Future lithium-ion batteries 978-1-78801-612-4, 1788016122, 978-1-78801-760-2, 1788017609, 978-1-78801-418-2

Citation preview

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP001

Future Lithium-ion Batteries

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP001

View Online

     

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP001

Future Lithium-ion Batteries

Edited by

Ali Eftekhari

Email: [email protected]

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP001

View Online

Print ISBN: 978-1-78801-418-2 PDF ISBN: 978-1-78801-612-4 EPUB ISBN: 978-1-78801-760-2 A catalogue record for this book is available from the British Library © The Royal Society of Chemistry 2019 All rights reserved Apart from fair dealing for the purposes of research for non-commercial purposes or for private study, criticism or review, as permitted under the Copyright, Designs and Patents Act 1988 and the Copyright and Related Rights Regulations 2003, this publication may not be reproduced, stored or transmitted, in any form or by any means, without the prior permission in writing of The Royal Society of Chemistry or the copyright owner, or in the case of reproduction in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK, or in accordance with the terms of the licences issued by the appropriate Reproduction Rights Organization outside the UK. Enquiries concerning reproduction outside the terms stated here should be sent to The Royal Society of Chemistry at the address printed on this page. Whilst this material has been produced with all due care, The Royal Society of Chemistry cannot be held responsible or liable for its accuracy and completeness, nor for any consequences arising from any errors or the use of the information contained in this publication. The publication of advertisements does not constitute any endorsement by The Royal Society of Chemistry or Authors of any products advertised. The views and opinions advanced by contributors do not necessarily reflect those of The Royal Society of Chemistry which shall not be liable for any resulting loss or damage arising as a result of reliance upon this material. The Royal Society of Chemistry is a charity, registered in England and Wales, Number 207890, and a company incorporated in England by Royal Charter (Registered No. RC000524), registered office: Burlington House, Piccadilly, London W1J 0BA, UK, Telephone: +44 (0) 20 7437 8656. Visit our website at www.rsc.org/books Printed in the United Kingdom by CPI Group (UK) Ltd, Croydon, CR0 4YY, UK

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP005

Preface The history of lithium-ion batteries is straightforward, as the pioneering works of Stanley Whittingham paved the path for the birth of intercalation chemistry. Soon after, LiCoO2 and graphite were introduced as potential materials for the cathode and the anode, respectively. This cell design was commercialised within ten years and remained the dominant architecture for decades. Numerous efforts have improved the cell technology, but the materials remained the same, although two alternative cathode materials, LiNi1/3Mn1/3Co1/3O2 and LiFePO4 were introduced in the 1990s. Economies of scale have reduced the price of lithium-ion batteries to minimize the cost of cell technology, and the current price is limited by the cost of the materials. Over the past two or three decades, the prime application of lithium-ion batteries was portable electronic devices such as mobile phones and laptops, but there is a big change ahead. Electric vehicles, which were considered as a fancy technology in the 1990s and even early 2000s, are now among the fastest growing technologies. This development depends on the rapid advancement of large-scale batteries. Many carmakers such as Tesla Motors use the same cell design (18 650 cell) that is utilised in laptops. In any case, the research area of lithium-ion batteries is now experiencing an unusual period, not only because of a sudden growth in the demand but also an unprecedented change in the application. As a result, the research strategy should be subtly planned to meet the current requirements rather than relying on the old-fashioned strategy of research for a gradual advancement of an emerging area. In a series of papers, I criticised the current strategy of research due to the lack of practical outlook on the current demand. For instance, I highlighted that energy efficiency is a critical factor in the investigation of new electrode materials, but it is ignored because it was not an issue for the classic intercalation materials (Sustainable Energy Fuels, 2017, 1, 2053). While the current   Future Lithium-ion Batteries Edited by Ali Eftekhari © The Royal Society of Chemistry 2019 Published by the Royal Society of Chemistry, www.rsc.org

v

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP005

vi

Preface

focus on materials research is to improve the specific capacity, which is a critical requirement for small-scale lithium-ion batteries, I stressed that the power density might be of imperative importance to the emerging market of electric vehicles to facilitate the transition from long-standing petrol-based vehicles (ACS Sustainable Chem. Eng., 2017, 5, 2799). In the same direction, I thought it might be beneficial to gather various opinions of eminent scientists about the future direction of lithium-ion batteries. Of course, the authors do not necessarily share my opinions and have their own stances. In other words, the current book is indeed a forum of different perspectives about the future of lithium-ion batteries. The book structure is rather simple, as the beginning chapters focus on the electrodes materials, followed by chapters on various electrodes, and finally discussing critical issues of cell design. Junjie Niu in the first chapter discusses new possibilities for anode materials, as we have already reached the theoretical capacity of graphite anode. The subsequent chapters by Seung-Taek Myung and Ulla Lassi explain the shift from the classic LiCoO2 towards the alternative structure of layered metal oxides. In the late 1970s, one of the key issues of lithium-ion batteries was the formation of a solid electrolyte interphase to avoid the formation of dendrites on the lithium anode (when the graphite anode was not yet introduced). Michael Armand introduced polymer electrolytes, which were not embraced for commercial development until recently. In Chapter 4, he unfolds the current situation and the road ahead of solid electrolytes leading us to safer batteries. Dong-Won Kim then talks about the flexibility of gel polymer electrolytes. In Chapter 6, Weishan Li explains the current status of electrode/ electrolyte structures for modifying non-aqueous electrolytes of high-voltage performance. Hideaki Horie sheds light on the emerging possibilities of polymer materials in novel cell designs. Although Li as the lightest metal is an excellent charge carrier, the opportunities of exploiting other alkali metals such as Na and K are of particular interest, as thoroughly discussed by Shinichi Komaba in Chapter 8. In addition to the cell design, a vital matter in developing lithium-ion batteries is to closely monitor their status during usage. This is why Peter Notten discusses the aging mechanism of lithium-ion batteries. In addition to the example of electric vehicle application of large-scale lithium-ion batteries, household energy storage is of particular importance to better utilise the electricity generated by solar panels, as described by Matthias Vetter in Chapter 10. Despite considerable advancement in cell technology over the past two decades, manufacturing is still a critical matter in the production of lithium-ion batteries, as elucidated by Emma Kendrick. Since the early days of lithium-ion batteries, safety has been a major concern, and catastrophic incidents have been common in commercial batteries due to the possible explosion of inflammable materials therein. In Chapter 12, William Q. Walker

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP005

Preface

vii

elaborates these concerns and the approaches we use to tackle the key safety issues in the development of lithium-ion batteries. As stressed above, the rapidly growing demand made the supply an imperative matter, as is explicated by Wolfgang Bernhart in Chapter 13. Along the same lines as Chapters 10 and 13, the last chapter discusses the potential market of large-scale lithium-ion batteries for energy storage to make household power generation less dependent on the grid. Overall, the book attempts to provide a concise outlook on various aspects of lithium-ion batteries. The authors have been deliberately chosen from different disciplines and sectors to build a more comprehensive perspective by looking at the issues from different angles. At the early stages of materialising the basic idea of this book, Leanne Marle was very helpful in shaping the structure. However, I must especially thank Connor Sheppard whose contribution was beyond his official responsibility. I confess that this book project could not have been completed without his efforts. Ali Eftekhari

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP005

View Online

      

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP009

Contents Chapter 1  New High-energy Anode Materials  Junjie Niu and Shuai Kang



1.1 Introduction  1.2 Carbide/Nitride Materials  1.2.1 Carbon Materials  1.2.2 Carbides and Nitrides: MXenes  1.3 Silicon, Metals, Metal Oxides and Metal Sulfides  1.3.1 Challenges of Volume Expansion and Solutions  1.3.2 Silicon Materials  1.3.3 Metals, Metal Oxides and Metal Sulfides  1.4 Li Metal  1.4.1 The Advantages of Li Metal as Anode  1.4.2 The Challenges of Li Metal as Anode  1.4.3 Solid-state Electrolyte and Solid-electrolyte Interphases for a Li Metal Anode  1.4.4 Manufacturing Techniques for Li Metal  1.5 Conclusions and Outlook  Acknowledgement  References 

Chapter 2 Layered Ni-rich Cathode Materials  Seung-Taek Myung, Chang-Heum Jo and Aishuak Konarov

2.1 Introduction  2.1.1 Cathode Material for the Next Generation Lithium-ion Battery 

  Future Lithium-ion Batteries Edited by Ali Eftekhari © The Royal Society of Chemistry 2019 Published by the Royal Society of Chemistry, www.rsc.org

ix

1 1 2 2 5 7 7 9 10 13 13 14 14 17 17 18 18 26 26 26

View Online

Contents

x



Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP009







2.1.2 Ni-rich Cathode Material  2.1.3 Solutions for Incomplete Ni-rich Cathode Material  2.1.4 Perspective for Next Generation Ni-rich Cathodes Materials  Acknowledgements  References 

39 40 40 44

3.1 Introduction  3.2 Common Layer-structured Materials  3.2.1 LCO  3.2.2 NCM  3.2.3 NCA  3.2.4 LLO and Blended Cathodes  3.3 Modification Methods  3.3.1 Coating  3.3.2 Doping  3.4 Conclusions  References 

44 46 46 48 49 50 51 53 60 66 68







29

Chapter 3 Modification of Layered Oxide Cathode Materials  J. Dong, M. Hietaniemi, J. Välikangas, T. Hu and U. Lassi

Chapter 4 Solid Electrolytes for Lithium Metal and Future Lithium-ion Batteries  Gebrekidan Gebresilassie Eshetu, Xabier Judez, Chunmei Li, Maria Martinez-Ibañez, Eduardo Sánchez-Diez, Lide M. Rodriguez-Martinez, Heng Zhang and Michel Armand



28



4.1 Introduction  4.2 Evolution and Recent Advancements in Solid Electrolytes  4.2.1 Ionic Transport in Solid Electrolytes  4.2.2 Solid Polymer Electrolytes  4.2.3 PEO-based SPEs  4.2.4 Inorganic Electrolytes  4.3 Composite/Hybrid Electrolyte  4.4 Electrode/Electrolyte Interfacial Chemistry and Compatibility  4.5 Conclusions and Outlook  References 

72

72 73 73 76 77 83 86 90 91 93

Chapter 5 Gel Polymer Electrolytes  Dong-Won Kim

102

5.1 Introduction  5.2 Types of Electrolytes 

102 103

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP009

Contents

xi





5.3 Components of Gel Polymer Electrolytes  5.3.1 Lithium Salt  5.3.2 Organic Solvent  5.3.3 Polymer  5.3.4 Inorganic Filler  5.4 Characteristics and Requirements of Gel Polymer Electrolytes  5.4.1 Ionic Conductivity  5.4.2 Electrochemical Stability  5.4.3 Lithium Transference Number  5.4.4 Mechanical Properties  5.5 Preparation of Gel Polymer Electrolytes  5.5.1 Solution Casting  5.5.2 Hot Melting  5.5.3 Immersion of Porous Membrane  5.5.4 In situ Cross-linking  5.6 Concluding Remarks  References 

Chapter 6 Liquid Non-aqueous Electrolytes for High Voltage Lithium Ion Batteries  Lidan Xing and Weishan Li



6.1 Introduction  6.2 Electrolyte Component of the High Voltage Electrode/Electrolyte Interphase  6.3 Critical Role of the Anion in the Interphasal Stability of a High Voltage Cathode/Electrolyte  6.4 Electrolytes for High Voltage Cathode Materials  6.4.1 Salt/Solvent Substitution  6.4.2 Film-forming Electrolyte Additives  6.5 Summary  Acknowledgements  References 

Chapter 7 Creation of a New Design Concept for All-polymer-structured Batteries  Hideaki Horie

7.1 Introduction  7.2 Improvement of Safety Is an Urgent Issue  7.3 Issues in the Current Manufacturing Method of Li Batteries  7.3.1 The Drying Process in Manufacturing Reduces Productivity  7.3.2 Metallic Current Collector Required for Press Process 

104 105 106 107 116 117 117 118 118 119 120 120 121 122 122 123 125 130 130 131 133 138 138 144 153 154 154 163 163 164 165 165 167

View Online

Contents

xii

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP009





7.4 Re-inventing the Structure of the Battery  7.4.1 Designing the Battery from the Direction of Current Flow  7.4.2 Advantages of the Bipolar Structure  7.4.3 Flexibility to Withstand Stress  7.5 The New Proposal of a Resin Base  7.5.1 Combining Resins with Various Functions  7.5.2 New Design Method for Active Material, Electrodes, and Cell  7.5.3 Surface Design to Ensure Electron Conductivity  7.5.4 Significant Reduction in Size and Weight with Resin  7.6 Implementation of a Simple Cell Manufacturing Process  7.7 Possibility of Radical Measures to Counter Fire Accidents  7.7.1 Resin Suppresses Expansion of Abuse Mode  7.7.2 Tests that Withstood Drilling  Acknowledgement  References 

Chapter 8 From Lithium to Sodium and Potassium Batteries  A. Shahul Hameed, Kei Kubota and Shinichi Komaba



8.1 Introduction  8.2 Material Resources and Battery Size of LIBs  8.3 Alternative Rechargeable Batteries  8.3.1 Sodium–Sulfur (Na–S) Batteries  8.3.2 Redox-flow Batteries  8.3.3 Mg Batteries  8.3.4 Na-ion Batteries  8.3.5 K-ion Batteries  8.4 Materials for Na-ion Batteries  8.4.1 Positive Electrode Materials  8.4.2 Negative Electrode Materials  8.5 Materials for Potassium-ion Batteries  8.5.1 Positive Electrode Materials  8.5.2 Negative Electrode Materials  8.6 Outlook  Acknowledgement  References 

168 168 169 170 171 172 172 174 174 175 177 178 178 179 179 181

181 182 184 185 185 187 187 188 190 190 197 201 201 204 207 210 210

View Online

Contents

xiii

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP009

Chapter 9 Understanding Battery Aging Mechanisms  Dongjiang Li, Dmitri L. Danilov, Henk Jan Bergveld, Rüdiger-A. Eichel and Peter H. L. Notten

9.1 Introduction  9.2 Working Principles and Cell Design of Li-ion Batteries  9.2.1 Working Principles of Li-ion Batteries  9.2.2 Design of Li-ion Batteries  9.3 Degradation Mechanisms of Li-ion Batteries  9.3.1 SEI Formation Inducing Cyclable Li Losses  9.3.2 Cathode Degradation  9.3.3 Anode Electrode Degradation  9.4 Electromotive Force (EMF) Determination  9.5 Calendar Aging  9.6 Cycling-induced Aging  9.6.1 Irreversible Capacity Loss (ΔQir)  9.6.2 Postmortem Analyses  9.6.3 Non-destructive Approach  9.6.4 Summary  9.7 Modeling Degradation Mechanisms of Li-ion Batteries  9.8 Future Outlook  References 



220

220 221 221 224 225 225 226 228 229 230 233 233 236 238 240 241 246 247

Chapter 10 Battery Storage for Grid Connected PV Applications  M. Vetter, S. Lux and J. Wüllner

251



251





10.1 Introduction  10.2 Residential Behind-the-meter PV Battery Storage Systems  10.2.1 DC Coupled PV Battery System  10.2.2 AC Coupled PV Battery System  10.2.3 Generator Coupled PV Battery System  10.2.4 Energy Management  10.3 Commercial Behind-the-meter PV Battery Storage Systems  10.3.1 Load Balancing  10.3.2 Increased Resilience  10.4 District Battery Storage Systems in Combination with PV  10.4.1 Sector Coupling  10.4.2 Grid Services  10.5 Key Factors Affecting Bankability of PV Battery Storage Projects  10.6 Conclusion  References 

253 253 254 254 255 256 256 257 257 257 258 258 260 261

View Online

Contents

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP009

xiv

Chapter 11 Advancements in Manufacturing  Emma Kendrick

262



262 264 264





11.1 Introduction  11.2 Electrode Manufacturing Processes  11.2.1 Materials  11.2.2 Preparation of Inks and Mixing Methods  11.2.3 Coating Techniques  11.2.4 Electrode Drying  11.3 Cell Manufacturing  11.3.1 Cell Balance  11.3.2 Electrolyte Filling, Formation and Conditioning  11.4 Summary and Conclusions  References 

266 274 278 281 281 282 285 286

Chapter 12 Lithium-ion Battery Safety  William Q. Walker, Omar A. Ali and Dwight H. Theriot

290



290





12.1 Introduction to Lithium-ion Battery Safety  12.1.1 Cell Level and System Level Factors that Influence Thermal Runaway Behavior  12.1.2 Characteristics of Thermal Runaway  12.1.3 Motivation to Focus on Safety  12.2 High Profile Thermal Runaway Field Failures  12.3 Primary Thermal Runaway Testing Techniques  12.3.1 Copper Slug Calorimetry (CSC)  12.3.2 Accelerating Rate Calorimetry (ARC)  12.3.3 Bomb Calorimetry  12.3.4 Fractional Calorimetry  12.3.5 Cone Calorimetry  12.3.6 Trigger Techniques  12.3.7 System Level Testing  12.4 Thermal Runaway Modeling Techniques  12.4.1 Thermal Runaway Kinetic Relationships  12.4.2 Benefits of 2D and 3D Modeling  12.5 Advanced Thermal Management Systems  12.5.1 Interstitial Heat Sinks  12.5.2 Interstitial Insulating Materials  12.5.3 Flame Arresting Devices and Shielding  12.5.4 Active Thermal Management  12.5.5 Phase Change Materials  12.6 Future Trends in Lithium-ion Battery Safety  References 

291 294 296 297 299 300 300 301 301 302 302 303 303 304 307 308 308 309 309 310 310 310 311

View Online

Contents

xv

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP009

Chapter 13 Challenges and Opportunities in Lithium-ion Battery Supply  Wolfgang Bernhart



13.1 Introduction: Lithium-ion Cell Market and Players  13.2 The Lithium-ion Value Chain  13.3 Technology and Cost Development  13.3.1 Lithium-ion Technology and Cost Roadmap  13.3.2 Prospects of “Post-lithium” (PLiT) and Solid-state Technologies for the Automotive Market  13.4 Demand and Supply of Pre-cursor and Raw Materials—Mining, Refining, Recycling  13.5 Implications and Conclusions  References 

316 316 317 319 319 322 326 333 333

Chapter 14 Emerging Market of Household Batteries  D. Parra

335



335 338





14.1 Why Household Batteries?  14.2 A Household Battery System  14.3 Electricity Prices and Battery Applications for Consumers  14.3.1 PV Self-consumption  14.3.2 Avoidance of Renewable Energy Curtailment  14.3.3 Demand Load-shifting  14.3.4 Demand Peak-shaving  14.3.5 Back-up Power  14.3.6 Combination of Applications  14.4 Different Lithium-ion Battery Technologies  14.5 The Community Scale  14.6 Global Impact of Household Batteries  14.6.1 Energy System Cost  14.6.2 Environmental Impacts  14.6.3 Security of Energy Supply  14.6.4 Wholesale Electricity Market  14.7 Outlook  Acknowledgement  References 

Subject Index 

339 342 344 345 346 348 348 349 351 353 353 354 355 355 356 358 358 361

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-FP009

View Online

      

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00001

Chapter 1

New High-energy Anode Materials Junjie Niu*a and Shuai Kangb a

University of Wisconsin-Milwaukee, Department of Materials Science and Engineering, Milwaukee, Wisconsin, 53211, USA; bChinese Academy of Sciences, Chongqing Institute of Green and Intelligent Technology, Intelligent Manufacturing Technology Institute, Chongqing, 400714, PR China *E-mail: [email protected]

1.1  Introduction The lithium-ion battery (LIB) is one of the most promising batteries that can meet the rapidly growing energy requirement in the next decade. In order to be competitive with fossil fuels, high-energy rechargeable batteries are perhaps the most important enabler in restoring renewable energy such as ubiquitous solar and wind power and supplying energy for electric vehicles.1,2 The current LIBs using graphite as the anode electrode coupled with metal oxide as the cathode electrode show a low-energy density of ∼150 Wh kg−1 (∼250 Wh L−1) and a high cost of $200–300 kW h−1.3–7 Therefore, it is required to explore new electrode materials that can display efficient and stable reduction/oxidation and possess higher energy density, higher power density and a longer lifetime than those used in state-of-the-art batteries.7,8 Different battery chemistries and lithium storage mechanisms of new electrode materials need to be fully understood, not only for offering great improvements in energy density, power density and stability,9 but also for optimizing the safety of batteries when subjected to abuse conditions.10   Future Lithium-ion Batteries Edited by Ali Eftekhari © The Royal Society of Chemistry 2019 Published by the Royal Society of Chemistry, www.rsc.org

1

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00001

2

Chapter 1

The conventional anode material is graphite, which has a specific capacity of 372 mAh g−1. Several new anode materials with much higher theoretical capacity have been reported, including different carbon materials, silicon, metal and metal oxides. Two major challenges exist in these new anode materials: large volume expansion and slow electron/ion transport. Various nanostructured configurations have been fabricated to address these challenges. In this chapter, we will review the current studies on anode materials including: (1) carbide and nitride materials, (2) silicon, metal oxides and metal sulfides, and (3) Li metal. We will discuss some challenges such as volume change in silicon-based materials (which is well known for its highest theoretical specific capacity of 4200 mAh g−1 to date), metals, metal oxides and metal sulfides, the unstable solid-electrolyte interphase (SEI) layers, the crack and fraction of the electrodes and the nanotechnological solutions. Finally, we give a brief conclusion and outlook for anode materials in LIBs.

1.2  Carbide/Nitride Materials Carbon is the most versatile element in the periodic table due to the nature of chemical bonds between carbon atoms or with other elements, and the variety of structures, textures and particle shapes. Hundreds of carbon materials and derivatives have been examined as anode materials of LIBs in the past few years. Here we review the anode electrode studies on carbon materials including graphite, carbon nanotubes (CNTs) and graphene, and carbon composites such as MXenes. The lithium storage mechanism, the existing problems and the corresponding solutions will also be discussed.

1.2.1  Carbon Materials Natural graphite has been categorized as a critical strategic material in the US and Europe.11 Even though graphite and its derivatives can be synthesized, a higher cost of about $13 rather than $8 for natural graphite (in 2016) is needed. The Li-ion storage mechanism of graphite is based on the intercalation that the Li-ions insert/extract the planes of graphite. Li-ions can combine on every second carbon hexagon in the graphite sheet, which limits the number of Li-ions to one for every six carbon atoms (Figure 1.1),12 leading to a theoretical capacity of 372 mAh g−1. The spacing between graphite layers before and after Li-ion intercalation is 0.335 nm and 0.35 nm,12,13 respectively, with an expansion of less than 10%. The intercalation process is highly reversible, which contributes to a long cycling performance of over 1000 cycles. However, typically the rate capability is not on a solid footing, for example, it takes up to eight hours to recharge electric vehicles, whereas it takes less than five minutes to refill a traditional internal combustion engine with gasoline. When the charging rate increases, the lifetime of graphite is significantly reduced14 due to the thickening of the electrode film,15 blocking

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00001

New High-energy Anode Materials

3

Figure 1.1  Li-ion  storage mechanisms of graphite. (a) Schematic drawing of the

crystal structure of hexagonal graphite showing the AB layer stacking sequence and the unit cell. (b) Structure of Li-ion intercalated graphite compound LiC6. Reprinted with permission from K. R. Kganyago and P. E. Ngoepe, Phys. Rev. B, 68, 205111, 2003, Copyright 2003 by the American Physical Society. http://dx.doi.org/10.1103/PhysRevB.68.205111.12

electrode pores,16 and lithium plating on the surface. In general the lithium plating, particularly at low temperatures, is caused by the slow lithium-ion diffusion and polarization in graphite.17 To date graphite with an out-ofplane aligned architecture using a low external magnetic field has been fabricated and showed an enhanced electrochemical performance with a high loading of about 10 mg cm−2 and 200 µm thickness.18 A specific charge capacity up to three times higher than that of non-architecture electrodes at a rate of 1 C was obtained due to the low tortuosity resulting from the simple and scalable magnetic alignment approach. CNTs are one-dimensional cylindrical tubules of graphite sheet with high conductivity of 106 S m−1 (single walled CNTs),19 low density, high rigidity20,21 and high tensile strength up to 60 GPa.22 CNTs are used as alternative anode materials where the insertion level of Li-ions can be increased from LiC6 in close-end single walled nanotubes (SWNTs) to LiC3 after etching.23 The reversible capacity of CNTs varies due to the different structure including diameters, number of layers, length and defects.24 Li-ions cannot get through the defect-free hexagonal ring because of the high energy-barrier, whereas they can diffuse into the CNT and are captured inside if there are large defect rings (Figure 1.2). In this case the low energy barrier from the defects contributes to the accumulation of Li-ions during the charging process. Once inserted, the Li-ions undergo a one-dimensional transport inside the tube. If the tube is too long, the diffusion is affected by a hard de-insertion.25 The major issues of CNTs as anode materials are the large irreversible capacity, which is caused by large surface area, and the lack of voltage plateau during discharging. However, anode electrodes with CNTs normally show a higher rate performance (from the high conductivity), better bendability and self-supporting ability (from the good mechanical properties).26–28

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00001

4

Chapter 1

Figure 1.2  Effect  of defects on Li insertion in (5,5) single walled carbon nanotubes.

Red balls indicate the defects, green balls indicate the initial position of the Li, and yellow balls indicate the trajectory of Li. Note: for each nanotube, both the side view and the top view are presented. Reprinted with permission from K. Nishidate and M. Hasegawa, Phys. Rev. B, 71, 245418, 2005, Copyright 2005 by the American Physical Society. http:// dx.doi.org/10.1103/PhysRevB.71.245418.24

Graphene is two-dimensional single or multiple layers of graphite sheets with a low electrical resistance of 10−6 Ω cm, a high Young's modulus of 1 T Pa and a high heat conductivity coefficient of 5300 W m−1 K−1,29–34 which give graphene the much higher reversible capacity of >500 mAh g−1 35–37 and a faster rate performance than graphite.38 Reports show that carbon atoms on the edges of graphene are considered as very active sites for Li-ion storage, providing much stronger binding energies for Li-ions (1.70 to 2.27 eV) with respect to the graphene basal plane (1.55 eV).39 In Figure 1.3, the energy barriers toward the edges of graphene are markedly lower than those for the channels along the ribbon axis and the diffusion coefficient can be up to two orders of magnitude larger than the inside graphene, thus, those Li atoms are able to overcome the first barrier, then easily diffuse toward the edges.39 Similar to CNTs, graphene anodes also displays a large irreversible capacity with a lack of voltage plateau. More researchers focus on graphene composites, which are reported to positively affect the Li-ion storage performance.40 In addition, the flexible and self-supporting configurations can be achieved by forming the composites with graphene.41,42

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00001

New High-energy Anode Materials

5

Figure 1.3  Comparisons  of energy barriers for path B in graphene and the zig-

zag graphene nanoribbon. Top left: schematic representation of the graphene cell used in the calculations and migration path B in graphene. H, M, and T represent the Li adsorption positions on top of the hexagon, on the middle of the bond, and on top of a C atom, respectively. Bottom left: schematic representation of the zigzag graphene nanoribbon cell and migration path B toward the edge. In all cases, TV represents the translational vector. Reprinted with permission from C. Uthaisar and V. Barone, Nano Lett., 2010, 10, 2838. Copyright 2010 American Chemical Society.39

1.2.2  Carbides and Nitrides: MXenes Transition metal carbides, carbonitrides and nitrides (MXenes) are produced by selectively etching the “A” metal from MAX phases (Figure 1.4a), in which M is an early transition metal, A represents a group A element, and X represents C and/or N.43–45 Several computational studies have investigated the effect of M, X and T, the number of M layers, and the lattice strain on the electronics, thermal and mechanical properties of MXenes.46–49 MXene shows promise in increasing Li-ion storage performance with a theoretical capacity of 447.8 mAh g−1 on Ti3C2.50 In terms of theoretical gravimetric capacity, MXenes with low molecular weights, such as Ti2C, Nb2C, V2C and Sc2C, are the most promising.51 Similar to graphite, Li-ions penetrate only between the MXene sheets (Figure 1.4b).52 The bonds between M and X are too strong to be broken easily,53 which can also explain why Ti2C has ∼50% higher gravimetric capacitance than Ti3C2 (with one inactive TiC layer).54,55 Among MXenes, V2CTx shows the highest Li-ion capacity of 280 mAh g−1 at 1 C and 125 mAh g−1 at 10 C.44 The surface terminations are one factor that can affect the Li-ion storage, for example, oxygen terminations are considered most favourable, whereas hydroxyls and fluorines result in lower capacity due to the impeded effect of Li-ion transport.52,55,56 There is a continuous change in

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00001

6

Chapter 1

Figure 1.4  (a)  Schematic of the exfoliation process for Ti3AlC2 to Ti3C2. Reproduced

from ref. 45 with permission from John Wiley and Sons, Copyright © 2011 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim.45 (b) Schematic illustration of the Ti3C2Tx lithiation process. The valence electron localization functions are shown with and without the additional lithium layer. Reprinted with permission from Y. Xie, M. Naguib, V. N. Mochalin, M. W. Barsoum, Y. Gogotsi, X. Yu, K.-W. Nam, X.-Q. Yang, A. I. Kolesnikov and P. R. C. Kent, J. Am. Chem. Soc., 2014, 136, 6385. Copyright 2014 American Chemical Society.52

the transition metal (here it is titanium) oxidation state during charge and discharge up to 0.5 V versus Li/Li+ using in situ X-ray absorption spectroscopy (XAS).52 Owing to the two-dimensional nature and high conductivity of MXenes, Li-atoms can reversibly form an additional layer during insertion/ extraction without a further translation into a change oxidation state.55,57,58 Theoretical studies show low diffusion barriers for Li-ion,36 which is in agreement with the remarkable high-rate performance observed in MXenes.44,54 Except for a relatively low specific capacity, MXenes display a non-plateau region in their charge–discharge profiles, similar to the behaviour of supercapacitors.53 However, MXene-based composites hold particular promise for high-performance, high-rate Li-ion batteries with dramatically increased

View Online

New High-energy Anode Materials

7

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00001

cyclability and stability, where MXenes provide a conductive matrix that accommodates expansions and contractions of particles while maintaining structural and electrical connectivity.59–61

1.3  Silicon, Metals, Metal Oxides and Metal Sulfides In order to further increase the specific capacity, materials with nanostructures that can deliver a much higher theoretical capacity are designed. The new high-capacity electrode materials usually have large volume changes due to the intake of large amounts of Li-ions. The volume expansions of alloytype anodes are as high as 420% for Si, 260% for Ge and Sn during the Li-ion insertion/extraction process, all much greater than the 10% for traditional graphite anodes and MXenes. These drastic volume changes induce mechanical degradation of both active materials and electrodes during electrochemical cycling, significantly shortening the battery cycling life.

1.3.1  Challenges of Volume Expansion and Solutions Volume expansion in anode materials such as silicon has been observed with a high density of dislocations emerging from the reaction front using in situ TEM technology (Figure 1.5a).62 During the Li-ion insertion, the crystalline structure was changed to amorphous, and the nanowire diameter was increased. Generally, three major issues exist in these anode materials, as illustrated in Figure 1.5b. First, cracking and fracture of particles occurs upon lithiation/delithiation, which is mainly caused by the large stresses inside the particles, especially at fast charging/discharging.63,64 Nanomaterials with particle sizes below the critical size such as nanopillars, nanoparticles, nanowires, nanotubes, nanorods and nanocomposites can mitigate these phenomena due to size confinement and strain accommodation.65–69 Second, a stable SEI layer that can ensure high Coulombic efficiency and long-term stability is needed.70 During charging, a passivating SEI layer forms on the anode electrode surface due to electrochemical reduction of the electrolyte. The reduction potential of organic carbonates (around ∼1 V versus Li+/Li) from the electrolyte71 is typically higher than the electrochemical working potential of anode materials. During the cycling, the volume change of the electrode makes the SEI layer form and disappear frequently. The nanostructures and secondary clusters are designed to accommodate the volume expansion, and thus to stabilize the SEI layer as well. Third, the electrode fracture is another issue, which is due to the accumulation effect of numerous particles and additives in the electrodes such as carbon black and polyvinylidene fluoride binder. Because of the high surface energy, nanoparticles tend to aggregate strongly together to form secondary larger sized particles. The expansion/shrinkage of the electrode over discharge/charge cycles eventually leads to mechanical failure and rapid capacity fading due to the electrical disconnection and physical delamination of active materials.

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00001

View Online

Figure 1.5  (a)  Ultrafast charging of carbon-coated and phosphorus-doped sili-

con nanowires. Left: morphology evolution in 72 s after the bias was applied. The spiral shape was observed only at very high lithiation rates (reaction front marked by red arrows). The average speed for the reaction front was 213 nm s−1. Right: high magnification time-lapse TEM images showing an immediate expansion after lithiation and the core– shell lithiation behaviour. Reprinted with permission from X. H. Liu, L. Q. Zhang, L. Zhong, Y. Liu, H. Zheng, J. W. Wang, J.-H. Cho, S. A. Dayeh, S. T. Picraux, J. P. Sullivan, S. X. Mao, Z. Z. Ye and J. Y. Huang, Nano Lett., 2011, 11, 2251. Copyright 2011 American Chemical Society.62 (b) Schematic of Si electrode failure mechanisms, including material pulverization, continuous SEI growth and detachment from current collectors. Reprinted from Nano Today, 7, H. Wu and Y. Cui, Designing nanostructured Si anodes for high energy lithium ion batteries, 414, Copyright 2012, with permission from Elsevier.76

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00001

New High-energy Anode Materials

9

The most straightforward solution is to use robust polymeric binders, for example, an alginate binder,72 a cross-linked polymeric binder,73 and a catechol-conjugated adhesive binder,74 instead of the traditional polyvinylidene fluoride binder, to mitigate detachment of active materials. In addition, active–inactive nanocomposites such as Si/metal nitride and Si/metal carbide have been demonstrated with positive effects on large-volume-change electrode materials.75 Nanomaterials bring essential advantages in terms of their short lithium-ion diffusion path compared with microparticles, therefore the majority of highly performing volume expansion anode materials are synthesized with nanosizes. A short and fast transport for Li+/e− is necessary to achieve a better rate capability. To improve the electron conductivity for individual particles, coating active particles with a conductive layer and embedding active particles into a conductive matrix are usually used.77–80 Active materials composited with 3D conductive networks such as carbon nanotube sponge-based 3D electrodes81 and graphene foam-based 3D electrodes82,83 have been successfully reported. Additionally, well-defined one-dimensional nanowires or nanofibres and two-dimensional nanoflakes can also facilitate efficient electron transport. In parallel, the electronic and ionic conductivity at the electrode level is also crucial, particularly for the high mass loading of active materials. Nanoparticles possess an increased porosity compared with bulk materials, which normally affects the overall conductivity of electrodes as due to large interparticle resistance. This brings the problems of a low tap density and volumetric mass loading, which are important for high energy-density batteries. To address this issue, electrodes without a binder are constructed, such as active materials growing on metal current collectors,84 self-supported nanowire arrays,85,86 interconnected hollow nanospheres87 and inverse opal nanostructures.88 A secondary cluster design can also increase the tap density of nanoparticles.89,90 However, to reduce the high cost of nanomaterials, which is caused by their complex synthesis procedures or expensive raw materials, it is critical to explore facile routes for large-scale synthesis. Decreasing the particle size to the nanometre scale creates a lot of interparticle space, which usually leads to a low initial Coulombic efficiency owing to their inherent feature of high surface area that needs more electrolyte decomposition and lithium consumption. A hybrid configuration with a secondary structure may create a controllable electrode/electrolyte surface area. With an electrolyte-blocking layer, the specific SEI area can be reduced significantly.89

1.3.2  Silicon Materials The highest theoretical capacity of Si is 4200 mAh g−1 (Li4.4Si), which makes Si one of the most promising anodes for next-generation LIBs. Si is environmentally friendly and the second richest element in earth.76 However, Si/SiOx anode undergoes drastic structural transformation and large

View Online

Chapter 1

10

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00001

9,91,92

volume change during their charge and discharge process. Meanwhile, ultra-fine nanoparticles can be embedded in binary lithium compounds including Li2O, Li2S, LiF, Li3P or Li3N matrix.93–95 Cracking, fracture and pulverization of active particles occur due to the inherently large volumetric expansion/contraction, which further lead to the loss of electrical contact, as shown in Figure 1.5b.96 Studies show that there is a critical size for fracture generation.91,97 The mechanical robustness of small Si nanoparticles is from the insufficient stored strain energy from electrochemical reactions to drive crack propagation.98 The critical fracture size of crystalline Si particles is about 150 nm,98 which is smaller than crystalline Si nanopillars (240–360 nm) and nanowires (300–400 nm), depending on the electrochemical reaction rate.91 At the electrode level, significant mechanical degradation also occurs through displacement over layers of particles across the electrode.92 The particles detach from the surrounding electrical connections and delaminate from the electrode, thus leading to rapid capacity decay and battery failure, as shown in Figure 1.5b. The SEI of the Si/SiOx anode breaks and forms frequently due to large volume changes,99,100 which will result in: (1) electrolyte and lithium salt being consumed, and thus has a low Coulombic efficiency.101 (2) A thick SEI layer leads to a long lithium-ion diffusion distance and a sluggish electrochemical process. (3) The mechanical stress generated in a thick SEI layer causes material degradation. Currently only a small mass fraction of Si or SiOx (4.5 V) that it limits actual use because current electrolytes cannot handle such high voltages.23 Structural modification is needed to make LLOs commercially viable. Blended cathodes are different from composite cathodes. They are a physical mixture of different layered cathode material powders, or a mixture of layered material and Mn-rich spinel made in to a single cathode foil. In blended cathodes, each material is in distinctly different particles, so the outward behaviour of the cathode is a sum of all the materials. There is some evidence of structural stabilization in blended cathodes. Lithium-ion batteries (LIBs) with blended cathodes are commercially available.24,25

3.3  Modification Methods Effects that degrade battery performance during cycling can be roughly divided into two categories, external ones related to unwanted reactions with electrolyte, and internal ones that are due to structural changes during the Li-ion intercalation. (In truth this is more complicated, for example metal

View Online

Chapter 3

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

52

dissolution can be caused by structural changes, etc.) The first problem can be addressed by coating the cathode material, although strictly speaking this is also a problem with the electrolyte and can also be addressed by developing more durable electrolyte salts, using solid electrolytes or CEI-forming additives. However, coating is very easy and effective fix to an otherwise complicated problem. Internal structural problems can be addressed with doping. In recent years, there has been an almost overwhelming wealth of studies about modified cathode materials, where some of the results are questionable. Usually these studies have a bafflingly poor “pristine” material as a comparison sample, very small differences in capacities between different samples, short cycling times, no parallel cells and some have such thin film thickness on the cathode foils that in a laboratory setting it is impossible to guarantee that the loading in different samples has been equal. There are a lot of possible ways that electrochemical measurements can be influenced during making of the cell, such as bad contact, moisture, etc. Small capacity differences (5–10%) without multiple parallel cells should always be assumed to be measurement errors. The easiest way to spot results that need to be handled with care is to be aware of the level of performance that unmodified layered cathode materials are capable of, and the effects that battery testing conditions have on the results. In Figure 3.4 is the cycling behaviour of completely unmodified NCM622 (precipitated by us) that has been lithiated in air and cycled between 3.0–4.2 V for 1200 cycles. Lithiation in pure oxygen would give slightly better capacity in a Ni-rich sample such as this. The programme

Figure 3.4  Specific  discharge capacity of two full cells made with the same unmodified NCM622 cathode during cycling between 3.0 and 4.2 V.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

Modification of Layered Oxide Cathode Materials

53

starts with cycles using charging speeds of 0.03 C, 0.1 C, 0.2 C and 1 C, then continues at 2 C charging and discharging rate, with checks at 0.2 C in between and at the end. As can be seen, the discharge capacity is directly proportional to the discharge speed, and the c-rate should always be included when comparing capacities. What can also be seen is the breakdown of our air ventilation unit at about 590 cycles, where the laboratory room temperature rose from around 22 °C to about 30 °C. Higher temperatures will give higher capacities, so samples cycled in different temperatures are not directly comparable. This sample gave capacity retention of 93% after 1200 cycles. One of the reasons for this good cycling behaviour is that the cut-off voltage is limited to 4.2 V, which is both within the electrolyte comfort zone and does not induce structural changes in the cathode material. When the cut-off voltage is raised higher, the capacity will of course also be higher. In such cases care should always be taken to demonstrate that this capacity increase is justified because of improved stability and does not come at the cost of breaking the cathode. It is also important to remember that particle size and morphology of the cathode material have an effect on the electrochemical properties, and different thermal and mechanical treatments may affect things like cation disorder. Physical differences between doped and undoped comparison samples should be taken into consideration when interpreting results. Table 3.1 lists the performance of some of the common unmodified materials.

3.3.1  Coating Battery performance drops during cycling because transition metals react with or dissolve in the electrolyte under heavy cycling. One of the detrimental reactions in a full battery cell is the water catalysed breakdown of lithium Table 3.1  Theoretical  and practical capacities of different cathode materials. Material LiCoO2 LiNiO2 LiMnO2 (layered) LiMn2O4 (spinel) LiNi0.8Co0.2O2 LiNi0.8Co0.15Al0.05O2 LiNi1/3Co1/3Mn1/3O2 LiFePO4 NCM333 NCM523 NCM424 NCM622 NCM811 Mn-rich He-NCM

Theoretical Practical capacity % of Li reversibly capacity (mAh g−1) (mAh g−1) removed Source 274 275 286 148 274 279 278 170 278 278 279 277 276

142 52 145 53 Converts to spinel during cycling 120 81 180 66 180–200 65–72 160–170 56–61 170 100 154 55 164 59 155 56 178 64 >185 >67 260

26 26 26 26 26 27 27 27 28 28 28 28 28 28

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

54

Chapter 3

electrolyte salt LiPF6 into HF, which then will dissolve the cathode. Another unwanted side reaction is the degradation of organic solvents in electrolyte when exposed to voltages higher than 4.2–4.5 V (depending on the electrolyte).29–31 The idea of coating is to provide a shield that protects the cathode material from direct contact with the electrolyte. Coating has been used on LCO,32,33 NCM34 and NCA.35 All of them show positive results on improving electrochemical performance and stability of the material. A coating layer can inhibit undesired reactions between the electrode and the electrolyte and at the same time suppress structure distortion caused by transition metal dissolution in the electrolyte. In some cases, when the discharge voltage is high enough to cause electrolyte decomposition on the electrode surface, having a coating layer that prevents direct contact between the electrolyte and cathode reduces the influence of electrolyte decomposition on the battery performance.36 Other mechanisms include preventing the transition metal dissolution during charging and improving electron conductivity and lithium ion diffusion rate on the cathode surface. A proper coating material with the above-mentioned mechanisms will improve largely the electrochemical performance of the cathode material as well as thermal stability related battery safety. Many factors must be taken into consideration in the surface modification such as the chemical and physical properties of both cathode and coating materials, the thickness of the coating material used and the coating method. These factors will determine the performance of the coating layer and affect the overall performance of the cathode material.

3.3.1.1 Coating Materials Coating materials are generally classified into metal oxides, metal fluorides and phosphates, lithium salts, carbon materials and others. Oxides like Al2O3, SiO2, SnO2 etc. can provide a good barrier between the cathode material and the electrolyte and reduce the negative reactions on the interface. However, they also reduce the transportability of lithium ions as well as electrons, limiting rate performance. Lithium salts and carbon coating can increase the conductivity. In some cases, a combination of more than one coating materials is used to provide a better result.37 The pre-modification of the coating material is also a way to improve the low conductivity.38,39 ZnO can grow on the LCO surface easily due to their similarity in structure, which makes it a good choice of coating material for layered LCO. To provide a better electronic conductivity for the coating material, Dai et al.38 coated an Al2O3-doped ZnO (AZO) layer on a LCO cathode via magnetron sputtering as a small amount of Al2O3 doping in ZnO does not change the main constitution yet makes ZnO a conductive material. The electrochemical test of the AZO coated LCO gave positive results. A remarkably high stability in cycling performance test was observed. A reversible capacity of 173 mAh g−1 (90% retention rate) is maintained after 150 cycles, which was almost twice as much as the uncoated LCO. It is noted that a proper thickness of AZO coating was very important to achieve the balance between surface protection and

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

Modification of Layered Oxide Cathode Materials

55

electrochemical performance. As in their experiment, a series of thicknesses were tested; 20 nm gave the best result, as shown in Figure 3.5. The 20 nm AZO coated LCO also had a better rate performance. The coated electrode exhibited considerably higher reversible capacities at all elevated rates compared to bare electrode.

Figure 3.5  Cycling  performance of the bare and AZO-coated LCO electrodes tested

between 3.0 V and 4.5 V at 0.2 C: (a) reversible discharge capacity for 50 cycles; (b) AZO thickness dependence of the capacity retention after 50 cycles; (c) initial charge–discharge profiles; (d) the 20th charge–discharge profiles; (e) discharge capacity of the bare LCO, LCO/AZO20 and LCO/ZnO17 for 150 cycles and (f) their capacity retention after 150 cycles. Reprinted from J. Power Sources, 298, X. Dai, A. Zhou, J. Xu, B. Yang, L. Wang and J. Li., Superior electrochemical performance of LiCoO2 electrodes enabled by conductive Al2O3-doped ZnO coating via magnetron sputtering, 114–122, Copyright 2015, with permission from Elsevier.38

View Online

Chapter 3

56

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

33

Shen et al. also studied ZnO as a coating material. Unlike Dai's work, the sol–gel method was used to prepare the AZO-coated LCO, which was believed to be more economically adoptable to battery applications. Moreover, Al element instead of Al2O3 was introduced into the ZnO lattice replacing Zn partially. The AZO layer was deposited evenly throughout the surface of LCO particles. The amount of AZO coating material was 0.5 wt%, 1 wt%, 2 wt% and 4 wt% of the LCO powder respectively. 2% AZO coating gave the best result for the cycling performance test in which it maintained a capacity retention of 92.4% (173 mAh g−1) after 200 cycles between 2.75 V and 4.5 V at a current density of 100 mA g−1. The coated material also showed much better rate performance. Discharge capacity at the rate of 8 C was 86.6% of that at 0.1 C for 2% AZO coated LCO. The result for bare LCO at the rate of 8 C was only 45.4% of that at 0.1 C. The rate performance of 2% AZO coated LCO showed the best result, delivering a reversible capacity of 156 mAh g−1 at 8 C. It is believed that beside a physical barrier and conductive networks, the AZO coating material also plays an important role of HF consumer, so that the surface structure is stabilized and excellent cycling capability is retained. From the two examples of modified ZnO coating material, it is concluded that doping Al2O3 or Al could improve the conductivity of the coating material, which provides the conductive network to decrease impedance and also act as a protection layer. However, the amount of the coating material (the thickness of the coating layer) affects the electrochemical performance of the material by blocking lithium ion transportation when too thick layer is coated. It is important to find a balance. As SiO2 has a particular thermal property and it provides a scavenging effect for the hydrogen fluoride (HF), which is a common product of the reaction between residual water and LiPF6 salt in the electrolyte, it is often chosen as the coating material for Ni-rich cathode material such as NCM622 40 and NCM523,41 because in these materials the high nickel content causes cell performance degradation over time due to side reactions such as dissolution of metal from the cathode into the electrolyte, SEI. And the degradations are greatly accelerated at elevated temperatures. By coating nano-sized SiO2 onto NCM622, Cho et al. found out that the thermal stability of the coated material was improved. The main exothermic peak of decomposition of the charged electrode at 4.3 V shifted from 275 °C to 288 °C and the related heat generation was decreased from 1882 J g−1 to 1217 J g−1 thanks to the high thermal stability of SiO2. The SiO2 coating can suppress the decomposition of electrolyte at the contact surface with an electrode that triggers the oxygen generation from NCM. As a result, the cycle performance of the electrode at high temperature is improved. The SiO2 also shows the ability of HF scavenging, which benefits the capacity retention. In the experiment of Cho et al., an electrolyte containing 1000 ppm of water was used to create an environment where HF could be generated by the reaction between water and LiPF6. The coated material maintained a 92% capacity retention after 2 cycles at 0.1 C and 48 cycles at 1 C, whereas the pristine NCM was able to maintain 68% of the capacity.

View Online

Modification of Layered Oxide Cathode Materials

57

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

41

In another example of SiO2 coating, Chen et al. coated SiO2 on NCM 523 via a wet chemistry method and found similar results that a 0.5 wt% SiO2coated sample showed higher capacity retention and better cycling stability even at a high cut-off voltage and a high current density. The result of the experiment also indicated that the SiO2 coating layer with a suitable amount of silica could enhance the Li+ diffusion rate at the interface of the electrode and the electrolyte during cycling. The amount of transition metal dissolved in the electrolyte was analysed by ICP and it was found out that the amount of SiO2-coated sample dissolved in the electrolyte is lower than that of the pristine sample. So it showed that SiO2 may suppress the transition metal dissolution and the silica layer could protect active material from HF attack. The thermal stability of the material was also improved, the pristine material had an exothermic peak at 436 °C and the reaction released heat of 278.7 J g−1 whereas the coated sample had a higher peak at 442 °C and less heat was released (171.5 J g−1). This illustrated that the SiO2 coating layer reduced heat generation and suppressed oxygen release. Other oxides coating materials have also been studied. Cao et al.42 studied Al2O3 coated NCM523 and found out that the coated material had high reversible capacity, good cycling stability and better rate capability compared to the bare NCM because the Al2O3 coating layer could suppress oxygen elimination, prevent HF corrosion and reduce interface impedance. Hildebrand et al.43 compared the effects of Al2O3, SiO2 and TiO2 coated NCA cathode material and concluded that all these coating materials improved thermal stability. However, the overall capacity was decreased due to their additional inactive mass. Min et al.44 coated Co3O4 on LiNi0.91Co0.06Mn0.03O2. In fact, Co3O4 is reactive with Li residue on the material surface forming a layer of LiCoO2. The coated NCM exhibited superior capacity retention and rate capability compared to bare material. Different oxides can result in quite different electrochemical properties of the cathode materials upon coating due to their unique chemical and physical properties. However, the general role of coating material is to provide a physical barrier between the electrode and electrolyte to prevent undesired reactions.

3.3.1.2 Li+ Conducting Coating Materials Lithium salt is another kind of coating material, some of which can also be considered as active electrode materials, such as LFP,45 LTO46 and Li2SiO3 (inactive material).47,48 Lithium salts, compared with the above-mentioned oxides and fluorides, have a unique advantage that can be described as high Li+ ion conduction. Meng et al.49 coated NCM811 with Li2TiO3 by coating TiO2 onto NCM hydroxide precursor and lithiating the material through physical grounding. The weight ratio of Li2TiO3 to NCM was 3 wt%. It was found that a Li2TiO3 coating layer could suppress the decomposition of LiPF6 in the electrolyte. The surface of the coated cathode comprised more organic decomposition products, which was beneficial to the transmission of Li-ions through the cathode surface. Li2TiO3 was also beneficial to

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

58

Chapter 3

Figure 3.6  Comparison  of bare and Li2TiO3 coated NCM after 170 cycles (bare NCM on the left and coated NCM on the right). Reprinted from Electrochim. Acta, 211, K. Meng, Z. Wang, H. Guo, X. Li and D. Wang, Improving the cycling performance of LiNi0.8Co0.1Mn0.1O2 by surface coating with Li2TiO3, 822–831, Copyright 2016, with permission from Elsevier.49

stabilize the structure of NCM material. Compared to the bare reference material, the coated cathode maintained an integrated morphology after 170 cycles, as shown in Figure 3.6. Li4Ti5O12 is a ternary Li–Ti–O oxide, of which the zero-strain characteristic can ensure high structural stability and fast Li ion diffusion during cycling.46 Zhang et al. coated this material on NCM523 through a solid state synthesis process using TiO2 and CH3COOLi. Additionally, Li7Ti5O12 was converted from Li4Ti5O12 during cycling, which collectively provided good electronic and Li ion conductivity. Among the series of samples, 1.0 wt% LTO coated NCM523 exhibited the best electrochemical performance. The LTO coating layer served as both a Li ion conductive layer and a protective layer, improving the structure stability and electronic diffusion. A capacity retention of 91% was achieved after 100 cycles at cut-off voltages of 4.5 V. Meanwhile, the coated material also showed an improved thermal stability at 60 °C. Compared to SiO2 coating material, Li2SiO3 is such a material that not only stabilizes the structure of the bulk cathode materials but also provides high Li ion conduction.47,48 Two groups of researchers (Hu et al. and Zhao et al.) coated Li2SiO3 on NCM523 and NCM811. They found out that the coating material had largely improved the Li ion diffusion rate and reduced the charge transfer resistance of the electrode. The cycling performance and rate capability have also been enhanced. Li2ZrO3 is a lithium ion conductive material that can be coated on layered cathode materials such as LCO32 and NCM.50 Like previously mentioned Li4Ti5O12, Li2ZrO3 can provide a low strain path for lithium ion diffusion, which can improve the high-rate performance of the material. In the work of Zhang et al., Li2ZrO3 was coated on LCO via a synchronous lithiation route. In this synthesis route, the coating material and cathode active material were produced by lithiation of the precursor ZrO2 coated CoC2O4·xH2O (cobalt

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

Modification of Layered Oxide Cathode Materials

59

Figure 3.7  Synthesis  of Li2ZrO3 coated LCO. oxalate). The synthesis is illustrated in Figure 3.7. The ZrO2 coating procedure was realized by the reaction between the crystal water released from CoC2O4·2H2O and Zr(OC4H9)4. The as-prepared cathode material was confirmed to be a single-phased layered structure with a space group of R3−m by XRD. The lattice parameter c showed an obvious increase, which indicated that Zr4+ had also been incorporated into the bulk structure. As a result of Li2ZrO3 coating, the capacity retention of the material was largely improved at both room temperature and an elevated temperature of 55 °C. The specific capacity at 10 C was also significantly increased from 33.9 mAh g−1 to 103 mAh g−1. The coating material acted as a protective layer that inhibited the side reaction and transition metal dissolution. Meanwhile, lithium ion diffusion and electronic conductivity was improved. As a result, the cycle performance, rate capability and thermal stability of the coated material achieved a large improvement. Liang et al. coated Li2ZrO3 on NCM811 via a similar synthesis route by lithiation of the Zr material coated precursor. The coated material exhibited an improved cycle performance and rate performance. A 2% Li2ZrO3 coated sample was able to achieve a capacity retention of 94.5% after 200 cycles. It must be noted that a 1% coating sample had a higher capacity than the 2% sample after 200 cycles even though the latter sample's capacity retention was higher because the initial discharge capacity difference caused by the different amount of inactive coating material used. 1% Li2ZrO3 sample show superior high rate capability compared with bare and other amount samples. Again, the amount of coating material plays an important role in balancing between protection, sacrifice of active material and electrochemical performance, as the previous samples have illustrated. Besides the above-mentioned coating materials, there are also some others that are quite interesting. Polypyrrole (PPy),34 metals,51 etc. are among these coating materials. Core–shell structured cathode materials could also be considered as coated materials. In core–shell materials, the particles are manufactured in two phases, first a core of high capacity material that is usually high nickel such as NCM811, and then on top of this a layer of more stable material such as NCM111 is precipitated. This prevents the high nickel phase from reacting with electrolyte and supports the particle surface, which is the most likely to be over drained from its lithium content during charge. Core–shell materials are another huge branch of research centred around improving the stability of cathodes, so they are not discussed here in more detail.52

View Online

Chapter 3

60

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

3.3.2  Doping The most important goal of doping layered cathode materials is to improve the stability of the cathode in its delithiated state. During charging, Li-ions are removed from the layer structure, leaving vacancies between the MeO6 layers (Figure 3.8). As more and more Li-ions are removed, the metal oxide layers will start to repulse each other more and more because of similar charge. At some level of lithium removal, the situation becomes energetically untenable, and metal ions from the transition metal layer will start to migrate into the lithium layer. The atoms will rearrange to form a spinel or rock salt structure instead of a layer structure, releasing the “extra” oxygen that is over stoichiometric for the spinel structure. This rearrangement is irreversible and therefore ruins the battery. More importantly, this rearrangement is accompanied by the release of oxygen and heat, so it is also quite dangerous.54 In addition to over charging, a rise in temperature can trigger this cathode structure rearrangement, and cathode material temperature resistance is directly affected by the degree of lithium removal. Higher voltage (higher degree of lithium removal) means that the battery is more vulnerable to thermal runaway. Because of this instability in the delithiated state, the whole theoretical capacity of cathode can never be utilized, only certain percentage of Li-ions can be removed safely. As can be seen in Table 3.1 (in Section 3.3), the differences in capacities of layered cathode materials are not due to theoretical capacity differences, but are instead related to the fraction of Li-ions that can be removed from the material during charging. For unmodified NCM333 this is about 70% of Li-ions, which limits the safe upper voltage limit to about

Figure 3.8  Li  vacancies and cation mixing in the Li layer. Reprinted from Nano Res.,

10(12), J. Shi, D. Xiao, X. Zhang, et al., Improving the structural stability of Li-rich cathode materials via reservation of cations in the Li-slab for Li-ion batteries, 4201–4209, Copyright 2017, with permission from Tsinghua University Press and Springer-Verlag Berlin Heidelberg.53

View Online

Modification of Layered Oxide Cathode Materials

61

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

55

4.4 V vs. Li. For LCO the percentage of Li that can be safely removed is 50%, which corresponds to about 4.2 V.56 Dopants act loosely speaking as support struts, they are electrochemically inactive components that make the layer structure more resistant to changes. This means that a larger portion of the theoretical capacity can be used, safely. Other effects to consider when doping are speed of Li-ion diffusion, cation mixing, oxidation potentials and lattice volume changes. Often the amount of dopant needs to be carefully optimized because a small amount can improve structural stability, but a large amount starts to inhibit Li-ion diffusion. Some dopants are able to inhibit cation mixing, which means transition metal atoms moving from their intended place to the lithium layer. Cation mixing is in itself an unwanted phenomenon as it prevents the battery from functioning optimally,57 but is also linked to easier oxygen release from the material. Oxidation potentials determine the operating voltage of the battery. Dopants that can lessen the lattice changes will prevent microcracking and improve cycle life. A larger c-parameter will also increase the Li-ion diffusion speed in the structure.57 Large amounts of dopant are always unwanted, because most dopants do not contribute to the electrochemical activity, and therefore doped materials generally have lower initial discharge capacity. The list of dopants that have been investigated includes, but is not limited to, Na, K, Mg, Al, Ti, Cu, Cr, Ga, Fe, Zn, Sn, Mo, Zr, W, Nb, SO42−, PO43− and F−. To make sense of this plethora of substances, the vague concept of “doping” needs to be divided into several different categories. Doping methods can be divided into late-stage doping and doping during precursor manufacturing. Dopants can be divided into three categories based on where they attach in the structure. The first group forms a solid solution with the transition metal, i.e. takes the place of Me in the MeO6 octahedra. The second group intercalates to Li+ sites instead. The third group are anionic dopants, which replace O atoms.

3.3.2.1 Doping Methods Doping can be divided in to two different types depending on at which point of manufacturing it is done. More common and easier to achieve is doping during precursor heat treatment (oxidation and lithiation). Typically this “late-stage” doping is done as simply as mixing the precursor in a mechanical mixer with the wanted dopant, and then heating the mixture in an oven somewhere around 500–1000 °C.58–60 Doping in such a late stage usually gives very uneven depth distribution of the dopant, because only the surface of the particles comes in touch with the dopant, although for example Li et al.61 report uniform cross-section distribution of dopant during lithiation phase using a hydrothermal process. In some cases it is unclear whether an improvement in electrochemical behaviour is actually achieved by doping, or if the effect comes just from having a coating layer on the particles. The sampling depth of XRD, for example, is too deep to notice a thin coating layer.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

62

Chapter 3

The second sort of doping is precursor doping during manufacturing. Most commonly this means co-precipitation, adding the wanted dopant in the reactor while precursor is being precipitated,62,63 although there are other manufacturing methods where doping can be done simultaneously, such as the sol–gel method64 and self-combustion.65 All of these methods begin with aqueous metal solutions, which put a lot of restrictions on the doping element. The dopant needs to have a salt that is at least somewhat soluble in a solvent that can be added to the reactor without ruining the product, and it needs to precipitate in the same conditions as the precursor. Doping during manufacturing gives a very even distribution of the dopant throughout the precursor particles. This means that there are no concentration gradients in the material and there is less strain on the particle because it behaves similarly throughout. Anion doping is done very similarly to cation doping, but usually in the late stages. It involves mechanically mixing the precursor with a reactant salt either in solid state66 or in solvent.67

3.3.2.2 Solid Solution Dopants These dopants are intended to form a solid solution with the main cathode material, preserving the layered α-NaFeO2 structure. This means that the doping metal occupies an identical position in the MeO6 octahedra as Co atom would in LCO. Usually transition metals are used, but other metals can also be incorporated into the structure. The list of all these metals is quite long, and for ease of notation, battery scientists seem to have re-appropriated the term “transition metal” to refer to all of these metals, including aluminium, magnesium and tin. Most of these dopants work by the same mechanisms, they improve the strength of the Me–O bonds (leading to improved structural stability) and stop cation disordering by physically blocking the path of the transition metal moving to the lithium position. The effects of a dopant depend on its size (most transition metal ions are about the same size) and valence electrons (there will be charge balance reactions between the different transition metal ions in order to have them at stable oxidation states so that the energy of the system is minimized). A complete detailed discussion of all the studied dopants would be very long, so the next section includes detailed information of only some common ones. One of the best studied dopants is aluminium. Aluminium is cheap, abundant and non-toxic. Al has also been successfully doped into the precursor during precipitation, so industrial manufacturing of evenly Al doped cathodes is possible. LiAlO2 by itself has α-NaFeO2 structure, but is not electrochemically active because the intercalated Li-ions cannot be easily removed (i.e. a battery made with LiAlO2 cannot be charged).68 Instead it has been of interest as a dopant from late 1990s. Al is in the structure as Al3+. It has been shown to counteract the instability of delithiated structure in most layer structured cathodes, even composite cathodes,69 and as a result Al doping can improve both safety and lifetime of an LIB. DFT calculations show that it is energetically favourable for Al3+ to form a solid solution with the transition

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

Modification of Layered Oxide Cathode Materials

63

metals in NCM structure, preserving a layered structure instead of forming a new phase. The structural stability improvement appears to be due to the strong Al–O bonds, which inhibit structural changes in the MeO6 octahedra.70 During deep charge Al3+ also migrates preferentially and is stable in a tetrahedral site in the lithium layer. This blocks other metal atoms from moving in to the tetrahedral position, and from there to an octahedral position in the lithium layer, which is the mechanism of a layered structure transforming into a spinel structure. The transformation of layer to spinel structure releases oxygen, so preventing or slowing this reaction improves safety.71 Calculations by Ceder et al.68 and the experimental data of Jang et al.72 of Al doping into LCO and LiNiO2 73 indicate that Al doping raises the Li-ion intercalation potential. This means that the operating voltage of the battery needs to be higher in order to utilize the same capacity as in an undoped cathode. This rise is proportional to the amount of doped Al. Jang et al. report that around 50% of Co in LCO can be replaced by Al, but stability improvements were not observed by raising the doping amount above 25%. 25% Al doping reduced lattice variation during cycling almost by half and was able to prevent layered to spinel transformation on cycling between 2.5–4.4 V, although particle strain was still observed. Al doped LiNiO2 also had a smaller degree of cation mixing. Al doping shrinks the lattice parameter (assumed to be due to shortening of Me–O bonds) and lengthens the c parameter (longer interlayer distance). Unfortunately, Al in the MeO6 also hinders Li-ion mobility,70 so its amount needs to be carefully optimized. Usually the amount of added Al is no more than 5%, because further increase will hinder rate performance. Al as a dopant also does not contribute to charge of the battery so it is lost capacity. Another cheap, abundant and non-toxic dopant is iron, and it can also be doped in via a co-precipitation method. There are somewhat discordant results about the benefits of Fe doping. There seems to be no benefit in replacing Co with Fe. Fe is in the structure as Fe3+. Holzapfel et al.74 found that a large amount of Fe (20%) in LCO dropped the initial discharge capacity (cycled to 4.4 V) by half and over 40% started to give very broad peaks in XRD, indicating the loss of a well-ordered structure. Liu et al.75 doped NCM333 with increasing amounts of Fe replacing Co, and compared it to Al replacing Co. Their results (also cycled to 4.4 V) indicate that Fe doping is not equally beneficial to Al doping, both dopants slightly lowering the initial discharge capacity, but the 5% Al doped sample was the only one that showed equal or better cycling behaviour to the undoped sample. Fe doping makes the lattice parameters c and a longer, presumably because the Fe3+ has larger ionic radius than Co3+. There is another consideration to iron doping that is not a worry with Al doping: a study by Alcántara et al.76 found that iron in LCO, even at low levels (0.5%) tends to cluster and these clusters distort the layer structure in such a way that Li-ion diffusion is severely affected. This effect was lessened by proper heat treatment (850 °C), which prevented Fe clustering. Fe doped into LCO can also induce cation mixing (Fe in Li sites). This effect was lessened with excess lithium during lithiation. There might be some benefit to Fe doping above 4.4 V.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

64

Chapter 3

Calculations show that when charged to voltages over 4.4 V, Fe could oxidize from 3+ to 4+, and therefore contribute to the capacity. The cycle life is however low. Li et al. studied co-precipitated iron doped NCM333, with 5% of Ni replaced by Fe.77 Iron doping was shown to suppress cation mixing effectively. Cells cycled between 2.5–4.5 V at 0.5 C had initial discharge capacity of 146 mAh g−1 and retained 87.3% of capacity after 50 cycles. The undoped comparison sample had initial discharge capacity of 131 mAh g−1 and had capacity retention of 86.7% after 50 cycles. More importantly, the capacity loss for the doped sample happened mostly during the first five cycles, after which the capacity loss was slow, but the undoped sample lost capacity at a constant rate. Ramesha et al.78 studied Li1.2Ni0.13Co0.13Mn0.54−xMxO2 doped with Co, Fe and Cr replacing Mn. The best stability and cycling improvements were observed with 10% added cobalt. 10% Cr also showed similar or slightly better performance. Fe doped sample showed similar or worse performance compared to the pristine sample, but was able to suppress oxygen release. Mohan et al.64 doped Cr into NiO2 by the sol–gel method. Cr was shown to be incorporated into the layered structure. Cr increased the c lattice and decreased the a parameter (interpreted as strengthened metal bonds). 10% Cr substitution gave better initial capacity (204 mAh g−1) than the non-doped sample (194 mAh g−1) and capacity retention after 50 cycles (95% vs. 92%). Increasing the Cr amount further started to decrease capacity and cyclability. Cr has many oxidation levels, so it would be good dopant if it were electrochemically active, although these results indicated that it was not participating. Also chromium is about as toxic as cobalt, so as a cobalt replacement it is not ideal. Tin is a very promising dopant candidate. Eilers-Rethwisch et al.62 compared NCM622 doped with Al, Fe and Sn by the co-precipitation method. Al and Fe were shown to be incorporated into the layered structure without forming a new phase. Tin mostly formed a solid solution, though there were signs of a possible second phase, which, however, did not seem to be detrimental to the battery. When cycled between 2.5 and 4.3 V at 1 C, the undoped comparison sample NCM622 showed better capacity (∼20 mAh g−1) but dropped down to SOH 80% faster than the undoped samples. Sn doped NCM622 had the biggest improvement, about 20% longer cycling time until SOH 80%. The structural stability improvements were confirmed in TG and DSC measurements, where all doped samples improved temperature resistance, but the Sn doped sample gave 50% less mass loss up to 650 °C and released only a fifth of the heat amount compared to the undoped sample. Perhaps most interestingly the Sn doped sample showed better performance at higher C rates than the Fe or Al doped samples (though still worse than undoped sample) indicating that Sn does not inhibit the Li-diffusion speed as severely as Al and Fe. Wilcox et al.79 compared NCM333 doped with Al, Fe and Ti replacing Co by a self-combustion method. They also noticed that Fe doped materials suffer from structural problems and poor rate capability. They identified 8% Ti doped material as the best sample, because the initial capacity was same

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

Modification of Layered Oxide Cathode Materials

65

as the undoped materials. Ti doped sample also had better rate capability at high C rates than undoped material, presumed to be because the interslab layer was widened, and when cycled between 2.0 and 47 V the Ti doped sample had better capacity retention than the undoped sample. They speculated that replacing Co3+ with Ti4+ (so-called aliovalent substitution because they have different valence states) will cause a charge compensation reaction where Mn4+ is partially reduced to Mn3+. Mn3+ oxidation back to Mn4+ happens within 3–3.5 V, which means that even though the added Ti is electrochemically inactive, it will be compensated for by the Mn that is made active. Kam and Doeff80 studied Ti doping into NCM333 further. They found, based on XRD diffraction patterns, that the Ti amount can only be increased to about 7% before a second phase starts to form. Best performance was gained with 2% Ti doping. When charged to 4.3 V the initial discharge capacity (196 mAh g−1) was slightly better than undoped (184 mAh g−1). Cycling between 2.0 and 4.7 V the 2 and 3% Ti doped samples had significantly higher capacities and better capacity retention than the undoped sample. As stated, this is not a complete list of dopants. Other dopants include Zr,81 which is also found to be a beneficial dopant, but hard to be doped-in during manufacturing, Ga,82 some rare metals, etc.

3.3.2.3 Li Site Dopants Some dopants intercalate into lithium positions instead of inhabiting transition metal positions. Such dopants are low oxidation number cations such as K.59 Elements intercalating into Li sites can affect the interlayer distance. Because Li is so small, replacing it with Na or K leads to widening of the interlayer distance. This widening leads to faster Li-diffusion speed. However, if the amount of dopant is too large, the effect is the opposite, hindering Li diffusion and, because they are not electrochemically active, reducing capacity and large amounts are very bad for battery performance. 1–3% seems to be somewhat optimal.59 With LLOs there are some studies that have achieved improved material characteristics by replacing the Li in the transition metal layer. This is different from replacing lithium in the lithium layer, but still included under Li site dopant. Sallard et al.83 studied substitution of Li+ by Mg2+ in the transition metal layer in Li-rich NCM. They found that 1% Mg substitution exhibits a lower voltage drop and a more stable structure during cycling than unsubstituted material. However, substitution of 2.5% or more resulted in a higher voltage drop during cycling, due to the faster formation of the spinel-like structure.

3.3.2.4 Anion Doping Anion doping means substituting oxygen with another anion. Similar structure stabilizing effects can be achieved with anion doping as cation doping, but unlike cationic dopants, anion dopant does not take the place of an

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

66

Chapter 3

electrochemically active species. The list of elements that can be substituted for oxygen is much shorter than the ones that can be substituted for transition metals. Basically they are halogens F−, Cl−, Br− and elements that can form polyanions with oxygen such as BO33−/BO45−, SO42− and PO43−. Anion doping with, for example, fluoride has been known to improve cathode performance for decades but has seen little commercial use. Recently LLOs have raised a new interest towards anionic dopants. The mechanisms of anion doping involve complicated orbital energy level changes, and the effects that different anions have cannot be easily predicted. DFT calculations are a valuable tool in evaluating possible materials. Kim et al.84 studied F−-doping to NCM333 during lithiation and found that the doped F− amount could be raised to 15% before another phase started to form. Their results also indicated that the good effects could be achieved with only a small amount of F doping (∼5%). Increasing the amount of F− doping increased both a and c lattice parameters. The strong Li–F bond was speculated to be the reason for this, as it could increase repulsion in the oxygen layer. The strong Li–F bond was also suspected to be the reason for the drop in initial capacity (Li-ions are harder to remove) and the improvement in the cycle life when cycled up to 4.6 V (the layer structure is more stable when some Li remains in the structure). All fluorinated samples (5, 10 and 15%) also showed improved rate capability, so much that at 1 C all the doped samples had better efficiency than the undoped sample. Doped samples also had better thermal resistance and safety in DSC measurements. Yan et al.85 used DFT calculations to investigate halogen doping into nickel-based LLO. They found that Cl doping could improve stability, rate performance and lower the voltage needed to fully charge the material. In contrast, F doping could raise the needed voltage, and Br doping could lead to formation of a second phase. Li et al.86 doped boracic acid into high Mn LLO using the sol–gel method to make a material with the chemical formula of Li1.2Ni0.13Co0.13Mn0.54(BO4)0.75x­ (BO3)0.25xO2−3.75x. Doped material showed initial capacity of 319 mAh g−1 when cycled to 4.6 V at 0.1 C and capacity retention of 89% after 300 cycles, which is good for LLO. The doping also increased the material's thermal resistance. Better stability is attributed to the lessened covalent nature of M–O bonds, which in very simplified terms means that Co atoms will participate more in the redox rather than shift all of the electron vacancies on oxygen atoms. This helps reduce the amount of O2− oxidizing to O2 gas and preserves the structure.

3.4  Conclusions There are several ways in which the qualities of layered cathode materials may still be fine-tuned. Table 3.2 lists some modified materials in no particular order. The situation with conventional LIBs is starting to move from needing cathode materials with better energy capacity to needing electrolytes that can handle higher voltages so that all the capacity can be used.

Published on 14 March 2019 on https://pubs.rsc.org | doi

Material

Doping method

LiNi0.6Mn0.2Co0.15Al0.05O2 LiNi0.6Mn0.2Co0.15Sn0.05O2 LiNi0.6Mn0.2Co0.15Fe0.05O2 LiGa0.05Co0.95O2 LiGa0.1Co0.9O2 LiGa0.25Co0.75O2 Li(Ni0.5Co0.2Mn0.3)0.99Mo0.01O2 0.01%Al–LiNi0.5Co0.2Mn0.3O2 Li1.2Ni0.13Co0.13−xMn0.54AlxO2(1−y)F2y LiAlyCo1−yO2 (y = 0−0.5) LiCo1−xFexO2 (x = 0.2) LiCo0.995Fe0.005O2 Li1.2Ni0.13Co0.13Mn0.44Cr0.1O2 Li1.2Ni0.13Co0.13Mn0.49Fe0.05O2 Li1.2Ni0.13Co0.23Mn0.44O2 LiCr0.1Ni0.9O2 LiCr0.2Ni0.8O2 LiNi0.6Mn0.2Co0.15Al0.025Fe0.025O2 Li1.08Ni0.92O1.9F0.1 Li(Ni0.5Co0.2Mn0.3)0.97V0.03O2 LiNi0.59Co0.2Mn0.2Mg0.01O2

Coprecipitation Coprecipitation Coprecipitation High pressure High pressure High pressure Hydrothermal Coprecipitation Coprecipitation Solid reaction Re-anneal Sol–gel Sol–gel Sol–gel Sol–gel Sol–gel Self-combustion Solid reaction Solid reaction Coprecipitation

Li1.19Ca0.005Ni0.13Co0.13Mn0.54O2

Sol–gel

Cycling voltage (V)

Initial discharge capacity (mAh g−1)

Capacity retention (%)

2.5–4.3 2.5–4.3 2.5–4.3 3.0–4.7 3.0–5.0 3.0–5.1 2.5–4.5 3.0–4.3 3.0–4.5 2.0–4.4 3.0–4.4 3.2–4.3 2.0–4.8 2.0–4.8 2.0–4.8 3.0–4.5 3.0–4.5 2.5–4.4 3.0–4.3 2.7–4.4 2.8–4.3 2.8–4.5 2.0–4.8

145 160 156 215 197 132 154 168 250 182 164 143 224 230 248 185 155 189 200 170.5 177.1 179.7 273

80 80 80

97 88.2 61 43 87 93.7 90.4 88.8 95.1 90.3 89.9 63 88.5 90.0 87.7 82.5

Cycles C-rate

Source

140 155 148 1 1 1 50 50 150 9 2 50 50 50 50 50 50 10 100 50 100 100 100

62 62 62 82 82 82 61 63 87 72 74 76 78 78 78 64 64 65 66 88 89 89 90

1 1 1 0.02 0.02 0.02 8 0.07 0.5 0.125 0.025 0.1 0.1 0.1 0.1 0.5 0.5 0.05 1 1 1 1 0.2

Modification of Layered Oxide Cathode Materials

Table 3.2  Capacities  of doped cathode materials.

67

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

68

Chapter 3

The price of cobalt is still a major problem, because the layered materials with best stability all have some cobalt in the structure. The demand for all sorts of rechargeable battery systems is growing, so the price is unlikely to come down any time soon. It is likely that there will not be one universal LIB cathode material, instead big stationary applications will utilize cheaper but less energy dense spinel materials (LiFePO4 and LiMn2O4) and the volume and weight sensitive applications such as EVs, laptops and cellphones will continue to use layered mixed metal oxide materials.

References 1. M. Stanley Whittingham, Chem. Rev., 2004, 104(10), 4271–4302. 2. K. Mizushima, P. C. Jones, P. J. Wiseman and J. B. Goodenough, Mater. Res. Bull., 1980, 15(6), 783–789. 3. M. M. Doeff, Battery cathodes, in Batteries for Sustainability, ed. R. J. Brodd, Springer-Verlag, New York, NY, 1st edn, 2013, ch. 2. 4. T. Ohzuku and Y. Makimura, Chem. Lett., 2001, 30(7), 642–643. 5. X. Wang, X. Zheng and Y. Liao, et al., J. Power Sources, 2017, 338, 108–116. 6. A. Basch, L. de Campo, J. H. Albering and J. W. White, J. Solid State Chem., 2014, 220, 102–110. 7. R. Fallahzadeh and N. Farhadian, Solid State Ionics, 2015, 280, 10–17. 8. G. Hu, K. Du and Z. Peng, Cathode Materials for Lithium Ion Batteries: Principle, Performance and Production Technology, Beijing Industrial Publisher, Beijing, 2017, pp. 103–133. 9. N. Nitta, F. Wu, J. T. Lee and G. Yushin, Mater. Today, 2015, 18(5), 252–264. 10. M. Dixit, M. Kosa, O. S. Lavi, B. Markovsky, D. Aurbach and D. T. Major, Phys. Chem. Chem. Phys., 2016, 18(9), 6799. 11. J. Li, L. Wang, Q. Zhang and X. He, J. Power Sources, 2009, 189(1), 28–33. 12. M. Kim, H. Shin, D. Shin and Y. Sun, J. Power Sources, 2006, 159(2), 1328–1333. 13. J. W. Fergus, J. Power Sources, 2010, 195(4), 939–954. 14. H. Wang, H. He, N. Zhou, G. Jin and Y. Tang, Trans. Nonferrous Met. Soc. China, 2014, 24(2), 415–422. 15. C. H. Chen, J. Liu, M. E. Stoll, G. Henriksen, D. R. Vissers and K. Amine, J. Power Sources, 2004, 128(2), 278–285. 16. K. A. Jarvis, Z. Deng, L. F. Allard, A. Manthiram and P. J. Ferreira, Chem. Mater., 2011, 23(16), 3614–3621. 17. J. M. Paulsen, C. M. Thomas and J. R. Dahn, J. Electrochem. Soc., 2000, 147(3), 861. 18. Z. Lu, D. D. MacNeil and J. R. Dahn, Electrochem. Solid-State Lett., 2001, 4(11), A194. 19. E. Shinova, E. Zhecheva, R. Stoyanova and G. D. Bromiley, J. Solid State Chem., 2005, 178(5), 1661–1669. 20. M. M. Thackeray, S. Kang, C. S. Johnson, J. T. Vaughey, R. Benedek and S. A. Hackney, J. Mater. Chem., 2007, 17(30), 3112. 21. G. Assat and J. Tarascon, Nat. Energy, 2018, 3, 373.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

Modification of Layered Oxide Cathode Materials

69

22. J. Xu, M. Sun and R. Qiao, et al., Nat. Commun., 2018, 9(1), 1–10. 23. K. Kleiner, B. Strehle and A. R. Baker, et al., Chem. Mater., 2018, 30(11), 3656–3667. 24. M. Lang, M. S. D. Darma and K. Kleiner, et al., J. Power Sources, 2016, 326, 397–409. 25. S. B. Chikkannanavar, D. M. Bernardi and L. Liu, J. Power Sources, 2014, 248, 91–100. 26. A. G. Ritchie, C. O. Giwa and J. C. Lee, et al., J. Power Sources, 1999, 80(1), 98–102. 27. K. C. Kam and M. M. Doeff, Mater. Matters, 2012, 7(4), 56. 28. M. Fetcenko, BASF - ANL collaboration on NCM cathode materials, Updated 2014, Accessed 7th June 2018. 29. D. R. Gallus, R. Schmitz and R. Wagner, et al., Electrochim. Acta, 2014, 134(Supplement C), 393–398. 30. H. Zheng, Q. Sun, G. Liu, X. Song and V. S. Battaglia, J. Power Sources, 2012, 207(Supplement C), 134–140. 31. M. Wohlfahrt-Mehrens, C. Vogler and J. Garche, J. Power Sources, 2004, 127(1), 58–64. 32. J. Zhang, R. Gao, L. Sun, H. Zhang, Z. Hu and X. Liu, Electrochim. Acta, 2016, 209, 102–110. 33. B. Shen, P. Zuo and Q. Li, et al., Electrochim. Acta, 2017, 224, 96–104. 34. D. Wang, X. Li, Z. Wang, H. Guo, Y. Xu and Y. Fan, Electrochim. Acta, 2016, 196, 101–109. 35. X. He, C. Du and B. Shen, et al., Electrochim. Acta, 2017, 236, 273–279. 36. D. Zuo, G. Tian, X. Li, D. Chen and K. Shu, J. Alloys Compd., 2017, 706, 24–40. 37. R. Zhao, J. Liang and J. Huang, et al., J. Alloys Compd., 2017, 724, 1109–1116. 38. X. Dai, A. Zhou, J. Xu, B. Yang, L. Wang and J. Li, J. Power Sources, 2015, 298, 114–122. 39. S. Liu, Z. Wang and Y. Huang, et al., J. Alloys Compd., 2018, 731, 636–645. 40. W. Cho, S. Kim and J. H. Song, et al., J. Power Sources, 2015, 282, 45–50. 41. C. Chen, T. Tao and W. Qi, et al., J. Alloys Compd., 2017, 709, 708–716. 42. Z. Cao, Y. Li and M. Shi, et al., J. Electrochem. Soc., 2017, 164(2), A481. 43. S. Hildebrand, C. Vollmer, M. Winter and F. M. Schappacher, J. Electrochem. Soc., 2017, 164(9), A2198. 44. K. Min, K. Park, S. Y. Park, S. Seo, B. Choi and E. Cho, J. Electrochem. Soc., 2018, 165(2), A85. 45. S.-B. Kim, K. J. Lee and W. J. Choi, et al., J. Solid State Electrochem., 2010, 14(6), 919–922. 46. M. Zhang, G. Hu, L. Wu, Z. Peng, K. Du and Y. Cao, Electrochim. Acta, 2017, 232, 80–88. 47. G. Hu, M. Zhang, L. Wu, Z. Peng, K. Du and Y. Cao, J. Alloys Compd., 2017, 690, 589–597. 48. E. Zhao, M. Chen, Z. Hu, D. Chen, L. Yang and X. Xiao, J. Power Sources, 2017, 343, 345–353.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

70

Chapter 3

49. K. Meng, Z. Wang, H. Guo, X. Li and D. Wang, Electrochim. Acta, 2016, 211, 822–831. 50. H. Liang, Z. Wang, H. Guo, J. Wang and J. Leng, Appl. Surf. Sci., 2017, 423, 1045–1053. 51. A. Tornheim, V. A. Maroni, M. He, D. J. Gosztola and Z. Zhang, J. Electrochem. Soc., 2017, 164(13), A3005. 52. P. Hou, H. Zhang, Z. Zi, L. Zhang and X. Xu, J. Mater. Chem. A, 2017, 5(9), 4254–4279. 53. J. Shi, D. Xiao and X. Zhang, et al., Nano Res., 2017, 10(12), 4201–4209. 54. S. Bak, K. Nam and W. Chang, et al., Chem. Mater., 2013, 25(3), 337–351. 55. S.-C. Yin, Y.-H. Rho, I. Swainson and L. F. Nazar, Chem. Mater., 2006, 18(7), 1901–1910. 56. G. G. Amatucci, J. M. Tarascon and L. C. Klein, Solid State Ionics, 1996, 83(1), 167–173. 57. K. Kang and G. Ceder, Phys. Rev. B, 2006, 74(9), 094105. 58. A. H. Tavakoli, H. Kondo, Y. Ukyo and A. Navrotsky, J. Electrochem. Soc., 2012, 160(2), A305. 59. Y. Sun, L. Zhang and Y. Zhou, et al., J. Electrochem. Soc., 2018, 165(2), A338. 60. Y. Zhang, Z. Wang and J. Lei, et al., Ceram. Int., 2015, 41(7), 9069–9077. 61. L. Li, Y. Li and P. Li, et al., Ceram. Int., 2017, 43(4), 3483–3488. 62. M. Eilers-Rethwisch, M. Winter and F. M. Schappacher, J. Power Sources, 2018, 387, 101–107. 63. D. Aurbach, O. Srur-Lavi and C. Ghanty, et al., J. Electrochem. Soc., 2015, 162(6), A1027. 64. P. Mohan, K. Kumar, G. Kalaignan and V. Muralidharan, J. Solid State Electrochem., 2012, 16(12), 3695–3702. 65. W. El Mofid, S. Ivanov, A. Konkin and A. Bund, J. Power Sources, 2014, 268, 414–422. 66. K. Kubo, M. Fujiwara, S. Yamada, S. Arai and M. Kanda, J. Power Sources, 1997, 68(2), 553–557. 67. S. Han, J. H. Song, T. Yim, Y. Kim, J. Yu and S. Yoon, J. Electrochem. Soc., 2016, 163(5), A750. 68. M. K. Aydinol, D. R. Sadoway, Y.-I. Jang, B. Huang, G. Ceder and Y.-M. Chiang, Nature, 1998, 392(6677), 694–696. 69. B. Seteni, N. Rapulenyane, J. C. Ngila and H. Luo, Mater. Today, 2018, 5(4, Part 2), 10479–10487. 70. M. Dixit, B. Markovsky, D. Aurbach and D. T. Major, J. Electrochem. Soc., 2017, 164(1), A6365. 71. M. Guilmard, L. Croguennec, D. Denux and C. Delmas, Chem. Mater., 2003, 15(23), 4476–4483. 72. Y. Jang, B. Huang and H. Wang, et al., J. Electrochem. Soc., 1999, 146(3), 862–868. 73. Y. Jang, B. Huang and H. Wang, et al., J. Power Sources, 1999, 81, 589–593. 74. M. Holzapfel, R. Schreiner and A. Ott, Electrochim. Acta, 2001, 46(7), 1063–1070.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00044

Modification of Layered Oxide Cathode Materials

71

75. D. Liu, Z. Wang and L. Chen, Electrochim. Acta, 2006, 51(20), 4199–4203. 76. R. Alcántara, G. Ortiz, J. L. Tirado, R. Stoyanova, E. Zhecheva and S. Ivanova, J. Power Sources, 2009, 194(1), 494–501. 77. L. Hongjian, G. Chen, B. Zhang and J. Xu, Solid State Commun., 2008, 146, 115. 78. K. Ramesha, R. N. Ramesha and C. P. Laisa, Electrochim. Acta, 2017, 249, 377–386. 79. J. Wilcox, S. Patoux and M. Doeff, J. Electrochem. Soc., 2009, 156(3), A198. 80. K. C. Kam and M. M. Doeff, J. Mater. Chem., 2011, 21(27), 9991–9993. 81. S. H. Oh, S. M. Lee, W. I. Cho and B. W. Cho, Electrochim. Acta, 2006, 51(18), 3637–3644. 82. R. Stoyanova, E. Zhecheva and G. Bromiley, et al., J. Mater. Chem., 2002, 12(8), 2501–2506. 83. S. Sallard, D. Sheptyakov and C. Villevieille, J. Power Sources, 2017, 359, 27–36. 84. G. Kim, M. Kim, S. Myung and Y. K. Sun, J. Power Sources, 2005, 146(1), 602–605. 85. H. Yan, B. Li, Z. Yu, W. Chu and D. Xia, J. Phys. Chem. C, 2017, 121(13), 7155–7163. 86. B. Li, H. Yan and J. Ma, et al., Adv. Funct. Mater., 2014, 24(32), 5112–5118. 87. B. Guo, J. Zhao and X. Fan, et al., Electrochim. Acta, 2017, 236, 171–179. 88. H. Zhu, T. Xie and Z. Chen, et al., Electrochim. Acta, 2014, 135, 77–85. 89. Z. Huang, Z. Wang and X. Zheng, et al., Electrochim. Acta, 2015, 182, 795–802. 90. C. P. Laisa, R. N. Ramesha and K. Ramesha, Electrochim. Acta, 2017, 256, 10–18.

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00072

Chapter 4

Solid Electrolytes for Lithium Metal and Future Lithium-ion Batteries Gebrekidan Gebresilassie Eshetu, Xabier Judez, Chunmei Li, Maria Martinez-Ibañez, Eduardo Sánchez-Diez, Lide M. Rodriguez-Martinez, Heng Zhang* and Michel Armand* CIC Energigune, Parque Tecnológico de Álava, Albert Einstein 48, 01510 Miñano, Álava, Spain *E-mail: [email protected], [email protected]

4.1  Introduction In a climate-neutral society, the need for electrified means of transportation (xEVs) and deployment of large-scale renewable sources is increasingly becoming indispensable. These have vast potential to reduce dependence on fossil fuels and trim down the emission of greenhouse gases, and thereby protect the environment and the Earth's inhabitants. However, the reliable integration of such innovative applications requires the utilization of efficient energy storage devices. Among existing electrochemical storage devices, lithium-ion batteries (LIBs) have been scrutinized as the best choice to power such large-scale applications. Nevertheless, they are still far from meeting the future xEVs stringent requirements on high energy density, excellent safety,

  Future Lithium-ion Batteries Edited by Ali Eftekhari © The Royal Society of Chemistry 2019 Published by the Royal Society of Chemistry, www.rsc.org

72

View Online

Solid Electrolytes for Lithium Metal and Future Lithium-ion Batteries

73

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00072

1–3

long lifecycle, and wide operating temperature range. Accordingly, the scientific community is extensively looking at all options in a cell that can leapfrog the overall cell performance. Amid others, the electrolyte, positioned between the highly reducing negative and oxidizing positive electrodes, is one of the key components towards the success of batteries for xEVs. Conventional LIBs utilize solid electrodes (e.g., LiCoO2, LiNixMnyCozO2) and highly flammable/combustible organic liquid electrolytes (e.g., LiPF6 in a mixture of cyclic and linear carbonates). Under abuse conditions, such as mechanical, thermal or electrical, liquid electrolytes show poor safety characteristics, one of the most prevailing technological barriers for the massive development of large-sized devices. One of the viable strategies to bypass the limitations of current LIBs is to replace the existing state-of-the-art, alkyl carbonate-based liquid electrolytes, with solid electrolytes (SEs), which enable the building up of intrinsically safer, and possibly eco-friendly, lithium battery systems. SE materials mainly embrace solid polymer electrolytes (SPEs), inorganic solid electrolytes (ISEs), and their composite/hybrid versions. Generally, SPEs present advantages like cost effectiveness, production, processing (shaping, patterning, and integration), flexible battery design, easier manipulation (including fabrication of ultra-thin films), etc. Despite all these intriguing features, their ionic conductivities (σ) are low and a high value (>10−4 S cm−1) at room temperature (RT) is yet to be demonstrated by true (no plasticizer) SPEs.4 In comparison, ISEs possess high ionic conductivities comparable to liquid electrolytes, as well as Li-ion transference numbers (TLi+) close to unity, which could be beneficial for avoiding concentration gradients within the cell while operating. Some of the major challenges of ISEs include inferior power densities (mainly attributed to the poor contact between ISEs and electrodes), difficult cell processing and preparation techniques, as well as interfacial stress-related to volume variations upon (de-)lithiation of commonly employed electrode materials.5 In this chapter, a brief introduction of transport properties and detailed surveys on the status of the research on SEs are reported. In particular, attention is paid to the very recent interesting findings and breakthroughs in the field of SEs, instead of screening/analyzing the physicochemical and electrochemical properties of reported electrolytes that have been scrutinized in recently published reviews.6–9 Moreover, remarks and thoughts on the existing challenges and future outlook are presented.

4.2  E  volution and Recent Advancements in Solid Electrolytes 4.2.1  Ionic Transport in Solid Electrolytes A sufficient ionic conductivity of Li+ ion is always pursued for achieving good electrochemical performances of rechargeable lithium metal batteries (LMBs) and LIBs. Prior to the extensive discussion on the ionic conductivity

View Online

Chapter 4

74

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00072

of SEs in the following sections, a brief and general guideline is given regarding the understanding of ionic transport of SEs at both macroscopic (i.e. ionic conductivity-temperature dependence) and microscopic (i.e. transport mechanism) levels.

4.2.1.1 Ionic Transport at Macroscopic Level In 1889, Arrhenius10 proposed a succinct empirical equation (eqn (4.1)) for the description of temperature (T, K) dependence of reaction rate (k, rate constant). This equation has been extended to express the relationship between ionic conductivity (σ, S cm−1) and temperature (eqn (4.2)). In both equations, A (or σ0), Ea, and R are the pre-exponential factor, activation energy (J mol−1), and ideal gas constant (J mol−1 K−1), respectively. In 2009, Frech et al.11 suggested a “compensated” Arrhenius equation that introduces the temperature dependence of the dielectric constant (ε(T)) into the pre-exponential factor (eqn (4.3)) in order to annul the non-Arrhenius factors.   

     

E k(T ) A exp   a   RT

  

E   (T )  0 exp   a  RT

(4.1)

  

(4.2)

E   (T , )  0 [ (T )]exp   a  .  RT 

(4.3)

   In general, an Arrhenius type law has been widely used for the expression of the σ = F(T) (or σT = F(T)) dependence in inorganic electrolytes with ordered structures (such as Li3N, LISICON and β-Al2O3 12), as well as in crystalline polymer electrolytes.13 However, for disordered glass, i.e. amorphous materials (such as silicates, phosphates and chalcogenide glasses12) and, in particular, polymer electrolytes, the σ–T dependences deviate from the typical Arrhenius behavior (i.e. a linear correlation between the logarithm of conductivity and the inverse of temperature). In these cases, the Vogel–Tamman–Fulcher (VTF, eqn (4.4))14–16 and Williams–Landel–Ferry (WLF, eqn (4.5))17 equations are adopted. In both expressions, A, B, C1, and C2 are constants, T0 is the ideal glass transition temperature, and Ts is a reference temperature. Effectively, the WLF equation can be transformed into the VTF equation when C1 = B/(T − Ts) and C2 = Ts − T0.   

     

  (T ) AT



1 2

 B exp    T T0 

  

log[σ(T)/σ(Ts)] = C1(T − Ts)/(C2 + T − Ts).

(4.4) (4.5)

View Online

Solid Electrolytes for Lithium Metal and Future Lithium-ion Batteries

75

18–21

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00072

In the 1980s, Armand et al. suggested the employment of the VTF equation to correlate the σ–T dependence of Li-ion conducting polymer electrolytes. Currently, VTF equations are extensively used for studying the ionic transport phenomena in SEs for rechargeable LMBs and LIBs.

4.2.1.2 Ionic Transport at Microscopic Level Several theoretical models have been suggested for characterizing the ion diffusion in SEs. For example, while the random walk theory,22 the two-state model,23 the lattice gas model,24 and the continuous stochastic model25 are considered for crystalline materials, the percolation model26–30 and the free-volume theory31,32 are often used to describe the amorphous states (inorganic glass and polymer electrolytes). Instead of a comprehensive description of those theories on ISEs, which are well discussed in previous work,12 the most commonly used and relevant models, i.e. the ion hopping mechanism and the dynamic percolation model, are given in this section. In general, the Arrhenius equation presents a microscopic illustration where ionic transport is achieved by the hopping of ionic species from currently occupied sites to vacant sites. As shown in Figure 4.1a, in the crystalline LiAsF6/poly(ethylene oxide) (PEO) electrolyte, the ion motion is mainly contributed by the hopping of Li+ cations in the PEO semi-helical skeleton. That is, ion transport is facilitated by intra- or interchain hopping via the Li–O bonds breaking/forming.13 By contrast, the VTF equation depicts the segmental motion of a polymer chain, instead of the diffusivity or mobility of the ion species. According to the free-volume theory developed by Cohen and Turnbull,31 molecular transport occurs only at the moment when voids have a volume larger than a certain critical volume. The expansivity of the materials with the increase of temperature yields free volume, which allows the ionic carriers, solvated molecules, or polymer chains to move. Ratner et al.32 developed a dynamic percolation model for describing ion transport in polymer electrolytes (PEs). This microscopic model characterizes the ionic motion via jumping between neighboring positions, where the Li+ cations are transferred by a complete exchange of ligand surrounded in the local coordination environment, as sketched in Figure 4.1b. As a remark, the mechanisms of ionic transport in both organic polymer and inorganic electrolytes are rather sophisticated, due to the co-existence of multiple phases and structures, thus different models are required for simulating the diffusion of ionic species and accompanying mechanism. Nevertheless, for PEs, there is a general agreement that lowering Tg is of great importance for increasing the mobility of polymer segments, thus leading to an improved ionic conductivity. This is one of the major pathways to enhance the ionic mobility and is detailed in the following sections.

View Online

Chapter 4

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00072

76

Figure 4.1  (a)  Schematic diffusion pathway of the Li+ cations based on the struc-

tures of LiAsF6/PEO. Adapted from ref. 13 with permission from American Chemical Society, Copyright 2003. (b) Schematic diffusion pathway of the Li+ cations based on the dynamic percolation model proposed by Ratner et al. Adapted from ref. 32 with permission from American Chemical Society, Copyright 1988.

4.2.2  Solid Polymer Electrolytes Polymer electrolytes (PEs) are generally comprised of lithium salts (Li+X−, X− = monovalent anion) with low dissociation energy, in polymer matrices with high donor number (DN), and/or solid additives, and low-molecular-weight plasticizers. SPEs, containing only lithium salts, polymer matrices, and/or other solid additives, have better mechanical properties and fewer fire-induced hazards (i.e. thermal and chemical threats), compared to those added with plasticizers (i.e. gel polymer electrolytes, GPEs), though the former tends to be less ionic conductive.33 Besides, SPEs have many advantages over conventional liquid electrolytes such as low reactivity (kinetically) with polarized electrodes, no risk of electrolyte leakage and release of harmful/toxic gases, slower Li dendrite growth rate. In contrast to ISEs, SPEs can also attenuate the interfacial resistance and improve the electrode–electrolyte compatibility.9,34 The discovery of SPEs dates back four decades ago when Wright et al.35,36 reported the conductivity in complexes formed by NaI and NaSCN salts and PEO. In 1979, Armand et al.37 proposed the application of these SPEs in electrochemical devices based on Li batteries in an attempt of developing all-solid-state LMBs. These findings, combined with the safety issues related to the state-of-the-art flammable/combustible liquid electrolytes, have triggered the development of large families of PEs. Accordingly,

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00072

Solid Electrolytes for Lithium Metal and Future Lithium-ion Batteries

77

PEs have attracted the attention of both the academia and industry sectors. The feasibility of the technological application of SPEs has been well demonstrated by the expansion of the Bluecar® commercialized by Bolloré, equipped with a Li°|SPE|LiFePO4 battery.6 The Bolloré BlueCar is equipped with a 30 kWh Li°|SPE|LiFePO4-metal battery pack offering a 250 km (160 miles) driving range (under normal urban driving conditions) and 150 km (93 miles) on highways on one charge at a C/4-rate, a much higher autonomy compared to their lithium-ion counterparts and a maximum speed of 130 km h−1 (81 mph). However, the following are among the main challenges of the existing LMB technology. (1) It demands large Li metal excess (three-fold); (2) the battery needs to be kept plugged to maintain the operating internal temperature of 60−80 °C when not in use, which is imposed by the low polyethylene (PE) ionic conductivity at RT; (3) the positive electrode materials choice is limited to LiFePO4, due to the narrower electrochemical window of PEO.

4.2.3  PEO-based SPEs Among all the investigated polymer hosts for SPEs, PEO has received more attention and accordingly extensive research. The repeating units of PEO, ethylene oxide (–CH2CH2O–, EO), are properly spaced ether solvating units (low entropy change) that have strong solvation abilities (DN = 22), thus allowing the formation of the favored lithium salt/PEO complex and providing a sufficient concentration of charge carriers. Moreover, it presents optimal conditions for Li+ dissociation, and high chain flexibility (lower barrier to bond rotation) resulting in rapid ion transport.20 However, PEO-based SPEs have significant drawbacks such as low ionic conductivities at RT ( LiPF6 > LiClO4 > LiBF4 > LiCF3SO3.

LiClO4 has been used as a lithium salt in primary lithium cells. Compared with other lithium salts, it is relatively less hygroscopic and stable to ambient moisture. It is commonly used in preparing gel polymer electrolyte because it is easy to handle, and the salt is of relatively low cost. However, LiClO4 has a safety concern because it readily reacts with most organic species under certain conditions such as high temperature and high current charge. Nevertheless, it is still being used in preparing gel polymer electrolytes. LiBF4 is a salt based on an inorganic super-acid anion. Among anions in common lithium salts, BF4− has the highest mobility, but its dissociation is significantly lower than other lithium salts, resulting in moderate ion conductivity in organic solvents.20 Its low ion conductivity has been a major obstacle to its battery application. LiPF6 has superior overall physicochemical characteristics including high ionic conductivity, good solubility, high electrochemical stability and chemical stability. A critical drawback of LiPF6 is low thermal stability and high moisture sensitivity as compared with other lithium salts. It tends to decompose at high temperature and produces highly Lewis acidic PF5 gas that reacts with water to generate HF. Accordingly, the gel polymer electrolytes prepared with LiPF6 exhibit high ionic conductivity but low thermal stability. Battery performance may deteriorate with side reactions and electrolyte decomposition when HF is produced in the presence of moisture Table 5.2  Physicochemical  properties of lithium salts for lithium-ion batteries. Salt

Tm (ºC)

Molecular weight Anion radius (g mol−1) (nm)

LiClO4 LiBF4 LiPF6 LiCF3SO3 Li(CF3SO2)2N

236 293−300 200 >300 236

106.4 93.8 151.9 156.0 286.9

0.237 0.229 0.254 0.270 0.325

σRT (ms cm−1) 1M in EC/DMC 8.4 4.9 10.7 3.1 9.0

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

106

Chapter 5

at elevated temperatures. LiCF3SO3 is thermally stable, non-toxic and insensitive to moisture as compared with LiPF6. Its major drawback is low ion conductivity as compared with other lithium salts, which is due to its low dissociation.21 Serious ion pairing in LiCF3SO3-based electrolytes occurs. Li(CF3SO2)2N is a highly ionic conductive, thermally stable and safe salt. Its high ion conductivity can be attributed to the high degree of dissociation. The high dispersion of anionic charge makes it more easily dissociated into cations and anions. It melts at around 236 °C without thermal decomposition.22 Despite these merits, the use of Li(CF3SO2)2N may cause corrosion of the aluminium current collector, thus making it inadequate for use as a significant amount of the salt in lithium-ion batteries.

5.3.2  Organic Solvent While there are many types of organic solvents, not all are suitable for preparing the gel polymer electrolyte for lithium-ion batteries. Organic solvent for rechargeable lithium battery application should meet the following requirements. (1) It should have a high dielectric constant to dissolve large amounts of lithium salts, because the dielectric constant of the solvent affects ionic dissociation and association of lithium salt. In general, organic solvents with dielectric constant larger than 15 are recommended. (2) Its viscosity should be low so that ions can easily migrate, because the ionic mobility in the gel polymer electrolytes may be affected by solvent viscosity. In this respect, organic solvents should have a low viscosity of 1 cP or less. (3) It should be electrochemically stable in the potential range of 0 to 4.5 V versus Li/Li+, because the electrode potentials approach 0–0.1 V and 4.2–4.5 V at charged state for the negative electrode and the positive electrode, respectively. (4) It should have low melting point (Tm) and high boiling point (Tb) to maintain liquid state over wide temperature ranges. The operating temperature of lithium-ion batteries is mainly determined by the melting and boiling points of the organic solvent, indicating that the solvent should remain in liquid state at a temperature as low as −20 °C. The organic solvent should also exhibit low volatility. (5) It should be safe, in other words, it should have high flash point (Tf ) and be less flammable. The most commonly used organic solvents for lithium-ion batteries along with their physicochemical properties are listed in Table 5.3. The donor number represents the nucleophilicity of the organic solvent, and the dissociation of lithium salt increases with increasing donor number. It can be seen from the data in Table 5.3 that cyclic carbonates such as ethylene carbonate (EC) and propylene carbonate (PC) have a high dielectric constant and high viscosity due to large mutual interactions between solvent molecules. While a high dielectric constant is required for organic solvents to possess high ionic conductivity, a higher dielectric constant leads to increased viscosity. In contrast, linear carbonates such as dimethyl carbonate (DMC) and diethyl carbonate (DEC) have a low dielectric constant and low viscosity. Thus, cyclic carbonates and linear carbonates are

View Online

Gel Polymer Electrolytes

107

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

Table 5.3  Physicochemical  properties of organic solvents for lithium-ion batteries.a Solvent

Dielectric Viscosity Donor Tm (°C) Tb (°C) Tf (°C) constant (cP) number

Ethylene carbonate (EC) Propylene carbonate (PC) Butylene carbonate (BC) γ-Butyrolactone (γ-BL) Dimethyl carbonate (DMC) Diethyl carbonate (DEC) Ethylmethyl carbonate (EMC) 1,2-Dimethoxyethane (DME) 1,3-Dioxolane (DOL) Methyl formate (MF) Methyl propionate (MP) Methyl acetate (MA) Ethyl acetate (EA)

39 −49 −53 −42 3 −43 −55 −58 −95 −99 −88 −98 −84

248 242 240 206 90 127 108 85 78 32 79 58 77

160 132 121 97 18 31 23 −2 −6 −19 6 −10 −3

89.6 64.4 53.0 39.1 3.1 2.8 3.0 7.2 6.8 8.5 6.2 6.7 6.0

1.9 2.5 3.2 1.8 0.6 0.8 0.7 0.5 0.6 0.3 0.4 0.4 0.5

16.4 15.1 — 18.0 17.2 15.1 — 20.0 21.2 — — 16.5 17.1

a

(Tm: melting point, Tb: boiling point, Tf: flash point).

combined to obtain desirable characteristics for organic solvents for lithium-ion batteries. A synergistic effect in ionic conductivity is achieved when cyclic carbonates and linear carbonates are mixed, because the merits of each individual solvent are imparted to the resultant mixture. Among various cyclic carbonate-based solvents for preparing the gel polymer electrolytes, EC is the core and indispensable component, because EC can form an effective protective film on a graphitic anode that prevents any further electrolyte decomposition on the surface of the anode. This protection cannot be realized with PC and the exfoliation of graphite occurs due to the co-intercalation of PC into the graphite.23 Organic solvents usually undergo reductive decomposition due to their tendency to reduce at low potential. This can be solved by adding a small amount of additives such as vinylene carbonate (VC) and fluoroethylene carbonate (FEC). They are common additives used to form a stable and ionic-conductive solid electrolyte interphase (SEI) layer at the surface of a carbon anode.24 They can produce a stable SEI layer during the initial charging process and improve the cycling stability by preventing carbon exfoliation and direct reactions with organic electrolyte.

5.3.3  Polymer In gel polymer electrolytes, the role of the polymer is to sustain a solid-state film that supports ion migration in organic solvents. Because many polymers are commercially available, the number of gel polymer electrolytes reported in the literature is quite high. Among them, several types of polymers, such as PEO, polyacrylonitrile (PAN), poly(vinylidene fluoride) (PVdF), poly(vinylidene fluoride-co-hexafluoropropylene) (PVdF–HFP), poly(methyl methacrylate) (PMMA) and poly(vinyl chloride) (PVC), have been mainly used in

View Online

Chapter 5

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

108

Figure 5.1  Chemical  structure of polymers used in preparing gel polymer electrolytes for lithium-ion batteries.

preparing the gel polymer electrolytes. Figure 5.1 shows the chemical structure of polymers that have been investigated as hosts in preparing gel polymer electrolytes.

5.3.3.1 Poly(Ethylene Oxide) (PEO) It was discovered in the mid-1960s that PEO could form complexes with potassium salts,25 but it was not until 1973 that these materials had noticeable ionic conductivity. These PEO-based solid polymer electrolytes exhibit ionic conductivities ranging from 10−8 to 10−6 S cm−1 at room temperature.26,27 The low ionic conductivity at ambient temperature originates from a high degree of crystallinity that is unfavourable for ionic conduction in solid polymer electrolytes.28 The ionic conductivities of PEO-based polymer electrolytes could be improved by the addition of plasticizers. In addition to reducing the crystallinity and increasing the segmental motion of the polymer chain, the organic plasticizers can also result in greater ion dissociation. Low-molecular-weight polyethers, organic solvents and ionic liquids are commonly used types of plasticizers for the purposes. The addition of poly(ethylene glycol) (PEG) as a low-molecular-weight polyether to PEOLiCF3SO3 complexes increased the ionic conductivity; however, the presence of hydroxyl end groups negatively affected the interfacial properties at the electrode and electrolyte interface.29–31 An increase of ionic conductivity with the addition of PEG can be ascribed to the reduction of crystallinity and the increase of free volume in PEO. A gel polymer electrolyte containing PEO and a conventional LiPF6-based liquid electrolyte was prepared by in situ polymerization using UV irradiation.32 The gel polymer electrolyte exhibited a high ionic conductivity of 3.3 × 10−3 S cm−1 and high lithium transference number of 0.76 at room temperature. A Li/LiFePO4 cell showed good cycling stability with a high capacity retention of 81% after 500 cycles. Jung et al. reported that the plasticized hybrid electrolytes composed of PEO, lithium

View Online

Gel Polymer Electrolytes

109

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

+

perchlorate, Li ion-conducting ceramic particles and succinonitrile (SN) showed high ionic conductivities ranging from 3.0 × 10−5 to 1.1 × 10−4 S cm−1 at room temperature and exhibited good electrochemical stability.33 They added a small amount of SN known as a plastic crystal with high polarity into the polymer electrolyte because of its ability to dissolve lithium salts. A Li/LiFePO4 cell assembled with hybrid electrolyte delivered a high discharge capacity and exhibited good capacity retention at ambient temperature. Nagasubramanian et al. investigated the effect of crown ether as a plasticizer on the ionic conductivity and interfacial kinetics of PEO–LiX (X= CF3SO3, BF4, ClO4) complexes.34 They reported that maximum ionic conductivity of 7.0 × 10−4 S cm−1 was obtained when the 12-crown-4 to lithium salt ratio was kept at 0.003. In addition, the 12-crown-4-incorporated gel polymer electrolyte exhibited a lower charge transfer resistance in a cell. Appetecchi et al. prepared the composite gel polymer electrolytes composed of polymers (PEO and poly(ethylene glycol) dimethyl ether (PEGDME)), plasticizers (DMC, DEC, EC, PC), a lithium salt (Li(CF3SO2)2N) and a ceramic filler (γ-LiAlO2).35,36 The composite gel polymer electrolyte showed high ionic conductivities, a wide electrochemical stability window and good stability towards the lithium metal. The improved stability of the lithium–electrolyte interface was attributed to the addition of γ-LiAlO2 filler. Zhu et al. reported the PEO-based gel polymer electrolytes containing two different types of ionic liquids, 1-ethyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide (EMI-TFSI) and N-methyl-N-propylpiperidinium bis(trifluoromethanesulfonyl)imide (PP13TFSI).37 The addition of EMI-TFSI and PP13-TFSI to P(EO)20LiTFSI polymer electrolyte resulted in the improvement of the ionic conductivity, electrochemical stability and interfacial stability. However, the most critical issue of the PEO-based gel polymer electrolytes is an anodic stability. Their oxidative stability is not so high that they cannot be applied in rechargeable lithium batteries with high-voltage cathodes such as LiCoO2, LiNiO2, LiMn2O4 and LiNixCoyMn1−x−yO2.

5.3.3.2 Polyacrylonitrile (PAN) Studies on gel polymer electrolytes based on PAN started when Feuillade and Perche prepared a plasticized PAN using an aprotic solution containing an alkali metal salt.38 The polar side chain in PAN attracts lithium ions and organic solvents, thus making it appropriate as a matrix polymer for preparing the gel polymer electrolyte. These PAN-based gel polymer electrolytes exhibit a wide electrochemical stability window, relatively good mechanical properties and chemical stability toward electrode materials. Watanabe et al. reported the ionic conductivities of the gel polymer electrolytes composed of an organic solvent and LiClO4 in PAN.39,40 They used EC, PC and N,N-dimethylformamide (DMF) as the organic solvent. They concluded that PAN is inactive for the ion transport in the gel polymer electrolyte but plays a role as a matrix for structural stability. Abraham et al. prepared a gel polymer electrolyte comprising of EC, PC, LiClO4 immobilized in PAN.41 It exhibited

View Online

Chapter 5

110

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

−3

−3

−1

high ionic conductivities of 1.1 × 10 and 1.7 × 10 S cm at −10 and 20 °C, respectively. They reported that the PAN-based gel polymer electrolytes had lithium transference numbers higher than 0.5 because of the absence of oxygen atoms in the PAN matrix. Croce et al. reported that the lithium electrode underwent passivation when in contact with PAN-based gel polymer electrolytes.42 They demonstrated the passivation of the lithium electrode by AC impedance analysis of symmetrical Li/GPE/Li cells. The expansion of the semicircles with time in the AC impedance spectra implied that the lithium electrode was passivated with time. A fire-retardant gel polymer electrolyte based on PAN was introduced by Akashi et al. by optimizing the ratio of the polymer with a combination of EC, PC and LiPF6.43 The CN triple bond in PAN was broken at around 200 °C with a carbonization reaction arising from thermal decomposition, and the ladder arrangement stiffened to form a graphene structure where carbon layers serve as protection against combustible gases. The incorporation of LiPF6 significantly reduced the carbonizing point of the gel polymer electrolytes and increased the residue of carbonaceous materials after burning. PAN-based gel polymer electrolytes could also be prepared by soaking the porous PAN membrane in organic electrolytes such as LiPF6–EC/DMC, LiBF4–EC/DMC and LiClO4–EC/DMC.44 High ionic conductivities of the order of 2.0 × 10−3 S cm−1 and sufficient electrochemical stability over 5.0 V were achieved, which allowed their application in high-voltage lithium-ion polymer batteries. Gel polymer electrolytes consisting of PAN as host polymer, LiCF3SO3 and LiBF4 as lithium salts and poly (ethylene glycol) borate ester (PEGB), and EC/PC as plasticizers were prepared by Kaynak et al.45 Ionic conductivities of 1.8 × 10−3 and 1.4 × 10−3 S cm−1 were obtained for gel polymer electrolytes containing 4PAN-10EC/PC-4LiBF4 and 4PAN-10PEGB-4LiCF3SO3, respectively, at room temperature. These gel polymer electrolytes showed good dimensional stability and a wide electrochemical stability window. The preparation of PAN-based gel polymer electrolytes usually involves the heating of PAN dissolved in EC or PC at high temperatures. When PAN and salt are completely dissolved in organic solvents, the solution is cast and then left to cool at room temperature. Despite several advantages of PAN-based gel polymer electrolytes, including high ionic conductivity, good electrochemical stability and high lithium transference number, the poor compatibility with lithium metal has impeded their practical applications.46

5.3.3.3 Poly(Vinylidene Fluoride) (PVdF) and Poly(Vinylidene Fluoride-co-Hexafluoropropylene) (PVdF–HFP) PVdF and its copolymer have been used as a host polymer for preparing gel polymer electrolytes due to their unique properties. PVdF has a high dielectric constant, which allows a high degree of salt dissociation and thus provides a high concentration of charge carriers. In addition, PVdFbased gel polymer electrolytes exhibit high oxidative stability due to the strongly electron-withdrawing group. Watanabe et al. reported that PVdF

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

Gel Polymer Electrolytes

111

could form homogeneous gel polymer electrolytes with a lithium salt, EC and PC in the proper proportions.47 Zhu et al. prepared a composite gel polymer electrolyte based on PVdF and glass fibre mats, which showed good mechanical properties.48 Its maximums stress and strain are 14.3 MPa and 1.8%, respectively. Electrospun PVdF was also used to prepare a gel polymer electrolyte.49 Tsuchida et al. reported that the ionic conductivities of the gel polymer electrolytes composed of PVdF, LiClO4 and plasticizers (EC, PC, DMF, γ-BL, PEG, PPG) strongly depended on the ionic mobility within the material.50,51 Thus, the viscosity rather than the dielectric constant of the plasticizers was the controlling parameter for achieving high ionic conductivity. Capiglia et al. prepared gel polymer electrolytes based on PVdF–HFP copolymer and an electrolyte solution of EC, DEC and LiN(CF3SO2)2.52 The change of salt concentration caused conductivity variation in the range 10−2–10−8 S cm−1. Tarascon et al. developed a reliable and practical rechargeable plastic lithium-ion battery using PVdF–HFP copolymer as a polymer matrix.53 PVdF–HFP contains amorphous domains able to trap large amounts of liquid electrolyte, and crystalline regions providing enough mechanical integrity for the processing of free-standing films, thereby eliminating the need for cross-linking. They used a Li salt-free plasticizer that is substituted by liquid electrolyte at the last stage of the cell processing through an extraction and activation step. That is, the porous PVdF–HFP membrane is firstly prepared under atmospheric conditions and gel polymer electrolyte is obtained by injecting liquid electrolyte into the porous membrane. While PVdF homopolymer does not swell well, the presence of hexafluoropropylene increased the electrolyte uptake. A solution casting of PVdF–HFP containing ca. 12% of hexafluoropropylene in carbonate solutions resulted in films capable of swelling up to 60% by volume, which maintained good mechanical properties and high ionic conductivity (1 × 10−3 S cm−1), as shown in Figure 5.2. To enhance further the electrolyte uptake and the ionic conductivity, inorganic fillers such as fumed silica were added into the polymer matrix. Cheng et al. reported the porous and chemically cross-linked gel polymer electrolytes based on PVdF–HFP copolymer as a polymer matrix, PEG as a plasticizer and polyethylene glycol dimethacrylate (PEGDMA) as a chemical cross-linking oligomer.54 Gel polymer electrolytes based on PVdF–HFP/PEG/ PEGDMA with a composition of 5/3/2 exhibited a high ambient ionic conductivity of 1 × 10−3 S cm−1 and a high tensile modulus of 52 MPa because of their porous and network structures. They are electrochemically stable up to 5.0 V versus Li/Li+ in the presence of 1 M LiPF6–EC/DEC. Kim's group reported a composite gel polymer electrolyte consisting of PVdF–HFP, organic electrolyte and SiO2(Li+) nanoparticles.55–57 In their work, core–shell structured SiO2(Li+) nanoparticles with controlled morphology were synthesized and used as functional fillers in Li+-conducting composite gel polymer electrolytes for lithium-ion polymer batteries. A gel polymer electrolyte was obtained by soaking the porous composite polymer membrane composed of PVdF–HFP and SiO2(Li+) in liquid electrolyte. Figure 5.3 is a schematic

View Online

Chapter 5

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

112

Figure 5.2  Ionic  conductivity of gel polymer electrolyte films based on P(VdFHFP) and 1 M LiPF6 in EC/PC. Reprinted from Solid State Ionics, 86, J. M. Tarascon, A. S. Gozdz, C. N. Schmutz, F. Shokoohi and P. C. Warren, Performance of Bellcore's plastic rechargeable Li-ion batteries, 49–54, Copyright 1996, with permission from Elsevier.53

representation of lithium-ion conduction in the gel polymer electrolyte containing SiO2(Li+) particles.56 As illustrated in the figure, the lithium ions can dissociate from the SiO2(Li+) particles, thus acting as charge carriers. Thus, the number of lithium ions contributing to ionic conductivity increases with increasing shell thickness of SiO2(Li+) particles, resulting in an increase of lithium transference number with shell thickness. In addition, the incorporation of SiO2(Li+) resulted in a filler network providing mechanical integrity to the gel polymer electrolyte. As a result, the preparation of a free-standing film with a thickness of 30 µm was possible, thereby eliminating the need for additional mechanical supports such as polyolefin separators. Electrochemical tests of lithium metal or lithium-ion batteries assembled with composite gel polymer electrolyte containing optimized SiO2(Li+) nanoparticles exhibited superior cycling performance in terms of discharge capacity, cycling stability and rate capability, as compared to the lithium-ion cell employing conventional liquid electrolyte.

5.3.3.4 Poly(Methyl Methacrylate) (PMMA) Gel polymer electrolyte based on PMMA was first reported by Iijima et al.58 They reported that PMMA could be used as a gelation agent and an ionic conductivity of 10−3 S cm−1 was obtained at 15 wt% PMMA. The rheological and electrochemical properties of gel polymer electrolytes based on PMMA and LiClO4/PC were reported by Bohnke et al.59 The addition of PMMA in

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

Gel Polymer Electrolytes

113

Figure 5.3  Schematic  representation of lithium-ion conduction in a composite gel

polymer electrolyte containing core–shell structured SiO2(Li+) nanoparticles with different shell thickness for a lithium-ion polymer battery. Reproduced from ref. 56 with permission from the Royal Society of Chemistry.

various proportions increased the viscosity of the solution and decreased the ionic conductivity. The ionic conductivity of elastic gel polymer electrolytes prepared by polymerization of methacrylate solution containing LiBF4 in PC was also reported.60 Appetecchi et al. investigated kinetics and stability of the lithium electrode in PMMA-based gel polymer electrolytes with different lithium salts (LiClO4, LiAsF6, Li(CF3SO2)2N).61 They concluded that the electrochemical stability of PMMA-based gel polymer electrolytes mainly depended on the polymer host and lithium salt. Vondrak et al. prepared the PMMA-based gel polymer electrolytes with PC as plasticizer and complexed with salts of various perchlorates of different cations, including lithium.62 The gel polymer electrolyte prepared with lithium salt exhibited the highest ionic conductivity and was attributed to smaller ionic radius. A main disadvantage of PMMA-based gel polymer electrolytes is their poor mechanical strength when plasticized by large amounts of organic solvents, which restricts their application as electrolytes in rechargeable lithium batteries.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

114

Chapter 5

Some studies demonstrated that copolymerization of PMMA with other polymers improved the mechanical and electrical properties of the gel polymer electrolyte. Gel polymer electrolytes prepared with PMMA copolymerized with polyacrylonitrile, polystyrene and polyethylene showed the enhanced mechanical strength to prepare the free-standing films.63–68 Gel polymer electrolytes composed of acrylonitrile(AN)-methyl methacrylate(MMA)-styrene(ST) terpolymer, 1 M LiClO4–EC/PC and inorganic fillers were prepared by Kim et al.63 They concluded that the relative AN/MMA/ST molar ratio in the terpolymer played an important role in determining the ionic conductivity, the capacity to retain electrolyte solution and the mechanical strength of the gel polymer electrolyte film. Depending on the composition of the components (polymer, liquid electrolyte and fumed SiO2), these gel polymer electrolytes exhibited a wide range of mechanical and electrical properties. The ionic conductivity reached 1.4 × 10−3 S cm−1 in the gel polymer electrolyte containing 27 wt% terpolymer, 64 wt% LiClO4–EC/PC and 9 wt% fumed silica at room temperature to give homogeneous films that exhibit good mechanical properties. The use of these gel polymer electrolytes in rechargeable lithium polymer batteries was demonstrated by the charge and discharge cycling of Li/LiMn2O4 cells. Gel polymer electrolytes composed of poly(ethylene-co-methyl acrylate) (PEMA) copolymer, LiBF4–EC/EMC/PC and fumed silica were reported.66 The ionic conductivity reached 5.8 × 10−4 S cm−1 in the gel polymer electrolyte containing 22 wt% PEMA, 65 wt% liquid electrolyte and 13 wt% SiO2 at room temperature to give free-standing films sufficient to prepare thin film for rechargeable lithium-ion polymer batteries. In general, PMMA-based gel polymer electrolytes satisfy the basic requirements as the electrolyte material; however, their poor mechanical strength has limited their practical application. The mechanical stability of the PMMA-based gel polymer electrolytes could be enhanced by forming the cross-linked network structure.69–71 Gel polymer gel electrolytes based on oligomeric polyether and cross-linked PMMA blends have been prepared via in situ polymerization.69 The synthesized cross-linked gel polymer electrolytes were free-standing films with excellent dimensional stability, mechanical integrity and strength. They exhibited high ionic conductivity at room temperature reaching 4.3 × 10−4 S cm−1 for the highest conducting sample and exceptional thermal stability. The oligomeric polyether and PMMA had molecular level interaction in the gel polymer electrolytes and the in situ polymerization process allowed precise control of the composition of the gel polymer electrolytes. Woo et al. demonstrated the enhancement of mechanical strength and cycling stability by synthesizing three-dimensional semi-interpenetrating polymer network (semi-IPN) composite gel polymer electrolyte based on PMMA.71 The semi-IPN composite gel polymer electrolyte was synthesized from PMMA, divinylbenzene, vinyl-functionalized silica and liquid electrolyte. The resulting semi-IPN composite gel polymer electrolyte was a free-standing flexible film, as depicted in Figure 5.4. It effectively encapsulated electrolyte solution and exhibited stable interfacial characteristics toward lithium electrodes.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

Gel Polymer Electrolytes

115

Figure 5.4  Schematic  illustration of the synthesis of a three-dimensional semi-in-

terpenetrating polymer network (semi-IPN) composite gel polymer electrolyte based on PMMA. Reproduced from ref. 71 with permission from American Chemical Society, Copyright 2018.

5.3.3.5 Poly(Vinyl Chloride) (PVC) An amorphous PVC has been used as a matrix polymer for gel polymer electrolyte, because an amorphous polymer structure is preferable for achieving high ionic conductivity. LiTFSI was complexed with PVC plasticized by organic solvents such as dibutyl phthalate (DBP) and dioctyl adipate (DOA).72 The plasticizing effect was reflected as a reduction in the mechanical modulus by about one to two orders of magnitude compared to pure PVC. Ionic conductivities ranged from 10−7 to 10−4 S cm−1 at 25 °C when the weight ratio of PVC was changed from 0.67 to 0.17. The lithium transference number ranged from 0.54 to 0.98 depending on the composition and temperature. The cycling behaviour of Li/LiMn2O4 cells with PVC-based gel polymer electrolytes was reported by Alamgir and Abraham.73 The gel polymer electrolyte prepared with 15 wt% PVC, 40 wt% EC, 40 wt% PC and 5 wt% LiClO4 exhibited an ionic conductivity of 1.2 × 10−3 S cm−1 at 20 °C. Shembel et al. investigated the influence of lithium salts on the thermal and electrochemical stability of gel polymer electrolyte based on chlorinated PVC.74 The conductivity of gel polymer electrolyte was the same as for the conventional

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

116

Chapter 5

liquid electrolyte systems and depended on the nature of lithium salts. The chlorinated PVC-based gel polymer electrolyte was characterized by a lower resistance of passivating films on the lithium surface as compared with PVC. PVC-based gel polymer electrolytes exhibit high mechanical strength even at high content of plasticizing solvents. PVC was also used to enhance the mechanical strength of PMMA-based gel polymer electrolytes.75 The gel polymer electrolytes based on PVC/PMMA blend offered a phase-separated structure consisting of a PVC-rich phase and an electrolyte-rich phase. At the blend ratio of PVC/PMMA = 6/4, mechanical properties and ionic conductivity increased with increasing the polarity of the liquid electrolyte together with the decrease of domain size of the electrolyte-rich phase. The lithium-ion cell assembled with carbon anode, blended gel polymer electrolyte and LiCoO2 cathode exhibited a discharge capacity of 111 mAh g−1 at 2.0 C rate with good capacity retention.

5.3.4  Inorganic Filler By adding a large amount of liquid electrolyte into a matrix polymer, gel polymer electrolytes usually exhibit high ionic conductivities exceeding 10−3 S cm−1 at room temperature. However, most efforts to increase ionic conductivity by incorporating larger amounts of liquid electrolyte are detrimental to mechanical properties and cause poor compatibility with the lithium electrode. Accordingly, to obtain the gel polymer electrolytes with improved electrical and mechanical properties, inorganic fillers such as SiO2, Al2O3, γ-LiAlO2, TiO2, ZrO2, CeO2 and BaTiO3 have been incorporated into gel polymer electrolytes.76–87 The use of an inorganic filler was firstly reported by Weston and Steele, who prepared composite polymer electrolytes by adding inorganic ceramic particles into the PEO-based solid polymer electrolyte.88 The addition of inorganic filler improves ionic conductivity and mechanical stability of gel polymer electrolytes. The particle size of the inorganic filler is an important factor that affects the ion conductivity. Park et al. reported that the addition of Al2O3 (activated acidic and neutral) enhanced the electrochemical stability of the PEO-based polymer electrolyte.81 The enhancement was associated with the hydrogen bonding between surface group of Al2O3 and the perchlorate anion. An improvement of the ionic conductivity for the composite polymer electrolyte prepared with two fillers (Al2O3 and BaTiO3) can be attributed to both the reduction of crystallinity and the increase in charge carrier concentration. Ferroelectric inorganic fillers such as BaTiO3 reduce the interfacial resistance between the electrode and electrolyte due to the permanent dipole of ferroelectric materials. Core–shell structured SiO2 nanoparticles containing lithium ions were synthesized and used as functional inorganic fillers in the composite gel polymer electrolytes for lithium-ion batteries.55–57,89 The composite gel polymer electrolytes prepared with PVdF–HFP, SiO2(Li+) particles and liquid electrolyte exhibited high ionic conductivity, good thermal stability and favourable interfacial characteristics. Shin et al. synthesized a cross-linked composite gel polymer electrolyte using vinyl- or methacrylate-functionalized SiO2 nanoparticles for

View Online

Gel Polymer Electrolytes

117

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

90,91

lithium-ion battery applications. The composite gel polymer electrolyte was prepared by an in situ cross-linking reaction between reactive SiO2 particles dispersed on the PAN membrane and electrolyte precursor containing tri(ethylene glycol) diacrylate, as schematically demonstrated in Figure 5.5. The cross-linked composite polymer electrolyte effectively encapsulated electrolyte solution without leakage. It exhibited good thermal stability as well as favourable interfacial characteristics toward electrodes. The lithium-ion cells assembled with cross-linked composite gel polymer electrolytes using fibrous PAN membrane and reactive SiO2 particles exhibited high discharge capacity and good capacity retention at both ambient temperature and elevated temperatures. Inorganic clay with high strength and stiffness can be also used as an inert filler for the reinforcement of mechanical strength of gel polymer electrolytes.

5.4  C  haracteristics and Requirements of Gel Polymer Electrolytes 5.4.1  Ionic Conductivity Ionic conductivity of a gel polymer electrolyte is the most important property that affects battery performance. The ionic conductivity of a gel polymer electrolyte for lithium-ion battery application should be higher than 10−3 S cm−1 at room temperature. The migration of lithium ions between two electrodes and diffusion of lithium ions in the electrodes are especially important when lithium-ion batteries are rapidly charged or discharged at high

Figure 5.5  Schematic  illustration of the synthesis of a cross-linked composite gel polymer electrolyte using fibrous PAN membrane, vinyl-functionalized SiO2 nanoparticles and TEGDA. Reproduced from ref. 90 with permission from the Royal Society of Chemistry.

View Online

Chapter 5

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

118

current rates. Electrode materials cannot deliver a high capacity if lithium ions are not facilely transported between two electrodes. Ionic conductivity of the gel polymer electrolyte can be measured by electrochemical impedance spectroscopy (EIS) in a certain frequency range using an AC impedance analyser. For conductivity measurement, gel polymer electrolyte film is sandwiched between two blocking electrodes such as stainless-steel electrodes, and the bulk resistance (Rb) of the gel polymer electrolyte is obtained by the intercept on the real axis in AC impedance spectrum (Nyquist plot) of the cell. The ionic conductivity can be calculated by eqn (5.1):   





t Rb A

(5.1)

   where σ is the ionic conductivity, t is the thickness of the gel polymer electrolyte film, Rb is the bulk resistance and A is the area of gel polymer electrolyte in contact with blocking electrodes.92,93

5.4.2  Electrochemical Stability The operating voltage of the lithium-ion battery is determined by the electrochemical stability window of the electrolyte. Linear sweep voltammetry (LSV) is used to evaluate the electrochemical stability window of the gel polymer electrolyte, which is the range of potential that does not participate in oxidative and reductive decomposition reactions. A potentiostat is used to scan the potential of a working electrode at a constant rate with respect to the reference electrode. When the current increases rapidly, it corresponds to the decomposition voltage of the electrolyte. A platinum or stainless-steel electrode is often used as a working electrode, and lithium metal is used as reference and counter electrodes. Gel polymer electrolytes should not be involved in decomposition arising from oxidation or reduction within the given working voltage range of the cathode and anode. Specifically, gel polymer electrolytes should be electrochemically stable up to at least 4.5 V versus Li/Li+, because the lithium-ion batteries with lithium metal oxide cathodes such as LiCoO2, LiNiO2 and LiNixCoyMn1−x−yO2 reach a voltage of 4.3 V when fully charged. Since the gel polymer electrolyte is interposed between anode and cathode in the cell, any side reactions should not occur when the gel polymer electrolytes come into direct contact with electrodes. In other words, the gel polymer electrolytes should remain inert to the charged surfaces of the electrodes during cell operation.

5.4.3  Lithium Transference Number Since the gel polymer electrolyte has more than one mobile charge carrier, it is important to determine not only the total conductivity but also the contribution of current that is carried by each species. Lithium ions are the charge carriers in lithium-ion batteries. Therefore, lithium ions should have a higher mobility than anions. The fraction of the current carried by lithium

View Online

Gel Polymer Electrolytes

119

ions in the gel polymer electrolyte can be determined from the lithium transference number (tLi+) and it is given by eqn (5.2).   

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102



 tLi





       

(5.2)

   In the above equation, the conductivity ratio can also be expressed in terms of ionic mobility because lithium salt dissociates into the same number of cations and anions. When the lithium transference number is low, the overall resistance of the cell increases due to the concentration polarization of the anions in the electrolyte, which results in reduced capacity and low power density. The lithium transference number in the gel polymer electrolytes is usually in the range 0.2–0.5. The cation transference number is close to 1.0 in hydrogen-ion conductors like the Nafion-based polymer electrolyte for proton exchange membrane fuel cells. The transference number can be measured using various methods such as alternating current impedance, direct current polarization measurement, Tubandt's method, Hittorf's method and the pulsed field gradient NMR (PFG NMR) technique. The most commonly used method for determining the lithium transference number in the gel polymer electrolyte is the combination of AC impedance and DC polarization measurements of the symmetric Li/gel polymer electrolyte/Li cell using the following equation:94,95   



t 

I s V  I o Ro  I o V  I s Rs 



(5.3)

   where V is the applied voltage to the cell, Io and Is are the initial and steadystate current, respectively, and Ro and Rs are the interfacial resistances before and after the polarization, respectively. The cell is polarized by applying a DC voltage of about 10 mV, and the time evolution of the current flow is monitored. The initial and steady-state currents flowing in the cell are measured. AC impedance measurements of the cell are performed before and after DC polarization, and the initial and steady-state interfacial resistances are measured. The lithium transference number is influenced by various factors including temperature, salt type and concentration, types of polymer, organic solvent and inorganic filler. To increase the lithium transference number, some polyelectrolytes in which anions are covalently bonded to the polymer backbone have been used as a matrix polymer. Since the anions are effectively immobilized by chemical bonding to the polymer chain, the lithium transference number nearly approaches 1.0 in polyelectrolytes.

5.4.4  Mechanical Properties Gel polymer electrolyte should function as not only an electrolyte but also as a separator between two electrodes in the cell. From a practical point of view, the mechanical strength of gel polymer electrolytes plays a vital role in the manufacture of batteries. They should facilitate easy fabrication of battery

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

120

Chapter 5

with reproducibility and sustain their mechanical integrity during assembling, storage and usage of the battery. Thus, the mechanical strength of gel polymer electrolyte is the most important factor to be considered when battery technology moves from laboratory to mass production. Unfortunately, most of the gel polymer electrolytes reported so far have exhibited poor mechanical properties, which has prevented their use without a supporting membrane in lithium-ion batteries. Increasing the amount of organic solvent improves ionic conductivity but leads to a decrease in mechanical strength. When considering this relationship, ionic conductivity and mechanical properties should be optimized within a suitable range. The mechanical strength of the gel polymer electrolytes can be enhanced by addition of inorganic fillers and in situ chemical cross-linking. The thickness of gel polymer electrolyte is an important factor that affects battery performance such as rate capability and energy density, which can be reduced by enhancing the mechanical strength of gel polymer electrolytes. In general, the film thickness of gel polymer electrolyte should be less than 30 µm in consideration of the thickness of conventional polyolefin separators currently used in lithium-ion batteries. The mechanical strength of a gel polymer electrolyte can be evaluated from the strain–stress curve by tensile test.96,97 For maintaining good interfacial characteristics and intimate contacts with electrodes in the cell during charge and discharge cycles, gel polymer electrolytes should also possess adequate adhesive properties and flexibility. Gel polymer electrolytes must be also thermally stable to have an appropriate temperature range of operation.

5.5  Preparation of Gel Polymer Electrolytes Gel polymer electrolytes can be prepared using various methods such as solution casting, hot melting, immersion of a porous membrane into liquid electrolyte and in situ cross-linking, which vary according to material characteristics and battery manufacturing process. The detailed characteristics of commonly used preparation methods are introduced below.

5.5.1  Solution Casting The easiest method for preparing gel polymer electrolytes is solution casting, which produces flexible film with various thickness by simply evaporating the volatile solvent from a casting solution of polymer and liquid electrolyte containing lithium salt. Figure 5.6 shows the process of obtaining a gel polymer electrolyte film using the solution casting method. A proper amount of polymer and liquid electrolyte is dissolved in anhydrous co-solvent with high volatility. The co-solvent must be free from water, volatile and able to dissolve both polymer and salt. A small amount of inorganic filler can be added into the solution. When complete homogenization of the mixture is achieved, the resulting viscous solution is cast with a doctor blade onto a glass plate, then left to evaporate the volatile solvent. The thickness of the

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

Gel Polymer Electrolytes

121

Figure 5.6  Preparation  of gel polymer electrolyte by solution casting. gel polymer electrolyte can be controlled by adjusting the gap of the doctor blade or changing the solid content in the polymer solution. After complete evaporation of the co-solvent, the gel polymer electrolyte film is removed from the glass plate. For the film to be easily peeled off, Teflon or release paper may be used instead of a glass plate. This simple and low-cost method is widely used in preparing gel polymer electrolytes. In the solution casting method, linear carbonate solvents such as DMC and EMC are difficult to be used since they may be easily evaporated with co-solvent together during the drying process. Accordingly, this method is limited to the preparation of gel polymer electrolytes that include organic electrolytes with solvents having high boiling points such as EC and PC. Gel polymer electrolytes based on PEO and PMMA are usually prepared by solution casting. Although the solution casting technique makes it easy to prepare gel polymer electrolyte film with proper thickness, it consumes a large amount of co-solvent to dissolve the polymer and liquid electrolyte, which makes this process tedious and environmentally non-benign. The preparation of a gel polymer electrolyte using solution casting should be also done in a moisture-free atmosphere such as a glove box and dry room, because the moisture sensitive lithium salt and hygroscopic polar solvents are used during the whole process.

5.5.2  Hot Melting Hot melting is a method of obtaining gel polymer electrolyte through direct dissolution of polymer in liquid electrolyte, followed by film casting and cooling at room temperature. This method has some advantages such as it is solvent-free, low cost and a quick method. There is no loss of components since gel polymer electrolytes can be directly prepared from organic electrolytes and polymer without highly volatile co-solvent. Organic solvents with low boiling points such as DMC and DEC cannot be used in the hot melting method, because polymer is usually dissolved in the liquid electrolyte at high temperatures. Hot melting is commonly used to prepare PAN-based gel polymer electrolytes with EC or PC as organic solvent.

View Online

122

Chapter 5

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

5.5.3  Immersion of Porous Membrane Porous membrane is prepared and immersed in an electrolyte solution to obtain gel polymer electrolyte. When soaking in the liquid electrolyte, organic solvents swell the porous membrane and eventually form a gel polymer electrolyte. Unlike hydrophobic and microporous polyolefin separators used in lithium-ion batteries, the porous membranes used in this method have a high affinity to liquid electrolyte to produce gel polymer electrolyte. PVdF–HFP is a representative polymer to be used in preparing the porous membrane. The preparation of gel polymer electrolyte using PVdF–HFP copolymer can be described as follows. First, a proper amount of polymer, plasticizer and inorganic filler is dissolved in an acetone solvent. Polar solvents with high boiling point such as DBP are suitable plasticizers while silica or alumina can be used as inert inorganic fillers. The inorganic filler added into the mixture not only enhances the mechanical strength of the porous membrane but also allows a greater amount of liquid electrolyte to be absorbed into the porous membrane. After casting the mixed solution on the glass plate at an appropriate thickness, the acetone is removed by evaporation at room temperature, and the flexible film containing a large amount of plasticizer is obtained. When this film is put into a non-solvent such as water, methanol and ether, the plasticizer exchanges with non-solvent. The porous membrane with suitable thickness is finally obtained by removing non-solvent in the film through vacuum drying. The obtained membrane is then immersed in liquid electrolyte to eventually form a gel polymer electrolyte. Instead of direct impregnation, gel polymer electrolytes can be obtained by injecting liquid electrolyte into the cell comprised of an anode/porous membrane/cathode. High porosity of the membrane increases both electrolyte uptake and ionic conductivity. However, an excessively porous structure of a polymer membrane may deteriorate the mechanical strength, leading to internal short circuit and eventually battery failure.98 Compared to the solution casting technique, this process requires a moisture-free environment only at the time of assembling the cells and its high mechanical strength can be retained during cell fabrication.

5.5.4  In situ Cross-linking Gel polymer electrolytes can be classified into two groups according to types of cross-linking for forming gel polymer electrolyte, i.e. physically and chemically cross-linked gel polymer electrolyte. Physical gel polymer electrolytes have a physically cross-linked structure when polymer chains entangle with one another or intermolecular interactions occur among polymer, salt and organic solvents. In this case, liquid electrolyte is confined in a matrix polymer without any chemical bonding. They gain mobility as the polymer chains become untangled upon heating, and then turn into gel when cooled. Accordingly, their mechanical strength may be weakened at high temperature. On the other hand, structural change is difficult

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

Gel Polymer Electrolytes

123

in chemically cross-linked gel polymer electrolytes since the three-dimensional network is formed based on chemical bonds instead of van der Waals forces. A reaction between cross-linking agents and reactive oligomers (or monomers) leads to the formation of a network structure in the gel polymer electrolytes. To fabricate lithium-ion batteries with chemically cross-linked gel polymer electrolytes, a precursor capable of chemical cross-linking is dissolved in an electrolyte solution and injected into a battery.99–105 This is followed by thermal or photo cross-linking to attain uniform gelation of the electrolyte precursor. The in situ cross-linking enables the electrolyte to firmly bond both electrodes together in the cell. The cross-linked polymer networks that are swelled with the liquid electrolyte show high ionic conductivity, favourable interfacial properties and good mechanical strength. The commonly used cross-linking agents are reactive monomers or oligomers such as tri(ethylene glycol) diacrylate (TEGDA) and polyethylene gylcol dimethacrylate (PEGDMA), which contain two reactive groups. Lee et al. prepared chemically cross-linked gel polymer electrolytes by polymerization of PEGDMA and alkyl monomer in the presence of liquid electrolyte.105 Flexibility and ionic conductivity of the gel polymer electrolyte could be improved by controlling the monomer content. Chemically crosslinked gel polymer electrolyte was also synthesized without any initiators by Lee et al.106 As cross-linking agents, they used polyethyleneimine (PEI) with amine groups and poly(ethylene glycol) diglycidyl ether (PEGDE) with epoxy groups. Figure 5.7 shows the synthetic scheme of the cross-linked gel polymer electrolyte. They investigated leakage behaviour of the electrolyte solution from the lithium-ion battery, as illustrated in Figure 5.8. The pouch cell was placed on a flat plate, and a constant pressure of 100 kgf cm−2 was applied to the pouch cell for 10 s. The lithium-ion battery assembled with the cross-linked gel polymer electrolyte exhibited no leakage after applying a constant load. In contrast, for the battery assembled with the liquid electrolyte, 21 wt% of the initial amount of liquid electrolyte exuded through the hole in the pouch bag, indicating that the electrolyte solution was well encapsulated in the lithium-ion battery assembled with in situ cross-linked gel polymer electrolyte. Due to the networked structure of chemically cross-linked gel polymer electrolytes, they undergo little structural change compared to those of physically cross-linked gel polymer electrolytes. As a result, the lithium-ion batteries assembled using in situ chemical cross-linking exhibited a higher discharge capacity and better capacity retention than liquid electrolyte-based lithium-ion batteries.

5.6  Concluding Remarks Gel polymer electrolytes exhibit high ionic conductivity, a wide electrochemical window, good thermal stability and good compatibility with electrodes. These superior properties make them the promising electrolyte for lithium-ion batteries. A proper design and selection of matrix polymer, organic solvent, lithium salt and inorganic filler is important to obtain high-performance gel

View Online

Chapter 5

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

124

Figure 5.7  Reaction  scheme for the synthesis of a cross-linked gel polymer electrolyte. (a) Fluorinated carbamate synthesized from the reaction between PEI and FEC, and (b) cross-linked polymer network obtained by a ring-opening reaction between fluorinated carbamate and PEGDE. Reprinted from Solid State Ionics, 255, Y.-W. Lee, W.-K. Shin and D.-W. Kim, Cycling performance of lithium-ion polymer batteries assembled using in situ chemical cross-linking without a free radical initiator, 6–12, Copyright 2014, with permission from Elsevier.106

Figure 5.8  Schematic  illustration of the leakage test of the pouch cell assembled with different electrolytes.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

Gel Polymer Electrolytes

125

polymer electrolytes. Despite numerous attempts made on the development of gel polymer electrolytes, the poor mechanical strength of gel polymer electrolytes has been a main barrier for their practical application, especially in large-scale lithium-ion batteries. Although many gel polymer electrolytes have been reported to be fabricated as free-standing film in the lab scale and exhibited various favourable electrochemical characteristics, their mechanical strength still requires further enhancement to allow manufacture by conventional large-scale battery processes. The addition of functional inorganic fillers and in situ chemical cross-linking may enhance the mechanical strength of the gel polymer electrolytes. In addition, the hybrid of gel polymer electrolyte with non-woven supporting membrane is suggested from the viewpoint of large-scale applications. The unique features of nonwoven membranes such as low cost, high porosity, high thermal and mechanical strength make them potential attractions to prepare gel polymer electrolyte with mechanical stability for lithium-ion batteries. The development of gel polymer electrolytes with high ionic conductivity and good mechanical stability can provide a cornerstone for high-performance lithium-ion batteries with enhanced safety. Moreover, due to the flexible nature of polymer material, the gel polymer electrolytes also allow the fabrication of lithium-ion batteries with adjustable shapes and design for wearable electronic devices in near future.

References

1. J.-M. Tarascon and M. Armand, Nature, 2001, 414, 359. 2. M. Armand and J.-M. Tarascon, Nature, 2008, 451, 652. 3. B. Dunn, H. Kamath and J.-M. Tarascon, Science, 2011, 334, 928. 4. V. Etacheri, R. Marom, R. Elazari, G. Salitra and D. Aurbach, Energy Environ. Sci., 2011, 4, 3243. 5. D. Larcher and J.-M. Tarascon, Nat. Chem., 2015, 7, 19. 6. D. D. Lecce, R. Verrelli and J. Hassoun, Green Chem., 2017, 19, 3442. 7. M. B. Armand, T. M. Chabagno and M. Duclot, Proceedings of the Second International Meeting of Solid Electrolytes, St. Andrews, Scotland, 20–22 September 1978 (Extended Abstract). 8. M. B. Armand, Solid State Ionics, 1994, 69, 309. 9. W. H. Meyer, Adv. Mater., 1998, 10, 439. 10. A. S. Arico, P. Bruce, B. Scrosati, J.-M. Tarascon and W. Van Schalkwijk, Nat. Mater., 2005, 4, 366. 11. L.-Z. Fan, Y.-S. Hu, A. J. Bhattacharyya and J. Maier, Adv. Funct. Mater., 2007, 17, 2800. 12. Y.-C. Jung, S.-M. Lee, J.-H. Choi, S. S. Jang and D.-W. Kim, J. Electrochem. Soc., 2015, 162, A704. 13. Y.-C. Jung, M.-S. Park, D.-H. Kim, M. Ue, A. Eftekhari and D.-W. Kim, Sci. Rep., 2017, 7, 17482. 14. J. Y. Song, Y. Y. Wang and C. C. Wan, J. Power Sources, 1999, 77, 183. 15. J. W. Fergus, J. Power Sources, 2010, 195, 4554.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

126

Chapter 5

16. S. Liang, W. Yan, X. Wu, Y. Zhang, Y. Zhu, H. Wang and Y. Wu, Solid State Ionics, 2018, 318, 2. 17. M. Schmidt, U. Heider, A. Kuehner, R. Oesten, M. Jungnitz, N. Ignatev and P. Sartori, J. Power Sources, 2001, 97/98, 557. 18. S. Hossain, in Handbook of Batteries, ed. D. Linden, McGraw-Hill, New York, 2nd edn, 1995, ch. 36. 19. C. W. Walker, J. D. Cox and M. Salomon, J. Electrochem. Soc., 1996, 143, L80. 20. M. Ue, J. Electrochem. Soc., 1995, 142, 2577. 21. F. Croce, A. D'Aprano, C. Nanjundiah, V. R. Koch, C. Walker and M. Salomon, J. Electrochem. Soc., 1996, 143, 154. 22. L. A. Dominey, V. R. Koch and T. Blakley, Electrochim. Acta, 1992, 37, 1551. 23. R. Fong, U. Von Sacken and J. R. Dahn, J. Electrochem. Soc., 1990, 137, 2009. 24. M. Nie, J. Demeaux, B. T. Young, D. R. Heskett, Y. Chen, A. Bose, J. C. Woicik and B. L. Luchta, J. Electrochem. Soc., 2015, 162, A7008. 25. J. Moacanin and E. F. Cuddihy, J. Polym. Sci., 1966, C14, 313. 26. D. E. Fenton, J. M. Parker and P. W. Wright, Polymer, 1973, 14, 589. 27. P. W. Wright, Br. Polym. J., 1975, 7, 319. 28. C. Berthier, W. Gorecki, M. Minier, M. B. Armand, J. M. Chabagno and P. Rigaud, Solid State Ionics, 1983, 11, 91. 29. J. E. Weston and B. C. H. Steele, Solid State Ionics, 1982, 7, 81. 30. I. E. Kelley, J. R. Owen and B. C. Steele, J. Power Sources, 1986, 14, 13. 31. Y. Ito, K. Kanehori, K. Miyauchi and T. Kodu, J. Mater. Sci., 1987, 22, 1845. 32. W. Li, Y. Pang, J. Liu, G. Liu, Y. Wang and Y. Xia, RSC Adv., 2017, 7, 23494. 33. Y.-C. Jung, M.-S. Park, C.-H. Doh and D.-W. Kim, Electrochim. Acta, 2016, 218, 271. 34. G. Nagasubramanian and S. D. Stefano, J. Electrochem. Soc., 1990, 137, 3830. 35. G. B. Appetecchi, G. Dautzenberg and B. Scrosati, J. Electrochem. Soc., 1996, 143, 6. 36. G. B. Apetecchi, F. Croce and B. Scrosati, J. Power Sources, 1997, 66, 77. 37. C. Zhu, H. Cheng and Y. Yang, J. Electrochem. Soc., 2008, 155, A569. 38. G. Feuillade and P. Perche, J. Appl. Electrochem., 1975, 5, 63. 39. M. Watanabe, M. Kanba, K. Nagaoka and I. Shinohara, J. Appl. Electrochem., 1982, 27, 4191. 40. M. Watanabe, M. Kanba, K. Nagaoka and I. Shinohara, J. Polym. Sci. Polym. Phys. Ed., 1983, 21, 939. 41. K. M. Abraham and M. Alamgir, J. Electrochem. Soc., 1990, 137, 1657. 42. F. Croce, F. Gerace, G. Dautzenberg, S. Passerini, G. B. Appetecchi and B. Scrosati, Electrochim. Acta, 1994, 39, 2187. 43. H. Akashi, K. Sekai and K. Tanaka, Electrochim. Acta, 1998, 43, 1193. 44. H.-S. Min, J.-M. Ko and D.-W. Kim, J. Power Sources, 2003, 119–121, 469.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

Gel Polymer Electrolytes

127

45. M. Kaynak, A. Yusuf, H. Aydın, M. Taşkıran and A. Bozkurt, Electrochim. Acta, 2015, 164, 108. 46. B. Scrosati, J. Electrochem. Soc., 1992, 132, 2276. 47. M. Watanabe, M. Kanba, H. Matsuda, K. Mizoguchi, I. Shinohara, E. Tsuchida and K. Tsunemi, Makromol. Chem., Rapid Commun., 1981, 2, 741. 48. Y. Zhu, F. Wang, L. Liu, S. Xiao, Y. Yang and Y. Wu, Sci. Rep., 2013, 3, 3187. 49. Y. Zhu, S. Xiao, Y. Shi, Y. Yang, Y. Hou and Y. Wu, Adv. Energy Mater., 2014, 4, 1300647. 50. E. Tsuchida, H. Ohno and K. Tsunemi, Electrochim. Acta, 1983, 28, 591. 51. K. Tsunemi, H. Ohno and E. Tsuchida, Electrochim. Acta, 1983, 28, 833. 52. C. Capiglia, Y. Saito, H. Kataoka, T. Kodama, E. Quartarone and P. Mustarelli, Solid State Ionics, 2001, 131, 291. 53. J. M. Tarascon, A. S. Gozdz, C. N. Schmutz, F. Shokoohi and P. C. Warren, Solid State Ionics, 1996, 86–88, 49. 54. C. L. Cheng, C. C. Wan and Y. Y. Wang, J. Power Sources, 2004, 134, 202. 55. Y.-S. Lee, S. H. Ju, J.-H. Kim, S. S. Hwang, J.-M. Choi, Y.-K. Sun, H. Kim, B. Scrosati and D.-W. Kim, Electrochem. Commun., 2012, 17, 18. 56. S. H. Ju, Y.-S. Lee, Y.-K. Sun and D.-W. Kim, J. Mater. Chem. A, 2013, 1, 395. 57. Y.-S. Lee, J. H. Lee, J.-A. Choi, W. Y. Yoon and D.-W. Kim, Adv. Funct. Mater., 2013, 23, 1019. 58. T. Iijima, Y. Toyoguchi and N. Eda, Denki Kagaku, 1985, 53, 619. 59. O. Bohnke, C. Rousselot, P. A. Gillet and C. Truche, J. Electrochem. Soc., 1992, 139, 1862. 60. J. Vondrak, M. Sedlaikova, J. Reitera and T. Hodalb, Electrochim. Acta, 1999, 44, 3067. 61. G. B. Appetecchi, F. Croce and B. Scrosati, Electrochim. Acta, 1995, 40, 991. 62. J. Vondrak, M. Sedlarikova, J. Velicka, B. Klapste, V. Novak and J. Reiter, Electrochim. Acta, 2001, 46, 2047. 63. D.-W. Kim and Y.-K. Sun, J. Electrochem. Soc., 1998, 145, 1958. 64. D.-W. Kim, Y.-R. Kim, J.-K. Park and S.-I. Moon, Solid State Ionics, 1998, 106, 329. 65. D.-W. Kim, B.-K. Oh and Y.-M. Choi, Solid State Ionics, 1999, 123, 243. 66. D.-W. Kim, J. Power Sources, 2000, 87, 78. 67. M. M. Rao, J. S. Liu, W. S. Li, Y. Liang and D. Y. Zhou, J. Membr. Sci., 2008, 322, 314. 68. P. Zhang, H. P. Zhang, G. C. Li, Z. H. Li and Y. P. Wu, Electrochem. Commun., 2008, 10, 1052. 69. J. J. Xu and H. Ye, Electrochem. Commun., 2005, 7, 829. 70. W.-L. Li, Y.-M. Gao and S.-M. Wang, J. Appl. Polym. Sci., 2012, 125, 1027. 71. H.-S. Woo, Y.-B. Moon, S. Seo, H.-T. Lee and D.-W. Kim, ACS Appl. Mater. Interfaces, 2018, 10, 687.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

128

Chapter 5

72. A. M. Sukeshini, A. Nishimoto and M. Watanabe, Solid State Ionics, 1996, 86–88, 385. 73. M. Alamgir and K. M. Abraham, J. Electrochem. Soc., 1993, 140, L96. 74. E. M. Shembel, O. V. Chervakov, L. I. Neduzhko, I. M. Maksyuta, Yu. V. Polischyk, D. E. Reisner, P. Novak and D. Meshri, J. Power Sources, 2001, 96, 20. 75. H.-T. Kim, K.-B. Kim, S.-W. Kim and J.-K. Park, Electrochim. Acta, 2000, 45, 4001. 76. M. C. Borghini, M. Mastragostino, S. Passerini and B. Scrosati, J. Electrochem. Soc., 1995, 142, 2118. 77. J. Fan and P. S. Fedkiw, J. Electrochem. Soc., 1997, 144, 399. 78. D. E. Strauss, D. Golodnitsky and E. Peled, Electrochem. Solid-State Lett., 1999, 2, 115. 79. F. Croce, G. B. Appetecchi, L. Persi and B. Scrosati, Nature, 1998, 394, 456. 80. T. Itoh, Y. Miyamura, Y. Ichikawa, T. Uno, M. Kubo and O. Yamamoto, J. Power Sources, 2003, 119, 403. 81. C. H. Park, D.-W. Kim, J. Prakash and Y.-K. Sun, Solid State Ionics, 2003, 159, 111. 82. Y. Liu, J. Y. Lee and L. Hong, J. Appl. Polym. Sci., 2003, 89, 2815. 83. H. Han, W. Liu, J. Zhang and X.-Z. Zhao, Adv. Funct. Mater., 2005, 15, 1940. 84. C. M. Yang, H. S. Kim, B. K. Na, K. S. Kum and B. W. Cho, J. Power Sources, 2006, 156, 574. 85. M. Moskwiak, I. Giska, R. Borkowska, A. Zalewska, M. Marczewski, H. Marczewska and W. Wieczorek, J. Power Sources, 2006, 159, 443. 86. J. Syzdek, M. Armand, M. Marcinek, A. Zalewska, G. Zukowska and W. Wieczorek, Electrochim. Acta, 2010, 55, 1314. 87. S. K. Das, S. S. Mandal and A. J. Bhattacharyya, Energy Environ. Sci., 2011, 4, 1391. 88. J. E. Weston and B. C. H. Steele, Solid State Ionics, 1982, 7, 81. 89. W.-K. Shin, S.-M. Park, Y.-S. Lee and D.-W. Kim, J. Electrochem. Soc., 2015, 162, A2628. 90. W.-K. Shin, J. H. Yoo, W. Choi, K. Y. Chung, S. S. Jang and D.-W. Kim, J. Mater. Chem. A, 2015, 3, 12163. 91. W.-K. Shin, J. Cho, A. G. Kannan, Y.-S. Lee and D.-W. Kim, Sci. Rep., 2016, 6, 26332. 92. D.-W. Kim, J.-M. Ko and J.-H. Chun, J. Power Sources, 2001, 93, 151. 93. J.-S. Oh, Y. Kang and D.-W. Kim, Electrochim. Acta, 2006, 52, 1567. 94. J. Evans, C. A. Vincent and P. G. Bruce, Polymer, 1987, 28, 2324. 95. Y.-C. Jung, M.-S. Park, D.-H. Kim, M. Ue, A. Eftekhari and D.-W. Kim, Sci. Rep., 2017, 7, 17482. 96. P. Zhang, L. C. Yang, L. L. Li, M. L. Ding, Y. P. Wu and R. Holze, J. Membr. Sci., 2011, 379, 80. 97. S.-R. Park, Y.-C. Jung, W.-K. Shin, K. H. Ahn, C. H. Lee and D.-W. Kim, J. Membr. Sci., 2017, 527, 129.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00102

Gel Polymer Electrolytes

129

98. X. Wang, C. Gong, D. He, Z. Xue, C. Chen, Y. Liao and X. Xie, J. Membr. Sci., 2014, 454, 298. 99. L. Kong, H. Zhan, Y. Li and Y. Zhou, Electrochem. Commun., 2007, 9, 2557. 100. W. Kim, J.-J. Cho, Y. Kang and D.-W. Kim, J. Power Sources, 2008, 178, 837. 101. J.-A. Choi, Y. Kang, H. Shim, D. W. Kim, H.-K. Song and D.-W. Kim, J. Power Sources, 2009, 189, 809. 102. C. Gerbaldi, J. R. Nair, G. Meligrana, R. Bongiovanni, S. Bodoardo and N. Penazzi, Electrochim. Acta, 2010, 55, 1460. 103. D. Zhou, L.-Z. Fan, H. Fan and Q. Shi, Electrochim. Acta, 2013, 89, 334. 104. J. A. Choi, Y. Kang and D.-W. Kim, Electrochim. Acta, 2013, 89, 359. 105. K. H. Lee, K. H. Kim and H. S. Lim, J. Electrochem. Soc., 2001, 148, A1148. 106. Y.-W. Lee, W.-K. Shin and D.-W. Kim, Solid State Ionics, 2014, 255, 6.

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

Chapter 6

Liquid Non-aqueous Electrolytes for High Voltage Lithium Ion Batteries Lidan Xing* and Weishan Li* School of Chemistry and Environment, South China Normal University, Guangzhou 510006, China *E-mail: [email protected], [email protected]

6.1  Introduction Since lithium-ion batteries (LIBs) were introduced for commercial use about a decade ago, they have quickly reshaped our daily life.1 The high energy density and power density of LIBs over traditional rechargeable batteries make them a promising power source for most of today's portable electronic devices, including laptops, mobile phones, portable power tools, etc. However, LIBs still require improvement in the energy and power density, cycle life and safety to extend their medium- and large-scale applications such as pure-electric vehicles. Increasing the specific capacity and operating potential of cathode materials is one of the key approaches to promote the energy and power density of LIBs.2 Therefore, great effort has been made by both the academic and industrial communities to design and optimize novel cathode materials with higher capacity and operating potential. The successfully proposed novel cathode materials mainly include nickel-rich layered oxides

  Future Lithium-ion Batteries Edited by Ali Eftekhari © The Royal Society of Chemistry 2019 Published by the Royal Society of Chemistry, www.rsc.org

130

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

Liquid Non-aqueous Electrolytes for High Voltage Lithium Ion Batteries

131

(LiNi1−xMxO2, M = Co, Mn and Al), lithium-rich layered oxides (Li1+xM1−xO2, M = Mn, Ni, Co, etc.) and high-voltage spinel oxides (LiNi0.5Mn1.5O4).3–6 Unfortunately, practical application of these high voltage cathode materials in large-scale LIBs is greatly hindered by their poor cyclic stability at high potential. It has been demonstrated that the continuous electrolyte oxidation is largely responsible for the capacity fading of high voltage LIBs. The commercialized carbonate-based electrolyte oxidizes when the cell operating voltage is over 4.5 V, leading to a continuous increase in the interfacial reaction resistance and generation of harmful by-products, especially HF, which accelerates the dissolution of transition metal ions from cathode materials.7–9 Increasing the interfacial stability of high voltage cathode/electrolyte is critical for the practical application of high voltage LIBs, which can be achieved by element doping, electrode surface coating, solvent substituting and film-forming electrolyte additives.10–16 In this chapter, we focus on understanding the electrolyte oxidation decomposition mechanism and recent research progress on developing novel high voltage electrolytes (not including solid state electrolytes).

6.2  E  lectrolyte Component of the High Voltage Electrode/Electrolyte Interphase It has been demonstrated that the redox stability of electrolytes (containing lithium salt and mix-solvents) is not simply determined by the lowest oxidation/reduction stability solvent, but the stability of the component that sits next to the electrode surface. Moreover, the composition of the interfacial electrolyte near the electrode depends strongly on the electrode.17,18 Vatamanu et al. studied firstly the electrolyte (the electrolyte consisted of 114 ethylene carbonate (EC) molecules, 256 dimethyl carbonate (DMC) molecules, 31 PF6− ions, and 31 Li+ ions, which approximates the most commonly used electrolyte 1 M LiPF6/EC : DMC = 3 : 7) component near a graphite electrode (within 6.0 Å from the electrode surface) as a function of electrode potential using molecular dynamic (MD) simulation.17 As presented in Figure 6.1, upon increasing the electrode potential (charging process), the less polar DMC solvent is partially replaced in the interfacial electrolyte layer by the more polar ethylene carbonate (EC) solvent. And at higher potential, the surface is richer in the minority EC than in DMC. This result well explains that the detected electrolyte oxidation decomposition products are mainly from EC instead of DMC, while the former shows higher oxidation stability than the latter.19–21 Sulfone-based electrolytes have attracted great attention due to their high oxidation stability compared to conventional carbonates.22–24 The interfacial electrolyte component of sulfolane (TMS)/DMC mixed-solvent doped with LiPF6 salt was also investigated subsequently.18 TMS shows higher polarity and stronger binding energy with Li+ than that of DMC, resulting in a higher proportion of TMS molecules on the negative electrode surface as the electrode potential becomes more negative. Interestingly, at the positive

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

132

Chapter 6

Figure 6.1  Chemical  structure of (a) EC, (b) dimethyl carbonate (DMC) and (c) PF6−. A snapshot of the 2D periodic simulation cell is shown in (d). The oxygen lone-pair on EC and DMC was modelled by placing off-atom centred charges as shown in red in (a) and (b). (e) Cumulated density of electrolyte species in the interfacial layer, i.e. within 6.0 Å of the surface as a function of the electrode potential. Reproduced from ref. 17 with permission from American Chemical Society, Copyright 2012.

Figure 6.2  Normalized  atomic density profiles of carbonyl oxygen of DMC and

TMS near the positive electrode surface (+2.4 V) (a), and a snapshot of the interfacial electrolyte structure positive electrode (b). Reproduced from ref. 18 with permission from American Chemical Society, Copyright 2012.

electrode surface, the ratio of TMS/DMC was found to be similar to the bulk components, which is different to the observation of the EC/DMC system.17 Molecular structure results show that the DMC carbonyl oxygen atom (=O) does not approach the positive electrode surface as close as TMS, as shown in Figure 6.2, and hence DMC solvents are less likely to be oxidized than TMS. On the other hand, in TMS/DMC/LiPF6 electrolyte, DMC is located approximately 0.8 Å further away from the positive electrode than in EC/DMC/LiPF6 electrolyte, which further indicates that it might be more difficult to oxidize

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

Liquid Non-aqueous Electrolytes for High Voltage Lithium Ion Batteries

133

DMC in the TMS-based electrolytes. This trend is in line with experimental observations that show that electrolytes containing TMS and less oxidation stability DMC solvent mixed electrolyte still show high oxidative stability, and that they are only slightly worse than the TMS-based electrolytes.25 The change in the interfacial electrolyte component as a function of electrode potential obtained from the MD simulation allows us to predict the electrochemical stability of the electrolyte, and more importantly, to design an ideal interfacial electrolyte layer with high reaction inertness, or to generate a high stability solid electrolyte interphase (SEI) film and cathode electrolyte interphase (CEI) film on the anode and cathode surfaces, respectively.

6.3  C  ritical Role of the Anion in the Interphasal Stability of a High Voltage Cathode/Electrolyte The most commonly used electrolyte for LIBs consists of lithium salt, such as LiPF6 in a cyclic and linear carbonate mixed solvent. The thermal and moisture stability of the PF6− anion has been extensively studied experimentally, while the role of the anion in the oxidation stability has received less attention.26–28 Recently, more and more references appear to reveal the critical role of anions in affecting the electrolyte oxidation stability. Arakawa and Kanamura demonstrated that the oxidation decomposition rate of propylene carbonate (PC)-based electrolyte greatly depends on the adopted lithium salts, indicating that salt anions are indeed involved in electrolyte oxidation.29,30 Moreover, MD simulation results show that apart from more polar solvent, a higher accumulation of PF6− anions on the positively charged surfaces can also be observed,17,18 indicating the possibility of anions participating in the oxidation reaction with the neighbouring polar solvent. To understand the influence of anions on the oxidation stability and decomposition of the solvent, the oxidation of PC solvent with and without PF6− and ClO4− anions has been investigated with the density functional theory (DFT) calculation.31 The calculation results show that the presence of anions significantly reduces PC oxidation stability, stabilizes the PC anion oxidation decomposition products, and changes the order of the oxidation decomposition paths. The calculated oxidation potential and decomposition energy barrier of the PC–ClO4− complex is lower than that of PC–PF6−, revealing that electrolyte containing LiClO4 salt shows lower onset oxidation potential and higher decomposition rate, which is indeed in good agreement with the experimental observation.29 Importantly, the decrease in oxidation potential of the PC–PF6− complex in comparison with isolate PC is attributed to its deprotonation and HF formation. HF and PF5 are widely believed to be generated only by the hydrolytic reaction of LiPF6 with a trace amount of H2O remaining in the electrolyte.7 Herein, we provide evidence that HF and PF5 can be also generated during the initial oxidation reaction of PC–PF6− complex without the existence of H2O. A similar result can be found from

View Online

Chapter 6

134

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

−

−

−

oxidation of the EC anion (including PF6 , BF4 and ClO4 ) complex reported in Borodin's work and our following papers.32,33 A detected increase in HF content after electrolyte storage at high voltage further confirms our DFT calculation result.34,35 Sulfone-based electrolytes have attracted a great attention due to their high oxidation stability compared to conventional carbonates. However, our DFT calculation results indicate that the oxidation stability of isolate sulfones (one molecular) is lower than that of isolate carbonate.36 This contradiction was found to be caused by the influence of anions and their neighbouring solvent. Specifically, the investigated sulfones, including 3,3,3-trifluoropropylmethyl sulfone (FPMS), ethyl-iso-propyl sulfone (EiBS), methoxyethyl methyl sulfone (MEMS), ethyl methoxyethyl sulfone (EMES), ethylmeth-oxyethoxy-ethyl sulfone (EMEES), 1-fluoro-2-(methyl-sulfonyl) benzene (FS) and ethylmethyl sulfone (EMS), showed surprisingly high oxidation stability toward the presence of anions and neighbouring solvents, as shown in Figure 6.3, resulting in overall higher calculated oxidation potentials than those of carbonates. While the other investigated sulfones, including sulfolane (SL), trimethylene sulfone (TriMS), 1-methyltrimethylene sulfone (MTS) and ethyl-iso-propyl sulfone (EiPS), behaved similarly with carbonate solvent, showing decreased oxidation stability toward the presence of anions and co-solvent. It can be noted from Figure 6.3 that the

Figure 6.3  Calculated  oxidation potentials (V vs. Li+/Li) of the investigated isolate solvents, anions and clusters. Reproduced from ref. 36 with permission from American Chemical Society, Copyright 2013.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

Liquid Non-aqueous Electrolytes for High Voltage Lithium Ion Batteries

135

calculated oxidation potential of SL, TriMS, MTS and EiPS is lower than that of carbonate solvents, indicating that the experimentally detected high oxidation stability of these sulfone-based electrolytes should be ascribed to CEI film generation. Indeed, in our subsequent work, we confirmed the oxidation and CEI film-forming generation of SL-based electrolytes.37 Recently, Xu et al. also indicated a generated protective CEI film on a high voltage LiNi0.5Mn1.5O4 cathode surface in SL-based electrolyte, which greatly improved the cyclic stability of the LiNi0.5Mn1.5O4 electrode, as shown in Figure 6.4.38

Figure 6.4  Cryo-TEM  images and EELS spectra of LNMO particle CEI after the 50th

discharge with 1.2 M LiPF6 (EC/EMC 3 : 7) (a), (c) and 3 M LiFSI-SL (b), (d). XPS of the pristine cathode surface (e), Gen-II CEI (f), and 3.25 M LiFSI-SL CEI (g) after the 40th discharge. Reproduced from ref. 38 with permission from Elsevier, Copyright 2018.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

136

Chapter 6

The influence of anions on nitriles was also investigated in our recent work.39 Nitriles have been widely used in high voltage LIBs as solvent and electrolyte additive because of their unique ability in stabilizing electrolytes against oxidation at high voltages.40 The experimentally detected anodic stability of nitriles-containing electrolytes at high voltage had been ascribed to the preferential chemisorbed of monolayer nitrile molecules on the transition-metal (TM) oxide surface, which generates a layer of (– C=N–TM) complexes that physically expels carbonate solvents from direct contact with the highly charged electrode surface.41–43 While in our recent work, we overturned this belief based on calculation and experimental results. As shown in Figure 6.5, in comparison with the EC, DMC and ethyl methyl carbonate (EMC) carbonate molecules, the calculated oxidation

Figure 6.5  Calculated  oxidation potential of solvents, solvent-Li+, solvent-PF6−, sol-

vent-Li+-PF6−, solvent-SN, and solvent-Li+-PF6−-SN complexes (a); snapshots of Li/soaked Co3O4 and Li/prescanned Co3O4 V-type cells before (b), (e) and after (c), (f) LSV scanning, together with the N 1s XPS spectra of soaked Co3O4 (d) and prescanned Co3O4 (g) after LSV scanning in the baseline electrolyte. Reproduced from ref. 39 with permission from American Chemical Society, Copyright 2017.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

Liquid Non-aqueous Electrolytes for High Voltage Lithium Ion Batteries

137

potential of isolate succinonitrile (SN) molecules is indeed higher. However, in the presence of salt anion and cosolvent, the oxidation stability of SN decreased dramatically and became even lower than those carbonate solvents. This conflicts with all the reports that claimed higher anodic stability of SN than carbonates. Importantly, the preferential oxidation of SN in SN-containing electrolyte has been confirmed by our electrochemical test and physical characterization, see Figure 6.5. Therefore, the high oxidation stability of nitrile-containing electrolytes is ascribed to the created protective (–C=N)-containing CEI film.39 It is worth mentioning that, different from carbonates, the observed HF-free oxidation process of the SN–PF6− complex might be another key for the improved cyclic stability of SN-containing electrolyte with high voltage cathodes. This mechanistic correction of high anodic stability nitrile solvent/additive would be of high significance in guiding the design of new electrolytes and interphases for future battery chemistries. Recently, it has been proved that salt anions not only affect the interphasal stability of high voltage cathodes, but also the stability of graphite/ electrolyte.44 The interphasal property of graphite/electrolyte is very sensitive to electrolyte components. Two interphasal extremities induced by only one methyl substituent of solvent is the most representative example of this sensitive interphase, known as the “EC–PC disparity”. EC forms an almost ideal SEI film on the graphitic anode, thus becoming the indispensable solvent in all LIBs manufactured today, while PC does not form any protective interphase, leading to catastrophic exfoliation of the graphitic structure.45,46 Combining DFT calculation and experiments, we identified the partial desolvation of the solvated Li+ at graphite edge sites as a critical step.44 Specifically, during initial charging process, for EC solvent, the partial desolvation of the solvated Li+ prefers to lose EC solvent and keep PF6− in its primary Li+ solvation sheath. The reduction decomposition products of the generated PF6−-containing Li+ solvation sheath creates a LiF-rich protective SEI film on graphite electrode surface. By contrast, for PC solvent, the anion-free solvation structure has a higher chance of occurring during the initial desolvation of the solvated Li+, which would predominantly lead to ingredients from the solvent reductive decomposition with LiF being a relative minority. In other words, the key difference between EC- and PC-based Li+ solvation sheaths is how they generate the precursor to co-intercalation at graphite; the former apparently tends to form a ternary graphite intercalation compound (GIC) with higher anion population than the latter. These different GIC intermediates naturally lead to different surface products, and what differentiates the two extremes of interphasal chemistries is the LiF content in each case. Increasing the salt concentration of LiPF6/LiTFSI would improve the F-containing anion population on the graphite surface in PC-based electrolytes, leading to a generated LiFrich SEI film and high interphasal stability of graphite/PC-based electrolyte. The detected LiF content on a cycled graphite electrode surface with various

View Online

Chapter 6

138

Table 6.1  LiF  content on the surface of a cycled graphite anode. Reproduced from

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

ref. 44 with permission from American Chemical Society, Copyright 2018.

Pristine Sample graphite

1.0 M LiPF6 EC

1.2 M LiPF6 EC : EMC

1.0 M LiPF6 PC

3.25 M LiPF6 PC

3.5 M LiFSI PC

LiF %a 0 PVdF % 100

67.11 32.89

65.47 34.53

54.53 45.47

68.94 31.06

70.14 29.86

a

Relative content vs. polyvinylidene fluoride (PvDF), which is used as binder in the graphite anode, and serves as a reference for the LiF content estimation.

electrolytes can be found in Table 6.1. In this sense, the reported “abnormal interphasal behaviour” observed at ultrahigh salt concentrations was rooted in the changed Li+ solvation structure.47–50

6.4  Electrolytes for High Voltage Cathode Materials EC-linear alkyl carbonate mixed-solvent doped with LiPF6, the most common used electrolytes, have dominated the electrolyte market of 4 V class LIBs since their first commercialization. However, these carbonated-based electrolytes oxidize when the operating voltage is beyond 4.5 V, and therefore show difficulties in satisfying the requirements of next-generation LIBs in terms of both cyclic stability and safety. Great efforts have been made to develop novel electrolytes for high voltage cathode materials, mainly including searching for a novel high anodic solvent/lithium salt to substitute ­carbonate solvents/LiPF6, increasing the salt concentration and application of film-forming additives.

6.4.1  Salt/Solvent Substitution Searching for a high anodic stability salt and a solvent to replace carbonate-based electrolytes and LiPF6 could radically solve the instability of the high voltage cathode/electrolyte interphase. However, the practical application of these novel electrolytes in LIBs takes longer since it must take into account the compatibility with the graphite anode and the Al current collector, and the influence on interphasal reaction kinetics, safety, cost etc. Interestingly, high salt concentration electrolytes drew great attention in recent investigations, and most of the novel non-carbonate/LiPF6 electrolytes appear in a highly concentrated state, which indeed shows improved anodic stability than in a dilute state. For instance, the improved oxidation stability of glyme-lithiumbis(trifluoromethylsulfonyl)-amide (Li[TFSA]) electrolyte by increasing the salt concentration from [Li(glyme)1][TFSA] to [Li(glyme)x]-[TFSA] (x > 1) was reported by Watanabe et al., which was confirmed by their theoretical and experimental investigations, as presented in Figure 6.6.51 The enhanced oxidative stability was proposed due to the donation of lone pairs of ether oxygen atoms to

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

Liquid Non-aqueous Electrolytes for High Voltage Lithium Ion Batteries

139

Figure 6.6  Linear  sweep voltammograms of [Li(glyme)x][TFSA] (x = 1, 4, 8, and 20) at a scan rate of 1 mV s−1 at 30 °C. Each inset depicts an enlarged view of current density. Reproduced from ref. 51, http://dx.doi.org/10.1021/ ja203983r, with permission from American Chemical Society, Copyright 2011.

the Li+ cation, resulting in the highest occupied molecular orbital (HOMO) energy level lowering of a glyme molecule. Because of the enhancement of the oxidative stability of glyme caused by the equimolar complexation with Li[TFSA], the Li/LiCoO2 cell could be operated for more than 200 charge/discharge cycles in the cell voltage range 3.0–4.2 V regardless of the use of an ether-based electrolyte. For the oxidation instability of carbonate-based electrolytes, Dahn et al. first proposed that the problems of current carbonate-based electrolytes for high voltage Li-ion cells, including gas generation, salt consumption and impedance growth, originate from EC solvent which is thought to be an essential electrolyte component for Li-ion cells.52,53 They showed that LiPF6/EC-free-linear alkyl carbonate-based electrolytes with a small amount of film-forming additive, such as vinylene carbonate (VC), achieved higher capacity retention of high voltage LIBs. Accordingly, they suggested that further optimizing the linear alkyl carbonate electrolytes with appropriate co-additives may represent a viable path to the successful commercial utilization of high voltage LIBs. Without using EC as co-solvent and replacing LiPF6 with a high stable salt lithium bis(fluorosulfonyl)-amide (LiFSA), Wang et al. achieved stable and fast charge/discharge cycling of 5 V class LiNi0.5Mn1.5O4/graphite cells.54 The proposed “super-high” concentration LiFSA/DMC electrolyte with salt-to-solvent molar ratio of 1 : 1.1 shows great stability, with high voltage cathode, graphite anode and an Al current collector, therefore resulting in a significantly improved cyclic performance of high voltage LIBs, as shown in Figure 6.7. On the other hand, dissolution of transition metal ions from the high voltage cathode during cycling is also effectively suppressed (also presented in Figure 6.7). Owing to the much lower content of organic solvents in the concentrated solutions, the 1 : 1.1 LiFSA/DMC solution does not burn as fiercely as the commercial dilute electrolyte. The superior thermal stability

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

140

Chapter 6

Figure 6.7  Performance  of a high-voltage LiNi0.5Mn1.5O4|natural graphite battery.

Charge–discharge voltage curves of LiNi0.5Mn1.5O4|graphite full cells using (a) a commercial 1.0 mol dm−3 LiPF6/EC : DMC (1 : 1 by vol.) electrolyte and (b) a lab-made superconcentrated 1 : 1.1 LiFSA/DMC electrolyte at a C/5 rate and 40 °C. The curves of the 2nd, 10th, 50th and 100th cycles are shown. (c) Discharge capacity retention of the full cells at a C/5 rate. The inset shows EDS spectra on the graphite electrode surface (200 × 200 µm2 area) after eight-day cycling tests, which is equivalent to an operating time of 100 and 20 cycles for the battery using the commercial and superconcentrated electrolytes, respectively. (d) Discharge capacity of the full cell at various C-rates and 25 °C. All charge–discharge cycling tests were conducted with a cut-off voltage of 3.5–4.8 V. The 1 C-rate corresponds to 147 mAg−1 on the weight basis of the LiNi0.5Mn1.5O4 electrode. Reproduced from ref. 54, https://doi. org/10.1038/ncomms12032, under the terms of the CC BY 4.0 license, https://creativecommons.org/licenses/by/4.0/.

and flame-retardant ability of the concentrated electrolytes would also contribute to the remarkably improved safety properties as compared with the dilute electrolytes. Inaba et al. showed that, without replacing the LiPF6 lithium salt, highly concentrated PC-based electrolyte solutions are highly stable against oxidation at LiNi0.5Mn1.5O4 positive electrodes.55 In the electrolyte, the number of PC molecules in the Li+(PC)n solvation shell decreases with the increase of electrolyte concentration; Li+(PC)3, Li+(PC)2, and Li+(PC) will be formed at the salt concentration around 3.26, 4.90, and 9.79 mol kg−1, respectively. The

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

Liquid Non-aqueous Electrolytes for High Voltage Lithium Ion Batteries

141

calculated HOMO energy of PC was evaluated to be −15.7, −16.3 and −16.9 eV for Li+(PC)3, Li+(PC)2, and Li+(PC), respectively, indicating that the oxidative stability of PC should be enhanced by increasing the salt concentration. The enhanced stability of the concentrated electrolyte solutions was proved experimentally by linear sweep voltammetry and charge/discharge tests, which showed that low irreversible capacities and high Coulombic efficiencies of the LiNi0.5Mn1.5O4 electrode during cycling can be achieved in 4.27 mol kg−1 LiPF6/ PC electrolyte. Later on, they proposed that the high electrode polarization in concentrated LiPF6/PC electrolyte can be effectively decreased by replacing LiPF6 with LiBF4.56 In comparison with LiPF6, the solubility of LiBF4 in PC was higher and a nearly saturated electrolyte solution (7.25 mol kg−1) can be obtained. Charge/discharge tests revealed that the rate capability in 7.25 mol kg−1 LiBF4/PC was higher than that obtained with the nearly saturated 4.3 mol kg−1 LiPF6/PC, although the viscosity was higher, and the ionic conductivity was lower than those in the latter. The viscosity of highly concentrated LiBF4/ PC electrolyte solutions were then decreased by application of diluents.57 The investigated diluents included 1,1,2,2-tetrafluoroethyl-2,2,3,3-tetrafluoropropyl ether (HFE), 1,1,2,3,3,3-hexafluoropropyl-2,2,3,3-tetrafluoropropyl ether (FE2), 2,2,3,3,3-pentafluoropropyl-1,1,2,2-tetrafluoroethyl ether (T1216), 1,1,2,2-tetrafluoroethyl-2,2,2-trifluoroethyl ether (T3057), hexafluoroisopropyl methyl ether (T7301) and 1,1,2,2-tetrafluoroethyl ethyl ether (T5202), among which HFE was proposed as the most suitable diluent for highly concentrated LiBF4/PC electrolyte. 2.50 mol kg−1 LiBF4/PC+HFE (2 : 1 by volume) electrolyte with a low viscosity (51.7 mPa s) and a low PC/Li molar ratio (2.39) was chosen for 5 V LiNi0.5Mn1.5O4 positive electrodes, which showed good charge/discharge performance with low irreversible capacities compared to 2.50 mol kg−1 LiBF4/PC, and that the highly concentrated LiBF4/PC system can be diluted with HFE without losing the high stability against oxidation. However, the mechanism of improved oxidation stability of these high concentrated electrolytes is still unclear. The low Coulombic efficiency of cells during initial cycling in highly concentrated electrolytes may be ascribed to the oxidation of electrolytes, suggesting that the improved oxidation stability of high voltage electrolyte/electrolyte interphase may be due to the CEI film formation. In a similar manner, Dahn et al. showed that great cyclic stability of Li[Ni0.33Mn0.33Co0.33]O2/graphite and Li[Ni0.42Mn0.42Co0.16]O2/graphite pouch cells at high voltage (up to 4.4 V) and 40 °C can be achieved by increasing the salt concentration of the ethyl acetate (EA) sole solvent-based electrolyte.58 On the other hand, in comparison with that of carbonate-based electrolytes, the cells cycling in LiPF6 : LiFSi : EA (3 : 32 : 65 molar ratio) electrolyte generated fewer gaseous products during the formation cycle and showed lower impedance even when charged to 4.4 V for 500 h. The incompatibility of sulfone-based electrolytes with graphite electrode stand in the way of their application in high voltage LIBs. Recently, a carbonate-free, sulfone-based concentrated electrolyte for high voltage LIBs was proposed by Xu et al.,38 which created an effective SEI and CEI on the

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

142

Chapter 6

graphite anode and LiNi0.5Mn1.5O4 (LNMO) cathode surface, respectively, leading to a long-term operation of a high-voltage (4.85 V) LNMO/graphite full cell with a capacity retention of 70% after 1000 cycles, as presented in Figure 6.8. This high concentration LiFSI dissolved in SL created a LiFrich interphase by early-onset reduction of the salt anion and effectively suppressed solvent co-intercalation and subsequent graphite exfoliation, enabling unprecedented and highly reversible graphite cycling in a pure sulfone system. Moreover, high concentrated SL-based electrolytes exhibited fast ion conduction, stability over a wide temperature range and non-flammability. Obviously, this work presents a promising new direction toward unlocking the potential of next generation Li-ion battery electrodes. Fluorinated carbonate solvents were evaluated with high voltage LiNi0.5Mn1.5O4 cathode materials at elevated temperature by Zhang et al., which showed higher oxidation stability and higher reduction potential than that of carbonate solvents via theoretical calculations and

Figure 6.8  Galvanostatic  cycling capacity for Gen II and 3 m LiFSI-SL at 30 °C (a).

Figures (b)–(d) show the Al current collector corrosion on the active material side of the current collector for the 3 M LiFSI-SL cells with (b) pristine, (c) 50th discharge and (d) 1000th discharge. Reproduced from ref. 38 with permission from Elsevier, Copyright 2018.

View Online

Liquid Non-aqueous Electrolytes for High Voltage Lithium Ion Batteries

143

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

59

electrochemical characterizations. The substitution of EMC solvent with fluorinated linear carbonate (F-EMC) and EC with fluorinated cyclic carbonate (F-AEC) greatly improves the voltage limits of the electrolyte, resulting in enhanced cycling performance of the all-fluorinated electrolyte in Li4Ti5O12(LTO)/LiNi0.5Mn1.5O4(LNMO) cells at elevated temperatures, as shown in Figure 6.9. Importantly, fluorinated carbonate solvents also showed great compatibility with graphite anode by quick SEI formation and were non-flammable.60 The stability of lithium salt plays a critical role in the redox and thermostability of electrolyte. LiPF6 creates LiF-rich SEI film on graphite electrode surface; however, it decomposes into HF and PF5 at high voltage, which leads to a cascade of unwanted side reactions, and also decreases the conductivity of the electrolyte solution.31,36,44 LiBF4 has been considered as the most likely candidate to replace LiPF6 in high voltage electrolytes, as it shows higher thermal and moisture stability.61,62 Importantly, LiBF4-based electrolytes decreased the interfacial reaction resistance of high voltage cathode and passivated the Al current collector corrosion at high potential, which led to improved Coulombic efficiency of the cathode electrode.61 While Dahn et al. revealed that when the operating potential of cathode is beyond 4.4 V the LiBF4 would lose its advantage, and electrolyte oxidation, gas generation and impedance growth can be observed.63,64 The failure of LiBF4-based electrolytes beyond 4.4 V is ascribed to the lower passivation capability of LiBF4 on a graphite electrode, which results in increased charge transfer resistance at the negative electrode. They also showed that application of an effective SEI film-forming electrolyte additive, such as prop-1-ene-1,3-sultone (PES), would improve the interphasal stability of the graphite electrode, making the cells with LiBF4 perform just as well as cells with LiPF6 in aggressive 4.5 V “barn” cycling tests, and high voltage, high temperature storage tests.

Figure 6.9  Cycling  capacity retention of LTO/LNMO cells with baseline electrolyte Gen2 (1.2 M LiPF6: EC/EMC = 3/7) and fluorinated electrolytes E3, E5 and E6 at 55 °C. Reproduced from ref. 59 with permission from the Royal Society of Chemistry.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

144

Chapter 6

Three novel half borate lithium salts, including lithium difluoro-2-methyl-2-fluoromalonaoborate (LiDFMFMB), lithium difluoro-2-ethyl-2-fluoromalonaoborate (LiDFEFMB), and lithium difluoro-2-propyl-2-fluoro malonaoborate (LiDFPFMB), have been synthesized and evaluated for application in high voltage lithium ion batteries by Sun et al.65 The ionic conductivity of the half borate lithium salts follows the order of LiDFMFMB > LiDFEFMB > LiDFPFMB, which is consistent with the decreasing mobility of the anions due to the increase in the alkyl chain length of the anion. Moreover, the ionic conductivity of 1.0 M LiDFMFMB at 20 °C is 2.24 × 10−3 S cm−1, which is the same as those of LiBF4- and LiPF6-based electrolyte. Cyclic voltammetry and electrochemical floating tests were performed to investigate the redox stability and passivation capability of these new salts on a natural graphite anode and a high voltage spinel LiNi0.5Mn1.5O4 (LNMO) cathode, which indicated that three new salts preferentially reduced/oxidized on the graphite/LNMO surface when the electrode potential initially decreased/ increased, and importantly, created SEI and CEI film on the graphite and LNMO surfaces, respectively. Therefore, these new salt-based electrolytes exhibited good cycling stability with high coulombic efficiencies in both graphite and LNMO based half-cells and full cells. However, further investigation of these salt-based electrolytes on high voltage LIBs in comparison with that of LiPF6 is still needed.

6.4.2  Film-forming Electrolyte Additives It has been well accepted that the most economical approach to improving the interphasal stability of a high voltage cathode based on cost should be from an electrolyte perspective, whose effectiveness has already been proven with the various electrolyte additives widely employed by the Li-ion battery industry to modify the interphasal chemistry on the graphite anode.66 In the past decade, great progress has been made in research on the development of CEI film-forming additives for high voltage LIBs. In comparison with salt/ solvent substitution, application of film-forming electrolyte additives has the advantage of improving the interphasal stability of high voltage cathode/electrolyte without changing the physicochemical properties of the electrolyte, including ionic conductivity, viscosity and compatibility with the graphite anode. Some of the additives exhibit multifunction, such as simultaneously improving the interphasal stability of graphite/electrolyte, enhancing the thermostability of LiPF6, capturing HF and reducing the flammability of electrolytes.

6.4.2.1 Additives for High Voltage Spinel Oxides LiNi0.5Mn1.5O4 Spinel oxides LiNi0.5Mn1.5O4 were studied when researchers tried to partially substitute Mn with other metal (including Co, Ni, Mg, Cu etc.) to hinder the Jahn–Teller effect on spinel LiMn2O4, and they were first reported as 3 V cathode materials by Amine et al. in 1996. Later on, Dahn et al. found that

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

Liquid Non-aqueous Electrolytes for High Voltage Lithium Ion Batteries

145

LiNi0.5Mn1.5O4 can charge up to 4.7 V with a theoretical capacity of 146.7 mAh g−1, which is 20% and 30% higher than that of LiCoO2 and LiFePO4, respectively, and therefore became a potential candidate to be used in high energy density LIBs.67 However, as mentioned in Section 6.1, the development and application of high voltage LiNi0.5Mn1.5O4 is limited by the instability of the electrode/electrolyte. Xu et al. conducted further investigation and identified that an electrolyte additive based on a highly fluorinated phosphate ester structure tris(hexafluoro-iso-propyl)phosphate (HFiP) was able to stabilize carbonate-based electrolytes on 5 V class LiNi0.5Mn1.5O4 cathode surfaces.66 It was proposed that, similarly to SEI film additive, the sacrificial oxidation of HFiP occurs before oxidation of the bulk electrolyte components, whose oxidation products passivate the catalytic sites of the cathode surface, resulting in improved cyclic stability of the LiNi0.5Mn1.5O4/Li half-cell. In addition, they demonstrated that HFiP also provided excellent protective SEI chemistry on the graphitic anode. In our previous work, we proposed that tris-(pentafluorophenyl) phosphine (TPFPP) as a novel electrolyte additive could be used to improve the cyclic stability of high voltage LiNi0.5Mn1.5O4 cathodes.68 The electrochemical behaviours and surface chemistry of LiNi0.5Mn1.5O4 with TPFPP additive were investigated via cyclic voltammetry, chronoamperometry, charge–discharge test, X-ray photoelectron spectroscopy, and DFT calculations, which indicated that TPFPP additive oxidized preferentially to the base electrolyte (1 M LiPF6 EC/DMC/DEC = 1/1/1 by volume) to create a CEI film on the LiNi0.5Mn1.5O4 surface, resulting in the capacity retention of the LiNi0.5Mn1.5O4/Li half-cell improved from 70.3% to 85.0% after 55 cycles at 0.2 C between 3.5 and 4.9 V. The optimal content of TPFPP additive for LiNi0.5Mn1.5O4 was found to be 0.5%; a higher concentration would lead to inferior cyclic performance. The film-forming reaction of TPFPP also improved the cyclic stability of the LiMn2O4/Li half-cell charging up to 4.5 V.69 Interestingly, our subsequent work demonstrated that TPFPP could also decrease the flammability of the base electrolyte by decreasing the self-extinguishing time, and it was even better than the other reported phosphorous flame retardant compounds that have been used conventionally in lithium-ion batteries.69 Other novel effective CEI film-forming electrolyte additives for high voltage LiNi0.5Mn1.5O4 proposed by our group included 4-(trifluoromethyl)-benzonitrile (4-TB), prop-1-ene-1,3-sultone (PES), dimethyl phenylphosphonite (DMPP), phenyl trifluoromethyl sulfide (PTS), 1,1′-sulfonyldiimidazole (SDM)70 and trimethylsilylcyclopentadiene (SE).71 Specifically, using 0.5% 4-TB additive, LiNi0.5Mn1.5O4 delivered an initial capacity of 133 mAh g−1 and achieved a capacity retention of 91% after 300 cycles compared to that of 128 mAh g−1 and 75% by using a base electrolyte (1 M LiPF6 in EC/DMC).72 The preferential oxidation and film-forming reaction of 4-TB was confirmed by DFT calculation, linear sweep voltammetry, scanning electron microscopy, energy dispersive spectroscopy, Fourier transform infrared and inductively coupled plasma characterizations. Additionally, we found that apart from

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

146

Chapter 6

creating CEI film on the high voltage cathode surface, 4-TB could also coordinate with trace impurities such as trace alcohol and water that are detrimental to the thermal stability of electrolyte, and therefore, application of 4-TB additive can inhibit the formation of HF to a great extent by simultaneously inhibiting the electrochemical oxidation and thermal decompositions of electrolyte, as presented in Figure 6.10.73 Due to the greatly suppressed electrolyte oxidation, HF generation, Mn dissolution and Al current collector corrosion by application of 0.5% 4-TB, the capacity retention of the LiMn2O4/

Figure 6.10  Changes  in colour (a), HF content (b) and GC-MS spectrum (c) of the

electrolytes after storage at 55 °C for 72 h. Cyclability of LiMn2O4 electrodes at 0.5 C for the first three cycles and at 1 C for the subsequent cycles under 25 °C, using the electrolytes after (d) and before (e) storage. Reproduced from ref. 73 with permission from the Royal Society of Chemistry.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

Liquid Non-aqueous Electrolytes for High Voltage Lithium Ion Batteries

147

Li half-cell was improved from 19% to 69% after 450 cycles between 3.0 to 4.5 V under 55 °C. Further investigation on the influence of 4-TB on graphite anode and high voltage LIBs is needed. PES was firstly proposed in our work to be an effective SEI film-forming additive for the graphite anode,74 and has been extensively studied as a film-forming additive for both anode and cathode materials (including LiNi0.5Mn1.5O4, LiNi1/3Mn1/3Co1/3O2, LiCoO2, LiNi0.4Mn0.4Co0.2O2 and LiMn2O4) in the years that followed.75–80 Electrolyte oxidation and electrode structural destruction of high voltage LiNi0.5Mn1.5O4 cathode during cycling can be greatly suppressed by the CEI film created by preferential oxidation of PES.81 By application of 1% PES, the capacity retention of the LiNi0.5Mn1.5O4/ Li half-cell was improved from 49% to 90% after 400 cycles between 3.5 and 4.95 V. By the addition of PES additive into the baseline electrolyte for LiNi1/3Mn1/3Co1/3O2/graphite pouch cells, Dahn et al. found that increasing the amount of PES up to 6% by weight would increase the charge transfer resistance of the cell, but reduce the gas evolution during cycling.78 During 60 °C storage experiments, PES-containing cells produced 90% less gas than VC-containing cells. The cyclic performance of high voltage LiNi0.5Mn1.5O4 at 50 °C was improved by application of 0.5% dimethyl phenylphosphonite (DMPP) electrolyte additive with a capacity retention from 42% to 91% after 100 cycles.82 Further increasing the content of DMPP resulted in decreased cyclic stability, which was ascribed to the redundant oxidation and thicker film covering on the LiNi0.5Mn1.5O4 surface. It is worth mentioning that the oxidation decomposition reaction kinetics is faster than the other reported CEI film-forming additive, because the obvious oxidation current peak of DMPP can be easily identified on the Pt, LiNi0.5Mn1.5O4 and LiMn2O4 electrode surfaces during initial charging.82,83 The easy oxidation of an electrolyte additive is important for the formation of a protective cathode interphase, which would more efficiently suppress the oxidation of the baseline electrolyte. A thin and compact CEI film can be also observed on the LiMn2O4 surface when cycling with 0.5% DMPP by scanning electron microscopy, transmission electron microscopy and Fourier transition infrared spectroscopy characterizations, which dramatically improved the cyclic stability of LiMn2O4/Li half-cell at 55 °C between 3 and 4.5 V.83 The influence of the DMPP additive on the graphite anode in the initial charge/discharge is negligible, while the effect of long term cycling of graphite electrode with DMPP additive is unclear. The DFT calculation plays an important role in developing and optimizing new energy storage and conversion materials, and has been widely used to study the redox activity of carbonate-based electrolytes. A PTS additive was proposed as a novel CEI film-forming electrolyte additive by the DFT calculation, which showed higher oxidation activity and weaker interaction ability with Li+, ensuring its preferential enrichment and oxidation on the cathode surface.84 The preferential oxidation and CEI film-forming reaction of PTS on high voltage LiNi0.5Mn1.5O4 were confirmed by subsequent electrochemical investigations. The addition of 0.5% PTS improves the capacity retention of

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

148

Chapter 6

Figure 6.11  Cyclic  performance of LiNi0.5Mn1.5O4/Li half-cells in STD- and PTS-containing electrolytes at room (a) and (b) and elevated temperature (c). Reproduced from ref. 84 with permission from Elsevier, Copyright 2015.

LiNi0.5Mn1.5O4 from 65% to 84% after 450 cycles at room temperature, and from 64% to 95% after 100 cycles at the evaluated temperature, as presented in Figure 6.11. Electrolyte oxidation and transition metal ion dissolution of the LiNi0.5Mn1.5O4 electrode during cycling were effectively suppressed in the PTS-containing electrolyte. Recently, a fluorinated phosphazene derivative, ethoxy-(pentafluoro)-cyclotriphosphazene (PFN), was proposed as a novel electrolyte additive for improving the electrochemical performance and safety of the LiNi0.5Mn1.5O4 cathode by Guo et al., which not only generated a thin and uniform CEI film on the cathode surface, but also created a highly synergistic flame-retardant effect.85 The generated dense, uniform, and thin CEI film on the surface of the cathode material suppressed the decomposition of electrolyte and electrode corrosion, correspondingly protecting the cathode from structural destruction and improving the cyclic stability of LiNi0.5Mn1.5O4. They proposed that due to the combined structure of the non-flammable cyclophosphazene and fluorine, the PFN additive, and a low content of PFN (5 wt%) can almost completely extinguish burning electrolyte, leading to excellent safety performance for the lithium-ion battery.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

Liquid Non-aqueous Electrolytes for High Voltage Lithium Ion Batteries

149

The advantage of lithium salt as a CEI film-forming additive compared to organic molecular additive is that the functionalized anionic additives are electron rich species, which are more easily attracted to the cathode surface upon cell polarization, increasing the concentration of the reactive species on the cathode surface. Novel series of lithium alkyl trimethyl borates and lithium aryl trimethyl borates were prepared and investigated as film-forming additives for a high voltage LiNi0.5Mn1.5O4 cathode by Lucht et al.86 Among the investigated additives, including lithium 2-fluorophenol trimethyl borate (LFPTB), lithium 4-pyridyl trimethyl borate (LPTB), lithium trimethylsilyl trimethyl borate (LTSTB) and lithium propargyl trimethyl borate (LPrTB), LPTB showed the greatest capability to enhance the cyclic stability of the graphite/LiNi0.5Mn1.5O4 cell, which was demonstrated to create SEI and CEI film on the graphite anode and LiNi0.5Mn1.5O4 cathode surface, respectively. In addition, the CEI film on the cathode surface derived from LPTB inhibits Mn and Ni dissolution from the cathode and subsequent deposition on the anode. The inhibition of transition metal dissolution hinders damage to the anode SEI, which likely contributes to the improved electrochemical performance. Later on, they investigated and compared three different lithium borate electrolyte additives, including lithium bis(oxalato)borate (LiBOB), lithium 4-pyridyl trimethyl borate (LPTB), and lithium catechol dimethyl borate (LiCDMB) as a CEI film-forming additive for the LiNi0.5Mn1.5O4 cathode.87 Interestingly, they found that the thickness of the generated CEI film and calculated HOMO energy increases as a function of the additive structure LiCDMB > LPTB > LiBOB, suggesting that the properties of the cathode surface film can be tailored by the additive structure and reactivity.

6.4.2.2 Additives for High-voltage Nickel-rich Layered Oxides Nowadays, Ni-rich layered oxides are considered as one of the most promising cathode materials for the next generation LIBs due to their high reversible capacity and low cost. The reversible specific capacity of Ni-rich layered oxides shows high dependence on the transition-metal ratio and increases with nickel content from ∼160 mAh g−1 for LiNi1/3Mn1/3Co1/3O2 (NMC111) to ∼200 mAh g−1 for LiNi0.8Mn0.1Co0.1O2 (NMC811) charging to 4.3 V versus Li/ Li+. However, the high voltage interfacial stability of these Ni-rich layered oxides decreases with the increasing of Ni-content. Specifically, Jung et al. compared the stability of a carbonate-based electrolyte on NMC111, NMC622 (LiNi0.6Mn0.2Co0.2O2) and NMC811 by using online electrochemical mass spectrometry to determine the onset decomposition potential of the electrolyte.88 They found that the electrolyte decomposition (CO2 and CO evolution) onset potentials depend on the electrode material and increase in the order NMC811 < NMC111 ≈ NMC622, which would lead to severe gas generation, impedance increase and capacity fading of NMC811 in comparison with other Ni-rich layered oxides. As presented in Figure 6.12, two fundamentally different oxidation mechanisms, chemical and electrochemical electrolyte

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00130

150

Chapter 6

Figure 6.12  Proposed  gassing mechanisms in high-voltage lithium-ion cells

involving (a) electrochemical electrolyte oxidation proportional to the exposed surface area and (b) chemical electrolyte oxidation due to release of lattice oxygen from layered oxide cathodes and its reaction with electrolyte. Reproduced from ref. 88 with permission from American Chemical Society, Copyright 2017.

oxidation, were proposed simultaneously, showing that the intrinsic oxidative stability of the electrolyte toward electro-oxidation is rather high at room temperature (>5.0 V), while the lattice oxygen release from layered oxidebased cathode active materials causes substantial chemical electrolyte oxidation at lower potentials (90 >90

— —

0.1–100 12 000– 14 000 5–20 60–85%

Vehicles, Hybrid electric Load leveling, back-up vehicles bulk storage power supply Low cost, high High power recyclability density, of Pb environmentally friendly Requires Fast self-­ maintenance discharge

K-ion

Mg metal

RFB (redox flow)

Pb–acid

Consumer Load leveling, Load leveling, Load leveling, Load electronics, bulk storage bulk storage bulk storage leveling electric vehicles High energy High Low cost, high Low cost, High volumetric Higher density, quick efficiency, efficiency high capacity, depth of response, high efficiency no dendrite discharge, efficient energy formation long cycle cycles density (better safety) life Safety issues High cost, Lower energy Lower energy — Low energy at high scarcity of density than density density, operating lithium LIBs than LIBs low temperature and cobalt efficiency

View Online

From Lithium to Sodium and Potassium Batteries

187

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

8.3.3  Mg Batteries Magnesium metal is another interesting anode material due to its high abundance in the Earth's crust (Figure 8.1) and in sea water. Research of Mg batteries started in the late 1980s and they are extensively studied for post Li-ion batteries. The Mg/Mg2+ redox couple provides almost double the volumetric capacity (∼3800 mAh cm−3) than Li, owing to two-electron redox reactions of magnesium and similarity in its ionic radius of 0.76 Å with 0.72 Å of Li+.29 Moreover, no dendrite formation on Mg during charge makes the Mg batteries safer.30 However, its development is plagued due to the lack of a stable electrolyte with a wide voltage window, slow diffusion of Mg2+ ions in the solid-state electrodes and at the interface between the electrodes and electrolyte, and large charge transfer resistance due to resistive passivation interphase on the negative electrode.30 Development of suitable electrolytes and high energy density positive electrode materials are important for practical realization.

8.3.4  Na-ion Batteries SIBs operate on the same principle as LIBs except that the charge carriers are Na+ instead of Li+. More than 30 years ago, sodium intercalation materials were developed at the same time as that of lithium intercalation materials.31,32 In 1987, Showa Denko K. K., Hitachi, Ltd. and Allied-Signal, Inc. demonstrated a Na-ion full cell with NaxCoO2 and Pb-poly(p-phenylene) as positive and negative electrode materials, respectively.33,34 However, higher energy densities demonstrated by Li-insertion materials like LiCoO2 and subsequent commercialization of LIBs by Sony in 1991 almost ceased the research activity on SIBs. Moreover, lack of high performance carbon negative electrode materials hindered the progress of SIBs. High capacity hard carbon was demonstrated by Dahn et al. in 2001,35 and a report on the enhanced cycle stability of hard carbon//NaNi1/2Mn1/2O2 full cells was published by our group in 2011.36 These reports rejuvenated the attention on SIBs. As shown in Figure 8.3a, the number of scientific publications in the field of SIBs has increased rapidly in recent years and significant efforts have been devoted to developing electrode and electrolyte materials for high-voltage and high-capacity.10 Owing to higher abundance of Na element in the Earth's crust (Figure 8.1) and sea water, Na resources such as Na2CO3, Na2SO4 and NaCl are all less expensive than lithium ones and are suitable for low-cost battery production in large scale. However, the higher redox potential (0.23 V vs. Li/Li+),40 higher atomic mass (23 g mol−1) and larger ionic radius (1.02 Å) of Na+ than those of lithium lead to lower energy density of electrode materials. Passerni et al. investigated the cost advantage of replacing lithium in the positive electrode (LiMn2O4) and the electrolyte with sodium (Figure 8.4).18 Since the other components of battery remain the same, a reduction in the cost is possible for shifting to Na-containing materials. Moreover, an additive cost reduction is achieved by replacing the expensive copper current collector of the LIB negative electrode with aluminum since the latter does not alloy with Na at

View Online

Chapter 8

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

188

Figure 8.3  Number  of scientific publications per year for (a) SIBs and (b) KIBs. The data was updated on May 29, 2018.

low potentials, in contrast to LIBs. Thus, theoretically replacing Li with Na and Cu with Al results in a cost reduction of ∼12.5% based on the calculation of raw material cost. However, there are lot of challenges to achieve a SIB with similar energy density to commercial LIBs. Progress on the development of positive and negative electrode materials is provided in Section 8.4.

8.3.5  K-ion Batteries KIBs are another class of rechargeable batteries that employ potassium ions as charge carriers. The high abundance of K element in the Earth's crust and sea water enables cheaper resources similar to sodium ones; hence KIBs

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

From Lithium to Sodium and Potassium Batteries

189

Figure 8.4  Theoretical  cost comparison of a model Li-ion battery with a synthetic graphite (sG) anode and LiMn2O4 (LMO) cathode with a model sodium-ion battery containing sG and NaMnO2 (NMO) by replacing lithium with sodium and copper with aluminum (data derived from ref. 18).

are interesting and attractive for large-scale energy storage applications. Although Na battery materials have been researched for over 50 years,41 potassium was overlooked as a charge carrier until recently.42 This is mainly due to the large ionic radius of potassium ions and the dangerous nature of metallic potassium that is used in half cells for assessment of individual performance of positive/negative electrode materials. Matsuda et al. reported that the Stokes radius of solvated K+ ions in propylene carbonate (PC) is smaller than those of Li+ and Na+ ions (Figure 8.5a), and weaker solvation of K+ ions is expected due to its weaker Lewis acidity.43 Indeed, potassium bis(fluorosulfonyl)amide (KFSA) salt in PC exhibits higher ionic conductivity (Figure 8.5b) and lower viscosity (Figure 8.5c) than Li and Na salts. Another striking advantage of KIBs over SIBs is the higher voltage of the former, owing to lower redox potentials of K metal in aprotic solvents. The standard redox potentials of Li, Na, and K vs. SHE, are −3.04, −2.92, and −2.72 V, respectively, in aqueous solutions. However, in propylene carbonate-based electrolytes, E° of K/K+ becomes lower than that of Li/Li+ by ∼0.1 V.44 Since the redox potential of the carbon (graphite) negative electrode is close to that of K metal,45 a high operation voltage is expected for KIBs similar to that of LIBs. These aspects have contributed to the rapid growth in exploration and development of KIB materials in recent years (Figure 8.3b).10–12 Although the high operating potential and low cost of potassium resources are attractive for practical applications, we should note the larger ionic size and atomic weight of K compared to Li, which result in lower volumetric energy density of KIBs than LIBs. Therefore, KIBs may be advantageous mainly to stationary EES and not to portable devices. To realize practical KIBs, positive and negative electrode materials with high capacity and working voltage are vital. The research progress in positive and negative electrode materials for KIBs is described in Section 8.4.

View Online

Chapter 8

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

190

Figure 8.5  (a)  Comparison of Shannon's ionic radii and Stokes radii of Li+, Na+,

and K+ ions in propylene carbonate solvent. (b) Concentration dependence of ionic conductivity of LiFSA, NaFSA, and KFSA salts in PC, and (c) concentration dependence of viscosity of LiFSA, NaFSA, and KFSA salts in PC. Reproduced from ref. 10 with permission from John Wiley and Sons, © 2018 The Chemical Society of Japan & Wiley‐VCH Verlag GmbH & Co. KGaA, Weinheim.

8.4  Materials for Na-ion Batteries 8.4.1  Positive Electrode Materials 8.4.1.1 Layered 3d Transition-metal Oxides (NaxMeO2, where Me = 3d Transition Metal) Layered 3d transition-metal oxides such as LiCoO2, LiNi0.85Co0.1Al0.05O2 and LiNixMnyCo1−x−yO2 are the most significant positive electrode materials for practical LIBs.46–48 Similarly, Na-containing layered oxides have been extensively studied as positive electrode materials for SIBs owing to high theoretical capacities (e.g. 244 mAh g−1 for NaMnO2) and ease of synthesis. Structural and electrochemical properties of many Na layered oxides were reported in the 1980s.31,32 However, the renewed interest in SIBs has inspired the reinvestigation of previously known compositions, owing to the recent availability of high purity and battery grade electrolytes, and advanced characterization

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

From Lithium to Sodium and Potassium Batteries

191

tools. In addition, new compositions have been investigated to achieve high energy density and long cycle life.5–7 A detailed review of the electrochemical properties of these oxides can be found elsewhere.8,9 NaxMeO2 crystal structure is mainly built up of alternate MeO2 slabs (formed by edge-sharing of MeO6 octahedra) and Na layers.5 Oxygen packing of the MeO2 slabs along the c-axis in different ways leads to polycrystals, as shown in Figure 8.6a and b. Proposed by Delmas et al. in 1980, the structures are represented as O2-, O3-, P2-, P3-type and so on.49 The letter in the acronym indicates the coordination polyhedra of Na, namely, octahedral (O), prismatic (P), and tetrahedral (T), while the number specifies the number of MeO2 slabs in a hexagonal unit cell. The O3- and P2-type phases are commonly obtained for Na layered metal oxides which will be discussed below. 8.4.1.1.1  O3-NaMeO2.  Maintaining the Na and Me stoichiometry during synthesis would favor the O3- or O′3- (with structural distortion) type oxides (cystallographically α-NaFeO2 type). The Men+ and Na+ ions occupy distinct octahedral sites, forming a layered rock-salt type structure (space group, R-3m). Although O3-type LiMe′O2 can be directly crystalized via high-temperature solid-state reaction only for Me′ = Co, Ni, Cr, and V, the sodium coun­ terparts are possible for 3d metals from Sc to Ni in the periodic table and have been extensively studied for SIBs.8,9 Among them, Fe and Mn are particularly important in the context of elemental abundance, since SIBs are of interest for low-cost and large-scale energy storage. O3-NaFeO2 (α-NaFeO2) can be obtained by conventional solid-state synthesis, and was first electrochemically examined in 1980 by Takeda et al. in Li// NaFeO2 cells50 and in 2013 by Okada et al. in Na cells.51 As shown in Figure 8.6c, NaFeO2 demonstrates Fe3+/4+ redox accompanied by Na extraction/insertion at ∼3.3 V vs. Na/Na+ with negligible polarization, while a similar redox reaction is not exhibited by O3-LiFeO2.52 In NaFeO2, only ∼0.3 mol of Na can be reversibly extracted in the range of 2.5–3.4 V. Increasing the upper cut-off voltage would lead to migration of Fe into the interslab space, which blocks the Na diffusion pathway and reduces the reversible capacity.53 In addition, the instability of O3-NaFeO2 in air renders it inappropriate for practical applications. O′3-NaMnO2 with monoclinic structure is an attractive material owing to elemental abundance. Ma et al. reported a reversible capacity of 185 mAh g−1 with extraction of ∼0.8 Na in the voltage range of 2.0–3.8 V (Figure 8.6d).54 However, capacity fading was observed during cycling despite any substantial structural change. In addition, its average working voltage is lower than the Fe3+/4+ redox couple in NaFeO2. O3-NaCoO2 55 and O′3-NaNiO2 56 also deliver capacities of ∼140 mAh g−1 with stepwise voltage profiles, owing to several phase transitions during charge/discharge and in-plane Na/vacancy ordering. In addition, no migration of transition metal into the interslab space was observed for Ni and Co.57 Reversible phase transition

View Online

Chapter 8

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

192

Figure 8.6  Schematic  representation of different phases of layered transition

metal oxides (NaxMeO2), namely, (a) O3- and (b) P2-type structures. (c)– (h) Galvanostatic charge/discharge curves of different layered 3d transition-metal oxides: (c) O3-NaFeO2, (d) O′3-NaMnO2, (e) O3-Na[Fe1/2Co1/2] O2, (f) P′2-Na2/3CoO2, (g) P′2-Na2/3MnO2, and (h) P2-Na2/3Ni1/3Mn1/2Ti1/6O2. Parts (a)–(e), (g), and (h) reproduced from ref. 5 with permission from American Chemical Society, Copyright 2014. Part (f) reproduced from ref. 8 with permission from Cambridge University Press, Copyright 2014.

from O3- (or O′3-) to P3-phase is observed during charge/discharge for Cr, Ni and Co. The P3-phase provides open pathways for sodium diffusion with a lower diffusion barrier than the O3-type phase. Computational studies suggested that the P3-Na0.5MeO2 phase is stabilized for Me = Ni, Mn, and

View Online

From Lithium to Sodium and Potassium Batteries

193

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

58

Co. Combination of different transition metals sitting at the Me site in NaMeO2 can stabilize the P3-type phase, which results in better electrochemical performance. High capacities of >150 mAh g−1 are obtained for O3-Na[Fe1/2Co1/2]O2 (Figure 8.6e), O3-Na[Ni1/3Fe1/3Co1/3]O2, etc.59,60 In addition, a tiny substitution of Li in Me site results in better performance. For example, O3-Na[Li0.05(Ni0.25Fe0.25Mn0.5)0.95]O2 offers a high capacity of 180 mAh g−1 in Na cells.61 The O3-type layered oxides are sensitive to moisture and water and storing in air deteriorates the performance. Recently, O3-type oxides with dopants like Mg2+ and Cu2+ ions in the Me site such as O3-Na0.9[Fe0.30­ Mn0.48Cu0.22]O2, O3-Na[Li0.05Mn0.50Ni0.30Cu0.10Mg0.05]O2, and O3-Na[Ni0.45­ Cu0.05Mn0.4Ti0.1]O2 have been identified to exhibit enhanced water-resistant properties.62,63 In addition, large irreversible capacity is experienced in the O3-type oxides due to decomposition of the electrolyte on the particle surface, which contribute to poor cycling stability.64 Hence to mitigate the electrolyte decomposition is another important strategy to improve the cycle life. Another family of layered metal oxides named P2-type NaxMeO2 are also extensively studied as positive electrodes and will be discussed in the next section. 8.4.1.1.2  P2-type NaxMeO2.  Preserving an off-stoichiometry of sodium, i.e. Na/Me atomic ratio = ∼0.6–0.7, during the material synthesis would stabilize the Na deficient phase P2-NaxMeO2, where 0.5 ≤ x ≤ 0.8 (crystallographically Cr2AlC type65 with space group: P63/mmc), as drawn in Figure 8.6b. The large Na+ ions are stabilized in the trigonal prismatic sites while P2-type Li layered oxides are not favored due to the small ionic size of lithium. In contrast to O3-NaMeO2, the crystallization of the P2-type phase is possible only for V, Co and Mn as single metal oxides. The P2-type oxides are of great interest as positive electrodes since they exhibit higher working potential in the Na range ∼0.67 < x < 0 compared to O3-type oxides. In addition, the P2-type oxides exhibit better Na diffusion kinetics compared to the O3-phase due to the presence of an open pathway for Na migration with a low energy barrier.66 The electrochemical properties of P2-NaxCoO2 and P2-NaxMnO2 were first reported in 1980 and 1985, respectively.55,67 The P2-NaxCoO2 (Figure 8.6f) and P2-NaxMnO2 (Figure 8.6g) exhibit stepwise voltage profiles owing to several phase transitions during Na insertion/extraction. Large Na+ ions in the interslab space introduce greater Na–Na electrostatic repulsion and elongation of Me–O bonds in the slabs than that in LiMeO2. This leads to sodium-ion ordering in the interslab space and a similar ordering is not seen in LixCoO2 except for Li1/2CoO2 due to weak Li–Li interactions.68,69 Reversible capacities of ∼120 and ∼190 mAh g−1 were obtained for P2-NaxCoO2 and P2-NaxMnO2, respectively.70 Substituting half of the Mn for Fe results in good reversibility, smooth voltage profiles and high capacity of 190 mAh g−1; however, with a low average potential of 2.75 V.71 Ni substitution helps in enhancing the voltage (e.g., 3.63 V for P2-Na2/3Ni1/3Mn2/3O2).72 However, it

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

194

Chapter 8

undergoes phase transition to O2-type structure upon charging up to 4.5 V, accompanied by significant volume shrinkage of ∼23%, which results in poor capacity retention. Restricting the charge cut-off voltage to 3.8 V results in better cycling stability with a trade-off in reversible capacity, delivering only 80 mAh g−1. In a recent work, our group revealed that Al-substitution in P2-Na2/3Ni1/3Mn2/3O2 improved the capacity retention despite a significant volume shrinkage of ∼20%.9 Slight Al doping suppresses the surface degradation of the particles and helps to achieve good cycle and rate performance. In addition, partial substitution of Ti in the P2-oxide was adapted as another strategy to improve cycling stability. Ti-substituted P2-Na2/3Ni1/3Mn2/3O2 demonstrated a tremendous decrease in the volume shrinkage to ∼12%.73 In the case of Ti substitution, Na2/3Ni1/3Mn1/2Ti1/6O2 delivers a reversible capacity of 127 mAh g−1 with long cycle life (Figure 8.6h). Moreover, the voltage profile changes from step-wise to continuous sloping profile with an average potential of 3.7 V, indicating the suppression of in-plane Na+-vacancy ordering. An ideal full cell with P2-Na2/3Ni1/3Mn1/2Ti1/6O2 as a positive electrode material and hard carbon as a negative electrode material (delivering a capacity of ∼350 mAh g−1) will demonstrate >300 Wh kg−1, calculated on the basis of their half-cell data. Recent advancements in the energy density of the Na layered oxides suggest that they have potential to compete with commercial Li-ion full cells. However, the huge irreversible capacity loss of anode in the first cycle inevitably reduces the energy density of full cells. As it is difficult to directly synthesize the P2-phase with high sodium content, practical capacity in full cells are considerably lower. Na loss can be compensated by addition of sacrificial salts, such as NaN3 which irreversibly decomposes to Na and N2 gas.74 Further improvements in the energy density, cycle life, and stability of the layered metal oxides are still required for practical SIBs.

8.4.1.2 Polyanionic Compounds Since the successful demonstration of LiFePO4 as an LIB cathode,75 research on polyanionic compounds for secondary batteries has increased tremendously. A wide variety of polyanionic compounds including phosphates, sulfates, silicates, etc., are available as Li-, Na-, and K-ion hosts.75–77 Generally, they deliver lower specific capacity compared to oxides, owing to the high formula weight of the polyanion. Nevertheless, the energy density is compensated to some extent by higher working potentials due to inductive effect.78 In addition, they exhibit outstanding safety features owing to high structural stability, actually, oxygen release associated with structural decomposition of layered transition metal oxides can be avoided in case of polyanion, which is broadly known from studies on LiFePO4. NaFePO4 is one of the most studied polyanion compounds and it crystallizes in triphylite and maricite phases. Maricite is the thermodynamically stable phase; however, it lacks Na migration channels and is electrochemically

View Online

From Lithium to Sodium and Potassium Batteries

195

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

79

inactive. The triphylite type phase provides a one-dimensional Na migration pathway along the b-axis (Figure 8.7a), similar to triphylite LiFePO4 and is electrochemically active. The triphylite NaFePO4 prepared by Na/Li-ion exchange of LiFePO4 offers a reversible capacity of ∼120 mAh g−1 at 2.8 V (Figure 8.7b).79–83 Recently, electrochemical activity of the maricite phase was revisited using nanoparticles, which allows hopping of Na+ ions with a substantially low energy barrier to deliver high reversible capacity of ∼140 mAh g−1 for the initial cycle and high capacity retention (∼95%) after 200 cycles.84 Interestingly, alluaudite-type Na2Fe2(SO4)3 reported by Barpanda et al. exhibits Fe2+/3+redox reaction at a high voltage of at 3.8 V.85 The material demonstrates reversible sodium extraction/insertion and energy density of >540 Wh kg−1 in Na cells, which is higher than those of LiMn2O4 (430 Wh kg−1) and LiFePO4 (500 Wh kg−1) in Li cells.85 NASICON-type Na3V2(PO4)3 with fast Na-ion diffusion in the channels (Figure 8.7c) is another promising material for practical use owing to its working potential of 3.4 V and long cycle life.86–90 However, conventional synthesis results in poor capacity owing to long diffusion path and low electronic conductivity.91 Hence, most of the studies on this material focused on improving the electronic conductivity by carbon coating, preparation of graphene composite, and so on.86,90,92,93 A porous carbon/Na3V2(PO4)3 composite prepared by Saravanan et al. exhibited a high energy density of >400 Wh kg−1 and good rate capability (Figure 8.7d).86 In addition, it demonstrated a long cycle life of over 30 000 cycles at 40 C, with 50% capacity retention (Figure 8.7e). However, the toxicity and high cost of vanadium must be a concern. A tetragonal phase of Na3V2(PO4)2F3 exhibits higher capacity (130 mAh g−1), high voltage (∼3.75 V vs. Na/Na+) compared to NASICON-type Na3V2(PO4)3, good capacity retention, and high-power performance of 70 mAh g−1 at 5 C rate.94,95 In addition, pyrophosphates,96,97 mixed phosphate-pyrophosphates,98,99 fluorophosphates,100–105 and fluorosulfates85 have been studied as positive electrodes for SIBs.

8.4.1.3 Prussian Blue Analogues Prussian blue analogues (PBAs), represented by a general formula AxMe1[Me2(CN)6]·nH2O, are investigated as positive electrode materials for SIBs, owing to their simple liquid-state synthesis and availability of tunnels for Na migration.106 PBAs are also attractive positive electrode materials for KIBs and will be discussed in detail in Section 8.5.1.3. In 2012, Goodenough et al. reported the Na insertion in various K-based PBAs, KMe[Fe(CN)6], where Me = Fe, Mn, Ni, Cu, Co and Zn, in carbonate-based electrolytes.107 PBAs with different compositions have been widely investigated as positive electrode materials for SIBs.108–119 Recently, our group comparatively studied the electrochemical performance of Na-rich hexacyanoferrates (HCF) with Mn, Fe, Co, and Ni.120 The redox potential of the PBAs depends on the transition metal ions. Mn-HCF provided the highest voltage among them

View Online

Chapter 8

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

196

Figure 8.7  (a)  Crystal structure representation of triphylite NaFePO4, demonstrat-

ing one-dimensional Na diffusion along the b-axis; (b) electrochemical curve for the synthesis of NaFePO4 and Na0.7FePO4 in PITT mode. Reproduced from ref. 83 with permission from American Chemical Society, Copyright 2014. (c) Crystal structure representation of NASICON Na3V2(PO4)3. (d) and (e) Rate capability and long term cycling of c-Na3V2(PO4)3, respectively. Reproduced from ref. 86 with permission from John Wiley and Sons, Copyright © 2013 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim.

View Online

From Lithium to Sodium and Potassium Batteries

197

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

+

(∼3.3 V vs. Na/Na on average) with initial discharge capacity of 126 mAh g−1. However, the cycling stability of Mn-HCF was poor while Ni-HCF in which Ni is inactive in the potential window demonstrated a high capacity retention of 66 mAh g−1 over 50 cycles. Na storage properties of the PBAs are promising; however, they suffer from low volumetric capacity owing to their low density. In general, Na deficiency and [Fe(CN)6] defects formed during the synthesis of PBAs due to inclusion of water in the crystal structure, lead to inferior electrochemical performance. Therefore, controlled synthesis to eliminate interstitial water in PBAs is detrimental to improving their performance.106

8.4.2  Negative Electrode Materials 8.4.2.1 Carbonaceous Materials In commercial LIBs, graphite is employed as negative electrodes owing to low cost, high initial energy efficiency, and long cycle life. Lithium intercalation into graphite occurs at a low potential of ∼0.1 V vs. Li, which is close to that of E°(Li/Li+), ensuring high voltage and high energy density of full cells. It also offers a high capacity of 372 mAh g−1 owing to formation of a stage-1 compound, LiC6. Exhibiting a similar mechanism, K-intercalation into graphite leads to the formation of a stage-1 compound, KC8 (this will be discussed in detail in Section 8.5.2.1), offering a theoretical capacity of 279 mAh g−1.45 In contrast, graphite cannot be utilized as a negative electrode for SIBs due to the thermodynamic instability of stage-1 Na-GICs such as NaC6 or NaC8.121 Recently, co-intercalation of Na coordinated with glyme solvent into graphite has been demonstrated with capacities of over 100 mAh g−1,122–125 which is considerably lower than that of graphite in LIBs. On the other hand, disordered carbons, namely, graphitizable carbon (soft carbon), non-graphitizable carbon (hard carbon), and graphene-related materials can accommodate sodium to a greater extent.126 Soft carbons have a distorted structure with short range order of sp2-carbon sheets and micropores in bulk,127 and they can be prepared by pyrolysis of precursors such as petroleum coke,128 polyvinyl chloride (PVC),129 pitch,130 etc. In 1993, Deoff et al. demonstrated that soft carbon obtained from petroleum coke can provide capacities over 100 mAh g−1 in Na cells.128 Later, Dahn et al. achieved over 200 mAh g−1 for soft carbons derived from pitch.131 The voltage profiles of soft carbons exhibit a sloping voltage region between ∼1.2 and 0 V vs. Na/Na+.126 Recently, higher capacities were reported for soft carbons obtained from perylene tetracarboxylic dianhydride (PTCDA), cellulose nanocrystals (CNC) and so on.132–136 Nevertheless, they exhibit unsatisfactory reversible capacity or low initial coulombic efficiency, which hinder their use in full cells. Hard carbon, which is also a form of disordered carbon, is regarded as the best available carbonaceous material for SIB applications.126 The hard

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

198

Chapter 8

carbon also has a disordered carbon network with randomly oriented and curved carbon layers and micropores in bulk.127 Hard carbon was used in commercial LIBs in the 1990s because hard carbons synthesized at relatively low temperatures were found to deliver higher gravimetric capacity than that of graphite in Li cells.137 However, the volumetric capacity of hard carbon is lower compared to graphite due to defective stacking and low density. Hard carbon was first studied in Na cells by Stevens and Dahn in 2001, and delivered a high capacity of ∼300 mAh g−1.35 Our group first succeeded in demonstrating long-cycle life hard carbon in a Na cell by optimizing the electrolyte solution.36 Figure 8.7a and b illustrate the voltage profiles and cycle stability, respectively, of hard carbon prepared from sucrose.138 As shown in Figure 8.8a, the charge/discharge curves consist mainly two regions; a sloping voltage region in 1.0–0.1 V, similar to that of soft carbon, and the plateau region below ∼0.1 V.138 Although various structure models and reaction mechanisms have been proposed to explain

Figure 8.8  (a)  Voltage profiles and (b) capacity retention plots of hard carbon

derived from sucrose at 1300 °C. Reproduced from ref. 138 with permission from the Royal Society of Chemistry. (c) Comparison of voltage profiles of Si, Pb and Sn in Na cells using PAA binder. Reproduced from ref. 142 with permission from Elsevier, Copyright 2012. (d) Voltage profiles of red phosphorous electrode in a Na-cell in the voltage window of 0–0.8 V. Reproduced from ref. 143 with permission from John Wiley and Sons, © 2014 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

From Lithium to Sodium and Potassium Batteries

199

the Na storage mechanism in hard carbons, most of them agree on the fact that the hard carbons contain defects, short-range carbon layers, and micropores. According to the early model proposed by Stevens and Dahn, the sloping voltage can be attributed to Na insertion between the carbon layers, while pseudo Na-metal clustering in the micropores occurs in the plateau region.131 The wider interlayer spacing in hard carbon compared to graphite and the presence of micropores are possibly advantageous for Na storage. Recently, Na insertion into the sites close to defects have been proposed for the Na insertion in the sloping voltage region.126 To date, hard carbons are the best carbon-based electrode for SIBs. Since hard carbon can be prepared from different carbohydrates and biomass wastes,139–141 low-cost production on a large scale is possible by finding an appropriate precursor. As the source and carbonization temperatures influence the performance of hard carbons, optimization of synthesis conditions is important. In addition, electrolyte salt and solvents, binders and electrolyte additives influence the reversibility of sodium insertion/extraction into/from hard carbon. Consequently, further improvements in the initial coulombic efficiency, cycle stability, and rate capability of hard carbons are expected to enable them to be first-generation negative electrodes in practical SIBs.

8.4.2.2 Titanium Oxides and Phosphates Non-carbonaceous materials that reversibly accommodate alkali ions are of interest as negative electrode materials. Spinel-type lithium titanate, Li4Ti5O12 is one of the most studied negative electrode materials and is utilized in LIBs produced by TOSHIBA under the product name SCiB™. Li4Ti5O12 exhibits long cycle life and good rate performance due to negligible volume change during charge/discharge. It is called a ‘zero-strain’ material and there is no SEI formation above 1.0 vs. Li/Li+.144,145 When Li4Ti5O12 is subjected to Na-insertion, segregation of Li and Na phases happens with formation of Li7Ti5O12 and Na6LiTi5O12, delivering capacities of ∼160 mAh g−1.146 Na insertion occurs at a lower discharge voltage of 0.7 V vs. Na/Na+ compared to 1.55 V vs. Li/Li+ in Li cells. Li4Ti5O12 in Na cells shows large kinetic barrier for Na+ ion diffusion, which leads to insufficient rate capability compared to that in Li cells. A layered titanate, Na2Ti3O7 has been reported as a negative electrode material for SIBs.147 The Na2Ti3O7/carbon composite electrodes delivered a high capacity of 177 mAh g−1 at 0.1 C rate at a low voltage of 0.3 V vs. Na/Na+. However, a low first coulombic efficiency of ca. 45% is a practical issue. Recently, our group reported a lepidocrosite-like C-base centered Na0.9[Ti1.7Li0.3]O4, which exhibits a higher initial coulombic efficiency of ca. 80% by using polyacrylonitrile binder.148 In addition to the titanium oxides, a NASICON-type titanium phosphate, NaTi2(PO4)3 has been studied as a negative electrode material. First proposed by Delmas et al.,149 the material shows Na insertion at a flat potential of ∼2.1 V

View Online

Chapter 8

200 + 150

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

vs. Na/Na . Two moles of Na per formula unit can be reversibly inserted into the structure owing to Ti3+/Ti4+ redox couple, delivering capacities close to the theoretical one of 133 mAh g−1.151

8.4.2.3 Alloying Materials Na alloys or binary materials with group 14 and 15 elements, such as Si, Ge, Sn, Pb, P, As, Sb, and Bi are of interest as negative electrodes for SIBs, owing to their large capacities compared to hard carbon and titanium based mat­ erials.152 However, the alloying materials suffer from huge volume change during charge/discharge, which induces mechanical stress in the electrode, leading to formation of cracks and loss of electrical contact between the particles and with the current collector. Strategies to mitigate the capacity fade include optimization of morphology of particles, choice of binders and electrolyte additives. Though Si offers a high theoretical capacity of ∼3600 mAh g−1 in Li cells,153 much smaller capacities are delivered in Na cells (Figure 8.8c). In contrast, Sn provides a higher practical capacity of 847 mAh g−1in Na cells. Sn undergoes sodiation in multiple steps forming intermediate phases such as NaSn5, NaSn, Na9Sn4 and Na15Sn4.154 Capacity decay is a crucial issue due to the volume change, which could be improved with functional binders such as Na polyacrylate and Na carboxymethyl cellulose (NaCMC) and electrolyte additive like fluoroethylene carbonate (FEC), as shown in Figure 8.8c.142 Sb undergoes an alloying reaction with three sodium per mole forming Na3Sb alloy (660 mAh g−1 at ∼0.6 V), similar to LIBs. Initially, an amorphous phase of NaxSb was found that transforms into hexagonal Na3Sb at the end of sodiation as evidenced by XRD patterns during cycling by Darwiche et al.155 Upon desodiation, the crystalline Na3Sb was found to transform into amorphous Sb. Excellent cycling stability was achieved, even with micron sized particles, by using FEC as electrolyte additive and NaCMC as binder.155 Phosphorous reversibly undergoes a three-electron redox reaction with Na, forming Na3P and offers a theoretical capacity of ∼2600 mAh g−1. Among the allotropes of phosphorus, white P is unstable while red and black P are relatively stable. Red P and black P have been studied as negative electrode materials for SIBs.156–158 Figure 8.8d shows the charge/discharge profiles of a red phosphorous electrode in a Na cell in the voltage window of 0–0.8 V vs. Na/Na+. In contrast to Sn and Sb, phosphorous experiences small volume changes owing to the covalent nature of Na in Na3P. Despite these advantages, its utilization in practical cells is restricted due to safety issue of toxic PH3 gas, generated by hydrolysis of Na3P. Among the alloying based negative electrode materials for LIBs, Si provides very high capacity and hence Si–graphite composites are actively researched as next generation negative electrodes. Similarly, a mixture composite of hard carbon and an alloying material should be a possible solution for increasing the energy density of future SIBs.

View Online

From Lithium to Sodium and Potassium Batteries

201

8.5  Materials for Potassium-ion Batteries

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

8.5.1  Positive Electrode Materials 8.5.1.1 Layered Transition Metal Oxides Commercial LIBs employ layered transition metal oxides like LiCoO2, LiNi0.85­ Co0.1Al0.05O2, and LiNixMnyCo1−x−yO2.46–48 Moreover, promising Na layered transition metal oxides are emerging as SIB cathodes.5–7 From this viewpoint, it is worth understanding the K-intercalation behavior of K layered transition metal oxides. The structures of some of the oxides were known since the 1980s;159–162 however, their electrochemical properties were not reported until our report. Electrochemical properties of P2-type K0.41CoO2 and P3-type K2/3CoO2 were studied by our group163 and that of P2-K0.6CoO2 was investigated by Ceder et al.164 Voltage profiles of KxCoO2 show numerous plateaus arising from many phase transitions (Figure 8.9a). This is due to strong K+–K+ repulsive interactions, which induce K/vacancy ordering, and this repulsive interaction follows the order of K+ > Na+ > Li+ following the order of ionic size. Nonetheless, K-intercalation in both P2- and P3-KxCoO2 was reversible, delivering ∼60 mAh g−1 with good capacity retention and rate performance.163 Apart from the limited capacity, KxCoO2 suffers from a profound decrease in the K cell voltage (Figure 8.9a) with increasing K content, resulting in low energy density of KIBs. This is due to low Lewis acidity of K+ and elongated Co–O bonds. Vaalma et al. investigated birnessite K0.3MnO2 in K half cells, which delivered an initial capacity of ∼65 mAh g−1 and 57% capacity retention after 685 cycles.166 Increasing the charge cut-off voltage from 3.5 to 4.0 V helps in enlarging the capacity; however results in severe capacity fading during cycles. Wang et al. reported a higher capacity in interconnected K0.7Fe0.5Mn0.5O2 nanowires (Figure 8.9b), which exhibited capacities of 178 and 125 mAh g−1 for the 1st and the 45th cycles, respectively, at 20 mA g−1 in K half cells. In addition, full cells with soft carbon anodes demonstrated a reversible capacity of 119 mAh g−1.165

Figure 8.9  (a)  Comparison of voltage profiles of O2-LiCoO2, P2-Na2/3CoO2 and P2-K0.41CoO2 in Li, Na and K cells, respectively, and (b) voltage profile of K0.7Fe0.5Mn0.5O2 nanowires in a K half cell. Reproduced from ref. 165 with permission from American Chemical Society, Copyright 2015.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

202

Chapter 8

The lithium and sodium layered transition metal oxides provide high gravimetric and volumetric energy densities in LIBs and SIBs, respectively. In contrast, the potassium ones exhibit poor K storage in KIBs due to the large ionic size of K+ ions. Material innovation with focus on high voltage, smooth charge–discharge profiles and high capacities are indispensable in achieving high energy density. For these limitations, the major research focus on positive electrodes is on polyanionic compounds and potassium iron or manganese hexacyanoferrates, so-called Prussian blue analogues, which will be discussed in the next sections.

8.5.1.2 Polyanionic Compounds As discussed above, polyanionic compounds attract huge interest as positive electrode materials for KIBs owing to the open framework structure with high diffusivity of alkali-metal ion, excellent structural stability and associated thermal stability, and higher working potential. Reversible K storage in FeSO4F was explored during the early stage of KIB research by Tarascon's group.167 Prepared by electrochemical de-potassiation of KFeSO4F in Li half cells, the FeSO4F host undergoes reversible K intercalation in a K half cell at ∼3.6 V vs. K/K+, which is higher than voltages of 3.4 V vs. Li/Li+ and 3.3 V vs. Na/Na+ observed for the Li and Na cells, respectively. The KTiOPO4 (KTP)-type KFeSO4F reversibly extracts and accommodates potassium via multiple phase transitions. Interestingly, KVPO4F with a KTP-type structure was studied in 2016 by Fedotov et al. in Li-cells after initial electrochemical extraction of K-ions in a Li half cell.168 In 2017, our group demonstrated reversible K extraction/insertion in KVPO4F (Figure 8.10a), which exhibited a reversible capacity of ∼90 mAh g−1 at a high potential of ∼4.1 V vs. K/K+ (V3+/V4+ redox couple) with acceptable capacity retention.169 In the same work, an isostructural KVOPO4 phase demonstrated good cycling stability at ∼4 V (Figure 8.10b) with almost the same capacity as that of KVPO4F. Another phosphate, an amorphous FePO4, was reported as a positive electrode material for KIBs by Mathew et al.170 The material exhibited a high capacity of 150 mAh g−1 at ∼2.5 V and high cycling stability for over 50 cycles. The polyanionic compounds can provide high voltages (for instance, 3.6 V for K/KFeSO4F cell) compared to layered oxides. As observed for LiFePO4 in LIBs and Na3V2(PO4)3 in SIBs, long cycle life and high energy density can be achieved for KIBs with polyanion compounds.

8.5.1.3 Prussian Blue Analogues Prussian blue analogues (PBAs) containing alkali metal are exciting positive electrode materials for KIBs. They are represented by a general formula, AxMe1[Me2(CN)6]·nH2O; where Me = 3d transition metal and A = Li, Na, K, Mg, Ca, and so on. They have a three-dimensional open framework structure consisting of alternatively bridged Me1N6 and Me2C6 octahedra, in which large ions are diffusible.171 While most of the PBAs reported for KIBs are composed of Fe

View Online

203

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

From Lithium to Sodium and Potassium Batteries

Figure 8.10  Voltage  profiles of the first five cycles of (a) KVPO4F and (b) KVOPO4.

Reproduced from ref. 169 with permission from the Royal Society of Chemistry. (c) Voltage profiles of K1.75Mn[Fe(CN)6]0.93·0.16H2O for selected cycles in K half cells. (d) Voltage profiles of K1.86Fe­ [Fe(CN)6]0.89·0.15H2O. Reproduced from ref. 178 with permission form the Royal Society of Chemistry. (e) Comparison of discharge curves of K1.75Mn[Fe(CN)6]0.93·0.16H2O and K1.86Fe[Fe(CN)6]0.89·0.15H2O.

or Mn at the Me2 site and different metals such as Fe, Cu, Mn, Ni, Co, etc., at the Me1 site.172 Though PBAs can undergo reversible Li and Na insertion,106,120,173 K extraction/insertion in the PBAs have been identified to be more favorable than Na and Li.174–176 In general, water molecules tend to occupy the [Me2(CN)6] or vacant sites during precipitation synthesis in aqueous solutions. This results in non-stoichiometric composition of AxMe1[Me2(CN)6]y·nH2O, where y < 1177 or large water content as in A2Me1[Me2(CN)6]·2H2O. Consequently, the gravimetric capacity density decreases and care must be taken

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

204

Chapter 8

to minimize the vacancy/water content in the material. Fortunately, the water content of potassium hexacyanoferrates (HCF) is smaller than those of Li and Na ones.172,178 Non-aqueous K cells with PBAs as positive electrodes were first demonstrated by Eftekhari in 2004 using thin films of KFe[Fe(CN)6], which exhibited respectable cycling stability and a capacity of ∼70 mAh g−1 after 500 cycles.42 In 2017, our group reported KIB positive electrodes of K-rich HCFs containing Mn and Fe prepared by a simple precipitation method.178 K1.75Mn[Fe(CN)6]0.93·0.16H2O (Mn-HCF) demonstrated an initial discharge capacity of 137 mAh g−1 with high capacity retention (Figure 8.10c). The material exhibited two reduction peaks at 3.81 and 3.92 V vs. K/K+ in the cyclic voltammogram curves, which correspond to low-spin Fe2+/3+ and high-spin Mn2+/3+ redox couples, respectively. K1.86Fe[Fe(CN)6]0.89·0.15H2O (Fe-HCF) also exhibited two peaks, at 3.11 and 3.80 V vs. K/K+ (Figure 8.10d). The difference in the electron spin states of Fe in high spin FeN6 and low spin FeC6 environments leads to the different redox potentials. The discharge curves of Fe-HCF and Mn-HCF are compared in Figure 8.10e. Wu et al. studied the influence of different transition metals (Fe, Co, Ni, Cu) on the electrochemical properties of HCFs in K cells.172 While Fe-HCF delivered a capacity over 100 mAh g−1, the Co-, Ni-, and Cu-HCFs delivered less than 65 mAh g−1. Various positive electrodes materials such as layered oxides, polyanion compounds, and PBAs have been investigated for KIBs. Among them, the most successful K storage was demonstrated by the PBAs so far. The KMnHCF delivers a capacity of ∼130 mAh g−1 at high voltage (∼4 V), enabling high energy density. The presence of Earth abundant Fe and Mn is also advantageous in terms of cost and sustainability. However, their unsatisfactory low density (2.5 g cm−3) leads to low volumetric capacity, which may hinder their application for portable devices or transportation applications. However, they look very promising for stationary energy storage owing to the high gravimetric energy density and ease of synthesis.

8.5.2  Negative Electrode Materials 8.5.2.1 Carbonaceous Materials K-graphite intercalated compounds (K-GICs) have been known since 1932 179 and were prepared by non-electrochemical methods.180,181 Intercalation happens in different stages and the staging mechanism was reported in 1962.181 Stage-n represents n layers of graphene sheets separating successive intercalant layers. In Li cells, graphite exhibits a low potential of ∼0.1 V vs. Li/Li+ with formation of LiC6 as stage-1 compound, offering practical capacities of ∼360 mAh g−1. However, graphite cannot be employed in SIBs due to thermodynamic instability of Na-GICs.121 Interestingly, formation of K-GICs is thermodynamically feasible despite the larger ionic size of potassium.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

From Lithium to Sodium and Potassium Batteries

205

Electrochemical K-intercalation into graphite was demonstrated by Jian et al.183 and our group in 2015,45 and in our patent literature earlier. Li-intercalation into graphite is evidenced by three sharp cathodic peaks at 0.19, 0.09 and 0.05 V vs. Li/Li+ (Figure 8.11b) in a cyclic voltammogram, which indicate the formation of stage-3, 2 and 1 Li-GICs, respectively from graphite. Similarly, the K-intercalation is realized via three steps at 0.34, 0.23 and 0.1 V vs. K/ K+ (Figure 8.11a).182 Jian et al. proved the reversible formation of KC36 (stage3), KC24 (stage-2), and KC8 (stage-1) from ex situ XRD data (Figure 8.11c and d).183 The K-intercalation occurs at slightly higher potential versus plating/ stripping of K-metal compared to the formation of Li-GICs at potential closer to the Li-plating/stripping. This turns out to be an advantage since it mitigates the risk of K-metal plating at high rates. High capacities close to the theoretical capacity of 279 mAh g−1 can be achieved in K cells.45 The choice of binder and electrolyte affects the first cycle coulombic efficiency, which is crucial for practical batteries. In our studies, sodium polyacrylate (PANa) binder and 1 M KFSA solution in ethylene carbonate (EC) : diethyl carbonate (DEC) as electrolyte demonstrated a higher first cycle efficiency (Figure 8.11e) for graphite than the conventional poly(vinylidene fluoride) (PVDF) binder. In addition, better capacity retention and excellent rate performance of 230 mAh g−1 at a high current rate of 15 C (4.2 A g−1) were achieved (Figure 8.11f). The intrinsic large volume change of ∼60% during charge/discharge should affect the cycle life. Detailed studies are desired to realize the practical use of graphite in KIBs. In addition to graphite, non-graphitic carbons such as soft carbon,183 hard carbon,184 N-doped hard carbon,185 N-doped carbon nanotubes,186 polynanocrystalline graphite,187 graphene based materials,188,189 etc. have been studied as negative electrode materials to achieve higher capacity. Some of them exhibit capacities higher than the theoretical value of graphite. However, they suffer from lower initial coulombic efficiency, larger polarization, higher working potential leading to lower energy density of the full cell, or lower density than graphite.190 However, larger capacity and random orientation of graphitic domains are advantageous for higher performance KIBs.

8.5.2.2 Alloying Materials As discussed in the previous section, graphite can reversibly intercalate K with good rate capability and can be utilized in first-generation KIBs. However, its theoretical capacity is limited to 279 mAh g−1 based on KC8 formation and is lower than that of graphite in practical LIBs. Therefore, alloying materials may be interesting since they offer large capacities in LIBs. For instance, a graphite–silicon composite has been actively researched as the next generation negative electrode for LIBs. As in the case of alloys of Li and Na, alloys of K with group 14 and 15 elements encounter large volume change during charge/discharge, which induces electrical isolation of the particles in the electrode, leading to deterioration of electrochemical performance.

View Online

Chapter 8

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

206

Figure 8.11  Cyclic  voltammograms of graphite in (a) K and (b) Li cells, respectively.

Reproduced from ref. 182 with permission from John Wiley and Sons, © 2016 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim. (c) Charge– discharge profile for the first cycle of a graphite electrode in a K cell. (d) XRD patterns at marked states of charges in figure (c). Reproduced from ref. 183 with permission from American Chemical Society, Copyright 2015. (e) and (f) Voltage profiles of graphite using PANa binder and rate capability studies, respectively. Reproduced from ref. 45 with permission from Elsevier, Copyright 2015.

Silicon exhibits a high theoretical capacity of over 3600 mAh g−1 in Li cells owing to the formation of overlithiated Li15Si4 crystals in the fully lithiated state.191 By contrast, silicon can theoretically deliver only 955 mAh g−1 in K cells since calculations suggest that KSi is the most K-rich phase, which can be thermodynamically stable with consideration of the

View Online

From Lithium to Sodium and Potassium Batteries

207

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

192,193

K–Si phase diagram. However, no satisfactory performance of Si in K cells has been reported yet.10 Similar to silicon, Sn offers a lower capacity of 226 mAh g−1 in K cells,194 while >800 mAh g−1 is possible in Li and Na cells due to the formation of Li17Sn4 by lithiation195 and Na15+δSn4 by sodiation.196 Wang et al. investigated Sn nanoparticles as anode material and reported a capacity of 197 mAh g−1 for the first charge.197 The K–Sn phase was observed at the end of potassiation process using XRD. However, the material underwent severe capacity fading in ten cycles due to pulverisation of Sn nanoparticles. In contrast to silicon and tin in K cells, antimony exhibits three electron reactions due to the formation of K3Sb alloy, delivering a high theoretical capacity of 660 mAh g−1, which is similar to that of Sb in Li and Na cells. McCulloch et al. evidenced the formation of a cubic K3Sb phase by using ex situ XRD data of an Sb/C composite electrode tested in a K cell.198 It demonstrated reversible K (de)alloying in the initial ten cycles, delivering high capacities of 660 mAh g−1, which is close to the theoretical value. However, the capacity fading was inevitable after 10 cycles due to mechanical disintegration of the electrodes. Phosphorous is another interesting candidate in this category, which can chemically react with K, similar to Li and Na, offering a high theoretical capacity of ∼2600 mAh g−1 based on three electron reactions (Li3P and Na3P). Theoretically, K3P formation is also expected in K cells, however, no satisfactory electrode performance has been reported yet.194 Interestingly, a binary compound of Sn and P, Sn4P3–C obtained by ball-milling of Sn and P in carbon black, offered a high capacity of 385 mAh g−1 with a sloping voltage centered at ∼0.5 V.199 While the Sn–C and P–C composites failed to provide stable capacities, the Sn4P3 electrodes delivered a stable capacity for 40 cycles and capacity fading was observed after 40 cycles. Negative electrode materials that alloy with Li and Na provide capacities in excess than those of graphite and hard carbon. Due to the impressively large capacity of silicon in Li cells, it is in the verge of commercialization for LIBs. In K cells, Si and Sn electrochemically store less K content, hence, much lower capacity is delivered in KIBs than LIBs. Moreover, Sb and P have not demonstrated satisfying performance so far in K cells. Despite, higher capacities than graphite can be achieved through optimization of binder and particle morphology.

8.6  Outlook Renewable energy production from solar/wind energy can partially help to reduce CO2 emissions arising from the usage of fossil fuels. Hence, there is increase in the investment on renewables by many countries. According to a survey by Bloomberg NRF, the amount of stored energy is expected to reach six times the current value by 2030.200 The demand for zero-emission vehicles is also increasing rapidly and predicted to dominate the transportation sector by 2040.201 Owing to the high energy density of Li-ion

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

208

Chapter 8

batteries, adaption from portable devices to transportation applications has been realized. However, the high cost of batteries leads to the overall high price of electric vehicles, which prevents a quick transition to EVs from the conventional ICE vehicles. Although improvement in the manufacturing of Li-ion batteries can decrease the cost of batteries, the feasibility of a sustainable supply of Li resources is not guaranteed for the future of LIBs. Geopolitical issues on the uneven distribution of Li resources and mismatch in the supply and demand for production of Li resources may lead to price rise. In addition, dependence on expensive cobalt for positive electrodes can be mitigated by utilization of cobalt-lean or cobalt-free positive electrodes such as LiNi0.85Co0.1Al0.05O2 and LiNi0.8Mn0.1Co0.1O2. However, this is not an ideal way for large-scale use since the abundance of nickel is also limited, as shown in Figure 8.1. Hence, these Ni-rich cathode materials may be applicable for only for transportation applications, as portable devices need high energy density materials. There is no apparent cost benefit of replacing Li by Na or K in the current scenario as they occupy only a few percent of the total cost of batteries. However, huge demand for LIBs in the future would strain the supply of Li resources and would eventually lead to price rise, and Na- or K-ion batteries would have the cost advantage. In addition, the replacement of Cu foil used as anode current collector in LIBs by Al foil in NIBs/KIBs results in cost and weight reduction, which increases the energy density of batteries. As discussed above, sodium and potassium elements are abundant and can be useful as charge carriers for large-scale rechargeable batteries. However, Na-ion (or K-ion) batteries should be cost effective (low $ kWh−1 of stored energy) to be availed as complementary devices to store energy along with LIBs, which are already proven and successful rechargeable energy storage devices. To realize the advancement in SIBs or KIBs, the development of high energy density positive and negative electrode materials is needed, along with optimization of separators, binders, electrolytes, and electrolyte additives to deliver the full potential of the electrodes. Layered 3d transition metal oxides containing sodium have been investigated as positive electrode materials for around 50 years. However, in recent years, sophisticated characterization tools and high purity electrolyte salts and solvents are available and are successfully contributing to renewing performances and behaviors better than previous studies. Better understanding of these materials enables material design of higher-performance materials. The research progress of SIBs has been plagued since suitable materials such as active electrode materials and electrolytes are different from that of LIBs. Although graphite is not available in SIBs, hard carbon delivers high capacities of ∼350 mAh g−1 and more than 400 mAh g−1 has been reported recently.8 In addition, binder, electrolyte solvent and electrolyte additive have had a decisive influence on cycle life and energy efficiency. Their optimization for SIBs is important. The low operation potential and high capacity of hard

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

From Lithium to Sodium and Potassium Batteries

209

carbons enable SIBs possessing high energy density. Thus, further improvement in the capacity of hard carbons or a hard carbon-alloy composite will lead to higher energy density.18 In addition to the negative electrode materials, the progress of positive electrode materials is key to realize practical SIBs. Layered transition metal oxides with O3-type and P2-type structures are extensively investigated. In spite of the difference in the redox potential between Li/Li+ and Na/Na+ being only ∼0.3 V, the difference in the average working potential between Na and Li layered oxides is apparent (for example, ∼1 V between Li/LiCoO2 and Na/ NaCoO2 cells). As shown in Figure 8.12, SIBs with energy densities higher than that of graphite//LiMn2O4 are now possible with hard carbon as negative electrode (assuming a capacity of 350 mAh g−1 and Eave = 0.3 V vs. Na/Na+) and layered oxides as positive electrodes and are approaching that of graphite// LiCoO2 cells, but it is still challenging. The O3-type materials exhibit lower working potential of ∼3.1 V vs. Na/Na+, compared to 4 V vs. Li/Li+ of LiCoO2 in LIBs. To obtain energy densities similar to that of LIBs, a higher reversible capacity than 140 mAh g−1 of LiCoO2 (charged to 4.2 V) is highly required. The P2-type materials with slightly higher working potentials have been reported, as shown in Figure 8.12. Further improvements in the energy density, cycle life, and the stability of the materials in air are expected in the future.

Figure 8.12  Average  voltage (V) and energy density (Wh kg−1) versus gravimetric

capacity (mAh g−1) for selected O3- and P2-type oxides for SIBs. Hard carbon with reversible capacity of 350 mAh g−1 and Eave = 0.3 V versus Na is chosen as negative electrode material. Reproduced from ref. 9 with permission from John Wiley and Sons, © 2018 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

210

Chapter 8

KIBs are a relatively new technology but hold many promising factors such as high voltage, high ionic conductivity, and high rate performance due to the weaker coulombic interaction and lower Lewis acidity of K ions. Currently, graphite has shown promising performance as a negative electrode material with capacity close to the theoretical capacity of 279 mAh g−1 and a high rate performance of 230 mAh g−1 at 15 C rate.45 However, research progress of positive electrode materials is relatively slow according to our understanding. Only a few studies on layered 3d metal oxides are found, which suggest that the working potential of KxMeO2 is even lower than the Na counterparts. On the other hand, polyanion compounds and Prussian blue analogues deliver a high voltage of ∼4 V in K cells. For instance, K2Mn[Fe(CN)6] provides a half-cell energy density of ∼530 Wh kg−1, which is similar to that of LiCoO2. The cost advantages of replacing Li by Na and K is debatable as materials showing desired properties such as high energy density, long life and so forth are not developed and established in practical and industrial points of view. However, ease of Na/K production compared to the environmental impact of Li extraction should also be considered for large scale use in the future. Hence, lithium resources can be efficiently utilized for purposes such as portable and transportation purposes where high energy density is of the utmost priority and post-Li ion chemistries need to be developed for utilization in stationary storage EES.

Acknowledgement This study was in part supported by MEXT program “Elements Strategy Initiative to Form Core Research Center” (since 2012), the JST through A‐STEP program (AS2614056L and AS282S001d), and JSPS KAKENHI Grant Number JP16K14103, JP16H04225, and JP18K14327.

References 1. M. S. Whittingham, Chem. Rev., 2004, 104, 4271–4302. 2. D. Larcher and J. M. Tarascon, Nat. Chem., 2014, 7, 19. 3. S. E. Kesler, P. W. Gruber, P. A. Medina, G. A. Keoleian, M. P. Everson and T. J. Wallington, Ore Geol. Rev., 2012, 48, 55–69. 4. B. Nykvist and M. Nilsson, Nat. Clim. Change, 2015, 5, 329. 5. N. Yabuuchi, K. Kubota, M. Dahbi and S. Komaba, Chem. Rev., 2014, 114, 11636–11682. 6. J.-Y. Hwang, S.-T. Myung and Y.-K. Sun, Chem. Soc. Rev., 2017, 46, 3529–3614. 7. K. Hyungsub, K. Haegyeom, D. Zhang, L. M. Hwan, L. Kyungmi, Y. Gabin and K. Kisuk, Adv. Energy Mater., 2016, 6, 1600943. 8. K. Kubota, N. Yabuuchi, H. Yoshida, M. Dahbi and S. Komaba, MRS Bull., 2014, 39, 416–422.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

From Lithium to Sodium and Potassium Batteries

211

9. K. Kubota, S. Kumakura, Y. Yoda, K. Kuroki and S. Komaba, Adv. Energy Mater., 2018, 8, 1703415. 10. K. Kubota, M. Dahbi, T. Hosaka, S. Kumakura and S. Komaba, Chem. Rec., 2018, 18, 459–479. 11. X. Wu, D. P. Leonard and X. Ji, Chem. Mater., 2017, 29, 5031–5042. 12. K. Haegyeom, K. J. Chul, B. Matteo, S. Dong-Hwa, R.-G. Jorge and C. Gerbrand, Adv. Energy Mater., 2018, 8, 1702384. 13. R. S. Carmichael, Practical Handbook of Physical Properties of Rocks and Minerals, CRC Press, Boca Raton, FL, 1989. 14. United States Geological Survey, Mineral Commodity Summaries 2018, USGS Mineral Resources Program, 2018, https://minerals.usgs.gov/ minerals/pubs/mcs/2018/mcs2018.pdf. 15. P. W. Gruber, P. A. Medina, G. A. Keoleian, S. E. Kesler, M. P. Everson and T. J. Wallington, J. Ind. Ecol., 2011, 15, 760–775. 16. United States Geological Survey, Mineral Commodity Summaries 2017, USGS Mineral Resources Program, 2017, https://minerals.usgs.gov/ minerals/pubs/mcs/2017/mcs2017.pdf. 17. M. Armand and J. M. Tarascon, Nature, 2008, 451, 652. 18. C. Vaalma, D. Buchholz, M. Weil and S. Passerini, Nat. Rev. Mater., 2018, 3, 18013. 19. P. Denholm, E. Ela, B. Kirby and M. Milligan, The Role of Energy Storage with Renewable Electricity Generation, 2010, https://www.nrel.gov/docs/ fy10osti/47187.pdf. 20. P. Kurzweil, J. Power Sources, 2010, 195, 4424–4434. 21. Y. Yung-Fang Yu and J. T. Kummer, J. Inorg. Nucl. Chem., 1967, 29, 2453–2475. 22. X. Lu, G. Xia, J. P. Lemmon and Z. Yang, J. Power Sources, 2010, 195, 2431–2442. 23. O. Taku, K. Masaharu and O. Akiyasu, Int. J. Appl. Ceram. Technol., 2004, 1, 269–276. 24. R. J. Bones, D. A. Teagle, S. D. Brooker and F. L. Cullen, J. Electrochem. Soc., 1989, 136, 1274–1277. 25. W. Kangro, Ger. pat., DE914264, 1949. 26. W. Kangro and H. Pieper, Electrochim. Acta, 1962, 7, 435–448. 27. J. Winsberg, T. Hagemann, T. Janoschka, M. D. Hager and U. S. Schubert, Angew. Chem., Int. Ed., 2017, 56, 686–711. 28. M. Skyllas-Kazacos, M. H. Chakrabarti, S. A. Hajimolana, F. S. Mjalli and M. Saleem, J. Electrochem. Soc., 2011, 158, R55–R79. 29. R. D. Shannon, Acta Crystallogr., Sect. A, 1976, 32, 751–767. 30. H. D. Yoo, I. Shterenberg, Y. Gofer, G. Gershinsky, N. Pour and D. Aurbach, Energy Environ. Sci., 2013, 6, 2265–2279. 31. C. Delmas, J. J. Braconnier, A. Maazaz and P. Hagenmuller, Rev. Chim. Miner., 1982, 19, 343–351. 32. K. M. Abraham, Solid State Ionics, 1982, 7, 199–212. 33. T. Shishikura and M. Takeuchi, Jp. Pat., 1899055, 1987. 34. L. W. Shacklette, L. B. Townsend and T. R. Jow, WO Pat. 89/004066, 1988.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

212

Chapter 8

35. D. A. Stevens and J. R. Dahn, J. Electrochem. Soc., 2000, 147, 1271–1273. 36. S. Komaba, W. Murata, T. Ishikawa, N. Yabuuchi, T. Ozeki, T. Nakayama, A. Ogata, K. Gotoh and K. Fujiwara, Adv. Funct. Mater., 2011, 21, 3859–3867. 37. A. Castillo and D. F. Gayme, Energy Convers. Manage., 2014, 87, 885–894. 38. World Energy Council E-Storage, 2016, https://www.worldenergy.org/ wp-content/uploads/2017/03/WEResources_E-storage_2016.pdf. 39. G. H. A. Akhil, A. B. Currier, B. C. Kaun, D. M. Rastler, S. B. Chen, A. L. Cotter, D. T. Bradshaw and W. D. Gauntlett, Electricity Storage Handbook in Collaboration with NRECA, 2013, https://www.energy.gov/sites/prod/ files/2013/08/f2/ElecStorageHndbk2013.pdf. 40. Y. Marcus, Pure Appl. Chem., 1985, 57, 1129–1132. 41. C. Delmas, Adv. Energy Mater., 2018, 8, 1703137. 42. A. Eftekhari, J. Power Sources, 2004, 126, 221–228. 43. Y. Matsuda, H. Nakashima, M. Morita and Y. Takasu, J. Electrochem. Soc., 1981, 128, 2552–2556. 44. M. Okoshi, Y. Yamada, S. Komaba, A. Yamada and H. Nakai, J. Electrochem. Soc., 2017, 164, A54–A60. 45. S. Komaba, T. Hasegawa, M. Dahbi and K. Kubota, Electrochem. Commun., 2015, 60, 172–175. 46. N. Yabuuchi and T. Ohzuku, J. Power Sources, 2003, 119–121, 171–174. 47. C. Delmas and I. Saadoune, Solid State Ionics, 1992, 53–56, 370–375. 48. K. Mizushima, P. C. Jones, P. J. Wiseman and J. B. Goodenough, Mater. Res. Bull., 1980, 15, 783–789. 49. C. Delmas, C. Fouassier and P. Hagenmuller, Physica B+C, 1980, 99, 81–85. 50. Y. Takeda, K. Nakahara, M. Nishijima, N. Imanishi, O. Yamamoto, M. Takano and R. Kanno, Mater. Res. Bull., 1994, 29, 659–666. 51. J. Zhao, L. Zhao, N. Dimov, S. Okada and T. Nishida, J. Electrochem. Soc., 2013, 160, A3077–A3081. 52. K. Ado, M. Tabuchi, H. Kobayashi, H. Kageyama, O. Nakamura, Y. Inaba, R. Kanno, M. Takagi and Y. Takeda, J. Electrochem. Soc., 1997, 144, L177–L180. 53. N. Yabuuchi, H. Yoshida and S. Komaba, Electrochemistry, 2012, 80, 716–719. 54. X. Ma, H. Chen and G. Ceder, J. Electrochem. Soc., 2011, 158, A1307–A1312. 55. J. J. Braconnier, C. Delmas, C. Fouassier and P. Hagenmuller, Mater. Res. Bull., 1980, 15, 1797–1804. 56. J. J. Braconnier, C. Delmas and P. Hagenmuller, Mater. Res. Bull., 1982, 17, 993–1000. 57. X. Li, Y. Wang, D. Wu, L. Liu, S.-H. Bo and G. Ceder, Chem. Mater., 2016, 28, 6575–6583. 58. S. Kim, X. Ma, S. P. Ong and G. Ceder, Phys. Chem. Chem. Phys., 2012, 14, 15571–15578.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

From Lithium to Sodium and Potassium Batteries

213

59. P. Vassilaras, A. J. Toumar and G. Ceder, Electrochem. Commun., 2014, 38, 79–81. 60. H. Yoshida, N. Yabuuchi and S. Komaba, Electrochem. Commun., 2013, 34, 60–63. 61. S.-M. Oh, S.-T. Myung, J.-Y. Hwang, B. Scrosati, K. Amine and Y.-K. Sun, Chem. Mater., 2014, 26, 6165–6171. 62. D. Jianqiu, L. Wen-Bin, L. Xiao, Y. Qingrong, W. Zhongmin, L. Hua-Kun, Z. Huaiying and D. Shi-Xue, Adv. Energy Mater., 2018, 8, 1701610. 63. H.-R. Yao, P.-F. Wang, Y. Gong, J. Zhang, X. Yu, L. Gu, C. OuYang, Y.-X. Yin, E. Hu, X.-Q. Yang, E. Stavitski, Y.-G. Guo and L.-J. Wan, J. Am. Chem. Soc., 2017, 139, 8440–8443. 64. J.-Y. Hwang, S.-T. Myung, J. U. Choi, C. S. Yoon, H. Yashiro and Y.-K. Sun, J. Mater. Chem. A, 2017, 5, 23671–23680. 65. W. Jeitschko, H. Nowotny and F. Benesovsky, Monatsh. Chem. Verw. Teile Anderer Wiss., 1963, 94, 672–676. 66. N. Yabuuchi, R. Hara, M. Kajiyama, K. Kubota, T. Ishigaki, A. Hoshikawa and S. Komaba, Adv. Energy Mater., 2014, 4, 1301453. 67. A. Mendiboure, C. Delmas and P. Hagenmuller, J. Solid State Chem., 1985, 57, 323–331. 68. Y. Hinuma, Y. S. Meng and G. Ceder, Phys. Rev. B, 2008, 77, 224111. 69. R. Berthelot, D. Carlier and C. Delmas, Nat. Mater., 2010, 10, 74. 70. S. Kumakura, Y. Tahara, K. Kubota, K. Chihara and S. Komaba, Angew. Chem., Int. Ed., 2016, 55, 12760–12763. 71. N. Yabuuchi, M. Kajiyama, J. Iwatate, H. Nishikawa, S. Hitomi, R. Okuyama, R. Usui, Y. Yamada and S. Komaba, Nat. Mater., 2012, 11, 512. 72. Z. Lu and J. R. Dahn, J. Electrochem. Soc., 2001, 148, A1225–A1229. 73. H. Yoshida, N. Yabuuchi, K. Kubota, I. Ikeuchi, A. Garsuch, M. Schulz-Dobrick and S. Komaba, Chem. Commun., 2014, 50, 3677–3680. 74. J. Martinez De Ilarduya, L. Otaegui, J. M. López del Amo, M. Armand and G. Singh, J. Power Sources, 2017, 337, 197–203. 75. A. K. Padhi, K. S. Nanjundaswamy and J. B. Goodenough, J. Electrochem. Soc., 1997, 144, 1188–1194. 76. C. Masquelier and L. Croguennec, Chem. Rev., 2013, 113, 6552–6591. 77. P. Barpanda, L. Lander, S. Nishimura and A. Yamada, Adv. Energy Mater., 2018, 8, 1703055. 78. A. K. Padhi, K. S. Nanjundaswamy, C. Masquelier and J. B. Goodenough, J. Electrochem. Soc., 1997, 144, 2581–2586. 79. P. Moreau, D. Guyomard, J. Gaubicher and F. Boucher, Chem. Mater., 2010, 22, 4126–4128. 80. M. Nakayama, S. Yamada, R. Jalem and T. Kasuga, Solid State Ionics, 2016, 286, 40–44. 81. K. Zaghib, J. Trottier, P. Hovington, F. Brochu, A. Guerfi, A. Mauger and C. M. Julien, J. Power Sources, 2011, 196, 9612–9617.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

214

Chapter 8

82. M. Avdeev, Z. Mohamed, C. D. Ling, J. Lu, M. Tamaru, A. Yamada and P. Barpanda, Inorg. Chem., 2013, 52, 8685–8693. 83. P. Moreau, D. Guyomard, J. Gaubicher and F. Boucher, Chem. Mater., 2010, 22, 4126–4128. 84. Y. Fang, L. Xiao, J. Qian, X. Ai, H. Yang and Y. Cao, Nano Lett., 2014, 14, 3539–3543. 85. P. Barpanda, G. Oyama, S.-I. Nishimura, S.-C. Chung and A. Yamada, Nat. Commun., 2014, 5, 4358. 86. K. Saravanan, C. W. Mason, A. Rudola, K. H. Wong and P. Balaya, Adv. Energy Mater., 2013, 3, 444–450. 87. Q. Wang, J. Xu, W. Zhang, M. Mao, Z. Wei, L. Wang, C. Cui, Y. Zhu and J. Ma, J. Mater. Chem. A, 2018, 6, 8815–8838. 88. S. P. Ong, V. L. Chevrier, G. Hautier, A. Jain, C. Moore, S. Kim, X. Ma and G. Ceder, Energy Environ. Sci., 2011, 4, 3680–3688. 89. Z. Jian, W. Han, X. Lu, H. Yang, Y.-S. Hu, J. Zhou, Z. Zhou, J. Li, W. Chen, D. Chen and L. Chen, Adv. Energy Mater., 2013, 3, 156–160. 90. X. Xiang, K. Zhang and J. Chen, Adv. Mater., 2015, 27, 5343–5364. 91. Z. Jian, L. Zhao, H. Pan, Y.-S. Hu, H. Li, W. Chen and L. Chen, Electrochem. Commun., 2012, 14, 86–89. 92. Y. H. Jung, C. H. Lim and D. K. Kim, J. Mater. Chem. A, 2013, 1, 11350–11354. 93. C. Zhu, K. Song, P. A. van Aken, J. Maier and Y. Yu, Nano Lett., 2014, 14, 2175–2180. 94. R. K. B. Gover, A. Bryan, P. Burns and J. Barker, Solid State Ionics, 2006, 177, 1495–1500. 95. A. Ponrouch, R. Dedryvere, D. Monti, A. E. Demet, J. M. Ateba Mba, L. Croguennec, C. Masquelier, P. Johansson and M. R. Palacin, Energy Environ. Sci., 2013, 6, 2361–2369. 96. H. Kim, R. A. Shakoor, C. Park, S. Y. Lim, J.-S. Kim, Y. N. Jo, W. Cho, K. Miyasaka, R. Kahraman, Y. Jung and J. W. Choi, Adv. Funct. Mater., 2013, 23, 1147–1155. 97. K.-H. Ha, S. H. Woo, D. Mok, N.-S. Choi, Y. Park, S. M. Oh, Y. Kim, J. Kim, J. Lee, L. F. Nazar and K. T. Lee, Adv. Energy Mater., 2013, 3, 770–776. 98. M. Nose, H. Nakayama, K. Nobuhara, H. Yamaguchi, S. Nakanishi and H. Iba, J. Power Sources, 2013, 234, 175–179. 99. M. Nose, S. Shiotani, H. Nakayama, K. Nobuhara, S. Nakanishi and H. Iba, Electrochem. Commun., 2013, 34, 266–269. 100. Y. Kawabe, N. Yabuuchi, M. Kajiyama, N. Fukuhara, T. Inamasu, R. Okuyama, I. Nakai and S. Komaba, Electrochem. Commun., 2011, 13, 1225–1228. 101. Y. Kawabe, N. Yabuuchi, M. Kajiyama, N. Fukuhara, T. Inamasu, R. Okuyama, I. Nakai and S. Komaba, Electrochemistry, 2012, 80, 80–84. 102. F. Sauvage, E. Quarez, J. M. Tarascon and E. Baudrin, Solid State Sci., 2006, 8, 1215–1221.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

From Lithium to Sodium and Potassium Batteries

215

103. R. A. Shakoor, D.-H. Seo, H. Kim, Y.-U. Park, J. Kim, S.-W. Kim, H. Gwon, S. Lee and K. Kang, J. Mater. Chem., 2012, 22, 20535–20541. 104. K. Chihara, A. Kitajou, I. D. Gocheva, S. Okada and J.-I. Yamaki, J. Power Sources, 2013, 227, 80–85. 105. R. Tripathi, S. M. Wood, M. S. Islam and L. F. Nazar, Energy Environ. Sci., 2013, 6, 2257–2264. 106. J. Qian, C. Wu, Y. Cao, Z. Ma, Y. Huang, X. Ai and H. Yang, Adv. Energy Mater., 2018, 8, 1870079. 107. Y. Lu, L. Wang, J. Cheng and J. B. Goodenough, Chem. Commun., 2012, 48, 6544–6546. 108. M. I. Jamesh and A. S. Prakash, J. Power Sources, 2018, 378, 268–300. 109. J. Peng, J. Wang, H. Yi, W. J. Hu, Y. Yu, J. Yin, Y. Shen, Y. Liu, J. Luo, Y. Xu, P. Wei, Y. Li, Y. Jin, Y. Ding, L. Miao, J. Jiang, J. Han and Y. Huang, Adv. Energy Mater., 2018, 8, 1870048. 110. P. Marzak, J. Yun, A. Dorsel, A. Kriele, R. Gilles, O. Schneider and A. S. Bandarenka, J. Phys. Chem. C, 2018, 122, 8760–8768. 111. B. Paulitsch, J. Yun, A. S. Bandarenka, J. Yun and A. S. Bandarenka, ACS Appl. Mater. Interfaces, 2017, 9, 8107–8112. 112. S. T. Senthilkumar, H. Bae, J. Han, Y. Kim and Y. Kim, Angew. Chem., Int. Ed., 2018, 57, 5335–5339. 113. Y. You, X.-L. Wu, Y.-X. Yin and Y.-G. Guo, Energy Environ. Sci., 2014, 7, 1643–1647. 114. Y. Yue, A. J. Binder, B. Guo, Z. Zhang, Z.-A. Qiao, C. Tian and S. Dai, Angew. Chem., Int. Ed., 2014, 53, 3134–3137. 115. Y. Liu, D. He, R. Han, G. Wei and Y. Qiao, Chem. Commun., 2017, 53, 5569–5572. 116. J. Qian, C. Wu, Y. Cao, Z. Ma, Y. Huang, X. Ai and H. Yang, Adv. Energy Mater., 2018, 8, 1702619. 117. X. Wu, W. Deng, J. Qian, Y. Cao, X. Ai and H. Yang, J. Mater. Chem. A, 2013, 1, 10130–10134. 118. Y. Liu, Y. Qiao, W. Zhang, Z. Li, X. Ji, L. Miao, L. Yuan, X. Hu and Y. Huang, Nano Energy, 2015, 12, 386–393. 119. Y. Mizuno, M. Okubo, E. Hosono, T. Kudo, H. Zhou and K. Oh-ishi, J. Phys. Chem. C, 2013, 117, 10877–10882. 120. X. Bie, K. Kubota, T. Hosaka, K. Chihara and S. Komaba, J. Power Sources, 2018, 378, 322–330. 121. Y. Okamoto, J. Phys. Chem. C, 2014, 118, 16–19. 122. J. Maibach, F. Jeschull, D. Brandell, K. Edstroem and M. Valvo, ACS Appl. Mater. Interfaces, 2017, 9, 12373–12381. 123. H.-Y. Song and S.-K. Jeong, J. Power Sources, 2018, 373, 110–118. 124. M. Cabello, X. Bai, T. Chyrka, G. F. Ortiz, P. Lavela, R. Alcántara and J. L. Tirado, J. Electrochem. Soc., 2017, 164, A3804–A3813. 125. B. Jache and P. Adelhelm, Angew. Chem., Int. Ed., 2014, 126, 10333–10337. 126. D. Saurel, B. Orayech, B. Xiao, D. Carriazo, X. Li and T. Rojo, Adv. Energy Mater., 2018, 8, 1703268.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

216

Chapter 8

127. H. Shi, J. N. Reimers and J. R. Dahn, J. Appl. Crystallogr., 1993, 26, 827–836. 128. M. M. Doeff, Y. Ma, S. J. Visco and L. C. De Jonghe, J. Electrochem. Soc., 1993, 140, L169–L170. 129. J. J. Kipling, J. N. Sherwood, P. V. Shooter and N. R. Thompson, Carbon, 1964, 1, 315–320. 130. Y. Miao, J. Zong and X. Liu, Mater. Lett., 2017, 188, 355–358. 131. D. A. Stevens and J. R. Dahn, J. Electrochem. Soc., 2001, 148, A803–A811. 132. H. Zhu, F. Shen, W. Luo, S. Zhu, M. Zhao, B. Natarajan, J. Dai, L. Zhou, X. Ji, R. S. Yassar, T. Li and L. Hu, Nano Energy, 2017, 33, 37–44. 133. B. Cao, H. Liu, B. Xu, Y. Lei, X. Chen and H. Song, J. Mater. Chem. A, 2016, 4, 6472–6478. 134. L. Yu, H. Song, Y. Li, Y. Chen, X. Chen, J. Zhou, Z. Ma, X. Wan, P. Tian and J. Wu, Electrochim. Acta, 2016, 218, 285–293. 135. F. Ling, L. Qian, C. Suhua, X. Zhi and L. Bingan, Adv. Energy Mater., 2017, 7, 1602778. 136. Z. Jian, C. Bommier, L. Luo, Z. Li, W. Wang, C. Wang, P. A. Greaney and X. Ji, Chem. Mater., 2017, 29, 2314–2320. 137. E. Buiel and J. R. Dahn, Electrochim. Acta, 1999, 45, 121–130. 138. M. Dahbi, N. Yabuuchi, K. Kubota, K. Tokiwa and S. Komaba, Phys. Chem. Chem. Phys., 2014, 16, 15007–15028. 139. P. Liu, Y. Li, Y.-S. Hu, H. Li, L. Chen and X. Huang, J. Mater. Chem. A, 2016, 4, 13046–13052. 140. W. Liming, B. Daniel, V. Christoph, G. A. Giffin and P. Stefano, ChemElectroChem, 2016, 3, 292–298. 141. M. Dahbi, M. Kiso, K. Kubota, T. Horiba, T. Chafik, K. Hida, T. Matsuyama and S. Komaba, J. Mater. Chem. A, 2017, 5, 9917–9928. 142. S. Komaba, Y. Matsuura, T. Ishikawa, N. Yabuuchi, W. Murata and S. Kuze, Electrochem. Commun., 2012, 21, 65–68. 143. N. Yabuuchi, Y. Matsuura, T. Ishikawa, S. Kuze, J.-Y. Son, Y.-T. Cui, H. Oji and S. Komaba, ChemElectroChem, 2014, 1, 580–589. 144. E. Ferg, R. J. Gummow, A. de Kock and M. M. Thackeray, J. Electrochem. Soc., 1994, 141, L147–L150. 145. T. Ohzuku, A. Ueda and N. Yamamoto, J. Electrochem. Soc., 1995, 142, 1431–1435. 146. Y. Sun, L. Zhao, H. Pan, X. Lu, L. Gu, Y.-S. Hu, H. Li, M. Armand, Y. Ikuhara, L. Chen and X. Huang, Nat. Commun., 2013, 4, 1870. 147. A. Rudola, K. Saravanan, C. W. Mason and P. Balaya, J. Mater. Chem. A, 2013, 1, 2653–2662. 148. A. Katogi, K. Kubota, K. Chihara, K. Miyamoto, T. Hasegawa and S. Komaba, ACS Appl. Energy Mater., 2018, 3630–3635. 149. C. Delmas, F. Cherkaoui, A. Nadiri and P. Hagenmuller, Mater. Res. Bull., 1987, 22, 631–639. 150. G. Pang, P. Nie, C. Yuan, L. Shen, X. Zhang, H. Li and C. Zhang, J. Mater. Chem. A, 2014, 2, 20659–20666.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

From Lithium to Sodium and Potassium Batteries

217

151. P. Senguttuvan, G. Rousse, M. E. Arroyo y de Dompablo, H. Vezin, J. M. Tarascon and M. R. Palacín, J. Am. Chem. Soc., 2013, 135, 3897–3903. 152. L. Mengmeng, Z. Yu, L. Wenbin, Y. Qingyu, S. Wenping and D. S. Xue, Adv. Mater., 2017, 29, 1700622. 153. A. Magasinski, P. Dixon, B. Hertzberg, A. Kvit, J. Ayala and G. Yushin, Nat. Mater., 2010, 9, 353–358. 154. V. L. Chevrier and G. Ceder, J. Electrochem. Soc., 2011, 158, A1011–A1014. 155. A. Darwiche, C. Marino, M. T. Sougrati, B. Fraisse, L. Stievano and L. Monconduit, J. Am. Chem. Soc., 2012, 134, 20805–20811. 156. M. Dahbi, N. Yabuuchi, M. Fukunishi, K. Kubota, K. Chihara, K. Tokiwa, X.-F. Yu, H. Ushiyama, K. Yamashita, J.-Y. Son, Y.-T. Cui, H. Oji and S. Komaba, Chem. Mater., 2016, 28, 1625–1635. 157. Y. Kim, Y. Park, A. Choi, N.-S. Choi, J. Kim, J. Lee, J. H. Ryu, S. M. Oh and K. T. Lee, Adv. Mater., 2013, 25, 3045–3049. 158. Y. Zhu, Y. Wen, X. Fan, T. Gao, F. Han, C. Luo, S.-C. Liou and C. Wang, ACS Nano, 2015, 9, 3254–3264. 159. C. Delmas, C. Fouassier and P. Hagenmuller, J. Solid State Chem., 1975, 13, 165–171. 160. M. Jansen, F. M. Chang and R. Hoppe, Z. Anorg. Allg. Chem., 1982, 490, 101–110. 161. N. Z. Ali, J. Nuss, D. Sheptyakov and M. Jansen, J. Solid State Chem., 2010, 183, 752–759. 162. D. Sheptyakov, N. Z. Ali and M. Jansen, J. Phys.: Condens. Matter, 2010, 22, 426001. 163. Y. Hironaka, K. Kubota and S. Komaba, Chem. Commun., 2017, 53, 3693–3696. 164. H. Kim, J. C. Kim, S.-H. Bo, T. Shi, D.-H. Kwon and G. Ceder, Adv. Energy Mater., 2017, 7, 1700098. 165. X. Wang, X. Xu, C. Niu, J. Meng, M. Huang, X. Liu, Z. Liu and L. Mai, Nano Lett., 2016, 17, 544–550. 166. C. Vaalma, G. A. Giffin, D. Buchholz and S. Passerini, J. Electrochem. Soc., 2016, 163, A1295–A1299. 167. N. Recham, G. Rousse, M. T. Sougrati, J.-N. Chotard, C. Frayret, S. Mariyappan, B. C. Melot, J.-C. Jumas and J.-M. Tarascon, Chem. Mater., 2012, 24, 4363–4370. 168. S. S. Fedotov, N. R. Khasanova, A. S. Samarin, O. A. Drozhzhin, D. Batuk, O. M. Karakulina, J. Hadermann, A. M. Abakumov and E. V. Antipov, Chem. Mater., 2016, 28, 411–415. 169. K. Chihara, A. Katogi, K. Kubota and S. Komaba, Chem. Commun., 2017, 53, 5208–5211. 170. V. Mathew, S. Kim, J. Kang, J. Gim, J. Song, J. P. Baboo, W. Park, D. Ahn, J. Han, L. Gu, Y. Wang, Y.-S. Hu, Y.-K. Sun and J. Kim, NPG Asia Mater., 2014, 6, e138. 171. H. Tokoro and S.-I. Ohkoshi, Dalton Trans., 2011, 40, 6825–6833. 172. X. Wu, Z. Jian, Z. Li and X. Ji, Electrochem. Commun., 2017, 77, 54–57.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

218

Chapter 8

173. S. Lian, W. Zhaoxiang and C. Liquan, Chem. - Eur. J., 2014, 20, 12559–12562. 174. J.-Y. Liao, Q. Hu, B.-K. Zou, J.-X. Xiang and C.-H. Chen, Electrochim. Acta, 2016, 220, 114–121. 175. M. J. Piernas-Muñoz, E. Castillo-Martínez, O. Bondarchuk, M. Armand and T. Rojo, J. Power Sources, 2016, 324, 766–773. 176. L. Xue, H. Gao, W. Zhou, S. Xin, K. Park, Y. Li and J. B. Goodenough, Adv. Mater., 2016, 28, 9608–9612. 177. F. Herren, P. Fischer, A. Ludi and W. Haelg, Inorg. Chem., 1980, 19, 956–959. 178. X. Bie, K. Kubota, T. Hosaka, K. Chihara and S. Komaba, J. Mater. Chem. A, 2017, 5, 4325–4330. 179. A. Schleede and M. Wellmann, Z. Kristallogr. - Cryst. Mater., 1932, 83, 148. 180. D. Liu, Z. Yang, W. Li, S. Qiu and Y. Luo, Electrochim. Acta, 2010, 55, 1013–1018. 181. E. Matuyama, Nature, 1962, 193, 61. 182. Z. Jin, Z. Xiaoxi, Z. Yujie, X. Yunhua and W. Chunsheng, Adv. Funct. Mater., 2016, 26, 8103–8110. 183. Z. Jian, W. Luo and X. Ji, J. Am. Chem. Soc., 2015, 137, 11566–11569. 184. J. Zelang, X. Zhenyu, B. Clement, L. Zhifei and J. Xiulei, Adv. Energy Mater., 2016, 6, 1501874. 185. C. Chen, Z. Wang, B. Zhang, L. Miao, J. Cai, L. Peng, Y. Huang, J. Jiang, Y. Huang, L. Zhang and J. Xie, Energy Storage Mater., 2017, 8, 161–168. 186. X. Peixun, Z. Xinxin and X. Yunhua, ChemSusChem, 2018, 11, 202–208. 187. Z. Xing, Y. Qi, Z. Jian and X. Ji, ACS Appl. Mater. Interfaces, 2017, 9, 4343–4351. 188. K. Share, A. P. Cohn, R. Carter, B. Rogers and C. L. Pint, ACS Nano, 2016, 10, 9738–9744. 189. W. Luo, J. Wan, B. Ozdemir, W. Bao, Y. Chen, J. Dai, H. Lin, Y. Xu, F. Gu, V. Barone and L. Hu, Nano Lett., 2015, 15, 7671–7677. 190. H. Kim, J. C. Kim, M. Bianchini, D.-H. Seo, J. Rodriguez-Garcia and G. Ceder, Adv. Energy Mater., 2018, 8, 1702384. 191. B. Key, R. Bhattacharyya, M. Morcrette, V. Seznéc, J.-M. Tarascon and C. P. Grey, J. Am. Chem. Soc., 2009, 131, 9239–9249. 192. J. Sangster, J. Phase Equilib. Diffus., 2006, 27, 190–191. 193. T. T. Tran and M. N. Obrovac, J. Electrochem. Soc., 2011, 158, A1411–A1416. 194. I. Sultana, M. M. Rahman, Y. Chen, A. M. Glushenkov, Adv. Funct. Mater., 2017, 28, 1703857. 195. M. N. Obrovac and V. L. Chevrier, Chem. Rev., 2014, 114, 11444–11502. 196. Z. Li, J. Ding and D. Mitlin, Acc. Chem. Res., 2015, 48, 1657–1665. 197. Q. Wang, X. Zhao, C. Ni, H. Tian, J. Li, Z. Zhang, S. X. Mao, J. Wang and Y. Xu, J. Phys. Chem. C, 2017, 121, 12652–12657.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00181

From Lithium to Sodium and Potassium Batteries

219

198. W. D. McCulloch, X. Ren, M. Yu, Z. Huang and Y. Wu, ACS Appl. Mater. Interfaces, 2015, 7, 26158–26166. 199. W. Zhang, J. Mao, S. Li, Z. Chen and Z. Guo, J. Am. Chem. Soc., 2017, 139, 3316–3319. 200. https://about.bnef.com/blog/global-storage-market-double-sixtimes-2030/. 201. https://about.bnef.com/electric-vehicle-outlook/.

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

Chapter 9

Understanding Battery Aging Mechanisms Dongjiang Lia, Dmitri L. Danilov a,b, Henk Jan Bergveldb,c, Rüdiger-A. Eichela,d and Peter H. L. Notten*a,b,e a

Forschungszentrum Jülich (IEK-9), D-52425 Jülich, Germany; bEindhoven University of Technology, Eindhoven, 5600 MB, The Netherlands; cNXP Semiconductors, 5600 KA Eindhoven, The Netherlands; dRWTH Aachen University, D-52074 Aachen, Germany; eUniversity of Technology Sydney, Broadway, Sydney, NSW 2007, Australia *E-mail: [email protected]

9.1  Introduction In past centuries the rapid development of our civilization was ensured by a vast amount of fossil fuels. For that reason, energy consumption and environmental pollution have dramatically increased and this is seriously threatening our human society. As promoted by many governments, electric vehicles (EVs) and hybrid electric vehicles (HEVs) have become the most promising candidates for the next generation of environmentally friendly transportation. Li-ion batteries, which provide the energy and power, play a decisive role in the development of EVs. Battery performance, such as high energy density, long cycle life and safety standards, has become the bottleneck in the introduction and successful commercialization of EVs. The typical structure of a Li-ion battery includes various parts such as cathode, anode, separator, electrolyte, etc. All battery components will   Future Lithium-ion Batteries Edited by Ali Eftekhari © The Royal Society of Chemistry 2019 Published by the Royal Society of Chemistry, www.rsc.org

220

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

Understanding Battery Aging Mechanisms

221

experience a different extent of decay during usage, known as aging. Battery aging leads to serious detrimental effects, e.g., reducing the battery lifespan, decreasing the energy and power efficiency, and therefore increases the failure risk and EV cost. Controlling and diminishing the aging processes is one of the most critical challenges in advanced battery development, which starts with a proper understanding of the underlying aging mechanisms. Since the introduction of LiFePO4 as cathode material in Li-ion batteries (LIBs) by Padhi et al.,1 LiFePO4-based batteries (LFP) are nowadays drawing much attention due to the many favorable characteristics, such as high safety, long lifespan, environmental friendliness, low cost and widespread material abundancy. Moreover, LFP batteries are excellent candidates for investigating battery aging as the cathode has outstanding stability properties at moderate temperatures and the focus can therefore be put on the graphite anode. Since graphite is used as the anode in most commercial Li-ion batteries, this study has a much wider impact on many battery systems. This chapter will introduce the aging mechanisms of LFP batteries under both calendar and cycling conditions by experimental and theoretical/modeling approaches. Firstly, the capacity loss, which is mainly attributed to the cyclable Li immobilization, has been experimentally measured. Conventional postmortem analyses such as scanning electron microscopy (SEM), X-ray photoelectron spectroscopy (XPS), and Raman spectroscopy are used to unravel the capacity loss and the electrode degradation mechanisms. A novel non-destructive approach based on electromotive force (EMF) determination is then proposed to quantify the graphite degradation. Subsequently, a comprehensive electrochemistry model is introduced to simulate the irreversible capacity loss under various aging conditions. The dependence of capacity loss on aging conditions such as storage state of charge (SoC), cycling currents and temperatures, etc., is simulated and the simulations are in good agreement with the experiments. The aging model allows researchers to have an in-depth understanding of aging mechanisms and therefore helps manufacturers to optimize manufacturing procedures and design a new generation of batteries with high cycling performance. Moreover, the model can be further used to predict the battery cycle life, which can be used to develop more accurate battery management systems (BMSs) to increase the battery efficiency and safety.

9.2  W  orking Principles and Cell Design of Li-ion Batteries 9.2.1  Working Principles of Li-ion Batteries Like all electrochemical cells, a commercial Li-ion cell is composed of two electrodes, an anode and a cathode. A porous separator membrane is placed between the two electrodes to prevent electrical contact between them. The cell is filled with electrolyte composed of non-aqueous solvents and Li salts,

View Online

Chapter 9

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

222

Figure 9.1  Basic  concept of a Li-ion battery. Reproduced from ref. 2. e.g., EC (ethylene carbonate)/DEC (diethyl carbonate)/DMC (dimethyl carbonate) (1 : 1 : 1), VC (vinylene carbonate) (2%), and LiPF6 in concentration of 1 M. Figure 9.1 represents the layout of a generic Li-ion cell, where the anode is assumed to be graphite (C6) and the cathode is assumed to be LiFePO4 in this section. As indicated by the red arrows in this figure, during charging (ch) electrons are extracted from the LiFePO4 cathode and flow into the graphite anode through the outer circuit. Simultaneously, Li+ ions are delithiated from the LiFePO4 cathode and transported through the electrolyte to the graphite anode, safeguarding charge neutrality in the electrodes. The electric energy converts into chemical energy during charging. The flows of electrons and Li+ ions are reversed during discharging (d) as indicated by the black arrows. The stored chemical energy converts into electrical energy during discharging. The main electrochemical storage reactions of this type of battery can be represented by   

  

  Li x C6 , C6  xLi   xe   

(9.1)

ch   Li1 x FePO4  xLi   xe  , LiFePO4  

(9.2)

ch d

d

   resulting in the following overall reaction   



ch   Li1 x FePO4  Li x C6 , C6  LiFePO4   d

(9.3)

   where x is the SoC, 0 ≤ x ≤ 1. The energy density and stability of Li-ion batteries are strongly determined by the electrode materials and the electrolyte. The relative electron energies in the electrodes and the electrolyte of a thermodynamically stable cell with

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

Understanding Battery Aging Mechanisms

223

Figure 9.2  Schematic  representation of the open-circuit energy diagram of a

non-aqueous cell. Eaf and Efc are the Fermi level of the anode and cathode, respectively. The red and blue lines show two cases of the LUMO and HOMO inside the batteries.

non-aqueous electrolyte are shown in Figure 9.2. The Fermi levels of the two electrodes are Eaf for the anode and Efc for the cathode. The open-circuit voltage of the cell (VOCP) is determined by   



VOCP 

  Efc  Efa  e

,

(9.4)

   where e is the electronic charge. The energy difference of the lowest unoccupied molecular orbital (LUMO) and the highest occupied molecular orbital (HOMO) of the electrolyte is the “stability window” of the electrolyte.3,4 As illustrated in Figure 9.2, in the case that LUMO* (red) is below Eaf, the electrolyte will be reduced at the anode; in the case that HOMO* (red) is above Eaf, the electrolyte will be oxidized at the cathode. Therefore, LUMO (blue) should be higher than Eaf and HOMO (blue) should preferably be lower than Efc and, consequently, the difference between Eaf and Efc should be within the stability window of the electrolyte. It is essential to develop advanced electrolytes in order to design stable and safe Li-ion batteries.

View Online

Chapter 9

224

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

9.2.2  Design of Li-ion Batteries Li-ion battery performance, such as the energy and power densities, the storage and cycle life, safety characteristics, etc., is not only dependent on the properties of its materials, but is also determined by the cell design. The shapes and components of various Li-ion battery configurations are summarized in Figure 9.3. Cylindrical batteries in most products follow standard models in terms of size, for example, the 18 650 and 26 650 cell, etc. Typical 18 650 cells in commercial Li-ion battery products were considered to be the most advanced cells, offering a volumetric energy density of 600–650 Wh L−1.5 Recently, Tesla Motors, Panasonic and Samsung have decided on the 21 700 cell for reasons of easy manufacturing, optimal capacity and other benefits. The 18 650 cell with a volume of 66 cm3 has a capacity of around 3 Ah, while the 21 700 cell with a volume of 97 cm3 is expected to deliver a capacity up to 6 Ah, essentially doubling the capacity with only a 50% increase in volume. Tesla Motors refers to their company's new 21 700 cell as the “highest energy-density cell that is also the cheapest”.6 Prismatic cells were firstly introduced in the early 1990s. The electrodes can be assembled by the layer-stacking or jelly-rolling approaches. The assembled cell is then encased by a metal or hard-plastic housing for stability. The capacity of a prismatic cell varies from several ampere-hours for cell phones, laptops, etc., to up to hundred ampere-hours for electric vehicles.

Figure 9.3  Representation  of the shape and components of various Li-ion battery configurations: cylindrical (a), prismatic (b), coin (c) and pouch cell (d). Reproduced from ref. 7 with permission from Springer Nature, Copyright 2018.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

Understanding Battery Aging Mechanisms

225

The prismatic cell improves space utilization and allows flexible design but it can be more expensive for manufacturing. The coin cell, also known as a button cell, provided power for compact portable devices such as medical implants, watches, hearing aids, car keys and memory backup from the 1980s. Although small and inexpensive to build, the stacked button cell fell out of favor and gave way to more conventional battery formats. A drawback of the button cell is that it cannot be equipped with safety defense devices therefore is not safe. When commercial pouch cells were first introduced in 1995, it surprised the battery world with a radical new design approach.8 Rather than using a metallic housing and glass-to-metal electrical feed-through, conductive foil tabs were welded to the electrodes and brought to the outside in a fully sealed way. Therefore, the pouch cell offers a simple, flexible and lightweight solution to battery design. Moreover, the pouch cell makes the most efficient use of space and achieves 90–95% packaging efficiency, which is the highest among battery packs. Eliminating the metal enclosure reduces the total cell weight, but additional support and allowance for swelling must be made.

9.3  Degradation Mechanisms of Li-ion Batteries It is well known that all types of Li-ion batteries suffer from capacity loss during both storage and cycling. The capacity loss can be classified as reversible (ΔQr) and irreversible capacity losses (ΔQir).9 ΔQr is defined as losses that can be fully recovered in the subsequent cycles under low-current conditions. The origin of ΔQr is still under debate. Some scientists attribute ΔQr to current-dependent battery polarization.10 Others have proposed the formation of a metastable “electron-ion-solvent” complex as the reason for ΔQr.11 Extensive studies have been performed to investigate ΔQir. It is generally understood that the degradation mechanisms strongly depend on the battery chemistries, e.g., LiFePO4 12 and LiMO2(M = Ni, Mn, Co).13–15 ΔQir does not originate from a single aging process but rather combines various processes and their interactions. The degree of influence of the individual processes on ΔQir can vary and it is therefore a challenge to identify the contributions of individual processes to the total battery capacity loss. The commonly known degradation processes occurring in Li-ion batteries are discussed in the following sections.

9.3.1  SEI Formation Inducing Cyclable Li Losses As discussed in Figure 9.2, electrons at the anode can transfer to the LUMO of the electrolyte when the LUMO is lower than the anode Fermi level, leading to solvent reduction. Meanwhile, the same number of Li ions stored in the anode electrode will be transferred to the reduced electrolyte in order to

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

226

Chapter 9

safeguard charge neutrality of the whole system. These cyclable Li ions are immobilized by the solid products from solvent reduction reactions, finally leading to irreversible capacity loss. These solid products from the solvent reduction can deposit on the anode surface, forming a so-called solid-electrolyte-interphase (SEI) layer. The SEI layer on the graphite anode plays a dual role in determining battery performance. On the one hand, it protects the graphite anode from exfoliation, induced by the solvent co-intercalation, and prevents further solvent decomposition. On the other hand, the continuous growth of SEI layers will lead to irreversible capacity loss. SEI formation is a common process in commercial Li-ion batteries and occurs at any battery usage conditions, including storage and cycling. Calendar-life studies have revealed that Li-immobilization in the SEI layer is the main source of ΔQir at moderate storage temperatures.16,17 SoC is considered to be an important factor influencing SEI formation at low temperatures.18 At elevated temperatures the cathode degradation may also have an impact on ΔQir.12 The influence of cathode degradation on ΔQir is strongly dependent on the specific degradation mechanisms. For example, metal ions dissolved from the cathode can be deposited on the graphite surface and, consequently, accelerate SEI formation.19 The structural transformations at the cathode surface will result in the kinetic decline, which also leads to battery capacity loss.14,15 More details on cathode degradation will be given in the next section. Similar to the case of storage, ΔQir has also been attributed to Li-immobilization in the SEI layer under cycling conditions. However, the cycling-induced effect has an additional contribution to ΔQir.18,19 Therefore, it is commonly accepted that ΔQir is larger during cycling than under storage conditions.19 The influences of current, temperature and cycling range on ΔQir are found to be more pronounced under cycling.19,20 The irreversible capacity losses are always accompanied by the degradation of the electrodes, ΔQc and ΔQC6 for cathode and anode, respectively.20 Therefore, electrode degradation under both storage and cycling conditions is important to address, which will be done in the next sections.

9.3.2  Cathode Degradation 9.3.2.1 Transition Metal Dissolution Metal dissolution in various cathode materials was reported by Choi and Manthiram,21 see Table 9.1. The dissolution measurements were performed by soaking the sample powders in the electrolyte (EC/DEC 1 : 1, LiPF6) at 55 °C for seven days, followed by analyzing the amount of metal ions in the electrolyte with atomic absorption spectroscopy (AAS). Although LiCoxNiyMnzO2 (NMC) material is considered to be more stable than LiCoO2 and LiNi0.5Mn0.5O2, considerable dissolution was still observed. From Table 9.1 it can be concluded that NMC(111) material is relatively more stable than other materials of the NMC family.

View Online

Understanding Battery Aging Mechanisms

227

Table 9.1  Comparison  of transition metal dissolution from various cathode mate-

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

rials at 55 °C.21

Sample number Composition 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

LiCoO2 LiNi0.5Mn0.5O2 LiNi0.425Mn0.425Co0.15O2 LiNi0.33Mn0.33Co0.33O2 LiNi0.29Mn0.29Co0.42O2 LiNi0.25Mn0.25Co0.5O2 LiNi0.21Mn0.21Co0.58O2 LiMn0.8Cr0.2O2 LiMnO2 LiMn2O4 LiMn1.5Ni0.5O4 Li1.05Mn1.53Ni0.42O4 LiMn1.5Ni0.42Zn0.08O4 LiMn1.42Ni0.42Co0.16O4 LiFePO4

Metal ion dissolution % (based on sample weight) Mn

Ni

0.4 0.3 0.2 0.4 0.4 0.3 2.6 3.2 3.2 0.3 0.2 0.4 0.3

0.7 0.8 0.4 1.1 0.9 0.8

0.3 0.1 0.3 0.3

Co

Fe

Total

0.5

0.8 1.1 1.1 0.9 1.8 1.8 1.6 2.6 3.2 3.2 0.6 0.3 0.7 0.6 0.5

0.8 0 0.3 0.3 0.5 0.5

Amine et al.12 investigated LFP batteries under both storage and cycling conditions at high temperatures. At the end of the aging experiments, over 640 ppm and 535 ppm of Fe ions were detected in the LiFePO4 and C–LiFePO4 electrolyte solutions, respectively, by inductively coupled plasma ICP measurements. Furthermore, iron deposition on the graphite electrode was also confirmed by energy dispersive X-ray analysis. A similar investigation is described in ref. 22, where the Fe deposition on graphite was confirmed by XPS. It is generally accepted that cathode electrode dissolution is related to H+ contamination in the electrolyte.23,24 For example, H+ ions can react with the LiFePO4 cathode by ion exchange reactions (2H+ ↔ Fe2+). The dissolved iron ions can be transported to the anode and subsequently be reduced at the graphite surface. Thereby, cathode dissolution will not be influenced by the accumulation of Fe2+ ions in the electrolyte since the transition metal ions are continuously removed from the solution due to their reaction on the anode side.23 Residual water inside the electrolyte is considered to be the origin of H+, according to   



elevated T H2 O  LiPF6   LiF  POF3  2H  2F.

(9.5)

   The details of the dissolution mechanism in NMC positive electrodes are still under debate. Some researchers attribute the disproportion reaction of Mn ions to be the main reason for metal dissolution21,25,26 while others consider the acidic corrosion to be the origin.27

View Online

228

Chapter 9

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

9.3.2.2 Structural Transformation It is known that the NMC(111) material experiences a phase transition from the rhombohedral space group R3̄m (initial “O3” phase) to the monoclinic space group C2/m (“O1” phase) beyond a charge voltage of 4.4 V vs. Li+/ Li.15,28,29 The “O1” LiNi1/3Co1/3Mn1/3O2 phase has been clearly observed at x ≈ 0.3.30 Cycling above this phase transition point at higher potentials will lead to a faster capacity decay of the cathode. Structural changes induced by Li–Ni site interchange is considered to be another detrimental effect on the electrode cycling performance.29,31 Due to the similar ionic radius of Ni2+ (0.67 Å) and Li+ (0.76 Å), there is always a possibility that these two ions exchange their crystallographic sites, which induces local disorder in the NMC(111) materials. The un-removable Ni ions in the Li layer will then block Li diffusion pathways, leading to a decrease in the cathode rate capability. High currents32 and voltages15 are considered as unfavorable, leading to distortion of the electrode surface. It is worth pointing out that the degradation mechanism is composition dependent, for example, Jung et al.15 reported that the degradation of NMC(532) (LiNi0.5Mn0.3Co0.2O2) material is attributed to the phase transformation from rhombohedral to spinel at the surface while the degradation of NMC(111) material is due to the phase transformation from the O3 to the O1 phase. Therefore, the degradation mechanisms of NMC materials should be carefully studied with respect to the specific compositions.

9.3.2.3 Particle Isolation Based on the fact that the impedance of cathode electrodes dramatically increases after aging, some researchers believe that particle isolation should be responsible for the capacity and power fades of cathode materials.33,34 The increase of the impedance has been attributed to a loss of the conductive carbon and/or cathode-electrolyte-interface (CEI) layers at the cathode surface.

9.3.3  Anode Electrode Degradation Although graphite is considered to be a stable anode material, structural degradation has still been observed. Kostecki et al.35 suggested that a non-uniform current distribution across and within the anode can lead to co-intercalation of ion aggregates, generating local graphite degradation or exfoliation. The inhomogeneity of the current density within the graphite electrode is more significant at elevated temperatures.35 Furthermore, the cycling range was found to have a significant influence on the graphite electrode decay.36 Typical Raman spectra of graphite electrodes reveal that the structural damage is more pronounced when cycling at lower SoC ranges, for example, at 0 < x < 0.16. After semi-quantitative analysis of Raman spectra it was concluded that the average particle size of the graphite electrode

View Online

Understanding Battery Aging Mechanisms

229

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

37

decreased. Particle cracks and flake exfoliation have been visually observed from the aged graphite electrodes by SEM.37 The stress evolution caused by the repeated volume expansion and contraction of the graphite anode during cycling is believed to be the origin of the particle fracture and structural damage. Mathematical models based on the electrode volume changes have been proposed to quantitatively describe graphite degradation upon cycling. Among these models the diffusion-induced-stress model38–40 was well accepted and it is widely applied to analyze the graphite deformation. The model suggests that particle fractures take place due to severe tangential stresses developed during Li extraction. However, the model proposed by Christensen and Newman suggests that the particle surface is likely to fracture at the end of the extraction while the center is most likely to fracture at the beginning of Li insertion.41,42 Apart from the mechanical stresses induced by Li (de)intercalation, metal dissolution from the cathode and the subsequent deposition on the graphite surface is considered to be another factor responsible for graphite degradation. Transition metal ions on the graphite surface have been detected by energy disperse X-ray spectroscopy12 and XPS.22,23 These metal clusters covering on the graphite surface hinder the Li intercalation process, leading to inaccessibility of the graphite anode.16

9.4  Electromotive Force (EMF) Determination The EMF is the battery voltage when the battery is in the equilibrium state. In practice, the value of EMF (VEMF) is equal to that of the open-circuit-potential (VOCP) after a long relaxation period. The EMF is a very useful tool in investigating aging of Li-ion batteries, especially in analyzing the degradation of the individual electrodes. The relationship between the EMF and the battery thermodynamic properties has been described by   



ΔG = −zFVEMF,

(9.6)

   where ΔG is the change in the Gibb's free energy, z is the number of the electrons involved in the basis charge-transfer reactions (eqn (9.1)–(9.3)), F is the Faraday constant and VEMF is the EMF voltage of the battery. Several methods have been used to measure the battery EMF, such as the galvanostatic intermittent titration technique, the potentiostatic intermittent titration technique, etc. A more flexible and convenient approach, based on the regression extrapolation of voltage discharge curves obtained under various loading conditions, has been proposed by Notten and co-workers.16,20,43–47 Figure 9.4 shows, as an example, a set of voltage discharge curves for a C6/LiFePO4 battery at 60 °C and, extracted from these results, the extrapolated EMF-curve (dotted line). Dependent on the SoC, the extrapolation is performed at either constant SoC (vertical extrapolation) at the beginning of discharging or at constant voltage (horizontal extrapolation) at the end of

View Online

Chapter 9

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

230

Figure 9.4  Set  of voltage discharge curves (solid lines) obtained during the char-

acterization process of an LFP battery at 60 °C with various indicated discharge currents. The extrapolated EMF curve is represented by the dotted curve. The insets show an example of the vertical (a) and horizontal (b) extrapolation procedure.

the discharge process. The insets show an example of such a vertical (a) and horizontal extrapolation (b) procedure. Linear relationships are observed between the current and voltage in the vertical direction and parabolic relationships are observed between the current and extracted amount of charge (Qout) in the horizontal direction. Since the EMF curve describes the equilibrium voltage as a function of t SoC, the maximum storage capacity of the battery (Qmax ) is easily determined from these extrapolated EMF curves, as indicated in Figure 9.4. The obtained t values for Qmax represents the total amount of the cyclable Li ions inside the batteries. The irreversible capacity loss can be accurately calculated on the t basis of Qmax .

9.5  Calendar Aging The calendar life of Li-ion batteries is of major importance, especially in applications such as EVs and HEVs that reside in the so-called parking mode most of their lifetime. Moreover, in-depth investigation of aging mechanisms during storage is helpful in understanding the battery fading phenomena under other operating conditions too, such as cycling.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

Understanding Battery Aging Mechanisms

231

Figure 9.5  The  irreversible capacity loss (ΔQir) of LFP batteries stored at the indi-

cated SoC at 20 °C (a), 40 °C (b) and 60 °C (c). Black symbols represent ΔQir measured at SoC = 10%, blue symbols SoC = 50% and red symbols SoC = 100%. Note that the axes are kept the same for all temperatures. Reproduced from ref. 2.

Irreversible capacity loss (ΔQir) under storage represents the immobilization of cyclable Li ions and is mainly attributed to SEI formation on the graphite electrode. ΔQir is calculated according to the maximum capacities obtained from the EMF curves. The development of ΔQir for an LFP battery as a function of storage time at various SoC and temperatures is illustrated in Figure 9.5. It can be seen that ΔQir increases with increasing SoC. The influence of the temperature on ΔQir is, however, more significant than the dependence on SoC. The irreversible capacity losses at 60 °C are significantly accelerated due to the cathode dissolution and the subsequent metal deposition at the graphite electrode. Figure 9.6 shows an example of SEM images of both a pristine (a) and an aged graphite anode (b). Quite some changes in the graphite electrode can be seen after storage. The particle surface of the pristine graphite electrode is smooth and clean (a), while it becomes rough and vague after storage (b). The changes in the graphite electrode morphology are normally attributed to SEI formation. The thickness of the SEI layers can be detected by XPS analysis. The growth of a SEI layer on the graphite electrode is critical in determining the battery performance; however, the measurement of the SEI thickness still remains a challenge. XPS depth profiling provides an opportunity to analyze the growth of the SEI layers on graphite anodes. Figure 9.7 shows for example the evolution of C1s spectra of a graphite anode after storage for 6000 h at 20 °C (a) and 60 °C (b) as a function of sputtering time. The C1s spectra at a sputtering time t = 60 s (red) is selected to discuss the growth of the SEI layer under different storage conditions. The intensity of the graphite C1s spectra at a sputtering time of 60 s is quite high, as indicated in Figure 9.7a, and becomes much weaker, as shown in Figure 9.7b. Therefore, it is concluded that the thickness of SEI layers on the graphite electrode increases as a function of temperature.16 The thickness of the SEI layer is proportional to the total irreversible capacity losses. The results observed in Figure 9.7 show that the irreversible

View Online

Chapter 9

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

232

Figure 9.6  SEM  images of the graphite electrode before (a) and after (b) storage at 60 °C for 6000 h. Reproduced from ref. 2.

Figure 9.7  C  1s spectra obtained from the graphite anode after storage for about 6000 h at 20 °C (a) and 60 °C (b). Reproduced from ref. 16 with permission from Elsevier, Copyright 2016.

capacity losses are larger at elevated temperatures than at lower temperatures, which is in line with the storage results shown in Figure 9.5. It should be noted that after rinsing some of the SEI layers can be dissolved or detached. Therefore the real thickness of the SEI layers might be thicker than that detected by XPS.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

Understanding Battery Aging Mechanisms

233

Figure 9.8  Fe  2p spectra obtained from the dismantled graphite anode after storage of A123 batteries at different temperatures. Reproduced from ref. 16 with permission from Elsevier, Copyright 2016.

Cathode dissolution occurs when batteries are stored at elevated temperatures. These dissolved transition metal ions can transfer to the anode and deposit on the anode surface. The iron deposition on the graphite anode has been reported before.16 Figure 9.8 shows the case of Fe 2p spectra obtained after 60 s sputtering of the dismantled graphite anodes, which have been stored for 6000 h at different temperatures at SoC = 10%. The peak at 707.6 eV is observed for the electrode stored at 60 °C, and has been assigned to the 2p3/2 peak of metallic iron. The Fe 2p3/2 core level peak is at 707.6 eV compared to a theoretical value of 706.8 eV for metallic Fe and 710 eV for Fe2+ ions.48 The slight shift towards higher binding energies as compared to metallic Fe suggests that the Fe atoms have some interactions with the SEI material, e.g., donating charge to the atoms having high affinity energies, such as O and F.49

9.6  Cycling-induced Aging 9.6.1  Irreversible Capacity Loss (ΔQir) Many studies have been carried out to investigate the cycling performance of Li-ion batteries in order to understand the underlying aging mechanisms. For example, Li et al.18,20,50 investigated the battery capacity degradation by both experiments and simulations. Dubarry et al.51 studied battery aging on the basis of dQ/dV analysis and Safari et al.17 investigated the capacity fade of LFP batteries at different cycling temperatures. Christensen et al.52 discussed the battery capacity loss by mathematical simulations and Smith

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

234

Chapter 9

Figure 9.9  Irreversible  capacity losses (ΔQir) of A123 (2.3 Ah) batteries as a function of cycle number (a) and time (b) at various cycling currents at 60 °C.

et al. investigated battery aging at low currents and elevated temperatures.53 Cyclable Li-ion losses and electrode material decay are considered to be the most important degradation mechanisms. In this section we will describe the degradation of LFP batteries under various cycling conditions, such as current, temperature, as well as cycle ranges. Usually, capacity losses are plotted as a function of cycle number, as can be seen in Figure 9.9a. From Figure 9.9a it can be concluded that the capacity loss increases with decreasing current. However, the conclusion becomes the opposite when the same results are plotted as a function of time (Figure 9.9b). This is because both cycle number and time can influence capacity losses independently. The individual influences of the cycle number and time on ΔQir were firstly discussed by Li et al.20 A mathematical extrapolation method was used to distinguish the calendar aging and the cycling-induced aging. The method is illustrated by Figure 9.10. From the 3D plot (Figure 9.10) it can be seen that ΔQir clearly increases with both cycle number and cycling time, as indicated by the red and blue arrows, respectively. The influence of the calendar time on the total irreversible capacity loss is identified with a mathematical extrapolation towards zero current. The corresponding irreversible capacity losses obtained ca by this extrapolation method is defined as calendar ageing (ΔQca ir). ΔQ ir depends, in a logarithmic way, on time but not on current or cycle number (see the red region in Figure 9.10a). Similarly, the influence of the cycle number on the total irreversible capacity loss is identified by extrapolation towards zero time, as is indicated by the red region in Figure 9.10b. The irreversible capacity losses due to cycling have been attributed to SEI formation on the graphite anode and the volumetric changes of this electrode upon (dis)charging. Capacity losses related to this new SEI formation process have been denoted as QcrSEI. Remarkably, a linear relationship

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

Understanding Battery Aging Mechanisms

235

Figure 9.10  Three-dimensional  representation of ΔQir at 0.1, 0.5, 1 and 2 C and plotted as a function of cycle number and time. Capacity losses at I = 0 cr C corresponds to ΔQca ir (a) and at t = 0 corresponds to Q SEI (b).

Figure 9.11  Irreversible  capacity losses (ΔQir) as a function of cycle number at various temperatures.

between QcrSEI and cycle number n is found in Figure 9.10b. The mathematical extrapolation technique, illustrated by Figure 9.10, offers a new simple and accurate method to distinguish between the individual contributions of calendar ageing and cycling-induced ageing during long-term cycling experiments. The influence of temperature on ΔQir is shown in Figure 9.11. Obviously, the capacity losses increase with increasing temperature. The cycling window also has a considerable influence on battery capacity loss, as can be seen in Figure 9.12. ΔQir obtained in the high SoC window is larger than that in the low SoC window at moderate cycling conditions.2 However, it has been

View Online

Chapter 9

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

236

Figure 9.12  The  irreversible capacity loss ΔQir in various cycling windows (0–30%, 35–65% and 70–100%) at 60 °C as a function of cycle number.

also reported that the capacity degradation becomes faster at a low SoC window after long-term cycling at elevated temperatures, when electrode degradation becomes more dominant.50

9.6.2  Postmortem Analyses The thickness of the SEI layers grown at various temperatures was analyzed by XPS depth profiling. Figure 9.13 shows the C1s spectra of the graphite anodes dismantled from the batteries after cycling at different temperatures. The C1s peak of C6 is located at 284.6 eV. Peaks located at higher binding energies are assigned to various components of the SEI layers. For example, the peak at 286.5 eV can be assigned to –(CH2CH2O)n–, which at 287.6 eV corresponds to CH3OLi, and the peak at 291 eV has been attributed to Li2CO3 or R–CH2OCO2Li.54 The intensity of the C1s spectra of C6 at a sputtering time of 60 s decreases as a function of temperature, indicating that the thickness of SEI layers on the graphite electrode increases as a function of temperature. Apart from temperature, the influence of the cycling SoC windows on SEI growth has also been investigated.2 It was reported that the thickness of the SEI layer is higher in the higher SoC window, which is in line with the trend of irreversible capacity losses.50 Similar to the storage experiments at elevated temperatures, cathode dissolution has also been observed in cycling conditions at 60 °C. The dissolved transition metal ions can be transported through the electrolyte and they subsequently deposit onto the anode leading to accelerated battery capacity losses. Figure 9.14 shows the development of the Fe 2p spectra (707.6 eV for Fe 2p3/2) of the graphite electrode cycled at 60 °C at different sputtering

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

Understanding Battery Aging Mechanisms

237

Figure 9.13  The  C1s spectra of the dismantled graphite anodes cycled at 20, 40 and 60 °C at 0.1 C after sputtering for 60 s.

Figure 9.14  Fe  2p spectra collected for a dismantled graphite anode after cycling at 60 °C with a current of 0.5 C at various sputtering times.

times. Hardly any iron is detected at t = 0 s, while the intensity becomes significant after t = 60 s, indicating the existence of SEI layers on the graphite electrode. However, the intensity of Fe 2p spectra decreases after longer sputtering times due to approaching the anode surface. It has been concluded that the iron clusters are embedded inside the SEI layers, covering the graphite electrode surface. Precipitated iron partially blocks the graphene layers leading to the inaccessibility of graphite particles. The degradation of the graphite electrode after cycling at various conditions has been confirmed by Raman spectra. The D and D′ bands in Raman

View Online

Chapter 9

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

238

Figure 9.15  Raman  spectra of the cycled graphite anodes cycled at 0.1 C at vari-

ous temperatures. The Gauss function was used to fit the peaks (bold lines). The intensities are normalized with respect to the G band. Reproduced from ref. 20 with permission from Elsevier, Copyright 2018.

spectra imply the defects of the graphite anodes. The higher the D band, the stronger graphite degradation will be. Li et al.20 reported the graphite anode degradation under various cycling temperatures, as shown in Figure 9.15. It was found that the structural degradation of graphite anodes at high-temperature cycling is stronger. Apart from the cycling temperature, the cycling SoC window has also been found to have a significant influence on the graphite degradation. It is reported that the graphite decay is stronger in the low cycling SoC windows.36

9.6.3  Non-destructive Approach Electrode degradation is one of the most important aspects of battery aging. Using reference electrodes is considered to be an effective way to determine the graphite anode capacity. However, a battery has to be opened in order to position a reference electrode inside the electrode package, which will influence the cycling performance. Li and coworkers50 proposed a non-destructive approach, which is based on the analyses of the plateaus on the battery EMF derivative (dVEMF/dQ) curves, to quantitatively determine the graphite electrode capacity QC6. The three plateaus on dVEMF/dQ curves are denoted as plateau I, II and III, as shown in Figure 9.16. The shrinkage of the complete dVEMF/dQ curve represents the total irreversible capacity loss (ΔQir) and the changes in the width of the second plateau on the dVEMF/dQ

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

Understanding Battery Aging Mechanisms

239

Figure 9.16  The  development of dVEMF/dQ curves with peak alignment of the first peak at about 0.8 Ah at 2 C rate at 60 °C as a function of the various indicated cycle numbers. Reproduced from ref. 2.

curves (ΔQLi,II) are considered to be an indicator of the graphite degradation according to50   



Q QC6 Li , xII

(9.7)

 x  QLi,I  Qir   1  III  QLi,II , xII  

(9.8)

   where xII represents the ratio of the second graphite plateau to the total graphite capacity. Apart from region II, a decline in region I is also observed, as indicated in Figure 9.16. The decline of regions I and II (ΔQLi,I, ΔQLi,II) provides an insight into the understanding of the aging mechanisms according to   



   see ref. 50. From eqn (9.8) several cases have been distinguished, which are related to different aging mechanisms:    (i) Plateau II does not change. This refers to cases where ΔQLi,II = 0, indicating that no graphite electrode decay takes place upon cycling. However, ΔQLi,I (= ΔQir) is significant in these cases. The decrease in ΔQLi,I can be attributed to the Li-immobilization process, trapping Li ions irreversible inside the SEI layer at the graphite electrode surface.

View Online

Chapter 9

240

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

(ii) Plateau II decreases. This refers to cases that ΔQLi,II > 0. Referring to eqn (9.8), three distinct cases can additionally be distinguished:  x a. When Qir   1  III xII  to decrease.

  QLi,II, then ΔQLi,I > 0 and plateau I is expected 

 x  b. When Qir   1  III  QLi,II, then ΔQLi,I < 0 and plateau I is expected xII   to increase, as has been found in ref. 50. The increase of plateau I will finally lead to Li plating on the graphite anode surface when the graphite anode becomes the capacity-limiting electrode in Li-ion batteries.  x  c. When Qir   1  III  QLi,II, plateau I remains constant (ΔQLi,I = 0). xII   Only the decrease in plateaus II and III can be observed.

   Graphite degradation is considered to be induced by (i) structural deformation of the graphite anode and (ii) iron deposition onto the graphite anode. Structural deformations are generally attributed to the volume changes in the graphite anode during cycling.38 These volume changes induce stress, which leads to a poor connection between the graphite particles as well as cracking of the graphite particles. The isolated/deformed graphite particles therefore become inaccessible and this will result in an increase in ΔQC6.

9.6.4  Summary The development of the total battery capacity and the individual electrode capacities are illustrated by Figure 9.17. The maximum storage capacity of 0 ) is considered to be equal to that of the cathode (QLiFePO4) the battery (Qmax as the capacity of the anode (QC6) is oversized. A decline in graphite anode capacity (ΔQC6), caused by electrode blockage or deformation, is indicated in Figure 9.17. The capacity losses of the cathode, ΔQLiFePO4, caused by iron dissolution, are considered to remain negligibly low and are therefore not indicated in the figure. It is well known that during the activation process, commenced after the LFP batteries have been manufactured, the SEI layers grow on the graphite anode. Consequently, lithium ions are immobilized in the SEI. This process continues during the subsequent operations (storage and/or cycling). The lithium immobilization process results in a reduction of the reversible capacity indicated by ΔQir. The total reversible capacity Qtmax is indicated by the red area in Figure 9.17. It has been reported that the graphite anode capacity decreases faster than the battery capacity, especially at higher temperatures.20 The second slope on the battery capacity degradation can be related to the case that the graphite electrode becomes the capacity-limiting electrode, i.e. QC6 − ΔQC6 < Qtmax.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

Understanding Battery Aging Mechanisms

241

Figure 9.17  Schematic  representation of the development of the total battery

capacity and the individual electrode capacities. Reproduced from ref. 16 with permission from Elsevier, Copyright 2016.

9.7  M  odeling Degradation Mechanisms of Li-ion Batteries It has been unraveled that the SEI formation on the anode is the major cause of battery capacity losses.55 Although considerable efforts were made to study the SEI experimentally, the understanding of lithium immobilization is still incomplete because the SEI formation reaction is complex, and highly dependent on the electrode voltage, electrode surface morphology and the composition of the electrolyte. Modeling can be a powerful tool to study the SEI formation process. Many mathematical simulations have been carried out on the basis of various proposed mechanisms. The first mathematical simulations of passivation films associated with the metallic Li electrode in an aqueous electrolyte was reported in 1976 by Bennion et al.56 Afterwards an electron-tunneling-based mathematical model was proposed by Peled to describe the SEI formation on the Li electrode in a non-aqueous electrolyte.57 Peled concluded that the SEI grows as a function of t1/2. This conclusion has been widely accepted to estimate the battery capacity loss caused by SEI formation. Christensen and Newman58 developed a continuum-scale mathematical model to simulate the growth of the SEI and transport of Li+ ions and electrons through the surface films. They attempted to combine the relevant mechanisms for the growth in a more general oxide-growth model that includes transport of cationic and anionic vacancies and interstitials as well as electrons through the film but with explicit emphasis on Li-ion systems with a graphic anode. Ploehn et al.59 proposed a continuum mechanic model with the assumption that a reactive solvent component diffuses through the SEI and undergoes two-electron reduction at the graphite surface. Pinson and Bazant60 developed a single-particle model that was further extended to a porous electrode model to describe the SEI formation and battery fading mechanisms. However, the model in that stage did not consider information about the composition of

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

242

Chapter 9

the SEI layers, which indeed plays an important role in determining the SEI formation mechanisms. Apart from SEI formation, cathode dissolution at elevated temperatures is considered to be another important process during battery aging.12,16,20,22–24 The dissolved transition metal ions can be transported to the anode and can subsequently be deposited on the graphite surface.16,20 Both the metal dissolution and the subsequent reduction can directly lead to a decrease of the battery capacity. Furthermore, the deposited metal clusters will speed up the SEI development by facilitating the transport of electrons.12,16,20 Li et al.18,19 further developed a comprehensive electrochemistry model by considering both the SEI formation and the cathode dissolution. The cycling-induced effects and the catalyst effects from the deposited metallic clusters are also explained in this model. The degradation mechanisms inside a Li-ion (LFP) battery are illustrated in Figure 9.18. The SEI formation is initiated when the voltage of the graphite electrode drops below approximately 1.0 V vs. Li+/Li.61 The SEI is formed during the activation procedure after the battery manufacturing process has been completed. The quality of the SEI formed in this period determines the battery cycling performance in its forthcoming usage. An ideal SEI layer formed during the activation procedure can dramatically decrease further the SEI formation rate and, therefore, maintain a high battery coulombic efficiency. However, the SEI layers normally continue to grow during battery usage, leading to irreversible capacity losses.

Figure 9.18  Schematic  representation of SEI formation on the graphite electrode, transition metal dissolution from the cathode and the subsequent deposition on the anode.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

Understanding Battery Aging Mechanisms

243

It is commonly accepted that the SEI layers are composed of inorganic and organic salts. These components constitute the dense inner layer and porous outer layer, respectively, as schematically represented in Figure 9.18. The inner SEI layer is considered to be an insulator for electrons. It also prevents solvents from passing through and co-intercalating into the graphene layers. Therefore, the inner layer plays an important role in protecting the graphite anode and limiting the storage capacity losses. Although the inner SEI layer is a good insulator, electrons can still tunnel through it when its thickness is sufficiently small ( Ef(50%) > Ef(10%) and ΔE(100%) < ΔE(50%) < ΔE(10%). Therefore, the capacity loss at high SoC is larger than that at low SoC. The influence of temperature on capacity loss includes two aspects. On the one hand, it affects the thickness of the inner SEI layer, which substantially

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

Understanding Battery Aging Mechanisms

245

Figure 9.21  The  development of the surface area of the precipitated Fe particles

on the graphite anode (a) and the capacity loss (QSEI,Fe) caused by SEI formation at the Fe surface at the various indicated SoC values upon storage (b).

determines the SEI formation rate. On the other hand, cathode dissolution occurs at elevated temperatures, which can lead to metal deposition at the anode, thereby accelerating the SEI formation. Figure 9.21a shows the development of the surface area of the precipitated Fe particles (AFe) on the graphite anode as a function of storage time. The cathode dissolution is only determined by the environment temperature and the H+ concentration in the electrolyte, independent of the testing conditions, such as SoC and cycling current. The SEI layers can also be formed on these Fe particles, leading to accelerated capacity loss (QSEI,Fe). The development of QSEI,Fe as a function of storage time at various SoC values at 60 °C can be seen from Figure 9.21b. QSEI,Fe increases with increasing SoC since the electron tunneling barrier on the Fe surface is smaller at higher SoC. Due to the volumetric changes in the graphite anode, the SEI formation during cycling has been classified into two cases. As shown in Figure 9.18, one process is the SEI formation on the covered surface areas Acov, denoted as Qcov SEI, which is time dependent and has been described by the equation describing SEI formation on C6 in Table 9.2. The other process is SEI formation on the fresh surface areas Afr caused by cracks, denoted as Qcr SEI, which is cycle-number dependent and has been described by the equation describing SEI formation on cracks during cycling. From Figure 9.22 it can

View Online

Chapter 9

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

246

Figure 9.22  Experimental  (symbols) and simulated total capacity losses ΔQcyir (black line) and the contribution of the covered SEI Qcov SEI (blue line) and the SEI due to cracks Qcr SEI (red line) as a function of cycle time at 1 C rate at 40 °C.

cr be seen that Qcov SEI increases logarithmically with t, while Q SEI increases linearly with t. Generally, the capacity loss under cycling conditions is larger than that under storage due to the contribution of Qcr SEI. Experimental detection of the SEI layer thickness is difficult due to its porous and fragile structure. Before conducting SEI characterization, the graphite anode is usually rinsed to remove the Li salt crystalized from the electrolyte. However, it has also been reported that the outer SEI layers can be partially dissolved in the electrolyte/solvent, which makes the measurements inaccurate. The simulations based on SEI models provide an efficient way to accurately estimate the growth of both the inner and outer SEI layers. Figure 9.23 shows the simulated growth of the inner and outer SEI layers upon storage at various indicated SoC values (a), and upon cycling at various indicated currents (b) at 20 °C. The thickness of inner layer was measured by Edström,54 using XPS. According to this study, the inner SEI layer thickness is around 20 Å, which is in good agreement with the simulated results. It is worth noting that the simulated SEI-layer thickness in Figure 9.23 should be regarded as an average thickness. Obviously, the thickness of the outer layer is much larger than that of the inner layer.

9.8  Future Outlook Aging is a common challenge for all Li-ion chemistries. In-depth understanding of aging mechanisms is of vital importance for the optimization of the manufacturing procedures and the improvement of the battery

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

Understanding Battery Aging Mechanisms

247

Figure 9.23  The  development of the inner and outer SEI layers upon storage

(a) and cycling (b) at various indicated SoC values (a) and currents (b) at 20 °C.

performance. Battery modeling provides a powerful tool to quantify the aging processes under various usage conditions. Based on the presented aging models, people can easily estimate the battery calendar and cycling life, design accurate battery management systems (BMSs) and produce a new generation of batteries with higher performance, which will finally facilitate the wide application of electric vehicles.

References 1. A. K. Padhi, K. S. Nanjundaswamy and J. B. Goodenough, J. Electrochem. Soc., 1997, 144, 1188–1194. 2. D. Li, Chemical Engineering and Chemistry, Eindhoven University of Technology, Eindhoven, 2017. 3. J. B. Goodenough and Y. Kim, Chem. Mater., 2010, 22, 587–603. 4. J. B. Goodenough and K. S. Park, J. Am. Chem. Soc., 2013, 135, 1167–1176. 5. J. W. Choi and D. Aurbach, Nat. Rev. Mater., 2016, 1, 1–16. 6. http://batteryuniversity.com/learn/article/types_of_battery_cells, in, Battery University, 2017. 7. J. M. Tarascon and M. Armand, Nature, 2001, 414, 359–367.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

248

Chapter 9

8. http://www.epectec.com/batteries/prismatic-pouch-packs.html, in, Epec, 2017. 9. D. Guyomard and J. M. Tarascon, Solid State Ionics, 1994, 69, 222–237. 10. P. Arora, R. E. White and M. Doyle, J. Electrochem. Soc., 1998, 145, 3647–3667. 11. R. Yazami and Y. F. Reynier, Electrochim. Acta, 2002, 47, 1217–1223. 12. K. Amine, J. Liu and I. Belharouak, Electrochem. Commun., 2005, 7, 669–673. 13. K. Amine, Z. H. Chen, Z. Zhang, J. Liu, W. Q. Lu, Y. Qin, J. Lu, L. Curtis and Y. K. Sun, J. Mater. Chem., 2011, 21, 17754–17759. 14. N. Yabuuchi, K. Yoshii, S. T. Myung, I. Nakai and S. Komaba, J. Am. Chem. Soc., 2011, 133, 4404–4419. 15. S. K. Jung, H. Gwon, J. Hong, K. Y. Park, D. H. Seo, H. Kim, J. Hyun, W. Yang and K. Kang, Adv. Energy Mater., 2014, 4. 16. D. J. Li, D. L. Danilov, J. Xie, L. Raijmakers, L. Gao, Y. Yang and P. H. L. Notten, Electrochim. Acta, 2016, 190, 1124–1133. 17. M. Safari and C. Delacourt, J. Electrochem. Soc., 2011, 158, A1123–A1135. 18. D. J. Li, D. Danilov, Z. R. Zhang, H. X. Chen, Y. Yang and P. H. L. Notten, J. Electrochem. Soc., 2015, 162, A858–A869. 19. D. Li, D. Danilov, B. Zwikirsch, M. Fichtner, Y. Yang, R.-A. Eichel and P. H. L. Notten, J. Power Sources, 2018, 106–117. 20. D. J. Li, D. L. Danilov, L. Gao, Y. Yang and P. H. L. Notten, Electrochim. Acta, 2016, 210, 445–455. 21. W. Choi and A. Manthiram, J. Electrochem. Soc., 2006, 153, A1760–A1764. 22. H. C. Wu, C. Y. Su, D. T. Shieh, M. H. Yang and N. L. Wu, Electrochem. Solid-State Lett., 2006, 9, A537–A541. 23. M. Koltypin, D. Aurbach, L. Nazar and B. Ellis, Electrochem. Solid-State Lett., 2007, 10, A40–A44. 24. J. J. Wang, Y. J. Tang, J. L. Yang, R. Y. Li, G. X. Liang and X. L. Sun, J. Power Sources, 2013, 238, 454–463. 25. D. H. Jang, Y. J. Shin and S. M. Oh, J. Electrochem. Soc., 1996, 143, 2204–2211. 26. Y. Y. Xia and M. Yoshio, J. Electrochem. Soc., 1996, 143, 825–833. 27. D. Aurbach, B. Markovsky, G. Salitra, E. Markevich, Y. Talyossef, M. Koltypin, L. Nazar, B. Ellis and D. Kovacheva, J. Power Sources, 2007, 165, 491–499. 28. R. Robert, C. Villevieille and P. Novak, J. Mater. Chem. A, 2014, 2, 8589–8598. 29. H. Gabrisch, T. H. Yi and R. Yazami, Electrochem. Solid-State Lett., 2008, 11, A119–A124. 30. S. C. Yin, Y. H. Rho, I. Swainson and L. F. Nazar, Chem. Mater., 2006, 18, 1901–1910. 31. H. Gabrisch and R. Yazami, Electrochem. Solid-State Lett., 2010, 13, A88–A90. 32. Y. B. He, F. Ning, Q. H. Yang, Q. S. Song, B. H. Li, F. Y. Su, H. D. Du, Z. Y. Tang and F. Y. Kang, J. Power Sources, 2011, 196, 10322–10327.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

Understanding Battery Aging Mechanisms

249

33. M. Kerlau, J. A. Reimer and E. J. Cairns, Electrochem. Commun., 2005, 7, 1249–1251. 34. J. Shim, R. Kostecki, T. Richardson, X. Song and K. A. Striebel, J. Power Sources, 2002, 112, 222–230. 35. R. Kostecki and F. McLarnon, J. Power Sources, 2003, 119, 550–554. 36. V. A. Sethuraman, L. J. Hardwick, V. Srinivasan and R. Kostecki, J. Power Sources, 2010, 195, 3655–3660. 37. E. Markervich, G. Salitra, M. D. Levi and D. Aurbach, J. Power Sources, 2005, 146, 146–150. 38. Y. Qi, H. B. Guo, L. G. Hector and A. Timmons, J. Electrochem. Soc., 2010, 157, A558–A566. 39. Y. T. Cheng and M. W. Verbrugge, J. Appl. Phys., 2008, 104. 40. R. Deshpande, Y. T. Cheng and M. W. Verbrugge, J. Power Sources, 2010, 195, 5081–5088. 41. J. Christensen and J. Newman, J. Solid State Electrochem., 2006, 10, 293–319. 42. J. Christensen and J. Newman, J. Electrochem. Soc., 2006, 153, A1019–A1030. 43. H. J. Bergveld, W. S. Kruijt and P. H. L. Notten, Battery Management Systems, Design by Modeling, Kluwer Academic Publishers, Boston, 2002. 44. V. Pop, H. J. Bergveld, J. H. G. Op het Veld, P. P. L. Regtien, D. Danilov and P. H. L. Notten, J. Electrochem. Soc., 2006, 153, A2013–A2022. 45. V. Pop, H. J. Bergveld, P. P. L. Regtien, J. H. G. Op het Veld, D. Danilov and P. H. L. Notten, J. Electrochem. Soc., 2007, 154, A744–A750. 46. D. Danilov, R. A. H. Niessen and P. H. L. Notten, J. Electrochem. Soc., 2011, 158, A215–A222. 47. M. S. Rad, D. L. Danilov, M. Baghalha, M. Kazemeini and P. H. L. Notten, Electrochim. Acta, 2013, 102, 183–195. 48. P. C. J. Graat and M. A. J. Somers, Appl. Surf. Sci., 1996, 100, 36–40. 49. M. Fahlman, H. Guan, J. A. O. Smallfield and A. J. Epstein, Annual Technical Conference - ANTEC, Soc Plast Eng, Atlanda, Georgia, United States, 1998. 50. D. J. Li, D. L. Danilov, L. Gao, Y. Yang and P. H. L. Notten, J. Electrochem. Soc., 2016, 163, A3016–A3021. 51. M. Dubarry and B. Y. Liaw, J. Power Sources, 2009, 194, 541–549. 52. J. Christensen and J. Newman, J. Electrochem. Soc., 2005, 152, A818–A829. 53. A. J. Smith, H. M. Dahn, J. C. Burns and J. R. Dahn, J. Electrochem. Soc., 2012, 159, A705–A710. 54. A. M. Andersson, M. Herstedt, A. G. Bishop and K. Edstrom, Electrochim. Acta, 2002, 47, 1885–1898. 55. M. Kassem and C. Delacourt, J. Power Sources, 2013, 235, 159–171. 56. D. N. Bennion and E. L. Littauer, J. Electrochem. Soc., 1976, 123, 1462–1469. 57. E. Peled, J. Electrochem. Soc., 1979, 126, 2047–2051.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00220

250

Chapter 9

58. J. Christensen and J. Newman, J. Electrochem. Soc., 2004, 151, A1977–A1988. 59. H. J. Ploehn, P. Ramadass and R. E. White, J. Electrochem. Soc., 2004, 151, A456–A462. 60. M. B. Pinson and M. Z. Bazant, J. Electrochem. Soc., 2013, 160, A243–A250. 61. C. S. Wang, X. W. Zhang, A. J. Appleby, X. L. Chen and F. E. Little, J. Power Sources, 2002, 112, 98–104.

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00251

Chapter 10

Battery Storage for Grid Connected PV Applications M. Vetter*, S. Lux and J. Wüllner Fraunhofer Institute for Solar Energy Systems ISE, Department of Electrical Energy Storage, Heidenhofstr. 2, Freiburg, 79100, Germany *E-mail: [email protected]

10.1  Introduction Integrating large fractions of fluctuating renewable energy sources like photovoltaic or wind generators in the electricity grid requires a couple of measures to secure the feed-in of this power and to preserve the quality and availability of the power supply. Besides grid extension and development of demand-side management solutions as well as electrifying non-electrified sectors such as mobility and the heating market, electric storage systems have become crucial to achieve political goals of an electricity system based mainly on fluctuating renewable energies. As an example of such developments, Germany can be considered. By the end of 2017, Germany's electrical energy mix already included approximately 39% of renewables, related to power consumption.1 Thereby photovoltaic systems achieved a share of 7.2% with an installed capacity of 43 GWp. The actual political goals are defined as 65% renewable energy shares by 2030, related again to power consumption. For achieving this goal an additional 4–5 GWp of photovoltaic systems are necessary, annually.2

  Future Lithium-ion Batteries Edited by Ali Eftekhari © The Royal Society of Chemistry 2019 Published by the Royal Society of Chemistry, www.rsc.org

251

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00251

252

Chapter 10

Taking into account Germany's load curve, with a variation between 45 GW and 85 GW, it becomes obvious that further increase in fluctuating renewable energies can only be realized by integrating storage capacities. Thereby storage systems are needed in different sizes and for different purposes. Some of them will be installed decentralized, e.g. in combination with photovoltaic (PV) systems to increase self-consumption and self-sufficiency rates in residential as well as in commercial applications, some of them will be installed as district storage systems in distribution grids and some of them will be coupled directly with large PV and wind power plants. Additionally the highest share of renewables above 80% require seasonal storage.3 Therefore various types of technologies are needed to cover all these different applications. Figure 10.1 provides an overview of different options and a classification. In principle, the required stationary storage solutions can be divided into three classes, which are mainly defined by the typical discharge time and the energy to power ratio.4    ●● Short-term storage: seconds to minutes, energy to power ratio 10.    Within this chapter grid-connected PV applications for lithium-ion batteries will be considered. Thereby the focus lies on decentralized systems such

Figure 10.1  Comparison  of rated power, energy content and discharge time of different electrical energy storage technologies. Reproduced from ref. 7 with permission from Fraunhofer ISE, © by Fraunhofer ISE.

View Online

Battery Storage for Grid Connected PV Applications

253

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00251

as residential and commercial as well as district storage systems, as these segments are foreseen as the dominant markets in the coming years for stationary lithium-ion batteries in combination with PV.

10.2  R  esidential Behind-the-meter PV Battery Storage Systems Combining PV systems with battery storage becomes more and more economically justifiable as electricity end-user tariffs are increasing and solar feed-in tariffs are decreasing. As an example Germany can be considered: in the year 2018 residential end-user electricity tariffs have reached 30€-Cent kWh−1 while feed-in tariffs are only at a level of 11.84 €-Cent kWh−1 for new small PV roof-top installations.1 Such huge differences motivate end-users to invest in battery storage when the decision is made to install PV modules on their roof top. As a result more than 100 000 residential PV battery home storage systems have been implemented in Germany.4 It is expected that declining battery costs will accelerate the growth rate as well as retrofit of old PV systems, for which the feed-in tariff phase will terminate after 20 years of operation. Such retrofits will start, in a larger scale, in the years 2020 and 2021 with PV systems falling out of the 100 000 roof programme, which was implemented in 1999. Whereas the first installations of residential PV battery storage systems used lead-acid blocks, today lithium-ion technologies are dominating this market with almost 100% market share in Germany. Thereby lithium-iron phosphate and nickel manganese cobalt are widely used cathode materials and graphite represents the material of choice as anode material. The storage capacities of such residential applications vary between 2 kWh up to 10 kWh, but even 15 kWh are available on the market. As a rule of thumb for a justifiable system design it can be defined in Germany as: 1 kWh of usable storage capacity per 1 kWp of installed nominal PV power, based on “typical” residential end-user behaviour. In the following sub-sections three established and market available system concepts will be described.

10.2.1  DC Coupled PV Battery System In DC coupled system topologies for grid-connected PV battery applications typically two DC/DC converters are used to connect the PV generator as well as the battery modules with the intermediate circuit of a transformer-less inverter, as shown in Figure 10.2. Thereby the maximum power point tracking for the PV generator is performed by the DC/DC converter connected to the PV modules. The DC/DC converter for the battery modules should offer a wide range of input voltages to provide high flexibility in terms of the battery system design, which might vary in the used cell chemistries and cell formats. The advantage of this configuration is the cost reduction

View Online

Chapter 10

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00251

254

Figure 10.2  DC  coupled system configuration for grid-connected PV battery appli-

cations using a DC/DC converter for maximum power point tracking for the PV generator and a DC/DC converter for the battery pack. Both converters are connected to the intermediate circuit of the transformer-less inverter.

potential for the system components as only one full inverter and one or two DC–DC converters are needed. On the other hand the available power of the battery storage is limited by this inverter, which is typically dimensioned for the nominal power of the PV generator. Furthermore existing PV systems cannot easily be upgraded with storage systems without replacing power electronics.

10.2.2  AC Coupled PV Battery System Systems based on AC coupled topologies use two separate inverters for the PV generator and the battery pack, as shown in Figure 10.3. In the early days such battery inverters were operated only with 24 V or 48 V as DC input voltage and used an internal transformer to provide the corresponding AC grid voltage. Such inverters offer no satisfying efficiencies in partial load. Therefore, state of the art inverters use transformer-less topologies enabling the highest efficiencies over a wide operating range comparable to PV inverters. Installed PV systems can easily be complemented by battery storage at a later point in time without any adaptation. Due to the modular concept, the sizing of the battery pack is almost independent of the size of the PV system components like the PV inverter. Especially for retrofitting installed PV systems, which run out of the feed-in tariff schemes, e.g. in Germany after 20 years, this configuration offers its benefits. The main disadvantage of this system concept is the limited cost reduction potential, as two separate inverters for the PV generator and the battery storage are needed.

10.2.3  Generator Coupled PV Battery System A comparably new configuration allows the integration of battery storage into the conversion path of a PV system, as shown in Figure 10.4. During charging, PV power is converted via a step-down converter to the battery

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00251

Battery Storage for Grid Connected PV Applications

255

Figure 10.3  AC  coupled system configuration for grid-connected PV battery applications using two separate inverters for the PV generator and the battery pack. State of the art designs consist of inverters with transformer-less topologies.

Figure 10.4  Generator  coupled configuration, which allows the integration of battery storage directly into the conversion path of a PV system.5

system voltage. For discharging, a step-up converter is used and battery power is fed into the PV path towards the inverter. In principle this system concept can be used together with several market available PV inverters. Furthermore, existing PV systems can be retrofitted with battery storage easily on this basis. On the other hand warranty issues come along, as new components are integrated into the traditional PV path and the warranty of the PV inverter could be affected. Such issues have to be clarified before a decision is made towards this system configuration.

10.2.4  Energy Management The integration of battery storage in PV applications requires an energy management system, which takes care of reasonable and optimized operation. In behind-the-meter applications, as an example, the corresponding operating control strategies have to secure the maximization of the self-sufficiency degree. Therefore it is necessary that the whole system reacts fast enough

View Online

Chapter 10

256

Table 10.1  Settling  times of four market available PV home storage systems. Results from intensive laboratory tests.9

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00251

Settling time PV home storage

PV varying/load constant

PV constant/load varying

Product A Product B Product C Product D

8.37 s 45.14 s 20.61 s 6.48 s

8.11 s 38.33 s 23.37 s 6.03 s

to PV generation changes and load changes. Besides effective components, which are able to follow the corresponding set points, it is necessary that the communication on the field bus level for data acquisition of the actual PV generation and the actual load as well as the communication between the energy management, battery inverter and battery management system works fast enough. Otherwise the feed-in of PV electricity and purchasing of electricity from the grid occurs at the wrong time, which can destroy the economics of the PV battery storage investment as in most countries feed-in tariffs are much lower than end-user tariffs. Table 10.1 provides exemplarily testing results of commercially available systems showing a huge difference in the specific settling times.

10.3  C  ommercial Behind-the-meter PV Battery Storage Systems Commercial PV systems are already widely profitable. This results out of two main factors: on the one hand because commercial buildings normally have flat roof areas, which allows easy and cost effective integration of PV. On the other hand, office and various industrial facilities have daylight dominated load curves and correlate well with the power production curve of solar energy. Adding a battery system with comparably small capacities to this setup—accordingly to the relatively small mismatch between PV generation and consumption—opens additional use cases and revenue streams to ease the profitability of the overall system.

10.3.1  Load Balancing Industry with a certain annual MWh consumption is benefiting from industrial energy tariffs. The tariff scheme is split and consists of an regular energy part, which relates to price per MWh, as well as a peak demand charge, which depends on the highest power that was consumed in the billing period. Battery storage systems are used to compensate power peaks, e.g. inrush currents of electrical motors and machines, to avoid high peak demand charges. As these charges can be regularly quite high, compensation with batteries can be profitable. Another common application is the integration of battery

View Online

Battery Storage for Grid Connected PV Applications

257

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00251

storage solutions in combination with time-of-use (TOU) tariffs. In this case shifting of energy demand from high tariff periods to low tariff periods with the support of battery systems can lead to significant savings and can support the profitability of commercial battery systems.

10.3.2  Increased Resilience Apart from economic arguments for the use of battery systems, there can be grid requirements, such as resilience and reliability, as the main drivers for integration of PV battery systems in commercial behind-the-meter applications. In the context of special grid requirements there is to mention of autonomous operation capabilities, such as achieving independence from the grid power supply, operational reliability such as uninterruptible power supply (UPS) capability and black-start capability, to allow starting up a facility without an external grid supply available. These capabilities can be crucial for certain industry sectors, such as medical, military or public safety.

10.4  D  istrict Battery Storage Systems in Combination with PV Looking to the district level, there are also some specific use cases. Actually, there are two important ones to mention, which are described in the following sub-sections.

10.4.1  Sector Coupling On a district level sector coupling is getting more and more attention. Sector coupling means that electricity is merged with the most important other sectors (mainly transportation and heat supply). A typical example application for such sector coupling is power-to-heat, mostly by using electrical driven heat pumps, which are combined with heat storage units. As an example a project can be considered that was implemented in the city of Weinsberg in the South of Germany. The power and heat supply consists of distributed roof-top PV systems, a combined heat and power unit, heat pumps and battery storage (Figure 10.5). An innovative energy management system enables a high degree of electrical self-sufficiency of this district. Thereby feed-in of excess PV and combined heat and power unit (CHP) power into the public grid is minimized as well as purchasing of electricity from the public grid. Monitoring results have shown that electrical self-sufficiency rates of more than 90% can be achieved in this district.9 Another interesting example is coupling of electricity to the transportation sector, in the context of electromobility. The implementation of electrical charging infrastructure on a large scale, especially fast and super-fast

View Online

Chapter 10

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00251

258

Figure 10.5  Example  of an innovative district power and heat supply consisting of

distributed roof-top PV systems, a CHP, heat pumps and battery storage. Reproduced from ref. 6 with permission from Kruck and Partner GmbH, ©by Kruck + Partner GmbH & Co. KG.

charging stations often require buffer storage to allow the required high charging currents. Here again a profitable use for battery storage can be identified.

10.4.2  Grid Services In the case where district storage systems are primarily installed for buffering excess PV power, there are periods during the year with fewer operating hours as a result of lack of sunshine hours. In such periods the economics of storage investment can be improved, if additional revenues can be achieved by offering grid services. Actually a very prominent possibility is identified by offering primary control power. According to the numbers of grid-scale battery systems that were implemented in recent years in this market sector, multiple usage scenarios for medium-sized district storage units are getting more and more into the focus of project developers. As an example the project Flex4Energy can be mentioned, in which various flexibility options including district battery storage have been investigated.8

10.5  K  ey Factors Affecting Bankability of PV Battery Storage Projects For the majority of storage projects investors like to get proof of concept and try to minimize the risk of losses. The term “bankability” is used to assure a secure base for investment in a project. There are several factors that

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00251

Battery Storage for Grid Connected PV Applications

259

Figure 10.6  Key  factors affecting bankability and insurability of storage projects. Reproduced from ref. 9 with permission from Fraunhofer ISE, © by Fraunhofer ISE.

influence the bankability of a storage project, as shown in Figure 10.6. The main factors are safety, reliability, performance and cost. Along with lithium-ion technology, which can be considered as the leading battery technology, there is especially the risk of thermal runaway, fire and explosion. This might lead to a total loss of the investment. Proper cell selection along with sound engineering, providing functional safety for the overall storage system consisting of modules, switches, fuses, battery management and thermal management as well as battery inverter and energy management is a guarantee for safe operation. In terms of reliability it is crucial to consider the typical operating conditions of single applications—e.g. ambient conditions and typical power rates—as well as the system behaviour with aged components. Therefore the increase in the inner resistance as well as the capacity fade has to be taken into account in particular. To achieve long time performance 24/7 the system has to be reliable, thus power electronics has to be selected very carefully, taking into account also the partial load characteristics, and cells should provide a long life time. Investigations at Fraunhofer ISE have shown that a cycle life between 4000 and 7000 equivalent full cycles is achievable with commercial cells. Another factor is performance. Within the public funded project “Safety First”10 PV home storage systems have been assessed in the laboratories of Fraunhofer ISE to identify the status quo of market available products and to identify solutions for technical improvement. As an example Figure 10.7 shows the efficiency of energy conversion for commercially available systems in Germany, consisting of battery modules, battery inverter, PV inverter and energy

View Online

Chapter 10

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00251

260

Figure 10.7  Power  conversion path from battery to grid of four commercially available PV home storage systems.10

management. The performance of system A seems to be superior to system D. However, as around 80% of the energy is transferred at a normalized rate under 0.3 in a typical household, the round trip efficiency found for a three day profile test was 87.7% for system D compared to 85% for system A. This example shows that system design rules have to be assessed very carefully. Cost determination is an essential factor for bankability. Thereby the calculation should include the losses in the system and a proper system simulation has to consider capacity fade and resistance increase in the battery cells as the design life of storage systems and payback time are actually calculated between 10 and 20 years.

10.6  Conclusion In this chapter the need for electrical energy storage in future power grids with high shares of fluctuating renewables is highlighted. Various storage technologies for different tasks are already on the market or close to market entry. In particular lithium-ion battery systems play in this context a very important role. They can be designed in small units to store solar energy in the residential sector, offer a high degree of flexibility in commercial applications and in district power supplies. Furthermore they are installed in bigger units for providing grid services such as primary control power and by integrating them in huge PV power plants smoothing of the feed-in of volatile generation is enabled. In this context quality assurance is essential, which has to address safety, reliability and performance and comes along with techno-economical optimization of battery storage design and dimensioning. Therefore it is essential that all components of battery storage are taken into account including power electronics and energy management with appropriate operating control strategies.

View Online

Battery Storage for Grid Connected PV Applications

261

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00251

References 1. H. Wirth, Aktuelle Fakten zur Photovoltaik in Deutschland, http://www. pv-fakten.de. 2. Stromnetze für 65 Prozent Erneuerbare bis 2030. Zwölf Maßnahmen für den synchronen Ausbau von Netzen und Erneuerbaren Energien, Agora Energiewende, Juli 2018. 3. N. Hartmann, PhD thesis, University of Stuttgart, 2013. 4. http://www.pv-magazine.de/2018/08/28/100-000-photovoltaik-speicher-in-deutschland-in-betrieb-genommen. 5. J. Weniger, et al., Vergleich verschiedener Kennzahlen zur Bewertung der energetischen Performance von PV-Batteriesystemen, 32 Symposium Photovoltaische Solarenergie, Bad Staffelstein, March 8–10, 2017. 6. http://www.kruck-partner.de/files/wohnen_der_zukunft_image_ar.pdf. 7. Electrical Energy Storage – White Paper, IEC, December 2011. 8. https://www.flex4energy.de. 9. M. Vetter, Deploying electrical storage systems to reduce risks and drive bankability for renewable energy projects, presentation at VDE clean energy dialogue Africa, Cape Town, May 14, 2018. 10. S. Lux, et al., Results of project Safety First, funded by the German Ministry of Economics and Energy, 2018.

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00262

Chapter 11

Advancements in Manufacturing Emma Kendrick* School of Metallurgy and Materials, University of Birmingham, Edgbaston, B15 2TT, UK *E-mail: [email protected]

11.1  Introduction Lithium ion battery (LIB) manufacturing has been performed in a similar manner since its concept in the 1990s, based upon tape casting (also known as knife or doctor blade coating) of electrodes.1,2 This manufacturing process is quite different from that of alkaline batteries, which were a single anode and cathode packed and moulded into a cylindrical cell. A typical process is shown in Figure 11.1. The active material components of the anode and cathode are mixed in a solvent with a conductive additive, usually carbon black, and a polymeric binder. The resulting ink is then coated onto a metallic current collector, dried, calendared and then stacked or wound into cells with a separator between the anode and cathode. The precise manner in which mixing, coating, drying and calendaring of the electrodes is done has a huge impact upon the final performance of a lithium ion battery. The performance of a cell is essentially based around the optimised electronic and ionic conduction properties of the cell components. For the electrodes a balance between particle size, porosity and the electronically conductive carbon pathway is struck in order to obtain the 3D electronic and ionic conducting   Future Lithium-ion Batteries Edited by Ali Eftekhari © The Royal Society of Chemistry 2019 Published by the Royal Society of Chemistry, www.rsc.org

262

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00262

Advancements in Manufacturing

263

Figure 11.1  Schematic  summary of the manufacturing processes for lithium ion batteries. Adapted from ref. 4 with permission from Dove Medical Press.

network required for cell performance. The cell is optimised for the application as typically there is a trade-off between the energy and power density of a cell. For a high energy cell the components and design are optimised to maximise active material mass loadings, such that the highest energy density (volumetric and gravimetric) is obtained without compromising other cell performance properties. Whereas for power, thinner electrodes, coatings are typically utilised to reduce the distance required to move the ions and electrons from the front of the electrode to the current collector. Volumetric and gravimetric power and energy densities have different priorities depending upon the application. The mass of the components controls gravimetric densities whereas the absolute packing densities of the components control the volumetric capacities.3 After assembling, the final step of the manufacturing process is formation and conditioning; this creates the interfaces between the components in the cell and has an additional impact upon performance, capacities and lifetime. All of the manufacturing processes are interlinked, and have an impact upon the final performance properties of the lithium ion battery. This complexity and interplay of processes means, typically, a large matrix of experiments must be performed and the processes are empirically optimised. Many different coating techniques can be utilised; however, although tape-casting techniques have been utilised for some time, these processes are still producing the highest energy density

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00262

264

Chapter 11

cells. Innovations are likely to come from an increased understanding of the mechanisms behind these processes and the impact of these upon the final cell performance. Future research will bring together the modelling of these processes to be able to predict the effect of the manufacturing upon the final performance. The ‘holy grail’ for lithium (and sodium) ion battery manufacturing is to develop a complete understanding of the relationship between processing, design of cells and formation protocols, to achieve an optimised lithium ion battery performance with high energy density, high power density and long lifetime.3,4 This chapter discusses the current state of the art and future manufacturing capabilities of lithium ion batteries from materials to final cells.

11.2  Electrode Manufacturing Processes The components of a battery: current collectors, separator, electrolyte, electrodes, tabs, are all key for determining the gravimetric and volumetric energy density of a lithium ion battery. The precise physical nature of the electrode, including its porosity, homogeneity, density, contact between the conductive additives, binder, active materials and current collectors determines its true performance. In an optimised electrode coating, the 3D electronic and ionic conductivity pathways are maximised, and resistances between components, at interfaces, minimised. This is controlled within a cell through the manufacturing of the electrode, the adhesion of the composite electrode to the current collector and the 3D structure of the electrode, the particle size distribution of the active materials, together with the network of the electronic conductive pathways, and the ionic pathways through the pores of the electrodes into the active materials. The composition, mixing and coating methodologies all contribute to maximise the transport properties of electrons and ions. The methodologies for achieving this are discussed in this section. The performance of the electrodes is highly dependent upon many factors and interdependencies that are involved in the manufacturing process: materials, mixing, coating, drying and calendaring, as shown in Figure 11.2.

11.2.1  Materials In terms of the electrodes, a porous composite of the active materials, conductive additives and binders are required to adhere to the current collector. The choice of active materials is of course one of the key parts of the lithium ion battery, as they determine the working voltages, capacities and energy densities of the cells. Conductive additives are chosen by balancing cost versus performance. Usually micro-graphite and carbon black are used as baseline additives, but often mixed with low levels of carbon nanotubes or graphene, depending upon the manufacturer, to enhance the conductivity. These formulations are typically confidential know-how of the manufacturer.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00262

Advancements in Manufacturing

265

Figure 11.2  The  interdependencies of cell manufacturing and the effect upon

the physical and electrochemical properties and the lithium ion cell performance.

Polymeric binders add a resistive component into the electrodes and are therefore minimised wherever possible, whilst enough to ensure good adhesion and cohesion of the electrode components to each other and the current collector. The type of materials is just one aspect of the optimisation of a lithium ion cell performance. To maximise performance, packing densities are also optimised, For example, in order to maximise volumetric energy densities, the void space or porosity is minimised within the electrode whilst maintaining enough porosity to allow percolation of the electrolyte throughout the electrode coating, enabling a fast ionic transport of lithium ions to the active materials. To maximise the packing densities of the materials in an electrode, high tap density powders are used. This often means that there is a bi-modal particle size distribution of the supplied powder, which enables smaller particles to pack into the voids left by the larger particles, thus maximising the possible active material content in the electrode layers. As stated previously, there is typically a trade-off between power and energy density requirements. High energy density requires heavy loadings of active materials, leading to larger particle sizes, thicker coatings, and fewer conductive additives and binders. Whereas for high power, smaller particles and thinner coatings are typically used to reduce the diffusion path lengths and improve the kinetics. Small particles are also used for high capacity materials such as silicon and tin as they can accommodate large volume changes with less likelihood of cracking and pulverisation.5 However smaller particles require more conductive additives and binder to electronically wire all the active

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00262

266

Chapter 11

materials together, leading to increased surface areas and the relative importance of side reactions of electrolyte with material surfaces. This can cause higher interface resistances and gassing, which reduce the cycle life of the cell. A summary of the materials processing for LIB manufacture is summarised by Li and Ludwig.6,7 The types of cathode, anode and electrolyte materials have been reviewed extensively.5,8–13 However, for all these materials, the stability or ‘shelf-life’ is also an important consideration, in particular for the layered oxide type materials, but also for other cathode and anode types. The majority of the current lithium ion batteries are manufactured using lithium transition metal layered oxides and a graphite anode. The cathode layered oxides are prone to water absorption because of the spacing between the transition metal layers, where water can easily intercalate, leading to surface reaction of the active powders to form hydroxides and carbonates.14,15 The water absorption and surface decomposition leads to changes in the reactivity of the components during manufacture, where difficulties in producing stable inks for electrode coatings have been observed, and so this leads to shelf-life considerations of the powder components.16 With the stability of the powders in air being questionable, the shelf life of the electrodes must also be considered during manufacturing. An electrode shelf-life study by Jung et al.17 on electrode coatings for the layered transition metal oxides, NMC811 and NMC111 (LiNi0.8Mn0.1Co0.1O2, and LiNi0.33Mn0.33Co0.33O2) compared the degradation of the two electrodes in air over a period of time. The NMC111 material showed very little change, however significant deterioration of the NMC811 material was observed, with a combination of hydroxides and carbonates forming on the surface of the particles. Basic components, with water impurities in a NMP-PVDF based ink will cause an instability and the viscosity will change over time and eventually gel. One method utilised in overcoming small levels of basicity of these cathode inks is to add an acid into the formulation during the mixing process.18

11.2.2  Preparation of Inks and Mixing Methods The electrodes are typically deposited using a slurry based composite, the rheological properties of which are determined by the method of mixing and the components: active material, solvent, binder, conductive additive and dispersant or surfactant. To date much of the mixing is know-how within the manufacturers. However with the requirement for improved repeatability and yields there are new emerging technologies within this field. Some examples of the formulations are shown in Table 11.1. The battery industry has typically used PVDF/NMP based polymer–solvent systems, in particular for cathodes. Water based CMC-SBR systems are starting to be utilised due to the lower toxicity of the solvent system. It is preferable to utilise water based systems over organics due to the cost of solvent and the solvent recovery system, and also the toxicity of the solvents utilised.19 Aqueous processing is challenging, this is because water has a

Published on 14 March 2019 on https://pubs.rsc.org | doi

Type Active material

Conductive additives

Solvent

Binder

Other additives

Material Cathodes: LiNixMyCo1−x−yO2 (NMC) LiMn2O4 (LMO) LiNiO2 (LNO) LiCoO2 (LCO) LiFePO4 (LFP) Li2FeS2 (LFS) Electrical: Graphite Carbon black Acetylene black Carbon nanotube (CNT) Graphene, copper Organic: N-Methyl-2-pyrrolidone (NMP) Trimethyl benzene (TMB) Xylene Isopropyl alcohol Polyvinylidene fluoride (PVdF) Polytetrafluoroethylene (PTFE)

Anodes: Graphite (natural, synthetic) Hard and soft carbon Metal alloys/nitrides (Sn, Si) Si/Graphite blends Li4Ti5O12 (LTO) Lithium Ionic: Zeolites MOFs

Aqueous: Deionised water

Ratio 80–90 wt% (pure solid weight ratio)

0.5–10 wt% (pure solid weight ratio)

Advancements in Manufacturing

Table 11.1  Typical  components of electrode slurries.

Amount dependent on mixing and coating process

Polysaccharide binders e.g. sodium carboxymethyl 4–10 wt% cellulose (CMC-Na), alginate, chitosan, pectin, (pure solid ratio) guar gum with styrene-butadiene rubber (SBR) (1 : 1 or 1 : 1.5) Poly(acrylic acid) salts (PAA-X)

Polyethylene oxide (PEO) Ethylenepropylenediene monomer (EPDM) Acids e.g. oxalic, maleic, phosphoric, Surfactants (control dispersion/wetting) e.g. PEI, stearic, oleic (control slurry pH PAA, Triton X-100, Tergitol and porosity)

190 Wh/kg, in Sustainable Aircraft Symposium, 2016. 56. E. C. Darcy, Challenges with Achieving 180 Wh/kg Li-ion Battery Modules that Don't Propagate Thermal Runaway or Emit Flames/Sparks, in The Battery Conference 46, 2015. 57. C. Huber, Phase Change Material in Battery Thermal Management Applications, An assessment of efficiency and safety, 2017.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00290

Lithium-ion Battery Safety

315

58. N. Anderson, M. Tran and E. C. Darcy, 18650 Cell Bottom Vent - Preliminary Evaluation into its Merits for Preventing Side Wall Rupture, in S&T Meeting, 2016, pp. 1–53. 59. E. C. Darcy, D. P. Finegan and M. Keyser, Merits of the Cell Bottom Vent Feature in 18650 Cells for Preventing Side Wall Rupture, in Battery Safety Conference, 2017. 60. E. C. Darcy, Thermal Runaway Severity Reduction Assessment for EVA Li-ion Batteries, in NASA Aerospace Battery Workshop 49, 2014. 61. E. C. Darcy, Safety Benefits of the 18650 Bottom Vent for Future Space Battery Applications NASAs Future Applications, in The Battery Conference, 2016.

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

Chapter 13

Challenges and Opportunities in Lithium-ion Battery Supply Wolfgang Bernhart* Roland Berger GmbH, Löffelstr. 46, Stuttgart, 70597, Germany *E-mail: [email protected]

13.1  I ntroduction: Lithium-ion Cell Market and Players With the advent of lithium-ion technologies, new market segments for rechargeable batteries have been opened up, especially in the transportation sector. Consequently, the share of lithium-ion technologies has increased for a couple of years now, and reached a market size of more than €65 billion in 2017.1 The lithium-ion share of rechargeable battery market in 2017 already accounts for ∼23% of the energy storage capacity shipped (approximately 520 GWh in total, thereof ∼390 GWh in lead acid batteries), and ∼36% of its value (EUR). With ∼70 GWh, the automotive industry, with its segments commercial vehicles (light commercial vehicles, trucks and buses) and cars (PV: passenger vehicles, SV: shared vehicles), counts for more than 55%, followed by electronic devices with 26% and industrial energy storage systems (ESSs) with ∼5 GWh. Automotive demand is the main driver for the future lithium-ion battery market. Driven mainly by regulations in Europe and China,2 automotive original equipment manufacturers (OEMs) need to sell a significant share of their   Future Lithium-ion Batteries Edited by Ali Eftekhari © The Royal Society of Chemistry 2019 Published by the Royal Society of Chemistry, www.rsc.org

316

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

Challenges and Opportunities in Lithium-ion Battery Supply

317

Figure 13.1  Growth  in PC/SV sales mainly driven by BEVs, PHEVs and MHs— replacement of conventional powertrain types leads to higher electrification share and battery demand.

vehicles as electrified ones, either as battery electric vehicles (BEVs), plug-inhybrids that combine an internal combustion engine with an electric powertrain including a battery that can be recharged from the grid (PHEV), a full hybrid (FH, as promoted mainly by Toyota Motors) or a 48 V mild-hybrid (MH), Figure 13.1.3 As a consequence, the overall market for lithium-ion batteries (LiBs) from 2025, potentially also including substituting “post-lithium ion technologies” (PLiT) such as advanced cathode chemistries combined with Li–metal anode foils and solid-state electrolytes, will grow significantly (Figure 13.2). Our market model shows an increase to more than 300 GWh until 2020 and further growth to more than 1000 GWh in 2025 and up to 2100 GWh in 2030. From 2025, lithium-ion technologies will dominate the rechargeable battery market and will have surpassed lead acid, which might be in addition subject to substitution. In particular starter-generator batteries for automotive applications might be substituted from the first half of the next decade—further increasing the LiB market beyond the numbers shown in Figure 13.2. The automotive industry will therefore be the most important growth driver, accounting for >80% of the rechargeable battery market in 2025. Beyond 2025, fuel cells might, on the other hand, take over some of the LiB share.

13.2  The Lithium-ion Value Chain The rapid growth in LiB demand, especially for passenger car cells, leads to a market increase by factor 20 until 2025, from around 35 GWh in 2017 to around 740 GWh in 2025 (Figure 13.3).

View Online

Chapter 13

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

318

Figure 13.2  The  market for rechargeable batteries is growing rapidly—automotive by far largest growth segment.

Figure 13.3  The  lithium-ion value chain for cells and active materials is dominated

by Asian players—growth of passenger car segment creates a few lith­ ium-ion giants.

Cell manufacturing is dominated by a few large players from Asia, with China getting more important. In order to finance this capacity expansion, the industry has to invest between €60 and €70 billion for cell manufacturing capacities alone. Until then, driven by aggressive growth policies, the market will be dominated by a few Asian players: CATL, Lishen and BYD from China, LG Chem and Samsung SDI from South Korea, and Panasonic from Japan, the latter mainly supplying TESLA and Toyota.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

Challenges and Opportunities in Lithium-ion Battery Supply

319

Figure 13.4  Overview—major  automotive battery cell manufacturers. The market for active materials shows a similar picture: despite Umicore and the German–Japanese joint venture BASF-Toda, the market is dominated by Chinese, Japanese and Korean companies. While the battery-related revenue of major players was still in the range of a couple of hundred millions to around €3 billion in 2017 (see the characteristics of automotive cell manufacturers in Figure 13.4), the total capacity of key cell suppliers will increase from today at around 150 GWh to around 420 GWh in 2020.

13.3  Technology and Cost Development A major prerequisite for the anticipated market growth is driven by technology development, to further improve current lithium-ion technologies (“LiT”) and by changing the composition of the cell using Lithium-metal anodes and solid-state electrolytes.

13.3.1  Lithium-ion Technology and Cost Roadmap While currently cell chemistries used in automotive applications have a volumetric energy density of only 220 Wh L−1 to 400 Wh L−1, interviews with major cell manufacturers indicate a massive increase to up to 1000 Wh L−1 with new blends of nickel–cobalt and magnesium–aluminium from 2022 onwards (Figure 13.5, left). Although the general technological development trends are well known, recent developments of Ni- and Mn-rich materials (with lower cobalt share and high specific capacities) will potentially enable an extension of the period of technological competitiveness of LiT cells compared to polymer-/ ceramic-based “solid-state” PLiT cells from 2022 to 2026.

View Online

Chapter 13

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

320

Figure 13.5  Automotive  cell chemistry roadmap and cost development 2017 to 2025.4

Nevertheless the targeted “cell end-game scenario” is likely based on ceramic PLiT cells enabling energy densities of more than 1000 Wh L−1 at high safety levels, with fast-charging capabilities and no restrictions for performance temperatures. These will combine known cathode active materials (CAM) with Li-metal anodes, but potentially also enable the use of CAM with less critical raw materials (e.g. LiMnPO4—lithium manganese phosphate oxide) and/or 5 V cell voltage. Cell costs will level out around €75 KWh−1 in the second half of the next decade. Our bottom-up cost model that covers various cell form factors and chemistries shows that by 2025 it can be expected that the cost of automotive cells can be reduced to levels between €75 and €95, depending on the cell form factor, the cell chemistry used and the specific performance requirements, see Figure 13.5, right. Direct manufacturing costs (including equipment depreciation, personal and energy costs) today compromise approximately a quarter of cell costs. Material costs between prismatic, pouch and cylindrical cell mainly differ with respect of housing and contacts, while active materials account for roughly 60% of manufacturing costs (Figure 13.6). Increased volumetric energy density is main cost reduction lever, but manufacturing processes also provide further cost reduction opportunities. The increasing volumetric energy density of new cathode materials (as shown in Figure 13.5), combined with an expected move away from high cobalt content to nickel-rich cathode materials, has the biggest effect on driving down costs. In addition to improving the cathode material, the volumetric energy density of the anode also needs to be improved. Various

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

Challenges and Opportunities in Lithium-ion Battery Supply

321

Figure 13.6  Manufacturing  cost split of different cell types.

Figure 13.7  LiT  manufacturing process steps and selected process improvements. development efforts are therefore targeted at reducing first cycle losses (up to 15% for C and C/Si anode active materials) by lithiation of anode active materials prior to the coating process. As a rule of thumb, the process speed of bottleneck equipment (typically the coater, Figure 13.7) does not change significantly when the cell chemistry is changed. Doubling the volumetric energy density of a given cell therefore also doubles the output of a manufacturing line in terms of MWh (while the number of cells produced is constant). Figure 13.7 shows the typical manufacturing process and further improvement levers applied currently. Three main levers are in the focus of further improving current lithium-ion technologies.    1. Increased coating speed (to 100 m min−1), energy optimized drying process.

View Online

Chapter 13

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

322

2. Dry coating of electrodes to remove the major drying step, which is a key lever to reduce CAPEX, required space and energy consumption. 3. High-speed stacking for pouch cells, further optimization of assembly.    Process technology development will scale volume of Cathode active material processing units, causing a severe cost disadvantage risk to small sized players. CAM is produced in a continuous process. Small players faced a significant cost disadvantage to larger players, since plant operators typically secure a stable utilization with larger customers at lower prices. Today, even smaller cell manufacturers have increased significantly in size and therefore can guarantee a stable demand and an uninterrupted utilization of equipment (Figure 13.8). In line with increasing demand for cathode materials, it can be expected that larger reactors are being used in the future to scale up CAM production and reduce process costs. At that point, players with lower market shares (we expect that to be around 8–10% of global market) will face again significant cost disadvantages to larger cell manufacturers.

13.3.2  P  rospects of “Post-lithium” (PLiT) and Solid-state Technologies for the Automotive Market Further significant increases in the energy content of a cell can be reached by increasing the capacity of the anode side. While this is being done by blending silicon in the graphite material in a first step, a large increase can be achieved when using a Li–metal anode to further increase the energy density of the cell by increasing the anode's ion storage capabilities (in line with an

Figure 13.8  Potential  CAM pricing points depends on individual CAM demand and technology scale.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

Challenges and Opportunities in Lithium-ion Battery Supply

323

Figure 13.9  Comparison  of liquid electrolyte vs. solid-state electrolyte cell.

Figure 13.10  Thin  film solid-state cell. increasing energy density of the cathode). Independent of the usage in solid-state cells, we expect to see the first applications of Li–metal anodes with the introduction of NCMA in the 2023 timeframe. Solid-state cells still face major industrialization hurdles—no revolutionary technology changes for lithium-ion batteries is expected, rather a gradual shift of technologies. In a solid-state cell, the insulation layer, which consists of a semi-permeable separator between anode and cathode and a liquid electrolyte to support ion conductivity, is replaced by a polymer or ceramic electrolyte (Figure 13.9). Solid-state cells are already used as thin-film cells, e.g. for pacemakers. While they have an extremely long cycle life and high power capabilities, energy content is low and costs are high due to the vaporization processes used, Figure 13.10. Thin-film cells are therefore not a solution for automotive traction battery applications. Today, instead two basic approaches are being looked at: the polymer approach, where ion transport occurs by solvation and de-solvation, and the ceramic approach, where ion transport takes place by “ion hopping”. Current applications of “solid polymer” cells, such as the Bollore Group's lithium–metal polymer batteries are used, e.g. in the group's vehicles that are used for car sharing or buses,5,6 have the disadvantage of a working

View Online

Chapter 13

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

324

temperature range of 60 to 80 °C and little or no fast-charging capabilities. On the other hand, they have a longer cycle life which could—even at significant higher costs per kWh—lead to lower costs over the life cycle of the vehicle and therefore make it a favorable solution for these applications even today. While the higher operating temperature can also have some advantages (e.g. no need for cooling of the battery), it is also a major hurdle for wide usage in typical passenger car applications. Research is therefore being aimed at increasing conductivity, e.g. through “polymer-in-salt” materials, in which super ionic glass electrolytes are mixed with small quantities of polymers or “nanostructured solid polymers”. Ceramic approaches focus on using either sulfites or oxide ceramics. All three approaches still face major hurdles for industrialization for automotive usage (Figure 13.11). While the high operating temperature with a need for heating power during phases of no use are the main disadvantages specific to the polymer approach, in both the sulphite and the oxide ceramic approaches the usage of rare earth materials (e.g. germanium) and special coating needs for current NCA/NCM cathode formulations in order to handle critical interfaces for good ionic conductivity are the main challenges from the materials side. For all approaches, the Li–metal anode, which requires dry processing and operation in a completely dry environment, poses additional challenges. Solid-state cells are in the focus of academic and industrial research, with lots of patents issued to Asian OEMs Toyota and Hyundai, but also established cell manufacturers such as Samsung SDI and Panasonic. However, because

Figure 13.11  Solid-state  approaches and challenges for large-scale automotive industrialization.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

Challenges and Opportunities in Lithium-ion Battery Supply

325

Figure 13.12  Changes  in manufacturing process—example: polymer-based solid state.

of the still significant industrialization hurdles, and the potential limited advantages of both the polymer and the oxide approaches with respect to key performance indicators, we do see a broader application at the earliest after 2025. Depending on the approach used, manufacturing investments can be reused at least partially. Figure 13.12 shows an overview for polymer-based technologies, using Seeo, Inc's process as a basis.7 Major changes in manufacturing processes include the following.    ●● Replacement of graphite or silicone based anodes with Li–metal foil removes an entire coating line with additional positive impact on CAPEX and OPEX, at the same time, additional investments are needed to handle the lithium-metal foil and to coat the electrolyte on the electrodes, ●● Solid-state PLiT cells will not require any electrolyte filling as well as only a very much shorter formation process causing significant CAPEX savings.    While (especially ceramic) solid-state technologies might have advantages especially with respect to higher inherent safety of the cells, it is not so obvious that important performance criteria such as fast charging capabilities, significant higher energy density and lower costs can be fulfilled. Also it can be expected that already established players, who have invested a significant amount of money in at that time at least partially written-off manufacturing facilities, will react with lower prices towards the market entry of a new player betting on solid-state technology.

View Online

326

Chapter 13

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

13.4  D  emand and Supply of Pre-cursor and Raw Materials—Mining, Refining, Recycling With demand for lithium-ion batteries increasing significantly in the next ten years (Figure 13.2), supply and price risks are increasingly attracting the attention of all players. Demand for (battery-grade) nickel, cobalt and lithium will increase significantly. In order to understand these risks and to apply countermeasures, Roland Berger developed an integrated supply-demand model whereas material demand and supply for Ni, Co and Li products are calculated on the basis of market models and in-depth supply projection analysis (Figure 13.13). Within this model we look at demand from the automotive as well as from other battery applications to calculate refined product demand, and from other applications to calculate demand on raw and intermediate materials. In a first step, the yearly demand for battery cells in GWh needs to be calculated. The basis for this calculation is our long-term “base-case” forecast for the different xEV technology (battery electric, plug-in, full hybrid, mild hybrid), as shown in Figure 13.1, and certain assumptions regarding vehicle segment distribution, average energy consumption, expected range and resulting average battery pack size (see Figure 13.14 for assumptions regarding BEVs). We assume that, globally, the vehicle electricity consumption rate will be similar, while slight regional differences will exist for the full electric mileage. The battery pack capacity is then calculated by multiplying the average consumption and the expected mileage per segment and region. To calculate demand for raw materials, the change in, and the share of, different cell chemistries as outlined in the technology roadmap shown in Figure 13.5 plays a crucial role, since they have different material intensities (Figure 13.15).

Figure 13.13  Supply–demand  model for critical lithium-ion materials.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

Challenges and Opportunities in Lithium-ion Battery Supply

327

Figure 13.14  Example:  assumptions regarding average battery size of BEVs.

Figure 13.15  Material  intensity of battery cathode chemistries (g metal equivalent per kWh).

There is a clear trend towards nickel-rich materials such as NCM712 or NCMA. Also manganese-rich materials such NCM217 might serve as an alternative here. Besides an increase in the specific energy density, these materials offer a pathway to substitute or reduce the usage of the most critical

View Online

Chapter 13

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

328

Figure 13.16  Battery  cell demand for passenger cars by cell chemistry (GWh per annum).

material cobalt. The currently widely used NCM622 is an intermediate solution, as improvement from NCM523 and NCM611. NCA18650 (as used in the Tesla S) is replaced with NCA21700 with a lower Co and a higher Ni content, e.g. in the new Tesla models. Based on the expected trends in cathode material development as well as on interviews with market participants, we expect the battery cell demand by cell chemistry to change significantly to a much lower share of materials with high cobalt content (Figure 13.16). However, the massively increasing cell demand results in a similar increasing demand on the refined and raw material level. For passenger cars alone, compound average growth rates in cobalt, lithium and nickel demand amount to 35% to 50% per year until 2025, reaching approximately 25% (Ni), 80% (Co) and 280% (Li) of the total 2017 raw material production (Figure 13.17). As shown in Figure 13.17, substantial political country risks also exist, especially for cobalt. For all three commodities, resources are not only concentrated in a few countries (see pie-charts8 in Figure 13.17), but there is also a high concentration on the supply side. Price risks due to high concentration levels for raw and refined materials and political country risks need to be mitigated, but supply shortages are unlikely. For nickel, Vale and Norilsk Nickel account for more than 30% of the global supply. Other global players are: Glencore, Jinchuan Nickel, BHP Billiton, and Sumitomo Metal Mining. A limited number of global cobalt producers controls a large share of the market (∼70%): Glencore, Umicore, Freeport Cobalt, Sumitomo, Jinchuan, Sherritt International. The lithium supply market is

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

Challenges and Opportunities in Lithium-ion Battery Supply

329

Figure 13.17  Nickel,  cobalt, lithium: projected demand for xEV application and current production.

even more highly concentrated and primarily controlled by a small number of suppliers such as Albermale, SQM, FMC and Tianqi. The total supply of nickel has increased by ∼35% from ∼1600 kt in 2010 to ∼2100 kt in 2017. Further investments are being made by large market players to extend production capacities and cater for increasing future demand. For cobalt, a small market surplus is forecasted for 2020. However, further capacity expansion pretty much depends on additional nickel mining projects, where cobalt is often a by-product. Substantial investments are also expected in the upcoming years to increase supply capacities. In our model, we compared then the aggregated demand for batteries as well as for other applications with the expected supply. All mining and refinery projects in the DFS (design feasibility stage) or in the project feasibility stage, as they are known from various reports (e.g. by Roskill9), have been included. These calculations also take into account the average time to enter the LiB supply chain.10    ●● Mining: five years and longer from exploration to construction plus one to two years until production. ●● Conversion/refining: one and a half to two years, e.g. for new plant to produce stabilized lithium chemicals. ●● Precursor/cathodes: one to one and a half years for a new plant. ●● Cell manufacturing: two to three years after the project decision.   

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

330

Chapter 13

Figure 13.18  Price  and supply risks for Ni, Co and Li, 2017–2025. Missing mining capacities with its longest lead times create price (and potentially) supply risks driven by a potential supply-demand imbalance (Figure 13.18). For Ni-rich chemistries, only Nickel Grade I can be used. Due to high demand from lithium-ion batteries that require this higher quality, low investments because of low market prices for Nickel that are below cash costs of lots of mines, and recent mine shut-down in Phillipines, a supply-demand-deficit is expected from 2022 onwards. Significant price risks (with prices doubling compared to 2018) can therefore be expected. For cobalt, no supply deficit is expected. The speculative price increase in 2018 led to significant investments on both the mining and refinery side. Also additional supply is available from recycling. Therefore an only moderate increase in prices is expected. Lithium supply capacities are expected to exceed global Li demand until 2024 due to new lithium mines starting production. Price increases might lead to the installation of new conversion capacities so that the currently visible supply deficit in 2025 will likely not materialize. In order to mitigate price and supply risks, the industry has a couple of instruments available—from futures (which might be available mid-term also for some critical raw materials, as they are already for steel and other commodities) to stockpiling (Figure 13.19). The need for risk mitigation will impact cell costs to a certain extent—but at the same time, coordinated actions between the partners in the supply chain are necessary to avoid “over-hedging”, and also in order not to stimulate speculation as observed in the last 18 months for both cobalt and lithium prices. Recycling (“urban mining”) is becoming increasingly important—from a cost perspective, as well as from the perspective of securing raw material supply—but there are still challenges. Regulations on battery recycling for xEV exist only in the EU and Japan, but can be expected in other markets as well (Figure 13.20). While a “second life” of the batteries is often discussed at OEMs and even implemented in pilot programs, its long-term success can be doubted.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

Challenges and Opportunities in Lithium-ion Battery Supply

331

Figure 13.19  Strategic  options to mitigate price and supply risks for raw materials and refined products.

Figure 13.20  Regulations  status for xEV battery recycling (June 2018). Requirements towards cycle life are higher than in automotive applications. Cells specifically developed for usage in stationary systems, e.g. TESLA's Wallbox, therefore often use different materials (e.g. NCM111 instead of NCA), have more open electrodes and more electrolytes, and are therefore cheaper per charging cycle than a used automotive cell. To comply with upcoming regulations in a cost-effective way, OEMs therefore need to implement efficient recycling logistics and ensure effective recycling processes. With NCM622 battery cells and aluminum-housed modules, recycling in Europe can be done cost neutrally—logistic costs for the already

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

332

Chapter 13

dismantled battery pack to the recycling facilities and costs of recycling are covered by the value of the recycled metals. However, with the move towards lower cobalt content and cheaper materials, more efficient recycling processes are needed, and “design for disassembly” becomes much more important when designing battery modules and packs. A couple of approaches are either already in production or being researched (Figure 13.21 11,12).    ●● Pyrometallurgical approaches exist at industrial scale—there is already enough capacity for 2020 in the EU. ●● Hydrometallurgical approaches already work at pilot scale, with some companies developing in the EU and at industrial scale.    Hydrometallurgy has better economic prospects at scale, but both processes may struggle with LFP or other batteries with a low value of raw materials. Current gate fees are by far a larger source of revenue than the materials. Additional challenges also still exist.    ●● Long-term performance of recycled materials. ●● Optimized processes to deal with different cell chemistries. ●● Economic feasibility with declining value of cells/decreasing battery residual value. ●● Compliance with health, safety and environmental regulations. ●● Increasing battery volumes at low standardization.    For economic feasibility of LIB recycling, new business models need to be realized by doing the following.

Figure 13.21  Recycling  process routes—overview.   

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

Challenges and Opportunities in Lithium-ion Battery Supply ●● ●● ●● ●● ●● ●● ●● ●●

333

Optimization of existing processing technology. Scale-up and commercialization of lab-scale processing technology. Proof of concept of existing technical solutions. Reduction of costs as well as improvement of economic efficiency. Investments in, and industrialization of, xEV-LIB battery recycling. Realization of efficient collection system lithium-ion batteries. Implementation of safety standards for non-collected LIBs. Recirculation of recovered raw materials in the value chain.

   Recycling can also be a good lever to secure material, but needs to be embedded in clear strategy, since market entry of new players faces significant hurdles as well.    ●● Low volumes: today there is no significant return flow but a delay of 10–15 years (lifetime), a focus only on traction batteries today is therefore not economical. ●● Competition and overcapacities: for today's volumes there is already a capacity surplus with established companies having a competence advantage. ●● Ongoing technological development: today there is no dominating process, alternative approaches are possible, and high investments are necessary.

13.5  Implications and Conclusions Driven by the increasing demand from the automotive industry, lithium-ion batteries have made tremendous progress in the last few years—in terms of energy density, safety, fast-charging capabilities and costs. For the next decade, that progress will slow down—it will be evolution, not revolution. OEMs need to decide to what level of vertical integration they want to realize—whether to buy the complete pack, modules (as most OEMs currently do), cells or if they even go into their own cell production. The latter will only work for very large OEMs—with the move towards larger scale in active material production, a significant market share (between 10% and 15% of global capacity) is needed to secure competitive material costs or to be attractive for close cooperation with materials manufacturers.

References 1. C. Pillot, The Battery Show Europe Hannover 2018, Conference Proceedings, 2018, 5, 6. 2. Roland Berger internal paper. 3. Roland Berger internal paper. 4. H. Takeshita, B3 report 17-18, Chapter 11 LIB Materials Market Bulletin, 2018, 41, 44, 45, 51.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00316

334

Chapter 13

5. https://www.blue-solutions.com/en/, accessed June 25th, 2018. 6. http://media.daimler.com/marsMediaSite/en/instance/ko/Daimler-­ Buses-gives-an-outlook-on-the-2018-IAA- Commercial-VehiclesElectrification-of-the-public-transport-network-World-premiere-ofthe-all-electric-Mercedes-Benz-eCitaro-city-bus.xhtml?oid=40685587, accessed July 19th, 2018. 7. H. Zarem, U. Grape, Solid state batteries for grid-scale energy storage, 2014, 11. 8. U.S. Geological Survey, Mineral commodity summaries, 2018. 9. Roskill Information Service, Cobalt: Global Industry, Markets and Outlook to 2026, 13th edn, 2017, 12ff. 10. B. Jones, Material demand for batteries and potential supply constraints, 2018, accessed via https://www.iea.org/media/Workshops/2018/Session3BenjaminJonesCRUgroup.pdf. 11. D. Kushnir, Lithium Ion Battery Recycling Technology, ESA REPORT # 2015:18, 2015. 12. T. Träger, B. Friedrich and R. Weyhe, Recovery Concept of Value Metals from Automotive lithium-ion Batteries, Chem. Ing. Tech., 2015, 87(11), 1550–1557. Special Issue: Thermische Prozesstechnik für Metallgewinnung und Recycling.

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

Chapter 14

Emerging Market of Household Batteries D. Parra* Energy Efficiency Group, Institute for Environmental Sciences and Forel Institute, University of Geneva, Boulevard Carl-Vogt 66, 1205 Genève, Switzerland *E-mail: [email protected]

14.1  Why Household Batteries? Solar photovoltaics (PV) was the most deployed generation technology in the years 2016 (72.4 GW) and 2017 (93.7 GW), reaching over 385 GW by the end of 2017.1 This accelerated development is also characterised by rapid fall­ ing costs, and since 2010 the cost of new solar PV has declined by 70%.2 Encouraging policies in the form of economic incentives such as feed-in tariffs (e.g., in Germany and Italy) and renewable portfolio standards (e.g., USA and China) has been a great mechanism to accelerate the deployment of PV technology. Feed-in tariffs (FiTs) are a guarantee for the owner of the PV system that the electricity injected into the main grid will be bought at a constant price for several years, which covers the investment costs, while renewable portfolio standards is a policy that ensures a minimum share of renewable generation for a region.3 Although economies of scale play a role in PV technology and the capital cost reduces with the system scale (e.g., PV plants at the utility scale are

  Future Lithium-ion Batteries Edited by Ali Eftekhari © The Royal Society of Chemistry 2019 Published by the Royal Society of Chemistry, www.rsc.org

335

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

336

Chapter 14

50% and 60% cheaper than residential PV from an initial investment and life cycle cost perspective respectively4), PV is the most modular technology available in the market. Interestingly, the ratio between the levelised cost and capital expenditure remains constant with the scale since solar irra­ diance is scale independent (opposite to wind technology since power is proportional to the cubic of wind speed). As a consequence, small PV instal­ lations of a few kilowatts have largely been embedded into consumption centres. Making use of the available area on roofs, residential PV installa­ tions have been an important contributor to the massive deployment of PV installations. According to the National Renewable Energy Laboratory (NREL) in the USA, 67% of new PV installations with a nominal capacity lower than 2 MW were integrated into the residential sector.5 This is also the case in Germany (with a current installed capacity of 75 GW, the second largest in the world after China) where PV rooftop plants contribute to more than 60% of the total installed capacity.6 In Switzerland, 6317 and 834 PV systems were installed in individual and collective buildings in Switzerland in 2016, corresponding to 71% and 9% of the total of new grid connected PV systems.7 These installations add 58 MWp and 22 MWp respectively, i.e. 22% and 9% of total capacity additions respectively. Following domestic driving forces, the second sector of the economy (i.e. industry and craftwork) is the main contributor in terms of capacity additions (111 MWp in 2016) followed by farming with 40 MW. After this successful deployment, policies have been adjusted and the value of FiTs has been reduced across many countries in the last few years in order to adapt it to the falling PV cost. For example, the FiT was divided by 3.5 from 2009 to 2016 in Germany, as shown in Figure 14.1, and its phase-out is expected in the coming years.8 A similar trend has also been observed in other countries such as Switzerland, Italy, UK and Spain.9 This reduces the profitability of residential PV installations but the electricity generated by residential PV installations can also be directly supplied to the local demand loads in the dwelling. In this case, PV generation avoids the purchase of retail electricity, the value of which is currently much higher (up to three times, as given in Figure 14.1) than the FiT across many geographies. In the case of Germany whose legislation has supported the massive penetration of renew­ able energy technologies and PV in particular together with the phase out of its nuclear plants, retail electricity prices increased by an average rate of 5.7% p.a. between 2010 and 2013.10 Other European countries have experi­ enced substantial increases in recent years including Spain (63% from 2008 to 2013, i.e. 10% p.a.) and UK (46% from 2007 to 2013, i.e. 6.5% p.a.) and in the European Union on average household electricity prices rose 4% p.a. between 2008 and 2012. However, only a small fraction of the PV generation meets the electric­ ity demand of dwellings due to its weather dependence, typically around 25–35% on an annual basis for a standard PV system in a region without important cooling needs throughout the year. In order to further increase PV self-consumption, electricity storage with residential batteries is a

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

Emerging Market of Household Batteries

337

Figure 14.1  Feed-in  tariff (FiT) for PV power as a function of the commission­ ing date, average remuneration of the bidding rounds of the Federal Network Agency in Germany. Reproduced with permission from the author of the following study: Recent Facts about Photovoltaics in Germany, Fraunhofer ISE, download from http://www.pv-fakten.de, version February 21, 2018.

becoming increasingly attractive for consumers with a PV system,9,11 also referred to as prosumers. Adding a household battery can increase the PV self-consumption from around 30% (for direct PV self-consumption) to 60% approximately, but final values depend on the system configuration and dwelling. Following the increase of retail prices, the reduction of PV cost and the reduction of FiT, household batteries have become a prime application for the energy storage industry (another important application is the use of frequency control). One of the world's most advanced residential storage markets is Germany, where around 50% of new PV installations include a battery system.12 Other emerging markets are geographies such as Australia, California and United Kingdom. However, household batteries performing PV self-consumption are not profitable yet. In Germany, the motivation of consumers who invested in a household battery was not primarily based on profitability though; their decision responded to different criteria such as hedging against increasing electricity costs, sustainability and interest in energy storage.12 This chapter gives an overview of various factors that can contribute towards improving the techno-economic and environmental attractiveness of household batteries and therefore facilitate their acceptance among house­ hold owners and accelerate their deployment. Renewable energy and battery technologies for off-grid dwellings and communities are also an important niche13 but this chapter focuses on household for batteries connected to a national grid. Topics such as various battery applications for consumers,

View Online

Chapter 14

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

338

different battery technologies and the community scale are covered. The overall cost, environmental and social implications of the massive deploy­ ment of a PV-coupled battery system are also discussed in Section 14.5. The final section is an interdisciplinary outlook including implications for vari­ ous stakeholders such as consumers and utility companies as well as some policy recommendations.

14.2  A Household Battery System Figure 14.2 is a schematic representation of the two different types of topol­ ogies that can be used for PV-coupled battery systems, namely a DC-cou­ pled topology and an AC-coupled topology. The main difference lays in the electronics, namely DC/DC converters and DC/AC inverters, which enable the integration of the PV-coupled battery system into the dwelling and the national grid. A DC-coupled topology comprises a single power conversion to store electricity through a battery charge controller, whereas the AC-coupled

Figure 14.2  Two  different topologies that can be applied for a PV-coupled battery system: DC-coupled topology (top) and AC-coupled topology (down).

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

Emerging Market of Household Batteries

339

topology requires two power conversions: firstly DC PV electricity to AC PV electricity through the PV inverter and then through a bi-directional inverter to the battery. An important implication is that the AC-coupled topology allows for the retrofitting of a PV system, i.e. the household battery could be installed after the PV system while the DC-coupled topology requires a joint installation together. The DC-coupled topology is advantageous for installations that have a limited share of direct PV self-consumption from the PV system (i.e. much electricity is stored in the battery) since there is only one electricity loss across the battery charger between the PV system and the battery. The nom­ inal efficiency of a DC/DC converter is around 98% and is higher than the nominal efficiency of an inverter (around 95%), which is used twice in the AC-coupled topology. Furthermore, the DC-coupled topology also allows the selection of a smaller inverter relative to the PV rating and to store otherwise clipped energy.14 Likewise, the position of the battery upstream of the PV inverter (where the curtailment occurs) can prevent curtailment of PV elec­ tricity when a regulatory threshold is in place, as in Germany (see Section 14.3.4). In this case, PV electricity can be stored in the battery while this is not the case of the AC-coupled topology since the battery is downstream of the inverter.

14.3  E  lectricity Prices and Battery Applications for Consumers This section introduces key applications for consumers in general (and resi­ dential consumers in particular) including their economic drivers. Consumer applications are those applications performed by a battery that have a direct positive impact on the electricity bill. Figure 14.3 is a schematic representa­ tion of the typical performance of a battery throughout a representative day when performing each consumer application, namely PV self-consumption, avoidance of PV curtailment, demand load-shifting and demand peak-shav­ ing (only back-up power is not represented). Interestingly, the economic driver for each application is associated with the various components of an electricity bill. An electricity bill typically includes a single electricity price for the energy related component (Pe) of the electricity, as given by eqn (14.1). Here Ed refers to the electricity consumption in energy terms.   



Bd = EdPe

(14.1)

   In addition to the profit margin, the energy price of the bill typically com­ prises three elements, namely wholesale electricity price, network charges and taxes or levies. The relative share of the wholesale electricity price, which used to be the main component, in the retail price of electricity has dimin­ ished over time across the European Union. Moreover, the network compo­ nent increased 36.5% in average between 2008 and 2012. The reasons for

Published on 14 March 2019 on https://pubs.rsc.org | doi

340

Figure 14.3  Schematic  representation of the performance of a household battery when performing various consumer applications on a representative day: (a) PV self-consumption; (b) avoidance of PV curtailment; (c) demand load-shifting; and (d) demand peak-shaving. The horizontal axis gives the time throughout the day in minutes.

Chapter 14

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

Emerging Market of Household Batteries

341

this may be related to network regulation and cost allocation practices but these are not always justified in detail by utility companies. Finally, taxes are also an important contributor to the final retail price across many Euro­ pean countries. European states usually include a tax per se (e.g., VAT) as well as levies for financing energy and climate policies. For example, the cost of financing the feed-in tariff was around 15.5% of the final retail price in Ger­ many15 (similar in Spain). There are many ongoing discussions about how electricity tariffs should evolve in order to, on the one hand, reflect real market prices and on the other hand give clear signals for energy efficiency and/or carbon savings to consumers.16 Capacity-based tariffs are those in which the power component becomes relevant and are currently widely used for larger consumers (typi­ cally industrial) and small and medium enterprises (SMEs, with a typical con­ sumption of 10 MWh per month) but not for small consumers.9 This could change in the medium term and households could be charged also based on their maximum power capacity usage of the network regarding both import and export, in particular with the further penetration of PV systems, heat pumps and electric vehicles. A report commissioned by the Energy Supply Association of Australia concluded that capacity-based tariffs reflect best the real costs of power suppliers and are most effective for demand peak-shav­ ing. Capacity tariffs would therefore allow utility companies to overcome profit losses due to PV self-consumption while consumers incorporating a PV-coupled battery system could also minimise the related expenses. The electricity bill would then also include an additional electricity price for the power related component (Pp)m which applies to the maximum import from the grid in power terms (Πd) as shown in the third term for the electricity imported from the grid of eqn (14.2). Moreover, the price for the energy component can vary with time, e.g. hav­ ing an off-peak peak (subscript ‘off’) and peak period (subscript ‘p’), and if a PV-coupled battery system is installed on-site, the electricity exported to the grid is compensated for with a FiT but curtailed if exceeding a threshold (like in Germany).17 Eqn (14.2) also includes these two extra components in the electricity bill, but the curtailment part is implicit within EPVgrid. Here the sub­ script ‘dgrid’ denotes the part of the demand met with electricity imported from the grid (opposite to be supplied by local PV).   



Bd = Edgrid−offPe−off + Edgrid−pPe−p + ΠdgridPp − EPVgridFiT

(14.2)

   Household batteries can also contribute towards the stability and flexibility of the electricity network by performing other applications such as distribu­ tion upgrade deferral, frequency control, voltage control, power management, restoration and islanding capability.18 However, the performance of these applications requires more complex management techniques and in some cases (e.g., frequency control) the figure of an aggregator that pools many

View Online

Chapter 14

342 19

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

residential batteries to enter into a national market. Therefore, this section only focuses on consumer applications when considering that they can also contribute indirectly to increase the flexibility and stability of the power sys­ tem and reduce the likelihood of the stability issues of the power network.

14.3.1  PV Self-consumption Household batteries can help to increase the amount of local PV generated and used on-site in a dwelling by charging and discharging when PV genera­ tion is higher and lower than the electricity demand load respectively.20 Dis­ charging a battery to supply local electricity demand only makes economic sense when the round trip efficiency of the battery (i.e. it includes all the energy losses within the battery system such as those associated with the storage medium and battery bi-directional inverter), η, is higher than the ratio between the price of the electricity sold (i.e. the FiT) and purchased (i.e. retail), as indicated by eqn (14.4), which is derived from eqn (14.3) to cal­ culate the revenue obtained from performing PV self-consumption, RevPVSC. Here EPVbat and Edbat refer to the battery charge from the PV system and bat­ tery discharge to meet the local electricity demand load respectively. There­ fore, the evolution of the ratio between retail price and FiT determines the revenue for this application together with the round trip efficiency of the battery. Without economic incentives, the FiT converges to the wholesale electricity price and this is expected to be the case for the energy transition once PV technology becomes fully mature. It is expected that at both levels electricity prices will increase across the energy transition but there is much uncertainty in specific trends, as indicated by Figure 14.4.   

RevPVsc = Edbat × Pe − EPVbat × FiT   

 FiT  Rev PVsc  EPVbat  Pe     . Pe  

(14.3) (14.4)

   For economic sense, adding a battery to a PV system should be pursued when the levelised cost of the battery discharge throughout its lifetime is lower than the difference between the retail electricity price and the FiT. The levelised cost of battery systems is mostly sensitive to the capital expenditure (CAPEX) and the relationship is linear with a slope of 30–40% depending on the specific characteristics of the PV-coupled battery sys­ tems.21 Another very important factor also reflected in Figure 14.4 is the amount of surplus annual PV generation, the annual electricity demand and its annual pattern. Current levelised cost values for household bat­ teries performing PV self-consumption are still higher than 0.3 $ kWh−1 for any battery technology, solar resource and demand profile but can approach this value for lithium nickel manganese cobalt oxide (NMC) tech­ nology (current cost 400 $ kWh−1) installed in dwellings with large PV gen­ eration and annual electricity demand.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

Emerging Market of Household Batteries

343

Figure 14.4  Factors  impacting the economic viability of a PV-coupled battery sys­ tem as a function of the importance and the associated uncertainty. This specific case was developed for the German market but could be in principle extended to other geographies. Reproduced from ref. 59 with permission from the author.

The performance schematically represented in Figure 14.1 maximises the economic revenue for a consumer when constant electricity prices apply for both electricity imported and exported and it is based on a simple schedule in which electricity is stored whenever PV generation is higher than the electricity demand. Table 14.1 gives some representative values for the distribution of PV generation and how the electricity demand is met for PV systems that integrate a battery depending on the type of build­ ing, namely domestic, commercial or industrial buildings. However, it has been argued that this basic control of a household battery performing PV self-consumption is not optimal from an energy system perspective, in par­ ticular for guaranteeing the stability of the electricity network.22 For exam­ ple, forecast of the PV electricity generation and the electricity demand can be used to relieve the electricity network with co-benefits to the distribu­ tion system operator while keeping the PV self-consumption share. Alter­ native control strategies to enhance the integration of PV-coupled battery systems include advanced battery management and a fully programmable PV production profile.18 The CAPEX reduction of lithium-ion batteries (see Section 14.3) is con­ tributing to the reduction in the profitability gap of household batteries performing PV self-consumption but there is still no general economic case worldwide. Countries with high retail electricity prices such as Germany and Australia are examples of the most interesting markets at the moment but the amount of annual solar irradiance is also an important factor to be considered. Also in Germany there has been a market incentive programme since 2013 issued by the Germany Federal Government to stimulate the

View Online

Chapter 14

344

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

Table 14.1  Upper  part: annual distribution of PV electricity generation by PV-cou­

pled battery systems integrated into buildings depending on the sector. The direct PV self-consumption is given by the column referred to as “to demand”. Lower part: annual contribution to the electricity demand depending on the sector. The self-sufficiency is given by the column referred to as “from PV”. Data refer to some representative values and but final values are greatly depend on the system configuration.

a

Domestic Commercialb Industrialc Domestic Commercial Industrial

To demand

To battery

To grid

30% 70% 30–70% From PV 30% 30% 10–30%

30% 10% 15–30% From battery 30% 5% 5–30%

40% 20% 15–40% From grid 40% 65% 40–85%

a

 ased on a single dwelling. Direct PV self-consumption increases for block of flats (common B buildings) due to rooftop constraints regarding the total demand. b Based on a PV system with a size of 100 kWp and a 40 kWh battery for a supermarket with a maximum load of 60 kWp. c Ranges based on the type of industry, namely continuous process or daily activity, i.e. lower or higher contribution of the PV and battery system to the demand (self-sufficiency).

market and boost the technology development of PV-coupled battery sys­ tems.23 It is projected that the break-even point could occur around 2020 across several countries.8,24 Despite lithium-ion batteries being the benchmark technology, other elec­ tricity storage technologies have been proposed for PV self-consumption but only hot water tanks with heat pumps or electric heating make economic sense for the conversion of PV electricity into domestic hot water and/or heating, in particular if the dwelling already uses electricity for these pur­ poses.25 Other technologies such as hydrogen storage and flow batteries are significantly less attractive due to lower round trip efficiency and higher CAPEX.26,27 These technologies would only make sense in a future scenario with important CAPEX reductions and large amounts of PV surplus electric­ ity to be managed beyond a daily mismatch, e.g., seasonal storage.

14.3.2  Avoidance of Renewable Energy Curtailment Electricity curtailment is the most straightforward solution for a local net­ work that cannot cope with high levels of PV export when the local electric­ ity demand is low. It has been put into practice across many geographies (e.g., Germany and UK) when weather conditions are very favourable and therefore renewable energy generation is high enough in a period with low demand to be able to destabilize the network frequency or the local volt­ age.28 Germany was the first country to introduce curtailment capability as a new requirement for PV plants in 2009.29 The curtailment capability refers a threshold relative to the installed PV capacity beyond which electricity can­ not be exported into the grid. Therefore, the revenue of a household battery

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

Emerging Market of Household Batteries

345

system avoiding PV curtailment (REVPVct) can be determined using eqn (14.5) where [EPVbat] is the fraction of the stored PV electricity that would be otherwise curtailed. The threshold value has been revised since it was first adopted (equal to 70%) and currently the maximum value of the electricity injected into the grid should be lower than 40% of the nominal installed PV capacity in order to be eligible to receive a FiT scheme. Related regulations have also been ruled across other geographies, e.g. in Japan and California.   

RevPVct = [EPVbat]Ct × η × Pe.

(14.5)

   The avoidance of PV curtailment is an application that is typically per­ formed together with PV self-consumption but the value generated by a household battery avoiding PV curtailment is also substantially higher than for PV self-consumption since in this case there no alternative for the PV electricity generated opposite to PV self-consumption for which PV electric­ ity could be exported and rewarded with the FiT. As a consequence, there are no conditions for the battery round trip efficiency to meet in order to per­ form the avoidance of PV curtailment. At the moment, the avoidance of PV curtailment is also referred to as a power application since periods when PV curtailment are required from a distribution system perspective are not so common yet. However, the frequency could increase throughout the energy transition and therefore this application may become an energy application from the discharge duration perspective (see Table 14.2).

14.3.3  Demand Load-shifting Residential batteries can also be used to shift the electricity demand of a consumer without changing customer habits and this is the key advantage over demand-side-management technologies.30 In this case, the economic driver for this application is related to time-varying electricity prices, which include peak and off-peak periods such as time-of-use tariffs or real time tar­ iffs (RTPs) depending on the hour of the day, day of the week or month of the year. In terms of battery management, it consists of charging it whenever the price of electricity is low (off-peak period) and using the electricity stored when the price of the purchased electricity is higher later on (peak period), as stated by eqn (14.6):   

RevDLS = Edbat × Pe−peak − Egridbat × Pe−off      

 P Rev DLS Edbat  Pe  peak    e  off  Pe  peak 

  . 

(14.6)

(14.7)

In order to obtain some revenue from performing demand load-shift­ ing, the round-trip efficiency of the battery should be higher than the ratio between the off-peak and the peak electricity price as derived from eqn (14.7).

View Online

Chapter 14

346

Table 14.2  Comparison  of various consumer applications that can be performed

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

by household batteries. Some representative values are given for the battery discharge but they would finally depend on the system configu­ ration and daily mismatch between the PV generation and the electric­ ity demand load. Discharge duration Classification Economic driver

PV self-­ 3h consumption Demand 4h load-shifting Avoidance of PV 0.2 h curtailment Demand 1h peak-shaving Backup power 0.1–5 h

Energy Difference between application retail electricity prices and feed-in tariff Energy Difference between application peak and off-peak electricity prices Power Monetary losses application associated with electricity curtailed Power Capacity-based tariffs application Energy Loss of value associated application with the outage

Value added Medium–high Low–medium High Very high Very high

Time-of-use tariffs have been offered, in particular to industrial consumers who are more sensitive to electricity prices, by utility companies across many geographies (e.g., United Kingdom, United States and Australia) for the last 40 years because demand load-shifting also offers several benefits to the overall energy system such as better utilisation of the existing generation assets and reduction of the total cost and emissions.31 At the same time, it is expected that real-time pricing becomes more relevant across the energy transition since new renewable generators such as solar and wind as well as new electricity demand loads such as heat pumps and, to a lesser extent, elective vehicles, show great spatiotemporal and inter-annual variability depending on weather conditions. Off-peak prices are substantially higher than FiT values and therefore demand load-shifting is an application that significantly adds less value than PV self-consumption.32 For example, the FiT and the off-peak price in Swit­ zerland were equal to 10 $ kWh−1 and 0.17 $ kWh−1 respectively. On the other hand, a battery performing demand load-shifting is not limited by the PV surplus electricity (e.g., on cloudy days) and it can reach high cycle activity. As a result, the levelised cost of batteries performing demand load-shifting is around 30% less than for PV self-consumption in regions with temperate climate.21

14.3.4  Demand Peak-shaving Residential batteries can be used to meet peaks in the electricity demand curve (e.g., the typical peaks during evenings) in the residential sector. For a household battery system, the logic is similar to demand load-shifting but

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

Emerging Market of Household Batteries

347

some key differences are that the peak demand in a single dwelling may only last for a few minutes and the time at which it occurs is random. Therefore, demand peak-shaving is recognised as a power application characterised for discharges of a few minutes while demand load-shifting is referred to as an energy application with associated discharges of a few hours9,21 (see Table 14.2). In the domestic sector, batteries can reduce the peaks during the after­ noons and/or evenings by performing demand peak-shaving. Regarding battery management, demand peak-shaving consists of hav­ ing enough battery charge when the maximum peak occurs on a daily basis, therefore some forecast is necessary unless the peak electricity demand becomes predictable. The revenue associated with demand peak-shaving (RevDPS) is proportional to the reduction in the peak demand and the capac­ ity-based component of the electricity bill as shown by eqn (14.8) (see also the third element of eqn (14.2) about the electricity bill). Demand charges are referred to the part of an electricity bill where charges are based on the power component but so far are more relevant for large consumers, e.g., industrial consumers that use high intensive electrical demand loads. However, they usually pay for a maximum subscribed power that is not often used. Batteries can be used to smooth the grid import and even reduce the subscribed power, especially when the value of the maximum demand load can be forecasted. It is expected that the rolling out of smart meters will give more information about the consumer electricity demand profile and contribute towards the use of demand charges for residential consumers.   

RevDPS = max[Πdgrid − (Πdgrid)bat] × Pp.

(14.8)

   Being a power storage application, the value associated with demand peak-shaving has been reported to be the highest from all consumer storage applications33 and, for example, the combination of PV self-consumption and demand peak-shaving has been reported to create a positive return of investment for household batteries in Germany assuming a hypothetical sce­ nario with a market mechanism for this application.9 The reduction in the peak electricity demand brings several associated advantages to the energy system and contributes towards a reduction in the overall energy system cost. The peak demand is met by the most costly (and sometimes less efficient) generators that only run during peak time. For any utility company, the reduction in the peak power means purchasing less elec­ tricity when the price is higher. Typically the most expensive power plants in the “merit order” run the least just to meet the peak demand. In addition to this, it allows the use of less expensive equipment in the house (such as mea­ surement and other auxiliary equipment) but also the deferral of investment in low voltage lines to increase its capacity, e.g., transformers and voltage lines that are sized to meet the peak demand. Therefore, demand peak-shav­ ing could be very effective for the further integration of electric heating (i.e. heat pumps) and electric mobility (i.e. electric vehicles) into the existing power network across the energy transition.

View Online

348

Chapter 14

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

14.3.5  Back-up Power Across geographies with power systems that are not very reliable and/or sta­ ble due to endogenous (ageing) or exogenous (storms and hurricanes) factors and in general for services or dwellings that must ensure a continuous power supply (e.g., hospitals and hotels), batteries can also provide electricity for momentary outages. In this case, the economic revenue created by the bat­ tery (Revback-up) is in reality an avoided cost from the consumer in an alterna­ tive technology (e.g., diesel generator) and/or the avoided loss of value related to the activity supplied by the electricity (e.g., hotels and data centres).

14.3.6  Combination of Applications Table 14.2 gives an overview of the five existing consumer applications for household batteries. Following the increase in retail prices, the reduction of PV cost and the reduction of FiTs, PV-coupled battery systems have emerged as a prime application for energy storage in addition to frequency control. This is already increasingly the case in Germany, one of the world's most advanced residential storage markets, where around 50% of new PV installa­ tions include a battery system.23 However, the motivation of consumers who have invested in a PV-coupled battery system in Germany was not primar­ ily based on profitability; their decision responded to different criteria such as hedging against increasing electricity costs, sustainability and interest in energy storage. Household batteries can combine the various applications introduced above to increase the economic attractiveness, i.e. adding the various reve­ nues introduced above, as shown in eqn (14.9). Current research shows that residential batteries would be close to economic viability if the four applica­ tions are combined within an appropriate regulatory framework. In particu­ lar, when the tariff structure includes a capacity tariff and a PV curtailment obligation. The combination of applications is always attractive, except for PV self-consumption and demand load-shifting for which dwellings should have a large annual demand. These are energy applications, and therefore charging from both the PV system and the grid makes economic sense when the PV generation is relatively low and/or the annual electricity demand is relatively high. For the combination of applications to become a reality, there are however some important developments to be accomplished. From the technology perspective, battery management systems should become more sophisticated, e.g., communicate with smart meters and integrate forecast­ ing techniques for balancing PV generation and electricity demand in real time considering various electricity prices. Furthermore, policy makers should upgrade market mechanisms for the various applications and avoid regulatory barriers.34   

Revbat = RevPVsc + RevPVct + RevDLS + RevDPS + Revback-up   

(14.9)

View Online

Emerging Market of Household Batteries

349

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

14.4  Different Lithium-ion Battery Technologies Overall, lithium-ion battery technology has become the most widespread electrochemical technology for PV integration and household storage given its overall high round trip efficiency (90% approximately) and suitability for short-term and mid-term storage cycles. This success has been driven by a strong reduction in the cost of lithium-ion batteries. For example, the installed price per usable capacity reduced by half in 2017 from around 3000 € kWh−1 in 2013 in Germany.35 Lithium-ion is however a family of various technologies based on a common material (i.e. lithium, in particular on the movement of lithium ions between the two electrodes during charging and discharge). Table 14.3 compares key characteristic of various existing lith­ ium-ion technologies in the market, namely NMC, lithium nickel cobalt aluminium oxide (NCA), lithium iron phosphate (LFP) and lithium titanate (LTO). Moreover, other battery technologies such as lead-acid and vanadium redox are also given for comparison purposes. Although the technology readiness of various lithium-ion batteries in Table 14.3 is already very high and they are already in the commercialisa­ tion phase,36 it is still expected that the cost and performance will markedly improve through the energy transition. Some key drivers are the mass pro­ duction of lithium-ion batteries for electric vehicles and/or renewable energy support and intense R&D efforts given their expected pivotal role in achiev­ ing a global low carbon energy system. Based on a thorough literature review and experts' opinions, The International Renewable Energy Agency (IRENA) provided improvement trajectories of various key performance indicators for different lithium-ion technologies taking the year 2030 as a reference, as shown in Figure 14.5. It is possible to see that CAPEX, equivalent full cycles and calendar life are the parameters for which greater improvements are projected across all technologies, while progress in other indicators such as energy density, depth of discharge and round trip efficiency seems to have already reached a plateau. In particular, CAPEX reductions between 50% and 60% are expected across all lithium-ion technologies by 2030.37 Table 14.3  Key  characteristics of various lithium-ion battery technologies based on various sources.37,58

Technology Cathode family material Li-ion

Lead-acid Flow

LFP NMC NCA LTO Traditional Advanced Vanadium redox

Round trip Maximum CAPEX efficiency Cycle life Calendar discharge 2016 (%) (EFC) life (years) rate ($ kWh−1) 92 95 97 96 91 3000 70

2500 2000 1000 10 000 1500 3000 13 000

12 12 12 15 9 10 12

2C 0.4 C 1C 4C 0.1 C 1C 1C

580 420 350 1050 260 750 347

Published on 14 March 2019 on https://pubs.rsc.org | doi

350

Figure 14.5  Properties  of selected chemistries of lithium-ion battery electricity storage systems, 2016 and 2030. Figure created and

Chapter 14

provided by ©IRENA in the following report: IRENA (2017), Electricity Storage and Renewables: Costs and Markets to 2030, International Renewable Energy Agency, Abu Dhabi.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

Emerging Market of Household Batteries

351

The technology choice should be made based on the application(s) to be performed and considering the optimal battery size and discharge rating. In the end, this decision is a trade-off considering that technologies that offer better performance in terms of number of equivalent full cycles and round trip efficiency (e.g., LTO) have a higher CAPEX. However, the profitabil­ ity of a household battery system is mostly sensitive to the CAPEX and this parameter is typically pivotal for the decision-making process of consumers. Interestingly, current research suggests that some technologies with higher CAPEX values are relatively more competitive when installed in dwelling with large demand and/or various applications are combined since they can pro­ vide more cycles. Regarding other competitors, flow batteries are an attractive solution for mid-term storage applications (from four hours to daily storage) because of their unique characteristic of decoupled energy and power rating (this is not the case for lithium-ion batteries). However, their current main applications such as wind energy capacity firming correspond to systems of a few MWs installed at the distribution and transmission networks38 and the household scale does not seem be the most appropriate for its deployment given some intrinsic cost dependence with the scale. When disregarding capacity tariffs, traditional lead-acid batteries are still competitive for demand load-shifting, which is the consumer application with the lowest power-to-energy ratio (i.e. longest discharge duration, as given in Table 14.2) since the battery capacity is sized according to the electricity demand load occurring during the peak period.30 However, Li-ion batteries are much more interesting for PV self-con­ sumption (with the battery sized according to surplus PV generation require­ ments) and demand peak-shaving.32 As the penetration of renewable energy and low carbon technologies increases during the energy transition, hybrid systems (comprising different types of electrochemical technologies, e.g., supercapacitors, Li-ion batter­ ies, flow batteries and/or hydrogen) may be required for some communities or districts in order to cover the full spectrum of applications and meet the associated storage cycles with different temporal scales (from seconds to weeks or months).18,39

14.5  The Community Scale Considering that batteries installed in dwellings are not profitable yet, the concept of community energy storage (CES) as a battery system located near residential consumer points (at the edge of the grid) and able to develop sev­ eral roles has been proposed.40,41 Some basic differences between CES and a household battery are the number of dwellings to which the battery is con­ nected to, the size and the number of applications which the battery system delivers. For the latter, CES is sometimes suggested under the smart grid concept42 and delivering applications such as distribution network upgrade deferral and/or ancillary services (e.g., frequency control), which are relevant for other stakeholders in addition to consumers.40

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

352

Chapter 14

Community schemes create similar benefits as batteries implemented at the level of individual consumers but the advantages of CES are improved economies of scale (especially in aspects such as power electronics, commu­ nications and control technologies) and the option of professional manage­ ment as well as system benefits at the level of the distribution grid. Moreover, the required power rating of the inverter of the community battery system and to a lesser extent the battery capacity reduces with the community size.43 For example, Barbour et al. reported that the optimum CES requires a total storage capacity that is 65% of that at the level of individual households for a total amount of 4574 dwellings in Cambridge (USA).42 Furthermore, com­ munity batteries are able to reduce significantly the interaction with the main grid in terms of imports and exports when connected into micro-grids despite the lower total installed capacity, in particular, improvement val­ ues in the range 64–94% were reported in the aforementioned study. Over­ all, lower CAPEX and operational expenditure (in particular maintenance) should be expected per kWh of installed capacity. Last, but not least, the community scale has proven to be a catalyst for the engagement of citizens in the energy transition in order to build a sustainable future, i.e. speed up renewable energy penetration, increase energy awareness and reduce the carbon footprint of communities.18 Given the currently increasing rates of battery adoption for individual dwell­ ings and the further potential for widespread adoption, energy policies should prioritise the development of market mechanisms that facilitate the deploy­ ment of shared community-level storage. This is especially important given the limited and capital-expensive nature of the resources required for battery man­ ufacture. Policy makers are already reacting to this evidence and community schemes are being suggested for PV owners without or with storage in some countries. For example in Switzerland, the current legislation allows the pos­ sibility of pooling self-consumers when the PV installed capacity is greater or equal to 10% of the maximum grid connection capacity. This would allow own­ ers who share the same production site the opportunity to join together as one final consumer. This is called a “community of self-consumers” and they will be considered by the grid operator as one final consumer. The law also specifies that the costs of “joint PV self-consumption” are at the expense of the owners. In the ordinance, the production site should be a property of the owner, who could develop a joint project with the neighbours using the surrounding land (e.g., another building) as long as the electricity network of the grid operator is not used between the production and the consumption. Then, the owner can also “force” the tenants to self-consume by writing it in the lease agreement of new buildings/apartments. If it is an existing lease agreement, the tenants can­ not be “forced”. Furthermore, a community of “PV self-consumers” that has a demand of more than 100 MWh per year has no longer the obligation to pur­ chase his electricity from the local utility and they have access to the electricity market. Battery storage is one of the technological options consumers have to increase the amount of local PV generation used by them.

View Online

Emerging Market of Household Batteries

353

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

14.6  Global Impact of Household Batteries The effects of the massive penetration of PV-coupled battery systems located behind the meter are manifold including both positive and negative impacts, as summarized in Table 14.4. In particular, the coupling of battery storage to PV systems can help to mitigate some of the negative effects associated with the intermittency of PV generation. In this section, the impacts of PV-cou­ pled battery systems installed behind the meter are assessed according to the following criteria: affordability, sustainability and energy security (these criteria also known as the energy trilemma). Finally, the impact on the whole­ sale electricity market is also discussed.

14.6.1  Energy System Cost In the mid-term, the cost of the overall energy system should increase if relying heavily on PV-coupled battery systems. As discussed in Section 14.1, the capital cost of PV has fallen dramatically in recent years (by two thirds at the utility scale) and it is expected to reduce further up to 50% by 2040,2 but its weather dependency prevents it from supplying larger fractions of demand. Despite operational cost savings compared to fossil supply, the levelised cost of PV systems is projected to remain higher due to low capac­ ity factors. The typical annual capacity factor a PV installation located in central Europe (e.g., Switzerland) is around 15%. Household batteries can increase the value of PV technology by providing high-value electricity on demand (e.g., up to 250 € MWh−1 for consumers considering the difference between retail and wholesale prices in Germany) but as the cost of solar technology drops, the cost of batteries must also drop to continue to add value.44 From the levelised cost perspective, the current levelised cost of batteries performing PV self-consumption still doubles the levelised cost of a small scale PV system across central Europe (slightly higher than 200 CHF MWh−1 for Switzerland). However, the price of battery storage has fallen more than four times for the last five years to reach 230 $ kWh−1 in 2017 and, according to Panasonic, it may be half of 2016's price by 2025. Fur­ thermore, several applications such as PV self-consumption and demand load-shifting can also be combined to reduce the levelised cost of batteries to below 300 $ MWh−1.21 However, if external costs are included to evaluate the attractiveness of PV technology, e.g., health costs due to pollution and various climate change impacts, previous studies have proven that non-renewable energy technolo­ gies are more expensive than renewables. Moreover, the case becomes even more attractive if other macro-economic impacts such as jobs and GDP are considered within the balance. For example, several studies show the posi­ tive net employment and GDP effects of renewable energy expansion, in par­ ticular up to +3% in GDP and +1% in net jobs for Germany.45 Although final impacts depend on the economic structure of the country, energy supply

View Online

Chapter 14

354

Table 14.4  Various  impacts of the penetration of PV and battery technologies

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

across the energy transition. The impact could become more marked for higher penetration levels. Cost data for PV and battery refer to small scale installations in households. Data come from various sources intro­ duced across Section 6.6 and mainly refers to the European context.

Parameter

Effect

Investment cost Increase (negative) Operational cost Levelised costb

Reason (s) More cost effective supply options, e.g., natural gas No fuel consumption

Current Data - PV: 2500 $ kWp−1 - Battery:a 500 $ kWh−1 and 500 $ kWp−1 - PV: 35 $ kWp−1 - Battery: 0 - PV: 210 $ kWh−1 - Battery:c 400 $ kWh−1

Decrease (very positive) Increase Low capacity (very negative) factor and high investment cost External cost Decrease Reduction on 25–30% reduction (very positive) emissions from fossil fuels Externalities Increase More distribution and +3% GDP such as GDP (positive) labour-­intensive +1% jobs and jobs jobs Environmental Decrease Very low GHG - PV: 60 g CO2 eq. impacts (positive) emissions but kWh−1 metal depletion - Battery: 250–350 g CO2 eq. kWh−1 Security of Stable (neutral) Trade-off between n.a. supply local supply and flexibly provided Social Increase Easy integration in the n.a. acceptance (very positive) built environment, quiet operation, safe and symbolic value a

I ncludes the battery bidirectional inverter. Using a discount factor equal to 8%. PV lifetime is 30 years and battery calendar life 15 years. Based on a battery performing PV self-consumption.

b c

alternatives and local manufacturing levels, the public funded UK Energy Research Centre (UKERC) has stated that solar PV creates at least twice the number of jobs per unit of electricity generated compared with natural gas.46 It is expected that the retrofitting of PV systems with batteries and their com­ bined installation further improves these values.

14.6.2  Environmental Impacts The average life-cycle emissions depend on the selected battery chemistry, but are in general lower than 100 g CO2 kWh−1 for PV self-consumption. It should be noted that the life cycle emissions can double for other applica­ tions based on electricity charging form the grid, in particular in geogra­ phies with high carbon intensity of the grid electricity such as Poland and

View Online

Emerging Market of Household Batteries

355

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

47

Australia. To put these values into context, the life cycle GHG emissions of rooftop PV installations alone is around one tenth of natural gas CHP units, the latter ranging between 500 and 700 g CO2 eq. kWh−1.48 Another important challenge is to sustain growth in the production of lithium-ion batteries, since their manufacture is expected to grow as much as ten times the current size.49 A recent perspective from Olivetti et al. con­ cluded that most materials contained within lithium-ion batteries will likely meet the demand in the mid-term. However, cobalt used for the cathode of NMC and NCA batteries could become a critical material.50 Current produc­ tion is mainly located in countries such as the Democratic Republic of the Congo and Zambia in Africa.

14.6.3  Security of Energy Supply The contribution of PV-coupled battery systems to security of supply can be assessed using three different criteria, namely resource supply, flexibility and availability, and it is therefore subject to trade-offs. Without a house­ hold battery, the PV contribution to these three criteria has been ranked as maximum, minimum and minimum respectively.51 However, household bat­ teries increase the flexibility of PV systems (e.g., by extending the PV supply a couple of hours in the evening) while not impacting on their availability. Finally, the acceptance of PV by citizens worldwide has been demonstrated by various interviews, and people almost unanimously hold a strongly posi­ tive view of solar power.52 This is also the case for household batteries but the concern about their environmental impact is higher based on the author's experience.

14.6.4  Wholesale Electricity Market Wholesale electricity markets (also referred to as spot markets) are also being markedly affected by the penetration of renewable energy technologies such as PV and wind. Due to its zero marginal cost and/or privilege position in the merit order, PV shifts to the right the electricity supply curve and reduces the price considerably in the power spot market. In the short term, this creates savings from the demand side as well as reducing generator profits.53 The spot market price reduced by 0.8–2.3 € MWh−1 in average per additional GW of renewable energy added in Germany between the years 2008 and 2012.54 Furthermore, the spread between peak and off-peak prices is also being reduced, and this also impacts the benefits of hydropower plants. Despite solar PV penetration reducing the business case for conventional flexible generators such as hydro and gas in the spot market, new opportunities arise for battery and other storage technologies storing very cheap electricity and/ or reducing renewable energy curtailment in the spot market and beyond. For example, batteries performing frequency control have been reported to be very close to economic viability at the moment. Although this application

View Online

356

Chapter 14

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

is more straightforward for utility batteries given the minimum size to enter into the market, there are business opportunities for household batteries to be aggregated and provide frequency response.19

14.7  Outlook With some policy support (e.g., FiT) and following the objective of maximis­ ing PV self-consumption, PV systems are already profitable across several geographies considering their lifetime of 30 years. However, this is not the case for household batteries yet, but a break-event point is projected to occur in the coming years (expected in the 2020s) in countries with high electricity prices and/or high solar resources (e.g., Germany and Australia) due to the declining technology cost. However, assuming an interesting regulatory con­ text for household batteries (e.g., Germany), the decision whether to invest in them depends on many different socio-economic criteria such as available financial resources, space availability for installing a large PV system (e.g., 5 kWp or higher) and electricity demand consumption. Citizens, businesses and services who own the building where they live and/or operate respec­ tively can be interesting targets for the PV and battery residential markets. Furthermore, direct PV self-consumption (including a battery system) will be much higher for activities that run throughout the whole year (e.g., hotels, in contrast to schools). Similar to other low carbon technologies, the diffusion of battery storage will have a strong dependence on the regulatory context and their uptake can be enhanced through financial incentives and regula­ tory frameworks established by policymakers. The remarkable success of PV and other renewable electricity coincides with the increase in electricity consumption in final energy demand—a trend that has been underway in parallel to the success of renewable energy technologies. Electrification is hence emerging as a key strategy for deep decarbonisation, supported by various catalytic technologies such as PV, bat­ teries, heat pumps and electric vehicles. Across some countries, electricity is already being successfully used for space heating and for example heat pumps accounted in Switzerland for more than one third of the market share of heating systems by 2013, and with a share of already 12% of installed space heating in 2012.55 Household batteries can be key to manage these new electricity demand loads such as heat pumps and electric vehicles together with PV generation if all different actors, including local authorities and the government, are on board. Given the advantages of shared CES (i.e. shared batteries) over individual household batteries in terms of enhanced match of local PV generation and electricity demand per unit of installed storage, reduced battery capacity requirements for a given area and lower CAPEX and operational cost, energy policies should prioritise the development of market mechanisms that facil­ itate the deployment of shared community-level batteries. Regarding the ownership and related location of PV-coupled battery systems, different

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

Emerging Market of Household Batteries

357

solutions may also coexist in addition to the conventional purchase and management by the consumer. PV-coupled battery systems can be offered by PV installers and/or house-builders and therefore installed in different households and communities (e.g., block of buildings) and paid for by con­ sumers. Alternatively, they can be operated and/or provided by utility com­ panies and distribution system operators while being connected to the PV plants and demand loads of the residential sector. The low voltage side of the utility transformers is already being used for installing batteries in USA. Regardless of the type of ownership model, investments should be profitable but also associated business models should develop win–win solutions for different stockholders involved in the project and avoid free riders. This is an important subject of future research. Two examples of win–win solutions are: (a) electricity tariffs with capacity components for both electricity import and export, which promote house­ hold batteries to combine applications such as PV self-consumption and demand peak-shaving and therefore help them to become profitable; and (b) shared business and/or ownership models (including both CAPEX and maintenance) when the value propositions include applications that benefit different stakeholders. For example, the optimum management of local PV generation benefits both the consumer (e.g., self-consumption is driven by the difference between the import and export electricity prices), and the util­ ity company and/or distribution system operators (e.g., the deferral of distri­ bution network investment). Moreover, utility companies and/or aggregators could also benefit from optimising the performance of PV-coupled battery systems for the electricity network and/or wholesale markets. Likewise, hier­ archical control techniques including both the household/community level, upper level (e.g. distribution network and/or wholesale market) and mainte­ nance should be applied by the utility company (or aggregator). Some schol­ ars are now even suggesting that peer-to-peer energy-trading platforms may be a better market mechanism than a top-down approach managed by utility companies and/or aggregators to assure social welfare.56,57 Some of the various impacts related to the penetration of PV and battery systems may create conflict with other elements of the energy system and some particular stakeholders. Downstream of the power sector, consumers who generate their own electricity with PV systems and who increase the amount of PV-self consumption with batteries reduce substantially their electricity bills. This implies that they save not only the wholesale electricity component of the bill but also network fees and taxes that together account for more than 50% of the bill for some European countries. However, they do still rely on the utility company and distribution grid for meeting max­ imum peak demand and/or during days with low solar irradiance. Overall, the benefits of utility companies are decreasing while they still need to oper­ ate the distribution network system or even upgrade it. The transition from an old centralised to a more distributed energy system brings new opportu­ nities for new actors such as aggregators, IT companies as well as municipal

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

358

Chapter 14

utilities. Utility companies already have a privilege position and therefore they could engage prosumers and propose solutions (e.g., new tariff design) that are win–win for the consumer and (in terms of electricity bills) for the society (in terms of carbon footprint), and for their own profitability. Other­ wise, prosumers may collaborate with new stakeholders such as aggregators or even organise themselves to extract as much value as possible our of their PV-coupled battery systems.

Acknowledgement The results and analysis of this chapter were possible thanks to the support of the Commission for Technology and Innovation in Switzerland (CTI, now called Swiss Innovation Agency) through the projects SCCER-HaE phase I and II. The author would also like to acknowledge his current collaborations in this topic with Martin Patel and Alejandro Pena-Bello at the University of Geneva, and previous research at the University of Nottingham in collabora­ tion with E.ON between 2011 and 2014.

References 1. IRENA, Renewable Capacity Statistics 2018, Abu Dhabi, 2018. 2. Interantional Energy Agency, World Energy Outlook 2016, Organisation for Economic Co-operation and Development, OECD, 2016. 3. H. Yin and N. Powers, Energy Policy, 2010, 38, 1140–1149. 4. Lazard, Lazard's levelized cost of energy analysis-Version 11.0, Lazard, New York, 2017. 5. R. Fu, D. J. Feldman, R. M. Margolis, M. A. Woodhouse and K. B. Ardani, US Solar Photovoltaic System Cost Benchmark: Q1 2017, National Renew­ able Energy Laboratory (NREL), Golden, CO, United States, 2017. 6. Bundesministerium für Wirtschaft und Energie, Marktanalyse Photovoltaik-Dachanlagen, Bundesministerium für Wirtschaft und Energie (BMWI), 2015. 7. SwissSolar, Le recensement du marché de l’énergie solaire en 2016, Swiss Federal Office of Energy, Berne, 2017. 8. V. Bertsch, J. Geldermann and T. Lühn, Appl. Energy, 2017, 204, 1–15. 9. D. Parra and M. K. Patel, Appl. Energy, 2016, 164, 175–187. 10. Bundesverband der Energie- und Wasserwirtschaft e. V. (BDEW), BDEW-Strompreisanalzse Februar 2017, Berlin, 2017. 11. J. Hoppmann, J. Volland, T. S. Schmidt and V. H. Hoffmann, Renewable Sustainable Energy Rev., 2014, 39, 1101–1118. 12. K.-P. Kairies, D. Magnor and D. U. Sauer, Energy Procedia, 2015, 73, 200–207. 13. R. Siddaiah and R. Saini, Renewable Sustainable Energy Rev., 2016, 58, 376–396. 14. P. L. Denholm, R. M. Margolis and J. D. Eichman, Evaluating the Technical and Economic Performance of PV Plus Storage Power Plants,

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

Emerging Market of Household Batteries

359

National Renewable Energy Laboratory (NREL), Golden, CO, United States, 2017. 15. Clean Energy Wire, What German households pay for power, https://www. cleanenergywire.org/factsheets/what-german-households-pay-power, Accessed on 24/01/2019. 16. S. Buryk, D. Mead, S. Mourato and J. Torriti, Energy Policy, 2015, 80, 190–195. 17. M. Leuthold, D. Marchal, R. Sitte, K.-P. Kairies, P. Guerrier, N. Netzel, E. Radvanyi and T. Lenck, Conference on storage in the service of energy transition, 2016. 18. D. Parra, M. Swierczynski, D. I. Stroe, S. A. Norman, A. Abdon, J. Wor­ litschek, T. O'Doherty, L. Rodrigues, M. Gillott and X. Zhang, Renewable Sustainable Energy Rev., 2017, 79, 730–749. 19. G. Litjens, E. Worrell and W. van Sark, Appl. Energy, 2018, 223, 172–187. 20. D. Parra, G. S. Walker and M. Gillott, Sustainable Cities Soc., 2014, 10, 1–10. 21. A. Pena-Bello, M. Burer, M. K. Patel and D. Parra, J. Energy Storage, 2017, 13, 58–72. 22. J. Moshövel, K.-P. Kairies, D. Magnor, M. Leuthold, M. Bost, S. Gährs, E. Szczechowicz, M. Cramer and D. U. Sauer, Appl. Energy, 2015, 137, 567–575. 23. K. P. Kairies, D. Magnor and D. U. Sauer, Energy Procedia, 2015, 73, 200–207. 24. F. M. Vieira, P. S. Moura and A. T. de Almeida, Renewable Energy, 2017, 103, 308–320. 25. D. Parra, G. S. Walker and M. Gillott, Energy Build., 2016, 116, 614–621. 26. L. Baumann and E. Boggasch, Int. J. Hydrogen Energy, 2016, 41, 740–751. 27. D. Parra, M. Gillott and G. S. Walker, Int. J. Hydrogen Energy, 2016, 41, 5215–5229. 28. L. Kane and G. Ault, Energy Policy, 2014, 72, 67–77. 29. J. Wittlinger, in Photovoltaikanlagen im Steuerrecht, Springer, 2012, pp. 29–40. 30. D. Parra, S. A. Norman, G. S. Walker and M. Gillott, Appl. Energy, 2016, 174, 130–143. 31. C. Batlle López and P. Rodilla Rodríguez, Electricity demand response tools: current status and outstanding issues, 2009. 32. D. Parra, S. A. Norman, G. S. Walker and M. Gillott, Appl. Energy, 2017, 200, 358–369. 33. J. Eyer and G. Corey, Sandia Natl. Lab., 2010, 20, 5. 34. D. Hart, L. Bertuccioli and X. Hansen, Policies for Storing Renewable Energy. A Scoping Study of Policy Considerations for Energy Storage (RE-STORAGE), IEA-RETD, 2016. 35. K.-P. Kairies, J. Figgener, D. Magnor, H. Axelsen, E. Waffenschmidtd and D. U. Sauer, Analysis of Funded PV Battery Systems in Germany: Prices, Design Choices and Purchase Motivation. 36. International Energy Agency, Technology Roadmap: Energy Storage, 2014.

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00335

360

Chapter 14

37. IRENA, Electricity Storage and Renewables: Cost and Markets to 2030, Inter­ national Renewable Energy Agency (IRENA), Abu Dhabi, 2017. 38. Sandia National Laboratories, DOE Global Energy Storage Database, Washington DC, USA, 2016. 39. D. Parra, M. Gillott and G. S. Walker, Int. J. Hydrogen Energy, 2014, 39, 4158–4169. 40. A. Nourai, R. Sastry and T. Walker, Power and Energy Society General Meeting, 2010 IEEE, 2010. 41. D. Parra Mendoza, Optimum community energy storage for end user applications, PhD Thesis, University of Nottingham, 2014. 42. E. Barbour, D. Parra, Z. Awwad and M. C. González, Appl. Energy, 2018, 212, 489–497. 43. D. Parra, M. Gillott, S. A. Norman and G. S. Walker, Appl. Energy, 2015, 137, 576–587. 44. W. A. Braff, J. M. Mueller and J. E. Trancik, Nat. Clim. Change, 2016, 6, 964–969. 45. U. Lehr, C. Lutz and D. Edler, Energy Policy, 2012, 47, 358–364. 46. R. Ferroukhi, A. Lopez-Peña, G. Kieffer, D. Nagpal, D. Hawila, A. Khalid, L. El-Katiri, S. Vinci and A. Fernandez, Renewable Energy Benefits: Measuring the Economics, IRENA, Abu Dhabi, 2016. 47. OECD, Greenhouse gas emissions, https://stats.oecd.org/Index.aspx?­ DataSetCode=AIR_GHG, Accessed on 24/01/2019. 48. E. G. Hertwich, T. Gibon, E. A. Bouman, A. Arvesen, S. Suh, G. A. Heath, J. D. Bergesen, A. Ramirez, M. I. Vega and L. Shi, Proc. Natl. Acad. Sci. U. S. A., 2015, 112, 6277–6282. 49. S. Jaffe, Joule, 2017, 1, 225–228. 50. E. A. Olivetti, G. Ceder, G. G. Gaustad and X. Fu, Joule, 2017, 1, 229–243. 51. K. Volkart, C. Bauer, P. Burgherr, S. Hirschberg, W. Schenler and M. Spada, Int. J. Greenhouse Gas Control, 2016, 54, 1–14. 52. B. Sütterlin and M. Siegrist, Energy Policy, 2017, 106, 356–366. 53. F. Sensfuß, M. Ragwitz and M. Genoese, Energy Policy, 2008, 36, 3086–3094. 54. J. Cludius, H. Hermann, F. C. Matthes and V. Graichen, Energy Econ., 2014, 44, 302–313. 55. S. Peterhans and F. Rognon, Heat Pumps in Switzerland. A Success Story. 56. E. Barbour and M. C. González, Appl. Energy, 2018, 215, 356–370. 57. T. Morstyn, N. Farrell, S. J. Darby and M. D. McCulloch, Nat. Energy, 2018, 3, 94. 58. O. Schmidt, A. Hawkes, A. Gambhir and I. Staffell, Nat. Energy, 2017, 2, 17110. 59. V. Jülch, Wirtschaftlichkeit von zukunftsfähigen Geschäftsmodellen dezentraler Stromspeicher, Dissertation, Europa-Universität Flensburg, unpublished.

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00361

Subject Index AC coupled PV battery system, 254 acrylonitrile(AN)-methyl methacrylate(MMA)-styrene(ST) terpolymer, 114 aging mechanisms calendar aging, 230–233 cycling-induced aging irreversible capacity loss (ΔQir), 233–236 non-destructive approach, 238–240 postmortem analyses, 236–238 degradation mechanisms, 240–246 anode electrode degradation, 228–229 cathode degradation, 226–228 SEI formation inducing cyclable Li losses, 225–226 design, Li-ion batteries, 224–225 electromotive force (EMF) determination, 229–230 working principles, Li-ion batteries, 221–223 Al2O3-doped ZnO (AZO) layer, 54 all-solid-state lithium batteries (ASSLBs), 83 alloying materials, 200 anode electrode degradation, 228–229 APU. See auxiliary power unit (APU)

ASSLBs. See all-solid-state lithium batteries (ASSLBs) atomic absorption spectroscopy (AAS), 226 auxiliary power unit (APU), 298 battery electric vehicles (BEVs), 317 battery management systems (BMSs), 221, 247 carbon materials, 2–5 carbon nanotubes (CNTs), 2 carbonaceous materials, 197–199 carbonitrides and nitrides, 5–7 cathode degradation, 226–228 cathode electrolyte interphase (CEI), 90 film, 133 layers, 45, 228 cell structuration with functional polymers (CSfP), 172 cellulose nanocrystals (CNC), 197 challenges and opportunities cost roadmap, 319–322 implications, 333 lithium-ion technology, 319–322 lithium-ion value chain, 317–319 market and players, 316–317 post-lithium (PLiT), 322–325 pre-cursor and raw materials, demand and supply of, 326–333 solid-state technologies, 325 361

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00361

362

coating layered cathode materials, 53–54 Li+ conducting coating materials, 57–59 materials, 54–57 combined heat and power unit (CHP), 257 community energy storage (CES), 351 composite/hybrid electrolyte, 86–90 conventional composition (CC) cathode, 33 core–shell concentration-gradient (CCG), 30 cross-linked gel polymer electrolyte, 124 cycling-induced aging irreversible capacity loss (ΔQir), 233–236 non-destructive approach, 238–240 postmortem analyses, 236–238 DC coupled PV battery system, 253–254 density functional theory (DFT), 133 dibutyl phthalate (DBP), 115 diethyl carbonate (DEC), 106 diethyl phenylphosphonite (DEPP), 152 dimethyl carbonate (DMC), 106, 131 dimethyl phenylphosphonite (DMPP), 145, 147 dioctyl adipate (DOA), 115 diphenyl carbonate (DPC), 151 donor number (DN), 76 doping layered cathode materials, 60–61 anion doping, 65–66 Li site dopants, 65 methods, 61–62 solid solution dopants, 62–65

Subject Index

electric vehicles (EVs), 26 electrochemical impedance spectroscopy (EIS), 118 electrode–electrolyte interphases (EEI), 90 electrode/electrolyte interfacial chemistry and compatibility, 90–91 electromotive force (EMF) determination, 229–230 electron conductivity, 174 electron-probe X-ray microanalysis (EPMA), 30 film-forming electrolyte additives high-voltage nickel-rich layered oxides, 149–153 spinel oxides LiNi0.5Mn1.5O4, 144–149 gel polymer electrolytes characteristics and requirements electrochemical stability, 118 ionic conductivity, 117–118 lithium transference number, 118–119 mechanical properties, 119–120 components of, 104 inorganic filler, 116–117 lithium salt, 105–106 organic solvent, 106–107 poly(ethylene oxide) (PEO), 108–109 poly(methyl methacrylate) (PMMA), 112–115 poly(vinyl chloride) (PVC), 115–116 poly(vinylidene fluoride-co-hexa­ fluoropropylene) (PVdF-HFP), 111–112

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00361

Subject Index

poly(vinylidene fluoride) (PVdF), 111–112 polyacrylonitrile (PAN), 109–110 polymer, 107–108 preparation of in situ cross-linking, 122–123 hot melting, 121 porous membrane, immersion of, 122 solution casting, 120–121 types of electrolytes, 103–104 generator coupled PV battery system, 254–255 graphite intercalation compound (GIC), 137 grid connected PV applications commercial behind-the-­meter PV battery storage systems increased resilience, 257 load balancing, 256–257 district battery storage systems grid services, 258 sector coupling, 257–258 factors affecting bankability of, 258–260 residential behind-the-meter PV battery storage systems, 253 AC coupled PV battery system, 254 DC coupled PV battery system, 253–254 energy management, 255–256 generator coupled PV battery system, 254–255 HAADFSTEM. See high-angle annular dark field scanning transmission electron microscopy (HAADFSTEM)

363

high voltage cathode materials film-forming electrolyte additives, 144–153 salt/solvent substitution, 138–144 high voltage cathode/electrolyte, interphasal stability of, 133–138 high voltage electrode/electrolyte interphase, 131–133 high-angle annular dark field scanning transmission electron microscopy (HAADFSTEM), 12 high-energy anode materials carbon materials, 2–5 carbonitrides and nitrides, 5–7 Li metal advantages of, 13–14 challenges of, 14 manufacturing techniques for, 17 solid-electrolyte interphases, 14–17 solid-state electrolyte, 14–17 metals, metal oxides and metal sulfides, 10–13 silicon materials, 9–10 volume expansion and solutions, 7–9 highest occupied molecular orbital (HOMO) energy, 139, 223 household batteries, 335–338 back-up power, 348 battery applications, 339–342 combination of applications, 348 community scale, 351–352 demand load-shifting, 345–346 demand peak-shaving, 346–347 different lithium-ion battery technologies, 349–351 electricity prices, 339–342 global impact

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00361

364

household batteries (continued) energy system cost, 353–354 environmental impacts, 354–355 security of energy supply, 355 wholesale electricity market, 355–356 PV self-consumption, 342–344 renewable energy curtailment, 344–345 system, 338–339 hybrid electric vehicles (HEVs), 220 inorganic electrolytes, solid electrolytes properties of, 83–84 research highlights in, 84–86 inorganic solid electrolytes (ISEs), 73 International Renewable Energy Agency (IRENA), 349 ionic conductivity, 117–118 ionic transport, solid electrolytes at macroscopic level, 74–75 at microscopic level, 75–76 irreversible capacity loss (ΔQir), 233–236 KIBs. See potassium-ion batteries (KIBs) layered Ni-rich cathode materials advanced doping approaches, 35–36 core–shell and concentration gradient, 29–35 diverse surface modification approaches, 36–39 next generation cathode material, 28–29 next generation lithium-ion battery, 26–27 perspective for, 39–40

Subject Index

layered oxide cathode materials blended cathodes, 50–51 LiCoO2 (LCO), 46–48 lithium rich layered oxides (LLOs), 50–51 modification methods, 50–53 coating, 53–59 doping, 60–66 NCA, 49–50 NCM, 48–49 liquid non-aqueous electrolytes high voltage cathode materials film-forming electrolyte additives, 144–153 salt/solvent substitution, 138–144 high voltage cathode/ electrolyte, interphasal stability of, 133–138 high voltage electrode/ electrolyte interphase, 131–133 lithium rich layered oxides (LLOs), 50–51 lithium salt, 105–106 lithium transference number, 118–119 lithium-ion conductive glass ceramics (LICGC), 86 manufacturing processes, 263 cell manufacturing cell balance, 281–282 electrolyte filling, 282–285 formation and conditioning, 282–285 electrode manufacturing processes coating techniques, 274–278 electrode drying, 278–281 inks and mixing methods, 266–274 materials, 264–266

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00361

Subject Index

metals, metal oxides and metal sulfides, 10–13 methyl phenyl carbonate (MPC), 151 Mg batteries, 187 MXenes, 5–7 negative electrode materials Na-ion batteries alloying materials, 200 carbonaceous materials, 197–199 titanium oxides and phosphates, 199–200 potassium-ion batteries (KIBs) alloying materials, 205–207 carbonaceous materials, 204–205 next generation cathode material, 28–29 next generation lithium-ion battery, 26–27 original equipment manufacturers (OEMs), 136 PBAs. See Prussian blue analogues (PBAs), 195–197, 202–204 perylene tetracarboxylic dianhydride (PTCDA), 197 poly(ethylene oxide) (PEO), 108–109 PEO-based SPEs, solid electrolytes, 77–80 non-PEO-based polymer electrolytes, 80–81 single lithium-ion conducting solid polymer electrolytes, 81–82 poly(methyl methacrylate) (PMMA), 112–115 poly(vinyl chloride) (PVC), 115–116 poly(vinylidene fluoride-co-hexafluoropropylene) (PVdF-HFP), 111–112

365

poly(vinylidene fluoride) (PVdF), 111–112 polyacrylonitrile (PAN), 109–110 polymer matrix, 87 polymer electrolytes (PEs), 76–77 polymer-structured batteries counter fire accidents, 177 resin suppresses expansion, 178 withstood drilling, 178–179 manufacturing method drying process in, 165–167 metallic current collector, 167–168 re-inventing the structure bipolar structure, 169–170 direction of current flow, 168–169 flexibility to withstand stress, 170–171 resin base, 171–172 active material, electrodes, and cell, 172–174 electron conductivity, 174 size and weight, 174–175 various functions, 172 safety improvement, 164–165 simple cell manufacturing process, 175–176 potassium batteries. See also potassium-ion batteries (KIBs) alternative rechargeable batteries, 186 K-ion batteries, 188–190 Mg batteries, 187 Na-ion batteries, 187–188 redox-flow battery (RFB), 185 sodium–sulfur (Na–S) batteries, 185 LIBs, battery size of, 182–184 material resources, 182–184

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00361

366

potassium-ion batteries (KIBs) energy density, 209 gravimetric capacity, 209 negative electrode materials alloying materials, 205–207 carbonaceous materials, 204–205 positive electrode materials layered transition metal oxides, 201–202 polyanionic compounds, 202 Prussian blue analogues (PBAs), 202–204 Prussian blue analogues (PBAs), 195–197, 202–204 safety advanced thermal management systems active thermal management, 310 flame arresting devices and shielding, 309 interstitial heat sinks, 308–309 interstitial insulating materials, 309 phase change materials (PCMs), 310 cell level and system level factors, 291–294 future trends in, 311–312 high profile thermal runaway field failures, 297–299 motivation to focus on, 296–297 primary thermal runaway testing techniques accelerating rate calorimetry (ARC), 300–301 bomb calorimetry, 301 cone calorimetry, 302 copper slug calorimetry (CSC), 300

Subject Index

fractional calorimetry, 301–302 system level testing, 303 trigger techniques, 302–303 thermal runaway behavior, 291–294 thermal runaway modeling techniques 2D and 3D modeling, 307–308 thermal runaway kinetic relationships, 304–307 thermal runaway, characteristics of, 294–296 scanning electron microscopy (SEM), 30 silicon materials, 9–10 sodium-ion batteries (SIBs), 182, 187–188, 189 negative electrode materials alloying materials, 200 carbonaceous materials, 197–199 titanium oxides and phosphates, 199–200 positive electrode materials layered 3d transitionmetal oxides, 190–194 polyanionic compounds, 194–195 Prussian blue analogues (PBAs), 195–197 solid electrolyte interphase (SEI) film, 133 solid electrolytes composite/hybrid electrolyte, 86–90 electrode/electrolyte interfacial chemistry and compatibility, 90–91 inorganic electrolytes properties of, 83–84

View Online

Published on 14 March 2019 on https://pubs.rsc.org | doi:10.1039/9781788016124-00361

Subject Index

research highlights in, 84–86 ionic transport in at macroscopic level, 74–75 at microscopic level, 75–76 PEO-based SPEs, 77–80 non-PEO-based polymer electrolytes, 80–81 single lithium-ion conducting solid polymer electrolytes, 81–82 solid polymer electrolytes, 76–77 solid polymer electrolytes (SPEs), 73 solid-electrolyte interphases, 14–17 solid-state electrolyte, 14–17 thermal runaway behavior, 291–294 characteristics of, 294–296 modeling techniques 2D and 3D modeling, 307–308 thermal runaway kinetic relationships, 304–307

367

titanium oxides and phosphates, 199–200 ultrathin BN/grapheme, 17 UV irradiation, 108, 280 vehicles battery electric vehicles (BEVs), 317 electric vehicles (EVs), 26 hybrid electric vehicles (HEVs), 220 viscosity, 82, 106, 111, 113, 189, 190, 269 voltage. see high voltage cathode materials volatile organic solvents, 103 vinylene carbonate (VC), 107, 151 weight reduction, 174-175, 208 wholesale electricity market, 355-356 X-ray photoelectron spectroscopy (XPS), 145, 221 Young’s modulus, 4, 77, 170, 171 ZEBRA (ZEolite Battery Research Africa), 185