Investigation of Polymers Used in Lithium. Oxygen Batteries as Electrolyte and. Cathode Materials

Transcription

1 Investigation of Polymers Used in Lithium Oxygen Batteries as Electrolyte and Cathode Materials A thesis presented for the degree of Master by Research By Jinqiang Zhang, B. Sc. University of Technology, Sydney 2013

2 Certificate of original authorship I certify that the work in this thesis has not previously been submitted for a degree nor has it been submitted as part of requirements for a degree except as fully acknowledged within the text. I also certify that the thesis has been written by me. Any help that I have received in my research work and the preparation of the thesis itself has been acknowledged. In addition, I certify that all information sources and literature used are indicated in the thesis. Jinqiang Zhang May 2013 i

3 Acknowledgements Although it s only been one and half year since I started my Master research, I have received many help from many people who supported me finishing this Master degree project. First of all, I would like to express my sincere gratitude to my supervisor, Prof Guoxiu Wang, for the guidance, support, and encouragement. I cannot thank you enough for all the advice that leads to the improvement of me, all the patience when I made a mistake, and all the concern about my research and life. I m really grateful to have you as my teacher and supervisor. I wish to thank all my colleagues in the research team of Centre for Clean Energy Technology, Dr Hao Liu, Dr Xiaodan Huang, Dr Bei Wang, Dr Bing Sun, Dr Ying Wang, Mr Dawei Su, Mr Kefei Li, Mr Anjon Kumar Mondal, Mr Shuangqiang Chen, Mr Yiying Wei, and Mr Xiuqiang Xie for the help both in my research and life during my Master period. Special thanks would address to Dr Yueping (Jane) Yao, for the administrative assistance and lab management as well as the great support for our life. It is a great pleasure to work with all of you and I wish you all the best luck. I wish to thank Rochelle Seneviratne for the assistance and patience during my study. I am also grateful for the training and support in Faculty of Science from Dr Ronald Shimmon, Dr Linda Xiao, and all the MAU staff. The help from all the teachers and professors from school of chemistry and forensic science are much appreciated. Financial support provided by the Australian Research Council (ARC) through the ARC Linkage project (LP ), and ARC Discovery Project (DP ) is gratefully acknowledged. ii

4 Finally I would like to thank my family, my parents and my brother, for your help and support for me. You always have faith in me even when I was confused about my life and future in the toughest time. It is your loves that make me the person I am and allow me to chase my dreams. Thank you and I love you all. iii

5 Table of Contents Certificate of original authorship... i Acknowledgements... ii Table of Contents... iv List of Figures... viii List of Tables... xiii Abstract... xiv Introduction... 1 Chapter 1 Literature Review Li-O 2 batteries Anode Electrolyte Cathode Catalyst Polymer electrolyte Solid polymer electrolyte Gel polymer electrolyte Conducting polymer Synthesis method Application Summary iv

6 Chapter 2 Experimental Methods Overview Materials and chemicals Material preparation In situ oxidation Solution casting method Material characterization X-ray Diffraction (XRD) Scanning electron microscope (SEM) Fourier transform infrared spectroscopy (FT-IR) Thermogravimetric Analysis (TGA) Electrode preparation and cell assembly Electrode preparation Cell assembly Electrochemical characterization Cyclic Voltammetry (CV) Electrochemical Impedance Spectroscopy (EIS) Linear Sweep Voltammetry (LSV) Galvanostatic Charge and Discharge Chapter 3 Low Molecular Weight Polyethylene Glycol Based Gel Polymer Electrolyte Used in Li-O 2 Batteries Introduction v

7 3.2 Experiment Preparation of PEG based GPEs Material characterization Electrochemical testing Results and discussion Summary Chapter 4 Investigation of PVDF-HFP Based Gel Polymer Electrolyte Used in Li-O 2 Batteries Introduction Experiment Preparation of PVDF-HFP based GPEs Material characterization Electrochemical testing Results and discussion Summary Chapter 5 Conducting Polymer-Doped Polypyrrole as An Effective Cathode Catalyst for Li-O 2 Batteries Introduction Experiment Synthesis of materials Characterization of samples Electrochemical measurements vi

8 5.3 Results and discussion Summary Chapter 6 Conducting Polymer Coated CNT Used in Li-O 2 Batteries with Enhanced Electrochemical Performance Introduction Experiment Synthesis of materials Characterization of samples Electrochemical measurements Results and discussion Summary Chapter 7 Conclusions General conclusion Outlook and future work References vii

9 List of Figures Figure 1-1 The gravimetric energy density of commonly used rechargeable batteries... 5 Figure 1-2 Schematic mechanism of Li-O 2 batteries during discharge and charge process Figure 1-3 Different types of Li-O 2 batteries based on different architectures... 9 Figure 1-4 Two models of reaction mechanisms of Li-O 2 batteries, (A) aqueous system and (B) non-aqueous system Figure 1-5 Three different types of electrolyte filling on cathodes, (A) flooding, (B) dry and (C) wetting Figure 1-6 Schematic mechanism of decomposition of PC electrolyte in Li-O 2 batteries Figure 1-7 Cycle performance of Li-O 2 batteries with TEGDME as electrolytes Figure 1-8 Schematic mechanism of discharge process on porous carbon cathodes Figure 1-9 The morphology study, discharge performance and discharge mechanism of a hierarchical graphene Figure 1-10 Schematic mechanism of (a) side reactions of carbon cathode and discharge products and (b) side reactions between electrolyte and carbon cathode Figure 1-11 Discharge/charge profiles (left) and cycle performance (right) of nano gold cathode in DMSO based electrolyte Figure 1-12 Discharge/charge profile (left) and cycle performance (right) of graphene cathode and carbon black cathode Figure 1-13 First galvanostatic charge of Li 2 O 2 oxidation for various Li O 2 cells Figure 1-14 Schematic mechanism of Li 2 O 2 and Li 2 O forming on MnO 2 catalyst Figure 1-15 Schematic mechanism of Li + movement through PEO based polymer electrolyte viii

10 Figure 1-16 Schematic mechanism of the addition of ceramic fillers and the effect of different particle sizes, (a) macro-size and (b) nano-size Figure 1-17 Schematic presentation for functional role of PDMITFSI ionic liquid on lithium deposition, (a) without and (b) with ionic liquid Figure 1-18 The structures of the most commonly used conducting polymers Figure 1-19 Conjugated orbitals formed in polyacetylene Figure 1-20 Schematic illustration of synthesis mechanism of PPy Figure 1-21 Schematic illustration of synthesis mechanism of (A) PPy nanotube and (B) PANI nanowire Figure 1-22 Cycling performance of PPy/FC at (a) constant current density of 50 mag -1 and (b) different current densities Figure 1-23 Discharge/charge profiles (left) and resistance (right) of the LiFePO 4 cathode (a) coated with PEDOT, (b) coated with PPy, (c) coated with C, and (d) pristine particles Figure 1-24 a) Electron-transfer pathway for LiFePO 4 particles partially coated with carbon. b) Designed ideal structure for LiFePO 4 particles with typical nano-size and a complete carbon coating. c) Preparation process for the LiFePO 4 /carbon composite including an in situ polymerization reaction and two typical restriction processes Figure 1-25 (A) Morphology and cycle performance of PPy cathode [70], (B) nitrogendoped graphene derived from PANI and (C) Performance of PEDOT catalyst Figure 2-1 Schematic illustration of the whole experiment process Figure 2-2 The preparation process of PPy when (NH 4 ) 2 S 2 O 8 was used as oxidant Figure 2-3 Schematic drawing of Bragg s law Figure 2-4 An example TGA result of polypyrrole coated silicon Figure 2-5 The structure of a Li-O 2 battery ix

11 Figure 2-6 A typical ESI Nyquist curve of a battery system Figure 2-7 A typical result of LSV measurement Figure 2-8 An example charge and discharge curve of a Li-O 2 battery Figure 3-1 The typical molecular structure of PEG or PEO Figure 3-2 Cyclic voltammetry results of Li/GPE/Li type cells with (a) PEG and (b) PEG with SiO 2 addition Figure 3-3 Linear sweep voltammetry results of (a) Li/GPE/SS and (b) Li/GPE/CB-air type cells Figure 3-4 The impedance spectra of PEG at different temperatures. (b) The calculated ionic conductivity of PEG at different temperatures Figure 3-5 (a) First discharge and charge profiles of Li-O 2 batteries with PEG, PEG- SiO 2, TEGDME as electrolytes. (b) Partial enlarged view of first discharge and charge profiles from mahg Figure 3-6 Discharge and charge profiles of Li-O 2 batteries with (a) PEG, (b) PEG- SiO 2, and (c) TEGDME as electrolytes at fixed capacity to 500 mahg Figure 3-7 Cycle profiles of Li-O 2 batteries with PEG, PEG-SiO 2, and TEGDME as electrolytes Figure 3-8 Structures of (a) PEG-based electrolyte and (b) PEG-SiO 2 -based electrolyte Figure 3-9 XRD pattern of PEG before and after made into polymer electrolyte Figure 3-10 XRD pattern of cathode after discharge in PEG polymer electrolyte Figure 4-1 The typical structure of PVDF-HFP Figure 4-2 The cyclic voltammetry curve of Li/GPE/Li typed cell with TEGDME based GPE as electrolyte x

12 Figure 4-3 Linear sweep voltammetry results of (a) Li/GPE/SS and (b) Li/GPE/CB-air type cells Figure 4-4 The calculated ionic conductivity of PEG at different temperatures Figure 4-5 The discharge and charge profiles in the first cycle of Li-O 2 batteries using different electrolyte Figure 4-6 Discharge and charge profiles of Li-O 2 batteries with TEGDME based GPE as electrolytes at fixed capacity to 500 mahg Figure 4-7 Cycle profiles of Li-O 2 batteries with PVDF-HFP based GPE and TEGDME as electrolytes Figure 4-8 Proposed structure of PVDF-HFP based GPE Figure 4-9 The discharge and charge profiles in the first cycle of PC (a) and DMSO (c) based GPEs and the cycling performance of PC (b) and DMSO (d) based GPEs Figure 5-1 The typical structure of PPy Figure 5-2 SEM images of the as-prepared (a) PPy-Cl and (b) PPy-ClO 4, and (c) FT-IR spectra of both PPy polymers Figure 5-3 The discharge-charge profiles and (b) cycling performance of PPy-Cl, PPy- ClO 4 and carbon black electrodes Figure 5-4 The mechanism of (a) oxygen activation of PPy and (b) doping-undoping process of PPy-Cl and PPy-ClO Figure 5-5 FT-IR spectra of (a) PPy-Cl, (b) PPy-ClO 4, and (c) carbon black electrodes before discharge, after discharge and after charge process Figure 5-6 (a) The charge-discharge profiles and (b) the cycling performance of carbon black electrodes with and without LiCl additive Figure 5-7 Schematic mechanism of discharge process on cathode with LiCl addition xi

13 Figure 6-1 The typical structure of PEDOT Figure 6-2 The SEM images of (a) the bare CNT, the as-prepared (b) PPy/CNT 1:2, (c) PPy/CNT 1:1, (d) PPy/CNT 2:1, (e) PEDOT/CNT 1:1, and (f) FT-IR spectra Figure 6-3 The TGA spectra of as-prepared (a) PPy/CNT 1:2, (b)ppy/cnt 1:1, (c) PPy/CNT 2:1, and (d) PEDOT/CNT 1: Figure 6-4 (a) The discharge and charge profiles of as-prepared PPy/CNT 1:2, PPy/CNT 1:1, PPy/CNT 2:1, PEDOT/CNT 1:1, and CNT electrodes. (b) Partially enlarged profiles of as-prepared electrodes with capacity of 500 mah g Figure 6-5 The cycling performance of as-prepared prepared PPy/CNT 1:2, PPy/CNT 1:1, PPy/CNT 2:1, PEDOT/CNT 1:1, and CNT electrodes Figure The schematic mechanism of (a) PPy/CNT during cycling, and (b) the block of O 2 from PEDOT structure xii

14 List of Tables Table 2-1 Materials and chemicals used in the research project Table 5-1 EDS results of PPy-Cl and PPy-ClO 4 electrodes after cycling xiii

15 Abstract It has been well established that the electrolytes and cathodes have a significant effect on the electrochemical performance of lithium oxygen batteries. In this Master project, polymers were employed as electrolyte and cathode materials due to their unique superior properties. Using different methods, we synthesized suitable gel polymer electrolytes and conducting polymer catalysts for lithium oxygen batteries. Techniques such as field emission gun scanning electron microscopy, X-ray diffraction, and Fourier transform infrared spectroscopy were used to characterize the physical properties. Electrochemical analyses including the galvanostatic discharge and charge method, the cyclic voltammetry, the linear sweep voltammetry and the impedance spectra were conducted to determine the electrochemical performance for the as-prepared materials. Gel polymer electrolytes based on low molecular weight polyethylene glycol were prepared and used as electrolyte in lithium oxygen batteries. The as-prepared polymer electrolytes showed improved stability compared with liquid electrolytes and exhibited good performance in lithium oxygen batteries. Additionally, the addition of ceramic filler SiO 2 was found to reduce the stability of polymer electrolyte towards oxygen reduction reaction although higher ionic conductivity was obtained. Polyethylene glycol based gel polymer electrolyte without SiO 2 addition exhibited excellent cycling performance and it could be used for achieving long-life lithium oxygen batteries. Poly(vinylidene fluoride-co-hexafluoropropylene) based gel polymer electrolytes were prepared by solvent casting and employed as electrolytes in lithium oxygen batteries. The stability of the gelled electrolyte with tetraethylene glycol dimethyl ether has been greatly increased than the liquid one. The as-prepared polymer electrolyte was demonstrated excellent cycling performances. This thesis also investigated the effect of xiv

16 different plasticizers on the performance of lithium oxygen batteries. The reason could lie on the interactions among the components when the gelled structure was set. The tetraethylene glycol dimethyl ether based gel polymer electrolyte showed the best electrochemical performance and can be used for long-life lithium oxygen batteries. Polypyrrole conducting polymers with different dopants have been synthesized and applied as the cathode catalysts in lithium oxygen batteries. Polypyrrole polymers exhibited an effective catalytic activity for oxygen reduction in lithium oxygen batteries. It was discovered that dopant significantly influenced the electrochemical performance of polypyrrole. The polypyrrole doped with Cl - demonstrated higher capactity and more stable cyclability than that doped with ClO - 4. Polypyrrole conducting polymers also exhibited higher capacity and better cycling performance than that of carbon catalyst. Conducting polymer coated carbon nanotubes were synthesized and used as catalysts in lithium oxygen batteries. It was found that both polypyrrole and poly(3,4- ethylenedioxythiophene) coated carbon nanotubes could provide high cycling performance while polypyrrole based one exhibited higher capacities. The ratio of conducting polymer coating also affected the electrochemical performance of lithium oxygen batteries. The conducting polymer coated carbon nanotubes also showed better performance than the bare carbon nanotubes. xv

17 Introduction Due to the wide use of fossil fuels as a source of the energy supply in the last few centuries, pollution, the limiting sources of fossil fuels and global warming are the main issues for today s society. Green energy devices such as batteries are in great demand to suppress CO 2 emission and air pollution, and electric vehicles are designed to replace internal combustion engine cars. Finding alternative energy storages has never been so important ever. Lithium-ion (Li-ion) batteries, are one form of green energy storage devices, and have been demonstrated as the most advanced alternative electrochemical power sources in the last two decades. However, conventional Li-ion batteries cannot meet all the requirements for usage in large-scaled applications such as the electric vehicles due to a limitation on its theoretical energy density. Lithium-oxygen (Li-O 2 ) batteries have attracted intensive research recently as an alternative choice. The use of oxygen directly drawn from air as the cathode reactant makes their theoretical energy densities almost equal to gasoline. However, the commercialization of Li-O 2 batteries has not yet been realized. The development of effective electrolytes and air cathodes are the critical challenges to achieve high performance Li-O 2 batteries. Many approaches have been made and polymers have been researched as the main components for Li-O 2 batteries due to their superior properties such as stability and unique electrochemical properties. The main purpose of this thesis is to find effective ways to improve the performance of Li-O 2 batteries by employing polymer electrolytes and conducting polymers in this battery system. The materials were successfully prepared and characterized as electrolyte and cathode materials. The outline of each chapter is listed below, 1

18 Chapter 1 contains an introduction to Li-O 2 batteries, polymer electrolytes, and conducting polymers. Recent progresses in anode, electrolyte, cathode and catalyst materials used in Li-O 2 batteries are reviewed. Research in polymer electrolytes, both solid and gel ones, used in lithium batteries are also introduced. This chapter includes recent studies on conducting polymers such as polypyrrole, polyamine, and poly(3,4-ethylenedioxythiophene) and the application in the lithium batteries at the same time. Chapter 2 provides the methodology of research applied in this thesis. The material preparation methods, physical characterization methods, the electrode preparation methods and electrochemical testing methods are illustrated in this chapter. The preparation methods were solution casting for polymer electrolytes and in situ chemical synthesis method for conducting polymers. Field emission gun scanning electron microscopy was used for investigating morphology and X-ray diffraction was employed to determine the crystal structures of asprepared polymer electrolytes, conducting polymers and discharge products. Thermogravimetric analysis was used to determine the content of each component in composite materials. Galvanostatic discharge and charge and a series of electrochemical characterizations such as cyclic voltammetry, linear sweep voltammetry, and impedance spectra were performed to evaluate the properties of as-prepared materials. Chapters 3 and 4 report the use of polyethylene glycol and poly(vinylidene fluoride-co-hexafluoropropylene) based gel polymer electrolytes used in Li-O 2 batteries, respectively. In Chapter 3, the polyethylene glycol based polymer electrolyte was prepared and investigated. The effect of nano-sized silica addition was also investigated. In Chapter 4, the electrochemical performances 2

19 of poly(vinylidene fluoride-co-hexafluoropropylene) based polymer electrolytes with different plasticizer such as propylene carbonate, dimethyl sulfoxide, and tetraethylene glycol dimethyl ether were investigated. Chapters 5 and 6 describe the experiments on the use of conducting polymers in Li-O 2 batteries. The synthesis and comparison of Polypyrrole with different dopants were reported in Chapter 5. The catalytic mechanism was also discussed. In Chapter 6, polypyrrole coated carbon nanotubes with different polymer ratio were prepared and tested for Li-O 2 batteries. Poly(3,4- ethylenedioxythiophene) coated carbon nanotubes were also synthesized and compared with polypyrrole based ones. Chapter 7 concludes this thesis, and provide a brief summary of the previous work on the gel polymer electrolytes and conducting polymers and an outlook for future research of polymers used in Li-O 2 batteries. 3

20 Chapter 1 Literature Review 1.1 Li-O 2 batteries Due to the fast growth of economy and technology, finding large-scale energy storage is much more important today than ever. Coal-based and oil-based sources have been used for centuries because they can provide large quantity of energy to meet energy demands. However, due to the limited sources of coal and oil and pollution by CO 2 and other waste gas emissions, there is the great urgency to find more sustainable and environmentally friendly energy. Therefore, electrochemical devices for energy storage and conversion become more and more important. Devices including super capacitors, fuel cells and batteries are used to solve the problems raised by using coal- and oilbased energy sources. Rechargeable batteries are considered as good candidates for next generation energy storage devices. Figure 1-1 shows the energy densities of commonly used rechargeable batteries. Among all energy storage technologies, lithium batteries have drawn great interest because of their high energy density and light weight [1, 2]. Li-ion batteries have been already used in all kinds of portable electronic devices such as cell phones and laptops [3]. However, the theoretical energy density of Li-ion batteries is not high enough for electrical vehicles, which has limited their application. In order to meet the demands of society, new generations of batteries are urgently needed. 4

21 Figure 1-1 The gravimetric energy density of commonly used rechargeable batteries. Figure 1-1 The gravimetric energy density of commonly used rechargeable batteries [4]. Li-O 2 batteries, also known as Li-air batteries, have shown potentials for large scale applications such as electric vehicles because of their extremely high theoretic energy density (11140 kw kg -1 ), which is comparable to gasoline [4-8]. Another advantage of Li-O 2 batteries is that they do not need to storage cathode reactants inside the batteries. The fact that the reactant is oxygen and oxygen can be replenished at any time from air provides the possibility to reduce the weight of batteries and make them suitable for portable devices. A Li-O 2 battery is the combination of a lithium battery and a fuel cell, which has a lithium ion system and an oxygen system at the same time. A typical Li-O 2 battery consists of a lithium foil used as the anode, electrolyte which provides lithium 5

22 ion pathways between the anode and the cathode, and an air electrode to allow the flow of air into the system. During the discharge process, oxygen is consumed on the cathode and forms discharge products while Li + ions in electrolyte can be replenished by lithium anode. The reaction can be reversed during the charging process. The whole process is exhibited in Figure 1-2. Figure 1-2 Schematic mechanism of Li-O 2 batteries during discharge and charge process. Figure 1-2 Schematic mechanism of Li-O 2 batteries during the discharge and charge process [4]. Despite all the superiorities they can provide, current Li-O 2 batteries are far from satisfactory [4-8]. The main issues of Li-O 2 batteries are listed below, Low practical capacity due to insufficient porosity in the cathode electrodes for accommodating the discharge products Poor cycleablility due to the decomposition of organic electrolytes High reactivity of lithium metal in aqueous electrolyte Large discharge-charge over-potentials Side-reactions due to contaminates such as CO 2 and H 2 O 6

23 In order to solve all the above issues, a proper knowledge Li-O 2 battery is necessary. In the next section, each component of Li-O 2 batteries will be reviewed Anode The anode in a Li-O 2 battery provides Li + for electrochemical reactions. Lithium metal is usually used directly as anode. The anode reaction is shown below, (1-1) During the discharge process, lithium metal is oxidized and releases Li + into the electrolyte. The reaction is reversed during the charge process. Although metallic lithium is light-weight and has a very high energy density, the direct use of lithium metal as the anode material in Li-O 2 batteries still causes problems including dendrite growth during cycling and side reactions towards electrolyte and O 2 crossover from the porous cathodes [2, 8]. This may have a detrimental influence on the electrochemical performance of Li-O 2 batteries. Some research groups tried to replace lithium metal anode with lithium alloying compounds such as Li x Si [9] or partially charged LiFeO 4 [10] which are proven to have good cycling performance. Anther strategy is to process the lithium metal before use. Bruce et al. used 0.1 M LiClO 4 - propylene carbonate (PC) electrolyte to process lithium foil before it was used as the anode in a Dimethyl sulfoxide (DMSO) electrolyte [11]. It is believed that this process is effective to stabilize the lithium metal in DMSO electrolyte and to ensure good cycling performance at the same time. Besides processing the anode materials, a lithium protection layer can also be used, especially in aqueous electrolyte system, because lithium metal can aggressively react with water electrolyte and cause severe safety issues. Many Li + conducting but 7

24 electronically insulating membranes, such as LiSICON glass ceramics [8, 12, 13], are employed in aqueous Li-O 2 batteries. Polymer electrolyte is also a good candidature and will be introduced in the following section Electrolyte Electrolytes serve as a Li + pathway as well as a separator between anodes and cathodes. Therefore it is believed that ideal electrolytes for Li-O 2 batteries should have superior properties including high conductivity, wide electrochemical stability, acceptable lithium transference number, compatibility with both electrodes, stability towards battery reactions, high oxygen solubility, and low volatility as well [4, 5, 7]. There are four types of Li-O 2 batteries, depending on the architecture differences as shown in Figure 1-3. They are aprotic, aprotic, aqueous, solid, and mixed aqueousaprotic systems [2, 4, 8]. Although in all cases the discharge process involves the reactions with oxygen, the mechanism of each system is different from others depending on the electrolyte used. For simplicity, all four systems are divided into two types, aqueous and non-aqueous systems. In the non-aqueous systems, the oxygen reduction reaction products are insoluble in the electrolytes, which mean the products will deposited in the porous structures of cathodes and may hinder the oxygen transference and cause termination of battery reactions. Therefore, the actual capacity of non-aqueous Li-O 2 batteries mainly depends on the porosity of cathodes. In aqueous systems, the discharge products are soluble in the water systems. Figure 1-4 shows two models of reaction mechanisms of Li-O 2 batteries, which illustrates the difference clearly. The reactions in aqueous systems are often described as three-phase reactions, consisting of solid phase (cathodes and catalysts), liquid phase (electrolyte), and gas phase (O 2 ), while the ones in non-aqueous systems are called two-phase reactions 8

25 without the gas phase which means only the oxygen dissolved in the liquid electrolyte may be involved in the oxygen reduction reactions. The difference here between two systems is the solubility of discharge products which has been mentioned above. The battery reactions are listed below [14-16], (Aqueous acidic media) (1-2) (Aqueous alkaline media) (1-3) (Non-aqueous media) (1-4) (Non-aqueous media) (1-5) Figure 1-3 Different types of Li-O 2 batteries based on different architectures. Figure 1-3 Different types of Li-O 2 batteries based on different architectures [4]. 9

26 Figure 1-4 Two models of reaction mechanisms of Li-O 2 batteries, (A) aqueous system and (B) non-aqueous system Figure 1-4 Two models of reaction mechanisms of Li-O 2 batteries, (A) aqueous system and (B) non-aqueous system [7]. Despite the difference of reactions of both non-aqueous and aqueous media, the performance of Li-O 2 batteries with both electrolytes are affected by the amount of electrolytes filling of the cathodes. Generally there are three kinds of filling, known as flooding, dry, and wetting. Figure 1-5 shows the three fillings of electrolytes. It is believed that both oxygen dissolved in electrolytes and in gas phase can participate in the battery reactions. However, it is obvious that oxygen in solution is less mobile than one in the gas phase. Therefore, when the electrolyte floods the cathode, the kinetic is very low because of the slow dissolving process of oxygen. The oxygen reduction reaction more likely happens on the air side of the cathode. On the other hand when the amount of electrolyte is not sufficient, or dry electrolyte, the gas oxygen is very easy to penetrate into cathode. However, the Li + cannot reach the interface of electrolyte and cathode, which also leads to the low capacity of Li-O 2 batteries. Therefore, the ideal amount of electrolyte is just wetting the cathode to provide sufficient Li + for oxygen reduction reaction and allow oxygen transferring into cathode at the same time [17]. Zhang et al. has provided evidence about the effects of electrolyte filling towards performance of Li-O 2 batteries [18]. 10

27 Figure 1-5 Three different types of electrolyte filling on cathodes, (A) flooding, (B) dry and (C) wetting Figure 1-5 Three different types of electrolyte filling on cathodes, (A) flooding, (B) dry and (C) wetting [7]. Since the first report on non-aqueous electrolytes used in Li-O 2 batteries by Abraham and Jiang [19], this field has been a heated topic and attracted much attention. Generally speaking, there are three different types of non-aqueous electrolytes, such as organic liquid solvent electrolyte, hydrophobic ionic liquids and polymer electrolyte. All three types share the same mechanism displayed in equation (4) and (5). Electrolytes based on alkyl carbonate solvents including propylene carbonate (PC), dimethyl carbonate (DMC), ethylene carbonate (EC) and so on, are initially used in Li- O 2 battery owing to their low volatility, good Li compatibility, high ionic conductivity, oxygen solubility, and large operating window [5]. These carbonate-based electrolytes have been widely used in the conventional Li-ion batteries However, recent studies proved that organic carbonates are not suitable for long-life Li-O 2 batteries because the discharge by-products other than Li 2 O 2 are formed in these systems [20-22]. They believe superoxide radicals are formed during discharge process and these species attack the carbonates through nucleophilic reactions. It results in the formation of H 2 O, CO 2, Li 2 CO 3 and lithium alky-carbonates including HCO 2 Li and CH 3 CO 2 Li. A possible mechanism of decomposition of PC is exhibited in Figure 1-6. Although these products 11

28 can be oxidized in the charge process with evolution of H 2 O and CO 2, the whole battery suffers from low cycleability and fading of capacity because of the starvation of electrolyte and accumulation of products on the cathode surface. Figure 1-6 Schematic mechanism of decomposition of PC electrolyte in Li-O 2 batteries Figure 1-6 Schematic mechanism of decomposition of PC electrolyte in Li-O 2 batteries [22]. Due to the side reactions exhibited above, electrolyte based on carbonate solvents are no longer favourable in Li-O 2 batteries despite its superiorities. In order to achieve long life Li-O 2 batteries, stability of electrolyte should be sufficient. The use of additives or cosolvents is a way to improve the stability and performance of carbonate-based electrolytes. For instance, trifluoroethlphosphates and phosphite have been added into the Li-triflate-PC/DME or Li-triflate-PC system and is found to have superior performance during discharge and charge process [23, 24]. However, addition of such additives in electrolyte causes increase of the viscosity which finally affects the ionic conductivity. Therefore finding a substitute for carbonate-based electrolyte is needed. Ether-based electrolytes, known for their better stability than carbonate-based ones, are tested in Li-O 2 batteries and they have been proven that Li 2 O 2 are formed in first few cycles [25, 26]. For instance, Read employed 1,3-dioxolane (DOL) and 1,2-12

29 dimethoxyethane (DME) in Li-O 2 batteries in 2006 [27]. Xu et al. also used several ether-based solvents such as DME, diethylene glycol dimethyl ether (DG), diethylene glycol diethyl ether (EDG), diethylene glycol dibutyl ether (BDG), 1,2-diethoxyethane (DEE), and 1-tert-butoxy-2-ethoxyethane (BEE) and compared the performance of each candidate. In 2012, tetra(ethylene) glycol dimethyl ether (TEGDME) was found to be stable towards oxygen reduction reaction [28] and Jun et al. could even achieve at least 30 cycles at 500 mag -1 [29]. The result is exhibited in Figure 1-7. However, ether-based electrolytes are found to have some negative properties limiting their applications in Li- O 2 batteries, such as high viscosity for TEGDME, and high volatility for DME. Moreover, recent researches indicate that ether-based electrolytes are still not stable enough towards superoxide radicals. They claimed that even though Li 2 O 2 was detected in first discharge, the amount of Li 2 O 2 reduces significantly and the discharge products were then dominated by electrolyte composition products [30]. Bryantsev et al. used theoretical calculation to predict the autoxidation stability of ether- and amide-based electrolyte solvents and found that every electrolyte, more or less, suffers from decomposition in Li-O 2 batteries [31]. In this way, ether-based electrolytes may not be the best choices for Li-O 2 batteries. Figure 1-7 Cycle performance of Li-O 2 batteries with TEGDME as electrolytes Figure 1-7 Cycle performance of Li-O 2 batteries with TEGDME as electrolytes [29]. Current density is 500mA g

30 Dimethyl sulfoxide (DMSO) based electrolyte was used by Xu et al. and showed the possibility of long-term use in Li-O 2 batteries [32]. Thotiyl et al. also demonstrated although accompanied by small decomposition during cycling, DMSO is much more stable towards oxygen reduction reaction than ether-based ones [33]. Xu et al. also used tetramethylene sulfone (TMS) based electrolytes in Li-O 2 batteries and showed the stability towards cycles [34]. Another approach is the use of silanes as electrolytes. It is believed that silane-based electrolytes can provide higher stability in Li-O 2 batteries due to the presence of Si-O groups [35, 36]. Hydrophobic ionic liquids (ILs) are also considered to be used in Li-O 2 batteries. ILs are known for their lithium compatibility, low volatility, and also stability and showed superior performance when employed in Li-O 2 batteries [37-39]. Recently, ILs have also been used in polymer electrolytes which will be mentioned later in polymer electrolyte section. Although stability of electrolytes is an important issue for practical usage, it is not the only factor that can affect the performance of Li-O 2 batteries. It is believed that oxygen solubility, oxygen diffusion rate [40], water contents [41], lithium salts [42-44], and even binders [45] all have great influence on the performance. Therefore, choosing suitable electrolytes for non-aqueous Li-O 2 batteries is complicated and still needs great efforts. Unlike non-aqueous electrolytes with so many choices, the choice of aqueous electrolyte is limited only to acidic or basic solutions, either weak or strong. The most attractive properties of Li-O 2 batteries with aqueous electrolytes are the stability of electrolyte solution and the solubility of discharge products which differs from products in non-aqueous ones. However, due to the highly reactive property of lithium metal 14

31 towards water solution, it is not surprised that most research focus on the protection of the lithium anode. Typically, Li + -conducting but electronically insulating membranes like LISICON or NASICON are used between anodes and electrolytes [46, 47]. Without these protecting layers, the aqueous Li-O 2 batteries cannot function. Another difference between aqueous electrolytes and non-aqueous electrolytes is the participation of electrolyte solvents in the battery reactions. As discharge process proceeds, the solvents are continuously consumed and the amount of discharge product LiOH gradually increases. However, although LiOH can dissolve into aqueous electrolyte, saturation can be finally reached as the battery reaction proceeds. As a result, the product LiOH precipitates on the surface of cathodes and protective ceramic membranes which eventually leads to the clogging of porous structure and termination of battery reaction. This often causes low capacity and efficiency. In order to improve the performance of aqueous Li-O 2 batteries, several solutions have been come up with [48]. A flow cell construction has been used to replenish the electrolyte with fresh solution continuously and LiOH is drawn from electrolyte to make lithium metal to replenish the lithium anode [12, 49]. In this way the battery capacity is highly improved and it can also achieve good efficiency. Besides mechanical method to refresh electrolytes, blocking membranes can be employed between cathodes and electrolytes. These blocking membranes are also known as anion-exchange membranes which allow OH - formed in the cathode to be transport out and keep blocking Li + from entering the cathode at the same time. Many have also been demonstrated to have the ability to block any carbonates which can form Li 2 CO 3 in the electrolyte [8]. Another approach is to add a third electrode into the system for oxygen evolution on recharging. It is found to be much more efficient oxidizing LiOH than allowing LiOH precipitate onto the cathodes [15]. Using an acidic media as electrolyte is also a way to prevent LiOH from blocking 15

32 the cathode porosity corresponding to equation (2). However, when the battery reaction proceeds, the ph of electrolyte gradually increases and finally leads to the precipitate of LiOH in the electrolytes. This means a great achievement for producing long-life and effective aqueous Li-O 2 batteries is still needed. Among all designs of Li-O 2 batteries, the system with non-aqueous media is more preferred because of the stability, rechargeability and also safety Cathode It is believed that cathodes, also known as air electrodes, play important roles in Li-O 2 batteries because the battery reactions, both oxygen reduction and oxygen evolution reactions, happen in the cathodes. The most commonly used cathode matrix is carbon due to its intrinsic conductivity and light weight. According to the mechanisms of Li-O 2 batteries shown in Figure 1-8, the characteristics such as porosity, surface area and morphology of carbon can affect the performance greatly. Since the discharge products of non-aqueous Li-O 2 batteries cannot dissolve into electrolyte and precipitate in the pores of cathodes which usually hinder the diffusion of oxygen, electrolyte and even electrons, the capacity of Li-O 2 batteries in practical use is far less than theoretical calculation. Therefore the porosity and pore volume of electrodes can be very important for maintain low over-potentials and high capacity [50, 51]. However, the relationship between surface area and discharge capacity is still not clear [52, 53]. 16

33 Figure 1-8 Schematic mechanism of discharge process on porous carbon cathodes Figure 1-8 Schematic mechanism of discharge process on porous carbon cathodes [4]. However, it is believed that the effective use of pores, not the size or total volume of pores determines the highest performance of Li-O 2 batteries. It is believed that ideal pore structures of cathodes should be able to accommodate a large amount of insoluble discharge products without hindering the transportation of oxygen, Li + and even electrons. Among all porous structures, mesoporous structures seem to have the most suitable properties for Li-O 2 batteries. For instance, mesoporous carbon template [54, 55], mesocellular carbon foam [52], and highly mesoporous nitrogen-doped carbon [56, 57] are used in Li-O 2 batteries and proven to have enhanced performance because of the mesoporous structures. Carbon nantubes [58-62], and carbon nanofibers [59, 63] are also used in Li-O 2 batteries. It is believed that 1-Dimensional structures like nanofibers or nanotubes can also help maintain structures with sufficient porosity for oxygen diffusion and accommodation of discharge products. Graphene nanosheets (GNSs), as one type of carbon materials, have attracted great attention for Li-O 2 battery application because of their unique morphology and 17

34 structure. Besides the microporous channels for oxygen transportation, GNSs also have plenty of defects which help facilitate the formation of discharge products which can provide very high capacity and good cycle performance at the same time [64-67]. Xiao et al. used a hierarchical graphene structure in Li-O 2 battery and achieved extremely high capacity [64]. The result is shown in Figure 1-9. Figure 1-9 The morphology study, discharge performance and discharge mechanism of a hierarchical graphene Figure 1-9 The morphology study, discharge performance and discharge mechanism of a hierarchical graphene [64]. Current density is 0.1 ma cm -2. However, there are also some issues for using carbon matrices in Li-O 2 batteries. The intrinsic reactive properties of carbon make it possible to react with electrolytes and discharge products. The side reactions affect the performance of Li-O 2 batteries greatly. McCloskey et al. have demonstrated that during discharge carbon matrix can react with Li 2 O 2 to form Li 2 CO 3 and this leads to higher charge over-potential and low cycleability [68]. This idea was further confirmed by Bruce s group [33]. They believed that the stability of carbon also depended on the hydrophobicity/hydrophilicity of the carbon surface. Both of these results are shown in Figure

35 (A) (B) Figure 1-10 Schematic mechanism of (A) side reactions of carbon cathode and discharge products and (B) side reactions between electrolyte and carbon cathode Figure 1-10 Schematic mechanism of (A) side reactions of carbon cathode and discharge products [68] and (B) side reactions between electrolyte and carbon cathode [33]. In order to overcome this problem, carbon-free electrodes are used to replace carbon as cathodes in Li-O 2 batteries. Cui et al. successfully prepared a free-standing Co 3 O 4 electrode and used it in Li-O 2 [69]. The capacity and cycle performance are acceptable. In 2012, the same group introduced a tubular polypyrrole based cathode for Li-O 2 and exhibited excellent performance [70]. Peng et al. used nano gold particles instead of carbon as the matrix of cathode and achieved 100 cycles with only 5% drop of capacity which is shown in Figure Using surface coating of carbon is also a good idea for preventing side reactions [71]. 19

36 Figure 1-11 Discharge/charge profiles (left) and cycle performance (right) of nano gold cathode in DMSO based electrolyte Figure 1-11 Discharge/charge profiles (left) and cycle performance (right) of nano gold cathode in DMSO based electrolyte [11]. Current density is 500 mag -1. Besides the decomposition of carbon matrix during discharge and charge, other factors also influence the performance of Li-O 2 batteries greatly. Discharge products deposited on the surface of cathodes are usually insulators. As a result, the resistance to electron transport of cathode continuously increases as the battery reaction proceeds. This leads to the increase of electrode polarization and over-potential [59, 72]. Other factors such as decomposition of binders also influence the performance greatly [73, 74] Catalyst Li-O 2 batteries nowadays still suffer from many drawbacks, limiting their applications, such as low cycleability, large over-potential, low discharge-charge efficiency, and low practical capacity. In order to achieve good performance, catalysts are employed in Li- O 2 batteries. It is known that electrocatalysts can sufficiently reduce the over-potential of charge and discharge reactions, thus increasing round trip efficiency and improving cycling performance [2, 75]. The most commonly used catalysts in Li-O 2 batteries are carbon catalysts, noble metals, transition metals and transition metal oxides. Some carbon materials have the ability to catalyse the battery reactions. Graphene (GE), known for its unique structure and property, has been widely studied in Li-O 2 batteries [65, 66]. The edge effect and defects of GE is considered to have promoting effect for 20

37 oxygen reduction reaction and formation of discharge products which is beneficent for the performance of Li-O 2 batteries [64]. For instance, Sun et al. reported the GE catalyst used in PC-based electrolyte and achieved good performance for at least 5 cycles with lower over-potential in both discharge and charge process [67] as shown in Figure GE composites such as GE/CNT [76] and MnO 2 /GE [77], have also been used and demonstrated to have superior properties. Figure 1-12 Discharge/charge profile (left) and cycle performance (right) of graphene cathode and carbon black cathode Figure 1-12 Discharge/charge profile (left) and cycle performance (right) of graphene cathode and carbon black cathode. Current density is 50 mag -1 [67]. Noble metals such as Pt and Au are employed as catalysts because of their intrinsic conductivity, stability and also excellent catalytic properties in fuel cells [78]. The ability of noble metals to reduce over-potential in Li-O 2 batteries is very obvious. Lu et al. compared the trend in the catalytic activity of noble metals for the oxygen reduction reactions in the Li-O 2 systems [79]. Same research was carried out by Gopi et al. and achieved similar results [80]. To further increase the performance of Li-O 2 batteries, noble metals composites are also used. Metal oxide composites [81, 82] showed excellent catalytic properties due to the bi-catalyst structures while carbon composites [83-85] due to the increase of conductivity. However, noble metals are often very expensive and heavy. The synthesis methods of nanostructured noble metals are usually 21

38 very complex. All the drawbacks limit the application of noble metals in Li-O 2 batteries. Therefore, finding cheap and effective substitutes are still a challenge for Li-O 2 batteries. Transition metal oxides are demonstrated to have good catalytic properties towards oxygen reduction reaction in fuel cells [78]. They are famous for their activity, availability, low cost, thermodynamic stability, low electrical resistance and environmental friendliness. Due to these superiorities, transition metal oxides are widely employed in Li-O 2 batteries. For instance, Debart et al. investigated the catalytic activity of Fe 2 O 3, Fe 3 O 4, CuO, CoFe 2 O 4, and Co 3 O 4 in O 2 cathodes and found Co 3 O 4 gave the best compromise between capacity and capacity retention [86]. Similar results about the cobalt oxides are represented by other groups [87, 88]. Sun et al. also used composites consisting of CoO and CMK-3 and achieved excellent cycling performance for at least 15 cycles [54]. V 2 O 5 is also a good candidate for Li-O 2 batteries and is reported several time used as oxygen reduction reaction catalyst [89, 90]. Besides employing single metal oxides, multi-metal oxide composites are also favourable for Li- O 2 batteries. Zhao et al. reported hierarchical mesoporous perovskite La 0.5 Sr 0.5 CoO 2.91 nanowires for using as catalysts in Li-O 2 batteries and achieved ultrahigh capacity up to more than mah g -1 [91]. Similar compound was shown by other groups [92]. Various catalysts have been come up with and provided more opportunities to find suitable catalysts for Li-O 2 batteries. Giordani et al. developed a method using the H 2 O 2 decomposition reaction as a tool to find better catalysts for oxygen reduction reaction [93]. They have investigated most of transition metal oxides using this method. The results are shown in Figure Due to the similar structure of H 2 O 2 and Li 2 O 2, this tool may be a reliable, useful, and fast screening tool for materials that promote the 22

39 charging process of the Li-O 2 batteries and may ultimately give insight into the charging mechanisms. Figure 1-13 First galvanostatic charge of Li 2 O 2 oxidation for various Li O 2 cells Figure 1-13 First galvanostatic charge of Li 2 O 2 oxidation for various Li O 2 cells [93]. Among all transition metals, manganese oxides are considered the best choice for employing as catalysts due to their easy preparation, excellent catalytic properties and variety. Numerous researches focusing on manganese oxides carried on these days [94]. Debart et al. examined different phases of MnO 2 used as catalysts and found α-mno 2 had the best catalytic property to catalyse the oxygen reduction reaction and support higher capacity retention than MnO 2 in other forms [95]. This indicates the crystal structure of MnO 2 plays very important role in affecting the catalytic performance. They also demonstrated nanostructured MnO 2 such as nanowires showed even better ability to form and decompose Li 2 O 2 and to support higher capacity at the same time than the amorphous ones. The properties of promoting of Li 2 O 2 formation were confirmed by Trahey et al. and they even believed the employment of MnO 2 in carbonate electrolytes 23

40 could help suppress the decomposition of electrolyte [96]. Result is shown in Figure Due to superior properties of MnO 2, all kinds of morphologies were made such as nanorods [97], hollow clews [98], and unique card-house-like structures [99]. Besides morphologies, composites such as GE/MnO 2 [77], Au-Pd/MnO 2 [81], and MnO 2 /C [100], were also researched. Doping methods are also good ways to improve the performance of MnO 2. Benbow et al. found the addition of Ni 2+ could enhance the catalytic properties of MnO 2 [101]. Lee et al. showed the similar result with the doping of Na + [102]. Figure 1-14 Schematic mechanism of Li 2 O 2 and Li 2 O forming on MnO 2 catalyst Figure 1-14 Schematic mechanism of Li 2 O 2 and Li 2 O forming on MnO 2 catalyst [96]. Both metals and metal oxides have excellent catalytic properties towards oxygen reduction reaction. Cheng et al. compared the performance of Li-O 2 batteries with metals and metal oxides as catalysts [103]. They found catalysts in their metal forms were more catalytic in first few cycles with high capacity and low discharge overpotential. However, after few cycles, the catalysts in oxide forms showed better performance than the metal ones. This is because the stability of oxides is superior to that of metals. It makes metal oxides more favourable in Li-O 2 batteries. As it is so hard to choose between metals and oxides, Thapa et al. mixed Pd with oxides together and 24

41 the mixture showed significant lower charge over-potential and good cycling performance [82, ]. Although using catalysts in Li-O 2 batteries can support high capacity and low overpotential, the mechanisms of catalytic reaction are still not fully understood. Some researchers even believed catalysts may be unnecessary or detrimental in non-aqueous electrolytes. The addition of catalysts in Li-O 2 batteries may lead to side reactions such as binder decomposition [73]. Moreover, some believed the beneficial effects of catalyst in charge process are not attributed to the catalysts but to the electrolyte decomposition [107]. These issues will certainly be paid great attention in future s research. In summary, Li-O 2 batteries have been studied over the last few years. Numerous progresses have been made to improve the performance of Li-O 2 batteries. However, the performance of current Li-O 2 batteries is still far from satisfactory and they are not ready for industrial application. Therefore a great breakthrough is still urgently needed. 1.2 Polymer electrolyte Considering the typical properties, electrolytes used in Li-O 2 batteries must be stable enough against oxygen reduction reactions and able to provide sufficient Li + for ionic conductivity as well as battery reactions. Polymer electrolytes, known for their stability, have great potential application for Li-O 2 batteries. Polymer electrolytes are considered to be divided into two classes [108]. Those based on polymers which serve as both solvents to dissolve salts and mechanical support are known as solid polymer electrolytes [109, 110]. Those based on gel polymer gels in which polymer matrix are encaged in liquid electrolyte solutions are called gel polymer electrolytes [111, 112]. Polymer electrolytes act as separators preventing cathodes and 25

42 anodes from contact with each other and also media to transport ions involved in discharging and charging process. Compared with liquid electrolytes, polymer electrolytes exhibit the superior mechanical, thermal, and electrochemical stability. Due to these superiorities, polymer electrolytes have been widely used in lithium batteries Solid polymer electrolyte Since the first discovery that ether-based polymer polyethylene oxide was able to dissolve inorganic salts and exhibit certain ionic conductivity was published in 1973, polymer electrolytes have been widely applied in batteries [113]. When compared with liquid ones, polymer electrolytes can offer excellent processability and flexibility that can help adjust various geometric shapes of batteries. Moreover, the stability of polymer electrolytes can not only ensure the safety of batteries when cycling, but also help eliminate the use of separator. The lack of organic liquid in solid polymer electrolytes provides the possibility to prevent the growth of lithium dendrites [108]. Due to all these advantages, solid polymer electrolytes show the possibility to be employed as electrolytes in all kinds of batteries. Since lithium batteries are widely researched during recent years, employing polymer electrolytes in lithium batteries become more and more heated. Ionic conductivity of solid polymer electrolytes is considered critical when applied to practical use. The ionic conductivity of solid polymer electrolytes can be roughly determined by the effective number of mobile ions, the elementary electric charge, and the ion mobility [114]. The mobile ions are usually known as free ions that are responsible for ionic conductivity. Therefore, a high degree of dissociation of the lithium salts in polymer matrix is needed in order to obtain high conductivity. Besides the dissociation ratio, the Li + transference number is also critical since large amount of 26

43 Li + are required in battery reactions [110, 115, 116]. The ionic motion of Li + in polymer matrix is closely associated with local segmental motion of polymer chains. It is believed the interaction between Li + and atoms such as oxygen and fluorine is the driving force. For instance, molecular dynamics simulations suggest that in poly(ethylene oxide) (PEO) polymer electrolytes, the best ratio between Li + and ether oxygen of PEO chain are approximately 1 to 5. It is believed Li + moves from PEO chain to chain through complexation between each other [116]. The schematic mechanism is shown in Figure Figure 1-15 Schematic mechanism of Li + movement through PEO based polymer electrolyte Figure 1-15 Schematic mechanism of Li + movement through PEO based polymer electrolyte [114]. The segmental motion of polymer chain is often characterized by the glass transition temperature (T g ) of polymer matrix, which is also responsible for the mechanical properties. Various polymers with low T g have been investigated. For instance, polypropylene oxide hosts are known for their amorphousness even at room temperature [108]. Polymers such as PEO, PPO with low T g have the conductivities that are comparable with some of the liquid solutions [ ]. However, those polymer electrolytes have some drawbacks such as complex chemistry and lack of mechanical stability. In order to solve this problem, comb-type polyethers with ether linkages attached as side-chains to the stiff backbone are employed to make polymer electrolytes 27

44 [120, 121]. This design makes it possible for polymer electrolytes to have more flexible segmental motion. There are many polymer matrices that have been employed to make solid polymer electrolytes. PEO-based polymer electrolytes are first discovered in 1973 [113]. However the ionic conductivity was far from satisfactory. The poly(acrylonitrile) (PAN) based electrolytes were investigated because of the acceptable ionic conductivity and wide electrochemical stability window of PAN. However, PAN based electrolytes are found not suitable for directly using in lithium batteries due to the severe passivation upon contact with lithium metal anodes [122, 123]. Other matrices such as poly(methyl methacrylate) (PMMA) and poly(vinylidene fluoride) (PVDF) are demonstrated as potential hosts for lithium batteries. Although so many possible polymer matrices can be used to improve the properties of polymer electrolytes, the conductivity of prepared polymer electrolytes is still big issue for practical use. Operating at high temperature is believed to be a solution, though it is not practical when employed in lithium batteries. The addition of ceramic fillers is also considered as an efficient solution. It has been demonstrated that the ceramic fillers can significantly improve the electrochemical properties including conductivity [ ]. The most commonly used ceramic fillers are Al 2 O 3, SiO 2, MgO, LiAl 2 O 3 and TiO 2.They believed the reason of such improvement is the acid-base type interactions involving oxygen atoms or fluorine atoms, filler acid or base centres and alkali metal cations [127]. Another hypothesis is the addition of ceramic fillers can significantly reduce the possible side reactions between lithium metals and polymer chains as shown in Figure 1-16 [128]. The particle size of fillers also plays an important role. 28

45 Figure 1-16 Schematic mechanism of the addition of ceramic fillers and the effect of different particle sizes, (a) macro-size and (b) nano-size Figure 1-16 Schematic mechanism of the addition of ceramic fillers and the effect of different particle sizes, (a) macro-size and (b) nano-size [128]. Although solid polymer electrolytes have so many superiorities, Anderman questioned the application of solid polymer electrolytes in a review article published in 1994 [129]. He believed that lithium batteries with solid polymer electrolytes didn t share such practical advantages considering the flexibility, thickness, manufacture and mechanical strength. Therefore, the application of solid polymer electrolytes in lithium batteries is still limited Gel polymer electrolyte Since the ionic conductivity of solid polymer electrolytes is the main issue which limits their application, liquid solutions are added as plasticizer in the system to form gel-like 29

46 structures which is known as gel polymer electrolytes. Compared with solid polymer electrolytes, gel polymer electrolytes have been commercialized in many lithium battery industries. Gel polymer electrolytes have been widely studied due to their superior properties including high ionic conductivity, electrochemical stability, safety and tolerance against mechanical and electric abuse [108]. It is believed that the use of gel polymer electrolytes in lithium batteries can effectively suppress the growth of lithium dendrite which is the main issue in lithium batteries since it can lower the cycling efficiency and cause internal short-circuiting. Polymer electrolytes can also have excellent ability to endure the volume change of electrodes during cycling which will further improve the flexibility of designed cells. Another advantage of gel polymer electrolytes used in lithium batteries is their ability to reduce the reactivity of liquid electrolytes towards battery reactions and lithium anodes. This makes gel polymer electrolytes more suitable than liquid electrolytes due to the safety issues. Besides, the manufacturing integrity is also an advantage. Similar to solid polymer electrolytes, PEO can also be used in gel polymer electrolytes. The addition of liquid plasticizer into the systems results in reducing of the crystalline content of PEO, and the increasing of polymer segmental mobility which lead to the increase of ionic conductivity. Generally speaking, polyethers with low molecular weight and polar organic solvents are usually used as plasticizers [ ]. Since the percentage of polymer is low in the whole electrolyte system, the function of polymer matrix is no longer a solvent for lithium salts. The polymer matrices swelling by the liquid solvents only act to provide dimensional stability [111]. This means the interactions between lithium ions and oxygen or fluorine atoms are no longer necessities to form electrolytes and the choices of polymer matrix become various. PAN based electrolytes have been used in the lithium batteries due to their low crystalline content. 30

47 Abraham and Jiang also used a PAN based electrolyte in rechargeable Li-O 2 battery and last for 3 cycles which is also known as the first publication of using organic electrolyte in Li-O 2 batteries [19]. The use of PMMA in the polymer electrolyte was first published in 1985 [133]. The conductivity of PMMA based electrolyte reached 10-3 S cm -1 at room temperature. Since then, PMMA based electrolytes have been widely studied [ ]. However, the poor mechanical strength of PMMA after swelling by liquid plasticizers limits its applications. Additives, such as poly(vinyl chloride) (PVC), have been used to improve the mechanical properties of PMMA [ ]. Other polymer matrices such as PVDF, PVC are also used in polymer electrolytes [ ]. Furthermore, poly(vinylidene fluoride-hex fluoropropylene) (PVDF-HFP) draws great attention because the amorphous phase of HFP can help encage large amount of liquid electrolytes and the crystalline phase of PVDF can provide mechanical support as polymer matrix. The research of PVDF-HFP based electrolytes has also been widely carried on in Li-ion and Li-O 2 batteries [ ]. However, most of the polymer hosts lose their mechanical properties after swelling by liquid plasticizers. Moreover the gain of conductivity is often accompanied by the loss of mechanical strength, the decrease of compatibility with lithium metals and the reducing of safety. In order to solve these problems, various ceramic fillers such as zeolites, ionites, and some neutral fillers are added into the systems [150]. They are known as composite polymer electrolytes [128]. It is believed that the addition of ceramic fillers can significantly improve the conductivity of polymer hosts and also their interfacial properties in contact with lithium metals by suppressing the degree of crystallinity of polymer matrices. 31

48 Ionic liquids (ILs) have also been used as plasticizers for gel polymer electrolytes. ILs are liquids comprised entirely of ions at room temperature. Their unique properties such as low vapour pressure, high ionic conductivity, and good thermal and electrochemical properties make them potential candidates as plasticizers in polymer electrolytes used in Li-ion batteries [151, 152]. The function of ILs towards lithium deposition has also been discussed which is shown in Figure 1-17 [152]. Moreover, polymer electrolytes based on ILs have also been used in Li-O 2 batteries [153, 154]. Although polymer electrolytes exhibit so many superiorities, the use of polymer electrolytes in Li-O 2 batteries is still far from satisfactory. Most batteries with polymer electrolytes can only last for 3 cycles or less [153, 154]. In order to obtain effective Li- O 2 batteries with extraordinary performance, great efforts are still needed in the future. Figure 1-17 Schematic presentation for functional role of PDMITFSI ionic liquid on lithium deposition, (a) without and (b) with ionic liquid Figure 1-17 Schematic presentation for functional role of PDMITFSI ionic liquid on lithium deposition, (a) without and (b) with ionic liquid [152]. 32

49 1.3 Conducting polymer Since the use of carbon may cause problems during cycling in Li-O 2 batteries, it is eager to find substitutes for carbon cathodes. Many solutions have been brought up, one of which is use conducting polymers to replace carbon. Conducting polymers have been widely used in all kinds of areas and it is possible to be used in Li-O 2 batteries. Conducting polymers, or known as intrinsically conducting polymers such as polypyrrole (PPy), polythiophene (PT), polyaniline (PANI), and poly(3,4- ethylenedioxythiophene) (PEDOT) have drawn great attention since the Noble Prize Award in The structures of most commonly used conducting polymers are shown in Figure The conducting polymers show superior properties and applications than other organic polymers because of the processability and conductivity. It is known that conducting polymers can possess very high electrical conductivity in their doped states. Figure 1-18 The structures of the most commonly used conducting polymers Figure 1-18 The structures of the most commonly used conducting polymers. 33

50 Most of the conducting polymers share the similar unique structures which differ from other organic compounds. The carbon atoms and sometimes other heteroatoms such as sulphur and nitrogen in the backbones are usually sp 2 hybridized. This makes all the atoms in the backbones on the same plane. All the atoms have p z orbitals which are orthogonal to the σ-bonds formed by sp 2 hybridised orbitals. The conjugated structure can be formed by all the p z orbitals when they are standing side by side obtaining a delocalized π-bond [155]. For instance, the polyacetylene possesses a long chain-like conjugated structure with all the p z orbitals standing side by side as shown in Figure Technically, the delocalized π-bonds provide the pathways for electrons to move from side to side. However, it is proved that the undoped conducting polymers are usually insulators or semiconductors, which is far from expectation. This is because the energy gaps of undoped conducting polymers are too high to let p z electrons move freely. Due to this reason, doped conducting polymers obtained either by oxidized or reduced from undoped ones are synthesized. By introducing electrons into (reducing) or removing electrons from (oxidizing) the conjugated structures of conducting polymer backbones, the energy gaps are greatly reduced and the delocalized electrons can be very easy to move along the chains form side to side. In this way, the polymers obtain high conductivity. Figure 1-19 Conjugated orbitals formed in polyacetylene Figure 1-19 Conjugated orbitals formed in polyacetylene. 34

51 There are two types of doping, p-doping and n-doping, referred to as oxidizing and reducing states, respectively [156]. P-doping means the polymers are the electron donors and electrons are taken away from the structure backbones. In order to keep electroneutrality, same amount of anions have to be introduced as counterions. Contrarily, n-doping means the polymers are the electron accepters with same amount of cations as counterions. The type of doping usually depends on the gaps, or known as the position of the HOMO and LUMO levels (HOMO corresponds to the upper edge of the valence band and LUMO to the lower edge of the conduction band) [157]. Some polymers such as polyacetylene can be n-doped and p-doped. However, most conducting polymers such as PPy have very high position of their HOMO and LUMO levels, which means it would be much easier to obtain p-doped ones and n-doping would be quite difficult. Moreover, the oxidized conducting polymers possess higher stability than the reduced ones [158]. Besides, the chemical synthesis methods of most conducting polymers involve oxidation of polymer monomers. Therefore, it is easier to achieve the oxidized p-doped forms than the reduced n-doped forms of conducting polymers. The doped conducting polymers have various properties such as stable, environmentally friendly and conductive. The most unique property of conducting polymers that distinguishing from other materials is the doping-dedoping redox performance. In certain environment, the doped conducting polymers can be reduced into their undoped form which can be oxidized quickly at the same time [159, 160]. The redox reaction is shown below where CP stands for conducting polymers and A for anions, (1-6) 35

52 Due to these superiorities of conducting polymers, they have been considered to have great potential applications in various areas such as solar cells [161], fuel cells [162], batteries [163], corrosion protection coatings [164] and supercapacitors [165]. Among all the conducting polymers that have been investigated, PPy, PANI, PEDOT are most commonly used in most areas Synthesis method Conducting polymers are usually synthesized from monomers such as pyrrole (Py), aniline (ANI), and 3, 4-ethylenedioxythiophene (EDOT). The synthesis processes can be through either electrochemical or chemical methods. However, the synthesis mechanism is still not fully understood. Lots of mechanisms have been proposed to explain the process [ ]. The most acceptable mechanism is that the polymer monomer is activated by removing an electron from the molecule and forms a radical monomer structure [166]. When two radical monomers encounter with each other, they quickly combine with each other and form a conjugated long chain. Then the chain can be easily oxidized to its doped phase with anions balancing the charge. In this way, the doped conducting polymer is obtained. The synthesis mechanism of PPy as an example is shown in Figure It is known that electrochemical synthesis produces thin films on an electrode surface while chemical oxidation provides grained materials. 36

53 Figure 1-20 Schematic illustration of synthesis mechanism of PPy Figure 1-20 Schematic illustration of synthesis mechanism of PPy [166]. Electrochemical Synthesis methods are considered quick and easy to control. Proposed conducting polymer films are achieved through electropolymerization, also known as electrodeposition of monomers on suitable substrates and working electrodes. The electropolymerization methods don t provide undoped polymers but oxidized conducting forms. The final polymer chain carries a positive charge every several monomer units and is counter balanced by same amount of anion. In order to achieve ideal polymer films, various electrochemical techniques are employed in electropolymerization. The most commonly used methods are cyclic voltammetry, potentiodynamic, galvanostatic, potentiostatic, and reversal potential pulsing technique. For instance, Wang et al. obtained highly flexible and bendable free-standing PF 6 doped PPy films by electrochemical polymerization method [170]. They also found out the morphology of PPy changed as the deposition time increased while the conductivity decreased with the increase of thickness of polymer film. Another example is the work Sultana et al. carried out [171]. They obtained similar soft, light-weighted, mechanically 37

54 robust and highly conductive free-standing para(toluene sulfonic acid) (ptsa) doped PPy film by electropolymerization methods. They discovered the porosity of PPy films increases while prolonging deposition time. Dubal et al. employed potentiodynamic methods to synthesize nanostructured PPy, such as nanobelts, nanobricks and nanosheets on stainless steel substrates by simply changing the scanning rate during deposition [165]. PEDOT with different doping anions have also been synthesized by Spanninga et al. through electrochemical deposition [ ]. They further discussed what ions are more favourable for counterions when the deposition was carried in solutions with different ions. However, electrochemical synthesis methods usually need equipment to gain conducting polymers, which greatly depends on many parameters such as deposition time, current density and scanning rates. This leads to the complex synthesis processes. Compared with electrochemical synthesis methods, chemical synthesis methods are often known as simpler ways to obtain proposed polymers. In a typical chemical synthesis process, conducting polymers are obtained through chemical oxidation by oxidants, such as FeCl 3, (NH 4 ) 2 S 2 O 8, H 2 O 2, KIO 3, and K 2 GrO 7 [ ]. Usually the polymerization is going in aqueous solutions with acidic additions. The anions from acids can be employed as counterions for oxidized polymers. It is believed that the participation of acids in the process of polymerization can significantly improve the conductivity and other properties of conducting polymers [178, 179]. The choice of acid can be various, such as HCl, H 2 SO 4 and ptsa. All of them provide the counterions as well as ph the polymerizations need. However, we can only obtain amorphous conducting polymers by using in situ chemical methods. The amorphous conducting polymers are usually not conductive enough. Thus 38

55 the application of conducting polymers may be limited. In order to achieve conducting polymers with higher conductivity and electrochemical properties, nanostructured materials are made for use. It is well known that the nanostructured materials have superior properties than bulk and amorphous ones due to the unique structures. It is believed the defects and edge effects of these structures can provide possibility for better performance in various applications. One-dimensional conducting polymers show unusual physical and chemical behaviour due to the nanosize effects [180, 181]. There are various ways to synthesize nanostructured conducting polymers, such as electrospining [ ], hard template synthesis [ ], soft template synthesis [ ] and a variety of lithography techniques [ ]. With rational synthesis design, nanostructured conducting polymers with different diameters, thickness, and length can be obtained under control. Jang et al. used reverse microemulsion polymerization by adding a surfactant, sodium bis(2-ethylhexyl) sulfosuccinate (AOT), into a apolar solution hexane and prepared very regular PPy with nanotube morphology [199]. The whole process is shown in Figure PEDOT nanotube has also been synthesized using the same soft template method. Zhang et al. employed this method with diluted concentration of AOT and EDOT monomer to synthesize PEDOT nanotubes [200]. Huang et al. presented a mechanistic study about the formation of PANI nanofiber [175]. They believed the aniline had intrinsic potential to form nanofiber- or nanowire-like structures. In order to obtain regular structure and prevent secondary growth of PANI, they used fast-mixing method to achieve PANI nanofiber which is also shown in Figure Interfacial method was also employed to synthesis PANI nanofiber by the same group [201]. Qi et al. also exhibited a freezing interfacial polymerization method to prepare highly conductive free standing PPy films [202]. 39

56 However, most of the methods need strict conditions and they are difficult to control the parameters of conducting polymers. Therefore, a more efficient, simple and controllable synthesis method is still urgently needed. (a) Figure 1-21 Schematic illustration of synthesis mechanism of (A) PPy nanotube and (B) PANI nanowire (b) Figure 1-21 Schematic illustration of synthesis mechanism of (a) PPy nanotube [199] and (b) PANI nanowire [175] Application Due to all the superior properties mentioned above, conducting polymers have been widely used in most areas such as solar cells, fuel cells, supercapacitors and rechargeable batteries. Since rechargeable lithium batteries draw great attentions these days, here we would like to focus mainly on the rechargeable lithium batteries. 40

57 Rechargeable lithium batteries are extensively studied due to their extremely high theoretical capacity, light weight, low toxicity and relatively high safety [181]. Typical lithium batteries consist of anodes, cathodes, electrolytes and separators. During discharge process, the lithium ions are released from anodes and inserted into or consumed on the cathodes and it reverses in the charge process. Usually, the anode and cathode materials are made of carbon or metal oxides which can perform the lithium insertion and extraction. Conducting polymers, known for their unique π-conjugated structures and electrochemical properties, can also be used for rechargeable lithium batteries. There are many examples of using conducting polymers directly as cathode materials in lithium batteries because of their intrinsic nature of doping-dedoping redox reactions. For instance, Osaka et al. have proposed a Li-PPy battery with PEO based polymer electrolyte and achieved at least 1500 cycles [203]. Zhou et al used Fe(CN) 4-6 doped PPy as cathode material in Li-ion battery and achieved greatly enhanced capacity [204]. The result is shown in Figure Other groups also obtained similar results when employed PPy as cathode materials in lithium batteries [170, 171]. 41

58 Figure 1-22 Cycling performance of PPy/FC at (a) constant current density of 50 mag -1 and (b) different current densities Figure 1-22 Cycling performance of PPy/FC at (a) constant current density of 50 mag -1 and (b) at different current densities [204]. Many research groups also used conducting polymers as coating materials on anodes and cathodes due to the high conductivity and chemical stability [205, 206]. It is obvious the resistance of batteries has been significantly reduced after the introducing of conducting polymers into the whole systems. For instance, LiFePO 4 has been widely studied in Li-ion batteries. However, the semi conductivity of LiFePO 4 has limited its performance. In order to improve the conductivity of LiFePO 4 electrodes, PEDOT was used and coated on the surface of LiFePO 4 and the charge transference resistances have been significantly reduced [207]. The performance of these batteries was considered very good as shown in Figure Another strategy to use conducting polymers in 42

59 lithium batteries is through preparing composites. Conducting polymers play as conductive media in the composites for electron movement. Composites such as LiFePO 4 /PPy [208], LiN 1/3 Co 1/3 Mn 1/3 O 2 /PPy [209], PPy/Fe/O [210], Sn/PPy [211], PEDOT/V 2 O 5 [212, 213], PEDOT/PDBM [214] and PPy/graphene [215] have been prepared and showed improved performance. Conducting polymers can also be used as media to obtain carbon coating or carbon composites. Wang et al. proposed a method to synthesize carbon composite using PANI as media [216]. The mechanism has been discussed and shown in Figure Figure 1-23 Discharge/charge profiles (left) and resistance (right) of the LiFePO 4 cathode (a) coated with PEDOT, (b) coated with PPy, (c) coated with C, and (d) pristine particles Figure 1-23 Discharge/charge profiles (left) and resistance (right) of the LiFePO 4 cathode (a) coated with PEDOT, (b) coated with PPy, (c) coated with carbon, and (d) pristine particles [207]. 43

60 Figure 1-24 a) Electron-transfer pathway for LiFePO 4 particles partially coated with carbon. b) Designed ideal structure for LiFePO 4 particles with typical nano-size and a complete carbon coating. c) Preparation process for the LiFePO 4 /carbon composite including an in situ polymerization reaction and two typical restriction processes Figure 1-24 a) Electron-transfer pathway for LiFePO 4 particles partially coated with carbon. b) Designed ideal structure for LiFePO 4 particles with typical nano-size and a complete carbon coating. c) Preparation process for the LiFePO 4 /carbon composite including an in situ polymerization reaction and two typical restriction processes [216]. Conducting polymers can also be used in Li-S batteries as coating materials on the surface of sulphur to reduce impedance of whole electrodes. Liang et al. showed the improvement of the cycling stability by introducing nanostructured ordered PPy/S 44

61 cathode in Li-S battery [217]. The same group then further investigated PPy/CNT/S in Li-S battery and achieved even better performance [218]. There have also been some reports about conducting polymers used in Li-O 2 batteries. Cui et al. employed a tubular PPy baed air electrode in Li-O 2 battery as the substitute of carbon based electrodes, and showed better cycleability and higher capacity due to the tubular structure and hydrophilic properties of PPy [219]. PANI has also been employed as waterproof barrier in Li-O 2 batteries [220] as well as media to make nitrogen-doped graphene [76]. Nasybulin et al. investigated PEDOT as catalyst in Li-O 2 batteries and found the charge over-potential was greatly reduced [221]. These results are shown in Figure (a) (b) (c) Figure 1-25 (A) Morphology and cycle performance of PPy cathode [70], (B) nitrogen-doped graphene derived from PANI and (C) Performance of PEDOT catalyst Figure 1-25 (a) Morphology and cycle performance of PPy cathode [70], (b) nitrogendoped graphene derived from PANI [76] and (c) Performance of PEDOT catalyst [221]. 45

62 However, despite the few reports listed above about the use of conducting polymers in Li-O 2 batteries, there have been seldom reports on this area. Due to the very unique properties of conducting polymers, they should have the potential applications in Li-O 2 batteries. Great efforts should be made to obtain highly efficient Li-O 2 batteries with conducting polymers in future research. 1.4 Summary Great efforts have been made in order to obtain Li-O 2 batteries with high specific capacity and reasonable cycle performance. According to the literature above, the electrolytes and cathode materials are the most critical parts in Li-O 2 batteries because the decomposition of electrolytes and low efficiency of catalysts greatly limit the performance. The stability of polymer electrolytes and unique redox properties of conducting polymers provides the possibility to be applied in high-performance Li-O 2 batteries. Polyethylene glycol (PEG) and PVDF-HFP were chosen as the matrices of polymer electrolytes and had been investigated in this work. Gel polymer electrolytes based on these two matrices with different additions were studied in Chapters 3 and 4. Furthermore, PPy was also chosen as the catalyst material in Li-O 2 batteries. Batteries based on PPy and PPy/CNT were investigated to have high catalytic performance which is shown in Chapters 5 and 6, respectively. 46

63 Chapter 2 Experimental Methods 2.1 Overview In order to obtain the usable results of proposed hypothesis, experiments should be conducted in the laboratory. The main process of each experiment can be divided into four parts as listed below (Figure 2-1), The preparation of proposed materials The characterization of as-prepared materials The electrochemical analysis of as-prepared materials The characterization of products and materials During most of the procedures, electric equipment such as SEM, XRD, FT-IR and CHI will be used for characterizing and analysing which will be introduced later in the following parts of this chapter. In situ reaction FT-IR Preparation Solution casting XRD Characterization CV SEM Performance Figure 2-1 Schematic illustration i of the whole experiment ent process EIS Characterization Discharge/charge Figure 2-1 Schematic illustration of the whole experiment process. 47

64 2.2 Materials and chemicals A list of materials and chemicals which were used in the research project are shown below in Table 2-1 along with their formula, purity and supplier. Table 2-1 Materials and chemicals used in the research project Materials and chemicals Formula Purity Supplier Lithium chloride LiCl 99.0% 3,4-Ethylenedioxythiophene (EDOT) C 6 H 6 O 2 S 97% Aldrich Ammonium persulphate (NH 4 ) 2 S 2 O 8 98% Aldrich Acetone CH 3 COCH % Sigma- Aldrich Bis(trifluoromethane)sulfonimide Sigma- CF 3 SO 2 NLiSO 2 CF % lithium salt (LiTFSI) Aldrich Carbon black (Super-P) C 99% Lexel Dimethyl carbonate (CH 3 O) 2 CO 99.0% Sigma- Aldrich Dimethyl sulfoxide (CH 3 ) 2 SO 99.5% Sigma- Aldrich Ethanol C 2 H 5 OH 95% Sigma- Aldrich Hydrochloric acid HCl 32% Sigma- Aldrich Iron chloride FeCl 3 97% Sigma- Aldrich Lithium perchlorate LiClO % Sigma- Aldrich Sigma- Aldrich 48

65 Lithium foil Li % Hohsen Corporation Japan Pyrrole C 4 H 5 N 98% Aldrich Guangzhou Poly(tetrafluoroethylene) Chunting (CF 2 CF 2 ) n 60% Industrial Co., Ltd AJAX Perchloric acid HClO 4 70% Chemicals PTY Limited Propylene carbonate C 4 H 6 O % Sigma- Aldrich Polyethylene glycol (CH 2 CH 2 O) n - Sigma- Aldrich Fluka Poly(vinylidene fluoride-cohexafluoropropylene) (CH 2 CF 2 ) x (CH 2 CF(CF 3 )) y - Sigma- Aldrich Silicon dioxide nanoparticle SiO % Tetraethylene glycol dimethyl ether CH 3 O(CH 2 CH 2 O) 4 CH 3 99% Sigma- Aldrich Sigma- Aldrich Table 2-1 Materials and chemicals used in the research project 2.3 Material preparation The experiments in this thesis are mainly focus on the conducting polymers and polymer electrolytes used in Li-O 2 batteries. The preparation methods of these two materials are in situ oxidation and solution casting method, respectively. 49

66 2.3.1 In situ oxidation In situ oxidation simply represents mixture oxidation reaction when applied in chemical reactions. In situ oxidation is carried on in a container such as a flask with mixed reactants. In this thesis, PPy was synthesized by in situ oxidation. First, certain amount of pyrrole monomers was added into aqueous solution consisting of dopand acids and the solution was kept for stirring for 30 min. At same time, a solution consisting of certain amount of oxidant and dopand acids was prepared and it was added into the previous solution drop by drop. The mixture was kept stirring at room temperature for certain time. The black precipitation was then filtered and washed with distilled water and ethanol for several times and dried under vacuum at a certain temperature for certain time. In this process, the amounts of reactants, temperature and time vary from different materials. The dopands can be HCl and HClO 4 while the oxidants can be FeCl 3 and (NH 4 ) 2 S 2 O 8 depending on the experiments. The whole process is illustrated in Figure 2-2 when using (NH 4 ) 2 S 2 O 8 as oxidant. Figure 2-2 The preparation process of PPy when (NH 4 ) 2 S 2 O 8 was used as oxidant Figure 2-2 The preparation process of PPy when (NH 4 ) 2 S 2 O 8 was used as oxidant Solution casting method Polymer electrolytes in this thesis were prepared by solution casting methods. Solution casting method simply represents casting a solution mixed with polymer matrices and solvents or other components onto a flat glass surface. However, the process differs according to the polymer matrix used. For polymers with low melting points such as 50

67 PEO and PEG, the solution casting method involves no solvents. Polymer matrix was first heated to a certain temperature to obtain a liquid phase. Then, a pre-prepared liquid electrolyte with certain concentration of lithium salt was added into the melting polymer system. The mixture was kept stirring until a homogeneous mixture was obtained. The mixture was then casted onto a flat glass surface and allowed to solidify into a gel-like membrane. This membrane was the as-prepared polymer electrolyte. On the other hand, for polymers with high melting points such as PVDF-HFP, solvents to dissolve the polymer matrix are necessary to obtain electrolyte membrane. The whole preparation process was similar to the one described above except for the addition of solvents and elimination of heating process. The solvent can be acetone when PVDF-HFP was used as polymer matrix. 2.4 Material characterization The following techniques were employed to characterize the status, phase, morphology and structure of the as-prepared materials. In this project, XRD, FT-IR and SEM were mainly used X-ray Diffraction (XRD) X-ray diffraction (XRD) is a non-destructive analytical characterization method that reveals detailed information about the chemical composition and the crystal phase and structure of a wide range of materials. An X-ray diffractometer generates an X-ray beam hitting a sample as a function of incident and scattered angle, polarization, and wavelength or energy. A certain sample has a particular atom arrangement within the unit cell and this will lead to particular relative intensities of the recorded diffraction peaks upon X-ray hitting. Therefore, the unit cell size and geometry may be resolved from the angular positions of the X-ray diffraction results. The resultant diffraction lines 51

68 with obvious peaks together are called an XRD pattern which can provide information on crystal stricture, chemical composition, and physical properties of materials and thin films. Each crystal has a unique XRD pattern according to Bragg s law, (2-1) Where n stands for integer, λ is the wavelength of the incident X-ray beam, d is the distance between atomic layers in a crystal and θ stands for the incident angle. The theory of Bragg s law is shown in Figure 2-3. Figure 2-3 Schematic drawing of Bragg s law Figure 2-3 Schematic drawing of Bragg s law [222]. The basic use of XRD in this research project was to determine the composition and phase of products by comparing the obtained XRD patterns to known standard diffraction lines. The XRD instrument used in this research project was Siemens D5000 with a monochromatized Cu Kα radiation (λ= nm) at a scan rate of 1º min -1 and step size of 0.02º. 52

69 2.4.2 Scanning electron microscope (SEM) The scanning electron microscope (SEM) is a characterization technique that can image a sample by scanning it in a raster scan pattern with the high-energy electron beam. After each scan, the sample producing signals that contain information about the sample s surface topography, composition, and other properties such as electrical conductivity were made up by electrons interacting with the sample atoms. Generally, SEM are used for preliminary analysis. The basic use of SEM in this research project was observing the morphology of asprepared materials and surface of electrodes during cycling. The SEM instrument used in this project was Zeiss Supra 55VP field emission SEM (FE-SEM) with an accelerating voltage of 5-20 kv and mm aperture. The images were taken by an in-lens secondary detector. A thin layer of carbon was used to deposite on the surface of the materials if the conductivity was too low Fourier transform infrared spectroscopy (FT-IR) Fourier transform infrared spectroscopy (FT-IR) is used for obtaining infrared spectra of absorption, emission, photoconductivity or Raman scattering of a solid, liquid or gas. An FTIR spectrometer simultaneously collects spectral data in a wide spectral range. This confers a significant advantage over a dispersive spectrometer which measures intensity over a narrow range of wavelengths at a time. The basic use of FT-IR in this research project was to analyse the structure of asprepared materials and products that were produced during cycling. The FT-IR instrument used in this project was Nicolet Magna 6700FT-IR spectrometer using 4 cm - 1 resolution and 64 scans at room temperature. 53

70 2.4.4 Thermogravimetric Analysis (TGA) Thermogravimetric analysis (TGA), also known as thermal gravimetric analysis (TGA), is a type of measurement usually used to determine the weight changes of a certain sample during the changing of temperature in a controlled atmosphere. Such analysis can detect the weight of a sample as temperature elevates accurately. The results are often concluded in a figure with a continuous line to identify weight loss processes which are related to the chemical reactions occurring. Figure 2-4 shows an example result of a TGA measurement. The TGA measurement can be conducted in atmosphere of air or noble gases for different applications. Figure 2-4 An example TGA result of polypyrrole coated silicon Figure 2-4 An example TGA result of polypyrrole coated silicon [223]. In this project, the TGA was used to determine the weight ratio of each component in as-prepared materials. The TGA instrument used in this project was Simultaneous TG- DTA (SDT 2960) with a platinum plate as the sample holder. The temperature was set to increase to 1000 ºC in air atmosphere with a speed of 5-10 ºC/min. 54

71 2.5 Electrode preparation and cell assembly Electrode preparation The cathode used in Li-O 2 batteries was prepared as follows: the catalyst mash was prepared by mixing the as-prepared materials, conducting matrices and binders together in certain ratio in isopropanol with continuous stirring. The mixture was then pressed onto the stainless steel mesh or a nickel mesh to form the air cathode. The cathode film was punched into discs with a diameter of 14 mm and dried at 80ºC in a vacuum oven for 12 h and then kept in the glove box. The typical loading of the air electrode is about 2 mg cm Cell assembly A Swagelok type cell with an air hole (0.785 cm 2 ) on the cathode side was used to investigate the electrochemical performance. The cell was assembled in Ar filled glove box (Mbraun) with water and oxygen level less than 0.1 ppm. The as-prepared air cathode was used as the working electrode and a lithium foil was used as the counter and reference electrode. The electrodes were separated by a glass microfiber filter (Watman). The electrolytes depended on the experiments operated. The cell was gastight except for the air side window that exposed the porous cathode film to the oxygen atmosphere. The brief diagram of the structure of the Li-O 2 battery is shown in Figure 2-5. All experiments were tested in 1 atm dry oxygen atmosphere to avoid any negative effects of humidity and CO 2. 55

72 Current collector Cathode Separator Lithium foil Figure 2-5 The structure of a Li-O 2 battery Figure 2-5 The structure of a Li-O 2 battery. 2.6 Electrochemical characterization The electrochemical properties of as-prepared materials are obtained by performance characterization techniques on the assembled batteries. These techniques usually include cyclic voltammetry, electrochemical impedance spectroscopy, linear sweep voltammetry, and galvanostatic charge-discharge testing cyclic voltammetry. The characterization of as-prepared materials can be further evaluated from these results Cyclic Voltammetry (CV) Cyclic voltammetry or known as CV is a type of potentiodynamic electrochemical measurement which has been widely used in electrochemical characterization of 56

73 materials in Li-ion and Li-O 2 batteries. It records the relationship of current and voltage when the potential of the working electrode is ramped linearly versus time. This ramping is known as scan rate (V/s). CV can be conducted in two electrode and three electrode systems. In a typical three electrode system, the potential is applied between a reference electrode and a working electrode and the current is measured between a working electrode and a counter electrode while in a typical two electrode system, the potential and current are both measured between a working electrode and counter electrode. In this project, the CV measurements were mainly performed by two electrode systems, where a lithium anode was used as both reference and counter electrode. During the scanning, an analyte can be reduced or oxidized and re-oxidized or re-reduced on the return scan, which is known as a sign of highly reversible redox couples. The peaks on the CV results indicate the oxidation and reduction potentials. The CV measurements in this project were mainly conducted via a CHI 660C or CHI 660D electrochemical workstation (CH Instrument, Cordova, TN) Electrochemical Impedance Spectroscopy (EIS) Electrochemical impedance spectroscopy (EIS) has also been known as Alternating current (AC) Impedance. EIS is often used to characterize the dynamics of an electrochemical process in terms of an electrochemical cell s response to an applied potential at different frequencies. By observing the current response, the resistance within different frequencies can be examined. The resistance can be read from a typical Nyquist curve. Figure 2-6 shows the typical impedance Nyquist curve of a battery system consisting of a compressed semicircle in a medium frequency region which represents the charge-transfer resistance (R ct ) and an inclined line in the low frequency range which represents the Warburg impedance attributed to a diffusion-controlled 57

74 process. The high frequency intercept at the real axis corresponds to the electrolyte bulk resistance and electronic resistance of the current collector [55, 59, 224]. Figure 2-6 A typical ESI Nyquist curve of a battery system Figure 2-6 A typical ESI Nyquist curve of a battery system. ESI data were mainly conducted form a CHI 660C or CHI660D electrochemical workstation (CH Instrument, Cordova, TN) in this project. The frequency range was set between 100 khz and 0.01 Hz at controlled temperature when the amplitude of the AC signal applied was set to 5 mv Linear Sweep Voltammetry (LSV) Linear sweep voltammetry (LSV) is also a commonly used potentiodynamic electrochemical measurement where the current at a working electrode is measured while the potential between the working electrode and a reference electrode is swept linearly in time. The whole scanning process is carried on through constant scanning rate (V/s). During the scanning, materials on the working electrode will undertake redox reactions. Oxidation or reduction of species is registered as a peak or trough in the current signal at the potential at which the species begins to be oxidized or reduced. The 58

75 scanning process can be operated only once or repeatedly. The basic mechanism is similar to CV measurement. Figure 2-7 shows the typical result of a LSV measurement. Figure 2-7 A typical result of LSV measurement Figure 2-7 A typical result of LSV measurement. The LSV measurements in this project were mainly conducted via a CHI 660 C or CHI 660 D electrochemical workstation (CH Instrument, Cordova, TN) Galvanostatic Charge and Discharge Generally, Galvanostatic charge and discharge is used as electrochemical tests in which a constant current density is used to determine the electrochemical performances such as capacity. The capacities during charge or discharge can be calculated through the following equation, (2-1) I represents for current density, t is the charge or discharge time and Q stands for the capacity during charge or discharge process. In a three electrode system, the galvanostatic charge and discharge performance is tested through a chronoamperometry technique on an electrochemistry workstation with an aqueous electrolyte in open 59

76 circumstances. For a two electrode system, a sealed or open battery cell is used for the testing. Usually the galvanostatic charge and discharge tests can exhibit electrochemical information such as capacities, charge/discharge profiles, Columbic efficiency, and cycle properties. Figure 2-8 shows an example discharge and charge curve obtained through galvanostatic charge and discharge tests. Figure 2-8 An example charge and discharge curve of a Li-O 2 battery Figure 2-8 An example charge and discharge curve of a Li-O 2 battery [225]. In this project, the galvanostatic charge and discharge measurements were conducted on a computer-controlled Neware battery testing system or a Land Battery testing system. 60

77 Chapter 3 Low Molecular Weight Polyethylene Glycol Based Gel Polymer Electrolyte Used in Li-O 2 Batteries 3.1 Introduction Polyethylene glycol (PEG) is a polyether compound with many applications from industry manufacture to medicine. It is also known as polyethylene oxide (PEO), depending on the molecular weight. Usually PEO refers to polymer with a molecular weight higher than 20,000 g/mol while PEG prefers oligomers and polymers with a molecular weight lower than 20,000 g/mol. PEG are liquids or solids with low melting point, also depending on the molecular weight. The typical molecular structure of PEG or PEO is shown in Figure 3-1. Figure 3-1 The typical molecular lar structure ure of PEG or PEO Figure 3-1 The typical molecular structure of PEG or PEO. PEO has been widely used in polymer electrolytes, both solid states and gel states due to its semicrystalline properties with low glass transition temperature and low melting point [226]. In the solid polymer electrolyte systems, several PEO n -LiX systems have been studied and the ion conduction mechanism is closely related to the oxygen-assisted hopping process with the long range segmental motion [227]. However, due to the low conductivity of solid polymer electrolytes, gel polymer electrolytes (GPEs) are more 61

78 acceptable for lithium batteries. PEO has also been widely used as polymer matrix in GPEs [ ]. Appetecchi et al. investigated the PEO based GPEs with Polyethylene glycol dimethyl ether (PEGDME) as plasticizer with different ratio [132]. However, most of these reports use PEO with high molecular weight as the polymer matrix. Few focused on the low molecular weight PEG based ones and even less was employed in Li-O 2 batteries. Moreover, the ones that were used in Li-O 2 batteries cannot provide sufficient capacity and cycling performance. In this chapter, GPEs based on low molecular weight PEG were prepared and used in Li-O 2 batteries. Tetraethylene glycol dimethyl ether (TEGDME) was used as plasticizer and solvent for lithium salts. Silica nanoparticle (SiO 2 ) was used as addictive to increase the ionic conductivity. The electrochemical performances of all GPEs have been tested and liquid TEGDME electrolyte was used for comparison. 3.2 Experiment Preparation of PEG based GPEs PEG-based gel polymer electrolyte (GPE) was prepared by hot solution casting method. The liquid TEGDME electrolyte was prepared by dissolving Bis(trifluoromethane)sulfonimide lithium salt (LiTFSI) into tetraethylene glycol dimethyl ether (TEGDME). The concentration was kept at 1 M. The solid PEG was heating up to 70 o C and melting completely. After mixing the melting PEG, liquid TEGDME electrolyte and certain amount of ceramic fillers, SiO 2 nanoparticle, together, the whole solution was mechanically stirred at 70 o C until a homogenous solution was obtained. The weight ratio of PEG, liquid solution and SiO 2 was kept at 20: 70: 5. The homogenous solution was then cast into a glass microfiber filter as the mechanical support at room temperature and allowed to solidify. This process resulted in a white gel 62

79 like membrane and it was ready to characterize. The same process was used to make SiO 2 -free GPE only without the addition of ceramic filler. All the chemicals were dried before use and all the process was operated in a glovebox with water and oxygen content lower than 0.1 ppm Material characterization X-ray diffraction (XRD) was conducted on a Siemens D5000 X-ray diffractionmeter. During the XRD analysis process, all materials and cathodes were protected without exposure to the ambient atmosphere Electrochemical testing All the electrochemical characterization of as-prepared GPEs was carried on electrochemical workstation. The lithium ion plating/stripping was characterised by cyclic voltammetry measurements using Li/GPE/Li cell over a wide voltage range (-4.5 V to 4.5 V) at medium scanning rate (10 mv s -1 ). Linear sweep voltammetry measurements were used to determine the stability of as-prepared GPEs with Li/GPE/stainless steel (SS) sealed cell and Li/GPE/carbon black electrode (CB) cell exposed to ambient oxygen over a wide scanning range (open circuit potential to 6.5 V) at slow scanning rate (1.0 mv s -1 ). The preparation of carbon black electrode will be mentioned later. Liquid TEGDME electrolyte was characterised in the same way for comparison. The carbon electrode slurries were prepared by mixing carbon black (90 wt%) and poly(tetrafluoroethylene) (PTFE) (10 wt%) together in propanol. The mixture was then coated on a nickel mesh substrate. The cathode film was then obtained by punched into disc and dried under vacuum at 80 o C for 12 h. A Swagelok cell with an air hole (0.785 cm 2 ) on the cathode side was used to investigate the discharge and charge performance. 63

80 The Li-air cells were assembled in a glovebox with water and oxygen level less than 0.1 ppm. The as-prepared GPEs were sandwiched between lithium foils and carbon black electrodes and a glass microfiber filter was used for keeping the structure from collapsing, as mentioned before. For comparison, a cell with a glass microfiber filter soaked in liquid TEGDME electrolyte was also made. All the cells were gas-tight except for the cathode side window. All the measurements were conducted in 1 atm in dry oxygen atmosphere to avoid any negative effects of humidity and CO Results and discussion The most important factors that affect the performance of electrolytes used in Li-O 2 batteries are Li + transportation, the stability, and the conductivity. When GPEs are used in lithium batteries, they have to have sufficient properties to support the acceptable performance. It is very essential for GPEs used in Li-O 2 batteries to have a very good reversibility for Li + plating and stripping. Good reversibility can ensure a smoothly running of battery reactions. Figure 3-2 shows the typical cyclic voltammetry curves obtained in a Li/GPE/Li typed cell for both GPEs. There was only single peak found during the cathodic and anodic reaction process, respectively, which was related to the Li + plating and stripping and the curves were symmetrical. The anodic and cathodic peaks were almost identical both for current and potential. All those results indicated the good reversibility of both GPEs and demonstrated the usability of them in the Li-O 2 batteries. From the comparison of Figure 3-2 (a) and (b), it was very clear that the current was higher when GPE with SiO 2 addition was used. This indicated that the addition of SiO 2 did help improve the conductivity as many references claimed [108, 111, 128]. However, after 20 cycles the current was greatly reduced when GPE with SiO 2 addition was used while the one was almost the same when no addition was added. 64

81 This is probably due to the addition of SiO 2 into the system could reduce the stability of the GPEs, according to some references [117, 128] Li/PEG/Li (a) Current (A) Figure 3-2 Cyclic voltammetry results of Li/GPE/Li type cells with (a) PEG and (b) PEG with SiO 2 addition Cycle 20 Cycle Voltage (V) Li/PEG-SiO 2 /Li (b) 0.01 Current (A) 0.00 Cycle Cycle Voltage (V) Figure 3-2 Cyclic voltammetry results of Li/GPE/Li type cells with (a) PEG and (b) PEG with SiO 2 addition. The scanning rate is 10 mv s -1. It is also very critical for electrolyte used in lithium rechargeable batteries to have a wide operating potential window to ensure minimal side reactions. During the charge process in Li-O 2 batteries, the large over-potentials cause many side reactions such as the decomposition of electrolytes. This makes the batteries very difficult to have reasonable cycling performance. In order to investigate the stability of both as-prepared PEG based GPEs, linear sweep voltammetry (LSV) method was used with 65

82 Li/GPE/cathode cells and the results are shown in Figure 3-3. Liquid TEGDME electrolyte was also measured through the same method for comparison. In Figure 3-3 (a), the measurement was carried out in a sealed cell with a stainless steel as cathode. Both the two GPEs exhibited high electrochemical stability up to 5.5 V vs. Li + /Li while the liquid TEGDME electrolyte showed increasing current since 5 V. This indicated electrolytes gelled with PEG polymer exhibited a significant increase of stability in a sealed system and normally a stability window up to 4 V is sufficient for usage in Li-ion batteries. However, because the cathodes and electrolytes are exposed to the ambient oxygen in the Li-O 2 batteries, the stability results of GPEs in sealed cells shown in Figure 3-3 (a) are not convincing enough. In Figure 3-3 (b), the measurement was carried in a cell with an open hole at the cathode side allowing the oxygen flowing inside of cell and a porous carbon black electrode was used as cathode. Both the GPEs were stable up to 4.6 V vs. Li + /Li while the liquid TEGDME electrolyte suffered from decomposition at lower voltage. It was very obvious that exposure to oxygen atmosphere caused the instability of electrolyte systems. This indicates that choosing electrolytes for Li-O 2 batteries is much more difficult than the regular Li-ion batteries and both GPEs showed the potential qualification to be used in Li-O 2 batteries. It was also clearly seen from Figure 3-3 that the addition of Nano-scaled SiO 2 could reduce the stability of the whole electrolyte systems. But the stability of GPEs with the addition of SiO 2 was still sufficient enough for usage as electrolyte in Li-O 2 batteries. 66

83 Li-PEG-SS Li-PEG-SiO 2 -SS Li-TEGDME-SS (a) Current (A) Potential (V) Figure 3-3 Linear sweep voltammetry results of (a) Li/GPE/SS and (b) Li/GPE/CB-air type cells Li-PEG-air Li-PEG-SiO 2 -air Li-TEGDME-air (b) Current (A) Potential (V) Figure 3-3 Linear sweep voltammetry results of (a) Li/GPE/SS and (b) Li/GPE/CB-air type cells. The scanning rate is 1 mv s -1. The ionic conductivity was also considered as an important concern. The conductivities were calculated from the membrane resistances obtained from impedance spectra. The 67

84 typical impedance spectra obtained at different temperature are shown in Figure 3-4 (a). It is clearly seen that the spectra have the shape of slanted straight lines and intercept the real axis on the high-frequency side. According to the literatures, this represents an equivalent circuit in which a resistor is in series with the electrode capacitance and the intercept on the real axis gives the resistance from which the conductivity of the electrolyte is calculated [149]. The ionic conductivity is calculated through Equation 3-1, (3-1) where σ stands for ionic conductivity, R b represents the bulk resistance, d is the thickness of the gelled polymer electrolyte, and S is the area of the electrodes. The calculated result is shown in Figure 3-4 (b). It was seen that the ionic conductivity slightly increased with the temperature. However there was a dramatic increase 55 to 75 ºC. This is probably due to the melting of PEG polymer matrix. The melting point of PEG was estimated at 60 to 70 ºC and the gelling process lowered the temperature furthermore. After the melting of PEG polymer matrix, the ionic conductivity was greatly increased because the gelled structure was destroyed and the number of free ions would greatly increase. After then the conductivity stayed still even the temperature was kept increasing. It also fits the theory that operating at high temperature is a good way to increase conductivity for polymer electrolytes. The conductivity of GPE shown in Figure 3-4 was demonstrated to be enough for Li-O 2 batteries at room temperature. 68

85 3,000 (a) Z" ( ) 2,000 1,000 Figure 3-4 The impedance spectra of PEG at different temperatures. (b) The calculated ionic conductivity of PEG at different temperatures 105 o C 95 o C 85 o C 75 o C 65 o C 55 o C 45 o C 35 o C 25 o C ,000 2,000 3,000 Z' ( ) PEG (b) Conductivity ( cm -1 ) Temperature ( o C) Figure 3-4 (a) The impedance spectra of PEG at different temperatures. (b) The calculated ionic conductivity of PEG at different temperatures. 69

86 The electrochemical performance of as-prepared GPEs was investigated through Li-O 2 batteries. Figure 3-5 (a) shows the discharge and charge profiles of Li-O 2 batteries using the as-prepared GPEs. The specific discharge capacity of battery with PEG based GPE was 3,667 mah g -1. The battery with SiO 2 addition showed much higher capacity of 6,477 mah g -1 while the one with liquid TEGDME electrolyte showed even higher capacity of 7,921 mah g -1. It was clearly seen that liquid electrolytes had better ability to support higher capacity than the gelled ones and the addition of SiO 2 could be a good choice to increase the specific capacity of polymer electrolytes. Figure 3-5 (b) displays the partial enlarged view of the discharge and charge profiles in Figure 3-5 (a) from 0 to 1,500 mah g -1 in the first cycle. The discharge plateau of battery with liquid TEGDME electrolyte was the highest while the one of battery with PEG based GPE without any addition was the lowest. Similar order was shown during the charge process. It was very obvious that although the GPEs could support acceptable electrochemical performance, the addition of SiO 2 into the polymer system could help further reduce the overpotential and increase specific capacity during the discharge and charge process. The possible reason should be related to their different intrinsic properties such as conductivity and phase structure. The TEGDME electrolyte is in its liquid phase which means the ions that dissolved in the solution can move freely through the whole electrolyte system and this provides relatively high ionic conductivity. It is believed the higher ionic conductivity can help support higher capacity and lower over-potential due to the increase of the oxygen reduction reaction kinetics. For both GPEs obtained in this experiment, the gelled structures would hinder the transfer of ions through the systems and this might lead to a lower conductivity. Furthermore, gel structures are in their solid phase which means they don t have the same penetrability as the liquid electrolytes. The contact surface between electrolytes and cathodes will be reduced and this may result in 70

87 the decrease of the active sites which can also cause the reduction of specific capacity. This could explain the differences of behaviour between GPEs and liquid TEGDME electrolyte. For PEG based GPEs, the addition of SiO 2 nanoparticle can help provide higher ionic conductivity to the GPE system which has been demonstrated [117, 119]. This leads to the reduction of over-potential and the improvement of discharge specific capacity (a) Voltage (V) PEG PEG-SiO 2 TEGDME Figure 3-5 (a) First discharge and charge profiles of Li-O 2 batteries with PEG, PEG-SiO 2, TEGDME as electrolytes. (b) Partial enlarged view of first discharge and charge profiles from mahg ,000 4,000 6,000 8,000 Specific capacity (mah g -1 ) (b) 4.0 Voltage (V) 3.5 PEG PEG-SiO 2 TEGDME Specific capacity (mah g -1 ) Figure 3-5 (a) First discharge and charge profiles of Li-O 2 batteries with PEG, PEG- SiO 2, TEGDME as electrolytes. (b) Partial enlarged view of first discharge and charge profiles from 0-1,500 mah g -1. The current density is 50 ma g

88 In order to determine whether the reversibility of batteries with different electrolytes is suitable for Li-O 2 batteries, the batteries were cycled at a fixed capacity of 500 mahg -1 for 25 cycles. Figure 3-6 shows the cycling performances of the batteries with different electrolytes. It is clearly seen that the battery with PEG based GPE without addition could survive at least 25 cycles with no obvious fading. This indicates the sufficient stability of PEG based GPE to support acceptable cycling performance for Li-O 2 batteries. However, the one with SiO 2 addition could only last for 21 cycles and the capacity faded quickly since then. This was probably due to the reduction of stability of electrolyte system when the SiO 2 nanoparticle was added. Therefore, when considering the cycling performance, SiO 2 addition is not suitable for Li-O 2 batteries. For comparison, the battery with liquid TEGDME was also cycled and found to be able to proceed at least 25 cycles. 5.0 (a) 4.5 Voltage (V) st Figure 3-6 Discharge and charge profiles of Li-O 2 5th batteries with (a) PEG, (b) PEG-SiO 2, and (c) TEGDME as electrolytes at fixed capacity to 500 mahg -1 10th 15th 20th 25th Specicif capacity (mah g -1 ) Figure 3-6 Discharge and charge profiles of Li-O 2 batteries with (a) PEG, (b) PEG- SiO 2, and (c) TEGDME as electrolytes at fixed capacity to 500 mahg -1. Current density is 50 mag

89 5.0 (b) 4.5 Voltage (V) st 5th 10th 15th 20th 21st Specific capacity (mah g -1 ) (c) Voltage (V) st 5th 10th 15th 20th 25th Specific capacity (mah g -1 ) Figure 3-6 Discharge and charge profiles of Li-O 2 batteries with (a) PEG, (b) PEG- SiO 2, and (c) TEGDME as electrolytes at fixed capacity to 500 mah g -1. Current density is 50 ma g

90 To further investigate the stability of PEG based GPEs, cycles at fully discharge and charge were performed. The results are displayed in Figure 3-7. The battery with PEG based GPE without any addition as electrolyte could last at least 5 cycles with 86% retention of discharge capacity, while the capacity of battery with liquid TEGDME electrolyte dropped quickly after the first cycle with only 38% retention even it possessed extremely high capacity at the first discharge process. This provides the evidence that the gelled polymer electrolyte has much higher stability than the plasticizer before the addition of polymer matrix. The stability leads to the stable cycling performance in Li-O 2 batteries. Figure 3-7 also shows the fast fading of battery with addition of SiO 2, which also demonstrated that the addition of SiO 2 could reduce the stability of the polymer electrolyte system. Specific capacity (mah g -1 ) 8,000 7,000 6,000 5,000 4,000 3,000 2,000 PEG PEG-SiO 2 TEGDME Figure 3-7 Cycle profiles of Li-O 2 batteries with PEG, PEG-SiO 2, and TEGDME as electrolytes Cycle numbers Figure 3-7 Cycle profiles of Li-O 2 batteries with PEG, PEG-SiO 2, and TEGDME as electrolytes. Current density is 50 ma g -1. The reason of such different performance of GPEs comparing to liquid electrolytes is due to their unique gelled structures. It is believed that the interaction between polymer 74

91 matrix and plasticizer through Li + can help maintain a stable gel structure [149]. In this report, TEGDME was used for plasticizer as well as solvent for lithium salt. The possible structure is displayed in Fighure 3-8 (a). Li + acted as a cross-link between polymer matrix PEG and plasticizer TEGDME. The whole structure was bonded by Lewis acid-base force which indicated that the Li + interacted with oxygen atoms at the polyether structures from polymer matrix and plasticizer. The gelled structure was held up in this way. Moreover, the stability of structures like this could also be improved greatly towards reactants, such as superoxide radicals. The interactions of oxygen atoms in the electrolyte systems with Li + might cause electrons donated from oxygen atoms to Li +. In this way, the stability of α-hydrogen atoms adjacent to oxygen atoms is greatly improved. According to the previous reports, the oxidation of polyether preferred to happen at the α-carbon to oxidize the α-hydrogen atoms [31, 232]. The resistance of the interacted structures would have higher capability to suffer from superoxide radicals. Bryantsev et al. proved that electrolyte with a similar structure consisting of 1,2- Dimethoxyethane (DME) and Li + was much more stable than the DME itself [31]. After the structure was set up, the excess Li + could move along the electrolyte system and provide ionic conductivity and ions for battery reactions. When SiO 2 nanoparticle was added into the electrolyte systems, partial Li + used in the structures would be replaced and the amount of excess Li + would be increased as SiO 2 is a known Lewis acid which is shown in Figure 3-8 (b). In this way, the ionic conductivity can be improved. However, the stability of SiO 2 -added GPE system towards superoxide radicals reduced greatly. The strong Lewis acid property made it possible for radicals to attack the electrolyte systems instead of free Li + in the electrolytes. This might lead to the collapse of the interacted structures in GPE systems, which explains the instability of GPE with the addition of SiO 2 nanoparticle. 75

92 (a) (b) Figure 3-8 Structures of (a) PEGbased electrolyte and (b) PEG-SiO 2 - based electrolyte Figure 3-8 Structures of (a) PEG-based electrolyte and (b) PEG-SiO 2 -based electrolyte. Figure 3-9 shows the XRD patterns of PEG and GPEs, respectively. The peak related to polymer matrix PEG broadened and the intensity decreased. It indicated that the crystallinity of PEG after made into GPEs was greatly reduced by introducing Li + and plasticizer in the structures. This provided the evidence that there might be interactions among these three components which is also consistent with the previous theory. Another possible explanation of such improved cycleability of battery with GPE as electrolyte than the one with liquid TEGDME electrolyte is the protection of lithium anode. As we know, the anodes in Li-O 2 batteries use lithium metals directly and lithium metals are considered very reactive towards oxygen. Ether based electrolytes are believed to have great ability to dissolve and transport oxygen, which may lead to the corrosion of lithium anode. This would cause the fast fading of capacity. The gelled 76

93 structure of GPE could help lower the kinetics of transporting oxygen through the battery system. This could help maintain a longer cycle life for Li-O 2 batteries. before after Intensity (a.u) Figure 3-9 XRD pattern of PEG before and after made into polymer electrolyte Figure 3-9 XRD pattern of PEG before and after made into polymer electrolyte. As showed above, the liquid TEGDME could help deliver higher initial specific capacity while PEG based GPE could provide better cycling performance. Since the main issue for applying rechargeable Li-O 2 batteries in practical application is their cycle life and stability [103]. It is believed that the cycleability of Li-O 2 batteries at a certain current is more important than the initial discharge capacity. Therefore, the use of PEG based GPE instead of liquid TEGDME is much more favourable for Li-O 2 batteries. Due to the same reason, Nano SiO 2 addition is also considered unfavourable for Li-O 2 batteries. To further investigate the usability of PEG based GPE in Li-O 2 batteries, XRD measurement was conducted to detect the discharge products after first discharge process. Figure 3-10 shows the XRD patterns of cathodes after first discharge in 77

94 batteries with PEG based GPE as electrolyte compared with pure Li 2 O 2 and PEG. It is clearly seen the discharge products were dominated by Li 2 O 2. This demonstrated the high stability of PEG based GPE towards oxygen reduction reaction and confirmed the usability in Li-O 2 batteries. Therefore, PEG based GPE is suitable for Li-O 2 batteries. Further investigation of increasing the capacity of such batteries is still going. PEG electrode Li 2 O 2 * PEG Intensity (a.u.) Figure 3-10 XRD pattern of cathode after discharge in PEG polymer electrolyte * * * * * Figure 3-10 XRD pattern of cathode after discharge in PEG polymer electrolyte. 3.4 Summary Gel polymer electrolytes with low molecular weight polyethylene glycol were successfully prepared by hot solution casting method. The Li + plating/stripping reversibility, stability and conductivity were considered acceptable for usage in Li-O 2 batteries. The cycling performance of batteries using PEG based GPE as electrolyte was better than the direct use of liquid tetraethylene glycol dimethyl ether plasticizer. The addition of SiO 2 nanoparticle into the GPE system has also been demonstrated not 78

95 favourable for Li-O 2 batteries because of the decrease of stability. Therefore, we expect that PEG based GPE could be used as promising electrolyte in long-life Li-O 2 batteries. 79

96 Chapter 4 Investigation of PVDF-HFP Based Gel Polymer Electrolyte Used in Li-O 2 Batteries 4.1 Introduction Poly(vinylidene fluoride-co-hexafluoropropene) (PVDF-HFP) is known as a copolymer of vinylidene fluoride with hexafluoropropene. The copolymer process provides low crystallinity and glass transition temperature and helps improve the flexibility. However, the content of HFP in the copolymer systems is very critical, which is basically kept 8% to 25%. PVDF-HFP has been largely employed in many industry applications including battery industry. Figure 4-1 shows the typical structure of PVDF- HFP. Figure 4-1 The typical structure re of PVDF-HFP P Figure 4-1 The typical structure of PVDF-HFP. Poly(vinylidene fluoride) (PVDF) has been widely applied as polymer host in gel polymer electrolytes because of the high stability due to the strong electro-withdrawing function group (-C-F), and high dielectric constant which can help provide a high concentration of charge carrier [111, ]. However, the mechanical properties of PVDF based GPE are often very weak and need cross-linking treatment. PVDF-HFP is 80

97 considered to have very good capability of trapping large amount of liquid electrolytes and maintaining sufficient mechanical integrity at the same time without the need for cross-linking process. This makes PVDF-HFP more favourable as polymer matrix for GPEs. It is known that the PVDF-HFP GPE system can be described as a heterogeneous, phase-separated, plasticized polymer electrolyte and there are four phases in the system. They are semi-crystalline polymer with low degree of crystallinity, amorphous part plasticized with the electrolyte solution, large volume of nanopores, and interfacial regions of the nanoparticle filler filled/coated by the liquid solution of electrolyte [111]. This system has already been employed in the battery industry [236]. In this chapter, GPE based on PVDF-HFP with Tetratheylene glycol dimethyl ether (TEGDME) as plasticizer and solvents for lithium salts was prepared and used in Li-O 2 batteries. The liquid TEGDME electrolyte was used for comparison. Other plasticizer such as propylene carbonate (PC) and dimethyl sulfoxide (DMSO) were also used since they were most commonly used liquid electrolytes in Li-O 2 batteries. 4.2 Experiment Preparation of PVDF-HFP based GPEs PVDF-HFP based polymer electrolyte (GPE) was prepared by solvent casting method. The liquid TEGDME electrolyte was prepared by dissolving Bis(trifluoromethane)sulfonimide lithium salt (LiTFSI) into tetraethylene glycol dimethyl ether (TEGDME). The concentration was kept at 1 M. The solid PVDF-HFP was dissolved in certain amount of acetone and kept stirring for 2 h. After all PVDF- HFP was dissolved, liquid TEGDME electrolyte was added into the solution and the solution was kept mechanically stirring for 2 h until a homogenous solution was 81

98 obtained. The weight ratio of PVDF-HFP and liquid solution was kept at 2: 7. The homogenous solution was then cast onto a flat glass at room temperature and allowed to solidify. This process resulted in a transparent gel like membrane and it was ready to characterize. The same process was used to make GPEs containing propylene carbonate (PC) and dimethyl sulfoxide (DMSO) as plasticizers. All the chemicals were dried before use and all the process was operated in a glovebox with water and oxygen content lower than 0.1 ppm Material characterization X-ray diffraction (XRD) was conducted on a Siemens D5000 X-ray diffractionmeter. During the XRD analysis process, all materials and cathodes were protected without exposure to the ambient atmosphere Electrochemical testing All the electrochemical characterization of as-prepared GPE was carried on electrochemical workstation. The lithium ion plating/stripping was characterised by cyclic voltammetry measurements using Li/GPE/Li cell over a wide voltage range (-4.5 V to 4.5 V) at medium scanning rate (10 mv s -1 ). Linear sweep voltammetry measurements were used to determine the stability of as-prepared GPEs with Li/GPE/stainless steel (SS) sealed cell and Li/GPE/carbon black electrode (CB) cell exposed to ambient oxygen over a wide scanning range (open circuit potential to 6.5 V) at slow scanning rate (1.0 mv s -1 ). The preparation of carbon black electrode will be mentioned later. Liquid TEGDME electrolyte was characterise the same way for comparison. The as-prepared GPE was sandwiched between two symmetrical stainless steel electrodes and sealed in a Swagelok cell. The impedance was conducted by electrochemical workstation CH Instrument 660D in the frequency range khz. 82

99 The carbon electrode slurries were prepared by mixing carbon black (90 wt%) and poly(tetrafluoroethylene) (PTFE) (10 wt%) together in propanol. The mixture was then coated on a nickel mesh substrate. The cathode film was then obtained by punched into disc and dried under vacuum at 80 o C for 12 h. A Swagelok cell with an air hole (0.785 cm 2 ) on the cathode side was used to investigate the discharge and charge performance. The Li-air cells were assembled in a glovebox with water and oxygen level less than 0.1 ppm. The as-prepared GPEs were sandwiched between lithium foils and carbon black electrodes and a glass microfiber filter was used for keeping the structure from collapsing. For comparison, a cell with a glass microfiber filter soaked in liquid TEGDME electrolyte was also made. Same measurements were also performed with GPEs with PC and DMSO as plasticizers. All the cells were gas-tight except for the cathode side window. All the measurements were conducted in 1 atm in dry oxygen atmosphere to avoid any negative effects of humidity and CO Results and discussion The most three important factors that determine whether one electrolyte can be used in lithium battery or not are the reversibility of Li + insertion and extraction reaction, electrochemical stability, and ionic conductivity. Figure 4-2 displays the cyclic voltammetry curve using TEGDME based GPE sandwiched between two lithium electrodes. The curve shows only one anodic peak and one cathodic peak and the values of potential and current density were almost the same, which indicated the reversibility of Li + insertion and extraction. The sufficient reversibility of Li + insertion and extraction reaction can ensure the smoothly running of battery reaction during cycling. Therefore, the as-prepared TEGDME based GPE could be applied in Li-O 2 batteries. 83

100 Li/PVDF-HFP/Li Current (A) Figure 4-2 The cyclic voltammetry curve of Li/GPE/Li typed cell with TEGDME based GPE as electrolyte Potential (V) Figure 4-2 The cyclic voltammetry curve of Li/GPE/Li typed cell with TEGDME based GPE as electrolyte The second concern is the electrochemical stability of an electrolyte when using in lithium batteries. The demand of wide operating window can eliminate unwanted side reactions and ensure an acceptable electrochemical performance. Figure 4-3 (a) shows the result of linear sweep voltammetry measurement using sealed battery with TEGDME based GPE sandwiched between a lithium electrode and a stainless steel electrode. For comparison, the one with liquid TEGDME electrolyte was also measured using the same method. TEGDME based GPE could stay stable up to 5.5 V while the liquid electrolyte suffered decomposition at the voltage of 4.5 V. This indicates the gelled polymer electrolyte had higher stability than the liquid one. Normally, a lithiumion battery requires a stability window up to 4 V and these results were sufficient enough. However, as Li-O 2 batteries must be operated in an open atmosphere because oxygen is the reactant for battery reactions, the stability data in the sealed batteries is 84

101 not persuadable enough. In order to investigate the stability, linear sweep voltammetry measurement was conducted with an open battery using TEGDME based GPE sandwiched between a lithium electrode and a porous carbon black electrode with a hole at the carbon black side as the gas channel. The liquid TEGDME electrolyte was also measured for comparison. The results are shown in Figure 4-3 (b). The current density of battery with TEGDME based GPE didn t increase until 4.5 V while the one of battery with liquid electrolyte increased from 4 V. The results indicate the exposure to oxygen atmosphere can decrease the stability of electrolyte systems greatly and choosing a suitable electrolyte for Li-O 2 battery is much difficult than Li-ion battery. The stability of TEGDME based GPE showed in Figure 4-3 ensured the usability in Li- O 2 batteries. Current (A) Li-PVDF-HFP-SS Li-TEGDME-SS Figure 4-3 Linear sweep voltammetry results of (a) Li/GPE/SS and (b) Li/GPE/CB-air type cells (a) Potential (V) 85

102 Li-PVDF-HFP-air Li-TEGDME-air (b) Current (A) Potential (V) Figure 4-3 Linear sweep voltammetry results of (a) Li/GPE/SS and (b) Li/GPE/CB-air type cells. The ionic conductivity is another concern for the usage of GPE in Li-O 2 systems. Sufficient ionic conductivity can ensure a smooth running of battery reactions. The ionic conductivity was measured through A.C Impedance measurement and the values were calculated from the membrane resistances obtained from the impedance spectra. The results are displayed in Figure 4-4. It was seen that the ionic conductivity increase with the increase of temperature. This is probably due to the reason that the capability of ions to move freely in the electrolyte system was greatly improved when the temperature was increased. The ionic conductivity at the room temperature (25 ºC) was roughly 10-3 Ωcm -1 and it was considered sufficient for Li-O 2 batteries. 86

103 0.005 PVDF-HFP Conductivity ( cm -1 ) Figure 4-4 The calculated ionic conductivity of PEG at different temperatures Temperature ( o C) Figure 4-4 The calculated ionic conductivity of PEG at different temperatures. All the measurement confirmed the properties of TEGDME based GPE could be used in Li-O 2 batteries. The electrochemical performance of as-prepared GPE was investigated in a battery with TEGDME based GPE sandwiched between a lithium anode and a porous carbon black electrode. Same battery construction was used for liquid TEGDME electrolyte. The results of discharge and charge profiles in first cycle are shown in Figure 4-5. Battery with TEGDME based GPE as electrolyte could deliver specific capacity of 2,988 mah g -1 and this result was acceptable according to the previous reports of other groups. However, the battery with liquid TEGDME as electrolyte showed even higher capacity of 7,921 mah g -1 and lower over-potential. The possible reason for such differences of electrochemical performance in the first cycle for both electrolytes is that the performance was greatly influenced by their intrinsic properties such as ionic conductivity and phase structure. The liquid TEGDME was in its liquid phase which means the ions in the solution could move freely and provide relatively high ionic conductivity. On the other hand, the GPE was in its solid phase which 87

104 indicated the movement of the ions in the electrolyte system could be restricted. Although the impedance measurement confirmed the ionic conductivity of GPE was sufficient for Li-O 2 batteries, it was still lower than the plasticizer before it was made into GPE. This could explain the lower capacity and higher over-potential of battery with TEGDME based GPE. Another reaction is the solid state GPE did not have the same penetrability as the liquid state TEGDME. This meant the contact area of electrolyte and cathode was smaller for GPE. The active site on the cathode for oxygen reduction reaction was reduced and this led to the reduction of specific capacity. Despite these drawbacks, the first cycle performance of battery with TEGDME based GPE was acceptable for Li-O 2 batteries Voltage (V) PVDF-HFP Figure 4-5 The discharge and charge profiles in the first cycle of TEGDME Li-O 2 batteries using different electrolyte ,000 4,000 6,000 8,000 Specific capacity (mah g -1 ) Figure 4-5 The discharge and charge profiles in the first cycle of Li-O 2 batteries using different electrolyte. The current density was 50 mah g -1. The cycling performance of the battery with TEGDME based GPE was also investigated with a fixed capacity of 500 mah g -1. The result is shown in Figure 4-6. It 88

105 is clearly seen that the battery with TEGDME based GPE could last 25 cycles with nearly no differences in discharge and charge potentials. The reversibility was great according to this result and this demonstrated the usability of TEGDME based GPE in Li-O 2 batteries Voltage (V) st cycle 5th cycle 10th cycle 15th cycle 20th cycle 25th cycle Figure 4-6 Discharge and charge profiles of Li-O 2 batteries with TEGDME based GPE as electrolytes at fixed capacity to 500 mahg Specifi capacity (mah g -1 ) Figure 4-6 Discharge and charge profiles of Li-O 2 batteries with TEGDME based GPE as electrolytes at fixed capacity to 500 mah g -1. Current density is 50 ma g -1. In order to further demonstrate the sufficient stability of GPE in Li-O 2 batteries, cycles with fully discharge and charge were operated. For comparison, the battery with liquid electrolyte was also been cycled under the same condition. The results are displayed in Figure 4-7. The battery with TEGDME based GPE could last for at least 5 cycles with only a slight drop of capacity. The capacity of battery with liquid TEGDME electrolyte faded quickly from the second cycle although it could provide extremely high initial discharge capacity. This provided the evidence that the stability of the plasticizer was 89

106 greatly improved after made in to GPE. The stability of TEGDME based GPE system made it possible to be used in Li-O 2 batteries. 9,000 Specific Capacity (mah g -1 ) 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 Figure 4-7 Cycle profiles of Li-O 2 batteries with PVDF-HFP based GPE and TEGDME as electrolytes PVDF-HFP TEGDME Cycle numbers Figure 4-7 Cycling profiles of Li-O 2 batteries with PVDF-HFP based GPE and TEGDME as electrolytes. Current density is 50 ma g -1. The superior cycling performance of TEGDME based GPE is probably due to the gelled structure of the polymer electrolyte system which is proposed to be set up by the interaction between polymer matrix PVDF-HFP and plasticizer TEGDME through Li +. A brief schematic mechanism is shown in Figure 4-8. According to the reference, the Li + serves as a cross-link between polyether and PVDF or HFP chains with the Li + bonding the former via oxygen atoms and the latter via fluorine atoms [149]. As Li + was used as cross-link in the structure, the amount of free Li + in the polymer electrolyte system was greatly reduced, which explained the decrease of ionic conductivity. At the same time, the interactions could also help increase the stability of the α-hydrogen and 90

107 α-carbon atoms. This resulted in the improving stability of the plasticizer TEGDME. The mechanism was detailedly discussed in Chapter 3. In this way the stability of the whole electrolyte system was significantly improved. Figure 4-8 Proposed structure of PVDF-HFP based GPE Figure 4-8 Proposed structure of PVDF-HFP based GPE In order to find out more about the interactions, DMSO and PC were used to replace TEGDME to break the interactions. The batteries with DMSO based and PC based GPEs were discharged and charged under the same condition. The results are shown in Figure 4-9. It was clearly seen that the specific capacities of both batteries increased but the cycleability was greatly reduced. This confirmed the existence of unique interactions between TEGDME and polymer matrix PVDF-HFP. The results listed in Figure 4-9 also indicate the plasticizer could affect the electrochemical performance. The battery with DMSO based GPE as electrolyte could deliver higher specific capacity than the 91

108 one with PC based GPE. The differences should lie on the different properties of DMSO and PC plasticizers. Voltage (V) Voltage (V) PC (a) Figure 4-9 The discharge and charge profiles in the first cycle of PC (a) and DMSO (c) based GPEs and the 1,000 cycling performance of PC (b) and DMSO (d) based GPEs 0 1,000 2,000 3,000 4,000 Specific capacity (mah g -1 ) DMSO (c) Specific capacity (mah g -1 ) Specific capacity (mahg -1 ) 5,000 PC 4,000 3,000 2, Cycle number 14,000 DMSO 12,000 10,000 8,000 6,000 4,000 2,000 (b) (d) ,500 5,000 7,500 10,000 12,500 15,000 Specific capacity (mah g -1 ) Cycle number Figure 4-9 The discharge and charge profiles in the first cycle of PC (a) and DMSO (c) based GPEs and the cycling performance of PC (b) and DMSO (d) based GPEs. The current density was 50 mah g -1. It is very difficult to choose between initial capacity and cycleability because they both characterize the performance of Li-O 2 batteries. According to the reference, the cycling performance is the main issue that limits the practical application of Li-O 2 batteries [103]. In that case, TEGDME based GPE showed better cycling performance and was demonstrated suitable for Li-O 2 batteries. 92

109 4.4 Summary Gel polymer electrolytes with PVDF-HFP as polymer matrix have been successfully prepared and applied in Li-O 2 batteries. PVDF-HFP based GPE showed sufficient reversibility of Li + insertion and extraction, good stability during discharge and charge process, and high ionic conductivity. The battery with PVDF-HFP based GPE exhibited high cycling performance which was much better than the liquid TEGDME electrolyte. It is believed the interactions between the components in GPE helped stabilize the electrolyte system. The electrochemical performance of batteries with GPEs based on different plasticizer was also investigated. It is found that the performances of GPEs were significantly affected by the use of plasticizer and the best choice is TEGDME as plasticizer when compared with PC and DMSO. Therefore, we expect that PVDF-HFP- TEGDME based GPE could be used as promising electrolyte in long-life Li-O 2 batteries. 93

110 Chapter 5 Conducting Polymer-Doped Polypyrrole as An Effective Cathode Catalyst for Li-O 2 Batteries 5.1 Introduction The Nobel Prize in Chemistry was awarded in 2000 for work on conductive polymers, including polypyrrole (PPy) [237, 238]. PPy is a conjugated compound formed from a number of connected pyrrole ring structures. The pure PPy is considered as insulator. However, the oxidized PPy is very good electrical conductor. The conductivity is mainly based on the conditions and reagents used when the polymerization process is carried on. The conductivity can be ranged from 1 to 100 Scm -1. The typical structure of PPy is displayed in Figure 5-1. Figure 5-1 The typical structure of PPy Figure 5-1 The typical structure of PPy. PPy has attracted extensive investigations recent years due to the high electrical conductivity, high mechanical strength, and high chemical stability. It has been applied into several areas such as electric devices and chemical sensors. Li-ion batteries are considered heated research areas recent days and PPy has been widely used in this areas [ ]. It is believed that the PPy can be used as the coating materials or directly used as cathode materials. Owing to the intrinsic reversible redox properties, doped PPy 94

111 could be potentially used as catalysts for lithium batteries. The electrode reaction is considered as below, (5-1) where A - stands for the dopant ions [155, 160, 247]. PPy is also considered to have great catalytic properties for oxygen reduction reactions [ ]. Cui et al. reported replacement of carbon with tubular PPy as catalyst for Li- O 2 batteries and achieved a very good cycling performance [70]. However, the influence of different dopants on the electrochemistry performance of PPy as catalyst in Li-O 2 batteries has not yet been reported. Despite all the superiorities PPy shared, there have been seldom reports in Li-O 2 batteries. In this chapter, we report the use of polypyrrole with different dopants as catalyst in non-aqueous Li-O 2 batteries for the first time. Polypyrrole doped with Cl - - and ClO 4 were synthesized and applied as the cathode catalyst in Li-O 2 batteries. 5.2 Experiment Synthesis of materials Polypyrrole with different dopants was synthesized by an in situ chemical polymerization method. In a typical synthesis process, 10 mmol pyrrole monomers were added into 40 ml aqueous solution consisting of 1 M dopant acid under stirring. The dopant acids are HCl and HClO 4, respectively. After stirring for 30 min, 20 ml aqueous solution consisting of 5 mmol (NH 4 ) 2 S 2 O 8 and 1 M dopant acid was added into the previous solution. The mixture was kept stirring at room temperature for 6 h. The black precipitate was filtered and washed with distilled water and ethanol for several times and then dried in a vacuum oven at 60 o C for 12 h. 95

112 5.2.2 Characterization of samples Field emission scanning electron microscope (FESEM, Zeiss Supra 55 VP) was used to investigate the morphology of the as-prepared PPy polymers. Infrared spectra were measured using a Nicolet Magna 6700 FT-IR spectrometer. All spectra were obtained using 4 cm -1 resolution and 64 scans at room temperature Electrochemical measurements Cathode slurry was prepared by mixing the as-prepared PPy, poly(tetrafluoroethylene) (PTFE) with Super-P carbon black together in isopropanol with the weight ratio of 60:10:30. The mixture was coated on a stainless steel mesh substrate and then cut into discs with a diameter of 14 mm. The electrodes were dried at 80 o C under vacuum for 12 h. The loading of the cathode materials is about 2 mg cm -2. For comparison, pure carbon black electrodes were fabricated by mixing PTFE and Super-P carbon black in isopropanol with the weight ratio of 10:90. Some carbon black electrodes were chosen to soak into 0.5 M LiCl ethanol solution under vacuum for 3 hours and dried under vacuum at 150 o C for 6 h. A Swagelok type cell with an air hole (0.785 cm 2 ) on the cathode side was used to test the electrochemical performance. The cell was assembled in an argon filled glove box (Mbrau) with water and oxygen level less than 0.1 ppm. A lithium foil was used as the anode and was separated from cathode by a glass microfiber filter (Whatman) soaked in electrolyte (1 M LiClO 4 in propylene carbonate). The cell was gas tight except for the cathode side window that exposed the cathode film to the oxygen atmosphere. All measurements were conducted in 1 atm in dry oxygen atmosphere to avoid any negative effects of humidity and CO 2. 96

113 5.3 Results and discussion Figure 5-2 (a) and (b) show the SEM images of the as-prepared PPy-Cl and PPy-ClO 4 polymers. Both of them showed similar morphology, which consists of nanoparticles with a diameter about nm. The FT-IR spectra of the as-prepared PPy are shown in Figure 5-2 (c). The pyrrole ring vibrations can be observed at 1,544 and 1,456 cm -1 and =C-H vibrations appeared at 1,298 and 1,042 cm -1 in the PPy spectra. The vibration at 1175 cm -1 can be assigned to the stretching of C-N group. Figure 5-2 SEM images of the as-prepared ar ed (a) PPy- Py Cl and (b) PPy-ClO Py-Cl 4, and (c) FT-IR spectra of both PPy polymers 97

114 (C) =C-H PPy-Cl PPy-ClO 4 Transmition (a. u.) Fudamental vibrations of pyrrole ring -C-N Wavenumbers (cm -1 ) Figure 5-2 SEM images of the as-prepared (a) PPy-Cl and (b) PPy-ClO 4, and (c) FT-IR spectra of both PPy polymers. The electrocatalytic activity of the as-prepared PPy was examined in Li-O 2 battery and compared with carbon black at room temperature. The charge-discharge voltage range was set between 2.0 and 4.5 V for all the measurements. The voltage profiles of the electrodes in the first cycle are shown in Figure 5-3 (a). The discharge capacities of PPy-Cl and PPy-ClO 4 electrodes are 6,208 mah g -1 and 5,164 mah g -1, respectively. The carbon black electrode delivered a lower capacity of 1,365 mah g -1. The discharge plateaus of both conducting polymer electrodes were higher than that of carbon black electrode indicating PPy had better catalytic activity towards ORR than carbon black. The charge potential plateaus of conducting polymer electrodes were also lower comparing to carbon black electrode indicating PPy had better catalytic activity towards oxygen evolution reaction (OER) than carbon black. The cycling performances of both PPy and carbon black electrodes are shown in Figure 5-3 (b). PPy-Cl and PPy-ClO 4 electrodes exhibited much better cycling stability than carbon black electrodes in the 98

115 first five cycles. Furthermore, PPy doped with Cl - showed better cycling stability than the PPy-ClO 4 electrode. (a) 4.5 Voltage (V) PPy-Cl PPy-ClO 4 Carbon black Figure 5-3 The discharge-charge profiles and (b) cycling performance of PPy-Cl, PPy-ClO 4 and carbon black electrodes 0 2,000 4,000 6,000 Specific capacity (ma h g -1 carbon ) (b) Specific capacity (mah g -1 carbon ) 8,000 6,000 4,000 2,000 0 carbon black PPy-Cl PPy-ClO Cycle number Figure 5-3 (a) The discharge-charge profiles and (b) cycling performance of PPy-Cl, PPy-ClO 4 and carbon black electrodes. The discharge-charge current density is 100 ma g -1 in 1 atm O 2 at room temperature. 99

116 The ORR reaction with the presence of PPy on cathode in Li-O 2 batteries could involve two steps: (5-2) (5-3) As shown in Figure 5-4 (a), the π-bond orbitals of the oxygen molecule have the tendency to accept one electron by overlapping with pyrrole rings, which consist of delocalized π-bonds formed by carbon atoms in the state of sp 2 hybridization, corresponding to the equation (3). Khomenko et al. demonstrated that the carbon atoms in 3 and 4 positions of pyrrole rings can combine with oxygen atoms. The polypyrrole is the electron density donor and the oxygen is electron density acceptor [155], which makes the ORR occurring on PPy catalysts. In the whole discharge process on the PPy- ClO 4 cathode, PPy-ClO 4 is firstly reduced into its undoped state and releases ClO - 4 ions - into the electrolyte and then the undoped PPy is oxidized to PPy-ClO 4 by O 2 with ClO 4 ions from electrolyte. After the reduction of PPy-Cl, Cl - will go into electrolyte and LiCl will precipitate on the surface of PPy. When the reduced PPy reacts with O 2, it will react with ClO - 4 from the electrolyte to form PPy-ClO 4 instead of PPy-Cl (Figure 5-4 (b)). To further investigate the ORR on the cathodes, the FTIR spectra of all electrodes after discharge and charge are shown in Figure 5-5. The discharge products were dominated by Li 2 CO 3 and after charge, the peaks corresponding to Li 2 CO 3 disappeared. Recently, many investigations demonstrated that ORR in alkyl carbonate electrolytes is more complicated than simply forming Li 2 O 2 as the discharge product [21, 26]. The discharge reaction also involves the decomposition of the electrolyte and the oxidation of carbon materials to form other products, such as Li 2 CO 3, CH 3 COOLi, HCOOLi, C 3 H 6 (OCOOLi) 2, CO 2 and H 2 O. Although these products can be decomposed in the 100

117 charge process, the loss of electrolyte and the accumulation of discharge products will cause capacity loss during cycling. Although the PPy-ClO 4 electrode showed better capacity retention than that of carbon black electrode, it still suffered from severe capacity fading during cycling. On the other hand, the PPy-Cl electrode exhibited a better cycling stability than the PPy-ClO 4 electrode, which is probably due to the formation of LiCl on the surface of the electrode materials. Figure 5-4 The mechanism of (a) oxygen activation tion of PPy and (b) doping-undoping process of PPy-Cl and PPy-ClO 4 Figure 5-4 The mechanism of (a) oxygen activation of PPy and (b) doping-undoping process of PPy-Cl and PPy-ClO 4. (a) * Pristine electrode Discharged electrode Charged electrode Li 2 CO 3 Transmition (a.u.) Figure 5-5 FT-IR spectra of (a) PPy-Cl, (b) PPy-ClO 4, and (c) carbon black electrodes before discharge, after discharge and after charge process * * * * * 1,800 1,600 1,400 1,200 1, Wavenumber (cm -1 ) 101

118 (b) * Pristine electrode Discharged electrode Charged electrode Li 2 CO 3 Transmition (a.u.) * * * * * 1,800 1,600 1,400 1,200 1, Wavenumber (cm -1 ) (c) * pristine electrode discharged electrode charged electrode Li 2 CO 3 Transmition (a.u.) * * * * * 1,800 1,600 1,400 1,200 1, wavenumber (cm -1 ) Figure 5-5 FT-IR spectra of (a) PPy-Cl, (b) PPy-ClO 4, and (c) carbon black electrodes before discharge, after discharge and after charge process. 102

119 As the amount of LiCl formation on the surface of PPy-Cl cathode is very small, it is very difficult to detect LiCl by X-ray diffraction analysis. In order to confirm the formation of LiCl on the PPy-Cl cathode after discharge and charge cycling, energy dispersive X-ray spectroscopy (EDS) was performed on the cycled electrodes. The EDS results of chloride element are shown in Table 5-1. We found that the content of Cl in PPy-Cl electrodes is about 4-5 times higher than that in PPy-ClO 4 electrodes. Therefore, a LiCl precipitation layer formed on the surface of PPy-Cl electrode during cycling process. Table 5-1 EDS results of PPy-Cl and PPy-ClO 4 electrodes after cycling Cl % PPy-Cl electrode Atomic percentage 3.36 Weight percentage 8.06 PPy-ClO 4 electrode Atomic percentage 0.77 Weight percentage 1.93 Table 5-1 EDS results of PPy-Cl and PPy-ClO 4 electrodes after cycling In order to explain the differences of the electrochemical performances between PPy-Cl and PPy-ClO 4, a carbon black electrode with the addition of LiCl was tested and compared with pure carbon black electrode. The charge-discharge voltage profiles and cycling performances are shown in Figure 5-6. After the addition of LiCl, the carbon 103

120 black electrode exhibited reduced charge-discharge over-potential and better cycling stability than the bare carbon black electrode. The improved performance could be attributed to the formation of LiCl layer on the surface of the electrode. It is believed the LiCl layer can help protect the electrode from reacting with the electrolyte to form byproducts [68]. (a) Voltage (V) Figure 5-6 (a) The charge-discharge profiles and (b) the cycling performance Carbon black + LiCl of carbon black electrodes with and Carbon black without LiCl additive ,000 1,500 2,000 Specific capacity (ma h -1 carbon ) (b) Specific capacity (mah g -1 carbon ) 3,000 2,500 2,000 1,500 1, carbon black + LiCl carbon black Cycle number Figure 5-6 (a) The charge-discharge profiles and (b) the cycling performance of carbon black electrodes with and without LiCl additive. The current density is 100 ma g -1 in 1 atm O 2 at room temperature. 104

121 The formation of LiCl layers on the cathode surface not only can protect carbon from reacting with electrolyte to form discharge byproducts, but also can participate in the discharge and charge reactions. It is well known that LiCl is a commonly used Lewis acid catalyst in many organic reactions. The O - 2 radical formed on the cathode is a strong Lewis base. Therefore, LiCl may interact with O - 2 and stabilize the superoxide radical. This process may lower the barrier of the reaction and provide a reduced overpotential. The schematic mechanism is shown in Figure 5-7. Therefore, the presence of LiCl layer on the cathode could lead to better cycling performance and lower overpotential. Figure 5-7 Schematic mechanism of discharge process on cathode with LiCl addition Figure 5-7 Schematic mechanism of discharge process on cathode with LiCl addition. 105

122 5.4 Summary Polypyrrole with different dopants have been synthesized and applied as cathode catalysts in Li-O 2 batteries in alkyl carbonated electrolyte. The PPy electrodes showed higher discharge capacities and lower over-potential than that of the carbon black electrode. The mechanism of this phenomenon suggested that the PPy has an excellent redox property and a capability to activate oxygen reduction reaction. Furthermore, PPy doped with Cl - exhibited better cycling stability than that of PPy doped with ClO - 4. We expect that PPy-Cl polymer could be a promising cathode catalyst for Li-O 2 batteries. 106

123 Chapter 6 Conducting Polymer Coated CNT Used in Li-O 2 Batteries with Enhanced Electrochemical Performance 6.1 Introduction Polypyrrole (PPy) and poly(3,4-ethylenedioxythiophene) (PEDOT) are considered two most commonly used conducting polymers. PPy is formed by a number of connected pyrrole rings. PEDOT is a derivative of polythiophen (PT) based on the connected structures of 3,4-ethylenedioxythiophene (EDOT) monomers. Figure 6-1 shows the typical structure of PEDOT. Both conducting polymers have great chemical, mechanical and electrochemical properties. Due to these superior properties, PPy and PEDOT have been widely applied in many areas such as electric devices, anticorrosion coating, and chemical sensors. Figure 6-1 The typical structure of PEDOT Figure 6-1 The typical structure of PEDOT. 107

124 The conducting polymers used as catalysts in electric devices are believed to be involved in a unique redox reaction, which is also known as doping-dedoping process. The doped conducting polymers can be reduced into their undoped forms and this process can be reversed. The process is show in the equation below, (6-1) where CP stands for conducting polymer, and A stands for anion. In Chapter 5, PPy was used as catalyst in Li-O 2 batteries and showed high catalytic activity towards battery reactions. However, from Chapter 5, it is easy to find out that the capacity of batteries with PPy as catalyst still faded after few cycles. One reason was the instability of propylene carbonate electrolyte towards oxygen reduction reaction. The other was the partial reversibility of the redox reaction of PPy. Because only oxidized PPy or known as doped PPy was conductive and it was very difficult to oxidize the PPy once it was reduced because the undoped one was an insulator. Due to this reason, finding a way to maintain the conductivity of PPy while cycling is very important. In order to overcome this drawback, conducting polymer coated carbon materials are prepared. Carbon materials can be used as conductive matrix for electron transportation and conducting polymers can be used as catalyst for battery reactions. In recent years, multiwall carbon nanotubes (CNT) have been studied intensively because of their high conductivity along with their mechanical strength. Combining with conducting polymers coated on the surface, the composite shows improved electrochemical properties. There have been many researches carried on in this area [ ]. Some of these researchers believed that CNT can be used as dopant for conducting polymers. This provides the possibility that composite materials consisting of CNT and conducting polymers can find specific applications in the field of energy 108

125 storage, optical limiting and electron field emission devices, transistors, conducting textiles, sensors, biomedical and food domains [260]. In this chapter, PPy and PEDOT were used to coat on the surface of CNT. The composites were employed as catalysts in the Li-O 2 batteries for the first time. The effects of different ratio and different conducting polymers were also investigated. 6.2 Experiment Synthesis of materials Polypyrrole coated carbon nanotubes (PPy/CNTs) with different ratio were prepared by an in situ chemical polymerization method. CNT was pre-treated by HNO 3 to add function groups such as COOH to the surface of CNT. Acid-treated CNT (30 mg) was dispersed into 20 ml aqueous solution consisting of 1 M HCl by the supersonic method. Pyrrole monomers (0.5 mmol) were added into this suspension. After mechanically stirring for 30 min, 20 ml aqueous solution consisting of 2 mmol FeCl 3 (4 times of pyrrole) and 1 M HCl was added into the previous suspension. The mixture was kept stirring at room temperature for 6 h. The black precipitate was filtered and washed with distilled water and ethanol for several times and then dried in a vacuum oven at 60 ºC for 12 h. The same process was used only with different amount of pyrrole and FeCl 3. The amount of pyrrole was kept at 0.25 mmol and 1mmol. The as-prepared PPy/CNTs were weighed after drying process and the results were roughly 60, 45 and 90 mg, respectively. Since the pristine CNT was 30 mg, the as-prepared PPy/CNTs were named PPy/CNT 1:1, PPy/CNT 1:2, and PPy/CNT 2:1, respectively. The same synthesis process was used to prepare poly(3,4-ethylenedioxythiophene) coated carbon nanotube (PEDOT/CNT) with the ratio of 1:1. The as-prepared materials were ready to characterize. 109

126 6.2.2 Characterization of samples Field emission scanning electron microscope (FESEM, Zeiss Supra 55 VP) was used to investigate the morphology of as-prepared materials. Infrared spectroscopy was conducted on a Nicolet Magna 6700 FT-IR spectrometer. All spectra were obtained using 4 cm -1 resolution and 64 scans at room temperature. The conducting polymers and CNT content in the composite materials was determined by TGA on a Mettler Toledo TGA/DSC instrument in air at 10 ºC min -1 at temperature range of ºC Electrochemical measurements Cathode slurry was prepared by mixing the as-prepared materials, poly(tetrafluoroethylene) (PTFE) with Super-P carbon black together in isopropanol with the weight ratio of 60:10:20. The mixture was coated on a stainless steel mesh substrate and the cut into discs with a diameter of 14 mm. The electrodes were dried at 80 ºC under vacuum for 12 h. The loading of the cathode materials is about 1 mg cm -2. For comparison, CNT cathode was also prepared with the same process. A Swagelok type cell with an air hole (0.785 cm 2 ) on the cathode side was used to test the electrochemical performance. The cell was assembled in an argon filled glove box (Mbrau) with water and oxygen level less than 0.1 ppm. A lithium foil was used as the anode and was separated from cathode by a glass microfiber filter (Whatman) soaked in electrolyte (1 M LiTFSI in Diethylene glycol dimethyl ether). The cell was gas tight except for the cathode side window that exposed the cathode film to the oxygen atmosphere. All measurements were conducted in 1 atm in dry oxygen atmosphere to avoid any negative effects of humidity and CO

127 6.3 Results and discussion Figure 6-2 (a)-(e) show the SEM images of the as-prepared PPy/CNTs and PEDOT/CNT. All of them showed the similar tubular morphology only with different diameters. The pristine CNT had the diameter of 50 nm and the PPy/CNT 1:2 shared the similar diameter. The diameters of PPy/CNT 1:1 and PPy/CNT 2:1 were larger than the previous ones, which are 100 and 200 nm, respectively. This indicated that PPy was successfully coated onto these CNTs and higher content of PPy resulted in larger diameter. The morphology of as-prepared PEDOT/CNT 1:1 was similar to PPy/CNT 1:1 with the same diameter. The FT-IR specta of the as-prepared conducting polymer coated CNTs are shown in Figure 6-2 (f). The pyrrole ring vibrations can be observed at 1,544 and 1,456 cm -1 and the thiophen ring vibrations can be seen at 1,365 cm -1. =C-H vibrations appeared at 1,298 and 1,042 cm -1 in the spectra. The vibration at 1,175 cm -1 can be assigned to the stretching of C-N group and the vibration at 1,100 cm -1 can be assigned to the stretching of C-O-C in the ethylenedioxy group. The results clearly indicate the successful synthesis of the conducting polymer coated CNTs. Although the weight ratios were roughly determined by the weight comparison before and after synthesis, TGA method was still employed to further confirm the exact results. The TGA spectra are shown in Figure 6-3. Due to the similar decomposition temperatures of PPy and CNT, it is quite difficult to tell the weight ratio only referring to the weight ratio curves which is shown in Figure 6-3 (a)-(c). However, the results can be still found in the temperature difference curves. It was found that there were two peaks in each curve and the position indicated the weight content of CNTs in the composites were 34%, 48%, and 65% for PPy/CNT 1:2, PPy/CNT 1:1, PPy/CNT 2:1, respectively, which are consistent with the previous results. For PEDOT/CNT 1:1, it is 111

128 easy to tell the weight ratio because of the large difference decomposition temperatures of PEDOT and CNT shown in Figure 6-3 (d). The ratio was roughly 47%, which was consistent of previous result. Figure 6-2 The SEM images of (a) the bare CNT, the as-prepared (b) PPy/CNT 1:2, (c) PPy/CNT 1:1, (d) PPy/CNT 2:1, (e) PEDOT/CNT 1:1, and (f) FT-IR spectra CNT PPy/CNT 1:2 (f) PPy/CNT 1:1 PPy/CNT 2:1 Transmition (a.u.) PEDOT/CNT 1: Wavenumber (cm -1 ) Figure 6-2 The SEM images of (a) the bare CNT, the as-prepared (b) PPy/CNT 1:2, (c) PPy/CNT 1:1, (d) PPy/CNT 2:1, (e) PEDOT/CNT 1:1, and (f) FT-IR spectra. 112