- Email: [email protected]
Effect of zwitterions on electrochemical properties of oligoether-based electrolytes
Electrochimica Acta 175 (2015) 209–213
Contents lists available at ScienceDirect
Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta
Effect of zwitterions on electrochemical properties of oligoether-based electrolytes Mitsutake Suematsu a , Masahiro Yoshizawa-Fujita a, * , Haijin Zhu b , Maria Forsyth b , Yuko Takeoka a , Masahiro Rikukawa a a b
Department of Materials and Life Sciences, Sophia University, 7-1 Kioi-cho, Chiyoda-ku, Tokyo 102-8554, Japan Institute for Frontier Materials and the ARC Centre of Excellence for Electromaterials Science, Deakin University, Geelong, VIC 3216, Australia
A R T I C L E I N F O
A B S T R A C T
Article history: Received 27 October 2014 Received in revised form 7 December 2014 Accepted 11 December 2014 Available online 12 December 2014
Solid polymer electrolytes show great potential in electrochemical devices. Poly(ethylene oxide) (PEO) has been studied as a matrix for solid polymer electrolytes because it has relatively high ionic conductivity. In order to investigate the effect of zwitterions on the electrochemical properties of poly (ethylene glycol) dimethyl ether (G5)/lithium bis(fluorosulfonyl) amide (LiFSA) electrolytes, a liquid zwitterion (ImZ2) was added to the G5-based electrolytes. In this study, G5, which is a small oligomer, was used as a model compound for PEO matrices. The thermal properties, ionic conductivity, and electrochemical stability of the electrolytes with ImZ2 were evaluated. The thermal stabilities of all the G5-based electrolytes with ImZ2 were above 150 C, and the ionic conductivity values were in the range of 0.8–3.0 mS cm1 at room temperature. When the electrolytes contained less than 5.5 wt% ImZ2, the ionic conductivity values were almost the same as that of the electrolyte without ImZ2. The electrochemical properties were improved with the incorporation of ImZ2. The anodic limit of the electrolyte with 5.5 wt% ImZ2 was 5.3 V vs. Li/Li+, which was over 1 V higher than that of G5/LiFSA. ã 2014 Elsevier Ltd. All rights reserved.
Keywords: Polymer electrolyte Zwitterion Oligoether Electrochemical window Lithium-ion battery
1. Introduction Polymer electrolytes have been investigated for their potential application in electrochemical devices, particularly in secondary lithium-ion batteries. Solid polymer electrolytes have several advantages, such as no leakage of electrolytes and no internal shorting. Poly(ethylene oxide) (PEO)/lithium salt complexes have been extensively studied as polymer electrolytes [1–7] because PEO has a high polarity, which promotes the dissociation of lithium salts, and high mobility, which can promote the transport of dissociated ions. However, the conductivity of these complexes is mainly due to the migration of anionic species; the lithium-ion transference number (tLi+) is generally low, with values below 0.4. These low values may result in a concentration polarization that limits the rate (power) of batteries [8]. In addition, the anodic limit of pristine PEO is about 4.0 V. It is difficult to use electrodes with high charge voltages, such as LiCoO2 and LiNi1/3Mn1/3Co1/3O2 (NMC), as cathode materials [9,10]. In order to improve the electrochemical stability of these materials, many kinds of additives have been investigated. For example, PEO-based
* Corresponding author. Tel.: +81 3 3238 3498; fax: +81 3 3238 3361. E-mail address: [email protected] (M. Yoshizawa-Fujita). http://dx.doi.org/10.1016/j.electacta.2014.12.067 0013-4686/ ã 2014 Elsevier Ltd. All rights reserved.
electrolytes exhibited an anodic stability of up to 4.5 V with addition of low molecular weight poly(ethylene glycol) dimethyl ether (PEGDME) [11], ceramic fillers [12,13], or PEG-borate ester [14]. Although the electrochemical stability of PEO-based electrolytes was improved with these additives, the ionic conductivity values were below 104 S cm1 at room temperature. It is still difficult to achieve high ionic conductivity and good electrochemical stability simultaneously because the additives cause an increase in the viscosity of matrices. It has been reported that the addition of zwitterions promotes the dissociation of lithium salts and facilitates the transport of lithium ions in ionic-liquid-based electrolytes and polymer electrolytes [15]. In addition to enhancing the conduction of lithium ions, battery cycle performances were improved by the formation of a solid electrolyte interface (SEI) film on the surface of the electrodes [16–18]. We previously reported that an imidazolium-based zwitterion containing two oxyethylene units (ImZ2) can be synthesized as a colorless liquid at room temperature [19], and that lithium bis(fluorosulfonyl) amide (LiFSA) mixed with a small amount of the liquid zwitterion exhibited high ionic conductivity and tLi+ values [20]. In order to improve the lithium transference number and oxidation resistance at around 4.0 V for PEO, ImZ2 was added to poly(ethylene glycol) dimethyl ether (G5)/LiFSA electrolytes. In
210
M. Suematsu et al. / Electrochimica Acta 175 (2015) 209–213
2. Experimental 2.1. Materials 3-(1-(2-Methoxyethoxyethyl)-1H-imidazol-3-ium-3-yl) propane-1-sulfonate (ImZ2) was synthesized according to the published procedure [19]. LiFSA (Piotrek, 99.0%) was used as received. Poly(ethylene glycol) dimethyl ether (G5) (Aldrich, Mn = 250) was dried in vacuo before use. The chemical structures of G5, LiFSA, and ImZ2 are shown in Fig. 1. G5 was mixed with LiFSA in an argon filled glove box. The molar ratio of ethylene oxide (EO) units to Li+ was fixed at 20 for the G5/LiFSA electrolytes [21]. ImZ2 was then added to the G5/LiFSA electrolytes, and the resulting mixtures were stirred at 60 C for 24 h to obtain G5/LiFSA/ImZ2(x), where x represents the content of ImZ2 in the electrolyte (x = 1.5, 5.5, 10, or 20 wt%). The water content in the G5/LiFSA/ImZ2(x) electrolytes was measured to be below 30 ppm using a Karl-Fisher moisture meter (756 KF Coulometer, Metrohm Co.). 2.2. Methods Thermogravimetric analysis (TGA) was carried out on a TG-DTA instrument (TG-DTA7200, Hitachi High-Technologies Corp.) in a nitrogen gas atmosphere between 25 and 500 C at a heating rate of 10 C min1. The thermal behavior was studied by differential scanning calorimetry (DSC) (DSC7020, Hitachi High-Technologies Corp.) between -150 and 80 C at a heating/cooling rate of 10 C min1. Density measurements were conducted using an Anton Paar DMA 38 density meter. Viscosity measurements were conducted using an Anton Paar AMVn automated microviscometer. Ionic conductivity measurements were carried out using a locally designed cell made from two platinum electrodes. The cell constants were calculated with a 0.01 mol L1 KCl aqueous solution. The ionic conductivity values were obtained by measuring the complex impedance between 5 and 1 MHz using an impedance analyzer (1260, Solatron) over the temperature range of -30 to 80 C. The temperature was controlled using a constant temperature oven (SU-262, Espec Corp.). Pulsed-field gradient NMR (PFG-NMR) diffusion measurements were performed on a Bruker 300 MHz UltraShield spectrometer with an Avance III console utilizing a Diff50 diffusion probe. The various constituents of the samples were measured using the 1H nuclei for G5 and ImZ2, 19F nuclei for the FSA anion, and 7Li nuclei for the lithium cation. The experiments were performed using the stimulated spin-echo (STE) pulse sequence by varying the gradient in eight steps up to the maximum gradient strength [22]. This data was fit to the Stejskal-Tanner diffusion Eq. (1) [23] using the nonlinear least-squares method. Samples were packed and sealed in 5 mm NMR tubes in an argon gas atmosphere. All measurements were conducted at 25 C. A ¼ A0 exp½ðgdGÞ2 ðD d=3ÞD
strength, D is the delay between the two gradients, d is the duration of one of the gradient pulses, and D is the diffusion coefficient. For each measurement, the gradient pulse duration and diffusion time were typically 5 ms and 20 ms, respectively, and the gradient strength (G = 600–1400 G cm1) was optimized according to the diffusion coefficients. The electrochemical stability of the electrolytes was evaluated by linear sweep voltammetry (LSV) at a sweep rate of 10 mV s1 in the potential range of 0–6 V vs. Li/Li+ at 25 C. A stainless steel cell (TYS-00DM01, Toyo System Co., Ltd.) was used as the LSV measurement cell. A platinum plate polished with 1.0, 0.3, and 0.05 mm Al2O3 powder was used as the working electrode. A battery-grade lithium foil was used as the counter electrode. The measurements were performed using a potentiostat/galvanostat (HZ-7000, Hokuto Denko Co.). The lithium plating/stripping performance of the Ni/electrolyte/Li cells was characterized using cyclic voltammetry (CV). A stainless steel cell (TYS-00DM01, Toyo System Co., Ltd.) was used as the CV measurement cell. A nickel plate polished with 1.0, 0.3, and 0.05 mm Al2O3 powder was used as the working electrode. A battery-grade lithium foil was used as the counter electrode. A separator film (Advantec GA-55) was soaked in the electrolytes before cell assembly. The cells were fabricated in an argon filled glove box. The CV measurements (scan rate: 10 mV s1) were performed between -0.25 and 1.0 V vs. Li/Li+ using an electrochemical system (CH Instruments potentiostat/galvanostat, Model 660B). 3. Results and discussion 3.1. Thermal properties TGA measurements were carried out for G5/LiFSA and G5/ LiFSA/ImZ2(x). Watanabe and co-workers reported that complexes of glyme and lithium bis(trifluoromethylsulfonyl) amide (LiTFSA) are thermally stable up to 200 C [24]. Although the onset thermal decomposition temperature (Td) values of G5/LiFSA and G5/LiFSA/ ImZ2(x) were about 150 C, G5/LiFSA and G5/LiFSA/ImZ2(x) exhibited sufficient thermal stability (> 100 C) for use as electrolytes in lithium-ion batteries. These lower Td values could be due to low thermal stability of LiFSA compared with that of LiTFSA [20]. The DSC curves of the electrolytes are presented in Fig. 2. The G5/ LiFSA and G5/LiFSA/ImZ2(x) electrolytes exhibited low glass transition temperatures (Tg) below -90 C. In all the electrolytes
x = 1.5
Endothermic
this study, we investigated the effects of ImZ2 on the ionic conductivity and electrochemical stability of G5/LiFSA electrolytes.
x = 5.5 x = 10
(1)
x = 20
where A0 is the signal area without a gradient, g is the gyromagnetic ratio of the nuclei under study, G is the gradient
G5/LiFSA -120 -90
-60
-30
0
30
60
Temperature / ºC Fig. 1. Chemical structures of G5, LiFSA, and ImZ2. Fig. 2. DSC curves of G5/LiFSA and G5/LiFSA/ImZ2(x).
M. Suematsu et al. / Electrochimica Acta 175 (2015) 209–213
investigated in this study, the melting and crystallization behavior of pure G5 was not observed, indicating that LiFSA and ImZ2 were soluble in the G5 matrix. An exothermic peak was observed around -15 C when the content of ImZ2 was 5.5 or 10 wt%. This observation was attributed to the crystallization of G5 and ImZ2 because the G5/LiFSA and ImZ2/LiFSA [20] composites exhibited no thermal event at the same temperature. As the enthalpy change of the exothermic peak was quite low, this thermal event did not affect the conductivity of the G5/LiFSA/ImZ2 composites (vide infra). 3.2. Ionic conductivities Fig. 3 shows the Arrhenius plots of the ionic conductivities for G5/LiFSA and G5/LiFSA/ImZ2(x). Convex upward curves were observed for all the electrolytes used in this study. The ion conduction process in these materials should be expressed by the Vogel-Fulcher-Tammann (VFT) Eq. (2) [25–27], which describes the temperature dependence of viscosity in amorphous materials.
s ¼ s 0 expð
B Þ T T0
(2)
The parameters s 0 and B are related to the number of carrier ions and the activation energy of ion conduction, respectively, and T0 is believed to be closely related to the glass transition temperature. The best-fit parameters are listed in Table 1. The ionic conductivity value of G5/LiFSA was 3.0 mS cm1 at 25 C. When the electrolytes contained less than 5.5 wt% ImZ2, the ionic conductivity values and VFT parameters were almost the same as those of G5/LiFSA; however, the ionic conductivity values of G5/ LiFSA/ImZ2 gradually decreased with increasing ImZ2 content. This decrease in ionic conductivity could be related to an increase in viscosity. The viscosity value of G5/LiFSA was 18 mPa s at 25 C, whereas the viscosity values of G5/LiFSA with 1.5 and 5.5 wt% ImZ2 were 20 and 24 mPa s, respectively. When the ImZ2 content was greater than 10 wt%, the viscosity values increased further; the viscosity values of the electrolytes with 10 and 20 wt% ImZ2 were 29 and 52 mPa s, respectively.
211
Table 1 VFT parameters for G5/LiFSA and G5/LiFSA/ImZ2(x)
G5/LiFSA/ImZ2(1.5) G5/LiFSA/ImZ2(5.5) G5/LiFSA/ImZ2(10) G5/LiFSA/ImZ2(20) G5/LiFSA
s o/S cm1
B/K
To/K
R2
0.17 0.16 0.14 0.09 0.16
548 535 581 575 525
170 172 172 175 170
0.99 0.99 0.99 0.99 0.99
3.3. Self-diffusion coefficients Fig. 4 shows the self-diffusion coefficients of G5/LiFSA/ImZ2(x) at 25 C as a function of the ImZ2 content. The self-diffusion coefficients of each component, G5 (DG5), Li+ (DLi), FSA (DFSA), and ImZ2 (DZw) in G5/LiFSA and G5/LiFSA/ImZ2(x) were measured by PFG-NMR measurements. The diffusion coefficient values of each component monotonically decreased with increasing amounts of ImZ2. The DG5 and DFSA values in the electrolytes were almost the same at each ImZ2 concentration. It has previously been reported by Watanabe and co-workers that the DG and DLi values were almost the same in complexes of glyme and LiTFSA [24]. G5/LiFSA/ ImZ2 showed the same behavior with conventional organic solutions, with the diffusivity showing the following order: solvent anion > Li+ [28]. Table 2 summarizes the molar conductivity (Limp) values derived from the conductivity values measured using the ACimpedance technique, the molar conductivity (LNE) [28] values derived from the Nernst-Einstein equation using the diffusion coefficient values measured by PFG-NMR, the Li+ transference number (tLi+) derived from Eq. (3), and the ionicity (Limp/LNE) [29] at 25 C. The ionic conductivities of G5/LiFSA and G5/LiFSA/ImZ2(x) were higher than those of complexes of glyme and LiTFSA [24]. It is known that ionic liquids (ILs) incorporating FSA anions exhibit higher ionic conductivities and lower viscosities than those with TFSA anions [30–34]. The same tendency was observed in G5based electrolytes. 3.4. Lithium-ion transference number The tLi+ values of G5/LiFSA/ImZ2(x) were calculated using Eq. (3).
Fig. 3. Arrhenius plots of ionic conductivity for G5/LiFSA and G5/LiFSA/ImZ2(x).
Fig. 4. Self-diffusion coefficients of G5/LiFSA and G5/LiFSA/ImZ2(x) as a function of ImZ2 content.
212
M. Suematsu et al. / Electrochimica Acta 175 (2015) 209–213
Table 2 Molar conductivity from AC-impedance experiments (Limp), from PFG-NMR experiments (LNE), Li+ transference number (tLi+), and Limp/LNE for G5/LiFSA and G5/LiFSA/ImZ2(x) at 25 C
G5/LiFSA/ImZ2(1.5) G5/LiFSA/ImZ2(5.5) G5/LiFSA/ImZ2(10) G5/LiFSA/ImZ2(20) G5/LiFSA
Limp/S cm2 mol1
tLi+
LNE/S cm2 mol1
Limp/LNE
2.13 1.85 1.21 0.79 2.33
0.38 0.38 0.34 0.25 0.38
3.27 2.74 2.27 1.46 3.55
0.65 0.67 0.53 0.50 0.65
tLiþ ¼ DLi =ðDLi þ DFSA Þ
(3)
The tLi+ value of G5/LiFSA was 0.38, which is a typical value in PEO electrolytes because PEO strongly interacts with cations through dipole-ion interactions. It should be noted that the selfdiffusion coefficients measured using the PFG-NMR method are derived not only from ionic species but also from non-ionic species (i.e., ion pairs), indicating that these tLi+ values would be different from those determined by electrochemical techniques. The tLi+ values were maintained up to 5.5 wt% ImZ2 and then decreased with increasing the ImZ2 content. It was noted that on addition of ImZ2, the DLi values were similar to the DZw values. These results suggested that the interaction between Li+ and ImZ2 could be stronger than that of Li+ and G5. The ionicity values of G5/LiFSA, G5/LiFSA/ImZ2(1.5), and G5/LiFSA/ImZ2(5.5) were over 0.6, which is comparable to those of ILs that exhibit high ionic conductivity. The ionicity value represents the proportion of ions that contribute to ionic conduction in the electrolytes. The number of carrier ions in these systems should be comparable to that in ILs. 3.5. Electrochemical stability on Pt electrode Fig. 5 shows the linear sweep voltammograms of the electrolytes incorporating various amounts of ImZ2 at a scan rate of 10 mV s1 at room temperature using Pt as the working electrode and lithium metal as the reference and counter electrodes. The cathodic and anodic limits of G5/LiFSA were observed at 0 and 3.8 V vs. Li/Li+, respectively. The anodic limit was consistent with the reported value of 4.0 V [9]. The anodic stability of G5 was poor even after addition of LiFSA. While the cathodic stabilities of the G5 electrolytes were almost the same before and after addition of
Current density / mA cm-2
5
3 2 1 0 -1 -2 -3
0
1
2
3
ImZ2, G5/LiFSA/ImZ2(1.5) and G5/LiFSA/ImZ2(5.5) exhibited anodic limits of 4.1 and 5.3 V vs. Li/Li+, respectively. These results could be attributed to the formation of an passivation layer on the surface of the electrode [17,35] or a decrease in the HOMO energy level [36] owing to an interaction between G5 and ImZ2. As a result, the oxidation reaction of G5 was prevented by adding ImZ2. 3.6. Li plating/stripping on Ni electrode Fig. 6 shows the cyclic voltammogram of G5/LiFSA/ImZ2(5.5). The lithium plating/stripping reaction in these electrolytes was examined at room temperature. The peak current density of lithium stripping for G5/LiFSA/ImZ2(5.5) maintained almost the same value as that of the first cycle, even after 50 cycles. This result could be explained by the generation of a passivation film on the Ni electrode surface during the first cycle, which inhibited the reduction of G5/LiFSA/ImZ2(5.5) [18,37]. Stable lithium plating/ stripping cycling was achieved by adding a small amount of ImZ2 to G5/LiFSA electrolytes. 4. Conclusion G5-based electrolytes (G5/LiFSA and G5/LiFSA/ImZ2(x)) were prepared and the effects of ImZ2 on the thermal and electrochemical properties were investigated. When the added amount of ImZ2 was below 5.5 wt%, the ionic conductivity values of the G5/LiFSA electrolytes were maintained above 103 S cm1 at room temperature. The electrochemical stability of the G5-based electrolytes was improved by the incorporation of ImZ2. G5/LiFSA/ImZ2(5.5) showed an electrochemical window in the potential range of 0– 5.3 V vs. Li/Li+. Stable lithium plating/stripping behavior in G5 was observed in cyclic voltammetry experiments. Thus, ImZ2 was shown to be effective at improving the electrochemical properties of G5/LiFSA electrolytes. G5/LiFSA/ImZ2 will perform as an electrolyte in lithium-ion batteries with LiCoO2 and NMC electrodes as cathode materials. These investigations are now under progress and will be reported elsewhere. The liquid zwitterion represents an interesting additive for electrochemical applications.
x = 5.5 x = 1.5 G5/LiFSA
4
Fig. 6. Cyclic voltammogram of the 5th cycle for G5/LiFSA/ImZ2(5.5) on Ni at 25 C and 10 mV s1. Inset: 10th, 20th, 30th, 40th, and 50th cycles for G5/LiFSA/ImZ2(5.5).
4
5
6
+
Potential vs. (Li/Li ) / V Fig. 5. Liner sweep voltammograms of G5/LiFSA, G5/LiFSA/ImZ2(1.5), and G5/ LiFSA/ImZ2(5.5) at 25 C on Pt at 10 mV s1.
Acknowledgement This study was supported by a Grant-in-Aid for Scientific Research (C) (No. 26410140) from Japan Society for the Promotion of Science (JSPS).
M. Suematsu et al. / Electrochimica Acta 175 (2015) 209–213
References [1] D. Fauteux, A. Massucco, M. McLin, M. Van Buren, J. Shi, Electrochim. Acta 40 (1995) 2185. [2] M. Gauthier, A. Bélanger, P. Bouchard, B. Kapfer, S. Ricard, G. Vassort, M. Armand, J.Y. Sanchez, L. Krause, J. Power Sources 54 (1995) 163. [3] F.M. Gray, Polymer Electrolytes (RSC Materials Monographs), Royal Society of Chemistry, 1997. [4] K. Murata, S. Izuchi, Y. Yoshihisa, Electrochim. Acta 45 (2000) 1501. [5] B. Scrosati, F. Croce, S. Panero, J. Power Sources 100 (2001) 93. [6] V. Di Noto, S. Lavina, G.A. Giffin, E. Negro, B. Scrosati, Electrochim. Acta 57 (2011) 4. [7] H. Zhang, C. Liu, L. Zheng, F. Xu, W. Feng, H. Li, X. Huang, M. Armand, J. Nie, Z. Zhou, Electrochim. Acta 133 (2014) 529. [8] P.G. Bruce, B. Scrosati, J.-M. Tarascon, Angew. Chem. Int. Ed. 47 (2008) 2930. [9] Y. Xia, T. Fujieda, K. Tatsumi, P.P. Prosini, T. Sakai, J. Power Sources 92 (2001) 234. [10] P.P. Prosini, S. Passerini, Eur. Polym. J. 37 (2001) 65. [11] Y. Kang, H.J. Kim, E. Kim, B. Oh, J.H. Cho, J. Power Sources 92 (2001) 255. [12] J.-H. Shin, S. Passerini, Electrochim. Acta 49 (2004) 1605. [13] Q. Li, N. Imanishi, A. Hirano, Y. Takeda, O. Yamamoto, J. Power Sources 110 (2002) 38. [14] Y. Kato, K. Hasumi, S. Yokoyama, T. Yabe, H. Ikuta, Y. Uchimoto, M. Wakihara, Solid State Ionics 150 (2002) 355. [15] C. Tiyapiboonchaiya, J.M. Pringle, J. Sun, N. Byrne, P.C. Howlett, D.R. MacFarlane, M. Forsyth, Nat. Mater. 3 (2004) 29. [16] D.Q. Nguyen, J. Hwang, J.S. Lee, H. Kim, H. Lee, M. Cheong, B. Lee, H.S. Kim, Electrochem. Commun. 9 (2007) 109. [17] D.Q. Nguyen, H.W. Bae, E.H. Jeon, J.S. Lee, M. Cheong, H. Kim, H.S. Kim, H. Lee, J. Power Sources 183 (2008) 303. [18] Z.H. Li, Q.L. Xia, L.L. Liu, G.T. Lei, Q.Z. Xiao, D.S. Gao, X.D. Zhou, Electrochim. Acta 56 (2010) 804.
213
[19] M. Yoshizawa-Fujita, T. Tamura, Y. Takeoka, M. Rikukawa, Chem. Commun. 47 (2011) 2345. [20] M. Suematsu, M. Yoshizawa-Fujita, T. Tamura, Y. Takeoka, M. Rikukawa, Int. J. Electrochem. Sci. 10 (2015) 248. [21] S. Lascaud, M. Perrier, A. Vallee, S. Besner, J. Prud'homme, M. Armand, Macromolecules 27 (1994) 7469. [22] P.M. Bayley, A.S. Best, D.R. MacFarlane, M. Forsyth, ChemPhysChem 12 (2011) 823. [23] E.O. Stejskal, J.E. Tanner, J. Chem. Phys. 42 (1965) 288. [24] T. Tamura, K. Yoshida, T. Hachida, M. Tsuchiya, M. Nakamura, Y. Kazue, N. Tachikawa, K. Dokko, M. Watanabe, Chem. Lett. 39 (2010) 753. [25] H. Vogel, Phys. Z. 22 (1921) 645. [26] G.S. Fulcher, J. Am. Ceram. Soc. 8 (1925) 339. [27] G. Tamman, W. Hesse, Z. Anorg. Allg. Chem. 156 (1926) 245. [28] K. Hayamizu, Y. Aihara, S. Arai, C. Garcia Martinez, J. Phys. Chem. B 103 (1999) 519. [29] K. Ueno, H. Tokuda, M. Watanabe, Phys. Chem. Chem. Phys. 12 (2010) 1649. [30] P. Johansson, L.E. Fast, A. Matic, G.B. Appetecchi, S. Passerini, J. Power Sources 195 (2010) 2074. [31] H. Matsumoto, H. Sakaebe, K. Tatsumi, M. Kikuta, E. Ishiko, M. Kono, J. Power Sources 160 (2006) 1308. [32] S. Seki, Y. Kobayashi, H. Miyashiro, Y. Ohno, Y. Mita, N. Terada, P. Charest, A. Guerfi, K. Zaghib, J. Phys. Chem. C 112 (2008) 16708. [33] M. Biso, M. Mastragostino, M. Montanino, S. Passerini, F. Soavi, Electrochim. Acta 53 (2008) 7967. [34] G.B. Appetecchi, M. Montanino, A. Balducci, S.F. Lux, M. Winter, S. Passerini, J. Power Sources 192 (2009) 599. [35] B. Li, M. Xu, T. Li, W. Li, S. Hu, Electrochem. Commun. 17 (2012) 92. [36] K. Yoshida, M. Nakamura, Y. Kazue, N. Tachikawa, S. Tsuzuki, S. Seki, K. Dokko, M. Watanabe, J. Am. Chem. Soc. 133 (2011) 13121. [37] N. Byrne, P.C. Howlett, D.R. MacFarlane, M. Forsyth, Adv. Mater. 17 (2005) 2497.
Contents lists available at ScienceDirect
Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta
Effect of zwitterions on electrochemical properties of oligoether-based electrolytes Mitsutake Suematsu a , Masahiro Yoshizawa-Fujita a, * , Haijin Zhu b , Maria Forsyth b , Yuko Takeoka a , Masahiro Rikukawa a a b
Department of Materials and Life Sciences, Sophia University, 7-1 Kioi-cho, Chiyoda-ku, Tokyo 102-8554, Japan Institute for Frontier Materials and the ARC Centre of Excellence for Electromaterials Science, Deakin University, Geelong, VIC 3216, Australia
A R T I C L E I N F O
A B S T R A C T
Article history: Received 27 October 2014 Received in revised form 7 December 2014 Accepted 11 December 2014 Available online 12 December 2014
Solid polymer electrolytes show great potential in electrochemical devices. Poly(ethylene oxide) (PEO) has been studied as a matrix for solid polymer electrolytes because it has relatively high ionic conductivity. In order to investigate the effect of zwitterions on the electrochemical properties of poly (ethylene glycol) dimethyl ether (G5)/lithium bis(fluorosulfonyl) amide (LiFSA) electrolytes, a liquid zwitterion (ImZ2) was added to the G5-based electrolytes. In this study, G5, which is a small oligomer, was used as a model compound for PEO matrices. The thermal properties, ionic conductivity, and electrochemical stability of the electrolytes with ImZ2 were evaluated. The thermal stabilities of all the G5-based electrolytes with ImZ2 were above 150 C, and the ionic conductivity values were in the range of 0.8–3.0 mS cm1 at room temperature. When the electrolytes contained less than 5.5 wt% ImZ2, the ionic conductivity values were almost the same as that of the electrolyte without ImZ2. The electrochemical properties were improved with the incorporation of ImZ2. The anodic limit of the electrolyte with 5.5 wt% ImZ2 was 5.3 V vs. Li/Li+, which was over 1 V higher than that of G5/LiFSA. ã 2014 Elsevier Ltd. All rights reserved.
Keywords: Polymer electrolyte Zwitterion Oligoether Electrochemical window Lithium-ion battery
1. Introduction Polymer electrolytes have been investigated for their potential application in electrochemical devices, particularly in secondary lithium-ion batteries. Solid polymer electrolytes have several advantages, such as no leakage of electrolytes and no internal shorting. Poly(ethylene oxide) (PEO)/lithium salt complexes have been extensively studied as polymer electrolytes [1–7] because PEO has a high polarity, which promotes the dissociation of lithium salts, and high mobility, which can promote the transport of dissociated ions. However, the conductivity of these complexes is mainly due to the migration of anionic species; the lithium-ion transference number (tLi+) is generally low, with values below 0.4. These low values may result in a concentration polarization that limits the rate (power) of batteries [8]. In addition, the anodic limit of pristine PEO is about 4.0 V. It is difficult to use electrodes with high charge voltages, such as LiCoO2 and LiNi1/3Mn1/3Co1/3O2 (NMC), as cathode materials [9,10]. In order to improve the electrochemical stability of these materials, many kinds of additives have been investigated. For example, PEO-based
* Corresponding author. Tel.: +81 3 3238 3498; fax: +81 3 3238 3361. E-mail address: [email protected] (M. Yoshizawa-Fujita). http://dx.doi.org/10.1016/j.electacta.2014.12.067 0013-4686/ ã 2014 Elsevier Ltd. All rights reserved.
electrolytes exhibited an anodic stability of up to 4.5 V with addition of low molecular weight poly(ethylene glycol) dimethyl ether (PEGDME) [11], ceramic fillers [12,13], or PEG-borate ester [14]. Although the electrochemical stability of PEO-based electrolytes was improved with these additives, the ionic conductivity values were below 104 S cm1 at room temperature. It is still difficult to achieve high ionic conductivity and good electrochemical stability simultaneously because the additives cause an increase in the viscosity of matrices. It has been reported that the addition of zwitterions promotes the dissociation of lithium salts and facilitates the transport of lithium ions in ionic-liquid-based electrolytes and polymer electrolytes [15]. In addition to enhancing the conduction of lithium ions, battery cycle performances were improved by the formation of a solid electrolyte interface (SEI) film on the surface of the electrodes [16–18]. We previously reported that an imidazolium-based zwitterion containing two oxyethylene units (ImZ2) can be synthesized as a colorless liquid at room temperature [19], and that lithium bis(fluorosulfonyl) amide (LiFSA) mixed with a small amount of the liquid zwitterion exhibited high ionic conductivity and tLi+ values [20]. In order to improve the lithium transference number and oxidation resistance at around 4.0 V for PEO, ImZ2 was added to poly(ethylene glycol) dimethyl ether (G5)/LiFSA electrolytes. In
210
M. Suematsu et al. / Electrochimica Acta 175 (2015) 209–213
2. Experimental 2.1. Materials 3-(1-(2-Methoxyethoxyethyl)-1H-imidazol-3-ium-3-yl) propane-1-sulfonate (ImZ2) was synthesized according to the published procedure [19]. LiFSA (Piotrek, 99.0%) was used as received. Poly(ethylene glycol) dimethyl ether (G5) (Aldrich, Mn = 250) was dried in vacuo before use. The chemical structures of G5, LiFSA, and ImZ2 are shown in Fig. 1. G5 was mixed with LiFSA in an argon filled glove box. The molar ratio of ethylene oxide (EO) units to Li+ was fixed at 20 for the G5/LiFSA electrolytes [21]. ImZ2 was then added to the G5/LiFSA electrolytes, and the resulting mixtures were stirred at 60 C for 24 h to obtain G5/LiFSA/ImZ2(x), where x represents the content of ImZ2 in the electrolyte (x = 1.5, 5.5, 10, or 20 wt%). The water content in the G5/LiFSA/ImZ2(x) electrolytes was measured to be below 30 ppm using a Karl-Fisher moisture meter (756 KF Coulometer, Metrohm Co.). 2.2. Methods Thermogravimetric analysis (TGA) was carried out on a TG-DTA instrument (TG-DTA7200, Hitachi High-Technologies Corp.) in a nitrogen gas atmosphere between 25 and 500 C at a heating rate of 10 C min1. The thermal behavior was studied by differential scanning calorimetry (DSC) (DSC7020, Hitachi High-Technologies Corp.) between -150 and 80 C at a heating/cooling rate of 10 C min1. Density measurements were conducted using an Anton Paar DMA 38 density meter. Viscosity measurements were conducted using an Anton Paar AMVn automated microviscometer. Ionic conductivity measurements were carried out using a locally designed cell made from two platinum electrodes. The cell constants were calculated with a 0.01 mol L1 KCl aqueous solution. The ionic conductivity values were obtained by measuring the complex impedance between 5 and 1 MHz using an impedance analyzer (1260, Solatron) over the temperature range of -30 to 80 C. The temperature was controlled using a constant temperature oven (SU-262, Espec Corp.). Pulsed-field gradient NMR (PFG-NMR) diffusion measurements were performed on a Bruker 300 MHz UltraShield spectrometer with an Avance III console utilizing a Diff50 diffusion probe. The various constituents of the samples were measured using the 1H nuclei for G5 and ImZ2, 19F nuclei for the FSA anion, and 7Li nuclei for the lithium cation. The experiments were performed using the stimulated spin-echo (STE) pulse sequence by varying the gradient in eight steps up to the maximum gradient strength [22]. This data was fit to the Stejskal-Tanner diffusion Eq. (1) [23] using the nonlinear least-squares method. Samples were packed and sealed in 5 mm NMR tubes in an argon gas atmosphere. All measurements were conducted at 25 C. A ¼ A0 exp½ðgdGÞ2 ðD d=3ÞD
strength, D is the delay between the two gradients, d is the duration of one of the gradient pulses, and D is the diffusion coefficient. For each measurement, the gradient pulse duration and diffusion time were typically 5 ms and 20 ms, respectively, and the gradient strength (G = 600–1400 G cm1) was optimized according to the diffusion coefficients. The electrochemical stability of the electrolytes was evaluated by linear sweep voltammetry (LSV) at a sweep rate of 10 mV s1 in the potential range of 0–6 V vs. Li/Li+ at 25 C. A stainless steel cell (TYS-00DM01, Toyo System Co., Ltd.) was used as the LSV measurement cell. A platinum plate polished with 1.0, 0.3, and 0.05 mm Al2O3 powder was used as the working electrode. A battery-grade lithium foil was used as the counter electrode. The measurements were performed using a potentiostat/galvanostat (HZ-7000, Hokuto Denko Co.). The lithium plating/stripping performance of the Ni/electrolyte/Li cells was characterized using cyclic voltammetry (CV). A stainless steel cell (TYS-00DM01, Toyo System Co., Ltd.) was used as the CV measurement cell. A nickel plate polished with 1.0, 0.3, and 0.05 mm Al2O3 powder was used as the working electrode. A battery-grade lithium foil was used as the counter electrode. A separator film (Advantec GA-55) was soaked in the electrolytes before cell assembly. The cells were fabricated in an argon filled glove box. The CV measurements (scan rate: 10 mV s1) were performed between -0.25 and 1.0 V vs. Li/Li+ using an electrochemical system (CH Instruments potentiostat/galvanostat, Model 660B). 3. Results and discussion 3.1. Thermal properties TGA measurements were carried out for G5/LiFSA and G5/ LiFSA/ImZ2(x). Watanabe and co-workers reported that complexes of glyme and lithium bis(trifluoromethylsulfonyl) amide (LiTFSA) are thermally stable up to 200 C [24]. Although the onset thermal decomposition temperature (Td) values of G5/LiFSA and G5/LiFSA/ ImZ2(x) were about 150 C, G5/LiFSA and G5/LiFSA/ImZ2(x) exhibited sufficient thermal stability (> 100 C) for use as electrolytes in lithium-ion batteries. These lower Td values could be due to low thermal stability of LiFSA compared with that of LiTFSA [20]. The DSC curves of the electrolytes are presented in Fig. 2. The G5/ LiFSA and G5/LiFSA/ImZ2(x) electrolytes exhibited low glass transition temperatures (Tg) below -90 C. In all the electrolytes
x = 1.5
Endothermic
this study, we investigated the effects of ImZ2 on the ionic conductivity and electrochemical stability of G5/LiFSA electrolytes.
x = 5.5 x = 10
(1)
x = 20
where A0 is the signal area without a gradient, g is the gyromagnetic ratio of the nuclei under study, G is the gradient
G5/LiFSA -120 -90
-60
-30
0
30
60
Temperature / ºC Fig. 1. Chemical structures of G5, LiFSA, and ImZ2. Fig. 2. DSC curves of G5/LiFSA and G5/LiFSA/ImZ2(x).
M. Suematsu et al. / Electrochimica Acta 175 (2015) 209–213
investigated in this study, the melting and crystallization behavior of pure G5 was not observed, indicating that LiFSA and ImZ2 were soluble in the G5 matrix. An exothermic peak was observed around -15 C when the content of ImZ2 was 5.5 or 10 wt%. This observation was attributed to the crystallization of G5 and ImZ2 because the G5/LiFSA and ImZ2/LiFSA [20] composites exhibited no thermal event at the same temperature. As the enthalpy change of the exothermic peak was quite low, this thermal event did not affect the conductivity of the G5/LiFSA/ImZ2 composites (vide infra). 3.2. Ionic conductivities Fig. 3 shows the Arrhenius plots of the ionic conductivities for G5/LiFSA and G5/LiFSA/ImZ2(x). Convex upward curves were observed for all the electrolytes used in this study. The ion conduction process in these materials should be expressed by the Vogel-Fulcher-Tammann (VFT) Eq. (2) [25–27], which describes the temperature dependence of viscosity in amorphous materials.
s ¼ s 0 expð
B Þ T T0
(2)
The parameters s 0 and B are related to the number of carrier ions and the activation energy of ion conduction, respectively, and T0 is believed to be closely related to the glass transition temperature. The best-fit parameters are listed in Table 1. The ionic conductivity value of G5/LiFSA was 3.0 mS cm1 at 25 C. When the electrolytes contained less than 5.5 wt% ImZ2, the ionic conductivity values and VFT parameters were almost the same as those of G5/LiFSA; however, the ionic conductivity values of G5/ LiFSA/ImZ2 gradually decreased with increasing ImZ2 content. This decrease in ionic conductivity could be related to an increase in viscosity. The viscosity value of G5/LiFSA was 18 mPa s at 25 C, whereas the viscosity values of G5/LiFSA with 1.5 and 5.5 wt% ImZ2 were 20 and 24 mPa s, respectively. When the ImZ2 content was greater than 10 wt%, the viscosity values increased further; the viscosity values of the electrolytes with 10 and 20 wt% ImZ2 were 29 and 52 mPa s, respectively.
211
Table 1 VFT parameters for G5/LiFSA and G5/LiFSA/ImZ2(x)
G5/LiFSA/ImZ2(1.5) G5/LiFSA/ImZ2(5.5) G5/LiFSA/ImZ2(10) G5/LiFSA/ImZ2(20) G5/LiFSA
s o/S cm1
B/K
To/K
R2
0.17 0.16 0.14 0.09 0.16
548 535 581 575 525
170 172 172 175 170
0.99 0.99 0.99 0.99 0.99
3.3. Self-diffusion coefficients Fig. 4 shows the self-diffusion coefficients of G5/LiFSA/ImZ2(x) at 25 C as a function of the ImZ2 content. The self-diffusion coefficients of each component, G5 (DG5), Li+ (DLi), FSA (DFSA), and ImZ2 (DZw) in G5/LiFSA and G5/LiFSA/ImZ2(x) were measured by PFG-NMR measurements. The diffusion coefficient values of each component monotonically decreased with increasing amounts of ImZ2. The DG5 and DFSA values in the electrolytes were almost the same at each ImZ2 concentration. It has previously been reported by Watanabe and co-workers that the DG and DLi values were almost the same in complexes of glyme and LiTFSA [24]. G5/LiFSA/ ImZ2 showed the same behavior with conventional organic solutions, with the diffusivity showing the following order: solvent anion > Li+ [28]. Table 2 summarizes the molar conductivity (Limp) values derived from the conductivity values measured using the ACimpedance technique, the molar conductivity (LNE) [28] values derived from the Nernst-Einstein equation using the diffusion coefficient values measured by PFG-NMR, the Li+ transference number (tLi+) derived from Eq. (3), and the ionicity (Limp/LNE) [29] at 25 C. The ionic conductivities of G5/LiFSA and G5/LiFSA/ImZ2(x) were higher than those of complexes of glyme and LiTFSA [24]. It is known that ionic liquids (ILs) incorporating FSA anions exhibit higher ionic conductivities and lower viscosities than those with TFSA anions [30–34]. The same tendency was observed in G5based electrolytes. 3.4. Lithium-ion transference number The tLi+ values of G5/LiFSA/ImZ2(x) were calculated using Eq. (3).
Fig. 3. Arrhenius plots of ionic conductivity for G5/LiFSA and G5/LiFSA/ImZ2(x).
Fig. 4. Self-diffusion coefficients of G5/LiFSA and G5/LiFSA/ImZ2(x) as a function of ImZ2 content.
212
M. Suematsu et al. / Electrochimica Acta 175 (2015) 209–213
Table 2 Molar conductivity from AC-impedance experiments (Limp), from PFG-NMR experiments (LNE), Li+ transference number (tLi+), and Limp/LNE for G5/LiFSA and G5/LiFSA/ImZ2(x) at 25 C
G5/LiFSA/ImZ2(1.5) G5/LiFSA/ImZ2(5.5) G5/LiFSA/ImZ2(10) G5/LiFSA/ImZ2(20) G5/LiFSA
Limp/S cm2 mol1
tLi+
LNE/S cm2 mol1
Limp/LNE
2.13 1.85 1.21 0.79 2.33
0.38 0.38 0.34 0.25 0.38
3.27 2.74 2.27 1.46 3.55
0.65 0.67 0.53 0.50 0.65
tLiþ ¼ DLi =ðDLi þ DFSA Þ
(3)
The tLi+ value of G5/LiFSA was 0.38, which is a typical value in PEO electrolytes because PEO strongly interacts with cations through dipole-ion interactions. It should be noted that the selfdiffusion coefficients measured using the PFG-NMR method are derived not only from ionic species but also from non-ionic species (i.e., ion pairs), indicating that these tLi+ values would be different from those determined by electrochemical techniques. The tLi+ values were maintained up to 5.5 wt% ImZ2 and then decreased with increasing the ImZ2 content. It was noted that on addition of ImZ2, the DLi values were similar to the DZw values. These results suggested that the interaction between Li+ and ImZ2 could be stronger than that of Li+ and G5. The ionicity values of G5/LiFSA, G5/LiFSA/ImZ2(1.5), and G5/LiFSA/ImZ2(5.5) were over 0.6, which is comparable to those of ILs that exhibit high ionic conductivity. The ionicity value represents the proportion of ions that contribute to ionic conduction in the electrolytes. The number of carrier ions in these systems should be comparable to that in ILs. 3.5. Electrochemical stability on Pt electrode Fig. 5 shows the linear sweep voltammograms of the electrolytes incorporating various amounts of ImZ2 at a scan rate of 10 mV s1 at room temperature using Pt as the working electrode and lithium metal as the reference and counter electrodes. The cathodic and anodic limits of G5/LiFSA were observed at 0 and 3.8 V vs. Li/Li+, respectively. The anodic limit was consistent with the reported value of 4.0 V [9]. The anodic stability of G5 was poor even after addition of LiFSA. While the cathodic stabilities of the G5 electrolytes were almost the same before and after addition of
Current density / mA cm-2
5
3 2 1 0 -1 -2 -3
0
1
2
3
ImZ2, G5/LiFSA/ImZ2(1.5) and G5/LiFSA/ImZ2(5.5) exhibited anodic limits of 4.1 and 5.3 V vs. Li/Li+, respectively. These results could be attributed to the formation of an passivation layer on the surface of the electrode [17,35] or a decrease in the HOMO energy level [36] owing to an interaction between G5 and ImZ2. As a result, the oxidation reaction of G5 was prevented by adding ImZ2. 3.6. Li plating/stripping on Ni electrode Fig. 6 shows the cyclic voltammogram of G5/LiFSA/ImZ2(5.5). The lithium plating/stripping reaction in these electrolytes was examined at room temperature. The peak current density of lithium stripping for G5/LiFSA/ImZ2(5.5) maintained almost the same value as that of the first cycle, even after 50 cycles. This result could be explained by the generation of a passivation film on the Ni electrode surface during the first cycle, which inhibited the reduction of G5/LiFSA/ImZ2(5.5) [18,37]. Stable lithium plating/ stripping cycling was achieved by adding a small amount of ImZ2 to G5/LiFSA electrolytes. 4. Conclusion G5-based electrolytes (G5/LiFSA and G5/LiFSA/ImZ2(x)) were prepared and the effects of ImZ2 on the thermal and electrochemical properties were investigated. When the added amount of ImZ2 was below 5.5 wt%, the ionic conductivity values of the G5/LiFSA electrolytes were maintained above 103 S cm1 at room temperature. The electrochemical stability of the G5-based electrolytes was improved by the incorporation of ImZ2. G5/LiFSA/ImZ2(5.5) showed an electrochemical window in the potential range of 0– 5.3 V vs. Li/Li+. Stable lithium plating/stripping behavior in G5 was observed in cyclic voltammetry experiments. Thus, ImZ2 was shown to be effective at improving the electrochemical properties of G5/LiFSA electrolytes. G5/LiFSA/ImZ2 will perform as an electrolyte in lithium-ion batteries with LiCoO2 and NMC electrodes as cathode materials. These investigations are now under progress and will be reported elsewhere. The liquid zwitterion represents an interesting additive for electrochemical applications.
x = 5.5 x = 1.5 G5/LiFSA
4
Fig. 6. Cyclic voltammogram of the 5th cycle for G5/LiFSA/ImZ2(5.5) on Ni at 25 C and 10 mV s1. Inset: 10th, 20th, 30th, 40th, and 50th cycles for G5/LiFSA/ImZ2(5.5).
4
5
6
+
Potential vs. (Li/Li ) / V Fig. 5. Liner sweep voltammograms of G5/LiFSA, G5/LiFSA/ImZ2(1.5), and G5/ LiFSA/ImZ2(5.5) at 25 C on Pt at 10 mV s1.
Acknowledgement This study was supported by a Grant-in-Aid for Scientific Research (C) (No. 26410140) from Japan Society for the Promotion of Science (JSPS).
M. Suematsu et al. / Electrochimica Acta 175 (2015) 209–213
References [1] D. Fauteux, A. Massucco, M. McLin, M. Van Buren, J. Shi, Electrochim. Acta 40 (1995) 2185. [2] M. Gauthier, A. Bélanger, P. Bouchard, B. Kapfer, S. Ricard, G. Vassort, M. Armand, J.Y. Sanchez, L. Krause, J. Power Sources 54 (1995) 163. [3] F.M. Gray, Polymer Electrolytes (RSC Materials Monographs), Royal Society of Chemistry, 1997. [4] K. Murata, S. Izuchi, Y. Yoshihisa, Electrochim. Acta 45 (2000) 1501. [5] B. Scrosati, F. Croce, S. Panero, J. Power Sources 100 (2001) 93. [6] V. Di Noto, S. Lavina, G.A. Giffin, E. Negro, B. Scrosati, Electrochim. Acta 57 (2011) 4. [7] H. Zhang, C. Liu, L. Zheng, F. Xu, W. Feng, H. Li, X. Huang, M. Armand, J. Nie, Z. Zhou, Electrochim. Acta 133 (2014) 529. [8] P.G. Bruce, B. Scrosati, J.-M. Tarascon, Angew. Chem. Int. Ed. 47 (2008) 2930. [9] Y. Xia, T. Fujieda, K. Tatsumi, P.P. Prosini, T. Sakai, J. Power Sources 92 (2001) 234. [10] P.P. Prosini, S. Passerini, Eur. Polym. J. 37 (2001) 65. [11] Y. Kang, H.J. Kim, E. Kim, B. Oh, J.H. Cho, J. Power Sources 92 (2001) 255. [12] J.-H. Shin, S. Passerini, Electrochim. Acta 49 (2004) 1605. [13] Q. Li, N. Imanishi, A. Hirano, Y. Takeda, O. Yamamoto, J. Power Sources 110 (2002) 38. [14] Y. Kato, K. Hasumi, S. Yokoyama, T. Yabe, H. Ikuta, Y. Uchimoto, M. Wakihara, Solid State Ionics 150 (2002) 355. [15] C. Tiyapiboonchaiya, J.M. Pringle, J. Sun, N. Byrne, P.C. Howlett, D.R. MacFarlane, M. Forsyth, Nat. Mater. 3 (2004) 29. [16] D.Q. Nguyen, J. Hwang, J.S. Lee, H. Kim, H. Lee, M. Cheong, B. Lee, H.S. Kim, Electrochem. Commun. 9 (2007) 109. [17] D.Q. Nguyen, H.W. Bae, E.H. Jeon, J.S. Lee, M. Cheong, H. Kim, H.S. Kim, H. Lee, J. Power Sources 183 (2008) 303. [18] Z.H. Li, Q.L. Xia, L.L. Liu, G.T. Lei, Q.Z. Xiao, D.S. Gao, X.D. Zhou, Electrochim. Acta 56 (2010) 804.
213
[19] M. Yoshizawa-Fujita, T. Tamura, Y. Takeoka, M. Rikukawa, Chem. Commun. 47 (2011) 2345. [20] M. Suematsu, M. Yoshizawa-Fujita, T. Tamura, Y. Takeoka, M. Rikukawa, Int. J. Electrochem. Sci. 10 (2015) 248. [21] S. Lascaud, M. Perrier, A. Vallee, S. Besner, J. Prud'homme, M. Armand, Macromolecules 27 (1994) 7469. [22] P.M. Bayley, A.S. Best, D.R. MacFarlane, M. Forsyth, ChemPhysChem 12 (2011) 823. [23] E.O. Stejskal, J.E. Tanner, J. Chem. Phys. 42 (1965) 288. [24] T. Tamura, K. Yoshida, T. Hachida, M. Tsuchiya, M. Nakamura, Y. Kazue, N. Tachikawa, K. Dokko, M. Watanabe, Chem. Lett. 39 (2010) 753. [25] H. Vogel, Phys. Z. 22 (1921) 645. [26] G.S. Fulcher, J. Am. Ceram. Soc. 8 (1925) 339. [27] G. Tamman, W. Hesse, Z. Anorg. Allg. Chem. 156 (1926) 245. [28] K. Hayamizu, Y. Aihara, S. Arai, C. Garcia Martinez, J. Phys. Chem. B 103 (1999) 519. [29] K. Ueno, H. Tokuda, M. Watanabe, Phys. Chem. Chem. Phys. 12 (2010) 1649. [30] P. Johansson, L.E. Fast, A. Matic, G.B. Appetecchi, S. Passerini, J. Power Sources 195 (2010) 2074. [31] H. Matsumoto, H. Sakaebe, K. Tatsumi, M. Kikuta, E. Ishiko, M. Kono, J. Power Sources 160 (2006) 1308. [32] S. Seki, Y. Kobayashi, H. Miyashiro, Y. Ohno, Y. Mita, N. Terada, P. Charest, A. Guerfi, K. Zaghib, J. Phys. Chem. C 112 (2008) 16708. [33] M. Biso, M. Mastragostino, M. Montanino, S. Passerini, F. Soavi, Electrochim. Acta 53 (2008) 7967. [34] G.B. Appetecchi, M. Montanino, A. Balducci, S.F. Lux, M. Winter, S. Passerini, J. Power Sources 192 (2009) 599. [35] B. Li, M. Xu, T. Li, W. Li, S. Hu, Electrochem. Commun. 17 (2012) 92. [36] K. Yoshida, M. Nakamura, Y. Kazue, N. Tachikawa, S. Tsuzuki, S. Seki, K. Dokko, M. Watanabe, J. Am. Chem. Soc. 133 (2011) 13121. [37] N. Byrne, P.C. Howlett, D.R. MacFarlane, M. Forsyth, Adv. Mater. 17 (2005) 2497.