- Email: [email protected]
graphite–tin for lithium-ion rechargeable batteries anode materials
Journal of Alloys and Compounds 596 (2014) 86–91
Contents lists available at ScienceDirect
Journal of Alloys and Compounds journal homepage: www.elsevier.com/locate/jalcom
A novel nano-structured interpenetrating phase composite of silicon/ graphite–tin for lithium-ion rechargeable batteries anode materials Jinbo Wu, Zhengwang Zhu, Hongwei Zhang, Huameng Fu, Hong Li, Aimin Wang, Haifeng Zhang ⇑, Zhuangqi Hu Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences, 72 Wenhua Road, Shenyang 110016, China
a r t i c l e
i n f o
Article history: Received 11 October 2013 Received in revised form 19 December 2013 Accepted 24 January 2014 Available online 4 February 2014 Keywords: Silicon/graphite–tin Lithium-ion batteries High energy mechanical milling Electrochemical performance Anode material Nano-structured interpenetrating phase composite
a b s t r a c t A novel nano-structured interpenetrating phase composite (NSIPC) of silicon/graphite–tin (SGM) anode material for lithium-ion rechargeable batteries is synthesized by high energy mechanical milling (HEMM). The structural and morphological characterizations have been carried out through X-ray diffraction (XRD), scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The electrochemical performances have been analyzed with reference to Li+/Li and the results are compared with silicon/graphite composites. The SGM NSIPC electrode exhibits the better cyclability than the SG composite electrodes. The initial discharge specific capacity of the as-prepared SGM NSIPC is relatively high around 1790 mA h g 1 with 1592 mA h g 1 reversible capacity retention in the following cycle at a current density of 237 mA g 1 in the voltage from 0.03 V to 1.5 V. In addition, the SGM NSIPC electrode shows the good rate capability and possesses the stable cycling performance even charging and discharging at the large current density. Consequently, SGM NSIPC can be the promising anode material for the next generation lithium ion rechargeable batteries. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Li-ion batteries (LIBs) are now considered to be applied as energy storage system for the electric vehicle (EV) and plug-in hybrid electric vehicle (HEV) due to their unique features, such as long cycle life, high operation voltage, low self-discharge, and environmental friendliness [1–5]. However, the present commercial LIBs using the carbonaceous materials, e.g. carbon and natural graphite as the anode materials that have low specific capacity, cannot meet the high energy density requirements [6,7]. In order to fulfill the automobile and military industrial demands, silicon is investigated as the candidate for anode material and has attracted the most interest due to its high theoretical specific capacity (4200 mA h g 1 for Li22Si5) [8–10]. Unfortunately, the practical application of the pure silicon has been hindered because of the drastic volume (300% volume changes) and structure changes during the Li+ insertion and extraction, which causes mechanical fracture and pulverization and results in rapid decrease in the capacity [11,12].
⇑ Corresponding author. Tel./fax: +86 024 23971783. E-mail address: [email protected] (H. Zhang). http://dx.doi.org/10.1016/j.jallcom.2014.01.187 0925-8388/Ó 2014 Elsevier B.V. All rights reserved.
Recently, there has been intense interests focus on the silicon/ graphite (SG) composites [13–19]. The carbonaceous material matrix is expected to function as the structure buffer layer and accommodates the volume change during the charging and discharging process of the anode materials. In normal SG composites, the graphite has low deformation capacity (10%), so large amount of graphite is added as buffer in order to accommodate the large scale volume change of the anode material during the process of charge and discharge. However, it heavily discounts the specific capacity of the electrode. Additionally, normal SG composites fail to form stable electric contact between active silicon and copper current collector which leads to polarization and capacity fading. To promote the utilization of the silicon and improve the electrochemical performance of the anodes, some metal has been added into silicon/graphite composite. However, the specific capacity remained to be too low [6,13,20]. Tin shall be incorporated into the SG nano-composite due to the following advantages. The theoretical specific capacity of tin is considerably high, indicating that the addition of tin does not heavily reduce the theoretical specific capacity of anodes. The fairly good ductility of metallic tin relaxes the stress mutation caused by the volume and structural changes and improves the fatigue property of the electrode during the repeating charge–discharge process, which is helpful to the
J. Wu et al. / Journal of Alloys and Compounds 596 (2014) 86–91
improvement of the cycle life of the anode. With compared to the graphite (parallel to the planes 3 106 S/m), tin has the higher electrical conductivity (9 106 S/m), which improves the electrical conductivity of the composite anode, and mitigate the polarization of the electrode, leading to increasing the cyclability of the anodes. Furthermore, the discharge potential (versus Li/Li+) of the tin is different from the silicon during the discharge process, while charging at the potential of silicon, tin can act as buffer. On the contrary, silicon acts as buffer while charging at the potential of tin. In this way, it will enhance the electrochemical performance of the composite electrodes. However, whether the aforementioned advantages are achieved or not lies on the microstructure of the composite, i.e., the correlation between silicon and tin in the composite. In this work, a novel nano-structured interpenetrating phase composite (NSIPC) of SGM was synthesized using HEMM method. In the composite, silicon and tin continuously extend in microstructure of material to form a completely homogeneously distribution. This novel structure ensures the well electric contact and mechanical integrity, leading to the better electrochemical performance than the present silicon/graphite anode composite [14,21,22]. 2. Experimental The commercial silicon powder (P99.99%, 300 mesh) and tin powder (P99.5%, 300 mesh) and graphite (P99.85%, 30 lm) were used as starting materials. The powders were mixed according to the predetermined nominal composition (listed in Table 1) and sealed into a 250 mL tempered steel bowl with 10 mm, 8 mm, 5 mm diameter bearing steel balls. The ball to powder mass ratio was 16:1. Mechanical milling was carried out in a Fritsch P5 planetary mill under argon atmosphere for 20 h with 30 min break every hour to remove the possible thermal effect. Milling time should not be longer than 20 h, because milling exceeds 20 h will cause contamination. The rotation speed of the milling is 250 rpm. X-ray diffraction (XRD) patterns of the as-synthesized composites were obtained using a Rigaku D/max-2500pc diffractometer (reflection 2 theta geometry, Cu Ka radiation). The morphologies and microstructures of the composites were observed by Supra 35 field emission scanning electron microscope (FESEM) and High-resolution transmission electron microscope (HR-TEM, FEI Tecnai F20 Super Twin FEG TEM). Electrochemical performances were evaluated using a CR2025-type coin half cell. The composite materials working electrodes were prepared by coating the slurries onto the copper foil (25 lm). The slurry was fabricated by mixing 60% composite materials, 20% carbon black (Super P) and 20% polyvinylidene fluoride (PVDF). After coating, the film was dried at 120 °C for 24 h under vacuum, then cut into sheets with 12 mm in diameter and compressed under a pressure of 2 105 Pa between two stainless steel plates. Lithium metal was used as the Li+ source, counter and reference electrode. 1 M LiPF6 (dissolved in ethylene carbonate and dimethyl carbonate with a 1:1 volume ratio) was applied as the electrolyte. A sheet with 16 mm in diameter Celgard 2400 membrane was utilized as the separator. The coin cell was assembled in an argon-filled glove box with the content of oxygen and moisture below 1 ppm. All the charging and discharging test were performed under different current density using a LAND-CT2001A battery test system (Jinnuo Wuhan Corp., China). Cyclic voltammetric measurements were performed in the voltage range from 2.5 to 0 V (versus Li/Li+) at a sweep rate of 0.001 V s 1 by an electrochemical station (CHI 770 CH Instruments. Shanghai).
3. Results and discussion Fig. 1 shows the XRD patterns of the SGM and SG composites synthesized by HEMM method. JCPDS #04-0673 for tin, JCPDS
Table 1 Composition of the as-prepared composites. Composites
SGM SG1 SG2
Nominal composition (At%) Silicon
Graphite
Tin
70 80 70
20 20 30
10
87
Fig. 1. XRD patterns of the SGM composite and SG composites with different silicon/graphite ratios.
#80-0018 for silicon and JCPDS #261079 for graphite are used to confirm that the XRD pattern of the SGM composite material is composed of the diffraction peaks of the element state silicon, tin and graphite. Additionally, the XRD pattern of the SGM shows very low intensities of broad reflections, which are associated with the amorphization of Si. This is caused by the high energy impact during the HEMM, which leads to deformation of the silicon particles and increases the crystal defects. These crystal defects help enhance the diffusion of the elements and form amorphous structure. However, the diffraction patterns of the SG composites with different weight ratios are indexed as just single silicon substance. High energy mechanical milling is an outstanding approach to obtain well-distribution silicon, tin and graphite. No silicon carbide and steel peaks can be seen on all the three patterns, indicating that HEMM do not induce silicon carbide compounds and metallic contamination. It is good for the achievement of the high specific capacity properties of the studied system without inactive carbide phase and metallic contamination. SEM micrographs of the as-synthesized composites are shown in Fig. 2. It can be seen from Fig. 2(a) and (b) that the SG composite consisted of large agglomerates in a shape of grape, which is typically found in the products prepared by HEMM. Actually, these agglomerates are composed of nano-particles. Because of the high energy impact results in cold welding and fracturing, the particles aggregate together. Similar morphology for SGM samples are observed via SEM shown in Fig. 2(c), the large chunks of 10 lm and more are carbon particles coated with silicon particles and tin particles. In order to get a further understanding about the distribution of the elements after adding tin, mapping analysis was done. It can be seen from Fig. 2(d) that, silicon and tin form a continuous net work. In fact, both of constituent silicon and tin in their standalone state have an open-cellular microstructure. Because of the fairly good plasticity, the additive tin shows large deformation when under high energy ball milling. The silicon nano-particles are entangled by the ductile tin and cold-welded together. After milling of 20 h, lots of silicon nano-particles and tin interpenetrate into each other and form clusters. Each of these clusters is confirmed to NSIPC. TEM and HR-TEM micrograph are shown in Fig. 3 to fully recognize the morphological features and microstructure of the SGM composite. Aggregates are shown in Fig. 3(a), which have no difference with the results observed by SEM. From the HR-TEM micrograph Fig. 3(b), two kinds of lattice fringes with distance of 0.311 nm and 0.279 nm can be visualized, which correspond to the (1 1 1) plane of cubic silicon (JCPDS#800018) and the (1 0 1) plane of tetragonal tin (JCPDS#04-0673), respectively. It indicates that Sn has no influence in the lattice
88
J. Wu et al. / Journal of Alloys and Compounds 596 (2014) 86–91
Fig. 2. (a and b) FESEM images of samples SG1 and SG2 respectively. (c) FESEM images of the SGM samples. (d) Mapping micrograph of the IPC of SGM.
Fig. 3. TEM image and HR-TEM image of IPC of SGM composite.
structure of Si. It is shown in the marked circle that these two kinds of constituents combine closed and interpenetrate into each other, which is consistent with the mapping analysis. In general, the SGM composites anode material is determined to be NSIPC by SEM and TEM. However, for the SG composites, such kind of interpenetrating microstructure cannot be found because there is no ductile metallic phase exists. To realize the electrochemical properties of the as-prepared composite anode materials, electrochemical tests were carried out. Fig. 4 shows the first three charge and discharge profiles of the SGM, SG1 and SG2 electrodes at the charge–discharge rate of 0.1 C. It can be seen that the SG1 delivers an initial discharge capacity of 1760 mA h g 1, which is 51% to the theoretical capacity of the nominal composition Si80C20 and the SG2 exhibits the lowest initial discharge capacity of 930 mA h g 1, which is only 30% to the theoretical capacity of Si70C30. However, the SGM shows highest percentage about 57% to the theoretical capacity of Si70C20Sn10 with an initial discharge capacity of 1390 mA h g 1. Therefore, among the as-prepared composite anode electrodes,
Fig. 4. The first three charge–discharge curves of the SGM, SG1 and SG2.
J. Wu et al. / Journal of Alloys and Compounds 596 (2014) 86–91
SGM composite shows the highest first columbic efficiency. Also, we can see severe capacity fading occurs in SG1 and SG2 electrodes. The second discharge capacity of SG1 is 901 mA h g 1, while the third discharge capacity decreases to 625 mA h g 1. Similarly, severe capacity degradation happens in the SG2 composite electrode. Nevertheless, the SGM exhibits better cycling performance. The discharge specific capacity of the SGM keeps at 890 mA h g 1 after the initial cycle. The reason for the difference of the initial specific capacity is that the SGM and SG composite anode materials contain different fraction of silicon. SG1 has the highest specific capacity due to its highest fraction of silicon powder, but the percentage to the theoretical capacity is lower than SGM. The volume of silicon changes by about 300% during cycling. Much of the active material tend to pulverize and loses contact with the Cu current collector, resulting in poor transport of electrons in SG electrodes. As for IPC of SGM electrode, silicon nanoparticles are entangled by the ductile tin. The expansion of the volume resulting from the lithiation can be accommodated by the interpenetrating microstructure between silicon and tin. The ductile tin can serve as the buffer layer in the SGM composite. Facile strain relaxation in the interpenetrating microstructure allows active silicon and tin to increase volume without pulverizing. It is known that the active silicon nano-particles and deformed tin will undergo a severe contracting while dealloying, which will probably deteriorate the electric contact between active material and the copper current collector [23]. Because the interfacial bond strength between silicon and tin is strong that the integrity of the interpenetrating microstructure can be maintained and the electric contact between active materials and the current collector can be kept in a stable state after delithiation, allowing for the effective electron transportation, hence, decreasing the capacity fading. Moreover, silicon and tin are all active material for lithiation. The standard electric potential of Li/Li+ is 3.04 V, while Sn/Sn2+ and Si/Si4+ are 0.151 V and 1.24 V, respectively [24]. In contrast to the SG electrodes in which the discharge and charge process mainly concentrate on the active silicon potential versus Li+, active silicon and active tin are combined closely, charge and discharge under different potential plateau in the interpenetrating structure of SGM electrode. Charging and discharging under different potential plateau can help decrease the stress relaxing and eliminate the stress mutation, thus prevent the micro-cracks from expanding in the electrode during the insertion and extraction of the lithium ion. The better cycling performance obtained in SGM composite electrode might benefit by other favourable influences such as stable self-limiting and organic solvent insoluble solid electrolyte interface (SEI) layer, which can not only allow the Li+ to go across freely but also prevent the solvent molecules from passing through. Uniform SEI formed during the first cycle alloying will suppress the further reaction between organic electrolyte and the anode materials, thus electrode can be preserved from destroying. It is beneficial to both the reversible capacity and cyclability. From the Fig. 4, we can see that there is a voltage potential plateau around 0.75 V in the SGM discharging curve, which is corresponding to the formation of the SEI. However, SG1 and SG2 composite electrodes first discharge curve do not show the related voltage plateau indicating that uniform SEI layer failed to be formed, so the capacity fades rapidly. However, all the three electrode materials showed large initial irreversible capacity loss. As shown in Fig. 4, the initial irreversible capacity loss of the SGM, SG1 and SG2 are 500 mA h g 1, 861 mA h g 1 and 440 mA h g 1, respectively. It is supposed that the large capacity loss mainly results from the formation of the SEI between silicon and electrolyte, which consumes the Li+. But, the so-called interpenetrating phase structure is designed to decreases the contact area between the silicon and electrolyte, hence, the SEI formation is reduced, which can decrease the
89
consumption of Li+. It is reasonable to believe that the initial irreversible capacity loss is caused by other possibilities. In order to find out the causes, the SGM composite electrode was examined under the low constant current density condition. In Fig. 5, the initial discharge specific capacity of the SGM composite electrode charging–discharging at a constant current density of 237 mA g 1 reaches to 1790 mA h g 1, which is higher than reported anode composite with adding other metallic i.e. Ni, Ge, Ag, etc. [6,13,25,26]. Polarization can be suppressed at such a low current density, thus reducing the initial capacity loss and avoiding severe capacity fading after the first cycle. It is interesting to find out that the second cycle charge capacity is higher than the first cycle charge capacity in Fig 5. Polyvinylidene fluoride (PVDF) binder has strong adherence, thus adheres on active material surface. Soakage of electrolyte in the inner area of the electrode material is impeded due to the poor absorption ability of the PVDF binder for electrolyte, therefore gives rise to a gradual activation of the electrode-active materials [27]. The initial capacity is higher than the SGM electrode charge–discharge at 0.1 C (about 400 mA g 1) indicating that the activation of the active material is not completely in the first cycle under relatively high charge and discharge current density. Generally speaking, the initial capacity loss and the activation of the electrode are influenced by the binder and the current density. From the above electrochemical tests, it can be known that electrodes charging and discharging under different current density conditions delivered different specific capacity. In order to get a deep understanding about the rate capability of the NSIPC of SGM composite, charge–discharge experiments were carried out under various C rate (0.5 C and 1 C, 0.5 C is about 1556 mA g 1 while 1 C is about 3113 mA g 1) on purpose. Several cycles(cycle 2, 31, 32 and 51) of the charge and discharge process for SGM composite electrode at 0.5 C and 1 C rate with a voltage range from 0.03 V to 1.50 V (versus Li/Li+) are presented in Fig. 6. At first, the electrode charges and discharges at 0.5 C for 30 cycles, then charges and discharges at 1 C for next 10 cycles. Lastly, the electrode charges and discharges at 0.5 C again. Pure silicon is a semi-conductor, when charging and discharging at high current density, the electrode composed by pure silicon polarizes rigorously, Li+ deposits on the interface of the electrode and fails to be utilized, thus has a low specific capacity even less than the carbonaceous materials, but as for the designed nano-structured IPC of SGM, active silicon particles are entangled by the good electronic conductor tin. The electrical conductivity of the tin (9 106 S/m)
Fig. 5. First and second charge–discharge profiles for SGM composite electrode with a low current density of 237 mA g 1 and the charge–discharge voltage range from 0.03 to 1.50 V.
90
J. Wu et al. / Journal of Alloys and Compounds 596 (2014) 86–91
Fig. 6. Rate capability of the SGM (C rate is given with respect to the SGM composite).
Fig. 8. Cyclic voltammogram of the SGM composite anode material.
is much higher than the electrically conducting graphite (3 106 S/m). 10% of tin was added to improve the electrical conductivity of the composite anode and provide much more pathways for the electron transportation than the SG anode material without inducing tin oxide (the amount of tin above 10% will induce impurities. i.e. SnO2), which mitigate the polarization of the electrode resulting from the poor conductivity of the composites, hence, improve the C rate ability of the anodes. The capacity charges and discharges at 0.5 C were kept at a stable stage of 880 mA h g 1, while the capacity charges and discharges at 1 C were maintained at 460 mA h g 1 higher than the carbonaceous material (372 mA h g 1). In addition, it is revealed from Fig. 6 that the activate process of the lithiation and delithiation reaction is repeatable. The discharge capacity can be maintained at 880 mA h g 1 at the beginning 30 cycles charge–discharge at 0.5 C, then further to 1 C, the capacity declined to 460 mA h g 1 in the next 10 cycles. Once the current density returned to 0.5 C, the capacity of 880 mA h g 1 can be recovered. Under each charge–discharge current density condition, SGM composite electrode material has its corresponding optimal state for charging and discharging, so the capacity can be kept at a stable value. Fig. 7 shows the cyclability of the NSIPC of SGM charge–discharge at relative high constant current density (650 mA g 1). A comparison of the cycling performance between SGM and SG composites is given by Fig. 7(a). It is obvious that the as-prepared IPC of
SGM has much better capacity retention than SG composites during cycling. The remarkable improvement of the cyclability is attributed to the interpenetrating microstructure, and good conductivity of the additive active material tin. As displayed, Fig. 7(b) exhibits the reversible capacity of composite anode electrode and charge–discharge coulombic efficiency of the SGM as a function of cycle number. The capacity is kept at 700 mA h g 1 with the charge–discharge coulombic efficiency around 100%. There is a ‘‘bump’’ in the specific capacity of the SGM and the corresponding coulombic efficiency is higher than 100%. This bump is probably just an experimental artifact. Cyclic voltammetry (CV) analysis was performed at a scan rate of 0.001 V s 1 in the range from 0 to 2.50 V to verify the electrochemical reactions occurred in the lithiation and delithiation. Fig. 8 shows the cyclic voltammogram of the SGM composite anode material. The upper part of the figure (positive currents) corresponds to the lithiation, the lower part to the delithiation. The reduction peak around 0.25 V (versus Li/Li+) is only showed in the first cycle can be responsible for the formation of the amorphous LixSi. During the first discharging process, the stable amorphous microstructure is formed, which causes large irreversible capacity loss. The pair of peaks A–a are due to the Li–C alloying/dealloying process. The pair of peaks B–b and D–d demonstrate that the reaction of Li–Si proceeds in a multi-step process. The pair of peaks C–c are associated with the reaction of the Li–Sn, which
Fig. 7. Cyclability and coulombic efficiency of the composite electrodes with a current density of 635 mA g SG2, (b) SGM.
1
under a voltage arrange from 0.03 to 1.50 V. (a) SGM, SG1 and
J. Wu et al. / Journal of Alloys and Compounds 596 (2014) 86–91
are not found on the CV diagram of SG composite [19]. Cyclic voltammetry cycles show the same profile except the first cycle indicating that these redox reactions were repeatable in the NSIPC of SGM electrode. Also, the area encircled by the redox reaction curve in the second and third cycles are all larger than the initial cycle, which suggests that the lithiation reaction is more completely in the later two cycles.
4. Conclusions In summary, a novel NSIPC of SGM anode material for lithiumion batteries was obtained via HEMM method. The SGM composite electrode exhibits better cyclability than the barely silicon/graphite composite electrodes due to the interpenetrating microstructure between silicon and ductile tin. The initial discharge specific capacity of the as-prepared SGM composite electrode is relatively high about 1790 mA h g 1 with 1592 mA h g 1 reversible capacity retention at a current density of 237 mA g 1 in the voltage arrangement from 0.03 V to 1.50 V. Additionally, the SGM composite electrode shows excellent rate capability and possesses stable cycling performance even at a large current density. Consequently, the relatively low cost and excellent electrochemical capability of the SGM composite would well meet the challenge in the rapid charge and discharge. SGM composite can be the promising candidate anode material for the next generation lithium rechargeable batteries.
Acknowledgements The financial support from the National Natural Science Foundation of China and Chinese Academy of Sciences are gratefully acknowledged.
91
References [1] M. Holzapfel, H. Buqa, L.J. Hardwick, M. Hahn, A. Würsig, W. Scheifele, P. Novák, R. Kötz, C. Veit, F.-M. Petrat, Electrochim. Acta 52 (2006) 973–978. [2] O.K. Park, Y. Cho, S. Lee, H.-C. Yoo, H.-K. Song, J. Cho, Energy Environ. Sci. 4 (2011) 1621. [3] B. Scrosati, J. Hassoun, Y.-K. Sun, Energy Environ. Sci. 4 (2011) 3287. [4] J.B. Goodenough, Y. Kim, Chem. Mater. 22 (2010) 587–603. [5] K. Kang, Y.S. Meng, J. Breger, C.P. Grey, G. Ceder, Science 311 (2006) 977–980. [6] C.-M. Hwang, J.-W. Park, Electrochim. Acta 56 (2011) 6737–6747. [7] J.Y. Howe, D.J. Burton, Y. Qi, H.M. Meyer, M. Nazri, G.A. Nazri, A.C. Palmer, P.D. Lake, J. Power Sources 221 (2013) 455–461. [8] Y.H. Xu, G.P. Yin, P.J. Zuo, Electrochim. Acta 54 (2008) 341–345. [9] H.K. Liu, Z.P. Guo, J.Z. Wang, K. Konstantinov, J. Mater. Chem. 20 (2010) 10055. [10] X. Xin, X. Zhou, F. Wang, X. Yao, X. Xu, Y. Zhu, Z. Liu, J. Mater. Chem. 22 (2012) 7724. [11] W.-J. Zhang, J. Power Sources 196 (2011) 13–24. [12] S. Rousselot, M. Gauthier, D. Mazouzi, B. Lestriez, D. Guyomard, L. Roué, J. Power Sources 202 (2012) 262–268. [13] M.-S. Park, Electrochem. Acta 50 (2005) 5561–5567. [14] J. Yang, B.F. Wang, K. Wang, Y. Liu, J.Y. Xie, Z.S. Wen, Electrochem. Solid-State Lett. 6 (2003) A154. [15] M.-S. Park, Y.-J. Lee, S. Rajendran, M.-S. Song, H.-S. Kim, J.-Y. Lee, Electrochim. Acta 50 (2005) 5561–5567. [16] Z.-J. Han, N. Yabuuchi, K. Shimomura, M. Murase, H. Yui, S. Komaba, Energy Environ. Sci. 5 (2012) 9014. [17] X. Shen, D. Mu, S. Chen, B. Xu, B. Wu, F. Wu, J. Alloys Comp. 552 (2013) 60–64. [18] K. Hanai, Y. Liu, N. Imanishi, A. Hirano, M. Matsumura, T. Ichikawa, Y. Takeda, J. Power Sources 146 (2005) 156–160. [19] W. Li, R. Yang, X. Wang, T. Wang, J. Zheng, X. Li, J. Power Sources 221 (2013) 242–246. [20] L. Zhang, X. Song, F. Wang, Q. Hu, Z. Sun, S. Yang, L. Wang, S. Sun, J. Solid State Electrochem. 16 (2012) 2159–2167. [21] H.-C. Tao, M. Huang, L.-Z. Fan, X. Qu, Electrochim. Acta 89 (2013) 394–399. [22] P. Zuo, W. Yang, X. Cheng, G. Yin, Ionics 17 (2010) 87–90. [23] Y.-L. Kim, Y.-K. Sun, S.-M. Lee, Electrochim. Acta 53 (2008) 4500–4504. [24] Y. Chen, J. Qian, Y. Cao, H. Yang, X. Ai, ACS Appl. Mater. Interfaces (2012). [25] M. Au, Y. He, Y. Zhao, H. Ghassemi, R.S. Yassar, B. Garcia-Diaz, T. Adams, J. Power Sources 196 (2011) 9640–9647. [26] J. Yao, Z. Jia, P. Zhang, C. Shen, J. Wang, K.-F. Aguey-Zinsou, L. Wang, Ionics 19 (2012) 401–407. [27] J.R. Szczech, S. Jin, Energy Environ. Sci. 4 (2011) 56.
Contents lists available at ScienceDirect
Journal of Alloys and Compounds journal homepage: www.elsevier.com/locate/jalcom
A novel nano-structured interpenetrating phase composite of silicon/ graphite–tin for lithium-ion rechargeable batteries anode materials Jinbo Wu, Zhengwang Zhu, Hongwei Zhang, Huameng Fu, Hong Li, Aimin Wang, Haifeng Zhang ⇑, Zhuangqi Hu Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences, 72 Wenhua Road, Shenyang 110016, China
a r t i c l e
i n f o
Article history: Received 11 October 2013 Received in revised form 19 December 2013 Accepted 24 January 2014 Available online 4 February 2014 Keywords: Silicon/graphite–tin Lithium-ion batteries High energy mechanical milling Electrochemical performance Anode material Nano-structured interpenetrating phase composite
a b s t r a c t A novel nano-structured interpenetrating phase composite (NSIPC) of silicon/graphite–tin (SGM) anode material for lithium-ion rechargeable batteries is synthesized by high energy mechanical milling (HEMM). The structural and morphological characterizations have been carried out through X-ray diffraction (XRD), scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The electrochemical performances have been analyzed with reference to Li+/Li and the results are compared with silicon/graphite composites. The SGM NSIPC electrode exhibits the better cyclability than the SG composite electrodes. The initial discharge specific capacity of the as-prepared SGM NSIPC is relatively high around 1790 mA h g 1 with 1592 mA h g 1 reversible capacity retention in the following cycle at a current density of 237 mA g 1 in the voltage from 0.03 V to 1.5 V. In addition, the SGM NSIPC electrode shows the good rate capability and possesses the stable cycling performance even charging and discharging at the large current density. Consequently, SGM NSIPC can be the promising anode material for the next generation lithium ion rechargeable batteries. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Li-ion batteries (LIBs) are now considered to be applied as energy storage system for the electric vehicle (EV) and plug-in hybrid electric vehicle (HEV) due to their unique features, such as long cycle life, high operation voltage, low self-discharge, and environmental friendliness [1–5]. However, the present commercial LIBs using the carbonaceous materials, e.g. carbon and natural graphite as the anode materials that have low specific capacity, cannot meet the high energy density requirements [6,7]. In order to fulfill the automobile and military industrial demands, silicon is investigated as the candidate for anode material and has attracted the most interest due to its high theoretical specific capacity (4200 mA h g 1 for Li22Si5) [8–10]. Unfortunately, the practical application of the pure silicon has been hindered because of the drastic volume (300% volume changes) and structure changes during the Li+ insertion and extraction, which causes mechanical fracture and pulverization and results in rapid decrease in the capacity [11,12].
⇑ Corresponding author. Tel./fax: +86 024 23971783. E-mail address: [email protected] (H. Zhang). http://dx.doi.org/10.1016/j.jallcom.2014.01.187 0925-8388/Ó 2014 Elsevier B.V. All rights reserved.
Recently, there has been intense interests focus on the silicon/ graphite (SG) composites [13–19]. The carbonaceous material matrix is expected to function as the structure buffer layer and accommodates the volume change during the charging and discharging process of the anode materials. In normal SG composites, the graphite has low deformation capacity (10%), so large amount of graphite is added as buffer in order to accommodate the large scale volume change of the anode material during the process of charge and discharge. However, it heavily discounts the specific capacity of the electrode. Additionally, normal SG composites fail to form stable electric contact between active silicon and copper current collector which leads to polarization and capacity fading. To promote the utilization of the silicon and improve the electrochemical performance of the anodes, some metal has been added into silicon/graphite composite. However, the specific capacity remained to be too low [6,13,20]. Tin shall be incorporated into the SG nano-composite due to the following advantages. The theoretical specific capacity of tin is considerably high, indicating that the addition of tin does not heavily reduce the theoretical specific capacity of anodes. The fairly good ductility of metallic tin relaxes the stress mutation caused by the volume and structural changes and improves the fatigue property of the electrode during the repeating charge–discharge process, which is helpful to the
J. Wu et al. / Journal of Alloys and Compounds 596 (2014) 86–91
improvement of the cycle life of the anode. With compared to the graphite (parallel to the planes 3 106 S/m), tin has the higher electrical conductivity (9 106 S/m), which improves the electrical conductivity of the composite anode, and mitigate the polarization of the electrode, leading to increasing the cyclability of the anodes. Furthermore, the discharge potential (versus Li/Li+) of the tin is different from the silicon during the discharge process, while charging at the potential of silicon, tin can act as buffer. On the contrary, silicon acts as buffer while charging at the potential of tin. In this way, it will enhance the electrochemical performance of the composite electrodes. However, whether the aforementioned advantages are achieved or not lies on the microstructure of the composite, i.e., the correlation between silicon and tin in the composite. In this work, a novel nano-structured interpenetrating phase composite (NSIPC) of SGM was synthesized using HEMM method. In the composite, silicon and tin continuously extend in microstructure of material to form a completely homogeneously distribution. This novel structure ensures the well electric contact and mechanical integrity, leading to the better electrochemical performance than the present silicon/graphite anode composite [14,21,22]. 2. Experimental The commercial silicon powder (P99.99%, 300 mesh) and tin powder (P99.5%, 300 mesh) and graphite (P99.85%, 30 lm) were used as starting materials. The powders were mixed according to the predetermined nominal composition (listed in Table 1) and sealed into a 250 mL tempered steel bowl with 10 mm, 8 mm, 5 mm diameter bearing steel balls. The ball to powder mass ratio was 16:1. Mechanical milling was carried out in a Fritsch P5 planetary mill under argon atmosphere for 20 h with 30 min break every hour to remove the possible thermal effect. Milling time should not be longer than 20 h, because milling exceeds 20 h will cause contamination. The rotation speed of the milling is 250 rpm. X-ray diffraction (XRD) patterns of the as-synthesized composites were obtained using a Rigaku D/max-2500pc diffractometer (reflection 2 theta geometry, Cu Ka radiation). The morphologies and microstructures of the composites were observed by Supra 35 field emission scanning electron microscope (FESEM) and High-resolution transmission electron microscope (HR-TEM, FEI Tecnai F20 Super Twin FEG TEM). Electrochemical performances were evaluated using a CR2025-type coin half cell. The composite materials working electrodes were prepared by coating the slurries onto the copper foil (25 lm). The slurry was fabricated by mixing 60% composite materials, 20% carbon black (Super P) and 20% polyvinylidene fluoride (PVDF). After coating, the film was dried at 120 °C for 24 h under vacuum, then cut into sheets with 12 mm in diameter and compressed under a pressure of 2 105 Pa between two stainless steel plates. Lithium metal was used as the Li+ source, counter and reference electrode. 1 M LiPF6 (dissolved in ethylene carbonate and dimethyl carbonate with a 1:1 volume ratio) was applied as the electrolyte. A sheet with 16 mm in diameter Celgard 2400 membrane was utilized as the separator. The coin cell was assembled in an argon-filled glove box with the content of oxygen and moisture below 1 ppm. All the charging and discharging test were performed under different current density using a LAND-CT2001A battery test system (Jinnuo Wuhan Corp., China). Cyclic voltammetric measurements were performed in the voltage range from 2.5 to 0 V (versus Li/Li+) at a sweep rate of 0.001 V s 1 by an electrochemical station (CHI 770 CH Instruments. Shanghai).
3. Results and discussion Fig. 1 shows the XRD patterns of the SGM and SG composites synthesized by HEMM method. JCPDS #04-0673 for tin, JCPDS
Table 1 Composition of the as-prepared composites. Composites
SGM SG1 SG2
Nominal composition (At%) Silicon
Graphite
Tin
70 80 70
20 20 30
10
87
Fig. 1. XRD patterns of the SGM composite and SG composites with different silicon/graphite ratios.
#80-0018 for silicon and JCPDS #261079 for graphite are used to confirm that the XRD pattern of the SGM composite material is composed of the diffraction peaks of the element state silicon, tin and graphite. Additionally, the XRD pattern of the SGM shows very low intensities of broad reflections, which are associated with the amorphization of Si. This is caused by the high energy impact during the HEMM, which leads to deformation of the silicon particles and increases the crystal defects. These crystal defects help enhance the diffusion of the elements and form amorphous structure. However, the diffraction patterns of the SG composites with different weight ratios are indexed as just single silicon substance. High energy mechanical milling is an outstanding approach to obtain well-distribution silicon, tin and graphite. No silicon carbide and steel peaks can be seen on all the three patterns, indicating that HEMM do not induce silicon carbide compounds and metallic contamination. It is good for the achievement of the high specific capacity properties of the studied system without inactive carbide phase and metallic contamination. SEM micrographs of the as-synthesized composites are shown in Fig. 2. It can be seen from Fig. 2(a) and (b) that the SG composite consisted of large agglomerates in a shape of grape, which is typically found in the products prepared by HEMM. Actually, these agglomerates are composed of nano-particles. Because of the high energy impact results in cold welding and fracturing, the particles aggregate together. Similar morphology for SGM samples are observed via SEM shown in Fig. 2(c), the large chunks of 10 lm and more are carbon particles coated with silicon particles and tin particles. In order to get a further understanding about the distribution of the elements after adding tin, mapping analysis was done. It can be seen from Fig. 2(d) that, silicon and tin form a continuous net work. In fact, both of constituent silicon and tin in their standalone state have an open-cellular microstructure. Because of the fairly good plasticity, the additive tin shows large deformation when under high energy ball milling. The silicon nano-particles are entangled by the ductile tin and cold-welded together. After milling of 20 h, lots of silicon nano-particles and tin interpenetrate into each other and form clusters. Each of these clusters is confirmed to NSIPC. TEM and HR-TEM micrograph are shown in Fig. 3 to fully recognize the morphological features and microstructure of the SGM composite. Aggregates are shown in Fig. 3(a), which have no difference with the results observed by SEM. From the HR-TEM micrograph Fig. 3(b), two kinds of lattice fringes with distance of 0.311 nm and 0.279 nm can be visualized, which correspond to the (1 1 1) plane of cubic silicon (JCPDS#800018) and the (1 0 1) plane of tetragonal tin (JCPDS#04-0673), respectively. It indicates that Sn has no influence in the lattice
88
J. Wu et al. / Journal of Alloys and Compounds 596 (2014) 86–91
Fig. 2. (a and b) FESEM images of samples SG1 and SG2 respectively. (c) FESEM images of the SGM samples. (d) Mapping micrograph of the IPC of SGM.
Fig. 3. TEM image and HR-TEM image of IPC of SGM composite.
structure of Si. It is shown in the marked circle that these two kinds of constituents combine closed and interpenetrate into each other, which is consistent with the mapping analysis. In general, the SGM composites anode material is determined to be NSIPC by SEM and TEM. However, for the SG composites, such kind of interpenetrating microstructure cannot be found because there is no ductile metallic phase exists. To realize the electrochemical properties of the as-prepared composite anode materials, electrochemical tests were carried out. Fig. 4 shows the first three charge and discharge profiles of the SGM, SG1 and SG2 electrodes at the charge–discharge rate of 0.1 C. It can be seen that the SG1 delivers an initial discharge capacity of 1760 mA h g 1, which is 51% to the theoretical capacity of the nominal composition Si80C20 and the SG2 exhibits the lowest initial discharge capacity of 930 mA h g 1, which is only 30% to the theoretical capacity of Si70C30. However, the SGM shows highest percentage about 57% to the theoretical capacity of Si70C20Sn10 with an initial discharge capacity of 1390 mA h g 1. Therefore, among the as-prepared composite anode electrodes,
Fig. 4. The first three charge–discharge curves of the SGM, SG1 and SG2.
J. Wu et al. / Journal of Alloys and Compounds 596 (2014) 86–91
SGM composite shows the highest first columbic efficiency. Also, we can see severe capacity fading occurs in SG1 and SG2 electrodes. The second discharge capacity of SG1 is 901 mA h g 1, while the third discharge capacity decreases to 625 mA h g 1. Similarly, severe capacity degradation happens in the SG2 composite electrode. Nevertheless, the SGM exhibits better cycling performance. The discharge specific capacity of the SGM keeps at 890 mA h g 1 after the initial cycle. The reason for the difference of the initial specific capacity is that the SGM and SG composite anode materials contain different fraction of silicon. SG1 has the highest specific capacity due to its highest fraction of silicon powder, but the percentage to the theoretical capacity is lower than SGM. The volume of silicon changes by about 300% during cycling. Much of the active material tend to pulverize and loses contact with the Cu current collector, resulting in poor transport of electrons in SG electrodes. As for IPC of SGM electrode, silicon nanoparticles are entangled by the ductile tin. The expansion of the volume resulting from the lithiation can be accommodated by the interpenetrating microstructure between silicon and tin. The ductile tin can serve as the buffer layer in the SGM composite. Facile strain relaxation in the interpenetrating microstructure allows active silicon and tin to increase volume without pulverizing. It is known that the active silicon nano-particles and deformed tin will undergo a severe contracting while dealloying, which will probably deteriorate the electric contact between active material and the copper current collector [23]. Because the interfacial bond strength between silicon and tin is strong that the integrity of the interpenetrating microstructure can be maintained and the electric contact between active materials and the current collector can be kept in a stable state after delithiation, allowing for the effective electron transportation, hence, decreasing the capacity fading. Moreover, silicon and tin are all active material for lithiation. The standard electric potential of Li/Li+ is 3.04 V, while Sn/Sn2+ and Si/Si4+ are 0.151 V and 1.24 V, respectively [24]. In contrast to the SG electrodes in which the discharge and charge process mainly concentrate on the active silicon potential versus Li+, active silicon and active tin are combined closely, charge and discharge under different potential plateau in the interpenetrating structure of SGM electrode. Charging and discharging under different potential plateau can help decrease the stress relaxing and eliminate the stress mutation, thus prevent the micro-cracks from expanding in the electrode during the insertion and extraction of the lithium ion. The better cycling performance obtained in SGM composite electrode might benefit by other favourable influences such as stable self-limiting and organic solvent insoluble solid electrolyte interface (SEI) layer, which can not only allow the Li+ to go across freely but also prevent the solvent molecules from passing through. Uniform SEI formed during the first cycle alloying will suppress the further reaction between organic electrolyte and the anode materials, thus electrode can be preserved from destroying. It is beneficial to both the reversible capacity and cyclability. From the Fig. 4, we can see that there is a voltage potential plateau around 0.75 V in the SGM discharging curve, which is corresponding to the formation of the SEI. However, SG1 and SG2 composite electrodes first discharge curve do not show the related voltage plateau indicating that uniform SEI layer failed to be formed, so the capacity fades rapidly. However, all the three electrode materials showed large initial irreversible capacity loss. As shown in Fig. 4, the initial irreversible capacity loss of the SGM, SG1 and SG2 are 500 mA h g 1, 861 mA h g 1 and 440 mA h g 1, respectively. It is supposed that the large capacity loss mainly results from the formation of the SEI between silicon and electrolyte, which consumes the Li+. But, the so-called interpenetrating phase structure is designed to decreases the contact area between the silicon and electrolyte, hence, the SEI formation is reduced, which can decrease the
89
consumption of Li+. It is reasonable to believe that the initial irreversible capacity loss is caused by other possibilities. In order to find out the causes, the SGM composite electrode was examined under the low constant current density condition. In Fig. 5, the initial discharge specific capacity of the SGM composite electrode charging–discharging at a constant current density of 237 mA g 1 reaches to 1790 mA h g 1, which is higher than reported anode composite with adding other metallic i.e. Ni, Ge, Ag, etc. [6,13,25,26]. Polarization can be suppressed at such a low current density, thus reducing the initial capacity loss and avoiding severe capacity fading after the first cycle. It is interesting to find out that the second cycle charge capacity is higher than the first cycle charge capacity in Fig 5. Polyvinylidene fluoride (PVDF) binder has strong adherence, thus adheres on active material surface. Soakage of electrolyte in the inner area of the electrode material is impeded due to the poor absorption ability of the PVDF binder for electrolyte, therefore gives rise to a gradual activation of the electrode-active materials [27]. The initial capacity is higher than the SGM electrode charge–discharge at 0.1 C (about 400 mA g 1) indicating that the activation of the active material is not completely in the first cycle under relatively high charge and discharge current density. Generally speaking, the initial capacity loss and the activation of the electrode are influenced by the binder and the current density. From the above electrochemical tests, it can be known that electrodes charging and discharging under different current density conditions delivered different specific capacity. In order to get a deep understanding about the rate capability of the NSIPC of SGM composite, charge–discharge experiments were carried out under various C rate (0.5 C and 1 C, 0.5 C is about 1556 mA g 1 while 1 C is about 3113 mA g 1) on purpose. Several cycles(cycle 2, 31, 32 and 51) of the charge and discharge process for SGM composite electrode at 0.5 C and 1 C rate with a voltage range from 0.03 V to 1.50 V (versus Li/Li+) are presented in Fig. 6. At first, the electrode charges and discharges at 0.5 C for 30 cycles, then charges and discharges at 1 C for next 10 cycles. Lastly, the electrode charges and discharges at 0.5 C again. Pure silicon is a semi-conductor, when charging and discharging at high current density, the electrode composed by pure silicon polarizes rigorously, Li+ deposits on the interface of the electrode and fails to be utilized, thus has a low specific capacity even less than the carbonaceous materials, but as for the designed nano-structured IPC of SGM, active silicon particles are entangled by the good electronic conductor tin. The electrical conductivity of the tin (9 106 S/m)
Fig. 5. First and second charge–discharge profiles for SGM composite electrode with a low current density of 237 mA g 1 and the charge–discharge voltage range from 0.03 to 1.50 V.
90
J. Wu et al. / Journal of Alloys and Compounds 596 (2014) 86–91
Fig. 6. Rate capability of the SGM (C rate is given with respect to the SGM composite).
Fig. 8. Cyclic voltammogram of the SGM composite anode material.
is much higher than the electrically conducting graphite (3 106 S/m). 10% of tin was added to improve the electrical conductivity of the composite anode and provide much more pathways for the electron transportation than the SG anode material without inducing tin oxide (the amount of tin above 10% will induce impurities. i.e. SnO2), which mitigate the polarization of the electrode resulting from the poor conductivity of the composites, hence, improve the C rate ability of the anodes. The capacity charges and discharges at 0.5 C were kept at a stable stage of 880 mA h g 1, while the capacity charges and discharges at 1 C were maintained at 460 mA h g 1 higher than the carbonaceous material (372 mA h g 1). In addition, it is revealed from Fig. 6 that the activate process of the lithiation and delithiation reaction is repeatable. The discharge capacity can be maintained at 880 mA h g 1 at the beginning 30 cycles charge–discharge at 0.5 C, then further to 1 C, the capacity declined to 460 mA h g 1 in the next 10 cycles. Once the current density returned to 0.5 C, the capacity of 880 mA h g 1 can be recovered. Under each charge–discharge current density condition, SGM composite electrode material has its corresponding optimal state for charging and discharging, so the capacity can be kept at a stable value. Fig. 7 shows the cyclability of the NSIPC of SGM charge–discharge at relative high constant current density (650 mA g 1). A comparison of the cycling performance between SGM and SG composites is given by Fig. 7(a). It is obvious that the as-prepared IPC of
SGM has much better capacity retention than SG composites during cycling. The remarkable improvement of the cyclability is attributed to the interpenetrating microstructure, and good conductivity of the additive active material tin. As displayed, Fig. 7(b) exhibits the reversible capacity of composite anode electrode and charge–discharge coulombic efficiency of the SGM as a function of cycle number. The capacity is kept at 700 mA h g 1 with the charge–discharge coulombic efficiency around 100%. There is a ‘‘bump’’ in the specific capacity of the SGM and the corresponding coulombic efficiency is higher than 100%. This bump is probably just an experimental artifact. Cyclic voltammetry (CV) analysis was performed at a scan rate of 0.001 V s 1 in the range from 0 to 2.50 V to verify the electrochemical reactions occurred in the lithiation and delithiation. Fig. 8 shows the cyclic voltammogram of the SGM composite anode material. The upper part of the figure (positive currents) corresponds to the lithiation, the lower part to the delithiation. The reduction peak around 0.25 V (versus Li/Li+) is only showed in the first cycle can be responsible for the formation of the amorphous LixSi. During the first discharging process, the stable amorphous microstructure is formed, which causes large irreversible capacity loss. The pair of peaks A–a are due to the Li–C alloying/dealloying process. The pair of peaks B–b and D–d demonstrate that the reaction of Li–Si proceeds in a multi-step process. The pair of peaks C–c are associated with the reaction of the Li–Sn, which
Fig. 7. Cyclability and coulombic efficiency of the composite electrodes with a current density of 635 mA g SG2, (b) SGM.
1
under a voltage arrange from 0.03 to 1.50 V. (a) SGM, SG1 and
J. Wu et al. / Journal of Alloys and Compounds 596 (2014) 86–91
are not found on the CV diagram of SG composite [19]. Cyclic voltammetry cycles show the same profile except the first cycle indicating that these redox reactions were repeatable in the NSIPC of SGM electrode. Also, the area encircled by the redox reaction curve in the second and third cycles are all larger than the initial cycle, which suggests that the lithiation reaction is more completely in the later two cycles.
4. Conclusions In summary, a novel NSIPC of SGM anode material for lithiumion batteries was obtained via HEMM method. The SGM composite electrode exhibits better cyclability than the barely silicon/graphite composite electrodes due to the interpenetrating microstructure between silicon and ductile tin. The initial discharge specific capacity of the as-prepared SGM composite electrode is relatively high about 1790 mA h g 1 with 1592 mA h g 1 reversible capacity retention at a current density of 237 mA g 1 in the voltage arrangement from 0.03 V to 1.50 V. Additionally, the SGM composite electrode shows excellent rate capability and possesses stable cycling performance even at a large current density. Consequently, the relatively low cost and excellent electrochemical capability of the SGM composite would well meet the challenge in the rapid charge and discharge. SGM composite can be the promising candidate anode material for the next generation lithium rechargeable batteries.
Acknowledgements The financial support from the National Natural Science Foundation of China and Chinese Academy of Sciences are gratefully acknowledged.
91
References [1] M. Holzapfel, H. Buqa, L.J. Hardwick, M. Hahn, A. Würsig, W. Scheifele, P. Novák, R. Kötz, C. Veit, F.-M. Petrat, Electrochim. Acta 52 (2006) 973–978. [2] O.K. Park, Y. Cho, S. Lee, H.-C. Yoo, H.-K. Song, J. Cho, Energy Environ. Sci. 4 (2011) 1621. [3] B. Scrosati, J. Hassoun, Y.-K. Sun, Energy Environ. Sci. 4 (2011) 3287. [4] J.B. Goodenough, Y. Kim, Chem. Mater. 22 (2010) 587–603. [5] K. Kang, Y.S. Meng, J. Breger, C.P. Grey, G. Ceder, Science 311 (2006) 977–980. [6] C.-M. Hwang, J.-W. Park, Electrochim. Acta 56 (2011) 6737–6747. [7] J.Y. Howe, D.J. Burton, Y. Qi, H.M. Meyer, M. Nazri, G.A. Nazri, A.C. Palmer, P.D. Lake, J. Power Sources 221 (2013) 455–461. [8] Y.H. Xu, G.P. Yin, P.J. Zuo, Electrochim. Acta 54 (2008) 341–345. [9] H.K. Liu, Z.P. Guo, J.Z. Wang, K. Konstantinov, J. Mater. Chem. 20 (2010) 10055. [10] X. Xin, X. Zhou, F. Wang, X. Yao, X. Xu, Y. Zhu, Z. Liu, J. Mater. Chem. 22 (2012) 7724. [11] W.-J. Zhang, J. Power Sources 196 (2011) 13–24. [12] S. Rousselot, M. Gauthier, D. Mazouzi, B. Lestriez, D. Guyomard, L. Roué, J. Power Sources 202 (2012) 262–268. [13] M.-S. Park, Electrochem. Acta 50 (2005) 5561–5567. [14] J. Yang, B.F. Wang, K. Wang, Y. Liu, J.Y. Xie, Z.S. Wen, Electrochem. Solid-State Lett. 6 (2003) A154. [15] M.-S. Park, Y.-J. Lee, S. Rajendran, M.-S. Song, H.-S. Kim, J.-Y. Lee, Electrochim. Acta 50 (2005) 5561–5567. [16] Z.-J. Han, N. Yabuuchi, K. Shimomura, M. Murase, H. Yui, S. Komaba, Energy Environ. Sci. 5 (2012) 9014. [17] X. Shen, D. Mu, S. Chen, B. Xu, B. Wu, F. Wu, J. Alloys Comp. 552 (2013) 60–64. [18] K. Hanai, Y. Liu, N. Imanishi, A. Hirano, M. Matsumura, T. Ichikawa, Y. Takeda, J. Power Sources 146 (2005) 156–160. [19] W. Li, R. Yang, X. Wang, T. Wang, J. Zheng, X. Li, J. Power Sources 221 (2013) 242–246. [20] L. Zhang, X. Song, F. Wang, Q. Hu, Z. Sun, S. Yang, L. Wang, S. Sun, J. Solid State Electrochem. 16 (2012) 2159–2167. [21] H.-C. Tao, M. Huang, L.-Z. Fan, X. Qu, Electrochim. Acta 89 (2013) 394–399. [22] P. Zuo, W. Yang, X. Cheng, G. Yin, Ionics 17 (2010) 87–90. [23] Y.-L. Kim, Y.-K. Sun, S.-M. Lee, Electrochim. Acta 53 (2008) 4500–4504. [24] Y. Chen, J. Qian, Y. Cao, H. Yang, X. Ai, ACS Appl. Mater. Interfaces (2012). [25] M. Au, Y. He, Y. Zhao, H. Ghassemi, R.S. Yassar, B. Garcia-Diaz, T. Adams, J. Power Sources 196 (2011) 9640–9647. [26] J. Yao, Z. Jia, P. Zhang, C. Shen, J. Wang, K.-F. Aguey-Zinsou, L. Wang, Ionics 19 (2012) 401–407. [27] J.R. Szczech, S. Jin, Energy Environ. Sci. 4 (2011) 56.