Electrochimica Acta 83 (2012) 53–58
Contents lists available at SciVerse ScienceDirect
Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta
Fe3 O4 nanoparticles embedded in carbon-framework as anode material for high performance lithium-ion batteries Yang Yu a , Yongchun Zhu a,∗ , Huaxu Gong a , Yanmei Ma a , Xing Zhang a , Na Li a , Yitai Qian a,b,∗ a b
Hefei National Laboratory for Physical Science at Microscale and Department of Chemistry, University of Science and Technology of China, Hefei, Anhui 230026, PR China School of Chemistry and Chemical Engineering, Shandong University, Jinan 250100, PR China
a r t i c l e
i n f o
Article history: Received 9 April 2012 Received in revised form 2 August 2012 Accepted 2 August 2012 Available online 10 August 2012 Keywords: Magnetite Carbon-framework Lithium-ion battery High-rate performance
a b s t r a c t Fe3 O4 /C composites have been prepared by sucrose calcining with Fe3 O4 particles obtained from ferrous oxalate decomposition. The scanning electron microscopy (SEM) images show that Fe3 O4 nanoparticles (Fe3 O4 NPS) with average size of 200 nm are embedded in the three-dimensional (3D) carbon-framework. As an anode material for rechargeable lithium-ion batteries, the Fe3 O4 /C composite delivers a reversible capacity of 773 mAh g−1 at a current density of 924 mA g−1 after 200 cycles, higher than that of the bare Fe3 O4 NPS which only retain a capacity of 350 mAh g−1 . When the current density rises to 1848 mA g−1 , Fe3 O4 /C material still remains 670 mAh g−1 even after 400 cycles. The enhanced high-rate performance can be attributed to the 3D carbon-framework, which improves the electric conductivity, relaxes the strain stress and prevents the aggregation of Fe3 O4 particles during the charge/discharge process. © 2012 Elsevier Ltd. All rights reserved.
1. Introduction In the last few decades, graphite is mostly used as a commercial anode material in the lithium-ion batteries (LIBs) with a theoretical capacity of 372 mAh g−1 [1,2]. With the development of the high performance LIBs, transition metal oxides (MO, where M is Fe, Co, Ni or Cu, etc.) have been studied as a new series of anode materials due to their higher speciﬁc capacity compared with that of graphite [3–11]. Among these available alternative anode materials, magnetite (Fe3 O4 ) has always been regarded as an appealing candidate due to its high theoretical speciﬁc capacity (∼924 mAh g−1 ), as well as nontoxicity, high corrosion resistance and low processing cost . As reported in the literatures [13–15], the Fe3 O4 materials main follow the conversion reaction mechanism and are reduced to small metal clusters accompanying with the Li+ uptake and release. The electrochemical reactions can be described as follows: disch arge
Fe3 O4 + 8Li+ + 8e −→ 3Fe + 4Li2 O charge
8Li+ + 8e −→ 8Li
Fe3 O4 + 8Li ↔ 3Fe + 4Li2 O
Fe3 O4 based anodes undergo a signiﬁcant volume change, resulting in large potential hysteresis, capacity fading and poor
cycling performance . In nowadays, carbon coating is known as one of simplest and the most effective strategies in improving the electric conductivity and restrain volume change during the charge/discharge process. For instance, a dispersed Fe3 O4 nanospindle coated with carbon can remain 530 mAh g−1 after 80 cycles at a current density of 460 mA g−1 . The Fe3 O4 /C nanoﬁbers exhibit a reversible capacity of 1000 mAh g−1 after 80 cycles at 200 mA g−1 . Fe3 O4 /C core–shell nanospheres present a capacity of 636 mAh g−1 over 50 cycles at 1000 mA g−1 . Graphene sheets modiﬁed Fe3 O4 NPS deliver a capacity of 550 mAh g−1 even after 300 cycles at 1000 mA g−1 . Herein, we prepared Fe3 O4 /C composites in which Fe3 O4 NPS were embedded in the 3D carbon-framework. The composites can deliver a high reversible capacity of 773 mAh g−1 at 924 mA g−1 after 200 cycles, together with a capacity of 670 mAh g−1 at a higher current density of 1848 mA g−1 until the 400th cycle. While, without carbon coating, the bare Fe3 O4 NPS can only deliver 350 mAh g−1 up to 200 cycles at a current density of 924 mA g−1 . The remarkable high-rate performance of the composites indicates its promising application as anode material for lithium-ion batteries. 2. Experimental 2.1. Preparation of Fe3 O4 NPS
∗ Corresponding authors at: Hefei National Laboratory for Physical Science at Microscale and Department of Chemistry, University of Science and Technology of China, Hefei, Anhui 230026, PR China. Tel.: +86 0551 360 1589; fax: +86 551 360 7402. E-mail addresses: [email protected]
(Y. Zhu), [email protected]
(Y. Qian). 0013-4686/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.electacta.2012.08.003
In a typical synthesis, a homogeneous solution containing FeSO4 ·7H2 O 8 mmol (2.224 g) and 5.043 g citric acid was ﬁrst prepared in 30 mL distilled water, meanwhile, another solution was prepared by dissolving H2 C2 O4 ·H2 O 10 mmol (1.08 g) in 10 mL
Y. Yu et al. / Electrochimica Acta 83 (2012) 53–58
distilled water, then, the oxalic acid solution was added to the former solution under continuous stirring. Subsequently, the yellow ferrous oxalate precipitate was ﬁltered and washed. Ferrous oxalate precipitate and 5 mL ethanol were sealed in a 20 mL stainless steel autoclave and calcined at 550 ◦ C for 5 h. After the autoclave cooled down, the black powders were washed with absolute ethanol and distilled water for several times. 2.2. Preparation of Fe3 O4 /C composites The obtained Fe3 O4 powders were milled with sucrose (Fe3 O4 /sucrose = 3:1, w/w) in a ball mill (400 r/s, 6 h). Then, the mixture was annealed in a quartz tube with a slow ramping rate of 2 ◦ C/min to 600 ◦ C for 5 h in Ar atmosphere. The Fe3 O4 /C composites were obtained. 2.3. Characterization X-ray powder diffraction (XRD) patterns of the products were recorded on a Philips X’pert X-ray diffractometer with Cu Ka ˚ X-ray photoelectron spectra (XPS) were radiation ( = 1.54178 A). tested on a VGESCA-LAB MKII X-ray photoelectron spectrometer, using non-monochromated Mg K␣ X-ray radiation as the excitation source. Raman spectrum was carried out on a JYLABRAM-HR Confocal Laser Micro-Raman spectrometer with 514.5 nm from an argon laser at room temperature. The scanning electron microscopy (SEM) images were taken by using a JEOL-JSM-6700F ﬁeld-emitting (FE) scanning electron microscope. The high-resolution transmission electron microscope (HRTEM) images were taken on a JEOL 2010 HRTEM at an acceleration voltage of 200 kV. EIS were performed by a Zahner Elektrik IM6 (Germany) impedance instrument over the frequency range from 100 kHz to 0.01 Hz. The active materials, acetylene black and poly (vinylidene diﬂuoride) (PVDF) with a weight ratio of 60:30:10 were mixed homogeneously with N-methyl-2-pyrrolidone (NMP), the obtained slurry was coated on a copper foil and dried at 100 ◦ C for 12 h in vacuum. The coin cells (size: 2016) were assembled in an argonﬁlled glove box with lithium foil as the anode, celgard 2400 as the separator, and a solution of 1.0 M LiPF6 in ethylene carbonate (EC)/diethyl carbonate (DEC) (1:1 by volume) as the electrolyte. The total loading of active material in the electrode is 1.77 mg cm−2 . The galvanostatic charge and discharge tests were carried out on a LAND-CT2001A instrument in the potential range of 0.01–3.00 V (versus Li/Li+ ). The cells were charged/discharged at 25 ◦ C in a DH3600 electro-thermal constant-temperature incubator (Taisite, China). 3. Results and discussion The crystallographic structures of the samples are identiﬁed by XRD. Fig. 1 shows the typical XRD patterns of the as-obtained products. The diffraction peaks of Fe3 O4 NPS (Fig. 1a) and the Fe3 O4 /C composite (Fig. 1b) can be readily indexed as face-centered cubic structure Fe3 O4 with a space group of Fd3m (JCPDS Card No: 750033). From Fig. 1b, no diffraction peak for carbon can be observed. In addition, in our XRD patterns, there are no typical ␥-Fe2 O3 (JCPDS Card No: 39-1346) peaks such as (1 1 0), (2 1 0), (2 1 1) (the intensities of these peaks are even more than (1 1 1) existing in the obtained XRD patterns), indicating the absence of ␥-Fe2 O3 . The representative Raman spectrum of the as-obtained Fe3 O4 /C composite is used to investigate the presence of carbon. As shown in Fig. 2a, the two peaks at 1350 and 1587 cm−1 are the characteristic peaks corresponding to the D-band and G-bond of graphite . D-band is associated with the vibration of carbon atoms with dangling bonds in-plane terminations of disordered graphite, and G-bond is related to the vibration of sp2 -bonded carbon atoms in a
Fig. 1. Typical XRD patterns of the as-obtained products (a) Fe3 O4 nanoparticles (b) the Fe3 O4 /C composite.
2D hexagonal lattice. Here, the ratio of ID /IG is 0.757, implying poor crystallinity of carbon . The composition of the as-obtained composite is further identiﬁed by XPS. Fig. 2b is the wide scan of product, and the sharp peaks of the C1s, O1s and Fe2p indicate the existence of carbon, oxygen and iron elements in the composite. Fig. 2c is the binding energies of Fe2p, the peaks located at 711 and 724.5 eV corresponding to Fe2p3/2 and Fe2p1/2 , respectively, which are in good agreement with the reported values of Fe3 O4 in the literatures [23,24]. The morphologies and structures of the as-obtained products are characterized by SEM and TEM. The typical SEM image of Fe3 O4 NPS is shown in Fig. 3a, which shows the particles with an average size of 200 nm. Fig. 3b is a representative SEM image at a low magniﬁcation of the composite. It indicates that the main products consist of some particles embedded in framework. Fig. 3c is the TEM image of the composite, which shows strong contrast between the light and dark parts in the enlarged area (inset in Fig. 3c). The HRTEM image (Fig. 3d) of the dark part in Fig. 3c shows clear crystal lattices with d-spacing of 0.294 nm, corresponding to the (2 2 0) plane of the face-centered cubic Fe3 O4 crystals. The light part is considered as carbon. To further investigate the morphology of carbon, we remove Fe3 O4 NPS by stirring the Fe3 O4 /C composites in the concentrated hydrochloric acid for 12 h. The left black powders are carbon. Fig. 3e is the SEM image of the as-remained carbon. It is found that the square area is the position of Fe3 O4 NPS and carbon layers interconnect to each other to form a 3D framework structure. Fig. 3f is the TEM image of the carbon-framework. The carbon edge and the void area indicate that the Fe3 O4 NPS are actually embedded uniformly in the interconnected carbon-framework. Sucrose can mix with Fe3 O4 particles uniformly after a long-milling. When the temperature is elevated, sucrose is gradually carbonized in situ. The Fe3 O4 particles disperse in a carbon matrix after the annealing process to form the 3D Fe3 O4 /C composite. The similar process has been reported in other researches . The as-prepared Fe3 O4 NPS and Fe3 O4 /C composite materials are assembled into coin cells to investigate their electrochemical behaviors. As shown in Fig. 4a, the charge/discharge proﬁles of the Fe3 O4 /C electrodes are tested at a current density of 92 mA g−1 . The ﬁrst discharge curve is with a long voltage plateau at about 0.75 V versus Li/Li+ , which is close to that reported in the literature for Fe3 O4 /C anodes  and could be attributed to the reduction of Fe2+ /Fe3+ to Fe0 , and the discharge speciﬁc capacity is as high as ∼1360 mAh g−1 . The over discharge capacity is approximately consistent with the sloping voltage below 0.37 V. It is attributed to the formation of the SEI ﬁlm and further lithium consumption via interfacial reactions . The voltage plateau of the ﬁrst charge curve is present at about 1.8 V, corresponding to the reversible
Y. Yu et al. / Electrochimica Acta 83 (2012) 53–58
Fig. 2. (a) Raman spectrum of the Fe3 O4 /C composite; XPS survey spectra of the Fe3 O4 /C composite: (b) wide scan and (c) Fe2p.
oxidation of Fe0 to Fe2+ /Fe3+ . The charge/discharge curves of the following cycles (2nd–5th) appear to overlap and the obtained speciﬁc capacity remains unchanged and stables at 924 mAh g−1 , which is almost consistent with the theoretical value. Similarly, the wide voltage plateau of these charge curves centered at 1.8 V is observed, while the plateau of the discharge curves is present at 1.0 V. Fig. 4b is the cycle performances of the bare Fe3 O4 NPS and Fe3 O4 /C composites. The electrodes are ﬁrstly tested at a current density of 92 mA g−1 for ﬁve cycles, then at 924 mA g−1 until the 200th cycle. Finally, the current density returns to 92 mA g−1 . Although the initial discharge capacity of the Fe3 O4 NPS electrode is as high as 1432 mAh g−1 , it decreases to 660 mAh g−1 at the 5th cycle and fades to 200 mAh g−1 immediately as the density rises to 924 mA g−1 , which is only about 71.4% and 21.6% of the theoretical speciﬁc capacity, respectively. It is note that the capacity can stabilize at ∼350 mAh g−1 (37.9% of the theoretical speciﬁc capacity) after the density returns to 92 mA g−1 , which may be attributed to its high crystallinity . By comparison, the ﬁrst ﬁve cycles of the Fe3 O4 /C composites is the same as Fig. 4a. It still can deliver a capacity of ∼924 mAh g−1 at the 5th cycle. As the density rises to 924 mA g−1 , the cell experiences a gradual capacity fading process. After the 40th cycle, the capacity rises steadily and reaches 773 mAh g−1 at the 200th cycle which is ∼83.7% of the theoretical capacity of Fe3 O4 and more than double theoretical capacity of graphite (372 mAh g−1 ). The stage of the capacity decrease at the initial 40 cycles may be explained by the structure re-organization of the carbon coatings . The phenomenon of the gradual increased capacity is attributed to the reversible growth of a polymeric gel-like ﬁlm resulting from kinetically activated electrolyte degradation, which is well-documented in the literatures [31–33] and similar results have been reported for many transition metal oxides [34,35]. In addition, the reversible capacity of the Fe3 O4 /C composites reaches about 1100 mAh g−1
after 200 cycles, which indicates the good cyclic stability of the composites. Rate-performance of the Fe3 O4 /C composite is further investigated. As shown in Fig. 5a, after 5 cycles, the current density is gradually increased with a few cycles at current densities of 924, 1848 and 2310 mA g−1 . The reversible capacities are 700, 540 and 500 mAh g−1 , which are 75.8%, 58.4% and 54.1% of the theoretical speciﬁc capacity, respectively. More importantly, when the current density is reduced to 924 and then 92 mA g−1 , the capacities swiftly approach to the same values which obtained at the same density in the previous cycles. Note that even at the highest current density of 2310 mA g−1 , the speciﬁc capacity is still higher than the theoretical capacity of graphite. We further test the cyclic stability of the Fe3 O4 /C composites at a higher density. A new cell is cycled at 92 mA g−1 for 3 cycles, followed by cycling at 1848 mA g−1 for 400 cycles (Fig. 5b). The reversible speciﬁc capacity is around 670 mAh g−1 (72.5% of the theoretical speciﬁc capacity) with virtually no capacity loss for the cycles, except the stage of the capacity decrease at the former 40 cycles which is the same as the phenomenon in Fig. 5b. The coulombic efﬁciency (the black dots in Fig. 5b) maintains consistently at ∼98% throughout the cycling. To gain further understanding of the 3D carbon-framework, the electrochemistry impedance spectra (EIS) of the Fe3 O4 particles and Fe3 O4 /C electrodes are investigated. Fig. 6 shows the typical Nyquist plots of the cells after 10 cycles at 924 mA g−1 . The intercept at the real (Z ) axis in high frequency range corresponds to the ohmic resistance (Re ). The semicircle in the middle frequency indicates the charge transfer resistance (Rct ). The inclined straight line relates to the Warburg impedance (Zw ). It is reported that the cell impedance is mainly determined by Rct . As can be seen in Fig. 6, the semicircle diameter of the Fe3 O4 /C composite is much smaller than that of Fe3 O4 particles. The result of the EIS analysis indicates that carbon-framework has an important
Y. Yu et al. / Electrochimica Acta 83 (2012) 53–58
Fig. 3. (a) SEM image of Fe3 O4 particles; (b) SEM image, (c) TEM image and (d) HRTEM image of the Fe3 O4 /C composite; (e) SEM image and (f) TEM image of the carbonframework; the inset in (c) is the enlarged TEM image of the selected area.
role in reducing the Rct of cells during charging/discharging. The good electrochemical performance of the Fe3 O4 /C composite is believed to be attributed to the unique carbon-framework structure. Firstly, the 3D carbon-framework not only remarkably enhances the electric conductivity, but also provides continuous
paths between Fe3 O4 NPS, thus ensures the fast and continuous transportation of electrons between Fe3 O4 NPS and carbon, which is favorable for electrons moving unimpeded over particles to attain a high rate capability . Secondly, the interconnected carbon layers and Fe3 O4 NPS build a special micro-nanostructure, which can
Fig. 4. (a) Charge/discharge proﬁles of the Fe3 O4 /C composite at a current density of 92 mA g−1 . (b) Cycle performance of Fe3 O4 /C composite and the bare Fe3 O4 particles at a current density of 924 mA g−1 for 200 cycles.
Y. Yu et al. / Electrochimica Acta 83 (2012) 53–58
Fig. 5. (a) Rate performance of the Fe3 O4 /C composite at various current densities. (b) Cycle performance and Coulomb efﬁciency of the Fe3 O4 /C composite at a current density of 1848 mA g−1 (the front three cycles at 92 mA g−1 ).
Acknowledgments This work was ﬁnancially supported by the National Natural Science Fund of China (No. 91022033), the 973 Project of China (No. 2011CB935901), the Fundamental Research Funds for the Central Universities (No. WK 2340000027) and Anhui Provincial Natural Science Foundation (1208085QE101). References
Fig. 6. Impedance spectra of the anodes obtained after 10 cycles.
stable the composite, prevent the active materials aggregation or pulverization and effectively improves the cyclic performance of materials . Thirdly, the uniformity of carbon distribution based on the framework can relax the strain stress, buffer the volume expansion and hence improve the cyclic stability of the materials. The above three important roles favor the high reversible capacity and superior cyclic performance of the Fe3 O4 /C anode material in lithium-ion batteries.
4. Conclusions In summary, we have prepared Fe3 O4 /C composites in which Fe3 O4 NPS embedded in the 3D carbon-framework. As an anode material for rechargeable lithium-ion batteries, the Fe3 O4 /C composites deliver a reversible capacity of 773 mAh g−1 at a current density of 924 mA g−1 , while the Fe3 O4 NPS only deliver 350 mAh g−1 after 200 cycles. More importantly, the Fe3 O4 /C composites deliver a high reversible capacity (670 mAh g−1 ) at a higher current density (1848 mA g−1 ) until the 400th cycle. The high-rate capability and good cyclic performance may be attributed to the 3D conductive carbon-framework, which can favor fast electrons transportation and high structural stability during the reversible charge/discharge process. This study suggests that the optimized Fe3 O4 /C composite is a promising anode material for high-power rechargeable lithiumion batteries. Meanwhile, it provides an approach to prepare the other carbon-coated materials. The further work is in progress.
 J.R. Dahn, T. Zheng, Y.H. Liu, J.S. Xue, Science 270 (1995) 590.  K. Sato, M. Noguchi, A. Demachi, N. Oki, M. Endo, Science 264 (1994) 556.  P. Poizot, S. Laruelle, S. Grugeon, L. Dupont, J.M. Tarascon, Nature 407 (2000) 496.  Z.Y. Wang, D.Y. Luan, S. Madhavi, Y. Hu, X.W. (David) Lou, Energy and Environmental Science 5 (2012) 5252.  W.J. Yu, P.X. Hou, L.L. Zhang, F. Li, C. Liu, H.M. Cheng, Chemical Communications 46 (2010) 8576.  B.J. Li, H.Q. Cao, J. Shao, M.Z. Qu, J.H. Warner, Journal of Materials Chemistry 21 (2011) 5069.  P.C. Lian, X.F. Zhu, H.F. Xiang, Z. Li, W.S. Yang, H.H. Wang, Electrochimica Acta 56 (2010) 834.  W.Y. Li, L.N. Xu, J. Chen, Advanced Functional Materials 15 (2005) 851.  Z.Y. Wang, D.Y. Luan, F.Y.C. Boey, X.W. (David) Lou, Journal of the American Chemical Society 133 (2011) 4738.  J. Chen, L.N. Xu, W.Y. Li, X.L. Gou, Advanced Materials 17 (2005) 582.  L.M. Jin, Y.C. Qiu, H. Deng, W.S. Li, H. Li, S.H. Yang, Electrochimica Acta 56 (2011) 9127.  S. Ito, K. Nakaoka, M. Kawamura, K. Ui, K. Fujimoto, N. Koura, Journal of Power Sources 146 (2005) 319.  J.S. Chen, Y.M. Zhang, X.W. (David) Lou, ACS Applied Materials Interfaces 3 (2011) 3276.  G.M. Zhou, D.W. Wang, F. Li, L.L. Zhang, N. Li, Z.S. Wu, L. Wen, G.Q. Lu, H.M. Cheng, Chemistry of Materials 22 (2010) 5306.  S.M. Yuan, J.X. Li, L.T. Yang, L.W. Su, L. Liu, Z. Zhou, ACS Applied Materials Interfaces 3 (2011) 705.  Z.M. Cui, L.Y. Jiang, W.G. Song, Y.G. Guo, Chemistry of Materials 21 (2009) 1162.  W.M. Zhang, X.L. Wu, J.S. Hu, Y.G. Guo, L.J. Wan, Advanced Functional Materials 18 (2008) 3941.  L. Wang, Y. Yu, P.C. Chen, D.W. Zhang, C.H. Chen, Journal of Power Sources 183 (2008) 717.  J. Liu, Y.C. Zhou, F. Liu, C.P. Liu, J.B. Wang, Y. Pan, D.F. Xue, RSC Advances 2 (2012) 2262.  J.S. Zhou, H.H. Song, L.L. Ma, X.H. Chen, RSC Advances 1 (2011) 782.  A.C. Ferrari, J. Robertson, Physical Review B 61 (2000) 14095.  T. Mei, T. Li, H.Y. Bi, L.B. Wang, Y.C. Zhu, Y.T. Qian, European Journal of Inorganic Chemistry 27 (2010) 4314.  T.Z. Yang, C.M. Shen, H.T. Yang, C.W. Xiao, Z.C. Xu, S.T. Chen, D.X. Shi, H.J. Gao, Surface and Interface Analysis 38 (2006) 1063.  J.H. Zhang, Q.H. Kong, J. Du, D.K. Ma, G.C. Xi, Y.T. Qian, Journal of Crystal Growth 308 (2007) 159.  X.L. Wu, L.Y. Jiang, F.F. Cao, Y.G. Guo, L.J. Wan, Advanced Materials 21 (2009) 2710.  X.Y. Li, X.L. Huang, D.P. Liu, X. Wang, S.Y. Song, L. Zhou, H.J. Zhang, Journal of Physical Chemistry C 115 (2011) 21567.  T. Zhu, J.S. Chen, X.W. (David) Lou, Journal of Physical Chemistry C 115 (2011) 9814.  J. Jamnik, J. Maier, Physical Chemistry Chemical Physics 5 (2003) 5215.  E. Hosono, T. Kudo, I. Honma, H. Matsuda, H.S. Zhou, Nano Letters 9 (2009) 1045.
Y. Yu et al. / Electrochimica Acta 83 (2012) 53–58
 J.S. Gnanaraj, M.D. Levi, E. Levi, G. Salitra, D. Aurbach, J.E. Fischer, A. Claye, Journal of the Electrochemical Society 148 (2001) A525.  S. Laruelle, S. Grugeon, P. Poizot, M. Dolle, L. Dupont, J.M. Tarascon, Journal of the Electrochemical Society 149 (2002) A627.  S. Grugeon, S. Laruelle, L. Dupont, J.M. Tarascon, Solid State Sciences 5 (2003) 895.  J.S. Do, C.H. Weng, Journal of Power Sources 146 (2005) 482.
 S.L. Chou, J.Z. Wang, D. Wexler, K. Konstantinov, C. Zhong, H.K. Liu, S.X. Dou, Journal of Materials Chemistry 20 (2010) 2092.  C.C. Li, X.M. Yin, L.B. Chen, Q.H. Li, T.H. Wang, Applied Physics Letters 97 (2010) 043501.  C.H. Chen, J. Liu, K. Amine, Journal of Power Sources 96 (2001) 321.  X.H. Huang, J.P. Tu, C.Q. Zhang, X.T. Chen, Y.F. Yuan, H.M. Wu, Electrochimica Acta 52 (2007) 4177.