Solid State Ionics 272 (2015) 24–29
Contents lists available at ScienceDirect
Solid State Ionics journal homepage: www.elsevier.com/locate/ssi
Electrochemical performance of Si/CeO2/Polyaniline composites as anode materials for lithium ion batteries Ying Bai a,b, Yang Tang a, Zhihui Wang b, Zhe Jia b, Feng Wu a, Chuan Wu a,⁎, Gao Liu b,⁎ a b
School of Chemical Engineering & Environment, Beijing Institute of Technology, Beijing 100081, China Environmental Energy Technologies Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, United States
a r t i c l e
i n f o
Article history: Received 27 November 2014 Received in revised form 19 December 2014 Accepted 25 December 2014 Available online 14 January 2015 Keywords: Lithium ion batteries Anode material Silicon Composite Electrochemical performance Lithium ion diffusion coefﬁcient
a b s t r a c t Si has very high theoretical speciﬁc capacity as an anode material in a lithium ion battery. However, its application is seriously restricted because of relatively undesirable conductivity and poor cycling stability. Here we report Si/CeO2/Polyaniline (SCP) composite as an anode material, which was synthesized by hydrothermal reaction and chemical polymerization. The structures and morphologies of the SCP composites are characterized by X-ray diffraction (XRD), scanning electronic microscopy (SEM) and transmission electron microscopy (TEM). It is shown that Si/CeO2 (SC) particles are well coated by PANI elastomer which has good electrical conductivity. The SCP shows larger reversible capacity and better cycling performance compared with pure Si. The ﬁrst reason is that CeO2 can protect Si from reacting with electrolyte. More importantly, the PANI elastomer can accommodate the volume change of the composite during Li-alloying/dealloying processes, so the pulverization of silicon would be signiﬁcantly reduced. The SCP material can retain a capacity nearly 775 mAh/g after 100 cycles, while pure Si only shows a capacity of 370 mAh/g after 100 cycles. © 2014 Elsevier B.V. All rights reserved.
1 . Introduction Lithium ion batteries are being widely used in the world now, and graphite is a common commercial anode material because it's cheap and exhibits perfect cycling performance. However, the theoretical speciﬁc capacity of graphite is only about 370 mAh/g, which limits the improvement of LIBs . Therefore, it's necessary to ﬁnd new anode materials with high speciﬁc capacities. Si has drawn much attention as an anode material for LIBs because of high theoretical speciﬁc capacity (nearly 4200 mAh/g). But its application is seriously restricted due to relatively undesirable conductivity and poor cycling stability which is attributed to signiﬁcant volume change during lithiation and delithiation processes . To overcome the problems, many researchers adopt lots of methods such as Si-C [3–12], Si-metal [13–20] and Si with special structure (core–shell [21–23], porous [24,25], nanotube , nanowire , and ﬁlm ). In Si-C and Si-metal composites abovementioned, the C and metal can enhance the conductivity and buffer the mechanical stress caused by volume change of Si as a matrix; and the special structure also provides enough space to accommodate the expansion of Si. Here in this work, we report a series of Si/CeO2/Polyaniline (SCP) composites as anode materials for LIBs which is synthesized through ⁎ Corresponding authors. E-mail addresses: [email protected]
(C. Wu), [email protected]
http://dx.doi.org/10.1016/j.ssi.2014.12.016 0167-2738/© 2014 Elsevier B.V. All rights reserved.
hydrothermal reaction and chemical polymerization. As we know, because of its good stability, CeO2 has been used for cathode material coating to protect the active material from the corrosion of the electrolyte . For anode materials, such as silicon, the side reaction with the electrolyte is more serious. Therefore, it is a way worth trying to modify pure silicon by CeO2. Furthermore, due to the direct and fast transformation between Ce (III) and Ce (IV), CeO2 has good electrical conductivity , which is superior to other oxides. It is expected that CeO2 can enable improving the coulombic efﬁciency as a protective layer between the Si and electrolyte. In addition, PANI elastomer will be used to buffer the mechanical stress caused by Si, as well as improve the conductivity. 2 . Experimental 2.1. Sample preparation The SCP material was synthesized through hydrothermal reaction and chemical polymerization. Cerium nitrate (Ce (NO3)3·6H20), crystalline silicon nanopowder (Si, 20–50 nm, 97%, Strem Chemicals) and cetyl trimethyl ammonium bromide (CTAB) were dissolved in a mixture of distilled water and alcohol (volume ratio of 1:1). The mixture was transferred into a stainless steel autoclave, and heated at 180 °C for 12 h, then cooled to room temperature. The product was centrifuged, washed (with distilled water and alcohol) and dried under vacuum at 60 °C for 20 h to get Si/CeO2 (SC). Five SC composites were synthesized with
Y. Bai et al. / Solid State Ionics 272 (2015) 24–29
different mole ratios of Ce. The mole ratios of Ce were 0%, 0.1%, 0.5%, 1% and 2%, which were marked as SC00, SC01, SC05, SC10 and SC20. After SC particles were prepared, the chemical polymerization was carried out to prepare SCP. 1.24 g dodecyl benzenesulfonic acid (ABS) was added into 50 mL distilled water and stirred for 4 h to form a homogeneous microemulsion. Then 0.24 g aniline was mixed with the homogeneous microemulsion by stirring for 4 h. Afterwards, 0.56 g SC particles were added into the mixture, which was then treated with ultrasonication for 1 h. The mixture was subsequently stirred for 30 min, and then 0.56 g ammonium persulphate (APS, 0.5 M) was added into it. After stirring for 24 h, the mixture was then centrifuged, washed (with distilled water) and dried in vacuum at 60 °C for 24 h. An emerald powder of SCP was then obtained, where the mass fraction of PANI is 30%. Five SCP composites were prepared with different mole ratios of Ce in SC. The mole ratios of Ce were 0%, 0.1%, 0.5%, 1% and 2%, which were marked as SCP00, SCP01, SCP05, SCP10 and SCP20. 2.2. Material characterizations and electrochemical experiments Fig. 1. XRD patterns of PANI and SCP.
The structures and morphologies of the materials were analyzed by X-ray diffraction (XRD, Rigaku Ultima IV-185), scanning electronic microscopy (SEM, Quanta FEG 250) and transmission electron microscopy (TEM, FEI Tecnai G2 F30). The electrochemical performance was tested using coin cells. The work electrode was prepared by mixing the active material, super-p and poly (vinylidene ﬂuoride) (PVDF) binder (N-methyl-2-pyrrolidone as solvent) with a mass ratio of 8:1:1 to form homogeneous slurry. The slurry was coated on a copper foil with thickness of 200 μm, and then dried at 80 °C for 12 h. The loading of the active material on the electrode is 1.68 mg/cm2. Coin cell was assembled in a glove box full of argon, while the counter electrode was lithium metal and the electrolyte was 1 M LiPF6 solution in a mixture of ethylene carbonate (EC), ethyl methyl carbonate (EMC) and diethyl carbonate (DEC) (volume ratio of 1:1:1). And the membrane was Celgard 2400 polypropylene membrane. The cells were galvanostatically charged and discharged at 100 mA/g in the voltage range of 0 ~ 2 V (vs. Li+/Li) (CT2001A Land instrument). All the speciﬁc capacities in this paper are calculated based on the mass of silicon. Cyclic voltammetry (CV, CHI660d) was carried out under a sweep speed of 0.1 mV/s between 0.01 V and 2 V (vs. Li+/Li). Electrochemical impedance spectroscopy (EIS, CHI660d) was tested in a frequency range of 0.01 Hz and 100 kHz with amplitude of 5 mV. All measurements were carried out at room temperature. 3 . Results and discussion 3.1 . Physical characterization The XRD patterns of SCP and PANI materials are shown in Fig. 1. The peaks near 2Theta = 28, 47, 56, 69, 76 and 88 indicate the crystalline structure of Si, which doesn't change after the addition of CeO2 and PANI, as no other peaks are observed. It also can be seen that PANI has a wide peak near 2Theta = 22, and the peak also exists in other plots. It indicates that the SCP materials actually contain PANI. As shown in Table 1, the lattice constant and unit cell volume of SCP increase a little with increasing content of Ce. It means more space for lithiation and delithiation which can improve the diffusion of lithium ion . Fig. 2 shows the SEM images of Si (a), SC05 (b) and SCP05 (c). The added CeO2 and PANI slightly affect the particles' morphologies. The average diameters of them are all around 50–60 nm, where the diameter of SC particles (60 nm) is larger than that of the Si particles (50 nm), while slight agglomeration can be observed because of the high stress in reaction. The boundary among particles becomes indistinct after PANI coating (Fig. 2c), because the PANI layer covers the surface of the particles so that the particles are connected by the PANI layer.
Fig. 3 shows the TEM images of Si (a, f), SC05 (b, d, g) and SCP05 (c, e, h). It can be seen that the nanoparticles are distributed well (Fig. 3a, b, c, with the same magniﬁcation of 80,000) and the diameters of most particles are about 50 nm. While we observe that some CeO2 particles adhere to the Si particles' surface (Fig. 3d) and the PANI coating on the surface of SC particle can be easily distinguished (Fig. 3e, the layer between two arrows). After the addition of CeO2, the interplanar spacing of (220) plane increases from 1.9141 Å to 1.9202 Å (Fig. 3f, g), which indicates that the lattice constant also increases, while the interplanar spacing of the (220) plane is kept at 1.9202 Å (Fig. 3h) with the coverage of PANI (3.57 nm), and the lattice constant doesn't change. 3.2 . Electrochemical performance Fig. 4a shows the cycling performance of SCP materials with different cerium contents and pure Si. The initial capacities of SCP00, SCP01, SCP05, SCP10 SCP20 and Si are 2783, 2357, 1679, 2105, 1783 and 3646 mAh/g, respectively. After 10 cycles, the capacities decrease to 700, 824, 895, 774, 607 and 487 mAh/g, which indicate the existence of volume change upon cycling as well as side reaction of silicon with the electrolyte. However, it is notable that the SCP samples with PANI have much lower capacity fading than pure Si in the ﬁrst ten cycles, which is attributed to the elastic PANI conducting network that can ﬁrmly grasp the Si particles, thus suppress the pulverization of particles, and result in better cycling performances. After the volume changes during lithiation and delithiation processes in the ﬁrst ten cycles, the capacities become relatively stable. At the 20th cycle, the capacities of SCP00, SCP01, SCP05, SCP10, SCP20 and Si are 649, 741, 838, 744, 574 and 412 mAh/g, respectively. SCP00, SCP01, SCP05, SCP10, SCP20 and Si retain the capacities of 595, 675, 775, 683, 517 and 369 mAh/g after 100 cycles. With increasing molar ratio of cerium to 0.5%, sample SCP05 shows the best cycling performance, namely, the capacity still retains nearly 775 mAh/g after 100 cycles, which is higher than those of other SCP samples and Si (369.4 mAh/g). The phenomenon may be attributed to the fact that CeO2 can prevent the SCP material from directly
Table 1 The lattice constant and unit cell volume of SCP. SCP
Lattice constant (Å)
Unit cell volume(Å)3
SCP00 SCP01 SCP05 SCP10 SCP20
5.42053 5.43301 5.43599 5.43707 5.43821
159.27 160.37 160.63 160.73 160.83
Y. Bai et al. / Solid State Ionics 272 (2015) 24–29
Fig. 2. SEM images of Si (a), SC05 (b) and SCP05 (c).
contacting with the electrolyte as a protective layer to restrain the reactions between them, so the cycling reversibility is promoted . As a result, the coulombic efﬁciency of SC05 (74.6%) is higher than that of SC00 (70.9%). However, when the molar ratio of cerium is further increased to 1–2%, the reversible capacity after the 10th cycle begins to descend, as CeO2 is not electrochemistry-active and the reversible capacity of SCP will dramatically decrease with excessive CeO2 . Therefore, the SCP05 displays larger reversible capacities and better cycling performance compared with other SCP materials. The effect of PANI on SCP is shown in Fig. 4b. Without PANI compared with SCP, all the SC samples exhibit lower discharge capacity than the SCP samples, because the huge volume change cannot be well inhibited so that pulverization of particles and destruction of structure occur. According to a previous report , PANI is like elastomer which effectively buffers the huge volume change of Si during lithiation and delithiation processes via providing enough room for the expansion. At the same time, the PANI also improves the conductivities of SCP composites here. In consideration of the important effect of PANI, then we'll focus on the electrochemical performance of Si and SCP samples.
Fig. 5 shows the apparent morphologies of the typical Si-based samples after 100 cycles. A serious aggregation of pure Si is clearly observed, where the silicon particles are connected with each other to form large blocks, and the edge of each particle becomes misty, as shown in Fig. 5a, which has very great changes when compared with Fig. 2a. And sample SC05 seems more dispersible than Si (Fig. 5b). With PANI addition, the aggregation of SCP05 is well inhibited, especially compared with Si (Fig. 5c). Distinctly, the small particle sizes of sample SC05 and SCP05 are kept well after the long-term cycling, which are in favor of the diffusion of Li+, and result in superior electrochemical performances to pure Si. Fig. 6a shows the initial discharge–charge curves of Si, SCP00 and SCP05 at the current density of 100 mA/g. During the ﬁrst discharge process (lithiation), there is a long discharge plateau between 0 and 0.1 V with Li corresponding to crystalline silicon forming LixSi alloy. During the charge process, there is an inclined plateau between 0.4 and 0.5 V, while the LixSi alloy changes to Li and amorphous silicon. It can be seen that the proﬁles of SCP00 and SCP05 are similar to that of Si, indicating that the existence of CeO2 and PANI does not change the lithiation and delithiation behaviors. The initial discharge capacities of Si, SCP00 and SCP05 are 3646, 2783 and 1679 mAh/g, while their initial charge
Fig. 3. TEM images of Si (a, f), SC05 (b, d, g) and SCP05 (c, e, h).
Y. Bai et al. / Solid State Ionics 272 (2015) 24–29
Fig. 4. Cycling performances of (a) SCP, Si and (b) SC.
capacities are 2453, 2135 and 1471 mAh/g, respectively. The irreversible capacity is caused by the formation of SEI ﬁlm and lithium ion caught in the crystal. The coulombic efﬁciencies of the PANI modiﬁed samples are markedly improved with increasing Ce content, namely, 76.7% for SCP00 and 87.6% for SCP05, which are much superior to that of pure Si, 67.3%. It could be attributed to PANI buffering the volume expansion
to prevent the pulverization of Si, and CeO2 suppressing the side reactions between Si and electrolyte, which are both beneﬁcial to improving the reversible capacity. The cyclic voltammograms of different samples are shown in Fig. 6(b)–(d). The cyclic voltammograms show no obvious changes after adding CeO2 and PANI, which means only Si serves as the active anode material. It's clear that there are two oxidation (delithiation) peaks at 0.37 V and 0.55 V in anodic curves, and two reduction (lithiation) peaks at 0.01 V and 0.12 V in cathodic curves. The wide peak between 0.5 V and 0.8 V in cathodic curve is observed in the ﬁrst cycle and disappears in the following cycles. It's attributed to the formation of solid electrolyte interface (SEI) ﬁlm due to the reaction of between composite and electrolyte on the surface of electrode. The crystalline silicon forms alloy with Li in the ﬁrst cathodic process, and turns to amorphous silicon after delithiation in the ﬁrst anodic process. In the following cathodic processes, amorphous silicon forms alloy with Li, which are different from the ﬁrst one, so the peak at 0.12 V doesn't exist in the curve of ﬁrst cycle [35,36]. Fig. 7 shows the cycling performance of samples at different current densities. For the SCP05 electrode, the discharge capacities are 836, 747, 714, 681 mAh/g and 662 mAh/g at the current densities of 100, 200, 300, 400 mA/g and 500 mA/g, respectively, and recovers to 718 mAh/g as the current density back to 100 mA/g again, namely, nearly 91.3% of the discharge capacity can be retained; while the capacity retention of sample SCP00 and pure Si are 88.4% and 83.2%, respectively. The results demonstrate that sample SCP05 exhibits better rate performance. That's because PANI could enhance the conductivity and keep the structure stable, which is beneﬁcial to maintaining the transmission channel of lithium ion steady, so the Li-alloying/dealloying processes will be faster and more full. And the existence of CeO2 also expands the transmission channel of lithium ion with increasing lattice constant. Fig. 8a shows the EIS analysis of the Si, SCP00 and SCP05 samples. The equivalent circuit is shown in Fig. 8b. Rs is the electrolyte resistance, Q1 and Q2 are constant phase elements, Rct is the charge-transfer resistance, W is the Warburg impedance and Rf is the resistance of SEI ﬁlm. The addition of PANI and CeO2 actually decreases the charge-transfer resistance with the smaller high frequency semicircle and medium frequency region, which is beneﬁcial to the Li-alloying/dealloying processes. The diffusion coefﬁcient of lithium ion can be calculated according to the following equation: 2 2 2 4 4 2 2 D ¼ R T = 2A n F C σ
In the equation, R is ideal gas constant, T is absolute temperature, n is the stoichiometric number of electron participating in the reaction, A is the area of the surface of electrode, F is Faraday constant, C is the molar
Fig. 5. SEM images of Si (a), SC05 (b) and SCP05 (c) after 100 cycles.
Y. Bai et al. / Solid State Ionics 272 (2015) 24–29
Fig. 6. The initial discharge–charge curves of (a) Si; SCP00; SCP05 and the cyclic voltammograms of (b) Si; (c) SCP00; (d) SCP05.
concentration of Li+, D is the diffusion coefﬁcient of lithium ion, and σ is the Warburg coefﬁcient which conforms to the equation as follows: 0
Z ¼ σω
þ Rs þ Rct
where Rs is the resistance of electrolyte, Rct is charge-transfer resistance and ω is the angular frequency in the low frequency region. As shown in Fig. 8c (ﬁtted using ZSimpwin 3.20 software), the Warburg coefﬁcients of Si, SCP00 and SCP05 are 100.7, 61.1 and 46.5 Ω·s−1/2. So the diffusion coefﬁcients of lithium ion of Si, SCP00 and SCP05 could be calculated to
be 4.3 × 10−17, 2.0 × 10−16 and 9.6 × 10−16 cm2/s. The increase of diffusion coefﬁcient should be attributed to the stable structure and expanded transmission channel of lithium ion, caused by PANI and CeO2. So the SCP05 sample has better rate performance than Si and SCP00.
4 . Conclusions The Si/CeO2/Polyaniline (SCP) composites have been synthesized through hydrothermal reaction and chemical polymerization. Some CeO2 particles distribute on the surface of Si as protective layer. The SCP nanoparticles with an average diameter of 50 nm are well coated by the PANI elastomer with excellent conductivity. The PANI layer of about 3.6 nm could be easily recognized and the SCP particles are connected with each other by the layer with good elasticity. SCP05 shows best cycling performance: its initial capacity can reach 1678.6 mAh/g with an initial coulombic efﬁciency of 87.6% and still retains 774.8 mAh/g after 100 cycles, 46.2% of the initial capacity kept; while the SCP00 and Si only keep 594.8 and 369.4 mAh/g after 100 cycles. The good cycling performance of SCP05 can be attributed to the CeO2 protecting the SCP material from reacting with the electrolyte, as well as improving the conductivity. Furthermore, the PANI elastomer can accommodate the volume change of the composite during Li-alloying/ dealloying processes, so the pulverization of silicon can be well suppressed.
Fig. 7. The cycling performance of Si, SCP00 and SCP05 at different current densities.
This work is supported by the National Basic Research Program of China (Contract No. 2015CB251100), the Program for New Century Excellent Talents in University (Contract NCET-13-0033), and Natural Science Foundation of China (Contract No. 21476027). Y. Bai acknowledges
Y. Bai et al. / Solid State Ionics 272 (2015) 24–29
b Fig. 8. (a) EIS analysis of Si, SCP00 and SCP05; (b) the equivalent circuit; (c) the relationship between Z′ and ω−1/2 at low frequency region of Si, SCP00 and SCP05.
the support from the State Scholarship Fund (No.201406035025) of the China Scholarship Council. References  N.S. Choi, Y. Yao, Y. Cui, J. Cho, J. Mater. Chem. 21 (2011) 9825.  J. Qu, H.Q. Li, John J. Henry Jr., S.K. Martha, N.J. Dudney, H.B. Xu, M.F. Chi, M.J. Lance, S.M. Mahurin, T.M. Besmann, S. Dai, J. Power Sources 198 (2012) 312–317.  S. Yang, G.R. Li, Q. Zhu, Q.M. Pan, J. Mater. Chem. 22 (2012) 3420.  J.Y. Luo, X. Zhao, J.S. Wu, H.D. Jang, H.H. Kung, J.X. Huang, J. Phys. Chem. Lett. 3 (2012) 1824–1829.  H.C. Tao, L.Z. Fan, Y.F. Mei, X.H. Qu, Electrochem. Commun. 13 (2011) 1332–1335.  J.H. Yao, Z.T. Jia, P.J. Zhang, C.Q. Shen, J.B. Wang, K.-F.A. Zinsou, L.B. Wang, Ionics 19 (2012) 401–407.  R.C. de Guzman, J.H. Yang, M.M.-C. Cheng, S.O. Salley, K.Y. Simon Ng, J. Mater. Sci. 48 (2013) 4823–4833.  B. Wang, X.L. Li, X.F. Zhang, B. Luo, M.H. Jin, M.H. Liang, S.A. Dayeh, S.T. Picraux, L.J. Zhi, ACS Nano 7 (2013) 1437–1445.  G.Y. Zhao, L. Zhang, Y.F. Meng, N.Q. Zhang, K.N. Sun, J. Power Sources 240 (2013) 212–218.  D. Lu, Y. Xiao, X.H. Yan, Y.R. Yang, Chem. Phys. Lett. 515 (2011) 263–268.  Y.H. Zhu, W. Liu, X.Y. Zhang, J.C. He, J.T. Chen, Y.P. Wang, T.B. Cao, Langmuir 29 (2013) 744–749.  X. Zhao, C.M. Hayner, M.C. Kung, H.H. Kung, Adv. Energy Mater. 1 (2011) 1079–1084.  Z. Edfouf, C.F. Georges, F. Cuevas, M. Latroche, T. Hézèque, G. Caillon, C. Jordy, M.T. Sougrati, J.C. Jumas, Electrochim. Acta 89 (2013) 365–371.  Z. Edfouf, F. Cuevas, M. Latroche, C. Georges, C. Jordy, T. Hézèque, G. Caillon, J.C. Jumas, M.T. Sougrati, J. Power Sources 196 (2011) 4762–4768.  Y.M. Liu, B.L. Chen, F. Cao, H.L.W. Chan, X.Z. Zhao, J.K. Yuan, J. Mater. Chem. 21 (2011) 17083.  D.Y. Chen, X. Mei, G. Ji, M.H. Lu, J.P. Xie, J.M. Lu, J.Y. Lee, Angew. Chem. Int. Ed. 51 (2012) 2409–2413.
 Y. Yu, L. Gu, C.B. Zhu, S. Tsukimoto, P.A. van Aken, J. Maier, Adv. Mater. 22 (2010) 2247–2250.  X.D. Wu, Z.X. Wang, L.Q. Chen, X.J. Huang, Electrochem. Commun. 5 (2003) 935–939.  E. Kwon, H.-S. Lim, Y.-K. Sun, K.-D. Suh, Solid State Ionics 237 (2013) 28–33.  Z.W. Lua, G. Wang, X.P. Gao, X.J. Liu, J.Q. Wang, J. Power Sources 189 (2009) 832–836.  Y. Chen, J.F. Qian, Y.L. Cao, H.X. Yang, X.P. Ai, ACS Appl. Mater. Interfaces 4 (2012) 3753–3758.  Y. Li, G.J. Xu, Y.F. Yao, L.G. Xue, M. Yanilmaz, H. Lee, X.W. Zhang, Solid State Ionics 258 (2014) 67–73.  S. Sim, P. Oh, S. Park, J. Cho, Adv. Mater. 25 (2013) 4498–4503.  H. Cheng, R. Xiao, H.D. Bian, Z. Li, Y.W. Zhan, C.K. Tsang, C.Y. Chung, Z.G. Lu, Y.Y. Li, Mater. Chem. Phys. 144 (2014) 25–30.  Z.L. Zhang, Y.H. Wang, W.F. Ren, Q.Q. Tan, Y.F. Chen, H. Li, Z.Y. Zhong, F.B. Su, Angew. Chem. Int. Ed. 53 (2014) 5165–5169.  J.P. Rong, X. Fang, M.Y. Ge, H.T. Chen, J. Xu, Nano Res. 6 (2013) 182–190.  X.X. Zhao, X.H. Rui, W.W. Zhou, L.P. Tan, Q.Y. Yan, Z.Y. Lu, H.H. Hng, J. Power Sources 250 (2014) 160–165.  H.X. Li, H.M. Bai, Z.L. Tao, J. Chen, J. Power Sources 217 (2012) 102–107.  H.-W. Ha, N.J. Yun, K. Kim, Electrochim. Acta 52 (2007) 3236–3241.  X.J. Yang, Y.D. Huang, X.C. Wang, D.Z. Jia, W.K. Pang, Z.P. Guo, X.C. Tang, J. Power Sources 257 (2014) 280–285.  Q.F. Wang, Y. Huang, J. Miao, Y. Zhao, Y. Wang, Electrochim. Acta 93 (2013) 120–130.  S.B. Xia, Y.J. Zhang, P. Dong, R.M. Yang, Y.N. Zhang, Chin. J. Inorg. Chem. 30 (2014) 529–535.  D. Arumugam, G.P. Kalaignan, Electrochim. Acta 55 (2010) 8709–8716.  M. Chen, C.Y. Du, L. Wang, G.P. Yin, P.F. Shi, Int. J. Electrochem. Sci. 7 (2012) 819–829.  M.K. Datta, P.N. Kumta, J. Power Sources 165 (2007) 368–378.  X.H. Hou, S.J. Hu, Q. Ru, Z.W. Zhang, Rare Metal Mater. Eng. 39 (2010) 2079–2083.