Chemical Energy Storage
Chemical Energy Storage
Chemical Energy Storage
Also of Interest
Biorenery
From Biomass to Chemicals and Fuels
Michele Aresta, Angela Dibenedetto, Franck Dumeignil (Eds.), 2012
ISBN 978-3-11-026023-6, e-ISBN 978-3-11-026028-1
Nanocarbon-Inorganic Hybrids
Next Generation Composites for Sustainable Energy Applications
Dominik Eder, Robert Schlgl (Eds.), 2013
ISBN 978-3-11-026971- 0, e-ISBN 978-3-11-026986-4
Microalgal Biotechnology:
Potential and Production
Clemens Posten, Christian Walter (Eds.), 2012
ISBN 978-3-11-022501-3, e-ISBN 978-3-11-022502-0
Microalgal Biotechnology:
Integration and Economy
Clemens Posten, Christian Walter (Eds.), 2012
ISBN 978-3-11-029827-7, e-ISBN 978-3-11-029832-1
Green
The International Journal of Sustainable Energy Conversion and Storage
Martin Stutzmann (Editor-in-Chief )
ISSN 1869-876X, e-ISSN 1869-8778
Chemical Energy
Storage
Edited by Robert Schlogl
DE GRUYTER
Editor
Prof. Dr. Robert Schlgl
Fritz Haber Institute of the Max Planck Society
Department of Inorganic Chemistry
Faradayweg 4-6
14195 Berlin
Germany
acsek@fhi-berlin.mpg.de
ISBN 978-3-11-026407-4
e-ISBN 978-3-11-026632-0
Library of Congress Cataloging-in-Publication Data
A CIP catalog record for this book has been applied for at the Library of Congress.
Bibliographic information published by the Deutsche Nationalbibliothek
The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliograe;
detailed bibliographic data are available in the Internet at http://dnb.dnb.de.
2013 Walter de Gruyter GmbH, Berlin/Boston.
The citation of registered names, trade names, trade marks, etc. in this work does not imply, even in
the absence of a specic statement, that such names are exempt from laws and regulations protecting
trade marks etc. and therefore free for general use.
Typesetting: Apex CoVantage, LLC, Herndon, Virginia, USA
Printing and binding: Hubert & Co. GmbH & Co. KG, Gttingen
Cover image: Istockphoto/Thinkstock
Printed on acid-free paper
Printed in Germany
www.degruyter.com
Contents
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
xv
1.1
1.1.1
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.2
1.1.3
1.1.4
16
1.1.5
17
1.1.6
20
1.1.7
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
30
1.2
35
1.2.1
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
35
1.2.2
General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
35
1.2.3
37
1.2.4
Grid-Scale Storage of Electrical Energy . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.4.1 Storage on the Transmission Grid Scale . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.4.2 Storage on Distribution and Medium-Voltage Grid Scale . . . . . . . . . . . . . . .
39
40
43
1.2.5
Energy Storage for Mobile Applications . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.5.1 Chemical Compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.5.2 Traction Batteries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
44
45
46
1.2.6
Systems Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
47
1.3
49
1.3.1
Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
57
1.3.2
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
57
2.1
59
2.1.1
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
59
vi
Contents
2.1.2
Sources of Biomass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
59
2.1.3
Lignocellulose as Feedstock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
62
2.1.4
63
2.1.5
66
2.1.6
68
2.1.7
72
2.1.8
Controlled Transformations of Carbohydrates into Novel Biofuels . . . . . . . .
2.1.8.1 Transformations Based on LA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.8.2 Biofuel Compounds Based on 5-HMF . . . . . . . . . . . . . . . . . . . . . . . . . . . .
76
77
79
2.1.9
81
2.1.10
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
82
2.1.11
Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
82
2.1.12
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
82
2.2
87
2.2.1
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
87
2.2.2
2.2.2.1
2.2.2.2
2.2.2.3
Classication of Biomass .
Lignocellulose . . . . . . . .
Lipids . . . . . . . . . . . . . .
Proteins . . . . . . . . . . . . .
.
.
.
.
88
89
94
98
2.2.3
Selected Key Chemicals. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.3.1 Cellulose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.3.2 Glycerol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
98
98
99
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2.2.4
2.2.5
2.2.6
Outlook. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
2.2.7
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
2.3
2.3.1
Torrefaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
2.3.2
2.3.2.1
2.3.2.2
2.3.2.3
2.3.2.4
Pyrolysis . . . . . . . . . . .
Introduction. . . . . . . . .
Pyrolysis Reactors . . . .
Biomass. . . . . . . . . . . .
Composition of Bio-Oil .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
112
112
113
114
114
Contents
vii
Gasication . . . . . . . .
Introduction. . . . . . . .
Gasication Reactors .
Energy in Gasication.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
116
116
117
118
2.3.4
2.3.4.1
2.3.4.2
2.3.4.3
Combustion . . . . . . . .
Introduction. . . . . . . .
Energy in Combustion
Co-combustion . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
118
118
119
119
2.3.5
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
2.3.6
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
2.4
2.4.1
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
2.4.2
HTC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
2.4.2.1 HTC of Biomass Waste for Environmentally Friendly
Carbon Sequestration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
2.4.2.2 HTC for Carbon-Negative Materials. . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
2.4.3
Mineralized Biomass as Energy Carrier. . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
2.4.3.1 Biocoal and Its Comparison to Other Biofuels, Biogas and Bioethanol . . . . 129
2.4.3.2 Carbon Fuel Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
2.4.4
2.4.5
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
3.1
3.1.1
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
3.1.2
3.1.3
3.1.4
3.1.4.1
3.1.4.2
3.1.4.3
............
............
............
the Equilibrium
............
. . . . . . . . . . . 140
. . . . . . . . . . . 140
. . . . . . . . . . . 142
. . . . . . . . . . . 144
3.1.5
Concentration Dependence of E: The Nernst Equation. . . . . . . . . . . . . . . . . 145
3.1.5.1 The Nernst Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
3.1.5.2 Concentration Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
viii
Contents
3.1.6
3.1.7
Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
3.1.8
Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
3.1.9
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
3.2
3.2.1
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
3.2.2
Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
3.2.3
3.2.4
3.2.5
3.2.5.1
3.2.5.2
3.2.5.3
3.2.5.4
3.2.6
Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
3.2.7
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
3.3
3.3.1
3.3.2
3.3.3
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3.3.4
The Electrocatalysis of Oxygen Reduction at Fuel Cell Cathodes
3.3.4.1 Understanding the Electrode Potential Dependence of the ORR .
3.3.4.2 Understanding and Predicting Trends in ORR Activity
on Transition-Metal Catalysts . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.4.3 Nanostructured Pt Core-Shell Electrocatalysts for the ORR . . . .
3.3.4.4 Noble-Metal-Free ORR PEMFC Electrocatalysts . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
154
155
156
158
159
. . . . . . . . . 173
. . . . . . . . . 173
. . . . . . . . . 174
. . . . . . . . . 177
. . . . . . . . . 182
3.3.5
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
3.3.6
Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
3.3.7
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
3.4
3.4.1
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Contents
ix
3.4.2
3.4.2.1
3.4.2.2
3.4.2.3
3.4.2.4
3.4.2.5
3.4.2.6
3.4.2.7
Water Oxidation. . . . . . . . . . . . . . . . . . . . . . . . . . . .
PSII . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Geometric Structure of the WOC . . . . . . . . . . . . . . . .
Electronic Structure of the WOC . . . . . . . . . . . . . . . .
Function of the WOC . . . . . . . . . . . . . . . . . . . . . . . .
Suggested Mechanisms of OO Bond Formation . . . . .
Summary: Principles of Photosynthetic Water Splitting.
Current Water-Splitting Catalysts . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
187
187
190
192
194
195
197
198
3.4.3
3.4.3.1
3.4.3.2
3.4.3.3
3.4.3.4
3.4.3.5
3.4.3.6
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
199
200
200
203
209
210
211
3.4.4
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
3.4.5
Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
3.4.6
Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
3.4.7
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
3.5
3.5.1
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
3.5.2
3.5.3
3.5.4
3.5.5
3.5.6
3.5.7
Outlook. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
3.5.8
Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
3.5.9
Note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
3.5.10
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
4.1
4.1.1
4.1.1.1
4.1.1.2
4.1.1.3
Theory . . . . . . . . . . . . . . . . .
Introduction. . . . . . . . . . . . . .
Course of a Catalytic Reaction
Reaction Kinetics . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
249
249
249
251
Contents
4.1.2
4.1.2.1
4.1.2.2
4.1.2.3
4.1.2.4
Practical Aspects . . . . . . . . . . .
Laboratory Reactors . . . . . . . . .
Preliminary Tests . . . . . . . . . . .
Comparative Studies . . . . . . . . .
Development of Kinetic Models.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
258
258
258
259
260
4.1.3
4.1.3.1
4.1.3.2
4.1.3.3
Examples . . . . . . . . . . . . . . . . .
Oxidative Coupling of Methane .
Decomposition of Ammonia . . .
Slurry Reaction . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
264
264
267
270
4.1.4
Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
4.1.5
Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
4.1.6
Abbreviations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
4.1.7
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
4.2
4.2.1
4.2.2
4.2.3
4.2.4
4.2.5
Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
4.2.6
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
4.3
4.3.1
4.3.2
4.3.3
4.3.4
4.3.5
4.3.6
Contents
xi
4.3.7
4.3.8
Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
4.3.9
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
4.4
4.4.1
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
4.4.2
4.4.3
4.4.4
4.4.5
4.4.6
Synopsis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
4.4.7
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
4.5
4.5.1
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
4.5.2
4.5.3
4.5.3.1
4.5.3.2
4.5.3.3
4.5.3.4
4.5.3.5
4.5.4
4.5.5
Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
4.5.6
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
5.1
5.1.1
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
5.1.2
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
357
357
360
364
368
370
xii
Contents
5.1.3
5.1.4
5.1.5
5.1.6
5.1.7
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
5.1.8
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
5.2
5.2.1
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
5.2.2
5.2.3
Solarthermal Cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
5.2.3.1 General Principles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
5.2.3.2 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
5.2.4
5.2.5
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411
5.2.6
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411
5.3
5.3.1
5.3.2
5.3.3
5.3.4
5.3.5
5.3.6
5.3.7
5.3.8
5.3.9
Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 435
5.3.10
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 435
Contents
xiii
5.4
5.4.1
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443
5.4.2
5.4.3
Applications of Synthesis Gas: H2 and Methanol. . . . . . . . . . . . . . . . . . . . . 445
5.4.3.1 Syngas to Hydrogen: The WGS Reaction . . . . . . . . . . . . . . . . . . . . . . . . . . 445
5.4.3.2 Syngas to Methanol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 446
5.4.4
Syngas to Synthetic Fuels: The Fischer-Tropsch Synthesis. . . . . . . . . . . . . . . 446
5.4.4.1 Chemistry and Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
5.4.5
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 455
Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
459
Author Index
xvi
Author Index
Author Index
xvii
xviii
Author Index
Author Index
xix
xx
Author Index
Dr. Shengfa Ye
Max-Planck Institute for
Chemical Energy Conversion
Department of Molecular Theory and
Spectroscopy
Stiftstrae 34-36
45470 Mlheim an der Ruhr
Germany
shengfa.ye@cec.mpg.de
Chapter 4.5
1.1.1 Introduction
The energy system is a fundamental characteristic of a human society. All its individual and collective actions require the availability of free energy in various forms
(mechanical, chemical, electrical). The sum of all man-made inorganic energy interconversion processes reaches at the end of the rst decade in the twenty-rst century
the same dimension [1] as the biological processes needed to generate the food for
humankind. Both process groups amount to about 15% of all biological solar energy
conversion on the planet Earth. It is obvious that we are dealing with enormous dimensions that are difcult to comprehend, as they are so much larger than our individual frame of dimensions. In the following, we discuss a few aspects of these
dimensions using the example of the German energy system as the necessary data
are well accessible [2] and its absolute size of approximately 1% of the global energy
system is in suitable dimension.
The German Energiewende in the aftermath of the Fukushima events is a highly
popularized impulse [3] on the future of an energy supply system. It comprises the
termination of nuclear power generation within 10 years leading to the necessity to
replace about 25% of the German power consumption of about 600 TWh by alternative sources. This by itself was not a difcult task as within 1 month after the decision the contribution of nuclear power fell from about 420 GWh on 02.01.2011 to
below 100 GWh on 05.23.2011 (day of nuclear moratorium 03.15.2011) (source:
BDEW, Auswirkungen des Moratoriums). The replacement came from extended
use of fossil fuels and to a minor extent from a switch of net power exports to
power imports much in the same way as in Japan the fast phasing out of nuclear
power within one service cycle of 13 months caused an increase in fossil fuel consumption. In Germany this trend leads to increased emission of 60100 million
t CO2 per annum depending on the fuel mix between coal and gas.
The Energiewende is thus partly in conict with the German national plan for a
change of the entire energy system issued in 2010 [4]. In this plan it is stated that
the total German CO2 emission should fall between 2010 and 2035 from about
1,000 million tons to 265 million tons. Electrical power generation, the subject of
the Energiewende, has a 50% share on the total emission. This should be achieved
by a massive implementation of renewable energy and a smaller contribution from
increased energy efciency through various measures but mainly by strong reduction
of house heating following improved building insulation. In terms of relative contributions to primary energy carriers the situation is expected to develop as shown in
Figure 1.1.1.
100
% contribution
80
Renewables
Nuclear
60
Gas
Coal
Oil
40
20
0
2000
2010
2022
2050
Year
Figure 1.1.1 Relative total primary energy input into the German energy system. The values
in 2022 relate to the completion of the Energiewende, the data for 2050 for the completion of
the National Energy Concept. Source: BMWI [2].
It is obvious from Figure 1.1.1 that achieving the ambitious goal of the National
Energy Concept will require a deep-reaching change of the whole energy system and
not only the replacement of nuclear energy in a given system. This qualies the Energiewende indeed as an initial step rather than the great change in the energy system.
The values for the year 2022 indicate that the political concept assumes a combination of three factors for replacing the nuclear energy: increased energy efciency in
parallel with more renewable energy generation and some additional power generation through gas combustion. In the following 35 years, the contribution of renewable energy is expected to rapidly increase and to replace all oil and most coal.
In order to estimate the dimension of changes necessary, it is instructive to study
the distribution of absolute primary energy consumption in Germany as indicated in
Figure 1.1.2.
The energy conversion losses are by far the largest sink of primary energy followed
by three almost equal areas of industrial production, mobility and transport and
domestic energy consumption. Distribution of goods and services and the use of fossil carriers as chemical feedstock are slightly minor applications. The temporal evolution shows some changes following the industrial change with the unication in
1989 and a change in lifestyle following this event. We see very little energy saving
or effects of energy efciency increase that may well be hidden by rebound effects
of increased mobility and more spacious housing. The apparent contradiction
between assumed increased energy efciency also felt by increasing changes toward
a more energy-efcient lifestyle and the data in Figure 1.1.2 becomes clearer when
the correlation between the German gross domestic product (GDP) and the CO2 emission (being proportional to the fossil energy input) is considered. In Figure 1.1.3 it is
1.1.1 Introduction
5000
4500
1990
2000
2010
Primary Energy PJ
4000
3500
3000
2500
2000
1500
1000
500
al
Di
Ch
em
ic
n
st
rib
m
Do
ut
io
es
tic
ty
ob
ili
tio
uc
od
Pr
Lo
s
se
s
Application
Figure 1.1.2
Distribution of primary energy streams into key application areas for Germany.
Source: BMWI database 2011 [2].
1
0.95
0.9
0.85
GDP
0.8
Production
0.75
0.7
0.8
0.9
Figure 1.1.3 Relation between CO2 emission of Germany and its economic activity. The nor-
malized values are given for the whole GDP and for the production sector only. Source:
BMWI database 2011 [2].
shown that the total GDP inversely scales with CO2 emission supporting apparently
the energy saving trend. If, however, one correlates the total CO2 emission with the
changes in the production sector one nds the expected positive trend: more production means more primary energy use and hence more CO2 emission. The trend inversion in Figure 1.1.3 is a clear sign of the strong contribution of nonproduction
economic activity to the GDP.
These coarse considerations indicate that energy systems [57] are interconnected
in complex manners with multiple nonscientic aspects of utilization, generation
and distribution. Simple extrapolations of trends into evolutionary targets for individual technologies contributing to an energy system are dangerous and misleading.
It is one role of politics to dene global evolutionary targets for an energy system
such as to be compatible with climate protection targets [8] or to refrain from nuclear
energy options. Great care should be exerted when these macroscopic targets are underlined with microscopic targets delineating development avenues for individual
and innovative technologies. On this level, insufcient fundamental or technical
information may lead to inadequate priorizations of technical options and generally
precludes grassroots innovation that may lead to completely unexpected solutions.
The evolution of biofuels [1, 9, 10] in different parts of the world are such an example as well as the quite specic expectations of the German energy concept that
were overruled within a few months by the exit decision from nuclear power. Unrealistic high expectations in the energy saving potentials of a society without severe
modications of economic [11] or behavioral boundary conditions may also prove
detrimental.
At this point we try dening the term energy system [1, 1214]. From the perspective
of a scientist or engineer the energy supply chain from resources via distribution to
application is merely a sequence of process steps of energy conversions. The free energy
stored in an energy carrier is converted to other energy carriers or to mechanical/
electrical energy. The challenge herby is to use a minimum number of steps to minimize
the losses of free energy to the universal heat sink and to likewise minimize such losses
within every step. We accept that it is impossible to convert energy without losses and
that the usefulness of thermal energy is related to its difference in temperature with
respect to the environment of the process.
Within this realm a large number of processes were suggested to satisfy our
energy needs with few of them found their way to relevant technology. Nontechnological inuences served as selection pressures and have thus to be considered as
elements of the energy system in a more complete sense. The nonscientic elements
of an energy system are depicted with the here relevant scientic/technical elements in
Figure 1.1.4.
It becomes evident that any meaningful treatment of science and technology for
the energy system must be in contact with the nonscientic aspects. Science and technology further must transport its insights to those responsible for decision-making.
This is clearly true for the aspects of implementing ready technologies. It is required
in addition already at the earliest stages of fundamental scientic knowledge in order
to account for the request of modern societies to be knowledge societies in which
all members participate in detailed decisions on the functioning of the society. As
energy supply is the cross-sectional enabling basis of all individual and collective activities of mankind it may be concluded that the evolution of the energy system is one
of the central activities of organizing a society.
Keeping this broad view it is the intention of the present book to cover a tiny
fraction of the scientic ground necessary for converting the present fossil-based
energy system into one where regenerative primary energy is the main source of
Society
Politics
Production
Application
Science technology
Regenerative
resources
Economics
Distribution
Behaviour
Regulations
Fossil
resources
Residues
Finance
Figure 1.1.4 Scientic and nonscientic elements of the energy system. Multiple interfaces
and control loops exist between the key elements. Science and technology are the enabling elements for all energy processes and additionally serve the important purpose of informing decision makers about necessary regulatory and behavioral boundary conditions.
free energy. The following text thus concentrates not on recommendations of how
to perform the Energiewende. It rather focuses on the contribution that chemistry
can make to the evolution of the energy system. This contribution is twofold acting
either on energy efciency strategies of existing processes or providing solutions for
the energy storage challenge being an essential ingredient into regenerative energy
systems.
Sustainable
production
Materials for
recovery,
electronics,
lighting
Conservation
Coal combustion
CO2 sequestration
Petrochemistry
Fossil
Chemical
feedstock
Materials
under extreme
conditions
Energy
Issues
Nuclear
Nuclear waste
chemistry
Chemical
storage,
hydrogen
Renewable
Batteries
fuel cells
Syn fuels
Chemical
feedstock
Process
Material
Solar energy
harvesting
Figure 1.1.5 Chemistry is a platform science in the energy challenge. All major areas of
energy conversion and use require materials or processes provided by chemistry.
Energy (kJ/mole)
state equals to a maximum conversion level that cannot be exceeded at given conditions. An example of such a system is the synthesis of ammonia [19], which is at
technical reaction conditions heavily limited by equilibrium conversion.
The theory of chemical kinetics describes reactions in a multidimensional landscape of hills and valleys given by energy as function of the atom coordinates of
the reacting system. A simplied energy prole occurs when the energy change is
mapped against the reaction coordinate describing in single number the changes of
atom coordinates of the reacting system. For a very simple reaction of the formation
or cleavage of a single bond between two entities (atoms, molecular fragments), the
resulting prole may look as shown in Figure 1.1.6.
The process in Figure 1.1.6 is called an elementary step as it refers to the distance
change of one single chemical bond. The energy level of the nonreacted components
(educts) is dened as relative zero point of the energy scale (dashed line in Figure
1.1.6). We see that approximation of the reactants rst requires energy (the activation
energy Ea) as the electron clouds are repelling the bodies. At a critical distance called
the bond distance, the electronic systems interact quantum chemically and form a
chemical bond whereby the bond energy is liberated. If we try to compress the system
further, the total energy is rapidly increasing as the repulsion of the core electron systems of the atoms begins to dominate. The quantity H is the enthalpy change of the
process that is related through the second law of thermodynamics to the free energy
change of the reaction.
The activation energy can be interpreted as resistance of the system to undergo a
reaction; the enthalpy change is related to the extent to which the reaction can occur.
In the present context, it is important to note that chemical reactions that generate
free energy from reacting energy carriers with oxygen (combustion) occur with
Ea
Reaction coordinate
Figure 1.1.6
Reaction prole of a single elementary step. The thin lines between the energy
levels are guides to the eye and do not indicate the actual course of the changes of atomic
coordinates.
large free enthalpy changes and thus always run to completion. This is unfortunately
not the case for energy storing reactions in which CO2 is reacted with a primary
energy carrier such as hydrogen to form an energy storage molecule. These reactions
that require the addition of energy both in the form of heat ux from the environment and in the form of chemical energy of the primary energy storage molecule
are strongly limited by equilibrium conversions and require special measures such
as removal of the product from the reaction system in order to run to signicant conversions. This is illustrated in Figure 1.1.7 for the relevant example of hydrogenation
of CO2 to methanol [20]:
CO2 + 3 H2 CH3OH + H2O
This reaction can also be considered to occur from a rst step of CO2 reduction to
CO followed by the methanol formation from CO being the more facile reaction:
CO2 + H2 CO + H2O
CO + 2 H2 CH3OH
Figure 1.1.7 indicates the maximal methanol yield as function of temperature for
various mixtures of CO2/CO at 50 bar pressure. We see that indeed severe limitations
exist for CO2 hydrogenation unless extremely efcient catalysts are found that
can operate at low temperatures. Such systems are presently unknown [2123] as
we typically operate at 523 K.
100%
CO2
90%
10:1
80%
5:1
2:1
Methanol Yield
70%
1:1
60%
1:2
50%
1:10
CO
40%
30%
20%
10%
0%
400
450
500
T(K)
550
600
Figure 1.1.7 Equilibrium calculation for the hydrogenation of CO2 to methanol at 50 bar
pressure. The parameter is the fraction of CO in CO2 facilitating the hydrogenation.
10
Desired activation
Catalytic activation
Undesired activation
Activation
Educts
Energy (kJ/mole)
Without going into the details of thermodynamics here, it should be noted that for
chemical energy storage reactions the enthalpy change is not the relevant quantity as
we are interested in the amount of work that can be done with a given amount of chemical
energy storage material. This useful work is dened by the free energy and realistically
by the exergy [1] of a process that relates the free energy to the actual process conditions.
The free energy is independent from these conditions and thus the relevant quantity for
scientic discussions; the exergy is clearly related to specic realizations of a process.
Catalysis is changing the course of a reaction by modifying the path of the system
through the energy hypersurface (the energy prole as function of all relevant atomic
coordinates). The elementary step reactions change under the inuence of a catalyst
but not the overall energetic difference between educts and products. This is equivalent to stating that a catalyst changes the velocity of a chemical reaction but not the
maximal extent to which it can occur.
The statement is easy to understand when the reaction has only one product (or set
of products). Typical examples are combustion reactions [24, 25] where CO2 and
water are always the nal products. As molecules with many atoms (such as hydrocarbons) cannot combust in a single elementary process (not even methane), there is
always the chance that several pathways may lead to several products even when
less than the maximal energy is gained from the process. A typical example is the
formation of CO as a toxic by-product of combustion.
The prototypical situation is shown in Figure 1.1.8. Here we see that an initial
molecular mix has several possibilities to react. In synthetic chemistry, we are usually
Target
Common
intermediate
Reaction coordinate
Figure 1.1.8 Energy prole of reaction with two product pathways and a common intermediate as primary activation product. Common to both reactions is further the stable leaving
group (green). A catalytic pathway (red) to the target product (blue) circumvents the undesired
unselective path to the most stable reaction product.
11
interested in target molecules that are intermediate [2629] in energy with respect to
the thermodynamically most favorable reaction product of an educt mix. An example would be the formation of an olen target molecule from an alkane feedstock
molecule accessible by oxidative dehydrogenation with oxygen. The driving force
of such a reaction would be production of the stable leaving group water as a second
product besides the target molecule. The undesired reaction would be the combustion
of the alkane to CO2 (and water).
In practice, the reaction starts by an initial activation of the educt that may represent a high hurdle [30] as the educt is a stable molecule. The system then nds a low
barrier to the undesired reaction and a higher barrier to the desired one that looses
kinetically against the faster undesired reaction. The result is a low selectivity to the
desired product. Catalysis can now interfere (see Figure 1.1.7) by changing the reaction coordinate of the activated intermediate [31] such that the easy pathway to the
undesired process is less accessible for only a small extra cost in activation energy.
Then the process will slow down in velocity but gain substantially in selectivity.
The secret of the action of the catalyst is that by forming an activated complex
with the activated reactant it changes the molecular structure of the reactant such
that the undesired easy reaction path becomes unattractive. It is the specicity of
the chemical bonding between substrate and catalyst that exerts the benecial effect.
In many catalytic processes, an additional effect greatly changes the reaction rate: If
more than one reactant is necessary, then not only atoms but also electrons and
atomic fragments need to be exchanged. A catalyst can provide the separation of
these elementary steps in space and time. It can rst activate one reactant and store
the activated species at a reaction site for later use with the other reagent, and it
can lend and receive back redox equivalents (electrons) to participating species [32].
In this case, the catalyst changes profoundly its own structure, and it must be designed
such that it can regain its initial structure when the catalytic turnover is completed.
A catalyst is thus a functional material that undergoes cyclic changes in its structure upon interaction with reactants. The design of catalysts thus has two targets:
1. It must ensure the specic modication of the reactant such that the desired reaction occurs with maximal selectivity and rate.
2. It must be exible enough that the structure of the catalyst can undergo reversible
changes ensuring that it is not consumed in the desired reaction. This stability criterion is additionally made more difcult as also the reaction products must not
strongly bind to the catalyst and thus poison it by irreversible interactions.
A central concept for the structure of catalysts is that of an active site. This active
site is an ensemble of atoms guaranteeing the desired geometry of the reactant-catalyst
bonding by a highly specic local arrangement of the constituting atoms. In order
to ensure the regeneration of the site at the end of the catalytic cycle, it is desirable
that the active sites are embedded in a stabilizing matrix and that the embedment
allows for geometric exibility in order to minimize the activation energy for the
rearrangement of the atoms in the active site.
This is either realized in molecular species where pocket structures of ligands
around an ensemble of active atoms enable both specicity and exibility. In solids
special high-energy sites are terminating a matrix of an active phase providing by
12
its rigid chemical and geometric structure directing interactions for the reactant binding (such as in pores or at steps of a solid). Examples of both strategies can be found
in chapters 3.4 and 4.4.
A possible catalytic cycle for the process of Figure 1.1.8 is shown in Figure 1.1.9. A
typical reaction would be
A2 + B C + D
where C is the target product and D the stable leaving group. For a selective oxidation, A2 would be oxygen, and D would then be water. A possible chemical realization of an active site could be based on vanadium as the active element. The active
site [31, 33] would be
[V2(Ot)2(Obr)(Osup)2]2[O2]
assuming that the active metal changes its oxidation state by one formal unit per catalytic cycle. The various oxygen ligands are Ot, terminal oxygen; Obr, bridging; and
Osup, supporting bridges to the active matrix. It becomes clear that active sites are in
general not single atoms or terminating regular entities of active phases. They are
rather special structures and often form only during catalytic operation.
A schematic representation for the energy-relevant process of methane combustion
[34, 35] is given in Figure 1.1.10. Here a projection of the energy barriers on a selection of reaction coordinates is shown. The map can be read like a topographic map.
The circular structures represent maxima (peaks) in the energy landscape, whereas
the arrows indicate pathways through saddle points (valleys) of the energy structure.
We see immediately that two alternative reaction pathways are possible with one
Reactant
A
Reactant
A
Active site
Deactivated
site
O
V
R
R
Products
C and D
Reactions
Site plus
activated
educt A
O
O
V
V
O
R
R
Reactant
B
Activated
reaction
complex
Figure 1.1.9 The catalytic cycle illustrates the steps of structural modication of a hypothetical active site represented in the center.
13
Energy
leading through decomposition of formaldehyde to the CO product. The other channel leads via complete dehydrogenation to deep combustion. The dashed lines indicate minor reaction paths via less deep saddle points. It becomes apparent how the
facile combustion reaction is a complex process as soon as elementary steps are considered. It is also clear those catalysts such as Pd or Pt that strongly favor the dehydrogenation path [36, 37] will reduce the selectivity to CO. Finally, it is obvious
that these catalysts carry the risk for deactivation when the intermediate C atoms
polymerize into soot covering the active surface.
The understanding of catalysis at a level of complexity necessary to analyze the
course of a chemical reaction is only achievable through the progress in studying
model systems both with experimental [15, 38] and theoretical [39, 40] methods. The
model approach was developed with structurally simple single-crystal surfaces [41]
allowing a comprehensive treatment of elementary step reactions. Most of these reactions were, however, adsorption reactions, and only very limited reaction steps
such as dissociation reactions were possible. Introducing steps as high-energy sites
improved the situation considerably as now reactions such as CO oxidation and
ammonia synthesis became amenable to rigorous descriptions. More complex reactions are only recently under study by model systems [42, 43] since a combination
of chemical complex oxide support systems with single-crystalline quality and their
decoration with nanoparticles from metals or oxides became experimentally accessible. In combination with an increasing understanding of the structure of free nanoparticles [44, 45] of active phases, it is now possible to accurately describe complex
models and study elementary step reactions of kinetically demanding reactions
such as alkane activation and C1 chemistry being of relevance in chemical energy
conversion processes. Using the well-understood CO oxidation reaction [46] as
proxy and probe reaction, it becomes increasingly possible to extrapolate structural
at
n
di
or
on
CO
H2O
co
CO2
H2O
i
ct
a
Re
CH4
O2
CH2O
H2O
C
H2O
Reaction coordinate 2
Figure 1.1.10 Schematic representation of a possible energy surface for methane combustion.
The graphical impression is a projection of energy peaks onto a plane of reaction coordinates.
The reactant systems (clouds) are not given with stoichiometric accuracy. There are many
more intermediates and reaction pathways in the real gas combustion process.
14
Energy change
features from model systems to complex high-performance catalysts where the direct
structural analysis with the tools of surface science are not amenable for physical
methodical limitations and theoretical models are still of insufcient complexity.
With this knowledge we have now access to elementary reaction sequences of individual molecules. This is unfortunately still insufcient to understand or design performance catalytic processes. The reason lies in the fact that an observable amount of
a chemical (energy carrier) contains a very large number of individuals: 1 mL of
water contains 1022 molecules. One consequence of this large scaling is that the analysis of observable catalytic reaction includes to a large extent effects of transport of
molecules and of heat to and from the active sites of the catalyst. These transport
effects scale themselves in macroscopic (dimension of reactor) and mesoscopic (dimensions of boundary layers and pore systems) dimensions. Elementary questions
such as the comparison of catalysts for their performance good versus poor
and the determination of kinetic constants and parameters for comparison with
model or theoretical predictions and the determination of structure-function correlation [47] by using in situ analytical techniques all are hampered by the unavoidable
incorporation of transport effects. This multiscale nature of catalysis is reected in
the reaction chain indicated in Figure 1.1.11.
The form of the activated complexes [48, 49] is substantially different in homogeneous and heterogeneous reactions. In a homogeneous reaction, the catalyst forms
a molecular complex between the substrate that is recognized by its molecular
Time
Homogeneous
Macro
transport
Micro
transport
Activated
complex
Reaction
Activated
complex
Micro
transport
Macro
transport
Micro
kinetics
Macro
kinetics
Figure 1.1.11 Steps of heterogeneous reactions. The individual processes comprising sequences of elementary step reactions are linked to a process sequence. The microscopic part
is described by microkinetics, and the observable macroscopic performance by macrokinetics.
A typical relative dimension of energy changes associated with the individual steps is indicated.
In homogeneous reactions, the transport parts are often ignored.
15
shape and the single active site offering a binding site in the form of the replica of the
substrate molecule. This is extremely specic in biocatalysts called enzymes and still
highly specic in molecular catalysts where a limited set of ligand molecules denes
the electronic and geometric constraints of substrate binding. In heterogeneous systems, this process is less specic. Reactants are rst chemisorbed on a broad range of
surface sites. This interaction is weak with respect to the strength of chemical bonds
to be activated. Only in cases of very simple molecules such as H2 and N2 are the
approximations valid that all adsorption sites are equal on a surface and are not
affected by neighboring adsorbed species. The total surface may thus be considered
as landing sites for reactant molecules. They have to diffuse to special sites where
reactions of breaking and making chemical bonds can occur. These high-energy sites
[50, 51] are the few active sites discriminating any surface that may adsorb reactant
molecules from catalysts that can convert their adsorbates into products more than
once. The essentially at surface location of active sites greatly reduces the specicity
that can be executed by the surrounding atoms, and chemical selectivity of reaction
must be achieved only by exactly matching the electronic structure [52] of the reaction site to the molecular electronic structure of the adsorbate. As a result, the timing
of processes is critical in heterogeneous reactions as a means to induce specicity.
This leads to an extreme sensitivity of the catalytic performance to the details of
the electronic structure of the solid controlled by many types of defects referred to
as structure sensitivity [53, 54]. A complete description of a catalytic cycle can
thus only be achieved by understanding both [55] the microkinetic and macrokinetic
aspects of the reaction.
The relative changes of energy with the individual steps in Figure 1.1.10 should not
lead to the false assumption that steps with small energy changes are less relevant
than the chemical reaction bringing about the usually largest energy change. If the
molecules cannot access the active sites, they cannot perform their role. If the reaction energy cannot be distributed to the environment, side reactions can occur
and poison the active site or damage can occur for the whole catalyst through the
formation of hot spots.
Macrokinetics is the description and analysis of the performance of the functional
unit catalyst plus reagents plus reactor. It leads to formal activation barriers called
apparent activation parameter representing the superposition of several elementary barriers with transport barriers. It further delivers formal reaction orders and
rates as function of the process conditions. These data can be modeled with formal
mechanisms of varying complexity. In any case, these data can well describe the
system performance but cannot be used to deduce the reaction mechanism.
This task is left to the microkinetic modeling [56] trying to observe a catalytic
process unfolded from transport limitations. Here sequences of elementary steps
are composed of reaction mechanisms with chemically meaningful process steps.
These usually complex systems of equations with many unknown parameters are
then approximated to performance observations. It is obvious that multiple observations obtained by prole reactor studies and/or by kinetic isotope exchange reactions will greatly enhance the signicance of a microkinetic model adaptation. This
approach is presently the state of the art in understanding the relevant energy storage
reactions [57].
16
17
18
Storage
in sea and
minerals
H2O
Reduction of CO2
Heterotrophic
Foodstuff
animals,
humans
Dissipative
processes
Sunlight
Charge separation
in water
Biomass
synthesis
CO2
Fossil energy
Hydrocarbons
Synthesis
of carbohydrates
Dissipative
processes
Biomass
synthesis
Oxidation of
carbohydrates
Autotrophic
Photosynthesis
Humic acid
Waste
death
Figure 1.1.12 A crude systemic picture of the biogeochemical processes from which fossil
fuels derive a as nonoxidized storage system of nature.
19
Table 1.1 Selected amounts of carbon stored in natural compartments of the Earth.
Geosphere compartment
Deep sea
Surface sea
Stored carbon in Gt
13,800
1,020
Sea atmosphere
Fossil energy carriers
92
4,000
Combustion to atmosphere
5.5
Atmosphere
750
Vegetation
610
Photosynthesis
Soil
carbon in Gt/a
121
1,580
has to be produced by photosynthetic processes and related food chains. This energy
amounts to about 25 TW. If we add our demands for cellulosic biomass as feedstock
(wood, pulp) and consider that approximately 65% of the biomass production on our
planet occurs in water, but our demands for food and feedstock are largely related to
land-based organisms, then we see that the calculation gets tighter. If we add the
most critical but difcult-to-assess factor of ecosystem stability allowing only a fraction (about 30% without sustained damage, see sh and grasslands, for example) of
any ecosystem to be managed by humans, then we see quickly that we cannot rely by
and large on biological processes for converting energy from the sun in useable
energy carriers for technologies. In this estimation, we have further to consider
that the global population is growing fast, which requires rst more food and then
additional technical energy carriers.
Another issue is the required energy density. One metric ton of coal as 1 t Steinkohleeinheit (SKE) contains 7,000,000 kcal. One ton of dry glucose contains 3,720,000 kcal.
Biomass does not occur as dry glucose but rather as diluted and/or polymeric forms of
glucose. Several steps of energy-intensive concentration and rening are required until
we obtain glucose for technical applications. The relevant chapters 1.2 and 2.1 to 2.4 in
the book discusses potentials and difculties of this often overlooked aspect. This
is also true for the many concepts of using algae or other aquatic plants requiring
harvesting drying and rening before they can be used as energy carrier material.
The apparent initial success of biomass-derived energy carriers used in biofuels [1,
10] and fuel additives (E10) stems from a nonsustainable form of biomass harvesting.
These materials are generated from the fruit of plants. These fruits are developed to
carry a high-energy biomass such as starch or oil in order to support the reproduction
of the plant. In fact, only a small fraction of the total biomass of the plant is concentrated in the fruits. In addition, the extensive biocatalytic processes of biomass formation for these fruits (see Figure 1.1.12) are highly energy-intensive for the plant:
a large fraction of the effectively harvested solar energy is thus wasted in the biological processes generating the fruits for generating biofuels (of the rst generation).
This form of energy farming is suitable as niche solution in areas where an excessive
20
abundance of arable land is available in relation to the needs of the regional population. As a global solution it is inappropriate also due to many nonscientic
consequences of energy farming (see also Figure 1.1.4).
Large-scale application of energy farming further raises the question about sustainability of land use, water use, and availability of mineral fertilizer components
such as phosphate and soluble nitrogen. It is foreseeable that competitive situations
of these limiting resources on the planet may occur against food production. Even if
we solve the land use issue by appropriate land management (not yet practiced), the
other resources may prevent the use of energy farming as a main source of renewable
energy generation.
Whatever biomass may be spared from either keeping the ecosystems stable or from
feeding the global population, it may be used as feedstock for chemical processing
rather than as fuel for combustion processes. In this way, the complex structures nature
has generated during biomass formation (see Figure 1.1.12) can be partly preserved
and may be brought to use in material applications such as polymer formation or
direct synthesis of intermediates with complex molecular structures. In order to follow
this highly constructive use of biomass, it is essential to understand the underlying
chemistry of biomass transformation [7476] being largely different from our wellstudied hydrocarbon transformation. In the latter case, we need to enhance the functionality of hydrocarbon structures (gas and oil), whereas in biomass transformations
we need to defunctionalize overfunctionalized molecules leading upon activation to
polycondensation processes with unwanted polyoxo structures (such as humic acid).
An in-depth analysis of the potential of biomass to sustainably cure the CO2 emission problem arising from burning fossil resources (see Figure 1.1.12) revealed that
this is no permanent measure as compared to reducing the irreversible emission of
CO2 from fossil sources. Although terrestrial ecosystems can be managed to reduce
carbon emissions and increase carbon sink size signicantly, such increased carbon
uptake can offset fossil fuel emissions only temporarily on a timescale from decades
to a century. Terrestrial carbon sinks are thus best viewed as buying valuable time to
address the most signicant anthropogenic perturbation of the carbon cycle fossil
fuel emissions [77].
21
the dilute form (200 W/m2), the volatility of its ux at ground is the main obstacle. It
is impossible to sustain uninterrupted energy ux at demand if we use solar energy
directly converted into primary electricity. This approach is despite its volatility
highly efcient, as we need about 50% of our primary energy carriers for generating
electricity. If this comes without a CO2 bill and without cost for fossil fuel, it would
greatly relieve the energy system. It is also quite efcient: In Germany 1 ha land with
solar panels generates about 1 GWh electricity. If the same land was planted with
corn, then the energy equivalent from biomass to biogas conversion plus conversion
into electricity at demand would generate 0.15 GWh. If this this eld were to host a
wind turbine in addition, then an additional solar power of 2 GWh could be added. If
we assume that such a combi-energy-eld could generate 3 GWh/ha, in 1 year then
in Germany an area of about 2,000 km2 could generate all electrical power consumed
in Germany. This is all possible with existing technology. The investment would be
substantial [80, 81] with about 2.000/kW for solar and about half as much for wind
and biogas, but as solar energy is free and thus low operation cost would partly
compensate for this investment that is crudely estimated to be twice as high as the
investment in conventional fossil power stations with a mix of coal and gas.
The remaining fundamental problem is the volatility. Even when solar power and
wind power are combined and exchanged over the area of central Europe, this would
not be sufcient to guarantee the uninterrupted supply. Solar power has in Europe a
pronounced annual uctuation [82], and wind is quite unpredictable onshore and still
not always available offshore.
These trends are indicated with data from the German electricity production in Figures 1.1.13 and 1.1.14. The total contribution of regenerative power to the German
energy system was about 16% in 2011 including, however, a signicant fraction of thermal energy from waste incineration. In Figure 1.1.13 an episode in the spring season is
20
15
10
5.4
3.4
1.4
30.3
28.3
26.3
24.3
22.3
20.3
18.3
16.3
14.3
12.3
10.3
25
Date in 2011
Figure 1.1.13 Regenerative power in Germany from wind (blue) and photovoltaics (red).
This episode accounts in its maxima for about 25% of the electricity demand. Source: Data
from BDEW.
22
Electricity (TWh)
2.5
2
1.5
1
0.5
0
10
11
12
Figure 1.1.14 Annual change of PV electricity production in Germany for the year 2011.
These data change somewhat with the weather over the years. Source: Data from BSW solar.
shown with its contributions from wind and photovoltaic (PV) electricity. Comparing
to an average total demand of about 70 GW, we see that on certain days the contribution of regenerative power is already signicant. There are also periods with no sunshine and no wind where little contribution comes from regenerative sources. In many
instances, it is fortunate to combine PV with wind power to equalize the volatility. A
critical problem is the gradient of change being very fast on the timescale of compensating measures with fossil power plants. Here import/export helps, but quantitatively
larger uctuations would tend to destabilize the whole energy system. This issue gets
worse when the temporal resolution is increased: PV power has a strong day-night
cycle that is predictable, whereas wind power is completely unpredictable in shorter
timescales. These uctuation cause challenges for management of the grid, power stations, and possibly the demand situation. It limits the amount of renewable power that
can be tolerated in the grid with its present structure.
The issue of volatility has also a dimension on much longer timescales than days.
This is shown in Figure 1.1.14 for PV data of 2011 in Germany. A strong wintersummer trend is clearly visible stretching over almost one order of magnitude of
power change. This needs equilibration with storage measures that can hold energy
in large amounts (production of the total solar installations) for several months. It is
noted that the electricity production of PV in Germany now already exceeds the
storage capacity of the national hydropower stations.
The data clearly reveal that future energy systems relying to a large fraction on
regenerative primary energy will need substantial capacities for energy storage.
These storage systems need different response times from minutes to months and
are thus combinations of physical (short term) and chemical (long term) storage solutions. It is also clear that an intelligent management structure of the energy system
is needed deciding which storage option is used at what demand/supply ratio. The
sheer amount of energy that needs to be stored certainly cerates strong boundary
conditions on economics, efciency, and scalability of the solution.
23
The data in gure 1.1.13 indicate what kind of temporal structure is needed for the
chemical processes storing energy. Energy conversion systems such as electrolysis need
response times of several hours between full load and low load. As dimensionally stable electrodes are still a great challenge [82] it is a combination of distributed systems
switched in cascades and an operation scheme leaving partial load on the electrodes for
a maximum time that may solve the challenge. This randomly chosen example illustrates the complex requirements for energy storage systems many of which are at present not met with proven reliability causing research needs both with applied systems
and on the fundamental level to understand the adverse effects of load uctuations.
In general, the substantial contribution of volatile power sources will change the
operation scheme of the whole energy system as the traditional distribution in continuous base load and discontinuous peak load does not apply. All elements of the
energy system will need faster response times creating not only issues for the economic operation of large-scale installations and transmission capacities, but also
new demands on structural und functional materials having to cope with frequent
changes of operational loads.
The solution of these issues will benet from the availability of large-scale chemical
storage of electrical energy [18, 83] in articial solar fuels that can counteract the volatility of primary electricity in central and distributed installations. However, this has
a price tag [11, 13, 80, 84] arising from complexity and new optimization criteria. In
Figure 1.1.15, a crude schematics of the present energy system in Germany is shown
to illustrate the existing key components.
By and large, we generate our free energy from combustion of fossil fuels with
largely irreversible CO2 emission. The existing grid manages fossil electricity with
nuclear electricity in a common system and guarantees the uninterrupted supply
Hydro
Wind
PV
Fossil
fuels
Photosynthesis
Biomass
Nuclear
Grid
CO2
Combustion
Mech
storage
Electricity
Mobility
Transportation
Heating
Process heat
CHP
24
25
Solar radiation
PV, wind
electricity
Solar
fuels
Artificial
photosynthesis
Chemical
conversion
Biomass
Combustion
Fossil
fuels
Thermomechanical
conversion
Fusion
Smart
grid
Thermomechanical
storage
Electricity demand
Figure 1.1.16 A possible structure of the electricity part of a future energy system.
26
Wind
energy
CSP
photons
catalyst
H2O
CxHy
Photons
catalyst
H2O
Charge
seperation
H2O
electrolysis
CSP
H2O
MxOy
Proton
reduction
MxOyO
Mineral
carbon
Solar hydrogen
Biomass
Combustion
N2
CO
CH3OH
CnH2n+2
CnOmH2n+1+m
Sugars
Chemicals
NH3
CH4
CnH2n
27
Photons
CO2 pool
Gasoline
Solar fuels
Figure 1.1.17 The solar renery as the conceptual contribution of chemistry by chemical
energy conversion to the sustainable use of renewable energy. The upstream part (hydrogen
generation) and the downstream parts need not to be colocalized in a practical realization.
CSP stands for concentrated solar power. Green boxes indicate solar fuel products; blue
boxes stand for intermediate platform chemicals. The red arrows indicate ows of solar hydrogen to a storage and transport system for large-scale applications. The blue arrows show the
major application lines for chemical production of solar fuels. The scheme also indicates the
role of fertilizers from ammonia required in sustained use of biomass for energetic applications.
The carbon cycle of the energy system must be closed to ensure sustainability of
energy supply. This can be done by using biomass as source of CO2 or for stationary
sources such as gas power stations by collecting the CO2 from combustion of solar
fuels in the same way as for fossil fuels (carbon capture and use, CCU). The emissions
from distributed sources can only in limited dimensions be collected by biomass as argued in the previous text and in the respective chapters 1.2 and 2.4 of the book.
Using other storage options such as batteries or thermo-mechanical solutions for
small stationary applications should reduce CO2 emissions. The energy system can alternatively omit the use of carbon-based solar fuels by either working on hydrogen when it
is also generated in distributed systems or by using ammonia [90, 91] as a non-carbonbased energy carrier. It is expected that all these options will be realized, in particular
when we consider energy systems for other parts of the world than central Europe.
A large variety of options indicated in the upstream hydrogen generation section
reveal that much room exists for innovative chemical approaches for water splitting.
It must be clear, however, that the predominant part of these reactions must split
water under liberation of di-oxygen or chemically speaking by using reducing
28
equivalents from oxo anions. We are not interested in the oxygen molecule representing a waste like in photosynthesis, but we need the electrons carried by two oxo
anions:
2 H2O 2 OH + 2 H+ + 2e
2 OH O2 + 2 H+ + 2e
Many creative solutions circumventing this most difcult elementary step in hydrogen production can only be niche solutions as long as they work on sustainable other
sources of electrons from, for example, biomass debris. Many intelligent physicochemical approaches that use stoichiometric electron sources from nonrenewable
sources are not suitable solutions. If we plan to use other sources of electrons such
as reduced metal oxides, then we must make sure by life cycle analysis that such options are competitive with straight electrolysis including the use of material and the
efciency losses through the need to operate at high temperature levels. Many challenges of hydrogen generation would be relieved if we had available a source of nonsolar but sustainable high temperature. This is indicated in Figure 1.1.16 with the use
of fusion power helping generating hydrogen from gas-phase electrochemistry and
using the resulting intermediate temperature of a heat carrier for conventional thermomechanical power generation. Photochemical water splitting [9294] with oxide
or nitride photoelectrodes is a wide eld of chemical research. It qualies as articial
photosynthesis [9598] in the scheme of Figure 1.1.16 even when only free hydrogen would result. Despite enormous progress there is still much to do from the basic
understanding of coupling electronic excitation with catalytic surface chemistry to
reproducible and scalable production of photoactive materials and suitable reactors
allowing separating hydrogen and oxygen gas products.
On the downstream side, the development of effective catalysts is the core challenge.
These systems should avoid noble metals and should allow process operations resulting
in a minimum of by-product formation in order to minimize separation and purication steps for both the intermediates and the nal products. In particular, at the level of
nal products, the admixture of trace impurities absent in present-day fossil energy carriers can cause enormous barriers of application. A random example is the admixture
of traces of hydrogen in solar gas or wind gas sources: its presence would cause hydrogen embrittlement in the medium-to-high pressure part of the pipeline and pumping
systems and thus upset the whole existing infrastructure. Taking this into consideration
and remembering the likely need to operate energy conversion processes in decentralized small-scale units and at varying process conditions of partial load illustrates
quickly that we are far from having robust solutions even for process steps that are
conducted in present chemical plants such as ammonia synthesis or Fischer-Tropsch
processes. The boundary conditions come from system requirements upstream and
from the components of the energy system shown in gure 1.1.16 and must be met
by the chemical process steps. We can expect that such an approach will take time
to be built into future chemical process and catalyst development efforts.
The question is critical about the time structure of changes in energy systems.
From the point of climate protection, we need to stop growing the CO2 emission
29
30
25
20
15
10
5
2030
2025
2020
2015
2010
2005
2000
1995
1990
1985
1980
1975
1965
1970
1960
1950
1955
1945
Time (y)
Figure 1.1.18 Peak oil. The graphics shows on a crude timescale the temporal evolution of oil
production. The low resolution was chosen to identify the main trend irrespective of multiple
events in shorter timescales such as the oil crisis. Units: 1 Gba is 1.56 1011 L crude oil.
Source: Data from http://www.peakoil.com.
30
Figure 1.1.18 that we are far from being able to replace fossil sources in foreseeable
times in the dimensions that they are consumed at present.
In conclusion, the present chapter gives a crude overview of what it takes from the
side of chemical science to engage in the energy challenge. Future energy systems will
be based substantially on renewable solar primary energy but cannot be operated
without a suite of technologies of chemical energy conversion dealing with storage
and interconversion of energy carriers. The strategic and central science in this
endeavor is catalysis much in the same way as it now enables the operation of chemical and petrochemical industries. The insight and progress in understanding this
physicochemical phenomenon allows predicting that the challenge will be met at a
level higher than that enabling the evolution of the present energy industry. Despite
promising activities, a still enormous scientic, technological, and economic effort is
needed to initiate the turn of the energy system away from the dominance of fossil
energy carriers. The technologies of physical energy conversion by PV and wind
and other innovative concepts are far ahead of the chemical technologies. The combined efforts of science can be made faster and better targeted if the multiple interfaces between science and nonscientic inuences are recognized and dealt with by
the individual persons carrying responsibility in the energy science challenge. May
this book help to create awareness for the interdisciplinary work needed to create
new and sustainable energy systems.
1.1.7 References
1. Liao W, Heijungs R, Huppes G. Natural resource demand of global biofuels in the
Anthropocene: a review. Renewable Sustainable Energy Rev. 2012;16(1):9961003.
2. BMWI. Available from: http://www.bmwi.de/Navigation/Technologie-und-Energie/Ener
giepolitik/energiedaten.html.
3. Available from: http://www.leopoldina.org/leadmin/user_upload/Politik/Empfehlungen/
Nationale_Empfehlungen/Ad-hoc-Stellungnahme_Energie_Juni_2011.pdf.
4. Bundesregierung Energiekonzept. 2010. Available from: http://www.bmu.de/les/pdfs/
allgemein/application/pdf/energiekonzept_bundesregierung.pdf.
5. Deshmukh MK, Deshmukh SS. Modeling of hybrid renewable energy systems. Renewable
Sustainable Energy Rev. 2008;12(1):23549.
6. Ghanadan R, Koomey JG. Using energy scenarios to explore alternative energy pathways
in California. Energy Policy. 2005;33(9): 111742.
7. Doroodian K, Boyd R. The linkage between oil price shocks and economic growth with
ination in the presence of technological advances: a CGE model. Energy Policy. 2003;31
(10):9891006.
8. Heimann M, Reichstein M. Terrestrial ecosystem carbon dynamics and climate feedbacks.
Nature. 2008;451(7176):28992.
9. Clark JH, Deswarte FEI, Farmer TJ. The integration of green chemistry into future bioreneries. Biofuels Bioprod Bioren. 2009;3(1):7290.
10. Demirbas A. Progress and recent trends in biofuels. Prog Energy Combust Sci. 2007;33
(1):118.
11. Campbell H, et al. Efcient energy utilization and environmental issues applied to power
planning. Energy Policy. 2011;39(6):36307.
1.1.7 References
31
12. Kemp R. Technology and the transition to environmental sustainability the problem of
technological regime shifts. Futures. 1994;26(10):102346.
13. Lewis NS. Toward cost-effective solar energy use. Science. 2007;315(5813):798801.
14. Riahi K, Gruebler A, Nakicenovic N. Scenarios of long-term socio-economic and environmental development under climate stabilization. Technol Forecasting Soc Change.
2007;74(7):887935.
15. Ertl G, Freund HJ. Catalysis and surface science. Phys Today. 1999;52(1):328.
16. Aleklett K, et al. The peak of the oil age analyzing the world oil production reference
scenario in World Energy Outlook 2008. Energy Policy. 2010;38(3):1398414.
17. Schlgl R. The role of chemistry in the energy challenge. ChemSusChem. 2010;3(2):20922.
18. Centi G, Perathoner S. Opportunities and prospects in the chemical recycling of carbon
dioxide to fuels. Catal Today. 2009;148(34):191205.
19. Schlgl R. Ammonia synthesis. In: Ertl G, F. Schth, J. Weitkamp, editors. Handbook of
heterogeneous catalysis. Weinheim: Wiley VCH Verlag; 2008. p. 250175.
20. Olah GA. Beyond oil and gas: the methanol economy. Angew Chem Int Ed. 2005;44
(18):26369.
21. Raudaskoski R, et al. Catalytic activation of CO2: use of secondary CO2 for the production of synthesis gas and for methanol synthesis over copper-based zirconia-containing
catalysts. Catal Today. 2009;144(34):31823.
22. Tang QL, Hong QJ, Liu ZP. CO2 xation into methanol at Cu/ZrO2 interface from rst
principles kinetic Monte Carlo. J Catal. 2009;263(1):11422.
23. Liu Y, et al. Efcient conversion of carbon dioxide to methanol using copper catalyst by a
new low-temperature hydrogenation process. Chem Lett. 2007;36:11823.
24. Baldi M, et al. Catalytic combustion of C3 hydrocarbons and oxygenates over Mn3O4.
Appl Catal B Environ. 1998;16(1):4351.
25. Favre A, et al. Catalytic combustion of methane over barium hexaferrites. Catal Lett.
1998;49(34):20711.
26. Zavyalova U, et al. Morphology and microstructure of Li/MgO catalysts for the oxidative
coupling of methane. ChemCatChem. 2011;3(6):94959.
27. Su DS, et al. Nanocarbons in selective oxidative dehydrogenation reaction. Catal Today.
2005;102:1104.
28. Frank B, et al. Oxygen insertion catalysis by sp(2) carbon. Angew Chem Int Ed. 2011;50
(43):1022630.
29. Sanz AC, et al. Dynamics of the MoVTeNb oxide M1 phase in propane oxidation.
J Phys Chem C. 2010;114(4):191221.
30. Rozanska X, Sauer J. Oxidative dehydrogenation of hydrocarbons by V3O7+ compared
to other vanadium oxide species. J Phys Chem A. 2009;113(43):1158694.
31. Schlogl R. Active sites for propane oxidation: some generic considerations. Top Catal.
2011;54(1012):62738.
32. Herrmann J-M. The electronic factor and related redox processes in oxidation catalysis.
Catal Today. 2006;112(14):737.
33. Gruene P, et al. Role of dispersion of vanadia on SBA-15 in the oxidative dehydrogenation of propane. Catal Today. 2010;157(14):13742.
34. Geske M, et al. In-situ investigation of gas phase radical chemistry in the catalytic partial
oxidation of methane on Pt. Catal Today. 2009;142:619.
35. Guelder OL, et al. Unied behaviour of maximum soot yields of methane, ethane and propane laminar diffusion ames at high pressures. Combust Flame. 2011;158(10):2037044.
36. Hickman DA, Schmidt LD. Production of syngas by direct catalytic-oxidation of methane. Science. 1993;259(5093):3436.
32
37. Gelin P, Primet M. Complete oxidation of methane at low temperature over noble metal
based catalysts: a review. Appl Catal B Environ. 2002;39(1):137.
38. Ertl G. Elementary steps in ammonia synthesis: the surface science approach. In: Jennings
JR, editor. Catalytic ammonia synthesis: fundamentals and practice, fundamental and
applied catalysis. New York: Plenum Press; 1991. p.
39. Reuter K, Schefer M. First-principles kinetic Monte Carlo simulations for heterogeneous
catalysis: application to the CO oxidation at RuO2(110). Phys Rev B. 2006;73(4).
40. Stamp C, et al. Catalysis and corrosion: the theoretical surface-science context. Surf Sci.
2002;500(13):36894.
41. Somorjai GA, Park JY. Concepts, instruments, and model systems that enabled the rapid
evolution of surface science. Surf Sci. 2009;603(1012):1293300.
42. Baron M, et al. Interaction of gold with cerium oxide supports: CeO2(111) thin lms vs
CeOx nanoparticles. J Phys Chem C. 2009;113(15):60429.
43. Freund HJ. Adsorption of gases on complex: solid surfaces. Angew Chem Int Ed Engl.
1997;36(5):45275.
44. Williams WD, et al. Metallic corner atoms in gold clusters supported on rutile are the
dominant active site during water-gas shift catalysis. J Am Chem Soc. 2010;132
(40):1401820.
45. Dobler J, Pritzsche M, Sauer J. Oxidation of methanol to formaldehyde on supported vanadium oxide catalysts compared to gas phase molecules. J Am Chem Soc. 2005;127
(31):108618.
46. Freund H-J, et al. CO oxidation as a prototypical reaction for heterogeneous processes.
Angew Chem Int Ed. 2011;50(43):1006494.
47. Schlgl R. In situ characterisation of practical heterogeneous catalysis. In: Baerns M, editor. Basic principles in applied catalysis. Berlin: Springer Verlag; 2004. p. 32160.
48. Thomas J, Raja R. The advantages and future potential of single-site heterogeneous catalysts. Top Catal. 2006;40(1):317.
49. Thomas JM. Turning-points in catalysis. Angew Chem Int Ed. 1994;3(9):91337.
50. Zambelli T, et al. Identication of the active sites of a surface-catalyzed reaction.
Science. 1996;273(5282):168890.
51. Homann K, Kuhlenbeck H, Freund HJ. N2 adsorption and discussion on thin iron lms
on W(110). Surf Sci. 1995;327:21624.
52. Dahl S, et al. Electronic factors in catalysis: the volcano curve and the effect of promotion
in catalytic ammonia synthesis. Appl Catal A Gen. 2001;222(12):1929.
53. Teschner D, et al. Alkyne hydrogenation over Pd catalysts: a new paradigm. J Catal.
2006;242(1):2637.
54. Goodman DW. Catalysis from single-crystals to the real-world. Surf Sci. 1994;299(13):
83748.
55. Cambell CT, et al. Micro- and macro-kinetics: their relationship in heterogeneous catalysis. Top Catal.1994;1(34):35366.
56. Aparicio LM, Dumesic, JA. Ammonia synthesis kinetics: surface chemistry, rate expressions, and kinetic analysis. Top Catal. 1994;1(34):23352.
57. Hellman A, et al. Predicting catalysis: understanding ammonia synthesis from rstprinciples calculations. J Phys Chem B. 2006;110(36):1771935.
58. Schlgl R. Combinatorial chemistry in heterogeneous catalysis: a new scientic approach
or the kings new clothes? Angew Chem Int Ed. 1998;37(17):23336.
59. Schlogl R. The role of chemistry in the energy challenge. ChemSusChem. 2010;3(2):20922.
60. Greeley J, et al. Computational high-throughput screening of electrocatalytic materials for
hydrogen evolution. Nat Mater. 2006;5(11):90913.
1.1.7 References
33
34
85. de Leon, CP, et al. Redox ow cells for energy conversion. J Power Sources. 2006;160
(1):71632.
86. Tarascon JM, Armand M. Issues and challenges facing rechargeable lithium batteries.
Nature. 2001;414(6861):35967.
87. Jamnik J, Maier J. Nanocrystallinity effects in lithium battery materials aspects of nanoionics. Part IV. Phys Chem Chem Phys. 2003;5(23):521520.
88. Rossmeisl J, et al. Electrolysis of water on oxide surfaces. J. Electroanal Chem. 2007;607
(12):839.
89. Rossmeisl J, et al. Comparing electrochemical and biological water splitting. J Phys Chem
C. 2007;111(51):188213.
90. Klerke A, et al. Ammonia for hydrogen storage: challenges and opportunities. J Mater
Chem. 2008;18(20):230410.
91. Schlgl R. Catalytic synthesis of ammonia a never-ending story? [Review]. Angew
Chem Int Ed. 2003;42(18):20048.
92. Woodhouse M, Parkinson BA. Combinatorial approaches for the identication and optimization of oxide semiconductors for efcient solar photoelectrolysis. Chem Soc Rev.
2009;38(1):197210.
93. Kanan MW, Nocera DG. In situ formation of an oxygen-evolving catalyst in neutral
water containing phosphate and Co2+. Science. 2008;321(5892):10725.
94. Maeda K, et al. GaN:ZnO solid solution as a photocatalyst for visible-light-driven overall
water splitting. J Am Chem Soc. 2005;127(23):82867.
95. Yagi M, et al. Molecular catalysts for water oxidation toward articial photosynthesis.
Photochem Photobiol Sci. 2009;8(2):13947.
96. Alstrum-Acevedo JH, Brennaman MK, Meyer TJ. Chemical approaches to articial photosynthesis. 2. Inorg Chem. 2005;44(20):680227.
97. Meyer TJ. Chemical approaches to articial photosynthesis. Acc Chem Res. 1989;22
(5):16370.
98. Gratzel M. Articial photosynthesis water cleavage into hydrogen and oxygen by visiblelight. Acc Chem Res. 1981;14(12):37684.
1.2.1 Introduction
In the light of dwindling reserves of fossil fuels, which serve not only as a source of
energy for our societies but also as the major form of storing energy, energy storage
has become a pressing and modern topic. However, energy storage is probably as old
as our civilization, although we mostly do not think about it in these terms. Collecting wood for maintaining a re or to provide heat during the winter is a form of
energy storage, and most of the features of modern energy storage are indeed present
in rewood: it has a high energy density (about one-third of the gravimetric energy
density of diesel fuel), it can be stored for practically an indenite amount of time
without losing its energy content, it is reasonably safe, and it can easily be converted
to the useful forms of energy for which it is stored (i.e. heat and/or light).
In modern societies, energy storage is required for a whole range of different timescales and size scales (Figure 1.2.1). Capacitors in electronic devices store small
amounts of energy for short times; batteries provide the energy for cell phones, laptops, or the propulsion of cars; gasoline, diesel fuel, and jet fuel provide the high storage density required for cars, trucks, or planes; and big subterranean caverns store
grid-scale amounts of energy in the form of oil and natural gas. With more and
more uctuating energy from wind and solar power in the electricity grid and decreasing reserves of fossil fuels, energy storage will be a key component in any future
energy infrastructure. In the following, the strategies for energy storage on different
size scales and timescales will be discussed in more detail.
36
Chemical compounds
grid scale
106
Li-ion
(electronics)
Engine
fuel Redox-Flow
Li-ion Pb battery
CAES
(traction)
NaS battery
(Pumped) hydro
1 year
1 month
1 day
104
1 hour
Flywheel
102
Capacitors
1
1 Wh
1 kWh
1 MWh
1 GWh
1 TWh
Figure 1.2.1 Typical time and size scales associated with sufcient storage technologies.
forms of energy can be generated at will and if needed. This is the reason why our
societies are predominantly relying on fossil energy, which is used as a kind of easily
convertible energy currency.
On a crude scale, primary energy serves to supply one of three consumption
sectors that is, mobility, heat, and electrical energy, the latter of which, in turn,
is converted to various forms of useful energy, such as mechanical energy, light,
heat, and many others. Very roughly, globally each of these sectors requires onethird of the primary energy, while electricity has a smaller fraction in the nal energy
consumption due to the conversion losses from primary energy to electrical energy.
Energy storage has to meet completely different requirements for each of these consumption sectors, and the different storage concepts and technologies have to come
together in a concerted manner to provide the basis of an energy system.
It should be noted here that in many cases energy storage is not without alternatives. Electricity has to be produced at any given time in exactly the same amount as
it is being consumed. In times of overproduction for instance, during periods of
strong winds electricity could be stored and then released in times of high demand
and low production. Alternatively, however, one could install overcapacities, which
would be shut down in case of overproduction. Excess energy would therefore not be
stored. This would not make full use of the available energy resources (i.e. of wind or
solar irradiation) but overall could be a more cost-efcient solution. The higher the
transfer capacities of electricity grids and the better the grids are connected, the lower
the need for additional storage since exchange of electricity between different regions
could help to balance supply and demand. Similarly, if most long-distance trafc
would make use of trains, fuels for internal combustion engines or batteries of electric
37
vehicles would not be needed to such an extent. Thus, as important as energy storage
in various ways will be in future energy systems, the optimization of the whole system
may rely on storage as well as on other technologies, with the balance between the
approaches based on cost, practicability, societal acceptance, and other factors. It
is expected that future energy systems will become increasingly electric, and the consumption sectors of mobility and electricity will merge to an appreciable extent, since
more and more of the transportation sector will rely on electrical energy. Heat
demand of buildings, especially of new buildings, will decrease due to improved insulation, and heat storage of solar thermal heat could become an important issue. The
heating sector could become largely decoupled from the rest of the energy system,
with the exception of cogeneration and purely electrical heating (heat pump or resistive) to meet the residual heat requirements of buildings. In the next section, therefore, the storage of heat energy will be discussed before we turn to the issue of
storing electrical energy in future energy systems.
Property
Comment
Type of storage
Enthalpy effect
Temperature range
Cycle stability
Kinetics
As fast as possible
Hysteresis
Ideally absent
Losses
Ideally absent
38
There are two major emerging application elds for heat storage in future energy
systems, storage on the scale of hundreds of MWh for solar thermal power plants and
on the order of several MWh for domestic heating or district heating. If solar thermal
plants should be able to provide base load power, then even in regions where there is
almost never any cloud coverage, energy needs to be provided during the nighttime
hours. Current state-of-the-art solar thermal plants have power on the order of
50 MW, which means that for bridging 8 h at night, 400 MWh of electrical energy
needs to be generated. With typical efciencies of steam turbines, heat storage capacities on the order of 1,000 MWh (thermal) are required. In the Andasol 50 MW
solar thermal plant in Spain, this amount of heat is stored in 27,500 t of 60%
NaNO3/40% KNO3 melt. During the day, part of the heat generated in the parabolic
trough mirror eld is used to power the steam turbine, while the remainder heats up
the molten salt to approximately 390C. At night, the salt melt is passed through a
heat exchanger and cooled to approximately 290C, and the extracted heat provides
the energy for the steam turbine. This technology has proved to work in the eld and
could be deployed at a large scale in solar thermal power plants. It is rugged
provided that the salt does not solidify and can be scaled so that it can be adapted
to differently sized power plants. However, the cost of the material alone is substantial: at typical fertilizer prices (the salt is used for this purpose), the cost is on the
order of $1520 million per plant. Alternatives thus appear to be attractive. Sensible
heat storage is possible in other storage media as well. The use of huge concrete
blocks with integrated heat exchangers is an option that is currently being explored.
Another interesting system is Mg/MgH2, which has very good heat storage properties
and was extensively investigated in the 1980s by Boris Bogdanovic. The hydrogenation/dehydrogenation reaction has a thermal signature of 0.9 kWh kg1. Thus, for a
1,000 MWh thermal capacity, only somewhat over 1,000 t of magnesium instead of
27,500 t of salt melt is required, at a materials cost of approximately $3 million.
However, a more complex system is necessary since the hydrogen also needs to be
handled. An additional, dedicated hydrogen storage facility would probably render
such a system more expensive than a magnesium based heat storage facility. There
could be situations, though, in which solar thermal plants are also used for the generation of hydrogen via thermochemical cycles so that a hydrogen infrastructure
would be in place. Under such circumstances, hydride heat storage systems could
be a competitive alternative, especially since the temperature level can be adjusted
by the hydrogen pressure and/or the use of other metals or alloys.
Domestic heating in future homes will probably not require fossil fuel red heating
systems in most regions of the world due to much better insulation standards. Lowtemperature heat can be stored in sensible or latent heat storage systems. The rst
installations are already in operation. In Munich, the complex Ackermannbogen
has a 6,000 m3 water sensible heat storage, which is heated during the summer by
solar panels. The stored energy is used to heat the 319 apartments until well into
January; from then on, district heating provides the rest of the heating energy.
This saves approximately 50% of the heating energy that would be required. In addition, the German parliament has a subterranean heat storage at a depth of approximately 300 m. Another water reservoir at 60 m depth is used to provide cooling
during the summer months. Such concepts will become more and more widespread
39
in modern buildings; the rest of the required heating energy will be provided by heat
pumps and/or direct electrical heating.
While water is a cheap and simple solution, higher storage densities can be
achieved by other technologies. Water stores 4.18 J K1 g1 in the temperature range
up to 100C (i.e. assuming a useful temperature range of 50 K, approximately
200 kJ kg1 can be stored), and water is one of the most efcient sensible heat storage
media. The hydration and dehydration of zeolites, crystalline porous alumosilicates,
can be used in a similar temperature range. However, depending on the type of zeolite,
the heats of hydration can be ve times as high, and such a system would operate at
constant temperature. Space requirements could be substantially reduced so that
amounts on the order of 1020 t, corresponding to one room in the cellar, would be
sufcient for the heating of a well-insulated house during the whole winter. The
costs of such systems are presently still not competitive, but with increasing energy
prices, this may change in the future.
Other latent heat storage systems are also under development. Sodium acetate trihydrate with a melting point of 58C is the heat storage medium in hand warmers.
When crystallization is induced, the heat of fusion is released. Such systems are
not only used on a small scale, such as in hand warmers, but also industrially.
Low-temperature waste heat from industrial plants is used to melt the salt in large
containers. These are then transported to heat consumers, such as swimming
pools, and there the heat generated during solidication is used to heat the water
of the pools. The rst business models are emerging in which such systems are the
key component, for instance, by the company LaTherm (Dortmund, Germany).
40
Property
Comment
Cycle efciency
Energy density
Power density
Cycle stability
Self discharge
Life time
As long as possible
Investment cost
Safety
High
one important cost factor, while the other is in the capital expenditure for building
the facility. Cost effectiveness of the overall system depends on both of these cost
factors and on the frequency at which the storage system is used. For seasonal storage, investment costs should be low since otherwise the cost per kWh stored becomes
prohibitively high. On the other hand, for short-term/high-frequency storage, one can
afford higher capital expenditure since it is distributed over many operation cycles;
however, cycle costs and losses should be small for such systems. The desirable
features of storage for electrical energy are listed in Table 1.2.2.
Complex electricity grids consist of different levels. The transmission grid is used
for long-distance, high-power electricity transport. It operates typically at voltage levels of around 400 kV. One level lower is the distribution grid, in Europe at 110 kV.
Transformers convert this voltage to lower values for the intermediate-voltage grid
operating at 10 kV, 20 kV, or 30 kV which nally serves the local low-voltage grids
to which the small-scale users of electricity are connected. Storage on the level of the
transmission grid and on the sublevels requires different storage technologies, and
thus they will be discussed separately.
41
means for stabilizing the grid since reaction times on the order of minutes are possible, which allows fast reaction to changing demand and supply situations, and the
pumps and turbines can be regulated over a wide range of power.
However, in densely populated areas or in regions with minimal height differences,
the capacities for hydroelectric dam power are rather low since mechanical energy
generated from height differences does not have high storage density. Assuming
a height difference of 300 m, which is not untypical, 1 m3 of water has an energy
content of
m . g . h = 1,000 kg . 9.81 m s2 . 300 m = 2,943 kJ
(i.e. only about 0.8 kWh). In Germany, the total capacity of pumped hydro is
approximately 40 GWh, which is very little compared to the electricity consumption
of Germany (on the order of 600 TWh) or the total primary energy demand of
Germany (approximately 4,000 TWh). Thus, if hypothetically one would shut
off the primary energy supply to Germany, the amount of stored energy in water reservoirs would be sufcient to cover the energy demand for a little bit longer than
5 minutes. Other countries have much higher storage capacities in hydroelectric
dams (although at present these are mostly not pumped, pumps can often be retrotted). Austria, for instance, can store 4.5 TWh in dams; Sweden, 33.8 TWh; and
Norway, 81.7 TWh.
However, the balancing capacity of hydroelectric dams can only be exploited if
there is a well-established grid with low energy losses over long distances. The
only present-day method for low-loss/long-distance electricity transport is highvoltage direct current (HVDC) technology. To use hydroelectric storage to cope
with more and more uctuating energy demand and supply, a continental HVDC
overlay grid structure seems indispensable. With such a grid structure available,
hydroelectric dams, especially if combined with pumps, can provide an important
storage element in future energy systems. The cycle efciency of pumped hydro is
on the order of 80%, and of all storage methods available nowadays, (pumped)
hydroelectric energy is also the most cost-efcient method, both for short- and for
long-term storage.
1.2.4.1.2 Compressed Air Energy Storage (CAES) Another grid-scale method for
energy storage relies on air compressed in large subterranean caverns. The advantage
over hydroelectric power is the availability of many possible sites and the low degree
of interference with the landscape since most of the system is underground.
In CAES storage systems, air is compressed into airtight underground caverns
(several hundred thousand cubic meters volume) at times of high electrical energy
supply. To recover the stored energy, the compressed air is used to charge a gas turbine, thus making the compressor of such turbines redundant. The cycle efciency of
the systems depends on the heat management: during compression, the air is heated.
If this heat is lost, the efciency of the system is below 50%. Full heat recovery (adiabatic operation) should lead to efciencies around 70%, but currently, no such systems are installed.
42
Two CAES plants are in operation at present, one in Huntorf in Germany (321 MW)
and one in MacIntosh, Alabama, in the United States (110 MW). The plants have
already been in operation for several decades, and this technology can be considered
well proven. However, since capital expenditure for pumped hydro and CAES are
estimated to be in the same range per kWh capacity, overall storage costs are higher
for CAES due to the lower efciencies, and the capacities are relatively low. The
300,000 m3 storage caverns of the Huntorf plant can be used to drive the gas turbine
for 2 h at full power, which corresponds to a storage capacity of approximately
500 MWh, considering that natural gas is also fed to drive the turbine. Nevertheless,
CAES systems could nd more widespread use if grid extension does not take place
or proceeds at a slow pace and/or resistance against pumped hydro plants is high.
1.2.4.1.3 Storage in Chemical Compounds Much higher energy densities than with
pumped hydro or CAES can be achieved if the electrical energy is converted to a
chemical compound. Such a compound could then either be reconverted to electrical
energy or used in applications, where chemical compounds are indispensable at
least for the time being such as for jet fuel. However, conversion to any chemical
compound proceeds at low efciency, and thus storage in the form of chemical
compounds would probably only be used for long-term (seasonal) storage.
Five different compounds or groups of compounds are being discussed for energy
storage (i.e. hydrogen, methane, hydrocarbons, methanol, and ethanol). Ethanol has
a special status in this list since it would probably not be produced from electrical
energy, but directly by fermentation of biomass. Thus, it will not be discussed further
here. The shortest route to a chemical compound from electricity is water electrolysis
to result in the formation of hydrogen. All other compounds are produced from
hydrogen with a carbon source, such as CO or CO2. However, as long as fossil
fuels are used to produce electricity, one should not hydrogenate CO2 to produce
methane, methanol, or liquid hydrocarbons via the Fischer-Tropsch process. Instead,
the hydrogen should be directly reconverted to electrical energy and the power of a
fossil-fuel power plant should be reduced because otherwise additional losses would
occur.
For grid-scale storage, hydrogen appears to be the best compound. It has a number of advantages that is, storage densities are very high, the production by electrolysis is an established technology, storage is possible in similar caverns as in CAES, it
can be reconverted to electrical energy by fuel cells or turbines, and storage is essentially possible for indenite periods of time (subterranean caverns are reported to
loose less than 0.01% per year). However, hydrogen has one major disadvantage,
the substantial energy losses during one cycle: electrolysis can be estimated to have
an efciency of 60%, transport and compression for storage may lead to another
10% losses (although some of this energy could possibly be recovered), and reconversion to electricity proceeds at most at 50% efciency in fuel cells on the systems level
(higher efciencies could be possible in turbines, especially if cogeneration of heat is
integrated). However, the overall cycle will most probably not be more efcient
than 30%. The major advantages are the high storage density compared to CAES,
43
the capacity of the caverns could be increased by a factor of about 100 and the
possibility of long-term storage.
One should keep in mind that chemical compounds for energy storage are also
accessible from other renewable, nonelectricity sources. Methane could be generated
by fermentation from biomass, which is one of the most efcient ways for its energetic use; ethanol is accessible from sugar by fermentation, also a rather efcient
pathway; and liquid hydrocarbons could be produced by various biomass-to-liquid
processes. These pathways appear to be superior on the system scale, compared to
hydrogenation of carbon compounds with hydrogen produced by electrolysis.
Thus, hydrogen in spite of all shortcomings is probably the only feasible solution
for directly storing electricity in the form of a chemical compound.
44
Flow batteries are hybrids between a conventional battery and a fuel cell. The
energy is provided by liquids at different electrochemical potentials, which can be
passed along an electrode to deliver the power. The most well-known system is the
vanadium redox-ow battery, in which in the charged state one electrolyte consists
of V5+ solution, and the other one of V2+ solution. Discharge leads to reduction
of V5+ to V4+ and oxidation of V2+ to V3+; in charging, the process is reversed.
Redox-ow batteries have a number of advantages: power (dependent on the electrode area) and capacity (dependent on electrolyte tank size) can be scaled independently, and there is basically no self-discharge. The cycle and calendric life of redoxow systems are good, and investment costs are moderate, somewhere in between
that of lead acid batteries and NaS batteries. However, the storage densities are
rather low, so redox-ow installations have rather large footprints.
Battery technology is one of the most rapidly developing elds in energy storage,
driven by the anticipated application in electromobility, so that in the future additional competitive technologies could become available for stationary applications
as well.
1.2.4.2.2 Flywheel Storage Intuitively, one would not expect ywheels to be suitable storage devices for larger amounts of electrical energy. However, advanced,
very fast spinning types with carbon wheels can store substantial amounts of electrical energy. Commercial units with about 1 m diameter and 2 m height can store
25 kWh of electrical energy at a rotational speed of 16,000 rpm. Storage densities
are substantially higher than for CAES and can reach almost the values of redoxow battery systems. However, investment costs are very high, and self-discharge is
rapid. Thus, such systems are only used for grid frequency regulation. A 20 MW
plant with 200 independent ywheel modules is in commercial operation in the United
States.
45
be used for this purpose. For regular cars, however, alternative options for onboard
energy storage are within reach, even if there are some disadvantages associated with
them; the different options shall briey be discussed in the following.
46
the necessity of additional infrastructure, the low energy density, and the toxicity of
methanol. Balancing all factors against each other, methanol does not seem to be a
promising energy storage compound for mobile applications or for energy storage on
a grid scale.
1.2.5.1.4 Methane Methane is already being used as fuel in some cars. It can be
combusted in internal combustion engines after adaptation, but since it is a gas,
an additional tank is required. While the gravimetric storage density of methane is
very high (13.9 kWh kg1), the volumetric density is low since it is a gas at ambient
temperatures. Even compressed methane does not reach the volumetric density of liquid hydrocarbons (0.0099 kWh L1 at 1 bar). Methane is accessible from synthesis
gas by the Sabatier process. However, it would probably be better to use anaerobic
fermentation of biomass to produce biogas, which could then be puried to methane.
There is well-established methane infrastructure available, with large storage facilities (in Germany, 20 billion m3, corresponding to one-fth of the annual consumption) and a pipeline network. However, on the systems level, it does not appear to be
advantageous to use methane to power cars, considering that there are other options
with probably better overall performance.
1.2.5.1.5 Hydrogen Hydrogen in connection with proton exchange membrane fuel
cells presently appears to be the only option that could provide similar driving ranges
as cars powered by the internal combustion engine. Hydrogen has the highest gravimetric storage capacity (33.3 kWh kg1) of all chemical compounds, but the volumetric capacity is very low due to hydrogen being a gas (0.003 kWh L1 at 1 bar).
Storage of hydrogen onboard cars is one of the most important problems. Highpressure storage (700 bar) is currently state of the art, and despite intensive research
efforts, no other option is forthcoming. Storage in liqueed form, adsorptive storage
at liquid nitrogen temperature, storage in the form of hydrides, and storage in liquid
chemical compounds (from which hydrogen is released by reforming) have been explored, but no viable solution superior to high-pressure storage is in sight. One
advantage of hydrogen as a fuel is the fact that no CO2 is emitted at the point of
use. In addition, the efciency of a fuel cell is appreciably higher (around 50% on
the systems level) than that of an internal combustion engine, which leads to energy
savings. However, as with hydrogen storage on a grid scale, one has to solve the
question of hydrogen generation, which is associated with high energy losses.
Whether hydrogen eventually will become an important fuel for mobile applications
depends on advances in all key components of this technology. Possibly, hybrid vehicles using batteries for short driving ranges and a hydrogen-powered fuel cell for
long-distance trafc will be the successful solution for transportation beyond the
internal combustion engine.
47
this technology is CO2 neutral. Depending on charge and discharge currents, efciencies on the order of 80% can be achieved. The energy consumption of electric vehicles
is an estimated 0.15 kWh km1, while internal combustion engines need about four
times as much energy. However, the efciency of electricity generation from primary
energy has to be taken into account as well, and thus the difference is not quite as
high as it appears.
The biggest disadvantage of batteries for mobile applications is the low energy
density of electrochemical energy storage. The most advanced battery types for
mobile applications are lithium-ion batteries, which are used in essentially all laptops
and cell phones. Energy densities for this type of battery are around 200 Wh kg1 and
about 500 Wh L1. The prospects for a substantial increase in energy density are not
too bright for this type of battery since about half of the weight is contributed by the
positive electrode material. Although different types of electrode materials are being
used, they all rely on transition metal oxides or phosphates or related compounds, in
which the Li ions are intercalated. Thus, material with a formula mass of around
100 g mol1 is used to store the Li ion with an atomic mass of about 7 g mol1,
but only the latter carries the energy. Unless the dead weight of the matrix can
be substantially reduced, the energy density of traction batteries will not improve
signicantly. In order to make progress, completely new battery concepts are needed.
One interesting aspect of large-scale deployment of traction batteries shall be mentioned: they could also be used as a distributed storage system for grid-scale electricity storage. Cars are parked most of the time, and if connected to the grid during this
time, they could be used to balance uctuating power. Whether this would be done
unidirectionally ( i.e. charging would only occur at high electrical energy supply) or
bidirectionally (i.e. batteries would also be discharged during times of high electricity
demand) depends on the cycle stability of traction batteries. If this is low, bidirectional use would be too expensive because it would lead to short battery life. Assuming a size of 30 kWh for traction batteries (between that of hybrid cars like the Opel
Ampera and that of an all-electric car like the Tesla) and a eet of 42 million cars in
Germany, there would be an overall storage capacity of 1.26 TWh. This is very substantial compared to the 40 GWh storage capacity of pumped hydro in Germany,
even if this capacity would not be fully utilized since the main purpose of cars is,
after all, to provide mobility.
48
associated with losses, and thus such conversion should be avoided. Grid extension
helps to this end, and it also helps to couple the cheapest and most efcient form
of transmission grid-scale storage, pumped hydro, into a continental electricity system. Mobility, especially passenger transport, should be tightly integrated into the
electricity system, and battery-powered electromobility appears to be the most efcient option although we will probably not be able to achieve the driving ranges
we are used to nowadays. Thus, some behavioral changes, such as increased use of
train transport, may be necessary. If excess electricity should be stored for long
time periods, centralized hydrogen production and storage in caverns is probably
the best solution. This hydrogen could be used to power fuel celldriven cars or
hybrid cars for long-distance car travel, or it could be directly reconverted to electricity, albeit at low cycle efciency. For seasonal storage, methane also is an attractive
option. However, methane should be produced from biomass via fermentation rather
than from electrically generated hydrogen hydrogen should be stored directly,
not further converted. Biomass is also a suitable starting point for the production
of hydrocarbons, which will be indispensable as energy carriers for some kinds of
transportation, such as aviation and heavy-duty trucks.
Finally, the heat market will be largely decoupled from the rest of the energy system. Heating energy demand will decrease due to better insulation standards, and
part of the heating and cooling energy requirements will be met by heat storage systems. The rest of the heating and cooling power will be generated by electricity, and
in combination with heat and cold storage, this will also have an equilibrating effect
on the electricity system.
While progress in science and technology can always lead to unforeseen step
changes, the planning of a new energy infrastructure has to rely on technologies
at hand or at least in sight. From what we have and see, it seems clear that energy
storage will be an important element in any future energy system.
The history of science was a history of curiosity [1] as a major weekly German
newspaper recently wrote.1 One has to recognize that curiosity is the root of all
human inventions and discoveries. It was not out of sheer reection that sailors
courageously sailed the seven seas, that anatomists carried out self-experimentation,
and that scientists exposed themselves to dangerous rays (hardly anybody knew
where his drive would take him), but out of curiosity. According the German newspaper Die Zeit, Albert Einstein was only honest when he once said: I am not particularly talented, but only passionately curious. The fact that today this curiosity
can develop freely (for Galileo Galilei and Giordano Bruno, this was more dangerous) is a promise of democracy and the secret of its success. Only where inventiveness
is allowed to spread and interact does it break unfamiliar and economically fruitful
new ground and create something original.
Umberto Ecos novel The Name of the Rose can also be read as a story of curiosity
and its limits. The old, blind monk Jorge was well aware of the reason he deprived his
(not only monastic) contemporaries of the Second Book of the Aristotelian poetics of
curiosity: After all, a philosophy of laughter would change the world and entail tolerance and open-mindedness for new ideas old, conservative, authoritarian, even
fanatic structures would be put to the test. Finally, curiosity wins out: in the lm
based on the book, Sean Connery in the role of Ecos former inquisitor William of
Baskerville solves the apocalyptic murder cases monastery and book, however,
burst into ames. Something new is coming.
But this new does not arise without failures. Sometimes it takes effort; sometimes
painful delivery processes are necessary in short, a crisis, which, according to Antonio Gramsci, consists precisely in the fact that the old is dying and the new cannot be
born [2].2 However, the new is unstoppable; at best, it can be prudently shaped.
It was mainly the natural sciences that established the new in an unparalleled triumphal procession through history and that pushed it forward again and again, nally
accompanied by a justiably strict scientic theory: Do not verify falsify! which
Karl Popper claimed in contrast to the positivists. Those who wanted to put forth
a hypothesis and wanted to prove it would mostly look for what supported it; contradictory ndings would gladly be overlooked. Popper elevated the search for error
to a principle. Consequently, there were no nal truths because the principle was to
attempt to disprove each new theory [3]. However, truth was approaching adaequatio
intellectus et rei, which was what Thomas Aquinas called it.
But science either remained in the ivory tower for centuries or threw itself into the
arms of the ruling class, which had the power and the means to satisfy the craving for
recognition and the urge to discover. At the same time, the ndings in science meant
50
improvements in the quality of life for humans for many centuries, if not for millennia; in short, science meant progress. This began to change, imperceptibly at rst,
with the Industrial Revolution, when energy became the most important fuel of economic development energy soon mainly obtained from coal. Later, a by far more
dangerous change was the discovery of nuclear ssion, at rst, in nave optimism,
misunderstood as the nal solution to the energy problem. With this and the advancing destruction of the environment, progress lost its positive meaning; it became
ambivalent with the accompanying consequences for the political powers that had
based their programs on progress as a development driver for a better world. What
science had discovered had in part become dangerous so dangerous that it even
threatened the continuity of humankind.
Currently, the period of time still left for humans to stop climate change and global
warming, in particular in the Northern Hemisphere, is dwindling dramatically.
There, global warming takes place considerably more quickly than in the Southern
Hemisphere. Meanwhile, there is controversy among climatologists regarding the
question of whether the sea level will rise by 1 m or 5 m. In November 2011, Fatih
Birol from the International Energy Agency, presenting the World Energy Outlook
in Berlin, declared that the door to confront climate change will close in the year
2017; after that, we will be locked in.
Not only Los Alamos, the Manhattan Project, Drrenmatts play The Physicists
have changed the (self-) conception of serious science or represent symbols for this
change of mind-set. Science is responsible for what it realizes, discovers, and invents.
It is responsible for the effects it has, for how it is used, and for what consequences it
has for humankind. Science must leave the ivory tower of unprejudiced but responsibility free (if not irresponsible) knowledge. Earth has reached a point where science
must assume responsibility for its further existence, where it must intervene. It is not
just a question of not being allowed to do all we are able to do (contrary to the viewpoint of the American Christian Right, which claims that God has created the world
in such a wonderful way that humankind will not be able to destroy it is disproved
daily), but it is also a question of science, primarily natural science, making greater
efforts than ever to explain what has been researched. Science must intervene but
those who want to intervene successfully must want to be understood.
Therefore, it is about time for scientists to start using a comprehensible language
when addressing the public. Certainly, science strives to be precise, unambiguous,
invulnerable, to leave as few uncertainties as possible. This effort for preciseness,
however, does not contribute to (general) understanding; it is more of a hindrance
not for nothing was scientic gibberish a swearword in the 1950s and 1960s. One
illustration is the following sentence from a textbook that is only a couple of years
old: A general description of the communication management of administration requires . . . an analysis of processes through which the relation of operation and structure is continually reproduced [4, p. 42].3 Apart from the participants of the authors
advanced seminar, hardly anyone will understand it, and it is even doubtful whether a
translation into the understandable is possible at all. In the natural sciences, it is no
different.
Translation is required and previously the effort to achieve mutual understanding. Scientists should not generally regard journalists as dangerous simpliers, and
51
52
behavior that deviate from the standard are real news and of serious interest. Those
who are not read, heard, and seen lose interest in being read, heard, and seen in the
long run.
Moreover, the mission of the often wrongly so-called Fourth Estate, the media,
bears an internal conict: Every day, and on the World Wide Web even every minute, press, radio, and television in their diversity practice freedom of expression, one
of the primordial rights of democracy, the temporary control of the power and the
powerful. However, the products of this Fourth Estate must hold their ground in
the market; they must sell and achieve high ratings at the same time. The voting
at the kiosk and with the remote control on the sofa is relentless (with the exception
of the public broadcasting corporations, although they paradoxically act even more
as slaves to the rating chart).
Science provides absolutely welcome approaches to comprehensively systematic
thinking (and researching) and to processing and communicating results in a way
understandable for the public. There is still a lot that has not been sufciently investigated so far. Apart from the strongly increased CO2-emissions since the beginning
of industrialization, there are many also regional factors that have had an impact on
the global climate however, many among them have not been really understood to
date. Causes and effects of the terrestrial gravity and magnetic eld, the biosphere
and its impact on the climate problem, and also the tipping points themselves are
still insufciently investigated. The same applies to the processes of energy storage
and water electrolysis, for example, by means of desert energy as proposed by Desertec (the Max-Planck-Gesellschaft [MPG] is an associated partner of the Desertec
Industrial Initiative GmbH), a relatively undisputed vision within the framework
of the development cooperation in the Europe, Middle East, and North Africa
(EUMENA) area. The possibilities of government support for the long-term security of energy supply in Europe, North Africa, and the Middle East by means of
renewables from the deserts require urgent examination.
According to the Stern Review, we are facing two challenges of the century
climate change and poverty and both must be seen and addressed together. We
can only be successful in both elds together [8, 9, p. 10]. This is a deeply political
constellation. It requires a holistic consideration.
Scientists are not the better politicians and should not behave as such they can
only give advice and they must dare to give advice because with this they assume
the responsibility for their work. The (detailed) work with natural climate parameters
may not be trivialized for reasons of specic interests: Not only in the United States is
skepticism growing (organized and unscrupulously funded by industry) toward the
probability of anthropogenic global warming postulated by the IPCC and the Stern
Review. In Europe, the climate change skeptics are busy conquering the regulars
tables and blogs. A Google search for climate deniers returns more than 24,000
entries; for climate realist, there are more than 9,000 entries.
Reliable energy supply, energy efciency, and the accelerated transition to the
postfossil fuel generation of energy, thus low-carbon energy management, are essential prerequisites for sustainable growth not only since the energy transition in
Germany. Energy policy must also consider the niteness of the resources and political dependencies due to unbalanced energy supply. Energy issues must be addressed
53
54
The global (and European) nancial system must urgently be reformed. Its deceleration, inter alia through controls, public presentation of externalized production costs,
a tax on nancial transactions (such as the Tobin tax), and the limitation of nonperformance-related benets, are favored by more and more countries. The sustainability of the capital expenditure and its recourse to environmental compatibility
must become the essential criterion of economic action. And it must be made accessible to journalistic thinking it must represent a motivation to reect on interesting
public-oriented forms of how to mutually agree, initiate, and implement the decisive
steps.
This requires the use of modern technologies for public information, mechanisms of
problem communication, of supporting public relations management, and of agendasetting methods long common in other elds in favor of the acceptance of the consequences arising from changes. The development and impartation of climate knowledge
and the acceptance of the consequences of the climate-protecting policy must urgently
be included in school curricula. Strategies for these goals must be developed. The individual commitment must be strengthened; higher public perception and more recognition are needed. Forms, techniques, and strategies for the implementation and
nancing of necessary ideas and changes, above all from the eld of nongovernmental
organizations, must be studied more intensively and be better available.
The Leopoldina recommendations thus propose that the post-Fukushima discussion should be used to reach, to the extent possible, a non-partisan and widely approved consensus on future energy policy in Germany [10, p. 8f ]. According to the
nal report of the German Ethics Committee (originated by the German federal government in March 2001, which had submitted its nal report at the end of May), entitled Germanys Energy Transition, the society-wide discussion on the energy
transition should be continued to sustain the motivation of the citizens even when
the memories of the accident in Fukushima fade. The subsequent recommendation
in this report to establish a National Forum for Energy Transition takes this into
account [11, p. 19]. This motivation would be supported if individuals were able
to take part in the energy transition. It would provide more self-determination and
would make it easier for them to control their own electricity consumption. A private
household could save electricity by using smart meters. With the aid of these and
comparable simple means of acquiring information, it is expected that many people
will decide, much faster than before, to replace their main energy consumers, generally the fridge and the heating system, by more efcient appliances [11, p. 29]. It
would not be a matter of a skilful creation of acceptance, but rather of public participation in the widely supported energy transition and a matter of a fair balance of
impact and benet [11, p. 20]. In this context, the contradiction that politicians
elected for four or ve years have to make decisions that have a binding effect far
beyond their term must be addressed. The Leopoldina recommendation soberly states:
The time scale involved in changing energy systems is so long that changing course in
time with Germanys four-year electoral term is counter-productive [10, p. 9]. This
means that the mentioned change of course must reach far beyond one election period.
However, does the democratic system designed for comparatively short election
periods and terms in ofce actually have the legitimacy to make long-term decisions
that will bind the successors for decades, even for centuries? The professor emeritus
55
from Cologne, Carl Christian von Weizscker, doubts this: How do sustainability
and democracy get along with each other? he asked in a newspaper article and diagnosed conicts between intergenerational plans for climate policy and innovation
power of democratic processes [12]. It is about no less than the tense relationship
between sustainability and democracy. Sustainability has become a dogma in the
political discourse, Weizscker bemoans. The same applies to democracy as a
form of government. And he asks whether these two principles are compatible
with each other. Sustainability policy would only be reasonable if practiced worldwide and in the long term. However, it would not be in accordance with the idea
of sovereignty of people, applied to future generations, if the present generation dictated [to] their successors the climate policy they would have to practice. Weizscker
argues that this is a usurpation of the decision-making power of the successors by
decision makers of today. A sustainability policy that binds future generations thus
contradicts the principle of democracy. The principle of the sovereignty of people
perfectly entitles the following generations to lead ad absurdum the current climate
policy by not continuing it. And he speculates that although todays climate policy,
which serves the majority opinion, is focused on CO2 reduction, it could soon turn
into a minority opinion, for instance due to a transition to geo-engineering or to
the control of methane gas and nitrous oxide emissions as a matter of priority.
Now, this could be dismissed as ridiculous, but such a verdict will not do justice
to such a serious discussion. What is important is a discussion at eye level, the unagitated rebuttal of the trivializing arguments. What Weizscker forgets or omits is the
following: The absolutely powerful principle of sovereignty of people he referred
to must end where the existence of millions of people or even their right to life is
threatened. Its weighing against a theoretical principle would be neoliberal and
cynical.
He correctly states that scientic evidence does not represent eternal truths, but
could be overthrown by additional research results at any time. In support, he
cites Paul Crutzen, who declared biofuels to be counterproductive because plant cultivation releases nitrous oxide, which harms the stabilization of the climate due to its
ability to produce 300 times more heat than CO2. Weizscker incorrectly asserts a
still high uncertainty about the real climate effects of the productive activities of
man and that the currently overwhelming majority . . . must therefore not grant
itself the power to establish its policy recommendations as sustainability dogma
for future generations. Climate deniers will be glad. However, it should be mentioned that he makes proposals that seem to be quite useful for example, uniformly
applicable prices for emission licenses determined by the states (e.g. in the form of
taxes) for the global standardization of the CO2 regime under continuous adaptation
to the respective level of knowledge.
However, he criminally plays down the problem when he hopes that future generations will more easily acquiesce to a not avoided climate change than generally
believed today. The . . . concerns . . . of ones own generation should not be overestimated for the course of world history. That is easily said for someone living in
Germany, where, if the worst comes to the worst, climate change will bring more
Sptlese in viticulture and palm trees at the Baltic Sea somewhere else, however,
the term climatic disaster would be absolutely realistic.
56
Those who talk about the primacy of the democratic principle over the sustainability dogma, as Weizscker does, either do not understand the hierarchy of
human and natural rights (which cannot be imputed to Weizscker) or deliberately
ignore preferential rights (life, integrity). The fact that according to Weizsckers
argumentation, the long-term policy of sustainability would affect the democratic
autonomy of decision making can well be classied as cura posterior in view of
the imminent disaster. We do not need recourse to future generations the individual disasters and system collapses threatening the world are threatening it now and
will continue to do so in the next two decades. We will live to see it all. On July 21,
2011, the United Nations Security Council agreed on a so-called Presidential Statement after days of erce discussions: in the text submitted by Germany, the 15 states,
for the rst time ever, admitted that global warming could represent a threat to world
peace (Negative consequences of climate change could enhance already existing
threats to peace and security in the world.)
The previous discourse shall be an example of the fact that neither scientists
Weizscker argues (apparently?) scientically nor the journalistic interpreters, or
the politicians or economic leaders called upon to decide and act consistently, should
arrogantly consider themselves the holders of the truth. As previously discussed,
some members of the IPCC gave this impression inadvertently, which does not
alter the fact that a stop must be put to this overanxiousness. First of all, some
straight talking with climate deniers is required; they must be taken seriously.
When they ask for evidence for the upcoming climate change only to say that if it
cannot be proven it does not exist, we must reply that this call for proof is unfair
and clarify the nature of scientic inquiry to those laypersons (i.e. that science
seeks to disprove rather than prove hypotheses). Nobody who argues seriously,
not even the IPCC, has ever claimed that there is evidence for an imminent climate
disaster. There are merely very high probabilities. And they are sufcient to increase
the pressure for corrective actions.
On the occasion of a speech in Munich, Robert Schlgl bemoaned that politics was
only insufciently successful in conveying information, establishing framework conditions and moderating the long-term dialogue in society within the context of the
climate discussion that followed Fukushima. Moreover, Schlgl exhorted politicians
to nally take over the command and not to cede the eld of energy policy to engineers
and scientists. As one newspaper reported, They are elected for that purpose [13].
Klaus Tpfer, whose entire political life is dedicated to the connection between
science, environment, and politics, held the eleventh Einstein Lecture Dahlem, entitled Scientic Knowledge Tolerance Responsibility, where he argued for the
increased integration of scientic knowledge in politics as well as for the concentration on long-term objectives [14]. The chairman of the German Ethics Committee for
reliable energy supply raised questions about transfer processes between politics and
science and suggested reection on humankind and nature: How is scientic
knowledge integrated in our society? The interaction of research, politics, and society has to work considerably better; mechanisms for reappraisal have to be developed. Negative consequences of far-reaching decisions that are referred to in
research results are too often imposed on next generations. Politics following the
dictates of short-termism demonstrates difculties in making decisions that are
1.3.2 References
57
necessary in the long run, for which sometimes short-term disadvantages have to be
accepted.
Following Tpfer, scientists and politicians have to and should be allowed to
admit their mistakes. Finally, it is absolutely impossible to decide on the basis of
complete information. If we questioned ourselves and our decisions in the way Einstein did, then tolerance in society would increase. Each decision entails responsibility. For example, so-called geoengineering (i.e. technical interference with natural life
cycles; also mentioned as an option by Weizscker) should at least be discussed in our
modern world because societies for example, in nuclear policy have to react to
potential mistakes of their neighbors, too. Tpfer demanded, above all, the early
integration of science in political decisions. With the Ethics Committee, Germany
has shown how possible interaction of science and politics can work.
In 2011, Marshall McLuhan would have been 100 years old. The early mastermind
of the media era, he warned against the discrimination of minorities: Our new world
forces us to engage ourselves and to participate. Today, whether we want it or not,
we share in the life of all the others and we are responsible for each other [15]. In the
end, only a key word remains: sustainability. It ranks before anything else, transdisciplinarily, as the criterion: What serves the future development of Earth; what harms
it? What helps reduce excesses and heal damages? What is in the interest of a better
life in future? Finally, human beings not only face the question of whether they will
survive as a biological species, but whether they will be able to survive without falling
back to a form of existence that does not seem worth living [16, p. 176].
1.3.1 Notes
1. This text was triggered by Murdochs bugging scandals.
2. And its denition applies without restriction to the looming energy crises [2].
3. The term internationally used to refer to a research and development concept designed
for maximum exibility is resilience. A resilient society is able to cope better with external
inuences and internal change; cf. [10, p. 8].
1.3.2 References
1. Greiner U. Medienabhrskandal, Die Zeit, 2011. Available from: http://www.zeit.de/2011/
29/01-Medien-Abhoerskandal.
2. Kronberger H. Geht uns aus der Sonne: Die Zukunft hat begonnen. Wien: Uranus; 2011.
3. Available from: http://www.philolex.de/popper.htm.
4. Hirscher G., Korte K-R. Information and decision, communication management of the
political guidance. Wiesbaden: Westdeutscher Verlag; 2003.
5. The Changing Earth [conference]; 2009 November 23; Berlin. Available from: http://
www.thechangingearth.de.
6. Available from: http://www.hm-treasury.gov.uk/stern_review_report.htm.
7. Green Recovery: Internationale Manahme fr Wachstum und Klimaschutz Innovationskonferenz IV of the Federal Ministry for Environment, Nature Conservation and
Nuclear Safety; June 22, 2009, Berlin.
58
8. Stern N. Climate lecture on the occasion of the presentation of the honorary doctorate of
the TU; April 11, 2009, Berlin.
9. Stern N. The global deal. London: The Bodley Head; 2009.
10. German National Academy of Sciences Leopoldina National Academy of Sciences.
Energy- and research-policy recommendations following the events in Fukushima. 2011.
Available from: http://www.leopoldina.org.
11. Ethics Commission for a Safe Energy Supply. Germanys energy transition: a collective
endeavour for the future. Berlin 2011.
12. Wie vertragen sich Nachhaltigkeit und Demokratie? Konikte zwischen generationenbergreifenden Plnen der Klimapolitik und der Innovationskraft demokratischer Prozesse. Neue Zrcher Zeitung. 2010 January 20.
13. Klare Positionierung, Landkreis-SPD sieht in der Energiewende viel Whlerpotenzial.
Sddeutsche Zeitung. 2011 July 16.
14. Bedrftig D. Abschied vom Diktat der Kurzfristigkeit. In: Campusleben. Available from:
http://www.fu-berlin.de/campusleben/campus/2011/110621_toepfer/index.html.
15. Marshall McLuhan Der Magier. Die Zeit. 2011 July 21.
16. Meadows DH, Meadows DL, Randers J, Behrens III, WW. The Limits to Growth. New
York: Universe Books; 1972.
2.1.1 Introduction
Todays economy is dominated by fossil fuels including oil, natural gas, and coal as
feedstocks for production of heat, energy, and chemicals. Various studies discussing
the potential availability of remaining fossil resources present rather different results.
Nevertheless, current studies estimate natural gas and coal can be utilized for another
100 years or even longer, while depletion of crude oil is expected to occur within
40 years [1]. Consequently, solutions for the sustainable production of energy in
the future need to be established, and promising technologies that include water
and wind power, solar thermal energy, and photovoltaic systems already exist. However, alternative carbon sources for the future production of chemicals and liquid
fuels are indispensable. Although electromobility via high-performance batteries or
fuel-cell systems are suitable for application in short-distance individual transportation, future air trafc and heavy duty vehicles will inevitably rely on liquid fuels,
which possess the high energy density required to guarantee sufcient cruising range
and performance.
Herein, an account on recent developments in the eld of biofuel production is
given, and technological challenges with regard to conventional biofuels such as biodiesel and bioethanol are discussed. Advances in the elds of syngas and FischerTropsch technology are only illustrated briey. Instead, major emphasis is given to
novel approaches aimed at the controlled (chemo-)catalytic transformation of lignocellulose into potential alternative biofuel compounds as energy storage media for
mobile applications.
60
Table 2.2.1 Cruising range with biofuels derived from 1 ha cropping area. (Table adapted
Natural resource
Biogas
67,600
Straw
64,000
Rapeseed oil
Rapeseed
23,300 (17,600*)
Biodiesel
Rapeseed
23,300 (17,600*)
Bioethanol
Corn
22,400 (14,400*)
Car mileage: Otto, 7.4 l/100 km; diesel, 6.1 l/100 km.
*Biogas production from crop remaining.
61
With regard to bioethanol, conventional processes are based on enzymatic hydrolysis of starch, which yields glucose that can be converted to ethanol via fermentation.
Signicantly, ethanol is an inhibitor for the producing microorganisms, allowing only
limited product concentrations. This results in diluted aqueous ethanol solutions. The
low product concentration, together with azeotrope formation, means that product
separation requires a signicant energy input. With the aim of optimizing bioethanol
production, numerous scientic investigations have been conducted focusing on
alternative separation technologies, bioengineering to accelerate product formation,
and increasing the accessible ethanol concentrations to reduce the overall energy
demand. In terms of fuel properties, even dry ethanol has only roughly two-thirds
of the energy content per unit volume found for gasoline. Nevertheless, oxygenates
such as ethanol increase knock resistance and reduce exhaust emissions making
them suitable for integration into todays gasoline technology. Ethanol can be
blended into gasoline in amounts up to 15% without the need for any engine modication. Alternatively, 85% ethanol blends can be used in modern ex cars that are
now commercially available.
Starting from the 1980s, Brazil has pursued a national strategy for ethanolic fuels
called the Prolcool program to abate the countrys strong dependence on fossil-oil
imports. However, in the 1990s, increasing world market prices for sugar nearly disrupted Brazils ethanol production, although in recent years a signicant upturn in
bioethanol production has again occurred. Until 2005, Brazil was the biggest producer and consumer of bioethanol derived from sugarcane. Combustion of residual
plant material for energy and heat production results in a net positive energy and
CO2 balance for such bioethanol. Interestingly, in recent years, the United States
has taken over as the largest bioethanol producer reaching production capacities
around 26 billion liters (2009) mainly derived from corn and crops. However, this
increase in production has been accompanied by governmental funding and has
furthered competition between food and fuel production. This precipitated the Tortilla Crisis in Mexico during 2007, when prices for corn-based ground staple food
increased by up to 400%. Consequently, a careful choice of suitable feedstocks is
essential with regard to biofuel yields, and competition between food and fuel
production must be avoided.
The potential yields per hectare cropping area for different biomass-derived biofuels are illustrated in Table 2.1.1. Biogas, which is not further discussed in this overview, is produced via anaerobic digestion or fermentation of energy crops such as
maize, silage, or biodegradable wastes, including sewage sludge, and food wastes.
The composition of biogas may vary depending on feedstock and microorganism;
nevertheless, it is mainly composed of methane (50%75%), followed by carbon dioxide (25%50%) and traces of other compounds such as N2, O2, H2S, H2, or NH3.
The data presented in Table 2.1.1 clearly illustrate the advantage of transformation technologies utilizing the whole plant material and not only the oil or sugar
and starch fraction.
An alternative to the utilization of food crops in biofuel production is lignocellulose, which can for instance be utilized in biomass-to-liquid (BTL) processes. That
way, a direct competition between food and fuel production can be avoided. This
feedstock is much more abundant than vegetable oils or sugar and starch crops.
62
Moreover, lignocellulose is not edible and could theoretically be utilized without any
impact on food production. The cellulose and hemicellulose fraction of lignocellulose
may serve for the production of cellulosic ethanol, which could be produced via acid
or enzymatic catalyzed hydrolysis of cellulose, followed by further fermentation to
yield ethanol. Alternatively, the whole plant can be gasied to yield syngas, followed
by methanol or dimethyl ether synthesis or Fischer-Tropsch technology that produces hydrocarbon fuels. Furthermore, controlled (bio-)chemical transformations
to novel fuel compounds based on cellulose, hemicellulose, or lignin are possible,
and numerous recent publications emphasize intense research in this direction.
OH
HO
OH
HO
O
OH
O
HO
HO
O
OH
OH
OH
OH
O
63
OH
HO
HO
HO
Cellulose
OH
O
O
HO
O
Lignin
HOOC
O
H3CO
HO
OH
OH
O
HO
HO
O
OH
O
O
OAc
O
O
O
HO
OAc
HO
O
OH
O
OH
OH
HO
HO
HO
HO
O
OH
Hemicellulose
Figure 2.1.1
conventional solvent system. Thus, enzymatic processes still suffer from rather low
space-time yields. There are also disadvantages associated with chemical hydrolysis,
which like traditional wood saccharication requires harsh conditions utilizing concentrated mineral acids that have high corrosion potentials and the capacity to form
undesirable salts.
Consequently, the controlled (bio-)chemical breakdown of lignocellulose, and particularly cellulose, requires the development of tailor-made catalysts. Furthermore,
new solvent systems will be required to facilitate future renery technologies based
on lignocellulosic feedstocks [13].
64
energy density than gasoline, the oxygen content of ethanol allows for more efcient
combustion, enabling similar overall fuel efciency. Additionally, exible fuel vehicles are now commercially available and are capable of utilizing gasoline or fuel
blends comprised of 85% ethanol and 15% gasoline. Currently there are approximately 8 million exible fuel vehicles in the United States, a reality that demonstrates
the compatibility of bioethanol with current automobile technologies and fuel distribution networks [17]. While bioethanol offers numerous advantages, it should be
noted that there are also potential disadvantages associated with this fuel additive,
with bioethanol being linked to a decrease in urban air quality, contamination of
water resources, and a reduction in biodiversity [18].
The production of bioethanol requires the sequential processing of biomass into
fermentable reducing sugars, fermentation of these sugars to ethanol, and then a separation of ethanol from the fermentation mixtures. Overall the net energy balance
(i.e. the energy required during processing vs. the energy obtained from bioethanol)
has been found to be highest in systems based on sugarcane and starch feedstocks [19].
This fact is arguably demonstrated by the success of bioethanol production in Brazil
from sugarcane. The capacity of bioethanol to partially replace gasoline was recognized by the Brazilian government in the 1970s, and a political decision was made to
stimulate bioethanol production [20]. Although bioethanol production was initially
subsidized, during the 1990s these subsidies were removed; in 2004, Brazilian
bioethanol became commercially competitive with gasoline on international markets.
This achievement can be attributed to both the rising cost of crude oil and efciencies
made in bioethanol production, which in 2007 reached 16 billion liters, requiring
5.6 million hectares of land. Overall, this only required 1% of the land available
for agriculture in Brazil, meaning there is signicant scope for expansion. This
could prove to be desirable as analysis indicates Brazilian bioethanol has an energy
yield ratio of 7.9 (bioenergy yield to fossil energy input), and each hectare of land
farmed annually could mitigate CO2 production by approximately 25,000 kg per
annum [21]. In contrast to bioethanol obtained from sugarcane, bioethanol produced
from corn in the United States offers lower overall net benets [22]. This is problematic as in 2006, U.S. production of bioethanol reached 1.2 billion liters, with virtually
all transportation ethanol being obtained from corn. With an associated energy yield
ratio of just 1.3, it is clear that the energy obtained from corn-derived bioethanol
is minimal. Furthermore, some high-prole analysis suggests that bioethanol production in the United States is leading to a substantial net increase in greenhouse
gas emissions [23]. Alternative analysis indicates that the greenhouse gas emissions
associated with maize-derived bioethanol can vary signicantly based on plant
design [24]. For example, processing based on using coal as a heat source results
in a 3% net increase in greenhouse gas emissions. In contrast, utilization of woodchips as a fuel in processing results in a 52% decrease in net greenhouse gas emissions. Evidently, more research is needed to reconsider the different methodologies
used to produce ethanol from corn, with the aim of developing more sustainable
processes.
While production of bioethanol from tropical sugar-rich crops has been demonstrated to be efcient, the expansion of such industries could eventually impact the
food supply. With this in mind, a substantial research effort has been focused on
65
66
67
different chain lengths being present in the vegetable oils of alternative plants. For
example, biodiesel fuels obtained from palm oil and soybeans have very different
viscosities, cloud points, and combustion properties, making these fuels suitable
for different applications. Furthermore, the vegetable oil feedstock has a signicant
impact on the economics of biodiesel production; as for rst-generation biodiesel, the
price of the feedstock is estimated as contributing up to 70%90% of the total operating costs. In 2007, it was reported that biodiesel could be produced from European
oilseed rape or Malaysian palm oil at a cost of US$0.80 per liter gasoline equivalent,
an amount that compares favorably with the US$0.80 per gallon production costs
associated with petroleum diesel in 2008. One factor that reduces the economic attractiveness of biodiesel production is the low value of the by-product glycerol,
which is typically obtained in low purities. Despite these difculties, the sale of glycerol
has been shown to result in a marginal decrease in biofuel production costs. Alternatively, glycerol could form the basis of a range of value-added chemicals including fuel
additives. The development of such products could further enhance the economic viability of biofuel production, by adding value to an otherwise low-value commodity
by-product and increasing the atom efciency of the overall process.
Although biodiesel production is currently being achieved by the utilization of
homogeneous base and acid catalysts, such approaches have a number of key disadvantages. For instance, systems based on homogeneous catalysts often require an
excess of alcohol and are more suited to less economic batch processing and additional separation steps. Secondly, solubilized homogeneous catalysts often migrate
into the glycerol by-product, which makes catalyst recovery and reuse problematic.
Solubilized homogeneous base catalysts are also particularly susceptible to saponication. In contrast, solid catalysts are more easily separated from reaction mixtures
and thus offer higher reusability and suitability for continuous processing. Solid catalysts also tend to have higher selectivity for biodiesel production and are more tolerant
to the water and FFAs that are invariably present in the biodiesel streams. Lastly, heterogeneous catalysts facilitate the production of glycerol in higher purities than homogeneous systems (99% vs. 85%), enabling the production of a more utilizable byproduct [34]. Although these advantages make heterogeneous catalysis highly attractive, it should be noted that in general solid catalysts tend to be less active, with systems
based on heterogeneous catalysis often requiring the use of higher temperatures and
pressures to achieve high conversions. One factor that can inhibit heterogeneous catalysis is the low extent of catalyst surface area available for the initiation of transesterication. For an effective solid catalyst to proceed, the catalyst surface cannot become
saturated with polar molecules such as water or glycerol. With this in mind, the most
effective catalysts tend to have hydrophobic surfaces that preferentially absorb triglyceride substrates. Overall, the solid catalyst must consist of a robust material that can
resist decomposition. Examples of materials currently being considered for biodiesel
production include basic zeolites, hydrotalcites, metal oxides, insoluble basic salts,
immobilized organic bases, supported basic oxides, and alkali earth oxides.
One approach that could enable the production of cheaper biodiesel is to utilize
low-value vegetable oils that contain a high content of FFA. This is particularly
attractive as such materials can be obtained from either waste or the cultivation of
alternative crops, something that would increase the biodiversity from which biofuels
68
are obtained. Signicantly, vegetable oils that contain mixtures of FFAs and triglycerides are incompatible with the base catalysts required for triglyceride transesterication, due to the propensity of FFAs to undergo saponication. Instead, such oils
would rst require treatment with alcohols in the presence of an acidic catalyst, with
the aim of converting the FFAs to the corresponding esters. Once the FFA content of
the oil has been removed, the triglyceride portion can then undergo base-catalyzed
transesterication without the undesirable formation of soaps. Such a two-step process requires the combination of acid and base catalysts, which is most easily
achieved when the initial esterication step is catalyzed by a solid acid. Furthermore,
utilization of solid catalysts for both steps has been proposed to facilitate more
efcient continuous processes that do not suffer from catalyst neutralization or the
production of signicant wastewater streams. Indeed, it has been suggested that
such a continuous process could enable a 40%50% reduction in capital costs and
a 30%60% reduction in energy compared to batch processes. With this in mind,
the identication of suitable esterication solid acid catalysts is increasingly being
viewed as an integral part of biodiesel development.
Esterication and transesterication are by no means the only methodologies by
which fuels can be obtained from biologically derived oils. An alternative route by
which the glycerol fragment of triglycerides can be removed is by thermocatalytic
cracking, a process that can render diesel-like oils. This has the advantage of producing more stable nonester fuels that have higher energy contents, and desirable combustion properties, as well as physical properties that make them compatible for use
as transportation fuels. Examples of alternative fuels that can be produced by this
approach include nonesteried renewable diesel and hydrotreated renewable jet.
The possibility of producing jet fuels from biomass is of interest to the U.S. military,
which has explored cracking oils derived from biomass in the presence of hydrogen.
Such processing enables deoxygenation and decarboxylation of triglyceride, producing less polar compounds more suited for use as jet fuels. In 2009, the U.S. military
announced that this technology would be utilized in the EniEconning process,
which will supply 2.23 million liters of jet fuel to the U.S. military. Such processes
cannot be viewed as being fully sustainable as they currently are reliant on hydrogen
that is obtained from nonrenewable resources. However, the EniEconning process clearly demonstrates that a range of methodologies can be employed to convert
oils derived from biomass into utilizable fuels.
OH OH
OH
OH
OH
HO
HO
OH OH
Levulinic acid
Xylitol
OH
OH
OH
Lactic acid
O
OH
HO
OH
OH OH
Sorbitol
OH
69
EtOH/BuOH
HO
O
O
Itaconic acid
5-Hydroxymethylfurfural
Furfural
Figure 2.1.2
generation of such value-added commodities could optimize the overall process economics. Figure 2.1.2 illustrates promising platform chemicals derived from cellulose
and hemicellulose that have been transformed into potential fuel compounds.
A detailed account of recent reports concerning the synthesis of such platform chemicals is given in Chapter on synthesis of platform chemicals from biomass, which discusses chemicals based on renewable feedstocks, and several recent reviews [3739].
Signicantly, boundary conditions relating to the potential production of platform chemicals should be set, and the following rhetorical question should always be considered:
What would future scenarios for the production of these molecules entail? With regard
to sustainable chemical processes, Anastas et al. have published a clear denition of the
requirements for sustainable chemical processes in the frame of green chemistry [40, 41].
Furthermore, aspects of health and safety, environmental protection, and economics
have been considered and summarized by Poliakoff et al. (see Figure 2.1.3):
Green Chemistry is the design of chemical products and processes that reduce or eliminate the
use and generation of hazardous substances [42].
Sustainable development can be broadly dened as those that meet present needs
without compromising the ability of future generations to meet their needs [43].
Alongside the principles of green chemistry and sustainable processing, this denition
provides a clear guideline for future technology developments. Many of the conditions associated with green chemistry and sustainable development can be met if catalytic processes and extensive process integration are utilized. Consequently, the
development of suitable catalysts and efcient reaction systems will be indispensable
in facilitating future sustainable transformation technologies. Thus, novel developments in the eld of biomass processing clearly have to be benchmarked against
these requirements. When designing new processes, emphasis must be placed on optimizing reaction temperature, solvent systems, hydrogen requirements, and carbon efciencies. In general, optimized conversions and selectivity are desirable. However,
the overall efciency of a chemical transformation can be heavily inuenced by
70
Degradable products
Innocuous solvents
Atom efficiency
Sustainability
Shorter synthesis
Waste prevention
Pollution prevention
Energy efficiency
Environmental
protection
Economics
Catalytic reagents
product separation and workup, and so these processes must always be considered.
These points should already be taken into account during basic research to facilitate
implementation of efcient technical processes. Interestingly, recent studies emphasize that chemical disintegration of lignocellulose into the individual biopolymers,
and the subsequent depolymerization into sugars and single aromatic motifs, will
be of great importance to the overall process economics.
Todays reneries are based on utilization of crude oil. Consequently, technologies and catalyst systems for the treatment of oxygen-decient molecules in hightemperature gas-phase processes have been established. In contrast, the main
constituents of lignocellulose are rich in oxygen and require chemical transformations in liquid-phase processes at rather low temperature. Additionally, neither cellulose nor lignin is soluble in conventional solvents, which further complicates their
controlled chemical breakdown. Figure 2.1.4 illustrates the different temperature
and pressure ranges of typical transformations in todays reneries compared to
potential technologies for the controlled utilization of biomass [44].
These new technologies mostly serve to increase the carbon chain length and decrease
the oxygen content of typical lignocellulose-based platform chemicals to deliver molecules of high energy content and suitable evaporation and combustion properties.
Overall, three main approaches to the controlled low-temperature transformation of
lignocellulose into biofuels can be distinguished (Figure 2.1.5). Besides the transformation of lignocellulose into bioalcohols, which has already been discussed, several
studies focus on the production of simple hydrocarbons to feed conventional petrochemical supply chains and todays automotive combustion systems [45]. However,
only a few approaches address the production of novel, potentially oxygen-containing
biofuel candidates [46, 47].
71
Supercritical
water gasification
Liquid phase
200
Hydrogenation
Hydrogenolysis
Vapor phase
p /atm
150
Liquefaction
Petrochemical
processes
100
50
Reforming
in water
Isomerization
Aldol-condensation
oxidation
673
873
Dehydration
Hydrolysis
1073
1273
T/K
Figure 2.1.4 Schematic representation illustrating typical temperature and pressure ranges
of transformations in todays renery in comparison to reactions utilizing renewable feedstocks (Adapted from [44]). (Reproduced by permission of the Royal Society of Chemistry.)
Biopolymers
Platform chemicals
Xylose
Hemicellulose
Xylitol
Furfural
Cellulose
Glucose
Sorbitol
5-HMF
IA
EtOH/BuOH
Biofuels
Chemical transformation
into hydrocarbons
Figure 2.1.5 Potential platform chemicals considered for the controlled chemical transformation into fuel compounds and possible approaches to obtain novel biofuel motifs
(5-HMF, 5-hydroxy methyl furfural; LA, levulinic acid; IA, itaconic acid; EtOH, ethanol).
72
C6H14 + 6 H2O
C6O6H14 + 6 H2O
6 CO2 + 13 H2
C6O6H14 + 6 H2
______________________________________________
19/13 C6O6H14
Based on these results and the disadvantageously high volatility of hexane, Huber
et al. decided to develop another approach utilizing 5-hydroxymethylfurfural (5-HMF)
or furfural, compounds that can be obtained from the dehydration of glucose or
xylose, respectively [5457]. The production of C8-C15 alkenes from these furans is
achievable by sequential aldol condensation with acetone followed by hydrogenation
73
and hydrodeoxygenation. These products are in the desired range for diesel and jet fuel
applications. Maximum yields were accomplished by applying different catalysts and reaction conditions in each reaction step. An overall process analysis starting from aqueous
solutions of waste hemicellulose indicate that the utilization of a four-step process is optimal. Such a process would comprise (1) combined acid hydrolysis and xylose dehydration,
(2) aldol condensation, (3) low-temperature hydrogenation, and (4) high-temperature hydrodeoxygenation. In the rst step, xylose oligomers are hydrolyzed and further dehydrated to furfural, which is continuously extracted from the aqueous solution into a
suitable organic solvent. Next, aldol condensation of furfural and acetone is catalyzed
by homogeneous NaOH or solid MgAl-oxides, enabling the formation of unsaturated alkane precursors. Finally, these thermally unstable compounds are hydrogenated at low
temperatures (110C130C), using 5 wt% Ru\C and H2 pressures of 55 bar, a procedure
that yields saturated thermally stable products. In the nal step, further hydrodeoxygenation over bifunctional 4 wt% Pt/Al2O3-SiO2 catalysts at 260C and 62 bar H2 enables the
production of jet and diesel fuel range hydrocarbons. These conditions allow alkane
yields of approximately 91% with tri- and dodecanes produced as main products.
Experimentally, 0.46 kg jet fuel per kg of xylose could be produced, which corresponds
to 76% of the theoretical yield (0.61 kg/kg). Interestingly, a sensitivity analysis emphasized that the price of the raw materials, the organic-to-aqueous mass ratio in the biphasic dehydration step, and the xylose concentration in the feed solution signicantly
affects the product cost, while investment and operation costs have less impact.
Recently, Corma et al. followed yet another approach aimed at producing branched
C15-alkanes [58]. In this investigation, 2-methylfuran served as the substrate undergoing solvent-free hydroalkylation and hydrodeoxygenation. 2-Methylfurfural can be
derived from furfural [5961], and can also be produced in the industrial production
of furfuryl alcohol [62], with 93% selectivity simply by raising the reaction temperature from 135C to 250C [63]. Hydroalkylation of 2-methylfurfural with butanol
can be catalyzed by soluble para-toluolsulfonic acid or solid Amberlyst-15. The
hydrodeoxygenation of the obtained 1,1-bisylvylalkanes (1,1-bisylvylbutane) was performed without any additional solvents, using a continuous xed-bed reactor alongside the catalysts Pt\C or Pt\Al2O3. Overall, 90% of all carbon ends up in the organic
phase with 95% selectivity to alkanes. Furthermore, other approaches allow for the
production of alkanes from 2-methylfuran, without the need for any additional reagents. In this system, 4-oxopentanal is generated in situ from 2-methylfuran via
acid-initiated ring opening. This intermediate then reacts with two other molecules
of 2-methylfuran to produce a trimer. Finally, hydrodeoxygenation of the intermediate yields pentadecane isomers as the main products, and overall fuel yields of up
to 87% are generated.The obtained organic fraction exhibits an excellent pour
point (90C) and cetane number (70.9), making this mixture suitable as a fuel. Alternatively, the product mixture could serve as a valuable blending agent, the utilization
of which could improve both the cetane number and low-temperature ow properties
of other fuels. Notably, all of these methods are demanding in hydrogen, a factor that
must be considered before their widespread utilization (Figure 2.1.6).
The utilization of levulinic acid (LA) as a platform chemical could enable the
implementation of routes to biofuels that demand less hydrogen. Furthermore, the
formic acid released during LA formation may serve as a hydrogen source. This
74
Cellulose
[H+]
H2
Glucose
HO
OH OH
HO
OH OH
HO
C1 C6 Alkanes
OH
Sorbitol
HO
Hydrodeoxygenation
H2O
OH
Isosorbide
OH
Hydrodeoxygenation
C9 C15 Alkanes
5-HMF
Oligomerization
O
O
OH
O
Vl
LA
C8 + Alkanes
Butenes
OH
C9 Alkanes
Valeric acid
5-Nonanone
O
OH
OH
Dehydration/Hydrogenation
C-C Coupling
OH
C4 C7 Alcohols
C4 C7 Ketones
Lactic acid
Hemicellulose
[H+]
Xylose
O
O
C8 C13 Alkanes
Furfural
O
C4H9
C9 C15 Alkanes
2-Methylfurane
Figure 2.1.6 Graphic summary of alternative approaches to alkanes starting from cellulose
75
mixture is oligomerized over solid acids such as Amberlyst-70. The reactions may be
integrated into a single process, which runs at 36 bar pressure. In the rst reactor
unit, a temperature of 375C is maintained, while in the second unit a lower temperature of 170C is employed. This approach avoids the need for additional compression or thermal energy. Overall, the system gives a 77% yield for C8+ alkenes (based
on the initial amount of Vl used), of which C8-C16 alkenes represent the major fraction [68, 69]. The main advantages of this approach are (1) no requirement for external hydrogen or other substrates, (2) potential integration into a single continuous
process, (3) moderate reaction conditions, and (4) the utilization of simple solid
acid catalysts that do not contain any precious metals. Notably, this system releases
CO2 in pressures of up to 36 bar. This by-product could be used for sequestration or
chemical synthesis without any need for additional compression energy.
Currently, the efcient dehydration of carbohydrates into LA limits overall process efciencies. However, several investigations have identied solid acid catalysts
that offer promising selectivities [7072]. Despite these advances, the use of sulfuric
acid catalysts for the hydrolysis and dehydration of cellulose remains superior with
regard to space-time yields, despite the associated problem of salt formation [73]. Consequently, potential product streams resulting from cellulose dehydration would comprise sulfuric acid containing an aqueous solution of LA. Therefore, LA would have to
be separated from the aqueous sulfuric acid solution before further transformation.
With this in mind, Grbz et al. developed the extraction of LA via etherication
with butane, which allows the development of an integrated process concept [74].
Alternatively, Braden et al. developed a sulfuric acidresistant catalyst for the transformation of LA into Vl. Conventional catalysts such as Ru/C can facilitate the
transfer hydrogenation of LA into Vl, utilizing formic acid as the hydrogen source.
However, Ru/C suffers from fast deactivation in the presence of sulfuric acid, prompting the development of a bimetallic RuRe/C catalyst (4:3 molar ration of Re to Ru).
This bimetallic catalyst offers superior activity for the simultaneous hydrogenation of
LA and decomposition of formic acid, as well as displaying good stability in the presence of sulfuric acid [75]. Interestingly, a techno-economic analysis of a complete process utilizing lignocellulosic feedstocks to produce liquid hydrocarbons, in comparison
to a state-of-the-art lignocellulosic ethanol production strategy, revealed that feedstocks
and total installation costs are the most sensitive process parameters. Although the proposed process appears to be economically viable, further investigations are required to
improve both the production of LA and its transformation into Vl.
Another means by which the oxygen content of LA can be reduced is conversion to
valeric acid. This is achieved by transformation of LA to Vl followed by ring opening
and hydrogenation over a water-stable bifunctional Pd/Nb2O5 catalyst at moderate
temperatures and pressures. This transformation could prove to be advantageous, as
it results in a hydrophobic and less reactive intermediate that is suitable for further processing [76]. One means by which valeric acid can be further modied is ketonization,
which generates 5-nonane and stoichiometric amounts of CO2 and water. The reaction
conditions can even be adapted to higher temperatures and lower weight hour space
velocity (WHSV), which enables both valeric acid production and ketonization
to be initiated by the same Pd/Nb2O5 catalyst in a single xed-bed reactor (350C,
35 bar, 0.1 h1). An organic layer that spontaneously separates from the aqueous
76
50 wt% Vl feed is produced, accounting for 81% of the reacted carbon with 60% carbon yield of C9 ketones together with unreacted valeric acid, C45 ketones, and lower
ketones. Although this strategy reduces the number of reaction steps and could reduce
capital and operating costs in a potential technical application, higher yields are
reached by separating both steps and individually optimizing catalyst and reaction conditions. In such systems, valeric acid is formed over Pd/Nb2O5 at 325C, 35 bar, and
1.2 h1 with a 95% carbon yield in the organic phase and 92% selectivity to valeric
acid. Next, valeric acid can be converted to 5-nonane via decarboxylation, which is catalyzed by a ceria-zirconia catalyst at 425C, 20 bar, and 1.1 h1 [7781]. In this cascade approach, a 5-nonane yield of approximately 84% is obtained. The product,
5-nonanone, can be further converted via hydrogenation/dehydration over Pt/Nb2O5
to linear nonane, which possesses a good cetane number and lubricity, making it suitable
for use as a diesel fuel blender. Additionally, 5-nonanone can be hydrogenated over zeolites such as USY, yielding 5-nonanol. This alcohol can then undergo dehydration/
isomerization to give hydrocarbons that are suitable for use as gas components [8283].
In addition to synthesis strategies based on sorbitol, 5-HMF, and LA, investigations have been made examining the conversion of lactic acid into hydrocarbons.
Lactic acid can be catalytically upgraded via combined dehydration/hydrogenation
and CC coupling reactions [8485]. In such systems, an aqueous solution of lactic
acid is processed over Pt supported on Nb2O5 causing the formation of an organic
layer composed of propanoic acid and C4-C7 ketones, which spontaneously separates
from the aqueous feed solution. Through subsequent hydrogenations, C4-C7 alcohols
are generated, which are suitable for use as liquid fuels and have high energy densities. It has been demonstrated that the niobia support catalyzes both the dehydration
and CC coupling reactions, while platinum provides catalytically active sites suitable
for hydrogenation. Adjustment of the reaction conditions even allows further oxygen
removal, again delivering hydrocarbon compounds as potential biomass-derived fuel
compounds. Signicantly, this example presents a transition between attempts purely
targeting the production of hydrocarbons in controlled transformations and the
utilization of similar approaches for the production of biomass-derived oxygenates
suitable for use in novel biofuel compounds.
Certainly, various alternative approaches are possible to deliver liquid hydrocarbons based on cellulosic feedstocks. However, various factors including the overall
yield, energy requirements, ease of product separation, and not least the demand
for additional hydrogen will determine the economic and ecological competitiveness
of these approaches. Nevertheless, as discussed previously, oxygenates exhibit several
advantages as fuels (e.g. knock resistance and reduced exhaust emissions). These advantages have furthered the integration of ethanol and lately ethyl tert-butyl ether
into todays gasoline chains.
77
78
79
Cellulose
[H+]
Glucose
+ ROH
H2O
OR
O
HO
O
H2O
+ MeOH
O2
5-HMF
+H2
H2O
+H2
2,5-DMF
OH
O
CO
2,5-DMTHF
+H2
+H2
H2O
2-MTHF
2-MF
+H2
O
O
Furfural
+ ROH
O
OR
O
+H2
H2O
OH
O
LA
Vl
+H2
H2O
OH
O
Valeric acid
OR
+ ROH
Figure 2.1.7
80
could open interesting perspectives to produce valeric biofuels by subsequent selective hydrogenation.
In another approach, glucose is dehydrated to 5-HMF, which undergoes
further hydrodeoxygenation or hydrodeoxygenation and hydrogenation to 2,5dimethylfuran (2,5-DMF) or 2,5-dimethyltetrahydrofuran (2,5-DMTHF). The synthesis of DMF has been investigated by Yang and coworkers, who developed a
multifunctional catalyst system for the production of DMF from biomass-derived
carbohydrates, such as fructose, glucose, and cellulose. Furthermore, DMF was produced from raw lignocellulose in the form of corn stover. Notably, DMF exhibits an
energy density of 31.8 MJ L-1 and a boiling point of 90C92C and is immiscible in
water, properties that make DMF suitable for use in fuel applications. It has been
demonstrated that 2,5-DMTHF can also be derived from carbohydrates obtained
from biomass, including cellulose. Combinations of rhodium chloride, HCl, NaI,
and benzene facilitated the conversion of cellulose into DMTHF, which was obtained in yields as high as 76% at 160C, 16 h, and 20 bar H2. When this system
is applied to untreated corn stover, DMTHF is again obtained, albeit in a lower
yield of 41%.
Due to their good storage and transport stability, the utilization of saturated
compounds as biofuels is receiving the most attention, while potential applications
of furanes are being relatively neglected. Nevertheless, 2,5-DMF presents a promising fuel compound that exhibits 40% higher energy density compared to ethanol, a
suitable boiling point of 93C, and immiscibility with water, which facilitates product
recovery based on aqueous synthesis systems and subsequent fuel storage.
Roman-Leshkov et al. developed a concept for continuous production of 2,5-DMF
based on fructose [108]. The rst step involves the acid-catalyzed dehydration of fructose to produce 5-HMF in a biphasic reactor. Subsequently, 5-HMF is extracted by
the organic phase of the biphasic reactor and nally converted into 2,5-DMF by hydrogenolysis of CO bonds over a copper-ruthenium (CuRu) catalyst. An aqueous
30 wt% fructose solution containing hydrochloric acid served as feed, while saturation with NaCl made phase separation using 1-butanol as the extracting solvent
possible, reaching 82% selectivity of 5-HMF at 85% conversion. Hydrogenolysis to
2,5-DMF was carried out in a continuous liquid- or vapor-phase ow reactor catalyzed by CuRu/C (3:2 Cu:Ru) with 76%79% yield of 2,5-DMF for 1.5 and 10 wt%
HMF feeds.
Alternatively, a synthesis method utilizing heteropoly acids has been demonstrated. Heteropoly acids such as 12-phosphomolybdic acid (12-PMA) showed
remarkable activity and selectivity in the dehydration of glucose to 5-HMF in the
ionic liquids ethyl- or butylmethylimidazolium chloride (EMIMCl or BMIMCl),
with acetonitrile used as a cosolvent. Up to 98% glucose conversion with an HMF
selectivity of 99% after 3 h at 120C could be achieved [109]. The addition of acetonitrile to EMIMCl suppressed the formation of humins from glucose. The high
5-HMF selectivity was ascribed to stabilization of 1,2-enediol and other intermediates involved in the dehydration of glucose, avoiding polymerization reactions involved in humine formation. Additionally, 5-HMF could be converted into DMF
without isolation by simply replacing 12-PMA with palladium supported on carbon.
However, within 1 h at 100C and 60 bar H2 pressure, only a 2,5-DMF yield of
81
16% with 47% conversion of 5-HMF could be reached, leaving room for further
improvements.
Overall, one can dene a set of potential lead structures of promising biofuel candidates currently discuss in academia. Thereby, tailored catalyst development allowed the selective transformation into levulinates, furane, and tetrahydrofurane
motifs such as butyl levulinate, 2-methylfuran, and 2- or 3-MTHF as well as derivates of 5-HMF. Essential progress could be reached concerning exible and highly
selective catalysts for efcient transformation of levulinic and itaconic acid into these
targets [100]. Novel catalysts for efcient CO cleavage allowed synthesis of 2- and
2,5-furane, respectively. Nevertheless, most synthesis strategies still suffer from low
space-time yields and low product concentrations. Additionally, low hydrogen
demand and efcient product recovery will be essential to make the technical implementation of such technologies feasible. Consequently, future biorenery concepts
will strongly rely on the development of suitable catalyst systems and efcient
reaction systems.
82
2.1.10 Summary
Clearly, in the nal analysis, market demand and market price will decide whether
high-temperature processes aiming for syngas or bio-oil production, selective enzymatic transformation yielding methane or bioethanol, or tailored chemical conversion technologies via catalytic processes can deliver the required products. Based
on todays diversity in the chemical industry, various technology platforms may be
utilized to cover the required product range. Nevertheless, efcient utilization of
available carbon sources both with respect to carbon and energy content and
resource management will be decisive in developing a sustainable future scenario
for carbon-based industry.
2.1.11 Acknowledgment
We acknowledge nancial support of the Robert Bosch Foundation in the frame of the
Robert Bosch Junior Professorship for sustainable utilization of natural renewable
resources.
2.1.12 References
1. BP Statistical Review of World Energy, June 2012. Available from: http://www.bp.com/
statisticalreview.
2. Hill J, Nelson E, Tilman D, Polasky S, Tiffany D. Proc Natl Acad Sci U S A. 2006;103
(30):1120610.
3. Demirbas A. Prog Energy Combustion Sci. 2007;33(1):118.
4. Adler PR, Del Grosso SJ, Parton WJ. Ecol Appl. 2007;17(3):67591.
5. Crutzen PJ, Mosier AR, Smith KA, Winiwarter W. Atmos Chem Phys. 2008;8(2):38995.
6. Tilman D, Hill J, Lehman C. Science. 2006;314(5805) :1598600.
7. Demirbas A. Energy Convers Manag. 2008;49(8):210616.
8. Behr A, Eilting J, Irawadi K, Leschinski J, Lindner F. Green Chem. 2008;10(1):1330.
2.1.12 References
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
83
Zhou C-HC, Beltramini JN, Fan Y-X, Lu GQM. Chem Soc Rev. 2008;37(3):52749.
Ma FR, Hanna MA. Bioresour Technol. 1999;70(1):115.
Meher C, Sagar DV, Naik SN. Renewable Sustainable Energy Rev. 2006;10(3):24868.
Kamm B. Biorenery.
Palkovits R. Chem Ing Tech. 2011;83(4):110.
von Blottuitz H, Curran MA. J Cleaner Prod. 2007;15:60719.
Available from: http://www.afdc.energy.gov/afdc/ethanol/e85.html.
Browner, CM. EPA Administrator, Press Conference, March 20, 2000. Available from:
http://www.epa.gov/otaq/consumer/fuels/mtbe/press34b.pdf.
Available from: http://www.afdc.energy.gov/afdc/vehicles/exible_fuel.html.
Niven RK. Renewable Sustainable Energy Rev. 2005;9(6):53555.
Hayes DJ. Catal Today. 2008 June.
Goldemberg J. Science. 2007;315:808.
De CarvalhoMacedo I. Biomass Bioenergy. 1998;14:7781.
Shapouri H, Dufeld J, Wang M. The Energy Balance of Corn Ethanol: An Update, in
Agricultural Economic Report No. 813. United States Department of Agriculture; 2002.
Searchinger T, Heimlich R, Houghton RA, et al. Science. 2008;319:123840.
Wang M, Wu M, Huo H. Environ Res Lett. 2007;2:113.
Sun Y, Cheng J. Bioresour Technol. 2002:83:111.
Kumar P, Barrett DM, Delwiche MJ, Stroeve P. Ind Eng Chem Res. 2009;48:371329.
Galbe M, Zacchi G. Appl Microbiol Biotechnol. 2002;59:61828.
Hayes, DJ. Catal Today. 2009;145(12):13851.
Garca V, Pkkil J, Ojamo H, Muurinen, Keiski RL. Renewable Sustainable Energy
Rev. 2011;15:96480.
Drre P. Biotechnol J. 2007;2:152534.
Gabriel CL. Ind Eng Chem. 1928;20:10637.
Jones DT, Woods DR. Microbiol Rev. 1986;50:484524.
Ni Y, Sun Z. Appl. Microbiol Biotechnol. 2009;83:41523.
Centi G, Trir F, Perathoner S, Cavani F. Sustainable Industrial Chemistry. Weinheim:
Wiley-VCH; 2009.
Werpy T, Petersen G. Top value added chemicals from biomass. Oak Ridge, TN: U.S.
Department of Energy; 2004. Vol. 1.
Bozell JJ, Petersen GR. Green Chem. 2010;2(12):53954.
Huber GW, Iborra S, Corma A. Chem Rev. 2006;106(9):404498.
Corma A, Iborra S, Velty A. Chem Rev. 2007;107:2411502.
Climent MJ, Corma A, Iborra S. Green Chem. 2011;13:52040.
Anastas PT, Warner JC. Green chemistry: theory and practice. New York: Oxford University Press USA; 2000.
Anastas P, Eghbali N. Chem Soc Rev. 2010;39:30112.
Tang SLY, Smith RL, Poliakoff M. Green Chem. 2005;7:761.
World Commission on Environment and Development. Our Common Future, Chapter 2:
Towards Sustainable Development. Available from: www.un-documents.net/ocf-02.htm/.
Chheda JN, Huber GW, Dumesic JA. Angew Chem Int Ed.2007;46:716483.
Serrano-Ruiz JC, Dumesic JA. Energy Environ Sci. 2011;4(1):8399.
Huber GW, Iborra S, Corma A. Chem Rev. 2006;106(9):404498.
Alonso DM, Bond JQ, Dumesic JA. Green Chem. 2010;12:1493513.
Davda RR, Shabaker JW, Huber GW, Dumesic JA. Appl Catal B Environ (Special Issue
on H2 Prod). 2005;56:17186.
Huber GW, Cortright RD, Dumesic JA. Angew Chem Int Ed. 2004;43:154951.
Lin N, Huber GW. J Catal. 2010;270(1):4859.
84
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
2.1.12 References
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
85
2.2.1 Introduction
In the earliest years of humankind, those auxiliaries today known as chemicals could
only be obtained from nature. This may have included the use of native coal, gas, or
oil, but heavy exploitation may not have occurred. Since industrialization, the use of
coal (since the nineteenth century), acetylene (19301950), and nally oil and gas
(since the 1950s) served as a base for fuels and chemicals as well.
In Europe, approximately 69 million tons of oil was used as the raw material for
the chemical industry in 2008 [1]. The total oil demand in Europe was 703 million
tons in that year [2]. In contrast, only approximately 5% of all industry feedstock
is of renewable origin [3]. Most of this reects direct use of natural products like cotton for textiles, wood pulp for papermaking, or different oils for special applications
and for oleochemistry in general (detergents, lubricants, etc.) [3].
In Germany, approximately 4.1% of all fossil raw materials and 14% of the
crude oil demand were used in 2008 as feedstock for material use in the chemical
industry, together 17 million t [4]. In contrast, the total energy demand in Germany
was 224 million t [5].
Therefore, the ecological and environmental effects of switching the base of chemicals to renewable sources will be diminished as long as fuels are produced from crude
oil. Conversely, the relatively low amount of crude oil needed for the chemical industry will be accessible for a rather long period of time at moderate expenditures and
exploitation cost.
As the production of fuels from biomass is described in Chapter 2.1, the authors
want to mention that chemicals may also be classied as energy storage materials
as all energy input during production is present in chemical bonds. As the formation
of chemical bonds, especially the CC bond and CO bond synthesized by nature,
is tedious, the structure given by nature should wherever possible be retained
(Figure 2.2.1). This holds especially true for chemicals, in which it is not the chemical
energy that is needed but the property of a special molecule.
Another point in terms of CO2 reduction is the fact that binding carbon from
biomass into chemicals is a more sustainable approach than converting CO2 to
fuel which is burnt afterward or capturing CO2 into subterranean caverns.
The use of biomass as feedstock for fuels means a big challenge in that biomass in
general has a signicantly higher oxygen content than crude oil and common fuels
(diesel, gasoline). They comprise more or less completely unfunctionalized hydrocarbons of variable chain length. In current oil reneries, crude oil is therefore mainly
fractionated directly or only slightly modied.
88
Enthalpy
Oil and
gas
CxHy
Naphtha
Various
biomass
CxHy
CxHzN
CxHzO
CxHzOy (OCHz)yLignin
CxHzOy Carbohydrates
Petrochemical
approach
Chemical
products
Bio-refinery
approach
Figure 2.2.1 The production of functionalized chemicals from naphtha or from biomass [6].
With biomass as feedstock, completely new strategies are needed to reduce the
oxygen content and the degree of functionalization in biomass to reach the levels
of common fuels. Alternatively, the fuel basis can be switched to fuels more similar
to the oxygen-rich biomass (e.g. bioethanol).
These points hold only partially for the production of chemicals from biomass.
The smallest part of bulk chemicals consists of unfunctionalized hydrocarbons.
Nevertheless, the oxygen content of biomass has to be reduced to the level needed
for typical bulk chemicals. Therefore, strategies are strongly needed [7].
Politics has recognized the need for massive research. With the Biomass R&D Act
of 2000, the U.S. Department of Energy has founded many studies regarding the use
of biomass for energy and chemical use. Similar programs have also been conducted
in Europe and other countries.
Funding the research on that topic has led to several major studies examining the
potential of biomass as feedstock for chemical production. This includes the identication of possible intermediate chemicals, state-of-the-art processes for their production, and their current economic chances. Examples of these studies are the Top
Value-Added Chemicals from Biomass [8] and, in Europe, the BREW project [9].
It is not the intention of this chapter to give an extensive review of the subject headlined but instead to discuss fundamentals, ways to select key chemicals, and, nally,
examples for the current status of industrial chemical production from biomass.
Growth
89
Photosynthesis
CO2
H2O
Fuels
Cellulose
Biomass
Hemicellulose
Carbohydrates
Seed Oil
Lignin
Aromatics
Waste
Oil
Glycerol
Fiber
Fatty acids
Starch
Biopolymers
Crops
Chemical/
biological/
mechanical
Conversions
Chemicals
and plastics
Heat/power
Extracted
feed
Processed
feed
Complex
operations
Refined
products
Figure 2.2.2 Overview of the processing of crude feedstocks to rened products in a sustainable biorenery [10].
homogenous and constant feed (Figure 2.2.2). Not mentioned here are further carbohydrates such as free sugars and starch, because these are mainly used directly for
human nutrition.
The direct use of biomass components may be subject to specialty and ne chemicals or pharmaceuticals. Processes are well established and mainly comprise extraction and simple transformation of biomass components. Examples are monoterpenes
or fatty alcohols. They can be summarized as products with a very similar molecular
structure compared to the deployed biomass.
When the structure of biomass components has to be quite drastically rearranged
compared to the building blocks needed, synthesis gas or different sugars can serve as
platform chemicals. The latter can be converted, for example, to sugar-derived building blocks (Figure 2.2.3) that is, to glycerol, sorbitol, levulinic acid, and furfural.
Besides chemocatalysis, biotechnological processes are very important for converting biomass into valuable chemical compounds, which are summarized in Figure 2.2.4.
2.2.2.1 Lignocellulose
Lignocellulose denotes the mixture of the carbohydrate biopolymers cellulose and
hemicellulose with the aromatic polymer lignin that is found in plants. Wooden raw
materials consist mainly of cellulose (3050 wt %), hemicellulose (1040 wt %), and lignin (1530 wt %). As the structure of cellulose (C6 carbohydrates) and hemicellulose
(C5 carbohydrates) is quite similar, they will be discussed together in Section
2.2.2.1.1, followed by lignin, which has a very different composition (Section 2.2.2.1.2).
Lignin
Lipids/Oil
Protein
Lipids, Oil
Protein
Sugar
-Glucose
-Fructose
-Xylose
SynGas
Platforms
Lignin
Cellulose
Hemicellulose
Starch
Carbohydrates
Precursors
direct
Polymers
Aromatics
C6
C5
C4
C3
C2
SynGas
C1
Caprolactam
Gallic acid
Phenolics
Carnitine
THF
Furane
Furfural
Lysine
1,3-PDO
Levulinic acid
Acrylate
Dilactid
Diacids, Esters
Glycerol
Lactic acid
Olefins
Polysaccharides
Resins
Polyurethanes
Nylons
......
......
Polyacrylate
PLA
Emulsifiers
Chemical
intermediates
Green solvents
Solvents
Fuel additives
Ether
......
Intermediates
Secondary
Chemicals
Ethanol
Methanol
Propionic acid
Building
Blocks
Health a.
Recreation
Housing
Communication
Environment
Safe Food
Textiles
Transportation
Industrial
Products/
Uses
Biomass
90
2.2 Biomass Conversion to Chemicals
Lactic acid
3-Hydroxypropionic acid
Glycerol
1,2-Propanediol
1,3-Propanediol
Propionic acid
Acetone
Fumaric acid
Succinic acid
Malic acid
Butyric acid
1-Butanol
2,3-Butanediol
Acetoin
Aspartic acid
1,2,4-butanetriol
Itaconic acid
Glutamic acid
Citric acid, Aconitic acid
Lysine
cis-cis muconic acid
Gluconic acid
Kojic acid
C4
C5
C6
Ethanol
Acetic acid
Glyoxylic scid
Oxalic acid
C3
C2
Vitamin A
Vitamin B1
Vitamin B2
Vitamin B12
Vitamin C
Biotin
Folic acid
Pantothenic acid
Indigo
Miscellaneous
Industrial enzymes
Alginate
Bacterial cellulose
Curdlan
Chondroitin
Cyanophycin
Gellan
Heparin
Hyaluronic acid
Poly-gamma-Glutamic acid
Poly-epsilon-lysine
Polyhydroxyalkanoates
Pullulan
Scleroglucan
Sphingan
Xanthan
L-Alanine
L-Glutamine
L-Histidine
L-Hydroxyproline
L-Isoleucine
L-Leucine
L-Proline
L-Serine
L-Valine
L-Arginine
L-Tryphophane
L-Aspartic acid
L-Phenylalanine
L-Threonine
L-Glutamic acid
L-Lysine
Vitamins
Polymers
Aminio acids
Fatty acids
Glycerol
Xylose
arabinose
Glucose
Fructose
Figure 2.2.4
Protein
OIls
Lignin
Hemicellulose
Cellulose
Starch
Sucrose
92
Soil
Salts
Water
Kraft
Sulfite
Organosolv
Pyrolysis
Steam
Explosion
AFEX
Hot water
Other
Pretreatment
OH
HO
OH
Lignin
OH
OH
O O
OH
OH
OH
Hemicellulose
Cellulose
OH HO
Step 1
Catalysis
New technology
Catalysis
Further
processing
phenol
OH
OH
OH
Step 2
Catalysis
Current technology
e.g. Fischer-Tropsch,
Methanol synthesis
OH
New technology
Platform
chemicals
Syngas
CO
H2
Figure 2.2.5 Lignocellulosic biorenery scheme with particular emphasis on the lignin stream [7].
Proteins
Lipids
Ash
Cellulose
Hemicellulose
Lignin
Biomass
Pyrolysis and
gasification
Biorefinery
Fuels,
bulk and
fine
chemicals
93
94
2.2.2.2 Lipids
Lipids include a variety of quite different substance classes, with a wider range than
carbohydrates or proteins. The International Union of Pure and Applied Chemistry
(IUPAC) denition is quite loose and says substances of biological origin that are
soluble in nonpolar solvents [19]. Numerous further classication systems can be
applied, but a recent initiative suggested eight different groups [20]. Industrially,
the most important lipids are glycerides, fatty acids, and waxes. Further classes consist of functional plant and animal lipids such as glycerophospholipids, sphingolipids,
prenol lipids, saccharolipids, and polyketides.
2.2.2.2.1 Glycerides Glycerides or O-acylglycerols are esters of glycerol and fatty
acids and are colloquially named oils and fats. In nature, monoglycerides, diglycerides, and triglycerides may occur, but 97% of all oils and fats are triesters of glycerol
with different fatty acids. They are distinguished by their physical state at room
temperature; fats are solid, whereas oils are liquid [21].
They comprise a major product in agriculture with worldwide production amounts
of 173 million tons in 2009/2010 [22]. During the last decade, the global production
rate increased by roughly 45%, which can be attributed to an increasing demand in
developing countries and especially the increased biodiesel production. The shares in
demand are 80% for human nutrition, 6% for animal nutrition, and approximately
14% for chemical industry [23]. Most of the raw materials come from vegetable
oils (85%); only 15% are of animal origin [23].
The part of chemistry dealing with oils and fats is commonly known as oleochemistry. Figure 2.2.6 gives an overview of typical processes and products.
Figure 2.2.6
Oils
and
Fats
Fatty acids
Glycerol from
transesterification and
splitting
Purification
Fatty acid
methyl esters
Spilitting
Transesterification
Fatty
alcohols
Esterification
Esterification
Hydrogenation
Hydrogenation
Amidation, sulfonation
Basic oleochemicals
Raw materials
Triacetine
Partial glycerides
Biodiesel
Fatty acid alkanolamides
Sulfo fatty acid esters
Derivatives
95
96
2.2.2.2.2 Fatty Acids and Fatty Alcohols Fatty acids are traditionally meant as aliphatic unbranched monocarboxylic acids, either saturated or unsaturated, but with a
chain length of 4 to 28 carbon atoms. Sometimes even shorter acyclic aliphatic carboxylic acids like acetic acid are named fatty acids, although they are not found in
oils and fats [19].
The most abundant fatty acids in vegetable oils and fats are palmitic acid (hexadecanoic acid or 16:0), oleic acid ([9Z]-octadec-9-enoic acid or 18:1 cis-9), and linoleic acid (cis, cis-9,12-octadecadienoic acid or 18:2 cis-9 cis-12) [21]. Other fatty acids
are found in special oils (e.g. 80%87% ricinoleic acid in castor oil) [23], but these oils
are quite rare. Castor oil, for example, has a production rate of 610,000 tons/year
compared to the top four: palm oil (46 million tons/year), soya oil (40 million
tons/year), rapeseed oil (24 million tons/year), and sunower oil (12 million tons/
year) [24]. Further sources of fatty acids are tall oils (2 million tons/year) [25] and
to a lesser degree synthetic fatty acids derived by mainly hydroformylation and hydrocarboxylation of olens [23]. The summed fatty acid production is estimated to be
8 million tons/year (2006) [23].
Fatty acids are obtained by fat splitting using water (hydrolysis), methanol (methanolysis), and base (saponication) of amines (aminolysis). Splitting with water or
methanol can be considered transesterication because glycerol is liberated. The
methanolysis is the reaction taking place in biodiesel production as the resulting
product is called fatty acid methyl ester.
The hydrolysis is industrially conducted by acid or base catalysis (temperatures
210C260C) or by enzymatic hydrolysis in the case of sensitive fatty acids. The
resulting crude product mixture is separated and puried mainly by means of
distillation or crystallization, and more rarely by adsorption or extraction.
Pure fatty acids can be further treated (e.g. by hydrogenation, isomerization, and
dimerization). The hydrogenation of fats and fatty acids leads to partially or fully
unsaturated molecules under retention of the carboxylic group. This is known as
hardening because the melting point is increased. Hardened products exhibit a higher
stability to air and thermal treatment. Hydrogenation is mostly catalyzed by nickel
catalysts and carried out under increased temperature and pressure [23].
Fatty alcohols are obtained by direct hydrogenation of fatty acids or by hydrogenation of fatty acid esters. Typically, this is performed over copper catalysts at elevated temperature (170C270C) and pressure (40300 bar hydrogen) [26]. By
this route, completely saturated fatty alcohols are produced. In the past, unsaturated
fatty alcohols were produced via hydrolysis of whale oil (a natural wax occurring in
whale blubber) or by reduction of waxes with sodium (Bouveault-Blanc reduction).
Today, they can be obtained by selective hydrogenation at even higher temperatures
(250C280C), but lower pressure up to 25 bar over metal oxides (zinc oxide, chromium oxide, iron oxide, or cadmium oxide) or partially deactivated copper chromite
catalysts [26].
Approximately 75% of the fatty alcohols produced (around 3 million tons/year in
2008) are used in surfactants; further uses are cosmetics, lubricants, polymer ingredients, and avors [26]. The same uses hold for fatty acids, either in pure form or as
derivatives [23].
Fatty
alcohol
+ Oxygen
Alkali melt
Carboxylic acids
+ Alkali
Dimeric alcohols
+Protons
Ether, olefins
+ Alkynes
Vinyl ethers
+ Carboxylic acids
Esters
+ Hydrogen halides
Alkyl halides
+ Ammonia/amines
Amines
+ Aldehydes/ketones
Acetals
+ Sulfides
Thiols
+ Alkoxides/H2S
Xanthates
+ Metals
Metal alkoxides
+ Ethylene oxide/SO3
Ethoxylates/ethersulfates
97
Because the terminal hydroxyl of the carboxylic group of fatty alcohols and
fatty acids, respectively, can act as an anchor for further transformation, it can be
used for a variety of reactions not yet commercialized but with great possibilities
(Figure 2.2.7) [27].
2.2.2.2.3 Waxes Waxes were classically dened as esters of a fatty acid with a fatty
alcohol (e.g. beeswax), but this no longer holds as many waxes are not covered by
that denition [28]. Therefore, a denition based on physical properties has been proposed [29]. The global market for waxes is relatively small (estimated to be 3.4 million tons/year), comprising mainly natural, fossil parafn waxes (90%) and synthetic
polyolen and Fischer-Tropsch waxes [28]. So-called recent natural waxes (animal
and plant waxes) make up a small percentage with only 28,000 tons/year.
Similar to all natural waxes is a hydrocarbon-like chain structure with chain
lengths of approximately 20 to 60 and hence certain hydrophobic properties. As
their numerous applications are driven by physical properties, only slight chemical
modications take place in the natural waxes. Additionally, due to the low production
98
rate, waxes are not further considered as feedstock for additional processing into, for
example, commodity chemicals.
2.2.2.3 Proteins
Proteins are found in wooden biomass (< 5 wt %), cereals (~10 wt %), and herbs (up to
20 wt %) [13, 30]. They can accumulate in large-scale biomass processes like bioethanol or biodiesel production with approximately 1% of the deployed biomass [6].
As the annual world production of food protein from plants accounted for
250 million tons in 1990 [29], a huge surplus generated in biofuel plants could be further used for the production of chemicals. Proteins from plants can be directly used
in animal and human nutrition, cosmetics, and health care products as well as
adhesives, coatings, and further special applications [30, 31].
Amino acids can be obtained by hydrolysis of proteins, chemical synthesis, fermentation (e.g. of sugars), and enzymatic processes [32]. They serve similar purposes as
proteins, but in addition, aspartic acid and glutamic acid were rated as top valueadded chemicals, which can be converted to further ne chemicals [8]. Another example is the production of building blocks like ethylenediamine and butanediamine
from serine and arginine, respectively [6].
99
2.2.3.2 Glycerol
In the past, glycerol was produced mainly from propene via allyl chloride and epichlorohydrin, a process developed by I. G. Farben and in operation since 1943.
Today, glycerol is obtained almost completely as a coproduct in oleochemistry (fat
splitting) and biodiesel production (transesterication) with 110 kg crude glycerol
or 100 kg pure glycerol per ton of biodiesel [37]. With the rise in biodiesel production,
the availability increased while the price decreased drastically by approximately 66%
within 15 years in the United States [38].
The crude glycerol obtained during biodiesel production has a purity of approximately 80% to 85% and includes water, salt, and further residues. This crude glycerol
can be puried to pharmaceutical grade glycerol (> 99.5%) by vacuum distillation
[39]. Heterogeneously catalyzed processes are known that give glycerol in higher
purity (< 98%) without further distillation [40].
Glycerol is a valuable C3 building block, whose modication possibilities were
extensively reviewed by Behr et al. [41] and Pagliaro et al. [42]. Examples of the
use of glycerol are shown in Figure 2.2.8.
The rst plants for the biotechnological production of ethanol, acetone, and isoprene directly from glycerol with a capacity of approximately 20,000 tons/year are
under construction [38].
An important intermediate from glycerol is acrolein, which has been examined
in academia and industry [43] and already runs on a pilot scale but is not commercialized yet [44]. Many research reports are also focused on selective glycerol oxidation to dihydroxyacetone or glyceric acid, glycerol hydrogenolysis to 1,2- or 1,3propandiol, and aqueous-phase reforming (APR) of glycerol to produce hydrogen.
It is also possible to convert glycerol to epichlorohydrine in a commercial route
via the Epicerol process.
H2
APR
OH
HO
OH
Dehydratisation
Etherification
O
Oxidation
HO
OH
Amination
O
Hydrogenolysis
OH
HO
HO
NH2
O
OH
100
OH
+ 2HCl
RCOOH
OH
Cl
Catalyst
OH
Cl
Figure 2.2.9
Cl
+ NaOH
Cl + Cl
Cl
Cl
OH + 2H2O
+ NaCl
pany) [46].
Streamcracker
Oxidation
Propene
Hydrolysis
1. Chlorohydrin
2. Halcon-Hydroperoxide
3. HPPO (H2O2)
Propylene oxide
Oxidation
Propene
1,2-Propanediol
Uses: unsaturated polyester,
food processing, cosmetics,
pharmaceuticals, antifreezes
Hydrolysis
Acrolein
3-Hydroxypropionaldehyde
Hydrogenation
1,3-Propanediol
Uses: polyester production,
pharmaceuticals, ...
Figure 2.2.10 Old way petrochemical route toward 1,2- and 1,3-propandiol.
OH
CH2
CH
CH3
101
+ H2O
1,2-propanediol
Propylene glycol
OH
OH OH
CH2
CH
CH2
OH
+ H2
CH2
OH
CH2
CH2
+ H2O
1,3-propanediol
OH
OH
CH2
CH2
+ CH3OH
1,2-ethylene glycol
Figure 2.2.11 New way hydrogenolysis of glycerol toward 1,2- and 1,3-propandiol.
based either on propylene oxide with subsequent hydrolysis (large excess of water) to
1,2-propanediol or acrolein hydrolysis to 3-hydroxypropionaldehyde followed by
hydrogenation to 1,3-propanediol.
However, nowadays 1,2- and 1,3-propanediols can be produced by selective hydrogenolysis with appropriate bifunctional catalysts exhibiting active sites for dehydration and hydrogenation (see Figure 2.2.11). Cu/ZnO-based catalysts are frequently
used to produce 1,2-propanediol with high selectivity and catalyst activity; however
they suffer from strong deactivation. When Cu/ZnO is modied by Ga2O3, a stable
catalyst can be obtained that operates even under harsh reaction conditions (e.g. 220C)
and in the presence of water. Thus, 1,2-propanediol can be produced with a high
space-time yield of 22.1 g/(gCu h) [47].
In contrast, selective hydrogenolysis of glycerol to 1,3-propanediol by means of
chemo catalysis is still a challenging task. Although several attempts do exist with,
for example, Pt/WO3/ZrO2 or Ir-ReOx/SiO2 catalysts [48, 49], the enzyme-catalyzed
route using bacterial strains is more efcient [42] and has been commercialized
(see Table 2.2.1).
2.2.3.2.3 Acrolein Zirconium and niobium mixed oxides have been shown to catalyze the dehydration of glycerol to acrolein, at 300C in the presence of water with
high selectivity (72%) at nearly total glycerol conversion [50]. Silica-supported niobia
catalysts can also be used with similar catalytic performance [51]. Catalytic results for
small-sized H-ZSM 5 zeolites showed that the high density of Brnsted acid sites favors acrolein production [52]. Acrolein production from glycerol has also been carried out in subcritical water at 360C and 34 MPa with catalytic quantities of ZnSO4
(791 ppm [g/g]) [52].
2.2.3.2.4 Oxidation Products In principle, glycerol oxidation leads, according to
the complex reaction network shown in Figure 2.2.12, to numerous products mainly
depending on the pH of the solution and the type of catalyst (metal) used. Glyceric
acid and dihydroxyacetone are the main target products because of their use for the
production of amino acids and tanning components, respectively. Oxidation of the
102
Table 2.2.1 Examples of the current status of industrial chemical production from biomass.
Product
Producer
Capacity
t/a
Feed
1,3-Propandiol
100,000
Glycerol
45,000
Start-up
2011
METabolic EXplorer
50,000
Crude glycerol
3-Hydroxypropionic
acid
OPX Biotechnologies
Demo scale
Sugar
Butanediol
BioAmber
23,000
2014
BioAmber
50,000
2014
Genomatica
Demo scale
Citric acid
Biotechconsult
Intended
Crude glycerol
2012
Glucaric acid
Rennovia
135,000
Glucose
2012
Rivertop Renewables
45
Glucose
2011
Rivertop Renewables
27,000
Glucose
2013
Isoprene
METabolic EXplorer
20,000
Crude glycerol
2012
Itaconic acid
Biotechconsult
Intended
Crude glycerol
Methanol
Methanex
45,000
Waste
2011
(Poly)Ethylene
Braskem
200,000
2010
(Poly)Propylene
Braskem
30,000
2013
Succinic acid
BASF/PURAC
75,000
BioAmber
2,000
BioAmber
20,000
2013
BioAmber
35,000
2013
BioAmber
65,000
2014
DSM/Roquette
300
2010
DSM/Roquette
10,000
Starch
2012
Myriant Technologies
13,000
Sorghum
2012
2013
Wheat/glucose
2011
secondary OH group is more difcult and can be achieved with Pt-Bi catalysts (low
pH, 50C60C) [5456] or with mono- and bimetallic gold catalysts (high pH,
presence of NaOH is necessary, 60C) also producing glyceric acid [5760]. Pt-Bi catalysts show a high initial selectivity to dihydroxyacetone in acidic media but exhibit a
strong deactivation during reaction as well. For example, in continuous experiments
using a trickle-bed reactor (Liquid Hourly Space Velocity (LHSV) = 0.151.5 h1,
60C, 10 bar, O2/glycerol = 1.40.14), glycerol conversion of 90% was achieved
with a very high initial selectivity toward dihydroxyacetone of 80%, which decreased
to 40% within 1,000 h time on stream [56]. It has been shown that glyceric acid,
formed by oxidation of the primary OH group, selectively blocks those kinds of
active sites that are predominantly responsible for dihydroxyacetone formation [61].
OH
HO
HBTS OH
DHA
O
OH OH
MOS
OH
HO
103
O
HO
Decarboxylation
OH
GLY
HO
OH
OH
OH
GLA
HO
GLS
OH
O
OH OH
TS
Decarboxylation
OH
O
O
GOX
OH
H+
OH
OH
OH
O
GOS
GYS
O
OH
OS
Figure 2.2.12 Reaction network of glycerol oxidation (GLY, glycerol; DHA, dihydroxyace-
tone; GLA, glyceric aldehyde; GLS, glyceric acid; HBT, hydroxypyruvic acid; MOS, mesoxalic
acid; TS, tartronic acid; GOX, glyoxal; GOS, glycolic acid; GYS, glyoxylic acid; OS, oxalic acid).
104
3. Use of existing know-how of petrochemistry, especially from the eld of chemocatalysis, keeping in mind that the processes leading to chemicals from biomass
are necessarily different from those that start from fossil feedstock.
However, there are some boundary conditions and problems that have to be considered and solved for a successful process design:
In most cases, the key reactions from bio-based educts to chemicals are gas/liquid/
solid (G/L/S) systems.
They often contain complex reaction networks with an inherent problem of controlling the selectivity to the desired product depending on the type of catalyst
and the reaction conditions.
There is a remarkable inuence of pH on activity, selectivity, and time-on-stream
behavior.
Complete quantitative analysis of the reaction products is complicated but necessary to close material balances.
Mass transport effects have to be considered to understand the catalysts reactivity.
Catalyst deactivation, for example, via leaching, dissolution of the catalyst, or
decrease of metal dispersion has to be resolved.
It is important to note that the results in terms of selectivity/yield, catalyst activity,
and stability have to be veried when crude educts, instead of pure components,
are used.
Actually, most reactions are still carried out in batch reactors; however, development of continuous processes is necessary (e.g. using trickle-bed reactors).
To understand the complex chemistry and the catalysts mode of action in situ, catalyst characterization/operando spectroscopy is necessary, but methods for G/L/S
systems are still lacking.
2.2.7 References
Table 2.2.2
105
Forecast for the production of bulk chemicals from biomass in the Rotterdam
region [43].
010 years
1020 years
2030 years
As Table 2.2.1 demonstrates, the chemical industry step by step has reached a
point where ne chemicals are produced from biomass, especially by means of
biotechnology. Beginning with the production of bulk chemicals, not only biotechnology but also chemocatalysis is needed to convert renewable feedstock into
products in the high quantities characteristic of bulk chemicals.
2.2.6 Outlook
A case study and forecast for the explicit production of bulk chemicals from biomass
in the region of Rotterdam has recently been provided by Van Haveren et al. [46]. As
technologies for the conversion of, for example, ethanol, glycerol, and sugars to glycols, iso-propanol, and acetone are well known and readily available, and as the
prices for feedstock are well below the selling prices for the named products, there
is a clear short-term potential for these bulk chemicals (Table 2.2.2).
Technological barriers exist for the production of, for example, acrylic acid from
biomass, while currently higher prices impede the production of N-containing chemicals (acrylonitrile, acrylamide, and -caprolactam) from proteins. These obstacles
will be overcome according to van Haveren et al. in the next 10 to 20 years.
A clear long-term development will be the production of aromatics from biomass.
As the direct extraction of aromatics from lignin still needs a breakthrough in
research, the detour via pyrolysis or gasication of lignin is the current alternative
for the production of aromatics.
Although these results are not transferable to all markets, the chemical industry is
aware of chances to switch the value chain to feedstocks from biomass, as their prices
are expected to decrease contrary to oil prices.
2.2.7 References
1. CEFIC The European Chemical Industry Council. Cec review 20092010 sustainability
and innovation driving chemistry solutions for the future. 2010. Available from: http://asp.
zone-secure.net/v2/index.jsp?id = 598/765/10024&lng = en (accessed 24 January 2012).
2. Europia. 2009 activity report. 2009. Available from: http://www.europia.com/DocShareNo
Frame/Common/GetFile.asp?PortalSource = 1362&DocID = 25002&mfd = off&pdoc = 1
(accessed 24 January 2012).
3. Turley DB. The chemical value of biomass. In: Clark J, Deswarte F, editors. Introduction
to chemicals from biomass. Chichester: John Wiley & Sons; 2008.
106
2.2.7 References
107
25. Norlin L-H. Tall oil. In: Ullmanns encyclopedia of industrial chemistry. Wiley-VCH
Verlag GmbH & Co. KGaA; 2000.
26. Noweck K, Grafahrend W. Fatty alcohols. In: Ullmanns encyclopedia of industrial chemistry. Wiley-VCH Verlag GmbH & Co. KGaA; 2000.
27. Behr A, Westfechtel A, Prez Gomes J. Catalytic processes for the technical use of natural
fats and oils. Chem Eng Technol. 2008;31:70014.
28. Wolfmeier U, Schmidt H, Heinrichs F-L, et al. Waxes. In: Ullmanns encyclopedia of
industrial chemistry. Wiley-VCH Verlag GmbH & Co. KGaA; 2000.
29. Seher A, Lange J. Gemeinschaftsarbeiten der DGF, 60. Mitteilung. Deutsche Einheitsmethoden zur Untersuchung von Fetten, Fettprodukten und verwandten Stoffen, 45. Mitt.: Analyse
von Wachsen und Wachsprodukten X. Fette, Seifen, Anstrichm. 1974;76:135.
30. Klostermeyer H, Schmandke H, Soeder CJ, et al. Proteins. In: Ullmanns encyclopedia of
industrial chemistry. Wiley-VCH Verlag GmbH & Co. KGaA; 2000.
31. Koutinas AA, Du C, Wang RH, Webb C. Production of chemicals from biomass. In: Clark
J, Deswarte F, editors. Introduction to chemicals from biomass. Chichester: John Wiley &
Sons; 2008.
32. Drauz K, Grayson I, Kleemann A, Krimmer H-P, Leuchtenberger W, Weckbecker C.
Amino acids. In: Ullmanns encyclopedia of industrial chemistry. Wiley-VCH Verlag
GmbH & Co. KGaA; 2000.
33. Liang G, Wu C, He L, et al. Selective conversion of concentrated microcrystalline
cellulose to isosorbide over Ru/C catalyst. Green Chem. 2011;13:83942.
34. Van de Vyver S, Geboers J, Jacobs PA, Sels BF. Recent advances in the catalytic
conversion of cellulose. ChemCatChem. 2011;3:8294.
35. Kobayashi H, Ito Y, Komanoya T, et al. Synthesis of sugar alcohols by hydrolytic
hydrogenation of cellulose over supported metal catalysts. Green Chem. 2011;13:
32633.
36. Cabiaca A, Guillona E, Chambonb F, Pinel C, Rataboulb F, Essayem N. Cellulose
reactivity and glycosidic bond cleavage in aqueous phase by catalytic and non catalytic
transformations. Appl Catal A Gen 2011;402:110.
37. Christoph R, Schmidt B, Steinberner U, Dilla W, Karinen R. Glycerol. In: Ullmanns
encyclopedia of industrial chemistry. 7. Electronic ed. Weinheim: Wiley-VCH; 2006.
38. Guzman DD, Wilson S, Halaln H, Lefebvre B, Taylor J, Seng S. Low glycerin price drives
research. ICIS Chem Business. 2010;278:223.
39. Viinikainen TS, Karinen RS, Krause AOI. Conversion of glycerol into trafc fuels. In:
Centi G, van Santen RA, editors. Catalysis for renewables. Weinheim: Wiley-VCH; 2007.
40. Bournay L, Casanave D, Delfort B, Hillion G, Chodorge JA. New heterogeneous process
for biodiesel production: a way to improve the quality and the value of the crude glycerin
produced by biodiesel plants. Catal Today. 2005;106:1902.
41. Behr A, Eilting J, Irawadi K, Leschinski J, Lindner F. Improved utilisation of renewable
resources: new important derivatives of glycerol. Green Chem. 2008;10:1330.
42. Pagliaro M, Ciriminna R, Kimura H., Rossi M, Della Pina C. From glycerol to valueadded products. Angew Chem Int Ed. 2007;46:443440.
43. Coons R. Arkema nds glycerol-to-acrylic acid catalysts, demo-scale plant possible. Chem
Week. 2009;171:36.
44. Boswell C. A bigger toolbox for acrylic acid. ICIS Chem Business. 2011;279:324.
45. Sienel G, Rieth R, Rowbottom KT. Epoxides. In: Ullmanns encyclopedia of industrial
chemistry. Wiley-VCH Verlag GmbH & Co. KGaA; 2000.
46. Bell BM, Briggs JR, Campbell RM, et al. Glycerin as a renewable feedstock for epichlorohydrin production. The GTE process. Clean Soil Air Water. 2008;36:65761.
108
This work was part of the activities at the bo Akademi University Process Chemistry Centre within the Finnish Centre of Excelence Programme (20002011) by the
Academy of Finland.
Thermal conversion involves the use of elevated temperature with or without the
presence of oxygen to break down the structure of the feedstock. It includes torrefaction, pyrolysis, gasication, and combustion. Thermal conversion of biomass can
also be carried out in a solvent (e.g. as in hydrothermal processing) [1], but in this
chapter, only torrefaction; fast pyrolysis; gasication with air, oxygen, or steam;
and combustion in air will be considered.
Combustion of solid fuels can be divided into three stages: drying, devolatilization/
pyrolysis, and char conversion (Figure 2.3.1). During drying, water physically present in the biomass is released. Devolatilization represents the release of volatile compounds. These consist of low-molecular-weight compounds in the biomass as well as
higher-molecular-weight volatile compounds formed through the breaking of bonds
in the biomass. This process is called pyrolysis when carried out in an oxygen-free
environment for the recovery of liquid (tars). After devolatilization, a solid carbonaceous residue called char remains. The char, a mixture of xed carbon and ashforming matter, can be oxidized by O2, H2O, and CO2 leaving ash. The difference
between combustion and gasication is that in gasication a substoichiometric
amount of oxygen is supplied so that H2 and CO represent signicant amounts of
the carbon and hydrogen inputs.
A diagram of some thermochemical pathways for conversion of biomass is given in
Figure 2.3.1. Torrefaction is a thermal pretreatment method for the production of
biocoke, which has improved fuel handling and grindability as described later. Pyrolysis can be used to produce oil for combustion, gasication, or catalytic upgrading
for use in a traditional renery, or it can be fractionated and upgraded for chemicals
production (Figure 2.3.2). Similarly, gas from the gasication of biomass can be
burned in a boiler, burned in a gas turbine for electricity production, or conditioned
and catalytically or enzymatically converted to fuels and chemicals.
Cellulosic biomass, which is considered in this chapter, is composed of cellulose,
hemicelluloses, lignin, and extractives. On a dry basis, biomass has a lower heating
value (LHV) similar to low-rank coals and approximately 50% of that of high-rank
coals such as anthracite and bitimus coals (Table 2.3.1). Heating value is the energy
content of the solid fuel on a mass basis. Lower heating value is the usable energy
content (i.e. the condensation of water formed is not included in heating value).
110
H2O
CO2 +H2O
Solid fuel
Air
CO2 +H2O
O2
CxHy
Drying
CO2 CO, H2
O2
O2, CO2,
H2O
Pyrolysis/
devolatilisation
and
gas combustion
Char
combustion
Ash
Torrefaction
and
pelletization
250
to
300
650
to
1050
Pyrolysis
Solid biomass
400
to
600
Gasification
Bio-coal
Gas/diesel
Combustion
Oil to a
Hydrotreating petroleum
Pyrolysis Oil
refinery
(acidic, high O2 content)
Gasification
Gasification
+O2, hv
Gas cleaning,
CO, CO2, WGS, C2O CO, H
2
H2, H2O
Removal
Catalytic
upgrading
Gas
turbine
Transport
Fuels (MeOH,
DME, FT)
Electricity
Combustion
900
to
1300
Figure 2.3.2
Combustion
(alone or
co-combustion
w/coal)
Steam
turbine
Electricity
Industry,
household,
electric vehicles
Higher heating value includes the heat released by condensing water. Since the water
in combustion gases is not condensed, this energy is not captured in combustion systems. Biomass can have moisture contents in excess of 50%, meaning the energy density of the original feedstock can be quite low when considering transportation and
storage.
111
Table 2.3.1 Lower heating value of some fuels. LHV analyzed for fuels studied at bo
Akademi University unless noted otherwise. Higher heating values from a [2], b [3], and c [4]
were converted to LHV using the concentration of hydrogen.
Fuel
Coal, anthracite
34.9a
Coal, Polish
28.3
Coal, brown
22.7
Peat
19.3
Wood chips
18.7
Bark, spruce
20.8
20.7
Bagasse, Thai
17.5
Switchgrass
17.6b
Corn stover
16.2b
Straw
17.4
21.0
Carbohydrates
12.3c
Lignin, softwood
25.7c
Lignin, hardwood
23.9c
Extractives
35.9c
Pine
Birch
Extractives
3%
29 %
6%
3%
Extractives
7%
24 %
Lignin
Lignin
35 %
Hemicellulose
27 %
Cellulose
32 %
42 %
29 %
34 %
Hemicellulose
21 %
40 %
Mass
Cellulose
30 %
Energy
40 %
Mass
Energy
Figure 2.3.3
Extractives and lignin have a higher energy content than cellulose and hemicelluloses due to the difference in oxygen content of these fractions. An example of an estimated energy distribution by fraction in two wood species is given in Figure 2.3.3.
112
This has practical implications in current and future bioreneries as many schemes
may involve fractionation of the wood to utilize the cellulose and hemicellulose fractions for liquid fuels or chemicals, leaving the lignin fraction to be burned to provide
the energy for the plant. An example of this is the Kraft pulp mill where lignin, extractives, and part of the hemicelluloses are removed from wood in the freeing of the
cellulose bers. The removed organics are subsequently burned in a special boiler
called a Kraft recovery boiler to recover the energy in the organics and the pulping
chemicals.
2.3.1 Torrefaction
Torrefaction is a thermal pretreatment method being studied extensively. It involves
the heating of biomass to between approximately 240C and 320C in the absence
of oxygen [5]. Torrefaction results in a reduction in mass of up to 30 wt% on a
dry basis with an energy loss of approximately 10% of the energy in the original biomass. This results in a biomass feedstock with a higher energy density [6]. In addition
to improving the energy density, torrefaction results in a fuel with signicantly improved grindability [7], a reduced hydroscopic nature [8], and higher resistance to
biological decay [9, 10]. The increased energy density and reduction in the hydroscopic nature of the fuel improves the transport and storage characteristics. The
improvement in grindability will improve biomass handling for pulverized-fuel applications such as coring with coal in a pulverized-fuel boiler and entrained ow
gasication.
2.3.2 Pyrolysis
2.3.2.1 Introduction
Pyrolysis is the thermal degradation of an organic material in the absence of oxygen,
also the initial step in combustion and gasication. The products formed in the pyrolysis are a solid residue called char, a liquid called bio-oil, and uncondensed gases. The
process can be optimized for maximal bio-oil production. Important requirements
in this optimization are the following:
2.3.2 Pyrolysis
113
In order to achieve rapid heating of the biomass, it should be nely ground. The
particle size of the biomass is dependent on the reactor type used. In particular, it
was reported [17] that ablative reactors can pyrolyze large-size feedstocks, whereas
the uidized beds require smaller (below 3 mm) particles [16]. The difference is due
to lower heat transfer through conduction in the uidized beds, which is even lower
in the entrained ow reactor, being in the range of 4% [17].
A typical uidized-bed reactor system consists of biomass preparation (i.e. drying
and grinding), a biomass feeding system, a uidized-bed pyrolyzer, a char removal
system, and a cooling system for bio-oil collection. Char removal from the pyrolysis
reactor is benecial since the char can catalyze vapor cracking, thereby decreasing
the bio-oil yield [11].
Metso Power (Finland) has built an integrated pyrolysis pilot plant, illustrated in
Figure 2.3.4 [18]. The integration was carried out by coupling a 2 MWfuel uidizedbed pyrolyzer with a 4 MWth circulating uidized-bed boiler. The pyrolyzer utilizes
hot sand from the boiler as a heat source. The bio-oil is condensed with product
bio-oil in a scrubber and condenser unit. The char and uncondensed gases are fed
to the boiler and then combusted. It is reported that the integrated concept is easy
and smooth to operate and has a high efciency compared to stand-alone pyrolysis
units [18].
Reports on commercial implementation of pyrolysis are scarce, while there are
several pilot-size units. For example, Dynamotive in Canada has built four uidized-bed units with a total capacity of 200 tons of dry wood per day [14].
Ensyn, also in Canada, has eight circulating uidized-bed units with a total capacity of 100 dry tons per day [14]. Biomass Technology Group from the Netherlands
has designed rotating cone pyrolyzers with a total capacity of 2 tons of dry material
per hour [14].
114
District
heating
Scrubber
CFB-Pilot
Condenser
Reactor
Cooling
water
Fuel
supply
Heat
exchanger
Liquid separator
Non-condensable gases to boiler
Figure 2.3.4
2.3.2.3 Biomass
In principle, most types of biomass can be used as a raw material in the pyrolysis process [14]. Most of the research has been carried out using different wood as feedstock,
although more than 100 different types of biomass have been tested [14]. Besides
wood, these materials include forest residues, such as bark; black liquor; and agricultural residues such as straw, olive pits, and nut shells [12, 14]. Additionally, pure biological polymers cellulose (linear polymer of D-glucose units), hemicellulose
(heteropolymers of different hexoses and pentoses), and lignin (heteropolymer of
p-coumaryl alcohol, coniferyl alcohol, and sinapyl alcohol [19]) have been tested
as raw materials in the pyrolysis.
Yields of the different pyrolysis products (i.e. char, bio-oil, and gases) are dependent
on the chemical composition of biomass. Biomass with high lignin content produces
more char compared to biomass with lower content [20]. Furthermore, biomass with
high alkali and alkali earth metal concentrations generates less bio-oil and instead
more char and gases [14, 15, 21]. Prior to pyrolysis, it is benecial to dry the biomass
to less than 10 wt% moisture in order to minimize the water content in the bio-oil [14].
2.3.2 Pyrolysis
115
composition of the bio-oil dependent on the raw material used in the pyrolysis. The
elemental composition of the bio-oil is close to the composition of biomass (e.g. 54
58 wt% carbon, 5.57.0 wt% hydrogen, 3540 wt% oxygen, and some nitrogen and
ash; furthermore, the sulfur content is very low) [22]. As the elemental composition of
the bio-oil resembles that of biomass, the heating values are also similar, 1619 MJ/kg,
although depending on the bio-oils water content [11]. Even if the pyrolysis liquid is
called bio-oil, it will not mix with hydrocarbons, being soluble, however, in polar solvents. Moreover, devolatilization of bio-oil renders a solid residue up to 39 wt% [23]
(i.e. it cannot be distilled).
Pyrolysis of pure cellulose produces large amounts of 1,6-anhydroglucose, levoglucosan [2427]. Impurities or ash-forming elements such as alkali metals shift the degradation of cellulose toward smaller molecules such as glycolaldehyde [12, 25, 26, 28].
Other polysaccharide-related pyrolysis degradation products are organic acids,
esters, alcohols, ketones, aldehydes, and furans [12, 13]. Thermal degradation of the
lignin produces a wide range of different monomeric phenols, but mostly high
molar mass phenolic oligomers. The most dominating softwood lignin-originating
monomeric compounds are alkyl- and methoxy-substituted phenols (i.e. cresols and
guaiacols) [11].
116
2.3.3 Gasication
2.3.3.1 Introduction
Gasication involves adding substoichiometric levels of oxygen for partial oxidation
of the char and the high-molecular-weight volatiles:
(C,H,O,S,N)fuel + O2(g) (CO,CO2,H2,H2O,CH4,H2S,NH3,HCN)gas
+ CxHyOz(tars)
While high-molecular-weight volatiles are a desirable product of pyrolysis, they are
undesirable in gasication as they can condense in downstream processing, causing
fouling of these downstream processes. Additionally, they represent an energy loss
2.3.3 Gasication
117
to the system as they are an intermediate that can be used. In pyrolysis, these highmolecular-weight compounds are denoted as pyrolysis oil, but in gasication, they
are called tars. Similarly, it is important to gasify the xed carbon in char to minimize
energy losses and thereby maximize the desired product.
Oxidant
Heat source
Temperature Pressure
Reactor
H2O
Indirectly
heated
Low
Air
Atm.
FB, M/FBed
Atm.
Pres.
FB, M/FBed
Atm.
EF, M/FBed
Pres.
EF, M/FBed
Low
Directly
heated
O2
High
Figure 2.3.5 Gasication technologies/routes (Atm. = atmospheric pressure; Pres. = pressurized; FB = uidized bed [both bubbling and circulating]; M/F Bed = moving or xed bed;
EF = entrained ow).
118
This minimizes the electricity needed for compression of the product gas and
maximizes the concentration of the reactive components, CO and H2.
Indirect heating can be accomplished through a heating mantle surrounding the
gasication chamber; through the use of tubes running through the gasier; or by
recycling a heat-transfer medium such as sand in a uidized bed. This is used, for
example, in steam gasication.
2.3.4 Combustion
2.3.4.1 Introduction
In combustion, sufcient oxygen is added so that the fuel is fully oxidized:
(C,H,O,S,N)fuel + O2(g) (CO2,H2O,SO2,NOx,N2)gas
Biomass combustion for electricity generation or for both heat and power, can be
accomplished by either burning the biomass in a dedicated furnace or by coring
the biomass with another fuel such as coal. Dedicated biomass boilers are predominantly uidized-bed boilers (bubbling or circulating) or grate-red boilers. Gratered boilers are generally less than 100 MWth, while circulating uidized-bed boilers
are larger than 100 MWth. Bubbling uidized-bed boilers are economical both smaller
2.3.4 Combustion
119
and larger than 100 MWth. Coring of biomass and coal has been carried out in both of
these boiler types as well. Coring has been achieved at relatively low levels in pulverized-coal boilers, but high levels are currently difcult to achieve because of the
high energy needed to grind biomass and the inability to get a similar size fraction
for biomass as pulverized coal.
2.3.4.3 Co-combustion
Co-combustion involves partial replacement of a fossil fuel such as coal with biomass. Reviews of co-combustion have been published by Leckner [46] and Sami
[47]. Co-combustion is being done successfully at the commercial scale [48, 49]. For
partial substitution of coal by biomass, the biomass can be fed with the coal through
the same burner/feeder, or it can be fed to a separate burner. Alternatively, the biomass
can be pyrolyzed/gasied rst, and then the products can be burned in the boiler. Biomass gasication followed by combustion in a boiler has been commercially applied
and is illustrated in Figure 2.3.6.
One factor limiting biomass use in pulverized-coal-red boilers is the difculty in
grinding biomass. It is energy intensive and results in a larger particle size and distribution compared to pulverized coal [3, 50]. Torrefaction of the biomass improves the
grindability of the biomass [7]. Fluidized beds have been successfully used for both
coring biomass with coal and for ring biomass alone [48]. Fluidized beds are
more fuel exible than pulverized-fuel boilers and do not require the size reduction
(i.e. grinding) of the fuel that is required for pulverized-fuel units. This is one of
the factors limiting the use of high levels of substitution in pulverized-coal boilers
where substitution levels are typically less than 10% on an energy basis; however,
trials up to 20% have been successfully run with only slight increases in slagging
[46, 49, 50]. Another challenge of high levels of biomass coring is the ability to
sell the ash for industrial use, such as use in concrete and cement [46, 49].
120
Biomass or
waste
CO2 reduction
10%
Processing
50 MW
Gasifier
Gas flame
Coal or natural
gas
Fly ash
Bottom ash
Figure 2.3.6 Illustration of Foster Wheeler gasier in Lahti, Finland, used for gasication of
waste or biomass. The gas from the gasier is then burned in a boiler, which can process coal
or natural gas.
2.3.5 Summary
Thermal conversion is one of the ways to utilize the energy stored in biomass. Thermal conversion can be used to convert the chemical energy in biomass to liquid transportation fuels, electricity, and heat. Torrefaction (250C to 300C) can be utilized to
improve the energy density and handling characteristics of solid biomass for combustion and gasication applications. Pyrolysis is utilized to produce a liquid fuel, which
can either be burned directly in a boiler or catalytically upgraded for transportation
fuel production. The temperature for pyrolysis has to be high enough to produce a
signicant amount of higher-molecular-weight volatile components (tars) while not
so high that the volatiles are broken down further to lower-molecular-weight gaseous
species. Gasication involves higher temperatures (650C to ~1,050C), and the addition of an oxidant is utilized to convert the biomass to gas, which can then be burned
in a boiler or catalytically upgraded to liquid transportation fuels. Combustion involves addition of sufcient air to fully oxidize the fuel. Biomass can be burned
alone or co-combusted with fossil fuels. The potential advantage of co-combustion
is that higher steam temperatures, and thus higher electricity generating efciencies,
can potentially be reached compared to a boiler burning only biomass. Furthermore,
in co-combustion there is the potential to burn pyrolysis oils, providing an opportunity to develop those technologies separately from liquid fuels production. The wide
variety of thermal conversion technologies allows for a versatile array of solutions for
the needs of different countries.
2.3.6 References
121
2.3.6 References
1. Peterson AA, Vogel F, Lachance RP, Frling, M, Antal MJ Jr, Tester JW. Thermochemical biofuel production in hydrothermal media: a review of sub- and supercritical
water technologies. Energy Environ Sci. 2008;1:3265.
2. Bartok W, Sarom AF. Fossil fuel combustion: a source book. New York: Wiley; 1991.
3. Mani S, Tabil LG, Sokhansanj S. Grinding performance and physical properties of wheat
and barley straws, corn stover and switchgrass. Biomass Bioenergy. 2004;27:33952.
4. Frederick WJ. Black liquor properties. In: Adams TN, editor. Kraft recovery boilers.
Atlanta, GA: TAPPI Press; 1998. p. 78.
5. Bergman PCA, Kiel JHA. Torrefaction for biomass upgrading. Report ECN RX
05180. Petten, the Netherlands: ECN; 2005.
6. Bergman PCA. Combined torrefaction and pelletisation. The top process. Report ECNC 05073. Petten, The Netherlands: ECN; 2005.
7. Arias B, Pevida C, Fermoso J, Plaza MG, Rubiera F, Pis JJ. Inuence of torrefaction on the
grindability and reactivity of woody biomass. Fuel Processing Technol. 2008;89:16975.
8. Acharjee T, Coronella CJ, Vasquez VR. Effect of thermal pretreatment on equilibrium
moisture content of lignocellulosic biomass. Bioresour Technol. 2011;102:484954.
9. Chaouch M, Ptrissans M, Ptrissans A, Grardin P. Use of wood elemental composition
to predict heat treatment intensity and decay resistance of different softwood and hardwood species. Polymer Degradation Stability. 2010;95:22559.
10. Hakkou M, Ptrissans M, Grardin P, Zoulalian A. Investigations of the reasons for fungal durability of heat-treated beech wood. Polymer Degradation Stability. 2006;91:3937.
11. Bridgwater AV, editor. Fast pyrolysis of biomass: a handbook. Newbury, United Kingdom: CPL Press; 2002. Vol. 2.
12. Huber GW, Iborra S, Corma A. Synthesis of transportation fuels from biomass: chemistry,
catalysts, and engineering. Chem Rev. 2006;106:404498.
13. Mohan D, Pittman CU Jr, Steele PH. Pyrolysis of wood/biomass for bio-oil: a critical
review. Energy Fuels. 2006;20:84889.
14. Bridgwater AV. Review of fast pyrolysis of biomass and product upgrading. Biomass
Bioenergy. 2011;38:6894.
15. Brown RC, editor. Thermochemical processing of biomass conversion into fuels, chemicals and power. Chichester, West Sussex, UK: John Wiley & Sons; 2011.
16. Crocker M, editor. Thermochemical conversion of biomass to liquid fuels and chemicals.
Cambridge, UK: Royal Society of Chemistry; 2010.
17. Bridgwater AV. An overview of fast pyrolysis of biomass. Org Geochem. 1999;30:147993.
18. Autio J, Lehto J, Heikkinen M, et al. Production of bio-oil and chemicals in an integrated
pyrolysis pilot unit. In: Proceedings of the Third Nordic Wood Biorenery Conference.
March 2224, 2011, Stockholm, Sweden. Innventia AB, p. 4752.
19. Fengel D, Wegener G. Wood, chemistry, ultrastructure, reaction. Berlin, Germany: Walter
de Gruyter; 1984.
20. Nowakowski DJ, Bridgwater AV, Elliott DC, Meier D, de Wild P. Lignin fast pyrolysis:
results from an international collaboration. J Anal Appl Pyrolysis. 2010;88:5372.
21. Chiaramonti D, Oasmaa A, Solantausta Y. Power generation using fast pyrolysis liquids
from biomass. Renewable Sustainable Energy Rev. 2007;11:105685.
22. Czernik S, Bridgwater AV. Overview of applications of biomass fast pyrolysis oil. Energy
Fuels. 2004;18:5908.
23. Branca C, Di Blasi C, Elefante R. Devolatilization and heterogeneous combustion of
wood fast pyrolysis oils. Ind Eng Chem Res. 2005;44:799810.
122
24. Essig M, Richards GN, Schenck E. Mechanisms of formation of the major volatile products
from the pyrolysis of cellulose. In: Schuerch C, editor. Cellulose and wood chemistry and
technology. Syracuse, NY: Wiley & Sons; 1988. p. 84162.
25. Evans RJ, Milne TA. Molecular characterization of the pyrolysis of biomass. 1. Fundamentals. Energy Fuels. 1987;1:1327.
26. Richards GN. Glycolaldehyde from pyrolysis of cellulose. J Anal Appl Pyrolysis. 1987;
10:2515.
27. Shazadeh F, Fu YL. Pyrolysis of cellulose. Carbohydr Res. 1973;29:11322.
28. Bridgwater AV. Production of high grade fuels and chemicals from catalytic pyrolysis of
biomass. Catal Today. 1996;29:28595.
29. RedArrow. 2011. Condensed Natural Smoke, Manufacturing Process. Available from:
http://www.redarrowusa.com/natural_smoke_condensates/manufacturing_process.html
(accessed 15 June 2011).
30. Solantausta Y, Nylund N-O, Westerholm M, Koljonen T, Oasmaa A. Wood-pyrolysis oil
as fuel in a diesel-power plant. Bioresour Technol. 1993;46:17788.
ejka J, van Bekkum H, Corma A, Schth A, editors. Introduction to zeolite science and
31. C
practice. 3rd revised ed. Amsterdam, the Netherlands: Elsevier Science; 2007.
32. Seyedeyn-Azad F, Salehi E, Abedi J, Harding T. Biomass to hydrogen via catalytic steam
reforming of bio-oil over Ni-supported alumina catalysts. Fuel Process Technol. 2011;
92:5639.
33. Atutxa A, Aguado R, Gayubo AG, Olazar M, Bilbao J. Kinetic description of the catalytic pyrolysis of biomass in a conical spouted bed reactor. Energy Fuels. 2005;19:76574.
34. Williams PT, Horne PA. Analysis of aromatic hydrocarbons in pyrolytic oil derived from
biomass. J Anal Appl Pyrolysis. 1995;31:1537.
35. Aho A, Kumar N, Ernen K, Salmi T, Hupa M, Murzin DY. Catalytic pyrolysis of
woody biomass in a uidized bed reactor: inuence of the zeolite structure. Fuel. 2008;87:
2493501.
36. Lappas AA, Samolada MC, Iatridis DK, Voutetakis SS, Vasalos IA. Biomass pyrolysis in
a circulating uid bed reactor for the production of fuels and chemicals. Fuel. 2002;81:
208795.
37. Aho A, Kumar N, Lashkul AV, et al. Catalytic upgrading of woody biomass derived
pyrolysis vapours over iron modied zeolites in a dual-uidized bed reactor. Fuel. 2010;
89:19922000.
38. Williams PT, Horne PA. The inuence of catalyst type on the composition of upgraded
biomass pyrolysis oil. J Anal Appl Pyrolysis. 1995;31:3961.
39. Vitolo S, Seggiani M, Frediani P, Ambrosini G, Politi L. Catalytic upgrading of pyrolysis
oils to fuel over different zeolites. Fuel. 1999;78:114759.
40. Francey S, Tran H, Jones A. Current status of alternative fuel use in lime kilns. 2008
TAPPI Engineering, Pulping and Environmental Conf.; 2008 Aug 2427; Portland, OR.
Atlanta: TAPPI Press; 2008.
41. Neathery JK. Chp 4 Biomass Gasication in Thermochemical Conversion of Biomass to
Liquid Fuels and Chemicals (ed. Crocker, M.). Cambridge,UK: RSC Publishing; 2010,
p. 734.
42. Umeki K, Yamamoto K, Namioka T, Yoshikawa K. High temperature steam-only gasication of woody biomass. Applied Energy 2010;87:7918.
43. IEA. Energy technology perspectives: scenarios and strategies to 2050. Paris: IEA; 2006.
44. Korens N, Simbeck DR, Wilhelm DJ. Process screening analysis of alternative gas treating and sulfur removal for gasication. Revised Final Report. U.S. Department of Energy.
Task Order No. 73965600100. Mountain View, CA: SFA Pacic, Inc.; 2002.
2.3.6 References
123
45. Bain RL. Material and energy balances for methanol from biomass using biomass gasiers. NREL/TP-51017098. Golden, CO: NREL; 1992.
46. Leckner B. Co-combustion a summary of technology. Thermal Sci. 2007;11(4):540.
47. Annamalai K, Wooldridge M. Co-ring of coal and biomass fuel blends. Prog Energy
Combust Sci. 2001;27:171214.
48. Kokko A, Nylund M. Biomass and coal co-combustion in utility scale operating experience of Alholmens Kraft. In: 18th International Conf. on FBC; 2005. New York:
ASME; 2005. p. 599609.
49. Overgaard P, Sander B, Junker H, Friborg K, Larsen O-H. Two years operational experience and further development of full-scale co-ring of straw. In: 2nd World Conf. on
Biomass for Energy, Industry and Climate Protection; 2004 May 1014; Rome, Italy.
p. 12614.
50. Pedersen L-S, Nielsen HP, Kiil S, et al. Full-scale co-ring of straw and coal. Fuel. 1996;
75(13):158490.
Low-value biomass (e.g. from waste streams or agricultural side products) is a potentially very valuable raw material source. For that, biomass has to be stabilized and
mineralized, and a process called hydrothermal carbonization (HTC) can be used for
densication of the material and energy content, but also as a unication technology.
As the resulting products bind carbon from natural, regrowing resources, they potentially inuence the CO2 balance sheet in a favorable, lowering fashion. It will be discussed how such novel usage cascades based on abundant biomasses can contribute
to global energy/materials generation or CO2-sequestration schemes.
2.4.1 Introduction
Our industrial society currently depends on a stable support of fossil crude oil for
energy generation, transportation, and the generation of chemical platform chemicals. The end of the oil age is, however, foreseeable, and the economic earthquakes
preceding the shortage of oil are already being felt today. A further downside of the
oil economy is the liberation of large amounts of CO2. From oil alone, 12.5 billion
tons (Gt) of CO2 are generated, with the known implications for climate and weather
extremes.
The replacement of smaller material streams by renewables is certainly desirable,
but essentially just window dressing, as it leaves the very basic problems untouched.
In particular, it cannot replace the current base load of fossil energy carriers, so
humankind is essentially employing a business as usual approach without effective
changes.
How could an actual useful instrument or technology be dened? It is very clear
that it would be worthwhile to not only lower further CO2 emissions (the current,
realistic base of all climate models) but also create processes that counteract previous
development by binding new atmospheric CO2 as well as that from early industrialization. The thought of climate management instead of climate consumption is on
the horizon, but currently still not the object of a public discussion. Realistically,
from the chemical point of view there is no bypass around such a technology: the
search for new carbon deposits is key, while part of our industrial products must
be turned into carbon-negative deposits (i.e. products that bind carbon rather than
generate it throughout their life cycle).
126
2.4.2 HTC
2.4.2.1 HTC of Biomass Waste for Environmentally Friendly
Carbon Sequestration
The simplest path to this target is presumably indirect, but close at hand. The most
effective and cost-free CO2 collector is Mother Nature herself, and she is also the
most massive sink for carbon. A rough estimate of the terrestrial biomass production
(not including the contributions of oceans) amounts to 120 Gt/year as dry matter,
these are approximately 60 Gt bound carbon or 220 Gt sequestered CO2 per year
[1, 2]. The natural CO2 cycle is therefore still one order of magnitude larger than
the anthropogenic one, except that nature has been in equilibrium for hundreds of
millions of years. This perpetuated binding and liberation of CO2 can indeed serve
as the role model for future chemistry.
The generated biomass is, however, just a short-term, temporal sink of CO2, as the
microbial degradation after plant death liberates exactly the amount of CO2 that was
previously bound in the plant material. Plant biomass, because of its chemical structure, contains 40 wt% (carbohydrates) and 50 wt% (lignocellulosic biomass) carbon.
Locking away just 8.5 wt% of the freshly produced biomass from the active ecosystem would indeed compensate for the complete CO2 generation from crude oil.
This nicely illustrates the overall scale of an actual effective climate operation, and
global application of biomass collection and conversion schemes might indeed not
be unrealistic for lowing atmospheric CO2 by direct human operations.
There are many reasons why the locking away will not occur as a slag heap, and
from the viewpoint of a material chemist or an economist, the carbon should instead
be used for a secondary purpose while being deposited. Examples of this are adding
carbon to large-scale technical products, as a component to improve concrete, as a
mixing additive for street pavements, for building insulation, or even as a lling soot
for mechanic or coloristic applications. For all that, we need another processing
step to turn biomass into coal, coke, carbon, or soot, to accomplish densication
and uniformity, and carbonization (i.e. the conversion of soft biomolecules into a
carbon-rich powder) is the name for this process.
Relevant to the present argument, there exists not only the hot ame carbonization, as a charcoal burner is using it, but also a more effective wet carbonization,
which takes place on the timescale of a few hundred years (throughout the formation
of peat) to some millions of years (black coal). In this pathway, biomass is dehydrated under slightly acidic conditions and the exclusion of oxygen, essentially creating only water and coal. In the modern scientic literature, some experiments are
described in which this coal formation is mimicked by faster technological processes.
The HTC was performed by Bergius and Specht as early as 1913. In those experiments, cellulose was converted hydrothermally in coal-like substances [3, 4]. Biomass
in the presence of water and potential catalysts is heated to 180C230C in a closed
autoclave, and the reaction still takes days to weeks. More modern versions still
accelerate those processes by additives and allow the generation of meaningful
micro- and nanostructures with special surface chemistry. The nally described acceleration of coalication by factors of 106109 down to the hour scale makes HTC a
2.4.2 HTC
127
biocoal
water
Scheme 2.4.1 Operating principle of HTC. Under temperature and catalysis, carbohydrates
(here glucose) are converted to biocoal and water only. The sum formula of biocoal is a
simplication and a matter of the reaction conditions.
In comparison to other biomass processes, HTC is not only fast but also simple
and effective. On the one hand, it inherently can use wet starting products, as the
reaction only effectively takes place in a water environment. Costly drying, therefore,
is not demanded. On the other hand, biocoal can be easily ltered off from the wastewater stream; the product isolation is rather simple. It is also favorable that the
majority of salts is kept in the water stream and can be donated back to the natural
cycle. Furthermore, under weak acidic conditions, at temperatures below 220C, and
using wastewater for processing, practically all carbon from the biomass ends up in
the biocoal (i.e. the carbon efciency [CE] is practically 1). Another benecial aspect
is that the reaction is exothermic and spontaneous (i.e. the process does not rely on
excessive energy input). The real energy output is still under debate, but depending
on the reaction scheme and degree of aromatization of the coal, between 5% and
30% of the original heat of combustion is liberated throughout the process. This
was already observed by Bergius in his setup: without cooling, some of his reactions
would have heated up to a few hundred degrees Celsius.
It is therefore my opinion that carbonization of waste biomass or of rapidly growing plants or algae (farmed just for this purpose) represents at the moment the most
effective process for the removal of atmospheric CO2. Ideally, however, this will be
coupled for protability with chemical material synthesis (make something useful
out of it) and presumably can be supported by political instruments encouraging
CO2-negative products.
128
tons, and the potentialities for a chemist to work with such molecules can be easily
illustrated.
By choice of the biomass or adding minor amounts of special chemical comonomers, a hydrophilic surface with special functional groups and a high capillarity as
a key property can be adjusted. Besides distinct chemical functionalities, a special
product texture is also benecial for such a purpose (i.e. the micro- and nanoarchitecture of the coal has to be optimized). In this way, and nevertheless using biological
waste as a starting product, the whole knowledge base of modern material chemistry
can be used to undertake such a project, and a superhumus could be synthesized,
as it is rather rarely produced in nature. Figure 2.4.1 depicts the inner structure of
this HTC product, which was synthesized from Berlin oak leaves [5]. The material
combines optimal accessibility with a highly functional surface chemistry and is
ideally suited for the capillary binding of water and specic ion binding, a carbon
sponge in gurative language.
For use as a soot replacement (e.g. for printing purposes or as a ller for car tires),
other disperse morphologies are regarded to be optimal (e.g. as those known from
emulsion polymerization processes), and a corresponding reaction engineering of
HTC including comonomers and stabilizers is providing exactly such products
(Fig. 2.4.2A). For use in gas lters, thermal insulation, mechanical strengthening,
and antistatic nishing, nanober architectures are most optimal, and again, HTC
can be directed to produce these materials (Fig. 2.4.2B). It is important to understand
that HTC is just another heterophase polymerization process that allows reapplication
of known principles of material design.
Making such products is a highly attractive task, especially when biomass waste
is directly used for product synthesis, without chemical deviations. The concept of
CO2-negative chemical products is new in industrial thinking, but companies
could improve their CO2 balance simply by making some of their products in this
manner [6].
200 nm
Figure 2.4.1 Electron microscopy picture of biocoal, which was made by hydrothermal treatment of oak leaves. The sponge-like pore structure with structural elements in the 2050 nm
region is nicely seen.
129
300 nm
2m
EHT = 3.00 kV
WD = 5 mm
Figure 2.4.2
130
2 C2H5OH
2 CO2
40C, 5 d
Digestion
C6H12O6
3 CH4
40C, 21 d
HTC
,,C2H4O2 +
C6H12O6
200C, 2-12 h
3 CO2
2720 kJ
2680 kJ
4 H2O
ca. 2600-3000
kJ
Figure 2.4.3 Comparison of different renewable energy pathways and carbon transfer
schemes from carbohydrates, as well as their typical conditions. Here, preservation of combustion energy and the carbon efciency (CE) of the transformation are compared. The combustion energy always concerns the complete side of the reaction equation. The sum
formula of the coalied plant material is a schematic simplication.
A graphical comparison between HTC and alcoholic fermentation or anaerobic digestion with respect to reaction schemes, typical conditions, and mass streams
is given in Figure 2.4.3.
All three conversions are, of course, exothermic and spontaneous and store about
the same level of energy in the nal products. The energy loss explains the simplicity
of conversion, independent of its biological or chemical origin. HTC generates water
instead of CO2 as a side product; coalication in these terms is just the elimination of
water under preservation of the carbon scaffolds.
Thus, using the HTC process to convert biomass into coal could represent a most
efcient tool for valozation of the energy content [7]. The already technically realized acceleration of the coalication down to the hour range and operation of continuous processes makes it a technically attractive, realistic instrument for generating
a transportable, dense, stable, and rather safe chemical energy carrier.
Current medium-scale operations for HTC calculate operating expenses on the
order of 200/ton biocoal, not including a forthcoming economy of scale and further
technical optimization. Even at this premature level, this corresponds to 2.7 (euro)
cents/kWh heat (i.e. it can nicely compete with current energy media, being carbon
neutral at the same time).
However, what needs to be considered is the societal convenience of liquid fuels,
which makes them a preferred choice even when coming with a high markup. Science
on energy storage also has to consider such issues (e.g. how to create convenience
with solid energy carriers).
Figure 2.4.4 depicts a technical realization of such an HTC machine on the pilot
scale (designed by Carbon Solutions Ltd.). Such a machine in principle ts into a
smaller building and can run 10,000 tons of input per year (i.e. good for a communal
or small industrial operation).
131
1 / mA
6
5
4
3
2
1
0
|EOC|/mV
500
400
300
200
0
5000
3.0
10000
t/s
15000
1 / mA
2.5
2.0
1.5
1.0
0.5
0
5000
10000
t/s
15000
20000
C
2.5
1 / mA
2.0
HTC coal
Vattenfall lignite
RWE anthracite
1.5
1.0
0.5
0.0
0
Figure 2.4.4 (A) Time-dependent electric current generated from the oxidation of HTC coal
in an indirect carbon fuel cell. Solutions of Fe III and VV were prepared in 0.5 mol L1 H2SO4.
(B) Development of open-circuit potential Eoc (up) and current I (down) due to Fe2+ formation
in the anodic half-cell via oxidation of HTC coal, indicating the reducing potential of bare
hydrothermal carbon (HC) dispersions. Charge equalization between the two half-cells
was assured by a salt bridge containing a saturated KCl solution. Carbon felt was used as
electrodes. (C) Comparison of hydrothermal and fossil carbon sources in the same setup.
132
2.4.5 References
133
2.4.5 References
1. Lieth H, Whittaker RH. Primary productivity of the biosphere. Berlin: Springer; 1975.
p. 2056.
2. Bobleter O. Hydrothermal degradation of polymers derived from plants. Prog Polym Sci.
1994;19:797841.
3. Bergius F, Specht, H. Die Anwendung hoher Drcke bei chemischen Vorgngen, Habilitation work. Halle; 1913.
4. Bergius F. Articles on the theory of coal formation. Naturwissenschaften. 1928;16:110.
5. Titirici MM, Thomas A, Yu SH, Antonietti M. A direct synthesis of mesoporous carbon with
bicontinuous pore morphology from crude plant material by hydrothermal carbonization.
Chem Mater. 2007;19:420512.
6. Titirici MM, Thomas A, Antonietti M. Back in the black: Hydrothermal carbonization of
plant material as an efcient chemical process to treat the CO2 problem? New J Chem.
2007;31:7879.
7. Hu B, Wang K, Wu LH, Yu SH, Antonietti M, Titirici MM. Engineering carbon materials from the hydrothermal carbonization process of biomass. Adv Mater. 2010;22:81328.
8. Jacques W. Method of converting potential energy of carbon into electrical energy. U.S.
Patent 555511, 26.06, 1896.
9. Horita T, Sakai N, Kawada T, Yokokawa H, Dokiya M. An investigation of anodes for
direct-oxidation of carbon in solid oxide fuel-cells. J Electrochem Soc. 1995;142.
10. Peelen WHA, Olivry M, Au SF, Fehribach JD, Hemmes K. Electrochemical oxidation of
carbon in a 62/38 mol % Li/K carbonate melt. J Appl Electrochem. 2000;30.
11. Paraknowitsch JP, Thomas A, Antonietti M. Carbon colloids prepared by hydrothermal
carbonization as efcient fuel for indirect carbon fuel cells. Chem Mater. 2009;21:11705.
3.1.1 Introduction
Generations of students have been introduced to electrochemistry with the glass
apparatus shown schematically in Figure 3.1.1.
Two platinum electrodes operate in a U-shaped electrochemical cell containing an
acidic electrolyte. If approximately 1.8 V are applied between the two electrodes, current ows that produces both oxygen and hydrogen according to the following
reactions:
2 H2O O2 + 4 H+ + 4 e
(1)
2 H+ + 2 e H2
(2)
G 0 = 474.4 KJ mol1
(3)
In Equation (3), the standard Gibbs free energy, G 0, is given. G 0 > 0 means that
the reaction is endothermic. The value of G 0 denes the thermodynamic (minimum)
electric energy required to decompose 2 mol liquid water according to the chemical
Equation (3) under standard conditions. An electrochemical process in which an endothermic reaction is forced to proceed by the imposition of an external voltage is
termed electrolysis, and the cell is said to operate as an electrolytic cell. The electric
energy W (in joules, J) consumed in an electrolytic process is dened by the product
of applied voltage U (in volts, V) and the electric charge Q (in coulombs, C) that has
passed through the cell, as shown in Equation (4). The electric charge is Q = I t (I is
the electric current in amperes, A, and t is the time in seconds, s).
W=QU
(4)
The electric work W required for an electrochemical process in a practical electrolytic cell will be larger than G 0. For example, the indicated value of G 0 in Equation (3) corresponds to a thermodynamic standard potential difference of E0 = 1.23 V,
while the voltage typically applied to the cell of Figure 3.1.1 is approximately U = 1.8 V.
136
O2
H2
The overvoltage (U E 0) is the driving force for the reaction to proceed at a desirable
rate. However, it also causes an energy loss, which leads to heating of the cell.
The electrochemical water decomposition reaction can be reversed according to
O2 + 2 H2 2 H2O
G 0 = 474.4 KJ mol1
(3a)
H2
137
O2
228.4
kJ mol1
H2O
Figure 3.1.2 The concept of (electro-)chemical energy storage [2]. The indicated energy value
corresponds to G 0 for the decomposition of 1 mol gaseous H2O according to Equation (3).
138
Ag
Electrolyte
Ag+NO3
Ag+
Ag+ + e Ag
Ag+ + e Ag
Figure 3.1.3 The Ag/Ag+ electrode: an example of a metal/metal ion electrode. The dynamic
equilibrium state is shown (i.e. both the forward and back reaction proceed at the same rate).
d. Concentration cells
Electrochemical cells may consist of two electrodes of the same type, but with different activities of the electroactive species in the electrolyte. Such systems will be
discussed subsequently in the context of the Nernst equation.
e. Ideal polarizable electrodes
If an inert conductor operates as an electrode in an inert electrolyte (e.g. KCl),
there exists a potential region in which no interfacial charge transfer takes
place. Still, the variation of the applied electrode potential is associated with
the ow of a capacitive (nonfaradaic) current. It is caused by charging and discharging of the electric double layer (i.e. the accumulation of electronic excess charge
in the metal surface and ionic excess charge of the opposite sign at the solution side
of the interface). The double-layer capacitance is on the order of 10100 F cm2.
High-surface-area electrodes (e.g. carbon) can store substantial quantities of electric
charge, and thus of electric energy.
139
The analysis of the Ag/Ag+ electrode of Figure 3.1.3 is presented in Figure 3.1.4.
The electrode is at open circuit (i.e. no external potential is imposed on the electrode).
At the metal/electrolyte interface, the reaction Ag+ + e Ag proceeds both in the
forward and backward directions. The following thermodynamic cycle leads to a
value of G 0 for the reaction Ag Ag+ + e. It is conducted using energy data
from various sources, in particular from Goodisman [5]. The rst step is the sublimation of 1 mol metallic silver. The silver atoms are then ionized, for which the ionization potential (IP) must be supplied. The electrons are transferred back into the metal
phase with the liberation of the free energy corresponding to the work function, . At
last, the standard free energy of solvation of Ag+ is determined from the corresponding
enthalpy (H 0) and entropy (S 0) values. Note that
G = H TS
(6)
The data given in Figure 3.1.4 may be summarized as follows in Table 3.1.1.
The previous analysis clearly shows that a substantial amount of energy has to be
supplied to the metallic silver in order to make it dissolve as Ag+. On the other hand,
Ag+ from the electrolyte will spontaneously react at the silver surface to form metallic Ag: G 0(Ag+ Ag) 95 kJ mol1. Indeed, this is why silver is considered a
noble metal.
The establishment of a stable equilibrium potential between the metal electrode
and the electrolyte can be straightforwardly explained as follows. As soon as the neutral silver electrode gets in contact with the electrolyte, the reaction Ag+ + e Ag
will proceed, while the rate of the back reaction is negligibly small. The excess positive charge injected into the silver electrode will render the potential of the electrode
Ag
e + Ag+
H = 464 kJ mol1
S = 115 JK1 mol1
W
o
(1 rk f
11 un
) ct
= ion
4.
74 of
eV
IP = 7.62 eV
Ag
H = 285 kJ mol1
S = 138 JK1 mol1
Ag vapor
Ag+(solv H2O)
Figure 3.1.4 Born-Haber cycle for determining the standard free energy change associated
with the reaction Ag Ag+(solvated). The data are shown as they are normally reported in
the literature: 1 eV/atom = 96.5 KJ/mol. For more details, see the text.
140
G(sublim)
+247 kJ mol1
IP
=
=
+735 kJ mol1
457 kJ mol1
430 kJ mol1
+95 kJ mol1
+1 eV/atom
G (solv Ag )
0
G (Ag Ag
0
solv)
(relative to the electrolyte solution) more and more positive. Consequently, the rate
of the reaction Ag+ + e Ag decreases while the back reaction is accelerated. Eventually, the dynamic equilibrium state is attained at which both the silver deposition
and dissolution reactions proceed at the same rate, as illustrated in Figure 3.1.3. This
state corresponds to a stable positive potential drop across the metal/electrolyte interface. The experimental investigation of such interfacial potential drops requires
methods as used in double-layer studies. While this is an important and interesting
theme, it is not the subject of this work.
Rather, electrochemical cells containing a second electrode will be considered in
the following. These are the electrochemical systems of practical interest, as shown
in Figure 3.1.1. Still, one should keep in mind that the behavior of a two-electrode
cell is the combination of the two individual electrodes. And indeed, experimental
methods have been developed, in particular potentiostatic methods, by which the effects of the second electrode are virtually eliminated so that one electrode of interest
may be studied [6].
(2a)
(7)
141
e
H2
LOAD
Ag
Electrolyte
Ag+NO3
Ag+
Ag+ + e Ag
H+NO3
NO3
Membrane
H+
H2 H+ + e
Figure 3.1.5 Electrochemical cell consisting of an Ag+/Ag and an H2/H+ electrode. Electronic current (e) ows through the external circuit containing the load. Inside the cell,
the electric charge is shown to be transported by nitrate ions (NO3 ).
from the outside circuit the electrons enter into it, causing cathodic reduction of an
electroactive species (Ag+, in the electrolyte). On the other hand, from the H2/H+
electrode the electrons exit into the external circuit. Thus, it constitutes the anode,
at which anodic oxidation of H2 takes place.
There is an important feature common to all electrochemical cells. Provisions have
to be made to prevent the electroactive components from reaching the opposite electrodes. Typically, this is achieved by a membrane between the anode and the cathode
compartments. The membrane must be permeable to ions (NO3 in Figure 3.1.5). Interestingly, the water decomposition cell of Figure 3.1.1 operates without a membrane. This has the consequence that some dissolved hydrogen and oxygen will be
transported by diffusion and convection to the respective opposite electrodes.
There, both reactions (1) and (2) will proceed in the back direction. In the device
of Figure 3.1.1, this is tolerable because this cell is designed for demonstration;
efciency is not an issue.
The two blue arrows, marked as NO3 (nitrate ions) and as e (electrons),
point to the continuous ow of negative electric charge across the entire electric circuit, consisting both of the cell and the external load. Ions are the charge carriers in
the electrolyte, while electrons transport the charge in the metal and the external
load. The transition from electronic to ionic charge transport occurs at the electrode/
electrolyte interface upon electron transfer between the electrode and an electron
acceptor or donor in the electrolyte.
In order to characterize the equilibrium state of the electrochemical cells, the
electronic conductor connecting the electrodes of Figure 3.1.5 is removed. Now,
both electrodes may establish their individual charge-transfer equilibrium state. A
stable equilibrium potential difference, E, is established between the two electrodes.
142
H2
Ag
Electrolyte
Ag+NO3
H+NO3
H+
Ag+
Ag+ + e Ag
Ag+ + e Ag
Membrane
H2 H+ + e
H2 H+ + e
Figure 3.1.6 The electrochemical cell of Figure 3.1.5 under equilibrium condition. The device
that measures the equilibrium cell voltage E has such large an internal resistance that the current ow across the cell is practically zero.
If both electrode processes operate under standard conditions, this voltage is E 0, the
equilibrium standard electrode potential difference. Values of E and E 0 may be conveniently measured with electrometers of so large an internal resistance that the current ow is nearly zero. Figure 3.1.6 illustrates the measurement and the equilibrium
state. The value of E 0 is a most signicant quantity characterizing the thermodynamics of an electrochemical cell. Various important features of E and E 0 will be
addressed in the following chapters.
Consider replacing the load of the cell of Figure 3.1.5 by an electric power supply.
If a voltage of U = E (or U = E 0) is applied to the electrodes, the cell equilibrium
voltage will be exactly compensated (i.e. one observes zero current ow), and the
equilibrium state will establish, as illustrated in Figure 3.1.6. If this power supply
is set to apply a voltage of U > E (or U > E 0) to the cell, the cell operation changes
in a fundamental way. The direction of the current ow is reversed. Both electrode
reactions are reversed (i.e. the H2/H+ electrode is now the cathode at which H2 is produced). At the Ag/Ag+ electrode, silver is oxidized to Ag+; it is now the anode. At
U > E, the cell operates as an electrolytic cell. If this system were used as a storage
battery, this operation at U > E would be termed charging the battery, while under
the conditions of Figure 3.1.5 the battery would be discharged at U < E.
143
(8)
The cell compartment containing the electron donor (in the system of Figure 3.1.5:
H2) is the one in which electron transfer proceeds from the solution into the electrode. The electrons may pass through the external load to the cathode in the second
half cell, with electron transfer to the acceptor (Ag+ in the electrolyte). Any electric
work generated in a galvanic cell is energy produced in the corresponding chemical
reaction. Consider a chemical reaction in the general form
aA + bB cC + dD
(9)
11
where G 0f represents the standard free energies of formation. According to Equation (6), G 0f values may be determined from standard enthalpies and entropies of
formation. These data are conveniently available in data collections such as the
Handbook of Chemistry and Physics (CRC Press). The data for the cell of Figure
3.1.5 are listed in the following Table 3.1.2.
Table 3.1.2 Values of the standard enthalpies and entropies of formation for the products
and reactants of the chemical reaction proceeding in the electrochemical cell of Figure 3.1.5,
see Equation (8). Note that the listed values for H+soln are zero because by convention the
hydrogen electrode under standard condition is the reference system.
Chem. reaction
Ag+soln
H2
H0f kJ mol1
105.8
S0f
J mol
73.3
Ag
H+soln
(130.6)
42.7
105.8
95.9
144
In the last column of Table 3.1.2 (), the standard reaction enthalpies and entropies (products reactants) are presented. Using the denition of the Gibbs free
energy (6), one obtains for this reaction at T = 298 K
G 0(reaction Equation [8]) = 77.2 kJ mol1
(12)
(13)
Considering Equation (4), it is clear that one has to operate a galvanic cell at the
maximum possible cell voltage in order to maximize the electric energy yield. As
shown previously (Figure 3.1.6), the maximum value of U is the equilibrium
electrode potential difference E or E 0. Thus, one may formulate the fundamental
relationship between chemical and electric energy:
G(reaction) = nFE
(14)
G 0(reaction) = nFE 0
(15)
Consider again our model system with both the Ag/Ag+ and H2/H+ electrodes at
equilibrium. Under standard conditions, the Gibbs reaction free energy was shown
above to be G 0 = 77.2 kJ (g-equivalent)1. Thus, according to Equation (15),
E 0 = G 0/nF = 0.80 V. And indeed, this standard equilibrium cell voltage may
be conveniently measured with the cell in the conguration of Figure 3.1.6.
One may short circuit galvanic cells (e.g. the one in Figure 3.1.5) by connecting the
electrodes with an electronic conductor of zero resistance. This forces the potential
difference between the electrodes to U = 0 V. In this situation, the electrode reactions
will proceed at maximum rate (i.e. at maximum galvanic current ow). This current
will be limited only by the internal resistance of the cell that is, by the interfacial
charge-transfer resistances and by the ohmic resistance of the electrolyte. Note
that under this condition, no electric energy is harvested, as W = Q U = 0. The energy
dissipates as in the regular chemical reaction. The cell heats up.
Clearly, in galvanic cells of practical interest (i.e. in batteries and fuel cells), the
voltage of operation U has to be a compromise. It must be smaller than E for obtaining an adequate current ow. On the other hand, it has to be close to E in order to
recover as much as possible electric work, W.
145
c d
RT
a a
ln Ca D
nF
a A a bB
(17)
146
To illustrate this, consider again the model cell of Figure 3.1.6. According to
Equation (17), the equilibrium cell potential, E, may be formulated as follows:
E=E0
=E0
RT
aAg aH+
ln
p
nF aAg+ pH2
(18a)
RT aAg
RT
aH+
ln
ln p
nF aAg+ nF
pH2
(18b)
In Equation (18b), the activity quotient is separated into the terms relating to the silver electrode and the hydrogen electrode. We assume that both electrodes (Ag+/Ag
and H+/H2) operate under the standard condition (i.e. the H+/H2 electrode of our
cell happens to constitute the SHE). This means that the equilibrium voltage of
the cell of Figure 3.1.6 is identical with the half-cell equilibrium potential E 0(Ag+/
Ag) = 0.80 V. Furthermore, we note that the activity of the element silver is per definition unity. As the stoichiometric number of electrons transferred is one, the Nernst
equation for the Ag+/Ag electrode can be formulated in the following convenient and
standard way:
E=E0+
RT
ln aAg+
F
(19)
(20)
147
Electrode reaction
Li+ + e Li
3.045
2.714
Na + e Na
+
E0 in Volt
Al3+ + 3 e Al
+ 2 e Zn
1.662
0.763
Pb++ + 2 e Pb
0.126
Zn
++
2 H + 2 e H2
0.000
Cu++ +2 e Cu
+0.337
I2 + 2 e 2 I
+0.535
Fe3+ + e Fe++
+0.770
Ag + e Ag
+
+0.799
Cl2 +2 e 2 Cl
F2 + 2 e 2 F
+1.360
+2.866
RT
pH(1) pH(2)
F
(21c)
148
purication of the metal. Some of the contaminants present in the anode copper will
remain in the electrolyte; others form a solid deposit in the cell, so that the copper
deposited on the cathode (electrolytic copper) contains fewer impurities than the
original metal.
(22)
If one makes the reasonable assumption that H(reaction) and S(reaction) do not
signicantly depend on the temperature, Equation (22) may be used conveniently to
calculate the temperature dependence of E. Consider, for example, Table 3.1.2, in
which the values of H 0(reaction) and S 0(reaction) for the cell consisting of the
Ag+/Ag and H2/H+ electrodes were determined. For room temperature (i.e. 298 K),
a value of E 0 = 0.80 V was found. Raising the temperature by 100 K (i.e. to T = 398 K),
a new value of E 0 = 0.70 V results. Note the signicant effect of temperature on
the equilibrium potential.
Differentiating Equation (22), one obtains
E/T = S(reaction)/nF
(23)
Clearly, the temperature dependence of the equilibrium cell voltage is associated with
the entropy change of the underlying chemical reaction. In the galvanic cell considered previously (Ag+/Ag and H2/H+ electrodes), hydrogen is transformed into dissolved H+. This means a reactant in the gas phase becomes a product in the
condensed phase. This is a typical process associated with a negative change of reaction entropy, and thus with a negative temperature coefcient of the equilibrium
potential.
Numerous electrode reactions require substantial overvoltages to proceed at
acceptable rates. A method for enhancing the reaction rate at a given overvoltage
is operation at higher temperature. One has to consider that this might be a futile
effort if the cell voltage (i.e. the equilibrium state of the system) changes unfavorably.
3.1.7 Conclusion
It was the purpose of this chapter to present fundamental thermodynamic concepts
for evaluating electrochemical processes for energy conversion and storage. To illustrate the merits of these concepts, two highly exothermic chemical reactions are
presented in Table 3.1.4, with the corresponding thermodynamic data.
3.1.8 Acknowledgment
149
Table 3.1.4 Selected thermodynamic data of chemical processes relevant for electrochemical
energy conversion.
H0(reaction)
kJ/mol
S0(reaction)
J/mol K
G0(reaction)
kJ/mol at 298 K
2H2 + O2 2H2O(liq)
572
326.3
C + O2 CO2(gas)
393
+2.9
n
4
4
(3a)
It constitutes the reaction underlying the most widely used fuel-cell system. Note rst
that this reaction is associated with a relatively large negative entropy change. According to Equation (6), the free energy that might be converted into electric work
at room temperature, G 0(reaction), is only 83% of the reaction enthalpy, H 0(reaction). This means that the gain in efciency resulting from energy conversion by the
electrochemical process relative to the heat engine is reduced substantially. Dividing
the reported value of G 0(reaction) by 4F, the standard equilibrium cell voltage
E0 = 1.23 V is obtained. Unfortunately, practical fuel cells operate at voltages on
the order of 0.85 V, which reduces the efciency further by approximately onethird. Operation of the fuel cell at elevated temperature will indeed accelerate the sluggish cathodic O2 reduction reaction. However, this effect will be partially compensated
by the negative temperature coefcient of the equilibrium cell voltage, caused by the
negative S 0(reaction) term. Thus, the evaluation of the oxygen/hydrogen fuel cell
on the basis of such fundamental concepts makes this system appear far from ideal.
By contrast, consider the reaction of carbon with oxygen:
C + O2 CO2(gas)
(24)
In this case, the entropy change is positive that is, the G 0(reaction) exceeds the
value of the reaction enthalpy. Thus, conducting this reaction in an efcient fuel
cell would mean an enormous gain in energy conversion efciency in comparison
with burning the coal in a heat engine. Clearly, it would be most desirable to develop
an efcient process in which carbon acts as an electron donor at an electrode. This
fact was pointed out by Wilhelm Ostwald as early as 1894 [7]. Unfortunately,
since the days of Ostwald, no signicant progress has been made in developing
such systems. It would certainly be a rewarding subject of future research.
3.1.8 Acknowledgment
Valuable comments on this work by Prof. Waldfried Plieth, Technical University
Dresden, are gratefully acknowledged.
150
3.1.9 References
1. Gobrecht H. Bergmann Schfer Lehrbuch der Experimentalphysik. Vol. II, Elektrizitt und
Magnetismus. Berlin, Germany: Walter de Gruyter & Co; 1971.
2. Dickerson RE, Geis I. Chemie. Weinheim, Germany: VCH; 1990.
3. Vielstich W, Lamm A, Gasteiger HA, editors. Handbook of fuel cells. Vol. I, Fundamentals
and survey of systems. Chichester, UK: John Wiley; 2003.
4. Bard AJ, Stratmann M, editors. Encyclopedia of electrochemistry. Vol. I, Thermodynamics
and electried interfaces. Vol. VII, Inorganic electrochemistry. Weinheim, Germany: WileyVCH; 2006.
5. Goodisman J. Electrochemistry, theoretical fundamentals. New York: John Wiley & Sons;
1987.
6. Bard AJ, Faulkner LR. Electrochemical methods: fundamentals and application. New
York: Wiley; 2001.
7. Ostwald W. Die wissenschaftliche Elektrochemie der Gegenwart und die technische der
Zukunft. Zeitschrift Elektrotechnik Elektrochemie. 1894;4:1225.
3.2.1 Introduction
The elements and concepts of electrocatalysis were already in place more than
50 years ago. However, the tools to investigate the electrochemical process have developed dramatically since that time. Recently, ab initio (mostly density functional
theory [DFT]) simulations on systems of a size relevant for electrochemistry have
emerged. DFT provides a method to calculate interaction energies, reaction barriers,
and reaction mechanisms from rst principles. Therefore, many of the fundamental
concepts of electrochemistry are now being revisited and broadened into a quantitative
form based on electronic structure simulations. The new microscopic insights can
hopefully help solve some of the technological challenges in electrocatalysis, especially
those related to electrochemical energy conversion. One of the main challenges is the
energy efciency of the water-splitting reaction [1]. See also Chapter 3.3 in this book.
The ab initio simulations should be used where they really can make a difference.
In this context, it is important to realize that trends and differences are much better
described in DFT than absolute numbers. Even semiquantitatively, trends can be obtained. Despite the importance of having an accurate description of the real electrochemical environment for obtaining absolute values, it seems that many trends and
relative features can be obtained within a somewhat simpler framework [2].
In this chapter, the focus will be on trends in electrocatalysis of the water-splitting
reaction or the oxygen evolution reaction (OER), which is the reaction at the
anode side in an electrolysis cell. Furthermore, simple framework for addressing
OER applying DFT simulations will be presented. For further reading, there are
two previous book chapters where the approach has been reviewed [3, 4].
3.2.2 Fundamentals
The overall splitting of water into hydrogen and oxygen reads
2 H2O(l) + energy O2(g) + 2 H2(g)
(1)
This simple reaction is maybe the most studied, and it has been known for a very
long time [5]. The amount of energy that has to enter in reaction (1) should at
least cancel the difference in enthalpy between water and hydrogen and oxygen,
which is 5.92 eV at standard conditions. As there are three diatomic molecules on
the product side and only two liquid water molecules on the initial side, the entropy
152
is higher at the product side. This means that some of the energy can be provided as
heat without violating the second law of thermodynamics. The total work that has to
be provided is 4.92 eV. This is, in most cases, the relevant energy to look at if the
reaction takes place at constant temperature and pressure. However, it can be a difcult engineering issue to keep the temperature constant inside the electrolysis cell as
the reaction is running if only the needed work is provided.
In the electrochemical cell, the redox reaction is separated into two half-cell reactions namely, the reduction at the cathode and oxidation at the anode. The sum of
the two reactions yields reaction (1):
4 H+ + 4e(cathode) 2 H2(g)
(2)
(3)
By applying a sufciently high potential difference between the cathode and the
anode, the chemical potential of the electrons can be changed so that both halfcell reactions become possible. At the anode electrode, the Fermi level has to be
low enough that it can accept electrons from the water molecules; on the cathode
electrode, the Fermi level has to be high so that the cathode can donate electrons
to protons forming hydrogen. In the overall reaction, four electrons are involved
per oxygen molecule. This means that the work that has to be done on each electron
is G e/4 = E e = 1.23 eV, which means that the potential difference that has to
be applied is at least 1.23 V. However, due to losses and sluggish electrocatalysis,
a substantial additional potential must in reality be applied to speed up the reaction.
This extra potential provides a driving force for the water-splitting reaction and is
called the overpotential, . The energy efciency of the electrolysis can be dened as
(E e + )/E e. It is therefore important to design the electrolysis cell so that the overpotential is as small as possible. At the same time, the current should be as high as possible, as the current denes how large, and thereby how expensive, the electrolysis cell
has to be in order to produce hydrogen at a given rate.
(4)
153
If the pH of the electrolyte and the pressure of hydrogen are constant, the chemical
potential of protons and hydrogen molecules are also constant. This means that
the chemical potential (the Fermi level) of the electrons is constant, which denes
a reference for the potential. When a bias is applied, the potentials are hereafter measured relative to this reversible hydrogen electrode reference, which evidentially is
independent of the electrode material.
The work function, however, inuences the amount of transferred charge when the
Fermi levels are aligned. This in turn determines the surface charge and the eld over
the narrow electrochemical interface. In general, it has to be tested to what degree the
surface chemistry is dependent on the eld. For the intermediates involved in water
electrolysis, the adsorption energies are only very weakly inuenced by the eld, and
normally, this effect can be neglected [6].
The reversible hydrogen electrode reference is useful for calculating the free energy
change of a reaction involving protons and electrons [7]. Instead of simulating protons in solution, we can reference the energy to the gas-phase hydrogen molecule.
The adsorption of hydrogen can of course also be referenced to the gas-phase hydrogen, which means that the free energy difference between protons in solution and
adsorbed hydrogen can be calculated.
Volmer
(5)
H+ + e + H* + * 2 H* H2 + 2*
Tafel
(6)
H+ + e + H* H2 + *
Herovsky
(7)
This results in two different possible reaction paths, Volmer-Tafel or VolmerHerovsky. No matter the reaction mechanism, the intermediate is always H*. This
makes hydrogen evolution a very simple case for rational catalyst material design.
If H binds too weakly to the surface, H+ cannot adsorb from the dissolved phase;
if it binds too strongly, it will have difculty leaving the surface for the gas phase.
One would expect the optimal rate when hydrogen at the surface is as stable as
gas-phase hydrogen which by denition has the same free energy as solvated protons and electrons at zero potential relative to the reversible hydrogen electrode [8].
Hydrogen evolution has been studied intensively using DFT [9, 10, 11], and new
catalyst materials have been suggested [1, 12]. However, the hydrogen evolution
catalysis is very efcient on many noble metal surfaces, and in most cases, it is not
the surface catalysis that limits the cathode reaction.
From a catalysis point of view, the hydrogen evolution reaction can in principle be
catalyzed perfectly, and in reality, it is seen that there is almost no overpotential on
154
the cathode side. The main challenge in terms of catalysis is related to the more
complex oxygen evolution reaction at the anode electrode [13, 14].
Solid
Liquid
2H2O
H+
O2 + 4H+ + 4e
Figure 3.2.1 Schematicpicture of the anode interface between electrode and electrolyte at
155
to get the four different oxidation states, a proton together with an electron is removed in each reaction step, and thereby the oxidation state of the oxygen atoms
on the surface is changed. Otherwise, the picture is similar to that of the molecular
catalyst. If the chemical potential is not sufciently low to directly accept the electrons from each of the states, then the reaction will have to overcome a barrier
and the current will be small.
On the electrode surface (see Figure 3.2.1), the four steps could either involve
an HOO* intermediate or the combination of two oxygen atoms on the surface to
form O2.
Reaction mechanism 1
H2O + * HO* + H+ + e
(8)
HO* O* + H+ + e
O* + H2O HOO* + H+ + e (associative)
HOO* O2 + * + H+ + e
Reaction mechanism 2
H2O + * HO* + H+ + e
(9)
HO* O* + H+ + e
O* + * + H2O HO* + O* + H+ + e (dissociative)
HO* + O* 2O* + H+ + e O2 + 2*
The reactions mechanisms the catalyst will follow may depend on the electrode
material. The difference is that in reaction mechanism 2, the last reaction step doesnt
include an electron and proton. It is clear that reaction mechanism 2 can only be relevant if the barrier for recombining the two oxygen atoms is small. Later, it will be
argued why including mechanism 2 in the analysis doesnt change the conclusion.
156
S0
S1
S2
S3
S4
U = 0V
G [eV]
3.2 eV
G2
2
1
U = 1.23 V
0
1
2
U = 1.60 V
2H2O
HO*+
H++e
O*+
2H++2e
HOO*+
3H++3e
O2+
4H++4e
Figure 3.2.2 The free energy diagram of oxygen evolution on 110 facets of rutile RuO2 at three
different potentials: U = 0 V versus the RHE; the equilibrium potential, E e, U = 1.23 V; and the
smallest potential where all reaction steps are downhill, G OER, U = 1.60 V. The 3.2 eV difference
between OH and HOO is shown by the arrow on the right; the smaller arrow to the left indicates
the universal descriptor for OER.
It is the state from which it is most difcult to remove a proton and electron that
determines the smallest potential needed to obtain a high current. This is often
referred to as the potential-determining step. The size of the potential-determining
step can be estimated by just the biggest change in free energy in the free energy
diagram [15]:
G OER = Max [G1, G2, G3, G4]
(10)
(11)
This is a lower limit for the size of the actual overpotential as there could be additional barriers for the reaction steps, but these are not included in the analysis. The
assumption is that trends are captured by the lower limit.
157
(12)
where ja is the kinetic anode current density, j0 is the exchange current, is the
overpotential, and is the transfer coefcient. Here, is considered, but
the model could be made for other values of . The overpotential is written as
; U0 U in order to rewrite the Tafel equation
ja = j0exp[(eU0 eU )/kT ] = jlimitexp[(eU OER eU )/kT ]
(13)
(14)
The physical meaning of the term jlimit is the current density achieved if all surface
reactions are exergonic (i.e. the highest possible turnover frequency per site). The
term jlimit is dependent on the number of active sites per area and potential independent surface reactions. This means that jlimit is dependent on the catalyst material.
However, if similar surfaces are compared (e.g. a set of 110 rutile oxide surfaces),
the number of active sites per area only varies with a few presents, as the lattice constants are very similar, and jlimit can effectively be considered material independent,
unlike exchange current density. In that case, trends in UOER should correlate with
trends in activity, ja.
The assumption made in many studies and in the previous discussion is that
eU OER can be substituted with its own lower bond namely, the size of the potentialdetermining step, G OER without changing trends. This corresponds to assuming
that additional barriers are material independent or at least scale with GOER.
There is a difference between the concept of the rate-determining step and that of
the potential-determining step. It is clear that the exponential in Equation (13) contains the potential-determining step, whereas jlimit could contain the rate-determining
step. The potential-determining step is also rate determining at low overpotentials,
but as the driving force of the reaction is increased via the potential, the size of
this step is decreased until the point where the reaction rate is no longer increased
by increasing the overpotential. At this point, the rate is determined by the ratedetermining step. In experiments, this is not seen very often since the rate at some
point is given by factors other than catalysis, such as transport.
To summarize, optimizing electrocatalytic activity is a matter of achieving the
highest possible current at the lowest possible overpotential. This means that there
are two things to consider: the highest possible rate and the potential at which the
reaction rate saturates. The highest possible current is often determined by transport
of reactants to and product from the surface, which is related to electrode structure,
158
particle size, and many other things, but less related to the catalyst material per se.
Differences in electrocatalytic performance between different materials are therefore
mostly related to differences in overpotential.
In this analysis, nding a good electrocatalyst becomes a matter of making the largest step G14 as small as possible. If it were possible to make the levels S13 independent of each other, then G OER could be only a bit larger than the equilibrium
potential. However, the energy levels are not independent of each other.
In the next section, the analysis from investigating one catalyst material, RuO2,
will be generalized to studying trends in binding as the catalyst material is varied.
G O*
RuO2
G HOO*
G(O2)
G HO*
G2
3.2 eV
G(H2O)
Strong binding
G HO*
Weak binding
Figure 3.2.3 The scaling relations G O*, GHOO* as functions of GHO* are represented with
the dashes lines. The constant difference between GHOO* and GHO* of 3.2 eV is seen. The
number for RuO2 is shown with the dotted line. The sizes of the reaction steps for RuO2 are the
four sections of the dotted line; it is the same information as in Figure 3.2.1 for U = 0 V. Moving the dotted line along the x-axis will give the reaction step sizes for different materials with
different binding of GHO*.
159
is related to the coupling strength, which in the end ensures that the scaling for
carbon species on the surface changes by 1/4 per hydrogen. This also ensures that
molecules scale in the same manner independent of coordination of the metal
atom to which they are binding.
The slope is thus determined just by counting the number of bonds to the surface.
The intercept of the scaling relation is affected by the binding site geometry. Imagine
a scaling relation between an intermediate that binds via two bonds to the surface
and an intermediate with one bond to the surface (e.g. G O*[GHO*]). The slope
will be 2 because HO* binds through one bond, and O* via two bonds. Oxygen has
a preference for hollow site binding, whereas HO* binds on top or via bridges. If
the most stable is plotted for each, the result is one intercept of the scaling relation.
Now the surface is changed and an atom is added. Both O* and HO* bind more
strongly to the added atom, but oxygen has to pay the price of moving to the ontop binding to gain the increased interaction with the under coordinated site. HO*,
on the other hand, already prefers on-top binding, so it gains in the whole interaction.
Please note that the slope of the scaling is conserved, which means that only the intercept changes. The scaling relation is therefore moved a bit toward strong HO* binding
at a given oxygen binding strength.
That means that only two things can change the scaling: the number of bonds and
the binding site. The number of bonds is related to the intermediates; the binding site
can be changed by defects or by isolating metals atoms like in oxides where only ontop binding is possible [18]. However, comparing two intermediates (e.g. HO* and
HOO*) that have the same number of bonds to the surface and that prefer the
same binding geometry, there is nothing that can be varied. The binding of both
HO* and HOO* can change, but the scaling relation, GHOO*(GHO*), stays the
same. This is a universal scaling relation. As HO* and HOO* bind via one bond
and prefer the same binding site, this means that GHOO*(GHO*) = GHO* +
3.2 eV on all metal and metal oxide surfaces. In Figure 3.2.3, G O*, GHOO* as functions of GHO* aresketched. The example is the GHO* value for RuO2; each crossing between the vertical line and the lines representing the free energy of H2O,
GHO*, G O*, GHOO*, and O2 is the size of the reaction step for RuO2 for U = 0
V, and it is the same information as in Figure 3.2.2.
160
outside the general trends, and there is no oxide-based material that provides an optimum binding of both HO* and HOO*. This is seen in Figure 3.2.3 by the fact that
there is no value of GHO* where the reactions steps all have G14 = 1.23 V. With
the universality of the scaling relation, there is no hope of obtaining the ideal catalyst just
by tuning the binding energy. The challenge is to nd a way to modify oxide surfaces or
the electrochemical interface, such that the relative stability of HOO* and HO* changes.
Maybe some molecular catalysts can do this by providing three-dimensional binuclear
reactive sites.
3.2.5.4.1 Descriptor and Activity Volcano Given the constant difference between
the HOO* and HO* levels, the variation in the overpotential, OER, from one surface
to the next is determined by the O* adsorption energy. This means that either step 2
(G2 in Figure 3.2.3 is the biggest) or step 3 (G3 in Figure 3.2.3 is the biggest) is
potential determining:
G OER = Max[G2, G3] = Max[(G O* GHO*), (GHOO* GO*)]
Max[(GO* GHO*), 3.2eV (GO* GHO*)]
(15)
The difference, G2 = (GO* GHO*), is therefore a unique descriptor for the OER
activity, and the theoretical overpotential at standard conditions is
OER = {Max[(GO* GHO*), 3.2 eV (GO* GHO*)]/e} 1.23 V
(16)
G3/e
Equilibrium potential
OER
Potential
RuO2
Descriptor (G2)
Figure 3.2.4 The volcano for OER. The potential and overpotential for RuO2 is indicated by
the circle and the arrow.
3.2.7 References
161
Please note that this only relates to the catalysis; materials may not be suited
for electrodes for many reasons, such as surface area, stability, and conductivity.
In reality, all of these issues are of course important.
When comparing to experiments, it has to be taken into consideration that no
measured quantity directly corresponds to OER as it is calculated; however, the assumptions were that differences in measured catalytic activity at least should scale
with the size of the potential-determining step and that barriers should not signicantly change that. In OER experiments on rutile and perovskites, the potential at
a given current density is seen to scale very nicely with the calculated OER [17].
Very often, the reaction path 1 applied in the analysis is considered another
assumption. However, it is not. Reaction mechanism 2 can only be relevant if
2O* O2 (2G O* > G [O2]; see Figure 3.2.3) is downhill in free energy, and this
is the case for weaker bindings than that on RuO2. But for weaker binding, it is
G2 that is potential determining. This means that the reaction mechanism could
go through 2O*, and on some of the surfaces with intermediate binding close to
the top of the volcano, it probably will. However, this will not change the analysis
of the potential-determining step.
3.2.6 Conclusion
It is possible to capture the trends in OER activities observed in experiments with the
simple analysis discussed previously. The hypothesis is that the main difference in
activity between two catalysts is related to the difference in the potential needed to
drive the most difcult reaction step. The main assumption is that barriers along
the reaction scale with the reaction free energies; see also the original literature
[13, 17, 21].
The somewhat depressing nding here is that the efciency of OER is limited by a
fundamental scaling relation between the binding of HO* and HOO*. And as RuO2
is close to the top of the volcano, there is no hope of nding a much better catalyst
just by tuning the binding to the surface. The challenge of making electrolysis
more energy efcient can thus be reformulated: the challenge is to break the scaling
relations and stabilize HOO* relative to HO*. This could possibly be done at
three-dimensional binuclear reactive sites.
3.2.7 References
1. Whitesides GM, Crabtree GW. Dont forget long-term fundamental research in energy
Science. Science. 2007;315:796.
2. Nrskov JK, Bligaard T, Rossmeisl J, Christensen CH. Towards the computational design
of solid catalysts. Nat Chem. 2009;1,37.
3. Rossmeisl J, Greeley J, Karlberg GS. Electrocatalysis and Catalyst Screening from Density Functional Theory Calculations. In: Koper M, editor. Fuel cell catalysis: a surface
science approach. Hoboken, NJ: Wiley-VCH; 2009. Chapter 3.
162
The rst half of the nineteenth century witnessed the advent of important electrochemical energy storage and energy conversion devices. In 1800, Volta developed his
famous pile that enabled, for the rst time in history, continuous electrical currents. Around the same time, Nicolson, Carlisle, and Ritter [1] used the Volta pile
to split water into its constituents, hydrogen and oxygen. The elusive origin of the
continuous electric currents generated by the Volta pile created a long-lasting controversy among leading scientists (Volta controversy) about whether electricity is
caused by metal contact action or by a chemical process. Scientists explored a
large number of variations of the original voltaic cell (voltaic battery) involving
many combinations of different electrode materials and electrolytes. These efforts resulted in the invention of the Leclanche element and the lead-acid battery, which are
the basis of two important types of energy storage batteries of our times.
In 1838, between court appearances, a 27-year-old lawyer and later judge, Sir
William Robert Grove, experimented with platinum wires, mineral-acid electrolytes,
and voltaic piles. In a postscript of a letter from December 1838 [2], in which Grove
reported additional experiments on the Voltaic Series, he describes an important
illustration of the combination of gases by platinum. Two electrochemically cleaned
platinum strips were brought in contact with a common mineral-acid electrolyte.
As soon as the upper half of one Pt strip was exposed to hydrogen gas, while the
other was exposed to oxygen, a galvanometer needle indicated the ow of electricity.
Grove called this galvanic cell device a gas voltaic battery [3]; today, we refer to
it as a hydrogen/oxygen fuel cell [1, 4, 5]. A few years later, Grove proved that his
gas voltaic battery, where hydrogen and oxygen combined to water, can be used
to drive the opposite reaction, the splitting of water into oxygen and hydrogen
(Figure 3.3.1). At around the same time, Christian Friedrich Schoenbein recognized
that it was a chemical action between the gases and the platinum wires rather than
the metallic contacts that caused the combination of hydrogen and oxygen [6, 7].
During the 1890s and the following decades, galvanic cells received full recognition
as highly efcient energy conversion devices. In a 1894 keynote address to chemists,
physicists, and chemical engineers, the German physical chemist Wilhelm Ostwald
characterized galvanic cells as a technical revolution far superior to the steam
engine [8]. In particular, scientists and engineers focused on the realization of coalair fuel cells, which remained elusive despite decades of intensive research work.
During the 1930s and 1940s, research work shifted from coal to hydrogen as fuel,
combined with alkaline liquid electrolytes. KOH electrolytes offered the advantage
of reduced corrosion and the possibility of employing less expensive nonnoble-metal
164
ox
hy
ox
hy
ox
hy
ox
hy
Figure 3.3.1
Scheme of a series of gas voltaic batteries (lower portion) consisting of 4 individual galvanic elements involving pairs of platinum wires (black lines), partially exposed to
hydrogen and oxygen gas (upper portion of wire) and partially exposed to a liquid mineral
acid (lower portion of wire). The cell voltage is sufcient to electrochemically split water
into hydrogen and oxygen in a separate set up (upper portion). Arrows are to indicate the
ow of electricity.
electrocatalysts such as Ni. In the early 1960s, nally, fuel cell technology reached a
turning point when the National Aeronautics and Space Administration decided to
develop alkaline fuel cells to power auxiliary units of the Apollo space modules. In
parallel, work on novel solid polymer membrane fuel cells intensied, which replaced
the liquid electrolyte. Early ion-exchange membranes were based on polystyrene and
carried acid sulfonic groups and thus represented a highly acidic proton exchange
ionomer. Now nonnoble metals had to be replaced by acid-stable noble metals,
mostly platinum. Subsequent milestones in polymer electrolyte membrane fuel cell
(PEMFC) technology included the development of chemically resistant peruorinated proton exchange membranes, such as Naon [9], and the use of Naoncontaining fuel cell electrodes incorporating high-surface area-supported platinum
nanoparticles rather than low-surface-area platinum black [10]. Since the early
1990s, driven by public and automotive industry funding, PEMFCs have been receiving
increased attention as an alternative power source for vehicles.
Oxygen
Fuel (H2)
Fuel Cell
H2 + O2 H2O
165
H2O (l/g)
Electricity
Figure 3.3.2 Principle of a fuel cell as an electrochemical energy conversion device. Inside a
fuel cell, fuel, e.g. hydrogen, and an oxidant, typically oxygen, combine electrochemically to
form products, e.g. water, and electricity and some excess heat (not shown). Figure adapted
from ref. [4]
well as some dissipative heat (Figure 3.3.2). Depending on the chemical nature of the
fuel and oxidant on which the fuel cell is operated, different chemical products can be
generated. A hydrogen-oxygen fuel cell shown in Figure 3.3.2 generates only water
as the by-product in the overall chemical reaction
H2 +
1
2
In accordance with thermodynamic laws, only the Gibbs free energy, G o, of the
overall fuel cell reaction can be converted into the equivalent electric cell potential,
Eo; these two quantities are linked via
Go = nF E o
(1)
where F is the Faraday constant and n denotes the number of electrons exchanged in
the overall chemical process.
The thermodynamic efciency of a fuel cell is dened as the ratio between Go and
the enthalpy of reaction, Ho, = Go/Ho, and is not, unlike thermal external or
internal combustion engines, limited by the ideal Carnot cycle.
166
H2
O2
Eo
Chemical
input
power
Proton
flow
H+
Electrolyte
Anode
Cathode
Figure 3.3.3
Principle of a hydrogen and oxygen fuel cell galvanic element. Hydrogen and
oxygen are fed to the anode and cathode electrocatalyst, respectively, giving rise to a cell
potential, E o. Upon closing the external circuit, protons and electrons are continuously created
at the anode. Protons migrate through the electrolyte to the cathode, while electrons ow
through the external wiring to the cathode where they recombine with oxygen and protons
to water. The ow of electrons generates an external electric current, i.
and the electrons that traveled through the external circuit to water in the
oxygen-reduction reaction (ORR) according to
1
2
O2 + 2 H+ + 2e H2O
Table 3.3.1 shows an overview of the most common types of todays fuel cells. Fuel
cells are typically categorized by the type of ion conductor or electrolyte employed.
An important class of fuel cells is based on proton-conducting (acidic) electrolytes,
either in the form of a solid membrane (PEMFCs) or a liquid acid, possibly absorbed
inside a polymer matrix (phosphoric acid fuel cells). The PEMFC has been the subject of tremendous research since the early 1990s, with impressive progress achieved
in the areas of stability of the ion-conducting polymer and in the area of design and
fundamental understanding of the Pt and Pt-alloy electrocatalysis of oxygen reduction,
hydrogen oxidation, and the oxidation of small organic molecules [1113].
A second class of fuel cells employs hydroxide-conducting (alkaline) electrolytes,
again either in form of a solid membrane (alkaline membrane fuel cells) or a liquid electrolyte (alkaline fuel cells). While the modern era of fuel cells began with the latter type,
the former type is under intense research today because a stable, highly conducting
alkaline membrane with good CO2 tolerance has remained elusive to date.
KOH
Solid polymer
Solid polymer
(Naon)
Phosphoric acid
Lithium and
potassium
carbonate
Solid oxide
electrolyte (yttria,
zirconia)
Alkaline membrane
fuel cell
Proton exchange
membrane fuel cella
Molten carbonate
fuel cell
Operating
Temperature
60C120C
50C100C
50C100C
~220C
~650C
~1000C
Charge
Carrier
OH
OH
H+
H+
CO2
3
O2
Electrolyte
Pure H2
Fuel
>50%
>50%
40%
35%45%
<35%
35%55%
Electric Efciency
(System)
2 kWMW range
Stationary
(200 kWMW)
Stationary (200kW)
Automotive, stationary
(5250 kW), portable,
under research
Power Range /
Application
Overview of the principal fuel cell types and their characteristics. Table adapted with permission from [1].
Table 3.3.1
167
168
Finally, there are two fuel cell types operated at such high temperatures that
organic polymers or liquid electrolytes could not withstand them. Their electrolytes
are either CO32-conducting molten carbonates () or O2-ion-conducting oxidic
solids. Depending on the electrolyte and temperature, different fuels and catalysts
can or must be employed. While hydrogen or a liquid organic compound are the
fuels of choice for the low-temperature fuel cells, high-temperature fuel cells can
be operated on a wider variety of fuels, including small hydrocarbons or COcontaminated feeds, because the high temperatures allow for additional conversion
chemistry of the hydrocarbon fuels and mitigate catalyst poisoning. Acidic electrolytes generally require more noble-metal electrocatalysts, while alkaline fuel cells
tolerate less expensive catalysts. High-temperature fuel cells often incorporate oxidic
electrocatalytic materials.
Carbon
2 nm
Pt
20 nm
Pt
Membrane
H2
H+
e
O2
e
Anode
H2 -> 2H+ + 2e
Hydrogen oxidation
reaction (HOR)
O2
Cathode
H+
H2O
Figure 3.3.4 (A) Scanning electron microscopic cross-sectional view of an individual MEA of
a low-temperature PEMFC. The Naon membrane is sandwiched between the electrode layers
(left: anode; right: cathode) followed by the two layers of porous carbon (gas diffusion layers).
Hydrogen diffuses across from left across the gas diffusion layer to the anode where the electrocatalytic oxidation of molecular hydrogen occurs. Protons migrate across the Naon membrane, and electrons travel through the external circuit (dotted arrows). Oxygen enters the
MEA from the right. (B) Enlargement of the cathode layer reveals the platinum nanoparticle
electrocatalyst supported on a porous carbon support material. (C) Further enlargement of the
cathode electrocatalyst shows an individual Pt nanoparticle consisting of well-ordered Pt
atoms at the edge of the carbon support. Black arrow indicates a Pt surface atom where the
electrocatalysis occurs.
169
Table 3.3.1 also indicates suitable power ranges of the various fuel cell types and
some of their typical applications. PEMFCs are clearly the most versatile class with
strong focus on portable and automotive applications. High-temperature fuel cells
are more often employed for stationary power generation.
Figure 3.3.4 provides electron microscopic cross-sectional images of an individual
membrane-electrode assembly (MEA) of a PEMFC (Figure 3.3.4A). The protonconducting membrane is at the center of the MEA with typical diameters of 25
100 m. On either side of the membrane are the electrocatalyst layers with thicknesses of only approximately 510 m. Next on either side of the catalyst layers
are the gas diffusion layers, typically porous carbon ber materials that help distribute the fuel and oxidants uniformly across the catalyst layer, help remove reaction
products, and, as electronic conductors, are part of the external circuit for the electron ow. Figure 3.3.4B reveals the detailed structure of the cathode electrocatalyst
layer showing nanoscale (210 nm) Pt particles that are supported on high-surfacearea carbon black supports. Figure 3.3.4C shows the atomic ne structure of an individual Pt nanoparticle. A regular arrangement of individual Pt atoms is evident.
Pt surface atoms (black arrow) act as active sites for the ORR at the cathode of
hydrogen/oxygen fuel cells.
The detailed chemical structure of the most common proton-conducting polymer,
Naon, is shown in Figure 3.3.5A. Stabilized and tethered using backbone chains
of peruorinated polyethylene (Teon), sulfonic groups are dangling inside the ionomer and help transport protons across the membrane. As illustrated schematically in
Figure 3.3.5B, the microscopic structure of a hydrated Naon membrane reveals hydrophilic water channels in between self-organized hydrophobic Teon domains. The
sulfonic groups are reaching into the water channels and mediate the diffusion and
migration transport of hydrated hydronium ions (H3O+). Ionomer membranes are
A
CF3
SO3
n
O S
H2O
H3O+
1 nm
O H+
(B) Schematic illustration of the microscopic structure of hydrated Naon membrane: peruorinated polyethylene backbone chains form spherical hydrophobic clusters. Sulfonic end
groups interface with water-lled channels and mediate the migration and diffusion of protons.
The channels are lled with water and hydronium ions. Figure adapted from [4].
170
Cell
Cathode
171
the membrane and the external circuit caused by their combined ionic and electronic
resistance, Rohmic. Concentration overpotentials become noticeable only at higher
current densities where large amounts of fuel and oxidants are converted at the catalysts. Under these conditions, mass transport of fuel and oxidant can become rate
limiting and result in a loss of cell voltage at a given current density. Figure 3.3.7 illustrates how the total overvoltage of a fuel cell (i.e. the difference between the theoretical cell voltage, Eo, and the real cell voltage, V ) splits into the three overvoltage
contributions. Mathematically, the experimentally observed cell voltage, V, can be
expressed and modeled according to the relation
V = Eo act,cathode,ORR act,anode,HOR iRohmic conc,anode
conc,cathode
(1)
Figure 3.3.7 demonstrates that the activation overvoltages are by far the largest cause
of loss of voltage and hence efciency of a fuel cell. This is why over the past decade a
great deal of research has focused on the discovery and the fundamental understanding of improved fuel cell electrocatalysts. The anode and the cathode reaction are not
contributing equally to the total activation overvoltage. This is illustrated in Figure
3.3.8, which plots the experimental current density versus the electrode potential near
their standard electrode potential for the hydrogen electrode (H2/H+, standard equilibrium potential 0 V) and the oxygen electrode (O2/H2O, standard equilibrium
potential 1.23 V). These current potential characteristics of redox systems can be
Theoretical cell behavior
Eo
Open circuit
potential
Cell voltage V
0.1
Activation polarization
(losses due to sluggish kinetics)
Total power loss
Ohmic polarization
(resistance loss)
0.5
Concentration polarization
(transport loss)
Real cell
behavior
Figure 3.3.7 Theoretical (dashed dotted) and real (solid) cell voltage (V ) current density (I )
performance characteristics of a fuel cell. Overpotentials are responsible for the difference
between theoretical and real performance and cause efciency losses. They split into (i) activation polarization overpotentials at anode and cathode due to slow chemical kinetics, (ii) ohmic
polarization overpotential due to ohmic voltage losses along the circuit, and (iii) concentration
polarization overpotentials due to mass-transport limitations. The activation overpotentials of
the cathode are typically the largest contribution to the total overvoltage.
172
4 H+ + 4 e
2 H2
O2 + 4 H+ + 4 e
j/mA cm2
HOR
2 H2 O
OER
jcell
V
0V
act, ORR
act, HOR
1.23 V
jcell
HER
ORR
oxygen fuel cell. The solid curves represent exponential analytic current densities versus electrode potential of the hydrogen electrode (standard potential 0 V) and the oxygen electrode
(standard potential 1.23 V). Relevant for a PEMFC fuel cell are the HOR (anode) and the
ORR (cathode) branches. To satisfy a cell current ( jcell), the anode potential moves more positive by act,HOR, while the cathode potential moves more negative by act,ORR. As a result of
this, the observed cell potential is V, which is smaller than 1.23 V. The shape of the individual
characteristics is such that the cathode overpotentials are larger than those at the anode.
well approximated by exponential expressions rst derived by Butler and Volmer, assuming the Gibbs free energy of activation of the forward and reverse reaction being
a linear function of the electrode potential [14]. Figure 3.3.8 schematically illustrates
how activation overpotentials of the individual electrode reaction depend on the
shape of their Butler-Volmer-type j-E relation. For the HOR, a given positive, anodic
current density, jcell, causes a relatively small overpotential, HOR, while for the much
less reversible ORR, the corresponding negative reductive cell current results in a
much larger ORR overpotential, ORR. In fact, ORR typically exceeds HOR by at
least two orders of magnitude. This is because there is no material for which there
is a measurable ORR current at the equilibrium potential of 1.23 V. This makes
the j-E curve of ORR very broad and at around its standard potential. Again, consistent with Figure 3.3.7, the difference between the two standard redox potentials of
1.23 V (equivalent to the standard cell voltage) and the two activation overpotentials
combined generates the experimental cell voltage, V. Figure 3.3.8 suggests that developing an improved ORR electrocatalyst offers much larger gains in cell voltages and
hence efciencies, which is why much effort in the area of fuel cell catalysis is currently directed toward the identication of more active electrocatalyst materials for
the ORR.
The rational design of improved electrocatalyst materials requires rst and foremost a fundamental understanding of the origin of the kinetic overpotential. Over
173
HOOad + 3 H+ + 3e
G1
(2)
HOOad + 3 H+ + 3e
H2O(l) + Oad + 2 H+ + 2e
G2
(3)
G3
(4)
H2O(l) + HOad + H+ + e
G4
(5)
2 H2O(l)
As the schematic illustration of this four-step mechanism in Figure 3.3.9 shows, there
are three reaction intermediates, HOOad, Oad, and HOad, which are adsorbed on the
surface of the electrocatalyst.
Based on Sabatiers principle [20] stating that the rate of a catalytic reaction is
maximized at intermediate binding strength between catalyst and reaction intermediates and drops at very large and very low catalyst-reactant interaction, diagrams of
the Gibbs free energy reaction pathway provide valuable insight into why the ORR
is not proceeding at its equilibrium potential. Figure 3.3.10A plots the Gibbs free
energy of the four elementary reaction steps (2) to (5) over the reaction coordinate
for a Pt-containing electrocatalyst. At the equilibrium electrode potential of +1.23 V
(red pathway), the Gibbs energies of the chemical and electrical species are such
that reactions (2), (4), and (5) are endergonic (positive driving force, Gi > 0). This
keeps practical reaction rates negligibly low. Only reaction (3) (G2 < 0) could proceed
spontaneously.
174
O2
G1
O2Had
e +H+
H2O
G2
Oad e +H+
OHad
G3
e +H+
G4
H2O
Catalyst
Figure 3.3.9 A simple four-step reaction mechanism of the ORR involving three adsorbed
To model the effect of the electrode potential, the electrochemical framework links
changes in the Gibbs free energy of the electrons, G, to variations in the electrode
potential, E, through a simple linear relation, G = nF E, where n is the number
of electrons on the left side of reactions (2) through (5). Thus a decrease of the
applied electrode potential lifts all the Gibbs energy levels in Figure 3.3.10A, however, at different rates. As a result of this, G1, G3, and G4 gradually decrease
until, at +0.81 V, all steps have become downhill in energy and hence can proceed
with measurable rates. Lowering the electrode potential further does not result in enhanced rates. In the context of this simple electrochemical framework, the last step to
become downhill is referred to as the potential-determining step [18, 19].
The dependence of the Gibbs free energy pathway on electrode potential (Figure
3.3.10A) manifests itself directly in the experimental current potential characteristic
illustrated in Figure 3.3.10B. At 1.23 V, no ORR current is measureable, while with
decreasing electrode potentials the ORR current increases exponentially until at
+0.81 V, processes other than surface kinetics (e.g. mass transport) begin to limit the
overall reaction rate. Figure 3.3.10B represents a typical performance characteristic
of a Pt or Pt-alloy electrocatalyst for the ORR.
O2
+4
+*
+4
H*
3H
+3
OO
3.0
V
Gibbs free energy/e
2.5
U = +0.81 V
2.0
175
+
O*
2H
e
+2
+H
2O
+
H*
1.5
+e
+H
2O
*+
1.0
2H
2O
0.5
G
G 1
0.0
U = +1.23 V
0.5
0.7
0.8
G3
1.0
0.9
1.0
Reactio
n
1.1
coordin
ate
1.2
0
E = 1.23 V
j/mA cm2
1
2
3
4
5
E = 0.81 V
6
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
1.3
Figure 3.3.10 (A) The electrode potential dependence of the Gibbs free energy reaction path-
way of the ORR. While the overall reaction has elementary steps that are energetically uphill
at +1.23 V (red pathway), all elementary steps become downhill at +0.81 V (yellow pathway)
(i.e. at an overpotential of approximately 0.42 V. At this point, the reaction is not limited by
kinetics anymore. (B) The experimentally observed current-potential ( j-E) relation of the ORR
is consistent with the computational conclusions from (A): between +1.23 V and +0.81 V the
j-E curve shows an exponential behavior, while at electrode potentials below +0.81 V, the
ORR reaction rate becomes oxygen mass-transport limited, which is reected by a at ( j-E)
prole. Figure adapted with permission from [19].
176
0.0
Pt
Pd
0.5
Ir
1.5
Rh
Ni
Ag
Cu
Co
Au
2.0
Mo
Fe
W
2.5
Pt-O bond
too weak
Stronger bond
EO (eV)
Weaker bond
Figure 3.3.11
dcompressed
Pt
Pt
177
dstretched
Pt
Interatomic distanced
Figure 3.3.12 The effect of surface lattice strain on adsorbate chemisorption energies. Lattice
strain modulates chemisorption and can be used to tune the reactivity of electrocatalysts. Compressed Pt surface layers (right portion) bind adsorbates more weakly; stretched Pt layers bind
adsorbates more strongly.
pure Pt (red vertical dotted line through top of volcano); however, there is no metal
in the periodic table that exhibits that optimal binding energy.
How can we modify or even deliberately tune surface chemisorption energies of
catalytic surfaces? A popular strategy is the mixing of two dissimilar metals in the
surface of the catalyst. The distinct electronic structures of the two metals interact
and, depending on their relative atomic ratio, can generate a surface with a new electronic structure, which is different from either one of the components. This alloying
strategy works well, though the interaction is not simple. For instance, mixing Au
and Co in various ratios would not necessarily generate a surface with intermediate
oxygen chemisorption properties closer to the top of the volcano.
A more controlled strategy to deliberately tune chemisorption energies is displayed
in Figure 3.3.12. Theory predicts that the interatomic distance of surface atoms (surface lattice constant) affects the chemisorption energies in a nearly linear fashion [23].
A surface layer of compressed (compressively strained) Pt atoms like the one shown
to the left of Figure 3.3.12 is predicted to bind oxygen weakly, while a stretched Pt
surface (tensile strain) is predicted to bind oxygen more strongly.
A way to stretch or compress metal surface atoms in a controlled way is to deposit
them on top of a substrate with similar crystal symmetry, yet with different atomic
diameter and lattice constant. Such a single monolayer of a metal supported on
another is called an overlayer. Metal overlayers strive to approach the lattice constant of their substrate without fully attaining it; hence, they are strained compared
to their own bulk state [24, 25]. The choice of suitable metal substrates enables tuning
of the strain in the overlayer and of the chemisorption energy of adsorbates. A Pt
monolayer on a Cu substrate, for instance, was shown to bind adsorbates much
weaker than bulk platinum due to compressive strain induced by the lattice mismatch
between Pt and Cu, with Cu being smaller [26].
178
Pt
Core
Shell
Cu
5 nm
Figure 3.3.13 (A) Pt core-shell nanoparticle ORR electrocatalysts are prepared by (electro)
chemical selective removal of a less noble metal, M, from a Pt-M nanoparticle alloy. The
near-surface region is Pt enriched, while the particle core remains bimetallic. (B) Transmission
electron microscopic elemental map of dealloyed Pt-Cu core-shell nanoparticle ORR
electrocatalysts.
treatment. The less noble metal Cu atoms are leached out from the surface of the particle down to a depth that can be controlled by the corrosion conditions. The leaching
of the less noble component from a uniform alloy is referred to as dealloying.
In the end, a nanostructured core-shell catalyst particle is obtained characterized
by an in expensive nonnoble-metal-rich inner core, surrounded by a noble-metal
particle shell of controlled thickness (core-shell architecture). Depending on the
thickness of the shell and the atomic diameter of the nonnoble element, the noblemetal surface atoms are kept in a state of compressive or tensile strain, depending
on whether the nonnoble-metal atom is smaller or larger than the noble one,
respectively.
Figure 3.3.14 displays the results of voltammetric ORR activity measurements of a
dealloyed Pt-Cu and a dealloyed Pt-Ni core-shell catalyst under fuel cell relevant
conditions. The typical sigmoidal current density electrode potential (i-E) shape
(compare to Figure 3.3.10B) of the Pt-based catalyst is clearly evident. The large
179
Dealloyed PtCu3/C
i [mA/cm2geometric]
Pt/C
Dealloyed PtNi3/C
6
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Figure 3.3.14 Experimental ORR activity of dealloyed Pt-Cu and Pt-Ni core-shell nanopar-
ticle ORR catalysts compared to a pure-Pt nanoparticle catalyst. All three catalyst particles are
supported on a high surface area carbon material indicated by the sufx /C. The shift of the j-E
curve of the core-shell catalysts indicates the onset of oxygen reduction catalysis at a more anodic electrode potential (equivalent to a lower overpotential) and hence represents improved
ORR reactivity compared to pure Pt.
shift of the i-E characteristic of the core-shell catalysts evidences a catalytic reactivity
at more positive electrode potentials, implying less potential loss versus the thermodynamic standard potential of 1.23 V for the oxygen electrode with reference to the
hydrogen electrode; this also implies improved efciency of the ORR process on the
dealloyed materials. More specically, the core-shell catalysts exhibit an ORR activity that is about six to eight times higher than commonly employed state-of-the-art
pure monometallic Pt nanoparticle electrocatalysts.
Subsequent deployment of the new catalyst in the cathode layer of small-area
MEAs rst, then large-area MEAs, and nally fuel cell stacks represents the typical
series of performance tests to check the practical viability of novel ORR electrocatalyst materials. Figure 3.3.15A shows the experimental cell voltagecurrent density
characteristics (compare to Figure 3.3.7) of three dealloyed Pt-M (M = Cu, Co, Ni)
nanoparticle ORR cathode electrocatalysts compared to a state-of-the-art pure-Pt catalyst. At current densities above 0.25 A/cm2, the Co- and Ni-containing cathode catalysts perform comparably to the pure-Pt standard catalyst, even though the amount
of noble metal inside the catalysts is lower than that of the pure-Pt catalyst by a factor
of two to three. The dealloyed Pt-Cu catalyst is even superior to Pt at reduced metal
loading.
To inspect and compare the activation overvoltage of the three catalysts in more
detail, so-called Tafel plots are used, which plot the cell voltage as a function of
the logarithm of the current density. Figure 3.3.16B shows the Tafel plots derived
from Figure 3.3.15A. At a cell voltage of 0.9 V, where the overall reaction rate is limited by the chemical surface catalysis, the dealloyed core-shell catalysts perform three
180
0.95
Cell voltage, V
0.90
0.5
1.0
1.5
2.0
B 0.96
45wt% Pt/C
0.94
Pt/Co3/C
PtNi3/C
0.92
Cell voltage, V
PtCu3/C
0.90
0.88
0.86
0.84
0.82
0.80
0.1
Figure 3.3.15 (A) Cell voltagecurrent density (E-j) proles of a 10 cm2 single PEMFC in-
corporating dealloyed Pt-M (M = Cu, Co, Ni) ORR catalysts at the cathode. The active catalysts were prepared by electrochemical dealloying of PtM3 alloys. The alloy performance is
shown in comparison to a reference PEMFC with a pure-Pt nanoparticle catalyst at the cathode. Dealloyed Pt-Cu nanoparticles outperform Pt and the other alloys over a wide current
density range. (B) (Elog j) Tafel plots of the PEMFC proles of (A). Pt-based mass activity
(A/mgPt) is chosen as activity measure. Dealloyed Pt-Cu and dealloyed Pt-Co outperform pure
Pt by a factor of three to four times at 900 mV cell voltage. Conditions: Tcell = Tanode =
Tcathode = 80C; H2/O2, fuel/oxygen excess = 2/10; total pressure = 150 kPa; MEA active
area = 10 cm2; membrane = NR212. Figure taken from [27].
181
A
1
Pt
3
4
5
6
5
6
0.0
0.2
0.4
0.6
0.8
N
Fe2+
Figure 3.3.16 (A) Current density-potential plots of the ORR activity of noble-metal free
182
to ve times more actively than pure Pt, conrming the ORR performance results
obtained in liquid electrolytes presented in Figure 3.3.14.
3.3.5 Conclusions
Fuel cells are versatile electrochemical energy conversion devices. They represent an
integral component of an emerging strategy toward a future sustainable energy
3.3.7 References
183
infrastructure for stationary and automotive applications, where hydrogen will play a
major role as chemical energy carrier. Major advantages of fuel cells over combustion engines involve fuel exibility and efciency. PEMFCs play a particularly
important role due to their wide power range.
The efciency of fuel cells is largely limited by the kinetic barriers of the surface
catalytic electrode reactions. In particular, the electroreduction of molecular oxygen
at a PEMFC cathode severely limits high reaction rates and hence currents near the
equilibrium cell voltage.
Modern DFT has become a powerful tool to understand, predict, and discover
electrochemical catalysts with improved ORR activity and stability. Computational
free energy reaction diagrams provide insight into the potential-determining elementary reaction step of the ORR as a function of atomistic descriptors (surfacerelated properties) of the catalyst material. DFT-based volcano relations have
been established pointing to improved catalyst systems.
From combined theoretical and experimental insights, nanostructured Pt coreshell electrocatalyst architectures have recently emerged as promising, cost-effective
cathode fuel cell catalysts. Pt-enriched multilayer surface shells surround Pt-poor
cores that modify the reactivity of the surface Pt layer.
Beyond human-made Pt- or noble-metal-based ORR electrocatalysts, there exist
very active biomimetic carbon-nitrogen-iron ORR electrocatalysts that show great
potential for use in PEMFC cathodes, even rivaling the catalytic activity of pure Pt.
3.3.6 Acknowledgments
The author thanks Professor Ifan Stephens, Professor Ib Chorkendorff, Dr. Shirlaine
Koh, Dr. Ratndeep Srivastava, Dr. Prasanna Mani, Nastaran Ranjbar, Dr. Lin
Gan, Camillo Spoeri, and Koteswara Vuyurru for help with the graphical material.
The author is deeply indebted to Professor Anders Nilsson for his past encouragement
and support.
3.3.7 References
1. Hoogers G. Fuel cell technology handbook. Boca Raton, FL: CRC Press; 2003.
2. Grove WR. On voltaic series and the combination of gases by platinum. Philos Mag J Sci
Ser 3. 1839;14:12730.
3. Grove WR. On a gaseous voltaic battery. Philos Mag J Sci Ser 3. 1842;21:41720.
4. OHayre R, Cha S-W, Colella W, Prinz FB. Fuel cell fundamentals. New York: Wiley;
2006.
5. Vielstich W, Lamm A, Gasteiger H. Handbook of fuel cells fundamentals, technology,
and applications. Chichester, UK: Wiley; 2003.
6. Schoenbein CF. On the voltaic polarization of certain solid and uid substances. Philos
Mag J Sci. 1839;14:435.
7. Schoenbein CF. On the theory of the gaseous voltaic battery. Philos Mag J Sci Ser 3.
1843;22:1656.
184
8. Ostwald W. Die Wissenschaftliche Elektrochemie der Gegenwart und die Technische der
Zukunft. Z Elektrotech Elektrochem. 1894;4:1225.
9. Grot W. Peruorierte Ionenaustauscher-Membrane von hoher chemischer und thermischer Stabilitt. Chem Ing Tech. 1972;44:1679.
10. Raistrick ID. Modied gas diffusion electrode for proton exchange membrane fuel cells.
In: Van Zee JW, White RE, Kinoshita K, Barney HS, editors. Proceedings of the Symposium on Diaphragms, Separators, and Ion-Exchange Membranes; Pennington, NJ: The
Electrochemical Society; 1986. Vol. PV-86. p. 172.
11. Lipkowski J, Ross PN. Electrocatalysis. New York: Wiley-VCH; 1998.
12. Koper M. Fuel cell catalysis a surface science approach. Hoboken: NJ: Wiley; 2009.
13. Vielstich W, Gasteiger HA, Yokokawa H. Handbook of fuel cells: advances in electrocatalysis, materials, diagnostics and durability. Chichester, UK: John Wiley & Sons; 2009.
Vol. 56.
14. Bard AJ, Faulkner LR. Electrochemical methods. New York: Wiley; 1980.
15. Norskov JK, Rossmeisl J, Logadottir A, et al. Origin of the overpotential for oxygen
reduction at a fuel-cell cathode. J Phys Chem B. 2004;108:1788692.
16. Hoare JP. The electrochemistry of oxygen. New York: Wiley; 1968.
17. Kinoshita K. Electrochemical oxygen technology. New York: Wiley; 1992. p. 448.
18. Rossmeisl J, Karlberg GS, Jaramillo TF, Nrskov JK. Steady state oxygen reduction and
cyclic voltammetry. Faraday Discuss. 2008;140:33746.
19. Stephens IEL, Bondarenko AS, Grnbjerg U, Rossmeisl J, Chorkendorff I. Understanding the electrocatalysis of oxygen reduction on platinum and its alloys. Energy Environ
Sci. 2012;5:674462.
20. Sabatier P. Hydrognations et deshydrognations par catalyse. Ber Dtsch Chem Ges.
1911;44:19842001.
21. Man IC, Su H-Y, Calle-Vallejo F, et al. Universality in oxygen evolution electrocatalysis
on oxide surfaces. ChemCatChem. 2011;3:115965.
22. Greeley J, Stephens IEL, Bondarenko AS, et al. Alloys of platinum and early transition
metals as oxygen reduction electrocatalysts. Nat Chem. 2009;1:5526.
23. Mavrikakis M, Hammer B, Norskov JK. Effect of strain on the reactivity of metal surfaces. Phys Rev Lett. 1998;81:281922.
24. Kibler LA, El-Aziz AM, Hoyer R, Kolb DM. Tuning reaction rates by lateral strain in a
palladium monolayer. Angew Chem Int Ed. 2005;44:20804.
25. Zhang JL, Vukmirovic MB, Xu Y, Mavrikakis M, Adzic RR. Controlling the catalytic
activity of platinum-monolayer electrocatalysts for oxygen reduction with different substrates. Angew Chem Int Ed. 2005;44:21325.
26. Strasser P, Koh S, Anniyev T, et al. Lattice-strain control of the activity in dealloyed core
shell fuel cell catalysts. Nat Chem. 2010;2:45460.
27. Mani P, Srivastava R, Strasser P. Dealloyed binary PtM3 (M = Cu, Co, Ni) and ternary
PtNi3M (M = Cu, Co, Fe, Cr) electrocatalysts for the oxygen reduction reaction: performance in polymer electrolyte membrane fuel cells. J Power Sources. 2011;196:66673.
28. Lefvre M, Proietti E, Jaouen F, Dodelet J-P. Iron-based catalysts with improved oxygen
reduction activity in polymer electrolyte fuel cells. Science. 2009;324:714.
3.4.1 Introduction
Photosynthesis is the most important chemical process on our planet [1]. The invention of water splitting by early photosynthetic organisms led to an oxygen-rich atmosphere and the protective ozone layer in the stratosphere, which was important for
the development of all higher life on Earth. Photosynthesis stores the suns energy
in the form of energy-rich organic compounds (carbohydrates) via CO2 reduction.
It is thus the only basic source of food, biomass (e.g. hay, wood, cotton, etc.), and
all fossil fuels such as coal, oil, and natural gas. The rapid exploitation of these fossil
fuels over the last two centuries, which are not renewable on a human timescale,
has led to inevitable shortages. Furthermore, the burning of carbon-rich fuels has
increased the concentration of the greenhouse gas CO2 in the atmosphere, leading
to climate changes with many adverse effects for our planet and human society.
Human population and economic growth, particularly in fast-developing countries like India and China, will lead to further increases in energy demand. The economic, social, and political consequences of a shortage of fossil fuels are already
being felt today. It is therefore imperative to identify alternative sustainable energy
sources. An obvious choice is to exploit solar energy, which has enormous potential
as a clean, abundant, and economical energy source. The problem is developing technologies that allow this energy source to be efciently captured and stored in a useful
form, such as heat, electricity, and in chemical fuels. Whereas the rst two forms are
already quite well developed, the use of solar energy to generate a useable chemical
fuel still represents a major challenge [2].
Chemical bonds are the best way to store energy. The efcient production of a
clean, storable solar fuel would represent a very important breakthrough in scientic research. Such a fuel must be made from abundant, inexpensive, and nontoxic
materials such as water, which could be split into molecular oxygen and hydrogen.
Hydrogen is considered the ideal primary fuel of the future [3] since its combustion
(e.g. in fuel cells) generates only water as a waste product. Furthermore, it can be
converted to many other important energy-rich materials (NH3, CH4, CH3OH) for
example, for further storage and transport. These fuels also have uses in other
industrial and agriculture sectors, such as ammonia for fertilizers (see Figure 3.4.1).
All these processes, however, require appropriate catalysts.
A promising starting point for the development of a synthetic light-driven watersplitting catalyst is to use natures approach for inspiration. A simple schematic of
186
ATP
NADPH
2x 2H+
Biomass/food
Rubisco
CO2 reduction
Hydrogenase
H2 production
Mn
Mn
NH3
CO2
CO2
CH4 CH3OH
Storage
O2-Cat
Mn
Mn Ca
2 H2O
N2
P
Biomimetics
H2 Dihydrogen
Solar Fuel
RC
Solar Fuel
H2-Cat
4H+ + 2O2
2 H2O
O2, 4H+
Solar fuel production systems. Left: Scheme for light-induced charge separation and water splitting in oxygenic photosynthesis (i.e. in the photosystem II [PSII] reaction
center, RC), which drives the production of energy carriers (ATP, NADPH) that are subsequently used to make energy-rich compounds via CO2 reduction (food, biomass). The electrons can also be redirected to reduce excess protons to dihydrogen via the enzyme
hydrogenase (biological H2 production). Right: Biomimetic system showing multiple lightinduced charge separation in a photo(electro)catalyst P and transport of redox equivalents
to a light-driven water-splitting catalyst (anode) and a light-driven hydrogen producing catalyst (cathode). The produced primary solar fuel is dihydrogen, in which the suns energy is
stored. H2 can be directly used (e.g. in combustion engines or in fuel cells to generate electricity) or can be converted for easier storage, transport, or other uses. The examples given include
ammonia production from nitrogen and reactions with carbon dioxide to yield methane or
methanol. These reactions also require efcient catalysts.
Figure 3.4.1
the multiple steps seen in oxygenic photosynthesis, the process by which water is split
into O2 and protons, and hydrogen production, the process by which protons are reduced to H2, is shown in Figure 3.4.1. Oxygenic photosynthesis primarily occurs in a
single biological supercomplex, photosystem II (PSII), which is found in green plants,
algae, and cyanobacteria. Many green algae and bacteria also contain an enzyme
that allows hydrogen evolution, hydrogenase. The efcient coupling of these two
biological processes in vitro and in vivo is currently being investigated in many
laboratories for the development of biohydrogen technology [4, 5]:
photosynthesis
hydrogenase
(1)
187
efcient way that is synthesize a catalyst, with low overpotential, high rates (turnover frequencies, TOFs), and long lifetimes (turnover numbers, TONs). One major
difculty is the little understood coupling of the light-induced one-photon/oneelectron charge separation to the multielectron catalytic processes leading to water
oxidation and fuel production. In addition, nature uses a smart matrix the protein
to optimize the enzymatic process. This matrix also provides protection and allows
repair in vivo. In this respect, chemically synthesized systems are currently by far inferior to the native ones. This shows that another problem is the sensitivity and limited
lifetime of the in vitro isolated native enzymes and current articial systems; for example, most of the hydrogenases and H2-generating catalysts are sensitive to molecular
oxygen, which is produced in the water-splitting process (Figure 3.4.1). The lesson
to be learnt from nature is that catalysts are not static on the contrary, they are
dynamically involved in the reaction process.
In recent years, high-resolution X-ray crystallographic structures of PSII [7] and
different hydrogenases [810] have been obtained, and additional spectroscopic
and electrochemical experiments have signicantly increased our knowledge about
the structure and function of the native enzymes [11, 12]. Slowly, a mechanistic understanding is developing for the underlying processes in all their details. It is clear that an
understanding of the basic principles of water oxidation and hydrogen conversion in
nature is a goal of major importance in scientic research and for our society. This
would allow us to design better photosynthetic organisms capable of (bio)hydrogen
production and at the same time advance the eld of articial photosynthesis, aimed
at synthesizing new catalysts for large-scale water splitting and hydrogen production and for energy storage in chemical bonds in general, processes that are of
key importance for a sustainable future hydrogen economy.
(2)
3.4.2.1 PSII
PSII is a multisubunit protein complex embedded in the thylakoid membrane of
photosynthetic organisms [17]. PSII puried from the thermophilic cyanobacterium
188
Cytoplasm
Cyt b-559
Cp43 D1 D2 Cp47
Cytoplasm
QA
Fe
QB
PheoD1
PheoD2
Cyt b-559
ChlzD1
ChlD1
PsbU
PsbO
Lumen
Pheo Pheo
cyt
b-559
P680
YZ
YD
Mn4O5Ca
D
H,O2
CP43 Lumen
CP47
Mn3Mn3
H2O
Thylakoid membrane
D2 Fe D1
QB
QA
2 H2O
4 H
H,e
h
H,e h
h
Mn
c550/24
S0
S1
4
12/16
Mn3Mn4
S4
Mn3Mn4
33
ChlzD2
Mn4O5Ca Cluster
Stroma
O2
Car
ChlD2 D2
TyrD
Lumen
C
Thylakoid membrane
Tyrz
CarD1
PsbV
PD1 P
D2
P680
Mn
4
or
Mn4Mn4
Mn4Mn4
S3
h
H,e
S2
Mn3Mn3
Mn4Mn4
e
Mn3Mn4
Mn4Mn4
H2O
Figure 3.4.2 (A) X -ray crystallographic structure of the PSII dimer from Thermosynechococcus vulcanus [7] at 1.9 resolution showing the major protein subunits in color and (B) the
arrangement of the cofactors in the central D1/D2 proteins involved in charge separation, electron transport, and water oxidation (for details, see text). (C) Simplied scheme of the core
components of the water-splitting PSII complex embedded in the thylakoid membrane of
chloroplasts or cyanobacteria. Only the relative positioning of the ve most important transmembrane subunits and of three extrinsic proteins are shown (the full complex consist of about
20 subunits). The two central subunits carrying the electron transport chain and the Mn4O5Ca
cluster (WOC) are the D1 and D2 proteins. Channels for water and for H+ and O2 are indicated starting from the sequestered WOC (white oval). (D) The S-state cycle proposed by
Kok [13]. The different states of the Mn4O5Ca cluster that are attained after various numbers
of light ashes (h) are termed Si states, where i = 0, . . . , 4 indicates the number of oxidizing
equivalents stored in the cluster. The oxidation states of the Mn ions are given for each S state
(high valance model) [14,15]. The release of protons is also indicated [15,16].
Thermosynechococcus (T.) vulcanus crystallizes as a dimer (Figure 3.4.2) that contains about 2,800 solvent water molecules [7]. Each monomer consists of about 20
protein subunits that harbor 77 cofactors: 35 chlorophylls (Chl a); 11 -carotenes;
2 plastoquinones (PQ); 2 pheophytines (Pheo a); 1 Mn4O5Ca complex; 2 heme Fe;
1 nonheme Fe; and 1 hydrogen carbonate, HCO3/CO3. The overall reaction of
PSII is that of a light-driven water:plastoquinone oxidoreductase [1]:
4hv
2 H2 O + 2 PQ + 4 H+stroma!
O2 + 2 PQH2 + 4 H+lumen
189
(3)
PSII uses visible light (400700 nm) to drive the water-splitting reaction. The
chemical energy derived from this process is temporally stored as reduced plastoquinol (PQH2) that is subsequently used to generate reduced nicotinamide adenine dinucleotide phosphate (NADPH), one of the two major energy carriers of biology.
Concomitant proton transfer (H+) across the thylakoid membrane results in a
buildup of a proton gradient across this membrane. This pH difference drives the
synthesis of the second major energy carrier of biology, adenosine triphosphate
(ATP). Both NADPH and ATP are required for carbon xation; thus PSII, together
with PSI (which is not addressed here), is primarily responsible for the majority of
biological energy storage that is derived from sunlight.
The reactions that occur in PSII can be divided into four processes:
1. Light harvesting and energy transfer by the chlorophyll (Chl) and carotenoid molecules of the antenna (A) complexes to the reaction center (RC) of PSII.
2. Excitation and primary charge separation of the RC, a multipigment assembly of
four Chl a and two pheophytin a (Pheo a) molecules (Figure 3.4.2B), which are
bound to the D1 and D2 protein scaffold. Light excitation initiates a single electron transfer within the multipigment assembly, resulting in a charge-separated
(radical pair) state [18]. The radical cation (primary donor) generated is thought
to predominantly reside on a single Chl a pigment (PD1), which is commonly
referred to as the pigment radical P680+ according to the position of its optical
absorption maximum. It has an estimated oxidizing potential of +1.2 to +1.3 V,
one of the highest known in biology [19]. The radical anion of the primary acceptor generated is the PheoD1 (the Pheo a bound to the D1 protein). This chargeseparated state is stabilized by subsequent electron/hole-transfer steps from both
the primary donor and acceptor. The PheoD1 passes its electron to the bound
PQ plastoquinone molecule QA, whereas P680+ is reduced by a redox active tyrosine side chain of the D1 protein, YZ. In this way, the distance between the electron and electron hole increases suppressing recombination reactions, leading to
an excellent quantum efciency of PSII for water splitting (>90%). However,
the multiple electron-transfer reactions lead to a decrease in G, reducing the
overall energy efciency of photosynthesis [20].
3. Reduction of plastoquinone QB by QA and protonation at the acceptor side of
PSII. The QA is tightly bound to the protein, acting as a one electron acceptor.
It passes electrons to a second plastoquinone, QB, which can accept two electrons
and two protons and acts as a mobile electron carrier connecting PSII to the next
complex of the photosynthetic apparatus (i.e. the cytochrome b6f complex). After
two electron-reductions and two protonation events, QBH2 leaves the reaction
center and is replaced by an oxidized quinone from the pool in the membrane.
4. Accumulation of oxidizing equivalents on the donor side of PSII and water splitting;
the S-state cycle (Figure 3.4.2D). For water oxidation and O2 generation to occur,
four electrons must be removed from two water molecules. As the RC only generates one electron hole per photon absorption, the PSII complex must store
190
Ala344
CP43-Glu354
C2
Asp342
O1
O2
W4
O4
C2
191
Glu189
D1
B3
2.60
O5
A4 2.50
1.79
1.90
Asp170
B3
D1
W3
A4
W1
2.89
W2
2.97
2.84
3.29
2.76
2.67
2.81
3.35
2.74
2.81
2.81
3.36
His332
Glu333
Figure 3.4.3 (A) The structure of the WOC as seen in the latest crystal structure of Umena
et al. [7] (B) Mn4O5Ca inorganic core seen in the crystal structure of Umena et al. (C) Best
geometry-optimized DFT structure of the WOC reported by Ames et al. [33]. (D) Optimized
DFT structure that contains a protonated O5 bridge. The best DFT structure (C) simultaneously reproduces the EXAFS constraints (i.e. Mn-Mn distances and so forth) and the
magnetic resonance data (see Figure 3.4.4). See Ames et al. [33] for further details. Purple
spheres Mn; red spheres oxygen; green sphere Ca.
192
ENDOR amplitude
S2: Mn4O5Ca(Sr)
3
2
JAB
4
A4
1H
B3
5 6
C2
JBC
Asp342
JCD
JBD
193
Glu189
D1
C2
D1
IV
III
Ca2+
His332
Sr2+
Mn
Ca Open Coord.
IV B3
Site (H2O/OH
O
N
A4 IV
55Mn
Data
Sim. (OEC)
Sim. (Mn2+)
4
1H 1
5 6
100
300
200
Radio frequency VRF (MHz)
C
W1, W2
W3, W4
EDNMR amplitude
S2: H216O/H217O
W4
W3
17O
D1
O5
W2
III
His332
O5
His332
IV
14N
IV
W1
A4
C2
IV
B3
10
Vmw - Vmw
20
(0)
30
(MHZ)
Figure 3.4.4
(A) 55Mn-ENDOR spectra (Q band) of the WOC of PSII from Thermosynechococcus elongatus native (Ca) and with Ca2+ replaced by Sr2+ [41]. Insert: Electronic connectivity ( Y-coupling scheme) of the WOC with four exchange couplings J between Mn ions [14].
(B) Structure of the WOC poised in the S2 state according to Cox et al. [41] EPR/55Mn
data together with theoretical modeling suggest the MnIII ion is within the cuboidal unit
(MnD1) and has a ve-coordinate tetragonally elongated ligand eld. (C) 17O ELDORdetected NMR spectra (94 GHz) of the WOC exchanged in H217O and poised in the S2 state
[46]. (D) Structure of the Mn4O5Ca cluster with -oxo bridges and bound water molecules W1
to W4 (color coded: Mn purple spheres; Ca yellow sphere; N blue; H white spheres;
O (nonexchangeable) red spheres; O (exchangeable) pink/orange/green).
194
195
TON. While the sites of substrate binding remain under debate, the kinetics of
substrate binding are well understood as determined from time-resolved membrane-inlet mass spectrometry (MIMS) studies, which employ a rapid jump in
H218O concentration within specic ash sequences [22]. These experiments
have identied a possible role for the Ca2+ in binding one substrate water, as substrate binding is inuenced by biochemical Ca2+/Sr2+ exchange. Similarly, they
also establish that the two substrate molecules bind at separate sites, as their
exchange rates vary independently with Si state. The slower-exchanging substrate
water (WS) can be detected in all Si states, suggesting that it binds earlier in the
S-state cycle (S0), while the faster-exchanging substrate (Wf) has only been resolved in the S2 and S3 states. Fourier transform infrared spectroscopy (FTIR)
measurements support these ndings, suggesting that one water molecule, possibly
Wf, binds during the S2 S3 transition [49].
EPR spectroscopy is a powerful tool for determining the identity (H2O/OH/O2)
and location of water-derived species bound in the vicinity of the Mn4O5Ca cluster.
This can be achieved via the detection of the coupling of magnetic nuclei [1H/2H (I =
1/2, 1) or 17O (I = 5/2)] to the paramagnetic states (S0, S2) of the Mn4O5Ca cluster.
The measured 1H/ 2H couplings for the WOC as detected via electron spin echo envelope modulation (ESEEM) or ENDOR are consistent with terminal Mn-H2O/ OH
ligands (i.e. 1H nuclei 23 from a Mn) [46, 50, 51], as seen in the recent crystal
structure (MnA4-OH/OH2), but inconsistent with a protonated bridge species, at
least in the S2 state. A more direct probe for water-derived ligands is via the direct
detection of oxygen using 17O labeling as this isotope is still sensitive to the fully deprotonated substrate (i.e. the incorporation of water into a Mn--oxo bridge). Recently, this topic has been addressed using high-eld 17Oelectron-electron-double
resonance detected NMR [46]; 17O couplings could be readily observed for the
WOC (Figure 3.4.4C). Importantly, the induction of these 17O signals the rate
with which the bulk solvent water (H217O) exchanged with the bound Mn-16O species occurred on the same timescale as observed in the MIMS experiment already
discussed. Three classes of nuclei were assigned (1) a -oxo bridge, (2) terminal
MnA4-OH/OH2, and (3) Ca/second shell-OH/H2O ligand(s) by comparison to
17
O-labeled model complexes. The exchangeable -oxo bridge was assigned to the
O5, which links the outer Mn to the Mn3O3Ca (open-cubodial) unit (Figures
3.4.3B and 3.4.4D). A similar -oxo bridge signal was observed in Q-band 17OENDOR experiments [52].
196
The experimental evidence already described limits the scope of question (1). At
least one substrate site, presumably the slow substrate WS bound in the early S states
(S0S2), must be either a ligand to the Ca (W3 or W4) or a ligand to the MnA4 (terminal H2O/OH (W1 or W2) or a 2-oxo bridge (O5). From these restrictions, the
subsequent mechanism by which the two substrates couple is grouped into two classes:
1. A nucleophilic-attack mechanism between two adjacent substrate oxygen atoms
bound to the Ca2+ and the MnA4 as a terminal or bridging ligand (class 1a; i.e.
W3 and W2/O5) [55, 56] or between substrate oxygen atoms bound to MnA4
(class 1b; i.e. W2 and W1/O5) [31, 57]; see Figure 3.4.5(I).
2. An oxo/oxyl radicalcoupling mechanism between the exchangeable -oxo bridge
(O5) and a second oxygen species (oxyl radical) not yet formed/bound in the
S2 state [32]; see Figure 3.4.5(II).
la
Ws = W3
W4
3e
2H
Wf = W2
D1
A4
O4
Mnv=O
C2
O5
W1
D1
A4
B3
lb
C2
B3
W4
W3
Ws = W2
C2
O5
Wf = W1
A4
O4
W4
D1
C2
A4
B3
Wf
D1
D1
W2
W1
3e
2H
B3
W3
II
Mnv=O
D1
Ws
A4
3e
2H
C2
O5
C2
A4
B3
B3
O4
55Mn
40Ca
Figure 3.4.5
16O
17O
1H
17O
(exchangeable, substrate)
197
Nucleophilic attack in the S4 state. A terminal oxygen, W2 or the 2-oxo bridge O5,
is progressively deprotonated during the S-state cycle such that in S4 it is a strong
electrophile. In the simplest (terminal ligand) case, this species can be considered a
Mn(V) = O; however, it may equally be described as a Mn(IV) ; O+ or Mn(IV)-O
species. During the S3YZ S4 S0 transition, the electrophile is attacked by a nucleophilic water molecule. This may either be a water/hydroxo bound to Ca (W3 or
W4) or bound to the Mn (W1) (Figure 3.4.5). This type of mechanism requires a Mncentered oxidation during the S2 S3 transition and a Mn-centered or ligandcentered (substrate) oxidation during the S3 S4 transition. The nucleophilic attack
mechanism has been previously observed in Mn model systems that perform OO
bond formation. However, these systems display turnover rates orders of magnitude
slower than that of the WOC [58].
Oxo/oxyl radical coupling in the S4 state. There are a number of class II type
radical-coupling mechanisms in the literature (for reviews, see references [53, 54]).
Here, we limit our discussion to the most detailed and rigorous proposal at present,
the mechanism proposed by Siegbahn [32]. In this catalytic cycle, the slowly exchanging substrate is considered to be the O5 -oxo bridge between MnA4 and MnB3. The
fast substrate water binds then at the open coordination site on the MnD1 as water/
hydroxo in S2/S3, forming an oxyl radical in S4 [32] (see Figure 3.4.5[II]). This type
of mechanism allows for the possibility that OO bond formation can occur in the
S3 state [59] that is, the WOC contains a complexed peroxide in the S3 state,
which presumably is in redox equilibrium with various other forms of S3 that may
include an oxygen radical and/or a formal Mn4(IV,IV,IV,IV) state. Upon formation
of the S3YZ state, only centers that contain the complexed peroxide conguration
are able to donate an electron to YZ and liberate O2. Thus, the rate of the S3 S4
transition reects the equilibrium constants between the different S3 redox states and
would follow directly the time course of YZ reduction as observed experimentally.
In contrast to the nucleophilic-attack-like mechanism, the radical-coupling mechanism
has no precedence in manganese chemistry. It is, however, the energetically favorable
pathway for efcient OO bond formation in rare-earth catalysts such as the ruthenium
blue dimer (see [60]).
198
essential for achieving nearly equivalent redox potentials for all oxidation steps
that match the oxidizing potential of the light-generated primary oxidant.
3. The release of the products of the water-splitting reaction (O2 and H+) from the
catalytic site is decoupled.
4. The Mn4O5Ca cluster undergoes several structural changes during the Kok cycle,
which are signicant for the mechanism.
5. The Mn4O5Ca cluster is readily assembled from abundant Mn2+/Ca2+ in solution,
apparently without chaperones or maturation factors. As the PSII protein itself must
be constantly repaired due to photodamage, rapid recovery of the water-splitting
capability (i.e. Mn4O5Ca cluster) is of critical importance.
199
catalyst durability. As the complex self-assembles in solution, like the Nocera catalyst,
it can be considered self-repairing. These systems display TOF of >5 s1 and TON ~75,
operating at mild pH (8) and temperature with an overpotential of 350 mV.
Manganese catalysts: Kurz and coworkers [70, 74] have recently shown that
Mn2O2Ca oxides, known to appear in nature, do display water-oxidation capacity,
at an albeit low TOF, 3.3 104 per Mn. A similar Mn mineral like material is
now thought to constitute the Dismukes catalyst [75], a tetranuclear Mn cubane
[Mn4O4L6] (L = diarylphosphinate) that when suspended in a naon matrix is
capable of efcient water oxidation at low overpotential. Recent X-ray spectroscopic
results [76] demonstrate that the cluster dissociates into MnII upon insertion into
naon, which is then reoxidized to form dispersed nanoparticle structure with
oxidation state MnIII/IV. It is this state that is catalytically active, generating
O2 upon visible photon absorption, returning to the MnII state. Nevertheless,
this photoelectrochemical material exhibits TOF of 0.00750.075 s1 and TON in
excess of 1,000 and has been engineered to operate within a dye-sensitizer cell
device [77].
While the goal of a synthetic WOC has yet to be achieved, the selected examples
given above show that great progress has been made in this eld over a relatively
short timescale.
(4)
The acidity of H2, which is extremely low, is dramatically increased upon interaction
with a metal site. Many of the currently used catalysts for anthropogenic utilization
of hydrogen require precious metals such as platinum, while natures catalysts are
based on cheap and abundant rst-row transition metals (Ni, Fe). An important feature of the native hydrogenase enzymes is the very high TOF, which often reaches
numbers close to 104 s1 [79]. Moreover, hydrogenases from (hyper)thermophilic organisms [80] exhibit good stability at elevated temperatures (up to ~ 100C). The pertinent problem of the high oxygen-sensitivity of many hydrogenases is solved in
species (e.g. Knallgas bacteria) with specically modied cofactors that provide
a greatly enhanced tolerance to the deleterious effects of molecular oxygen [81].
200
201
B
His
Cys
Cys
Cys
Cys Cys
[4Fe-4S]H
CO
H-cluster
Ni
Glu
Active center
Fe
Cys
Arg
CN
Glu
Cys
CN
Fep
Fed [2Fe]H
Cys CO
Pro
CN
Arg Ser
CO
CO
Pro
Ala
Large
subunit
[4Fe-4S]pro
[3Fe-4S]med
Small
subunit
[4Fe-4S]dis
[4Fe-4S]
[4Fe-4S]
Figure 3.4.6 (A) Three-dimensional structure of the [NiFe] hydrogenase from Desulfovibrio
vulgaris Miyazaki F. The small and large subunits are shown in light blue and green, respectively. Top: The [NiFe] active site and its surrounding amino acid residues. The possible hydrogen bonds are depicted with black dotted lines. The third bridging ligand between Ni and Fe is
denoted by X [PDB 1WUI] [11, 91]. The open coordination site at Ni is indicated (green
arrow). The three FeS clusters for electron transport are depicted. (B) Structure of the
[FeFe] hydrogenase from Desulfovibrio desulfuricans [89]. (Top) Ball-and-stick representation
of the H cluster and the surrounding amino acid residues based on the hydrogenases oxidized
form [PDB 1HFE]. The open coordination site at Fed is indicated (green arrow).
3.4.3.2.1 [NiFe] Hydrogenases In Figure 3.4.6, the structure of a [NiFe] hydrogenase is shown as an example of catalytic hydrogenases. The active site is found in
the large subunit (~65 kDa) of the enzyme, whereas the smaller subunit (~38 kDa)
harbors one [3Fe-4S] cluster anked by two [4Fe-4S] centers. The distance between
the FeS clusters is in the range of 12 , which is consistent with biological electron
transfer [92]. The Ni atom in the active site is bound by four cysteine residues of the
protein, two of which form a bridge to the Fe atom, which, in turn, is coordinated by
two CN and one CO ligands, originally identied by FTIR spectroscopy [93]. In the
most oxidized state of the enzyme (as isolated), an additional bridging ligand X
between Fe and Ni is modeled in the crystal structure [94], while in the active reduced
states this electron density is absent. Spectroscopic evidence suggested that this is an
202
oxygen species [95], but the presence of sulfur may also be possible under certain conditions [96]. The Ni atom has a distorted square pyramidal coordination with a
vacant site believed to be the primary binding position of the H2 substrate. The hydrophobic gas channel leads directly from the surface of the protein to the open coordination site of the Ni atom [97]. It is assumed that this gas channel is also involved
in the reactions with CO and O2, which inhibit the enzyme. Indeed, comparative
crystallography show that CO and O2 inhibit [NiFe] hydrogenases, as evidenced
by additional diatomic electron density at the open fth coordination site of the
Ni [98]. In the oxidized unready state of the enzyme, an additional elongated electron density was observed at the bridging ligand X, suggesting it to be a hydroperoxide OOH [91, 99]. It is suggested that the peroxo bridging ligand causes the
enzyme to be locked in the unready state, which can be reactivated only very
slowly. Another metal center has been assigned to a magnesium ion located at the
C terminus [84] and is probably involved in proton transfer between the [NiFe] active
site and the molecular surface.
The subclass of [NiFeSe] hydrogenases contains a selenocysteine replacing one
of the terminal cysteines coordinating the Ni. The ratio of H2-production to H2-oxidation
activity of this enzyme is higher than that for the other [NiFe] hydrogenases, which
makes this enzyme particularly interesting for biohydrogen technologies. Furthermore, the [NiFeSe] hydrogenases exhibit a higher oxygen tolerance and can be
more easily activated than the [NiFe] hydrogenases. The overall structure of this
[NiFeSe] hydrogenase [86] is very similar to that of the [NiFe] hydrogenases (Figure
3.4.6), except for the presence of three [4Fe-4S] clusters in the electron-transfer chain
and an iron ion replacing the magnesium.
For electron transfer to exogenous acceptors [8, 84] three iron-sulfur clusters are
located in an almost linear fashion in the small subunit within a distance of
~10 . Desulfovibrio hydrogenases possess one [3Fe-4S] and two [4Fe-4S] clusters
[8, 84] (Figure 3.4.6A). The distal cluster is coordinated by three cysteine and one
histidine residues in all known species; the histidine residue is exposed to the solvent
accessible region. The [NiFeSe] hydrogenase harbors three [4Fe-4S] clusters. In
oxygen-tolerant hydrogenases (e.g. from Aquifex aeolicus, R. eutropha) [100, 101],
a novel type of proximal [4Fe] cluster has been found that is ligated by six cysteines
and has a [4Fe-3S] core. It has special redox properties and is essential for the defense
mechanism against molecular oxygen [101, 102].
3.4.3.2.2 [FeFe] Hydrogenases Crystal structures of the cytoplasmic [FeFe] hydrogenase from the gram-positive bacterium Clostridium (C.) pasteurianum [90] and the
periplasmic [FeFe] hydrogenase from D. desulfuricans (DdH) [89, 103] have been
determined, the latter in two redox states (oxidized and reduced). Figure 3.4.6B
shows the structure of the [FeFe] hydrogenase isolated from D. desulfuricans [104].
The active site, the hydrogen-converting cluster (H cluster) a unique [6Fe-6S]
cluster is located in the center of the protein. Two additional [4Fe-4S] cubanes
(F clusters) form the electron pathway to the external redox partner of the enzyme.
The [FeFe] hydrogenases differ in the number of iron-sulfur centers. In C. pasteurianum, three [4Fe-4S] and one [2Fe-2S] clusters serve this task, while the algal
203
hydrogenases appear to lack the accessory FeS clusters altogether [105]. In these hydrogenases, the H cluster interacts directly with the ferredoxin redox partner. A hydrophobic gas channel has also been modeled [104]. It ends at the most remote Fe
atom (Fed) of the active site (see arrow), where extrinsic CO also binds [106]. Fed
is thus believed to be the site where H2 initially binds (or H2 is released). The gasaccess channel is much shorter than that of [NiFe] hydrogenase. Finally, a chain
of amino acid residues is identied that could act as proton pathways [104].
The H cluster harbors a classical cubane [4Fe-4S] subcluster connected to the catalytic [2Fe]H subcluster through a cysteine bridge. This subcluster consists of two
octahedrally ligated Fe atoms and ve diatomic nonprotein ligands (i.e. CO,
CN). The CO and CN ligands in the crystal structure have been assigned, making
use of FTIR spectroscopy as well as taking into account the possibility of hydrogen
bonding of the CN groups to nearby amino acid residues [89]. In the case of D. desulfuricans hydrogenase, the CO ligands are located in a hydrophobic pocket,
whereas specic amino acid residues provide H bonds to the CN ligands (Figure
3.4.6B). It is noteworthy that the [2Fe]H subcluster is not bound to the protein
except for the H bonds to the CN ligands. Furthermore, a dithiolate bridge exists
between the two iron atoms. The dithiolate ligand was assigned in D. desulfuricans
to a propanedithiolate, which was later revised to a dithiomethylamine [103]. In
the dithiolate bridge of D. desulfuricans, a nitrogen was conrmed by application
of magnetic resonance techniques as the apical atom that could be involved in
the catalytic cycle of the enzyme [107]. The distance between the distal Fe and the
[4Fe-4S] cluster is ~4 , which is much shorter than the respective distance in the
[NiFe] hydrogenase. This is reasonable for a fast electron-transfer process and for
the high turnover rate of H2 production of the [FeFe] hydrogenases.
204
|||(OH)Fe||
Ni
Ni-B
e, H
H2O
Ni|| Fe||
Ni-Sla
e, H
Ni|||[O]Fe||
Ni-A
Ni|||(H)Fe||
Inactive states
Ni-C
e,
H
Ni||(H)Fe||
Ni-R(1,2,3)
Catalytic cycle
H
Ni||
Fe||
S
H2
Fe||
H
Ni-R
Ni-Sla
e
H
X
Ni||
Ni|||
H
H
H
Fe||
e
H
H2
e
H
e
H
S
Ni|||
Ni-C
Fe||
H
Figure 3.4.7
(A) Proposed reaction scheme for a catalytic [NiFe] hydrogenase detailing the
catalytic cycle and the activation starting from the oxidized state(s). The formal oxidation
states of the Ni and the Fe are given; furthermore, the type of bridging ligand that changes
identity in the cycle is indicated (see text for details). Paramagnetic (EPR-active) states
are in red; EPR-silent states are in blue. (B) Two models showing the catalytic cycle of
[NiFe] hydrogenase proposed by Fontecilla-Camps et al. [114] for the catalytic cycle of
[NiFe] hydrogenase; one cycle is indicated by green, the other by blue arrows (see text).
(Figure 3.4.7). The activation process has been studied in detail using EPR and other
spectroscopies [109, 112]. Ni-A is the unready state of the enzyme and takes up to
an hour to reactivate. In contrast, Ni-B can be reactivated within a few minutes. It
should be noted that in oxygen-tolerant species (e.g. R. eutropha, A. aeolicus) the
Ni-A state is not observed. It is therefore considered the oxygen-inhibited state.
Upon reduction of Ni-B, the enzyme passes through several EPR-silent states to
reach Ni-SIa (silent active). After a two-electron reduction, the EPR-active Ni-C
state is reached. In this redox state, a hydrogen species has been identied using advanced EPR (hyperne sub-level correlation [HYSCORE], ENDOR) spectroscopic
techniques [115, 116]. Figure 3.4.8 shows the characteristic EPR spectrum of Ni-C
of hydrogenase I of A. aeolicus as an example [117]. In addition, the ENDOR and
HYSCORE spectra from the hydrogenase activated using D2 in D2O are depicted.
The obtained 2H hyperne coupling is consistent with a hydride bridging the NiFe unit. Ni-C is light sensitive and upon light exposure forms the Ni-L state, in
which the hydride bridge is reversibly lost [115, 118]. A last reduction step from
Ni-C produces the EPR-silent Ni-R state, which is the most reduced form of the
active site (Figure 3.4.7A). CO gas will inhibit the enzyme. The Ni-CO state is
also photosensitive, and low-temperature illumination results in the same Ni-L
state as obtained from Ni-C. Density functional theory (DFT) calculations of the
electronic structure of the Ni-Fe cluster have been published with possible redox
states of the individual Ni and Fe atoms [119, 120]. It is shown that Fe remains
Fe(II) throughout all redox-accessible states. In the most oxidized (inactive) states
(Ni-A and Ni-B), the Ni is proposed to be Ni(III), while in the intermediate
reduced states (e.g. Ni-SIa), it is Ni(II). Ni-C on the other hand is shown to have
2.21
EPR
9 GHz
2.15
2H
ENDOR
T=7K
1 0.5 0 0.5
VRF V2H [MHz]
205
1106.1
B0, mT
1120.8
1129.2
1134.5
2.01 Ni-C
T=130K
1156.4
v2, MHz
2H/14N
HYSCORE
T=7K
Ni-C Model
D
CN
CN
H
H
14N
4
2H
Fe
Ni
0
0
CO
A. aeolicus
HN(His)
D. vulgaris MF
2
4
v1, MHz
Figure 3.4.8 Spectroscopic characterization of the active Ni-C state of [NiFe] hydrogenase I
of A. aeolicus [117]. (A) EPR (X band, frozen state) showing three characteristic g tensor
components; (B) deuterium ENDOR spectra at different elds yielding two 2H hyperne couplings (red/green shaded area), (C) HYSCORE (X band at g = 2.15) showing resonances from
a coupled 14N (histidine) and the deuterium in the bridge between Ni and Fe (see panel D).
(D) Structural model of the Ni-C state with a bridging hydride, based on the EPR,
ENDOR, and HYSCORE experiments, comparison with data from the NiFe hydrogenase
of D. vulgaris [116]. Adapted from [117].
206
carry the hydride bridge between Ni(III) and Fe(II) [115, 116, 125]. It is then proposed
that further reduction of Ni-C by another H2 molecule leads to the Ni-R state, which
still carries the hydride [126]. A proposal for the mechanistic details of the reaction
with H2 in the actual catalytic cycle has recently been made, which is based on earlier
experiments [124]. The nal step would then be the release of another proton and electron to give rise to the initial Ni-SIa state with an open bridge, ready for the next turnover. This last step could be preceded by relocation of the proton from the bridge to
the terminal cysteine, creating a state that resembles Ni-L [127].
In another model (Figure 3.4.7B) proposed by Fontecilla-Camps et al. [114], the
Ni-R state is formed directly from Ni-SIa with H2. In the catalytic cycle, the hydride
remains in the bridge between Ni and Fe and acts as a base for the next incoming H2.
In this mechanism, the hydrogenase cycles between Ni-R, Ni-C, and a transient Ni-X
state; the latter has a second hydride bound to the Ni. This model assumes that the
two protons are released in two subsequent oxidation steps. H2 production will occur
through the same reverse pathway.
3.4.3.3.2 [FeFe] Hydrogenase Less information on the electronic structure is available from EPR on [FeFe] hydrogenase as compared to [NiFe] hydrogenase since only
the Hox state (and Hox-CO) is paramagnetic. In particular, the hydride-carrying species (Hred) cannot be characterized by EPR techniques. However, FTIR and Mssbauer spectroscopy have revealed important information on the intermediate states
of the H cluster [128, 129].
In Figure 3.4.9A, a scheme for the activation, catalysis, and inhibition of the H
cluster is shown for D. desulfuricans [FeFe] hydrogenase as example. In contrast to
most other [FeFe] hydrogenases, this enzyme can be isolated aerobically. The hydrogenase is then inactive ( Hox
air ) and has to be reductively activated (e.g. by exposing it
to hydrogen gas); this process has been followed spectroscopically [129]. In Figure
3.4.9A, the different states of the protein that were identied are listed together
with the proposed oxidation states of the iron atoms in the binuclear subcluster. In
the as-isolated form, the protein is in an overoxidized state in which both
irons are Fe(II). The free coordination site at the distal iron could be occupied by
a water or OH ligand. This state of the protein is diamagnetic, and the attached cubane subcluster is also oxidized. Activation under hydrogen leads to formation of the
active oxidized state (Hox), which has a characteristic rhombic EPR spectrum. It is
assumed that the binuclear subcluster is in the [Fep(I)Fed (II)] conguration. Continued reductive activation ultimately produces the active reduced state (Hred), which is
EPR silent but can be identied in FTIR and Mssbauer spectroscopy. It is believed
that this state is characterized by a [Fep (I)Fed (I)] conguration. As is indicated in
Figure 3.4.9A, the Hox and Hred states are considered to take part in the catalytic
cycle. The short-lived intermediates that close the reaction cycle are not identied
yet, but clearly, these have to involve species with a bound dihydrogen and/or
hydride. For Chlamydomonas (C.) reinhardtii hydrogenase HydA1, an additional reduced state, Hsred (superreduced), has been identied using FTIR-monitored spectroelectrochemistry [130], which is included in Figure 3.4.9A. It is suggested that
here, similar to the Htrans state, the cubane subcluster is reduced to [4Fe4S]+.2
CO-inhibited
HoxCO
Fe|Fe||[CO]
CO
2
CO 2
H2O
H2
H2
NC
OC
Hsred
Fe|Fe|[H]
e
Hred
Fe|Fe|[H]
2
2
CO
CN
e, H
Fe|Fe||[H2]
C
O
Fed
Hred?
Fe||Fe||[H]
S
Fep
[4Fe4S] Cys
e, H active
Hox
Fe|Fe||[ ]
2e?
H
Htrans
Fe||Fe||[OH]
e
Hox
air
Fe||Fe||[OH]
Inactive
2
2
Fe|
Fe|
Hred
2
2
C
O
H
N
Fe|| [H]
H
N H
Fe|| [H2]
H
N
Fe|I Fe||[H]
O
e
H
H
N
Fe
Fe|
2
2
Active
H2
Fe| [ ]
[]
H
NH
Fe||
H
N
Fe| [H]
Fe|
e
H
(Hred)
2
Hox
ets. The formal oxidation states of the irons in the [2Fe]H and the charge of the [4Fe-4S] subcluster are given. The paramagnetic (EPR-active)
states are in red; the EPR-silent states, in blue; and the transient states, in black. Note that Htrans has not yet been observed by EPR, Hsred is only
observed as a stable intermediate in hydrogenase from algae [130].2 See text for further details. (B) Proposal for H2 heterolytic cleavage/formation
at the active site involving Fed and the amine function. (C) Detailed catalytic cycle according to [11].
Figure 3.4.9 (A) Intermediate states identied for the H cluster. Proposed ligands to the exchangeable site at Fed are indicated between brack-
2
207
208
In this state, FTIR shows that the CO bridge in D. desulfuricans between the Fe ions
has opened, whereas it is still intact in Hred of C. reinhardtii, a nding relevant for the
catalytic mechanism of hydrogen oxidation and evolution at the H cluster.
The enzymatic activity can be effectively inhibited by CO, which binds to the open
coordination site at Fed. The (Hox-CO) state has the same electronic conguration as
Hox and shows a characteristic axial EPR spectrum. The inhibiting CO ligand is
photolabile and can be photodissociated (<40K) leading to formation of the Hox
state [131, 132]. The spectroscopic investigations of the H cluster in its various states
show that the ligand environment of the H cluster is exible and that the electrons
move to and from the binuclear subcluster via the cubane subcluster to the accessory
F clusters or the ferredoxin redox partner (Figure 3.4.6). In addition, 57Fe ENDOR
and ESEEM studies [133] have shown that the two subclusters of the H cluster are in
intimate electronic contact.
An important breakthrough in the understanding of the active site has been the
spectroscopic detection of a 14N atom in the bridgehead of the dithiolate moiety
of the [2Fe]H subcluster using pulse EPR (HYSCORE) methods [107]. This nitrogen
is placed in the perfect position to act as a pendant base for the heterolytic splitting of
molecular hydrogen or for the reverse reaction that is, formation of hydrogen at the
H cluster (Figure 3.4.9B).
The possible catalytic mechanism of [FeFe] hydrogenases is not well established,
since there are fewer redox states accessible to spectroscopy (e.g. by EPR), the
proton-accepting base is still being debated, and the protein structure shows a larger
variability. Also, the DFT modeling of the reaction cycle is more complicated
because the covalently attached cubane [4Fe-4S] subcluster, which seems to play
an important role in the electron shufing, is difcult to include in the calculations.
An overall consensus oxidation/reduction-reaction scheme of [FeFe] hydrogenases is depicted in Figure 3.4.9A, in which it is assumed that electron and proton
transport take place in (almost) simultaneous steps. An important issue is the stability of the hydride intermediate in which the hydride is bound to the Fe(II)Fe(II)
subcluster, either terminally to the distal iron [134136] or in a bridging position
between the two irons [137, 138]. It turns out that both congurations are feasible,
but the bridging conguration is thermodynamically more stable, which might suggest that the hydrogen splitting and proton reduction will occur in the bridging
position. At the same time, however, it has been demonstrated by Hall et al.
[134] that the nitrogen in the azadithiolate (ADT) in the bridging ligand would
be an excellent proton acceptor in the catalytic cycle (Figure 3.4.9B). In addition,
the crystal structure provides a proton pathway through the protein leading to the
ADT bridging ligand [103]. If the hydride would bind in the bridging position, the
dithiolate sulfur ligands could act as proton acceptors, but it would also require
the bridging CO ligand to ip over to the terminal position on the distal iron. Thermodynamically, this rearrangement would be feasible, but it is not clear if the
active site provides sufcient exibility to allow such a ligand-exchange process
at high rates (i.e. compatible with the high turnover rates of the enzyme). The catalytic cycle depicted in Figure 3.4.9C shows H2 conversion via a terminal hydride
(at Fed) with the dithiolamine acting as internal base; the transition state is depicted
in Figure 3.4.9B.
209
210
O2 4 H 4 e 2H2O
H
H, e
e Rapid reactivation
H2O
H
Cys-S
Cys-S
Fe||
Ni|||
S
Cys
CN
CN
Cys-S
CO
Fe||
Ni||
Cys-S
CN
CN
OH
Cys-S
Cys-S
CO
Cys Cys
Cys
H
e
H2
[4Fe-3S]-6Cys
CO
Cys
H 2O
+
3e
O2
s Fe
Fe
CN
CN
3H
Cys-S
Cys-S
S
Cys
Cys-S Fe
Proximal cluster
Fe||
Ni|||
Fe
S
S-Cys
S-Cys
Cys-S
Figure 3.4.10 Simplied scheme for O2-tolerant hydrogenases that show both hydrogen oxi-
dation and oxygen reduction activity [101, 140]. The Ni-SIa state (center) reacts with H2 and
forms the Ni-C state (left) that returns (via Ni-R, not shown) to Ni-SIa; thereby 2 H+ and 2e
are released. Upon O2 binding to Ni-SIa, one electron from Ni(II) and three from the FeS clusters (two from the proximal cluster!) are delivered quickly, together with 3 H+, to reduce oxygen to water. OH remains in the active side bridging Ni and Fe in the Ni-B state (right). A key
role is played by the novel proximal [4Fe-3S] cluster. Efcient electron delivery from the other
FeS clusters also assures rapid reactivation of the enzyme from the oxidized state.
The oxygen sensitivity of [FeFe] hydrogenases is even more severe since the H cluster is irreversibly destroyed upon reaction with oxygen [145,146]). The Fe centers in
the subcluster are probably oxidized to Fe(III) and lose their CO ligands. Nevertheless, some [FeFe] hydrogenases seem to show a remarkable oxygen tolerance
[147, 148]. This behavior might be correlated with the properties of the hydrophobic
gas channel. It is assumed that H2 can diffuse more easily through the protein and
reach the active site than O2 and CO. Therefore, much effort is put into the modication of the amino acids in the gas channel, in particular those of the hydrogenase
from C. reinhardtii [149]. But this is certainly not the only effect responsible for the
oxygen tolerance of several hydrogenases; the geometric and electronic structure of
the active site also plays a signicant role.
211
The two groups of hydrogenases have similar active-site structures and mechanisms, which shows that they had a convergent evolution. However, they are totally
unrelated to each other with respect to protein sequence/structure, maturation mechanisms, and distribution ([NiFe] hydrogenases are found in archaea and bacteria;
[FeFe] hydrogenases are found in anaerobic bacteria and in mitochondrial eukaryotes) [82]. The [FeFe] hydrogenases are, in general, most active in H2 production,
while [NiFe] hydrogenases are more tuned to H2 oxidation. Both types are, however,
bidirectional due to a lack of over-potential. [FeFe] hydrogenases are extremely oxygen sensitive and irreversibly inhibited under O2. [NiFe] hydrogenases are, in general,
more oxygen tolerant, and some enzymes (e.g. from Knallgas bacteria) even oxidize
H2 in the presence of O2. Next to the different structures of the active sites of these
two classes of hydrogenases, a difference lies also in the FeS centers of the electron
transport chain. It is remarkable that the dinuclear iron center of the [FeFe] hydrogenase is directly linked to a [4Fe-4S] cluster, which acts as an efcient redox ligand
and affects the electronic structure of the [2Fe]H site. Another difference between the
two types of hydrogenases is the structure of the metal-hydride complex formed, a
key intermediate in the H2 conversion/generation. In [NiFe] hydrogenases, a hydride
bridge between the two metals is found, whereas in [FeFe] hydrogenases, the hydride
is purportedly terminally bound to one of the iron ions. The latter is more reactive
it readily forms H2 with H+ whereas the bridging hydride is much more stable. This
might explain the different catalytic rates of the two types of hydrogenases. It has to
be kept in mind, however, that an intelligent protein matrix has many possibilities to
adjust the thermodynamic and kinetic properties of these intermediates and thus
modify their reactivity.
There are also similarities between the two hydrogenases: Both enzymes employ a
bimetallic active center for their catalysis with a rather short distance, indicating a
metal-metal bond; they have a buttery-shaped core, in which the two metals are
bridged by SR ligands. Only one of the metal ions is redox active (Ni in [NiFe] and
Fed in [FeFe] hydrogenase). This metal has an open coordination site where H2 is believed to bind or be released. In both catalytic sites, the Fe atoms are kept at low
valence states by the strongly donating ligands CN and CO. The H/D isotope exchange shows that in both cases, the H2 splitting is heterolytic. In both active sites,
a sulfur or nitrogen/oxygen ligand probably acts as base to accept or donate the H+.
Both enzymes are O2 sensitive and inhibited by CO (exceptions are known). These
interesting features have served as guidelines for the construction of biomimetic
hydrogenase models, as discussed next.
212
(high TOF), and (3) a low overpotential. However, these criteria are not easy to meet
in model systems [150].
The work on biomimetic models for [NiFe] and [FeFe] hydrogenase has been described in several review articles [150158]. In this work, many of the structural features important for proper function found in the native systems have been
successfully incorporated for example, the bimetallic Ni-Fe or Fe-Fe core with
rather short metal-metal distances and an open coordination site at one metal, the
sulfur-rich environment (terminal and bridging thiolate ligands), CO/CN ligation
of the iron(s), and the incorporation of a base for acceptance of the proton and,
more recently, of hydride bridges. Still-existing problems of many model systems
are the O2 sensitivity, the high overpotentials, and lack of activity (low turnover
rates).
Despite the large number of structural analogues for hydrogenases synthesized
during the last decade, until very recently, no good functional models were known.
For [NiFe] hydrogenases, the situation changed when Rauchfuss and coworkers
[159] described a complex that features a Ni(diphosphine) group linked to a Fe(CO)3
unit by two bridging thiolate ligands [160]. In this complex, a bridging hydride could
be established as the key functional intermediate that reacts with acids and evolves
dihydrogen.
Signicant progress has also been made with [FeFe] hydrogenase models, in which
the issue of Fe valence states, pendant base and hydride chemistry, has been addressed. It was also recognized that the attached [4Fe-4S] cluster in the native enzyme
is important; a rst synthetic model has been described by Pickett et al. [161] A recent
complete system combines all three major functional components of the active site
that couples acid-base chemistry to redox reactions: (1) a reactive di-iron center
with open coordination site, (2) an amine base (azadithiolate) as proton relay, and
(3) a redox-active ligand modeling the [4Fe-4S] center of the native H cluster (see
Figure 3.4.11A). The system shows catalytic activation of hydrogen [162].
Inspiration from the native hydrogenases suggests that catalytic activity might
require a bimetallic site. However, a closer look shows that at least for the
A
NBn
Fe|||
2
Ph
Et2P
OC Fe||
OC
e
Fe|
C
O
H2
Ph2
Ph
P
P Ph
H
Ni
2
Ph
N
Ph
P
P
Ph
Ph
P
Ph2
H
Ph
Figure 3.4.11 Synthetic molecular catalysts for hydrogen conversion. (A) [FeFe] hydrogenase
model with functionalities for substrate binding (H2) and management of redox and proton
equivalents (adapted from [162]). (B) Synthetic mononuclear Ni electrocatalyst with pending
amines that function as proton relays; proposed transition state for heterolytic H2 splitting
and formation (adapted from [165]).
3.4.4 Conclusions
213
3.4.4 Conclusions
During recent years, our understanding of the WOC and the hydrogenase enzymes
has made considerable progress. This is due to intensive efforts in these important
scientic elds of basic research. The availability and combination of structural
214
results from reliable high-resolution X-ray crystallography with structural and functional multifrequency spectroscopy, electrochemistry, and kinetic measurements
has proved to be extremely valuable. Furthermore, the enormous progress in theoretical chemistry now allows us to calculate structure and properties of reaction
intermediates and reaction mechanisms even of larger (biological) systems [32, 173].
Based on studies of natural systems, much has also been learned concerning the
design principles required for biomimetic catalysis of water splitting and hydrogen
evolution. In summary, these include use of abundant and inexpensive metals, the
effective protection of the active sites in functional environments, repair/replacement
of active components in case of damage, and the optimization of reaction rates. Biomimetic chemistry aims to mimic all these features; many labs are working toward
this goal by developing new approaches in the design and synthesis of such systems,
encompassing not only the catalytic center, but also smart matrices and assembly via
self-organization [67, 71, 167].
More stable catalysts that do not require self-repair may be obtained from fully
articial (inorganic) catalytic systems that are totally different from the biological
ones and only apply some basic principles learned from nature. Metals other than
Mn/Ca, Fe, and Ni could be used (e.g. Co) in new ligand spheres and other matrices.
For light harvesting, charge separation/stabilization, and the effective coupling of the
oxidizing/reducing equivalents to the redox catalysts, different methods have been
proposed for example, covalently linked molecular donor-acceptor systems, photovoltaic devices, semiconductor-based systems, and photoactive metal complexes (for
a review, see [174]).
The aim of all these approaches is to develop catalytic systems that split water with
sunlight into hydrogen and oxygen while displaying high efciency and long-term stability. Such a system either biological, biomimetic, or bioinspired has the potential
to be used on a large scale to produce solar fuels (e.g. hydrogen or secondary
products thereof ).
3.4.5 Acknowledgments
M.-E. Pandelia (Pennsylvania State University, USA) and A. W. Rutherford (Imperial College, London, UK) are gratefully acknowledged for their critical review of the
manuscript, and R. Groever and B. Deckers for their extensive help with the text and
artwork. The authors own work cited in the manuscript was nancially supported by
the EU/Energy Network project SOLAR-H2 (FP7 contract 212508) and BMBF
(03SF0355C) and the Max Planck Society.
3.4.6 Notes
1. A new theoretical study [175] of the OEC suggests the S2 state exists in two near isoenergetic
structural forms which differ in terms of the facile movement of the O5 bridge. O5 acts as
either i) a linkage of the outer Mn (MnA4) to the cuboidal unit, where the O5 forms part of
a bis--oxo coordination between MnA4 and MnB3; or ii) as a corner/vertex of the cuboidal
3.4.7 References
215
unit, bridging MnD1, MnC2 and MnB3. This exibility of O5's coordination supports its
assignment to the exchangable -oxo bridge signal as described in the text and its role
as a potential substrate water.
2. Recently the Hsred state has been trapped and characterized by EPR spectroscopy, and
is consistent with a FeIFeI, reduced 4FeS4 (1+) cofactor [176]. This super-reduced state
is proposed as an intermediate of the catalytic cycle.
3. A phylogenetic study suggests this novel [4Fe-3S] cluster is a common oxygen tolerance
adaptation in membrane bound hydrgoenases [177].
3.4.7 References
1. Wydrzynski TJ, Satoh K, editors. Photosystem II. The light-driven water: Plastoquinone
oxidoreductase. As a series in: Govindjee, editor. Advances in photosynthesis and respiration, Volume 22. Dordrecht, The Netherlands: Springer; 2005.
2. Wydrzynski TJ, Hillier W, eds. Molecular solar fuels. Cambridge, UK: RSC Publishing;
2012.
3. Rand DAJ, Dell RM. Hydrogen energy. Challenges and prospects. Cambridge: RSC Publishing; 2008.
4. Stripp ST, Happe T. How algae produce hydrogen-news from the photosynthetic hydrogenase. Dalton Trans. 2009;(45):99609.
5. Cammack R, Frey M, Robson R, editors. Hydrogen as a fuel: learning from nature.
London: Taylor & Francis; 2001.
6. Collings AF, Critchley C, editors. Articial photosynthesis: from basic biology to industrial application. Weinheim: Wiley-VCH Verlag GmbH; 2005.
7. Umena Y, Kawakami K, Shen JR, Kamiya N. Crystal structure of oxygen-evolving
photosystem II at a resolution of 1.9. Nature. 2011;473(7345):5560.
8. Fontecilla-Camps JC, Volbeda A, Cavazza C, Nicolet Y. Structure/function relationships
of [NiFe]- and [FeFe]-hydrogenases. Chem Rev. 2007;107(10):4273303.
9. Fritsch F, Scheerer P, Frielingsdorf S, et al. The crystal structure of an oxygen-tolerant
hydrogenase unvocers a novel iron-sulphur centre. Nature. 2011;479(7372):24952.
10. Shomura Y, Yoon KS, Nishihara H, Higuchi Y. Structural basis for [4Fe-3S] cluster in the
oxygen-tolerant membrane-bound [NiFe] hydrogenase. Nature. 2011;479(7372):2537.
11. Lubitz W, Ogata H, Reijerse E, Higuchi Y. Structure and function of hydrogenase enzymes. In: Wydrzynski TJ, Hillier W, editors. Molecular solar fuels. Cambridge, UK:
The Royal Society of Chemistry; 2012. p. 288325.
12. Messinger J, Noguchi T, Yano J. Photosynthetic O2 evolution. In: Wydrzynski TJ, Hillier W,
editors. Molecular solar fuels. Cambridge, UK: The Royal Society of Chemistry; 2012.
p. 163207.
13. Kok B, Forbush B, McGloin M. Cooperation of charges in photosynthetic O2 evolution.
Photochem Photobiol. 1970;11(6):45776.
14. Kulik LV, Epel B, Lubitz W, Messinger J. Electronic structure of the Mn4OxCa cluster in
the S0 and S2 states of the oxygen-evolving complex of photosystem II based on pulse
55
Mn ENDOR and EPR spectroscopy. J Am Chem Soc. 2007;129(44):1342135.
15. Renger G, editor. Primary processes of photosynthesis, part 2. Principles and apparatus.
Cambridge, UK: RSC Publishing; 2008.
16. Dau H, Haumann M. The manganese complex of photosystem II in its reaction cyclebasic framework and possible realization at the atomic level. Coord Chem Rev.
2008;252(34):27395.
216
17. Kargul J, Barber J. Structure and function of photosynthetic reaction centres. In: Wydrzynski TJ, Hillier W, editors. Molecular solar fuels. Cambridge, UK: The Royal Society of
Chemistry; 2012. p. 10742.
18. Holzwarth AR, Mller MG, Reus M, Nowaczyk M, Sander J, Rgner M. Kinetics and
mechanism of electron transfer in intact photosystem II and in the isolated reaction center:
pheophytin is the primary electron acceptor. Proc Natl Acad Sci U S A. 2006;103(18):
6895900.
19. Rappaport F, Guergova-Kuras M, Nixon PJ, Diner BA, Lavergne J. Kinetics and pathways of charge recombination in photosystem II. Biochemistry. 2002;41(26):851827.
20. Messinger J, Renger G. Photosynthetic water splitting. In: Renger G, editor. Comprehensive series in photochemical and photobiological sciences. Cambridge, UK: The Royal
Society of Chemistry; 2008. Volume 9. p. 291349.
21. Tommos C, Babcock GT. Oxygen production in nature: a light-driven metalloradical
enzyme process. Acc Chem Res. 1998;31(1):1825.
22. Hillier W, Wydrzynski TJ. 18O-water exchange in photosystem II: substrate binding and
intermediates of the water splitting cycle. Coord Chem Rev. 2008;252(34):30617.
23. Styring S, Rutherford AW. In the oxygen-evolving complex of photosystem II the S0 state
is oxidized to the S1 state by D+ (signal IIslow). Biochemistry. 1987;26(9):24015.
24. Renger G, Holzwarth AR. Primary electron transfer. In: Wydrzynski TJ, Satoh K, editors.
Photosystem II. The light-driven water: plastoquinone oxidoreductase. Dordrecht, The
Netherlands: Springer; 2005. p. 13975.
25. Boussac A, Rutherford AW. Nature of the inhibition of the oxygen-evolving enzyme of
photosystem II induced by sodium chloride washing and reversed by the addition of
Ca2+ or Sr2+. Biochemistry. 1988;27(9):347683.
26. Boussac A, Setif P, Rutherford AW. Inhibition of tyrosine Z photooxidation after formation of the S3-state in Ca-depleted and Cl--depleted photosystem-II. Biochemistry. 1992;
31(4):122434.
27. Ishida N, Sugiura M, Rappaport F, Lai TL, Rutherford AW, Boussac A. Biosynthetic
exchange of bromide for chloride and strontium for calcium in the photosystem II
oxygen-evolving enzymes. J Biol Chem. 2008;283(19):1333040.
28. Yano J, Kern J, Sauer K, et al. Where water is oxidized to dioxygen: structure of the
photosynthetic Mn4Ca cluster. Science. 2006;314(5800):8215.
29. Messinger J, Wacker U, Renger G. Unusual low reactivity of the water oxidase in redox
state S3 toward exogenous reductants. Analysis of the NH2OH- and NH2NH2-induced
modications of ash-induced oxygen evolution in isolated spinach thylakoids. Biochemistry. 1991;30(31):785262.
30. Schansker G, Goussias C, Petrouleas V, Rutherford AW. Reduction of the Mn cluster of
the water-oxidizing enzyme by nitric oxide: formation of an S2 state. Biochemistry.
2002;41(9):305764.
31. Kusunoki M. Mono-manganese mechanism of the photosytem II water splitting reaction
by a unique Mn4Ca cluster. Biochim Biophys Acta Bioenerg. 2007;1767(6):48492.
32. Siegbahn PEM. Structures and energetics for O2 formation in photosystem II. Acc Chem
Res. 2009;42(12):187180.
33. Ames W, Pantazis DA, Krewald V, et al. Theoretical evaluation of structural models of
the S2 state in the oxygen evolving complex of photosystem II: protonation states and
magnetic interactions. J Am Chem Soc. 2011;133(49):1974357.
34. Robblee, JH, Messinger, J, Cinco, RM, et al. The Mn cluster in the S0 state of the oxygenevolving complex of photosystem II studied by EXAFS spectroscopy: are there three
Di--oxo-bridged Mn2 moieties in the tetranuclear Mn complex? J Am Chem Soc.
2002;124(25):745971.
3.4.7 References
217
35. Liang WC, Roelofs TA, Cinco RM, et al. Structural change of the Mn cluster during the
S2 S3 state transition of the oxygen-evolving complex of photosystem II. Does it reect
the onset of water/substrate oxidation? Determination by Mn X-ray absorption spectroscopy. J Am Chem Soc. 2000;122(14):3399412.
36. Haumann M., Mller C., Liebisch P, et al. Structural and oxidation state changes of the
photosystem II manganese complex in four transitions of the water oxidation cycle (S0
S1, S1 S2, S2 S3, and S3,4 S0) characterized by X-ray absorption spectroscopy at
20 K and room temperature. Biochemistry. 2005;44(9):1894908.
37. Dismukes GC, Siderer Y. Intermediates of a polynuclear manganese center involved in
photosynthetic oxidation of water. Proc Natl Acad Sci U S A Biophys. 1981;78(1):2748.
38. Haddy A. EPR spectroscopy of the manganese cluster of photosystem II. Photosynth Res.
2007;92(3):35768.
39. Randall DW, Chan MK, Armstrong WH, Britt RD. Pulsed 1H and 55Mn ENDOR studies of dinuclear Mn(III)Mn(IV) model complexes. Mol Phys. 1998;96(6):128394.
40. Kulik LV, Lubitz W. Electron-nuclear double resonance. Photosynth Res. 2009;102(23):
391401.
41. Cox N, Rapatskiy L, Su JH, et al. Effect of Ca2+/Sr2+ substitution on the electronic structure
of the oxygen-evolving complex of photosystem II: a combined multifrequency EPR, 55MnENDOR, and DFT study of the S2 state. J Am Chem Soc. 2011;133(10):363548.
42. Teutloff C, Pudollek S, Keen S, Broser M, Zouni A, Bittl R. Electronic structure of the
tyrosine D radical and the water-splitting complex from pulsed ENDOR spectroscopy on
photosystem II single crystals. Phys Chem Chem Phys. 2009;11(31):671526.
43. Stich TA, Yeagle GJ, Service RF, Debus RJ, Britt RD. Ligation of D1-His332 and
D1-Asp170 to the manganese cluster of photosystem II from Synechocystis assessed by
multifrequency pulse EPR spectroscopy. Biochemistry. 2011;50(34):7390404.
44. Messinger J, Robblee JH, Bergmann U, et al. Absence of Mn-centered oxidation in the
S2 S3 transition: implications for the mechanism of photosynthetic water oxidation.
J Am Chem Soc. 2001;123(32):780420.
45. Glatzel P, Bergmann U, Yano J, et al. The electronic structure of Mn in oxides, coordination complexes, and the oxygen-evolving complex of photosystem II studied by resonant
inelastic X-ray scattering. J Am Chem Soc. 2004;126(32):994659.
46. Rapatskiy L, Cox N, Ames W, et al. Detection of the water binding sites of the oxygenevolving complex of photosystem II using W-band 17O ELDOR-detected NMR spectroscopy. J Am Chem Soc 2011; Epub date: August 31, 2012.
47. Rappaport F, Diner BA. Primary photochemistry and energetics leading to the oxidation
of the Mn4Ca cluster and to the evolution of molecular oxygen in photosystem II. Coord
Chem Rev. 2008;252(34):25972.
48. Dasgupta J, Ananyev GM, Dismukes GC. Photoassembly of the water-oxidizing complex
in photosystem II. Coord Chem Rev. 2008;252(34):34760.
49. Noguchi T. FTIR detection of water reactions in the oxygen-evolving centre of photosystem II. Phil Trans R Soc B. 2008;363(1494):118995.
50. Kawamori A, Inui T, Takaaki O, Yorinao I. ENDOR study on the position of hydrogens
close to the manganese cluster in S2 state of photosystem II. FEBS Lett. 1989;254(12):21924.
51. Aznar CP, Britt RD. Simulations of the 1H electron spin echoelectron nuclear double resonance and 2H electron spin echo envelope modulation spectra of exchangeable hydrogen
nuclei coupled to the S2-state photosystem II manganese cluster. Phil Trans R Soc B.
2002;357(1426):135966.
52. McConnell IL, Grigoryants VM, Scholes CP, et al. EPRENDOR characterization of
(17O, 1H, 2H) water in manganese catalase and its relevance to the oxygen-evolving complex of photosystem II. J Am Chem Soc. 2012;134(3):150412.
218
53. Messinger J. Evaluation of different mechanistic proposals for water oxidation in photosynthesis on the basis of Mn4OxCa structures for the catalytic site and spectroscopic data.
Phys Chem Chem Phys. 2004;6(20):476471.
54. McEvoy JP, Brudvig GW. Water-splitting chemistry of photosystem II. Chem Rev.
2006;106(11):445583.
55. Barber J, Ferreira K, Maghlaoui K, Iwata S. Structural model of the oxygen-evolving centre of photosystem II with mechanistic implications. Phys Chem Chem Phys. 2004;6
(20):473742.
56. Sproviero EM, Gascn JA, McEvoy JP, Brudvig GW, Batista VS. Quantum mechanics/
molecular mechanics study of the catalytic cycle of water splitting in photosystem II. J Am
Chem Soc. 2008;130(11):342842.
57. Yamanaka S, Isobe H, Kanda K, et al. Possible mechanisms for the OO bond formation
in oxygen evolution reaction at the CaMn4O5(H2O)4 cluster of PSII rened to 1.9 X-ray
resolution. Chem Phys Lett. 2011;511(13):13845.
58. Gao Y, kermark T, Liu J, Sun LC, kermark B. Nucleophilic attack of hydroxide on a
Mnv oxo complex: a model of the OO bond formation in the oxygen evolving complex of
photosystem II. J Am Chem Soc. 2009;131(25):87267.
59. Renger G. Oxidative photosynthetic water splitting: energetics, kinetics and mechanism.
Photosynth Res. 2007;92(3):40725.
60. Liu F, Concepcion JJ, Jurss JW, Cardolaccia T, Templeton JL, Meyer TJ. Mechanisms of
water oxidation from the blue dimer to photosystem II. Inorg Chem. 2008;47(6):172752.
61. Romain S, Vigara L, Llobet A. Oxygenoxygen bond formation pathways promoted by
ruthenium complexes. Acc Chem Res. 2009;42(12):194453.
62. McDaniel ND, Coughlin FJ, Tinker LL, Bernhard S. Cyclometalated iridium(III) aquo
complexes: efcient and tunable catalysts for the homogeneous oxidation of water.
J Am Chem Soc. 2008;130(1):2107.
63. Limburg J, Vrettos JS, Liable-Sands LM, Rheingold AL, Crabtree RH, Brudvig GW. A
functional model for OO bond formation by the O2 evolving complex in photosystem II.
Science. 1999;283(5407):15247.
64. Mukhopadhyay S, Mandal SK, Bhaduri S, Armstrong WH. Manganese clusters with relevance to photosystem II. Chem Rev. 2004;104(9):39814026.
65. Beckmann K, Uchtenhagen H, Berggren G, et al. Formation of stoichiometrically
18
O-labelled oxygen from the oxidation of 18O-enriched water mediated by a dinuclear
manganese complex a mass spectrometry and EPR study. Energy Environ Sci. 2008;1(6):
66876.
66. Naruta Y, Sasayama M, Sasaki T. Oxygen evolution by oxidation of water with manganese porphyrin dimers. Angew Chem Int Ed. 1994;33(18):1839841.
67. Kanan MW, Nocera DG. In situ formation of an oxygen-evolving catalyst in neutral
water containing phosphate and Co2+. Science. 2008;321(5892):10725.
68. Jiao F, Frei H. Nanostructured cobalt oxide clusters in mesoporous silica as efcient
oxygen-evolving catalysts. Angew Chem Int Ed. 2009;48(10):18414.
69. Yin Q, Tan JM, Besson C, et al. A fast soluble carbon-free molecular water oxidation
catalyst based on abundant metals. Science. 2010;328(5976):3425.
70. Najafpour MM, Ehrenberg T, Wiechen M, Kurz P. Calcium manganese(III) oxides
(CaMn2O4x H2O) as biomimetic oxygen-evolving catalysts. Angew Chem Int Ed.
2010;49(12):22337.
71. Lutterman DA, Surendranath Y, Nocera DG. A self-healing oxygen-evolving catalyst.
J Am Chem Soc. 2009;131(11):38389.
72. Risch M, Khare V, Zaharieva I, Gerencser L, Chernev P, Dau H. Cobaltoxo core of a
water-oxidizing catalyst lm. J Am Chem Soc. 2009;131(20):69367.
3.4.7 References
219
73. Kanan MW, Yano J, Surendranath Y, Dinca M, Yachandra VK, Nocera DG. Structure
and valency of a cobaltphosphate water oxidation catalyst determined by in situ X-ray
spectroscopy. J Am Chem Soc. 2010;132(39):13692701.
74. Zaharieva I, Najafpour MM, Wiechen M, Haumann M, Kurz P, Dau H. Synthetic manganese-calcium oxides mimic the water-oxidizing complex of photosynthesis functionally
and structurally. Energy Environ Sci. 2011;4(7):24008.
75. Brimblecombe R, Swiegers GF, Dismukes GC, Spiccia L. Sustained water oxidation photocatalysis by a bioinspired manganese cluster. Angew Chem Int Ed. 2008;47(38):73358.
76. Hocking RK, Brimblecombe R, Chang LY, et al. Water-oxidation catalysis by manganese
in a geochemical-like cycle. Nat Chem. 2011;3(6):4616.
77. Brimblecombe R, Koo A, Dismukes GC, Swiegers GF, Spiccia L. Solar driven water oxidation by a bioinspired manganese molecular catalyst. J Am Chem Soc. 2010;132(9):28924.
78. Vignais PM, Billoud B, Meyer J. Classication and phylogeny of hydrogenases. FEMS
Microbiol Rev. 2001;25(4):455501.
79. Cammack R. Hydrogenase sophistication. Nature. 1999;397(6716):2145.
80. Eberly JO, Ely RL. Thermotolerant hydrogenases: biological diversity, properties, and
biotechnological applications. Crit Rev Microbiol. 2008;34(34):11730.
81. Lenz O, Ludwig M, Schubert T, et al. H2 conversion in the presence of O2 as performed
by the membrane-bound [NiFe]-hydrogenase of Ralstonia eutropha. ChemPhysChem.
2010;11(6):110719.
82. Vignais PM, Billoud B. Occurrence, classication, and biological function of hydrogenases: an overview. Chem Rev. 2007;107(10):420672.
83. Shima S, Pilak O, Vogt S, et al. The crystal structure of [Fe]-hydrogenase reveals the
geometry of the active site. Science. 2008;321(5888):5725.
84. Ogata H, Lubitz W, Higuchi Y. [NiFe] Hydrogenases: structural and spectroscopic studies
of the reaction mechanism. Dalton Trans. 2009;(27):757787.
85. Ogata H, Kellers P, Lubitz W. The crystal structure of the [NiFe] hydrogenase from the
photosynthetic bacterium Allochromatium vinosum: characterization of the oxidized
enzyme (Ni-A state). J Mol Biol. 2010;402(2):42844.
86. Garcin E, Vernde X, Hatchikian EC, Volbeda A, Frey M, Fontecilla-Camps JC. The
crystal structure of a reduced [NiFeSe] hydrogenase provides an image of the activated
catalytic center. Structure. 1999;7(5):55766.
87. Marques MC, Coelho R, de Lacey AL, Pereira IAC, Matias PM. The three-dimensional
structure of [NiFeSe] hydrogenase from Desulfovibrio vulgaris Hildenborough: a hydrogenase without a bridging ligand in the active site in its oxidised, as-isolated state.
J Mol Biol. 2010;396(4):893907.
88. Volbeda A, Amara P, Darnault C, Mouesca JM, Parkin A, Roessler MM, Armstrong FA,
Fontecilla-Camps JC. X-ray crystallographic and computational studies of the O2-tolerant
[NiFe]-hydrogenase 1 from Escherichia coli. Proc Natl Acad Sci USA 2012;109(14):530510.
89. Nicolet Y, Piras C, Legrand P, Hatchikian CE, Fontecilla-Camps JC. Desulfovibrio desulfuricans iron hydrogenase: the structure shows unusual coordination to an active site Fe
binuclear center. Struct Fold Des. 1999;7(1):1323.
90. Peters JW, Lanzilotta WN, Lemon BJ, Seefeldt LC. X-ray crystal structure of the Fe-only
hydrogenase (Cpl) from Clostridium pasteurianum to 1.8 ngstrom resolution. Science.
1998;282(5395):18538.
91. Ogata H, Hirota S, Nakahara A, et al. Activation process of [NiFe] hydrogenase elucidated by high-resolution X-ray analyses: conversion of the ready to the unready state.
Structure. 2005;13(11):163542.
92. Page CC, Moser CC, Chen XX, Dutton PL. Natural engineering principles of electron
tunnelling in biological oxidation-reduction. Nature. 1999;402(6757):4752.
220
93. Happe RP, Roseboom W, Pierik AJ, Albracht SPJ, Bagley KA. Biological activation of
hydrogen. Nature. 1997;385(6612):126.
94. Higuchi Y, Yagi T, Yasuoka N. Unusual ligand structure in Ni-Fe active center and an
additional Mg site in hydrogenase revealed by high resolution X-ray structure analysis.
Structure 1997;5(12):167180.
95. Carepo M, Tierney DL, Brondino CD, et al. 17O ENDOR detection of a solvent-derived
Ni-(OHx)-Fe bridge that is lost upon activation of the hydrogenase from Desulfovibrio
gigas. J Am Chem Soc. 2002;124(2):2816.
96. Vincent KA, Belsey NA, Lubitz W, Armstrong FA. Rapid and reversible reactions of
[NiFe]-hydrogenases with sulde. J Am Chem Soc. 2006;128(23):74489.
97. Volbeda A, Fontecilla-Camps JC. The active site and catalytic mechanism of NiFe hydrogenases. Dalton Trans. 2003;(21):40308.
98. Ogata H, Mizogushi Y, Mizuno N, et al. Structural studies of the carbon monoxide complex of [NiFe]hydrogenase from Desulfovibrio vulgaris Miyazaki F: suggestion for the
initial activation site for dihydrogen. J Am Chem Soc. 2002;124(39):1162835.
99. Volbeda A, Martin L, Cavazza C, et al. Structural difference between the ready and
unready oxidized states of [NiFe] hydrogenases. J Biol Inorg Chem. 2005;10(5):23949.
100. Pandelia ME, Nitschke W, Infossi P, Giudici-Orticoni MT, Bill E, Lubitz W. Characterization of a unique [FeS] cluster in the electron transfer chain of the oxygen tolerant [NiFe]
hydrogenase from Aquifex aeolicus. Proc Nat Acad Sci U S A. 2011;108(15):6097102.
101. Goris T, Wait AF, Saggu M, et al. A unique iron-sulfur cluster is crucial for oxygen tolerance of a [NiFe]-hydrogenase. Nat Chem Biol. 2011;7(5):3108.
102. Lukey MJ, Roessler MM, Parkin A, et al. Oxygen-tolerant [NiFe]-hydrogenases: the
individual and collective importance of supernumerary cysteines at the proximal Fe-S
cluster. J Am Chem Soc. 2011;133(42):1688192.
103. Nicolet Y, de Lacey AL, Vernde X, Fernndez VM, Hatchikian EC, Fontecilla-Camps
JC. Crystallographic and FTIR spectroscopic evidence of changes in Fe coordination
upon reduction of the active site of the Fe-only hydrogenase from Desulfovibrio desulfuricans. J Am Chem Soc. 2001;123(8):1596601.
104. Nicolet Y, Cavazza C, Fontecilla-Camps JC. Fe-only hydrogenases: structure, function
and evolution. J Inorg Biochem. 2002;91(1):18.
105. Florin L, Tsokoglou A, Happe T. A novel type of iron hydrogenase in the green alga
Scenedesmus obliquus is linked to the photosynthetic electron transport chain. J Biol
Inorg Chem. 2001;276(9):612532.
106. Lemon BJ, Peters JW. Binding of exogenously added carbon monoxide at the active site
of the iron-only hydrogenase (CpI) from Clostridium pasteurianum. Biochemistry.
1999;38(40):1296973.
107. Silakov A, Wenk B, Reijerse E, Lubitz W. 14N HYSCORE investigation of the H-cluster
of [FeFe] hydrogenase: evidence for a nitrogen in the dithiol bridge. Phys Chem Chem
Phys. 2009;11(31):65929.
108. Pandelia ME, Ogata H, Lubitz W. Intermediates in the catalytic cycle of [NiFe]
hydrogenase: functional spectroscopy of the active site. ChemPhysChem. 2010;11(6):
112740.
109. Lubitz W, Reijerse E, van Gastel M. [NiFe] and [FeFe] hydrogenases studied by advanced magnetic resonance techniques. Chem Rev. 2007;107(10):433165.
110. Happe RP, Roseboom W, Pierik AJ, Albracht SPJ, Bagley KA. Biological activation of
hydrogen. Nature 1997;385(6612):126.
111. Bagley KA, Duin EC, Roseboom W, Albracht SPJ, Woodruff WH. Infrared-detectable
groups sense changes in charge density on the nickel center in hydrogenase from Chromatium vinosum. Biochemistry. 1995;34(16):552735.
3.4.7 References
221
112. de Lacey AL, Fernndez VM, Rousset M, Cammack R. Activation and inactivation of
hydrogenase function and the catalytic cycle: spectroelectrochemical studies. Chem Rev.
2007;107(10):430430.
113. Vincent KA, Parkin A, Armstrong FA. Investigating and exploiting the electrocatalytic
properties of hydrogenases. Chem Rev. 2007;107(10):4366413.
114. Fontecilla-Camps JC, Amara P, Cavazza C, Nicolet Y, Volbeda A. Structure-function relationships of anaerobic gas-processing metalloenzymes. Nature 2009;460(7257):81422.
115. Brecht M, van Gastel M, Buhrke T, Friedrich B, Lubitz W. Direct detection of a hydrogen ligand in the [NiFe] center of the regulatory H2-sensing hydrogenase from Ralstonia
eutropha in its reduced state by HYSCORE and ENDOR spectroscopy. J Am Chem
Soc. 2003;125(43):1307583.
116. Foerster S, van Gastel M, Brecht M, Lubitz W. An orientation-selected ENDOR and
HYSCORE study of the Ni-C active state of Desulfovibrio vulgaris Miyazaki F hydrogenase. J Biol Inorg Chem. 2005;10(1):5162.
117. Pandelia ME, Infossi P, Stein M, Giudici-Orticoni MT, Lubitz W. Spectroscopic characterization of the key catalytic intermediate Ni-C in the O2-tolerant [NiFe] hydrogenase I from
Aquifex aeolicus: evidence of a weakly bound hydride. Chem Comm. 2012;48(6):8235.
118. Medina M, Williams R, Cammack R, Hatchikian EC. Studies of light-induced nickel
EPR signals in Desulfovibrio gigas hydrogenase. J Chem Soc. Faraday Trans. 1994;90(19):
29214.
119. Stein M, Lubitz W. Quantum chemical calculations of [NiFe] hydrogenase. Curr Opin
Chem Biol. 2002;6(2):2439.
120. Stein M, Lubitz W. Relativistic DFT calculation of the reaction cycle intermediates
of [NiFe] hydrogenase: a contribution to understanding the enzymatic mechanism.
J Inorg Biochem. 2004;98(5):86277.
121. Wang H, Patil DS, Gu W, et al. L-edge X-ray absorption spectroscopy of some Ni enzymes: probe of Ni electronic structure. J Electron Spectros Relat Phenomena.
2001;114116:85563.
122. Pardo A, de Lacey AL, Fernndez VM, Fan H. J, Fan Y, Hall MB. Density functional
study of the catalytic cycle of nickel-iron [NiFe] hydrogenases and the involvement of
high-spin nickel(II). J Biol Inorg Chem. 2006;11(3):286306.
123. George SJ, Kurkin S, Thorneley RNF, Albracht SPJ. Reactions of H2, CO, and O2 with
active [NiFe]-hydrogenase from Allochromatium vinosum. A stopped-ow infrared
study. Biochemistry. 2004;43(21):680819.
124. Kurkin S, George SJ, Thorneley RNF, Albracht SPJ. Hydrogen-induced activation of
the [NiFe]-hydrogenase from Allochromatium vinosum as studied by stopped-ow infrared spectroscopy. Biochemistry. 2004;43(21):682031.
125. Foerster S, Stein M, Brecht M, Ogata H, Higuchi Y, Lubitz W. Single crystal EPR
studies of the reduced active site of [NiFe] hydrogenase from Desulovibrio vulgaris
Miyazaki F. J Am Chem Soc. 2003;125(1):8393.
126. Volbeda A, Fontecilla-Camps J C. Catalytic nickel-iron-sulfur clusters: from minerals to
enzymes. Top Organomet Chem. 2006;17:5782.
127. de Lacey AL, Fernndez VM, Rousset M. Native and mutant nickel-iron hydrogenases:
unravelling structure and function. Coord Chem Rev. 2005;249(1516):1596608.
128. Mnck E, Popescu CV. Mbauer studies of exchange coupled cluster assemblies in biological systems. Hyperne Interact. 2000;126(14):5967.
129. Roseboom W, de Lacey AL, Fernndez VM, Hatchikian C, Albracht SPJ. The active site
of the [FeFe]-hydrogenase from Desulfovibrio desulfuricans. II. Redox properties, light
sensitivity and CO-ligand exchange as observed via infrared spectroscopy. J Biol Inorg
Chem. 2006;11(1):10218.
222
130. Silakov A, Kamp C, Reijerse E, Happe T, Lubitz W. Spectroelectrochemical characterization of the active site of the [FeFe] hydrogenase HydA1 from Chlamydomonas reinhardtii. Biochemistry. 2009;48(33):77806.
131. Albracht SPJ, Roseboom W, Hatchikian E. The active site of the [FeFe]-hydrogenase
from Desulfovibrio desulfuricans. I. Light sensitivity and magnetic hyperne interactions
as observed by electron paramagnetic resonance. J Biol Inorg Chem. 2006;11(1):88101.
132. Silakov A, Wenk B, Reijerse E, Albracht SPJ, Lubitz W. Spin distribution of the Hcluster in the Hox-CO state of the [FeFe] hydrogenase from Desulfovibrio desulfuricans:
HYSCORE and ENDOR study of 14N and 13C nuclear interactions. J Biol Inorg Chem.
2009;14(2):30113.
133. Silakov A, Reijerse EJ, Albracht SPJ, Hatchikian EC, Lubitz W. The electronic structure
of the H-cluster in the [FeFe]-hydrogenase from Desulfovibrio desulfuricans: a Q-band
57
Fe ENDOR and HYSCORE study. J Am Chem Soc. 2007;129(37):1144758.
134. Fan HJ, Hall MB. A capable bridging ligand for Fe-only hydrogenase: density functional
calculations of a low-energy route for heterolytic cleavage and formation of dihydrogen.
J Am Chem Soc. 2001;123(16):38289.
135. Cao ZX, Hall MB. Modeling the active sites in metalloenzymes. 3. Density functional
calculations on models for [Fe]-hydrogenase: structures and vibrational frequencies
of the observed redox forms and the reaction mechanism at the diiron active center.
J Am Chem Soc. 2001;123(16):373442.
136. Liu ZP, Hu P. A density functional theory study on the active center of Fe-only hydrogenase: characterization and electronic structure of the redox states. J Am Chem Soc.
2002;124(18):517582.
137. Bruschi M, Zampella G, Fantucci P, De Gioia L. DFT investigations of models related
to the active site of [NiFe] and [Fe] hydrogenases. Coord Chem Rev. 2005;249(1516):
162040.
138. Zhou TJ, Mo YR, Zhou ZH, Tsal K. Density functional study on dihydrogen activation
at the H cluster in Fe-only hydrogenases. Inorg Chem. 2005;44(14):49416.
139. Esper B, Badura A, Rgner M. Photosynthesis as a power supply for (bio-)hydrogen production. Trends Plant Sc. 2006;11(11):5439.
140. Armstrong FA, Belsey NA, Cracknell JA, et al. Dynamic electrochemical investigations
of hydrogen oxidation and production by enzymes and implications for future technology. Chem Soc Rev. 2009;38(1):3651.
141. Buhrke T, Lenz O, Krauss N, Friedrich B. Oxygen tolerance of the H2-sensing [NiFe]
hydrogenase from Ralstonia eutropha H16 is based on limited access of oxygen to the
active site. J Biol Chem. 2005;280(25):237916.
142. van der Linden E, Burgdorf T, de Lacey AL, et al. An improved purication procedure
for the soluble [NiFe]-hydrogenase of Ralstonia eutropha: new insights into its (in)stability and spectroscopic properties. J Biol Inorg Chem. 2006;11(2):24760.
143. Vincent KA, Cracknell JA, Lenz O, Zebger I, Friedrich B, Armstrong FA. Electrocatalytic hydrogen oxidation by an enzyme at high carbon monoxide or oxygen levels. Proc
Nat Acad Sci U S A. 2005;102(47):169514.
144. Pandelia ME, Fourmond V, Tron-Infossi P, et al. Membrane-bound hydrogenase I from
the hyperthermophilic bacterium Aquifex aeolicus: enzyme activation, redox intermediates and oxygen tolerance. J Am Chem Soc. 2010;132(20):69917004.
145. Stiebritz MT, Reiher M. Hydrogenases and oxygen. Chem Sci 2012;3(6):173951.
146. Stripp ST, Goldet G, Brandmayr C, et al. How oxygen attacks [FeFe] hydrogenases from
photosynthetic organisms. Proc Natl Acad Sci U S A. 2009;106(41):173316.
147. Baffert C, Demuez M, Cournac L, et al. Hydrogen-activating enzymes: activity does not
correlate with oxygen sensitivity. Angew Chem Int Ed. 2008;47(11):20524.
3.4.7 References
223
224
168. Zhang W, Hong J, Zheng J, Huang Z, Zhou J., Xu R. Nickelthiolate complex catalyst
assembled in one step in water for solar H2 production. J Am Chem Soc. 2011; 133
(51):206803.
169. Kubas GJ. Fundamentals of H2 binding and reactivity on transition metals underlying
hydrogenase function and H2 production and storage. Chem Rev. 2007;107(10):4152
205.
170. Sala X, Escriche L, Llobet A. Molecular Ru and Ir complexes capable of acting as
water oxidation catalysts. In: Wydrzynski TJ, Hillier W, editors. Molecular solar fuels.
Cambridge, UK: The Royal Society of Chemistry; 2012. p. 27387.
171. Matsumoto T, Kim K, Ogo S. Molecular catalysis in a fuel cell. Angew Chem Int Ed.
2011;50(47):112025.
172. Dempsey JL, Brunschwig BS, Winkler JR, Gray HB Hydrogen evolution catalyzed by
cobaloxines. Acc Chem Res. 2009;42(12):19952004.
173. Neese F. Prediction of molecular properties and molecular spectroscopy with density
functional theory: from fundamental theory to exchange-coupling. Coord Chem Rev.
2009;253(56):52663.
174. Navarro RM, Alvarez-Galvan MC, Villora de la Mano JA, Al-Zahrani SM, Fierro JL.
A framework for visible-light water splitting. Energy Environ Sci. 2010;3(12):186582.
175. Pantazis DA, Ames W, Cox N, Lubitz W, Neese F. Two interconvertible structures that
explain the spectroscopic properties of the oxygen evolving complex of photosystem II in
the S2 state. Angew Chem Int Ed 2012;51(39):993540.
176. Adamska A, Silakov A, Lambertz C, Rdiger O, Happe T, Reijerse EJ, Lubitz W. Identication and characterization of the super-reduced state of the H-cluster in [FeFe]
hydrogenase: a new building block for the catalytic cycle? Angew Chem Int Ed 2012;
in press.
177. Pandelia ME, Lubitz W, Nitschke W. Evolution and diversication of Group 1 [NiFe]
hydrogenases. Is there a phylogenetic marker for O2-tolerance? Biochim Biophys Acta
2012;1817(9):156575.
3.5.1 Introduction
Imagine for a moment our world without electricity, and the signicance of electrical
energy in daily life will be quickly understood. This not only applies to energy supply
or conversion as such, but also indirectly to powering modern communication technologies and thereby the basis of the nervous system of our society. As far as its
key role for information, communication, and mobility is concerned, electricity
is comparable with the role that bioelectrochemistry plays for our individual
activity [1].
In the context of chemical energy storage, the transformation of electrical energy
into chemical energy must be a matter of highest priority, the more so since electricity
is hard to store and conversion into chemical energy represents a highly effective
storage mode. This is particularly attractive if, within the same device, this can be
reversed. Exactly this is possible in a rechargeable battery (also termed secondary
battery or accumulator) [2].
Looking at the basic mechanism of energy storage in batteries, a simple comparison to a pumped-storage hydroelectric power plant can be made (see Figure 3.5.1).
In this kind of power plant, water is stored at two different levels of gravitational
energy. In the discharged state, all water is stored in a downhill basin. Upon charging, water is pumped to an uphill reservoir. In terms of energy, electrical energy
drives a pump motor that produces mechanical energy, which ultimately is converted
to potential energy. The discharge process works vice versa: water from the uphill
basin drives a turbine connected to a generator producing electrical energy.
At a rst glance, energy storage in batteries works in a very similar way. Just as the
hydropower plant offers two levels of gravitational potential for storage of water, a
battery offers two levels of chemical potential for an electrochemically active species,
such as Li in a lithium battery. The chemical potential can be understood as a measure of how much a certain increase of species is disliked in a given system and therefore states how much (free) energy is released if the species is removed from that
system (see Section 3.5.4 for more details).
The uphill reservoir, which is the state of high (gravitational) potential, corresponds to an electrochemical reservoir (= electrode), where an electrochemically
active species is stored with high chemical potential (see Figure 3.5.10). This electrode is called negative electrode in a battery context, and during discharge, the
anodic reaction takes place here.1
The corresponding downhill reservoir corresponds to an electrode where the electrochemically active species can be stored with a low respective chemical potential.
226
Uphill
reservoir
Pump/
turbine
Downhill
reservoir
Negative
electrode
Negative
electrode
e
tra
ns
po
Positive
electrode
Li +
rt
tr a
nsp
ort
Positive
electrode
Figure 3.5.2
This is the positive electrode of a battery, with the cathodic reaction taking place
during discharge [1].
In a hydroelectric power plant, the uphill and downhill reservoirs are connected via
pipes and valves. Such a seemingly simple, direct connection does not apply for an
electrochemical energy storage device. If, in a charged battery, anode and cathode
electrode came in direct contact, Li as the electrochemically active species would
immediately ow from the negative to the positive electrode. The difference in chemical potential would be released in the form of heat. This heat could be used again to
produce electricity, but only in a Carnot-limited process. The same would be true if
the connecting phase (yellow in Figure 3.5.2A) was a material that is conductive for
both Li+ and e.
The secret of a battery is now to restrict the transport between positive and negative electrodes in a way that only the ionic species, but not electrons, can pass. For
this, an electrolyte is put in place between the electrodes.
The electrons are then forced to ow through an external wire, where they directly
can drive an external load (Figure 3.5.2B). In this way, the whole free-energy content
of the electrochemical reaction theoretically can be converted to electrical energy.
Besides the efciency, the enormous difference in energy density (per mass or volume) deserves to be emphasized. Assuming a height difference of 100 m, the gravimetric energy density is 981 N m kg1, or 0.27 W h kg1 in units used in battery
3.5.1 Introduction
227
applications.
Device
Conversion mode
Typical application
Electrolyzers
Synthesis, storage
Primary
batteries
Fuel cells
Power plants
Heat-power devices
(efcient fuel conversion)
Primary metalair
Fleet applications
Electrochemical
capacitors
Secondary
batteries
228
2 Li
Li+
Fe(OH)2 + 2
Li+
+2
2 Li+ + 2 e
2 Zn(OH)42 + 4 e
Zn2+ + 2 e
2 Na+ + 2 e
FePO4 +
Li+
+xS+2
2 Li+ + 2 e + O2 (g)
O2 + 2 H2O + 4 e
2 Ni(OH)2 + 2
LiSx
Li2O2
4 OH
VO2+ + 3 H2O
OH
2 Ni(OH)2 + 2 OH
2 Ni(OH)2 + 2
OH
PbSO4 + 6 H2O
2 NaCl + Ni
NaSx
2 Br
VO2+ + 2 H3O+ + e
Br2 (liq) + 2 e
2 Na+ + NiCl2 + 2 e
2 Na+ + x S + 2 e
2 NiO(OH) + 2 H2O + 2
2 NiO(OH) + 2 H2O + 2 e
LiCoO2
LiFePO4
2 NiO(OH) + 2 H2O + 2
Li+
(*): The wide variety of Li-Ion battery chemistries is discussed in the following section "Lithium batteries"
M = (La,Ce,Nd,Pr)(Ni,Co,Mn,Al)5 alloy
2 Li
Li+
LiPF6 in EMS
Li-Air
Li-sulphur
2 Zn + 8 OH
OH
KOH (aq)
Zn-Air
V2+
H+
Nafion
Vanadium redox
flow
Zn
Zn2+, Br
ZnBr (aq)
Zn-bromine flow
battery
2 Na
Na+
Na--alumina,
molten NaAlCl4
Na/NiCI2 (ZEBRA)
2 Na
Na+
2 M + 2 H2O + 2
Cd(OH)2 + 2 e
2 Na+ + 2 e
Fe + 2
Na--alumina
OH
OH
Na-sulphur
Cd + 2 OH
OH
KOH (aq)
KOH (aq)
2 MH + 2
Li+
PbSO4 + 2 e
OH
6C +
OH
LiC6
Pb + SO42
Ni-Fe
KOH (aq)
NiMH
Li+
SO42
Mobile
species in
electrolyte
NiCd
H2SO4 (aq)
LiPF6 in EC/DMC
Lead acid
Li-Ion (*)
Electrolyte
Battery technology
2.0
2.7
1.6
1.2-1.6
1.8
2.6
2.1
1.2
1.2
1.2
4.2
3.6
2.1
Standard cell
potential / V
350
750-2000
450
25-35
75-85
90
150
50
40-60
30-80
150
150-250
30-40
Gravimetric
energy density/
Wh kg1
"Chemical" rechargeability
by replacing the Zn anode.
Operating temperature:
~250 C
Operating temperature:
~300 C
Notes
Table 3.5.2 Compilation of the most important secondary battery technologies. Energy densities compiled from various manufacturers' data sheets and
references [4,7-9] (EC/DMC, ethylene carbonate/dimethyl carbonate; EMS (ethyl methyl sulphone)).
230
1000
IC Engine
6
4
Fuel Cells
100 h
EV goal
2
Li-ion
PHEV goal
100
6
4
2
Ni-MH
Lead-Acid
HEV goal
10 h
10
Capacitors
6
4
2
1
100
1h
0.1 h
101
36 s
102
Specific Power (W/kg)
3.6 s
103
104
Figure 3.5.3
Ragone plot for different energy storage solutions. The stars show specic
energy and power goals for different vehicle technologies (EV: electric vehicle; PHEV: plugin hybrid EV; HEV: hybrid EV); compared to internal combustion (IC) engine. (V. Srinivasan,
Berkeley Electrochemical Research Council). Used with authors permission.
A
Terminal
231
Electrical insulation
Alpha-alumina seal
Sodium chamber
Na/NiCl2/NaCl/NaAlCl4
positive electrode
Na--alumina separator
Metal insert
Sodium electrode
Solid electrolyte
Case
Sulfur electrode
Cell container
Figure 3.5.4 (A) Sodium sulfur cell. Source: NASA John Glenn Research Center, public
domain. (B) Na/NiCl2 cell. Source: [10].
V3
V5
Positive
electrode
storage
tank
V4
Diffusion
layers
V2
Negative
electrode
storage
tank
Proton conducting
membrane
232
These types of batteries fulll the requirements for reversible storage clusters, which
are needed for buffering uctuations of solar or wind power plants. Here, cost, safety
and energy density, rather than power density, are the primary considerations.
(1)
As the storage on the anode side is similar to the storage on the cathode side if
TiS2 is used, the lithium is rocked backward and forward between two intercalation
electrodes. This gave rise to the name rocking chair battery. Nowadays, such
batteries are termed Li-ion battery, even though it is always Li = Li+ + e that is
stored.
The replacement of pure metallic Li by Li metal encapsulated in carbon as the
anode led to a technologically important difference. While originally metallic lithium
was the lithium source in the cell, it proved advantageous to use the positive electrode as the Li source. In other words, lithiated compounds were used as positive
electrodes, and the cell was manufactured in the discharged state.
The rst large-scale commercial lithium battery was built in the 1980s by Sony.
It employed graphite as the negative electrode and LiCoO2 as the positive electrode
(see Figure 3.5.6). This is still the prevailing battery concept, especially in consumer
electronics.
233
4.2
200A/cm2
4.1
4.0
3.9
3.8
3.7
3.6
0.5
0.6
0.7
0.8
0.9
1.0
Figure 3.5.6 Discharge curve of LixCoO2 positive electrode material, with Li metal as neg-
ative electrode. Reprinted from [12] with permission, copyright 1980 Elsevier.
4
3
2
LiCoPO4
LiMn2O4
LiCoO2
Li2SO4
Water
Li(NiCoMn)O2
LiFePO4
Li / O
Li / S
TiO2
Soggy
LiPF6
EC/
Sand
LiBF4 DMC
LiPON
IL
Positive electrode
Electrolyte
Sn
Carbon
Si
Li metal
Negative electrode
Figure 3.5.7 Selected cathode and anode materials and electrolytes used in todays lithium
batteries (darker colors) and possibly useful future technologies (lighter colors). For electrode
materials, the width of the boxes displays demonstrated capacities (note the different scales for
positive (cathode) and negative (anode) electrode materials). The height of the boxes corresponds to the half-cell potential and its variation for different states of charge. References:
[1220].
Numerous other electrode materials for both negative and positive electrodes have
been developed since then. A compilation of the most signicant developments is
shown in Figure 3.5.7.
Among the most signicant novel developments for positive electrode materials
are the layered oxide LiNi1/3Co1/3Mn1/3O2 (LiNCM) [20, 21] and the olivinestructured LiFePO4 [14]. A main reason for moving away from LiCoO2 (besides
environmental concerns) is safety aspects. LiCoO2 tends to release O2 gas under
overcharge conditions and even worse under high thermal load, possibly leading
to exothermic reactions with the electrolyte inside the cell that increase the temperature even further (the so-called thermal runaway, see Figure 3.5.8). LiFePO4 does
234
400
LiCoO2
350
Gen2: LiNi0.8Co0.15Al0.05O2
300
250
Gen3: Li1.1(Ni1/3Co1/3Mn1/3)0.9O2
200
150
LiMn2O4
100
LiFePO4
50
0
0
100
200
300
400
Temperature (C)
Figure 3.5.8 Thermal runaway behavior of the most common positive electrode materials
upon heating at 5 K min1. Source: E. Peter Roth, Power Sources Technology Group, Sandia
National Laboratories [22]. Used with authors permission.
not exhibit such behavior, and for LiNCM, the onset temperature of thermal runaway
is increased by almost 70 K compared to LiCoO2.
As shown in Figure 3.5.7, the capacities for most cathode materials do not differ
drastically. From the 1980s when LiCoO2 (125 mA h g1) was produced to recently,
with the development of LiNCM (180 mA h g1), capacities were increased by less
than 50%. From chemical and structural considerations, there is a clear reason not
to expect drastic increases in electrode capacities within these classes of materials
(and storage mechanisms). For example, the lithium insertion and extraction reaction
in LiFePO4 reads
LiFePO4 [ FePO4 + Li+ + e
(2)
235
236
Separator
membrane
Negative
electrode
Positive
electrode
Copper
Aluminium
Graphite
Liquid electrolyte
Lithium
iron phosphate
electrodes are prevented from touching by a polymer separator, which is soaked with
the liquid electrolyte. As mentioned in the introduction, a direct contact of positive
and negative electrodes would lead to an uncontrolled release of the chemical energy
stored in the battery in the form of heat. Preventing direct contact between positive
and negative electrodes is therefore extremely important to cell safety, and multiple
optimizations such as using polymer/ceramic composite materials as separators have
evolved [40]. Alternatively, using a shapable electrolyte such as the abovementioned
soggy sand electrolytes, or polymer electrolytes [41], could make the separator
dispensable.
In the next section, we will also address nonconventional storage modes (conversion reaction, interfacial storage), which can provide signicant energy densities. Yet
it remains to be seen whether or not they can be made reversible enough as to
compete with single- and two-phase storage.
237
e
Li
Li
Li(rhs)
e(rhs)
Li(lhs)
(lhs)
e
Positive electrode
Negative electrode
Electrolyte
Li, e during
discharge
Li, e during
charge
EF = ( Li[lhs] Li[rhs])
(3)
This intelligible equation is obtained rigorously when splitting the chemical potential
of lithium into the contributions of the lithium ions and the electrons (Li = ~Li+ + ~e);
the difference of the rst vanishes due to the presence of the electrolyte, and the
difference of the second determines the voltage measured.
The bracketed difference is the Gibbs energy of the reaction that would occur
on direct contact. As the Gibbs energy differs from the reaction enthalpy by TS
(T: thermodynamic temperature, S: reaction entropy), a positive S would result
in theoretical efciencies greater than 100% (then the environment would cool).
Usually, those efciencies are approximately 100%, as S is negligible. This high
theoretical efciency is another asset of electrochemical conversion devices.
The second quantity besides cell voltage that is important here is the equilibrium
storage capacity. This quantity follows from the defect chemistry and phase
thermodynamics.
Figure 3.5.11 shows the storage modes of relevance. The already mentioned intercalation mode belongs to the class of single-phase mechanism. This means homogeneous storage by a phase-absorbing Li (example LixCoO2). This is both a redox
reaction (accumulation of e by introducing excess electrons and/or annihilating electron holes; in praxi it almost always coincides with reduction of transitions metal elements) as well as an acid-base reaction (in the sense of [42]; accommodation of Li+ by
occupying interstitial sites in the lattice and/or annihilating vacancies).
Let us for simplicity assume the storage to solely occur through
Li [ Li+ (interstitial site) + e (excess electron)
(4)
If the two defects on the right-hand side are dilute, an ideal mass action law can be
formulated. The defect concentrations can be connected with the nonstoichiometry
( = i ; i: interstitial Lithium concentration; : Lithium vacancy concentration),
which is proportional to the storage capacity, while the lithium activity (see the
238
MOn (Li)
MOn LiMOn
Single phase
Phase change
MOn
Li2O
Decomposition
Li
e
Li2O
Interfacial
Figure 3.5.11 Relevant storage modes in lithium batteries. Reprinted from [73] with
permission.
left side of Equation [4]) relates to the cell voltage. The maximum capacity (maximum nonstoichiometry) of the single-phase mechanism is reached when the phase
transforms into a Li-richer phase (or if not existent, when Li segregates as metal
layer). These concentrations of coexistence derive from the phase thermodynamics
that eventually is also based on defect chemistry.
The end of solubility is not the end of storage as the transformation into neighboring phases can also be used as a storage mode. A prominent example is FePO4/
LiFePO4. If only Li is mobile, the voltage stays constant during the transformation,
as Li is invariant. It proves simpler in such cases to connect E with the phase reaction (Li(s) + FePO4 LiFePO4). This is, however, identical with Equation (3)
as the chemical potential of Li in the electrode can be written as the difference of
the chemical potentials of the two phases. During the phase reaction, the individual
nonstoichiometries of FePO4 (LiFePO4) and LiFePO4 (Li1FePO4) do not
change; rather it is the phase fraction that varies and determines the storage capacity. As the local nonstoichiometries dene the transport coefcients, defect chemistry is also relevant for this storage mode, a point that is taken up in the
Section 3.5.5.
If the Li-poor component has been fully transformed, then another single-phase
mechanism sets in (dissolution of Li in the Li-richer phase) before a further phase
reaction occurs. The abovementioned phase reaction is very simple as only FePO4/
LiFePO4 interfaces are generated. The same occurs for RuO2/LiRuO2; further Li
incorporation in the last case, however, leads to a true decomposition, namely, reduction to Ru and Li2O. Here three phases are in equilibrium and a complex morphology evolves [4346]. This conversion reaction to be used in a rechargeable mode is
kinetically extremely demanding. Fortunately, these conversion reactions automatically lead to extreme nanostructuring. Nonetheless, RuO2, Fe2O3, and FeF3 are the
only examples where a reasonable reversibility could be achieved. Performing further
research is highly relevant as the storage capacity of this mechanism is huge. Having
239
In other words, under realistic conditions (I = 0), entropy is produced, with the positive entropy production being given by uxes and forces related to the process j. Equation (5) assumes that all these processes are in series, which is mostly correct. The most
obvious contributions are transport resistances, due to the nite conductivities of Li+
and e in electrolyte and electrodes. For usual geometries, these resistances are constant to a good approximation, while for resistances stemming from impeded charge
transfer and phase boundaries, the dependence on current can be severe.
Diffusion resistances can occur for Li in the electrode, but also for the salt in the
electrolyte (if anion conductivity in the electrolyte is signicant). Further effects are
due to depletion of carriers at a phase boundary. In such cases, time dependencies of
the electrical properties occur (in addition to Rs, effective capacitances Cs also
appear). The same is true for impeded nucleation processes. Since any potential
step of the electrochemical potential can be connected with current-dependent effective resistances and capacitances, the kinetic description is typically very specic and
complex. As the storage processes in Li-based batteries are solid-state processes, the
240
A'h
A',h
D,V 'Li
VxLi
VxLi
V 'Li,h
V 'Li,h
V 'Li
A'
V 'Li
241
DV 'Li
LiFePO4
A',h
D,V 'Li
A'h DV 'Li
log [A]
log [D]
A'Lii De'
e'
A'
Lii,e'
A',Lii
A'Lii
A',Lii
Lixi
Lii
D
Lii,e'
Lixi
FePO4
D,e'
D,e'
De'
Figure 3.5.12 The variation of charge carrier (defect) concentrations in LiFePO4 and
FePO4 by acceptor [A] or donor [D] doping shows a wide range of modied properties for
the materials. This is to illustrate the sheer complexity of the defect equilibria in battery
materials. For details, the reader is referred to the references [53, 54].
allowing for heterogeneously doping materials [56]. This of course works best in nanomaterials, where the interfacial density is high [51, 52]. Again, the effect of two oppositely charged carriers is opposite. Yet the potential of heterogeneous doping has
been revealed in the generation of soggy sand electrolytes [19, 57]. In these semisolid-liquid materials, anions are adsorbed at the surface of oxide particles admixed
to the salt-containing organic liquids. Hence, Li+ conductivity is increased. The
decreased counter-ion conductivity is very favorable here.
In most cases, morphological optimizations are done not in order to vary materials
properties, but rather to vary the electrodes network topology. The effects achieved
are quite often really drastic. As mentioned, solid-state chemical diffusion of lithium
is a critical step for almost all solid battery electrodes. Hence, shortened transport
length for solid-state diffusion is a key to better electrode kinetics and is particularly
important for high-power applications. Note that downsizing electrodes from 1 mm
to 10 nm reduces the diffusion time by 10 orders of magnitude (Figure 3.5.13A),
but only if ions and electrons are brought quickly enough to the squillions of particles. This requires a very efcient ionic/electronic wiring. So nanostructuring
combined with morphological preparation will be a key strategy for improving
power density.
242
~ Li / cm2 s1
D
t / s for 1 mm particles
t / s for 10 nm particles
108
5 105: ~1 week
5 105
5 10 : ~1 hour
5 107
50: ~1 minute
10
20
1015
10
10
10
104
17
Electroactive material
Electronic conductor
Pore filled with electrolyte
Carbon black
electrolyte
current collector
Li electrolyte
ecurrent collector
Figure 3.5.13 (A) Equilibrium times for diffusion on macroscopic (1 mm) and nanoscopic
(10 nm) length scales. (B) Illustration of ionic and electronic wiring, with hierarchical porosity
as Li+ distribution network and a carbon second-phase e distribution network. Reprinted
from [58] with permission, copyright 2007 John Wiley & Sons.
243
2 m
50 nm
1 m
Figure 3.5.14 Carbon monolith with hierarchical porosity. Reprinted from [69] with permis-
244
1200
1 m
1000
800
600
400
200
0
0
20
40
60
80
100 120
Cycle number
140
160
100
200
Figure 3.5.15 Tin nanoparticles in an electrospun carbon ber for use as a negative electrode
(lhs reprinted with permission from [34], copyright 2009 John Wiley & Sons; rhs reprinted with
permission from [33], copyright 2009 American Chemical Society).
transport occurs perpendicular to the surface; only a short distance in the Li-permeable carbon is to be traversed, leading to an almost perfect electrochemical
coupling. (4) Tin particles are small enough to ensure quick Li transport therein.
(5) The carbon ber acts as a binder and guarantees morphological stability.
(6) Carbon provides a stable passivation layer in contact with the electrolyte.
(7) The growth of the tin nanoparticles is suspended; Carbon is a good Li+ and
e, but certainly not a good Sn4+ conductor. Therefore, no electrochemical Ostwald ripening can take place that otherwise would cause the elemental particles
to grow [72].
3.5.7 Outlook
The eld of rechargeable batteries is even though having experienced an enormous
push not at all a new technology. Owing to limits set by the periodic table, energy
densities cannot be increased by orders of magnitude, unless one gives up the enormous reversibility achievable in modern batteries. Great improvements are, however,
possible as regards kinetics and hence power density. The progress in this exciting
eld is to be achieved by the highly interdisciplinary research embracing chemistry,
physics, and materials science, whereby solid-state electrochemistry will play the
key role.
3.5.8 Acknowledgments
The authors would like to thank Dr Venkat Srinivasan (Lawrence Berkeley National
Laboratory), Dr E. Peter Roth (Sandia National Laboratory), Dr Roland Wengenmayr and Changbao Zhu (MPI-FKF) for providing gures. Thanks also to
Dr Jelena Popovic for a critical reading of the manuscript.
3.5.10 References
245
3.5.9 Note
1. Please note that the negative electrode strictly speaking acts as the anode only upon
discharge, upon charge it acts as cathode. Still, in almost all of the battery literature, the
terms anode and negative electrode are used synonymously, independent of the direction of current. Similarly, cathode and positive electrode are used synonymously as well.
3.5.10 References
1. Gruss P, Schth F, editors. Die Zukunft der Energie. Mnchen: C.H. Beck; 2008.
2. Hamann CH, Hamnett A, Vielstich W. Electrochemistry. Weinheim: Wiley-VCH; 1998.
3. Mallela VS, Ilankumaran V, Rao NS. Trends in cardiac pacemaker batteries. Ind Pacing
Electrophysiol J. 2004;4(4):201212.
4. Winter M, Besenhard JO. Wiederauadbare Batterien. Teil I: Akkumulatoren mit
wriger Elektrolytlsung. Chemie in unserer Zeit. 1999;33(5):25266.
5. Ritter JW. Beitrge zur nheren Kenntni des Galvanismus und der Resultate seiner
Untersuchung. Jena: Frommann; 18001805.
6. Plant G. Storage of electrical energy. Birmingham: Parker & Hill; 1859.
7. Winter M, Besenhard JO. Wiederauadbare Batterien. Teil II: Akkumulatoren mit nichtwriger Elektrolytlsung. Chemie in unserer Zeit. 1999;33(6):32032.
8. Linden D, Reddy TB, editors. Handbook of batteries. 3rd ed. New York: McGraw-Hill;
2002.
9. Hamann CH, Vielstich W. Elektrochemie. 3rd ed. Weinheim: Wiley-VCH; 1998.
10. Dvorak W, Sodium-nickel-chloride cell. In Wikimedia Commons CC-BY, 2011. Available
from: http://commons.wikimedia.org/wiki/File:Sodium-nickel-chloride_cell.svg.
11. Whittingham MS. Lithium batteries and cathode materials. Chem Rev. 2004;104
(10):4271301.
12. Mizushima K, Jones PC, Wiseman PJ, Goodenough JB. LixCoO2 (0<x1): a new cathode
material for batteries of high energy density. Mater Res Bull. 1980;15(6):7839.
13. Tarascon JM, Armand M. Issues and challenges facing rechargeable lithium batteries.
Nature. 2001;414(6861):35967.
14. Padhi AK, Nanjundaswamy KS, Goodenough JB. Phospho-olivines as positive-electrode
material for rechargable lithium batteries. J Electrochem Soc. 1997;144(4):118894.
15. Wessels C, Ruff R, Huggins RA, Cui Y. Investigations of the electrochemical stability of
aqueous electrolytes for lithium battery applications. Electrochem Solid-State Lett.
2010;13:A5961.
16. Knauth P. Inorganic solid Li ion conductors: an overview. Solid State Ionics. 2009;180
(1416):9116.
17. West WC, Whitacre JF, Lim JR. Chemical stability enhancement of lithium conducting
solid electrolyte plates using sputtered LiPON thin lms. J Power Sources. 2004;126(12):
1348.
18. Lee JM, Kim SH, Tak Y, Yoon YS. Study on the LLT solid electrolyte thin lm with LiPON
interlayer intervening between LLT and electrodes. J Power Sources. 2006;163(1): 1739.
19. Bhattacharyya AJ, Maier J. Second phase effects on the conductivity of non-aqueous salt
solutions: soggy sand electrolytes. Adv Mater. 2004;16:8114.
20. Thackeray MM, Kang S-H, Johnson CS, Vaughey JT, Benedek R, Hackney SA. Li2
MnO3-stabilized LiMO2 (M = Mn, Ni, Co) electrodes for high energy lithium-ion batteries. J Mater Chem. 2007;17:311225.
246
21. Kang K, Meng YS, Brger J, Grey CP, Ceder G. Electrodes with high power and high
capacity for rechargeable lithium batteries. Science. 2006;311(5763):977980.
22. Roth EP. Comparative abuse response of Li-ion cells with LiFePO4 and LiMn2O4 cathodes. Presented at: Fall 2008 ECS meeting; 2008; Honululu, HI. 2008.
23. Dominko R, Bele M, Kokalj A, Gaberscek M, Jamnik J. Li2MnSiO4 as a potential
Li-battery cathode material. J Power Sources. 2007;174(2):45761.
24. Nyten A, Kamali S, Haggstrom L, Gustafsson T, Thomas JO. The lithium extraction/
insertion mechanism in Li2FeSiO4. J Mater Chem. 2006;16(23):226672.
25. Armand M, Endres F, MacFarlane DR, Ohno H, Scrosati B. Ionic-liquid materials for the
electrochemical challenges of the future. Nat Mater. 2009;8(8):6219.
26. Huang SY, Kavan L, Exnar I, Gratzel M. Rocking chair lithium battery based on nanocrystalline TiO2 (anatase). J Electrochem Soc. 1995;142(9):L1424.
27. Kavan L, Prochzka J, Spitler TM, et al. Li insertion into Li4Ti5O12 (xpinel). J Electrochem Soc. 2003;150(7):A10007.
28. Hu Y-S, Kienle L, Guo Y-G, Maier J. High lithium electroactivity of nanometer-sized
rutile TiO2. Adv Mater 2006;18:14216.
29. Armstrong AR, Armstrong G, Canales J, Carca R, Bruce PG. Lithium-ion intercalation
into TiO2-B nanowires. Adv Mater. 2005;17:8625.
30. Wang J, Polleux J, Lim J, Dunn B. Pseudocapacitive contributions to electrochemical
energy storage in TiO2 (anatase) nanoparticles. J Phys Chem C. 2007;111:1492531.
31. Shin J-Y, Samuelis D, Maier J. Sustained lithium storage performance of hierarchical nanoporous anatase TiO2 at high rates: emphasis on interfacial storage phenomena. Adv
Funct Mater 2011; 21(18): 346472.
32. Shin J-Y, Joo J, Samuelis D, Maier J. Oxygen-decient TiO2- nanoparticles via hydrogen
reduction for high rate capability lithium batteries. Chem Mater. 2012; 24(3):5431.
33. Yu Y, Gu L, Wang C, Dhanabalan A, Van Aken PA, Maier J. Encapsulation of
Sn@carbon nanoparticles in bamboo-like hollow carbon nanobers as an anode material
in lithium-based batteries. Angew Chem Int Ed. 2009;48:64859.
34. Yu Y, Gu L, Zhu C, Van Aken PA, Maier J. Tin nanoparticles encapsulated in porous
multichannel carbon microtubes: preparation by single-nozzle electrospinning and application as anode material for high-performance Li-based batteries. J Am Chem Soc. 2009;
131:159845.
35. Peled E. The electrochemical behavior of alkali and alkaline earth metals in nonaqueous
battery systems the solid electrolyte interphase model. J Electrochem Soc. 1979;126:
204751.
36. Balbuena PB, Wang Y. Lithium-ion batteries: solid-electrolyte interphase. 1st ed. London,
UK: Imperial College Press; 2004.
37. Armand MB, Bruce PG, Forsyth M, Scrosati B, Wieczorek W. Polymer electrolytes. In:
Bruce DW, OHare D, Walton RI, editors. Energy materials. Chichester, UK: John Wiley
& Sons; 2011.
38. Kreuer K-D, Wohlfarth A, de Araujo CC, Fuchs A, Maier J. Single alkaline-ion (Li+, Na+)
conductors by ion exchange of proton-conducting ionomers and polyelectrolytes. ChemPhysChem. 2011;12(14):255860.
39. Maier J. Concentration polarization of salt-containing liquid electrolytes. Adv Funct
Mater. 2011;21(8):144855.
40. Arora P, Zhang Z. Battery separators. Chem Rev. 2004;104:441962.
41. Armand M, Chabagno JM, Duclot M. Poly-ethers as solid electrolytes. In: Vashishita P,
Mundy JN, Shenoy GK, editors. Fast ion transport in solids. Amsterdam: North-Holland
Publishing; 1979.
3.5.10 References
247
42. Maier J. Acidbase centers and acidbase scales in ionic solids. Chem Eur J. 2001;7
(22):476270.
43. Badway F, Cosandey F, Pereira N, Amatucci GG. Carbon metal uoride nanocomposites. J Electrochem Soc. 2003;150(10):A131827.
44. Idota Y, Kubota T, Matsufuji A, Maekawa Y, Miyasaka T. Tin-based amorphous oxide:
a high-capacity lithium-ion-storage material. Science. 1997;276(5317):13957.
45. Poizot P, Laruelle S, Grugeon S, Dupont L, Tarascon JM. Nano-sized transition-metal
oxides as negative-electrode materials for lithium-ion batteries. Nature. 2000;407(6803):
4969.
46. Li H, Balaya P, Maier J. Li-storage via heterogeneous reaction in selected binary metal
uorides and oxides. J Electrochem Soc. 2004;151(11):A187885.
47. Delmer O, Balaya P, Kienle L, Maier J. Enhanced potential of amorphous electrode
materials: case study of RuO2. Adv Mater. 2008;20(3):5015.
48. Delmer O, Maier J. On the chemical potential of a component in a metastable phaseapplication to Li-storage in the RuO2-Li system. Phys Chem Chem Phys. 2009;11(30):
6424-6429.
49. Balaya P, Li H, Kienle L, Maier J. Fully reversible homogeneous and heterogeneous
Li storage in RuO2 with high capacity. Adv Funct Mater 2003;13:6215.
50. Maier J. Mass storage in space charge regions of nano-sized systems (Nano-ionics.
Part V). Faraday Discuss. 2007;134:5166.
51. Maier J. Nanoionics: ion transport and electrochemical storage in conned systems.
Nature Mater. 2005;4(11):80515.
52. Kaskhedikar N, Maier J. Lithium storage in carbon nanostructures. Adv Mater 2009;21
(2526):266480.
53. Amin R, Maier J, Balaya P, Chen DP, Lin CT. Ionic and electronic transport in single
crystalline LiFePO4 grown by optical oating zone technique. Solid State Ionics.
2008;179(2732):16837.
54. Maier J, Amin R. Defect chemistry of LiFePO4. J Electrochem Soc. 2008;155(4):
A33944.
55. Sigle W, Amin R, Weichert K, van Aken PA, Maier J. Delithiation study of LiFePO4
crystals using electron energy-loss spectroscopy. Electrochem Solid-State Lett. 2009;12
(8):A1514.
56. Maier J. Ionic conduction in space charge regions. Prog Solid State Chem. 1995;23(3):
171263.
57. Balaya P, Bhattacharyya AJ, Jamnik J, Zhukovskii YF, Kotomin EA, Maier J. Nanoionics in the context of lithium batteries. J Power Sources. 2006;159(1):1718.
58. Guo Y-G, Hu YS, Sigle W, Maier J. Superior electrode performance of nanostructured
mesoporous TiO2 (anatase) through efcient hierarchical mixed conducting networks.
Adv Mater 2007;19(16):208791.
59. Maier J. Thermodynamics of nanosystems with a special view to charge carriers. Adv
Mater 2009;21(2526):257185.
60. Wagemaker M, Singh DP, Borghols WJH, et al. Dynamic solubility limits in nanosized
olivine LiFePO4. J Am Chem Soc. 2011;133(26):102228.
61. Gu L, Zhu C, Li H, et al. Direct observation of lithium staging in partially delithiated
LiFePO4 at atomic resolution. J Am Chem Soc. 2011;133(13):46613.
62. Van der Ven A, Garikipati K, Kim S, Wagemaker M. The role of coherency strains on
phase stability in LixFePO4: needle crystallites minimize coherency strain and overpotential. J Electrochem Soc. 2009;156:A94957.
63. Tang M, Carter WC, Belak JF, Chiang Y-M. Modeling the competing phase transition
pathways in nanoscale olivine electrodes. Electrochim Acta. 2010;56:96976.
248
64. Meethong N, Huang H, Speakman S, Carter WC, Chiang Y-M. Strain Accommodation
during phase transformations in olivine-based cathodes as a materials selection criterion
for high-power rechargeable batteries. Adv Funct Mater 2007;17:111523.
65. Dreyer W, Jamnik J, Guhlke C, Huth R, Mokon J, Gabercek M. The thermodynamic
origin of hysteresis in insertion batteries. Nature Mater. 2010;9:44853.
66. Delmas C, Maccario M, Croguennec L, Le Cras F, Weill F. Lithium deintercalation in
LiFePO4 nanoparticles via a domino-cascade model. Nature Mater. 2008;7:66571.
67. Chen G, Song X, Richardson TJ. Metastable solid solution phases in the LiFePO4/FePO4
system. J Electrochem Soc. 2007;154:A62732.
68. Wagemaker M, Mulder FM, Van der Ven A. The Role of Surface and Interface Energy on
Phase Stability of Nanosized Insertion Compounds. Adv Mater. 2009;21(2526):27039.
69. Hu YS, Adelhelm P, Smarsly B, Hore S, Antonietti M, Maier J. Synthesis of hierarchically
porous carbon monoliths with highly ordered microstructure and their application in
rechargeable lithium batteries with high-rate capability. Adv Funct Mater. 2007;17(12):
18738.
70. Hu YS, Guo YG, Sigle W, Hore S, Balaya P, Maier J. Electrochemical lithiation synthesis
of nanoporous materials with superior catalytic and capacitive activity. Nature Mater.
2006;5(9):7137.
71. Guo Y-G, Hu Y-S, Sigle W, Maier J. Superior electrode performance of nanostructured
mesoporous TiO2 (anatase) through efcient hierarchical mixed conducting networks.
Adv Mater. 2007;19:208791.
72. Schrder A, Fleig J, Gryaznov D, Maier J, Sitte W. Quantitative model of electrochemical
Ostwald ripening and its application to the time-dependent electrode potential of nanocrystalline metals. J Phys Chem B. 2006;110(25):1227480.
73. Maier J. Thermodynamics of electrochemical lithium storage. Angew Chem Int Ed,
in print.
4.1.1 Theory
4.1.1.1 Introduction
Any chemical conversion requires time. The time-dependent development of a reaction,
the reaction rate r, is expressed by its kinetics, which describes the correlation of r and the
determining factors, such as concentration, pressure, temperature, diffusion, catalyst,
mass and heat transfer, and so forth.
The reaction rate determines the application potential and/or the need of improvement to accelerate the conversion. For example, the production of biogas via fermentation of biomass and the onboard production of H2 via steam reforming of CH3OH
require different times for a certain degree of conversion. However, for the biogas
production the time is of less importance, as one can simply wait and/or increase
the size of the fermentation vessel. This is obviously not possible for the onboard
production of H2 in a vehicle.
Therefore, it is important to determine a reaction rate. Moreover, it is important to
know how to determine it because the reaction rate is inuenced by several factors,
hampering the comparison of reactions, reactors and catalysts. The next sections
will shortly introduce the basic concepts for the description of the kinetics of a heterogeneously catalyzed reaction and their practical applications; followed by three
real-research examples demonstrating the application and possible problems in detail.
250
Gas flow
7
6
1
Reactor bed
1. Film diffusion
Catalyst particle
2. Pore diffusion
Catalyst pore
3. Adsorption
345
4. Reaction
5. Desorption
Boundary layer
6. Pore diffusion
Active center
7. Film diffusion
Figure 4.1.1 Delineation of the steps of a catalyzed chemical reaction. In an ideal case, the
reactant molecule diffuses through the boundary layer and the pores. Near the active center, it
adsorbs, reacts, and desorbs and is the being transported back via pore and lm diffusion.
4.1.1 Theory
Boundary
layer
251
Solid
Concentration
Gas
a
Local position
Figure 4.1.2
Concentration prole in the catalyst particle if the reaction is much faster (a) or
of comparable rate (b) or much slower (c) than the transport of the reactants to the active site.
(1)
(2)
252
Energy
ATS
ATS
A
Aads.
B
Reaction coordinate
Figure 4.1.3
For the uncatalyzed reaction (blue), the activation energy required for the reaction is high. If catalysis takes place (red), a transition state in which the reactant has adsorbed
on the catalyst surface exists. In such a case, the required activation energy for the conversion
of A to B is considerably lower.
If a reaction has to be divided into more than one elementary reaction, it is called a
reaction network. The complexity of such reaction networks can be very different,
ranging from just two elementary reactions to a network consisting of parallel-,
side-, subsequent-, and equilibrium reactions. Details about more complicated reactions, such as bimolecular reactions, reversible reaction steps and reactions with different kinds of adsorption (chemical, physical, dissociative, etc.), can be found in the
typical literature [14].
4.1.1.3.2 Reaction Rate The time dependent evolution of the concentrations of reactant cA or product cB, taking the stoichiometric coefcients into account, is a measure for the reaction rate or its velocity, cf. Equation (3). The convention requires a
negative sign for reactants, indicating consumption, and a positive sign for the
products, indicating formation:
r=
1 dcA 1 dcB
=
a dt b dt
(3)
It is also possible to take additional factors into account to further characterize the
reaction, for example, F being the reaction volume, the mass of the catalyst, the surface of the catalyst, or the interface of a two-phase system, and so forth, as shown in
Equation (4):
r=
1 dcA
1 dcB
=
a F dt b F dt
(4)
4.1.1 Theory
253
any kind of calculations, inter- and extrapolations. An analytical solution of Equation (3) is possible for elementary reactions, with a being the stoichiometric coefcient and the exponent of the concentration in the rate law. For a reaction
network consisting of more than one elementary reaction, it is still possible to
build a model with an appropriate number of differential equations for the key compounds. However, it might not be analytically but only numerically solvable. Details
can be found in many text books on physical chemistry or catalysis [1, 3]. A common
approach to express Equation (3) has the form of a polynomial with empirical
exponent for the concentration, as shown in Equation (5):
r = k cA n
(5)
P = pre-exponential factor
Ea = activation energy [J/mol]
R = universal gas constant, 8.314 [J/mol K]
T = absolute temperature [K].
(6)
254
For a reaction with a preceding adsorption step, two rate constants for the equilibrium reaction and one rate constant for the conversion to B exist, as shown in
Equation (7):
k1
k2
Aads: !B
AA
k
(7)
It is a reasonable assumption that the equilibrium reaction proceeds much faster than
the conversion. The reaction rate of B is shown in Equation (8):
r=
dcB
= k2 cAads:
dt
(8)
The concentration of the adsorbed species can be obtained by compiling the formation term and the consumption term of the rate for the adsorbed species, as displayed
in Equations (9) and (10):
Formation :
dcAads:
= k1 cA
dt
Consumption :
dcAads:
= k1 cAads: + k2 cAads:
dt
(9)
(10)
In the steady state, these two equations are equal. The resulting equation can be
converted, giving the concentration cAads:
k1 cA = k1 cAads: k2 cAads:
cAads: =
k1
cA
k1 + k2
(11)
(12)
The obtained expression for cAads: can now be used to obtain the reaction rate for B,
as shown in Equation (13). If the rate constant of desorption (k1) is much higher
than the rate of conversion to the product (k2), cAads: can be expressed from the
adsorption constant K by cAads: = K cA .
k1
(13)
cA
r = k2 cAads: = k2
k1 + k2
k2 K cA
(14)
= keff cA
(15)
The product k2 K can now be comprised to a new rate constant keff , which is a
rate constant for the overall reaction. In order to account for the temperature dependency of keff , the temperature dependencies for both, k2 and K, have to be considered.
While k2 follows the temperature according to the Arrhenius equation, Equation (6),
the temperature dependency of K is given by the vant Hoff Law, cf. Equation (16):
4.1.1 Theory
d( ln K) H O
=
dT
RT 2
255
(16)
keff = Peff e RT
(17)
kapp = Papp e RT
(18)
A linearization of the Arrhenius Equation (6), (17), or (18) as shown in Equation (19)
is called an Arrhenius plot. It is very useful for the discussion of several different
aspects:
ln (kapp ) = ln (Papp )
Eapp 1
T
R
(19)
Ln(k)
The slope of Equation (19) is the activation energy or the effective or apparent activation energy. As shown in Figure 4.1.4, a comparison of different catalysts by
Suitable catalyst
Less suitable
catalyst
Without
catalyst
1/Temperature
Figure 4.1.4 Delineation of activation energy for different catalysts. The uncatalyzed one has
the highest slope and therefore the highest activation energy. The better catalyst has a smaller
slope and therefore a smaller activation energy, as compared to the less suitable catalyst.
256
Limited by pore
diffusion
Ln(k)
No limitation
by mass
transport
1/Temperature
Figure 4.1.5 The activation energy for a given reaction in the case of mass transport limitation (boundary layer diffusion and pore diffusion, respectively) and in the case of the absence
of mass transport limitation.
comparing their activation energy is possible. A low slope, and therefore a small activation energy, is usually a desirable feature of a good catalyst. The uncatalyzed
thermal reaction usually exhibits the highest activation energy.
There are two options for the utilization of a catalyst providing a low activation
energy for a reaction. It can be operated at the same temperature than other catalysts, achieving a higher productivity or the temperature of the reaction may be
decreased.
However, care has to be taken in such cases because a different slope can also be
an indication for different regimes of transport limitation, as shown in Figure 4.1.5.
At very high temperatures (small values for 1/T ), the reaction is so fast that the mass
transport into the particles is the limiting factor for the reaction. Because this process
is nearly independent of the temperature, a temperature change does not affect the
observed rate. At lower temperatures, the transport inside the pores becomes the limiting factor. The observed activation energy is approximately 1/2 of the real activation energy. However, at even lower temperatures (or higher 1/T values) the reaction
rate is determined by the chemical reaction itself and not by any transport limitations. In this case, the slope is the highest and the determined activation energy is
the real one.
4.1.1.3.4 Limitations The interaction of the kinetics of the chemical reaction and
transport phenomena of the reactions are described as macrokinetics, in which transport phenomena, such as mass and heat transfer, adsorption, and desorption have a
substantial impact on the time-dependent development of the reaction. The macrokinetic is important for reactor engineering, construction, operation, and safety.
However, for development and comparison of heterogeneous catalysts, the macrokinetic is not suitable, as the comparison of different catalytic materials is often
hampered by the transport phenomena.
4.1.1 Theory
257
The intrinsic kinetics describes a reaction rate that is not inuenced by such transport phenomena; therefore, it only depends on the factors concentration, pressure,
temperature, and catalyst. For the comparison of the catalytic activity and the investigation of different catalysts, it is necessary to adjust the experimental conditions
such that only the intrinsic kinetics is determined. If this is not the case, none of
the obtained data are of use. The microkinetics is equivalent to the intrinsic-kinetic,
with the difference that it consists of the elementary reactions.
4.1.1.3.4.1 Mass Transport
(20)
The heat transport is treated similar to the transport of mass. There are three ways
for heat transport: thermal conduction, convection and radiation. If the heat transport is limited, hot spots can occur in the catalyst bed, causing deactivation of the
catalyst. Moreover, if a hot-spot occurs, the temperature at which is the reaction
occurs is unknown. However, in typical laboratory equipment (small reactor, small
particle size, diluted catalyst, and limited conversion), this is usually not a problem.
4.1.1.3.4.3 Estimation Criteria
There are several criteria equations described in literature to check for the presence
or absence of transport limitations. An overview of different criteria can be found
here [46]. However, there are many more textbooks listing these criteria.
258
259
n+1
re c
1
2
De ci;0
(21)
[ mol
m ]
3
260
It is noteworthy that catalytic activity is not a general term with a xed denition
and no absolute quantity. In the literature this term can have several meanings. Usually, the activity of a catalyst is used to describe how much reactant is converted. Improving the activity means the catalyst converts more reactant, not considering if the
desired product or an unwanted by-product is produced. Thus the term catalytic performance can be used, which is closely related to the yield, taking into account how
much of the desired product was formed. But it also does not have an absolute quantity. A catalyst with a high conversion and a low selectivity and one with a low conversion and a high selectivity can give the same yield. The latter catalyst has the
better catalytic performance because it produces fewer by-products.
For many kinetic studies the selectivities of different catalysts are compared at the
same level of conversion. However, the main drawback is that sometimes very different reaction conditions (mass of catalyst, residence time, reactor temperature, feed
gas composition) are required to achieve this. Another option is the comparison of
conversion and selectivity under a given set of reaction conditions. The main drawback of this approach appears if catalysts show full or very different conversions,
then the comparison of the selectivities is not appropriate. Despite this, it is still possible to conclude if a catalyst heavily favors one of several reaction pathways, for
example, if partial or total oxidation is prevailing.
Therefore, to compare the catalytic activity, or better the catalytic performance,
of a series of catalysts, several factors such as reaction conditions, conversion, and
selectivity must be taken into account.
261
SC2 H6 +C2 H4 =
si,0 si
si,0
(
P
(Products) + si ) si
P
(Products) + si
sP1 sP1;0
Ratio of stoichiometric coefficients
si;0 si
=P
sP1 sP1;0
Ratio of stoichiometric coefficients
(Products) + si
(22)
(23)
(24)
(25)
The dilution of the feed can be useful to prevent hot-spots in the catalytic bed or to
enable experiments at small conversions, which are necessary for the determination
of kinetic parameters. However, for such cases it is important to check if the
catalyst is stable with and without dilution of the feed.
However, measurements at small conversions are also limited by another factor,
the analytic power. Regardless of the applied analytic methods (GC, MS, HPLC, etc.),
there is a detection limit. For small conversions of reactants, the amount of products
can be rather small, falling near the detection limit.
It is very useful to determine a mass balance for at least one element, comparing
the amount of unconverted reactants and the amount of products. For a proper
determination of kinetic parameters the mass balance should be closed at least up
to 95%. Commonly carbon is used for the determination of the mass balance, as
all carbon containing materials are rather easy to be quantied, compared to H2
or H2O, for example.
Moreover, it is useful to detect and quantify as many reaction products as possible
and as accurate as possible. A gas chromatograph is usually to be preferred compared to a mass spectrometer and the applied detectors should be as sensitive as possible. For example, for the detection of hydrocarbons and COx, an FID in
combination with a TCD + methanizer should be preferred to only a TCD, due to
the higher sensitivity.
Furthermore, it is necessary to ensure that after sample taking, the reaction is
being quenched immediately. This is usually not a big problem for reactions at higher
temperatures, for example, the oxidative conversion of hydrocarbons. However, for
reactions carried out in the liquid phase near room temperature or at only slightly
elevated temperatures, this might very well be a problem.
4.1.2.4.2 Determination of Time The exact determination of the kinetic parameters
requires not only the exact determination of dc but also of the elapsed time dt.
Taking samples from a batch reactor, the determination of dt is rather simple: it is
262
the time span from the start of the reaction until the sampling, provided that the reaction is quenched in the sample.
If there is a volume ow into and out of the reactor, the residence time is taken
into account, as dened in Equation (26). A very important remark is that the volume ow of a gas changes with the applied reactor temperature, having a great
impact on the actual residence time:
=
VR
V_
(26)
= residence time [ 1s ]
VR = volume of reactor [m3 ]
3
V_ = volume ow [ ms ]
However, if the reactor is lled, for example, with a catalyst, the situation becomes
more complicated. The VR would be the empty volume of the reactor, which is then
difcult to determine, for instance, using settled apparent densities. The residence
time can also be experimentally determined, usually resulting in a residence time distribution; however, the experimental effort for such experiments is often large.
Therefore, it is useful to apply a modied residence time, as shown in Equation
(27), which denes the ratio of the mass of the catalyst and the gas ow, two easily
measurable values:
modified =
mCatalyst
V_
(27)
1
modied = modied residence time [ skg
]
m = mass [kg]
It has to be stressed that the exact determination of dc and dt is necessary. If
only one parameter is accurately measured and the other one not, the error of the
calculated kinetic parameters will still be large.
263
For the simple elementary reaction, A B, the reaction rate r is correlated to the
reaction conditions by the rate law composed of the rate constant k, the reactant
concentration cA, and the order of the reaction a. This is analytically solvable for
elementary reactions:
1 dcA
= k (cA )a
r=
a
dt
(5)
In a power law approach for the description of a more complex reaction, a numerical
t of the data points is necessary and a is exchanged for n, the apparent order of
the reaction. Moreover, if two and more reactants take part in the reaction, all of
them will have their own, respective reaction order. A separate series of experiments
is necessary for the determination of each reaction order.
4.1.2.4.3.3 Activation Energy
For the determination of the activation energy, it is necessary to determine the rate
constant k for several different temperatures, as described previously. The linearization of Equation (6), for an elementary reaction, or Equation (18) for an overall reaction gives the activation energy or the apparent activation energy as the slope of an
Arrhenius plot:
ln (k) =
ln (keff ) =
Ea 1
R T
(28)
Eapp 1
T
R
(29)
4.1.2.4.3.4 Miscellaneous
The turnover frequency (TOF) and the turnover number (TON), as shown in Equations (31) and (30), are frequently used parameters for the characterization of the catalyst performance. However, while the determination of the number of active centers
is not too difcult for homogeneous catalysts, it can be a very intricate subject for
heterogeneous catalysts. If it is useful to determine a TOF or TON for a catalyst,
whose number of active center is unclear, is left to the readers opinion.
TON =
(30)
TOF =
(31)
264
4.1.3 Examples
4.1.3.1 Oxidative Coupling of Methane
The oxidative coupling of methane (OCM) to ethane and ethylene is a very attractive
reaction for the conversion of natural and biogas, see Equation (32). However, the
reaction suffers from severe drawbacks, such as high reaction temperatures, lack
of stable and selective catalysts, and an extremely complex reaction mechanism.
2 CH4 +
1
2
O2 C2H6 + H2O
(32)
re =
(33)
p
The concentration cCH4;0 is obtained via the ideal gas law to Vs = RT
. For the diffusion
of gases, the two most important cases are molecular diffusion (diffusion in the gas
phase) and Knudsen diffusion (diffusion through pores, while the number of collisions between the pore wall is larger than the number of collisions between the gas
4.1.3 Examples
265
molecules). Since OCM catalysts have typically low surface areas and a low porosity,
it is reasonable to assume that the importance of the Knudsen diffusion is low and
the diffusion coefcient for the molecular diffusion is used. The molecular diffusion
can easily be calculated with the help of the kinetic gas theory; however, typical
values are within the range of 107m2/s to 105m2/s, with the latter value being the
worst case scenario. For porous catalysts, the effective diffusion coefcient has
to be considered, as shown in Equation (34):
De = D
(34)
= porosity
= labyrinth factor
Moreover, for high temperatures as applied in the OCM, the diffusion coefcients
can be expected to be even higher.
= (0:0002 m)2
4+1
2
mol
kg
1,300 3
kg s
m
1
2
m
mol
5
5:3 3
10
m
s
38:5 103
(35)
= 0:01 1
l
n
c
mol
= 38:5 103 kgs
observed effective reaction rate
The obtained value is two orders of magnitude below 1; therefore, mass transport limitations can be considered to be absent. It is worth mentioning that the calculation for
values such as c is not absolutely correct; in particular, the estimation of d is not
exact. But it is sufcient for an estimation, as done in with the Weisz Prater criterion.
For the catalyst testing, several important points were taken into consideration.
The catalyst was diluted in quartz sand in order to avoid hot spots and to allow a
proper plug-ow of the gas through the reactor bed. The reaction conditions (amount
of catalyst, feed gas composition, temperature) were chosen, such that the conversion
of CH4 and O2 is limited. The so called kinetic regime is valid. Usually, several preliminary experiments or a lot of experience is necessary to determine the appropriate
reaction conditions, but this is hardly mentioned in the nal publication of the results.
Once this was done, the stability of the catalysts was determined by time on stream
experiments, as shown in Figure 4.1.6 for Li/MgO samples prepared via wet impregnation. The time on stream experiments revealed that the Li-doped MgO catalyst is
instable, irrespective of the preparation procedure or the Li-loading. After 40 h time
266
10
10
15
20
Time [h]
25
30
35
40
Figure 4.1.6 The CH4 conversion versus time on stream for Li-doped MgO with different
loadings of Li, prepared via wet impregnation. Reaction conditions: 750C; 100 mg catalyst;
feed gas composition, CH4:O2:N2 = 4:1:4. The strong deactivation is evident, apparent stability is only due to the scaling of the axis. A stable range has not been reached yet. Unchanged
gure adapted from [11]. Reprinted from Topics in Catalysis Vol. 54, S. Arndt, U. Simon, S.
Heitz, A. Berthold, B. Beck, O. Grke, J. D. Epping, T. Otremba, Y. Aksu, E. Irran, G. Laugel, M. Driess, H. Schubert, R. Schomcker, Li-doped MgO from different preparative routes
for the oxidative coupling of methane, 12661285, Copyright (2011), with permission from
Springer via the Copyright Clearance Center.
on stream, none of the catalysts had reached a stable state. Some catalysts seem to be
stable after 20 h; however, this is only due to the scaling of the axis.
A gap in the trajectories of 4 wt% and 8 wt% Li/MgO appears at approximately
7 h. This can be explained by looking into the calculation of conversion and selectivity in detail. The denition of the conversion X of CH4 is given in Equation (24), and
the selectivity for the desired products in Equation (25).
There are two ways to obtain si,0 and si. In the rst way, CH4 is measured in the
bypass mode before or during the reaction and a comparison with the amount of
CH4 after the gas passed the reactor is done. However, this procedure has drawbacks: the gas ow through the reactor is never absolutely stable due to uncertainties
of the mass ow controllers; additionally, during such measurements one cannot
measure the outlet gas of the reactor. Therefore, the obtained values have large errors, due to summation, especially at low conversions. In the second way, the sum
of the remaining amount of CH4 and all the products is used to obtain si,0, and
the sum of the reaction products is used to obtain si,0 si. Therefore, the calculation
can result in much lower uncertainties. For the detection of the hydrocarbon species,
a ame ionization detector (FID), and for CO and CO2 a thermal conductivity detector (TCD), was used. When COx falls below the detection limit of the TCD, it drops
4.1.3 Examples
267
out of the sum to calculate si,0 and si,0 si. In that moment, a part of the actual conversion is not detected/calculated anymore; therefore, the conversion suddenly drops
as observed for 4 wt% and 8 wt% Li/MgO at 7 h time on stream. The FID is more
sensitive than a TCD by approximately a factor of 1,000, which means that hydrocarbon products are still detected, while COx species are not, due to the small
amounts. Therefore, the calculation of the selectivity is meaningless, as without the
detection of the COx species, it is 100%. However, this is only a problem at very
small conversions.
The duration of the time on stream experiments for the determination of the
catalyst stability depends on the intended application. More than 12 h is often useful.
Due to the strong deactivation, it is close to impossible to determine kinetic parameters because all data measured are only snapshots on the deactivation trajectory.
Moreover, since the active center of the Li-doped MgO catalyst is unknown [12], it
is not reasonable to calculate values such as TON or TOF. That the apparent
activation energy also depends on the state of deactivation is also shown in [11].
It is worth mentioning that the Li-doped MgO reacts with the quartz reactor,
which is known in the literature [12]. The Li leaves the catalysts rather quickly as
the volatile LiOH and it then reacts with the SiO2 of the inert diluent or the reactor
to form Li2SiO3. Therefore, the inertness of the reactor is another important subject,
which can be, depending on the required reaction conditions and catalyst, a rather
difcult task.
In the case of kinetic studies with Li/MgO the catalyst is only stable in the presence
of 80% inert gas in the feed, only such conditions allow the evaluation of the
kinetic parameters.
268
B
H
H2
H
N
N2
NH2
N
N
Zr
Zr
Figure 4.1.7 The proposed reaction mechanism of Soerijanto et al. assuming a Mars
van-Krevelenlike reaction mechanism considering the structure of the catalyst, especially
the oxygen vacancies. In (A), NH3 adsorbs into a vacancy adjacent to a nitrogen atom of
the bulk. After delivering of the hydrogen atoms, the remaining nitrogen reacts irreversibly
with the neighboring atom to N2 (B). In contrast to the surface reaction mechanism, the
formed N2 molecule consists of nitrogen from NH3 and the catalyst. As a result of this reaction
step, another vacancy is formed on the catalysts surface, which is lled with nitrogen of
another NH3 molecule, delivering more H2. Unchanged gure adapted from [13]. Reprinted
from Applied Catalysis A: General, Vol. 392, T. Otremba, N. Frenzel, M. Lerch, T. Ressler,
and R. Schomcker, Kinetic studies on ammonia decomposition over zirconium oxynitride,
103110, Copyright (2011), with permission from Elsevier via the Copyright Clearance Center.
Reaction 3 was considered to be the RDS. Therefore, the reaction rates of 1, 2, 4, and
5 are zero, and the reaction rate of 3 represents the rate of the catalytic cycle. A rate
for the overall reaction depending on the assumed mechanism can be calculated. To
simplify the tting, the reaction rate was further simplied by combining the rst and
the last reaction step and the according reaction rate, see Equation (36), or by taking
the reaction steps 1, 2, 3 and a combination of 4 and 5 into account, see Equation
(37), with k3* = k3 [cs].
R=
R=
K k* K p(NH3 )2
12 3 3 45
K45 p(NH3 ) p(H2 )2 + K12 p(NH3 ) + p(H2 )3
(36)
3
2
(37)
+ p(H2 )3
The development of the concentration as a function of the residence time was experimentally determined for several different reaction conditions, such as modied residence time, temperature, and feed gas composition. This is shown in Figure 4.1.8.
Moreover, it is important to perform experiments adding product to the feed gas
stream, to investigate if a product inhibition of the reaction can occur.
The t of the experimental data to the developed model is done by varying the
equilibrium constants and one rate constant, until the best possible t is obtained
by minimizing the least square method (the most commonly applied method).
Such tting can be done with a variety of simulation software tools.
Step
k5, f
5, r
3
N00 V 00
(V0 (NH3 )0 )0 a
0 + 2 H2 (g)
k
4,r
k4, f
(V0 (NH3 )0 )0
(V0 V0 )0 + NH3 (g) a
k
2,r
1,r
2
r5 = k5, f [(V0 (NH3 )0 )0 ] k5,r [N00 V 00
0 ] p(H2 ) = 0
r3 = k3 [(N0 N0 )0 ] = 0
k3
3
0
2
r2 = k2, f N00 (NH3 ) 00
0 k2,r [(N0 N0 ) p (H2 ) ] = 0
N00 (NH3 ) 00
(N0 N0 )0 + 32 H2 (g)
0 a
k
k2, f
p(NH3 ) k1,r [N00 (NH3 )00
r1 = k1, f N00 V 00
0 ]=0
0
k1, f
Reaction Rate
Elementary Reaction
Table 4.1.1 Illustration of the derivation of the elementary steps. Step 1: Adsorption of NH3 on catalyst surface; step 2: rst release of H2; step 3:
N2-formation of one N atom from catalyst and one from former NH3; steps 4 and 5: lling of one vacancy with another N atom from NH3.
Unchanged table adapted from [13]. Reprinted from Applied Catalysis A: General, Vol. 392, T. Otremba, N. Frenzel, M. Lerch, T. Ressler,
and R. Schomcker, Kinetic studies on ammonia decomposition over zirconium oxynitride, 103110, Copyright (2011), with permission from
Elsevier via the Copyright Clearance Center.
4.1.3 Examples
269
270
800
700
600
500
400
300
0.0
475 C
525 C
550 C
575 C
600 C
0.2
0.4
0.6
0.8
1.0
Modified residence time / s.g/ml
1.2
1.4
Figure 4.1.8
Residual NH3 partial pressures after leaving reactor as a function of the modied residence time ( x )/mcat. The comparison of the three curves shows the inuence of H2
on the catalyst. (V = 50 mL/min, p = 1 bar (a), T = 600C, mcat= 2.0 g ZrON catalyst). Unchanged gure adapted from [13]. Reprinted from Applied Catalysis A: General, Vol. 392,
T. Otremba, N. Frenzel, M. Lerch, T. Ressler, and R. Schomcker, Kinetic studies on ammonia decomposition over zirconium oxynitride, 103110, Copyright (2011), with permission
from Elsevier via the Copyright Clearance Center.
If such a t is impossible, the assumed reaction mechanism is incorrect and a different one has to be chosen or developed. However, in the present case, a suitable
agreement was reported. Moreover, the enthalpies and their errors could be determined. Based on the reaction mechanism and the determined enthalpies, a reaction
pathway was proposed, as shown in Figure 4.1.9. Its agreement with theoretical
predictions supports the proposed mechanism.
4.1.3 Examples
271
150
100
EA
[(NONO)]*
NH3(g)
3/2 H2(g)
NO'VO
N2(g)
3 H2(g)
50
2 NH3(g)
NO'VO
0
50
100
150
NH3(g)
3/2 H2(g)
(NONO)
H1
NH3(g)
NO'(NH3)O
H2
H5
H4
(VOVO)
N2(g)
NH3(g)
3/2 H2(g)
((NH3)OVO)
N2(g)
3/2 H2(g)
200
Figure 4.1.9 The energy level diagram for the assumed reaction mechanism illustrates the
energy of the system as a function of the reaction coordinate. The dashed lines represent the
error bars of the determined values. At the beginning, the energy of the system is zero, containing the active site and two NH3 molecules. In the rst step, NH3 adsorbs at the active site and
the energy H1 is freed. Next, the energy of H2 is needed to release the H2 molecules. This
energy level corresponds to H1,2 because this is the sum of H1 and H2. After that, the activation energy Ea lifts the system into an excited stage. N2 is released, and the systems energy
drops. NH3 adsorbs to the double vacancy, and the energy H4 is released. The last step is the
release of H2, which needs an energy of H1. Figure adapted from [13]. Reprinted from
Applied Catalysis A: General, Vol. 392, T. Otremba, N. Frenzel, M. Lerch, T. Ressler, and
R. Schomcker, Kinetic studies on ammonia decomposition over zirconium oxynitride,
103110, Copyright (2011), with permission from Elsevier via the Copyright Clearance Center.
The saturation concentration of the gaseous reactant in the liquid phase can be
calculated via Henrys law. However, vigorous stirring is necessary to avoid the formation of a concentration gradient of the molecules of the gas phase in the liquid
phase. If this is not ensured, the obtained kinetic data are mostly useless because
the saturation concentration in the liquid phase cannot be accurately determined.
The hydrogenation of 1,5-cyclo-octadien (COD) to cyclo-octene (COE) is performed in a slurry reactor. The reaction is relevant because the product is an intermediate for the production of special polymers. However, this reaction suffers
from the drawback that the hydrogenation does not stop at cyclo-octene, because
a full hydrogenation to cyclo-octane (COA) is possible, as shown in Figure 4.1.12.
Schmidt and Schomcker studied the hydrogenation of COD on Pd/-Al2O3 in a
slurry reactor [16]. H2, as gaseous reactant, was pressed into the reactor, but mass
transfer of the H2 from the gas phase to the liquid phase is necessary. To avoid a concentration gradient, the reactor was vigorously stirred. Only when the reaction rate is
272
Gas phase
Liquid phase
Catalyst particle
Stirrer
Figure 4.1.10
A slurry reactor for a three-phase reaction, between a liquid and a solid, catalyzed by a solid catalyst.
Boundary
layer
Liquid
Boundary
layer
Concentration
Gas
Solid
a
Local position
H2
H2
4.1.4 Notes
273
100
Selectivity COE/%
Conversion COD/%
100
80
60
0.2 MPa
0.5 MPa
1.0 MPa
40
20
0
95
90
0.2 MPa
0.5 MPa
1.0 MPa
85
80
20
40
60
Time/min
80
100
0
B
20
40
60
80
Conversion COD/%
100
Figure 4.1.13 COD conversion (A) and COE selectivity (B) at different hydrogen pressures,
50C, and cCOD,0: 0.41 mol/L. Simulation (lines) and experimental data (symbols). Unchanged
gure adapted from [16]. Reprinted from Industrial & Engineering Chemistry Research, Vol.
46, A. Schmidt and R. Schomcker, Kinetics of 1,5-cyclooctadiene hydrogenation on Pd/Al2O3, 103110, Copyright (2007), with permission from Elsevier via the Copyright Clearance
Center.
independent of the stirring rate there is no mass transfer limitation and the solution is
saturated with H2.
Schmidt and coworkers tried to determine the activation energy for this reaction,
by measuring the reaction rate for different temperatures. The expectation was an accelerated reaction rate at higher temperatures. However, for this system, the additional factors (the mass transfer of H2 into the liquid phase and the solubility of
H2 in the liquid phase) also need to be taken into account. They are also temperature
dependent and they do not necessarily increase with increasing temperature.
It was found that the observed reaction rate did increase with increasing temperatures, see Figure 4.1.13, but by far not as much as expected considering the activation
energy. This is a result of the reduced concentration of H2 in the liquid phase, due to
the increased temperatures.
For the kinetic investigation of every chemical reaction an individual procedure
has to be developed, considering all special aspects of the reaction and the analytics
of its components. No general procedure or standard equipment can be recommended. The three presented examples were selected to show the diversity of kinetic
studies combined with aspects common to all catalytic reactions.
4.1.4 Notes
1. There is no agreement in the literature as to whether this is an ideal reactor.
2. The amount of substance is usually abbreviated with n; however, since in this manuscript n
has been used for the reaction order, s has been chosen.
3. Usually the labyrinth factor is abbreviated wit ; however, this symbol was already used for
the residence time.
274
4.1.5 Acknowledgment
This work is part of the Cluster of Excellence Unifying Concepts in Catalysis coordinated by the Technische Universitt Berlin. Financial support by the Deutsche
Forschungsgemeinschaft (DFG) within the framework of the German Initiative for
Excellence is gratefully acknowledged.
4.1.6 Abbreviations
c
ci,0
Concentration
Concentration of the component i at the time t = 0
cs
COA
Surface concentration
Cylco-octane
COD
COE
1,5-Cyclo-octadien
Cyclo-octene
CSTR
D
De
Ea
Porosity
Activation energy
Eapp
Eeff
FID
GC
H
Gas chromatography
Enthalpy
HPLC
K
Rate constant
kapp
keff
l
MS
Characteristic length
Mass spectrometry
m
n
Mass
Order of the reaction
OCM
Thiele modulus
Weisz-Prater criterion
4.1.7 References
Catalyst density
Labyrinth factor
p
P
Pressure
Preexponential factor
Papp
Peff
PFTR
r
re
R
RDS
S
s
Selectivity
Amount of substance
time
Residence time
Modied residence time
modied
T
TCD
Temperature [in K or C]
Thermal conductivity detector
TOF
TON
Turnover frequency
Turnover number
X
V
Conversion
VR
275
Volume ow
Volume of reactor
4.1.7 References
1. Chorkendorff I, Niemantsverdriet JW. Concepts of modern catalysis and kinetics. Weinheim: Wiley-VCH Verlag GmbH & Co. KGaA; 2007.
2. Thomas JM, Thomas WJ. Principles and practice of heterogenous catalysis. Weinheim:
VCH; 1996.
3. van Santen RA, Niemantsverdriet JW. Chemical kinetics and catalysis. New York:
Plenum Press; 1995.
4. Baerns M, Behr A, Brehm A, et al. Technische Chemie. Weinheim: Wiley-VCH Verlag
GmbH; 2006.
5. Dittmeyer R, Emig G. Simultaneous heat and mass transfer and chemical reaction. In:
Ertl G, Knzinger H, Weitkamp J, editors. Handbook of heterogeneous catalysis. 2nd
ed. Weinheim: Wiley-VCH Verlag GmbH & Co. KGaA; 2008. p. 172784.
6. Kapteijn F, Moulijn JA. Laboratory testing of solid catalysts. In: Ertl G, Knzinger H,
Weitkamp J, editors. Handbook of heterogeneous catalysis. 2nd ed. Weinheim: WileyVCH Verlag GmbH & Co. KGaA; 2008. p. 201945.
276
7. Hagen J. Industrial catalysis. Weinheim: Wiley-VCH Verlag GmbH & Co. KGaA; 2006.
8. Ertl G, Knzinger H, Weitkamp J, editors. Handbook of heterogeneous catalysis. Weinheim: Wiley-VCH Verlag GmbH & Co. KGaA; 2008.
9. Fogler HF. Elements of chemical reaction engineering. Upper Saddle River: Prentice Hall
International; 2004.
10. Moulijn JA, van Diepen AE, Kapteijn F. Deactivation and regeneration. In: Ertl G, Knzinger H, Weitkamp J, editors. Handbook of heterogeneous catalysis. 2nd ed. Weinheim:
Wiley-VCH Verlag GmbH & Co. KGaA; 2008. p. 118.
11. Arndt S, Simon U, Heitz S, et al. Li-doped MgO from different preparative routes for the
oxidative coupling of methane. Top Catal. 2011;54(1618):126685.
12. Arndt S, Laugel G, Levchenko S, et al. A critical assessment of Li/MgO-based catalysts
for the oxidative coupling of methane. Catal Rev Sci Eng. 2011;53(4):424514.
13. Otremba T, Frenzel N, Lerch M, Ressler T, Schomcker R. Kinetic studies on ammonia
decomposition over zirconium oxynitride. Appl Catal A Gen. 2011;392(12):10310.
14. Soerijanto H, Rdel C, Wild U, et al. The impact of nitrogen mobility on the activity of
zirconium oxynitride catalysts for ammonia decomposition. J Catal. 2007;250(1):1924.
15. Soerijanto H. Catalytic on-board hydrogen production from methanol and ammonia for
mobile application [PhD thesis]. Berlin: Technische Universitt; 2007.
16. Schmidt A, Schomcker R. Kinetics of 1,5-cyclooctadiene hydrogenation on Pd/-Al2O3.
Ind Eng Chem Res. 2007;46(6):167781.
278
hydrogen as sustainable energy carrier depends, therefore, on the availability of efcient, robust, and low-priced catalysts composed of abundant elements making
great demands on materials and catalyst synthesis.
The synthesis of a heterogeneous catalyst signicantly inuences the catalytic performance, which is generally described by activity, selectivity, and lifetime of the catalyst (Figure 4.2.1) [32]. The linkage is highly sensitive toward changes in the
preparation procedure due to the very complex interplay between solid state, surface,
and gas- or liquid-phase chemistry during catalyst pretreatment and under conditions
of the catalytic reaction (Figure 4.2.1). In the emerging eld of photocatalysis, the
relationship between parameters like chemical composition, structure, crystallinity,
electronic properties, morphology, and the performance of the material were hardly
ever analyzed in detail. Electrocatalysis involves catalytic processes occurring at the
surface of an electrode. Also in this case, the chemical nature of the electrode material,
crystallographic structure, band structure, morphology of the constituent particles,
and surface defects signicantly affect the efciency of the catalytic process [33].
Electrodes and solid catalysts applied in the synthesis of chemicals or in emission
control are, generally, hierarchical systems comprising dimensions ranging from millimeter to nanometer scale, allowing for mass and heat transport within a reactor,
molecular transport of reactants and products through a pore system, and chemical
reactions on nanostructured, frequently multifunctional surface sites as illustrated in
Figure 4.2.2. Catalyst preparation always yields a catalyst precursor, whereas the
active phase is only formed in contact with the feed of the substrate molecules in
Raw materials
Synthesis techniques
Solid-state structure
Chemical composition
Surface structure
Surface composition
Electronic properties
Acid-base properties
Dynamic response to
reaction conditions
Morphology
Pore size/structure
Particle-size distribution
Mechanical stability
Chemical stability
Shape
Reaction/process
conditions
Figure 4.2.1 Catalyst performance is controlled by multiple synthesis and reaction parameters.
Catalyst body
Porous particle
nm
M=
O
mm
M-
279
Active site
Figure 4.2.2 Hierarchical catalyst structure with the schematic representation of active sites.
timescales ranging from seconds to weeks. Therefore, the nature of the active site
nally established at steady state depends not only on the structure of the precursor,
but also on the operation conditions (i.e. reaction temperature and chemical potential
of the gas phase or the liquid phase, e.g. determined by the pH value in an aqueous
solution). The reacting substrate molecule is activated on the catalyst surface by
undergoing a chemical interaction implying that in the course of the catalytic cycle
the molecular structure of the active surface species is subjected to electronic and
structural modications as well (i.e. heterogeneous catalysis is necessarily a highly
dynamic phenomenon with respect to the structure of the active site and the solidstate catalyst structure that accommodates this site). The single-site active center
on the surface or the ensemble of active atoms is, therefore, embedded in a exible
matrix that prevents irreversible structural transformations, which might be one reason for catalyst deactivation. Synthetic strategies aimed at stabilization of active-site
precursors signicantly affect catalyst stability and performance. This is in particular
difcult when the catalytic reactions are performed at high temperatures at which
the atoms in the lattice of the solid feature already considerable mobility. Under
these conditions, contributions from homogeneous gas-phase reactions have to be
taken into account as well. Formation and depletion of involved radical species
might be mediated by defective surface sites, which complicate the elucidation of
structure-reactivity relationships.
In this chapter, general aspects of catalyst synthesis are addressed and discussed
covering the entire scale of the catalyst body involving the macro-, meso-, and nanostructure. The unit operations of catalyst synthesis are summarized in a simplied
manner in Figure 4.2.3. The rst step involves the synthesis of the inorganic solid
by, for instance, precipitation or sol-gel synthesis including ltration, washing, and
drying procedures. After the manufacture of a shaped body in the second step, thermal treatment nally results in the catalyst or in a carrier. In the synthesis of coated
catalysts, the active phase is deposited by, for instance, impregnation, atomic layer
280
Solid-state synthesis
1) Polycondensation of
molecular precursor or
solid-state reactions
2) Phase separation
3) Removal of by-products
4) Drying
Shaping
Mechanical treatment
Mixing
Drying
Bulk catalyst
Support
Supported catalyst
deposition (ALD), or deposition precipitation onto the support, again involving separation, purication, and drying operations. The nal step that yields a supported
catalyst is another thermal treatment. All these unit operations have a strong impact
on catalyst performance, sometimes described with the term chemical memory.
Reproducible synthesis of catalysts, therefore, requires strict process control by application of analytical tools. Understanding the synthesis and rational catalyst improvements is only possible by scientic analysis of all unit operations considering the
catalyst as a hierarchical system and the catalyst design as a loop that comprises
solid-state synthesis and activation. Within the scope of this short overview, a detailed description of the numerous synthetic approaches is impossible. Instead,
some of the most important basic principles are outlined, referring to literature for
elaboration of the concepts. The examples that have been selected are primarily
studies performed in our laboratory.
281
which the active sites are embedded, is given by the catalytic reaction and represents
one crucial, but not the only, aspect of catalyst design. The number of potential
active sites strongly depends on the specic surface area of the catalyst that can be
controlled by synthetic means. The surface area is related to crystallite or particle
size, particle morphology, surface texturing, and porosity. The latter is referred to
as the fraction of the total void volume with respect to the volume of the catalyst
grain. The accessibility of the active sites requires pores within the solid that allow
molecular transport. Pores are formed by voids between agglomerated or aggregated
primary catalyst particles that represent the catalyst grains. Larger pores are determined by the packing of these catalyst grains. In addition, many catalysts possess
regular or irregular pore patterns within amorphous catalyst particles or structural
porosity of primary crystallites that means that the pore system is spanned by the
atomic arrangements in the crystal structure of the solid, like in zeolites or metalorganic frameworks [34]. The intraparticular pores involve pore diameters in subnanometer or nanometer scale. Generally, the pores within a catalyst are classied into
macropores with d > 50 nm, mesopores with 2 nm d 50 nm, and micropores with
d < 2 nm. Figure 4.2.4 shows a transmission electron microscopy (TEM) image of
mesostructured amorphous silica SBA-15 that comprises a regular, hexagonal pore
structure with pore diameters of approximately 7 nm for the present example.
Pore size, pore-size distribution, and shape of the pores constitute the texture of a
catalyst.
In addition to chemical composition, particle morphology, and texture, the preparation of industrially relevant catalysts requires the consideration of the process
conditions at an early stage of the catalyst development because the macroscopic
shape of a catalyst body in the micro- to millimeter scale depends on the reactor
operation (Figure 4.2.1). Generally, shaping of catalysts for chemical synthesis
5 nm
282
in the gas or liquid phase is performed to improve control over the following
parameters:
Mechanical stability
Thermal conductivity
Pressure drop in xed-bed reactors
Mass transport
In xed-bed reactors, the catalyst is situated within heated reactor tubes. Size, shape,
and porosity of the catalyst bodies inuence the pressure drop along the catalyst
bed, the efciency of the radial heat transfer, internal or external mass-transport limitations, and the packing density. All these factors have impact on the kinetics of the
chemical reaction and nally on the space-time yield of the product (gproduct gcat1 h1)
that can be achieved in the process. Sufcient mechanical and thermal stability is necessary to avoid attrition and crumbling while charging the reactor and crushing
under the weight of the catalyst bed and due to thermal stress under operation.
For example, the selective oxidation of propene to acrolein and acrylic acid is performed in tubular xed-bed reactors with an inner diameter of 2025 mm and a
length of 57 m. Tens of thousands of these reactors are bundled and placed within
a molten salt bath as the heat-transfer medium. Common technologies to produce
granular catalyst particles with diameters between 1 and 50 mm for xed-bed and
moving-bed applications are pelletizing, granulation, or extrusion [35]. In eggshell
catalysts, the active mass is coated as a thin layer in the micrometer range on the surface of an inert core usually composed of high-temperature ceramic. The composition of such a catalyst is illustrated in Figure 4.2.5, which shows scanning electron
microscopy (SEM) images of a mixed-oxide catalyst that has been coated on steatite
spheres with a diameter of approximately 5 mm (Fig. 4.2.5, left). The primary, crystalline catalyst particles consist of needles, which are aggregated and/or agglomerated in spheres that have been formed as a result of the spray-drying process
applied for drying after precursor precipitation (Fig. 4.2.5, right). These spheres
Coated carrier
Catalyst
embedded in
the binder
1 mm
30 m
Agglomerates and
aggregates of primary
particles
3 m
Figure 4.2.5 Cross section of steatite spheres, coated with a mixed oxide having primary,
283
with a diameter of a few micrometers are embedded in a porous matrix of an inorganic binder, resulting in a thin layer of about 100 m covering the steatite sphere
(Fig. 4.2.5, center). Monoliths [36, 37], foams [38], and gauzes [39, 40] are other examples of shaped catalysts applied in xed-bed reactors. Catalysts for uidized-bed
or slurry applications consist of microspheroidal particles with diameters between
20 and 100 m. These particles are subjected to strong mechanical stress within
the uidized bed. Attrition resistance is achieved either by embedding the catalyst
particles homogeneously within a mechanically stable and porous matrix of an inorganic binder, which is composed of amorphous alumina, amorphous silica-alumina,
or clays (e.g. kaolin, bentonite), or covering the active mass with a porous, protective
layer of the binder (egg-yolk catalyst). Catalysts for uid catalytic cracking of petroleum fractions are generally composed of zeolite crystallites embedded within a mixture
of amorphous alumosilicates and clays. Another prominent example is promoted
bismuth molybdate for ammoxidation of propene, which is protected by a silica layer.
The spray-drying technique is based on rapid evaporation of water from droplets,
which are formed by feeding an aqueous sol, gel, or suspension through (e.g. in the
simplest construction) a nozzle together with a ow of hot air resulting in spherical
catalyst particles of 10100 m diameter. A spray of droplets can also be generated
by discharging the liquid at high speed into hot air from a rotating wheel or disk. The
shape of the particles and the particle-size distribution depends on the operation
conditions and the characteristics of the liquid and is generally inhomogeneous.
Figure 4.2.6 shows an image of spherical mixed-oxide particles obtained by spraydraying a diluted aqueous solution prepared by dissolution of Mo, V, Te, and Nb
oxides in oxalic acid.
2 m
oxide.
284
The mechanical stress applied during shaping of solid catalysts can originate
changes in the specic surface area and pore structure. Stabilization techniques [41]
or the direct synthesis of shaped catalyst bodies [41, 42] are therefore benecial, in
particular, with regard to shaping strategies for meso- and microporous catalysts.
The usage of formulation additives in catalyst shaping technologies has a strong feedback to the accessibility and nature of active sites [43]. Active sites might be modied
or deactivated by the interaction with solvents, lubricants, binders, or carriers
through leaching, coverage, or chemical reaction. As an example, electron micrographs of carbon nanotubes (CNTs) that have been subjected to shaping by extrusion
applying two different binders and procedures are shown in Figure 4.2.7. The structures of the CNTs remain intact in both cases. However, the arrows in the lower two
images indicate that residues of the binder partially cover the CNTs shaped using the
10 nm
10 nm
10 nm
5 nm
Figure 4.2.7 High-resolution TEM images of CNT extrudates obtained by two different tech-
niques (method I, upper part, A and B; method II, lower part, A and B).
285
second procedure, which will have an effect on the accessibility of surface sites for the
reactants.
In addition to the requirements with respect to size, shape, and mechanical stability, the nature of the active phase also has to be adopted when the same catalyst
is applied in different reactor concepts mainly due to differing process conditions.
Vanadium phosphorous oxide composed of the vanadyl pyrophosphate phase
(VO)2P2O7 is an excellent catalyst for selective oxidation of n-butane to maleic anhydride [4447]. This type of catalyst has been operated in, for example, xed-bed reactors and uidized-bed-riser reactors [48]. In the different reactor types, different
feedstock is applied, the feed being more rich in n-butane (i.e. more reducible) in
the riser-reactor technology, which requires different catalyst characteristics [49].
The transport of electrons, protons, and gas are crucial design elements in the optimization of electrochemical reactions in electrodes [50]. The same holds for the
transport of charge carriers in batteries [25]. The main types of fuel cells are proton
exchange membrane fuel cells (PEMFCs), alkaline fuel cells, phosphoric acid fuel
cells, molten carbonate fuel cells, and solid oxide fuel cells (SOFCs). The latter
two types operate at high temperatures between 650C and 1,000C. Due to the
low operation temperature, the high power density, and the rapid response to changing loads, the proton exchange membrane (PEM) fuel cell is preferred for use in
transportation and many other applications. In PEM fuel cells, the transport processes comprise the electron transport, the diffusion of protons from the membrane
to the catalyst, and the diffusion of the reactant gases to the catalyst, which are controlled by the catalyst structure, the structure of the gas diffusion layer, the design of
the electrodes, and the overall fuel cell design [51].
Supported catalysts
Deposition precipitation
Sol-gel techniques
Solvothermal synthesis
Ship-in-bottle synthesis
Heterogenization of complexes and enzymes
286
the catalyst is carrier free or composed of nanostructured metal oxides or metal particles supported on a carrier. The support is not always inert, but contributes at
least due to its specic interaction with the guest species signicantly to catalysis
if not by own additional functionalities.
In catalyst manufacture, generally, the bulk material or support is prepared rst,
which is then subjected to shaping (Figure 4.2.3). If required, the highly dispersed
metal or metal-oxide species are deposited on the surface of the shaped catalyst
bodies, for instance, by impregnation, ion exchange, deposition precipitation, grafting, CVD or atomic layer deposition (ALD). The most frequently applied support
materials in catalysis are alumina, silica, amorphous or nanoporous alumosilicates,
and carbon. In battery materials [52] and PEMFCs or direct methanol fuel cells
[50, 51, 53], nanostructured carbon supports are most commonly used, while ceramics or metals are applied, for example, in SOFCs [54, 55]. In Table 4.2.2, the advantages and disadvantages of the most frequently applied techniques for synthesis
of nanostructured bulk catalysts or supports are summarized.
High-temperature solid-state (ceramic) techniques are advantageous when multicomponent single-phase catalysts with atomically homogeneous distribution of the
constituents are desired. In this method, solid reactants are intimately mixed in dry
form and subjected to a temperature treatment. The chemical reaction in solid state requires the diffusion of ions, atoms, or molecules within the bulk of the solid, on its
surface or via the gas phase. Since the diffusion coefcient of, for example, ions in
a solid is typically 1013 cm2 s1 at 300 K but increases exponentially with temperature, the reactions are performed at very high temperatures that are usually higher
than two-thirds of the melting temperature of the reactant with the highest melting
point in the mixture. One example is the synthesis of carbides from the elements,
like vanadium carbide that is synthesized above 1,673 K. Consequently, thermodynamically stable and dense phases with very small specic surface areas and crystallographic rather perfect particles with low abundance of defects are obtained. In the
laboratory, the exothermic and stoichiometric reactions are performed in crucibles or
Table 4.2.2 Assessment of techniques used in the synthesis of bulk catalysts and carriers.
Method
Advantage
Disadvantage
Precipitation and
coprecipitation
Synthesis of defect-rich
materials, easy to perform
Microemulsions
Particle-size control
Sol-gel techniques
Homogeneity of the
product
Solvothermal synthesis
Predictions difcult
Solid-state reactions
Flame pyrolysis
Particle-size control
287
ampoules in the presence of a gas phase or under vacuum. The control of the reaction
and product separation is difcult. The selectivity to a desired product can be optimized by systematic variation of the mixture of reactants, the reaction times, temperature, and the heating and cooling rates. Since the reaction starts at the interface
between particles and defects have a strong inuence on the diffusion, ball milling
enhances the rate of reaction. Examples for catalysts synthesized by ceramic methods
are nitrogen-conducting zirconium oxynitrides that have been investigated in the
decomposition of ammonia [56] or as nonplatinum fuel cell catalysts [57].
Synthesis at high temperatures is in particular benecial in view of thermal stability and controlled crystallinity. Supports and catalysts prepared in ames combine
these properties with purity, small particle-size distribution, controlled aggregate
size, well-dened surface chemistry, and high specic surface area. Flame technologies, such as ame hydrolysis have been used in the large-scale manufacture of binary
oxides, like fumed silica, alumina, and titania. The standard TiO2 photocatalyst P25
(Degussa) is manufactured using this method. In the synthesis of fumed silica, SiCl4,
air and hydrogen are reacted in a continuously operated ame. The continuous operation is one advantage in view of large-scale production. The size of the nonporous
primary nanoparticles can be controlled by the concentration and residence time of
the reactants and the ame temperature. Flame synthesis allows the mixing of various components giving access to multicomponent systems and nanocomposites [58].
CVD is an established method for the synthesis of CNTs [59, 60]. Supported transition metals that catalyze the growth of CNTs, such as iron, nickel, or cobalt, are
situated in a tubular reactor, and CNTs are grown at elevated temperature on the
surface of the catalyst particles by decomposition of a carbon-containing precursor.
The catalyst particles have to be removed by chemical treatment /washing in order to
obtain a metal-free nal product.
An important technique in the synthesis of support materials and catalysts is precipitation [61]. Binary oxides, like silica, and active aluminas are produced in large
scale using this technique and applied as catalyst supports. Precipitation comprises
phase separation of a solid from homogeneous solution induced by either physical
or chemical means. In catalyst preparation, usually a precipitating agent is added
that leads to high supersaturation by chemical reaction and often to the formation
of an amorphous solid. Disadvantages associated with this technique are the necessary product separation and large volumes of salt-containing solutions that are generated. Nevertheless, precipitation is the most important method used in synthesis of
catalyst supports. Coprecipitation is performed in the synthesis of multicomponent
materials. Prominent examples of technical catalysts manufactured by coprecipitation
are Cu/ZnO/Al2O3 catalysts for methanol synthesis and vanadium pyrophosphate
(VO)2P2O7 catalysts for selective oxidation of n-butane to maleic anhydride. Coprecipitation is applied, in particular, when a homogeneous distribution of different
catalyst components in the catalyst precursor is required.
In heterogeneous catalysis, a high catalytic activity per unit volume of catalyst is
generally the target, which accounts for the stabilization of very small particles under
reaction conditions. Stabilization against thermal sintering can be achieved by depositing the nanoparticles on the surface of a thermally stable, porous oxide as support.
When loadings of the active mass higher than 1020 wt% are required, deposition
288
precipitation is the method of choice [62]. Here, the precipitating agent is often generated in situ by hydrolysis of, for instance, urea CO(NH2)2 initiated by increasing
the temperature above 60C.
Despite the importance, fundamental understanding of precipitation processes
remains fragmentary, and the control of product properties is limited, which is due
to restricted experimental access to the key steps of solid formation and to the multitude of inuencing parameters. Precipitation basically involves four steps, including
the following:
1. Formation of a precursor (e.g. the hydrolysis product of a metal ion in solution,
which is able to condense and to form a solid phase)
2. Creation of nuclei through condensation of precursor molecules
3. Growth of the nuclei through the addition of matter, until the primary particle
stage is reached
4. Aging of the particles in suspension
In aqueous solution, the solvated aquo, hydroxo, or oxo complexes (Equation 1) of
metal ions undergo hydroxylation yielding hydroxy precursors (Equation 2) that are
subjected to inorganic polymerization reactions. Depending on the chemical nature
of the metal ion, either olation (Equation 3) or oxolation (Equation 4) happens. Olation results in the formation of hydroxo bridges between the central metal atoms involving polycations as intermediates, while oxo bridges are formed in oxolation via
polyanions.
H2O
[M(OH2)]z+
[M(OH)](z1)+ + H+aq
H2O
[M-O](z2)+ + 2 H+aq
aquo
hydroxo
oxo
pH < 7, z 4
pH > 7, z 4
z>4
[M-O] + H3O+
M-OH + H2O
[M-OH2]+ + HO
(1)
(2)
precursor
M-OH + M-OH2
M-OH-M + H2O
(3)
289
12
12
Oxolation of CuO
10
pH
pH
10
Precipitation of Cu2+
0.0
0.5
1.0
nCO2 /nCu2+
1.5
Precipitation of Zn2+
0.0
0.5
C
12
1.0
nCO2 /nCu2+
3
10
pH
8
6
Precipitation of Zn2+
4
2
Precipitation of Cu2+
0
0.0
0.5
1.0
1.5
2.0
Figure 4.2.8 Precipitation titration of (A) Cu, (B) Zn, and (C) Cu-Zn solutions.
1.5
290
291
Figure 4.2.9 SEM image of polycrystalline MoVTeNb oxide composed of the crystalline
phases M1 (red) and M2 (yellow). The crystals were colored articially for better visualization.
292
nuclei can be changed by the degree of supersaturation and the physicochemical conditions of the medium that affect the surface tension of the small particles via pH of
the solution, ionic strength, or the adsorption of ions [61].
Subsequent to nucleation, the particles grow by incorporation of precursor molecules, which might be the same complex, polynuclear species that were involved in
the nucleation or simple monomers. Growth and nucleation can also occur simultaneously. The rate of nucleation and particle growth are strongly temperature and
concentration dependent. The size distribution of the nal particles is determined
by the relative rates of nucleation and growth, which depends on the growing mechanism. The kinetic description of the formation of complex catalyst precursors applying classical nucleation and growth models often fails because simple assumptions,
like a spherical shape of the nucleus, are not always applicable in the case of complex
systems.
In precipitation of complex catalyst precursors, amorphous products are often
formed in a kinetically controlled manner. Aging the precipitate in the mother liquor
allows the system to reach increased thermodynamic stability. During aging, various
changes may happen, such as further particle growth by dissolution precipitation or
aggregation processes, crystallization or phase transformations, and changes in the
particle morphology. Figure 4.2.10 illustrates the modications that happen when coprecipitated Cu-Zn hydroxy carbonates are allowed to age in the mother liquor at a
temperature of 65C. The process was monitored in an automated laboratory reactor
(LabMax, Mettler-Toledo) using probes for measuring temperature, pH, conductivity (not shown), and turbidity. Crystallinity and average particle size were analyzed
250
70
Crystallization
Aging
68
66
200
100
64
50
62
10
20
30
40 50
2
60
70
80
60
30
BET: 32 m2/g
500 m
60
Aging time/min
90
120
BET: 57 m2/g
after pH drop
Intensity/a.u.
150
before pH drop
Intensity/a.u.
300
co-precipitation
Turbidity/a.u.
Temperature/C
350
pH
Particle diameter/nm
CuZn co-precipitation
10
500 m
20
30
40 50
2
60
70
80
293
294
Figure 4.2.11 Hydrogel obtained by acrylamide gelation that contains Mo, V, Te, and Nb
ions.
strength, and the pH and may last for minutes or days. The gel is a three-dimensional
(3-D) network of polymeric chains that captures the entire preparation vessel and is
characterized by elasticity when touched (Figure 4.2.11). Within the polymeric network of the gel, solvent molecules and by-products of the condensation reactions
are trapped. Before drying, the by-products are discharged by washing. Xerogels
are obtained when the solvent molecules are removed by evaporation. The solventfree products obtained after supercritical drying are called aerogels. The versatile
methods of sol-gel processing are widely applied today in the synthesis of binary oxides, multimetal mixed oxides, and composites; the preparation of coated monoliths;
and the synthesis of membranes and thin lms. Catalytically active molecular structures and nanoparticles may be entrapped during synthesis [91]. One disadvantage of
the method is the organic chemistry that is often involved and that may cause undesired contamination of the nal catalyst with carbon. In particular, with respect to
the synthesis of transition-metal oxides, residual carbon may interfere with the
redox chemistry of the metals giving rise to unwanted reduction and, in some cases,
also to unexpected results. In the sol-gel synthesis of mixed MoVTeNb oxides, for
instance, acrylamide and N,N-methylenebisacrylamide cross-linker were added to a
clear solution of Mo/Te/V/Nb oxalates/citrates, and the formation of a 3-D tangled
polymer network was initiated. Gelation occurs rapidly leading to the formation of
295
296
other hand, oxidizing agents, like H2O2, HClO4, HNO3, Cl2, Br2, and O2 can be
added to adjust the redox potential. Reactants that do not dissolve easily go into
solution by addition of mineralizers, which are substances that help to dissolve by
formation of complexes. Alkali metal hydroxides, for instance, are used as mineralizers in the synthesis of silicate materials. Specic structures, morphologies, and
porosities can be achieved by using structure-directing agents or templates. The
template-assisted hydrothermal synthesis of crystalline, microporous alumosilicates
(zeolites) is the classic example [95, 96, 105]. Larger pores in the mesopore range
are obtained by applying sol-gel-related techniques under solvothermal or normal
conditions, while templating relies on supramolecular arrays, in which micellar systems are formed using surfactants or block copolymers [105109]. Solvothermal
treatment can also be used to improve the crystallinity, to transform amorphous
material into crystalline material, to induce phase changes, or for spontaneous formation of porous oxides in the absence of templates [110]. Microwave radiation
can signicantly reduce the time that is required to complete the hydrothermal reaction [111115]. Combined sonochemical/hydrothermal treatments can accelerate
crystallization, increase the content of thermodynamically stable phases, and initiate
redox reactions [115117]. The synthesis of metal-oxide nanoparticles has also been
studied under supercritical conditions [89, 93, 118122].
Most of the investigations devoted to the synthesis of inorganic materials by solvothermal synthesis are restricted to the impact of the starting material or the reaction conditions on nal product properties. In order to have the ability to control the
physical and chemical properties of the synthesis products, the reaction paths and
kinetics of the inorganic reactions must be understood. This requires the application,
adaptation, or development of advanced in situ analysis tools. The analysis of corresponding spectroscopic data is generally a challenging task in view of the complexity
of many systems of practical relevance [123]. Sol-gel chemistry, nucleation, and crystallization have been extensively investigated in zeolite synthesis by sampling and in
situ techniques applying analytical methods like NMR, X-ray scattering, dynamic
light scattering, attenuated total reection infrared spectroscopy (ATR-IR), and electron microscopy [96]. NMR and mass spectroscopy have been shown to be useful to
study the early stages of polycondensation reactions, in particular in zeolite synthesis
[80, 124]. Crystallization processes are analyzed by X-ray and neutron diffraction
[125129]. Raman and ultraviolet-visible spectroscopy can probe molecular precursors and reaction intermediates [84, 123, 130, 131]. Especially with respect to the
synthesis of new materials for electrochemical application, such investigations
are seldom performed. However, a more detailed understanding of the reaction mechanisms is required for the rational development of next-generation materials with
advanced specications.
297
support [132]. Supported catalysts are prepared when the precursor, which contains
the catalytically active metal, is costly or when a monolayer of the active phase is
desired. In Table 4.2.1, some frequently used methods for the preparation of supported catalysts are summarized. In the following, the most important techniques
are described briey.
When the surface of a support bears charged ions, ion exchange can be employed
to coat the carrier with the active element. An equilibrium is established between an
ion inherently attached to the solid support and an ion in a diluted solution followed
by ltration and washing. Zeolites or clays are cation exchangers as a consequence of
their structure containing protons to compensate the charge of the framework, which
can be exchanged by positively charged ions. The same holds for hydrotalcites as
anion exchangers. Such structural ion exchangers possess a constant number of
exchangeable surface sites, irrespective of the pH of the aqueous solution in which
the ion exchange is performed. The surface of binary oxides, like silica, can be functionalized in order to achieve ion-exchange capacity [42, 133, 134]. In Figure 4.2.12,
an example is shown. Grafting of 3-aminopropyltrimethoxysilane on the surface SiOH (silanol) groups gives surface amino groups that can be converted into anchored
propylammoniumchloride species by reaction with hydrochloric acid resulting in a
surface that is capable of exchanging anions [135, 136]. Chloride can be exchanged,
for example, by vanadate or molybdate species resulting in a high dispersion of the
metal-oxide species on the surface of the support [136138].
Electrostatic adsorption rather than ion exchange is possible on the surface of
nonfunctionalized mineral oxides. In aqueous solution, the oxides are covered by hydroxyl groups, which undergo hydrolysis resulting in a pH-dependent adsorption
capacity based on electrostatic interactions. Depending on the pH of the aqueous
solution and the nature of the metal atom, the hydroxyl groups are either protonated
(positively charged) or deprotonated (negatively charged). The pH at which a particle
of a given oxide is not charged overall (i.e. the pH at which the number of positively
and negatively charged surface groups is equal) is called the point of zero charge
(PZC). At pH > PZC, the surface is negatively charged (hydroxyl groups are deprotonated). Such a surface absorbs cations by electrostatic interaction. PZC values of
common silicon oxides vary between 1.5 and 3 [139, 140]. This means that the surface
of silica is negatively charged over almost the entire pH range, and practically only
cations can be adsorbed. Equilibrium adsorption is the method of choice when small
amounts of expensive noble metals should be deposited onto the support.
In the case of higher loadings, impregnation is used. In this technique, the pores of
the support are lled with a solution that contains a precursor of the active species,
OR
Si
OH
Si
Si
OH
Si
Si
OH 1. (H3CO)3Si(CH2)3NH2
Si
Si
Si
Si
OH
OH
Si
O
O
Si
OH
Si
2. HCl
OR
Si(CH2)3NH3+ Cl
Si(CH2)3NH3+ Cl
O
Si(CH2)3NH3+ Cl
3. Mo7O246
Si
Si
Si
Si
Si
O
O
Si
Si(CH2)3NH3+
Si(CH2)3NH3+
O
Mo7O246
Si(CH2)3NH3+
MoOx/SBA-15
298
and the solvent is then removed by drying without previous washing. Two different
procedures are frequently used. In the so-called dry-impregnation or incipientwetness technique, the volume of the precursor solution corresponds to the volume
of the pores. The solution lls the pores due to capillary forces. To reduce the
mechanical stress, which the escaping gas bubbles exert on the pore walls, incipient
wetness is either performed in a vacuum, or surfactants are added to the solution. In
an alternative technique (wet impregnation), the pores are already lled with the solvent, and the driving force for the diffusion of the precursor into the pores is the concentration gradient. When the interaction between the surface and the precursor is
weak, the process can last hours. The strength of the interaction between the precursor and the support surface is affected by numerous parameters, including the pH of
the solution, the concentration, and the temperature. All these factors determine the
charging of the surface on the one hand and the speciation of the precursor on the
other hand, which is, for example, important for oxo species of transition metals
(like molybdates, chromates, or vanadates) that can undergo pH and concentrationdependent condensation reactions. Other important factors that affect the precursorsupport interaction are the nature of ligands or the presence of complexing agents.
Competitive adsorption has to be taken into account when the impregnation of
more than one metal is intended. Furthermore, the type of the counterion has to be
considered in particular in view of oxidizing (e.g. NO3) or reducing (e.g. NH4+)
properties during the subsequent calcination that may change the nature of the active
species on the surface of the nal catalyst.
An improved control with respect to coverage of the surface with an active species
is possible when the precursor is grafted by specic chemical reactions with surface
functional groups, such as surface hydroxyl groups, forming a strong chemical
bond [141143]. The grafting of titanium and vanadium on silica is shown as an
example in Equations 5 and 6, respectively. The geometry of the surface species
can be controlled via the surface density of the functional groups, which can be adjusted in the synthesis of the support. The average density of, for instance, surface hydroxyls can be estimated by using analytical methods, like thermal analysis, infrared
spectroscopy, or NMR spectroscopy. The determination of the amount of residual
functional groups on the surface by spectroscopic techniques allows the measurement
of adsorption isotherms and enables conclusions with respect to the surface structure
of the grafted complex that can be deduced from the stoichiometry of the reaction. The
molecular structure will, however, change during the subsequent calcination that is
necessary to remove residual ligands.
OCH(CH3)2
SiO
SiOH + [Ti(OCH(CH3)2)4]
Ti
(CH3)2CHO
SiO
2 SiOH + [O=V(OCH(CH3)2)3]
(5)
+ 2 CH(CH3)2OH
(6)
OCH(CH3)2
V
SiO
+ CH(CH3)2OH
OCH(CH3)2
299
The as-prepared catalyst precursors are usually not yet active for the desired catalytic reaction. The step in catalyst preparation that transforms the precursor into an
active material is called activation [144]. Sometimes, in particular when the activation is performed in the feed of the reactants of the target reaction, the term formation is also used. The activation is a crucial process in catalyst synthesis and strongly
inuences the catalyst performance and stability. Generally, the activation involves a
thermal treatment applying temperatures higher than the reaction temperature in
catalysis. Calcination represents a heat treatment in oxygen-containing gas atmosphere that is generally applied in the activation of oxides and supported oxides.
The purpose of calcination is the transformation of carbonates or hydroxides in precipitated bulk-catalyst precursors into oxides or the oxidative or thermal removal of
counterions in oxide-supported catalysts. The reaction during calcination of supported species may be restricted to the interface or may involve diffusion-controlled
solid-state reactions that lead to the formation of mixed phases, like the formation of
spinels in alumina-supported catalysts that contain transition metals like Ni or Co
(e.g. NiAl2O4 or CoAl2O4), or orthovanadates Mg3V2O8 in MgO-supported V catalysts. Another typical activation procedure is the reduction of metal-oxide species
to highly dispersed metal nanoparticles on the surface of a support for hydrogenation
or electrochemical reactions [9, 145, 146], or highly defective metal nanoparticles
100
100
60
(mL/
80
flow
80
40
min)
20
nt (%
onte
M1 c
90
70
0
60
600
610
620
630
640 650
660
670
Figure 4.2.13 Content (wt%) of the crystalline M1 phase obtained by activation of a single,
amorphous MoVTeNb-oxide precursor in argon in a rotating oven at different temperatures
and argon ow rates.
300
The nanostructuring of the surface is essential in many applications, but nanostructured surfaces are not necessarily required in every catalytic reaction.
The phenomenological knowledge in catalyst synthesis is immense, in particular if
the secrets kept in industry are included. Basic concepts, however, are rare because
this implies fundamental understanding of the elemental steps. To establish the
design of catalysts on a rational basis, systematic investigation of catalyst synthesis
applying analytical tools including in situ studies is required to a much greater
extent. This also includes scaling up of the synthesis batches and the processes of catalyst shaping and activation. The latter two aspects have been considered with less
attention so far, in particular in academic research.
Also at the laboratory scale, the synthesis of reasonable quantities is necessary to
ensure comprehensive characterization of the nal product. In modern laboratories,
4.2.6 References
301
4.2.5 Acknowledgments
All coworkers of the Department of Inorganic Chemistry at the Fritz-Haber-Institut
der Max-Planck-Gesellschaft and our collaboration partners (see http://www.fhiberlin.mpg.de/acnew/welcome.epl) are greatly acknowledged for their contributions,
which have been included in this chapter.
4.2.6 References
1. Cavani F, Teles JH. Sustainability in catalytic oxidation: An alternative approach or a
structural evolution? ChemSusChem. 2009;2(6):508534.
2. Centi G, Perathoner S. Catalysis: Role and challenges for a sustainable energy. Topics in
Catalysis. 2009;52(8):948961.
3. Schlgl R. The role of chemistry in the energy challenge. ChemSusChem. 2010;3(2):
209222.
4. Somorjai GA, Park JY. Molecular factors of catalytic selectivity. Angewandte Chemie
International Edition. 2008;47(48):92129228.
302
5. Britovsek GJP, Gibson VC, Wass DF. The search for new-generation olen polymerization catalysts: Life beyond metallocenes. Angewandte Chemie International Edition.
1999;38(4):428447.
6. Resconi L, Cavallo L, Fait A, Piemontesi F. Selectivity in propene polymerization with
metallocene catalysts. Chemical Reviews. 2000;100(4):12531345.
7. Grubbs RH. Olen metathesis. Tetrahedron. 2004;60(34):71177140.
8. Schlgl R, Hess C. Characteristics of selective oxidation reactions. In: Hess C, Schlgl R,
editors. Nanostructured catalysts: Selective oxidation reactions. Cambridge: RSC Nanoscience & Nanotechnology; 2011. p. 355397.
9. Teschner D, Borsodi J, Wootsch A, Revay Z, Hvecker M, Knop-Gericke A, Jackson SD,
Schlgl R. The roles of subsurface carbon and hydrogen in palladium-catalyzed alkyne
hydrogenation. Science. 2008;320(5872):8689.
10. Perego C, Ingallina P. Recent advances in the industrial alkylation of aromatics: new
catalysts and new processes. Catalysis Today. 2002;73(12):322.
11. Klier K. Methanol synthesis. Advances in Catalysis. 1982;31:243313.
12. Kung HH. Methanol synthesis. Catalysis Reviews-Science and Engineering. 1980;22(2):
235259.
13. Kasatkin I, Kurr P, Kniep B, Trunschke A, Schlgl R. Role of lattice strain and defects
in copper particles on the activity of Cu/ZnO/Al2O3 catalysts for methanol synthesis.
Angewandte Chemie, International Edition. 2007;46(38):73247327.
14. Ertl G. Reactions at surfaces: From atoms to complexity (Nobel Lecture). Angewandte
Chemie International Edition. 2008;47(19):35243535.
15. Climent MJ, Corma A, Iborra S. Heterogeneous catalysts for the one-pot synthesis of
chemicals and ne chemicals. Chemical Reviews. 2010;111(2):10721133.
16. Granger P, Parvulescu VI. Catalytic NOx abatement systems for mobile sources: From threeway to lean burn after-treatment technologies. Chemical Reviews. 2011;111(5):31553207.
17. Speight JG. New approaches to hydroprocessing. Catalysis Today. 2004;98(12):5560.
18. Rinaldi R, Schth F. Design of solid catalysts for the conversion of biomass. Energy &
Environmental Science. 2009;2(6):610626.
19. Corma A, Iborra S, Velty A. Chemical routes for the transformation of biomass into
chemicals. Chemical Reviews. 2007;107(6):24112502.
20. Chheda JN, Huber GW, Dumesic JA. Liquid-phase catalytic processing of biomassderived oxygenated hydrocarbons to fuels and chemicals. Angewandte Chemie International Edition. 2007;46(38):71647183.
21. Quadrelli EA, Centi G, Duplan J-L, Perathoner S. Carbon dioxide recycling: Emerging
large-scale technologies with industrial potential. ChemSusChem. 2011;4(9):11941215.
22. Arakawa H, Aresta M, Armor JN, Barteau MA, Beckman EJ, Bell AT, Bercaw JE,
Creutz C, Dinjus E, Dixon DA, Domen K, DuBois DL, Eckert J, Fujita E, Gibson DH,
Goddard WA, Goodman DW, Keller J, Kubas GJ, Kung HH, Lyons JE, Manzer LE,
Marks TJ, Morokuma K, Nicholas KM, Periana R, Que L, Rostrup-Nielson J, Sachtler
WMH, Schmidt LD, Sen A, Somorjai GA, Stair PC, Stults BR, Tumas W. Catalysis research
of relevance to carbon management: Progress, challenges, and opportunities. Chemical
Reviews. 2001;101(4):953996.
23. Su DS, Schlgl R. Nanostructured carbon and carbon nanocomposites for electrochemical
energy storage applications. ChemSusChem. 2010;3(2):136168.
24. Schth F. Chemical compounds for energy storage. Chemie Ingenieur Technik. 2011;83
(11):19841993.
25. Goodenough JB, Kim Y. Challenges for rechargeable Li batteries. Chemistry of Materials.
2010;22(3):587603.
4.2.6 References
303
26. Whittingham MS. Lithium batteries and cathode materials. Chemical Reviews. 2004;
104(10):42714302.
27. Artero V, Chavarot-Kerlidou M, Fontecave M. Splitting water with cobalt. Angewandte
Chemie International Edition. 2011;50(32):72387266.
28. Walter MG, Warren EL, McKone JR, Boettcher SW, Mi Q, Santori EA, Lewis NS. Solar
Water splitting cells. Chemical Reviews. 2010;110(11):64466473.
29. Maeda K, Domen K. Photocatalytic water splitting: Recent progress and future challenges. The Journal of Physical Chemistry Letters. 2010;1(18):26552661.
30. Chen X, Shen S, Guo L, Mao SS. Semiconductor-based photocatalytic hydrogen generation. Chemical Reviews. 2010;110(11):65036570.
31. Balzani V, Credi A, Venturi M. Photochemical conversion of solar energy. ChemSusChem.
2008;1(12):2658.
32. Gallei EF, Hesse M, Schwab E. Development of Industrial Catalysts. In: Ertl G, Knzinger
H, Schth F, Weitkamp J, editors. Handbook of heterogeneous catalysis. Weinheim: WileyVCH Verlag GmbH & Co. KGaA; 2008. p. 5766.
33. Guerrini E, Trasatti S. Recent developments in understanding factors of electrocatalysis.
Russian Journal of Electrochemistry. 2006;42(10):10171025.
34. Cheetham AK, Frey G, Loiseau T. Open-framework inorganic materials. Angewandte
Chemie International Edition. 1999;38(22):32683292.
35. Schth F, Hesse M. Catalyst Forming. In: Ertl G, Knzinger H, Schth F, Weitkamp J,
editors. Handbook of heterogeneous catalysis. Weinheim: Wiley-VCH Verlag GmbH &
Co. KGaA; 2008. p. 676699.
36. Heck RM, Gulati S, Farrauto RJ. The application of monoliths for gas phase catalytic
reactions. Chemical Engineering Journal. 2001;82(13):149156.
37. Avila P, Montes M, Mir EE. Monolithic reactors for environmental applications: A
review on preparation technologies. Chemical Engineering Journal. 2005;109(13):1136.
38. Twigg MV, Richardson JT. Fundamentals and applications of structured ceramic foam
catalysts. Industrial & Engineering Chemistry Research. 2007;46(12):41664177.
39. Sadykov VA, Isupova LA, Zolotarskii IA, Bobrova LN, Noskov AS, Parmon VN,
Brushtein EA, Telyatnikova TV, Chernyshev VI, Lunin VV. Oxide catalysts for ammonia
oxidation in nitric acid production: properties and perspectives. Applied Catalysis A:
General. 2000;204(1):5987.
40. Goetsch DA, Schmidt LD. Microsecond catalytic partial oxidation of alkanes. Science.
1996;271(5255):15601562.
41. Herbert R, Wang D, Schomcker R, Schlgl R, Hess C. Stabilization of mesoporous silica
SBA-15 by surface functionalization. ChemPhysChem. 2009;10(13):22302233.
42. Schth F. Engineered porous catalytic materials. Annual Review of Materials Research.
2005;35(1):209238.
43. Baldovino-Medrano VG, Farin B, Gaigneaux EM. Establishing the role of graphite as a
shaping agent of vanadium-aluminum mixed (hydr)oxides and their physicochemical
properties and catalytic functionalities. ACS Catalysis. 2012;2(3):322336.
44. Bluhm H, Hvecker M, Kleimenov E, Knop-Gericke A, Liskowski A, Schlgl R, Su DS.
In situ surface analysis in selective oxidation catalysis: n-butane conversion over VPP.
Topics in Catalysis. 2003;23(1):99107.
45. Hvecker M, Mayer RW, Knop-Gericke A, Bluhm H, Kleimenov E, Liskowski A, Su D,
Follath R, Requejo FG, Ogletree DF, Salmeron M, Lopez-Sanchez JA, Bartley JK,
Hutchings GJ, Schlgl R. In situ investigation of the nature of the active surface of a vanadyl pyrophosphate catalyst during n-butane oxidation to maleic anhydride. The Journal
of Physical Chemistry B. 2003;107(19):45874596.
304
46. Hutchings GJ. Vanadium phosphate: a new look at the active components of catalysts for
the oxidation of butane to maleic anhydride. J. Mater. Chem. 2004;14(23):33853395.
47. Carreon MA, Guliants VV. Chapter 6 selective oxidation of n-butane over vanadiumphosphorous oxide. Nanostructured Catalysts: Selective Oxidations The Royal Society
of Chemistry; 2011. p. 141168.
48. Lintz H-G, Reitzmann A. Alternative reaction engineering concepts in partial oxidations
on oxidic catalysts. Catalysis Reviews: Science & Engineering. 2007;49(1):132.
49. Centi G. Vanadyl pyrophosphate A critical overview. Catalysis Today. 1993;16(1):
526.
50. Litster S, McLean G. PEM fuel cell electrodes. Journal of Power Sources. 2004;130(12):
6176.
51. Mehta V, Cooper JS. Review and analysis of PEM fuel cell design and manufacturing.
Journal of Power Sources. 2003;114(1):3253.
52. de las Casas C, Li W. A review of application of carbon nanotubes for lithium ion battery
anode material. Journal of Power Sources. 2012;208(0):7485.
53. Sharma S, Pollet BG. Support materials for PEMFC and DMFC electrocatalysts A
review. Journal of Power Sources. 2012;208(0):96119.
54. Sun C, Stimming U. Recent anode advances in solid oxide fuel cells. Journal of Power
Sources. 2007;171(2):247260.
55. Tucker MC. Progress in metal-supported solid oxide fuel cells: A review. Journal of Power
Sources. 2010;195(15):45704582.
56. Soerijanto H, Rdel C, Wild U, Lerch M, Schomcker R, Schlgl R, Ressler T. The
impact of nitrogen mobility on the activity of zirconium oxynitride catalysts for ammonia
decomposition. Journal of Catalysis. 2007;250(1):1924.
57. Wang B. Recent development of non-platinum catalysts for oxygen reduction reaction.
Journal of Power Sources. 2005;152(0):115.
58. Strobel R, Baiker A, Pratsinis SE. Aerosol ame synthesis of catalysts. Advanced Powder
Technology. 2006;17:457480.
59. Tessonnier J-P, Rosenthal D, Hansen TW, Hess C, Schuster ME, Blume R, Girgsdies F,
Pfnder N, Timpe O, Su DS, Schlgl R. Analysis of the structure and chemical properties
of some commercial carbon nanostructures. Carbon. 2009;47(7):17791798.
60. Tessonnier J-P, Su DS. Recent progress on the growth mechanism of carbon nanotubes:
A review. ChemSusChem. 2011;4(7):824847.
61. Schth F, Hesse M, Unger KK. Precipitation and coprecipitation. Handbook of Heterogeneous Catalysis. Weinheim; Wiley-VCH Verlag GmbH & Co. KGaA; 2008. p. 100119.
62. Geus JW, van Dillen AJ. Preparation of supported catalysts by deposition precipitation.
Handbook of Heterogeneous Catalysis. Weinheim; Wiley-VCH Verlag GmbH & Co.
KGaA; 2008. p. 428467.
63. Jolivet J-P. Metal Oxide Chemistry and synthesis From solution to solid state; John
Wiley & Sons Ltd: Chichester, 2000.
64. Cavani F, Trir F, Vaccari A. Hydrotalcite-type anionic clays: Preparation, properties
and applications. Catalysis Today. 1991;11(2):173301.
65. Behrens M, Brennecke D, Girgsdies F, Kissner S, Trunschke A, Nasrudin N, Zakaria S,
Idris NF, Hamid SBA, Kniep B, Fischer R, Busser W, Muhler M, Schlgl R. Understanding the complexity of a catalyst synthesis: Co-precipitation of mixed Cu,Zn,Al hydroxycarbonate precursors for Cu/ZnO/Al2O3 catalysts investigated by titration experiments.
Applied Catalysis A: General. 2011;392(12):93102.
66. Baltes C, Vukojevic S, Schth F. Correlations between synthesis, precursor, and catalyst
structure and activity of a large set of CuO/ZnO/Al2O3 catalysts for methanol synthesis.
Journal of Catalysis. 2008;258(2):334344.
4.2.6 References
305
67. Li X, Buttrey D, Blom D, Vogt T. Improvement of the Structural Model for the M1 Phase
MoVNbTeO Propane (Amm)oxidation Catalyst. Topics in Catalysis. 2011;54(10):
614626.
68. DeSanto P, Jr., Buttrey DJ, Grasselli RK, Lugmair CG, Volpe AF, Jr., Toby BH, Vogt T.
Structural aspects of the M1 and M2 phases in MoVNbTeO propane ammoxidation
catalysts. Zeitschrift fuer Kristallographie. 2004;219(3):152165.
69. Baca M, Pigamo A, Dubois JL, Millet JMM. Propane oxidation on MoVTeNbO mixed
oxide catalysts: Study of the phase composition of active and selective catalysts. Topics in
Catalysis. 2003;23(14):3946.
70. Grasselli RK, Andersson A, Buttrey DJ, Burrington JD, Lugmair CG, Volpe AF. The
role of site isolation and phase cooperation in selective light parafn (amm)oxidation
catalysis. In: Abstracts of Papers, 228th ACS National Meeting, Philadelphia, PA, United
States, August 2226, 2004, p COLL-245.
71. Trunschke A. Propane selective oxidation to acrylic acid. In: Hess C, Schlgl R, editors.
Nanostructured catalysts: Selective oxidation reactions. Cambridge: RSC Nanoscience &
Nanotechnology; 2011. p. 5695.
72. Lin MM. Complex metal oxide catalysts for selective oxidation of propane and derivatives. II. The relationship among catalyst preparation, structure and catalytic properties.
Applied Catalysis, A: General. 2003;250(2):287303.
73. Lin MM. Complex metal-oxide catalysts for selective oxidation of propane and derivatives. I. Catalysts preparation and application in propane selective oxidation to acrylic
acid. Applied Catalysis, A: General. 2003;250(2):305318.
74. Oliver JM, Lopez Nieto JM, Botella P, Mifsud A. The effect of pH on structural and
catalytic properties of MoVTeNbO catalysts. Applied Catalysis, A: General. 2004;257(1):
6776.
75. Tu X, Furuta N, Sumida Y, Takahashi M, Niiduma H. A new approach to the preparation of MoVNbTe mixed oxide catalysts for the oxidation of propane to acrylic acid.
Catalysis Today. 2006;117(13):259264.
76. Popova GY, Andrushkevich TV, Aleshina GI, Plyasova LM, Khramov MI. Effect of oxalic acid content and medium of thermal treatment on physicochemical and catalytic properties of MoVTeNb oxide catalysts in propane ammoxidation. Applied Catalysis A:
General. 2007;328(2):195200.
77. Zhu Y, Lu W, Li H, Wan H. Selective modication of surface and bulk V5+/V4+ ratios
and its effects on the catalytic performance of MoVTeO catalysts. Journal of Catalysis.
2007;246(2):382389.
78. Ivars F, Solsona B, Hernandez S, Lopez Nieto JM. Inuence of gel composition in the
synthesis of MoVTeNb catalysts over their catalytic performance in partial propane and
propylene oxidation. Catalysis Today. 2010;149(34):260266.
79. Beato P, Blume A, Girgsdies F, Jentoft RE, Schlgl R, Timpe O, Trunschke A, Weinberg
G, Basher Q, Hamid FA, Hamid SBA, Omar E, Mohd Salim L. Analysis of structural
transformations during the synthesis of a MoVTeNb mixed oxide catalyst. Applied Catalysis, A: General. 2006;307(1):137147.
80. Maksimovskaya RI, Bondareva VM, Aleshina GI. NMR spectroscopic studies of interactions in solution during the synthesis of MoVTeNb oxide catalysts. European Journal of
Inorganic Chemistry. 2008;2008(31):49064914.
81. Popova GY, Andrushkevich TV, Dovlitova LS, Aleshina GA, Chesalov YA, Ishenko AV,
Ishenko EV, Plyasova LM, Malakhov VV, Khramov MI. The investigation of chemical and phase composition of solid precursor of MoVTeNb oxide catalyst and its transformation during the thermal treatment. Applied Catalysis A: General. 2009;353(2):
249257.
306
82. Girgsdies F, Schlgl R, Trunschke A. In-situ X-ray diffraction study of phase crystallization from an amorphous MoVTeNb oxide catalyst precursor. Catalysis Communications. 2012;18(0):6062.
83. Ueda W, Vitry D, Kato T, Watanabe N, Endo Y. Key aspects of crystalline Mo-V-Obased catalysts active in the selective oxidation of propane. Research on Chemical Intermediates. 2006;32(34):217233.
84. Celaya Sanz A, Hansen TW, Girgsdies F, Timpe O, Rdel E, Ressler T, Trunschke A,
Schlgl R. Preparation of phase-pure M1 MoVTeNb oxide catalysts by hydrothermal
synthesis-inuence of reaction parameters on structure and morphology. Topics in Catalysis. 2008;50(14):1932.
85. Kolenko YV, Zhang W, Naumann dAlnoncourt R, Girgsdies F, Hansen TW,
Wolfram T, Schlgl R, Trunschke A. Synthesis of MoVTeNb oxide catalysts with tunable particle dimensions. ChemCatChem. 2011;3(10):15971606.
86. Perego C, Villa P. Catalyst preparation methods. Catalysis Today. 1997;34(34):281305.
87. Landau MV. SolGel Process. In: Ertl G, Knzinger H, Schth F, Weitkamp J, editors.
Handbook of heterogeneous catalysis. Weinheim: Wiley-VCH Verlag GmbH & Co.
KGaA; 2008. p. 119160.
88. Hench LL, West JK. The sol-gel process. Chemical Reviews. 1990;90(1):3372.
89. Sui R, Charpentier P. Synthesis of metal oxide nanostructures by direct Sol-Gel chemistry
in supercritical uids. Chemical Reviews. 2012;112(6):30573082.
90. Niederberger M. Nonaqueous Sol-Gel routes to metal oxide nanoparticles. Accounts of
Chemical Research. 2007;40(9):793800.
91. Arnal PM, Comotti M, Schth F. High-temperature-stable catalysts by hollow sphere
encapsulation. Angewandte Chemie International Edition. 2006;45(48):82248227.
92. Kolenko YV, Amakawa K, Naumann dAlnoncourt R, Girgsdies F, Weinberg G,
Schlgl R, Trunschke A. Unusual phase evolution in MoVTeNb oxide catalysts prepared
by a novel acrylamide-gelation route. ChemCatChem. 2012;4(4):495503.
93. Aymonier C, Loppinet-Serani A, Reveron H, Garrabos Y, Cansell F. Review of supercritical uids in inorganic materials science. Journal of Supercritical Fluids. 2006;38(2):
242251.
94. Cheetham AK, Rao CNR, Feller RK. Structural diversity and chemical trends in hybrid
inorganic-organic framework materials. Chemical Communications. 2006;42(46):47804795.
95. Cundy CS, Cox PA. The hydrothermal synthesis of zeolites: History and development
from the earliest days to the present time. Chemical Reviews. 2003;103(3):663702.
96. Cundy CS, Cox PA. The hydrothermal synthesis of zeolites: Precursors, intermediates
and reaction mechanism. Microporous and Mesoporous Materials. 2005;82(12):178.
97. Cushing BL, Kolesnichenko VL, OConnor CJ. Recent advances in the liquid-phase
syntheses of inorganic nanoparticles. Chemical Reviews. 2004;104(9):38933946.
98. Demazeau G. Solvothermal reactions: an original route for the synthesis of novel materials. Journal of Materials Science. 2008;43(7):21042114.
99. Parnham ER, Morris RE. Ionothermal synthesis of zeolites, metal-organic frameworks,
and inorganic-organic hybrids. Accounts of Chemical Research. 2007;40:10051013.
100. Wakihara T, Okubo T. Hydrothermal synthesis and characterization of zeolites. Chemistry Letters. 2005;34(3):276281.
101. Kirschhock CEA, Feijen EJP, Jacobs PA, Martens JA. Hydrothermal Zeolite Synthesis.
In: Ertl G, Knzinger H, Schth F, Weitkamp J, editors. Handbook of heterogeneous
catalysis. Weinheim: Wiley-VCH Verlag GmbH & Co. KGaA; 2008. p. 160178.
102. Reddy CVS, Wicker SSA, Walker JEH, Williams QL, Kalluru RR. Vanadium oxide
nanorods for Li-Ion battery applications. Journal of The Electrochemical Society.
2008;155(8):A599-A602.
4.2.6 References
307
308
124. Schth F. Nucleation and crystallization of solids from solution. Current Opinion in
Solid State and Materials Science. 2001;5(5):389395.
125. Norby P. In-situ XRD as a tool to understanding zeolite crystallization. Current Opinion
in Colloid & Interface Science. 2006;11(23):118125.
126. Michailovski A, Kiebach R, Bensch W, Grunwaldt J-D, Baiker A, Komarneni S, Patzke
GR. Morphological and kinetic studies on hexagonal tungstates. Chemistry of Materials.
2006;19(2):185197.
127. Beale AM, Sankar G. In situ study of the formation of crystalline bismuth molybdate
materials under hydrothermal conditions. Chemistry of Materials. 2002;15(1):146153.
128. Beale AM, Sankar G. Understanding the crystallization of nanosized cobalt aluminate
spinel from ion-exchanged zeolites using combined in situ QEXAFS/XRD. Chemistry
of Materials. 2005;18(2):263272.
129. Kiebach R, Pienack N, Bensch W, Grunwaldt J-D, Michailovski A, Baiker A, Fox T,
Zhou Y, Patzke GR. Hydrothermal formation of W/Mo-Oxides: A multidisciplinary
study of growth and shape. Chemistry of Materials. 2008;20(9):30223033.
130. Kongmark C, Martis V, Rubbens A, Pirovano C, Lofberg A, Sankar G, Bordes-Richard
E, Vannier R-N, Van Beek W. Elucidating the genesis of Bi2MoO6 catalyst by combination of synchrotron radiation experiments and Raman scattering. Chemical Communications. 2009;45(32):48504852.
131. Sadakane M, Yamagata K, Kodato K, Endo K, Toriumi K, Ozawa Y, Ozeki T, Nagai
T, Matsui Y, Sakaguchi N, Pyrz WD, Buttrey DJ, Blom DA, Vogt T, Ueda W. Synthesis
of orthorhombic Mo-V-Sb oxide species by assembly of pentagonal Mo6O21 polyoxometalate building blocks. Angewandte Chemie International Edition. 2009;48(21):37823786.
132. Lambert JF, Che M. The molecular approach to supported catalysts synthesis: state of
the art and future challenges. Journal of Molecular Catalysis A: Chemical. 2000;162
(12):518.
133. Stein A, Melde BJ, Schroden RC. Hybrid inorganicorganic mesoporous silicates
nanoscopic reactors coming of age. Advanced Materials. 2000;12(19):14031419.
134. Hoffmann F, Cornelius M, Morell J, Frba M. Silica-based mesoporous organic
inorganic hybrid materials. Angewandte Chemie International Edition. 2006;45(20):3216
3251.
135. Carvalho WA, Wallau M, Schuchardt U. Iron and copper immobilised on mesoporous
MCM-41 molecular sieves as catalysts for the oxidation of cyclohexane. Journal of
Molecular Catalysis A: Chemical. 1999;144(1):9199.
136. Hess C, Wild U, Schlgl R. The mechanism for the controlled synthesis of highly dispersed vanadia supported on silica SBA-15. Microporous and Mesoporous Materials.
2006;95(13):339349.
137. Guo CS, Hermann K, Hvecker M, Thielemann JP, Kube P, Gregoriades LJ, Trunschke
A, Sauer J, Schlgl R. Structural analysis of silica-supported molybdena based on x-ray
spectroscopy: Quantum theory and experiment. The Journal of Physical Chemistry C.
2011;115(31):1544915458.
138. Grne P, Wolfram T, Pelzer K, Schlgl R, Trunschke A. Role of dispersion of vanadia
on SBA-15 in the oxidative dehydrogenation of propane. Catalysis Today. 2010;157(14):
137142.
139. Parks GA. The isoelectric points of solid oxides, solid hydroxides, and aqueous hydroxo
complex systems. Chemical Reviews. 1965;65(2):177198.
140. Parks GA, Bruyn PLd. The zero point of charge of oxides. The Journal of Physical
Chemistry. 1962;66(6):967973.
141. Bond GC, Tahir SF. Vanadium oxide monolayer catalysts Preparation, characterization
and catalytic activity. Applied Catalysis. 1991;71(1):131.
4.2.6 References
309
312
Length scale
Heat and
Mass transport
Ageing
- Macroscopic
- Atomic scale
information
Ignition/transient
processess
Adsorption and
Surface reactions
Morphological
Changes
Time scale
Figure 4.3.1 Catalyst characterization under reaction conditions spans all length scales from
real reactors to the atomic structure with the catalytically active site as well as many orders in
timescales considering long-term behavior in a real process and rapid changes of the structure
and extremely fast surface processes.
On the other hand, morphological changes can occur on the minute scale [8], or
transformations during activation of a catalyst (temperature-programmed reaction/
reduction/suldation), ignition of a reaction, or oscillations can even occur on the
subsecond timescale [11, 14, 15].
Hence, the task is not only to cover the different length scales from the atomic
scale to the macroscale but also to study the structure at different timescales as highlighted by Figure 4.3.1. This will allow establishing a bridge between surface science
and theoretical studies on the atomic scale and real-world catalysts in catalytic converters where reactor simulation and modeling on the macroscale, including description of transient processes, is required. This approach to monitoring the atomic scale,
the mesoscale, and the macroscale is analogous to the approach in reactor modeling
from rst principles [16]. Atomic-scale information is required for modeling the surface reactions, the mesoscale for spillover and similar effects, and the macroscale to
account for heat and mass transport.
313
Physical Structure:
Surface Area
Surface
Composition
Overall
Composition
Elemental analysis,
XRF, laser ablation
BET
Method
Crystalline
Phases
Amorphous
Phases
Surface
Morphology
XRD, neutron
diffraction
SEM
Pore
Structure
Mechanical
Strength
Mercury porosimetry,
gas adsorption (Kelvin
equation models
SAXS, X-Ray
microscopy)
Crush and
pressure
test
Surface
Acidity
Surface
Coordination,
Valence
Thermal desorption,
IR
XPS, IR
UV-vis
Dispersion
Tomography
Electron tomography
X-Ray tomography
Figure 4.3.2 Schematic strategy for catalyst characterization; analysis by a number of differ-
ent and complementary techniques is required. The full names of the respective characterization techniques are given in the text.
314
paramagnetic resonance (EPR) are very powerful methods, and the reader is referred
to other books for a detailed description of the mentioned methods [1719].
In order to gain insight into the meso- and macrostructure, similar methods can be
applied. One of the most important aspects is the porosity that can either be measured
by physisorption or by mercury porosimetry. In addition, tomographic imaging is of
great importance; X-ray and NMR tomography are among the few methods that do
not require destruction of the catalyst bed and even allow imaging of reactors
[6, 13, 2025]. X-ray contrast exploiting diffraction, uorescence, or absorption spectroscopy even allows distinguishing different species and phases. Finally, the shaped
catalysts can also be cut and analyzed by microspectroscopic methods like micro-IR,
micro-Raman, and microultraviolet-visible (UV-vis) spectroscopy [6, 24].
Number of publications
500
400
300
200
2010
2009
2008
2007
2006
2005
2004
2003
2001
2002
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
100
Year
Figure 4.3.3 Number of papers published containing the words in situ or operando in connection with catalysts and spectroscopy (ISI Web of Knowledge).
315
298 C
302 C
500 C
331 C
80
oxidized
Rh
m/e = 2
60
m/e = 32
40
20
0
500
450
400
350
323 C
1.0
RT
0.8
0.6
Time
Ignition and Extinction
0.4
0.2
Rh K-edge
0.0
23.20
23.25
23.30
23.35
23.40
300
250
200
T/C
Absorption/a.u.
327 C
1.2
m/e = 15
MS-signals/a.u.
294 C
Extinction
100
Ignition
100
oxidized
Rh
80
60
m/e = 2
MS-signals/a.u.
100 C
m/e = 15
40
20
m/e = 32
0
250
300
350
400
450
500
T/C
E/keV
Figure 4.3.4 Example of the importance of in situ studies; the structural changes unraveled
by selected XANES spectra at the Rh K edge during ignition and extinction of the catalytic
partial oxidation of methane over 2.5% Pt2.5% Rh/Al2O3 in 6% CH4/3% O2/He demonstrate
that the structure changes during heating up and cooling down; relative concentrations for
hydrogen (m/e = 2), methane (m/e = 15), and oxygen (m/e = 32) and the fraction of oxidized
Rh (by linear combination analysis) are given on the right, cf. refs. [11, 29].
316
with a 0.5 mm 0.5 mm beam uncovered that the structure in the beginning of the
catalytic reactor (oxidized) was different from the end of the catalyst bed (reduced
state, cf. [11, 29]). This shows the requirement to (a) measure in certain cases in a
time-resolved way and (b) watch the catalytic reactor not only in an integral manner
but also spatially resolved or even in a spatiotemporal manner [30].
IR
UV
Neutrons
lons, e
X-Rays
soft and hard
Laser light
T
UV/VIS spectroscopy
Neutrons scattering and diffraction
Mbauer
XPS, AES, ISS,
TEM, SEM
TPR, TPS, TPD,...
Products
On-line Analysis
by IR, MS, GC
Figure 4.3.5 The task for in situ/operando studies; techniques that are blue shaded are difcult to realize under reaction conditions since they deal with ions or electrons; extremely thin
window materials and differential pumping are presently designed for these techniques.
317
These challenges are underlined by two extreme cases in Figure 4.3.6. On the left, a
catalyst pellet is embedded in a catalytic reactor that allows optimal transmission of
X-rays. However, the dead volume is large and dynamic studies are not so easy,
whereas a xed-bed reactor allows minimizing it. Further, using a sieved fraction allows establishing optimal conditions for reactions and dynamic changes of the reaction conditions. In fact, analyzing the reduction of Cu/ZnO diluted with boron
nitride as pellet (2 mm thickness, left reactor) and as sieved fraction (80 m particle
diameter) uncovered a reduction of more than 40C lower temperature in the rst
case [31] and the presence of a much higher Cu(I) concentration than in the second
case (ca. 15 min reduction time at a ramp rate of 3C/min). Calculations (details, c.
f. [31]) on the basis of an effective diffusion coefcient of De = 106 to 108 m2/s
lead to the conclusion that internal mass transport is much more limited in the pellet
(0.5 to 60 min) compared to the sieved fraction (0.02 to 2 s). In the case of gas-phase
reactions, the external mass transport is usually negligible but may become very
important in the case of liquid-phase reactions. Furthermore, the previous calculations
demonstrate that under optimal conditions dynamic studies with subsecond time
1. Catalyst Wafer
(a) Principle
X-rays
2 mm
C (CH4)/Cs (CH4)
Gas
Inlet
0.8
0.6
0.4
0.2
0.0
0.0
Catalytic
Performance
2. Sieved Catalyst
(a) Principle
X-rays
Gas
Inlet
2 mm
C (CH4)/Cs (CH4)
1.0
Catalytic
Performance
0.8
0.6
0.4
0.2
0.0
0.0
16 32 48 64 80
Diameter of Particle/m
Figure 4.3.6 Principle of two in situ cells that are optimized for the spectroscopic studies and
the catalytic experiments; the X-ray transmission image is recorded by an X-ray eye, and the
calculated prole of the methane concentration is estimated for 1% CH4/4% O2/He on a pellet
of 1 mm thickness and catalyst particles of 100 m at 500C (taken from ref. [29]).
318
resolution are still reasonable; using powder catalysts, it is, however, not necessary to
go below the subsecond timescale, except if special designs are used.
Not all catalytic reactions embrace rapid structural changes. Often one may only
be interested in stationary conditions. Still, if measuring a catalytic reaction, one has
to be very careful with the mass- and heat-transport effects. An illustrative example is
the total oxidation of hydrocarbons over Pd particles [31, 35]. In catalytic converters,
thin wash-coat layers are used to exploit the noble metals catalytic activity efciently, and the Thiele modulus is used for its estimation. What happens when the
catalyst is not used efciently enough is demonstrated in 1(c) of Figure 4.3.6. For
the pellet-like catalyst, methane is rapidly consumed on the outer part since the catalytic reaction is much faster than the diffusion of the catalyst. If 1% CH4 and 4% O2
in Helium is used, this means that in the top 50100 m layer methane is completely
combusted and the inner part (more than 90% of the catalyst pellet) is exposed
mainly to 2% oxygen without methane. Shining the X-rays through the catalyst pellet
therefore results in reality in monitoring most of the catalyst in 2% oxygen instead of
1% CH4 and 4% O2 in Helium.
Despite the catalytic reactor on the right-hand side being excellently suited for gasphase reactions, it still has its limitations: quartz-glass wall thicknesses below 20 m
should be used if X-rays with photon energy below 8 keV are applied. Other designs
are required not only at lower energy but also if reactions are followed at higher pressure, in liquid phase, or even when the preparation of catalysts is studied. X-rays are
best suited for this purpose since they can penetrate a number of materials (Be, Kapton, graphite, aluminum foil, silicon nitride, quartz, etc.) rather well but it remains
a challenge to nd the best compromise of the in situ cell for the specic scientic
questions, and therefore new areas require the design of new in situ cells [31, 32, 34].
The same is true for spectroscopic studies based on IR, UV light, and laser light
for Raman spectroscopy; the methods are outlined in Figure 4.3.5. The application
of methods based on the use or detection of electrons or ions is difcult. This is especially true for TEM, as well as surface-sensitive techniques like XPS, ISS, SIMS, and
so forth, which are preferentially performed in ultrahigh vacuum. Thanks to the
development of thin window materials, based on silicon nitride membranes, as
well as efcient differential pumping and high-intensity beams, nowadays samples
can be exposed to atmospheres in the mbar regime [10, 3638].
Absorption/a.u.
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
24.30
770 C
735 C
XRD
319
Raman
400 C
705 C
740 C
670 C
500 C
770 C
Intensity/a.u.
800 C
XAS
810 C
70 C
~100 C/spectrum
770 C
500 C
RT
24.35
24.40
E/keV
24.45
24.50 12.0
12.5
13.0
2/
13.5
800 C
14.0
Figure 4.3.7
which the structural changes of a 10 wt. % Pd/ZrO2 catalyst were monitored during
the total oxidation of methane by EXAFS, XRD, and Raman spectroscopy [35, 39].
Interestingly, a deactivation of the catalyst is found at 750C combined with the
autoreduction of the Pd catalyst and a strong sintering due to the fact that intense
reections in the XRD pattern are observed. The Raman spectrum during the rst
heat up does not show any contributions demonstrating a very good dispersion of
Pd in the catalyst as prepared, whereas during the second heat up, a band due to
PdO is clearly visible.
Under reaction conditions, changes in oxidation state or sintering effects can occur,
and morphological changes have also been observed [8, 9, 40, 41]. Again, the direct
relation of catalytic activity and structure using complementary methods is benecial
[2, 8]. One prominent illustration in the eld of hydrogen management is the hydrogenation of CO2. Copper particles are one of the keys in this kind of reaction and they
are already frequently used in methanol synthesis. EXAFS and XRD studies have demonstrated that the shape of such particles, as well as the microstructure, strongly
depend on the reaction conditions [8, 40]. This is depicted in Figure 4.3.8. Under
more reducing conditions (either a lower CO2/CO ratio or less water in the feed),
the Cu-Cu coordination number in EXAFS is smaller than when including water
(higher oxidation potential). Stronger reducing conditions probably lead to vacancies
on the ZnO surface and therefore more Cu-surface atoms (or a spreading out of the
particles as depicted in the scheme). Both methods indicated at even stronger reducing
conditions the formation of Cu-Zn alloys, also evidenced by IR spectroscopy [42, 43].
Based on the model derived from these studies (Figure 4.3.8C), the system was further
studied by electron microscopy where specic reaction conditions could be mimicked
in the mbar region [9, 41]. Despite the fact that real reaction conditions cannot be used
(yet) in the electron microscope and particles can be investigated only very locally, the
model could be further substantiated. The gure also demonstrates the complementarity of these local imaging methods under idealized conditions and the integral spectroscopic and scattering methods under real reaction conditions. Note that further effects
like strain in crystals may additionally play an important role [12].
320
CN (Cu-Cu)
+H2O
9
dry +H2O
+H2O 6
dry
5
4
3
2
1
0
1
2
3
Dry-Wet Syngas Cycle
(c) Model
Oxidizing
Methanol Formation/a.u.
10
I)
ZnO
II)
ZnO
III)
5 nm
5 nm
ZnO
5 nm
IV)
Reducing
ZnO
Oxygen Vacancies
Reduced Zn
Figure 4.3.8 Dynamics of Cu/ZnO particles shown by in situ EXAFS (left) and in situ electron microscopy (10 mbar, middle) and the model derived from EXAFS (right); synthesis gas
with more CO2 or H2O in the feed is more oxidizing and leads to a higher Cu-Cu coordination
number and more round-shaped particles (reproduced with permission from refs. [8, 9]).
321
Table 4.3.1 Examples of combining complementary spectroscopic methods while the cataly-
Reaction studied
Reference
EXAFS/XRD
[2, 8]
UV-vis/EPR
EPR/UV-vis/Raman
Propane dehydrogenation
[48, 49]
UV-vis/Raman/XAFS
Propane dehydrogenation
[50]
XANES/XRD
XANES/XRD/Raman
[35, 39]
NMR/UV-vis
[51]
EXAFS/DRIFTS
[5254]
WAXS/XANES and
UV-vis
[55]
XAFS/WAXS
[56]
EPR/UV-vis
[57]
Raman/IR/EPR
[58]
ATR-IR/UV-vis
[59]
EXAFS/DRIFTS
[60]
322
-Fe was formed after reduction apart from Fe(II)silicates and iron oxide [63].
Monitoring in the 10 nm100 nm regime appears important to grab the reactivity
of different phases, spillover, aging, redispersion, and support effects. While X-rays
with a focus of 20 nm and less have been available in the soft X-ray regime since
the 1990s [64, 65], spot sizes below 100 nm have only very recently been achieved
with hard X-rays [66, 67]. Thus, only recently, X-ray microscopy has been achieved
in the submicrometer region using hard X-rays, which allow in situ spectroscopic
cells at normal or elevated pressure as well as in liquid phase [13].
Equally important are structural changes on the micrometer and the millimeter
length scales, and eventually real reactors on a centimeter or even meter scale. In the
rst cases, full-eld XAS is a well-matched method for in situ monitoring [6, 13, 25].
An illustrative example is the catalytic partial oxidation (CPO) of methane to synthesis
gas, a relevant reaction in the futures solid-oxide fuel cells. In Figure 4.3.4, it was
G
Oxygen K edge
-Fe2O3
100%
50 nm
705
710
715
720
725
H
Fe0 Fe2SiO4 Fe3O4
25 25
50
Normalized absorption
Normalized absorption
E
100%
33
50 17 100%
2
50 nm
705
710
715
720
Fe3O4
20
100%
75
725
Fe0/FexCy Fe2SiO4
40
50
100%
2
50 nm
705 710 715 720 725
X-ray photon energy (eV)
-Fe2O3
Fe3O4
SiO2
Fe2SiO4
Fe /FexCy
Figure 4.3.9 Chemical contour maps (AC) and X-ray absorption spectra (DI) of a FischerTropsch catalyst during the following different stages of reaction: before treatment (top), after
2 h in H2 at 350C (middle), after 4 h in synthesis gas at 250C (bottom). Chemical contour
maps (A, B, C) of a 400 nm 3,750 nm region and corresponding iron L2 and L3 edge (D, E, F)
and oxygen K edge (G, H, I). Specic sampling regions and the corresponding X-ray absorption spectra are indicated in the gures (for details, cf. ref. [63], reproduced with permission,
Copyright Nature Publishing Group, 2008).
200 m
323
1.2
1
1.0
Absorption/a.u.
0.8
2
0.6
0.4
0.2
0.0
23.20
23.22
23.24
23.26
E/keV
23.28
23.30
Figure 4.3.10 Full-eld X-ray microscopy on a 5 wt% Rh/Al2O3 catalyst during catalytic
partial oxidation of methane: (A) amount of oxidized Rh species (corresponds to XANES species 1 in [D]), (B) reduced Rh species (reduced species 2 in [D]), (C) the distribution of other
elements that show a featureless absorption spectrum in the given energy range, and (D) spectra used for X-ray absorption contrast (original image taken by X-ray camera was 3.0 mm
1.5 mm; the reaction gas mixture 6% CH4/3% O2/He enters from the left) (reproduced with
permission from ref. [69], Copyright ACS, 2006).
demonstrated that the noble metal atoms are in a reduced state under the reaction conditions. However, full-eld and scanning X-ray microscopy studies with XAS contrast
on Pt-, Rh-, and Pd-based catalysts uncovered that the noble metals are in an oxidized
state in the beginning of the catalyst bed where the total oxidation of methane occurs,
and in the reduced state in the second part of the reactor where reforming of the remaining methane with the formed water and carbon dioxide occurs [68, 69]. Some of the results are shown in Figure 4.3.10. This is in line with spatially resolved product analysis
together with temperature proles [70, 71]. As in integral approaches, the aspect of
dynamic variations is also important during spatially resolved studies. The ignition
process of the catalytic partial oxidation of methane as depicted in Figure 4.3.4 has
been recently monitored by full-eld X-ray microscopy unraveling changes on the
micrometer scale on the subsecond timescale [15, 72]. These recent studies demonstrate
that in certain cases atomic-scale information on a macroscale can be obtained and is
needed [24, 73].
324
them to studying catalytic materials in their working state. This is important since
each study with an in situ cell that tries to combine real-world catalysis with the
most advanced characterization tools is a compromise. In certain cases, the reaction
rates are high, and internal mass transport is important to consider during the spectroscopic study; in other cases, the reaction conditions with high temperature or pressure are demanding, the structure of the catalyst is very dynamic, or it is very difcult
to grab the structure by only one characterization method. Hence, also with new materials and/or reactions in chemical energy conversion, new approaches have to be
developed to come up with the best design for combining one or several spectroscopic
techniques with the right timescale and length scale, as well as online measurement of
the catalytic activity. The last section of this chapter has shown that spatially resolved
and spatiotemporal studies pave the way to understand the dynamics on the atomicscale structure in catalytic reactors (see overview and perspective in ref. [74]), which
will have an important impact on the modeling of the catalytic reactions and reactors.
4.3.8 Acknowledgment
The author would like to thank all co-workers at the Institute for Chemical Technology and Polymer Chemistry (ITCP) & Institute for Catalysis Research and Technology (IKFT) at Karlsruhe Institute of Technology (KIT), co-workers at DTU, A.
Baiker (ETH Zurich), R. Frahm (University Wuppertal), C.G. Schroer (TU Dresden) and their group members for stimulating discussions and support as well as HASYLAB at DESY (beamlines X1, C, BW1 and BW2, Hamburg, Germany), the APS
(ID-1, Chicago, USA), ESRF (ID26 and SNBL, Grenoble, France), ANKA
(ANKA-XAS), and SLS (superXAS and microXAS, Villigen, Switzerland) for granting beamtime for the different projects shown in this chapter. Financial support by
KIT, the BMBF projects Materials in Action and X-ray microscopy for operando
investigations of chemical reactions, the European Community - Research Infrastructure Action under the FP6: Structuring the European Research Area (Integrating Activity on Synchrotron and Free Electron Laser Science (IA-SFS) RII3CT-2004-506008), DANSCATT, ESRF, SLS and the Karl Winnacker grant are gratefully acknowledged.
4.3.9 References
1. Thomas JM. Design, synthesis and in situ characterization of new solid catalysts. Angew
Chem Int Ed. 1999;38:3588.
2. Grunwaldt J-D, Clausen BS. Combining XRD and EXAFS with on-line catalytic studies
for the in situ characterization of catalysts. Top Catal. 2002;18:37.
3. Weckhuysen BM. Determining the active site in a catalytic process: operando spectroscopy is more than a buzzword. Phys Chem Chem Phys. 2003;5:4351.
4. Topse H. Developments in operando studies and in situ characterization of heterogeneous catalysts. J Catal. 2003;216:155.
4.3.9 References
325
5. Banares MA. Operando methodology: combination of in situ spectroscopy and simultaneous activity measurements under catalytic reaction conditions. Catal Today. 2005;100:71.
6. Grunwaldt JD, Kimmerle B, Baiker A, et al. Catalysts at work: from integral to spatially
resolved X-ray absorption spectroscopy. Catal Today. 2009;145:267.
7. vant Blik HFJ, Van Zon JBA, Huiziga T, Vis JC, Koningsberger DC, Prins R. An extended x-ray absorption ne structure spectroscopy study of a highly dispersed rhodium/
aluminum oxide catalyst: the inuence of carbon monoxide chemisorption on the topology
of rhodium. J Phys Chem. 1983;87:2264.
8. Grunwaldt J-D, Molenbroek AM, Topse N-Y, Topse H, Clausen BS. In situ investigations of structural changes in Cu/ZnO catalysts. J Catal. 2000;194:452.
9. Hansen PL, Wagner JB, Helveg S, Rostrup-Nielsen JR, Clausen BS, Topse H. Atomresolved imaging of dynamic shape changes in supported copper nanocrystals. Science.
2002;295:2053.
10. Helveg S, Lopez-Cartes C, Sehested J, et al. Atomic-scale imaging of carbon nanobre
growth. Nature. 2004;427:426.
11. Grunwaldt J-D, Baiker A. Axial variation of the oxidation state of Pt-Rh/Al2O3 during
partial methane oxidation in a xed-bed reactor: an in situ X-ray absorption spectroscopy
study. Catal Lett. 2005;99:5.
12. Behrens M. Meso- and nano-structuring of industrial Cu/ZnO/(Al2O3) catalysts. J Catal.
2009;267:24.
13. Grunwaldt JD, Schroer CG. Hard and soft X-ray microscopy and tomography in catalysis: bridging the different time and length scales. Chem Soc Rev. 2010;39:4741.
14. Ressler T, Hagelstein M, Hatje U, Metz W. In situ X-ray absorption spectroscopy studies
on chemical oscillations in the CO/O2 system on supported Pd catalysts. J Phys Chem B.
1997;101:6680.
15. Kimmerle B, Baiker A, Grunwaldt JD. Oscillatory behaviour of catalytic properties, structure and temperature during the catalytic partial oxidation of methane on Pd/Al2O3. Phys
Chem Chem Phys. 2010;12:2288.
16. Maestri M. Atomic-scale understanding of complex chemical kinetics. In: Pignataro B,
editor. New strategies in chemical synthesis and catalysis. Wiley VCH; 2011.
17. Baiker A, Kohler M. Characterization of heterogeneous catalysts. In: Cheremisinoff PN,
editor. Handbook of heat and mass transfer. Houston: Gulf Publishing; 1989, Vol. 3, p. 3.
18. Niemantsverdriet JW. Spectroscopy in catalysis. Weinheim: Wiley-VCH; 2007.
19. Ertl G, Knzinger H, Schth F, Weitkamp J, editors. Handbook of heterogeneous catalysis. Weinheim: Wiley-VCH; 2008.
20. Schroer CG, Kuhlmann M, Gnzler TF, et al. Mapping the chemical states of an element
inside a sample using tomographic X-ray absorption spectroscopy. Appl Phys Lett.
2003;82:3360.
21. Gladden LF. Recent advances in MRI studies of chemical reactors: ultrafast imaging of
multiphase ows. Topics Catal. 2003; 24:19.
22. Koptyug IV, Lysova AA, Sagdeev RZ, Kirillov VA, Kulikov AV, Parmon VN. In situ
MRI of the structure and function of multiphase catalytic reactors. Catal Today.
2005;105:464.
23. Beale AM, Jacques SDM, Bergwerff JA, Barnes P, Weckhuysen BM. Tomographic
energy dispersive diffraction imaging as a tool to prole in three dimensions the distribution and composition of metal oxide species in catalyst bodies. Angew Chem Int Ed.
2007;46:8832.
24. Weckhuysen BM. Chemical imaging of spatial heterogeneities in catalytic solids at different length and time scales. Angew Chem Int Ed. 2009;48:491043.
25. Schroer CG, Grunwaldt J-D. X-ray tomography. Synchrotron Radiat News. 2009;22:23.
326
26. Basini L, Guarinoni A, Aragno A. Molecular and temperature aspects in catalytic partial
oxidation of methane. J Catal. 2000;190:284.
27. Grunwaldt J-D, Basini L, Clausen BS. In situ EXAFS study of Rh/Al2O3 catalysts for catalytic partial oxidation of methane. J Catal. 2001;200:321.
28. Hannemann S, Grunwaldt J-D, Gnther D, Krumeich F, Lienemann P, Baiker A. Combination of ame synthesis and high throughput experimentation: preparation of alumina
supported noble metal particles and their application in the catalytic partial oxidation of
methane. Appl Catal A. 2007;316:226.
29. Grunwaldt J-D. Shining X-rays on catalysts at work. J Phys Conf Ser. 2009;190:012151.
30. Grunwaldt JD, Beier M, Kimmerle B, et al. Structural changes of noble metal catalysts
during ignition and extinction of the partial oxidation of methane studied by advanced
QEXAFS techniques. Phys Chem Chem Phys. 2009;11:8779.
31. Grunwaldt J-D, Caravati M, Hannemann S, Baiker A. X-ray absorption spectroscopy
under reaction conditions opportunities and limitations of in situ monitoring and
time-resolved studies of heterogeneous catalysts. Phys Chem Chem Phys. 2004;6:3037.
32. Bare SR, Ressler T. Characterization of catalysts in reactive atmospheres by X-ray
absorption spectroscopy. In: Advances in Catalysis. 2009;52:339.
33. Grunwaldt J-D, Baiker A. Time-resolved and operando XAS studies on heterogeneous
catalysts from the gas phase towards reactions in supercritical uids. In: Hedman B, Oianetta P, editors. AIP Conference Proceedings X-ray Absorption Fine Structure XAFS13;
2007: Vol. CP 882, p. 57781.
34. Meunier FC. The design and testing of kinetically-appropriate operando spectroscopic
cells for investigating heterogeneous catalytic reactions. Chem Soc Rev. 2010;39:4602.
35. Grunwaldt JD, van Vegten N, Baiker A. Insight into the structure of supported palladium
catalysts during the total oxidation of methane. Chem Commun. 2007:4635.
36. Creemer JF, Helveg S, Hoveling GH, et al. Atomic-scale electron microscopy at ambient
pressure. Ultramicroscopy. 2008;108:993.
37. Hansen PL, Helveg S, Datye A. Atomic-scale imaging of supported metal nanocluster catalysts in the working state. J Catal. 2006;50:77.
38. Bluhm H, Hvecker M, Knop-Gericke A, Kiskinova M, Schlgl R, Salmeron M. In situ
X-ray photoelectron spectroscopy studies of gassolid interfaces at near-ambient conditions. MRS Bull. 2007;32:1022.
39. Grunwaldt JD, van Vegten N, Baiker A, van Beek W. Insight into the structure of Pd/
ZrO2 during the total oxidation of methane using combined in situ XRD, X-ray absorption and Raman spectroscopy. J Phys Conf Ser. 2009;190:012160.
40. Clausen BS, Schitz J, Grbek L, et al. Wetting/non-wetting phenomena during catalysis:
evidence from in situ on-line EXAFS studies of Cu-based catalysts. Top Catal. 1994;1:
367.
41. Simonsen SB, Dahl S, Johnson E, Helveg S. Ceria-catalyzed soot oxidation studied by
environmental transmission electron microscopy. J Catal. 2008;255:1.
42. Fujitani T, Nakamura J. The effect of ZnO in methanol synthesis catalysts on Cu dispersion and the specic activity. Catal Lett. 1998;56:119.
43. Topse N-Y, Topse H. On the nature of surface structural changes in Cu/ZnO methanol
synthesis catalysts. Topics Catal. 1999;8:267.
44. Grunwaldt J-D, Ramin M, Rohr M, Michailowski A, Patzke GR, Baiker A. High pressure in situ x-ray absorption spectroscopy cell for studying simultaneously the liquid phase
and the solid/liquid interface. Rev Sci Instrum. 2005;76:054104.
45. Clausen BS, Topse H, Frahm R. Application of combined X-ray diffraction and absorption techniques for in situ catalyst characterization. Adv Catal. 1998;42:315.
4.3.9 References
327
46. Brckner A, Kondratenko E. Simultaneous operando EPR/UV-vis/laser-Raman spectroscopy a powerful tool for monitoring transition metal oxide catalysts during reaction.
Catal Today. 2006;113:16.
47. Bentrup U. Combining in situ characterization methods in one set-up: looking with more
eyes into the intricate chemistry of the synthesis and working of heterogeneous catalysts.
Chem Soc Rev. 2010;39:4718.
48. Brckner A. Simultaneous combination of in situ-EPR/UV-VIS/on line GC: a novel setup
for investigating transition metal oxide catalysts under working conditions. Chem Commun. 2001:2122.
49. Brckner A. Killing three birds with one stone simultaneous operando EPR/UV-vis/
Raman spectroscopy for monitoring catalytic reactions. Chem Commun. 2005:1761.
50. Beale AM, van der Eerden AMJ, Kervinen K, Newton MA, Weckhuysen BM. Adding a
third dimension to operando spectroscopy: a combined UV-Vis, Raman and XAFS setup
to study heterogeneous catalysts under working conditions. Chem Commun. 2005:3015.
51. Hunger M, Wang W. Formation of cyclic compounds and carbenium ions by conversion
of methanol on weakly dealuminated zeolite H-ZSM-5 investigated via a novel in situ CF
MAS NMR/UV-Vis technique. Chem Commun. 2004.
52. Newton MA, Jyoti B, Dent AJ, Fiddy SG, Evans J. Synchronous, time resolved, diffuse
reectance FT-IR, energy dispersive EXAFS (EDE) and mass spectrometric investigation
of the behaviour of Rh catalysts during NO reduction by CO. Chem Commun. 2004:2382.
53. Newton MA, Dent AJ, Fiddy SG, Jyoti B, Evans J. Combining diffuse reectance infrared
spectroscopy (DRIFTS), dispersive EXAFS, and mass spectrometry with high time resolution: potential, limitations, and application to the study of NO interaction with supported Rh catalysts. Catal Today. 2007;126:64.
54. Ferri D, Kumar MS, Wirz R, et al. First steps in combining modulation excitation spectroscopy with synchronous dispersive EXAFS/DRIFTS/mass spectrometry for in situ time
resolved study of heterogeneous catalysts. Phys Chem Chem Phys. 2010;12:5634.
55. OBrien MG, Beale AM, Jacques SDM, Weckhuysen BM. A combined multi-technique in
situ approach used to probe the stability of iron molybdate catalysts during redox cycling.
Top Catal. 2009;52:1400.
56. de Smit E, Beale AM, Nikitenko S, Weckhuysen BM. Local and long range order in promoted iron-based Fischer-Tropsch catalysts: a combined in situ X-ray absorption spectroscopy/wide angle X-ray scattering study. J Catal. 2009;262:244.
57. Kumar MS, Schwidder M, Grnert W, Brckner A. On the nature of different iron sites
and their catalytic role in Fe-ZSM-5 DeNO(x) catalysts: new insights by a combined EPR
and UV/VIS spectroscopic approach. J Catal. 2004;227:384.
58. Brandhorst M, Cristol S, Capron M, et al. Catalytic oxidation of methanol on Mo/Al2O3
catalyst: an EPR and Raman/infrared operando spectroscopies study. Catal Today.
2006;113:34.
59. Brgi T. Combined in situ attenuated total reection infrared and UV-vis spectroscopic
study of alcohol oxidation over Pd/Al2O3. J Catal. 2005;229:55.
60. Kubacka A, Martnez-Arias A, Fernndez-Garca M, Newton MA. Dynamic operando
observation of 10 wt% Pd-based TWCs: simultaneous XAS/DRIFTS/mass spectrometry
analysis of the effects of Ce0.5Zr0.5O2 loading on structure, reactivity and performance.
Catal Today. 2009;145:288.
61. Thomas JM, Hernandez-Garrido JC. Probing solid catalysts under operating conditions:
electrons or X-rays? Angew Chem Int Ed. 2009;48:3904.
62. Urakawa A, Baiker A. Space-resolved proling relevant in heterogeneous catalysis. Top
Catal. 2009;52:1312.
328
63. de Smit E, Swart I, Creemer JF, et al. Nanoscale chemical imaging of a working scanning
transmission X-ray microscopy. Nature. 2008;456:222.
64. Chao WL, Harteneck BD, Liddle JA, Anderson EH, Attwood DT. Soft X-ray microscopy
at a spatial resolution better than 15nm. Nature. 2005;435:1210.
65. Vila-Comamala J, Jemovs K, Raabe J, et al. Advanced thin lm technology for ultrahigh
resolution X-ray microscopy. Ultramicroscopy. 2009;109:1360.
66. Mimura H, Handa S, Kimura T, et al. Breaking the 10 nm barrier in hard-X-ray focusing.
Nat Phys. 2010;6:57.
67. Schroer CG, Lengeler B. Focusing hard x rays to nanometer dimensions by adiabatically
focusing lenses. Phys Rev Lett. 2005;94:054802.
68. Hannemann S, Grunwaldt JD, Kimmerle B, Baiker A, Boye P, Schroer C. Axial changes
of catalyst structure and temperature in a xed-bed microreactor during noble metal catalysed partial oxidation of methane. Top Catal. 2009;52:1360.
69. Grunwaldt J-D, Hannemann S, Schroer CG, Baiker A. 2D-Mapping of the catalyst structure inside a catalytic microreactor at work: partial oxidation of methane over Rh/Al2O3.
J Phys Chem B. 2006;110:8674.
70. Horn R, Degenstein NJ, Williams KA, Schmidt LD. Spatial and temporal proles in millisecond partial oxidation processes. Catal Lett. 2006;110:169.
71. Basile F, Fornasari G, Trir F, Vaccari A. Effect of reaction parameters and catalyst
composition on the thermal prole and heat distribution. Catal Today. 2001;64:21.
72. Kimmerle B, Grunwaldt J-D, Baiker A, et al. Visualizing the ignition of catalytic partial
oxidation of methane. J Phys Chem C. 2009;113:3037.
73. Schroer CG, Boye P, Feldkamp JM, et al. Hard X-ray microscopy with elemental, chemical, and structural contrast. Acta Phys Pol A. 2010;117:357.
74. Grunwaldt J-D, Wagner JB, Dunin-Borkowski RE. Imaging Catalysts at Work: A Hierarchical Approach from the Macro- to the Meso- and Nano-scale. ChemCatChem. 2013;5,
DOI: 10.1002/cctc.201200356.
4.4.1 Introduction
Harvesting, saving, and storing energy are key issues in todays general energy discussions [1]. Harvesting energy through photochemistry and storing energy in compounds using thermal energyefcient processes are attractive solutions to some
problems within the grand picture [2].
Heterogeneous thermal and photocatalysis are bound to play a key role in those
solutions. However, while thermal heterogeneous catalysis is widespread in chemical
industries, and photocatalysis not yet, there is still a strong demand in both areas for
research in order to understand the process and its elementary steps as well as to
rationally design the catalytic material.
Investigation of model catalysts can play a decisive role in a rational approach to
understand heterogeneous catalysis, and this is the topic of this chapter [36].
In the introduction, the model systems will be dened to familiarize the reader with
the approach in order to appreciate the connection to real-world catalysis. Following
the introduction, we will demonstrate via four case studies various fundamental aspects in thermal and photocatalysis whereby studies on model systems might become
important to unravel the foundations of reaction mechanisms.
The model systems have been created under the premises that it is possible to
investigate them at the atomic scale using the toolbox that has been developed in
surface science during the second half of the last century up to now [7].
The model systems are based on the idea to grow well-ordered oxide layers or
thin lms, representing the bulk material, and as such the catalyst support, on singlecrystal metal surfaces using the concepts of epitaxial growth [8]. This, in turn, allows
one to investigate bulk insulators without having to worry about charging when
charged information carriers such as electrons or ions are used to investigate
those systems. One thus avoids one of the key difculties hampering the detailed
study of real catalysts. Also the application of scanning probe techniques including
electron tunneling is possible for thin-lm systems. The fact that the oxides are
grown on a metal support also ensures the easy applicability of infrared reection
absorption spectroscopy at such systems. Moreover, controlling the growth parameters of the oxide lm also allows one to vary the density of defects at the surface
of the oxide lms, a factor of importance when nanoparticles are grown on the
oxide support.
330
The rst case study presented in this paper is related to a model system that is schematically represented in Figure 4.4.1A [9]. An oxide lm representing the bulk material is grown on a metal surface. The growth of metal nanoparticles (MNPs), and
in particular their morphologies, which may distinctively inuence the catalytic performance, are compared on pristine oxide lms and lms that have been doped
with metal ions that may assume higher valences than the metal of the host material. Using the model system, even the effect of doping as a function of the depth
(Figure 4.4.1A) at which the dopant is deposited below the surface may be studied,
and thus questions of the presence of dopants in materials used to prepare supports
in real catalysis may be addressed.
Real dispersed metal catalysts are prepared by precipitation from solution.
Investigating processes involving oxide surfaces in solution is a challenging topic.
Figure 4.4.1B schematically shows the type of system investigated [10]. Particles
are deposited from an aqueous solution of a metal chloride as a function the pH
vacuum/gas
solution
Cl
M+
ligands 4
M+
dopant
oxide
oxide
metal
metal
vacuum/gas
vacuum
h
overlayer
oxide
oxide
metal
metal
Figure 4.4.1 Schematic representation of the model systems discussed within the chapter:
(A) nanoparticle growth inuenced by dopants in the support, (B) nanoparticle deposition
from solution, (C) strong metal support interaction, and (D) photochemistry at supported
nanoparticles as a function of size.
331
of the solution. The pH strongly inuences the outcome of the impregnation process.
Understanding the details of such processes is of utmost importance for preparation
of technical catalysts, and the rst steps in this direction are being taken. This
represents the second case study.
In a third case study, the effect of a reducible support material, such as magnetite,
on Pt nanoparticles anchored onto the magnetite surface after a heat treatment is investigated. In this system, a typical strong metalsupport interaction (SMSI) that is,
the migration of a thin oxide lm from the support onto the metal nanoparticles inuencing catalytic performance has been observed. An SMSI effect usually renders
the system inactive in oxidation reactions such as CO oxidation. Here it is shown that
depending on the chemical potential of oxygen in the gas phase, the SMSI system
schematically shown in Figure 4.4.1C may become more active than the clean metal
at the given temperature. Some general design principles may be deduced from
those studies [11, 12].
In the fourth and last case study, the interaction of laser light with aluminasupported silver nanoparticles (Ag NPs) and the consequences for the desorption
of molecules were studied. In particular, the question of tuning the light frequency
into the plasmon frequency of the particles and thus optimizing the harvesting of
energy has been investigated as a function of particle size. Also, the consequences
of varying the duration of the laser pulse from nanoseconds to femtoseconds, addressing more fundamental questions, are discussed. This is schematically indicated in
Figure 4.4.1D [1315].
This chapter does not provide the space to discuss all aspects of model catalyst studies in detail, but the four case studies exemplify the breadth and depth that is
provided by such investigations.
332
CO oxidation [22, 23]. A similar conclusion was drawn from scanning tunneling
microscopy (STM) measurements on the Au/TiO2 system, where again bilayer
high deposits turned out to be the most active ones [24, 25]. Naturally, a close structure-reactivity relationship is not restricted to Au particles but can be found for other
metal species in a multitude of chemical reactions [26].
Given the importance of the particle geometry, a good fraction of catalysis
research has been devoted to exploring different means to optimize this parameter.
In principle, particle morphologies are governed by the strength of the metal-support
interaction with respect to the surface-free energies of the separated systems. For
oxide supports, as commonly used in heterogeneous catalysis, this usually leads to
the formation of three-dimensional (3-D) deposits, as both the free energy of the
oxide surface and the interface adhesion of the ad-metal are small (Figure 4.4.2A).
Moreover, the metal particles tend to mimic the symmetry of the oxide lattice in
order to establish interfacial bonds. This causes, for instance, the Ni lattice to switch
from face centered cubic (fcc) in the bulk to hexagonal close packed (hcp) in small
deposits grown on the MgO(001) surface [27]. To overcome these growth restrictions,
several procedures have been developed that enable a certain control of particle equilibrium shapes on oxide supports. Most of them are based on a targeted modication
of the substrate morphology, for example, by introducing defects [28] and suitable
ad-species (i.e. hydroxyl groups) [29]. Their function is to anchor the metal particles
on the inert oxide surface, inhibiting sintering and ripening processes (Figure 4.4.2A).
Alternatively, oxide surfaces with polar character might be used because they feature
high surface-free energies and therefore an enhanced adhesion of the ad-metal
[30, 31]. This approach has been demonstrated for the polar ZnO(0001) surface,
where Cu was found to grow in a layer-by-layer fashion [32]. Finally, electron transfer processes across the metal-oxide interface can be exploited to alter the equilibrium
shape of the ad-particles [33]. The charge exchange between metal deposits and support is stimulated by different mechanisms. On thin oxide lms grown on metal substrates, it often occurs spontaneously due to the different Fermi levels in support and
ad-metal. The accumulation of excess charges in the deposit thereby activates new
charge-mediated coupling schemes between both systems, such as polaronic and Coulomb-type interactions. The enhanced interface adhesion promotes a spreading of the
metal on the oxide support and might, in an extreme case, switch the growth regime
from 3-D to two-dimensional (2-D) when going from neutral to charged ad-species
[34, 35]. The charge transfer is, however, restricted to ultrathin oxide lms, as it
requires electron tunneling through the insulating layer.
A charge-induced formation of low-dimensional metal aggregates has been identied on various oxide thin lms, employing mostly STM as the imaging tool. On alumina lms grown on NiAl(110), for example, the formation of one-dimensional
(1-D) gold chains was observed (Figure 4.4.2C) [36]. The linear nature of the aggregates enables good spatial separation of the extra electrons inside the particle
electronic system, which in turn lowers the internal Coulomb repulsion. With increasing atom number, the equilibrium shape changes to planar 2-D islands, as observed
for gold on magnesia thin lms [34]. By analyzing the electronic structure of the 2-D
islands, their charge state could be quantied directly [37]. For a planar Au18 aggregate on 2 monolayer (ML) MgO/Ag(001) for example, four extra electrons were
Au3
Au4
Au5
333
Au7
Conductance: 0.2 V
Figure 4.4.2 (A) STM images of Ag particles on pristine and sputtered Al2O3 lms grown on
NiAl(110) (1 V, 100 100 nm2). Note the different particle density and shape on the defectrich oxide surface. (B) Atomically resolved image of the top facet of a Pd particle on Al2O3/
NiAl(110) (0.05 V, 4 4 nm2). The facet is well ordered and corresponds to a Pd(111) plane.
(C) One-dimensional Au clusters on ultrathin alumina lms that develop as a result of a charge
transfer out of the support (1.0 V, 5 5 nm2). (D) Two-dimensional Au island on a 2 monolayer (ML) MgO/Ag(001) lm. Whereas the central panel displays the island morphology
(0.2 V, 25 25 nm2), the left and right panels are conductance images and reect the state
density at the given bias voltage. (E) Large monolayer high Au sheets on MgO thin lms
with characteristic meandering shapes (0.2 V, 25 25 nm2).
found to reside on the deposit, providing a rough idea of the efciency of charge
transfer processes across the metal-oxide interface. The relationship between the
preferred growth geometry in the presence of excess charges and the need of the system to minimize the internal electron-electron repulsion has been visualized best for
even larger Au islands on the MgO support [38]. The transferred charges from the
support localize exclusively at the perimeter of the metal island, where electronrich edge states with Au 6s character become visible in tunneling spectroscopy
(Figure 4.4.2D). As the electron-storage capacity of those edge states is limited,
the gold islands develop a peculiar meandering shape that keeps the ratio between
inner and edge atoms constant (Figure 4.4.2E). The electron-rich nature of their
perimeter renders charged monolayer Au islands particularly attractive for adsorbates. Not surprisingly, CO molecules dosed onto the Au/MgO thin-lm system
were found to bind exclusively to the low-coordinated edge atoms, while no adsorption takes place in the island interior [39]. According to density functional theory
334
(DFT) calculations, the island perimeter is also prone to bind and dissociate O2 molecules, being the initial step to form gold oxide at the island perimeter [40].
Although charge-mediated growth schemes are ideally suited to alter the equilibrium shape of metal particles on thin oxide lms, the mechanism breaks down on
thick, bulk-like oxide supports as used in real catalysts. The reason is the limited tunneling length of electrons through an insulating oxide, being on the order of 12 nm.
An obvious approach to overcome this restriction is to insert the charge donors
directly into the oxide material, preferentially into a near-surface region. By this
means, charge transfer into the ad-metal can be sustained even for arbitrarily thick
oxide slabs, preserving the previously discussed possibility to tailor the particle
shapes. There are two types of donor-like impurities in an oxide material. The rst
one comprises intrinsic lattice defects that are able to trap and release electrons [41].
Prominent examples are single O vacancies that contain one or two excess electrons
(Fo or F+ centers) and structural lattice defect with electron-trapping abilities. For
example, grain boundaries that are present in most oxide lms were found to trap
large quantities of extra electrons (up to 5 per nm defect line) [42, 43]. The subsequent
transfer of electrons out of these defects into suitable metal deposits has been demonstrated for the Au/MgO1x system, employing infrared absorption spectroscopy with
CO as probe molecule [39, 44]. The main disadvantage of using intrinsic lattice defects as electron donors is the strong variation of their concentration with the
oxide preparation conditions. Whereas the abundance of oxygen vacancies depends
on the O2 chemical potential during preparation or reaction conditions, the density
of structural defects is given by the mechanical strain exerted on the oxide lattice during growth. Both inuences cannot easily be controlled in real materials used for catalysis and energy storage and are hence unsuitable to tailor the equilibrium geometry of
ad-particles.
The second approach exploits new charge degrees of freedom that can be inserted
into an oxide lattice by replacing some of the intrinsic ions with foreign species
[4550]. Whereas doping has proved to be extremely successful for tailoring the properties of semiconductors, a transfer of this technique to oxides is challenging due to
the higher complexity of the latter material. Two general approaches can be distinguished, which include overvalent doping (i.e. replacing a lattice ion with a species
of higher charge state) and undervalent doping where an electron-poor ion substitutes the original species. While the dopants have donor character in the rst case
(i.e. they release electrons to suitable adsorbates), they act as electron acceptors in
the latter by creating hole states in the oxide that stimulate electron transfer out of
the adsorbates. In both scenarios, the localization of excess charges in the admetal might lead to distinct changes of its geometric and electronic properties, in a
similar fashion as described for thin oxide lms before.
The impact of doping on the growth geometry of metal particles has been demonstrated for CaO lms doped with Mo impurities in the 1%2% range [9]. After dosing
gold onto the atomically at CaO(001) surface, a perfectly 2-D growth regime was
detected with the metal spreading into large monolayer high islands on the oxide support (Figure 4.4.3). This unusual growth behavior is not observed on nondoped
lms, where tall 3-D Au deposits develop as expected from the weak interaction
between the wide-gap insulator and the ad-metal. The observed crossover in growth
335
E
CaO
CaOLi>Mo
D
CaOMo>Li
CaOMo
0.0
0.2
Aspect Ration
0.4
Figure 4.4.3 STM images of 0.5 monolayer (ML) Au deposited onto (A) pristine CaO,
(B) CaO doped with 4% Mo, (C) CaO doped with 4% Mo plus 2% Li, and (D) CaO doped
with 4% Mo plus 8% Li (6.0 V, 50 50 nm2). Whereas the presence of Mo donors changes
the Au morphology to 2-D, codoping with Li reinstalls the original 3-D growth mode. Note
that monolayer Au islands in (C) appear as depressions, as their conductance is lower than
the one of bare CaO(001). (E) Histogram of aspect ratios for Au particles grown on the differently doped CaO lms. The low aspect ratios observed for Mo-rich samples correspond to a
wetting growth of gold, as induced by electron transfer from the donor-type impurities.
336
compensation of the two dopants, in which the excess electrons of the donors are
trapped by the hole states inserted by the acceptor species [51]. As a consequence,
no electron transfer into the Au ad-metal takes place anymore for the codoped
CaO lms, and the Au deposits adopt the original 3-D equilibrium shape that is typical for pristine CaO (Figure 4.4.3C, D). Apparently, the desired doping effect gets
cancelled out if dopants with opposite charge state coexist in the oxide host.
Interestingly, not every overvalent dopant is able to donate charges into an admetal. Cr ions embedded in an MgO matrix, for instance, are inappropriate electron
donors, and the Au growth morphology is hardly affected in this case [52]. The reason is that Cr in the MgO lattice cannot be stabilized in the 2+ charge state because it
prefers a high-spin conguration and consequently puts electrons into both the lowlying t2g and the high-lying eg state manifold. The latter already overlaps with the
MgO conduction band, so that the electrons move away from the dopant and
become trapped in other oxide defects, preferentially in Mg vacancies sites [53].
The Cr adopts a 3+ charge state right from the beginning and cannot donate further
charges to surface species. An active dopant therefore needs to fulll two requirements. On the one hand, it should occur in two stable charge congurations in the
oxide host. On the other, the top-most occupied state needs to be high enough in
energy to render charge transfer into an adsorbate afnity level energetically feasible.
The same considerations also hold for acceptor-like dopants, only that the hole state
produced upon doping localizes in an adjacent oxygen ion. It should be noted that
holes in the oxygen 2p states are intrinsically unstable against charge transfer from
competing electron sources, such as electron-rich O defects (Fo centers) or donortype molecules that are always present on oxide surfaces (water, hydrogen) [54, 55].
Hole doping as a means to tailor the properties of metal ad-particles is hence more difcult to realize than electron doping with donor-type impurities.
In conclusion, the doping of oxide materials opens promising new routes to change
the morphology and electronic properties of supported metal particles as used in heterogeneous catalysis. Thin oxide lms are ideally suited to elucidate such doping effects, as they can be explored by means of conventional surface science techniques at
a fundamental level. The identied mechanisms can be transferred to real catalysts
later, as the doping approach is not based on specic thin-lm effects.
337
preparation procedures (e.g. impregnation) make use of metal complexes as the precursor, primarily in the form of salts dissolved in aqueous solution, which adsorb at
the support-solution interface in the initial step [58, 59]. The pH of the solution is one
of the important parameters, as it controls the interaction of precursor complexes
with the support by determining the speciation of the solution complexes as well
as the surface charge of the support. Based on the knowledge of the pH-dependent
precursor speciation in solution on one hand, and the properties of the support
(point of zero charge, PZC, density of hydroxyl groups) on the other, this interaction
can be described phenomenologically by complexation models [60]. The transformation of the adsorbed precursor complex into the catalytically active compound then
usually involves a calcination step followed by reduction. The performance in a catalytic test reaction is the ultimate criterion for the usefulness of a certain preparation
procedure. While prereaction, in situ, and postreaction characterization of the catalyst provides relevant information about the active sites in the catalytic reaction, initial nucleation and decomposition of the adsorbed precursor into metal nanoparticles
are difcult to track experimentally with powder catalysts and remain largely
unknown aspects of catalyst preparation. As will be shown in the following, singlecrystalline oxide thin lms are ideally suited for studying these processes, provided
they are stable at typical impregnation conditions.
We have chosen to use Fe3O4(111) lms grown on Pt(111) for studying the
deposition of palladium from a PdCl2 precursor. Fe3O4(111) lms with a thickness
of ~10 nm were prepared in a UHV chamber following a recipe described in the literature [61]. The surface of the lms was subsequently covered by precursor solution,
which was prepared from an acidic (5 M HCl) PdCl2 stock solution by dilution to the
desired concentration and pH adjustment with NaOH. Following washing with
water and drying under a stream of He, the sample was then subjected to thermal
treatment in UHV to transform the adsorbed Pd precursor into nanoparticles.
Shown in Figure 4.4.4 are STM images (taken in air) of Fe3O4(111)-supported Pd
particles formed by impregnation with 5 mM Pd precursor solutions of different pH
ranging from 1.3 to 10 and after a nal thermal treatment at 600 K in UHV. In the
acidic pH range (pH 1.3 to pH 4.7), we observe a substantial variation of Pd coverage, particle size, and particle dispersion. At the lowest pH, only a few Pd particles
with a diameter of 1 nm are present (note that after impregnation, the surface of the
Fe3O4[111] lm is affected by heating leading to the observation of meandering features, most prominently seen for pH 1.3; on top of this surface, the Pd particles
appear as small spots). As the pH of the impregnation solution is changed to
1.6 and 2.5, higher Pd coverage is obtained with a slight increase of particle diameter,
however, with still uniform size distribution. The behavior in this pH range reects
the results of typical uptake curves [62] for negatively charged precursor complexes
(here PdCl42) on positively charged support surfaces (PZC[Fe3O4] 8). The retardation of Pd uptake in strongly acidic medium may be ascribed to (i) competitive
adsorption of chloride [63] or (ii) a decrease of the adsorption equilibrium constant
at high ionic strength [64]. As the pH is increased to 4.7, the very unfavorable situation of uncontrolled deposition of large and nonuniformly sized Pd particles occurs.
Such deposition characteristic has been observed with various near-neutral pH
precursor solutions and can be explained by deposition of solution-precipitated
338
pH 1.3
pH 2.5
pH 4.7
pH 10
Figure 4.4.4 STM images (100 100 nm2) of Pd particles formed on Fe3O4(111)/Pt(111) thin
lms by impregnation with Pd precursor solution of different pH prepared from PdCl2,aq and
subsequent annealing in UHV to 600 K.
Pd particles. At alkaline pH (pH 10, Figure 4.4.1) the Pd particle size distribution
in the nal catalyst is again uniform. However, compared to the acidic pH range,
a signicantly higher Pd loading is achieved.
A more detailed account on the nucleation process of Pd nanoparticles from
adsorbed precursor complexes is given in Fig. 4.4.5, which compares STM and
X-ray photoelectron spectroscopy (XPS) results obtained at different annealing temperatures after impregnation of Fe3O4(111)/Pt(111) with precursor solutions of
pH 1.3 (15 mM Pd2+) and pH 10 (2 mM Pd2+), respectively. The combination of
STM and XPS allows the thermal evolution of morphological features with the
chemical nature of the precursor species to be directly correlated. In the case of acidic
precursor solution (Figure 4.4.5A, [10]), the STM image obtained directly after removing the precursor solution without additional heating (room temperature, RT)
shows a low density of particles on a seemingly clean Fe3O4 substrate. As the temperature is increased to 390 K, the particle density is increased, and at a nal temperature of 600 K, particles uniformly cover the Fe3O4 surface (Figure 4.4.5A). Note
that the particle size (2 nm) does not change from RT to 600 K, showing that
600 K is well below the onset of Ostwald ripening under the present experimental
conditions. The corresponding Pd 3d XPS spectra (Figure 4.4.5B) reveal the presence
of two Pd species with binding energies (BEs) of 337.8 and 336.2 eV (Pd 3d5/2) on the
Fe3O4 sample directly after removing the pH 1 solution. With increasing temperature, the Pd component at higher BE is gradually transformed into the lower BE
component. This behavior is perfectly in line with the STM observations and allows
the high BE component to be assigned to adsorbed Pd precursor complexes, which
cannot be resolved with STM. The low BE component, on the other hand, correlates
with the emergence of nucleated particles at increasing annealing temperature. The
chemical identity of the surface species may be inferred from knowledge of the solution speciation of Pd. It is safe to assume that in pH 1.3 (HCl) solution, Pd is present
as PdCl42 complex. The high Pd 3d5/2 BE obtained for the adsorbed species at
RT (337.8 eV, compared to 335.2 eV for bulk Pd) is in line with previous observations [65, 66] and can, therefore, be attributed to adsorbed PdCl42 or aquochloro
complexes of the form Pd(H2O)xClyn. According to STM results shown in Figure
4.4.5A, these adsorbed precursor complexes are thermally decomposed into Pd particles. At a nal annealing temperature of 600 K, the BE of the particles is 335.7 eV,
which is signicantly higher than expected for metallic Pd particles. XPS analysis
B) pH 1.3, XPS
RT
Pd 3d
C) pH 10, STM
339
D) pH 10, XPS
RT
Pd 3d
337.8
337.8
336.2
RT
336.5
390 K
336.0
390 K
337.1
250
390 K
RT
335.9
390 K
335.7
335.4
600 K
600 K
5000
600 K
600 K
350
340
335
345
binding energy/eV
330
350
345
340
335
binding energy/eV
330
Figure 4.4.5 STM and XPS results for Pd deposited on Fe3O4(111)/Pt(111) surfaces from
suggests that this BE shift results from remaining chlorine adsorbed on the Pd
particles rather than from a particle size effect [10].
A slightly different behavior is observed on the Fe3O4 sample contacted with
pH 10 precursor solution (Figure 4.4.5C and 4.4.5D, [67]). In the initial state, no particles are observed, suggesting that the surface is uniformly covered by adsorbed
precursor. Only moderate drying at 390 K leads to the formation of small particles
in the size range 12 nm covering the entire surface. The corresponding XPS spectrum
shows the dominant abundance of a single Pd species with a Pd 3d5/2 BE of 335.9 eV.
Signicant particle sintering (average diameter 45 nm) occurs upon further annealing
with a concomitant shift of the Pd 3d5/2 component to 335.4 eV. The different nucleation behavior observed on the pH 10 sample as compared to the pH 1.3 case is a result
of the different speciation of Pd complexes. At pH 10, the adsorbed Pd species are hydroxo complexes. Their thermal decomposition into Pd nanoparticles proceeds via the
formation of PdO particles as an intermediate step (observed at 390 K).
Based on the results presented in Figure 4.4.4 and Figure 4.4.5, the deposition of
Pd on single-crystalline Fe3O4(111) thin lms from aqueous precursor solutions can
be divided into three different regimes: electrostatic adsorption of PdCl42, precipitation, and adsorption of Pd hydroxide. In the acidic pH range (< pH 3), the interaction is controlled by electrostatic adsorption of the negatively charged precursor
340
complex (PdCl42) and the positively charged surface. If we apply this so called
strong electrostatic adsorption concept to the present case we expect the following
behavior: A minimum uptake at the PZC of the oxide is predicted, followed by a
maximum below the PZC, and, again, a small uptake at small, i.e. acidic, pH [60].
The latter case, which is usually ascribed to the effect of high ionic strength, has
been observed here. In this regime, small and uniformly sized Pd particles are formed
by thermal decomposition of the adsorbed precursor. The regime of maximum
adsorption due to strong electrostatic interaction could not be reached because uncontrolled deposition of precipitated Pd (most probably PdCl2) species sets in at
near-neutral pH. Shifting the solution pH into the basic range (pH 10) leads again
to the formation of Pd particles with uniform size distribution after decomposition
of the precursor. Strong electrostatic adsorption is not possible in this regime because
both the solution complexes (Pd[OH]42) and the oxide surface are negatively charged.
A chemical interaction between precursor and support (e.g. hydrolytic adsorption of
Pd hydroxide) is more likely the dominant interaction at the support-solution interface
at basic pH.
In summary, the way is paved to look at oxide-supported metal nanoparticles, prepared in solution, and to understand the formation of MNPs through calcination and
reduction. However, there is still a way to go to identify the elementary steps in the
interaction of the species from solution at the solid-liquid interface. Of course, this is
what we really want.
5 nm
341
FeO(111)
Pt
Fe3O4(111)
Fe
O
Pt(111)
Figure 4.4.6 (A) Typical STM image of Pt deposited onto Fe3O4(111) lm and subsequently
annealed in UHV at 850 K. Atomically resolved STM image of the top facet is shown in the
inset. (B) The cross view of the FeO(111)/Pt(111) interface and schematic presentation of the Pt
particles encapsulated by the FeO(111) overlayer.
at top facets, which are Pt(111) in nature, owing to the epitaxial relationships
between Fe3O4(111) and Pt(111) [77]. It has turned out, however, that hightemperature annealing leads to a dramatic attenuation of the CO uptake compared
to the samples annealed at 600 K, which cannot be assigned solely to Pt sintering
[77, 78]. This behavior is, in fact, a classical manifestation of the SMSI [68].
High-resolution STM study showed that the top facets exhibit the hexagonal lattice of protrusions with a ~3 periodicity, in turn forming a superstructure with
a ~25 periodicity (see inset in Figure 4.4.6). This structure is well-documented in
the literature on thin iron oxide lms grown on Pt(111) [79] and can unambiguously
be assigned to an FeO(111) lm, which consists of close-packed layers of iron and
oxygen stacked as O-Fe-Pt(111). The Moir superstructure originates from a mismatch between the Pt(111) and FeO(111) lattices. Since FeO(111) lms can be
grown on the Pt(100) surface as well [80], it seems plausible that the encapsulation
by the FeO(111) layer also occurs on the side facets, which, according to the particles
habitus, expose (111) and (100) surfaces.
The mechanism of the encapsulation is still unknown. Nonetheless, the encapsulation implies high adhesion energy between Pt and iron oxide, which could, in principle, be derived from the structural information, obtained by STM on the particle size
and shape, using the modied Wulff construction [81].The analysis yielded an energy
in the range of 3.84.2 J/m2, which is, indeed, considerably larger than those obtained
for Pd particles on Fe3O4(111) and alumina lms (i.e. 3.13.3 J/m2), for which the
encapsulation has not been observed [77]. Note also, that CO adsorption experiments
indicated Fe-Pt surface intermixing with the onset at ca. 600 K [82], probably as the
rst step in the encapsulation.
The well-dened Pt/Fe3O4(111) systems were examined in the CO oxidation reaction at near-atmospheric pressures and relatively low temperatures (~450 K) [84].
Figure 4.4.2A shows kinetic curves of CO2 production under O-rich conditions
(e.g. 10 mbar CO + 50 mbar O2, He balance to 1 bar) over two samples possessing
the same amount of Pt, but annealed either at 600 K (i.e. exposing clean Pt surface)
or 850 K (i.e. encapsulated by FeO) prior to the reaction. The results for pristine
342
Fe3O4(111) lms and clean Pt(111) under the same conditions are also shown, for
comparison.
It is clear that the encapsulated Pt particles exhibit higher reactivity than the clean
Pt particles. The difference must be even higher if one normalizes the reactivity to the
particles surface area, which obviously decreases at 850 K due to particle sintering,
albeit not measured in those experiments. The same effect was observed also at the
stoichiometric ratio (40 mbar CO + 20 mbar O2) [83]. Such promotional effect of encapsulation seems counterintuitive since the FeO lm covering Pt particles and exposing a close-packed O layer must be essentially inert. In order to rationalize
these ndings on highly dispersed systems, we have to address the structure-reactivity
relationships observed for extended, well-ordered FeO(111) lms on Pt(111).
The FeO(111) lm is, indeed, extremely stable and chemically inert under conditions typically used in UHV-based experiments. However, the situation changes
dramatically in the mbar range of pressures. At low temperatures studied here
(400450 K), Pt(111) is inactive in CO oxidation due to the well-known blocking
effect of CO on O2 dissociation. The nanometer-thick Fe3O4(111) lms shows
some activity, but it is negligible as compared to ultrathin FeO(111) lms, which
showed an order of magnitude higher reaction rate under the same conditions
(Figure 4.4.7A). Therefore, it is the thin FeO overlayer on Pt that is responsible
for the enhanced reactivity of encapsulated Pt particles in CO oxidation.
The experimental results in combination with DFT calculations provided compelling evidence that at elevated pressures the FeO(111) lm transforms into a different
structure containing an Fe layer sandwiched between two oxygen layers like O-Fe-O
lm (Figure 4.4.7B) [11, 12]. The mechanism for this transformation starts by O2
adsorption on an Fe atom pulled out of the pristine FeO lm. Because of local lowering of the work function by this process, electrons are transferred from the oxide/
metal interface to oxygen, resulting in a transient superoxo species, which dissociates,
thus forming the O-Fe-O structure. It appears, however, that the formation of the
trilayer structure depends on the registry to underlying Pt(111), ultimately resulting
in close-packed islands with a FeO2 stoichiometry (see Figure 4.4.7B) rather than a
continuous O-Fe-O lm. Nonetheless, the topmost O atoms in the resulting FeO2-x
lms are more weakly bound than those in the original FeO layer and readily
react with incoming CO to form CO2, which desorbs and leaves an oxygen vacancy
behind. The overall activation barrier for CO2 formation on the ideal O-Fe-O overlayer, as determined by DFT (~0.3 eV), is considerably lower than the computed
barrier (~1 eV) for the CO oxidation reaction on Pt(111) and as such may explain
higher reactivity of FeO(111)/Pt(111) than pure Pt(111) [12]. Certainly, to end the
catalytic cycle, the oxygen vacancies must be replenished via the reaction with O2
from the gas phase. Recent STM studies provided strong evidence for this mechanism of the Mars-van Krevelen type [84]. Interestingly, NO transforms the FeO
lm into the trilayer lm in the same way as O2. Comparison of the CO + O2 and
CO + NO reactions over the FeO(111)/Pt(111) surface showed that the replenishment
of oxygen vacancies is the rate-limiting step that proceeds much faster with O2 than
NO [85].
It is important to note that both the transformation of the FeO into FeO2-like lm
and the oxygen vacancy replenishment under the reaction conditions involve the
[CO2] production
10 mbar CO + 50 mbar O2
450 k
FeO(111)/Pt(111)
343
ann. 850 k
Pt/Fe3O4(111)
ann. 600 k
Fe3O4(111)
Pt(111)
0
20
40
60
80
Time, min
100
120
5 nm
Pt/Fe3O4(111)
FeO(111)/Pt(111)
CO
CO2
O
Fe
O
Figure 4.4.7 (A) CO2 production over Pt(111), iron oxide lms on Pt(111), and Pt/
Fe3O4(111) annealed to 600 and 850 K. (B) STM images of the encapsulated Pt particle on
Fe3O4(111) and, for comparison, of the FeO(111)/Pt(111) lm, both exposed to 20 mbar O2
at 450 K. The scheme illustrates the reaction mechanism. (See the text.)
charge transfer accompanied by a lattice distortion. Both factors favor the reaction
on ultrathin lms.
Apparently, the same scenario holds true for the encapsulated Pt particles.
Figure 4.4.7B shows STM images of the encapsulated Pt particle on Fe3O4(111)
and, for comparison, of the FeO(111)/Pt(111) lm, both exposed to 20 mbar O2 at
450 K. The close similarities between these two systems with respect to the surface
morphology and reactivity indicate the absence of the material gap, suggesting
that the results and conclusions drawn for extended surfaces can be transferred to
the supported nanoparticles.
In the experiments, presented in Figure 4.4.7A, the initial reaction rate over the
encapsulated particles is almost identical to that measured on the FeO(111) lm
because the particle surface area (including both top and side facets) at the high Pt
344
loadings studied here is close to the surface area of the FeO lm. Whereas the rate
over the FeO lm is almost constant until all CO in the ambient is consumed and
the reaction stops, the reaction slows down over the encapsulated particles indicating
catalyst deactivation. The latter may include carbon deposition, but this issue is
beyond the scope of the present chapter.
This example demonstrates that ultrathin oxide lms may enhance reactivity of
metal catalysts, particularly in oxidation reactions in the low-temperature regime,
where pure metal catalysts may suffer from site-blocking effects and strong chemisorption of reactants. A continuously growing body of studies on reactivity of ultrathin oxide lms leads us to believe in a rational design of monolayer oxidation
catalysts by combining different ultrathin lms oxides with different metals, thereby
controlling the charge transfer.
345
In the past 7 years, we have examined this range of questions for an apparently
simple (though in fact quite complex) system. Since one of the focal points was the
inuence of Mie plasmons, the chosen material was silver. The deposition of Ag
NPs on ultrathin alumina lms (on AlNi alloy surfaces) has been studied, and
their properties have been characterized in detail [89]. The preparation of narrowly
dened particle sizes in the range of 210 nm is possible. These particles possess a
strong plasmonic mode at ~3.5 eV (polarized perpendicular to the surface; the
lower energy parallel mode is screened by the close metal substrate), which has
been observed and characterized by photon STM [89] (see Figure 4.4.8A) and by
two-photon photoemission [90]. The thin alumina lm decouples the Ag NPs quite
efciently electronically from the metal substrate without leading to charge-up, so
that connement effects might be expected. This is in contrast to the situation of
Ag NPs on strongly reduced TiO2 surfaces, where plasmon excitation leads to electronhole pairs in the TiO2 which decay radiatively [91]. NO was chosen as adsorbate
because of the ease of its state-selective detection. Its adsorption on silver surfaces
is, however, more complex than usually observed on (transition) metal surfaces.
Because of its weak interaction with the noble metal silver, the adsorption has to
be done below 80 K; in this range, NO dimers are formed on the surface. On
Ag(111), the resulting adsorption layers have been characterized in great detail
[92], including their photochemistry [93, 94], so that we were able to start on a
good basis. The formation of dimers makes the photochemical reaction channels
more complex: besides breaking of the ON-NO bond and the bond to the surface
simultaneously, which leads to NO desorption into the gas phase (and some NO
left on the surface), the dimers can also react to N2O + O, which stays on the surface
(N2O can be detected by subsequent thermal desorption; adsorbed O leads to a stronger bond of the NO monomer, which desorbs thermally at a much higher temperature and has a much smaller photoreaction cross section than the dimer). Even the
desorption of (extremely fast) N2 molecules has been observed under certain conditions [95]. However, NO desorption is the strongest channel; it has been investigated
in great detail. Besides measuring yields and desorption cross sections, the energy distributions over the translational [14] and internal modes (rotation, vibration) [96] of
the desorbing NO have been measured. The inuences of photon energy and polarization, particle size, and laser pulse duration have been investigated. In all cases,
direct comparison to Ag(111) in the same experimental system and with the same
methods has been made.
In the following, some of the main results are listed:
1. We found that the cross section, , for NO desorption from (NO)2 monolayers on
Ag NPs is indeed strongly enhanced by excitation of the plasmon, which is known
to lie at approximately 3.5 eV (with a weak dependence on NP size [89, 90]). Compared to Ag(111), an enhancement of by up to a factor 40 (depending on NP
size; see item 3 in this list) has been found [14].
2. There is also an enhancement of off the plasmon resonance, albeit weaker,
which we interpret as due to connement of excitations, here of hot electrons in
the Ag NP. The photochemical mechanism of all the processes seen is believed
to involve TNI states [87, 94], which is consistent with detailed characterizations
B
Cross section, X1017 [cm2]
Intensity (a.u.)
346
2.5
V
IV
III
IV
II
III
V
II
I
3
size: 30 30nm
20
Poton energy
15
10
2.3 eV p-pol.
3.5 eV p-pol.
4.7 eV p-pol.
Ag(111)
5
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
1/R [nm1]
Figure 4.4.8
(A) Radiation emitted from individual Ag NPs of various sizes, observed with
the photon STM. Inset: STM micrograph of the NP ensemble investigated. After [90]. (B)
Variation of desorption cross sections, , of NO from (NO)2 monolayers on Ag NPs as a function of mean particle size (top abscissa), for the three photon energies given. The plasmon is
excited at 3.5 eV. The bottom abscissa gives the inverse radius and emphasizes the approximate scaling off the plasmon resonance. After [14].
of the desorbing NO ([96]; see item 4 in this list). This means that a hot electron of
sufcient energy which is excited in an NP by the absorbed light tunnels from the
NP into an adsorbed (NO)2; in the formed TNI, the N-N and N-O bonds are weakened, and the bond to the surface is strengthened by the charge-image charge
interaction, so that the corresponding wave packet is put into areas with potential
gradients in all dimensions. When the electron jumps back into the NP after a very
short residence time, a considerable fraction of the molecules has evolved sufciently along the gradients to lead to breaking of the N-N and the NO-Ag
bonds and to translational, rotational, and vibrational excitation of the NO
that appears in the gas phase.
3. This cross-section enhancement is also size dependent (Figure 4.4.8B, [14]). It generally increases with decreasing NP size (diameter d), which we ascribe to connement of the primary hot electrons in the NP ( ~ 1/d, going with the surface/
volume ratio S/V; see item 2 in this list). For excitation in the plasmon resonance,
a strong maximum of at a certain NP size (here at d ~ 5 nm) is found, which is
explained by two counteracting inuences: on the one hand, the conversion of
plasmons into hot electrons increases with decreasing NP size against the competing radiative decay; on the other hand, the total number of photons collected
decays with the number of atoms in the particle [14, 96].
4. On the other hand, the dynamics of the bond dissociations (i.e. the motion of the
representative wave packet on the potential energy surfaces of ground and excited
states for the various molecular entities) turns out to be always the same. This is
concluded from the fact that the nal state of the desorbing NO stays constant: all
energy distributions, translational, rotational, and vibrational, as well as their correlations, are identical despite the strong variations of cross sections [14, 96]. These
characteristics are compatible with the proposed TNI mechanism (see item 2 in
347
this list). The only exception to this constancy of mechanism is found for excitation with the highest photon energy used (4.7 eV) and the smallest NPs (d < 4 nm).
For these conditions, the mean translational energy of NO was considerably
enhanced in fact, a new, much faster component was found in addition to the
normal one [14, 96]. We have explained this new path by excitation of a transient
positive ion of the NO dimer that becomes accessible at an excitation energy sufcient to produce holes in the Ag d-band or direct excitations from a lled adsorbate state to empty states of the NP. That this path becomes observable only for
very small NPs is due to the fact that its contribution is proportional to the surface
area, while the TNI contribution goes down with the volume [14, 96].
5. All these experiments have been carried out with nanosecond lasers under uences
that lead to linear behavior of the photochemical yield with uence (i.e. the determined cross sections are independent of uence). This means that the excitations
do not interact, even when they are conned in the NP. Very different behavior is
found with femtosecond (~ 100 fs) lasers of even smaller uences, for which the
photons of a pulse arrive in a much shorter time so that the spatiotemporal photon density is increased by at least a factor of 104 [15]. Here the cross section is
considerably enhanced and increases with uence (i.e. the yield has nonlinear
dependence on the uence). With two-pulse measurements, in which one pulse
is split into two and the two are delayed with respect to each other, it is found
that the memory of the system is conned to the time of overlapping pulses. So
there is a very short-lived connement of interacting hot electrons. We have explained this by a repumping of hot electrons within the same pulse [15]. However,
despite this dramatic change of reactivity, it is found again that the mechanism of
desorption, as indicated by NO nal state energies, is unchanged.
To summarize, we nd that nanoparticles induce strong increases of photochemical
cross sections if they possess strong new excitations, in particular of plasmon type.
Also, connement of excitations leads to (more gradual) enhancement of cross sections. If this connement is also temporally constrained additional, nonlinear effects
can be caused. The timescales of the electronic processes (including excitation and
decay of plasmons) are very short, probably in the range below 10 fs (our results
can only impose an upper limit of 100 fs). The dynamics of evolving molecular states,
however, and thus the mechanism, appear essentially unchanged. This makes sense
since these processes occur on much longer timescales. It cannot be excluded that
there are systems in which the timescales can mix and, for example, the plasmon excitation has a direct photochemical inuence; in fact, we have seen one such case [97],
but we consider this as a probably rare, somewhat exotic case. While these results
have been obtained with a particular system, we believe that these main conclusions
will be fairly general, since they do not depend on any particular system properties.
As to the messages that might be carried over to practical photochemical processes, the main emphasis of the use of NPs should lie on the improved light harvesting made possible by plasmon excitations. The fact that plasmon resonances are
fairly narrow, even if variations with particle size are utilized by using broadly varying ensembles, limits the usable part that can be caught of a broadband source like
sun light. If one aims at photochemical reactions that require a certain energy, it will
348
be important to use NPs with plasmons in the required energy range. The additional
effects offered by connement will depend more strongly on the particular system
since it is strongly inuenced by the photochemical mechanism, which will not
always be describable by the TNI process. Generally, we do not expect that particular new mechanisms will open up on NPs. The subject of photochemistry at nanoparticles is still in its infancy and deserves to be investigated more fully on a broader
front, both for model systems and practical setups.
4.4.6 Synopsis
Via four case studies, this chapter has demonstrated how model studies on complex
materials related to heterogeneous catalysis may help to unravel an atomistic view of
processes at solid-vacuum, solid-gas, and solid-liquid interfaces. Going beyond metal
single-crystal surfaces toward modeling some of the real complexity of catalytic materials is an important step forward to help designing energy-efcient catalysts by
providing information on principles more than on specic systems. We are approaching a situation where the interplay between work on real catalytic material and work
on specically designed catalyst models may lead to a detailed understanding of the
relation between structure-morphology and catalytic activity and selectivity.
4.4.7 References
1. Renn J, Schlgl R, Zenner H-P. Herausforderung Energie. Berlin: Max Planck Research
Library for the History and Development of Knowledge, Proceedings 1; 2011.
2. Behrens M, Schlgl R. Energie ist Chemie Katalyse als Schlsseltechnik. Energie von
Morgen: Eine Momentaufnahme. In: Renn J, Schlgl R, Zenner H-P, editors. Herausforderung Energie. Berlin: Max Planck Research Library for the History and Development of
Knowledge, Proceedings 1); 2011. p. 16374.
3. Freund H-J. Model studies in heterogeneous catalysis. Chem Eur J. 2010;16:938497.
4. Shaikhutdinov, S, and Freund, H-J. Ultrathin oxide Films on metal supports: structurereactivity relations. Ann Rev Phys Chem. 2012;63:61933.
5. Freund H-J, Pacchioni G. Oxide ultra-thin lms on metals: new materials for the design of
supported metal catalysts. Chem Soc Rev. 2008;37:222442.
6. Freund H-J. Clusters and islands on oxides: from catalysis via electronics and magnetism
to optics. Surf Sci. 2002;500:27199.
7. Duke CB. Surface science: the rst thirty years. Surf Sci. 1994;299/300:viiviii.
8. Freund H-J, Goodman DW. Ultrathin oxide lms. In: Ertl G, Knzinger H, Schth F,
Weitkamp J, editors. Handbook of Heterogeneous Catalysis. Weinheim: Wiley-VCH
Verlagsgesellschaft GmbH; 2008. p. 130938.
9. Shao X, Prada S, Giordano L, Pacchioni G, Nilius N, Freund H-J. Tailoring the shape of
metal ad-particles by doping the oxide support. Angew Chem Int Ed. 2011;50:115257.
10. Wang H, Ariga H, Dowler R, Sterrer M, Freund H-J. Surface science approach to supported metal catalyst preparation Pd deposition onto thin Fe3O4(111) lms from
PdCl2 precursor. J Catal. 2012;286:15.
11. Sun YN, Qin ZH, Lewandowski M, et al. Monolayer iron oxide lm on platinum
promotes low temperature CO oxidation. J Catal. 2009;266:35968.
4.4.7 References
349
12. Sun YN, Giordano L, Goniakowski J, et al. The interplay between structure and CO
oxidation catalysis on metal-supported ultrathin oxide lms. Angew Chem Int Ed.
2010;49:441821.
13. Watanabe K, Menzel D, Nilius N, Freund H-J. Photochemistry on metal nanoparticles.
Chem Rev. 2006;106:430120.
14. Mulugeta D, Kim KH, Watanabe K, Menzel D, Freund H-J. Size effects in thermal and
photochemistry of (NO)2 on Ag nanoparticles. Phys Rev Lett. 2008;101:146103-1-4.
15. Kim KH, Watanabe K, Mulugeta D, Freund H-J, Menzel D. Enhanced photoinduced
desorption from metal nanoparticles by photoexcitation of conned hot electrons using
femtosecond laser pulses. Phys Rev Lett. 2011;107:047401.
16. Molina LM, Hammer B. Some recent theoretical advances in the understanding of the
catalytic activity of Au. Appl Catal A. 2005;291:2131.
17. Green IX, Tang WJ, Neurock M, Yates JT. Spectroscopic observation of dual catalytic
sites during oxidation of CO on a Au/TiO2 catalyst. Science. 2011;333:7369.
18. Abbet S, Riedo E, Brune H, et al. Identication of defect sites on MgO(100) thin lms by
decoration with Pd atoms and studying CO adsorption properties. J Am Chem Soc.
2001;123:61728.
19. Ekardt W. Metal clusters. Chichester: John Wiley; 1999.
20. Kreibig U, Vollmer W. Optical properties of metal clusters. Berlin: Springer-Verlag;
1995.
21. Risse T, Shaikhutdinov S, Nilius N, Sterrer M, Freund H-J. Gold supported on thin oxide
lms: from single atoms to nanoparticles. Acc Chem Res. 2008;41:94956.
22. Hashmi ASK, Hutchings GJ. Gold catalysis. Angew Chem Int Ed. 2006;45:7896936.
23. Herzing AA, Kiely CJ, Carley AF, Landon P, Hutchings GJ. Identication of active gold
nanoclusters on iron oxide supports for CO oxidation. Science. 2008;321:13315.
24. Valden M, Lai X, Goodman DW. Onset of catalytic activity of gold clusters on titania
with the appearance of nonmetallic properties. Science. 1998;281:164750.
25. Chen MS, Goodman DW. The structure of catalytically active Au on titania. Science.
2004;306:2525.
26. Mostafa S, Behafarid F, Croy JR, et al. Shape-dependent catalytic properties of Pt
nanoparticles. J Am Chem Soc. 2010;132:157149.
27. Benedetti S, Myrach P, di Bona A, Valeri S, Nilius N, Freund H-J. Growth and morphology of metal particles on MgO/Mo(001): a comparative STM and diffraction study. Phys
Rev B. 2011;83:125423 (110).
28. Benia HM, Nilius N, Freund HJ. Effect of electromagnetic interactions on plasmon
excitations in silver particle ensembles. Surf Sci. 2006;600:12833.
29. Brown M, Fujimori Y, Ringleb F, et al. Oxidation of Au by surface OH-nucleation
and electronic structure of gold on hydrohylated MgO(001). J Am Chem Soc. 2011;133:
1066810676.
30. Goniakowski J, Noguera C. Electronic states and Schottky barrier height at metal/MgO
(100) interfaces. Interface Sci. 2004;12:93103.
31. Goniakowski J, Noguera C. Polarity of oxide surfaces and nanostructures. Rep Prog Phys.
2008;71:016501 (155).
32. Koplitz LV, Dulub O, Diebold U. STM study of copper growth on ZnO(0001)Zn and
ZnO(0001)O surfaces. J Phys Chem B. 2003;107:1058390.
33. Pacchioni G, Giordano L, Baistrocchi M. Charging of metal atoms on ultrathin MgO/Mo
(100) lms. Phys Rev Lett. 2005;94:226104 (14).
34. Sterrer M, Risse T, Heyde M, Rust H-P, Freund H-J. Crossover from three-dimensional
to two-dimensional geometries of Au nanostructures on thin MgO(001) lms: a conrmation of theoretical predictions. Phys Rev Lett. 2007;98:2061034.
350
4.4.7 References
351
56. Campbell CT. Ultrathin metal lms and particles on oxide surfaces: structural, electronic
and chemisorptive properties. Surf Sci Rep. 1997;27:1111.
57. Fu Q, Wagner T. Interaction of nanostructured metal overlayers with oxide surfaces. Surf
Sci Rep. 2007;62:43198.
58. Ertl G, Knzinger H, Schth F, Weitkamp J. Handbook of heterogeneous catalysis.
Weinheim: Wiley-VCH; 2008.
59. de Jong KP. Synthesis of solid catalysts. Weinheim: Wiley-VCH; 2009.
60. Regalbuto JR, Navada A, Shadid S, Bricker ML, Chen Q. An experimental verication of
the physical nature of Pt adsorption onto alumina. J Catal. 1999;184:33548.
61. Weiss W, Ritter M. Metal oxide heteroepitaxy: Stranski-Krastanov growth of iron oxides
on Pt(111). Phys Rev B. 1999;59:520113.
62. Hao X, Spieker WA, Regalbuto JR. A further simplication of the revised physical
adsorption (RPA) model. J Colloid Interface Sci. 2003;267:25964.
63. Olsbye U, Wendelbo R, Akporiaye D. Stucy of Pt/alumina catalyst preparation. Appl
Catal A. 1997;152:12741.
64. Spieker WA, Regalbuto JR. A fundamental model of platinum impregnation onto
alumina. Chem Eng Sci. 2001;56:3491504.
65. Bozon-Verduraz F, Omar A, Escard J, Pontvianne B. Chemical state and reactivity of supported
palladium: I. characterization by XPS and UV-visible spectroscopy. J Catal. 1978;53:126134.
66. Fleisch TH, Hicks RF, Bell AT. An XPS study of metal-support interactions on PdSiO2
and PdLa2O3. J Catal. 1984;87:398.
67. Wang, H-F, Kaden, WE, Dowler, R, Sterrer, M, and Freund, H-J. Model oxidesupported metal catalysts comparison of ultrahigh vacuum and solution based preparation of Pd nanoparticles on a single-crystalline oxide substrate. Phys Chem Chem Phys.
2012;14:1152511533.
68. Tauster SJ. Strong metal-support interactions. Acc Chem Res. 1987;20:38994.
69. Tauster SJ, Fung SC, Garten RL. Strong metal-support interactions. Group 8 noble
metals supported on titanium dioxide. J Am Chem Soc. 1978;100:1705.
70. Ko EI, Garten RL. Ethane hydrogenolysis studies of TiO2-supported group VIII metal
catalysts. J Catal. 1981;68:2336.
71. Haller GL, Resasco DE. Metal-support interactions between Group VIII metals and
reducible oxides. Adv Catal. 1989;36:173.
72. Bernal S, Calvino JJ, Cauqui MA, et al. Some recent results on metal/support interaction
effects in NM/CeO2 (NM: noble metal) catalysts. Catal Today. 1999;50:175206.
73. Dulub O, Hebenstreit W, Diebold U. Imaging cluster surfaces with atomic resolution: the
strong metal-support interaction state of Pt supported on TiO2(110). Phys Rev Lett.
2000;84:3646.
74. Bowker M, Stone P, Morrall P, et al. Model catalyst studies of the strong metalsupport
interaction: surface structure identied by STM on Pd nanoparticles on TiO2(110).
J Catal. 2005;234:17281.
75. Barcaro G, Agnoli S, Sedona F, Rizzi GA, Fortunelli A, Granozzi G. Structure of reduced
ultrathin TiOx polar lms on Pt(111). J Phys Chem C. 2009;113:57219.
76. Netzer FP, Allegretti F, Surnev S. Low-dimensional oxide nanostructures on metals:
hybrid systems with novel properties. J Vac Sci Technol B. 2010;28:116.
77. Qin ZH, Lewandowski M, Sun YN, Shaikhutdinov S, Freund HJ. Encapsulation of
Pt nanoparticles as a result of strong metal-support interaction with Fe3O4(111). J Phys
Chem C. 2008;112:1020913.
78. Qin ZH, Lewandowski M, Sun YN, Shaikhutdinov S, Freund HJ. Morphology and CO
adsorption on platinum supported on thin Fe3O4(111) lms. J Phys Condens Matter.
2009;21:134019.
352
79. Weiss W, Ranke W. Surface chemistry and catalysis on well-dened epitaxial iron-oxide
layers. Prog Surf Sci. 2002;70:1151.
80. Shaikhutdinov S, Ritter M, Weiss W. Hexagonal heterolayers on a square lattice: a
combined STM and LEED study of FeO(111) on Pt(100). Phys Rev B. 2000;62:753541.
81. Hansen KH, Worren T, Stempel S, et al. Palladium nanocrystals on Al2O3: structure and
adhesion energy. Phys Rev Lett. 1999;83:41203.
82. Sun YN, Qin ZH, Lewandowski M, Shaikhutdinov S, Freund HJ. CO adsorption and
dissociation on iron oxide supported Pt particles. Surf Sci. 2009;603:3099103.
83. Lewandowski M, Sun YN, Qin ZH, Shaikhutdinov S, Freund HJ. Promotional effect of
metal encapsulation on reactivity of iron oxide supported Pt catalysts. Appl Catal A.
2011;391:40710.
84. Lewandowski M, Groot IMN, Shaikhutdinov S, Freund HJ. Scanning tunneling microscopy evidence for the Mars-van Krevelen type mechanism of low temperature CO oxidation on an FeO(111) lm on Pt(111). Catal. Today. 2012;181:5255.
85. Lei Y, Lewandowski M, Sun Y-N, et al. CO+NO versus CO+O2 reaction on monolayer
FeO(111) lms on Pt(111). ChemCatChem. 2011;3:6714.
86. Dai H-L, Ho W. Laser spectroscopy and photochemistry on metal surfaces. Singapore:
World Scientic; 1995.
87. Zimmermann FM, Ho W. State resolved studies of photochemical dynamics at surfaces.
Surf Sci Rep. 1995;22:127247.
88. Menzel D. Electronically induced surface chemistry: localised bond breaking versus
delocalisation. Surf Interface Anal. 2006;38:170211.
89. Nilius N, Ernst N, Freund HJ. Photon emission spectroscopy of individual oxidesupported silver clusters in a scanning tunneling microscope. Phys Rev Lett. 2000;84:39947.
90. Evers F, Rakete C, Watanabe K, Menzel D, Freund H-J. Two-photon photoemission
from silver nanoparticles on thin alumina lms: role of plasmon excitation. Surf Sci.
2005;593:438.
91. Nilius N, Ernst N, Freund H-J. On energy transfer processes at cluster-oxide interfaces:
silver on titania. Chem Phys Lett. 2001;349:3517.
92. Carlisle CI, King DA. Direct molecular imaging of NO monomers and dimers and a
surface reaction on Ag(111). J Phys Chem B. 2001;105:388693.
93. Vondrak T, Burke DJ, Meech SR. The dynamics and origin of NO photodesorbed from
NO/Ag(111). Chem Phys Lett. 2000;327:13742.
94. So SK, Franchy R, Ho W. Photodesorption of NO from Ag(111) and Cu(111). J Chem
Phys. 1991;95:138599.
95. Kim KH, Watanabe K, Menzel D, Freund H-J. Photoinduced abstraction reactions
within NO dimers on Ag(111). J Am Chem Soc. 2009;131:16601.
96. Mulugeta D, Watanabe K, Menzel D, Freund H-J. State-resolved investigation of the
photodesorption dynamics of NO from (NO)2 on Ag nanoparticles of various sizes in
comparison with Ag(111). J Chem Phys. 2011;134:164702 (111).
97. Watanabe K, Kim KH, Menzel D, Freund H-J. Hyperthermal chaotic photodesorption
of xenon from alumina-supported silver nanoparticles: plasmonic coupling and plasmoninduced desorption. Phys Rev Lett. 2007;99:225501.
4.5.1 Introduction
The leading theme of this volume is to document the state of energy research in
chemistry. It is evident that the challenges associated with renewable, sustainable,
and scalable energy reactions touch all elds of chemistry, ranging from the most
fundamental aspects of mechanistic catalysis to intricate questions of large scalechemical engineering and process design. The efcient and reversible storage and
release of energy in chemical bonds is indeed an essential aspect, and a great challenge, in energy research. Ideally, such energy-conserving chemistry follows the capture of photon energy provided by sunlight. In this respect, the most impressive
chemistry of renewable energy occurs in green plants, algae, and photosynthetic bacteria, all of which convert CO2 and H2O into O2 and energy-rich sugar molecules
under the action of sunlight [1].
The relevant processes all involve the activation of small, largely inert molecules.
The most important elementary processes are summarized by the following six
reactions:
2 H+ + 2e H2
(1)
2 H2O O2 + 4 H+ + 4e
(2)
O2 + 4 H+ + 4e 2 H2O
(3)
CO2 + 2 H+ + 2e HCOOH
(4)
O2 H3COH
(5)
CH4 +
1
2
N2 + 6 H+ + 6e 2 NH3
(6)
Reaction (1) is one of the most elementary chemical reactions and is of central importance for energy research as it provides H2, which either serves as a fuel by itself or
acts as a precursor for further reduction reactions (e.g. in the activation of O2, CO2,
and N2 reactions [3], [4], and [6]). Ideally, the protons and electrons required for
354
reaction (1) are provided by reaction (2), the oxidation of water. The latter process
has been found to be very difcult to achieve in chemical catalysis, and its technological realization by homogenous, heterogeneous, or electrochemical means is of fundamental importance for energy research. The oxidation chemistry of methane (or
higher reduced hydrocarbons, reaction [5]) is a paradigm for the conversion of biomass to fuel. It involves the activation of nonactivated C-H bonds in a controlled
manner, again a process that is exceedingly difcult to achieve in catalysis. Finally,
the activation of the triple bond of dinitrogen (reaction [6]) is a pivotal step of fundamental importance for the large-scale production of fertilizer and consequently
feedstock, and thus it is also an integral part of energy research.
The chemistry underlying reactions (1)(6) has, of course, been intensely studied in
various elds of chemistry. Most of these reactions can be achieved with varying degrees of efciency in biological, homogenous, or heterogeneous catalysis. Of these
elds, nature is unsurpassed in terms of catalyzing all of these reactions with stunning
efciency and selectivity in enzymatic catalysis. In addition, in all cases, nature utilizes highly abundant transition metals in the enzyme active site in order to carry out
these processes. Any chemistry that aims at scalability must also be based on abundant elements, which basically restricts the relevant transition metals to the rst transition row and molybdenum (which is also sufciently abundant in nature that it
could be used for large-scale chemical processes). Many heterogeneous or homogenous processes are based on noble metals from the second or third row. While
many impressive chemical conversions have been achieved on the basis of these metals, they will not be covered in this chapter, as these processes will eventually fall
short of the requirement of scalability.
It seems clear that progress toward the efcient, selective, scalable, sustainable,
and cost-effective catalysis of reactions (1)(6) will require a concerted effort by
the chemical community, as well as strong synergy with neighboring disciplines.
The fundamental basis for rational process design is a thorough understanding of
the chemical principles that involve the key reactions of energy research.
The aim of this book is to summarize the state of the art of the efforts within individual chemical subdisciplines toward achieving the common goals associated with
energy research. By bringing together the various communities, this will hopefully
develop into an internal joint research effort that benets from cross-fertilization. In
this chapter, we will focus on the biological and molecular aspects of energy research.
Obtaining a basic understanding of the chemistry of natural systems has certainly
been a benecial outcome of research in the eld of biological catalysis. However, to
date, it has not been possible to design biomimetic catalysts that resemble the natural
active sites and have similar activity. In fact, enzyme active sites have evolved to a
very high degree of sophistication in which the placement of individual hydrogen
bonds or active-site water molecules is absolutely essential for catalytic activity
and selectivity. If one aims at the construction of low-molecular-weight catalysts
that reproduce all of the geometric features of enzyme active sites, the result
would very likely be an extremely complex synthesis that would produce molecules
that are as fragile, or even more fragile, than the enzymes themselves. The design
of chemically highly complex ligands counters the requirement of scalability, as eventually these catalysts must be produced in kilogram quantities. Thus, it appears to be
355
important to take mechanistic inspiration from studying enzyme active sites, while
judiciously searching for reactions that utilize readily available base metals within
simple ligand frameworks.
Understanding reactions (1)(6), catalyzed by rst-row transition metals, is a complex undertaking. The chemistry of the rst transition row is particularly difcult as
most of the ions involved can exist in multiple redox and spin states. Hence, in understanding these reactions, input from theoretical chemistry is essential, and there are
ample opportunities for the interaction between theory and experiment. In the next
section, we describe a research strategy that emphasizes the interaction between
high-level spectroscopy and quantum chemistry. This approach has been found to
be particularly fruitful in bioinorganic chemistry and catalysis, and we expect that
it will prove equally useful in facing the many challenges associated with energy
research.
356
357
358
Presently, the most common method for the production of molecular hydrogen on
an industrial basis is steam reforming, a process in which steam is allowed to react
with fossil fuels at high temperatures, according to the following reaction:
CH4 + H2O CO + 3 H2
The energy required (i.e. the enthalpy change H) for methane steam reforming, for
example, is +49 kcal/mol [25]. A second well-established and widely applied method
for hydrogen production, introduced during the early days of electrochemistry in the
1800s and that has recently become commercially available, concerns electrolysis of
water,
2 H2O 2 H2 + O2
which requires +116 kcal/mol. For comparison, the bond dissociation energy of H2
is +104 kcal/mol [25]. These numbers indicate that the production of molecular
hydrogen, the key ingredient of the hydrogen economy, is by no means a trivial
task, and that the presently used industrial processes to produce molecular hydrogen
will in the long term likely not be sustainable.
(7)
The family of hydrogenases is divided into three classes, depending on the metal content of the active site [26]. The classes comprise [NiFe], [FeFe], and [Fe] hydrogenases.
The turnover numbers for hydrogen production of the [FeFe] hydrogenases amount to
9,000 molecules per second [27]. A disadvantage of all the enzyme systems is their oxygen sensitivity; this is particularly problematic for the [FeFe] hydrogenases. The active
sites of the [NiFe] and [FeFe] hydrogenases display an unusual arrangement, which includes inorganic CO and CN ligands (Figure 4.5.1). In the case of the [FeFe] hydrogenase, the enzyme contains a [4Fe4S] cluster coupled to a [2Fe2S] cluster, both of
which are linked to the protein by the cysteine thiolate ligands.
In general, the active sites of all hydrogenases have one open coordination position,
which is most likely where substrate interacts [26]. The [NiFe] hydrogenases display a
rich redox structure in which the nickel atom cycles between the 3+ and 2+ redox state
and the iron is 2+, low spin. Using Hyperne Sublevel Correlation (HYSCORE) spectroscopy, a hydride ligand has been detected [26] providing direct evidence that coordination of the substrate occurs at the nickel. Currently, there is much interest in
improving the issue of oxygen sensitivity. In this respect, the hydrogenases from extremophile organisms are promising candidates since these enzymes are much more
robust, oxygen insensitive, and even function at elevated temperatures.
Cys
CN
359
Cys
Cys
Ni
CO
Cys
Fe
CN
Cys Fe
Fe
Cys
Cys
Cys
Figure 4.5.1
(A) Structure of [NiFe] hydrogenase from D. gigas (PDB 1FRV). The two subunits are indicated in blue (small subunit) and green (large subunit). The structure of the active
site is shown enlarged at the bottom (see text). The arrow indicates the sixth coordination site
at Ni, which is found to be unoccupied. (B) Structure of [FeFe] hydrogenase from Desulfovibrio desulfuricans (PDB 1HFE). The H cluster (hydrogen-activating cluster) and the two additional [4Fe-4S] clusters are all located in the large subunit. The molecular structure of the
H cluster with the cubane [4Fe-4S]H and the dinuclear [2Fe]H subclusters is shown enlarged
at the bottom. The arrow indicates the free coordination site at the distal iron Fed. The gure
is reproduced from Lubitz W, Reijerse EJ, van Gastel M. Chem Rev. 2007;107:433165.
In a broader sense, in addition to the H-H bond, nature often stores energy in the
chemical bonds of reduced molecules. Well-known examples are nicotinamide adenine dinucleotide (NADH) or nicotinamide adenine dinucleotide phosphate. As
with molecular hydrogen, the energy stored in the respective C-H bonds may be released in an oxidizing environment. Moreover, by storing energy in chemical bonds
as opposed to utilizing charge separation, nature has greatly simplied the issue of
energy storage. Besides NADH, another common molecule is adenosine triphosphate (ATP), in which energy (approx. 7 kcal/mol) is stored in the P-O bond.
These molecules are common fuels employed by nature to store energy. Additionally, in plant photosynthesis, light energy is converted and stored as sugars (see
Section 4.5.3.2).
360
4.5.3.1.2 Molecular Models for Hydrogen Production The active sites of hydrogenases have been a focal point for inorganic chemists with the aim of preparing biomimetic inorganic molecules that possess catalytic activity. One of the rst molecules,
which showed catalytic activity for hydrogen splitting in aqueous solution was reported by Ogo et al. [28]. This molecule features a nickel-ruthenium dinuclear
metal center, which is bridged by a hydride (as demonstrated by neutron-diffraction
studies). The hydride was additionally shown to be the product of the heterolytic
cleavage of molecular hydrogen. Other systems that show catalytic activity in organic
solvents involve homonuclear metal centers featuring ruthenium or iridium. Yet,
despite intense efforts, it has proved to be very difcult to synthesize catalytically
active complexes featuring nickel and iron, or only iron.
The challenges for the future are thus signicant in number. The oxygen sensitivity
of the enzyme systems remains an issue of crucial importance. While the synthesis
of inorganic complexes has recently resulted in catalytic activity toward hydrogen
splitting, no model systems that catalyze hydrogen production currently exist. This
feat so far remains limited to hydrogenases. Research efforts are also being directed
toward storing energy in the form of other reduced molecules (i.e. sugars and biomass)
in a broader sense. It is clear that in the general area of hydrogen splitting and production signicant research opportunities and challenges exist as we move toward an
understanding and broader utilization of this fundamental chemical reaction.
1.77 (pH 0)
H2O2 + 2e 2 OH 0.88
O2 + 4 H+ + 4e 2 H2O
1.23 (pH 0)
361
Ala344
Asp170
Glu189
Asp342
Glu333
Figure 4.5.2
His332
His337
Structure of the Mn4O5Ca core and the immediate protein surrounding the
WOC of PSII in its S2 state, as obtained from DFT geometry optimization of the 1.9 crystal
structure (Mn: purple; Ca: green; C: gray; N: blue; O: red).
362
The WOC is oxidized stepwise by a nearby tyrosine residue (Tyrz), which is itself
oxidized by the chlorophyll cation radical P680+ (formed by light-induced charge
separation). The electrons are eventually used by PSII for the reduction of plastoquinone. After the WOC has lost four electrons, the accumulated oxidizing power drives
the formation of molecular oxygen from two substrate water molecules, and the catalytic system is reset. The sequence of the four electron-transfer steps is summarized
in the Kok cycle [32] of Figure 4.5.3, where the most probable spectroscopically derived oxidation states of the Mn ions [33] are shown for each of the ve redox state
intermediates Sn (n = 04).
A prominent feature of the natural system is the separation of components and
their nely controlled spatial organization. This distribution is coupled to efcient
electron- and proton-transfer pathways, which are dened, for example, by extended
hydrogen-bond networks or accurately spaced donors and acceptors. The whole of
the protein is organized in such a way that the direction of several events becomes
practically irreversible; for example, there is a very low probability of charge recombination after the initial charge separation step at P680. The separation of primary
(P680) and secondary (Tyrz) oxidants and the WOC cluster are critical for regulating
the efciency of each catalytic step. Of particular importance is the functional role of
Tyrz as a proton-coupled redox active link. It has been shown that slight variations in
proton-transfer distance and hydrogen-bond dynamics can change the rate of protoncoupled electron transfer by orders of magnitude [34]. Clearly, if this design is to be
carried over to synthetic systems, it is necessary to have simultaneously efcient dedicated components and delicate control over their spatial distribution and interconnection even at the molecular scale. It remains to be seen whether this can
be achieved through a supramolecular approach or through the combination of
different phases.
In terms of the specic components, advancement to each step of the cycle is possible because the absorption of light and the transfer of excitation energy to P680
leads to formation of an extremely strong oxidant. The preparation of an analogous
durable synthetic system that can be selectively switched into a strong oxidizing form
H+
(MnIII)2(MnIV)2
e
(MnIII)3(MnIV)
S1
III
IV
S2 (Mn )(Mn )3
S5
H+
H+
e
O2
2 H2O
e
e
P680
Tyrz
S3
S4
e H+
Figure 4.5.3
hv
OEC
The Kok cycle showing the sequence of redox and proton- and electron-transfer
steps that lead to the oxidation of water in the WOC (left), and the electron-transfer steps in
PSII (right).
363
4.5.3.2.2 Molecular Models for Water Oxidation The inorganic cluster itself satises a number of requirements necessary for its role: it can accommodate the four necessary oxidation equivalents, it remains intact in all states of the Kok cycle, and it
avoids at all steps the Mn(II) oxidation state that is labile. Unsurprisingly, there
have been many attempts to produce structural mimics of the WOC (i.e. complexes
containing manganese ions and possibly calcium) [39]. One of the most recent ones is
a Mn3Ca cubane model [40], which reproduces the bonding of part of the WOC quite
successfully. Until now, however, the success of such systems in terms of catalytic
efciency, stability, and turnover has not been impressive. This is rather unsurprising
since it is not only the metal stoichiometry and connectivity of the WOC core that
determine its catalytic properties, but also a functionally important protein matrix
that stabilizes the various states of the core and offers specic transfer pathways.
Even the presence of extrinsic regulatory proteins affects the stability of the
Mn4O5Ca cluster [38]. The ways that the surrounding protein modies the intrinsic
properties of the inorganic core of the WOC and directs its catalytic activity are not
known in detail, but an understanding of the specic microenvironment effects is
crucial for the design of articial water-oxidation systems.
From a mechanistic point of view, the details of all reaction steps in the WOC are
not certain, but most probably, radical coupling is involved at the critical O-O bondformation step [41]. Alternative mechanisms have been proposed, including direct coupling of two hydroxo groups and nucleophilic attack of a free or loosely bound water
molecule onto a terminal oxo group [38]. Even though only a radical mechanism appears at this point to be operative in the WOC, nucleophilic attack has been shown to
be the dominant, if not exclusive, mechanism in some synthetic catalysts, notably
many ruthenium-based ones [42, 43]. This suggests that no pathway can be excluded
a priori: different mechanisms are operative in different systems, depending on the
nature of the metal and other aspects of the electronic structure or the setup of components. Therefore, there is more than one way to achieve water oxidation, and it is
crucial to recognize which one should be optimized for any given catalytic system.
In conclusion, the challenge in terms of fundamental chemistry is broadly understood; there are multiple potential ways to address it, but the best way is by no means
364
obvious. It might be nave to expect that a synthetic catalyst will mimic the WOC
both structurally and functionally. It is quite probable that the WOC will offer
insight into the fundamental chemical and physical aspects of water-oxidation catalysis but will not be a synthetic blueprint. For instance, the requirement for a fourelectron oxidation-state span and the avoidance of labile oxidation states can be
satised in multiple ways. Thus, the goal can be achieved with metals other than
manganese and with different chemistry, as has been shown, for example, by ruthenium, platinum, and iridium complexes [4245]. It should be borne in mind, though,
that it is uncertain whether systems based on rare or precious metals can be developed into practical platforms for applications. Cobalt-based systems are an attractive
alternative in this respect [46]. Besides, an appealing approach for large-scale applications is the development of catalytic surfaces or nanoclusters, relying on the carefully controlled composition and the properties of the surface rather than on the
homogeneous chemistry of an expensive molecular catalyst. Recent demonstrations
of activity by, among others, cobalt polyoxometalate [47] and cobalt-phosphate catalytic systems [48, 49] have shown that this might be a promising way forward in
terms of ease of development and implementation.
365
radical. In fact, one can think of a biological cell as an elaborate fuel cell in which the
energy released by oxygen reduction is stored in a proton gradient across the mitochondrial membrane that is in turn used to drive the synthesis of energy-rich ATP
molecules a paradigm for storing energy in chemical bonds. The terminal oxidase
in this cell respiration process is the enzyme cytochrome c oxidase (COX), which is a
large enzyme that spans the mitochondrial membrane. Using electrons provided by
metabolic processes that oxidatively break down sugars or fatty acids, COX reduces
dioxygen to water while pumping protons across the membrane. As found invariably
throughout biology, oxygen activation is achieved by an active site containing rstrow transition metals. In the case of COX, the active site consists of a mixed-metal
heme-iron/copper center.
The reaction of COX starts with the fully reduced active site containing an Fe(II)/
Cu(I) binuclear center. This unit serves as a two-electron donor. Hence, binding of
O2 leads to a simultaneous two-electron transfer creating an Fe(III)-(O22)-Cu(I)
intermediate (Figure 4.5.4). In this way, the sluggish rst reduction step from dioxygen to superoxide is avoided. The strategy to use a simultaneous two-electron transfer to create a bridging peroxide species is widespread in biochemistry in enzymes
with binuclear metal centers. The O-O bond is known to cleave after injection of
the third electron. This is sensible from an electronic structure point of view because
this electron has to enter a strongly antibonding sigma* orbital with respect to the OO bond. Hence, stabilizing this orbital is a prerequisite for O-O bond cleavage. Obviously, coordination of peroxide to transition-metal cations is a good way to achieve
this objective. A second strategy is protonation of the peroxide to yield a hydroperoxide, which is known to have a much deeper-lying sigma* orbital. In the case of
COX, it is not known whether protonation of the peroxide occurs prior to or after
O-O bond cleavage. However, it is agreed upon that the third electron is provided
by the heme center, thus creating a high-valent heme-Fe(IV)=O intermediate
together with a Cu(II)-OH conguration in the active site. The heme-Fe(IV)=O species is a powerful oxidant itself. Hence, the last step of the reaction consists of the
oxidation of a nearby tyrosine residue in a proton-coupled electron transfer, which
creates an Fe(III)-OH species. The nal stages of the reaction involve the recharging
of the active site with electrons concomitant with protonation of the hydroxide
ligands and product release.
The COX mechanism serves as a paradigm for oxygen activation at binuclear
active sites. A key feature is that the active site is constructed such that a peroxide
molecule (O-O distance ~1.4 ) ts comfortably between the two metal centers.
His
His
HisCu1+HisTyrOH
His
HisCu2+HisTyrO
OH
HisCu2+HisTyrO
OH
Fe3+
Fe3+
Fe4+
His
His
His
Figure 4.5.4 Dioxygen bond activation by cytochome c oxidase (adapted from [60]).
366
One particular feature of this mechanism is that the two metals in the active site can
provide three electrons, which means that a fourth electron has to come from a different source, in this case a redox active tyrosine residue. A different solution to the
same problem has been found in the enzyme methane monoxygenase, which involves
a binuclear iron center and will be discussed subsequently.
In summary, the metal centers serve the dual roles of providing electrons and activating the O-O bond for cleavage through lowering of the energy of the critical
sigma* orbital. The second key feature of enzymatic O2 activation concerns the
role of protons. Protons represent essentially naked positive charges and hence
are very aggressive chemical species that greatly inuence the electronic structure
of reactants. Thus, nely controlled acid/base chemistry is a key feature of enzyme
reactions. Recalling that a change of the pKa by one unit only involves an energy
change of about 1.2 kcal/mol, it becomes evident how nely tuned nature orchestrates the ow of charge in enzyme active sites and how difcult it will be to try to
achieve comparable subtlety in the design of low-molecular-weight catalysts.
Obviously, oxygen activation is more difcult to achieve at mononuclear transition-metal centers. Assuming that a given metal center cannot provide more than
two electrons in the biologically accessible potential window of at most 1.27 V
[61], it becomes clear that oxygen activation either has to stop at the peroxide
level or has to rely on other electron sources to reach completion. The paradigm
for such a reaction is provided by the family of cytochrome P450 (CYP) enzymes,
for which the reaction mechanism has been investigated in great detail (Figure 4.5.5)
[58, 60, 6264]. The active site of CYP consists of a heme-iron center with a thiolate
axial ligand. Binding of O2 to the ferrous enzyme leads to a very characteristic intermediate in the chemistry of O2 with rst-row transition metals namely, an end-on
Fe(III)O2 unit in an antiferromagnetic alignment. This intermediate cannot be further processed before another external electron and a proton enter the reaction, thus
creating an Fe(III)-OOH intermediate. This intermediate spontaneously decays under
heterolytic cleavage of the O-O bond to yield a very characteristic intermediate that
is called compound I and consists of a high-valent Fe(IV)=O center coupled to a
porphyrin radical. Thus, overall the metal and porphyrin together have contributed
three electrons toward O-O cleavage, while a fourth electron must be provided externally. Compound I is known to be a magical oxidant and reacts with unactivated
C-H bonds to yield, following a rebound step, hydroxylated products.
The characteristic feature that becomes apparent is natures use of high-valent
metal-oxo species that form during dioxygen activation. Such species are aggressive
oxidants. Hence, their creation is tightly coupled to the presence of substrate in a
conned space, which is absolutely critical for the reactions to proceed in a controlled
manner.
We should nally note the importance of the overall spin state in reactions involving dioxygen. The fact that both the open-shell transition metal and the dioxygen
unit can exist as open-shell fragments leads to a signicant variety of possible spin
couplings. For example, a quintet high-spin Fe(II) center coupled to a triplet dioxygen fragment gives rise to triplet, quintet, or septet nal states. Each spin state is
characterized by a different electronic structure and hence different reactivity. In
fact, the important concept of two-state reactivity states that a chemical reaction
367
H
O*
R-H
R-H
FeIII
R-H
O*
FeIV
FeIII
S-Cys
coordinated
product
R-OH
S-Cys
resting Fe(III) e
R-H
H2O*2
FeII
H2O*
S-Cys
oxy-Fe(IV)
porph cation radical
H+
O*2
OH
R-H
*O
FeIII
e
H+
S-Cys
hydroperoxo-Fe(III)
R-H
S-Cys
Fe(II)
O*2
FeII
S-Cys
oxy-Fe(II)
may proceed on several of those surfaces as the barriers for the elementary reaction
steps may differ widely for each spin state. It is not possible to go into further detail
here.
4.5.3.3.2 Molecular Models for Oxygen Activation The design of oxygen-activating
catalysts on the basis of rst-row transition metals has been greatly inspired by the
insights obtained in biology. In fact, there has been signicant progress in the design
of catalysts on the basis of relatively simple amine- and pyridine-based ligands (such
as tripyridine-amine, triazacyclononane). Also, reaction mechanisms and reaction intermediates have been carefully investigated, leading to a variety of transition-metal
complexes involving Fe(II) and Fe(III) peroxides and hydroperoxides as well as highvalent Fe(IV)=O and Fe(V)=O oxo units that could even be stabilized to the point
where crystal structures could be obtained. In addition to iron, a widely studied
motif is provided by dicopper complexes that bind O2 in the Cu(II) state to produce
either Cu(II)2(O22) or Cu(III)2(bis--oxo) species, which can be thought of as a fascinating case of valence isomerism. Either oxidation state features distinct reactivity.
The relative stability of the two states can be controlled via the supporting ligand
framework.
Unfortunately, it is not possible here to go into further detail about the rich chemistry involving rst-row transition metals. It, however, hopefully became apparent
that there are a number of subtle electronic and geometric structure design criteria
that must be met for efciently catalyzing dioxygen activation. Many of those factors
are fairly well understood. In the context of energy research, it remains to be seen
how these insights could be applied to the elds of heterogeneous catalysis and
368
H = 213 kcal/mol
However, at normal temperature and pressure, methane is a gas; therefore, it is difcult to transport.
At present, the Fischer-Tropsch process is employed in industry to convert standard natural gas into synthetic gasoline, diesel, or jet fuel [65]. It involves a series
of chemical reactions that lead to a variety of hydrocarbons:
H2O + CH4 CO + 3 H2
(2n+1) H2 + n CO CnH2n+2 + n H2O
The reactions are catalyzed by transition metals (cobalt, iron, and ruthenium) on
high-surface-area silica, alumina, or zeolite supports. However, the exact chemical
identity of the catalysts is unknown, and their characterization presents challenges
as these transformations are carried out under very harsh reaction conditions.
Typically, the Fischer-Tropsch process is operated in the temperature range of 150C
300C and in the pressure range of one to several tens of atmospheres [66]. Thus, the
entire process is costly and inefcient and even produces waste [67]. Hence, development of more economical and sustainable strategies for the gas-to-liquid conversion
of methane is highly desirable.
4.5.3.4.1 Methane Oxidation in Nature In nature, transformation of methane to
methanol can be very efciently carried out at ambient conditions. Methanotrophic
bacteria perform a methane hydroxylation reaction using metalloenzymes called
methane monooxygenases (MMOs) through dioxygen-activation mechanisms.
There are two types of MMOs that have been discovered so far. Soluble MMO
(sMMO) [56, 68] oxidizes methane with a well-characterized catalytic di-iron center
(Figure 4.5.6A). The intermediate called Q that is kinetically competent for methane
hydroxylation has been experimentally characterized using a variety of spectroscopic
methods. The extended X-ray absorption spectroscopy suggests that intermediate
Q likely contains an [FeIV2(-O)2]4+ diamond core [69]. The second type of MMO
is particulate MMO (pMMO), which is an integral membrane metalloprotein that
requires copper for catalytic activity. Despite the long-term intensive research
work on pMMO [70, 71], only recently was the active site for methane oxidation
369
O2
O
FeIV
FeIV
O
O2
O
FeIV
Figure 4.5.6 (A) Structure of sMMO (PDB code 1MTY) and (B) Structure of TauD (PDB
code 1OS7). The active sites and the key intermediates responsible for C-H bond oxidation
are shown at the bottom.
shown to be a di-copper center [72]. In addition to di-iron enzymes, a range of mononuclear iron enzymes are capable of effectively modifying unactivated C-H bonds
with Bond Dissociation Energies (BDEs) comparable to methane, such as taurine/
-ketoglutarate dioxygenase (TauD) (shown in Figure 4.5.6B). Experiments have revealed that the C-H bond cleaving intermediates in their catalytic circles all involve a
high-valent metal-oxo unit [73]. These metalloenzymes thus provide the optimal model
for an efcient, environmentally friendly catalyst.
4.5.3.4.2 Molecular Models for Methane Oxidation Recently, the synthetic precedent for the [FeIV2(-O)2]4+ diamond core proposed for intermediate Q has been reported [74, 75]. Surprisingly, this complex is a sluggish oxidant and is only able to
oxidize weak C-H bonds (BDE < 80 kcal/mol) [74]. However, its one-electron reduced species possessing an open [HO-FeIII-O-FeIV=O]2+ core structure displays
much higher reactivity toward C-H bond activation, and the reaction rate is at
least 105-fold faster than that determined for the diamond complex with the higher
oxidation states of the metal centers [76]. Interestingly, the oxidation state does
not appear to dictate their relative reactivities; hence, there must be more critical factors, such as spin states of ferryl centers (see subsequent discussion), that need to be
taken into account in order to fully understand high-valent di-iron chemistry.
370
371
Figure 4.5.7 Structure of the FeMoco in the MoFe protein (PDB: 1M1N). Orange spheres
372
models that are capable of catalytic dinitrogen reduction are known. The only homogeneous catalyst known to reduce dinitrogen is a molybdenum complex, again lending
intrigue to the role of the heterometal.
4.5.3.5.2 Molecular Model for Nitrogen Fixation Since the rst report of an N2bound transition-metal complex in 1965, chemists have been hopeful that molecular
catalysts could be made to mimic the chemistry of nitrogenase. However, it was not
until 2003 that Schrock and coworkers reported the rst discrete transition-metal catalyst that reduces dinitrogen to ammonia in the presence of electrons and protons
[98]. The complex is a monomeric Mo species that utilizes a triamidoamine ligand
framework and cycles between oxidation states of Mo(III) to Mo(VI) during the
course of catalysis. Several intermediates in the reaction cycle have been isolated
and characterized, and the suggested mechanism parallels the original proposals of
Chatt for low-valent Mo and W species.
Notable progress in ammonia synthesis has also been made by Chirik and coworkers, who have shown that ammonia can be evolved from a zirconium metallocene
complex. Using a tetramethylated-bis-Cp-dichloride complex, they were able to rst
add N2 to form a dimeric Zr complex and subsequently add excess H2, with heating,
to evolve ammonia [99]. Their work, though not catalytic, also provides basic
mechanistic insights into transition-metal-mediated N-H bond catalysis.
In contrast, a molecular iron catalyst for conversion of dinitrogen into ammonia
has remained elusive, despite its relevance to both the heterogeneous process and biological catalysis. A number of iron-dinitrogen complexes have been isolated and
characterized with varying degrees N-N bond activation [100]. Both iron(0) and
iron(I) complexes have been shown to mediate conversion of N2 to ammonia, however, in low yields and not catalytically. Recent studies by Holland and coworkers
demonstrate that a trimeric iron complex can evolve ammonia stoichiometrically,
though not catalytically [101]. With much continued work in this area, there is promise that a molecular iron catalyst is on the horizon and that there is still much to learn
at the interface of chemical and biological ammonia synthesis.
4.5.6 References
373
4.5.5 Acknowledgments
The authors thank the Max Planck Society for funding.
4.5.6 References
1. Blankenship RE. Molecular mechanisms of photosynthesis. Oxford: Blackwell; 2001.
2. Ballou DP, Palmer GA. Practical rapid quenching instrument for study of reactionmechanisms by electron-paramagnetic resonance spectroscopy. Anal Chem. 1974;46:
124853.
3. Bollinger JM, Krebs C. Stalking intermediates in oxygen activation by iron enzymes:
Motivation and method. J Inorg Biochem. 2006;100:586605.
4. Bollinger JM, Tong WH, Ravi N, Huynh BH, Edmondson DE, Stubbe J. Redox-active
amino acids in biology. 1995, 258, 278.
5. Cherepanov AV, de Vries S. Microsecond freeze-hyperquenching: development of a new
ultrafast micro-mixing and sampling technology and application to enzyme catalysis. Biochim Biophys Acta Bioenerg. 2004;1656:131.
6. Krebs C, Bollinger JM Jr. Freeze-quench (57)Fe-Mssbauer spectroscopy: trapping reactive intermediates. Photosynth Res. 2009;102:295304.
7. Krebs C, Price JC, Baldwin J, Saleh L, Green MT, Bollinger JM. Rapid freeze-quench
57
Fe Mssbauer spectroscopy: monitoring changes of an iron-containing active site during
a biochemical reaction. Inorg Chem. 2005;44:74257.
8. Mitic N, Saleh L, Schenk G, Bollinger JM, Solomon EI. Rapid-freeze-quench magnetic
circular dichroism of intermediate X in ribonucleotide reductase: new structural insight.
J Am Chem Soc. 2003;125:112001.
9. Koeningsberger DC, Prins R. X-ray absorption: principles, applications and techniques of
EXAFS, SEXAFS and XANES. New York: John Wiley & Sons; 1988.
10. Buhl M, Reimann C, Pantazis DA, Bredow T, Neese F. Geometries of third-row transitionmetal complexes from density-functional theory. J Chem Theory Comput. 2008;4:
144959.
374
11. Ray K, Begum A, Weyhermller T, et al. The electronic structure of the isoelectronic,
square-planar complexes [FeII(L)2]2 and [CoIII(L Bu)2] (L2 and (L Bu)2=benzene-1,
2-dithiolates): an experimental and density functional theoretical study. J Am Chem Soc.
2005;127:440315.
12. Schoneboom JC, Neese F, Thiel W. Toward identication of the compound I reactive
intermediate in cytochrome P450 chemistry: a QM/MM study of its EPR and Mssbauer
parameters. J Am Chem Soc. 2005;127:58403.
13. Berry JF, Bill E, Bothe E, et al. An Octahedral Coordination Complex of Iron(VI).
Science. 2006;312:193741.
14. Curtiss LA, Raghavachari K, Redfern PC, Pople JA. Assessment of Gaussian-2 and
density functional theories for the computation of enthalpies of formation. J Chem Phys.
1997;106:106379.
15. Neese F. A critical evaluation of DFT, including time-dependent DFT, applied to bioinorganic chemistry. J Biol Inorg Chem. 2006;11:70211.
16. Shaik S, Cohen S, Wang Y, Chen H, Kumar D, Thiel W. P450 enzymes: their structure, reactivity, and selectivity-modeled by QM/MM calculations. Chem Rev. 2010;110:9491017.
17. Siegbahn PEM, Eriksson L, Himo F, Pavlov M. Hydrogen Atom Transfer in Ribonucleotide Reductase (RNR). J Phys Chem B. 1998;102:106229.
18. Holm RH, Kennepohl P, Solomon EI. Structural and Functional Aspects of Metal Sites in
Biology. Chem Rev. 1996;96:22392314.
19. Solomon EI, Lever ABP, editors. Inorganic electronic structure and spectroscopy. New
York: John Wiley & Sons; 1999. Volumes 12.
20. Jensen F. Introduction to computational chemistry. New York: Wiley; 2009.
21. Neese F. Prediction of molecular properties and molecular spectroscopy with density functional theory: From fundamental theory to exchange-coupling. Coord Chem Rev. 2009;
253:52663.
22. Petrenko T, DeBeer George S, Aliaga-Alcalde N, et al. Characterization of a genuine iron
(V)-nitrido species by nuclear resonant vibrational spectroscopy coupled to density functional calculations. J Am Chem Soc. 2007;129:1105360.
23. Petrenko T, Krylova O, Neese F, Sokolowski M. Optical Absorption and Emission Properties of Rubrene: Insight by a Combined Experimental and Theoretical Study. New J
Phys. 2009;11:1105360.
24. Bratus VY, Petrenko TT, Okulov SM, Petrenko TL. Positively charged carbon, vacancy
in three inequivalent lattice sites of 6H-SiC: Combined EPR and density functional theory
study. Phys Rev B. 2005;71:123.
25. Handbook of chemistry and physics. 81st ed. New York: Taylor and Francis Group; 2000.
26. Michl, J. Chem Rev Hydrogen Issue. 2007;107:38994435.
27. Cammack R. Hydrogenase sophistication. Nature. 1999;397:2145.
28. Ogo S, Kabe R, Uehara K, et al. A dinuclear Ni(mu-H)Ru complex derived from H2.
Science. 2007;316, 5857.
29. Inoue H, Shimada T, Kou Y, et al. The water oxidation bottleneck in articial photosynthesis: how can we get through it? An alternative route involving a two-electron process.
ChemSusChem 2011;4:1739.
30. Umena Y, Kawakami K, Shen J.-R, Kamiya N. Crystal structure of oxygen-evolving
photosystem II at a resolution of 1.9 . Nature. 2011;473:5560.
31. Ames W, Pantazis DA, Krewald V, et al. Theoretical evaluation of structural models of
the S2 state in the oxygen evolving complex of Photosystem II: protonation states and
magnetic interactions. J Am Chem Soc. 2011;133:1974357.
32. Kok B, Forbush B, McGloin M. Cooperation of charges in photosynthetic O2 evolution-I.
A linear four step mechanism. Photochem Photobiol. 1970;11:45775.
4.5.6 References
375
33. Kulik LV, Epel B, Lubitz W, Messinger J. 55Mn pulse ENDOR at 34 GHz of the S0 and
S2 states of the oxygen-evolving complex in photosystem II. J Am Chem Soc. 2005;
127:23923.
34. Hammarstrm L, Styring S. Proton-coupled electron transfer of tyrosines in Photosystem II
and model systems for articial photosynthesis: the role of a redox-active link between catalyst and photosensitizer. Energy Environ Sci. 2011;4:237988.
35. Dau H, Haumann M. Eight steps preceding O-O bond formation in oxygenic photosynthesisA basic reaction cycle of the Photosystem II manganese complex. Biochim Biophys
Acta Bioenerg. 2007;1767:47283.
36. Dau H, Haumann M. The manganese complex of photosystem II in its reaction cycle
Basic framework and possible realization at the atomic level. Coord Chem Rev. 2008;
252:27395.
37. McEvoy JP, Brudvig GW. Water-splitting chemistry of photosystem II. Chem Rev. 2006;
106:445583.
38. Messinger J, Renger G. In: Renger G, editor. Primary processes of photosynthesis, part 2:
Principles and apparatus. Cambridge: The Royal Society of Chemistry; 2008. Volume 9. p. 291.
39. Mukhopadhyay S, Mandal SK, Bhaduri S, Armstrong WH. Manganese clusters with relevance to photosystem II. Chem Rev. 2004;104:39814026.
40. Kanady JS, Tsui EY, Day MW, Agapie T. A synthetic model of the Mn3Ca subsite of the
oxygen-evolving complex in photosystem II. Science. 2011;333:7336.
41. Siegbahn PEM. Structures and energetics for O2 formation in photosystem II. Acc Chem
Res. 2009;42:187180.
42. Romain S, Vigara L, Llobet A. Oxygen-oxygen bond formation pathways promoted by
ruthenium complexes. Acc Chem Res. 2009;42:194453.
43. Concepcion JJ, Jurss JW, Brennaman MK, et al. Making oxygen with ruthenium complexes. Acc Chem Res. 2009;42:195465.
44. Sala X, Romero I, Rodrguez M, Escriche L, Llobet A. Molecular catalysts that oxidize
water to dioxygen. Angew Chem Int Ed. 2009;48:284252.
45. McDaniel ND, Coughlin FJ, Tinker LL, Bernhard S. Cyclometalated iridium(III) Aquo
complexes: efcient and tunable catalysts for the homogeneous oxidation of water.
J Am Chem Soc. 2008;130:2107.
46. Artero V, Chavarot-Kerlidou M, Fontecave M. Splitting Water with Cobalt. Angew
Chem Int Ed. 2011;50:723866.
47. Yin Q, Tan JM, Besson C, et al. A fast soluble carbon-free molecular water oxidation catalyst based on abundant metals. Science. 2010;328:3425.
48. Kanan MW, Nocera DG. In situ formation of an oxygen-evolving catalyst in neutral
water containing phosphate and Co2+. Science. 2008;321:10725.
49. Lutterman DA, Surendranath Y, Nocera DG. A self-healing oxygen-evolving catalyst.
J Am Chem Soc. 2009;131:38389.
50. Costas M, Mehn MP, Jensen MP, Que L. Dioxygen activation at mononuclear nonheme
iron active sites: enzymes, models, and intermediates. Chem Rev. 2004;104:93986.
51. Fox BG, Lyle KS, Rogge CE. Reactions of the diiron enzyme stearoyl-acyl carrier protein
desaturase. Acc Chem Res. 2004;37:4219.
52. Friedle S, Reisner E, Lippard SJ. Current challenges of modeling diiron enzyme active
sites for dioxygen activation by biomimetic synthetic complexes. Chem Soc Rev. 2010;
39:276879.
53. Kovaleva EG, Lipscomb JD. Versatility of biological non-heme Fe(II) centers in oxygen
activation reactions. Nat Chem Biol. 2008;4:18693.
54. Krebs C, Galonica Fujimori D, Walsh CT, Bollinger JM. Non-heme Fe(IV)-oxo intermediates. Acc Chem Res. 2007;40:48492.
376
55. Siegbahn PEM, Borowski T. Modeling enzymatic reactions involving transition metals.
Acc Chem Res. 2006;39:72938.
56. Tinberg CE, Lippard SJ. Dioxygen activation in soluble methane monooxygenase.
Acc Chem Res. 2011;44:28088.
57. Stone KL, Borovik AS. Lessons from nature: unraveling biological CH bond activation.
Curr Opin Chem Biol. 2009;13:1148.
58. Groves JT. High-valent iron in chemical and biological oxidations. J Inorg Biochem.
2006;100:43447.
59. Nam W. High-valent iron(IV)-oxo complexes of heme and non-heme ligands in oxygenation reactions. Acc Chem Res. 2007;40:52231.
60. Bassan A, Blomberg MRA, Borowski T, Siegbahn PEM. Theoretical studies of enzyme
mechanisms involving high-valent iron intermediates. J Inorg Biochem. 2006;100:72743.
61. Bakac A. Oxygen activation with transition-metal complexes in aqueous solution. Inorg
Chem. 2010;49:358493.
62. Jung C. The mystery of cytochrome P450 Compound I: a mini-review dedicated to Klaus
Ruckpaul. Biochim Biophys Acta Proteins Proteomics. 2011;1814:4657.
63. Ortiz de Montellano PR. Hydrocarbon hydroxylation by cytochrome P450 enzymes.
Chem Rev. 2009;110:93248.
64. Shaik S, Cohen S, Wang Y, Chen H, Kumar D, Thiel W. P450 enzymes: their structure, reactivity, and selectivity-modeled by QM/MM calculations. Chem Rev. 2009;110:9491017.
65. Hamelinck CN, Faaij APC, den Uil H, Boerrigter H. Production of FT transportation
fuels from biomass; technical options, process analysis and optimisation, and development
potential. Energy. 2004;29:174371.
66. Hermans I, Spier ES, Neuenschwander U, Turra N, Baiker A. Selective Oxidation Catalysis: Opportunities and Challenges. Top Catal. 2009;52:116274.
67. Arakawa H, Aresta M, Armor JN, et al. Catalysis research of relevance to carbon management: progress, challenges, and opportunities. Chem Rev. 2001;101:95396.
68. Que L Jr, Dong YH. Modeling the oxygen activation chemistry of methane monooxygenase and ribonucleotide reductase. Acc Chem Res. 1996;29:1906.
69. Shu LJ, Nesheim JC, Kauffmann K, Mnck E, Lipscomb JD, Que L. An Fe2IVO2 diamond core structure for the key intermediate Q of methane monooxygenase. Science.
1997;275:5158.
70. Lieberman RL, Rosenzweig AC. Crystal structure of a membrane-bound metalloenzyme
that catalyses the biological oxidation of methane. Nature. 2005;434:17782.
71. Chan SI, Yu SSF. Controlled oxidation of hydrocarbons by the membrane-bound methane monooxygenase: the case for a tricopper cluster. Acc Chem Res. 2008;41:96979.
72. Balasubramanian R, Smith SM, Rawat S, Yatsunyk LA, Stemmler TL, Rosenzweig AC.
Oxidation of methane by a biological dicopper centre. Nature. 2010;465:1159.
73. Krebs C, Fujimori DG, Walsh CT, Bollinger JM Jr. Non-heme Fe(IV)-oxo intermediates.
Acc Chem Res. 2007;40:48492.
74. Xue G, Wang D, De Hont R, et al. A synthetic precedent for the [FeIV2(mu-O)2] diamond
core proposed for methane monooxygenase intermediate Q. Proc Natl Acad Sci U S A.
2007;104:2071318.
75. Wang D, Farquhar ER, Stubna A, Mnck E, Que L Jr. A diiron(IV) complex that cleaves
strong C-H and O-H bonds. Nat Chem. 2009;1:14550.
76. Xue G, De Hont R, Mnck E, Que L Jr. Million-fold activation of the [Fe(2)(micro-O)(2)]
diamond core for C-H bond cleavage. Nat Chem. 2010;2:4005.
77. Que L Jr. The road to non-heme oxoferryls and beyond. Acc Chem Res. 2007;40:493500.
78. Louwerse MJ, Baerends EJ. Oxidative properties of FeO2+: electronic structure and
solvation effects. Phys Chem Chem Phys. 2007;9:15666.
4.5.6 References
377
79. Shaik S, Hirao H, Kumar D. Reactivity of high-valent iron-oxo species in enzymes and
synthetic reagents: a tale of many states. Acc Chem Res. 2007;40:53242.
80. Ye S, Neese F. Quantum chemical studies of C-H activation reactions by high-valent
nonheme iron centers. Curr Opin Chem Biol. 2009;13:8998.
81. Solomon EI, Wong SD, Liu LV, Decker A, Chow MS. Peroxo and oxo intermediates in
mononuclear nonheme iron enzymes and related active sites. Curr Opin Chem Biol.
2009;13:99113.
82. Ye S, Neese F. Nonheme oxo-iron(IV) intermediates form an oxyl radical upon approaching the C-H bond activation transition state. Proc Natl Acad Sci U S A. 2011;
108:122833.
83. Smith BE. Nitrogenase reveals its inner secrets. Science. 2002;297:165455.
84. Fryzuk MD. Inorganic chemistry: ammonia transformed. Nature. 2004;427:4989.
85. Burgess BK, Lowe DL. Mechanism of Molybdenum Nitrogenase. Chem Rev. 1996;96:
29833012.
86. Howard JB, Rees DC. Structural Basis of Biological Nitrogen Fixation. Chem Rev.
1996;96:296582.
87. Hu YL, Lee CC, Ribbe MW. Extending the carbon chain: hydrocarbon formation catalyzed by vanadium/molybdenum nitrogenases. Science. 2011;333:7535.
88. Lee CC, Hu YL, Ribbe MW. Vanadium nitrogenase reduces CO. Science. 2010;329:642.
89. Einsle O, Tezcan FA, Andrade SLA, et al. Nitrogenase MoFe-protein at 1.16 A resolution: a central ligand in the FeMo-cofactor. Science. 2002;297:16961700.
90. George SJ, Igarashi RY, Xiao Y, et al. Extended X-ray absorption ne structure and
nuclear resonance vibrational spectroscopy reveal that NifB-co, a FeMo-co precursor,
comprises a 6Fe core with an interstitial light atom. J Am Chem Soc. 2008;130:567380.
91. Xiao YM, Fisher K, Smith MC, et al. How nitrogenase shakesinitial information about
P-cluster and FeMo-cofactor normal modes from nuclear resonance vibrational spectroscopy (NRVS). J Am Chem Soc. 2006;128:760812.
92. Hoffman BM, Dean DR, Seefeldt LC. Climbing nitrogenase: toward a mechanism of
enzymatic nitrogen xation. Acc Chem Res. 2009;42:60919.
93. Lukoyanov D, Pelmenschikov V, Maeser N, et al. Testing if the interstitial atom, X, of the
nitrogenase molybdenum-iron cofactor is N or C: ENDOR, ESEEM, and DFT studies of
the S = 3/2 resting state in multiple environments. Inorg Chem. 2007;46:1143749.
94. Harris TV, Szilagyi RK. Comparative Assessment of the Composition and Charge State
of Nitrogenase FeMo-Cofactor. Inorg Chem. 2011;50:481124.
95. Lancaster KM, Roemelt M, Ettenhuber P, et al. X-ray emission spectroscopy evidences a
central carbon in the nitrogenase iron-molybdenum cofactor. Science. 2011;334:9747.
96. Spatzal T, Aksoyoglu M, Zhang LM, et al. Evidence for interstitial carbon in nitrogenase
FeMo cofactor. Science. 2011;334:940.
97. Lukoyanov D, Dikanov SA, Yang ZY, et al. ENDOR/HYSCORE studies of the common intermediate trapped during nitrogenase reduction of N2H2, CH3N2H, and N2H4
support an alternating reaction pathway for N2 reduction. J Am Chem Soc. 2011:133:
1165564.
98. Yandulov DV, Schrock RR. Catalytic reduction of dinitrogen to ammonia at a single
molybdenum center. Science. 2003;301:768.
99. Pool JA, Lobkovsky E, Chirik PJ. Hydrogenation and cleavage of dinitrogen to ammonia with a zirconium complex. Nature. 2004;427:52730.
100. Hazari N. Homogeneous iron complexes for the conversion of dinitrogen into ammonia
and hydrazine. Chem Soc Rev. 2010;39:404456.
101. Rodriguez MM, Bill E, Brennessel WW, Holland PL. N(2) reduction and hydrogenation
to ammonia by a molecular iron-potassium complex. Science. 2011;334:7803.
5.1.1 Introduction
Sunlight is the largest energy ow entering the terrestrial ecosystem (~162 PW) and
accounts for more than three orders of magnitude of the current global consumption
of primary energy (~15 TW). However, solar energy must be converted to electrical,
thermomechanical, or chemical energy to be used. In addition, energy must be
stored/transported in an energy-efcient approach to be used when and where is necessary. Chemical energy storage is the most efcient way for the storage and transport of energy, in terms of energy density and cost-effectiveness. This is the main
motivation in the interest in energy vectors. Our society and energy scenario is largely
based on the use of liquid hydrocarbons as energy vector, derived mainly from the
rening of oil. In year 2010, approximately half of the world energy consumption
was based on oil and derived liquid fuels (gasoline, diesel, jet fuels, gasoil, etc.),
while only less than one-fth was based on electrical energy. Current forecasts by
the International Energy Agency [1] indicate for year 2050 that chemical energy
will still play a dominant role with respect to other forms of energy (electrical,
etc.). While energy could be interconverted between the different forms, as schematically shown in Figure 5.1.1, it is evident that the possibility of using directly
sunlight to store energy in a chemical form to facilitate its storage and transport
(solar fuels [2]) is the objective to obtain better efciency (avoiding the losses in
the energy conversion) and improve the sustainability of the energy value chain.
Suitable energy vectors must fulll a number of requirements:
1. They must have a high energy density both by volume and by weight.
2. They must be easy to store without need of high pressure at room temperature.
3. They must be of low toxicity and safe in handling and show limited risks in their
distributed (nontechnical) use.
4. They must show good integration into the actual energy infrastructure without
need of new dedicated equipment.
5. They must have a low impact on the environment in both their production
and use.
H2 has been often presented as the ideal energy vector [3, 4], but its energy density,
even considering future possible developments in storage materials, will be still a
main issue for practical large-scale use. The second critical problem is related to
the very large investments necessary to change the energy infrastructure, when
380
Electrical
Energy
Chemical
Energy
Thermal
Energy
Photovoltaics
Photosynthetic,
Photo(electro)chemical
Solar Thermal
Mechanical
Energy
Thermochemical
processes
Thermoelectric,
Thermo-ionic, etc.
there is a change in the dominant energy vector. Solar H2 [5] may be a better and
more sustainable alternative, when combined with the possibility of forming liquid
fuels easily transportable and with high energy density [6, 7]. This possibility is offered from the use of solar H2 to produce fuels from CO2, or even better to integrate
directly the solar cell able to produce protons/electron (the equivalent of H2) from
water oxidation using sunlight with an electrocatalyst able to use the protons/electrons to efciently convert CO2 to fuels. This is the photoelectrocatalytic (PEC)
approach, in which the two reactions of water oxidation using sunlight and CO2
reduction using the electrons/protons generated in the light-illuminated side occurs
in two different cell compartments separated through a proton-conducting membrane [8]. Note that we introduce the concept of photoelectrocatalysis, an alternative
to photoelectrochemistry, because catalysis is the critical element to reduce the overpotentials in the reactions and enable an effective device. Producing solar fuels via
recycling CO2 is a carbon-neutral approach to store and transport solar energy,
which can be well integrated into the current energy infrastructure. It is an effective
path to introduce renewable energy into the energy chain and, as discussed recently,
also in the chemical production chain [9]. Photoelectrochemical CO2 activation is
thus a critical component to enable a sustainable energy portfolio and society.
381
harvesting and photoinduced charge separation, and catalysis modules to split water
into oxygen and hydrogen (in the form of reducing equivalents: protons and electrons) and then use these protons and electrons to reduce CO2. Plants and algae can
split water exploiting sunlight via the enzyme photosystem II (PSII), whose active element is composed of four manganese atoms (Mn) and a calcium ion (Ca2+) surrounded by a highly conserved protein matrix. The electrons and protons can be
also transferred to photosystem I (PSI) to produce a reduced form of hydrogen in
the form of nicotinamide adenine dinucleotide phosphate (NADPH), which is the cofactor used by other enzymes to reduce CO2. However, reproduce by a synthetic
(articial, biomimetic) approach this complex machinery never has resulted in a practical application [1012], due to the too low cost-effectiveness of the approach, even if
the scientic relevance of these studies has to be remarked.
The design of efcient, cost-effective articial systems requires maintaining the
main elements in the leaf hierarchical structure (capture of sunlight photons, electronhole separation with long lifetimes, energy transduction, etc.) while developing a
new functional and robust design that realizes two goals: (1) intensifying the process,
thus allowing higher productivity and efciency in converting sunlight (in the plants,
the quantum yield is typically below 1%) and (2) using solid components that keep
functionalities but are more robust, scalable, and cost-effective. The PEC solar cell
discussed in the previous section is the preferable starting point on which to construct
the new functional design to realize an articial leaf. In a simplied approach, the
dark side of this articial leaf can be based instead on a catalyst for proton/electron
recombination to evolve H2. The PEC solar cell device (articial leaf ) may be thus
used also to produce H2 and O2 in separate compartments from water splitting
using sunlight. Figure 5.1.2 shows this concept. An articial leaf should be thus
composed of the following main elements:
An anode exposed to sunlight carrying a photocatalysts able to oxidize water
and supported on a conductive substrate, which allows the fast collection of the
2H2O O2 + "2H2"
design taking inspiration from nature, but developing conceptually new and robust devices
that overcome the limits of natural leaves.
382
anode
e
383
cathode
e
H+
CO2(g)
2H2O
O2
4H+
+ 4e
(CH2O)
nH+
+ ne
membrane
H2O
O2
(CH2O)
H+
CO2
electrocatalyst
electrode
electrocatalyst
e
electrode
porous
semiconductor
h+
Ebandgap
H+
membrane
H+
Figure 5.1.3 PEC solar cell device with a scheme of the photoelectrolysis process at the ano-
dic side and the electrocatalytic process of CO2 reduction at the cathodic side. (CH2O) indicates in general the products of CO2 hydrogenation.
384
385
Current density
Sel H2
Sel CO
Sel HCOO
10
100
Sel C2H4
Sel CH4
1
60
40
Selectivity, %
80
0.1
20
0.01
0.8
0
1.0
1.2
Potential (V vs SHE)
1.4
Figure 5.1.4 Dependence on the potential (with respect to SHE) of the current density and
selectivities toward products in the CO2 reduction in methanol under pressure and using
Cu-foil electrode. Adapted from [16a].
it is necessary to run the cathode very negative (i.e. at a high overpotential) for the
reaction to occur. However, by using suitable solvents or catalytic pathways, it is
possible to lower the potential for formation of the CO2-type intermediate,
which then reacts with H+ on the cathode to produce CO (Figure 5.1.5).
Masel et al. [19] have shown recently that it is possible to electrocatalytically
reduce CO2 to carbon monoxide (CO) at overpotentials below 0.2 V using an
ionic liquid electrolyte to lower the energy of the CO2-type intermediate, most
likely by complexation, and thereby lower the initial reduction barrier. Using a silver
cathode, it is possible then to catalyze the formation of the nal products. Formation
of gaseous CO is rst observed at an applied voltage of 1.5 V, just slightly above the
minimum (i.e. equilibrium) voltage of 1.33 V. The system continued producing CO
for at least 7 hours at faradaic efciencies greater than 96% [19].
There are two main approaches, depending whether the conversion of CO2 is studied in aqueous or nonaqueous solutions. Formic acid is the main reaction product in
electrolysis of aqueous solutions of CO2. A problem in the utilization of CO2 in aqueous solution derives from its low solubility in water at standard temperature and
pressure. At the surface of the electrode, there are very small amounts of CO2 available for the reaction to proceed. For aqueous solutions, in order to speed the reaction
rate, the pressure must be increased.
386
+ 2H
0.8
0.6
Overpotential
0.4
CO + H2O
0.2
0
CO2
Catalytic
Route
0.2
0.4
Reaction Coordinate
Figure 5.1.5 A scheme of how the free energy of the system changes during the reaction CO2
Numerous studies have been made on the electrochemical reduction of CO2 under
high pressure on various electrodes in an aqueous electrolyte. Productivity increases
substantially at 30 bar of CO2 with respect to that at 1 bar of CO2. However, the
total cathodic current barely increases with increasing CO2 pressure. In terms of
products, CO, H2, and formic acid are mainly observed. A fast deactivation is also
typically present.
Solvents with high solubility for CO2 are used in the nonaqueous electrochemical
reduction of CO2. Carbon-dioxide concentration in dimethyl-formamide is about 20
times higher than in aqueous solutions, and in propylene carbonate and methanol,
the CO2 solubility is about 8 and 5 times higher, respectively. However, high CO2
solubility requires larger current density, but low electrolytic conductivity leads to
high ohmic losses. For this reason, methanol is often used to balance these two aspects. Cu-based foils give the best results as electrodes. Another problem is that
very high current densities at the copper cathode are necessary to maximize the formation of hydrocarbons, but a fast deactivation is present in these conditions [20].
The formation of products as a function of the potential and partial current densities
is shown in Figure 5.1.4.
Recent advances on the electrochemical reduction of CO2 to a range of possible
product are summarized in Figure 5.1.6, showing the efciencies and current densities for CO2 reduction from literature data compared to the typical values for water
electrolyzers [17]. Typical efciencies for commercial water electrolyzers are in the
56%73% range, with alkaline electrolyzers running at 110300 mAcm2 and
PEM electrolyzers running at much higher current densities (8001,600 mAcm2)
[2123]. These efciencies for the electrolyzers are total system efciencies (including
losses from system peripherals). However, because no complete systems for CO2
387
100
90
Energy Efficiency
80
Alkaline Electrolyzers
PEM Electrolyzers
70
60
Formic Acid
CO2
50
Syngas
40
Hydrocarbons
CO
30
formic acid
20
10
ethylene
methane
0
0
200
400
600
800
1000
1200
1400
1600
Figure 5.1.6 Comparison of the energy efciencies and current densities for CO2 reduction to
formic acid, syngas, and hydrocarbons (methane and ethylene) reported in the literature with
those of water electrolyzers. Efciencies of electrolyzers are total system efciencies, while the
CO2 conversion efciencies only include cathode losses and neglect anode and system losses.
Adapted from [17].
reduction are still available, the efciencies plotted for CO2 reduction only include
losses on the cathode and ignore anode and system losses.
Figure 5.1.6 evidences that electrochemical conversion of CO2 shows moderate efciencies and reasonably high current densities, although not at the same time. While
researchers have reported high faradaic efciency for many products (typically >90%
for formic acid and carbon monoxide, and 6570% for methane and ethylene), high
overpotentials are a major hindrance to improving energy efciency.
The direct electrochemical reduction to hydrocarbons (methane and ethylene) is
particularly interesting but suffers from low energy efciency due to high overpotentials. In contrast, the electrochemical reduction to CO has high energy efciencies,
and the main side product, hydrogen, is included in the syngas product. However,
because hydrogen evolution has a higher (less negative) theoretical reduction potential than carbon monoxide, it is more efcient to optimize the electrolysis cell for CO
production and supply the hydrogen from a different renewable source.
The reaction rate, as measured by the current density, is also an important parameter as it determines the reactor size and thus capital cost of the process. To date, researchers have reported moderate to high current densities (200600 mAcm2) using
gas diffusion electrodes (GDEs) similar to those used in fuel cells. Further work is
thus needed to signicantly improve the energetic efciency and current densities
for CO2 reduction.
As commented before regarding the electrolysis in ionic liquids [19] and the use of
catalytic pathways to reduce the overpotential (Figure 5.1.5), it is possible to reduce
388
the overpotential, increase the efciency, and control the selectivity, as well as stability of operations, by a proper design of the electrode and control of the reaction conditions. This is the rationale for framing the discussion in terms of electrocatalytic
rather than electrochemical conversion, because overcoming the classical design
of electrodes to introduce catalytic functionalities is the approach necessary for
signicant progress in this eld. This will be discussed in the following section.
The importance of these concepts can be illustrated from the analysis of recent
data on the role of the surface nature of copper electrodes in the electroreduction
of CO2. Shibata et al. [24] rst observed that alkanes and alkenes up to C6 can be
obtained in CO2 electroreduction at room temperature and atmospheric pressure
by application of a commercially available Cu electrode, provided that pretreatment
by electropolishing (usually made in preparing electrodes) is avoided. When the same
electrode material was pretreated by electropolishing it behaved like a pure Cu electrode, giving mainly methane and ethene. Gonalves et al. [25] also recently observed
in the electrochemical conversion of CO2 on copper electrodes that the presence of
the electrodeposits drastically changes the performances and product distribution
for example, selectively producing ethylene instead of methane and also largely
increasing the stability of the electrode. While no clear explanation of these observations was given in the cited papers, it may be reasonably interpreted that the effect is
due to the presence on the copper electrode surface of catalyst nanoparticles, which
modify the overpotential in the reactions and thus selectivity and faradaic efciency.
A better understanding of these effects will thus be the key for the design of efcient
cathodic electrocatalysts for CO2 reduction.
389
carbon dioxide to products such as formic acid, formaldehyde, and methanol. Importantly, high faradaic yields for methanol have been observed in both electrochemical
and photoelectrochemical systems at low reaction overpotentials. At hydrogenated
Pd electrodes, faradaic efciencies for the reduction of CO2 to methanol of ~30%
were observed at overpotentials of only ~200 mV [32]. Using instead a p-GaP photoelectrochemical system yielded nearly 100% in faradaic efciency for methanol at underpotentials as high as 300 mV (below the thermodynamic potential of 0.52 V vs.
saturated calomel electrode [SCE] at pH of 5.2) could be reached instead [33].
The mechanism of pyridinium-catalyzed CO2 reduction to methanol was investigated recently by Barton Cole et al. [30]. At metal electrodes, formic acid and
formaldehyde were observed to be intermediate products along the pathway to the
6 e reduced product of methanol, with the pyridinium radical playing a role in
the reduction of both intermediate products. Differently from the multielectron
transfer necessary in metals to achieve highly reduced products such as methanol,
the pyridinium cation is a simple 1 e electrocatalyst but is nevertheless capable of
reducing many different chemical species en route to methanol through six sequential
electron transfers. The mechanism of the reduction proceeds through various coordinative interactions between the pyridinium radical and carbon dioxide, formaldehyde, and related species. This suggests an inner-sphere-type electron transfer from
the pyridinium radical to the substrate for various mechanistic steps where the
pyridinium radical covalently binds to intermediates and radical species.
Pyridinium ion can thus reduce CO2 to highly reduced species through multiple
electron transfers without the need for a metal-based multielectron transfer. This
nding stands in contrast to current thinking that assumes fuel-forming reactions
will only be energetically favorable if a multielectron charge-transfer reagent is employed. Pyridinium acts as an electron shuttle, and this property probably explains
high faradaic yields. It is thus evident that there are several possibilities for overcoming current limitations in the performances of electrochemical conversion of CO2 by
a proper and novel design of catalytic advanced electrodes.
Another key limiting factor for the conversion of CO2 is the low mass-transfer
rates of CO2 at the electrode/catalyst surface, which is exacerbated by the low solubility of CO2 in many electrolytes. Ionic liquids [34] may help to solve this problem,
as Zhao et al. [35] originally proposed.
In electrodes, these mass-transfer limitations could be overcome using GDEs,
which create a three-phase interface between the gaseous reactants, the solid catalyst,
and the electrolyte. Thus, optimization of the electrode will be the key to improving
current densities. The extensive work on GDE optimization for fuel cells over the
past decades will greatly accelerate progress in this area. Furthermore, as seen in
fuel-cell work, temperature is the key to enhance the performance of fuel cells, indicating that optimizing the reaction temperature will yield signicant improvements
for CO2 reduction. This is another further possibility to increase productivity,
which is not possible in the photochemical approach.
In addition, to nd a suitable mechanism for an energy-efcient electron transfer
that avoids overpotentials, the technical feasibility of CO2 conversion to fuels from
the exploitation perspective depends on the possibility of forming under solventless
conditions liquid fuels such as long-chain hydrocarbons and/or alcohols, which can
390
be easily collected without the need to distillate from liquid solutions (a quite energyintensive process). It is thus necessary to pass from conventional electrochemical
conversion of CO2 in aqueous or nonaqueous solutions to the electrocatalytic conversion of CO2 in gaseous phase over solid electrodes. There are two possibilities,
related to the use of high or low temperatures. High-temperature CO2 conversions
are typically carried out using variations of the solid oxide fuel cell (SOFC), whereas
low-temperature systems largely utilize a variation of polymer electrolyte membrane
(PEM) fuel cells. A recent review [36] has discussed the use of SOFC-type cells for
the electrochemical CO2 conversion in comparison with low-temperature systems
based on transition-metal electrodes in aqueous or nonaqueous electrolytes. However, the low-temperature approach avoiding the use of a liquid electrolyte was
not discussed, even if it was shown that new breakthrough possibilities in terms of
type of products (long-carbon-chain hydrocarbons and alcohols) exist from this
approach [37].
There are very few studies based on this approach, but it was demonstrated that
using nanostructured carbon-based electrodes, it is possible to electrocatalytically
reduce CO2 in the gas phase using the protons owing through a membrane [38].
Long-chain hydrocarbons and alcohols up to C9C10 are formed, with preferential
formation of isopropanol using carbon-nanotube-based electrodes [14, 39]. Productivities are still limited, but these results demonstrate the concept of a new approach
to recycle CO2 back to fuels.
The features of the electrode used in this gas-phase electrocatalytic reduction of
CO2 are close to those used in PEM fuel cells [37, 40, 41] (e.g. a carbon cloth/Pt
or Fe on carbon black/Naon assembled electrode, GDE). The electrocatalysts are
Pt or Fe nanoparticles supported on nanocarbon (doped carbon nanotubes), which
is then deposited on a conductive carbon cloth to allow the electrical contact and
the diffusion of gas phase CO2 to the electrocatalyst. The metal nanoparticles are
at the contact of Naon, through which protons diffuse. On the metal nanoparticles,
the gas-phase CO2 reacts with the electrons and protons to be reduced to longerchain hydrocarbons and alcohols, the relative distributions of which depend on the
reaction temperature and type of metal nanoparticles. Isopropanol forms selectively
from the electrocatalytic reduction of CO2 using a gas diffusion electrode based on
an Fe/N carbon nanotube (Fe/N-CNT) [14, 39, 40]. Not only the nature of carbon
is relevant, but also the presence of nanocavities, which could favor the consecutive
conversion of intermediates with formation of C-C bonds.
The SOFC cell approach is instead used for methane wet reforming with CO2 to
produce syngas [42]. These high-temperature cells typically utilize a yttria-stabilized
zirconia (YSZ) electrolyte or tube with various metal and/or mixed-metal catalysts at
the anode and cathode. When CO2 is used as the fuel without CH4, CO alone can be
synthesized at the cathode along with oxygen at the anode [43]. Current densities of
more than 1 Acm2 could be obtained by this high-temperature CO2 electrolysis
approach, but the high temperature necessary (more than 900C to have enough
ionic conductivity in the YSZ membrane) and the possibility of obtaining only CO
or syngas, which should be thus further converted, greatly limit the applicability of
this approach.
391
392
production), were discussed in a chapter of a recent book on sustainable energy technologies [49]. However, this type of approach may be adapted with difculty to the
reduction of CO2, and up to now, the attempts in this direction were quite limited.
In general, tandem or multijunction cells are particularly interesting since the
absorption efciency can be enhanced by complementary spectrum absorption,
while a corrosion-resistant material can be used as the top junction (e.g. WO3 on
GaInP2). For development of efcient tandem cell structures that are based on nonoxidic semiconductors, stabilization strategies are the key challenge because semiconductors are, in general, not thermodynamically stable at the reactive electrolyte
interface. Physical surface passivation could be realized using the nanoemitter
concept [46], rst developed for n-Si for operation in the photovoltaic mode of an
electrochemical solar cell [50] yielding efciencies greater than 10%.
When in contact with redox electrolytes, the Fermi level of the metallic nanoemitters adjusts to that of the solution. If the metal nanoislands (size less than approximately 50 nm) form rectifying junctions without Fermi-level pinning, the system is
characterized by a rectifying junction between the electrolyte and the semiconductor
absorber. The decoupling of the metal/semiconductor junction, which is often impeded by metal-induced gap states, can be achieved by spatial separation introducing, for example, an ultrathin interfacial lm that inhibits interpenetration of
evanescent metal states into the semiconductor energy gap. Upon illumination, the
transparent passive layer transmits the sunlight, which is absorbed in the underlying
semiconductor. Using p-Si and p-InP thin lms, solar-to-hydrogen efciencies reach
12.1% for homoepitaxial InP thin lms covered with Rh nanoislands [46]. There are
other strategies for solving the issue of stability of the photoanode in PEC cells, such
as (a) chemical passivation of the surface, as, for instance [46], by H-termination or
methylization, and (b) stabilization by energy band structure properties. The latter
has been found in transition-metal chalcogenide materials where the upper valence
band and the bottom of the conduction band are composed predominantly of metallic
d-states.
The hydrogen production efciency in a photoelectrochemical cell was already demonstrated a decade ago to be more than 12% by Khaselev and Turner [51] (i.e. two
to three times higher than that reported by the recent cited approach by Nocera and
coworkers [45]). However, performances of the Turner cell (schematically reported in
Figure 5.1.7) fast degrade with time. However, the approach described previously
will be a good way to have more stable performances. The Turners PEC/ PV cell
shown in Figure 5.1.7 could be modied to have separate O2 and H2 production introducing a membrane (which introduces extraresistances), but transferring this kind
of cell to an articial leaf device, which could be produced on a larger scale, is difcult, due to the many difculties in sealing a system with a corrosive electrolyte
and with evolving gases.
These strategies are interesting for the production of H2 and O2 from water by a
photoelectrochemical approach, but in the case of CO2, it is necessary to (1) avoid
the use of a liquid electrolyte (to eliminate problems of CO2 solubility, diffusion limitation due to double layer, solvent competition, and to simplify cell sealing and facilitate product recovery eliminating the solvent), (2) have the anodic and cathodic
reactions in separate compartments (reduce separation costs and eliminate safety
393
H2
p-GalnP2
Pt
n-GaAs
p-GaAs
O2
3M
H2SO4
Ohmic contact
Interconnet
Figure 5.1.7 Scheme of the Turners PEC/PV device. Adapted from [51].
issues), and (3) have a cell design suitable for ease of scaling-up and application (a
design like that shown in Figure 5.1.7 is not suited for at-type articial leaf solar
cells). Most of the articial-leaf-type solar cells proposed in literature show similar
issues [52].
A new PEC design was originally proposed by the authors of this chapter [3840]
as a further development of the original concept proposed by Hitachi Green Centre researchers [53]. The design ts the previously mentioned requirements and
can be used for the reduction of carbon dioxide using the protons/electrons generated
at a photoanode. The advantage is that this solar-fuel-cell has a conguration close
to commercial PEM fuel cells, which are already optimized in terms of engineering
and mass/charge transfer. There is a signicant difference in this new design with
respect to the PEC cell proposed Hitachi Green Centre [53]: the CO2 electrocatalyst operates in gas phase, instead of liquid phase. The difference in cell operations
leads not only to an easier product recovery, but also to a different type of electrode
(a at-type copper electrode in the case of Ichikawa et al. [53], while metallic nanoparticles supported on nanocarbon materials (as described in the previous section)
were used by the authors of this chapter [37, 40, 54].
The scheme of this solar fuel cell is discussed in the following section, but it should
be remarked that similarly to some other recent advanced solar leaf cell designs [52],
it is necessary to have a porous photoanode that combines high effectiveness in sunlight harvesting with fast collection of the electrons produced and an effective transport of the protons generated to the proton-diffusion membrane on which the
photoanode is integrated. This is a key property determining the cell performances
and that is not present in the PEC schemes similar in design to the cell shown in
394
Figure 5.1.7, including variations of the dye-sensitized solar cells originally proposed
by Grtzel [55] and expected for use in combination with hematite (-Fe2O3)
photoelectrode material for producing H2/O2 by water photoelectrolysis [56].
A porous photoanode for PEC solar cells must t a number of requirements:
It must be cost-effective and easily scaled to larger sizes.
It must be robust for stable operations.
It must have an optimal nanostructure allowing (1) an enhanced light harvesting
(possible over the entire sunlight spectrum, with effective use of the radiation
for creating a photocurrent); (2) low rate of charge recombination and reduced interfaces/grain boundaries, which favors the charge recombination; (3) negligible defects and centers that favor thermal or radiative pathways (which reduce quantum
efciency); and (4) fast transport of the electrons to a conductive substrate.
It must have a porous nanostructure that allows a fast transport of protons (generated from water oxidation) to the underlying proton-conductive membrane,
avoid surface recombination between protons and electrons (which should have
different paths of transport), and have an optimal interface with the membrane.
It is thus necessary to use knowledge in the synthesis and mastering of semiconductor nano-objects to realize these demanding characteristics in the photoanode [57]. In
particular, it is necessary to realize a semiconductor lm characterized by the presence of an array of vertically aligned nanotubes that allow good light harvesting,
limit charge recombination at the grain boundaries with respect to an assembly of
nanoparticles while maintaining a high geometrical surface area necessary to
improve the photoresponse, have in principle separate paths for electron and proton
movement, and have a good porosity and interface with the membrane to allow a
good transport of the protons [13, 58]. This concept is schematically presented in
Figure 5.1.8.
A good method to achieve these characteristics and suitable also for scale-up is the
anodic oxidation of titanium thin foils to form ordered arrays of vertically aligned
titania nanotubes [5862].
The process of anodic anodization used to prepare the nanostructured titania thin
lms may be summarized as follows. When the Ti foil, after the preliminary cleaning
treatment, is immersed in the electrolyte for the anodization process, a fast surface
oxidation occurs with the formation of a thin TiO2 layer. This process may be monitored by a fast decrease of the current, the TiO2 layer being not conductive. Due to
the presence of an aqueous solution of HF as an electrolyte, the solubilization of Ti4+
ions and/or of small TiO2 particles starts simultaneously with the formation of the
oxide layer. These processes lead to the formation of holes, which locally modify
the electrical eld and induce from one side the acceleration of the process of dissolution, due to eld-enhanced effects, and from the other side the oriented growing of
the one-dimensional structure (nanotubes, nanorods, etc.). The electrolyte and conditions of anodization have a relevant inuence not only on the lm thickness but also
on the specic nanostructure obtained [5862], which in turn strongly inuence both
the stationary and transient photoresponse. The thicker walls prepared in the
organic electrolyte allow a fast electron transport, but it is also necessary to avoid
the presence of defects in the titania, which trap electrons and reduce the photocurrent.
395
20 m
electrode
200 nm
electrode
Figure 5.1.8 Scanning electron microscopy (SEM) images (cross section) of a TiO2 lm de-
posited over a conductive support by the solgel method (A) or produced by anodic oxidation
of a Ti foil in ethylene glycol containing 0.3 wt% NH4F and 2 vol% H2O applying a 50 V
potential for 6 h. The thickness of the titania nanostructured lm is approximately 14 m;
tube internal diameter is approximately 40 nm, while tube external diameter is approximately
100 nm. (The latter SEM image is adapted from [14].) On the right side of the gure, schematic
model show the differences between TiO2 thin lms and the aligned nanotube array of TiO2.
Adapted from [58].
396
Filter
Assembly of photoanode
with the naflon membrane
Solar
simulator
PEC cell
e
CO2
reduction
to fuels
or
CO2
H2O
CH3CH(OH)CH3
2H+
H2
production
H2
H+
e
e
e
e
e
O2
Light
Electrocatalyst
Photoanode
PhotoMembrane
Figure 5.1.9
PEC solar cell. Bottom: scheme of the cell with electron microscopy images of a
particular of the TiO2-nanotube array electrode and of the Fe nanoparticles on N-doped carbon nanotubes, used as a photocatalyst for water oxidation and an electrocatalyst for CO2
reduction, respectively. It is also shown that it may be possible to use this cell for the production of H2/O2 in separate compartments by water photoelectrolysis. Top: photo of the experimental cell and of the assembly of the photoanode with the Naon membrane. Adapted from
[14, 40, 52].
and especially (4) changes the type of products that are formed. The electrochemical
reduction of CO2, either in aqueous or organic electrolytes, gives mainly C1 products
of reduction and is limited in productivity by the solubility of CO2 in the electrolyte
and the mass transfer in solution [1518]. Conversely, gas-phase operations allow formation of >C2 products [14, 3840] (never detected in liquid-phase operations) and
eliminate the problem of solubility of CO2 and mass transfer in liquid phase, as
well avoid the formation of double-layer and related effects. The electrocatalyst,
however, should be different. Instead of using conventional electrodes, the solar
fuel cell shown in Figure 5.1.9 uses an electrode based on metal nanoparticles dispersed over conductive doped CNTs and then deposited over a carbon-cloth conductive material acting as electron transport net [40, 41]. Using a Fe/N-CNT-based gas
5.1.7 Conclusions
397
diffusion electrode, it was shown that isopropanol is formed, being the main reaction
product of CO2 electrocatalytic reduction [14, 3840].
The photoanode (Section 5.1.5) is instead based on an array of vertically aligned
doped titania nanotubes, produced by anodic oxidation of thin Ti layers [13, 14,
5860], in order to meet the demanding requirements discussed in Section 5.1.4.
Due to self-doping during preparation and/or the creation of surface phononic heterostructures (by deposition of very small gold nanoparticles), these titania nanotube ordered thin lms are active in the visible light region [63], although still
not with optimal performances. Commercial Naon is used as the proton-conductive
membrane. There are still a number of problems at the interface between this membrane and the titania photanode, and thus effectiveness in the transport of protons
is still limited. In addition, transient measurements [14] indicate the presence of a
signicant surface quenching process (associated to surface peroxo-species), which
limits the steady-state productivity in water oxidation. This cell design could be
transferred to application as articial leaf, but productivity of both electrodes has
to be improved in terms of (1) the response to visible light, (2) reducing surface
self-quenching during reaction, (3) the presence of several interfaces that limit the
mass/charge-transfer and cell efciency, and (4) the rate of CO2 reduction. A new
advanced cell design has been proposed and is under investigation to solve these
issues [52].
5.1.7 Conclusions
Photoelectrochemical activation of CO2 in articial-leaf-type PEC cells is still a longterm goal for the practical, widespread use of the great potential of sunlight and to
collect energy naturally, but we have shown here how it is potentially the solution to
make sustainable energy an approach compatible with the large investments made
for the energy chain and infrastructure. To avoid intermittency of solar energy, it is necessary to design systems that directly capture CO2 and convert it into liquid solar fuels,
which can be easily stored. Chemical energy storage is the key for sustainable use of
solar energy.
Even if there is great interest in research on these topics, we recommend the need
for a radically different approach, particularly in terms of system design. This is
because often current studies, although scientically valuable, are not focused on
the key issues, particularly regarding the question of transferring the results to a practical use. Due to the complexity of the problems, a fundamental understanding is the
key for advancing, but taking into consideration the system engineering and integration and the following three keywords to designing articial leaves: smart, cheap, and
robust. The fast advances in the development of nanotailored materials will be certainly pivotal to progress in this eld, when combined with the integration between
catalysis and electrode concepts, and breakthroughs in the understanding of the reaction mechanisms of these fast surface processes. This requires integration of theory/
modeling to new experimental tools and designed experiments on model systems, but
with the guidelines given from a dened design of advanced PEC articial-leaf-type
solar cells.
398
5.1.8 References
1. International Energy Agency (IEA). World energy outlook 2011. Paris, France: IEA Pub;
2011.
2. Centi G, Perathoner S. Towards solar fuels from water and CO2. ChemSusChem. 2010;
3:195208.
3. Zttel A, Borgschulte A, Schlapbach L. Hydrogen as a future energy carrier. Weinheim,
Germany: Wiley-VCH; 2008.
4. Farrauto RJ. Building the hydrogen economy. Hydrocarbon Eng. 2009;14:2530.
5. Nowotny J, Sheppard LR. Solar-hydrogen. Int J Hydrogen Energy. 2007;32:26078.
6. Bockris OM. Hydrogen no longer a high cost solution to global warming: new ideas. Int J
Hydrogen Energy. 2008;33:212931.
7. Olah GA, Goeppert A, Surya Prakash GK. Beyond oil and gas: the methanol economy.
2nd ed. Weinheim, Germany: Wiley-VCH; 2009.
8. Centi G, Perathoner S, Passalacqua R, Ampelli C. Solar production of fuels from water
and CO2. In: Muradov NZ, Veziroglu TN, editors. Carbon-neutral fuels and energy carriers. Boca Raton, FL: CRC Press (Taylor & Francis Group); 2012. p. 291323.
9. Centi G, Iaquaniello G, Perathoner S. Can we afford to waste carbon dioxide? Carbon
doxide as a valuable source of carbon for the production of light olens. ChemSusChem.
2011;4:126573.
10. Balzani V, Credi A, Venturi M. Photochemical conversion of solar energy. ChemSusChem.
2008;1:2658.
11. Amao Y. Solar fuel production based on the articial photosynthesis system. ChemCatChem. 2011;3:45874.
12. Moore GF, Brudvig GW. Energy conversion in photosynthesis: a paradigm for solar fuel
production. Annu Rev Condens Matter Phys. 2011;2:30327.
13. Centi G, Passalacqua R, Perathoner S, Su DS, Weinberg G, Schlgl R. Oxide thin lms
based on ordered arrays of one-dimensional nanostructure. A possible approach toward
bridging material gap in catalysis. Phys Chem Chem Phys. 2007;9:49308.
14. Ampelli C, Centi G, Passalacqua R, Perathoner S. Synthesis of solar fuels by a novel
photoelectrocatalytic approach. Energy Environ Sci. 2010;3:292301.
15. Lvov SN, Beck JR, LaBarbera MS. Electrochemical reduction of CO2 to fuels. In Muradov NZ, Veziroglu TN, editors. Carbon-neutral fuels and energy carriers. Boca Raton,
FL: CRC Press (Taylor & Francis Group); 2012. p. 363400.
16. Gattrell M, Gupta N, Co, A. A review of the aqueous electrochemical reduction of CO2 to
hydrocarbons at copper. J Electroanal Chem. 2006;594:119.
17. Whipple DT, Kenis PJA. Prospects of CO2 utilization via direct heterogeneous electrochemical reduction. J Phys Chem Lett. 2010;1:34518.
18. Bockris JOM, Wass JC. The photoelectrocatalytic reduction of carbon dioxide. J Electrochem Soc. 1989;136:25218.
19. Rosen BA, Salehi-Khojin A, Thorson MR, et al. Ionic liquidmediated selective conversion of CO2 to CO at low overpotentials. Science. 2011;334:6434.
20. Hori Y, Konishi H, Futamura T, et al. Deactivation of copper electrode in electrochemical reduction of CO2. Electrochim Acta. 2005;50:535469.
21. Millet P, Dragoe D, Grigoriev S, Fateev V, Etievant C. GenHyPEM: a research program
on PEM water electroysis supported by the European Commission. Int J Hydrogen
Energy. 2009;34:497482.
22. Turner J, Sverdrup G, Mann MK, et al. Renewable hydrogen production. Int J Energy
Res. 2008;32:379407.
5.1.8 References
399
23. Gandia LM, Oroz R, Ursua A, Sanchis P, Dieguez PM. Renewable hydrogen production:
performance of an alkaline water electrolyzer working under emulated wind conditions.
Energy Fuels. 2007;21:1699706.
24. Shibata H, Moulijn JA, Mul G. Enabling electrocatalytic Fischer-Tropsch synthesis from
carbon dioxide over copper-based electrodes. Catal Lett. 2008;123:18692.
25. Gonalves MR, Gomes A, Condeo J, et al. Selective electrochemical conversion of CO2
to C2 hydrocarbons. Energy Conv Manag. 2010;51:302.
26. Cokoja M, Bruckmeier C, Rieger B, Herrmann WA, Khn FE. Transformation of carbon
dioxide with homogeneous transition-metal catalysts: a molecular solution to a global
challenge? Angew Chem Int Ed. 2011;50:851037.
27. Quadrelli EA, Centi G, Duplan JL, Perathoner S. Carbon dioxide recycling: emerging
large-scale technologies with industrial potential. ChemSusChem. 2011;4:1194215.
28. Centi G, Perathoner S. Opportunities and prospects in the chemical recycling of carbon
dioxide to fuels. Catal Today. 2009;148:191205.
29. Sakakura T, Choi JC, Yasuda H. Transformation of carbon dioxide. Chem Rev. 2007;
107:236587.
30. Morris AJ, McGibbon RT, Bocarsly AB. Electrocatalytic carbon dioxide activation:
the rate-determining step of pyridinium-catalyzed CO2 reduction. ChemSusChem. 2011;
4:1916.
31. Tanaka K, Ooyama D. Multi-electron reduction of CO2 via Ru-CO2, -C(O)OH, -CO,
-CHO, and -CH2OH species. Coord Chem Rev. 2002;226:2118.
32. Seshadri G, Chao L, Bocarsly AB. A new homogeneous electrocatalyst for the reduction
of carbon dioxide to methanol at low overpotential. J Electroanal Chem. 1994;372:14550.
33. Barton EE, Rampulla DM, Bocarsly AB. Selective solar-driven reduction of CO2 to methanol using a catalyzed p-GaP based photoelectrochemical cell. J Am Chem Soc. 2008;
130:63424.
34. Brennecke JF, Gurkan BE. Ionic liquids for CO2 capture and emission reduction. J Phys
Chem Lett. 2010;1:345964.
35. Zhao GY, Jiang T, Han BX, et al. Electrochemical reduction of supercritical carbon dioxide in ionic liquid 1-n-butyl-3-methylimidazolium hexauorophosphate. J Supercrit Fluids.
2004;32:28791.
36. Spinner NS, Vega JA, Mustain WE. Recent progress in the electrochemical conversion
and utilization of CO2. Catal Sci Technol. 2012;2:1928.
37. Centi G, Perathoner S. Carbon nanotubes for sustainable energy applications. ChemSusChem. 2011;4:91325.
38. Centi G, Perathoner S, Wine G, Gangeri M. Electrocatalytic conversion of CO2 to long
carbon-chain hydrocarbons. Green Chem. 2007;9:6718.
39. Gangeri M, Perathoner S, Caudo S, et al. Fe and Pt carbon nanotubes for the electrocatalytic conversion of carbon dioxide to oxygenates. Catal Today. 2009;143:5763.
40. Centi G, Perathoner S. Nanostructured electrodes and devices for converting carbon dioxide back to fuels: advances and perspectives. In: Zang L, editor. Energy efciency and
renewable energy through nanotechnology. London, UK: Springer-Verlag; 2011. p. 56184.
41. Centi G, Perathoner S. The role of nanostructure in improving the performance of electrodes for energy storage and conversion. Eur J Inorg Chem. 2009;26:385178.
42. Kim T, Moon S, Hong S-I. Internal carbon dioxide reforming by methane over Ni-YSZCeO2 catalyst electrode in electrochemical cell. Appl Catal A Gen. 2002;224:11120.
43. Bidrawn F, Kim G, Corre G, Irvine J, Vohs J, Gorte R. Efcient reduction of CO2 in a
solid oxide electrolyzer. Electrochem Solid-State Lett. 2008;11:B16770.
44. Reece SY, Hamel JA, Sung K, et al. Wireless solar water splitting using silicon-based
semiconductors and earth-abundant catalysts. Science. 2011;334:6458.
400
45. Pijpers JJH, Winkler, MT, Surendranath Y, Buonassisi T, Nocera DG. Light-induced
water oxidation at silicon electrodes functionalized with a cobalt oxygen-evolving catalyst.
Proc Natl Acad Sci U S A. 2011;108:1005661.
46. Lewerenz HJ, Heine C, Skorupska K, et al. Photoelectrocatalysis: principles, nanoemitter
applications and routes to bio-inspired systems. Energy Environ Sci. 2010;3:74860.
47. Currao A. Photoelectrochemical water splitting. Chimia. 2007;61:8159.
48. Neumann B, Bogdanoff P, Tributsch H. TiO2-protected photoelectrochemical tandem
Cu(In,Ga)Se2 thin lm membrane for light-induced water splitting and hydrogen evolution. J Phys Chem C. 2009;113:209809.
49. Van de Krol R, Schoonman J. Photo-electrochemical production of hydrogen. In: Hanjali K, Van de Krol R, Leki A, editors.Sustainable energy technologies. Dordrecht, Netherlands: Springer; 2008. p. 12142.
50. Stempel T, Aggour M, Skorupska K, Munoz A, Lewerenz H-J. Efcient photoelectrochemical nanoemitter solar cell. Electrochem Commun. 2008;10:11846.
51. Khaselev O, Turner JA. A monolithic photovoltaic-photoelectrochemical device for hydrogen production via water splitting. Science. 1998;280:4257.
52. Bensaid S, Centi G, Garrone E, Perathoner S, Saracco G. Articial leaves for solar fuels
from CO2. ChemSusChem. 2012;5:50021.
53. Ichikawa S, Doi R. Hydrogen production from water and conversion of carbon dioxide to
useful chemicals by room temperature photoelectrocatalysis. Catal Today. 1996;27:2717.
54. Su D, Centi G. Carbon nanotubes for energy applications. In: Rios G, Centi G, Kanellopoulos N, editors. Nanoporous materials for energy and the environment. Singapore: Pan
Stanford Pub.; 2012. p. 173202.
55. Grtzel M. Dye-sensitized solar cells. J Photochem Photobiol C Photochem Rev. 2003;
4:14553.
56. Sivula K, Le Formal F, Grtzel M. Solar water splitting: progress using hematite (-Fe2O3)
photoelectrodes. ChemSusChem. 2011;4:43249.
57. Centi G, Perathoner S. Creating and mastering nano-objects to design advanced catalytic
materials. Coord Chem Rev. 2011;255:148098.
58. Centi G, Perathoner S. Nano-architecture and reactivity of titania catalytic materials. Part 2.
Bidimensional nanostructured lms. Catalysis. 2009;21:82130.
59. Ampelli C, Passalacqua R, Perathoner S, Centi G, Su DS, Weinberg G. Synthesis of TiO2
thin lms: relationship between preparation conditions and nanostructure. Top Catal.
2008;50:13344.
60. Perathoner S, Passalacqua R, Centi G, Su DS, Weinberg G. Photoactive titania nanostructured thin lms. Synthesis and characteristics of ordered helical nanocoil array.
Catal Today. 2007;122:313.
61. Grimes CA, Mor GK. TiO2 nanotube arrays: synthesis, properties, and applications. Heidelberg: Springer; 2009.
62. Schmuki P. Self-organized oxide nanotube layers on titanium and other transition metals.
In: Schmuki P, Virtanen S, editors. Electrochemistry at the nanoscale. New York: Springer
Science; 2009. p. 43566.
63. Centi G, Perathoner S. Nanostructured titania thin lms for solar use in energy applications. In: Rios G, Centi G, Kanellopoulos N, editors. Nanoporous materials for energy
and the environment. Singapore: Pan Stanford Pub.; 2012. p. 25782.
5.2.1 Introduction
Using sunlight in concentrated solar power (CSP) systems is an environmentally
friendly and sustainable way to generate electricity. These systems consist of a multitude of sun-tracking mirrors (heliostats) to concentrate the incident sunshine onto a
small area, where a receiver is placed. In this receiver, the reected energy, or more
precisely a part of it, is absorbed into a working uid, which usually drives a turbine
to generate electrical power. Alternatively, the concentrated solar power can provide
heat for energy-demanding chemical processes, as extremely high temperatures up to
3,773 K can be reached. Therefore, CSP can be a sustainable energy source for endothermic high-temperature reactions such as the splitting of CO2 (Equation [1]) or
H2O (Equation [2]):
1
O2
2
1
H2O H2 + O2
2
CO2 CO +
(1)
(2)
The promising carbon-dioxide splitting (CDS) reaction is discussed in detail in the following. It is generally performed as a two-step process using thermochemical cycles requiring a suitable metal oxide, which can be easily reduced and reoxidized. In the rst
step of the redox cycle, the metal oxide is heated by CSP to a high temperature, which
is sufcient for a fast release of lattice oxygen (Equation [3]). In the second step, the
reduced species is cooled down and reoxidized by CO2 resulting in the formation of
CO and the initial metal oxide (Equation [4]). The reduced species can either be the
metal or a stoichiometric or nonstoichiometric oxide. This CO2 cycle is completely
analogous to CSP-based water splitting for the production of solar hydrogen:
y
MOx ! MOxy + O2
2
MOxy + yCO2 ! MOx + yCO
1
CO2 ! CO + O2
2
(3)
(4)
(3) + (4)
In this way, the unwanted back reaction and the separation of the reaction products are
avoided. Before presenting planned industrial applications of solarthermal cycles for the
CO2 splitting, the fundamental aspects are discussed to understand the challenges and
advantages of this process.
402
In every industrial process, efciency is the main goal. Particularly for the production of a commodity such as solar fuel, the productivity should be as high as possible.
The denition of productivity rP in chemical processes is given by Equation (5):
rP =
nP,out nP,in
tr + td
(5)
The produced molar amount of the product P determined by the difference of the
amount at the end (nP,out) and the beginning (nP,in) of the reaction is divided by
the total operating time of the reactor. The overall operating time is the sum of
the time for the reaction (tr) and the downtime (td), which is needed to ll and
empty a reactor, and to heat it up and cool it down. For the production of CO by
thermochemical CO2 splitting only during the second cycle (Equation [4]), the valuable
product CO is generated, but the whole time for reduction in step one (Equation [3])
and for heating and cooling also has to be taken into account for the productivity. This
is a remarkable drawback of all cyclic processes.
403
and reacts with it forming a product, which is also a uid (Equation [4]), the solid
particles may either remain unchanged in size during the reaction, because they
form a rm product material (MOx), or they may shrink in size. The reaction occurs
rst at the outer surface of the solid particle, then passing into the bulk of a solid
material, leaving behind completely converted material (e.g. inert solid), which is
called ash [1]. At any time during the reaction, an unreacted core exists, which
shrinks in size [1]. Gas-solid reactions proceeding according to the shrinking
unreacted core model usually occur in ve steps [1]:
1. Diffusion of gaseous reactant (e.g. CO2) through the lm surrounding the particle
to the surface of the solid
2. Diffusion of the reactant through the blanket of ash (e.g. MOx) to the surface of
the unreacted core (e.g. MOx-y; representing the reaction surface)
3. The chemical reaction of the reactant with the solid
4. Diffusion of the gaseous product (e.g. CO) through the ash back to the surface of
the solid
5. Diffusion of the gaseous product through the gas lm into the main body of the uid
Solid-state diffusion, which is involved in the release of oxygen, proceeds generally
through the movement of point defects. The vacancy mechanism, the interstitial
mechanism, and the interstitialcy mechanism can occur depending on the distortion
of the solid lattice and the nature of the diffusing species. When one of the steps 15
is the slowest step representing the major resistance, that step is the rate-controlling
one, which is not necessarily the chemical reaction (step 3).
Additionally, the rate of heat transfer may also become important. Nonuniform
temperature distributions within the solid particles result in differing local rates of
reaction, as the reaction rates are strongly depending on the temperature according
to the Arrhenius law. Heat- and mass-transfer effects become increasingly important
with increasing rates of reaction [1]. Whereas the macroscopic kinetics describe the
rate of a chemical reaction, thermodynamics determines the maximum extent to
which reactions can occur. Provided that the rate of reaction is sufciently fast,
the thermodynamical equilibrium can be reached.
404
reaction into a two-step process involving the redox cycling of an appropriate metal
oxide as mentioned previously. The concept of a simplied metal oxide redox cycle
for CO2 splitting is illustrated in Figure 5.2.1. The metal oxide is heated to a high
temperature, which is sufcient for the release of O2 (Equation [3]). The reduced species is then cooled down and reoxidized by CO2 resulting in formation of CO and the
initial metal oxide (Equation [4]). In this way, oxygen is removed from the gas phase
forming strongly bound lattice oxygen.
The reduction of any metal oxide according to Equation (3) is generally endothermic and has a positive reaction entropy. As a consequence, high temperatures are
needed for the release of O2 yielding the reduced state. The reaction of the reduced
species with CO2 is slightly exothermic. Both reactions sum up to the overall CO2splitting reaction (Equation [1]). Accordingly, GCDS, the Gibbs enthalpy change
of the carbon dioxidesplitting reaction, is the sum of the two reactions:
GCDS = G3 + G4
(6)
rfuel Hfuel
Psolar + rinert Einert
O2
(7)
MOx-y
CO2
Concentrated
solar radiation
MOx
CO
(+C)
405
300
250
G/kJ mol 1
FeO
Fe3O4
200
150
100
50
0
50
500
1000
1500
Temperature/K
2000
2500
Figure 5.2.2
Plot of Gibbs free energy changes of the thermal reduction (green line) (Fe3O4 =
3 FeO + 1/2 O2), FeO oxidation (red line) (3 FeO + CO2 = Fe3O4 + CO), and overall reaction
(blue line) (CO2 = CO + 1/2 O2) for the iron-oxide-based cycle. The vertical lines show the melting points of the indicated iron-oxide phases.
It depends on the molar fuel production rate (rfuel), the heating value of the fuel
(Hfuel), the incident solar radiation power (Psolar), the ow rate of the inert gas during thermal reduction (rinert), and the energy (Einert) that is required for separation. In
a rough efciency analysis (second-law analysis), several simplifying assumptions are
usually made [3]:
1.
2.
3.
4.
5.
The main experimentally measured energy losses can be due to heat conduction
through the reactor walls or reradiation through the aperture and are therefore
mainly related to the reactor design [2]. A concomitant advantage of cycling is the
simplied downstream processing. Because CO and O2 are obtained in different
steps, an additional product separation is not needed. While the general concept
seems to be promising due to its simplicity, the practical realization of the cyclic process
is associated with some problems caused by the severe reaction conditions, especially in
the reduction of the metal oxide (Equation [3]).
First, we dene an ideal cycle and the material properties from a chemists point of
view that should be met for high performance and economic efciency. The most
important parameters for the evaluation of these processes with respect to efciency
are productivity (Equation [5]) and stability. However, it appears that with respect to
materials chemistry and physics these properties represent opposing factors that is,
the optimization of productivity is likely to decrease stability and lifetime, and vice
versa.
406
Of course, the basic requirement of the metal oxide applied in the thermochemical
cycle for CO2 splitting is the feasibility of the thermal reduction and the reoxidation
by CO2 in reasonable temperature ranges. Regardless of the impact of very high temperatures on important properties of the thermochemical cycle, this demand is also
related to the availability of refractory construction materials in terms of mechanical,
but also chemical stability under these conditions.
In addition to the productivity, the stability of the working oxide is a key issue.
Ideally, the cycle possesses unlimited repeatability without loss of performance and
capacity [4]. This requirement implies that the working oxide must not be trapped
in an inactive form as a consequence of high-temperature reduction and oxidation
by CO2. Scenarios accounting for deactivation are phase changes, for example the
formation of more dense phases with lower gas diffusivity, or the formation of
mixed oxides by the reaction with surrounding insulation materials. These are usually
less prone to thermal reduction or possess unfavorable material properties.
An additional important property is the oxygen storage/release capacity of the applied
metal oxide. It quanties the amount of oxygen, which can be released during the reduction step and stored during the reoxidation step. During the thermal reduction for any
real system the oxygen evolution is terminated at some point caused by thermodynamic
or kinetic limitations. On the one hand, the oxygen release may be limited by reaching
thermodynamic equilibrium conversion resulting in no further net oxygen evolution, or,
in other words, oxygen incorporation with identical reaction rate. On the other hand, the
resulting metal or metal oxide layer produced during the thermal reduction may act as a
barrier for the oxygen anions to diffuse out of the metal oxide bulk.
The kinetics of the process is important in view of economic viability, because it is
the crucial factor for productivity. For an assessment the complete macrokinetics of
both reaction steps have to be considered (i.e. heat and mass transfer as well as the
rates of the gas-solid reactions). Thermal reduction of an oxide and its reoxidation by
CO2 are reversible gas-solid reactions. The reaction rate of a solid reacting with a gas
to form another solid material is described by Equation (8):
r = k f () f ( px)
(8)
Factors inuencing the rate of reduction are the temperature, the partial pressure of
O2 in the gas phase and , a factor representing the degree of reduction of the solid,
which depends on the chosen oxide and its reduction mechanism. It is quite clear
that fast kinetics demand temperatures as high as possible. Due to the temperature
dependence of k according to the Arrhenius law the slope of k versus T decreases at
very high temperatures, which means that a further increase of temperature leads to
an underproportional gain. High O2 concentrations result in high rates of the reverse
reaction, that is, oxidation. High local O2 concentrations can originate from nonideal
ow conditions in the reactor such as dead volumes and bypass. Mass transfer limitations can also lead to high O2 concentrations and are partially resolved by increasing
the ow rate of the inert carrier gas. High ow rates also result in a thinner stagnating
boundary layer, in which transport only occurs via slow molecular diffusion.
The most important factor for the process design is the metal oxide-dependent
term f(). In a simplied scheme the whole accessible surface of the particle takes
407
part in the reaction, i.e. the complete gas-solid interface. Consequently, large surface
areas result in fast kinetics. Here, an intrinsic problem becomes obvious: while high
temperatures are required to enable thermal reduction and to enhance reaction rates,
they also lead to increased sintering close to the Tammann temperature and loss of
surface area resulting in lower rates. The gas-solid reaction results in the formation of
vacancies at the surface, and the diffusion of these into the bulk of the grain and the
countercurrent diffusion of oxygen anions to the surface can be a limiting factor.
Furthermore, the formation of a dense surface layer can effectively suppress the
ongoing reaction and limit the conversion in each step.
Cycle duration can also emerge as a critical factor. It may be regarded as advantageous to perform short cycles limiting the reduction to surface layers, which are
more easily reduced, thus limiting the required temperatures and avoiding bulk diffusion limitations. The counterpart is complete reduction at high temperatures with
eventually slow rates at high degrees of conversion. While the former strategy operates at high reaction rates, it only uses a small fraction of the active metal oxide and
leads to high thermal stress in the material due to fast heating and cooling. The latter
uses most of the active mass and therefore has a higher yield per cycle, but the mean
reaction rates and productivity may be lower.
The reaction rates of the thermal reduction and the reoxidation by CO2 are increased by high oxygen anion conductivity and high surface areas. Oxygen anion
conductivity is a function of temperature, crystal structure, and defects. Because cycling results in stoichiometric gas-solid reactions, the gas-solid interface can be a crucial parameter depending on the reaction conditions. Whether gas-solid, intraparticle
mass transfer, or surface chemical processes are rate-limiting is primarily determined
by the reaction temperature and gas ow rates.
Another crucial point is the stability of the material, because reduction requires
sufcient mobility of lattice oxygen anions, which is usually observed at temperatures, where substantial sintering or volatilization of reactant or product oxides or
metals occurs. It is evident that oxygen mobility and the stability against sintering
are at least in part conicting, and it has already been mentioned that high temperatures can be sufcient to facilitate undesired solid state reactions with the surrounding reactor materials. Obviously, it is of limited use to employ a high surface area
oxide with excellent reducibility in terms of oxygen release capacity and reaction
rate, which undergoes extreme sintering and phases segregation within the rst cycles. Oxide structures are needed that can accommodate high oxygen defect densities
without structural collapse/disintegration. The phase stabilities of oxides can be improved by doping with other metal ions. Solid solutions are generated, which may
have a better stability and a higher oxygen mobility resulting in improved reducibilities and oxygen storage/release capacities. A famous example is the ceria-based system, where solid solutions of e.g. CexZr1-xO2, CexLa1-xO2, or CexHf1-xO2 are well
investigated [47].
Finally, specic material properties are required for efcient solarthermal cycles for CO2 activation. For solarthermal applications the metal oxides must
effectively absorb solar radiation and convert it into heat without substantial reradiation. Therefore, bright materials are unfavorable because they always cause
reradiation.
408
5.2.3.2 Examples
In the following three examples are presented, which are judged to be practicable. They
illustrate the different experimental approaches to meet the stringent requirements of
cyclic thermochemical CDS.
5.2.3.2.1 The ZnO/Zn System The solarthermal activation of ZnO has attracted
considerable attention for energy storage. In addition to thermochemical cycles the
production of elemental Zn for battery applications has been investigated within the
EU-funded project SolZinc. Thermochemical cycles for solar H2, CO, and combined
syngas production have been proposed:
ZnO(s) Zn(g) +
1
O2,g
2
(9)
(10)
The reduction was experimentally performed at approximately 1,900 K [8]. The melting and boiling points of elemental zinc are at 693 K and 1,180 K at 1 bar, respectively. Accordingly, the reaction products of ZnO dissociation are Zn(g) and O2,g,
which have to be quenched at high temperatures to prevent recombination. Obviously, the ZnO/Zn system is not in line with the demand for material stability of
the working oxide and requires extended processing to handle the multiple material
streams. For example, Chambon et al. [8] developed a moving-front reactor for solarthermal ZnO dissociation. This concept enabled the recovery of approximately
30% of the volatilized product as a powder consisting of nanoparticles containing
up to 50% of reduced Zn species.
The thermodynamics of the CO2-splitting reaction using Zn have been recently examined by Galvez et al. [3] including a second-law analysis to assess the theoretical
system efciency. For the complete conversion of Zn temperatures below 1,000 K are
mandatory, which are, however, far above the melting point and would lead to considerable vaporization of Zn. Concerning the formation of carbon or of CO, the
exothermic Boudouard equilibrium (2 CO S C + CO2) has to be taken into account.
The formation of elemental carbon is favorable for temperatures up to 1,273 K, while
CO formation is predicted to be feasible at temperatures higher than 800 K. The
second-law analysis yielded 39% solar-to-chemical energy conversion efciency for
an idealized cycle [3].
The kinetics of Zn oxidation by CO2 has been recently studied by Steinfeld and coworkers [9, 10] and by Abanades and Chambon [11]. The former group studied commercial Zn powders with low-surface area and found that conversion was incomplete at
temperatures between 673 and 748 K including the melting point of Zn because of
the formation of a dense ZnO layer inhibiting further oxidation of the Zn particles. Correspondingly, the application of low oxidant concentrations was found to be mandatory.
Furthermore, these authors observed a transition from a fast interface-controlled to a
slow diffusion-controlled regime, when the reactions reached completion. In contrast,
Abanades and Chambon [11] using high surface area Zn samples obtained from solarthermal reduction of ZnO found the reaction to be complete within several minutes.
409
In both cases the effect of temperature on the initial reaction rates in the studied range
was weak, while it became strong in the diffusion-controlled regime. Although thermodynamics predict the formation of some carbon for temperatures up to 973 K, this was
not observed experimentally indicating that its formation is kinetically hindered in the
studied temperature range.
Although the ZnO/Zn thermochemical cycle does not meet the initially formulated
criteria with respect to material stability, its properties render it attractive and promising. The volatilization of Zn (Equation [9]) imposes difculties in terms of product
separation and aerosol processing and is accompanied with an energy penalty for
quenching leading to low thermal efciencies. On the other hand, the solar reactor
for oxide dissociation can be operated continuously without temperature cycling
and reproducibly yields Zn metal with high surface area and favorable properties
for subsequent CO2 splitting, avoiding the degradation of the working oxide.
5.2.3.2.2 The Fe3O4 /FeO System The basic ferrite cycle operating with iron oxidation states between that of magnetite (Fe3O4) and wustite (FeO) is another promising
option for CO2 activation:
Fe3O4 3 FeO +
1
O2
2
(11)
(12)
410
enabling efcient heat transfer in the temperature cycling of the working oxide. An
increase of efciency up to 76% is predicted applying this technique [4].
The kinetics of FeO oxidation using H2O/CO2 mixtures was studied by Stamatiou
et al. [9, 10] in the temperature range from 973 to 1,273 K. The initial reaction rates increased with increasing H2O concentration and also with temperature. Degrees of conversion of FeO up to 0.8 were obtained in the diffusion-controlled reaction with full
conversion appearing to be within reach. A transition from a fast interface-controlled
to a slow diffusion-controlled regime was observed indicating that materials with
low-surface area or materials, which are prone to sintering, cannot be used.
Miller et al. [4] were able to perform successive cycles of thermal reduction and CO2
splitting without signicant loss of activity using a cobalt ferrite Co0.67Fe2.33O4 supported on yttria-stabilized zirconia (YSZ). The results indicated that a favorable temperature range for performing the reoxidation is from 1,373 to 1,473 K, because slow
kinetics limit below 1,373 K and unfavorable thermodynamics limit above 1,473 K.
The stability of YSZ-supported iron oxides was investigated by Coker et al. [12]
using in situ X-ray diffraction (XRD). It was found that utilization and reduction of
Fe oxide was especially effective, when it was dispersed in the YSZ matrix.
5.2.3.2.3 The CeO2/Ce2O3 System The redox properties of ceria and ceria-based materials have been studied in great detail because of their oxygen storage and release capacities, which are highly relevant for their application in modern three-way automotive
catalysts [6, 7]. Unfortunately, the temperatures used in these studies are usually far
lower than those necessary for thermal reduction in inert atmospheres. Therefore, the
applicability of these data must be regarded as limited. Some case studies of ceria-based
systems in solarthermal cycles are published in the literature [2, 4, 14]:
2 CeO2 Ce2O3 +
1
O2
2
(13)
(14)
Thermodynamic calculations showed that the reduction of CeO2 to Ce2O3 (Equation [13]) takes place spontaneously at temperatures above 2,623 K, which is in
between the melting points of the reactant (2,873 K) and the product (2,503 K). Nevertheless, partial reduction of ceria to suboxides can be performed at lower temperatures [4]. Typical drawbacks of ceria for the solarthermal cycling are its tendency to
sinter and its surface-limited oxygen storage/release capacity. Modication with dopants such as Zr helps to overcome these drawbacks [4]. Experimentally, solarthermal cycles with Zr-doped ceria (CeO2)0.25(ZrO2)0.75 were successfully performed
and the rate of the CDS was limited by surface processes. Higher ceria contents
lead to increased sintering of the materials and therefore to a loss in activity during
the reduction step (Equation [13]). For pure ceria, thermodynamic analysis results in
solar-to-fuel energy conversion efciencies of maximum 16% to 19% without heat
recovery [2]. However, only efciencies as low as 0.8% were experimentally achieved
in the same study. As previously mentioned, the main energy losses were due to heat
conduction and reradiation.
5.2.6 References
411
(15)
2 CO + 2 H2 CO2 + CH4
(16)
The dry reforming reaction (Equation [15]) with HR = +247.2 kJ mol1 is associated with a high energy demand. The reaction becomes thermodynamically favorable
at temperatures higher than 700C, and full conversion is obtained at temperatures
around 900C. In addition to the formation of synthesis gas for Fischer-Tropsch or
methanol synthesis the process offers another alternative in terms of energy storage,
because its application can be combined with long-distance heat pipelines [15]. The
general principle is solar-to-chemical energy conversion and transport of the cooled
product stream to a consumer, where the conversion to CH4 and CO2 (Equation [16])
releases the stored heat. In this way, the thermochemical cycle is closed. Because the
back reaction has to be carried out at signicantly lower temperatures to achieve
high degrees of conversion, the temperature level at which the energy is nally obtained
and, correspondingly, the efciency of downstream processes are slightly reduced.
5.2.5 Summary
Two-step thermochemical cycles based on metal oxides and concentrated solar power
are a highly interesting pathway for the activation of CO2. Their feasibility has been
demonstrated for ZnO and for Fe3O4 and CeO2, which can be further optimized by
adding structural promoters. The long-term stability of the metal oxide and the construction materials under the harsh conditions remains a substantial challenge for
materials science. The efciency and the CO productivity can be further increased
by reactor optimization and heat recuperation between the reduction of the metal
oxide at very high temperatures and its reoxidation by CO2 at low temperatures.
Correspondingly, further studies are needed focusing on the thermodynamic, kinetic,
and reaction engineering aspects of the two-step thermochemical cycles.
5.2.6 References
1. Levenspiel O. Chemical reaction engineering. 3rd ed. New York: John Wiley & Sons;
1998.
2. Chueh WC, Falter C, Abbott M, et al. Science. 2010;330:1797.
412
The authors thank the members of the Inorganic Chemistry Department at the Fritz
Haber Institute Berlin for their contributions to this chapter. Sd-Chemie AG Bruckmhl (now Clariant) and Martin Muher and his group at Ruhr-University Bochum are
acknowledged for a long-standing and fruitful collaboration
414
Methanol-containing fuels (MW50, 50% methanol and 50% water) have also been
used as a synthetic fuel for temporarily increased performance of aircrafts during
the Second World War.
Methanol is particularly interesting because it can be produced by hydrogenation
of the greenhouse gas CO2 (see Section 5.3.5). Thus, anthropogenic CO2 (e.g. from
industrial exhaust or coal power plants) could be used for its manufacture. The
amount of CO2 emitted upon methanol combustion is then equal to the amount consumed during its production. Such CO2 recycling via methanol has been proposed by
Olah et al. [2] as a carbon-neutral methanol economy.
Another promising aspect of the use of methanol in energy applications is that its
synthesis is already a large-scale industrial process (see Section 5.3.3). Thus, regarding a further upscaling of methanol production, which is necessary for energy-related
application, mature technology and long-lasting experience already exist. This is an
important advantage compared to completely new approaches that have to be developed from scratch like renewable hydrogen or biofuels. Today, methanol is an
important platform molecule in the chemical industry. It is further upgraded into
other chemical intermediates like formaldehyde or acetic acid. Zeolite-catalyzed
methanol-to-olens and methanol-to-gasoline processes, developed by Mobil, open
a pathway for direct conversion of methanol into fuels and chemicals currently derived exclusively from petroleum. However, these are only expected to become economically viable in the case of high petroleum prices. Among the direct products of
methanol conversion, methyl-tert-butyl-ether (MTBE) and dimethyl ether (DME)
are of relevance for energy applications. MTBE is used as an octane booster to
improve the antiknocking properties of gasoline but recently has been banned by
some states due to environmental concerns. DME has been proposed as a potential
substitute for diesel fuel.
Methanol production today is not a sustainable process but is part of a petrochemical route for conversion of fossil carbon into chemicals and fuels (see Section 5.3.3).
It has to be emphasized that a one-to-one upscaling of existing industrial methanol
synthesis capacities for fuel production is not useful. This is mainly because the current industrial process has not been developed and optimized under the boundary
conditions of conversion of anthropogenic CO2, but rather for synthesis gas feeds derived from fossil sources such as natural gas or coal. The switch to an efcient largescale methanol synthesis with a neutral CO2 footprint is still a major scientic and
engineering challenge, and further research and catalyst and process optimization
is urgently needed to realize the idea of a sustainable methanol economy.
Obviously, the other reactant for methanol production by CO2 hydrogenation is
hydrogen. Thus, all considerations of the use of methanol as a sustainable fuel are
only useful if a regenerative source of hydrogen is available. The electrolysis of water
(see Chapter 3.2) through renewably derived electricity such as hydropower or wind
power, or alternatively gasication of biomass (see Chapter 2.3), may serve as such a
source [2]. Hence, the primary chemical energy storage has to happen in the form of
hydrogen production. The further conversion of hydrogen into methanol is associated
with a gain in volumetric and a loss in gravimetric energy density, but also with much
easier handling of the energy carrier [6]. This implies not only distribution and storage
without pressurized or cryogenic containers, but also safety issues. It should not be
415
concealed that methanol itself is toxic and ammable, with risks and safety measures for
everyday use that are comparable to those of gasoline [2].
This chapter focuses on the catalytic aspects of methanol chemistry and covers
thermodynamic, kinetic, chemical engineering, and materials science aspects. It provides brief introductions into these topics with the aim of establishing an overview of
the state of the art of methanol chemistry with only a snapshot of the relevant literature. It highlights what the authors think are the most relevant aspects and future
challenges for energy-related catalytic reactions of methanol. It is not meant to provide
a complete literature overview on methanol synthesis and reforming.
(1)
(2)
(3)
Methanol synthesis from CO2 (Equation [1]) and CO (Equation [2]) is mildly
exothermic and results in volumetric contraction. Methanol steam reforming
(MSR) refers to the inverse of reaction (1), and the inverse of reaction (2) is conventionally referred to as methanol decomposition an undesired side reaction to MSR.
The slightly endothermic reverse water-gas shift (rWGS) reaction (Equation [3]) occurs as a side reaction to methanol synthesis and MSR. According to Le Chateliers
principle, high pressures and low temperatures would favor methanol synthesis,
whereas the opposite set of conditions would favor MSR and methanol decomposition. It should be noted that any two of the three reactions are linearly independent
and therefore sufcient in describing the compositions of equilibrated mixtures.
A precise quantitative description of the methanol synthesis equilibrium must
account for the nonideal behaviors of the gas species, in particular those of water
and methanol. Skrzypek et al. show that the Soave-Redlich-Kwong equation gives
good agreement with experiment [7]. Although nonidealities lead to higher methanol
yields compared to those expected from ideal gas behavior [8], noncorrected equilibrium calculations sufce in qualitatively illustrating the equilibrium behavior as a
function of conditions. Figure 5.3.1A and B shows the ideal gas methanol and CO
yields, respectively, for a 3:1 H2:CO2 feed mixture. The methanol yield displays a
positive dependence on pressure and an inverse dependence on temperature, whereas
the rWGS yield shows weak pressure dependence and increase with increasing
416
A 80%
B
400 K
450 K
500 K
550 K
600 K
70%
50%
25%
20%
CO Yield
Methanol Yield
60%
80%
40%
30%
400 K
450 K
500 K
550 K
600 K
15%
10%
20%
5%
10%
0%
0%
0
10
P(MPa)
10
P(MPa)
Figure 5.3.1 Equilibrium methanol and CO yields from 3:1 H2:CO2 mixture.
O
H
H3C
H
CH3
O
H3C
OH H3C
1109
1107
CO
O
OCH3CH3OH H3C
1105
CH4
C2H6
C2H5OH
C(s)
CH3
1103
1101
1101
1103
1105
1107
Keq/CO2 reacted
Figure 5.3.2
temperature. The addition of CO to the feed mixture has a positive effect on the
equilibrium yield of methanol, as will be discussed in detail in Section 5.3.5.
The current low-pressure synthesis process (see Section 5.3.3) yields methanol
with greater than 99% selectivity on a CO-free basis. The equilibrium constants
(per CO2 molecule reacted) for the formation of common by-products at 250C
are shown in Figure 5.3.2. The high methanol selectivity is fairly astonishing, considering that the formation of ethers, ketones, and alkane impurities found in industrial
methanol is more thermodynamically favored than is the formation of methanol.
Similarly, coke another thermodynamically favored product and common cause
of catalyst deactivation is never observed [9]. This observation unambiguously
indicates that the aforementioned products are kinetically inaccessible on the used
Cu/ZnO-based catalyst that will be described in more detail in Section 5.3.5. From
these thermodynamic considerations, the need for a more active CO2 hydrogenation
catalyst becomes apparent. Even at temperatures associated with the low-pressure
process, K1 lies between 105 and 106, allowing for a single-pass methanol yield
of 15%25% and thus necessitating the implementation of costly recycling loops.
MSR is also carried out on methanol synthesis catalysts at similar temperatures
(see Section 5.3.7), but unlike methanol synthesis, it is not subject to thermodynamic constraints. Thermodynamic considerations play a lesser role in MSR, as the
inverse of reactions (1) and (2) can be considered irreversible at atmospheric pressure.
417
However, lower-temperature operation would thermodynamically hinder CO formation via methanol decomposition and rWGS. A low CO content is desired for
MSR-PEMFC combinations (see Section 5.3.7). To that end, the development
of catalysts active at lower temperatures still remains the central goal of methanol
catalysis research.
418
(4)
Gas mixtures with a modulus value, M (Equation [5]), of around 2 satisfy the
stoichiometric requirements:
M=
[H2 ][CO2 ]
[CO] + [CO2 ]
(5)
Countries with large domestic coal reserves, such as China and South Africa, rely
primarily on coal gasication to produce synthesis gas. This synthesis gas is hydrogen
decient (M < 2) and must undergo a further water-gas shift (WGS) step to yield
a CO2-rich mixture [2]. Methanol synthesis from CO2 and CO2-rich mixtures provides special catalyst and reactor design challenges, which will be further discussed
in more detail.
Industrial methanol synthesis is carried out in xed-bed ow reactors, which are
designed to achieve effective removal or dissipation of the heat generated during
this exothermic reaction. The original adiabatic quench reactor developed by ICI
consists of a single catalyst bed, with cold syngas injected at several points along
the axial direction of the bed. The Kellogg and Haldor-Topse reactor designs consist of a series of catalyst beds with interstage cooling of the products. A quasiisothermal reactor developed by LURGI consists of several tubular catalyst beds
surrounded by an outer shell of boiling water [8]. The pressure in the shell is used
to control the reaction temperature. The ICI and LURGI congurations account
for approximately 60% and 30% of global methanol production, respectively.
Crude methanol leaving the reactor contains volatile impurities, such dissolved
gases, light hydrocarbons, esters, and ketones (Figure 5.3.2), which are removed in
an initial stripping step. Less volatile impurities such as water and heavier alkanes
(C8+) are removed as bottoms in subsequent distillation steps [2, 8, 15]. A simplied
schematic of a methanol synthesis plant is shown in Figure 5.3.3.
Low-pressure methanol synthesis relies almost exclusively on catalysts based on
copper, zinc oxide, and alumina. The catalysts are produced by ICI (now Johnson
Matthay), Sdchemie (now Clariant), Haldor Topse, in the past also by BASF,
and other chemical enterprises and contain 5070 atomic % CuO, 20%50% ZnO,
and 5%20% Al2O3. Instead of alumina, chromium oxide and rare earth oxides
have also been used. The mixed oxide catalysts are usually shipped as 46 mm cylindrical pellets with specic surface area of 60100 m2/g. The catalysts are activated
in situ with dilute hydrogen, often derived from off-gases from synthesis gas
Methanol
Light ends
Purge gas
Flue gas
419
Recycle gas
A
C
CH4
Synthesis
Gas
Steam
Water and
heavy products
Figure 5.3.3 A simplied schematic of a methanol synthesis plant: (A) gasier, (B) compressor,
(C) methanol reactor, (D) ash drum, (E) light-ends column, and (F) methanol column.
production. The activation procedure takes place at 190C230C, completely reducing copper oxide to metallic crystallites interspersed by a ZnO-Al2O3 matrix. More
details on the preparation and properties of Cu/ZnO-based methanol synthesis
catalysts will be given in Section 5.3.6.
Typical Cu/ZnO/Al2O3 catalyst lifetimes are several years, with one-third of the
total activity loss occurring during the rst 1,000 hours of operation. The loss of
activity with time is compensated by increasing the reaction temperature. Deactivation occurs through loss of copper dispersion during particle growth or poisoning by
impurities. High partial pressures of water, associated with CO2-rich gas mixtures,
have also been shown to accelerate particle growth. However, the complete removal
of CO2 leads to an even faster deactivation by dispersion loss [17]. Sulfur is a potent
poison for Cu catalysts; however, sulfur poisoning is seldom a problem as syngas
feeds are desulfurized to less than 0.5 ppm. The ZnO catalyst component provides
some protection against sulfur poisoning by scavenging sulfur irreversibly as ZnS,
thereby preserving a large fraction of catalyst activity even at sulfur loading of
several percent [8, 17].
420
predicted a zero rate of methanol production in the absence of CO2, and it was assumed that CO2 and water prevented overreduction of Cu and thus helped maintain
a population of active Cu+ species [18, 19]. In the 1980s, experiments conducted by
Razovskii and later by Chinchen et al. involving the use of 14CO or 14CO2 tracers
in methanol synthesis from CO2/CO/H2 mixtures over commercial catalysts proved
conclusively that CO2 was the primary methanol source [20, 21]. Chinchen et al. measured the radioactivity of reaction products at the outlet of a reactor operating with a
14
CO2/12CO feed. As shown in Figure 5.3.4, at lower space velocities, scrambling of
carbon isotopes between CO and CO2 through WGS resulted in the incorporation of
both isotopes into methanol allowing no identication of the carbon sources at these
conditions. However, at high space velocities, where the rate of scrambling is negligible and conversion is low, methanol retained the specic radioactivity of the 14CO2,
indicating that only carbon from CO2 was incorporated. Chinchen et al. showed that
even when present at very low concentrations (100 ppm), CO2 was the primary carbon
source for methanol.
In situ spectroscopic studies have identied a variety of species, such as formate,
dioxymethylene, carbonate, and methoxide, to coexist under methanol synthesis conditions on Cu/ZnO-based catalysts [22, 23]. Fourier transform infrared spectroscopy
studies of CuZn-based catalysts under H2/CO2 identied the presence of formate
bound to both Cu and ZnO, whereas methoxide was found on ZnO only. Carbonates
were found to form via CO2 adsorption on ZnO [24] and partially oxidized Cu [23],
and were quickly converted into formate via Cu-activated hydrogen. Upon exposure
to CO mixtures, only zinc-bound formate was observed [22]. The hydrogenation of
these formates to methoxide is thought to be rate determining in methanol synthesis.
0.14
Inlet CO2
Specific Radioactivity C/m mol
0.12
0.10
CH3OH (Exit)
CO2 (Exit)
0.08
0.06
0.04
CO (Exit)
0.02
Inlet CO
5 104
Space Velocity (h1)
10104
Figure 5.3.4 Effect of space velocity on 14C distribution in methanol synthesis products, from [20].
421
However, this hydrogenation may not be direct, as even in the presence of hydrogen the rate of methanol synthesis on Cu/SiO2 from formate was negligible compared to the rates of formate decomposition into CO2 and H2 [25]. The presence
of water and/or hydroxyl groups was found to be critical for methanol formation
[26]. However, the absence of direct coverage by reducible oxygen species (and therefore of Cu+) was conrmed by CO pulse experiments on catalysts in the working
state [27].
Surface science studies on Cu single crystals also identify formate an abundant surface species but report a wide range of activities for methanol synthesis. The intrinsic
reaction rates reported by Szanyi and Goodman for the hydrogenation of CO/CO2
mixtures on Cu(100) were four orders of magnitude below those reported for highsurface-area Cu/ZnO-based catalysts [2829]. Rasmussen et al. report rates two to
three orders of magnitude higher than those reported by Zanyi and Goodman for
Cu(100), but using CO2/H2 mixtures. Yoshihara et al. report rates comparable to
those on high-surface-area catalysts on polycrystalline copper [30], and even rates
three times higher on the more open Cu(110) surface [31]. Yoshihara and Rasmussen
both conrm the absence of oxygen on the Cu surface and assume metallic Cu to be
the active phase [31, 32]. Furthermore, ZnO is assumed to maintain metallic Cu in a
form that more closely resembles an open surface, thereby acting as a promoter [30].
Direct deposition of up to 0.19 ML of Zn on a polycrystalline Cu surface has been
shown increase CO2 hydrogenation activity by a factor of six, thus suggesting a more
direct role for ZnO [33].
The structure sensitivity of the rWGS reaction that accompanies methanol synthesis was even more pronounced, with Cu(110) being an order of magnitude more
active than polycrystalline Cu. Furthermore, kinetic studies of CO2 hydrogenation
on Cu/ZnO catalysts yielded a temperature-dependent nonzero CO selectivity at
the limit of zero CO2 conversion [22, 31, 34]. These ndings suggest that both reactions do not share a common intermediate. Whereas methanol formation is associated with formate hydrogenation, the rate of the rWGS reaction is controlled by
direct dissociation of CO2, a direct redox process that is more favorable on open
surfaces [31].
Theoretical studies conrm the importance of formate as a surface intermediate
and its sequential hydrogenation as being rate determining. A highly simplied version of the mechanism proposed by Askgaard et al. for methanol synthesis from CO2
on Cu(111) is given in equations 611 [35], where steps described by equations 811
may each represent a series of reaction events. The formation of formate from CO2
with hydrogen is a fast process, and CO2 is thought to bind directly to adsorbed
hydrogen in an Eley-Rideal step [36], instead of forming carbonate by binding to surface oxygen species [37]. Asgaard et al. propose that the hydrogenation of H2COO*
is rate limiting, whereas Yang et al., Hu et al., and Grabow and Mavrikakis identify
the hydrogenation of HCOO* and CH3O* respectively as rate determining [3638].
The presence of adsorbed Zn may promote methanol synthesis on Cu(111) by stabilizing formate and associated transition states [39]. Zhao et al. show that methanol
cannot be formed directly though formate hydrogenation on Cu(111) but is instead
formed by a hydrogen-transfer reaction requiring the presence of water [40]. According to Grabow and Mavrikakis, the main role of CO in promoting the hydrogenation
422
of CO2 is the removal of site-blocking OH species via rWGS, although the direct
hydrogenation of CO is said to account for one-third of the methanol produced
under industrial conditions [37].
H2 + 2* A 2 H*
(6)
H2 O + 2* A H* + OH*
(7)
CO2 + 2* A CO* + O*
(8)
CO2 + H* A HCOO*
(9)
(10)
(11)
423
100%
CO2
90%
10:1
CO2:CO
ratio
80%
5:1
2:1
Methanol Yield
70%
1:1
60%
1:2
50%
1:10
CO
40%
30%
20%
10%
0%
400
450
500
T(K)
550
600
Figure 5.3.5 Equilibrium yield of methanol at 50 bar from 3:1 CO:CO2/H2 mixtures as a
function of CO2/CO ratio.
Although CO2 has been shown to undergo hydrogenation faster than CO, kinetic
limitations would arise in a process operating at high conversions, as in industrial applications. Several investigators have shown that a maximum in methanol production
with CO2/CO ratio occurs at CO2 concentrations of 25 mol% of total carbon
[19, 34, 42]. Klier et al. ascribed this behavior to active Cu+ sites being created by oxidation at low CO2 concentrations, and surface poisoning by strongly bound CO2 at
higher concentrations, thus resulting in an activity maximum [19]. Further investigations have shown that such a maximum in the synthesis rate is only exhibited at high
reactant conversions, thus indicating that product inhibition plays an important role
[34, 4244]. Shibzada et al. have studied the effect of methanol synthesis rate versus
CO2/CO ratio in an integral internal recycle reactor compared to a differentially operated down-ow microreactor. Their results (Figure 5.3.6) show that a maximum
exists at 2% CO2 in the integral reactor, while in the differential reactor (operated
at < 0.3% methanol yield) an almost linear relationship between the methanol synthesis rate and CO2 concentration was observed. These investigators have correlated
the methanol synthesis rate at integral conditions with the concentration of water
in the product stream during CO2 hydrogenation. Furthermore, they have shown
that cofeeding of water to a differential reactor in concentrations similar to those
formed during integral operation results in a 10-fold decrease of the methanol synthesis rate [34]. These results show conclusively that water acts to inhibit methanol synthesis from CO2, and that the promotional effects of CO at integral conversion arises
from its ability to scavenge excess water via WGS [23, 45].
Indirect evidence for the inhibitory role of water in methanol synthesis can be
drawn from a simple experiment involving the variation of space velocities. In
Figure 5.3.7a, the methanol synthesis and rWGS rates as a function of space velocity
424
15
10
20
0.25
0.2
0.15
differential
conversion
0.1
0.05
finite conversion
0
0
0.2
0.4
0.6
0.8
Figure 5.3.6 Methanol synthesis at differential and integral conditions at 250C and 50 bar,
from [34].
for two Cu/ZnO-based catalysts are shown. The more active catalyst is derived from
a zincian malachite (MA) precursor [46] (see Section 5.3.6) and promoted with
Al2O3. It possesses a Cu surface area of 30 m2/g, while the less active catalyst is derived from a hydrotalcite (HT) precursor [47] and possesses a smaller Cu surface area
(7 m2/g). In both catalysts, the rate of rWGS shows little variation with increasing
space velocity, whereas the rate of methanol synthesis increases dramatically. The
concentration of the reactants does not vary signicantly throughout this space
velocity range as the maximum CO2 conversion is not higher than 14% for the
MA catalyst. However, the concentrations of water and methanol decrease by a factor of three when increasing the space velocity from 10 to 80 mmol/gcatmin, as a
result of product dilution. In the absence of external mass transport limitations,
these results imply product inhibition of methanol synthesis. The rWGS reaction
does not seem to be product inhibited. Furthermore the rWGS rate seems to remain
constant even as the concentration of methanol decreases at higher space velocities,
also due to product dilution. The lack of correlation between methanol partial pressure and CO production rate implies that CO does not form from methanol.
Although methanol decomposition (K = 200 at 230C) is more thermodynamically
favorable than rWGS (K = 0.01 at 230C) at these conditions, the former has
been shown to be kinetically unfavorable on Cu [48].
When comparing the intrinsic methanol synthesis activity of both catalysts (turnover
frequency [TOF] per surface Cu site measured with N2O-reactive frontal chromatography) at the space velocity of 20 mmol/gcatmin, the HT catalyst (0.54 min1) appears
twice as active as the MA catalyst (0.24 min1). Although both catalysts are exposed
to nearly the same reactant concentration, the MA catalyst operates at 11% CO2
Rate (mol/min-gcat)
Same S.V. HT
MA MeOH
600
HT CO
500
Same S.V. MA
HT MeOH
425
MA CO
400
300
200
100
0
0.55
Same Conv. MA
0.50
Same Conv. HT
0.45
Promoted-HT
0.40
Ga Al Promoted-MA
Cr
0.35
0.30
MA-bin
0.25
0.20
20
40
60
S.V. (mmol/gcat-min)
80
100
10
20
30
40
50
Figure 5.3.7 The variation of CO and methanol production rates with space velocity (A) and
methanol synthesis rates on promoted CuZn-based catalysts derived from hydrotalcite and
Malachite precursors (B). Reaction conditions: 200 mg catalyst, 30 bar, 230C, 3:1 H2:CO2.
conversion and is exposed to three and a half times as much reaction-inhibiting water
as the HT catalyst (X = 3.3%). If both catalysts are tested at the same CO2 conversion
(5%6%), they show a similar intrinsic activity (TOF ca. 0.35 min1). These results suggest that the interpretation of catalytic activity data for the hydrogenation of CO2 is a
strong function of the measurement conditions, and therefore great caution must be
taken when comparing and interpreting reactivity data for different materials. It is
noted that for methanol synthesis from CO-containing feed gases, these considerations
are less important, as water is scavenged by CO.
The selectivity to methanol is another important consideration in CO2 hydrogenation. Methanol synthesis and rWGS are independent reaction channels on Cu-based
catalysts (see Sections 5.3.2, 5.3.4). On Cu-based catalysts, the apparent activation
energies are reported around 120 kJ/mol and 70 kJ/mol for rWGS and methanol synthesis, respectively [25, 49, 50]. As a result of this difference, selectivity toward methanol decreases with increasing temperature. The selectivity to methanol also increases
with total reaction pressure, as the total pressure increases the rate of methanol synthesis, but the rate of rWGS is weakly affected [51]. How the properties of Cu/ZnObased catalysts can be tailored in order to favor methanol synthesis over rWGS is a
key question for catalyst development for CO2 hydrogenation. Recently Liao et al.
showed that methanol selectivity can be controlled by the shape of ZnO crystallites
in catalysts consisting of physical mixtures of Cu and ZnO crystallites [52]. They
suggest that a stronger Cu-ZnO interaction observed for plate-like crystallites showing
the polar (001) face as opposed to nonpolar rod-like crystallites is responsible for
higher methanol selectivity. These results suggest that tuning the so-called synergy between Cu and ZnO (see Section 5.3.6) is a promising approach to improve the selectivity of Cu toward methanol synthesis. The undesired rWGS is known to be more
structure sensitive than methanol synthesis on copper. Therefore, another selectivity
control strategy may involve controlling Cu particle size and shape [28]. Another approach is the use of promoters. Arena et al. have suggested that adding promoters
such as ZrO2 that increase the hydrophobicity of the catalyst surface would lead to
426
better activity in CO2 hydrogenation [49, 53] due to less product inhibition by water.
Oxides such as Al2O3, Cr2O3, Ga2O3, or ZrO2 are known to promote methanol synthesis [54]. They may improve Cu dispersion and stability [9], but also the intrinsic catalytic properties of the exposed Cu surface [55]. The latter is probably related to an
adjustment of the properties of the ZnO crystallites and related to modied interactions between Cu and ZnO. A clear picture of the nature of the promoting effect of different oxides and their inuence on the selectivity in CO2 hydrogenation is still lacking.
When studying the effect of promoters on the methanol synthesis activity of differently promoted catalysts, again great caution is needed for a reliable comparison of
performance data. An example is shown in Figure 5.3.7B. The intrinsic methanol synthesis activities for a series of MA-derived catalysts and a series of HT-derived catalysts with difference promoters are plotted as a function of Cu surface area and
compared to an unpromoted binary MA-derived catalyst (MA-bin). The activity
was measured at the either the same space velocity or the same CO2 conversion.
When considering activities measured at the same space velocity, all HT-based catalysts appear intrinsically more active than the MA-derived ones. All promoted MAbased catalysts show the same intrinsic activity as MA-bin under these conditions.
One might thus presume that the promoting effect is limited to improvement of the
Cu dispersion within this series of samples. However, isoconversion activities of
both promoted series fall in the same range (0.30.35 min1) and are signicantly
greater than that of the unpromoted sample (0.25 min1), showing clearly that the presence of the promoters also improves the intrinsic activities of the MA-derived catalysts.
Cu
Cu
Cu
ZnO
ZnO
ZnO
Cu
427
ZnO
Cu
ZnO
Cu
5 nm
ZnO
Cu
50 nm
100
110
Twin boundaries
111
Disordered
ZnOx overlayer
2 nm
Figure 5.3.8
amounts to 2535 m2gcat1. If reliable data of the average Cu particle size are available (e.g. by sufcient TEM observations [57]), the degree of oxide coverage of the
Cu particles (i.e. the average ratio of interface area to surface area) can be calculated
[61]. For industrial Cu/ZnO/Al2O3 catalysts, this value is around 35%. The favorable
microstructure and the proper balance of Cu dispersion and loading in this type of
Cu/ZnO/(Al2O3) catalyst leads to a large SACu, which is probably the most important
property of a methanol synthesis catalyst.
428
The SACu has been observed to scale linearly with the activity for sample families
with a similar preparation history [55]. However, between these families considerably
different intrinsic activities (i.e. activities normalized by SACu) can be found [61]. Thus,
in agreement with the structure sensitivity of methanol synthesis over Cu [31], different
qualities of Cu surfaces can be prepared, which vary in the activity of their active
sites and/or in the concentration of these sites. Differences in intrinsic activity of the
exposed SACu can be related to defects and disorder in the Cu nanoparticles and to
the role of ZnO. Clearly, one role of ZnO that is apparent from Figure 5.3.8B is to
act as spacer and stabilizer, avoiding direct contact of the Cu particles and preventing
them from sintering [62]. But it is generally agreed that the role of ZnO in Cu-based
methanol synthesis catalysts exceeds the function of a mere physical stabilizer. In addition to this geometric function, a so-called Cu-ZnO synergy is described in literature
for methanol synthesis [8, 63, 64]. The nature of this synergy and the contribution
of ZnO to the active site of methanol synthesis are strongly debated, and several models have been proposed for example, Cu+ ions in the ZnO matrix [65], ZnO segregated on Cu+ [66], electron-rich Cu at the Cu-ZnO heterojunction [66], CuZn
surface alloy formation [68], or Cu metal supported on ZnO [23]. Strong metaloxide-interactions (SMSI) between Cu and ZnO were observed at highly reducing conditions [69, 70], and it was suggested that partially reduced ZnOx migrates onto the
surface of the Cu particles under methanol synthesis conditions [71]. On a supported
Cu/ZnO model catalyst, reversible wetting/dewetting was observed as the reduction
potential of the gas phase was varied [72], an observation not made on Cu/SiO2.
Another contribution to variations of intrinsic activity is the different number of
defects and amount of disorder in the metallic Cu phase. This disorder can manifest
itself in the form of lattice strain detectable, for example, by line prole analysis of
X-ray diffraction (XRD) peaks [73], 63Cunuclear magnetic resonance lines [74],
or as an increased disorder parameter (Debye-Waller factor) derived from extended
X-ray absorption ne structure spectroscopy [75]. Strained copper has been shown
theoretically [76] and experimentally [77] to have different adsorptive properties compared to unstrained surfaces. Strain (i.e. local variation in the lattice parameter) is
known to shift the center of the d-band and alter the interactions of metal surface
and absorbate [78]. The origin of strain and defects in Cu/ZnO is probably related
to the crystallization of kinetically trapped nonideal Cu in close interfacial contact
to the oxide during catalyst activation at mild conditions. A correlation of the concentration of planar defects in the Cu particles with the catalytic activity in methanol
synthesis was observed in a series of industrial Cu/ZnO/Al2O3 catalysts by Kasatkin
et al. [57]. Planar defects like stacking faults and twin boundaries can also be
observed by HRTEM and are marked with arrows in Figure 5.3.8C [58].
Recently, experimental and theoretical evidence for a model of the active site of
industrial methanol synthesis that combines the role of ZnO and defects in Cu has
been presented [58]. Planar defects have been shown to lead to changes in surface faceting of the Cu nanoparticles (Figure 5.3.8C) associated with formation of steps and
kinks that were assumed to represent high-energy surface sites of special catalytic
activity. For a series of Cu/ZnO-based catalysts, a linear correlation of the defect
concentration with the intrinsic activity of the exposed Cu surface was observed.
In addition, (partial) surface decoration of Cu with ZnOx by SMSI has been
429
A
Calcination
Ageing
Meso-structuring
Reduction
Nano-structuring
Activation
5 nm
Cu
Cio
Cio
Cu
Cu
ZnO
ZnO
Zio
ZnO
Cu
ZnO
Cu
Zio
Cu
140
120
ercial
% comm ctivity
produ
catalyst
Thin needles
size ca. 20 200 nm
Cu
5 nm
ZnO
ZnO
10 0
80
60
40
20
70
(p 60
re
ci
pi 50
ta
tio 40
n)
3
[
C] 0
10
pH
Figure 5.3.9 (A) Simplied geometric model [46, 89] for the preparation of industrial Cu/ZnO
catalysts comprising subsequent meso- and nanostructuring of the material from [56]. In a rst
microstructure directing step (mesostructuring), the Cu,Zn coprecipitate crystallizes in the
form of thin needles of the zincian malachite precursor, (Cu,Zn)2(OH)CO3. In a second
step, the individual needles are decomposed and demix into CuO and ZnO. The effectiveness
of this nanostructuring step depends critically on a high Zn content in the precursor, which in
zincian malachite is limited to Cu:Zn ca. 70:30 due to solid-state chemical constraints [75].
Finally, interdispersed CuO/ZnO is reduced to yield active Cu/ZnO. (B) Chemical memory:
Dependence of catalytic activity in methanol synthesis on the conditions of the coprecipitation
and aging steps, from [85].
430
5.3.7 MSR
431
of 200300 C, and unlike methanol decomposition does not directly produce CO,
which acts as a poison for the downstream PEMFC anode catalyst. MSR was rst described in 1921 by Christiansen [90], and research on its application for hydrogen production has a long history [91]. The renewed interest in the late 20th century was
triggered by the development of fuel-cell technology requiring clean and preferably
renewable hydrogen. A number of general overview articles and reviews are available
addressing the role of MSR in this context [12, 9196]. In areas where steam reforming
of natural gas is not an option, MSR is also applied in the methanol-to-hydrogen
process to produce hydrogen in relatively small-sized units.
MSR is an endothermic reaction (see Section 5.3.2) and requires external heating.
It is sometimes used in combination with exothermic partial oxidation of methanol
(autothermal reforming or oxidative steam reforming) [97, 98] or combustion of
methanol [94] in order to generate the necessary heat. The endothermicity of MSR
is much weaker compared to steam reforming of other hydrocarbons or higher alcohols [94], and reformer units can be relatively small, enabling the onboard combination with PEMFCs. A comparison between methanol and other molecules as
reactants for onboard hydrogen production can be found in the comprehensive
review by Palo et al. [12]. The detrimental role of CO in the efuent for the downstream PEMFC is to be emphasized. CO chemisorbs irreversibly on Pt-based fuelcell catalysts and causes irreversible site blocking. Its concentration has to be
below ca. 20 ppm in order to prevent poisoning, which is usually not achieved in
the reformer outlet gas. Thus, in technical applications, a gas purication step has
to be introduced between reformer and fuel cell. The CO concentration in the gas
stream can be lowered by means of the WGS reaction, preferential oxidation, or
using Pd membranes, which in all cases complicates the setup and generates costs
[99]. Generally, a low selectivity to CO in addition to high activity and stability
is thus a major and particular requirement for a successful MSR catalyst to be used
for onboard hydrogen production. Methanol-reforming catalysts should be particularly stable toward abrupt changes of the conditions of reforming that is, work reliably in transient situations like on-off operations as well as in steady state to produce
sufcient amounts of hydrogen on demand.
Commercial industrial Cu/ZnO-based catalysts for methanol synthesis (see Section
5.3.6) or WGS are also active in MSR. Cu/ZnO/Al2O3 catalysts or the unpromoted
binary Cu/ZnO model system were thus employed in many studies of MSR [56].
While preparation and composition of the industrial Cu/ZnO/Al2O3 catalysts have
been adjusted for application of methanol synthesis and WGS, modications of the
Cu/ZnO/X system turned out to improve the properties for use in MSR. In particular,
choosing another second oxide phase X like rare earths [100] or ZrO2 [101105], or
employing new catalyst precursors like layered double hydroxides [106108], or even
changing to ZnO-free samples and using ZrO2 and/or CeO2 [110114] for preparation
of Cu-based catalysts was reported to lead to interesting MSR performance.
Several studies are available addressing the mechanism and kinetics of the MSR
reaction over Cu-based catalysts [115121]. There is agreement nowadays that
CO2 is a direct product of the MSR reaction and not of a sequence of methanol
decomposition and WGS reactions. The main source of CO is the rWGS reaction
taking place as a secondary reaction after MSR. Frank et al. [121] presented a
432
comprehensive microkinetic analysis of the MSR reaction based on the work of Peppley et al. [117]. They investigated several Cu-based catalysts with various oxide components showing considerably different activities. Similar activation energies support the
idea that the surface chemistry is independent of the oxide materials used in their
study (with the exception of Cu/Cr2O3/Fe2O3, which behaved differently). Dehydrogenation of methoxy groups is the rate-limiting step, and by means of diffuse reectance
infrared fourier transform spectroscopy (DRIFTS) experiments, methoxy and formate
species were found as the dominating species at the surface. Two distinct kinds of active
sites were considered, one for the adsorption and desorption of oxygenates and one
for hydrogen. Two reaction pathways of the methoxyl intermediate are discussed via
dioxomethylene/formate, intermediates of the reverse methanol synthesis reaction,
or methyl formate. Recent theoretical studies favor the former pathway [122, 123].
There are many similarities between methanol synthesis and MSR [118]. This is
often accounted for by the concept of microscopic reversibility, as MSR formally is
the reverse reaction of methanol synthesis from CO2 (see Section 5.3.2). It has to
be considered, however, that the different reactant gas mixtures used for MSR and
methanol synthesis will affect the surface state of the catalyst, which consequently
will be different under highly reducing methanol synthesis conditions compared to
the less reducing MSR feed. Thus, unlike forward and reverse reactions at equilibrium, methanol synthesis and MSR probably may take place over practically different catalytic surfaces. This general limit of the application of the concept of
microscopic reversibility has been pointed out by Spencer for WGS and rWGS
[124] and, accordingly, is also valid for methanol synthesis and MSR [9]. One may
conclude that an optimized methanol synthesis catalyst, for which the ne-tuning
of preparation and operation conditions is far more advanced, will also be active
in MSR due to its generally large SACu and represents a powerful reference system,
but it does not necessarily represent the optimal catalyst for this reaction [9]. Finding
Cu/ZnO/X systems with a composition and microstructure optimized for the MSR
reaction is thus the major current challenge in development of an MSR catalyst
for energy applications. In this context, it is interesting to note that the Cu-ZnO synergy (see Section 5.3.6) does not seem to be as critical a factor in the case of MSR
compared to methanol synthesis [125], or this synergistic effect is not as strictly limited to ZnO. Highly active Cu-based MSR catalysts can also be prepared in the
absence of ZnO (e.g. as Cu/ZrO2).
The question as to what is the active site of Cu-based catalysts in MSR is still unclear and debated in the literature. Similar to the methanol synthesis reaction, either
metallic Cu0 sites, oxidized Cu+ sites dispersed on the oxide component or at the Cuoxide interface, or a combination of both kinds of sites are thought to contribute to
the active ensembles at the Cu surface. Furthermore, the oxidic surface of the refractory component may take part in the catalytic reaction and provide adsorption sites
for the oxygenate-bonded species [126], whereas hydrogen is probably adsorbed
at the metallic Cu surface. Similar to methanol synthesis, factors intrinsic to the
Cu phase also contribute to the MSR activity in addition to SACu. There are two
major views discussed in the literature relating these intrinsic factors either to the
variable oxidation state of Cu, in particular to the in situ adjustment of the Cu0/Cu+
ratio at the catalysts surface [102, 107, 127132], or to the defect structure and varying
433
434
future. Although most researchers agree that copper-based catalysts will remain the
industry standard, numerous possibilities exist for the improvement of the activity,
selectivity, and stability of these materials through the incorporation of promoters,
more efcient exploitation of the Cu-ZnO synergy, and implementation of new synthesis strategies. Therefore, a deeper understanding of the active sites for methanol
synthesis, MSR, and rWGS on Cu-based catalysts is desirable and essential to further
knowledge-based catalyst development. Here, combined effort from the elds of
theory, material science, model catalysis, and reaction kinetics is needed. Several
studies have exhaustively characterized model systems whose structure does not
represent the unique interspersed structure of industrial Cu/ZnO-based catalysts.
Although these studies shed some light on the general nature of the Cu-Zn interaction, they can only provide limited insight into what makes the current state of the
industry catalyst active, and how it could be improved. With that in mind, complementary in situ structural und surface studies on working industrial catalysts would
be benecial to validate the conclusions drawn from model studies and promise
further progress in the understanding and optimization of methanol catalysts.
In contrast to traditional methanol synthesis, in which selectivity to CO formation
is not a major issue, diminishing the rate of rWGS is critical to the efcient methanol
synthesis from CO2 as well as MSR. In that respect, properties such as particle size,
interaction with oxide matrix, and identity of exposed Cu crystal planes should be
investigated in more detail with regard to CO selectivity. Furthermore, targeted synthesis strategies can be developed to take these factors into account and in the end
allow for tailoring the CO2 and CO reaction channels on Cu surfaces. Due to the relatively harsh reaction conditions that hinder application of some in situ techniques
and the entangled reaction network of relevant C1 conversions, the contribution
from theory to mechanistic understanding is critical. Because such studies should
be consolidated with experimental data carefully determined under relevant conditions, a stronger focus on the real multicomponent functional catalyst is also
needed. Although it has been conclusively established that CO2 is the source of methanol on Cu/ZnO, a very recent detailed theoretical study on Cu surfaces suggests that
CO may not only promote CO2 hydrogenation but also directly contribute to methanol formation [37]. Since this study does not take into account Zn promotion, it is
difcult to extrapolate their conclusions to the industrial system. One solution to this
discrepancy would be to construct models based on input from characterization studies of industrial materials. As a second option, experimentalists could approach the
current models and validate them by determining the methanol carbon source on Cu
particles dispersed on inert supports (irreducible carbon, silica, alumina). Both approaches would yield signicant insight into the role of ZnO, Cu as well as the respective
synthesis gas components in the H2/CO2/CO/CH3OH/H2O reaction network.
Future widespread use of anthropogenic CO2 in combination with renewable
hydrogen as well as the implementation of coal, biomass, and other nonconventional
sources of synthesis gas will lead to suboptimal synthesis gas compositions. Efcient
incorporation of these synthesis gas mixtures into the current methanol synthesis infrastructure will necessitate the redevelopment of catalysts to perform stably under
high concentrations of CO2, water, and impurities. To that end, advanced characterization methods must be implemented to discriminate between surface area loss by
5.3.10 References
435
sintering, loss of active sites by defect annealing, and poisoning by impurities. The
freely available data basis of Cu/ZnO catalyst deactivation is still relatively sparse,
and more investigations are needed to draw rm conclusions on the role of different
deactivation mechanisms. Additionally, current high-throughput testing methods for
methanol synthesis catalysts would need to be modied to take product inhibition
from water into account. Deeper understanding of the performance of a catalyst at
industrially relevant integral conditions can be gained through activity measurements
at differential conditions in the presence and absence of water, as performed by
Sahibzada et al. [34]. In this manner, the catalysts intrinsic activity can be deconvoluted from its susceptibility to inhibition, thus allowing both properties to be
optimized independently.
In summary, as one among other energy storage strategies, methanol has great
potential as a sustainable synthetic fuel and seems particularly promising for the
transportation sector. The industrial methanol synthesis process with Cu/ZnObased catalysts represents a good starting point for implementation of methanol
chemistry in a future energy scenario. However, the thermodynamic constraints of
CO2 hydrogenation require further research. The key remains the development of
catalysts with better low-temperature activity. Despite long-lasting experience with
the industrial processes, methanol chemistry is scientically not yet mature. Elaboration of a rm scientic basis for effective catalyst design by resolving the open questions as to the mechanism and nature of the active sites of the relevant reactions is a
major challenge for the future.
5.3.9 Notes
1. Part of this section is based on work previously published in [56].
2. This section is based on the work previously published in [56].
5.3.10 References
1. Moffat SM. Methanol powered cars get ready to hit the road. Science. 1991;251
(4993):5145.
2. Olah GA, Goeppert A, Prakash GKS. Beyond oil and gas: the methanol economy. Weinheim an der Bergstrasse, Germany: Wiley-VCH; 2006.
3. Aldrich B. ABCs of AFVs: a guide to alternative fuel vehicles. Diane Pub.; 1995. http://
infohouse.p2ric.org/ref/26/25156.pdf.
4. Methanol: the clear alternative for transportation. 2011. Available from: http://www.
methanol.org.
5. Reed TB, Lerner RM. Methanol versatile fuel for immediate use. Science. 1973;182
(4119):1299304.
6. Rihko-Struckmann LK, Peschel A, Hanke-Rauschenbach R, Sundmacher K. Assesment
of methanol synthesis utilizing exhaust CO2 for chemical storage of electrical energy.
Ind Eng Chem Res. 2010;49:1107311078.
7. Skrzypek J, et al. Thermodynamics and kinetics of low pressure methanol synthesis. Chem
Eng J. 1995;58:1018.
436
8. Hansen JB, Nielsen PEH. Methanol synthesis. In: Ert, et al., editors. Handbook of heterogenous catalysis. Weinheim, Germany: Wiley-VCH; 2008.
9. Twigg MV, Spencer MS. Deactivation of copper metal catalysts for methanol decomposition, methanol steam reforming and methanol synthesis. Top Catal. 2003;22(34):
191203.
10. Bernton H, Kovarik W, Sklar S. A history of power alcohol. Lincoln, NE: Bison Press;
2010.
11. Global B, editor. Statistical review of world energy. BP Global, 2011.
12. Palo DR, Dagle RA, Holladay JD. Methanol steam reforming for hydrogen production.
Chem Rev. 2007;107(10):39924021.
13. Stiles A. Methanol, past, present, and speculation on the future. AIChE J. 1977;23(3):
36275.
14. Lewis WK, Frolich PK. Synthesis of methanol from carbon monoxide and hydrogen. Ind
Eng Chem. 1928;20(285):28590.
15. Weissermel K, Arpe H-J. Industrial organic chemistry. 4th completely rev. ed. Weinheim,
Germany: Wiley-VCH; 2003.
16. Fiedler E, et al. Methanol in Ullmanns Encyclopedia of Industrial Chemistry. Weinheim,
Germany: Wiley-VCH; 2011.
17. Kung HH. Deactivation of methanol synthesis catalysts a review. Catal Today. 1992;11(4):
44353.
18. Kung HH. Methanol synthesis. Catal Rev. 1980;22(2):23559.
19. Klier K, et al. Catalytic synthesis of methanol from CO/H2: IV. The effect of carbon
dioxide. J Catal. 1981;74:34360.
20. Chinchen GC, et al. Mechanism of methanol synthesis from CO2/CO/H2 mixtures over
copper/zinc oxide/alumina catalysts use of 14C-labeled reactants. Appl Catal. 1987;30(2):
3338.
21. Rozovskii AY. Kinetika i Kataliz. 1980;21:87.
22. Fujita S, et al. Mechanism of the formation of precursors for the Cu/ZnO methanol synthesis catalysts by a coprecipitation method. Catal Lett. 1995;34(12):8592.
23. Waugh KC. Methanol synthesis. Catal Today. 1992;15(1):5175.
24. Saussey J, Lavalley JC, Bovet CJ. Chem Soc Faraday Trans. 1982;87:1457.
25. Yang Y, et al. Isotope effects in methanol synthesis and the reactivity of copper formates
on a Cu/SiO2 catalyst. Catal Lett. 2008;125(34):2018.
26. Yang Y, et al. (Non)formation of methanol by direct hydrogenation of formate on copper
catalysts. J Phys Chem C. 2010;114(40):1720511.
27. Muhler M, et al. On the role of adsorbed atomic oxygen and CO2 in copper-based methanol synthesis catalysts. Catal Lett. 1994;25(12):110.
28. Nakamura J, Campbell JM, Campbell CT. Kinetics and mechanism of the water-gas shift
reaction catalyzed by the clean and Cs-promoted Cu(110) surface a comparison with Cu
(111). J Chem Soc Faraday Trans. 1990;86(15):272534.
29. Szanyi J. Goodman DW. Methanol synthesis on a Cu(100) catalyst. Catal Lett. 1991;10
(56):38390.
30. Yoshihara J, et al. Methanol synthesis and reverse water-gas shift kinetics over clean
polycrystalline copper. Catal Lett. 1995;31(4):31324.
31. Yoshihara J, Campbell CT. Methanol synthesis and reverse water-gas shift kinetics over
Cu(110) model catalysts: structural sensitivity. J Catal. 1996;161(2):77682.
32. Rasmussen PB, et al. Methanol synthesis on Cu(100) from a binary gas-mixture of CO2
and H2. Catal Lett. 1994;26(34):37381.
33. Nakamura J, et al. Methanol synthesis over a Zn-deposited copper model catalyst. Catal
Lett. 1995;31(4):32531.
5.3.10 References
437
34. Sahibzada M, Metcalfe IS, Chadwick D. Methanol synthesis from CO/CO2/H-2 over
Cu/ZnO/Al2O3 at differential and nite conversions. J Catal. 1998;174(2):1118.
35. Askgaard TS, et al. A kinetic-model of methanol synthesis. J Catal. 1995;156(2):22942.
36. Yang YX, et al. Fundamental studies of methanol synthesis from CO2 hydrogenation on
Cu(111), Cu clusters, and Cu/ZnO(0001). Phys Chem Chem Phys. 2010;12(33): 990917.
37. Grabow LC, Mavrikakis M. Mechanism of methanol synthesis on Cu through CO(2) and
CO hydrogenation. ACS Catal. 2011;1(4):36584.
38. Hu ZM, Takahashi K, Nakatsuji H. Mechanism of the hydrogenation of CO2 to methanol
on a Cu(100) surface: dipped adcluster model study. Surface Sci. 1999;442(1):90106.
39. Morikawa, Y., K. Iwata, and K. Terakura, Theoretical study of hydrogenation process of
formate on clean and Zn deposited Cu(111) surfaces. Appl Surface Sci. 2001;170:1115.
40. Zhao YF, et al. Insight into methanol synthesis from CO2 hydrogenation on Cu(111):
complex reaction network and the effects of H2O. J Catal. 2011;281(2):199211.
41. Saito M. R&D activities in Japan on methanol synthesis from CO2 and H2. Catal Surv
Jpn. 1998;2:17584.
42. Lee JS, et al. A comparative-study of methanol synthesis from CO2/H2 and CO/H2 over a
Cu/ZnO/Al2O3 catalyst. J Catal. 1993;144(2):41424.
43. Chanchlani KG, Hudgins RR, Silveston PL. Methanol synthesis from H2, CO, and CO2
over Cu/ZnO catalysts. J Catal. 1992;136(1):5975.
44. Liu G, et al. The rate of methanol production on a copper-zinc oxide catalyst the dependence on the feed composition. J Catal. 1984;90(1):13946.
45. Bart JC, Sneeden RPA. Catal Today. 1987;2:1.
46. Behrens M. Meso- and nano-structuring of industrial Cu/ZnO/(Al2O3) catalysts. J Catal.
2009;267(1):249.
47. Behrens M, et al. Phase-pure Cu,Zn,Al hydrotalcite-like materials as precursors for copper
rich Cu/ZnO/Al2O3 catalysts. Chem Mater. 2009;22(2):38697.
48. Greeley J, Mavrikakis M. Methanol decomposition on Cu(111): a DFT study. J Catal.
2002;208(2):291300.
49. Arena F, et al. Solid-state interactions, adsorption sites and functionality of Cu-ZnO/ZrO2
catalysts in the CO2 hydrogenation to CH3OH. Appl Catal A. 2008;350:1623.
50. Soczynski J, et al. Effect of metal oxide additives on the activity and stability of Cu/ZnO/
ZrO2 catalysts in the synthesis of methanol from CO2 and H2. Appl Catal A. 2006;
310:12737.
51. An X, et al. Methanol synthesis from CO2 hydrogenation with a Cu/Zn/Al/Zr brous
catalyst. Chin J Chem Eng. 2009;17(1):8894.
52. Liao F, et al. Morphology-dependent interactions of ZnO with Cu nanoparticles at the
materials interface in selective hydrogenation of CO2 to CH3OH. Angew Chem Int Ed.
2011;50:21625.
53. Arena F, et al. Synthesis, characterization and activity pattern of Cu-ZnO/ZrO2 catalysts
in the hydrogenation of carbon dioxide to methanol. J Catal. 2007;249(2):18594.
54. Saito M, et al. Development of copper/zinc oxide-based multicomponent catalysts for methanol synthesis from carbon dioxide and hydrogen. Appl Catal A Gen. 1996;138(2):31118.
55. Kurtz M, et al. New synthetic routes to more active Cu/ZnO catalysts used for methanol
synthesis. Catal Lett. 2004;92(12):4952.
56. Behrens M., Armbrster M. Methanol steam reforming. In: Guczi L, Erdohelyi A, editors.
Catalysis for alternative energy generation. New York: Springer; 2012:175236.
57. Kasatkin I, et al. Role of lattice strain and defects in copper particles on the activity of
Cu/ZnO/Al2O3 catalysts for methanol synthesis. Angew Chem Int Ed. 2007;46(38):73247.
58. Behrens M, et al. Methanol synthesis over Cu/ZnO/Al2O3: the active site in industrial
catalysis. Science, 336(6083):893897.
438
59. Chinchen GC, et al. The measurement of copper surface-areas by reactive frontal chromatography. J Catal. 1987;103(1):7986.
60. Hinrichsen O, Genger T, Muhler M. Chemisorption of N2O and H2 for the surface determination of copper catalysts. Chem Eng Technol. 2000;23(11):9569.
61. Behrens M, et al. The potential of microstructural optimization in metal/oxide catalysts:
higher intrinsic activity of copper by partial embedding of copper nanoparticles. ChemCatChem. 2010;2(7):81618.
62. Spencer MS. The role of zinc oxide in Cu ZnO catalysts for methanol synthesis and the
water-gas shift reaction. Top Catal. 1999;8(34):25966.
63. Burch R, Chappell RJ, Golunski SE. Synergy between copper and zinc-oxide during methanol synthesis transfer of activating species. J Chem Soc Faraday Trans I. 1989;85:356978.
64. Kanai Y, et al. The synergy between Cu and ZnO in methanol synthesis catalysts. Catal
Lett. 1996;38(34):15763.
65. Klier K. Methanol synthesis. Adv Catal. 1982;31:243313.
66. Jansen WPA, et al. Dynamic behavior of the surface structure of Cu/ZnO/SiO2 catalysis.
J Catal. 2002;210(1):22936.
67. Frost JC. Junction effect interactions in methanol synthesis catalysts. Nature 1988;334
(6183):57780.
68. Nakamura J, Choi Y, Fujitani T. On the issue of the active site and the role of ZnO in
Cu/ZnO methanol synthesis catalysts. Top Catal. 2003;22(34):27785.
69. Topsoe NY, Topsoe H. On the nature of surface structural changes in Cu ZnO methanol
synthesis catalysts. Top Catal. 1999;8(34):26770.
70. dAlnoncourt RN, et al. The inuence of strongly reducing conditions on strong metalsupport interactions in Cu/ZnO catalysts used for methanol synthesis. Phys Chem Chem
Phys. 2006;8(13):152538.
71. Grunwaldt JD, et al. In situ investigations of structural changes in Cu/ZnO catalysts.
J Catal. 2000;194(2):45260.
72. Hansen PL, et al. Atom-resolved imaging of dynamic shape changes in supported copper
nanocrystals. Science. 2002;295(5562):20535.
73. Kurr P, et al. Microstructural characterization of Cu/ZnO/Al2O3 catalysts for methanol
steam reforming a comparative study. Appl Catal A Gen. 2008;348(2):15364.
74. Kniep BL, et al. Rational design of nanostructured copper-zinc oxide catalysts for the
steam reforming of methanol. Angew Chem Int Ed. 2004;43(1):11215.
75. Gunter MM, et al. Redox behavior of copper oxide/zinc oxide catalysts in the steam reforming of methanol studied by in situ X-ray diffraction and absorption spectroscopy.
J Catal. 2001;203(1):13349.
76. Sakong S, Gross A. Dissociative adsorption of hydrogen on strained Cu surfaces. Surface
Sci. 2003;525(13):10718.
77. Girgsdies F, et al. Strained thin copper lms as model catalysts in the materials gap. Catal
Lett. 2005;102(12):917.
78. Hammer B, Norskov JK. Electronic factors determining the reactivity of metal surfaces.
Surface Sci. 1995;343(3):21120.
79. Behrens M, et al. Understanding the complexity of a catalyst synthesis: co-precipitation of
mixed Cu,Zn,Al hydroxycarbonate precursors for Cu/ZnO/Al2O3 catalysts investigated by
titration experiments. Appl Catal A Gen. 2011;392(12):93102.
80. Schimpf S. Methanol catalysts. In: de Jong KP, editor. Synthesis of solid catalysts. Weinheim, Germany: Wiley-VCH; 2009. p. 329.
81. Waller D, et al. Copper-zinc oxide catalysts. Activity in relation to precursor structure and
morphology. Faraday Discuss Chem Soc. 1989;87:10720.
5.3.10 References
439
82. Whittle DM, et al. Co-precipitated copper zinc oxide catalysts for ambient temperature
carbon monoxide oxidation: effect of precipitate ageing on catalyst activity. Phys Chem
Chem Phys. 2002;4(23):591520.
83. Fierro G, et al. Study of the reducibility of copper in CuO-ZnO catalysts by temperatureprogrammed reduction. Appl Catal A Gen. 1996;137(2):32748.
84. Gunter MM, et al. Implication of the microstructure of binary Cu/ZnO catalysts for their
catalytic activity in methanol synthesis. Catal Lett. 2001;71(12):3744.
85. Baltes C, Vukojevic S, Schuth F. Correlations between synthesis, precursor, and catalyst
structure and activity of a large set of CuO/ZnO/Al2O3 catalysts for methanol synthesis.
J Catal. 2008;258(2):33444.
86. Li JL, Inui T. Characterization of precursors of methanol synthesis catalysts, copper zinc
aluminum oxides, precipitated at different pHs and temperatures. Appl Catal A Gen.
1996;137(1):10517.
87. Bems B, et al. Relations between synthesis and microstructural properties of copper/zinc
hydroxycarbonates. Chem Eur J. 2003;9(9):203952.
88. Shen GC, Fujita SI, Takezawa N. Preparation of precursors for the Cu/ZnO methanol
synthesis catalysts by coprecipitation methods effects of the preparation conditions
upon the structures of the precursors. J Catal. 1992;138(2):7548.
89. Behrens M, et al. Knowledge-based development of a nitrate-free synthesis route for Cu/
ZnO methanol synthesis catalysts via formate precursors. Chem Commun. 2011;47(6):
17013.
90. Christiansen JA. A reaction between methyl alcohol and water and some related reactions. J Am Chem Soc. 1921;43:16702.
91. Prigent M. On board hydrogen generation for fuel cell powered electric cars a review of
various available techniques. Rev Inst Fr Petrole. 1997;52(3):34960.
92. Navarro RM, Pena MA, Fierro JLG. Hydrogen production reactions from carbon feedstocks: fossils fuels and biomass. Chem Rev. 2007;107(10):395291.
93. Sa S, et al. Catalysts for methanol steam reforming-a review. Appl Catal B Environ.
2010;99(12):4357.
94. de Wild PJ, Verhaak MJFM. Catalytic production of hydrogen from methanol. Catal
Today. 2000;60(12):310.
95. Joensen F, Rostrup-Nielsen JR. Conversion of hydrocarbons and alcohols for fuel cells.
J Power Sources. 2002;105(2):195201.
96. Cheekatamarla PK, Finnerty CM. Reforming catalysts for hydrogen generation in fuel
cell applications. J Power Sources. 2006;160(1):4909.
97. Velu S, Suzuki K, Osaki T. Oxidative steam reforming of methanol over CuZnAl(Zr)oxide catalysts; a new and efcient method for the production of CO-free hydrogen
for fuel cells. Chem Commun. 1999(23):23412.
98. Lattner JR, Harold MP. Autothermal reforming of methanol: experiments and modeling. Catal Today. 2007;120(1):7889.
99. Park ED, Lee D, Lee HC. Recent progress in selective CO removal in a H2-rich stream.
Catal Today. 2009;139(4):28090.
100. Tsai MC, et al. Promotion of a copper-zinc catalyst with rare earth for the steam reforming of methanol at low temperatures. J Catal. 2011;279(2):2415.
101. Breen JP, Ross JRH. Methanol reforming for fuel-cell applications: development of
zirconia-containing Cu-Zn-Al catalysts. Catal Today. 1999;51(34):52133.
102. Agrell J, et al. Production of hydrogen from methanol over Cu/ZnO catalysts promoted
by ZrO2 and Al2O3. J Catal. 2003;219(2):389403.
103. Matsumura Y, Ishibe H. High temperature steam reforming of methanol over Cu/ZnO/
ZrO2 catalysts. Appl Catal B Environ 2009;91(12):52432.
440
104. Jones SD, Hagelin-Weaver HE. Steam reforming of methanol over CeO2- and ZrO2promoted Cu-ZnO catalysts supported on nanoparticle Al2O3. Appl Catal B Environ.
2009;90(12):195204.
105. Velu S, Suzuki K. Selective production of hydrogen for fuel cells via oxidative steam reforming of methanol over CuZnAl oxide catalysts: effect of substitution of zirconium and
cerium on the catalytic performance. Top Catal. 2003;22(34):23544.
106. Velu S, et al. Oxidative steam reforming of methanol over CuZnAl(Zr)-oxide catalysts
for the selective production of hydrogen for fuel cells: catalyst characterization and performance evaluation. J Catal. 2000;194(2):37384.
107. Costantino U, et al. Cu-Zn-Al hydrotalcites as precursors of catalysts for the production
of hydrogen from methanol. Solid State Ionics. 2005;176(3940):291722.
108. Turco M, et al. Production of hydrogen from oxidative steam reforming of methanol
I. Preparation and characterization of Cu/ZnO/Al2O3 catalysts from a hydrotalcite-like
LDH precursor. J Catal. 2004;228(1):4355.
109. Turco M, et al. Production of hydrogen from oxidative steam reforming of methanol II.
Catalytic activity and reaction mechanism on Cu/ZnO/Al2O3 hydrotalcite-derived catalysts. J Catal. 2004;228(1):5665.
110. Ritzkopf I, et al. Decreased CO production in methanol steam reforming over Cu/ZrO2
catalysts prepared by the microemulsion technique. Appl Catal A Gen. 2006;302(2):
21523.
111. Purnama H, et al. Activity and selectivity of a nanostructured CuO/ZrO2 catalyst in the
steam reforming of methanol. Catal Lett. 2004;94(12):618.
112. Liu YY, et al. Highly active copper/ceria catalysts for steam reforming of methanol. Appl
Catal A Gen. 2002;223(12):13745.
113. Szizybalski A, et al. In situ investigations of structure-activity relationships of a Cu/ZrO2
catalyst for the steam reforming of methanol. J Catal. 2005;233(2):297307.
114. Mastalir A, et al. Steam reforming of methanol over Cu/ZrO2/CeO2 catalysts: a kinetic
study. J Catal. 2005;230(2):46475.
115. Jiang CJ, et al. Kinetic mechanism for the reaction between methanol and water over a
Cu-ZnO-Al2O3 catalyst. Appl Catal A Gen. 1993;97(2):14558.
116. Takezawa N, Iwasa N. Steam reforming and dehydrogenation of methanol: difference in
the catalytic functions of copper and group VIII metals. Catal Today. 1997;36(1):4556.
117. Peppley BA, et al. Methanol-steam reforming on Cu/ZnO/Al2O3 catalysts. Part 2. A
comprehensive kinetic model. Appl Catal A Gen. 1999;179(12):3149.
118. Rozovskii AY, Lin GI. Fundamentals of methanol synthesis and decomposition. Top
Catal. 2003;22(34):13750.
119. Lee JK, Ko JB, Kim DH. Methanol steam reforming over Cu/ZnO/Al2O3 catalyst: kinetics and effectiveness factor. Appl Catal A Gen. 2004;278(1):2535.
120. Peppley BA, et al. Methanol-steam reforming on Cu/ZnO/Al2O3. Part 1: The reaction
network. Appl Catal A Gen. 1999;179(12):219.
121. Frank B, et al. Steam reforming of methanol over copper-containing catalysts: inuence
of support material on microkinetics. J Catal. 2007;246(1):17792.
122. Lin S, et al. Pathways for methanol steam reforming involving adsorbed formaldehyde
and hydroxyl intermediates on Cu(111): density functional theory studies. Phys Chem
Chem Phys. 2011;13(20):962231.
123. Lin S, Xie DQ, Guo H. Methyl formate pathway in methanol steam reforming on copper: density functional calculations. ACS Catal. 2011;1(10):126371.
124. Spencer MS. On the activation-energies of the forward and reverse water-gas shift reaction. Catal Lett. 1995;32(12):913.
5.3.10 References
441
125. Saito M, et al. Effects of ZnO contained in supported Cu-based catalysts on their activities for several reactions. Catal Lett. 2002;83(12):14.
126. Noei H, et al. The identication of hydroxyl groups on ZnO nanoparticles by infrared
spectroscopy. Phys Chem Chem Phys. 2008;10(47):70927.
127. Goodby BE, Pemberton JE. XPS characterization of a commercial Cu/ZnO/Al2O3
catalyst effects of oxidation, reduction, and the steam reformation of methanol. Appl
Spectrosc. 1988;42(5):75460.
128. Raimondi F, et al. Structural changes of model Cu/ZnO catalysts during exposure to
methanol reforming conditions. Surface Sci. 2003;532:3839.
129. Raimondi F, Wambach J, Wokaun A. Structural properties of Cu/ZnO/Si methanol reforming catalysts: inuence of the composition of the reactant mixture and of the Cu
island size. Phys Chem Chem Phys. 2003;5(18):401524.
130. Reitz TL, et al. Time-resolved XANES investigation of CuO/ZnO in the oxidative methanol reforming reaction. J Catal. 2001;199(2):193201.
131. Busca G, et al. Methanol steam reforming over ex-hydrotalcite Cu-Zn-Al catalysts. Appl
Catal A Gen. 2006;310:708.
132. Busca G, et al. Hydrogen from alcohols: IR and ow reactor studies. Catal Today.
2009;143(12):28.
133. Vargas MAL, et al. An IR study of methanol steam reforming over ex-hydrotalcite
Cu-Zn-Al catalysts. J Mol Catal A Chem. 2007;266(12):18897.
134. Zhang XR, et al. A unique microwave effect on the microstructural modication of
Cu/ZnO/Al2O3 catalysts for steam reforming of methanol. Chem Commun. 2005(32):
41046.
135. Wang LC, et al. Production of hydrogen by steam reforming of methanol over Cu/ZnO catalysts prepared via a practical soft reactive grinding route based on dry oxalate-precursor
synthesis. J Catal. 2007;246(1):193204.
136. Lofer DG, McDermott SD, Renn CN. Activity and durability of water-gas shift catalysts used for the steam reforming of methanol. J Power Sources. 2003;114(1):1520.
137. Thurgood CP, et al. Deactivation of Cu/ZnO/Al2O3 catalyst: evolution of site concentrations with time. Top Catal. 2003;22(34):2539.
138. Agrell J, Birgersson H, Boutonnet M. Steam reforming of methanol over a Cu/ZnO/Al2O3
catalyst: a kinetic analysis and strategies for suppression of CO formation. J Power
Sources. 2002;106(12):24957.
139. Iwasa N, Masuda S, Takezawa N. Steam reforming of methanol over Ni, Co, Pd and Pt
supported on ZnO. React Kinet Catal Lett. 1995;55(2):34953.
140. Penner S, et al. Growth and structural stability of well-ordered PdZn alloy nanoparticles.
J Catal. 2006;241(1):1419.
141. Karim A, Conant T, Datye A. The role of PdZn alloy formation and particle size on the
selectivity for steam reforming of methanol. J Catal., 2006;243(2):4207.
142. Liu ST, et al. Hydrogen production by oxidative methanol reforming on Pd/ZnO: catalyst preparation and supporting materials. Catal Today. 2007;129(34):28792.
143. Fottinger K, et al. Dynamic structure of a working methanol steam reforming catalyst: in
situ quick-EXAFS on Pd/ZnO nanoparticles. J Phys Chem Lett. 2011;2(5):42833.
144. Friedrich M, Teschner D, Knop-Gericke A, Armbrster M. Inuence of bulk composition of the intermetallic compound ZnPd on surface composition and methanol steam
reforming properties. J Catal. 2012;285:4147.
145. Ota A, Kunkes EL, Kasatkin I, Groppo E, Ferri D, Poceiro B, Navarro Yerga RM,
Behrens M. Comparative study of hydrotalcite-derived supported Pd2Ga and PdZn intermetallic nanoparticles as methanol synthesis and methanol steam reforming catalysts.
J Catal. 2012;293:2738.
5.4.1 Introduction
Synthesis gas, or syngas, is a mixture of carbon monoxide, carbon dioxide, and
hydrogen. Syngas can be produced from many sources, including natural gas,
coal, biomass, or virtually any hydrocarbon feedstock, by gasifying it with an oxidant such as O2 or steam. Syngas is a versatile intermediate for the production of
hydrogen, ammonia, methanol, and in Fischer-Tropsch synthesis (FTS), the production of synthetic hydrocarbon fuels. As such, syngas is also a key component in
present and future sustainable energy technology. Figure 5.4.1 summarizes the conversions. In this chapter, we describe how syngas can be produced and how it is
converted into substances that are in essence storage media for chemical forms of
energy.
444
Gas
Steam
reforming
Oil
Partial
Oxidation
Coal
Synthesis gas
Water-gas-shift
CO + H2
Fischer-Tropsch
Synfuels
synthesis
(Syngas)
Gasification
Bio
mass
H2 (NH3)
reaction
Methanol
CH3OH
dimethylether
synthesis
Reforming reactions
CH4 + H2O S CO + 3 H2
CnHm + m H2O S n CO + (n+ m) H2
+206 kJ/mol
endothermic
CO + H2O S CO2 + H2
Oxidation reactions
CH4
CH4
CH4
CH4
+
+
+
+
O2 S CO + 2 H2
O2 S CO + H2O + H2
1 O2 S CO + 2 H2O
2 O2 S CO2 + 2 H2O
36 kJ/mol
278 kJ/mol
519 kJ/mol
802 kJ/mol
complete combustion
+ 75 kJ/mol
172 kJ/mol
131 kJ/mol
endothermic
Boudouard reaction
Auto thermal
reforming
Catalytic partial
oxidation
Partial
oxidation
CPO
POX
ATR
O2
O2
mixing
Combustion:
generation
of heat
Catalyst
bed
CH4
Catalyst
bed
Thermal and
catalytic reforming
Heat recovery
section
Steam
CH4
CH4 + H2O
Burners
Hydrocarbon
feed
O2
Fuel
Purification
Syngas
Syngas
Syngas
Syngas
445
carbon than methane does, and therefore, catalysts for the reforming of naphtha
need special promoters and are generally more complex than those for reforming
of natural gas.
Partial oxidation of methane (or hydrocarbons) is another option to produce syngas [4]. This process, which runs without a catalyst, needs high temperatures for high
CH4 conversion and to suppress soot formation. The process can handle other feedstocks, such as heavy oil factions and biomass, and yields syngas with a H2/CO ratio
of about 2. The process is eminently suitable for large-scale production of syngas
(e.g. for gas-to-liquids [GTL] plants).
A catalytic version of partial oxidation (CPO) exists as well [5, 6]. It is based on
short-contact time conversion of methane, hydrocarbons, or biomass on, for example,
rhodium catalysts. This process is suitable for small-scale applications.
An often-applied hybrid process combining SMR and oxidation is the autothermalreforming (ATR) process [1]. Here, the heat for the reforming is generated inside the
reactor by oxidizing some of the methane. Syngas forms both by catalytic reforming
(over a nickel catalyst) and by thermal partial oxidation. The H2/CO ratio can be varied between 2 and 3.5. As partial oxidation (POX), ATR is also suitable for large-scale
production of syngas for GTL or larger-scale methanol synthesis processes.
An alternative route to synthesis gas, which may become important in the future, is
by reducing CO2 (e.g. from ue gas) with H2 originating from the electrolytic splitting of water. This option is also interesting from the viewpoint of storage of wind or
solar energy [7].
In the next section, we describe the major applications of syngas: production of
hydrogen, methanol, and synthetic fuels, with emphasis on the latter.
41 kJ/mol
(1)
is not only implicitly present in the production of syngas, but can also be applied separately to adjust the H2/CO ratio of the syngas, or to shift the latter entirely to H2.
The principle is that CO is a strong reducing agent, which can be used to convert
water into H2. Being exothermic, the reaction runs at much lower temperatures
than the reforming and oxidation reactions described previously. The reaction uses
a catalyst. The high temperature WGS runs at 300C500 and uses an iron-oxide
based catalyst, which is robust in the sense that it can handle impurities such as sulfur. It is often followed by a more delicate low-temperature process (around 200C),
which is based on a copper-zinc oxide catalyst. The latter allows for a higher conversion to hydrogen but is less tolerant with respect to impurities. As this catalyst is also
a successful methanol synthesis catalyst, it needs to be promoted to suppress methanol by-product formation. The exit gas contains unconverted methane, which is
446
removed by an adsorptive process, called pressure swing adsorption. In this way, the
product hydrogen can be upgraded to more than 99.99% purity [1].
(2)
Combining it with the WGS reaction (1) discussed previously, one obtains
CO + 2 H2 S CH3OH 91 kJ/mol
(3)
The process is exothermic, and the heat released by the methanol synthesis is used to
partially heat up the natural gas for the endothermic synthesis gas generation.
The modern methanol synthesis catalyst consists of copper, zinc oxide, and alumina.
Copper metal is seen as the catalytically active phase, and ZnO as the promoter. It is
well known that the interaction between the two components is essential for achieving a high activity, but the nature of the promoting effect is still a matter of debate.
Loss of activity is caused by sintering of the Cu crystallites, and, if the feed gas
contains impurities such as chlorine and sulfur, by poisoning.
447
The formal overall reaction equations are straightforward, but they hide a
tremendous amount of mechanistic detail as we discuss later on in this chapter:
n CO + (2n+1) H2 CnH2n+2 + n H2O
(4)
n CO + 2n H2 CnH2n + n H2O
(5)
n CO + 2n H2 CnH2n+1OH + (n 1) H2O
(6)
Typical reaction conditions include temperatures between 200C and 350C and
pressures between 20 and 50 bar. The reactions are strongly exothermic, and dealing
with the heat is an important engineering consideration when designing reactors for
the process.
Water is an important by-product of the FTS. With cobalt catalysts, essentially all
oxygen from CO dissociation (typically around 99%) is discarded as water. In this
case, the FTS as represented by Equations (4)(6) needs a H2:CO ratio of at least 2.
Iron catalysts, however, possess activity for the WGS reaction (1), and hence a significant portion of the oxygen is also discarded as CO2. In this case, the H2:CO ratio may
be as low as 0.5. This becomes evident if one adds the symbolic FTS reaction equation
per one CO molecule:
CO + 2 H2 = CH2 + H2O
(7)
(8)
In this way, iron generates hydrogen from CO and water, the important implication
being that processes based on iron can handle syngas with a low H2/CO ratio as is
obtained from coal.
The low WGS activity of cobalt makes it the preferred catalyst for GTL technology since the H2/CO ratio of syngas from natural gas is close to the desired
composition for hydrocarbon formation.
448
The mechanism of the FTS is still a subject of scientic debate. In the simplest
mechanism [17], CO adsorbs and dissociates, C atoms are hydrogenated to CHx species, which then couple to form longer hydrocarbons. We illustrate this with the most
straightforward but not necessarily correct! sequence, in which * denotes an
active site on the catalysts surface:
Adsorption and dissociation
CO + * S COads
(9)
(10)
H2 + 2* S 2 Hads
(11)
(12)
(13)
C hydrogenation
Cads + Hads S CHads + *
(14)
(15)
(16)
Chain growth
CH3,ads + CH2,ads S C2H5,ads + *
(17)
(18)
Termination
CH3,ads + Hads S CH4; + 2*
(19)
(20)
(21)
Other schemes exist as well, for example, where the initial CO dissociation step is
replaced by a hydrogen-assisted step (Figure 5.4.3) [1720], or where the chain
growth is proposed to occur by insertion of CO, while the C-O bond breaks after
incorporation into the growing chain. Computational modeling can investigate the
449
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe Fe Fe Fe Fe
Fe Fe Fe Fe Fe
Fe Fe Fe Fe Fe
Fe Fe Fe Fe Fe
Fe Fe Fe Fe Fe
Fe Fe Fe Fe Fe
+1.09 eV
0.29 eV
0.93 eV
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe
C
Fe Fe Fe Fe Fe
Fe Fe Fe Fe Fe
0.16 eV
+0.94 eV
Fe
Fe
O
CFe
Fe
Fe
Fe
Fe Fe Fe Fe Fe
+0.72 eV
Fe
Fe
Fe
Fe
Fe CFe Fe
Fe
Fe
Fe
Fe
O
Fe CFe Fe
Fe
Fe
Fe
Fe
C
Fe
Fe
Fe
O Fe Fe
Fe Fe Fe
0.0 eV
Fe
Fe
Fe
Fe
Fe Fe OFe Fe Fe
C
Fe
O
Fe CFe Fe
Fe
Fe Fe Fe Fe Fe
Fe Fe Fe Fe Fe
+1.36 eV
0.24 eV
Fe
Fe
Fe Fe Fe Fe Fe
1.0 eV
Fe
Fe
Fe
Fe
Fe
Fe O
FeC Fe Fe Fe Fe
Fe
Fe Fe Fe Fe Fe
Fe
Fe Fe Fe Fe Fe
Fe
Fe
Fe
Fe
Fe
Fe
1.45
1.36
1.09
0.94 HC--O
0.93
0.16
(,)
Fe
C
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe Fe Fe Fe Fe
+1.45 eV
+0.16 eV
Figure 5.4.3
Fe
Fe CFe Fe
+1.12 eV
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe Fe Fe Fe Fe
+2.82 eV
Fe
Fe
Fe
Fe
O
Fe
Fe
Fe Fe Fe Fe Fe
0.31 eV
Fe
C--OH
2.82
Fe
Fe
CO--H
H
+
C--O
HC--O
COH
1.12
HCO
0.72
0.24
(,)
CO+H
0.0(,)*
(bridge)
C+OH
0.31
1.0
C+O+H (,,)
HC+O (,)
0.29 (,,)
450
likelihood of the reaction steps by predicting the energies of the adsorbed intermediates
and the barriers for the transition states. This an active area of research nowadays.
5.4.4.1.2 Hydrocarbon Product Distribution Regardless of what the exact elementary steps at the molecular level are, the FTS as expressed by the general reaction equations (1) and (2) is a polymerization process. And even without knowing the exact
nature of the monomer and the initiating species, we can derive a general expression
for the distribution of chain lengths in the products [14]. We will assume that the
monomer is a species containing one carbon atom, and the chain initiator is a onecarbon species as well. They may be the same but do not have to be. For convenience,
we follow the scheme of reactions (9)(21) discussed previously and consider the initiating species as an adsorbed methyl, CH3, and the monomer as adsorbed CH2. Chain
growth would then correspond to the insertion of the methylene species into the metalalkyl bond, as expressed in reaction steps (17) and (18) discussed previously. The
derivation, however, does not require that we specify the mechanism in detail.
An alkyl species on the surface, schematically indicated as M-CnH2n+1,can either
grow (probability ; 0 < <1) or terminate (probability 1 ) and be hydrogenated
to an alkane, or be dehydrogenated and desorb as an alkene. Obviously, a high value
of results in long hydrocarbon products. We can now derive a general expression
for the probability that a hydrocarbon will terminate as a molecule with n carbon
atoms, by realizing that the chances for formation of methane, ethane/ethylene, and
propane/propylene are equal to C1 = (1 ), C2 = 2(1 ), and C3 = 3(1 )2,
respectively. In general, the chance that a molecule with n carbon atoms forms, is
Cn = n(1 )n1. The total amount of carbon in the product spectrum then forms
a convergent innite sum, which has the solution
Cn =
n (1) n1 =
1
1
(22)
Now we can express the selectivity toward a product with n carbon atoms as
Sn =
Cn
= n (1)2 n1
P
Cn
(23)
and hence a plot of ln Snn versus carbon number (n) yields a straight line with a
slope equal to ln(). The insert of Figure 5.4.4 shows a few examples of ASF plots
for different values of the chain-growth probability.
451
0
5
= 0.95
In (Sn/n)
10
25
= 0.5
35
40
90
Carbon atom selectivity %
15
20
30
100
80
= 0.7
CH4
10
20
30
Carbon number, n
40
50
C5+
C20+
waxes
70
C 2 C4
60
C5 C11
gasoline
50
40
C9 C22
dieseldistillates
30
20
10
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Chain-growth probability,
tions as a function of the chain-growth probability, , as calculated with Equation (23). The
insert shows a few ASF plots according to Equation (24).
Although the ASF distribution is usually a quite good description of the chainlength distribution, deviations occur as well. Often, the methane yield is higher,
and the C2 yield lower, than predicted, and it is also common to observe a higher
chain-growth probability for longer hydrocarbons than for shorter ones [21].
5.4.4.1.3 Fischer-Tropsch Technology FTS can be carried out in several different
reactor types xed bed, uidized bed, or slurry phase and at different temperatures. The high-temperature Fischer-Tropsch (HTFT) synthesis runs at 320C350C,
at which temperatures typically all products are in the gas phase [22]. HTFT is operated in uidized-bed reactors, with iron catalysts. Selectivities correspond to chaingrowth probabilities in the range of 0.700.75 and are ideal for gasoline production,
but olens and oxygenates are formed as well and are used as chemicals.
Low-temperature Fischer-Tropsch (LTFT) synthesis runs at temperatures between
200C and 250C [2325]. The chain-growth probability at these conditions is much
higher than for the HTFT, and as a consequence, the product distribution extends
well into the liquid waxes. LTFT reactors are thus three-phase systems: solid catalysts, gaseous reactants, and gaseous and liquid products. Both cobalt and iron
452
Moving-bed
reactors
Low temperature FTS
200250C
Three-phase system:
gas-liquid-solid
= 0.850.95
Products:
waz, diesel, naphta
Catalysts:
supported cobalt or
precipitated iron
Stationary-bed
reactors
Multitubular
fixed bed
6,000 b/d
Slurry bubble
column,
with wax
24,000 b/d
Circulating
fluidized bed
7,000 b/d
Fixed
fluidized bed
20,000 b/d
Figure 5.4.5 Overview of Fischer-Tropsch technologies and reactor types used. Capacities
are indicated in barrels per day (b/d).
catalysts are applied, although cobalt catalysts are generally used in all modern GTL
plants. Reactors can be xed bed or liquid slurry phase, where the catalyst is
suspended in the liquid product wax.
Although the slurry reactor offers advantages in terms of efcient heat removal
and temperature homogeneity, an efcient solid-liquid separation step is required
to recover the product. The catalyst needs to be highly mechanically stable to offer
sufcient resistance against attrition.
The LTFT synthesis is ideally suited for the production of high-quality middle distillates (diesel and jet fuel) after hydrocracking of the long-chain waxes. The waxes
themselves can be used as specialty waxes. Finally, the product contains naphtha,
which is a feedstock for chemicals.
5.4.4.1.4 Fischer-Tropsch Catalysts For the HTFT process, iron-based catalysts
are used, which are promoted with alkali metals to improve activity and selectivity
[15, 22]. They are normally prepared by fusion at temperatures in excess of 1000C,
during which the mill scale (iron oxide) and the promoters are fused together, followed
by casting into ingots and cooling, milling (to obtain the correct particle sizes), and
reduction/activation in hydrogen.
Iron catalysts can also be used for the LTFT process [15, 23, 26]. Iron-based LTFT
catalysts contain promoters like potassium and copper as well as a structural
453
5 nm
Cobalt
particle
Silica
support
Figure 5.4.6 Silica supported cobalt catalyst with a 10 nm cobalt crystallite (adapted from
Location
Sasolburg, South
Africa
Secunda, South
Africa
Bintulu, Malaysia
Mosselbay, South
Africa
RasLaffan, Qatar
RasLaffan, Qatar
Escravos, Nigeria
Company
Sasol
Sasol
Shell
PetroSA
Sasol-QP
(Oryx)
Shell (Pearl)
ChevronSasol
Natural gas
Natural gas
Natural gas
Natural gas
Natural gas
Carbon feedstock
Multi-tubularxed bed
Slurry phase
Co/Al2O3
Slurry phase
Co/Al2O3
Co/TiO2
1993
Fused Fe/K
2013
2011
2007
1992
19801999
1995
34 000
140 000
34 000
22 000
14 500
160 000
1955ca. 1985
1955
5 000
1993
capacity
(barrel per
day)
Start-up date
Reactor type
Multi-tubularxed bed
Co/SiO2 Co/TiO2
Fused Fe/K
Fused Fe/K
Precipitated Fe/K
Precipitated Fe/K
(spray dried)
Catalyst type
Table 5.4.2
454
5.4 Synthesis Gas to Hydrogen, Methanol, and Synthetic Fuels
5.4.5 References
455
5.4.4.1.5 Historical Development and Future Perspectives The Fischer-Tropsch process dates back to the early 1920s when Franz Fischer and Hans Tropsch demonstrated the conversion of synthesis gas into a mixture of higher hydrocarbons, with
cobalt and iron as a catalyst [35, 36]. Some 20 years earlier, Sabatier had already discovered the reaction from synthesis gas to methane catalyzed by nickel [37]. The FTS
played an important role in the Second World War, as it supplied Germany and
Japan with synthetic fuel. The plants used mainly cobalt catalysts supported on a
silica support called kieselguhr and promoted by magnesia and thoria.
After the war, the German Fischer-Tropsch technology came into the hands of the
Allied forces, and the entire Fischer-Tropsch technology was extensively investigated
at the U.S. Bureau of Mines. The classical textbook by Storch, Golumbic, and
Anderson [12] originates from this period. In the postwar period, Sasol South Africa
was the only user of the technology [38], until in the 1970s two oil crises in combination with the report of the Club of Rome led to the awareness that fossil fuel sources
are nite. The FTS went through a revival, and several companies (ExxonMobil, Gulf,
BP, Shell, Sasol, and Statoil) developed new versions of the FTS, which were mainly
based on cobalt catalysts and production of long hydrocarbon chains, which can, for
example, be hydrocracked to diesel fuel [13, 15, 25, 39].
Nowadays, FTS is regarded as proven technology and an important factor in the
monetization of natural gas. Table 5.4.2 shows an overview of the current and
planned commercial Fischer-Tropsch plants for 2012. Note that capacity totals
only some 400,000 barrels per day in 2013, compared to a total crude oil production
of about 85 million barrels per day.
Applications of Fischer-Tropsch technology in the utilization of biomass are explored as well [4042]. Here the challenge is to cope with higher impurity levels of
biomass-derived syngas. Nevertheless, demonstration plants have been built, and it
is expected that applications will grow in the future.
5.4.5 References
1. Rostrup-Nielsen J, Christiansen LJ. Concepts in syngas manufacture. London: Imperial
College Press; 2011. (Catalytic science series volume 10.)
2. Trimm DL. Coke formation and minimisation during steam reforming reactions. Catal
Today. 1997;37(3):2338.
3. Bengaard HS, Norskov JK, Sehested J, et al. Steam reforming and graphite formation on
Ni catalysts. J Catal. 2002;209(2):36584.
4. ter Haar LW, Vogel JE. In: Proceedings of the Sixth World Petroleum Congress; Frankfurt, 1963.
5. Bharadwaj SS, Schmidt LD. Catalytic partial oxidation of natural gas to syngas. Fuel Processing Technol. 1995;42(23):10927.
6. Hu YH, Ruckenstein E. Catalytic conversion of methane to synthesis gas by partial oxidation and CO2 reforming. Adv Catal. 2004;48:297345.
7. Centi G, Perathoner S. Opportunities and prospects in the chemical recycling of carbon
dioxide to fuels. Catal Today. 2009;148(34):191205.
8. Newsome DS. The water-gas shift reaction. Catal Rev Sci Eng. 1980;2(2)1:275318.
456
9. Armor JN. The multiple roles for catalysis in the production of H-2. Appl Catal A Gen.
1999;176(2):15976.
10. Chinchen GC, Denny PJ, Jennings JR, Spencer MS, Waugh KC. Synthesis of methanol,
1. Catalysts and kinetics. Appl Catal. 1988;36(1):165.
11. Klier K. Methanol synthesis. Adv Catal. 1982;31:243313.
12. Storch H, Golumbic N, Anderson RB. The Fischer-Tropsch and related syntheses. New
York: Wiley; 1951.
13. Iglesia E, Reyes SC, Madon RJ, Soled SL. Selectivity control and catalyst design in the
Fischer-Tropsch synthesis sites, pellets, and reactors. Adv Catal. 1993;39:221302.
14. van der Laan GP, Beenackers AACM. Kinetics and selectivity of the Fischer-Tropsch
synthesis: a literature review. Catal Rev Sci Eng. 1999;41(34):255318.
15. Steynberg AP, Dry ME. Fischer-Tropsch technology. Amsterdam: Elsevier; 2004. (Studies
in surface science and catalysis, volume 152.)
16. Dry ME, Hoogendoorn JC. Technology of the Fischer-Tropsch process. Catal Rev Sci
Eng. 1981;23(12):26578.
17. Biloen P, Sachtler WMH. Mechanism of hydrocarbon synthesis over Fischer-Tropsch catalysts. Adv Catal. 1981;30:165216.
18. Inderwildi OR, Jenkins SJ, King DA. Fischer-Tropsch mechanism revisited: alternative
pathways for the production of higher hydrocarbons from synthesis gas. J Phys Chem
C. 2008;112(5):13057.
19. Ojeda M, Nabar R, Nilekar AU, Ishikawa A, Mavrikakis M, Iglesia E. CO activation
pathways and the mechanism of Fischer-Tropsch synthesis. J Catal. 2010;272(2):28797.
20. Elahifard MR, Perez-Jigato M, Niemantsverdriet JW. Direct versus hydrogen-assisted CO
dissociation on the Fe(100) surface: a DFT study. ChemPhysChem. 2012;13(1):8991.
21. Donnelly TJ, Yates IC, Sattereld CN. Analysis and prediction of product distributions of
the Fischer-Tropsch synthesis. Energy Fuels. 1988;2(6):7349.
22. Steynberg AP, Espinoza RL, Jager B, Vosloo AC. High temperature Fischer-Tropsch synthesis in commercial practice. Appl Catal A Gen. 1999;186(12):4154.
23. Espinoza RL, Steynberg AP, Jager B, Vosloo AC. Low temperature Fischer-Tropsch synthesis from a Sasol perspective. Appl Catal A Gen. 1999;186(12):1326.
24. Geerlings JJC, Wilson JH, Kramer GJ, Kuipers H, Hoek A, Huisman HM. FischerTropsch technology from active site to commercial process. Appl Catal A Gen. 1999;
186(12):2740.
25. Sie ST. Process development and scale up: IV. Case history of the development of a
Fischer-Tropsch synthesis process. Rev Chem Eng. 1998;14(2):10957.
26. Davis BH. Fischer-Tropsch synthesis: relationship between iron catalyst composition and
process variables. Catal Today. 2003;84(12):8398.
27. Niemantsverdriet JW, Van der Kraan AM, Van Dijk WL, Van der Baan HS. Behavior
of metallic iron catalysts during Fischer-Tropsch synthesis studied with Mossbauerspectroscopy, X-ray-diffraction, carbon content determination, and reaction kinetic measurements. J Phys Chem. 1980;84(25):336370.
28. de Smit E, Weckhuysen BM. The renaissance of iron-based Fischer-Tropsch synthesis: on
the multifaceted catalyst deactivation behaviour. Chem Soc Rev. 2008;37(12):275881.
29. Khodakov AY, Chu W, Fongarland P. Advances in the development of novel cobalt
Fischer-Tropsch catalysts for synthesis of long-chain hydrocarbons and clean fuels.
Chem Rev. 2007;107(5):1692744.
30. Tsakoumis NE, Ronning M, Borg O, Rytter E, Holmen A. Deactivation of cobalt based
Fischer-Tropsch catalysts: a review. Catal Today. 2010;154(34):16282.
31. van de Loosdrecht J, Bazhinimaev B, Dalmon JA, et al. Cobalt Fischer-Tropsch synthesis:
deactivation by oxidation? Catalysis Today. 2007;123(14):293302.
5.4.5 References
457
32. Bezemer GL, Bitter JH, Kuipers H, et al. Cobalt particle size effects in the FischerTropsch reaction studied with carbon nanober supported catalysts. J Am Chem Soc.
2006;128(12):395664.
33. Saib AM, Moodley DJ, Ciobica IM, et al. Fundamental understanding of deactivation
and regeneration of cobalt Fischer-Tropsch synthesis catalysts. Catal Today. 2010;154
(34):27182.
34. Saib AM, Borgna A, van de Loosdrecht J, van Berge PJ, Geus JW, Niemantsverdriet JW.
Preparation and characterisation of spherical CO/SiO2 model catalysts with well-dened
nano-sized cobalt crystallites and a comparison of their stability against oxidation with
water. J Catal. 2006;239(2):32639.
35. Fischer F, Tropsch H. The preparation of synthetic oil mixtures (synthol) from carbon
monoxide and hydrogen. Brennstoff Chem. 1923;4:27685.
36. Fischer F, Tropsch H. The synthesis of petroleum at atmospheric pressures from gasication products of coal. Brennstoff Chem. 1926;7:97104.
37. Sabatier P, Senderens JB. Hydrogenation of CO over nickel to produce methane. Hebd
Seances Acad Sci. 1902;134:514.
38. Dry ME. The Fischer-Tropsch process: 19502000. Catalysis Today. 2002;71(34):22741.
39. Oukaci R, Singleton AH, Goodwin JG. Comparison of patented CoF-T catalysts using
xed-bed and slurry bubble column reactors. Appl Catal A Gen. 1999;186(12):12944.
40. van Steen E, Claeys M. Fischer-Tropsch catalysts for the biomass-to-liquid process. Chem
Eng Technol. 2008;31(5):65566.
41. Tristantini D, Logdberg S, Gevert B, Borg O, Holmen A. The effect of synthesis gas composition on the Fischer-Tropsch synthesis over Co/gamma-Al2O3 and Co-Re/gammaAl2O3 catalysts. Fuel Processing Technol. 2007;88(7):6439.
42. Tijmensen MJA, Faaij APC, Hamelinck CN, van Hardeveld MRM. Exploration of the
possibilities for production of Fischer Tropsch liquids and power via biomass gasication.
Biomass Bioenergy. 2002;23(2):12952.
Index
alkenes
production from furans 7273
production from -valerolactone 74, 77
alkyl valerates, for fuel applications 7778
alloying
in catalyst design 177
as single-phase mechanism for negative
electrodes 239
alumina-supported catalysts 116, 299,
331, 345
aluminas, active, synthesis by
precipitation 287
aluminosilicates, microporous see zeolites
Amberlyst catalysts 73, 75, 78
amino acids, production from biomass 98
ammonia
decomposition 267270
production from nitrogen 353, 354,
370373
Anderson-Flory-Schultz (ASF)
distribution 450
annealing
and loss of active sites 435
thermal 181, 182, 338, 340, 341
anode
denition 141
half-cell reaction 154155
Sn 244245
see also electrodes; photoanodes
apparent activation parameter 15
aromatic compounds, production from
lignin 81, 94
Arrhenius equation 253254
Arrhenius plot 255, 263
aspartic acid, production from biomass 98
atomic layer deposition (ALD) 279280,
285, 286
Au
on doped oxide supports 334335
low-dimensional aggregates 332333
460
Index
bioethanol
advantages and disavantages 6364
feedstocks 6465
as fuel additive 6364
net energy balance 64
production 6061, 64
as sustainable biofuel 6366
see also ethanol
biofuels
based on 5-HMF 7981
economic considerations 60, 61, 62
energy storage densities 59
niche solution 1920
novel 7681
production from lignocellulose 6871
biogas
economic considerations 21
production and composition 61
in production of methane 46, 264
biological processes, insufcient for human
energy needs 1819
biomass
boundary conditions and problems
104
as chemical feedstock 20, 8788, 102,
104, 105
classication 8898
combustion for heat and/or power
118119
competition between food and fuel
60, 61
conversion to fossil fuels 6
degradation to C1 building blocks 103
economic considerations 104105
energy density 19
functionalization 103
in future energy systems 48
gasication 109, 118
heating value 109110, 111
high-pressure transformation
reactions 320
hydrodeoxygenation 103
low-value 125
moisture content 110
oxygen content 8788
oxygenates, advantages as fuels 78
potential utilization technologies
7071
Index
461
462
Index
Index
463
cobalt see Co
cocombustion 119
cold storage 3739
combustion
of biomass 118119
energy 119
of methane 13, 321
of methanol 413
of solid fuels 109, 110
CoMo catalysts, sulded, in treatment of
bio-oil 116
compressed air energy storage (CAES)
4142
concentrated solar power (CSP) systems,
sustainable 401
concentration, determination 260261
concentration cells 138, 147148
concentration prole, in catalyst particles
250, 251
continuous xed-bed reactor systems, in situ
studies 321
coprecipitation, in the synthesis of
multicomponent materials 287, 289
cost effectiveness
of catalysis 354
of energy storage 3940
Coulomb-type interactions 332
counter-ion conductivity 235
Cr, as dopant 336
Crutzen, Paul 55
Cu catalysts
active site 428430, 432
drawback 432
lattice defects 428
in methanol production 417, 418,
422, 434, 446
model systems 434
in MSR 431433
in oxygen activation 367
specic Cu surface area 426428
Cu electrodes, in electroreduction of CO2
386, 388
Cu-Ru catalyst 80
Cu-Zn hydroxycarbonates, aging 292293
Cu/ZnO catalysts
deactivation mechanisms 434435
growth on polar surface 332
in methanol synthesis 424
464
Index
Index
electrochemical cells
concept 140141
equilibrium condition 141142,
144145, 148, 236
maximum electric energy 144145
permeable membrane 141
use in energy conversion/storage
applications 227
electrochemical processes
chemical and electrical energy 142144
classical experiment 135136
high theoretical efciency 237
investigations 151
optimization 285
stoichiometry 142143
thermodynamics 236239
electrode potential dependence, of the
oxidation reduction reaction 173
electrode reactions, energetics 138140
electrodes
enhanced conductivity 243
materials 233, 240, 242243
types 137138
electrolysis
binding energies of intermediates
158159
catalysis 159160
energy efciency 161
of water 26, 2728, 135136
electrolytes
aqueous 230
based on ionic liquids 235, 385,
387388, 389
phase states 228
electron microscopy, in atomic-scale
studies 321
electron nuclear double resonance
(ENDOR) 192193, 195, 204205,
208, 357
electron paramagnetic resonance (EPR)
spectroscopy
in catalyst characterization 313314
of hydrogenases 203, 204, 205, 206,
207, 208
of reaction mechanisms 357
of WOC 192, 193, 195
electron transfer, in positive electrodes
234
465
466
Index
Index
467
468
Index
glucose
dehydration to 5-HMF 80
as energy storage molecule 18
hydrogenation 72
glutamic acid 98
glycerides, major agricultural product 94
glycerol
by-product of biodiesel production
60, 66, 67, 99
coproduct in oleochemistry 99
oxidation products 101103
uses 67, 99
gold see Au
grain boundaries
electron trapping 334
in porous photoanodes 394
granulation, of granular catalyst particles
282
graphite, as negative electrode in Li
batteries 232, 233
green chemistry 69
green plants see photosynthesis
greenhouse gas emissions see CO2
emissions
Haber-Bosch process 370
half cells, concept 140, 146
heart pacemaker battery 227
heat storage
emerging applications 38
media 38
systems 3739, 48
traditional methods 37
heat transport 257
heating energy
demand will decrease 48
domestic/district heating 3839
heating value, of biomass 109110, 111
heliostats 401
heme group 182
hemicellulose 61, 62
as feedstock 61, 62, 68, 69, 74, 77, 92
structure 62, 63, 89, 114
waste 73
heteropoly acids 80
hexose fermentation 65
high temperature techniques, in catalyst
preparation 286287
Index
469
470
Index
lead-acid batteries
aqueous electrolyte 230
for grid storage 43
history and importance 228
leaves, articial 380382
Leclanch element 227
Leopoldina report 53, 54
levoglucosan 115
levulinic acid (LA)
cellulosic feedstocks 77
conversion to valeric acid 75
esters as potential fuel candidates
7677
as platform chemical 7375
Li
chemical potential 236237
as dopant 335336
Li batteries
advantages and disadvantages 232
cell voltage 235
energy storage, kinetics 239240
history and development 232233
positive electrode materials 233234
single-phase mechanism 237239
storage modes 237238
thermal runaway 233, 234
transition metal constituent 234
typical structure 235236
Li-doped MgO, in oxidative coupling of
methane 264267
Li-ion batteries 47, 232
Li2MnSiO4 234
LiCoO2 232, 233, 234
LiFePO4 233234, 235, 238, 240, 241
lifestyles, energy-efcient 2
light harvesting
improved 347348
see also photovoltaic cells; solar energy
lignin
in biomass 62
composition 92
as feedstock 9294
as fuel in pulp mills 92, 112
industrial production 92
precursor to coal 17
transformation into potential fuel
compounds 81
lignocellulose
avoids competition with food use
61, 62
in biofuel production 6871
chemical disintegration and
depolymerization 65, 70
chemical structure 6263
composition 89
as feedstock 6263
techno-economic analysis 75
lignosulfonates 92
LiNCM 233, 234
LiPF6, as liquid electrolyte 235
lipids, industrially important types 94
LiPON, solid electrolyte 235
liquid fuels, from CO2 reduction 389390
liquid-solid composite electrolytes 235
lithium see Li
low-temperature Fischer-Tropsch (LTFT)
synthesis 451452
McLuhan, Marshall 57
macrokinetics
and intrinsic kinetics 256257
modeling 1516
magnesium hydrogenation/
dehydrogenation reaction, for heat
storage 38
malachite catalyst
for methanol production 289
zinc content 289, 290, 293, 424426
maleic anhydride 285, 287
manganese see Mn
Mars-van Krevelen type reaction 268, 342
mass balance, determining 260261
mass transport
kinetics 402404
limiting process 257
in situ analysis 316318
medium-voltage grid scale, energy
storage 4344
membrane-electrode assembly (MEA)
168, 169, 170
mesoporous catalysts, shaping strategies
284
mesoporous materials, in treatment of
bio-oil 116
Index
471
472
Index
Index
calculation 156
in catalytic H2 formation 213
decreasing 387388, 389
denition 152, 170
and electrocatalytic performance 158
in electrolysis 152, 154, 384385
in fuel cells 170172
low 157, 187, 194, 198, 212, 389
lower limit 159
in oxygen reduction reaction 173,
175, 179, 364
variation 160
in water oxidation reaction 198, 199,
361
oxidation equivalent accumulation 194
oxidation state, changes under reaction
conditions 319
oxide catalysts, in methanol production
425426
oxide supports
in model systems 329330
three-dimensional deposits 332
oxides
donor-like impurities 334
as negative electrodes 235
ultrathin lms 340344
oxygen
activation 364368
evolution
catalyst (OEC) 391
descriptor and activity volcano
160161
free energy 155158
limits 161
in nature 364367
reduction
electrocatalysis of at fuel cell
cathodes 173182
intermediates 154, 158159,
173174
transition-metal catalysis 174177
sensitivity/tolerance of hydrogenases
209210
oxygen anion conductivity 407
oxygen-evolving complex (OEC) see
water-oxidizing complex (WOC)
oxygen-reduction reaction (ORR) 166,
173182
473
474
Index
photoelectrochemical cells
for CO2 reduction 392397
hydrogen production efciency 392
photoelectronic process, timescale 347
photon STM, of Ag NPs 345, 346
photosynthesis
articial 26, 28, 186187, 197199,
381382, 391395
catalysis 186187, 197199
energy storage process 185, 359
natural 1719, 380382
water oxidation 361
photosystem I (PSI) 381
photosystem II (PSII) 186, 187190, 194,
361362, 381
photovoltaic cells 391
plant materials see biomass
plasmon excitation, photochemical
inuence 346347
platform chemicals, derived from cellulose
and hemicellulose 69, 71
platinum see Pt
point of zero charge (PZC) 296, 337
polar surfaces, growth on 332
polaronic interactions 332
policy-makers, understanding of science
51, 52
politics
role in energy policy 4
and scientic knowledge 51, 52, 56
polymer electrolyte membrane (PEM) fuel
cells, for low-temperature CO2
reduction 390
pore diffusion, limiting process 257
porosity
denition 281
hierarchical 242243, 278
measurement 314
porphyrin compounds 182
precipitates, aging 292
precipitation
investigation by titration 288289
mechanism 288290
in synthetic processes 287293
precursor-support interaction 298
primary batteries, non-rechargeable 227
primary energy, consumption sectors 2, 36
process design, rational 354
Index
pyrolysis
in bio-oil production 112115
of biomass 109
by-products 112113
degradation products 115
reactors 113114
quantum chemical methods
355357
475
Rh catalysts
for cellulose conversion 80
for CO/NO conversion 321
Rh surfaces, in CO atmospheres 311
Rh/Al2O3 catalysts, for partial oxidation of
methane 315
Ru catalysts
for degradation of lignin 81
for water oxidation 363
Ru/C catalysts, for hydrogenation of LA
75
RuO2-modied Pt electrodes, in tandem
cells 391
RuO2/LiRuO2, phase reaction 238
RuRe/C bimetallic catalyst 75
S-state cycle 188, 189, 190
Sabatiers principle 173, 176
salt melt, for heat storage 38
saponication, in biodiesel production 66,
67, 68
scaling relationships 16, 158159, 176
scanning probe techniques, in catalyst
modelling 329
scanning tunneling microscopy (STM)
of Pt/Fe3O4 ultrathin lm 341
in study of model systems 338
Schlgl, Robert 56
science
responsibilities 50
understanding 5051, 52
scientic knowledge, and politics 56
sea level rise 50
secondary battery systems
electrochemical recharging process
227228
history 228
important technologies 229
secondary ion mass spectroscopy (SIMS)
313
semiconductor-liquid junctions (SCLJ)
391
semiconductors, thermodynamic
instability 392
separators, in lithium batteries 236
silica
fumed (SiCl4), manufacture 287
functionalized (SBA-15) 297
476
Index
Index
477
478
Index
function 193195
geometric structure 190192
O-O bond formation 195197
reassembly/repair 194
selective binding 194195
waxes 9798
Weisz-Prater criterion 258259, 264, 265
von Weizscker, Carl Christian 5456
wind energy, in Germany 2122
wind farms, Na-S batteries 43
WO3/GaInP2, corrosion-resistant
material 392
wood
energy distribution and mass
composition 111112
as feedstock for pyrolysis 114, 115
gasication for methanol production
118
X-ray absorption near edge structure
(XANES)
in catalyst characterization 313, 315
in situ studies 316, 321, 323
X-ray absorption spectroscopy (XAS)
for catalyst characterization 357
in situ studies 319, 320, 321, 322, 323
X-ray diffraction (XRD)
for catalyst characterization 313
in combined methods 319, 320, 321
of lattice defects 428
in methane oxidation 319
of YSZ-supported iron oxides 410
X-ray uorescence (XRF) 313
X-ray microscopy, in the submicrometer
region 321
X-ray photoelectron spectroscopy (XPS)
in catalyst characterization 313
of Fe3O4(111)/Pt(111) 338339
for in situ studies 316, 430
surface-sensitive technique 318
X-ray techniques
in situ studies 315, 318, 319, 320
for transient changes 320
see also extended X-ray absorption ne
structure (EXAFS)
xerogels 294, 295
xylose, dehydration 7273
Index
479
zinc see Zn
zirconium see Zr
Zn-air battery 227
Zn-Br ow 231
ZnO/Zn system, for solarthermal
applications 408409
Zr catalysts, in nitrogen xation 372
Zr oxide catalysts, for HMF 78
Zr/Nb mixed oxides, in acrolein
production 101
ZrON catalyst
for ammonia decomposition 267270
synthesis 287