Influence of Battery Cell Components and Water on the Thermal and Chemical Stability of LiPF6 Based Lithium Ion Battery Electrolytes

Influence of Battery Cell Components and Water on the Thermal and Chemical Stability of LiPF6 Based Lithium Ion Battery Electrolytes

G Model EA 28390 No. of Pages 5 Electrochimica Acta xxx (2016) xxx–xxx Contents lists available at ScienceDirect Electrochimica Acta journal homepa...

851KB Sizes 0 Downloads 34 Views

G Model EA 28390 No. of Pages 5

Electrochimica Acta xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta

Influence of Battery Cell Components and Water on the Thermal and Chemical Stability of LiPF6 Based Lithium Ion Battery Electrolytes Simon Wiemers-Meyera , Sebastian Jeremiasa , Martin Wintera,b , Sascha Nowaka,* a b

University of Münster, MEET Battery Research Center, Institute of Physical Chemistry, Corrensstraße 46, 48149, Münster, Germany Helmholtz-Institute Münster (HI MS), IEK-12 Forschungszentrum Jülich, Corrensstraße 46, 48149, Münster, Germany

A R T I C L E I N F O

Article history: Received 13 October 2016 Received in revised form 14 November 2016 Accepted 17 November 2016 Available online xxx Keywords: electrolyte degradation lithium ion battery electrolyte NMR spectroscopy thermal stability

A B S T R A C T

Lithium ion battery electrolytes based on LiPF6 and organic solvents are known to degrade at elevated temperatures. The degradation reactions can be caused either chemically e.g. by simple contact with battery cell components and/or electrochemically during cycling. This study is focused on thermally induced chemical reactions of the electrolyte with different battery cell components. These reactions are monitored by means of quantitative NMR spectroscopy. The results allow for categorizing the influences of the components according to their reactivity against HF. Inert materials (graphite, carbon black, polyvinylidene difluoride, polyolefinic and ceramic separator) do not show any observable influence on the thermal stability of the electrolyte. If the materials react with HF but the reaction does not form water in significant amounts (Li metal and LiNi1/3Co1/3Mn1/3O2), there is also no influence observable. In contrast to that, materials, which can form water in contact with HF at significant rates (glass fiber separator, Si and LiFePO4), can lead to a slightly increased or even severe electrolyte degradation. However, if the material neutralizes the acid HF (carboxymethyl cellulose), it stabilizes LiPF6 against water sources. Furthermore, the results of this study show that LiPF6 is stable at temperatures up to 80 C, if no water sources are present. This stability is most likely also given for even higher temperatures. ã 2016 Elsevier Ltd. All rights reserved.

1. Introduction Lithium ion batteries (LIBs) are considered as a key technology for electric and hybrid electric vehicles as well as stationary energy storage systems [1]. They usually consist of a graphite anode and a transition metal oxide cathode with a separator in between, soaked with electrolyte. The LIB technology has undergone extensive changes and advancements throughout the last two and a half decades [1–6]. The appearance and development of new electrode materials enlarged the specific capacity of the batteries and prolonged their cycle life [7,8]. However, the cycle life is still limited, due to several aging phenomena [9]. One of the major aging processes is attributed to degradation of LIB electrolytes [10]. The electrolyte is usually based on solutions of LiPF6 in mixtures of organic carbonates, e.g. ethylene carbonate (EC) and dimethyl carbonate (DMC). These compounds tend to decompose in different ways, namely, thermally at elevated temperatures [11–13], chemically e.g. in contact with glass sample containers [14] and electrochemically at

* Corresponding author. Fax: +49 251 83 36032. E-mail address: [email protected] (S. Nowak).

low and at high electrode potentials at anode and cathode, respectively [15,16]. In a battery, these processes are taking place simultaneously. Understanding of the individual influences of these processes requires a separate elucidation. Thermal aging of LIB electrolytes is already thoroughly investigated [10–14,16–20]. Since reactions of LiPF6 based electrolytes with glass sample containers have been observed in our previous publication [14], it is assumed that also some battery cell components can exhibit a similar behavior. The reaction with glass is caused by HF which is a degradation product of LiPF6 [14]. A comprehensive investigation of these reactions requires not only a qualitative identification of the degradation products but also quantitative data. Nuclear magnetic resonance (NMR) spectroscopy appeared to be a very powerful tool to investigate LIB electrolyte stability in a qualitative and quantitative manner [14]. The quantification method applied in this work is based on using EC as a heteronuclear quantification standard, which is explained in more detail in a previous publication [14]. In this work typical LIB cell components are investigated. Three different kinds of anode materials (Li metal, graphite and silicon) are analyzed. In half cell investigations of LIB cell components Li metal is typically used as reference and counter electrode material. It has to be noted that Li metal is an inherently charged electrode material. Since the

http://dx.doi.org/10.1016/j.electacta.2016.11.100 0013-4686/ã 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: S. Wiemers-Meyer, et al., Influence of Battery Cell Components and Water on the Thermal and Chemical Stability of LiPF6 Based Lithium Ion Battery Electrolytes, Electrochim. Acta (2016), http://dx.doi.org/10.1016/j.electacta.2016.11.100

G Model EA 28390 No. of Pages 5

2

S. Wiemers-Meyer et al. / Electrochimica Acta xxx (2016) xxx–xxx

electrode potential of Li metal is outside of the electrochemical stability window of the electrolyte, it will inevitably undergo a redox reaction with the electrolyte, which has to be taken into account in this study [5]. In contrast to that graphite and silicon as well as the analyzed cathode materials (LiNi1/3Co1/3Mn1/3O2 (NCM) and LiFePO4 (LFP)) are used in the discharged state. Therefore, no redox reactions are expected in these samples. Furthermore, three different kinds of separators (ceramic, polyolefinic and glass fiber) are investigated. Polymeric separators are assumed to be inert against the electrolyte and its decomposition products, while ceramic and especially glass separators might not be chemically stable against HF, due to their oxide groups. In addition, there are two binder materials (polyvinylidene difluoride (PVdF) and sodium carboxymethyl cellulose (CMC)) and the conductive agent carbon black investigated in this study. PVdF as a fluorinated polymer and carbon black are assumed to be inert. However, CMC is a Brønsted base, hence it can react with HF. The objective of this study is to classify and explain the impact of these cell components on the electrolyte stability. This means to find out whether they stabilize, destabilize or do not affect the electrolyte composition and furthermore, to identify the responsible chemical reactions. Clarification of the chemical reactions and decomposition products in a LIB after cell assembly is the starting point for further studies focusing on subsequent electrochemical reactions and the resulting decomposition products. 2. Experimental 2.1. Sample preparation The investigated electrolyte is battery grade SelectiLyteTM LP30 (BASF) which consists of a mixture of EC and DMC (1:1 by weight) and 1 mol L 1 LiPF6. The samples were stored in PTFE-FEP (polytetrafluoroethylene-fluorinated ethylene polypropylene copolymer) NMR tube liners (Wilmad-LabGlass) inside of flame sealed NMR glass tubes at 60  C. The cell components are listed in Table 1 They were dried under vacuum over one night and added to the electrolyte in a dry room (dew point < 65  C). Three samples of each kind were prepared. The amounts of different kinds of cell components differed because of their varying tap densities. The mass of silicon is not precisely determined due to large weighting errors caused by electrostatic interactions of the Si nanopowder. However, this does not affect the interpretations stated in this work. To prevent a reaction of lithium metal with the PTFE-FEP NMR tube liners polyolefinic separator was wrapped around the lithium. The number of separator layers in the other samples was chosen accordingly to the amount of the specific kind of separator usually used for cell assembling [21–23]. Furthermore,

LP30 electrolyte samples containing molecular sieve (3A, SigmaAldrich) were prepared. The molecular sieve was dried at 300 C under vacuum and afterwards placed inside of the glass but outside of the PFTE tubes. In addition, reference samples without molecular sieve were prepared. Each sample was stored at 60 C and 80 C before analysis. 2.2. Measurements The samples were aged for 42 days in total and measured periodically. The NMR measurements were performed on a Bruker Avance III HD spectrometer working at 400MHz (1H) with a broadband probe (PA BBO, 400MHz). The measurements were carried out at 15 C to slow down chemical reactions and to achieve sharper signals. 2.3. Data processing The software controlling the NMR spectrometer is TopSpinTM 3.2 (Bruker). This software and also MestReNova 10.0 (Mestrelab research) were used for NMR data processing. OriginPro 2015 9.2 (OriginLab) was used to plot data. 3. Results and discussion The degradation products found in the stored samples are HF, OPF3, OPF2OH, OPF2OCH3, BF4 and BF3-OPF2OH. The boron containing compounds are only formed in the glass fiber separator samples, where boron oxide is part of the separator material. These compounds are typical hydrolysis products of LiPF6 based electrolytes [14]. This indicates that the investigated cell components do not cause electrochemical degradation of the electrolyte under the given conditions. For reasons of comparability of the electrolyte stability of different samples, the losses of PF6 concentration are considered. These values are calculated as the sum of all PF6 degradation products containing phosphorous. The final values after 42days of aging at 60 C are depicted in Fig. 1. The logarithmic scale allows to compare values over a wide range. The samples containing glass fiber separators are much more degraded than all the other samples. A concentration of 215  27 mmolL 1 means that more than 20% of the initially contained LiPF6 is decomposed. The limit of detection (LOD) of LiPF6 degradation is 0.02% of the initial LiPF6 concentration. In our previous publication it was shown that if the electrolyte is stored in glass vials, HF reacts with glass and forms water that further hydrolyzes PF6 , creating a reaction cycle that keeps the degradation rate on a constantly high level [14]. The borate groups of the glass are very reactive against HF. As a result

Table 1 Investigated cell components, amounts and drying temperatures. Material

Supplier

Amount

Drying temperature/ C

Li metal foil Graphite (Timrex1 SFG 6) Si nanopowder LiNi1/3Co1/3Mn1/3O2 LiFePO4 (LFP P2 non coated) Ceramic separator (OZ-S30)

Rockwood Lithium Imerys Graphite & Carbon Evonik Creavis Toda Südchemie Mitsubishi

– 150 150 150 150 80

Polyolefinic separator (FS2226)

Freudenberg

Glass fiber separator (GF/D)

Whatman

PVdF (Kynar Flex1 761) Na-CMC (WalocelTM CRT2000 PA) Carbon black (SuperC651)

Arkema Dow Wolff Cellulosics Imerys Graphite & Carbon

8.1  0.3mg 3.7  0.2mg 3  2mg 18.8  0.2mg 5.5  0.2mg One layer, Diameter: 13mm Six layers, Diameter: 13mm One layer, Diameter: 13mm 11.8  0.2mg 15.3  1.4mg 2.7  0.1mg

80 300 150 150 150

Please cite this article in press as: S. Wiemers-Meyer, et al., Influence of Battery Cell Components and Water on the Thermal and Chemical Stability of LiPF6 Based Lithium Ion Battery Electrolytes, Electrochim. Acta (2016), http://dx.doi.org/10.1016/j.electacta.2016.11.100

G Model EA 28390 No. of Pages 5

S. Wiemers-Meyer et al. / Electrochimica Acta xxx (2016) xxx–xxx

Fig. 1. Concentrations c of degraded LiPF6 after 42 days at 60 C on a logarithmic scale calculated as the sum of all phosphorous containing aging products. Concentration in samples without addition of any cell component: 3.09  0.12 mmol L 1 [14]. The concentration values, obtained in this study are given in each row.

no HF but increasing concentrations of BF4 and BF3-OPF2OH are found in the samples containing glass fiber. It is expected that these reactions are slower at room temperature but are still taking place, which means that separators made of borate glass have a negative impact on the stability of LiPF6, which is in agreement with previously published data [18]. The second highest amount of degradation products is found in the electrolytes containing Si nanopowder (17.4  0.6 mmol L 1). Similar to the glass fiber containing samples these samples do also not contain HF. It can be concluded that the surface of the silicon particles is oxidized and undergoes a reaction with HF. This reaction forms water that accelerates the electrolyte degradation, as discussed above. The degradation proceeds as long as silicon oxides are present. The amount of hydrogen atoms either bound to fluorine or oxygen is constant and originates from initially contained water in the samples. Therefore, it is assumed that the initially present concentration of water determines the rate of water formation and PF6 degradation. When water forming sources like silicon oxide or borate glass are present in LIBs with LiPF6 based electrolytes, it is of particular importance to keep traces of water as low as possible. Previous publications report a beneficial effect of silicon oxides on the cycle life of silicon based anodes by reducing the mechanical stress during volume changes of the particles [24–26]. However, the results of this work propose potentially negative aging aspects of silicon oxides in LiPF6 based electrolytes. The LFP containing samples exhibit a final concentration of 6.16  0.30 mmol L 1 indicating a slightly increased degradation rate. The HF concentration of these samples (Fig. 2) increases during the first day by consumption of water traces and subsequently decreases. After 15 and 21 days, respectively, the HF concentration is below the LOD (0.3 mmolL 1). However, after 28 days it starts to rise again. LFP is known to slowly form LiOH in contact with moisture [27]. Decreasing HF concentrations are most likely caused by small amounts of LiOH on the LFP particle surface undergoing a reaction with HF thus forming water and solid LiF. It is assumed that no significant amounts of LiOH are formed via reaction of LFP with water during the measurement period. The formation of water is in accordance with slightly increased degradation rates. The sudden rise of the HF

3

Fig. 2. HF concentration curves of selected LP30 electrolyte samples. c(HF)
concentration after 28 days, resulting in values higher than during the first days, indicates total consumption of LiOH. However, the obtained NMR data does not indicate an explanation for the unexpectedly low rate of this process. The electrolyte samples containing polyolefinic separator (3.46  0.19mmolL 1), Li metal (2.45  0.12mmolL 1), ceramic separator (2.050.28 mmo lL 1), PVdF (1.84  0.58 mmol L 1), NCM (1.57  0.01 mmo lL 1), graphite (1.190.06mmolL 1) and carbon black (1.08 0.10mmolL 1) exhibit similar concentrations of phosphorous containing degradation products (Fig. 1). Without addition of any cell component the concentration values are in the same range [14]. Small differences might be explained by unequal drying behaviors resulting in slightly different, but still overall small amounts of water in the samples. Though the PF6 degradation rates are similar in these samples, the HF concentration curves might be different. The HF concentration in the lithium metal containing samples (Fig. 2) slowly decreases after seven days, most likely caused by a slow reaction of HF with lithium. In contrast to the lithium metal samples the HF concentration in the electrolytes containing NCM decreases rather fast. After 15 days the concentration is below the LOD. This, in combination with the fact that the electrolyte degradation rate of these samples is low indicates a medium fast reaction of HF with NCM which does not release significant amounts of water within the examined time scale. The HF concentration in the graphite containing samples steadily increases throughout the monitored time span. After a fast initial increase during the first day, the concentration rises at a lower rate in the following days. The same behavior is observed in the samples containing polyolefinic and ceramic separators, PVdF and carbon black. Furthermore, when no cell components are added to LP30 electrolytes, the HF concentration increases in the same way [14]. Therefore, it can be concluded that these cell components are inert against HF and do not influence the electrolyte stability. Though the above mentioned glass fiber and Si nanopowder samples as well as the CMC containing samples do not contain HF at concentrations above the LOD, not all of these samples exhibit an increased degradation rate. In fact, the CMC samples show the lowest degradation rate of all cell components investigated within this study. It is assumed that the base CMC neutralizes the acid HF,

Please cite this article in press as: S. Wiemers-Meyer, et al., Influence of Battery Cell Components and Water on the Thermal and Chemical Stability of LiPF6 Based Lithium Ion Battery Electrolytes, Electrochim. Acta (2016), http://dx.doi.org/10.1016/j.electacta.2016.11.100

G Model EA 28390 No. of Pages 5

4

S. Wiemers-Meyer et al. / Electrochimica Acta xxx (2016) xxx–xxx

which leads to precipitation of LiF. Furthermore, it is assumed that low amounts of HF (when present in the electrolyte sample) escape from the PTFE tube, react with the glass tube and water permeates back into the electrolyte, resulting in a reaction cycle which causes slow but steady electrolyte degradation (Fig. 3). In order to prove this assumption, an experimental set-up using molecular sieve was designed. Similar to the above mentioned samples the PFTE tubes containing the electrolyte were stored in NMR glass tubes without direct contact between the electrolyte and the glass. In addition, molecular sieve was placed inside the glass but outside the PFTE tubes to be able to adsorb water and to interrupt the reaction cycle without being in contact and thus not interfering with the electrolyte. Fig. 4 shows concentration curves of degraded LiPF6 of samples with and reference samples without molecular sieve. Single determinations for qualitative discussions have been performed. The reference samples exhibit clearly observable degradation. In particular, the reference sample stored at 80 C ages at a high rate and reaches a concentration of degraded LiPF6 of approximately 20mmolL1 after 70 days. In contrast to the reference set-up, the molecular sieve containing set-up results in only low degradation product concentrations. After reaching values above the limit of quantification the concentrations stay almost constant, clearly indicating the reaction cycle to be interrupted. Therefore, it can be concluded that LiPF6 is thermally stable in the absence of water at least up to temperatures of 80 C. It can be assumed that LiPF6 is thermally stable at even higher temperatures. Inevitably, there is an equilibrium between H2O vapor and H2O adsorbed by the molecular sieve. Higher temperatures push the equilibrium to the vapor side [28]. This suggests increasing degradation rates with increasing temperature, caused by a reduced drying ability of the molecular sieve. Though this experiment is not able to determine the limit of thermal stability of LiPF6, it confirms the proposed reaction cycle and identifies water as the main and probably even only source of LiPF6 degradation in the investigated temperature range. This confirms the assumption of a previous publication about a high thermal stability of LiPF6 based electrolytes [29]. However, the conclusions of this work are not in favor of all the

Fig. 4. Concentration curves of degraded LiPF6 of reference samples and samples containing molecular sieve (Mol. sieve) at 60 C and 80 C. Concentration value of reference sample after 70 days at 80 C: 20 mmol L 1 (not shown).

Fig. 5. Scheme depicting different kinds of influences of cell components on the thermal stability of LiPF6 in organic carbonates dependent on the reactivity against HF.

results of the above mentioned article that attributes thermal stability to the electrolyte even in the presence of water. This was investigated by adding 300ppm H2O to the electrolyte prior to thermal treatment, followed by infrared spectroscopical measurements. It should be noted that the amount of water is below the estimated limit of detection (1% by molar fraction). This means the observed stability of the electrolyte in the presence of water is due to an insufficient sensitivity of the applied measurement technique which explains the contradiction between this and the above mentioned article. 4. Conclusions

Fig. 3. HF H2O reaction cycle. HF and H2O permeate through the PTFE NMR tube liner causing slow but steady degradation reactions. X: position where molecular sieve breaks the reaction cycle.

Li metal and lithium ion battery cell components are able to influence the stability of LiPF6 based electrolytes. Type and extent of the influence depend on the reaction of HF with the respective cell component (Fig. 5). When the material is inert against HF (as in the case of graphite, carbon black, PVdF, polyolefinic and ceramic separators), no influence on the electrolyte stability is observed.

Please cite this article in press as: S. Wiemers-Meyer, et al., Influence of Battery Cell Components and Water on the Thermal and Chemical Stability of LiPF6 Based Lithium Ion Battery Electrolytes, Electrochim. Acta (2016), http://dx.doi.org/10.1016/j.electacta.2016.11.100

G Model EA 28390 No. of Pages 5

S. Wiemers-Meyer et al. / Electrochimica Acta xxx (2016) xxx–xxx

If the reaction with HF or the potentially resulting formation of water is slow (as for Li metal and NCM), there is also no significant influence observable. However, when the reaction with HF is fast and leads to the formation of water (as for glass fiber separator, oxidized Si nanopowder and LiOH containing LFP), LiPF6 can exhibit significant degradation rates. In contrast to that, if the reaction with HF is fast, but does not produce water (as with CMC), the cell component material stabilizes LiPF6 against water sources. Furthermore, a molecular sieve containing set-up proves that in the absence of water or water sources, LiPF6 is stable at least at temperatures up to 80 C and most likely also at even higher temperatures. There is obviously no strong or even no influence of elevated temperatures, when no water is present. Dry electrolytes should not produce HF at these elevated temperatures. The degradation products found in the samples are typical hydrolysis products of LiPF6, which indicates that the monitored reactions are caused by chemical degradation. Therefore, the results of this study can be considered as starting point for further investigations about electrochemical degradation reactions of the electrolyte in a battery. Acknowledgements The authors would like to thank the German Federal Ministry of Education and Research (BMBF) for financial support. This work was conducted within the context of the project “Elektrolytlabor4e” (03  4632). References [1] D. Andre, S.J. Kim, P. Lamp, S.F. Lux, F. Maglia, O. Paschos, B. Stiaszny, J. Mater Chem A 3 (2015) 6709–6732. [2] R. Fong, U. von Sacken, J.R. Dahn, J. Electrochem. Soc. 137 (1990) 2009–2013. [3] M.S. Whittingham, Chem. Rev. 104 (2004) 4271–4302.

5

[4] M. Wakihara, Mater. Sci. Eng. R-Rep. 33 (2001) 109–134. [5] J.B. Goodenough, Y. Kim, Chem. Mater. 22 (2009) 587–603. [6] R. Wagner, N. Preschitschek, S. Passerini, J. Leker, M. Winter, J Appl Electrochem 43 (2013) 481–496. [7] T. Ohzuku, R.J. Brodd, J. Power Sources 174 (2007) 449–456. [8] M. Hu, X. Pang, Z. Zhou, J. Power Sources 237 (2013) 229–242. [9] M. Grutzke, V. Kraft, B. Hoffmann, S. Klamor, J. Diekmann, A. Kwade, M. Winter, S. Nowak, J. Power Sources 273 (2015) 83–88. [10] V. Kraft, M. Grützke, W. Weber, J. Menzel, S. Wiemers-Meyer, M. Winter, S. Nowak, J. Chrom. A 1409 (2015) 201–209. [11] C.L. Campion, W. Li, W.B. Euler, B.L. Lucht, B. Ravdel, J.F. DiCarlo, R. Gitzendanner, K.M. Abraham, Electrochem. Solid State Lett. 7 (2004) A194–A197. [12] C.L. Campion, W. Li, B.L. Lucht, J. Electrochem. Soc. 152 (2005) A2327–A2334. [13] B. Ravdel, K.M. Abraham, R. Gitzendanner, J. DiCarlo, B. Lucht, C. Campion, J. Power Sources 119–121 (2003) 805–810. [14] S. Wiemers-Meyer, M. Winter, S. Nowak, Phys. Chem. Chem. Phys 18 (2016) 26595–26601. [15] J.O. Besenhard, M. Winter, J. Yang, W. Biberacher, J. Power Sources 54 (1995) 228–231. [16] V. Kraft, W. Weber, B. Streipert, R. Wagner, C. Schultz, M. Winter, S. Nowak, RSC Adv. 6 (2016) 8–17. [17] S.E. Sloop, J.K. Pugh, S. Wang, J.B. Kerr, K. Kinoshita, Electrochem. Solid State Lett. 4 (2001) A42–A44. [18] V. Kraft, W. Weber, M. Grutzke, M. Winter, S. Nowak, RSC Adv. 5 (2015) 80150–80157. [19] V. Kraft, M. Grutzke, W. Weber, M. Winter, S. Nowak, J. Chrom. A 1354 (2014) 92–100. [20] A.V. Plakhotnyk, L. Ernst, R. Schmutzler, J. Fluor. Chem. 126 (2005) 27–31. [21] C. Dippel, S. Krueger, R. Kloepsch, P. Niehoff, B. Hoffmann, S. Nowak, S. Passerini, M. Winter, J. Li, Electrochim. Acta 85 (2012) 66–71. [22] T. Placke, V. Siozios, R. Schmitz, S.F. Lux, P. Bieker, C. Colle, H.W. Meyer, S. Passerini, M. Winter, J. Power Sources 200 (2012) 83–91. [23] J.-A. Choi, S.H. Kim, D.-W. Kim, J. Power Sources 195 (2010) 6192–6196. [24] Y.S. Hu, R. Demir-Cakan, M.M. Titirici, J.O. Muller, R. Schlogl, M. Antonietti, J. Maier, Angew. Chem. Int. Edit. 47 (2008) 1645–1649. [25] L. Su, Z. Zhou, M. Ren, Chem. Commun. 46 (2010) 2590–2592. [26] M.T. McDowell, S.W. Lee, I. Ryu, H. Wu, W.D. Nix, J.W. Choi, Y. Cui, Nano Lett. 11 (2011) 4018–4025. [27] S.F. Lux, F. Schappacher, A. Balducci, S. Passerini, M. Winter, J. Electrochem. Soc. 157 (2010) A320–A325. [28] Y. Wang, M.D. LeVan, J. Chem. Eng. Data 54 (2009) 2839–2844. [29] H. Yang, G.V. Zhuang, P.N. Ross Jr., J. Power Sources 161 (2006) 573–579.

Please cite this article in press as: S. Wiemers-Meyer, et al., Influence of Battery Cell Components and Water on the Thermal and Chemical Stability of LiPF6 Based Lithium Ion Battery Electrolytes, Electrochim. Acta (2016), http://dx.doi.org/10.1016/j.electacta.2016.11.100