible
to study multistep energy transfer, i.e., exciton transport kinetics.
IN
Fig. 1:
%.
Schematic description of
a
ea
and
polyvinylnaphthalene guest chains embedded in PMMA. The dashed
lines are the "short-cuts" possible for singlet ET but not for triplet ET.
Long-range triplet exciton transport can be monitored by trapping at
impurities or defects such as triplet excimers [8-10]. However, the most effective
method seems to be that of triplet exciton fusion, which can take place anywhere,
including at a trapping site. This fusion process is monitored not only by its effect
on the triplet phosphorescence, but also by its creation of delayed fluorescence, with
its hi.ly increased quantum efficiency and its unmistakable time characteristics.
Inadition,.fusion is a photophysical process which can chara terize the geometry
by which it is confined (see below), over a range of about 1000 A or larger, but with
a definition of about 5 A (Fig. 1).
FRACTAL-LIKE TRANSPORT AND KINETICS
Recent insights into transport and reaction dynamics in restricted geometries
and heterogeneous media have led to a "fractal" approach to diffusion controlled
247
kinetics [8].
Specifically, the rate coefficient for a bimolecular or pseudomonomolecular reaction can be written as
k = klt-h
0<h < 1
(1)
where h is the "heterogeneity exponent," and k1 is roughly the rate of a single hop.
For classical processes (homogeneous, 3-dimensional media) h = 0 and k is the
classical "rate constant." For a strictly one-dimensional topology h = 1i2. For most
other situations (effective topology between one- and three-dimensional),
0 < h < 12. Underlying this result is a non-Poissonian distribution of the reacting
particles [8], due to the fact that diffusion in low dimensions is an inefficient
stirring mechanism. (This is an expression of the non-recurrent nature of random
walk in low dimensions.) The exponent h is thus sensitive to the local morphology.
In addition, there are other short-time effects that are sensitive to the local
morphology [8].
We note that eq.(1) is appropriate for both bimolecular reactions (such as
-+Product) and pseudomonomolecular reactions (A + C -+ C + Product). In
particular we deal here with: 1. Triplet homofusion:
T + T -+ S
-* hv
(2)
where T is a triplet excitation, S a singlet excitation and hv a photon of fluorescence
(or a phonon). 2. Triplet heterofusion:
T + T - T + hv
(3)
where T' is a defect (excimer) excitation. Both equations (2) and (3) are somewhat
oversimplified [9] but are kinetically sound. In addition, there is the true
monomolecular decay:
T -+ hv'
(4)
where hv"is a photon of phosphorescence (or a phonon). The homofusion process is
thus described by
F = ki't-hP2
(5)
where F is delayed fluorescence and P phosphorescence, while the heterofusion
process is given by
F f ki"t-hP
(6)
where h is the same and depends on the medium but not on the process. We note
that proper time modulation of the laser pump and probing procedure will lead to
either homofusion or heterofusion.
EXPERIMENTAL
Both P1VN (Mw - 100,000) and PMMA (Mw - 154,000) used in this study
are purchased from Polyscience Inc. They are all further purified by multiple
precipitation (3 times). For PMMA we use toluene as the solvent and methanol as
the non-solvent, while for P1VN the solvent is dichloromethane and the non-solvent
is methanol. The solvents used in the purifications are all spectral grade. It is
believed that the purified polymers are free of monomers and low molecular weight
impurities from the polymerization. The final precipitates are left in the air to dry
for 1-2 days and then vacuum dried at about 70 OC (below their T. a) for another 2
days ensuring that there is no solvent left in the polymer samples.
248
Film samples are cast by using the doctor blade technique. Different portions
of polymers are carefully weighted. For the concentrations below 0.5%, we first
dissolve 5.0 mg of PIVN in 50 ml toluene and then by taking out 0.5, 1.0, 2.5, and
5.0 ml of this solution, get corresponding weights of P1VN equal to 0.05, 0.1, 0.25,
and 0.5 milligrams, respectively. These solutions are evaporated so that another
solvent can be used. By adding different amounts of PMMA, we get 0.005%, 0.01%,
0.05%, 0.1%, (Wt/Wt) compositions. For each gram of solid polymer 4 ml
dichloromethane is used as the solvent. The solutions are well stirred for several
hours in order to get homogeneous mixtures. The substrates on which the films are
cast are chosen to be aluminum plates. The casting surfaces are well washed and/or
rinsed with water, soap, ethyl alcohol and dichloromethane. Films are cast in a
nitrogen dry box. First, the well-stirred solutions are pored into the hollow of the
blade which is on the surface of the substrate. By sliding the blade towards the
other end a thin film is made through the groove of the blade. The newly cast films
are left in the nitrogen dry box for several hours and then further dried in the
vacuum oven at 70 *C for 2 days. The thickness of the finished film is about 20 $im.
The sample preparation of naphthalene-doped PMMA was described before [I1].
When experiments are performed at liquid nitrogen temperature (77K), the
samples are immersed in a liquid nitrogen bath inside a quartz cryostat. Excitation
light for this study is a Lambda Physics Excimer laser, with XeC1; it gives about a
10 nanosecond
pulse the
at the
wavelength
nimis and
line-withan
th is
several
angstroms.
Usually
output
power of of
the308
laser
muchthe
stronger
required;
tensity filters are used. For consistency, each set of experiments
isperformed at the same strength of excitation light.
For the kinetic studies the samples inside the cryostat are usually placed at
45* to the incident excitation and the emission signals are picked up by 2 UV grade
lenses, which focus the light on the entrance slit of a SPEX 0.5 M double
spectrometer. Inorder to pick up more signal and increase the signal-to-noise ratio,
both the entrance and exit slits are open wide (about 1000 microng or more). For
phosphorescence decay signals the monochrometer is set at 5250 A, while for the
delayed fluorescence the signals are collected at around 3400 A. These settings are
based on the steady state emission spectra of PIVN/PMMA. At the exit slit a
regular EMI PMT is attached. The samples with concentrations lower than 2%, due
to the fact that the decay signals are weak and the films are relatively clear, are
placed at 900 to the excitation light. The emission signals are collected at the back
of the sample (similarly to the way in which absorption spectrum is taken). Both
delayed fluorescence and phosphorescence decay emissions are first passed through
a CuSO solution (00g CuSO4 oSH2 0/liter) to filter out the residual laser light.
Then a 340 nm band pass filter is used before the PMT to collect the delayed
fluorescence. For phosphorescence decay a 520 nm band pass filter is used. The
current output from the PMT is fed into a Signal Averager (Princeton Applied
Research Mode! 4202). The termination resistor is selected to be small enough to
eliminate the intrinsic effect of the apparatus on the fast decay of the samples. A
Wavetek Model 187 function generator provides the external trigger. For the early
time decay data collections, the rising edge of an about 40 microseconds wide pulse
is used to trigger the excimer laser, while the falling edge of the pulse is used to
trigger the signal averaqer. The pulse-width is selected as a parameter to exclude
the laser residual which is reflected into the spectrometer for the higher
concentration samples. This setting together with the termination resistor and the
resolution of the signal averager enable us to detect nothing while a blank stainless
steel plate in used. Therefore, all the decay data reported here for the early time
kinetics
haveabout
a 40 microseconds
offset.
longer
are coll ected
10 milliseconds
afterFo
tf-ethe
laser
fires.time kinetics studies the data
249
15J&.
120FL
283.1
-25.09
g83.88
781.0
,x57&2
~375.4
172.56
-3M.22
.COBO
1.234
2.451
3.689
&.916
6.144
Tim(C Millise-ccd
Fig. 2 Delayed fluorescence (top) and phosphoremence
decay for 50%blend
250
MICROSECOND LUMINESCENCE OF BLENDS
While the laser pulse is only 10 nanoseconds wide, the data collection begins
after a delay of 40 microseconds. Cutting out shorter time responses eliminates: (1)
prompt fluorescence; (2) artifacts due to phototube saturation; (3) early transient
kinetic effects. The latter allows the formation of a stationary, kinetic
ordered
particle (free exciton) distribution [8]. Typical phosphorescence 8V delayed
fluorescence decays in this microsecond (and early millisecond) regime are shown in
Figure 2 for a 50% sample of PIVN/PMMA. We note that the decay of the delayed
fluorescence is extremely abrupt, compared to the phosphorescence. A doublelogarithmic representation of these data, Ln(F/P2 ) vs. Ln(t), following eq.(5), is
given in Figure 3. The curve has a definite negative slope (h = 0.30 ±0.04). Similar
plots are shown for the sample of 100% (Fig. 4). We see that the last sample gives
an essentially horizontal (classical) slope (h = 0.01±0.01).
MILLISECOND LUMINESCENCE OF BLENDS
In these experiments the data monitoring is started 10 ms after the exciting
pulse (10 ns). The rationale is the assumption that by this time most energy "traps'
(e.g., excimer forming sites) have been "filled" by the freely moving excitons, while
the free exciton population has been depleted by both fusion and trapping (and to a
lesser degree by natural decay). Double-logarithmic curves of Ln(F) vs. Ln(t),
following eq.(6), for these millisecond (and early second) decays are given for
samples of 0.01% (Fig. 5) and 50% (Fig. 6). We notice that the linear fits (on the LnLn scale) are good but have definite negative slopes. These results, as well as the
early-time (microsecond) decay time results, are plotted as a function of blend
concentration (Fig. 7). For comparison, we have added to this plot the older
(millisecond to second) results of Li and Kopelman [10], which were performed on
some of the same blend samples, but with xenon lamp excitation (lower intensity),
by first creating a steady-state excitation (illumination over about 10 sec) and then
cutting off the excitation and monitoring over a time interval of 10 millisecond to 10
sec.
DISCUSSION OF BLEND MORPHOLOGY
The last figure (Fig. 7) shows a consistent trend of heterogeneity exponents
(h) over four and a half decades of blend concentration. We first notice that we
obtain the classical (h f 0) result for the non-blended material (100% PIVN).
Whatever the exact morphology, it is obvious that here the exciton will be able to
jump onto neighborin naphthyl groups in all directions (three-dimensional
topology, on the average.
Going to the other extreme of highly diluted blends (0.005% and 0.01%), we
may expect isolated P1VN chains. Provided that these are not highly folded (into
"little balls"
micelle-like structures), the triplet-naphthyl excitons, whose jump. onlyorabout
5 A, will usually find themselves moving in a one-dimensional
topolog (consisting of a chain of several hundred monomers). Rememberin that
for a one-dimensional topology our theory predicts h = 1/2, our very dilute blend
data are fully consistent with a model of largely isolated, largely open P1VN chains.
Previous papers by Frank and collaborators [3,4] have shown that for
P2VN/PMMA (and similar) blends, there appeared to be a phase-separation
occurring at about 0.3% weight, for blends with orders of magnitude of guest and
host molecular weights similar to those in our experiments (but note that we use
P1VN, not P2VN). We thus expect some guest chain aggregation to start at about
0.1%. This is consistent with our decreasing h values (see Fig. 7). In addition, the
monomer/excimer fluorescence ratios (R) and the optical scattering data for our
PIVN/PMMA cast films (10] show a very similar concentration dependence to the
251
-5.772
-5.728
-6.287
.
-6.155
% -6.5.-,
•
.
".
-.
M &
......-
*,82..,
".
..
,
.
-6.601
. **'
".
.
.-.
,..
*..
......
-7.009-j
*.
7.436
-7.53
-&50T -2j77 -2.06? -1.347
-..
-.6265 -.0934
-1.001
.-.
0555 -
a
I
-2.77 -2.06? -1.347 -.4265 -D4
"3.50?
..
-T.39
-343
-3.170
-2.556 -1.901 -1.267 -.6324 -.0019
-.6916
,
1553A-k
4
-W
.
.
.
,
-1.267 -6524 -.0019
Ln [Time (millisecond)]
Fig. 4: 100% Blend (homofusion).
Fig. 8: 50% Blend (homofusion).
.
IfFl
V
-1.002
-3.170 -2.536 -1.9"
Ln [Time (millisecond)]
• .0403
,I
,-.1143;
.3955
-4541
-6767
.
-I473
-.9579
-1.976
-2402
-:.239[
-l.520
-&298 -4699 -4.100 -3.504 -902
.
"%9'65
.
.
.
-2.303
.2/"
r,'
-01o51" ,
M! -256
.
-4.703 -3.884 -3.065 -2.245 -1.427
-5.521
.
98
.
.
.
.
.
."...
0 .. .'".'
IB,,
-o
.
.
.
________19
-5298 -4699
-4100 -&50t -2.902 -1303
Ln (Time (second)]
Fig. 5: 0.01% Blend (heterofuion).
-5&21
-4703 -884
-3065 -2.245 -. 427
Ln [Time (second)]
Fig. 6: 50% Blend (heterofusion),
252
Frank et al. [3,4] P2VN/PMMA cast films (with similar molecular weights). Based
on the last two criteria one would assume that when the PVN guest concentration
reaches about 10% weight, a fairly complete phase separation should occur.
Furthermore, based on the fact that at 10% and higher no monomer" fluorescence
is observed (only excimer fluorescence), one wouldconclude that the PMMA rich
phase contains little PVN.
Accepting the above picture, it is natural to expect the heterogeneity
exponent h to decrease monotonically with increasing PVN concentration, and
approach zero around about 10%. However, the opposite is observed (Fig. 7). A
non-monotonic behavior is seen. This results in an h increase, reaching a maximum
(close to 0.5) at about 10% PVN, then falling slowly and reaching h = 0 (the classical
value) only at concentrations approaching 100% (but definitely not at 50%). We
note that DSC measurements at 50% show a clear two-phase pattern only, after
partial annealing of the films [10].
It has been recognized earlier [3] that kinetic restrictions during solvent
casting could affect the attainment of equilibrium. In particular, the P2VN/PMMA
blends appeared to be more miscible than predicted from equilibrium models
(especially for higher molecular weights of PMMA). For simplicity, we assume here
a "hairy" interface model, where the different phases are separated by a "fuzz
interphase domain, i.e., PVN chain that penetrate significantly into the PMI
phases (and, possibly, vice versa). On the short-range scale typical of the triplettransfer, these hairs have an effectively one-dimensional topology, while on the
The observed
longer range singlet-transfer scale this is no longer true.
excimer/monomer ratios (R) appear to be affected by monomer-monomer singlet
energy transfer. However, the triplet-triplet fusion is dominated by the short-range
monomer-monomer triplet energy transfer. The triplet fusion kinetics are thus
more sensitive to the topological details of the interface regions. A similar
sensitivity of this method to domain interface topology has been demonstrated for
vapor-deposited crystalline naphthalene films [9]. We also note that the fusion rate
in the core regions (with h = 0) is faster than that at the interface regions (h > 0), so
Our
that the observed decays may be biased by the interfacial domains.
heterogeneity exponent h is thus very sensitive to local heterogeneities and less
sensitive to local homogeneities. On the other hand, a sample with macroscopic
heterogeneities consisting of separated domains with sharp, rather than fuzzy,
interfaces, would have resulted in local environments that are mostly homogeneous
and thus would have given near-zero values for the heterogeneity exponent (h).
MOLECULAR DOPED POLYMERS (MDP)
The naphthalene doped PMMA films begin to scatter light at about 20%
concentration (by weight). It is generally agreed [11] that at this concentration
Under these
there are segregated, micron-sized, naphthalene crystallites.
conditions we expect to see classical behavior, i.e., h = 0. This is indeed observed:
Figure 8 shows the naphthalene concentration dependence of the heterogeneity
exponent h. Below 20% there is a monotonic increase as the concentration
decreases. This excludes the total segregation model, for which h = 0. The
naphthalene aggregates must be smaller than 1000 A (at least in some directions)
to give h > 0, based on an exciton mean free path of about 500 A [8]. However, the
aggregates also do not appear to be random (percolation) clusters. For such random
on the order of 10% or higher, based
clusters the percolation concentration would
on mixed crystal work [8]. At such a percolation concentration we expect h - 1/3,
approximately. However, in the MDP samples, the value of h = 1/3 is reached only
at about 1%naphthalene. Whatever form of aggregation that is responsible for this
h = 0.3 value cannot be a random aggregation (under random aggregation we would
have mostly monomers at 1%). We conclude that the dilute MDP has non-random
guest aggregates, with sizes less than 1000 A, at concentrations below 20%.
253
0.01
lop
1.0
0.1
100 (%
.4= -
.1000-
•
I I
' ,
.mO_
-3.00
-2.000
,
-1.000
I
.0000
'
x
1.000
2.000
Log CC)
--X
Late time decay with pulse excitationeterofueoln
rly tim decay with pulse excitatianNomfusion
0 - Sted state excitation with Xnan lapIHeteofusion
-
-
Fi. 7: Heterogeneity Exponent v. Blend Concentration.
254
z..-I
.I
\
[aE~m\
.AnD
&.0
10.L
ILSLOB
.
I0.0
Fig. 8: Heterogeneity Exponent vs. MDP Concentration.
ACKNOWLEDGMENT
This research was supported in part by NSF Grant No. DMR 8801120 and in
by the Petroleum Research Fund administered by the American Chemical
Society Grant No. 18791-AC5,6.
REFERENCES
1.
2.
3.
4.
M. A. Winnik, in Pbotonhysial and Photochemical Tools in Polymer Science
(NATO ASI Series C182), edited by M. A. Winnik (Reidel, Dordrecht, 1986),
p. 611.
S. Tazuke and M. A. Winnik, in Photohs
A d and Plhtochemical Tools in
*
(NATO
ASI
Series
C182),
edited by M. A. Winnik (Reidel,
Dordrecht, 19M6), p. 15.
C. W. Frank et al., in PhotonhMcland hotochemial Tools in Polymer
fdom (NATO ASI Series C182), edited by M. & Winnik (Reidel,
Dordrecht, 1986), p. 523.
C. W. Frank and Gelles, in Phota~hywcal and Photochemical Tools in
S
(NATO ASI Series C182), edited by M. A. Winnik (Reidel,
Dordrecht, 1986), p. 561.
5.
6.
7.
H. Morawetz, Science 24, 172 (1988).
K A. Peterson et al., Macromolecules 20, 168 (1987).
N. Kim and S. E. Webber, Macromolecules 13. 1233 (1980).
8.
R. Kopetman, Science 241, 1620 (1988).
9.
10.
11.
L A. Harmon and K. Kopelman, J. Phys. Chem. (in press).
C. S. Li and K. Kopelman, Macromolecules (in press).
E. I. Newhouse and R. Kopelman. Chemi Phys. Lett. 143. 106 (1988).
255
THE ORDERED BICONTINUOUS DOUBLE DIAMOND
STRUCTURE IN BINARY BLENDS OF
DIBLOCK COPOLYMER AND HOMOPOLYMER.
Karen I. Winey* and Edwin L. Thomas*
* University of Massachusetts, Polymer Science and Engineering
Department, Amherst, MA 01003
** Massachusetts Institute of Technology, Materials Science and
Engineering Department, Cambridge, MA 02139
3We
ABSTRACT
report the observation of the ordered bicontinuous double diamond
(OBDD) structure in binary blends of poly(styrene-isoprene) diblock copolymer
and homopolystyrene. The overall polystyrene volume fraction range is 64 67 PSvol% for the OBDD structure in binary blends of a lamellar diblock
(SI 27/22) and a homopolymer (14.0 hPS). This composition range is
approximately within the polystyrene volume fraction range established for pure
diblock copolymers in the strong segregation regime having the OBDD
structure. Ordered lamellse are observed at approximately 65 PSvol% when the
homopolystyrene molecular weight is greater than the molecular weight of the
polystyrene block of the copolymer. This observation is discussed in terms of the
decreased degree of mixing between the homopolymer and the corresponding
block and the resultant effect on the interfacial curvature.
INTRODUCTION
Diblock copolymers contain two polymer chains which are covalently
bonded to one another at one end. Immiscible polymer blends macrophase
separate, whereas the connectivity within diblock copolymers limits the size of
the phase separated domains and gives rise to periodic microphase separated
morphologies. In the strong segregation limit the volume percent of the
components in the diblock copolymer determines which morphology is observed:
spheres on a body-centered cubic lattice, cylinders on a hexagonal lattice, the
ordered bicontinuous double diamond (OBDD) morphology, or lamellae. The
polymer chains are amorphous within these ordered domains.
The ordered bicontinuous double diamond microstructure is periodic in
three dimensions and bicontinuous in that both the majority and minority
components are continuous throughout the microstructure. The minority
component is divided into two interpenetrating networks having diamond cubic
symmetry. A color computer-generated image will assist the reader in
visualizing this microstructure1 . Both minority channels are separated from
the majority component by a surface of approximately constant mean curvature.
Similar bicontinuous structures have been proposed for surfactant-oil-water
systems 2 . The OBDD microstructure has been observed previously in linear
diblock copolymers and star diblock copolymers 3 . This paper will discuss a new
method of producing the OBDD microstructure: binary blends of diblock
copolymer and homopolymer.
Mat. MO.O.
Soc. Symp. Proc. Vol. 171.
1
MaeOdWASI,
Remrch Society
256
EXPERIMENTAL
We limit our discussion to diblock copolymers which by themselves exhibit
the lamellar morphology. The diblock copolymers used in this study were
anionically synthesized by Dr. L. J. Fetters of Exxon. Poly(styrene-isoprene) and
poly(styrene-butadiene) diblock copolymers are designated as SI and SB,
respectively, followed by nominal block molecular weights given in kg/mol: SI
27/22 and SB 20/20. The polydispersities of the polystyrene block and the
copolymer are less than 1.05. The homopolystyrenes from Pressure Chemical
were characterized by Dr. Fetters and are designated as x bPS, where x is the
molecular weight given in kg/mol. The homopolystyrene molecular weights
range from 5.9 to 30.1 kg/mol and have a polydispersity of less than 1.08. The
copolymer content of the blends examined ranges from 60 wt% to 80 wt%
copolymer.
Our experimental procedure is more fully discussed in a previous paper 4 .
Briefly, we prepared dilute solutions of the blend in toluene, a nonselective
solvent. Bulk samples were made by allowing the solvent to evaporate slowly at
room temperature to form 1 mm thick films. The samples are then annealed at
125'C for one week and quenched to room temperature in liquid nitrogen. This
experimental procedure is designed to obtain near-equilibrium conditions
characteristic of 125'C which is in the strong segregation limit for this system.
Thin sections for transmission electron microscopy were cryomicrotomed at
approximately -110'C and stained with aqueous Os04 vapors.
RESULTS AND DISCUSSION
Figure 1 exhibits an example of the ordered bicontinuous double diamond
microstructure in the blend of 60.0 wt% SB 20/20 in 17.2 hPS. The micrograph
exhibits the [1111 projection of the OBDD as confirmed by computer simulations
compared to digitized TEM data5 . The [1111 projection has been called the
"wagon wheel" projection, since six spokes are seen to protrude from the axle, if
one imagines the axle of a wheel perpendicular to the page; the spokes are light
and the axle is dark at this composition. The added homopolystyrene selectively
swells the polystyrene block of the copolymer to induce an order-order transition
from lamellae to the ordered bicontinuous double diamond morphology. The
homopolystyrene and the PS blocks are contained in the matrix region of the
OBDD microstructure, while the PB blocks are in the channels.
There are four experimental parameters necessary to describe binary
blends of homopolymer (A) and diblock copolymer (AB) at a fixed temperature.
The AB diblock copolymer is fully described by two parameters, for instance, the
molecular weight of the A block and the A composition of the copolymer. The
copolymer content in the blend will be given as the weight percent of the
copolymer. Finally, the homopolymer molecular weight must be specified to
fully define these binary blends. We have explored the effect of the copolymer
content in the blend and the homopolymer molecular weight.
Figure 2 shows a copolymer content series for the blend system of S1 27/22
and 14.0 hPS. At the lowest copolymer weight percent shown (64.0 wt%),
polyisoprene cylinders are observed on a hexagonal lattice. The dark regions
indicate the preferentially stained polyisoprene domains. The intermediate
concentrations of 68.0 wt% and 70.0 wt% copolymer exhibit the OBDD structure
with PI diamond channels, while a higher copolymer content (71.8 wt%)
257
100
nm
Figure 1: Transmission electron micrograph of 60.0 wt% SB 20/20 in 17.2 hPS
exhibits the ordered bicontinuous double diamond microstructure.
4N
Figure 2: Copolymer content series of the blend system SI 27/22 and 14.0 hPS:
a.) 64.0 wt% copolymer, cylinders on a hexagonal lattice; b.) 68.0 wt% copolymer,
OBDD structure; c.) 70.0 wt% copolymer, OBDD; d.) 71.8 wt% copolymer,
lamellae.
produces ordered lamellae. Thus, increasing the copolymer content of this
blend system induces two transitions from cylinders to OBDD and from OBDD to
lamellae.
For the purpose of comparison with pure linear diblock copolymers the
copolymer content of the blend was converted to the overall polystyrene volume
percent ( 0 pS). The value of OPS increases as the copolymer content in the blends
decreases. Figure 3 illustrates the morphology observed as function of Ops for
both the binary blend and the linear diblock copolymer. The squares indicate
blends which have been prepared and their observed morphologies are
indicated. The same sequence of morphologies are observed for both the diblock
258
Overall PS Volume Percent
Binary Blends
of SI 27/22
nd14.0PS
Ordered
Lamellae
OB
DD
Hexagonal
Cylinders
Pure
SI 27/22
50.0
inear
SIDiblocks
55.0
Ordered
Lamellae
60.0
65.0
OBDD
70.0
75.0
Hexagonal
Cylinders
Figure 3: Observed morphologies as a function of overall polystyrene volume
percent for linear SI diblock copolymers and binary blends [squares] of SI 27/22
and 14.0 hPS.
250 nm
Figure 4: Homopolymer molecular weight series for 70.0 wt% SI 27/22 and
various hPS: a.) 5.9 hPS, OBDD structure; b.) 14.0 hPS, OBDD structure; c.) 30.1
hPS, lamellae.
259
copolymer and the binary blend as 4 pS increases: lamellae, the OBDD structure,
and cylinders. The overall polystyrene volume percent windows for the OBDD
microstructure are similar for the copolymer and the copolymer/homopolymer
blend: approximately 62 - 66 vol% for the pure SI diblock copolymer 6 and
approximately 64 - 67 vol% for the binary blend of SI 27/22 and 14.0 hPS. One
might expect a biphasic region between the various ordered morphologies in the
binary blend indicative of a first order phase transition. We have not observed
such biphasic regions at this time.
Figure 4 shows a homopolymer molecular weight series in which the
overall polystyrene volume percent is constant at - 65 vol%. The binary blends of
70.0 wt% SI 27/22 and either 5.9 hPS or 14.0 hPS exhibit the OBDD
microstructure. Increasing the homopolymer molecular weight to 30.1 hPS
results in a lamellar morpbology.
Leibler, Orland and Wheeler developed a diblock copolymer/homopolymer
blend theory for the case of dilute, spherical micelles in a matrix of
homopolymer 7 . Their free energy expression for a micelle includes the entropy
of mixing in the corona region between the homopolymer and the corresponding
block of the copolymer. This term is assumed to be inversely proportional to the
degree of polymerization of the homopolymer. Kinning experimentally
confirmed this trend for dilute, spherical micelles in blends of poly(styrenebutadiene) and homopolystyrene 4 .
The degree of mixing within the corona influences the size of the corona
and the mean curvature of the core/corona interface. We have previously
reported a shape transition in a blend of 12.5 wt% SB 40/40 and hPS as a function
of the homopolymer molecular weight. Blends made with 2.9 hPS and 7.4 hPS
exhibit disordered spherical micelles, while a blend with 17.0 hPS exhibits
disordered cylindrical micelles 8 . This shape change from spherical to
cylindrical disordered micelles exhibits decreasing interfacial mean curvature
as the degree of mixing in the swollen corona region decreases by increasing the
homopolymer molecular weight. The experimental results discussed here are
at considerably higher copolymer content, so that the corona-corona interaction
induces long range order. At a low degree of mixing of the hPS into the PS
block, the PS-PI interface in the SI/hPS blend maintains the flat, zero curvature
lamellae of the pure copolymer, Figure 4c. A highly swollen PS block (via the
addition of lower molecular weight homopolystyrene) causes the PS-PI interface
to exhibit non-zero mean curvature, Figure 4a and 4b. The degree of mixing in
the corona controls the interfacial curvature in such a way to obtain either
lamellae or the OBDD microstructure at the same overall polystyrene
composition.
Although not directly applicable to our present work, Wang and Safran
have calculated the phase behavior of microemulsion systems of diblock
copolymer (AB) separating A and B homopolymers 9 . Their model assumes no
interaction between diblock interfaces (low copolymer content in the ternary
blend) and no penetration of the homopolymers into the copolymer interface
("dry brush"). The transitions between cylinders, lamellae and bicontinuous
structures were found by minimizing the curvature elastic free energy of the
interface which is defined by the following coefficients: K as the bending elastic
modulus, K as the saddle-splay elastic modulus, and co as the spontaneous
curvature. K was found to be always negative indicating that saddle shaped
deformation is energetically unfavorable. In the Wang and Safran model, the
OBDD structure, which exhibits saddle deformation, transforms to lamellae or
cylinders, which do not exhibit saddle deformation as the magnitude of X
increases. The OBDD structure in our experimental binary blends became
lamellae as the homopolymer molecular weight or the copolymer content
increases and became cylinders as the copolymer content decreases.
CONCLUSIONS
The ordered bicontinuous double diamond microstructure can be
prepared in blends of diblock copolymer and homopolymer. The
homopolystyrene selectively swells the polystyrene block of the SI copolymer to
induce a' transition from lamellae to the OBDD morphology. The overall
polystyrene composition in OBDD blends is approximately 64 - 67 PSvol% in the
case of SI 27/22 and 14.0 hPS; this composition is comparable to that of the OBDD
structure in pure SI diblock copolymers. Finally, the ordered lamellae
morphology is observed rather than the OBDD structure when the homopolymer
molecular weight is too large which can be discussed in terms of preferred
interfacial curvature.
ACKNOWLEDGEMENTS
The National Science Foundation supported this work with a Graduate
Fellowship to K. I. W. and Grant No. DMR 89-07433 (Polymers Program). We
also thank the NSF Materials Research Laboratory at the University of
Massachusetts for facilities. The authors thank Dr. Lewis J. Fetters of Exxon
Corporation for synthesizing the diblock copolymers used in this work.
REFERENCES
1. E. L. Thomas, D. M. Anderson, C. S. Henkee and D. Hoffman, Nature 334
(6184), 598(1988).
2. S. M. Gruner, et al., Biochemistry 27 (8), 2853 (1988).
3. D. B. Alward, D. J. Kinning, E. L. Thomas and L. J. Fetters, Macromolecules
19 (1), 215 (1986); E. L. Thomas, D. B. Alward, D. J. Kinning, D. C. Martin,
D. L. Handlin and L. J. Fetters, ibid., 12 (8), 2197 (1986); II. Hasegawa, H.
Tanaka, K. Yamasaki and T. Hashimoto, ibid., 2& (7), 1651 (1987).
4. D. J. Kinning, E. L. Thomas and L. J. Fetters, J. Chem. Phys. 90 (10), 5806
(1989).
5. D. M. Anderson and E. L. Thomas, Macromolecules 21 (11), 3230 (1988).
6. D. A. Gobran, PhD thesis, University of Massachusetts, forthcoming.
7. L. Leibler, H. Orland and J. C. Wheeler, J. Chem. Phys. 79 (7), 3550 (1983).
8. D. J. Kinning, K. I. Winey and E. L. Thomas, Macromolecules 21 (12), 3502
(1988).
9. Z.-G. Wang and S. A. Safran, J. de Physique, in press; Europhys. Lett., in
press.
261
STUDIES ON THE EXCESS FREE ENERGY AND THE EARLY STATE OF
SPINODAL DECOMPOSITION OF THE BLEND d-PS/PVME AND THE
ISOTOPIC BLEND d-PS/PS WITH SMALL ANGLE NEUTRON SCATTERING
D. SCHWAHN, T. SPRINGER, K. HAHN*, AND J. STREIB*
Institut fiir Festk6rperforschung der Kernforschungsanlage Jiilich GmbH,
Postfach 1913, D-5170 Jiilich, Federal Republic of Germany
*BASF Aktiengesellschaft,
D-6700 Ludwigshafen, Federal Republic of Germany
INTRODUCTION
The article deals with the phase diagram and spinodal decomposition of two
polymers blends, namely d-PS/PVME and d(deutero)-PS/PS, investigated by small
angle neutron scattering (SANS). The result of the static experiments is the excess free
energy and the phase diagram. This is used as a basis for studies of non-equilibrium
phenomena as spinodal decomposition. In polymer blends the Cahn-Hilliard-Cook
theory of the early state of spinodal decomposition can be tested easily, because the
blends
have rather low relaxation rates; and they are meanfield systems [2,6] which
makes interpretation simple, except in a very narrow temperature region near the
critical point [6]. The kinetics in the isotopic blend d-PS/PS are so slow that the
early states of spinodal decomposition can be studied within minutes. The presented
SANS results have been performed at the KWS I small angle instrument at the FRJ-2
reactorin the KFA Jiilch.
DETERMINATION
OF THE
EXCESS FREE ENERGY WITH
CRITICAL
SCATTERING
The excess free energy of a mixture can be determined from composition
fluctuations in the mixed state which is studied by SANS. The static structure factor
S(Q) for critical fluctuations of a blend in the Zimm approximation is given by [1,2]
1
2
1
Q is the scattering vector for the neutron, g the Gibbs free energy of mixing in units of
kT, 0 the concentration of one component, Rg is the radius of gyration, and V. the
weight average of the molecular volume, assuming the same volume for both
components. The second derivative of g with respect to the composition 0 is calculated
from the Flory-Huggins model of polymer blend [1,2], namely
"W
r .
(2)
r is the generalized Flory-Huggins parameter (see eq. (4)) which describes the
difference of the segmental interaction energy (or enthalpy). It depends only on the
segmental properties of the molecular chains. r will be obtained from the scattering
experiment.
In Fig. 1 results of critical scattering at different temperatures for a d-PS/PS
2
3
=0.48 and V,, = 0.91- 10 cm /mol are plotted vs. Q (Zimm
representation). The fitted straight lines show that the experimental data follow the
Zimm approximation eq. (i) below Q = 5. 10-3 -. The extrapolated value of
S(Q-0) gives 02g0 2 and r by eq. (2). Similar experiments were performed for other
compositions (see Ref. [3]).
mixture with
Mat. Res. Soc. Symp. Proc. Vol. 171. '19"0 Materials Research Society
3I
I
i
I
I
d-PSIPS
0 0.48
T=205
F=
0
0
T= 185t
T=166 0 C
___
2
Vo1
Ts
2
3
4,
5
Q [10- 5V ]
Fig. 1
Critical scattering of the d-PS/PS mixture at T = 166 OC, 185 0C and
205 0C. The two dashed lines indicate the scattering curves at the spinodal,
and at the compensation temperature, respectively.
The experiments show that r can be written as
r =
- ro
(3)
with an enthalpic (rh) and an entropic (ra) term, which both depend on 0 [3].
Because the SANS experiments determine
Og/O2, 1(o) is related to the usual
Flory-Huggins parameter by
P= X-(1 - 20)0-7 0( - O)W
and X can only be calculated from r if the 0 dependence of
(4)
x is known [3].
In Fig. 2 the generalized Flory-Huggins parameter r of d-PS/PS (0 = 0.5) and
d-PS/PVME (0 = 0.2) is plotted as a function of the inverse temperature, using
SANS results [3,6,7 (see also Ref. [4,5]). The spinodal temperature is defined by
02g/0 = 0 (eq.(2.
Especially, the evaluated T, of two molecular volumes are
shown in Fig. 2. The slope of F, which is given by the enthalpy term rh in r, is
positive for d-PS/PS and negative for d-PS PVME and accounts for an upper or for a
lower critical solution temperature, respectively. The absolute value of Fh for
d-PS/PVME is two orders of magnitude larger than for the isotopic mixture. The
spinodal lines for the investigated d-PS/PVME and d-PS/PS mixtures are presented
ii Fig 3 and 5, respectively.
)is
2' (V =2 "105 cm3/mot)
T= 140C
-10--V/
Tc :60°C
-
o d-PS/PVME
E
10
5-
Fig. 2
factor
msi
Excess
0
3
6
/ moL)
CM
2(VW =10
dPIP
eeenergy omin
) in
T
(s
parameter)as a function of
thalse
pt a temperatur e
()Qs teh
=
i
Tyisdineo
for the case of an upper or a
are responsible
or negative
Positive
solutionvalues
temperature,
respectively (see Fig. 3 and 5), as
lower critical
obtained from eq. (2) for Mg/02 = 0.
i
EARLY STATE OF SPINODAL DECOMPOSITION
~A
'q
blend is rapidly quenched from an equilibrium state at To with the structure
factor STo(Q) to a final state kept at a temperature T with ST(Q). If T is in the
miscibility gap this state is unstable and the corresponding S(Q) is the virtual
structure factor [2]. The Cahn-Hilliard--Cook (C-H-C) [g,I0] theory extended by
Binder [2] for polymers leads to the time dependent structure factor
ST(Q,t) = ST(Q) + [STo(Q) - ST(Q) exp[- 2r- t]
.
(5)
The relaxation rate r- is calculated with a nonlocal Onsager coefficient because of the
large size of the molecules, namely for Q < i/Rg
r-'(Q) _ 6 TR-1 Rg2 Q2 {
r ( +R
)
(6)
where 7 R is the characteristic time of molecular diffusion. The relaxation rate can be
connected with the collective diffusion constant
Doiil = 1/r Q2 for Q - 0.
(7)
264
SPINODAL DECOMPOSITION IN d-PS/PVME
The spinodat of a d-PSIPVME mixture with the molecular volumes of
= 227000 cm3/mol and V,(PVME) = 53000 CM31 IMol is plotted in Fig. 3.
The spinodal
has been calculated from r in Fig. 2 by means of eq. (2), using
2
2g/O
=0. The spinodal temperature of the investigated specimen (0 0.2) is
150 OC. The system was heated from the mixed (homogeneous) state at 149 OC up to
152 OC, about 2 K above Tc. The SANS measurement was immediately started when
the temperature was reached; and took 30 sec. From then on the measurements were
performed consecutively in periods of 30 sec.
V.(d-PS)
170
1600 -
spinodaL
150-
04
PVME
Fig. 3
0.2
0.4
0.6
'0. 8
Spiniodal, of d-PS/PVME mixture.
V.(PVME) = 53000 cm3 /mol).
V,(d-PS)
1.0
d-PS
=227000
cm3/mol
and
0.20
PSD 10.2 -PVMEI 096
0.15
0T:1S2*C
0,10
00
& 0.050-0
0
Fig. 4
1
2
a uio31
3
4
j
Static structure factor at 149 OC C--)and structure factor -o-- measured
at 15 OG during the first 30 sec. of the decomposition process.
26$
Fig. 4 shows the structure factors in the homogeneous region at T = 149 OC and after
heating the blend into the -region of the miscibility gap at 152 DC (see arrow in Fig. 2).
The peak at Q, = 1.6 • 10 3 A' is expected to correspond to the maximum growth rate
i.e. the maximum of (- r)- in eq. (8). Using the equilibrium data, the calculated value
of
Q. = [1.5/Ri2 (fr/s - 1)1 1/2
(8)
agrees roughly with the measured one. At later times a shift of the peak position to
smaller Q is observed due tp coarsening which is not described by eq. (5). Peak
positions as small as 5 • 10-5 - have been observed with the neutron double crystal
diffractometer 17,8]. This is the first measurement of the early state of decomposition
in this system. More experiments on this system are in preparation.
SPINODAL DECOMPOSITION IN d-PS/PS
The spinodal and the binodal of an isotopic d-PS/PS mixture (with
V, = 0.91. 106 cm3/mol) were calculated with F from the SANS experiments by
eq. (2) in Fig. 5. We have investigated a 0 = 0.48 mixture with a TS = (135 * 5) OC
Fig. 6 shows the SANS results of demixing. It started from the homogeneous
equilibrium state at 160 0C and stopped after 3 weeks at 125 0C. All SANS
measurements were performed at room temperature after a rapid quench below
TG = 100 0C. The data could be analyzed with eq. (5) and the solid lines in Fig. 6
shows
the
best
fit.
We
find
a
collective
diffusion
constant
of
Dcoll= - 1.4 - 10- 8 cm 2 /s. This system is in an earlier state of decomposition as.it
was found in d-PS/PVME. In this region the observed peak at 1.4. 10-3 A-'
approaches the value expected from eq. (8) namely Q. = 7.6 . 10 4 A. Before
coarsening occurs there is a change of the relaxation behaviour due to the presence of
the glass transition [12]. The relaxation rate given by diffusion (eq. (6)) couples to
additional and slow structural changes. This effect was observed in this system in a
relaxation experiment where the initial and final temperatures were both in the
homogeneous region. Only in the limit of small Q and small annealing time the system
could be described in terms of C-H-C theory [3].
200
150-
-
Spinodal
TG
Bincdol
100-
0
PS
Fig. 5
0.2
0.4
0.6
0
0.8
1.0
d-PS
Phase diagram of d-PS/PS with an average Vw = 0.91 • 106 cm3/mol.
1.2-
%
l
V
d-PSIPS
4=0.48
0 1%
0.8t
T =125 0C
= 21.4days
0.4 - T-16 0°C/
0
2
4
6
8
10
a (10-3 -1)
Fig. 6
Static structure factor at 160 DC and S(Q,t) after 3 week at 125 0C. The
dashed line is the virtual structure factor at 125 0 C.
CONCLUSIONS
1) By means of SANS experiments we have determined the excess free energy
(Flory-Huggins parameter) of the d-PS/PVME and d-PS/PS and calculated the
phase diagrams with this quantity.
2) Kinetic experiments were carried out by quenching from the homogeneous to
the unstable state. In d-PS/PVME the peak position consistent with C-H theory was
observed. The pubsequent coarsening with a shift of the peak position from 10-3 to
about 5 • 10-5 A-1 is found in Ref. [7,8]. For d-PS/PS the very early stage could be
measured which occur before the appearance of the unstable peak at Qm.
REFERENCES
1]
2]
P.G. de Gennes, Scaling Concepts tn Polymer Physics (Cornell Univ. Press,
Ithaca 1979)
K. Binder, J. Chem. Phys. 79, 6387 (1983)
D. Schwahn, K. Hahn, J. Streib, and T. Springer, submitted to publication
F.S. Bates and G.D. Wignall, Macromolecules 19, 932 (1986)
P.F. Green and B.L. Dayle, Phys. Rev. Lett. 57, 2407 (1986)
D. Schwahn, K. Mortensen, T. Springer, H. Yee-Madeira, and R. Thomas,
J. Chem. Phys. 87, 6978 (1987)
e-Mdra, JUL-Report Nr. 2268 (1989)
7]H.
D. Schwahn and H. Yee-Madeira, Colloid & Polymer Sci. 265, 867 (1987)
9
10
II
J.W. Cahn, Acta Metal]. 9, 795 (1961)
H.E. Cook, Acta Metall. 18, 297 (1970)
G.R. Strobl, Macromolecules 18, 558 (1985)
J. Jackle and M. Pieroth, Z. Phys. B
, 25 (1988)
267
MECHANICAL PROPERTIES AND STRUCTURE OF MELAMINE
FORMALDEHYDE/POLY(VINYL ALCOHOL) MOLECULAR COMPOSITES
KECHENG GONG AND XINGHUA ZHANG
Dept. of Polymer Science & Tech., South China Univ. of Tech.,
Guangzhou, P.R. China
ABSTRACT
The mechanical properties and structure of melamine formaldehyde (MF)/poly(vinyl alcohol) (PVA) composites were studied in
this paper. When PVA content was less than a certain value(about
20 weight %), both flexible strength and impact resistance were
improved obviously. While the impact resistacce improvement remained the previous trend the flexible strength didn't increase
so rapidly as before when the PVA content was more than 20%.
The morphology and reactivity of the prepolymer powder and
the morphology of the finished specimens were investigated by
means of microscope, infrared spectrum and transmission electron
microscope(TEM). The results indicated that molecular composite
structure was formed in this two-component system. The well-distributed PVA in this system was beneficial to the formation of
the molecular composite structure. That could explain the effects
of PVA on the mechanical properties and showed that the molecular
composite structure are favourable for making full use of the
macromolecular potentiality.
INTRODUCT ION
4the
Since melamine was synthesized in 1834 by Liebig Il] its
reactions with other chemicals have been studied. Melamine formaldehyde resin (2) extended the application of melamine. Like
urea formaldehyde (UP) and phenolic formaldehyde (PF) resin, MF
resin can be used as moulding and laminating materials, adhesive
and coating. In order to improve its toughness and to lower its
cost, a lot of modified, reinforced or filled products are developed for practical uses (3]. MF resin and its related products
are often used in paper and textile industry [4). Particularly,
the aqueous solution of MF and PVA is mainly used in papermaking
Sf5. The mechanical properties and structure of MF/PVA as moulding material were studied in this paper. The results showed that
the molecular composite structure benefited the improvement of
mechanical properties.
EXPERIMENTAL
Preparation of prepolymers
When one mol 37% formaldehyde solution with pH 8.0-8.5 was
heated to 60*C, 1/3 mol melamine was added into the solution.
As melamine solved completely, the temperature was elevated to
and kept at 850C. As soon as one milliliter of this solution became tu:-bid when it was mixed with three milliliters of water at
room temperature, 5% aqueous solution of PVA was mixed with it.
The liquid mixture was stirred at 856C for a period of time until it became a homogeneous mixture. After that, the liquid was
put in a vacuum oven to eliminate water at 60C for 72 hours.
Mat. Res. S"c. Symp. Proc. Vol. 171. - 19O Materials Research Society
266
Then
the solid obtained from the oven was grinded into prepoly-
mer powder (<100-mesh).
Measurement of Mechanical Properties
the poder
was moulded
intokg/cm
spesimens
with dimension Atof165"C
MOWx0
me
under
about 1500
pressure
for 12
minutes.
The flexible strength was measured at speed of 15ram/min
according to
the impact
GB1042-79 (National Standard of P. R. China).
resistance
of
the
samples
measured according to GB1043-79.
without
notches
And
was
Investigation of Structure
The homogeneous solution of MP and PTA was coated on an obJect carrier. After being dried,
the sample on the carrier was
observed under a microscope.
The prepolymer powder, after necessary preparation, was
tested with infrared spectrum.
In addition, TER was used to investigate the morphology of
the moulded samples.
RESULTS AND DISCUSSION
Mechanical Properties
Both flexible strength and impact resistance of the composite were improved by the addition of PTA. The flexible strength
and impact resistance were plotted in figure 1 against the PVA
content (weight %). When PTA content was less than about 20%, as
shown in Fig.l, the flexible strength of the material increased
rapidly as the content incredsed. When P1A content was more than
20%, the flexible strength increased slightly. Not quite the same,
the impact resistance straight increased all
the way to 40% PTA.
As we know, in most polymer blend and copolymer systems, the
improvement of the impact resistance is achieved with the loss of
0
50
~4
304
10o 0
20
30
40
o"
=
VA content (%)
Figure I Relation of flexible strenth
and impact resistance to FVA content
(w%) In the composites
l , Lmm mm N
m m m mm(lm mmmmmm
%
269
the flexible _ind tensile strength. The effect of IVA in considerably wide content range on impact resistance and flexible strength
depended on the structure of the composite.
Structure Analysis
The reactions, in which melamine is involved, have been
studied for more thAn one hundred years. Until recent years some
papers on the reactions of melamine and formaldehyde were still
published[6]. Under the experimental condition of this piper,the
reaction of melamine and formaldehyde was is follows
NH'2. 0
NHI2 .
NH2
O
= Melamine=
whe+
where
-NH 2
NH2,
NH2
0
Nc'CH=OI
N\
C-Nll 2
NH2 C-N'
proAlthough trihydroxymethylate could be gotten, the reaction
duct (MF prepolymer) was a mixture of several substitution derivates. Etherification alone with some other reactions occurred
when MF prepolymer was heated.The product of these reactions was
a piece of crosslinded solid that was unmeltable and unsolvable.
Fig.2 showed the infrared spectrums of MF and MP/PVA(90/10)
at different temperatures. The temperature elevated 10 degrees in
one minute. Fig.3 was the differential spectrums at room temperature and at 180'C. Fig.4 showed the peak area of about 3500cm
area
in Fig.2. The area was indicated by the ratios of the peak
at different temperatures to that at room temperature.
16 OC
IE I
140b
6cC
14(t
120*0
Bct
oft
3A
28
19
4
9.0
.0
38
Wavenumbers (cm
(a)
-I
8
x10
19
1"
9.0
-2 )
(b,
Fig.2 IR spectrums at different temperaturer
Temperature elevited 10 C/min
,a) MY/PVA-100/0; (b) MF/PVA901lO
*RT=room temperature
d
4.0
270
OH groups joined the
reactions in this system5
1489
15
(a)
when the prepolymer was
heated. Fig.2 inqicated
1 50
1 51
that the 3500cm- peak
(the IR absorption of O-H
expansion vibration) re985
duced as temperature el1652
1610
1as
it
Fig.3,
evated. In
found that the 990cm1600
1900
0100
80
peak (C-OH expansion vi1)
Wavenumbers (cm
bration) became smaller
for higher temperature.
1487
In addition, whether PVA
(b)
1552
was put in or not, OH
groups(Fig.4) reacted to
1
1053
a quite high level(>90%).
Ether bonds formed
in processing. This was
proved by jhe increment
1659 ---of 1050cm-lpeak(C-0-C
1800
1500
1200
900
asymmetric expansion vi(cm-1
Wavenumbers
bration) in Fig.3. Furthermore, we could know Fig.3 Differential spectrums between IR
that the increme t
f
spectrums at 180 C and at room temperature
1050cm-lpeak in (b)10%
PVA) was obviously larger (a) MF/PVA=100/0; (b) MF/PYA=90/10.
than that in (a)(O% PVA).
That was to say, when FVA was put in the system, more ether bonds
formed.
The INA phase was well distributed in this system. We could
see that in Fig.5. It was significant that PVA phase distributed
as a continuous network phase even only 10% PVA was put in
(a).
This may be caused by the effect of PYA like a surfactant. When
PVA was more than 20%(b), MF phase took spherical shape. If PYA
content was 40% (c) or more, the MF phase spheres were kept apart
absolutely by the PVA phase.
The general reactivity of OH groups was decreased by PVA.
The OH groups on the PVA molecular chain has lower reactivity
than that in MF prepolymer. For the prepolymer containing low PVA
content (e.g. 10%), in addition to that reason, the distribution
of VA phase played a important part. Because PVA phase was continuous, so MP phase was divided into fine drops. As a result,
the general OH reactivity was reduced. The result was proved by
the IR spectrums (Fig.4). In Fig.4, the 3500cm- 1 peak area of the
two samples decreased with the elevated temperature as the result
1O0
I
*
\Fig.4
Variation of A/A*
with temperature
A=3500 cm 1 peak (IR)
area at a certain
50
temperature
A*- A at room tem0
RT
50
100
150
T mperature ('0)
200
perature (RT)
--------
MP/PVA=100/0
MP/PVA-90/1O
271
00_
Z_
100 A.
Fig.5 Microscope photographs of different MF/PVA ratios
MF/PVA: (a) 90/10;
(b) 80/20; (c) 60/40.
of the reactions. For 0% PVA sample, the area changed most rapidly at about 90°C, and for 10% PVA one, at about 120"C.
Molecular composite structure was formed in this composite
system. From our previous experimental result, we knew that PVA
phase was well-distributed even only 10% PVA was put in and the
OR groups reacted to a relatively high extent. So the crosslinkage of the material increased especially in low PVA content range.
The close linkage of the two phase formed molecular composites.
TEM photographs of two samples were exhibited in Fig.6. Sample (a)
was the mould specimen of this experiment and sample (b) was made
from the specimen that HF prepolymer powder mixed directly with
PYA powder. The PVA phase (white) in (a) was fine and regular
whereas that was coarse and irregular in (b). Obviously, this
related to the distribution of PVA phase. In the prepolymer of
(a), the PVA phase was dispersed as the shape showed in Fig.5(b).
After the processing under certain pressure and temperature, the
two phase linked closely and the PVA phase was fine. On the other
hand, in sample (b), after the same processing, the two kinds of
powder combined together but not so closely as (a). Inevitably,
the mechanical properties of the composite were affected by their
molecular composite structure. Both flexible strength and impact
resistance were improved in comparison with the case of polymer
blends and copolymers.
For the samples containing more than 20% PVA, the MF phase
acted as a reinforcer in PVA. Because the PVA content was high
and the HF phase was in the shape of separated spheres, the molecular composite structure was not so significart as in the low
PVA content samples. Nevertheless, the impact resistance and
flexible strength were still improved.
0.25
0.25_ u
,
(a)
(b)
Fig.6 TEN photographs of the MF/PVA(80/20) composites
The composites were made from
(a) solution mixture and (b) powder mixture.
i iiiiil
] mmm
/'lln
Im
NOUNS.
272
SUMMARY
MF/PVA
Because molecular composite structure was formed in
composites, the flexible strength dnd impact resistance was improved significantly when PYA content was less than 20 weight %.
As PVA content increased further until 40%, the impact resistance
increased continuously while the flexible strength didn't increase
so quickly as PVA content lower than 20%. In addition, the addi-
tion of PYA can change the morphology and reactivity of prepolymers and the morphology of the finished polymers.
REFERENCE
1. C. Goldschmidt, Chem. Ztg., 21, 460 (1897).
2. D. Braun, et al, Osterr. Chemie-Zeitschrift, Juli/August,
188-196 (1985).
3. R.J. Schupp, Modern Plastics Encyclopia, 1986-1987, 17-18.
4. G.X. Song, Suliao KeJi, 1, 20-24 (1988).
Adelman, Wilmington, Del., U.S. Patent No.4,461,858
5. Robert
(24 JulyL.1984).
6. Ani2 Xumar, Polymer, 28
Mt)155 (1987).
PART VI
Synthesis-Electro- Optical
Properties
275
MORPHOLOGICAL CONSEQUENCES OF
CATALYTIC HYDROGENATION OF POLYMERS IN THE BULK
LAURA R. GILLIOM, DALE W. SCHAEFER AND JAMES E. MARK*
Sandia National Laboratories, Albuquerque, NM 87185
ABSTRACT
When suitable catalysts are molecularly dispersed in polymers, the
polymers can be modified without added solvent. This paper describes
studies on the morphology of samples of trans-l,4-polybutadiene and sn1,2-polybutadieni which have been partially deuterated in the bulk. The
development of a peak in the SANS data for the 1,2-polybutadiene suggests
the formation of small domains upon deuteration. Possible explanations
for this observation, including chemical and physical heterogeneity, are
evaluated. Results of SAXS and thermal measurements are also considered.
INTRODUCTION
We have previously shown that olefinic polymers can be hydrogenated in
the bulk with transition metal catalysts [1].
The absence of added
solvent is in contrast to more conventional solvent-based methods for
polymer modification [2].
Both the molecularly dispersed (dissolved)
catalyst and the flexible polymer chain provide the mobility required for
high conversions to hydrogenated polymer. The reaction exotherm must be
completely adsorbed by the polymer. Given the unique reaction conditions,
it is appropriate to investigate the morphology of the product polymer at
intermediate levels of hydrogenation.
Small-angle x-ray (SAXS) and small-angle neutron (SANS) scattering
provide structural information on the 5-50OX length scale. Differential
scanning calorimetry (DSC) probes physical transitions in materials. We
have used these techniques to assess physical state and chemical
heterogeneities resulting from the bulk modification. Typically, polymers
were only partially deuterated. Since the huge difference in coherent
scattering lengths of H and D enhances contrast between deuterated and
undeuterated polymer regions, SANS is sensitive to chemical heterogeneity
produced by localized deuteration. SAXS, on the other hand, is relatively
insensitive to deuteratlon because of the small difference in x-ray
scattering length of H and D. SAXS is, however, sensitive to density
variations (e.g., crystallinity) on the 5-500A length scale. This paper
describes our studies using these techniques on trans-l,4-polybutadiene
and =yn-l,2-polybutadiene.
EXPERIMENTAL
Sample Prenaration: Materials: i-l,2-polybutadiene (1,2-PB) was
purchased from Polysciences. NMR analysis suggests approximately 10%
1,4-addition and 90% 1,2-addition. Purchased material was described as
281 crystalline. Trans-l,4-polybutadiene (1,4-PB), a gift of Gencorp, was
88% trans. Both polymers were purified by precipitation from toluene
solution prior to use. Crabtree catalyst jIr(COD)(py)(tcyp)]PF 6
(COD-l,5-cyclooctadiene; py-pyridine; tcyp-tricyclohexylphosphine) was
prepared according to literature procedures [3].
Sample Preparation:
To
polymer (l.5g) dissolved in 25 mL benzene was added 15 mg catalyst. The
Mal. Res. Soc. Symp. Proc. Vol. 171. c 1990 Materials
Research Society
276
solution was poured into an open aluminum mold and was frozen in an
acetone slurry. After demolding, solvent was removed from the solid
mixture under vacuum at a rate sufficient to prevent melting.
Deuterations: Deuterations were performed in a pressure reaction vessel
(V-l8OmL) consisting of a glass sample container, pressure gauge and gas
inlet port. Foam samples were initially pressurized to 40psig deuterium
pressure. Samples of partially deuterated 1,2-PB were removed at 25, 66,
and 85% conversion as determined by pressure drop and confirmed by sample
weight gain. Samples of 1,4-PB were removed at 35 and 70% conversion.
Meaurmnt: SAXS profiles were measured using the lOm SAXS facility
at Oak Ridge National Laboratory. SANS data were taken at Los Alamos
National Laboratory using the LQD camera. Observed slopes have error
limits of ±0.4. Thermal measurements were obtained on a Perkin-Elmer DSC7 system. Surface area measurements were based on nitrogen desorption
(BET) with a Quantachrom Monosorb instrument.
RESULTS
Materils
Polymer samples were fabricated and used as foams in order to maximize
the accessibility of hydrogen throughout the material. The foams were
solution cast as described in the Experimental Section. Scanning electron
micrographs of the foams showed a directional, columnar morphology with
spacing between columns of tens of microns. Since no indications of
catalyst crystallites were observed by scanning or transmission electron
microscopy, the catalyst is believed to be fully solvated in the polymer.
BET studies of the surface area of foams of 1,4-PB indicate a surface area
between 7 and 11 m2 /g. Deuterations of the olefinic bonds in the foam
samples were performed directly under mild conditions.
Small-Anale X-Ray Scattering (SAXS)
Figure 1 shows the SAXS data for the undeuterated and partially
deuterated PB's. These curves represent the scattering cross section I(Q)
as a function of the scattering magnitude of the wave vector Q. The data
have been radially averaged. The abscissa is related to the scattering
angle 8 as Q - (4x/A) sin (0/2). Within the errors introduced by sample
irreproducibility, the curves overlap. Therefore the data sets have been
shifted to demonstrate that the shape of the curves is unchanged on
deuteration. This result is consistent with the insensitivity of x-rays
to low-Z elements.
Although there is some tailing off of the 66% 1,2-PB data at large Q,
all of the curves in Fig. 1 are power-law with exponents near -4. Slopes
of -4 are consistent with Porod's law and are the signature of sharp
interfaces between phases. In this case, the interface in question is the
polymer/air interface of the underlying foam structure. In fact, the
specific interfacial area, a, can be calculated from Porod's constant Kp,
obtained from the magnitude of I(Q) in the Q-4 regime [4j,
Kp - Q 4 I(Q)
(I)
and
a - Kp/2wpo(Ap)
2
(2)
277
0%
.,
V)
1-4
-4.0
66% 1-2
CIA 1-2
'U
0
0.01
o(A".)
0.001
Figure 1.
0.1
SAXS data for unreacted and partially hydrogeiia-
syn-l.2-polybutadiene
(1-2) and trans-1,4-polybutadiene
in
where pa is the foam density and Ap is the scattering0 contrast
2
scattering length between the phases (Ap - 9.9 2x 10-1 cm- ). Using (2)
2
we find a - 3 m /g for the 1,2-PB foams and 8 m /g for the 1,4-PB foams.
These results for the 1,4-PB agree well with the BET measured surface
areas.
Small-Angle Neutron-Scattering (SANS)
Figures 2 and 3 show the development of SANS for 1,4-PB and 1,2-PB
samples, respectively, as a function of deuteration. The data for the
Q flattening out at
unreacted samples show power-law profiles at small
2
2
large Q. Using Eq. (2) we find that a - 6.7 m /g for 1,?-PB and 15.4 m /g
for 1,4-PB. These values of foam surface area are approximately twice
that found by SAXS. It should be emphasized that the errors in these
measurements are approximately 50% because of nonuniform sample thickness
and sample inhomogeneities with respect to foam density.
Upon deuteration, only minor changes are observed for 1.4-PB samples.
The slopes in the low-Q region become steeper indicating a non-distinct
interface. Assuming a sigmoidal contrast erofile, the interfacial
thickness is calculated [5] to be 100+/-10A. Presumably, the pore
boundaries are preferentially deuterated leading to the observed
interfacial contrast gradient. The 2measured foam surface area of
partially deuterated samples is 4 m /g, somewhat smaller than that found
for 0% deuteration. While this difference may be due to sample-to-sample
inhomogeneity, some annealing of the porosity may be caused by the
reaction exotherm.
278
0%
.4
0a
00
~~
*
0*
00
0*
00000
z
00
00
0.01
0.001
011.0
O)(A-I)
70%
1
(C
000
a
~~
00
0.00
0.01.
0'
>iue2-
A4
aa
o
zrn-,-oyuain:
nece
()rwdt:
n
prilydueae
()cretddt
279
6%
as%
0.°1.01
..
.
S25%
,.
.
r
o% 0
16%
,*
.*%
%
z
,
No %
%.0
-
0..
*.OooS.
1
0.0
oIA -,1)
Figure 3.
SANS dat~a for unreacted and partially deUe[dL~d
= -1+,2-polybutadin: (a) raw data;
(b) corrected data.
280
pTo
Deuteration also leads to enhanced scattering in the high Q region.
more clearly reveal this change, the 0% and 35% data in Fig. 2a are
replotted in Fig. 2b. Assuming the background is due to incoherent
scattering, the data were first adjusted to match at high Q before
background subtraction. This procedure scales the data to the same number
of scatterers in the beam, compensating for unknown scale factors due to
sample inhomogeneity. The featureless profile in Fig. 2a is at most
indicative of a weakly phase separated system.
SANS results indicate a qualitatively different structure for the
domains in the 1,2-PB system. A peak is now observed near Q - .06A-1 in
the scaled, background-subtracted data shown in Fig. 3b. This peak is
consistent with the formation of phase-separated domains upon deuteration.
Limiting slopes of roughly -4 on the high-Q side of the peak indicate a
distinct phase boundary. Since the contrast factor Ap in equation (2) is
uncertain here, the interfacial surface area was calculated [4] from Kp
defined in equation (1) and an invariant, q,
a
-
Kp7rO(l-0)/poq
2
q - fQ I(Q)dQ
(3)
(4)
where 0 is equal to the fraction of deuterated double bonds. This value
of 0 rests on the assumption that contrast arises solely from deuterated
domains and not from density fluctuations (see Discussion). To avoid
contributions from the pore interface, I(Q) is assumed to be flat below Q
2
Resulting surface areas are 402, 933, and 1094 m /g for 25%,
- .05 A-.
66% and 85% deuteration. The mean chord, or characteristic lengths,
associated with these surface areas are 25A, 28A, and 31A, respectively.
The large increase in surface area with essentially no change in
characteristic length implies that domain growth proceeds by development
of new domains with minimal growth beyond 30A. Similar behavior is
reported for silica/siloxane systems [6]. The presence of a peak in I(Q)
near .06A 1 implies that the domains are correlated in space. Although
such a peak is usually associated with spinodal decomposition, a peak is
possible if phase separation takes place by nucleation-and-growth. From
Bragg's law, a domain spacing of I00k is found. This value is somewhat
larger than twice the characteristic lengths associated with the domains
calculated from a.
The low-Q data for 1,2-PB are consistent with foam pore surface areas
2
of about 4 m /g. The surface area appears unchanged upon deuteration. An
indication of preferential deuteration at pore boundaries was observed
similar to that seen for 1,2-PB.
Thermal Measurements
Because of their stereoregularity, both the 1,2-PB and the 1,4-PB
samples used in this study are semi-crystalline. Since the deuteration
should not affect the polymer stereochemistry [7], the product saturated
polymers were also expected to be crystalline. The glass transition
temperatures and melt data of the unreacted and partially reacted polymers
provide information on the polymer morphology.
Table I suammarizes the thermal data obtained for the 1,2-PB samples.
A single glass transition was observed at all levels of deuteration. The
glass transition onset temperature descended gradually with deuteration.
The melting temperature also decreased at higher levels of deuteration
from that of the unreacted 1,2-PB. The reduced heat of fusion seen at 66
and 85% deuteration -- as well as the high conversion -- implies reaction
of double bonds originally in crystalline domains. No melting transition
281
ascribable to the product =yn-l-butene (the hydrogenation product of
1,2-PB) was observed in the range 25-175"C. Doi reports a melt
temperature for =-l-butene of 45*C [8]. That region was featureless in
our scans.
Table 1. Results of DSC Analysis of Partially Deuterated syn-l,2Polybutadiene
ID2
0
25
IgLLU
0
-14
LAJlzs
106
105
15.6
15.1
66
-26
95
7.2
85
-30
90
4.1
Two endotherms, one at 420 C and one at 680 C, are present in the DSC
plot for unreacted 1,4-PB, consistent with previous studies[9,10). Upon
deuteration, a substantial new endotherm appears due to the hydrogenated
polybutadiene product which is similar to LDPE
(1[1. The product at 70%
deuteration melts at 990C. The original 420 C endotherm is still present
0
although considerably reduced in size. The 68 C peak may be covered by
the tail of the product melt. Unreacted 1,4-PB had a Tg at -750 C. As
expected for linear polyethylene, no glass transition was observed at high
deuteration.
DISCUSSION
The most striking result of those described above is the difference in
the SANS plots obtained for partially deuterated 1,2-PB and 1,4-PB.
Specifically, the presence of a peak in the 1,2-PB data is unusual. This
peak is similar to that observed for block copolymers and, thus, suggests
domain formation. Four possible explanations for the origin of this peak
have been considered:
Chemical Heterogeneity- The most interesting explanation is the
possibility of a catalyst-localized reaction. Initially the catalyst is
dissolved in the polymer matrix. If preferential reaction occurred on
double bonds near the catalyst molecule and catalyst mobility were
limited, regions of deuterated polymer in a matrix of unreacted polymer
would appear. Such chemical heterogeneity could give rise to the observed
SANS data, although the peak should disappear as the reaction goes to
completion.
The data presented above imply the formation of domains of radius on
the order of 301.
The large increase in the surface area between domains
with no change in characteristic length implies that deuteration proceeds
by nucleation of new domains with minimal growth beyond 30A. The DSC
results on 1,2-PB do not show two distinct glass transitions as is
expected for phase separated materials. It should be emphasized, however,
that the domains under consideration are very small. It is not clear that
such tiny domains would show a distinct glass transition.
282
Crystal-Excluding Reaction- If the amorphous regions of the polymer are
deuterated preferentially as compared to the crystalline regions, the
contrast between those regions may be enhanced in SANS. Although
crystallites are deuterated during the course of the reaction, the DSC
data for 1,2-PB suggest that reaction of amorphous regions occurs
preferentially. Specifically, there is minimal change in the heat of
fusion on going from 0% to 25% conversion. The size of the observed peak
in SANS, however, increases substantially on going to 66% conversion. At
this conversion level, crystallites must have reacted suggesting the
domains are not the result of preferential deuteration of the amorphous
phase. Furthermore, crystallites in these polymers would be expected to
be microns in size. Structure on this length scale would not be detected
by SANS.
Crystalline Product Formation- As mentioned above, the reaction product,
1n-l-butene, may be semi-crystalline.
If reaction occurs randomly
throughout the material but only highly deuterated regions crystallize,
the observed domains could be attributable to the contrast between highly
deuterated crystalline domains and less deuterated amorphous regions.
Further, SANS is known to be sensitive to density fluctuations. The peak
may not be related to differences in deuteration but only to the density
difference between the developing crystalline and amorphous regions. If
this is the case, the size of the peak should increase with conversion as
observed. In contrast, if the peak is related only to differences in
deuteration, it should disappear at 100% reaction. No DSC peak
attributable to che melting of product crystallites was observed; however,
whether the melt transition of tiny crystallites could be observed is not
clear. The SAXS data for the 66% 1,2-PB sample bhows tailing in the high
Q (low size) regime, possibly indicating the presence of tiny polydisperse
crystallites.
Deuteration-induced phase separation- The observed peak in SANS might
signify phase separation via spinodal-decomposition. In general, however,
the peak would be expected to move to smaller Q as phase-separation
progressed. The surface area would also be expected to decrease as
coarsening progressed. Finally, the fully deuterated material would not
be phase-separated and the peak should decrease beyond 50% deuteration.
The restrictions imposed by the polyseric nature of the system could, of
course, lead to unusual kinetic behavior.
If a rubbery system became
glassy, for example, miscibility at high deuteration would be kinetically
impeded.
Miscellaneous- The peak in the SANS data may be due to catalyst
crystallites, some feature of the foam fabrication, or of some other
experimental variable. The fact that 1,4-PB had no such peak limits our
concern about such miscellancus factors.
jCONCLISON
Unfortunately none of the above interpretations is completely
satisfactory. Models which attribute domain formation to preferential
deuteration do not account for the persistence of the peak in the SANS
data at high deuteration levels (where contrast should decrease). On the
other hand, models that attribute the SANS peak to reaction-induced
density differences (e. g. formation of lamellar crystals) fail to account
for the absence of a similar peak in the SAXS profile. Further data on
283
fully deuterated and solution-reacted samples will be required to settle
the issue, Irrespective of its origin, the presence of such a peak
suggests interesting morphological consequences of catalytic hydrogenation
of polymers in the bulk.
ACKNOWLEDGENKNTS
This work was supported by the U, S. Department of Energy under
contract DE-AC04-76DPOO.'89. The authors acknowledge use of the SAXS
facility at Oak Ridge National Laboratory and the SANS facility at Los
Alamos National Laboratory. We are grateful to Dr. 1. G. Hargis of
GenCorp for providing the 1,4-PB and to Jeff Kawola, Ed Russick, and Ron
Weagley for technical assistance. We thank Bob Lagasse for suggestions
concerning spinodal decomposition, We thank Phil Seeger for important
contributions to collection of the SANS data.
REFERENCES AND NOTES
* Permanent address:
Department of Chemistry and The Polymer Research
Center, The University of Cincinnati, Cincinnati, OH 45221
I.
L. R. Gilliom. Macromolecules 22, 662 (1989),
2.
See, for example, E. Marecnal "Chemical Modification of Synthetic
Polymers" in Comoreh. Polym r Sci., vol. 6, edited by G. Allen and J
Berington (Permagon Press, Oxford, 1989), p. 1.
3.
R. H. Crabtree, S. M. Morehouse, J. M. Quirk, Inc. g. Synch, 2.4, 173
(1986).
4.
A. J. Hurd, D. W. Schaefer and A. M. Clines, J. Appl
(1988).
5.
J. T. Koberstein, B. Mora and R. S. Stein, J. Appl. Cryst
(1980).
6.
D. W. Schaefer, J. E. Mark, D. McCarthy;, Li. Jian C. -C. Sun and B.,
Farago, in Polymer Based Molecular Comooslstes, edited by D. W.
Schaefer and J. E. Mark (Mat. Res. Soc. Symp. Proc., &, Pittsburgh,
PA 1990) pp. XX-XX.
7.
NMR analysis of the hydrogenation products in both cases indicated
that no change in stereochemistry had occurred. The analysis of the
spectra for 1,2-PB, however, was hindered by the fact that the
starting material contained approximately 10% 1,4-connectlvity,
8.
Y. Dol, A Yano, K. Soga, and D. R. Burfield, Macromolecules, .L, 2409
(1986).
9.
M. Berger and D. J. Buckley J. Polym. Sci., A, 1, 2945 (1963).
10, P. Wang and A. E. Woodward, Macromolecules,
11.
2_Q,
2718
Cryst
21,
8b,
13, 3.
(1987)
A similar endotherm was observed in the solution hydrogenation of cis1,4-polybutadiene. J. M. C. Cowie and 1. J. McEwen, Ibid.. 1,0,1124
(1977).
285
SYNTHETIC CONTROL OF MOLECULAR STRUCTURE IN ORGANIC AEROGELS
RICHARD W. PEKALA
Lawrence Livermore National Laboratory, Livermore, CA 94550
ABSTRACT
Organic aerogels have been formed from the aqueous, sol-gel polymerization of
resorcinol with formaldehyde.
These materials are transparent and have
continuous porosity with cell/pore sizes of less than 1000 A. Their microstructure is
composed of interconnected colloidal-like particles wit'i diameters of 30-200 A. The
particle size, cell size, surface area, and density of the aerogels are predominantly
controlled by the catalyst concentration used in gel preparation.
INTRODUCTION
The sol-gel processing of metal alkoxides (e.g. tetramethoxy silane, aluminum
sec-butylate) is a convenient method for tailoring the properties of inorganic
materials at the molecular level Sol-gel research has principally focused on the
manipulation of silicate precursors to form polymeric or colloidal structures in
solution. The hypercritical drying of crosslinked silica gels leads to the formation of
a special class of open-celled foanis referred to as aerogels. Aerogels have an
ultrafine cell/pore size (< 1000 A) and a solid matrix composed of interconnected
colloidal-like particles or lightly crosslinked polymer chains, These particles or
chains havr. characieristic diameters of less than 100 A. The above microstructure is
responsible for the unusual optical, thermal, and acoustic properties of these
materials (1,2).
Our research has focused on organic syntheses which proceed through a sol-gel
transition and can be controlled to give aerogels with specific properties. Organic
aerogels have been synthesized from the base catalyzed, aqueous reaction of
resorcinol with formaldehyde. In this reaction, resorcinol (1,3 dihydroxy benzene) is
a trifunctional monomer capable of adding formaldehyde in the 2,4, and/or 6 ring
s
positions. These intermediate products condense into polymeric "clusters" with
diameters -anging from 30-200 A The resorcinol-fornialdehyde (RF) "clusters"
contain surface functional groups (-CH 2 OH) which lead to further crosslinking and
eventual gel formation.
RF gels are dark red in color and transparent, indicative of their ultrafine pore
size. If *he solvent in the pores of the gel is exchanged with a monomer such as
methyl methacrylate, it is possible to form transparent molecular composites by
polymerizing the monomer with a free-radical initiator or UV light. The final
molecular composite consists of two continuous phases with poly(methyl
Mat. A#*. Soc. Symp. Proc. Vol. Ill. I 10 Matedail Research Society
1
286
methacrylate) being the dominant phase (80-95% by volume).
In order to obtain organic aerogels, the RF gels are hypercritically dried from
carbon dioxide. The resultant aerogels are dark red in color and transmit light.
Because RF aerogels consist of a highly crosslinked aromatic polymer, they can be
pyrolyzed in an inert atmosphere to form vitreous carbon aerogels. In a sense, both
RF and carbon aerogels are molecular composites with air being the dominant
phase.
The particle size, cell size, density, surface area, and modulus of organic aerogels
largely depend upon the catalyst concentration (i.e. Na 2 CO 3 ) used in the sol-gel
polymerization.
In this paper, the chemical manipulation of the aerogel
microstructure will be discussed in detail. Characterization methods include TEM,
BET nitrogen adsorption, and small angle scattering.
GEL PREPARATION AND DRYING
Resorcinol (99% purity), formaldehyde (37.5%; methanol stabilized), and sodium
carbonate were used as received from commercial suppliers. All solutions were
prepared from water which had been deionized and distilled. A typical gel
formulation contained 0.29M resorcinol, 0.58M formaldehyde, and 1-6 mM sodium
carbonate for a total of 5% solids. The molar ratio of [formaldehyde]/(resorcinoll was
held constant at a value of 2.0 for all formulations.
RF solutions were poured into glass vials, sealed, and cured for 7 days at 85-95 *C.
Depending upon the % solids and the catalyst level, gel times varied from several
hours to days. After curing, the RF gels were removed from their glass containers
and placed in a dilute acid solution (pH-2) at 50 'C to promote further crosslinking.
Gel modulii showed a noticeable increase after the acid treatment.
In preparation for hypercritical drying, the gels were exchanged into an organic
solvent (e.g. acetone). The solvent-filled gels were then placed in a jacketed pressure
vessel (Polaron® , Watford, England) which was subsequently filled with carbon
dioxide. After several days of exchanging the RF gels with fresh carbon dioxide, the
vessel was taken above the critical point (Tc= 31 *C; PC= 1100 psi). All samples were
maintained at a temperature of 45 IC for a minimum of 4 hours before the pressure
was slowly released over a period of 16 hours. At atmospheric pressure, the aerogels
were removed from the vessel and characterized.
RESULTS AND DISCUSSION
In a manner similar to silica aerogels, the structure and properties of organic
aerogels depend upon the amount of catalyst used in the sol-gel polymerization.
287
Because the pH of the RF solution decreases as the polymerization proceeds, all
formulations are referenced by the [Resorcinol]/[Catalyst] ratio in the mixture. R/C
ratios of 50-300 provide an acceptable range in which transparent gels are formed.
Resorcinol reacts with formaldehyde under alkaline conditions to form mixtures
of addition and condensation products which react further to form a crosslinked
network. The major reactions include: (1) the formation of hydroxymethyl
derivatives of resorcinol, (2) the condensation of the hydroxymethyl derivatives to
form methylene and methylene ether bridged compounds, and (3) the
disproportionation of the methylene ether bridges to form methylene bridges plus
formaldehyde as a byproduct. These reactions have been studied extensively by
NMR 13,4).
In our current model of the RF polymerization, the base catalyst abstracts a
proton from one of the hydroxyl groups on resorcinol to form the corresponding
anion. Because the resorcinol anion is much more reactive than free resorcinol in
solution, it quickly adds formaldehyde in an electrophilic aromatic substitution
reaction at the 2,4 and/or 6 ring positions. Electrostatic repulsion retards interaction
between the charged, substituted resorcinol molecules; however, additional
monomers (formaldehyde and resorcinol) are able to add to this cluster. The
functionality of the cluster increases as resorcinol is covalently attached; therefore,
the probability of additional monomers reacting with the cluster increases.
Titration, chromatography, and NMR data show that resorcinol is the first
monomer to be completely consumed. As a consequence, only formaldehye is
available to react at the outer surface of the RF clusters forming numerous
hydroxymethyl groups. The above process is best described as reaction limited
monomer-cluster growth.
The R/C ratio in a particular formulation is the primary factor which controls
the number of clusters formed in solution and the size to which they grow. In the
late stages of polymerization, cluster-cluster growth is ultimately responsible for gel
formation through the condensation of surface hydroxymethyl groups. In terms of
growth processes, the RF reaction is similar to the base catalyzed polymerization of
TMOS [5-71.
The R/C ratio affects both the density and surface area of the aerogels. Table 1
shows the final density of both RF and carbon aerogels made from solutions
containing 5% solids but varying R/C ratios. In all cases, some shrinkage occurs
during hypercritical drying and the final densities exceed the target densities. The
greatest amount of densification occurs for gels prepared under high catalyst
conditions -- i.e. low R/C ratios. The relationship between the final aerogel density
and the R/C ratio is not linear. In fact, the final densities plateau at a value -15%
higher than the theoretical value for R/C > 150.
28
TABLE I
Target
Density
[Res]/[Cat]
RF
Density
Carbon
Density
50
0.05 g/cc
0.088 g/cc
0.205 g/cc
100
0.05
0.067
0.123
150
0.05
0.057
0.098
200
0.05
0.057
0.085
300
0.05
0.064
0.084
looss
900
CR
arb n
$CO-,
*
400
0
50
100 150 200 250 300 356
[Resorcinoll I[Catalyst]
Figure 1. Aerogel surface areas as a function of the R/C ratio.
289
If the RF aerogels are pyrolyzed at 1050 IC in argon, the final carbon densities are
still a function of the initial R/C ratio. The %mass loss is -50% for all samples, yet
aerogels synthesized under high catalyst conditions experience much greater
volumetric shrinkage (-75%) during pyrolysis leading to higher final densities.
The BET surface areas of the above RF and carbon aerogels are shown in Figure 1.
The observed trend shows that surface areas increase as the amount of catalyst
increases in a formulation. A slightly stronger dependence of surface area upon the
R/C ratio is observed for the RF aerogels as compared to the carbon aerogels.
The effects of the R/C ratio upon both density and surface area suggest
differences in the RF gel structure which translate to the final dried aerogel. Figure
2 shows the microstructure of two RF aerogels synthesized at the extremes of our
catalyst conditions. Each aerogel is composed of interconnected colloidal-like
particles which were referred to as "clusters" in solution. At R/C=300, the particles
have diameters of 160-200 A and are connected in a "string of pearls" fashion. At
R/C=50, the particles have diameters of 30-50 A. These particles are fused together
in such a manner that it is sometimes difficult to visualize individual particles or
beads. This high degree of interconnection between the RF particles is reflected in
mechanical property data which show that R/C=50 aerogels are loX stiffer than
R/C=300 aerogels at equivalent densities [8].
(a)
(b)
SoA
Figure 2. Transmission electron micrographs of RF aerogels prepared
at 5% solids with different [resorcinol](catalyst
ratios: (a) 50 (b) 300.
290
100000-
000
&Carbon
Ooo
0
0 0
0'?
R/C-300
RF R/C-300
10000
aa
.-
1000
0
C
a,
10
,m-.
u
10o-
.00 1
.0 1
.1
q 1/1
Figure 3. SAXS curves for selected RF and carbon aerogeis.
Densities equal 0.065 and 0.085 g/cc, respectively.
Disordered materials such as aerogels often display "dilation symmetry" which
means that they are geometrically self-similar over a given range of size scales.
Small angle scattering can be used to evaluate the fractal dimensions (D,D s ) of a
material [9-111. In this method, fractals show a power law dependence of the
scattering intensity I(q) on the scattering vector, q. The latter quantity is defined
such that q= 4irX sin (0/2) where X is the wavelength and 0 is the scattering angle.
For fractals, the following relationships hold:
Mass Fractal
I(q)
-
q -D
1< D <3
(1)
Surface Fractal
I(q)
-
q (6-Ds)
3< 6-D s <4
(2)
Figure 3 shows the SAXS curves for RF and carbon aerogels synthesized at R/C=300.
At large q, both curves roll over to a slope of -4 which indicates that these materials
have smooth surfaces on a microscopic scale. At intermediate q values, I(q) does not
291
show a power law dependence for either curve. The above data indicate that these
materials are not surface or mass fractals. The same finding applies to aerogels
synthesized at other catalyst conditions. Only the correlation range, which relates to
the median pore size, changes as a function of the R/C ratio. This data and other
SAXS results will appear in a future publication.
SUMMARY
Organic aerogels with tailored physical properties (e.g. surface area, compressive
modulus) can be formed from the sol-gel polymerization of resorcinol with
formaldehyde. In this reaction , the [Resorcinol]/ [Catalyst] ratio is responsible for
the final aerogel microstructure. Although TEM reveals the microstructure of
organic aerogels to be similar to silica aerogels, SAXS data show that these materials
differ in that organic aerogels are not mass or surface fractals.
ACKNOWLEDGMENTS
The author would like to thank Cynthia T. Alviso for preparation of the gels
used in this project. Special thanks to J.S. Lin (ORNL), S. Spooner (ORNL), and
D.W. Schaefer (SNLA) for their assistance in acquiring the SAXS data. This research
was performed under the auspices of the U.S. Department of Energy by Lawrence
Livermore National Laboratory under contract # W-7405-ENG-48.
REFERENCES
M. Gronauer, A. Kadur, and J. Fricke, in Aerogels edited by J. Fricke (SpringerVerlag, New York, 1986) pp. 167-173; 0. Nilsson, A. Fransson, and 0. Sandberg,
ibid. pp. 121-126.
[2] J. Fricke, Sci. Am. -2(5),92 (1988).
[3] D.D. Werstler, Polymer 27, 757 (1986).
[4] A. Sebenik, U. Osredkur, and I. Vizovisek, Polymer 22, 804 (1981).
[5] C.J. Brinker and G.W. Scherer, J.Non-Crystalline Solids 70, 301 (1985).
[6] D.W. Schaefer, Science 243, 1023 (1989).
[7] D.W. Schaefer, MRS Bulletin 13(2), 22 (1988).
[81 J.D. LeMay, presented at the First Pacific Polymer Conference, Maui, Hawaii,
December 1989 (unpublished).
[9] A. Craievich, M.A. Aegerter, D.I. dos Santos, T. Woignier, and J. Zarzycki, J.
Non-Crystalline Solids 86, 394 (1986).
[10] R. Vacher, T. Woignier, J. Pelous, and E. Courtens, Phys. Rev. B 7(1), 6500
(1988).
[111 D.W. Schaefer and K.D. Keefer, Phys. Rev. Lett. 5(20), 2199 (1986).
[1
293
SYNTHETIC PROCEDURES FOR PREPARING CROSS-LINKABLE ACRYLIC
COMB-LIKE COPOLYMERS VIA MACROMONOMERS
GANG-FUNG CHEN* AND FRANK N. JONES
Polymers & Coatings Dept., North Dakota State University, Fargo, ND
58102. *Present address: Ashland Chemical Company, Columbus, Ohio
ABSTRACT
A versatile procedure was developed for synthesis of acrylic comb-like
copolymers in three steps: (1) Hydroxyl terminated oligomers were
synthesized from methyl methacrylate, butyl acrylate and glycidyl acrylate
by free-radical initiated addition polymerization using a functional chain
transfer agent, 2-mercaptoethanol, and very low initiator levels. (2) The
oligomers were converted to macromonomers by reaction with isocyanatoethyl
methacrylate. (3) The macromonomers were polymerized by free-radical
initiation. Conditions during the first stage must be carefully selected to
minimize formation of difunctional material which could cause gelation in
the third stage. A variety of structures can be made such as comb-like
copolymers with homopolymer tines or comb-like homopolymers with copolymer
tines. Functional groups can be introduced by copolymerizing glycidyl
acrylate into the macromonomer. Assignment of comb-like structures is not
rigorously proven but is strongly supported by the synthetic route and by
DSC, FT-IR and chromatographic data.
INTRODUCTION
Comb-like polymers are distinguished from graft polymers by the close
and regular spacing of their side chains and from star polymers by the
attachment of the side chains to a polymeric chain rather than to a central
core. They share with star polymers the potentially useful property of
having much lower solution and bulk viscosities than linear polymers of
similar molecular weight. Comb-like copolymers are generally synthesized by
polymerization of macromonomers -- oligomers or polymers that contain a
single polymerizable group on one end. Macromonomers have been synthesized
by a variety of techniques [] involving anionic [2], cationic [3], and
group transfer [4,5] polymerization.
Until recently there have been few reports of synthesis of
macromonomers by routes involving free-radical chain polymerization. If
such a process could be developed it would open an economical route to a
very wide variety of comb-like polymers including, possibly, structures that
have reactive sites for subsequent modification or cross-linking. A major
obstacle to development of such a process is the difficulty of attaching a
polymerizable group to one end of the great majority of macromonomer
molecules while suppressing the formation of molecules having two or more
reactive groups. A macromonomer contaminated with with polyfunctional
material can be expected to gel when polymerization is attempted.
In 1986 Albrecht and Wunderlich shcwed that these difficulties can be
overcome in the case of oly(methyl methacrylate) (6]. They synthesized
PMMA macromonomers with M 's of 6500 to 23,000 by free-radical
polymerization of MMA in the presence of a functional chain transfer agent,
2-mercaptoethanol (2-ME), followed by fractionation and reaction of the
mono-hydroxy functional products with isocyanatoethyl methacrylate (ICEM).
Polymerization of hhese macromonomers yielded comb-like polymers with M 's
on the order of 10 as determined by light scattering. The necessity oy
fractionation may limit the utility of this metnod to relatively high T
materials. Thus it would be desirable to find a procedure that does nof
involve fractionation.
M t. Res.Soc. Symp. Proc. Vol, 171. 'i1O Materials Research Society
294
Here we report the development of such a procedure -- one which appears
to be relatively general with respect to monomer selection and product T
It wi~l be shown that comb-like copolymers of methyl methacrylate (MMA), g
buty! acrylate (BA) and glycidyl methacrylate (GMA) can be synthesized by a
three-step procedure that can be idealized as follows:
I& + M + HOXSH --> I10 MKMNMMH + HOXSMMMMMMMMMMH + termination
1,nol 1000mol 100mol
Imol
98moi
products
HOXSMMMMI4IOI + CH2 fC(Me)COOYN-C-0 --> CH 2 'C(Me)COOYNHCOOXSMMMMMMMMMH
ICEM
CH2 .C(Me)COOYNHCOOXSMMMMMN
mac romonomer
+ I- --
comb-like copolymer
[I- = initiator, M = mixed monomers, X and Y = -CH2 CH2 -]
The first step, macromonomer synthesis, was performed under these conditions:
(1)
(2)
(3)
(4)
Solution polymerization in toluene at 108 - 110 °C or in methyl
isobutyl ketone at 80 to 90 °C.
Monomer starved conditions.
Low levels of initiator and high levels of chain transfer agent -for example a 1000/100/I or a 1000/50/1 mol ratio of
monomer/2-ME/initiator, and
Postheating the product to 140 to 175 °C in most cases.
These conditions were chosen to minimize termination by combination
(expected to lead to difunctional materials) and to virtually eliminate
unreacted initiator before addition of ICEM. Deviation from these
conditions generally gave materials that gelled during attempted
polymerization of the macromonomer.
By adhering to these conditions a variety of structures including
homopolymeric backbones with copolymeric tines and copolymeric backbones
with homo- and copolymeric tines were synthesized.
EXPERIMENTAL DETAILS
Details can be found in the Ph.D. dissertation of G-F. Chen [7].
synthetic procedures described here are representative.
The
Materials. Reactants were generally 97 to 99 % grades obtained or
purchased from commercial sources and were used without further
purification. Isocyanatoethyl methacrylate (ICEM) was obtained from Dow
Chemical Company, which no longer supplies this material; other sources are
said to exist. "Aromatic 100" is a mixed alkyl benzene solvent, bp 155 177 °C, supplied by Exxon Chemical Company.
Synthesis of Hydroxyl Terminated Homo- and Co-polymers of Methyl
Methacrylate (MMA) and Butyl Acrlate (BA). Solutions of 2-mercaptoethanol
(2-ME) and 2,2"-azobis(iaobutyronitrile) (AIEN) in the monomer(s) having a
monomer/2-ME/AIBN mol ratio of 1000/100/1 or 1000/50/1 were added
continuously during 2.5 h to a 3-necked reaction flask containing refluxing
toluene in a N atmosphere. The weight of toluene approximately equalled
the weight of reactants. Stirring and refluxing were continued throughout
the addition and for 4 h therafter. A sample was withdrawn to
gravimetrically estimate conversion of monomer to non-volatile material;
295
conversion varied somewhat but typically ran 60 to 80 %. The resulting
solution was gradually heated to 160 - 170
and stirred under a nitrogen
purge to volatilize solvent and unreacted monomers and to decomposeYesidual
initiator. The FT-IR specta had substantial peaks at about 3500 cm
(OH
stretching) and were otherwise consistent with the assigned structures.
Hydroxyl number determinations by the pyromellitic dianhydride/dimethylformamide method [81 showed in most cases that all or almost all of the 2-ME
could be accounted for in the non-volatile material; in a few cases small
amounts (< 10 %) of unreacted 2-ME was found in the volatile solvent/monomer
distillate. After removal of volatiles the residues were dissolved in
sufficient xylene to make 30 wt-% solutions.
In other experiments methyl isobutyl ketone was substituted for toluene
in the above procedure and the reaction was carried out at 80 to 90 °C.
Conversions of up to 90 % were attained by this procedure.
Synthesis of MMA and BA Homo- and Copolymer Macromonomers. To the
above solutions were added isocyanatoethyl methacrylate (ICEM) and dibutyl
tin dilaurate (DBTDL); the amount of ICEM was a 10 mol % excess over the
amount of 2-M used in the above synthesis, and the concentration of DBTDL
was 4.7 x 10
wt %. The total concentrations of reactants in xylene was
about 30 wt %. The solutions were heated a 55 C with stirring overnight.
Disappearance of the FT-IR peak at 3500 cm
(OH) and appearance of peaks at
3350 (NH) and 1640 cm
(urethane C=O) indicated complete reaction.
Synthesis of Comb-like Homo- and Copolymers. AIBN (2.5 mol/L) was
added to the above solutions, and the solutions were heated with stirring
under N to 60 °C. The solutions became very viscous within 5 to 10 min,
and heaing was continued briefly and the solutions were cooled. The
solutions often gelled when lower AIBN concentrations or higher temperatures
were used. GPC retention times of the comb-like polymers were relatively
short, typically 22 to 24 minutes; in comparison, retention times of the
hydroxyl terminated polymers from which they were made had retention times
in the range 31 to 35 min. The comb-like copolymers contained small to
modest fractions of long elution time material, presumably unreacted
macromonomer or an unreactive fraction present in the macromonomer. Typical
GPC traces are shown in Figure 1.
Synthesis of Hydroxyl Terminated Copolymers of BA with Glycidyl
Acrylate (GA) and Conversion to Macromonomers. BA/GA mixtures (83/17 and
67/33 w/w) were polymerized by procedures essentially the same as those
described above using MIBK as solvent. Reaction temperatures were 80 to 90
0
C, and no postheating step was used to remove solvent and unreacted
monomers. Conversions were 84 to 88 %. These hydroxyl terminated
copolymers were reacted with ICEM essentially as described above to yield
macromonomers.
Synthesis of Comb-like Copolymers with Mixed Tines. Mixtures of homoand copolymer macromonomers prepared as described above were polymerized
with AIBN under conditions similar to those described above. In some
instances it was necessary to heat the reaction mixture to about 90 °C to
complete polymerization, as indicated by GPC.
Test procedures. Differential scanning calorimetry (DSC) was effected
with a du Pont Model 990 Thermal Analyzer using samples that had been heated
under N 2 for 2 h to remove volatiles. Heating rate was 10 °C/min. T was
taken as the onset of the endothermic deflection. Gel permeation
g
chromatography (GPC) using Waters 100A, 500A, 1000A and 10000A columns in
series; the solvent was tetrahydrofuran, and a Waters Model R401 refractive
index detector was used.
296
RESULTS AND DISCUSSION
The synthetic conditions for hydroxyl terminated polymers were devised
after a good deal of trial and error; they seem reproducible except that
monomer conversions vary. This variation is attributable to the unusually
low levels (0.07 wt %) of AIBN; 0.5 to 1.0 wt % is commonly used. Five such
materials made with 1000/100/1 monomer/2-ME/AIBN ratios had M of 800 to
1000 and
/
of 1.6 to 2.2 as measured by GPC; conversions ere 68 to 100
Five simi~ar materials prepared with 1000/50/1 ratios had M of 1400 to
%.
n
2100 and
/M of 1.5 to 2.4 at conversions of 60 to 99 %.
w nn
Conversion of the hydroxyl terminated polymers to macromonomers by
reaction with ICEM and polymerization of the macromonomers is relatively
straightforward providing hydroxyl terminated precursors are prepared as
described. Substantial deviation from the recommended procedure generally
afforded precursors that caused gelation during the third stage. The
recommended procedures yield macromonomers that do not gel, although they
become very viscous in 30 wt % solution. Some difunctional macromonomer is
presumably present and causes bridging of comb-like structures, but its
level is apparently low enough that gelation can be avoided.
GPC traces of a hydroxyl terminated poly(NMA/BA) (70/30 w/w) polymer
and of the comb-like copolymer made from it are shown in Figure 1. In this
case the monomer/2-ME/AIBN ratio was 1000/100/1. Retention times are marked
on the Figure, and molecular weights of polystyrene calibration standards
are also indicated. While it is likely that the hydroxyl terminated
precursor has a molecular weight close to the indicated level, it is seems
unlikely that the molecular weight of the comb-like copolymer can even be
estimated in this way, as discussed below.
Figure 1.
GPC Trace of an Hydroxyl Terminated poly(MMA/BA) Copolymer
[A] and of A Comb-like Copolymer made from it [B]
[Al
[B]
10
10
10'
10'
Upp lin Retentlon tlm, mlin.
Lower lw.Moleculer weight of polystyreie standards
With the objective of making comb-like copolymers bearing functional
groups as potential cross-linking sites we investigated copolymerization of
glycidyl acrylate into the hydroxyl terminated precursors. Somewhat milder
reaction conditions, described in the EXPERIMENTAL SECTION, were used in
order to minimize the potential for side reactions. One side reaction of
concern is the possibility of reaction of the thiol group of 2-ME with the
oxirane ring of the glycidyl acrylate, an occurance that would increase
297
functionality of the hydroxyl terminated precursor and probably lead to
gelation in step three. It proved possible to prepare hydroxyl terminated
precursors of composition BA/GA in ratios 83/17 and 67/33. Successful
conversion of these materials to comb-like copolymers, indicated that the
chain transfer reaction is much faster than the oxirane addition reaction
and consumes virtually all of the 2-ME.
Availability of different macromonomers opens possibilities for
synthesis of a wide variety of structures involving copolymerization in the
first stage or the third stage, or both. For example, copolymerization in
the first stage and homopolymerization in the third will yield a relatively
uniform comb-like copolymer. On the other hand, if two dissimilar
macromonomers from the first stage are blended and then polymerized in the
third stage, a comb-like copolymer with homopolymer tines can be expected.
Evidence for the formation of such materials was provided by differential
scanning calorimetry (DSC) studies (Table 1). Comb-like 1MA/BA copolymers
made from single copolymeric precursors displayed a single T , while comblike copolymers of similar overall composition made from mixfures of
homopolymeric precursors ("comb-like copolymers with homopolymer tines")
displayed two T 's, one presumably related to the T of the BA precursor and
the other to thl T of the IMKA precursor. However,gonly the lower T 's
could be detected In the comb-like copolymers with a homopolymeric MMA tine
and copolymeric BA/GA tines.
Table I.
T of Selected Hydroxyl Terminated Precursors and
of Comb-Like Copolymers
Tg
Description
Composition (wt ratio)
P-MMA precursor
P-BA precursor
BA/GA precursor A
BA/GA precursor B
OH
OH
OH
OH
Comb-like
Comb-like
Comb-like
Comb-like
Comb-like
Comb-like
Comb-like
Comb-like
Control
-15
MMAIBA-30/70, copolymer tines
8
MMA/BA-50/50, copolymer tines
24
MMA/BA-70/30, copolymer tines
-44, 39
MMA/BA-30/70, homopolymer tines
-43, 29
MMA/BA-50/50, homopolymer tines
-49, 39
MMA/BA-70/30, homopolymer tines
-22
MMA homopol., BA/GA a3/17 copolymer
-17
MMA homopol., BA/GA 67/33 copolymer
26
KMA/BA/HEMA-40/40/20 random copolymer
terminated
terminated
terminated
terminated
PMMA,
PBA,
BA/GA
BA/GA
M
ca. 1000
n ca. 1000
93/7 copolymer
67/33 copolymer
40
-60
-40
-35
Gel permeation chromatography calibrated with linea5 polystyrene
standards indicated molecular weights on the order of 10 for the comb-like
It is
copolymers, but little meaning is assigned to the absolute values.
well known [6,9] that the hydrodynamic volume of highly branched polymers is
much lower than that of similar polymers of the same molecular weight. Thus
correlation of GPC retention times of comb-like polymers with those of
linear standards can be expected to understate the molecular weight of the
comb-like materials. For example, Albrecht ang Wunderlich measured
molecular weight of their comb-like PMM as 10 by light scattering but only
4.5 x I0 by GPC, and Masson reported similar differences.
Retention times were shorter for polymers made at lower (60 vs. 90 °0
T
298
stage-three polymerization temperatures, indicating that, on a relative
basis, lower temperatures favor higher molecular weight comb-like
copolymers. A recent kinetic study of free-radical copolymerization of
macromonomers showed that the reactivity of a growing chain with A
macromonomer end toward a second macromonomer decreases as molecular weight
increases.[10] This observation is attributed to a kinetic excluded volume
effect. In the context of the present study, this result is consistent with
the assignment of comb-like structures for the third-step products but
suggests that further study of how reaction conditions affect molecular
weight are warranted.
CONCLUSIONS
Here we have demonstrated an apparently versatile procedure for
synthesis of short chain acrylic macromonomers by a free-radical initiated
chain transfer process. These materials can be converted to an almost
limitless variety of comb-like polymers and copolymers. Reactive functional
groups can be incorporated by copolymerization of glycidyl acrylate.
Assignment of comb-like structures is not rigorously proven but is strongly
supported by the synthetic route and by DSC, FT-IR and chromatographic data.
ACKNOWLEDGEMENT
Support of this work by a National Science Foundation EPSCoR/ASEND
grant is gratefully acknowledged.
REFERENCES
4(1986).
[1]
P. R. Rempp and E. Franta, Adv. Polym. Sci. 58, 1 (1984).
[2]
G. 0. Schultz and R. Milkovich, J. Appl. Polym. Sci..
[3]
B. Ivan, J. P. Kennedy and V. Chang, J. Polym. Sci. (Chem.
1523, 1539, 3177 (1980).
[4]
F. P. Boettcher, J. Macromol. Sci.-Chem
[5]
D. Y. Sogah and 0. W. Webster, J. Polym. Sci., Polym Lett. Ed..
(1983).
[6]
K.
[7]
G-F. Chen, Ph.D. Dissertation, North Dakota State University, Fargo, ND
58105, pp 185 - 289.
[8]
B. M. H. Kingston, J. J. Garey and W. B. Hellwig, Analy. Chem..
86, (1969).
[9[
Z. Grubisic, P. Rempp and H. Benoit, J. Polym. Sci.. Polym. Lett. Ed.,
5, 753, (1967).
27, 4773 (1982).
Ed.).
18,
A22(5-7), 665 (1985).
Albrecht and W. Wunderlich, Ansew. Makromol. Chemie.
[101 Y. Nabeshima and T. Tsuruta, Makromol. Chem., 190,
21, 927
145/146,
1635, (1989).
89,
41(1),
299
SYNTHESIS AND CHARACTERIZATION OF SEGMENTED
COPOLYMERS OF A METHYLATED POLYAMIDE AND A
THERMOTROPIC LIQUID CRYSTALLINE POLYESTER
Gregory T. Pawlikowski, R. A. Weiss and S. J. Huang
Box U-136, Institute of Materials Science, University of Connecticut, Storrs, CT
06269
ABSTRACT
A block copolymer consisting of liquid crystalline polyester segments
and methylated polyamide segments has been synthesized. Solution polycondensation of acid chloride end-capped poly(terephthaloyl
phenylhydroquinone) (LCP portion) with an amine terminated poly(N,N'dimethylethylene sebacamide) was utilized to prepare the block copolymer.
Characterization by differential scanning calorimetry, infrared spectroscopy,
thermogravimetric analysis, optical microscopy and elemental analysis has
been performed to verify the existence of the block copolymer that may have
potential as a molecular composite material or self-reinforcing thermoplastic.
INTRODUCTION
Much of the high strength/high modulus composites currently available
stem from the use of macrophase reinforced materials such as fiber reinforced
composites and self reinforced thermoplastics where the reinforcement fiber
phase is on the order of microns in diameter. Although these composite
systems have attained widespread use as high strength materials, their
strengths have been limited due to fiber imperfections, the tendency of the
fibers to fibrillate and poor interfacial adhesion between the fiber
reinforcement and matrix phase.1 It is believed that reinforcement on a much
finer scale (10 - 30 nm or less) would result in more efficient reinforcement and
create stronger materials. 2 These types of materials have been termed
molecular composites.
Molecular composites have been found to possess very high strength and
stiffness. Hwang and coworkers 3 prepared molecular composites by solution
blending poly(p-phenylenebenzobisthiazole) (PPBT) and poly(2,5(6')-benzimidazole) (ABPBI) (30/70). Spun fibers of this blend were found to have a Young's
modulus of 117.0 GPa and a tensile strength of 1270 MPa. Molecular
composites also require much less reinforcement loading for equivalent
strength.
Takayanagi 2 used solution blending techniques to prepare
molecular composites of poly(p-phenyleneterephthalamide) (PPTA) with nylon
6 and nylon 66. He reported that this material, which contained only a few
percent rigid-rod molecules, displayed the same mechanical properties as
fiber reinforced plastics which employ 40 percent fiber reinforcement.
The majority of work on molecular composites is being done on solution
blended systems. While these systems have been successful in achieving high
strength molecular composites, the technique has limited process" g potential
and requires large amounts of harsh solvents. It would be more t.esirable to
have a truly melt processible molecular composite capable of 'in situ' fiber
formation of the reinforcement phase during processing. One possible route
for obtaining such a material may be through the use of block copolymers of
the reinforcing rigid-rod phase with the flexible-coil matrix phase. This
chemical attachment could act to anchor the reinforcement phase to the
matrix, creating a strong interfacial bond between the two phases. Also,
Mat. Rev. Soc. Syrup. Proc. Vol. 171. '1990 Materials Research Society
300
chemically linking the two incompatible polymers 4would provide the driving
force for intermixing, resulting in better dispersion.
The work presented here involves the solution polymerization of a block
copolymer of a rigid-rod segment and a flexible-coil segment. The rigid-rod
polymer used was a liquid crystalline polyester (LCP) (III) of phenylhydroquinone (PHQ) (I) and terephthaloyl chloride (TPC) (I) and the flexiblecoil polymer was poly(N,N'-dimethylethylene sebacamide) (VI) prepared from
sebacoyl chloride (IV) and N,N'-dimethylethylenediamine (V). This work
represents the initial effort in our attempt to create a melt processible
molecular composite material where the reinforcing phase forms an 'in situ'
fiberous network during processing with diameters in the range of 10 - 30 nm
or less.
EXPERIMENTAL
An oven dried, 200 ml round bottom, three necked flask was equipped
with a gas inlet valve, a mechanical stirrer and a stopper. A 0.06 wt %
PHQ/tetrachloroethane (TCE) solution was slowly added to a 0.04 wt %
TPC/TCE solution containing 10 % molar excess TPC and 150% pyridine (acid
acceptor) over a period of about 1 to 1.5 hours. The reaction mixture was left to
stir for about 24 hours. One half of the LCP solution was precipitated in about
400 ml acetone. The precipitate was then washed with acetone and dried. The
remaining half of the reaction mixture was immediately used to prepare the
block copolymer with half of the N-methylated polyamide reaction mixture as
described below.
o
HO
gOH
o
1
C
C1_--.O.--
+
I
C
-
1
C-
1o
4
II
0 o
oh
C
0III
An oven dried, 200 ml round bottom, three necked flask was equipped
with a gas inlet valve, a mechanical stirrer and a stopper. A 0.04 wt % N,N'dimethylethylenediamine/TCE solution containing 150% pyridine was cooled
down to approximately -15 *C. To the resulting cooled solution, 10% molar
excess sebacoyl chloride was quickly added while rapidly stirring. The
stirring was continued for approximately 24 hours while the reaction mixture
was allowed to gradually warm to room temperature. Half of the polyamide
solution was worked up for analysis. The remaining half was immediately
used to prepare the block copoly- er with half of the LCP reaction mixture
from the reaction described previously.
0
0
4
1H
- _-CH-CHr-N-H
X_
CN.CI*W
+
IV
V
VI
Preparation of the block copolymer was achieved through a nucleophilic
acyl substitution reaction between the acid chloride terminated liquid
crystalline polyester and the amine terminated polyamide oligomers. The half
of the polyamide reaction mixture remaining in the reaction flask was stirred
while the remaining liquid crystalline polyester portion was quantitatively
transferred to the polyamide reaction flask. After 24 hours, the copolymer
product was worked up for analysis by precipitating in approximately 500 ml
301
acetone. The precipitate was filtered, washed with acetone and dried in a
vacuum oven.
III
+ VI-
C- CH
[ C-O-C---.0
C-N-CHrCHrN
CH
VII
3
RESULTS AND DISCUSSION
Infrared snectrosco0v
Infrared spectra were obtained on a Nicolet 60 SX FT-IR spectrometer.
The most significant absorption of the N-methylated polyamide homopolymer
was due to the N-C=O carbonyl stretch of the amide group at 1643 cm-1 (Fig.
1). Also significant in this spectrum is the absence of an absorption at 1715
cm- 1 which would correspond to carboxylic acid endgroups. In contrast to the
'P showed a
spectrum of the N-methylated polyamide, the spectrum of thr
tretching of
strong absorption band at 1734 cm- 1 due to the O-C=O carboi
the ester group (Fig. 1). A small but important absorption occurred at 1793 cm1. This was due to the Cl-C=O acid chloride carbonyl stretch of the acid
chloride endgroups.
C,- C.0
o
--
N-C-C
ii
LU
LCP HOMOPOLYMER
10
POLYAMIDE HOMOPOLYMER
COPOLYMER PRODUCT
a6oo
1 88
1776
1 6
15 5
WAVENUMBER
Fig. 1 Carbonyl stretch region of FT-IR spectra
for homopolymers and block copolymer.
302
In addition to other characteristic absorption bands from each of the
homopolymers, the spectrum of the block copolymer shows absorptions due to
the carbonyl stretch of both the LC polyester, at 1734 cm-1, and the Nmethylated polyamide, at 1643 cm- 1 (Fig. 1). The absence of the 1793 cm- 1
absorption due to the acid chloride carbonyl stretch of the LC polyester
endgroups indicates that copolymerization of the two oligomers has occurred.
In addition, the inability to extract the polyamide from the product clearly
demonstrates that it is indeed a copolymer. Because of the extremely high
solubility of the polyamide in many common solvents, any oligomer not
chemically joined to the LCP block would have been washed away during
precipitation of the copolymer product or during Soxhlet extraction of the
copolymer product. In fact, Soxhlet extraction in methanol, a very good solvent
for the polyamide, resulted in no significant weight loss. These two points
combined with the FT-IR spectra provide strong evidence that the two
segments were indeed chemically bonded.
Differential scanning calorimetry
Differential scanning calorimetry was performed on a Perkin-Elmer
Series 7000 Thermal Analyzer equipped with a DSC cell. The rate at which the
experiments were carried out was 20 0C/min. The DSC thermograms of the
second heating and cooling for the homopolymers and the block copolymer is
shown in Fig 2.
The thermograms of the N-methylated polyamide
homopolymer showed no first order transitions, indicating that this is an
amorphous material with a glass transition temperature of approximately -10
'C. This is expected because of the disruption of hydrogen bonding due to the
N-methyl groups.
The DSC thermograms of the LCP homopolymer showed a small glass
transition at a temperature of 135 *C. On heating, the thermograms showed
an endotherm due to the crystal to nematic transition and an exotherm on
cooling corresponding to the reverse nematic to crystal transition. The
thermogram of the second heating of the block copolymer indicated a glass
transition temperature of 120 OC. At higher temperatures, the thermogram
showed only a small, broad crystal to nematic endotherm if any at
approximately 300 *C. On cooling, The thermogram of the block copolymer did
not show any clear transitions, but the absence of a recrystallization exotherm
is noted. At this time it is not fully understood why a recrystallization
exotherm is not seen upon cooling.
Ontical microscopy
Dark field microscopy was performed on a Nikon optical microscope
equipped with cross polars, a hot stage and a Lintam TH600 temperature
controller. As in the DSC experiments, the samples were heated and cooled at
a rate of 20 OC/min. The microscopy of the liquid crystalline polyester
homopolymer showed a birefringent material that remained solid to a
temperature of about 330 0C at which point it started to melt. The molten liquid
remained highly birefringent with a texture similar to nematic liquid
crystalline polymers. Furthermore, this birefringence can be enhanced upon
shearing the LCP between two glass coverslides.
The microscopy of the block copolymer was much more complex than
the LCP homopolymer. At approximately 270 0C there existed an isotropic
liquid phase surrounding a solid anisotropic phase. When the temperature
303
reached 330 'C, the anisotropic phase began to fragment and disperse into the
isotropic phase. As the temperature was held isothermally at 340 °C, the LCP
phase slowly dissolved into the isotropic phase to create a one phase, slightly
birefringent system. Shearing the block copolymer at this temperature did not
enhance the birefringence. Upon cooling, however, the birefringence of the
LCP was able to be shear induced starting at about 300 C. In the absence of
shear, small highly birefringent specks of the LCP phase began to precipitate
out of the one phase 'solution' at approximately 285 'C.
POLYAMIDE HOMOPOLYMER
-
mmmlidmlli
,.min
il
mT9
Temeature (C)
Fig. 2 DSC thermograms of 2nid heat and cool
for homopolymers and block copolymer.
304
Elemental sAallmi
The elemental analysis results of the LCP homopolymer showed that it
contained the following percentages of each element: 72.47 % carbon; 3.66 %
hydrogen; 19.54 % oxygen and 4.25 % chlorine. These percentages correspond
most closely with the number of repeat units being five or n = 5. This indicates
that the LCP homopolymer has an Mn of approximately 1780 g/mol. The
percentages of the elements contained in the N-methylated polyamide
homopolymer were found to be as follows: 50.67 % carbon, 10.00 % hydrogen,
18.20 % oxygen and 10.69 %nitrogen. Because this polymer does not contain
unique endgroup elements, the results for this segment are not conclusive.
However, these percentages best correspond to an N-methylated polyamide
with ten repeat units or m = 10, which corresponds to an Mn of about 2630
g/mol.
CONCLUSION
It has been demonstrated that a solution polycondensation reaction
between sebacoyl chloride and ten percent excess of N,N'-dimethylethylene
diamine has been employed to prepare a low molecular weight, telechelic
poly(N,N'-dimethylethylene sebacamide) with reactive amine endgroups. In a
similar reaction, a low molecular weight, telechelic liquid crystalline polyester
with reactive acid chloride ends was prepared from phenyihydroquinone and
terephthaloyl chloride.
Due to a stoichiometric imbalance in the
copolymerization reaction, an ABA type block copolymer composed of flexible
polyamide blocks (B) and rigid LCP blocks (A) has successfully been prepared
through a solution polycondensation reaction between the reactive endgroups
of the telechelic homopolymers. The existence of the block copolymer was
verified by a combination of infrared spectroscopy, differential scanning
calorimetry, optical microscopy, elemental analysis and solvent extraction. In
each of these techniques, the results of the block copolymer were quite different
from those of a physical blend of the two homopolymers.
REFERENCES
1.
2.
3.
4.
B. C. Auman and V. Percec, Polyer, 29, 938 (1988).
M. Takayanagi, Pure Anal. Chem., 55(5), 819 (1983).
W-F. Hwang, D. R. Wiff, C. Verschoore, G. E. Price, T. E.
Helminiak and W. W. Adams,
y- E
, 23(14), 784 (1983).
M. Takayanagi, T. Ogata, M. Morikawa and T. Kai, J. Macromol.
Sci..Phs., B17(4), 591 (1980).
306
AGGREGATION STRUCTURE AND ELECTRO-OPTICAL PROPERTIES OF
(LIQUID CRYSTALLINE POLYMER)/(LOW MOLECULAR WEIGHT
LIQUID CRYSTAL) COMPOSITE SYSTEM
TISATO KAJIYAMA, HIROTSUGU KIKUCHI, AKIRA MIYAMOTO, SATORU MORITOMI AND
JENN-CHIU HWANG
Department of Applied Chemistry, Faculty of Engineering, Kyushu University,
Hakozaki, Higashi-ku, Fukuoka 812, Japan
ABSTRACT
A series of thin films composed of liquid crystalline polymer (LCP)
and low molecular weight liquid crystal (LMWLC) was prepared by a solventcasting method or by a bar-coating method. LCPs were of mesogenic side
chain type with strong or weak polar terminal groups in the side chain
portion.
A mixture of smectic LCP (LCP with side chain of strong polar
end) and nematic LMWLC formed a smectic phase in a LCP weight fraction
range above 50 %. Also, a mixture of nematic LCP (LCP with side chain of
weak polar end) and nematic LMWLC with strong polar group induced a new
smectic phase in a LCP molar fraction range of 20-80 %.
Reversible and
bistable electro-optical effects based on light scattering were recognized
for a smectic phase of a binary composite composed of LCP and LMWLC.
A
light scattering state caused by many fragmented smectic lamellae appeared
in the case of application of an a.c. electric field below a threshold
frequency (,1 Hz).
Furthermore, application of a 100 Vpp s.c. field of
I kHz made the transmission light intensity increased to 94 2 within a few
seconds. The optical heterogeneity in a smectic layer composed of the side
chain group of LCP was caused by the difference of two forces based on both
dielectric anisotropy of the side chain and electrohydrodynamic motion of
the main chain. Since application of a low frequency electric field causes
an ionic current throughout the mixture film, it is reasonable to consider
that an induced turbulent flow of main chains by an ionic current collapsed
a fairly well organized large smectic layer into many small fragments,
resulting in an increase in light scattering. The response speed of LCP
upon application of an electric field increased remarkably by mixing LMWLC.
In the case of a smectic mesophase, turbid and transparent states remained
unchanged as it was, even though after removing an electric field.
Such a bistable and reversible light switching driven by two different
frequencies
of
electric field could be newly
realized
by
both
characteristics of turbulent effect of a well organized large smectic layer
of LCP and rapid response of LMWLC. We believe that the LCP/LMWLC mixture
system is promissing as a novel type of "light valve" exhibiting memory
effect (bistable light switching).
Key
Words:
polymer/(liquid crystal) composite, self-supported
liquid
crystalline film, light-intensity control, bistable light
switching, memory light valve, large area display
INTRODUCTION
Recently, functional characteristics of low molecular weight
liquid
crystals (LMWLCa) have been studied in many fields because of their unique
orientation behavior and hydrodynamic properties. Orientation of nematic
LMWLCs is easily controlled by applying electric or magnetic fields.
The
Mat. Res. Soc. Symp. Proc. Vol. 171. 11990 Msaterils Research Society
306
molecular axes of LMWLCs with positive dielectric anisotropy orient along
the direction of an applied electric field.
The authors have reported on preparation methods of polymer/LMWLC
composite films as novel permselective membranes[1-7 1. PolymerLMWLC
composite films, in which LMWLC is embedded in a three dimensional network
of polymer matrix, exhibit a distinct jump in permeability since an abrupt
change in thermal molecular motion occurs at the crystal-liquid crystal
phase transition temperature, TKN. The characteristic orientation and the
hydrodynamic properties of nematic LMWLC materials can be applied for the
control of permeation of molecules or ions[8-15). The authors have also
demonstrated that K+ permeation through a crown-ether-containing composite
film is largely governed by the dispersion state (homogeneous or phaseseparated) of the immobilized crown ether[16-18].
Thermotropic liquid crystalline polymers with mesogenic side chain
groups have both characteristics of polymer and liquid crystal.
Recently,
electro-optical
properties
of thermotropic LCP have
been
studied
extensively. Since LCP in a mesomorphic state are more viscous than LUWLC,
the magnitude of response time of LCP to an external stimulation such as
electric or magnetic fields is much greater than that of LMWLC. Reversible
and
bistable electro-optical effect based on light scattering
was
recognized for a smectic phase of a binary composite composed of LCP
and
LMWLC[19-211.
Also, various types of polymer/LMWLC composite system have
been reported as large-area and flexible light-intensity controllable films
(light valve)[22-27].
In this study, large-area and flexible self-supported LC films in a
smectic liquid crystalline state were prepared from a mixture of side chain
type LCP and LMWLC. Aggregation state, phase transition behaviors and
reversible and bistable electro-optical effect of the mixture system have
been investigated.
EXPERIMENTAL
The chemical structures of the side chain type LCP and LMWLC are given
in Figure 1. LCP with side chain of strong polar end was PCPHS of which
the main chain of poly(methylsiloxane) and the side chain of mesogenic
group were linked by a flexible aliphatic spacer (m=6). PCPHS was used in a
smectic liquid crystalline state. Also, LCP with side chain of weak polar
end was PMPPS of which liquid crystalline state was nematic. LMWLCs used
were CPHOB and 50CB. CPHOB have a chemical structure similar to that of
the side mesogenic group of PCPHS.
The phase transition behaviors and the aggregation state of the
LCP/LMWLC composites were investigated on the basis of differential
scanning calorimetry (DSC), polarizing optical microscopy (POM) and X-ray
diffra:tion study.
In
order to investigate electro-optical effect, the
LCP/LMWLC
composites were sandwiched between the two ITO-coated glass plates which
were transparent electrodes. The distance between the two electrodes were
.
maintained by the PET spacer of 4 jm thick with a hole of about 0.16 cm
Under an a.c. electric field, a change of the transmission light intensity
through the cell was detected by means of a photodiode and was recorded
with digital storage oscilloscope. The incident light source was He-Ne
laser.
307
1. liquid crystalline polymer
(a) m=3 poly(4-cyanophenyl 4'-propyloxy
benzoate methyl slloxane) (PCPPS)
S-36 I-I (Tg=293K)
m=6 poly(4-cyanophenyl 4'-hexyloxy
benzoate methyl sliloxane) (PCPHS)
K-297-S-421-1 (Tg=260K)
~H3
Si-O
0
(b) poly(4-methoxyphenyl 4'-propyloxy
benzoate methyl siloxane) (PMPPS)
N-334-1 (Tg=288K)
H
Si-O0
C,),_OOC.O-CH
11
0n
Figure 1 Chemical structures
of liquid crystalline polymers
and low molecular weight
3
liquid crystals.
2. liquid crystal
(a) 4-cyanophenyl 4'-hexyloxy benzoate
CHy-(CH 2 )-0-GC-00GCN
(CPHOB)
K-339-N-354-1
0
(b) 4-cyano 4'-pentyloxy biphenyl
C5Hit- 0
-CN
(50CB)
K-321-N-341-I
RESULTS AND DISCUSSION
I. The composite system composed of PCPIIS/CPHOB
Figure 2 shows the phase diagram of the PCPHS/CPHOB composite which
was obtained on the basis of DSC, POM and X-ray studies.
The glass
transition temperature, Tg of PCPHS decreased with an increase of CPHOB
ratio.
This decrease in Tg might be caused by plasticizing effect of
CPHOB.
Endothermic peaks attributed to
both the crystal-mesophase
transition, TKM of the side chain of PCPHS and that of CPHOB were observed
in temperature ranges of around 295-305 K and 315-335 K, respectively.
This result suggests that the composite forms the phase-separated structure
T
below
KM of CPHOB, TKM(M), that is, the PCPHS and the CPHOB rich phases.
Since
only one endothermic peak attributed to the mesophase-isotropic
transition, TMI was observed, the composite forms homogeneously mixed
mesomorphic phase (molecular dispersion state) above TKM(M).
Therefore,
CPHOB is miscible over a whole concentration range ot PCPHS in both
isotropic and mesomorphic states. These conclusions were also confirmed by
POM observation.
In order to investigate the aggregation state of the
PCPHS/CPHOB composite, the X-ray diffraction studies were carried out at a
temperature range above TKM(M) and below TMI (denotej by "(B) Mesophase" in
Figure 2). The sharp low angle X-ray diffraction 'ith a d-spacing of about
3.06 nm corresponding to a smectic layer spacing, was observed in a weight
fraction range of CPHOB below 40 wt%. In the case above 60 wt% of CPHOB,
this sharp X-ray diffraction disappeared and the diffuse diffraction was
observed because the mesomorphic state of the comosite was nematic. In the
306
10__
_PCPHSICPHOB
- T =- 3 K
\10T
f=lkHz
P.P
400
•
r
by P#4
-PO
DSC
"Moby
o
10
Ub
( SA) eaeoPt
( a)
\
1
W~
40160
hao
TK1
a
10-
o
\
10~
(E) Cystlln(P+M)
\20 80
1,300
a Crystlfte
_M.__D__01100_
(P)
n
In
(D
0
()
I
..
1
S(0 0I.sy(P)#Cryar-,-(P#+M)
0.5
Weight fraction of CPHOB
Figure 2 The phase diagram of the
PCPHS/CPHOB composite system.
1.0
10
100
VoltagelV
Figure 3 applied voltage
dependence of TR for the
PCPHS/CPHOB composites
Tc is the pahse transition
temperature between mesomorphic
and istropic states.
cases of PCPHS and the PCPHS/CPHOB-(80/20) composite, the weak second order
diffraction corresponding to a smectic layer was also observed. The second
order diffraction intensity of the (80/20) composite became stronger than
that of PCPHS. These results indicate that mixing CPHOB of about 20 wt% to
PCPHS makes the regularity of smectic layer increased. Then, a smectic
phase certainly existed in a mesophaae region below the CPHOB fraction of
40 wt% and in the case above the CPHOB fraction of 60 wt%, the mesophase of
the composite was nematic.
We investigated an electro-optical effect of the PCPHS/CPHOB mixture
system based on birefringence. The composite film was sandwiched between
the two ITO-coated glass plates of which the surfaces had been rubbed in
one direction to obtain a homogeneous alignment of the mesogenic group. An
application of s.c. electric field of I kHz to the mixture film resulted in
a decrease of transmission light intensity under the crossed polarizers due
to reorientation of the mesogenic group from a homogeneous alignment to a
homeotropic one. The response time, TR, was defined as the time period for
which transmission light intensity dropped to 50 % after applying an
electric field. Figure 3 shows log TR-log V plots for various PCPHS/CPHOE
mixture systems with different mixing ratios. TR increased remarkably by
mixing CPHOB. The response time was inversely proportional to the 2nd power
of the magnitude of electric field. This response characteristic indicates
that driving force for reorientation of the mesogenic group of crystalline
309
PCPHS/CPHOB Mixture Systems
(1)80/20 (Smectic)
Memory Effect
200V 0.1Hz
64%
10sec
6%
200V lkHz
7.5rrin
Memory Effect
(2) 60/40 (Smectic)
100V 1Hz>
Isec
94%
5
100V lkHz
Figure 4 Change of turbid-
3sec
(3) 50/50
94%
3W
3ov 1k
transparent states and riseresponse times for the
No Memory Effect
43%
1kNZ
ff
Off/i
D
/6'3ov D.C.
D
smectic and nematic PCPHS/
CPHOB composites under
various conditions of an
a.c. and d.c. electric fields.
700ms
87% /
600ms
No Memory Effect
(4) 40/60 (Nernmtic.
95%
-D..
30V D.C.
30V 1kHz
l0Oms
40%
84%
50ms
molecules is attributable to the dielectric anisotropy of a LC molecule.
The electro-optical effect based on light scattering was also studied
The samples were
under various conditions of an s.c. electric field.
between
the two non-treated
ITO-coated
glass
plates.
sandwiched
Transmission light intensity of He-Ne laser through the mixture film of
PCPHS/CPHOB without any optical polarizers was measured by photodiode. The
distance between the mixture film and the photodiode was 305 mm. In this
measurement, the rise-response time was conventionally defined as the time
period being necessary for a change from 10 % to 90 % of transmission light
intensity. The as-cast PCPHS/CPHOB (60/40) mixture system of 4jim thick in
As
a smectic state transmitted 86 % of an incident light of He-Ne laser.
shown in Figure 4, the transmittance of incident light strikingly decreased
to 5 % after applying an a.c. 100 V of 1 Hz due to a remarkable increase
of light scattering. Furthermore, application of an a.c. 100 V of 1 kHz
made the transmittance increased to 94 % within a few seconds. The degree
of light intensity difference (contrast) between light scattering (turbid)
and non-scattering (transparent) states could be changed reversibly by
imposing an electric field with different frequencies. Even if an electric
field was removed, each turbid and transparent state remained unchanged as
In the case of
the
it was, indicating bistable light switching.
PCPHS/CPHOB (40/60) mixture film in a nematic state, the response speed
became much faster than that for the PCPHS/CPHOB (60/40) mixture film in a
smectic state. However, both the transparent and turbid states could not
That
be maintained without continuous application of an electric field.
is, there was no memory effect in the case of the PCPHS/CPHOB mixture film
in a nematic state.
The proposed molecular aggregation states are
transparent
schematically illustrated in Figure 5 for the turbid and
cases. Since an application of a low frequency electric field induces an
310
Reversible .3 Bistable Light Switching
by (Liq. Cryst. Polym.)/( Low Mol.Wt. Liq. Cryst.) Comp. Systems
> Low Freq. <
(ionic current)
As-cast
Liq. Cryst. Polym.
->
""s
High Freq.
(no ionic current )
0,0
000
0
101010100
0010010
:~1
'tooi 10 1ooto |
LowMoWt.Lq.Cryst.
(1) Random
T=
86%
(2)Turbulent Flow
5 e%
(3)Homeotropic
94%
Figure 5 Schematic illustration of turbid and transparent
states for the PCPHS/CPHOB composite system under different
frequencies of an a.c. electric field.
ionic
current throughout the mixture
film, it
is reasonable to
consider
that an induced turbulent flow by an ionic current collapsed a fairly well
organized large smectic layer into many small fragments. This induces an
increase in light scattering and also, a decrease of transmittance up to 5
% (from (1) to (2) in Figure 5). Since a high frequency electric field
does not induce an ionic current, a large scale homeotropic alignment of
smectic layer is easily formed by dielectric characteristics of the
mesogenic group of the composite, increasing the transmittance up to 94 %
owing to a remarkable reduction of spatial distortion of smectic directors
and/or optical
boundaries (mismatch of refractive indices) (from (2) to
(3)).
Such a bistable and reversible light switching driven by two
different frequencies could be newly realized by both characteristics of
turbulent effect of LCP main chain and rapid response of LMWLC. We believe
that the LCP/LMWLC mixture system is useful as a novel type of "light
valve" exhibiting memory effect (bistable light switching).
2. The composite system of PMPPS/50CB
PMPPS is nematic liquid crystalline polymer with side chain of weak
polar end. It has been reported that a binary mixture of nematic LMWLCs
with both strong and weak polar ends gives an induced smectic phase.
In
tilisstudy, it has been investigated that this rule can be applicable to a
binary mixture of side chain type liquid crystalline polymer (LCP) with
side chlain of weak polar end and LMWLC with strong polar group, which
individually exhibit a nematic phase. LCP was polysiloxane which attached
I terminal weak polar mesogenic unit of side chain (PMPPS) and LMWLC was
',tL:11.
The chemical structures of PMPPS and 5OCB are shown in Figure
1.
i
cliirocteristic X-ray diffraction patterns of nematic PMPPS and
50CB
w.r,
hroad
and obscure as shown in Figure 6.
The PMPPS/5OCB (50/50 in
311
0 DSC
38
1 P= PMPPS
360
-O.44nm
- O.44nm
Ts
:
(~5je
,-2.6gnm
-(C
:'. .320
PMPPS
0 POM
-lorol
M=5CB
50CB
E
300
~(F)esophasl(P)
Cyst0lM
j
-0.44nm
n
2.0 m
"-4.2nm
PMPPS6OCB
I
280
1.3m
260
[
Tg
240
t
I
t
50/50
PMPPS
Figure 6 Schematic representation
of x-ray diffraction patterns of
PMPPS (nematic), 50CB (nematic)
and the PMPPS/5OCB (50/50 mot %)
(induced smectic) composite.
50
50CB
50CB mo%
Figure 7 The phase diagram of the
PMPPS/50CR composite.
mol%) composite exhibits sharp Debye rings which indicate an induced
smectic phase.
A d-spacing of 4.2 nm may correspond to a repeating
distance of a mesogenic side group composed of PMPPS and 50CB. The sharp
X-ray diffraction rings resulting from a smectic layer were observed for
the PMPPS/50CB composite systems of which the fraction ranges 80/20-20/80
(mol%).
Therefore, PMPPS is fairly miscible with 50CB and the binary
mixture exhibits an induced smectic phase over a wide range of component
fraction.
The phase diagram of PMPPS/50CB of Figure 7 was obtained on the basis
of DSC and POM studies. Tg of LCP decreased from 288 K to 244 K by
addition of 50CB. A decrease of Tg may be caused by a plasticizing effect
of 50CB. The narrow biphasic nematic-isotropic region of about 3 K wide,
where the nematic and isotropic phases coexist, could be recognized by POM
observation. An induced smectic phase appears in a range of 80/20-20/80 of
PMPPS/5OCB.
The electro-optical effect of the PMPPS/5OCB mixture based on light
f an a.c.
scattering was also investigated under various conditions
electric field. Figure 8 shows the electro-optical effects of PMPPS, 50CB
and the PMPPS/50CB (50/50) composite. In the cases of electro-optical
measurements of PMPPS and 50CB, there have been observed no distinguishable
optical change and a little change of transmittance of about 8 % upon onand off- electric fields, respectively. On the other hand, the PMPPS/50CB
312
*
*
,
01Hz
,
Figure 8 Electro-optical
effect of PMPPS, 50CB and
the PMPPS/5OCB (50/50)
composite under low and
high frequencies of an
100.
5
o PMPPS
at
'
,
The applied voltage was
200 Vp-p and the cell
thickness was 16 jm.
50
A
0
0
PMPPSIOC:
,
,
20
0
650
10
20
,
0
10
20
Timelsec.
4
(50/50) composite exhibited a bistable and reversible light switching
driven by two different frequencies of an a.c. electric field in a similar
fashion to the PCPHS/CPHOB composite system. In a low frequency range of
about 0.01 Hz, an application of an electric field above the threshold
voltage induced strong light scattering state due to a remarkable increase
of optical heterogeneity being attributable to many collapsed fragments of
smectic layer.
This might be caused by a strong turbulent flow of main
chain by ionic materials. Also, in a high frequency range above 10 Hz, the
transmittance increased up to 98 % since a well organized homeotropic
alignment was formed by dielectric characteristics of the composite.
The
reversible turbid- transparent change upon an application of electric field
of low and high frequencies, respectively, can be explained schematically
by Figure 5. Both transparent and turbid (light scattering) states could
be stored stably, eventhough an electric field was turned off.
Both rise
and decay response times were in a range of several seconds.
The above results indicate that the rule of an induced smectic phase
to a binary LMWLC mixture is also maintained to the PMPPS/5OCB mixture
system and an induced smectic phase of the PMPPS/LWLC composite system
makes us expect that its electro-optical effect is promissing as a novel
type of memory display (bistable light switching).
REFERENCES
I T. Kajiyama, Y. Nagata, E. Maamura and M. Takayanagi, Chem. Lett.,
1979, 679(1979).
2 T. Kajiyama, Y. Nagata, S. Washizu and M. Takayanagi, J. Membrane Sci.,
II, 39(1982).
3 . Washizu, T. Kajiyama and M. Takayanagi, J. Chem. Soc. Jpn., Chem.
Ind. Chem., 1983, 838(1983).
4 S. Washizu, I. Terada, T. KaJiyama and M. Takaysnagi, Polym. J. 16,
307(1984).
5 T. Kajiyama, S. Washizu and M. Takayanagi, J. Appl. Polym. Sci., 29,
3955(1984).
m4
313
6 T. KaJiyama, S. Washizu, A. Kumano, I. Terada, M. Takayanagi and
S. Shinkai, J. Appl. Polym. Sci., Appl. Polym. Symp., 41, 327(1985).
7 H. Kikuchi, A. Kumano, T. Kajiyama, M. Takayanagi and S. Shinkai, J.
Chem. Soc. Jpn., Chem. Ind. Chem., 1987, 423(1987).
8 Y. Ohmori and T. Kajiyama, J. Chem. Soc. Jpn., Chem. Ind. Chem., 1985,
1897(1985).
9 T. Kajiyama, S. Washizu and Y. Ohmori, J. Membrane Sci., 24, 73(1985).
10 A. Kumano, 0. Niwa, T. Kajiyama, M. Takayanagi, K. Kono and S. Shinkai,
Chem. Lett., 1983, 1327(1983).
11 H. Kikuchi, M. Katayose, A. Takahara, S. Shinkai and T. Kajiyama,
Kobunshi Ronbunshu, 43, 669(1986).
12 T. Kajiyama, H. Kikuchi and S. Shinkai, J. Membrane, Sci., ?6,
243(1988).
13 T. Kajiyama, H. Kikuchi, M. Katayose and S. Shinkai., New Polym. Mater.,
1, 99(1988).
14 T. Kajiyama, J. Mqcromol. Sci.-Chem., A25. 583(1988).
15 H. Kikuchi, J. Hattori, Y. Mori and T. Kajiyama, Ka~aku Kogaku
Ronbunshu, 15, 617(1989).
16 S. Shinkai, K. Torigoe, 0. Manabe and T. Kajiyama, J. Chem. Soc., Chem.
Commun., 1986, 933(1986).
17 S. Shinkai, S. Nakamura, K. Ohara, S. Tachiki, 0. Manabe and
T. KaJiyama, Macromolecules, 20, 21(1987).
18 S. Shinkai, K. Torigoe, 0. Manabe and T. KaJiyama, J. Am. Chem. Soc.,
109, 4458(1987).
19 §. Moritomi, A. Miyamoto, H. Kikuchi and T. Kajiyama, ..pt. Progr.
Polym. Phys. Jpn., 31, 191(1988).
20 T. Kajiyama, H. Kikuchi, A. Miyamoto an,' Y. ',orimura, "Frontiers of
Macromolecular Science", IUPAC 505(1989).
21 T. Kajiyama, H. Kikuchi, A. Miyamoto, S. Moritomi and J.C. Hwang, Chem.
Lett., 1989, 817(1989).
22 H.G. Craighead, J. Cheng and S. Hackwood, Appl. Phvs.
. 40,
22(1982).
23 J.L. Fergason, SID Int. Symp. Dig. Tech., 16, 68(1985).
24 P.S. Drazic, J. Appl. Phys., 60, 2142(1986).
25 J.W. Doane, N.A. Vaz, B.-G. Wu and S. Zumer, Appl. Phys. Lett.,48,
27(1986).
26 H. Kikuchi, A. Miyamoto, A. Takahara, T. Furukawa and T. Kajiyama,
Prepr. 2nd SPSJ Int. Polym. Conf. 33(1986).
27 T. Kajiyama, A. Miyamoto, H. Kikuchi and Y. Morimura, Chem. Lett., 1989,
813(1989).
PART V1I
Interfaces/ Mechanical Properties
317
DIBLOCK COPOLYMERS AT SURFACES*
Peter F. Green
Thomas M. Christensen*
Sanda National Laboratories
Albuquerque N.M. 87185-5800
Thomas P. Russell
Spiros H. Anastasiadis
IBM Research Division
Almaden Research Center
650 Harry Road
San Jose, California 95120-6099
Abstract
The surface properties of symmetric microphase separated diblock copolymers of polystyrene
(PS) and polymethylmethacrylate (PMMA) were investigated using X-ray photoelectron
spectroscopy (XPS), the specular reflectivity of neutrons and se- -ndary ion mass spectrometry
(SIMS). PS, the lower surface energy component, exhibited a preferential affinity for the free
surface. For copolymers that are far from the bulk microphase separation transition (MST), the
surface consists of a layer of pure PS. When the system is close to the MST the surface is a
-1 r2
dependence,
mixture of PS and PMMA. The PS surface excess can be described by a N
where N is the number of segments that comprise the copolymer chain. It is shown that the
surface undergoes an ordering transition at a temperature T, that is above that of the bulk MST.
The ordering of the bulk lamellar morphology is induced by an ordering at the surface. This is
analogous to the ferromagnetic order observed in systems such as Gd at temperatures above the
bulk Curie temperature. The results here are discussed in light of previous work on copolymer
surfaces and in light of mean field theory.
----------------------
# Current Address: Dept. of Physics, Univ. of Colorado, Colorado Springs, Colorado 80933
Mal.1As. Soc- Symp. Proc. Vol. 171. 0190 Materials Research Society
318
Introduction
Diblock copolymers are technologically an important class of materials. They
behave like surfactants, analogous to the way in which soap molecules behave at oil water
interfaces. Because of this they are used to promote adhesion between the phases of immiscible
polymers and, hence to improve the mechanical properties of these systems. Due to their
unique surface properties block copolymers are used in a variety of biomedical and
microelectronic applications.
The inherent incompatibility between the components of a block copolymer chain
necessarily induces microphase separation at temperatures below the bulk microphase
separation transition (MST). Unlike A-B polymer-polymer mixtures the phases are unable to
grow very large because of the connectivity of the components. The resulting spatially periodic
phases that form exhibit varying symmetries (cubic, cylindrical, double diamond and lamellar)
that lower the free energy of the system [1-31. The symmetry of a given copolymer system is
dictated by the relative length of each constituent.The total number of segments, N, that
comprise each chain, determine the size of the phases. The bulk properties of diblock
copolymers are well understood. However, the surface properties of these materials are less
well understood.
It is well documented in multicomponent systems that the lowest surface energy
component preferentially segregates to a free surface in order to minimize the total free energy
of the system [4]. Block copolymers are no exception. Indeed, the surface segregation of
polymeric molecules is expected to be more severe than that which occurs in small molecule
systems since the combinatorial entropy of mixing in these systems is very small (-I/N).
Much of the early work on copolymer surfaces was performed using contact angle
[51 and surface tension measurements [6]. The data obtained using these techniques, though
useful, provided limited information. For example, information concerning concentration
profiles in the near surface region of the copolymers could not be obtained. More recent
techniques include X-ray photoelectron spectroscopy (XPS) [7-11], ion scattering
spectroscopy (ISS), transmission electron microscopy (TEM) [12,131, neutron reflectivity
[14,15] and secondary ion mass spectrometry (SIMS) [161. These techniques provide an
additional means by which polymer surfaces may be better understood.
The experiments on copolymer systems, using TEM, XPS, contact angle and
surface tension measurements, showed that the lower surface energy component of the
copolymer chain exhibits a preferential affinity for the free surface. In some systems, such as
the polystyrene/polydimethyl siloxane (PS/PDMS) [10] system, XPS studies showed that the
lower surface energy component, PDMS in this case, completely covered the surface. In other
systems, such as the PS/polyethylene oxide (PEO) [81 system, both phases were observed to
coexist at the surface. The extent of the segregation in these and all other systems studied
previously can not be correlated with the surface energy differences between the components as
one might anticipate. There have been a number of reasons presented to explain this
behavior.The morphology of the system could play a crucial role. In the case of the PS/PEO
system, PEO crystallizes upon evaporation of the solvent used to cast the films and the
crystallization kinetics, evidently, affects the mobility of the PS chains thereby prohibiting the
PS chains from migrating to the surface. While this may be true of systems where one
component crystallizes, there are other cases, such as the bisphenol-A-polycarbonate
(BAPC)/PDMS system [7], in which both components are observed at the surface. Gaines [171
suggested that the criterion for liquid spreading should explain the descrepancy of whether the
surface is covered by one component or both. We will show later that this criterion is
inadequate for a description of copolymer surfaces.
Here the general question of what controls the surface composition of diblock
copolymers is addressed. Why in some cases do both components appear at the surface while
in others only one component is observed? It is also shown that in this diblock system the
interactions between the molecules in the near surface region are modified such that they are
weaker than in the bulk. The near surface region is shown to order at a temperature above the
bulk MST. This phenomenon has been observed in rare earth metals that exhibit ferromagnetic
order [18,19].
319
Experimental
A series of symmetric diblock copolymers of PS and polymethylmethacrylate
(PMMA), each comprised of N segments, where N ranged from 275 to 5600, were studied.
Further details about the polymers may be found elsewhere [21]. These copolymers were
chosen since they all exhibit a lamellar morphology below the bulk MST and the glass
transition temperatures of PS and PMMA are similar. The condition for microphase separation,
which varies depending on the bulk morphology, is, for a lamellar phase, XN=10.5, where X
is the Flory-Huggins interaction parameter [201.
The bulk properties of these systems were well characterized using small angle
X-ray scattering (SAXS) [21,221, secondary ion mass spectrometry (SIMS) [ 161 and neutron
reflectivity [14,151. The surface and near surface properties of these systems were
characterized using XPS [23,24] and neutron reflectivity [14,151.
XPS
The XPS measurements have been described in detail elsewhere [23,24], therefore
the description will be brief. High resolution XPS spectra were obtained using a Kratos
XSAM800 spectrometer (fixed retardation ratio mode) and a hemispherical analyzer with a high
resolution band pass energy of 10eV. A non-monochromatized 300 watt MgKa source was
used.
Samples were prepared using three different procedures. Films -5gam thick were
prepared using solvent casting techniques. In some cases the mutual solvent , toluene, was
allowed to evaporate rapidly in air while in other cases the solvent was allowed to evaporate
slowly. Slow evaporation of the solvent was accomplished by placing films in an environment
where the vapor pressure of toluene was high. Another series of films were prepared by first
allowing the solvent to evaporate and then annealing them at 170 0C for many hours so that they
reached equilibrium.
Spectra were recorded at normal angles to the sample surface and at 700, 450 and
200 with respect to the surface normal. The spectra of pure PS and of pure PMMA served as a
basis for the analysis of the PS-PMMA copolymers. The spectra obtained were in good
quantitative agreement with those found in the literature [251.
Following X-ray satellite subtraction and the subtraction of the background, the
spectra were fitted using Gaussian line shapes with a FWHM of 1.5eV. Figure la shows the
Ols doublet peaks of pure PMMA. The solid line drawn through the data is a convolution of
two peaks, one at energy 533.8 eV and the other at 535.4 eV, representing the C=O and C-O-C
bonding environments, respectively. Figure lb shows a typical Ols doublet peak of a
PS-PMMA copolymer. The relative fractions of both components at the surface were
determined using two independent procedures which yielded results that were in excellent
agreement. The first procedure was to divide the area of the Ols profile of the copolymer with
that of pure PMMA; this yielded the surface fraction of PMMA. The second procedure
involved recognizing that the contribution of PS to the C Is profile is twice that of PMMA and
that the fraction of PS plus the fraction of PMMA is unity.
Neutron Reflectivity
The technique of neutron reflectivity is briefly described below since its use in the
area of polymers is new. A more detailed description may be found in the literature 114, 15,
26, 27]. Consider the diagram shown in Figure 2 where neutrons impinge on a sample surface
at glancing angles, 0. Neutrons are reflected at the corresponding angle e.The neutron
momentum normal to the sample surface is
2x sin 0
(I)
where X is the wavelength of the neutrons. If the neutrons penetrate a medium characterized by
320
Fiur
1:
a
omlage(0=
covluin ftw
Fiur
1:pa)aNormrmalangle
Figure 2:
X
SOspofl
ymerc
0 IP
fP
eksrpesnin
s profile of
M
.Ti
oi
iei
heC0C n CObodn
PMMAly.hsoldin
Schematic of th,- Neutron Reflectivity experiment.
321
a scattering length density of (b/V)j then the neutron momentum in the medium is
1/2
kjz= [kO~z - 4x V)i](2)
We may now consider the case where the scattering length density in the film
varies as a function of depth in the film. Here the film is divided intoj layers, each of thickness
d and of scattering length density (bIV)j; the substrate is designated as the j+l th layer. d,
should be sufficiently thin in order to closely approximate the varying scattering length density
of the real system, If the reflectance, or, equivalently, the reflection amplitude, from an
interface is r'JJ+,
where r',j+i= (kj, -kj. 1 ,,)/(kj.,+kj..hz), then the reflection coefficient of the j
th layer is
rJ
r'. + r
exp(2 id.k.)
1 + rfh r'jj+
1 exp(2 i dJk.)
(3)
This recursion relation can be used to calculate the reflection coefficient of the j- I th layer by
replacing j- I with j-2 and j with j- 1.The recursion is continued until the reflection coefficient.
ro.1,
at the free surface is obtained. The reflectivity is
(4)
R(ko) = r0. r0.
One obtains from this experiment the reflectivity, R, as a function of the neutron momentum,
k0 , which can be compared to the experimentally determined profile.
The reflectivity measurements were performed at the National Institute of
Standards and Technology. The neutrons were of wavelength X=2.35A with AX/X--0.02. The
reflectivity profiles were obtained by varying 0 in order to change the neutron momentum.
More details of the experiment may be found elsewhere [14,15].
Results and Discussion
Bulk properties of the Copolymers
All the copolymers studied here exhibit a lamellar morphology in the bulk. This
has been verified using SAXS 1221, SIMS [161 and neutron reflectivity [14,151. When films of
these copolymers are cast from toluene, a mutual solvent, the lamellae are randomly oriented,
as shown by ion beam analysis measurements [16,21,221. The degree of orientation of the
lamellae improves if the solvent evaporation rate is very slow. Under equilibrium conditions
the lamellae are aligned parallel to the sample/substrate and air(vacuum)/sample interfaces, as
shown by SIMS and reflectivity. The kinetics of the alignment process has been studied using
SIMS and neutron reflectivity.The alignment process begins at the interfaces and then
propagates through the bulk. The kinetics of the alignment process is a strong function of N,
the rate decreases as N increases.
It is also clear from these measurements that for a copolymer with an interlamellahr
spacing of L and volume fractions of PS and PMMA of 0.5. the thickness of the layer'
adjacent to the interfaces is L/4, see Fig. 3. The width of the interfaces between the phases is
E=50A and is independent of N. Fig. 4 shows a typical reflectivity profile of a microphasc
separated copolymer. The inset in the figure shows the scattering length density profile used to
calculate the reflectivity profile, represented by the line in the figure. The SAXS and reflectivity
measurements show that L increases with NO, as shown in Fig. 5. The solid line drawn
through the data has a slope of 2/3 and the broken line 1/2. The value of a, obtained from the
slope, should, in principle, indicate the degree of segregation in these systems, ca= 1/2for a
322
AIFVVACUUM
1000-
1ligUre 4:
Typaic refe
t profile
12
4
I-g
yialre fl'ei it
cofa my eicro s separ at e
uibriumcpo.
o
-
rofaleofar spacng ase efunc
tionof N.o\ier
boc
I1
323
weakly segregated system and a=2/3 for a highly segregated system. However the limited
range of N precludes a precise determination of this exponent.
A very important question, related to the subsequent discussion on surfaces.
concerns the proximity to criticality (MST) of some of these systems. An indication of this can
be obtained by looking at the ratio 2E/L, the volume fraction of the sample occupied by the
interface, as a function of N (Figure 4). For the N=275 chain the ratio is 0.6 and for the
longest chain 0.1. Based on measurements of the scattering length densities, the shorter
copolymer chain systems were found to be phase mixed. This is strong evidence that these
shorter chain systems are weakly segregated, ie. close to the MST. The much longer chains are
in the strong segregation limit. In principle one could determine the degree of segregation by an
analysis based on the value of a. The analysis of the 2E/L ratio and the fact that the
components are phase mixed is a much more accurate indicator of the proximity of the system
to the MST than the determination of the exponent a.
Surface Ordering in Diblock Copolymers (Theory)
Recently Fredrickson [27] developed a theory that addresses surface ordering in
diblock copolymer systems near the weak segregation limit (2E/L is large). The system is
considered to exhibit a lamellar morphology below the bulk below the MST. An expansion of
the free energy is made in terms of an order parameter W(r), which is the local deviation of the
average density of the A component from its bulk value. The perturbing effect of the surface o1n
the bulk is expressed in terms of a "bare" surface energy term, expressed in terms of a surfa,c
free energy density
f,(Vw)= (-H 41(r) + a 42kr)/21N)
(5)
H and a are phenomenological parameters that describe the effects of the surface potential on
the bulk thermodynamic potential. Both H and a are defined such that they are proportional to
N t27]. H is a field that may be related to a chemical potential, or surface energy difference,
that favors one component at the surface; it is assumed to be small. In the PS/PMMA systeill
the surface energy difference is less than I dyne/cm 1291. The parameter a describes the way in
which the surface potential modifies the interactions between the molecules in the near surface
region.
It is often justified to neglect the effect of surfaces on the bulk properties of
materials since the number of atoms at a surface is considerably lower than that in the bulk.
Near critical points, however, where the correlation length, or, equivalently, the concentration
fluctuations, become very large, decoupling of the surface and bulk properties may not be
justified. An example of this occurs when a>0. Here the local interaction parameter, in the
vicinity of the surface, XIto,<Xbulk, where XbuI is the interaction parameter in the bulk. In this
case the surface orders at a temperature, Ts, above Tb, that of the bulk (X-I/T). The ordering at
the surface can then induce ordering in the bulk. On the other hand the interactions at the
surface may be sufficiently enhanced that a<O or a=0. For a<O, Xi.oIa>XbUlk (Tb>T,) and the
bulk may induce ordering at the surface. For a--O, the surface and bulk order spontaneously.
There are analogies between polymer surfaces and Ising ferromagnets 1281. In
surface critical behavior of some systems that exhibit magnetic order it is customary to refer t)
a surface extrapolation length, X [181. In the case of polymer surfaces X=B/a, where
Bt/2~NI/2 is a measure of the range of interactions between the molecules in the systeitt.
Physically, X represents the depth of the region beneath the surface where effects due to the
surface potential are important. Xshould be on the order of a statistical segment length.
The manner in which the N dependence of the surface excess, W, of the more
surface active species, was calculated from Fredrickson's work was discussed previousl'
123,241. It was shown that
324
A-13 N-tn
(6)
2
The coefficients A=HAIB=H/a and 13=2.7A X. It is clear that the slope of a Wl versus N-1
plot should show how the interactions in the vicinity of the surface are modified.
Surface Composition (Experiment)
The surface composition of the copolymer films was found to be controlled by the
rate at which the solvent evaporated during the preparation of the films, as shown in Figure 6.
The triangles represent the PS surface fraction, 0t, of the films produced by the rapid solvent
evaporation process and the circles represent that of the samples produced using the slow
solvent evaporation process. The average surface fraction of PS obtained for the samples
produced using the slow solvent evaporation process is higher than that obtained for the other
solvent cast films. As mentioned earlier, the systems produced using a rapid solvent
evaporation process are characterized by randomly oriented microdomains [21,22] where some
are oriented perpendicular to the free surface. The systems produced by very slow extraction of
solvents, or annealed for long periods of time, are characterized by highly oriented lamellar
microdomains which are parallel to the free surface where a PS rich layer is preferentially
located. This explains why the average surface fraction of PS is higher for the films produced
using the slow solvent extraction process. The data in this figure also shows a trend where 0i
increases with N. This is strong evidence that as the degree of incompatibility increases the
driving force for surface segregation increases. Evidently the reduction of the surface free
energy that results from the surface segregation of the lower surface energy component to the
free surface more than offsets the concomitant increase in the bulk free energy (decrease in
combinatorial entropy). The total free energy of the system is therefore minimized.
Represented in Figure 7, by the triangles, is the surface excess of PS, Wv
- /2
(Wlt=Os-Ob' where Os is the surface composition and Obis the bulk composition) versus N 1
plot for the samples produced by the slow solvent evaporation process (near equilibrium
structures). The data scales well with a N-4 2 dependence. In the theory, this N dependence
arises from the fact that Bi/2,which is a measure of the interactions between the molecules
varies as Nt/2 .This dependence applies in the range of N where the theory should not apply
which suggests that the N-12 is not associated with criticality. An estimate of the slope may be
obtained if the constant term in equation 6 is set equal to 0.5, a value which the surface excess
should approach for very large N. Based on the slope it is clear that X,or equivalently a,is not
only positive but ~3. This suggests that the local X parameter in the vicinity of the surface is
lower than that in the bulk (Xtoat<Xbutk). Furthermore, the surface potential modifies the
interactions between the molecules at a depth of 3 segment lengths beneath the surface which is
entirely reasonable.
The circles in Figure 7 represent the V, versus N 1/2 dependence of films annealed
at 170VC for long periods of time under high vacuum conditions [311. At very large N, where
the system is in the strong segregation limit, the surface is composed only of PS, the lower
surface energy component. For smaller values of N where the systems are weakly segregated
the surface is composed of both PS and PMMA [30]. The difference in the relative values of
the surface excess for the annealed and solvent cast films could be associated with the way in
which the solvent alters the relative thermodynamic interactions between PS and PMMA.
Since X=3, then a>0 which suggests that Ts>Tb (X-l/T). Neutron reflectivity
icesurements on the shortest copolymer chain show that at a temperature above the bulk MST
the surface begins to order, Figure 8, while the bulk is disordered. The existence of oscillations
near the interfaces is evidence of surface ordering. In the bulk the oscillations are absent
indicating the absence of order. As the temperature approaches the bulk MST, order begins to
325
10
-
evaporation
slowsolvent
09
0
ps
.f
0.8
evaporation
solvent
0 7ast
06
0oo
0
0
2000
4000
3000
6000
6000
N
Figure 6:
The surface fraction of PS, D, as a function of N.
0.5
IT
0.2
0.1
0.0 0
0.00
Figure 7:
0.06
0.02
0.08
- r
from
Surface excess of PS, W, as a function of N 12. The circles represent data
samples that were annealed until they reached equilibrium and the triangles that o
samples that were produced using the slow solvent evaporation process.
6.0
< 4.0
2.0
0
400
800
1200
z(A)
Figure 8:
ISI
pS-PMMA copolymer abxve the bulk
Scattering length density profile of a
326
propagate through the bulk. This is clear experimental evidence that near a critical point the
surface orders and induces ordering in the bulk. This phenomenon has been observed in
ferromagnetic systems where the spins in the bulk and in the surface order at different
temperatures [ 19].
The application of the criterion for liquid spreading to the understanding of
copolymer surfaces is too simplistic. It suggests that a liquid, A, will not completely wet
another liquid, B, if the surface energy difference between them is smaller than the interfacial
energy between the liquids. For long chains, certainly in the range of the chain lengths used
here, the surface and the interfacial energies are independent of N. This criterion could not be
correct since both complete and incomplete coverage is observed in this system. Furthermore,
the main condition for applicability of this criterion is that the forces between the A and B
components should be primarily van der Walls forces ( no chemical bonds). In the case of
copolymers, A and B are covalently bonded!
Conclusion
It has been shown that in copolymer systems the lower surface energy component
exhibits a preferential affinity for the free surface whether the systems are prepared using
mutual solvents or are annealed at sufficiently long times to produce equilibrium structures. In
the case of the solvent cast films the surface composition is controlled, in part, by the rate at
which the solvent evaporates from the film. The surface excess of PS increases as the degree of
incompatibility between the segments of the copolymer increase for both solvent cast and
annealed films. For the annealed systems at equilibrium, at very large N (strong segregation
limit), the PS component completely covers the surface. At smaller N (weak segregation limit)
both PS and PMMA are found at the surface. For the solvent cast films that were near
equilibrium complete dominance of PS was not observed and this could be related to the fact
that the solvent altered the thermodynamic interaction between the PS and PMMA segments.
It was also demonstrated that the critical points of the surface and the bulk are
different. The surface orders at a temperature, Ts, above that of the bulk. As the temperature
approaches Tb the correlation length becomes very large, consequently, the bulk microstructure
orders.This is analogous to the ferromagnetic order observed in systems such as Gd 19 at
temperatures above the bulk Curie temperature.
*This work was supported in part by U. S. DOE under Contract DE-AC046-DP00789.
References
1I. Goodman, Ed., Developments in Block copolymers (Applied Science, New York, 1982).
2 L. Leibler, Macromolecules 13, 1602 (1980).
3 T. Ohta, K, Kawasaki, Macromolecules 19, 2621 (1986). A.N. Semenov, Soy. Phys.
JETP, 61, 733 (1985).
4 J. M. Blakley, in Chemistry and Physics of Solid Surfaces, Vol 2, edited by R. Vanselow
(CRC Press, Boca Raton, FL, 1979).
5 M.J. Owen and T.C. Kendrick, Macromolecules, 3, 458 (1970).
6 A.K. Rostogi and L.E. St. Pierre, J. Colloid and Inf. Sce. 31, 168 (1969).
7 R.L. Schmitt, J.A. Gardella Jr., J.H. Magill, L. Salvati, Jr., and R.L. Chin.
Macromolecules 18, 2675 (1985)
H.R. Thomas and J.J. O'Malley, Macromolecules 12, 323 (1979).
' N.M. Patel, D.W. Dwight, J.L. Hendrick, D.C. Webster and J.E. McGarth,
Macromolecules, 21, 2689 (1988).
1I J.E. McGarth, D.W. Dwight. J.S. Riffle, T.F. Davidson, D.C. Webber and R.
Vishwanathan, Polym. Prepr. Am Chem. Soc. Div. Polym. Chem 20(2), 528 (1979)
12 C.S. Henkee, E.L. Thomas and L.J. Fetters, J. Materials Sci. 23, 1685 (1988).
11I. Hasegawa, T. Hashimoto, Macromolecules 18, 589 (1985).
1-1S.H. Anastasiadis, T.P. Russell, S.K. Satija and C.K. Majkzrak, Physical Review Letters
(2, 1852 (1989).
1 S.!1. Anastasiadis, T.P. Russell, S.K. Satija and C.K. Majkzrak, J. Chem. Phys.
327
1016
(submitted).
G. Coulon, T.P. Russell, V.R. Deline and P.F. Green, Macromolecules, 22, 2581 (1989).
17G.L. Gaines, Macromolecules, 14, 208 (1981)
IS K. Binder, In Phase Transitions and Critical Phenomena, C. Domb, J.L. Lebowitz, Eds.;
Academic Press, New York, Vol. 8 (1983).
19 C. Rau and M. Robert, Physical Review Letters, 58, 2714 (1987).
20 P.G. deGennes, Scaling Concepts in Polymer Physics (Cornell University Press, Ithaca,
New York, 1979).
21 P.F. Green, T.P. Russell, R. Jerome and M. Granville, Macromolecules, 21, 3266 (1988).
22 P.F. Green, T.P. Russell, R. Jerome and M. Granville, Macromolecules, 22, 908 (1989).
23 P.F. Green, T.M. Christensen, T.P. Russell and R. Jerome, Macromolecules, 22, 2189
(1989).
24 P.F. Green, T.M. Christensen, T.P. Russell and R. Jerome, J. Chem. Phys. (in press).
25 D.T. Clark, J. Peeling and J.M. O'Malley, J. Polym. Sci., Polym. Chem, Ed. 14, 543
(1976). D.T. Clark and H.R. Thomas, J. Polym. Sce. 16, 791 (1978)
26 J. Als-Nielson, In Structure and Dynamics of Surfaces 11, edited by M. Schommers and P.
von Blanckenhagen (springer-Verlag, Berlin, 1987).
27 G.H. Fredrickson, Macromolecules 20, 2535 (1987).
28S. Wu, Polymer Interfaces and Adhesion (Marcel Dekker, New York, New York, 1982) p.
73.
29 The angle resolved measurements of these samples made by the slow solvent evaporation
process showed variations only for the shortest copolymer system, the PMMA fraction varied
from 0.25 to 0.18 in going from 900 to 200.
30 The angle resolved measurements showed variations only for the shortest copolymer
system, the PMMA fraction varied from 0.22 to 0.17 in going from 900 to 200.
31 Many of the films, particularly the high N systems, that were not annealed long enough
yielded values ofWt that were closer to 0.5. After longer anneals they reached their equilibrium
values.As N increased the time required to reach equilibrium increased appreciaby.
4
'1
I.
329
ON THE SCALE OF DIFFUSION LENGTHS OBSERVABLE BY NEUTRON
REFLECTION: APPLICATION TO POLYMERS.
A. Karim, A. Mansour and G.P. Felcher Argonne National Laboratory, Argonne, IL 60439,
T.P. Russell, IBM Research Division, Almaden Research Center, San Jose, CA 95120.
ABSTRACT
A systematic approach has been applied to neutron reflectivity data to study interdiffusion
across an interface. It is shown that with this technique it is possible to probe interface
broadening from -I0A to upward of 200 A, the upper limit being already within the range of
observation of other techniques such as Rutherford backscattering spectrometry (RBS), forward
recoil spectrometry (FRES) and secondary ions mass spectroscopy (SIMS). As example is
analyzed the interdiffusion of a bilayered polymer system: a deuterated polystyrene (d-PS) layer
on protonated polystyrene (h-PS).
Introduction
I
Neutron reflectivity has emerged as a useful technique to probe interface broadening
caused by diffusion across an interface or. the nanometer scale. The technique relies on the
contrast of the neutrons scattering length densities of the layers facing the interface. The quantity
obtained in such experiments is the reflectivity (R) of the sample: this is a function of the
component of the neutron momentum perpendicular to the surface, kz . R(kz) is an optical
transform of the scattering length density, as measured as a function of the depth z from the
surface. In this paper, we define the range of interface widths measurable by neutron
reflectivity, by taking the reflectivity curves at subsequent times of the diffusion process.
Limits of resolution
Consider a homogeneous layer of thickness z 2 on an infinitely thick lower layer, the indices
1, 2 and 3 indicate the vacuum above, homogeneous material and lower layer respectively. Then
the reflectivity R is given by 2,3
R= r? 2 +p2 3 +2r 1 2P 23 cos 2k2 z 2
l+r?2 p23+2rl2P23cos 2k 2z 2
-The
reflectance at the i,i+l boundary is:
f
ri,i+l
1
ki+ki~l
where ki = [ k2 - (b/v)i 1/2 and (b/v
k=2itSin0/X. If symmetric interdiffusion,
(2)
i5 -
scattering length density of the ith layer, while
occurs between layers 2 and 3,
Mat. Res. Soc. Symp. Proc. Vol. 171. c 1990 Materials Research Society
I-
(1)
SilAklllli en illmmmf"-
.
..
330
P23=r23exp -2k2kso
2
(3)
2
diffusion, the diffusion coefficient D is a function of the width 0 (t) and of time,
In conventional
6
namely
D=
2
(t)
4t
(4)
In this paper we would like to examine how to obtain a 2 systematically from the experimental
data in order to obtain the measurable range of diffusion coefficients.
The z2 averaged expression for the reflectivity [ eq(l)J, is:
<R>- r?2+P232r1222
1.r2 p23
(5)
which for large values of k can be approximated by the expression:
<R>- r12+Ph3
-
[t 2 /k14 ] [(b/v)
2
2
+ ((b/v) 3 - (b/v) 2 )2
exp - 4k12 o 2 ]
(6)
Thus RO4 vs k tends, for large values of k, to an asymptotic value. This value is different for an
infinitely sharp 2,3 interface compared to when some interdiffusion has occurred across such a
boundary; the decay to the new asymptote reaches a value l/e for kI=( 1/4aF2 ) 1/2.
It is easy to see that the determination of a becomes progressively ill defined for large a.
This is because the lie point occur for smaller values of k1. Even if the more correct eq. (5) is
used. rather than its asymptotic form, the relation between the experimental resolution, Ak, the
precision by which aYcan be determined:
dk/k = - do/
(7)
shows that extremely stringent requirements are progressively important on the angular
resolution and the wavelength definition of the incident beam. Practically speaking, this method
of analysis is convenient for values of a less than 150A. However it is possible to monitor
diffusion over significantly longer distances by using derivatives of the reflectivity. In the
normalized difference analysis, the infinitesimal changes in R with o2 is given by R-I(dRdo 2).
expression, it is
Rather than giving the rather cumbersome - but easily derivable - complete
worthwhile to point out that the z-averaged expression of R'l(dR/daF2) is simply zero. On this
flat base line, the envelope between the maxima and the minima of the reflectivity becomes:
331
aR
A(_ -
)= 4rl2p23k2k3 [)
A
2-23
1-r12P23
(8)
the exponential term which gives the broadening of the interface is now contained in the linear
expression of P23 at the numerator of eq.(8); which means that the range over which a can be
measured is extended by a factor of 1/ '42.
In practical terms the difference between the reflectivities at times t', t' can be taken. It is easy to
see that the relation holds:
R(t")-R(t')
R(t")+R(t')
-
2D (aR ) (t"- t')
2(9)
where D is the diffusion coefficient as measured in the t"-t' interval.
Interdiffusion of olvmers
We would like to apply directly this method of analysis to the study of the interdiffusion between
two polymers. Neutron reflection can readily observe diffusion even over short time, when
eq.(9) is not applicable. It is worthwhile then to compare the value of D obtained after relatively
long annealing time with neutron reflectivity with that obtained by other techniques which have
intrinsically less spatial resolution. Unfortunately, it has not been possible yet to apply directly
eq.(9) on two accounts: the instrumental resolution has to be obtained more accurately than
hitherto has been done, and the interval of annealing time t"-t' must be shorter than that presently
empioyed. However we would like to compare the normalized difference plot with those
calculated on the basis of a model, in which only the interface thickness has been changed.
Specific samples were prepared for the purpose, in the form of polymer bilayers on a flat
round of silicon of 2"diameter. The lower layer of h-PS( Mw=233,000; thickness >3000A )
was spin coated on a silicon substrate from its solution in toluene and annealed for 48 hours at
160C. A layer of d-PS (Mw=203,000;thickness-600A) was then separately spin coated on a
glass slide and subsequently floated off on to a pool of water. The floating d-PS film was then
picked up from beneath by the silicon substrate with the coated h-PS to form a bilayer. The
bilayer was then placed under vacuum for 12 hours at room temperature to remove any water
molecules that may have been trapped between the layers.
The neutron reflectivity experiments were perfc. med at the Intense Pulsed Neutron Source
at Argonne National Laboratory. The experiment basically consisted of reflecting neutrons
pulses between wavelengths 2-16 A from the sample surface at very small angles of incidence
(typically Idegree) and measuring the number of reflected neutrons as a function of wavelength.
The number of reflected neutrons normalized by the incident neutrons gives the reflectivity R, as
a function of the neutron momentum k.
In Figure 1,experimental results are presented for the sample "as cast", and then annealed
22 and 53 minutes at 135C. Figure 1 shows also fits for the thickness of the layers d-PS/h-PS
on Si, as well as the thickness of the interfaces. The fit to Figure I(A) indicates a top layer dPS of thickness 590A on a "thick" lower layer. The thickness of the lower layer is certainly
332
Fig. I
4
(A)--------1-----1.0
R(k)k as afunction of kforabilayred samlple
--------of d-PS Mw=233,000) on top of h-PS
(Mw=203,000) before beating [1 (A)] and after------heating for 22 minutes [1(B)] and 53 minutes----I-[1(C). The solid points are the experimental----I--data whereas the solid lines were calculatedI.........:----:--using a bilayered model with a symmetricA
0.1\
interface between the uppradowrly.
-------
09
4
10
------------------
0.
4 ----- .4.. -------- -
--
---- ------
------01-
--
-----
-------------- ...----- -- ---
t --
001--- 0024--0030
000- 0012---
enu
.... Mo.........
Neuro
333
larger than 3000A, which is the maximum distance which can be resolved for a beam of angular
resolution of the order of 0.020. The interface between the two polystyrene layers has initially a
thickness Y= 10A. The fit I(B) shows that the interface has broadened to 135A, as a result of
diffusion after the first annealing; after the second annealing the broadening is 200k. The
reflectivity in Figure I(C) is almost featureless. In order to observe clearly how well the
broadening is determined we replot in Figure (2) the experimental and calculated normalized
difference of 1(C) with its previous frame 1(B). The hollow circles indicate the experimental
difference points whereas the lines are normalized difference curves for three different values of
o (=175, 200 and 250 A) for fits to I(C) with X (= 4.6, 4.4, 5.2) respectively. Of the three a
values, o=200A has the minimum X (the fit using this value of ais shown in Figure I1(Q). The
offset to the normalized difference from zero at large k in Figure2 is caused by a small error in
normalization of reflectivity for l(B). We also see that if the offset is neglected, the integral of the
normalized fluctuations is nearly zero, in accordance with equation (8) - that the total area under
the normalized difference curve be zero.
From the value of a=200A, is obtained a diffusion coefficient (cfr eq.(4) ) D=3.43 x 1016 cm 2/s at T=135 0 C. This value is already in reasonable agreement with that obtained for the
16
7
same molecular weight at the same temperature from FRES measurements : D=4.35x 10-
0-_______
o
o
ZOo
0
D0
o
C 000
000
LLJ
0
I
0
z
_
(I
0
I.
0.006
0.009
0.012
0.015
0.018
0.021
NEUTRON MOMENTUM (ANG-1)
0.024
Fig. 2 Normalized difference curves as a function of k.
The open circles are the experimental data
obtained from the measurements I(B) and
1(C). The calculated reflectivity difference
curves with I(B) are shown by the solid line
(o=200A), the dashed line(o=250A) and the
dotted line (O=175A) for fits to I(C).
0.027
0.030
334
cm 2/s and it is to be hoped that further improvements in the sophistication of the two techniques
will yield results in yet better agreement and/or be able to focus on the causes of the discrepancy.
ACKNOWLEDGEMENT
The authors wish to thank Rick Goyette for his scientific assistance.
This work was performed under the auspices of the U.S. Department of Energy, Division
of Materials Sciences, Office of Basic Energy Sciences, under Contract W-31-109-ENG-38.
REFERENCES
[1]
S.A. Werner, and A.G. Klein, in Neutron Scattering, edited by D.L. Price, and K. Skold
(Academic Press, Ney York, 1989).
[21 M. Born and E. Wolf, Principles of Optics ( Pergamon, Oxford, 1975).
[3] O.S. Heavens, Optical Principles of Thin Solid Films (Dover Pubi., New York, 1965).
141 T.P. Russell, A. Karim, A. Mansour and G.P. FelcherMacromolecules, 1890 (21) 1988.
[5]
M.L. Fernandez, J.S. Higgins, J.Penfold, R.C. Ward, C. Shackleton, and D.J. Walsh,
Polymer, 1923 (29)1988.
[6]
J. Crank, The Mathematics of Diffusion, Oxford 1983.
[7]
P.F. Green, CJ. Palmstrom, J.W. Mayer and E.J. Kramer,Macromolecules, 501 (18)
1985.
335
NEUTRON REFLECTION STUDY OF SURFACE ENRICHMENT IN AN
ISOTOPIC POLYMER BLEND
R.A.L. JONES*, L.J. NORTON**, E.J. KRAMER**,R.J. COMPOSTOt, R.S.
STEINt,T.P. RUSSELLtt, G.P. FELCHERt, A. MANSOURt and A.
KARIMt
*Present address: Cavendish Laboratory, Cambridge University,
Madingley Road, Cambridge, CB3 0HZ, UK
**Materials Science and Engineering, Cornell University, Ithaca, NY
14850
tDepartment of Polymer Science and Engineering, University of
Massachusetts, Amherst, MA 01003
ttIBM Almaden Research Center, San Jose, CA 95120
tMaterials Science Division, Argonne National Laboratory, Argonne, IL
60439-4814
We have measured neutron reflectivities from the surface of films of
deuterated polystyrene (d-PS) and protonated polystyrene (PS) blends before
and after annealing, and used the results to determine the concentration
versus depth profile of the films. After annealing, the surface is enriched
in d-PS, with a surface excess proportional to the bulk concentration of dPS, in agreement with previous measurements using forward recoil
spectrometry[l]. The decay of the enhanced concentration into the bulk
occurs over a length approximately equal to the bulk correlation length
(-200 A), in close agreement with that predicted by current mean-field
theory[2]. However, the agreement between the experimental reflectivity
curves and the fitted curves is not completely adequate. Figure 1, a plot of
k4 R against k, demonstrates this point for a 15% d-PS sample. The dashed
line is the best fit assuming the mean field profile. The inset shows the
corresponding concentration profile and a trial profile, solid line, which fits
the data much better. The small deviation between the theoretical and trial
profiles may be due to the assumption of the theory that surface
interactions leading to enrichment is short ranged.
Mat. Res. Soc. Symp. Proc. Vol. 171. 1 1990 Materials Research Society
3rFigure 1
10
f
1.0
0.5-
0.0
0
500
1000
VI
0.000
0.00?
0.014
0.021
0.028
0.035
1.- R.A.L. Jones, E.J. Kramer, M.H. Rafailovich, J. Sokolov, S.A. Schwarz,
Phys. Rev. Let. 62, 280 (1989).
2. 1. Schmidt and K. Binder, J. Physique Ifi, 1631 (1985).
3. R.A.L. Jones, L.J. Norton, E.J. Kramer, R.J. Composto, R.S. Stein, T.P.
Russell, A. Mansour, A. Karim, G.P. Feicher, M.H. Rafailovich, J.
Sokolov, X. Zhao, S.A. Schwarz, submitted for puplication.
337
p
X-RAY REFLECTIVITY AND FLUORESCENCE MEASUREMENTS FROM
POLYSTYRENE- CO- BROMOSTYRENE/POLYSTYRENE INTERFACES
2
3
J. SOKOLOV', M. RAFAILOVICH', X. ZHAO', W.B. YUN , R.A.L. JONES ,
3
E.J. KRAMER , R.J. COMPOSTO4, R.S. STEIN', A.BOMMANNAVAR',M. ENGBRETSON
1) QUEENS COLLEGE, FLUSHING, NY
2) ARGONNE NATIONAL LABORATORY, ARGONNE, IL
3) CORNELL UNIVERSITY, ITHACA, NY
4) UNIVERSITY OF MASSACHUSETTS AT AMHERST, AMHERST, MA
5) BROOKLYN COLLEGE, BROOKLYN, NY
6) OAK RIDGE NATIONAL LABORATORY, OAK RIDGE, TENN
ABSTRACT
X-ray fluorescence using synchrotron radiation at glancing angles of
incidence was used to measure interface widths for the highly immiscible
polymer blend system polystyrene/polybromostyrene (PS/PBrS, with
bromination level x - 0.8 bromine atoms per monomer). The interfacial
widths are for bilayers annealed for 4 hours and 24 hours at Pl',llt are
found to be lO0±20A and llO±20A respectively.
Introduction
The characterization of interface in polymer blend systems gives
valuable information on mechanical and thermodynamic properties of blends
and can be used to discriminate between the various proposed theories of
11
polymer phase behavior. 1 When dealing with highly immiscible blends, such
as PS/PBrS (for high levels of bromination), the interfaces between phases
may be of the order of lOOA or less and measurements of interfacial
profiles require high spatial resolution. We report here the result. of a
study using the techniques of x-ray fluorescence under conditions of near
2
total external reflection (NTEF)A 3 In the NTEF technique, the angular
dependence of the x-ray-excited fluorescence of a labeled species is
measured for angles of incidence near the critical angle 6, for total
reflection - for polymers 0,
l-2mrad (see Table 1).
Below 6_. only
evanescent waves penetrate the medium and the fluorescence is sensitive to
the surface region. In the case of 15keV x-rays incident on polymers the
sampling depth is typically 50-00A. As the incident angle is increased to
# ) 0, deeper layers are probed, the total depth sensed being a function of
0 and the extinction length (Table I) for x-rays in the scattering medium
The total fluorescence(2 signal from the labeled species, in this case
bromine, is given by,
IBr(6)
-
K14'r(Z)
I.(Z,
8) eXp(-AerZ)dz
(1)
where 9 is the x-ray incidence angle, z is the depth into the sample., e,(c)
is the concentration of Br atoms of z, I,(z, 8) is the x-ray intensity at
depth z, Pr is the linear absorption coefficient for the characteristic Pr
x-rays in the polymer [as (l/psc) ) 100pm and the bilayers studied w(,r( 4
thickness ( 0.Spm, uB, was taken as zero] and K is a constant taking aco.nt
of the atomic fluorescence cros,-Fection and geometric factors sti.h a,.,
detector acceptance angle, efficiency, etc. Calculations of the ll,
i-
Mat. Res. Soc. Symp. Proc. Vol. 171.
190 Materals Research Society
338
field intensity Ix(z, 9) were done using the standard matrix propagation
3
method 1, in which the scattering medium is divided up into layers parallel
to the surface of constant (within each layer) refraction index n - Re(n)
+ ilm(n). The plane wave solutions of Maxwell's equations for each layer
are joined so as to satisfy the boundary conditions of continuous
tangential E and B fields. For the present calculations, 200-400 layers
were used to determine the scattering properties of the -0.5gm thick
polymer samples. Typical computation times were approximately 100s on an
IBM 3090 for an angular scan spectrum of 120 values of 9.
TABLE I
PS
Critical Angle (mrad)
1.45
PBrO 8S
Si(substrate)
I/e Extinction Length (pm)
13,110.
1.65
2.07
160.
418.
Experimental
Polystyrene (mw - 670K, polydispersity index - 1.15) was solution:
brominated following the procedures outlined by Kambotir and Bendler' to
obtain PBrS with x = 0.8, which is known to be highly immiscible with PS
5
of mw - 670K. 1 The degree of bromination was determined independently by
mass microanalysis. PS/PBrS bilayers were deposited on 2 inch diameter
polished silicon wafers in two steps: a) a 4000-5000A thick PBrS film was
spin-cast from a toluene solution onto the Si wafer and b) a 1200-20OOA PS
layer, floated on water, was placed on top of the PBrS layer. Samples were
annealed under vacuum for up to 24 hours at 141C.
X-ray reflectivity and
fluorescence measurements were made at beamline XI8B of the Brookhaven
National Synchrotron Light Source. A monochromated x-ray beam of energy
15keV. selected to be above the Br Ka absorption edge of 13.474KeV.
impinged on the polymer sample at angles 0 varying fromn 0.5 to 5.0 mend.
The bilayer samples were mounted vertically with the beam collimated to
-0.9mm in the vertical direction and in the horizontal direction set large
enough to engure complete coverage of the 0.9mm x 2 inch 'footprint' for
all angles of incidence. The reflectivity (mainly used to calibrate the
scattering angle and to measure sample thickness) was determined using
argon-filled gas detectors for both the incident and reflected beams while
the Br fluorescence was measured with a solid-state germanium detector.
Each angular spectrum of 120 data points took 20 minutes, or lOs/pt. No
detectable radiation damage was observed for these runs.
Analysis
Figures 1-3 show a comparison of calculated and experimental
fluorescence spectra for PS/PBrS bilayers annealed for 4 and 24 hours at
141!±C. The interface was modeled with a hyperbolic tangent profile of
339
variable with w:
OB-(Z)
(O./2)tanh((z-t)/w) + ((4/2)
-0
z
(
d
(2'
z)d
where t is the thickness of the PS layer and d is the combined thickness of
the PS and PBrS layers. The interface width w was allowed to vary from IDA
to 200A while t and d were varied ±300A about their nominal values. The
best values for the PS layer thickne:s t was found to be 1200±50A for both
samples; the PBrS layer thickness (d-t) was 4800±150A for the 4 hour
anneal sample and 4400±150A for the 24 hour anneal sample. Figure I shows
calculations for the 24 hour sample for widths w of 70. 110, 150 antd200A.
The main spectral feature which is sensitive to w is the shoulder at about
1.75mrad (this feature tends to be reduced for samples with thicker PS
layers-bilayers with t - 1900A exhibited only one peak and proved difficult
to analyze accurately). The optimum agreement between theoretical and
experimental spectra was found for
w - 100±20A [4 hour, 141C
and 110±20A [24 hour, 141C
anneal]
anneal]
Overall agreement is quite satisfactory.
Conclusion
The technique of x-ray near-total external fluorescence (NTEF) his
been demonstrated to be capable of studying interface formation on the
small length scales of highly immiscible polymer blend systems. The
PS/PBrS system is also suitable (by using
deuterated PS or PBrS) for
6
experiments using neutron reflectivity , the only other current]% availahie
technique used for high resolution polymer interface studies. Meisurement s
of the same samples will allow direct comparisons between the two
techniques, enhancing the reliability of both methods.
340
4000
....
4 0'0 0' 'l . . . ......
.....................
'3500
:t
.
. .
' . ....
........
...................
...................
.. .............
b
............
...
...
........
.".....................
;,
xi~...
C....
..
.........
....
...............
10,3000C
c 2500
2000
a..l
1.5
2.0
2.5
3.0
3.5
Angle (mrad)
Fig. i Calculated fluorescence spectra of PS/PBrS interface of width w
The PS layer is 1200A thick
equal to a) 70A, b) IIOA, c) 150A and d) 200A.
and the PBrS layer 4400A for all spectra. The zero of each spectrum has
been shifted in multiples of 100 intensity units.
341
4000
- 3500
6,~3000-
41
C2500
201.5
2.0
Angle (mnrad) 25
3.0
Fig. 2 Experimental (dotted) and calculated (solid line) spectra for the
24 hour, l41*C anneal sample.
PS and PBrS layer thicknesses as in F~ig.1.
Interfacial width w
110iA. The zero of intensity has been shifted 100
units
.
00
-U)3500
Z23000
C
200
1.5
2.0
2.5
Angle (mrad)
3.0
Fig. 3 Experimental (dotted) and calculated (solid line) spectra for t)Jc
hour, 141*C anneal samplo.. PS layer 1200A, ParS layer 1.800A and
interfacial width w - 100A. The zero of intensity has been shifted1 100
units.
342
References
l.P.G, deGennes, J. Chem. Phys. 72, 4756 (1980), P. Pincus, J. Chem.
Phys.75, 1896 (1981), E. Helfand, in Polymer Compatibility and
Incompatibility, ed. K. Soic, MMI Press, Harwood Academic Publishers,
Chur (1982) p. 143, I. Schmidt and K. Binder, J. Physique 46, 1631
(1985).
2.J.M. Bloch, M. Sansone, F. Rondelez, D.G. Peiffer, P. Pincus, M.W. Kim
and P.M. Eisenberger, Phys. Rev. Lett. 54, 1039 (1985), M.J. Bedzyk, G.M.
Bommarito and J.S. Schildkraut, Phys. Rev. Lett. 62, 1376 (1989).
3.M.
Born and E. Wolf, Principles of Optics (Pergamon, NY 1983).
4.R.P. Kambour and J.T. Bendler, Macromolecules 19, 2679 (1986).
5.C.R. Stobl, J.T. Bendler, R.P. Kambour and A.R. Schultz, Macromolecules.
19 2683 (1926) and Ref. 4.
6.M.L. Fernandez, J.S. Higgins, J. Penfold, R.C. Ward, C. Shackleton and
D.3. Walsh, Polymer 29_ 1923 (1988), S.H. Anastasiadis,
T.P. Russell, S.K. Satija and C.F. Majkrzak, Phys. Rev. Lett. 62, 1852
(1989).
343
INTERFACIAL SEGREGATION EFFECTS IN MIXTURES
OF HOMOPOLYMERS WITH COPOLYMERS
VIJAY S. WAKHARKAR, THOMAS P. RUSSELL and VAUGHN R. DELINE
IBM Research Division. Almaden Research Center, San Jose, CA 95120.
ABSTRACT
Secondary Ion Mass Spectroscopy (SIMS) has been used to study the surface and interfacial
segregation of diblock copolymers in mixtures of the copolymer in the homopolymer. Symmetric.
diblock copolymers of polystyrene (PS) and poly(methyl methacrylate) (PMMA), in either the PS or
the PMMA homopolymers were investigated.
In mixtures of the copolymer with the PS
homopolymer, systematic surface enrichment of the copolymer as well as segregation of the
copolymer to the polymer/Si interface or In the case of bilayered films to the polymer/polymer
interface occurs with annealing treatments. These segregation effects persist over a large range
of hornopolymer molecular weights with changes in the kinetics of the segregation process being
predominant.
INTRODUCTION
The structure and composition of the polymeric surface/interface and of the near-surface
region controls some of the important properties and therefore the applications of such polymers.
Hence an in-depth understanding of the surface and interfacial behavior of copolymers is essential
towards optimizing the use of these polymers in various applications, Presently, various analytical
techniques such as X-ray Photoelectron Spectroscopy (XPS) (1,2), Forward Recoil Spectrometry
(FRES) (3) and Electron Microscopy 14,5) are being used to probe the structure of polymeric
surface/interface and near surface i.e. interphase regions Secondary Ion Mass Spectrometry
(SIMS) has been applied previously in polymeric materials to study the interface induced orientation effects of diblock copolymers in thin films (6,7) and in diffusion studies (8) In this article we
report on the use of SIMS to Investigate the segregation effects at polymer Interfaces. We have
applied depth profiling in polymeric films to study surface segregation; segregation effects at
polymer/substrate interfaces or in the case of multilayered films, segregation effects at
polymer/polymer interfaces. SIMS has been used to obtain segregation profiles in real space as
well as to study the kinetics of the segregation process.
EXPERIMENTAL
The polymers used in this study were purchased from Polymer Laboratories
1, 3 and 10%
mixtures of the copolymer in the PS homopolymer were prepared in toluene solutions. In the
diblock copolymer used in this study, both the PS and the PMMA blocks were deuterated and is
designated as PSD/PMMAD The average molecular weight of the diblock copolymer was approxi,
mately 1.0 x 101 Films about 1600 A thick were cast directly on cleaned silicon substrates. These
films were then annealed in a vacuum oven at 170°C. which is above the glass transition temperature of either component. The annealing times ranged from a few minutes to tens of hours
A
thin layer of Au (
20-A) was vapor evaporated on all the films In order to serve as marker for the
air/polymer interface in the SIMS depth profiles. Subsequently, a PS layer of about 200-A was spin
coated on a glass substrate, floated on water, and then retrieved onto all the polymer films being
analyzed via SIMS
This PS film prevents hydrogen contamination and acts as a buffer layer
during the initial sputtering transition of the ion beam
The SIMS depth profiles were obtained using a Cameca IMS-4F secondary ion microscope A
3 KeV O4 primary ion source provides a means of sputtering the specimen surface
The secondary ions sputtered trom an 80 oim diameter area of the specimen are extracted, mass separated and analy/ed using a mass spectrometer
The specifit. conditions for SIMS analysis are
shown in table I It has been shown that these conditions provide high sensitivity and good depth
resolution while depth profiling organic films (9) During the SIMS experiments the intensity of the
secondary ion counts of
H'. I H*.
2C', and "'Aur'
are measured as a function otf puttelrinq
time tot each litm The sputtering rate in homopolymers PS and PMMA was determined by using
standard specimens of accuralely known thirknes
Fo the mixtures lieing used it was noted that
the sputterinq rate does not vary stqnifirantly from that of ie respe( live homopolymers Howevei
th stmiltering rate was ye r dept nrdenl upon ttie expertio.t;l (tI
onlditons and hen( e it is n es,.
sary to calibrate the sputtering rate (titincm each e rlfiririt
Hoi'ni
knowing the spiltterint r t.
,
therintensity of tlip mrea stirpd secondary ions vet i
'
'
l 11014i
titti e,
it rliv (onveiled intm a
dptSh pr.Ml
e
t
Mal. Res. Soc. Symp. Prec. Val. 1?1.
l1
Itlletis Research lotclty
344
Table I
Cameca IMS 41
AiparaI,
Primary Ions,
200 300 nA
Bean CuirnrI
Raster
0,
-500
Selected Area Aperlure
Beam Size
Effective Depth Resolution
Detected Secondary Ions -
" X 500 p
8U p diameter
200p
12 5 nm
I .'11H.
"C.. _Au
The relative proportion of the protonated and deuterated polymers as a function of depth is
t4
C* signal is used as a matrix monitor
indicated by monitoring the 'IH and 'H* signals. The
ensuring that the sputtering rate is constant during the depth profiling. The monitoring of the
t
""Au* signal serves as an interface marker i.e. indicating the location of the original air/polymer
interface prior to the placement of the PS buffer layer. It was noted that the measured sputtering
rate of gold was approximately the same as that of PS under the conditions used. Hence the presence of gold serves only as an inert marker and does not perturb the sputtering rate.
RESULTS AND DISCUSSION
A typical SIMS depth profile for an as cast film of a mixture of 10% fully deuterated diblock
copolymer PSD/PMMAD in the PS homopolymer is shown in figure 1 The position of the original
air/polymer interface is indicated by the vertical dashed line on the left hand side. corresponding
to the peak in the Au signal, while the polymer/silicon interface is indicated by the vertical dashed
line on the right hand side. The initial transient in the signal where the secondary ion counts of
carbon (C), hydrogen (H), deuterium (D) counts vary markedly occurs in the buffer layer of PS
Note that a steady state sputtering is achieved by the time the ion beam has reached the polymer
film under investigation. The C counts, which serve as a matrix monitor, remain relatively constant
throughout the entire thickness of the film. In this example, since we have a mixture of a fully
deuterated copolymer in a protonated homopolymer, an examination of the deuterium (D) signal
indicates where the copolymer is residing. Thus in the as cast film the copolymer has preferentially segregated to the polymer/silicon interface as indicated by the relatively high deuterium
Also note that there is a small surface enrichment of the copolymer as
signal at that interface
compared to the bulk The H signal, which was relatively constant throughout the entire film thickness. drops at the polymer/silicon interface due to the segregation of the fullydeuterated
copolymer to this interface
The depth profile for an identical film thatwas annealed at 170' C tn vacuum for twenty hours
Note that the C
ts; shown in figure 2 The dramatic change in the deuterium (D) ptlile is evident
high
t
The relatlv
signal remain- invariant throughout the sputtering of the film aq before
detferim signal at Ihe air/polymer interface now clearly shows thi' s r-ace onio hment of the
t
on
at
the
The
ep
to
the
annealing
ftealnent
copolymer that has o((ttred due
,
tn it
cat lee cleat ly seen t ttitte
Tht oeg reta hi
polytnerl ;tlron interface it ;lso enhtattced
which is a conposteo ot th two sttetttlrfttOl)d deuteriut (D) profile, Thus at utfhciently lotn
er
bulk tlt))
i hull low the copolvni
ttt Ilti,
annealing times the toncentfattn of the otlytnti
I
tit itlt-;l(
ijtivt.i-..
e )r
it
ItIt
the atr/polytte
now b iri presortl either ;It
345
107
106
Air
Surface
Si
Surface
,1o5
cc
H
0
o
102
Au
101
0
500
1000
1500
2000
2500
Depth (A)
Figure 1. A SIMS depth profile obtained from an as cast film of a mixture of 10% PSD/PMMAD in
PS (127 K).
107 .....
106
.............
.....
Air
Si
Surface
Surface
105
C104
H
-o
o
Co3
10
102
t01
At,,
0
500
1000
1!00
200
2200
Depth iA)
Figure 2 A SIM' depth profile obtained from an annealed fim of a tixltre of 10 , PSt)/PMMAD in
PS (127 K) afler annealing af 170^ C for two hours
346
14
12
2 1.0
0008
0
o
0.6
Air
Surface
0.4
I
02 -
0.0
0
500
1000
1500
2000
2500
Depth (A)
Figure 3. A composite of two superimposed Deuterium (D) profiles corresponding to a mixture of
10% PSD/PMMAD in PS (127 K). The solid line corresponds to an as cast film while the dotted line
corresponds to a specimen annealed for 20 hours at 170C.
The time dependence of the interfacial copolymer segregation was followed over a wide
range of annealing times. Figure 4 shows the superimposed depth profiles at different annealing
times for a 3% PSDIPMMAD copolymer dissolved in a PS homopolymer of 1.27 x 105 average
molecular weight. The systematic surface enrichment of the copolymer due to the annealing treatment is clearly seen. At the polymer/silicon interface similar effects were observed. In this particular example it was noted that even a small annealing time, such as five minutes, was sufficient to
observe detectable migration of the copolymer to the interfaces.
2.5
.
Air
Surface
.4
20x
lit
C
11
c 15
o
''
c
21(
) O0
0
0
,f11 , . .fBI .
81
,400
80O0
600
[.~lh
(A)
10()(1
Figure 4 A compo,,ite of four superimposed Deuterium (D) profles" in the ne;ir surface region This
rorrrsponds to filmsiof ;r inoture of 3/,, PSD/PMMAD in PS (127 K) atd annealed ;it 171'C ars mdirated (A) As C;it (R) 5 min (C) 3n n0 n (D) 12n nmn (from ref 1111
347
8.0
-i 6.0
Air
Air
0
Surface
c:4.0
(0
ow2.0
hOi
0.0
0
500
II
1000
1500
2000
Depth (A)
Figure S. A composite of two superimposed Deuterium (D) profiles corresponding to a mixture of
3% PSDIPMMAD in PS (1150 K). The solid line corresponds to an as cast film while the dotted line
corresponds to a specimen annealed for 20 hours at 170*C.
The molecular weight of the homopolymer PS in the mixture was varied from 1.75 x 10' to 115
x 106. Surface enrichment of the copolymer as well as its segregation to the polymer/silicon interface persists over the entire range of homopolymer molecular weight variation. Figure 5 shows a
composite of two superimposed deuterium profiles corresponding to a mixture of 3% PSO/PMMAD
in PS of average molecular weight of 1.15 x 1Os. In this case it was necessary to anneal the specimen for several hours before any detectable migration of the copolymer to the interfaces was
observed,
SIMS has also been used to examine polymer/polymer interfaces by depth profiling multilayered polymer films. For the bilayer films analyzed, the lower film corresponds to a mixture of 1%
PSO/PMMAD in PMMA while the top layer corresponds to a mixture of 1% PSOPMMAD in PS
The top layer was coatedl on a glass substrate, floated on water and then retrieved onto the first
polymer layer, This bilayer film was then annealed and prepared for SIMS as described before
Figure 6 shows a typical depth profile for a bilayered film that was annealed for 48 hours. Thus in
this case, in addition to Ihe surface enrichment, segregation of the copolymer to the
polymer/polymer inlerface also occurs.
CONCLUSIONS
SIMS has been used to investigate the intrfacial segregation ot diblock copolymers in mixlures of the copolymers with homopolymers
It was found that annealing treatments result in the
systematic surface enrichmenl of the copolymer as well as its segregalion to the polymer/silicon
interface or in the case of bilayers to the polymer/polymer interlace The segregation effects were
observed regardless of the molecular weight of it,- homopo(ymer and the bulk volume fraction of
the copolymer in the mixture
The predominant effect of these rharges was fo influence the
kinetics of the segregation process
348
107
16 Air
Surface
PS
PMMA
PSrD/PMMAO PSO/PMMA()
r
,105
C
10
c
H
4
0
.13
0
102
101
0
Au
500 1000 1500 2000 2500 3000 3500
Time (sec)
Figure 6. A SIMS depth profile obtained from a bilayer film that was annealed at 170*C for 48
hours.
REFERENCES.
1.
PdJ.Mills. P.F. Green, CA. Palmstorm, J.W. Mayer and EdJ.Kramer. App. Phys. Lett.. 45.
9, (1984).
2.
H.R. Thomas and JJ O'Malley, Macromolecules 12, 323. (1979).
3.
P.F. Green, CdJ. Palmslrom, J.W. Mayer and E.J. Kramer. Macromolecules. 18.
501,(1985).
4.
H. Hasegawa and T. Hashimoto, Macromolecules, 18, 589, (1983).
5.
C-S. Henkee, E.L.Thomas and L.d. Fetters, J. Mater. Sci., 23, 1885, (1983).
6.
G. Coulon. I.P. Russell, V. R. Deline and P. F. Green, Macromolecules 22. 2581, (1989).
7.
T.P. Russell, V. R. Define, V.S. Wakharkar and G. Coulon, Materials Research Society
Bulletin. VoI.XIV (No.10), 33,1989.
8.
S.d. Whitlow and R.P. Wool, Macromolecules, 22. 2648, (1989)
9.
GdJ. Scilla. submitted to Anal. Chem..
10.
V.S. Wakharkar, V. R. Define and T.P. Russell. Proceedings of the VII International SIMS
Conference, September 3-8, Monterey, CA.. (John Wiley 8Sons Limited Publishers.
1989), In press
349
ELLIPSONETER
A NEW VARIABLE ANGLE FT-IR
J.L STEHLE * , O.T THOMAS *
J.H.LECAT *, L.C HAMMOND **
,
J.P. PIEL
*
,
P.EVRARD *
•
SOPRA, 26/68 Rue Pierre Joigneaux , 92270 BOIS-COLOMBES,
FRANCE
•* ARIES/QEI, 5A1 Damonmill Square, Concord MA 01742, USA
ABSTRACT
Use of a Fourier-transform interferometer integrated with
a variable incidence angle ellipsometer extends the spectral
range and the capabilities of spectroscopic ellipsometry into
the infrared. With a spectral range of 600 to 6600 cm-l, thick
layers, such as epitaxial doped layers and polymers can be
analysed.
A
full
description
of
this
novel
instrument
will
be
given. Incidence angle can be varied automatically to enhance
signal/noise
and
the
ellipsometric
data
can
be
obtained
together with vibrational absorption bands information to give
a characteristic " fingerprint " of the layers.
Examples of spectra of HCN polymer on nickel, DMHS on
aluminium and PMMA on silicon will be presented for various
incidence angles and layer thickness.
I :
INTRODUCTION
Since the year 1950, a great number of materials have
been characterized by mean of an infrared ellipsometer (1)
As
the
rotation
of
the
polarizer
was
manual,
these
measurements were tedious and time-consuming . The development
by Stobie, Rao and Dignam of the first photometric infrared
ellipsometer has been a great improvement (2). They show the
great sensitivity of this technique by the measurement of
formic acid chemisorbed on silver.
With the development of Fourier-transform spectrometers,
a new generation of infrared ellipsometers is born, using a
combination of a Michelson interferometer and a photometric
ellipsometer. The first one was achieved by Roseler (3) in
1981. Today, it is an instrument in development at SOPRA. The
first results and the possible applications will be shown.
II
: THEORY
Ellipsometry is a non-destructive technique, determining
the change of state of polarization of the light reflected on
the surface of a sample. More precisely, the complex
reflectance ratio r /r. is measured, where r and r are the
complex
reflectands
of
p- and
s- pglarizeg
light,
respectively.
rp/rs = tan
ei
A
The ellipsometric parameters tanV'
and cosA
are a
function of the refractive index and the thickness of each
layer of the sample. In comparison with reflectometry, two
parameters instead of one are obtained per wavelength, and
this allows the determination of both real and imaginary parts
of
the
materials,
without
using
a
Kramers-Kronig
transformation. It means that there is no need to extrapolate
the spectrum values outside the measuring range.
Mat. Aes. Soc.Symp. Proc. Vol. 171. 1990 Materials
Research Society
350
physical
to
get
a
are
used
Different
models
spectrum. The most
interpretation of the tan W and cosA
general one is a multilayer model, where thickness and
composition of each layer are introduced as parameters in a
regression program.
Beside this general approach, it is possible to use some
data reductions concerning the particular case of one layer
over a substrate. For a transparent layer, the real part of
the index and the thickness can be calculated for each
the
substrate composition. If the
wavelength, knowing
thickness of the layer is known, a direct computation of both
real and imaginary part of the index can be obtained.
Workinq in the infrared allows some simplification of the
ellipsometric equations, because the thickness of the layer
can often be considered an small, compared with the
wavelength. Therefore, it is useful to introduce the following
quantity (4)
D = ln( tan'4
/tan 4' ) + i(
A
--
A are the values of a bare substrate in the
where tanW and
same conditions.
When the substrate is transparent in the infrared, the
variations of the dielectric function can be directly seen on
the spectrum of the real and imaginary part of the relative
complex optical density D .
III:
CHARACTERISTICS AND PERFORMANCES OF THE ELLIPSOEM
The advantage of this technique, compared with reflectometry is that ellipsometry does not measure intensities but
a polarization state of light. Thus, the precision of the
measurement is not degraded by intensity fluctuations of the
source, or electronic drifts. No reference spectrum is needed.
A Fourier transform has the following advantages on a
monochromator :
- it is a multiplex configuration, because all the
wavelengths are simultaneously measured. Therefore, a spectrum
is obtained faster and with a better signal to noise ratio
than with a monochromator ( Fellgett advantage (5) ). A
measuring time in the minute range is possible.
there is an energetic disadvantage of the
monochromators
advantage (5) ).compared to an interferometer ( Jacquinot
The Michelson interferometer of a Fourier transform
spectrometer is followed by a photometric ellipsometer of
rotating analyzer type ( fig 1 ). Ge Brewster angle polarizers
have been developed against grid polarizers to obtain a better
degree of polarization over the whole spectral range. The
spectral range goes from 6600 to 600 cm-1, and the resolution
can be selected from to 2 to 32 cm-1.
The sample reflects a collimated beam whose divergence
can also be selected from 0.5 to 6 degrees by the software. As
the sensitivity of the ellipsometric measurements is greatly
dependent on the incidence angle, it can be automatically
modified by the user, from 40 to 90 ,+/- 0.01 degrees.
351
At the end of the optical train, the liquid nitrogen
cooled detector HgCdTe delivers an electrical signal. The
processing of this signal contains two Fourier transforms, the
first one to get the signal per wavelength, the second one to
compute the ellipeometric paramieters.
Midisu Intrferorroeter
41
APERWN
POAIZER
FIXED
61 Brewster
Ld cooted
MCI detectr
- ~7
AUk ~
SELECTABLE
G1Bewster
10S- 1- 2 - 41
FocusIng
fg1:Optical
MIrro
Scheme of the Elllpsometer
352
0.2
0.37S
400d
$G0oo
.000
--
1000
2000
1004 700 (:k-l
f
o, zo
0.,o0-
-
.4
i
Aluminium.
0000
tanW
, o0
and cosA of HMDS( Hexamethyldisilane) on
,000
o1000
o00
00
70001.
0.00
tanq4
silicon
and cosA
spectrum of PMMA resist on
353
.000.
$"a
4000.
3600
3000
l00
100
1
0.0
"
-
-0.0
b
L
Fig
!7)
tanP
and cosA
three incidence angles :
spectrum of HCN polymer on Ni for
a : 65
b: 70
"; C: 75
V : APPLICATION TO SURFACES STUDIES
Dielectric function of bulk materials, such as metals,
can have their n,k values directly
silicides,
glasses,
calculated for each energy from the ellipsometric parameters
. For instance, optical properties of Au have
tanW and COSA
been studied by Jungk and Roseler using FTIR Ellipsometry (6).
Ellipsbmetric techniques have already proved their high
sensivity in the visible spectral range for the study of film
Recently, experimental studies by
processes (7).
growth
R.BENFERHAT and B.DREVILLON using an infrared spectroscopic
phase-modulated ellipsometer have showed that it was possible
to get evidence of anisotropic orientation of one single
The vibrational
monolayer of Langmuir-Blodgett films (8).
properties of hydrogenated amorphous silicon ultrathin films
by
this technique
obtained
(less than 100 A) have also been
(9).
the
ellipsometer,
IR
a
Fourier-Transform
Using
densification and the porosity in low temperature deposited
oxide have been determined (10). The mean Si-O-Si bond angle
was calculated from the peak IR values and related to the
oxide density.
The Fourier Transform Ellipsometry can also offer a way
of characterization of thick layered materials(Fig. 2,3 and 4).
Measurements of slightly doped Si Layers epitaxied on heavily
doped Si substrates have been performed (11). With a 2 cm-1
spectral resolution, a 100 micron-thick layer can be studied.
Thick porous silicon layers are also a possible application of
this 1lipsometer.
354
REFERENCES
(1)
J.R BEATTIE
Phil. Mag. 46 235 (1955)
(2)
R.W STOBIE, B.RAO, and M.J. DIGNAM
Applied Optics Vol 14, No 4, (1975) 999
(3)
A.ROSELER
Infrared Physics 21 349 (1981)
(4)
M.J. DIGNAM, B.RAO, M. MOSKOVITS, and R.W. STOBIE
Can. J. Chem. 49, 1115 (1971)
(5)
(6)
A.E MARTIN
Vol 9 in Vibrational Spectra and Structure
Ed J.R Durig, Elsevier.
G.JUNGK and A. ROSELER
Phys.Stat.Sol (b) 137, 117,
(1986)
(7)
S.ANDRIEU, F. ARNAUD D'AVITAYA
Surface Science 219 (1989) 277 - 293
(8)
R.BENFEHRAT,B.DREVILLON and P.ROBIN
MRS BOSTON Fall Meeting 1-6 Dec 1986.
(9)
B.DREVILLON and R.BENFERHAT
J.Appl.Phys. 63 (1988) 5088
(10) F.FERRIEU and R.A.B DEVINE
To be published in J.Non-Crystalline Solids
(11)
F.FERRIEU
Rev. Sci. Instrum. 60 (1989) 3212
A
355
POLYMER MOLECULES AT INTERFACES:
STUDIES BY SMALL-ANGLE NEUTRON SCATTERING
W.C. Forsman and B.E. Latshaw*
*Department of Chemical Engineering, University of Pennsylvania, Philadelphia,
PA 19104.
D.T. Wu**
** Marshall Research & Development Laboratory, E.l. DuPont de Nemours & Company, 3500 Grays Ferry Avenue, Philadelphia, PA 19146
Introduction -- Our Kxoerimental System
The behavior of polymer molecules at interfaces is an essential asp ct of
a wide variety of physical-chemical phenomena. For example, polymer molecules
in the fiber-polymer interface play an important role in determining the
mechanical properties of composites, and polymer molecules at the liquid-solid
interface are critical in the stabilization of colloidal dispersions. In some
cases polymer molecules are physically adsorbed at the interface and in other
cases they are chemically bonded to the surface at one or more positions along
the chain. To simplify the terminology, however, both cases will be referred
to as polymer adsorption. The chains in the second case are said to be
grafted to the surface. It is an example of this second case that we are
concerned with in this paper, and in particular with polymer grafted to a
finely divided substrate dispersed in a liquid.
Among the parameters of interest in this example of polymer adsorption
are the number of polymer molecules adsorbed per unit surface area, and the
polymer density profile -_ that Is, the density of polymer repeat units as a
function of the distance from the adsorbing surface.
In principle, the length scales characterizing most density profiles
makes them amenable
study by small-angle
neutronof
scattering.
Indeed, neuthree
laboratories
have nowto demonstrated
the feasibility
using small-angle
tron scattering for determining density profiles for polymers adsorbed at
solid-solution interfaces of high surface-area substrates. Cosgrove and
1 2 3 4. . 7
associates . . . 3
have been studying the density profile of poly(ethylene
oxide) adsorbed on the9 surface of polystyrene latex particles. Auvray, Cotton
and their associatesO. have been examining acrylic polymers adsorbed on the
inner surface of porous silica.
We recently completed
a set of preliminary experiments at Brookhaven
5
National Laboratory . Our system was poly(n-butyl methacrylate) chemically
bonded at one end of each chain to the surface of nearly monodisperse, 2006
A
2.
silica spheres. The grafting density was one polymer molecule per 900 A
The polymer was prepared by group transfer polymerization, with a narrow
molecular weight distribution and an average molecular weight of 53,000.
Scattering measurements were made on dispersions of bare particles and
particles with grafted polymer. We did contrast variation experiments using
mixtures of hydrogenous and deuterated isopropanol (IPA) for the dispersant
phase. Since IPA is a solvent for the polymer (albeit, a poor solvent), we
assume that if the poly(butyl methacrylate) were not grafted to the surface it
would have leached off of the substrate. We saw no evidence of leaching
during our neutron scattering experiments.
Scattered intensities (after correcting for background) were directly
proportional to the particle concentration for 1, 5 and 10 volume % dispersions. We took this to mean that there is no contribution to the scattered
intensity from interparticle interference over our range of experimental
scattering angles. This is to be expected for particles as large as 2000 A.
Autal:
Uncoated Particles
Scattering from homogeneous spheres of radius R Is determined by their
scattering function, FS(QJ, which is written:
Mat. Ro. 8oc. SyMp. Proc. Vol. 171. '1990 MaterAll Research Society
356
FS"-F,[
(- 5 .
[n(QR)- QR)cos(QR)]
(1)
where
V s - The volume of the sphere
and
Q- 4sn(e/2)
(2)
In equation 2, X is the wavelength of the radiation and e is the angle between
the incident and scattered beams.
For most experiments, the scattered intensity, 1(Q), as a function of Q
and the intensity of the incident radiation, I0, is written as
I(Q)- 1oa'F'(Q)
(3)
The quantity as is the contrast between the sphere and the solvent (or dispersent). The contrast of any scattering object is given by
a,-p,-p,.,
where
(4)
p,-Scattering length density of component i. and
p_-Scattering length density of solvent.
For spheres with diameters as large as 2000 A the scattering function is
well within the large.Q asymptotic scattering regime over our entire range of
experimental Q. Furthermore, the sinusoidal oscillations are beyond the
experimental resolution of Q. Consequently, the instrument gives the scattered intensity averaged over the individual detector elements. For 2000
spheres, equation 3 must be replaced by
I(Q)- I
PQ<F(Q)>
(5)
which can be written as
1(Q)
Constant
Q4
(6)
This same form 3 of 1(Q) would apply to polydisperse spheres as long as
they were all of l0 A length scale or larger.
Scatterine by the "flare" Particles
If the silica particles had been homogeneous, our experimental I(Q) would
have been described by Equation 6. That is not, however, what we measured.
Our experiments demonstrated significant deviations from Equation 6 at both
ends of our experimental range of Q.
Modeling of scattering from spheres leads to the conclusion that deviations from equation 6 are the result of non-uniform particle composition.
There are two quite different aspects to this non-uniformity -- aspects we
could call "course grain" and "fine grain.
They can be described as follows:
(1) The Si02 concentration is not a constant throughout the particle, but is
a smooth function of the radius -- this is the "course grain" non-uniformity.
At this level of detail, we can think of a particle as a core surrounded by a
shell, with a transition from core to shell that takes place smoothly over
radii from about 100 A to about 200 A. The Si0 2 concentration is relatively
constant for all radii greater than 200 A. This "course grain" non-uniformity
contributes deviations from equation 6 at the small-Q end of our experimental
range.
357
(2)
$particles.
We also found that there are random density fluctuations throughout the
The size scale of the fluctuations is about 9 A -- i.e., the "fine
grain" non-uniformity. This is exactly the same type of random, short-range
density fluctuation one finds in other glassy materials. This "fine grain"
non-uniformity contributes deviations from Equation 6 at the large-Q end of
1
our experimental range. 1
These structural features are consistent with what one would expect, considering how the silica particles are made in the laboratory. 2 The most
important result from chis analysis is, however, that neither the "fine grain"
nor the "course grain" non-uniformities interfere with interpretation of the
scattering data in terms of the density profile of the grafted polymer.
In addition to scattering from the structural features associated with
the body of the silica spheres, we found a significant scattering from a thin
shell surrounding the particles, which we assume to be surface hydroxyl groups
and an associated solvation layer. The scattering function for this shell is
written
3 j[(sjnQR)(sinQD) (cosQR)(l -cosQD)]
F 3 H - V3[
where D is the thickness of the shell.
We conclude that the following equation adequately accounts for the scattering of "bare" particles:
-a
<
1(Q)
F > .
z
'(Q), , '>+q<F 2M>
ooF
5 5 ,
2asos,,<FF,,>
(8)
where <Ft> now takes into account the non-homogeneities of the sphere.
The Polymer Density Profile:
Theoratical Analysis
The scattering function for the grafted polymer layer can be written
F,
-
V,[3
{
] sin(QR)I.
cos(RQ)1 j
(9)
where
c"
QJ(Z)os(QZ)dZ
and
Is - Qf(Z)sn(QZ)dZ
The function #(Z) is the polymer density profile expressed as volume fraction
of polymer, and is a function of the distance from the surface, Z. The two
integrals will be called the cosine and sine transforms of the polymer density
profile. Expressing the transforms as Taylor series gives
(10)
-d
I,-Q<7%.-Q
--
2'
and
<31
I
358
The zeroth moment of the density profile, <Z'>, is just the integral over (Z)
from 0 to -, and gives the polymer coverage of the substrate. The product of
0
<Z > and the specific volume of the polymer gives the coverage in mass of
°
polymer per unit area of the substrate. The ratio <Z>/<Z > is the average
thickness of the polymer layer. It is clear, therefore, that much can be
learned about the adsorbed layer just from the limiting behavior of the transforms as Q approaches zero.
Scatterine from Soheres with Grafted Polymer
When the spheres have a grafted polymer layer, the scattered intensity is
written
(11)
F>
2ar<F3F> +2aas<FsFm>+ 2apasm<FpFsH >
<'
By taking the difference between equations 11 and 8, we obtain the terms
associated with the polymer layer.
2asa,<FeFs> + 2apUsH<FpFso>
-0
. .2<F2>
(12)
By evaluating the terms in equation 12 we get
(QR)4A/
9V21
a(12+ 12
)-2oaul_
P
+ 2auo
[(sinDQ)lc+(I -cosQD)13 ]
(13)
Given D, for every experimental value of Q, there are two unknowns in equation
13, Ic and Is. In principle, jy doing scattering experiments in only two
mixtures of hydrogeneous and deuterated dispersants, and thus having data at
two different contrasts, one has enough information to solve for the two
transforms at every value of Q. This procedure is, even in principle, valid
only if the nature of the polymer adsorption is independent of isotopic substitution in the dispersant. In practice, if Ic and Is are to be determined
with any reasonable level of confidence, one should have scattering data with
several mixtures of hydrogeneous and deuterated dispersants.
The Ouestion of a Deoletion Layer.
Consider polymer chains grafted to a surface by one end. Suppose that
the net polymer-surface interaction is not favorable for adsorption -- indeed,
that if the chains were not grafted to the surface they would not adsorb.
Under these cirrumstances there is theoretical reason to believe that the
polymer density profile would be one with a maximum some distance out from the
substrate surface. We could define a region from Z - 0 to some value of Z
less than the value for the maximum in the density profile as a depletion
layer -_ I.e. a layer within the domain of adsorbed polymer that is both close
to the substrate surface and significantly depleted of polymer.
For example, consider the two hypothetical polymer density profiles given
below and shown in figure 1.
f-
(constant)e-*
z,-(constant)Z'eoz
(14)
359
Figure 1
250
(A
.. ,
........
0
200
0
CL
150
z
100
LaJ
0
~
0 v/
0
20
40
60
80
100
DISTANCE FROM ADSORBING SURFACE
Polymer density profiles with and without a depletion
layer. The solid line represents *o(Z) and the dashed
line 42(Z). Units are arbitrary.
Profile *O(Z) shows no depletion layer. The density of polymer repeat
units drops off steadily with increasing Z. Profile 02(Z), on the other hand,
shows a depletion region. Definition of the depletion layer is rather arbitrary, but clearly there is little polymer between Z - 0 and Z - 10 (in
arbitrary units).
Figures 2 and 3 show the sine and cosine transforms for the hypothetical
density profiles defined above.
Notice that, in the case of *2 (Z), for increasing Q, both transforms go
through a maximum, cross the Q axis, demonstrate a minimum and then increase
towards zero. Behavior of the transforms for *0 (Z), which shows no depletion
layer, is entirely different. They both demonstrate a maximum and then
decrease monotonically toward zero with increasing Q.
At least for the family of hypothetical density profiles of the sort
given in equation 14, a depletion layer is characterized by profile transforms
that cross the Q axis and demonstrate both a maximum and a minimum. Further
analysis is under way. But in the interim, we suggest that these criteria may
serve to "fingerprint" the presence of a depletion layer.
360
Figure 2
600
7
400 -i
200
o
o
0
-400
0
-600
0
0.05
0.1
0.15
0.2
0 [reciprocal AngstromsJ
Sine and cosine transforms of 0,(Z), the hypothetical
segment density profile that demonstrates a depletion
layer. Solid line; cosine transform. Dotted line;
sine transform.
Figure 3
1.2
0.8
0.6
0.4
C
0.2
0
0.05
0.1
0.15
0.2
0 [reciprocal Angstromsl
Sine and cosine transforms of *O(Z), the hypothetical
segment density profile that does not demonstrate a
depletion layer. Solid line; cosine transform.
Dotted line; sine transform.
361
Experimental Resulta
The Grafted Polymer Laver
We used equation 13 to determine Ic and Is using scattering data taken
from dispersions in 50, 60, 70 and 100% deuterated IPA. Preliminary calculations indicated that, as long as the scattering length density of the spheres
is not too close to that of the dispersant (I.e., not near the matching
point), the first term on the right-hand side of equation 13 is negligible.
This implies that scattering effects due to interference between the polymer
layer and the hydroxyl shell, and the polymer layer and the body of the sphere
are greater than the effect of scattering from the polymer layer itself. This
is not unreasonable, since there is less mass of grafted polymer then mass of
either sphere or hydroxyl layer.
Figure 4
•
0.4
C
0.2
0
-6,- 1001 and 701 DIPA
701 and 601 DIPA
c-e-
-0.2
601 and 501 DIPA
o-
-0.4
-0.6
0.01
0.02
0.03
0.04
0.05
0.06
0 Ireciprocal Angsfrom$n
Experimental sine transform over a range of small Q.
Figure 5
S10
Z
-6- 100% and 70% DIPA
c 0
-5
E
a
-----
701 and 60% OIPA
601 and 50% DIPA
-10
o
15
-200
0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16
0 [reciprocol Angstromsl
Experimental sine transform over an intermediate
range of Q.
362
In the analysis shown here, we used combinations of 50 and 60%, 60 and
70%, and 70 and 100% deuterated IPA to determine three different Is(Q) functions. The average of these three were then used, along with experimental
data to determine Ic(Q) for each of the four dispersant mixtures.
The results are shown in figures 4, 5 and 6.
Figure 6
e
5
0
0
-e0
o
E
-10
10 -t5
-20
0
100% OIPA
--- 70% DIPA
• ' '
0
' '
'
-8-
60% DIPA
--
50% DIPA
'
0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16
0 [reciprocal Angstroms]
Experimental cosine transform over an intermediate
range of Q.
Agreement of the various transforms computed from the scattering results in
the various dispersant mixtures was surprisingly good at low Q. However, the
agreement between the Is(Q) for the three pairs of contrasts became quite
unsatisfactory at higher Q. The poor agreement between Is(Q) determined from
experiments using 50 and 60% DIPA and the other two sets (see figure 4) may be
because these measurements were made too near the matching point of the
spheres for the first term on the right hand side of equation 13 to be negligible.
Both transforms demonstrate a maximum and then become negative with
increasing Q. The cosine transform then demonstrates a clear minimum. The
behavior of Is(Q) with increasing Q is not so clear. Nonetheless, there is
evidence of a minimum, especially if the 50-60% DIPA data (that which is
suspect) is ignored.
We conclude that, to the extent we can rely on the criteria described
above, the results indicate that the polymer density profile in our experimental system demonstrates a depletion layer. Additional scattering experiments
and analysis are required before we can quantify the polymer density profile
in detail.
383
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
K.C. Barnett, T. Cosgrove, B. Vincent, A.N. Burgess, E.L. Crow] y, T.
King, J.D. Turner and Th.F. Tadros, Polymer, 22, 283 (1981).
Cosgrove, T.L. Crowley, B. Vincent, K.G. Barnett and Th.F. Tadros, Faraday
Symp. Chem. Soc., 16, 101 (1982).
T. Cosgrove, T.L. Crowley and B. Vincent, in kdsorption from Solution,
edited by R.H. Ottewill, C.H. Rochester and A.L. Smith, (Academic Press,
1983).
T. Cos rove, T.C. Heath, K. Ryan and T.L. Crowley, Macromolecules, 20.
2879 (f987).
T. Cosgrove, T.G Heath, K. Ryan and B. van Lent, Polymer, 28, 64 (1987).
T. Cosgrove, T.C. Heath, and K. Ryan, Extended abstracts. American Institute of Chemical Engineers Annual Meeting, paper 4D (1987).
T. Cosgrove, T.L Crowley, L.M Mallagh, K. Ryan and J.R.P. Webster, Polymer
Preprints, 3, 370 (1989).
L. Auvray, and J.P. Cotton, Macromolecules, 20, 202 (1987).
L. Auvray, and J.P. Cotton, Extended abstracts, American Institute of
Chemical Engineers Annual Meeting, paper 4E (1987).
10. W.C. Porsman and B. Latshaw, Paper presented at the annual meeting of the
American Institute of Chemical Engineers, San Francisco, CA, November,
1989 (unpublished).
11. P. Debye and A.M. Bueche, J. Appl. Phys., 20, 518 (1949).
12. W. Stober, A. Fink and E. Bohn, J. Colloid Interface Sci., 26, 62 (1968).
365
MECHANICALLY INDUCED SILICA-SILOXANE MIXTURES.
STRUCTURE OF THE ADSORBED LAYER AND PROPERTIES OF THE NETWORK STRUCTURE
J.P. COHEN-ADDAD
Laboratoire de Spectrom~trie Physique associ6 au CNRS, Universit6 Joseph
Fourier, Grenoble I, B.P. 87 - 38402 St Martin d'Hres Cedex, France
ABSTRACT
Properties specific to silica-siloxane mixtures are analysed and discussed. The effect of polymer adsorption is interpreted from the Gaussian statistics of chains : the amount of adsorbed polymer Qr is proportional to the
square root of the chain molecular weight. The kinetics of adsorption is described as a process of surface saturation. It is discussed as a function of
the silica concentration. The effect of swelling of the mixtures is interpreted within the statistical framework proposed by Flory for ordinary gels.
I. INTRODUCTION
This description deals with typical properties observed from siloxane
polymers filled with silica particles. The addition of the particulate filler
to poly(dimethylsiloxane) (PDMS) chains was achieved from a mechanical mixing
. The contact of the polymer melt with the surface of silica gives
rise to a process of adsorption through the formation of hydrogen-bonds.
Each hydrogen-bond is supposed to link one silanol group of the surface to
one oxygen atom of the 3keleton of a PDMS chain. Several hydrogen-bonds can
be formed ,etween the silica surface and one end-methylated chain. This undergoes ar uniform adsorption. An end-adsorption occurs when siloxane chains
are hydrovyl terminated, in this case a double hydrogen bond is established
between or.echain-end and one silanol group of the surface.
Inve. tigations concerning properties of silica-siloxane systems are de:cording to three main topics.
veloped
i) Tm.! first one deals with the interface formed by monomeric units bound
to the silica surface. This interface must be characterized from the number
r of coi.act points of one chain with the surface. It is also characterized
from the listribution of positions of these contact points upon the surface
of partic'' s. The number r contributes to the determination of the free enthalpy of adsorption of one chain.
ii) -he second topic concerns the interphase formed by loops and tails.
The interphase determines the transition between the solid state of silica
particles and the liquid state of the polymer defined from siloxane chains
which ar' more freely because they are not adsorbed. The interphase must be
characterzed from the distribution of lengths of loops. Coated particles
which are not bridged to one another govern the flow process of mixtures chinracterizt
by low silica concentrations.
iii) An infinite cluster is formed when the concentration of particles
is high •iough and the polymer chains have an appropriate length. This infinite clui er results from a percolation process. It behaves like a permanent
gel when 11 free chains have been removed from the initial mixture. It must
be characderized from an effect of swelling induced by a good solvent or
from an elfect of uniaxial stretching.
II. THE INTERFACE
Fumed silica was bought from Degussa. Aggregate3 are made from elementary beads. The average diameter of one bead is about 140 A. Therefore, the
ideal specific area is 200 m
red from particles is 150 m
2
g -.
1
g-
However, the real specific area AT measu-
(B.E.T.). The discrepancy corresponds to the
Mal. Res. Soc. Symp. Proc. Vol. 171. '1990 Materials Research Society
reduction of surface used to build aggregates or agglomerates. The estimate
of the average number of silanol groups participating in the adsorption pro-
cess is
-
1 = 1.8 x 10 2 A 2. Silica particles have a fractal character ine
vestigated from neutron scattering [1]. The fractal exponent is D = 1.9.
II.1 - Average number of contact points
The average number of contact points of one chain with the silica surface can be determined from either of two experimental ways.
The first one corresponds to the measurement of the residual amount of
polymer Q
left bound to silica after removing all free chains [2]. This
r
quantity is measured as a function of the polymer molecular weight Mn, at
constant silica concentration Cs.. The weight Q0Z is defined per unit mass
n
sr
of silica ; it is found to obey a linear dependence upon the square root of
Mn :
Or = Xa M(1)
-3
with Xa = 4 x 10
1/2
(g/mole)
. This law holds for chain molecular weights
ranging from Mn = 1800 to Mn = 3600 x 10 2g mole
-1
. A simple analysis of this
result is given by assuming that the average number of contact points of one
chain is
< r
>
=
caVN
(2)
c
N is the number of skeletal bonds and Ea accounts for the effect of chain
stiffness. Then,
a
= Ea
AT
V--Mm
a
A (l.
e
A is the Avogadro number and
)3
a
Pa accounts for an effect of bridging between
particles ; Mm is the average molar weight of one skeletal bond. Values of
ca and P
are close to one.
a
The other experimental way used to determine <r c > is the magnetic relaxation of protons linked to PDMS chains. The time dependence of the relaxation function of the transverse magnetization exhibits two well defined parts.
The first one varies rapidly with time ; it is associated with protons linked to adsorbed monomeric units. Its amplitude is m. The second one shows
evolution with time. It corresponds to monomeric units forming loops and
tails ; its amplitude is mf. The ratio TB = mB/M B + mf gives the fraction
of monomeric units frozen by the adsorption process. For one chain
TB = (l+pa)y v ca /N /N
(4)
on average ; Yv accounts for frozen neighbours of monomeric units linked to
the surface. Another expression of TB is
TB
A T Mm Yv
T
e Or
=
(5)
367
The experimental value of the slope of the straight line describing iB as a
linear function of l/Q. is 0.09. The experimental value of the slope (i p a
Ea Y. in formula (4) is 3.3 . Considering that AT = 150 m g
and Mm = 37 g
I
mole- and pa = 4xi03 then the estimate of ca (l+P ) is 0.75 while yV
5.
a
It is considered that the condition of adsorption illustrated by formula (2) is firmly established from experimental results.
11.2 - Law of adsorption
The formation of the adsorbed layer results from a full immersion of
all particles in the polymer melt. Consequently, aggregates and agglomerates
are rapidly surrounded by polymer chains. This adsorption process is very
contrasted to that observed from particles in suspension in a polymer solution
. It is characterized by two main features.
i) The whole surface of silica is instantaneously in contact with PDMS
chains.
ii) PDMS chains compete with one another to be adsorbed on the surface.
Chains in a molten polymer are known to obey a Gaussian statistics ; any
chain formed from N skeletal bonds is swollen by a number of other chains
equal to VN, on average. Any time one chain is in contact with the surface,
/N other chains are also in contact with the same area of the surface.
The condition of adsorption <r > = ca/N has been recently given a simc
ple interpretation by assuming that any adsorbed polymer chain pictures the
random flight of a fictitious particle colliding a plane [3j. Considering an
area A T of the silica surface, the number of chains of N skeletal bonds forming rc contact points is :
A
vc(rc,N)
=
/Z i
cc
ae
£
) c
) N,1/2
.
- r/2N
e
(6)
a
Formula (6) is the simplified expression of a more complicated distribution
function (4]. Then, the total number of adsorbed chains is
vr
=
vc(rc,N) dr c
or
/2-/ AT
£T
T
(7)
e
this leads to formula (3) and (4) with
a
c=
CV-/2.
aa
11.3 - Clusters of contact points
The desc:iption of the adsorption process of one chain relies not only
on the determination of the number r of contact points with the surface of
c
particles but also upon the way these points are spread on silica. Positions
R (.j= 1, 2, ... rc of these contact points form a cluster. The radius nf
gyration of one cluster can be defined in the usual way
388
(AR )2
12r
1
(8)
i~i
The average of (G )2 over all adsorbed chains leads to the mean square size
of one cluster
2
<(A )
G
=
-2
c
(9)
A correlation function g(P) can also be defined to characterize the cluster
representing the binding site of one chain
g(o) is the probability that
a contact point a distance 0 apart from another contact point belongs to the
same chain. Although the detailed expression of g(p) is not known, yet, it
obeys simple properties :g(o) = 1 and
L g( )
a
Pa
/
(10)
The simplest guess about the distribution of contact points of one chain
is to assumed that it represents a two-dimensional random flight. Considering,
for the sake of simplicity, that the mean number of bonds in each loop is
2
T'N/Ea,
then the mean square step of the random flight is /N b /Ea ; consequently the mean area associated with the two-dimensional random flight is
2
E'
ll<r> V.N b /6ca
C
R N b'/6
=
c
a
=
(ii)
b is the average length of one skeletal bonds ; the quantity
pared with the area actually covered by one chain
0
<r > o
= Ea u VN
c
e
e
e
Thus,
a must be com-
2
/c
R N b /6<a * e
(12)
a c7e
This result shows that about /N chains are adsorbed upon the area N b'2
of the silica surface.
Ca
111.
THE INTERPHASE
Properties of the interphase. are determined by the structure of loops
.ind tai 1s, i .,.
by the distribution of lengths of corresponding chain segments.
o11.1
- Averag,
thickness
NeuIlct in? the effect of bridging of particles, due to polymer
chains,
the avorage thickness e ot' the adsorbed lay,r is estimated from the specific
area AT and th- amoljnt of polymer Qr bound to silica
Or/A T a
e
it
mus;t
11.2
!r
-
truct ur>
of
)f the chain molecular weight.
Isp
By noglocting tails,
loops carl b(, described within the framework used
:2ection 11.2. Thus, a given configiration of one adsorbed chain is
-harac-
leriz1d
sd
vary as the square root
(13)
from
t,/ rf. crnttict
r.
f
skoeetil
rood
poinis
hndoi
fom
n.iR;
i.e.
r
f)
c
loops
the multiplicity
-
..
1,
L,).
The total
of a
loop
for-
number of dif-
389
ferent kinds of loops is LC . Then
L
c
8J
j=l B.
r
=
n B
and
N.
The statistical weight of this configuration is
r/2
r2
f(N,nj,B;rc )
with
=
"
c,
36./2
, Lc i/(B!(n)
j=l
j
)
(14)
L
C
c n(N,njB.;rc)
j=l
and
L
r
ad (N;r
c
)
cc
=
R (N;r,)
(15)
/2/s (r /N'/z) exp(-r 2
/2N)
c
The distribution function (14) leads to the description of specific properties of the adsorbed layer such as the total number of loops longer than
a given number n , for example.
The adsorbe layer plays a crucial role in the determination of viscoelastic properties of silica-siloxane mixtures. The degree of stiffness of
loops controls the elasticity response of these systems.
IV. GEL BEHAVIOUR
An infinite cluster is formed through a percolation effect when the concentration of silica particles is high enough and polymer chains are not too
short. For example, strong gels
obtained from silica concentrations are
higher than .09, .17 and .29 (w/w) when chain molecular weights are 3600x10',
500xlO' and 70xlO' g/mole-' , respectively.
The random lattice formed from silica aggregates behaves like a permanent gel after removing all free chains. The gel is permanent as long as the
constraint applied to the mixture is weaker than a given threshold determined
from the number of contact points and the lengths of loops. One of the most
characteristic property of these permanent gels is the swelling effect observed in the presence of a good solvent. The maximum swelling ratio Om depends
upon the texture of loops connecting silica particles. The sw 'ling ratio Qm
is defined by dividing the volume of swollen polymer in the mixture by the
volume of dry polymer. The volume of silica is not taken into consideration
to calculate Om. As it is well known from the simplest descriptions of the
swelling effects of ordinary polymeric gels, an estimate of Om is obtained
by considering that the effect of osmotic pressure is in equilibrium with
the effect of elasticity of chain segments. The main difficulty encountered
in describing the swelling process of silica-siloxane mixtures concerns the
determination of the length of active chain segments involved in the cohesion
of the gel. A simple way used to overcome this difficulty is to consider that
the number of active segments is equal to the total number of hydrogen bonds
wA = A T/ e formed on the silica surface ; while the average length of a segment is given by Q/Ppeo, ep is the pure polymer density
. Then. according
to the theory proposed by Flory, the swelling effect is well described from
Q'/' _
m
102 = Qlvpoo(6
Or/vsP H
(I)
where vs is the molecular volume of solvent. This description has been found
to be in agreement with experimental by varying both Or and the number of
links between the silica surface and POMS chains. This number is v;aried by
treating the silica surface to prevent silanol groups from forming hydr,,n
bonds.
370
V. KINETICS OF ADSORPTION
The amount of polymer adsorbed on the silica surface is a time dependent
function Orit) until an equilibrium is reached. At the end of the mechanical
mi xing most part of the surface is covered by PDM'S chains. The surface, is
nearly saturated by this initial amount of polymer hereafter called Qr(o).
However, chain segments defining loops may be too stretched because of the
presence of constraints exerted during the mechanical mixing. The structure
of the binding site of any given chain is supposed to vary with time until
an equilibrium between retractive forces and hydrogen-bonds is obtained. This
corresponds to the completion of the surface saturation. The description of
the adsorption process relies upon two hypothesis. It is considered as resulting from a conventional diffusion process occuring in the presence of an
. Also, it is supposed to obey an excluded surface efabsorbing screen
fect : the adsorption process occurs on free parts of the surface, only. Then,
Qr(t)
Or - (Or-Or(o))exp -/t0
17)
,
typical values of t are l0-1O hours ; they ,pend upon the temperature of
adsorption. End-hydroxylated chains are more idpidly adsorbed than end-methylated one i5 . Th,- increase of the concentration of silica has been found
to increase the rate of adsorption of PDMS chains. This experimental result
has been given a simple interpretation. Considering long polymer chains, toe
addition of silica particles induces an effect of bridging between particles.
The limit value of th, amount of adsorbed polymer is expressed as :
IL
I'
)
(l..
)
181
r
ai
r
a
where Q correspo)nds to 'h-' maximum amo unt of pol'ymer adsorbed on the surface,
without any bridging effect. The parameter sa depends upon the concentratln
of silica. Then, it can be shown that the time constant t is also a function
. More precisely
of
a
a
0 is measured in the absence of bridges between particles. The value of the
ratio
)
a
r a
has b-,n estimat-1. f-r thr-e concentrations of silica : 3., 2.5 and 3.2 for
C5
.)9, .17 and .2) w/w, resp-ctively. The increase of the rate of adsorption upon ;Adiionorf si li,-aresults mainly from the effect of bridging induc,,dby polymer chains. The activation energly of adsorption corresponds to
about 18 Kcu l moleV[ . CONCLUSI0N
,
Th mixing of silica particles to siloxane chains does not correspond
!, a simple addition of hard objects to a fluid medium. The elementary process
of adsorption, is ,-ll locilized by the formation of one hydrogen bond ; th',
rrerrsponding energy ,f adsorption is hardly higher than the thermal energy,
a, room temperature. This process induces a polymer-silica interaction well
appropriate to the det'rmination of a broad range of properties of the mixtures. The fractal Tharacter of particles is also involved in the specific behaviour (f this finely divided matter.
REFFFNCF
(1) -;,hif,,r, !LW.,
Mt."Bulletin (11986),
(' Cohen-Adad, .1. .,
() Cohen-Addad,
(41 Feller, W. "An
, Cohen-Addad,
j.P.,
Roby, ('.
XIIi, 22.
and Sauviat,
Polymer (1q89),
M.,
Polymer (19815),
26,
1231.
30, 1820.
Iltroduction to Probability Theory and its App)ications"(1968)
I.P.,
lluehot,
P., Jost, P.,
Pouchelon, A.,
lolym.(1989),30,143.
371
THE EFFECT OF MASKED ISOCYANATES ON THE MECHANICAL
PROPERTIES OF MY720/DDS EPOXY RESIN
N. RUNGSIMUNTAKUL, S.V. LONIKAR, R. E. FORNES*, AND R.D. GILBERT
North Carolina State University, Fiber and Polymer Science Program, and
*Physics Department, P. 0. Box 8202, Raleigh, NC 27695-8202
ABSTRACT
The mechanical properties of epoxy resins and epoxy resin/graphite fiber
composites are adversely affected by moisture absorption. Incorporation of
masked isocyanates that unmask to release isocyanates in situ at the cure
temperatures (150-177°C) reduce the equilibrium absorption up to -70%.
Dynamic mechanical analyses and stress-strain properties of epoxy resins
containing masked isocyanates were examined to determine their effect on
mechanical properties. The ultimat Tg of the epoxy is reduced by incorporation
of masked isocyanate, but the actual Tg is comparable to the "as cured" Tg of the
epoxy. The dynamic moduli up to the Tg are relatively unaffected. In a number
of cases, the initial modulus, elongation at break and peak stress are equal or
better than the unmodified rcsins.
INTRODUCTION
In structural and aerospace applications, epoxy resins are used as the
matrices in fiber reinforced composites that have high strengths and moduli and
are light weight. The adverse effect of moisture on mechanical properties is one
major drawback of epoxy resins. Upon curing, the epoxy-amine systems generate
hydrophilic groups such as hydroxyl, secondary and tertiary amine groups and
residual oxirane and primary amine groups (and in some cases sulfone) that can
interact with water. Water is also absorbed into the unoccupied volume of the
cured epoxy resins. The sorbed water plasticizes the resins, lowers the glass
transition temperature (Tg), causes swelling, induces stresses, chemical bond
cleavage and debonding of the fiber-matrix interface. Enhanced craze initiation
and crack propagation reduce the serviceability of composites. Attempts to
reduce moisture absorption of epoxy have been reported by several workers. (See
Fisher et al.[ 11, Hu et al. [2] and references therein.)
Fisher et al. [1 showed that the polar functional groups in cured
tetraglycidyl-4,4'-diaminodiphenyl methane (TGDDM) - diaminodiphenyl
sulfone (DDS) (73/27 wt%) thin films reacted with a, o, ox-trifluoro-m-tolyl
isocyanate (MTFPI) in dimethylsulfoxide (DMSO) at 700C, and reported a 54 %
reduction in moisture absorption. Hu et al. [21 obtained 75 and 69 % reductions in
equilibrium water absorption by reacting the hydroxyl, amine, and epoxide
functional groups of cured TGDDM-DDS epoxy thin film (20-50 pm thick) with
pentafluorobenzoyl chloride and 2, 4-difluorophenyl isocyanate in N,
N'-dimethylacetamide, respectively. The reactants were allowed to diffuse into
Mal. Ai.
So.
Symp. Proc. Vol. 171. , 1990 Materials Research Society
I,,
4
372
the resin films which had been previously swollen with solvent. This technique is
not feasible with the thicker films >150 pm) because the films crack due to
successive swelling.
The objectives of this study were to study the effect of masked isocyanates,
incorporated into TGDDM-DDS epoxy prior to curing, on reduction in moisture
absorption and on the mechanical properties of the er')xy and its composite.
Dynamic mechanical analysis and tensile tests were employed.
EXPERIMENTAL
ii
-Instron
MY720 Araldite® Epoxy (primarily TGDDM) was placed into a mixing jar
at 105-1 10*C. The calculated amount of DDS (27 wt%) was added slowly into the
mechanically stirred TGDDM. The mixture was stirred until all the DDS was
apparently dissolved. The prepolymer was degassed at 110*C for 30 min, cooled
in a desiccator, removed and stored in nitrogen filled bags in a refrigerator at
50C.
The masked isocyanates were synthesized according to procedures reported
by Lonikar et al. [3] from cyclohexyl isocyanate, hexamethylene diisocyanate or
phenyl isocyanate masked with alcohols, phenols or fluorinated phenols (Table 1).
They were incorporated into the epoxy by slowly adding to the heated prepolymer
(105-118*C) and stirred until completely dissolved.
Thin, bubble-free films with uniform thickness were prepared by curing
the prepolymer between the extra smooth Teflon( sheets. The sample thickness
was controlled by using a MylarO film as a spacer (0.02cm thick). The samples
were degassed at 1100 C for 30-45 min and cured at 150'C for 1 h and 177°C for
5 h under a nitrogen atmosphere.
The graphite/epoxy composite samples were prepared by hand-winding
graphite fiber tows around a 9cm x 5cm x 0.1cm Teflon® window to keep the
graphite in place. The prepolymer was applied on both side of the tows and cured
in the same manner as the resins.
Water absorption of the cured epoxies specimens immersed for 6 months
was measured gravimetrically at 25°C. The dynamic mechanical analyses (DMA)
were made with a Rheovibron Viscoelastometer DDV-II-C (Toyo Baldwin Co.,
Japan) equipped with a Autovibron automatic system (Imass, U.S.A.), at a
frequency of I I Hz and a heating rate of 2.5*C/min over the temperature range of
- 120 to 300'C. The Tg was defined by the temperature at the maximum of the
loss tangent spectrum (a-transition). The tensile tests were performed using a
tensile tester at a speed of 5 mm/min (ASTM D 882) and a gauge length of
2.54 cm.
RESULTS AND DISCUSSION
DSC and IR verified hat the masked isocyanates reacted with the epoxy [31.
No unreacted epoxides remained in the system following the cure at 177'C. The
I
373
DMA and reductions in water absorption are listed in Table 2. In general,
significant reductions of water absorption were obtained. The highest reduction,
70%, was obtained with masked isocyanate #9.
A typical DMA spectrum of the "control" epoxy is shown in Figure 1a
where the storage modulus (E') and the loss tangent (tan 5) are plotted as a
function of temperature. The DMA results agree well with those previously
reported [4]. There are 3 transitions in the tan 8 spectrum: a y- transition around
-60°C, attributed to the crank shaft rotational motion of the glycidyl portion of
the epoxy after reaction with curing agent [4,51; an a-transition at 180-200'C or
the "as cured Tg" associated with vitrification before the epoxides fully react; and
an a-transition around 285°C or the ultimate Tg (Tgoo) associated with the fully
crosslinked system. The average elastic modulus E' at 25°C is 2.16 x 1010
dynes/cm 2 (for a Length/Area (L/A) = 901 cm- 1) and is 1.71 x 1010 dynes/cm 2
(for LA= 702 cm-1). L/A affects the modulus value: the higher the L/A the
higher the apparent modulus [4] . Approximately the same L/A was used in order
to compare the moduli of the "control" and the modified epoxies.
In general, the DMA spectra of the epoxies containing masked isocyanates
have 2 transition peaks: a y-transition around -60'C and an oa-transition
approximately 38 to 130'C lower than the Tgoo of the "control" epoxy. The
a-transition peak of most of the modified epoxies is broader and has a higher
magnitude than that of the "control" epoxy. The broadening of the transition
region on the tan 8 spectrum can be attributed to a broader molecular weight
distribution and a plasticizing effect of the masked isocyanate [6]. (The magnitude
of tan 8 is related to the distance between crosslinks, Mc, as well as the presence of
plasticizers.) The increase in magnitude of tan 5 max with increasing Mc (or lower
crosslink density) reflects the increasing ability of polymer to absorb energy as
the molecular constraints are reduced; the greater separation of crosslinks permits
greater mobility of chain segments 17 1. The incorporation of the masked
isocyanates #2, #7, #8, and #9 caused a higher magnitude of the tan 5 peak.
Therefore, the epoxies containing these masked isocyanates probably have a lower
degree of crosslinking than the "control" epoxy. The epoxies containing masked
isocyanates #5 and #10 have a lower magnitude of tan 5 and a much broader peak.
This may be due to a broad molecular weight distribution between crosslinks or
to heterogeneous crosslinked structures. Since both masked isocyanates have low
unmasking temperatures, they probably unmasked and reacted with TGDDM or
DDS before all the DDS reacted with TGDDM.
The DMA spectrum of the epoxy containing 26 mole % of masked
isocyanate #7 is shown in Figure lb. The Tgoo is about 177°C, and the E' is 2.35 x
1010 dynes/cm 2 (L/A = 1066 cm- 1). Masked isocyanate #7 unmasks at about
180'C. At this temperature, most of the TGDDM probably has reacted with DDS.
The masked isocyanate #7 therefore reacts with the OH and NH groups on the
epoxy rather than reacting with the TGDDM. The result is a more uniform Mc
than the epoxy containing masked isocyanate #5, and the tan 5 peak at transition is
narrower. The Tg- is lower than the "control" epoxy, possibly because of the
plasticization of the polymer network and the flexibility of the crosslinks
involving hexamethylene groups of the diisocyanate. However, the room
374
Table I
Masked isocyanates Data
Masking Agent
No.
Isocyanate
1
a,t,-Trifluoro-otolyl isocyanate
Phenyl isocyanate
Phenyl isocyanate
Hexamethylene
diisocyanate
Hexamethylene
diisocyanate
Hexamethylene
diisocyanate
Hexamethylene
diisocyanate
Cyclohexyl isocyanate
Cyclohexyl isocyanate
2
3
4
5
6
7
73
Phenol
FC-10**
FC-10**
125
88
105
130
130
185
n-Butanol
88
120
Pentafluorophenol
169
170
Phenol
136
180
137
.148
130
130
83
120
11
12
*
**
Solution unmasking temperature
FC-10 = Ci.5F16SO2N(C2H5)CH2CH20H
10
Unmasking
Temp.(OC)
Pentafluorophenol
Phenol
Pentafluorothiophenol
Cyclohexyl isocyanate 2,4-Difluorobenzylalcohol
ATBN
Phenyl isocyanate
Phenyl isocyanate
DDS
8
9
Melting
Temp.(°C)
-
-
-
173
175
Table 2
Ultimate glass transition temperatures, loss tangents, elastic moduli and reduction in
water absorption of the "control" epoxy and epoxies containing masked
isocyanates, and the Tg* of the composites
Masked
Isocyanate
Number
Mole
(%)
Wt
(%)
Control
Resin
Tg =
(OC)*
tan 5 max
285**** 0.48
2
5
20
20
28
22
183
247**
1.11
0.32
7
26
29
177
0.80
49
20
20
20
44
28
37
33
9
35
155
192
169
246**
228
243
0.93
0.99
1.04
0.39
0.77
0.65
8
9
10
11
12
*
*
**
-
Temperature at which tan 8 is maximum
Broad tan 8 peak
Short autovibron sample (L/A <718 cm -I)
The "as-cured" Tg is -180 0 C
E' at 250C
(x10-10
2
dynes/cm )
2.16
1.71 **
1.51***
2.30
1.78***
2.35
2.09***
1.83***
2.15
2.10
1.83
1.53
1.50
% Reduction
in water
absorption
Composite
Tg0 °
(OC)*
-
286
40
17
187
226
39
186
55
45
70
30
0
11
186
171
199
242
246
375
Table 3
Tensile properties of "control" epoxy and epoxies containing masked
isocyanates
Masked Wt% Peak stress
Iso(x 10-8
cyanate
dynes/cm 2 )
Number
Control2
28
5
22
7
29
8
28
9
37
10
33
11
9
12
35
Change
in
Stress
(%)
6.5
6.7
6.5
8.1
4.0
3.5
6.2
4.6
6.2
-
+3
0
+25
-38
-46
-5
-29
-5
Peak Change
Strain in
(%)
Strain
(%)
Initial
Modulus
(x 10-10
2
dynes/cm )
Change
in
Modulus
(%)
8.2
8.5
8.2
8.4
4.7
4.6
8.2
7.8
10.2
1.6
1.7
1.6
1.6
1.5
1.7
1.6
1.2
1.2
+6
0
0
-6
+6
0
-25
-25
-
+4
0
+2
-43
-44
0
-5
+24
00
-
-
-'
o
Figure 1
,
,
a
...
g
Dynamic mechanical spectra of (a) "control" epoxy (b) epoxy containing masked
isocyanate #7
I
isocyanate #7
0c
376
temperature E's are higher than the "control" epoxy, possibly because of a
reduction in the matrix free volume which would restrict the short-range motion
of the matrix polymer chains [8-10].
The DMA spectra of T300/epoxy composites containing masked isocyanate
#7 are shown in Figure 2. The Tgo, tan 5 and E' of the composites are
summarized in Table 2. The Tgoo's of most of the T300/epoxy containing masked
isocyanates are about the same as the epoxy resins except for those containing
masked isocyanates #5 and #10. The Tgo's of the latter two are -20'C lower than
the Tg's of most the modified epoxies. The tan 8 peaks of these two composites
are narrower than those of the corresponding resins. The reason for this change
is not clear.
The magnitudes of the tan 8 peaks at the a-transition of all the composites
are slightly lower than those of the corresponding resins. However, the
magnitudes of the tan 8 peaks of the composites containing masked isocyanates are
higher than the "control" composite, which suggests that the former had a lower
extent of cure.
The E' of all the composites found in this study are lower than those of the
corresponding resins. This may be due to 1) low length/area ratio, 2) the
composites were manually prepared and therefore they were non-uniform, and
perhaps 3) poor adhesion of the epoxy to the fibers. Also, the E' of the modified
composites should not be compared with the "control" composite because of
difficulty in controlling the weight ratio of the graphite fiber to epoxy. However,
from the DMA data of these hand-prepared composites, no major adverse effects
are caused by the incorporation of the masked isocyanates on the interface
between the fibers and the matrix.
The peak stresses, peak strains and initial moduli form tensile
measurements of films of the modified epoxy resins are reported in Table 3. The
typical stress-strain diagram of TGDDM-DDS epoxy reveals a yield point, which
was also observed by Morgan [11 ]. The initial moduli of epoxy films containing
masked isocyanates #2 and #9 were about 6% higher than the "control" epoxy.
The initial modului obtained with the specimens having masked isocyanate #5, #7
and #10 incorporated were about the same as that of the "control" epoxy, but for
the specimen containing masked isocyanate #8 was about 6% lower. The lowest
moduli were obtained from epoxies containing 9 wt% of ATBN-PhNCO and 35%
DDS-PhNCO (-25%). In general, the tensile moduli are in relative agreement
with the DMA results.
Epoxies containing masked isocyanates #2 and #7 had higher peak stresses,
peak strains and initial moduli than the "control" epoxy, and therefore higher
toughness than the"control" epoxy. Epoxies containing masked isocyanates #5
and #10 had about the same peak stresses, peak strains and initial moduli as the
"control" epoxy, or about the same toughness. Epoxies containing masked
isocyanates #8, #9 and #11 had lower peak stresses and peak strains than the
"control" epoxy. Epoxy containing #12 had a lower peak stress and initial
modulus but higher peak strain than the "control" epoxy.
377
SUMMARY
The incorporation of a series of masked isocyanates into TGDDM-DDS
epoxy results in specimens with properties: 1) with Tg's that are 38- 130°C lower
than the Tgoo, but are about the same as the "as cured" Tg of the "control" epoxy,
2) with dynamic elastic moduli at room temperature up to the Tg and initial tensile
moduli of the epoxies containing masked isocyanates that are generally about the
same or higher than the "control" epoxy.
The DMA of T300 graphite fiber/epoxy composites (90°-orientation) made
from epoxies containing masked isocyanates showed that the Tg's were about the
same as the corresponding unfilled epoxies. There was no evidence of adverse
effects of -iie masked isocyanates on the interfacial region.
REFERENCES
1. C.M. Fisher, R.D. Gilbert, R. E. Fornes, and J. D. Memory, J. Polym. Sci.:
Polym. Chem. Ed., 23, 2931 (1985).
2. H.P. Hu, R. D. Gilbert, and R. E. Fornes, J. Polym. Sci., Part A, Polym.
Chem. Ed., 25 (5), 1235 (1987).
3. S. V. Lonikar, N. Rungsimuntakul, R. D. Gilbert and R. E. Fomes, J. Polym.
Sci., Polym. Chem. Ed., accepted for publication.
4. T. W. Wilson, R. E. Fornes, R. D. Gilbert, and J. D. Memory, J. Polym. Sci.:
Part B: Polym. Phys., 26, 2029 (1988).
5. J. D. Keenan, Master Thesis, University of Washington, Seattle, 1979.
6. L. E. Nielsen, Mechanical Properties of Polymers and Composites, Vol. 2,
Marcel Dekker Inc., New York, 1974.
7. T. Murayama and J. P. Bell, J. Polym. Sci., Part A-2, a, 437 (1970).
8. N. Hata, R. Yamauchi and J. Kumanotani, J. Appl. Polym. Sci., 17, 2173
(1973).
9. N. Hata and J. Kumanotani, J. Appl. Polym. Sci., 21, 1257 (1977).
10. J. A. Sauer, J. Polym. Sci.,Part C: Polym. Symp., 32, 69 (1971).
11. R. J. Morgan, in "Advances in Polymer Science No. 72", K. Dusek, Ed.,
Springer-Verlag, New York, 1985, p. 1 .
379
A STUDY OF SHORT METAL FIBER REINFORCED COMPOSITE MATERIALS
W. C. Chung
Division of Technology, One Washington Square, San Jose State
University, San Jose, CA 95192-0061
ABSTRACT
Over the years, the conventional involvement of short
fiber reinforced composites in electrical applications has
been as electrical insulation. Contrary to this approach,
with the increasing need of bettei electromagnetic
interference (EMI), radio-frequency interference (RFI)
shielding and control of electrostatic charge distribution
(ESD) for computer, defense, space exploration and some
high-tech structural components, it is expected that the
development of conductive polymeric composite materials will
grow strongly and significantly.
An experimental investigation is designed to
systematically evaluate the mechanical properties and
electrical properties of metal fiber reinforced composites
subjected to various loading conditions. In this study,
chopped Inconel 601 (nickel base) metal fiber with a fixed
fiber aspect ratio (length/diameter ratio) is used to
reinforce commercially available thermoset polyester resin.
Mechanical testing of custom made samples, failure analyses
using visual inspection, light microscopy and SEM are
conducted to understand the fracture behaviors and possible
failure causes in such composites. The feasibility of using
metal fiber polymeric composites in structural/electrical
applications is discussed in this paper.
INTRODUCTION
Polymers have been well known for their electrical
insulating properties and great strides have been made in
electrical and electronic applications, mainly related to
electrical insulation. Consequently, research has been
directed to improve the dielectric strength of polymers so
that they can be used for better insulators. In the past few
years, with the advent of electrically conductive polymers,
their potential to perform as active roles in conducting
electricity has been discovered and realized (ref. 1).
Recent
polymer researches have revealed that polymers can indeed
conduct electricity as well as metals. Now the electrically
conductive polymers can be used as antistatic coatings, fuel
cell catalysts, solar elcctrical cells, photoelectrodes in a
photogalvanic cell, protective coatings on electrodes in
photoelectro-chemical cells, and as light weight, inexpensive
batteries.
Research Society
Mat. Re. Soc. Syrp. Proc.Vol. 171. '1990 Materials
380
Due to the increasing need of light weight, low cost,
moldable, and high specific strength for defense and high-tech
applications, it is expected that the development of
electrically conductive polymeric materials will grow
rapidly. It ir understood that conductive plastic housings
and molded parts can be beneficial to the controls of
electromagnetic interference (EMI), radio-frequency
interference (RFI) shielding and electrostatic charge
discharge (ESD) distribution. Advanced research studies have
shown that there are three possible methods to make polymers
conductive. The first approach is to apply a thin conductive
coating onto the molded part. This approach, however, is
costly and not efficient because of involving a two-step
operation which increases the difficulties in obtaining a good
adhesion as well as a uniform coating. The second approach is
Synthesis is
often held by synthesis or by doping (ref. 2-5).
done by side reactions. One of the major side reactions
involves the benzene ring. Other reactions lead to branched
and cross-linked polymers. Doping involves oxidation and
reduction reactions. This method, usually produces polymeric
compounds such as polyacetylene and polyphenylene, although is
proved to be effective and has been widely used, problems rise
from conductive polymers themselves such as their
processability, stability, mechanical and physical properties,
etc. The last approach, proposed in this study, is to
incorporate electrically conductive fillers in the polymeric
resin matrix. Many conductive materials surh as :arbon,
metals, metal-coated fillers in the form of powders, flakes,
particles, particulates, and fibers can be randomly dispersed
into a resin matrix and form a so-called "conductive
This approach so far appears to be a viable
composite."
solution to the development of conductive polymers. Due to
the lack of systematic research study in this area, material
properties are hardly found in the application of such
materials. Much research is urgently needed to fully
understand the interrelationships among structure, property
and processing prior to their commercial utilizations (ref.
1-6).
The conductive polymeric composite was first presented in
He used silver particles,
1966 by Garland (ref. 7).
approximately 50 to 200 microns in diameter, to reinforce a
thermoset phenol-formaldehyde (Bakelite) resin matrix. His
experimental data indicated that metal-filled polymers undorgo
a sharp transition from an insulator to a conductor at a
In his study
critical volume concentration of metal fillers.
the electrical resistivity remained almost constant until the
silver volume concentration of 38% is reached - then it droped
drastically and the whole composite became an electrical
conductor. Since Gut land's work, many other researchers have
reported different sharp transition from insulators to
conductors at different volume concentrations (ref. 8-15).
Among their studies, Dearaujo and his co-investigator (ref.
10) had found that normally at least 40% volume fraction of
metal fillers was needed in order to make a composite
conductive.
381
Recently, because of short fiber reinforced composites
can offer design flexibility, weight reduction, energy savings
and high-volume production for structural applications, they
are widely used in automotive, recreation, business machinery,
electrical appliance, and military applications. Metal fiber
reinforced composites become highly desirable to meet the
aforementioned requirements not only for load-bearing
capability but electrical conductivity as well, which normally
metal particle reinforcement cannot achieve. However, not
much work has been done in this area. Experimental data were
found only limited to individual cases. Davenport (ref. 16)
mentioned in his study that the metal fiber length (L) to
diameter (D) ratio (known as aspect ratio) in a composite must
have 100 or more in order to induce electrical conductivity.
He demonstrated the electrical conductivity should be a
function of L/D. In addition, the fiber packing density is a
significant factor which is closely associated with the ratio
of L/D (ref. 17).
Bigg and stutz investigated a stainless
steel fiber (8 microns in diameter, aspect ratio: 750)
reinforced ABS system, and found that the composite had an
electrical resistivity of 0.70 chm-cm at the fiber volume
concentration of 1% (ref. 18).
They also claimed in their
research that a highly conductive composite can be achieved
with a low concentration of metal fibers by simply using high
aspect ratio fibers. Their work, although seems very
promising, yet needs to be proved. Most of the metal fiber
reinforced composites were emphasizing on the electrical
properties rather than the mechanical properties.
Nickel has long been considered as preferred metal
because of its low electrical resistivity. In this study,
Inconel 601 nickel based fiber with a dimeter of 8 microns and
an aspect ratio of 125 was heavily used to reinforce a
commercially available thermoset polyester resin. Composite
samples were made in coupon shapes depending on the test
requirements. Both mechanical and electrical measurements
were further conducted to help understand the micromechanical
behavior as welL as electrical conductivity.
SPECIMEN PREPARATION AND TESTING
Chopped Inconel 601 metal fibers were donated by Bekaert
Fiber Technologies. Tu prevent the sizing effect from the
interfacial bonding between fiber and rein matrix, a thin
watex-soluble PVA (polyvinyl 4lcohol) coating originally
attiched to fibers was removed from Inconel fibers prior to
the process. Fiber volume cocentration, varied from 0% to
50%, was carefully controlled as material parameter to conduct
this study. Metal fibers were completely mixed with
appropriate amount of polycster resin and MEKP (Methyl Ehtyl
Ketone Peroxide) c:uring agent in a cheamical beaker based on a
predetermined volume ratio. The mixure was then poured into
an aluminum mold for cure. Traditional compression molding
practice was employed in the curing process. pressure was
around 17 psi (1.17 x 10- Pa) and temperature war set at
3561F (180^C). Specimen dimensions were carefully prepared
accoiJing to ASTM standard test methods.
382
RESULTS
Table 1.
Tensile Test:
Fiber Ratio
(v%)
0
5
10
15
20
25
30
35
40
Table 2.
Tensile Strength
(MPa)
55
50
41
42
44
51
47
43
38
Young's modulus
(Gpa)
23
23
23
25
23
26
25
29
29
Impact & Flexure Tests:
Fiber Ratio Impact Strength
(J/m)
(V%)
17
0
17
5
Flexural Strength
(MPa)
65
68
10
18
66
15
20
25
30
35
40
18
21
19
22
25
23
68
70
71
71
71
72
Electrical Measurement:
Resistivity: 6.79 x 10-6 ohm-cm
Very high resistance at fiber ratios below 30%
Resistivity -- 1.0 ohm-cm at 45%
383
The tensile and flexure tests were performed in a
screw-driven computer-asisted Satec testing machine. A
testing speed of 0.1 in./min. (2.54 mm/min.) was used for
tensile and flexure tests. The ASTM method D257 was also
followed to measure the volume resistance of each sample.
The test data were collected and discussed in the following
sections.
RESULTS AND DISCUSSION
Tensile test data, as shown in Table 1, have demonstrated
that fiber concentration can indeed increase the tensile
strength of the composite. Young's modulus is also improved
as well. It is interesting to note that the small fiber
concentration at the ratio lower than 10 volume percent will
not contribute to the increase of entire tensile strength.
According to the study, fiber concentration at 25 % has the
maximum UTS.
It is found that fiber orientation along the
pulling direction will have significant effect to tensile
properties. Since the specimens are prepared through a
casting process, the fiber orientation in all direction is
assumed the equal.
Impact test data (in Table 2) reveal that
impact strength increases with the addition of metal fibers.
However, there is a limitation set at 35%.
Low fiber
concentration impairs the impact strength of the composite.
Optical microscopy indicates that because of the existing of
metal fibers, small air bubbles are attached to fiber ends,
which is believed to be responsible for the degraded impact
strength. Three point (flexture) test data show that fiber
fillers can improve the flexural strenth of the composite, as
shown in Table 2. It is also noticed that metal fibers can
dissipate some energy in a crack propagation. In other words,
with the addition of metal fibers the crack pattern of a given
composite shifted from a pure tension failure mode toward a
more shear failure mode, which increases the flextural
properties. Fiber pull-outs and fiber breakage are some
evidence. In the electrical measurements, a critical fiber
concentration is recorded. Electrical resistivity of 1.0
ohm-cm is measured at the fiber volume ratio of 45%, which is
unexpected high. However, when fiber concentration falls
below 30% the electrical resistance remains almost constant,
that is the composite is an electrical insulator. In this
study, metal fiber reinforced composites did undergo a sharp
transition which is inconcert of Garland's work (ref. 7).
CONCLUSION
Because of excellent electrical conductors, metal fibers
are suitable additives for inducing electrical conductivity in
traditionally known insulators, polymer materials. Inconel
metal fibers, although proved to be effective reinforcing
elements, are considerably denser than expected.
384
It is found, during this study, that the explanation of
the fracture behavior in a metal/polymer composite is often
diff.::ult to make, because it involves with many unseen
factors such as stress concentration, orientation effect,
viscoelastic behavior utc..
While significant progress has
been made, much work still needs to be done. A systematic
approch including experimental and theoretical techniques
should be developed to help understand the micromechanisms and
to elucide the interrelationships among structure, property
and processing.
Several factors such as fiber concentration,
fiber aspect ratio, interface compatibility between fiber and
marix can then be studied accordingly. The aforementioned
suggestions, if applicable, may lead to a complete data bank
setup which may eventually benefit all the designers,
engineers, and scientists who are using conductive composites
in their work.
ACKNOWLEDGEMENTS
The author gratefully acknowledge the support provided by
Mr. Steven J. Kidd of Bekaert Fiber Technologies, Marietta,
Georgia. for Allowing the use of Inconel 601 steel fibers.
The author is indebted to Mr. Jeff Warnock and Mr. Bart
Wensink of Sat. Jose State University students for their
dedicated work on specimen preparation. Informative
discussionrs regArdinig this work with Mr. Steven J. Kidd are
also gratrfully appreciated.
REFERENCES
1. Kanur, P. S. and Hac:Diarmid, A. G., P!:t~cs That Conduct
ElectLicity, Fci. Amer., Fob. 1938, pp. 106.
2. Blythe, A. R., E1~ctrizal Properties of _Polym,,rs,
CasrLid(;,. "ni:,.'r'ity Press, New York, 1979
2.
Chidscy. C. arid Murray, R. W..
Elctruactive Polymers ind
Macromoleculai Eectronics, Science, Jan. 3. 1986, pp.
25.
4. Davidson, T.,
Poljmers
in El.ztronics, American
Society, Washinqtczn P. C.,
Chemical
19 s'1.
5.
Ferrar,),
1. F.
arid W;iiazs-,
J. M.,
Introduction to
synth,~ti: E.c.,icu'
Ccnductr,
,\cad-nic: Piers. New
: s.
y: k . 1)81.
-.
Skuthein ..
.. ,andbouk
2nd 2
MaIL1l D,_)kvr,
"
Ai, i
ir :.|.
,
.,
rr
'.
M. ,
ti
J(;'
.
New
1
of Conducting
Y.rk, 193C.
1:e. Soc.
uc
App!.
Phys.
AI2IE.
,
42,
Polymers,
2 '5,
1972,
1966,
vol.
pp.
pp. 2463.
1
642.
385
9. Kusy, R. P. and Corneliussen, R. D.,
1975, pp. 107.
Polym. Eng. Sci.,
10. DeAraujo, F. T. and Rostenberg, H. M.,
D, Appl. Phys., 9, 1976, pp. 1025.
11. Nicodemo, L.,
Jour. Phys.,
15,
Sec.
18, 1978, pp. 293.
et al., Polym. Eng. Sci.,
12. Shorokhova, V. I. and Kuzmin, L. L., Soy. Plast., 3, 1965,
pp. 26.
13. Scheer, J. E. and Turner, D. J., Adv. Chem., 99, 1971.
14. Bigg, D. M.,
composites, 10, 1979, pp. 95.
15. Kwan, S. H.,
et al., Jour. Matl. Sci.,
16. Davenport, D. E.,
17. Milewski, J. V.,
Polym. Sci. Tech.,
15, 1980, pp. 2978.
15, 1981, pp. 39.
Ph.D. Thesis, Rutgers University, 1973.
18. Bigg, D. M. and Stutz, D. E.,
Polym. Comp.,
19. Edwards, J. H. and Feast, W. J.,
Polymer,
4, 1983,pp.40.
327, 1984.
20. Reyonlds, J. R., Chemtech, July 1988.
21. Krieger, J.,
Chem. Eng. News, June 1987.
22. Cotts, D. B. and Reyes, Z., Elec. Conduc. Organ. Polym.
Noyes Data Corp., New Jersey, 1986.
dv. Appl.,
PART VIII
Miscellaneous /Conventionlal
Composites
389
DEFORMATION BEHAVIOR 0OF POLYMER GEL",
IN ELECTRIC FIELD
Toshio Kurauchi. Tohru Shiga. Yoshiharu Hirose and Akane O(kada
TOYOTA Central Research and Development laboratories Inc.
41-I. Nagakute, Aichi, 480-11, JAPAN
ABSTRACT
The deformation of polylacrylic acid)-co-(arylamide) gels
and Polylvinyl alcohol )-poly(acryli- acid) gels under an
electric field was investigated. Bending ot these io)nic, gels
was induced by an electic field. Using this deformation, we
constructed a prototype oif a robot hand having soft fingers.
and an attificial
fish able to swim.
INTRODUCT ION
A polymer gel is a crosslinked polymer network swollen in a
liquid medium. Polymer gels. "solid-liquid coexistant materials",
are candidate "biomimetic materials". Recently. th, ir
mechanical strength becomes very close to living muscle. fi a
soft structure of gels, a motion of polymer network and a
diffusion of ions take place easily by an external stimulus.
Therefore, polymer- gels have various p ossibi lilies as advanced
tunctional polymers.
A typical function of a gel containing ionic groups, is to
bend reversibly under the influence of an elec-tric- field'.
making it useful in some actuators driven by all elec,(trit- fIeld.
In this paper, deformation of the gels undor- ant ele(t ro iel
is presented.
DEFORMAT ION OF POLYMER GELS UNDER AN ELF:CTR 1C FI El 1)
Acrylic acid-acrylamide copolymer- gels (l'AAm gels) swollen
in aqueous electrolyte solution show three types oft defo(rmation
under the influence of a d.c. electric field ; hrinking.
)
swel Iing and bend ing, as shown i n Fig.lI.'IThetype ofi,
deformation depends on the fract ion of sro iurn acry late IAAN~t
in the PAAm gel, the shape of the PAArn gel and the ;Nisiti-nT o'F
A
the gel between the positive and the negative viectruie(iv.
PAAm gel
-
with a
low fraction of
AANa
facing to the positive electrodie (Hig.
shrinks on the siittfact,
(a))
and a gel wi ti.a
it
side (F-ig.
lcv
high fract ion of AANa swells, on the tosi live
1(b)). When a rectangular gel is placed par-al lol to the
elect rodes,. be-ndIng of the gel takes place Ii warut the vx s iitivi
electrode (Fils. I I) I or toward the negat ive elct olie (fig.l~d)).
ion uof
the cri tic-al frail
according to the fraction of AANa.
AANa is abo.ut 25mul'Z. This bending behavior i, similar to the
buckling of a himetallIic- strip suhmit tiil to at variation of
temperature. Because we obseorved the be~ndiung ulefutmal ion with
gels having different ionic groups, such as SONa and NR,(l.
the phenomenon is considered to he a general jpr qt v uift ini
Mat. Res. Soc. Symp. Proc. Vol. 171., 1990 Materials Research Society
390
'U
.
00
(a)
Fig.1
(b)
•
/
..- ,
0
0
(c)
(d)
Deformations of PAAm gels under I.c. electric field.
(a) and (c):for small fraction of AAna
(b) and (d):for large fraction of AAna
gels. Of course the direction of the bending is inversed, when
the sign of the ionic groups is changed.
BENDING OF PAAm GELS WITH HIGH FRACTION OF AANa
In Fig.2, the weight gain of a PAAm gel (rectangular bar
8x8x80mm) is plotted as a function of the strain in bending.
When the PAAm gel bends semicircular, the strain in bending is
UI
02
200
,,'
20
-0
0
0
a. S04
4-'II
0
,0
0
10
0
.1
0N
0'
0
0
0 0
AU
"A0
0.1
0
,,-,
02
03
,'
70
t'.
.2
Weight gain of PAAm
geIs plotted as a
iunction ()f strain
i n bend i fig.
0
0.
I
15
2
2.5
Salt Concentration (10-1 mol/1)
Fig.3 BendinR speed plotted
as a function of salt
concent ration.
391
soo
\
*
eo
0000*, S
...
0.1
00000000000000
o
4
000000
00
0
(9
000000000000000000
Cracks
4.)
4d)
005
In
Time (min.)
05
1.5
1.-
0
10
20
30
40
Cycle Number
Fig.4 Time dependence of
the strain when the
polarity of the
applying voltage
is changed suddenly.
Fig.5 Fatigue-rupture.
In Figs.4 and 5, symbols
show (O):NaOH, (O):NaZSO,,
(0):NaCl and (A):Na2CO,.
0.225. At that time, a weight gain of 15% was observed, as
shown in Fig.2.
The bending speed of the PAAm gel depends on three factors.
The bending speed is proportional to the field strength and the
concentration of -COO- in the gel. The speed is plotted as a
function of salt concentration outside the gel in Fig.3, where
the strain after 30 seconds from the beginning of d.c. supply
of 50V is taken as the vertical axis. There is a maximum in the
bending speed, and the salt concentration giving the maximum
depends on the valence of the ions.
By the change of polarity (+50V , -5OV), the strain can be
recovered as shown in Fig.4. This suggests that it can be bent
repeatedly by an a.c.electric field.
The number of cycles to "fatigue-rupture" depends on the
kind of electrolytes in the solution. In basic electrolyte
solutions, such as NaOH and NaCO, , the bending speed of PAAm
gel is constant up to 50 cycles. In neutral salt solutions,
such as NaCl and NaSO, , the speed slows gradually, as shown in
Fig.5. Some cracks occur on the surface of the gel after 14 or
15 cycles.
BENDING MECHANISM
The volume of a gel is controlled by osmotic pressure". The
osmotic pressure V is given as the sum of V, V,.and V, which
correspond to the osmotic pressure due to the solubility of the
solvent in the polymer chain, rubber elasticity and ion
roncentration difference between the inside and the outside of
the gel, respectively".
50
392
,r=-(Ln(l-v)+v+Xv2 )RT/V,
i -v/2)
+ (v'
RTL-. /V.
+ (XCi
-2C,
) RT
where v is the volume fraction of the polymer network, x is the
solubility parameter. V. is the volume of the polymer network
under the dry condition, P. is the number of chains, V. is the
molar volume of the solvent, C; and Cr are the ion
concentrations inside and outside of the gel. respectively, R
is the gas constant and T is the temperature.
At equilibrium, the osmotic pressure ' of the gel is equal
to that of the surrounding aqueous solution, 7to. Then, fi,
K 2
and it; have the definite values, respectively. When a d.c.
electric field is applied on the gel in the aqueous solution,
the counter ion and the free ion can drift to the positive or
the negative electrode, whereas the polyion can not move. Then,
7r, varies and the value of 'F deviates from 'Fe. The swelling or
the shrinking of the gel occurs until the gel reaches its new
equilibrium state. Since the ions drift at the different speeds,
depending on their size and valencey, the osmotic pressure of
the positive side is unequal to that of the negative side, and
bending of the gel occurs.
PVA-PAA GEL
Although PAAm gels bend like a finger under an electric
field, because they contain 99% water, their mechanical
strength is so weak that they break easily. It is necessary to
improve the mechanical properties of the polymer gels, to apply
them as actuators.
A PVA-PAA (poly(vinyl alcohol)-poly(acric acid)) gel is
prepared as follows ; Poly(vinyl alcohol) saponified over 95%
and Poly(acric acid) were desolved in a mixed solvent of water
and DMSO. The pregel solution was frozen and thawed repeatedly.
The resultant gel, which turns white at this point. was
PA A
crystalline
Fig.6 Conceptional
picture of
45.*
.
s.64
20
Fig.7 Diffracted X-ray
from the dry
PVA-PAA gel.
393
Q3
2 (2
PVA-PAA gel
0.1
PMA gel1
0
1
E
Fig.8 Tensile stressstrain curves.
Fig.9 Bending of the
PVA-PAA gel.
immersed in a NaOH aqueous solution for a long time in order to
dissociate the carboxyl groups.
The PVA-PAA gel obtained contains 90% water. Polymers in
this PVA-PAA gel may form an interpenetrated polymer network,
as shown in Fig.6. The X-ray diffraction profile from a dry gel
is shown in Fig.7.
Fig.8 shows the stress - strain curve of the PVA-PAA gel in
a tensile test. The tensile strength and the elongation to the
rupture of the PVA-PAA gel are more than 50 times as large as
those of the PAAm gel, and are close to those of rubber
vulcanizates. The PVA-PAA gel appeared to possess high shear
strength and toughness.
The PVA-PAA gel bends toward the negative side under a d.c.
electric field, as shown in Fig.9. The bending speed of the
PVA-PAA gel is almost equal to that of the PAAm gel.
POLYMER GEL ACTUATOR
Fig.10 shows a robot hand with four gel fingers composed of
a PVA-PAA gel. The each gel is rectangular bar whose sizes are
60mm in length and 6x7mm in cross section. To bend the gel, a
pair of electrodes is located near the gel. The positive
electrode is fixed with a spacer apart from the gel. The
negative electrode is embeded on the surface of the other side
of the gel. The robot hand can pick up a fragile egg in safety
in a Na2 CO 3 solution by applying an electric field of 50V.
Fig.ll shows an artificial fish composed of a plastic plate
as a float and a PVA-PAA gel as a tail (30mm in length, 10mm in
width and Imm in thickness). In Na 2CO, solution with electrodes
set on each side, the fish swims by waving the tail as the
polarity of the electric field (50V) is changed alternately.
The swimming speed is about 2cm/s.
We demonstrated that a body can be transfered by mechanical
deformation of polymeric material. This exhibits a new type of
functionality of a polymeric material.
394
Fig.1l
Artificial fish.
Fig.10 Robot hand.
CONCLUSION
The mechanical response of the PAAm and the PVA-PAA gels
upon the electric field was examined experimentally. The gels
can be bent by appling a d.c. electric field across the gels.
The deflection of the gel is considered to occur due to the
drift of ions under the influence of the applied electric field.
Using this deformation, we constructed a prototype of a robot
hand having soft fingers, and an artificial fish having a soft
tail, both driven by the electric field. The next subjects are
to realize the more rapid response of the gel and to increase
the modulus in order to increase a work to the external.
References
1. T. Shiga and T. Kurauchi, Japanese Polymer Preprints. 36,
2894 (1987).
2. T. Tanaka, Science 218, 467 (1982).
3. P.J. Flory, Principles of Polymer Chemistry, (Cornell Univ.
Press. Ithaca, 1953).
4. 1. Ohmine and T. Tanaka, J. Chem. Phys., 77. 5725 (1982).
395
BIAXIAL EXTRUSION OF POLYIMIDE LARC-TPI AND LARC-TPI BLENDS
R. ROSS HAGHIGHAT*, LUCY ELANDJIAN, AND RICHARD N. LUSIGNEA
*Foster Miller, Inc., 350 Second Avenue, Waltham, MA 02254
ABSTRACT
Blaxial films of polyimide LARC-TPI and LARC-TPI/liquid crystal polymer
Xydare were extruded directly from the melt for the first time via an
innovative new extrusion technique. Three types of films, neat LARC-TPI,
LARC-TPI/1O wt percent and 30 wt percent blends were processed as a part of
this NASA funded program. This new process offers an alternative technique
to costly post-processing stretching of both solution cast and sheet
extruded films. The post-processing step is often required to enhance
certain properties. Processability was greatly enhanced by incorporating
Xydar. The coefficient of thermal expansion was reduced from 34 ppm/°C for
the neat LARC-TPI to 15 ppm/°C for the 10 wt percent Xydar blend and
ultimately down to 1 to 3 ppm/°C for the 30 wt percent blend films in the
direction of extrusion. The maximum improvement in stiffness was realized
by incorporating 10 wt percent Xydar (2.8 GPa up to 4.9 GPa). Tensile
strength, however, experienced a drop as a result of Xydar addition,
probably caused by inefficient mixing of the two phases.
INTRODUCTION
Today's growing demand for high temperature, high performance, and low
coefficient of thermal expansion (CTE) polymers requires the development of
new polymeric systems and novel processing techniques [1]. One such
polymeric system drawing much recent attention Is the polyimide (2].
In our study, polyimide LARC-TPI (Figure 1) developed by NASA was
processed both in the neat form and with additives Into blaxial films by
melt extrusion. To our knowledge, this marks the first time a polyimide
has been melt extruded directly from the fully imidized powder into blaxial
films. Our film processing made use of an innovative extrusion technique
which imparts biaxial orientation in the films during processing, rendering
post-processing orientation unneccessary. Post-processing stretching is an
often required step for property enhancement. Through this innovative
processing, films with a variety of orientations from near uniaxial to
balanced biaxial can be extruded in a one-step process without the
shortfalls of other competing film processing techniques.
Fiber reinforced polymer composites have been shown to considerably
improve the engineering performance of various polymers and have found
0
0
C-TPI
LAR
Thermoplostlc Polylmlde
Figure 1. Chemical Composition of LARC-TPI
Mat. Nos. Soc. Symp. Proc. Vol. 171.
1990 Malterill Research Society
396
numerous applications in aircraft, automotive, and marine industries. The
traditional inorganic filaments such as glass and graphite are consistently
challenged by newly developed high modulus and high strength polymeric
systems. Examples of such systems include polyaramide (Kevlars), and more
recently thermotropic liquid crystalline polymers (LCP).
LCPs can be
processed in the melt state and are capable of forming highly oriented
crystalline structures when subjected to shear above their nelting point
[3]. Their rod-like molecular conformation and stiff backbone chains allow
the LCPs to form fibrous chains. Examples of LCPs include Xydar and
Vectra® manufactured by Ammoco and Hoechst Celanese, respectively.
Experimental
LARC-TPI is produced under a licensing agreement with NASA by Rogers
Corporation of USA, and Mitsui Toatsu Chemical of Japan.
Both
manufacturers produce polyamic acid solutions for film casting and fully
imidized powders for injection molding. Mitsui Toatsu also manufactures
LARC-TPI 1500 extrusion grade polyimide which exhibits a modestly lower
melt viscosity than the injection molding grade LARC-TPI 1000 and is
thermally more stable. Mitsui LARC-TPI 1500 was the only grade of LARC-TPI
successfully extruded into films both in the neat and the blend forms.
Thermotropic liquid crystal polymers (LCP), Xydar, and Vectra were
considered as additives to LARC-TPI. Their presence is thought to (a)
enhance processability of LARC-TPI by reducing the melt viscosity; (b)
create a fibrous network which in itself gives rise to a self-reinforcing
mechanism in the composite similar to fiber reinforced composites; (c)
contribute to molecular order within the composite causing a rise in the
stiffness and resulting in a decrease In the coefficient of thermal
expansion (CTE) due to restricted movement of the molecular chains.
The LCP selection process included considering compatibility in the
following areas: temperature, rheology, and particle size. Xydar was the
most suitable candidate LCP available for use in this study. In all, the
following types of films were successfully biaxially extruded:
" Neat LARC-TPI
" LARC-TPI/I0 Xydar
* LARC-TPI/30 Xydar
The individual powders were seperately dried in an N2 -purged oven at
120 0 C for 12 hours followed by dry mixing of the two components and
redrying under the same conditions prior to extrusion.
Extrusion was carried out with a specially designed laboratory scale
blown-film apparatus. Using this extruder, biaxially oriented polymeric
films were formed by adjusting a series of easily controllable variables
such as shear rate, feed rate and take-up speed. Orientation here refers
to controlling the direction of the polymer chain molecules to tailor
properties in the plane of the film. Visual inspection of the transparent
films revealed preferred biaxial orientation. Angular orientation was
verified by both optical microscopy and by manually measuring the
molecular chain orientation with respect to the machine direction.
Results
Table I summarizes the studied properties of the extruded films. Comparing
the tensile strength values of the near uniaxial and the ±24 deg films, the
397
effect of orientation becomes clear. The tensile strength of the near
uniaxial film of neat LARC-TPI was 28 percent higher than the ±24 deg film
This property improvement was more
(126 MPa compared to 98 MPa).
pronouncea with increasing Xydar content (45 percent increase at 10 wt
percent Xydar and 60 percent at 30 wt fraction). The stiffness also
followed a similar trend. Figure 2 plots the stiffness versus wt fraction
LCP. A higher than expected increase in stiffness was seen at
10 wt percent Xydar. Although we are currently studying the mechanism
giving rise to this phenomenon, it is possible that 10 wt fraction is the
loading level at which the Xydar fibril network Is most efficiently
oriented under shear.
Table 1. Properties of Extruded LARC-TPI and LARC-TPI/Xydar Blends
Tensile Strength
MPa
MD*
TD**
Film
Stiffness
GPa
TD
MD
CTE
ppm/°C
MD
Neat LARC-TPI
t24 deg orient.
98
104
2.8
2.8
34
Neat LARC-TPI
near uniaxial
126
-
2.6
-
28
LARC-TPI/10. Xydar
t24 deg orient.
80
52
2.1
2.3
16-18
LARC-TPI/10. Xydar
near uniaxial
113
83
4.8
2.7
14
LARC-TPI/30". Xydar
±24 deg orient.
63
-
2.0
-
5-7
LARC-TPI/30% Xydar
near uniaxial
108
-
3.3
-
1-3
*Machine direction
**Transverse direction
STIFFNMESSRMs
0.99
RULE-O-MXUE
R?
4.5
6.5
0.616
AULE.OF.UXTURES
3.5
0.33
0
25
so
75
10
W.FRACTION LCP
Figure 2.
Stiffness versus Nt Percent LCP
Figure 3 Is a plot of the CTE versus wt fraction LCP for the near
uniaxial extruded films.
Addition of even small amounts of Xydar
dramatically decreased the CTE. At 10 wt percent, the CTE was reduced by
greater than 55 percent. The results suggest that the Xydar addition
alters the mechanism and probably substantially hinders molecular chain
movement within the extruded films. Further work is underway in this area
which will address these initial findings.
Figures 4 and 5 are photomicrographs of tensile failed films of the
neat and the 30 percent Xydar, respectively. The morphologies, as
suggested by the micrographs, are noticeably different. The failure
mechanism, as suggested by the micrographs, was matrix failure followed by
fiber failure. There is evidence of fiber pull-out to support this theory.
CONCLUSIONS
Through a novel process, we successfully extruded biaxial films of
LARC-TPI polyimide and blends of LARC-TPI with the LCP Xydar. To our
knowledge this Is the first time a thermoplastic polyimide has been
extruded directly from the melt into biaxial films. This simple one-step
process presents an alternate means of producing high performance polyimide
films In high volumes and cost-effectively.
It eliminates the often
neccessary post-processing film stretching to enhance certain properties.
It also eliminates the often toxic devolatilization of solvents associated
with casting operations.
He have shown that through LCP blending the tensile modulus can be
increased. LCP incorporation also lowered the expansion coefficient (CTE)
from 34 ppm/°C in the neat film down to 1 to 3 ppm/°C at 30 wt percent
Xydar loading.
40
30
RULE-OF-MIXTURES
CTE
ppmC
2
O
10
25
50
75
100
WT FRACTION LCP
•10 -
Figure 3. Coefficient of Thermal Expansion versus Nt Fraction LCP
399
Work is currently underway on parallel programs to further study
polyimides. We seek to better understand their processing parameters and
means of enhancing their properties.
4
Figure 4. SEW Micrograph of Biaxially Extruded Neat LARC-TPI
Figure 5. SEW Wicrograph of Biaxially Extruded
LARC-TPI/30 percent Xydar Blend Film
400
REFERENCES
I. Walker, C.C., Proceedings
Polymides, pp. 429 (1985).
Second
International
Conference
on
2. Sherman, D.C., et al., "Neat Resin and Composite Properties of
Durimid High Temperature Thermoplastic Polyimides," Proceedings of
the 33rd International SAMPE Symposium, March 1988.
3. "How Nell Do Various Blends of LCP and Nylon 12 Work?", Plastics
Engineering, pp. 39-41, October 1987.
ACKNOLNEDGEMENT
This work was supported by NASA-Langley Research Center, Contract
No. NASI-18527, and monitored at NASA by Dr. Terry St. Clair. The authors
would also like to thank Mr. Roland Wallis for his technical contributions.
401
J
STRUCTURAL STUDIES OF SEMIFLEXII3LE FLUOROCARBON CHAINS
CONTAINING AN AROMATICCORE
A. SCHULTE. V. M. HALLMARK. R. TWVIEG, K. SONG AND .I. F. RABOL'I
113M Research Di'isiont. Almnadent Research Center. San -Jose, C'A 95120 6099
ABSTRACT
The niolet'nlar struri'tirc of a series of perfliioroalkatie tdigoiirs havinig two
chains at tachedi to a phen-ivi groiip ill the paira, posit ion has' hea i stutdiet
%v
Fou rier Transform aid r'tilventit ini Ranlianl sped ro so pv at a oilhtout alit loiw tet'ilic'
at ures. Fronm comiparison withii motlel t'oupo iis bands at trihtiable to Imb thi t iiisill '
F( CF 2 )
a
a~~t
itteti phenivI ring aind the( pt'r-fiuttoaikane ch ainis coutld bei assigned. In particida r.
the lo w frequency regin waii nvest iga teduas a ftuct i (of tenil it'ci it altoI pisutreiii
oirder toi assess the( imipact oif tL' rigid ariatatic core ont tite laniusliapitsatut fr('t'itii''.
I. INTRODUCTION
atheir
I
The uise- of flutorocarbion oligouii'rs ais ilioihi systemts to st oiy setiilfi'll potlyliers is well estaldished J1.2,31 and hias piro videdt a sigitivatit intsighit ito thi ititlt' of
backbont' tconfotrmtation
oil t'iii stifftes. Ini addli itin toi irfltitoin ta rlo
iiii
ligimri ati
sets (t thillock [4.31 atid t rilock (61 st'iiifiitritiatt't alkauivi's
liviii Ynititc~,ir aridl
crystal and conforinat jonarietnt' t studied extensively Of part icuilar iliterist
ini thle t ribloi'k series was thle effect (if a sentifle~xile hivi riucari a
cinteir block i thei
ftt'qtv- anti intensity of thtf' ltuw~ frt'tItit'v u Raatin
i t ivi' Iitigitithlinal ;ti'ti
v
od'.it
(LAM ). Restilt a indtica te that at rootmn ttilprti
ti the ren't'etir llt ik itirtii'iptt ' withI
hie tittoicarlo
tin'td
idtttks. in thle atctcirionii like' mot in riiihtarai tetrist ii ofi L AM. Alitv
tie mteltitng point, til thle other hand, t hit ilisi nlirtiI iIn ircaittiiii sigu tti
v'
as Ai
~~~~weak
cotupling spring pert urnKtiglii LA NI freipittic
if t'nt'eni
biiilltcuk.
It is the puirptost of tit', present stit ti to rej unt' t ill, .'iltifi'Xiilt'hytn
it~doint ii ii
block withi a rigid aromia tit' Core anti t hen assess its effe'it tin both t ht int eniity andii
freqttency' tf LAMI.
2. EXPERIMENTAL MIETHODS
The, p-ip''~uitak'ieiiiswere iri'pari'i liy coppeir liai'iitt'i itiuttittiat itti
of the apirop~riate' perlutiralky' ioide with p)tiiiitlii iitatiu ini ilinituiyilftxiii. T6hu
acetic acitd.
0,Diffe're'nt
4
4tof
ial scaitminig atti tut ry ( DSC iti'wtitntntsa
Ii icfi'o i:iii'i c i a Dii
Pont 911) DSC withI a 11090 ri ttri tlt r A stauti i g ratru'if 10 O(C / 01ii i wits t is 'ii
Scanintg Rtattani ti'a',triu tit wetri ri''rilu'i iv iiig
a .liiiiu \'vi ii HG 2,;
dotlei mtiiuthtrtuuatm ntiigttre't for iphotoin rititjg
miil ititifaiiiii ti ia Nititlvt 11811
data systemi. X'aiiahil' temitarattir' stuitt' wire 'rriitl wiit ini a vitital flarua'' Milli
cull. Tite siampie trrtiuratlir' wtas itutiti itd C'ontant to withli 1 O('C
Ramtan e-xlterimttii til samples ittiir- iire'sstirt' vent' 1 ii'i-foirdi'i tin itiiikstat
tiring geittietry uising it Wiaittilliy diamuond anvil i-il withIi hut i guiskit . 'fit' 11~t,"1114
wam do',rnuinu'i ini sitvi frtint jt'e peak potsitioin tif itttyutiti i's-i'iii'
litw withIin ito
± 1 kflar.
MetI.Resn.Soc. Symp. Proc. Vol. 171.
'
1"0O Materials Research Society
.
402
Forier transjformt
Bluimai spaectra were collected witll) a 131anenil DA3.02 ii ite
feroiiter lisilig iitlieriioceetrically
iaSpectro
ModiNbael
SLcA0 ew \'N
coo~led IiiG~iAs detector [7]. The cxcitaition oic
AG laser.
3. RESULTS AND DISCUSSION
Thermal analysis.
show strong
DSC cu rves of the serilifli o rilnit ed plieii Irill cks ( FrPliFi
ieltinig eiildothierrils between 601 mnd 180 ''
as, shownii i Fig. 1.
0
0 20
40
60
80
100
120
140
160
180
TemperatUre ('C)
Fig. I DSC thiermiogr'li
/iiloin.
of sonie seiiitiiorimitei
it ;alanes SCedi
;it ;I i t
of 10'
"Ieiiipeiatuires refer to die peak po(sitiolls.
The melt ing temliparat ire ilierekises withI the lengtli of the finorarlioti cliitiiis
il
siiiihir to thle parfhmalo alkamies. However, the comoin i F7PhiF7 exhibit,
lin aliolmloiisly higher imltinig tellihalatlirvid
n its DSC curve is typifiedl by at Neaker
einhothlari momiid 17 'C.characteristir of ;I solidh-solidh idiase trais'itioli,
The liit s of fiisioii iir('ise, coiitiiiosly with thle lenigthI of th fli miro;iii
dliml. fiom 31 -,(
FGPIiFGi to 60(1
ii
~IF12PhiF12.
Ranman
hda'nil*1
measurement s.
t ribleaks
;m, ciutiiuin
Ili Figmre 2, the Piilliili sped tn of three sexii1ifi orinlittId
withI those of per'flitoro- ii-ilkmres having the smile fhtaili
403
~FSPhF8
,
!
SF7PhF7
m
C43
F6PhF6
C,2 F2
100
6
IL
Ij
1500
1000
500
Frequency (cm
Fig. 2 Unpolarized Raman spec)
c'm1700
(50
tra
1111
of P-)isp)erflinor)alkyll)('lzezfes and
Te asterisk
perfluoro-n-alkmes.
delzlott's a spurious vimissionl lin('
F12 Ph F 12
due to the room light.
FOPh FlO
*,>,
F8 Ph F8
spec
3
Raman
Fig.
tra of P-Iislarlmmroalkvll)(' es
.
-.
...
"
I
recorded with Fourier transform
F 7 Ph F
spectrometer. A resolition of 4
was ised with a laser pwer
rm
of 700 m\W at 1.064 pin..-
'" "
.
I
-..
F6 Ph Fi
1500
1000
500
Fr-
wtn ,
,
404
Most clearly, a new hand due to the C = C st ret ch of the phieiyl ring occurs
aroundl 1625 cm-'
In addition, there are bands due to the ring at 1095. 695 and 635
rill' ,which are not presenit in the perfinoroalkanies aild do not shift with the chain
length. This conclusion is emphasized by the spectra in Fig. 3 measured with Fourier
transform Ramnan spectroscopy. This technique uses excitation in the near infrared and
allows the Ranman spectrumn to be obtained in the absence of fluorescence. lIn tis way
lie Ranian spectra of all the FnPhFnl compounids synt hesized. which displayed varying
ainloinits of fflorescence couild be measured. As ab~ove, bands attributable to the phenyl
ring call be recognized. The sharp bands in the 700 -750 cin- region arid the smaller
(haes between 300 and 330 cm- are assigned to various -CF, mnodes. In the region fromn
800 to 1100 cm-' a nanier of bands whose frequencies change with chain length can be
semii. For inistanice. the hand at 915 curl- shifts to 788 co'when going from F6PhF6
to F1OPhilO. These hands can he used to plot out portion,; of dispersion curves for the5
polymer [11.
The low frequency spectra (below 200 cm-' ) are in the case of the perfinoroalkalies dominated byvanl intense sharp band with a peak frequiency varying inversely with
Chain lengthI. It has been assigned to the llamai-act ive longituditnal acoustic mode
(LAMI-i). Onl the( Other hand. tie low frequnency spectra of the FnPhFt oligo
show
ail inteinse buat rat her biroad hansd. wit hoot a syst einat ic correlation with ile cli
--cngthI.
This 1 id~ m
iay lbe at tribut able to a lattice modi~e involving librat ions of the phenyl ring.
Tluis the insert ion oif thle ring iinto the fluorocarbon chaiin seems to have pert urbedh
lie vib ra t iinal lioia hi node struicturrc iangh to reiiov(' (or dliminishi thle LAMNmode
ilt ensity.
Previous studmoies (61 oil trilhck semifhuioriuated ni-atkanes have shoiwn that no
deci iip ing if thle chi ln vib ra tion occurs ktt lie, junctulre between heclical and planar
zig/iig iiiforuiiatious. Ii aiddition. there waks re asoiiailih' agreenrenit between thieciservedl
LA NI valueiaindI thIose cal cu la ted with ilie skeletal ando poin~t ilass approximiatins [U],
Preil il Ia vy
riiial mlode calciilatioiis were pecrfo rnied kiiing t hese lines to investigate
the effect of a rigid core onl thle LA NI ill pihieiiy t ril locks withi seinifluorina ted ri- alkane
chains,. Tia' results iindicate that the geoiietrical chainge has a significant effect oil the(
LA NI frequeincy iad iiteiisity. thlou gh it iiay tiot decouple LA NI vibrat ionis over thle
enitire, cloiiii.
Temperature anid Pressure Studies.
Decreasing the temipleratuire prdue oidy
ixpecteud changes, ill vibrit ionmal Iinewidthis iii the frequency region iibove 15(0 rilliFig. -1). By far thle biagest change iccu rs iii thie very iintenise low frequency hand for
the iiiiipinmd F7PhiF7. Whein loweriiig tlla temperature froim 22 'C' to -71 'C the peaik
pi st iiil is shliftl li
by an it 30 (-ill-' to higher wa veiimiibers. Also the widthI of thle
'id brcollic' iiarruiwer. which is conisistenit withI a restriction of the ring motion.
F i i thea idonl i~ liii iinlof thle series exhiibit ing a iat her sub stamitial change
li1114ia
16j1ii'tii J0111;is will ;is a thermial t ransition blohw rooiim tempiheratuire. There(01n4it al
lacsi Ibc ini ;I lilbretlit liist, ;It rooiii tceiiiaatre than its eveui chain lenigth
;lauli,g,. pcrliaji, dii to, ;I diff'eent rrystal si it ire or toi ifferenit conformiation of the(
tli i- hicuilui
m )l
ilii.
om.lvhi
woliiluI Ivjplt
that siiiiilmr ihiiges as iiliihi cooilinig occur wheii apipi1 ilm, 1,ii ,mci. -incl, in boith cases the( ;ickimig increases. The Ramian sliectrmnli of
I.71%1 71 iiiuid-1I;it various pi-essirec (Fig. 5) shows two nmajor effects oil the inteinse
I'm fli uiiu
iaiill At ;I 1uir-esi near I kBai tlie, hial tiarniows mid~ iii(A
to liowei
jgans ;Iii~
l ioii'of ihe ring lilmitiou . AlaiVV 9 kBar a new banat i
;1,, ;%1us
\N1-1 lims li4,li -assignedi to the planari Agzag forii of fluoroiarlioii
11,11 116i- 111416,;1h., fit
tIs huinuiia i port ion if li, illdeiilics have iulergomaf
405
C
-71'1,
'AF8PhtF8
E
F7PhF7
t7~
FOP?,F6
71C
*T
500000)
Soo
150
Fequency
c-i'
Fig. 4 Ramuan spertira (30 - 1700) cr
o f Fiu~lun at ibicut aiia) low
The asterisk dteoiovs a siioix ciionau linei (11i iv ti t roo iii ght.
tcmijaii;tli'.
C
0,
10
0
1000BA
Frequency (cm I)
Fig. 3 Ramuan spiiitra (20
1300 cmiiof Fuiflhi a.,i a fii'rioti (if pio'nsuri.
tippe r spect rti was t aken Ifttr 1,01varsilig th lensdil.'
t,
406
a conformiational change. Correspondingly the low frequency band broadens again and
shifts tip in frequency, ats the lattice mode changes in response to thle new molecular
packing of the fluorocarbon chains.
4. CONCLUSIONS
The molecular structure of a series of semifluorinated alkane chains containing
a rigid core have been characterized by Ramnan spectroscopy in the visibile and in the
near infrared. Vibrational bands due to the phenyl ring and the alkane chains have
been identified. The low frequency spectrum is dominated by a broad b~and. which
we attribute to a lattice mode involving ring libration. This band undergoes large
frequency and bandwidth changes. as a fuiiction of temperature, particularly for Ihe case
of F7PhF7 which exhibits a solid-solid phase transition. Although preliminary normal
mode calculations indicate thlat insertion of a ring into the fluorocarbon chain does not
necessarily decouple LAM vibrations over thle entire chain, significant perturbations in
thle LAM position and intensity are possible. This mtay account for the fact that no
hand attributable to LAM is obse.rved for these compounds.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
J. F. Ilabolt andl B. Fanconi. Macromolecules 11, 740 (1978).
G. Masetti. F. Cahassi. G. Morelli, G. Zerbi, Macromolecules 6. 700 (1973).
A. 3. Hannon. F. J. Boerio. J. L. Koenig. J. Chieni. Phys. 50. 2829 (1969).
II. Twieg and J1.F. Raholt, J1.Polyn. Sci., Polvi. Phi~~. Ed. 21. 901 (1983).
G. Miniji atid G. Zerbi. J1.Polymn. Sci.. Polymn. Lett. Ed. 22, 533 (1984).
R. Twicg and J1.F. Rabolt. Macromolecules 21, 1806 (1988).
C'. G. Ziniha, V. MX.Hallmark, J1.D. Swalvii, .J. F. Rtabolt. AppI. Speatr. 41. 721
(1987).
T. ShliatiOUtCld M
N. Tasuiti. Indian 3. Pure Appl. Pltys. 9. 958 (1971).
C. K. Wit andi~NI. Nivol. Cheiin. Phys. Lett., 21. 153 (1973).
407
THE EFFECT OF LOW POWER AMMONIA
ON CARBON FIBRE SURFACES
AND NITROGEN PLASMAS
**C.JONES AND *E.SAMMANN
National Centre for Composite Materials Research, University of Illinois, Urbana,
Ill.61801;*Materials Research Laboratory, University of lllinois.(**Now at Liverpool
University, England)
ABSTRACT
The effect of low power nitrogen and ammonia plasmas on carbon fibre surfaces
has been studied using X-ray photoelectron spectroscopy(XPS) and scanning electron
microscopy (SEM). A comparison is made between two polyacrylonitrile based fibres and
a pitch based fibre. Grazing angle techniques have been exploited to probe only the first
12-15A of the fibre surface. Plasma treatments were carried out in an insitu plasma
treatment cell which was attached to a PHI 5400 X-ray photoelectron spectrometer
enabling the immediate effects of the plasma to be studied before the treated surface was
exposed to air.
INTRODUCTION
The properties of composite materials are not only governed by the properties of
the individual components but also by the interface between them. Successful
reinforcement of composite materials is only achieved by obtaining sufficient stress
transfer between fibre and matrix. This can be realised by physical and/or chemical
adhesion between the two. However, for recent applications an extremely strong interface
is not always desirable. It would be ideal to be able to design the interfacial properties to
suit a particular application for example increase stiffness or promote toughness at the
interface. One method of achieving this would be to chemically graft a monomer with a
suitable backbone onto the fibre surface. It would also be necessary for this monomer to
be fully compatible with the resin. An ideal example would be another epoxy with the
desired backbone. Some of the most common curing agents for epoxy resins are amines.
These allow curing of the resin at room temperature. It would therefore seem beneficial to
introduce amines onto the fibre surface which would certainly have the potential to react
with the epoxy coating at room temperature. The effect of ammonia plasmas on fibre
reinforcement has been examined by several research groups[l-5], however, a thorough
understanding of the chemical changes induced on fibre surfaces and whether or not they
play a a role in fibre resin adhesion is yet to be determined.
A method of selectively introducing amines onto the surface would be a great
advantage. Introducing the desired fibre surface chemistry without destroying the
mechanical properties of the fibre itself would also prove very useful, Most of the plasma
treatments used to date however, remove a substantial amount of material causing pits to
form in the fibre surface [e.g.2,3]. This has led the authors to develop a low power plasma
(<IW) treatment which only alters the chemistry of the immediate surface layers without
causing severe etching of the fibre[6]. In this paper we report the selective introduction of
C/N functionality onto three types of carbon fibre.
EXPERIMENTAL
The fibres used in this study included untreated and unsized T300 fibres (supplied
by Amoco Performance Products) and HMU fibres (supplied by Hercules). Sized but
untreated pitch based fibres from Amoco, thoroughly washed to remove any sizing, were
also examined.
Plasma treatments were carried out in an in situ plasma chamber attached to a
PHI 5400 X-ray photoelectron spectrometer. A half wavelength helical resonator was
formed from a 100- turn coil wound directly on the outside of the tube and centred within a
Mat. RN.. Soc. Symp. Proc. Vol. 171.
1990 Materials Research Society
408
shield made from 3" diameter brass tubing. A simple self-excited two transistor oscillator
delivered radio energy to the centre of the helix at resonant frequency (approx. 15MHz).
Low power levels (<IW) were sufficient to sustain a plasma within the tube. Continuous
gas flow was maintained from a leak valve at one end of the tube to a butterfly valve at
the other, which opened to a turbo molecular pump. The pressure within the cell was kept
constant at 0.1Torr during plasma treatment. A more detailed description of the cell is
given elsewhere[6]. This enabled the immediate effect of the plasma on the fibre surfaces
to be studied before being exposed to air. The fibre samples were electrically isolated
from the supporting rod, using a piece of mica sheet, allowing a controlled bias, with
respect to the grounded supporting rod, to be applied to the fibres independent of the RF
power exciting the plasma. Fibre surfaces were examined using X-ray photoelectron
spectroscopy before and after treatment. The spectra were collected at grazing angles to
enhance the signal from the immediate surface layers of the fibre[71. The binding energies
reported in this paper have been reference to the main graphitic peak in the carbon Is
spectrum which isgiven as 284.3eV and all intensity ratios have incorporated the relative
sensitivities calculated for a hemispherical analyser (Cls 0.296, NIs 0.477, and Ols
0.711) [8]. The nitrogen ls spectra were curve fitted using a Gaussian/ Lorentzian peaks
shape in a non linear least squares curve fitting program [8].
Scanning electron micrographs were obtained with a Hitachi S800 microscope.
With its field emission gun magnifications of 100-300 thousand could be achieved.
RESULTS
AND DISCUSSION
Fig. I shows the spectra obtained from fibres exposed to an ammonia plasma for 60
seconds. This is the first time that C/N functional groups have been observed with out the
presences of oxygen containing groups. Widescan spectra (Fig.la) taken at both bulk and
surface sensitive angle indicate that the nitrogen containing functionality introduced is
predominantly within the first 12-15A of the fibre surface. The N:C ratio did not change
with exposure time to the plasma and the difference between that obtained from surface
and bulk sensitive data remained. This indicates that the reaction remains on the surface
of the fibre and does not proceed deeper within the graphite layers as electrochemical
reactions [71. Significant etching of the fibre surface was not observed in the SEM
micrographs suggesting that saturation of surface sites has taken place.
The nitrogen ls spectrum reveals that the plasma introduces three type of nitrogen
containing species onto the fibre surface. Signals at binding energies 398.9eV, 400.4eV
and 402.8eV were observed. The two main signals are very similar to those produced on
electrochemically treating the fibres in ammonium salt electrolyte [9]. However, in this
case peak assignment to amide functionality is invalid since an oxygen Is signal was not
detected. The binding energy of PhNH 2 has previously been reported as 399.1eV [10].
The relaxation energy associated with the ejection of a photoelectron is far greater for a
graphitic like lattice than that obtained from a benzene ring [11,12] and hence a shift of
-0.2eV from that observed with PhNH 2 is not unlikely for amines groups on carbon fibre
surfaces. -C=NH groups could also give rise to a signal around 398.8eV. The signal at
400.4eV arises from aliphatic amine functionality. The peak at highest binding energy is
most probably due to a positively charged ammonium species similar to that observed by
Chang and Navalov (13]. Cyano (-C-N) were not detected. These groups are unlikely to
remain on the fibre surface as they would be prone to electron assisted desorption.
All the chemically shifted shifted signal intensity seen in the surface sensitive
carbon Is spectrum (Fig. lc) is due to C/N functionality. The chemical shift for -C-NH 2 is
expected to be around 1.3-1.5eV from the main peak. -C=NH would lead to a chemically
shifted signal between 2.2 and 2.5eV from the main peak. There is certainly a large
amount of signal intensity in both these region confirming the nitrogen Is data.The signal
intensity around 6eV from the main peak is due in part to a Jr-x* shakeup satellite.
Nitrogen plasma treatments yielded very similar results to those obtained from
4W9
'
a
a)
Cli
NIS
SSC 520
.90 1.60
.3o 4W
370
340
310 280
293
b)
2963 29i.7 293.1 2915 28%9 2883 2867 2851 2835 231920-3
c)
4084 4070 4056 400.2
40
,0.
397.2 39%
Binding EnergyleV)
.00K
0 3"
39.
Figure 1. XPS Spectra From T300 Carbon Fibres Treated With An Ammonia Plasma For 5
Minutes a) Widescan, b) Carbon Is. and c) Nitrogen Is (curve fitted)
fibres exposed to an ammonia plasma. However, the nitrogen plasmas were unable to
remove all the C/O from the fibre surface and the amount of nitrogen containing groups
was slightly less than those obtained from ammonia plasma treatment. The nitrogen Is
spectrum obtained from these fibres consisted of three signal at identical binding energies
as those described above but with a different relative ratio. The relative intensity of the
signal at 398.9eV was less for fibres treated in a nitrogen plasma. The hydrogen
contained within these functional groups is thought to arise from the external surface of
the fibres in the form of C-H.
410
CIS
Nis
NI4
T T300
HM
P100
550
520
.90
.
730 ;10 310 280
) *60370
Binding Erergy(eVI
250
Figure 2. Widescan XPS Spectra of Three Different Types of Carbon Fibres Treated With An
Ammonia Plasma
Fig.2 shows widescan spectra (taken at a take off angle of 150 -surface sensitive angle)
from three different types of fibres after exposure to an ammonia plasma for 5 minutes.
Each fibre has behaved differently to the plasma, the lower modulus PAN based fibre
being the most reactive. The main structural elements of carbon fibre surfaces are
graphitic plates or crystallites which are roughly aligned along the fibre axis. Their size
and degree of alignment increase with increasing graphitization temperature during
processing. The P100 fibres have a very similar structure to that of highly orientated
pyrolytic graphite yet the crystallites on the PAN-based T300 surfaces (the fibres being
graphitized at a much lower temperature) are much smaller and less aligned and hence
the number of edge sites an defects is far greater in the latter case. (The HM fibres have
a surface structure in between the two.)The lack of reactivity shown by the pitch based
fibres compared to the T300 fibres tends to indicate that chemical change has only
occurred on the edge sites (and defects) and not on the basal planes. As the percentage
of edge sites increases, so does the amount of "nitrogen" incorporated onto the fibre
surface during treatment.
This lack of reactivity of the higher modulus fibres is undesirable since the aim of
the treatment was to introduce amine groups in sufficient concentration to promote
chemical bonding between the fibre and resin. The method suggested by Stoller and Allred
III was to pretreat the fibres in an argon plasma for approximately 10 seconds prior to
ammonia plasma exposure. This resulted in a 30% increase in the amount of amine
functionality incorporated onto polyaramid fibres. In our case, the number of nitrogen
containing groups did increase on the higher modulus fibres with Ar pretreatments but
only by a very small amount. This is due to the lack of sputtering taking place in these
low power plasmas. Pretreatments in an air plasma were also performed in a hope that
the oxygen functionality on the surface would be substituted by nitrogen containing
groups. In these cases, most of the C/O groups were removed and only a very small
number of nitrogen containing species incorporated.
A more successful method of increasing the number of am, 'e groups introduced
was to accelerate ions from the plasma to increase their impact energy on the fibre
surfaces. This was done by biasing the fibre samples to a negative potential with respect
to a ground electrode within the plasma. Fig.3 shows the widescan spectra of HMU
411
NIs
5SO
S20
9C
.6C
43C
300
70
340
310
21C,
280
Bndlng Energy(eV
Figure 3. Widescan XPS Spectra Of HMU Fibres Treated In An Ammonia Plasma a)With
And b) Without Bias During Treatment.
fibres that have been exposed to an ammonia plasma with and without sample biasing.
The N:C ratio has increased from 0.07 to 0.18 on applying the sample bias and is
approaching that of lower modulus fibre (0.22). A increase in chemically shifted species is
also observed in the carbon ls spectrum of these treated fibres. The nitrogen Is spectra of
these treated fibres is very similar to the lower modulus fibres in that three species are
detected at 398.9eV, 400.3eV and 402.8eV. Their relative intensity ratios are very similar
to the unbiased lower modulus fibres for both nitrogen and ammonia plasma exposures. A
similar result was obtained for nitrogen plasma treatments. Examination of SEM
micrographs from these treated samples reveals that little damage has been done to the
fibre surface. Although there was an increase in the amount of nitrogen functionality
incorporated onto pitch fibres by biasing the fibres during treatment (N:C 0.01 to 0.05),
the overall number of nitrogen containing species remained very small compared to the
PAN based fibre and will be expected to have little effect on the interfacial shear strength
of the composites.
I
Cls
sNls
a)
b)
550 520 490460
430
40
R
340
310
280 250
Binding Energy IeV)
Figure 4. Widescan XPS Spectra Of T300 Carbon Fibres Treated In An Ammonia Plasma
a)Before, And b) After Exposure To Air.
412
Exposure To Air
Fig.4 shows the widescan spectra of fibres treated with an ammonia plasma before
and after exposure to air. There is a significant pickup of oxygen upon air exposure.
However, no change in the carbon Is or nitrogen Is spectra is detected e.g. evidence of
-COH or -NO signal intensity, suggesting that oxidation has not taken place. It was
concluded that this oxygen pickup was due to a strongly bound physisorbed layer. The
oxygen Is spectrum consisted of a single species with binding energy 532.4eV which is
characteristic of a strongly bound -OH species and in this case most probably arises from
moisture in the air. The amount of 'oxygen' pickup was directly related to the number of
C/N functional groups introduced onto the fibre during surface treatment. This physisorbed
layer would certainly inhibit the desired fibre/epoxy chemical bonding in composites. Both
Loh [4]and Evans [141 tried to compare the effects of ammonia plasma on structurally
different carbon surfaces. Unfortunately, in both cases their samples were exposed to air
before chemical analysis was carried out.
4.CONCLUSIONS
Both ammonia and nitrogen plasma were successful in introducing amine
functionality onto carbon fibre surfaces. The reactivity of a fibre to a plasma was shown to
depend largely on the structure of its surface. Reactivity increased along the series
P100-HM-T300. The number of nitrogen containing groups was dramatically increased by
biasing the HM fibres to a negative voltage (10-30V). This increase occurred for both
ammonia and nitrogen plasmas. Exposure of the treated fibres to air resulted in a strongly
physisorbed layer onto the fibre surface most probably arising from moisture in the
environment. This may inhibit any desired reaction between fibre and resin during
composite processing. It is therefore suggested that treated fibres be immersed into resin
or a suitable coating material prior to exposure to air.
Acknowledgements- This work was sponsored by the Office of Naval Research under
contract No. N00014-86-K-0799, and was carried out in the Materials Research
Laboratory at the University of Illinois supported by the U.S.Department of Energy under
contract DE-AC 02-76ER 01198.
REFERENCES
1. H.M.Stoller and R.E.Allred, Proc. of l8th Internat.SAMPE Tech.Conf., 993 (1986).
2. J.B.Donnet, M.Brendie, T.L.Dhami and O.P.Bahl, Carbon, 24, No.6,757 (1986)
3. Sun Mujin, Hu Baorong, Wu Yisheng, Tang Ying, Huang Weiqui and Da Youxian,
Composites Science and Technology, 34, 353 (1989)
4. I.H.Loh, R.E.Cohen, and R.F.Baddour, J.Mat.Sci., 22, No.8, 2937 (1987).
5." J.C.Goan, US Patent 3,776,820 Dec.4 1973 assigned to the Great Lakes Corporation.
6. C.Jones and E.Sammann, in press, Carbon (1989).
7. C.Jones, Carbon, 27, No.3, 487 (1989)
8. Perkin Elmer 5400 XPS Software (1989).
9. C.Kozlowski and P.M.A.Sherwood, Carbon, 24, No.3, 357, (1986)
10. R.Nordberg, H.Brecht, R.G.Aldridge, A.Fahlmann and J.R.Van Wazer, Inorg.Chem., 9,
2469 (1970)
11. C.Kozlowski, Ph.D Thesis, Inorganic Chemistry, (1984) (Newcastle Upon Tyne,
England)
12. D.W.Davis and D.A.Shirley, J.Elec.Spec. and Rel.Phenom., 3, 137, (1974).
13. S.G.Chang and T.Novakov, Amorphous Enviroment,9, 495, (1975).
14. J.F.Evans and T.Kuwana, Analytical Chemistry, S1, No.3, 358 (1979).
413
DETERMINATION OF PARTICLE SIZE OF A DISPERSED PHASE
BY SMALL-ANGLE X-RAY SCATTERING
FRANK C. WILSON
Polymer Products Dept.,
Wilmington, DE 19880-0323
DuPont
Experimental
Station,
P 0
Box
80323,
ABSTRACT
A method for determining particle diameters up to ca 500 nm is described.
X-ray data
are
obtained
with
an ultra-high resolution Bonse-Hart
diffractometer and subsequently desmeared. The resultant data, viewed as the
invariant argument h I(h), are interpreted as arising from a log-normal
distribution of independent spherical particles.
The distribution is
characterized by its median value and breadth.
INTRODUCTION
Small-angle x-ray scattering has long been used to obtain estimates of
particle szes.
Very little has been published recently in terms of the
methodology, but utilization of, and references to the technique are very
common.
Guinier and Fournet [] and Hosemann and Bagchi [2) provide
comprehensi-. discussions of the technique as commonly cited today.
The
methods were developed before the easy availability of computers and rely on
the interpretation of
plots
of
the data, typically with tangent
extrapolations.
With
"conventional"
small-angle diffractometers, the
practical upper limit of particle-size determination is of the order of 100
nanometers, but the Bonse-Hart diffractometer [3,4] extends this range to
almost a micrometer.
The impetus for this work was the need to develop a procedure for
measuring particle sizes in "rubber-toughened" polymers, where a dispersed
elastomeric phase is employed to enhance the impact resistance of a host
polymer.
The particle sizes are typically a few tenths of a micrometer
(requiring the Bonse-Hart instrument) and the method is totally dependent upon
the availability of a computer to synthesize the modeled curves. Although the
method was developed for polymer blends, it is not restricted to them and can
be employed for any two-phase system where the dispersed phase is present as
approximately spherical "particles".
The
only requirement is that a
reasonable electron-density difference exists between the two phases. As well
as polymer blends (solid-in-solid), we have characterized discrete particles,
voids, emulsions and dispersions (solid-in-gas, gas-in-solid, liquid-in-liquid
and solid-in-liquid). The sizes which can be accommodated range from about 10
to 500 or more nanometers, depending upon the breadth of the distribution.
DISCUSSION
When observed
Bonse-Hart
at
small-angle
sufficiently high resolution, as is possible
x-ray diffractometer, the scattering
from
with
a
dispersed
particles smaller than about one micrometer in diameter conveys obvious
information about the size and distribution of size of such particles. Figure
1 shows the small-angle scattering from four polymer blends of nearly
equivalent composition (an elastomer dispersed in a semi-crystalline polymer)
but different processing history. The data are displayed as the small-angle
invariant, which is obtained by
multiplying the desmeared small-angle
intensity by the square of the scattering angle (which is proportional to h).
It is apparent that
the
scattering
from all four samples differs
significantly, and it would be useful to quantify the differences among the
smples with regards to some average size and size distribution.
Mal. Ras. Sot. Symp. Proe. Vol. 171. ' 1990 Materials Research Sociy
414
60
40
20
.... .......
"
10
0
.02
.04
.06
t0
.1
Two-Theta (dog)
Figure 1
Invariant Plots for Blends
-
The scattering observed from independently scattering spheres of radius
"r" is given by the equation 1. the Rayleigh equation (5], where u - hr:
)(1
- co2 u 2
1(h) - p1W - -9.min
The limit as "u" becomes large (since sin(u)/u approaches
approaches 1/2) is given by equation 2 [6]:
-
2
2
,A(h)
(u) -
9cos u
-- -
9
1
I '•
--
zero
and
cos u
(2)
which shows an inverse 4th power relationship with angle at higher angles. The
validity of this relationship is demonstrated by Figure 2, which shows the
scattering observed from polystyrene particle-size standards. The ordinate is
the third power of "h" times the observed (smeared) intensity, since slit
smearing changes an inverse 4th power relationship to an inverse 3rd power. As
"u" approaches zero, the Rayleigh equation can be approximated by a Gaussian
relationship, as in equation 3:
I(h)-
S
p'(u)
(
)
(3)
which is the well known Guinier approximation [7].
4
.. 'V 01 2g' : 3
j
Vw,,V,,
,
jv4~
0
.1
VV.. .
0.481~
.2
.3
4
5
.6
Two-Theta
Figure 2 - Observed Data from PS Standards
415
Since the ripples in the scattering, which are implicit in the Rayleigh
equation, are obliterated when there exists any significant particle-size
distribution, the scattering from spheres of a given size can be modeled by a
curve which is Gaussian at low angles and inverse 4th power at high angles.
Since the Gaussian and inverse 4th power curves cross twice, a smooth
interpolation can be used to join them. Figure 3 shows the Gaussian, inverse
4th power and interpolated curves displayed as the invariant argument, where
they are most easily and usefully visualized. Since the modeled intensity is
multiplied by a number proportional to the square of the angle, the Gaussian
curve starts at zero and the inverse 4th power curve becomes an inverse
square.
125
.,
too
2
£
.
75 -
Gowssion
Inverse 4th Power
-Interplaod
Mode
50
25
a
.02
.06
.04
.08
Two-Theta (deg)
Figure 3
-
Invariant Plots for Dio. = 250 nm
In Figure 4, this interpolated approximation is plotted with curves
derived from the Rayleigh equation (1) and showing the scattering expected
from a monodisperse population and from a narrow particle-size distribution.
Beyond the first maxi", the approximation effectively bisects the ripples
arising from the cos term of equation 2 for both curves derived from the
Rayleigh equation.
177. Intsmal
2
td Mo"k
I
oytse"
.o
.1
Two-Theta
....... Manodispee
(Dsrbto Ratio -
1.15)
(from qayleigh equation)
1.5
SI
0
Figure 4
I
-
.15
(dog)
.2
Invariant Plots for Oia. - 250 nm
416
the
finite distribution,
the scattering from a
In order
to predict
logarithm of
the diameter of the particles
is
assumed
to be
normally
distributed and this
is
approximated by summing the contributions
of 23
fractions,
equally spaced with regard to the logarithm of the relative
size,
between plus and minus 3-sigma. The logarithm at 1-sigma differs from that at
the
center of
the distribution by an amount corresponding to the ratio
of
sizes which characterizes the breadth of the distribution. This "Distribution
to
Ratio" and the
size (diameter) at the center of the distribution serve
characterize
the
log-normal
distribution fully. Since the
area of the
invariant curve
is proportional to
the amount of that
fraction present
the
particle-size
(hence
the
name "invariant"),
the
regardless
of
contributions can be summed directly.
The center of the distribution is
a
median value, with half the mass of the dispersed phase consisting of
larger
Figure 5 shows
particles and half the mass consisting of smaller particles.
the
expected scattering from distributions with a characteristic diameter of
250
nm and distribution ratios varying from one (monodisperse) to three.
The
contributions from the larger particles causes the invariant maximum to shift
to
lower angles while the contributions from the smaller particles raises the
level at higher angles.
100
Dia.
.... 250.0
250.0
.......
250.0
80
---
20
0
D.R.
3.0
2.5
2.0
..
0
.02
.04
06
.08
Two-Theta (deg)
Figure 5
Invariant Curves with Same Median
Figure 6
is similar to figure 5 except that each curve has its maximum
where it occurs
for
a 250 nm monodisperse population.
The method
for
determining median particle-size and distribution-breadth is implicit in this
figure:
The position of the maximum is inversely proportional to the diameter
of a monodisperse particle size; the breadth of the peak (taken as the
ratio
of the position of the high-angle half-maximum to that of the maximum) enables
the empirical calculation of the median diameter and distribution ratio.
As
the method has evolved in our laboratory, the characterization of a blend is
accomplished by displaying
the invariant curve on a graphics
terminal and
locating the position of the maximum with the graphics cursor. The computer is
programmed
to
locate the position of the high-angle half-maximum,
calculate
the median diameter and distribution ratio, and then calculate and overlay the
expected
invariant curve for this distribution. At times, a slight movement
of
the estimated position of the maximum provides a more satisfactory visual
match between the observed and calculated invariant curves, but the estimated
median diameters and distribution ratios seldom change significantly.
417
100
Dia.
63.8
93.2
136.6
------197.1
250.0
---
so,
,
60",
Froc.
D.R.
.255
3.0
.373
.546
.788
1.00
2.5
2.0
1.5
1.0
20
0
0
Figure 6
.04
.08
Two-Theta (dog)
"
.12
.16
Invariant Curves with Same Maximum
Figure 7 shows the match between observed and calculated invariant curves
for two of the samples of Figure 1. The positions of the maxima are nearly
the same, but the shapes (or widths at half-maximum) are obviously different.
This difference in shape leads to significant differences in the estimated
8 shows
median diameter and distribution ratio as noted in the figure. Figure
histogram of particle sizes that is generally furnished with the results
a
the
from a sample. Although there is no real additional information beyond
skewed
mean diameter and distribution ratio, a visual presentation of the useful.
distribution and enumeration of the mode and mean values has proved
The individual bars represent the mass of particles in a given 25 nm range.
The mode is contained in the tallest bar and is always less than the median,
figure and
and the mean is located at the horizontal center of gravity of thebetween
the
is always greater than the median. The median is near the line
of
Comparison
bars with the fine shading and those that are cross-hatched.
generally
results such as these with those from electron microscopy are in assumptions
good agreement, usually within a factor of two. Considering the
Typically
and averaging implicit in both techniques, this is satisfactory.
the electron diffraction results are smaller, indicating a favoring of the
more common particles--nearer the mode. One should bear in mind that an
electron microscopy result is usually based on observation of a few hundred
irradiat
particles and is commolly an "eyeball" estimate; for a typical about
10
volume of about 10 mm , the x-ray technique averages over
particles.
60
Median ODo. - 165 nm
Oist.Ratio
t.52
50
40
"X3 20
4O
:/
/
MedianRatio
Oio.,"2.13
-104
Dist
nrn
to
0
0A
3
.02
.04
.06
.08
.A
Two-Theta (dog)
Figure 7 - Fitted Invariants for Blends
418
0WD
Median = 165 nm
Ratio - 1.52
14D.0
Mode = 138 nm
=
CMean
a-40.0
Wa.
-
179 nm
i
20
00
200
Figure 8 -
400
wo
Diameter (rm)
8000
Bo
00 0
Histogram of Particle Sizes
CONCLUSIONS
of polymer
The method described here provides a useful characterization
are nearly
blends when the dispersed phase consists of particles which
not
The assumption of a log-normal particle-size distribtion has
spherical.
A surprisingly good fit between
proved to be particularly restrictive.
of cases, and
observed and calculated curves is obtained in a great majority trends among
of
the kind of information generally desired is usually that
are sufficient
generally similar samples. The x-ray data themselves commonly
properties due to a
to answer the question: "Are the observed differences in
two items of
difference in particle size (and/or distribution)?". In essence,
input, the position of the invariant maximum and the high-angle half-height,
and the
are used to determine two items of output, the median particle-size
great variety of particle size
a
Granted that
distribution ratio.
for
allowing
curves,
distributions could lead to indistinguishable scattering
of freedom as to
models other than the log-normal would offer so many degrees
make results much less useful in general.
REFERENCES
1.
A. Guinier and G. Fournet, Small-Anle -catterina of X-RavY,
(John Wiley & Sons, Inc., N.Y., 1955), Chapter 4
2.
of Diffraction
R. Hosemann and S. N. Bagchi, Direct AnalysisChapter
XVII
(Interscience Publishers Inc., N.Y.. 1962),
3.
Koffman, Advances in X-ray Analysis, 2U, 334-338, 1967
4.
0. Clatter and 0. Kratky, Small Anele X-Ray Scatterina,
(Academic Press, London, New York, 1982), pps 69-72
5.
Lord Rayleigh, Proc. Roy. Soc. (London), AN, 219 (1914)
6. A. Guinier and G. Fournet, op. cit., p 17
7.
A. Guinier, Ann. phys., UZ, 161-237 (1939)
by Matter,
419
SYNTHESIS AND CHARACTERIZATION OF A THERMOTROPIC POLYALKANOATE OF 4,4'DIHYDROXY-a,a'-DIMETHYLBENZALAZINE
H.FRIUTWALA, A.L.CIMECIOGLU AND R.A. WEISS
Polymer Science Program, Institute of Materials Science, University of Connecticut, Storrs,
CT 06269-3136
INTRODUCTION
Several research projects in our laboratory have used an LCP based on 4,4'-dihydroxya-a'
dimethylbenzalazine, that was first synthesized by Roviello and Sirigu 1,2. Copolymers with the
followinq general structure,
C = N-
N= C-
(CH
.. C
- (CHz) 1o-
have been used inself-reinforcing polymer blends3 and are currently being used in small angle neutron
scattering studies of the molecular conformation of an LCP in a nematic field. In this latter study, it is
imperative that the polymer be well characterized, especially with respect to molar mass and thermal
stability. This paper describes the effect of molar mass and thermal history on the thermal stability
and transitions of this particular LCP.
EXPERIBMAL
Hydrazine monohyrate (Eastman Kodak), 4-hydroxyacetophenone (Aldrich), pyridine
(Anhydrous, Aldrich), triethylamine (Gold Label, Aldrich) and chloroform (HPLC, Aldrich) were used as
supplied. N-methylpyrrolidone (NMP) (Aldrich) was dried by refluxing over calcium hydride, distilled and
was stored over molecular sieves Type 3A. Sebacoyl chloride and dodecanedioyl dichloride (Aldrich)
were purified by vacuum distillation and stored under nitrogen prior to use. All other solvents used were
reagent grade and were employed without further purification.
4,4'-diIdroxy-a,a'-dimethylbenzalazine was prepared according to the reported procedure, M.
pt. 225-226 *C
Solution polymerization was carried out ina three-neck round bottom flask (500 mL) equipped
with an addition funnel, a reflux condenser and a mechanical stirring assembly under nitrogen
atmosphere, according to the following general procedure. The diol (20-35 mmol) was dissolved in a
solution of chloroform (70 mL) containing an excess of the acid acceptor (pyridine or triethylamine; 5-25
mL) and, insome cases, a small amount of NMP (5-10 mL). The latter helped to facilitate dissolution of
the otherwise insoluble azine inchloroform. An equivalent or aslight excess (1 mole%) of a 50/50 molar
mixture of the acid chlorides (sebacoy/dodecanedioyl; 20-35 mmol) inchloroform (30 mL) was added to
the solution over a period of 10 min-1.5 h. with vigorous stirring. The temperature was maintained
between 0-60 OC. Following completion of the addition, the solution was stirred vigorously for a further
4h.
Mat. Res. Soc. Symp. Proc. Vol. 171., 1990 Materials Research Society
420
After completion of the reaction, the mixture was diluted with chloroform (ca. 50 mL) to reduce
its viscosity. The products were then recovered by precipitation into methanol (500-1000 mL). They
were broken up and washed in a high speed blender, filtered, further washed with aliquots of methanol,
methano/water, and water and were finally dried in vacuo at 60-70 0C for at least 24 h. In all cases,
the yields were near quantitative
A number of analytical techniques were used to characterize these liquid crystalline polymers
(LCP), including differential scanning calorimetry (DSC), thermal gravimetric analysis (TGA),and gel
permeation chromatography (GPC).
2M A Perkin Elmer DSC-7 was used to measure the enthalpy changes and specific heat changes
associated with the thermal transitions. The samples were heated at 20 0C and cooled at 10 0C in all
experiments. The cell was continuously flushed with dry nitrogen.
TGA. A Perkin Elmer TGA-7 was used to study the thermal stability of the LCP.in a dry nitrogen.
Both dynamic and isothermal experiments were made. Following the TGA experiments, GPC
measurements were made on the same samples in order to assess the effect of thermal history on the
molar mass of the polymer.
RESULTS AND DISCUSSION
Characterization of thermal transitions. The DSC thermograms of the LCP are shown in the Fig. 1.
The polymer had a glass transition temperature (Tg) of 24 0C, and a nematic to isotropic transition
(TN.II at 245 PC. The polymer showed two peaks, T1 & T2 , in the temperature range of 140-160°C
and exhibits a nematic mesophase above 1600C and up to TN.. The neutron scattering studies (Fig. 2)
indicate the possibility of an additional mesophase just before the solid to nematic transition, which
would give a peak in the OSC thermograms. As seen in Fig. 1 the cooling curve does not show two
crystallization peaks. A detailed investigation of this somewhat surprising behaviour is currently
underway in our laboratory.
The effect of the weight average molar mass on the thermal transitions is shown in Fig. 3.
Below a molar mass about 50000 (polystyrene equivalent), T1 , T2 and TN.,, increases with increasing
molar mass and the transition temperatures plateau at higher Mw's. There is relatively small change in
Tg with changes in molar mass.
The effect of thermal history on the transitions is shown in Fig. 4. Prior to each run the sample
was annealed at elevated temperature. The sample was then cooled back to its starting temperature
and the subsequent heating scans are shown in Fig. 4. Annealing at elevated temperature raised Tg.
This was a consequence of the reduced crystallinity of the polymer. There was no significant influence
of annealing time on T1 and T2 when the sample was annealed at 2200C. The peaks moved to lower
temperatures as the sample was annealed for longer period at temperatures close to its and above the
TN.. The transitions were broadened as the annealing time was increased. Thus, the order of the melt
and the kinetics of ordering are affected by annealing, and this did not appear to be due to polymer
degradation as indicated by only minor changes in molar mass as discussed below. This was also
evident in the neutron scattering contour plots (Fig.3), during which sample was annealed for long
periods of time at elevated temperatures.
Characterization of Thermal Stability-. The LCP was stable up to 350 0C as shown by the TGA
thermogram in Fig. 5 As with the DSC experiments, the LCP was exposed to various annealing
temperatures for varying longths of time. The details of these experiments are summarized in Table I.
In most cases the mass loss was < 1%. After TGA analyses, the same samples were dissolved in
chloroform for GPC analyses. These results are also given in Table I. Annealing at 220 0C actually
appeared to increase the molar mass, which suggests that some additional polymerization or
421
FIG. 1
1.40
A
.
T T2
IA
0..T"~M
1.A typical
-0.
080 thermogram of the LCP under investigation.
FIG. 2
0.C
2. Neutron scattered intensity inthe plane perpendicular to the incident beam at (a)2500C, (b)2200C.
(c)2000C, (d)1700C, (a)1600C and (I) 1500C. Magnetic field direction ison the horizontal axis.
422
equilibration of the molar mass took place. At 240 OC, which is in the broad clearing transition, and at
260 oc,which is in the isotropic region, the molar mass decreased with increasing annealling time. The
molar masses reached with annealing for one hour at each temperature were similar and corresponded
to about a 10% reduction in molar mass from the starting polymer (i.e., no high temperature annealing).
In any event, the molar mass of the LCP after all the annealing histories given in Table I were above
50,000, and therefore, the changes in molar mass were most likely not responsible for the changes in
crystallization and mesophase formation described earlier.
FIG. 3
ZO
INo'r AICLIQID
T
,IT,
W.0
**
CRYSTAL,
LIQUID
T
1
TM.,
it
GLASS + CRYSTAL
a
100000
200000
uimatmwaa
3. Effect of the molecular weight on the thermal transitions.
423
FIG.*4
6..
4..
t
LIA
5.8
*AA
40
l.
l.
S
4.Effect of annealing on the thermal transitions.
l
424
FIG.5
7 Seri-
PEN UK"-a
Thwal Analysis Systm
4.0
LO,
7°0.
LO"
La"
IBMS.
.. a -
ft
9&- W
5. TGA curve of the polymer.
TABLE
I
NE.AIUN
TEMP t Q)
TIME (Min)
220
10
700
3.2
220
30
77000
3.55
220
60
78700
3.69
240
10
75000
3.46
240
30
66000
3.36
240
60
57050
3.08
260
10
69600
3.57
260
30
64700
3.87
260
60
60400
3.79
Nit
•
77700
2.39
1. Effect of ennaing temiet-p
Mw
MwIMn
me on molecular wight of IN polymer.
425
CONCLUSION
The liquid crystalline polymer discussed in this report is relatively stable at high temperatures,
and a small reduction in its weight and molecular mass was only observed after prolonged exposure to
such temperatures. Neutron scattering results seem to indicate that there is a second mesophase prior
to the nematic transition. The nature of this additional phase is not clearly identified as yet and is
currently under investigation.
REFERNCE
1.Roviello A.and Sirigu R., J.Poly Sci., poly. lett. ed. 13 455 (1975)
2. Roviello A.and Sirigu R., Eur. Poly. J. 15 61 (1979)
3.Weiss R,Huh W.S., Nicolais L, Pol. Eng. Sci. 27 684 (1987)
4. Blout ER,Eager V.W., and Golfstein R.M. J.Am. Chem. Soc. 68 1983 (1946)
427
Author Index
Ackerman, Jerome L., 65
Anastajiadis, Spiros H.,
Aubrey, Norman E., 197
317
Bauer, Barry J., 203
Berry, G.C., 141
Blackson, J.H., 159
Bommannavar, A., 337
Brennan, A.B., 15
Briber, Robert M., 203
Brostow, Witold, 177, 183
Candau, FrancoiSe, 71
Chau, C.C., 159
Chen, Gang-Fung, 293
Chen, Tsuey Ing, 211
Christensen, Thomas M., 317
Chu, Benjamin, 237
Chung, W.C., 379
Cimecioglu, A.L., 419
Cohen Addad, Silvie, 125
Cohen-Addad, j.P., 365
Composto, R.J., 335, 337
Crevecoeur, Guido, 165
Deline, Vaughn R., 343
Doremus, Robert H., 79
Dowrey, A.E., 117
Dziemianowicz, Theodore S.,
177
Elandjian, Lucy, 395
Engbretson, M., 337
Evrard, P., 349
Farago, Bela, 57
Feicher, G.P., 329, 335
Fornes, R.E., 371
Forsman, W.C., 355
Frisch, Harry L., 231
Friutwala, H., 419
Garrido, Leoncio, 65
Giannelis, E.P., 39
Gilbert, R.D., 371
Gillion, Laura R., 275
Gong, Kecheng, 267
Green, Peter F., 317
Groeninckx, Gabriel, 165
Grothaus, J.T., 117
Guadiana, Russell, 125
Gvozdic, N., 105
Haghighat, R. Ross,
Hahn, K., 261
395
Hallmark, V.M., 401
Hammond, L.C., 349
Hess, Michael, 177, 183
Hirose, Yoshiharu,' 389
Hsiao, Benjamin S., 125
Huang, Hao-Hsin, 15
Huang , S.J., 299
Huang , Y.S., 217
Hurley, Jr., William J., 79
Hwang, Jenn-Chiu, 305
Hiwang, wen-Fang, 131, 159
Ibenesi, J., 105
Interrante, Leonard V., 79
Jian, Li, 57
Jones, C., 407
Jones, Frank N., 293
Jones, R.A.L., 335, 337
Kajiyama, Tisato, 305
Kaxnigaito, Osami, 45
Karasz, Frank E., 197
Karim, A., 329, 335
Kawasumi, Masaya, 45
Kikuchi, Hirotsugu, 305
Klassen, H.E., 159
Kojima, Yoshitsugu, 45
Kopelman, Raoul, 245
Kosfeld, Robert, 177. 183
Kramer, E.J., 335, 337
Krause, Stephen J., 131
Kuemin, M., 105
Kurauchi, Toshio, 45, 389
Kyu, Themn, 211
Lamne. Richard M., 31
Latshaw, B.E., 355
Lecat, J.H., 349
Li, Ching-Shan, 245
Lonikar, S.V., 371
Lusiglea, Richard W., 395
Macnight, William J.. 197
Mahler, Walter, 237
Mansour, A., 329, 335
Marcott, C., 117
Mark, James E., 51, 57, 65.
275
McCarthy, David, 57
Mehrotra, V., 39
Meier, D.J., 105
Miyamoto, Akira, 305
Moritcxni, Satoru, 305
428
Noda, I., 117
Norton, L.J., 335
Nyitray, Alice M.,
Uy, William C.,
99
O'Reilly, James M., 225
Okada, Akane, 45, 389
Onn, David, 99
Paleos, Constantinos M., 87
Parker, Alex J., 189
Pawlikowski, Gregory T., 299
Pekala, Richard W., 285
Perusich, E.R., 153
Piel, J.P., 349
Prilutski, Gerard M., 147
Rabolt, J.F., 401
Rafailovich, M., 337
Rahn, Jeffrey A., 31
Roach, J.F., 217
Rodrigues, David, 15
Rojstaczer, S.R., 171
Rungsimuntakul, N., 371
Russell, Thomas P., 317, 329,
335, 343
Sammann, E., 407
Sanford, W. Michael, 147
Schaefer, Dale W., 51, 57,
275
Schmidt, Helmut, 3
Schmidt, Wayde R., 79
Schubert, Frank, 183
Schulte, A., 401
Schwahn, D., 261
Sedita, Joseph S., 225
Shi, Zhong-You, 245
Shiga, Tohru, 389
Smith, B.A., 171
Smith, S.D., 117
Sokolov, J., 337
Song, K., 401
Springer, T., 261
Stehle, J.L., 349
Stein, Richard S., 125, 335,
337
Stille, John K., 189
Streib, J., 261
Subramanian, R., 217
Sukumar, Vijay, 79
Sullivan, V.J., 141
Sun, C.-C., 57
Tarshiani, Y., 105
Thomas, Edwin L., 255
Thomas, O.T., 349
Tsang, Joseph W., 189
Twieg, R., 401
Usuki, Arimitsu, 45
Volksen, W.,
153
171
Wakharkar, Vijay S., 343
Wang, Bing, 15
Weeks, Norman, 125
Weiss, R.A., 299, 419
Wiff, D.R., 217
Wilkes, G.L., 15
Williams, Joel M., 99
Wilson, Frank C., 413
Winey, Karen I., 255
Witek, Adam, 99
Wu, D.T., 355
Wu, Dan Q., 237
Wu, Wen-Li, 189
Yamaoka, Hiroaki, 197
Yang, Jan Chang, 211
Yoon, D.Y., 171
Yun, W.B., 337
Zhang, Xinghua, 267
Zhang, Zhi-Fan, 31
Zhao, X., 337
429
Subject Index
acrylamide, 72
adsorbed molecules, 349
adsorption, kinetics of, 365
aerogels, organic, 285
aerosol OT, 72
aerospace, 371
aggregation, 125
alcoholysis, 33
aspect ratio, 379
bilayer films, 343
blends, 177, 183, 197, 211,
245, 255
isotopic polymer, 335
calorimetry, 184
carbon fibre, 407
catalytic, 31
ceramer, 15
ceramic materials, 51
ceramics, 5
comb-like copolymers, 293
compatibility, 189
compliances, 141
composites, 79, 379
in situ, 165
molecular, 131, 153
compressive strength, 99
conductive polymeric
composite, 379
conductivity, electrical, 39
conductors, synthetic, 39
constitutive equation, 141
copolymers, 31, 105, 177,
189, 317
isotope-labeled block, 117
methylsilsesquioxane, 32
styrene-isoprene diblock,
117
correlation function
analysis, 237
crosslink, 203
cross-linkable acrylic, 293
i
deformation, 389
density, 99
depth profiles, 343
diblock copolymer, 255, 343
dielectric studies, 225
diffusion lengths, 329
diphenylsiloxane, 105
doped polymers, 245
dynamic, 141
mechanical analysis, 184
elastomers, 57
filled, 51
electric field, 389
electron microscopy, 51
ellipsometry, 349
emulsion,. 71
entanglements, 231
enthalpic relaxation times,
225
epoxy, 371
excellent properties, 45
fiber, 159
composites, 371
fibril, 159
fillers, 51
flexible strength, 267
flocculation, 71
fluorescence, 245, 337
fluorohectorite, 39
foams, 99
formaldehyde, 285
Fourier-transform, 349, 401
fractal, 159
freeradical initiated, 293
volume parameters, 225
FTIR, 217
fusion kinetics, 245
gels, 389
glasses, 5
grafted polymer, 355
graphite, 371
heat distortion temperature,
45
high temperature stability, 31
homopolymer, 255
imaging, 65
impact resistance, 267
Inconel 601, 379
infrared, linear dichroism,
117
injection moulding, 165
in situ precipitated SiO 2 , 65
intercalation, 39
inteifaces, 337, 355
interlayer distance, 47
interphase, 117
ionic aggregates, 237
iomoners, morphology of, 237
IPN, 231
I
43
48O
latexes, 73
layered silicate, 39
LCP, 419
light scattering, 125, 305
liquid
crystalline polyester, 299
crystals, 177, 183, 305
longitudinal acoustic mode,
401
mean-field theory, 335
mechanical properties, 141,
165
melamine formaldehyde, 267
micelles, 71
microcomposites, 153
polyamide thermoplastic,
153
thermoplastic, 147
microphase separation
transition, 317
microscopy, 184
microwave, 15
miscibility, 217
miscible, 225
polymer blends, 197
moduli, 371
molar mass, 419
molecular
composites, 125, 147, 171,
299
rigid rod, 131
conformation, 419
dispersion, 189
montmorillonite, 45
morphology, 131, 159, 256,
267
multilayers, 39
-
nanocomposites, 39, 45
nematic, 305
networks, 231
interpenetrating, 5
model reinforced, 65
neutron
reflection, 329, 335
scattering, 51
6w materials, 39
NMR TGA and DTA, 31
nucleation, 75
57
-and-growth,
nylon
6-clay hybrid (NCH), 45
ordered bicontinuous double
diamond, 255
ordering, transition, 317
organic-inorganic, 15
organometallic, 79
organosiloxanes, 105
.4
particle size, 413
particulates, 57
PBZT, 153
PBZT/PEKK, 147
PDMS, 51
perfluoroalkane oligomers, 401
phase
diagrams, 177
segregation, 211
separation, 57, 159, 171
phosphorescence, 245
plasmas, 407
PMMA, 217
polyaniline, 39
polybenzimidazole, 197
polydimethylsiloxane, 65
polyester
/polycarbonate, 225
stiff chain, 125
polymer(s), 131, 329
adsorption, 365
blends, 225
conjugated, 39
high performance, 197
modified silica glasses, 65
morphology, 245
rigid-rod heterocyclic, 153
synthesis, 419
thermotropic liquid
crystalline, 165
toughened glasses, 57
polymeric precursor, 79
polymerization, 293
polymerized
micelles, 87
vesicles, 91
polyimides, 171
fluorine-containing, 197
poly (p-phenylene
benzobisthiazole), 131, 147
polyquinolines, 189
polysilsesquioxanes, 31
polystyrene, 203
poly(vinyl alcohol) (PVA), 267
PPTA/AN, 211
pressure, 401
processing, 217
properties, mechanical, 79
PS/PS, 261
PS/PVME, 261
Raman spectroscopy, 401
reinforcement, 51, 177
resorcinol, 285
rheology, 141
rigid rod, polymer, 153
robot, 389
SAN, 217
SANS, 189, 203, 355, 419
431
SAXS, 15, 189, 237, 413
anomalous, 237
scattering, 51, 211
scratch, 8
secondary ion mass
spectroscopy (SIMS), 343
segmental interactions, 117
segmented copolymers, 299
segregation, 343
semiflexible, 401
shear-thinning, 75
silica, 355
-siloxane, 365
small angle scattering, 189
smectic, 305
sol gel, 3, 15, 285
spectroscopy
infrared, 39
Raman, 39
spinning, 165
spinodal, 57, 212, 217, 261
stress-strain, 371
surface, 317
area, 79, 99
enrichment, 335
saturation, 365
swelling, 365
synchrotron, 337
synthesis, 31, 299
thermal
conductivity, 99
history, 419
stability, 419
thin film coatings, 171
transmission electron
microscopy, 256
viscosity, 141
x-ray photoelectron
spectroscopy, 407
MATERIALS RESEARCH SOCIETY SYMPOSIUM PROCEEDINGS
ISSN 0272 - 9172
Volume 1-Laser and Electron-Beam Solid Interactions and Materials Processing,
J. F. Gibbons, L. D. Hess, T. W. Sigmon, 1981, ISBN 0-444-00595-1
Volume 2-Defects in Semiconductors, J. Narayan, T. Y. Tan, 1981,
ISBN 0-444-00596-X
Volume 3-Nuclear and Electron Resonance Spectroscopies Applied to Materials
Science, E. N. Kaufmann, G. K. Shenoy, 1981, ISBN 0-444-00597-8
Volume 4-Laser and Electron-Beam Interactions with Solids, B. R. Appleton,
G. K. Celler, 1982, ISBN 0-444-00693-1
Volume 5-Grain Boundaries in Semiconductors, H. J. Leamy, G. E. Pike,
C. H. Seager, 1982, ISBN 0-444-00697-4
Volume 6-Scientific Basis for Nuclear Waste Management IV, S. V. Topr 1982,
ISBN 0-444-00699-0
Volume 7-Metastable Materials Formation by Ion Implantation, S. T. Picraux,
W. J. Choyke, 1982, ISBN 0-444-00692-3
Volume 8-Rapidly Solidified Amorphous and Crystalline Alloys, B. H. Kear,
B. C. Giessen, M. Cohen, 1982, ISBN 0-444-00698-2
Volume 9-Materials Processing in the Reduced Gravity Environment of Space,
G. E. Rindone, 1982, ISBN 0-444-00691-5
Volume 10-Thin Films and Interfaces, P. S. Ho, K.-N. Tu, 1982, ISBN 0-444-00774-1
Volume 11-Scientific Basis for Nuclear Waste Management V, W. Lutze, 1Q)82,
ISBN 0-444-00725-3
Volume 12-In Situ Composites IV, F. D. Lemkey, H. E. Cline, M. McLean, 1982,
ISBN 0-444-00726-1
Volume 13-Laser-Solid Interactions and Transient Thermal Processing of Materials,
J. Narayan, W. L. Brown, R. A. Lemons, 1983, ISBN 0-444-00788-1
Volume 14-Defects in Semiconductors I, S. Mahajan, J. W. Corbett, 1983,
ISBN 0-444-00812-8
Volume 15-Scientific Basis for Nuclear Waste Management VI, D. G. Brookins, 1983,
ISBN 0-444-00780-6
Volume 16-Nuclear Radiation Detector Materials, E. E. Hailer, H. W. Kraner,
W. A. Higinbotham, 1983, ISBN 0-444-00787-3
Volume 17-Laser Diagnostics and Photochemical Processing for Semiconductor
Devices, R. M. Osgood, S. R. J. Brueck, H. R. Schlossberg, 1983.
ISBN 0-444-00782-2
Volume 18-Interfaces and Contacts, R. Ludeke, K. Rose, 1983, ISBN 0-444-00820-0
Volume 19-Alloy Phase Diagrams, L. H. Bennett, T. B. Massalski, B. C. Giessen,
1983, ISBN 0-444-00809-8
Volume 20-Intercalated Graphite, M. S. Dresselhaus, G. Dresselhaus, I E Fischer,
M. J. Moran, 1983, ISBN 0-444-00781-4
Volume 21-Phase Transformations in Solids, T. Tsakalakos, I984, ISBN 0-444-00001-0
Volume 22-High Pressure in Science and Technology, C. Homan, R. K MacCrone,
E. Whalley, 1984, ISBN 0-444-00032-9 (3 part set)
Volume 23-Energy Beam-Solid Interactions and Transient Thermal Processing,
J. C. C. Fan, N. M. Johnson, 1984, ISBN 0-444-00903-5
Volume 24-Defect Properties and Processing of High-Technology Nonmetallic
Materials, 1. H. Crawford, Jr., Y Chen, W. A. Sibley, I084,
ISBN 0-444-00904-3
Volume 25-Thin Films and Interfaces i, I E. E.Baglin, D. R Campbell, W K (hu,
1984, ISBN 0-444-00905-1
MATERIALS RESEARCH SOCIETY SYMPOSIUM PROCEEDINGS
Volume 20-Scientific Basis for Nuclear Waste Management VII, G. L. McVay, 1984,
ISBN 0-444-00906-X
Volume 27-lon Implantation and Ion Beam Processing of Materials, G. K. Hubler,
0. W. Holland, C. R. Clayton, C. W. White, 1984, ISBN 0-444-00869-1
Volume 28-Rapidly Solidified Metastable Materials, B. H. Kear, B. C. Giessen, 1984,
ISBN 0-444-00935-3
Volume 29-Laser-Controlled Chemical Processing of Surfaces, A. W. Johnson,
D J.Ehrlich, H. R. Schlossberg, 1984, ISBN 0-444-00894-2
Volume 30-Plasma Processing and Synthesis of Materials, J. Szekely, D. Apelian,
1984, ISBN 0-444-00895-0
Volume 31-Electron Microscopy of Materials, W. Krakow, D. A. Smith, L. W. Hobbs,
1Q84, ISBN 0-444-00898-7
Volume 32-Better Ceramics Through Chemistry, C. J. Brinker, D. E. Clark,
D. R. Ulrich, 1Q84, ISBN 0-444-00898-5
Volume 33-Comparison of Thin Film Transistor and SOI Technologies, H. W. Lam,
M. 1.Thompson, 1984, ISBN 0-444-00899-3
Volume 34-Physical Metallurgy of Cast Iron, H. Fredriksson, M. Hillerts, 1985,
ISBN 0-444-00938-8
Volume 35-Energy Beam-Solid Interactions and Transient Thermal Processing/ 1984,
D. K. Biegelsen, G. A. Rozgonyi, C. V. Shank, 1985, ISBN 0-931837-00-b
Volume 30-Impurity Diffusion and Gettering in Silicon, R. B. Fair, C. W. Pearce,
J. Washburn, 1985, ISBN 0-931837-01-4
Volume 37-Layered Structures, Epitaxy, and Interfaces, J.M. Gibson, L. R. Dawson,
1985, ISBN 0-)31837-02-2
Volume 38-Plasma Synthesis and Etching of Electronic Materials, R. P. H. Chang,
B. Abeles, 1085, ISBN 0-Q31837-03-0
Volume 30-High-Temperature Ordered Intermetallic Alloys, C. C. Koch, C T. Liu,
N. S. Stoloff, 1985, ISBN 0-931837-04-q
Volume 40--Electronic Packaging Materials Science, E. A. Giess, K.-N. Tu,
D R. Uhlmann, 1985, ISBN 0-931837-05-7
Volume 41 -Advanced Photon and Particle Techniques for the Characterization of
Defects in Solids, 1. B. Roberto, R. W. Carpenter, M. C. Wittels, 1985,
ISBN 0-Q31837-0o-5
Volume 42 -Very High Strength Cement-Based Materials, J. F. Young, 1085,
ISBN 0-931837-07-3
Volume 43
Fly Ash and Coal Conversion By-Products: Characterization, Utilization,
and Disposal I, C. 1. McCarthy, R. J. Lauf, 1)85, ISBN 0-931837-08-1
Volume 44 -Scientific Basis for Nuclear Waste Management VIII, C. M. Jantzen,
I A. Stone, R. C. Ewing, 1Q85, ISBN 0-931837-OQ-X
Ion Beam Processes in Advanced Electronic Materials and Device
Technology, B. R.Appleton, F.H. Eisen, T. W. Sigmon, 1Q85, ISBN 0-931837-10-3
Volume 4o -Microscopic Identification of Electronic Defects in Semiconductors,
N. M. Johnson, S. G. Bishop, G. D. Watkins, 1985, ISBN 0-931837-11-1
Volume 45
Thin Films: The Relationship of Structure to Properties, C. R. Aita,
K. S SreeHarsha, 1985, ISBN 0-931837-12-X
Volume 48 Applied Materials Characterization, W. Katz, P. Williams, 1985,
ISBN 0-031837-13-8
Volume 49- Materials Issues in Applications of Amorphous Silicon Technology,
) Adler, A Madan, M. 1. Thompson, 1985, ISBN 0-931837-14-6
Volume 47
MATERIALS RESEARCH SOCIETY SYMPOSIUM PROCEEDINGS
Volume 50-Scientific Basis for Nuclear Waste Management IX, L. 0. Werme, 1986,
ISBN 0-931837-15-4
Volume 51-Beam-Solid Interactions and Phase Transformations, H. Kurz, G. L. Olson,
J. M. Poate, 1986, ISBN 0-931837-16-2
Volume 52-Rapid Thermal Processing, T. 0. Sedgwick, T. E. Seidel, B.-Y. Tsaur, 198o,
ISBN 0-931837-17-0
Volume 53-Semiconductor-on-Insulator and Thin Film Transistor Technology,
A. Chiang. M. W. Geis, L. Pfeiffer, 1986, ISBN 0-931837-18-9
Volume 54-Thin Films-Interfaces and Phenomena, R. J.Nemanich, P. S. Ho, S. S. Lau,
1986, ISBN 0-931837-19-7
Volume 55-Biomedical Materials, J. M. Williams, M. F. Nichols, W. Zingg, 198b,
ISBN 0-931837-20-0
Volume 56-Layered Structures and Epitaxy, J.M. Gibson, G. C. Osbourn, R. M. Tromp,
1986, ISBN 0-931837-21-9
Volume 57-Phase Transitions in Condensed Systems-Experiments and Theory,
G. S. Cargill I1, F. Spaepen, K.-N. Tu, 1987, ISBN 0-931837-22-7
Volume 58-Rapidly Solidified Alloys and Their Mechanical and Magnetic Properties,
B. C. Giessen, D. E. Polk, A. I. Taub, 1986, ISBN 0-931837-23-5
Volume 59-Oxygen, Carbon, Hydrogen, and Nitrogen in Crystalline Silicon,
J. C. Mikkelsen, Jr., S. J.Pearton, 1.W. Corbett, S. 1. Pennycook, 108to,
ISBN 0-931837-24-3
Volume 60-Defect Properties and Processing of High-Technology Nonmetallic
Materials, Y. Chen, W. D. Kingery, R. J.Stokes, lO8o, ISBN 0-931837-25-1
Volume 61 -Defects in Glasses, F. L. Galeener, D. L. Griscom, M. ). Weber, 1O!;0,
ISBN 0-931837-2b-X
Volume 62-Materials Problem Solving with the Transmission Electron Microscope,
L. W. Hobbs, K. H. Westmacott, D. B. Williams, 1986, ISBN 0-031837-27-8
Volume 63-Computer-Based Microscopic Description of the Structure and Properties of
Materials, J. Broughton, W. Krakow, S. T. Pantelides, 1 8o, ISBN 0-031837-28-o
Volume 64-Cement-Based Composites: Strain Rate Effects on Fracture, S. Mindess,
S. P. Shah, 1986, ISBN 0-931837-29-4
Volume 65-Fly Ash and Coal Conversion By-Products: Characterization, Utilization and
0
Disposal 11,G. J.McCarthy, F.P. Glasser, D. M. Roy, 1 8o, ISBN 0-031837-30-8
Volume 66-Frontiers in Materials Education, L. W. Hobbs, C. L. Liedl, 108(,
ISBN 0-931837-31-b
Volume 67-Heteroepitaxy on Silicon, I.C.C. Fan, J. M. Poate, 198o, ISBN 0-031837-33-2
Volume 68-Plasma Processing, I. W. Coburn, R. A. Gottscho, D WI Hess, 1086,
ISBN 0-931837-34-0
Volume 69-MaterialsCharacterization, N. I.Cheung, M.-A Nicolet, 198,6, ISBN 0-031837-35-0
Volume 70-Materials Issues in Amorphous-Semiconductor Technology, D. Adler,
Y. Hamakawa, A. Madan, 1986, ISBN 0-Q31837-3o-7
Volume 71-Materials Issues in Silicon Integrated Circuit Processing, M. Wittmer,
). Stimme)l, M. Strathman, 1986, ISBN 0-931837-37-5
Volume 72-Electronic Packaging Materials Science II, K. A. Jackson, R. C. Pohanka,
D. R. Uhlmann, D. R. Ulrich, 1986, ISBN 0-931837-38-3
Volume 73-Better Ceramics Through Chemistry II, C. 1. Brinker, D E. Clark,
D. R. Ulrich, 1986, ISBN 0-931837-30-1
Volume 74-Beam-Solid Interactions and Transient Processes, M. 0. Thompson,
S. T. Picraux, J. S. Williams, 1987, ISBN 0-931837-40-5
MATERIALS RESEARCH SOCIETY SYMPOSIUM PROCEEDINGS
Volume 75-Photon, Beam and Plasma Stimulated Chemical Processes at Surfaces,
V. M. Donnelly, I. P. Herman, M. Hirose, 1987, ISBN 0-931837-41-3
Volume 76-Science and Technology of Microfabrication, R. E. Howard, E. L. Hu,
S. Namba, S. Pang, 1987, ISBN 0-931837-42-1
Volume 77-Interfaces, Superlattices, and Thin Films, J. D. Dow, I. K. Schuller, 1987,
ISBN 0-931837-56-1
Volume 78-Advances in Structural Ceramics, P. F. Becher, M. V. Swain, S. S6miya,
1987, ISBN 0-931837-43-X
Volume 79-Scattering, Deformation and Fracture in Polymers, G. D. Wignall, B. Crist,
T. P. Russell, E. L. Thomas, 1987. ISBN 0-931837-44-8
Volume 80-Science and Technology of Rapidly Quenched Alloys, M. Tenhover,
W. L. Johnson, L. E. Tanner, 1987, ISBN 0-931837-45-6
N. S. Stoloff,
Volume 81-High-Temperature Ordered Intermetallic Alloys, 11,
C. C. Koch, C. T. Liu, 0. Izumi, 1987, ISBN 0-931837-46-4
Volume 82-Characterization of Defects in Materials, R. W. Siegel, J. R. Weertman,
R. Sinclair, 1987, ISBN 0-931837-47-2
Volume 83-Physical and Chemical Properties of Thin Metal Overlayers and Alloy
Surfaces, D. M. Zehner, D. W. Goodman, 1987, ISBN 0-931837-48-0
Volume 84-Scientific Basis for Nuclear Waste Management X, J. K. Bates,
W. B. Seefeldt, 1987, ISBN 0-931837-49-9
Volume 85-Microstructural Development During the Hydration of Cement, L. Struble,
P. Brown, 1987, ISBN 0-931837-50-2
Volume 86-Fly Ash and Coal Conversion By-Products Characterization, Utilization and
Disposal Ill, G. J. McCarthy, F. P. Glasser, D. M. Roy, S. Diamond, 1987,
ISBN 0-931837-51-0
Volume 87-Materials Processing in the Reduced Gravity Environment of Space,
R. H. Doremus, P. C. Nordine, 1987, ISBN 0-931837-52-9
Volume 88-Optical Fiber Materials and Properties, S. R. Nagel, J. W. Fleming, G. Sigel,
D. A. Thompson, 1987, ISBN 0-931837-53-7
Volume 89-Diluted Magnetic (Semimagnetic) Semiconductors, R. L. Aggarwal,
J. K. Furdyna, S. von Molnar, 1987, ISBN 0-931837-54-5
Volume 90-Materials for Infrared Detectors and Sources, R. F. C. Farrow,
J. F. Schetzina, J. T. Cheung, 1987, ISBN 0-931837-55-3
Volume 91-Heteroepitaxy on Silicon II, J. C. C. Fan, J. M. Phillips, B.-Y. Tsaur, 1987,
ISBN 0-931837-58-8
Volume 92-Rapid Thermal Processing of Electronic Materials, S. R. Wilson,
R. A. Powell, D. E. Davies, 1987, ISBN 0-931837-59-6
Volume Q3-Materials Modification and Growth Using Ion Beams, U. Gibson,
A. E.White, P. P. Pronko, 1987, ISBN 0-931837-60-X
Volume Q4--Initial Stages of Epitaxial Growth, R. Hull, J. M. Gibson, David A. Smith,
1087, ISBN 0-931837-61-8
Volume 05-Amorphous Silicon Semiconductors-Pure and Hydrogenated, A. Madan,
M. Thompson, D. Adler, Y. Hamakawa, 1987, ISBN 0-931837-62-6
Volume Qo- Permanent Magnet Materials, S. G. Sankar, J. F. Herbst, N. C. Koon, 1987,
ISBN 0-031837-63-4
Volume 07 -Novel Refractory Semiconductors, D. Emin, T. Aselage, C. Wood, 1987,
ISBN 0-031837-b4-2
olum 08 I Plasma 'rocessing and Synthesis of Materials, D. Apelian, J. Szekely, 1987,
ISBN 0-031837-o5-0
MATERIALS RESEARCH SOCIETY SYMPOSIUM PROCEEDINGS
Volume 99-High-Temperature Superconductors, M. B. Brodsky, R. C. Dynes,
K. Kitazawa, H. L. Tuller, 1988, ISBN 0-931837-67-7
Volume 100-Fundamentals of Beam-Solid Interactions and Transient Thermal
Processing, M. J. Aziz, L. E. Rehn, B. Stritzker, 1988, ISBN 0-931837-68-5
Volume 101-Laser and Particle-Beam Chemical Processing for Microelectronics,
D.J. Ehrlich, G.S. Higashi, M.M. Oprysko, 1988, ISBN 0-931837-69-3
Volume 102-Epitaxy of Semiconductor Layered Structures, R. T. Tung, L. R. Dawson,
R. L. Gunshor, 1988, ISBN 0-931837-70-7
Volume 103-Multilayers: Synthesis, Properties, and Nonelectronic Applications,
T. W. Barbee Jr., F. Spaepen, L. Greer, 1988, ISBN 0-931837-71-5
Volume 104-Defects in Electronic Materials, M. Stavola, S. J. Pearton, G. Davies, 1988,
ISBN 0-931837-72-3
Volume 105-SiO, and Its Interfaces, G. Lucovsky, S. T. Pantelides, 1988,
ISBN 0-931837-73-1
Volume 106-Polysilicon Films and Interfaces, C.Y. Wong, C.V. Thompson, K-N. Tu,
1988, ISBN 0-931837-74-X
Volume 107-Silicon-on-Insulator and Buried Metals in Semiconductors, J. C. Sturm,
C. K. Chen, L. Pfeiffer, P. L. F. Hemment, 1988, ISBN 0-931837-75-8
Volume 108-Electronic Packaging Materials Science I, R. C. Sundahl, R. Jaccodine,
K. A. Jackson, 1988, ISBN 0-931837-76-6
Volume 109-Nonlinear Optical Properties of Polymers, A. J. Heeger, J. Orenstein,
D. R. Ulrich, 1988, ISBN 0-931837-77-4
Volume 110-Biomedical Materials and Devices, J. S. Hanker, B. L. Giammara, 1q88,
ISBN 0-931837-78-2
Volume 111-Microstructure and Properties of Catalysts, M. M. J. Treacy,
J. M. Thomas, J. M. White, 1988, ISBN 0-931837-79-0
Volume 112-Scientific Basis for Nuclear Waste Management XI, M J. Apted,
R. E. Westerman, 1988, ISBN 0-9'"337-80-4
Volume 113-Fly Ash and Coal Conversion By-Products: Characterization, Utilization,
and Disposal IV, G. J. McCarthy, D. M. Roy, F. P. Glasser,
R. T. Hemmings, 1988, ISBN 0-931837-81-2
Volume 114-Bonding in Cementitious Composites, S. Mindess, S. P. Shah, 1988,
ISBN 0-931837-82-0
Volume 115-Specimen Preparation for Transmission Electron Microscopy of Materials,
J. C. Bravman, R. Anderson, M. L. McDonald, 1988, ISBN 0-031837-83-0
Volume 116-Heteroepitaxy on Silicon: Fundamentals, Structuresand Devices,
H.K. Choi, H. Ishiwara, R. Hull, R.J. Nemanich, 1988, ISBN: 0-931837-8o-3
Volume 117-Process Diagnostics: Materials, Combustion, Fusion, K. Hays,
A.C. Eckbreth, G.A. Campbell, 1988, ISBN: 0-Q31837-87-1
Volume 118-Amorphous Silicon Technology, A. Madan, M.J. Thompson, P.C. Taylor,
P.G. LeComber, Y. Hamakawa, 1988, ISBN: 0-931837-88-X
Volume 119-Adhesion in Solids, D.M. Mattox, C. Batich, J.E.E. Baglin, R.I. Gottschall,
1988, ISBN: 0-931837-89-8
Volume 120-High-Temperature/High-Performance Composites, F.D. Lemkey,
A.G. Evans, S.G. Fishman, J.R. Strife, 1988, ISBN: 0-931837-90-1
Volume 121-Better Ceramics Through Chemistry Ill, C.I. Brinker, D.E. Clark,
D.R. Ulrich, 1988, ISBN: 0-931837-91-X
MATERIALS RESEARCH SOCIETY SYMPOSIUM PROCEEDINGS
Volume 122-Interfacial Structure, Properties, and Design, M.H. Yoo, W.A.T.Clark,
C.L. Briant, 1988, ISBN: 0-931837-92-8
Volume 123-Materials Issues in Art and Archaeology, E.V. Sayre, P. Vandiver,
J. Druzik, C. Stevenson, 1988, ISBN: 0-931837-93-6
Volume 124-Microwave-Processing of Materials, M.H. Brooks, I.J. Chabinsky,
W.H. Sutton, 1988, ISBN: 0-931837-94-4
Volume 125-Materials Stability and Environmental Degradation, A. Barkatt,
L.R. Smith, E. Verink, 1988, ISBN: 0-931837-95-2
Volume 126-Advanced Surface Processes for Optoelectronics, S. Bernasek,
T. Venkatesan, H. Temkin, 1988, ISBN: 0-931837-96-0
Volume 127-Scientific Basis for Nuclear Waste Management XII, W. Lutze,
R.C. Ewing, 1989, ISBN: 0-931837-97-9
Volume 128-Processing and Characterization of Materials Using Ion Beams, L.E. Rehn,
J. Greene, F.A. Smidt, 1989, ISBN: 1-55899-001-1
Volume 129-Laser and Particle-Beam Modification of Chemical Processes on Surfaces,
A.W.Johnson, GL. Loper, T.W. Sigmon, 1989, ISBN: 1-55899-002-X
Volume 130-Thin Films: Stresses and Mechanical Properties, J.C. Bravman, W.D. Nix,
D.M. Barnett, D.A. Smith, 1989, ISBN: 1-55899-003-8
Volume 131-Chemical Perspectives of Microelectronic Materials, M.E. Gross,
J. Jasinski, J.T. Yates, Jr., 1989, ISBN: 1-55899-004-6
Volume 132-Multicomponent Ultrafine Microstructures, L.E. McCandlish, B.H. Kear,
D.E. Polk, and R.W. Siegel, 1989, ISBN: 1-55899-005-4
Volume 133-High Temperature Ordered Intermetallic Alloys III, C.T. Liu, A.I. Taub,
N.S. Stoloff, C.C. Koch, 1989, ISBN: 1-55899-006-2
Volume 134-The Materials Science and Engineering of Rigid-Rod Polymers,
W.W. Adams, R.K. Eby, D.E. McLemore, 1989, ISBN: 1-55899-007-0
Volume 135-Solid State lonics, G. Nazri, R.A. Huggins, D.F. Shriver, 1989,
ISBN: 1-55899-008-9
Volume 136-Fly Ash and Coal Conversion By-Products: Characterization, Utilization
and Disposal V, R.T. Hemmings, E.E. Berry, G.J. McCarthy, F.P. Glasser,
1989, ISBN: 1-55899-009-7
Volume 137-Pore Structure and Permeability of Cementitious Materials, L.R. Roberts,
J.P. Skalny, 1989, ISBN: i-55899-010-0
Volume 138-Characterization of the Structure and Chemistry of Defects in Materials,
B.C. Larson, M. Ruhle, D.N. Seidman, 1989, ISBN: 1-55899-011-9
Volume 139-High Resolution Microscopy of Materials, W. Krakow, F.A. Ponce,
D.J. Smith, 1989, ISBN: 1-55899-012-7
Volume 140-New Materials Approaches to Tribology: Theory and Applications,
L.E. Pope, L. Fehrenbacher, W.O. Winer, 1989, ISBN: 1-55899-013-5
Volume 141 -Atomic Scale Calculations in Materials Science, J. Tersoff, D. Vanderbilt,
V. Vitek, 1989, ISBN: 1-55899-014-3
Volume 142-Nondestructive Monitoring of Materials Properties, 1. Holbrook,
1 Bussiere, 1989, ISBN: 1-55899-015-1
Volume 143-Synchrotron Radiation in Materials Research, R. Clarke, J. Gland,
I H. Weaver, 1989, ISBN: 1-55899-016-X
Volume 144-Advances in Materials, Processing and Devices in III-V Compound
Semiconductors, D.K. Sadana, L. Eastman, R. Dupuis, 1989,
ISBN: 1-55899-017-8
R
'ei
Matera/
A irnl Resrar, h Sx
Proceedings listed in thrfront.
MATERIALS RESEARCH SOCIETY CONFERENCE PROCEEDINGS
Tungsten and Other Refractory Metals for VLSI Applications, Robert S. Blewer, 1986;
ISSN 0886-7860; ISBN 0-931837-32-4
Tungsten and Other Refractory Metals for VLSI Applications II, Eliot K. Broadbent,
1987; ISSN 0886-7860; ISBN 0-931837-66-9
Ternaryand Multinary Compounds, Satyen K. Deb, Alex Zunger, 1987; ISBN 0-931837-57-X
Tungsten and Other Refractory Metals for VLSI Applications III, Victor A. Wells, 1988;
ISSN 0886-7860; ISBN 0-931837-84-7
Atomic and Molecular Processing of Electronic and Ceramic Materials: Preparation,
Characterization and Properties, Ilhan A. Aksay, Gary L. McVay, Thomas G. Stoebe,
J.F. Wager, 1988; ISBN 0-931837-85-5
Materials Futures: Strategies and Opportunities, R. Byron Pipes, U.S. Organizing
Committee, Rune Lagneborg, Swedish Organizing Committee, 1988: ISBN 1-55899-000-3
Tungsten and Other Refractory Metals for VLSI Applications IV, Robert S. Blewer,
Carol M. McConica, 1989; ISSN 0886-7860; ISBN 0-931837-98-7
Tungsten and Other Advanced Metals for VLSI/ULSI Applications V, S. Simon Wong,
Seijiro Furukawa, 1990; ISSN 1048-0854; ISBN 1-55899-086-2
High Energy and Heavy Ion Beams in Materials Analysis, Joseph R. Tesmer, Carl J.
Maggiore, Michael Nastasi, J.Charles Barbour, James W. Mayer, 1990; ISBN 1-55899091-7
Physical Metallurgy of Cast Iron IV, Goro Ohira, Takaji Kusakawa, Eisuke Niyama,
1990; ISBN 1-55899-090-9
4
I
mm, m
I