Fuel Cell Systems Explained

Fuel Cell Systems Explained Fuel Cell Systems Explained Third Edition Andrew L. Dicks Griffith University Brisbane, Australia David A. J. Rand CSIRO Energy Melbourne, Australia This edition first published 2018 © 2018 John Wiley & Sons Ltd Edition History John Wiley & Sons Ltd (1e, 2000); John Wiley & Sons Ltd (2e, 2003) All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions. The right of Andrew L. Dicks and David A. J. Rand to be identified as the authors of this work has been asserted in accordance with law. Registered Offices John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK Editorial Office The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com. Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats. Limit of Liability/Disclaimer of Warranty While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. Library of Congress Cataloging‐in‐Publication Data Names: Dicks, Andrew L., author. | Rand, David A. J., 1942– author. Title: Fuel cell systems explained / Andrew L. Dicks, Griffith University, Brisbane, Australia, David A. J. Rand, CSIRO Energy, Melbourne, Australia. Description: Third edition. | Hoboken, NJ, USA : Wiley, [2018] | Includes bibliographical references and index. | Identifiers: LCCN 2017054489 (print) | LCCN 2017058097 (ebook) | ISBN 9781118706978 (pdf ) | ISBN 9781118706961 (epub) | ISBN 9781118613528 (cloth) Subjects: LCSH: Fuel cells. Classification: LCC TK2931 (ebook) | LCC TK2931 .L37 2017 (print) | DDC 621.31/2429–dc23 LC record available at https://lccn.loc.gov/2017054489 Cover design by Wiley Cover images: Top Image: © Iain Masterton/Alamy Stock Photo; Bottom Image: Courtesy of FuelCell Energy, Inc. Set in 10/12pt Warnock by SPi Global, Pondicherry, India Printed in the UK by Bell & Bain Ltd, Glasgow 10 9 8 7 6 5 4 3 2 1 v Contents Brief Biographies xiii Preface xv Acknowledgments xvii Acronyms and Initialisms xix Symbols and Units xxv 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.7.1 1.7.2 1.7.3 1.8 1.9 1.10 2 2.1 2.2 2.3 2.4 2.5 2.5.1 2.5.2 2.5.3 2.5.4 2.6 Introducing Fuel Cells 1 Historical Perspective 1 Fuel‐Cell Basics 7 Electrode Reaction Rates 9 Stack Design 11 Gas Supply and Cooling 14 Principal Technologies 17 Mechanically Rechargeable Batteries and Other Fuel Cells 19 Metal–Air Cells 20 Redox Flow Cells 20 Biological Fuel Cells 23 Balance‐of‐Plant Components 23 Fuel‐Cell Systems: Key Parameters 24 Advantages and Applications 25 Further Reading 26 Efficiency and Open‐Circuit Voltage 27 Open‐Circuit Voltage: Hydrogen Fuel Cell 27 Open‐Circuit Voltage: Other Fuel Cells and Batteries 31 Efficiency and Its Limits 32 Efficiency and Voltage 35 Influence of Pressure and Gas Concentration 36 Nernst Equation 36 Hydrogen Partial Pressure 38 Fuel and Oxidant Utilization 39 System Pressure 39 Summary 40 Further Reading 41 vi Contents 3 Operational Fuel‐Cell Voltages 43 3.1 3.2 3.3 3.4 3.4.1 3.4.2 3.4.3 3.5 3.6 3.7 3.8 3.9 3.10 3.10.1 3.10.2 3.10.3 Fundamental Voltage: Current Behaviour 43 Terminology 44 Fuel‐Cell Irreversibilities 46 Activation Losses 46 The Tafel Equation 46 The Constants in the Tafel Equation 48 Reducing the Activation Overpotential 51 Internal Currents and Fuel Crossover 52 Ohmic Losses 54 Mass‐Transport Losses 55 Combining the Irreversibilities 57 The Electrical Double-Layer 58 Techniques for Distinguishing Irreversibilities 60 Cyclic Voltammetry 60 AC Impedance Spectroscopy 61 Current Interruption 65 Further Reading 68 4 Proton‐Exchange Membrane Fuel Cells 69 4.1 4.2 4.2.1 4.2.2 4.2.3 4.2.4 4.2.5 4.3 4.3.1 4.3.2 4.3.2.1 4.3.2.2 4.3.2.3 4.3.2.4 4.3.2.5 4.3.2.6 4.3.3 4.3.4 4.3.5 4.4 4.4.1 4.4.2 4.4.3 4.4.4 4.4.5 4.4.6 4.5 Overview 69 Polymer Electrolyte: Principles of Operation 72 Perfluorinated Sulfonic Acid Membrane 72 Modified Perfluorinated Sulfonic Acid Membranes 76 Alternative Sulfonated and Non‐Sulfonated Membranes 77 Acid–Base Complexes and Ionic Liquids 79 High‐Temperature Proton Conductors 80 Electrodes and Electrode Structure 81 Catalyst Layers: Platinum‐Based Catalysts 82 Catalyst Layers: Alternative Catalysts for Oxygen Reduction 85 Macrocyclics 86 Chalcogenides 87 Conductive Polymers 87 Nitrides 87 Functionalized Carbons 87 Heteropolyacids 88 Catalyst Layer: Negative Electrode 88 Catalyst Durability 88 Gas‐Diffusion Layer 89 Water Management 92 Hydration and Water Movement 92 Air Flow and Water Evaporation 94 Air Humidity 96 Self‐Humidified Cells 98 External Humidification: Principles 100 External Humidification: Methods 102 Cooling and Air Supply 104 Contents 4.5.1 4.5.2 4.5.3 4.6 4.6.1 4.6.2 4.6.3 4.6.4 4.6.5 4.6.6 4.7 4.7.1 4.7.2 4.7.2.1 4.7.3 4.8 4.8.1 4.8.2 4.9 4.9.1 4.9.2 4.9.3 4.10 4.10.1 4.10.2 4.10.3 4.11 Cooling with Cathode Air Supply 104 Separate Reactant and Cooling Air 104 Water Cooling 105 Stack Construction Methods 107 Introduction 107 Carbon Bipolar Plates 107 Metal Bipolar Plates 109 Flow‐Field Patterns 110 Other Topologies 112 Mixed Reactant Cells 114 Operating Pressure 115 Technical Issues 115 Benefits of High Operating Pressures 117 Current 117 Other Factors 120 Fuel Types 120 Reformed Hydrocarbons 120 Alcohols and Other Liquid Fuels 121 Practical and Commercial Systems 122 Small‐Scale Systems 122 Medium‐Scale for Stationary Applications 123 Transport System Applications 125 System Design, Stack Lifetime and Related Issues 129 Membrane Degradation 129 Catalyst Degradation 129 System Control 129 Unitized Regenerative Fuel Cells 130 Further Reading 132 5 Alkaline Fuel Cells 135 Principles of Operation 135 System Designs 137 Circulating Electrolyte Solution 137 Static Electrolyte Solution 140 Dissolved Fuel 142 Anion‐Exchange Membrane Fuel Cells 144 Electrodes 147 Sintered Nickel Powder 147 Raney Metals 147 Rolled Carbon 148 Catalysts 150 Stack Designs 151 Monopolar and Bipolar 151 Other Stack Designs 152 Operating Pressure and Temperature 152 Opportunities and Challenges 155 Further Reading 156 5.1 5.2 5.2.1 5.2.2 5.2.3 5.2.4 5.3 5.3.1 5.3.2 5.3.3 5.3.4 5.4 5.4.1 5.4.2 5.5 5.6 vii viii Contents 6 Direct Liquid Fuel Cells 157 6.1 6.1.1 6.1.2 6.1.3 6.1.4 6.1.5 6.1.6 6.1.7 6.1.8 6.1.9 6.1.10 6.1.11 6.2 6.2.1 6.2.2 6.2.3 6.2.4 6.3 6.4 6.4.1 6.4.2 6.4.3 6.5 6.5.1 6.5.2 6.6 6.6.1 6.6.2 6.7 Direct Methanol Fuel Cells 157 Principles of Operation 160 Electrode Reactions with a Proton‐Exchange Membrane Electrolyte 160 Electrode Reactions with an Alkaline Electrolyte 162 Anode Catalysts 162 Cathode Catalysts 163 System Designs 164 Fuel Crossover 165 Mitigating Fuel Crossover: Standard Techniques 166 Mitigating Fuel Crossover: Prospective Techniques 167 Methanol Production 168 Methanol Safety and Storage 168 Direct Ethanol Fuel Cells 169 Principles of Operation 170 Ethanol Oxidation, Catalyst and Reaction Mechanism 170 Low‐Temperature Operation: Performance and Challenges 172 High‐Temperature Direct Ethanol Fuel Cells 173 Direct Propanol Fuel Cells 173 Direct Ethylene Glycol Fuel Cells 174 Principles of Operation 174 Ethylene Glycol: Anodic Oxidation 175 Cell Performance 176 Formic Acid Fuel Cells 176 Formic Acid: Anodic Oxidation 177 Cell Performance 177 Borohydride Fuel Cells 178 Anode Catalysts 180 Challenges 180 Application of Direct Liquid Fuel Cells 182 Further Reading 184 7 Phosphoric Acid Fuel Cells 187 7.1 7.2 7.2.1 7.2.2 7.2.3 7.2.3.1 7.2.3.2 7.2.3.3 7.3 7.3.1 7.3.2 7.3.3 7.3.4 7.4 High‐Temperature Fuel‐Cell Systems 187 System Design 188 Fuel Processing 188 Fuel Utilization 189 Heat‐Exchangers 192 Designs 193 Exergy Analysis 193 Pinch Analysis 194 Principles of Operation 196 Electrolyte 196 Electrodes and Catalysts 198 Stack Construction 199 Stack Cooling and Manifolding 200 Performance 201 Contents 7.4.1 7.4.2 7.4.3 7.4.4 7.5 Operating Pressure 202 Operating Temperature 202 Effects of Fuel and Oxidant Composition 203 Effects of Carbon Monoxide and Sulfur 204 Technological Developments 204 Further Reading 206 8 Molten Carbonate Fuel Cells 207 8.1 8.2 8.2.1 8.2.2 8.2.3 8.2.4 8.3 8.3.1 8.3.2 8.4 8.4.1 8.4.2 8.5 8.5.1 8.5.2 8.5.3 8.5.4 8.6 8.7 8.8 Principles of Operation 207 Cell Components 210 Electrolyte 211 Anode 213 Cathode 214 Non‐Porous Components 215 Stack Configuration and Sealing 215 Manifolding 216 Internal and External Reforming 218 Performance 220 Influence of Pressure 220 Influence of Temperature 222 Practical Systems 223 Fuel Cell Energy (USA) 223 Fuel Cell Energy Solutions (Europe) 225 Facilities in Japan 228 Facilities in South Korea 228 Future Research and Development 229 Hydrogen Production and Carbon Dioxide Separation 230 Direct Carbon Fuel Cell 231 Further Reading 234 9 Solid Oxide Fuel Cells 235 Principles of Operation 235 High‐Temperature (HT) Cells 235 Low‐Temperature (IT) Cells 237 Components 238 Zirconia Electrolyte for HT‐Cells 238 Electrolytes for IT‐Cells 240 Ceria 240 Perovskites 241 Other Materials 243 Anodes 243 Nickel‐YSZ 243 Cathode 245 Mixed Ionic–Electronic Conductor Anode Cathode 247 Interconnect Material 247 Sealing Materials 248 9.1 9.1.1 9.1.2 9.2 9.2.1 9.2.2 9.2.2.1 9.2.2.2 9.2.2.3 9.2.3 9.2.3.1 9.2.3.2 9.2.3.3 9.2.4 9.2.5 9.2.6 246 ix x Contents 9.3 9.3.1 9.3.2 9.4 9.5 9.5.1 9.5.2 9.6 Practical Design and Stacking Arrangements 249 Tubular Design 249 Planar Design 251 Performance 253 Developmental and Commercial Systems 254 Tubular SOFCs 255 Planar SOFCs 256 Combined‐Cycle and Other Systems 258 Further Reading 260 10 Fuels for Fuel Cells 263 Introduction 263 Fossil Fuels 266 Petroleum 266 Petroleum from Tar Sands, Oil Shales and Gas Hydrates 268 Coal and Coal Gases 268 Natural Gas and Coal‐Bed Methane (Coal‐Seam Gas) 270 Biofuels 272 Basics of Fuel Processing 275 Fuel‐Cell Requirements 275 Desulfurization 275 Steam Reforming 277 Carbon Formation and Pre‐Reforming 280 Internal Reforming 281 Indirect Internal Reforming (IIR) 283 Direct Internal Reforming (DIR) 283 Direct Hydrocarbon Oxidation 284 Partial Oxidation and Autothermal Reforming 285 Solar–Thermal Reforming 286 Sorbent‐Enhanced Reforming 287 Hydrogen Generation by Pyrolysis or Thermal Cracking of Hydrocarbons 289 Further Fuel Processing: Removal of Carbon Monoxide 290 Membrane Developments for Gas Separation 293 Non‐Porous Metal Membranes 293 Non‐Porous Ceramic Membranes 294 Porous Membranes 294 Oxygen Separation 295 Practical Fuel Processing: Stationary Applications 295 Industrial Steam Reforming 295 Fuel‐Cell Plants Operating with Steam Reforming of Natural Gas 296 Reformer and Partial Oxidation Designs 298 Conventional Packed‐Bed Catalytic Reactors 298 Compact Reformers 299 Plate Reformers and Microchannel Reformers 300 Membrane Reactors 301 Non‐Catalytic Partial Oxidation Reactors 302 10.1 10.2 10.2.1 10.2.2 10.2.3 10.2.4 10.3 10.4 10.4.1 10.4.2 10.4.3 10.4.4 10.4.5 10.4.5.1 10.4.5.2 10.4.6 10.4.7 10.4.8 10.4.9 10.4.10 10.4.11 10.5 10.5.1 10.5.2 10.5.3 10.5.4 10.6 10.6.1 10.6.2 10.6.3 10.6.3.1 10.6.3.2 10.6.3.3 10.6.3.4 10.6.3.5 Contents 10.6.3.6 10.7 10.8 10.8.1 10.8.2 10.8.3 10.8.4 10.9 10.9.1 10.9.2 10.10 10.10.1 10.10.2 10.10.3 10.10.4 Catalytic Partial Oxidation Reactors 303 Practical Fuel Processing: Mobile Applications 304 Electrolysers 305 Operation of Electrolysers 305 Applications 307 Electrolyser Efficiency 312 Photoelectrochemical Cells 312 Thermochemical Hydrogen Production and Chemical Looping 314 Thermochemical Cycles 314 Chemical Looping 317 Biological Production of Hydrogen 318 Introduction 318 Photosynthesis and Water Splitting 318 Biological Shift Reaction 320 Digestion Processes 320 Further Reading 321 11 Hydrogen Storage 11.1 11.2 11.3 11.3.1 11.3.2 11.3.3 11.3.4 11.4 11.5 11.6 11.6.1 11.6.2 11.6.3 11.7 11.7.1 11.7.2 11.8 11.9 323 Strategic Considerations 323 Safety 326 Compressed Hydrogen 327 Storage Cylinders 327 Storage Efficiency 329 Costs of Stored Hydrogen 330 Safety Aspects 330 Liquid Hydrogen 331 Reversible Metal Hydrides 333 Simple Hydrogen‐Bearing Chemicals 338 Organic Chemicals 338 Alkali Metal Hydrides 339 Ammonia, Amines and Ammonia Borane 340 Complex Chemical Hydrides 341 Alanates 342 Borohydrides 342 Nanostructured Materials 344 Evaluation of Hydrogen Storage Methods 347 Further Reading 350 12 The Complete System and Its Future 12.1 12.1.1 12.1.1.1 12.1.1.2 12.1.1.3 12.1.1.4 12.1.2 12.1.3 351 Mechanical Balance‐of‐Plant Components 351 Compressors 351 Efficiency 354 Power 356 Performance Charts 356 Selection 359 Turbines 361 Ejector Circulators 362 xi xii Contents 12.1.4 12.1.5 12.2 12.2.1 12.2.2 12.2.3 12.2.4 12.2.4.1 12.2.4.2 12.2.5 12.2.6 12.3 12.4 12.4.1 12.4.2 12.4.3 12.4.4 12.4.5 12.5 12.5.1 12.5.2 12.6 Fans and Blowers 363 Pumps 364 Power Electronics 365 DC Regulators (Converters) and Electronic Switches 366 Step‐Down Regulators 368 Step‐Up Regulators 370 Inverters 371 Single Phase 372 Three Phase 376 Fuel‐Cell Interface and Grid Connection Issues 378 Power Factor and Power Factor Correction 378 Hybrid Fuel‐Cell + Battery Systems 380 Analysis of Fuel‐Cell Systems 384 Well‐to‐Wheels Analysis 385 Power‐Train Analysis 387 Life‐Cycle Assessment 388 Process Modelling 389 Further Modelling 392 Commercial Reality 394 Back to Basics 394 Commercial Progress 395 Future Prospects: The Crystal Ball Remains Cloudy 397 Further Reading 399 Appendix 1 A1.1 A1.2 Appendix 2 A2.1 A2.2 A2.3 A2.4 A2.5 A2.6 Useful Fuel‐Cell Equations 405 Introduction 405 Oxygen and Air Usage 406 Exit Air Flow Rate 407 Hydrogen Usage 407 Rate of Water Production 408 Heat Production 409 Appendix 3 A3.1 A3.2 Calculations of the Change in Molar Gibbs Free Energy 401 Hydrogen Fuel Cell 401 Carbon Monoxide Fuel Cell 403 Calculation of Power Required by Air Compressor and Power Recoverable by Turbine in Fuel‐Cell Exhaust 411 Power Required by Air Compressor 411 Power Recoverable from Fuel‐Cell Exhaust with a Turbine Glossary of Terms 415 Index 437 412 xiii Brief Biographies Andrew L. Dicks Andrew L. Dicks, PhD, CChem, FRSC, was educated in England and graduated from Loughborough University before starting a career in the corporate laboratories of the UK gas industry. His first research projects focused on heterogeneous catalysts in gas‐making processes, for which he was awarded a doctorate in 1981. In the mid‐1980s, BG appointed Andrew to lead a research effort on fuel cells that was directed predominantly towards molten carbonate and solid oxide systems. The team pioneered the application of process modelling to fuel‐cell systems, especially those that featured internal reforming. This work, which was supported by the European Commission during the 1990s, involved collaboration with leading fuel‐cell developers throughout Europe and North America. In 1994, Andrew was jointly awarded the Sir Henry Jones (London) Medal of the Institution of Gas Engineers and Managers for his studies on high‐temperature systems. He also took an interest in proton‐exchange membrane fuel cells and became the chair of a project at the University of Victoria, British Columbia, in which Ballard Power Systems was the industrial partner. In 2001, he was awarded a Senior Research Fellowship at the University of Queensland, Australia, that enabled further pursuit of his interest in catalysis and the application of nanomaterials in fuel‐ cell systems. Since moving to Australia, he has continued to promote hydrogen and fuel‐cell technology, as director of the CSIRO National Hydrogen Materials Alliance and as a director of the Australian Institute of Energy. He is now consulted on energy and clean technology issues by governments and funding agencies worldwide. xiv Brief Biographies David A. J. Rand David A. J. Rand, AM, BA, MA, PhD, ScD, FTSE, was educated at the University of Cambridge where, after graduation, he conducted research on low‐ temperature fuel cells. In 1969, he joined the Australian government’s CSIRO laboratories in Melbourne. After further exploration of fuel‐cell mechanisms and then electrochemical studies of mineral beneficiation, he formed the CSIRO Novel Battery Technologies Group in the late 1970s and remained its leader until 2003. He was one of the six scientists who established the US‐ based Advanced Lead–Acid Battery Consortium in 1992 and served as its manager in 1994. He is the co‐inventor of the UltraBatteryTM, which finds service in hybrid electric vehicle and renewable energy storage applications. As a chief research scientist, he fulfilled the role of CSIRO’s scientific advisor on hydrogen and renewable energy until his retirement in 2008. He remains active within the organisation as an Honorary Research Fellow and has served as the chief energy scientist of the World Solar Challenge since its inception in 1987. He was awarded the Faraday Medal by the Royal Society of Chemistry (United Kingdom) in 1991, the UNESCO Gaston Planté Medal by the Bulgarian Academy of Sciences in 1996 and the R.H. Stokes Medal by the Royal Australian Chemical Institute in 2006. He was elected a fellow of the Australian Academy of Technological Sciences and Engineering in 1998 and became a member of the Order of Australia in 2013 for service to science and technological development in the field of energy storage. xv Preface Since publication of the first edition of Fuel Cell Systems Explained, three compelling drivers have supported the continuing development of fuel‐cell technology, namely: ● ● ● The need to maintain energy security in an energy‐hungry world. The desire to reduce urban air pollution from vehicles. The mitigation of climate change by lowering anthropogenic emissions of carbon dioxide. New materials for fuel cells, together with improvements in the performance and lifetimes of stacks, are underpinning the emergence of the first truly commercial systems in applications that range from forklift trucks to power sources for mobile phone towers. Leading vehicle manufacturers have embraced the use of electric drivetrains and now see hydrogen fuel cells complementing the new battery technologies that have also emerged over the past few years. After many decades of laboratory development, a global — but fragile — fuel‐cell industry is bringing the first products to market. To assist those who are unfamiliar with fuel‐cell electrochemistry, Chapter 1 of this third edition has been expanded to include a more detailed account of the evolution of the fuel cell and its accompanying terminology. In the following chapters, extensive revision of the preceding publication has removed material that is no longer relevant to the understanding of modern fuel‐cell systems and has also introduced the latest research findings and technological advances. For example, there are now sections devoted to fuel‐cell characterization, new materials for low‐temperature hydrogen and liquid‐fuelled systems, and a review of system commercialization. Separate chapters on fuel processing and hydrogen storage have been introduced to emphasize how hydrogen may gain importance both in future transport systems and in providing the means for storing renewable energy. The objective of each chapter is to encourage the reader to explore the subject in more depth. For this reason, references have been included as footnotes when it is necessary to substantiate or reinforce the text. To stimulate further interest, however, some recommended further reading may be given at the end of a chapter. There are now several books and electronic resources available to engineers and scientists new to fuel‐cell systems. The third edition of Fuel Cell Systems Explained does not intend to compete with specialist texts that can easily be accessed via the Internet. Rather, it is expected that the book will continue to provide an introduction and overview for students and teachers at universities and technical schools and act as xvi Preface a primer for postgraduate researchers who have chosen to enter this field of technology. Indeed, it is hoped that all readers — be they practitioners, researchers and students in electrical, power, chemical and automotive engineering disciplines — will continue to benefit from this essential guide to the principles, design and implementation of fuel‐cell systems. December 2017 Andrew L. Dicks, Brisbane, Australia David A. J. Rand, Melbourne, Australia xvii Acknowledgments As emphasized throughout this publication, the research and development of fuel cells is highly interdisciplinary in that it encompasses many aspects of science and engineering. This fact is reflected in the number and diversity of companies and organizations that have willingly provided advice and information or given permission to use their images in the third edition of Fuel Cell Systems Explained. Accordingly, the authors are indebted to the following contributors: Avantica plc (formerly BG Technology Ltd), UK Ballard Power Systems Inc., USA CNR ITAE, Italy Coregas, Australia Cygnus Atratus, UK Daimler AG, Germany Doosan Fuel Cell, USA Eaton Corporation, USA Forschungszentrum Jülich GmbH, Germany Fuel Cell Energy, USA Horizon Fuel Cells, Singapore Hydrogenics Corporation, Canada Hyundai Motor Company, Australia Pty Ltd Intelligent Energy, UK International Fuel Cells, USA ITM Power, UK Johnsons Matthey plc, UK Kawasaki Heavy Industries, Japan Kyocera, Japan NDC Power, USA Osaka Gas, Japan Proton Energy Systems, USA Proton Motor Systems, GmbH, Germany Redflow Ltd, Australia Serenergy, Denmark Siemens Westinghouse Power Corporation, USA In addition, the authors acknowledge the work of James Larminie, who instigated the first edition of this book, as well as the assistance of others engaged in the advancement xviii Acknowledgments of fuel cells, namely, John Appleby (Texas A&M University, USA), Nigel Brandon and David Hart (Imperial College, UK), John Andrews (RMIT University, Australia), Evan Gray (Griffith University, Australia), Ian Gregg (Consultant, Australia) and Chris Hodrien (University of Warwick, UK). The authors also wish to express their thanks for the support and encouragement given by family, friends and colleagues during the course of this project. xix Acronyms and Initialisms ABPBI AC ADP AEM AEMFC AES AFC AMFC ANL APEMFC APU ASR phosphoric acid doped poly(2,5‐benzimidazole) alternating current adenosine 5’-triphosphate alkaline‐electrolyte membrane alkaline‐electrolyte membrane fuel cell air‐electrode supported alkaline fuel cell anion‐exchange membrane fuel cell Argonne National Laboratory alkaline proton‐exchange membrane fuel cell auxiliary power unit area specific resistance BCN BG BIMEVOX BOP BPS BSF Dutch Fuel Cell Corporation British Gas bismuth metal vanadium oxide (Bi4V2O11) balance-of-plant Ballard Power Systems Boudouard Safety Factor CAN bus CBM CCS CFCL CGO CHP CLC CNR CNT CODH-1 CPE CPO CRG CSG CSIRO Controller Area Network coal‐bed methane carbon capture and storage Ceramic Fuel Cells Ltd cerium–gadolinium oxide (same as GDC) combined heat and power chemical looping combustion Consiglio Nazionale delle Ricerche (Italy) carbon nanotube carbon monoxide dehydrogenase constant phase element catalytic partial oxidation catalytic rich gas coal‐seam gas Commonwealth Scientific and Industrial Research Organisation xx Acronyms and Initialisms CSO CSZ CV CVD cerium‐samarium oxide (same as SDC) calcia‐stabilized zirconia cyclic voltammetry chemical vapour deposition DBFC DC DCFC DEFC DEGFC DFAFC DFT DG DIR DIVRR DLFC DMFC DOE DPFC DPFC(2) DSSC direct borohydride fuel cell direct current direct carbon fuel cell direct ethanol fuel cell direct ethylene glycol fuel cell direct formic acid fuel cell (also formic acid fuel cell, FAFC) density functional theory distributed generator direct internal reforming directly irradiated, volumetric receiver–reactor direct liquid fuel cell direct methanol fuel cell Department of Energy (United States) direct propanol fuel cell direct propan‐2‐ol fuel cell dye‐sensitized solar cell EC ECN EFOY EIS EPFL EU EVD EW evaporatively cooled Energy Research Centre of the Netherlands Energy for You electrochemical impedance spectroscopy Swiss Federal Institute of Technology European Union electrochemical vapour deposition membrane equivalent weight FCE FCES FCV FRA FT Fuel Cell Energy Inc. Fuel Cell Energy Solutions GmbH fuel cell vehicle frequency response analyser Fischer–Tropsch GDC GDL GE GHG GM GPS GTL GTO gadolinium‐doped ceria/gadolinia‐doped ceria (same as CGO) gas-diffusion layer General Electric greenhouse gas General Motors Global Positioning System gas‐to‐liquid gate turn‐off (thyristor) HAZID HAZOP hazard identification hazard and operability study Acronyms and Initialisms HCNG HDS HEMFC HEV HHV HOR HPE hydrogen-compressed natural gas hydrodesulfurization hydroxide‐exchange polymer membrane fuel cell hybrid electric vehicle higher heating value hydrogen oxidation reaction high‐pressure proton‐exchange membrane electrolyser IBFC ICE ICEV IFC IGBT IHI IHP IIR ITM IT‐SOFC IUPAC indirect borohydride fuel cell internal combustion engine internal combustion engine vehicle International Fuel Cells insulated‐gate bipolar transistor Ishikawajima‐Harima Heavy Industries Co., Ltd inner Helmholtz plane indirect internal reforming (also known as ‘integrated reforming’) ion transport membrane, also refers to company ITM Power intermediate‐temperature solid oxide fuel cell International Union of Pure and Applied Chemistry KEPCO KIST Korea Electric Power Corporation Korea Institute of Science and Technology LAMOX LCA lanthanum molybdate (La2Mo2O9) life‐cycle assessment (also known as ‘life‐cycle analysis’ and ‘cradle‐to‐grave analysis’) LCOE levelized cost of electricity LH2 liquid hydrogen LHV lower heating value LNG liquefied natural gas LPG liquefied petroleum gas LSCF lanthanum strontium cobaltite ferrite LSCV strontium‐doped lanthanum vanadate LSGM lanthanum gallate (LaSrGaMgO3) LSM strontium‐doped lanthanum manganite LT‐SOFC low‐temperature solid oxide fuel cell MCFC MCR MEA MEMS METI MFC MFF MHPS MIEC MOF MOSFET molten carbonate fuel cell microchannel reactor membrane–electrode assembly microelectromechanical systems Ministry of Economy, Trade and Industry (Japan) microbial fuel cell mass flow factor Mitsubishi Hitachi Power Systems mixed ionic–electronic conductor (oxides) metal–organic framework metal‐oxide‐semiconductor field‐effect transistor xxi xxii Acronyms and Initialisms MPMDMS MRFC MSW MTBF MWCNT (3‐mercaptopropyl)methyldimethoxysilane mixed‐reactant fuel cell municipal solid waste mean time between failures multiwalled carbon nanotube NADP NASA NCPO NEDO NOMO NTP nicotinamide adenine dinucleotide phosphate National Aeronautics and Space Administration non-catalytic partial oxidation New Energy Development Organization (Japan) Notice of Market Opportunities normal temperature and pressure OCV OEM OER OHP ORR open‐circuit voltage original equipment manufacturer oxygen evolution reaction outer Helmholtz plane oxygen reduction reaction P2G P3MT PAFC PANI PAR PBI PBSS PC PCT PEC PEMFC PET PF PFD PFSA plc POX PPA PPBP Ppy PROX PrOx PSA PTFE PV PWM power‐to‐gas poly(3‐methylthiophene) phosphoric acid fuel cell polyaniline photosynthetically active radiation polybenzimidazole poly(benzylsulfonic acid)siloxane phthalocyanine pressure composition isotherm photoelectrochemical cell proton‐exchange membrane fuel cell (also called ‘polymer electrolyte membrane fuel cell’ and same as SPEFC and SPFC) polyethylene terephthalate power factor, also PFC power factor correction process flow diagram perfluorinated sulfonic acid programmable logic controller partial oxidation polyphosphoric acid poly(1,4‐phenylene), poly(4 phenoxybenzoyl‐1,4‐phenylene) polypyrrole preferential oxidation preferential oxidation reactor pressure swing adsorption polytetrafluoroethylene photovoltaic pulse width modulation QA quaternary ammonium Acronyms and Initialisms RDE RFB RH RHE RRDE RSF rotating disc electrode redox flow battery relative humidity reversible hydrogen electrode rotating ring‐disc electrode rotational speed factor SATP SCG SCT‐CPO SDC SECA SFCM SHE SI SLM SMR SNG SOFC m-SPAEEN-60 SPEEK SPEFC SPFC SPOF STP SWPC standard ambient temperature and pressure simulated coal gas short contact time catalytic partial oxidation samarium‐doped ceria/samaria‐doped ceria (same as CSO) Solid State Energy Conversion Alliance standard cubic foot per minute standard hydrogen electrode International System of Units (French: Système international d’unités) standard litre per minute steam reforming reaction substitute natural gas (also synthetic natural gas) solid oxide fuel cell sulfonated poly(arylene ether ether nitrile) sulfonated polyether ether ketone solid polymer electrolyte fuel cell (same as PEMFC) solid polymer fuel cell (same as PEMFC) single point of failure standard temperature and pressure Siemens Westinghouse Power Corporation TAA THT TMPP TPP TPTZ TTW tetraazaannulene tetrahydrothiophene tetramethoxyphenylporphyrin tetraphenylporphyrin 2, 4, 6‐tris(2‐pyridyl)‐1,3,5‐triazine tank‐to‐wheel UCC UK ULP UPS URFC USA USB UTC UV Union Carbide Corporation United Kingdom unleaded petrol uninterruptible power system; also uninterruptible power supply unitized regenerative fuel cell United States of America universal serial bus United Technologies Corporation ultraviolet WGS WTT WTW water–gas shift well‐to‐tank well‐to‐wheels XPS X‐ray photoelectron spectroscopy xxiii xxv Symbols and Units Subunits d c m μ n A A Multiple units deci centi milli micro nano 10−1 10−2 10−3 10−6 10−9 k M G T P kilo mega giga tera peta 103 106 109 1012 1015 ampere electrode area (cm2), also coefficient in natural logarithm form of the Tafel equation Ah ampere hour a chemical activity; also coefficient in base 10 logarithm form of the Tafel equation ax chemical activity of species x atm atmosphere (=101.325 kPa) B exergy (J) ΔB change in exergy (J) bbl barrel of oil: 35 imperial gallons (159.113 L), or 42 US gallons (158.987 L) bar unit of pressure (=100 kPa) bhp brake horsepower (=745.7 W) C constant in various equations; also coulomb (=1A s), the unit of electric charge °C degree Celsius CP specific heat capacity at constant pressure (J kg−1 K−1) CV specific heat capacity at constant volume (J kg−1 K−1) cP molar heat capacity at constant pressure (J mol−1 K−1) cV molar heat capacity at constant volume (J mol−1 K−1) cm centimetre Dm diffusion coefficient (m2 s−1) d separation of charge layers in a capacitor (mm) E electrode potential (V) E° standard electrode potential (V) Er reversible electrode potential (V) E r standard reversible electrode potential (V) xxvi Symbols and Units EW e− ΔEact F F ft G ΔG ΔG° G f G f g g g gf g f g g H ΔH ΔH° H f H f h h h hf h f h IR e/ IR t/ I i ic il io J K L MFF m ṁ mx mEq mol N 0003367618.INDD 26 (membrane) equivalent weight electron, or the charge on one electron (=1.602 × 10−19 coulombs) activation overpotential (V) farad, unit of electrical capacitance (s4 A2 m−2 kg−1) Faraday constant (=96 458 coulombs mol−1) foot (linear measurement = 305 mm) Gibbs free energy (J) change in Gibbs free energy (J) change in standard Gibbs free energy (J) standard Gibbs free energy of formation (J) change in standard Gibbs free energy of formation (J) molar Gibbs free energy (J mol−1) change in molar Gibbs free energy (J mol−1) change in standard molar Gibbs free energy (J mol−1) change in molar Gibbs free energy of formation (J mol−1) change in standard molar Gibbs free energy of formation (J mol−1) gram acceleration due to gravity (m s−2) enthalpy (J) change in enthalpy (J) change in standard enthalpy (J) standard enthalpy of formation (J) change in standard enthalpy (heat) of formation (J) molar enthalpy (J mol−1) change in molar enthalpy (J mol−1) change in standard molar enthalpy (J mol−1) change molar enthalpy of formation (J mol−1) change in standard molar enthalpy of formation (J mol−1) hour resistive loss in electrolyte (Ω) total resistive loss in electrodes (Ω) current (A) current density, i.e., current per unit area (usually expressed in mA cm−2) crossover current (A) limiting current density (usually expressed in mA cm−2) exchange-current density (usually expressed in mA cm−2) joule (=1 W s) kelvin (used as a measure of absolute temperature) litre mass flow factor (kg s−1 K1/2 bar−1) metre mass flow rate, e.g., of gas (kg s−1) or of a liquid (ml min−1) mass of substance x (g) milliequivalent (weight) (mg L−1) mole, i.e., mass of 6.022 × 1023 elementary units (atoms, molecules, etc.) of a substance newton (unit of force = 1 kg m s−2) 2/24/2018 9:01:39 AM Symbols and Units N NA N‐m3 n ni n x P Pe P° PSAT Px Pa ppb pH ppm R R/ RDS,on RH ® r S S ΔS ΔS° S f ∆S f s ∆s s ∆s f ∆ s f s SLM T TM t t1/2 V Vc Vr Vr ΔVgain ΔVloss vol.% rotor speed of fan (revolutions per minute) Avogadro’s number, 6.022140857 × 1023 normal cubic metre of gas (i.e., that measured at NTP) number of units (electrons, atoms, molecules) involved in a chemical or electrochemical reaction; also number of cells in fuel‐cell stack number of units or moles of species i molar flow rate of species x (mol s−1) pressure (in Pa, or bar) power (W), only used when context is clear that pressure is not under discussion standard pressure (=100 kPa) saturated vapour pressure partial pressure of species x pascal (1 Pa = 1 N m−2 = 9.869 × 10−6 atm) parts per billion numerical scale used to specify the acidity or basicity of an aqueous solution parts per million gas constant (=8.1345 J K−1 mol−1) resistance (Ω) internal resistance of a transistor relative humidity (%); also denoted by the symbol ϕ (v.i.) registered trademark/copyright area specific resistance (Ω cm2) siemens, unit of conductance (Ω−1) entropy (J K−1) change in entropy (J K−1) change in standard entropy (J K−1) standard entropy of formation (J K−1) change in standard entropy of formation (J K−1) molar entropy (J K−1 mol−1) change in molar entropy (J K−1 mol−1) change in standard molar entropy (J mol−1) change in molar entropy of formation (J mol−1) change in standard molar entropy of formation (J mol−1) second standard litre per minute temperature trademark tonne half‐life volt cell voltage (V) reversible cell voltage; also known as ‘open‐circuit voltage’ (V) reversible cell voltage (V) under standard conditions of temperature (298.15 K) and pressure (101.325 kPa) voltage gain (V) voltage loss (V) volume percent xxvii xxviii Symbols and Units W W′ W Wel Wth Wh wt.% xi Z z work done, e.g., in compressing a gas (J) isentropic work (J) watt watt, electrical power watt, thermal power watt‐hour weight percent mole fraction of species i in solution impedance (Ω) number of units (electrons, atoms, molecules) involved in a chemical or electrochemical reaction α γ δm ɛ ξ η η+ η− ηC ηf ϑ λ μf μi μ ϕ ρ ω charge transfer coefficient ratio of the specific heats of a gas CP:CV thickness of proton exchange membrane (cm) electrical permittivity (F m−1) electro‐osmotic coefficient electrode overpotential (V); also efficiency (%) (e.g., of a fuel cell) overpotential at a positive electrode (V) overpotential at a negative electrode (V) isentropic compressor efficiency (%) fuel utilization coefficient (%), a ‘figure of merit’ for DMFCs phase angle stoichiometric ratio fuel utilization coefficient chemical potential of species i (J kg−1 or J mol−1) gas viscosity (centipoise, cP = 0.001 kg m−1 s−1) relative humidity (usually expressed as a percentage); also denoted by RH gas density (kg m−3) humidity ratio, also known as ‘absolute humidity’ and ‘specific humidity’; symbol also used for radial frequency ohm Ω 1 1 Introducing Fuel Cells 1.1 Historical Perspective This book is an introduction to fuel‐cell systems; it aims to provide an understanding of the technology — what it is, how it works and what are its applications. Essentially, a fuel cell can be defined as a device that produces electrical power directly from a fuel via an electrochemical process. In some respects, this operation is similar to that of a conventional battery except that the reactants are stored outside the cell. Therefore, the performance of the device is limited only by the availability of the fuel and oxidant supply and not by the cell design. For this reason, fuel cells are rated by their power output (kW) rather than by their capacity (kWh). Before addressing the technology in depth, it is necessary to understand that by virtue of being electrochemical, fuel cells have both chemical and electrical characteristics. Accordingly, their development has been inextricably linked with the development of electrochemistry as a distinct branch of physical chemistry. At the beginning of the 19th century, it was recognized that an ‘electrochemical cell’ (nowadays, commonly called a ‘battery’) could be made by placing two dissimilar metals in an aqueous salt solution. This discovery was made by Alessandro Volta, the professor of experimental physics at Pavia University, who constructed a pile of alternating discs of copper (or silver or brass) and zinc (or tin) that were separated by pasteboard discs (or ‘any other spongy matter’) soaked in brine. When the top and bottom of the pile were connected by a wire, the assembly delivered, for the first time in history, a more or less steady flow of electricity. Volta introduced the terms ‘electric current’ and ‘electromotive force’, the latter to denote the physical phenomenon that causes the current to flow. In due course, he conveyed his findings in a letter dated 20 March 1800 to Joseph Banks, the then president of the Royal Society. Known as the ‘Volta (or Voltaic) pile’, this was the first ‘primary’ (or non‐rechargeable) power source, as opposed to a ‘secondary’ (or rechargeable) power source. Sir Humphry Davy, who was working at the Royal Institution in London, soon recognized that the Volta pile produces electricity via chemical reactions at the metal– solution interfaces — hydrogen is evolved on the ‘positive’ copper disc, and zinc is consumed at the ‘negative’ disc. Indeed, this recognition of the relationship between chemical and electrical effects prompted Davy to coin the word ‘electrochemical’, from which sprang the science of ‘electrochemistry’. He gave warning that Volta’s work was ‘an alarm bell to experimenters all over Europe’. His prediction was soon to be verified. Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 2 Fuel Cell Systems Explained Volta had sent his letter to the Royal Society in two parts because he anticipated problems with its delivery given that correspondence from Italy had to pass through France, which was then at war with Britain. While waiting for the second part to arrive, Joseph Banks had shown the first few pages to Anthony Carlisle (a fashionable London surgeon) who, in turn, with the assistance William Nicholson (a competent amateur scientist) assembled on 30 April 1800 the first pile to be constructed in England. Almost immediately, on 2 May 1800, the two investigators found that the current from their device when passed through a dilute salt solution via two platinum wires was capable of decomposing water into its constituents — hydrogen at one wire and oxygen at the other. Details of the discovery were published in Nicholson’s own journal in July of the same year. Thus, the new technique of ‘molecular splitting’ — to be coined ‘electrolysis’ by Michael Faraday much later in 1834 and derived from the Greek ‘lysis’ = separation — was demonstrated before Volta’s own account of the pile was made public in September 1800. A schematic representation of the process is shown in Figure 1.1a. It was left to Michael Faraday, Davy’s brilliant student, to identify the mechanisms of the processes that take place within ‘electrolytic’ cells and to give them a quantitative basis. In addition, he was also the guiding force behind the nomenclature that is still in use today. First, Faraday with the assistance of Whitlock Nicholl (his personal physician and accomplished linguist) devised the name ‘electrode’ to describe a solid substance at which an electrochemical reaction occurs and ‘electrolyte’ to describe the chemical compound that provides an electrically conductive medium between electrodes. (Note that in the case of dissolved materials, it is fundamentally incorrect to refer to the ‘electrolyte solution’ as the ‘electrolyte’; nevertheless, the latter terminology has become common practice.) To distinguish between the electrode by which conventional current (i.e., the reverse flow of electrons) enters an electrolytic cell and the electrode by which (a) Electrolysis cell Current (b) Fuel cell – Current + External power source e– e– + – – External load e– + – – Anion Anion + + Cation Cathode e– Cation Anode Electrolyte solution Anode Cathode Electrolyte solution Figure 1.1 Terminology employed in operation of (a) electrolysis cells and (b) fuel cells. Introducing Fuel Cells it leaves, Faraday sought the assistance of the polymath William Whewell, the Master of Trinity College at the University of Cambridge. In a letter dated 24 April 1834, he asked Whewell: ‘Can you help me out to two good names not depending upon the idea of a current in one direction only or upon positive or negative?’ In other words, he wanted terms that would be unaffected by any later change in the convention adopted for the direction of current. Eventually, they settled on calling the positive electrode an ‘anode’ and the negative electrode a ‘cathode’, which were coined from Greek ‘ano‐dos’ (‘upwards’–‘a way’) to represent the path of electrons from the positive electrode to the negative and ‘katho‐dos’ (‘downwards’–‘a way’) to represent the counter direction. For an electrolytic cell, then, the anode is where the current enters the electrolyte and the cathode is where the current leaves the electrolyte. Thus the positive electrode sustains an oxidation (or ‘anodic’) reaction with the liberation of electrons, while a reduction (or ‘cathodic’) reaction takes place at the negative electrode with the uptake of electrons. With use of the Greek neutral present participle ‘ion’ — ‘a moving thing’ — to describe the migrating particles in electrolysis, two further terms were obtained, namely, ‘anion’, i.e., the negatively charged species that goes to the anode against the current (or with the flow of negative charge), and ‘cation’, i.e., the positively charged species that goes to the cathode with the current (or against the flow of negative charge). The operation of an electrolysis cell is shown in Figure 1.1a. It should be noted that the anode–cathode terminology for an ‘electrolytic cell’ applies to a ‘battery under charge’ (secondary system). A fuel cell operates in the reverse manner to an electrolysis cell, i.e., it is a ‘galvanic’ cell that spontaneously produces a voltage (similar to a ‘battery under discharge’). The anode of the electrolysis cell now becomes the cathode and the cathode becomes the anode; see Figure 1.1b. Nevertheless, the directions of the migration of anions and cations with respect to current flow are unchanged such that the positive electrode remains a positive electrode and the negative electrode remains a negative electrode. Thus, in a fuel cell, the fuel is always oxidized at the anode (positive electrode), and the oxidant is reduced at the cathode (negative electrode). There is some debate over who discovered the principle of the fuel cell. In a letter written in December 1838 and published on page 43 of the January issue of the January– June 1839 Volume XIV of The London and Edinburgh Philosophical Magazine and Journal of Science, the German scientist Christian Friedrich Schönbein described his investigations on fluids that were separated from each other by a membrane and connected to a galvanometer by means of platina wires. In the 10th of 14 reported tests, one compartment contained dilute sulfuric acid that held some hydrogen, whereas the other compartment contained dilute sulfuric acid that was exposed to air. Schönbein detected a current and concluded that this was caused ‘by the combination of hydrogen with (the) oxygen (contained dissolved in water)’. This discovery was largely overlooked, however, after the publication of a letter from William Robert Grove, a Welsh lawyer and a scientist at the Royal Institution; see Figure 1.2a. The letter, which was dated 14 December 1838, appeared on page 127 of the February issue of the aforementioned Volume XIV and described his evaluation of electrode and electrolyte materials for use 3 4 Fuel Cell Systems Explained (a) (b) Ox Hy Current Ox Hy Ox Hy Ox Hy Ox Hy Figure 1.2 (a) William Robert Grove (1811–1896) and (b) Grove’s sketch of four cells of his gaseous voltaic battery’ (1842). (Source: https://commons.wikimedia.org/w/index.php?curid=20390734.Used under CC BY‐SA 3.0; https://creativecommons.org/licenses/by‐sa/3.0/.) in batteries. Unfortunately, the order in which these two letters had been written is unknown as Schönbein did not date his letter in full — he gave the month, but not the day. In fact, this chronology is of little importance given the following postscript that Grove had added to his letter in January 1839: ‘I should have pursued these experiments further, and with other metals, but was led aside by some experiments with different solutions separated by a diaphragm and connected by platinum plates; in many of these I have been anticipated.’ In the same postscript, Grove went on to speculate that by connecting such cells in series sufficient voltage could be created to decompose water (by electrolysis). Grove carried out many experiments that demonstrated the principle of the fuel cell. In 1842, he realized that the reaction at the electrodes was dependent on an area of contact between the gas reactant and a layer of liquid that was sufficiently thin to allow the gas to diffuse to the solid electrode (today, this requirement is commonly related to the formation of a ‘three‐phase boundary’ or ‘triple‐point junction’ where gas, electrolyte and electrocatalyst come into simultaneous contact, v.i.). At that time, Grove was the professor of experimental chemistry at the London Institution in Finsbury Circus, and in the same communication he reported the invention of a ‘gaseous voltaic battery’. The device employed two platinized platinum electrodes (to increase the real surface area), and a series of fifty such pairs when semi‐immersed in dilute sulfuric acid solution was found ‘to whirl round’ the needle of a galvanometer, to give a painful shock to five persons joining hands, to give a brilliant spark between charcoal points, and to decompose hydrochloric acid, potassium iodide and acidulated water. An original sketch of four such cells is reproduced in Figure 1.2b. It was also found that 26 cells were the minimum number required to electrolyse water. Grove had indeed realized Introducing Fuel Cells the desire expressed in his 1839 postscript in that he had achieved the beautiful symmetry inherent in the ‘decomposition of water by means of its composition’. The aforementioned apparatus became widely recognized as the first fuel cell and Grove was designated as the ‘Father of the Fuel Cell’. Historically, this title is not fully justified. More accurately, Schönbein should be credited with the discovery of the fuel‐ cell effect in 1838 and Grove with the invention of the first working prototype in 1842. Happily, such accreditations were of little concern to the two scientists and they became close friends. For almost 30 years, they exchanged ideas and developments via a dynamic correspondence and visited each other frequently. It is interesting to note that many latter‐day authors have attributed the introduction of the term ‘fuel cell’ to Ludwig Mond and Charles Langer in their description of a new form of gas battery in 1889. Remarkably, however, there is no mention of ‘fuel cell’ in this communication. Other claims that William W. Jacques, in reporting his experiments to produce electricity from coal, coined the name are equally ill founded. A. J. Allmand in his book The Principles of Applied Electrochemistry, published in 1912, appears to attribute the appellation ‘fuel cell’ to the Nobel Laureate Friedrich Wilhelm Ostwald in 1894. Grove concluded his short paper in 1842 with the following modest entreaty: ‘Many other notions crowd upon my mind, but I have occupied sufficient space and must leave them for the present, hoping that other experimenters will think the subject worth pursuing.’ Unfortunately, however, the invention of the first internal combustion engine to become commercially successful by Jean Joseph Étienne Lenoir in 1859, coupled ironically with Faraday’s earlier discovery of electromagnetic force, diverted the course of electricity generation from electrochemical to electromagnetic methods. As a result, the fuel cell became merely an object of scientific curiosity during much of the next half‐century. Meanwhile, knowledge of electrochemical conversion and storage of energy progressed largely through the development of battery technologies. In 1894, a well‐documented criticism against heat engines was expressed by Friedrich Ostwald, who drew attention to the poor efficiency and polluting problems associated with producing electrical power via the combustion of fossil fuels rather than by direct electrochemical oxidation. A fuel cell is inherently a more thermodynamically efficient device since, unlike an engine in which heat is converted to mechanical work, the cell is not subject to the rules of the Carnot cycle. By virtue of this cycle, the efficiency of a thermal engine is always lowered to a value far below 100%, as determined by the difference between the temperature at which heat is taken in by the working fluid and the temperature at which it is rejected. On this basis, Ostwald advocated that: ‘The path which will help to solve this biggest technical problem of all, this path must be found by the electrochemistry. If we have a galvanic element which directly delivers electrical power from coal and oxygen, […] we are facing a technical revolution that must push back the one of the invention of the steam engine. Imagine how […] the appearance of our industrial places will change! No more smoke, no more soot, no more steam engine, even no more fire, […] since fire will now only be needed for the few processes that cannot be accomplished electrically, and those will daily diminish. […] Until this task shall be tackled, some time will pass by.’ 5 6 Fuel Cell Systems Explained Regrettably, Ostwald was proven to be correct as regards his closing prediction for although attempts were made at the turn of the century to develop fuel cells that could convert coal or carbon into electricity (for instance, the work of William W. Jacques in the United States), the need for an expensive platinum catalyst and its poisoning by carbon monoxide formed during the coal gasification limited cell affordability, usefulness and lifetime. Consequently, interest in such ‘direct carbon fuel cells’ dwindled. In the 1930s, Emil Bauer and H. Preis in Switzerland experimented with solid oxide fuel cells (SOFCs). Given the limitations of solid oxides at that time (i.e., poor electrical conductivity and chemical stability), G.H.J. Broers and J.A.A. Ketelaar in the late 1950s turned to the use of fused salts as electrolytes. The work gave birth to the molten carbonate fuel cell (MCFC), which eventually became one of the main types of fuel cell in commercial production. The renaissance of the fuel‐cell concept in the 20th century can be attributed largely to the work of Englishman F.T. (Tom) Bacon. He was an engineer by profession and thus appreciated the many potential advantages of the fuel cell over both the internal combustion engine and the steam turbine as a source of electrical power. His interest in fuel cells dated as far back as 1932, and he ploughed a lone furrow, with little support or backing, but showed enormous dedication to the challenge of developing practical cells. Early in his career, Bacon elected to study the alkaline‐ electrolyte fuel cell (AFC), which used nickel‐based electrodes, in the belief that platinum‐group electrocatalysts would never become commercially viable. In addition, it was known that the oxygen electrode is more readily reversible in alkaline solution than in acid. This choice of electrolyte and electrodes necessitated operating the cell at moderate temperatures (100–200°C) and high gas pressures. Bacon restricted himself to the use of pure hydrogen and oxygen as reactants. Eventually, in August 1959, he demonstrated the first workable fuel cell — a 40‐cell system that could produce about 6 kW of power, which was sufficient to run a forklift truck and to operate a welding machine as well as a circular saw. A major opportunity to apply fuel cells arose in the early 1960s with the advent of space exploration. In the United States, fuel cells were first employed to provide spacecraft power during the fifth mission of Project Gemini. Batteries had been employed for this purpose in the four earlier flights, as well as in those conducted in the preceding Project Mercury. This switch in technology was undertaken because payload mass is a critical parameter for rocket‐launched satellites, and it was judged that fuel cells, complete with gas supplies, would weigh less than batteries. Moreover, the objective of Project Gemini was to evolve techniques for advanced space travel — notably, the extravehicular activity and the orbital manoeuvres (rendezvous, docking, etc.) required for the moon landing planned in the following Project Apollo. Thus, lunar flights demand a source of power of longer duration than that available from batteries. A proton‐exchange membrane fuel cell (PEMFC) system manufactured by the General Electric Company was adopted for the Gemini missions (two modules, each with a maximum power of about 1 kW), but this was replaced in Project Apollo by an AFC of circulating electrolyte design, as pioneered by Bacon and developed by the Pratt and Whitney Aircraft Company (later the United Technologies Corporation). Both Introducing Fuel Cells types of system were fuelled by hydrogen and oxygen from cryogenic tanks. The AFC could supply 1.5 kW of continuous power, and its in‐flight performance during all 18 Apollo missions was exemplary. In the 1970s, International Fuel Cells (a division of United Technologies Corporation) produced an improved AFC for the Space Shuttle orbiter that delivered eight times more power than the Apollo version and weighed 18 kg less. The system provided all of the electricity, as well as drinking water, when the Space Shuttle was in flight. The successful exploitation of fuel cells in the space programme drove research activity worldwide during the 1970s to develop systems that would generate power with high efficiency and low emissions for terrestrial applications. Research was stimulated further by the hiatus in the global oil supply in 1974. What followed was the emergence of various national initiatives on fuel‐cell development. In the United States, demonstrations of phosphoric acid fuel cell (PAFC) technology by the American Gas Association led to a Notice of Market Opportunities (NOMO) initiative. This activity, in turn, renewed interest in the MCFC by US researchers, and in the mid‐1980s, national research and development programmes were established in Japan and Europe. Renewed interest in the PEMFC was championed in the late 1980s by Geoffrey Ballard, a Canadian pioneer, who saw the potential for the technology to replace internal combustion engines. Since then, this system has been the subject of much advancement for a variety of applications, so much so that it merits two chapters in this book. 1.2 Fuel‐Cell Basics To understand how the reaction between hydrogen and oxygen produces an electric current, and where the electrons are released, it is necessary to consider the reaction that takes place at each electrode. The reactions vary for different types of fuel cell, but it is convenient to start with a cell based around an acid electrolyte, not only because this system was used by Grove but also because it is the simplest and still the most chosen for commercial applications. At the anode of an acid fuel cell, hydrogen is oxidized and thereby releases electrons and creates H+ ions, as expressed by: 2H 2 4H 4e (1.1) This reaction also releases energy in the form of heat. At the cathode, oxygen reacts with electrons taken from the electrode, and H+ ions from the electrolyte, to form water, i.e., O2 4 e 4H 2H 2 O (1.2) Thus the overall cell reaction is: 2H 2 O 2 2H2O heat (1.3) 7 8 Fuel Cell Systems Explained Clearly, for both the electrode reactions to proceed continuously, electrons produced at the negative electrode must pass through an electrical circuit to the positive. Also, H+ ions must pass through the electrolyte solution — an acid is a fluid with free H+ ions and so serves this purpose very well. Certain polymers and ceramic materials can also be made to contain mobile H+ ions. These materials are commonly called ‘proton‐ exchange membranes’, as an H+ ion is also known as a proton. The PEMFC is examined in detail in Chapter 4. The cell reaction (1.3) shows that two hydrogen molecules will be needed for each oxygen molecule if the system is to be kept in balance. The operating principle is illustrated in Figure 1.3. In a fuel cell with an alkaline electrolyte (AFC), the overall reaction of hydrogen oxidation is the same, but the reactions at each electrode are different. In an alkaline solution, hydroxyl (OH−) ions are available and mobile. At the anode, these ions react with hydrogen to release electrons and energy (heat) together with the production of water: 2H2 4OH 4 H2O 4 e (1.4) At the cathode, oxygen reacts with electrons taken from the electrode, and water in the electrolyte and thereby forms new OH− ions: O2 4 e 2H 2 O 4OH (1.5) Comparing equations (1.4) and (1.5) shows that, as with an acid electrolyte, twice as much hydrogen is required compared with oxygen. The operating principle of the AFC is presented in Figure 1.4. There are many other types of fuel cell, each distinguished by its electrolyte and the reactions that take place on the electrodes. The different systems are described in detail in the following chapters. Hydrogen fuel – Anode 4H+ 2H2 + 4e– Load e.g., electric motor H+ Ions through electrolyte + Cathode O2 + 4e– + 4H+ Oxygen, usually from the air 2H2O Electrons flow round the external circuit Figure 1.3 Electrode reactions and charge flow for fuel cell with an acid electrolyte. Note that although the negative electrons flow from the anode to cathode, the ‘conventional positive current’ flows from cathode to anode. Introducing Fuel Cells Hydrogen fuel – Anode 2H2 + 4OH– 4H2O + 4e– Load e.g., electric motor OH– Ions through electrolyte + Cathode O2 + 4e– + 4OH– 2H2O Electrons flow round the external circuit Oxygen, usually from the air Figure 1.4 Electrode reactions and charge flow for a fuel cell with an alkaline electrolyte. Electrons flow from negative anode to positive cathode, but ‘conventional positive current’ flows from cathode to anode. 1.3 Electrode Reaction Rates The oxidation of hydrogen at the negative electrode liberates chemical energy. It does not follow, however, that the reaction proceeds at an unlimited rate; rather, it has the ‘classical’ energy form of most chemical reactions, as shown in Figure 1.5. The schematic represents the fact that some energy must be used to excite the atoms or molecules sufficiently to start the chemical reaction — the so‐called ‘activation energy’. This energy can be in the form of heat, electromagnetic radiation or electrical energy. In visual terms, the activation energy helps the reactant to overcome an ‘energy hill’, Energy Activation energy Energy released Stage of reaction Figure 1.5 Classical energy diagram for a simple exothermic chemical reaction. 9 10 Fuel Cell Systems Explained and once the reaction starts, everything rolls downhill. Thus, if the probability of an atom or molecule having sufficient energy is low, then the reaction will only proceed slowly. This is indeed the case for fuel‐cell reactions, unless very high temperatures are employed. The three main ways of dealing with the slow reaction rates are to (i) use catalysts, (ii) raise the temperature and (iii) increase the electrode area. Whereas the first two options can be applied to any chemical reaction, the electrode area has a special significance for electrochemical cells. The electrochemical reactions take place at the location where the gas molecules (hydrogen or oxygen) meet the solid electrode and the electrolyte (whether solid or liquid). The point at which this occurs is often referred to as the ‘three‐phase boundary/junction’ or the ‘triple‐phase boundary/ junction’ (v.s.). Clearly, the rate at which either electrode reaction proceeds will be proportional to the area of the respective electrode. Indeed, electrode area is such an important issue that the performance of fuel cells is usually quoted in terms of the current per cm2. Nevertheless, the geometric area (length × width) is not the only issue. The electrode is made highly porous so as to provide a great increase in the ‘effective’ surface area for the electrochemical reactions. The surface area of electrodes in modern fuel cells, such as that shown in Figure 1.6, can be two to three orders of magnitude greater than the geometric area. The electrodes may also have to incorporate a catalyst and endure high temperatures in a corrosive environment; catalysts are discussed in Chapter 3. Figure 1.6 Transmission electron microscope image of a fuel‐cell catalyst. The black spots are the catalyst particles that are finely divided over a carbon support. The structure clearly has a large surface area. (Source: Courtesy of Johnson Matthey Plc.) 75 nm Introducing Fuel Cells 1.4 Stack Design Because a fuel cell functions at a low voltage (i.e., well below 1 V), it is customary to build up the voltage to the desired level by electrically connecting cells in series to form a ‘stack’. There are a number of different designs of fuel cell, but in each case the unit cell has certain components in common. These are as follows: ● ● ● ● ● An electrolyte medium that conducts ions. This may be a porous solid that contains a liquid electrolyte (acid, alkali or fused salt) or a thin solid membrane that may be a polymer or a ceramic. The membrane must be an electronic insulator as well as a good ionic conductor and must be stable under both strong oxidizing and strong reducing conditions. A negative fuel electrode (anode) that incorporates an electrocatalyst, which is dispersed on an electronically conducting material. The electrode is fabricated so that the electrocatalyst, the electrolyte and the fuel come into simultaneous contact at a three‐phase boundary (v.s.). A positive electrode (cathode), also with a triple‐point electrocatalyst, at which the incoming oxygen (either alone or in air) is reduced by uptake of electrons from the external circuit. A means of electrically connecting individual cells together. The design of interconnector depends on the geometry adopted for the cells. Seals that keep the gases apart and also prevent cell‐to‐cell seepage of liquid electrolyte, which otherwise would give rise to partial short-circuits. A stack also has current-collectors that are located at the two ends of the stack and are connected by end‐plate assemblies. Historically, the flat plate is by far the preferred geometry for fuel cells, and one way of assembling such cells in series is to connect the edge of each negative electrode to the positive of the next cell through the string, as illustrated in Figure 1.7. (For simplicity, the diagram ignores the difficulty of supplying gas to the electrodes.) The problem with this method, however, is that the electrons have to flow across the face of the electrode to the current collection point at the edge. The electrodes might be quite good conductors, but if each cell is only operating at about 0.7 V, even a small voltage drop can be significant. Consequently, this type of stack design is not used unless the current flows are very low, the electrodes are particularly good conductors and/or the dimensions of the stack are small. A much better method of cell interconnection for planar fuel cells is to use a ‘bipolar plate’. This is an electrically conducting plate that contacts the surfaces of the positive electrode of one cell and the negative electrode of the next cell (hence the term ‘bipolar’). At the same time, the bipolar plate serves as a means of feeding oxygen to the negative anode and fuel gas to the positive cathode of the adjacent cells. This is achieved by having channels machined or moulded on either side of the plate along which the gases can flow and the products, i.e., pure water in the case of hydrogen fuel, can exit. Various designs of channel geometry have been proposed to maximize the access of gases and the removal of water, e.g., pin‐type, series–parallel, serpentine, integrated and interdigitated flow-fields. The different types are described in later chapters when considering the stacking arrangement of each type of fuel cell. The arrangement of the 11 12 Fuel Cell Systems Explained Load Oxygen fed to each cathode Hydrogen fed to each anode Cathode Electrolyte Anode For reactions in this part the electrons have to pass all along the face of the electrode Figure 1.7 Simple edge connection of three‐planar fuel cells in series. When the electrolyte is a membrane, the cathode–electrolyte–anode unit is generally known as a membrane–electrode assembly (MEA). channels (also known as the ‘flow-field’) leads the bipolar plate to be also known as the flow‐field plate. Bipolar plates must also be relatively impermeable to gases, sufficiently strong to withstand stack assembly and easily mass produced. They are made of a good electronic conductor such as graphite or stainless steel. For transport applications, low weight and low volume are essential. The method of connecting two plates to a single cell is illustrated in Figure 1.8; the respective gases are supplied orthogonally. To connect several cells in series, anode–electrolyte–cathode assemblies have to be prepared. These are then ‘stacked’ together with bipolar plates placed between each pair of cells. In the particular arrangement shown in Figure 1.9, the stack has vertical channels for feeding hydrogen over the anodes and horizontal channels for feeding oxygen (or air) over the cathodes. The result is a solid block, in which the electric current passes efficiently more or less straight through the cells, rather than over the surface of each electrode one after the other. The electrodes and electrolytes are also well supported, and the whole structure is clamped together to give a strong and robust device. Although simple in principle, the design of the bipolar plate has a significant effect on fuel‐cell performance. If the electrical connection between cells is to be optimized, then the area of contact points should be as large as possible, but this would mitigate good gas flow over the electrodes. If the contact points have to be small, at least they should be frequent. This may render the plate more complex, difficult and expensive to manufacture, as well as fragile. Ideally, bipolar plates should be as thin as possible so as to minimize both the electrical resistance between individual cells and the stack size. On the other hand, such an Introducing Fuel Cells Anode Electrolyte Cathode Hydrogen fed along these channels Negative connection Air or oxygen fed to cathode Positive connection Figure 1.8 Single cell with end-plates for collecting current from the whole face of the adjacent electrode and applying gases to each electrode. Hydrogen fed along these vertical channels over the anodes Negative connection Positive connection Air or oxygen fed over the cathodes through these channels Figure 1.9 A three‐cell stack showing how bipolar plates connect the anode of one cell to the cathode of its neighbour. approach would narrow the gas channels and thereby place greater demands on the pumps for supplying gases. High rates of flow are sometimes required, especially when using air instead of pure oxygen at the positive electrode. For low‐temperature fuel cells, the circulating air has to evaporate and carry away the product water. Moreover, in many 13 14 Fuel Cell Systems Explained cases, additional channels have to pass through the bipolar plate to carry a cooling fluid. Some further challenges for the bipolar plate are considered in the next section. 1.5 Gas Supply and Cooling The arrangement given in Figure 1.9 has been simplified to show the basic principle of the bipolar plate. In practice, however, the twin problems of gas supply and preventing leaks mean that the design is somewhat more complex. Because the electrodes must be porous (to permit the access of gas), they allow leakage of the gas through their edges. Consequently, the edges must be sealed. Sometimes this is done by making the electrolyte compartment slightly larger than one, or both, of the electrodes and fitting a gasket around each electrode, as presented in Figure 1.10. Such assemblies can then be made into a stack in which the fuel and oxygen can then be supplied to the electrodes using the external manifolds as shown disassembled in Figure 1.11. With this arrangement, the hydrogen should only come into contact with the anodes as it is fed vertically through the fuel‐cell stack. Similarly, the oxygen (or air) fed horizontally through the stack should only contact the cathodes and certainly not the edges of the anodes. Such would not be the case for the basic design illustrated in Figure 1.9. The externally manifolded design suffers from two major disadvantages. The first is that it is difficult to cool the stack. Fuel cells are far from 100% efficiency, and considerable quantities of heat are generated, as well as electrical power. In practice, the cells in this type of stack have to be cooled by the reactant air passing over the positive electrodes. This means that air has to be supplied at a higher rate than that demanded by the cell chemistry — sometimes the flow is sufficient to cool the cell, but it is wasteful of energy. The second disadvantage of external manifolding is that there is uneven pressure over the gasket round the edge of the electrodes, i.e., at the points where there Edge sealing gasket Electrolyte Edge sealing gasket Anode Assembly Cathode Figure 1.10 The construction of cathode–electrolyte–anode units with edge seals that prevent the gases leaking in or out through the edges of the porous electrodes. Introducing Fuel Cells Manifolds Cathode – electrolyte – anode assemblies Figure 1.11 Three‐cell stack, with external manifolds. Unlike the stack shown Figure 1.9, the electrodes now have edge seals. is a channel and the gasket is not pressed firmly onto the electrode. This increases the probability of leakage of the reactant gases. ‘Internal manifolding’ is a more common stack arrangement and requires a more complex design of bipolar plate, such as that displayed schematically in Figure 1.12. In this arrangement, the plates are made larger relative to the electrodes and have extra channels running through the stack for the delivery of fuel and oxygen to the electrodes. Holes are carefully positioned to feed the reactants into the channels that run over the surface of the electrodes. Reactant gases are fed in at the ends of the stack where the respective positive and negative electrical connections are also made. An example of a commercial fuel‐cell stack is shown in Figure 1.13. A stack with internal manifolding can be cooled in various ways. The most practical method is to circulate a liquid coolant through electrically conductive metal plates that are inserted between groups of cells. In this passive approach, the heat within the plane of the plate must be conducted out to one or more of the edges of the plate for transfer to a heat-exchanger external to the fuel‐cell stack. Alternatively the bipolar plates themselves can be made thicker and machined to incorporate extra channels that allow passage of cooling air or water. The preferred cooling method varies greatly with the type of fuel cell and is addressed in later chapters. From the foregoing discussion, it should be apparent that the bipolar plate is a key component of a fuel‐cell stack. As well as being a fairly intricate item to manufacture, the choice of material for its construction raises issues. For low‐temperature fuel 15 16 Fuel Cell Systems Explained Air supplied through here Air removed through here Hydrogen removed through here Channel for distributing air over cathode Hydrogen supplied through here Channel for supplying hydrogen to surface of anode Figure 1.12 Internal manifolding. A more complex bipolar plate allows reactant gases to be fed to electrodes through internal tubes. (Source: Courtesy of Ballard Power Systems.) Figure 1.13 A 96‐cell, water‐cooled PEMFC stack that produces up to 8.4 kW and weighs 1.4 kg. (Source: Courtesy of Proton Motor GmbH.) cells, graphite was one of the first materials to be employed, but it is difficult to work and brittle and, consequently, has now largely been replaced by various carbon composite materials. Stainless steel can also be used, but it will corrode in some types of fuel cell. Ceramic materials have found application in fuel cells that operate at high temperatures. The bipolar plate nearly always is a major contributor to the capital cost of a fuel cell. Introducing Fuel Cells 1.6 Principal Technologies Setting aside practical issues such as manufacturing and materials costs, the two fundamental technical problems with fuel cells are: ● ● The slow reaction rates, particularly for the oxygen reduction reaction, which lead to low levels of current and power. The fact that hydrogen is not a readily available fuel1. To address these problems, many different types of fuel cell have been developed and tested. The systems are usually distinguished by the electrolyte that is used and the operating temperature, though there are always other important differences as well. There are six principal types of fuel cell, namely: ● ● ● Low temperature (50–150°C): alkaline electrolyte (AFC), proton‐exchange membrane (PEMFC), direct methanol (DMFC) and other liquid‐fed fuel cells. Medium temperature (around 200°C): PAFC. High temperature (600–1000°C): molten carbonate (MCFC) and SOFC. Some operational data on each type are given in Table 1.1. There are other less well‐ known types such as the direct borohydride (DBFC) and direct carbon fuel cells (DCFC); the former operates at low temperatures and the latter at high temperatures. Table 1.1 Principal types of fuel cell. Fuel cell type Operating Mobile ion temperature (°C) Fuel Alkaline (AFC) OH– 50–200 Pure H2 Space vehicles, e.g., Apollo, Shuttle Proton‐exchange membrane (PEMFC) H+ 30–100 + a Pure H2 Vehicles and mobile applications, and for lower power CHP systems Direct methanol (DMFC) H+ 20–90 Methanol Portable electronic systems of low power, running for long times Phosphoric acid (PAFC) H+ ~220 H2, (low S, low CO, tolerant to CO2) Large numbers of 200‐kW CHP systems in use ~650 H2, various hydrocarbon fuels (no S) Medium‐ to large‐scale CHP systems, up to MW capacity 500–1000 Impure H2, variety All sizes of CHP systems, of hydrocarbon fuels 2 kW to multi MW Molten carbonate CO32− (MCFC) Solid oxide (SOFC) O2− Applications and notes CHP, combined heat and power. a) New electrolyte materials as described in Chapter 4 are enabling higher operating temperatures for the PEMFC. 1 Although hydrogen is preferred for most types of fuel cell, other fuels can be used for some technologies. For example, methanol is employed in the direct methanol fuel cell (DMFC) and carbon as the fuel in the direct carbon fuel cell (DCFC). 17 18 Fuel Cell Systems Explained To date, the PEMFC has proved to be the most successful commercially. The electrolyte is a solid polymer, in which protons are mobile. The chemistry is the same as that shown Figure 1.3 for an acid‐electrolyte system. The PEMFC runs at relatively low temperatures, so the problem of slow reaction rates is addressed by using sophisticated catalysts and electrodes. Platinum has been the preferred catalyst. It is an expensive metal but, through improvements in materials, only minute amounts are now required. Consequently, in modern PEMFC designs, the platinum makes a relatively small contribution to the total cost of the fuel‐cell system. More recent research suggests that in some cases platinum can be eliminated from the catalyst. Further discussion of the PEMFC is given in Chapter 4. The PEMFC has to be fuelled with hydrogen of high purity, and methods for meeting this requirement are discussed in Chapter 10. The DMFC is a variant of the PEMFC. The technology differs from the PEMFC only in that methanol in its native liquid form is used as fuel. Other liquid fuels such as ethanol and formic acid may also be viable for some applications. Unfortunately, most of these liquid‐fuelled cells produce very low levels of power, but, even with this limitation, there are many potential applications for such devices in the rapidly growing area of portable electronics devices. Such cells, for the foreseeable future at least, will remain low‐power units and will therefore suit applications that require slow and steady consumption of electricity over long periods. As mentioned earlier, an AFC system was chosen for the Apollo and Space Shuttle orbiter craft. The problem of slow reaction rate was overcome by using highly porous electrodes, with a platinum catalyst, and sometimes by operating at quite high pressures. Although some historically important AFCs have been run at about 200°C, the systems usually operate below 100°C. Unfortunately, the AFC is susceptible to poisoning by the carbon dioxide in the atmosphere. Thus the air and fuel supplies must be free from this gas, or else pure oxygen and hydrogen must be supplied. The PAFC was the first type of fuel cell to reach commercialization and the technology enjoyed a reasonable degree of widespread terrestrial use in the period 1980–2000. Many 200‐kW systems, manufactured by the International Fuel Cells Corporation, were installed in the United States and Europe. Other systems were produced by Japanese companies. In the PAFC, porous electrodes, platinum catalysts and a moderately high temperature (~220°C) help to boost the reaction rate to a reasonable level. Such PAFC systems were fuelled with natural gas, which is converted to hydrogen within the fuel‐ cell system by steam reforming. The required equipment for steam reforming unfortunately adds considerably to the costs, complexity and size of the fuel‐cell system. Nevertheless, PAFC systems have demonstrated good performance in the field, for instance, units have run for periods in excess of 12 months without any maintenance that has required shutdown or human intervention. A typical installation of a 400 kW PAFC system is shown in Figure 1.14. The most common form of SOFC operates in the region of 600–1000°C. These high temperatures permit high reaction rates to be achieved without the need for expensive platinum catalysts. At these elevated temperatures, fuels such as natural gas can be used directly (internally reformed) within the fuel cell without the need for a separate processing unit. The SOFC thus addresses the aforementioned key problems (viz. slow reaction rates and hydrogen supply) and takes full advantage of the inherent simplicity of the fuel‐cell concept. Nevertheless, SOFCs are made from thin ceramic materials that are difficult to handle and therefore are expensive to manufacture. In addition, Introducing Fuel Cells Figure 1.14 Phosphoric acid fuel cell for stationary power‐plant applications (Source: Creative commons – Courtesy of UTC.) a large amount of extra equipment is needed to make a full SOFC system, e.g., air and fuel preheaters, heat-exchangers and pumps. Also the cooling system is more complex than for low‐temperature fuel cells. Care also has to be taken during start‐up and shutdown of SOFC systems, on account of the intrinsic fragile nature of the ceramic materials in the stacks. The MCFC has an interesting and distinguishing feature in that it requires carbon dioxide to be fed to the positive electrode, as well as oxygen. This is usually achieved by recycling some of the exhaust gas from the anode to the cathode inlet. The high temperature means that a good reaction rate is achieved with a comparatively inexpensive catalyst — nickel. Like the SOFC, an MCFC system can be fuelled directly with gases, such as methane and coal gas (a mixture of hydrogen and carbon monoxide), without the need for an external reformer. This advantage for the MCFC is somewhat offset, however, by the nature of the electrolyte, namely, a hot and corrosive molten mixture of lithium, potassium and sodium carbonates. 1.7 Mechanically Rechargeable Batteries and Other Fuel Cells At the start of this book, a fuel cell was defined as an electrochemical device that converts a fuel to electrical energy (and heat) continuously, as long as reactants are supplied to its electrodes. The implication is that neither the electrodes nor the electrolyte is consumed by operation of the cell. Of course, in all fuel cells the electrodes and 19 20 Fuel Cell Systems Explained electrolytes are degraded and subject to ‘wear and tear’ during service. The first two technologies under consideration in this section are often misleadingly described as fuel cells and employ electrodes that are entirely consumed during use. 1.7.1 Metal–Air Cells The most common type of cell of this category is the zinc–air battery, though aluminium– air and magnesium–air cells have been produced commercially. In all cases, the basic operation is the same. At the negative electrode, the metal reacts with hydroxyl ions in an alkaline electrolyte to form the metal oxide or hydroxide. For example, the reaction with a zinc fuel is given by: Zn 2OH ZnO H2O 2e (1.6) The electrons thus released pass round the external electric circuit to the air electrode where they are available for the reaction between water and oxygen to form more hydroxyl ions. Thus at the air electrode the reaction is exactly the same as equation (1.5) for the AFC. Cells using a salt solution (e.g., seawater) as the electrolyte solution also work reasonably well when using aluminium or magnesium as the fuel. Metal–air cells have a very high specific energy (Wh kg−1). Zinc–air batteries are employed widely in devices that require long running times at low currents, such as hearing aids. Some interest has also been shown in the development of units with higher power for application in electric vehicles. Such systems can also be ‘refuelled’ by replenishing the metal consumed at the negative electrode — which is why the technology is sometimes promoted as a ‘fuel cell’. This claim is also supported by the fact that the reaction at the positive electrode is exactly the same as for a fuel cell, and indeed the same electrodes can be used. It should be noted, however, that removal of the metal oxide will also necessitate renewal of the electrolyte solution. Thus, the metal–air systems cannot properly be described as fuel cells and are best classified as ‘mechanically rechargeable batteries’. 1.7.2 Redox Flow Cells Another type of electrochemical power source that is sometimes taken to be a fuel cell is the ‘redox flow cell’ (or ‘flow cell’); a multicell unit is usually referred to as a ‘flow battery’. It is useful at this point to define two types of flow cell, as several different chemistries are under development: 1) Flow batteries, in which there is a decoupling of cell power and cell capacity, e.g., the bromine–polysulfide cell and the vanadium redox cell. 2) Hybrid flow batteries, in which there is no decoupling of cell power and cell capacity, e.g., the zinc–bromine battery. The first category is different from all other fuel cells in that the oxidant is not air, and therefore it cannot be said that the fuel is ‘combusted’. In this type of cell, there is one reactant (which can be called the fuel) that is oxidized and a complementary reactant that serves as the oxidant. These are removed from the electrode compartments when the cell is being charged and stored in tanks. The capacity of such cells can thus be very large. Discharge is undertaken by resupplying the reactants to the electrodes. Introducing Fuel Cells Two flow batteries have been the subject of much research, namely, the sodium‐ bromide–sodium‐polysulfide cell and the vanadium redox cell. The former cell was introduced in the 1990s by Regenesys Technologies Limited in the United Kingdom. After a utility‐scale demonstration at a power station by the National Power in Cambridgeshire, United Kingdom, the development was taken over by RWE and subsequently by Prudent Energy to complement its own work on a vanadium battery. No further studies or trials of the Regenesys have been reported. The vanadium redox battery was pioneered in the 1980s at the University of New South Wales in Sydney, Australia, and the Japanese Electrotechnical Laboratory. The operating principle of the system is illustrated in Figure 1.15. The two reactants are flowing aqueous solutions of vanadium sulfate and the electrode reactions are as follows. At the positive electrode: VO2 Discharge 2H e Charge VO2 (1.7) H2O At the negative electrode: V2 Discharge Charge V3 (1.8) e Thus, in the charged state, the positive‐electrolyte loop contains a solution of V5+ and the negative loop contains a solution of V2+. On discharging, the former solution is reduced to V4+ and the latter is oxidized to V3+. The difference in the oxidation state of e– e– Load or power source e e d ro El Tank V2+/V3+ e an d ro br t ec t ec em Charge V2+ V3+ M Charge V4+ El Tank V5+/ V 4+ V5+ + – Discharge Ions Pump Discharge Pump Figure 1.15 Operating principle of the vanadium redox battery. 21 22 Fuel Cell Systems Explained vanadium in the two reactant solutions produces 1.2–1.6 V across the membrane, as determined by the electrolyte solution, temperature and state-of-charge. Regeneration takes place by reversing the flow of the solutions and applying a potential across the cell to restore the original oxidation states in the solutions. It can easily be seen that (i) this is a reversible cell and (ii) the capacity of the cell (e.g. as measured in kWh) is determined by the amount of liquid pumped, i.e., the size of the storage tanks, and not by the dimensions of the electrodes as would be the case in a normal battery. Furthermore, the more the cells and the faster the flow of electrolyte solutions, the higher is the power rating. This approach enables economies of scale in both manufacturing and energy–power capacity. The vanadium redox cell shares many characteristics with now‐abandoned Regenesys. Numerous companies and organizations have been involved in funding and developing the vanadium technology, and several large field trials have been conducted around the world. Research and development is continuing. In the hybrid form of flow cell, one or more of the electroactive components are deposited as a solid layer. Consequently, the system may be viewed as a combination of one battery electrode and one fuel‐cell electrode. The zinc–bromine system is the best‐ known example of such technology. A modern version developed by Redflow Limited, an Australian‐based company, is shown in Figure 1.16. As with the vanadium redox cell, the zinc–bromine cell is comprised of two fluids that pass carbon‐plastic electrodes that are each placed in a half‐cell either side of a microporous polyolefin membrane. During discharge, zinc and bromine combine into zinc bromide and thereby generate 1.8 V across each cell. During charge, metallic zinc will be drawn out of solution and deposited (plated) as a thin film on one side of the negative electrode. Meanwhile, bromine evolves as a dilute solution at the positive electrode on the other side of the membrane. Because bromine is a highly volatile and reactive liquid, it is complexed with an organic reagent to form a poly‐bromo compound, which is an oil and is immiscible with the aqueous electrolyte solution. The oil sinks down to the bottom of the electrolytic Figure 1.16 Redox zinc–bromine battery. (Source: Courtesy of Redflow Pty Ltd.) Introducing Fuel Cells tank and is separated and stored in a special compartment in the external reservoir of the positive electrode until required again for discharge. The capacity of the cell is limited by the amount of zinc that can be plated on the negative electrode. 1.7.3 Biological Fuel Cells Finally, it should be noted that, although not yet a principal technology, the biological fuel cell is attracting interest as a long‐term prospect. The cell would normally operate with an organic fuel, such as methanol or ethanol. The distinctive ‘biological’ aspect is that the electrode reactions are promoted by enzymes present in microbes, rather than by conventional ‘chemical’ catalysts such as platinum. Hence, these systems — also known as ‘microbial fuel cells (MFCs)’ — replicate nature in the way that energy is derived from organic fuels. Biological or microbial fuel cells should be distinguished from biological methods for generating hydrogen, which is then used in a conventional fuel cell. Such methods of hydrogen production are discussed in Chapter 10. Research into advanced microfluidics, new bacterial strains, more robust separator membranes and efficient electrodes is the key to unlocking the potential of MFCs. 1.8 Balance‐of‐Plant Components It should be evident that a practical fuel‐cell system requires not only a readily available fuel but also a means of cooling the stack, an ability to employ the heat produced to do useful work and an application for the direct current (dc) power that is produced by the stack(s). For a fuel‐cell stack to function effectively, various other components are necessary. The exact composition of this so‐called balance-of-plant depends on the type of fuel cell, the available fuel and its purity and the desired outputs of electricity and heat. Typical auxiliary subsystems are:- (i) fuel clean‐up processor, e.g., for sulfur removal — so‐called desulfurization; (ii) steam reformer and shift reactor for the fuel; (iii) carbon dioxide separator; (iv) humidifier; (v) fuel and air delivery units; (vi) power‐conditioning equipment, e.g., for inverting dc to alternating current (ac) and then transforming to line voltage; (vii) facilities for the management of heat and water; (viii) overall control and safety systems; and (ix) thermal insulation and packaging. Individual components include fuel storage tanks and pumps, compressors, pressure regulators and control valves, fuel and/or air pre‐heaters, heat-exchangers and radiators, voltage regulators, motors and batteries (to provide power for pumps on start‐up). These important subsystem issues are described in much more detail in Chapter 12. The requirements for a fuel‐cell system for a stationary power application and a vehicle are very different. In a stationary power plant system, such as shown in Figure 1.14, the fuel‐cell stack is, in terms of size, a small part of the installation that is dominated by the fuel and heat‐processing systems and the power‐conditioning equipment. This will nearly always be the case for combined heat and power (CHP) facilities that run on a conventional fuel such as natural gas. By contrast, a fuel‐cell power source for a car is shown in Figure 1.17. The unit operates on gaseous hydrogen fuel that is stored on the vehicle, and the waste heat is only used to warm the car interior. The fuel‐cell stack occupies the bulk of the compartment 23 24 Fuel Cell Systems Explained Figure 1.17 Hyundai fuel‐cell system located under the car hood. (Source: Courtesy of Hyundai Motor Company, Australia.) that would normally be filled with an internal combustion engine (ICE). Other components of a hydrogen fuel‐cell ‘engine’ in a vehicle, i.e., pumps, humidifier, power electronics and compressor, are generally much less bulky than those of a CHP system. 1.9 Fuel‐Cell Systems: Key Parameters To compare the performance of fuel‐cell systems with each other and with other electric power generators, some key operating parameters must be considered. For electrodes and electrolytes, the key criterion is the current per unit area, which is always known as the ‘current density’ and usually expressed in terms of mA cm−2, except in the United States where A ft−2 is frequently adopted (the two units are quite similar, i.e., 1.0 mA cm−2 = 0.8 A ft−2). The current density should be reported at a specific operating voltage, typically about 0.6 or 0.7 V. The values for current density and selected voltage can then be multiplied to give the power per unit area, in mW cm−2. A note of caution should be made here, namely, that electrodes frequently do not ‘scale up’ properly. That is, if the area is doubled the current will often not double. The reasons for this are varied but generally relate to issues such as the even delivery of reactants to, and removal of products from, the entire face of the electrode. Specific power (kW kg−1) and power density (kW m−3 or kW L−1) are key ‘figures of merit’ for comparing electrical generators. Note that whereas power is measured in kW, energy is simply power delivered over a certain period of time and is measured in kWh. The capital cost of a fuel‐cell system is obviously an important parameter and is usually quoted in US$ per kW for ease of comparison. The lifetime of a fuel cell is rather difficult to specify. Standard engineering measures such as ‘mean time between failures’ (MTBF) are not entirely applicable given that the performance of a fuel cell always gradually deteriorates and the power drops fairly Introducing Fuel Cells steadily with time as the electrodes and the electrolyte solution both age. The degradation of a fuel cell is sometimes reported as a decline in cell voltage, given in units of mV per 1000 h. Formally, the life of a fuel cell is considered to be over when it can no longer deliver the rated power, e.g., when a 10‐kW fuel cell can no longer deliver 10 kW. It should be noted that, when new, a fuel cell may be capable of providing more than the rated power, e.g., an extra 25% is not unusual. The remaining fuel‐cell characteristic of key importance is the efficiency, i.e., the electrical energy delivered by the system compared with the energy supplied as fuel. When making comparisons between systems in terms of efficiency, care should be taken that the data are expressed on the same basis. Efficiency is addressed in Chapter 2. In the automotive industry, primary issues are the cost per kW and the power density. In round figures, current ICE technology costs US$10 per kW and delivers 1 kW L−1. Such a power source should last at least 4000 h, i.e., about 1 h of duty each day for over 10 years. For CHP plant, the capital cost is still important, but a much higher target of US$1000 per kW is generally accepted. The higher cost is due to the extra balance of plant that is required and to the fact that the system must have a substantially longer lifetime. A period of 40 000 h would be a minimum. For stationary power‐generation systems, the levelized cost of electricity (LCOE) is often used as a measure of performance. The LCOE is the price at which electricity must be generated from a specific source to break even over the lifetime of the project. It is an economic assessment of the cost of the generating system and includes all the costs over its lifetime, namely, capital cost, operations and maintenance, and cost of fuel. The LCOE enables analysts to compare the costs of fuel‐cell systems with other forms of power generation. 1.10 Advantages and Applications For all types of fuel cell, a significant disadvantage or barrier to commercialization is the capital cost. There are, however, various advantages that feature more or less strongly for the different systems and lead to fuel cells being attractive for different applications. These include the following: ● ● ● ● Efficiency. As explained in Chapter 2, fuel cells are generally more efficient than piston‐ or turbine‐based combustion engines. A further benefit is that small fuel‐cell systems can be just as efficient as large ones. This capability opens up a market opportunity for small‐scale cogeneration (CHP) that cannot be satisfied with turbine‐ or engine‐based systems. Simplicity. The essentials of a fuel cell involve few, if any, moving parts. This can lead to highly reliable and long‐lasting systems. Low emissions. When hydrogen is the fuel, pure water is the by‐product of the main reaction of the fuel cell. Consequently, the power source is essentially ‘zero emission’. This is a particularly attractive for vehicle applications, as there is a requirement to reduce emissions and even eliminate them within cities. Nevertheless, it should be noted that, at present, emissions of carbon dioxide are nearly always involved in the production of the hydrogen. Silence. Fuel cells are very quiet — even those with extensive extra fuel‐processing equipment. Quietness is very important in both portable‐power applications and for local power generation via CHP schemes. 25 26 Fuel Cell Systems Explained Ironically, the fact that hydrogen is the preferred fuel is, in the main, one of the principal disadvantages of fuel cells. On the other hand, many envisage that as fossil fuels run out, hydrogen will become a major fuel and energy vector throughout the world. It could be generated, for example, by electrolysing water using electricity provided by massive arrays of photovoltaic (solar) cells. Indeed, the so‐called hydrogen economy may emerge in future decades. In the meantime, it is more likely that ‘hydrogen energy’ will have only a very small impact globally as it is most economically produced by the steam reforming of natural gas (see Chapter 10). In summary, the advantages of fuel cells impact particularly strongly on CHP systems (both large and small scales) and on mobile power systems ― especially for vehicles and electronic equipment such as portable computers, mobile telephones and military communications equipment. A notable feature of the technology is the very wide range in system sizes, i.e., from a few watts up to several megawatts. In this respect, fuel cells are unique as energy converters. Further Reading Bossel, U, 2000, The Birth of the Fuel Cell 1835‐1845, European Fuel Cell Forum, Oberrohndorf. Hoogers, G, 2003, Fuel Cell Technology Handbook, CRC Press, Boca Raton, FL. ISBN 0‐8493‐0877‐1. 27 2 Efficiency and Open‐Circuit Voltage This chapter examines the efficiency of fuel cells—how it is defined and calculated and what are the limits. The energy considerations provide information about the open‐ circuit voltage (OCV) of a fuel cell, and the associated formulae yield important details of the effect on the voltage of factors such as pressure, gas concentration and temperature. 2.1 Open‐Circuit Voltage: Hydrogen Fuel Cell The inputs and outputs of energy in a fuel‐cell system are shown schematically in Figure 2.1. The electrical power and energy output are easily calculated from the following well‐known formulae: Power V I (2.1) Energy V I t (2.2) where V is voltage, I is current and t is time. By contrast, the energies of the chemical inputs and outputs are less easily defined. In simple terms, it could be said that the ‘chemical energies’ of the hydrogen, oxygen and water are involved. The problem is that ‘chemical energy’ can be defined in different ways—terms such as enthalpy, Helmholtz function and Gibbs free energy are used. In recent years, the term ‘exergy’ has also become popular,1 and this is particularly useful when considering the operation of high‐temperature fuel cells. The reader will also come across older terms such as ‘heating value’ or ‘calorific value’ in the literature. In the case of fuel cells, it is the ‘Gibbs free energy’ that is fundamentally important. This can be defined as the energy liberated or absorbed in a reversible process at constant pressure and constant temperature. Put another way, it is the minimum thermodynamic work (at constant pressure) required to drive a chemical reaction (or, if negative, the maximum work that can be done by the reaction). Thus, the Gibbs free energy is a quantity that can be used to determine if a reaction is thermodynamically viable or not. The change in free energy, ΔG, in a chemical reaction (i.e., the difference between the Gibbs free energies of the reactants and products) is given by ΔG = ΔH − TΔS, 1 In thermodynamics, the exergy of a system is the maximum useful work available during a process that brings the system into equilibrium with its surroundings. Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 28 Fuel Cell Systems Explained Hydrogen Energy = ? Electricity Energy = V × I × t Fuel cell Oxygen Energy = ? Figure 2.1 Inputs and outputs of a fuel cell. Heat Water where ΔH is the change in enthalpy, ΔS is the change in entropy between reactants and products, and T is the absolute temperature. This expression is known as the ‘Gibbs equation’. At this point, it should be remarked that absolute values of the properties of thermodynamic functions such as enthalpy and entropy are unknown. Only changes in values caused by parameters such as temperature and pressure can be determined. It is therefore important to define a baseline for substances, to which the effect of such variations may be referred. The ‘standard state’ is such a baseline and defines the standard conditions for temperature and pressure. The International Union of Pure and Applied Chemistry (IUPAC) has two standards: (i) standard temperature and pressure, abbreviated as ‘STP’, specifies a temperature of 273.15 K and an absolute pressure of 100 kPa (1 bar) and (ii) standard ambient temperature and pressure, abbreviated as ‘SATP’, specifies a temperature of 298.15 K and an absolute pressure of 100 kPa (1 bar).2 It is customary to use the superscript ° to denote that a given quantity is in its reference state and the subscript f to indicate that a compound is formed from its elements. Thus G °f is the Gibbs free energy of formation of a compound under standard state conditions and, therefore, is more frequently referred to as the ‘standard free energy of formation’. Pure elements are taken to have a free energy of formation of zero at the reference state. Thus, for an ordinary hydrogen fuel cell operating at STP, the Gibbs free energy of each reactant (hydrogen and oxygen) is zero—a useful simplification. When using thermodynamic functions such as free energy, care should be taken that the reference states are clearly defined. The standard Gibbs free energy of formation of a compound, G °f , is the change in Gibbs free energy that accompanies the formation of 1 mol of a substance in its standard state from its constituent elements in their standard states. Often, the standard state for gases is taken as 298.15 K or 25°C rather than the recommended 0°C. In most cases, confusion is avoided so long as all of the quantities are referred to the same standard conditions. In a fuel cell, it is the change in the Gibbs free energy of formation ΔGf that generates the electrical energy released by the cell. This change is the difference between the Gibbs free energy of formation of the products and that of the inputs or reactants, namely, Gf G f products G f reactants (2.3) 2 An earlier IUPAC definition of STP in terms of 273.15 K and 1 atm (101.325 kPa) was discontinued in 1982. Reference may also be made to a normal temperature and pressure (NTP), which is usually taken to be 20°C (293.15 K) and 1 atm (101.325 kPa). Efficiency and Open‐Circuit Voltage Box 2.1 Molar Mass and the Mole The ‘mole’ (abbreviation, ‘mol’) is the unit of measurement in the International System of Units (French: Système international d’unités, SI), which expresses the amount of a given material. One mole is defined as the number of atoms in precisely 0.012 kg (i.e., 12 g) of carbon‐12, the most common naturally occurring isotope of the element carbon. This dimensionless number is equal to approximately 6.022140857 × 1023 and is also called ‘Avogadro’s number’ or the ‘Avogadro constant’. It is represented by the letter NA or L. The SI unit for molar mass is kg mol−1. For historical reasons, however, molar masses are almost always expressed in g mol−1. The ‘unified atomic mass unit’ (symbol, u) is numerically equivalent to 1 g mol−1, i.e., one‐twelfth the mass of one atom of carbon‐12. (Note that the ‘atomic mass unit’—symbol, amu—without the ‘unified’ prefix is a technically obsolete unit based on oxygen‐16, but most uses of this term actually refer to the unified atomic mass unit.) For example, it follows that the molar mass of H2 is 2.0 u, and therefore 1 g mol of H2 is 2.0 g and 1 kg mol is 2.0 kg. Similarly, the molecular mass of H2O is 18 u, so 18 g is 1 g mol and 18 kg is 1 kg mol. A mole of any substance always has the same number of entities (atoms, molecules, ions, electrons, photons) so that a mole of electrons is 6.022140857 × 1023 electrons. The charge is NA e−, where e− is 1.60217662 × 10−19 coulombs—the charge on one electron. This quantity is called the ‘Faraday constant’, which is designated by the letter F and has the following value: F N Ae 96 485 coulombs To make comparisons easier, it is nearly always most convenient to consider these quantities in their ‘per mole’ form, as discussed in Box 2.1. These can be indicated by a bar over a lower‐case letter, e.g., ( g f )H2O represents the molar Gibbs free energy of formation for water. Consider the basic reaction for the hydrogen–oxygen fuel cell: 2H 2 O 2 (2.4) 2H 2 O This is equivalent to: 1 O 2 2 H2 (2.5) H2O The ‘product’ is 1 mol of H2O, and the ‘reactants’ are 1 mol of H2 and 1/2 mol of O2. Thus (2.6) 1 g g g g f f HO 2 f H 2 2 f O 2 This equation seems straightforward and simple enough. The Gibbs free energy of formation is not constant, however, but changes with temperature and state (liquid or gas). Values of ∆ g f for the basic reaction of the hydrogen fuel cell under a number of different conditions are listed in Table 2.1. The method used to calculate these data is outlined in Appendix 1. Note that the values are negative, and therefore, by convention, this indicates that energy is released by the reaction. For the hydrogen fuel cell, two electrons pass round the external circuit for each water molecule produced and each hydrogen molecule used. Thus, for each mole of hydrogen 29 30 Fuel Cell Systems Explained Table 2.1 ∆ g f for the reaction H2 1 O 2 2 H2O at various temperatures. (kJ mol−1) Form of water product Temperature (°C) ∆g f Liquid 25 −237.2 Liquid 80 −228.2 Gas 80 −226.1 Gas 100 −225.2 Gas 200 −220.4 Gas 400 −210.3 Gas 600 −199.6 Gas 800 −188.6 Gas 1000 −177.4 consumed, 2NA electrons pass round the external circuit. Given that each electron carries a unit negative charge (e−), the corresponding charge, in coulombs (C), that flows is 2 N Ae (2.7) 2F where F is the Faraday constant or the charge on 1 mol of electrons (see Box 2.1). If V is the voltage of the fuel cell, then the electrical work, in joules (J), expended in moving this charge round the circuit is Electrical work done charge voltage 2FV (2.8) If the system is thermodynamically reversible (i.e., it has no energy losses), then the electrical work done will be equal to the Gibbs free energy released by the fuel‐cell reaction ∆ g f . Thus: gf 2 FVr or Vr gf 2F (2.9) This fundamental equation gives the ‘reversible voltage’, Vr, or ‘OCV’ across the terminals of the cell when there is no net current flow. Under standard conditions, this is the ‘standard cell voltage’ , Vr°. When the fuel is hydrogen, the reversible voltage under standard conditions (STP) is 1.229 V at 25°C. If the cell operates at 200°C, then ∆ g f = −220.4 kJ (from Table 2.1), and thus: Vr 22.04 103 2 9 6485 1.14 V (2.10) Note that this value assumes no ‘irreversibilities’ and that pure hydrogen and oxygen are supplied at standard pressure (100 kPa). In practice, the voltage would be lower than this because of the voltage losses discussed in Chapter 3. Some of these irreversibilities Efficiency and Open‐Circuit Voltage Box 2.2 Reversible Processes, Irreversibilities and Losses An example of a simple reversible process is that shown in Figure 2.2, which depicts a ball of mass m about to roll down a hill. In position A, the ball has no kinetic energy, but a potential energy given by m × g × h, where g is the acceleration due to gravity. If m is expressed in kg, g in m s−2 and h in m, then the energy is expressed in joules. In position B, the potential energy has been converted into kinetic energy. If there is no rolling resistance or wind resistance, then the process is ‘reversible’, i.e., the ball can roll up the other side and recover its potential energy. In practice, however, some of the potential energy will be converted into heat because of friction and wind resistance. The process is now ‘irreversible’ as the heat cannot be converted back into kinetic or potential energy. It might be tempting to describe this as a ‘loss’ of energy but that would not be very precise. In a sense, the potential energy is no more ‘lost’ to heat than it is ‘lost’ to kinetic energy. So, the term ‘irreversible energy loss’ or ‘irreversibility’ is a rather more precise description of situations that many would describe as a ‘loss of energy’. Ball of mass m A h B Figure 2.2 Simple reversible process. even exert a slight influence when no current is drawn, so the OCV of a fuel cell will usually be lower than the value given by equation (2.9). Further explanation of ‘reversible’ and ‘irreversible’ processes is given in Box 2.2. 2.2 Open‐Circuit Voltage: Other Fuel Cells and Batteries Equation (2.9) derived for the OCV of the hydrogen fuel cell is also applicable to other reactions. The only step in the derivation that was specific to the hydrogen fuel cell was the two electrons for each molecule of fuel consumed. In general terms, therefore, equation (2.9) can be written as Vr g zF f (2.11) where z is the number of electrons transferred for each molecule of fuel. The derivation is also not specific to fuel cells and applies equally well to other electrochemical power sources, particularly primary and secondary batteries. For example, the primary alkaline cell that is used widely for domestic applications employs electrodes of zinc and manganese dioxide. The overall cell reaction in this battery can be expressed simply by 31 32 Fuel Cell Systems Explained Zn 2MnO2 H2O ZnO 2MnOOH (2.12) for which ∆ g f is −277 kJ mol−1. At the negative electrode the reaction can be given as: Zn 2OH ZnO H2O 2e (2.13) and at the positive electrode as: 2MnO2 2H2O 2e 2MnOOH 2OH (2.14) Thus two electrons are passed round the circuit, and the OCV is expressed according to equation (2.11), namely, Vr 277 103 1.44 V 2 96 485 (2.15) Another example is the methanol fuel cell, which is discussed in Chapter 6. The overall reaction is: 2CH3OH 3O2 4 H2O 2CO2 (2.16) and involves the passage of 12 electrons from the negative to the positive electrode, i.e., 6 electrons for each molecule of methanol. For the methanol reaction, g f is −698.2 kJ mol−1. Substituting this information into equation (2.11) gives: Vr 698 103 1.21 V 6 96 485 (2.17) It is to be noted that this is similar to the OCV for the hydrogen fuel cell. 2.3 Efficiency and Its Limits The efficiency of a fuel cell—the fraction of the energy in the fuel that is converted into useful electrical output—is a critical issue. Much is made of the fact that fuel cells are not heat engines, so their efficiency is not limited by the Carnot cycle3 and therefore should be high. This reasoning has driven much of the interest and investment in the technology. The Carnot theorem as applied to a heat engine can be expressed as: heat engine W H T1 T2 T1 (2.18) where W is the generated work, ΔH is the heat of combustion of the fuel and T1 and T2 are the absolute temperatures between which the heat engine operates. In practice, heat engines are irreversible and normally operate with the lower temperature (T2) at room temperature and with the upper temperature (T1) imposed by the materials of construction of the engine. Thus, the efficiency of a heat engine is limited and depends on the 3 The Carnot cycle states that only a fraction of the heat produced by an engine can perform work and that the remainder dissipates into the engine, its compartment and the environment. Efficiency and Open‐Circuit Voltage temperatures at which heat is supplied and withdrawn. As an example, for a steam turbine operating at 400°C (673 K) with the water exhausted through a condenser at 50°C (323 K), the Carnot efficiency limit is: 673 323 673 0.52 or 52% (2.19) For a fuel cell working ideally under isothermal conditions, the free energy change of the reaction may be totally converted into electrical energy with a (maximum) efficiency given by: max Wmax H G H 1 T S H (2.20) where Wmax is the maximum work delivered. The term TΔS is the heat exchanged with the surroundings. Thus, under reversible conditions, the reaction enthalpy is converted into electrical energy, except for an entropy term. The ΔH is usually larger in magnitude than ΔG to such an extent that the ideal efficiency of a fuel cell, on a thermal basis, is usually in the region of 90%, i.e., superior to that of a heat engine. It should be noted that, for positive reaction entropies, the efficiency may become greater than 100% because under isothermal conditions heat energy would be absorbed from the surroundings and converted into electricity. The theoretical maximum efficiency of a fuel cell (ηmax) is sometimes called the ‘thermodynamic efficiency’. Unfortunately, the previously mentioned definition of efficiency is not without its ambiguities, as there are two different values that can be used for the ΔH term. For the conventional oxidation of hydrogen: H2 hf 1 O 2 2 H2O(steam ) 241.83 kJ mol (2.21) 1 If the product water is condensed back to liquid, the reaction is: H2 hf 1 O 2 2 H2O liquid 285.84 kJ mol (2.22) 1 The difference between these two values for ∆h f (44.01 kJ mol−1) is the molar enthalpy of vapourization4 of water. The higher figure is called the ‘higher heating value’ (HHV) and the lower, quite logically, the ‘lower heating value’ (LHV). Any statement of efficiency should say whether it relates to the HHV or LHV of the fuel. When comparing the efficiencies of various appliances that are using the same fuel, it is convenient to take the LHV since this is usually the maximum amount of heat that can be recovered in the appliance itself. The difference between LHV and HHV varies with the fuel. Generally, the sensible heat5 is small, and it is the heat of condensation of steam that predominates. It follows that the richer the fossil fuel is in hydrogen, the greater the deviation is between the LHV and the HHV. For example, the ratio of LHV 4 This used to be known as the ‘molar latent heat’. 5 Sensible heat is heat exchanged by a body or thermodynamic system in which the exchange of heat changes the temperature of the body or system without causing a phase change. 33 Fuel Cell Systems Explained Table 2.2 ∆ g f , maximum open‐circuit voltage and thermodynamic efficiency limit (HHV) for hydrogen fuel cells. Form of water product Liquid Temperature (°C) 25 Liquid ∆g f (kJ mol−1) −237.2 Maximum open‐ circuit voltage (V) Efficiency limit (HHV) (%) 1.23 83 80 −228.2 1.18 80 Gas 100 −225.3 1.17 79 Gas 200 −220.4 1.14 77 Gas 400 −210.3 1.09 74 Gas 600 −199.6 1.04 70 Gas 800 −188.6 0.98 66 Gas 1000 −177.4 0.92 62 90 Fuel cell, liquid product 80 Efficiency limit/% 34 70 Fuel cell, steam product 60 Carnot limit, 50°C exhaust 50 40 30 0 200 400 600 800 1000 Operating temperature/°C Figure 2.3 Maximum efficiency (HHV) of the hydrogen fuel cell at standard pressure. By way of comparison, the Carnot limit is shown for a 50°C exhaust temperature. to HHV is almost 1.0 for carbon monoxide (no hydrogen), 0.98 for coal (a little hydrogen), 0.91 for petrol, 0.90 for methane and 0.85 for hydrogen. The values of the efficiency limit, relative to the HHV, for a hydrogen fuel cell are listed in Table 2.2. The maximum OCVs, from equation (2.11), are also given. Plots in Figure 2.3 show how efficiencies vary with temperature and how they compare with the ‘Carnot limit’. The following three important points should be noted: 1) Although the information displayed in Figure 2.3 and Table 2.2 would suggest that lower fuel cell operating temperatures are better, the voltage losses are nearly always less at higher temperatures (these losses are discussed in detail in Chapter 3). In practice, therefore, fuel‐cell operating voltages are usually higher at higher temperatures. Efficiency and Open‐Circuit Voltage 2) Any energy in the fuel that is not converted into electricity in the fuel cell appears as waste heat (as with any heat engine). The waste heat from high‐temperature cells is more useful than that from low‐temperature cells. 3) Contrary to statements often made by their supporters, fuel cells do not always have a higher efficiency limit than heat engines.6 The decline in maximum possible efficiency with temperature associated with the hydrogen fuel cell does not occur in exactly the same way with other types of fuel cell. For example, when using carbon monoxide: CO 1 O 2 2 CO2 (2.23) The value of ∆ g changes even more rapidly with temperature, and the maximum possible efficiency falls from about 82% at 100°C to 52% at 1000°C. On the other hand, for the reaction CH 4 2O2 CO2 2H2O (2.24) ∆ g is fairly constant with temperature, and therefore there is very little change in the maximum possible efficiency. Fuel‐cell efficiency is a topic that has given rise to much confusion in the literature. In addition to the losses that originate in the cell stack, there are other system losses or external inefficiencies to be taken into account. These include electrical losses in compressing the incoming hydrogen and air and in converting the low‐voltage DC output to high‐voltage AC. The total effect is a significant reduction in overall system efficiency. Finally, if the fuel cells are to be used to propel electric vehicles, for example, there are also inefficiencies in the electric motors and the drivetrain to be considered. 2.4 Efficiency and Voltage It is clear from data given in Table 2.2 that there is a connection between the maximum voltage of a cell and its maximum efficiency. The operating voltage of a fuel cell can also be very easily related to its efficiency. This can be shown by adapting equation (2.9). If all the energy from the hydrogen fuel, i.e., the heating value, or enthalpy of formation, were transformed into electrical energy, the voltage would then be given by: Vr hf 2F (2.25) and have a value of 1.48 V and 1.35 V based on the HHV and the LHV, respectively. These are the voltages that would be obtained for a 100% efficient system. Consequently, the true efficiency of the cell is the actual voltage, Vc, divided by these values, e.g., Cell efficiency Vc 100% HHV 1.48 (2.26) 6 In Chapter 8, it is shown how a heat engine and a high‐temperature fuel cell can be combined into a particularly efficient system. 35 36 Fuel Cell Systems Explained In practice, however, it is found that not all the fuel can be used, for reasons discussed later; some of it usually has to pass through unreacted. A fuel utilization coefficient, μf, can be defined as: mass of fuel reacted in cell f (2.27) mass of fuel input to cell This parameter is equivalent to the ratio of the current delivered by the fuel cell to that which would be obtained if all the fuel were reacted. The fuel‐cell efficiency, η, is therefore given by: f Vc 100% 1.48 (2.28) If a figure relative to the LHV is required, 1.25 instead of 1.48 should be used in the previously mentioned formula. A good estimate for μf is 0.95, which allows the efficiency of a fuel cell to be estimated accurately from the very simple measurement of its voltage. The efficiency can be a great deal less in some circumstances, as is discussed in Section 2.5.3 and later in Chapter 6. 2.5 Influence of Pressure and Gas Concentration 2.5.1 Nernst Equation As discussed in Section 2.1, the Gibbs free energy changes in a chemical reaction vary with temperature. Equally important, though more complex, is the influence of both reactant pressure and concentration. Consider, for example, a general reaction such as: jA kB mC (2.29) where j moles of A react with k moles of B to produce m moles of C. Each of the reactants, as well as the product, has an associated ‘activity’,7 which is designated by the symbol a. Accordingly, aA and aB represent the activities of the respective reactants, and aC the activity of the product. For the case of gases behaving as ‘ideal gases’, it can be shown that a P P (2.30) where P is the pressure, or partial pressure, of the gas and P° is the standard pressure, namely, 100 kPa. Since fuel cells are generally gas reactors, this simple equation is very useful. The activity of a gaseous component in the system can be taken to be proportional to partial pressure, whereas for dissolved chemicals, the activity is linked to the molarity (‘strength’) of the solution, which is usually expressed in mol dm−3. The case 7 The thermodynamic activity of a species is a measure of the ‘effective concentration’ of a species in a reacting system. By convention, it is a dimensionless quantity. The activity of pure substances in condensed phases (liquids or solids) is taken as unity. Activity depends principally on the temperature, pressure and composition of the system. In reactions that involve real gases and mixtures, the effective partial pressure of a constituent gas is usually referred to as ‘fugacity’. Efficiency and Open‐Circuit Voltage of the water produced in fuel cells is somewhat difficult, since this can be as either steam or liquid. For steam, the following can be written: PH2O PH2O aH2O (2.31) where PH2O is the vapour pressure of the steam at the temperature concerned; values for this parameter can be obtained readily from published steam tables. When liquid water is the product, it is a reasonable approximation to assume that aH2O 1. The activities of the reactants and products modify the Gibbs free energy change of a reaction. By using thermodynamic principles, for a chemical reaction such as the general example given in equation (2.29), the following holds: gf gf RT ln aAj .aBk aCm (2.32) where g f is the change in molar Gibbs free energy of formation at standard pressure. For the reaction in a hydrogen fuel cell, equation (2.32) becomes: 1 gf gf RT ln aH2 aO2 2 (2.33) aH2O The standard free energy change for the reaction ( g f ) is the quantity given in Tables 2.1 and 2.2. Thus, if the activity of the reactants increases, ∆ g f becomes more negative, i.e., more energy is released. On the other hand, if the activity of the product increases, ∆ g f increases and becomes less negative, and less energy is released. To see how activity influences the cell voltage, ∆ g f can be substituted into equation (2.9) to obtain: 1 Vr gf 2F aH2 aO2 2 RT ln 2F aH2O (2.34) 1 Vr aH2 aO2 2 RT ln 2F aH2O where Vr° is the OCV at STP. The equation shows precisely how raising the activity of the reactants increases the voltage; it is known as the Nernst equation. Note that this relationship is equally applicable to individual electrodes with potentials Er and E r° replacing voltages Vr and Vr°, respectively. The Nernst equation can be manipulated to investigate the influence of different parameters on the operation and/or performance of a fuel cell. For example, in reaction (2.21), namely, H2 1 O 2 2 H2O steam (2.21) it can be assumed that the steam behaves as an ideal gas, and so: aH2 PH2 , aO2 P PO2 , aH2O P PH2O P (2.35) 37 38 Fuel Cell Systems Explained Then the Nernst equation will become: Vr RT ln 2F Vr PH2 PO2 P P 1 2 (2.36) PH2O P In nearly all cases, the pressures will be partial pressures; that is, the gases will be components of a mixture. For example, the hydrogen gas might be part of a mixture of hydrogen and carbon dioxide from a fuel reformer, together with product steam. Oxygen will nearly always be a component of air. It is also often the case that the total pressure on both the positive and negative electrodes is approximately the same as this simplifies the cell design. If the system pressure is P, then P , PO2 PH2 P , PHO P (2.37) where α, β and δ are constants that depend on the molar masses and concentrations of H2, O2 and H2O, respectively. The Nernst equation then becomes Vr Vr RT ln 2F Vr RT ln 2F 1 2 1 (2.38) P2 1 2 RT ln P 4F This relationship and equation (2.36) provide a theoretical basis for, and a quantitative indication of, the relative importance of a large number of variables in design and operation of a fuel cell. These variables are discussed in more detail in later chapters, but some points are considered briefly here to help introduce the technology. 2.5.2 Hydrogen Partial Pressure Hydrogen can be supplied either pure or as part of a mixture. Isolation of the hydrogen pressure term in equation (2.38) yields 1 Vr Vr P2 RT ln O2 PH2O 2F RT ln PH2 2F (2.39) So, if the hydrogen partial pressure changes, say, from P1 to P2, with PO2 and PH2O unchanged, then the resulting change in voltage ΔV will be given by V RT ln P2 2F P RT ln 2 2F P1 RT ln P1 2F (2.40) The use of hydrogen mixed with carbon dioxide occurs particularly in phosphoric acid fuel cells (PAFCs) that operate at about 200°C (473 K). Substituting the values for R, T and F in equation (2.40) yields: Efficiency and Open‐Circuit Voltage V P2 P1 0.02 ln (2.41) This relationship gives values that are in good agreement with experimental results, which correlate best with a factor of 0.024 instead of 0.020. As an example, changing from pure hydrogen to a 50% H2–50% CO2 mixture causes a reduction of 0.015 V per cell. 2.5.3 Fuel and Oxidant Utilization As air passes through the positive‐electrode (cathode) compartment of a fuel cell, oxygen is consumed, and thereby its partial pressure is reduced. Similarly, the partial pressure of the fuel will often decline in the negative‐electrode compartment. Referring to equation (2.39), it can be seen that α and β decrease, whereas δ increases. Consequently, the following term in equation (2.38): RT ln 2F 1 2 (2.42) becomes smaller as fuel and oxidant are consumed as they pass through the cell, and so the cell voltage would be expected to fall between the inlet and the outlet of the cell. In most stack designs, it is not actually possible to have variations in voltage throughout a cell—the fact that the electrodes are good electronic conductors ensures that the voltage is approximately uniform throughout each cell. Accordingly, it is the current density that changes throughout the cell. The current density will be lowest nearer the exit where the fuel concentration is lower.8 The RT term in equation (2.42) also dictates that the drop in cell voltage (or current density where the voltage cannot change) due to fuel and oxidant being consumed will be greater in high‐temperature fuel cells. Obviously, for a system to exhibit high efficiency, the fuel utilization should be as high as possible. On the other hand, equation (2.39) also suggests that high fuel utilization will lead to low average cell voltage or current density. The effect of low current density can be compensated by increasing the size of the cell, but this will increase the cost. In a practical system, therefore, it is always necessary to reach a compromise between fuel utilization and stack size (i.e., cost). This issue is most important with high‐temperature cells and is considered further in Chapters 7–9. 2.5.4 System Pressure The Nernst equation also demonstrates that the system pressure can increase the voltage of a fuel cell according to the term: RT ln P 4F (2.43) 8 The current density distribution in a stack will depend also on the orientation of the fuel and oxidant channels. Where the flows are parallel and in the same direction (co‐flow), the current density will be lowest at the outlet of the cells. This is not the case for counter‐flow or cross‐flow configurations. Modern flow‐ field design is focused on optimizing current density distribution throughout the stack. 39 40 Fuel Cell Systems Explained For instance, if the pressure changes from P1 to P2, there will be change in voltage: V RT P ln 2 4F P1 (2.44) For a solid oxide fuel cell operating at 1000°C, the equation would give: V 0.027 ln P2 P1 (2.45) This relationship has been found to be in very good agreement with reported results for high‐temperature cells, but not for other fuel cells that work at lower temperatures. For example, whereas a PAFC at 200°C should be affected by system pressure according to: V 0.010 ln P2 P1 (2.46) published data deliver a different correlation, namely, V 0.063 ln P2 P1 (2.47) In other words, at lower temperatures, the benefits of raising system pressure are much greater than predicted by the Nernst equation. This discrepancy in performance is because, except for very high‐temperature cells, increasing the pressure also reduces the losses at the electrodes, especially at the positive electrode; see Chapter 3. A similar outcome occurs when changing the oxidant from air to oxygen. This action effectively changes β in equation (2.38) from a value of 0.21 (21% oxygen in air) to 1.0 (pure oxygen). Isolating β in this equation gives: Vr Vr RT ln 4F RT ln 2F RT ln P 4F (2.48) The change in β from 0.21 to 1.0, with all other factors remaining constant, yields: V RT 1.0 ln 4F 0.21 (2.49) For a PEMFC at 80°C, the change in voltage would be 0.012 V. In fact, studies have demonstrated a much larger change; namely, 0.05 V is commonplace. Again, this is due to the reduction in overpotential at the cathode (positive electrode) as a result of high oxygen pressure. 2.6 Summary The OCV (also known as the reversible voltage) for a hydrogen fuel cell is given by: Vr G 2F (2.50) Efficiency and Open‐Circuit Voltage where ΔG is the free energy change for the fuel‐cell reaction. In general, for a reaction where z electrons are transferred for each molecule of fuel, the OCV is: Vr G zF (2.51) The Gibbs free energy change, ΔG, varies with temperature and other factors. The maximum efficiency is given by the expression: G 100% H max (2.52) The efficiency (HHV) of a working hydrogen fuel cell can be found by using the following simple formula: f Vc 100% 1.48 (2.53) where μf is the fuel utilization (typically about 0.95) and Vc is the voltage of a single cell. The pressure and concentration of the reactants also influence the change in Gibbs free energy, and thus the voltage. This is expressed in the Nernst equation, which can take many forms. For example, if the water product is in the form of steam, then: 1 Vr Vr PH2 PO2 2 RT ln 2F PH2O (2.54) where Vr° is the cell OCV at standard pressure. In most of this chapter, equations have been given for the voltage of a cell, or its OCV. In practice the operating voltage is less than that predicted and in some cases much less. This is the result of losses or ‘irreversibilities’, which are explained more fully in the next chapter. Further Reading Barclay, FJ, 2006, Fuel Cells, Engines and Hydrogen: An Exergy Approach, John Wiley & Sons, Ltd, Chichester. ISBN: 978‐0‐470‐01904‐7. EG&G Technical Services, Inc., under contract to US Department of Energy, 2016, Fuel Cell Handbook (Seventh Edition), National Energy Technology Laboratory, Morgantown, WV. Srinivasan, S, 2006, Fuel Cells. From Fundamentals to Applications, Springer, New York. ISBN: 9781441937728. Stolten, D (ed.), 2010, Hydrogen and Fuel Cells – Fundamentals, Technologies and Applications, Wiley‐VCH, Verlag GmbH & Co. KGaA, Weinheim. ISBN: 978‐3‐527‐32711. 41 43 3 Operational Fuel‐Cell Voltages 3.1 Fundamental Voltage: Current Behaviour As shown in Chapter 2, the theoretical value of the ‘no-loss’ open‐circuit voltage of a hydrogen fuel cell is expressed by equation (2.9): Vr gf 2F (2.9) where ∆ g f is the change in free energy for the cell reaction (i.e., the difference between the free energy of formation of the reactants and the free energy of formation of the products) and F is the Faraday constant. This gives a value of about 1.2 V for a cell that is operating below 100°C. When, however, a fuel cell is put to use, it is found that the ‘operational voltage’ is less than this, indeed often considerably less. The voltage versus current density1 performance of a single cell of typical design and operating at 40°C and normal air pressure is presented in Figure 3.1. The key points are as follows: ● ● ● ● Even the open‐circuit voltage is less than the theoretical value. There is a rapid initial drop in voltage. The voltage then falls less slowly and more linearly. A more rapid decline in voltage may be observed at higher current densities. There are two marked changes in the previously mentioned performance characteristics when a fuel cell is operated at higher temperatures, namely: ● ● As shown in Chapter 2, the reversible (‘no loss’) voltage falls, and thereby its value usually becomes closer to that of the actual operating voltage. The initial drop in voltage as current is drawn from the cell is greatly reduced. The performance for a typical solid oxide fuel cell (SOFC) that is operating at about 800°C is given in Figure 3.2 and has the following significant features: ● ● ● The open‐circuit voltage is equal to, or only a very little less than, the theoretical value. The initial drop in voltage is very small, and the graph is considerably more linear. There may be a higher current density at which the voltage falls rapidly away, as found for fuel cells that run at lower temperatures. 1 It is common practice to refer to current density, or current per unit area, rather than just current so that it is easier to compare the performance of cells of different size. Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. Fuel Cell Systems Explained ‘No-loss’ voltage of 1.2 V 1.2 Even the open-circuit voltage is less than the theoretical no-loss value 1.0 Rapid initial fall in voltage Cell voltage/V 44 Voltage falls more slowly, and graph is fairly linear 0.8 0.6 0.4 Voltage begins to fall faster at higher currents 0.2 0 0 200 400 600 800 Current density/mA cm–2 1000 Figure 3.1 Voltage versus current density performance of a typical fuel cell operating at low temperature and air pressure. Comparison of the two sets of data reveals that although the reversible voltage is lower for the cell running at the higher temperature, the actual operating voltage is generally greater, because the voltage drop or ‘irreversibilities’ are smaller. This chapter examines the factors that are responsible for the voltage falling below the reversible value and consider ways to ameliorate their adverse effects. 3.2 Terminology Efforts to develop fuel‐cell systems are highly interdisciplinary. Success requires the skills of chemists, electrochemists, materials scientists, thermodynamicists, electrical and chemical engineers, control and instrumentation engineers and others. Not surprisingly, there are occasions when these various disciplines have their own names for what is often essentially the same performance parameter. The main topic of this chapter — fuel‐cell voltage — is a case in point. The graphs of Figures 3.1 and 3.2 show the difference between the voltage that is expected from a fuel cell operating reversibly (ideally) and the voltage that is observed in practice. Remarkably, five names are commonly used to denote the voltage difference: ● ‘Overvoltage’ is a term often adopted by electrochemists to describe the nonideal behaviour of electrolysers, fuel cells and batteries. Similarly, ‘overpotential’ signifies differences in potentials that are generated at electrode|electrolyte interfaces. Unfortunately, the form of the word overvoltage tends to imply that the observed voltage is larger than the value predicted by theory, whereas in fuel cells the observed voltage is smaller. Operational Fuel‐Cell Voltages 1.2 ‘No-loss’ voltage of 1.0 V 1.0 Cell voltage/V Graph is fairly linear 0.8 0.6 Very small initial fall in voltage, and open-circuit voltage is very close to theoretical value 0.4 Voltage begins to fall faster at higher currents 0.2 0 0 200 400 600 800 1000 Current density/mA cm–2 Figure 3.2 Voltage versus current density performance of a typical fuel cell operating at about 800°C and air pressure. ● ● ● ● ‘Polarization’ is another term that has been employed by electrochemists, but it is misleading on several counts and is generally best avoided. ‘Irreversibility’ is the best term from a thermodynamics point of view. Nonetheless, it is perhaps not sufficiently specific to fuel cells and does not connect well with the main effect under consideration here, namely, that which gives rise to a reduction in cell voltage. ‘Voltage loss’ may be taken as a simple way to indicate that a practical fuel cell exhibits a voltage that is less than would be expected from thermodynamic considerations. A discussion of ‘reversibility, irreversibility and losses’ is given in Section 2.1, Chapter 2. ‘Voltage drop’ is certainly not scientifically precise, but it does convey the effect observed and is readily understood by electrical engineers. These alternative terms, which demonstrate the richness of the English language often having many words for the same subject, will be encountered during the course of this book. It is also worth remarking that ‘potential’ and ‘voltage’ are often misleadingly used interchangeably. As with the thermodynamic properties G, H and S that were introduced in the last chapter, electric potential is only able to be measured as a potential difference between two electrodes. Since a standard state is defined for thermodynamic properties, electrochemists have adopted the standard hydrogen electrode (SHE) as the reference against which the potential of an electrode can be measured. In this book, E is used to denote the potential of an electrode (i.e., with reference to the SHE), and E° is the electrode potential under standard conditions. The voltage difference between two electrodes in a cell is represented by the symbol V. 45 46 Fuel Cell Systems Explained 3.3 Fuel‐Cell Irreversibilities The characteristic shape of the voltage versus current density relationships shown in Figures 3.1 and 3.2 is the result of four major irreversibilities. The concomitant voltage loses will be outlined briefly here before being considered in more detail later, namely: 1) Activation losses. These represent the slowness of the reactions taking place on the surface of the electrodes. A proportion of the voltage generated is lost in driving the chemical reaction that transfers the electrons to or from the electrode. As discussed in Section 3.4, the resulting effect on the voltage is highly non‐linear. 2) Internal currents and fuel crossover. This voltage loss results from a small amount of fuel passing through the electrolyte from the anode to the cathode and, to a lesser extent, from electron conduction through the electrolyte. In an ideal situation, the electrolyte should only transport ions through the cell, as illustrated in Figures 1.3 and 1.4, Chapter 1. In practice, however, a certain amount of fuel diffusion and electron flow will always be possible. Generally, the fuel loss and current are both small, and thereby the net effect is usually not very important. Crossover does, however, have a marked influence on the open‐circuit voltage of low‐temperature cells, as will be examined in Section 3.5. 3) Ohmic losses. This voltage loss is the straightforward resistance to the flow of electrons through the material of the electrodes and the various interconnections, as well as the resistance to the flow of ions through the electrolyte. The voltage drop is essentially linearly proportional to the current density and therefore is sometimes also called resistive losses. 4) Concentration or mass‐transport losses. These losses arise from the change in concentration of the reactants at the surface of the electrodes as the fuel is consumed. Since reactant concentration affects the voltage, this type of irreversibility is sometimes referred to as concentration losses. Because the effect is the result of a failure to transport sufficient reactant to the electrode surface, the term mass‐transport losses is also used. There is even a third name — Nernstian losses — that evolved following modelling of the effects of concentration by the Nernst equation. The four categories of irreversibility are considered, in turn, in the sections that follow. 3.4 Activation Losses 3.4.1 The Tafel Equation In considering the overvoltage at any one electrode, the activation loss (ΔEact) can be defined as: E act E Eeq (3.1) where E is the measured electrode potential and Eeq is the theoretical equilibrium electrode potential. As a result of experiments rather than theoretical considerations, Julius Tafel observed and reported in 1905 that the variation in potential (later to be given Operational Fuel‐Cell Voltages Equation of best-fit line is V = a log (i/io) 0.6 0.5 Overpotential/V 0.4 Fast reaction Slow reaction 0.3 0.2 0.1 0 0 1 2 3 4 5 Log (current density)/mA cm–2 Best-fit line intercepts the current density axis at io Figure 3.3 Tafel plots for slow and fast electrochemical reactions. the term overpotential2) at the surface of an electrode followed a similar pattern for a great variety of electrochemical reactions. This general behaviour, which is displayed in Figure 3.3, shows that if overpotential is plotted against the log of current density, then, for most values of overpotential, the relationship approximates to a straight line. Such a graph is known as a ‘Tafel plot’ and the linear relationship is represented by the expression: E act a log i io (3.2) where a is a constant, commonly referred to as the ‘Tafel slope’, i is the current density and io is the ‘exchange‐current density’, i.e., the current density at zero overpotential or that at which the overpotential begins to manifest itself. The exchange‐current density io can be visualized as follows. The reaction at the oxygen electrode of a proton‐exchange membrane or acid electrolyte fuel cell is: O2 4 e 4H 2H 2 O (3.3) At zero current density, it may be assumed that there is no activity at the electrode, and therefore this reaction does not take place. In fact, this is not so. The reaction is 2 Agar, JN and Bowden, FP, 1938, The kinetics of electrode reactions I and II, Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences, vol. 169 (937), pp. 206–234. 47 48 Fuel Cell Systems Explained occurring, but the reverse reaction is also proceeding at the same rate. There is an equilibrium, which is expressed as: 4 H  2H 2 O O2 4 e (3.4) Thus, there is a continual backwards and forwards flow of electrons from and to the electrolyte that constitutes the exchange‐current density, io. If the value of io is high, then the surface of the electrode can be said to be more ‘active’, leading to a low activation loss when current is drawn. If the value of io is low, the activation overpotential will be high. Equation (3.2) is known as the Tafel equation and can be expressed in many forms. One simple and preferred variation is to use natural logarithms instead of base‐10 logarithms, i.e., Eact A ln i io (3.5) The constant A is higher for an electrochemical reaction that is slow. It is important to remember that the Tafel equation only holds true when i > io. 3.4.2 The Constants in the Tafel Equation Although it was originally deduced from experimental results, the Tafel equation also has a theoretical basis. For a hydrogen fuel cell, the constant A in equation (3.5) is given by: A RT 2 F (3.6) where R is the universal gas constant (= 8.314 472 J K−1 mol−1) and T is the temperature in Kelvin (K). The parameter α is called the ‘charge‐transfer coefficient’ and is the proportion of the electrical energy applied that is harnessed in changing the rate of an electrochemical reaction. Its value depends on the reaction involved and the material used for the electrode, but it must be in the range 0–1.0. For the hydrogen electrode, α is about 0.5 for a wide variety of electrode materials. At the oxygen electrode, the charge‐transfer coefficient shows more variation but is still between about 0.1 and 0.5 in most circumstances. In short, experimenting with different materials to get the best possible value for A will make little impact. The presence of T in equation (3.6) might give the impression that raising the temperature increases the overpotential. In fact, this is very rarely the case as the effect of increases in io with temperature far outweighs any increase in A. Indeed, the key to making the activation overpotential as low as possible is the value of io, as this can vary by several orders of magnitude. Furthermore, io is influenced by several parameters other than the material used for the electrode. In summary, the exchange‐current density is crucial in controlling the performance of a fuel‐cell electrode. Equations (3.5) and (3.6) can be rearranged to describe the cell current as a function of voltage. This is achieved by converting from the logarithmic to the exponential form, to give: i io exp 2 F Eact RT (3.7) Operational Fuel‐Cell Voltages Electrochemists will recognize this as a form of the Butler–Volmer equation that is more fully expressed as: i io exp n aF RT Eact exp n c F Eact RT (3.8) where n is the number of electrons transferred in the electrochemical reaction and αa and αc are the charge‐transfer coefficients at the negative and positive electrodes, respectively. The Butler–Volmer equation is one of the most fundamental equations in electrochemistry as it expresses the current produced by an electrochemical reaction in terms of the rates of reactions at the two electrodes. The equation was derived from kinetic theory and provides a sound basis for its simpler but empirical relative — the Tafel equation, which only holds when the exchange‐current density is very much smaller than the measured current density (i >> io). Even so, the Tafel equation is adequate for understanding and expressing the performance of most practical fuel‐cell systems. For a fuel cell that has no losses except for the activation overpotential on one electrode, the cell voltage would be given by: Vc Vr A ln i io (3.9) where Vr is the open‐circuit voltage given by equation (2.9). Plots of cell voltage (Vc) versus current density (i) obtained using equation (3.9) with values of io of 0.01, 1.0 and 100 mA cm−1 and using a typical value for A of 0.06 V are presented in Figure 3.4. ‘No-loss’ voltage of 1.2 V 1.2 100 1.0 Cell voltage/V 0.8 1.0 0.6 0.01 0.4 0.2 0 0 200 400 600 800 1000 Current density/mA cm–2 Figure 3.4 Cell voltage versus current density, assuming losses due only to the activation overpotential at one electrode, for exchange‐current density io values of 0.01, 1.0 and 100 mA cm–2. 49 50 Fuel Cell Systems Explained The importance of io can be clearly seen. High values of io give the highest actual cell voltages, and low values result in the lowest cell voltages. For most values of current density, the actual cell voltage is fairly constant for each value of io. Note that when io is 100 mA cm−2, there is no voltage drop until the current density i is greater than 100 mA cm−2. It is possible to measure the overpotential at each electrode, either with reference electrodes within a working fuel cell, or by using half cells, as described later. The values of io for the hydrogen electrode at 25°C for various metal substrates are given in Table 3.1; the measurements were conducted on flat smooth electrodes. The great variation in exchange-current indicates that some metals are more catalytically active than others. There is often inconsistency between values obtained by different researchers, which suggests that there are several influencing factors. The io for the cathode also varies appreciably and is generally lower than that for the anode by a factor of about 105. For a cathode, therefore, the exchange-current is of the order of 10−8 A cm−2, even when using a platinum catalyst, i.e., far lower than the lowest curve in Figure 3.4. Fortunately, in practice, the value of io for a fuel‐cell electrode is much higher than those given in Table 3.1 because the roughness of the electrode makes the ‘real’ surface area many times larger (typically, by at least three orders of magnitude) than the nominal length × width. The differences in values of io between the two electrodes reflect the different rates of the reactions that take place on either side of the cell. The hydrogen oxidation reaction (HOR) on the anode is a very fast and simple reaction. By contrast, the oxygen reduction reaction (ORR) on the cathode is many times slower because it is more complex, i.e., several reaction steps are involved. It is generally considered that the overpotential at the anode is negligible compared with that at the cathode, at least in the case of hydrogen fuel cells. Table 3.1 Values of io for the hydrogen electrode for various metals in an acid electrolyte. Metal io (A cm−2) Pb 2.5 × 10−13 Hg 3 × 10−12 Zn 3 × 10−11 Cd 8 × 10−10 Mn 1 × 10−11 Ti 2 × 10−8 Ta 1 × 10−7 Mo 1 × 10−7 Fe 1 × 10−6 Ag 4 × 10−7 Ni 6 × 10−6 Pt 5 × 10−4 Pd 4 × 10−3 Operational Fuel‐Cell Voltages In other fuel cells, for example, the direct methanol fuel cell (DMFC), the overpotential at the anode is by no means negligible. In these systems, the equation for the total activation overvoltage would combine contributions from both electrode polarities, namely, Activation overvoltage Aa ln i ioa Ac ln i (3.10) ioc where ioa and ioc are the exchange‐current densities at the anode and cathode, respectively. This equation can be expressed as: V A ln i b (3.11) where ΔV is the total drop in voltage due to the combined activation overpotentials and A Aa Ac Aa Ac and b ioaA iocA (3.12) Note that the equation (3.12) is only valid for i > b. The relationship mimics equation (3.5), which expresses the overpotential for one electrode. So whether the activation overpotential arises mainly at one electrode only, or both, the equation that models the voltage is of a similar form. Moreover, in all cases, the term in the equation that shows the most variation is the exchange‐current density io, rather than the parameter A. Further discussion of electrode kinetics for different fuel‐cell types will appear in later chapters. 3.4.3 Reducing the Activation Overpotential Improving fuel‐cell performance via increasing the value of io can be accomplished in various ways: ● ● ● ● ● Raising the cell temperature. This action fully explains the different shape of the voltage versus current density graphs of low‐ and high‐temperature fuel cells illustrated in Figures 3.1 and 3.2. For a low‐temperature cell, the io at the positive electrode will be about 0.1 mA cm−2, whereas for a typical 800°C cell, it will be about 10 mA cm−2 — a 100‐fold improvement! Using more effective catalysts. The effect of different metals in the electrode is shown clearly by the data given in Table 3.1 where the precious metals platinum and palladium are much more active for hydrogen activation than base metals such as zinc and lead. In recent years, major efforts have been made to develop superior catalysts through the use of alloys. Increasing the roughness of the electrodes. This technique increases the real surface area of each nominal 1 cm2 that, in turn, enhances the io. Increasing reactant concentration, e.g., using pure oxygen instead of air. Such action enables the catalyst sites to be more effectively occupied by reactants. As demonstrated in Chapter 2, this also increases the open‐circuit voltage. Increasing the pressure. This approach is also considered to be effective through enhancing the reactant occupancy of catalyst sites. Similar to enhancing the reactant concentration, the strategy produces a ‘double benefit’ through increasing the open‐ circuit voltage. 51 52 Fuel Cell Systems Explained The last two points in this list explain the discrepancy between the theoretical and the actual open‐circuit voltage that has been discussed in Section 2.5.4, Chapter 2. It is useful to reflect that the activity of the catalyst, the electrode roughness and the issues of pressure and reactant concentration all exert an influence on the reaction rate and, consequently, on the performance of the fuel cell. The electrode reactions take place at a triple‐phase boundary, and therefore cell performance is highly dependent on the design and distribution of the catalyst and its interaction with the electrode (i.e., the catalyst topology). Greater consideration will be given to such requirements when examining each type of fuel cell in later chapters and will include the introduction of advanced solid‐state materials such as mixed ionic–electronic conductors. 3.5 Internal Currents and Fuel Crossover Although the electrolyte of a fuel cell will have been chosen for its ion‐conducting properties, it will invariably possess some electronic conductivity. Minute internal currents due to conduction of electrons will reduce the cell voltage by a small amount. Probably more important in a practical fuel cell is that some hydrogen will diffuse from the anode, through the electrolyte, to the cathode. The hydrogen will react directly with oxygen on the cathode catalyst to be consumed and thereby generate no current from the cell. The wasted fuel that migrates in this manner through the electrolyte is known as ‘fuel crossover’. The previously mentioned two adverse effects are essentially equivalent. The crossing over of one hydrogen molecule wastes two electrons and amounts to exactly the same as two electrons crossing internally in the opposite direction rather than as an external current. Furthermore, if the major loss in the cell is the transfer of electrons at the interface of the cathode, which is the case for hydrogen fuel cells, then the effect of both these phenomena on the cell voltage is also the same. Internal electron flow or fuel crossover will typically be the equivalent of only a few mA cm−2. In terms of energy loss, the irreversibility is not very important. In low‐ temperature cells, however, it does cause a very noticeable voltage drop under open‐ circuit conditions. Users of fuel cells can readily accept that the working voltage of a cell will be less than the theoretical ‘no loss’ reversible voltage. In an open circuit, however, when no work is being done, it may be expected that the cell voltage will be the same as the reversible voltage. For low‐temperature cells, such as proton‐exchange membrane fuel cells (PEMFCs), when operating on air at ambient pressure, the open‐circuit voltage will usually be at least 0.3 V less than the reversible voltage (~1.2 V), due to internal currents or crossover. If, as in the last section, the losses in a fuel cell are assumed to be caused only by the ‘activation overpotential’ at the cathode, then the cell voltage (Vc) will be reduced only by the amount given by equation (3.9), namely: Vc Vr A ln i io (3.9) For a PEMFC operating at about 30°C and using air at atmospheric pressure, reasonable values for the parameters in equation (3.9) are V = 1.2 V, A = 0.06 V and io = 0.04 mA cm−2. Operational Fuel‐Cell Voltages Table 3.2 PEMFC voltages at low current densities. Current density (mA cm−2) Voltage (V) 0 1.2 0.25 1.05 0.5 1.01 1.0 0.97 2.0 0.92 3.0 0.90 4.0 0.88 5.0 0.87 6.0 0.86 7.0 0.85 8.0 0.84 9.0 0.83 If the internal current density is 1.0 mA cm–2, then the open-circuit voltage will drop to 0.97 V Using these values, the cell voltages for a range of low current densities have been calculated and are listed in Table 3.2. Because of the internal currents, the current density is not zero, even if the cell is at open circuit. For instance, if the internal current density is 2 mA cm−2, then the open‐ circuit voltage would be 0.92 V, i.e., nearly 0.3 V (or 25%) less than the theoretical value. This appreciable loss in voltage is a consequence of the very steep initial fall that is shown by the data in Figure 3.4 (v.s.). The steepness of the curve also explains why the open‐circuit voltage of low‐temperature fuel cells is highly variable. The information given in Table 3.2 and Figure 3.4 demonstrates that a small change in fuel crossover and/ or internal current caused, for example, by a change in the humidity of the electrolyte, can promote a large change in open‐circuit voltage. Obviously, it is not easy to measure the fuel crossover and the internal current — an ammeter cannot be inserted in the circuit! One method, however, is to determine the consumption of reactant gases at open circuit. For single cells and small stacks, the very low rates of gas usage cannot be measured by means of normal gas flow meters, so that bubble counting, gas syringes, or similar have to be employed. For example, at open circuit, a small PEM cell of area 10 cm2 might have a hydrogen consumption of 0.0034 cm3 s−1, at normal temperature and pressure (author’s measurement performed on a commercial cell). According to Avogadro’s law, the volume of 1 mol of any gas is 2.24 × 104 cm3 at standard temperature and pressure (STP), and therefore the gas usage is 1.52 × 10−7 mol s−1. Equation (A2.13) in Appendix 2 shows that the rate of hydrogen fuel usage in a single cell (n = 1) is related to the current (I) by the formula: Gas usage I mol s 2F 1 (3.13) The previously mentioned losses therefore correspond to a current of 1.52 × 10−7 × 2 × 9.65 × 104 = 29 mA. Given that the cell area is 10 cm2, the current density is 2.9 mA cm−2 53 Fuel Cell Systems Explained and is the sum of the current equivalent of fuel lost from crossover and the actual internal current density. If in is the value of this internal current density, then equation (3.9) used to express the cell voltage can be refined to: Vc Vr A ln i in io (3.14) Taking typical values for a low‐temperature cell, namely, V = 1.2 V, A = 0.06 V, io = 0.04 mA cm−2 and in = 3 mA cm−2, yields a graph of cell voltage against current density of the form displayed in Figure 3.5; the relationship is quite similar to that shown in Figure 3.4. The importance of the internal current is considerably less for high‐ temperature cells because the exchange‐current density io is very much greater and, consequently, the initial fall in voltage is less dramatic. 3.6 Ohmic Losses The losses in cell voltage due to the electrical resistance of the electrodes, and to the resistance to the flow of ions in the electrolyte, are the simplest to understand and to model. The size of the voltage drop (ΔV) is simply proportional to current, i.e., as given by Ohm’s law: V (3.15) IR In most fuel cells, the resistance — R in equation (3.15) — mainly emanates from the electrolyte, though the cell interconnects or bipolar plates (see Section 1.3, Chapter 1) can also be important contributors. ‘No-loss’ voltage of 1.2 V 1.2 1.0 0.8 Cell voltage / V 54 0.6 0.4 0.2 0 0 200 400 600 Current density/mA 800 1000 cm–2 Figure 3.5 Fuel‐cell voltage modelled using activation and fuel crossover/internal current losses only. Operational Fuel‐Cell Voltages To be consistent with the other equations for voltage loss, equation (3.15) should be expressed in terms of current density. To do this, it is necessary to introduce the concept of the resistance corresponding to 1 cm2 of the cell. The parameter is called the ‘area specific resistance’ (ASR) and can be represented by the symbol r. The equation for the voltage drop then becomes: (3.16) V ir where i is, as usual, the current density in mA cm−2 and therefore r should be given in kΩ cm2. Using the methods described in Section 3.10, it is possible to distinguish this particular irreversibility from the others. For instance, it can be shown that the ‘ohmic loss’ of voltage is significant in all types of cell and is especially important in the case of the SOFC. Three ways of reducing the internal resistance of a cell are as follows: ● ● ● The use of electrodes with the highest possible conductivity. Optimization of the design and choice of materials for the bipolar plates or cell interconnects. This issue has already been addressed in Section 1.3, Chapter 1. Making the electrolyte as thin as possible. Unfortunately, such an approach is often difficult given that if a solid electrolyte is employed, it sometimes has to be fairly thick as it is the support on which the electrodes are built. Also where the electrolyte is a liquid, e.g., in the alkaline fuel cell, the separation of electrodes has to be sufficiently wide to allow a circulating flow of electrolyte between them. The electrolyte in a SOFC can be made very thin but still must have adequate thickness to prevent internal shorting between electrodes, a requirement that implies a certain level of physical robustness. 3.7 Mass‐Transport Losses If the oxygen at the positive electrode of a fuel cell is supplied in the form of air, then it is self‐evident that during operation, there will be a slight reduction in the concentration of the oxygen in the region of the electrode, as the reactant gas is extracted. The extent of the change in concentration, which reduces the partial pressure of oxygen, will depend on the current being taken from the fuel cell and on physical factors that relate to how well the air around the electrode can circulate and how quickly the oxygen can be replenished. Similarly, if the negative fuel electrode is supplied with a gas mixture that contains hydrogen (such as a reformed gas containing carbon oxides), then there will be a fall in hydrogen partial pressure as the hydrogen is consumed by the cell. Whether addressing a reduction in the absolute pressure or in the partial pressure, the same principles apply, and the net result will be a reduction in voltage. There is no analytical solution to modelling the change in cell voltage as a function of the hydrogen partial pressure. One approach is to revisit the Nernst equation, i.e., use equation (2.40) in Chapter 2: V RT P ln 2 2F P1 (2.40) 55 56 Fuel Cell Systems Explained Note that this particular equation relates the increase in cell voltage due to increasing pressure from P1 to P2. The equation can be used to estimate the voltage drop as a result of the decrease in pressure caused by consumption of the fuel gas as follows. Consider a limiting value of current density, il, at which fuel is consumed at a rate equally to its maximum supply rate. Clearly, the current density cannot rise above this value because the fuel gas cannot be supplied at a greater rate. At this current density, the pressure of the hydrogen supply will have just fallen to zero. If P1 is the pressure when the current density is zero, and it is assumed that the pressure falls linearly down to zero at the current density il, and then the pressure P2 at any current density i is given by: P2 P1 1 i il (3.17) Substitution of this relationship into equation (2.40), given earlier, yields the voltage change due to the concentration (mass‐transport) losses, namely: V RT i ln 1 il 2F (3.18) It should be noted that care must be taken over the signs, i.e., equations (2.40) and (3.18) are written in terms of a voltage gain, and the term inside the brackets is always less than 1. Consequently, the equation for voltage drop should be written as: V i RT ln 1 il 2F (3.19) More generally, the concentration (mass‐transport) losses are given by: V B ln 1 i il (3.20) where B is a parameter that depends on the fuel cell and its operating state. For example, if B is set to 0.05 V and il to 1000 mA cm−2, then quite a good fit is obtained to curves such as those in Figures 3.1 and 3.2. Nevertheless, this theoretical approach has many weaknesses, especially in the case of the vast majority of fuel cells, which are supplied with air rather than oxygen. An alternative way to quantify the voltage loss is to use an empirical relationship, e.g., V m exp ni (3.21) where m and n are constants. Using values of m = 3 × 10−5 V and n = 8 × 10−3 cm2 mA−1, the voltage change predicted by equations (3.20) and (3.21) are very similar. In particular, equation (3.21) is found to give a good fit with voltage losses that are measured experimentally and is widely accepted in the fuel‐cell community. It will be used in the sections that follow. The overvoltage due to concentration (mass‐transport) losses is particularly important in cases where the hydrogen is supplied from a reformer or generator; as such an arrangement might have difficulty in adjusting the rate of supply of hydrogen sufficiently rapidly to meet changes in demand. The nitrogen left behind after oxygen is consumed Operational Fuel‐Cell Voltages at the air electrode can also hinder mass transport at high currents — it effectively blocks the oxygen supply. 3.8 Combining the Irreversibilities It is useful to construct an equation that brings together all the irreversibilities associated with fuel cells. Such an exercise results in the following relationship between operating voltage and current density: Vc Vr i in r A ln i in io B ln 1 i in il (3.22) where: ● ● ● ● ● ● ● Vr is the reversible open‐circuit voltage given by equation (2.9), Chapter 2 in is the sum current density equivalent of fuel crossover and the internal current density, as described in Section 3.5 A is the slope of the Tafel line, as described in Section 3.4.2 io is either the exchange‐current density at the positive electrode if the overpotential is much greater than that of the negative electrode, or it is a function of both exchange‐ current densities, as given in equation (3.11) B is the parameter in the mass-transfer overvoltage equation (3.21), as discussed in Section 3.7 il is the limiting current density at the electrode that has the lowest limiting current density, as discussed in Section 3.7 r is the ASR, as described in Section 3.6 Example values of the constants are given in Table 3.3 for two different types of fuel cell. It is possible to model equation (3.22) by means of a spreadsheet (such as EXCEL), a program such as MATLAB, or a graphics calculator. It must be borne in mind that there may be problems at low current densities, as the third term in the equation is only valid when (i + in) >> io. Also, the equation is not valid when the limiting current density is exceeded, i.e., (i + in) > il. Given these caveats, the reader should be able to generate graphs very similar to those displayed in Figures 3.1 and 3.2 by using the data provided in Table 3.3. Table 3.3 Example values of parameters for equation (3.22). Parameter Low temperature (e.g., PEMFC) High temperature (e.g., SOFC) Vr (V) 1.2 1.0 –2 in (mA cm ) 2 2 r (kΩ cm2) 30 × 10−6 300 × 10−6 io (mA cm–2) 0.067 300 A (V) 0.06 0.03 B (V) 0.05 0.08 il (mA cm–2) 900 900 57 58 Fuel Cell Systems Explained 3.9 The Electrical Double-Layer The inquisitive newcomer to fuel cells is often prompted to ask further about the nature of the processes occurring at the electrodes. Granted that chemical reactions — oxidation and reduction — are occurring at the electrodes, it is pertinent to enquire about the nature of the interaction between the reacting species and the electrode and electrolyte materials at the molecular or the atomic level. To explore this subject, it is necessary to invoke a concept known as the ‘electrical double-layer’. First described by Helmholtz as far back as 1853, the concept has helped to explain the properties of many everyday substances from colloids such as milk or paint to electrical devices such as capacitors and batteries. Whenever two such different materials are in contact, there is a build‐up of electrical charge on the surface at the interface between the materials or a charge transfer from one to the other. In semiconductors, for example, there is a diffusion of positive ‘holes’ and negative electrons across junctions between n‐type and p‐type materials that are in contact. This forms a ‘double-layer’ at the junction (of electrons in the p‐type region and ‘holes’ in the n‐type) that plays a fundamental role in semiconductor devices such as diodes, transistors, photosensors and solar cells. In electrochemical systems, the double-layer forms in part due to diffusion effects (as in semiconductors) associated with the reactions between the electrons in the electrodes and the ions in the electrolyte, and also as a result of applied voltages. For example, the situation depicted in Figure 3.6 might arise at the cathode of a fuel cell with an acid electrolyte. Electrons will collect at the surface of the electrode, and H+ ions will be attracted from the bulk to the surface of the electrolyte. The electrons and ions, together with the oxygen supplied to the positive electrode, will take part in the reaction given by equation (3.4), i.e., Electrode Electroyte Figure 3.6 Charge double layer at the cathode surface of a fuel cell. Operational Fuel‐Cell Voltages O2 4 e 4H (3.4) 2H 2 O The accumulation of positive charge on the surface of the cathode as a result of the migrated H+ ions and a relatively lower charge in the surrounding electrolyte results in the formation of an electrical double-layer. The layer has a complex structure with (i) an inner Helmholtz plane (IHP), which is the layer of absorbed ions on the surface of the electrode (H+ ions in the case of Figure 3.6), and (ii) an outer Helmholtz plane (OHP), which represents the position of the ions in the electrolyte closest to the electrode surface. As discussed later in Chapter 4, all the ions in (i) and (ii) are hydrated ions in PEMFCs. Beyond the OHP, there are ions in the electrolyte that can interact via long‐ range electrostatic forces. The probability of a reaction taking place depends on the density of the charges, electrons and H+ ions on the electrode and electrolyte surfaces. Any collection of charge will generate a difference in electrical potential between the electrode and electrolyte — this is the ‘activation overpotential’, which was discussed in Section 3.4. The layer of charge on or near the electrode|electrolyte interface is a store of electrical energy, and as such behaves much like an electrical capacitor. If the current changes, it will take some time for the charge (and its associated voltage) to dissipate (if the current reduces) or build up (if there is a current increase). Consequently, unlike an ohmic loss in voltage, the activation overpotential does not immediately change with the current. Consider now the combined effect of the overpotentials on two electrodes of a complete fuel cell. If the current through the fuel cell suddenly changes, the operating voltage will show an immediate change due to the internal resistance that is followed by a fairly slow progress to its final equilibrium value. The behaviour can be modelled by using an equivalent circuit, with the double-layer represented by an electrical capacitor. The capacitance of a capacitor, C, is given by the formula: C A d where ε is the electrical permittivity, A is the surface area and d is the separation of the plates. For a fuel cell, A is the real surface area of the electrode, which is several thousand times greater than its length × width. The separation, d, is very small, i.e., typically only a few nanometres. Consequently, the capacitance in some fuel cells will be of the order of a few Farads, which is high in terms of capacitance values. (In electrical circuits, a 1‐μF capacitor is relatively large.) The connection between this capacitance, the charge stored in it and the resulting activation overpotential leads to an equivalent circuit, as shown in Figure 3.7. The resistor Rr simulates the ohmic losses. A change in current gives an immediate change in the voltage drop across this resistor. The resistor Ra models the activation overpotential, and the capacitor ‘smooths’ any voltage drop across this resistor. If the concentration overpotential were to be included, it would be incorporated in Ra. (3.23) Rr Ra E Figure 3.7 Simple equivalent circuit model of a fuel cell. 59 60 Fuel Cell Systems Explained Generally speaking, the capacitance that results from the double layer gives the fuel cell a ‘good’ dynamic performance, in that the voltage moves gently and smoothly to a new value in response to a change in current demand. It also permits a simple and effective way to distinguish between the main types of voltage drop, and hence to analyse the performance of a fuel cell, as described in the next section. 3.10 Techniques for Distinguishing Irreversibilities At various points in this chapter, it has been asserted that a certain distinctive type of overpotential/overvoltage is dominant under a given condition. For example, it has been said that for an SOFC, the ohmic voltage drop is more important than activation losses. Much of the evidence that supports this assertion comes from fundamental experimental measurements. The following describes some of the techniques that are frequently employed for experimentally characterizing electrochemical cells — first those relating to individual electrodes and then those that are applied to complete cells. 3.10.1 Cyclic Voltammetry Cyclic voltammetry (CV) is widely employed in the investigation of electrochemical reactions on individual electrodes. Most commonly, a three‐electrode cell is used with a liquid electrolyte, as illustrated in simplified form in Figure 3.8. The setup is comprised of the following components: ● ● A ‘working electrode’ that usually consists of a highly-polished, glassy carbon substrate on which the electrode or catalyst material to be investigated is deposited. A ‘counter electrode’, usually a flag of platinum of sufficient area to ensure that any electrochemical reaction occurring at this electrode, does not influence the performance of the working electrode. Working electrode Counter-electrode Provision for O2 gas to saturate the electrolyte solution Reference electrode (e.g., Calomel) Luggin capillary Cell contains electrolyte solution Figure 3.8 Simple 3‐electrode setup for cyclic voltammetry. Operational Fuel‐Cell Voltages ● ● A ‘reference electrode’ against which voltage measurements are made; examples are Pt|H2|H+ (SHE), Hg|Hg2SO4 (mercury/mercurous sulfate), Ag|AgCl|Cl– (silver|silver chloride) and Hg|Hg2Cl|Cl– (saturated calomel electrode). The reference electrode is located close to the working electrode, usually via a small Luggin capillary. Provision may be made for admitting oxygen to the electrolyte solution, e.g., for performing CV for the ORR on PEMFC catalysts. The principle of CV operation is as follows. The material of interest, e.g., a carbonsupported platinum catalyst material for the negative electrode of a PEMFC, is prepared as a fine powder and dispersed in a solvent such as dilute ethanol. A material, e.g., NafionTM, may be added to facilitate good adhesion to the electrode. The mixture is finely dispersed by agitation or ultrasonication, deposited on the surface of the working electrode and then allowed to dry in air. Using, typically, a dilute sulfuric acid solution (0.01–0.1 M) as the liquid electrolyte, the cell is assembled and the experiment commenced. A potential difference is applied between the working and reference electrodes and scanned at a fixed rate towards higher or lower values, as dictated by the reaction of interest. The current flowing between the working and counter electrodes is recorded as a function of the applied voltage. Voltage is controlled and measured in CV experiments by means of a potentiostat, which is an instrument that draws no current from the reference electrode. When the reaction at the electrode is complete, the voltage is scanned in the opposite direction (hence the term ‘CV’). If the reaction of interest is reversible, then the reverse sweep will show this as a current flowing in the opposite direction. The plot of current versus applied voltage is known as a ‘cyclic voltammogram’; examples will be examined later in the chapters devoted to low‐temperature fuel cells. A voltammogram provides information about the oxidation–reduction potential and the rates of the electrochemical reactions occurring at a given electrode. The technique is particularly valuable in allowing measurement of the activity of fuel‐cell catalysts without the need to assemble complete fuel cells or half cells. The rotating disc electrode (RDE) is an extension of the CV method. This device employs the same three‐electrode experimental setup as for CV, except that the working electrode is able to rotate at high speeds. If the electrochemical reaction occurring on the surface of the electrode is limited by diffusion, this is shown by a change in the voltammogram as the speed of rotation increases. Above a certain speed, the effects of diffusion to the surface are minimized. The RDE technique can be employed to probe the reaction mechanism, for example, to distinguish between 2‐ and 4‐electron transfer with materials employed as cathode catalysts in PEMFCs. A slight variant of the RDE is the rotating ring–disc electrode (RRDE), which enables reaction mechanisms to be elucidated in more detail. Unlike the CV method, in which the electrode is stationary and the reaction is essentially reversed in the return sweep, this does not take place with the RDE/RRDE techniques since the surface layer on the catalyst is disturbed by rotation of the electrode. With rotating electrodes, it is therefore only possible to carry out linear voltage sweeps and not cyclic sweeps as in CV. 3.10.2 AC Impedance Spectroscopy AC impedance spectroscopy, sometimes referred to as electrochemical impedance spectroscopy (EIS), has become a popular means for characterizing both half-cells and complete fuel cells. In contrast with most other electrochemical methods, this 61 62 Fuel Cell Systems Explained technique can be applied in situ to a working fuel cell. It is fairly straightforward to understand in principle, but care must be taken in the analysis of data since many factors can affect the results. The procedure essentially involves driving a small variable‐frequency alternating current (AC) through the fuel cell and measuring the resulting AC voltage across the cell, from which the impedance of the cell can be determined. Since the AC frequency can be quite low, it is important that the fuel cell is operating under steady conditions, for example, in the absence of catalyst activation or deactivation. As with internal resistances, several impedances can be distinguished in a working cell and attributed to the electrolyte, electrodes and interfaces.3 The essentials of AC impedance spectroscopy have been known since the 1950s, but it is really since the emergence of advanced computing systems and frequency response analyzers (FRAs) in the 1980s that the technique has achieved a routine status in electrochemistry. The FRA generates a reference voltage sine wave of given amplitude and frequency, and the magnitude and phase of the resulting AC current is measured and recorded. Sweeping a range of frequencies gives an ‘impedance spectrum’, which can be presented as a ‘Bode plot’ of current versus frequency. The technique is capable of a high degree of precision since unwanted signals in the spectrum can be filtered out by carrying out measurements over a large number of cycles. The measured AC current through the fuel cell is phase‐shifted with respect to the applied AC voltage sine wave by a phase angle θ. If a radial frequency ω (measured in radians per second) is defined as: (3.24) 2 f where f is the frequency (in Hertz) of the applied voltage, an expression analogous to Ohm’s law for resistors can be derived, namely: Z Et It Eo sin t ) I o sin( t Zo sin t sin t (3.25) where Z is the impedance of the system, Et and It are the voltage and current at time t and Zo is the impedance of the system when both current and voltage are in phase (θ = 0). It is also possible to show that the impedance can be represented by a complex number, i.e., Z Zo cos j sin (3.26) Mathematically, this means that Z can be represented by a real and imaginary component. Plotting the real part (Zre) on the x‐axis and the imaginary part (Zimag) on the y‐axis of a chart produces a so‐called ‘Nyquist plot’ — an example obtained for a 3 The total internal resistance of a working fuel cell is the sum of the resistances of the various cell components. It varies with current density, and, as with a battery, the maximum power delivered by a fuel cell is achieved when the total internal resistance is equal to the sum of the resistances in the external circuit. The magnitude of the internal resistance is the same as the magnitude of the cell impedance as measured by EIS. Impedance, however, also has a phase dimension as alternating current is involved. Operational Fuel‐Cell Voltages (a) 0.75 0.00 E2 E3 E4 E-2 –Zim /Ω cm2 0.25 E1 0.50 100 0.50 0.75 1.00 1.25 1.50 50 0 0 50 100 150 Zre /Ω cm2 (b) 10–3 10–1 101 103 –75 –50 10 ϕ/degrees |Z| /Ω cm2 100 –25 1 0 10–3 10–1 101 103 f/Hz Figure 3.9 (a) Nyquist plot of a practical SOFC anode in 97% CH4 and 3% H2O, at 932°C. Inset: Zoom at frequencies higher than 1 Hz. (b) Corresponding Bode plots. (Source: Kelaidopoulou, A, Siddle, A, Dicks, AL, Kaiser, A and Irvine, JTS, 2001, Anodic behaviour of Y0.20Ti0.18Zr0.62O1.90 towards hydrogen electro‐oxidation in a high temperature solid oxide fuel cell, Fuel Cells, vol 1(3–4), pp. 226–232.) anode in an SOFC is shown in Figure 3.9a. The advantages of presenting data in this form are as follows: ● ● ● Ohmic resistance (Rr) is displayed in the left‐hand region where the semicircle reaches (or extrapolates to the x‐axis); this represents the point of zero frequency. In the example under consideration, Rr = 0.48 Ω cm2. The y‐axis represents the capacitive elements of the cell. Activation‐controlled processes with distinct time constants show up as unique impedance arcs, and the shape of the curve provides insight into a possible reaction mechanism or governing phenomena. In the example shown in Figure 3.9a, two processes were detected, namely, (i) at high frequency, there was a charge‐transfer 63 Fuel Cell Systems Explained ● process (Ox + ne− → red), and (ii) at low frequency, careful deconvolution of the curve distinguished three separate resistive contributions.4 The main disadvantage of the Nyquist plot is that frequency is not directly plotted so it is difficult to determine the frequency of a point on a Nyquist plot. This can be overcome by displaying data in the form of a Bode plot, e.g., Figure 3.9b, in which the impedance (either the real or imaginary component) or phase angle is plotted versus frequency. It is possible to fit AC impedance curves to an equivalent electrical circuit made up of individual resistors and capacitors, such as shown in Figure 3.7. Computer software is now available that will fulfil this task. Impedance (Nyquist) plots shown in Figure 3.10a are typical spectra for the ORR at a cathode platinum|Nafion interface of a PEMFC at different potentials.5 There are two pronounced arcs, accounting for the charge transfer at high frequencies and masstransfer processes at low frequencies. The charge‐transfer arc decreases at higher voltages due to the increased rate of the electrochemical reaction, and the arc due to mass‐transport impedance becomes more dominant. The representative equivalent (a) 100 –Z″/Ω cm2 64 50 0 0 50 100 150 200 Z′/Ω cm2 (b) CPE RΩ Rct WS Figure 3.10 (a) Effect of overpotential on impedance plots for Nafion 117. Applied DC potential: (ο) 0.775 V, (▲) 0.75 V, (∙) 0.725 V and (•) 0.70 V. Temperature 303 K and oxygen pressure 207 kPa. Solid lines represent fits of the equivalent circuits. (b) Typical equivalent circuit of PEMFC for ORR at Pt/Nafion interface. (Source: Xie, Z and Holdcroft, S 2004, Polarization‐dependent mass transport parameters for ORR in perfluorosulfonic acid ionomer membranes: an EIS study using microelectrodes, Journal of Electroanalytical Chemistry, vol 568, pp. 247–260. Reproduced with the permission of Elsevier.) 4 Kelaidopoulou, K, Siddle, A, Dicks, AL, Kaiser, A and Irvine, JTS, 2001, Methane electro‐oxidation on a Y0.2Ti0.18 Zr0.62O0.19 anode in a high temperature solid oxide fuel cell, Fuel Cells, vol. 1(3–4), pp. 219–225. 5 Yuan, X, Wang, H, Sun, JC and Zhangm, J, 2007, AC impedance technique in PEM fuel cell diagnosis—A review, International Journal of Hydrogen Energy, vol. 32, pp. 4365–4380. Operational Fuel‐Cell Voltages circuit is depicted in Figure 3.10b to simulate the typical plot, where R, Rct and Ws represent the ohmic resistance, charge‐transfer resistance and finite‐length Warburg impedance, respectively. The last-mentioned represents diffusion of the reacting species (in this case, oxygen). The conventional double‐layer capacitance is replaced by a constant phase element (CPE) because the capacitance caused by the double‐layer charging is distributed along the length of the pores in the porous electrode. The circuit of Figure 3.10b is known as the Randles circuit and is one the simplest commonly used model of fuel‐cell electrodes. By carrying out AC impedance on both the anode and the cathode, the user may be able to determine the contributions made to the overpotentials by individual materials or processes occurring within the cell. Usually, impedance spectra are more complex than that illustrated in Figure 3.10; procedures for their interpretation are outside the scope of this book. Fortunately, AC impedance has become a more popular tool in the characterization of fuel‐cell systems, and consequently there are many references on the experimental techniques and interpretation of data.6 3.10.3 Current Interruption The current‐interrupt technique not only provides accurate quantitative results but also delivers rapid qualitative indications of internal losses in working fuel cells. Unlike impedance spectroscopy, it can be performed using standard, low‐cost, electronic equipment; the basic setup is shown in Figure 3.11. To understand current interruption, consider a fuel cell that is providing a current at which the concentration (or mass transport) overpotential is negligible, and, therefore, the voltage drop is caused by ohmic losses and the activation overpotential. If the current is suddenly cut off, the charge double layer will take some time to disperse and so will the associated overpotential. By contrast, the ohmic losses will immediately reduce to zero. The resulting change in voltage measured at the fuel‐cell terminals when a load is suddenly disconnected is shown schematically in Figure 3.12. Fuel cell A Digital storage oscilloscope Figure 3.11 Simple circuit for performing a current‐interrupt test. 6 Example: ‘Basics of Electrochemical Impedance Spectroscopy’, published by Gamry Instruments — available online at http://www.gamry.com/application‐notes/EIS/basics‐of‐electrochemical‐impedance‐spectroscopy/ 65 66 Fuel Cell Systems Explained Voltage Slow final rise to OCV,Va Immediate rise in voltage, Vr OCV– open-circuit voltage Va – activation overpotential Vr - voltage rise due to ohmic losses Time Time of current interruption Figure 3.12 Schematic of voltage against time for a fuel cell after a current‐interrupt test. Measurement of current interruption is as follows. The switch in the circuit illustrated in Figure 3.11 is closed, and the load resistor is adjusted until the desired test current is flowing. The storage oscilloscope is set to a suitable time base, and the load current is then switched off. The oscilloscope triggering will need to be set so that the instrument moves into ‘hold’ mode — though with some cells, the system is so slow that the procedure can be done by hand. The two voltages Vr and Va, shown in Figure 3.12, are then read off the screen. Although the method is simple, care must be taken when obtaining quantitative results as it is possible to overestimate Vr by missing the point where the immediate rise in voltage ends. The setting of the oscilloscope time base will vary for different types of fuel cell, as determined by the capacitance. The current‐interrupt test is easy to perform with single cells and small fuel‐cell stacks. With larger cells and stacks, the switching of high currents can be problematic. Typical results from three current‐interrupt tests, as shown in Figures 3.13, 3.14 and 3.15, provide a clear qualitative indication of the importance of the different types of voltage increases that can be observed. Because oscilloscopes do not show vertical lines, the appearance of each trace is slightly different from that given in Figure 3.12, namely, there is no vertical line corresponding to Vr. The tests were performed on three different types of fuel cell: a PEMFC, a DMFC and a SOFC. In each case, the total voltage drop is about the same (Vr + Va), though the current density certainly is not. The three examples give a good summary of the causes of voltage losses in fuel cells. Concentration or mass‐transport losses are important only at higher currents, whereas in a well‐designed system with a good supply of fuel and oxygen, they should be very small over the range of cell operating currents. In low‐temperature hydrogen fuel cells, anode activation can be ignored, and the dominant voltage loss is due to the activation losses at the cathode, especially at low currents (below about 50 mA cm−2). At higher currents, namely, above about 50 mA cm−2, the activation and the ohmic losses are similar — see Figure 3.13. In cells using fuels such as methanol, there are considerable activation losses at both the anode and cathode, and therefore the activation overpotential Operational Fuel‐Cell Voltages Figure 3.13 Current‐interrupt test for a low‐temperature, ambient pressure, hydrogen fuel cell. Ohmic and activation overpotentials are similar. (Time scale 0.2 s per division; i = 100 mA cm−2.) Va Vr Figure 3.14 Current‐interrupt test for a direct methanol fuel cell. There are large activation losses at both electrodes so that, by comparison, the ohmic losses are barely discernable. (Time scale 2 s per division; i = 10 mA cm–2.) Va Vr Figure 3.15 Current‐interrupt test for a small SOFC working at about 700°C. The large immediate rise in voltage shows that ohmic losses are responsible for most of the voltage drop. (Time scale 0.02 s per division; i = 100 mA cm–2.) Va Vr 67 68 Fuel Cell Systems Explained dominates at all times, as demonstrated in Figure 3.14. On the other hand, the activation losses become much less dominant in cells that operate at high temperatures, such as SOFCs at 700°C, and thereby ohmic losses become the main concern, as shown Figure 3.15. The aim of the opening three chapters has been to provide a sound understanding of the general principles of fuel‐cell operation. The following chapters delve more deeply into the construction, operation and application of the main types of fuel‐cell system. Further Reading Büchi, FN, Marek, A and Scherer, GG, 1995, In‐situ membrane resistance measurements in polymer electrolyte fuel cells by fast auxiliary current pulses, Journal of The Electrochemical Society, vol 142(6), pp. 1895–1901. Greef, R, Peat, R, Peter, LM, Pletcher, D and Robinson, J, 2002, Instrumental Methods in Electrochemistry, Ellis Horwood, Oxford. ISBN‐13: 978‐1898563808. Hamann, CH, Hamnett, A and Vielstich, W, 2007, Electrochemistry (Second, Completely Revised and Updated Edition), Wiley‐VCH, Weinheim. ISBN: 978‐3‐527‐31069‐2. Yuam X‐Zi, Song, C, Wang, H and Zhang, J, 2010, Electrochemical Impedance Spectroscopy in PEM Fuel Cells, Springer, London. ISBN: 978‐1‐84882‐845‐2 (Print) 978‐1‐84882‐846‐9 (Online). Zhang, J, (ed.), 2008, PEM Fuel Cell Electrocatalysts and Catalyst Layers, Fundamentals and Applications, Springer‐Verlag, London. ISBN 978‐1‐84800‐935‐6; DOI 10.1007/978-1-84800-936-3. Zhang, J, Wu, J, Zhang, H and Zhang, J, 2013 PEM Fuel Cell Testing and Diagnosis, Elsevier, Burlington, VT. ISBN 978‐0‐44453‐689‐1. 69 4 Proton‐Exchange Membrane Fuel Cells 4.1 Overview The proton‐exchange membrane fuel cell (PEMFC) — also called the ‘polymer electrolyte membrane fuel cell’ (PEMFC), ‘solid polymer electrolyte fuel cell (SPEFC)’ and solid polymer fuel cell (SPFC) — is the most widely known version of acid fuel cell. The design was first developed by General Electric (GE) in the United States for use by the National Aeronautics and Space Administration (NASA) in the Gemini manned spacecraft of the 1960s. Instead of the liquid proton‐conducting electrolyte used in early experimental acid fuel cells, a solid or quasi‐solid ‘membrane’ material was used. The first PEMFCs employed electrolytes that were based on polymers such as polyethylene; for instance, the initial NASA fuel cells operated with polystyrene sulfonic acid. In 1967, DuPont introduced a novel fluorinated polymer based on a polytetrafluoroethylene (PTFE) structure with the trademark Nafion™ (hereafter, simply referred to as ‘Nafion’). The PTFE material is used to coat non‐stick cookware and is highly hydrophobic (i.e., it is not wetted by water). Nafion constituted a major advance for fuel cells, and it has become an industry standard against which new polymer membranes are judged. Nevertheless, Nafion is expensive to produce and has certain limitations, such as the need to be hydrated and therefore is functional only below about 80°C. For these reasons, many alternative electrolyte materials have been investigated; the more common varieties are reviewed in Section 4.2. Apart from the solid oxide fuel cell (SOFC) described later in Chapter 9, the PEMFC is unique in that it uses a solid sheet of electrolyte that is bound on both sides to sheets of catalysed porous electrodes. The negative|electrolyte|positive assembly is thus one item, is very thin and is commonly referred to as the ‘membrane‐electrode assembly’ (MEA). A PEMFC stack comprises several such MEAs connected in series, usually by means of bipolar plates, as shown in Figure 1.9, Chapter 1. The charge carrier in the polymer electrolyte is an H+ ion (also known as a proton), and the basic operation of the cell is essentially as described for the generic acid electrolyte fuel cell illustrated in Figure 1.3, Chapter 1. The usual polymer membranes operate at near‐ambient temperatures. This enables the PEMFC to start up quickly. The absence of corrosive and hazardous fluids that are present in the electrolytes of alkaline (AFC), phosphoric acid (PAFC) and molten Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 70 Fuel Cell Systems Explained Figure 4.1 Four PEMFC stacks that illustrate development by Ballard Power Systems through the 1990s. The left‐hand stack, the 1989 model, has a power density of 100 W L−1. The right‐hand, 1996 model delivers 1.1 kW L−1 (Source: By kind permission of Ballard Power systems.) carbonate (MCFC) fuel‐cell systems means that the PEMFC is able to function in any orientation. Furthermore, the thinness of modern MEAs also enables the production of compact fuel cells with very high power densities (W L−1). These attributes combine to make the PEMFC very robust and especially suitable for use in road vehicles and as a power source for portable electrical and electronic applications. The early versions of the PEMFC, as used in the Gemini spacecraft, had a lifetime of only about 500 h, but that was sufficient for those limited early missions. Concern arose, however, over the reliability of water management in the electrolyte (which is considered in some detail in Section 4.4), such that NASA selected the AFC for use in the following Apollo spacecraft. General Electric also chose not to pursue commercial development of the PEMFC, probably because the costs were seen as higher than those for other types of fuel cell, such as the PAFC, which was under development for stationary power applications. At that time, catalyst technology was such that 28 mg of platinum were required for each cm2 of electrode — compared with 0.2 mg cm−2 or less today. The development of PEMFCs passed, more or less, into abeyance in the 1970s and early 1980s. A renaissance of interest began in the latter half of the 1980s and early 1990s, and much of the credit for this must go to Ballard Power Systems of Vancouver, Canada, and to the Los Alamos National Laboratory in the United States. Developments in more recent years have brought current densities up to 1 A cm−2 or more, while at the same time the amount of platinum used in the catalysts has been reduced by two orders of magnitude. These improvements have led to a huge reduction in cost per kW of power, and a major increase in power density, as demonstrated in Figure 4.1, which shows the progress made by Ballard Power Systems during the 1990s. Both the specific power (W kg−1) and area‐specific power density (W cm−2) of the PEMFC are higher than for any other type of fuel cell. It is also worth noting that the 2010 performance targets of 650 W kg−1 and 650 W L−1 set by the United States Proton‐Exchange Membrane Fuel Cells Figure 4.2 Honda FC stack and Gearbox (exhibited at the Tokyo Motor Show 2007). Department of Energy (US DOE) for an 80‐kW PEMFC stack were achieved in 2006 by Honda with a novel 100‐kW vertical‐flow stack that is used in the current FCX Clarity car. The stack has a volumetric power density of almost 2.0 kW L−1 and a specific power of 1.6 kW kg−1. In 2008, Nissan also claimed to have achieved 1.9 kW L−1. Since then, Honda has increased the power density to over 3 kW L−1 (Figure 4.2).1 Proton‐exchange membrane fuel cells are being developed for a very wide range of operations. For instance, systems as small as a few watts are being marketed for charging mobile phones and other consumer electronic devices, stationary units of several kW are now in service as for remote telecommunications towers and data centres2, and others are employed as the power source for domestic‐scale combined heat and power (CHP) ‘cogeneration’ systems. Nonetheless, their application in road vehicles, such as cars and buses, has brought the attention of PEMFCs to a greater public. It could be argued that PEMFCs exceed all other electricity generators in the range of their possible uses. In all applications, the three most important distinguishing features of PEMFCs are as follows: ● ● ● The type of electrolyte (Section 4.2). The electrode structure (Section 4.3). The catalyst (Section 4.3). 1 It is worth cautioning that power density is not the only parameter by which a stack should be judged. Increasing the current density may improve power density but at the expense of stack lifetime. Therefore, an important consideration is the required duty and lifetime of the stack for each particular application. 2 Ballard Power Systems has built a 1‐MW stationary power generation system. 71 72 Fuel Cell Systems Explained Other aspects of system design vary greatly depending on the end use, boundary conditions and skills of the designer. The most important of these are as follows: Water management (Section 4.4). The method of cooling (Section 4.5). The method of connecting cells in series — bipolar plate designs vary greatly, and some fuel cells employ markedly different methods (Section 4.6). The operating pressure (Section 4.7). The reactants used — pure hydrogen is not the only possible fuel, and oxygen can be used instead of air (briefly discussed in Section 4.11). ● ● ● ● ● Some examples of PEMFC systems are examined in Section 4.9. Over and above technical issues of the PEMFC, cost is the perhaps the most challenging of the barriers to widespread commercialization. The first commercial PEMFC systems are now on the market for around US$3000 per kW, i.e., significantly greater than alternative power sources such as engine‐ or turbine‐based generators. Widely reported cost targets for stationary PEMFC systems of around US$1000 per kW continue to be a challenge and are discussed further elsewhere.3 4.2 Polymer Electrolyte: Principles of Operation 4.2.1 Perfluorinated Sulfonic Acid Membrane Nafion, for many years the industry standard membrane electrolyte in PEMFCs, is a particular type of perfluorinated sulfonic acid (PFSA). The starting material for Nafion is the synthetic polymer commonly known as polyethylene or simply polythene. The molecular structures of ethylene and polyethylene are shown in Figure 4.3. Polyethylene is modified by chemically substituting fluorine for the hydrogen atoms, to create a ‘perfluorinated’ polymer. The modified polymer, shown in Figure 4.4, is H H C C H H H H H H H H H H H H H H H H H C C C C C C C C C C C C C C C H H H H H H H H H H H H H H H Ethylene Polyethylene (or polythene) Figure 4.3 Chemical structure of polyethylene. F F C F C F Tetrafluoroethylene F F F F F F F F F F F F F F F C C C C C C C C C C C C C C C F F F F F F F F F F F F F F F Polytetrafluoroethylene (PTFE) Figure 4.4 Chemical structure of PTFE. 3 Staffell, I and Green, R, 2013, The cost of domestic fuel cell micro‐CHP systems, International Journal of Hydrogen Energy, vol. 38(2), pp. 1088–1022. Proton‐Exchange Membrane Fuel Cells PTFE; ‘tetra’ indicates that all four hydrogen atoms in each ethylene group have been replaced by fluorine. First produced in 1938 and sold by the DuPont Corporation under the trade name TeflonTM, this remarkable material has exerted a key influence in the development of fuel cells. The strong bonds between the fluorine and carbon atoms make PTFE exceptionally resistant to chemical attack and thereby very durable. Moreover, PTFE is also strongly hydrophobic (i.e., it repels water). Consequently, it is used in fuel‐cell electrodes to drive product water out of the electrode and thereby prevent flooding. For the same reason, PTFE is also employed in AFCs and PAFCs. To make an ion‐conducting electrolyte, PTFE requires further chemical modification, namely, it has to be ‘sulfonated’. This treatment adds side-chains to the PTFE molecular backbone, and each of these is terminated with a sulfonic acid (─SO3H) group; there are several procedures and are mostly proprietary to the membrane manufacturers. An example of a side-chain structure is given in Figure 4.5 — the details vary both for different types of Nafion and other PFSAs. In contrast to the creation of side-chains, the sulfonation of complex molecules is a widely adopted and understood chemical process. It is used, for example, in the manufacture of detergent. In practice, Nafion is terminated with side-chains of SO3− ions that are balanced by Na+ ions. In other words, Nafion may more accurately be considered to be a sodium salt. The ─SO3H groups that feature for use in PEMFCs are generated by boiling the Nafion with concentrated sulfuric acid in a final preparative step during which the sodium is discarded as sodium sulfate. When the sulfonated polymer is converted to the acidic form, the ─SO3H group is ionic and so the end of the side-chain is actually an SO3− ion, in which the sulfur atom is bound to the carbon chain. For this reason, the resulting polymer structure possesses ionic character and is called an ‘ionomer’. Due to the presence of SO3− and accompanying H+ ions, there is a strong mutual attraction between the positive and negative ions from each molecule. Consequently, the side-chains tend to ‘cluster’ within the overall structure of the material. A key property of sulfonic acid is that it is highly hydrophilic, that is, it attracts water.4 In Nafion, therefore, the effect is that hydrophilic regions are Figure 4.5 Chemical structure of a sulfonated fluoroethylene, also called ‘perfluorosulfonic acid PFTE copolymer’. F F F F F F F F F F F F F F F C C C C C C C C C C C C C C C F F F F F F F O F F F F F F F F C F F C F O F C F F C F O S O O– H+ 4 This is why most detergents are sulfonates. In a detergent molecule, such as an alkylbenzene sulfonate, the ionic sulfonic end of the molecule mixes readily with water, whereas the polar end of the molecule (the alkylbenzene) is attracted to the fat, grease and dirt. 73 74 Fuel Cell Systems Explained Table 4.1 Structure of Nafion and characteristics of other PFSAs. ( CF2CF2)x (CF2CF )y ( OCF2CF )m O (CF2 )n SO3H CF3 Structure parameter m = 1, x = 5–13.5, n = 2, y = 1 Trade name and type Equivalent weight Thickness (µm) Nafion 120 1200 260 Nafion 117 1100 175 Nafion 115 1100 125 Nafion 112 1100 80 Flemion ‐ T 1000 120 Flemion ‐ S 1000 80 Dupont Asashi Glass m = 0, 1, n = 1–5 Flemion ‐ R 1000 50 m = 0, n = 2–5, x = 1.5–14 Asashi Chemicals Aciplex ‐ S 1000–1200 25–100 m = 0, n = 2, x = 3.6–10 Dow Chemical Dow 800 125 Source: Lee, JS, Quan, ND, Hwang, JM et al., 2006, Polymer electrolyte membranes for fuel cells. Journal of Industrial Engineering Chemistry, vol. 12(2), pp. 175–183. created within a generally hydrophobic substance. As mentioned earlier, Nafion is a specific type of PFSA, and there are many other PFSAs that have been used as fuel‐cell membranes. Some examples are shown in Table 4.1. The hydrophilic regions around the clusters of sulfonated side-chains in Nafion and other PFSAs can lead to the absorption of large quantities of water, which can increase the dry weight of the material by up to 50%. Within these hydrated regions, the H+ ions are weakly attracted to the SO3− groups and are therefore mobile; essentially a dilute acid is created. The resulting material has different microdomains within the macromolecular structure, namely, dilute acid regions, in which H+ ions are attached to water molecules to create hydronium ions (H3O+), within a tough and strong hydrophobic structure, as illustrated in Figure 4.6. Although the hydrated regions are somewhat separate, it is still possible for the H+ ions to move through the supporting long molecule structure. The proton conductivity of membranes is, nevertheless, higher than would be expected by simple migration of H3O+ ions. This has led to the view that proton conduction is via a Grotthus mechanism in which H+ ions move by ‘hopping’ from one water cluster to the next, a process made easy by the weak hydrogen bonds that have to Proton‐Exchange Membrane Fuel Cells Water collects around the clusters of hydrophylic sulfonate sidechains Figure 4.6 Structure of PFSA‐type membrane materials: long‐chain molecules containing hydrated regions around the sulfonated side-chains. be made and broken with each ion movement. The mechanism was confirmed by experimental studies reported in 2006.5 For application in fuel cells, Nafion and other PFSA ionomers offer the following attractive features in that they are: ● ● ● ● ● Resistant to chemical attack and stable in both oxidizing and reducing environments. Mechanically strong, on account of the durable PTFE backbone, and so can be made into very thin films, down to 50 µm. Acidic. Able to absorb large quantities of water. Good proton conductors when well hydrated, to allow H+ ions to move quite freely within the material. The ionic conductivity of Nafion depends not only on the degree of hydration, which is influenced by the temperature and operating pressure, but also on the availability of the sulfonic acid sites. For example, the conductivity of Nafion membranes quoted in the literature varies widely depending on the system, pretreatment and equilibrium parameters used. At 100% relative humidity (RH), the conductivity is generally between 0.01 and 0.1 S cm−1 and falls by several orders of magnitude as the humidity decreases. Therefore, the degree of hydration has a very marked influence on the ionic conductivity of the membrane and thereby on performance of the fuel cell. In contrast, the availability of sulfonic acid sites, usually expressed as the membrane equivalent weight6 (EW), is relatively unimportant. Values of EW between 800 and 1100 (equivalent to acid capacities of between 1.25 and ~0.90 mEq g−1) are acceptable for most 5 Tushima S, Teranishi K and Hirai S, 2006, Experimental elucidation of proton-conducting mechanism in a polymer electrolyte fuel cell by nuclei labelling MRI, ECS Transactions, vol. 3(1), pp. 91–96. 6 Membrane equivalent weight (EW) is defined as the weight of polymer (in terms of molecular mass) per sulfonic acid group. Ion‐exchange capacity, or acid capacity for PFSAs, is the reciprocal of EW. 75 76 Fuel Cell Systems Explained membranes because investigations have shown that the maximum ionic conductivity can be obtained in this range. It may be also expected that the proton conductivity of a PFSA can be improved by reducing the thickness of the material. In addition to thickness, however, the proton conductivity depends on the water content and structural variables such as porosity, tortuosity, distribution of protons and various diffusion coefficients for the proton conduction processes. Therefore, whereas making thinner membranes may improve conductivity, other factors should be taken into consideration such as the fact that thin materials are inherently less robust and small amounts of fuel crossover can occur with consequent reduction in the observed cell voltage. For these reasons, membrane thicknesses of between 80 and 150 µm have been found to be optimum for most PEMFCs. Despite being used widely by developers of fuel cells, PFSA membranes suffer from two major disadvantages, namely: (i) high cost, due to the inherent expense of the fluorination step in the synthesis of the ionomer and (ii) inability to operate above about 80°C at atmospheric pressure due to evaporation of water from the membrane. With respect to the latter, higher operating temperatures can be achieved by running the cells at elevated pressures, but this has a negative effect on system efficiency due mainly to the additional electrical power required to pressurize the gases. Above 120–130°C, the PFSA materials undergo a glass transition (i.e., a structural change from an amorphous plastic phase to a more brittle state) that also severely limits their usefulness. Membranes that could operate at higher temperatures without the need for pressurization could therefore bring the following significant benefits: ● ● ● ● Carbon monoxide concentrations in excess of about 10 ppm at low temperatures (<80°C) will poison the electrocatalyst used in the Nafion‐based PEMFC. As the operating temperature increases, so the carbon monoxide tolerance of the platinum catalyst improves. Operating at high temperatures has the advantage of creating a greater driving force for more efficient cooling of the stack. This is particularly important for transport applications to reduce the need for balance‐of‐plant equipment (e.g., radiators). High‐grade exhaust heat from high‐temperature fuel cells can be useful for fuel processing, or in CHP applications. Increasing the operating temperature will allow the use of a catalyst with lower activity. Thus the cost penalty associated with expensive platinum catalysts could be reduced or even avoided. 4.2.2 Modified Perfluorinated Sulfonic Acid Membranes It was noted earlier that although thin membranes may bring the advantage of reduced internal ionic resistance, the need to have mechanically strong materials limits how thin they can be made. To overcome this restriction, some materials — for example, the Gore Select™ membrane that is favoured by some PEMFC developers — are formed using a very thin microporous‐based material of expanded PTFE into which an ion‐ exchange resin is incorporated, typically a PFSA or a perfluorinated carboxylic acid. This technique has enabled membranes with acceptable mechanical properties to be made as thin as 5–30 µm. Proton‐Exchange Membrane Fuel Cells An alternative approach has been to modify chemically the molecular structure of the polymer so as to increase the porosity at the nanoscale and thereby allow a greater retention of water. This objective has been achieved by incorporating co‐monomers with bulky side groups, or by using block copolymers. The resulting membrane material has high proton conductivity, even at relatively low humidity. Another simple method for attaining the same outcome is to add a second proton‐conducting material alongside the polymer. The earliest examples of this approach were the inclusion of small particles of inorganic, proton‐conducting oxides such as silica (SiO2) or titania (TiO2). Sol–gel techniques were employed to introduce the oxides with the aim of absorbing water on the oxide surface so as to limit water loss from the cell via ‘electro‐osmotic drag’. Unfortunately, such a technique has normally led to significant reduction in the proton conductivity of the PFSA. Better results have been obtained by incorporating silica‐ supported phosphotungstic acid and silicotungstic acid, zirconium phosphates and silica alkoxides produced using, e.g., (3‐mercaptopropyl)‐methyldimethoxysilane (MPMDMS).7 4.2.3 Alternative Sulfonated and Non‐Sulfonated Membranes The high cost of manufacturing the PFSAs has led researchers to seek alternative membrane materials for PEMFCs (particularly for high‐temperature operation) and also for application in direct methanol fuel cells (DMFCs). Severe methanol crossover can occur through traditional PFSA membranes from the anode to the cathode when used in DMFCs. Many hydrocarbon polymers have attracted attention in recent years despite the fact that materials such as phenol sulfonic acid resin and poly(trifluorostyrene sulfonic acid) were considered during the 1960s but later fell out of favour on the account of their low thermal and chemical stability. Polymer chemists have evaluated materials such as trifluorostyrene, copolymer‐based α,β,β‐trifluorostyrene monomer and radiation‐grafted polymer membranes. Of the non‐fluorinated polymers, the most studied are sulfonated poly(phenyl quinoxalines), poly(2,6‐diphenyl‐4‐phenylene oxide), poly(aryl ether sulfone), acid‐doped polybenzimidazole (PBI), sulfonated polyether ether ketone (SPEEK), poly(benzyl sulfonic acid)siloxane (PBSS), poly(1,4‐phenylene), poly(4phenoxybenzoyl‐1,4‐phenylene) (PPBP) and polyphenylene sulfide. The range of candidate polymers examined during the 1990s by the Advanced Materials Division of Ballard Power Systems and by other fuel‐cell companies and organizations such as the Stanford Research Institute in the United States has been quite exhaustive. The physical robustness of many polymers can be enhanced via chemical modification to give increased entanglement of the side chains. Some of these materials have improved thermal stability, but unfortunately most have generally lower ionic conductivities than Nafion at comparable equivalent weights. Many others are also more susceptible than Nafion to oxidative8 or acid‐catalysed degradation. 7 Ladewig, BP, Knott, RB, Hill, AJ, Riches, JD, White, JW, Martin, DJ, Diniz da Costa, JC and Lu, GQ, 2007, Physical and electrochemical characterization of nanocomposite membranes of Nafion and functionalized silicon oxide, Chemistry of Materials, vol. 19(9), pp. 2372–2381. 8 Depending on the catalyst used on the oxygen side of the PEMFC, highly oxidizing peroxide species may be formed. These can attack the membrane and thereby significantly reduce its lifetime. Furthermore, platinum particles may dissolve in the acid membrane and recombine to form an electrically conductive pathway through the polymer that serves to reduce the open‐circuit voltage and therefore the cell performance. 77 78 Fuel Cell Systems Explained L2 N N Y Z N N L1 Figure 4.7 Chemical structure of polybenzimidazole (PBI). Perhaps the most investigated and representative of the non‐fluorinated hydrocarbon polymers is PBI (Figure 4.7), which is a heat‐resistant (melting point >600°C), non‐ sulfonated, basic material made by condensation of 3,3′diaminobenzidine and diphenyl isophthalate. The inherent proton conductivity of PBI is very low, but it is easily doped with strong acids to form a single‐phase polymer electrolyte with high conductivity. Phosphoric acid has been found to be most stable and cost‐effective for this purpose. At high temperatures, it exhibits good thermal stability, adequate mechanical properties with low gas permeability and low electro‐osmotic drag of water. A membrane enhanced with phosphoric acid can be prepared either by infusing a cast film of PBI with phosphoric acid or by actually polymerizing the monomers directly in polyphosphoric acid (PPA). This acid can then be hydrolysed to phosphoric acid to provide a membrane of high mechanical stability and with a high loading of phosphoric acid. Given the latter feature, the proton conductivity approaches that of Nafion and increases with temperature. When using a phosphoric acid‐doped membrane, a typical MEA operates in the temperature range of 150–180°C, and the proton conduction is essentially through the phosphoric acid rather than through water as in the conventional PEMFC. Because the temperature is so high, water that is produced by the fuel cell is released as steam, and therefore the pores of the catalyst layer or gas‐ diffusion layer (GDL) are much less likely to be prone to flooding. Phosphoric acid is allowed to penetrate the catalyst layers (see Section 4.3) and, because it is mobile, care needs to be taken to ensure that there is sufficient access and egress of the reactant gases and products on both sides of the fuel cell. The high‐temperature PEMFCs based around phosphoric acid‐doped polymers require no external humidification and, in principle, can lead to a much simpler system compared with the traditional design. One disadvantage, however, is that the catalytic reaction on both electrodes is slower with phosphoric acid than with PFSAs, and therefore the cell voltage is generally lower for such high‐temperature PEMFCs. Consequently, catalysts with slightly higher loadings of platinum are required. Another problem is that although the phosphoric acid is immobilized on the basic sites of the PBI, the very high loadings cause the polymer to lose its chemical stability. This shortcoming can be lessened by casting the PBI with a polyphenolic resin such as polybenzoxazine (PBOA). High‐temperature polymer membranes for the PEMFC have been developed by several research groups in numerous universities and companies. Examples are those investigated by BASF, the Paul Scherrer Institute, and Sartorius (later Elcomax). Complete high‐temperature PEMFC systems are now also being commercialized by the Danish company Serenergy for small‐scale stationary power applications (Figure 4.8). Proton‐Exchange Membrane Fuel Cells (a) (b) Figure 4.8 Serenegy liquid‐cooled high‐temperature PEMFC: (a) stack and (b) system. (Source: Reproduced with permission of Serenergy.) 4.2.4 Acid–Base Complexes and Ionic Liquids Two other classes of potentially useful materials for fuel‐cell membranes are acid–base complexes and ionic liquids. The first category comprises traditional inorganic acids, such as sulfuric, phosphoric or hydrochloric acid that are embedded within a polymer. The polymer has to be chemically basic so that the acid is chemically bound within the structure. Of the many possible complexes, perhaps the one with the most suitable 79 80 Fuel Cell Systems Explained characteristics for PEMFCs, or especially for DMFCs, is phosphoric acid combined with PBI or ABPBI (a simpler version of the polymer without the phenylene groups). The most widely employed ionic liquids in the fuel‐cell industry are the molten alkali carbonates used in the MCFC, as described in Chapter 8. Although these may be referred to as ionic liquids, they are usually known as molten salts because at normal temperatures they are solid and have to be heated above their melting point to have significant ionic conductivity. The term ‘ionic liquid’ most often is used to describe a liquid that possesses ionic character but at room temperature. Many organic compounds fit into this category, and some of these are also being evaluated as a source of possible membranes for low‐temperature fuel cells. To date, none of these ionic liquids have progressed from the laboratory into commercial fuel‐cell systems. 4.2.5 High‐Temperature Proton Conductors As mentioned earlier, phosphoric acid is a good proton conductor and the PAFC is described in Chapter 5. There are other materials that exhibit proton conduction, but at much higher temperatures. The most favoured are ceramics with a perovskite structure,9 notably doped barium and strontium cerates and their mixtures. These exhibit good proton conductivity (in the order of 10 mS cm−1) in the temperature range 500–900°C. Unfortunately, due to their basic character, they are unstable in gas atmospheres that contain H2O, or CO2, H2S, SO2 or SO3, and form Ba(OH)2, BaCO3, BaS or BaSO4, or the strontium equivalents. The poor chemical stability limits these materials for application as electrolytes in hydrogen‐only fuel cells. The barium and strontium cerates are the simplest among many perovskite oxides that have been investigated for high‐temperature proton conduction, the others being: ● ● ● II–IV type oxides, e.g., (Ca, Sr, Ba) (Ce, Zr, Ti) O3. I–V type oxides, e.g., (K Ta O3). III–III type oxides, e.g., (La Y O3). The roman numbers refer to groups in the periodic table where the included elements are to be found. To make such materials proton-conducting, generally they must be doped with an element of lower valency than the B‐site atom, e.g., Y in BaCeO3, to increase the concentration of charged species. Some more complex perovskites such as: ● ● II2–(III/V) type oxides (e.g., Sr2ScNbO6). II3–(II/V2) type oxides(e.g., Ba3CaNb2O9). can also be made proton-conducting by making them non‐stoichiometric, e.g., Ba3Ca1.18Nb1.82O9‐δ (also known as BCN18). A range of dopants have been investigated for inclusion in the barium and strontium cerates (e.g., Y, Tm, Yb, Lu, In or Sc), and it has been found that the larger the ionic radius or the more basic the dopant, the greater the conductivity for the same dopant level.10 9 Perovskite materials are discussed further in Chapter 9 as many of these materials are also good oxygen‐ion conductors, suitable as electrolytes for SOFCs. 10 Matsumoto, H, Kawasaki, Y, Ito, N, Enoki, M and Ishihara, T, 2007, Relation between electrical conductivity and chemical stability of BaCeO3‐based proton conductors with different trivalent dopants, Electrochem. Solid State Letters, vol. 10, pp. B77–B80. Proton‐Exchange Membrane Fuel Cells Alternatives with comparable proton conductivity to the cerium compounds include fluorite‐related structures, e.g., the tungstates La5.8WO11.7 and La5.7Ca0.3WO11.85, but their chemical stability has yet to be established. Other materials such as pyrochlores, e.g., La1.95Ca0.05Zr2O6.975, appear to be more chemically stable, but only show good proton conductivity up to about 600°C. Development of high‐temperature proton‐conducting ceramics has been carried out by groups in the United States under a DOE‐led programme and by a consortium of research teams in Europe under the EU‐FP7 project. Notably, research in Norway led to the spin‐out company Protia AS in 2008, with the purpose of commercializing proton conductors and mixed proton–electron conductors. Applications include hydrogen separation membranes and enhanced steam reforming systems in addition to proton ceramic fuel cells. Further to the lanthanum tungstate mentioned earlier, lanthanum niobate LaNbO4 has also been studied by the researchers in Norway and promises to be an important material for the future. Another member of the ABO4 family, namely, lanthanum vanadate LaVO4 was identified by workers in North Europe as a high‐ temperature proton conductor when doped with calcium. Critical to the optimization of these new materials is the required level of dopants to achieve adequate proton conduction with good mechanical and chemical stability. To date, conductivity is significantly lower than alternative oxygen‐ion conductors so that quite thin electrolytes of the order of 1–10 µm are required. Although a commercial proton ceramic fuel cell has yet to emerge, by operating up to 900°C platinum‐containing electrode catalysts will clearly not be required. The porous carbon GDL and carbon‐supported catalyst layer used in PEMFCs will also not feature in such fuel cells. Rather, the electrodes will need to resemble more closely those in SOFCs that operate over a similar temperature range. In addition, and in contrast to SOFCs that feature an oxygen‐ion‐conducting electrolyte, water is produced at the cathode of the proton ceramic fuel cell. Therefore, although the temperature of operation may be similar to the intermediate-temperature SOFCs described in Section 9.1.1, Chapter 9, the system design is likely to be significantly different. 4.3 Electrodes and Electrode Structure Platinum is the metal with the greatest catalytic activity for both electrode reactions in the PEMFC. In the early days of the development of this fuel cell, around 28 mg of platinum was required per cm2 of electrode surface area for each electrode. This high rate of usage led to the belief, still widely held, that platinum is a major factor in the cost of PEMFC and that the world’s supply of the metal is not adequate to satisfy the market for fuel‐cell vehicles should they become widely adopted. Both observations are misleading. The reality is that platinum usage has been reduced to below 0.2 mg cm−2 and, moreover, yields much better performance in fuel cells today than catalysts of 10 years ago. At such low ‘loadings’, the basic raw material cost at the present prices of platinum in a 1‐kW PEMFC would be about US$10, so the prospects of mass commercialization appears to have greatly increased. Even so, refinement of catalysts to give further improvements in performance and lifetime will be necessary for the PEMFC to achieve widespread commercial acceptance. 81 82 Fuel Cell Systems Explained Bipolar plate with flow-field channels for oxidant and fuel Gas-diffusion layer Catalyst layers Electrolyte membrane Membrane electrode assembly Figure 4.9 Basic structure of a low‐temperature PEMFC with a simple configuration of bipolar plate. The basic structure of the electrode in different designs of PEMFC is very similar, despite variations in the details. The negative and positive electrodes are also essentially the same, and in many PEMFCs they are identical. The main features of a typical planar PEMFC in which layers of catalyst are sandwiched between the electrolyte membrane and a porous GDL are shown in Figure 4.9. The GDL, in turn, is in direct contact with the bipolar flow‐field plate. The following subsections separately provide a description of each of these components. 4.3.1 Catalyst Layers: Platinum‐Based Catalysts In a typical PEMFC, the catalyst layer on each GDL has a thickness of around 10 µm and comprises very small particles of platinum metal on the surface of finely divided carbon of a somewhat larger particle size. The requirements for the fuel side (negative) and air side (positive) of the PEMFC are very different. As has been remarked in Section 3.4.2, Chapter 3, the rate of the oxygen reduction reaction (ORR) is much slower than that of the hydrogen oxidation reaction. Typically, the exchange‐current density for hydrogen oxidation is three orders of magnitude higher than that for oxygen reduction, e.g., 1 mA cm−2 (H2) versus 10−3 mA cm−2 (O2). At a representative operating current density of 400 mA cm−2, the voltage loss at the anode is about 10 mV, while that at the cathode is over 400 mV. For this reason, the platinum loading in the catalyst layer on the air electrode (cathode) is usually much higher than that in the layer on the fuel electrode (anode). The carbon in the catalyst layers is usually produced by the pyrolysis of hydrocarbons to yield a highly porous, nanostructured powder with a high surface area (800– 2000 m2 g−1). An example of such a commercially available powder is Vulcan XC72 (Cabot), which is found in many industrial applications. In the fuel cell, the carbon serves not only to disperse the active metal but also to provide good electronic conductivity to enable a high current to be drawn. The method of depositing the platinum on the carbon usually starts with a precursor solution (e.g., chloroplatinic acid, or another water‐soluble platinum compound), which is absorbed on the surface of the carbon. The absorbed precursor may then be chemically reduced (e.g., using sodium borohydride) or simply heated to decompose the compound and release the metal as finely divided particles on the surface of carbon clusters. The result of loading the carbon with platinum, in a somewhat idealized form, is shown in Figure 4.10. This should be compared with Figure 1.6 in Chapter 1 that shows an electron micrograph of an actual supported ® Proton‐Exchange Membrane Fuel Cells Catalyst particles Carbon support Figure 4.10 Structure, idealized, of a carbon‐supported platinum catalyst. catalyst. The platinum is well dispersed on the carbon particles, so that a very high proportion of the surface area of the metal will be in contact with the gas‐phase reactants. This high degree of dispersion maximizes the ‘three‐phase boundary’ described in Section 1.3, Chapter 1. Two alternative methods are generally employed for depositing the catalyst layers in the MEA. Either the catalyst is first bonded to the appropriate GDL and then to the electrolyte or it is bonded first to the electrolyte and the GDLs are added afterwards. The end result is essentially the same in both cases. Usually, the first requirement is to produce a dispersion of the platinized carbon powder in a polar and volatile solvent such as ethanol. A small amount of Nafion solution is normally added to the mixture for reasons that will become apparent later. PTFE will often be added also to the catalyst layer; during operation of the fuel cell, this hygroscopic material serves to expel product water to the electrode surface where it can evaporate. Ultrasonic agitation of the catalyst/ethanol suspension disperses the powder and creates an ‘ink’ that enables deposition of the catalyst onto the appropriate cell component (GDL or electrolyte membrane) by a suitable method such as painting, printing or rolling. The solvent in the ink is allowed to evaporate to leave the solid catalyst adhering to the given component. If the catalyst is first deposited on the two GDLs, the resulting two electrodes are then bonded to either side of the polymer electrolyte membrane by means of the following common procedure: ● ● ● The electrolyte membrane is cleaned by immersion in boiling 3 vol.% hydrogen peroxide in water typically for 1 h, and then in boiling sulfuric acid for the same time, to ensure as full a protonation of the sulfonate groups as possible (and removal of sodium ions). The membrane is rinsed in boiling de‐ionized water for another hour to remove any remaining acid. The electrodes are placed on the electrolyte membrane, and the assembly is hot pressed for 3 min at 140°C and high pressure. The result is a complete MEA. If the catalyst ink is first deposited directly to the protonated electrolyte rather than the respective GDLs, then both GDLs must be applied afterwards. This method tends 83 Fuel Cell Systems Explained (a) (b) 0.4 0.4 0.2 0.2 0.0 ji/mA cm–2 ji/mA cm–2 84 1 –0.2 2 –0.4 3 4 6 0.0 1 –0.2 2 3 4 6 –0.4 –0.6 –0.6 5 5 –0.8 –0.8 –0.2 0.0 0.2 0.4 0.6 E/ V vs. SCE 0.8 1.0 1.2 –0.2 0.0 0.2 0.4 0.6 E/ V vs. SCE 0.8 1.0 1.2 Figure 4.11 Cyclic voltammograms for thin‐film platinum electrodes (curves 1–5) and bulk platinum (curve 6) in (a) argon‐saturated 0.1 M HClO4 and (b) 0.05 M H2SO4. Film thickness: (1) 0.25, (2) 0.5, (3) 1, (4) 2 and (5) 10 nm. Sweep rate: 100 mV s−1. Note the negative peak at 0.40–0.45 V is caused by the ORR. (Source: Adapted from Thompsett, D, 2003, Catalyst for the proton‐exchange membrane fuel cell, Chapter 6, in Fuel Cell Technology Handbook, CRC Press, Boca Raton, FL, ISBN: 978‐1‐4200‐4155‐2.) to result in a thinner catalyst layer and may be preferred for some applications, but otherwise the MEA gives similar results to that produced by the alternative procedure outlined earlier. Both of the procedures for assembling a PEMFC, while being low cost and amenable to volume production, have the disadvantage of producing relatively thick layers of catalyst in which platinum is underutilized. More recently, other means of depositing the active metal onto carbon have been investigated with a view to improving its effectiveness. Emerging methods include various modified thin‐film techniques, electrodeposition and sputter deposition, dual ion beam‐assisted deposition, electroless deposition, electrospray processes, and direct deposition of platinum sols. For example, platinum particles of less than 5 nm in diameter can be plasma sputtered directly onto carbon nanofibres11 to produce a catalyst with a loading of between 0.01 and 0.1 mg cm−2. The performance of platinum catalysts depends very much on the active surface area, i.e., on the degree of dispersion and the particle sizes. Cyclic voltammograms for thin films of platinum on carbon are displayed in Figure 4.11. The data show that the intensity of the oxygen reduction peak increases for film thicknesses between 2 and 10 nm. This complies with data published elsewhere that suggest the optimum size of platinum particles supported on carbon for catalysing the ORR is between 2 and 4 nm. Although carbon blacks have been well proven and continue to be used in practical PEMFCs, both single‐walled and multiwalled carbon nanotubes, as well as graphene, have been investigated recently as alternatives. The single disadvantage of these forms of carbon is that they all have intrinsically low surface areas, which is a feature that does not favour the production of a very active catalyst. On the other hand, the highly ordered surfaces of carbon nanotubes and graphene do appear beneficial to some non‐precious metal catalysts, as discussed in the next section. 11 Caillard, A, Charles, C, Boswell, R and Brault, P, 2008, Improvement of the sputtered platinum utilization in proton exchange membrane fuel cells using plasma‐based carbon nanofibres, Journal of Physics D‐Applied Physics, vol. 41(18), pp. 1–10. Proton‐Exchange Membrane Fuel Cells Cyclic voltammetry and especially the characterization of catalysts using rotating disc electrodes have identified that two fundamental reactions can occur at the positive electrode in the PEMFC. The first is the more normal oxygen reduction via a 4‐electron transfer process, namely: O2 4 H 4e 2H 2 O (4.1) The second reaction is via a 2‐electron transfer intermediate reaction as follows: O 2 2H 2e H2O2 (4.2) The peroxide reaction (4.2) is favoured at cathode potentials of less than 0.5 V with respect to the standard hydrogen electrode. Peroxide formation may also occur if hydrogen can crossover (see Section 3.3, Chapter 3) through the membrane and then become oxidized directly on the cathode. Peroxide reacts with the electrolyte and can accelerate electrode degradation. It is important therefore to ensure that in the PEMFC crossover is minimized and that the electrode potentials are maintained within safe limits. 4.3.2 Catalyst Layers: Alternative Catalysts for Oxygen Reduction The high cost of platinum has spurred researchers both to reduce its loading in catalysts and to seek cheaper alternatives. The reason platinum is so active for both hydrogen oxidation and oxygen reduction has puzzled chemists for many years.12 Current understanding is that the high activity arises partly because the metal loosely adsorbs molecules such as oxygen or hydrogen on its surface and also because it can ease the dissociation or splitting of the adsorbed molecules into adsorbed atoms, which are then able to react.13 The strength of the chemical bond between oxygen atoms and atoms on the metal surface is dependent on the crystal plane and edges exposed. The oxygen–metal bond strength can also be influenced by alloying platinum with other metals. To this end, nickel, rhodium, iridium, cobalt and other transition elements have been combined with platinum to promote the dissociative adsorption of oxygen. The influence of combining different metals with platinum has been calculated from first principles, and experiments have verified that metals such as ruthenium have positive effects on the surface properties of platinum. A more recent development has involved depositing the platinum as a shell of single‐atom thickness (a monolayer), or even as small islands a few atoms thick, on particles of other metals such as ruthenium or rhodium. These procedures increase the dispersion of the active metal and potentially reduce the cost of the catalyst. To lower the costs of the catalyst further, other materials have been evaluated that do not involve platinum or platinum‐group metals. Over the past 10 years, the alternatives that have received the most attention are as follows. 12 Platinum is very active for hydrogen oxidation but less active for oxygen reduction in the PEMFC. As will be shown, other metals work well for oxygen reduction in the alkaline fuel cell, where the reduction reaction mechanism is somewhat different. 13 Holton, OT and Stevenson, JW, 2013, The role of platinum in proton exchange membrane fuel cells, Platinum Metals Rev., vol. 57(4), pp. 259–271. 85 86 Fuel Cell Systems Explained 4.3.2.1 Macrocyclics Transition metal macrocyclic compounds have been examined as potential ORR catalysts since the early 1960s. The compounds have a molecular structure in which a central transition metal atom is enclosed within a much larger cyclic organic molecule. Often the metal atom is linked to nitrogen atoms, and a common characteristic is the MN4 structure, in which the metal atom is bound to four nitrogen atoms. The structure is an example of chelation, and therefore the molecule is also known as a ‘chelate’. Among the many series of macrocyclic compounds, phthalocyanines (Pc) complexed with various transition metals such as iron, cobalt, nickel and copper have been thoroughly investigated as oxygen reduction catalysts. Phthalocyanines have been known since the beginning of the 20th century and are widely used as dyes. Of the various phthalocyanines evaluated for the reduction of oxygen in fuel cells, the complexes with cobalt and copper appear to be the most stable, whereas those with iron and cobalt seem to have the best combination of activity and stability. The chelates of 5,14‐dihydro‐5,9,14,18‐dibenzotetraaza[14]annulene, or ‘tetraazaanulene’ (TAA) (see Figure 4.12) are another class of macrocyclic complex that exhibits good potential for ORR catalysis. Porphyrins are the second major group of macrocyclics that have been considered as non‐precious metal catalysts; some examples are tetraphenylporphyrin (TPP) and tetramethoxyphenyl‐porphyrin (TMPP), also shown in Figure 4.12. In many cases, these are absorbed onto the carbon carrier, which is then heated to a high temperature (typically, 800–900°C) to decompose the porphyrin molecule and thereby allow the metal to bind directly to nitrogen that has, in turn, become bound to the carbon surface. Indeed, functionalization of the carbon support surface with nitrogen only (e.g., by treating with concentrated nitric acid) has been found to enhance the activity of TPP and TMPP catalysts. The study of macrocyclic compounds as ORR catalysts is currently a very active field of research. N N M N N OMe TAA NH N NH N MeO N HN TPP OMe N HN OMe TMPP Figure 4.12 Molecular structures of macrocyclic organic frameworks used for ORR catalysts: tetraazaanulene (TAA), tetraphenylporphyrin (TPP), tetramethoxyphenyl‐porphyrin (TMPP). Proton‐Exchange Membrane Fuel Cells 4.3.2.2 Chalcogenides In the context of ORR catalysts, ‘chalcogenides’ refer to either sulfides or selenides of various transition metals. Among the many examples examined to date, Co3S4 and CoSe2 supported on carbon, as well as various ternary variants such as W–Co–Se, are all claimed to have high ORR activity. 4.3.2.3 Conductive Polymers Polymers such as polyaniline (pani), polypyrrole (Ppy) and poly(3‐methylthiophene) (P3MT) can be used to prepare electronically conducting materials in which metal atoms, e.g., iron, cobalt or nickel are bonded to the nitrogen atoms within the polymer. Several of these have been shown to have appreciable catalytic activity for oxygen reduction. 4.3.2.4 Nitrides Building on the notion that nitrogen is required for macrocyclics and conductive polymers to be endowed with catalytic activity, some researchers have explored the prospects of transition metal nitrides. Whereas tungsten and molybdenum nitrides supported on carbon exhibit some promise, it has yet to be determined whether these candidates can be engineered with sufficient activity and longevity to compete with the established platinum catalysts. 4.3.2.5 Functionalized Carbons There are essentially two ways by which carbons can be treated or ‘functionalized’14 to achieve nitrogen on the surface. The first is simply by treatment with nitric acid or by heating the carbon in nitrogen or ammonia. The process can be conducted either before the active metal is added or after it is impregnated by a salt of the appropriate metal, e.g., cobalt acetate or nickel nitrate (acetate and nitrates are easily decomposed by heating, to leave the metal atoms bound to the carbon). The second and widely adopted method of functionalizing carbon is to use a transition metal complex as the source of metal that when decomposed will ensure a high dispersion of metal atoms on the carbon surface. Perhaps the most frequently chosen complex has been 2, 4, 6‐tris(2‐pyridyl)‐1,3,5‐triazine (TPTZ). Typically, the metal‐TPTZ complex is impregnated into the porous carbon, carbon nanotube or graphene and then decomposed by heating in the absence of air. Highly active N catalysts have been obtained by this procedure. It has been found that the activity is dependent on the experimental conditions N N (e.g., temperature, heating rate), as well as on the type of N N carbon. Interestingly, the more ordered the carbon structure, N the higher the activity of the catalyst is. Catalysts prepared using Fe‐TPTZ (Figure 4.13) for example, can approach the same activity as platinum‐based catalysts, but these are supported Figure 4.13 Molecular on expensive graphene materials and no doubt much work will structure of 2,4,6‐tris have to be undertaken if they are to compete with platinum‐ (2‐pyridyl)‐1,3,5‐triazine (TPTZ). based materials in terms of long‐term activity in fuel cells. 14 Functionalization, broadly, is the addition of functional groups onto the surface of a material by chemical synthesis methods. A functional group is a small number of atoms or bonds within a molecule that determines the chemical properties of the group and of the molecule to which it is attached. 87 88 Fuel Cell Systems Explained 4.3.2.6 Heteropolyacids Indicating that the field of novel ORR catalysts is by no means exhausted, mention should be made of recent work on a particular class of inorganic compounds known as heteropolyacids. Some of these, such as H3PMo12O40 and H3PW12O40, have received particular attention due to their acidic and redox properties, stability at elevated temperatures, commercial availability and relative ease of synthesis. Heteropolyacids are also proton conductors — a feature which may be exploited in the design of the PEMFC. 4.3.3 Catalyst Layer: Negative Electrode As mentioned earlier, there is less incentive for developers to seek non‐platinum‐based catalysts for the negative electrode of the fuel cell because less platinum is required for the hydrogen oxidation reaction. On the other hand, the anode catalyst is susceptible to poisoning by sulfur and CO, both of which may be present in the hydrogen fed to the fuel cell, particularly if it has been produced from hydrocarbons. If CO is in the fuel entering the fuel cell at a concentration of more than a few ppm, it will preferentially adsorb on the surface platinum atoms and reduce the activity of the catalyst. If the partial pressure of CO in the fuel stream is low (i.e., below a few ppm), its adsorption on the anode catalyst is reversible. In such a situation, the catalyst can be kept active, for example, by regularly purging the fuel side with a small amount of oxygen or briefly applying a negative potential to the electrode. This technique has been applied in some practical fuel‐cell systems. Another method of increasing the allowable concentration of CO on the negative electrode is to use an alloy of platinum and ruthenium, rather than simply platinum, as the catalyst. 4.3.4 Catalyst Durability Early PEMFCs had lifetimes that were limited not only by the stability of the membrane but also by the durability of the catalysts. Over the past 10 years, remarkable progress has been made in the understanding of catalyst durability and has resulted in a significant increase in the expected lifetime. Catalyst degradation is now known to occur, variously, through the sintering of platinum particles, dissolution of platinum and corrosion of the carbon support. Sintering of platinum particles on the carbon support decreases the catalytically-active surface area. It may take place by a dissolution– precipitation mechanism in which small metal particles of the catalyst may dissolve into the acidic operating environment and then precipitate onto larger metal particles and thus promote particle growth, or the particles may directly coalesce with each other due to movement on the carbon surface. Both mechanisms occur to some degree, with dissolution–precipitation being more prevalent when load changing or shutdown/ start‐up occurs. The sintering of the catalyst may be reduced by strengthening the interaction between the catalyst metal and the supporting carbon. For example, grafting polyaniline to the carbon surface has been found to decrease the mobility of the metal particles. The nitrogen moiety in the polyaniline has a lone pair of electrons that can anchor the platinum particles. Dissolution of platinum is accelerated by voltage cycling, which, for example, is experienced in a fuel‐cell vehicle where acceleration and braking give rise to a varying load on the PEMFC stack. One way of reducing the effect is to hybridize the fuel‐cell system in a vehicle with a battery. Corrosion of the carbon support is generally not an issue for fuel‐cell systems that operate at steady-state but becomes an important degradation mechanism with repeated Proton‐Exchange Membrane Fuel Cells start–stop cycles. When a fuel supply is shut off, air may leak into the anode compartments of a stack and cause a hydrogen–air front to form in the flow-fields of the separator plates. The result is spatially separated oxygen and hydrogen on the negative side of the cells that can cause a lowering of the potential at the positive electrodes, leading to oxidation of carbon in the cathode GDL and/or catalyst layers. Corrosion of carbon on the cathode side results in a thinning of the catalyst layer that can be mitigated by changing the type of carbon in the catalyst to a more stable graphitic form than the conventional pyrolytic variety, by purging hydrogen quickly from the negative side of the cell on shutdown or by purging oxygen from the positive side.15 4.3.5 Gas‐Diffusion Layer Commercial GDLs are made of porous conductive material — usually carbon fibres — in the form of paper or thin fabric/cloth and typically have a thickness of 100–400 µm. ‘Gas‐diffusion layer’ is a slightly misleading name for this part of the electrode, as it does much more than simply provide a porous structure so that reactant and product gases can diffuse respectively to and from the catalysts. Namely, it also forms an electrical connection between the carbon‐supported catalyst and the bipolar plate, or other current‐collector. In addition, the GDL carries the product water away from the electrolyte surface and forms a protective layer over the very thin layer of catalyst. The structure of the electrolyte | catalyst layer and the GDL is shown in idealized form, in Figure 4.14. The carbon‐supported catalyst particles are joined to the electrolyte Platinum particles supported on carbon Gas-diffusion layer, e.g., carbon cloth fibres Main bulk of electrolyte Figure 4.14 Simplified structure of a PEMFC electrode. 15 Hydrogenics has a patented method of removing oxygen from the positive electrode on shutdown that involves sealing the supply of air to the cathode of the fuel‐cell stack and electrochemically reducing the remaining oxygen by using a buffer of hydrogen adjacent to the anode that is reserved for this purpose. 89 90 Fuel Cell Systems Explained A thin layer of electrolyte polymer adheres to catalyst metal particles, promoting the three-phase contact between electrolyte, reactant gas (hydrogen at the anode and oxygen at the cathode) and the catalyst surface Main bulk of the electrolyte Figure 4.15 Enlargement of part of Figure 4.14 to show that the electrolyte reaches out to the catalyst particles. on one side, and the GDL (current‐collecting, water‐removing, physical support) on the other. The hydrophobic PTFE that is needed to remove water from the catalyst is not shown explicitly in the figure, but will almost always be present. Two further points need to be made. The first relates to the impregnation of the electrodes with electrolyte material. A section of the catalyst–electrode region is shown enlarged in schematic form in Figure 4.15. The electrolyte material extends out to the catalyst particles sufficiently to provide proton transport to and from the catalyst, which is where the electrode reactions take place. An important point is that only the catalyst which is in direct contact with both the electrolyte and the reactant gas can be active for the electrochemical reactions at the electrodes (reactions only occur at the three‐phase boundary). To maximize catalyst activity, therefore, the catalyst layer of each electrode is lightly covered with electrolyte, usually by brushing the surface with a solubilized form of the electrolyte. In the case of the ‘separate electrode’ method of MEA preparation, this procedure is conducted before the electrode is hot pressed onto the membrane. By contrast, the alternative ‘integral membrane–electrode’ process is undertaken before the GDL is added. The second point relates to the selection of the GDL, which is generally either a carbon paper or carbon cloth material. Carbon paper (e.g., Toray paper) is chosen when it is required to make the cell as thin as possible in compact designs. Such paper, which is made by pyrolysing a non‐woven carbon fibre sheet, has good conductive properties but tends to be brittle and fragile. By virtue of their greater thickness, carbon cloths will absorb a little more water than paper and thereby will be less prone to flooding. Cloths also simplify the mechanical assembly, since they are inherently more flexible and will deform under compressive forces. Consequently, cloths can fill small gaps and irregularities in the manufacture and assembly of bipolar plates. On the other hand, cloths may slightly deform out into the gas‐diffusion channels on the bipolar plates and thereby may restrict gas flow through the channels. The Elat range produced by Nuvant Systems Inc. is a commercial cloth that is a popular choice for GDLs. The cloth is prepared by loading highly conductive carbon fibres with carbon black. Another GDL innovation is the addition of a very thin microporous hydrophobic layer of carbon between the GDL and the catalyst layer. The layer consists principally of carbon black with a 10–40 wt.% loading of PTFE as determined by the type of carbon. Carbon plays an important role in both the catalyst layer and the GDL of both the PEMFC and, as shown later, the PAFC. It is not surprising, therefore, that the discovery of ® ® Proton‐Exchange Membrane Fuel Cells Figure 4.16 Example of a membrane electrode assembly (MEA). The membrane is a little larger than the electrodes that are attached. The 10 cm2 membrane is typically 0.05–0.1 mm thick, the electrodes are about 0.03 mm thick, and the GDL is between 0.2 and 0.5 mm thick. carbon nanotubes and graphene at the end of the 20th century stimulated considerable research activity for their application in fuel‐cell systems.16 In the case of the carbon GDL, the type of carbon, its structure, thickness and electrical conductivity are all influenced by cell operating conditions. For example, the GDLs in conventional low‐ temperature PEMFCs must be able to accommodate water movement as discussed in Section 4.4. They tend therefore to have an open‐pore structure to allow for water diffusion out of the cell. By contrast, the GDLs in PEMFCS operating at high temperatures (above 100°C) do not need to have such open porosity as liquid water is less likely to condense out and flood the electrodes. In summary, the MEA is the key component of a PEMFC and irrespective of its method of manufacture, every MEA possesses positive and negative electrodes, each of which incorporates catalyst material. A practical 10‐cm2 MEA is shown in Figure 4.16. The nature of the electrolyte may differ according to the operating temperature of the stack. Also important is the way in which the MEA is incorporated in the construction of fuel‐cell stacks. The design does vary significantly between manufacturers and is influenced by the application as discussed in the sections that follow. 16 Dicks, AL, 2006, The role of carbon in fuel cells, Journal of Power Sources, vol. 156(2), pp. 128–141. 91 92 Fuel Cell Systems Explained 4.4 Water Management 4.4.1 Hydration and Water Movement It will be clear from the description of proton‐exchange membranes given in Section 4.2 that, particularly for the PFSA versions, there must be sufficient water in the polymer electrolyte to maintain high proton conductivity. At the same time, the water content must be managed to prevent flooding in either the catalyst layer or the GDL. In an ideal PEMFC, the water that forms at the positive electrode would be expected to keep the electrolyte at the correct level of hydration. Air would be blown over the electrode and, as well as supplying the necessary oxygen, it would dry out any excess water. Because the membrane electrolyte is so thin, water would diffuse from the positive side to the negative, and throughout the whole electrolyte a suitable state of hydration would be achieved without any special difficulty. This preferred situation can sometimes be achieved but relies on a good engineering design. There are several complications. During operation of the cell, the H+ ions moving from the negative to the positive electrode pull water molecules with them — a process usually referred to as ‘electro‐osmotic drag’. Typically, between 1 and 2.5 water molecules are conveyed for each proton. This means that, especially at high current densities, the negative side of the electrolyte can become dried out, even if the positive side is well hydrated. Another major problem is the drying effect of air at high temperatures; this issue is discussed quantitatively in Section 4.4.2. Suffice it to say, at temperatures of over about 60°C, the air will always dry out the electrodes faster than water is produced by the hydrogen–oxygen reaction. It is customary to keep the membrane sufficiently hydrated by humidifying the air, or the hydrogen, or both, before entry into the fuel cell. Such action may seem bizarre, as it effectively adds by‐product to the reactants, but sometimes it is necessary and, moreover, it can greatly improve the performance of the fuel cell. To achieve uniform proton conductivity throughout the whole fuel cell, the degree of hydration in the electrolyte must similarly be uniform. In practice, some parts may be correctly hydrated, others too dry and others overhydrated or flooded. For example, consider that air entering the cell may be quite dry, but by the time it has passed over some of the electrode, it may have achieved the optimum level of humidity. On reaching the exit, however, the air may be so saturated that it is unable to remove any more of the water produced. This is a particular problem when designing larger cells and stacks. The different water movements are shown in Figure 4.17; fortunately they are predictable and controllable.17 For example, water productions at the cathode and electro‐ osmotic drag are both directly proportional to the current. Both water generation and drag can lead to build-up of liquid water at the cathode of an operating cell. This build‐up creates a driving force for back diffusion of water from the cathode to the anode, which helps to keep the membrane uniformly hydrated. The water lost from the cell by evaporation is governed by the RH of the gases on either side of the cell and can also be 17 Further quantitative discussion of the various water fluxes through the polymer membrane is provided in: Kumbur, EC and Mench, M, 2009, Water management, in Garche, J, Dyer, C, Moseley, P, Ogum, Z, Rand, DAJ and Scrosati, B (eds.), Encyclopedia of Electrochemical Power Sources, vol. 2, pp. 828–847, Elsevier, Amsterdam. Proton‐Exchange Membrane Fuel Cells Anode Electrolyte Cathode H2O Water may back diffuse from the cathode to the anode, if the cathode side holds more water Water maybe supplied by externally humidifying the hydrogen supply Water will be produced within the cathode Water will be dragged from the anode to the cathode side by protons moving through the electrolyte Water will be removed by the O2-depleted air leaving the fuel cell Water may be supplied by externally humidifying the air/O2 supply Water may be removed by circulating hydrogen Figure 4.17 Schematic illustration of the different water movements to, within and from the electrolyte of a PEMFC. predicted with care, using the theory outlined in the following Section 4.4.2. Therefore external humidification of the reactant gases prior to entry to the fuel cell, if employed, can be controlled and can also help to achieve uniform humidification throughout the MEA. As operating experience with PEMFC systems has grown, so has the understanding that excessive water in the MEA not only has an immediate negative effect on the cell performance but also can have long‐term damaging effects. These are caused by the following actions: ● ● Internal stresses within the electrolyte. Since water causes swelling of the membrane, uneven distribution throughout the electrolyte can invoke physical stresses and degradation of the electrolyte and catalyst layers. Contamination through water‐soluble ionic species. Post‐service analysis of fuel cells has shown appreciable accumulation of calcium, iron oxides, copper magnesium and other metals. 93 94 Fuel Cell Systems Explained ● Freeze‐out damage. If a small quantity of liquid water remains present in the GDL or catalyst layers when a PEMFC stack is shut down and the temperature falls below the freezing point, permanent damage can be caused to these layers. It is essential to purge any excess water from the PEMFC to prevent such an occurrence. 4.4.2 Air Flow and Water Evaporation Except for the special case of PEMFCs supplied with pure oxygen, it is universally the practice to remove the product water via the air that flows through the cell. Consequently, the air will always be fed through the cell at a rate faster than that needed just to supply the necessary oxygen. If it were fed at exactly at the ‘stoichiometric’ rate, there would be substantial ‘concentration losses’, as described in the Section 3.7, Chapter 3, because the exit air would be completely depleted of oxygen. In practice, the stoichiometry (λ) will be at least 2. Section A2.2, Appendix 2, provides the derivation of the useful equation (A2.10), which relates the air flow rate, the power of a fuel cell and the stoichiometry. Problems arise because the drying effect of air has a non‐linear relationship with temperature. To understand this characteristic, consideration must be given to the precise meaning and quantitative effects of terms such as ‘RH’, ‘water content’ and ‘saturated vapour pressure’. The partial pressures of the various gases that make up air have been given in Section 2.5, Chapter 2. The analysis, however, ignored the fact that air also contains water vapour. A straightforward way of measuring and describing the amount of water vapour is to give the ratio of water to the other gases, namely, nitrogen, oxygen, carbon dioxide and others that make up ‘dry air’. This quantity is usually given the symbol ω and is known variously as the ‘humidity ratio’, the ‘absolute humidity’ or the ‘specific humidity’; it is defined as: mw ma (4.3) where mw is the mass of water present in the mixture and ma is the mass of dry air, i.e., the total mass of the air is mw + ma. The humidity ratio does not, however, give a very good idea of the drying effect, or the ‘feel’, of the air. Warm air with quite high water content can feel very dry and indeed have a very strong drying effect. On the other hand, cold air with low water content can feel very damp. This characteristic is due to changes in the ‘saturated vapour pressure’ of the water vapour, which is the partial pressure of the water when a mixture of air and liquid water is at equilibrium, i.e., when the rate of evaporation of water in the air is equal to the rate of condensation. When the air cannot hold any more water vapour at a given temperature and pressure, it is said to be ‘saturated’. Air that has no ‘drying effect’ is fully saturated with water and could reasonably be said to be ‘fully humidified’. This state is achieved when P w = Psat, where P w is the partial pressure of the water and Psat is the saturated vapour pressure of the water. The ratio of these two pressures is the ‘RH’, namely: Pw Psat (4.4) Proton‐Exchange Membrane Fuel Cells Typical relative humidities vary from about 0.3 (or 30% RH) in the ultra‐dry conditions of the Sahara desert to about 0.7 (or 70% RH) in a city such as Brisbane or New York on an ‘average day’. Very important for fuel cells is the fact that the drying effect of air, or the rate of evaporation of water, is directly proportional to the difference between the water partial pressure P w and the saturated vapour pressure Psat. The complication for PEMFCs is that the saturated vapour pressure varies with temperature in a highly non‐linear way, i.e., Psat increases more rapidly at higher temperatures. The saturated vapour pressure for a range of temperatures is listed in Table 4.2. Given the rapid rise in P sat with temperature, air that might be only moderately drying (say 70% RH) at ambient temperature can be fiercely drying when heated to about 60°C. For example, for air at 20°C and 70% RH, the pressure of the water vapour in the mixture is: Pw 0.70 Psat 0.70 2.338 1.64 kPa (4.5) If this air is then heated to 60°C, at constant pressure, without adding water, then Pw will not change and so the new RH will be: Pw Psat 1.64 19.94 0.08 8% (4.6) This is very dry — more extreme than in the Sahara desert, for example, where the RH is typically about 30%. Such a condition would have a catastrophic effect on polymer electrolyte membranes, which not only require high water content but also are very thin and therefore prone to drying out rapidly. The ‘dew point’ is an alternative way of describing the water content. This is the temperature to which the air should be cooled to reach saturation. For example, if the partial pressure of the water in air is 12.35 kPa, then, referring to Table 4.2, the dew point would be 50°C. As indicated in the previous section, it is sometimes necessary to humidify the gases going into a fuel cell to ensure an adequate level of hydration throughout the electrolyte Table 4.2 Saturated vapour pressure of water at selected temperatures. T (°C) Saturated vapour pressure (kPa) 15 1.705 20 2.338 30 4.246 40 7.383 50 12.35 60 19.94 70 31.19 80 47.39 90 70.13 95 96 Fuel Cell Systems Explained membrane. To do this in a controlled way might involve calculation of the mass of water that must be added to the air so that the required humidity is achieved at any pressure and temperature. Given that the mass of any species in a gas mixture is proportional to the product of the molecular mass and the partial pressure and that the molecular mass of air is usually taken to be 28.97, equation (4.3) yields: 18.016 Pw 28.97 Pa mw ma 0.622 Pw Pa (4.7) The total air pressure P is the sum of the dry air Pa and water vapour pressures P w, so: P Pa Pw Pa P Pw (4.8) Substitution of equation (4.8) into equation (4.7) and subsequent rearrangement yields: mw 0.622 Pw ma P Pw (4.9) The water vapour pressure, P w, can be obtained by using data from Table 4.2. The mass of dry air per second required by a fuel cell can be found from equation (A2.10) in Appendix 2. Note that the mass of water needed is inversely proportional to the total air pressure P. Higher pressure systems require less added water to achieve the same humidity. A worked example using equation (4.9) is given in Section 4.4.5. 4.4.3 Air Humidity Previous sections have noted that the humidity of gases within the PEMFC has to be controlled carefully to achieve the optimum level of hydration throughout the membrane electrolyte. Fortunately, it is not difficult to derive a simple formula for the humidity of the exit air. This is given by: Pw Pexit Pw Pexit number of water molecules total number of molecules nw nw nO2 nrest (4.10) where nw is the number of moles of water leaving the cell per second nO2 is the number of moles of oxygen leaving the cell per second nrest is the number of moles of the ‘non‐oxygen’ component of air per second, mainly nitrogen P w is the vapour pressure of the water Pexit is the total air pressure at the exit of the fuel cell If it is assumed that all of the product water from the cell is removed by the cathode air, then equation (A2.17) in Appendix 2 can be used, namely: nw Pe 2Vc F (4.11) Proton‐Exchange Membrane Fuel Cells where Pe is the power of the fuel‐cell stack and Vc is the voltage of each cell. From equation (A2.7), the rate of use of oxygen can be expressed as: nO2 rate of supply of O2 rate of use of O2 Hence: nO2 1 Pe 4Vc F (4.12) where λ is the air stoichiometry. The exit flow rate of the inert components of air (mainly nitrogen) will be the same as at the inlet. These components amount to 79% by volume of air so the flow rate will be proportionately greater than the oxygen molar flow rate, i.e., by the ratio 0.79/0.21 = 3.76, so that: nrest 3.76 Pe 4Vc F (4.13) Substituting equations (4.11), (4.12) and (4.13) into equation (4.10) yields: Pw Pexit Pe 2Vc F Pe 1 4Vc F Pe 2Vc F 2 2 1 3.76 3.76 Pe 4Vc F (4.14) 2 1 4.76 The relationship reduces to: Pw 0.42 Pexit 0.21 (4.15) Thus, it is seen that the vapour pressure of water at the cathode outlet depends only on the air stoichiometry and the air pressure at the exit Pexit. In this derivation, any water vapour in the inlet air has been ignored, and, consequently, the formula represents the ‘worst case’ situation, with dry inlet air. As an example, consider a fuel cell that is operating with an exit air pressure of 110 kPa, a temperature of 70°C and an air stoichiometry of λ = 2. If the humidity of the inlet is low, i.e., any inlet water can be ignored, then substituting these values into equation (4.15) gives: Pw 0.42 110 2 0.21 20.91 kPa (4.16) Referring to the data in Table 4.2 and using equation (4.4), the RH of the exit air is: Pw Psat 20.91 31.19 0.67 67% (4.17) 97 98 Fuel Cell Systems Explained This would be judged too dry and therefore would require attention. The humidity in the cell could be increased by the following: ● ● ● Lowering the cell temperature, which would increase losses. Lowering the rate of air flow and hence λ, which would help a little, but would reduce cathode performance. Increasing the air (and fuel) pressure, which would require energy to run the compressors. Another option is to condense the water from the exit gas and use it to humidify the inlet air. This has an obvious penalty in terms of extra equipment, weight, size and cost, but it may be justified by the increase in performance that is possible. If the water content of the inlet is not negligible, it can be shown that the pressure of the outlet water vapour is given by the slightly more complex formula: Pw 0.42 1 Pexit 0.21 (4.18) where ψ is a coefficient with a value given by: PWin Pin PWin (4.19) where Pin is the total inlet air pressure, which will usually be slightly greater than Pexit and PWin is the vapour pressure of water at the inlet. Equations (4.15), (4.18) and (4.19) therefore provide a means of ensuring that there is adequate humidity in an operating PEMFC. 4.4.4 Self‐Humidified Cells In the example given in the previous section, the exit air from the fuel cell was too dry. By choosing suitable operating temperatures and air flow rates, it is possible to operate a PEMFC that has adequate internal humidification, i.e., self‐humidified. The exit air humidity at different temperatures and air flow rates can be obtained from equation (4.11) or equation (4.13), together with the saturated water vapour pressure taken from Table 4.1. Examples of the exit humidity at air stoichiometries of 2 and 4 are shown in Figure 4.18 for a cell that is operating at 100 kPa. Some selected values are also given in Table 4.4. It can readily be seen that there is a band of operating conditions within which an adequate level of humidification can be achieved. As would be expected, the RH is low at high rates of air flow and falls sharply at high temperatures. The cell will dry out, if the RH of the exit air is much less than 100% since the data of Figure 4.18 are calculated assuming that all the water produced by the cell evaporates. If the calculated RH is above 100%, water will condense out in the electrodes, which will then become flooded. Consequently, in practical terms, there is a narrow band of operating conditions imposed by the requirement to maintain an adequate level of humidification. Providing that the temperature of the cell is maintained below about 60°C, there will be an air flow rate that achieves an RH of around 100%. Some of the conditions are given in Table 4.4. Proton‐Exchange Membrane Fuel Cells 200 180 2 4 Relative humidity/ % 160 140 120 Too wet 100 80 Too dry 60 2 40 20 0 20 4 30 40 50 60 70 90 80 Temperature/°C Figure 4.18 A graph of relative humidity versus temperature for the exit air of a PEMFC with an air stoichiometry of 2 and 4. The entry air is assumed to be dry, and the total pressure is 100 kPa. Table 4.3 Theoretical exit air relative humidities at selected temperatures and stoichiometries. The inlet air is assumed to be at 20 °C and 70% relative humidity; the exit air pressure is assumed to be at 100 kPa. The blanks in the Table are where the relative humidity is too high or too low. Temperature (°C) λ = 1.5 λ=2 λ=3 λ=6 20 30 194 40 λ =12 λ = 24 213 142 117 78 273 195 112 68 45 50 208 164 118 67 40 26 41 60 129 101 72 70 82 65 46 80 54 43 30 90 37 28 An important conclusion from Figure 4.18 and Table 4.4 is that at temperatures above about 60°C (at atmospheric pressure) the RH of the exit air is below, or well below, 100% at all reasonable values of air stoichiometry. In other words, self‐humidification can be achieved for a cell operating at or below 60°C, but for a PMEFC operating above this temperature, extra humidification is usually essential. This feature makes for difficulties in choosing the optimum operating temperature for a PEMFC. The higher the temperature, the better the performance is — mainly because of a reduction in the overpotential at the positive electrode. Once above 60°C, however, the extra weight and cost of the 99 100 Fuel Cell Systems Explained Water circulation Dry air Damp air c e a Damp hydrogen Water circulation Dry hydrogen Membrane electrode assembly (MEA) Figure 4.19 Contra-flow of reactant gases to even the humidification throughout a cell. additional equipment required to humidify the cell can outweigh the benefits of a simple system as is required, for example, for a small air‐breathing fuel cell.18 One of the several ways to achieve self‐humidification is to employ a countercurrent flow arrangement of hydrogen and air, i.e., the air and hydrogen flow in opposite directions across the MEA, as illustrated in Figure 4.19. The water that flows through the membrane from anode to cathode is fairly uniform across the cell, since it is driven by the ‘electro‐osmotic drag’, which is directly proportional to the current density. The back diffusion of water from cathode to anode decreases from the anode inlet to the anode outlet. Even distribution of humidity is also encouraged by the use of thin electrodes and thick GDLs to hold more water and by recycling the fuel gas. Through application of these measures and control of the flow rate of air to match the demands of the load, it is possible to define a band of operating parameters in which a PEMFC stack can be self‐humidified. These requirements, however, are difficult to accomplish for systems above a few watts capacity, and therefore many developers have opted for external humidification as described in the following section. 4.4.5 External Humidification: Principles It has been shown that operating temperatures of over 60°C are desirable to reduce losses, especially the cathode activation overpotential described in Section 3.4, Chapter 3. This objective can be achieved with external humidification and can be demonstrated by revisiting equation (4.18). Consider, for example, a fuel cell that is operating at 90°C with a moderate inlet pressure of 220 kPa, an exit pressure of 200 kPa and a typical air stoichiometry of 2.0. Assuming the air at the cathode inlet to be at 20°C and 70% RH, then equation (4.16) together with values from Table 4.2 yields the following information: ● ● ● ● Inlet water vapour pressure PWin is 1.64 kPa. The Ψ term is 0.00751. Vapour pressure of the water at the exit is 39.1 kPa. Exit air humidity is 56%. 18 An air‐breathing cell is one in which the air electrode is open to the atmosphere and no forced air flow is provided. Such devices are currently being produced for charging mobile phones. Proton‐Exchange Membrane Fuel Cells Under these conditions, the exit humidity is far too low and the membrane would dry out very quickly. If, however, the inlet air is warm and damp, say, 80°C and 90% RH, the following conditions are established: ● ● ● ● Inlet water vapour pressure PWin is 42.65 kPa. The Ψ term is 0.2405. Vapour pressure of the water at the exit is 66.96 kPa. Exit air humidity is 95%. The amount of water that must be added to the inlet air to achieve the exit humidity of 95% can be determined from equation (4.9). For example, if the given fuel cell that is operating at 220 kPa, an exit pressure of 200 kPa and a typical air stoichiometry of 2.0 was a 10‐kW cell, then using equation (A2.10) from Appendix 2 the mass flow rate of dry air (kg s−1) into the cell is given by: a m 3.57 10 7 2 10 103 10 103 0.65 0.011 (4.20) The desired water vapour pressure is 42.7 kPa; see equation (4.2) and Table 4.2. The amount of water (kg s−1) that has to be added to the air, as given by equation (4.9), is therefore: w m 0.622 42.7 0.011 0.0016 220 42.7 (4.21) The rate is approximately equivalent to 100 ml min−1. From where can this water be obtained? Having the water as an extra input to the fuel‐cell system is obviously not desirable so the next best option is to separate it from the cathode exit gas, since the fuel cell generates water by virtue of the electrochemical reactions. For the above 10‐kW cell, equation (A2.17) in Appendix 2 predicts the rate of water production to be 0.0014 kg s−1. The total flow of water in the cathode exhaust is therefore 0.0016 + 0.0014 = 0.003 kg s−1. Given that the water is expelled as a vapour, a condensation or separation system in the exit path must extract rather more than half of the entrained water so that it can be recycled for use in a humidifier. Such a supply arrangement can also ensure that the purity of the water is maintained, but it does make the system more complicated. Before considering some of the practicalities that are involved in humidifying the PEMFC reactants, three factors have to be considered, as follows: 1) It is often not the case that only the air is humidified. To ensure that humidity is even within the cell, the hydrogen fuel may be humidified as well. 2) Humidification involves evaporating water in the incoming gas. The process will cool the gas, as the energy required to evaporate the water will come from the air. This feature is helpful in pressurized systems because the temperature rises when air is compressed. Consequently, the humidification process is an ideal way of lowering the air temperature to match the inlet requirements of the stack (compressors are discussed in Section 12.1.1, Chapter 12). 3) The quantities of water added to the air and the resulting benefits in terms of increased humidity are all much improved by raising the operating pressure. Conversely, lowering the pressure causes more problems. Recalculation of all the 101 102 Fuel Cell Systems Explained values in the example used in this section, for example, but with the 10‐kW stack operating at pressures of 140 kPa (inlet) and 120 kPa (outlet), shows that the mass of water to be added to the inlet air stream becomes 0.003 kg s−1, and yet the exit humidity is hopelessly inadequate at 34%. The operating pressure therefore clearly has a major influence on humidification and is considered further in Section 4.5. 4.4.6 External Humidification: Methods No standard method is used to humidify the reactant gases of a PEMFC stack. The following procedures require a supply of water to be available: 1) Bubbling the gases through water at a controlled temperature. The process is known as ‘sparging’, and it is generally assumed that the dew point of the humidified air is the same as the temperature of the water it has bubbled through, which makes control straightforward. The procedure is suitable for conducting experimental and test work in the laboratory, but is not a preferred method for practical systems. 2) Direct injection of water into the feed gas(es) as a spray. This technique has the advantage that the cold water will cool the gas, an action that will be necessary if the gas has been compressed or if it is hot through being produced by reforming some other fuel. The method requires a pump to pressurize the water together with a valve to open/close the injector. It is therefore fairly expensive in terms of equipment and parasitic consumption of energy. Nevertheless, the practice is based on a mature technology and is widely used, especially for larger fuel‐cell systems. 3) Direct injection as a fine water spray through a metal foam. This approach has the advantage that only a pump is required to move the water — the water is injected passively. 4) Humidification of the GDL via a series of wicks. The wicks dip into the water and draw it directly to the GDL. The system is somewhat self‐regulating, as no water will be extracted if the wicks are saturated. Unfortunately, the method creates a gas‐sealing issue, namely, the wick offers an easy exit route for reactant gases. The possibility of cooling the incoming air is also lost. 5) Direct injection of liquid water into the fuel cell. Normally, such action would lead to flooding of the electrode and consequent failure of the cell. The technique, however, is combined with a bipolar plate that has an ‘interdigitated’ flow‐field design (see Section 4.6.4), which forces the reactant gases to blow the water through the cell and over the entire electrode. The ‘flow‐field’ cut into the bipolar plate is like a maze with no exit, as illustrated in Figure 4.20. The gas is forced under the bipolar plate and into the electrode and thereby drives the water with it. If the flow-field is well designed, a uniform distribution of water will be obtained all over the electrode. Good results have been reported for direct water injection; although there are concerns that it may degrade the electrodes over time. In addition, cooling the incoming air is not possible with this method. Alternatively, the following three methods enable water to be recycled from the cathode exhaust gas: 1) Water in the cathode exit gas is used without condensation to liquid. The practice involves the use of a rotating wheel that contains water‐absorbing or desiccant material. The device is usually called an ‘enthalpy wheel’ and is applied in other technologies, Proton‐Exchange Membrane Fuel Cells Out In Top view Bipolar plate Gas and water driven through electrode Electrolyte Side view, enlarged Figure 4.20 Diagrams to show the principle of humidification using interdigitated flow-fields. (Source: After Wood, DL, Yi, JS and Nguyen, TV, 1998, Effect of direct liquid water injection and interdigitated flow-field on the performance of proton exchange membrane fuel cells, Electrochimica Acta, vol. 43(24), pp. 3795–3809. Reproduced with permission of Elsevier.) such as air‐conditioning systems. Water in the exhaust gas is absorbed on the material, which then rotates so that it is introduced into the path of the dry cathode inlet. The process is continuous — it constantly delivers water from exit to inlet gases and transfers heat from the exhaust stream to the inlet stream. The method suffers from the disadvantages of being fairly bulky and requiring power and control system for its operation. 2) A more sophisticated system, first disclosed by the Paul Scherrer Institute in Switzerland, also uses the exit water without condensation. In this case, a membrane is placed between the cathode exit and the cathode inlet streams. Water vapour in the exit stream condenses on the membrane and then passes through it to the dry inlet side. The membrane can be the same material as the PEMFC electrolyte, and some manufacturers have employed this technique for every cell within a stack. 3) Many developers have sought to modify the PEMFC membrane to enhance water retention. One approach is to generate water in situ. The membrane is modified, not only to retain water but also to produce water. Retention is enabled by impregnating the electrolyte with particles of silica (SiO2) and/or titania (TiO2), together with nanocrystals of platinum. If the membrane is sufficiently thin, the platinum catalyses the reaction between the incoming hydrogen and oxygen to generate water. The reaction of course uses up some valuable hydrogen gas, but it is claimed that the improved performance of the electrolyte justifies the parasitic loss of fuel. 103 104 Fuel Cell Systems Explained 4.5 Cooling and Air Supply 4.5.1 Cooling with Cathode Air Supply An electrical generation efficiency of around 50% may be achieved when converting the chemical energy in hydrogen into electricity in a PEMFC, i.e., heat and electricity are produced in approximately equal amounts. Removal of the heat is essential as the cell has to be cooled to maintain the required operating temperature. If the product water is evaporated within the cell, the heat produced is given by (see Section A2.6, Appendix 2): Heating rate Pe 1.25 1 Vc (A2.21) The way this heat is removed depends greatly on the size of the fuel cell or stack. For fuel cells below about 100 W, it is possible to use naturally flowing air to cool the cell and to evaporate the water produced without recourse to a fan. Similar convective cooling can be applied to stacks that have a fairly open structure with a spacing of between 5 and 10 mm per cell. With a more compact design of fuel cell, small ‘fans’ can be employed to blow excess air through the cell cathodes, though a large proportion of the heat will still be lost through natural convection and radiation. For small systems, such an air fan only imposes a minor parasitic loss of power on the system, namely, about 1%, for a well‐designed system. For systems with stacks producing more than about 100 W, proportionately less heat is lost by natural convection and radiation from and around the external surfaces of the cells. Larger systems therefore require forced cooling in addition to that provided by the cathode air to maintain the necessary low operating temperature. 4.5.2 Separate Reactant and Cooling Air The need to separate the reactant air and the cooling air for anything but the smallest of PEMFCs can be demonstrated by working through a specific example where the reactant gas and the cooling gas are combined. Consider, therefore, a fuel cell of power Pe watts running at 50°C with each cell in the stack at an average voltage of 0.6 V. Suppose that cooling air enters the cell at 20°C and leaves at 50°C. (In practice the temperature change will probably not be so great, but it is instructive to take the best possible case for the present.) Assume also that only 40% of the heat generated by the fuel cell is removed by the air — the rest is radiated or naturally lost by convection from the outer surfaces. The rate of removal of heat by air of specific heat capacity CP flowing at a rate of  kg s−1, and subject to a temperature change ΔT will be the same as the heat produced m according to equation (A2.21). It follows that: 0.4 Pe 1.25 1 Vc  P T mC (4.22) On substituting known values, i.e., CP = 1004 J kg−1 K−1, ΔT = 30 K and Vc = 0.6 V and then rearranging, the following equation is obtained for the flow rate of the cooling air:  1.4 10 m 5 Pe (4.23) Proton‐Exchange Membrane Fuel Cells In Section A2.2, Appendix 2, it is shown that the flow rate of reactant air is:  3.58 10 m 7 Pe Vc (4.24) If the reactant air and the cooling air are one and the same, then these two flow rates are equal. Therefore, combining equation (4.23) and equation (4.24), cancelling Pe, substituting Vc = 0.6 V and solving for λ yields: 14 0.6 0.357 24 (4.25) Reference to Table 4.2 will show that at 50°C, this air stoichiometry produces an exit air humidity of 27%, i.e., dryer than the Sahara desert! The data in Table 4.2 assume that the entry air has a humidity of 70%. Consequently, the RH is decreasing as the air goes through the cell, and this will promote rapid drying out of the PEM. If the assumptions made at the beginning of this section are made more realistic, that is, more heat has to be taken out by the air because the cell efficiency is lower, and the allowable temperature rise in the cell is also lower to maintain adequate humidity, then the situation becomes even worse. The only way to reduce λ, which should be somewhere between 3 and 6 at 50°C in order to prevent the cell from drying out, is to decrease the rate of the air flowing over the electrodes and have a separate cooling system. Such action is necessary when more than about 25% of the heat generated by a fuel cell has to be removed by a cooling fluid. In practice, this applies to cells of about 100 W in size. Fuel cells of greater output will generally need a separate forced supply of reactant air and a cooling system. The usual way to cool cells with outputs from about 100 W to 1 kW is to make extra channels in the bipolar plates through which cooling air can be blown, as shown in Figure 4.21. An alternative approach is to add separate cooling plates for the passage of air. A commercial air‐cooled stack is shown in Figure 4.22. For systems larger than 1–5 kW, air is no longer adequate and water cooling is preferred. 4.5.3 Water Cooling Cooling with air is the simplest option for PEMFC stacks and is generally adopted for those below about 2 kW. Above about 5 kW, the common preference is to employ water cooling given its advantage of not requiring large channels for the cooling medium — 1 kg of water can be pumped through a much smaller channel than 1 kg of air, and the cooling effect of water is much greater. Water‐cooled stacks can therefore be more compact for a given kW size and, additionally, bring the benefit of being able to employ the heat generated by the fuel cell, e.g., in a domestic CHP system. With an air‐cooled system, heat is lost to the atmosphere, whereas heat from a water‐cooled system can be put to practical use. The method of water cooling a fuel cell is essentially the same as for air in Figure 4.21, except that water is pumped through the cooling channels. In practice, such channels are not always necessary or provided at every bipolar plate. The following section considers cooling in more detail for manufactured systems. 105 106 Fuel Cell Systems Explained MEA with sealing gasket on each side Cooling air through these channels Reactant air feed channels Hydrogen feed channels Figure 4.21 Three‐cell PEMFC stack with bipolar plates with separate reactant and cooling air channels. Horizon 5000 W Air-breathing fuel-cell stack Number of cells................................................................120 Rated performance...............................................72 V@70 A Reactants...................................................Hydrogen and Air Ambient temperature................................5 – 30°C (41 – 86°F) Max stack temperature....................................65°C (149°F) Hydrogen pressure.........................................0.45 – 0.55 Bar Humidification...............................................Self-humidified Cooling.......................................Air (integrated cooling fan) Weight (with fan and casing).............................30 kg (±200 g) Controller weight...........................................2500 g (±100 g) Stack size........350 x 212 x 650 mm (13.8 × 8.3 × 25.6 in) Flow rate at max output............................................65 L/min Start-up time.............................≤30 s(ambient temperature) Efficiency of system.............................................40% @72 V Figure 4.22 Example of a commercial air‐cooled stack manufactured by Horizon Fuel Cells. (Source: Reproduced with permission of Horizon Fuel Cells.) Proton‐Exchange Membrane Fuel Cells 4.6 Stack Construction Methods 4.6.1 Introduction Most PEMFC stacks are constructed along the general lines of multiple cells connected in series with bipolar plates, as outlined in Section 1.3, Chapter 1 and illustrated in Figure 1.7. The bipolar plate has to collect and conduct the current from the anode of one cell to the cathode of the next, while distributing the fuel gas over the surface of the anode, and the oxygen/air over the surface of the cathode. Furthermore, the plate often has to carry a cooling fluid though the stack and keep all the reactant gases and cooling fluids apart. The distribution of the reactant gases over the electrodes is achieved with a ‘flow-field’ formed into the surface of the plate. The flow field usually has a fairly complex serpentine pattern. Bipolar plates contribute a high proportion of the cost of a PEMFC stack and have to satisfy several requirements, namely: ● ● ● ● ● ● Good electrical conductivity (>100 S cm−1). High thermal conductivity — this should exceed 20 W m−1 K−1 for normal integrated cooling fluids or must exceed 100 W m−1 K−1 if heat is to be removed only from the edge of the plate. High resistance to chemical attack and corrosion. High mechanical stability, especially under compression (flexural strength >25 MPa). Low gas permeability (<10−5 Pa L s−1 cm−2). Low density — to minimize both weight and volume of the stack. The methods of forming the plates, as well as the materials from which they are fabricated, vary considerably. Similar to humidification, which was considered in the previous section, there is no single method or material that is optimum for every application. Before examining the materials and the manufacturing methods, it should be understood that in most cases the plate is made in two halves. Cooling channels are cut on the back of one of the half‐plates. Whereas this design simplifies the incorporation of cooling within the assembled plates, a significant electrical resistance can arise between the two half‐plates when they are pressed together. The latter feature is especially a problem for carbon plates. For metal plates, there are several methods (e.g., welding, diffusion bonding) that can be used to join half‐plates. The first bipolar plates were machined from sheets of graphitic carbon, and this material remains a good choice for stationary applications of fuel cells where longevity is required and compactness is less of a priority. For service in vehicles, stainless steel tends to be favoured given that it can be formed into very thin sheets by means of processes comparable with those practised in the automotive industry. Metal bipolar plates usually require a surface coating to protect against both chemical attack and corrosion. Nearly all PEMFC stacks incorporate either carbon or metal bipolar plates, as described in the following two subsections. Other types of cell connection and topologies have been investigated for some smaller systems, and examples are given at the end of this section. 4.6.2 Carbon Bipolar Plates Graphite has high electrical and thermal conductivity. It also has a very low density, i.e., less than for any metal that might be considered suitable for a bipolar plate, as well as good resistance to chemical attack. Consequently, the earliest PEMFCs had graphite 107 108 Fuel Cell Systems Explained bipolar plates into which the flow‐field channels were machined. Graphite does, however, have the following three disadvantages: 1) The plates must have a thickness of few mm to provide the mechanical integrity required for machining and handling, even though the latter activity may be automatic; cutting of graphite is time‐consuming with an expensive milling machine. 2) Graphite is brittle, so the resulting cell demands careful handling. 3) Graphite is quite porous, so plates have to be coated and made sufficiently thick to ensure that the reactant gases are separated. Therefore, although the density of graphite is low, the final bipolar plate may not be particularly light in weight. These disadvantages are addressed by using composite materials that combine graphite powder with a polymer binder. Such materials have also been used for PAFCs. Most state‐of‐the‐art carbon bipolar plates for the PEMFC are made from a composite that consists of a high loading of conductive carbon (e.g., graphite, carbon black or nanotubes)19 with a commercial thermoplastic polymer binder (e.g., polyethylene, polypropylene or polyphenylene sulfide) or thermosetting resins (phenolic or epoxy resins). Carbon fibres are often added to strengthen the finished product. Since the physical properties of the composite carbon materials are largely determined by the polymer binder, bipolar plates can be formed either by compression moulding or by injection moulding. The former procedure requires the use of a thermoplastic polymer and is usually the chosen method for limited quantities of small stacks. The mould has a top and bottom part. Graphite powder mixed with polymer is spread over the lower part of the mould. The top is lowered in place and pressure applied. The temperature is then raised above the melting point of the polymer so that the materials mix and flow to fill the mould. After cooling, the product can be released. High carbon content can be achieved with compression moulding. The process is simple but unfortunately slow — production times of 15 min are typically needed for each half‐plate. Variations of the compression moulding technique have been published. For example, the porous material that normally results from compression moulding can be made less permeable by applying a coating of solid carbon on the back of the plate. This modification is realized by chemical vapour infiltration, which is a standard technique easily adapted to mass production. In another procedure, carbon black is taken as the raw material for compression moulding, and the resulting plate is heated to a high temperature (above 2500°C) to cause graphitization. Although this method can improve the electrical conductivity, it can also give rise to warped and brittle sheets. Injection moulding is an attractive proposition for larger and more complex components and involves the participation of a thermosetting polymer. The process is, however, very demanding in that the composite has to be sufficiently fluid to flow into the mould, while at the same time have an adequate carbon loading to achieve good electrical conductivity. Thermosetting polymers also have very different properties to their thermoplastic counterparts. Consequently, composites of the former are injected as a powder into a mould at a temperature below the melting point of the thermosetting polymer, whereas the latter composites are heated above the melting point of the thermoplastic polymer and injected into a mould that is maintained at a lower temperature. 19 Natural graphite flakes have also been used; these are processed to produce a continuous foil that can then be made gas‐tight by incorporating a polymer binder. Proton‐Exchange Membrane Fuel Cells Whereas thermosetting polymers may allow higher loadings of carbon, a time‐ consuming curing step is required after the moulding is complete. With both techniques, the surface of the resulting injection‐moulded plate has to be cleaned, e.g., with an abrasive, to remove any film of polymer that would limit the electrical contact with the GDL. Although many of the details are proprietary, several companies are now manufacturing bipolar plates with thermoplastic‐bonded graphite structures. These alternatives offer a route to reduced costs through mass production. 4.6.3 Metal Bipolar Plates Metals have advantages over carbon in that they are good conductors of heat and electricity, can be machined easily and are not porous. The main disadvantages are that they have higher density and are prone to corrosion — the hot oxygen and water vapour inside a PEMFC is quite a corrosive environment. In addition, there is sometimes a problem of acid leaching out of the MEAs, and therefore it is common practice to coat metal bipolar plates with a corrosion‐resistant material. Plates made of stainless steel, titanium, aluminium and several alloys have been tested with separate coatings of a conductive carbon–polymer material, a transition metal that forms a passive oxide layer (e.g., molybdenum, vanadium or niobium), or a noble metal (e.g., gold). Metal sheets can be machined in similar fashion to graphite sheet, but the process is expensive. Nevertheless, the fact that metals are not as brittle and porous as carbon means that thinner plates can be made. If the plates are sufficiently thin, flow fields can be made by stamping the patterns onto the metal plates. Although stamping is widely practised industrially, it is difficult to achieve extremely narrow channels (with depth and width measured in mm) by stamping the metal. An alternative approach is to use perforated metal or metal foam to form the flow fields. One such method has been described by Murphy et al.20 who have chosen titanium as the metal. The electrical conductivity of titanium is relatively low compared with metals such as copper that exhibit good conductivity — nevertheless, it is some thirty times greater than graphite. As a material for a bipolar plate, titanium can also be made sufficiently corrosion‐resistant by coating with titanium nitride. Such an electrically conductive coating can be applied inexpensively on a large scale. The method adopted by Murphy for creating a bipolar plate is to employ two sheets of metal foam with a thin layer of solid metal between them. The concept is shown in Figure 4.23. The pores or voids in the foam sheets serve as pathways for the diffusion of gas to the electrodes. The reactant gases are fed to the edges of the foam sheets via porous plastic gaskets that are sealed to the periphery of the plate. Metal foam can also be used to effect the cooling of a stack. To do this, one sheet of the metal foam is placed between two sheets of solid (but thin) metal sheets. Water is passed through the metal foam to carry away the heat. The method offers the advantage of using readily available materials (metal foam sheet is made for other applications) to make fuel‐cell components that are thin, lightweight, highly conductive and serve to separate the reacting gases. Furthermore, the only manufacturing processes involved in employing foams are the cutting and moulding of the plastic edge seals. There are many 20 Murphy, OJ, Cisar, A and Clarke, E, 1998, Low cost light weight high power density PEM fuel‐cell stack, Electrochim. Acta, vol. 43(24), pp. 3829–3840. 109 110 Fuel Cell Systems Explained Metal sheet Edge seals with holes to feed gas into metal foam Complete bipolar plate Metal foam Metal foam Figure 4.23 Diagram showing bipolar plate construction from metal foam. (Source: After Murphy, OJ, Cisar, A and Clarke, E, 1998, Low cost light weight high power density PEM fuel‐cell stack, Electrochim Acta, vol. 43(24), pp. 3829–3840. Reproduced with permission of Elsevier.) ways in which this basic idea can be adapted, but in each case the foam needs to be coated to protect against corrosion. In summary, coated metal or metal foam bipolar plates have been successfully applied to vehicle fuel‐cell stacks where corrosion stability is not as demanding as for stationary power generation. For vehicles, a minimum stack lifetime of 2 000 h is required, whereas for stationary systems 40 000 h is expected and is now routinely achieved with carbon bipolar plates. The rapid growth in the application of 3‐D printing in recent years has not been ignored by fuel‐cell researchers. Many research groups have reported the fabrication of both carbon composite and metal bipolar plates using 3‐D printing methods, which are particularly promising for printing the flow‐field patterns described in Section 4.6.4. Unfortunately, current methods lack the precision for fabricating large numbers of components and have generally been employed only for creating prototype stacks.21 4.6.4 Flow‐Field Patterns In the bipolar plate illustrated in Figure 1.12, Chapter 1, the reactant gases are fed over the electrode in a simple pattern of parallel grooves. The other three basic types of flow‐ field pattern chosen for bipolar plates are the ‘pin (or grid)’, ‘interdigitated’ and ‘serpentine’ varieties; the four alternatives are illustrated in Figure 4.24a–d. The design of flow field is influenced both by the plate material itself and by the adjoining GDL. The aim with any design of flow-field is to ensure that humidity is balanced throughout the cell and that gases can flow readily to and from each GDL. It is also desirable to minimize 21 Gould, BD, Rodgers, JA, Schuette, M, Bethune, K, Louis, S, Rocheleau, R and Swider‐Lyonsa, K, 2015, Performance and limitations of 3D‐printed bipolar plates in fuel cells, ECS Journal of Solid State Science and Technology, vol. 4(4), pp. 3063–3068. Proton‐Exchange Membrane Fuel Cells (a) (b) (c) (d) (e) (f) (g) (h) Figure 4.24 Main flow‐field patterns used in PEMFC bipolar plates: (a) pin or grid type, (b) parallel channels, (c) interdigitated, (d) single serpentine channel (e–h) oxygen concentration in the cathode GDL for each of the flow‐field patterns (schematic). Red, high oxygen concentration; blue, low oxygen concentration. (Source: From Heinzel, A, Mahlendorf, F and Jansen, C, 2009, Bipolar plates, in Garche, J, Dyer, C, Moseley, P, Ogum, Z, Rand DAJ and Scrosati B (eds.), Encyclopedia of Electrochemical Power Sources, pp. 810–816, vol. 2, Elsevier, Amsterdam.) the pressure drop through the flow-field. Inevitably, some compromise has to be reached. As has been remarked earlier, the configuration of flows on either side of the MEA, i.e., co‐flow, crossflow or counterflow, also has an important bearing on cell performance. The parallel‐channel arrangement, Figure 4.24b, can be used when there is little likelihood of the formation of water droplets, which otherwise could accumulate, blocking some the channels and thereby cause poor distribution of current throughout the cell and stack. The development of droplets can be minimized by ensuring that flows are large, e.g., in the case of the fuel flow channel, by recycling of the fuel gas. The pin flow-field, Figure 4.24a, is also best suited to applications that require high reactant flows and low utilization of fuel and oxygen. In this arrangement, gases can swirl all over the face of the electrode. Unfortunately, any slight perturbation in the flow can lead to a path of least resistance through the flow field. Such behaviour can engender uneven distribution of reactants, especially on the oxygen side. The interdigitated flow field, Figure 4.24c, consists of a large number of dead‐ended channels. In this arrangement, the gases are forced to flow through the GDL, which therefore has to have adequate porosity and hydrophobicity. Since the efficacy of these two parameters may decline as the GDL ages, this pattern is not generally favoured. The serpentine pattern is preferred by most PEMFC manufacturers. It offers a good compromise between the issues of pressure drop and water removal. The large number of turns in the flow path means that pressure drop is compromised, and this is overcome usually by employing multiple parallel serpentines rather than the single example shown in Figure 4.24d. 111 112 Fuel Cell Systems Explained The flow of fluids within the channels of a bipolar plate is difficult to measure directly. Accordingly, considerable work has been undertaken to model flows with the assistance of fluid‐flow finite‐element methods. Many simulations of gas flows within PEMFC stacks have been published in the literature. For example, the concentration of oxygen at the cathode GDL predicted for the four different flow-fields is presented in Figure 4.24e–f. Several research teams have also developed the technology to measure the current density distribution within an operating fuel‐cell stack. This is commonly done by inserting a thin sensor plate between adjacent cells. The plate usually has a large number (typically over 100) of individual segments that are electrically insulated from each other, but can be connected by fine wires to a measuring device. Each segment imposes a small electrical resistance at regular intervals between adjacent cells, and measurement of the voltage developed across the resistances enables determination of the local current density. Data obtained from working stacks not only help to validate models but also assist in optimizing the design of flow-fields and other cell components. 4.6.5 Other Topologies The construction of fuel‐cell stacks through the application of bipolar plates gives very good electrical connection between one cell and the next. On the other hand, the use of bipolar plates necessarily means that there are many joints with the potential for causing leakage of both reactant gas and cooling fluid. The supply of reactant gas to each and every positive electrode has to be kept separate from that to each and every negative electrode. The entire edge of each anode and cathode is also respectively subject to leakage. Good quality control in the production of components minimizes the risk of leakage, but necessarily means high fabrication costs. In situations where a fuel cell is operated at a fairly low current density, it is often useful to compromise the electrical resistance of the cell interconnects in the interest of simpler and cheaper manufacturing methods. This option is available because the flexibility and ease of handling of the MEAs used in PEMFCs allows for different types of construction other than the traditional bipolar plate. One such design is the system of three cells shown in Figure 4.25. The main body of the unit (depicted in light grey) would normally be made of plastic material. There is only one chamber containing air and only one containing hydrogen. By passing a metal connector strip through the reactant gas separator, the cells are joined in series with the edge of one cathode connected to the edge of the next anode. For even less chance of leaks, the connection MEA with gas-diffusion layers above and below. Perforated metal current collectors are attached to the gas diffusion layers c c Air c a a Hydrogen a + Figure 4.25 Method of connecting fuel cells in series to simplify reactant gas supply. Proton‐Exchange Membrane Fuel Cells could be made externally but this would increase the current path. In this design, the potential for leaks is greatly reduced as the only seals are those around the edges of the MEAs, and the challenge of maintaining uniform humidification of the cells is simplified since there is fairly free circulation of reactant gases in the cell. In practice, however, the arrangement is not compact and is therefore only suitable for low‐power systems. Small PEMFCs can also employ a cylindrical stack design. For example, a microtubular design was designed by NASA for small electronics applications. In most cases, however, the fuel cell is preferably assembled on a planar substrate as a two‐dimensional device with edge connections between cells. Printed‐circuit and microelectromechanical system (MEMS) technologies are preferred as these are amenable to well‐established mass-production processes. The MEMS approach has given rise to designs such as those illustrated in Figure 4.26. Such small systems are beginning to enter the market after many years of research and development.22 For example, Ultracell has an exclusive licence with Lawrence Livermore National Laboratory for a micro‐fuel‐cell system targeted principally at military applications. Subtle variations of the traditional planar fuel cell have been investigated in an attempt to increase the power density of stacks. Intelligent Energy has pursued the design of an air‐breathing stack in which each cell is constructed from a stainless steel anode current-collector with flow-fields for the hydrogen on top of which sits the MEA, and (a) (b) Air-breathing holes Current collector Computational domain Figure 4.26 Examples of MEMS‐based air‐breathing PEMFCs. (a) Six cells sealed with epoxy resin (Si and glass wafers, 1.2 cm × 1.2 cm single‐cell active area, ~4 cm3, >140 mW cm−2 peak power with H2)23. (b) PCB‐based air‐breathing PEMFC with other integrated components24. 22 Pichonat, T and Gauthier‐Manuel, B, 2006, Recent Developments in MEMS‐based micro fuel cells, DTIP, Stresa, Lago Maggiore, Italy. TIMA Editions 6p. <hal‐00189312>. Available https://hal.archives‐ overtes.fr/hal‐00189312 (accessed on 15 August 2017). 23 Reprinted from Zhang, XG, Wang, T, Zheng, D, Zhang, J, Zhang, Y, Zhu, L, Chen, C, Yan, J, Liu, HH, Lou, YW, Li, XX and Xia, BJ, 2007, Design, fabrication and performance characterization of a miniature PEMFC stack based on MEMS technology, International Journal of Electrochemical Science, vol. 2, pp. 618–626. 24 Reprinted from Hwang, JJ and Chao, CH, 2007, Species‐electrochemical transports in a free‐breathing cathode of a PCB‐based fuel cell, Electrochimica Acta, vol. 52, pp. 1942–1950. 113 114 Fuel Cell Systems Explained Metallic, porous, conductive and water-retaining cathode current collector More cells Channels through which hydrogen flows Insulator Stainless steel cell interconnector Stainless steel anode current collector and flow-field for hydrogen MEA with gasdiffusion layers Open structure allows free circulation of cooling and reactant air More cells Figure 4.27 Structure PEMFC demonstrated by Intelligent Energy. then a porous metal current collector sits on top of the cathode. Patented and proprietary techniques are employed to fabricate the cathode current-collector from sintered, stainless steel powder of carefully graded size. The result is a material that is metallic, corrosion resistant, porous, strong, conductive and water retaining. A fuel‐cell stack is assembled by placing the self‐contained cells one on top of the other; a simple piece of folded stainless steel connects the anode of one cell to the cathode of the next. The arrangement is shown in Figure 4.27. Hydrogen is piped via thin plastic tubing to each anode. The open structure of the cell allows for free circulation of air, though this may be fan assisted. 4.6.6 Mixed Reactant Cells In all conventional fuel cells, fuel and oxidant are supplied as separate streams to the anode and cathode, respectively. By comparison, in a mixed reactant fuel cell (MRFC) a mixture of fuel and oxidant flows through the cell as a single stream. The concept first appeared in the literature in the 1960s and is attractive in that there is no requirement for the gas‐tight seals, which are necessary for manifolds and for separating air and fuel systems in conventional stacks. By avoiding the cost and weight of bipolar plates in the design, significant cost reduction should also be possible. Similarly, some simplification of the balance‐of‐plant is to be expected. An MRFC requires the following properties: ● The cathode catalyst should support the reduction of oxygen and not the oxidation of fuel, i.e., mixed potentials should not be possible. Proton‐Exchange Membrane Fuel Cells ● ● The cell should operate at a low enough temperature to avoid the spontaneous thermochemical reaction between fuel and oxidation that may occur in the bulk reaction mixture or on catalyst surfaces. The electrode structures, i.e., the GDLs, should enable the fuel and oxidation to reach the anode and cathode catalyst layers, respectively, by controlling the diffusion of the species. Alternatively, the electrode catalysts should have sufficiently different reaction kinetics to ensure that the fuel oxidation reaction and the ORR are separated. In reality, none of these three properties can be 100% effective, with the result that the cell voltage and energy efficiency of the MRFC is compromised. Rather, the issue is whether such deficiencies are offset by potentially lower capital costs and higher power densities that, in some applications, may favour MRFCs over conventional systems. In principle, several different types of MRFC could be constructed according to the type of electrolyte and cell reactions. This includes cells based on PEMFC, AFC and SOFC materials. One of the first MRFCs to incorporate PEMFC materials was reported in 2002.25 There followed in 2004 a direct methanol MRFC with a Pt–Ru–C anode catalyst and a Ru–Se–C cathode catalyst, from which power densities of approximately 50 and 20 mW cm–2 could be obtained at 90°C with oxygen and air fed cathodes, respectively.26 The mixed reactant DMFCs did not exhibit parasitic direct reaction of methanol with oxygen. 4.7 Operating Pressure 4.7.1 Technical Issues Although small PEMFC stacks are operated at normal air pressure; larger stacks of 10 kW or more are sometimes run at higher pressures. Increasing the operating temperature increases the cell voltage, but as mentioned in Section 4.2.1, the PSFA membranes need to remain hydrated. At atmospheric pressure this limits the operating temperature to about 80°C. Raising the pressure enables the temperature to be increased. Energy is consumed, however, in compressing the fuel and air, and may not be recovered from the fuel‐cell exhaust streams. The simplest type of pressurized PEMFC system is that in which the hydrogen is supplied from a high‐pressure cylinder. Such as system, as employed for example by Hydrogenics, is shown in Figure 4.28. Only the air has to be compressed. The hydrogen gas is fed from a pressurized storage container to the fuel‐cell anodes. The fuel side of the stack is ‘dead-ended’, i.e., there is no exhaust stream for the fuel gas; it is all consumed by the cell.27 The compressor for the air has to be driven by an electric 25 Priestnall, MA, Kotzeva, VP, Fish, DJ, and Nilsson, EM, 2002, Compact mixed‐reactant fuel cells, Journal of Power Sources, vol. 106, pp. 21–30. 26 Scott, K, Shukla, AK, Jackson, CL, Meuleman, WRA, 2004, A mixed‐reactants solid‐polymer‐electrolyte direct methanol fuel cell, Journal of Power Sources, vol. 126(1–2), pp. 67–75. 27 Dead‐ended systems usually release a very small amount of the fuel gas from the system at regular intervals to avoid build‐up of contaminants in the negative electrode. It is also common to recycle some of the fuel gas back to the inlet of the stack (shown by dotted lines in Figure 4.28). Again, this helps to purge contaminants and maintain uniformity of humidity throughout the negative electrode. 115 116 Fuel Cell Systems Explained Anode pressure regulator Hydrogen recycle compressor Hydrogen purge PEM fuel cell Anode Compressed hydrogen storage Motor Electrolyte Cooler/ humidifier Cathode Exhaust air and water vapour Air compressor Air intake Figure 4.28 Schematic representation of a simple PEMFC system, as employed in Hydrogenics fuel-cell modules. (Source: Reproduced with permission of Hydrogenics.) motor, which of course uses up some of the valuable electricity generated by the fuel cell. Note that for the system outlined in Figure 4.28, the pressure of the hydrogen at the anode side of the stack can be controlled as a function of the pressure on the cathode side that, in turn, is determined by the power delivered to the air compressor. Consequently, the differential pressure developed between the two sides of the stack can be maintained at a constant low level to minimize the risk of gas crossover. In a worked example in Appendix 3, it is shown that the typical power consumption by the air compressor will be about 20% of the fuel‐cell power for a 100‐kW system. Compression also raises the temperature of the air so that cooling may be necessary before its entry to a PEMFC, so‐called ‘intercoolers’ operate similarly in internal combustion engines. When the hydrogen fuel is derived from other hydrocarbons, such as methane, the situation is much more complex. Depending on the design of the reformer (described more fully in Chapter 10), the fuel gas is likely to contain other components in addition to hydrogen. In such a situation, running the fuel cell ‘dead-ended’ is therefore not an option, and the exhaust gas stream from the anode may contain a significant amount of unconverted hydrogen. Clearly, this fuel cannot be wasted, and the role of the fuel‐cell system designer is to make sure that any energy in the exhaust stream is utilized effectively. For example, unreacted hydrogen may be burned and the energy release directed to compression of the fuel gas, or it may provide heat for the endothermic reforming reaction. Proton‐Exchange Membrane Fuel Cells 4.7.2 4.7.2.1 Benefits of High Operating Pressures Current The increase in power that results from operating a PEMFC at elevated pressure is mainly the result of the reduction in the cathode activation overpotential, as discussed in Section 3.4, Chapter 3. The increased pressure raises the exchange‐current density, which in turn causes an increase in the open‐circuit voltage (OCV) of the cell, as shown in Figure 3.4, Chapter 3. Note that, however, there is sometimes a reduction in the masstransport losses, with the result that the cell voltage begins to decline at a high current density. The influence of pressure on cell performance can be appreciated from the graph of voltage against current given in Figure 4.29. In simple terms, for most values of current density, the voltage is raised by a fixed value. Although not shown by the graph, this voltage ‘boost’, ΔV, is proportional to the logarithm of the pressure rise. The feature is observed experimentally and has a theoretical basis. In Section 2.5.4, Chapter 2, it was noted that the rise in OCV due to the change in Gibbs free energy can be expressed as: RT P ln 2 4F P1 V (2.45) As given by equation (3.8) in Chapter 3, the activation overpotential is related to the exchange‐current by a logarithmic function. Therefore, to a first approximation, it follows that an increase in pressure from P1 to P2 will promote an increase or gain in voltage, i.e., V gain C ln P2 P1 (4.26) Cell voltage Higher pressure P2 The voltage change is fairly constant at most currents Normal atmospheric pressure P1 Higher current before masstransport losses become important Current Figure 4.29 Effect of increasing pressure on the voltage versus current relationship for a typical fuel cell. 117 118 Fuel Cell Systems Explained where C is a parameter with a value that depends not only on how the exchange‐current density, io, is affected by pressure, but also on the temperature. Various values for C of between 0.03 and 0.10 V are quoted in the literature; this parameter is also influenced by the level of cell humidification. The simple system shown in Figure 4.28 is a useful basis for reaching an understanding of the cost benefit offered by pressurization. For this system, the advantage lies in the greater electrical power obtained from the fuel cell. The increase in voltage for each cell in the stack, ΔVgain, is expressed by equation (4.26). To quantify the power gain, consider a current of I amps flowing through a stack of n cells. The increase in power (watts) is then given by: Power gain C ln P2 I n P1 (4.27) Some of the power produced by the fuel‐cell stack is required to drive the air compressor. As shown later by equation (12.10), Chapter 12, an equation can be written for the power consumed in terms of the compressor efficiency, ηc, the entry temperature of the air T1 and the pressure ratio P2: P1, namely: 1 Compressor power C P T1 C P2 P1  1 m (12.10)  is the flow rate of the air, in kg s−1. This is the power required by the In this equation, m compressor’s rotor. If the efficiency of the motor and drive system is expressed as ηm, then the electrical power required by the compressor will be greater by a factor of 1/ηm. Therefore, the electrical power required by the compressor to achieve the desired pressure ratio P2: P1 will be given by: 1 Power required by compressor C P T1 P2 P1 m C  1 m (4.28) As already discussed earlier in this chapter, equation (A2.10) in Appendix 2 shows  is related to the fuel‐cell electrical power output, the average cell that the parameter m voltage and the air stoichiometry, i.e.,  3.58 10 m 7 Pe Vc (A2.10) Substituting this relationship, electrical power Pe air into equation (4.28) yields: Compressor power 3.58 10 4 T1 m C P2 P1 nIVc and the values of CP and γ for 0.286 1 In (4.29) Proton‐Exchange Membrane Fuel Cells The effect of the loss due to the compressor can also be expressed as a voltage loss, ΔVloss, simply by dividing the power given in equation (4.29) by the total current, I, and for the number of cells in the stack, n, thus: Vloss 3.58 10 4 T1 m C P2 P1 0.286 1 (4.30) The equations now provide a quantitative means of estimating whether a pressure increase will improve the net performance of the fuel‐cell system. Equation (4.26) provides the voltage gain by the fuel‐cell stack, and equation (4.30) can be used to estimate the voltage loss due to the compressor. It is possible to plot values of: Net V V gain (4.31) Vloss for different values of P2/P1, and two examples are given in Figure 4.30, one case is designated ‘optimistic’, the other ‘realistic’. For these examples, the values of the various parameters required in equations (4.26) and (4.30), i.e., C, T1, ηm,ηC and λ are given in Table 4.4. 0.02 ‘Optimistic’ model Net voltage change/V 0.015 0.01 0.005 0 1 2 3 4 5 6 7 Pressure ratio –0.005 –0.01 –0.015 ‘Realistic’ model –0.02 Figure 4.30 Net voltage change that results from operating at higher pressure — for two different PEMFC designs. Table 4.4 Parameters for the examples given in Figure 4.30. Optimistic model Voltage gain constant (C), V Inlet gas temperature, °C 0.10 15 Realistic model 0.06 15 Efficiency of drive for electric compressor (ηm) 0.95 0.90 Compressor efficiency (ηC) 0.75 0.70 Air stoichiometry (λ) 1.75 2.0 119 120 Fuel Cell Systems Explained For the optimistic model, there is a net gain of about 17 mV per cell when the pressure is boosted by a ratio of about 3, but the gain diminishes at higher pressures. For the more ‘realistic’ model, however, there is always a net loss as a result of the higher pressure. The power gained is always exceeded by the power needed to drive the compressor. This shows clearly why operating at above atmospheric pressure is by no means beneficial even with larger PEMFCs. 4.7.3 Other Factors From the elementary analysis just given, one may wonder why pressurized operation should be considered at all. The reason is that although it is the simplest to quantify, the voltage boost is not the only benefit from operating at higher pressure. Similarly, the power required by the compressor is not the only loss. High pressure can also enhance fuel reforming. Whereas thermodynamics shows that hydrogen production by the steam reforming of liquid hydrocarbons is favoured by operating at low pressures, the required size and therefore the cost of the reactor hardware are reduced if the operating pressure is increased. Humidification of the reactant air is also favoured by pressurization. Less water is required to achieve the same level humidity in the cell at elevated pressures compared with air at atmospheric pressure — refer back to Section 4.4.3 where equation (4.15) shows that the humidity of the cathode exhaust air is dependent on the cell operating pressure. With large fuel cells, the flow paths will be quite long and narrow. Therefore, the reactant gases have to be pressurized to overcome frictional losses. A challenge for the fuel‐cell system designer is in selecting a blower or compressor that matches the flow rate required and the pressure drop imposed by the architecture of the stack. For example, other than for simple PEMFCs of low power, an air blower or fan will always be required to overcome the pressure drop through the cathode flow-fields of a stack. Such a fan has to be replaced with a generally more expensive compressor for a system operating at pressure. From a practical point of view, therefore, the extra size, weight and cost of high‐pressure compressors compared with low‐pressure blowers have to be considered. In the discussion thus far, it has been assumed that air supplies the cathode with oxygen. There are, however, some fuel cells — notably, in space applications — that run on pure oxygen from pressurized cylinders. In such systems, the operating pressure of the stack will be chosen by balancing the advantage of the higher performance at elevated pressure against the increased weight of the stack that is necessary mechanically to withstand the high internal pressure. The optimum pressure will probably be much higher than for air systems. 4.8 Fuel Types 4.8.1 Reformed Hydrocarbons Up to now in this chapter, it has been generally assumed that the PEMFCs have been running on pure hydrogen gas as the fuel and air as the oxidant. In small systems, this will usually be the case. In larger systems, however, the hydrogen will frequently come from a fuel processing or reforming system that produces carbon monoxide (CO) as a by‐product. A prime example is the steam reforming reaction between methane and steam, i.e., CH 4 H2O 3H2 CO (4.32) Proton‐Exchange Membrane Fuel Cells Whereas some of the high‐temperature fuel cells described in later chapters can use this CO as a fuel, this does not apply to the PEMFC. Any CO in the fuel stream of a PEMFC will be preferentially absorbed on the platinum catalyst in the anode electrode. Consequently, hydrogen fuel is prevented from reaching the active platinum sites, thereby inhibiting the oxidation reaction on the anode. Experience shows that a CO concentration even as low as 10 ppm in the fuel gas degrades the performance of a PEMFC. Therefore, if a reformed hydrocarbon is to serve as a fuel, the CO has to be removed or at least reduced to a very low level. The extraction process is usually carried out in several stages. Initially, CO and steam are passed over a catalyst that promotes the water–gas shift reaction: CO H2O (4.33) H2 CO2 Not all of the CO is converted by this reaction — an equilibrium point, governed by the process conditions, is reached at 250°C, for example, the product gas from a shift reactor will contain 1–2 vol.% of CO. Further process steps are therefore required for reducing the concentration of CO to levels below a few ppm; these steps are described in detail in Chapter 10. The shift reactor and additional processing steps add considerably to the cost and size of a PEMFC system. In some cases, the requirement to remove CO can be made somewhat less demanding by the addition of small quantities of oxygen or air to the fuel stream that is being fed to the PEMFC. At the catalyst sites on the fuel electrode, CO is converted directly to CO2 by reaction with the oxygen. Reported results show, for example, that adding 2 vol.% oxygen to a hydrogen gas stream containing 100 ppm CO eliminates the poisoning effect. On the other hand, any oxygen not reacting with CO will certainly react with hydrogen and thus waste fuel. Also, the method can only be used for CO concentrations below about 100 ppm, which are not the levels found in the product stream of a typical fuel reformer. In addition, the system required to feed precisely controlled amounts of air or oxygen will be fairly complex, as the flow rate has to match carefully the hydrogen supply rate. Another important point to note is that the problem with CO intensifies with hydrocarbons of increasing molecular length. The initial methane (CH4) reforming reaction (4.32) produces three molecules of hydrogen. By contrast, the processing of a fuel such as n‐octane (C8H18): C 8H18 8H2O (4.34) 17H2 8CO results in a gas where the ratio of H2 to CO is now about 2 : 1. 4.8.2 Alcohols and Other Liquid Fuels For any type of fuel cell, an ideal fuel would be a liquid that is already in regular use, such as petrol or diesel. Unfortunately, these two fuels simply do not react at a sufficient rate to warrant consideration for PEMFC systems. Possible alternatives to hydrogen in a PEMFC are methanol and, to a lesser extent, ethanol; both are widely available commercially. Methanol reacts at the anode of a PEMFC, albeit slowly, according to the equation: CH3OH H2O 6H 6e CO2 (4.35) 121 122 Fuel Cell Systems Explained Note that the methanol needs to be mixed with water and that six electrons are produced for each methanol molecule and that the reaction does not directly produce CO. This is the operational reaction of the DMFC which, together with other fuel cells that operate directly on liquid fuels, is discussed further in Chapter 6. The following section presents three typical applications of PEMFs. All the systems employ stacks that operate with approximately ambient air pressure and use pure hydrogen as the fuel and air as the oxidant. 4.9 Practical and Commercial Systems 4.9.1 Small‐Scale Systems A class of PEMFC stacks with outputs of between a few watts and 1 kW is found in a variety of applications, namely: (i) as battery chargers for portable electronic equipment (e.g., mobile phones and laptop computers), (ii) for military use as personal power sources and (iii) as stationary backup power supplies. Some of the smallest systems use methanol and are described in Section 6.1, Chapter 6. Horizon Fuel Cells, in collaboration with the associated company Horizon Energy Systems of Singapore, has championed small, hydrogen‐fuelled PEMFC systems for several years and now markets a range of air‐breathing stacks with outputs of 12 W to 1.0 kW for portable and educational systems; examples are shown in Figure 4.31. The company also produces the ‘Mini‐pak’ (a) (b) (c) Figure 4.31 Horizon Fuel Cell products: (a) 12‐W ‘H‐Series’, (b) 1‐kW ‘H‐Series’ and (c) ‘Mini‐Pak’ phone charger. (Source: Reproduced with permission of Horizon Fuel Cells.) Proton‐Exchange Membrane Fuel Cells Figure 4.32 Mobile phone charger from Intelligent Energy (the ‘UPP’). fuel cell for charging electronic devices; it uses an air‐breathing stack and a cartridge that contains hydrogen stored as a hydride. Following the introduction of the ‘Mini-pak’ into the US camping and outdoor markets in the United States and Europe, Horizon Fuel Cells teamed up with Brunton to produce the ‘Brunton Hydrogen Reactor’ for charging most pocket devices such as smartphones, iPads, camera batteries, UV water purifiers, rechargeable lights and GPS units. A similar product, known as the ‘UPP’, from Intelligent Energy in the UK is presented in Figure 4.32. Both products use small air-breathing fuel-cell stacks. The UPP device employs a hydride cartridge (90.5 mm × 40 mm × 48 mm, weight 385 g) that can deliver 25 Wh of energy. The stack is a 5‐W PEMFC that is able to produce up to 1000 mA at 5 V. One fuel cartridge will therefore provide a smartphone with approximately five full charges, and it is approved for carriage onboard aircraft. Each cartridge has a life of 9 years and is therefore ready to meet any emergency well within the expected lifetime of the smartphone. 4.9.2 Medium‐Scale for Stationary Applications Several companies are marketing PEMFC systems for backup or stationary power systems, for example, for remote telecommunications towers and data centres. Systems below about 5 kW such as those produced by PlugPower/Relion and Altergy employ air‐cooled stacks. For the reasons given in Section 4.5.3, systems above 5 kW such as those produced by Ballard/Dantherm, Hydrogenics and M‐Field are water cooled. The following description of a Hydrogenics fuel‐cell power module (Figure 4.33) is given as an illustration of a PEMFC product designed for stationary applications such as for data centres. The module employs a water‐cooled stack of fairly conventional design and is composed of 60 cells, each with an active area of 500 cm2, and bipolar plates fabricated from a compression‐moulded carbon–polymer composite. The stack produces about 12 kW at current of 350 A and a nominal voltage of 35–58 V. It is self‐ humidified, that is, there is no external humidification of either the fuel or air streams. The essential balance-of-plant (BOP) items in the Hydrogenics power module is as indicated in the schematic process flow diagram of Figure 4.28. Careful control of gas flow rates and stack temperature (i.e., through flow of cooling water) keeps the stack at 123 124 Fuel Cell Systems Explained Anode recycle compressor Stack Air filter and blower Pressure control valve Figure 4.33 Hydrogenics, rack‐mounted, ‘HYPM’ module, covers removed to show the stack and balance‐of‐plant items. (Source: Reproduced with permission of Hydrogenics.) optimum humidity. Recirculation of the anode exhaust gas helps to maintain even humidity on the anode side of the stack. A differential pressure control valve ensures that the pressure on the air side of the stack closely follows that on the fuel side. There is a pump that circulates the hydrogen through the anodes of the stack, and the fuel loop is ‘dead-ended’. A relief valve purges this line at intervals to prevent a build‐up of contaminants within the anodes. Air is supplied via a blower, which is regulated to provide the correct stoichiometry over the full operating regime of the system. When the module is shut down, both the air and the fuel supplies are switched off by solenoid valves. A small buffer vessel contains sufficient hydrogen so that any remaining oxygen in the cathode air is self‐consumed by the system, and thereby only inert gas remains in the shutdown state. This procedure is claimed to limit degradation and prolong the life of the stack. A process control system is coupled to the Hydrogenics power module and is embedded with software both to monitor the performance of the stack and to adjust parameters, such as hydrogen and air flow, in response to the electrical demand imposed. Other process parameters that are fed to the controller include the stack temperature, cell voltages and the pressure at the fuel side of the stack. The Sankey diagram is a useful way to indicate the various energy flows and power losses in a power‐generating system such as a fuel cell. The energy flows in an earlier version of the Hydrogenics module are represented in the form of a Sankey diagram in Figure 4.34. The diagram shows that only 10 kW of the 25.3 kW of energy embedded in the hydrogen fed to the module appears as useful electrical power, i.e., the module has an efficiency of 39% with respect to the lower heating value (LHV). Most of the energy Proton‐Exchange Membrane Fuel Cells 10.9 kW heat removed via cooling water 2.4 kW heat lost via exhaust gas 2 kW waste heat 25.3 kW hydrogen 12 kW DC power from stack 10 kW total DC electrical power 200 W controller 700 W 1.06 kW UPS loss DC-DC converter Figure 4.34 Sankey diagram of energy flows in a Hydrogenics fuel‐cell power module. that is not converted to electricity is discharged as heat in the cooling water or exhaust gas, or lost to the environment. Since the voltage produced by the stack varies according to the load imposed, a DC–DC converter is employed to increase the voltage to a useful and stable value. For stationary power applications, the DC output is usually converted to AC so as to be compatible with the local network. The Sankey diagram in this case shows that there are electrical losses associated with the DC–DC conversion and in providing power to the system controller and battery uninterruptible power system (UPS) that manages the module. 4.9.3 Transport System Applications When Ballard Power Systems (BPS), a Canadian company, showcased its first PEMFC stacks in the late 1980s, it became clear that this type of fuel cell was well suited for application in electric vehicles. The high power density of PEMFC stacks, together with zero‐emissions when fuelled by hydrogen, attracted companies such as DaimlerChrysler and Shell who bought shares in BPS in 1994. New ventures were set up by Daimler to develop the stacks and drivetrains for vehicles. Daimler built its first vehicle, the NECAR (‘new electric car’), in 1994 and spent the next 20 years improving the fuel‐cell technology through optimization of both the stack and the drivetrain components. The ensuing developments led to the Mercedes B‐class F‐CELL which, in 2009, was the first fuel‐cell car in series production. Geoffrey Ballard, founder of the Canadian company, realized that buses provided a unique opportunity to demonstrate his technology. Buses all refuel at a central depot, 125 126 Fuel Cell Systems Explained they are amenable to be used with novel fuels, and they operate in cities where air pollution is often a major issue. In August 1993, a 21‐seat Ballard bus carried its first public passengers at the Commonwealth Games in Vancouver. This vehicle contained only a small lead–acid starter battery as Ballard wanted to demonstrate that the fuel cell could provide the motive power by itself. Seventeen years later, in 2010, BC Transit had sufficient confidence in the technology to order twenty 12‐m, low‐floor, fuel‐cell buses to carry participants between Vancouver and Whistler for the Winter Olympics. These featured 130‐kW fuel‐cell stacks that were each supplied with hydrogen stored in a tank at a pressure of 36 MPa. Hybridized with a nickel–metal-hydride battery, the buses had a range of some 500 km. Government‐supported demonstration programmes of fuel‐cell buses are now in place throughout the developed world with manufacturers in Europe, Japan and North America. Most recently, both China and India have become involved in such activity through the development of their own PEMFC technology. Many of the fuel‐cell buses on the road are hybrid vehicles and the latest model by Mercedes‐Benz, the Citaro FuelCELL‐Hybrid, incorporates lithium batteries alongside the PEMFC stacks; one of the buses is shown in Figure 4.35. As with previous Mercedes‐Benz buses, hydrogen for the fuel‐cell stacks is stored at pressure in cylinders in the roof shell of the vehicle. The number of cylinders required has been reduced from 9 to 7 (Figure 4.36) on account of improved system efficiency and the use of the lithium‐ion batteries. These batteries (which like the PEMFC stacks are water cooled) have a capacity of 27 kWh that is sufficient to power the wheel‐hub electric motors at a constant 120 kW (165 hp). The fuel consumption is 11–33 kg hydrogen per 100 km — i.e., 50% less in comparison with its predecessor, the Citaro F‐CELL — and the range of the vehicle is 250 km. Figure 4.35 Mercedes‐Benz Citaro FuelCELL‐Hybrid bus. (Source: Reproduced with permission of Daimler.) Proton‐Exchange Membrane Fuel Cells Figure 4.36 Roof compartment of the Mercedes‐Benz Citaro FuelCELL‐Hybrid bus (showing seven storage tanks and lithium batteries behind them). (Source: Reproduced with permission of Daimler.) Fuel‐cell stacks for some of the early converted buses were situated where the engine would be found in a diesel counterpart. The latest Citaro buses, however, have the fuel cells located in the roof shell, at the rear of the bus behind the hydrogen cylinders. The batteries sit between the hydrogen cylinders and the fuel‐cell stacks (see Figures 4.36 and 4.37). Thus all of the drivetrain is essentially mounted in the roof of the vehicle. Other mechanical components required for operation of the bus, such as air‐conditioning pumps, electric‐steering pump, air pump and inverter for auxiliaries, are placed in what otherwise would be the rear engine compartment of a conventional diesel bus. This placement allows easy access for servicing. Vehicle manufacturers adopted Ballard stacks to demonstrate PEMFC technology in cars. A 75‐kW design was the standard for most of the early Daimler fuel‐cell cars. As confidence in the technology grew, however, automotive companies developed their own stack technology. Examples include those built by General Motors, Honda (see Figure 4.2), Hyundai, Nissan, PSA Citroen Peugeot, Toyota and Volkswagen. The Toyota Mirai, launched in 2015, employs a 115‐kW stack that delivers power to a single 114‐kW electric motor. Hydrogen is contained at pressure in two tanks with a combined volume of 122.4 L, and the manufacturer claims that this storage will give a range of up to 650 km. The Hyundai ix35 fuel‐cell car is fitted with a 100‐kW stack and promises a range of 594 km from one charge of hydrogen at 70 MPa. In the United Kingdom, Intelligent Energy is developing fuel cells with various companies, which include the motorbike manufacturer Suzuki, and has announced their own innovative 100‐kW water‐cooled system for vehicles; see Figure 4.38. As explained by the company, the 100‐kW platform takes full advantage of Intelligent Energy’s stack technology 127 128 Fuel Cell Systems Explained Figure 4.37 Roof compartment of the Mercedes‐Benz Citaro FuelCELL‐hybrid bus (showing the fuel cell stacks and associated BOP). (Source: Reproduced with permission of Intelligent Energy.) (a) (b) Figure 4.38 Intelligent Energy 100‐kW fuel‐cell system for vehicles: (a) water‐cooled stack and (b) packaged system. (Source: Reproduced with permission of Intelligent Energy.) that offers a power density of 3.5 kW L−1 and a specific power of 3.0 kW kg−1, while being engineered for low‐cost, high‐volume series production. The key to this performance is said to be the proprietary, evaporatively cooled (EC) technology. The stack employs metal separator plates, and compared with conventional liquid‐cooled fuel‐cell stacks, the EC design is said to remove the need for individual cooling channels between each cell and thereby delivers considerable advantages in terms of the reduction in both stack volume and mass. The technology also indicates that there continues to be room for innovation in PEMFC technology, which bodes well for the future of applications such as fuel‐cell vehicles. Proton‐Exchange Membrane Fuel Cells 4.10 System Design, Stack Lifetime and Related Issues Many years of research have shown that durable, high‐performance and low‐cost PEMFCs can be achieved through the appropriate combination of materials, design and operating conditions. Investigations have also helped to identify the means by which the following modes of cell degradation may eventuate. 4.10.1 Membrane Degradation Mechanical degradation may be caused by swelling of the membrane through, for example, poor water management. The membrane can also breakdown as a result of chemical reaction by foreign elements such as precipitated platinum from the catalysts or iron from metal bipolar plates. Degradation can occur through peroxide formation by the ORR. It has been shown that the aggressive action of peroxide is accelerated by the presence of iron and is believed to be due to the generation of hydroxyl (OH–) and hydroperoxyl (HOO–) radicals, which attack the acidic moieties in the polymer membrane. In summary, to avoid degradation of the membrane, it is important to address at least one of the following actions: ● ● ● ● ● ● Reduce peroxide generation or accelerate in situ peroxide decomposition. Remove or passivate iron and other undesirable metal contaminants. Enhance the oxidative stability of the membrane. Improve water management. Reduce time spent at >0.9 V. Ensure membrane is sufficiently hydrated, e.g., by limiting high‐temperature operation. 4.10.2 Catalyst Degradation On the cathode side, sintering or dissolution of platinum may be reduced by alloying with other elements, e.g., cobalt and iridium. More stable catalyst supports, such as graphitized carbon, are also required. Operational benefits include introducing and maintaining hydrogen at the anode when the cell is ‘off ’, i.e., no load, and shorting or applying an immediate load on cell start‐up to remove air from the cathode. Care also needs to be taken to extract any potential catalyst poisons from the fuel and air streams. 4.10.3 System Control Much effort has been devoted to designing appropriate control technology to ensure that a fuel‐cell system operates under conditions that best prolong the lifetime of the MEA. This task requires the performance of the stack to be accurately monitored, and the most convenient method is to measure the voltages of individual cells, or groups of cells, in real time. A microprocessor or a programmable logic controller (plc) system can be used to read, record and analyse the voltages in response to changes in demand by the load. The controller can actuate valves and other devices to change operating parameters such as gas flows, humidifier temperature, and system pressures. Consequently, the controller may, for example, maintain the correct stoichiometries of air or oxygen to enable the cell voltages to remain constant within narrowly defined limits. If the voltage 129 130 Fuel Cell Systems Explained of one cell falls significantly, an alarm situation can be enunciated to indicate that some remedial action may be required to prevent the one cell from causing a reduction in the voltage of the whole stack and thereby accelerate stack degradation. Similar provisions are embodied in the battery management systems of advanced lithium‐ion batteries. If the PEMFC system employs several stacks in modules, the plc will oversee the operation of the whole system to ensure that each module is operating cohesively. For example, when the fuel‐cell system receives a start‐up signal from an operator, the controller gives instructions so that the system follows a pre‐established boot‐up procedure that starts each module at an appropriate time. The strength of using a plc in this role is that, by programming suitable algorithms, it can be employed to detect stack malfunctions, such as unusual variations in cell voltage due to fluctuating rates of gas flow caused by blocked channels in the bipolar plates. To enable stacks to be integrated within vehicle systems, the PEMFC controller is usually made to communicate with other components via a controller area network (CAN bus).28 4.11 Unitized Regenerative Fuel Cells A unitized regenerative fuel cell (URFC) is a reversible cell that is able to operate as a conventional fuel cell and, in regenerative mode, as an electrolyser. When in electrolyser mode, the URFC generates hydrogen and oxygen by the electrolysis of water (see Section 10.8, Chapter 10). Both modes are carried out with the same fuel‐cell stack. In comparison with a separate fuel cell and electrolyser, the combination of these duties in the same hardware holds several advantages, such as lower capital cost, simpler structure, higher specific energy and no need for auxiliary heating. Although both the AFC and SOFC have received some attention as reversible fuel cells, systems based on PEMFC stacks are the most mature. Designs of URFC have already been employed in aerospace applications. Rechargeable secondary batteries are widely used for energy storage purposes due to their high round‐trip efficiency (around 80%), but they suffer from some obvious drawbacks. The durability of lead–acid batteries is not very satisfactory when faced with deep cycling, and their specific energy is constrained by the heavy weight. Lithium‐ ion batteries promise to be much more durable with respect to cycling, but are subject safety issues. Redox flow batteries (RFBs), as described in Section 1.7.2, Chapter 1, have attracted interest because they provide the means of decoupling energy storage capacity and rated power. By enlarging the electrolyte storage tank, the capacity can easily be increased, while the rated power can be enhanced by using electrodes of greater area or through stacking. On the other hand, due to the bulk electrolyte solution contained in the system, the specific energy of RFBs is generally much lower. 28 A controller area network (CAN) bus is a vehicle electronic serial bus standard that is designed to allow microcontrollers and other electrical or electronic devices to communicate with each other in applications without the need for a host computer. CAN bus is a message‐based protocol and, although designed originally for automotive applications, it is used in many other contexts. The modern automobile has many electronic control units for various subsystems. Typically, the biggest processor is the engine control unit. Others are used for items such as transmission, airbags, anti‐lock braking/anti‐skid braking system (ABS), cruise control, electric power steering, audio systems, power windows, doors, mirror adjustment and battery recharging systems for hybrid/electric cars. Proton‐Exchange Membrane Fuel Cells Box 4.1 Pure oxygen versus air in a PEMFC Running a PEMFC with oxygen rather than air as the cathode gas markedly improves cell performance by virtue of the following three effects: 1) The ‘no loss’ open‐circuit voltage rises on account of the increase in oxygen partial pressure, as predicted by the Nernst equation; see Section 2.5, Chapter 2. 2) The activation overpotential reduces through better use of catalyst sites; see Section 3.4.3, Chapter 3. 3) The limiting current increases and thus reduces the mass-transport or concentration overpotential losses. This benefit is due to the removal of the nitrogen gas, which is a major contributor to such losses at high current densities; see Section 3.7, Chapter 3. Depending on the design of PEMFC, a change from air to oxygen can increase the power of the stack by about 30%. In particular, a stack with poor reactant air flow will benefit more from a switch to oxygen. For a URFC system that involves the storage of oxygen and hydrogen, the use of pure oxygen has a significant impact; it may increase the round‐trip efficiency from a typical 35 to 50%. As with flow batteries, URFCs also store the fuel and oxidant, generally H2 and O2, externally in separated gas tanks and therefore offer the ability to decouple storage capacity and output power. By contrast, however, their specific energy is much higher than that of RFBs, i.e., about 0.4–1.0 kWh kg−1 (including the mass of the hydrogen and oxygen gas tanks29) compared with 0.01–0.02 kWh kg−1 for a vanadium redox battery. In addition, URFCs can be totally charged and discharged without damaging the durability of the fuel cell. These advantages have made URFCs very competitive against secondary batteries and flow batteries. On the debit side, however, URFCs generally achieve lower round‐trip efficiency than batteries (typically below 40%) due to the sluggish reactions for oxygen evolution and oxygen reduction. Efficiency can be increased if hydrogen and oxygen are stored and used (see Box 4.1). The low efficiency would also be more tolerable if the URFC could be employed in a cogeneration system, where the heat that is generated in fuel‐cell mode could be harnessed. Other issues such as high cost, hydrogen storage, and relatively low technology readiness, have also hindered their exploitation. In practice, there are other technical issues concerning PEMFC‐based URFCs. These mainly concern the bifunctional catalyst that has to service both the ORR and the oxygen evolution reaction (OER). To date, most of the bifunctional catalysts utilized in URFCs are based on noble metals. Platinum (Pt), the preferred catalyst for the ORR, is not suitable for the OER. Moreover, the preferred catalysts for the OER, such as ruthenium (Ru), iridium (Ir) and the oxides of the two metals, are not suitable for the ORR. Consequently, a compromise has to be made from the combination of these candidate metals and oxides that delivers the best performance, as composite catalysts. The combination of Pt and Ir or its oxides is currently the preferred choice for a bifunctional 29 Mitlitsky F, Myers B, and Weisberg AH, 1988, Regenerative fuel cell systems, Energy Fuels, vol. 12, pp. 56–71. 131 132 Fuel Cell Systems Explained catalyst. Numerous studies on the optimization of the two metals (e.g., elemental ratio, method of catalyst preparation, microstructure) have been conducted. Carbon, which is the preferred catalyst support material for PEMFCs, is less suitable for URFCs because, on the oxygen side of the cell, carbon corrosion is promoted under electrolysis conditions. For this reason other support materials such as titania, titanium carbide or nitride have been investigated. As noted in Section 4.3.2, the study of catalyst materials, particularly non‐precious metals, for the cathode in PEMFCs is a very active research area, one from which URFCs could also benefit. Similarly, the carbon‐based GDL that is employed in the PEMFC is not suitable for the URFC, and alternatives are under investigation. Despite these issues, several developers have produced URFCs that employ PEMFC‐ type stacks and include the following: ● ● ● ● Distributed Energy Systems (Connecticut, USA) has constructed a multi‐kW, closed‐ loop, lightweight URFC for high‐altitude airships, that can generate pressurized hydrogen and oxygen electrochemically without mechanical compression.30 The NASA Glenn Research Center demonstrated a closed‐loop URF for a solar electric aircraft in 2006.31 The system could store the input electrical energy and output a steady electrical power of 5 kW for at least 8 h. IHI (Japan) collaborated with Boeing to develop a URFC for aircraft auxiliary power units (APUs). Demonstration systems have been produced in the United States by Giner Inc., Lynntech, Lawrence Livermore National Laboratory and Proton Energy Systems Inc. Further challenges for the URFC arise from the management of water consumption in electrolysis mode or production in fuel‐cell mode. Water management is therefore even more complex than that for the PEMFC as described in Section 4.4. Further Reading Barbir, F, 2012, PEM Fuel Cells: Theory and Practice, Academic Press, Waltham, MA. Behling, N, 2012, History of proton exchange membrane fuel cells and direct methanol fuel cells, in Fuel Cells: Current Technology Challenges and Future Research Needs, pp. 423–600, Elsevier, Amsterdam. Koppel, T, 1999, Powering the Future – The Ballard Fuel Cell and the Race to Change the World, John Wiley & Sons, Inc., New York. Gasteiger, HA, Baker, DR, Carter, RN, Gu, W, Liu, Y, Wagner FT and Yu PT, 2010, Electrocatalysis and catalyst degradation challenges in proton exchange membrane fuel cells, in Stolten D (ed.), Hydrogen and Fuel Cells, Fundamentals, Technologies and Applications, pp. 3–16, Wiley‐VCH, Weinheim. Reijers, R and Haije, W, 2008, Literature review on high temperature proton conducting materials: Electrolyte for fuel cell or mixed conducting membrane for H2 separation, 30 Funding, demo for regenerative fuel cell, 2004 Fuel Cells Bulletin, 2004, pp. 7–8. 31 Bents, DJ, Scullin, VJ, Chang, BJ, Johnson, DW, Garcia, CP and Jakupca, IJ, 2006, PEM hydrogen‐oxygen regenerative fuel cell development at NASA Glenn Research Center, Fuel Cells Bulletin, vol. 2006, pp. 12–14. Proton‐Exchange Membrane Fuel Cells Report no. ECN‐E‐‐08‐091, prepared under the KIMEX project no. 7.0330, ECN Research Centre, Petten, the Netherlands. Zhang, J, Xie, Z, Zhang, J, Tang, Y, Song, C, Navessin, T, Shi, Z, Song, D, Wang, H, Wilkinson, DP and Liu, ZS, 2006, High temperature PEM fuel cells, Journal of Power Sources, vol. 160(2), pp. 872–891. Wang, Y, Leung, DYC, Xuan, J and Wang, H, 2016, A review on unitized regenerative fuel cell technologies, part‐A: Unitized regenerative proton exchange membrane fuel cells, Renewable and Sustainable Energy Reviews, vol. 65, pp. 961–977. 133 135 5 Alkaline Fuel Cells 5.1 Principles of Operation The basic chemistry of the alkaline fuel cell (AFC) has been explained in Figure 1.4, Chapter 1. The reaction at the anode is: 2H 2 4OH 4 H2 O 4 e E 0.282 V (5.1) where E° is the standard electrode potential. The electrons released travel round the external circuit to the cathode, where they react to form new OH ions, i.e., O2 4e 2 H2 O 4 OH E 0.40 V (5.2) References to the AFC can be traced back at least to 1902,1 but it was the work of F. T. (Tom) Bacon, first at the University of Cambridge (1946–1955) and then at Marshall of Cambridge Limited (1956–1961), which led to the first practical demonstration of the technology. The Bacon cell was adopted for the Apollo space programme — an example is shown in Figure 5.1 — and this created the general impression that the AFC was an expensive and specialized system. Later, however, Kordesch at Union Carbide (Cleveland, Ohio) and Justi and Winsel at Siemens (Erlangen, Germany) showed that an atmospheric pressure hydrogen–air AFC could work very effectively, with the proviso that carbon dioxide (CO2) must not be present in the fuel or oxidant unless means of either purification or replacement of the electrolyte solution were included. Experimental AFCs were tested for their ability to power agricultural tractors, cars, offshore navigational equipment, boats, forklift trucks and various other applications during the 1960s and early 1970s. Although many of the systems worked reasonably well as demonstrations of proof‐of‐concept, issues such as cost, reliability, ease of use, ruggedness and safety proved to be a challenge. During the 1980s and 1990s, prospects appeared poor for the AFC when compared with other emerging fuel cells. Consequently, research was scaled down so that by the close of the century only a couple of companies were actively working on AFCs. According to most analysts, the emergence of the proton‐exchange membrane fuel cell (PEMFC) heralded the final demise of the AFC, especially when a decision was taken in 1997 to replace the system that had been used for the Space Shuttle Orbiter vehicles with PEMFCs for future missions. 1 Reid, JH, 1902, US Patent no. 736 016 017. Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 136 Fuel Cell Systems Explained Figure 5.1 Alkaline fuel‐cell system (1.5 kW) employing 32 circular fuel cells (200 mm diameter) as used in the Apollo spacecraft. The cells were in the lower container that was purged with nitrogen to remove waste heat. The stack area of 465 cm2 gave 0.86 V at 470 mA cm−2 (4.0 kW m−2). Three of these units were connected in parallel to provide redundancy, each weighed 109 kg. The fuel cells provided the electrical power, as well as much of the potable water, for the craft that took mankind to the moon. (Courtesy of International Fuel Cells.) Despite the lack of interest in the AFC by many developers, it should be pointed out that this technology does offer some technical advantages over the more successful PEMFC and phosphoric acid fuel cell (PAFC) alternatives. The activation overpotential at the cathode of an AFC is generally lower than that in acid fuel cells, and the electrode reactions are faster. It is therefore not essential to use platinum‐based catalysts in the AFC. Furthermore, the efficiency of electrical generation by the AFC is generally greater than that of the PEMFC on account of the lower overpotential at the cathode. Indeed it was the high efficiency of energy conversion — typically, 70% (LHV) — that led to the National Aeronautics and Space Administration (NASA) selecting the AFC for deployment in the US space programme. Conventional AFCs use an alkali electrolyte dissolved in water. Sodium hydroxide (NaOH) and potassium hydroxide (KOH) — the most abundant and cheapest alkaline hydroxides — were the prime candidates for the early AFCs. Unfortunately, however, CO2 present either in the fuel or oxidant streams can react with such hydroxides and cause the formation of potassium or sodium carbonate in the electrolyte solution, for example: 2KOH CO2 K 2 CO3 H2 O (5.3) Alkaline Fuel Cells Such a reaction has the following adverse impacts: ● ● ● ● ● Decrease in the OH− concentration in the electrolyte solution, thereby interfering with the kinetics of the cell reaction. Increase in the viscosity of the electrolyte solution, thereby resulting in a slower diffusion rate and lower limiting currents. Precipitation of carbonate salts in the porous electrode, thereby reducing mass transport. Reduction in oxygen solubility. Decrease in the conductivity of the electrolyte solution. The end result is a severe deterioration in cell performance. Of the two candidate hydroxides, KOH is generally preferred because its carbonate is much more soluble in water than the sodium counterpart. Degradation of the AFC by CO2 was a significant contributor to the abovementioned declining interest in AFCs. Nowadays, however, there is a renewed enthusiasm for the AFC that has been brought about by an improved understanding of the influence of CO2 on cell performance and by the emergence of anionic polymer membranes to replace the traditional electrolyte solutions. 5.2 System Designs 5.2.1 Circulating Electrolyte Solution The design of AFC in which the electrolyte solution is circulated was pioneered by Bacon and subsequently used by Pratt and Whitney (later International Fuel Cells) in the 1950s for the Apollo missions. A schematic representation of essentials of the system is presented in Figure 5.2. In the Bacon design, an aqueous solution of the electrolyte (typically 33 wt.% KOH) is pumped through the fuel cell. Hydrogen is supplied to the anode but must be circulated as water is produced at this electrode.2 When pressurized to 500 kPa, the cell operates at 200°C, and therefore the product water is actually steam and has to be condensed out from the circulating hydrogen. For the Apollo missions, the pressure was reduced to 330 kPa, and the concentration of the electrolyte solution was increased to 85 wt.% KOH. As shown in Figure 5.2, hydrogen is supplied from a compressed gas cylinder. Unlike the fuel‐cell systems employed in the US space programme that were supplied with pure oxygen, stationary AFC power plant invariably operates with air as the oxidant. This can be supplied with a blower since the pressure drop over a cell is usually low (around 2.0 kPa). To avoid cell degradation, a scrubber is installed in the air line to remove CO2 to a level well below 50 ppm. For small‐scale AFCs, the scrubber can simply be a vessel that contains soda lime, which has to be discarded once it has absorbed its full capacity of CO2. More substantial systems can use a regenerative scrubber that employs an amine‐based material in two parallel reactors that can be alternated between absorption and desorption cycles. One reactor absorbs the CO2 present in the air flowing towards the stack. Meanwhile, in the other reactor, the exhausted amine material, which has been used previously for treatment of the incoming air, is rejuvenated by desorbing the CO2 with the excess air that is exiting the stack. The CO2 is weakly bound to the amine material and can be released by simply raising the temperature of the reactor. 2 As with PEMFCs, it is possible to run the AFC dead ended, but this is not preferred as the anode has to be purged frequently to remove the product water and any contaminants. 137 138 Fuel Cell Systems Explained Air pump Air in Ejector circulator Carbon dioxide removal – H2 H2 A n o d e H2 E l e c t r o l y t e + Electrical power output C a A t i h r o d e Cooling air Hydrogen cooler and water condenser Coolant circulation pump Air out Electrolyte cooler Electrolyte circulation pump Figure 5.2 Diagram of an alkaline fuel cell with circulating electrolyte solution, which also serves as a coolant for the fuel cell. The disadvantages of a circulating electrolyte solution lie in the extra equipment that is required in the form of pumps and a cooler. The pipework necessary to achieve the circulation is prone to leak on account of the low surface tension of the aqueous KOH. It is also a challenge to design a system that will function in any orientation. In a facility that employs multicell stacks, the design must be such that the circulating electrolyte solution does not provide an unwanted current path between cells.3 Systems with circulating electrolyte solutions do have advantages, however, and the principle benefits are as follows: ● ● The circulating electrolyte solution can serve as a cooling system for the fuel cell. The electrolyte solution is continuously stirred and mixed. Reactions (5.1) and (5.2) show that twice as much water is produced at the anode as is consumed at the cathode. Without intervention, this will result in the electrolyte solution becoming too 3 This unwanted ‘internal’ or ‘shunt’ current can be determined by measuring the hydrogen consumption at open circuit. For example, in the cells used by Kordesch (1971), the internal current density was found to be about 1.5 mA cm−2. Alkaline Fuel Cells ● ● ● ● concentrated at the cathode — in fact, so concentrated that it solidifies. Stirring reduces this problem. Excess electrolyte solution can be stored in a vessel external to the stack. This solution can be heated, if necessary, to drive off any additional water that has been absorbed. It is comparatively straightforward to pump out all the electrolyte solution and replace it with a fresh solution. Start‐up and shutdown are both simple — for a cold start‐up, only the reservoir holding the electrolyte solution needs to be heated, rather than the whole stack. The cell can be monopolar, which enables a stack design that is easier to build than one that employs bipolar plates to interconnect the cells. Moreover, there is greater flexibility to configure a stack in terms of the desired voltage and current. Although monopolar designs may be easier to build than bipolar versions, edge collection of current can lead to lower performance as a result of the cumulative resistance, or voltage drop, between the centre and edges of an electrode. The consequence is a lower average current density over the whole electrode surface. This adverse effect becomes more serious as stacks are scaled up. Following the successful application of the AFC in spacecraft, a few companies have pursued the technology for other applications; almost universally the cells have operated with circulating electrolyte solutions. In the United States, Allis‐Chalmers had worked on the technology during the 1960s and Union Carbide continued some work in 1970s. Fuji Electric was the only Japanese company to support AFC technology for an appreciable period. Consequently, it was left mainly to European companies — notably, Siemens and later Elenco — to address the challenges of the AFC in the final years of the century. Elenco located in Belgium was owned by the Belgium company Bekaert and the Dutch State Mines until 1995, when its financing ended. Together with partners in the Netherlands and France, Elenco built AFC systems for articulated city buses under a EUREKA project that was supported by the European Union from 1991 to 1994. The work showcased 40‐kW AFC systems. The rise of interest in the PEMFC in the 1990s, however, signalled the end of Elenco and in a last‐bid effort to rescue the AFC, the company was taken over by Zetek, a UK technology venture. Further interest was generated by Zetek who retrofitted several vehicles, which included a London taxi, with AFC systems that were based on the Zetek Mk2 stack illustrated in Figure 5.3a. Each stack consisted of a series–parallel configuration of 24 individual cells that delivered a power output of 434 W and an output current of 108 A at 4 V. Current densities of up to 120 mA cm−2 were measured at an operating temperature of 70°C. The stack was normally run at 100 mA cm−2 at an average cell voltage of 0.67 V. Rebranded as Zevco, Zetek continued its AFC development until 2001. The Elenco– Zetek–Zevco technology worked extremely well and showed multiple year lifetimes. Unfortunately, the company was caught in the turmoil after the September 11 attacks and the expected financial closure with investors failed to materialize due to the disruption of trading on international markets. In the fall‐out following the collapse of Zetek, new companies were established to take on the intellectual property and move the technology forward. These include AFC Energy (originally Eneco Ltd.) and, more recently, Cygnus Atratus. The AFC Energy systems have a bipolar stack arrangement that has been adapted from the original Zetek monopolar stack design. Low‐cost polymer 139 140 Fuel Cell Systems Explained (a) (b) Fuel-cell stack Figure 5.3 Alkaline fuel cells: (a) Schematic of Zetek Mk2 stack and (b) Intensys–Vito 6‐kW system. (Source: Reproduced with permission of Cygnus Atratus.) frames are used in both the AFC Energy and Cygnus Atratus technologies. In 2008, the Flemish Technical Institute, VITO, in collaboration with Intensys introduced 6‐kW systems that are also based on the Elenco–Zetek technology; an example is shown in Figure 5.3b. 5.2.2 Static Electrolyte Solution An alternative design of AFC, in which each cell in the stack has its own, separate, electrolyte solution that is held in a matrix material between the two porous gas‐diffusion electrodes, is shown schematically (Figure 5.4). The arrangement is clearly less complex than that required for a circulating electrolyte solution and, as will be seen in Chapters 7 and 8, is similar to that used in the PAFC or the molten carbonate fuel cell (MCFC). Furthermore, with a static electrolyte solution, the AFC stack can be used in any orientation, and there is no risk of internal short circuits arising, as is the case with a circulating electrolyte solution. It was these key advantages of simplicity of design and the ability to work in any orientation that led to the static electrolyte solution AFC that was produced by United Technologies Corporation for the Space Shuttle Orbiter (Figure 5.5). The state‐of‐the‐art alkaline fuel‐cell stacks in the Orbiter were rectangular (38 × 114 × 35 cm), weighed 118 kg and produced a peak power of 12 kW at a minimum of 27.5 V (end of life) and an average power of 7 kW. The stacks operated at a similar pressure to the Apollo versions (400 kPa) but at a lower temperature (85–95 vs. 200°C). Unfortunately, the lower operating temperature necessitated the use of Pt catalysts.4 There are, however, some challenges for this type of AFC with respect to the durability and the robustness required for commercial terrestrial applications. As the electrolyte solution matrix can neither be removed nor be completely replaced once a cell has been 4 For the Orbiter fuel cell, gold‐plated nickel electrodes were employed onto which the catalyst was deposited. The catalyst loading on each electrode was 20 mg cm−2 of Au–Pt alloy on the cathode and 10 mg cm−2 Pt on the anode. Alkaline Fuel Cells Electrical power output – + Ejector circulator O2 H2 H2 A n o d e H2 Cooling air E l e c t r o l y t e O2 C a t h o d e Next cell O2 Hydrogen cooler and water condenser Coolant also flows through stack Coolant circulation pump Figure 5.4 Alkaline fuel cell with static electrolyte solution held in a matrix. The system uses pure hydrogen and oxygen, e.g., as employed in spacecraft. Figure 5.5 Alkaline fuel‐cell module used in Space Shuttle Orbiter. 141 142 Fuel Cell Systems Explained assembled, any impurities or carbonates formed within the electrolyte solution will inevitably accumulate. This can drastically reduce the cell performance. The electrolyte solution also cannot be used for cell cooling although this may be achieved via the phase change of water to steam, through the evaporation of water in the anode and/or cathode gas streams. Alternatively a separate cooling system may be employed, as illustrated in Figure 5.4, which was the approach taken for the Apollo and Orbiter spacecrafts. The Apollo AFCs were cooled using a mixture of ethylene glycol and water, as is used in car engines. In the Orbiter systems, the cooling fluid was a fluorinated hydrocarbon dielectric liquid. The system represented by Figure 5.4 uses pure oxygen at the cathode, though this is not obligatory for a matrix‐held electrolyte solution. As in the design that operates with a pumped electrolyte solution (see Figure 5.2), the hydrogen is circulated to remove the product water. In spacecraft systems, the product water is used for drinking, cooking and cabin humidification. Water management is, however, an issue and essentially is similar to that for PEMFCs, though ‘inverted’ in that water is produced at the anode and removed from the cathode. (In the PEMFC, water is produced at the cathode and removed from the anode by electro‐osmotic drag, as explained in Section 4.4, Chapter 4.) The AFC system must be designed so that the water content of the cathode region is kept sufficiently high by diffusion from the anode. In general, the problem of water management is much less severe than with the PEMFC. For a start, the rise in the saturated vapour pressure of KOH solution with temperature is less rapid than that shown by pure water, as will be discussed in Section 5.4. Accordingly, the rate of evaporation is much slower. In the earliest forms of AFCs with static electrolyte solution, the KOH solution was held in a matrix made of asbestos that had excellent porosity, strength and corrosion resistance. Given recognition of the health hazards associated with the deployment of asbestos, alternative materials were developed for spacecraft. For instance, butyl‐ bonded microporous potassium titanate [(K2O)x•(TiO2)z z/x ≈ 8] or K2TinO(2n+1) (n = 4.0–11.0) was used in the space shuttle fuel cell. Ceria and zirconium phosphate have also been proposed, but as yet it appears that no substitute has proved to be universally acceptable for the porous matrix. In addition, for terrestrial service, renewal of the electrolyte solution from the matrix must be possible given that the problem of CO2 contamination is bound to occur. For other applications, the use of AFCs with static electrolyte solution may be overtaken by cells that employ anion‐exchange membranes, as described in Section 5.2.4. 5.2.3 Dissolved Fuel A fuel cell that operates with dissolved fuel is unlikely to be employed for serious power generation but is included here as the design is the simplest to manufacture. The dissolved fuel AFC, in particular, was popular for demonstrating the operating principle of fuel cells and featured in early textbooks, before the widespread availability of small‐scale educational PEMFC systems for schools and colleges. The underlying concept is in Figure 5.6. The KOH electrolyte solution is mixed with a fuel, such as hydrazine, ammonia or sodium borohydride. The fuel anode is along the lines discussed in Section 5.3.4, with a platinum catalyst. The fuel is also fully in contact with the cathode. Whereas this would markedly increase the severity of the ‘fuel Alkaline Fuel Cells Electrical power output – + Waste gasses Air cathode Electrolyte and fuel mixture Fuel anode Figure 5.6 Schematic representation of a dissolved fuel AFC, arguably the simplest of all types, it has a selective catalyst on the cathode that does not react with the fuel. An alternative design has a membrane within the electrolyte solution that isolates the fuel from the air cathode, but adds to cost and complexity. crossover’ problem (discussed in Section 3.5, Chapter 3), it is of no consequence here as the cathode catalyst is not platinum and therefore the rate of reaction of the fuel is very low. Furthermore, there is only one seal that could leak, namely, a very low pressure joint around the cathode. The cell is re‐fuelled simply by adding more fuel to the electrolyte solution. Hydrazine, H2NNH2, is an ideal fuel for this type of cell because it dissociates into hydrogen and nitrogen at the anode; the hydrogen that is formed reacts according to equation (5.1). Sodium borohydride (NaBH4) can also be used as a fuel. In Chapter 11, this compound will be considered as a material for hydrogen storage. As a fuel, it can be dissolved in the AFC electrolyte solution, and it reacts at the anode according to: NaBH 4 8OH NaBO2 (5.4) 6H2 O 8e The impressive fact to note is that eight electrons are formed by this reaction for just one molecule of fuel. Even more interesting is the large change in Gibbs free energy (expressed as ∆ g f kJ mol−1, see Section 2.1, Chapter 2) and therefore the high reversible voltage (Vr) of the cell. The reaction of air at the cathode reaction is exactly the same as for the hydrogen fuel cell, i.e., equation (5.2). The overall reaction is thus: (5.5) NaBH 4 2O2 NaBO2 2H2 O For this reaction: 920.7 Gf 2 237.2 123.9 1271.2 kJ mol 1 (5.6) Therefore, from equation (2.9), Chapter 2: Vr Vr g f (2.9) zF gf zF 1271.2 103 8 96 485 1.64 V (5.7) 143 144 Fuel Cell Systems Explained This theoretical voltage is significantly higher than that obtained with hydrogen, and at eight electrons per molecule, it indicates a fuel of remarkable potency. Unfortunately, the voltage actually obtained with a borohydride fuel cell is not so different from that with a hydrogen cell, because the catalysts that facilitate the direct borohydride oxidation, reaction (5.4), also promote the following hydrolysis reaction: NaBH 4 2H 2 O NaBO2 4 H2 (5.8) This reaction was the main reason for the abandonment of the technology in the 1960s. The electrodes available at the time were not able to utilize the hydrogen effectively, so the loss of hydrogen via reaction (5.8) made the borohydride cell inefficient. This is not the case with modern electrodes, which, even with low platinum loadings, will promote direct hydrogen oxidation. Furthermore, if the concentration of borohydride in the electrolyte solution is low, the rate of reaction (5.8) is reduced significantly with the net effect of an improvement in cell voltage. As a fuel, sodium borohydride is an expensive but convenient means of providing hydrogen. Further discussion of borohydride fuel cells is given in Section 6.5, Chapter 6. 5.2.4 Anion‐Exchange Membrane Fuel Cells In contrast to the PEMFC, the AFC exhibits facile kinetics for both the anode and cathode reactions. Consequently, cheaper non‐noble metal catalysts can be used in the electrodes. As noted in the previous sections, however, the AFC has a significant drawback in that degradation of both the electrolyte solution and the electrodes can occur through the formation of carbonate/bicarbonate (CO32−/HCO3−) via reaction between OH− ions and CO2 in the oxidant gas stream. A more recent variant of the AFC is the anion‐exchange membrane fuel cell (AMFC), in which the KOH electrolyte solution is replaced by a solid alkaline‐electrolyte membrane (AEM).5 The AEM is a polymer material, and the fuel cell is effectively an alkaline analogue of the PEMFC. The AMFC thus retains the electrocatalytic advantages of the AFC, but introduces a CO2‐tolerant electrolyte. In general, an AEM is composed of a polymer backbone on which cationic sites are tethered. These cationic moieties are not carbonate ions with free mobility as in a liquid electrolyte. Thus carbonate precipitates cannot form in the AMFC. Transport of OH− ions within the AEM is between the cationic sites in an analogous manner to the H+ ions that are transported between sulfonic acid sites in the PEMFC membrane. The AMFC shares the advantage of the PEMFC of being a solid‐state device (there is no liquid electrolyte to leak) and is built up using catalyst and gas diffusion layers (GDLs) in much the same way as a PEMFC. Furthermore, corrosion of the bipolar plate is less of an issue and therefore permits the use of thin and easily manufactured hardware. 5 Several terms and acronyms have been proposed for this type of fuel cell. Anion‐exchange membrane (AEM) is consistent with the use of PEM in the case of proton‐exchange membrane fuel cells and will be used throughout this book. The reader will also find references in the literature to alkaline‐electrolyte membrane fuel cell (AEMFC), hydroxide‐exchange polymer membrane fuel cell (HEMFC) and alkaline proton‐exchange membrane fuel cell (APEMFC). Alkaline Fuel Cells For many years, polymeric anion‐exchange membranes have been employed in seawater desalination plants, the recovery of metal ions from waste waters, electrodialysis and bio‐separation processes. Unfortunately, however, most of these membranes possess ionic conductivities that are too low to be considered for AMFC application. Also, most AEM polymers have poor solubility in the solvents employed in the production of NafionTM, the membrane that is used in most PEMFCs (see Section 4.2.1, Chapter 4). The low solubility complicates the fabrication of an AMFC in that, unlike Nafion, it is more difficult to incorporate an anion‐exchange polymer as a binder in the electrode layers. An example of an AEM is that formed by functionalization of a polysulfone via chloromethylation, followed by reaction with an amine (quaternization) or phosphine to yield a quaternary ammonium (QA) or phosphonium salt. The salt form of the membrane can then be treated with KOH to yield a hydroxide‐ion‐conducting AEM in much the same way that the sodium form of a PEM membrane (e.g., Nafion) can be treated with sulfuric acid to yield a proton‐conducting membrane. The synthesis reactions involved in the production of a commercial polysulfone (Udel® from Solvay Advanced Polymers LLC) are summarized in Figure 5.7. Membranes prepared by QA chemistry have been the most studied for fuel‐cell applications, and they have reasonable stability in alkaline environments (especially CH3 O O S CH3 Chloromethylation O CH3 O S CH3 O n O CH2Cl (CH3)3N Quaternization O CH3 O S CH3 –Cl+(H C) NH C 3 3 2 O n O CH2N(CH3)3+Cl– 1M KOH Ion exchange O CH3 O –OH+(H C) NH C 3 3 2 n O (CH3)3SiCl + (CH2O)n + SnCl4 CIH2C O CH3 CH2N(CH3)3+OH– S O O n Figure 5.7 Chemical reaction steps to convert a polysulfone into an anion‐exchange membrane polymer. 145 146 Fuel Cell Systems Explained membranes that contain benzyltrimethyl ammonium exchange sites). The general issues with such AEMs are as follows6: ● ● ● The diffusion coefficient and mobilities of OH− anions are typically one‐third to one‐ half less than those of H+ in most media, and QA ionic groups are less dissociated than the typical sulfonic acid groups. Thus, there were concerns that AMEs would not possess intrinsic ionic conductivities high enough for application in fuel cells. The OH− ions are effective nucleophiles that potentially cause degradation of the polymer via (i) a direct nucleophilic displacement and/or (ii) a Hofman elimination reaction when a β‐hydrogen is present and possibly (iii) a mechanism that involves an ylide intermediate.7 The AEM must have the chemical stability to withstand the final step in the preparation, i.e., typically, the exchange of chloride (Cl−) ions with OH− ions in a strongly alkaline solution of NaOH or KOH. All of the polymer degradation mechanisms are enhanced at high temperatures, and therefore most AMFC developers are targeting operation at room temperature. Various starting materials are under investigation for synthesizing anion‐conducting polymers. Examples are polybenzimidazole (PBI), poly‐ether ketones, polyphenylene oxides and polyvinyl alcohol grafted with 2,3‐epoxypropyltrimethylammonium chloride. A range of quaternizing agents are also being evaluated. Synthetic routes to AEMs other than by quaternization are under investigation, and, as with PEMFC membranes, there are several well‐practiced methodologies for the preparation of membranes.8 The polymer can be synthesized directly from a functionalized monomer, polymerized from a monomer with subsequent functionalization or prepared by functionalizing a commercial polymer. A body of literature is emerging from which some general remarks concerning AEMs can be made as follows. Fluorine‐containing polymers generally show higher thermal stabilities than hydrocarbon polymers. Irradiation of polymer films using X‐rays, γ‐rays or electron beams is a flexible way to introduce functional groups, but the easiest synthesis route is to dope inert polymers directly with concentrated KOH solution. For instance, polar polymers (e.g., polyethylene oxide) can be doped with alkali hydroxides (e.g., KOH), or ammonium hydroxides such as tetrabutyl ammonium hydroxide. Polybenzimidazole doped with KOH shows a very high ionic conductivity compared with proton conductivity in Nafion but could suffer from carbonate precipitation as experienced in conventional AFC electrolyte solutions. The development of AMFCs is in its infancy. Single‐cell AMFCs have been built and tested in the laboratory, but to date no stack demonstrations at the kW scale have been built. 6 Slade, RCT, Kizewski JP, Poynton, SD and Varcoe JR, 2013, Alkaline membrane fuel cells, in Meyers, RA (ed.), Encyclopedia of Sustainability Science and Technology, Springer Science + Business Media, New York. 7 Chempath, S, Einsla, BR, Pratt, LR, Macomber, CS, Boncella, JM, Rau, JA and Pivovar, BS, 2008, Mechanism of tetra‐alkyl ammonium head group degradation in alkaline fuel cell membranes, Journal of Physical Chemistry C vol. 1123, pp. 3179–3182. 8 Couture, G, Alaaddine, B, Boscheti, F and Amedur, B, 2011, Polymeric materials as anion‐exchange membranes for alkaline fuel cells, Progress in Polymer Science, vol. 36, pp. 1521–1557. Alkaline Fuel Cells 5.3 Electrodes As discussed earlier, although AFCs can be operated over a wide range of temperatures and pressures, the extent of their applications is quite restricted. Accordingly, there is no standard type of electrode for the AFC, and different approaches are taken as determined by performance requirements, operating temperature and pressure and cost limits. Different catalysts can be used, but this does not necessarily affect the electrode structure. For example, a platinum catalyst is effective with any of the main electrode structures that are described here. 5.3.1 Sintered Nickel Powder When F. T. Bacon designed his pioneering fuel cells in the 1940s and 1950s, he opted for nickel‐based electrodes in the belief that the expensive platinum‐group electrocatalysts would never become commercially viable. His electrodes were made porous through fabrication from powdered nickel, which was then sintered to make a rigid structure. To enable a good three‐phase contact between the reactant gas, the electrolyte solution and the solid electrode, the nickel electrode was made in two layers from two sizes of nickel powder. The procedure gave a wetted, fine‐pore structure for the liquid side and more open pores for the gas side. Very good results were achieved, though careful control of the differential pressure between the gas and the electrolyte solution was necessary to ensure that the liquid gas boundary was anchored to the electrode (note: wet‐proofing materials, such as polytetrafluoroethylene (PTFE), were not available at that time). This electrode structure was also selected for the fuel cells employed in Apollo missions. In both the Bacon and Apollo cells, the anodes employed plain nickel powders, whereas the nickel oxide cathodes were treated with lithium salts to generate LiNO2 on the surface to provide chemical stability. 5.3.2 Raney Metals An alternative method for obtaining a very active and porous form of a metal is the use of Raney metals; it has been a common practice for AFCs from the 1960s to the present. The metals are prepared by mixing the required active metal (e.g., nickel) with an inactive metal, usually aluminium. The mixing is performed in such a way that distinct regions of aluminium and the host metal are maintained, i.e., the material is not a true alloy. The mixture is then treated with a strong alkali that dissolves out the aluminium to leave a porous product with a very high surface area. The process gives scope for changing the pore size by altering the proportions of the two metals and by adding small amounts of other metals such as chromium, molybdenum or zinc. Raney nickel electrodes were employed in many of the demonstrations of fuel cells that were reviewed in the opening (Section 5.1). Often Raney nickel was chosen for the anode and silver for the cathode. In the early 1990s, this combination of electrodes was a feature of the AFC built by Siemens for service in submarines. Raney metals have also been used as catalyst in a ground‐up form, for the rolled electrodes that are described in the following section. 147 148 Fuel Cell Systems Explained 5.3.3 Rolled Carbon Most modern AFCs employ carbon electrodes that are similar to those used in the PEMFC. In the late 1950s, Karl Kordesch carried out the initial development of carbon electrodes while he was employed by the Union Carbide Corporation (UCC). The first UCC electrodes were built up of several layers of carbon black, PTFE and pore‐forming additives. Catalyst metals included not only nickel but also silver and cobalt. Cell voltages of around 0.6 V were achieved in air at current densities up to 200 mA cm2. The work of Kordesch culminated in the demonstration of a hydrazine‐fuelled AFC that powered a converted Austin A40 van. The vehicle is now in the London Science Museum. The latest carbon electrodes now invariably employ carbon‐supported catalyst metals mixed with PTFE, which are then rolled onto a material such as nickel mesh. The PTFE acts as a binder, and its hydrophobic properties prevent flooding of the electrode and provide for controlled permeation of the electrode by the electrolyte solution. A thin layer of PTFE will often be placed over the surface of the electrode for two reasons: (i) to control further the porosity and (ii) to impede the electrolyte solution from passing through the electrode, without the need to pressurize the reactant gases, a requirement that is necessary with porous metal electrodes. Carbon fibre is sometimes added to the mix to increase the strength, conductivity and porosity of the resulting electrode. Modified papermaking machines can be used to manufacture rolled electrodes at quite low cost. Such electrodes find application not only in fuel cells but also in metal– air batteries, for which the cathode reaction is much the same as for an alkali fuel cell. For example, the same electrode can act as the cathode in a zinc–air battery (e.g., for hearing aids) and an aluminium–air battery (e.g., to provide reserve power for telecommunications). Such an electrode is shown in Figure 5.8. The carbon‐supported catalyst is of the same structure as that presented in idealized form in Figure 4.11, Chapter 4. The catalyst may not always be platinum. Manganese, for example, is an effective Figure 5.8 The structure of a rolled AFC electrode. The catalyst is mixed with a PTFE binder and rolled onto nickel mesh. The thin layer of PTFE on the gas side is shown partially rolled back. Alkaline Fuel Cells cathode catalyst in both metal–air batteries and AFCs. Commercial rolled electrodes with a non‐platinum catalyst are readily available at about US$0.01 per cm2 or around US$10 per ft2, i.e., at a cost that is very low compared with other fuel‐cell materials. Adding a platinum catalyst increases the cost in line with the loading, but it might only be by a factor of about three, which, with respect to fuel cells, still gives a very inexpensive electrode. There are, however, problems elsewhere. One issue is that because the electrode is covered with a layer of PTFE, the surface is non‐conductive, and thus a bipolar plate cannot be employed for cell interconnection. Instead, the cells are normally edge connected. Fortunately, this is not too much of a constraint given that the nickel mesh running right through the electrode results in a higher than normal conductivity across the plane of the electrode and thereby renders edge connection to be a practical option. Edge connection gives a certain flexibility to stack design in that it is not necessary to connect the positive of one cell to the negative of the adjacent cell, as must occur with bipolar plates. Instead, series–parallel electrical connections can be made and inherently improve the performance of the cells by reducing internal current losses. The problem of internal shunt currents within an AFC stack is a unique feature of using a liquid electrolyte that is circulated throughout the stack. The ion‐conducting electrolyte is in contact with all cells within the stack and can therefore provide an ionic current pathway between adjacent cells. The path is short if the cells are configured electrically in series, using conventional bipolar plates between each cell. Note that this problem does not exist in an MCFC because the current collectors and flow‐field plates separate the electrolyte of each cell. In the AFC, however, there is no such separation between cells. By electrically connecting AFCs in parallel, the current path between cells is elongated so that any loss of voltage caused by movement of electrolyte between cells is minimized. In practice AFCs may be joined together with a mixture of series and parallel connections to minimize the losses due to electrolyte circulating between cells. Apart from the serious problem that crystals of carbonate can form in the pores of the electrodes from CO2 in the fuel or oxidant gases, there has also been a suggestion that some carbon dissolution can occur in the AFC catalysts, as happens with PEMFC carbon cathodes. Extensive studies,9 indicated that the operational life of air electrodes (PTFE‐bonded carbon electrodes on porous nickel substrates) with CO2‐containing air at 65°C ranged from 1600 to 3400 h at a current density of 65 mA cm−2, compared with 4000–5500 h when using CO2‐free air under similar conditions. The current density was not particularly high in these tests, and lifetime was less at higher currents. It was also found that lower temperatures shorten life, presumably due to a decrease in the solubility of the carbonate. Note that a lifetime of 3400 h is only 142 days and implies that such electrodes are only suitable for a limited number of applications. Gulzow10 describes an anode based on granules of Raney nickel mixed with PTFE that is rolled onto a metal net in much the same way as the PTFE/carbon‐supported catalyst. A cathode was prepared likewise, only using silver instead of nickel. It was claimed that such electrodes are not degraded by CO2. 9 Kordesch, K, Gsellmann, J and Kraetschmer, B, 1983, Studies of the performance and life‐limiting processes in alkaline fuel cell electrodes, Power Sources, vol. 9, p. 379, ed. By Thompson, J, Academic Press, New York. 10 Gulzow, E, 1996 Alkaline fuel cells: a critical view, Journal of Power Sources, vol. 61, pp. 99–104. 149 150 Fuel Cell Systems Explained 5.3.4 Catalysts The AFC stacks in the Orbiter had very high loadings of noble metal catalyst in the electrodes: a 80 wt.% Pt + 20 wt.% Pd anode catalyst was loaded at 10 mg cm−2 and a 90 wt.% Au + 10 wt.% Pt cathode catalyst was loaded at 20 mg cm−2, each on a silver‐ plated nickel screen. Both catalysts were bonded with PTFE to achieve high performance at 85–95°C. The aggressive nature of the alkaline electrolyte is much less than that of acid in PAFCs or PEMFCs and thereby enables the selection of a broader range of catalysts. Very high surface area (Raney) nickel can be used at the cathode instead of platinum. The nickel can, in turn, be enhanced by a catalyst that consists of high surface area active carbon doped with silver, and iron (or cobalt) macrocyclics such as heat‐treated cobalt tetra‐phenoxymethyl porphyrins on carbon. As with PEMFC cathode catalysts, the selection of porphyrins for the oxygen reduction reaction (ORR) has been stimulated by knowledge of compounds that are involved in the reduction of oxygen in biological systems. By raising the temperature, most developers of stationary AFC systems since the 1960s have opted for ‘classic’ non‐noble metals for the catalysts (Raney nickel for the anode, silver and/or manganese dioxide for the cathode). With a nickel anode catalyst, activation overpotential is dominant at low current densities, whereas transport processes significantly increase the overpotential at very high current densities. Therefore, as with the PAFC, it is essential for the AFC to operate within these limits. Unlike the platinum anode catalyst in the PEMFC, the nickel catalyst in the AFC or PAFC can undergo permanent oxidation if the current density is allowed to go too high. To avoid such issues, researchers have been exploring other ion‐conducting materials as catalysts, specifically spinels and perovskites that are able to tolerate cycling between oxidizing and reducing conditions. The cathode catalyst of the AFC is of particular interest since the overpotential at this electrode contributes most to the voltage loss in the cell. Silver has the highest electrical conductivity of any element and is approximately fifty times less expensive than platinum. Moreover, silver is one of the most active catalysts for the ORR — the metal is competitive to platinum in highly concentrated alkaline media, as well as on a cost/performance basis. Incidentally, it is interesting to note that cathodes loaded with silver have also given a longer lifetime (3 years) in alkaline electrolyzers than platinum‐based cathodes (1 year) under practical chlor‐alkali electrolysis conditions. The impregnation of silver into a carbon support via the in situ reduction of silver nitrate (AgNO3) has been shown to produce very fine particles that constitute a high surface area catalyst for optimum cathode performance. Research has shown11 that carbonate deactivation can be avoided by replacing the porous carbon support with porous silver of the form used in commercial water purification membranes. The catalytic activity of the silver electrode for the ORR can be improved by including a platinum or manganese dioxide (MnO2) catalyst. Gas accessibility through the silver is also enhanced by impregnating the pores near the gas surface with Teflon AF (a microporous form of PTFE). 11 Bidault, F, Kucernak, A, 2011, Cathode development for alkaline fuel cells based on a porous silver membrane, Journal of Power Sources, vol. 196(11), pp. 4950–4956. Alkaline Fuel Cells 5.4 Stack Designs 5.4.1 Monopolar and Bipolar In a monopolar stack, each cell is connected in series with the next by using an electrically conducing meal strip or wire — the anode of one cell is linked to the cathode of the next. The arrangement is illustrated in Figure 5.9a and is necessary in the case of recirculating electrolyte solution cells where the electrodes are coated with non‐conductive PTFE. (a) H2 O2 H2 O2 H2 O2 + – (b) Seal Membrane–electrode assembly (MEA) Bipolar plate + – H2 O2 H2 O2 H2 O2 H2 O2 Figure 5.9 AFC stack configurations: (a) monopolar and (b) bipolar. 151 152 Fuel Cell Systems Explained Bipolar AFC stacks are similar in configuration to most PEMFC stacks, as shown in Figure 5.9b. The bipolar arrangement is more suitable for fuel cells with static electrolyte solutions where there is no prospect of short circuits occurring because the electrolyte is held in a matrix material that separates the two electrodes. The downside of this design is that the matrix has to be relatively thick, which increases the ohmic loss compared with the relatively thin liquid electrolyte film that can be employed in a cell with recirculating electrolyte solution. If successful AEMs can be produced, the bipolar design will be more attractive again, in that it may be possible to use thinner electrolyte films. 5.4.2 Other Stack Designs A variation on the more conventional design of static electrolyte solution cell has been produced by Hydrocell, a Finnish company. The technology uses a gel electrolyte and has cylindrical geometry, which requires the cells to be connected externally either electrically in series and/or in parallel according to the required voltage, rather than in a bipolar arrangement. Another type of stack is the falling‐film fuel cell developed by Hoechst AG in Germany. The configuration is the same as that of the cell with the recirculating electrolyte solution shown in Figure 5.2, except that the flow of liquid through the cell is entirely gravity driven. Therefore, the typical height‐dependent hydrostatic pressure of a column of liquid does not develop, and consequently, the hydrostatic pressure is the same at the inlet and outlet of the cell. The great advantage of the falling‐film fuel cell is that the pressure difference between the electrolyte on the front side of the electrode and the gas on the rear remains constant over the whole area of the vertical electrode and hence is uniform throughout the cell. The absence of a pressure driving force leads to a stable three‐phase boundary of electrolyte within the GDLs and thereby minimizes any potential loss of electrolyte solution, with the result that the gap between the two electrodes can be made very narrow, typically about 0.5 mm. Cells as large as 0.25 × 1 m have been constructed, which, on account of the thin electrolyte layer, exhibit very large current densities of up to 2.5 A cm−2. 5.5 Operating Pressure and Temperature Historically, most AFCs have operated well above ambient pressure and temperature. These two parameters, together with information about the electrode catalyst, are given for a selection of important types of AFC in Table 5.1. The choice of operating pressure is dependent on the system design. In general, cells that employ recirculating electrolyte solution and the falling‐film cell operate at near‐ambient pressure. For spacecraft fuel cells that used static electrolyte solutions, higher pressures were more common, namely, from 300 kPa to more than 1 MPa. The conductivity of the OH− ions is dependent on the temperature and concentration of the electrolyte solution. Conductivity increases with temperature. The traditional AFC can be started below 0°C since the freezing point of the electrolyte solution at typical concentrations of around 30 wt.% is well below that of water. In fact, the concentration required to achieve the maximum ionic conductivity only increases from about 30 wt.% at 0°C to 34 wt.% at 80°C. Similarly, the boiling point of the electrolyte Alkaline Fuel Cells Table 5.1 Operating parameters for certain AFCs. The pressure data are approximate as there are usually small differences between each reactant gas. Fuel cell Pressure (kPa) Temperature (°C) KOH (wt.%) Anode catalyst Cathode catalyst Bacon 500 200 30 Ni NiO Apollo 350 230 75 Ni NiO Orbiter 410 93 35 Pt–Pd Au–Pt Siemens 220 80 n/a Ni Ag Data from: Warshay, M and Prokopius, PR, 1990, The fuel cell in space: Yesterday today and tomorrow, Journal of Power Sources, vol. 29, pp. 193–200, and Strasser, K, 1990, The design of alkaline fuel cells, Journal of Power Sources, vol. 29, pp.149–166. solution is elevated and thus enables the cell to operate up to 230°C if the electrolyte concentration is increased to 85 wt.%. The advantages of higher pressure have been considered in Chapter 2 where it was shown, in Section 2.5.4, that the open‐circuit voltage, Vr, of a fuel cell is raised when the pressure increases from P1 to P2 according to the relationship: V RT P ln 2 4F P1 (5.9) The demonstration cell of F. T. Bacon operated around 500 kPa and 200°C. Even these high pressures, however, would only raise the voltage by about 0.04 V if this ‘Nernstian’ effect was the sole benefit. A rise in pressure (and/or temperature) also increases the exchange‐current density and thereby reduces the activation overpotential at the cathode (see Section 3.4, Chapter 3). Consequently, the benefit of increased pressure is much more than equation (5.9) would predict. For example, the very high pressure gave the Bacon cell a performance that even today would be considered to be remarkable, namely, 400 mA cm−2 at 0.85 V or 1 A cm−2 at 0.8 V. The choices of operating pressure, KOH concentration and catalyst are interrelated. A good example is the transition from the Bacon cell to the system developed for the Apollo spacecraft. Although the Bacon cell gave an impressive performance, it was a heavily engineered design that operated at a very high pressure. To reduce mass for space applications, the pressure had to be lowered. Consequently, the temperature had to be raised to maintain the performance at an acceptable level. It was then necessary to increase the concentration to 75 wt.% KOH; otherwise the electrolyte solution would have boiled. Unfortunately, increasing the concentration considerably lowers the vapour pressure, as can be seen from Figure 5.10. At ambient temperature, a 75 wt.% KOH solution is solid, and therefore it was necessary to provide heaters to start the fuel cell. In the Orbiter system, the concentration was reduced back to 32 wt.%, and the temperature was set at 93°C. In many applications of AFCs, the reactant gases are contained in pressurized or cryogenic storage vessels. In such cases, it is necessary to reduce the supply pressure of each gas to match the operating conditions of the stack. This requires accurate control to avoid a large differential pressure between anode and cathode compartments. When pressurized gases are supplied, there also is a risk that leaks may develop. Apart from 153 154 Fuel Cell Systems Explained 4.0 Water Pressure/MPa 40 wt.% KOH solution 3.0 2.0 1.0 80 wt.% KOH solution 50 200 250 300 Temperature/°C Figure 5.10 Change in vapour pressure with temperature for different concentrations of KOH solution. the waste of gas, leakage could lead to the build‐up of explosive mixtures of hydrogen and oxygen, especially when the fuel cell is for use in confined spaces such as submarines. One solution to this problem is to provide an outer envelope for the fuel‐cell stack that is filled with nitrogen at a higher pressure than that of each of the reactant gases. In a Siemens submarine system, for example, the hydrogen was supplied at 0.23 MPa and the oxygen at 0.21 MPa, with the surrounding nitrogen gas at 0.27 MPa. Any leak would result in a flow of nitrogen into the cells that would reduce the performance but would prevent an outflow of reactant gas. In AFCs, there is often a difference in the pressure of the reactant gases and/or in the vapour pressure of the electrolyte solution. For instance, the hydrogen pressure in the aforementioned Siemens AFC was slightly higher than that of the oxygen. In the Orbiter fuel cell, the hydrogen gas was kept at 35 kPa below the oxygen pressure. By contrast, the gases in the Apollo system were at the same pressure, but both were about 70 kPa above the vapour pressure of the electrolyte solution. There are no rules governing the setting of reactant pressure — small differences will be required for a variety of reasons, e.g., to maintain the boundary of the electrolyte solution and gas in the GDLs. Raising the temperature actually reduces the open‐circuit voltage of a fuel cell, as explained in Section 2.3, Chapter 2. In practice, however, the magnitude of this effect is far exceeded by the reduction in the activation overpotential, especially at the cathode. As a result, increasing the temperature increases the voltage of an AFC. From a wide survey of results, it has been concluded12 that below about 60°C there is a very large benefit to raising the temperature, namely, as much as 4 mV per °C for each cell. At this rate, increasing the temperature from 30 to 60°C would lift the cell voltage by about 12 Hirschenhofer, JH, Stauffer, DB and Engleman, RR, 1995, Fuel Cells: A Handbook, revision 3, pp. 6‐10 to 6‐15, Business/Technology Books, Orinda, CA. Alkaline Fuel Cells 0.12 V — a major improvement in the context of fuel cells that operate at about 0.6 V per cell. There is still a noticeable advantage at higher temperatures, but only in the region of 0.5 mV per °C. It would appear, therefore, that about 60°C would be a minimum operating temperature for an AFC. At higher values, the choice would depend strongly on the power of the cell (and thus any heat losses), the pressure and the effect of the concentration of the electrolyte solution on the rate of evaporation of water. 5.6 Opportunities and Challenges The AFC is one of the most efficient energy conversion devices, employing a low‐cost electrolyte and potentially inexpensive electrodes, and is capable of operating at near‐ ambient temperature and pressure. It could be concluded, therefore, that such attributes would make the technology attractive for many applications. Unfortunately, successful exploitation of the AFC for terrestrial applications has been blocked by the incompatibility between the alkaline electrolyte and CO2. Work carried out in recent years by AFC Energy and others has shown that modern gas‐diffusion electrodes have somewhat better tolerance to CO2 than earlier porous metal electrodes. Nevertheless, challenges in separating product water and in mechanically circulating the electrolyte solution or constraining it within a matrix have continued to hinder the development of the AFC compared with the PEMFC, which does not suffer from such fundamental technical issues. If robust anionic membranes with high OH− conductivity can be produced easily (i.e., at low cost), then perhaps the AMFC can compete with the PEMFC in applications such as fuel‐cell vehicles. The ability to run at higher temperatures may also see a new type of AFC competing with PAFCs for stationary power generation. For an AFC to perform reliably over a long period, it is essential to remove the CO2 from the air. Although this is possible, using processes that are practiced industrially (e.g., the Benfield process or absorption in aqueous alkanolamine solution) such procedures would substantially increase the cost, complexity, mass and size of the system. Ahuja and Green have proposed a novel method13 that would only be feasible when hydrogen is stored as a liquid. Their method takes advantage of the fact that heat exchangers are needed to warm the hydrogen and cool the fuel cell. The system is designed in such a way that the incoming air is cooled in a heat exchanger by the liquid hydrogen as it vapourizes, thereby freezing out CO2 from the air, which can be separated. The cold air can then be used to cool the cell, and in doing so its temperature is raised to that required at the cathode inlet. Alternative methods that have received serious consideration for CO2 removal have included the utilization of zeolite separation membranes. Another possibility for AFC application, and which is actually what Bacon had in mind when developing his AFC designs in the mid‐20th century, is to incorporate the cells into a regenerative system. Electricity from renewable sources is used to electrolyze water, and the fuel cell turns the hydrogen and oxygen so produced back into electricity 13 Ahuja, V & Green, R 1988, Carbon dioxide removal from air for alkaline fuel cells operating with liquid hydrogen – a synergistic advantage, International Journal of Hydrogen Energy, vol. 23(20), pp. 131–137. 155 156 Fuel Cell Systems Explained as needed. Of course, other types of fuel cell could be employed in such a system, but here the disadvantages of the AFC would be largely removed, since both reagents would be free of CO2. Further Reading Arges, CG, Ramani, V and Pintauro, PN, 2010, Anion exchange membrane fuel cells, The Electrochemical Society Interface, vol. 19, pp. 31–35. Kordesch, KV, 1971, Hydrogen‐air/lead battery hybrid system for vehicle propulsion, Journal of the Electrochemical Society, vol.118(5), pp. 812–817. Kordesch, KV and Cifrain, M, 2004, Advances, aging mechanism and lifetime in AFCs with circulating electrolytes, Journal of Power Sources, vol. 127, pp. 234–242. McLean, GF, Niet, T, Prince‐Richard, S and Djilali, N, 2002, An assessment of alkaline fuel cell technology, International Journal of Hydrogen Energy, vol. 27(5), pp. 507–526. Mulder, G, 2009, Fuel cells –alkaline fuel cells, in Garche, J, Dyer, CK, Moseley, PT, Ogumi, Z, Rand, DAJ and Scrosati, B (eds.), Encyclopedia of Electrochemical Power Sources, pp. 321–328. Elsevier, Amsterdam. 157 6 Direct Liquid Fuel Cells A direct liquid fuel cell (DLFC) generates electricity via the oxidation of a liquid fuel that requires no preliminary preparation. Most DLFCs use a proton‐exchange membrane (PEM) as the electrolyte and are therefore closely related to the proton‐exchange membrane fuel cell (PEMFC). The direct methanol fuel cell (DMFC) is the most mature version of this technology and is therefore described in the opening section of this chapter. The cell is available commercially for some low‐power applications; for example, over 35 000 battery chargers employing DMFC systems have been produced by SFG Energy AG, under the trademark of Energy for You (EFOY). The remainder of the chapter is devoted to types of low‐temperature fuel cells that run on alternative fuels that are liquids under normal conditions. Potential candidates include many alcohols (e.g., ethanol, propanol, propan‐2‐ol) and other organic liquids (e.g., ethylene glycol, acetaldehyde, formic acid). Some characteristics of these fuels are given in Table 6.1. Sodium borohydride (as a solution), which has already been discussed in Chapter 5, will be also considered as an inorganic fuel for a DLFC. 6.1 Direct Methanol Fuel Cells Methanol (CH3OH) is a simple alcohol that is liquid at normal temperatures and pressures (boiling point 64.7°C) and is miscible with water. It is readily available but has a specific energy (Wh kg−1) that is only half that of gasoline. Nonetheless, in the early 1990s, methanol was proposed for fuel‐cell vehicles given that it is relatively easy to reform directly into hydrogen. For instance, Daimler built a demonstration car — the Necar 3 — that employed an on‐board methanol reformer to generate hydrogen that fed a PEMFC. If methanol can serve as a fuel, then all the problems associated with storing hydrogen in a vehicle are swept aside. As has been mentioned already at various points in Chapters 2 and 3, methanol can, in principle, also be used directly in fuel cells. The DMFC, in which the methanol is oxidized directly at the anode, has the advantage of not requiring a fuel processor to convert the methanol to hydrogen. Consequently, the DMFC could potentially be attractive for small portable systems where weight can be an issue. The DMFC was pioneered by Shell Research Limited in England and Exxon‐Alsthom in France during the 1960s and 1970s. Shell chose a sulfuric acid electrolyte, while Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 158 Fuel Cell Systems Explained Table 6.1 Thermodynamic characteristics of PEMFC and some DLFCs at 25°C and 101.325 kPa. Fuel cell Fuel Weight (g mol−1) Number of electrons involved Standard Theoretical energy cell voltage density (V) (Wh mL−1) Maximum efficiency (%) PEMFC Hydrogen 2.01 2 1.23 1.55 (at 70 MPa) 83 DMFC Methanol 32.04 6 1.21 4.33 97 DEFC Ethanol 46.07 12 1.15 5.80 97 DEGFC Ethylene glycol 62.07 10 1.15 5.85 99 DFAFC Formic acid 46.03 2 1.41 1.88 106 DPFC Propanol DPFC(2) Propan‐2‐ol 60.1 18 1.13 7.35 97 60.1 18 1.12 7.10 97 DEFC, direct ethanol fuel cell; DEGFC, direct ethylene glycol fuel cell; DFAFC, direct formic acid fuel cell; DMFC, direct methanol fuel cell; DPFC, direct propanol fuel cell and DPFC(2), direct propan‐2‐old fuel cell. Exxon‐Alsthom pursued an alkaline approach. Despite some good work with alkaline and buffer electrolyte technology, Exxon terminated its research programme in the late 1970s. The Shell research continued until the early 1980s when, as a result of the curtailed growth in oil consumption that resulted from the conservation measures taken after the 1973 crisis, it became clear that the fears of an imminent oil shortage were unfounded. The drop in oil prices had pushed the target cost for the DMFC out of reach. Nevertheless, substantial progress had been made by teams at the Shell Thornton Research Centre in Chester, United Kingdom, and at the Koninklijke Shell Laboratorium in Amsterdam, the Netherlands. During the period 1973−1981, the UK effort led to an improvement in the performance of the fuel electrode by over two orders of magnitude, and a more detailed understanding of the mechanism of the methanol oxidation reaction emerged. This outcome is discussed in Section 6.2.2. At the same time, the Amsterdam laboratory made considerable progress in the development of stable, active, non‐noble metal catalysts for the air electrode. Some of this work will be addressed later. The DMFC attracted little attention during the 1980s until the PEMFC emerged as viable technology at the end of the decade. At this point, several university research groups, particularly in the United States, started to conduct investigations of DMFCs based on PEMs. The work carried out by these groups provided the basis for the evolution of current DMFC technology. The net specific energy of methanol is higher than other means of storing hydrogen, particularly as compressed gas or as metal hydride, as indicated in Table 6.2. In general, liquid fuels have much higher specific energies than gases, and this is an important factor for a fuel‐cell system destined for transport applications if a long driving range is to be achieved on a single tank of fuel. Other advantages of the DMFC are ease of handling of methanol, rapid refuelling and simplicity of design. A negative aspect of the DMFC is that the oxidation of methanol at the anode is a much slower reaction than the oxidation of hydrogen at the anode of a PEMFC, as explained in Section 6.2. Consequently, a DMFC has reduced power output compared with a PEMFC of similar size and using the same membrane-electrode assembly (MEA). Direct Liquid Fuel Cells Table 6.2 Comparison of specific energy (LHV) for common energy storage materials and the most important hydrogen storage technologies. Storage method H2 at 30 MPa in composite cylinders H2 in metal hydride cylinders Specific energy of fuel 119.9 MJ kg−1 Storage efficiencya (%) 0.6 33.3 kWh kg−1 119.9 MJ kg −1 H2 from methanol — ‘indirect methanol’b Methanol in strong plastic tanks for direct use as fuel Ethanol 119.9 MJ kg 0.65 19.9 MJ kg 6.9 24 MJ kg 95 Gasoline 46.4 MJ kg 95 Diesel 48 MJ kg 13.33 kWh kg−1 22.8 MJ kg−1 6.34 kWh kg−1 95 12.0 kWh kg−1 −1 18.9 MJ kg−1 5.26 kWh kg−1 6.67 kWh kg−1 −1 8.27 MJ kg−1 2.3 kWh kg−1 c 5.54 kWh kg−1 −1 0.78 MJ kg−1 0.22 kWh kg−1 33.3 kWh kg−1 −1 0.72 MJ kg−1 0.20 kWh kg−1 33.3 kWh kg−1 −1 Net specific energy 44.27 MJ kg−1 11.4 kWh kg−1 95 45.6 MJ kg−1 12.66 kWh kg−1 a Storage efficiency is defined here as the weight of hydrogen stored per kg of total system. For example, a compressed gas cylinder that weighs 500 g will store 0.06 × 500 = 30 g of hydrogen, whereas as vessel containing hydride and weighing the same will contain 32.5 g of hydrogen. b An estimated mass of the reformer is included in the case of ‘indirect methanol’, where it is chemically reacted to produce hydrogen. c The storage of these liquid fuels is assumed to be 95% efficient, i.e., the mass of the liquid is 95% of the total mass of the liquid and storage vessel. A state‐of‐the‐art DMFC operating at about 50°C with a cell voltage of 0.4 V will produce around 5 mW cm−2. Raising the temperature to 70–80°C can lead to power densities of 80–100 mW cm−2. Despite these relatively low figures, DMFCs are attractive for some stationary applications of the small‐to‐medium scale, i.e., up to about 5 kW. Fuel crossover is another issue with the DMFC and has been discussed briefly in Section 3.5, Chapter 3. The phenomenon is particularly acute in the DMFC if the electrolyte is a perfluorosulfonic acid (PFSA) membrane, as described for PEMFCs in Section 4.2.1, Chapter 4. The water that provides the proton conductivity pathway in PFSA membranes can readily absorb methanol, and, as a result, the methanol can quickly migrate from the anode to the cathode. Such action reduces the open‐circuit voltage of the cell that, in turn, adversely affects the performance of the fuel cell at all currents. A comparison of the performance of a state‐of‐the art DMFC with that of a PEMFC is given in Figure 6.1. The shape of the two graphs is broadly similar, but the voltages and current densities of the DMFC are considerably lower. Electrolytes for the DMFC and the problem of fuel crossover are further discussed in Section 6.3. The implications in terms of potential applications for the DMFC are covered in Section 6.4. 159 Fuel Cell Systems Explained 1.2 A typical performance curve of H2-PEMFCs 1.0 Cell voltage/V 160 0.8 0.6 A typical performance curve of direct methanol fuel cells 0.4 0.2 0 100 200 300 400 Current density/mA cm–2 Figure 6.1 Voltage versus current density performance of a 2010 state‐of‐the‐art DMFC and a typical PEMFC when operating under ambient conditions. 6.1.1 Principles of Operation The overall reaction in the DMFC can be expressed as follows: 2CH3 OH 3O2 4 H2 O 2CO2 (6.1) As noted in Section 2.2, Chapter 2, the change in standard Gibbs free energy, g f , for this reaction is −698.2 kJ mol−1. Six electrons are transferred for each molecule of methanol that is consumed, and, from equation (2.11), the reversible cell voltage is therefore given by: Vr gf zF 698.2 1000 1.21 V 6 96 485 (6.2) The practical voltages obtained are considerably less than this, and the losses are greater than those for other types of fuel cell. Indeed, one feature that sets apart the DMFC is that there is considerable voltage loss at both the anode and the cathode. The anode reaction of the DMFC is discussed in more detail in the following section. 6.1.2 Electrode Reactions with a Proton‐Exchange Membrane Electrolyte Methanol can be used as a fuel for both the PEMFC and the alkaline fuel cell (AFC) as described in Chapters and 5, respectively. For the DMFC with a PEM electrolyte, the overall anode reaction is: CH3 OH H2 O CO2 6H 6e (6.3) Direct Liquid Fuel Cells CH3OH HCHOH CHOH CHOH 1 1 HCHO CHO CO 2 2 HCOOH 3 COOH 3 CO2 Figure 6.2 Stepwise reaction paths for oxidation of methanol at a DMFC anode. The H+ ions move through the electrolyte and the electrons travel round the external circuit. Note that water is required at the anode, though it is produced more rapidly at the cathode via the accompanying reaction: 1½O2 + 6H + + 6e − → 3H2 O (6.4) Unlike the direct electrochemical oxidation of hydrogen in a PEMFC, reaction (6.3) takes place in several steps that can take a variety of routes. The first step is the dissociative adsorption of methanol on the platinum (Pt) catalyst with the release of six protons and six electrons that gives rise to a large electric current. The product of the dissociation is a methanolic residue that remains on the catalyst surface, the precise composition of which is still debated. This surface residue is slowly oxidized to CO2 via reaction with water or other adsorbed oxygenated species. Despite a plethora of research carried out by Shell and others before 19801 and by several university research groups since then, the true mechanism of methanol electro‐oxidation has still to be resolved. What can generally be agreed is that following the initial dissociative adsorption step, dehydrogenation involves reaction of the adsorbed species with adsorbed OH groups. The chart in Figure 6.2 is an attempt to illustrate the steps and possible reaction routes that may take place during methanol electro‐oxidation. At the top left of the diagram is methanol; at the bottom right is the main reaction product — carbon dioxide. The lateral steps from left to right involve ‘hydrogen stripping’ or dehydrogenation, i.e., the removal of a hydrogen atom and the generation of a proton (H+) and electron (e−) pair. The downward‐ moving steps not only involve the removal of a hydrogen atom and the generation of a proton–electron pair but also include the addition or destruction of an OH group. 1 Hampson, NA, Willars, MJ, McNicol, BD, 1979, The methanol‐air fuel cell: A selective review of methanol oxidation mechanisms at platinum electrodes in acid electrolytes, Journal of Power Sources, vol. 4(3), pp. 191–201. 161 162 Fuel Cell Systems Explained Any reaction route through the compounds shown in Figure 6.2 from top left to bottom right is possible, and all have the same net result, namely, the oxidation of methanol to carbon dioxide and six proton–electron pairs. The compounds connected by the red arrows are stable compounds, and moving along this sequence might be considered a ‘preferred’ route. The route can be divided neatly into three steps. First, the methanol is converted to methanal (formaldehyde), HCHO, i.e., CH3 OH HCHO 2H 2e (6.5) The methanal then reacts to form methanoic (formic) acid, HCOOH: HCHO H2 O HCOOH 2H 2e (6.6) Finally, the formic acid is oxidized to carbon dioxide: HCOOH CO2 2H 2e (6.7) The sum of the reactions (6.5)–(6.7) is the same as (6.3). The fact that oxidation proceeds via a number of steps leads to the relatively low reaction rates for direct oxidation of methanol. It can also be seen that the formation of carbon monoxide is possible and thereby influences the choice of catalyst — an issue that is discussed in Section 6.1.4. It should be noted in passing that either of the two stable intermediate compounds — formaldehyde or formic acid — could be used as fuels instead of methanol. Their specific energies would be considerably less, however, as only four or two electrons would be produced, respectively, for each molecule of these two fuels. 6.1.3 Electrode Reactions with an Alkaline Electrolyte If an alkaline electrolyte is used for the DMFC, the anode reaction is: CH3 OH 6OH CO2 5H2 O 6e (6.8) The OH− ions are generated at the cathode by the reduction of oxygen: 1½O2 + 3H2 O + 6e − → 6H + (6.9) Unfortunately, the CO2 produced at the anode will react with a hydroxide electrolyte to form carbonates and therefore will rule out any prospect of a DMFC based on the conventional AFC. The advent of anion‐exchange membranes has produced some renewed interest, which is founded largely on the expectation that a more direct anodic oxidation will lead to lower voltage losses and also enable the employment of low‐cost catalysts. Reported power densities from such alkaline DMFCs are, however, much lower than those for equivalent PEM DMFCs (<10 mW cm−2). Direct methanol fuel cells based on PEM technology therefore continue to be the preferred option. 6.1.4 Anode Catalysts Unlike the direct oxidation of hydrogen in the PEMFC, the stepwise oxidation of methanol leads to a considerable activation overpotential at the DMFC anode. Investigations by Shell Research in the 1960s found that platinum by itself was rapidly poisoned by Direct Liquid Fuel Cells adsorbed reaction products and therefore was unsuitable as a catalyst for methanol oxidation. As a consequence, a wide range of platinum alloys was examined for catalytic activity.2 The stepwise oxidation reactions suggest that a bimetallic catalyst may be suitable, each metal promoting the different types of reaction. Significant enhancements in activity were found for platinum modified with rhenium, ruthenium, tin or titanium.3 In particular, Shell researchers found that both platinum–ruthenium (Pt–Ru) and platinum–rhodium (Pt–Rh) performed well as catalysts, but the former was favoured. Electrodes with catalyst loadings of 10 mg cm−2 gave results that encouraged the construction of stacks. A Pt–Ru catalyst has to this day continued to be preferred for methanol oxidation, although some enhancements by additives such as tungstophosphoric acid have been reported. Methanol is believed to be converted to carbonyl species on the Pt (the ‘left‐to‐right horizontal sequences’ in Figure 6.2), and these species are further oxidized by OH− groups that are absorbed on the Ru. As in the PEMFC, the anode catalyst is in the form of finely divided metal particles supported on carbon black and is prepared using the same procedure as that outlined for the PEMFC in Section 4.3.1, Chapter 4. To suppress the activation overpotential, the catalyst loading in DMFCs (often 2–10 mg cm−2) is usually much higher than the corresponding loading in PEMFCs (0.05–0.5 mg cm−2). A small amount of electrolyte ionomer may be added to the DMFC catalyst layer to encourage the product gas (CO2) to be expelled quickly while allowing the methanol–water mixture to penetrate the porous structure. There is scope for improving the DMFC anode catalyst. For example, catalysts supported on carbon nanotubes (CNTS) have recently been found to exhibit higher electrochemical activity than those based on conventional amorphous carbons. Membrane–electrode assemblies with Pt–Ru+CNT anode catalysts have been demonstrated to yield greater power densities than reference samples that employ a Vulcan XC‐72 substrate.4 Electronically conducting oxides (such as oxides of titanium, tin or tungsten) have also been found to enhance the activity of Pt–carbon catalysts. 6.1.5 Cathode Catalysts The oxygen reduction reaction at the cathode of a DMFC, as expressed by equation (6.9), is essentially the same as that for the hydrogen fuel cell with acid electrolyte (reaction (1.2), Chapter 1) except that six electrons are transferred per molecule of methanol, compared with four electrons per molecule of hydrogen. Consequently, the same catalyst can be employed, i.e., platinum supported on carbon. Unfortunately, however, platinum also catalyzes the oxidation of methanol, albeit slowly, and if there is significant crossover from the anode, there will be a substantial deterioration in cell performance. A wide range of alternative oxygen reduction catalysts has been evaluated for use in the DMFC (cf., Section 4.3.2, Chapter 4), but no material has so far proved to be significantly more methanol tolerant than platinum. 2 Andrew, MR and Glazebrook, RW, 1966, in Williams, KR (ed.), An Introduction to Fuel Cells, p. 127, Elsevier, Amsterdam. 3 McNicol, BD, Rand, DAJ and Williams, KR, 1999, Direct methanol–air fuel cells for road transportation, Journal of Power Sources, vol. 83, pp. 15–31. 4 Gan, L, Lu, R, Du, H, Li, B and Kang, F, 2009, High loading of Pt–Ru nanocatalysts by pentagon defects introduced in a bamboo‐shaped carbon nanotube support for high performance anode of direct methanol fuel cells, Electrochemistry Communications, vol. 11(2), pp. 355–358. 163 164 Fuel Cell Systems Explained Unlike the PEMFC, it is not necessary to humidify the air supply to the cathode. Indeed, excess liquid water must be removed at a rate that is sufficient to prevent flooding of the cathode. The mass-transfer resistance of oxygen increases when liquid water is retained in the pores of the macroporous gas-diffusion layer (GDL) and thereby lowers the cathode performance. Excess water may be removed effectively by adjusting the hydrophobic properties and pore distribution of the GDL. The procedure involves bonding a hydrophobic, microporous (pore size 100–500 nm) carbon–polytetrafluoroethylene (PTFE) layer of 10–30 µm thickness onto the GDL. The small hydrophobic pores result in a low permeability of liquid water — a feature that hinders the transport of liquid water from the catalyst layer and forces more liquid water to be transported back to the anode. The outcome is a lower saturation level in the microporous layer and a higher rate of oxygen transport into the catalyst layer. 6.1.6 System Designs There are essentially two different approaches to the design of a DMFC system: ‘active’ and ‘passive’. In the case of active systems, as illustrated in Figure 6.3, pumps, fans and heat-exchangers are used to provide the stack with a controlled supply of Negative terminal Positive terminal Carbon dioxide out Methanol and water mixture Air and water vapour out Water separator and store Methanol sensor Methanol tank Anode Pump or control valve Control valve for water resupply system Cathode Pump Electrolyte Air in Figure 6.3 The main components of an ‘active’ DMFC system. Not all the components will always be present. Larger systems may have additional components, such as heat-exchangers in the fuel system for cooling, air pumps and methanol condensers in the CO2 outlet pipe. Note that the electrode connections are shown at the edge for simplicity; normally, the current will be taken off the whole face of the electrode, as in Figure 1.8, Chapter 1. Direct Liquid Fuel Cells reactants and to remove waste heat and product water. In practice, pure methanol has to be supplemented with a supply of water. As shown previously in reaction (6.1), water is produced within the fuel cell. The product water will evaporate when air passes over the cathode; indeed it is likely that the rate of water evaporation will exceed the rate of water production. (Water management is discussed at length for the PEMFC in Section 4.4, Chapter 4.) In an active DMFC system, water can be recovered from the cathode exhaust, stored and resupplied as required with fuel to the anode. Although recycling water does add system complexity, the use of a dilute solution of methanol as fuel gives two advantages. First, it helps to reduce the propensity for crossover as considered in Section 6.3. The concentration has to be about 1 M (ca. 3 wt.%) to limit methanol crossover. It is important therefore to control the methanol feed rate to ensure that the optimum concentration is maintained, either in response to a methanol flow sensor in the feed system or in response to the output power of the cell. The second advantage with a dilute methanol solution is that water is in contact with the MEA and thus ensures that the membrane is hydrated, which is of course important for PFSA membranes. It should be noted that water management in an active DMFC system is less complex than for most hydrogen PEMFCs. In a passive DMFC system, the cells are usually supplied with reactants by means of diffusion, convection, evaporation and capillary forces. There is no forced recirculation of the methanol–water mixture or water recovery from the cathode, and, therefore, passive designs are much simpler than their active counterparts. Passive systems normally operate with lower power densities and are more suitable for small portable devices. Methanol is usually fed to the cell as a vapour rather than a liquid. As with active systems, water management remains an important issue, and, since water is not supplied to the anode, the MEA in a passive system has to be designed in such a way that sufficient water is transmitted from the cathode to the anode through the membrane itself. It is osmotic drag that moves water from the anode to the cathode in both the PEMFC and DMFC. In the case of the PEMFC, water dragged from the anode to the cathode leads to a build‐up at the cathode and thus to a driving force for back-diffusion to the anode. Due to the high level of water at the anode in normal liquid‐fed DMFCs, there is no driving force for back-diffusion of water, so the cathode tends to become flooded with water. In both designs of DMFC system, the MEA is very similar to that in the PEMFC. It is made by hot‐pressing the membrane between the anode and cathode catalyst layers that are each supported by a carbon GDL. 6.1.7 Fuel Crossover Fuel crossover was considered in general in Section 3.5, Chapter 3, as it occurs to some extent in all types of fuel cell and with all fuels. The problem is particularly severe with a DMFC employing a PEM electrolyte. Since methanol mixes very readily with water, it is able to penetrate the hydrated membrane and therefore migrate from the anode to the cathode. The cathode uses a platinum catalyst that will oxidize the fuel, although not so readily as the Pt–Ru catalyst on the anode. The reaction of the fuel at the cathode is not only a waste of fuel — it will also reduce the cell voltage, for the reasons explained in Section 3.5, Chapter 3. 165 166 Fuel Cell Systems Explained The loss of methanol is often quantified as a ‘crossover current’, i.e., the current that would have been produced had the methanol reacted fully on the anode. The crossover current ic is defined by: ic nFADm Ci I xi (6.10) m where I is the discharge current, n is the number of electrons involved, F is the Faraday constant, A is the electrode area, Dm is the diffusion coefficient of methanol in the PEM, Ci is the methanol concentration at the anode–PEM interface, δm is the thickness of the PEM, ξ is the electro‐osmotic coefficient and xi is the mole fraction of methanol in the solution. The crossover current can be compared with the useful output current i to give a ‘figure of merit’ for a DMFC that is expressed as the fuel utilization coefficient ηf. The coefficient gives the ratio of the fuel that is actually reacted on the anode to the total fuel supplied, namely: i f i ic (6.11) Using the techniques described in the following text, it is possible to raise the figure of merit up to 0.85 or even 0.90, though 0.80 (or 80%) would probably be a more realistic value to expect. 6.1.8 Mitigating Fuel Crossover: Standard Techniques There are four principal ways to reduce fuel crossover: 1) The anode catalyst is made as active as possible, within the bounds of reasonable cost, so that less methanol is available for diffusion through the electrolyte to the cathode. 2) The fuel feed to the anode is controlled, so that methanol is not present in excess at low currents. Clearly, the lower the methanol concentration at the anode, the lower will it be in the electrolyte, and hence at the cathode, as shown in Figure 6.4. A concentration of 1 M is regarded as optimum for most DMFC applications. 3) Thick PEM electrolytes should be used (i.e., thicker than what is normal for PEMFCs). These will not only reduce crossover but also increase the cell resistance, and thus a compromise will need to be sought. For a DMFC, the membrane normally has a thickness of between 0.15 and 0.20 mm5 compared with between 0.05 and 0.10 mm for a hydrogen PEMFC.6 4) In addition to its thickness, the composition of the PEM also has an effect. Studies have shown that the diffusion and water uptake for 1100EW Nafion is about half that for 1200EW Nafion. (EW is the weight in grammes of Nafion, in terms of molecular mass, per sulfonic acid group.) 5 For example, Du Pont’s Nafion 117 at 0.18 mm. 6 For example, Du Pont’s Nafion 112 at 0.05 mm. Direct Liquid Fuel Cells 120 Crossover current/mA cm–2 100 80 1.0 M 60 40 0.5 M 20 0 0.25 M 0 100 200 300 400 500 600 700 800 Useful output current density/mA cm–2 Figure 6.4 Graph showing how the crossover of methanol to the cathode changes with fuel concentration at the anode and with load current. (Source: Ren, X, Zelanay, P, Thomas, S, Davey, J and Gottesfeld, S, 2000, Recent advances in direct methanol fuel cells at Los Alamos National Laboratory, Journal of Power Sources, vol. 86, pp. 111–116.) Techniques 1 and 2 reduce the likelihood of crossover by promoting reaction of methanol at the anode. The anode reaction can also be boosted by increasing the current drawn from the cell. 6.1.9 Mitigating Fuel Crossover: Prospective Techniques In addition to the (almost) universally applied four standard techniques, there are other methods under investigation that are more experimental or at a very early stage of development. Among these efforts are the following: 1) Use of selective (non‐platinum) cathode catalysts. These materials will stop the fuel from reacting on the cathode and so eliminate the voltage drop due to the mixed potential that results from the simultaneous reduction of oxygen and oxidation of methanol. Examples of non‐precious metal catalysts for the PEMFC are given in Section 4.3.2, Chapter 4. Organic transition metal complexes appear to be suitable to some extent except they tend to break down at high temperatures and high potentials. Transition metal chalcogenides have also been shown to be more tolerant to methanol than platinum. A few inorganic materials have been proposed as suitable substitutes, including transition metal sulfides (MoxRuySz, MoxRhySz) or chalcogenides (Ru1−xMoxSeOz). The methanol tolerance mechanism of these materials is due to the absence of adsorption sites for methanol dehydrogenation. On the other hand, the catalytic activity of these materials towards the oxygen reduction reaction is still much lower than that of platinum. In addition, although the methanol that has crossed over to the cathode may not react on such materials, it will probably just evaporate and therefore be wasted. This approach therefore does not offer a complete solution. 167 168 Fuel Cell Systems Explained 2) Insertion of a layer in the electrolyte that is porous to protons but less so to methanol. If such a material could be found, then this would obviously be a solution to the problem. Some of the ideas being tried in this area include treating the surface of the Nafion membrane and also coating it with a very thin layer of palladium. The idea of using bilayer membranes is now well established as is the incorporation of additives in the PEM that discourages methanol crossover. The problem with such approaches is that proton conductivity is usually compromised. 3) Development of more conductive PEMs. In general, it appears that a reduction in the size and amount of water clusters inside the ionomer will lead to reduced methanol crossover while retaining good proton conductivity. In terms of alternative non‐ fluorinated PEM materials, blends of polybenzimidazole (PBI) with sulfonated poly(ether ether ketone) (SPEEK) or sulfonated poly(arylene ether sulfone) (BPSH‐30) or sulfonated poly(arylene ether ether nitrile) (m‐SPAEEN‐60) show the most promising characteristics for DMFCs. 6.1.10 Methanol Production The potential of the DMFC, as well as the steam reforming of methanol given consideration in Chapter 8, relies on the fact that methanol is produced in bulk at a reasonable cost. Production in 2013 amounted to 70 million tonnes and global demand in 2016 is over 90 million tonnes. According to the US Methanol Institute, about 30% of the global demand is for producing formaldehyde (employed in the preparation of urea–formaldehyde and phenol–formaldehyde resins for particle board and other construction materials). Methanol is also required for the production of other important industrial chemicals such as acetic acid, various cleaners and windshield washer fluid for cars. Only about 2% is currently used as a fuel, typically blended with other hydrocarbons. Methanol can be formed by the reaction of hydrogen and carbon monoxide over a suitable catalyst at high pressures and, therefore, can be obtained by the steam reforming of natural gas or another hydrocarbon fuel or biofuel (see Chapter 10). As methanol is required in large quantities for industrial purposes, considerable effort has been expended on making the production process as efficient as possible. The overall efficiency is ~70% and ~60% for conversion from natural gas and renewable fuels, respectively. Such processing efficiencies mean that the cost of methanol is principally governed by the cost of the raw fuel. Posted prices by Methanex in July 2016 range from US$240 to 275 per tonne or between US$0.72 and 0.82 per gallon, which puts the fuel at a price level comparable with that of gasoline. 6.1.11 Methanol Safety and Storage The use of methanol in consumer products, such as power supplies for portable electronics equipment, raises potential safety concerns. The first issue is flammability and the fact that methanol burns with an invisible flame. The same situation, of course, applies to hydrogen, and studies have shown that hydrogen and methanol are both equal from a safety point of view and both are better than gasoline. The second issue relates to the toxicity of methanol. The chemical is a poison, made worse by the fact that it can mix easily with water. Furthermore, it does not have a taste Direct Liquid Fuel Cells that would make it immediately repellent. This ‘drinkability’ problem causes methanol to be considerably more dangerous in everyday use than other fuels — such as gasoline — which are also poisonous and in wide circulation. The safety arguments are fairly complex; however, because methanol is naturally present in the human body it is perfectly safe in small quantities. Indeed, methanol is produced in the body by the digestion of a wide range of natural products (e.g., fruit) as well as man‐made additives such as the aspartame sweetener used in ‘diet’ drinks. By contrast, when liquid methanol enters the body by direct contact with the skin, by drinking or by inhaling the vapour, it transfers to the liver where it is converted to formaldehyde. In turn, the formaldehyde oxidizes to formic acid that interferes with the function of mitochondria and thereby causes toxic injury to the retina and optic nerve that frequently results in blindness and, in extreme cases, death. Data collected by the American Association of Pollution Control Centers suggest that most deaths from methanol poisoning are suicides or the result of self‐harm actions. Methanol systems must therefore be designed in such a way that deliberate and premeditated drinking of methanol is very difficult — stopping accidental consumption is not enough. The usual practice of including an additive with the liquid that makes it undrinkable could be problematic if the additive interferes with the operation of the fuel cell. Inhaling methanol vapour is particularly dangerous and needs to be considered in any system in which methanol is vapourized. The storage of methanol poses few problems. The liquid should be kept in airtight stainless steel or plastic containers because it can absorb moisture from the atmosphere, and even in dilute form a methanol–water mixture is quite corrosive. Methanol can also dissolve some polymers, so care must be taken over the choice of materials for the storage vessel, gaskets and tubing. For small‐scale or portable applications, storage concerns have largely been addressed and both the International Civil Aviation Organization’s Dangerous Goods Panel and the US Department of Transportation now permit approved fuel cells with installed methanol cartridges to be carried by passengers and crew on aircraft. Large‐scale methanol storage requires special measures to be taken in terms of venting and fire safety, but these are beyond the scope of this book. 6.2 Direct Ethanol Fuel Cells A direct ethanol fuel cell (DEFC) is more acceptable to consumers and for commercial markets compared with cells using toxic methanol. Ethanol has a higher specific energy (6.67 kWh kg−1, LHV) than methanol (5.54 kWh kg−1, LHV) and a higher boiling point (78°C). It is potentially attractive for both stationary and mobile fuel‐cell systems. Ethanol can be obtained from renewable plant material such as food crops (e.g., cereal grains, sugarcane, sugar beets) or other forms of biomass such as switch grass and forest residue. It can therefore be deemed to be a ‘zero‐carbon fuel’, unlike methanol made from natural gas, since the CO2 produced when bioethanol is oxidized is returned to the atmosphere to be consumed by renewable plants. Commercially, ethanol is produced by three main routes: fermentation of starch and sugar using yeast, processing of biomass with bacteria and reaction of ethene (which can be produced from oil) with steam over a catalyst. 169 170 Fuel Cell Systems Explained 6.2.1 Principles of Operation In the DEFC, ethanol in an anhydrous form or diluted with water, either as liquid or vapour, is fed directly into the anode, with air supplied to the cathode. Conceptually, as with the DMFC, the DEFC may employ either acid or alkaline electrolytes. The direct overall oxidation of ethanol can be represented as: C 2 H5 OH 3O2 G 2CO2 1325 kJ mol 1 3H2 O Vr (6.12) 1.145 V The standard reversible voltage at 1.145 V is less than that of the DMFC (1.21 V), but unlike the DMFC in which there are 6 electrons involved in the electrochemical reactions with ethanol oxidation, there are 12 electrons. In an acid DEFC, the anode reaction is: C 2 H5 OH 3O2 2CO2 12H 12e (6.13) The corresponding oxygen reduction at the cathode is: 3O2 12e 12H 6CH2 O (6.14) As was the case with the DMFC, the earliest work on DEFCs was with an alkaline electrolyte (aqueous KOH). In this version, hydroxide ions are the electroactive species, and the electrode reactions are as follows: Anode : C 2 H5 OH 12OH Cathode : 3O2 6CH2 O 12e 2CO2 9H2 O 2e 12OH (6.15) (6.16) Unfortunately, as with the DMFC and AFCs in general, there is the problem of contamination of the electrolyte by CO2 that is generated at the anode of the DEFC in addition to being present in the air supplied to the cathode. Some work has been conducted in which the aqueous KOH electrolyte is replaced by an OH− ion‐conducting membrane. To date, published research suggests that with membranes such as alkali‐doped PBI, the benefits may not be substantially better than the performance obtained from a DEFC with a proton‐conducting membrane. 6.2.2 Ethanol Oxidation, Catalyst and Reaction Mechanism The main challenge in the electrochemical oxidation of ethanol is the breaking of the C─C bond in the molecule. Given that PEM electrolytes function at low temperatures (below 100°C), the requirements placed on the catalyst are even more demanding than those for methanol oxidation. Nevertheless, both carbon‐supported Pt–Ru and Pt–Sn catalysts have been found to be suitable. As with methanol oxidation, bifunctional catalysts are required because Pt can easily be poisoned by the carbonyl reaction intermediates. In this respect, the reaction mechanism for ethanol oxidation is more complex and even more uncertain than that for methanol. The present understanding of the reaction in an acidic DEFC can be summarized as follows: The reaction proceeds via a multistep mechanism that involves a number of Direct Liquid Fuel Cells CH3 Ethanol HO H H H2O+ OH2 H H3C O H OH2 CH3 H3O+ O 1e– 0 Acetaldehyde O H 0.8 1e– Adsorbed acetyl CH3 OH2 H3 O+ O C Acetaldehyde EV (RHE) CH3 O CH3 H Acetic acid OH OH 1e– H3O+ + 1e– 2H2O Adsorbed acetyl CH3 H CH4 E < 0.2 V (RHE) H 3C C – O – – O C H O OH2 H2O+ O C O E > 0.5 V (RHE) 1e– Figure 6.5 Mechanism of ethanol electro‐oxidation at platinum surface in acid medium. (RHE = reversible hydrogen electrode.) (Source: Reproduced with permission from Vigier, F, Rousseau, S, Coutancean, C, Leger, J‐M and Lamy, C, 2006, Electrocatalysis for the direct alcohol fuel cell, Topics in Catalysis, vol. 40(1), pp. 111–121. Reproduced with permission of Springer.) adsorbed reaction intermediates and by‐products, which emanate from the incomplete oxidation of ethanol. The major intermediates have been identified as adsorbed carbon monoxide (CO) and C1 and C2 hydrocarbon residues, whereas acetaldehyde and acetic acid have been detected as the main by‐products. Information obtained from electrochemical and spectro‐electrochemical studies has led to the proposal and general acceptance of the reaction mechanism outlined in Figure 6.5.7 The first reaction product from the dissociative adsorption of ethanol on platinum is acetaldehyde, which requires the transfer of only two electrons per ethanol molecule. Acetaldehyde has to re‐adsorb on the catalyst to complete its oxidation into either acetic acid (CH3COOH) or CO2 with methane produced at low potentials. Unlike acetaldehyde, it is difficult to oxidize acetic acid further at low temperatures, so it becomes a ‘dead-end’ in the reaction. The possible formation of many intermediate products rather than the complete conversion of ethanol to CO2 leads to a significantly higher overpotential at the DEFC 7 Vigier, F, Rousseau, S, Coutancean, C, Leger J‐M and Lamy, C, 2006, Electrocatalysis for the direct alcohol fuel cell, Topics in Catalysis, vol. 40(1), pp. 111–121. 171 172 Fuel Cell Systems Explained anode in addition to the drying of Nafion® and other perfluorinated sulfonic acid (PSFA) membranes with concomitant loss in proton conductivity. A slight improvement in the rate of ethanol oxidation can be achieved by increasing the operating temperature, and, in 1998, the first reasonable performance given by a DEFC was reported.8 The cell employed a Pt–Ru anode catalyst and Pt cathode catalyst, both supported on carbon, together with a composite membrane synthesized from Nafion and silica (SiO2). Although the power density of 110 mW cm−2 was about half of that obtainable with a DMFC under the same conditions (0.6 A cm−2 and 0.4 V at 550 kPa and 145°C), the high selectivity towards the formation of CO2 (95%) enhanced the prospect of a viable DEFC. Research on different catalyst materials suggests that the first step of the oxidation reaction occurs on the Pt surface and is not enhanced by alloying the Pt with other metals. Indeed, the combination of Ru with Pt appears to inhibit cleavage of the C─C bond. On the other hand, the addition of Ru restricts the formation of unwanted intermediates and thereby improves the selectivity towards CO2. A substantial body of literature on ethanol oxidation catalysts has emerged over the past 20 years and has shown that Pt modified with Sn and/or Ru are effective combinations. In general, and in contrast to the DMFC, Pt in conjunction with Sn is presently the more active binary catalyst for ethanol oxidation. Catalysts comprising Pt–Sn–Ru with a nominal Ru:Sn atomic ratio of less than 1 appear overall to be the most promising ternary anode catalysts, despite the fact that the Sn and Ru both inhibit cleavage of the C─C bond. For a DEFC with an alkaline electrolyte, the anode reaction is subtly different from that with an acid electrolyte, cf., reactions (6.13) and (6.15), although obviously the cleavage of the C─C bond is a significant issue in both cells. Interestingly, platinum on carbon has twice the activity for ethanol oxidation in an AFC than it does in an acid cell. In early work, non‐precious metal catalysts (iron, nickel and/or cobalt supported on carbon) showed promise as anode catalysts in alkaline systems, but they were less active than Pt–C. Palladium, which gives no activity at all in acid fuel cells, has recently been found to be effective in alkaline DEFCs, especially if combined with certain oxides such as ceria (CeO2) or titania (TiO2). Unfortunately, most of the catalysts in alkaline systems do not oxidize the ethanol completely to carbon dioxide but typically stop at the oxidation level of acetic acid, and this may be acceptable for some applications. Despite this limitation, NDC Power based in Cheyenne, USA, has developed a platinum‐free DEFC that has been scaled up to 10‐kW prototype stacks for military applications. 6.2.3 Low‐Temperature Operation: Performance and Challenges The performance of state‐of‐the‐art DEFCs is significantly inferior to that of DMFCs. The main challenge is the slow kinetics of electrochemical reactions on both electrodes, especially at the anode. The high overpotential at this electrode is due to both the difficulty in breaking the strong C─C bond in ethanol and the slow reaction rate given the number of reaction steps, coupled with the low selectivity to complete oxidation and CO2 production. Although bi‐/tri‐metallic Pt‐based catalysts have been extensively 8 Arico, AS, Creti, P, Antonucci, PL, and Antonucci, V, 1998, Comparison of ethanol and methanol oxidation in a liquid‐feed solid polymer electrolyte fuel cell at high temperature, Electrochemical and Solid‐State Letters, vol. 1, pp. 66–68. Direct Liquid Fuel Cells studied for the anode, the activity of the best catalysts is still too low for practical application. Compared with the anode, oxygen reduction at the cathode is relatively fast; nevertheless, there is still room for improvement of the Pt–C catalyst that is the most often employed. As with the DMFC, another challenge is the crossover of ethanol from anode to cathode. Not surprisingly, raising the temperature of operation improves the performance of the DEFC, as has been demonstrated9 for a cell with a composite Nafion–silica membrane — the peak power density increases from about 60 mW cm−2 at 90°C to 90 mW cm−2 at 130°C. 6.2.4 High‐Temperature Direct Ethanol Fuel Cells Both molten carbonate (MCFC) and solid oxide (SOFC) fuel cells are able to accept alcohols directly as fuels. When fed to the anode in either of these two technologies, ethanol is able to react on the nickel‐containing anode or catalyst via a number of reactions, such as steam reforming, partial oxidation, autothermal reforming and dry (CO2) reforming. The reactions are discussed in Chapter 10. Unfortunately, carbon deposition can be an issue, particularly in high‐temperature SOFCs (>800°C). Ethanol may decompose and deposit carbon either in the inlet channels of the SOFC or on the nickel anode material. Unless sufficient steam is present to suppress the reaction, carbon formation can prove to be a significant problem for both MCFCs and SOFCs. There are many factors that determine how and where carbon may occur and include, for example, the temperature of operation and the composition of the anode catalyst. To reduce the operating temperature for ethanol oxidation, some work has been undertaken with a combination of SOFC and MCFC electrolytes, namely, incorporating molten carbonate in a solid oxygen ion‐conducting matrix. With such material, power densities as high as 500 mW cm−2 have been reported for ethanol at 580°C.10 6.3 Direct Propanol Fuel Cells In the search for ways to reduce fuel crossover, alcohols other than methanol and ethanol have been evaluated on Pt or Pt–Ru electrodes. For example, the electro‐oxidation of propan‐2‐ol on platinum electrodes was studied by several research groups during the 1990s. The potential advantages of this fuel are as follows: (i) it is relatively less toxic than other alcohols, (ii) at low potentials, it is less prone to anode poisoning and (iii) it has better resistance to crossover and cathode poisoning. The performance of the fuel cell for a given catalyst system has been associated with the effect of parameters such as 2‐propanol concentration, anode and cathode fuel flow rates, cell temperature and oxidant back-pressure. Optimization studies have shown that a power density of 45 mW cm−2 can be achieved with 1.5 M propan‐2‐ol at a cell temperature of 80°C when 9 Di Blasi, A, Baglio, V, Stassi, A, D’Urso, C, Antonucci, V and Aricò, AS, 2006, Composite polymer electrolyte for direct ethanol fuel cell application, ECS Transactions, vol. 3(1), pp. 1317–1323. 10 Mat, DM, Liu, X, Zhu, Z and Zhu, B, 2007, Development of cathodes for methanol and ethanol fuelled low temperature (300–360°C) solid oxide fuel cells, International Journal of Hydrogen Energy, vol. 32, pp. 796–801. 173 174 Fuel Cell Systems Explained using a Nafion 117 PEM and anode and cathode loadings of 4 mg Pt–Ru cm−2 and 1 mg Pt cm−2, respectively. Direct propan‐2‐ol fuel cells have also been tested with both Nafion and H3PO4‐doped PBI membrane electrolytes bonded with Pt–Ru catalysts. The initial product of oxidation at low potentials (<0.4 V vs. RHE) is acetone, which undergoes further oxidation at higher potentials. Obviously, with alcohols of higher molecular weight, there is the prospect of more intermediates forming at the anode catalyst that have the potential to poison or degrade the catalyst. Consequently, it remains to be seen whether systems employing propanol or other alcohols will fare any better than methanol or ethanol. 6.4 Direct Ethylene Glycol Fuel Cells Ethylene glycol (CH2OH)2 is another alcohol that is manufactured on a large scale; globally more than seven million tonnes are produced each year. It is widely available as antifreeze for the radiators of automobiles and as an ingredient of popular plastics, such as polyethylene terephthalate (PET). It has a number of features that make it superior to methanol for fuel cells, namely: ● ● ● ● High boiling point 198°C (cf., methanol 64.7°C) and low vapour pressure. High energy capacity 4.8 Ah ml−1 (cf., methanol 4.0 Ah ml−1). An established distribution infrastructure for the automobile industry. An OH group on each carbon atom, therefore more easily oxidized than ethanol. 6.4.1 Principles of Operation The operation of direct ethylene glycol fuel cell (DEGFC) is exactly the same as that of both the previous direct alcohol fuel cells. With an acid electrolyte, the oxidation of fuel at the anode can be represented by: CH2 OH 2 2H 2 O 2CO2 10H 10e (6.17) Thus, if ethylene glycol could be oxidized to CO2 completely, 10 electrons would be obtained from one molecule of the fuel. Even with the most active anode catalyst, however, complete oxidation is rarely achieved, and, as with ethanol, the energy required to break the C─C bond is the main barrier. Most DEGFCs have been based on PEM electrolytes in which protons transfer from the anode to the cathode. When such cells operate at room temperature and potentials below about 0.9 V, ethylene glycol is degraded to oxalic acid, which cannot be oxidized further. Electro‐osmotic drag also transfers water from the anode to the cathode; consequently fuel crossover is also an issue. A mixed potential can be formed between the oxidation of transferred ethylene glycol and the reduction of oxygen at the cathode. In the 1970s, Siemens built an alkaline DEGFC system that employed circulating KOH solution as the electrolyte. The stack consisted of 52 cells that produced 28 V at 4.5 A (normal power output 125 W) and 16 V at 14 A (225 W peak power). The anode catalyst was a platinum–palladium–bismuth alloy with a Pt + Pd loading of about 5 mg cm−2. In the alkaline version of the cell, the anode reaction can be expressed as: Direct Liquid Fuel Cells CH2 OH 2 10OH 2CO2 8H2 O 10e (6.18) As with other alkaline cells, CO2 is formed at the anode and thereby has an immediate impact on the local pH of the electrolyte. If the liquid electrolyte is replaced by an anion‐exchange membrane, stability under reduced pH has also proved to be an issue, and further effort must be undertaken to develop an anionic membrane that is robust enough for this type of fuel cell. 6.4.2 Ethylene Glycol: Anodic Oxidation The probable pathway for ethylene glycol oxidation is shown in Figure 6.6.11 As with other alcohols, the oxidation process involves several consecutive and parallel adsorption–desorption reaction steps that feature intermediates. The reaction can progress easily and stop with the formation of oxalic acid or oxalates since the C─C bond cannot be cleaved. Some C2 products have also been detected in the product stream, notably formic acid and formaldehyde. The partial oxidation of ethylene glycol up to oxalic acid (involving the transfer of eight electrons) yields 3840 Ah L−3 (3450 Ah kg−1); this value is comparable with that of the complete oxidation of methanol to CO2, namely, 3970 Ah L−3 (5019 Ah kg−1). This is also the case with ethanol. Partial oxidation to intermediate products is certainly not desirable — not only because of the loss of energy but also on account of the toxicity of substances such as glyoxal, glocyoxylic acid and oxalic acid. Raising the platinum loading on the anode catalyst increases the probability of re‐adsorption of such intermediates and leads to more complete oxidation. The alternative approach, as practised with the other direct alcohol fuel cells, is to use platinum alloys such as Pt–Ru and Pt–Sn as oxidation catalysts. Many bimetallic and tri‐metallic catalysts supported on traditional amorphous carbons or multi‐walled carbon nanotubes (MWCNTs) have COOH 2H+ CH2OH CH2OH Ethyelene glycol 2H+ 2e– CHO CH2OH Glycolic acid 2e– 2e– 2H+ COOH CHO Glyoxalic acid CH2OH Glycol aldehyde 2e– 2H+ 2e– COOH COOH Oxalic acid 2H+ 2CO2 2e– 2e– CHO 2H+ CHO Glyoxal 2H+ Figure 6.6 Stepwise reaction scheme for the oxidation of ethylene glycol. 11 Ogumi, Z and Miyazaki, K, 2009, Direct ethanol glycol fuel cells, in Garche, J, Dyer, CK, Moseley, PT, Ogumi, Z, Rand, DAJ and Scrosati, B (eds.), Encyclopedia of Electrochemical Power Sources, pp. 412–419, Elsevier, Amsterdam. 175 176 Fuel Cell Systems Explained been tested for ethylene glycol oxidation. The reaction scheme shown in Figure 6.6 is notably more complex than that for methanol or ethanol, and the ability of the different catalyst compositions to oxidize the various reaction intermediates is not well established. Elucidation of both the nature of the reaction pathway and the role of the catalyst components requires further research. 6.4.3 Cell Performance Since the boiling point of ethylene glycol is higher than that of either methanol or ethanol and its vapour pressure is lower, DEGFCs are able to operate at higher temperatures. As the temperature is raised, so the rates of the electrochemical reactions increase, together with an enhancement in the proton conductivity of the membrane, but with the downside of an escalation in crossover. Improvements to the selectivity for oxygen reduction by the cathode catalyst will therefore be necessary if the advantages of higher temperatures are to be realized. To date, most of the development of DEGFCs has focused on the use of PFSA membranes, principally the various forms of Nafion. Few studies have employed membranes that are capable of operation at temperatures above about 90°C. Two other factors have been found to affect the performance of DEGFCs, namely, pH and ethylene glycol concentration. In an alkaline cell with a KOH electrolyte, the generation of CO2 at the anode can reduce the pH of the electrode and thereby lower the cell performance through adsorption of carbonate ions on the platinum catalyst. This adverse effect can be mitigated by employing bimetallic catalysts (e.g., Pt–Ru). In alkaline electrolytes, high concentrations of ethylene glycol can lead to increased crossover, all other factors being equal. At the same time, it has been reported that increasing the concentration from 1 to 6 M delivers higher current densities. 6.5 Formic Acid Fuel Cells Formic acid (also known as methanoic acid), HCOOH, is the simplest of the carboxylic acids with just one carbon atom in the molecule. As noted in the previous sections, it is formed as an intermediate product in the direct oxidation of other alcohols and can be used as a fuel in its own right. Formic acid has the lowest energy density among the liquid alcohols listed earlier in Table 6.1, but this is offset by the higher theoretical cell voltage (1.41 V) compared with other direct acid fuel cells. The overall reaction of the direct formic acid fuel cell (DFAFC) can be represented by HCOOH + ½O2 → CO2 + H2 O (6.19) As with the alcohols considered previously, interest in formic acid has been primarily for the low‐temperature fuel cells, although many other systems could probably operate with this fuel. Formic acid will react with alkaline electrolytes to produce salts (formates) that cause cell degradation. This effectively rules out using the fuel with alkaline cells and leaves PEMFC systems as the preferred option. The total direct anodic oxidation of formic acid involves the transfer of only two electrons per molecule, i.e., HCOOH 2H CO2 2e (6.20) Direct Liquid Fuel Cells Unlike the higher alcohols, in which many electrons are transferred in multiple steps, there are few steps in the electro‐oxidation reaction of formic acid. This feature increases the likelihood for complete oxidation to CO2 and therefore better fuel utilization. Note that, unlike methanol (reaction (6.3)) or ethanol (reaction (6.14)), water is not required for the oxidation of formic acid, and there is no C─C bond to be cleaved. In the case of the DMFC, for example, methanol concentrations of 0.5 M are required, whereas formic acid can, in theory, be used in the fuel cell without dilution. Although it is necessary to keep PFSA membranes hydrated, the complex water management provision that is usually required with PEMFC‐related fuel cells is not an issue when employing formic acid as the fuel. 6.5.1 Formic Acid: Anodic Oxidation Formic acid displays higher activity for oxidation than the alcohols discussed earlier, and therefore both platinum and palladium are effective as anode catalysts in DFACs. It is now well established that the oxidation of formic acid on a platinum metal surface takes place through a dual‐pathway mechanism.12 The direct path can be represented by the fast reaction (6.20), which proceeds via highly reactive intermediates. By contrast, the oxidation can follow an indirect pathway that involves site blocking or poisoning intermediates, such as the strongly adsorbed ─COOH intermediate also encountered during the electro‐oxidation of methanol on platinum, i.e., Pt HCOOH Pt COOH x H (6.21) e The adsorbed ─COOH intermediate is then oxidized by adsorbed Pt–OH species, namely: Pt COOH x Pt OH 2Pt CO2 H2 O (6.22) The direct and indirect pathways are manifested in cyclic voltammograms performed on platinum electrodes, as shown in Figure 6.7. The peak current at 0.5–0.6 V (with respect to the RHE) corresponds to the direct pathway, whereas the peak approximately 0.9 V corresponds to the indirect pathway. In most practical cases of formic acid oxidation, the indirect mechanism predominates. As with DMFCs, Pt–Ru is found to perform well as an anode catalyst, particularly at low temperatures. With palladium catalysts, the direct pathway is dominant, and no adsorbed reaction intermediates are detected. The most studied alloy for formic acid is Pd–Pt that gives a high activity (conversion efficiency of about 50%) which is comparable with that of hydrogen on a Pt electrode in a PEMFC. 6.5.2 Cell Performance The solubility of formic acid in water is of a similar order to that of methanol. Nevertheless, it poses less of a crossover issue with PFSA membranes because it partially ionizes to create formate ions (HCOO−) that are repelled by the negatively-charged sulfonate groups on the membrane molecules. Formic acid is much less hygroscopic than 12 Capon, A and Parsons, R, 1973, The oxidation of formic acid on noble metal electrodes III, intermediates and mechanism on platinum electrodes, Journal of Electroanalytical Chemistry, vol. 45, pp. 205–231. 177 Fuel Cell Systems Explained 150 Current / mA 178 100 50 0.8 0 0.2 1.2 Electrode potential/ V vs. RHE Figure 6.7 Cyclic voltammogram for a smooth Pt bead electrode (diameter = 0.15 mm) at 25°C in 0.5 M H2SO4 containing 0.1 M HCOOH. Scan rate = 140 mV s−1. (–) continuous reading and (‐‐‐) after 5 min at open circuit. (Source: Reproduced from Capon, A and Parsons, R, 1973, The oxidation of formic acid on noble metal electrodes III. Intermediates and mechanism on platinum electrodes, Journal of Electroanalytical Chemistry, vol. 45, pp. 205–231. Reproduced with the permission of Elsevier.) methanol, and this also helps to reduce crossover since, even in its most concentrated form, it will not wet hygroscopic materials such as the membrane and GDLs. The high‐power density for formic acid in combination with high concentration and low crossover results in high current density being achieved for DFACs compared with other liquid‐fuelled cells. For instance, a DFAC with an open‐circuit voltage of 0.72 V has been shown to deliver a maximum current density of 134 mA cm−2 and power outputs of up to 49 mW cm−2 with 12 M formic acid.13 6.6 Borohydride Fuel Cells The possibility of using sodium borohydride as a fuel for AFCs was introduced in Chapter 5, but there are few published references to borohydride fuel cells before 2000. As a fuel, sodium borohydride can safely be shipped as a white solid or as a 30 wt.% aqueous solution. The chemical has wide application in chemistry as a reducing agent for converting sulfur dioxide to sodium dithionate that serves as a bleaching agent for wood pulp and in the dye industry. The chemical is also used in the synthesis of vitamin A and for the reduction of aldehydes to ketones and alcohols in the preparation of various antibiotics. In acid solution or in the presence of a catalyst, sodium borohydride will react with water to produce hydrogen: NaBH 4 2H 2 O NaBO2 4 H2 (6.23) 13 Rice, C, Ha, S, Masel, RI, Waszczuk, P, Wieckowski, A and Barnard, T, 2002, Direct formic acid fuel cells, Journal of Power Sources, vol. 111(1), pp. 83–89. Direct Liquid Fuel Cells It could therefore act as a source of hydrogen for a conventional PEMFC or AFC. The term ‘indirect borohydride fuel cell (IBFC)’ is used to define such a fuel cell that is supplied with hydrogen generated in a separate reactor through the reaction of borohydride and water. By contrast, the ‘direct borohydride fuel cell (DBFC)’ refers to an AFC in which the borohydride is decomposed and oxidized directly on the anode. In theory, hydrogen should not be produced in the DBFC by borohydride decomposition and consequently a higher efficiency may be achieved. The direct decomposition of sodium borohydride on the anode of a DBFC takes place as follows: NaBH 4 8OH NaBO2 6H 2 O e E 1.24 V (6.24) The electrolyte in the DBFC is an alkali through which OH− ions migrate to the anode from the cathode where they are produced by reduction of oxygen as in the traditional AFC, i.e., O2 2H 2 O 4 e 4OH 0.4 V E (6.25) The combination of reactions (6.24) and (6.25) yields a theoretical cell voltage of 1.64 V, which is higher than any of the other DLFCs discussed in this chapter. The sodium borate (NaBO2) produced by the direct anode reaction (6.24) is environmentally friendly and can be recycled back to sodium borohydride. The high cell potential of the DBFC gives rise to a theoretical specific energy of 9.3 kWh kg−1, which is greater than that of methanol (5.54 kWh kg−1). The specific energy will be lower when using borohydride solutions despite the fact that these can be prepared up to a level of 30 wt.% in concentrated alkaline aqueous solutions (>6 M). Unfortunately, the high anode potential predicted by reaction (6.24) is rarely achieved, because on most metals borohydride spontaneously hydrolyses to generate a hydroxyl borohydride intermediate and then hydrogen according to BH 4 H2 O BH3 OH BH3 OH H2 O e H2 BO2 e (6.26) 3H2 (6.27) The presence of atomic hydrogen on the DBFC anode gives this electrode a mixed potential as a consequence of reaction (6.24) and the competing reaction: H2 2OH 2H 2 O 2e E 0.828 V (6.28) The observed anode potential is therefore between −1.24 and −0.828 V. Note that reactions (6.26) and (6.27) form the basis of the IBFC. They can be carried out in a reactor that is fed with borohydride solution and are favoured by neutral or acid conditions and promoted by catalysts. An alternate means of generating OH− ions at the cathode of the DBFC, or indeed any AFC, is by the reduction of hydrogen peroxide; the method has been proposed for application in underwater vehicles and anaerobic systems. Research has demonstrated that alkali systems employing hydrogen peroxide produce cell voltages that are higher than those obtained when using air‐breathing cathodes. 179 180 Fuel Cell Systems Explained 6.6.1 Anode Catalysts The electrochemical oxidation of borohydride was considered for fuel‐cell applications in the 1960s using porous nickel and palladium anodes. Direct oxidation requires selective anode catalysts with high activity for reaction (6.24) and low activity for the hydrolysis reaction (6.23). Recent electrode materials that have been investigated include N2B, Pd–Ni, Au, colloidal Au, Au alloys with Pt and Pd, MnO2, mischmetal,14 AB5‐type hydrogen storage alloys (see Chapter 11), Raney Ni, Cu, colloidal Osmium and Osmium alloys. Only gold appears to achieve the transfer of eight electrons that is predicted by reaction (6.24) and leads to the highest cell voltage. Nickel, while still active, enables the transfer of just four electrons, i.e., about half of the theoretical energy value from the fuel can be obtained with nickel anodes. As gold is not able to absorb hydrogen, it is proposed that the oxidation reaction on this metal proceeds via a mechanism in which the first step is the creation of borohydride radicals BH4• by the extraction of electrons, i.e., BH 4 e BH 4 • (6.29) The second step involves the oxidation of the radical to BH3− and water, followed by the formation of diborane, B2H6, which undergoes further electron transfers. On other metals such as Ni, Pt or Pd, the borohydride radical is dissociated on the surface according to the following reaction where M represents the catalyst metal: M BH 4 • M BH3 M H (6.30) The adsorbed BH3− is then oxidized by surface and electron-transfer reactions. Most metals do not achieve the transfer of eight electrons, which indicates partial oxidation to intermediate products, or high rates of reaction (6.23). A number of hydrogen storage alloys have been proposed as anode catalysts for the borohydride fuel cell, e.g., ZrCr0.8Ni1.2, MmNi3.55Al0.3Mn0.4Co0.79 (Mm = mischmetal). Hydrogen is generated by the borohydride, which is then stored in the lattice of the alloy. Such electrodes have exhibited moderately high current densities (up to 300 mA cm−1) at 0.7 V but with low efficiency (i.e., only four electrons are transferred). 6.6.2 Challenges Both cation‐ and anion‐permeable membranes have been tested for use in DBFCs containing NaOH electrolyte. Each type of membrane leads to different chemistries within the cell. Cation permeable membranes (i.e., permeable to Na+ ions) lead to a chemical imbalance — the oxidation of 1 mol of NaBH4 transfers 8 mol of Na+ ions across the membrane and thereby increases the concentration of NaOH in the cathode region with concomitant decrease in the anode. Given the latter situation, extended operation of the cell could raise problems because BH4− ions are stable only in solutions with strong alkali concentration. With a cation membrane, it is therefore necessary to 14 Mm denotes mischmetal—an alloy of cerium, lanthanum, neodymium and other rare earth metals. Direct Liquid Fuel Cells (a) Load + – OH– OH– 6H2O BO2– 8e– 12OH– 8e– 8Na+ Anode 8OH– BH4– 2O2 4H2O Cathode OH– OH– Cation-permeable membrane (b) Load + – OH– OH– 8e– Anode 6H2O BO2– 8OH– BH4– 12OH– 8OH– OH– OH– 2O2 4H2O 8e– Cathode Anion-permeable membrane Figure 6.8 Principles of operation of a DBFC with (a) a cation‐conducting membrane and (b) an anion‐ conducting membrane. Drawn to emphasize the chemical balance of the reactions at the electrodes. (Source: From Ponce de Leon, C and Walsh, FC, 2009, Sodium borohydride fuel cells, in Garche, J, Dyer, CK, Moseley, PT, Ogumi, Z, Rand, DAJ and Scrosati, B (eds.), Encyclopedia of Electrochemical Power Sources, pp. 192–205, Elsevier, Amsterdam. Reproduced with the permission of Elsevier.) recycle NaOH from the cathode to the anode. A diagrammatic representation of the operation is given in Figure 6.8a.15 By contrast, an anion permeable membrane transfers 8 mol of OH− from cathode to anode across the membrane for each mole of borohydride that is oxidized. The chemistry using this membrane is in balance, and to maintain power production only borohydride needs to be supplied to the anode, as shown in Figure 6.8b. The cation permeable membranes that have been the most widely investigated for DBFCs are the various Nafion materials used in PEMFCs. These are stable in alkaline solution. By contrast, at present, there is no commercially available anionic membrane material that can survive the high alkalinity necessary to maintain NaBH4 in solution without it being hydrolyzed to hydrogen. It is, of course, possible to operate a DBFC with a liquid alkaline electrolyte and no ion‐conducting membrane. Sodium or potassium hydroxide solutions with concentrations 15 Ponce de Leon, C and Walsh, FC, 2009, Sodium borohydride fuel cells, in Garche, J, Dyer, CK, Moseley, PT, Ogumi, Z, Rand, DAJ and Scrosati, B (eds.), Encyclopedia of Electrochemical Power Sources, pp. 192–205, Elsevier, Amsterdam. 181 182 Fuel Cell Systems Explained between 10 and 40 wt.% work well (in which borohydride solutions of 10–30 wt.% are used as fuel). With these electrolytes, however, crossover becomes a problem but is less of a concern with cells containing anionic membranes. In anionic membrane cells, BH4− ions are prevented from reaching the cathode by the membrane. Although there is tendency, particularly at low currents, for BH4− ions to flow towards the cathode, flow of OH− ions in the opposite direction keeps the cell charges in balance. 6.7 Application of Direct Liquid Fuel Cells Currently, the DMFC is the only DLFC that can be said to have reached a stage of sustained commercialization. Several educational kits and specialized systems of other direct liquid systems, such as the DEFC marketed by Horizon Fuel Cells, can also be purchased, but these do not constitute a large market segment. State‐of‐the‐art DMFCs can achieve power densities of up to 60 mW cm−2. This is considerably lower than the performance of hydrogen fuel cells and constrains the area of application to duties where the power density can be low, but the energy density must be high. To put it another way, DMFCs are suited to services where the average power is only a few watts, but that power must be provided for a very long time — typically, for several days. Example applications include mobile phones, laptops, remote monitoring and sensing equipment, and mobile homes. In the case of consumer electronics, increasing computing power is placing heavy demands on batteries that are driving improvements in lithium‐ion technology. The best state‐of‐the‐art lithium batteries pack around 0.6 Wh mL−1 of energy. By comparing this performance with the high energy densities of fuels given in Table 6.1, it is not difficult to see why the DLFCs are also of great interest. The energy density of the liquid alone is significantly higher than that of the battery. Even when the conversion efficiency of the fuel cell is taken into account, the energy densities of fuel‐cell systems are still higher than those of batteries. Moreover, this ignores a distinguishing feature of the fuel cell, namely, that it will continue to produce power as long as fuel is supplied. The size of the fuel cell is therefore determined by the maximum power in watts that it needs to supply16 for the particular application. By contrast, the size of a rechargeable battery is governed by the watt‐hours that it needs to provide. It may take several hours to recharge a lithium battery that has been discharged to, say, a 10% state-of-charge. There is no issue with the recharging of a fuel cell. A canister of methanol that may power a laptop through a DMFC system can be replaced in less than a minute once it becomes drained. Many DMFC stacks that employ Nafion membranes have been demonstrated for portable applications by organizations such as Motorola Labs, Energy Related Devices, Samsung Advanced Institute of Technology, Los Alamos National Laboratory and the Jet Propulsion Laboratory, and by various research groups in universities. SFC Energy AG in Germany has the longest track record in the commercialization of 16 Flow batteries such as the vanadium redox battery or the zinc bromine battery, introduced in Chapter 1, are also sized according to kW rather than kWh — another reason they are often categorized as ‘fuel cells’ rather than as ‘batteries’. Direct Liquid Fuel Cells (a) (b) Figure 6.9 (a) EFOY Comfort fuel cell installed in a mobile home with a container of methanol on the left and (b) range of EFOY Pro series (800–2400 W). (Source: Reproduced with permission of Elsevier.) DMFC systems. The EFOY Comfort series of DMFCs produced by this company covers systems from 40 to 85 W nominal output that consume methanol at a rate of 0.9 L kWh−1. One of these systems located in a mobile home is shown in Figure 6.9a, together with a range of larger EFOY Pro series DMFCs in Figure 6.9b. Oorja Protonics (USA) is supplying 1‐kW DMFC systems for stationary power and materials handling applications (forklift trucks), and a 5‐kW system has been developed under a European ‘Dreamcar’ project17 for application in vehicle auxiliary power units, as shown in Figure 6.10. 17 Liu, H and Zhang, J (eds.), 2009, Electrocatalysis of Direct Methanol Fuel Cells, WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim. 183 184 Fuel Cell Systems Explained Figure 6.10 5‐kW DMFC stack developed within the framework of the European ‘Dreamcar’ project. (Source: Reproduced from Arico, AS, Baglio, V and Antonucci, V, 2009, Direct methanol fuel cells: history, status and perspectives, in Liu, H and Zhang, J (eds.), Electrocatalysis of Direct Methanol Fuel Cells, WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim. Reproduced with permission of Wiley‐VCH.) Further Reading Adamson, K‐A and Pearson, P, 2000, Hydrogen and methanol; a comparison of safety, economics, efficiencies, and emissions, Journal of Power Sources, vol. 86, pp. 548–555. Arico, AS, Baglio, V and Antonucci, V, 2009, Direct methanol fuel cells: history, status and perspectives, in Liu, H and Zhang, J (eds.), Electrocatalysis of Direct Methanol Fuel Cells, WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim. Badwal, S, Giddey, S, Kulkarni, A and Jyoti, G, 2015, Direct ethanol fuel cells for transport and stationary applications – a comprehensive review, Applied Energy, vol. 145, pp. 80–103. Choi, WC, Kim, JD and Woo, SI, 2001, Modification of proton conducting membrane for reducing fuel crossover in a direct methanol fuel cell, Journal of Power Sources, vol. 96, pp. 411–414. Dohle, H, Divisek, J and Jung, R, 2000, Process engineering of the direct methanol fuel cell, Journal of Power Sources, vol. 86, pp. 469–477. Dohle, H, Schmitz, H, Bewer, T, Mergel, J and Stolten, D, 2002, Development of a compact 500W class direct methanol fuel cell stack, Journal of Power Sources, vol. 106, pp. 313–322. Dyer, CK, 2002, Fuel cells for portable applications, Journal of Power Sources, vol. 106, pp. 31–34. Hamnett, A 1997 Mechanism and electrocatalysis in the direct methanol fuel cell, Catalysis Today, vol. 38, pp. 445–457. Direct Liquid Fuel Cells Jarvis, LP, Terrill, BA and Cygan, PJ, 1999, Fuel cell/electrochemical capacitor hybrid for intermittent high power applications, Journal of Power Sources, vol. 79, pp. 60–63. Jung, DH, Cho, S, Peck, DH, Shin, D and Kim, JJ, 2002, Performance evaluation of a Nafion/silicon oxide hybrid membrane for direct methanol fuel cell, Journal of Power Sources, vol. 106, pp. 173–177. 185 187 7 Phosphoric Acid Fuel Cells 7.1 High‐Temperature Fuel‐Cell Systems In Chapter 2, it was noted that the open‐circuit voltage for a hydrogen fuel cell decreases at higher temperatures. Indeed, above about 800°C, the theoretical maximum efficiency of a fuel cell is actually less than that of a heat engine. On this basis, one may question why fuel cells should be operated at higher temperatures? The reason is that, in many cases, high temperatures bring the following benefits that outweigh the disadvantages: ● ● ● ● Electrochemical reactions proceed more rapidly at higher temperatures, and thus voltage losses due to electrokinetic (‘activation’) effects are lower. Consequently, precious metal catalysts are often not required. The exhaust gases from the fuel‐cell stacks are sufficiently hot to facilitate the generation of hydrogen from other fuels that are readily available, e.g., natural gas. The exhaust gases are at a high temperature and therefore a valuable source of heat for buildings, processes and facilities near the fuel‐cell installation. In other words, these types of fuel cell make excellent ‘combined heat and power’ (CHP) systems. Heat extracted from exhaust gases and cooling fluids can be employed to drive turbines and generators to produce more electricity. When a turbine uses waste heat from a generator, such as a fuel cell, the scheme is known as a ‘bottoming cycle’.1 A combination of a fuel cell and heat engine allows the complementary characteristics of each to be exploited to great advantage, so that electricity can be generated with a higher level of efficiency. The phosphoric acid fuel cell (PAFC) is the most developed of the common competing types of technology that operate at temperatures above about 200°C. Many 200‐kW PAFC CHP systems are installed throughout the world at hospitals, military bases, leisure centres, offices, factories and even prisons. Their performance and behaviour are well understood. The moderate operating temperature of the PAFC requires the use of noble metal catalysts, and, as with the PEMFC, these will be poisoned by any carbon monoxide (CO) that may be in the fuel gas. A somewhat complex fuel‐processing system is required to achieve acceptably low levels of CO. 1 Conversely in a ‘topping cycle’, electricity is produced primarily from a steam turbine. The exhaust steam from the turbine is condensed, and the heat released is utilized in external applications such as district heating or water desalination. Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 188 Fuel Cell Systems Explained Some system design issues are common to all high‐temperature fuel cells and are worth addressing before examining PAFC systems in detail. These issues relate principally to the fate of the heat generated by the fuel cells, namely, whether the heat is used to reform fuels, drive engines or facilitate practical applications. Thus in evaluating the utility of PAFC, molten carbonate fuel cell (MCFC) and solid oxide fuel cell (SOFC) stacks, each of the three technologies should not be considered in isolation but rather as an integral component of a complete system that generates both heat and power. The common features for the high‐temperature fuel cells are as follows: ● ● ● ● A PAFC, MCFC or SOFC will nearly always use a fuel that will need refining or processing. A detailed review of fuel processing is given in Chapter 10, but the basics of how this operation is integrated into the fuel‐cell system and subsequently impinges on overall performance is explained in Section 7.2.1. The fuel will invariably be a mixture of hydrogen, carbon oxides and other gases. During passage of the fuel gas through the stack, hydrogen will be consumed, and the resulting reduction in its concentration in the mixture will lower the local current density. ‘Fuel utilization’ is an important operating parameter and is discussed in Section 7.2.2. The high‐temperature exhaust gases carry large amounts of heat energy that can be employed in a bottoming cycle with a turbine or other heat engines. How this combination of fuel cell and heat engine can lead to very high levels of efficiency is considered in Section 7.2.3. Heat from the exhaust gases can also serve to preheat fuel and oxidant with the aid of suitable heat-exchangers. The best use of heat within high‐temperature fuel‐cell systems is an important aspect of system design and is often referred to as ‘process integration’ by chemical engineers. To achieve high electrical and thermal efficiencies, systems need to be designed to minimize exergy loss, and designers may introduce ‘pinch technology’ to achieve the best outcome for process integration. Such system heat management aspects are covered in Section 7.2.3. 7.2 System Design 7.2.1 Fuel Processing As this topic is described in detail in Chapter 10, it is sufficient at this stage to say that the production of hydrogen from a hydrocarbon usually involves the process of ‘steam reforming’. The procedure should not be confused with the reforming of hydrocarbons as practised in the petroleum industry. In the case of methane, the steam reforming reaction (often referred to as SMR) may be written as: CH 4 H2 O 3H2 CO (7.1) A general expression to include other hydrocarbons, represented by C xHy, can be written as: CxHy xH 2 O x y H2 2 xCO (7.2) Phosphoric Acid Fuel Cells In most cases, and certainly with natural gas, the SMR is ‘endothermic’. That is, heat needs to be supplied to drive the reaction forward to produce hydrogen. Again, for virtually all fuels, the reforming has to be conducted at relatively high temperatures, usually well above about 500°C. With medium‐ and high‐temperature fuel cells, heat required by the reforming reactions can be provided, at least in part, from the fuel cell itself, i.e., from the exhaust gases. In the case of the PAFC, the heat at around 200°C has to be supplemented by burning fresh fuel gas. This requirement lowers the efficiency of the overall system so that, for the PAFC, the upper limit falls to 40–45% (LHV). By comparison, heat carried by the exhaust gases from both the MCFC and the SOFC is available at much higher temperatures. If all of the exhaust heat from the MCFC or SOFC stack is used to promote the SMR (especially when the process is performed inside the stack), the outcome is high system efficiency. Typically >50% (LHV) efficiency is achievable for MCFC or SOFC systems. For the PAFC, as with the PEMFC, the gas mixture produced by steam reforming must be further processed to reduce the concentration of CO in the mixture. The ‘water-gas shift’ reaction (usually abbreviated as the ‘shift reaction’), whereby CO is converted to carbon dioxide, is employed, namely: CO H2 O CO2 H2 (7.3) This process is generally carried out in two stages (see Section 10.4.9, Chapter 10) in reactors that operate at different temperatures— to achieve levels of CO that are sufficiently low to be acceptable for PAFC stacks. A further complication is that fuels such as natural gas nearly always contain small amounts of sulfur or sulfur‐containing compounds. Sulfur is a well‐known catalyst poison, i.e., it will absorb preferentially on the catalyst metal and reduce the activity for both steam reforming and shift reaction. In a similar manner, sulfur will also deactivate the electrode catalysts of all types of fuel cell. Consequently, it is essential that this impurity is removed from the fuel gas before it is fed to the reformer or stack. Desulfurization is well established industrially and is featured in many hydrocarbon processes, not just for fuel cells; the process is discussed further in Section 10.4.2, Chapter 10. 7.2.2 Fuel Utilization The issue of fuel utilization arises whenever the hydrogen for a fuel cell is supplied as one component of a reactant gas or becomes a component of the gas mixture due to internal reforming. Consider a purified fuel gas for a PAFC containing hydrogen, carbon dioxide and water vapour. As this gas mixture flows through a cell, the hydrogen is consumed electrochemically, and CO2 and H2O simply pass through without reacting. The result is that the partial pressure of hydrogen falls as the fuel gas travels from cell inlet to outlet. A similar effect is observed with oxygen in the air on the cathode side of the cell. The effects of pressure and gas concentration on the open‐circuit voltage of a fuel cell have been examined in Section 2.5, Chapter 2. The Nernst relationship, introduced as equation (2.36), relates the open‐circuit voltage, Vr, and the partial pressures of hydrogen, oxygen and steam as follows: 189 190 Fuel Cell Systems Explained Vr Vr PH PO 2 2 . RT   P ln P 2F PH O 1 2 (2.36) 2 P If only the partial pressure of hydrogen is considered and the pressure changes from Pin to Pout, then the change in cell voltage is expressed by: V P RT ln out 2F Pin (7.4) Given that the partial pressure of hydrogen in the fuel gas is falling due to the reaction taking place within the cell, Pout is always less than Pin, and thus ΔV will always be negative. The open‐circuit cell voltage, and therefore the voltage under load, could be expected to fall on moving from inlet to outlet. This clearly cannot be the case since bipolar plates are good electronic conductors, and therefore the voltage difference between the two electrodes of a fuel cell must be the same over the whole area of the cell. The local cell voltage under load measured at the fuel inlet must be the same as that measured at the fuel outlet. For this situation to occur, the local current density must be lower at the outlet of the cell than at the inlet to accommodate the fact that less hydrogen is available to react at the outlet of the cell compared with the inlet. The above situation is assumed to hold especially for the SOFC and the MCFC in which the activation overpotentials at each electrode are relatively small and the internal ohmic losses are taken to be uniform throughout the cell. Recent research has found that this is not necessarily the case for PEMFCs. Careful in situ measurements have shown variations in both current density and local impedance for these fuel cells according to position in the cell. Example data for the two parameters in a cross‐flow PEMFC are given in Figure 7.1. Current density is highest towards the corner of fuel and oxidant inlets and also high towards the corner where both fuel and oxidant exit the cell— confirmed as segments 17 and 33 in Figure 7.1a. By contrast, the AC impedance spectra for these two segments given in Figure 7.1b are clearly very different, showing that there is difference in impedance depending on cell position. Indeed, in this example, the segment in the centre of the cell (25) exhibits a similar impedance spectrum to that at the cell outlet (33). Equation (7.4) shows that ΔV is also dependent on temperature, which means that the expected open‐circuit voltage drop, and hence the reduction in current density as a result of the falling partial pressure of hydrogen through the anode, will be greater for fuel cells operating at higher temperatures. At the cathode of the fuel cell, the partial pressure of oxygen in the air will also reduce as it passes through the cell. This is less of a problem in practical terms since the cell voltage is dependent on the square root of the partial pressure of oxygen, as indicated by equation (2.36). The influence of fuel and oxygen utilization on the open‐circuit voltage, Vr, is illustrated in Figure 7.2. The uppermost dashed line shows the voltage of a typical hydrogen fuel cell, which is operating at 100 kPa and supplied with pure hydrogen and oxygen. The lower dashed line is for a cell using air at the cathode and a Phosphoric Acid Fuel Cells (a) λair = 3 0.050 Seg.33 H2 inlet Seg.25 0.150 Seg.17 Acm–2 0.100 0.200 0.250 0.300 Air inlet (b) 0.10 λair = 3 Seg. 17 Seg. 25 Seg. 33 0.05 Im(Z)/Ωcm2 0.00 –0.05 –0.10 –0.15 –0.20 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Re(Z)/Ωcm2 Figure 7.1 (a) Current density distribution and (b) electrochemical impedance spectroscopy Nyquist plot at selected cell segments for a PEMFC at 200°C; cathode stoichiometry of λ = 3. (Source: Bergmann, A, Kurz, T, Gerteisen, D and Hebling, C, 2010, Spatially resolved impedance spectroscopy in PEM fuel cells up to 200°C, in: Stolten, D. and Grube, T. (eds.), 18th World Hydrogen Energy Conference, WHEC 2010, Parallel Sessions Book 1: Fuel Cell Basics / Fuel Infrastructures, Proceedings of the WHEC, May 16–21.Reproduced with permission of Forschungszentrum J¸lich GmbH, Zentralbibliothek, Verlag.) mixture of four parts hydrogen to one part carbon dioxide at the anode to simulate the gas mixture that would be obtained from reformed methane. The upper and lower solid lines are plots of the ‘open‐circuit exit voltage’ for 80 and 90% fuel utilization, respectively, with 50% air consumption in both cases. The data are for a situation in which both the air and fuel are flowing in the same direction (co‐flow). Under such conditions, the drop in the open‐circuit voltage is significant for a co‐flow configuration and, as expected, increases with both temperature and fuel utilization. Sometimes, the current density distribution through a high‐temperature cell can be made more uniform by feeding the air and fuel through the cell in opposite directions, i.e., a counter‐flow operation. With this arrangement, the fuel outlet region of the cell has the highest partial pressure of oxygen. It should be noted, however, that the particular 191 Fuel Cell Systems Explained 1.2 Pure reactants 1.1 Open-circuit voltage / V 192 Entry voltage for air and reformed methane fuel 1.0 0.9 Voltage at exit, 80% fuel utilization 0.8 Open-circuit voltages at cell exit Voltage at exit, 90% fuel utilization 0.7 0.6 0 200 400 600 800 1000 Temperature/°C Figure 7.2 Open‐circuit voltage of a hydrogen fuel cell under different conditions. The two curves for the voltage at the exit show how the voltage depends on fuel utilization and temperature. In both cases, the oxygen utilization is 50%. configuration adopted for a fuel cell is also dependent on how the fuel and air flows influence the temperature distribution within the stack. This, in turn, is influenced by the method adopted for stack cooling. It should be remembered that steam is produced at the anode in both an MCFC and an SOFC rather than at the cathode as in a PEMFC and a PAFC. In other words, in an MCFC or SOFC, the hydrogen in the fuel is essentially replaced, as it is consumed, by steam. Accordingly, if the partial pressure of the hydrogen decreases as it passes through the fuel‐cell anode compartment, the partial pressure of steam will increase, and, as previously indicated by equation (2.36), the outcome will be a fall in the open‐circuit voltage. Unfortunately, the effect of this behaviour is difficult to model because, for example, some of the steam may be employed in internal fuel reforming. The situation is therefore liable, in practice, to be worse than Figure 7.2 would indicate. It can be concluded that, in the case of a reformed fuel containing carbon dioxide or when internal reforming is applied, it is impossible to consume all the hydrogen in the fuel‐cell stack. Some of the hydrogen must therefore pass straight through the cell unconverted, to be used later to provide energy to process the fuel or to be burnt to increase the heat energy available for further operations, as discussed in Section 7.2.3. In the early days of fuel‐cell development, optimum values of both fuel and air utilization in PAFCs, MCFCs and SOFCs were determined experimentally. The task has been made easier in recent years by the advent of computer models that can simulate entire fuel‐cell systems. 7.2.3 Heat‐Exchangers Not surprisingly, there are challenges in the way in which the various components of a fuel‐cell system are integrated. The situation applies to all types of fuel cell, but is Phosphoric Acid Fuel Cells particularly notable for PAFC, MCFC and SOFC systems where several of the balance‐ of‐plant items operate at high temperatures. Examples of such items are the desulfurizer, reformer reactor, shift reactors, heat-exchangers, recycle compressors and ejectors. In some of these components, heat may be generated or consumed. The challenge for the system designer is to arrange the various components in a manner that minimizes heat losses to the external environment and at the same time ensures that heat is utilized within the system in the best possible way (i.e., by avoiding unnecessary losses). 7.2.3.1 Designs In any fuel‐cell system, heat is required by several process streams, e.g., for preheating the fuel fed to the reformer, for running the reformer itself and for raising and superheating steam. There are also areas that need to be cooled, e.g., the fuel‐cell stack and, in the case of a PEMFC, also the outlet of the shift reactor(s). Heat transfer from one process stream to another is carried out by means of a heat-exchanger. The gas (or liquid) to be heated passes through pipework that is heated by the gas (or liquid) to be cooled. A commonly used symbol for a heat-exchanger is shown in Figure 7.3. When the exit fluids from a process are employed to heat incoming fluids, the heat-exchanger is often called a ‘recuperator’. There are several types of heat-exchanger that include the shell‐and‐tube, plate‐fin and printed‐circuit designs. The selection for any particular application will be governed by the temperature range of operation, the fluids involved (e.g., liquid or gas phase), the fluid throughput and the cost. The materials of construction, the method of fabrication and the heat-transfer area required for the application determine the cost of the exchanger. 7.2.3.2 Exergy Analysis Exergy is the maximum amount of work that can be done by a system as it approaches thermodynamic equilibrium with its surroundings by a sequence of reversible processes. Consequently, the exergy of a system can be considered as a measure of its ‘distance’ from equilibrium with the surroundings. When the system and its surroundings are in equilibrium, the exergy of the system is zero. Therefore, thermal exergy is simply ‘available heat’. Potential energy and Fluid losing heat kinetic energy, as classically defined, are also forms of exergy as is the Gibbs free energy of combustion of a fuel (with the sign changed). Energy is conserved in all processes (first law of thermodynamics) whereas exergy is conserved in processes Fluid gaining that are reversible. Real processes are of heat course irreversible, so that exergy is always partly consumed to give enthalpy. In a power conversion device such as a fuel‐cell system where a chemical reaction is continually taking place, the state of the system cannot be defined merely from the Figure 7.3 Common symbol for a heattemperature, the volume and the pressure. exchanger. The fluid to be heated passes through In recognition of this, Gibbs defined a the zigzag element. 193 194 Fuel Cell Systems Explained property, μ, known as the chemical potential of a substance or system. When there is an energy change in a system that also involves a chemical reaction, the change in Gibbs free energy can be represented formally by: G V P S T i (7.5) ni where V, S and ni denote extensive parameters of the system (volume, entropy and number of moles of different chemical components, respectively); P, T and μi are intensive parameters of the environment (pressure, temperature and chemical potential of the components). It can be shown2 that the change in exergy (ΔB) of a system in going from an initial state to a reference state (subscript o) is given by B S T To V P Po ni i o (7.6) Clearly the higher the temperature, the greater is the exergy of the system. Consider, for instance, the case where a PEMFC and an SOFC system have the same power output and efficiency. The heat, i.e., the enthalpy content, of the exhaust streams from both systems will be the same. The heat that is liberated in a PEMFC is at a temperature of around 80°C and thereby is of limited value both within the system and for external applications. For the latter, it may be applied to space heating in buildings or possibly integration with an absorption cooling system to provide air cooling. In the design of a PEMFC system, care should be taken to ensure that heat is used efficiently so that there is maximum available exhaust heat. By contrast, the heat produced by the SOFC will be at a much higher temperature and therefore will have a higher exergy and, consequently, be more valuable for further utilization than that from a PEMFC. For example, the exhaust heat from an SOFC could enable the powering of a steam turbine in a bottoming cycle. All fuel‐cell systems should therefore be configured in such a way that exergy loss is minimized. This is especially important for PEMFC and PAFC systems, which operate at moderate temperatures, where any heat utilized inefficiently within the system will have a more deleterious effect on the amount of exhaust heat that is available externally. 7.2.3.3 Pinch Analysis Pinch analysis, or pinch technology, is a methodology that can be applied to fuel‐cell systems for deciding the optimum arrangement of heat-exchangers and other units so as to minimize loss of exergy. It was originally designed by chemical engineers as a tool for defining energy‐saving options, particularly in heat-exchanger networks, but has since been applied in the development of fuel‐cell systems. The concept is fairly straightforward but for complex systems, sophisticated computer models are required. The procedure for pinch technology is broadly as follows. In any fuel‐cell plant, there will be process streams that require heating (cold streams) and cooling (hot streams), irrespective of where heat-exchangers are located. The first stage in system design is therefore to establish the basic chemical‐processing requirements and to produce a configuration that shows and defines all of the cold 2 Dincer, I and Cengel, YA, 2001, Energy, entropy and exergy concepts and their roles in thermal engineering, Entropy, vol. 3, pp. 116–149. Phosphoric Acid Fuel Cells 700 600 Cathode feed Cathode effluent Cathode feed Temperature/°C 500 400 Combustion air preheat Process + motive ss Flue gas 300 Process ss 200 Recycle gas preheat Feed + fuel gas 100 preheat Bfw preheat Wet flue gas Dearation preheat 0 0 2 4 6 8 10 12 14 16 Enthalpy/kW x 100 18 20 22 24 26 Figure 7.4 Hot and cold heating plots for a conceptual 3.25‐MW MCFC system with high‐pressure steam generation. ss, steam superheat; Bfw, boiler feed water. and hot streams. Calculation of the heat and mass balances enables the engineer to determine the enthalpy content of each stream. Heating and cooling curves can then be produced from knowledge of the required temperatures of each stream; examples for an MCFC system are given in Figure 7.4. The individual cooling and heating curves are then summed together to make two composite plots—one shows the total heating required by all of the streams that need heating and the other shows the total cooling required by the streams that require cooling. The composite plots obtained by summing the curves of Figure 7.4 are given in Figure 7.5. The composite plots are brought together by sliding along the x‐axis and where they ‘pinch’ together with a minimum temperature difference of, for example, 50°C; the temperature is noted. This so‐called pinch point defines the target for optimum process design, since in a real system heat cannot be transferred from above or below this pinch temperature. Once the pinch point is known, heat-exchangers can be positioned in such a way that maximum transfer of heat is achieved between units that need heating and those that require cooling. In some fuel‐cell systems, a pinch temperature is not found, in which case the problem becomes one of defining an upper limit of temperature for the system. Either way, pinch technology provides an excellent method for system optimization. Many computer models are available for calculating the heat and material flows around the system and for calculating the pinch‐point temperature.3 Once such an analysis has been undertaken, the required heat-exchangers and reactors can be designed. 3 Aspen Technology Inc., for example, produces a suite of software packages including Aspen Plus® and AspenTech Pinch™ that are widely used for process and system design. 195 Fuel Cell Systems Explained 800 Cold stream Hot stream 700 600 Temperature/°C 196 500 Pinch point 400 300 200 100 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Enthalpy/kW x 100 Figure 7.5 Composite curves derived from data presented in Figure 7.4. Of course, other considerations need to be taken into account when designing a fuel‐ cell plant, for example, the choice of materials for the balance‐of‐plant components and the mechanical layout of the system. Configurations are drawn up initially as a process flow diagram (PFD) that shows the logical arrangement of the fuel‐cell stack and the associated balance‐of‐plant components. Many other techniques that are available to the system designer are outside the scope of this book. Nonetheless, it is hoped that this consideration of common elements of high‐temperature fuel‐cell systems will be a good starting point. 7.3 Principles of Operation The PAFC works in a similar fashion to the PEMFC, as described in Chapter 4. A proton‐conducting electrolyte is employed, and the reactions that occur on the anode and cathode are those given in Figure 1.3, Chapter 1. The electrochemical reactions take place on highly dispersed electrocatalyst particles that are supported on carbon black. As with the PEMFC, the PAFC uses platinum (Pt) alloys as the catalyst at both electrodes. The electrolyte is an inorganic acid, concentrated phosphoric acid (100 wt.%). 7.3.1 Electrolyte Phosphoric acid (H3PO4) is the only common inorganic acid that, above 150°C, has satisfactory thermal, chemical and electrochemical stability and sufficiently low volatility to be considered as an electrolyte for fuel cells. Most importantly, phosphoric acid is tolerant to carbon dioxide in the fuel and oxidant, unlike the electrolyte solution in the alkaline fuel cell. The acid was therefore chosen by United Technologies Phosphoric Acid Fuel Cells (a US company, later becoming the spin‐off ONSI Corporation) in the 1970s as the preferred electrolyte for fuel‐cell power plants in terrestrial applications. Phosphoric acid is a colourless, viscous, hygroscopic liquid. It is contained in the PAFC by capillary action (it has a contact angle >90°) within the pores of a matrix made of silicon carbide (SiC) particles of about 1 µm that are held together with a small amount of PTFE. Pure 100% phosphoric acid, which has been used in fuel cells since the early 1980s, has a freezing point of 42°C. Therefore, to avoid the development of stresses due to freezing and rethawing, PAFC stacks are usually maintained above this temperature once they have been commissioned. Although the vapour pressure is low, some acid is lost during normal fuel‐cell duty over long periods at high temperature; the amount depends on the operating conditions, especially gas flow velocities and current density. It is therefore necessary either to replenish electrolyte during service or to ensure that at the start of operation, there is sufficient reserve of acid in the cell to sustain the projected lifetime. The SiC matrix is thin enough (0.1–0.2 mm) to keep ohmic losses at a reasonably low level (i.e., to give high cell voltages) while having adequate mechanical strength and the ability to prevent crossover of reactant gases from one side of the cell to the other. This latter property is a challenge for all fuel cells with liquid‐based electrolytes. Under some conditions, the pressure difference between the anode and the cathode can rise considerably, as determined by the design of the system. Loss of phosphoric acid from the fuel cell can occur through volume change or evaporation and transfer by electrochemical pumping. During operation, the volume of phosphoric acid electrolyte expands and contracts according to temperature, pressure, changes in load and the humidity of the reacting gases. A similar effect occurs with the electrolyte in the MCFC, as will be discussed in Section 8.2, Chapter 8. To replace electrolyte that may be lost through expansion or volume change, the porous, carbon‐ribbed, flow‐field plates in the PAFC act as reservoirs for excess electrolyte. The porosity and pore‐size distribution of these plates are deliberately chosen to accommodate any volume changes of the electrolyte. Loss of electrolyte through evaporation is minimized by keeping the stack operating temperature reasonably low, but even at 200°C there is some escape of electrolyte through the air channels. In practical PAFC stacks, evaporative loss is curtailed by ensuring that the cathode exit gases pass through a cool condensation zone at the edge of the cell. With additional cooling, the zone is maintained between 160 and 180°C, which is sufficiently low to condense out most of the electrolyte vapour. Electrochemical pumping is a phenomenon that occurs per se with fuel cells that employ any liquid electrolyte or dissolved electrolyte. In the case of the PAFC, the electrolyte dissociates into positively charged cations (H+) and negatively charged anions (H2PO4−). During operation, the protons move from anode to cathode, whereas the anions move in the other direction. Phosphate anions therefore build up at the anode and can react with hydrogen to form phosphoric acid, thereby leading to an accumulation of electrolyte at the cathode of each cell. Not surprisingly, a similar effect can occur in the MCFC where pumping of alkali metal ions from the anode to cathode causes a build‐up of carbonate electrolyte at the cathode of the fuel cell. Electrochemical pumping can be minimized by the optimization of the porosity and pore‐size distribution of the porous components, i.e., the ribbed plates and the electrodes in the case of the PAFC and the MCFC, respectively. Degradation of the separator plate in the PAFC can lead to migration of electrolyte from one cell to the next with consequent 197 198 Fuel Cell Systems Explained catastrophic loss of stack voltage. Migration of acid can also occur through breakdown of the manifold seals, again resulting in serious stack failure. 7.3.2 Electrodes and Catalysts Like the PEMFC, the PAFC has gas‐diffusion electrodes, in which the catalyst is platinum supported on carbon black. This catalyst has replaced the PTFE‐bonded platinum black that was used in the first PAFC stacks that were built in the mid‐1960s. In a modern PAFC, the catalyst layers contain 30–50 wt.% PTFE to act as a binder for the creation of a porous structure. Meanwhile, the carbon catalyst support provides the following functions similar to those fulfilled by the support in PEMFC catalysts: ● ● ● To disperse the platinum to ensure good utilization. To provide micropores in the electrode for maximum gas diffusion to the catalyst and the electrode–electrolyte interface. To increase the electrical conductivity of the catalyst layer. The activity of the PAFC catalysts in both positive and negative electrodes depends on the nature of the Pt, i.e., its crystallite size and specific surface area. In state‐of‐the‐art stacks, the loadings are currently about 0.10 and about 0.50 mg Pt cm−2 in the anode and cathode, respectively. The low loadings are, in part, the result of advances in nanotechnology—the ability to prepare small crystallite sizes of around 2 nm in diameter with high specific surface areas of up to 100 m2 g−1; see Figure 1.6, Chapter 1. Each catalyst layer in the PAFC is usually bonded to a thin gas‐diffusion layer (GDL) or substrate made of carbon paper. A typical GDL used in PAFCs has carbon fibres of 10 mm in length that are embedded in a graphitic resin. The paper has an initial porosity of about 90%, which is reduced to about 60% by impregnation with 40 wt.% PTFE. The resulting wet‐proof carbon paper contains macropores of 3–50 µm diameter (median pore diameter of about 12.5 µm), which can serve as a reservoir for phosphoric acid, and micropores with a median pore diameter of about 3.4 nm to permit gas permeability. The composite structure of a carbon black + PTFE layer on a carbon paper substrate forms a stable, three‐phase interface in the fuel cell, with electrolyte on the electrocatalyst side and the reactant gas environment on the other (carbon paper) side. The choice of carbon for the catalyst layer is important, as is the method of dispersing platinum, and much of the expertise in these two areas is proprietary to the manufacturers of fuel cells. Through many decades of proof of operation in the field, the PAFC has demonstrated good long‐term reliability. For instance, stack operating times can extend well beyond 40 000 h before decay in electrode performance has reached an unacceptably low level. Phosphoric acid electrodes can be poisoned by carbon monoxide although the tolerance is significantly greater than for PEMFC catalysts. Thus compared with the anode catalyst of a PEMFC, which can accept up to only a few ppm of CO in the fuel gas, the PAFC anode catalyst can tolerate typically up to about 2 mol.% at 200°C. In addition to sulfur, which also poisons the catalyst, small amounts of ammonia and chlorides, even at the ppm level in the fuel, degrade cell performance. These do not inhibit the platinum catalyst per se but react with the phosphoric acid to form salts that decrease the acidity of the electrolyte and can precipitate and block the porous electrodes. To avoid Phosphoric Acid Fuel Cells unacceptable performance losses, the concentration of ammonium phosphate ((NH4)H2PO4) in the host electrolyte must be kept below 0.2 mol.%. To achieve this requirement, an ammonia trap is usually inserted between the outlet of the fuel processor and the inlet of the anodes to prevent ammonia from entering the stacks. The PAFC catalysts can also degrade through the agglomeration of platinum particles. During operation, the particles have a tendency to migrate to the surface of carbon and combine to form larger particles, thereby decreasing the available active surface area. The rate of this type of degradation depends mainly on the operating temperature. An unusual difficulty is that corrosion of carbon becomes a problem at high cell voltages (above about 0.8 V). For practical applications, low current densities with cell voltages above 0.8 V and hot idling at open circuit are therefore best avoided with the PAFC. 7.3.3 Stack Construction The PAFC stack consists of a repeating arrangement of a ribbed bipolar plate, the anode, the electrolyte matrix and the cathode. In a similar manner to that described for the PEMFC, the ribbed bipolar plate serves to separate the individual cells and to connect them electrically in series while providing the gas supply to the anode and cathode, respectively, as shown in Figure 1.9. As discussed previously, there is an additional requirement in the PAFC to build in a reservoir of phosphoric acid. This feature can be located in the electrode substrates or GDL. When made from porous graphitic carbon, the ribbed bipolar plates can also serve as a reservoir for excess phosphoric acid. This capability is realized in a modern PAFC stack by building a ‘multilayer’ bipolar plate, in which a graphitic flow‐field plate is bonded on either side of a thin non‐porous carbon layer which forms the gas barrier between adjacent cells. A generic arrangement is illustrated in Figure 7.6 and the configuration deployed in the water‐cooled PAFC stacks produced by International Fuel Cells (IFC), a subsidiary of United Technologies Corporation, is shown in Figure 7.7. In the IFC design, the catalyst layers are deposited onto porous carbon paper substrates for both cathodes and anodes. These are in turn aec aec Standard one piece bipolar plate Two-cell stack Figure 7.6 Cell arrangement using ribbed substrates (bipolar plates): a, anode; e, electrolyte; and c, cathode layers. 199 200 Fuel Cell Systems Explained Oxidant gas channels Gas-tight metal plate Cathode catalyst layer Electrolyte matrix Anode catalyst layer Fuel gas channels Cooling water channel Single cell Gas flow channels Figure 7.7 Schematic basic design of water‐cooled PAFC stacks produced by International Fuel Cells (IFC). The figure shows a cross section through two cells of the stack. (Source: Adapted from Kurzweil, P, 2003, Fuel Cell Technology, Vieweg, Teubner, Wiesbaden.) bonded, using a polymer that decomposes on heating, to the flow‐field plates into which the channels are pressed. The resulting ‘multilayer’ bipolar plate has the following advantages over previously adopted stack configurations: ● ● ● The surfaces between catalyst layer and GDL substrate promote uniform gas diffusion to the electrode. The plate is amenable to a continuous manufacturing process since the ribs on each substrate run in only one direction; cross‐flow configuration, if required, can be easily accommodated. The substrate and flow‐field plates can act as a reservoir for phosphoric acid and thereby offer a means to increase the lifetime of the stack. A typical PAFC stack may contain 50 or more cells connected in series to obtain the practical voltage level required. 7.3.4 Stack Cooling and Manifolding Phosphoric acid fuel‐cell stacks can be cooled by liquid (usually water or antifreeze solution), a dielectric (oil), or air. Cooling channels or pipes can be located between groups of cells in the stack. As shown in Figure 7.7, cooling can also most easily be achieved by circulating the cooling fluid between the gas‐tight components of the bipolar plates. Note that it is not necessary for the coolant to flow between every cell— usually, between about every fifth cell is sufficient. Air‐cooled PAFC stacks have also been produced and offer the advantages of simplicity, reliability and low cost. The channels in air‐cooled stacks are, however, large and this imposes a limit to the practical size of stacks. Better heat removal is achieved with liquids that require only narrow Phosphoric Acid Fuel Cells channels, leading to more compact stack design. Conversely, narrow channels may be complex to design and costly to fabricate. Whereas small PAFC stacks may be cooled with air, stacks above about 50 kW invariably employ either boiling or pressurized water as the coolant. With the former method, the heat of vaporization of water is used to remove the heat from the cells. Since the average cell temperature is around 180–200°C, the temperature of the cooling water will be about 150–180°C. Reasonably uniform temperature in the stack can be attained with boiling water and thereby leads to increased cell efficiency. If the alternative of pressurized water is employed, the heat is only removed from the stack by the heat capacity of the liquid water, so a greater flow of coolant is required. Nevertheless, pressurized water is easier to control and, while not so efficient as boiling water, it provides a better overall performance than that obtained with oil (dielectric) or air as the cooling medium. The main disadvantage of water cooling is that water treatment is necessary to prevent the corrosion of cooling pipes and the formation of blockages in the cooling loops. The water quality required is similar to that demanded by boilers in conventional thermal power stations. Although not difficult to achieve with ion‐exchange resins, such water treatment adds to the capital cost of PAFC systems. All PAFC stacks are fitted with manifolds that are usually attached to the outside of the stacks; these are so‐called external manifolds. (It will be noted in Section 8.4.1, Chapter 8 that an alternative ‘internal manifold’ arrangement is preferred by some developers of MCFC systems.) Respective inlet and outlet manifold systems enable fuel gas and oxidant to be circulated through each cell of a particular stack. To minimize temperature variations within the stack and thereby ensure long lifetimes, the inlet manifold for the fuel gas is carefully designed to provide a uniform supply to each cell. Often a stack is made of several sub‐stacks, in which the plates are mounted horizontally on top of each other with separate fuel supplies to each sub‐stack.4 If the fuel‐cell stack is to be operated at high pressure, the whole stack assembly has to be located within a vessel that is filled with nitrogen gas at a pressure slightly above that of the reactants. 7.4 Performance The performance (voltage–current) curve for a typical PAFC is similar to that shown in Figure 3.1 for medium to low‐temperature cells, although the current density of PAFC stacks is usually in the range 150–400 mA cm−2. When operating at atmospheric pressure, the output gives a cell voltage of between 600 and 800 mV. As with the PEMFC, the major voltage losses occur at the cathode, and the overpotential is greater with air (typically 560 mV at 300 mA cm−2) than with pure oxygen (typically 480 mV at 300 mA cm−2) because of the dilution of oxygen with nitrogen in the former. The voltage losses at the anode are very low (ca. 4 mV per 100 mA cm−2) with pure hydrogen, which increases 4 Reference will often be found in the fuel‐cell industry to ‘sub‐stack’ and ‘short stack’. These two terms each describe a small group of full‐size cells (i.e. the same cell area as used in completely assembled stacks). Manufacturers routinely carry out lifetime tests on ‘short stacks’ to avoid the cost of manufacturing a full stack of cells. The performances of a short stack and a full stack are expected to be very similar. 201 202 Fuel Cell Systems Explained when carbon monoxide is present in the fuel gas. The ohmic loss in PAFCs is also relatively small, namely, about 12 mV per 100 mA cm−2. 7.4.1 Operating Pressure For any type of fuel cell, performance is a function of pressure, temperature, and composition and utilization of the reactant gas. It is well known that an increase in the operating pressure boosts the performance of the PAFC and, indeed, all other candidate fuel cells. The increase in cell voltage resulting from a change in system pressure from P1 to P2 is given by the formula (see Section 2.5.4, Chapter 2): V P RT ln 2 4F P1 (2.44) The change in voltage is not, however, the only benefit of a higher pressure. At the operating temperature of the PAFC, raising the pressure also decreases the activation overpotential at the cathode, due to the concomitant increase in the partial pressure of both oxygen and product water. If the partial pressure of water is allowed to increase, a lower phosphoric acid concentration will cause a slight enhancement of the ionic conductivity that, in turn, will bring about a higher exchange-current density. This important beneficial effect, which has been discussed in detail in Section 3.4.2, Chapter 3, promotes further reduction of the activation overpotential, and the greater conductivity lessens the ohmic losses. The end result is that, for the PAFC, the increase in voltage with pressure is much higher than what is predicted by equation (2.44). From experimental data collected over some period, the US Department of Energy Fuel Cell Handbook5 suggests that the formula: V 63.5 ln P2 P1 mV (7.7) is a more reasonable approximation for a temperature range of 177°C < T < 218°C and a pressure range of 0.1 MPa < P < 1.0 MPa. 7.4.2 Operating Temperature The reversible voltage of a hydrogen fuel cell decreases as the temperature increases; see Section 2.3, Chapter 2. Over the possible temperature range of the PAFC, the effect is a decrease of 0.27 mV per °C under standard conditions (at which the product of hydrogen oxidation is water vapour). On the other hand, an increase in temperature has a beneficial effect on cell performance because activation overpotential, mass-transfer overpotential and ohmic losses are all reduced, as discussed in Chapter 3. The kinetics for the reduction of oxygen on platinum also improves as the cell temperature increases. The abovementioned Fuel Cell Handbook states that for a PAFC running on air and 5 EG&G Technical Services, Inc., 2004, Fuel Cell Handbook (7th Edition), US Department of Energy, Office of Fossil Energy, National Energy Technology Laboratory, P.O. Box 880, Morgantown, West Virginia 26507‐0880. Phosphoric Acid Fuel Cells 750 Cell voltage/mV 700 H2 650 H2 + H2S 600 H2 + CO SCG 550 190 170 210 230 Temperature/°C Figure 7.8 Effect of temperature on PAFC cell voltage for different fuels: H2, H2 + 20 ppm H2S, H2 + CO and simulated coal gas (SCG). (Source: Reproduced from Jalan, V, Poirier, J, Desai, M and Morrisean, B, 1990, Development of CO and H2S tolerant PAFC anode catalysts, Proceedings of the Second Annual Fuel Cell Contractors Review Meeting, 2–3 May 1990, Morgantown, WV.) pure hydrogen at a mid-range operating load (~250 mA cm−2), the voltage gain (ΔVT) with increasing temperature is given by: VT 1.15 T2 T1 mV (7.8) The data collected to derive this equation suggests that it is reasonably valid for a temperature range of 18°C < T < 25°C. The relationship shows that each degree increase in cell temperature should produce a performance increase of 1.15 mV. Other data indicate that the coefficient may actually be in the range of 0.55–0.75, rather than 1.15. Although temperature has only a minimal effect on the hydrogen oxidation reaction at the anode, this operational parameter is important in terms of anode poisoning. Increasing the cell temperature results in increasing anode tolerance to carbon monoxide, as demonstrated in Figure 7.8. The benefit is a result of reduced adsorption of the gas. A strong temperature effect for simulated coal gas (SCG) is also seen in Figure 7.8. 7.4.3 Effects of Fuel and Oxidant Composition As mentioned in Section 7.2, fuel and oxidant utilizations are important operating parameters for the PAFC and indeed for all types of fuel cell. In a fuel gas that is obtained, for example, by steam reforming of natural gas, the CO2 and unreacted hydrocarbons (e.g., CH4) are electrochemically inert and act as diluents. Because the anode reaction is nearly reversible, the fuel composition and hydrogen utilization generally do not strongly influence cell performance. Cell voltage will, however, be influenced by a change in the partial pressure of hydrogen that can result from a change in either the composition or the utilization of the fuel. This effect can be described by a relationship similar to equation (7.7), namely: V 55 ln P2 P1 mV (7.9) 203 204 Fuel Cell Systems Explained On the cathode side, the use of air with ~21 vol.% oxygen instead of pure oxygen results in a decrease in the current density by a factor of about three at constant electrode potential. The overpotential at the cathode increases with an increasing consumption of oxygen. 7.4.4 Effects of Carbon Monoxide and Sulfur It has been stated already that platinum in the anode catalyst of the PAFC may be poisoned by carbon monoxide in the fuel gas. At low concentrations of CO, absorption on the anode electrocatalyst is reversible, and CO will be desorbed if the temperature is raised. The methods undertaken to limit the CO concentration are discussed in Section 10.4.11, Chapter 10. Sulfur in the fuel stream, usually present as hydrogen sulfide (H2S), will similarly poison the anode of a PAFC. State‐of‐the‐art PAFC stacks are able to tolerate up to 50 ppm of sulfur in the fuel. Sulfur poisoning does not affect the cathode, and moderately poisoned anodes can be reactivated by increasing the temperature. 7.5 Technological Developments Until recently, the PAFC was the only fuel‐cell technology that could be said to be available commercially. Systems are ready to meet market specifications and are supplied with guarantees. Many of the systems built by IFC have now run for several years, and so there is a wealth of operating experience from which developers and endusers can draw. One important aspect that has arisen from field trials of the early PAFC plants is the reliability of the stack and the quality of power produced by the systems. These dual attributes have led to systems being preferred for so‐called ‘premium power’ applications, such as in banks, hospitals and computing facilities. Worldwide, PAFC plants with a total installed capacity in excess of 65 MW have been tested, are being tested, or are being fabricated. Most of the systems are in the capacity range of 50–200 kW, but large versions of 1 and 5 MW have also been constructed. The largest plant operated to date has been that built by IFC and Toshiba for Tokyo Electric Power. This facility can generate 11 MW of grid‐quality AC power. Efforts in the United States and Japan are now concentrating on the improvement of PAFCs for stationary dispersed power and on‐site cogeneration (CHP). The major industrial developers are Doosan Fuel Cell America Inc. in the United States and Fuji Electric, Toshiba and Mitsubishi Electric Corporation in Japan. Although the PAFC has now reached a level of maturity where customer confidence can be guaranteed, the technology is still too costly to be economic compared with alternative power‐generation systems, except in the niche premium power applications referred to in the previous text. There is a need to increase the power density of the cells and reduce capital costs; both issues are inextricably linked. System optimization is also a key issue. Much of the recent development in the technology is proprietary, but the following overview gives an indication of progress made during the past few years. During the early 1990s, the goal of the research and development of PAFCs in the United States was to design and demonstrate a large stack with a power density of 0.188 W cm−2, a practical life of 40 000 h and a stack cost of less than US$400 per kW. Phosphoric Acid Fuel Cells A conceptual design of an improved technology stack operating at 820 kPa and 200°C was produced. The stack would be composed of 355 × 1 m2 cells to produce over 1 MW of DC power in the same physical envelope as the unit 670‐kW stack for the 11‐MW PAFC plant built for Tokyo Electric Power. The improvements made to the design were tested in single cells, as well as in sub‐scale and full‐size short stacks. The results of these tests were outstanding. The power density goal was exceeded; namely, 0.323 W cm−2 was achieved in single cells when operating at 645 mA cm−2 and up to 0.66 V per cell. A cell performance of 0.307 W cm−2 was obtained from stacks, with an average of 0.71 V per cell at 431 mA cm−2. By comparison, in 1991 the 11‐MW Tokyo Electric Power’s system gave an average cell performance of approximately 0.75 V per cell at 190 mA cm−2. The rate of performance degradation of the stacks was less than 4 mV per 1000 h during a test of 4500 h. The results from this programme represent the highest performance of full‐size phosphoric acid cells and short stacks published to date. Mitsubishi Electric Corporation has also demonstrated an enhanced performance of 0.65 V at 300 mA cm−2 in single cells. Component improvements by Mitsubishi have resulted in the lowest rate of PAFC degradation to be publicly acknowledged, namely, 2 mV per 1000 h for 10 000 h at 200–250 mA cm−2 in a short stack with cells of 3600 cm2 area. Catalyst development continues to be an important aspect of the future of the PAFC. Over the past 10 years, several non‐precious metal catalysts for cathodes have been investigated and are similar to those for PEMFC catalysts, as described in Section 4.3.2, Chapter 4. These include transition metal (e.g., iron or cobalt) organic macrocycles of tetramethoxyphenylporphyrin (TMPP), phthalocyanines (PC), tetraazaannulene (TAA) and tetraphenylporphyrin (TPP). Another approach has been to alloy platinum with transition metals such as nickel, titanium, chromium, vanadium, zirconium and tantalum. Notable work by Johnson Matthey during the early 2000s showed that platinum–nickel alloy catalysts yielded a 49 wt.% increase in specific activity over pure platinum. This advance is translated into a 39 mV improvement in the performance of the air electrode at 200 mA cm−2. Other recent significant advances in PAFC technology are improved construction of the gas‐diffusion electrode and materials that offer greater protection against carbon corrosion. Of course, there is scope for many changes in the system design, with better engineered balance‐of‐plant components such as the reformer, shift reactors, heatexchangers and burners. Power electronics for DC to DC or DC to AC conversion have improved significantly since early units built in the 1990s, in terms of both size and performance (see Section 12.2.1, Chapter 12). Nonetheless, in the 400‐kW systems now supplied by Doosan, the footprint and weight of the fuel processing components are significant in comparison with the stack modules. Some of the earliest PAFC system demonstrations were conducted in the 1970s under the ‘Target Program’ funded by the American Gas Association. Many organizations have been involved in the development of PAFCs, but in recent years the thrust of commercialization has been borne mainly by two companies, namely, UTC Fuel Cells (formerly trading under the name ONSI or International Fuel Cells) based in Connecticut, USA, and Fuji Electric in Japan. The latter has been developing PAFCs since the 1980s and in the 1990s started to supply 50‐kW and 100‐kW systems worldwide. Over 100 such systems have been commissioned—a testament to their 205 206 Fuel Cell Systems Explained reliability and durability. Current research at Fuji is directed towards raising performance through better reforming catalysts and reducing costs, especially with respect to the balance‐of‐plant equipment. UTC Fuel Cells has supplied globally several hundred model PC25 200‐kW systems. The technology has found a niche application in high‐value locations such as a post office facility in Alaska, a science centre in Japan, the New York City Police Department and the First National Bank of Omaha. The last‐mentioned application is especially interesting as several fuel cells are linked together with other generation equipment to create an ultra‐reliable power system to sustain a critical load. Sure Power Corporation who put together the equipment has guaranteed 99.9999% reliability of the whole power system, thereby capitalizing on the reliability and robustness of PAFC technology. There are no single points of failure,6 and the Sure Power product is extremely fault tolerant. In 1987, Bharat Heavy Electricals Ltd. of India started to fund the research and development of PAFC systems. In 2001, the company had built a 50‐kW prototype, but shortly afterwards the company cut back its effort on PAFCs in favour of PEMFCs. Similarly, Caltex Oil Corporation in South Korea engaged in the construction of a 50‐kW system during the 1990s. This undertaking, too, does not appear to have been taken further. The Japanese companies Sanyo, Toshiba and Mitsubishi Electric all produced PAFC stacks during the 1980s and early 1990s, but there is little evidence that these companies have continued their efforts on PAFCs. As technical progress in PEMFC technology has moved forward, many organizations once involved in PAFCs have shelved their activities or used them to enhance their own PEMFC development. Further Reading Behling, N, 2012, History of phosphoric acid fuel cells, in Fuel Cells: Current Technology Challenges and Future Research Needs, pp. 53–135, Elsevier, Amsterdam, the Netherlands. Sammes, N, Bove, R and Stahl, K, 2004, Phosphoric Acid Fuel Cells: Fundamentals and Applications, Current Opinion in Solid State and Materials Science, vol. 8(5), pp 372–378. 6 A single point of failure (SPOF) is a part of a system that, if it fails, will stop the entire system from working. SPOFs are undesirable in any system with a goal of high availability or reliability, be it a business practice, software application or other industrial systems. 207 8 Molten Carbonate Fuel Cells 8.1 Principles of Operation The electrolyte of the molten carbonate fuel cell (MCFC) is a molten mixture of alkali metal carbonates — usually a binary mixture of lithium and potassium, or lithium and sodium carbonates — which is retained in a ceramic matrix of lithium aluminate (LiAlO2). At the high operating temperatures (typically 600–700°C), the alkali carbonates form a highly conductive molten salt, with carbonate CO32− ions providing ionic conduction. The anode and cathode reactions are shown schematically in Figure 8.1. Note that, unlike other common types of fuel cell, carbon dioxide (CO2) has to be supplied to the cathode as well as oxygen, and this becomes converted to carbonate ions, which migrate to the anode where reconversion to CO2 occurs. For every mole of hydrogen that is oxidized in the cell, there is therefore a net transfer of one mole of CO2 along with two Faradays of charge or two moles of electrons between the two electrodes. Note that the requirement for CO2 to be supplied to the MCFC contrasts with the alkaline fuel cell (AFC) from which CO2 must be excluded. The overall reaction of the MCFC is therefore: H2 1 O2 2 CO2 cathode H2 O CO2 anode (8.1) The Nernst reversible voltage for an MCFC, taking into account the transfer of CO2, is given by the equation 1 Vr Vr RT ln 2F PH2 . PO22 PH2 O PCO2c RT ln 2F PCO2a (8.2) where sub‐subscripts a and c refer to the anode and cathode gas compartments, respectively. Usually, there is a difference in the partial pressures of CO2 between the two electrodes, but when these pressures are identical, the cell potential depends only on the partial pressures of H2, O2 and H2O. In an MCFC system, it is normal practice to recycle externally the CO2 generated at the cell anodes to the cathodes where it is consumed. Whereas the recycling might at first seem to be an added complication and therefore place the MCFC at a disadvantage Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 208 Fuel Cell Systems Explained Hydrogen fuel Anode 2H2 + 2CO32– → 2H2O + 2CO2 + 4e– CO32– ions through electrolyte Cathode O2 + 2CO2 + 4e– 2CO32– Load (e.g., electric motor) Electrons flow round the external circuit Oxygen and carbon dioxide Figure 8.1 Anode and cathode reactions for a MCFC using hydrogen fuel. Note that the product water is at the anode and that both carbon dioxide and oxygen have to be supplied to the cathode. CH4 CO2 + CH4 + steam Anode Fuel in Electrolyte Cool air Cathode Hot air + CO2 + steam Burner Hot oxygen depleted air + CO2 + steam Figure 8.2 Adding carbon dioxide to the cathode gas stream does not add complexity. compared with other fuel‐cell types, it can be achieved by feeding the anode exhaust gas to a combustor (burner), which converts any unused hydrogen or fuel gas (e.g., methane, CH4) into water and CO2. The exhaust gas from the combustor is then mixed with fresh air and fed to the cathode inlet, as shown in Figure 8.2. The process is no more complex than that for other types of high‐temperature fuel cell, and, moreover, it also serves to preheat the reactant air, to burn the unused fuel and to bring the waste heat into one stream for use in a bottoming cycle or for other purposes. Another possible method of supplying CO2 to the cathode inlet is to use a device, such as a membrane, that will separate CO2 from the anode exit gas and allow it to be transferred to the cathode inlet gas. The advantage of using such a ‘transfer device’ is that any unused fuel gas can be recycled to the anode inlet or used for other operations such as providing heat for fuel processing. A further alternative would be to supply CO2 from an external source, especially where a ready supply of the gas is available. At the operating temperature of MCFCs, nickel (anode) and nickel oxide (cathode), respectively, are adequate catalysts to promote the two electrochemical reactions. Unlike the phosphoric acid fuel cell (PAFC) or the proton‐exchange membrane fuel cell Molten Carbonate Fuel Cells (PEMFC), noble metals are not required. Other important differences between the MCFC and the PAFC or PEMFC are its ability to perform direct electrochemical conversion of carbon monoxide (CO) and to reform hydrocarbon fuels internally. If CO were to be fed as fuel to the MCFC, the reactions at each electrode given in Figure 8.3 would occur. The voltage of the fuel cell when operating with CO is calculated in exactly the same way as for the hydrogen fuel cell, as described in the Section 2.1, Chapter 2. Two electrons are released for each molecule of CO, as illustrated in Figure 8.3, just as two electrons are released for each molecule of H2. Consequently, the formula for the open‐ circuit voltage is identical, i.e., Vr gf (8.3) 2F The method of calculating ∆ g f is given in Appendix 1. Notably, the values for hydrogen and carbon monoxide are remarkably similar at 650°C, as shown in Table 8.1. In practical applications, it is most unlikely that pure CO would be used as a fuel. It is more practical for the fuel gas to contain both H2O and CO, and in such cases the electrochemical oxidation of the CO would probably proceed via the water-gas shift reaction, i.e., equation (7.3), Chapter 7, which is a fast reaction that occurs on the nickel electrocatalyst of the anode. The shift reaction converts CO and steam to hydrogen that then oxidizes rapidly. The two reactions — direct oxidation of CO or shift reaction and then the oxidation of H2 — are entirely equivalent. Carbon monoxide fuel Anode 2CO + 2CO32– → 4CO2 + 4e– Load (e.g., electric motor CO32– ions through electrolyte Cathode O2 + 2CO2 + 4e– 2CO32– Electrons flow round the external circuit Oxygen and carbon dioxide Figure 8.3 Anode and cathode reactions for an MCFC using carbon monoxide fuel. Table 8.1 Values of ∆ g fo and Vro for hydrogen and carbon monoxide fuel cells at 650°C. Fuel ∆ g fo (kJ mol−1) Vro (V) H2 –197 1.02 CO –201 1.04 209 210 Fuel Cell Systems Explained Unlike the PEMFC, AFC and PAFC, the MCFC operates at a temperature that is sufficiently high to enable internal reforming of hydrocarbon fuels such as methane. This is a particular strong feature of both the MCFC and, as shall be discussed later in Chapter 9, the solid oxide fuel cell (SOFC). In internal reforming, steam is added to the fuel gas before it enters the stack. Inside the stack, the fuel and steam react in the presence of a suitable catalyst according to steam reforming reactions such as those given by equations (7.1) and (7.2), Chapter 7. Heat for the endothermic reforming reactions is supplied by the electrochemical reactions of the cell. Internal reforming is discussed in more detail in Chapter 10. Compared with low‐temperature fuel cells, the high operating temperature of the MCFC provides the opportunity for achieving superior overall system efficiencies and greater flexibility in the use of available fuels. Unfortunately, however, the higher temperature places severe demands on the corrosion resistance and life of cell components in the aggressive environment of the molten carbonate electrolyte. The PAFC and MCFC are similar in that they both use a liquid electrolyte that is immobilized within a porous solid matrix. As discussed in Section 7.3.2, Chapter 7, PTFE employed in a PAFC serves as both a binder and a wet‐proofing agent to maintain the integrity of the electrode structure and to establish a stable electrolyte|gas interface in the porous electrodes. The phosphoric acid is retained in a matrix of SiC doped with PTFE that is sandwiched between the anode and cathode. Since there are no materials comparable with PTFE that are able to withstand MCFC temperatures, a different approach is necessary to establish a stable electrolyte|gas interface in the porous electrodes of the MCFC. Stable interfacial boundaries have been achieved through a balance in capillary pressures. The porous electrolyte matrix generally has a narrow pore‐size distribution of relatively small pores. By contrast, the electrodes are characterized by a much wider pore‐size distribution of pores with much larger diameters. These different characteristics between electrolyte and electrodes allow the electrolyte matrix to remain completely filled with molten carbonate, whereas the porous electrodes are only partially filled. The pores in the electrodes will be partially filled in inverse proportion to the pore size, namely, the larger the pore, the less they are filled. Electrolyte management, i.e., the control over the optimum distribution of electrolyte in the different cell components, is critical for achieving high performance and good service life from MCFCs. This operational feature is extremely specific to MCFCs, and it should be emphasized that the use of a liquid electrolyte constrained within the matrix by capillary forces can give rise to various undesirable processes. Examples are the consumption of electrolyte by corrosion reactions, potential‐driven migration or creepage of liquid electrolyte and vapourization of carbonates or hydroxide species, all of which can contribute to the redistribution or loss of molten carbonate from the cells. 8.2 Cell Components In the early days of MCFC development, precious metals were generally used as the electrode materials. During the 1960s and 1970s, however, nickel‐based alloys became the preferred choice for the anode and nickel oxide for the cathode. Since that time, Molten Carbonate Fuel Cells Table 8.2 Evolution of cell component technology for molten carbonate fuel cells. Component ~1965 ~1975 Current status Anode Pt, Pd or Ni Ni ‐ 10 wt.% Cr Ni–Cr or Ni–Al 3–6 µm pore size 45–70% initial porosity 0.20–1.5 mm thickness 0.1–1 m2 g−1 Cathode Ag2O or lithiated NiO Lithiated NiO Lithiated NiO 7–15 µm pore size 70–80% initial porosity 60–65% after lithiation and oxidation 0.5–1 mm thickness 0.5 m2 g−1 Electrolyte support MgO Mixture of α‐, β‐, γ‐LiAlO2 10–20 m2 g−1 Electrolyte a α‐LiAlO2 β‐LiAlO2 0.1–12 m2 g−1 0.5–1 mm thickness 52 Li–48 Na 62 Li–38 K 62 Li–38 K 43.5 Li–31.5 Na–25 K ~60–65 wt.% 50 Li–50 Na Hot press ‘tile’ ~50 wt.% 1.8 mm thickness Tape cast ‘Paste’ 0.5–1 mm thickness Source: Hirschenhofer, JJ, Stauffer, DB and Engelman, RR, 1998, Fuel Cell Handbook, 4th edition, Report No. DOE–FETC‐99/1076, Parsons Corporation, for U.S. Department of Energy. a) Figures in this entry are in mol.% unless stated otherwise. both the electrode materials and the electrolyte structure (molten carbonate in a ceramic matrix of LiAlO2) have remained essentially unchanged. On the other hand, from the 1980s onwards, there has been an evolution in the methods employed for fabricating the electrolyte structures. Some of the principal materials used in MCFCs since the 1960s are summarized in Table 8.2. 8.2.1 Electrolyte The MCFC electrolyte is generally a mixture of alkali metal carbonates. Lithium offers the best ionic conductivity due to its low atomic weight, and in the beginning eutectic mixtures of Li–Na carbonate (Li2CO3–Na2CO3) were used; then attention shifted to Li–K (Li2CO2–K2CO3), especially the eutectic mixture of 62 : 38 mol.% Li2CO2–K2CO3, which proved effective for atmospheric pressure operation due to its increased reactivity. Ternary mixtures of Li–K–Na have also been investigated, but the present trend is to go back to mixtures of Li–Na, due to their lower vapour pressure that reduces electrolyte loss and their increased basicity compared with Li–K mixtures. 211 212 Fuel Cell Systems Explained The molten electrolyte is constrained in a ceramic fibrous matrix of high surface area submicron lithium aluminate (LiAlO2). Historically, the aluminate has been prepared from the low‐temperature γ form of alumina, and developers are comfortable in using this for the preferred tape‐casting manufacturing process. The matrix may be enhanced by the addition of particles of α‐alumina or zirconia to improve its mechanical strength. The material appears to be reasonably stable when supporting electrolytes based on lithium and potassium carbonate. If sodium carbonate is employed, e.g., as a Li–Na mixture, however, the particle size of γ‐LiAlO2 has been shown to increase over a period of time along with a phase change to α‐LiAlO2. The reverse phase change has also been noted in systems where α‐LiAlO2 has been employed with Li–K carbonate mixtures. Such phase changes can lead to a loss of structural integrity within the matrix, giving rise to cracks and cell degradation. No universal solution to this problem has yet been identified, although some improvements in the mechanical strength of the matrix have been reported by adding particles of alumina or even micro‐sized particles of aluminium to the slurry before it is tape‐cast.1 Until the 1990s, the matrix was often fabricated into a tile by hot‐pressing the powdered material, and it is still often referred to as the electrolyte ‘tile’. Nowadays, the matrix is invariably made using tape‐casting methods that are commonly employed in the ceramics and electronics industries. The basic process involves dispersing the ceramic materials in a ‘solvent’ that contains dissolved binders (typically, organic compounds), plasticizers and additives to achieve the desired viscosity and rheology of the resulting mixture or ‘slip’. The slip material is then cast in the form of a thin film over a moving smooth surface, and the required thickness is obtained by shearing with an adjustable blade (a so‐called ‘doctor blade’ assembly). After drying, the film is heated further in air and any organic binder is burnt off at 250–300°C. The semi‐stiff ‘green’ product is then assembled into the stack structure. Tape casting of the electrolyte provides an effective means of fabricating components of large area. The methods described here can also be applied to the cathode and anode materials for the ready manufacture of stacks with an electrode area up to about 1 m2. Compared with most other fuel cells, the ohmic resistance of the MCFC electrolyte, and especially the ceramic matrix, has a major influence on the operating voltage. Under typical operating conditions, the electrolyte has been found to account for 70% of the ohmic losses in a MCFC.2 Furthermore, the losses are dependent on the thickness of electrolyte according to the formula: Vr 0.533t (8.4) where t is the thickness in cm. The relationship shows that a fuel cell with an electrolyte structure of 0.025 cm thickness would operate at a cell voltage that is 82 mV higher than an identical cell with a structure of 0.18 cm thickness. Using tape‐casting methods, electrolyte matrices can now be made quite thin (0.25–0.5 mm), which is a development that offers a significant advantage in reducing the ohmic resistance. There is a trade‐off, 1 Kim, J, Patil, K, Han, J, Yoon, SP, Nam, S, Lim, T, Hong, S, Kim, H and Lim, H, 2009, Using aluminum and Li2CO3 particles to reinforce the α‐LiAlO2 matrix for molten carbonate fuel cells, International Journal of Hydrogen Energy, vol. 34, pp. 9227–9232. 2 Yuh, C and Farooque, JR, 1992, Understanding of carbonate fuel cell resistances in MCFCs. Proceedings of the Fourth Annual Fuel Cell Contractors Review meeting, U.S. DOE–METC, pp. 53–57. Molten Carbonate Fuel Cells however, between low resistance and the long‐term stability that is otherwise obtained with thicker materials. With respect to the electrolyte, there is an important difference between the MCFC and all other types of fuel cell. That is, the final preparation of the cell is carried out once the stack components are assembled. Layers of electrodes, electrolyte and matrix, together with various non‐porous components (current-collectors and bipolar plates), are assembled together, and the whole package is heated slowly up to the operating temperature of the fuel cell. As the carbonate reaches its melt temperature (over 450°C), it becomes absorbed into the ceramic matrix. This process results in a significant shrinkage of the stack that must be accommodated by careful mechanical design of the total assembly. During the heating, a reducing gas has to be supplied to the anode side of the cell to ensure that the nickel anode remains in the chemically reduced state. 8.2.2 Anode State‐of‐the‐art MCFC anodes are made of porous sintered nickel alloyed with a small amount of chromium and/or aluminium (see Table 8.2). The flat planar anode usually has a thickness of 0.4–0.8 mm, a porosity of between 55 and 75% and a mean pore diameter of 4–6 µm. Fabrication involves either hot‐pressing finely divided powder or tape‐casting a slurry of the powdered material mixed with additives to achieve the desired fluid properties of the powder–binder mixture. Tape casting is a low‐cost wet process that produces a thinner anode with more control over the thickness and pore‐ size distribution than the alternative procedure of hot pressing a powder. The addition of chromium or aluminium (usually 10–20 wt.%) improves the mechanical stability of the porous nickel by reducing the sintering of nickel particles during cell operation. Unless controlled, sintering can become a major problem in the MCFC anode because it leads to growth in pore size, a reduction in surface area and a loss of carbonate from the electrolyte. The change in pore structure can also result in mechanical deformation of the anode under the compressive load in the stack that, in turn, decreases electrochemical performance and may cause cracking of the electrolyte. Although Ni–Cr or Ni–Al alloy anodes have achieved acceptable stability for commercial applications, the cost is relatively high, and consequently developers have been investigating alternative materials. Partial substitution of the nickel with copper, for example, can go some way to reducing the cost of the alloys, but complete substitution is not feasible as copper exhibits more creep than nickel. In an attempt to improve the tolerance to sulfur in the fuel stream, various ceramic anodes are also being investigated. These include LiFeO2 with and without the doping of manganese or niobium. The anode of the MCFC has to provide more than just electrocatalytic activity. Because the anode reaction is relatively fast at MCFC temperatures, a high surface area is not required, compared with the cathode. Partial flooding of the anode with molten carbonate is therefore acceptable, and this feature is used to good effect, not only to act as a reservoir for carbonate (i.e., much in the same way that the porous carbon substrate functions in the PAFC) but also to replenish carbonate that may be lost from a stack during prolonged use. In some of the earlier MCFC stacks, a so‐called bubble barrier was located between the anode and the electrolyte. The bubble barrier consisted of a thin layer of Ni or LiAlO2 that had only small pores. The component served to prevent a flow of electrolyte 213 214 Fuel Cell Systems Explained to the anode, as well as to lower the risk of gas crossover. As discussed earlier, this latter problem is common to all liquid fuel cells in which an excess of pressure on one side of the cell may cause the fuel gas to cross the electrolyte. Nowadays, the use of a tape‐cast structure enables control of the pore distribution in anode materials during manufacture so that small pores are found closest to the electrolyte and larger pores are nearer to the gas channels. Long‐term electrolyte loss is, however, still a significant problem with the MCFC, and a totally satisfactory approach to electrolyte management is yet to be achieved. 8.2.3 Cathode Pure nickel oxide (NiO), which is an n‐type semiconductor, is the preferred choice for the cathode material. In the MCFC environment, the oxide becomes doped with Li+ ions from the electrolyte with the concomitant creation of extra electron–hole pairs for conduction by replacing Ni2+ with Ni3+ that, in turn, enhances the electrical conductivity. One of the major difficulties in employing NiO for the MCFC cathode, however, is that it has a small, but significant, solubility in molten carbonates. Consequently, some nickel ions are formed in the electrolyte and tend to diffuse towards the anode. As the ions approach the chemically reducing conditions at the anode (note: hydrogen is present from the fuel gas), metallic nickel can precipitate out in the electrolyte and cause internal short circuits with subsequent loss in the power output of the cell. Furthermore, the deposited nickel can act as a sink for the ions and thereby promote the further dissolution of the metal from the cathode. The leaching of nickel is intensified at high partial pressures of CO2 through the following reaction: NiO CO2 Ni 2 CO32 (8.5) The problem is mitigated if the more basic, rather than acidic, carbonates are used in the electrolyte. The basicity of the common alkali metal carbonates decreases in the order: (basic) Li2CO3 > Na2CO3 > K2CO3 (acidic). The lowest nickel oxide dissolution rates have been found for the eutectic mixtures of 62 wt.% Li2CO3 + 38 wt.% K2CO3 and 52 wt.% Li2CO3 + 48 wt.% Na2CO3. The addition of some alkaline earth oxides (CaO, SrO, BaO) to the carbonates has been shown to reduce the solubility of NiO by up to 50%. It also has been reported that lanthanum oxide3 can reduce the solubility still further, and this is thought to be due to the formation of oxy‐carbonates such as La2O2CO3 that increase the basicity of the electrolyte melt. Such rare earth oxide addition may also improve the rate of the oxygen reduction reaction in the MCFC; it has been observed that with addition of 0.5 wt.% CeO2 and 0.5 wt.% La2O3, the charge‐ transfer resistance in Li–K carbonate melt decreases by an order of magnitude.4 With state‐of‐the‐art NiO cathodes, nickel dissolution can be minimized by (i) using a basic carbonate, (ii) operating at atmospheric pressure and keeping the CO2 partial pressure in the cathode compartment low and (iii) employing a relatively thick electrolyte matrix to increase the Ni2+ diffusion path. By these means, cell lifetimes in excess of 3 Ota, KI, Matsuda, Y, Matsuzawa, K, Mitsushura, S and Karnia, N, 2006, Effect of rare earth oxides for improvement of MCFC, Journal of Power Sources, vol. 160(2), pp. 811–815. 4 Scaccia, S, Frangini, S, Dellepiane, S, 2008, Enhanced oxygen solubility by Re2O3, Journal of Molecular Liquids, vol. 138, pp. 107–112. Molten Carbonate Fuel Cells 40 000 h have been demonstrated under atmospheric pressure operation. For operation at higher pressure, alternative cathode materials have been investigated. LiCoO2 and LiFeO2 have attracted the most attention, and the former has the lower dissolution rate, which is an order of magnitude lower than NiO at atmospheric pressure. Dissolution of LiCoO2 also shows a lower dependency on CO2 partial pressure than NiO. Initial work on LiCoO2 in the early 1990s focused on using it simply as an alternative to NiO. Unfortunately, the relatively higher cost of cobalt and the lower mechanical strength of LiCoO2 compared with NiO discouraged developers to adopt it as a single‐ phase replacement. More success was obtained when LiCoO2 was combined with NiO or when both LiCoO2 and LiFeO2 were combined with NiO. Many cathodes made with NiO particles coated with oxides (i.e., core–shell structure) or with oxides finely dispersed on NiO particles have been evaluated in recent years. It has been shown, for example, that finely dispersed Ce and Co on NiO particles prepared via a polymeric precursor route significantly reduces NiO dissolution over the short term (up to a few hundred hours).5 It remains to be seen, in commercially operating stacks, whether nickel oxide dissolution from Ce–Co–NiO cathodes becomes a dominating factor in cell degradation in the longer term. 8.2.4 Non‐Porous Components The bipolar plates for the MCFC are usually fabricated from thin sheets of stainless steel. The anode side of the plate is coated with nickel. The coating is stable in the reducing environment of the anode, provides a conducting path for current collection and is not wetted by electrolyte that may migrate from the anode. Gas‐tight sealing of the cell is achieved by allowing the electrolyte from the matrix to contact the bipolar plate at the edge of each cell outside the electrochemically active area, as illustrated schematically in Figure 8.4. To avoid corrosion of the stainless steel in this ‘wet‐seal’ area, the bipolar plate is coated with a thin layer of aluminium, which forms a protective layer of γ‐LiAlO2 through reaction with Li2CO3 in the electrolyte. There are many designs of bipolar plate, as determined by whether the gases are manifolded externally or internally. Some of the bipolar plates developed for internal reforming have the reforming catalyst incorporated within the anode gas flow-field (see Figure 10.4, Chapter 10). 8.3 Stack Configuration and Sealing The stack configuration for the MCFC is very different from those described in previous chapters for the PEM, AFC and PAFC although there are, of course, some similarities. The most important difference is in the method of sealing. As described in the previous section, the MCFC stack is composed of various porous components (matrix and electrodes) and non‐porous components (current-collector and bipolar plate). In assembling and sealing these components, it is essential to ensure an equitable distribution of gas flows between individual cells, uniform distribution within each cell and good 5 Kim, MH, Hong MZ, Kim, YS, Park, E, Lee, H and Ha, W, 2006, Cobalt and cerium coated Ni powder as a new candidate cathode material for MCFC, Electrochimica Acta, vol. 51, pp. 6145–6151. 215 216 Fuel Cell Systems Explained Bipolar plate Cathode currentcollector Cathode Matrix Wet seal Anode Anode currentcollector Figure 8.4 Schematic cross-section of an MCFC: cell components exposed to high‐temperature hot corrosion environment showing the location of the wet seals on the anode and cathode sides of the electrolyte support matrix. thermal management to reduce temperature gradients throughout the stack. Whereas several proprietary techniques have been developed for constructing stacks, some generic aspects are described in the following text, with examples from practical systems. 8.3.1 Manifolding Reactant gases have to be supplied in parallel to all cells in the same stack via common manifolds. The basic arrangement for external manifolding is shown in Figure 1.11, Chapter 1. The electrodes are about the same area as the bipolar plates, and the reactant gases are fed into, and removed from, the appropriate faces of the fuel‐cell stack. One advantage of external manifolding is its simplicity in enabling a low‐pressure drop in the manifold and a good flow distribution between cells. A disadvantage is that the two gas flows are at right angles to each other, i.e., there is a cross‐flow, and this can cause uneven temperature distribution over the faces of the electrodes. Other problems have been gas leakage and migration (‘ion pumping’) of the electrolyte. Each external manifold must have an insulating gasket to form a seal with the edges of the stack. This is usually made from zirconia felt, which provides a small amount of elasticity to ensure a good seal. Note that most stack developers arrange the cells to lie horizontally so that the fuel and the oxidant are supplied to the vertical sides of the stack in a cross‐flow arrangement. The ‘Hot Module’ is an alternative arrangement pioneered by MTU Friedrichshafen and has vertically mounted cells with the anode inlet manifold located underneath the stack. In this way, sealing with the gasket located at the anode inlet is enhanced by the weight of the whole stack. Internal manifolding refers to a means of gas distribution via ducts that penetrate through the cells within the stack. The arrangement is illustrated in Figure 1.12, Chapter 1. An important advantage of internal manifolding is the greater flexibility in Molten Carbonate Fuel Cells the direction of flow of the gases. For even temperature distribution, co‐flow or counter‐ flow can be used, as discussed in Section 7.2.2, Chapter 7. The ducts that constitute the internal manifold in an MCFC stack are formed by holes in each separator plate that line up with each other once the stack components are assembled. The bipolar separator plates in an MCFC stack can be quite complex mechanically as illustrated by the IMHEX design shown schematically in Figure 8.5b and c. The parallel arrangement of channels in this design allows for either co‐flow or counter‐flow gas configurations. The IMHEX bipolar plate consists of several thin sheets of metal. Two of the sheets are corrugated to form flow-fields. These sit above and below a plane solid sheet of stainless steel that separates the fuel and oxidant gases from adjacent cells. Above and below the corrugated sheets are perforated current‐collector plates, and above and below these are plates that serve to act as the holders for the anode and cathode components of adjacent cells. The electrolyte matrix extends out to the extremities of the separator plates, and all of the plates have holes for the manifold gas ducts that align when the Straight rib (a) Active area Thickness: 0.020″ Pitch: ~0.200″ Height: ~0.100″ (b) 2 3 Current-collector Gas inlet/outlet Anode Electrolyte plate 1 Cathode Current-collector 3 2 Separator Gas channel (c) Electrolyte plate Cathode Separator Corrugated plate Anode Manifold Perforated plate Fuel Figure 8.5 Examples of practical separator-plate designs with internal manifolding: (a) IMHEX design of ECN, (b) multiple‐cell stack of Hitachi and (c) cross-section of wet‐seal area in an internally manifolded MCFC stack. 217 218 Fuel Cell Systems Explained stack is assembled. Thus the electrolyte itself, when molten, acts as the means of sealing both around the internal gas ducts and around the perimeter of each cell. Using the electrolyte in this way creates a ‘wet seal’ that prevents leakage or cross‐contamination of gases so long as the pressure of gas inside the stack is close to that of the external atmosphere. The advantages of using internal manifolding are therefore offset to some extent by the more complex design of bipolar plate compared with that required for an externally-manifolded stack. 8.3.2 Internal and External Reforming Internal reforming has been championed in the MCFC from the early 1960s. If a mixture of methane and steam (2 : 1 by volume) is reformed at the normal 650°C operating temperature of the MCFC and the product gas reaches thermodynamic equilibrium, then typically the methane conversion is about 85%. This conversion can be obtained via indirect internal-reforming (IIR), namely, by simply inserting reforming plates between groups of cells. Each reforming plate supports a conventional metal catalyst. By contrast, direct internal reforming (DIR) is achieved by inserting supported metal catalyst particles within the anode compartments of cells, i.e., within the flow‐ field channels of the bipolar-plates on the anode side. Direct internal reforming achieves 100% conversion of methane and much better heat utilization. Perhaps the most successful application has been the combination of both IIR and DIR technologies, as illustrated schematically in Figure 8.6. This approach is used in the Direct FuelCell™ products that are available from Fuel Cell Energy Inc. (FCE) in the United States. Oxidant Fuel DIR catalyst Natural gas nit ru e orm f Re e ag Partially reformed fuel k ac ll p IR ce D Oxidant IIR catalyst Figure 8.6 Combination of IIR and DIR achieves very high electrical conversion efficiency and high conversion of fuel. Molten Carbonate Fuel Cells Obviously, internal reforming eliminates the cost of an external reformer, and system efficiency is improved but, as mentioned earlier, at the expense of potentially greater complexity in cell configuration and issues with catalyst lifetime. Thus there is an economic compromise or trade‐off to be made between internal and external reforming. Internal reforming can only be conducted in an MCFC stack if a steam reforming catalyst is incorporated. This requirement arises because, although nickel is a good reforming catalyst, the conventional porous nickel anode has a low surface area and hence insufficient catalytic activity in itself to support the steam reforming reaction at the operating temperature (650°C). As will be discussed in Chapter 9, this is not the case in the SOFC, in which complete internal reforming may be carried out directly on the anode. For the DIR‐MCFC, the reforming catalyst needs to be close to the anode to enable the reaction to occur at a sufficiently high rate. Several research groups demonstrated internal reforming in the MCFC during the 1960s and identified the major problem areas to be associated with catalyst degradation, which was caused by carbon deposition, sintering and catalyst poisoning by alkali from the electrolyte. Internal reforming was studied extensively in the 1990s by BG Technology in a European Union‐supported programme led by BCN (Dutch Fuel Cell Corporation). This project identified novel catalyst compositions that could tolerate the presence of carbonate from the MCFC electrolyte. Key requirements for MCFC reforming catalysts are as follows: ● ● ● Sustained activity to achieve the desired cell performance and lifetime. For the catalyst to provide sufficient activity over the desired lifetime of the stack, any catalyst degradation must be less than the degradation in the electrochemical performance of the cell. The reforming reaction is strongly endothermic and thus causes a pronounced dip in the temperature profile of a cell with internal reforming. This behaviour is exceptionally severe with the DIR version. Consequently, optimization of the activity of the reforming catalyst is important to ensure that such temperature variations are kept to a minimum so as to reduce thermal stress and thus contribute towards a long stack life. Improvements in temperature distribution across the stack may also be achieved through the recycling of either the anode gas or the cathode gas, or both. Resistance to poisons in the fuel. Most raw hydrocarbon fuels that may be used in MCFC systems (including natural gas) contain impurities (e.g., sulfur compounds) that are harmful to both the anode and the reforming catalyst. In particular, the tolerance of most of the catalysts to sulfur is very low, i.e., typically in the parts per billion (ppb) range. Resistance to alkali or carbonate poisoning. With DIR, in which the catalyst is located close to the anode, there is a risk of catalyst degradation through reaction with carbonate or alkali from the electrolyte. Supported nickel catalysts are generally preferred, although supported ruthenium has also undergone testing for DIR‐MCFC application. Poisoning of DIR‐MCFC reforming catalysts is now known to occur through contact with liquid molten carbonate that arrives via two principal routes: (i) creepage along the metallic cell components and (ii) transport in the gas phase in the form of alkali hydroxyl species. The problem is illustrated schematically in Figure 8.7, which also indicates one possible remedial action, namely, the insertion of a protective porous shield between the anode and the catalyst. 219 220 Fuel Cell Systems Explained Electrolyte Anode (containing molten Li2CO3/K2CO3) Metal cell housing Gas phase transport LiOH(g) KOH(g) Li2CO3(I) K2CO3(I) Liquid-phase creep Protective shield Catalyst pellets Figure 8.7 Alkali transport mechanisms in the DIR‐MCFC. 8.4 Performance The operating conditions for an MCFC are selected essentially on the same basis as those for a PAFC. The stack size, efficiency, voltage level, load requirement and cost are all important, and a trade‐off between these factors is usually sought. The performance curve (voltage vs. current density) is defined by gas composition and utilization, cell pressure and temperature. State‐of‐the‐art MCFCs generally operate in the range of 100–200 mA cm−2 at 750–900 mV per cell. As with the PAFC, there is a significant overpotential at the cathode in the MCFC. This is noticeable if the cell performance when using air as oxidant is compared with that when using pure oxygen. The resulting behaviour is presented in Figure 8.8 that shows a cathode performance curve obtained at 650°C with an oxidant that comprised of oxygen, nitrogen and CO2 as typically used in MCFCs,6 and a curve obtained using a baseline nitrogen‐free composition. The baseline composition contains the reactants O2 and CO2 in the stoichiometric ratio that is required for the electrochemical reaction at the cathode (see Figure 8.1). With this gas composition, little or no diffusion limitations occur in the cathode because the reactants are provided primarily by bulk flow. The other (more realistic) gas composition yields a cathode performance that is limited by gas diffusion and by the lower partial pressure of oxygen in the mixture. 8.4.1 Influence of Pressure There is a performance improvement to be made by increasing the operating pressure of the MCFC. As shown in Section 2.5.4, Chapter 2, for a change in system pressure from P1 to P2, the change in reversible voltage according to the Nernst equation is given by: V RT P ln 2 4F P1 (8.6) 6 The gas composition is the result of burning anode exhaust gas with fresh air. This is the normal means of supplying CO2 to the cathode inlet of the MCFC. It yields a mixture with a lower oxygen‐to‐nitrogen ratio than fresh air. Molten Carbonate Fuel Cells Baseline cathode gas inlet composition: 33 vol.% O2 + 67 vol.% N2 Cathode overpotential/V 0 –100 Typical MCFC cathode gas inlet composition: 12.6 vol.% O2 + 18.4 vol.% CO2 + 69 vol.%N2 –200 0 100 200 300 400 Current density/mA cm–2 Figure 8.8 Influence of oxidant gas composition on cathode overpotential in a MCFC at 650°C. (Source: Adapted from Bregoli, LJ and Kunz, HR, 1982, The effect of thickness on molten carbonate fuel cell cathodes, Journal of the Electrochemical Society, vol. 129(12), pp. 2711–2715.) From this relationship, it can be shown that, at 650°C, a fivefold and a tenfold increase in pressure should yield a gain in the open‐circuit voltage of 32 and 46 mV, respectively. In practice, the increase is somewhat greater because of a reduction in cathode overpotential. An increase in the operating pressure of MCFCs results in enhanced cell voltages because of the accompanying increases in the partial pressure of the reactants, the gas solubilities and the mass‐transport rates. As was shown when considering the PEMFC at pressure (see Section 4.7.2, Chapter 4), parasitic power is required to compress reactant gases. Also opposing the benefits of increased pressure are the effects on undesirable side-reactions such as carbon deposition (via the Boudouard reaction, as discussed in Section 10.4.4, Chapter 10). Furthermore, higher pressure inhibits the steam reforming reaction, i.e., equation (7.1), Chapter 7, which is a disadvantage if internal reforming is being used. These effects, as will be described in Chapter 10, can be minimized by increasing the steam content of the fuel stream. In practice, the benefits of pressurized operation are significant only up to about 0.5 MPa. The problem of ‘differential pressure’ is another factor to consider. To reduce the risk of gas crossover between the anode and the cathode in the MCFC, the difference in pressure between the two sides of each cell should be kept as low as possible. For safety reasons, the cathode is usually maintained at a slightly higher pressure than the anode (a few kPa). The ceramic matrix that constrains the electrolyte is a fragile material that is susceptible to cracking if subjected to stresses induced either through thermal cycling, temperature variations or excessive pressure differences between the anode and the cathode. The pressure difference between the anode and the cathode compartments in stacks has always been a concern of system designers, since recycling of anode burn‐off gas to the cathode is normally required. Inevitably, there is pressure loss associated with such gas recycling. The control of small differences in pressure has also militated against running the stacks at elevated pressures even though there may be advantages from an efficiency standpoint. 221 222 Fuel Cell Systems Explained It is also necessary to minimize the pressure difference between the cell compartments and the outside of the stack when running at atmospheric or elevated pressure since the molten carbonate itself provides the gas‐tight wet seal between the compartments and the outside. Consequently, if an MCFC stack is to operate at elevated pressure, it must be enclosed within a pressure vessel that is filled a non‐reactive pressurizing gas, which is usually nitrogen. Another issue relating to the choice of pressure level is that an improvement in the overall efficiency of high‐temperature fuel‐cell systems may be achieved through combination with gas turbines. The latter require hot gas at typically 500 kPa. Solid oxide fuel cells are very suitable for this application, as they can run in a pressurized mode and have a high exhaust gas temperature. Molten carbonate fuel cells could also be combined with gas turbines even though the stack exhaust temperature is lower. For the reason described earlier, however, the MCFC is not so amenable to operating at high pressure. Accordingly, although some conceptual systems were devised in the 1990s, MCFC–turbine facilities are unlikely to be developed. 8.4.2 Influence of Temperature Simple thermodynamic calculations indicate that the reversible voltage of an MCFC should decrease with increasing temperature. This relationship is a function of the change in Gibbs free energy (see Sections 2.1 and 2.2, Chapter 2) and the change in gas composition at the anode. The main reason for the latter influence is that the gas composition depends on the equilibrium of the shift reaction, i.e., equation (7.3), Chapter 7, and this equilibrium is rapidly achieved. The equilibrium constant (Keq) for the shift reaction increases with temperature; the gas composition therefore changes with temperature and utilization to affect the cell voltage as illustrated in Table 8.3. Under actual cell operating conditions, the influence of temperature is most often dominated by the cathode overpotential. As the temperature is increased, this overpotential is reduced considerably. The net effect is that the operating voltage of the Table 8.3 Equilibrium constant (Kc) and equilibrium gas compositions for fuel gas and reversible cell voltage (Vr) calculated using the Nernst equation for an initial anode gas composition of 77.5 vol.% H2, 19.4 vol.% CO2 and 3.1 vol.% H2O at 0.1 MPa and a cathode composition of 30 vol.% O2, 60 vol. % CO2 and 10 vol.% N2. Temperature (K) Parameter 800 900 1000 PH2 0.669 0.649 0.641 PCO2 0.088 0.068 0.052 PCO 0.106 0.126 0.138 PH2 O 0.137 0.157 0.168 Vr (V) 1.155 1.143 1.133 Kc 0.247 0.48 0.711 Molten Carbonate Fuel Cells MCFC usually increases with temperature. Above 650°C, however, this effect is very slight, i.e., only about 0.25 mV per °C. Since higher temperatures also increase the rate of all the undesired processes, particularly electrolyte evaporation and material corrosion, 650°C is generally regarded as an optimum operating temperature. 8.5 Practical Systems 8.5.1 Fuel Cell Energy (USA) In the final years of the 20th century, MCFC development in the United States was conducted by two companies, namely, MC Power and FCE. An earlier programme undertaken by United Technologies Corporation (UTC) was concluded in 1992, and, with the consent of the US Department of Energy, know‐how was transferred to the Italian company Ansaldo. The MC Power effort, which grew out of work performed in the 1960s by the Gas Technology Institute of Chicago, finished in 2000. Consequently, FCE became the sole US manufacturer of MCFC systems. Figure 8.9 Fuel‐cell stack manufactured by The company has its headquarters in Fuel Cell Energy. (Source: Reproduced with Danbury (CT) and operates a manufacturpermission of Fuel Cell Energy.) ing plant in Torrington (CT) with a capacity of 90 MW per year. All of the products manufactured by FCE incorporate MCFC stacks that typically comprise 300–400 cells; an example is shown in Figure 8.9. One MCFC stack is used in the compact DFC300 system (see Figure 8.10) that measures 6 × 4.5 × 6 m3, weighs 19 tons and generates up to 300 kW at 480 V. Exhaust gas flow rate is 1800 kg h−1 at about 370°C. The system offers a cogeneration capacity between 140 and 235 kW. The DFC1500 plant (Figure 8.11) produces nominally 1.4 MW of power, measures 16 × 12 × 6 m3 and is built around a fuel‐cell module that houses four stacks. The plant also features other process modules such as a water treatment skid, a main process skid, electric balance‐of‐plant and a fuel pretreatment and desulfurization unit. Up to 1100 kW of heat is available from the exhaust gas flow of 8300 kg h−1. The largest of the FCE products, the 2.8‐MW DFC3000 system (Figure 8.12), is built around two 4‐stack modules. In 2014, FCE had installed over 80 sub‐MW and MW‐class DFC power plants around the world. The facilities have operated successfully on a variety of fuels, such as natural gas, biogas (digester gas) derived from industrial and/or municipal wastewater, propane ® ® ® ® 223 224 Fuel Cell Systems Explained Figure 8.10 A 300‐kW DFC‐300® system. (Source: Reproduced with permission of Fuel Cell Energy.) Exhaust stack Water treatment and control panel Blower, heater, preconverter Fuel treatment desulfurizer Module Switchgear Inverter Figure 8.11 DFC1500® 1.4‐MW system showing the different process units. (Source: Reproduced with permission of Fuel Cell Energy.) Molten Carbonate Fuel Cells Figure 8.12 DFC3000® 2.8‐MW system. (Source: Reproduced with permission of Fuel Cell Energy.) and coal gas. The ‘coal gas’ fuel here includes gas from active and abandoned coal mines as well as synthesis gas processed from coal. Biogas offers a unique opportunity for the MCFC. In a wastewater treatment facility, for instance, the methane‐rich biogas produced by the anaerobic digestion of sludge becomes the fuel to generate electricity to power the plant, and the resulting exhaust gas from the fuel cell can be used to heat the sludge to accelerate the anaerobic digestion. Moreover, biogas is a renewable fuel that is eligible for incentive funding for various project installations throughout the world. In 2012, FCE had field‐tested several units on biogas, of which 70% were wastewater treatment applications, the largest being a 1‐MW DFC1500 unit at King County (WA, USA). Unlike natural gas, which is very consistent in quality, the composition of anaerobic biogas is influenced by the chemical composition and treatment of the sludge. To achieve a consistent feed required for stable operation of MCFC stacks, FCE has designed systems in which the biogas is automatically blended with natural gas. ® 8.5.2 Fuel Cell Energy Solutions (Europe) Work carried out in the United States during the 1970s and early 1980s by UTC and the Gas Technology Institute provided the stimulus for European researchers to start their own research and demonstration programmes in the mid‐1980s. Over the following two decades, several research and industry groups were formed and delivered world‐ class innovations, but there was reluctance on the part of industry to provide investment for scale‐up and commercialization. Over a period of 20 years, a body of research was conducted separately by the Energy Research Centre of the Netherlands (ECN) and 225 226 Fuel Cell Systems Explained Ansaldo (Italy). Supported by the European Commission under the Framework Programmes for Research and Technological Development, these organizations collaborated with others including Gaz de France, Sydkraft (Sweden), the Consiglio Nazionale delle Ricerche (CNR) (Italy), British Gas (BG) Technology (UK), Stork Engineering (the Netherlands) and the Royal Dutch Schelde Group. Later, the German company MTU Friedrichshafen started its own MCFC programme in collaboration with FCE in the United States. Unfortunately, the European MCFC effort was eventually wound down as follows: ● ● ● In 2005, ECN sold its interests and intellectual property in MCFC technology to FCE in the United States. MTU Friedrichshafen rebranded as CFC Solutions. The prime interest of CFC Solutions lay in repackaging MCFC technology from the United States for the European market. The company went out of business in 2010. Ansaldo work also ended in 2010 with test facilities and intellectual property disbursed among university research groups in Italy and Poland. There has, however, been one positive outcome. In May 2012, FuelCell Energy Solutions GmbH (FCES) was formed as a joint venture between FCE (75%) and the Fraunhofer Institute for Ceramic Technologies and Systems (Fraunhofer IKTS) (25%). The venture continues efforts to enhance MCFC technology by combining the strength of the Direct FuelCell technology developed by FCE with that of the ‘EuroCell’ technology that will be licensed to the company by Fraunhofer IKTS. The latter builds on the patents, assets and intellectual property retrieved from the former CFC Solutions and now held by Fraunhofer IKTS. The assets include the 250‐kW Hot Modules constructed by MTU and installed throughout Europe. The European venture therefore provides a platform for FCE in Europe, adding value by accessing former German MCFC research and development and knowledge of ceramic materials and processing. FuelCell Energy Solutions GmbH has manufacturing facilities for new European MCFC systems in Ottobrunn, Germany. Unlike the FCE stack (shown in Figure 8.9) in which the cells are horizontal, the European Hot Module is based on the original design by MTU (see Section 8.4.1) and therefore uses stacks in which the cells are arranged vertically. The features of the MTU stack arrangement and the Hot Module counterpart, illustrated in Figure 8.13, are as follows: ● ● ● ● ● ● ● All hot components are integrated into one common vessel. Common thermal insulation with internal air-recycling for best temperature levelling. Minimum flow resistance and pressure differences. Horizontal fuel‐cell block with internal reforming. Gravity‐sealed fuel manifold. Simple and elegant mechanics. Transportable by standard truck in sizes up to 400 kW. The design enables all the processes that need to be run at elevated temperature to be located within the Hot Module, whereas auxiliaries such as power‐conditioning and natural gas compression are located outside. In the Hot Module system, a 292‐cell stack produces about 280 kW DC, which translates to 250 kW AC when power conversion (a) Endplate Insulation Gas manifold Fuel Bipolar plate Anode Matrix Cathode Oxidator gas Tension rod (b) External exaust recycle Hot module Cat burner Mixing zone Dosing pump Preheater FC Nitrogen for start and emergency DFC DC/AC convert grid connector Fresh air Nonretum flap Startheater Heat utilization (optional) Steam Boiler feed water Hot module periphery Fuel gas Desulfurizer Electrical Preconverter startheater Control system (c) Figure 8.13 Example of MCFC system assembly. MTU ‘Hot Module’ cogeneration system showing (a) stack construction, (b) simplified flow diagram and (c) early demonstration units under construction. (Source: Reproduced with permission of Fuel Cell Energy.) 228 Fuel Cell Systems Explained and parasitic losses are taken into account. Exhaust heat is available at about 450°C and an LHV efficiency of 49% is reported for systems that are fuelled by natural gas. Two 250‐kW ‘Eurocell’ systems (denoted as DFC 250 EU Reference Plant) have been deployed at the Elektrizitätswerk Zürich and the German Federal Ministry for Research. A further two systems (DFC 300 EU Reference Plant) are being installed in London at two new prestigious developments in Regent Street and Fenchurch Street. ® ® 8.5.3 Facilities in Japan A MCFC development programme was conducted in Japan from 1981 to 2004. It was a major undertaking, almost equal to that of the United States. Funding was provided by Japan’s Ministry of Economy, Trade and Industry (METI) with a total budget of about US$470 million. The programme was managed by the New Energy Development Organization (NEDO) and several companies took part, including Fuji Electric, Hitachi, Ishikawajima‐Harima Heavy Industries (IHI), Mitsubishi Electric, Sanyo Electric and Toshiba. Many of the participants sought partnership with MCFC developers in the United States to gain access to expertise. Mitsubishi Electric, for example, formed an alliance with FCE, and IHI collaborated with MC Power. The Japanese MCFC programme was dubbed the ‘Moonlight Project’, and there were the following three phases: Phase 1: (1981–1986) focused on developing 10‐kW stack demonstrations. Phase 2: (1987–1999) aimed to develop cells and stacks of up to 200 kW, and balance‐of‐ plant systems. Phase 3: (2000–2004) aimed to develop of a series of high‐pressure, high‐efficiency short stacks to ultimately develop a 750‐kW system. A notable success of the Japanese effort was the operation of pressurized stacks at current densities of 200 mA cm−2 with voltage degradation rates of less than 0.3% per 1000 h, which encouraged the prospect of lifetimes in excess of 40 000 h. Unfortunately, the programme — particularly the last phase — suffered from various technical setbacks among the interlinked projects. By the end of the NEDO work in 2004, a projected 7‐MW plant was abandoned as, with the exception of IHI, all of the partners had withdrawn from the programme. As the sole remaining Japanese company engaged in MCFC development, IHI started to market MCFC products in 2005, but few were sold.7 It appears that IHI is also no longer engaged in the MCFC business. 8.5.4 Facilities in South Korea POSCO Energy is the largest private power‐generating company in Korea, with over 40 years of experience in building and operating power plants. The company has promoted the advancement of MCFCs since early 2000 with government support and in collaboration with the Korea Electric Power Corporation (KEPCO). Work involved the development of external‐reforming MCFC technology that in 2010 resulted in the 7 There have been a total of 23 MCFC pre‐commercial prototypes or commercial power plants installed in Japan. As of April 2012, IHI had four 300‐kW plants in service. At the same time, there were 13 FCE plants installed, all based on 250‐kW systems. For more details, refer to Fuel Cells by Nobura Behling, listed at the end of this chapter. Molten Carbonate Fuel Cells Figure 8.14 World’s largest operating MCFC power plant (59 MW) located in Hwaseong, South Korea, comprises 21 DFC‐3000® systems and is owned by Gyeonggi Green Energy. (Source: Reproduced with permission of Fuel Cell Energy.) successful demonstration of a 125‐kW plant. In 2007, POSCO Energy was granted a licence to manufacture, and distribute within Korea, FCE systems that used stacks supplied from the United States. The aim of the arrangement was to reduce costs by producing balance-of-plant at a new facility in Pohang. A research and development centre was established at this site, and in March 2011 a cell manufacturing facility was added. Current expansion is expected to increase annual output to 170 MW in 2016. POSCO Power has provided the provinces of Gyeonggi, Geonra, Gyeongsang and Chungchung with 8.8‐MW MCFCs, and, in 2011, 14‐MW counterparts were supplied to Suncheon, Dangjin, Ilsan, and Incheon. A 11.2‐MW plant in Daegu City followed in 2012, and the world’s largest operating fuel‐cell plant was commissioned in 2013 — a 59‐MW MCFC (see Figure 8.14) that delivers power and district heating to the city of Hwaseong. As of late 2014, a total of 144.6 MW is being generated by MCFC plants at 18 sites in Korea. 8.6 Future Research and Development It is clear that the MCFC products now marketed by FCE and associated companies fulfil a number of commercial requirements. Of all the types of fuel‐cell system, MCFC power plants are the largest and with over 80 DFC installations worldwide that, to date, have produced over 2 billion kWh of power. The DFC products are designed to give a 20‐year lifespan with stack lifetimes of between 5 and 7 years. Systems now demonstrate 50% (LHV) efficiencies, low noise (below 65 dBA) and very low emissions (negligible ® ® 229 230 Fuel Cell Systems Explained with regard to sulfur oxides, SOx, and particulates and less than 10 ppm of nitrogen oxides, NOx). Nevertheless, with all the advantages that these systems have over alternative technologies, it is still difficult to make a compelling case for the commercial adoption of MCFC power plant. It became clear at the close of the 20th century that the capital cost of MCFC systems, although lower than some other fuel‐cell types, was not low enough for the technology to be commercially competitive. Investment in scale‐up and manufacturing was required, and venture capital and other funding sources were difficult to secure. For these reasons, the programmes in Europe were halted, even though good technical progress had been made during the previous 20 or so years. A similar situation eventuated in Japan. If MCFC systems are to become commercially viable in significant numbers, improvements must be made both in the materials from which cells and stacks are constructed and in the design of systems. A fundamental issue is the interdisciplinary nature of fuel‐cell research, as will be discussed in the closing Chapter 12 of this book. It is clear, however, that some areas of MCFC technology would benefit from further research and investigation, as follows: ● ● ● Power density. Compared with all other candidate fuel cells, the power density of a state‐of‐the‐art MCFC is very low. For an internally reforming cell, the power density is typically 0.16 W cm−2, but with pressurized stacks 0.5 W cm−2 should not be an unreasonable target. Cathode overpotential. As with low‐temperature fuel cells, the rate of oxygen reduction is low and thus leads to poor cell performance. Alternative cathode materials should be investigated, including metals, oxides and semiconductors. Wetting of materials by the molten carbonate electrolyte. The corrosion of metals and alloys by molten salts reduces cell lifetime. A better understanding of passive layer formation and dissolution–deposition mechanisms is required, especially with respect to porous or layered composite substrates. The present configurations of FCE systems are the product of progressive enhancement in the performance of component materials but with the essential features of internal reforming that were outlined by Baker and colleagues in the 1960s. As mentioned earlier, during the late 1990s, the ECN brought in other industrial partners to develop an advanced DIR‐MCFC system for the European cogeneration market. The consortium devised several concepts that included the novel method of stack connection illustrated in Figure 8.15. In this arrangement, three stacks are connected together in series on the cathode side and in parallel on the anode side. The anode gas is recycled. Calculations show that the system eliminates the need for all major heat-exchangers and provides a high efficiency. The concept is also applicable to systems that are comprised of two or more stacks. Research on various aspects of MCFC technology is being conducted by some universities and institutes in Europe (Italy, Sweden, Poland, Germany and France), the United States (the University of Connecticut and Illinois Institute of Technology) and Korea (KIST). 8.7 Hydrogen Production and Carbon Dioxide Separation For an MCFC system that incorporates the reforming of natural gas or biogas, hydrogen produced at the anode side is normally consumed by the cell to produce electricity. If the supply of fuel is maintained but the electrical demand on the system is reduced, Molten Carbonate Fuel Cells Recirculation pump CH4 Fuel in Anode Anode Anode Electrolyte Electrolyte Electrolyte Cathode Cathode Cathode Combustor Exhaust Air in Figure 8.15 Stack networking can simplify system design and increase overall performance. proportionately more hydrogen appears in the anode exhaust gas. Consumption of this extra gas in the normal cathode burner would increase the temperature at the inlet and thereby intensify the demand for air cooling. A potentially more useful option is to separate the hydrogen from the anode exhaust gas so that a change in the cathode air flow is not required, and the high energy efficiency of the system is maintained. This is the basis on which FCE has been promoting the DFC systems as hydrogen generators. Put simply the MCFC can either act as a power generator that is operating at full electrical load or if the load demand falls, excess hydrogen that would otherwise be routed to the burner is separated and stored for alternative use (e.g., for fuelling fuel‐cell vehicles). The concept is attractive for applications where the MCFC system is run in parallel with an intermittent renewable energy source, e.g., a wind farm. In the MCFC, there is a transfer of CO2 from the cathode to the anode in the form of carbonate ions. This transfer can also be put to practical use in that the MCFC can be employed for separating CO2 from power‐plant flue gases. The idea is also being promoted by FCE and relies on the fact that the exhaust gas from the cathode of typical MCFC stacks contains about 1 vol.% CO2 compared with about 10 vol.% at the cathode inlet. If the cathode is therefore supplied with flue gas from a fossil‐fuel power plant, most of the CO2 is extracted by the fuel cell and appears in concentrated form at the outlet of the anode. At the anode outlet, the CO2 may be more amenable to separation and capture. The concept is illustrated in Figure 8.16. ® 8.8 Direct Carbon Fuel Cell In 1891, the famous Thomas Edison filed a patent in the United States for the direct electrochemical conversion of carbon to electricity. The proving of the idea was contracted to William W. Jacques and Lowell Briggs of the American Bell company, who in 1896 demonstrated a device for directly producing DC power from baked coal. The apparatus (Figure 8.17) consisted of 100 cells connected in series and placed on top of a 231 232 Fuel Cell Systems Explained Hydrogen-rich fuel CO2 separator CO2 capture (~90% CO2) Recycle or sell Water CO2 - depleted flue gas (~1% CO2) Direct fuel-cell (DFC) A n o d e C a t h o d e Supplemental fuel Fossil fuel Power Flue gas (~10% CO2) Power plant or process emitting CO2 greenhouse gas Power Air Figure 8.16 Electricity production and carbon separation. (Source: Reproduced with permission of Fuel Cell Energy.).) (a) (b) Figure 8.17 Illustration of the Jacques and Briggs carbon battery: (a) 100 cells sitting on top of a coal‐ fired furnace. (b) Details of an individual cell: a carbon, C, is plunged into a solution of caustic soda, E; a pump, A, forces air into a perforated nozzle, R, which distributes the air uniformly in the electrolyte. The positive pole is fixed upon an iron receiver, I, that contains the solution, and the negative pole, B, upon the carbon that is supported and insulated from the receiver by a collar, S. Two tubes, o and i, serve for the admission and discharge of the electrolyte solution. Molten Carbonate Fuel Cells furnace that kept the temperature of the caustic soda electrolyte at between 400 and 500°C. The output was measured as 16 A at 90 V. Since the carbon was consumed rather than continuously supplied and the hydroxide electrolyte reacted to form carbonate according to equation (8.7), the system was really a battery rather than a fuel cell. Current densities as high as 100 mA cm−2 were reported for the earliest cells: C 2NaOH O2 Na 2 CO3 H2 O Er (8.7) 1.42 V The direct electrochemical oxidation of carbon to CO2 can, in theory, proceed with very high conversion efficiency since the entropy change of the complete oxidation of carbon (ΔS°) is very small in comparison with the entropy change for the oxidation of hydrocarbons: C O2 G/ H CO2 1.00; (8.8) S 2.5 J mol 1K 1 ; Er 1.0 V ; T 800 C Since air is the preferred oxidant, the early fused hydroxide electrolytes used in the Jacques and Briggs cells were later replaced by molten carbonates. Such so‐called direct carbon fuel cells (DCFCs) employing molten carbonate electrolyte operate at 700–900°C.8 The carbon is oxidized directly to CO2 to yield four electrons per carbon atom. The half‐cell reactions are as follows: Anode : C 2CO32 Cathode : O2 2CO2 3CO2 4e 4e 2CO32 (8.9) (8.10) Unlike most other types of fuel cell in which both the anode reactants and products are gases, the cell potential of the DCFC is not dependent on the degree of conversion of the carbon. Thus, in theory up to 100% carbon conversion could be achieved in a single‐pass operation. By contrast, if carbon is treated to produce syngas (e.g., by gasification of coal with steam and oxygen), the syngas converted to hydrogen and the hydrogen used in fuel cells, then the energy and exergy losses in each step in the process can add up to a sizeable figure. After the initial work by Jacques, interest in the carbonate DCFC virtually disappeared in the first half of the 20th century. It re‐emerged in the late 1990s, prompted mainly by the high theoretical efficiency of direct carbon oxidation and because of the growing interest in carbon capture and storage (CCS). Carbon dioxide is the only product at the anode of the DCFC and thus facilitates its capture, in contrast to conventional coal gasification or combustion systems that require specific processes for the separation and capture of CO2. 8 Alternative types of DCFC that have been investigated are based around SOFC materials, that is, with O2− being the ion that migrates from cathode to anode through a solid yttria‐stabilized zirconia electrolyte. The operating temperature for this type of cell is in the 800–1000°C range. In such cells the cathode is lanthanum strontium manganite (LSM), and the anode can be a nickel‐based solid that interacts directly with fluidized carbon particles. Alternatively the anode can be a molten metal, such as tin, or molten carbonates into which the carbon fuel is supplied. The latter approach is essentially a hybrid type of molten carbon/solid oxide fuel cell. 233 234 Fuel Cell Systems Explained As with the conventional MCFC, CO32− is the ion that migrates from the cathode to the anode in the carbonate DCFC. In a laboratory‐scale single‐cell DCFC, the ions are generated by bubbling air through the molten carbonate electrolyte where they interact with a nickel oxide cathode. It may be more practical in a scaled‐up system to employ a porous ceramic oxide for the cathode material as in the SOFC, e.g., lanthanum strontium manganite. The challenges in developing a practical carbonate DCFC have been the build‐up of ash in the electrolyte, low anode reaction rates and the high cost of producing suitable carbon and its transport to the fuel cells. There are also mechanical issues concerning the distribution of carbon in the fuel cell. The form of carbon has been found to be particularly influential on the cell performance. It has been shown that carbon with a highly disordered structure at the nanometre scale (2–30 nm) is much more reactive than graphite. Such material, commonly referred to as ‘turbostratic’ carbon, can be obtained by the controlled thermal decomposition of coal, petroleum or natural gas. The principal challenge for the DCFC is scaling up the concept from a laboratory batch process to that of a ‘fuel cell’, in which carbon is continually fed and any impurities such as ash are removed. If the carbonate DCFC is to be developed further, it is also necessary to show how the structure of carbon is determined by the method of its preparation and hence how best to produce carbons with high electrochemical reactivity. Specifically, the key requirement for the DCFC is to devise an efficient and affordable means whereby a source of carbon, such as coal or biochar, can be transformed into a low‐ash carbon to serve as a premium fuel. Further Reading Behling, N, 2012, History of molten carbonate fuel cells, in Fuel Cells: Current Technology Challenges and Future Research Needs, pp. 137–221, Elsevier, Amsterdam. Farooque, M and Maru, H, 2009, Full scale prototypes, in Garche, J, Dyer, CK, Moseley, PT, Ogumi, Z, Rand, DAJ and Scrosati, B (eds.), Encyclopedia of Electrochemical Power Sources, vol. 2, pp. 508–518, Elsevier, Amsterdam. Leto, L, Della Pietra, M, Cigolotti, V and Moreno, A, 2015, International Status of Molten Carbonate Fuel Cells Technology, Advanced Fuel Cells Implementing Agreement, IEA Energy Technology Network. McPhail, SJ, Aarva, A, Devianto, H, Bove, R and Moreno, A, 2011, SOFC and MCFC: Commonalities and opportunities for integrated research, International Journal of Hydrogen Energy, vol. 36, pp. 10337–10345. Selman, JR, 2006, Molten‐salt fuel cells—Technical and economic challenges, Journal of Power Sources, vol. 160, pp. 852–857. 235 9 Solid Oxide Fuel Cells 9.1 Principles of Operation 9.1.1 High‐Temperature (HT) Cells The solid oxide fuel cell (SOFC) is a completely solid‐state device that operates with an oxide ion‐conducting ceramic material as the electrolyte. It is therefore simpler in concept than all of the other types of fuel‐cell system as only two phases (gas and solid) are involved. The electrolyte management issues that arise with the phosphoric acid (PAFC) and molten carbonate (MCFC) fuel cells do not occur, and the high operating temperatures mean that precious metal electrocatalysts are not necessary. As with the MCFC, both hydrogen and carbon monoxide (CO) can serve as fuels for the SOFC, as shown in Figure 9.1.1 The SOFC is similar to the MCFC in that a negatively charged ion (O2−) is transferred from the cathode through the electrolyte to the anode. Consequently, water is produced at the anode. Development can be traced back to 1899 when Nernst was the first to recognize that zirconia (ZrO2) is a conductor of oxygen ions. Until recently, all SOFCs have been based on an electrolyte of zirconia stabilized with the addition of a small percentage of yttria (Y2O3). Above a temperature of about 700°C, zirconia becomes a conductor of oxygen ions (O2−), and state‐of‐the‐art zirconia‐based SOFCs function between 800 and 1100°C. This is the highest operating temperature range of all fuel cells and thereby presents extra challenges in terms of construction and durability. Solid oxide fuel cells typically exhibit good electrical efficiencies, i.e., >50% (LHV), and even better performance in combined‐cycle schemes. Indeed, both simple‐cycle and hybrid SOFC plants have demonstrated some of the best efficiencies of any power‐generation system. The anode of the SOFC is usually a cermet2 of yttria‐stabilized zirconia (YSZ) and nickel. The nickel is chosen principally because it has a good electronic conductivity and is resilient under chemically reducing conditions. Moreover, it is also far more sulfur resistant than the precious metal catalysts employed in low‐temperature fuel cells and is 1 The high temperatures and presence of steam also means that hydrogen production via the shift reaction (equation (7.3), Chapter 7) invariably occurs in practical systems, as with the MCFC. The use of the carbon monoxide may thus be more indirect but just as valuable, as shown in Figure 9.1. 2 A cermet is a composite material that is composed of ceramic (cer) and metallic (met) materials. Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 236 Fuel Cell Systems Explained Product water as steam, available for steam reformation of fuel Hydrogen fuel Anode 2H2 + 2O2– → 2H2O + 4e– O2– ions through electrolyte Cathode O2 + 4e– → Load 2O2– Electrons flow round the external circuit Oxygen, usually from the air Carbon monoxide fuel Anode 2CO + 2O2– → 2CO2 + O2– ions through electrolyte Cathode O2 + 4e– → 4e– Load 2O2– Electrons flow round the external circuit Oxygen, usually from the air Figure 9.1 Anode and cathode reactions for the SOFC, when using hydrogen and carbon monoxide fuel. not poisoned by carbon monoxide. Consequently, the SOFC can accept a wide range of converted (reformed) fuels, which include coal‐derived gases. Indeed, the presence of nickel can be used to advantage as a catalyst for internal reforming — it is possible to perform this process on the anode of the SOFC. By contrast, finding a suitable material for the cathode has proved to be challenging. In the early days of SOFC development, noble metals were used but have since fallen out of favour on the grounds of cost. Most cathodes are now made from electronically conducting oxides or a ceramic material that possesses both ionic and electronic conductivities. The most common cathode material of the latter type is strontium‐doped lanthanum manganite (LSM), La₁₋ₓSrₓMnO₃. Solid Oxide Fuel Cells Unlike the MCFC, the SOFC requires no recycling of CO2, which leads to system simplification, as shown in Figure 9.1. The absence of CO2 at the cathode means that the open‐circuit voltage (OCV) of the cell is given by the simple form of the Nernst equation, i.e., equation (2.34), Chapter 2. Nevertheless, the SOFC has one disadvantage compared with the MCFC in that at the higher operating temperature, the Gibbs free energy of formation of water is less negative. Consequently, the OCV at 1000°C is about 100 mV less than that of the MCFC at 650°C (see Chapter 2, particularly Figure 2.3 and Table 2.2). The lower OCV would be expected to decrease the efficiency of the SOFC. In practice, however, the effect is partly offset by the lower internal resistance of the SOFC and by the ability to use thinner electrolytes than those required by the MCFC. Consequently, the SOFC can be operated at relatively high current densities (i.e., up to 1000 mA cm−2). Development of the SOFC escalated in the 1960s with the introduction of the tubular design by Westinghouse Electric Corporation. Compared with the planar counterpart, the tubular shape is more tolerant of the thermal stresses that are associated with the high temperature of operation. Initially, lanthanum chromite (LaCrO3), a ceramic material, was used to interconnect the tubular cells. Unfortunately, however, the chromium in this material can migrate into the cathode and cause serious degradation of the cell. This behaviour is one of the reasons for the more recent development of intermediate‐ temperature SOFCs (IT‐SOFC). The high‐temperature SOFC (HT‐SOFC) has been championed by companies such as Westinghouse Electric, Siemens AG and Rolls‐Royce for application in large‐scale (base‐load) power‐generation facilities fuelled by natural gas. The ability to carry out internal reforming potentially simplifies system design and enables an efficiency that is significantly higher than that obtained by PAFC systems that operate on natural gas. Furthermore, the heat available from the stack can be used for large‐scale cogeneration or a combined‐cycle plant. 9.1.2 Low‐Temperature (IT) Cells Research on HT‐SOFCs during the 1980s and 1990s identified several long‐term problem areas especially in relation to planar stacks. Delamination of the electrodes from the electrolyte occurred due to a mismatch between thermal expansion coefficients that was exacerbated by the thermal stresses encountered at the high temperature of operation. Sealing was also an issue between adjacent cells and the metal bipolar plates, as well as between the cells and the metal supporting hardware (see Section 9.2.3.3). Operating temperatures above 800°C require the use of expensive metal alloys, such as inconel steels, for the stack hardware and bipolar plates. The problems stimulated researchers to find ways of reducing the operating temperature of the SOFC. New materials were identified for the electrolyte and electrodes, and the outcome has been a technology that operates typically between 600 and 800°C and is now generally known as the ‘IT‐ SOFC’. In pursuing low‐temperature operation, the following advantages were identified: ● Metal interconnects, which are subject to severe corrosion at elevated temperatures, may now be used instead of LaCrO3‐based oxide interconnects; see Section 9.2.3.2. The use of metals is expected to lead to significant cost reduction and longer lifetime. 237 238 Fuel Cell Systems Explained ● ● ● ● ● Thermodynamic conversion efficiency increases for reformed gas (a mixture of carbon monoxide (CO) and hydrogen). More options for the sealing of cell components are available. Low‐cost austenitic steels can be employed in stack construction (i.e., for the bipolar plates in planar SOFCs) rather than exotic alloys such as inconel metals. Cell components are less prone to delaminating through differences in thermal expansion. For small systems, radiation heat loss from the stack becomes less. Hence, heat management becomes easier at lower temperatures. On the other hand, decreasing the operation temperature gives rise to additional issues with the component materials: ● ● ● ● The oxide‐ion conductivity of electrolyte material decreases rapidly as the temperature is reduced. It is therefore essential to have faster oxide‐ion conductors or to develop a good method of fabricating a thin electrolyte film. Consequently, anode‐supported electrolytes have proved to be a promising alternative to the more conventional electrolyte‐supported cell, as discussed in Section 9.3.2. It is necessary to use electrode materials that have greater activity. For example, scandium‐doped zirconia is more conductive than YSZ, permitting a further reduction of the operating temperature of an IT‐SOFC by 50–100°C. Sulfur poisoning of nickel, which is still the best choice for the anode, becomes more severe. Chromium poisoning of the lanthanum strontium manganite used for cathodes intensifies, against expectations. In more recent years, another category of cells that operate below 600°C has been designated as ‘low‐temperature SOFCs’ (LT‐SOFCs).3 Affording the prospect of faster start‐up and more robust operation compared with high‐temperature counterparts, these will not be discussed further here as they are at an early stage of research. 9.2 Components 9.2.1 Zirconia Electrolyte for HT‐Cells In an SOFC, the electrolyte is exposed to both oxidizing (air side) and reducing (fuel side) species at high temperatures. Hence, a successful long‐term SOFC operation requires the electrolyte to have the following properties: ● ● Sufficient ionic conductivity to minimize ohmic loss and with little electronic conductivity. A dense structure, i.e., impermeable to gas. It is difficult to fabricate dense, thin layers of electrolyte if the porous anode or the cathode is used as the support. 3 Gao, Z, Mogni, LV, Miller, EC, Railsback, JG and Barnett, SA, 2016, A perspective on low‐temperature solid oxide fuel cells, Energy and Environmental Science, vol. 9, pp. 1602–1644. Solid Oxide Fuel Cells ● ● Chemical stability — the electrolyte is exposed to both the air and the fuel at elevated temperatures, so it must endure both oxidation and reduction processes. Mechanical compatibility with both electrodes, i.e., the thermal expansion coefficients, must match at the interfaces. Zirconia doped with 8–10 mol.% yttria (YSZ) continues to be the most effective electrolyte for HT‐SOFCs although several other oxides have been investigated, for example, Bi2O3, CeO2 and Ta2O5. Pure zirconium dioxide (ZrO2) has a monoclinic crystal structure and is a poor ionic conductor at room temperature. When heated above 1173°C, it undergoes a phase transformation from monoclinic to tetragonal and then, on further heating to 2370°C, changes to a cubic fluorite structure. These phases are ion conductors. The phase change from monoclinic to tetragonal is accompanied by an appreciable change in volume (about 9%). To stabilize the cubic structure at lower temperatures and to increase the concentration of oxygen vacancies (which are required for the conduction of oxygen ions via vacancy hopping) acceptor4 dopants are introduced into the cation sublattice. Example dopants are Ca2+ and Y3+, which produce calcia‐stabilized zirconia (CSZ) and YSZ, respectively, as illustrated in Figure 9.2. Each dopant stabilizes the cubic fluorite structure and improves the oxygen‐ion conductivity of the zirconia. Y3+ Zr4+ O2– ZrO2 Y2O3 Oxygen vacancy YSZ (yttria-stabilized zirconia) cubic fluorite structure Figure 9.2 Structure of yttria‐stabilized zirconia. 4 In semiconductor physics, a donor atom is one that, when introduced into a semiconductor crystal, increases the density of electrons to create a so‐called ‘n‐type’ region. The electron density increases because the donor atom has more electrons in its outer shell than the metal that it is replacing. Conversely, an acceptor atom creates a deficiency of electrons (known as holes) in a so‐called ‘p‐type’ region. Both donors and acceptors enhance the electronic conductivity of the material. By analogy, a positively charged ion, for example, Y3+, is an acceptor dopant for zirconia because it increases the number of oxygen‐ion vacancies in the crystal lattice. 239 240 Fuel Cell Systems Explained Zirconia is extremely stable in both the reducing and oxidizing environments, which are to be found at the anode and cathode regions, respectively, and is a fast oxygen‐ion conductor above about 700°C. The ability to conduct O2− ions is due to the replacement of some Zr4+ ions with Y3+ ions in the fluorite structure. When this exchange of ions occurs, a number of oxygen‐ion sites become vacant because three O2− ions replace four O2− ions. Transport of oxygen ions occurs between vacancies located at tetrahedral sites in the lattice, a process that is now well understood at both the atomic and the molecular levels. The ionic conductivity of YSZ (0.02 S cm−1 at 800°C and 0.1 S cm−1 at 1000°C) is similar to that of liquid electrolytes. Moreover, YSZ can be made very thin (25–50 µm) and thereby ensures that the ohmic loss in the SOFC is comparable with that of other types of fuel cell. A small amount of alumina may be added to the YSZ to improve its mechanical stability, and tetragonal phase zirconia has also been used likewise to strengthen the electrolyte structure so that even thinner layers may be fabricated. The oxygen‐ion conductivity of zirconia can also be influenced by doping. The atomic number, ionic radius and concentration of the dopant are all found to influence the conductivity. Among the materials that have been investigated, scandium has frequently been promoted as a dopant for YSZ, because the ionic radius of the Sc3+ ion in scandium oxide is 0.87 Å and therefore is close to that of the Zr4+ cation in cubic zirconia (0.84 Å). Although doping YSZ with a small quantity of Sc2O3 does improve its ionic conductivity, the long‐term performance of this material does not match that of un‐doped YSZ. Consequently, scandia is normally ruled out as a dopant on the grounds of cost. An alternative approach to enhancing the ionic conductivity and performance of zirconia is to employ a thin film of the electrolyte. Thicknesses as low as 40 µm can be obtained by electrochemical vapour deposition (EVD), as well as by tape casting and other ceramic processing techniques. The EVD process was pioneered by Westinghouse Electric Corporation to produce thin layers of refractory oxides suitable for the electrolyte, the anode and the interconnection employed in the tubular SOFC design (v.i.). Now, however, the technique is only applied for fabrication of the electrolyte for tubular SOFCs. The metal chloride vapour to form the electrolyte is introduced on one side of the tube surface and an oxygen–steam mixture on the other side. The gas environments on both sides of the tube act to form two galvanic couples. The net result is the formation of a dense and uniform metal oxide layer on the tube. The deposition rate is controlled by the diffusion rate of ionic species and the concentration of electric charge carriers. 9.2.2 9.2.2.1 Electrolytes for IT‐Cells Ceria Zirconia‐based electrolytes are suitable for SOFCs because they exhibit pure anionic conductivity. Some materials, such as ceria (CeO2) and bismuth oxide (Bi2O3), also adopt the same fluorite crystal structure. Both of these oxides have greater oxygen‐ion conductivities than YSZ, but they are less stable at the low partial pressures of oxygen that pertain at the anode. This condition gives rise to the formation of defects in the oxides with concomitant increase in the electronic conductivity that, in turn, lowers the cell voltage. Solid Oxide Fuel Cells The ionic conductivity of cerium oxide (CeO2), also known as ceria, can be increased by doping with gadolinium and can reach a level comparable with that of YSZ at 600°C.5 The doped ceria is known variously in the literature as gadolinium‐doped ceria (GDC), gadolinia‐doped ceria and cerium‐gadolinium oxide (CGO). Samarium oxide (Sm2O3), also known as samaria, can equally be used as a dopant to produce samarium‐doped ceria (SDC) also referred to as samaria‐doped ceria and as cerium‐samarium oxide (CSO). It is the similar ionic radii of Gd3+/Sm3+ and Ce4+ that gives rise to the enhanced ionic conductivity of GDC and SDC materials. The main challenge with employing doped ceria is the reduction of Ce4+ to Ce3+ with the onset of electronic conduction in reducing conditions at temperatures greater than about 650°C. Further increase in the reduction of Ce4+ to Ce3+ can also lead to lattice expansion and the development of microcracks in the electrolyte. For bismuth oxide, certain dopants can improve the chemical stability and enhance ion conductivity, examples being lanthanum, vanadium, copper and zinc. Vanadium‐ and copper‐doped bismuth oxide, for example, has a conductivity that is greater than that of doped ceria at temperature above about 600°C, but it becomes less stable as the temperature is increased. 9.2.2.2 Perovskites As well as oxides with a cubic fluorite structure, perovskites also have a crystal structure in which oxygen‐ion transport is favoured. The cubic perovskite structure is represented by the general formula ABO3, where an A‐site ion on the corners of the lattice is usually an alkaline earth or a rare‐earth element and the B‐site ions in the centre of the lattice are 3d, 4d or 5d transition metal elements. Either the A or B cation can be substituted by introducing other cations of the same or different valency. Perhaps the most notable of the perovskite structures that have been used as electrolytes is that of lanthanum gallate, illustrated in Figure 9.3. This material, doped with strontium in the A‐sites and magnesium in the B‐sites, i.e., LaSrGaMgO3, was first reported in 1994.6 Lanthanum gallate is a superior oxide‐ion electrolyte that exhibits pure ionic conductivity over a very wide range of oxygen partial pressures (10−20 < pO2 < 1). In the form of strontium–magnesium‐doped lanthanum gallate (LSGM) at 800°C, it provides a performance that is comparable with that of YSZ at 1000°C, as shown in Figure 9.4. Unfortunately, the performance declines sharply at lower temperatures as shown in Figure 9.5. On the other hand, LSGM is also attractive as an electrolyte because it is compatible with a variety of active cathode materials; hence excellent electrochemical performance has been reported. The challenges with employing LSGM are the difficulty in making single‐phase materials and improving both its chemical and mechanical stabilities. Gallium is volatile at elevated temperatures, and lanthanum can react with nickel (as used in the conventional SOFC anode) to produce an ionically insulating LaNiO3 phase. Electrolytes based on gadolinium‐doped ceria, samarium‐doped ceria, LSGM or other materials such as bismuth metal vanadium oxide (BIMEVOX, Bi4V2O11) have enabled researchers to focus their attention on IT‐SOFCs and explore a wide range of materials and fabrication methods for the electrodes. 5 The ionic conductivity of Ce0.9Gd0.1O1.95 is 0.025 Ω−1 cm−1 at 600°C compared with 0.005 Ω−1 cm−1 for YSZ. 6 Ishihara, T, Matsuda, H and Takat, Y, 1994, Doped LaGaO3 Perovskite type oxide as a new oxide ionic conductor, Journal of the American Chemical Society, vol. 116, pp. 3801–3803; also Feng, M and Goodenough, JB, 1994, European Journal of Solid State and Inorganic Chemistry, T31, pp. 663–672. 241 Fuel Cell Systems Explained (a) (b) Figure 9.3 Two representations of the cubic perovskite structure of LaGaO3; (a) La‐centred unit cell; (b) corner‐shared GaO6 octahedra (Ga in the centre surrounded by 6 atoms) with La centred on 12 coordinate sites. Large red spheres = O2− ions; light green spheres = La3+ ions; small blue spheres = Ga3+. (Source: Ishihara 1994. Reproduced with permission of American Chemical Society.) Temperature/°C 800 700 600 500 400 300 Stainless steel Cr-Fe(Y2O3), inconel-Al2O3 Bipolar plate material La(Ca)CrO3 0 Log σ (S cm–1) 242 1500 µm R0 = L/σ = 0.15 Ω cm2 Bi2 V –1 150 µm 0.9 Cu 0.1 O 5.35 –2 Self-supported electrolytes 15 µm La Ce 0.9 Gd (Z 0.9 S r0 rO 0.1 O . G 1.95 1 2) a 0.9 0.8 M (Y g0 2O .2 O 3 )0 2.8 5 .1 –3 –4 0.8 1.0 1.2 1.4 1.6 1.8 0.5 µm Supported electrolytes 0.15 µm 2.0 1000/Temperature/K–1 Figure 9.4 Specific conductivity versus reciprocal temperature for selected solid oxide electrolytes. The conductivities of some SOFC electrolyte materials are compared in Figure 9.4. To ensure that the total internal resistance of the cell (i.e., of the electrolyte and electrodes) is sufficiently small, the target value for the area‐specific resistance (ASR)7 of the electrolyte in Figure 9.4 is set at 0.15 Ω cm2. Films of oxide electrolytes can be produced reliably using cheap, conventional ceramic fabrication routes at thicknesses down to approximately 7 The area‐specific resistance (ASR) of a fuel cell is its resistance normalized by its area, ASR = R × A, and is measured in Ω cm2. An ASR is more fundamental than a resistance because fuel cells are compared on a per‐area basis. The ohmic losses can be found by multiplying a current density by the ASR. Good fuel cells target an ASR of 0.1 Ω cm2. Solid Oxide Fuel Cells 1200 1100 Cell voltage/mV 1000 900 800 700 800°C 600 500 600°C 650°C 700°C 750°C 400 0 200 600 400 800 –2 Current density/mA cm Figure 9.5 Typical single‐cell performance of LSGM electrolyte (500 µm thick) over a range of temperatures. 15 µm. It follows that to achieve the aforementioned target for the ASR, the specific conductivity of the electrolyte must exceed 10−2 S cm−1. This requirement can be met at 500°C for the Ce0.9Gd0.1O1.95 electrolyte and at 700°C for the (ZrO2)0.9(Y2O3)0.1 electrolyte. Although the Bi2V0.9Cu0.1O5.35 electrolyte exhibits higher conductivity, it is not stable in the reducing environment imposed by the fuel in the anode region of an SOFC. Other materials based on perovskite‐related structures that should be mentioned include Ln2NiO4+δ (where Ln = lanthanum, neodymium or praseodymium). Although Ln2NiO4+δ materials are mixed ionic and electronic conductors, there is expectation that it may be possible to suppress their electronic conductivity and thereby open up a new range of prospective electrolyte materials. 9.2.2.3 Other Materials In the past few years, two new classes of oxide‐ion conductors have emerged that may overtake doped ceria and perovskite materials for IT‐SOFC electrolytes. The first is lanthanum molybdenum oxide (LAMOX, La2Mo2O9). This material exists in several phases, and a high‐temperature cubic form possesses exceptionally good oxygen‐ion conductivity in comparison with most of the other perovskite materials. Apatite‐type oxides are the other class of alternative electrolyte materials. The general formula of apatite oxides can be written as M10(XO4)6O2±y where M is a rare‐earth or alkaline earth cation; X is a p‐block element such as phosphorous, silicon or germanium; and y is the amount of oxygen non‐stoichiometry. These oxides have the same structure as hydroxyapatite materials found in bones and teeth. The apatites that incorporate silicon or germanium are among the most promising of the new fast‐ion conductors, as shown in Figure 9.6. 9.2.3 9.2.3.1 Anodes Nickel–YSZ In the nickel–YSZ cermet that forms the anode of state‐of‐the‐art HT‐SOFCs, the porous YSZ serves to inhibit the coarsening or sintering of the metal particles (that otherwise would lead to loss of surface area) and provides the anode with a thermal 243 Fuel Cell Systems Explained –1 Log σ (S cm–1) 244 –2 YSZ Bi2V0.9Cu0.1O5.35 Cerium-gadolinium oxide –3 Ge-apatite Silicon-apatite –4 LSGM LAMOX –5 0.8 1.0 1.2 1.4 1.6 1000/Temperature/K–1 Figure 9.6 Total conductivities of several well‐known oxide‐ion conductors as a function of reciprocal temperature: CGO, Ce0.8Gd0.2O1.9; Ge‐apatite, La10(GeO4)6O3; LAMOX, La2Mo2O9; LSGM, La0.9Sr0.1Ga0.8 Mg0.2O2.85; Si‐apatite, La10(SiO4)6O3 and YSZ, (ZrO2)0.92(Y2O3)0.08. expansion coefficient similar to that of the electrolyte. At least 30 vol.% of nickel is generally used in the cermet in order to achieve adequate electronic conductivity while maintaining the required porosity so that mass transport of reactant and product gases is not compromised. Nickel and YSZ are non‐reactive over a wide range of temperature and are also essentially immiscible in each other. Both these properties simplify the synthesis of YSZ cermets and allow the preparation through conventional sintering of NiO and YSZ powders. Reduction of NiO to Ni in situ leads to a highly porous (20– 40%) YSZ structure that contains connected Ni particles on the surface of the YSZ pores. The structure provides a substantial three‐phase boundary for oxidation of fuels together with paths for electronic conduction from the reaction zone to the SOFC current-collector. Traditionally, the anode is generally prepared by painting an ink made from the NiO and YSZ powders directly onto the YSZ electrolyte tube or plate, according to the cell configuration required. The painted electrolyte is dried and then sintered at approximately 1400°C in air during which the aforementioned porous anode structure forms as the result of the growth of the NiO and YSZ particles. Sintering at such an elevated temperature also produces an excellent bond between the resulting porous anode layer and the dense YSZ electrolyte layer. Anode‐supported electrolyte cells are a relatively recent innovation for the IT‐SOFC. In these cells, the anode is prepared by extruding or pressing NiO and YSZ powders into the required shape, which is then dried sintered to produce a relatively thick plate that can support a thin layer of electrolyte. Powdered electrolyte material is painted onto the anode and the two layers are co‐fired as before. Factors that affect the performance of a Ni–YSZ cermet anode include the properties of the starting materials, the sintering temperature and the nickel content. In addition, modification of the microstructure and doping with various materials can improve long‐term performance by lowering the degree of sintering of nickel particles. To this Solid Oxide Fuel Cells end, MgO, TiO2, Mn3O4 and Cr2O3 have all been shown to reduce nickel sintering and also to act as anchor sites for nickel at the anode|electrolyte interface. As noted previously, hydrocarbon fuels can be reformed directly on the SOFC anode. While this feature carries advantages in terms of greater system efficiency, there are risks involved: notably, the formation of carbon either by the Boudouard reaction or through cracking (pyrolysis) on the nickel (see Section 10.4.4, Chapter 10). Investigations have shown that the propensity for carbon formation on an SOFC anode can be reduced by doping the cermet with molybdenum or gold. There is some ohmic loss at the interface between the anode and the electrolyte, and several developers have investigated bilayer anodes in which a small amount of ceria is added to the Ni–YSZ cermet. This improves the tolerance of the anode to temperature variations and to cycling between oxidation and reducing conditions on the surface (changing the anode gas from a reducing fuel gas to an oxidizing gas and vice versa). Control of the particle size of the YSZ can also improve the stability of the anode towards oxidation and reducing conditions. 9.2.3.2 Cathode Ceria itself possesses a substantial surface concentration of O2− ions that renders it a good catalyst for hydrocarbon oxidation. By doping ceria with certain rare‐earth metals, the catalytic activity for oxidation is enhanced. In addition, the doped ceria has a fluorite structure with not only conductivity for oxygen ions but also some electronic conductivity. Consequently, ceria has been pursued as an anode material for SOFCs in its own right. Although doped ceria has a good ionic conductivity, the electronic conductivity is relatively low, and therefore the material has most commonly been employed in the form of a cermet. The combination of doped ceria and metal provides excellent ionic conductivity at low temperature and a high electronic conductivity, and therefore the material is an ideal candidate for IT‐SOFC applications. Gadolinia‐doped ceria combined with nickel to form a Ni–GDC cermet has been shown to have high electrochemical activity towards the steam reforming of methane at temperatures as low as 500°C without coking of the catalyst. The principal problem with the material is the possibility of mechanical degradation through expansion of the crystal lattice, which results from the bulk transition of Ce4+ to Ce3+ that may occur in a low oxygen partial‐pressure environment. This is of particular concern when a YSZ electrolyte is used, as delamination at the electrolyte|electrode interface could occur. Doping the ceria with cations of low valency such as Sm3+, Y3+ or Gd3+ may help to prevent the degradation, and doping a ceria anode with 40–50 at.% of gadolinia, e.g., Ce0.6Gd0.4O1.8, can provide a good compromise between electronic conductivity and mechanical stability. Doped‐ceria cermets have been shown to sustain multiple, rapid thermal cycles and full oxidation–reduction cycles without a decline in performance. They can be employed in both HT‐SOFCs and IT‐SOFCs. Reaction between the cermet and YSZ electrolyte in the HT‐SOFC can be prevented by inserting a thin interlayer of ceria doped with both gadolinia and zirconia between the electrolyte and the ceria cermet anode. A problem with using cermets that incorporate ceria in conjunction with a ceria electrolyte is the reduction of Ce4+ to Ce3+, mentioned previously, and the ensuing increase in electronic conductivity. For this reason some developers insert a thin interlayer of ion‐conducting oxide material at the junction between the anode and the electrolyte of an IT‐SOFC. Alternatively the surface of the anode that bonds with the 245 246 Fuel Cell Systems Explained electrolyte can be functionalized to limit its electronic conductivity. Both approaches protect the ceria electrolyte from the reducing conditions at the anode. When using hydrocarbon fuels, the nickel in anode cermets, while providing good electronic conductivity, suffers the disadvantage of promoting carbon formation. Nickel is also poisoned by sulfur. Both issues are a concern if internal reforming of hydrocarbons is undertaken. To eliminate these problems, nickel can be replaced by copper in ceria cermets doped with rare earths, as described in Section 10.4.6, Chapter 10. A composite anode consisting of copper, ceria and YSZ has shown activity for direct electrochemical oxidation of hydrocarbons without any performance degradation by carbon deposition. When compared with a Ni–YSZ anode, the composite shows superior performance with both carbon monoxide and hydrocarbon fuels. The material can be improved further by incorporating noble metals such as platinum, rhodium or palladium. 9.2.3.3 Mixed Ionic–Electronic Conductor Anode With the development of electrolyte materials for IT‐SOFCs, attention has been given to alternative anode materials to the conventional cermets. Certain perovskites possess both electronic and ionic conductivities. Such materials are known as ‘mixed ionic–electronic conductors’ (MIECs). Examples that have been investigated include titanates such as strontium titanate (SrTiO3) doped with either yttrium or lanthanum, A‐site deficient lanthanides doped with neodymium and praseodymium and iron‐doped barium titanate. Perovskites based on chromites have also been examined, e.g., strontium‐doped lanthanum chromite doped with transition metals such as Co, Cu, Ni and Mn. Vanadium complexes and ferrites have shown some promise as anodes, e.g., strontium‐doped lanthanum vanadate La1–xSrxVO3 (denoted as LSV), as have ferrites such as lanthanum strontium cobaltite ferrite (LSCF). Needless to say, the search for effective and stable SOFC anodes is a very active field of research. Apart from the elimination of the metal (e.g., nickel) and therefore a reduction in the chance of carbon deposition, MIEC materials provide a means of extending the three‐phase boundary between anode and electrolyte, as shown in Figure 9.7. (a) Electronically conducting cermet (b) Mixed ionic–electronic conductor H2 H2 H2O H2O O= e– Electrolyte e– Particle of a pure electronic conductor Three-phase boundary region O= O= e– H+ H+ Electrolyte e– Particle of a mixed ionic and electronic conductor Figure 9.7 (a) Illustration of the three‐phase boundary regions of different SOFC anode materials. (b) Extension of the boundary is obtained with mixed conducting cathode materials. Solid Oxide Fuel Cells 9.2.4 Cathode The very first SOFCs used platinum for the cathode, but as soon as electronically conducting ceramics became available, these were favourably received. The first to attract serious attention was the perovskite lanthanum cobaltite (LaCoO3). Initially, this compound exhibited a good performance, but longer tests showed that it reacted with the YSZ electrolyte to cause permanent degradation. Although the ionic conductivity could be improved by doping with strontium, the LaCoO3 remained reactive towards YSZ. It is non‐reactive with doped‐ceria electrolytes but has a poor match with the expansion coefficient of YSZ. Other cobaltite materials were investigated and included gadolinium strontium cobaltite Gd1‐xSrxCoO3 that was also suitable for use with ceria electrolytes. After exploring the cobaltites for cathode materials, researchers moved to the investigation of manganites, and LSM emerged as the preferred material. Despite being susceptible to some reaction with YSZ, particularly at elevated temperatures, strontium‐doped LSM has become the most widely used cathode material for the SOFC over the past 20 years for both HT‐ and IT‐SOFCs. As with many perovskites, the crystal structure of LSM undergoes a phase change from an orthorhombic structure at room temperature to a rhombohedral form above 600°C. The mixed ionic–electronic conductivity of LSM can be enhanced by replacing some of the A‐sites with Sr2+. It is particularly notable that the defect structure and oxygen non‐stoichiometry of LSM depend on the applied partial pressure of oxygen. The fact that LSM can have oxygen excess or oxygen deficiency is unusual in comparison with most perovskite oxides. The material is stable over a wide range of oxygen partial pressures, but at very low levels, it can decompose to form two phases, namely, La2O and MnO. Although LSM has proven itself for most HT‐SOFCs, other materials have also been found suitable for the cathode, particularly those with perovskite structures that exhibit mixed ionic and electronic conductivity. Mixed conductivity is especially important for operation at lower temperatures since the cathode overpotential increases significantly with reduction of the SOFC temperature. The advantages of using oxides with mixed conductivity become especially noticeable in cells that operate at 650°C. Lanthanum strontium ferrite and lanthanum strontium cobaltite, which are all n‐type semiconductors, are better electrocatalysts than the state‐of‐the‐ art lanthanum strontium manganite at IT‐SOFC temperatures. 9.2.5 Interconnect Material The interconnect is the means by which neighbouring fuel cells are joined together. In the planar design of cell, this is the bipolar plate, but the arrangement is different for tubular geometries, as will be described in Section 9.3.1. Doped lanthanum chromite has been the material of choice for the interconnect in HT‐SOFCs, notably in the Westinghouse Electric tubular stacks. The electronic conductivity of pure chromite is very low but may be enhanced by substituting ions of lower valency (e.g., calcium, magnesium, strontium) on either the La3+ or the Cr3+ sites of the lanthanum chromite lattice. Unfortunately, the material has to be sintered at quite high temperatures (1625°C) to produce a dense phase. This requirement exposes one of the major problem areas with the SOFC, namely, the means of fabrication. All of the cell components must 247 248 Fuel Cell Systems Explained be compatible with respect to chemical stability and mechanical compliance (e.g., they must all have similar thermal expansion coefficients). The various layers have to be deposited in such a way that good adherence is achieved without degrading the material through the use of too high a sintering temperature. Many of the methods of fabrication are proprietary, and considerable research is being conducted on the processing of SOFC materials. In cells intended for operation at lower temperatures (<800°C), it is possible to use oxidation‐resistant metallic alloys for the interconnects. Ferritic steels are currently the preferred option. Compared with lanthanum chromite ceramic, metallic alloys offer advantages such as improved manufacturability, significantly lower raw material and fabrication costs and superior electrical and thermal conductivity. Alloys, such as the Cr‐5Fe‐Y2O3 Plansee material and Crofer22 APU (a high‐temperature ferritic stainless steel developed by VDM Metals GmbH), have been engineered with thermal expansion coefficients to match those of SOFC ceramic components. Unfortunately, under high oxygen partial pressure at the cathode, the chromium in such alloys has a tendency to vapourize and deposit at the LSM–YSZ three‐phase boundary and thereby cause permanent poisoning of the cathode. Another advantage for IT‐SOFCs is that cheaper low‐nickel materials, such as austenitic steels, may also be used in the manufacture of cells. 9.2.6 Sealing Materials A key issue with SOFCs, particularly planar configurations, is the method of sealing the ceramic and metal components to obtain gas tightness. The wet‐seal arrangement employed in the MCFC cannot be used, and the wide temperature range of SOFC operation poses particular problems. The seal needs to be thermochemically stable and compatible with other cell components; it should also withstand thermal cycling between room temperature and the operating temperature. To satisfy such requirements, a number of different sealing approaches have been developed — some more successful than others — and include rigid bonded seals (e.g., glass ceramics and brazes), compliant seals (e.g., glasses) and compressive seals (e.g., mica‐based composites). The most common practice has been to use glasses that have transition temperatures close to the operating temperature of the cell. These materials soften as the cells are heated up and form a seal all around the cell. Glass seals are employed in planar stack designs in which, for example, a number of cells may be assembled in one layer (v.i.). A particular problem associated with the use of glass seals is the migration of silica from the glass, especially onto the anodes with resultant degradation in cell performance. Glass ceramics have been adopted for all‐ceramic stacks, but migration of the silica component can still be a problem on both the anode and cathode sides. Brazes have not been featured strongly as sealants because of the ease by which suitable metals can oxidize at the elevated temperatures necessary for carrying out the brazing operation (i.e., above 800°C). Oxidation can be reduced by adding a metal such as titanium or zirconium to the brazing metal, but the need for a reducing atmosphere during the brazing process can reduce the activity of cathode components. Solid Oxide Fuel Cells Compliant compressive seals (i.e., gaskets) have had limited application in SOFCs as most of the appropriate metals tend to oxidize or deform excessively at the operating temperature under a sustained compressive force. This is usually less of a problem with short‐term laboratory tests where, for example, gold gaskets have been routinely used. Mica composite materials of greater resilience have been employed for single‐cell tests and may find application in future planar stack configurations. 9.3 Practical Design and Stacking Arrangements 9.3.1 Tubular Design The tubular SOFC was pioneered in the United States by the Westinghouse Electric Corporation (now the Siemens Westinghouse Power Corporation). The original design used a porous CSZ support tube of 1–2 mm thickness and about 20 mm internal diameter, onto which the cylindrical anodes were deposited. By a process of masking, the electrolyte, the interconnect and finally the fuel electrode were deposited on top of the anode. The procedure was reversed in the early 1980s so that the air electrode became the first layer to be deposited on the zirconia tube, and the fuel electrode was on the outside. This tubular fuel cell became the standard for the next 15 years. From its inception, the tubular design has suffered from the major problems such as low‐power density and punitive fabrication costs. The low‐power density resulted from both the long path for the electrical power through each cell, as depicted in Figure 9.8, and the large voids within the stack structure (i.e., between the tubes). The unfavourable costs arise from the preparation of the electrolyte and electrode via electrostatic vapour deposition, which is a batch process conducted in a vacuum chamber. More recently, the zirconia support tube has been replaced by the one made by the extrusion of a porous form of doped LSM onto which the electrolyte is deposited by EVD, followed by plasma spraying of the anode. The arrangement is known as an ‘air‐ electrode supported’ (AES) fuel cell; an example is shown in Figure 9.9. One significant advantage of the tubular design of SOFC is that high‐temperature gas‐tight seals are eliminated, as illustrated in Figure 9.10. Each fuel‐cell tube is fabricated in the form of a large test tube, sealed at one end. Fuel flows along the outside of the tube (anode side) towards the open end. Air is fed through a thin alumina supply tube that is located centrally inside each tubular fuel cell. Heat generated within the cell brings the air up to the operating temperature. The air then flows through the fuel cell back up to the open end. At this point, air and unspent fuel from the anode exhaust are instantly combusted so that the cell exit is above 1000°C. This combustion provides additional heat to preheat the air supply tube. Thus, the tubular SOFC has a built‐in air preheater and anode exhaust gas combustor, as well as no requirement for high‐ temperature seals. Moreover, by allowing imperfect sealing around the fuel‐cell tube, some recirculation of anode product gas (which contains both steam and CO2) occurs and thereby allows internal reforming of fuel gas on the SOFC anode. The current SOFC tubes produced by the Siemens Westinghouse Power Corporation are 150 cm in length and 2.2 cm in diameter. The voltage versus current density and power versus current density characteristics of a single‐tube cell at 900, 940 and 1000°C and operating with 89 vol.% H2 + 11 vol.% H2O fuel (85% fuel utilization) and air as 249 250 Fuel Cell Systems Explained More cells Fuel Fuel Electrolyte Air Typical path of an electron from the anode of one cell to the cathode of the next Cell interconnect Fuel Fuel Anode Air Cathode More cells Figure 9.8 End view of tubular SOFC produced by Siemens Westinghouse. Both electrolyte and anode are built onto the air cathode. Interconnect Electrolyte Cathode Fuel flow Cell bundle of 3 × 8 tubes Airflow Anode Figure 9.9 Small stack of 24 tubular SOFCs. Each tube has a diameter of 22 mm and is about 150 cm long. (Photograph reproduced by kind permission of Siemens Westinghouse Power Corporation). Solid Oxide Fuel Cells This seal or joint is quite straightforward to maintain Exhaust Air Fuel Combustion Deliberately ‘imperfect’ seals around the tube Recirculation A i r Tubular fuel cell Fuel A i r Internally reformed using product steam Figure 9.10 Diagram showing how the tubular‐type SOFC can be constructed with (almost) no seals. oxidant are presented in Figure 9.11. Such tubular cells have a power density at 1000°C of about 0.2 W cm−2, i.e., much lower than that of planar cells. Individual tubular cells are arranged in series–parallel stacks of 24 tubes, as displayed in Figure 9.9. Several other organizations — notably, Mitsubishi Heavy Industries and TOTO Ltd. in Japan, Adelan Ltd. in the United Kingdom and both Acumentrics and Watt Fuel Cell in the United States — have been developing tubular SOFC designs. To avoid the expensive EVD process, TOTO Ltd. has adopted wet sintering as a method of cell fabrication. 9.3.2 Planar Design Alternatives to the tubular SOFC have been pursued for several years, especially several types of planar configuration and a monolithic design. In the planar designs, cells are thin flat plates that are electrically connected to achieve the required stack voltage and current. The earliest planar cells featured the electrolyte as the support onto which the electrodes were deposited. More recently and especially for IT‐SOFCs, the anode‐ supported structure is preferred in which the anode and the electrolyte are deposited directly onto a metal bipolar plate and above the anode, respectively. There are several 251 Fuel Cell Systems Explained 1.2 800 1.0 600 0.9 0.8 400 0.7 800 750 700 650 600 0.6 0.5 0.4 0 200 400 600 800 Current density/mA cm–2 1000 Power density /mW cm–2 1.1 Voltage/V 252 200 0 1200 Figure 9.11 Influence of temperature on the performance of Siemens Westinghouse tubular fuel cells. variants of the planar design. In one example, promoted by the Swiss company Sulzer Hexis AG, the SOFC is in the form of a circular disc fed with fuel from the central axis, whereas in another design, preferred by Siemens AG and others, the cell employs a square plate that is fed from the edges. Planar designs offer several prospective advantages that include simpler and less expensive manufacturing processes and power densities that are better than those achievable with tubular SOFCs. Unlike tubular cells, however, planar designs require high‐temperature gas‐tight seals between the components in the SOFC stack. Sealing therefore remains one of the most significant technical barriers to the commercialization of planar SOFCs. Also of concern are the thermal stresses at the interfaces between the different cell and stack materials that can cause mechanical degradation. Particularly challenging is the brittleness of thin planar SOFCs in tension. Thermal cycling is a further problem for the planar SOFC; by contrast, the tubular cell is thermally more robust. Finally, the issue of thermal stresses and the fabrication of very thin components place a major constraint on the size of planar SOFCs. Early configurations employed a thick electrolyte as the support, which required an operating temperature that was often taken above 900°C to achieve adequate current densities. Advances in ceramic processing have allowed reproducible fabrication of thin electrolytes, i.e., 10 µm or thinner, by low‐cost conventional ceramic processing techniques such as tape casting, tape calendaring, slurry sintering and screen printing or by plasma spraying. The ability to create thin electrolytes led to the interest in anode‐ supported cells. For many years, the maximum size of single planar SOFCs was 5 cm × 5 cm. Now larger cells can be routinely manufactured and assembled into a stack of much greater area by building them into a window‐frame arrangement such that several cells are located in one layer. An example arrangement that uses four cells in each layer is shown in Figure 9.12. Solid Oxide Fuel Cells Figure 9.12 Stack with four 10 × 10 cm2 SOFCs in one layer (window‐frame design FY520). (Source: From Blum, L, Batfalsky, P, Fang, Q, deHaart, LGJ, Malzbender, J, Margaritis, N, Menzler, NH and Peters, R, 2015, SOFC stack and system development at Forschungszentrum Jülich, Journal of the Electrochemical Society, vol. 162(10), pp. F1199–F1205.) 9.4 Performance When hydrogen is the fuel, the OCV of the SOFC is lower than that of the MCFC and PAFC. Nevertheless, at the operating temperature of the SOFC, the overpotential at the cathode is much lower and thereby provides the technology with a superior operational voltage. The voltage losses in SOFCs are a function mainly of the resistance of the cell components, which include those associated with current collection. The contributions to ohmic voltage loss in a tubular cell are typically some 45% from the cathode, 18% from the anode, 12% from the electrolyte and 25% from the interconnect, when these components have a thickness of 2.2, 0.1, 0.04 and 0.085 mm and resistivities at 1000°C of 0.013, 3 × 10−6, 10 and 1 Ω cm, respectively. With a tubular SOFC, the ohmic loss at the cathode dominates despite the greater resistivities of both the electrolyte and the cell interconnection. This situation arises because the conduction path through these latter two components is much shorter than the current path in the plane of the cathode, as demonstrated in Figure 9.5. With planar cells, the long current path is absent, and therefore it is possible to achieve higher power outputs from the stacks. As with other types of fuel cell, SOFCs show enhanced performance with increase in cell pressure. Unlike low‐ and medium‐temperature cells, however, the improvement is mainly due to the increase in the Nernst voltage. In Section 2.5.4, Chapter 2, it was shown that the voltage change for an increase in pressure from P1 to P2 follows very closely the theoretical equation V 0.027 ln P2 P1 (2.45) The relationship was borne out in practice when Siemens Westinghouse in conjunction with Ontario Hydro Technologies tested AES tubular cells at pressures of up to 15 atm (1.52 MPa) on both hydrogen and natural gas. Operation at such pressures is particularly advantageous when using the SOFC in a combined‐cycle system with a gas turbine. In other cases, as with the proton‐exchange membrane fuel cell (PEMFC), the power costs involved in compressing the reactants render the benefits marginal. 253 254 Fuel Cell Systems Explained The temperature of an SOFC has a very marked influence on its performance, though the details will vary greatly between cell types and the materials used. The improvement in the performance of the Westinghouse tubular fuel cells on increasing the temperature from 900 to 1000°C is demonstrated by the data given in Figure 9.11. The predominant effect is that high temperatures increase the conductivity of the materials, and this reduces the ohmic losses within the cell that, as discussed in Section 3.65 Chapter 3, are the most important type of loss in the SOFC. For SOFC–combined‐cycle and hybrid systems, it is beneficial to maintain a high operating temperature. For other applications, such as cogeneration and possible transport applications (e.g., as an auxiliary power supply for vehicles), it is more beneficial to operate at lower temperatures, as the higher temperatures bring material and construction difficulties. Unfortunately, as Figure 9.11 clearly shows, for a given set of cell components, the performance decreases substantially for SOFCs as the temperature is lowered. 9.5 Developmental and Commercial Systems Reference was made in Chapter 8 to the significant MCFC development programmes in Europe and Japan that have been abandoned despite the investment of substantial public and private funding over many years. A similar theme has been seen with SOFC programmes. In 1977, the US Department of Energy commenced its funding for the development of the Westinghouse Electric tubular fuel cell. Given the issues described in the previous section (e.g., a long current path that led to low stack power and low power-density), Westinghouse made good progress throughout the 1980s, and the company believed that commercialization would be possible by the early 1990s. A number of set‐backs, which were partly technical and partly due to restructuring of the company at the turn of the century, slowed down the progress of the technology, but work continued until 2010. By that time, as noted previously, Siemens had taken over the fuel‐cell interests of Westinghouse Electric (with the creation of the Siemens Westinghouse Power Corporation), and, after many reviews of their progress, it was decided that SOFC development was not the core business for the parent company (Siemens in Germany). Consequently, further activities were stopped and the fuel‐cell enterprise was put up for sale. More details of the history of Siemens and various SOFC developers are catalogued by Behling8. Continued interest in the technology was assisted by the US Department of Energy, which in 2001 instigated a programme that focused on planar SOFCs. Known as the Solid State Energy Conversion Alliance (SECA), the programme was intended to develop an SOFC system that would cost as little as US$400 per kW by 2010. Whereas on most counts the SECA programme failed to live up to expectations, it did serve to restimulate activity in SOFCs, especially for planar systems — both in the United States and around the world. The present status of the SOFC is very different from that of the MCFC. In the case of MCFC development, few players are left, whereas for the SOFC many are active in 8 Behling, N, 2023, History of solid oxide fuel cells, in Fuel Cells, Current Technology Challenges and Future Research Needs, Elsevier, Amsterdam. ISBN 9780444563255 Solid Oxide Fuel Cells research and demonstration as well as commercialization, and, moreover, several of the participants are relatively new to the field. The following selected SOFC systems are currently under development or being commercialized. The examples are not intended to be an exhaustive list but to illustrate the wide variety of approaches that companies have taken. As well as Siemens Westinghouse Power Corporation, some other companies that appeared to be well advanced have pulled out of SOFC development in recent years. It therefore should be recognized that the industry is very fragile. The business landscape could change rapidly over the next few years as some research teams push forward and others fall from view. 9.5.1 Tubular SOFCs Despite the exit of Westinghouse from the fuel‐cell scene, there remain several developers of tubular SOFC. Most activities are focused on small systems for remote backup power, auxiliary power units for vehicles, and portable power supplies. The participating companies include Watt Fuel Cell, Acumentrics, Inc. and Protonex Technology Corporation in the United States and Adelan Ltd. in the United Kingdom. Acumentrics, Inc. was formed in 1994 to develop rugged uninterruptible power systems for use in harsh environments, and the company was involved in the SECA programme until 2010. Acumentrics has one of the most advanced tubular SOFC technologies that is perhaps closest in form to the Westinghouse tubular cells. The company has deployed over 350 remote power generators in a variety of sites throughout North America, and in 2015 the successful fuel‐cell activities were divested into a new venture — Atrex Energy, Inc. With a portfolio of four SOFC products that have outputs of 250, 500, 1000 and 1500 W, respectively, Atrex Energy has shipped systems to Europe and Asia as well as fulfilling a growing demand in North America. The technology differs from the original Westinghouse tubular cell in that fuel, rather than air, is fed through the centre of the tube, and therefore, the anode is on the inside of the tube and the cathode on the outside. Moreover, the anode (Ni–YSZ) extends the full internal length of the tube, whereas the cathode (LSM) is segmented into discrete partitions so as to enable efficient current collection from the anode at different stages. This arrangement avoids electrons having to flow throughout the whole length of the cell that, otherwise, would lead to greater ohmic losses. LaCrO3 wires coiled around the electrode partitions act as current-collectors; moreover, a tight coiling around the anode prevents an undesirable exposure to the air stream that could potentially result in reoxidation of the nickel matrix and, in turn, cause mechanical stress. A bundle of Atrex Energy/Acumentrics tubes is shown in Figure 9.13. Solid oxide fuel cells produced by Atrex Energy and others such as Protonex are intended for remote stationary systems and are fuelled with propane. Several companies in Japan have ongoing SOFC programmes. Kyocera Corporation and Mitsubishi Hitachi Power Systems (MHPS) have been in the fuel‐cell field for several decades and, with support from central government funds, have developed various different technologies. Kyocera has worked with other Japanese companies, for example, Osaka Gas Co., Ltd., to produce a ‘flat tubular’ SOFC that is comprised of a series of parallel tubes within one ceramic tablet. This robust configuration, shown in Figure 9.14, is used as the core fuel‐cell technology for a number of Japanese micro combined heat and power (CHP) units. The strategy of MHPS, by contrast, is to 255 256 Fuel Cell Systems Explained Figure 9.13 Acumentrics tubular SOFC bundle as used in the RP1500, 1.5‐kW system. (Source: Reproduced with permission of Acumentrics.) Figure 9.14 (a) Kyocera–Osaka gas power‐generation unit (left) and heating unit that uses SOFC exhaust heat (right) and (b) Kyocera flat tubular cells. (Source: Reproduced with permission of Kyocera.) implement SOFC stacks in combined‐cycle or hybrid systems, as described in Section 9.6. Both the Kyocera and MHPS stacks are designed to be fuelled with natural gas. Mitsubishi Heavy Industries, Ltd. (Nagasaki) and Tokyo Gas Co., Ltd. have also collaborated on tubular SOFC technology. In 2003, a ‘flat tubular’ SOFC built by NGK Insulators demonstrated a power density of 0.6 W cm−2 at 650°C and 1.6 W cm−2 at 750°C — a world record at that time. TOTO started a programme on tubular SOFCs in 1989 and by 2001 had produced 10‐kW stacks. The TOTO fuel cells were similar in appearance to those of Westinghouse tubular SOFCs, but the cell components were deposited using a wet chemical process rather than by EVD. 9.5.2 Planar SOFCs The organizations currently involved in the research, development and commercialization of planar SOFCs are too numerous to list here. Much of the early work Solid Oxide Fuel Cells was undertaken in the United States to be followed in the 1980s by that conducted in Europe and Japan. Some of the early players in the United States, such as Allied Signal Aerospace Co. and SofCo EFS,9 are no longer in the fuel‐cell business, whereas the others, such as Versa Power Systems, Inc., Ceramatec Inc. and Delphi Automotive LLP, which were supported under the SECA programme, are still active. In terms of commercialization status, the major planar SOFC provider in the United States is now Bloom Energy. This company was founded in 2001 with the name Ion America and is based in California (USA). It changed its name in 2006 after attracting US$400 million in investments. In the same year, Bloom shipped its first 5‐kW demonstration unit to the University of Tennessee, Chattanooga. After 2 years of field trials in Tennessee, California and Alaska, the first pre‐commercial prototype product was shipped to Google in July 2008. The company has continued increasing the size of their systems during the last few years through the production of the following models of Bloom’s Energy Server®: ES‐5000, ES‐5400 and ES‐5700 that generate 100, 105 and 210 kW, respectively. Each power generator is built up with 1‐kW stacks that are composed of 40 × 25‐W cells, fuelled with natural gas. Stacks are combined to provide a given power output, and the system is marketed as a ‘Bloom Box’. Many systems have been installed for eminent clients throughout the United States. In January 2011, Bloom started to offer a bold innovative service called ‘Bloom Electrons’ that would allow customers to purchase the electricity provided by the Bloom Box at a set price for 10 years without incurring any other costs. In Japan, developers of planar SOFCs include Fuji Electric, Co., Ltd.; Tokyo Gas, Co., Ltd.; Mitsubishi Heavy Industries, Ltd.; Mitsui Engineering and Shipbuilding, Co., Ltd.; Murata Manufacturing, Co., Ltd.; Sanyo Electric, Co., Ltd.; Tokyo Gas Co., Ltd.; and Tonen. In Europe, innovators of planar SOFCs with good track records have been a consortium of Haldor Topsoe A/S and Riso National Laboratory (Denmark), Forschungszentrum Jülich in collaboration with Sunfire GmbH (Germany), Wärtsilä Corporation (Finland), Hexis Ltd. (Switzerland) and Ceres Power (United Kingdom). There are of course many differences between the various technologies under development. Ceres Power, for example, is focused on an IT‐SOFC known as the ‘Steel Cell’. Whereas many planar cells use anode‐supported structures, the components in the Steel Cell are deposited directly onto a porous stainless‐steel substrate. Intended to be fuelled by natural gas, the cell operates in the range 500–600°C, is able to be started quickly and will endure thermal cycling better than ceramic‐supported SOFCs. Ceres Power is currently supplying stacks to original equipment manufacturers (OEMs) partners in the United Kingdom, Korea and Japan. The Rolls‐Royce ‘Integrated Planar’ SOFC is an especially distinctive technology. Work on the concept started in Derby, UK, in the mid‐1980s. From the outset, the aim was to target low‐cost manufacturing that could enable affordable MW‐class systems to be built. The Rolls‐Royce cells are screen printed onto a porous extruded ceramic substrate that gives a narrow cell pitch to reduce the ohmic losses; a schematic diagram of the arrangement is given in Figure 9.15. Steady progress was made in the United Kingdom during the 1990s, and by 2001 a 300‐mm, 40‐cell proof‐of‐concept module generated 27.3 W and achieved a power density of 155 mW cm−2 at an average cell 9 SOFCo EFS Holdings LLC was acquired by Rolls‐Royce Fuel Cells from McDermott International in 2007. 257 258 Fuel Cell Systems Explained Anode current-collector Cathode Air Microporous barrier layer YSZ Anode Interconnect Fuel Porous ceramic substrate Figure 9.15 Rolls‐Royce Integrated Planar SOFC concept. voltage of 0.62 V with 43% utilization of simulated reformed fuel gas. In June 2012, LG Electronics Inc., a South Korean company, acquired the Rolls‐Royce fuel‐cell business, which is now known as LG Fuel Cell Systems Inc. A final mention should be given to SOLIDpower S.p.A., an Italian company that was formed in 2006 following the acquisition of HTCeramix SA, a private company in Yverdon‐les‐Bains, Switzerland, that had commenced work on SOFCs four years earlier. The latter was a spin‐off of the Swiss Federal Institute of Technology in Lausanne (EPFL). In 2015, SOLIDpower S.p.A. also acquired the European assets and employees of Ceramic Fuel Cells Ltd. (CFCL), an Australian SOFC developer that, through lack of funding, had ceased trading earlier in the year. The company, which was a spin‐off in 1992 from the Commonwealth Scientific and Industrial Research Organisation (CSIRO), had produced for the domestic market a packaged CHP system (2.5 kWe, 2.0 kWth) — the ‘BlueGen’ — fuelled by natural gas. 9.6 Combined‐Cycle and Other Systems It has been mentioned in previous chapters that a high‐temperature fuel cell can be combined with a steam turbine in a bottoming cycle. The ability to use both gas turbines and steam turbines in a combined cycle with an SOFC has been known in concept for many years. It is only recently, however, that pressurized operation of SOFC stacks has been demonstrated for prolonged periods, thereby making the SOFC–gas turbine (SOFC–GT) system feasible practically. Pioneered by Siemens Westinghouse Power Corporation in their SureCell™ concept, the ideas of combined SOFC–GT are now being explored by other developers such as Mitsubishi Heavy Industries; the essential process features are illustrated diagrammatically in Figure 9.16. In this review of the status of SOFCs, it is worthwhile remarking that there are many opportunities for novel system design, as well as scope for considerable creativity by the systems engineer. Many examples have been reported in the literature. For instance, by connecting stacks in series, a multistage SOFC — the ‘UltraFuelCell’ — was developed under a programme supported by the US Department of Energy. A novel hybrid system concept has also been described in which SOFC and PEMFC technologies are Solid Oxide Fuel Cells After burner Natural gas C a t h o d e A n o d e Air Gas turbine Alternator (200 kW) Fuel cell (800 kW) Exhaust Preheaters Desulfuriser Figure 9.16 System concept for an SOFC–GT combined cycle. combined.10 The advantages of each type of fuel cell are enhanced by operating in synergy; the system is shown in Figure 9.17. The IT‐SOFC is run under conditions that give low fuel utilization and thereby enables a high‐power output for a relatively small stack size. Unspent reformed fuel appears in the anode exhaust where it undergoes a shift reaction, followed by a process stage when the final traces of carbon monoxide are removed. At this stage, the gas comprises mainly hydrogen and carbon dioxide, with some steam. This gas, once it is cooled, is suitable as a fuel in the PEMFC stack. The gas compositions for the system are listed in Table 9.1. The use of two stacks of different types for power generation results in increased overall electrical efficiency. The system is particularly attractive in terms of economics. Preliminary calculations reveal that the system is more competitive than an SOFC‐only system because of the anticipated relatively low cost of the PEMFC stack. On the other hand, the system would offer a much higher efficiency than could be achieved by a PEMFC‐only system when operating on natural gas. In the following chapter, it is shown that the fuel‐processing technology for running a PEMFC on natural gas is complex, bulky and expensive. How much better, then, to use an SOFC as the fuel processor! 10 Dicks, AL, Fellows, RG, Mescal, CM, Seymour, C 2000, A study of SOFC–PEM hybrid systems, Journal of Power Sources, vol. 86(1–2), pp. 501–506. 259 260 Fuel Cell Systems Explained Exhaust air Fuel in (e.g. methane) Solid oxide fuel cell Shift reactors Selective oxidation Air in PEM fuel cell Burner Exhaust air Heat addition to system Heat removal from system Figure 9.17 SOFC–PEMFC hybrid system (see also Table 9.1). Table 9.1 Summary of output powers for the hybrid system shown in Figure 9.17. SOFC stack power (kW) 369.3 PEM stack power (kW) 146.7 Turbine power (kW) Compressor power (kW) 100.3 –100.8 Net power output (kW) 515.5 Electrical output (kW) 489.7 Overall efficiency, % (LHV) 61 Further Reading Atkinson, A, Barnett, S, Gorte, RJ, Irvine, JTS, McEvoy, AJ, Mogensen, M, Singhal, SC and Vohs, J, 2004, Advanced anodes for high‐temperature fuel cells, Nature Materials, vol. 3, pp. 17–27. Cowin, P, Petit, C, Lan, R, Irvine JTS and Tao, S, 2011, Recent progress in the development of anode materials for solid oxide fuel cells, Advanced Engineering Materials, vol. 1, pp. 314–312. Solid Oxide Fuel Cells Fergus, JW, 2005, Metallic interconnects for solid oxide fuel cells, Materials Science and Engineering: A, vol. 397, pp. 271–283. Irvine, JTS, Neagu, D, Verbaeke, MC, Chatzichristodoulou, C, Graves, C and Mogensen, MB, 2016, Evolution of the electrochemical interface in high‐temperature fuel cells and electrolysers, Nature Energy, vol. 1, 15014, available online at http://palgrave.nature. com/articles/nenergy201514 (accessed 24 September 2017). Oishi, N, Rudkin, RA, Steele, BCH and Brandon, NP, 2002, Thick Film Stainless Steel Supported IT‐SOFCs for Operation at 500‐600°C, Scientific Advances in Fuel Cells, Elsevier Science Ltd, Amsterdam. Singhal, SC, 2007, Solid oxide fuel cells, The Electrochemical Society, Interface, Winter 2007, pp. 41–44. Singhal, SC and Kendall, K (eds.), 2003, High Temperature Solid Oxide Fuel Cells – Fundamentals, Design and Application, Elsevier, B.V, Amsterdam. Steele, BCH and Heinzel A, 2001, Materials for fuel cell technologies, Nature, vol. 414, pp. 345–352. Wei, T, Singh, P, Gong, Y, Goodenough, J, Huang, Y and Huang, K, 2014, Sr3‐3x Na3x Si3O9‐1.5x (x=0.45) as a superior solid oxide ion electrolyte for intermediate temperature solid oxide fuel cells. Energy and Environmental Science, vol. 7, pp. 1680–1684. 261 263 10 Fuels for Fuel Cells 10.1 Introduction This chapter considers fuels that can be used for the principal types of fuel cell. Hydrogen has been promoted worldwide as a panacea for energy problems in that it may eventually replace, or at least greatly reduce, the reliance on fossil fuels while being itself a clean‐burning fuel that releases no greenhouse gases into the atmosphere. Although the most abundant element in the universe — the stuff from which stars are made — hydrogen does not occur freely on earth, but is predominantly found in combination with oxygen as water and with carbon as fossil fuels. Chemical, thermal or electrical energy has to be expended to extract hydrogen from these sources. Hydrogen is therefore not a new form of primary energy, but a vector (or carrier) for storing and transporting energy from any one of a myriad of sources to where it may be utilized. Basic chemical and physical data on hydrogen and some of the other fuels that may be suitable for fuel cells are given in Table 10.1. Hydrogen has the advantage over all other fuels in that it is easily oxidized at the fuel cell anode and that the only chemical product is water. For this reason, hydrogen has become preferred for fuel‐cell vehicles (FCVs). Cars and other vehicles that employ hydrogen‐fuelled proton‐exchange membrane fuel cells (PEMFCs) are ‘zero emission’, because the only exhaust from the vehicle tailpipe is water or water vapour.1 Various primary fossil fuels that can be used to generate hydrogen are discussed in some detail in Section 10.2. These hydrocarbon fuels include natural gas, petroleum products (such as gasoline and diesel), coal gas and coal. Biofuels are also a possible source of hydrogen, and the options for using such fuels are reviewed separately in Section 10.3. The processes and chemical‐conversion technologies for obtaining hydrogen from hydrocarbons, whether these are from fossil fuels or biofuels, are well established industrially and are covered in Sections 10.4–10.6. For stationary fuel‐cell power plants, there are strong arguments for producing hydrogen as close to the fuel‐cell stack as possible. This is because, apart from safety considerations, the heat generated in the fuel‐ cell stack may then be conveniently used for some of the fuel processing. Integration of 1 In a combustion engine or gas turbine, not only is hydrogen oxidized to water (or steam), but also nitrogen in the air is oxidized to yield nitrogen oxides (NOx). Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 264 Fuel Cell Systems Explained Table 10.1 Some properties of hydrogen and other candidate fuels for fuel‐cell systems. Hydrogen (H2) Molecular weight (g) 2.016 Methane (CH4) 16.04 Ammonia (NH3) 17.03 Methanol (CH3OH) 32.04 Ethanol (C2H5OH) 46.07 Gasoline (C8H18)a 114.2 Freezing point (°C) −259.2 −182.5 −77.7 −98.8 −114.1 –56.8 Boiling point (°C) −252.77 −161.5 −33.4 64.7 78.3 125.7 Net enthalpy of combustion at 25°C (kJ mol−1) 241.8 802.5 316.3 638.5 1275.9 5512.0 Heat of vapourization (kJ kg−1) 445.6 510 1371 1129 839.3 368.1 Liquid density (kg L−1) 77 425 674 786 789 702 Specific heat (STP) (J mol−1 K−1) 28.8 112.4 188.9 Flammability limits in air (%) Auto‐ignition temperature in air (°C) a) 34.1 36.4 76.6 4–77 4–16 15–28 6–36 4–19 1–6 571 632 651 464 423 220 Gasoline is a blend of hydrocarbons and varies with producer, application and season. N‐octane is reasonably representative of properties except for vapour pressure, which is intentionally raised by introducing light petroleum fractions. the fuel‐cell stack and fuel processor is therefore an important aspect of system design, as outlined in Section 10.5. In the early days of FCV development, some of the concept cars and buses ran on methanol (CH3OH) — an on‐board reformer converted the methanol into a hydrogen‐ rich gas for the fuel cell. Even on‐board reforming of gasoline was investigated with support from the US Department of Energy (DOE). In recent years, however, the practice has largely fallen out of favour because such generation of hydrogen can be a complex and not particularly efficient process. Moreover, dispensing with on‐board fuel processing greatly simplifies the design of the vehicle drivetrain and ensures that water is the only emission. The implications for moving from on‐board to off‐board reforming of fuel are discussed in Section 10.7. Hydrogen may, of course, be generated by the electrolysis of water — the reverse of a fuel cell. Since the purpose of a fuel cell is to produce electricity, this may at first seem perverse. In many cases, however, electrolysis is a very convenient means of providing hydrogen for small mobile fuel cells and even vehicles. Attention is currently being given to the large‐scale generation of hydrogen by electrolysis as a means of storing renewable energy. To this end, several experimental power‐to‐gas (P2G) schemes have been built in which solar or wind power is used to produce hydrogen by electrolysis, as reported in Section 10.8. Thermochemical methods of hydrogen generation are evaluated in Section 10.9. The chapter concludes (Section 10.10) with an appraisal of hydrogen generation by biological systems that employ processes that involve enzymes, bacteria and sunlight. Fuels for Fuel Cells Fossil fuels (e.g., gas, oil, coal; Section 10.2) Large chemical plants reforming fuels to hydrogen (Section 10.4) Hydrogen stored as gas at pressure or cryogenic liquid (Section 11.3 and 11.4) Renewable fuels (e.g., biofuels, waste; Section 10.3) Biological hydrogen generation systems (Section 10.10) Hydrogen stored as synthetic fuel—methanol, ammonia, sodium borohydride (Sections 11.5 and 11.6) Central or dispersed power generation (Chapter 12) Thermochemical hydrogen production (Section 10.9) Electrolysis (Section 10.8) Small-scale hydrogen storage (e.g., high pressure cylinder or metal hydride cannister; Section 11.7) Small-scale fuel reformer producing hydrogen (Section 10.6) Stationary fuel-cell system with MCFC or SOFC using fuel directly (Chapters 8 and 9) Electricity generated by renewable energy, solar, wind, wave, hydro, etc, Nuclear Stationary fuel-cell system using hydrogen Stationary or portable fuelcell system using synthetic fuel Hydrogen fuel-cell system for vehicles (Chapter 4) Figure 10.1 The many different pathways by which hydrogen can be supplied to fuel cells. From the aforementioned overview, it will be evident that there are many possible ways of producing hydrogen for a fuel cell. An illustration of how the different fuel‐processing pathways are related is presented in Figure 10.1, together with directions as to the appropriate information that can be found in this chapter and others. 265 266 Fuel Cell Systems Explained 10.2 Fossil Fuels 10.2.1 Petroleum Petroleum is a mixture of gaseous, liquid and solid hydrocarbon‐based chemical compounds that occur in sedimentary rock deposits throughout the world. Crude liquid petroleum (crude oil) has little value, but when refined, it provides high‐value liquid feeds, solvents, lubricants and other products. Petroleum‐derived fuels account for up to one‐half of the world’s total energy supply and include gasoline (petrol),2 diesel fuel, aviation fuel and kerosene. Simple distillation is able to separate various components of crude petroleum into generic fractions with different boiling ranges, as shown in Figure 10.2. The amounts of the fractions that are obtained from any particular variety of crude oil depend on the origin of supply. Each fraction of petroleum contains a different proportion of chemical compounds, these being normal and branched paraffins or alkanes, monocyclic and polycyclic C1 to C4 20°C C5 to C3 Fractions decreasing in density and boiling point 70°C C5 to C10 120°C C10 to C16 Fractionating column 170°C Liquefied petroleum gas Chemicals Petrol for vehicles Jet fuel Paraffin for lighting and heating Diesel fuels C14 to C20 270°C Crude oil C20 to C50 Fractions increasing in density and boiling point Lubricating oils, waxes, polishes C20 to C70 Fuels for ships, factories and central heating 600°C >C70 residue Bitumen for roads and roofing Figure 10.2 Distillation products of crude petroleum. 2 Colloquially the word ‘petrol’ is used to describe the fuel for cars in the United Kingdom, Australia and other countries, whereas the term ‘gasoline’ is more commonly used for the same fuel in the United States. Fuels for Fuel Cells paraffins (naphthenes) and mononuclear and polynuclear aromatic hydrocarbons. The naphtha fraction (C7–C11) principally comprises normal alkanes and some monocyclic alkanes (e.g., cyclopentane and cyclohexane). Low boiling fractions generally have more of these low molecular weight alkanes than the high boiling fractions. Similarly the content of polycyclic alkanes and aromatic hydrocarbons increases on moving from low boiling to high boiling fractions. The various fractions can be used without further refining and are referred to as ‘straight‐run’ fractions. In the early days of motoring, for example, gasoline was a straight‐run distillate. As demand for fuel increased, more gasoline (which typically comprises alkanes with between 4 and 12 carbon atoms in each molecule) was produced by chemically cracking the fractions that contained hydrocarbons of high molecular weight and blending the product with the straight‐run distillate. Nowadays, fuels for vehicles are blended from several petroleum refinery process streams that are derived by the following methods: direct distillation of crude oil, catalytic and thermal cracking, hydrocracking, catalytic reforming, alkylation and polymerization. In addition, gasoline can be doped with various compounds to improve engine lubrication, reduce corrosion and lessen the risk of ‘knock’ or pre‐ignition in internal combustion engines.3 Other engine performance enhancers that may be added include antioxidants, metal deactivators, lead scavengers, anti‐rust agents, anti‐icing agents, detergents and dyes. At the end of the refining process, finished gasoline typically contains more than 150 separate compounds although as many as 1000 compounds have been identified in some blends. Clearly gasoline and other petroleum products can be chemically complex mixtures, but further discussion of their production and composition is outside the scope of this book. The importance of discussing the composition of vehicle fuels is that the choice of the process required for generating hydrogen is determined by the chemical composition and the physical and combustion characteristics of the fuel. In the case of catalytic conversion, the presence of various trace compounds in the fuel may be of serious concern as they may act as poisons for the conversion catalysts and indeed also for anode catalyst in the fuel‐cell stack. In this context, the problematic trace compounds in fossil fuels are organic compounds that contain silica and sulfur, and organometallic compounds, such as various porphyrins. Sulfur is a persistent poison for catalysts and its removal from raw fuels is discussed specifically in Section 10.4.2. The gasoline fraction of petroleum, together with the heavier diesel, is widely distributed as a fuel for vehicles that range from passenger cars to heavy‐duty trucks, buses and rail locomotives. Because the infrastructure for delivering such fuels is mature, there are valid reasons for using these fuels for FCVs. Components that are currently added to gasoline or diesel for anti‐knock or lubrication purposes would not be required for such vehicles so that future refinery operations could be simplified. Techno‐economic studies and ‘well‐to‐wheel’ analyses indicate, however, that the benefits of using gasoline in FCVs, compared with internal combustion engines, are less than those when operating with hydrogen or even methanol (cf. Section 12.4, Chapter 12). 3 Lead tetraethyl, (CH3CH2)4Pb, was added to gasoline or petrol in the 1920s to improve the octane rating to avoid pre‐ignition. Such leaded fuel was gradually phased out from the 1970s because of concern over its toxicity. Modern refining operations can produce an unleaded gasoline with a sufficiently high octane rating (98 octane) suitable for all high‐compression engines. 267 268 Fuel Cell Systems Explained Progressively more demanding emissions standards for vehicles have led to lower sulfur contents in distributed gasoline and diesel fuels in both Europe and the United States. This trend is expected to continue and it may lead to more widespread use of synthetic fuels, or biofuels. Synthetic diesel fuel made, for example, by the transesterification of vegetable oils or animal fats is intrinsically low in sulfur. Other biological sources are being investigated for future alternative fuels, notably algae and some lipids from waste waters. Biofuels can be used not only for land‐based vehicles, such as cars, buses, trucks and trains, but also for aircraft. Vehicles running on biofuels offer a demonstrable method of reducing carbon emissions. Other hydrocarbons of low molecular weight, such as propane (C3H8) and butane (C4H10), are often found associated in crude oil deposits. When this material is distilled, these hydrocarbons emerge as a gaseous product, which is marketed as liquefied petroleum gas (LPG). Available therefore as a by‐product from oil refineries, LPG is a gas at atmospheric pressure that will liquefy at a moderately elevated pressure and thus can be transported easily. The fuel is used extensively for applications as diverse as camping gas stoves and vehicle propulsion. In addition, LPG is attractive for fuel‐cell systems that may be used as remote stationary power sources where there is no pipeline source of gas and possibly also for some FCV applications. 10.2.2 Petroleum from Tar Sands, Oil Shales and Gas Hydrates Petroleum substances can be found in the Earth’s crust as deposits of solid or near‐solid material in sandstone at depths that are usually less than 2000 m. Some may also be found as an outcrop on the surface. Huge amounts of such ‘tar sands’ are present in various parts of Canada and the United States, but the high bitumen content (of very high molecular weight) makes recovery less attractive than from more conventional petroleum deposits. Similarly, ‘oil shales’ comprise a significant and largely untapped source of petroleum materials. Oil shales are compact laminated rocks of sedimentary origin in which petroleum is locked. The oil can be obtained from the rock by distillation. Global deposits of shale oil are estimated to be around 5 × 1012 barrels, that is, significantly more than crude oil resources, which stand at only 1.5 × 1012 barrels.4 Compared with crude oil, however, only a small fraction of oil from shales is easily recoverable by conventional technologies; much of the oil is of high molecular weight and bituminous in nature. While oil shales may become an important energy resource into the future, a discussion of the processing of such materials is outside the scope of this book. In natural petroleum reservoirs where the prevailing pressures are high and temperatures low (e.g., under permafrost), methane (CH4) and other normally gaseous hydrocarbons form ice‐like hydrogen‐bonded complexes with water that are known as ‘gas hydrates’. Methane hydrates have been considered as a method of transporting natural gas from remote fields. 10.2.3 Coal and Coal Gases Coal is the most abundant of all fossil fuels. Chemically, the resource is the most complex given that it is formed from the compaction and induration of various plant remains 4 World Energy Resources, 2013, World Energy Council. ISBN: 978 0 946121 29 8. Fuels for Fuel Cells that were laid down in geological time, especially during the Carboniferous Period between 345 and 280 million years ago. The first product of decay and consolidation is peat, which has a relatively low carbon content and a high moisture content. Under forces of heat and pressure, peat gradually converts first into bituminous coal and, ultimately, to hard coal (anthracite). The diversity of the original plants, the variations in the depositional environment and the age of the coal since it was laid down (the ‘rank’ of the coal) have resulted in an exhaustive literature that classifies coal by its appearance (macroscopic and microscopic), its chemical composition, the occluded mineral matter, and its physical properties. It is worth pointing out that apart from combustion, further processing of coal to produce liquids, gases and coke is highly dependent on the properties of the raw coal material. For example, coals with 20–30 wt.% volatile organic matter are primarily suitable for producing coke. By contrast, bituminous coals are best suited for carbonization, i.e., heating to temperatures of 750–1500°C to form ‘coal gas’, also known as ‘town gas’. Most types of coal can be gasified, although the rank and other characteristics of the coal will influence the product mix from the various designs of gasifier. Coal carbonization was the original method for producing coal gas in the 19th century in both the United Kingdom and North America. Simple carbonization, i.e., the destructive distillation of coal in the absence of air, yielded gas (a mixture mainly of hydrogen and carbon oxides), organic liquids (tars and phenolics) and a residual coke. Partial pyrolysis of coal was also carried out in coke ovens where the main objective was to produce coke for the steel industry. Such ovens generated combustible gas that could also be used in industry. Another type of coal gas, known as ‘producer gas’, was obtained by blowing air and steam over hot coke at high temperatures. Most of these gas‐making processes are now obsolete. In the 1950s, methods of manufacturing gas from oil were developed in the United Kingdom and the United States. The production of coal gas by the so‐called catalytic rich gas (CRG) process had been proven at the pilot scale by engineers in the United Kingdom, but with the advent of natural gas from the North Sea, further development was curtailed. Nevertheless, the CRG process is still operated in various locations around the world where manufactured gas is required. Nowadays, coal carbonization has been superseded for large‐scale gas production by various coal gasification processes. Gasification differs from carbonization in that the heated coal is reacted with steam and oxygen (or air) at high temperatures. The products of primary coal gasification are mainly gases, together with smaller amounts of liquids and solids. The relative proportion of products depends on the type of coal, the temperature and pressure of reaction and the relative amounts of steam or oxygen that are injected into the gasifier. Coke is not formed and the only waste material is an inert ash or slag. Further chemical processing of the raw gas can be carried out, for example, to increase the CH4 content or to alter the hydrogen‐to‐carbon monoxide (CO) ratio to suit the final application for the coal gas. The numerous gasification systems available today can be broadly classified into three basic types of gasifier: (i) moving bed, (ii) entrained bed and (iii) fluidized bed. The moving‐ bed gasifiers produce a gas at low temperature (450–650°C) that contains CH4 and ethane (C2H6), which arise from devolatilization of the coal, together with a hydrocarbon liquid stream that contains naphtha, tars, oils and phenolic liquids. Entrained‐bed gasifiers produce gas at higher temperatures (>1200°C) with virtually no hydrocarbons in the resultant gas stream and much lower amounts of liquid hydrocarbons. In fact, the 269 270 Fuel Cell Systems Explained Table 10.2 Typical coal gas compositions (mole per cent basis). BG–Lurgi gasifier (non‐ slagging) BG–Lurgi slagging gasifier Moving‐bed O2‐blown (BG–Lurgi) Fluidized‐ bed (Winkler) Coal Pittsburgh No.8 seam Pittsburgh No.8 seam Illinois No.6 trace Illinois Texas lignite Illinois No.6 Illinois No.6 No.6 Ar Trace Trace Trace CH4 9.43 6.55 3.26 Entrained‐ bed O2‐ blown (Texaco) 0.90 1.10 5.93 0.12 Entrained‐ bed air‐ blown Trace 1.08 Entrained‐ bed O2‐ blown (Shell) 1.15 0.00 C2H6 0.78 0.09 0.10 0.00 0.00 0.00 0.00 C2H4 0.33 0.18 0.20 0.00 0.00 0.00 0.00 H2 32.30 35.49 20.75 36.47 37.18 9.68 27.99 CO 19.98 50.50 5.73 42.65 48.59 17.20 66.14 CO2 34.52 3.55 11.66 19.97 13.25 6.45 1.57 2.66 3.64 0.20 0.77 0.86 66.67 4.30 N2 NH4 0.00 0.00 0.40 0.13 0.12 0.00 0.00 H2O 0.00 0.00 61.07 21.65 20.25 5.38 2.10 H2S Total 0.00 0.00 0.49 0.26 1.23 0.00 1.36 100.00 100.00 100.00 100.00 100.00 100.00 100.00 entrained‐bed gas is composed almost entirely of hydrogen, carbon monoxide and carbon dioxide (CO2). In terms of composition and temperature (925–1050°C), the gas from the fluidized‐bed gasifier falls somewhere between those obtained from the two other designs of reactor. For all three gasifiers, the heat required for the reaction of coal and steam is effectively provided by the partial oxidation (POX) of the coal. The temperature, and therefore the composition, of the gas is dependent on the proportions of oxidant and steam, as well as the design of the reactor. The chemical compositions of typical coal gases from some of the leading types of gasifier are given in Table 10.2. The gases invariably contain contaminants that must be removed before use in fuel cells. Methods of gas clean‐up are described in the next section. 10.2.4 Natural Gas and Coal‐Bed Methane (Coal‐Seam Gas) Natural gas is combustible and occurs in porous rocks such as sandstone in the Earth’s crust. It is found with, or close to, crude oil reserves but may also be present in separate reservoirs. Most commonly, it is trapped between liquid petroleum and an impervious rock layer (‘cap rock’) in a petroleum reservoir. If the pressure is sufficiently high, the gas will be intimately mixed with, or dissolved in, the crude oil. Natural gas is comprised of a mixture of hydrocarbons of low boiling point. Usually, methane is present in the greatest concentration, with smaller amounts of ethane, propane and higher‐order alkanes. In addition to hydrocarbons, natural gas contains Fuels for Fuel Cells Table 10.3 Typical compositions of natural gases from different geographic regions. Component North Sea Qatar the Netherlands Pakistan Ekofisk, Norway Indonesia CH4 94.86 76.6 81.4 93.48 85.5 84.88 C2H6 3.90 12.59 2.9 0.24 8.36 7.54 2.38 0.4 0.24 2.85 1.60 0.04 0.86 0.03 0.1 0.06 C3H8 i‐C4H10 0.15 n‐C4H10 C5+ N2 S 0.11 0.21 0.02 0.79 0.24 14.2 4 ppm 1.02 1 ppm 0.12 0.41 0.22 1.82 4.02 0.43 4.0 30 ppm 2 ppm N/A Values are vol.%, unless otherwise stated. various quantities of nitrogen and carbon dioxide (CO2), together with traces of other gases such as helium (often present in commercially recoverable quantities). Sulfur is also present to a greater or lesser extent, mostly in the form of hydrogen sulfide (H2S). Natural gas is often described as being dry or lean (containing mostly CH4), wet (containing considerable concentrations of higher molecular weight hydrocarbons), sour gas (with significant levels of H2S), sweet gas (low in H2S) and casinghead gas (obtained from an oil well by extraction at the surface). Some typical compositions of natural gases from different regions of the world are listed in Table 10.3. Coal‐bed methane (CBM), sometimes known as coal‐seam gas (CSG), is found absorbed in the solid matrix of underground deposits of coal. It is only in recent years that CBM has been recognized as a huge energy resource. In Queensland, Australia, for example, investment of over US$20 billion has been spent on the infrastructure to extract the gas from vast areas of underground coals and then combine and transfer the outputs from different wells to a liquefied natural gas (LNG) terminal. Located in Gladstone, the LNG terminal has been built to supply a growing demand in Asia for the fuel. Unlike natural gas, CBM is usually very low in sulfur and other hydrocarbons such as ethane and propane. Sometimes, there are significant concentrations of CO2 and N2 associated with CBM that may have to be removed before the gas is transported. The chief problem, however, lies with the water that is brought to the surface when the gas is released from the coal. This water has to be separated and dispatched via irrigation, evaporation or reinjection into underground reservoirs. Some processing of natural gas or CBM may therefore be necessary close to the point of extraction before it enters a transmission system. Examples with natural gas are the bulk removal of sulfur (sweetening) and the removal of high molecular weight hydrocarbons, nitrogen, acid gases, liquid water and liquid hydrocarbons. With CBM, the only processing usually required is the removal of nitrogen or acid gases and water. There are wide variations in the composition of natural gas fed to transmission systems around the world, as indicated in Table 10.3. Even within geographic regions 271 272 Fuel Cell Systems Explained and also according to the season, there may be significant variations in composition according to the field from which the gas is obtained. In colder months, the heating value of the gas may drop due to a fall in concentration of higher hydrocarbons. At such times it is common practice to enrich the gas by blending in mixtures of ethane, propane and butane to ensure that the distributed gas has consistent burning characteristics and heating value throughout the year. This raises a fundamental problem for developers of fuel cells. That is, the design of a fuel process is influenced by the composition of the gas rather than by its combustion properties. The situation is exacerbated by gas companies that may enrich the gas with mixtures of propane–air or butane–air in order to boost the heating value during times of seasonal peak demand. Certainly, reforming catalysts — such as the Johnson Matthey CRG catalysts (v.s.) widely used in the steam reforming of natural gas (see Section 10.4.3) — will only tolerate a very small percentage of oxygen in the feed gas. Natural gas has no distinctive odour (except for very sour gases), and for safety reasons, therefore, pipeline companies and utilities usually odourize the gas5 either as it enters the transmission system or within local distribution zones. Various odourants may be used and the most common are thiophenes and mercaptans. Tetrahydrothiophene (THT) is widely used throughout Europe and in the United States (where it is known as Pennwall’s odourant), whereas in the United Kingdom and Australia, for example, a cocktail of compounds is used (a combination of ethyl mercaptan, tertiary butyl mercaptan and diethyl sulfide). Coal‐bed methane is quite different from natural gas in that the composition (typically greater than 92% CH4 on a dry basis) is remarkably consistent throughout the lifetime of the well. The gas is also naturally low in sulfur, and therefore once any associated water or acid gases are removed, only the odourant that may be injected is likely to be a concern of fuel‐cell engineers. 10.3 Biofuels Biomatter (or biomass) is a catch‐all term for the natural organic material associated with living organisms that include terrestrial and marine vegetable matter — everything from algae to trees — together with animal tissue and manure. On a global basis, it is estimated that over 150 Gt of vegetable biomatter is generated annually. The production of biomass material is often expressed in tonnes per hectare, (t ha−1); 1 ha = 104 m2. This yield ranges from about 13 t ha−1 for water hyacinth to 120 t ha−1 for Napier grass. In view of its high energy content, biomass represents an important source of a renewable fuel that can be obtained via the following routes: ● ● ● ● ● Direct combustion. Conversion to biogas via pyrolysis, hydrogasification or anaerobic digestion. Conversion to ethanol (C2H5OH) via fermentation. Conversion to syngas thermochemically. Conversion to liquid hydrocarbons by hydrogenation. 5 LPG is also odourized, usually with ethyl mercaptan. Fuels for Fuel Cells Box 10.1 Syngas Syngas (also known as ‘synthesis gas’) is the product gas from a steam reformer and consists primarily of hydrogen, carbon monoxide and often some carbon dioxide. In fact, the term can be used for any gas mixture that contains hydrogen and carbon monoxide, e.g., the gas produced by gasification of coal. The name refers to its use in the synthesis of substitute (or synthetic) natural gas (SNG) and ammonia or in the production of synthetic liquid hydrocarbon fuels by the Fischer–Tropsch process, e.g., synthetic diesel and synthetic gasoline. The latter is now more commonly known as the ‘gas‐to‐liquid (GTL)’ process and is carried out over a nickel‐, cobalt‐ or thorium‐based catalyst at 200°C, i.e., nCO 2n 1 H2 C nH2 n 2 nH 2 O Another source of biofuel is municipal waste, i.e., sewage sludge and municipal solid waste (MSW). The latter is a general term applied to solid household or commercial garbage. Methane has been produced by the biological digestion of sludge in waste‐water treatment plants for many years. In many cases, the methane is used to fuel generators that provide supplementary power to assist the running of the treatment plant. Fuel‐cell systems have been employed for such power generation in several demonstrations throughout the Europe Union (EU) and the United States. Gaseous fuels that emanate from landfill sites and other forms of refuse digestion can also constitute a useful source of energy that is well suited to fuel‐cell systems. According to the European Commission, in the EU as a whole, approximately 42% of municipal waste is currently recycled, 24% is incinerated for energy generation and the remaining 34% is consigned to landfill sites. In some of the member states — notably Germany, Belgium, Austria and the Netherlands — landfill has almost been eliminated through strong legislation that encourages recycling and energy generation.6 In the United States, by contrast, only 23% of household waste is recycled, 6.4% is composted, 7.6% is incinerated to generate energy and 63% goes to landfill.7 The combustible component of MSW may be extracted either chemically by gasification (to yield gases, liquids and char) or by anaerobic pyrolysis. Alternatively, anaerobic digestion may be employed to generate methane from solid waste through the action of specific bacteria. Anaerobic digestion as currently practised requires a wet waste of relatively high nitrogen content. Consequently, the nitrogen‐to‐carbon ratio of about 0.03 that is typical of most plant‐based biomass is increased to 0.07 by the addition of animal manure, sewage sludge or other nitrogen‐rich waste. Anaerobic digesters can be built at a relatively small scale (a few kW), compared with pyrolysis gasifiers that normally only become attractive at the MW scale. Such anaerobic digesters are finding increasing use for generating power in remote areas of countries (e.g., India) that cannot afford to be connected to an electricity distribution network. Biogases produced directly from landfill sites, or by the anaerobic digestion of biomass, contain mixtures of CH4, CO2 and N2, together with various other organic 6 Eurostat 2012 website. http://ec.europa.eu/eurostat 7 Shin, D, 2014, Generation and Disposition of Municipal Solid Waste (MSW) in the United States –A National Survey, Columbia University Earth Engineering Center. 273 274 Fuel Cell Systems Explained Table 10.4 Example compositions of biogases. Biogasa Source Agricultural sludge Methane (vol.%) 55–65 Ethane (vol.%) Biogasb Landfill gase Biogasc Biogasd Agricultural sludge Brewery effluent 50–70 65–75 57 30–45 30–40 25–35 37 55–70 0 Propane (vol.%) 0 Carbon dioxide (vol.%) Nitrogen (vol.%) 33–43 2–1 0–2 Small Hydrogen sulfide (ppm) <2000 ~500 Small Ammonia (ppm) <1000 ~100 Hydrogen (vol.%) 6 <5000 <1 Small Higher heating value MJ (Nm3)−1 23.3 3 −1 Density kg (Nm ) >20 1.16 a) b) Paper BP‐12 20th World Gas conference 1997. Jemsen, J, Tafdrup, S and Chrisensen, J, 1997, Combined utilization of biogas and natural gas, 20th World Gas Conference, 1997, Paper BO‐06, c) Renewable Energy World, March 1999, p. 75. d) CADDET, Renewable Energy Newsletter, July 1999, pp. 14–16 (Biogas used in Toshiba 200‐kW phosphoric acid fuel cell.) e) CADDET, Renewable Energy Technical Brochure, No. 32, 1996. compounds. The respective compositions vary widely and, in the case of landfill, depend on the age of the site. A new site usually produces gas with a high heating value, but this tends to decrease over a period of time. Some compositions of biogases are listed in Table 10.4. Given the relatively high levels of carbon oxides and nitrogen, most biogases have low heats of combustion and are thus unattractive for use in gas engines or turbines. Fortunately, this is not a major issue for fuel cells, particularly in the case of molten carbonate (MCFC) and solid oxide fuel cell (SOFC) systems that are able to handle very high concentrations of carbon oxides. The same situation holds for phosphoric acid fuel cells (PAFCs) but to a lesser extent. Many MCFC and PAFC systems fuelled with landfill gas and/or gas from waste‐water treatment plants have been built and successfully operated. Bio‐liquids, such as methanol and ethanol, are acceptable for some fuel‐cell systems, as described in Sections 6.1 and 6.2, Chapter 6. Methanol can be obtained from syngas that may be derived from biomass or natural gas, whereas ethanol can be produced directly by the fermentation of biomass. Alcohols are also attractive because of the ease by which they can be reformed into hydrogen‐rich gas. This makes the alcohols suitable for applications such as stationary power backup systems, e.g., for remote telecommunications towers. Fuels for Fuel Cells 10.4 Basics of Fuel Processing 10.4.1 Fuel‐Cell Requirements Fuel processing may be defined as the conversion of the raw primary fuel supplied to a fuel‐cell system into the fuel gas required by the stack. Each type of stack requires fuel of a particular quality, as summarized in Table 10.5. Essentially, the lower the operating temperature of the stack, the more stringent is the level of fuel quality, and therefore the greater the demand placed on fuel processing. For example, fuel fed to a PAFC stack needs to be hydrogen‐rich and contains less than about 0.5 vol.% CO, whereas a PEMFC has to be essentially free of CO. By contrast, both MCFC and SOFC fuel cells are capable of utilizing this gas internally through the ‘water-gas shift (WGS)’ reaction. Additionally, unlike PAFCs and PEMFCs, SOFCs and internal‐reforming MCFCs can operate with methane. It is not widely known, however, that PEMFCs can run directly on some hydrocarbons such as propane, although the performance is poor.8 Basic explanations of the various technologies for fuel processing are given in the following sections. Some of the detailed design of individual reactors and systems is proprietary, of course, but a wealth of information is available from various organizations who are involved in the development of fuel‐cell systems. 10.4.2 Desulfurization As noted earlier, natural gas and petroleum liquids contain organic sulfur‐containing compounds that normally have to be removed before further processing of the given fuel. Some deactivation of catalysts used in steam reforming can occur with fuel containing less than 0.2 ppm sulfur, and WGS catalysts are even more intolerant. For the fuel cell itself, it has been shown that levels of only 1 ppb are sufficient to poison permanently the anode catalyst in a PEMFC. Fossil fuels and biofuels often contain a range of sulfur compounds. In the case of natural gas, sulfur may exist in the form of H2S, and it may be present in the odourant Table 10.5 Fuel quality for principal types of fuel cell. Gas species PEMFC AFC PAFC MCFC SOFC H2 Fuel Fuel Fuel Fuel Fuel CO Poison (>10 ppm) Poison Poison (>0.5%) Fuel a Fuela CH4 Diluent Diluent Diluent Diluent Diluentb CO2 and H2O Diluent Poison Diluent Diluent Diluent S (as H2S and COS) Few studies, to date Unknown Poison (>50 ppm) Poison (>0.5 ppm) Poison (>1.0 ppm) a) b) b In reality, CO reacts with H2O producing H2 and CO2 via the shift reaction (see reaction (8.3) in Chapter 8), and CH4 reacts with H2O to produce H2 and CO faster than it oxidizes as a fuel at the anode. Methane is a fuel for the internal‐reforming MCFC and SOFC. 8 Baker, BS, 1965, Hydrocarbon Fuel Cell Technology, Academic Press, New York and London. 275 276 Fuel Cell Systems Explained that has been introduced by the utility company for safety reasons. With petroleum fractions, the sulfur compounds may be highly aromatic in nature, and gasoline currently contains some 300–400 ppm of sulfur as organic compounds. In the drive to reduce emissions from vehicles, regulations have been introduced to limit the sulfur in both gasoline and diesel fuels. In the United States, for example, the 2004 Tier 2 Gasoline Sulfur Program permitted refiners to produce fuel with a range of sulfur levels, provided the annual corporate average remained below 30 ppm and no individual batch exceeded 80 ppm. The current Tier 3 programme lowered the sulfur content of all gasoline to a maximum of 10 ppm from the beginning of 2017. The Euro 3 standard issued by the EU in January 2000 set a sulfur limit of 350 ppm for diesel and 150 ppm for gasoline. The Euro 5 came into effect in 2009 and restricts sulfur in both fuels to 10 ppm. Further reduction of the allowable levels may require refinement of the methods that are presently used to desulfurize such fuels — even to the stage of complete removal of sulfur. There are essentially two methods of desulfurizing fuels. The most common industrial approach is a process known as hydrodesulfurization (HDS). In the HDS reactor, any organic sulfur‐containing compounds are converted over a supported nickel– molybdenum oxide or cobalt–molybdenum oxide catalyst to hydrogen sulfide via hydrogenolysis reactions such as: C 2 H 5 2 S 2H 2 2C 2 H6 H2S (10.1) The rate of hydrogenolysis increases with temperature. At operating temperatures of 300–400°C and in the presence of excess hydrogen, the reaction essentially goes to completion. It should also be noted that, at a given temperature, the lighter sulfur compounds easily undergo hydrogenolysis, whereas the corresponding reaction rate for odourants such as thiophene (C4H4S) and THT (C4H8S) is much slower. The H2S that is formed by such reactions is subsequently absorbed by a bed of zinc oxide and therein converted to zinc sulfide, i.e., H2S ZnO ZnS H2O (10.2) The operating conditions and composition of the feed gas determine the choice between nickel or cobalt catalysts. The optimum temperature for most HDS catalysts lies between 350 and 400°C, and the catalyst and zinc oxide may be placed in the same vessel. A popular variation of the traditional industrial HDS process, known as the PURASPEC™ process, is marketed by Johnson Matthey Process Technologies. Hydrodesulfurization as a means of removing sulfur to very low levels is ideally suited to PEMFC or PAFC systems. In such technology, the hydrogen required by reaction (10.1) is obtained by recycling a small amount of the reformer product, which is rich in hydrogen, back to the HDS reactor upstream of the reformer. Unfortunately, HDS cannot easily be applied to internal‐reforming MCFC or SOFC systems, since there is no hydrogen‐rich stream to feed to the reactor. At least one developer of high‐temperature fuel cells overcame this problem by including a small reformer reactor into their plant for the sole purpose of generating hydrogen for HDS. If HDS is not feasible, then it may be possible to remove sulfur‐containing compounds from the fuel gas with the aid of an absorbent material. Activated carbon is especially suitable for use in small systems and can be impregnated with metallic promoters to enhance the absorption of specific compounds such as hydrogen sulfide. Molecular Fuels for Fuel Cells sieves such as the ZSM‐5‐type and faujasite‐type zeolites may also be employed. The absorption capacity of such materials is, however, quite low, and the beds of absorbent have to be replaced at regular intervals. These issues may result in a serious economic disadvantage for large systems. Some organizations claim to have developed sulfur‐tolerant catalysts for reforming or POX applications. Argonne National Laboratory, for example, has developed a sulfur‐ tolerant catalyst for their autothermal diesel reformer, but this is only intended to cope with processing of commercial low‐sulfur diesel fuel. If POX is used to process fuel, rather than steam reforming, it is likely that the product gas will still contain a few ppm of sulfur, and its removal will be an essential, but non‐trivial, additional processing step before feeding the gas to the anode of a PEMFC. Zinc oxide (ZnO) may be employed to remove traces of H2S that appear in the outlet of a POX reactor, and although the oxide can be regenerated, it will degrade over time. Consequently, there is a need to devise desulfurization systems that can operate at moderate temperatures and with high concentrations of steam in the fuel stream such that they can be used to clean up the syngas produced by POX. To address this issue, McDermott in the United States conducted tests on a regenerable zinc oxide bed, and several research groups have also attempted to improve the thermal durability of highly active ZnO nanoparticles by supporting them on alumina, silica and carbon. Ceramic monoliths coated with zinc oxide have also been examined for mobile fuel‐cell systems. 10.4.3 Steam Reforming Steam reforming is a mature technology that is practised industrially on a large scale for hydrogen production. Several detailed reviews of the technology have been published,9 and useful data for system design is provided by Twigg.10 The respective basic reforming reactions for methane and a generic hydrocarbon CnHm are: CH 4 H2O C n H m nH 2 O CO H2O h f CO 3H2 nCO CO2 H2 m 2 h f 206 kJ mol (10.3) (10.4) n H2 41 kJ mol 1 1 (10.5) The reforming reactions (10.3) and (10.4) and the associated WGS reaction (10.5) are usually conducted over a supported nickel catalyst at an elevated temperature, typically above 500°C. Reactions (10.3) and (10.5) are reversible and normally reach equilibrium over an active catalyst because the rates of reaction are very fast at such high temperatures. Furthermore, a catalyst that is active for reaction (10.3) nearly always promotes reaction (10.5). The combination of the two reactions taking place means that the overall product gas is a mixture of CO, CO2 and H2, together with unconverted CH4 and steam. 9 Rostrup‐Nielsen, JR, 1993, Production of synthesis gas, Catalysis Today, vol. 18, pp. 305–324; also: Trimm, DR, 2009, Fuel cells — hydrogen production — natural gas conventional steam‐reforming, in Garche, J, Dyer, CK, Moseley, PT, Ogumi, Z, Rand, DAJ and Scrosati, B (eds.), Encyclopedia of Electrochemical Power Sources, pp. 203–299, Elsevier, Amsterdam. 10 Twigg, M, 1989, Catalyst Handbook, 2nd edition, Wolfe, London. 277 Fuel Cell Systems Explained Figure 10.3 Equilibrium concentrations of products and reactants for steam reforming of methane at 100 kPa as a function of temperature. 60 H2 50 Component / vol.% 278 40 H2O 30 20 CH4 10 CO CO2 0 500 600 700 Temperature/°C 800 900 The actual composition of the product from the reformer is governed by the temperature of the reactor, the operating pressure, the composition of the feed gas and the proportion of the accompanying steam. Graphs and computer models derived from thermodynamic data are available to determine the composition of the equilibrium product gas under different operating conditions. By way of example, the composition of the output at 100 kPa is shown in Figure 10.3. As shown by reaction (10.3), three molecules of H2 and one molecule of CO are produced for every molecule of CH4 reacted. According to Le Chatelier’s principle, the equilibrium of this reaction will be moved to the right (i.e., in favour of hydrogen) if the pressure in the reactor is kept low. Conversely, increasing the pressure during the course of reforming will favour the formation of CH4, since moving to the left of the equilibrium reduces the number of molecules in the system. By contrast, the influence of pressure on the equilibrium position of the WGS reaction (10.5) is very small. Reactions (10.3) and (10.4) are usually strongly endothermic, and therefore heat needs to be supplied to the reforming reaction to drive it forward to produce H2 and CO. High temperatures (up to 800°C) therefore favour H2 formation, as shown in Figure 10.3. It is important to note that although the WGS reaction (10.5) occurs simultaneously with the steam reforming reaction on most catalysts, at the high temperatures that are necessary for hydrogen generation, the equilibrium point for the reaction is well to the left of the equation. Accordingly, by no means all the CO will be converted to CO2. Consequently, processing will be required for fuel‐cell systems that require low levels of CO and is examined further in Section 10.4.11. Remarkably, steam reforming is not always endothermic. For example, in the case of a petroleum hydrocarbon such as naphtha, with the empirical formula CH2.2, the reaction is most endothermic when the hydrocarbon reacts with steam to give only oxides of carbon and hydrogen. Steam reforming of naphtha is therefore most endothermic when carried out at high temperatures. It is less endothermic and eventually exothermic (liberates heat) as the temperature is lowered because this favours the reverse of reaction (10.3), i.e., the formation of CH4. The fact that, depending on the temperature and pressure of the reactor, the steam reforming of naphtha can be either endothermic or exothermic is demonstrated by the reactions listed in Table 10.6. Fuels for Fuel Cells Table 10.6 Typical heats of reaction in naphtha reforming at different temperatures, pressure and steam‐to‐carbon ratios. Pressure (kPa) Temperature (°C) Steam‐ to‐carbon ratioa ∆H (25°C) (kJ mol−1 CH2.2) 2 070 800 3.0 CH2.2 3H2O 0.2CH 4 0.4 CO2 0.4 CO 1.94 H2 1.81H2O +102.5 2 760 750 3.0 CH2.2 3H2O 0.35CH 4 0.4 CO2 0.25CO 1.5H2 1.95H2O +75.0 31 050 450 2.0 CH2.2 2H2O 0.75CH 4 0.25CO2 0.14 H2 1.5H2O –48.0 Reaction Source: Kramer, GJ, Wieldraaijer, W, Biesheuvel, PM & Kuipers, HCPE, 2001, The determining factor for catalysts selectivity in Shell’s catalytic partial oxidation process, American Chemical Society, Fuel Chemistry Division Preprints, vol. 46(2), pp. 659–660. a) Ratio of the number of moles of steam to the number of moles of carbon in the steam + fuel fed to the reactor. In summary, steam reforming has the following implications for fuel‐cell systems. The reforming of natural gas will invariably be endothermic, and heat will have to be supplied to the reformer at a sufficiently high temperature to ensure a reasonable degree of conversion. Naphtha and similar petroleum fractions (e.g., gasoline, diesel) will also react under endothermic conditions when hydrogen is the preferred product and the operating temperature is kept high. If, however, steam reforming of naphtha is carried out at more moderate temperatures (up to 600°C), the product will contain a significant concentration of CH4 and the reaction will be less endothermic. ‘Dry reforming’ (also known as ‘CO2 reforming’) can be carried out if there is no ready source of steam as follows: CH 4 CO2 2CO 2H2 h f 247 kJ mol 1 (10.6) This reaction may occur in internal‐reforming fuel cells, for example, the MCFC, when anode exhaust gas, which contains CO2 and water, is recycled to the inlet of the fuel cell. ‘Mixed reforming’ is a term that is sometimes used to describe a hybrid approach where both steam and CO2 are used to reform the fuel. Dry and mixed reforming have energy and environmental advantages compared with traditional steam reforming. The reactions are catalysed by nickel, but deactivation due to carbon formation and nickel sintering can be particularly severe. Hydrocarbons such as methane, light distillates and naphtha are not the only fuels suitable for steam reforming. Alcohols will also react with steam to produce hydrogen and carbon oxides. Methanol, for example, yields: CH3OH H2O 3H2 CO2 h f 49.7 kJ mol 1 (10.7) The fact that this reaction is mildly endothermic is one of the reasons why methanol has been favoured by some vehicle manufacturers as a possible fuel for FCVs. 279 280 Fuel Cell Systems Explained In addition, methanol is a suitable fuel for remote stationary power systems. Little heat needs to be supplied both to overcome the inherent heat losses and to sustain the methanol reforming reaction, which will readily occur at modest temperatures (e.g., 250–300°C) over catalysts of mild activity such as copper supported on zinc oxide. Although CO does not feature as a principal product of reaction (10.7), this does not mean that it will not be present. The WGS reaction (10.5) is reversible, and with an active catalyst, even at moderate temperatures, some CO will be produced from CO2 via the reverse shift reaction. Nonetheless, the CO level in reformed methanol will be acceptable for PAFCs and high‐temperature PEMFCs. For low‐temperature PEMFCs, the CO content has to be lowered using one of the methods described in Section 10.4.11. 10.4.4 Carbon Formation and Pre‐Reforming Carbon formation by decomposition of the fuel gas is one of the most critical problems that may be encountered during the operation of fuel‐cell systems. This reaction can take place in several areas of the system where hot fuel gas is present. Natural gas, for example, will decompose when heated in the absence of air or steam at temperatures above about 650°C via pyrolysis reactions such as: CH 4 C 2H 2 h f 75 kJ mol 1 (10.8) Higher hydrocarbons tend to decompose more easily than methane, and therefore the likelihood of carbon formation is greater with vapourized liquid petroleum fuels than with natural gas. Another source of carbon formation is from the disproportionation of CO via the Boudouard reaction, i.e., 2CO C CO2 (10.9) The reaction is catalysed by metals such as nickel, and consequently there is a high probability that it will occur on steam reforming catalysts that contain nickel and on the walls of stainless‐steel reactors. Fortunately, there is a simple expedient to reduce the degree of carbon formation via reactions (10.8) and (10.9), namely, to add excess steam to the fuel stream. The principal effect of this is to promote the WGS reaction (10.5), which results in a reduction of the partial pressure of CO in the fuel gas stream. Moreover, steam encourages the carbon gasification reaction, which is also very fast, as expressed by: C H2 O CO H2 (10.10) By considering the thermodynamics of the system, it is possible to calculate the minimum amount of steam that has to be added to a hydrocarbon fuel gas to avoid carbon deposition. The procedure is based on the assumption that a given mixture of fuel gas and steam interacts through reactions (10.3), (10.4) and (10.5) to produce a gas that is in thermodynamic equilibrium with respect to reactions (10.3) and (10.5) at the particular temperature and pressure of operation. The measured or observed partial pressures of CO and CO2 are used to calculate an equilibrium constant for the Boudouard reaction (10.9). If, for the observed temperature, the calculated constant is greater than the theoretical value, then carbon deposition is expected on thermodynamic grounds. Fuels for Fuel Cells If, however, the calculated constant is lower than theory predicts, then the gas is said to be in a safe region and carbon deposition will not occur. In practice, a steam‐to‐carbon ratio of 2.0–3.0 is normally employed in steam reforming systems so that carbon deposition may be avoided with an adequate margin of safety. A particular type of carbon formation that can occur on metals is known as ‘carburization’ and can lead to spalling of metal, which is referred to as ‘metal dusting’. Again, it is important to reduce the possibility of this phenomenon developing in fuel‐ cell systems, and some developers have used copper‐coated stainless steel in their fuel gas preheaters to minimize its occurrence. Carbon formation on steam reforming catalysts has been the subject of intense study and is well understood. Carbon produced via the pyrolysis reaction (10.8) and the Boudouard reaction (10.9) adopts different forms, of which the most damaging are filaments that appear to ‘grow’ on the nickel crystallites within the catalyst. Such carbon filaments can develop very rapidly, for example, if the steam supply to the reformer reactor is cut off suddenly. In such an event, the consequences can be disastrous with carbon formation occurring within seconds, thereby causing permanent breakdown of the catalyst that, in turn, leads to plugging of the reactor. This hazard emphasizes the importance of process control during the operation of fuel‐cell systems, especially if the steam is obtained by recycling product from the stack. Commercial steam reforming catalysts contain elements such as potassium and molybdenum that are known to inhibit carbon formation on the catalyst surface. The propensity for carbon formation can also be diminished by ‘pre‐reforming’ the fuel gas before it is fed to the reformer reactor. Pre‐reforming is a term commonly used in industry to describe the preferential conversion of high molecular weight hydrocarbons to hydrogen and carbon oxides via steam reforming at relatively low temperatures (typically 250–500°C). This process step is carried out in an adiabatic reactor, i.e., one that has neither external heat nor cooling applied. The gas from the exit of a pre‐reformer therefore consists mainly of methane and steam, together with small amounts of hydrogen and carbon oxides; the exact composition depends on the temperature of the reactor. A pre‐reformer arrangement that was designed for the demonstration of an internal‐ reforming SOFC stack is shown in Figure 10.4. The facility was built not only to reform the higher hydrocarbons from the natural gas but also to convert some 15% of the CH4 and thereby provide sufficient hydrogen at the stack inlet to maintain the SOFC anodes in a reduced condition. The pre‐reforming also had the beneficial effect of reducing thermal stresses within the stack since only 85% of CH4 was reformed internally. 10.4.5 Internal Reforming In the foregoing discussion of fuel processing, it has been assumed that steam reforming is conducted in one or more reactors that are external to the fuel‐cell stack — a practice therefore known as ‘external reforming’. For many years, developers have been aware that the heat to sustain the endothermic reforming of low molecular weight hydrocarbons (e.g., natural gas) can be provided by the electrochemical reaction in the stack. This feature has led to various elegant concepts of internal reforming that have been applied to MCFCs and SOFCs, on account of their high operating temperatures. The option is not available for low‐temperature PEMFC, alkaline fuel cell (AFC) and PAFC systems. In contrast to the steam reforming reactions (10.3) and (10.4), the fuel‐cell reactions are always exothermic, mainly due to heat production in the cell caused by 281 282 Fuel Cell Systems Explained 490 °C Natural gas 2.74 kW 520 °C 15 °C Anode exhaust burner and steam recovery subsystem Catalytic adiabatic pre-reformer Desulfurizer (activated carbon) 371 °C 618 °C 850 °C SOFC anodes 5.23 kW 700 °C Figure 10.4 Pre‐reformer system devised for a Siemens 50‐kW SOFC demonstration. internal resistances. Under practical conditions, with a cell voltage of 0.78 V, the heat liberated in a fuel cell amounts to 470 kJ per mol CH4. By carrying out steam reforming of CH4 within the MCFC or SOFC stack, about half of the heat liberated by the fuel‐cell reactions can therefore be utilized by the steam reforming reaction (10.3). This obviously reduces significantly the need for stack cooling, which is normally achieved by flowing excess air through the cathode. The lower airflow required by internal‐reforming stacks significantly improves the electrical efficiency of the overall system. Developers of internal‐reforming fuel cells have generally adopted one of two approaches. These are usually referred to as direct internal reforming (DIR) and indirect internal reforming (IIR) and are illustrated schematically in Figure 10.5. In some cases, a combination of both approaches has been adopted. The application of internal reforming offers several additional advantages compared with external reforming, as follows: ● ● ● ● ● System cost is reduced because a separate external reactor is not required. Less steam is necessary with DIR since the anode reaction in both the SOFC and the MCFC produces steam. The generation of hydrogen over the anode of a DIR cell may lead to a more even distribution of temperature throughout the cell. The methane conversion is high, especially in DIR systems where the cell consumes hydrogen as it is produced. The efficiency of the system is high because internal reforming can provide an elegant method of cooling the stack and thereby decrease the need for excess air at the cathode. This advantage, in turn, lowers the requirement for air compression and recirculation. Fuels for Fuel Cells Reforming catalyst CH4 + steam Anode exhaust CO2 + steam Indirect reforming Direct reforming Heat H2 and CO-rich gas Anode CO3= Electrolyte Cathode Air + CO2 Cathode exhaust Figure 10.5 Schematic representation of direct and indirect internal reforming. 10.4.5.1 Indirect Internal Reforming (IIR) Also known as ‘integrated reforming’, IIR involves conversion of CH4 in reformer reactors that are positioned in close thermal contact with the stack. Thus the reforming reaction and electrochemical reactions are separated. An example of this type of arrangement alternates plate reformers with groups of five or six cells. The reformate from each plate is fed to a neighbouring group of fuel cells. Although the high rate of heat transfer between cells and reformer plates provides a benefit, IIR systems suffer from the fact that heat is transferred effectively only from cells most adjacent to the reformers and also that steam for reforming must be raised separately. A variation in design places the reforming catalyst in the gas‐distribution path of each cell. 10.4.5.2 Direct Internal Reforming (DIR) In the DIR approach, the reforming reactions are performed in close proximity to the anode of each cell within the stack, as shown in Figure 10.5. In the case of the MCFC, reforming catalysts are placed within the gas flow channels of the anode. For the SOFC, the higher temperature of operation and the fact that the anode usually has a high nickel content and high surface area mean that steam reforming can occur directly on the anode, and therefore a reforming catalyst may not actually be required. Direct internal reforming has the distinct advantage of not only affording good heat transfer between the fuel cell and the reforming reaction but also ensuring chemical integration, that is, product steam from the anode reaction is used directly by the reforming reaction. Although endothermic reforming absorbs some of the heat produced by MCFC or SOFC stacks with internal reforming, it is not sufficient to remove completely the need to cool the stack. Indeed, the combination of endothermic and exothermic reactions that occur in a stack with internal reforming can promote large temperature gradients within the cell and stack hardware. Such behaviour can arise because steam reforming is both fast and endothermic and thus cause a sharp fall in temperature close to the 283 284 Fuel Cell Systems Explained anode inlet. By contrast, the exothermic fuel‐cell reactions lead to a rise in temperature towards the anode outlet. To minimize such variations in temperature within an internal‐ reforming MCFC or SOFC stack, a counter‐flow configuration of fuel and oxidant is usually employed rather than co‐flow. Finally, it should be understood that internal reforming may, in principle, be applied to several hydrocarbon fuels that include natural gas and vapourized liquids such as naphtha and kerosenes. Coal gases are particularly attractive for internal‐reforming MCFC and SOFC stacks, since not only are CO and H2 consumed directly as fuels, but residual CH4 (as may be present in the product from the BG–Lurgi slagging gasifier; see Table 10.2) is also internally reformed. 10.4.6 Direct Hydrocarbon Oxidation Direct hydrocarbon oxidation refers to the conversion of a hydrocarbon fuel directly to steam and CO2 without first undergoing conversion to hydrogen by steam reforming or POX. The Gibbs free energy change for the direct oxidation reaction of CH4 to CO2 and steam is −796.5 kJ mol−1, which is very close to the change in enthalpy (∆h  = −802.5 kJ mol−1). In other words, if CH4 could be oxidized directly, most of the heat of reaction would be converted directly into electricity, with a maximum efficiency of g h 796.5 100 99.2% 802.5 (10.11) The high efficiency should be qualified because H2O and CO2 produced by the direct oxidation of methane can react with fresh methane via the reforming reaction over the nickel catalyst in an MCFC or the nickel‐containing anode in the SOFC. The only way therefore to establish whether direct oxidation is actually taking place in a fuel cell is to use an anode catalyst that does not promote the steam reforming reaction. The first demonstration of the feasibility of direct methane oxidation was provided by replacing the nickel cermet that was normally employed as the anode in SOFCs with a copper cermet.11 The copper cermet also addressed the issue of carbon formation. As discussed in Section 10.4.4, carbon has a propensity to deposit on nickel‐containing materials at temperatures even as low as 600°C. With a copper‐cermet anode, however, carbon formation is avoided and direct oxidation can occur. Furthermore, it has been found that copper cermets are stable in hydrocarbon environments as long as the temperature is limited. Apart from the increased conversion efficiency afforded by direct oxidation, a further benefit is that steam does not need be supplied with the fuel and thereby leads to systems of much simpler design. The benefits of high conversion efficiency and lower steam requirement and resistance to carbon fouling have encouraged various investigations of novel ceramic anodes for SOFCs made of mixed ionic–electronic conductor (MIEC) materials, as has been discussed in Section 9.2.3.3, Chapter 9. 11 Park, S, Vohs, JM and Gorte, RJ, 2000, Direct oxidation of hydrocarbons in a solid‐oxide fuel cell, Nature, vol. 404, pp. 265–267. Fuels for Fuel Cells It is also known that direct hydrocarbon oxidation can occur in fuel cells that use aqueous acid electrolyte. For instance, in the 1960s, propane was found to decompose on platinum catalysts at moderate temperatures (below 200°C) to form protons, electrons and CO2 (via reaction with water from the aqueous acid electrolyte). As in both PEMFC and PAFC systems, the protons migrate to the cathode where they are oxidized with air and electrons to form water. Researchers may find it worthwhile to revisit some of this earlier work to examine whether it holds the key to new ways of fuelling fuel cells. 10.4.7 Partial Oxidation and Autothermal Reforming As an alternative to steam reforming, methane and other hydrocarbons may be converted to syngas and hence hydrogen for fuel cells by ‘partial oxidation’, i.e., CH 4 1 O 2 2 CO 2H2 h f 247 kJ mol 1 (10.12) Partial oxidation can be carried out at high temperatures (typically 1200–1500°C) in the gas phase without a catalyst. The process has the advantage over catalytic processes in that it is not necessary to remove materials such as sulfur compounds, although sulfur would have to be extracted from the product at a later stage (as sulfur dioxide, SO2) if it is to be fed to a fuel‐cell stack. High‐temperature POX can also handle much heavier petroleum fractions than catalytic processes and is therefore attractive for processing diesel, other liquid fuels and residual oil fractions. Gas‐phase POX has been performed in large facilities by several companies, but it does not scale down well, and management of the reaction is problematic. A greater level of control of the POX reaction can be achieved by lowering the temperature and employing a catalyst. The process is then termed ‘catalytic partial oxidation’ (CPO). Catalysts for CPO tend to be either a platinum‐group metal or nickel supported on a ceramic oxide. At temperatures of around 1000°C, the CPO reaction is very fast, and an industrial version of the process developed by Eni S.p.A is known as ‘short contact time catalytic partial oxidation (SCT‐CPO)’. The CPO reaction occurs within a few milliseconds in a high‐temperature and thin (<1 µm) solid–gas interphase zone that surrounds the catalyst particles and does not propagate into the gas phase. These conditions favour the formation of the primary reaction products (H2 and CO) and allow the use of several hydrocarbon feedstocks that may also contain sulfur or aromatic compounds. Catalytic partial oxidation has been the subject of much research over the past 10–15 years with Haldor Topsøe A/S and Eni S.p.A establishing the technology for industrial production of hydrogen, gas‐to‐liquid (GTL) processes and refinery operations. It should be noted that the POX reaction (10.12) produces less H2 per molecule of CH4 than the steam reforming reaction (10.3). Accordingly, POX (either gas phase or CPO) is usually less efficient than steam reforming for fuel‐cell applications. Reaction (10.12) is effectively the summation of the steam reforming and oxidation reactions; it can be considered that about half of the fuel that is converted into hydrogen is oxidized to provide heat for steam reforming. Heat from the fuel cell is not utilized by POX or CPO reactors, and the net effect is a reduction in overall system efficiency. Another disadvantage of POX occurs when air is used as the oxidant. The presence of nitrogen from the air reduces the partial pressure of hydrogen at the fuel cell that, in turn, lowers the cell voltage (as determined by the Nernst equation) so that again there is a decrease 285 286 Fuel Cell Systems Explained in system efficiency. To offset these negative aspects, the key advantages of POX with air are that steam is not required and the process can be achieved in a much smaller reactor than that used for steam reforming. Consequently, POX may be considered for applications where system simplicity is regarded as more important than high electrical conversion efficiency as, for example, in small‐scale cogeneration systems. If the steam reforming and CPO reactions are carried out in the same reactor and over the same catalyst, the process is generally known as ‘autothermal reforming’. It should be pointed out, however, that the terms ‘autothermal reforming’ and ‘partial oxidation’ are often loosely employed in the literature and care needs to be exercised when comparing data from reports of different fuel‐processing systems. Autothermal reforming involves feeding fuel, steam and oxygen (or air) simultaneously to the reactor. By adjusting the flows of these reactants, as well as allowing for heat transferred out of the reactor in the product gas and any inevitable heat losses to the environment, a thermoneutral condition can be sustained in which the heat generated by POX exactly matches the heat consumed by steam reforming. In terms of reaction mechanisms, several studies have determined the relative rates of the steam reforming and the CPO reactions when conducted simultaneously over different catalysts. From such work, it has been concluded that in many cases the CPO reaction is brought to equilibrium much faster than the steam reforming reaction. This sequence, which has been termed the ‘indirect mechanism’ of POX, is promoted by supported nickel catalysts. By contrast, supported ruthenium and rhodium catalysts activate an alternative ‘direct mechanism’ whereby oxidation, steam reforming and shift reactions occur in parallel. In some processes that have been described as CPO, both steam and oxidant are fed with the fuel. The Shell POX process is one example of such practice (Kramer et al., 2001).12 The process uses a proprietary vertical tubular reactor design that contains a bed of platinum‐ group catalyst. The partial oxidation reaction occurs at the top of the bed where the rate is limited by mass transfer of the reactants. Further down the bed, the steam reforming and WGS reactions bring the gas to equilibrium. In other examples of CPO where the steam oxidation and reforming reactions operate in parallel, the rates of the reactions are not limited by mass transfer and are brought to equilibrium without gain or loss of heat. The advantages of autothermal reforming and CPO are that less steam is required compared with conventional steam reforming and that all of the heat for the reforming reaction is provided by partial combustion of the fuel. Both approaches are attractive for application with PEMFCs and AFCs as these cells do generate heat at a temperature that is sufficiently high for the steam reforming of fuel. 10.4.8 Solar–Thermal Reforming The heat for steam reforming of hydrocarbons such as methane can, in principle, be obtained directly from the sun. As the resulting syngas would contain a substantial amount of embodied solar energy (up to 25%), solar–thermal reforming offers the 12 Kramer, GJ, Wieldraaijer, W, Biesheuvel, PM and Kuipers, HCPE, 2001, The determining factor for catalysts selectivity in Shell’s catalytic partial oxidation process, American Chemical Society, Fuel Chemistry Division Preprints, vol. 46(2), pp. 659–660. Fuels for Fuel Cells prospects of high thermal efficiencies and greatly reduced emissions of carbon dioxide. Moreover, the emissions would be in concentrated form and thus more amenable to gas separation. Over the past 20 years, a considerable amount of work has been undertaken on solar–thermal reforming in several countries, and broadly three types of solar receiver–reactor have been proposed, 13 namely: 1) Molten sodium, reflux heat‐pipe receiver–reactor. 2) Directly irradiated, volumetric receiver–reactor (DIVRR). 3) Cavity receiver with conventional, tubular catalytic reactors. In each design, solar energy is directly focused on the receiver–reactor or transmitted via the use of heat‐transfer fluid from the solar receiver to the reactor. To achieve the required high temperature, three common methods are employed: 1) A simple parabolic dish that focuses the sun’s rays on to a thermal receiver mounted above the dish at its focal point. 2) An array of thousands of individual mirrors (‘heliostats’) around a central receiver set on top of a tall tower. 3) Parabolic trough mirrors that track the sun as it crosses the sky and that have receivers located at their foci. The three methods are illustrated in Figure 10.6. Solar–thermal reforming has a number of challenges. The first is that a large area of land has to be available to accommodate the dish, troughs or heliostats. Another concern is that solar energy is highly variable through the day and is not available at night. Therefore, some means of energy or heat storage has to be provided to ensure continuous production of hydrogen. There are also issues with the reforming catalyst in that temperatures above 850°C are required with conventional reforming catalysts to obtain an adequate yield of hydrogen, and both system performance and catalyst integrity are compromised if the temperature is allowed to fall and fluctuate excessively. The situation can be improved by employing catalysts that incorporate more active precious metals, e.g., ruthenium or rhodium, or by carrying out the reforming in a membrane reactor to achieve direct separation of hydrogen. 10.4.9 Sorbent‐Enhanced Reforming By mixing a solid absorbent for CO2 (e.g., calcined dolomite, CaO) with the reforming catalyst, it is possible to combine both the steam reforming and WGS reactions into a single step and, at the same time, reduce the temperature of the former from about 900 to 400–500°C. Removal of CO2 from the reaction zone influences the equilibrium of the combined reaction so that the production of hydrogen is enhanced, while CO is oxidized to CO2. A typical product gas from the sorbent‐enhanced reforming of CH4 is composed of 90 vol.% H2 and about 10 vol.% unreacted CH4, with a small percentage of CO2 and a trace of CO. 13 Stein, W, Edwards, J, Hinkley, J and Sattler, C, 2012, Natural gas: solar‐thermal steam‐reforming, in Garche, J, Dyer, CK, Moseley, PT, Ogumi, Z, Rand, DAJ and Scrosati, B (eds.), Encyclopedia of Electrochemical Power Sources, pp. 300–312, Elsevier, Amsterdam. 287 288 Fuel Cell Systems Explained (a) Receiver or engine/receiver Concentrator reflective surface (b) Receiver Heliostats Tower (c) Concentrator reflective surface Receiver Tracking mechanism Figure 10.6 Three main designs of solar furnace: (a) parabolic dish, (b) central receiver and (c) parabolic trough. When the sorbent has become saturated with CO2, it is regenerated by purging with steam. The exit stream is condensed and the released CO2 can be captured ready for compression and conveyance to underground storage. For continuous operation, two parallel reactors are required — one bed undergoes reaction, while the other is regenerated. The sorbent‐enhanced process offers an elegant approach in that it removes the need for a second shift and for additional gas‐separation steps. Also, the lower temperature of operation results in reduced heat loss and permits the use of cheaper materials of construction. Sorbent‐enhanced reforming has been the subject of research in many countries for the past twenty or so years, but there are technical challenges to be overcome before it can become a commercial proposition. These include the long‐term durability of the sorbent and catalyst and the scale‐up and switching of reactant streams to the two reactors. Fuels for Fuel Cells 10.4.10 Hydrogen Generation by Pyrolysis or Thermal Cracking of Hydrocarbons An alternative to all of the aforementioned methods of generating hydrogen from hydrocarbons is simply to heat the fuel in the absence of air, a process commonly referred to as ‘pyrolysis’. The hydrocarbon ‘cracks’ or decomposes into hydrogen and solid carbon (soot). The process is ideally suited to simple hydrocarbon fuels; otherwise various by‐ products may be formed (e.g., acetylene and other alkenes). The advantage of thermal cracking is that the hydrogen produced can be very pure and, if no catalyst is present, the carbon can be separated as a solid, which is easier to handle than gaseous CO2. On the other hand, if pyrolysis is enhanced with a catalyst, then extraction of carbon becomes a challenge. In principle, removal can be achieved by shutting off the supply of fuel and admitting air to the reactor to burn off the carbon as CO2. Switching the flow of fuel and oxidant is simple in theory, but there are real difficulties, not least of which are the safety implications of admitting fuel and air into a reactor at high temperatures. Control of the pyrolysis is critical; otherwise too much carbon will accumulate and consequently will irrevocably damage any catalyst that may be employed. Excessive carbon may also be formed in the absence of a catalyst — indeed, to such an extent that the reactor becomes plugged and accordingly creates difficulty with the removal of the solid and/or with establishing a sufficient flow of oxidizing gas to burn off the deposited material. Despite these substantial problems, pyrolysis is being considered seriously as an option for some fuel‐cell systems. For instance, thermal cracking of propane has been proposed as a source of hydrogen for small PEMFC systems.14 The carbon build‐up problem with thermal pyrolysis can be avoided by the application of thermal plasma technology (see Box 10.2). This is because a thermal plasma, which is created by an electric arc, is characterized by temperatures of 3000–10 000°C, and under such conditions no catalyst is required. A further advantage is that the reactions are very fast and therefore allow a much more compact and lighter design than that for a conventional catalytic processor. By adding steam and oxygen, a plasma reactor can operate as a reformer and produce a syngas. Non‐thermal plasmas (notably the sliding discharge reactor; see Box 10.2) have also been investigated for reforming fuels.15 A compelling feature of plasmas is their ability to break down high molecular weight hydrocarbons. In most cases, however, control of the plasma is both difficult and energy demanding and thereby results in poor conversion and variable product composition. Recently, researchers have sought to overcome these limitations by combining a reforming catalyst either within or external to the plasma reactor. Although still at the laboratory stage, such approaches may lead to an energy‐efficient device that fulfils the needs of a small‐scale hydrogen generator for fuel‐cell systems.16 14 Wang, Y, Shah, N and Huffman, GP, 2003, Production of pure hydrogen and novel carbon nanotube structures by catalytic decomposition of propane and cyclohexane, Prepr. Pap.‐Am. Chem. Soc., Div. Fuel Chem., vol. 48(2), p. 901. 15 Paulmier, T and Fulcheri, L, 2005, Use of non‐thermal plasma for hydrocarbon reforming, Chemical Engineering Journal, vol. 6, pp. 59–71. 16 Tu, X and Whitehead, C, 2014, Plasma dry reforming of methane in an atmospheric pressure AC gliding arc discharge: Co‐generation of syngas and carbon nanomaterial, International Journal of Hydrogen Energy, vol. 39(18), pp. 9658–9669. 289 290 Fuel Cell Systems Explained Box 10.2 What is a plasma? A plasma is a partially or fully ionized gas composed of positive ions and free electrons in proportions that result in more or less no overall electric charge. Based on the relative temperature of the electrons, ions and neutral particles, plasmas are classified as ‘thermal’ or ‘non‐thermal’. Thermal plasmas have the electrons and heavier particles at the same temperature (i.e., they are in thermal equilibrium with each other), whereas non‐thermal plasmas have the ions and neutrals at much lower temperatures (e.g., room temperature) and the electrons are much ‘hotter’. A thermal plasma is formed when an arc is discharged between two electrodes. Examples of non‐thermal plasmas are as follows. ● ● ● ● ● Glow discharges are produced when a high DC voltage is applied between metal electrodes in a reactor in which gas is at a low pressure (0.01–1.0 kPa) — the familiar neon and fluorescent light tubes are good examples. A glow discharge can also be formed with an alternating current at radio frequencies. Silent discharges, also known as ‘corona’ discharges, are obtained in gases at low or moderate pressures in which the discharge is struck between conductors with rounded or pointed tips. Dielectric barrier discharge is created when a high voltage is applied between two electrodes that have non‐conductive coatings to prevent an arc from developing. Capacitative discharges are generated when radio‐frequency (RF) power is applied to one electrode with a grounded electrode held at a small distance of separation, typically 1 cm. Electrodeless or microwave discharges are produced by an electric field from a coil wrapped around the reactor. 10.4.11 Further Fuel Processing: Removal of Carbon Monoxide A steam reformer running on natural gas and operating at atmospheric pressure with an outlet temperature of 800°C produces a gas that consists of some 75 vol.% H2, 15 vol.% CO and 10 vol.% CO2 on a dry basis. It is necessary to lower the concentration of CO in such a gas before it can be fed to a PAFC or a PEMFC. As noted earlier, the WGS reaction (10.5) takes place on active catalysts at the same time as the basic steam reforming reaction (10.4). If both reactions are at thermodynamic equilibrium, high temperatures favour the production of CO, i.e., reaction (10.5) is shifted to the left. The first approach to reducing the CO content of a reformed fuel is therefore to cool the product gas from the steam reformer and pass it through a reactor with a catalyst that only promotes the WGS reaction. This has the effect of converting CO into CO2. Depending on the composition of the product gas from the reformer, more than one shift reactor may be necessary to lower the CO to an acceptable level. An iron–chromium catalyst is found to be effective for promoting the WGS reaction at relatively high temperatures (400–500°C), and this may be followed by further cooling of the gas before passing to a second low‐temperature reactor (200–250°C) with a copper catalyst. At this lower temperature, the proportion of CO exiting the reactor will typically be about 0.25–0.5 vol.%, and so these two stages of shift conversion are sufficient to decrease the CO content to meet the needs of the PAFC. On the other Fuels for Fuel Cells hand, the level is equivalent to 2500–5000 ppm and exceeds the limit for typical PEMFCs by two orders of magnitude. It is similar to the level of CO in the product from a methanol reformer, and therefore further fuel processing is required for a PEMFC system. Until recently, the WGS reactors in fuel‐cell systems at the kW-scale have utilized industrial iron and copper catalysts. These are easily poisoned by sulfur and also present a safety hazard as the catalyst in the reduced state becomes pyrophoric when exposed to air. In addition, the reactors are large in comparison with the stacks, the reformer reactor and other balance‐of‐plant components. To improve the situation, some developers of fuel cells are working on novel catalysts that are able to operate with higher space velocities (i.e., are more active) and at low temperatures. The Selectra™ Shift catalyst supplied by Engelhard is a non‐pyrophoric base metal material that is an alternative to the traditional low‐temperature ZnO catalyst. Cobalt oxide and molybdenum oxide supported on alumina has also been used as a low‐cost, low‐ temperature WGS catalyst in a sulfided form. This particular composition is insensitive to sulfur, a feature that makes it even more attractive in comparison with the traditional zinc–copper materials. Precious metal catalysts, e.g., platinum on ceria (CeO2) developed by Nextech in the United States, platinum on mixed cerium–lanthanum oxides and gold on ceria, are non‐pyrophoric, are tolerant to sulfur and may become viable options, provided costs can be kept reasonably low. Some transition metal carbides also provide sulfur tolerance and relatively high WGS activity. Nevertheless, there is clearly room for improvement in WGS catalysts for fuel‐cell systems, and consequently research in this area is quite active at the moment. In the end, there will probably be a trade‐off between cost and performance, as well as for the other catalytic steps in fuel processing. For the PEMFC, further CO removal is essential after the WGS reactors. This is usually carried out by one of the following three methods: 1) Preferential oxidation (PROX) in a reactor to which a small amount of air (around 2 vol.%) is added to the fuel stream, which then passes over a precious metal catalyst. Typical catalysts include Pt–Al2O3, Ru–Al2O3, Rh–Al2O3, Au–MnOx, Pt‐Ru–Al2O3 and Ir‐based materials such as 5 wt.% Ir–(CoOx–Al2O3)–carbon. The catalyst preferentially absorbs the CO, rather than the H2, and reacts with the oxygen in the air to produce CO2. As well as the obvious issue of the cost of the precious metal catalysts, the PROX reaction is exothermic, and therefore the reactor may require cooling so that temperature control can be problematic. Since there is the presence of H2, CO and O2 at an elevated temperature with a noble metal catalyst, measures must be taken to ensure that an explosive mixture is not produced. This eventuality could be a particular concern where the flow rate of the gas is highly variable, such as with a PEMFC system on board a vehicle. 2) Methanation of the CO is an approach that reduces the danger of producing explosive gas mixtures. The reaction is the opposite of the steam reforming of CH4, i.e., reaction (10.3), namely, CO 3H2 CH 4 H2O hf 206 kJ mol 1 (10.13) The reaction is conducted in a small catalytic reactor placed close to the fuel‐cell inlet. Methanation has the obvious disadvantage that it consumes some hydrogen, which will reduce the electrical efficiency of the fuel‐cell system by a small amount. 291 292 Fuel Cell Systems Explained The methane produced by reaction (10.13) does not poison the PEMFC catalyst but simply dilutes the fuel within the stack. Such minor effects can be tolerated in a PEMFC system in which a simple methanation catalyst operating at about 200°C can reduce the level of CO to less than 10 ppm. In the case of reformed methanol, the methanation catalyst will also ensure that any unconverted methanol from the reformer reactor is converted variously to CH4, H2 and CO2. 3) Palladium or platinum membranes can be used to separate and purify the H2. This is mature technology that has been used for many years to produce H2 of exceptional purity, despite the fact that such membranes are expensive. Further discussion of membranes is given in Section 10.5. Workers at the National Renewable Energy Laboratory in the United States have been attempting to improve the WGS reaction through the use of bacteria, and this research has been taken up by other groups in recent years. If this method proves to be successful, there may be less need for the final CO clean‐up stage. To date, however, the reaction rates for the biological procedure have been two orders of magnitude below those experienced with the traditional catalytic systems. Further discussion of biological systems is given in Section 10.10. Electrochemical oxidation is a further and very different approach to CO removal for the PEMFC. Two methods have been investigated, namely, (i) oxidation in a reactor placed before the fuel cell17,18 and (ii) oxidation within the anode compartment of the PEMFC stack itself.19 Both methods involve two reaction steps: absorption of CO on the catalyst followed by oxidation of the absorbed CO to CO2, which is desorbed. In method (i) the catalyst is the anode of an electrochemical cell in which surface oxygen is generated by passage of an electric current. In method (ii) the first step is absorption of CO directly on the supported Pt of the fuel‐cell anode, and the second step is instigated by momentarily disconnecting the load on the fuel cell and applying a positive potential to the anode. The latter serves to generate oxygen directly on the surface of the anode electrocatalyst (essentially by electrolysis of water within the cell, i.e., the cell is switched from fuel cell to electrolyser mode). Atomic oxygen on the anode directly oxidizes the CO to CO2, and when this process is complete, the fuel‐cell mode is resumed. Electrochemical oxidation within the cell is a simple concept but has its own challenges, namely, increased catalyst degradation caused by switching between oxidizing and reducing conditions and reduced cell efficiency since the oxidation to CO2 effectively imposes an additional parasitic power load on the fuel‐cell stack. Pressure swing adsorption (PSA) is a further method of hydrogen purification that can be applied to reformates. In this process, the reformer product gas is passed into a reactor that contains a material that preferentially absorbs hydrogen. After a set time, the reactor is isolated and the feed gas is diverted into a parallel reactor. The first reactor is then depressurized and thereby enables pure H2 to desorb from the material. The process is repeated and the two reactors are alternately pressurized and depressurized. 17 Balasubramanian, S, 2011, Electrochemical oxidation of carbon monoxide in reformate hydrogen for PEM fuel cells, PhD Thesis, University of South Carolina. 18 US Patent 6245214 B1, Electro‐catalytic oxidation (ECO) device to remove CO from reformate for fuel cell application. 19 US Patent 5601936, A Method of operating a fuel cell. Fuels for Fuel Cells 10.5 Membrane Developments for Gas Separation As an alternative to PSA and the other methods of gas purification described in Section 10.4.11, research is in progress to find an effective means to separate hydrogen from CO2 subsequent to the WGS reaction. In particular, it is desirable to separate the gases while hot and so conserve heat energy. The work is directed primarily towards the use of membranes that are selective to the diffusion of H2 (a small molecule) while excluding CO2 and other species. By using ceramic membranes, it should be possible to effect the separation at close to the temperature of the WGS reaction or even the reforming reaction. Various membrane reformers and membrane WGS reactors have been proposed, and although these are already applied to some extent, there remains great scope for improving the performance and lowering the costs of all types of membrane used for gas separation. In general, membranes may be classified as (i) non‐porous, e.g., membranes based on metals, alloys, metal oxides or metal–ceramic composites, or (ii) ordered microporous materials, e.g., dense silica, zeolites and polymers. 10.5.1 Non‐Porous Metal Membranes Metal‐based, non‐porous membranes can produce an H2 stream of very high purity that can be used directly in a fuel cell. The separation process relies on the ability of the metal to allow only the diffusion of H2. The permeation of H2 through metals such as palladium and its alloys is thought to proceed via several steps, namely, adsorption of molecular H2, dissociation to the monatomic form, ionization, diffusion of the hydrogen ions or atoms under a concentration gradient through interstices within the metal lattice, reassociation and, finally, desorption. It is the surface properties of palladium and its alloys that give rise to high catalytic activity for absorption, dissociation and desorption. Hydrogen flux density (and hence membrane performance) is a function of the inherent diffusion characteristics of the material, and membranes above about 10 µm in thickness are limited in performance by diffusion. In recent years, therefore, research has been directed towards making thinner palladium and palladium alloy membranes. This also has a benefit in terms of cost reduction. Very thin films (<10 µm) of palladium or its alloys can be supported on a porous metal or ceramic substrate. Improved performance can be obtained by raising the metal temperature. Indeed, it has been found advantageous to maintain palladium and its alloys above the critical temperature (293°C for pure palladium) to avoid stresses otherwise caused by the interaction between the two forms of palladium hydride (PdH) with different crystal structures that can coexist at low temperatures. Such interactions can lead to mechanical degradation and failure of the metal — a process known as ‘hydrogen embrittlement’. Unlike palladium and its alloys that are crystalline in nature, a class of amorphous alloy membranes is emerging for H2 separation at high temperature. These are composed primarily of nickel and early transition metals (i.e., titanium, zirconium, niobium, hafnium and tantalum). Such alloys have a random atomic configuration, and, as with crystalline alloy membranes, H2 migrates through the alloy via interstices between the metal atoms. The first nickel‐based amorphous alloy membrane to be reported was Ni64Zr36. The permeability of this alloy is about 10% that of palladium under similar conditions, and its operation is limited to relatively low temperatures, typically below 400°C. 293 294 Fuel Cell Systems Explained 10.5.2 Non‐Porous Ceramic Membranes A further class of non‐porous membrane is based on proton‐conducting metal oxides from the perovskite family. These ceramic materials have the general formula ABO3 or A1 x A x B1 y B y O3 where x and y are fractions of dopants in the A‐ and B‐sites, respectively, and δ is the number of oxygen vacancies. Considerable research has been undertaken on SrCeO3 and BaCeO3 that have been variously doped with trivalent cations such as those of yttrium, ytterbium or gadolinium. These oxides can operate at much higher temperatures (up to 800°C) than metal membranes, which makes them suitable for use in membrane reformer reactors. Unfortunately, however, the oxides are difficult to make and suffer from low mechanical strength, poor H2 flux and low chemical stability in the presence of both CO2 and water. These features make them less applicable for medium‐temperature separation of H2, for example, in conjunction with a WGS reactor. By contrast, zirconates (e.g., barium zirconate (BaZrO3)) have better chemical stability but lower protonic conductivity. Cerate– zirconate solid solutions that combine the favourable attributes of each have been developed. Ceramic–metallic (‘cermet’) materials have also been investigated as H2 separation membranes. As with the MIEC oxides discussed in Section 9.2.3.3, Chapter 9, the cermet can exhibit improved performance through the incorporation of an electronically conducting metal to enhance the proton conductivity of the ceramic oxide. 10.5.3 Porous Membranes A porous hydrogen separation membrane usually consists of a thin layer of a microporous sieve material such as silica, carbon or zeolite on a thicker and highly porous support. To maximize the flux, the microporous material is made much thinner than the dense membranes discussed earlier, i.e., typically of the order of tens to hundreds of nanometres. Hydrogen is transported predominantly through the pores of the membrane by molecular diffusion, which is a purely physical process with a performance that is determined by the pore diameter of the membrane. To separate H2 effectively, the pores must be less than 1 nm in diameter. A variety of established manufacturing techniques can be used to fabricate microporous membranes on either metallic or ceramic macroporous components. Silica membranes, for example, are made by coating the surface of a porous material with a silica‐based chemical precursor. Two procedures can be employed. One method involves the dipping of a suitable porous support into a sol–gel that contains a silica precursor such as tetramethyl orthosilicate or tetraethyl orthosilicate. Multiple applications are usually required to eliminate pinhole defects in the membrane. Similar precursors are used in the alternative method but are applied by means of chemical vapour deposition (CVD). In either procedure, the silica that is formed is partially densified by heating to form a xerogel, with the desired pore‐size distribution. If heated further, the micropores will close and render the membrane ineffective. For this reason, microporous membranes are usually limited to gas separations below about 600°C. Membrane separators are usually produced in ‘tube‐and‐shell’ configurations that are assembled in multi‐tube modules for efficient distribution of feed and product gases; see Figure 10.7. Fuels for Fuel Cells Hydrogen-selective membrane film CO CO H Hydrogen Carbon dioxide H H H H H H 20 nm Hydrogen Carbon dioxide Figure 10.7 Schematic of a membrane reactor. 10.5.4 Oxygen Separation The membranes so far discussed are used to separate H2 from a gas mixture. Also relevant for fuel‐cell systems are membranes that are able to separate oxygen from air. If oxygen, rather than air, is supplied to a CPO reactor or autothermal reformer, the syngas product contains no nitrogen. This is beneficial on two counts. First, the absence of nitrogen ensures that no ammonia can be formed in the processing steps. Ammonia, which is discussed further in Section 11.6.3, Chapter 11, is harmful to PEMFCs and can cause permanent degradation. Nitrogen by itself is generally a diluent in the fuel gas stream and passes through most stacks without undergoing any chemical reaction. Nevertheless, the diluting effect of nitrogen on the reactive gases inevitably reduces the performance of all fuel cells, as indicated by the Nernst equation. Nitrogen also lowers the effectiveness of other processes that may utilize the syngas, such as GTL. Industrially, oxygen is separated from air by cryogenic processes or by PSA in which the gas is absorbed on beds of zeolite that are alternatively pressurized or depressurized. Pressure swing absorption is also used for oxygen concentrators used in medicine. Oxygen‐ion‐conducting membranes, as used in SOFC electrolytes, can also be employed for oxygen separation and are starting to be adopted in fuel‐cell systems. To avoid a charge distribution developing across the membrane caused by the migration of ions, an MIEC material is preferred for such duties. 10.6 Practical Fuel Processing: Stationary Applications 10.6.1 Industrial Steam Reforming Before practical fuel processing for fuel‐cell systems is considered, it is instructive to consider the operation of a typical industrial steam reforming plant. Such facilities have been built for many decades to provide H2 for both oil refineries and chemical plants 295 296 Fuel Cell Systems Explained (mainly to produce ammonia for the fertilizer industry). Industrial reformer systems usually produce between 7 and 30 million normal cubic metres (Nm3) of H2 per day. The systems, which consist of a number of tubular reactors packed with catalyst pellets, operate at temperatures up to 850°C and pressures up to 2500 kPa. The reactors are generally around 12 m in length and must be made from expensive alloy steels to endure both the high temperatures and the reducing gas conditions. Such reformers can be scaled down reasonably easily to give H2 outputs of some 0.1–0.3 million Nm3 day−1. It was noted in Section 2.3, Chapter 2, that the LHV for the enthalpy of combustion of H2 is −241.8 kJ mol−1. Hydrogen supplied at a rate of 1.0 Nm3 h−1 when combusted will therefore produce about 3 kW of heat energy. If the H2, rather than being combusted, is fed to a fuel‐cell system that has an overall electrical efficiency of 40% (LHV), then clearly the energy produced by such a fuel‐cell systems fed with 1.0 Nm3 h−1 of H2, for example, would be 0.4 × 3 = 2.4 kW. Unfortunately, scaling down an industrial reformer that is sized for several million Nm3 per day to one that provides only a few Nm3 h−1 of H2 is not a practical proposition. Conventional tubular reformers are expensive because of the need to run at high temperatures and pressures, and they are large in terms of footprint area and weight. Accordingly, alternatives have to be provided to suit the much smaller demands of fuel‐ cell systems. 10.6.2 Fuel‐Cell Plants Operating with Steam Reforming of Natural Gas For stationary power plants that employ PEMFC and PAFC stacks, steam reforming of natural gas is the preferred option for generating H2 because it gives high fuel‐conversion efficiency for the system as a whole. The technology has been applied for many years in facilities of between about 50 kW and several MW. In both PEMFC and PAFC systems, the sulfur removal can be achieved by HDS. For PAFC systems, two stages of shift are required for lowering the level of CO in the reformed gas. In the case of a PEMFC, however, a further CO removal step will also be necessary. With such systems, a degree of process integration is required so that heat from the fuel cell is utilized for various preheating duties. The chemical processes (desulfurization, steam reforming, WGS and CO removal) each take place at a different temperature. Consequently, there are a number of temperature changes to be made. The minimum requirements are as follows: ● ● ● ● ● ● Initial heating of the dry fuel gas to approximately 300°C prior to HDS. Further heating of gas and steam prior to steam reforming at 600°C or higher. Cooling of the reformer product gas to approximately 400°C for the high‐temperature WGS reaction. Further cooling to approximately 200°C for the low‐temperature WGS reaction. Temperature adjustment prior to entry into the CO removal step or directly into the fuel cell (depending on the fuel‐cell type). Heating of water to produce the steam required for the steam reformer. In addition to these six temperature changes, steam reforming demands high‐ temperature heat. This requirement can be met by burning the anode exhaust gas, which always contains some unconverted fuel. Further heating of the anode exhaust gas may also prove advantageous. Similarly, preheating the air to the burner will yield a higher combustion temperature. In such fuel‐cell systems, therefore, some gases have to Fuels for Fuel Cells Steam in Exhaust gas out 120 400 To fuel cell Anode A Air in 300 20 B 220 250 Steam reformer 150 850 Lowtemperature shift converter Burner 400 In the fuel cell most, but not all, of the hydrogen is used 250 800 C D 650 220 Steam 450 Desulfurizer Hightemperature shift converter 250 280 450 650 E 20 Natural gas F 380 220 Anode exhaust gas Figure 10.8 Diagram of a fuel‐processing system for a phosphoric acid fuel cell. The numbers indicate approximate temperatures (°C). be heated and others have to be cooled. Heating and cooling can be combined using heat-exchangers (see Section 7.2.3, Chapter 7). A simplified flow diagram of a fuel‐processing system for a PAFC powered by natural gas is shown in Figure 10.8. The PAFC requires reformed fuel gas at about 220°C, with a CO content below about 0.5 vol.%. The following is an explanation of the process: ● The natural gas enters the fuel processor at around 20°C and is heated in heat‐ exchanger E to a temperature that is suitable for desulfurization (280°C). Steam, sufficient for both the reforming and the WGS reaction, is then mixed with the desulfurized fuel. The steam–methane mixture is further heated by heat‐exchanger 297 298 Fuel Cell Systems Explained ● ● ● ● ● ● C before being fed to the steam reformer. Here, it is heated to around 850°C by the burner and is converted to a syngas product. Note that the syngas also contains some unreacted steam. The syngas then passes through the other side of heat‐exchanger C and loses heat to the incoming fuel gas. Further heat is lost to both the incoming gas at E and the anode exhaust gas at F. The gas is now sufficiently cool for the first WGS converter, where the majority of the CO is converted to CO2. At D, the gas is further cooled by giving up its heat to the incoming steam and then passes to the low‐temperature WGS converter for conversion of the remaining CO to CO2. The final cooling is accomplished at B, where the incoming steam is heated. The H2‐rich fuel gas is then sent to the PAFC stack. Here most, but not all, of the H2 is converted to electrical energy. The anode exhaust gas, still at about 220°C, is sent to heat‐exchanger F where it is preheated prior to reaching the burner. The burner is also fed with air, which will have been preheated by heat‐exchanger A through the use of energy from the burner exhaust gas. The steam arriving at about 120°C at heat‐exchanger B can be generated from water by using heat from the cooling system of the fuel cell. The burner exhaust gas, still very hot, can also be employed to raise steam to power any compressors that are required to drive the process. There are many other possible ways of configuring the gas flows and heat-exchangers to achieve the desired result, but the process flow diagrams of commercial systems are usually proprietary. A stationary fuel‐cell system is analysed further in Section 12.4.4, Chapter 12. 10.6.3 10.6.3.1 Reformer and Partial Oxidation Designs Conventional Packed‐Bed Catalytic Reactors Early PAFC plants, such as those developed by United Technologies Corporation (UTC), International Fuel Cells and Fuji Electric in the 1980s, employed fairly traditional designs of fuel processor that consisted of fixed catalytic beds for the desulfurization, steam reforming and WGS reactions. Heat from burning natural gas in a conventional burner is transferred mainly by radiation to the reformer reactor(s) that, as in a typical large‐scale refinery installation, is(are) operated at above 850 C. In 1989, researchers at WS Reformer GmbH discovered that stable combustion of natural gas with air could be obtained by increasing the throughput of the burner and recirculating the resulting product gas. The procedure became known as flameless oxidation or FLOX™ and, compared with conventional burner designs, offered the advantages of lower and more uniform temperatures throughout the combustor and a significantly lower level of nitrogen oxides (NO x) in the burner off‐gas. The FLOX™ concept was applied by WS Reformer GmbH to systems for generating H2 for the fuel‐cell buses that were undergoing trials in the HyFLEET Clean Urban Transport Europe (CUTE) programme. FLOX™ reformers have also been demonstrated with the 1‐kW high‐temperature PEMFC manufactured by Serenegy in Sweden. Fuels for Fuel Cells 10.6.3.2 Compact Reformers The desulfurizer, WGS reactors and CO clean‐up systems in all of the reformer systems may be packed‐bed catalytic units of traditional design. In many cases, the pellets or extrudates that are the common forms of catalyst used in the petrochemical industry have been replaced by coated ceramic monoliths for fuel cells. With respect to the design of the reformer reactor for fuel‐cell systems, several novel features are being pursued, particularly in integrating some of the heat‐transfer duties. A compact reformer produced by Haldor Topsøe for PAFC systems is shown in Figure 10.9. In this particular design, heat for the reforming reaction is provided by combustion of the lean anode exhaust gas, which may be supplemented with fresh fuel gas. Fuel is combusted at a pressure of some 450 kPa in a central burner that is located in the bottom of a pressure vessel. Feed gas is passed downwards through the first catalyst bed where it is heated to around 675°C by convection from a combined countercurrent of the combustion products and the reformed product gas. On leaving the first bed of catalyst, the partially reformed gas is transferred through a set of tubes to the top of the second reforming stage. The gas flows down through the catalyst and is heated typically to 830°C by convection from the co‐currently flowing combustion products and also by radiation from the combustion tube. The combination of co‐current and countercurrent heat transfer helps moderate the temperature of the reactor, an important consideration in high‐temperature reformer design. The advantages of such a reformer for fuel‐cell applications are (i) small size and suitability for small‐scale use, (ii) pressurized combustion of lean anode exhaust gas gives good process integration with the fuel cell, Reformed fuel out Fuel to be reformed Flue gas out First catalyst bed Second catalyst bed Anode off-gas containing unreacted fuel Burner Air Figure 10.9 Haldor Topsøe heat‐exchange reformer. 299 300 Fuel Cell Systems Explained (iii) improved load following and (iv) lower cost. In addition to Haldor Topsøe, several companies have been developing reformers of this type, namely, International Fuel Cells (now Doosan Fuel Cell), Ballard Power Systems, Sanyo Electric, Osaka Gas and ChevronTexaco. 10.6.3.3 Plate Reformers and Microchannel Reformers A plate reformer is built up of a number of box-like reactors stacked one on top of the other. Thin metal plates separate each reactor compartment. The compartments are alternately filled with suitable catalysts to promote the combustion and steam reforming reactions. In another approach, each separating plate is coated with a steam reforming catalyst on one side and a combustion catalyst on the other. The heat from the combustion reaction is used to drive the reforming reaction. Plate reformers have the dual advantage of being very compact and offering a means of maximizing heat transfer. The use of a combustion catalyst enables gases with low heating values (e.g., anode exhaust gases) to be burnt without the need for a supplementary fuel. Plate reformers were first developed by Ishikawajima‐Harima Heavy Industries Co., Ltd (IHI) in the early 1980s; the catalyst was in the form of spherical pellets located on either side of the heat‐ exchanger surface.20 Gastec, Plug Power, Osaka Gas and several other companies have since built plate reformers for fuel‐cell systems. The most advanced types of plate reformer employ compact heat‐exchanger hardware on which the catalyst is coated directly in the form of a thin film with a thickness of a few microns.21 The concept is shown in Figure 10.10. Such devices were developed in the United States by researchers at Pacific Northwest National Laboratory who have demonstrated a 1‐kW steam reformer and at Argonne National Laboratory who are developing a reformer with a monolithic catalyst for the processing of diesel. Plate reformers for methanol have been constructed by several organizations that include CH4 + O2 CO2 +H4O Active sites Porous catalyst support 10 – 50 μm 1 mm Heat 10 – 50 μm Stainless-steel substrate Porous catalyst support CH4 + H2O H2 + COx Figure 10.10 Plate or microchannel reformer concept. Catalyst is coated as thin film onto one or both sides of a heat‐exchange material. 20 Hamada, K, Mizusawa, M and Koga K, 1997, Plate reformer, US Patent No. 5,609,834. 21 Goulding, PS, Judd, RW and Dicks, AL, 2001, Compact reactor, Patent No. WO/2001/010773. Fuels for Fuel Cells (a) (b) (c) Figure 10.11 Experimental compact reformer reactors. (a) and (b) Pacific Northwest National Laboratory liquid fuel reformer (courtesy of Pacific Northwest National Laboratory). (c) Proof‐of‐concept Advantica natural gas reformer, a diffusion‐bonded multichannel reactor block (3 × 3 × 10 cm). (Source: Courtesy of Advantica Technologies Ltd.) IdaTech, Mitsubishi Electric, InnovaTek Inc., NTT Telecommunications Laboratory and Honeywell. Examples of the hardware are shown in Figure 10.11. A microchannel reactor (MCR) is another term for the compact reactor technology that could be applied to other units of a fuel processor such as the fuel vapourizers and the gas clean‐up reactors. Unfortunately, however, MCR systems suffer from two notable drawbacks, namely, (i) plugging of the channels due to catalyst degradation and carbon deposition, and (ii) the catalyst is incorporated into the reactor for life and therefore cannot easily be replaced when it becomes degraded. 10.6.3.4 Membrane Reactors An example of a membrane reactor developed by a commercial company is the ion transport membrane (ITM) technology developed by Air Products that combines air separation and high‐temperature generation of synthesis gas (via autothermal reforming) in a single ceramic membrane reactor. The ITM Syngas process employs a planar membrane of MIEC oxides. In operation, oxygen from a hot air stream is reduced at one surface of the ITM membrane to oxygen ions, which diffuse through the membrane under a chemical 301 302 Fuel Cell Systems Explained Scanning electron microscope cross-section H2 separation membrane Diffusion barrier (YSZ) H2 separation membrane (Pd/Al2O3) H2 Catalyst support Ni/ YSZ H2 H2 Reactor body Catalytic support 20 μm H2 H2 H2 H2O + CH4 CO + CO2 + H2O H2 H2 H2 H2 H2 H2 Figure 10.12 Concept of membrane reformer system under development by Tokyo Gas. potential gradient. At the opposite surface of the membrane, the oxygen partially oxidizes a preformed mixture of hot natural gas and steam to form syngas. Praxair is exploring a similar strategy that differs mainly in that a tubular membrane is used. Although such technologies have been targeted for GTL operations, there is no reason in principle why they should not be applied to fuel‐cell systems. In 2009, Tokyo Gas demonstrated a membrane reformer system that was capable of producing H2 from natural gas at a rate of 40 Nm3 h−1. The reactor system, including insulation, measured 1200 × 50 × 1350 mm and was comprised of 112 tubular membrane modules, each made of porous stainless steel onto which was coated a 20‐µm layer of palladium. The H2 production efficiency was claimed to be over 81%, which is significantly higher than that obtained with a conventional reformer reactor. Current research by Tokyo Gas is focused on using a thin‐film palladium–silver alloy that is deposited on yttria‐stabilized zirconia that acts as both a diffusion barrier for carbon oxides and a support to the reformer catalyst. The work programme to 2020 is aimed mainly at H2 production systems for FCVs and is based on the use of a tubular reactor, as illustrated schematically in Figure 10.12. 10.6.3.5 Non‐Catalytic Partial Oxidation Reactors Non‐catalytic partial oxidation (NCPO) is applied industrially by Texaco and Shell for the conversion of heavy oils to syngas. In the Shell process, liquid fuel is fed to a reactor together with oxygen and steam. A partial combustion takes place in the reactor and yields a product at around 1150°C. It is this high temperature that poses a particular problem for conventional POX. The reactor has to be made of expensive materials, and the product gas needs to be cooled to allow unreacted carbon material to be separated from the gas stream. The high temperature also means that expensive materials of construction are required for the heat-exchangers. In addition, the effluent from the reactors invariably contains contaminants (e.g., sulfur compounds) as well as carbon Fuels for Fuel Cells and ash, all of which require proper disposal. Therefore, given its extra cost and complex operation, NCPO has not been a preferred option for fuel‐cell applications. 10.6.3.6 Catalytic Partial Oxidation Reactors Catalytic partial oxidation reactors can be very simple in design given the requirement for only one bed of catalyst into which the fuel and oxidant (usually air) are injected. Often steam is added as well, in which case some conventional reforming also occurs. As mentioned in Section 10.4.7, the combination of CPO and steam reforming is usually referred to as ‘autothermal reforming’ because there is no net heat supplied to, or extracted from, the reactor. All of the heat for reforming is provided by partial combustion of the fuel. Depending on the nature of the fuel and the application, two types of catalyst are sometimes used — one primarily for the CPO reaction and the other to promote steam reforming. The HotSpot™ reactor, developed during the late 1990s, is an example of a CPO reactor. The technology was promoted by Johnson Matthey as a means of generating H2 from natural gas for small‐scale stationary systems, as well as for the reforming of liquid fuels on vehicles. The reactor employed a platinum–chromium oxide catalyst on a ceramic support. Three reactors are shown in Figure 10.13. As its name implies, the novel feature of the reactor was the hotspot caused by point injection of the air– hydrocarbon mixture through a narrow tube inserted into the centre of the catalyst bed. The arrangement eliminated the need for preheating the fuel gas and air during operation, although, for start‐up on natural gas, the fuel had to be preheated to around 500°C. Alternatively, the reactor could be started from ambient temperature by introducing an initiating fuel such as methanol or an H2‐rich gas. These fuels are oxidized by air at ambient temperature over the catalysts and thereby serve to raise the bed to the temperature required for natural gas to react (typically over 450°C). Figure 10.13 Johnson Matthey HotSpot™ reactor. These were made in different forms for methanol, methane or gasoline processing. (Source: By courtesy of Johnson Matthey plc.) 303 304 Fuel Cell Systems Explained 10.7 Practical Fuel Processing: Mobile Applications The motivation for the reforming of fuel on vehicles was probably the world oil crisis of 1974 that stimulated support for the advancement of all types of fuel cell, as well as interest in the possibility of a ‘hydrogen economy’. The rapid development of the PEMFC during the 1990s and its uptake by leading vehicle manufacturers led to some impressive research and development programmes being mounted by DaimlerChrysler, General Motors, Ford and others in collaboration with manufacturers of PEMFC stacks. Methanol was proposed as a suitable energy carrier because, in addition to being a liquid and therefore readily transportable, it reacts with steam over a catalyst at relatively low temperatures. On‐board reforming of methanol was therefore pursued during the 1990s by several groups and organizations that included Johnson Matthey, DaimlerChrysler, General Motors, Ballard Power System, Nissan and Toyota. Many demonstration vehicles were built, and towards the end of the decade, Arthur D. Little, ExxonMobil, Nuvera and Shell had also investigated catalysts and processes for on‐board reforming of gasoline. As the 21st century arrived, many questioned the wisdom of on‐board fuel reforming. This was due not only to the technical difficulties associated with the time and energy required at start‐up and the poor transient response but also to the high costs involved. In addition, various ‘well‐to‐wheel’ studies carried out in Europe and North America showed that on‐board gasoline reforming for an FCV did not achieve a particularly high overall energy‐conversion efficiency in comparison with an internal combustion engine vehicle (ICEV) or a hybrid electric vehicle (HEV). In 2004, the US DOE brought the various stakeholders together for a go/no‐go decision for on‐board reforming.22 The DOE committee unanimously decided to abandon all support for on‐board reforming projects, an outcome that was generally accepted worldwide. The only notable exceptions to the DOE decision are the on‐board fuel processors that form a component of auxiliary power units (APUs) for heavy‐duty trucks, military vehicles and recreational vehicles such as campervans. These all have significant power requirements even if the vehicle is not moving, and they make a sizeable contribution to global emissions and wasted energy. For example, idling diesel vehicles (including trailers and buses) are estimated to burn a billion gallons of diesel fuel every year in the United States alone. Diesel fuel or heavier logistic fuel can be converted to a syngas in a CPO reactor, and, apart from sulfur removal, this facility alone is sufficient to fuel an on‐board system for generating power with an SOFC. Any other fuel cell would be too problematic in terms of fuel processing for this application. A combination of CPO and SOFC is therefore an attractive proposition for producing the auxiliary power for heavy‐duty trucks and similar sizes of vehicle. Especially when combined with some battery storage, a CPO–SOFC need not exhibit fast dynamic response as the auxiliary systems demand a fairly constant power load. Delphi Corporation has been one of the leading proponents of on‐board SOFC–APU technology with an integrated fuel processor. The company worked in partnership with 22 DOE Team Decision Report, August 2004, On‐board fuel processing go/no‐go decision. Available online: http://www1.eere.energy.gov/hydrogenandfuelcells/pdfs/committee_report.pdf (accessed on 27 September 2017). Fuels for Fuel Cells BMW, Los Alamos National Laboratory, Battelle and Global Thermoelectric to develop a diesel‐fuelled APU in the 1990s. In 2001, a gasoline‐fuelled APU was demonstrated by BMW in a 7‐series sedan. In addition to APU systems for vehicles, the processing of gasoline and diesel has been suggested as a means of providing power for the hotel load on ships, i.e., the power that continues to be required by marine craft when they are anchored in harbour. Consequently, a large body of literature exists on the steam reforming and CPO of these fuels.23 The catalysts are much more demanding than those for processing simpler fuels such as natural gas or methanol. Higher temperatures are necessary, and a greater resistance to fouling by carbon or sulfur generally rules out simple supported nickel catalysts. In general, the steam reforming of fuels such as kerosenes involves the removal of sulfur compounds and the use of high steam‐to‐carbon ratios. In the case of CPO, the absence of steam gives rise to lower concentrations of H2 in the reformate and a higher risk of carbon formation. The influence of both the oxygen‐to‐carbon ratio and the highly exothermic CPO reaction has led to the development of catalysts that are more tolerant of carbon formation. For these catalysts, the preferred active metals are rhodium and ruthenium, supported on ceria. Compared with the more conventional supported nickel catalysts, the platinum‐group metal is less active for the reactions that form carbon, and the ceria provides an oxygen‐rich surface functionality for the cracking of highly aromatic hydrocarbons. 10.8 Electrolysers 10.8.1 Operation of Electrolysers Electrolysers use electricity to split water into hydrogen and oxygen. They are thus the opposite of a fuel cell. The basic theory and the reactions taking place at the electrodes are the same for electrolysers as for fuel cells — except that the reactions are reversed. Different electrolytes can be used, just as for fuel cells, and in order to minimize electricity consumption, it is important to choose an electrolyte of maximum conductivity. Electrolyser technology was developed in the 1800s, and by the beginning of the 20th century, there were more than 400 industrial water electrolysis units in operation. In 1939, the first large water electrolysis plant with an H2 output of up to 10 000 Nm3 h−1, built by the Norwegian company Norsk Hydro Electrolyzers, went into operation. Most industrial electrolysers employ an alkaline electrolyte, and, as with the AFC, this is usually an aqueous solution of potassium hydroxide (30–40 wt.%). The solution must be prepared from very pure water; otherwise impurities will accumulate during electrolysis. The chloride ion, which is usually present in water, is particularly harmful in that it causes pitting of the protective films that form on metal surfaces in alkaline solutions. Industrial alkaline electrolysers have been employed to generate hydrogen for diverse applications that range from the hydrogenation of fats in the food industry to the 23 Schwank, JW and Tadd, AR, 2010, Catalytic reforming of liquid hydrocarbons for on‐board solid oxide fuel cell auxiliary power, Catalysis, vol. 22, pp. 56–93. 305 306 Fuel Cell Systems Explained cooling of large gas turbines engaged in central electricity generation.24 In principle, electrolysers are well suited for use with electricity generated from renewable energy sources such as wind, solar and hydropower. In practice, however, alkaline electrolysers are designed to operate with a fairly constant power supply, and therefore the intermittent nature of renewable sources may necessitate the development of specialized power control and conditioning equipment. Commercial electrolysers are also not particularly efficient. A large industrial plant capable of producing 500 Nm3 h−1 of H2 would demand about 2.3 MW of power. With a capital cost of US$600–700 per kW, this makes the generation of H2 by electrolysis uneconomic unless low‐cost electricity is available. Using renewable electricity, it is estimated that H2 can be produced at around US$7–10 per kg, i.e., three to five times above that from fossil fuels.25 Traditional alkaline electrolysers may be constructed with either monopolar or bipolar configurations. The monopolar (or ‘tank‐type’) unit consists of alternating positive and negative electrodes that are held apart by microporous separators. The positives are all connected together in parallel, as are the negative electrodes, and the whole assembly is immersed in a single electrolyte bath or tank to form a unit cell. By contrast, a bipolar unit uses a metal bipolar plate to join adjacent cells, as in a PEMFC stack. In a bipolar alkaline electrolyser, the electrocatalyst for the negative electrode is coated on one face of the bipolar plate, and that for the positive electrode of the adjacent cell is on the reverse face. A series‐connected stack of bipolar cells forms a module that operates at a higher voltage and lower current density than the monopolar design. In an alkaline electrolyser, the reactions occur through the ionization of water into protons and hydroxyl ions, i.e., H2O H( aq ) OH( aq ) (10.14) The OH( aq ) ions migrate to the positive electrode to produce electrons and release oxygen: 4OH( aq ) 2H 2 O O 2 4 e (10.15) Concomitantly, hydrogen is produced at the negative electrode: 2H( aq ) 2e H2 (10.16) The reactions are, of course, the opposite of those occurring in the AFC. Despite widespread use of the alkaline electrolyser, much interest has been shown in two other technologies that offer certain advantages, namely, the proton‐exchange membrane (PEM) electrolyser and the high‐temperature steam electrolyser. 24 In a synchronous generator or alternator, hydrogen gas is circulated by blowers and fans through the rotor and stator and then passed over cooling coils inside the generator casing. The coils carry oil or water to extract heat from the circulating hydrogen. Hydrogen has the advantage of low density (7% that of air) and high thermal conductivity (6.7 times that of air). It offers several advantages over air for cooling alternators or generators, for example, it allows a machine of the same dimension to have 20–25% greater output capacity, and hydrogen cooled alternators required 20% less active material (steel and copper) than air-cooled machines. 25 The Hydrogen Economy: Opportunities, Barriers and R&D Needs, 2004, National Academic Press, Washington, DC. ISBN: 978‐0‐309‐09163‐3. Fuels for Fuel Cells The first PEM electrolyser was produced by General Electric in 1966. The basic structure is the same as the PEMFC, although the electrodes have different requirements. The reactions for water electrolysis in an acid electrolyte, such as a PEM, are as follows: At the negative electrode: 4H 4e 2H2 (10.17) At the positive electrode: 2H2O O2 4H+ 2e (10.18) These reactions are the opposite of those shown in Figure 1.3. One reason for the success of PEM electrolysers is that many of the problems associated with PEMFCs do not apply. Cooling is fairly trivial, as the water supplied to the cathode can be pumped around the cell to remove heat. Water management, another key problem of the PEMFC, is also massively simplified, as the positive electrode must be flooded with water. Electrolysers using PEM membranes also offer advantages over their alkaline counterparts in terms of: ● ● ● ● ● Wide operating range capable of meeting large variations in demand. Greater safety through the absence of alkaline solutions. More compact design due to higher current densities. Ability to operate at higher pressures. Minimal maintenance required. The H2 produced by such electrolysers will have a high purity but may have a high humidity due to protons dragging water molecules through the electrolyte. Indeed, the water content of H2 can be so high that condensation occurs and can prove to be a serious issue if it is intended to store the H2 at pressure or as a solid‐state hydride. Since the energy to compress water is less than that required to compress H2, the high‐pressure PEM electrolyser (HPE) has been the subject of increasing development in recent years. Hydrogen pressures of 12–20 MPa can be achieved and thus eliminate the need for a gas compressor that can be both expensive to maintain and inefficient. Stacks employed in HPEs are shown in Figure 10.14. Theory predicts that elevating the temperature of electrolysis would improve efficiency, since some of the energy used to split the water is provided as heat, and would also reduce the overpotentials at the electrodes. Increasing the temperature substantially to 700–1000°C is possible if a sold ceramic electrolyte is used, e.g., stabilized zirconia, as in the SOFC. The concept of such a high‐temperature ‘steam electrolyser’ has been investigated for many years in parallel with the development of the SOFC and other high‐temperature electrochemical reactors.26 10.8.2 Applications At first, it would seem to be perverse to use electricity to make H2 for use in fuel cells to turn it back into electricity. In each conversion step there are losses in efficiency, such that the ‘round‐trip’ efficiency of converting electricity to H2 and then back again can 26 Zahid, M, Schefold, J and Brisse, A, 2010, High‐temperature water electrolysis using planar solid oxide fuel cell technology: a review, in Stolten, D (ed.), Hydrogen and Fuel Cells, pp. 227–231, Wiley‐VCH, Weinheim. 307 308 Fuel Cell Systems Explained Figure 10.14 ITM Power HGas electrolyser stacks, each operating at 8 MPa pressure. (Source: Reproduced with permission of ITM Power.) be as low as about 40%. If the objective, however, is to avoid the wastage of surplus electricity (v.i.) through using it to make hydrogen that can be stored to meet a future energy demand, then it may be possible to make an economic case for electrolysis. In general, the H2 cost is a function of both the cost of electricity and the capital cost of the associated equipment, i.e., the electrolyser and the compressor. At the small scale, solar or wind power can be used to generate H2 for use in FCVs. In this case, the cost of electricity is governed principally by the capital cost of the photovoltaic (PV) panels and associated control and inverter equipment. Several manufacturers are providing systems for generating H2 for FCV filling stations and include ITM Power, Hydrogenics and Proton OnSite. Two examples are shown in Figure 10.15. On a larger scale, electrolysis can be used to generate H2 from the excess or off‐peak power generated by wind or solar farms that can then be stored for future use. The concept has become known as power‐to‐gas (often abbreviated to P2G). The H2 produced in P2G systems may be converted further to substitute natural gas (SNG) via methanation of CO2 or to syngas and liquid fuels by the Fischer–Tropsch (GTL) process (v.s.). On the other hand, the H2 may simply be compressed and blended with natural gas in the gas transmission or distribution network.27 Natural gas pipelines can provide a strategic means of bulk energy storage and can accept low concentrations of hydrogen. In Germany, for example, the capacity of the natural gas network is more than 200 000 GWh, which is sufficient to satisfy the national energy demand for several months. By comparison, the 27 Whereas substitute natural gas can be blended directly into a gas transmission network, there is a limiting concentration to which hydrogen can be blended, typically a few per cent by volume. The allowable concentration of hydrogen varies according to the country and point of injection. (a) (b) Figure 10.15 Hydrogen filling stations using (a) ITM Power PEM electrolysers coupled with wind turbine at the Advanced Manufacturing Park, South Yorkshire, United Kingdom. (b) Hydrogenics HySTAT alkaline electrolysers 130 kg day−1 dispensing 70 MPa H2, Stuttgart, Germany. (Source: Reproduced with permission of ITM Power.) 310 Fuel Cell Systems Explained capacity of all German pumped‐storage utilities amounts to only about 40 GWh. Note that natural gas pipelines cannot be used safely to distribute pure hydrogen without being modified to ensure against progressive degradation of the materials employed in construction of these networks. High carbon steels employed in transmission systems can be prone to hydrogen embrittlement and decarburization when subject to pressure cycling in pure hydrogen, and polyethylene used in low‐pressure distribution systems is not compatible with hydrogen. For P2G systems, the PEM electrolyser is favoured because of its ability to operate at sufficiently high pressure so that the hydrogen produced can be blended directly into the natural gas network. A further advantage is the faster response of the PEM electrolyser compared with an alkaline system. As more renewable energy sources are fed into a traditional electricity supply network, there will come a point where the intermittent renewable power is out of step with the demand from consumers. Traditional turbogenerators are unable to react sufficiently rapidly to balance the load on the network with the consequence that voltages on the network may rise or fall and the AC frequency may shift outside the prescribed limits, with potentially catastrophic results. It is generally recognized that some form of grid storage will therefore be required as renewable power becomes more widespread. Unfortunately, however, most storage options, such as rechargeable batteries, are very expensive per kWh or, in the case of pumped hydro, respond too slowly to changes in demand. To address the load‐balancing issue for power networks, ITM Power has produced the HGas rapid‐response PEM electrolyser shown in Figure 10.16. The HGas system is modular and can be supplied in a range of sizes upwards from 70 kW (producing 20 kg H2 day−1). The HGas system can be turned on within a few seconds to deliver H2 at 8 MPa that can be either stored or fed directly into the natural gas network without further compression. Over the past 20 or so years, many projects have been initiated to demonstrate the feasibility of using PV or wind to supply an electrolyser to generate H2 that can then be stored, either as compressed gas or in the form of a hydride. A fuel‐cell system — typically an AFC, PEMFC or PAFC — is coupled with the hydride. In most cases, lead– acid batteries have also been incorporated into the systems to assist with load levelling. Example projects were the HARI demonstration undertaken by Beacon Energy and Loughborough University in the United Kingdom and the RES2H2 project conducted in Greece. Some common issues with these systems were identified: ● ● ● ● ● ● Hydrogen from an alkaline electrolyser must be purified to be acceptable for hydride storage (see Section 11.5, Chapter 11). Steps need to be taken to minimize contamination of the H2 when the equipment is in standby mode. The electrolyser can cope with intermittent power from PV or wind turbines, provided a battery is used as a storage buffer. Careful process integration is required to minimize heat loss and ensure high overall efficiency. The requirement for water purification could be largely reduced, or even eliminated, by recycling product water from a PEMFC. Problems are more likely to arise with the auxiliary mechanical equipment (e.g., water demineralizer, air compressor, inert gas supply) than with the electrolyser and storage components. Fuels for Fuel Cells (a) (b) Figure 10.16 ITM Power HGas rapid‐response electrolyser: (a) installed in a P2G system in Frankfurt, Germany and (b) stack of PEM cells. (Source: Reproduced with permission of ITM Power.) 311 312 Fuel Cell Systems Explained 10.8.3 Electrolyser Efficiency The thermodynamically ‘reversible’ voltage Vr for water electrolysis under isothermal conditions and at standard temperature (298.15 K) and pressure (101.325 kPa) is 1.229 V. This value decreases almost linearly with increasing temperature to 1.0 V at 573 K (300°C). The concomitant decrease in free energy, ΔG, is largely offset by an increase in the entropy term, TΔS, so that the enthalpy of reaction, ΔH, is almost independent of temperature. The efficiency of an electrolyser is calculated in almost the same way as that for a fuel cell. If Vc is the operating voltage for one cell of a fuel‐cell stack, then it was shown in Section 2.4, Chapter 2, that the efficiency (on a higher heating value (HHV) basis) is given by: Vc 1.48 (10.19) In the case of an electrolyser, the formula is simply the inverse of this expression, namely: 1.48 Vc (10.20) Apart from there being no issue of fuel crossover, the voltage losses in electrolysers follow exactly the same pattern as those for fuel cells that are described in Section 3.3, Chapter 3, i.e., activation losses (overpotentials) at the positive and negative electrodes and the total resistive (‘ohmic’) losses within the electrodes and electrolyte. An alkaline electrolyser when operating at 90°C and atmospheric pressure with non‐precious metal electrodes typically requires 2.1 V overall to yield a current density of 200 mA cm2. Because the problems of cooling and water management are so much more easily solved in the PEM electrolyser, the performance of this technology is routinely higher than that for an alkaline electrolyser; in fact, it can match that of the best PEMFCs, with current densities of around 1.0 A cm−2 usually being achieved. To keep capital costs low, it is necessary to operate electrolysers at as high a current density as possible, but, as with PEMFCs, such practice has to be traded against lower cell efficiency. Industrial alkaline electrolysers are generally 60–75% efficient, whereas best practice with small‐scale systems is claimed to be closer to 80–85%. The German HOT ELLY high‐temperature electrolyser, which uses a zirconia electrolyte, reaches an efficiency of over 90%. Despite this high efficiency, high‐ temperature electrolysers still produce H2 at about four times the cost of that obtained from the steam reforming of natural gas. 10.8.4 Photoelectrochemical Cells Sunlight may be harvested by PV cells to generate DC power for the electrolysis of water. Photovoltaic solar cells, which are becoming widely used in their own right for power generation, are formed from a thin layer of a semiconductor material, such as silicon. The material is doped in such a way that one side is negatively charged (n‐type) and the other positively charged (p‐type). When light strikes the n‐type semiconductor, loosely held electrons are released, and, if a current-collector is attached, these electrons can be sent via an external circuit to the p‐type side where they will be accepted by Fuels for Fuel Cells vacancies or ‘holes’ in the material. The resulting flow of current can be utilized, but the voltage created by a single p–n junction is low, and because not all of the energy in the light is captured (i.e., only a part of the spectrum has sufficient energy to release electrons from the ‘conduction band’ to the ‘valence band’), the efficiency of PV cells tends to be low. Other semiconductors that have been employed in PV cells include gallium arsenide (GaAs), cadmium telluride (CdTe) and copper indium gallium diselenide (Cu(In,Ga)Se2). It is possible to connect PV cells in series to generate a sufficiently high voltage to electrolyse water. There is an attraction, however, in developing a system that harvests solar energy in a single cell to split the water molecules directly. The process is called ‘electrochemical photolysis’, or simply ‘photolysis’, and is effectively carried out in the leaves of every living plant — the first step in the process known as ‘photosynthesis’. The voltage required for water electrolysis of 1.293 V under standard conditions rules out most metallic oxides and sulfides. A search has been ongoing for many years to identify semiconductor materials that are able to exhibit at least the 1.6–1.7 V that is required for water splitting under normal operating conditions. In 1972, Fujishima and Honda were the first to demonstrate that hydrogen and oxygen could be produced directly in a photoelectrochemical cell (PEC).28 Their cell used a single‐crystal titanium dioxide (TiO2) electrode that was connected via a wire to a counter electrode of platinum where hydrogen was released. Unfortunately, titanium dioxide only absorbs light in the UV region, i.e., below a wavelength of about 385 nm. To encourage absorption in the visible region, the dye‐ sensitized solar cell (DSSC) was developed. The fundamental operation involves the absorption of a dye, usually ruthenium based, on the surface of the porous titania electrode. When light strikes the dye, electrons are released into the conduction band of the titania at the same time as the dye is oxidized. The electrons can then jump into the valence band of the titania, which is intimately mixed with a current-collector that is usually tin oxide coated on glass. The electrons flow around an external circuit, and the dye is reduced via a ‘redox mediator’, which is typically the iodide–tri‐iodide couple (I––I3–) dissolved in acetonitrile or some other organic solvent. The mediator diffuses to the negatively charged electrode where it is reduced by electrons that have travelled round the external circuit. The DSSC can be used together with a conventional PV cell to produce a ‘tandem’ cell in which a standard PEC is placed in front of a DSSC (Figure 10.17). The photoelectrode in the PEC absorbs the high‐energy UV and blue light in sunlight to liberate oxygen, while radiation of longer wavelengths in the green‐to‐red region of the spectrum passes through and is absorbed by the DSSC. This boosts the flow of electrons that are fed back to a counter electrode in the PEC to produce hydrogen. With such an arrangement, photon efficiencies as high as 12% have been reported. Since the discovery of the photochemical activity of titania, several other materials have been found to exhibit the required activity, although much work remains to produce candidates that have the required resistance to long‐term degradation. A gadolinium semiconductor (Ga0.82Zn0.18)(N0.82O0.18) loaded with rhodium and chromium photocatalyst was reported in 2008 to achieve a quantum yield in visible light of 5.9%. 28 Fujishima, AK and Honda, K, 1972, Electrochemical photolysis of water at a semiconductor electrode, Nature (London) vol. 238, pp. 37–38. 313 314 Fuel Cell Systems Explained e– Dye on semiconductor Platinum electrode Conducting glass Gas outlet H2 O2 H2O Violet Conducting glass Green Red Aqueous electrolyte Optical window Semiconductor Conducting glass Electrolyte (mediator/organic solvent) e– Photoelectrochemical cell Dye-sensitized solar cell Figure 10.17 Operating principles of a tandem cell for enhanced hydrogen production. Nevertheless, such materials display little absorption activity beyond about 440 nm and thereby have an overall solar‐to‐hydrogen efficiency of only about 0.1%. In 2015, a new low‐cost photoelectrochemical catalyst based on cobalt oxide was shown to have a superior efficiency of around 5% and therefore may pave the way for other affordable semiconductor nanomaterials.29 10.9 Thermochemical Hydrogen Production and Chemical Looping 10.9.1 Thermochemical Cycles Because of the stability of the water molecule and consequently the very high temperatures that are required to split it thermally, attempts have been made to accomplish the process at a more moderate upper temperature (i.e., <1000°C) by means of an indirect route. The general idea is to decompose water by reacting it with one or more chemicals that are regenerated via a series of cyclic thermochemical 29 Liao, L, Zhang, Q, Su, Z, Zhao, Z, Wang, Y, Li, T, Lu, X, Wei, D, Feng, G, Yu, Q, Cai, X, Zhao, J, Ren, Z, Fang, H, Robles‐Hermandex, F, Baldelli, S and Bao, J, 2014, Efficient solar water‐splitting using a nanocrystalline CoO photocatalyst, Nature Nanotechnology, vol. 9, pp. 60–73. Fuels for Fuel Cells reactions. In this way, the hydrogen and oxygen evolution reactions are separated. Clearly, on practical and efficiency grounds, the fewer reactions involved the better. Although the idea is sound from a thermodynamic point of view, there are both engineering and materials issues. The sulfur–iodine cycle has been studied extensively by the nuclear industry since the mid‐1970s. In this cycle, iodine is employed to promote the oxidation of S(IV) to S(VI) as follows: I2 SO2 2H2O 2HI + H2SO 4 (10.21) H2SO 4 → SO2 + H2O + ½O2 Endothermic : 850 − 900°C (10.22) 2HI (10.23) I2 H2 Endothermic : 300 450 C Two immiscible phases are first formed. The upper phase contains almost all of the sulfuric acid, and the lower and dense phase holds most of the hydrogen iodide and iodine. These are separated and the upper phase is then decomposed via reaction (10.22), while the hydrogen iodide in the lower phase is converted into hydrogen and iodine according to reaction (10.23). The products of the two reactions (SO2 and I2, respectively) are then recycled to reaction (10.21). In an attempt to reduce the number of steps in the thermochemical cycle from three to two and at the same time to avoid the use of highly corrosive reagents such as sulfuric acid, attention is on simpler cycles in which water is reduced to hydrogen by metals (M) or metal oxides (MOred) in a lower oxidation state, i.e., M/MO red H2O MO ox +H2 (10.24) In the second step of the cycle, the product oxide in its higher oxidation state is thermally decomposed to liberate oxygen and revert to its original form, namely: MO ox → M/MO red + ½O2 (10.25) The classic example is that based on iron oxides as follows: 3FeO H2O Fe3O 4 H2 Exothermic Fe3O 4 (l ) → 3FeO(l ) + ½O2 Endothermic : > 1600°C (10.26) (10.27) Note that the oxides are liquid at 1600°C. To avoid the high temperatures required by reaction (10.27), the manganite (Fe3O4) can be replaced by the mixed oxide (Ni0.5Mn0.5) Fe2O4, which is partially reduced to an oxygen‐deficient state. The first step in the cycle can be carried out at about 800°C to regenerate the oxide and liberate hydrogen. Even though the temperatures can be lowered, there are challenging issues in engineering a reactor that can be cycled from oxidizing to reducing environments, as well as in identifying material that does not degrade significantly over time with repeated oxidation and reduction. The many thermochemical cycles that have been investigated generally fall into four groups, as shown in Table 10.7. Recent work is focused on low‐temperature reactions, such as the copper chloride system. 315 316 Fuel Cell Systems Explained Table 10.7 Thermochemical cycles currently under consideration. Number of reaction steps Maximum temperature (°C) LHV efficiency (%) 2 900 43 Sulfur cycles Hybrid sulfur (Westinghouse ISPRA, Mark 11) (1150 without catalyst) Sulfur–iodine 3 900 (General Atomics, ISPRA Mark 16) 38 (1150 without catalyst) Volatile metal/oxide cycles Zinc/zinc oxide 2 1800 45 Hybrid cadmium 2 1600 42 Non‐volatile metal oxide cycles Iron oxide 2 2200 42 Cerium oxide 2 2000 68 Ferrites 2 1100–1800 43 4 530 39 Low‐temperature cycles Hybrid copper chloride A variation of the thermochemical cycle is to use a hydrocarbon gas as the means of reducing the oxide back to the metal. An example would be replacing reaction (10.24) with CH 4 MO x M/MO x CO2 2H2O (10.28) and specifically in the case of iron/iron oxide, this becomes CH 4 Fe3O 4 3Fe CO2 2H2O (10.29) If the reduced iron is heated, it can be used to reduce steam to pure hydrogen, i.e., 3Fe 4H2O 4H2 Fe3O 4 (10.30) Equations (10.29) and (10.30) therefore provide the basis for an excellent means of generating pure hydrogen from a hydrocarbon such as methane or natural gas. Compared with the processes examined in this chapter for dealing with hydrocarbon fuels, the use of a hydrocarbon gas offers the following three advantages: ● ● ● Saving of investment costs associated with CO removal and other purification units by cyclic operation of a single reactor. Saving of investment and operational costs by using iron oxide as a cheap material. High quality of the produced hydrogen. One of the problems with the simple iron/iron‐oxide process is that the reduction reaction (10.29) is very slow. Recent research has focused on improving the rate of this Fuels for Fuel Cells reaction, while maintaining moderate temperatures, by including other materials or promoters either in combination with the iron or as a separate layer of catalyst in the reactor. A mixture of ceria and zirconia has been shown to promote the rapid POX of methane to CO and H2, which will then reduce the iron material according to: 4CO Fe3O 4 4CO2 3Fe 4H2 Fe3O 4 4H2O 3Fe (10.31) (10.32) Even with such enhancements, there remains concern over the long‐term performance and mechanical integrity of the iron/iron oxide. 10.9.2 Chemical Looping Chemical looping combustion (CLC) is a process that is analogous to thermochemical hydrogen production. Instead of performing hydrocarbon combustion in a single reaction stage, two (or conceivably more) reactions are used. An additional species is required and circulates between the two reactions. This additional species is typically a metal and serves to carry oxygen between the reactions, that is, essentially the same function as the iron/iron oxide described earlier. As an example, consider the following two reactions, representing the CLC of methane, using a nickel-based reaction scheme: 4 Ni 2O2 CH 4 4 NiO 4 NiO (10.33) CO2 2H2O 4 Ni (10.34) If reactions (10.32) and (10.33) are added together, the nickel simply circulates between the two reactions; hence, from the perspective of an overall mass and energy balance, the two reactions simplify to the basic methane oxidation reaction, i.e., CH 4 2O2 CO2 2H 2 O (10.35) If both reactions (10.33) and (10.34) are arranged to take place in separate vessels, the oxygen does not come into contact with the fuel. This gives an immediate advantage when the source of oxygen is air, as will usually be the case. The result is that the product gas is not diluted with nitrogen, i.e., the process incorporates a method of separating CO2 from the combustion products (water is relatively easy to remove from the products of reaction (10.34)). For this reason CLC has been investigated in detail for use in industrial combustion processes as a means of separating CO2 prior to sequestration. Chemical looping combustion has also been proposed in conjunction with gasification of hydrocarbons, including coal, for the production of hydrogen. Such a process may include the following steps: 1) Generation of hydrogen from steam using suitable oxygen carriers. 2) Fuel gasification in the presence of an H2–steam mixture. 3) Combustion of the fuel off‐gas from the gasification process in the presence of oxygen carriers. 4) Regeneration of oxygen carriers. 317 318 Fuel Cell Systems Explained For the proposed system, the oxygen carrier is iron-based (Fe2O3/Fe3O4), and tests have shown that the process offers advantages over conventional coal gasification with CLC in terms of lower gasifier temperature (1068°C vs. 1700°C), increased hydrogen production and improved cold gas efficiency.30 The process has the further dual advantages of providing a ready means of CO2 separation prior to sequestration and the elimination of an air separation unit for the gasifier. Again, the use of oxygen carrier ensures that the hydrogen is produced with high purity and thereby is suitable for FCVs and other applications of PEMFCs. 10.10 Biological Production of Hydrogen 10.10.1 Introduction Reference has already been made in this chapter to the generation of hydrogen from biofuels (see Section 10.3). Consideration was given to various biofuels and how they could be used in fuel cells, mainly by converting to hydrogen‐rich gas via steam reforming and other processes reviewed in Sections 10.4, 10.5 and 10.6. By contrast, this section examines how biological methods might be employed to extract the hydrogen from any fuel — natural gas as well as biofuels. These biological systems proceed through one of three types of metabolic process as follows: 1) Photosynthesis by unicellular microorganisms that utilize either hydrogenase or nitrogenase reactions. 2) Digestion through the action of bacteria to produce hydrogen anaerobically. 3) Various stepwise processes that use a combination of bacteria to predigest complex organic molecules to make less complex organic material that can then be transformed with hydrogen‐producing organisms. Progress in developing these processes has been slow due to the following issues: ● ● ● ● The growth of organisms is inhibited by catabolites formed in microbiological cultures. Hydrogen production is limited because growth of organisms often slows down as hydrogen concentrations build up. There is only a narrow range of feedstocks on which the organisms can function. The rate of production of other gases is high, or the rate of production of hydrogen is low. Biological hydrogen generation is an active field of research, and, consequently, biologically active enzymes are now able to be isolated and modified. Results are starting to emerge that could have a profound influence on the perceptive of hydrogen as a future fuel. 10.10.2 Photosynthesis and Water Splitting Photosynthesis consists of two processes: (i) the conversion of light energy to biochemical energy by a photochemical reaction and (ii) the reduction of atmospheric carbon dioxide 30 Zhang, Y, Doroodchi, E and Moghtaderi, B, 2012, Thermodynamic assessment of a novel concept for integrated gasification chemical looping combustion of solid fuels, Energy & Fuels, vol. 26, pp. 287–295. Fuels for Fuel Cells to organic compounds such as sugars. In the first process, light is absorbed by chlorophyll, which acts as a mediator in the oxidation of water: (10.36) 2H2O + 2h 4H + 4e + O2 In the second process, the organic compound nicotinamide adenine dinucleotide phosphate (NADP) is reduced by electrons to a state generally designated as NADPH. Along with adenosine 5′‐triphosphate (ATP), NADPH is an important intermediary in the photosynthetic fixation of carbon dioxide. The protons and electrons react with carbon dioxide via these two mediators to produce sugars. The overall photo biochemical process taking place in green plants is represented by: nCO2 + 2nH2O ATP NADPH n(CH2O) nH2 O nO2 (10.37) where n is defined according to the structure of the resulting carbohydrate. Like the chlorophyll in plants, the pigments in some types of algae can absorb solar energy under certain conditions. A few groups of algae and cyanobacteria (formerly known as ‘blue‐green algae’) produce gaseous hydrogen rather than sugars via a photosynthesis route. Cyanobacteria contain hydrogenase or nitrogenase enzymes, and it is these that have the ability to catalyse hydrogen formation. In 1942 it was observed that the green alga Scenedesmus produces hydrogen when exposed to light after being kept in the dark and under anaerobic conditions.31 Further work directed towards elucidating the mechanism of this process identified hydrogenase as the key enzyme, which reduces water to hydrogen with concomitant oxidation of an electron carrier, ferredoxin. The green algae therefore became known as ‘water‐splitting’ organisms, and the process conducted in the photo biochemical cells was referred to as ‘biophotolysis’. Unfortunately, the hydrogenase in green algae is very sensitive to oxygen, which can rapidly deactivate the enzyme’s activity. In 2007, researchers found that production of oxygen by the hydrogenase could be blocked by adding copper to the algae. Sulfur had a similar effect. At the same time the Solar Biofuels Consortium — a collaboration between scientists at the University of Bielefeld (Germany) and Queensland (Australia) — managed to genetically modify the single‐cell green alga Chlamydomonas reinhardtii in such a way that it produces an especially large amount of hydrogen. The work demonstrated production rates up to five times the volume made by the wild form of alga and up to 1.6–2.0% energy efficiency.32 Although other species have exhibited good hydrogen production activity, e.g., Chlamydomonas moewusii and Anabaena cylindrica, C. reinhardtii has received the most attention from researchers, and scale‐up of algal bioreactors is now in progress. The unicellular aerobic nitrogen fixer Synechococcus sp. Miami BG043511 is another biological system that exhibits photosynthesis activity. It has provided a conversion efficiency estimated at around 3.5% based on photosynthetically active radiation (PAR) (i.e., light of energy 400–700 nm in wavelength) and using an artificial 31 Gaffron H. and Rubin J, 1942, Fermentative and photochemical production of hydrogen in algae, Journal of General Physiology, vol. 26, pp. 219–240. 32 Hankamer, B, Lehr, F, Rupprecht, J, Mussgnug, JH, Posten, C and Kruse, O, 2007, Photosynthetic biomass and H2 production by green algae: from bioengineering to bioreactor scale up. Physiologia Plantarum, vol. 131, pp. 10–21. 319 320 Fuel Cell Systems Explained light source. Certain other bacteria are also able to photosynthesize, but not via water oxidation. These function with either organic compounds or reduced sulfur compounds as the electron donors. The conversion efficiency of light energy to hydrogen in such systems can be much higher than that achieved with cyanobacteria. For example, efficiencies of 6–8% have been exhibited by Rhodobacter sp. in the laboratory. It is likely that solar efficiencies of 10% will soon be achieved by such bacteria. 10.10.3 Biological Shift Reaction In 1997, it was demonstrated that certain bacteria (Rhodospirillum rubrum) can utilize carbon monoxide and water to produce carbon dioxide and hydrogen gas via the WGS reaction. Subsequently, the US DOE has funded the research and development of this type of biological process. The implications of carrying out the WGS reaction in a bioreactor are important since normally catalysts operating at high temperatures are required to ensure reasonable rates. If the reaction could be carried out at near-ambient temperature, the product gas would contain little CO and therefore further expensive gas clean‐up for PEMFCs could be eliminated or at least greatly reduced. In a recent study, the WGS reaction was separated into two half‐cell electrochemical reactions, namely, H+ reduction and CO oxidation. The former reaction was catalysed by a hydrogenase, Hyd‐2, from Escherichia coli, and the latter reaction by a carbon monoxide dehydrogenase (CODH I) from Carboxydothermus hydrogenoformans. Both enzymes were attached to conducting carbon particles. The resulting electrocatalyst proved to be highly active for WGS when compared with more conventional, high‐ temperature, supported‐metal WGS catalysts.33 10.10.4 Digestion Processes Hydrogen can be produced by microbial digestion of organic matter in the absence of light energy. Under relatively mild conditions of temperature and pressure, many bacteria will readily produce hydrogen together with acetic acid and other low molecular weight organic acids. Nevertheless, the rates of reaction are usually low, and hydrogen is not produced in significant amounts because of two mitigating factors. First, inhibition of the microbial hydrogenase may occur as hydrogen builds up. Second, hydrogen may react with other organic species that may also be present, or with carbon dioxide in the system, to cause the generation of methane. As the partial pressure of hydrogen increases, the forward reaction of organic matter to hydrogen becomes thermodynamically unfavourable. The challenge in using digestion processes is therefore to increase the rate of hydrogen production while preventing methane formation. Hydrogen production generally occurs by fermentation of carbohydrate‐ rich material in the organic waste and is carried out by anaerobic bacteria belonging to species such as Enterobacter, Bacillus and Clostridium. More recently, the thermophilic Thermotoga neapolitana has also demonstrated considerable promise. 33 Lazarus, O, Woolerton, TW, Parkin, A, Lukey, MJ, Reisner, E, Seravalli, J, Pierce, E, Ragsdale, SW, Sargent, F and Armstrong, FA, 2009, Water-gas shift reaction catalyzed by redox enzymes on conducting graphite platelets, Journal of the American Chemical Society, vol. 131(40), pp. 14154–14155. Fuels for Fuel Cells This bacterium has the potential to utilize a variety of organic wastes and to offer a cost‐effective method of generating significant quantities of hydrogen. Microbial digestion in the absence of light (‘dark fermentation’) produces a mixed biogas containing primarily hydrogen and carbon dioxide. To maximize hydrogen production, it is necessary to optimize the activity of the enzyme hydrogenase. In this respect, recent studies have shown that the pH should be maintained in the range of 5–6.5, with an optimum value of 5.5.34 Digestion is attractive for treating sewage sludge as well as agricultural waste that would otherwise have low value, such as cheese whey and dairy manure. Unfortunately, the yield of hydrogen (e.g., as compared with methane) from anaerobic digestion is low. Although biological processes for hydrogen production are actively being developed, many questions remain concerning the fundamental biochemical processes that are occurring. Photosynthesis‐algae or photosynthesis‐bacteria systems seem to be the best candidates for the first technical applications. Present indications are that hydrogen production costs of 12 cents per kWh H2 or less are achievable. Further Reading Brown, RC and Stevens, C, 2011, Thermochemical Processing of Biomass: Conversion into Fuels, Chemicals and Power, John Wiley & Sons, Inc., Hoboken, NJ. ISBN:978‐0‐470‐ 72111‐7. Carmo, M, Fritz, DL, Mergel, J and Stolten, D, 2013, A comprehensive review on PEM water electrolysis, International Journal of Hydrogen Energy, vol. 38(12), pp. 4901–4934. Dincer, I and Joshi, AS, 2013, Solar Based Hydrogen Production Systems, Springer, New York. DOI 10.1007/978‐1‐4614‐7431‐9_2. ISBN 978‐1‐4614‐7430‐2. Hallenbeck, PC (ed.), 2012, Microbial Technologies in Advanced Biofuels Production, 15, Springer US, Boston, MA. DOI 10.1007/978‐1‐4614‐1208‐3_2. Hoogers, G, 2003, Fuel Cell Technology Handbook, CRC Press, Boca Raton, FL. ISBN 0‐8493‐0877‐1. HTGR‐integrated hydrogen production via steam methane reforming (SMR) process analysis, 2010, Technical Evaluation Study Project No. 23843, Idaho National Laboratory. Kahn, MR, 2011, Advances in Clean Hydrocarbon Fuel Processing: Science and Technology, Series in Energy, Woodhead Publishing, Philadelphia, PA. ISBN‐10: 1845697278. Kidnay, AJ, Parrish, WR and McCartney, DG, 2010, Fundamentals of Natural Gas Processing, 2nd edition, CRC Press, Boca Raton, FL. ISBN‐13:978‐1420085198. Kolb, G, 2010, Fuel Processing: For Fuel Cells, Wiley‐VCH, Weinheim. ISBN: 978‐3‐527‐31581‐9. Rand, DAJ and Dell, RM, 2008, Hydrogen Energy: Challenges and Prospects, The Royal Society of Chemistry, Cambridge. ISBN: 978‐0‐85404‐597‐6. 34 Valdez–Vazquez, I and Poggi–Varaldo, HM, 2009, Hydrogen production by fermentative consortia, Renewable and Sustainable Energy Reviews, vol. 13, pp. 1000–1113. 321 323 11 Hydrogen Storage 11.1 Strategic Considerations Society has readily adapted to many different fuels such as petroleum, diesel oil, naphtha, coal and biofuels (e.g., synthetic diesel, ethanol–gasoline blends) — but hydrogen is different. As a gas under normal temperature and pressure, hydrogen presents its own challenges as a fuel. It has to be generated either by extraction from another fuel or by the decomposition of water by electrolysis or photo‐electrolysis. Unlike electricity, however, that has become ubiquitous and the backbone of conventional energy‐supply networks, hydrogen can be stored relatively easily in bulk. There should be no surprise therefore that hydrogen has been promoted in recent years as a prospective means of storing renewable energy through offering long‐term advantages over battery systems. Over the past decade, the issue of climate change has made the case for zero‐ emission vehicles so strong that hydrogen has almost universally been adopted as the fuel of choice for fuel‐cell vehicles (FCVs). This is remarkable considering the substantial technical, environmental and fiscal challenges associated with the development of hydrogen infrastructure. A summary of some applications for which hydrogen is already finding use, together with current example systems, is presented in Table 11.1. Small fuel‐cell systems of a few kW or below are generally designed to run on stored hydrogen because it is difficult to scale down conventional methods of hydrogen generation. Whereas for systems of about 50 kW and greater it is cost‐effective to couple the fuel processor directly with the fuel‐cell stack, this is not the case for small proton‐ exchange membrane fuel cell (PEMFC) systems. A small local store of hydrogen is therefore an essential component of fuel‐cell systems for portable applications, unless the direct methanol fuel cell (DMFC) is being employed. Hydrogen can also be a reasonable way of storing electrical energy from sources such as wind‐driven generators and hydroelectric power, where generation might well be out of line with consumption. Electrolysers convert the electrical energy to hydrogen during times of high supply and low demand. In general, the storage of hydrogen for stationary applications is less demanding than for transportation systems in which there are more severe constraints in terms of acceptable mass and volume, speed of charge and discharge and, for some storage systems, heat management. Finding a satisfactory solution for the on‐board containment of hydrogen has proven to be a major challenge in the development of FCVs. Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 324 Fuel Cell Systems Explained Table 11.1 Applications that employ hydrogen storage. Application Method of hydrogen storage Example system Fuel‐cell car Compressed gas, 70 MPa Toyota Mirai, 2015 Hydrogen internal combustion‐engined car Liquid hydrogen, −252°C BMW series 7 sedan Fuel‐cell bus Compressed gas, 70 MPa Mercedes‐Benz Citaro Fuel‐cell bicycle/ motorcycle Compressed gas Palcan, Intelligent Energy Locomotive Low‐temperature hydride Anglo‐American Platinum, mining locomotive, Dishaba Mine, South Africa Airplane Liquid hydrogen Boeing Phantom‐Eye Fuel‐cell battery charger Low‐temperature hydride Intelligent Energy, Upp, and Horizon MiniPak portable chargers Renewable energy storage Low‐temperature hydride Sir Samuel Griffith Centre (30 MWh), Australia Compressed gas, 70 MPa INGRID project (39 MWh), Italy Natural gas + hydrogen blend Compressed gas ENEA, Regione Emilia Romagna, Italy, Hydrogen‐Compressed Natural Gas (HCNG) bus trials Industrial gas distribution Medium‐temperature hydride Hydrexia, Ni–Mg hydride Spacecraft Liquid hydrogen Government space agencies of China, Europe, India, Russia and the United States McPhy, Mg hydrogen Although hydrogen has very high specific energy (Wh kg−1), which makes it the fuel of choice for space missions, its volumetric energy density (Wh m−3) is very low. The latter feature is a great disadvantage compared with most other fuels. The energy density can be improved by compressing the gas, and pressures up to 70 MPa are routinely used for storage on FCVs. Unlike liquefied petroleum gas (LPG) or butane, which can be liquefied at ambient temperature by raising the pressure, hydrogen can only be liquefied by cooling the gas down to about 22 K. As a liquid, the energy density is quite low, 71 kg m−3. It has to be stored in a thermally insulating ‘Dewar’ vessel, but even with the best designs some loss (‘boil‐off ’) of liquid hydrogen by evaporation is inevitable. Hydrogen can also be stored in various chemical compounds that, on a gravimetric basis, can hold quite large quantities of hydrogen. To be useful, a given compound must pass the following three tests: 1) The compound must readily give up the hydrogen — otherwise there is no advantage over using a reformed fuel in one of the ways described previously in Chapter 10. 2) The manufacturing process must be simple and absorb little energy — in other words the energy and financial costs of hydrogen becoming incorporated in the compound must be low. 3) The compound must be safe to handle. Hydrogen Storage Table 11.2 Possible materials for hydrogen storage. Formula Percent hydrogen Liquid hydrogen H2 100 0.07 14.0 Cold, −252°C Liquid methane CH4 25.13 0.422 9.6 Cold, −175°C Ammonia NH4 17.76 0.682 8.5 Toxic, >100 ppm Name Density (kg L−1) Volume (L) to store 1 kg H2a Notes Liquids/gases Water H2O 11.11 1.00 8.9 Hydrazine N2H4 12.58 1.011 7.8 Methanol CH3OH 12.50 0.79 10.0 Ethanol C2H5OH 13.00 0.790 9.7 6.30 1.060 15.0 30 wt.% sodium NaHBH3 + H2O borohydride solution Toxic, >10 ppm Expensive, but works well Simple hydrides Lithium hydride LiH 12.68 0.82 6.50 Sodium hydride NaH 4.30 0.92 25.9 Diborane B2H6 21.86 0.417 11.0 Beryllium hydride BeH2 18.28 0.67 8.2 Silane SiH4 12.55 0.68 12.0 Calcium hydride CaH2 5.00 1.90 11.0 Aluminium hydride AlH3 1.30 7.1 Potassium hydride KH 2.51 1.47 27.1 Titanium hydride TiH2 4.40 3.90 5.8 10.8 Caustic Caustic but cheap Toxic Highly toxic Toxic, >5 ppm Caustic Complex hydrides Lithium borohydride LiBH4 18.51 0.666 8.1 Mildly toxic Aluminium borohydride Al(BH4)3 16.91 0.545 11.0 Mildly toxic Lithium aluminium hydride LiAlH4 10.62 0.917 10.0 Palladium hydride Pd2H Titanium–iron hydride TiFeH2 a) 0.47 10.78 20.0 1.87 5.47 9.8 All the extra equipment required to hold or process the compound is excluded. Consequently, each entry is not a total number and should only serve as a guide. For example, all the alkali metal hydrides require large quantities of water, from which some of the hydrogen is also released. A large number of promising chemicals have been suggested or tested; some examples, together with their key properties, are listed in Table 11.2. Unfortunately, many of the candidates do not warrant a great deal of consideration, as they fail one or more of the three tests. Hydrazine, for instance, passes the first test very well (it has been successfully demonstrated in fuel‐cell systems — see Section 5.2.3, Chapter 5) but is highly toxic and 325 326 Fuel Cell Systems Explained very energy intensive to manufacture, and so it fails the second and third tests. Nevertheless, several other hydrogen compounds have found practical application and are described in detail later on. The most important of these involves the use of metal hydrides. There are two principal forms: a ‘rare earth’ metal hydride compound that can reversibly store and deliver hydrogen and alkali metal hydrides that react with water to yield hydrogen gas. The principal methods of storing hydrogen that will be described in this chapter are therefore: Compression in gas cylinders. As a cryogenic liquid. As a reversible metal hydride. As a metal hydride that reacts with water. ● ● ● ● Before discussing these methods, which each have advantages and disadvantages, consideration must be given to issues of safety. 11.2 Safety Compared with all gases, hydrogen has (i) the lowest molecular weight, (ii) the highest thermal conductivity, velocity of sound and mean molecular velocity and (iii) the lowest viscosity and density. Accordingly, hydrogen leaks through small orifices faster than all other gases, e.g., 2.8 and 3.3 times faster than methane and air, respectively. In addition, hydrogen is a highly volatile and flammable gas, and in certain circumstances mixtures of hydrogen and air can detonate. Safety considerations must therefore feature strongly in the design of any fuel‐cell system. All measures should be taken to avoid the risk of hydrogen escape, and systems should be equipped with sensors to shut off gas supplies and alert personnel should leaks occur. Although safety should be a priority, it is important to stress that hydrogen is no more dangerous, and in some respects it is rather less dangerous, than various other conventional fuels. The key properties relevant to the safety of hydrogen and two other gaseous fuels that are widely available — methane and propane — are listed in Table 11.3. The lower concentration limit for ignition of hydrogen is much the same Table 11.3 Properties relevant to safety for hydrogen and two other commonly used gaseous fuels. Density (kg m−3 at NTPa) a Hydrogen Methane Propane 0.084 0.65 2.01 Ignition limits in air (vol.% at NTP ) 4.0–77 4.4–16.5 1.7–10.9 Ignition temperature (°C) 560 540 487 Minimum ignition energy in air (MJ) 0.02 0.3 0.26 Maximum combustion rate in air (m s−1) 3.46 0.43 0.47 Detonation limits in air (vol.%) 18–59 6.3–14 1.1–1.3 Stoichiometric ratio in air 29.5 9.5 4.0 a) NTP, normal temperature and pressure is 20°C and 101.325 kPa. Hydrogen Storage as for methane. For propane a lower concentration is necessary for ignition. The ignition temperatures for hydrogen and methane are similar, but both are higher than that for propane. The minimum ignition energy of hydrogen is, however, very low, and thereby it suggests that a fire could be started very easily. In fact, the ignition energies of all the gases are actually lower than would be encountered in most practical cases. A spark could ignite any of these three fuels. Furthermore, against this must be set the much higher minimum concentration needed for detonation of hydrogen in air. Another potential hazard arises from the rather greater range of concentrations required to cause detonation of hydrogen. Care must therefore be taken to prevent the accumulation of hydrogen in confined spaces. Fortunately, the high buoyancy and high average molecular velocity mean that of all common gases, hydrogen disperses most rapidly and is therefore less likely than other gases to build up to a level required for detonation. Comparing the data given in Table 11.3, hydrogen appears much the same as the other fuels from the point of view of potential danger. It is the much lower density that gives hydrogen an advantage from a safety point of view. The density of methane is similar to air, which means it does not disperse quickly but tends to mix with air. Propane has a lower density than air, which tends to make it sink and collect at low points, such as in basements, drains and the hulls of boats, where it can explode or be set alight with devastating effects. Hydrogen, on the other hand, is so light that it rapidly disperses upwards. The concentration levels necessary for ignition or detonation are therefore much less likely to be achieved for hydrogen in comparison with the other fuels. Hydrogen, like all fuels, must be handled carefully. All things considered, however, it does not present any greater hazard than any other flammable liquids or gases encountered today. The one unique characteristic that should always be borne in mind by developers or users of fuel cells is that, once ignited, hydrogen burns with an invisible flame. 11.3 Compressed Hydrogen 11.3.1 Storage Cylinders When hydrogen is produced at a central facility, it can either be stored in bulk before dispatch to customers or first distributed and then stored locally on‐site until needed. Whatever form delivery is taken, e.g., as gas in cylinders or in pipelines or as liquid hydrogen, it is first necessary to compress the gas. To do this, work has to be done, and energy expended. Broadly, the energy required to compress hydrogen lies between 5 and 15% of the higher heating value (HHV), as determined by the final pressure, and whether the process is carried out adiabatically or isothermally. In practice, a multistage compression process will probably be adopted, and if mechanical and electrical losses are included, the total energy wasted in compression can amount to 20%. The main advantages of storing hydrogen as a compressed gas are: ● ● ● Simplicity. Indefinite storage time. No purity limits on the hydrogen. 327 328 Fuel Cell Systems Explained Storing hydrogen gas at pressure in steel cylinders is the most technically straightforward method of holding hydrogen and the most widely adopted for small amounts of gas. Tube trailers, in which the compressed gas is contained in horizontal steel cylinders, are the normal delivery method for merchant hydrogen, i.e., where the quantities involved are considerable. The steel cylinders are permanently fixed to the trailer and are discharged in situ, i.e., not offloaded; a trailer is shown in Figure 11.1. Small‐scale consumers, such as laboratories, employ cylinders with storage capacities of just a few cubic metres (pressurized to 20 MPa) that can be manhandled with simple trolleys. The specifications of one such cylinder are compared in Table 11.4 with those of a larger cylinder used for storage on a bus or other road vehicle. The latter is constructed with an aluminium inner liner of 6 mm thickness, around which is wrapped a composite of aramid fibre and epoxy resin. This material has a high ductility and thereby has good burst behaviour in that it rips apart rather than disintegrating into many pieces. The burst pressure is 120 MPa.1 Figure 11.1 Tube trailer for delivery of compressed hydrogen gas — 30 tubes and 1225 m3 (105 kg) capacity. (Source: Reproduced with permission of Coregas.) 1 It should be noted, however, that at present composite cylinders are about three times the cost of steel cylinders of the same capacity. Hydrogen Storage Table 11.4 Comparative data for two cylinders used to store hydrogen at high pressure. 2‐L steel (20 MPa) 147‐L composite (30 MPa) Mass of empty cylinder (kg) 3.0 100 Mass of hydrogen stored (kg) 0.036 3.1 Storage efficiency (wt.% H2) 1.2 3.1 Specific energy (kWh kg−1) 0.47 1.2 a 3 Volume of tank (approximately), L (m ) 2.2 (0.0022) Mass of H2 (kg L−1) 0.016 a) 220 (0.22) 0.014 The storage efficiency here is defined as the total mass of hydrogen stored divided by the mass of the empty cylinder, expressed as wt.% H2. 11.3.2 Storage Efficiency In considering the storage of hydrogen on an FCV, it would be expected that a larger storage system would be a great deal more efficient than a smaller one in terms of specific mass of hydrogen stored. Therefore, the larger vessel (147 L) in Table 11.4 would be expected to hold proportionately more hydrogen than the (2 L) vessel. It should be remembered, however, that large tanks have to be secured in the vehicle, and consequently the weight of the supporting structure should be taken into account in assessing efficiency, which is defined as the mass of hydrogen stored divided by the mass of the storage medium (in this case the empty vessel). In one of the early European buses, in which hydrogen was used to fuel an internal combustion engine (ICE), 13 composite tanks were mounted in the roof space. The total mass of the tanks and the bus structure reinforcements was 2550 kg, or 196 kg per tank. This should be compared with the mass of an empty composite tank given in Table 11.4 of 100 kg. The increase in mass from 100 to 196 kg has the effect of reducing the ‘storage efficiency’ of this particular system to 1.6 wt.%, i.e., not so very different from that of the (2 L) steel cylinder listed in Table 11.4. Another point is that the weight of connecting valves, pressure‐reducing regulators and other essential hardware has also to be taken into account whether the system is built around steel or composite cylinders. When employing a 2.2 L steel cylinder, these components would typically add about 2.15 kg to the mass of the system and thus reduce the storage efficiency from 1.2 to 0.7 wt.%.2 The low density of hydrogen is responsible for the poor mass storage efficiency, even at such high pressures. The density of hydrogen gas at ambient temperature and pressure is 0.084 kg m−3, whereas air is about 1.2 kg m−3. In practice, compressed hydrogen often accounts for less than 2% of the total mass of the storage system. 2 Kahrom, H, 1999, Clean hydrogen for portable fuel cells, Proceedings of the European Fuel Cell Forum Portable Fuel Cells Conference, 21–24 June 1999, Lucerne, pp. 159–170. 329 330 Fuel Cell Systems Explained 11.3.3 Costs of Stored Hydrogen For small‐scale fuel cells, by taking into account all expenditure, e.g., depreciation of cylinders, administration and purchase of pressure‐reducing valves, it has been estimated that the cost of hydrogen fuel is about US$2.2 per g2. Using data given in Section A2.4, Appendix 2, this expenditure corresponds to about US$56 per kWh, or about US$125 per kWh for the electricity from a fuel cell of 45% efficiency. This is absurdly expensive when compared with mains electricity but is noticeably cheaper than current battery storage.3 11.3.4 Safety Aspects The metal employed in the fabrication of the hydrogen storage vessel requires careful selection. Hydrogen is a very small molecule and is capable of diffusing into materials that are impermeable to other gases. This is compounded by the fact that a very small fraction of the hydrogen gas molecules may dissociate on the surface of the material. Diffusion of atomic hydrogen into the material may then occur and compromise its mechanical integrity. Gaseous hydrogen can build up to form internal blisters in the material that, in turn, can lead to crack promotion. With carbonaceous metals, such as steel, the hydrogen can react with the carbon to produce bubbles of entrapped methane. The gas pressure in the resulting internal voids can generate an internal stress that is sufficient to cause fissures, cracks or blisters in the steel. The phenomenon is well known and is termed ‘hydrogen embrittlement’. Fortunately certain chromium‐rich steels and chromium–molybdenum alloys are resistant to hydrogen embrittlement, and composite reinforced plastic materials can also be used for larger tanks, as mentioned previously. As well as the problem of the very high mass associated with the storage vessel, there are significant hazards associated with storing hydrogen at high pressure. A leak from such a cylinder would generate very large forces as the gas is released. It is possible for such cylinders to become essentially jet‐propelled torpedoes and thereby inflict considerable damage. Furthermore, vessel fracture would most likely be accompanied by auto‐ignition of the released hydrogen in air and give rise to a fire that lasts until the contents of the ruptured, or accidentally opened, vessel are consumed. Nevertheless, compressed hydrogen is widely and safely used by correctly following established procedures and guidelines. In vehicles, for example, pressure‐relief valves or rupture discs (see following Section 11.4) are fitted and will safely vent gas in the event of a fire. Similarly, pressure regulators attached to hydrogen cylinders are equipped with flame traps to prevent ignition of the gas. Compressed hydrogen storage is most widely employed in places where there is moderate but variable demand. It is also used for road vehicles, both those with ICEs and those with fuel‐cell systems. Figure 11.2 shows an example of composite compressed hydrogen storage tanks on an FCV. 3 Lead–acid and lithium‐ion batteries cost around US$250 and US$300 per kWh, respectively, in 2015. Nykvist, B and Nilsson, M, 2015, Rapidly falling costs of battery packs for electric vehicles, Nature Climate Change, vol. 5, pp. 329–332. Hydrogen Storage Figure 11.2 Composite hydrogen storage tanks on the Honda FCX fuel‐cell car chassis. (Source: https://zh.wikipedia.org/wiki/%E6%B0%AB%E6%B0%A3%E7%AE%B1. CC BY‐SA 3.0.) 11.4 Liquid Hydrogen The storage of hydrogen as a liquid (denoted as LH2) at about 20 K is currently the only method that is widely practised for storing substantial quantities of hydrogen. A gas cooled to the liquid state in this way is known as a ‘cryogenic liquid’, and large amounts are currently required for processes such as petroleum refining and ammonia production. NASA is another notable customer and has huge 3200 m3 (850 000 US gallon) tanks to ensure a continuous supply for space exploration (see Figure 11.3a). The container or tank for storing cryogenic liquid hydrogen is a large, strongly reinforced vacuum (Dewar) flask. Given that the container is never perfectly insulated, liquid hydrogen will slowly evaporate and cause the pressure in the container to rise. The maximum allowable pressure is normally below 300 kPa, though some larger tanks may be designed for greater pressures. If the rate of evaporation exceeds the demand, then the tank is occasionally vented to make sure the pressure does not rise too high. A spring‐loaded valve will release and then close again when the pressure falls. The small amounts of hydrogen involved are routinely released to the atmosphere, though in very large systems they may be vented out through a flare stack and burnt. A rupture disc is commonly fitted as a backup safety feature. This device consists of a ring covered with a membrane of controlled thickness that will withstand a certain safe pressure. When the pressure in the system exceeds the allowable pressure of the rupture disc, the membrane bursts, and gas is released and safely vented away until the disc is replaced. Usually, the fault cannot be rectified until all of the gas has been vented from the storage system. When an LH2 tank is being filled, and when fuel is being withdrawn, it is most important that air is not allowed into the system; otherwise, an explosive mixture 331 332 Fuel Cell Systems Explained (a) (b) (c) (d) Figure 11.3 Examples of cryogenic LH2 tanks: (a) NASA bulk storage for space missions. (b) installation for AC Transit fuelling station at Emeryville, California. (c) Tank in the rear of BMW series 7 car. (d) Proposed bulk ocean transport from Australia to Japan by Kawasaki Heavy Industries. (Source: Reproduced with permission of Kawasaki Heavy Industries.) could form. The tank therefore should be purged with nitrogen before filling. Figure 11.3b shows an example of an LH2 tank installed as part of vehicle filling station for AC Transit in California. Considerable effort has gone into the design and development of LH2 tanks for cars, though not specifically for FCVs. Several automobile companies have invested in hydrogen‐fuelled ICEs, most of which have employed LH2. The tank in the BMW hydrogen cars, for example (see Figure 11.3c), is cylindrical in shape and has the conventional double‐wall vacuum flask type of construction. The walls are about 3 cm thick and consist of 70 layers of aluminium foil interlaced with fibreglass matting. The maximum operating pressure is 500 kPa. The tank stores 120 L of cryogenic hydrogen that, as the density of LH2 is very low (~71 kg m−3), weighs only 8.5 kg. The key features of the BMW cryogenic tank are given in Table 11.5. One of the problems associated with cryogenic hydrogen is that the liquefaction process is very energy intensive. Several stages are required. After an initial stage of compression, the gas is cooled to about 78 K under the action of liquid nitrogen. The hydrogen then is cooled further by expansion through a turbine. Finally, a magnetocaloric Hydrogen Storage Table 11.5 Details of a cryogenic hydrogen container for BMW cars. Mass of empty container (kg) 51.5 Mass of hydrogen stored (kg) 8.5 Storage efficiency (% mass H2) 14.2 Specific energy (kWh kg−1) 5.57 3 Volume of tank (approximately, m ) 0.2 Mass of H2 (kg L−1) 0.0425 process is performed to convert ortho‐H2 to para‐H2.4 The total energy required to liquefy the gas is about 40% of the specific heating value of the hydrogen. In addition to the regular safety concerns with hydrogen, there are a number of specific difficulties concerned with cryogenic storage. All pipes containing the fluid must be insulated, as must any parts in good thermal contact with these pipes. This precaution is necessary to minimize the chance of frostbite if human skin is brought into contact with the system. Insulation is also necessary to prevent the surrounding air from condensing on the pipes; otherwise an explosion could develop if liquid air drips onto nearby combustibles. Asphalt, for example, can ignite in the presence of liquid air. Thus, concrete paving is laid around static installations. In general, however, the hazards of hydrogen are somewhat less with LH2 than with pressurized gas. For instance, if there is a failure of the container, the fuel tends to remain in place and vent to the atmosphere more slowly. Certainly, LH2 tanks have been approved for use in cars in Europe. Kawasaki has designed an ocean‐going tanker to transport hydrogen from Australia to Japan (the concept is shown in Figure 11.3d). Filling stations for fuel‐cell vehicles are starting to become well established in numerous locations — the United States, Japan and Europe. Most of these incorporate high‐pressure storage and some also store liquid hydrogen. Figure 11.4 shows an example of a filling station in the United Kingdom that delivers compressed hydrogen generated by electrolysis. 11.5 Reversible Metal Hydrides Certain metals and alloys have the ability to absorb and release gaseous hydrogen, reversibly, via the formation of hydrides, i.e., M x H2 2 MH x heat (11.1) The hydrogen molecule first dissociates into its two atoms, which are chemisorbed on the surface of the metal/alloy, and then diffuses into the bulk lattice. The dissolved atoms can take the form of a random solid solution or react to produce a hydride of fixed stoichiometric composition. The quantity of hydrogen absorbed is expressed in terms of the hydride composition, either on a molar basis (MHx) or on a wt.% basis. 4 Both atoms in the molecule of ortho‐hydrogen have nuclear spins in parallel. In para‐hydrogen, the spins are antiparallel. The process of converting ortho‐ to para‐hydrogen is exothermic and if allowed to take place naturally would cause boil‐off of the liquid. 333 334 Fuel Cell Systems Explained Figure 11.4 Refilling with H2. (Source: http://theconversation.com/hydrogen‐car‐progress‐hasnt‐ stalled‐yet‐15821. CC BY‐ND 4.0.) The rates of absorption and desorption can be controlled by adjusting the temperature or pressure. (Note: Heat is liberated during formation of the hydride and therefore must be added to effect its subsequent decomposition with discharge of the hydrogen.) Hydrides can be tailored to operate over a wide range of temperatures and pressures. To be most useful as a hydrogen store, the metal or alloy should react with the gas at, or near, ambient temperature and at not too great a pressure. The enthalpy of absorption should also not be too high. Otherwise, heat transfer becomes a problem, especially in large hydride beds. Finally, the system should be capable of sustaining a practical number of absorption–desorption cycles without deterioration. The remarkable hydriding properties of LaNi5 were discovered in about 1969 at the Philips laboratories in the Netherlands. It has served as a benchmark for many hydriding alloys to be developed by other researchers. Metal alloys that have been designed and developed in succeeding years generally fall into one of the following types: AB5, A2B7, AB3, AB2 (Laves phase), AB and A2B. In these alloys, A represents a metal element with a strong affinity for hydrogen (i.e., an ability to absorb hydrogen), and B is a metallic element with weak hydrogen affinity but strong activity for catalysing the dissociation of H2 molecules to H atoms. Of these classes of materials, the most studied are probably the AB5 alloys in which A is calcium or a rare earth element and B is customarily a transition metal (e.g., Fe, Co, Ni, V, Mn, Cr). These alloys are remarkable due to the fact that their hydriding properties can be ‘fine‐tuned’ by alloying A or B with other transition metal elements. The hydrides are reversible with good kinetics and moderate operating pressures. Hydrogen Storage One example of a well‐studied material is the last entry in Table 11.2, v.s. titanium– iron hydride. In terms of mass, this does not appear to be a very promising material. Rather, it is the volumetric measure that gives such hydrides an advantage over other storage methods. Among all the examples in Table 11.2, the titanium–iron hydride requires one of the lowest volumes to store 1 kg. This hydride actually holds more hydrogen per unit volume than pure liquid hydrogen and this appear counterintuitive. In liquid hydrogen the molecules have a relatively high mobility that gives rise to the relatively low density of the material — only 0.07 kg L−1. By contrast, hydrogen molecules that are bonded to metal atoms in a hydride are bound closer together and thus give rise to the high storage capacity per unit volume, despite the fact that the density of the hydride material is higher at 5.47 kg L−1. A typical reversible metal hydride system functions as follows. Hydrogen is supplied at slightly above atmospheric pressure to the metal alloy (in the form of a powder) inside a container. The reaction (11.1) proceeds to the right to form the hydride. Given that the process is mildly exothermic, for large systems, the hydride container has to be cooled. The hydride‐forming stage takes place at approximately constant pressure and will require a few minutes, depending on the size of the system and how well the container is cooled. In this context, if the pressure of gas in the container P (or, as a rule, log P) is plotted versus the ratio of the quantity of hydrogen absorbed, Hads, to the maximum uptake of the material, M, then an absorption isotherm is obtained and is referred to as a ‘pressure composition isotherm (PCT)’ curve. An example of the typical behaviour exhibited by metal hydrides is shown in Figure 11.5. Once all the metal has reacted with the hydrogen, i.e., Hads/M = 1.0, the pressure in the container will begin to rise. At this point, the hydrogen supply is disconnected as the vessel has reached its capacity and needs to be sealed. When the stored hydrogen is required, the vessel is connected to, for example, a PEMFC. Hydrogen will be released so long as the pressure of the fuel cell is lower than the desorption pressure, Pd. If the pressure rises above Pd, the reaction will slow down or stop. The process is now endothermic, so heat must be provided. This is taken from 5.8 5.6 Log P (P in Pa) Figure 11.5 PCT curve for hydrogen absorption and desorption with a LaNi5 alloy. Note that absorption and desorption have different pressures (Pa, Pd) in the almost horizontal portion of the relationship; pressure is given in Pa. Pa 5.4 5.2 5.0 0.0 Pd 0.2 0.4 0.6 Hads / M 0.8 1.0 335 Fuel Cell Systems Explained the surroundings — the vessel will cool slightly during discharge of the hydrogen. It can be warmed slightly to increase the rate of supply by using, for example, the hot water or air from the cooling system of the fuel cell. Once the reaction has been completed and all the hydrogen has been withdrawn, the whole procedure can be repeated. Different hydrides will exhibit their own characteristic PCT. Each curve usually demonstrates hysteresis between absorption and desorption, i.e., Pa is generally higher than Pd. Also note that the useful part of the isotherm is the almost horizontal section that shows the range of hydrogen uptake (Hads/M) over which hydrogen can be absorbed and desorbed. This range is always less than the maximum amount that can be absorbed by the material. The enthalpy change, ΔH, for the hydriding reaction (11.1) can be related to the equilibrium dissociation pressure, P, of the hydride by the van’t Hoff 5 equation, i.e., ln P H RT S R (11.2) where ΔS is the change in entropy, T is the absolute temperature and R is the gas constant. A logarithmic plot of the dissociation pressure against the reciprocal of the absolute temperature should therefore be linear with a slope that is a measure of the heat that has to be supplied to form or decompose the hydride. A series of plots for various hydrides over the temperature range −20 to 400°C is presented in Figure 11.6. The data clearly show the LaNi5H6, TeFeH and MmNi5H6 hydrides exhibit good 700 500 Temperature/°C 200 150 100 300 10 6 Na3AIH6 Mg2FeH6 50 2 20 0 –20 Na3AIH4 MmNi5H6 4 Pressure/MPa 336 TiCr1.8H1.7 Mg2NiH4 PdH0.6 1 CaNi5H4 TiFeH 6 4 2 0.1 6 4 LaNi5H6 MgH2 2 LaNi3.5AI1.5H5 0.01 1.0 1.5 LaNi4AIH5 2.0 2.5 3.0 1000/Temperature/K–1 3.5 4.0 Figure 11.6 Dissociation pressures of various metal hydrides. Note that Mm in MmNi5H6 denotes mischmetal, which is a mixture of alkali earth elements. 5 Jacobus Henricus van’t Hoff (1852–1911) was one of the founders of modern physical chemistry. He was the first recipient of the Nobel Prize in Chemistry in 1901. Hydrogen Storage characteristics at close to normal temperatures and pressures. These are examples of materials that have received the most attention for application in low‐temperature fuel cells. High‐temperature materials, such as the various magnesium hydrides, are attractive in that the storage capacity is greater than that for low‐temperature materials. They are, however, more prone to degradation due to the repeated expansion and contraction of the metal lattice during adsorption–desorption cycling. Indeed, provision must be made in the design of container to allow for this behaviour. Much research has been carried out on improving both the long‐term performance and the kinetics of absorption–desorption by exploring various metal alloys. For example, alloying magnesium and nickel to give Mg2Ni provides a material that is more easily activated than pure magnesium to yield the hydride Mg2NiH4 (3.6 wt.% H2). Of the hydride materials that have been explored for hydrogen storage at low and medium temperatures, the alloys of aluminium (alanates) have probably received the most attention from researchers. Generally, hydride materials can sustain several hundred charge–discharge cycles. Nevertheless, as with rechargeable batteries, these systems can be abused. For example, if the hydride is filled at too high pressure, the charging reaction will proceed too fast, and the material may overheat and degrade. Hydrides can also be damaged by impurities in the hydrogen, for instance, hydrogen produced by an alkaline electrolyser must be purified to remove any traces of water, alkali or oxygen. The vessel in which the hydride is contained should be able to withstand a reasonably high pressure, especially if it is likely to be filled from a high‐pressure supply. For example, the small HydrostikTM unit shown in Figure 11.7 and manufactured by Horizon fuel cells has a rated charging pressure of 3.0 MPa. The unit is intended to be used with a ‘MiniPak’ USB charger for fuel cells, as discussed in Section 4.9.1, Chapter 4. The Horizon technology employs an AB5‐type metal hydride with an expected lifetime of 10 years and, it is claimed, a stored energy that is equivalent to 10 disposable AA batteries, but of course the unit can be recharged. Further details are given in Table 11.6. The volumetric measure (i.e., mass of hydrogen per litre) is nearly as good as that for LH2, and the gravimetric measure is very much the same as for a small cylinder of compressed gas. Figure 11.7 Hydrostik hydrogen canister (left) that accompanies the Horizon ‘MiniPak’ fuel‐cell USB charger (right), employed to power portable electronic equipment such as smartphones. (Source: Reproduced with permission of Horizon fuel cells.) 337 338 Fuel Cell Systems Explained Table 11.6 Details of Horizon Hydrostik hydrogen container for portable electronics equipment. Dimensions of container (mm) 22 (diameter) × 8 (length) Mass of empty container (g) 105 Mass of hydrogen stored (g) 1 Storage efficiency (wt.% H2) 0.9 −1 Specific energy (Wh g ) 0.133 Volume of tank (approximately, L) 0.033 Mass of H2 (kg L−1) 0.03 Proponents of hydrogen storage via a reversible metal hydride note that it is safer than a pressurized cylinder that could rapidly and dangerously discharge in the event of failure or damage. If a leak develops in a hydride vessel, the temperature of the hydride will fall, which will inhibit the release of the gas. It should, however, be cautioned that there could be ingress of air that could lead to combustion of the hydride powder. When coupled with a fuel cell, the low pressure required by hydrides helps to simplify the design of the hydrogen supply system. Consequently, hydrides are attractive for a very wide range of applications where small quantities of hydrogen are stored. The materials are also particularly suited to applications where space rather than weight is a problem. In fuel‐cell powered boats, for instance, the hydride‐storage vessel can be located near the bottom of the hull where additional weight is often an advantage but where space is at a premium. The disadvantages of reversible hydrides are particularly noticeable when larger quantities of hydrogen are to be stored, for example, in vehicles. The specific energy is poor. Also, the problem of heating during filling and cooling during the release of hydrogen becomes more acute as the amount of hydride increases. Consequently, the reactant bed should have a high thermal conductivity. Large systems have been tested for vehicles, and a typical refill time is about 1 h for a tank of approximately 5 kg. As noted previously, the other main disadvantage of metal hydrides is that the hydrogen must be of very high purity. 11.6 Simple Hydrogen‐Bearing Chemicals 11.6.1 Organic Chemicals Certain organic chemicals contain significant atomic proportions of hydrogen that can be recovered and therefore may be considered as prospective hydrogen‐storage materials. Cyclohexane (C6H12), for example, has been proposed as it is easily decomposed catalytically into hydrogen and benzene (C6H6) according to: C 6 H12 C 6 H6 3H2 (11.3) Though, in theory, this reaction can be carried out in the gas phase cleanly at a moderate temperature, it is normally achieved over a catalyst at 500–600°C. The yield is high, Hydrogen Storage but there is a risk of cracking to give unwanted by‐products and the nature of the products is dependent on the catalyst. By contrast, the reverse reaction (i.e., the reaction between benzene and hydrogen) readily occurs at quite modest temperatures (150–200°C) either in the liquid or vapour phase over a platinum catalyst. Both benzene and cyclohexane are liquids at ambient temperature and pressure. The amount of hydrogen stored by the conversion of cyclohexane to benzene would be 7.1 wt.% and therefore has received serious attention from researchers, as has methylcyclohexane (C7H8) and decalin (C10H18) with a yield of 6.1 and 7.2 wt.% H2, respectively. Until recently, the high temperatures involved, control of the reactions and the risk of by‐product formation have generally ruled out such materials for on‐board vehicles. Nonetheless, a modified form of cyclohexane — bis‐BN cyclohexane (C2B2N2H12) — may offer better prospects. This material, which appears to be remarkably stable up to 150°C and has a storage capacity of 4.7 wt.% H2,6 has been found to decompose easily at room temperature over a catalyst to yield hydrogen and no detectable by‐products. Heterocyclic compounds, principally n‐ethylcarbazole, and dibenzyl toluene, have been investigated as storage materials. Although the potential storage efficiency for such compounds is high, there are a myriad of issues in terms of their practical application, such as low reaction kinetics, toxicity and difficulty in efficiently reversing the dehydrogenation. In fact, reversing dehydrogenation has ruled out several organic chemicals. For example, formic acid stores 4.3 wt.% H2 and can be decomposed over a catalyst into hydrogen and carbon monoxide (CO), but reversing the process, i.e., generating formic acid from hydrogen and CO or CO2 is not a trivial process. 11.6.2 Alkali Metal Hydrides A calcium hydride system for hydrogen production has been proposed.7 The reaction is as follows: CaH2 2H2O Ca OH 2 2H 2 (11.4) It could be said that the hydrogen is being released from the water by the hydride. Both sodium and lithium hydrides also react with water to release hydrogen and were investigated as storage materials in the late 1990s with support from the US Department of Energy. In each case, the alkali metal hydride is highly reactive and has to be protected from accidental contact with water from the atmosphere. Accordingly, one proposal was to encase pellets of sodium hydride in polyethylene and then cut them open under water. Another approach was to slurry lithium hydride with an organic material such as a light mineral oil. The hydrogen content of this lithium hydride is the highest for any hydride and is three times that of sodium hydride. 6 Chen, G, Zakharov, LN, Bowden, ME, Karkamkar, AJ, Whittemore, SM, Garner, EB, Mikulas, TC, Dixon, DA, Autrey, T and Liu, S‐Y, 2015, Bis‐BN cyclohexane: a remarkably kinetically stable chemical hydrogen storage material, Journal of the American Chemical Society, vol. 137(1), pp. 134–137. 7 Bossel, UG, 1999, Portable fuel cell battery charger with integrated hydrogen generator, Proceedings of the European Fuel Cell Forum Portable Fuel Cells conference, 21–24 June 1999, Lucerne, pp. 79–84. 339 340 Fuel Cell Systems Explained Unfortunately, however, no commercial products based on alkali metal hydrides have emerged, principally on account of the difficulty and costs involved in producing the hydride and in recycling the spent material. 11.6.3 Ammonia, Amines and Ammonia Borane Ammonia has often been proposed as a means of hydrogen distribution and storage. The molecular formula is NH3, which immediately indicates its potential as a hydrogen carrier (it contains 17.7 wt.% of accessible H2). Under normal conditions, ammonia is a highly toxic, colourless gas with a pungent choking smell that is easy to recognize. It is produced in huge quantities — currently, at around 100 million tonnes per annum (including a little over 16 million in the United States alone). The manufacture of fertilizer is the most important of the many uses of ammonia in the chemical industry and accounts for about 80% of its consumption. Ammonia is a liquid at room temperature under a few atmospheres pressure and is therefore attractive for transport in cylinders and pipelines in the same way as liquid petroleum gas or propane is distributed today. The recovery of hydrogen from ammonia involves simple dissociation, i.e., 2NH3 N 2 3H2 H 46.4 kJmol 1 (11.5) For this reaction to occur at a practical rate, the ammonia has to be heated to between 600 and 800°C and passed over a catalyst. Higher temperatures are necessary if the output from the converter is to have residual levels of ammonia at the ppm level as required by PEMFC or phosphoric acid fuel cell (PAFC) systems. Note that the dissociation is endothermic, and if the ammonia is supplied initially as liquid, a significant amount of energy is also required to vapourize it into a gas (+ΔH = 23.3 kJ mol−1), which is why it is employed as a refrigerant. If the molar specific heat of ammonia is taken to be 36.4 kJ mol−1, then the heat required to raise the temperature from 0 to 800°C for reaction (11.5) is: H 800 36.4 29.1kJmol 1 (11.6) The process results in the production of 1.5 mol of hydrogen for every mole of NH3, for which the molar enthalpy of formation (HHV) is −285.84 kJ mol−1. The best possible process efficiency for the decomposition of ammonia to hydrogen would therefore be: 285.4 1.5 23.3 29.1 46.4 285.4 1.5 0.77(or 77%) (11.7) This should be regarded as the upper limit to efficiency for using ammonia as a storage medium as it does not take into account the production of ammonia from hydrogen and nitrogen. Also, the heat required to raise ammonia to the high temperature for reaction (11.5) and the heat of reaction itself are likely to prove challenging for low‐temperature fuel‐cell systems. A potential game changer for ammonia as a hydrogen provider is an alternative ‘cracking’ pathway that can be conducted at a lower temperature and, moreover, not with a precious metal catalyst but with the bulk chemical sodium amide (NaNH2) effectively Hydrogen Storage acting as catalyst. Recently, researchers at Oxford, United Kingdom, have demonstrated this method by powering a small PEMFC.8 The reactions involved are: NaNH2 Na s Na s NH3 g 1 N2 g 2 NaNH s H2 g 1 H2 g 2 (11.8) (11.9) (s, solid state; g, gaseous state) Essentially, an alkaline imide (e.g., NaNH2) decomposes to yield the amide (e.g., NaNH) and hydrogen. The ammonia is a ‘mediator’ in the two reactions. Subsequent experiments have shown that imide‐forming amides are highly active, with the lithium amide–imide system exhibiting superior activity per unit mass than the sodium counterpart. Two other issues arise when decomposing ammonia. The first is that nitrogen formed by the decomposition reaction will need to be separated from the hydrogen. Otherwise, the nitrogen will act as a diluent in the fuel cell and cause a loss of system efficiency, as mentioned in Section 10.5, Chapter 10. The second problem is that if a PEMFC or a PAFC is being operated, then any ammonia remaining in the product gas can potentially react with the acid electrolyte in these fuel cells and thereby lead to eventual failure of the system. One method of overcoming some of the difficulties of using ammonia in liquid form is to convert it into an ammine such as Mg(NH3)6Cl2, which is an inert solid that holds 51 wt.% NH3 and thus 9.1 wt.% H2. This compound is safe to handle and can be compacted into dense tablets that have a low vapour pressure of ammonia at ambient temperature and, on a volumetric basis, contain 60–70% more hydrogen than LH2. In contrast to ammonia, hydrazine hydrate (N2H4∙H2O), which has a recoverable hydrogen content of 80 wt.%, is an endothermic compound and therefore easier to decompose than ammonia. Indeed, it has been known to decompose explosively. Although difficult to manufacture or handle in bulk, hydrazine has served as a rocket fuel and, many years ago, was tested in an experimental alkaline fuel cell. Ammonia borane (NH3BH3) has long been promoted as a hydrogen carrier and storage medium. The molecule of ammonia borane is similar in chemical character to that of ethane (CH3CH3) but is a solid rather than a gas. On decomposition at 100–200°C, it yields up to 12 wt.% H2 and thereby forms polymeric iminoborane (NHBH)n that can, in principle, be reconverted to ammonia borane. In practice, however, the form of the polymer is highly dependent on the decomposition conditions and the choice of catalyst so that the reconversion process presents a major chemical challenge.9 11.7 Complex Chemical Hydrides There is a class of inorganic metal hydrides that are ionic rather than metallic in nature. For instance, the elements boron and aluminium form the hydride ions [BH4]− and [AlBH4]−, respectively. When combined with alkali metal cations, soluble ionic salts are 8 Hunter, H, Makepeace, J, Wood, T, Kibble, M, Nutter, J, Jones, M and David, B, 2015, Demonstrating hydrogen production from ammonia — powering a 100 W PEM fuel cell, Proceedings of the World Hydrogen Technology Convention, 11–14 October 2015, Sydney. 9 Peng, B and Chen J, 2008, Ammonia borane as an efficient and lightweight hydrogen storage medium, Energy and Environmental Science, vol. 1, pp. 479–483. 341 342 Fuel Cell Systems Explained formed, e.g., LiBH4, LiAlH4, NaBH4 and NaAlBH4. These compounds are generally known as ‘complex chemical hydrides’. Lithium and sodium borohydrides and the corresponding alumino hydrides are used in organic chemistry as reducing agents. 11.7.1 Alanates For hydrogen storage, the aluminohydrides (the so‐called alanates) are generally preferred to the borohydrides. Thermal decomposition of the alanate NaAlH4 takes place in two steps as follows: 3NaAlH 4 Na 3 AlH6 2Al 3H2 Na 3 AlH6 3NaH Al 3 H2 2 (11.10) (11.11) The reactions for the pure compounds are reversible, but only at temperatures above the melting point of NaAlH4 (183°C) and at hydrogen pressures of 10–40 MPa, which are impracticable. Fortunately, the temperatures for discharge and recharge of hydrogen may be reduced significantly by the inclusion of a titanium catalyst in the alanate. Titanium‐catalysed NaAlH4 has thermodynamic properties that are comparable with those of the classical low‐temperature hydrides, e.g., LaNi5H6 and TiFeH (see Figure 11.6). Reaction (11.10) is carried out at 50–100°C and corresponds to the release of 3.7 wt.% H2 and the second step, reaction (11.11), at 130–180°C yields a further 1.9 wt.% H2. Moreover, even if only the first reaction step can be utilized, the gravimetric hydrogen‐storage density of NaAlH4 is greater than that offered by most of the simple metal hydrides. By contrast, since the Na3AlH6 requires higher temperatures for hydrogen liberation, it might prove suitable for applications other than fuel cells, such as heat pumps and heat storage. The study of NaAlH4 is still at a preliminary stage, and there is considerable scope for the development of alternative catalysts and the optimization of their performance. Two of the main problems to be addressed with these materials are that they are pyrophoric (i.e., their propensity to ignite spontaneously) and they are costly to produce. 11.7.2 Borohydrides In recent years, developers of fuel cells have shown considerable interest in sodium tetrahydroborate, which is more commonly referred to as sodium borohydride, NaBH4. This compound was introduced in Section 5.2.3, Chapter 5, as a potential fuel for AFCs and again in Section 6.5, Chapter 6, as the fuel for borohydride fuel cells. As a general rule, the alkali metal borohydrides contain more hydrogen than the alanates, e.g., 18.5 and 10.6 wt.% H2 for the respective lithium (LiBH4) and sodium (NaBH4) analogues, but are more stable and therefore less practical as accessible hydrogen carriers. The compounds only decompose at relatively high temperatures (LiBH4 above 300°C and NaBH4 above 350°C), but it may be possible to lower these temperatures by incorporating catalysts in the materials. Sodium borohydride can be supplied as a solid, in which case it is often mixed with cobalt chloride, which acts as a desiccant. The material is hazardous and can spontaneously Hydrogen Storage give off hydrogen if it accidentally comes into contact with water. It is, however, able to dissolve without reaction in aqueous alkaline solutions (e.g., sodium hydroxide) and in this form is stable for long periods. Given the limitations with the decomposition of the solid form, i.e., the high temperature requirement and use of catalyst, sodium borohydride has attracted the most attention through its reaction with water to form hydrogen, as expressed by: 3NaBH 4 2H2O NaBO2 4 H2 H 218kJmol 1 (11.12) The reaction is not reversible but has the advantage that 50% of the hydrogen comes from the water — in effect, NaBH4 is a ‘water‐splitting’ agent. Given that alkaline solutions of NaBH4 in water are quite stable, a catalyst is usually required to promote the decomposition, and consequently the generation of hydrogen is quite controllable. Notable features of reaction (11.12) are as follows: ● ● ● It is exothermic, at a rate of 54.5 kJ mol−1 of hydrogen. Hydrogen is the only gas produced, i.e., no gas separation is necessary as may the case, for example, with ammonia or methanol decomposition. If the water is heated, then water vapour will be mixed with the hydrogen, which is a desirable feature for PEMFC systems. Weaker solutions of sodium borohydride are more stable than strong solutions, but their effectiveness as a hydrogen carrier diminishes. Such solutions of borohydride are also stable for long periods, though hydrogen evolution does occur slowly. The ‘half‐life’ of such solutions has empirically been shown to follow the relationship: log10 (t 1 2 ) pH 0.34T 192 (11.13) where the half‐life (t½) is in minutes and the temperature T is in Kelvin. A solution of 30 wt.% NaBH4 + 3 wt.% NaOH has a half‐life of about 2 years at 20°C. One litre of such a solution will yield 67 g of hydrogen on hydrolysis, which is a better yield than that obtained from any of the practical metal hydrides. Another approach to preparing a borohydride medium for hydrogen storage is to mix dry solid NaBH4 powder with light mineral oil and a dispersant to produce an ‘organic slurry’. The oil coats the solid particles and protects them from inadvertent contact with water during handling and transport and also moderates the decomposition rate of the hydride when water is introduced. More development work is required to control the reaction kinetics so as to yield hydrogen at the desired rate. In addition to the previous list of advantages afforded by reaction (11.12), the following benefits of the NaBH4 solution should be recognized: ● ● ● ● It is arguably the safest of all hydrogen‐containing liquids to transport. Apart from cryogenic hydrogen, it is the only liquid that gives pure hydrogen as the product. The reactor for the release of hydrogen requires no energy input and can operate at ambient temperature and pressure. The rate of hydrogen production can be simply controlled. Unfortunately, however, there are two significant disadvantages that arise in employing borohydrides for hydrogen storage. The first is that the product — sodium borate — cannot 343 344 Fuel Cell Systems Explained be reused in situ, which means that it has to be replaced with fresh material once all of the hydrogen has been released. It has been shown that this procedure on a vehicle could be quite rapid with the borate transported to a processing plant for regeneration. Nevertheless, the regeneration process is not only expensive but also requires a large amount of energy and, therefore, constitutes the major obstacle to the widespread uptake of borohydrides. It would appear that the cost of the regeneration has to fall by a factor of approximately 50 for NaBH4 to be acceptable as a fuel for vehicles. Nonetheless, from the standpoint of both mass and volume, the system appears superficially satisfactory as a hydrogen‐storage scheme for FCVs. DaimlerChrysler demonstrated that a NaBH4 system, designed by Millennium Cell in the United States, could provide a minivan (the Natrium) with a range of 480 km. Despite this performance, however, Millennium Cell ceased development of the NaBH4 system in 2008. In the past few years, the challenge of using borohydrides has been taken up by Cella Energy Ltd in the United Kingdom. The hydrogen‐storage material developed by this company is based on a proprietary complex chemical hydride that is formed into pellets with a polymer admixture that provide mechanical integrity to the material. The pellets are coated with a thin (<50 µm) film of a common hydrogen‐permeable or permselective polymer, e.g., polyethylene or polymethyl methacrylate. The protective film allows only hydrogen to diffuse out of the material once it is heated. The Cella Energy material can be easily transported and dispensed into vehicle fuel tanks. The protective polymer film also prevents ingress of oxygen or water and thereby gives the material a long shelf-life. 11.8 Nanostructured Materials In 1998, a study was published on the absorption of hydrogen in carbon nanofibres.10 The authors presented results that suggested that these materials could absorb in excess of 67 wt.% H2. This amazingly large amount took the academic world by storm, as it offered the prospects of levels of hydrogen storage in material at ambient temperatures and pressures that would ensure hydrogen a certain future for vehicles. The initial euphoria was tempered when other research groups tried to repeat the findings and found that there were measurement errors due to the presence of metal contaminants and/or water absorption. Nevertheless, so‐called ‘nanostructured’ materials continue to be evaluated for hydrogen storage and other catalytic properties. Nanotechnology has become a broad area of applied science that focuses on the control and exploitation of materials with characteristic geometric dimensions below 10−7 m and new properties that result from the nanostructure. It draws on fields as diverse as colloid science, device physics, molecular biology and supramolecular chemistry to address a wide range of potential applications. By virtue of their large surface‐to‐volume ratios, certain nanomaterials can adsorb considerable amounts of hydrogen in the molecular state via weak molecule–surface interactions (so‐called physisorption). This is in contrast to the chemisorption process on metal and complex 10 Chambers, A, Park, C, Baker, RTK and Rodriguez, NM, 1998, Hydrogen storage in graphite nanofibre, Journal of Physical Chemistry B, vol. 102, pp. 4253–4256. Hydrogen Storage hydrides in which the hydrogen is dissociated into atoms that chemically bond with the lattice of the storage medium. Obviously, physisorption is preferred as it would moderate the temperature and pressure required for the respective uptake and release of hydrogen. Moreover, there is no major heat‐transfer problem because physisorption bonds are weak, typically with enthalpies of adsorption from −10 to −20 kJ mol−1 H2. On the other hand, significant uptake of hydrogen is normally seen only at cryogenic temperatures, and this is a major inconvenience. Particular attention has been paid to the possibility of hydrogen storage in carbon‐ based materials that take the form of nanofibres or nanotubes. These are structures that derive from the fundamental carbon entity C60 (buckminsterfullerene) that has a spherical cage‐like structure that is made up of hexagons and pentagons, as shown schematically in Figure 11.8a.11 The generic term ‘fullerene’ is used to describe a pure carbon molecule that consists of an empty cage of 60 or more carbon atoms. Graphitic nanofibres are prepared by the decomposition of hydrocarbons or carbon monoxide over metal catalyst and are made up of graphene sheets aligned in a particular direction as determined by the choice of catalyst. Three distinct structures may be formed: platelet, ribbon and herringbone (see Figure 11.8b). The structures are flexible and can expand to accommodate the hydrogen. Graphitic nanofibres vary from around 50 to 1000 nm in length and 5 to 100 nm in diameter. Carbon nanotubes are cylindrical or toroidal varieties of fullerene and have lengths of between 10 and 100 µm. ‘Single‐walled’ nanotubes are composed of only one graphene sheet and have typical diameters of up to 2 nm. ‘Multiwalled’ nanotubes consist of concentric rings (diameters of 30–50 nm) or spirals of graphene sheets; different types of nanotube are illustrated in Figure 11.8c. Carbon nanotubes were first identified by Iijkima in 1991 who obtained them by accident when using an electric arc drawn between two carbon electrodes. Nowadays, laser ablation and chemical vapour deposition, which are more controlled procedures, are employed. Unfortunately, such high‐tech methods of preparation mean that carbon nanotubes are expensive materials. Since the first edition of this textbook in 2001, many groups have investigated carbon nanofibres and nanotubes with mixed results. One of the problems is the difficulty in determining hydrogen uptake experimentally on small amounts of materials, largely due to inaccuracies in measuring sample volumes.12 Whereas physisorption has been clearly demonstrated, useful levels of hydrogen uptake (up to about 3 wt.%) have only been achieved at cryogenic temperatures with materials of very high surface area. Some workers claim that uptakes of about 1 wt.% H2 can be achieved on materials doped with titanium or platinum, but long‐term consistent performance has yet to be achieved for carbon nanofibres and nanotubes. Molecular modelling has suggested that graphene sheets, perhaps spaced wider than in graphite and suitably functionalized, could provide the required characteristics to meet the targets for hydrogen storage at ambient conditions, but this prediction has not been verified experimentally.13 11 In theory, a molecule of buckminsterfullerene can be hydrogenated up to C60H60, which corresponds to an absorption of almost 7.7 wt.% H2. In the United States, MER Corporation has demonstrated absorption up to 6.7 wt.% H2 on fullerenes, but the absorption is very slow and to date no practical application has been found. 12 Webb, CJ and Gray, EMacA, 2014, The effect of inaccurate volume calibrations on hydrogen uptake measured by the Sieverts method, International Journal of Hydrogen Energy, vol. 39, pp. 2168–2174. 13 Tozzini, V and Pellegrini, V, 2013, Prospects for hydrogen storage in graphene, Physical Chemistry Chemical Physics, vol. 15, pp. 80–89. 345 346 Fuel Cell Systems Explained (a) (c) (b) Platelet Ribbon Herringbone (d) Figure 11.8 Schematic representations of (a) buckminsterfullerene, (b) carbon nanofibres, (c) single‐ and multiwalled carbon nanotubes and (d) metal oxide framework. In parallel with research on carbon, many other porous materials and composites with high surface areas are being investigated as possible storage media. These include the following: ● ● Zeolites: Complex aluminosilicates with engineered pore sizes and high surface areas. Metal–organic frameworks (MOFs): Typically zinc oxide structures bridged with benzene rings; an example is illustrated schematically in Figure 11.8d. These materials have extremely high surface areas, are highly versatile and allow for many Hydrogen Storage ● structural modifications. As with carbon structures, MOFs tend to absorb significant quantities of hydrogen only at high pressure and cryogenic temperatures, typically 77 K. Clathrate hydrates: Water (ice) cage‐like structures that often contain guest molecules such as methane and carbon dioxide (see Section 10.2.2, Chapter 10). Hydrogen caged in a clathrate hydrate was first reported in 2002 and requires very high pressures to be stable. In 2004, researchers showed solid H2‐containing hydrates could be formed at ambient temperature and high pressure by adding small amounts of promoting substances such as tetrahydrofuran.14 These clathrates have a theoretical maximum hydrogen absorption of around 5 wt.% and 40 kg m−3. 11.9 Evaluation of Hydrogen Storage Methods Fuel‐cell vehicles are the most demanding application for hydrogen storage. At present, the only practical means of storing hydrogen on a road vehicle is as a high‐ pressure gas in a cylinder. As discussed previously, modern lightweight composite cylinders are a distinct improvement on conventional steel vessels. Other options such as cryogenic liquid hydrogen, organic chemicals or metal hydrides are likely to be uneconomic, at least in the short-term. Irreversible chemical hydrides that react with water may find some success, but manufacturing and recycle costs need to be reduced dramatically. To this end, further investigation is necessary both at the fundamental scientific level and in terms of vehicle engineering, hydrogen recovery plant and overall logistics. The task of displacing liquid hydrocarbon fuels by hydrogen is formidable. This is illustrated pictorially in Figure 11.9 and numerically in Table 11.7. The latter presents a comparison of the probable masses and volumes of systems to accommodate the amount of hydrogen (11.75 kg) that is equivalent in energy content (1.4 GJ) to 45 L of gasoline, which is typically the capacity of the fuel tank in a car. Columns 3 and 5 relate these masses and volumes, respectively, with gasoline taken as unity. In the case of methanol, the mass and volume of an on‐board reformer is also taken into account. The data may only be approximate, because of the complexity of making due allowance for the containers and ancillary equipment, but the information does give some indication of the magnitude of the problem. It is seen that gasoline is well ahead of the various options for hydrogen storage, both in terms of mass and volume. The analysis also reveals why on‐board methanol reforming has been seriously considered for FCVs and why borohydride solutions would be attractive were it not for the difficulty in handling the material and the cost of regeneration. In this context, it should be pointed out that the tank‐to‐wheel efficiency (see Section 12.4.1, Chapter 12) of a fuel‐cell car is significantly better than that of either a conventional gasoline or diesel engine.15 14 Florusse, LJ, Peters, CJ, Schoonman, J, Hester, KC, Koh, CA, Dec, SF, Marsh, KN and Sloan, ED, 2004, Stable low‐pressure hydrogen clusters stored in a binary clathrate hydrate, Science, vol. 306(5695), pp. 469–471. 15 Davis, C, Edelstein, W, Evenson, W, Brecher, A and Cox, D, 2003, Hydrogen Fuel Cell Vehicle Study, A Report Prepared for the Panel on Public Affairs (POPA), American Physical Society, College Park, MD. 347 348 Fuel Cell Systems Explained Figure 11.9 Schematic representation of the relative volumes of two hydrides, liquid hydrogen and compressed gas (20 MPa pressure) to contain 4 kg of hydrogen for a vehicle driving range of 400 km. Table 11.7 Approximate comparison of mass and volume of gasoline and stored forms of hydrogen for equivalent energy content (1.4 GJ). Mass of store and fuel (kg) Gasoline (45 L) Index Volume of store (L) Index 41 1 45 1 Compressed H2 (20 MPa) conventional steel cylinder ~1150 28 ~1080 24 Compressed H2 (70 MPa) composite cylinder ~200 4.9 ~170 3.8 LH2 in cryostat ~100 2.4 ~350 7.8 Ti–Fe hydride bed ~1050 25.6 ~275 ~6.1 Methanol and reformer ~1140 27.7 ~294 ~6.5 NaBH4 solution ~553 13.5 192 ~4.3 In conclusion, mention should be made of energy stored in rechargeable batteries and how this compares with energy stored as hydrogen. As an example, consider the small metal hydride container shown in Figure 11.10, which holds 1.7 g of hydrogen. Taking the data given in Section A2.4, Appendix 2, for a PEMFC operating at Hydrogen Storage Figure 11.10 Small metal hydride hydrogen store for fuel cells used with small portable electronic equipment. Vc = 0.6 V, which corresponds to an efficiency of 40% (HHV), the hydride container will deliver 26.8 0.6 1.7 27 Wh of electrical energy (the effective specific electrical energy of hydrogen given in Table A2.1 is 26.8 × Vc kWh kg−1). This is approximately the same as the capacity of six D‐size nickel–cadmium cells, each with roughly the same volume as the hydrogen store of Figure 11.10. In other words, the stored hydrogen has obtained approximately six times the energy density (Wh L−1) of nickel–cadmium batteries. The most advanced lithium‐ion batteries exhibit a specific energy of around 900 kJ kg−1. Hydrogen itself has a specific energy of 142 MJ kg−1, which would indicate that hydrogen offers an amazingly better energy storage option. When the weight of the storage vessel is taken into account the figures are less compelling. Manipulation of the data in Table 11.7 will show that compressed hydrogen in a conventional cylinder at 20 MPa has a specific energy of some 1.2 MJ kg−1. If the storage pressure is raised to 70 MPa, the specific energy is increased to 7 MJ kg−1. This is still eight times the specific energy of the lithium‐ion battery and is one of the reasons why hydrogen is such an attractive means of storing electrical energy on vehicles. Further consideration shows that the specific energy of hydrogen stored in the form of hydrides is potentially even higher, but, as mentioned earlier, there are other considerations (principally heat management and slow reaction rates) that make hydrides currently a poor choice for storage on vehicles. Fuel processing and hydrogen storage are vital aspects of fuel‐cell system design, but they are not the only subsystems that have to be added to the fuel‐cell stack. The next chapter addresses issues associated with moving the reactant gases through the system. 349 350 Fuel Cell Systems Explained Further Reading Broom, DP, 2011, Hydrogen Storage Materials — The Characterisation of Their Storage Properties (Green Energy and Technology), Springer‐Verlag, London. ISBN: 978‐0‐85729‐220‐9. Gray, EMacA, 2007, Hydrogen storage — status and prospects, Advances in Applied Ceramics, vol. 106(1–2), pp. 25–28. Hirscher, M, ed., 2010, Handbook of Hydrogen Storage: New Materials for Future Energy Storage, Wiley‐VCH Verlag GmbH, Weinheim. ISBN: 978‐3‐527‐62981‐7. Rand, DAJ and Dell, RM, 2008, Hydrogen Energy — Challenges and Prospects, RSC Publishing, Cambridge. ISBN: 978‐0‐85404‐597‐6. Thomas, CE, 2009, Fuel cell and battery electric vehicles compared, International Journal of Hydrogen Energy, vol. 34, pp. 6005–6020. Varin, RA, Czujko, T and Wronski, ZS, 2009, Nanomaterials for Solid State Hydrogen Storage (Fuel Cells and Hydrogen Energy), Springer, New York. ISBN: 978‐0‐387‐77711‐5. Walker, G, ed., 2008, Solid‐State Hydrogen Storage: Materials and Chemistry, Woodhead Publishing, Cambridge. ebook ISBN: 9781845694944. Zhang, JZ, Li, J, Li, Y and Zhao, Y, 2014, Hydrogen Generation, Storage and Utilization, Wiley‐Science Wise Co‐publication, John Wiley & Sons, Inc., New York. ISBN: 978‐1‐118‐14063‐5. 351 12 The Complete System and Its Future It has been said that a fuel‐cell system comprises three elements: (i) the processor for providing the fuel to the stack, (ii) the stack itself and (iii) the power conditioner for converting raw DC power from the stack to a useful AC voltage. The stack is the heart of the system, but many other components are required for it to function in a real‐world application. For instance, balance‐of‐plant (BoP) items are necessary to provide cooling by moving air around the system and to supply oxygen to the cathode. Pumps, fan compressors and blowers are required to deliver this service. In addition, the energy in the stack exhaust gases can sometimes be harnessed instead of simply going to waste. The technology for such mechanical equipment is very mature given its development and long‐term use for other applications. Designers of fuel‐cell systems will therefore choose devices principally custom‐made for products in other markets. ‘Gas‐moving devices’ can vary greatly in size and application; therefore, in considering their suitability for a given fuel‐cell system, it is necessary to examine a wide range of candidate equipment. Consequently, the first section of this chapter will consider the various mechanical BoP items required to move both air and fuel gas, i.e., compressors, turbines, ejectors, fans/ blowers and pumps. Section 12.2.2 then addresses issues associated with electrical components, or how the DC power from a fuel‐cell stack is converted to a more useful AC. Section 12.4.1 covers the integration of batteries with fuel cells — an area of growing importance not only for road vehicles but also for stationary renewable power systems. Section 12.4.2 presents an analysis of fuel‐cell systems and brings together information from all of the previous 11 chapters. The final sections review the commercial status and future prospects of complete fuel‐cell systems. 12.1 Mechanical Balance‐of‐Plant Components 12.1.1 Compressors The principal operation of each of the four main types of compressor that are used in fuel‐cell systems is represented schematically in Figure 12.1. The basic Roots compressor, depicted in Figure 12.1a, is one of the simplest of positive‐displacement pumps and is frequently employed as a supercharger in diesel engines where it is driven directly from the engine crankshaft via a belt, chain or gears. The device operates by pumping a gas with a pair of meshing lobes, not unlike a set of stretched gears. Fluid is trapped Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 352 Fuel Cell Systems Explained (a) (b) (c) (d) Volute Rotor Figure 12.1 Some different types of compressor: (a) Roots positive‐displacement pump, (b) twin‐screw Lysholm, (c) centrifugal (radial) design and (d) axial flow design. in pockets surrounding the lobes and is carried from the intake side to the exhaust. The Roots compressor is quite cheap to produce and works over a wide range of flow rates. However, it only has a modest efficiency when delivering a small pressure lift — typically, about 90% at a pressure ratio (Pout/Pin) of 1.1 — and is therefore normally employed for moving large volumes of gas. The performance can be enhanced by increasing the number of rotors from 2 to 3 or 4, and by twisting the rotors by 60% to form a partial helix. Such improvements, while increasing the cost, also greatly reduce pressure fluctuations in the gas and provide a marginal increase in efficiency. Even so, such compressors are still only suitable for small pressure increases, typically up to a pressure ratio of 1.8. The Lysholm screw compressor of Figure 12.1b has two screws, which counter rotate and thereby drive the gas up through the region between the two screws while compressing it at the same time. The device can be thought of as a refinement of the ‘Archimedes screw’, which has been used for pumping water since ancient times. There are two variations of the twin‐screw compressor. In the first, an external motor drives only one rotor, and the second rotor is turned by the first. The arrangement requires the rotors to be in contact and therefore lubricated with oil. The small amounts of oil that is inevitably carried forward with the air are of no concern for many industrial The Complete System and Its Future (a) (b) Figure 12.2 (a) Typical centrifugal air‐compressor rotor and (b) twin‐screw compressor. applications; the twin‐screw compressor finds wide application in providing compressed air for pneumatic tools. In the second type of twin‐screw compressor, the two rotors are connected by a synchronizing gear — a separate pair of cogs provides the driving link from one rotor to the other. The counter rotating screws do not come into contact, though for a good efficiency they will run very close to each other. This version gives an oil‐free output, which is necessary for a fuel‐cell system; it is also used in other facilities, e.g., for circulating the fluid in refrigeration systems. By changing the length and pitch of the screws, the Lysholm compressor can be designed to cover an extensive number of compression ratios — the exit pressure can be up to eight times the input pressure. Another advantage is that the efficiency remains high over a broad range of flow rates. Unfortunately, however, this type of compressor is expensive to manufacture on account of the precision required by the rotors, synchronizing gears and bearings (there are high lateral and axial mechanical loadings on the axles of screw compressors). One of the more common types of compressor or blower is the centrifugal or radial design illustrated in Figure 12.1c. The gas is drawn in at the centre of the impeller and forced out at high speed to the surrounding volute. Here the kinetic energy is ‘converted’ into a pressure increase. The centrifugal compressor is used in the vast majority of engine turbocharging systems. An example rotor is shown in Figure 12.2a. Although the shape may be complex, it can usually be cast as one piece. This type of compressor is thus low cost, well developed and available to suit a wide range of flow rates. In addition, the efficiency compares well with that of other types of compressor, but operation must kept within well‐defined flow rates and pressure changes to obtain high efficiencies. Indeed, the centrifugal compressor cannot operate at all at low flow rates, as is explained in Section 12.2. Another problem is that the rotor must rotate at a very high speed (80 000 rpm is typical), whereas twin‐screw and Roots compressors (such as shown in Figure 12.2b) are limited to about 15 000 rpm. Care must therefore be taken with the design and lubrication of bearings in a centrifugal compressor. 353 354 Fuel Cell Systems Explained P1 T1 P1 T1 P2 T2 P2 T2 Figure 12.3 Symbols for (a) a compressor and (b) a turbine. The axial flow compressor, as shown in Figure 12.1d, drives the gas by rotating a large number of blades at high speed. It is, in essence, the inverse of the turbine that is commonly operated in thermal power plants. As with turbines, the gap between the ends of the blades and the housing must be as small as possible. This requirement adds considerably to the manufacturing cost. The efficiency is high, but only over a fairly narrow range of flow rates. Experience with diesel systems suggests that the axial flow compressor will be the best option for fuel cells that have an output above a few megawatts and that are only operated between full power and half power at any given time. The rotating vane compressor is used for air management in some industrial operations. The device claims advantages in terms of cost over the screw compressor. Nevertheless, it is unlikely to be used for fuel‐cell systems, since the tips of the rotating vanes must be lubricated with a thin film of oil, and so, even after filtering, there will always be some oil in the output gas that, as noted earlier, is not generally acceptable. The symbol adopted in process flow diagrams (PFDs) for compressors of all types is as shown in Figure 12.3a with the complementary symbol for turbines shown in Figure 12.3b. 12.1.1.1 Efficiency As with the efficiency of a fuel cell, care needs to be taken in defining the efficiency of a compressor. Whenever a gas is compressed, work is done on the gas and so its temperature will rise, unless the compression is enacted very slowly or there is substantial cooling. In a process that is reversible and also adiabatic (no heat loss), it can readily be shown that if the pressure changes from P1 to P2, then the temperature will change from T1 to T2′ according to the following relationship: 1 T2 T1 P2 P1 (12.1) The Complete System and Its Future where γ = CP/CV, i.e., the ratio of the specific heat capacities of the gas at constant pressure (CP) and constant volume (CV). This formula gives the temperature of the gas as a consequence of the change in pressure, as is indicated by the prime (ʹ) on T2′, for an isentropic process, i.e., one in which the entropy of the system remains unchanged. In practice, the new temperature, T2′, will be higher than given by equation (12.1) since some of the motion of the moving blades and vanes will be ineffective in compressing the gas and may be observed as a small increase in gas temperature. Also, some of the gas might ‘churn’ around the compressor and, thereby, become hotter without being compressed. If the actual new temperature is T2, then the ratio between the following two quantities will enable derivation of the compressor efficiency: 1) The actual work done to raise the pressure from P1 to P2. 2) The work that would have been done if the process had been reversible or isentropic — the ‘isentropic work’. To find these two quantities, the following assumptions, which are generally valid, can be made: ● ● ● The heat flow from the compressor is negligible. There is no change in kinetic energy of the gas as it flows into and out of the compressor, or at least any change is negligible. The gas is a ‘perfect gas’ and so the specific heat at constant pressure, CP, is constant. With these assumptions, the work done, W, is simply the change in enthalpy of the gas, namely: W (12.2) C P T2 T1 m where m is the mass of gas compressed. The isentropic work done, W′, is given by: W (12.3) C P T2 T1 m The isentropic efficiency, ηc, is the ratio of these two quantities, W′/W. Hence: isentropic work c C P T2 T1 m C P T2 T1 m real work T2 T1 T2 T1 (12.4) If the isentropic temperature T2′ from equation (12.4) is substituted into equation (12.1), then: c T1 (T2 T1 ) 1 P2 P1 (12.5) 1 Equation (12.5) can also be rearranged to give the change in temperature on compression, as follows: T T2 T1 T1 c 1 P2 P1 1 (12.6) 355 356 Fuel Cell Systems Explained This definition of isentropic efficiency does not consider the work done on the shaft driving the compressor. To include this work, the mechanical efficiency, ηm, should also be considered, as this takes into account the friction in the bearings, or between the rotors and the outer casing (if any). In the case of centrifugal and axial compressors, the mechanical efficiency is very high, typically over 98%, so that the total efficiency, ηT, can reasonably be expressed by: T m (12.7) c The isentropic efficiency, ηc, is the most useful measure of efficiency because it is related directly to the rise in temperature (ΔT), and this can be quite high. For example, using air at 20°C (293 K) for which γ = 1.4, a doubling of the pressure and a typical value for ηc of 0.6, when substituted into equation (12.6), gives: T 293 0.286 2 1 0.6 170 K (12.8) For some fuel cells, the temperature rise is beneficial as it preheats the reactants. On the other hand, for low‐temperature fuel cells, it means that the compressed gas needs cooling. Such coolers, which are located between the compressor and the fuel cell, are often called ‘intercoolers’. 12.1.1.2 Power The power needed to drive a compressor can be readily found from the change in temperature and knowledge of the heat capacity of the gas. Thus: Power C P  T m (12.9) where ṁ is the rate of flow of gas in kg s−1. The temperature difference, ΔT, is given by equation (12.6). Therefore 1 Power C P T1 P2 P1 c 1  m (12.10) In the case of an air compressor, which is a feature of most fuel‐cell systems, the CP for air can be taken to be 1004 J kg−1 K−1 and γ = 1.4, so that power required by the air compressor is given by: Power 1004 T1 c 12.1.1.3 P2 P1 0.286 1  m (12.11) Performance Charts The efficiency and performance of a compressor will depend on many factors that include the following: ● ● Inlet pressure, P1. Outlet pressure, P2. The Complete System and Its Future ● ● ● ● ● Inlet temperature, T1. Gas density, ρ. Gas viscosity, μ. As flow rate, ṁ. Compressor rotor speed, N. With all these variables, to tabulate or draw some kind of chart of compressor performance would clearly be a difficult, if not impossible, task. Thus, it is necessary to eliminate or group together the variables. This exercise is often performed in the following way: ● ● ● The inlet and outlet pressures are combined into one variable, i.e., the pressure ratio P2/P1. For any gas, the density is given by ρ = P/RT and can therefore be ignored as P and T are being considered. The viscosity of the gas, bearing in mind the limited range of gases normally used, can also be ignored. Further simplification is done by a process of dimensional analysis, which can be found in textbooks on turbines and turbochargers. The result is to group together variables in ‘non‐dimensional’ groups. The two groups are:  m T1 P1 N T1 and They are known, respectively, as the ‘mass flow factor’ (MFF) and the ‘rotational speed factor’ (RSF). They are sometimes also referred to as the ‘non‐dimensional mass flow’ and ‘rotational speed’. Charts of the efficiency for different pressure ratios and MFFs and lines of constant RSF are plotted. The chart for a typical twin‐screw compressor (Lysholm) is given in Figure 12.4. The lines of constant efficiency are similar to the contours of a geographical map — instead of indicating hills, however, the lines indicate areas of increasing efficiency of operation. The unit generally used for P1 is the bar and for temperature, the Kelvin. The MFF can be related to the power of a fuel cell as follows. Assuming typical operating conditions for the fuel cell (i.e., air stoichiometry = 2, average voltage = 0.6 V), then, from equation (A2.9) in Appendix 2, the flow rate of the air (air usage) for a 250‐kW fuel cell is: 3.58 10 7 Vc Pe kg s (A2.9) 1 where Pe is the power of the fuel cell (watts), λ is the air stoichiometry and Vc is the cell voltage. Thus:  m 3.58 10 7 2 250000 0.6 0.3 kg s 1 (12.12) If standard conditions are assumed for the air (i.e., P1 = 1 bar, T = 298 K), the MFF is: MFF 0.3 298 1.0 5.18 kg s 1 K1/2 bar 1 (12.13) 357 Fuel Cell Systems Explained Rotor speed factor/rev. min–1 K–1/2 600 η= η=0 .4 4.0 0 .5 400 Pressure ratio 358 800 η= 0.6 1000 η= 0.7 1200 .75 η=0 3.0 2.0 1.0 0 1.0 2.0 3.0 4.0 5.0 Mass flow factor/kg s–1 K1/2 bar–1 Figure 12.4 Performance chart for a typical twin‐screw compressor. A mass flow factor of 5 corresponds to the air needs of a fuel cell with an output of about 250 kW. Consequently, the horizontal x‐axis of Figure 12.4 corresponds to the air flow requirements of fuel cells of power approximately 0–250 kW. Similarly, if the rotor speed factor is 1000, this will correspond to a speed of about 17 300 rpm. The use of these ‘non‐dimensional’ quantities is standard practice in textbooks on compressors and turbines, but do not feature in many manufacturers’ data sheets. In the latter case, standard condition values are applied (P1 = 1.0 bar, T = 298 K), and the MFF is replaced with the mass flow rate or even the volume flow rate, and the RSF is replaced with speed in rev min−1 (rpm). Generally speaking, such charts will give satisfactory results — except in the case of multistage compressors. When gas has been compressed through the first of a series of stages, its temperature and pressure would obviously have changed markedly, and so the MFF will be quite different, even though the actual mass flow is unchanged. The performance chart for a typical centrifugal compressor is shown in Figure 12.5. The chart is different in form from that of a screw‐type or Lysholm compressor. Two points to note are as follows: 1) There are regions of high efficiency, but these are very narrow. The constant efficiency ‘contours’ are very close together when moving across the chart for a given pressure ratio. 2) There is a distinct ‘surge line’ and to the left of this the compressor is unstable and should not be used. As mentioned earlier, the centrifugal compressor works by accelerating the gas out from the centre of the unit. If there is no gas to pump (i.e., the inlet pressure is too low), then the pressurized gas will flow back to the inlet, only to be pumped again. This action will cause the pressure to become erratic and the gas will start to heat up. A centrifugal compressor, therefore, must not be operated in this ‘low flow rate’ region. If the flow rate has to be low, the pressure ratio must also be reduced. The Complete System and Its Future 3.4 6000 K –1/2 act or/ rev .m in –1 η = 0.68 5500 5000 η= 0.7 0 1 lin e 0.7 η= rge 2.2 Su 4500 spe ed f Pressure ratio 2.6 η = 0.65 η = 0.60 η = 0.55 3.0 1.8 Ro tor 4000 1.4 3500 3000 2500 1.0 0 1.0 2.0 3.0 Mass flow factor/kg s–1 4.0 K1/2 bar–1 5.0 ~250 kW Figure 12.5 Performance chart for a typical centrifugal compressor. As a consequence of these two features, it is difficult to maintain constant pressure with centrifugal compressors without compromising efficiency. To achieve optimum performance, the pressure should be allowed to rise and fall as the gas flow rate increases or decreases so as to follow the maximum efficiency regions shown in Figure 12.5. For most applications of centrifugal compressors (e.g., as turbocharger for an internal combustion engine), the issue of variable pressure is not a problem. This is also the case with some proton‐exchange membrane fuel cells (PEMFCs). In many fuel‐cell systems (especially the molten carbonate fuel cell (MCFC)), however, the pressure of oxidant and fuel gas within the stack needs to be closely matched and therefore makes the centrifugal compressor a poor choice. 12.1.1.4 Selection For pressure ratios in the range of 1.4 to around 3, the best option is to employ compressors that are designed for internal combustion engines. For example, the Eaton supercharger shown in Figure 12.6 is produced for large petrol engines with gas flows in the range of 50–100 L s−1. The rate corresponds to a power range of 50–150 kW for the fuel cell.1 1 The power was calculated using the equation (A2.4) derived in Appendix 2 with typical values of stoichiometry (λ = 2) and an average cell voltage of 0.6 V. 359 360 Fuel Cell Systems Explained Figure 12.6 Eaton supercharger. The unit is about 25 cm long and will boost pressure to between 36 and 70 kPa. (Source: Reproduced by kind permission of Eaton Corporation (http://www. enginetechnologyinternational.com/eaton.php).) Pulley Figure 12.7 Lysholm compressor (Model 1200 AX). Air enters a hole (not visible) on the left and exits via the six holes on the visible face. The pulley on the right drives the screws. For higher pressure ratios, a twin‐screw compressor is the first choice for efficiency and flexibility, as is demonstrated by its performance chart given in Figure 12.4. An example of a small Lysholm twin‐screw compressor (Model 1200 AX) is shown in Figure 12.7. This particular unit measures 260 × 176 × 120 mm, weighs just 5 kg and is designed for flow rates up to about 0.12 kg s−1, which corresponds to about 100 kW for a typical fuel cell operating at λ = 2 and an average cell voltage of 0.6 V. For systems of higher power, the fuel‐cell designer can choose from a wide range of commercially available twin‐screw compressors that have been tried and tested over the years. They are produced in large numbers at low cost by several manufacturers for the automotive industry, both for original equipment manufacturers (OEMs) and as after‐market products. Twin‐screw compressors are also frequently used to replace piston The Complete System and Its Future compressors where large volumes of high‐pressure air are needed, either for large industrial applications or to operate high‐power air tools such as jackhammers. It should be noted that most automotive pumps or superchargers are driven by a pulley from the engine crankshaft. In fuel‐cell systems, the compressor will be driven by an electric motor. By controlling the speed of the motor, the gas flow rate and pressure lift can be varied and thus provide another method of optimizing the stack performance according to the load or electrical demand on the system. 12.1.2 Turbines In a bottoming cycle for a solid oxide fuel cell (SOFC) or an MCFC system, a turbine is used to harness the energy in the hot exhaust gas to produce additional power. In some cases, the turbine can also turn a compressor to compress the incoming air or fuel gas. Two types of turbines can be combined with fuel‐cell systems. The first is the centripetal or radial turbine, which is essentially the inverse of the centrifugal compressor discussed earlier. This technology is the preferred choice unless the power involved is greater than about 500 kW, when the axial turbine, which is the standard technology employed in gas and steam turbine power generation sets, may be considered. In both cases, the symbol for the turbine is the inverse of that for the compressor and is shown in Figure 12.3b. It is possible to mount a turbine and compressor side by side on the same shaft. With the surrounding housing common to both, this makes for a very compact and simple unit. Such an arrangement is generally known as a ‘turbocharger’, because its main application is in the supercharging of engines using a turbine driven by the exhaust gases. The efficiency of the turbine is treated in a similar manner to that for compressors, with the same assumptions. If the turbine works isentropically, then the outlet temperature will fall from T1 to T2′, whereas with the compressor: 1 T2 T1 P2 P1 (12.1) In practice, however, because some of the energy will not be transferred to the turbine shaft, but will stay with the gas, the outlet temperature will be higher than T2′. The actual work done will therefore be less than the isentropic work (note: for the compressor it was more). Consequently, the isentropic efficiency of a turbine is defined as: actual work done c (12.14) isentropic work Making the same assumptions about ideal gases, as were made for compressors, equation (12.14) becomes: c T1 T2 T1 T2 T1 T2 1 T1 1 P2 P1 (12.15) 361 362 Fuel Cell Systems Explained By rearranging equation (12.15), the change in the temperature is given by: 1 T T2 T1 c T1 P2 P1 1 (12.16) Note that because P2 < P1, ΔT will always be negative. This expression enables the derivation of a formula for the power available from the turbine. Applying the same reasoning and simplifying the assumptions as undertaken in Section 12.4.2 yields: 1 Power C P  C P cT1 T m P2 P1 1  m (12.17) To obtain the power available to drive an external load, the power should be multiplied by the mechanical efficiency which, as for compressors, should be 0.98 or above. Turbine performance can be represented using charts in exactly the same way as for compressors, except that in the case of turbines the vertical axis becomes P1/P2 instead of P2/P1 and the direction of the RSF lines is completely different. An example of a chart for a radial turbine is presented in Figure 12.8. For any given turbine speed, the mass flow rate rises as the pressure drop increases, as would be expected, but tends towards a maximum value, which is called the ‘choking limit’. Naturally enough, the value of the choking limit depends largely on the diameter of the turbine housing. Worked examples of compressor and turbine calculations are given in Appendix 3. 12.1.3 Ejector Circulators With no moving parts, the ejector is the simplest of all types of pump. In a fuel cell that operates with gaseous fuels stored at pressure, an ejector harnesses the incorporated mechanical energy to circulate the given fuel through the stack. The device is widely chosen to perform this function in hydrogen fuel‐cell systems and some SOFCs. A diagram of a simple ejector is shown in Figure 12.9. A gas or liquid passes through the narrow pipe A and enters the venturi B. It acquires a high velocity at B and hence produces suction in pipe C. The fluid passing through A thus entrains the fluid from C and discharges it at D. The fluid from A, which must be at a higher pressure than that in C, B or D, does not have to be the same as that in pipe C, B or D. Ejectors are frequently found in steam systems, with steam being the fluid passing through the narrow pipe and jet A. Ejectors can also be used to pump air, to maintain a vacuum in the condensers of steam turbines or to pump water into boilers. In a hydrogen fuel‐cell system, an ejector circulates the fuel gas through the stack. Hydrogen is supplied at high pressure at A, and the energy of the expanding gas in the ejector draws in gas from the anode exhaust at C and, together with the fresh hydrogen, sends it on through D to the anode inlet. The pressure difference generated will be sufficient to drive the gas through the cell as well as through any humidification equipment in the case of a PEMFC. The internal diameters of pipes A, C and D and the The Complete System and Its Future η = 0.70 4.0 3.5 2.5 6000 η=0 .65 Rotor speed factor/ rev. min–1 K–½ η = 0. 55 η=0 .60 Pressure ratio 3.0 5000 2.0 4000 1.5 3000 1.0 0 0.25 0.5 0.75 Mass flow factor/kg s–1 K½ 1.0 1.25 bar–1 Figure 12.8 Performance chart for a typical small radial turbine. A B D C Figure 12.9 Diagram of a simple ejector circulation pump. mixing region B to suit the pressure differences and flow rates associated with the required duty can be obtained from chemical engineering reference books. 12.1.4 Fans and Blowers Fans or blowers deliver straightforward air cooling in a wide range of equipment from desktop computers to cars. The axial fan, which is employed for cooling electronic 363 364 Fuel Cell Systems Explained equipment, is an excellent device for moving air but provides only a very small pressure lift. For example, air blown at a rate of about 0.1 kg s−1 by a small axial fan may drop to zero if the back‐pressure rises even to 50 Pa (0.5 cm water pressure). Nevertheless, axial fans have been included in a few open designs of PEMFC. By contrast, the centrifugal fan, which blows air through air‐conditioning systems, provides a somewhat greater pressure lift. This type of fan is not too dissimilar to the centrifugal compressors described earlier, except that it runs at much lower speeds (by a factor of several hundred), has much longer blades and a considerably more open construction. There are various types of centrifugal fan that differ in the design and orientation of the blades. Blowers and fans are normally employed to assist in the removal of heat (cooling) from the gas stream rather than increasing the kinetic energy of the gas. In a cooling system, the effectiveness is a function not only of the power consumed by the fan or blower but also of the design of heat exchanger employed to cool the gas. Rather than using the term efficiency in such cases, it is more practical to define effectiveness as: rate of heat removal kWh h Cooling system effectiveness 1 (12.18) electrical power consumed (kW) As an example, consider a small axial fan with a 120‐mm diameter that is often used to cool electrical equipment. Such a fan might move air at a rate of 0.084 kg s−1 and consume 15 W of electrical power. If the air it blows rises in temperature by 10°C, then the rate of removal of heat will be given by:  1004 10 0.084 843 W Power C P DT m (12.19) That is, 843 W of heat is removed for just 15 W of electrical power, so the cooling system effectiveness is 843/15 = 56. There is always a balance, or trade‐off, in cooling systems between the flow rate of air and the electrical power consumed. Higher flow rates improve heat transfer, but at the expense of more power consumed by the fan. 12.1.5 Pumps The blowers and fans considered earlier are best suited to high flow rates of gas with very small pressure lift. With small‐to‐medium‐sized PEMFC stacks of, say, 200 W to 2 kW, the back‐pressure on both the air and fuel is too high for blowers and fans. On the other hand, the pressures and flow rates will be too low for any of the commercial compressors that are discussed earlier. For small PEMFC systems another type of pump is required, the main features of which should include the following: ● ● ● Low cost. Silent. Reliable long‐term operation. The Complete System and Its Future Figure 12.10 Diagram of a diaphragm pump — a device with the advantages of low‐cost, quiet operation and long‐term reliability. Soft rubber diaphragm Soft rubber valves ● ● Available in a range of sizes that will also accommodate small‐scale PEMFC systems, e.g., gas flow rates from about 2.5 × 10−4 to 2.5 × 10−3 kg s−1, i.e., 12–120 standard litre per minute (SLM).2 Efficient, low‐power consumption. The most suitable pumps that address these requirements are either small vane or diaphragm designs. The diaphragm pump, as commonly employed in fish tanks, meets most of the earlier requirements. There are many variations in design, but the basic operating principle can be understood from Figure 12.10. The diaphragm is moved up and down by an electrical motor to shift the air through the system by way of two valves in an obvious mechanical manner. The diaphragm is made from soft rubber and thereby provides quiet operation over long periods, although the pressure lift is limited to 10–20 kPa. The flow rate can be controlled easily by modulating the force applied by the pump actuator. For instance, it has been reported that a 300‐W PEMFC has employed a diaphragm pump to supply a required reactant air flow of 10–20 SLM. This flow rate is delivered at a pressure of between 110 and 115 kPa, and the motor–pump combination consumes between 14 and 19 W. The parasitic system power loss due to the air pump is thus about 6%, which is acceptable for such a small system. 12.2 Power Electronics The electrical output of a stack usually needs to be conditioned to match the demands of the particular application. Some operations require a constant or near‐constant voltage and others a conversion of the DC output to AC. Power electronic components 2 The standard litre per minute (SLM) is a unit of volumetric flow rate of a gas corrected to ‘standardized’ conditions of temperature and pressure, which can vary considerably between different fields of science and engineering. 365 Fuel Cell Systems Explained can be added to achieve these requirements. A voltage regulator is used to stabilize the DC voltage, and an inverter is employed to convert DC into AC. These items of power electronics are described in the following subsections. 12.2.1 DC Regulators (Converters) and Electronic Switches As shown in Figures 3.1 and 3.2 in Chapter 3, the voltage from a fuel cell falls with increasing current density. By way of example, data from a 250‐kW PEM fuel‐cell system that powered a bus3 are presented in Figure 12.11. The stack voltage varied from about 400 to over 750 V during operation and had different values at the same current. The variability arises because, as well as current, the voltage is also dependent on many other factors such as operating temperature and air pressure. Such behaviour is not compatible with most electronic and electrical devices as they normally require a power supply with a fairly constant voltage. In most fuel‐cell systems, it is therefore necessary to stabilize the voltage supplied from the stack to the electrical equipment, either by dropping the voltage down to a fixed value below the operating range of the fuel cell or by boosting it up to a fixed value above the operating range. Changes in DC voltage are achieved by using ‘switching’ or ‘chopping’ circuits, which are described in the following text. These circuits, as well as the inverters discussed in Subsection 12.2.4.2, operate with electronic switches. Although the particular type of electronic switch is not of great concern, it is useful to outline the main characteristics 800 Depending on conditions, the voltage/current points can be anywhere in this region 600 Stack voltage (V) 366 400 200 50 100 150 200 Stack current (A) Figure 12.11 Data from a 250‐kW fuel‐cell system designed to power a bus. (Source: Derived from data in Spiegel, RJ, Gilchrist, T and House, DE, 1999, Fuel cell bus operation at high altitude, Proceedings of the Institution of Mechanical Engineers, Part A, vol. 213, pp. 57–68. Reproduced with permission of Sage Publications.) 3 From: Spiegel, RJ, Gilchrist, T and House, DE, 1999, Fuel cell bus operation at high altitude, Proceedings of the Institution of Mechanical Engineers, Part A, vol. 213, pp. 57–68. The Complete System and Its Future Table 12.1 Key data for the main types of electronic switch used in power electronics. Type Thyristor MOSFET Symbol IGBT d c g g s e Maximum voltage (V) 4500 1000 1700 Maximum current (A) 4000 50 600 Switching time (µs) 10–0.5 0.3–0.5 1–4 IGBT, insulated‐gate bipolar transistor; MOSFET, metal-oxide-semiconductor field‐effect transistor. of those that are most commonly employed, as given in Table 12.1, and to discuss their respective advantages and disadvantages. The metal-oxide-semiconductor field‐effect transistor (MOSFET) is a solid‐state switch that is turned on by applying a voltage, usually between 5 and 10 V, to the gate ‘g’ shown in symbol displayed in Table 12.1. When in the ‘on’ or ‘closed’ state, the resistance between the drain ‘d’ and the source ‘s’ is very low. The power required to ensure a very low resistance is small, as the gate current is low. The gate does, however, have a considerable capacitance, so special drive circuits are usually required. The current path behaves like a resistor, with a certain built‐in value of the transistor’s internal resistance, RDS,on, in the ‘on’ state. In voltage regulation circuits, the value of RDS,on for a MOSFET can be as low as 0.01 Ω. Such low values are only possible with devices that can switch low voltages, in the region of up to 50 V. Devices that can switch higher voltages have RDS,on values of about 0.1 Ω that lead to higher power losses. Consequently, MOSFETS are generally employed in low‐voltage power electronics systems of less than about 1 kW. The integrated gate bipolar transistor (IGBT) is essentially a three‐terminal integrated circuit that combines a conventional bipolar transistor and a MOSFET — therefore, it has the advantages of both. A low voltage with negligible current is applied to the gate to switch it to the ‘on’ state. The main current flow is from the collector ‘c’ to the emitter ‘e’, as indicated by the symbol in Table 12.1, and this path has the characteristics of a p–n junction. The result is that the voltage does not increase much above 0.6 V at all currents within the rating of the device. This feature makes the IGBT the preferred choice for systems in which the current is greater than about 50 A. The IGBT can also be made to withstand higher voltages. The longer switching times compared with the MOSFET, as given in Table 12.1, are a disadvantage in low‐power systems. Nevertheless, the IGBT is now almost universally the electronic switch of choice in systems from 1 kW up to several hundred kW, with the ‘upper’ limit rising each year with ongoing improvement in the technology. The thyristor has been the electronic switch most commonly adopted in small‐scale power electronics such as dimmer switches for electric lights. Unlike the MOSFET and the IGBT, the thyristor can only serve as an electronic switch — it has no other application. The transition from the ‘off’ to the ‘on’ state is triggered by a pulse of current into the gate. The device then remains in the ‘on’ or conducting state until the current flowing through it falls to zero. This characteristic makes the thyristor particularly useful in circuits 367 368 Fuel Cell Systems Explained Figure 12.12 Symbol used for an electronically operated switch. dedicated to the rectification of AC. Various types of thyristor, in particular the gate turn‐ off (GTO) version, can be switched off even when current is flowing by the application of a negative current pulse to the gate. Despite the fact that the switching is achieved by just a pulse of current, the energy required to effect the switching is much greater for the thyristor than for the MOSFET or the IGBT. Furthermore, the switching times are markedly longer. The only advantage shown by the thyristor (in its various forms) for DC switching is that higher currents and voltages can be accommodated. The circuit symbol for a switching device, whether MOSFET, IGBT or thyristor, is generally the ‘device independent’ symbol shown in Figure 12.12. In all cases, it is essential that the switch moves as quickly as possible from the conducting to the blocking state, and vice versa. No energy is dissipated in the switch when it is open and only very little energy is lost when it is fully closed; it is during the transition between open and closed states that the product of voltage and current is non‐zero and power is lost. This latter loss appears as heat and therefore in all cases some cooling of the switching device is required, and is usually provided by mounting it on a heat sink of large area that may be cooled with an air blower. 12.2.2 Step‐Down Regulators The electronic switches that have just been described are used in regulators such as the switch‐mode ‘step‐down’ or ‘buck’ switching regulator (or chopper), as shown diagrammatically in Figure 12.13. The essential components are an electronic switch with an associated drive circuit, a diode and an inductor. When the switch is on, the current from the fuel‐cell stack flows through the inductor and the load, as indicated by the circuit in Figure 12.13a. A voltage drop across the inductor, due to the induced magnetic field inside the coil, initially limits the current through the circuit.4 The current gradually rises over a short time as the inductor becomes fully magnetized. The switch is then turned off, and the energy stored in the inductor keeps the current flowing through the load via the diode, as illustrated by the circuit in Figure 12.13b. The different currents during each part of this on–off cycle are shown in Figure 12.14. If necessary, the voltage across the load can be further smoothed through the use of capacitors. 4 The effect is often described as a ‘back‐voltage’, because it is the voltage that ‘pushes back’ against the current that induces it. The back‐voltage is the voltage drop in an AC circuit caused by magnetic induction. The Complete System and Its Future Inductor (a) V1 Fuel cell On Load Current path when switch is on Inductor (b) V1 Fuel cell Load Off Current path when switch is off Figure 12.13 Circuit diagrams showing the operation of a switch‐mode, step‐down, voltage regulator: (a) current path when switch is on and (b) current path when switch is off. On On On Off Off Off Current supplied by fuel cell during the on time Current circulating through diode during the off time Current through load, being the sum of these two components Figure 12.14 Currents in the switch‐mode, step‐down, regulator circuit. 369 370 Fuel Cell Systems Explained If V1 is the supply voltage and the ‘on’ and ‘off ’ times for the electronic switch are ton and toff, then it can be shown that the output voltage, i.e., the voltage across the load, V2, is given by: V2 ton ton toff V1 (12.20) The regular variations in voltage across the load are known as ripple, and these are influenced by the frequency of switching — at higher frequency, the ripple is less. Nonetheless, each turn‐on and turn‐off involves the loss of some energy, and therefore the frequency should not be too high. A control circuit is needed to adjust ton to achieve the desired output voltage, and such circuits are readily available from manufacturers. The main energy losses in the step‐down chopper circuit are as follows: ● ● ● ● Switching losses in the electronic switch. Power lost in the switch while it is turned on (0.6 × I for an IGBT or RDS,on × I2 for a MOSFET). Power lost because of the resistance of the inductor. Losses in the diode, 0.6 × I where I is the current in the circuit and the value of 0.6 is taken to be the voltage loss over the diode. In practice, all of these losses can be reduced to very low levels. The efficiency of such a step‐down chopper circuit should be at least 90% and in systems with voltages of 100 V or higher, efficiencies of over 98% are routinely achieved. As a note of caution, mention should be made of an alternative to the step‐down regulator, namely, the ‘linear’ regulator circuit. A transistor that has been designed to provide a variable resistance between the emitter and the collector, rather than an on– off function, is employed in this circuit. The gate voltage is adjusted so that the device resistance is at the correct value required to drop the cell voltage. The linear circuit is commonly used in some low‐cost electronic devices but is inefficient and is a poor choice for fuel‐cell systems. 12.2.3 Step‐Up Regulators Since fuel cells are low‐voltage devices, it is usually desirable to step up or boost the DC voltage. This can also be simply and efficiently done by using electronic‐switching circuits. The circuit of a typical switch‐mode, step‐up voltage regulator is shown in Figure 12.15, and its operation is as follows. The starting assumption is that there is charge in the capacitor. When the switch is on, an electric current builds up in the inductor, as shown in Figure 12.15a. The load is supplied by the discharge of the capacitor. The diode prevents the charge from the capacitor flowing back through the switch. When the switch is off, as in Figure 12.15b, the voltage across the inductor increases sharply because the current is falling. Once the voltage rises above that of the capacitor (plus about 0.6 V for the diode), the current will flow through the diode, charge up the capacitor and pass through the load. This will continue as long as there is still energy in the inductor. The switch is then closed again. The Complete System and Its Future (a) Inductor V1 Fuel cell On Load Current paths while switch is on (b) Inductor V1 Fuel cell Off Load Current flow while switch is off Figure 12.15 Circuit diagram to show the operation of a switch‐mode, step‐up, voltage regulator: (a) current path when switch is on and (b) current path when switch is off. High voltages across the load are achieved by having the switch turned off for a shorter period than when in the ‘on’ position. For an ideal switch‐mode regulator with no losses, the voltage across the load (V2) is given by: V2 ton toff toff V1 (12.21) where toff and toff, as before, are the times when the switch is on or off. In practice, the output voltage will be somewhat less, for similar reasons as those for the step‐down regulator. Control circuits are readily available from many manufacturers for both step‐down and step‐up voltage regulators. The losses in the step‐up regulator arise from the same sources as in the step‐down regulator. Since, however, the currents through the inductor and switch are higher than the current through the load, the energy losses are also higher. Also, as all the charge passes through the diode, there will be an energy loss associated with this current. Given these disadvantages, step‐up converters are generally less efficient than step‐down converters, although values of 95% or more are achievable. The step‐up and step‐down switching or chopping circuits are generally referred to as ‘DC–DC converters’, and commercial packages are readily available covering a wide range of power and voltage requirements. 12.2.4 Inverters Fuel‐cell systems that are designed to supply electricity to homes and businesses need to generate AC power. For small, single dwellings, single‐phase AC is required, whereas for more substantial installations a three‐phase supply is generally required. In some 371 372 Fuel Cell Systems Explained cases, the fuel‐cell system may be operating in parallel with a conventional power supply grid; in other situations, it may be a stand‐alone system or grid independent. The rapid emergence of solar photovoltaic (PV) systems in recent years has encouraged power electronics companies to develop inverters that will accept DC voltages from solar panels, and many different types and classes of inverter are on the market. Therefore, given the considerable body of literature that is now available, only a basic description of the operating principle of an inverter will be presented here. 12.2.4.1 Single Phase The basic circuit diagram for a single‐phase inverter is shown in Figure 12.16. There are four electronic switches, labelled A, B, C and D, that are connected in a so‐called H‐bridge. Across each switch is a diode. A resistor and an inductor represent the load through which the AC is to be driven. The circuit operates as follows. First, switches A and D are turned on (B and C are open) and a current flows to the right through the load. These two switches are then turned off, and switches B and C are turned on, thereby causing a current to flow in the opposite direction through the load, i.e., right to left. The diodes across the four switches provide a safe path for any charge to dissipate when the switches are turned off. The resulting current waveform is shown in Figure 12.17. In some situations, though increasingly few, this waveform will be adequate. Most homes and businesses are currently supplied with electricity, which is generated in thermal power stations by means of rotating equipment that produces AC of varying voltage in the form of an almost perfect sine wave at a fundamental frequency of 50 or 60 Hz.5 A B Fuel cell Load C D Figure 12.16 H‐bridge inverter circuit for producing single‐phase alternating current. 5 The frequency of distributed power was determined by the rotation of steam‐powered generators at the end of the 19th century. In the early days of electrification, so many frequencies were used that not one value prevailed (London in 1918 had 10 different frequencies). The proliferation of frequencies grew out of the rapid development of electrical machines in the period 1880–1900. In the early incandescent lighting period, single‐phase AC was common and typical generators were 8‐pole machines operated at 2000 rpm that gave a frequency of 133 cycles per second. As distribution networks developed, it became necessary to standardize the frequency. In Europe, generator manufacturers chose 50 Hz, whereas, for the most part, manufacturers in the United States adopted the 60 Hz standard. The Complete System and Its Future Current in load A and D on B and C on A and D on Time Figure 12.17 Current versus time graph for a square‐wave, switched mode, single‐phase inverter. By contrast, the more square waveform delivered by the simple circuit of Figure 12.16 is incompatible with distributed electricity and the appliances that it powers. The incompatibility arises because a square voltage waveform is composed of many different frequencies, or harmonics, in addition to the fundamental frequency. The higher frequency harmonics can have harmful effects on other equipment connected to the distribution network (grid), and on cables, transformers and switchgear. Harmonics can cause inefficiencies in electric motors and damage to computers and other electronic equipment. The possibility of such adverse behaviour now requires networks to impose strict regulations concerning the ‘purity’ of the waveform of any AC power source that may be connected to the grid. The standards vary between countries and inverter manufacturers have to ensure that the AC output is as near to a pure sine wave as possible with a minimum of harmonics, or ‘harmonic distortion’. To achieve purity in the AC generated by an inverter, pulse width modulation (PWM) is employed to control the IGBTs or other electronic switches employed. More recently the ‘tolerance band’ technique has been applied (v.s.). The principle of PWM is shown in Figure 12.18 and relates to the same circuit as shown in Figure 12.16. In the positive cycle, only switch D is on all the time, while switch A is active only intermittently (i.e., pulsed on–off ). When A is on, current starts to flow through the load and switch D. The load inductance causes a back‐voltage, which ensures that the current is initially small and increases over a very short time. When A is turned off, the current that has built up continues to flow around the bottom right loop of the circuit due to the load inductance through switch D and the ‘free‐wheeling’ diode in parallel with switch C. A similar process occurs in the negative cycle, except that switch B is on all the time and switch C is ‘pulsed’. When C is on, current builds up in the load, and when C is off, it continues to flow (though declining) through the upper loop in the circuit and through the diode in parallel with switch A. Control of switches A and C is carried out by an electronic circuit that modulates the on–off times, i.e., pulse widths according to a pre‐ determined sequence to generate a varying voltage that best approaches that of a sine wave. The precise shape of the voltage waveform generated over the load will depend on the nature (resistance, inductance and capacitance) of the load as well as the pulsing of switches A and C, but a typical half‐cycle is shown in Figure 12.19. The waveform is still 373 374 Fuel Cell Systems Explained Time Sinusoidal current required Control voltages on electronic switches of H bridge A D C B Figure 12.18 Pulse‐width modulation switching sequence for producing an approximately sinusoidal alternating current from the circuit of Figure 12.17. Voltage Time Figure 12.19 Typical voltage versus time waveforms from a pulse‐modulated inverter. The Complete System and Its Future not a sine wave but is much closer than that of Figure 12.17. Clearly, the more pulses there are in each cycle, the closer will be the wave to a pure sine wave and the weaker will be the harmonics. A common standard is 12 pulses per cycle. This requires the switches to be able to operate at frequencies twelve times those of normal a 50 Hz supply, i.e., at least 600 Hz. One of the issues with PWM is that to improve the sine wave output, i.e., limit the harmonics, high frequency switching is required. This can place a strain on the electronic switches and can lead to low inverter efficiency. In modern inverters, the switching pulses are generated by microprocessor circuits, and some very sophisticated approaches have been developed that are outside the scope of this book. One example is the abovementioned tolerance band pulse method, in which switching is controlled so that the output voltage at any time mirrors that of a waveform generated by the microprocessor circuit, within a certain tolerance band of upper and lower voltage limits. The method is illustrated in Figure 12.20. The output voltage is continuously monitored and compared with an internal ‘upper limit’ and ‘lower limit’, which are sinusoidal functions of time. In the positive cycle switch D (in Figure 12.16) is on all the time. Switch A is turned on, and the current through the load rises. When it reaches the upper limit, A is turned off, and the current flows on, though declining, through the diode in parallel with C, as before. When the lower limit is reached, switch A is turned on again, and the current begins to build up again. This process is continuously repeated, with the voltage rising and falling within the tolerance band. In Figure 12.20, the on/off cycle is shown in (a) for a wide tolerance band and in (b) for a narrow tolerance band. It should be appreciated that the resistance and inductance of the load will also affect the waveform and hence the frequency at which switching occurs. The method is therefore an adaptive system that always keeps the same deviation from a sine wave and hence limits the unwanted harmonics below fixed levels. Tolerance band methods are employed in multilevel inverters that are built to raise AC voltages to high levels for power distribution on networks. (a) (b) C On C Off C On On On Off Upper limit On Off Off Upper limit Lower limit Lower limit Figure 12.20 Graph summarizing some data from a real 250‐kW fuel‐cell system used to power a bus: (a) narrow tolerance band and (b) wide tolerance band. (Source: Derived from data in Spiegel, RJ, Gilchrist, T and House, DE, 1999, Fuel cell bus operation at high altitude, Proceedings of the Institution of Mechanical Engineers, Part A, 213, 57–58.) 375 376 Fuel Cell Systems Explained 12.2.4.2 Three Phase In almost all parts of the world, AC electricity is generated and distributed using three parallel circuits, the voltage in each one being out of phase with the next by 120°. While most homes are supplied with just one phase, most industrial establishments have all three phases available. For industrial combined heat and power (CHP) systems, for example, the DC from the fuel cell will need to be converted to three‐phase AC. A three‐phase inverter is only a little more complicated than the single‐phase device. The basic circuit is illustrated in Figure 12.21a. The output of the inverter is shown connected to the primary of a three‐phase transformer. Six switches, with free‐wheeling diodes, are connected to the primary windings of a three‐phase transformer on the right, which represents the load. Equally, the load could be a three‐phase motor such as would be found on an industrial pump or compressor, each of which could be BoP components in a MW‐scale fuel‐cell system. Adopting the same principle described for the single‐phase circuit of Figure 12.17, the switches enable the generation of three (a) Three-phase transformer primary (b) Voltage A B C Time Time Time Figure 12.21 (a) Circuit diagram of a simple three‐phase DC–AC inverter and (b) current versus time graph for the inverter assuming a purely resistive load. One complete cycle for each phase is shown. Current flowing out from the common point is taken as positive. The Complete System and Its Future similar, but out‐of‐phase, voltage waveforms. Each cycle can be divided into six steps. The graphs in Figure 12.21b show how the current in each of the three phases changes with time through employing this simple arrangement. As with the single‐phase inverters described earlier, the switching sequence in a three‐phase inverter is modified — by means of pulse‐width modulation or tolerance band methods — to achieve current and voltage waveforms that closely approach that of a sine wave at the required frequency. The modern three‐phase ‘Universal’ inverter is built along similar lines whether it is for high or low power, and whether it is ‘line‐commutated’ (i.e., the timing signals are derived from the grid to which it is connected) or ‘self‐commutated’ (i.e., independent of the grid). Indeed, the same basic circuit is used irrespective of the modulation method. Particularly with the growth in solar PV, both single‐phase and three‐phase inverters have become commodity items. The circuit is shown in Figure 12.21(a). Signals to turn the switches on and off are taken from a microprocessor. Voltage‐ and current‐ sensing signals may be taken from the three phases, the input, each switch or other places. Digital signals from sensors may also be employed and both instructions and information may be sent to and received from various parts of the system. In all cases the hardware will essentially be the same, i.e., as shown in Figure 12.22. Inverter units have thus become like many other electronic systems — a standard piece of hardware that can be programmed for diverse applications. Analogue-toVoltage digital and current converter sense signals Microprocessor Digital signals e.g., alarms Data to and from a supervisory controller Figure 12.22 ‘Universal’ three‐phase inverter. To gates of the six electronic switches 377 378 Fuel Cell Systems Explained 12.2.5 Fuel‐Cell Interface and Grid Connection Issues The point has been made throughout this book that the power output of a fuel‐cell stack, as measured by the voltage and current, is dependent on various operating parameters such as fuel and oxidant flow rates, pressure and temperature. The output of the stack, in turn, impacts on the DC–DC converter, which may also influence the inverter. Unlike the inverter in a power generation system such as a PV array or wind turbine, however, the performance of the inverter in a fuel‐cell system can actually affect the fuel‐cell stack. The main issue is that excessive ripple currents, which may be caused by faulty PWM in the inverter, are known to cause degradation of the fuel‐cell catalysts. Provision must therefore be made for the monitoring of such currents and the annunciation of an alarm if the levels get too high. When designing a fuel‐cell system for stationary application, e.g., cogeneration systems from a few kW up to many MW, the power that is generated has to meet certain quality standards before the facility can be connected to the electricity distribution grid. Again, an important issue is the level of harmonics generated in the power electronics circuits. Devices must also be in place to protect the fuel cell, the inverter and the grid from faults such as short-circuits (surge protection) and lightning strikes. Such faults can develop with any power generation equipment (e.g., privately owned solar PV or wind turbines) and are not specific to fuel‐cell systems. Any grid‐connected inverter also has to produce AC power that matches the existing power presented on the grid. In particular, a grid‐interactive inverter must match the voltage, frequency and phase of the power line to which it is connected. There are numerous technical requirements to the accuracy of this tracking. Home and business owners that have installed solar PV arrays may be able to sell excess power from their system back into the grid, i.e., when the production exceeds the demands of the home or business load. For PV systems, this situation may occur in the middle of the day when solar incidence is at its highest. Utilities have set feed‐in tariffs that encourage such flow back into the grid. Given the right circumstances, power returned to the grid from a distributed generator (DG) such as a fuel‐cell system can reduce the need for upgrading the local grid. Indeed, although a fuel cell may be sized to meet the load of the home or business, there may be occasions when it becomes profitable for excess power generated locally to be exported into the grid. Grid‐ connected systems also benefit in that if the DG fails, the grid provides emergency backup. If, however, the grid fails, e.g., through a lightning strike at some location on the high‐voltage transmission system, a situation can arise in which a DG continues to power a section of the distribution system. The incident is known as ‘islanding’ and can be dangerous to utility workers, who may not realize that a circuit is still powered. When an islanded fuel cell reconnects to the grid, it is essential that the inverter has provision to lock the phase of its AC voltage waveform to that of the grid. 12.2.6 Power Factor and Power Factor Correction The voltage and current of an AC circuit are sinusoidal in nature (see Figure 12.23), i.e., the amplitude of both the current and the voltage of an AC circuit constantly changes over time. In a purely resistive AC circuit, voltage and current waveforms are in step (or in phase) and thus change polarity at the same instant in each cycle. All the power The Complete System and Its Future Voltage Current Figure 12.23 Voltage and current out of phase. The reactive power can be generated locally by distribution systems such as fuel cells. VA r/k nt re pa e ow p Ap Reactive power/kVAR Phase angle (ϕ) Real or true power/kW Figure 12.24 Relationship between true power, apparent power and reactive power. entering the load is consumed (or dissipated). Such power is defined as the ‘true power’, also known as ‘active power’, ‘real power’ or ‘useful power’. This ideal situation is rarely met in practice because loads are reactive, i.e., they have elements such as capacitors or inductors present, and consequently create a phase difference between the current and voltage waveforms. During each cycle of the AC voltage, extra energy — in addition to any energy consumed in the load — is temporarily stored in the load and then returned to the grid in a fraction of the cycle later. The extra energy is known as the ‘reactive power’. The combination of reactive power and true power gives the ‘apparent power’, which, in an AC circuit, is the total of all the power both dissipated in the load and absorbed or returned to the grid. When the circuit is purely resistive, the apparent power is equal to true power, but in an inductive or capacitive circuit, the apparent power is greater than the true power. Therefore, the following are encountered in AC circuits: ● ● ● True (active, real or useful) power, expressed in watts (kW). Reactive power, usually expressed in reactive volt‐amperes (kvar). Apparent power, usually expressed in volt‐amperes (kVA). To help understand the relationship between true power and reactive power, the two parameters are shown as two sides of a right‐angled triangle, as depicted in Figure 12.24. The angle Ф describes the phase shift between the voltage and current. The larger the phase angle, the greater the reactive power is generated by the system. The hypotenuse of the triangle is the apparent power, i.e., the vector sum of the 379 380 Fuel Cell Systems Explained true power and the reactive power. The ratio between the true power and the apparent power is known as the ‘power factor’ (PF) and is the cosine of Ф. The power factor is a dimensionless number between −1 and 1. When PF = 0, the energy flow is entirely reactive and stored energy in the load returns to the source on each cycle. When the power factor is 1, all the energy supplied by the source is consumed by the load (i.e., the load behaves as a pure resistor with no capacitance or inductance). A power factor of −1 can result from returning power to the grid, such as in the case of a building fitted with solar panels when their power is not being fully utilized and the surplus is fed back into the supply network. Power factors are usually stated as ‘leading’ or ‘lagging’ to show the sign of the phase angle. Capacitive loads are leading (current leads voltage), and inductive loads are lagging (current lags voltage). A low‐power factor (for which the consumer may be penalized by the network operator) can be increased to an acceptable level (usually above 0.9) by installing a power factor correction (PFC) unit close to the load. The most widely adopted of the PFC units are passive and comprise a series of capacitors, which are incrementally brought on stream by contactors that engage as the load changes. Power factor correction can also be achieved with an active power electronics system that incorporates inverter technology, as described earlier. The switching of the pulses in inverter circuits can be programmed by the user, and in most modern inverters the phase angle can be adjusted when the inverter is installed. This feature is particularly beneficial for fuel‐cell systems since they can often be installed close to the load, unlike other distributed power generation systems such as large‐scale solar or wind farms. 12.3 Hybrid Fuel‐Cell + Battery Systems In general, all types of fuel‐cell stack perform best if they are run under a constant load. Large fluctuating cell voltages can accelerate degradation of both the anode and cathode catalysts and thereby reduce stack lifetime. Combining a rechargeable battery or supercapacitor6 with a fuel cell creates a hybrid system that can accommodate the variations in load that occur during service. When the total power requirements from a hybrid system are low, the surplus electrical energy is stored in the rechargeable battery or capacitor. Conversely, when the power demand exceeds that available from the fuel cell, energy is taken from the battery or capacitor. In essence, hybrid systems are using the fuel cell as a battery charger. Two extreme types of hybrid may be considered. One of these facilities relates to a situation where the electrical power demand varies in a predictable manner and, therefore, the hybrid system is mainly in ‘standby’ mode for fairly long periods, and the fuel cell serves to recharge the battery. Whereas the battery is required to supply most of the power during ‘transmit’ or peak periods, the fuel cell operates more or less continuously in providing average power to charge the battery. In turn, the battery must deliver sufficient power, hold enough energy for the regular and frequent loads 6 The advantages of the capacitor are that the charge–discharge cycle is more efficient and much faster. The disadvantages are that capacitors store much less energy for a given space and are more expensive per watt‐ hour stored. The Complete System and Its Future DC–DC inverter Rechargeable battery Fuel-cell stack L o a d EN Battery fully charged sense Figure 12.25 Diagram of a simple fuel‐cell–battery hybrid system. and be conducive to recharging in the periods in between. Examples of this set of requirements are to be found with certain data‐logging devices, telecommunications systems and land‐ or buoy‐based navigation equipment. The other extreme type of hybrid is designed for situations in which the power demand can be highly irregular and unpredictable as encountered, for instance, with mobile phones. In most cases, the combination of battery and fuel cell offers the additional benefit of a lower cost compared with a fuel‐cell system that is sized to meet the highest demand. A hybrid concept that employs a fuel cell and a battery is shown schematically in Figure 12.25. In addition to the components shown, a controller is required to prevent overcharging of the battery. As discussed in Section 12.2.3, a DC–DC converter may be required to match the output voltage of the fuel‐cell stack with that of the battery/load. Such a system would be attractive for application with direct methanol fuel cells (DMFCs) or other cells that run on liquid fuels, from all of which the average power is very low. Full‐hybrid arrangements (also known as ‘hard hybrids’) are those in which the peak power that may be required is much greater than the average power that can be delivered by the fuel cell and therefore takes full advantage of power that is stored in the battery. By contrast, a mild (or ‘soft’)‐hybrid system is one in which the battery power and energy storage are quite low compared with the power delivered by the fuel cell. Such systems may be found in vehicles and boats, for example, where there is little difference between the peak and average power demands. An example of the variation in system power with time for a mild hybrid is shown in Figure 12.26a. The fuel‐cell power is sufficient for most of the time, but the battery can ‘shave’ or ‘lop’ the peaks off the power requirement and thus can substantially reduce the required fuel‐cell capacity. During times when the fuel‐cell power exceeds the power demand, the battery is recharged. This profile is typical of urban electric vehicles, where the peaks correspond to occasions such as accelerating from traffic lights, but for most of the time the vehicle will be proceeding slowly and steadily, or else it will be stationary. An additional option is illustrated in the power–time graph of Figure 12.26b. Here an electric vehicle is using the motor as a brake. The electric motor functions as a generator to convert the motion energy into electrical energy. The energy can either be passed through a resistor and dissipated as heat (referred to as ‘dynamic braking’) or can be passed to a rechargeable battery, to be used later to run the motor. The latter 381 Fuel Cell Systems Explained Power (a) Average power Time (b) Maximum fuel-cell power Power 382 Braking Time Figure 12.26 Power versus time graphs for systems suitable for a soft hybrid electric vehicle: (a) without and (b) with regenerative braking. method is known as ‘regenerative braking’ and is clearly the better option from the point of view of system efficiency, but it does presuppose a hybrid system with a rechargeable battery. A fairly sophisticated control system with a responsive fuel cell is required when regenerative braking is incorporated in a vehicle. The battery needs to be operated in a partial state‐of‐charge condition so that it can absorb the electrical energy supplied by the motor during braking. The flow of power from the fuel cell to the battery will have to be properly controlled, since the duty cycle of the vehicle is much more variable than one without regenerative braking. That is, the fuel cell needs to respond quickly depending on whether it is required to provide a small amount of power to top up the battery or whether a burst of energy is demanded because the battery is at a low state-of-charge and the vehicle needs to accelerate. The motor controller also has to be a full ‘four‐quadrant’ type.7 The resulting hybrid system is illustrated diagrammatically in Figure 12.27. A measurement of the battery state-of-charge is necessary, rather than just a ‘fully-charged’ indication. The energy flows to and from the different components of the system that can be quite complex. Given that the battery in the hybrid system just discussed is supplying fairly short‐term power peaks and also absorbing power from regenerative braking, the rate of transfer of charge in and out of the battery is likely to create an operational problem. This is where the introduction of a ‘supercapacitor’ promises to be particularly beneficial. The specific 7 Four quadrants refer to the four possible modes of motoring, namely, forwards accelerating, braking while going forwards, backwards accelerating and braking while going backwards. The Complete System and Its Future Four-quadrant inverter Fuel-cell stack M Control Rechargeable battery State-of-charge measurement Power switching and control Control Power/braking demand Figure 12.27 Diagram of a hybrid fuel‐cell–battery system for a vehicle with regenerative braking. Bold arrows indicate energy flows. A current‐sensing resistor in series with the battery measures the input and output of charge. energy (Wh kg−1) of such a device is much less than that of rechargeable batteries, but the power density is very much greater — typically 2.5 kW kg−1. Supercapacitors can also sustain at least 500 000 charge–discharge cycles. The UltraBattery™, invented by the Commonwealth Scientific and Industrial Research Organisation (CSIRO) in Australia,8 is the first practical example of combing the attributes of a rechargeable battery — in this case, lead–acid chemistry — and a supercapacitor into an efficient and affordable energy-storage device. The technology was primarily designed to meet the high‐rate partial state‐of‐charge duty required from hybrid electric vehicles (HEVs) but can be reconfigured for a variety of applications, for example, power tools, forklift trucks, high‐power uninterruptible power supplies, remote‐area power supplies and grid frequency regulation. The CSIRO technology has been taken up by the Furukawa Battery Co., Ltd., Japan, and the East Penn Manufacturing Co., Inc., USA, and is under mass production for both HEV and renewable energy applications. The UltraBattery™ is under evaluation by many carmakers worldwide and it has been adopted in both the Honda Odyssey Absolute and the Honda StepWGN hybrid models as original equipment. Hybrid electric vehicles have grown in popularity in recent years — driven by improved fuel utilization and emissions reduction. When fuel cells are employed in hybrid vehicles, the two basic configurations available are those illustrated in Figure 12.28. In addition to fuel cell + battery hybrids, there are a large number of other hybrid options that can include a fuel‐cell stack. An example is that of a solar PV array, fuel cell and battery. Such systems have been successfully deployed in roadside variable message signs and could also serve as remote power supplies. The fuel cell may compensate for the somewhat unreliable nature of solar energy in many situations. 8 Lam, LT, Haigh, NP, Phyland, CG and Rand, DAJ, 2005, High performance energy storage devices, International Patent WO/2005/027255. 383 384 Fuel Cell Systems Explained (a) Battery or supercapacitor Ancillary devices Regeneration, recharge battery when needed Assist fuel cell when needed Stored hydrogen Fuel-cell stack DC/DC converter Inverter Motor (b) Ancillary devices Stored hydrogen Fuel-cell stack DC/DC converter Battery Inverter Motor Figure 12.28 Diagrams of (a) parallel and (b) series versions of hybrid fuel‐cell vehicle. 12.4 Analysis of Fuel‐Cell Systems Throughout this book, it has been stressed that many different components are required to build a complete fuel‐cell system. Whereas the choice of stack technology will be influenced by the application, the availability of fuel will determine the extent of any fuel processing that may be required. The application will also dictate the level of power electronics that is necessary to operate the complete system. Over and above the issues of system design and construction, the cost of materials and fabrication are keys to the successful commercialization of fuel‐cell systems. Despite all of their advantages in terms of environmental impact and performance with respect to other technologies, most fuel‐ cell systems have competitors. Moreover, many of these alternative technologies are well‐established in mature markets. In the transport sector, for example, it is the internal combustion engine — a highly engineered technology that is of relatively low cost compared with fuel‐cell systems. In the stationary power market, large generators based on gas or steam turbines with lower capital costs can produce electricity more cheaply The Complete System and Its Future than most fuel‐cell systems. Such competing technologies set the cost targets that must be met if fuel‐cell systems are to become commercially viable. Various analytical methods can be employed to establish the competitiveness of different designs of fuel cell. Although economic modelling is outside the scope of this book, the following approaches will be briefly examined as they can help place fuel cells within the wider context of alternative technologies: ● ● ● ● Well‐to‐wheels analysis. Power‐train or drive‐train analysis. System life‐cycle assessment. Flowsheet or process modelling. 12.4.1 Well‐to‐Wheels Analysis The infrastructure required to generate, distribute and store hydrogen for fuel‐cell vehicles (FCVs) is critical to their commercial success. A number of different routes for generating hydrogen have been examined in Chapter 10. Several hydrogen supply options exist for vehicles: ● ● ● ● ● Hydrogen generated by steam reforming of natural gas in a large centralized plant and then delivered as liquid hydrogen by trailer to filling stations. Hydrogen generated by steam reforming of natural gas in a large centralized plant and then delivered as compressed gas by pipeline to filling stations. Hydrogen generated as a by‐product, e.g., from oil refineries and industrial ammonia production plants. Hydrogen produced at filling stations by small‐scale steam reformers that run on pipeline natural gas. Hydrogen produced at filling stations by electrolysers. The consensus view from several studies in the United States is that, in the near term, hydrogen generated by steam reforming of natural gas in small, localized reformers is the best option. Where there is no natural gas supply, hydrogen produced by electrolysis may be the preferred alternative (especially when using renewable electricity). As a hydrogen infrastructure emerges in the future, there may come a time when centralized production would be economic, with the added bonus of being able to collect and sequester the carbon dioxide that is generated from the reformer. To consider the various scenarios for road transport applications, well‐to‐wheels (WTW) analyses may be carried out to quantify the options for generating hydrogen (from the well), its conversion to electricity on a vehicle and transfer of the power to mechanical energy for driving the wheels. Consequently, WTW analysis determines the energy efficiency of converting the energy (in the well) to energy required at the wheel of a vehicle. In a well‐documented WTW analysis carried out in 2001 by General Motors (GM), the Argonne National Laboratory (ANL) and others9 for the North American market, 15 different vehicles were investigated. These included conventional vehicles and HEVs 9 Well‐to‐Wheel Energy Use and Greenhouse Gas Emissions of Advanced Fuel/Vehicle Systems, General Motors, Argonne National Laboratory, BP, Exxon/Mobil and Shell, April 2001. Available from http://www.ipd.anl.gov/anlpubs/2001/04/39097.pdf (accessed 28 September 2017). 385 386 Fuel Cell Systems Explained 8 Gasoline Diesel Naptha Natural gas North American and non-North American 39 Compressed natural gas Methanol Hydrogen Liquid fuels made using Fisher-Tropsch reactors Renewable Corn Woody Herbaceous 12 Ethanol 16 Hydrogen compressed and liquid Crude oil Electricity Combined cycle, hydro nuclear North Europe, CA, and US mixture Figure 12.29 The 75 different pathways investigated in the North American well‐to‐wheels study conducted in 2001 by GM, ANL, BP, Exxon/Mobil and Shell. (Source: Reproduced with the permission of Elsevier.) with spark‐ignition and compression‐ignition engines, as well as hybridized and non‐ hybridized FCVs with and without on‐board fuel processors. All 15 vehicles were configured to meet the same performance requirements. Thirteen fuels, selected from 75 different fuelling options or pathways, were considered in detail. These included low‐sulfur gasoline, low‐sulfur diesel, crude oil‐based naphtha, Fischer–Tropsch (FT) naphtha,10 liquid or compressed gaseous hydrogen (based on five different pathways for production), compressed natural gas, methanol and neat and blended (E85) ethanol. The 75 different pathways for the WTW analyses undertaken in the study, as shown Figure 12.29. There are some key findings from the North American study that are worth examining here, since they may influence the development of infrastructure in many countries that are seeking to adopt hydrogen vehicles. These findings are as follows: Total energy use. For the same amount of energy delivered to the vehicle tank, petroleum‐based fuels and compressed natural gas exhibit the lowest energy losses from the well-to-tank (WTT). FT naphtha, FT diesel, gaseous hydrogen from natural gas, methanol and corn‐based ethanol are all subject to moderate WTT energy losses. By contrast, liquid hydrogen from natural gas, hydrogen from electrolysis (gaseous and liquid), electricity generation and cellulosic bioethanol suffer large WTT energy losses. 10 The Fischer–Tropsch process is used for making fuels artificially. Basically a fuel such as biomass, or even natural gas, is steam reformed using the methods described in Chapter 10. The resulting hydrogen and carbon dioxide products are then reacted, over catalysts developed by Fischer and Tropsch, to produce liquid fuels such as octane (C8H18), nonane (C9H20) and decane (C10H22). The Complete System and Its Future Greenhouse gas (GHG) emissions. Liquid hydrogen (produced in both central plants and filling stations) and compressed gaseous hydrogen obtained by electrolysis can both be energy inefficient and lead to large emissions of GHGs. Ethanol (derived from renewable cellulose sources such as corn) offers a significant reduction in GHG emissions. Other fuel options were found to have moderate energy efficiencies and GHG emissions. Tank‐to‐wheels (TTW) efficiency. Fuel‐cell systems consume less energy than conventional power-trains, because of the intrinsic high efficiency of the stacks. Fuel‐cell vehicles operating directly on liquid or compressed gaseous hydrogen exhibit significantly higher fuel economy than those employing on‐board fuel processors. Overall well‐to‐wheels efficiency. Hybrid systems offer consistently higher fuel economy than conventional vehicles. There has been a plethora of WTW analyses of the North American vehicle fleet in the years that have followed the earlier investigation conducted by General Motors et al. With respect to FCVs, there have been no significant changes in the findings. Well‐to‐wheels studies that have been carried out for European countries, Japan and elsewhere have often lead to similar conclusions.11 That is, optimum results are realized when renewable energy sources such as wind, solar or biomass are used in the production of hydrogen. To a lesser extent, natural gas vehicles offer improvements relative to hybrid vehicles and are less problematic in engineering terms than conventional engines adopted to run on hydrogen. In the context of vehicle efficiency, it is noteworthy that the addition of a small amount of hydrogen can enhance the combustion of liquid fuels in gasoline and diesel engines. Although not yet promoted among leading vehicle makers, several studies have shown that engine efficiency can be improved slightly and emissions reduced by injecting hydrogen with the fuel. Research is underway to generate the hydrogen required for this treatment by utilizing heat within the exhaust gas to reform some of the liquid fuel that is carried on board. 12.4.2 Power‐Train Analysis In a conventional gasoline‐ or diesel‐fuelled vehicle, energy from the fuel is transmitted from the engine to the wheels by a mechanical power-train. In an FCV, the power-train is electrical — energy from the fuel is converted into DC electricity (which may be converted to AC) to power the motor(s). In an HEV, the power-train may involve a combination of electrical and mechanical energy conversion. A power‐train or drive‐ train analysis is simply quantification of the energy transfer in the vehicle from fuel to wheels — it is the final stage in the WTW pathway. The analysis can help define the relative sizes of the components required for specific needs. For example, urban delivery vehicles may employ a series hybrid power-train (cf. Figure 12.28b) in which the battery needs to be large as it is providing most of the power for stopping and starting (short journeys), and therefore the fuel cell can be small. In such a system, the fuel cell is acting principally as a charger to keep the battery topped up. By contrast, in a parallel hybrid drive-train (cf. Figure 12.28a), most of the power may be produced by the fuel cell, which serves also to keep the battery charged for whenever rapid acceleration is 11 Grube, T, Hohlein, B, Stiller, C and Weindorf, W, 2010, Systems analysis and well‐to‐wheels studies, in Stolten, D (Ed.), Hydrogen and Fuel Cells, Wiley‐VCH Verlag GmbH, Weinheim, pp. 831–852. 387 388 Fuel Cell Systems Explained required. Thus, the parallel hybrid may be a better option for long‐range vehicles. In power‐train analysis, other issues may also be considered. For example, whether one electric motor is to be employed, with the power transmitted mechanically to the wheels via a conventional axle differential, or whether it is beneficial to power each wheel with a dedicated hub motor, thereby reducing mechanical energy losses. 12.4.3 Life‐Cycle Assessment WTW and power‐train analyses are specific examples of what is generally referred to as a ‘life‐cycle assessment’ (LCA) (also known as ‘life‐cycle analysis’ and ‘cradle‐to‐grave’ analysis). The technique assesses environmental impacts associated with all the stages of a product’s life from cradle-to-grave, i.e., from raw material extraction to materials processing, manufacture, distribution, use, repair and maintenance, and disposal or recycling at the end of life. Life‐cycle assessments are employed widely to: ● ● Draw up an inventory of relevant energy and material inputs, as well as emissions to the environment. Evaluate the potential consequences associated with identified inputs and outputs of a process or product. The information obtained from a LCA study is used to improve processes, support policy and provide a sound basis for informed decisions on, for example, technology development. Fuel‐cell systems are candidates for LCAs because they impact the environment in terms of inputs (fuel) and outputs (emissions) throughout their working life and because the materials of construction have a cost associated with extraction, assembly and disposal at the end of life. For example, in an LCA of a PEMFC system, the combined cost of the mining and extraction of platinum, which is required for the catalysts, and of that incurred in the recovery of the metal at the end of the stack lifetime has to be factored in to the life‐cycle cost of the system. The same consideration has to be given to other catalysts employed in the fuel‐processing stages of a system. In this respect, the material in an MCFC, for example, will involve lower costs in extraction and recovery after service than those required for a PEMFC or a phosphoric acid fuel cell (PAFC). In addition, the power that can be generated per kg of catalyst may be much more for the MCFC than for the PEMFC, and therefore the life‐cycle cost of the MCFC may be lower than that of the PEMFC. Such issues have to be considered in a full LCA of any fuel‐cell system and may help identify the best pathway for system development. The procedures undertaken in an LCA are described in ISO 14040 : 2006 and 14044 : 2006, which are included in the ISO 14000 family of Environmental Management Standards set by the International Organization for Standardization (ISO). The standards distinguish four phases of an LCA: 1) Goal defining and scope ● Define the purpose of the activity. ● Define boundary conditions. ● Define the life cycle of the product, process or activity. ● Recognize the general material flow in the life cycle. ● Identify all operations that contribute to the life cycle and fall within the system boundaries, including time and spatial boundaries. The Complete System and Its Future 2) Inventory ● Quantify energy, raw materials and environmental releases throughout all stages of the life cycle. ● Define scope and boundaries by gathering data and running computer models. ● Analyse results and draw conclusions. 3) Impact assessment ● Characterize the effects of resource requirements and environmental loadings identified. ● Address ecological and human health impacts. 4) Improvement assessment ● Evaluate needs and opportunities to reduce environmental burdens throughout the whole life cycle. Since a full LCA can be both a costly and time‐consuming exercise, only well‐established and large organizations with a budget for sustainability studies are capable of completing such a task. More often, WTW or process studies are undertaken as a means of comparing one fuel‐cell system with another. 12.4.4 Process Modelling Fuel‐cell power plants for the stationary market can be divided broadly into two categories: power‐only systems and cogeneration systems. The former include backup or uninterruptible power systems running on hydrogen, as well as various niche systems that may be fuelled with alternative fuels such as ammonia or methanol. Examples of power‐only systems can be found operating in telecommunication towers and similar remote applications. By contrast, cogeneration systems produce useful heat (or cooling) as well as electricity and are therefore more complex. Portable versions of these systems for consumer electronic items present their own unique issues relating to miniaturization of the fuel processing and close integration of the stack and control system, such issues have been discussed in Chapters 4–6. In process modelling, the designer starts with a PFD and carries out an analysis of the steady‐state operation of the plant. Once this is complete, consideration is given to how the stack is to be started up and shut down, how it responds to load changes during operation and what eventuates as the cell degrades. All of these features can be modelled using computer programmes such as Hysis™, MATLAB®, Simulink®, Cycle‐Tempo (Technical University of Delft) and Aspentech™. The last‐mentioned programme will model steady‐state processes as well as dynamic changes. TRNSYS is a useful platform for dynamic modelling of thermal energy systems and can show how fuel cells will integrate with other energy generators, e.g., PV cells, wind turbines and electrolysers. Once the fuel‐cell system has been modelled, the basic requirements for the stack and each item of the BoP will be understood. Consequently, a specification can be drawn up for the plant and detailed design can begin. To illustrate the issues involved in the genesis of a system, the design of a stationary PEMFC system for application at a scale of around 100 kW is considered in the following text. The PFD is shown in Figure 12.30. Various assumptions are made as follows: ● Fuel input — natural gas fed at a flow rate of 1 kmol h−1, i.e., approximately equivalent to a 100‐kW supply. For simplicity, the system will be modelled in the following for pure methane. 389 390 Fuel Cell Systems Explained 20 HX-12 Natural gas Water Steam Reformer 3 2 HX-2 AIR HX-11 Pre-reformer 18 4 5 7 6 HX-4 HX-3 8 HX-5 HTS 9 12 HX-6 LTS 14 PROX HX-8 Combustion Cooling water 19 15 Exhaust 16 Fuel cell Fuel cell off-gas 17 Stream Temp/°C Pressure/Bar Mole flow/kmolh Mass flow/kgh Enthalpy/kW Mole flows/kmolh CH CO CO H O H O N Nat. Gas Water Air 2 5 6 7 9 3 25 25 20 300 285.5 750 400 400 150 1.2 1.7 1 1.69 1.66 1.65 1.64 1.62 1.68 1 3 13.2 1 4.021 5.958 5.958 5.958 5.959 16.0 54.0 379.3 16.0 70.1 70.1 70.1 70.1 70.1 –20.92 –239.4 0 –17.43 –210.4 –131.3 –151.1 –155.8 –170.8 1 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 2.88 0 10.25 1 0 0 0 0 0 0 0.989 0.011 0 2.989 0.032 0 0 0.021 0.595 0.384 1.637 3.321 0 0 0.021 0.595 0.384 1.637 3.321 0 0 0.021 0.193 0.787 1.234 3.723 0 0 0.021 0.005 0.976 1.045 3.912 0 0 12 15 16 120 80 70 1.61 1.59 1.59 5.959 5.958 11.09 70.1 71.1 318.6 17 18 80 300 1.58 1.57 15.41 15.41 389.8 389.8 19 70 1.6 2.04 58.9 20 775 1.56 17.21 449.6 0.021 0.021 0.005 0 0.976 0.981 1.045 1.1 3.912 3.856 0 0 0 0 0.021 0 0.981 4.379 0.578 0.361 9.09 0 0 0 0 0 0.45 1.59 0 0 1.053 5.1 0 0.38 10.68 0 0 0 0 0 2 9.09 0.021 0 0.981 4.379 0.578 0.361 9.09 Figure 12.30 Process flowsheet for a 100 kW PEMFC system operating in a stationary power application of about 100 kW output. Note that, for clarity, some stream numbers have been omitted from the table. HX, heat-exchanger; HTS, high‐temperature shift; LTS, low‐temperature shift. (Source: Reproduced with the permission of Elsevier.) ● ● ● ● ● ● ● Steam reforming will be the means of conversion, with an initial steam–carbon ratio of 3. This is higher than the minimum required thermodynamically to prevent carbon deposition (see Section 10.4.4, Chapter 10). The reformer operates with an outlet temperature of 750°C to maximize hydrogen production. The reactions are all modelled assuming that there are no kinetic limitations, i.e., reactions reach a state of thermodynamic equilibrium. Desulfurization of the natural gas is carried out at atmospheric pressure and temperature by an absorber and is not therefore considered in this example system. Pre‐reforming is carried out in an adiabatic reactor, operating between 300 and 250°C, to reduce the concentration of hydrocarbons of high molecular weight in the feed gas to the main reformer reactor. The feed pressure for natural gas is set at 170 kPa to allow for pressure drops through each of the fuel‐processor elements. There are two stages of shift: (i) a high‐temperature reactor with a catalyst of iron oxide that is operating at an inlet temperature of 400°C and (ii) a low‐temperature reactor that contains a 30 wt.% CuO, 33 wt.% ZnO and 30 wt.% alumina catalyst at 200°C. A preferential oxidation (PROX) unit is modelled in the Aspen flowsheet code by two stoichiometric reactors, one to perform the PROX of carbon monoxide and the other to remove the remaining oxygen via reaction with hydrogen. These have been removed in the flowsheet shown in Figure 12.30, and the PROX reactor is shown as a single unit, as would be experienced in practice. The Complete System and Its Future ● Heat for the reforming reaction is provided by combustion of exhaust gas from the anode of the fuel cell and is supplemented by fresh natural gas, as required. A material balance for the flowsheet was obtained by simulating the whole process using Aspentech software and is summarized in the stream data shown below the PFD in Figure 12.30. Pinch analysis12 was applied to optimize the layout and connection of the heat-exchangers. Knowing the flow rates of gas through each of the process units and the respective catalysts employed, the sizes of the reactors can be calculated and a preliminary mechanical design formulated. Also once the flow rates and heat loads are known, the system efficiency can be found from: Efficiency AC output power of system Power of fuel supplieed Power of fuel (stack efficiency inverter efficiency) paarasitic loads Power of fuel supplied The hydrogen is supplied to the stack at a rate of 3856 kmol h−1 (stream 15). This is equivalent to 259 kW, with respect to the lower heating value (LHV) of 241.83 kJ mol−1. If it is assumed that the fuel utilization in the stack is 85% and that an operating voltage of 0.65 V is selected, the efficiency of the stack is found from equation (2.28), Chapter 2 as follows: Efficiency f Vc 1.25 0.85 0.65 1.25 0.442 44.2% (LHV) (12.22) Therefore, the electrical output (DC) of the stack is 0.442 × 259 = 114.5 kW. Taking a representative efficiency of 95% for an inverter, the gross AC power produced by the stack will be 114.5 × 0.95 = 108.8 kW. Assuming a parasitic power requirement of 5.89 kW for compressors and pumps, the net AC power delivered by the system can therefore be expected to be 108.8 − 5.89 = 102.9 kW. Given that 1 kg mol h−1 of methane as supplied to the system has a combustion enthalpy of 802.6 kJ mol−1 (LHV), the net efficiency of the whole system is therefore: Efficiency 102.9 3600 802.6 1000 0.462 46.2% (LHV) (12.23) It is possible to estimate the size of the stack required as follows. The total current, I, that has to be delivered by the stack is given by: I PDC Vc 114.5 1000 176 154 A 0.65 (12.24) where PDC is the electrical DC output of the stack and Vc is the stack operating voltage. Referring to Figure 3.1, Chapter 3, which gives the expected performance from a good 12 Pinch analysis was introduced in Section 7.2.3. 391 392 Fuel Cell Systems Explained PEMFC, the current density is expected to be about 600 mA cm−2 if the cell voltage is 0.65 V. The total area of cells that make up the stack will therefore be: Area current current density 176 154 0.6 293 590cm 2 (12.25) This may seem like a large area, but if the stack is built with cells of a reasonable size, say, 50 × 50 cm, each cell area would be 2500 cm2, and the total number of cells required would be 293 590/2500, i.e., approximately 120 cells. In practice, such a system would be made up of a perhaps two stacks wired in parallel, each containing 60 cells. Limiting the number of cells in a single stack to around 60 also keeps the stack voltage to a safe level — in this case, 60 × 0.65 = 39 V. Once a steady‐state system has been established, the designer of the fuel cell will need to focus attention on questions such as the following: ● ● ● ● ● ● Is additional heat required to raise the temperature of some reactors (e.g., the reformer) up to the operating point on start‐up? Is a supply of hydrogen required to activate the reforming catalyst before start‐up? How will the system perform at part‐load or under changing loads? Some dynamic modelling may be required. Is a purge gas required when the stack is shutdown or held in a hot standby condition? What control elements are required (e.g., control valves, thermocouples and other sensors)? What type of control system is needed? After such issues have been resolved, detailed mechanical and electrical drawings can be produced. Normally, and especially for large installations, a hazard identification (HAZID) and/or a hazard and operability (HAZOP)13 study or similar safety risk assessment and analysis will also be conducted. 12.4.5 Further Modelling The approach taken in this book has been a pragmatic one. The fuel‐cell systems that have been described use technologies that have emerged during the course of the 20th century, and chapters have been devoted to particular types of fuel cell. There are, of course, common features to all fuel‐cell technologies in terms of cell components (catalyst layers, electrodes, electrolyte) and the chemical and physical reactions that occur at the interfaces between these components. The processes that occur within a 13 A hazard and operability (HAZOP) study is a standard hazard analysis technique enacted worldwide by process industries for the preliminary safety assessment of new systems or modifications to existing ones. The HAZOP study is a detailed examination, by a group of specialists, of components within a system to determine what would happen if that component were to operate outside its normal design mode. Each component will have one or more parameters associated with its operation such as pressure, flow rate or electrical power. The HAZOP study looks at each parameter in turn and imposes guide words to list possible off‐normal behaviour such as ‘more’, ‘less’, ‘high’, ‘low’, ‘yes’ or ‘no’. The effect of such behaviour is then assessed. The Complete System and Its Future fuel cell can be characterized by using mathematical models. In terms of mass transport, for example, it is possible to model the following actions: (i) flow velocity within a fuel‐ cell channel, (ii) mass transport in the gas‐diffusion layer (GDL) and catalyst layers and (iii) proton and water transport in the membrane. Furthermore, heat transfer within the membrane and electrode layers can also be described. A one‐dimensional (1D) model is the simplest type of mathematical representation of a fuel cell. For instance, it ignores losses due to reactant transport in the feed channels and is a relatively easy to derive using an Excel spreadsheet given the Nernst equation and the theory discussed in Chapters 2 and 3. A simple 1D model will predict quite accurately the open‐circuit voltage of cells running on hydrogen, on the assumption that the cell is isothermal. A two‐ or three‐dimensional model is a more complex proposition on account of the voltage losses that occur during operation a fuel cell, as discussed in Chapter 3. Two approaches can be taken in the development of a model that will account for these losses: 1) Simply use data that has been obtained experimentally to devise an empirical relationship between voltage and current (as a function of the other operating parameters such as temperature and pressure). Although widely adopted in the literature, this procedure does not help in understanding the processes that are occurring within the fuel cell. 2) Apply the Butler–Volmer (or Tafel) equation and knowledge of the materials’ properties to estimate the total cell voltage loss by summing all of the individual losses at the electrodes and through the electrolyte. Approach 2 has been adopted with some success for PEMFCs, but a particular difficulty arises in trying to account for the transport of water.14 With high‐temperature fuel cells, there is no water‐management issue, and the anode overpotential can be virtually ignored. The voltage of an SOFC running on hydrogen, for example, can be determined reasonably accurately from equation (3.10), Chapter 3. Modelling of internal reforming SOFCs (or MCFCs) is a much more difficult proposition. The mechanism of internal reforming is subject to much debate, and there are conflicting models in the literature. A simple approach is to assume that the steam reforming reaction is fast and comes to equilibrium over the anode (or internal reforming catalyst in the case of the MCFC) before the occurrence of any electrochemical reactions. Equation (3.10) can then be invoked to calculate the cell voltage that arises from the reformed equilibrium gas mixture over the whole of the anode surface. Unfortunately, this approach is rather naive because in a functional internal reforming cell, the fuel gas composition changes along the channels from anode inlet to outlet. Water is produced by the anode electrochemical reaction as hydrogen is consumed, and both these processes affect the reforming reaction. It can be expected, therefore, that the concentration of hydrogen at the surface of the anode will decrease on moving from the inlet to the outlet of the cell. While it is possible to measure the composition of the anode exhaust gas, it is virtually impossible to measure the composition of the anode gas along the anode channels of an SOFC or a MCFC. Certainly there is much more to learn about the fundamental processes at the heart of the fuel cell. 14 Berning, T., Lu, D.M. and Djilali, N., 2002, Three‐dimensional computational analysis of transport phenomena in a PEM fuel cell, Journal of Power Sources, vol. 106, pp. 284–294. 393 394 Fuel Cell Systems Explained 12.5 Commercial Reality 12.5.1 Back to Basics In her book, Fuel Cells — Current Technology Challenges and Future Research Needs (Elsevier, 2012), Noriko Behling has argued that: ‘The difficulty of the technology is rooted in the complexities of how fuel cells work, which involve multiple chemical and physical interactions at the atomic level. Perhaps no advanced technology on the market today including airplanes, computers, or even nuclear reactors require the scale, magnitude, and range of scientific, physical, and engineering knowledge that fuel cell technology requires.’ Why is it that fuel cells are always a ‘few years away’ from commercialization, and why, despite the injection of massive amounts of funding for research and development, do they never seem to quite make it past the demonstration stage — except in a few high‐value niche markets? The plain fact is that fuel‐cell systems are complex and therefore require the input of expertise from many disciplines in science, engineering and technology. In particular, it has often proved difficult to bring together the necessary skills to focus on the fundamental issues that could lead to the required reduction in costs. The reasons for this situation are many; for example, there has often been a lack of sufficient funding sustained over long periods and an impatience on the part of developers to launch a product to market, even when there is no apparent market pull. Collectively, the various chapters of this book have assessed the status of various types of fuel‐cell system in terms of their design, components, mode of operation and performance. Although much is understood by developers in terms of system engineering, there is still considerably more to learn about the processes taking place in the individual fuel cell at the atomic and molecular level. What, for example, are the physical and chemical principles that control the exchange‐current density of a catalyst material, and why is platinum such an active catalyst for hydrogen oxidation compared with other metals? In the early days of fuel cells, in the 19th century, electrochemistry was just emerging, and the physics and chemistry known at the time was unable to address such questions. In more recent years, with the advent of powerful techniques such as X‐ray photoelectron spectroscopy (XPS), which is able to identify surface species on catalyst materials, a greatly enhanced understanding of the processes occurring within fuel cells has been achieved. For example, XPS has been applied to elucidate how non‐platinum group catalysts function for PEMFCs. Another tool available to the scientist in the 21st century that did not exist beforehand is the computing power necessary to carry out complex mathematical calculations. For instance, this facility enables chemical and physical processes to be modelled at the atomic and molecular level through the application of quantum mechanical methods such as density functional theory (DFT). Without such detailed basic research, it is feared that no matter how much money and time may be poured into product development, progress will be haphazard and ultimate success will be doubtful. Consequently, practical and affordable fuel cells will remain always ‘a few years away’. The Complete System and Its Future 12.5.2 Commercial Progress The attributes of fuel‐cell systems — high efficiency, low emissions, silent operation — that were touted as unique features by their pioneering proponents did not prove sufficient to establish the technology as an alternative to other forms of power generation. A breakthrough came with the PAFC in the 1990s when another beneficial property became apparent, namely, that of reliability. With banks and other financial institutions requiring ‘5 nines’ reliability, i.e., 99.999%, to avoid costly power outages, a PAFC (or preferably two running in parallel to provide some redundancy) could easily achieve this target. The high capital cost of the PAFC systems was small in comparison with the losses that would arise from even a few seconds of power outage for the financial institutions. A business case for installing systems could therefore be made. More recently, a PEMFC with stored hydrogen has been shown to operate in sub‐zero temperatures, and thereby a compelling case can be made for forklift trucks to be powered by fuel cells in refrigerated warehouses. Both of these high‐quality power and material‐handling applications are niche markets where the fuel cell has gained some commercial status. In many other cases, notably road vehicles, similar success has yet to be achieved. In 1998, Daimler‐Benz announced that it would be producing between 40 000 and 100 000 FCVs by 2004.15 This was shortly after commercialization agreements had been reached with Ballard Power Systems and other parties. At that time, Ballard was concentrating on demonstrating their fuel‐cell systems in buses. Progress has been much slower than expected — in terms of public transport, at the end of 2015 there were globally less than 200 fuel‐cell buses operating and less than 50 minibuses. At the same time, fewer than 3000 fuel‐cell cars were on the roads worldwide, with OEMs promoting the cars on a city‐by‐city basis in Japan, the United States (notably California) and the EU, rather than by country.16 A good case may be made for increased production of fuel‐cell buses, and if progress continues at the present rate, there are encouraging prospects that the cost could be brought down to a competitive position. Vehicle manufacturers recognize that, important though they are, battery electric vehicles will not address all the needs of the private motorist. Accordingly, there is renewed commitment to fuel‐cell cars by most of the major automakers and growing optimism that such vehicles will prove cost‐effective once fuelling issues have been addressed. A case can also be made for fuel‐ cell systems in rail transport. For instance, in May 2015, Hydrogenics Corporation signed a 10‐year agreement to supply PEMFC systems to Alstom Transport, the France‐based train manufacturer. This has led to the unveiling of the Alstom Coradia iLINT, a hydrogen‐ fuelled multiple unit, in September 2016; see Figure 12.31. As of January 2018, Alstom is building 14 trains for deployment from December 2021 in Lower Saxony. The long time that fuel cells are taking to transform into commercial products should be put into perspective by making a comparison with other game‐changing or disruptive technologies. Perhaps one of the most informative examples is the evolution of PV cells. Interestingly, it was in 1839 — the same year Grove assembled the first fuel cell — that a 19‐year‐old Frenchman, Edmond Becquerel, found that electricity could be produced directly from sunlight. He measured a voltage and could draw a current between two platinum electrodes immersed in an acidic solution of silver chloride when the solution 15 All, J., 1998, Auto makers race to sell cars powered by fuel cells, Wall Street Journal, 15 March 1998. 16 This information was provided by Kerry‐Ann Adamson of 4th Energy Wave, 2015. 395 396 Fuel Cell Systems Explained Fuel-cell composition Traction motor Traction inverter and DC/DC-converter Auxiliary converter Hydrogen fuel tank Battery composition Figure 12.31 The Alstom Coradia iLint zero‐emission fuel‐cell train. was illuminated by sunlight. This discovery was followed in 1873 by the English engineer Willoughby Smith who showed that selenium possesses photoconductivity. A decade later, the first selenium photocell was produced by the American inventor Charles Fritts, and it continued to be used as a light sensor until the 1960s. Meanwhile, the advent of a more practical PV cell awaited the research conducted by Gordon Teal and John Little at the Bell Telephone Laboratories (nickname ‘Bell Labs’), USA, in the early1950s.17 The two chemists were the first to grow single crystals of germanium and, later, silicon. Their work heralded the world of transistors and other semiconductors. Bell Labs exhibited the first high‐power silicon PV cell in 1954. The first satellite to use solar power, the US Vanguard 1, was launched in 1958; it employed a 100‐cm2 solar panel that delivered 0.1 W. Since then, solar cell technology has progressed in terms of enhanced efficiency, greater manufacturing ability, lower failure rate, improved lifetime and, most importantly, reduced capital cost. The Japanese company, Kyocera, was the first manufacturer in the world to mass-produce polycrystalline silicon solar cells, which were made via a casting method that is today’s industry standard. With commercial solar systems now achieving significant penetration in the electricity market, it is easy to overlook the long gestation period for the technology and to forget that the first niche market, in around 1978, was to power pocket calculators. It is also important to realize that, as with the different types of fuel cell, there are also several different types of silicon PV cell, e.g., crystalline, polycrystalline and amorphous. Each 17 Teal, GK and Little, JB, 1950, Growth of germanium single crystals, Physical Review, vol. 78, p. 647. The Complete System and Its Future version of silicon solar cell has had its own line of development and the progressive advancement of one technology has influenced another. Studies since 1990 have shown that the impressive cost‐reduction of silicon cells from US$76 per watt in 1977 to US$0.2 per watt in 2017 has taken place in a number of steps, as the science and manufacturing capability improved. Perhaps there is a lesson here for developers of fuel cells. 12.6 Future Prospects: The Crystal Ball Remains Cloudy A short history of the fuel cell in the 19th and early 20th centuries has been given in the opening chapter of this book. Much of the development of the technology took place in Europe and later in the United States. The work culminated with the fuel cells developed for the US space programme of the 1960s. United Technologies Corporation (UTC) was set up by the US‐based firm of Pratt and Whitney in 1958 to develop the alkaline fuel cell pioneered in the United Kingdom by Francis Bacon. In 1966, the company supplied the fuel cells to the National Aeronautics and Space Administration (NASA) for the Apollo project and later for the Space Shuttle missions until 2010. Having developed PAFCs during the 1970s and early 1980s, the UTC fuel‐cell team became focused on the technology business and in 1985 formed a wholly owned trading subsidiary under the name International Fuel Cells (IFC). The enterprise was subsequently renamed UTC Fuel Cells and finally UTC Power in 2001. The activities of UTC fuel cells were targeted towards stationary systems and resulted in the marketing of the PC25 200‐kW and later 400‐kW packaged systems. The company later expanded its interests to include PEMFCs for transport. Thus UTC Power could trace its history back to the pioneering work of Bacon, but by the early years of the 21st century, it was clear that the PAFCs produced by the company were not economically viable for most applications for which they were designed (i.e., cogeneration). Only in some niche markets (e.g., quality power for data centres and banks) could a robust business case be made. A brand new venture, Quantum Leap Technology, was formed in 2003 by the innovator Brett Vinsant who had developed a PEMFC in his garage in Hillsboro, Oregon. In August 2005, Quantum Leap changed its name to ClearEdge Power and directed its efforts to the development of novel PEMFC systems that were built using semiconductor technology. The company went through several rounds of seed and venture funding, so in May 2007, it had grown to having 20 employees and had raised US$10 million in venture capital. In early 2008, ClearEdge sold and installed its first fuel‐cell system and in January 2009 raised another US$11 million in venture capital. Doubling its workforce in less than 12 months, further successful funding rounds continued in 2010 with orders for 5‐kW backup power plants. In June 2010, ClearEdge signed a US$40 million deal to supply 800 fuel‐cell systems over a 3‐year period to LS Industrial Systems, a subsidiary of the South Korean LS Group. To undertake and complete this assignment, even more venture capital was raised (over US$70 million). ClearEdge presented itself as a rapidly growing clean technology start‐up business. Several other examples can be found in the fuel‐cell industry, some more successful than others. Companies cannot, however, continue to be supported by venture capital indefinitely. Hungry for expansion, ClearEdge acquired some of the assets of UTC Power in February 2013. In all such deals, there are winners and losers, and in this case 397 398 Fuel Cell Systems Explained it was UTC Power who ended up filing for bankruptcy protection in 2014. This story ended in July 2014 when ClearEdge was purchased by the Doosan Group, a South Korean conglomerate. The purpose of relating the fate of UTC Power and Clear Edge Power is not to highlight particular companies, but to show that the fuel‐cell industry is far from mature. It is a fragile industry because few of the companies have firm order books and the infrastructure to deliver mass‐produced products. A small number of the companies that are currently trading have focused on fuel‐cell systems for more than 25 years, e.g., Rolls‐ Royce Fuel Cells, Fuel Cell Energy and Ballard Power Systems. Many more companies are very much younger and have been spun out of research groups who have claimed a particular breakthrough, or they have been set up by entrepreneurs seeking to exploit the technology to fulfil a new business model in a carbon‐constrained world. Fuel‐cell technology — even the advanced PAFC and PEMFC systems — finds it difficult to compete in the market place with alternative entrenched power-generation technologies (e.g., gas micro‐turbines) despite some apparent advantages. The reader can explore the fate of other fuel‐cell companies, such as the various SOFC developers (e.g., Siemens, Westinghouse, Sulzer Hexis) and MCFC developers (MC‐Power and MTU‐Onsite), who have fallen by the wayside. For many years, particularly in the closing years of the 20th century, fuel‐cell systems were oversold by zealous engineers. Hype led to the prospect of commercial systems and unfulfilled expectations led to disillusionment by government and other funding agencies, and to an extent, the public at large. Companies and research organizations that remain active in the field continue to battle against competing technologies (e.g., the fuel‐cell car vs. the battery car,18 or stationary natural gas fuel‐ cell systems vs. gas turbines). At the same time, there is a realization that fuel cells will only achieve widespread commercial success as the understanding of the science and technology grows. When the first edition of this book was published (2000), the expected lifetime of a PEMFC stack was around 2000 h. As the mechanisms of cell degradation have become more understood so has the proven lifetime increased – it can now be in excess of 20 000 or even 100 000 h in stationary energy-storage applications. At the same time, improved materials and manufacturing methods are helping to reduce costs. Consequently, the automotive industry has been encouraged to introduce FCVs such as the Nexo, which has an impressive range of 600 km and was launched by Hyundai in early January 2018. We hope, therefore, that this third edition will provide a timely account of the basic science and technology of fuel-cell systems, together with an understanding of their challenges and prospects. 18 There is a somewhat spurious argument between proponents of battery electric vehicles and fuel‐cell vehicles. Both are electric vehicles, have much in common and are anticipated to satisfy different market sectors. Battery‐only electric vehicles are likely to be commercially viable for short journeys given the required weight of batteries and limitations in the rate of charging. Hydrogen fuel‐cell vehicles, by contrast, can be charged quickly and can be expected to have a longer range. The Complete System and Its Future Further Reading Bagotsky, VS, 2012, Fuel Cells: Problems and Solutions, 2nd ed., John Wiley & Sons, Inc., Hoboken, NJ. Behling, N, 2012, Fuel Cells: Current Technology and Future Research Needs, Elsevier, Burlington, MA. Dell, RM and Rand, DAJ, 2004, Clean Energy, The Royal Society of Chemistry, Cambridge. Dell, RM, Moseley, PT and Rand, DAJ, 2014, Towards Sustainable Road Transport, Elsevier, Amsterdam. Evers, AA, 2010, The Hydrogen Society…More Than Just a Vision? Hydrogeit Verlag, Oberkraemer. Kulikovsky, AA, 2010, Analytical Modelling of Fuel Cells, Elsevier BV, Amsterdam. Rand, DAJ and Dell, RM, 2008, Hydrogen Energy: Challenges and Prospects, The Royal Society of Chemistry, Cambridge. Romm, JJ, 2005, The Hype about Hydrogen: Fact and Fiction in the Race to Save the Climate, Island Press, Washington, DC. Schewe, PF, 2007, The Grid: A Journey Through the Heart of Our Electrified World, Joseph Henry Press, Washington, DC. Scott, DS, 2007, Smelling Land: The Hydrogen Defense Against Climate Catastrophe, Canadian Hydrogen Association, Vancouver, BC. Sigfusson, TI, 2008, Planet Hydrogen — The Taming of the Proton, Coxmoor Publishing Company, Oxford. Sperling, D and Cannon, JS, 1994, The Hydrogen Energy Transition: Cutting Carbon from Transportation, Elsevier Academic Press Pt Inc., San Diego, CA. Topler, J and Jochen, L, 2016, Hydrogen and Fuel Cell: Technologies and Market Perspectives, Springer, Berlin. ISBN: 978‐3662449714. Weiss, MA, Heywood, JB, Drake, EM, Schafer, A and Yeung, AFF, 2000, On the Road on 2020 A Life‐Cycle Analysis of New Automobile Technologies, Energy Laboratory Report # MIT EL 00‐003, Energy Laboratory, Massachusetts Institute of Technology, Cambridge, MA, 02139–04307. 399 401 Appendix 1 Calculations of the Change in Molar Gibbs Free Energy A1.1 Hydrogen Fuel Cell This appendix shows how to calculate the change in molar Gibbs free energy for the reaction: H2 1 2 O2 H2O (A1.1) The Gibbs free energy (G) of a system (also known as Gibbs energy or Gibbs function) is defined in terms of the enthalpy (H), temperature (T ) and entropy (S) according to the relationship: G (A1.2) H TS Similarly, the molar Gibbs free energy of formation ( g ), the molar enthalpy of formation (h f ) and the molar entropy (s)1 are connected by the equation: f gf hf (A1.3) Ts In the case of the hydrogen oxidation reaction (A1.1), it is the change in energy that is important, i.e., the difference in energy between the reactants (hydrogen and oxygen) and the product (water or steam). Also, in a fuel cell, the temperature can be taken as constant2 and, therefore, the following holds: gf hf (A1.4) T s The value of ∆h f is the difference between h f of the products and h f of the reactants. Thus, for the hydrogen oxidation reaction: hf hf H2 O hf 1 H2 2 hf O2 (A1.5) 1 Because entropy can be measured as an absolute value, that is, not relative to those of the elements in their reference states, there is no need to use the term ‘entropy of formation’, but simply use the absolute entropies for products and reactants. 2 The heat generated by a fuel cell would give rise to an increase in temperature were it not for the fact that the cell is cooled by the cathode air, by virtue of conducting internal reforming or by a combination of methods. Enough cooling must be applied to ensure that there are no large temperature gradients that could cause stresses and therefore degradation of cell materials. Given these requirements, to a first approximation, the cell may be considered to be at constant temperature. Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 402 Appendix 1: Calculations of the Change in Molar Gibbs Free Energy Similarly, Δs is the difference between s of the products and s of the reactants. Consequently, for the reaction under consideration: s s H2 O s 1 H2 2 s (A1.6) O2 The values of h f and s vary with temperature according to equations (A1.7) and (A1.8) given in the following text. These standard equations are derived using thermodynamic theory, and their proof can be found in textbooks on engineering thermodynamics.3 The subscript to h and s is the temperature, and cP is the molar heat capacity at constant pressure. Standard temperature is taken as 298.15 K. The molar enthalpy of formation at temperature T is given by T hT h298.15 (A1.7) cP dT 298.15 The molar entropy is given by: T sT s298.15 1 cP dT T 298.15 (A1.8) The values for both the molar enthalpy of formation and the molar entropy of formation at 298.15 K are obtainable from thermodynamics tables. Typical data at standard pressure are given in Table A1.1.4 The molar heat capacity varies with temperature. To use equations (A1.7) and (A1.8), therefore, it is necessary to know the c p values at constant pressure over a range of temperatures. Fortunately, empirical equations for the dependence of c p on temperature are available in many thermodynamics texts.5 The results given by the following three equations are accurate to within 0.6% over the range 300–3500 K. Table A1.1 Values of hf are in J mol−1 and those for s in J mol−1 K−1, both at 298.15 K, for the hydrogen fuel‐cell reaction (A1.1). hf s H2O (liquid) −285 838 70.05 H2O (steam) −241 827 188.83 H2 0 130.59 O2 0 205.14 3 For example: Balmer, RT, 2011, Modern Engineering Thermodynamics, Academic Press, New York; Smith, JM, Van Ness, HC and Abbott, MN, 2005, Introduction to Chemical Engineering Thermodynamics, 7th edition, McGraw Hill Higher Education, Boston, MA. 4 From: Keenan, JH and Kaye, J, 1948, Gas Tables, John Wiley & Sons, Inc., New York. 5 For example: Van Wylen, GJ and Sonntag, RE, 1986, Fundamentals of Classical Thermodynamics, 3rd edition, John Wiley & Sons, Inc., New York, p. 688. Appendix 1: Calculations of the Change in Molar Gibbs Free Energy Table A1.2 Sample values for Δhf and Δgf in J mol−1, and Δs in J mol−1 K−1, for the reaction (A1.2). Temperature Δhf Δgf Δs 100 −242.6 −0.0466 −225.2 300 −244.5 −0.0507 −215.4 500 −246.2 −0.0533 −205.0 700 −247.6 −0.0549 −194.2 900 −248.8 −0.0561 −183.1 Temperatures are in celsius. For steam: 143.05 58.040T 0.25 8.2751T 0.5 0.036989T cP (A1.9) For hydrogen: 56.505 22 222.6T cP 0.75 116 500T 1 560 700T 1.5 (A1.10) For oxygen: 37.432 2.0102 10 5T 1.5 178 570T cP 1.5 2 368 800T 2 (A1.11) All of the equations for cP are in J mol−1 K−1. The values can be substituted into equations (A1.7) and (A1.8) to yield functions that can be readily integrated and thus evaluated at any temperature T. This mathematics is undertaken to derive values for Δh and Δs for steam, hydrogen and oxygen. The values are then substituted into equations (A1.5) and (A1.6) to give values for ∆h f and Δs that are finally substituted into equation (A1.4) to calculate the change in molar Gibbs energy of formation, ∆g f . Sample values are given in Table A1.2. For liquid water, the standard values for h f and s at 25°C are taken from Table A1.1. To find h f s and s at 80°C, equations (A1.7) and (A1.8) are again employed, but since the temperature range (25–80°C) is small, it can be assumed that cP is constant. A1.2 Carbon Monoxide Fuel Cell It is possible that in the high‐temperature fuel cells introduced in Chapter 6, the carbon monoxide generated from steam reforming of a fuel (e.g., methane) is directly oxidized. The reaction is: CO 1 2 O2 CO2 (A1.12) The method used, and the theory employed, for calculating the change in molar Gibbs free energy, is exactly the same as for the hydrogen fuel cell, except that the equations are 403 404 Appendix 1: Calculations of the Change in Molar Gibbs Free Energy Table A1.3 Values of hf in J mol−1 and s in J mol−1 K−1, both at 298.15 K, for the carbon monoxide fuel‐cell reaction (A1.12). hf O2 s 0 205.14 CO −110 529 197.65 CO2 −393 522 213.80 altered to fit the new reaction. Oxygen features as shown in equation (A1.11), and the values of the molar heat capacity for carbon monoxide and carbon dioxide are given by: For carbon monoxide: cP 69.145 0.022282T 0.75 2007.7T 0.5 5589.64T 0.75 (A1.13) For carbon dioxide: cP 3.7357 3.0529T 0.5 0.041034T 2.4198 10 6 T 2 (A1.14) Together with values from Table A1.3, these equations are used with equations (A1.7) and (A1.8) to determine the molar enthalpies and molar entropies for the three gases; see Table A1.3. The change in the molar enthalpy and the molar entropy is then determined, respectively, from the following two equations: hf s hf s hf CO2 CO2 s 1 CO 1 CO 2 s 2 hf (A1.15) O2 (A1.16) O2 The change in molar Gibbs free energy of formation is then calculated, as for the hydrogen fuel cell, via equation (A1.4). Some example results are given below in Table A1.4. Table A1.4 Sample values for Δhf and Δgf in J mol−1, and Δs in J mol−1 K−1, for the carbon monoxide fuel‐cell reaction (A1.12). Temperature Δhf Δgf Δs 100 −283.4 −250.7 −0.0877 300 −283.7 −232.7 −0.0888 500 −283.4 −214.6 −0.0890 700 −281.8 −196.5 −0.0877 900 −281.0 −178.5 −0.0822 Temperatures are in celsius. 405 Appendix 2 Useful Fuel‐Cell Equations A2.1 Introduction This appendix presents the derivation of useful equations that relate to the following fuel‐cell parameters: ● ● ● ● ● ● Oxygen and air usage rate. Inlet air flow rate. Exit air flow rate. Hydrogen usage and energy content of hydrogen. Rate of water production. Heat production. The term ‘stoichiometric’ is encountered in the ensuing discussion. Its meaning could be defined as ‘just the right amount’. For instance, in the simple fuel‐cell reaction: H2 1 2 O2 H2O (A2.1) exactly two moles of hydrogen would be provided for each mole of oxygen. These reagents would produce exactly 4F of charge since 2 moles of electrons are transferred for each mole of hydrogen. Either hydrogen, oxygen or both are often supplied at greater than the stoichiometric rate. This is especially the case with oxygen if it is being supplied as air. Otherwise, the air leaving the cell would be completely devoid of oxygen. Note also that reactants cannot be supplied at less than the stoichiometric rate. The stoichiometry can be expressed as a variable that is normally denoted by the symbol λ. Thus, if the rate of usage of a chemical in a reaction is ṅ moles per second, then the rate of supply is λṅ moles per second. To increase the usefulness of the equations derived in the following, they have been formulated in terms of the electrical power of the whole fuel‐cell stack, Pe, and the average voltage of each cell in the stack, Vc. The electrical power will nearly always be known because it is the most basic and important information about a fuel‐cell system. If Vc is not given, it can be assumed to be between 0.6 and 0.7 V — most fuel cells operate in this region (see Figures 3.1 and 3.2, Chapter 3). The value of Vc could be calculated by using equation (2.5), Chapter 2, provided the efficiency is known; otherwise, taking Vc as 0.65 V would be a good approximation. If, however, the fuel cell is pressurized, then a somewhat higher estimate of Vc should be taken. Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 406 Appendix 2: Useful Fuel‐Cell Equations A2.2 Oxygen and Air Usage In the basic operation of the hydrogen fuel cell, four electrons are transferred for each mole of oxygen; see equation (1.3), Chapter 1. Hence, for a single cell: Charge transferred 4 F amount of O2 (A2.2) Dividing by time and the rearranging gives: Oxygen usage I mol s 4F 1 (A2.3) where I represents the current. For a stack of n cells: Oxygen usage In mol s 4F 1 (A2.4) It would be more useful, however, to have the formula in kg s−1 so that it is not necessary to know the number of cells, as well as in terms of power rather than current. If the voltage of each cell in the stack is Vc, then: Power, Pe Vc I n (A2.5) Thus, the current is given by I Pe Vc n (A2.6) Substituting this expression into equation (A2.4) gives Oxygen usage Pe mol s 4Vc F 1 (A2.7) Changing from mol s−1 to kg s−1 Oxygen usage 32 10 3 Pe 4 Vc F 8.29 10 Vc 8 Pe kg s 1 (A2.8) This formula allows determination of oxygen usage by any fuel‐cell system of given power. When Vc is not known, it can be calculated from the efficiency, and if this parameter is not given, as noted earlier, 0.65 V can be used as a good approximation. The oxygen will normally be supplied in the form of air, and therefore it is necessary to adapt equation (A2.7) to air usage. The molar proportion of air that is oxygen is 0.21, and the molar mass of air is 28.97 × 10−3 kg mol−1. Consequently, equation (A2.7) becomes: Air usage 28.97 10 3 Pe 0.21 4 Vc F 3.58 10 Vc 7 Pe kg s 1 (A2.9) If the air was used at this rate, then as it left the cell it would be completely devoid of any oxygen. This is impractical and consequently the airflow is set at a value well above Appendix 2: Useful Fuel‐Cell Equations the stoichiometric requirement — typically, twice as much. If the stoichiometry is λ, equation (2.9) becomes Air usage 3.58 10 7 Vc Pe kg s 1 (A2.10) The kilogram per second is not, in fact, a very commonly used unit of mass flow. The following conversions of the mass flow unit to volume under standard conditions will be found more useful. The mass flow rate from equation (A2.10) should be multiplied by: ● ● ● ● 3050 to give flow rate in standard m3 h−1. 1795 to give flow rate in standard ft3 min−1, abbreviated as SCFM (i.e., standard cubic foot per minute). 5.1 × 104 to give flow rate in standard L min−1. 847 to give flow rate in standard L s−1. A2.3 Exit Air Flow Rate It is sometimes important to distinguish between the inlet flow rate of the air, which is given by equation (A2.10), and the outlet flow rate. This is particularly important when calculating the humidity, which is an issue in certain types of fuel cell — especially proton‐exchange membrane fuel cells (PEMFCs). The difference is caused by the consumption of oxygen. There will usually be more water vapour in the exit air, but ‘dry air’ is being considered at this stage of the discussion. Water production is examined later in Section A2.5. Clearly: Exit air flow rate inlet air flow rate oxygen usage (A2.11) Using equations (A2.10) and (A2.8), equation (A2.11) becomes Exit air flow rate 3.5 10 A2.4 7 Pe Vc 8.29 10 8 Pe kg s Vc 1 (A2.12) Hydrogen Usage The rate of usage of hydrogen is derived in a way similar to that for oxygen, except that there are two electrons from each mole of hydrogen. Equations (A2.4) and (A2.7) thus become, respectively: H2 usage In mol s 2F H2 usage Pe mol s 2Vc F 1 (A2.13) 1 (A2.14) 407 408 Appendix 2: Useful Fuel‐Cell Equations Table A2.1 Energy content of hydrogen fuel expressed in different forms. Form Energy content Specific enthalpy (HHV) 1.43 × 108 J kg−1 Specific enthalpy (HHV) 39.7 kWh kg−1 Effective specific electrical energy 26.8 × Vc kWh kg−1 Energy density as STP (HHV) 3.20 kWh m–3 = 3.20 Wh SL−1 Energy density as STP (HHV) 3.29 kWh m–3 = 3.29 Wh SL−1 SL, standard litre. To obtain the lower heating value (LHV), multiply the HHV by 0.846. The molar mass of hydrogen is 2.02 × 10−3 kg mol−1, so equation (A2.14) becomes: H2 usage 2.02 10 3 Pe 2Vc F 1.05 10 8 Pe kg s Vc 1 (A2.15) under stoichiometric conditions. Obviously, this formula only applies to a hydrogen‐fed fuel cell. In the case of a mixture of hydrogen and carbon monoxide derived from a reformed hydrocarbon, the situation will be different and dependent on the proportion of carbon monoxide present. The result can be transformed to a volume rate by using the density of hydrogen, which is 0.084 kg m−3 at normal temperature and pressure (NTP, 293.15 K and 1 atm). In addition to the rate of consumption of hydrogen, it is often also important to know the electrical energy that could be produced from a given mass or volume of hydrogen. The list in Table A2.1 gives the energy in kilowatt‐hours (kWh), rather than in Joules, as this measure is commonly used for electrical power systems. In addition to the ‘raw’ energy per kilogram and standard litre, there is an ‘effective’ energy that takes into account the efficiency of the cell and is expressed in terms of Vc, the mean voltage of each cell. If efficiency of a hydrogen fuel cell has to be taken into account, then the formula for efficiency derived in Section 2.5, Chapter 2, can be used, namely: Efficiency Vc 1.48 (A2.16) Note that in equation (A2.16), the term for fuel utilization is not included as most pure hydrogen fuel cells will be assumed to run with 100% fuel utilization. A2.5 Rate of Water Production In a hydrogen‐fed fuel cell, water is produced at the rate of one mole for every two electrons (see Section 1.1, Chapter 1) and can be expressed by adapting equation (A2.7) to obtain: Water production Pe mol s 2Vc F 1 (A2.17) Appendix 2: Useful Fuel‐Cell Equations The molecular mass of water is 18.02 × 10−3 kg mol−1; therefore: Water production 9.34 10 Vc 8 Pe kg s 1 (A2.18) In a hydrogen‐fed fuel cell, the rate of water production is approximately stoichiometric. If, however, the fuel is a mixture of carbon monoxide with hydrogen, then the water production would be less, namely, in proportion to the amount of carbon monoxide present in the mixture. For a hydrocarbon fuel that was internally reformed, some of the product water would be used in the reformation process. For instance, it was shown in Chapter 9 that if methane is internally reformed, then half the product water is used in the reforming process, thus halving the exit rate of water from the fuel cell. By way of an example, consider a 1‐kW fuel cell that operates for 1 h at a cell voltage of 0.7 V. This performance corresponds to an efficiency of 47% (with respect to the HHV), as given by equation (A2.16). Substituting this value into equation (A2.17) gives: Water production 9.34 10 8 1000 1.33 10 4 kg s 0.7 1 (A2.19) It follows that the mass of water produced in 1 h is 1.33 × 10−4 × 60 × 60 = 0.48 kg. Since the density of water is 1.0 g cm−3, this mass corresponds to 480 cm3. As a rough guide, therefore, a 1‐kW fuel cell will produce about 0.5 L of water per hour. A2.6 Heat Production Heat is produced when a fuel cell operates. It was noted in Section 2.4, Chapter 2, that if all the enthalpy of reaction of a hydrogen fuel cell was converted into electrical energy, then the output voltage would be 1.48 V or 1.25 V if the water product was in liquid form or vapour form, respectively. Cleary, it follows that the difference between the actual cell voltage and either of these two voltages represents the energy that is not converted into electricity, that is, the energy that is transformed into heat. Since, in most cases, the water is not produced in liquid form, the following analysis assumes that product water is in the vapour phase and that no account is taken of the cooling effect of water evaporation. It also means that energy is leaving the fuel cell in three forms, namely, as electricity, as ordinary ‘sensible’ heat and as the heat of vapourization (latent heat) of water. For a stack of n cells at a current I, the heat generated, in watts, is given by: Heating rate nI 1.25 Vc (A2.20) In terms of electrical power, in watts, this becomes: Heating rate Pe 1.25 1 Vc (A2.21) 409 411 Appendix 3 Calculation of Power Required by Air Compressor and Power Recoverable by Turbine in Fuel‐Cell Exhaust A3.1 Power Required by Air Compressor The following worked example for compressor power is for a 100‐kW fuel‐cell stack pressurized at 300 kPa (3 bar). Air is fed to the stack using the Lysholm compressor whose chart is shown in Figure 12.4, Chapter 12. The air inlet to the compressor is at 100 kPa (1 bar) and 20°C. The fuel cell is operated at an air stoichiometry of 2.0, and the average cell voltage is 0.65 V, which corresponds to an efficiency of 52% (LHV). The chart of Figure 12.4, Chapter 12, is used to determine the values of the following parameters: ● ● ● ● Required rotational speed of the air compressor. Efficiency of the compressor. Temperature of the air as it leaves the compressor. Power of the electric motor required to drive the compressor. First, it is necessary to find the mass flow rate of air that will be consumed by the cell using equation (A2.9): Air usage 3.58 10 7 2 100 000 = 0.11 kg s 0.65 1 (A3.1) This value is then converted to the mass flow factor: Mass flow factor 0.11 293 1 = 1.9 kg s 1 K 2 bar 1.0 1 (A3.2) Note that the pressure units are in bar, i.e., the same as in the chart given in Figure 12.4, Chapter 12. The chart can now be used to determine the speed and efficiency of the compressor, that is, the intercept of a horizontal line drawn from pressure ratio =3 and a vertical line starting from the x‐axis at a mass flow factor =1.9. The result will be very close to the 600 rotor speed factor line and the 0.7 ‘efficiency contour’. Thus, the rotor speed can be taken to be: 600 293 10 300 rpm Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. (A3.3) 412 Appendix 3: Calculation of Power Required by Air Compressor and Power Recoverable by Turbine in Fuel‐Cell Exhaust The efficiency of the compressor and the mass flow rate are used to find the temperature rise and the compressor power. The former is obtained from equation (12.6), Chapter 12, namely: T 293 0.7 30.286 1 155 K (A3.4) Since the entry temperature is 20°C, the exit temperature is therefore 175°C. Note that if the system is a PEMFC, cooling will be necessary. Alternatively, if it is a PAFC, then the compressor would enable the fuel gas to be preheated. The power required for the compressor can be determined from equation (12.10), Chapter 12: Power 1004 293 0.7 30.286 1 0.11 17.1 kW (A3.5) This is the power for the compressor without considering any mechanical losses in the bearings and driveshafts. The electric motor also will not be 100% efficient — a reasonable estimate of its power would be about 20 kW. It is important to note the following: ● ● The 20 kW of electrical power will have to be provided by the 100‐kW fuel cell, i.e., by consuming 20% of its output. This parasitic load is a major problem when running systems at pressure; its importance for PEMFCs is discussed in Section 4.7.2, Chapter 4. In this worked example, the assumption is that the air is not humidified, i.e., it has a low water content. As pointed out in Section 4.4, Chapter 4, the inlet of a PEMFC is sometimes humidified. This action alters both the specific heat capacity and the ratio of the heat capacities, γ, and will influence the performance of the compressor. Humidification, if required, is usually undertaken after compression because the air is hotter at this stage. A3.2 Power Recoverable from Fuel‐Cell Exhaust with a Turbine The power available from the exit gases of the 100‐kW fuel cell, and recoverable by using a turbine, can be found as follows. The mass of the cathode exit gas is increased by the presence of water in the cells, but since this is the result of replacing O2 with 2H2O, the mass change will be insignificant, as the mass of hydrogen is so small. The mass flow rate, ṁ , will therefore still be taken as 0.11 kg s−1. The exit temperature can be estimated as 90°C for a typical PEMFC, and the entry pressure is 300 kPa (3 bar). The exit pressure must be a little less than this, and assuming it to be 280 kPa, the mass flow factor can be calculated as: Mass flow factor 0.11 363 2.8 0.75 kg s 1 K1/2 bar 1 (A3.6) Appendix 3: Calculation of Power Required by Air Compressor and Power Recoverable by Turbine in Fuel‐Cell Exhaust The speed and efficiency of the turbine can be determined from the performance chart given in Figure 12.8, Chapter 12. The intercept on the chart between 0.75 on the x‐axis and 2.8 on the pressure ratio axis is close to the rotor speed factor line of 5000, and in the efficiency region of 0.7 or 70%. Consequently, the required rotor speed is predicted to be: 5 000 363 95 000 rpm (A3.7) This very high speed is suitable for directly driving a centrifugal compressor on the same shaft, but not for a screw compressor. The power available from the turbine can be obtained from equation (12.10), Chapter 12, i.e., 1 Power C P T1 C P2 P1  1 m (12.10) The exit gas is not normal air; it has less oxygen and a changed specific heat capacity. For engines, standard values are 1150 J kg−1 K−1 for CP and 1.33 for γ. In the case of a fuel cell, the change in gas composition is not so great, and a value of 1100 J kg−1 K−1 will be used for CP and 1.33 for γ. The constant (γ − 1/γ) thus becomes 0.275. The temperature T1 is 363 K, so equation (12.10) becomes: Power available 100 0.7 363 10.275 1 2.8 0.11 7.6 kW (A3.8) The minus sign indicates that power is given out by the turbine. This power is a useful addition to the 100 kW of electrical output of the fuel cell, but note that it provides less than half of the power required to drive the compressor, as calculated above. Furthermore, this example is the best possible result — turbine efficiencies will usually be somewhat lower than the 0.7 assumed here. As can be seen from the turbine performance chart in Figure 12.9, Chapter 12, much of the operating region is at greatly lower efficiency. 413 415 Glossary of Terms AB5 A range of metal alloys (e.g., LaNi5) capable of undergoing a reversible hydrogen absorption–desorption reaction. Absorption The process by which a liquid or gas is drawn into the permeable pores of a solid material. See: Adsorption; Chemisorption; Physisorption. Activation energy The energy needed to initiate a chemical reaction. Also known as ‘Arrhenius energy’. Activation overpotential The overpotential that results from the restrictions imposed by the kinetics of charge transfer at the electrode|electrolyte interface. Activity A measure of the ‘effective concentration’ of a species in a reacting system. By convention, it is a dimensionless quantity. The activity of pure substances in condensed phases (liquids or solids) is taken as unity. Activity depends principally on the temperature, pressure and composition of the system. In reactions involving real gases and mixtures, the effective partial pressure of a constituent gas is usually referred to as ‘fugacity’. See: Fugacity. Adiabatic process A process (e.g., expansion of a gas) that takes place without heat entering or leaving the system; in reversible adiabatic expansion, as a gas cools, its internal energy is reduced by the amount of work done by the gas on the environment. Adsorbate A material that has been or is capable of being adsorbed. Adsorbent A material having capacity or tendency to adsorb another substance. Adsorption The adhesion of molecules of gases, dissolved substances or liquids to the surface of solids or liquids with which they are in contact; distinguished from absorption, a process in which one substance actually penetrates into the inner structure of the other. Thus, adsorb and adsorbent. See: Chemisorption; Physisorption. Alanates Aluminium hydrides of alkali metals or alkaline earth metals, e.g., LiAlH4, NaAlH4, Mg(AlH4)2. Alternating current Electric current that flows for an interval of time (half‐period) in one direction and then for the same time in the opposite direction; the normal waveform is sinusoidal. Alternating current is easier to transmit over long distances than direct current, and it is the form of electricity used in most homes and business. See: Direct current. Amines Organic compounds that contain nitrogen as a key atom. The compounds are similar in structure to ammonia, with one or more of the hydrogen atoms replaced by organic groups, and exhibit a wide range of properties. Amine scrubbing is commercially used for the removal of carbon dioxide from natural gas. Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 416 Glossary of Terms Amorphous material A solid material in which there is no long‐range order of the positions of the atoms. Anaerobic Any process (usually chemical or biological) that takes place without the presence of air or oxygen. Anode An electrode at which an oxidation process, i.e., loss of electrons, is occurring. In a fuel cell, the anode is the negative electrode where hydrogen is consumed. During electrolysis, the anode is the positive electrode where oxygen is evolved. In a secondary battery (or cell), the anode is the positive electrode on charge and the negative electrode on discharge. Anthracite The highest rank of coal in terms of hydrocarbon content (carbon content 85–95 wt.%) with a glossy, black appearance. It is used primarily for residential and commercial space heating. Also known as ‘hard coal’. See: Bituminous coal; Lignite; Peat; Sub‐bituminous coal. Anthropogenic emission Emission caused, directly or indirectly, by human activities. The emission of sulfur dioxide due to the use of fossil fuels is an example of a direct cause of emissions, and the emission of nitrogen oxides from farmland as a function of fertilizer application is an example of an indirect cause. Area‐specific resistance The electrical resistance of a sample multiplied by its geometric area. Austenitic steel Steel alloys based on austenitic iron (γ‐phase iron). Austenitic stainless steel contains a maximum of 0.15 wt.% carbon, a minimum of 16 wt.% chromium and sufficient nickel and/or manganese to retain its elasticity even at cryogenic temperatures. Autothermal reforming An energy‐efficient reforming process that uses heat generated from combining partial oxidation and catalytic steam reforming of a hydrocarbon feedstock (methane or liquid fuel) in a single step; this can significantly reduce emissions of carbon dioxide. See: Partial oxidation; Steam reforming. Balance-of-plant The sum of those components additional to and integrated with the primary power module of a fuel cell (which is individual fuel cells connected in a stack) to make up the entire operational system. These components can include a fuel processor or fuel reformer, power‐conditioning equipment (such as inverters and voltage controls), motors, compressors, blowers and fans, valves and piping, fuel storage medium and even conventional batteries complementary to the fuel‐cell stack. Barrel A measure of crude oil (petroleum), approximately 159 dm3. Base load The typical minimum electrical power demand placed on a power‐ generating system. Battery A multiple of electrochemical cells of the same chemistry, connected in series or in parallel and housed in a single container. (Note that the term is often used to indicate a single cell, particularly in the case of primary systems.) BET‐specific surface area The total surface area of a specimen per unit of mass, usually expressed in m2 g−1, obtained by applying the Brunauer–Emmett–Teller model to gas adsorption isotherms. Pore volume and pore size-distribution can also be obtained by this method. Binder A substance added to the active material of an electrode to enhance mechanical strength. Bio‐electrochemical fuel cell A fuel cell that exploits biological species as catalysts to facilitate the generation of electricity. There are two major classes: ‘enzymatic’ fuel cells and ‘microbial’ fuel cells. The former employ enzymes as catalysts, whereas Glossary of Terms the latter employ microorganisms to convert the chemical energy of biofuels (e.g., glucose, other sugars, alcohols) into electrical energy or hydrogen. Also known as a ‘biological fuel cell’. Biofuel A gaseous, liquid or solid fuel that is derived from a biological source. Biofuel may be in its natural form (e.g., wood, peat) or a commercially produced form (e.g., ethanol from sugarcane residue, diesel fuel from waste vegetable oils). Biogas A gaseous fuel of medium energy content, composed of methane (typically 50–60 vol.%) and carbon dioxide, that results from the anaerobic decomposition of waste matter. Also known as ‘anaerobic digester gas’. See: Anaerobic. Biomass A collective term used to describe all biologically produced matter at the end of its life that can be converted to a solid fuel, a renewable liquid fuel (‘biofuel’) or a gaseous fuel (‘biogas’, such as methane or hydrogen). Biomass can be derived from forest and mill residues, agricultural crops and wastes (e.g., corn stover, alfalfa stems, obsolete seed corn, hulls and nut shells, fibre from sugarcane, straw from rice and wheat), wood and wood wastes (sawdust, timber slash mill scrap), animal wastes, livestock operation residues, aquatic plants, fast‐growing trees and plants and municipal and industrial wastes. See: Biofuel; Biogas. Biophotolysis The storage and use of electrons produced by the first stages of photosynthesis that can then be used to produce free hydrogen. Bipolar plate A dense electronic (but not ionic) conductor that electrically connects the positive electrode in one cell to the negative in the adjacent. The cells are series connected and so allow the voltage to be built up. Bipolar plates also serve as a means to distribute fuel or air to the electrodes, to remove reaction products and to transfer heat. Depending on the type of electrochemical cell, the plate may be made out of carbon, metal or a conductive polymer (which may be a carbon‐filled composite). See: End-plate; Flow-field. Bituminous coal A dense coal (carbon content 45–85 wt.%) that is black, but sometimes dark brown, often with well‐defined bands of bright and dull material. It serves primarily as fuel in electricity generation, with substantial quantities also used for heat and power applications in manufacturing and to make coke or coking coal, an essential ingredient in making steel. See: Anthracite; Lignite; Peat; Sub‐bituminous coal. Butler–Volmer equation The relationship between the current flowing through an electrode and the potential across the electrode|electrolyte solution interface. At low overpotentials, it can be very well approximated by a linear relationship, and at high overpotentials by the Tafel equation. See: Tafel equation. Calcination Thermal treatment process for solids to induce thermal decomposition or phase transitions or eliminate volatile components. Capacitance or Capacity The electric charge stored in a capacitor, measured in Farads. Capacitor A device for the temporary storage of electrical charge. Carbon black An amorphous form of carbon, produced commercially by thermal or oxidative decomposition of hydrocarbons. It has a high surface area‐to‐volume ratio, although this ratio is low compared with activated carbon. Often used to support electrocatalysts. Carnot cycle The most efficient (‘ideal’) cycle of operation for a reversible heat engine. It consists of four successive reversible operations, as in the four‐stroke internal combustion engine, namely, isothermal expansion and heat transfer to the 417 418 Glossary of Terms system from a high‐temperature reservoir, adiabatic expansion, isothermal compression and heat transfer from the system to a low‐temperature reservoir, and adiabatic compression that restores the system to its original state. See: Adiabatic process. Carnot efficiency The maximum efficiency with which thermodynamic work can be produced from thermal energy flowing across a temperature gradient. Catalyst A substance that increases the rate of a chemical reaction but that is not itself permanently changed. Cathode An electrode at which a reduction process, i.e., gain of electrons, is occurring. In a fuel cell, the cathode is the positive electrode where oxygen is consumed. During electrolysis, the cathode is the negative electrode where hydrogen is evolved. In a secondary battery, the cathode is the negative electrode on charge and the positive electrode on discharge. Cell voltage The algebraic difference in voltage between the positive and negative electrodes of an electrochemical cell. Cell voltage usually refers to non‐equilibrium conditions, that is, when current is flowing through the cell. The term ‘voltage’ is usually reserved for the case when an electrochemical cell is under consideration, while the term ‘potential’ is usually reserved for the case when an electrode is considered. Unfortunately, the two terms are sometimes used interchangeably. Cermet A composite material composed of ceramic (cer) and metallic (met) components. A cermet is ideally designed to have the optimum properties of both a ceramic, such as high‐temperature resistance and hardness, and those of a metal, such as the ability to undergo plastic deformation. It is typically used as the negative electrode (anode) in a solid oxide fuel cell. Chalcogenide A chemical compound that consists of at least one chalcogen ion and at least one more electropositive element. Although all group 16 elements of the periodic table are defined as chalcogens, the term is more commonly reserved for sulfides, selenides and tellurides rather than oxides. Charge‐transfer coefficient An important parameter in the Butler–Volmer equation for kinetic treatment of electrochemical reactions. The parameter signifies the fraction of the interfacial potential at an electrode|electrolyte solution interface that helps to lower the free energy barrier for the electrochemical reaction. See: Butler–Volmer equation; Gibbs free energy. Chelate An inorganic complex in which a ligand is coordinated to a metal ion at two or more sites. Chemical potential A form of potential that can be absorbed or released during a chemical reaction or phase change. For a given component in a mixture, the change in Gibbs free energy with respect to change in amount of the component, with temperature, pressure and amounts of other components being constant. See: Gibbs free energy. Chemical vapour deposition A chemical process used to produce high‐purity solid materials, usually as thin deposits. Abbreviated as ‘CVD’. Chemisorption A process whereby atoms or molecules of the adsorbed substance (gas or liquid) are held to the surface of a solid material by covalent bonds. See: Adsorption; Physisorption. Clathrates A substance in which the molecules of one compound are encapsulated in lattices or cage‐like structures within another compound. For example, crystalline Glossary of Terms clathrates are formed between certain gases (e.g., carbon dioxide, hydrogen sulfide, methane) and water at low temperatures and high pressures. Climate change A statistically significant change of climate that is attributed directly or indirectly to human activity, which alters the composition of the global atmosphere and is in addition to natural climate variability observed over comparable time periods. Note that climate is usually defined as the ‘average weather’, which, in turn, means using statistics to describe weather (temperature, precipitation and wind) in terms of the mean and variability over a period of time. The World Meteorological Organization uses periods of 30 years, but periods can be as short as months or as long as tens of thousands of years. See: Greenhouse effect; Greenhouse gases. Coal gas A fuel gas, usually rich in methane, that is produced when coal is heated in the absence of air (so‐called destructive distillation) or pyrolysis. It is a by‐product in the preparation of coke and coal tar. Coal gas was a major source of energy in the late 19th and early 20th centuries and was also known as ‘town gas’. The use of this gas declined with the increasing availability of natural gas. See: Pyrolysis. Cogeneration See: Combined heat and power system. Combined cycle A technology to improve the thermal efficiency of a power station that uses natural gas as fuel. The gas is first burnt in a gas turbine, which drives a generator to produce electric power. The waste heat contained in the exhaust gases is then recovered and used to raise high‐pressure steam, which is expanded through a steam turbine to drive another electric generator to produce additional power. Combined‐cycle systems generate electricity in a more efficient and environmentally sound manner. See: Thermal efficiency. Combined heat and power system An installation where there is simultaneous generation of power (either electrical or mechanical) and useful heat (e.g., process steam) in a single process. Also known as ‘cogeneration’. Composite The combining or compositing of best performance benefits from two or more different materials in one component. In a polymer electrolyte fuel cell, for example, a polymer composite of carbon fibre–epoxy may be used in the bipolar plates; in a solid oxide fuel cell, the plates and membrane are ceramic, while the interconnects may be metallic (making the ultimate fuel‐cell stack of composite construction). Composite membrane An ionically conducting membrane, usually constructed in the form of a film made from two or more materials, for use in some types of battery and fuel cell. Concentration overpotential The potential difference caused by differences in the concentration of the charge carriers between the bulk solution and the electrode surface. It occurs when the electrochemical reaction is sufficiently rapid to lower the surface concentration of the charge carriers below that of bulk solution. The rate of reaction is then dependent on the ability (‘mass transfer’) of the charge carriers to reach the electrode surface. Also known as ‘mass‐transport overpotential’ or, less commonly, as ‘diffusion overpotential’. Counter electrode An electrode in an electrochemical system that is used only to make an electrical connection to the electrolyte solution so that a current can be applied to the working electrode. The processes occurring on the counter electrode are unimportant; it is usually made of inert materials (noble metals or carbon/graphite) to avoid its dissolution. Also called an ‘auxiliary electrode’. See: Working electrode. 419 420 Glossary of Terms Crude oil A mixture of hydrocarbons that exists in the liquid phase in natural underground reservoirs and remains liquid at atmospheric pressure after passing through surface‐separating facilities. It occurs in many varieties, distinguished by specific gravity, concentrations of the component hydrocarbons, volatility, heating value and sulfur content. Fuels such as motor petrol (also called gasoline), diesel fuel and jet fuel are derived from crude oil, as well as a variety of materials known as petrochemicals. Cryogenic A term applied to low‐temperature substances and apparatus; usually referring to the temperature range below 77 K. Current density In an electrochemical cell, the current flowing per unit electrode area. Cyclic voltammetry See: Voltammetry. Dead‐end fuel cell A unit fuel cell, or fuel‐cell stack, without fuel and/or oxidant outlet ports. In a dead‐end operation, all of the reactants fed to the cell, or stack, are consumed. Care must be taken, however, to allow continuous removal of the reaction product from the cell/stack. Performance losses are often observed when one or both reactants are supplied in dead‐end mode. This is caused by non‐ optimum flow distribution, as well as by the accumulation of contaminants or inert gases. Density functional theory A quantum mechanical theory used in physics and chemistry to investigate the electronic structure of many‐body systems, in particular atoms, molecules and condensed phases. Desorption Opposite of adsorption, where molecules separate from the solid surface. See: Adsorption. Dielectric A substance (solid, liquid or gas) that is a non‐conductor of electricity (i.e., an insulator). An electric field in a dielectric substance gives rise to no net flow of electricity. Rather, an applied field causes electrons within the substance to be displaced and, thereby, creates an electric charge on the surface of the substance. This phenomenon is used in capacitors to store charge. See: Capacitor. Diesel fuel A combustible distillate of petroleum used as a fuel for diesel (compression ignition) engines; usually the fraction of crude oil that is distilled after kerosene. See: Crude oil. Diffusion Net spontaneous and random movement of molecules, particles or ions in a fluid (gas or liquid) from a region in which they are at a high concentration to a region of lower concentration, until a uniform concentration is achieved throughout. The difference in concentration between two such regions is called the ‘concentration gradient’. Diffusion coefficient The coefficient of proportionality between the flux of a substance and its concentration gradient. Diode A solid‐state electrical device that only allows current to flow in one direction. Direct current Electric current that flows in one direction only, although it may have appreciable pulsations in its magnitude. It is the form of electricity produced by electrochemical cells. See: Alternating current. Direct internal reforming Production of a desired product (e.g., hydrogen) within a unit fuel cell, or fuel‐cell stack, from a hydrocarbon‐based fuel (e.g., diesel, methanol, natural gas) fed to the cell, or stack. See: External reforming; Indirect internal reforming. Disproportionation A chemical reaction in which a single substance acts as both an oxidizing and a reducing agent to produce dissimilar substances. For example, carbon Glossary of Terms monoxide can decompose over a catalyst to form both solid carbon and carbon dioxide; this particular disproportionation is known as the Boudouard reaction. Dissociation, dissociation constant In chemistry and biochemistry, a general process in which ionic compounds (complexes, molecules or salts) separate or split into smaller molecules, ions or radicals, usually in a reversible manner. The equilibrium constant of a reversible dissociation is called the ‘dissociation constant’. It is the ratio of the product of the concentrations of the dissociated species to the concentration of the undissociated compound. Distributed energy A network of power generation, storage and metering/control systems that allow power to be used and managed in a distributed and small‐scale manner, thereby placing the supply close to the load, rather than from a large, centralized power plant, in order to minimize electricity transmission and maximize waste heat utilization. Also as ‘distributed power generation’ or ‘embedded generation’. Drive-train The elements of the propulsion system (including engine, transmission, driveshaft and differential) that deliver mechanical energy from the power source to drive the wheels of a given vehicle. Dye‐sensitized solar cell A photoelectrochemical cell that uses a dye‐impregnated layer of titanium dioxide to generate a voltage by means of light energy rather than the semiconducting materials used in most photovoltaic cells. See: Photovoltaic cell. Electrical double-layer A model of the ionic environment (charge accumulation) at the interface between an electrode and an electrolyte solution close to it. In general terms, the structure is comprised of a compact charged layer adjacent to the electrode surface and a diffuse region of charge extending into the electrolyte solution. Note that there are several theoretical treatments of the solid|liquid interface. Also known simply as a ‘double layer’. Electric vehicle A vehicle that is powered solely by an electrochemical power source, such as a battery or fuel cell. Power assistance may also be provided by a supercapacitor. Electrocatalyst A substance that accelerates the rate of an electrochemical (electrode) reaction but that is not itself permanently changed. Electrochemical capacitor A capacitor that stores charge in the form of ions (rather than electrons), adsorbed on materials of high surface area. The ions undergo redox reactions during charge and discharge. The device is also known as an ‘electrochemical double‐layer capacitor’, a ‘supercapacitor’ or an ‘ultracapacitor’. Electrochemical (AC) impedance spectroscopy An investigative technique for the examination of processes that occur at electrode surfaces. An alternating‐current (sinusoidal) excitation signal (potential or current), of small amplitude and covering a wide range of frequencies, is applied to the system under investigation, and the response (current or voltage or another signal of interest) is recorded. Given the small amplitude of the excitation signal, data can be obtained without significantly disturbing the normal operation of the system. By conducting measurements over a wide range of frequencies, a complex sequence of coupled processes, such as electron transfer, mass transport and electrochemical reaction, can often be separated and evaluated. The technique is routinely used to study electrode kinetics and reaction mechanisms and for the characterization of battery, fuel cell and corrosion phenomena. The term is abbreviated as ‘EIS’. See: Impedance. 421 422 Glossary of Terms Electrode An electronic conductor that acts as a source or sink of electrons that are involved in electrochemical reactions. Electrode potential The voltage developed by a single electrode, either the positive or the negative; usually related to the standard potential of the hydrogen electrode (which is arbitrarily set at 0 V). The International Union of Pure and Applied Chemistry defines electrode potential as the voltage of a cell in which the electrode on the left is a standard hydrogen electrode and the electrode on the right is the electrode in question. See: Standard hydrogen electrode. Electrolysis The passage of an electric current through an ionic substance (an electrolyte, either dissolved in a suitable solvent or molten) that results in chemical reactions at the electrodes and separation of materials. Note that high‐temperature electrolysis (also known as steam electrolysis), which is currently being investigated for the production of hydrogen from water, employs a solid electrolyte of yttria‐stabilized zirconia. See: Electrolyser. Electrolyte A chemical compound that ionizes when dissolved or molten to produce an electrically conductive medium; also solid materials that are conducive due to the movement of ions through voids, or empty crystallographic positions, in their crystal lattice structure, e.g., yttria‐stabilized zirconia used primarily in solid oxide fuel cells. Note that in the case of dissolved materials, it is fundamentally incorrect to refer to the ‘electrolyte solution’ as the ‘electrolyte’. Nevertheless, the former terminology has become common practice. Electrolyser An electrochemical plant designed to effect the process of electrolysis. Electron An elementary particle with negative electric charge of 1.602 × 10−19 coulombs and a mass of 9.109 × 10−31 kg. Electron microscope A form of microscope that uses a beam of electrons instead of a beam of light (as in the optical microscope) to form a large image of a very small object. Electro‐osmotic drag In certain types of fuel cell, the flux of water that, due to its attraction to protons, is transported through the electrolyte medium from the negative electrode (anode) to the positive electrode (cathode). The flux is driven by the electric field between the electrodes. Endothermic reaction A chemical reaction that takes place with an uptake of energy (heat) from the environment. See: Enthalpy; Exothermic reaction. End-plate A flat metal plate at each end of a fuel‐cell stack, through which tie bolts are used to compress the cells and cooling plates in the stack, thereby rendering the whole a continuous electronic conductor. Either the end-plates or the tie bars must be electrically insulated. If the latter, the end-plate may be the take‐off point for electric current; otherwise, a separate take‐off on the stack side of each end plate may be used. See: Bipolar plate. Energy The ability to do work or produce heat (measured in Joules). Energy density The accessible stored energy per unit volume of an electrochemical cell, usually expressed as Wh L−1 or Wh dm−3 (MJ m−3 or kWh m−3 for large storage facilities). See: Theoretical energy density. Energy efficiency The ratio of the energy output from a device to the energy input, usually expressed as a percentage. Enthalpy A thermodynamic quantity (H) equal to the total energy content of a system when it is at constant pressure. The gain or loss of energy of a system when it reacts at constant pressure is expressed by the change in enthalpy, symbolized by ΔH. Glossary of Terms When all the energy change appears as heat (Q), the change in enthalpy is equal to the heat of reaction at constant pressure, i.e., ΔH = Q. The values of ΔH and Q are negative for exothermic reactions (heat evolved from system) and positive for endothermic reactions (heat absorbed by system). Entropy A thermodynamic quantity that represents the amount of energy in a system that is no longer available to do useful work. When a closed system undergoes a reversible change, the entropy change (ΔS) equals the energy lost from, or transferred to, the system by heat (Q) divided by the absolute temperature (T) at which this occurs, i.e., ΔS = Q/T. At constant pressure, the amount of heat (Q) is equal to the change in enthalpy (ΔH). Equivalent weight The weight of a substance that will combine with or displace 1 g of hydrogen (or 8 g of oxygen) in a chemical reaction. For an element, it is the relative atomic mass divided by the valency. For a compound, it depends on the reaction considered. Eutectic mixture A solid solution that consists of two or more substances and has the lowest freezing point of any possible mixture of these components. The minimum freezing point for a set of components is called the ‘eutectic point’. Low‐ melting‐point alloys are usually eutectic mixtures. Exchange‐current density The current per unit area that flows equally in the forward and backward directions when an electrode reaction is in equilibrium. Exergy A thermodynamic property representing the thermal and chemical energy that is available to do useful work, i.e., it expresses the quality of an energy source. Specifically, exergy is a measure of the energy difference between some process state (i.e., pressure, temperature and composition) and a reference state (typically atmospheric conditions). Exergetic efficiency measures entropy production and hence represents irreversible losses associated with chemical and thermal processes. Exothermic reaction A chemical reaction that takes place with a release of energy (heat) to the environment. See: Endothermic reaction; Enthalpy. External reforming The production of hydrogen from a hydrocarbon fuel (e.g., methanol, natural gas, propane) prior to entry to a unit fuel cell or fuel‐cell stack. See: Direct internal reforming; Indirect internal reforming. Fermentation The chemical decomposition of a complex substance, especially a carbohydrate, into simpler chemical products that is brought about by the action of enzymes, bacteria, yeasts or molds, generally in the absence of oxygen. May be a natural process or one promoted or enhanced technically to produce a desired end product, e.g., corn products, to yield ethanol. Fischer–Tropsch process A catalysed chemical reaction in which synthesis gas, a mixture of carbon monoxide and hydrogen, is converted into liquid hydrocarbons of various forms. Also known as ‘Fischer–Tropsch synthesis’. See: Synthesis gas. Flow battery A form of rechargeable battery in which the electroactive materials (usually redox couples) of both electrode polarities are dissolved in a solvent (usually water) to form electrolyte solutions that are stored externally and pumped to the cells of the battery during operation. Flow batteries can be rapidly ‘recharged’ by replacing the electrolyte solutions while simultaneously recovering (‘re-energizing’) the spent material for re‐entry to the battery. Also known as a ‘redox battery’. 423 424 Glossary of Terms Flow-field The structure of channels in the bipolar plate/separator of some types of fuel cell that distributes reactants across the surface of a catalysed fuel‐cell membrane–electrode assembly and also removes the products of the electrochemical reaction and excess reactants (including inert components of the reactant inlet stream, e.g., nitrogen). When flow-fields are on one side only, the plate is called a ‘separator-plate’ or ‘current-collector’ and is not considered bipolar such as the end-plates in a fuel‐cell stack or the two end-plates in a single cell that is not integrated into a stack. The most common designs of flow-field are ‘parallel’ (an array of parallel channels), ‘serpentine’ (gas channels are not straight but contain bends) and ‘interdigitated’ (a comblike arrangement of discontinuous, so‐called dead‐ended, channels, with staggered channels respectively connected to the gas inlet and outlet). The channels are either machined or moulded into flat plates made of metal, ceramic, graphite or composite. See: Bipolar plate; End-plate; Membrane–electrode assembly. Fossil fuels Carbonaceous deposits (solid, liquid or gaseous) that derive from the decay of vegetable matter over geological time spans. Fuel‐cell vehicle An electric‐drive vehicle that derives the power for its motor(s) from a fuel‐cell system. Fugacity A thermodynamic function that is introduced as an effective substitute for pressure to allow a real gas system to be considered by the same equations that apply to an ideal gas. See: Activity. Gas‐diffusion layer A fuel‐cell component that serves two main functions: it must permit the passage of gases and be sufficiently conductive to allow the transfer of electrons. The layer also provides a support for the catalyst, and its structure promotes the removal of generated water that may hinder (‘flood’) the electrochemical reaction. The layer is very thin, usually with a thickness of 0.25–0.40 mm and with pore sizes of 4–50 µm. It is typically made from carbon cloth, carbon paper or Toray paper (a carbon–carbon composite paper of high strength). Gasification A special type of pyrolysis where thermal decomposition takes place in the presence of a small amount of air or oxygen. See: Coal gas; Pyrolysis. Gasoline Same as Petrol. Gibbs free energy The energy liberated or absorbed in a reversible process at constant pressure and constant temperature. Put another way, it is the minimum thermodynamic work (at constant pressure) needed to drive a chemical reaction (or, if negative, the maximum work that can be done by the reaction). Thus, the Gibbs free energy is a thermodynamic quantity that can be used to determine if a reaction is spontaneous or not. The change in free energy, ΔG, in a chemical reaction is given by ΔG = ΔH − TΔS, where ΔH is the change in enthalpy and ΔS is the change in entropy. This is known as the ‘Gibbs equation’. See: Enthalpy; Entropy. Grain boundary The interface between two regions of a solid that have different crystal orientations. Greenhouse effect The trapping of heat by greenhouse gases that allow incoming solar radiation to pass through the Earth’s atmosphere but prevent the escape to outer space of a portion of the outgoing infrared radiation from the surface and lower atmosphere. This process has kept the Earth’s atmosphere about 33°C warmer than it would be otherwise; it occurs naturally but may also be enhanced by certain human activities, e.g., the burning of fossil fuels. See: Greenhouse gases. Glossary of Terms Greenhouse gases Any of the gaseous constituents of the atmosphere, both natural and anthropogenic, that absorb and re‐emit radiation at specific wavelengths within the spectrum of infrared radiation emitted by the Earth’s surface, the atmosphere and clouds. Greenhouse gases include water vapour, carbon dioxide, methane, nitrous oxide, halogenated fluorocarbons, ozone, perfluorinated carbons and hydrofluorocarbons. See: Greenhouse effect. Half‐cell reaction The electrochemical reaction at an electrode. Heat-exchanger A device in which heat is transferred from one fluid stream to another without mixing. Heat‐exchanger operations are most efficient when the temperature differentials are greater. Higher heating value of a fuel, HHV The amount of heat released by the complete combustion of a unit volume or weight of a fuel (initially at 25°C) with all the products brought back to the original temperature. Thus, the value takes into account recovery of the latent heat of vapourization of the water formed by combustion and is useful in calculating heating values for fuels where condensation of the reaction products is practical (e.g., in a gas‐fired boiler used for space heating). Also known as ‘gross calorific value’ or ‘gross energy’. See: Lower heating value of a fuel. Hole The vacancy where an electron would normally exist in a solid. A hole is an electric charge carrier with a positive charge, equal in magnitude but opposite in polarity to the charge on the electron. Holes and electrons are the two types of charge carriers in semiconductor materials; holes are induced into a semiconductor by adding small quantities of an acceptor dopant to the host crystal. Under the application of an electric field, holes move in the opposite direction from electrons, thus producing an electric current. See: Semiconductor. Hybrid electric vehicle A vehicle that derives part of its propulsion power from an internal combustion engine and part of its propulsion power from an electric motor or that uses an internal combustion engine to power a generator to charge a battery that in turns powers one or more electric‐drive motors. See: Parallel hybrid electric vehicle; Plug‐in hybrid electric vehicle; Series hybrid electric vehicle. Hydrogen economy The concept of an energy system based primarily on the use of hydrogen as an energy carrier and fuel, especially for transportation vehicles and distributed power generation. See: Distributed energy. Hydrogen embrittlement A process whereby a metal becomes brittle on exposure to hydrogen. Embrittlement arises from (i) the formation of metal hydride phases with different lattice parameters to the non‐hydrided metal, which creates stresses within the metal lattice, and (ii) recombination of atomic hydrogen to molecular hydrogen in defects within the metal. Hydrophilic Having an affinity for water. Hydrophobic Lacking an affinity for water. Impedance The analogue of the resistance when applied to alternating current. It is a measure of the inability of a circuit to carry the electrical current. In many cases, the impedance varies with the changes in the frequency of the applied electrical potential due to the properties of the conducting liquid or solid. In electrochemistry, the impedance of the electrodes is also frequency dependent. Inconel An austenitic nickel‐based alloy designed for use in high‐temperature applications. Composed primarily of nickel, chromium, iron and molybdenum. See: Austenitic steel. 425 426 Glossary of Terms Indirect internal reforming The reformer unit is separated but adjacent to the fuel‐cell negative electrode (anode). This arrangement takes advantage of the close‐coupled thermal benefit from the exothermic reaction of the fuel cell to support the endothermic reforming reaction. See: External reforming; Direct internal reforming. Inductance The magnitude of the capability of a component (inductor), such as a wire loop or coil, in an electrical circuit to store energy in the form of a magnetic field. An inductance of 1 H is produced when 1 V is induced by a change in current of 1 A s−1. Insulated‐gate bipolar transistor A three‐terminal power semiconductor device, noted for high efficiency and fast switching. It switches electric power in many modern appliances, such as fuel cells, electric cars, variable speed refrigerators and air conditioners. Abbreviated as ‘IGBT’. Interconnect In a solid oxide fuel cell, the interconnect is either a metallic or ceramic material that typically lies between each individual cell to allow the cells to be connected in series and also to allow the passage of fuel and air to the negative electrode (anode) and positive electrode (cathode), respectively. Internal resistance The opposition to current flow that results from the various electronic and ionic resistances within an electrochemical or photoelectrochemical cell. Internal short-circuit Same as Short-circuit. Inverter An electronic device that converts low‐voltage direct current to high‐voltage alternating current. Ion An atom that has lost or gained one or more orbiting electrons and thus becomes electrically charged. Ion‐exchange membrane A plastic film formed from ion‐exchange resin. The utility of such membranes is based on the fact that they are permeable preferentially only to either positive ions (cation‐exchange membrane) or negative ions (anion‐ exchange membrane). Ionic liquid A liquid that essentially contains only ions. In the broad sense, the term includes all molten salts. Nowadays, however, the term ‘ionic liquid’ is commonly used for salts whose melting point is below 100°C. In particular, the salts that are liquid at room temperature are called ‘room‐temperature ionic liquids’ or ‘RTILs’. Ionization Any process by which an atom, molecule or ion gains or loses electrons. Ionomer An ionomer is a polymer that comprises repeat units of both electrically neutral repeating units and a fraction of ionized units. The ionic groups lead to novel physical properties of the polymer, such as electrical conductivity and isoviscosity (an increase in ionomer solution viscosity with increase in temperature). kVAR Unit of reactive power. See VAR. Latent heat The heat absorbed or released by a substance when it changes state (e.g., from solid to liquid, or vice versa) at constant temperature and pressure. The term Specific latent heat denotes the heat absorbed or released per unit mass of a substance in the course of its change of state. Life‐cycle analysis A method for evaluating ‘the whole life of a product’. That is, all the stages involved, such as raw materials acquisition, manufacturing, distribution and retail, use and reuse and maintenance, recycling and waste management in order to create less environmentally harmful products. The process consists of three parts: inventory analysis (selecting items for evaluation and quantitative analysis), Glossary of Terms impact analysis (evaluation of impacts on the ecosystem) and improvement analysis (evaluation of measures to reduce environmental loads). According to the prime objective of the exercise, also known as ‘cradle‐to‐grave analysis’, ‘dust‐to‐dust energy cost’, ‘ecobalance’ and ‘well‐to‐wheel analysis’. Lignite The lowest rank of coal (carbon content 25–35 wt.%) and used almost exclusively as fuel for electric power generation; also referred to as ‘brown coal’. See: Anthracite; Bituminous coal; Peat; Sub‐bituminous coal. Liquefied petroleum gas Various petroleum gases, principally propane and butane, stored as a liquid under pressure. Abbreviated as ‘LPG’. Load The total demand for electrical power on a supply such as a battery, fuel cell or supercapacitor. Load following (fuel cell) Method of operating a fuel‐cell system to generate a varying amount of power, depending upon the load demanded. Due to an inherent lag time between the change in base load and the response of peripheral balance‐of‐plant components that support the operation of a fuel cell, load following can be enhanced with a buffer such as a battery or supercapacitor. See: Balance-of-pant; Base load. Lower heating value of a fuel, LHV The amount of heat released by the complete combustion of a unit volume or weight of a fuel assuming that all products remain in the gaseous state. Thus, the latent heat of vapourization of the water formed by combustion is not taken into account. The value is useful in comparing fuels where condensation of the combustion products is impractical or heat at a temperature below 150°C cannot be put to use. Also known as the ‘net calorific value’. See: Higher heating value of a fuel. Luggin capillary A salt bridge with a thin capillary tip at one end that is used to connect the working and reference electrode compartments of a three‐electrode cell. Placement of the capillary tip very close to the surface of the working electrode defines a clear sensing point for the reference electrode and serves to minimize the solution IR drop. Also known as a ‘Luggin tip’, ‘Luggin probe’ or ‘Luggin‐Haber capillary’. Mass transport Transfer of materials consumed or formed in an electrode process to or from the electrode surface. The mechanism of mass transport may include diffusion, convection and electromigration. Membrane A layer of material that serves as a selective barrier between two phases and remains impermeable to specific particles, molecules or substances when exposed to the action of a driving force. Some components are allowed passage by the membrane into a permeate stream, whereas others are retained by it and accumulate in the retentate stream. In a fuel cell, the membrane acts as an electrolyte (ion exchanger), as well as a barrier film separating the gases in the positive (cathode) and negative (anode) electrode compartments. See: Ion‐exchange membrane. Membrane–electrode assembly A core component of the structure of a proton‐ exchange membrane fuel cell that consists of a polymer electrolyte membrane coated with catalyst–carbon–binder layers (‘electrodes’) and sandwiched by two microporous conductive layers that function as gas‐diffusion layers and currentcollectors; the assembly is placed between bipolar plates to form the basic unit of a fuel‐cell stack. Electrochemical reactions occur when a fuel (e.g., hydrogen) and an oxidant (e.g., oxygen) are applied, respectively, to the negative‐electrode (anode) and the positive‐electrode (cathode) sides of the assembly. See: Bipolar cell; Gas‐diffusion layer. 427 428 Glossary of Terms Microbial fuel cell A bio‐electrochemical system that exploits living microorganisms as catalysts to facilitate the generation of electricity. Also known as a ‘biological fuel cell’ or a ‘bio‐electrochemical fuel cell’. Abbreviated as ‘MFC’. Micro‐electromechanical system The integration of mechanical elements, sensors, actuators and electronics on a common silicon substrate through microfabrication technology. Abbreviated as ‘MEMS’. Micro fuel cells Fuel cells sized for small portable devices such as cell phones, cameras and laptop computers. Mixed potential The electrode potential when two electrode reactions occur on the same electrode surface. The mixed potential has a value in between the equilibrium potentials of the two electrode reactions; it is a steady‐state phenomenon. Molar Terminology to denote that an extensive physical property is being expressed per mole of a substance. (An extensive variable is proportional to the size of the system, e.g., volume, mass, energy.) Molarity The concentration of a solution expressed as the number of moles of dissolved substance dissolved per unit volume of solvent, usually expressed as mol dm−3. Mole The amount of a substance (in grams) that contains as many elementary units as there are atoms (6.02 × 1023) in 0.012 kg of the carbon isotope 12C. The elementary units may be atoms, molecules, ions or electrons. Mole fraction In a system of mixed constituents, the ratio of the number of moles of a single constituent in a given volume to the total number of moles of all constituents in that volume. Monomer A compound whose simple molecules can be joined together (polymerize) to form a giant polymer molecule. Monopolar The conventional method of battery construction in which the component cells are discrete and are externally connected to each other. Municipal solid waste Solid domestic/household waste. Nafion™ A brand name used by DuPont for a series of fluorinated sulfonic acid copolymers, the first synthetic ionic polymer. The material is resistant to chemical breakdown, and thus it is useful for membranes in proton‐exchange membrane fuels cells. Nernst equation A thermodynamic equation demonstrating that the voltage developed in an electrochemical cell is determined by the activities of the reacting species, the reaction temperature and the standard free energy change of the overall reaction. See: Gibbs free energy. n‐Type semiconductor A semiconductor in which electrical conduction is due mainly to the movement of electrons. Nyquist diagram or Nyquist plot A graphical illustration of the data obtained from electrochemical impedance spectroscopy. See: Electrochemical (AC) impedance spectroscopy; Impedance. Ohmic loss The decrease in the voltage of a fuel cell (or battery) that results from current flow through the internal resistance. Oil shale Rocks rich in organic material (kerogen) from which petroleum may be recovered by dry distillation. Open‐circuit voltage The voltage of a power source, such as a battery, fuel cell or photovoltaic cell, when there is no net current flow. Glossary of Terms Original equipment manufacturer A confusing term that has two meanings. Originally, an original equipment manufacturer was a company that supplied equipment to other companies to resell or incorporate into other products using the respective resellers’ brand names. A number of companies, both equipment suppliers and equipment resellers, still use this meaning. More recently, the term is used to refer to the company that acquires a product or component and reuses or incorporates it into a new product with its own brand name. Abbreviated as ‘OEM’. Overpotential The shift in the potential of an electrode from its equilibrium value as a result of current flow. Overvoltage The difference between the cell voltage (with a current flowing) and the open‐circuit voltage. The overvoltage represents the extra energy needed (an energy loss that appears as heat) to force the cell reaction to proceed at a required rate. Consequently, the cell voltage of an electrochemical cell (e.g., a rechargeable battery during discharge) is always less than the open‐circuit value, while the cell voltage of an electrolytic cell (e.g., a rechargeable battery during charge) is always more than the open‐circuit value. The overvoltage is the sum of the overpotentials of the two electrodes of the cell and the ohmic loss of the cell. Unfortunately, the terms ‘overvoltage’ and ‘overpotential’ are sometimes used interchangeably. Moreover, overvoltage is also referred to as ‘polarization’ of the cell, and overpotential as polarization of the electrode. This is an ill‐defined and misleading term, as exemplified by the many different definitions to be found in dictionaries. See: Open‐circuit voltage. Parallel connection The connection of like terminals of cells or batteries to form a system of greater capacity, but with the same voltage. Parallel hybrid electric vehicle A type of hybrid electric vehicle in which the alternative power unit is capable of producing motive force and is mechanically linked to the power-train. See: Power-train; Series hybrid electric vehicle. Parasitic load Power consumed by the balance‐of‐plant equipment that is necessary to operate a fuel‐cell system. See: Balance-of-plant; Self‐discharge. Partial oxidation A combustion process in which just sufficient oxygen is supplied to oxidize a hydrocarbon fuel to carbon monoxide and hydrogen rather than fully to carbon dioxide and water. This is accomplished by injecting air with the fuel stream prior to the reformer. The advantage of partial oxidation over steam reforming of the fuel is that it is an exothermic reaction rather than an endothermic reaction and therefore generates its own heat. The hydrogen‐rich gaseous product can then be put to further use, for example, in a certain types of fuel cell. Peat A precursor of coal that has industrial importance as a fuel in some regions, for example, Ireland and Finland. Permeability The rate of diffusion of gas or liquid through a porous material. Expressed, for a thin material, as the rate per unit area and, for a thicker material, as the rate per unit area of unit thickness. Perovskite Any material with the same type of crystal structure as calcium titanium oxide (CaTiO3). The general chemical formula for perovskite compounds is ABX3, where A and B are two cations of very different sizes and X is an anion that bonds to both the cations. The A atoms are larger than the B atoms. Perovskite compounds are widely used as positive electrodes (cathodes) in solid oxide fuel cells. 429 430 Glossary of Terms Petrol Term used in the United Kingdom for a light hydrocarbon liquid fuel obtained by refining petroleum that is used in most spark‐ignition internal combustion engines. Other terms for such fuel are ‘gas’, ‘gasoline’ and ‘motor spirit’. See: Crude oil. Petroleum A collective term for crude oil, natural gas, natural gas liquids and other related products (both hydrocarbon and non‐hydrocarbon compounds). It is usually found in deposits beneath the Earth’s surface and thought to have originated from plant and animal remains of the geologic past. See: Crude oil. pH A measure of the acidity/alkalinity (basicity) of a solution. The pH scale extends from 0 to 14 (in aqueous solutions at room temperature). A pH value of 7 indicates a neutral solution. A pH value of less than 7 indicates an acidic solution; the acidity increases with decreasing pH value. A pH value of more than 7 indicates an alkaline solution; the basicity or alkalinity increases with increasing pH value. Photobiological hydrogen production The production of hydrogen by three classes of organisms, namely, photosynthetic bacteria, cyanobacteria and green algae. These organisms use their photosynthetic properties to absorb sunlight and convert it into chemical energy. Photoelectrochemical cell Solar cells that extract electrical energy from light, including visible light. Each cell consists of a photosensitive electrode and a conducting counter electrode immersed in an electrolyte solution. Some photoelectrochemical cells simply produce a direct current, while others liberate hydrogen in a process similar to the conventional electrolysis of water. Photolysis A chemical reaction (often a decomposition) caused by exposure to light. Photovoltaic Relating to or designating devices that absorb solar radiation and transform it directly into electricity. Photovoltaic cell A semiconductor device for converting light energy into low‐ voltage direct‐current electricity. Physisorption Adsorption of gases on solid surfaces whereby the bonding is by means of a weak intermolecular (van der Waals) attraction rather than by chemical bonding. See: Adsorption; Chemisorption. Platinum‐group metals The three members of the second and third transition series immediately preceding silver and gold, i.e., ruthenium, rhodium, palladium and osmium, iridium, platinum. Plug‐in hybrid electric vehicle A hybrid electric vehicle with batteries that can be recharged by connecting a plug to an electric power source. Thereby, it shares the characteristics of both traditional hybrid electric vehicles, by having an electric motor and an internal combustion engine, and of battery electric vehicles. See: Electric vehicle; Hybrid electric vehicle. Polarization An ill‐defined and misleading term used for overpotential and for overvoltage. See: Overvoltage. Porosity The ratio of the accessible volume of a porous body to the total volume, usually expressed as a percentage. Porosity features such as overall open porosity, pore shape, size and size distribution are key properties of battery and fuel‐cell electrodes that significantly influence cell performance. Potentiostat Electronic hardware required to control a three‐electrode cell and run most electroanalytical experiments. The system functions by maintaining the potential of the working electrode at a constant level with respect to the reference electrode by adjusting the current at an auxiliary electrode. See: Reference electrode; Working electrode. Glossary of Terms Power conditioner The subsystem that converts the direct‐current power from a fuel‐cell stack subsystem to direct‐current or alternating‐current power that is required by the application. Power density The power output of an electrochemical cell per unit volume, usually expressed as W L−1 or W dm−3. Power factor The ratio between the total real power (measured in watts or kilowatts) and the total apparent power (the product of the root‐mean‐square voltage and the root‐mean‐square current, measured in volt‐amperes or kilovolt‐amperes), expressed as either a decimal fraction or a percentage. Power-train The elements of a vehicle propulsion system that include all drivetrain components plus an electrical power inverter and/or controller but not the battery or fuel‐cell system. See: Drive-train. Preferential oxidation A reaction that preferentially oxidizes a gas on a catalyst. For example, the oxidation of carbon monoxide to carbon dioxide using a heterogeneous catalyst placed on a ceramic support; a reaction of considerable interest in fuel‐cell design. Also known as ‘selective oxidation’. Abbreviated as ‘PROX’. Pressure swing adsorption A technology used to separate some gas species from a mixture under pressure according to the molecular characteristics of a given gas and its affinity for an adsorbent material. The process operates at near‐ambient temperature. Primary battery (or cell) A battery (or cell) that contains a fixed amount of stored energy when manufactured and that cannot be recharged after that energy is withdrawn. Producer Gas A mixture of carbon monoxide and nitrogen made by passing air over very hot carbon. The gas is used as a fuel in some industrial processes. Proton An elementary particle that is stable, carries a positive charge equal in magnitude to the negative charge of the electron and has a mass of 1.672 × 10−27 kg (i.e., ~1836 times that of the electron). Also, the nucleus of an ordinary or light hydrogen atom. Protons are constituents of all atomic nuclei. Proton‐exchange membrane The polymer‐based component in a proton‐exchange membrane fuel cell that acts as an electrolyte through which protons, but not electrons, can pass (to move along the electrode and generate a current), as well as a barrier film to separate the hydrogen‐rich feed in the positive‐electrode (cathode) compartment of the cell from the oxygen‐rich negative electrode (anode) side. Proton‐exchange membranes are also employed in certain designs of electrolysis cells. Also known as a ‘polymer electrolyte membrane’ or a ‘solid polymer membrane’. Pyrolysis Thermal decomposition of a substance at elevated temperatures in the absence of air or oxygen. Quantum yield For photocells, the fractional number of electrons generated per photon incident on the cell or the ratio of the number of photon‐induced reactions occurring to the total number of incident photons. Rechargeable battery See: Secondary battery (or cell). Redox battery A battery in which the chemical energy is stored as dissolved redox reagents. The electrodes are contained in compartments that are typically separated by an ion‐exchange membrane. See: Flow battery. Reference electrode An electrode with a reproducible, well‐established potential, against which potentials of other electrodes can be measured. Reformate The product of a hydrocarbon reforming process. See: Steam reforming. 431 432 Glossary of Terms Regenerative braking The recovery of some fraction of the energy normally dissipated as heat during braking of a vehicle and its return to a battery or some other energy‐storage device. The process of slowing a vehicle involves drawing kinetic energy into a motor so that it acts as an electric generator and thereby exerts a rotational drag on the wheels. Most hybrid electric vehicles employ regenerative braking. Regenerative fuel cell A type of fuel cell in which the chemical reactants undergo reversible reactions, such that the cell may be recharged with a separate power source if desired. For example, the hydrogen–oxygen fuel cell may be recharged for the production of hydrogen via water electrolysis. Also called a ‘reversible fuel cell’. Abbreviated as ‘RFC’. See: Unitized regenerative fuel cell. Relative humidity The ratio of the actual amount of moisture in the air to the amount needed for saturation at the same temperature. Renewable energy Forms of energy (such as geothermal heat, hydropower, sunlight, tidal energy, wave power, wind and organic matter) that flow through the Earth’s biosphere and are available for human use indefinitely, provided that the physical basis of their flow is not destroyed. Also known as ‘renewables’. Reversible potential The potential of an electrode when there is no net current flowing through the cell. Reversible voltage The difference in the reversible potentials of the two electrodes that make up the cell. See: Reversible potential. Round‐trip efficiency Same as Energy efficiency. Saturated calomel electrode A reference electrode based on the reaction between elemental mercury and mercury (I) chloride (Hg2Cl2, ‘calomel’). The aqueous phase in contact with the mercury and the mercury (I) chloride is a saturated solution of potassium chloride in water. The electrode is normally linked via a porous frit (‘salt bridge’) to the solution in which the other electrode is immersed. The equilibrium electrode potential is a function of the chloride concentration in the internal electrolyte solution. At 25°C, the potential of the saturated calomel electrode is +241.2 mV versus the standard hydrogen electrode. Secondary battery (or cell) A battery (or cell) that is capable of repeated charging and discharging. Also known as a ‘rechargeable battery (or cell)’. Selective oxidation See: Preferential oxidation. Semiconductor A solid‐state crystalline material that has a value of electrical resistivity intermediate between those of metals and insulators. The conductivity of semiconductors can be controlled by adding very small amounts of foreign elements called ‘dopants’. Conductivity is facilitated not only by negatively-charged electrons but also by positively-charged holes, and it is sensitive to temperature, illumination and a magnetic field. See: Hole. Sensible heat The heat absorbed by a substance that gives rise to an increase in temperature of the substance. See: Latent heat. Separator An electronically non‐conductive, but ion‐permeable, material that prevents electrodes of opposite polarity from making contact. Sequestration The capture of carbon dioxide from streams of mixed gases and its subsequent indefinite storage. Also known as ‘carbon capture and storage’. Series hybrid electric vehicle A type of hybrid electric vehicle that runs on battery power like a pure electric vehicle until the batteries discharge to a set level, when an Glossary of Terms alternative power unit turns on to recharge the battery. See: Parallel hybrid electric vehicle. Short circuit The direct connection of positive and negative electrodes either internal or external to the battery. Short‐circuit current Current flowing freely through an external circuit that has no load or resistance; the maximum current possible. Sintering A method for making objects from powder by heating the material (below its melting point — solid‐state sintering) until its particles adhere to each other. Sintering is employed in the manufacture of membranes for solid oxide fuel cells. Sol–gel synthesis A method of preparing single or mixed oxides at low temperature that involves the formation of a sol (a colloidal suspension or solution of the precursor), which is converted to a gel (a continuously linked network of oxide with a continuous interstitial liquid phase) before drying to form a xerogel. See: Xerogel. Solid electrolyte A solid‐state ion conductor in which the electrical conductivity is due to the movement of ions (cations or anions) through voids or interstitial spaces in the lattice structure. Also known as ‘fast ion conductors’ or ‘superionic conductors’. Specific energy The accessible stored energy per unit mass of an electrochemical cell, expressed as MJ kg−1, Wh kg−1 or kWh kg−1. See: Theoretical specific energy. Specific heat The quantity of heat that unit mass of a substance requires to raise its temperature by one degree, expressed as J kg−1 K−1. Specific power The power output of an electrochemical cell per unit weight, usually expressed as W kg−1. Specific surface area Total surface area of a material divided by the mass of the material, usually expressed as m2 g−1. See: BET‐specific surface area. Sputtering A method of depositing a thin layer of one material on to a substrate. A target material is bombarded by charged particles (typically argon) that dislodge atoms from the target and deposit them on a substrate. The technique is a form of physical vapour deposition. Standard conditions for temperature and pressure A standard set of conditions for experimental measurements to allow comparisons to be made between different sets of data. The International Union of Pure and Applied Chemistry (IUPAC) has established two standards: (i) standard temperature and pressure, abbreviated as ‘STP’, specifies a temperature of 273.15 K and an absolute pressure of 100 kPa (1 bar) and (ii) standard ambient temperature and pressure, abbreviated as ‘SATP’, specifies a temperature of 298.15 K and an absolute pressure of 100 kPa (1 bar). By contrast, the standard formulated by the National Institute of Science and Technology (NIST) in the United States stipulates a temperature of 293.15 K and an absolute pressure of 101.325 kPa (1 atm), abbreviated as ‘NTP’. Standard electrode potential The reversible potential of an electrode with all the active materials in their standard states. The standard states usually adopted by electrochemists specify an absolute pressure of 101.325 kPa (1 atm) for gases and unit activity for elements, solids and 1 mol dm−3 solutions — all at a temperature of 298.15 K. See: Reversible potential. Standard hydrogen electrode A standard reference electrode, usually consisting of a platinum electrode coated with platinum black that is bathed with a stream of hydrogen gas bubbles and immersed in a solution of hydrogen ions (typically 433 434 Glossary of Terms sulfuric acid). Its potential is declared to be 0 V at all temperatures when the activity of all species is unity. A zero point is needed since the potential of a single electrode cannot be measured — only the difference of two electrode potentials is measurable. All electrode potentials are expressed on this ‘hydrogen scale’. In practice, a unit concentration (rather than unit activity) of hydrogen ions and unit pressure (rather than unit fugacity) of hydrogen gas are used. Other reference electrodes (e.g., calomel or silver|silver chloride) are often employed, but the measured electrode potentials can be converted to the hydrogen scale. See: Activity; Fugacity; Saturated calomel electrode. Steam reforming The reaction of fossil fuels with steam at high temperature to generate a mixture of hydrogen and carbon monoxide (‘synthesis gas’). It is the dominant method of commercial hydrogen production and is based on reacting methane (natural gas) with water; carbon dioxide is formed as a by‐product. See: Synthesis gas. Steam‐to‐carbon ratio The number of moles of water per mole of carbon in either the reformate or the fuel streams. This term is used when steam is injected into the reformate stream for the water–gas shift reaction or into the fuel for steam reforming. See: Reformate; Steam reforming; Water–gas shift reaction. Stoichiometric ratio The perfect oxidant to fuel ratio in a reaction such that all of the oxidant exactly reacts with all of the fuel. Stoichiometry The branch of chemistry concerned with the exact or fixed relative proportions of elements in a chemical compound or of reactants to produce a compound. For example, in carbon dioxide, the stoichiometric ratio of carbon atoms to oxygen is 1 : 2. Stoichiometric amounts satisfy a balanced chemical reaction with no excess of reactants or products. Sub‐bituminous coal A medium‐soft coal (carbon content 35–45 wt.%) with properties that range from those of lignite to those of bituminous coal. It is used primarily as fuel for electricity generation and is an important source of light aromatic hydrocarbons for the chemical synthesis industry. See: Anthracite; Bituminous coal; Lignite; Peat. Substitute natural gas A fuel gas with similar properties to those of natural gas and that can be produced from fossil fuels (such as lignite coal) or from biofuels. It can be distributed in the natural gas grid, provided it fulfils the strict criteria for net gas feeding. Abbreviated as ‘SNG’. Sustainable energy (‘sustainability’) A set of energy technologies that meets humanity’s needs on an indefinite basis without producing irreversible environmental effects. (Note that various definitions exist in the literature, but they all convey the same message.) Synthesis gas A gas mixture that contains varying amounts of carbon monoxide and hydrogen. Examples of production methods include the steam reforming of natural gas or liquid hydrocarbons, the gasification of coal or biomass and some types of waste‐to‐energy gasification processes. Also known as ‘syngas’. See: Steam reforming. Synthetic natural gas Methane produced by the catalytic reaction of carbon monoxide with hydrogen or from coal by reaction with hydrogen. Tafel equation The relationship between the current flowing and the overpotential of an electrode. A plot of electrode potential versus the logarithm of current density is called the ‘Tafel plot’ and the resulting straight line the ‘Tafel line’. The slope Glossary of Terms provides information on the mechanism of the electrochemical reaction, and the intercept on the current axis (abscissa) provides information on the rate constant (and exchange‐current density) of the reaction. See: Butler–Volmer equation; Exchange‐current density. Tape casting A method for the production of thin, flat ceramics. The ceramic powder is blended with a liquid and binder (this mixture is referred to as a ‘slip’) and then deposited on a moving flat surface that is passed under a flat blade to create a continuous tape. The tape is heat treated to remove the liquid phase and binder and sintered to promote bonding between the ceramic particles. Theoretical energy density The energy output of an electrochemical cell referred to the volume of only the active materials and a 100% utilization of these materials, expressed as Wh L−1 or Wh dm−3. Theoretical specific energy The energy output of an electrochemical cell referred to the weight of only the active materials and a 100% utilization of these materials, expressed as Wh kg−1. Thermal efficiency For a heat engine, the ratio of the useful work done by the engine in a given time interval to the mechanical equivalent of the heat energy supplied in the steam or fuel during the same time interval. Thermal expansion coefficient A parameter used to express the dimensional response of a given solid material to a given unit change of temperature, specifically, the ratio of change in dimensions to original dimensions per degree rise in temperature. Thermochemical (hydrogen) cycle A multistep chemical reaction that sums to the overall production of hydrogen (and oxygen) by water decomposition; the Carnot efficiency of the step performed at the highest temperature places a theoretical limit on the overall hydrogen production efficiency of such a cycle. See: Carnot efficiency. Three‐phase boundary The gas|electrocatalyst|electrolyte interface formed in a fuel‐cell electrode such that the electrocatalyst has simultaneous contact with the reactant gas, the ionic conductor (electrolyte) and the electron conductor. The electrochemical reactions occur at these points of simultaneous contact. Also known as a ‘triple‐phase boundary’. Tortuosity The distance a molecule or ion must travel to get through a substance film divided by the thickness of the substance. Town gas See: Coal gas. Transformer An electrical device that steps up or steps down voltage. Transformers work with alternating current only. Transport number The fraction of the total current flowing in an electrolyte phase that is carried by a particular ion. Also known as ‘transference number’. Triple‐phase boundary See: Three‐phase boundary. Unitized regenerative fuel cell A unitized regenerative fuel cell based on a proton‐ exchange membrane that can perform the electrolysis of water in regenerative mode and function in the other mode as a fuel cell by recombining oxygen and hydrogen to produce electricity. Abbreviated as ‘URFC’. See: Regenerative fuel cell. Valence A number that indicates the combining power of one atom with others, that is, the number of other atoms with which it can combine. VAR (volt‐amperes reactive) A unit of reactive power in a circuit that is carrying a sinusoidal current. A VAR equals the amount of reactive power in the circuit when 435 436 Glossary of Terms the product of the root‐mean‐square value of the voltage (volts) by the root‐mean value of the current (amperes), and by the sine of the phase angle between the voltage and the current, equals 1. Viscosity The resistance of a fluid to shear forces and hence to flow. Voltage The difference in potential between the two electrodes of a cell or the two terminals of a battery. Voltammetry An electrochemical measuring technique used for electrochemical analysis, for the determination of the kinetics and mechanism of electrode reactions, and for corrosion studies. ‘Voltammetry’ is a family of techniques with the common characteristics that the potential of the working electrode is controlled (typically with a potentiostat) and the current flowing through the electrode is measured. ‘Linear‐sweep voltammetry’ involves scanning the potential linearly in time (the plots are known as ‘voltammograms’). ‘Cyclic voltammetry’ is a linear‐sweep voltammetry with the scan continued in the reverse direction at the end of the first scan; this cycle can be repeated a number of times. In alternating current (AC) voltammetry, an alternating voltage is superimposed on the direct‐current ramp. Water-gas A mixture composed primarily of hydrogen and carbon monoxide produced by passing steam over incandescent carbon, usually from anthracite coal or coke. The reaction is strongly endothermic but may be combined with the exothermic reaction for producer gas. Used for lighting (mainly during the 19th to early 20th century) and as a fuel (well into the 20th century). See: Producer gas; Steam reforming; Water-gas shift reaction. Water-gas shift reaction The reaction of water–gas with steam to yield hydrogen and carbon dioxide. Well‐to‐wheels analysis See: Life‐cycle analysis Working electrode The electrode in an electrochemical system on which the reaction of interest is taking place. The kinetics and mechanism of the reaction may be under investigation, or the reaction occurring on the working electrode may be used to perform an electrochemical analysis of the electrolyte solution. The electrode can serve either as a positive or a negative according to the applied polarity. See: Counter electrode. Xerogel A solid formed from a gel by drying with unhindered shrinkage. Xerogels usually retain high porosity (25%) and high surface area (150–900 m2 g−1), along with a very small pore size (1–10 nm). X‐ray photoelectron spectroscopy A quantitative technique used to determine the elemental composition, empirical formula, chemical state and electronic state of the elements on the surface of a material. The analysis is performed under ultrahigh vacuum conditions. Spectra are obtained by irradiating a specimen with a beam of X‐rays while simultaneously measuring the kinetic energy and number of electrons that escape from the top 1–10 nm of the material under investigation. Also known as ‘electron spectroscopy for chemical analysis’ (ESCA). Abbreviated as ‘XPS’. Zeolite Any one of a family of hydrous aluminium silicate minerals with a cage‐like molecular structure; used chiefly as molecular filters, ion‐exchange agents and catalysts (either used directly, e.g., in petroleum refineries, or loaded with catalysts for other chemical reactions). 437 Index a Absorption 74, 137, 155, 204, 276, 293, 313, 334–7, 345 Acetaldehyde 157, 171 Acetonitrile 313 Acid electrolyte 7–8, 50, 58, 161, 163, 172, 174, 307, 341 Acid fuel cell 7, 69, 136, 172 AC impedance 65 spectroscopy 61–2, 64, 190 AC power 371, 378 source 373 Activation losses 46, 60, 66–8, 312 Activation overpotential 48–9, 51–2, 59, 65–7, 153–4, 162–3 Activity 36–7, 86–8, 172–3, 205–6, 254–5, 415, 433–4 electrocatalytic 213 Acumentrics, Inc. 255–6 AC voltage 62, 375, 378, 379 Adsorption 88, 176, 293, 337, 345, 415, 418, 420, 430 dissociative 85, 161, 171 pressure swing 292, 431 Advantica Technologies Ltd. 301 Air‐breathing fuel‐cell stack 106 Air compressor 116, 118, 310, 356, 411–13 Aircraft 169, 268 Air electrode 20, 57, 82, 100, 149, 158, 205, 249 Air‐electrode supported (AES) 249 Air flow 94, 98, 124, 364 Air humidity 96, 100 Air usage 357, 405–7, 411 Air utilization 192 Alanates 337, 342, 415 Algae 268, 272, 319 Alkali metal carbonates 207, 211 hydrides 325–6, 339–40, 415 Alkaline electrolyser 150, 306, 310, 312, 337 Alkaline electrolyte 8–9, 17, 20, 69, 150, 155, 162, 170, 172, 176, 305 Alkaline fuel cell (AFC) 6–8, 17–18, 69–70, 135–47, 149–56, 160, 178–9, 207, 305–6, 341–2 Apollo 7, 17–18, 136, 140–2, 153–4 Bacon 147 catalysts 149 dissolved fuel 142–3 membranes 144, 146 stacks 140, 149–51 systems 18, 139, 142, 310 Alkaline proton‐exchange membrane fuel cell (APEMFC) 144 Alkanes 266–7 Alkylbenzene 73 Allied Signal Aerospace Co. 257 Allis‐Chalmers 139 Alloys, AB5, 180, 334, 415 Alstom 395 Altergy 123 Alternating current (AC) circuit 368, 378–9 Fuel Cell Systems Explained, Third Edition. Andrew L. Dicks and David A. J. Rand. © 2018 John Wiley & Sons Ltd. Published 2018 by John Wiley & Sons Ltd. 438 Index American Gas Association 7, 205 Amide 341 Amine 145, 340 scrubbing 415 Ammonia 87, 142, 198–9, 264–5, 273–4, 295–6, 325, 343, 389, 415 borane 340–1 Anaerobic 273, 319 Analogue‐to‐digital converter 377 Anion‐exchange membrane (AEM) fuel cells 144–6, 155 Anionic membranes 175, 182 Anion‐permeable membrane 180–1 Anode 3, 7–9, 11–16, 142–4, 157–67, 179–81, 207–10, 219–22, 233–6 alloy 213 catalyst 88, 115, 162–3, 166, 172–5, 177–8, 180, 198, 204, 267 ceria cermet 245 composite 246 copper cermet 284 cylindrical 249 palladium 180 reaction 160, 162, 167, 170, 172, 174, 203, 213, 282–3 Anode‐supported electrolyte cells 244 Anthracite 269, 416–17, 427, 434 Apparent power/kVA 379–90 Area specific resistance (ASR) 55, 57, 242–3, 416 Arrhenius 415 Asbestos 142 Atomic mass unit (amu) 29 Austenitic steel 248, 425 Auto‐ignition 264, 330 Autothermal 173, 301, 416 reforming 285–6, 303 Auxiliary power units (APUs) 132, 183, 255, 304 Avogadro’s law 53 Avogadro’s number 29 Axial fans 363–4 Axial flow design 352 b Bacon 6, 135, 137, 147, 153, 155, 397 Bacteria 169, 264, 273, 292, 318, 423 anaerobic 320 photosynthetic 430 Baker 132, 230, 275, 344 Balance‐of‐plant (BoP) 114, 128, 351, 389 components 23, 196, 228, 291, 376 Ballard, Geoffrey 7, 125, 126 Ballard Power Systems (BPS) 16, 70–1, 77, 125, 300, 304, 395, 398 Barium zirconate 294 Barriers 25, 72, 306 bubble 213 selective 427 Battery 1, 4, 6, 31, 182, 350–1, 387, 398, 419, 421, 425 bromine 20, 22 cadmium 349 chargers 122, 157, 380 gaseous voltaic 4 lithium 126–7, 182 metal hydride 126 primary 431 redox 423 secondary 31, 131, 416, 418, 431–2 vanadium 21 BIMEVOX 241 Binder 145, 148, 198, 210, 416, 434–5 mixture 213 organic 212 Bioethanol 169 cellulosic 386 Biofuels 168, 263, 265, 268, 272–4, 275, 318, 323, 417, 434 Biogas 223, 230, 272–4, 417 anaerobic 225 compositions of 274 Biological digestion 273 fuel cells 23, 417, 428 hydrogen generation 318 shift reaction 320 Biomass 169, 272–4, 321, 386–7, 417, 434 Biophotolysis 319, 417 Bio‐reactor 320 Bio‐separation processes 145 Index Bipolar plates 11–16, 54–5, 89–90, 102–3, 105–12, 114, 149, 151, 213, 237–8 AFC stacks 152 alkaline electrolyser 306 construction 110 designs 72, 215 flow‐field plate 82 metal 109 multilayer 199–200 ribbed 199 Bloom 257 Blowers 120, 124, 137, 224, 306, 351, 353, 363–4, 416 BMW 305, 324, 333–4 Bode plot 62, 64 Boilers 195, 201, 227, 362 gas‐fired 425 Borohydride 144, 179–81, 342–4 alkali metal 342 decomposition 179 Boudouard reaction 221, 245, 280–1, 421 Bromine 20, 22 Buckminsterfullerene 345–6 Bus 71, 125–6, 130, 264, 267–8, 304, 328, 366, 375, 395 articulated city 139 diesel 127 Butane 268, 272, 324, 427 Butler–Volmer equation 49, 393, 417–18, 434 c Cadmium telluride 313 Calcia‐stabilized zirconia (CSZ) 239 Calomel 60, 432, 434 Calorific value 27 Caltex Oil Corporation 206 Capacitance 59–60, 65–6, 367, 373, 380, 417 double‐layer 65, 421 Capacitative discharges 290 Carbohydrate 319, 423 Carbon 16–17, 163–4, 231–4, 277–8 activated 276, 282, 417 bipolar plates 107, 110 capture and storage 432 catalyst 163, 198 cloth 90 corrosion 89, 132, 199, 205 deposition 173, 219, 221, 246, 280–1, 301, 390 electrodes 148, 345 fibre 89, 108, 148, 198, 419 filaments 281 functionalizing 90 graphitic 107 graphitized 129 multiwalled 84, 346 nanofibres 344–6 nanomaterial 289 nanotubes 87, 91, 163, 345 paper 90, 198, 320 turbostratic 234 Carbonates 70, 137, 142, 149, 210–11, 213–14 deactivation 150 electrolyte 197 mixtures 212 molten 233 poisoning 219 precipitates 144 Carbon dioxide 161–2, 207–9, 270–1, 273–4, 319–21, 431–2, 434 atmospheric 318 emissions 25, 416 removal 138, 155, 415 separation 23, 230 Carbon monoxide 202–4, 209, 235–6, 273, 403–4, 408–9, 434 adsorbed 171 dehydrogenase 320 fuel cell 209, 403, 404 tolerance 76 Carnot cycle 5, 32, 417 efficiency 418, 435 limit 33, 34 Catalyst 52, 61, 114–15, 143, 147–8, 149–50, 163–7, 168–74, 176, 200, 218–20, 275–8, 280–1, 288–92, 338–41, 380 carbon‐supported 89, 148 chromium 290, 303 cobalt 276 439 440 Index Catalyst (cont’d ) composite 131 copper 290–1 degradation 88, 129, 132, 219, 301 layers 68, 78, 82–3, 85, 92–4, 164, 198–200, 392–3 molybdenum oxide 276 nickel alloy 205 non‐precious metal 84, 86, 167, 172, 205 overpotential 100, 117, 221–2, 230, 247, 393 palladium 177 particles 10, 83, 90, 285 photoelectrochemical 314 platinum‐based 82, 87, 136, 286 poison 189, 219 precious metal 187, 235, 291, 340 reactions 143–4, 148, 207–9, 236 thorium‐based 273 titanium 342 Catalytic partial oxidation (CPO) 285, 303–5 Catalytic rich gas (CRG) 269 Cathode 7–9, 11–16, 50–2, 159–67, 195–9, 207–10, 233–8 Cation 2–3, 180–1, 241, 245, 429, 433 Cation‐permeable membrane 181 Cella Energy Ltd. 344 Ceramatec Inc. 257 Ceramic Fuel Cells Ltd. (CFCL) 258 Ceramic materials 8, 16, 19, 212, 226, 235–7, 294, 426 Ceramic processing techniques 240 Ceramic substrate 258, 293 extruded 257 Cerate 294 Ceria 172, 240–1, 245–6, 291, 305, 317 cerium gadolinium‐doped (CGO) 241, 244, 245 samarium‐doped 241 Cerium 180, 215 Cermet 235, 244–6, 294, 418 Chalcogenide 87, 167, 418 Channels 13–15, 90, 105–6, 114, 301, 393 cooling water 200 flow‐field 82, 108, 218 staggered 424 Charge 65, 123, 127, 370–2, 380, 382–3, 405–6, 416, 418, 429 accumulation 421 carriers 69, 419, 425 distribution 295 electric 240, 425 electrical 58, 417 flow 8–9 state 22, 182 transfer 48–9, 58, 64, 382, 415 Chelates 86, 418 Chemical looping 314, 317–18 Chemical structure 72–3, 78 Chemical vapour deposition (CVD) 240, 294, 345, 418 Chemical vapour infiltration 108 Chemisorption 415, 418, 430 Chloride concentration 432 Chloride ion 305 Chloromethylation 145 Chlorophyll 319 Choking limit 362 Chromite 247 Chromium 147, 205, 213, 237, 248, 416 photocatalyst 313 poisoning 238 steels 330 Clathrate 347, 418–19 Clean Urban Transport Europe (CUTE) 298 ClearEdge Power 397–8 Climate 399, 419 Coal 5–6, 34, 225, 234, 263, 265, 268–71, 284, 317, 323, 419, 424, 429, gas 19 anthracite 436 bituminous 269, 416–17, 427, 434 brown 427 coking 417 gasification 6, 233, 273, 434 hard 269, 416 lignite 434 tar 419 Coal‐bed methane (CBM) 270–2 Coal seam gas (CSG), see Coal‐bed methane (CBM) Coating 108–9, 168, 215, 294 Index Cobalt 85–7, 129, 148, 150, 172, 205, 215, 273, 276 Cobalt chloride 342 Cobalt oxide 291, 314 Co‐flow 39, 111, 191, 217, 284 Cogeneration 254, 286, 378, 389, 397, 419 Coke 269, 417, 419, 436 Collector 367, 370 current 11, 112–14, 149, 213, 215–17, 244, 255, 258, 312–13, 424 Combined‐cycle systems 253, 419 Combined heat and power, see Cogeneration Combustion 5, 32, 193, 249, 251, 264, 269, 296, 338, 387, 390–1 catalyst 300 chemical looping 317–18 enthalpy 391 partial 286, 302–3 pressurized 299 products 299, 317, 427 reaction 300 Combustor 208, 231, 298 Commercialization 18, 25, 72, 205, 225, 254–5, 384, 394 sustained 182 Commonwealth Scientific and Industrial Research Organisation (CSIRO) 258, 383 Compact regenerative reformers 299 Composite curves 196 Composite cylinders 159, 328–9, 347–8 Composites 108, 346, 419 cylinder 137, 159 hydrogen storage 330 metal‐ceramic 293 mica‐based 248 Compressed gas 158, 310, 324, 327–8, 337, 348, 356, 385 Compression 107, 116, 288, 310, 326–7, 332, 354–5, 412 isothermal 418 mechanical 132 moulding 108 ratios 353 Compressor 23–4, 98, 118–20, 351–6, 358–9, 361–2, 411–13 anode recycle 124 axial 354–6 centrifugal 353–4, 358–9, 361, 364 efficiency 118, 355 multistage 358 recycle 193 rotating vane 354 screw 353–4, 413 Concentration gradient 293, 420 Concentration losses 46, 94 Concentration overpotential 59, 131, 419 Conduction 214, 239 band 313 electronic 241, 244 Conductivity 75–6, 80–1, 137, 148, 152, 202, 240–3, 245, 254, 432 mixed 247 proton 294 Conductor 235, 290, 417 fast oxygen‐ion 240 mixed ionic–electronic conducting (MIEC) 243, 246, 284, 294–5, 301 oxide‐ion 238, 243–4 proton 81 Contaminants 124, 137, 270, 302, 420 Control circuits 370–1 elements 392 switches 373 valves 23, 164, 392 Controller 124–5, 377, 381, 431 programmable logic controller (PLC) 129 Convection 104, 165, 299, 427 Conversion efficiency 177, 182, 307, 319–20 Cooling 102, 107–9, 114, 116, 197, 290–1, 296–7, 351, 354, 356 air 15, 104–6, 138, 141, 194, 231, 363 channels 105, 107, 128, 200 coils 306 curves 195 fluid 14, 105, 107, 112, 142, 187, 200 Copper 1, 86, 109, 213, 246, 280, 306, 319 cermet 284 441 442 Index Corrosion 88, 107, 109–10, 114, 144, 201, 215, 230, 237, 267 resistance 142, 210 Cost 25, 39, 70, 72, 85, 98–9, 120–1, 149, 168, 193, 213, 219–20, 291, 330, 394–5 life‐cycle 388 reduction 293 target 158, 385 Coulombs 29–30, 422 Counter‐current 299 Counter‐electrode 60 Counter‐flow 39, 111, 217 Cracks 212, 289, 330 Creepage 210, 219 Cross‐flow configurations 39, 200 Crossover 46, 52, 54, 85, 143, 165–7, 173, 176–8, 182, 197 Crude oil 266–70 Cryogenic 420 liquid 265, 326, 331 storage vessels 153 Crystal structure 241, 247, 293, 429 fluorite 240 Current density 24, 43, 46–7, 49–50, 67, 70–1, 139, 148–50, 201–2, 204, 312 distribution 39, 112, 191 limiting 57 local 112, 188, 190 Current‐interruption 65 Current losses 54 Current path 113, 138, 149, 253, 367, 369, 371 Cyanobacteria 319–20, 430 Cyclic voltammetry (CV) 60–1, 84–5, 355, 420, 436 Cyclohexane 267, 289, 338–9 Cyclopentane 267 d Daimler 125–7, 157, 344, 395 Dark fermentation 321 Davy, Sir Humphry 1, 2 Dead ended 115–16, 124, 420 Decomposition 5, 280, 323, 334, 345, 430 anaerobic 417 chemical 423 direct 179 oxidative 417 rate 343 Degradation 25, 93, 129, 137, 144, 146, 197, 199, 219, 237, 245, 248, 337, 378, 380 acid‐catalysed 77 mechanical 129, 245, 252, 293 permanent 247, 295 progressive 310 Delphi Automotive LLP 257 Density functional theory (DFT) 394, 420 Desorption 293, 415 cycles 137, 334–7 pressure 335 reaction steps 175 Desulfurization 23, 189, 275, 296–8, 390 Desulfurizer 193, 224, 227, 269, 282, 297, 299 Detergents 73, 267 Devolatilization 269 Dewar 324, 331 Dew point 95, 102 Diaphragm pump 365 Dibenzyl toluene 339 Diborane 325 Dicks, A. L. 63–4, 91, 259, 300 Dielectric 200–1, 420 Diesel 121, 159, 263, 267, 276, 279, 300 auxilliary power unit (APU) 305 engines 347, 351, 387 fuels 266, 268, 304, 417, 420 low‐sulfur 277, 386 oil 323 synthetic 273, 323 systems 354 Diethyl sulfide 272 Differential pressure 116, 147, 153, 221 control valve 124 Diffusion 65, 115, 142, 427 barrier 302 bonding 107 characteristics 293 coefficient 76, 146, 166, 420 limitations 220 molecular 294 Index overpotential 419 rate 240 water 91–2, 100, 165 Digester gas 223 anaerobic 417 Digestion 169, 273, 318 anaerobic 225, 272–3, 321 microbial 320–1 Digital storage oscilloscope 65 Diode 58, 368–73, 375, 420 Diphenyl isophthalate 78 Direct anodic oxidation 162 Direct borohydride fuel cell (DBFC) 17, 179–81 Direct carbon fuel cells (DCFC) 6, 17, 233–4 Direct combustion 272 Direct current (DC) 23, 368, 376, 378, 387, 391, 399, 415, 420, 426, 430 DC/AC conversion 125, 371, 378, 381, 384, 396 inverter 381 output 125, 365 power 205, 312, 351 regulators 366 voltage 35, 366, 370, 372 Direct ethanol fuel cell (DEFC) 158, 169–73, 182, 184 acidic 170 platinum‐free 172 Direct ethylene glycol fuel fells (DEGFC) 158, 174, 176 Direct formic acid fuel cell (DFAFC) 158, 176, 178 Direct fuel cell (DFC) 218, 227–9, 231–2 Direct internal reforming (DIR) 218–19, 282–3, 420, 423, 426 Direct liquid fuel cell (DLFC) 157–9, 161, 163, 165, 167, 169, 171–9, 181–3 Direct methanol fuel cells (DMFC) 17, 51, 67, 77, 157–8, 160, 162–3, 167, 183–5 anode 161–3 mixed reactant 115 passive 165 stacks 182 system 164–5, 182–3 Direct oxidation of hydrocarbons 284 Direct propanol fuel cell (DPFC) 158, 173, 174 Discharge 20–3, 334, 336, 338, 416, 418, 421 atmospheric pressure gliding arc 289 corona 290 dielectric barrier 290 microwave 290 Disproportionation 280, 420 Dissociation 85, 161, 293, 334 constant 421 Distributed generator (DG) 321, 378 Distribution networks 308, 372–3 Dopants 80–1, 239–41, 294, 432 acceptor 239, 425 Doped ceria 241, 243, 245 cermet 245 Doped perovskite type oxide 241 lanthanum chromite 247 Doping 213, 240–1, 244–5, 247 Dosing pump 227 Double‐layer 58–60, 65, 421 Dreamcar project 183–4 Drive‐train 35, 125, 127, 421, 431 Dry air 94, 100–1, 407 Drying 95, 172, 212, 433, 436 effect 94 rapid 105 Dry reforming 279 Dupont Corporation 69, 74, 428 Dye‐sensitized solar cell 313–14 Dynamic braking 381 e Eaton supercharger 359–60 Edge connections 113, 149 Edge seals 14–15, 110 plastic 109 Edison, Thomas 231 Electronic conductivity 52, 236, 238–40, 243–7 Efficiency 25, 32–3, 35–6, 187–9, 312–13, 352–4, 356–7, 360–1, 391, 408–9, 411–13 cold gas 318 electrical 259, 282, 291, 296 443 444 Index Efficiency (cont’d ) energy‐conversion 304 engine 387 fuel‐cell 35–6 isentropic 355–6, 361 limit 34 mechanical 356, 362 photon 313 solar 320 tank‐to‐wheels (TTW) 347 thermodynamic 33 well‐to‐wheels (WTW) 387 Electrical permittivity 59 Electric vehicles 20, 35, 125, 330, 350, 381, 395, 398, 421, 430 Electrocatalysis 171, 184 Electrocatalyst 4, 11, 76, 247, 306, 320, 421 Electrochemical impedance spectroscopy (EIS) 61–2, 64–5, 68, 191, 421, 428 Electrochemical oxidation 170, 180, 209, 292 Electrochemical vapour deposition (EVD) 240, 249, 256 Electrode 10–16, 18–22, 45–52, 57–62, 88–92, 147–52, 431–6 area 10, 166, 212 catalysts 115, 152, 189 counter 60–1, 313, 419, 430, 436 degradation 85 kinetics 51, 421 photosensitive 430 potentials 85, 433 reactions 8–10, 21, 23, 47, 52, 81, 90, 136, 160, 162, 170, 423, 428, 436 rolled 147–8 titania 313 Electrolyser 44, 130, 261, 305–10, 312, 323, 385, 389, 422 high‐temperature 312 Electrolysis 2–4, 264–5, 305–8, 323, 333, 385–7, 416, 418, 422 Electrolyte 2–4, 6–15, 52–5, 89–93, 196–9, 207–16, 218–21, 234–46 Electromagnetic force 5 Electron 7–8, 29–32, 52, 58–9, 160–3, 170–1, 180, 312–13, 420–2, 424–6 donors 320 free 290 micrograph 82, 422 moles of 29, 207, 405 pairs 161–2 Electro‐osmotic drag 77, 92, 100, 142, 174, 422 Electro‐oxidation 173 Electrospray processes 84 Elenco 139–40 Embrittlement 425 Emissions 25, 184, 264, 276, 287, 387–8 reduction 383 Energy 7–9, 157–9, 348–51, 379–83, 387–9, 422–4, 431–3 activation 9, 415 density 182, 324, 349, 408, 422 Energy for You (EFOY) 157 Enthalpy 27–8, 35, 193, 296, 334, 345, 401, 408, 422–4 Entrained‐bed gasifiers 269 Entropy term 33, 312 Equilibrium 27, 48, 94, 193, 222, 277–8, 286–7, 393, 423 constant 222, 280, 421 thermodynamic 218, 280, 290, 390 Equivalent weight (EW) 74–5, 77, 166, 423 Ethanol 18, 23, 61, 83, 157, 159, 169–77, 264, 272, 274, 323, 325, 386–7, 417, 423 Ethylene glycol 142, 157, 174–6 Eutectic mixtures 211, 214, 423 Exchange‐current density 47–9, 51, 54, 57, 82, 117–18, 153, 394, 423, 434 Exergy 27, 193–4, 423 Expansion coefficient 244, 247 f Fans 104, 106, 114, 120, 164, 306, 351 centrifugal 364 Faraday constant 29–30, 43, 166 Faraday, Michael 2–3, 5 Farads 59, 417 Fermentation 169, 272, 274, 320, 423 Ferritic steels 248 Index Filling stations 308–9, 332–3, 385, 387 Fisher–Tropsch reactors 386 Flame traps 330 Flammability 168, 264 Flooding 73, 78, 90, 92, 102, 148, 164 Flow batteries 20–1, 131, 182, 423, 431 Flow‐field 12, 89, 102, 107, 109, 111–14 anode gas 215 interdigitated 11, 100, 103, 111 Flowsheet 385, 390–1 Fluorine‐containing polymers 146 Fluorite structure 240, 245 Formaldehyde 162, 168–9, 175 Formic acid 18, 157, 162, 169, 175, 177–8 Formic acid fuel cells (FAFC) 158, 176 fossil fuels 5, 26, 33, 232, 263, 265–8, 306, 416, 424, 434 Four‐quadrant inverter 383 Free energy 27–30, 37, 41, 117, 143, 160, 193–4, 222, 237, 404, 424, 428 barrier 418 change in molar Gibbs 401–4 of formation 28, 43 standard 28 Frequency response analysers (FRAs) 62 Fuel cartridge 123 Fuel cell 1–11, 13–21, 23–8, 30–41, 54–60, 91–8, 187–90, 378–85, 389–99 bio‐electrochemical 416, 428 borohydride 144, 178–9, 180, 342 circular 136 compact 70 direct acid 176 direct alcohol 171, 174–5 direct carbon 6, 17, 231, 233 mixed reactant 114 mobile 169, 277 portable applications 184 regenerative 132–3, 432, 435 stationary 265, 298 systems 192–5, 273–5, 278–9, 289, 291, 295–6, 351, 370–2, 378, 384, 387–9, 394–5, 397–8 Fuel cell vehicles (FCVs) 155, 157, 263, 267, 302, 304, 308, 323–4, 329–30, 332–3, 344, 347, 384–95 Fuel infrastructures 191 Fuel pretreatment 223 Fuel processing 76, 120, 188, 208, 263, 275, 281, 290–1, 321, 384, 389 practical 295, 304 Fuel utilization 39, 41, 177, 188–9, 191–2, 249, 391, 408 Fugacity 36, 415, 424, 434 Fullerenes 345 Functionalization 86–7, 145–6 g Gadolinium 241, 294 Gadolinium‐doped ceria (GDC) 241 Gadolinium strontium cobaltite 247 Gallium 241 Gallium arsenide 313 Gas adsorption isotherms 416 Gas battery 5 Gas crossover 116, 214, 221 Gas density 357 Gas‐diffusion electrodes 198, 205 modern 155 Gas‐diffusion layer (GDL) 82, 89, 112, 114, 144 macroporous 164 Gas engines 274 Gas hydrates 268 Gasification 269, 273, 317, 424 Gas manifold 227 Gas micro‐turbines 398 Gasoline 157, 159, 168–9, 263–4, 266–7, 276, 279, 304–5, 347–8, 387, 420, 424, 429 Gas phase transport 220 Gas purification 293 Gas separation 287, 293–4, 343 Gas shift reaction 121, 189, 209, 275, 320, 434, 436 Gas‐tight seals 114 high‐temperature 249, 252 Gas‐to‐liquid 285 Gas turbines 222, 253, 258–9, 263, 398, 419 Gate turn‐off (GTO) 368 Gemini 6, 69–70 General Atomics 316 General Electric (GE) 69–70, 307 445 446 Index General Motors (GM) 127, 304, 385–7 Gibbs 27–30, 36–7, 41, 117, 143, 193–4, 222, 237, 284, 418, 424, 428 energy 401 equation 28, 424 free energy of combustion 193 free energy of formation 28–9 function 401 Glass ceramics 248 Glow discharges 290 Glycolic acid 175 Gold 109, 180, 245, 291, 430 Grain boundary 424 Graphene 87, 91, 345 Graphite 12, 16, 107–9, 234, 345, 424 Greenhouse effect 419, 424–5 Greenhouse gases 263, 387, 419, 424–5 Grid 110, 372–3, 377–80, 399 electricity distribution 378 local 378 Grid‐connected systems 378 Grotthus mechanism 74 Grove 4–5, 7, 395 h Half‐cell reaction 233, 425 Halogenated fluorocarbons 425 Harmonic distortion 373 Hat management 23, 238, 323, 349 Hazard identification (HAZID) 392 H‐bridge inverter circuit 372 Heat capacities 201, 355–6, 412–13 molar 402, 404 Heat engines 5, 32–3, 35, 187–8, 435 reversible 417 Heat‐exchangers 15, 19, 23, 155, 164, 188, 193–5, 205, 297–8, 302, 364, 390–1, 425 Heating and cooling curves 195 Heat losses 155, 193, 280, 286, 310, 354 Heat of reaction 284, 340, 423 Heat of vapourization 264, 409 Heat pumps 342 Heat sink 368 Heat storage 287, 342 Heat transfer area 193 Heat utilization 218, 227 Heliostats 287–8 Helmholtz 58 Heterocyclic compounds 339 Heteropolyacids 88 Hexis Ltd. 257 Higher heating value (HHV) 33, 274, 312, 327, 425 Hitachi 217, 228 Hole 15, 58, 110, 217, 239, 313, 360, 425 pairs 214 Honda 71, 127, 313 Honeywell 301 Hopping 74 vacancy 239 Horizon Fuel Cells 106, 116, 122–3, 182, 337 Hot Module 216, 226–7 Hotspot 303 Hot stream 194–6 Humidification 101–2, 107, 120, 412 cabin 142 Humidifier 23–4, 101, 116 Humidity 53, 92, 96–8, 100–1, 105, 110, 115, 120, 124, 197, 407 ratio 94 Hybrid cadmium cycle 316 Hybrid drive‐train 387 Hybrid electric vehicle (HEVs) 304, 383, 385, 387, 425, 429–32 Hybrid flow batteries 20 Hybrid sulfur cycle 316 Hybrid systems 156, 254, 256, 259–60, 380–82, 387 Hydrazine 142–3, 325, 341 Hydrides 122, 159, 310, 333–40, 343, 348–9 Hydrocarbon fuels 17, 168, 210, 245–6, 263, 284, 316, 409, 423, 429 synthetic liquid 273 Hydrocarbon oxidation 233, 245 direct 284–5 Hydrocarbon polymers 77 non‐fluorinated 78 Hydrocell 152 Hydrocracking 267 Hydrogen 7–10, 16–19, 25–30, 112–16, 124–7, 187–92, 263–5, 308–10, 317–21, 323–45, 347–9, 434–6 Index alloys storage 180 carrier 340–1, 343 consumption 53, 138 distribution 340 economy 26, 304, 306, 425 electrode 48, 50, 422 embrittlement 293, 310, 330 energy 26, 28, 64, 72, 155–6, 158, 173, 212, 234, 289, 321, 345, 349, 350, 399, 408 fuel cell 27, 29, 31–2, 34–5, 37, 40, 43, 48, 50, 52, 182, 401, 403–4, 406 fuel‐cell 402 generation 264, 278, 289, 323, 350 in graphite nanofibre 344 ions 293, 433 oxidation 9, 82, 85, 158, 202, 394 production 23, 120, 235, 265, 277, 318, 320–1, 339, 341, 343 purification 292 safety 326 separation membranes 81 storage 131, 143, 323–5, 327, 329, 331–5, 337–9, 341–5, 347, 349–50 usage 405, 407 utilization 203 Hydrogenase 318–21 Hydrogenation 272, 305, 425 Hydrogenics 89, 115, 123–4, 308, 395 Hydrogenolysis 276 Hydrogen oxidation reaction (HOR) 8, 50, 82, 88, 203, 401 Hydrogen peroxide 83, 179 Hydrogen sulfide 204, 271, 274, 276, 419 Hydrolyse 179, 343 Hydrophilic 73 Hydrophobic 69, 73, 111, 164 Hydrostik hydrogen canister 337 Hydrous aluminium silicate minerals 436 HyFLEET 298 Hyundai 24, 127 i IdaTech 301 Ignite 327, 333, 342 Ignition 326–7, 330 Ignition temperature 326–7 IMHEX bipolar plate 217 Imide‐forming amides 341 Impedance 62, 64, 190, 421, 425, 428 finite‐length Warburg 65 local 190 mass‐transport 64 Impedance spectroscopy 65, 421, 428 electrochemical 61, 65, 68, 428 Impregnation 90, 150, 198 Impurities 142, 189, 219, 234, 305, 337 Inconel 237, 425 Indirect internal reforming (IIR) 218, 282–3, 420, 423, 426 Inductance 373, 375, 380, 426 Inductor 368–72, 379, 426 Injection 308, 394 direct liquid water 102–3 moulding 108 Ink 83, 244 Inner Helmholtz plane (IHP) 59 Insulation 227, 302, 333 Insulators 114, 420, 432 electronic 11 Integrated gate bipolar transistor (IGBT) 367–8, 370, 373, 426 Integrated Planar SOFC 257 Integrated reforming 283 Intelligent Energy 113–14, 123, 127, 324 Interconnect 139, 237, 247–50, 253, 419, 426 Intercoolers 116, 356 Interdigitated flow‐field design 102 Interfaces 52, 58, 62, 239, 252, 261, 392, 424 anode|electrolyte 245 cathode platinum–Nafion 64 electrochemical 261 electrode|electrolyte 44, 59, 415, 417–18 electrolyte|electrode 245 gas|electrocatalyst|electrolyte 435 solid|liquid 421 stable electrolyte|gas 210 three‐phase 198 Interfacial 418 Intermediary, important 319 Intermediate products 171, 176 partial oxidation to 175, 180 447 448 Index Intermediate temperature solid oxide fuel cell (IT‐SOFC) 237–8, 241, 244–8, 251, 257, 259 applications 245 electrolytes 243 temperatures 247 Internal combustion engine (ICE) 6–7, 116, 267, 329, 359, 384, 417, 425, 429–30 Internal combustion engines vehicle (ICEV) 304 Internal currents 52–3 Internal manifolding 15–16, 216–18 Internal reforming 281 MCFC 275–6, 284 SOFC 393 Internal resistances 55, 59, 62, 282, 367, 426 Internal short‐circuit 140, 214, 426 Internal shunt currents 149 International Civil Aviation Organization’s Dangerous Goods Panel 169 International Fuel Cells (IFC) 7, 18, 136–7, 199–200, 204–5, 298, 300, 397 International Union of Pure and Applied Chemistry (IUPAC) 28, 422, 433 Inverter 127, 224, 366, 371–3, 376, 378, 380, 384, 391, 416, 426 circuits 380 grid‐connected 378 multilevel 375 pulse‐modulated 374 single‐phase 372–3, 377 Iodine 315–16 Ion conductivity 241 Ion conductors 239 fast 433 Ion‐exchange resins 201, 426 Ionic character 73, 80 Ionic compounds 421 Ionic conductivity 75, 145, 202, 240–1, 246 intrinsic 146 Ionic conductor 239, 435 Ionic liquids 79–80 room‐temperature 426 Ionic radius 80, 240–1 Ionomer 73, 76, 168, 426 Ion transport membrane (ITM) 301 Iridium 85, 129, 131, 430 Iron 86–7, 129, 150, 172, 205, 290, 317–18 austenitic 416 hydride 325, 335 oxides 93, 315–16, 390 receiver 232 Irreversible processes 31 Isentropic temperature 355 Ishikawajima‐Harima Heavy Industries (IHI) 132, 228, 300 Islanding 378 Isotherm 336 pressure composition 335 Isothermal conditions 33, 312 ITM Power 308–11 ITM Syngas process 301 j Jacques 232–3 Jet Propulsion Laboratory 182 Junction 10, 58, 245, 313, 367 triple‐point 4 k Kawasaki 80, 332–3 Kerosenes 266, 284, 305, 420 Ketones 178 poly‐ether 146 KIMEX project 133 Kinetic energy 31, 193, 353, 355, 364, 431, 436 Kordesch, K. 135, 148–9, 156 Korea Electric Power Corporation (KEPCO) 228 Kyocera Corporation 255–6, 396 l Lanthanum 180, 241, 243, 246 chromite 237, 248 gallate 241 lanthanum strontium manganite (LSGM) 233–4, 236, 238, 247–8, 255 molybdenum oxide 243 niobate 81 strontium cobaltite 247 Index tungstate 81 vanadate 81 Lanthanum strontium cobaltite ferrite (LSCF) 246 Laser ablation 345 Laves phase 334 Levelized cost 25 LG Electronics Inc. 258 Life‐cycle analysis 388–9 Lignite 416–17, 427, 434 Liquefied petroleum gas (LPG) 266, 268, 272, 324, 427 Liquid air 333 Liquid hydrocarbon 120, 269, 271–2, 305, 423, 434 fuels 347 Liquid hydrogen 155, 324, 327, 331, 333, 335, 348, 385–7 cryogenic 347 Lithium 19, 211–12 aluminate 207 amide 341 hydride 325 Lithium‐ion batteries 126, 130, 349 Load‐levelling 310 Losses 31, 34–5, 40–1, 43, 45–6, 49, 52–4, 56, 98, 100, 119–20, 210, 212–14, 370–1, 393 Lower heating value (LHV) 33, 35–6, 124, 136, 159, 169, 189, 229, 235, 296 Lquefied natural gas (LNG) 271 Luggin capillary 60, 427 Lurgi slagging gasifier 270, 284 Lysholm compressor 352–3, 357–8, 360, 411 m Macrocyclics 86–7, 150 Magnesium 20, 241, 247, 337 Magnetic field 426, 432 induced 368 Manganese 148, 213, 416 Manganese dioxide 31, 150 Manganites 247, 315 strontium‐doped lanthanum 236 Manifolds 14–15, 114, 201, 216–17 Mass balances 195 Mass flow 407 factor 357–8, 411–12 non‐dimensional 357 Mass storage efficiency 329 Mass transfer 64, 286, 419 Mass transport 56, 65–7, 117, 131, 244, 393, 421, 427 Materials handling applications 183 Matrix 141–2, 152, 155, 197, 210, 212–13, 215–16 ceramic 207, 211–13, 221 electrolyte support 216 oxygen ion‐conducting 173 solid 210, 271 McDermott International 257 MC Power 228, 398 Mean time between failures 24 Mechanical energy 385, 421 Mechanical integrity 108, 317, 330, 344 Mechanical stability 213, 240–1, 245 Mechanical strength 197, 212, 416 Membrane 21–2, 74–8, 90–3, 103, 129, 145–6, 165–6, 180–2, 292–5 acid‐doped 78 anion‐conducting 181 bilayer 168 cation 180 cation‐conducting 181 cation‐exchange 426 ceramic 293 composite 172, 419 electrode 100 electrolyte 92, 96 hybrid 185 hydrated 165 ion‐conducting 181 ion‐exchange 426–7, 431 microporous 294 nanocomposite 77 non‐porous 293–4 performance 293 permeable 180–1 planar 301 proton‐conducting 145, 170 reactors 287, 295, 301 tubular 302 WGS reactors 293 449 450 Index Membrane electrode assembly (MEA) 12, 69, 83–4, 91, 93, 100, 109, 111–14, 165 Mercedes B‐class F‐CELL 125 Mercedes‐Benz buses 126 Metal alloys 334–5, 337, 415 Metal bipolar plates 107, 109–10, 129, 237, 251, 306 Metal dusting 281 Metal foam 102, 109–10 Metal hydrides 158, 326, 335–6, 338, 347 inorganic 341 practical 343 rare earth 326 simple 342 Metal interconnects 237 Metal membranes 294 Metal oxide 20, 293, 315 framework 346 proton‐conducting 294 Metal–oxide–semiconductor field] effect transistor (MOSFET) 367–8, 370 Methanal 162 Methanation catalyst 291, 292, 308 Methane 208, 210, 218, 268, 270, 273, 275, 277–81, 284–6, 289, 292, 316–17, 320–1, 326–7, 416–17, 419 conversion 218, 282 electro‐oxidation 64 entrapped 330 hydrates 268 reformed 191 Methanex 168 Methanol 17–18, 121–2, 157–63, 165–70, 173–9, 182–4, 264–5, 267, 274, 279–80, 300, 303–5, 347–8, 386, 389 crossover 77, 167–8 decomposition 343 dehydrogenation 167 electro‐oxidation 161, 177 fuel cell 32 on‐board 347 oxidation 158, 161–3, 167, 170, 172 production 168 reformed 280, 292 reformer 291 safety and storage 168–9 sensor 164 Methylcyclohexane 339 Methyldimethoxysilane 77 Microbial fuel cell 23, 416, 428 Microchannel reactor (MCR) 301 Microcontrollers 130 Microcracks 241 Microdomains 74 Microelectrodes 64 Micro‐electromechanical system (MEMS) 113, 428 Microorganisms 417 living 428 Micropores 198, 294 Microporous 164, 258, 427 polyolefin membrane 22 separators 306 Microprocessor 129, 377 Microscope cross‐section 302 optical 422 Microstructure 132, 244 Migration 3, 197–8, 216, 248, 295 potential‐driven 210 simple 74 Military bases 187 Millennium Cell 344 Mini‐pak 122, 123, 337 Ministry of Economy, Trade and Industry (METI) 228 Mischmetal 180, 336 Mitochondria 169 Mitsubishi Electric 204–6, 228, 301 Mitsubishi Heavy Industries 251, 256–8 Mitsubishi Hitachi Power Systems (MHPS) 255 Mitsui Engineering 257 Mixed reactant fuel cell (MRFC) 114–15 Mixed reforming 279 Mobile applications 17, 304 Mobile homes 182–3 Mobile phones 122, 182, 381 charging 71, 100 Index Model 51, 54, 70, 99, 112, 119, 126, 192, 360, 421 dynamic 389, 392 economic 385 fuel‐cell electrodes 65 hybrid 383 mathematical 393 molecular 345 Module 6, 123–5, 130, 224, 306 multi‐tube 294 tubular membrane 302 Molarity 36, 428 Molecular biology 344 Molecular liquids 214 Molecular weight 264 Mole fraction 166, 428 Moles 29, 36, 96, 181, 194, 207, 279, 340, 390, 405, 408, 428, 434 Molten carbonate 17, 173, 210–11, 213–14, 222, 233, 235, 274 Molten carbonate fuel cell (MCFC) 207, 209, 211–13, 215, 217, 219, 221–3, 225, 233–4 anode 213 cathode 214 electrolyte 173, 211–12, 219 stack 215, 217, 219, 222–3, 225 systems 19, 195, 201, 207, 230–1, 361 Molten sodium 287 Molybdenum 109, 147, 245, 281, 425 Molybdenum nitrides 87 Monolayer 85 Monomers 78, 146, 428 α,β,β‐trifluorostyrene 77 functionalized 146 Monopolar 139, 151, 306, 428 Moon 6, 136 Moonlight Project 228 Motor 23, 116, 118, 361, 365, 381–2, 384, 387, 416, 424, 431 dedicated hub 388 electrical 365 electric‐drive 425 external 352 three‐phase 376 Motorbike manufacturer Suzuki 127 Motor controller 382 Motorola Labs 182 Multiwalled nanotubes 345 Municipal solid waste (MSW) 273, 428 Municipal wastewater 223 Murata Manufacturing 257 n Nafion 64, 69, 72–5, 77–8, 145–6, 166, 172, 174, 176, 428 composite 173 solution 83 structure and characteristics 74 Nafion/silicon oxide 185 Nanocrystalline CoO photocatalyst 314 Nanocrystals 103 Nanomaterials 344 Nanomaterials for Solid State Hydrogen 350 Nanotechnology 198, 344 Nanotubes 345 Naphtha 269, 278–9, 284, 323, 386 crude oil‐based 386 steam reforming of 278–9 Napier grass 272 National Aeronautics and Space Administration (NASA) 69–70, 136, 331, 332 National Institute of Science and Technology (NIST) 433 National Renewable Energy Laboratory 292 Natural gas 189, 218–19, 225, 237, 256–9, 268–72, 274–5, 279–82, 297–8, 302–3, 308, 389–90, 419–20 composition 271 compressed 226, 386 grid 434 liquefied 271 liquids 430 pipelines 308, 310 steam reforming of 26, 168, 203, 272, 296, 312, 385, 434 synthetic natural gas (SNG) 273, 308, 434 vehicles 387 451 452 Index NDC Power 172 Necar 125, 157 Neodymium 180, 243, 246 Nernst equation 36–41, 46, 55, 131, 207, 220, 222, 235, 237, 285, 295, 393, 428 Nernstian effect 153 Nernstian losses 46 Nernst relationship 189 Nernst voltage 253 n‐ethylcarbazole 339 Networks 310, 323, 373, 375, 421 controller area 130 gas transmission 308 linked 433 local 125 New Energy Development Organization (NEDO) 228 Nickel 85–7, 147–50, 205, 208, 213–15, 219, 235–6, 238, 241, 243–6, 273, 276, 279–80, 285, 317 amorphous alloy membrane 293 anode 150, 173, 180, 213 catalyst 150, 284 cathode dissolution 214–15 cermet 284 crystallites 281 electrocatalyst 209 electrode 147 ions 214 matrix 255 mesh 148–9 porous anode 219 sintering 245, 279 Nickel nitrate 87 Nickel oxide (NiO) 153, 208, 210–11, 214–15, 244, 317 Nickel‐YSZ 243 Nicotinamide adenine dinucleotide phosphate (NADP) 319 Niobium 109, 213, 293 Niquist plot 428 Nissan 71, 127, 304 Nitric acid 87 Nitrides 87, 132 Nitrogen 86–7, 94, 96–7, 154, 220, 271, 274, 295, 340–1 Nitrogen oxides 230, 263, 298, 416 Noble metals 109, 131, 209, 236, 419 incorporating 246 Non‐catalytic partial oxidation (NCPO) 302–3 Non‐conductive coatings 290 Non‐noble metal catalysts 158 Non‐porous ceramic membrane 294 metal membrane 293 Non‐pyrophoric 291 Non‐retum flap 227 Non‐slagging 270 Non‐stoichiometric 80 Non‐sulfonated membranes 77 Non‐volatile metal oxide cycles 316 Notice of Market Opportunities (NOMO) 7 Nucleophiles 146 Nyquist plot 62–4, 428 o Odourants 272, 275–6 Ohmic loss 46, 54–5, 59, 65–8, 152, 197, 202, 212, 238, 240, 242, 245, 253–5, 257, 428–9 internal 190 Oil shales 268 Open‐circuit voltage (OCV) 27, 29–35, 37, 39–41, 43–6, 49, 51–3, 66, 117, 153–4, 159, 187, 189–92, 237, 428–9 Operating voltage 24, 35, 41, 57, 59, 212, 222, 312, 391 Orbiter 140, 150, 153 Organic frameworks 86, 346 Original equipment manufacturers (OEMs) 257, 360, 395, 429 Osmotic drag 77, 78 Outer Helmholtz plane (OHP) 59 Overpotential 40, 44, 47–8, 50–1, 57, 59, 64–5, 99, 150, 201, 204, 307, 312, 429–30, 434 Oxide ion 240 Oxygen carriers 317–18 Oxygen concentrators 295 Oxygen evolution reaction (OER) 131, 315 Oxygen‐ion‐conducting membranes 295 Index Oxygen reduction reaction (ORR) 87, 131, 170, 173, 176, 230 Oxygen transport 164 Oxygen usage 406–7 Oxygen utilization 190, 192 82, 85, p Palladium 51, 168, 172, 174, 177, 246, 292–93, 302, 430 Parabolic dish 288 Paraffins 267 branched 266 Para‐hydrogen 333 Parasitic losses 103, 228, 365 Parasitic power load 292 Partial oxidation (POX) 173, 175, 180, 270, 285–6, 416, 429 and autothermal reforming 285 reactors 303 Partial pressure 36, 38–9, 55, 94–6, 131, 189, 192, 202, 207, 214–15, 221, 241, 247–8 Passive systems 165 PCT curve for hydrogen absorption and desorption 335 Perfluorinated sulfonic acid (PFSA) membrane 72 Performance charts 356, 358–60, 363, 413 Permeation 293 controlled 148 Perovskite 150, 241, 246–7, 429 complex 80 cubic 241–2 lanthanum cobaltite 247 structures 80, 241, 247 Petroleum 234, 266–8, 323, 331, 416, 420, 428, 430 crude 266 fractions 276, 279, 285 refining 429 PFSA‐type membrane materials, structure of 75 Phase angle 62, 64, 379–80, 435 Phosphoric acid 17, 69, 78, 80, 196–200, 210, 235 concentrated 196 Phosphoric acid fuel cell (PAFC) 17–19, 38, 69–70, 136, 150, 187–9, 191–3, 195–9, 201–6, 208–10, 220, 274–5, 297, 340–1 catalysts 198–9 degradation 205 development of 204–5 high‐temperature PEMFCs 280 stacks 189, 197, 199, 201, 275 stationary power 19 Phosphorous 243 Photobiological hydrogen production 430 Photoconductivity 396 Photoelectrochemical cell solar 430 Photoelectrode 313 Photoelectrolysis 323 Photolysis 313, 430 electrochemical 313 Photosensors 58 Photosynthesis 313, 318, 417 Photosynthesis‐algae 321 Photovoltaic (PV) 26, 308, 310, 372, 391 array 378 cells 312–13, 389, 395 Phthalocyanines 86, 205 Physisorption 344–5, 415, 418 Pinch point 195–6 Planar configurations 248, 251 Planar SOFCs 238, 252, 254, 256–7 commercialization of 252, 256 single 252 Planar stack configurations 249 Plasma 84, 289 non‐thermal 289–90 spraying 249, 252 Plate reformers 283, 300 Platinum 18, 70, 81–5, 88, 148, 150, 161–3, 167, 171–2, 174, 177, 198, 204–5, 246–7, 291 alloy 85, 163, 175, 205 anode catalyst 103, 150 bulk 84 catalyst 18, 50, 61, 76, 84, 87, 121, 142, 147, 149, 165, 176, 198, 285, 339 dispersing 198 453 454 Index Platinum (cont’d ) dissolution of 88, 129 loading 82, 175 particles 77, 84, 88–9, 199 water‐soluble 82 Plug Power 123, 300 Polarization 45, 429–30 Polyaniline (PANI) 87–8 Polybenzimidazole (PBI) 77–8, 146, 168 acid‐doped 77 alkali‐doped 170 Polybenzoxazine (PBOA) 78 Polycrystalline 396 Polyethylene 69, 72, 108, 310, 339, 344 Polyethylene oxide 146 Polyethylene terephthalate (PET) 174 Poly‐fluorinated sulfonic acid (PFSA) 72–8, 92, 159 ionomers 75 membranes 76, 159, 165, 176–8 Polymer 8, 11, 69, 75, 77–80, 87, 108–9, 123, 169, 200, 293, 341, 419, 426 acid‐doped 78 anion‐exchange 145 anionic 137 films 146 hydrocarbon 146 membranes 69, 92, 129 non‐fluorinated 77 perfluorinated 72 permselective 344 Polymethyl methacrylate 344 Polyphenolic resin 78 Polyphenylene sulfide 77, 108 Polypropylene 108 Polypyrrole (Ppy) 87 Polysulfone 145 Polysulfonic acid 172 membranes 115 Polytetrafluoroethylene (PTFE) 69, 72–3, 83, 90, 147–50, 164, 197–8, 210 Porphyrins 86, 150, 267 Potassium 19, 207, 281 Potassium hydride 325 Potassium hydroxide 136, 305 Potentiostat 61, 430, 436 Power capacity 22 Power conditioner 226, 351, 430 Power density 24–5, 70–1, 113, 115, 128, 159, 163, 172–3, 182, 204, 230, 251, 252, 256–7, 383, 431 Power distribution 375 Power electronics 24, 365–7, 384 Power factor 380, 431 Power factor correction (PFC) 378, 380 Power generation 25, 142, 273, 312, 395, 421 dispersed 265 distributed 421, 425 Power‐to‐gas (P2G) 308 Praseodymium 243, 246 Praxair 302 Preferential oxidation (PROX) 291, 390, 431–2 Preheaters 227, 259 Preheat 193, 208, 249, 296, 303, 356 combustion air 195 recycle gas 195 Pre‐ignition 267 Pre‐reforming 280–2, 390 Pressure 27–8, 36, 56, 94–6, 115–20, 126–7, 153–5, 201–2, 221, 268–70, 278–80, 323–4, 331, 334–5, 358–9 standard 30, 34, 36–7, 41, 402 Pressure lift 361, 364–5 Pressure ratio 119, 352, 357–9, 363, 411 Pressure swing adsorption (PSA) 292–3, 295, 431 Process flow diagram (PFD) 196, 298, 354, 389, 391 schematic 123 Process flowsheet 390 Process integration 188, 296, 310 Propane 223, 255, 268, 270–2, 274–5, 285, 289, 326–227, 340, 423, 427 Propanol 157–8, 173 Proton conductivity 74, 76–8, 81, 146, 168, 172, 176, 294 Proton Energy Systems Inc. 132 Protonex 255 Proton‐exchange membrane (PEM) 6, 8, 17, 69, 71, 73, 83–85, 91–93, 105, 144, 157–58, 166, 306–7, 427–8, 431 Index catalysts 61, 198, 205, 292 cells 311 conductive 168 crossover 85 electrode 89 electrolyser 307, 310, 312 electrolytes 103, 160, 165, 170, 174 low‐temperature 82, 280–1 stacks 69–70, 88, 94, 100, 102, 105, 107, 112, 125–6, 130, 152, 259, 292, 304, 306 Proton‐exchange membrane fuel cell (PEMFC) 17–18, 69–73, 75–85, 87–91, 93–9, 103–5, 107–9, 115–17, 119–21, 131–3, 135–7, 144–5, 157–67, 189–94, 275–7 Proton OnSite 308 Proton transfer 174 Pulse‐width modulation (PWM) 373, 375, 378 Pumping 197, 351 electrochemical 197 water 352 Pumps 13, 19, 21, 23–4, 102, 124, 138–9, 164, 231–2, 351, 358, 362, 364–5 PURASPECTM, 276 Purification 135 Pyrochlores 81 Pyrolysis 82, 245, 272, 289, 419, 424 anaerobic 273 Pyrophoric 291, 342 q Quality standards 378 Quantum 313, 394, 420, 431 Quantum Leap Technology 397 Quaternary ammonium (QA) 145 r Radial turbine 361–3 Radiation 104, 298–9, 313 electromagnetic 9 infrared 424–5 Radiation heat loss 238 Radiators 23, 76, 174 Radicals 129, 421 Radio‐frequency (RF) 290 Rail locomotives 267 Rand, D. A. J. 111, 156, 163, 175, 181, 187, 234, 277, 287, 321, 323, 350–1, 387, 399 Raney metals 147, 149–50 Rate constant 434 Reaction intermediates 176 adsorbed 171, 177 carbonyl 170 Reaction kinetics 115, 343 Reaction mechanisms 61, 170–1, 286, 421 Reactive power/kVAR 379 Real power 379 Rechargeable batteries 19–20, 130, 182, 310, 337, 348, 380–3, 423, 429, 431 Rectification 368 Recuperator 193 Redflow Pty Ltd. 22 Redox couples 423 Redox flow batteries (RFBs) 20, 130–1 Reference states 28, 194, 401, 423 Reformate 283, 292, 305, 431, 434 Reformed hydrocarbons 120–1, 408 Reformer 56, 116, 159, 189, 193, 205, 276, 278–9, 283, 289–90, 296, 298–300, 385, 390, 392 autothermal 295 catalyst 302 external 19, 219 industrial 296 on‐board 264, 347 plates 283 product gas 276, 292, 296 reactor 193, 281, 283, 291–2, 298–9 Reforming reaction 120–1, 188–9, 210, 219, 221, 277–8, 282–6, 299–300 Regenerative braking 382 Regenerative fuel cell systems 131 Regulator circuit 369–70 Relative humidity (RH) 75, 92, 94–5, 97–101, 105, 163, 291, 432 Renewable energy sources 155, 306, 310, 351, 387 storage 324 Reversible adiabatic expansion 415 455 456 Index Reversible metal hydrides 326, 333, 338 Reversible processes 27, 31, 193, 424 Reversible voltage 312 Rolled carbon 148 Rolls‐Royce 237, 257–8 Roots compressors 352–3 Rotating disc electrode (RDE) 61, 85 Roughness 50–1 Round‐trip efficiency 131, 307, 432 Rupture discs 330–1 Ruthenium 85, 88, 131, 163, 287, 305, 313, 430 s Safety 326 Samarium‐doped ceria (SDC) 241 Sankey diagram 124–5 Saturated vapour pressure 95 Screen 66, 257 printing 252 silver‐plated nickel 150 Scrubber 137 Sealing 89, 215–16, 218, 237–8, 252 gasket 106 Seals 11, 113, 143, 151, 216, 248, 251 Selectivity 172, 176 Selenides 87, 418 Self‐discharge 429 Self‐humidification 99–106 Semiconductors 58, 230, 396, 425, 428 crystal 239 gadolinium 313 nanomaterials 314 n‐type 214, 247, 312, 428 three‐terminal power 426 Sensors 326, 377, 392, 428 Separation membranes 294, 302 Separator 217, 432 Sequestration 317–18, 432 Serpentine 11, 110, 111, 424 Shift 189, 209, 222, 235, 259, 275, 286 catalyst 291 converter 297 reaction, reverse 280 Short‐circuit 152, 378, 426, 432 partial 11 Shuttle 17 Siemens 135, 139, 147, 153, 174, 254, 398 Siemens Westinghouse Power Corporation (SWPC) 249–50, 258 Silane 325 Silent discharges 290 Silica 77, 103, 172, 243, 248, 267, 277, 294, 312, 396 membranes 173, 294 Silicon PV cell 396 Silver 1, 61, 147–50, 430 Simulated coal gas (SCG) 203 Single‐walled nanotubes 345 Sintering 88, 129, 213, 219, 243–4, 433 Sliding discharge reactor 289 Sludge 225, 273 sewage 273, 321 Slurry 212–13 organic 343 Smartphones 123, 337 Soda lime 137 Sodium 73, 181, 339, 342–3 Sodium borate 179 Sodium borohydride 82, 142–4, 157, 178–9, 265, 325, 342–3 fuel cells 181 Sodium bromide 21 Sodium carbonate 19, 136, 207, 212 Sodium dithionate 178 Sodium hydride 325 Sodium hydroxide 136, 343 Sodium sulfate 73 Sodium tetrahydroborate 342 Solar 26, 132, 264–5, 306, 308, 387 embodied 286 energy 319, 383 hydrogen production 321 panels 372, 380 thermal 287 Solid oxide fuel cell (SOFC) 17–19, 173, 188–90, 234–41, 243–5, 247–9, 251–5, 257–61, 281–4 auxiliary power 305 high‐temperature 173, 237 intermediate‐temperature 237 internal‐reforming 281 low‐temperature 238 Index multistage 258 stacks 189, 252–3, 258, 282–384 tubular 251 Solid polymer fuel cell (SPFC) 69 Solid State Energy Conversion Alliance (SECA) 254 Solid‐state hydrogen storage 350 Sorbent‐enhanced reforming 287 Space applications 120, 153 Spacecraft 139, 141–2, 324 manned 69 Spark‐ignition 386, 429 Speed factor 358, 363 rotational 357 Stack 11–12, 14–15, 103–4, 115–16, 118–20, 122–5, 128–231, 283, 391–2 area 136 components 213, 217 configuration 215 cooling 192, 200, 282 networking 231 Stainless steel 12, 16, 107, 109, 215, 217, 242 Standard temperature and pressure (STP), conditions 28, 30, 45, 53, 202, 264, 357, 407, 408, 433–7 State‐of‐charge measurement 383 Stationary power applications 23, 70, 125, 390 Steam electrolyser 307 Steam electrolysis 422 Steam methane reforming (SMR) 188–9, 321 Steam reformer 23, 273, 290, 296–8 Steam reforming 188, 277 Steam turbines 6, 33, 187, 194, 258, 362, 384, 419 Step‐down converters 368, 370–1 Step‐up converters 371 Stoichiometric 124, 129, 326, 405, 407–9 Storage 5, 116, 127, 131, 159, 168–9, 233, 323–4, 328–9, 331, 340, 349–50, 417 composite tank 330 cryogenic 333 high‐pressure 333 tank 22, 127 temporary 417 underground 288 vessel 159, 169, 330, 349 Stored energy 337, 380, 431 accessible 422, 433 Strontium 241, 247 cerates 80 Strontium titanate 246 Sub‐bituminous coal 416–17, 427, 434 Submarines 147, 154 Sub‐stacks 201 Sulfides 87, 313, 418 Sulfonated fluoroethylene 73 Sulfonated polyether ether ketone (SPEEK) 77, 168 Sulfonate groups 83 charged 177 Sulfonation 73 Sulfonic acid 73 benzyl 77 group 75, 144, 166 perfluorinated 72, 172 polystyrene 69 trifluorostyrene 77 Sulfur 88, 189, 198, 204, 213, 219, 235, 267–8, 271–2, 275–7, 285, 291, 305, 315–16, 319 compounds 189, 276 content 276, 420 organic 275–6 poisoning 204, 238 removal 23, 296, 304 tolerance 291 Sulfur dioxide 285, 416 Sulfuric acid 3, 145, 315, 433 concentrated 73 Sulfur oxides 230 Sulzer Hexis AG 398 Supercapacitor 383–4, 421, 427 Superchargers 351, 361 Supported electrolytes 242 Supported nickel catalysts 219, 277, 286, 305 Supported Pt 292 Support electrocatalysts 417 457 458 Index Surface area 50, 59, 83, 198, 213, 243, 416, 433 active 84, 88 available active 199 effective 10 Surface coating 107 Surface concentration 245, 419 Surface properties 85, 293 Sweep rate 84 Switching regulator 368 Switching sequence 374, 377 Switch‐mode 368–9, 371 Synthesis gas (syngas) 225, 233, 272–4, 277, 285–6, 289, 295, 298–304 Synthetic diesel fuel 268 Synthetic natural gas (SNG) 273, 308, 434 t Tafel 393 equation 46, 48–9, 417, 434 plots 47, 57, 434 Tank‐to‐wheels 387 Tantalum 205, 293 Tape calendaring 252 Tape‐casting 213 Target program 205 Tars 269 Teflon 150 Telecommunications 148 towers, remote 71, 123, 274, 389 Temperature 27–30, 33–7, 94–9, 152–5, 189–97, 201–4, 222–3, 241–5, 268–70, 278–82, 284–7, 342–3, 354–7, 401–4, 432–3 distribution 192, 217, 219 gradients 216, 418 standard conditions 28, 433 Terminals 30, 429, 435 Tetraazaannulene (TAA) 86, 205 Tetraethyl orthosilicate 294 Tetrafluoroethylene 72 Tetrahydrofuran 347 Tetrahydrothiophene (THT) 272, 276 Tetramethoxyphenylporphyrin (TMPP) 86, 205 Tetramethyl orthosilicate 294 Tetraphenylporphyrin (TPP) 86, 205 Thermal cracking, of hydrocarbons 289 Thermal cycling 221, 252, 257 Thermal decomposition 342, 417, 424 controlled 234 Thermal durability 277 Thermal efficiency 188, 287, 419, 435 Thermal expansion coefficient 237, 239, 248, 435 Thermal plasmas 289–90 Thermochemical methods, of hydrogen generation 264 Thermoneutral condition 286 Three‐phase boundary 4, 11, 83, 90, 244, 246, 435 Three‐phase inverters 376–7 Thyristor 367–8 Tile 211–12 Titanium 109, 163, 205, 248, 293, 335 Titanium carbide 132 Titanium dioxide 77, 103, 172, 313, 421 Titanium hydride 325 Titanium nitride 109 Tokyo Electric Power 204–5 Tokyo Gas Co., Ltd. 257, 302 Tolerance band 375 Toray paper 424 Tortuosity 76, 435 Transistors 58, 367, 370, 396 Transition metal 109, 205, 246, 293, 334 chalcogenide 167 macrocyclic compounds 86 nitrides 87 Transmission systems 271–2, 310 high‐voltage 378 Trifluorostyrene 77 Tube‐and‐shell configurations 294 Tube trailer 328 Tubular cells 237, 251–3, 255 flat 256 Turbine 25, 187–8, 274, 332, 351, 354, 357–8 axial 361 calculations 362 Index performance chart 413 wind 309–10, 378, 389 Turbochargers 357, 359, 361 Twin‐screw compressor 352–3, 360 u UltraBatteryTM 383 Ultracapacitor 421 Ultracell 113 UltraFuelCell 258 Unicellular 318–19 Uninterruptible power system (UPS) 125, 255, 389 Union Carbide 135, 139, 148 United Technologies Corporation (UTC) 6–7, 19, 140, 199, 205–6, 223, 225, 298, 397 Unitized regenerative fuel cell (URFC) 130, 133, 432, 435 Universal three‐phase inverter 377 Utilities 188, 272, 378, 426 Utilization 155, 194, 202–3, 220, 222, 258, 350, 435 UV region 313 UV water purifiers 123 v Vacancies 240, 313, 425 oxygen‐ion 239 Valency 241, 423, 435 Vanadium 22, 109, 205, 241 redox battery 21, 131, 182 Vapour 101, 165, 169–70 metal chloride 240 Vapourization 33, 210, 264, 409, 425, 427 Vapour phase 339, 409 Vapour pressure 37, 96–8, 100–1, 153–4, 176, 197, 264 Volmer, see Butler–Volmer equation Voltage 35–7, 39–41, 43–6, 50–6, 59–62, 65–7, 117–18, 129–30, 192, 366–8, 370–80, 417–18, 421–2, 428–30, 435 degradation rates 228 fuel‐cell 34, 44, 54 losses 30, 34, 45–6, 55–6, 67, 82, 119, 149–50, 160 open‐circuit cell 190–1 regulator 23, 366, 369, 371 reversible cell 160, 222 Voltammetry 420, 436 linear‐sweep 436 Voltammogram 61, 436 cyclic 61, 177, 178 Volt–amperes 379, 431, 435 reactive 379 Vulcan XC72, 82, 163 w Waste 154, 165, 265, 351, 417 agricultural 321 household 273 municipal 273 organic 320–1 Water‐cooled PAFC stacks 199–200 Water cooling 105, 123, 125, 128, 201, 390 Water evaporation 94, 165, 409 Water management 70, 72, 92, 129, 132, 142, 165, 307, 312 Water production 92, 407–9 rate of 101, 165, 405, 408–9 Water splitting 313, 318, 343 Water vapour 94–5, 97, 103, 109, 116, 164, 189, 202, 263, 343, 407, 425 pressure 96, 101 Wavelengths 313, 319, 425 Well‐to‐wheels 304, 385, 387, 389 Westinghouse 237, 254–6, 398 Wet‐proofing agent 210 Wet‐seal 215–18, 222 Wind 265, 306, 308, 310, 387, 419, 432 Wind farms 231, 380 Window‐frame design 253 Wood 103, 341, 417 pulp 178 wastes 417 x Xerogel 294, 433, 436 X‐ray 146, 436 X‐ray photoelectron spectroscopy 436 394, 459 460 Index y Yeasts 169, 423 Yttria 235, 239 stabilized zirconia 233, 238–41, 244–7, 255, 258, 302, 422 z Zeolites 293–5, 346, 436 faujasite‐type 277 separation membranes 155 Zetek–Zevco technology 139 Zinc 1, 20, 22–3, 31, 51, 147–8, 241 Zinc–bromine battery 182 Zinc oxide 276–7, 280 Zirconia 212, 216, 235, 239–40, 245, 317 calcia‐stabilized 239 cubic 240 tetragonal 240 Zirconium phosphates 77, 142