Porous micro-spherical LiFePO₄/graphene nanocomposites for high-performance Li ion battery cathode materials

Abstract

In this work, we designed a novel cathode material, porous micro-spherical LiFePO₄/graphene nanocomposites with low graphene content, via a hydrothermal route followed by a lithiation process. The nanocomposites consist of LiFePO₄ microspheres (∼3 μm) with their primary particles (∼100 nm) sandwiched between layers of graphene nanosheets. Mesopores (∼25 nm) are widely present in the microspheres. TGA analysis indicates that only 2.65% graphene was loaded in the composites. When studied as a cathode material for Li-ion batteries, with the absence of a conductive agent, the nanocomposites can display a high capacity of 118.2 mA h g⁻¹ at 20 C.

---

1 Introduction

Since the early 1990s, rechargeable lithium ion batteries (LIBs) have been enjoyed by people all over the world in daily life due to their high energy density, long cycle life and great convenience. For new generations of LIBs, they are applied not only to consumer electronics but also to large electrical appliances such as power stations, electric vehicles (EVs) and hybrid electric vehicles (HEVs). The key point of further breakthrough in the use of LIBs is electrode materials especially cathode materials, because cathode materials not only directly affect the performance of LIBs but also determine the cost of the battery.¹ Among reported cathode materials, LiFePO₄ (LFP), with many appealing features, such as long cycle life, high capacity (170 mA h g⁻¹), low cost, suitable voltage (3.4 V versus Li⁺/Li), environmental friendliness and high thermal stability,^1–8 has been considered as a promising cathode candidate for the next generation LIBs. However, the intrinsically low electronic conductivity, sluggish Li⁺ transport and low tap density degrade the rate performance and energy density of LFP, and thus limit its further application.^1–8

Considerable effort has been made to overcome the ionic and electronic transport limitations by cationic doping,^9,10 decreasing the particle size,^1,2,8,11,12 or coating with conductive materials.^13,14 Particularly, decreasing the LFP particle size to the nanometer level is now widely accepted to enhance Li ion diffusion, while enwrapping or coating LFP particles with carbonaceous materials can improve their conductivity. Among such carbonaceous materials, graphene attracts people's attention most, due to its numerous intriguing properties such as high specific surface area, excellent conductivity and good mechanical flexibility,^15–17 which make it suitable for fabrication into high performance composites with other electrode materials for LIBs. For LFP, the carbon content can greatly affect the final tap density. Because of the low density of carbonaceous materials, on the one hand, too much carbon will seriously decrease the tap density of LFP; however, too little carbon cannot guarantee good electronic conductivity and thus can diminish the electrochemical performance. Hence, it is necessary to optimize the carbon content. However, in many of the reported LFP/graphene nanocomposites, either the graphene content is too high (>10%),^18,19 or a high weight ratio of conductive agent was used in the electrodes,^20,21 which greatly decreases the tap density and energy density of LFP. Therefore, it is necessary to control the graphene content at a low value under the premise of good conductivity. Of course, it is a challenge to scientists. To increase the tap density of LFP, controlling its morphology is another possible way. For a powder material, the tap density depends not only on the theoretical density, but also on the particle morphology, including particle size, geometric shape, and particle distribution. As for LFP, micro-sized spherical particles exhibit higher tap density than irregular particles because they avoid a lot of vacant space between the particles and bad fluidity of the powders.⁸ For example, we adopted a hydrothermal method to prepare micro-sized spherical LFP that shows a tap density of 1.5 g cm^{−3 2}; the tap density of spherical LFP microspheres reported by Xie et al.¹³ reaches up to 1.8 g cm⁻³. These values are much higher than the actual tap density of most LFP materials (∼1 g cm⁻³).² However, the micro-sized spherical LFP suffers another drawback, which is the poor rate performance because of the following reasons: (1) the long Li ion transport pathway from the inner part to the surface of the sphere and (2) the poor conductivity of the inner part of the LFP sphere as conductive materials are usually coated on the surface.

In order to overcome these two drawbacks, we developed a hydrothermal method followed by a lithiation process to design and synthesize porous micro-spherical LFP/graphene nanocomposites (LFP/GNs) with a graphene content of only 2.65%. In the LFP/GNs, most LFP primary nanoparticles were sandwiched between layers of graphene nanosheets, thus constructing a perfect interconnected conductive network. Therefore, when tested as cathode materials for LIBs, with the absence of a conductive agent, at a high current rate of 20 C, the cathode can still deliver a capacity of 118.2 mA h g⁻¹. Since the synthetic method is simple and can be easily scaled up and the graphene content is low, which greatly reduces the cost of the composite cathode, it is reasonable to claim that the as-designed LFP/GNs are a promising possible cathode candidate for the next generation Li-ion batteries.

2 Experiments

Preparation of materials

Micro-sized FePO₄/graphene composites (FP/GNs) were synthesized via a hydrothermal route. Typically, 10 mmol Fe(NO₃)₃·9H₂O (≥99.5%, Aladdin) was first dissolved in 40 ml GO (1.5 mg ml⁻¹) solution, and the GO solution was prepared according to a method reported elsewhere.²² Thereafter, the solution was stirred and ultrasonically exposed for 10 min, and then 10 mmol NH₄H₂PO₄ (≥99.5%, Aladdin) was introduced to the above solution under stirring. After ultrasonic treatment for another 10 min, the mixture was transferred into a 50 ml PTFE inner steel autoclave and heated at 180 °C for 10 h. The autoclave was naturally cooled to room temperature. The obtained FP/GNs slurry was washed with de-ionized water several times and dried under vacuum at 80 °C.

The lithiation process was achieved as follows: 10 mmol LiOH·H₂O was dissolved in 10 ml de-ionized water, and then the as-prepared FP/GN powder was added. The mixture solution was ultrasonically exposed for 1 min and then vaporized in vacuum at 60 °C. The obtained mixture powder was grounded carefully with a mortar and a pestle and then transferred to a tube furnace, where it was annealed under a reducing atmosphere. The tube was purged with 10% hydrogen and 90% argon for 0.5 h before heating started (flow rate: 150 cm³ min⁻¹). The powders were first heated at the rate of 5 °C min⁻¹ to 300 °C and kept for 1 h. Secondly, the temperature was increased to 700 °C at the same heating rate and held for 10 h. Finally, LFP/GNs were obtained after natural cooling to ambient temperature.

Characterizations

The X-ray diffraction (XRD) patterns of the sample were recorded by a Rigaku Dmax 2200 X-ray diffractometer with Cu-Kα radiation (λ = 1.5416 Å). The morphologies of the samples and the elemental maps were obtained using a scanning electron microscope (SEM, Hitachi S-7500, 5.0 kV) and energy dispersive spectrum (EDS). TEM investigation was carried out by a JEOL JEM-2100F microscope at the acceleration voltage of 20 kV. Specific surface areas were measured at 77 K by Brunauer–Emmett–Teller (BET) N₂ adsorption–desorption (NOVA 2200e, Quantachrome, USA), and pore size distributions were calculated from the desorption branch of the N₂ adsorption isotherm using the Barrett–Joyner–Halenda (BJH) formula. Raman spectra were obtained from a RM-1000 Renishaw confocal Raman micro-spectroscope with 514.5 nm laser radiation at a laser power of 0.48 mW in the range of 500–2000 cm⁻¹. TG was performed in an oxygen atmosphere using a Pyris Diamond TG/DTA (PerkinElemer Inc., USA). The samples were heated from 50 °C to 650 °C at 10 °C min⁻¹. The tap density of the LFP/GNs was measured on a type ZS-102 tap density meter (Liaoning Instrument Research Institute Co. Ltd.).

Electrochemical performance

Electrochemical reactions of samples with lithium were investigated using a simple two-electrode cell. The working electrode consists of 80 wt% LFP/GN active material, 15% carbon black as the conductive agent, 5 wt% polyvinylidene fluoride (PVDF) as the binder, and Al foils as the substrate (current collector). A N-methyl pyrrolidinone (NMP) slurry consisting of the above mixture was uniformly coated on an aluminum disk 14 mm in diameter. The disk electrodes were dried overnight at 60 °C under vacuum followed by compression at 1.0 × 10⁶ Pa. The 2016 type coin cells were assembled in an Ar-filled glove box using polypropylene (PP) micro-porous film as the separator, a solution of 1 M LiPF₆ in ethylene carbonate (EC)/dimethyl carbonate (DMC) (1 : 1, v/v) as the electrolyte and metallic lithium foil as the counter electrode. The electrochemical tests were performed on Land CT2001A battery testing systems (Jinnuo Electronics Co. Ltd., China). The charging and discharging tests were performed galvanostatically at various currents and at a constant temperature of 25 °C in the voltage range of 2.0 to 4.2 V. Cyclic voltammograms (CVs) were obtained on a CHI660D electrochemical workstation (Shanghai Chenhua Co. Ltd., China) at a scan rate of 0.2 mV s⁻¹, and the electrochemical impedance measurements were also performed on the CHI660D electrochemical workstation with an AC voltage of 5 mV amplitude in the 100 kHz to 0.01 Hz range immediately as the electrodes ran over 200 cycles (at 2.0 V).

3 Results and discussion

FP/GNs were prepared by an in situ hydrothermal method, and LFP/GNs were further obtained via a high-temperature chemical lithiation process. The synthetic mechanism is schematically illustrated in Fig. 1. In brief, Fe³⁺ ions dissolved in the solution were first anchored on graphene oxide (GO) nanosheet surfaces by the electrostatic interaction between Fe³⁺ ions and the negatively charged GO nanosheet surfaces, because GO nanosheets contain epoxy and hydroxyl groups on the basal planes and carboxylic acid groups on the edges.²³ After that, PO₄³⁻ was introduced into the solution. During the subsequent hydrothermal process, FePO₄ nanoparticles were formed and assembled into microspheres with GO nanosheets simultaneously embedded, and the driving force is the reduction of the surface tension of the highly dispersed FePO₄ nanoparticles. After a high temperature lithiation process, the FePO₄/GN precursor was transformed to LFP/GNs.

Fig. 1 Schematic illustration of the synthetic procedure for LFP/GNs.

Fig. 2 shows the XRD pattern of FePO₄/GNs, from which it can be seen that well crystallized FePO₄ with high purity was synthesized (matches with JCPDS card no. 29-715). No graphene-related peaks can be seen, and this is because of the low content of graphene in the composites (2.65% determined by TGA in later section).²⁴

Fig. 2 XRD pattern of FP/GNs.

Fig. 3a shows a SEM image of FP/GNs, from which it can be found that the as-prepared product is formed with a spherical morphology. The mono-dispersed spheres have a particle size in the range of 1–3 μm in diameter. In fact, the diameter of more than 80% of these spheres is around 3 μm. In addition, no graphene nanosheets are found in the SEM field of vision, indicating that they may be embedded in the FePO₄ microspheres. Fig. 3b shows an individual FP/GN microsphere. The EDS elemental mappings (Fig. 3c–f) of the FP/GN microsphere show a uniform distribution of C, Fe, O and P, which confirms the existence and homogeneous distribution of graphene in the FePO₄ microspheres.

Fig. 3 (a) Low magnification SEM image of FP/GN microspheres, (b) SEM image of an individual FP/GN microsphere, (c)–(f) EDS maps corresponding to the C, Fe, O, P elements of an individual FP/GN microsphere.

After the high-temperature chemical lithiation process, orthorhombic phase LFP (space group Pnma, JCPDS card no.: 81-1173) with high purity was formed, as confirmed by the XRD pattern shown in Fig. 4a. The average particle size, determined based on the peak broadening, is 96 nm. The reduction of GO nanosheets during the hydrothermal and calcination process was characterized by Raman spectroscopy, as it is a powerful tool to characterize carbonaceous materials.^25,26 The significant structural changes from GO to graphene are reflected in their Raman spectra, as shown in Fig. 4b. The Raman spectrum of GO contains both a G band (1587 cm⁻¹, E_2g phonon of C sp² atoms) and D bands (1358 cm⁻¹, κ-point phonons of A_1g symmetry).²⁷ The Raman spectra of the LFP/GNs also contain both G and D bands. Moreover, the small weak peak at around 950 cm⁻¹ is caused by the intramolecular stretching of the PO₃⁻⁴ anion.²⁸ The D/G intensity ratios are 0.92 for GO and 1.46 for LFP/GNs, respectively. Obviously, the D/G intensity ratio for LFP/GNs is larger than that for GO, which suggests a decrease in the average size of the sp² domains upon reduction of the exfoliated GO.²⁹

Fig. 4 (a) XRD pattern of LFP/GNs, (b) Raman spectra of LFP/GNs and GO solution.

The morphology of the LFP/GNs is shown in Fig. 5a–c. Fig. 5a shows an overview of the composites, and it can be seen that after the high temperature lithiation process most of the LFP/GNs still retain the spherical morphology of their precursor, although a few broken spheres exist. Fig. 5b is a close view of an individual LFP/GN microsphere, and it clearly shows that small pores widely exist in the microsphere. The porous structure of LFP/GNs is further confirmed by the adsorption–desorption isotherms of the sample, which indicate a typical hysteresis mesoporous system, as shown in Fig. 5d. According to the BET analysis, a total specific surface area of 54.39 m² g⁻¹ is obtained. The Barrett–Joyner–Halenda (BJH) pore-size distribution, shown in the inset of Fig. 5d, indicates the LFP/GNs have an average pore diameter of about 25 nm. Fig. 5c is the partially enlarged image of Fig. 5b, and it can be seen that the LFP/GN microsphere has a primary LFP particle size of ∼100 nm, which is consistent with the XRD analysis. More importantly, it can be observed that graphene nanosheets are attached on the primary LFP nanoparticles.

Fig. 5 (a) SEM image of the LFP/GN microspheres, (b) SEM image of an individual LFP/GN microsphere, (c) partially enlarged image of (b), (d) nitrogen adsorption–desorption isotherms of the LFP/GNs. Inset: the pore size distribution plot calculated by the BJH formula in the desorption branch isotherm.

For a further verification, the TEM image of a broken LFP/GNs sphere is supplied. From Fig. 6 it can be clearly observed that LFP primary nanoparticles with a diameter of ∼100 nm are anchored on graphene nanosheets. The introduction of graphene dramatically enhanced the electrical conductivity of the nanocomposites to 9.9 S cm⁻¹, which is nearly 10 orders of magnitude higher than that of pure LFP (10⁻⁹ S cm⁻¹). The greatly improved electrical conductivity of LFP also implies the well constructed graphene conducting network. Based on the synthetic mechanism as well as SEM and TEM analysis, we found that the LFP primary particles are not likely wrapped by graphene nanosheets but more likely sandwiched between graphene nanosheets. In fact, such a composite is better than the former in terms of a high rate performance, because lithium ions cannot pass through the carbon atomic arrays in two-dimensional sheets of graphene. Therefore, in the wrapped systems lithium ions have to tortuously detour to reach the electrolyte, which results in a longer diffusion distance and undermines the high rate performance of the electrode.³⁰

Fig. 6 TEM image of a broken LFP/GN microsphere.

The graphene loading of LFP/GNs was determined by TGA as shown in Fig. 7. In order to completely remove the graphene component, the LFP/GNs were sintered under flowing oxygen with a temperature range from 50 to 650 °C. For the LFP/GNs, the graphene in the composites was oxidized to CO₂ gas at 460 °C, leading to an obvious weight loss. At 560 °C, the oxidation of both LFP and carbon is completed, and the weight of the sample stays constant. Thus, by taking into account the weight gain of pure LFP (5.07%) and the actual weight gain of the LFP/C composite (2.30%) during TG measurement, the total mass ratio of 2.65% of carbon in the composite can be calculated as follows:^2,31,32

C wt% = 1 − (1 + 2.3%)/(1 + 5.07%) = 2.65%

Fig. 7 TGA patterns of the LFP/GNs heated in oxygen from 50 to 650 °C.

The perfectly designed morphology coupled with the optimized graphene content confer the LFP/GNs with a high tap density. As determined by a tap density meter, the LFP/GNs show a high tap density of 1.56 g cm⁻³.

To demonstrate this successful design of the LFP/GN cathode material, their electrochemical performance including cyclic voltammogram properties, high rate capability, cycling performance and electrochemical impedance were tested. Fig. 8a shows CV curves for the LFP/GN composites. The reduction and oxidation peaks appear at around 3.3 V and 3.55 V, respectively. The potential separation between the oxidation and reduction peaks is only 0.25 V. Except for the 1^st cycle, the CV curves of the 2^nd, 5^th, and 10^th cycles overlap very well. The well-defined peaks, small potential interval and good overlap of CV curves suggest the sample has high electrochemical reaction activity and reversibility. Fig. 8b shows the charge–discharge capacities measured at various C rates of the LFP/GNs. The LFP/GNs exhibits an excellent capacity as high as 151.3 mA h g⁻¹ at 1 C. At 5 C, the LFP/GNs can deliver a capacity of 139.2 mA h g⁻¹. Even at the high current rate of 10 C and 20 C, the LFP/GNs can still achieve capacities as high as 129.1 and 118.2 mA h g⁻¹, respectively. These values are among the best rate capacities reported for LFP/graphene systems.^19,20,33,34 Considering our LFP/GNs have the lowest graphene content, and no conductive agent was used in the electrode, such a high rate capability is really surprising. In order to make a comparison to further demonstrate the excellent electrochemical performance of the LFP/GN cathode, we also prepared LFP microspheres using the same method with 2.65% carbon coated on the surface (denote as LFP/C). Shown in Fig. 7c is the comparison of the rate performance of LFP/GNs and LFP/C. At a low rate, for example 1 C, the LFP/C displays a discharge capacity of 116.2 mA h g⁻¹, while the LFP/GNs displays a capacity of 151.3 mA h g⁻¹, indicating the specific capacity was improved by incorporation of graphene. At higher rates, the superior capacity characteristics of the LFP/GN composite are even clearer; for example at 20 C, the LFP/GN composite retains a capacity of 118.2 mA h g⁻¹ in contrast to 40.1 mA h g⁻¹ for LFP/C. In order to test the cycle stability of LFP/GNs at a high rate, we performed long-term charge–discharge cycling at the current rate of 5 C with LFP/C for comparison as shown in Fig. 7d. The LFP/GNs can deliver an initial capacity of 138.6 mA h g⁻¹, and after 500 cycles, the capacity retention is 89.2% of the initial capacity. This result confirms that our LFP/GN composites are tolerant to high charge–discharge currents, which is a desirable characteristic required for high power applications such as in EVs and HEVs. In contrast, the LFP/C shows much worse cycling stability at the high rate. The LFP/C has a capacity retention of 61.6% when charging at 5 C for only 200 cycles.

Fig. 8 (a) Cyclic voltammograms of the 1 (black line), 2 (blue line), 5 (green line) and 10 (red line) cycles of LFP/GNs at a scan rate of 0.2 mV s⁻¹, (b) the charge–discharge profiles of LFP/GNs at different rates; (c) the capacities of LFP/GNs and LFP/C at different rates; (d) the capacity retention of LFP/GNs and LFP/C at 5 C.

The excellent electrochemical performance of the LFP/GN cathodes indicate their good kinetics for Li ion insertion and extraction, and therefore, we performed an electrochemical impedance test with LFP/C for comparison at the end of 200 cycles. As can be seen from Fig. 9, according to the Nyquist plots of the two electrodes, the LFP/GN electrode has a R_ct value of 51 Ω, which is much smaller than that of LFP/C electrode (357 Ω), revealing a lower charge-transfer resistance in the LFP/GN electrode.

Fig. 9 The impedance spectra of the LFP/GNs and LFP/C electrodes at the end of 200 cycles. The inset equivalent circuit was used to fit the impedance data with R_s, resistance of the electrolytes; R_ct, charge-transfer resistance; Z_w, Warburg resistance; CPE, constant phase element.

The test results clearly demonstrate the excellent electrochemical performance of LFP/GNs, and the distinct improvement in the electrochemical performance could be ascribed to their successfully designed structural features: (1) the porous structure of the LFP/GNs, for which we believe that, when the cathode is assembled in lithium cells, liquid electrolyte can easily flood into the mesopores, ensuring the high surface area contact with active LFP material and reducing the diffusion distance of the Li ions.^35–37 This effect induces fast lithium ion diffusion. (2) The well constructed graphene conducting network enables both Li ions and electrons to migrate and reach each active LFP nanoparticle, which successfully solves the biggest problem that hindered the electrochemical performance of LFP, i.e. the poor conductivity, and hence realizes the full potential of the LFP active materials. (3) The graphene nanosheet, which has a large surface area and good mechanical flexibility, can also serve as an electrolyte container for high rate operation as well as an elastic buffer to relieve the strain during the Li ion insertion/de-insertion process that has a beneficial effect on cycling performance.

4 Conclusions

In conclusion, we successfully designed a LFP/GN cathode for Li ion batteries. The as-prepared composite material has a graphene loading weight percentage as low as 2.65%. Electrochemical testing showed that, in the absence of A conductive agent, the LFP/GN electrode can deliver a capacity as high as 118.2 mA h g⁻¹ at the high current rate of 20 C. The porous structure as well as the well constructed graphene conducting network promise the composite will have excellent electrochemical performance. As the synthetic method is simple and can easily be scaled up, which coupled with the low content of graphene greatly reduces the cost of the composite cathode, and thus the LFP/GNs are a promising potential candidate cathode for LIBs that can be used in EVs and HEVs.

Acknowledgements

This work was financially supported by the National Basic Research Program of China (2010CB934700) and the National Natural Science Foundation of China (11079002 and 20973019).

Notes and references

1. M. S. Whittingham, Chem. Rev., 2004, 104, 4271.
2. W. Wei, D. Chen, R. Wang and L. Guo, Nanotechnology, 2012, 23, 475401.
3. Y. Wu, Z. Wen and J. Li, Adv. Mater., 2011, 23, 1126.
4. Y. Wu, Z. Wen, H. Feng and J. Li, Chem.–Eur. J., 2013, 19, 5631.
5. Y. H. Huang and J. B. Goodenough, Chem. Mater., 2008, 20, 7237.
6. A. K. Padhi, K. S. Nanjundaswamy and J. B. Goodenough, J. Electrochem. Soc., 1997, 144, 1188.
7. C. Nan, J. Lu, C. Chen, Q. Peng and Y. Li, J. Mater. Chem., 2011, 21, 9994.
8. L. X. Yuan, Z. H. Wang, W. X. Zhang, X. L. Hu, J. T. Chen, Y. H. Huang and J. B. Goodenough, Energy Environ. Sci., 2011, 4, 269.
9. S. Y. Chung, J. T. Bloking and Y. M. Chiang, Nat. Mater., 2002, 1, 123.
10. I. Bilecka, A. Hintennach, M. D. Rosell, D. Xie, P. Novák and M. Niederberger, J. Mater. Chem., 2011, 21, 5881.
11. J. Liu, T. E. Conry, X. Song, M. M. Doeff and T. J. Richardson, Energy Environ. Sci., 2011, 4, 885.
12. D. Rangappa, K. Sone, M. Ichihara, T. Kudo and I. Honma, Chem. Commun., 2010, 46, 7548.
13. H. M. Xie, R. S. Wang, J. R. Ying, L. Y. Zhang, A. F. Jalbout, H. Y. Yu, G. L. Yang, X. M. Pan and Z. M. Su, Adv. Mater., 2006, 18, 2609.
14. X. L. Wu, L. Y. Jiang, F. F. Cao, Y. G. Guo and L. J. Wan, Adv. Mater., 2009, 21, 2710.
15. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666.
16. A. K. Geim and K. S. Novoselov, Nat. Mater., 2007, 6, 183.
17. A. K. Geim, Science, 2009, 324, 1530.
18. J. Mun, H. W. Ha and W. Choi, J. Power Sources, 2014, 251, 386.
19. Y. Shi, S. L. Chou, J. Z. Wang, D. Wexler, H. J. Li, H. K. Liu and Y. Wu, J. Mater. Chem., 2012, 22, 16465.
20. X. Guo, Q. Fan, L. Yu, J. Liang, W. Ji, L. Peng, X. Guo, W. Ding and Y. Chen, J. Mater. Chem. A, 2013, 1, 11534.
21. G. Qin, S. Xue, Q. Ma and C. Wang, CrystEngComm, 2014, 16, 260.
22. N. I. Kovtyukhova, P. J. Ollivier, B. R. Martin, T. E. Mallouk, S. A. Chizhik, E. V. Buzaneva and A. D. Gorchinskiy, Chem. Mater., 1999, 11, 771.
23. S. Stankovich, D. A. Dikin, G. H. B. Dommett, K. M. Kohlhaas, E. Z. Zimney, E. A. Stach, R. D. Piner, S. T. Nguyen and R. S. Ruoff, Nature, 2006, 442, 282.
24. M. Zhang, D. Lei, Z. Du, X. Yin, L. Chen, Q. Li, Y. Wang and T. Wang, J. Mater. Chem., 2011, 21, 1673.
25. J. Liang, W. Wei, D. Zhong, Q. Yang, L. Li and L. Guo, ACS Appl. Mater. Interfaces, 2012, 4, 454.
26. W. Wei and L. Guo, Part. Part. Syst. Charact., 2013, 30, 658.
27. A. C. Ferrari and J. Robertson, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 61, 14095.
28. J. Qian, M. Zhou, Y. Cao, X. Ai and H. Yang, J. Phys. Chem. C, 2010, 114, 3477.
29. F. Tuinstra and J. L. Koenig, J. Chem. Phys., 1970, 53, 1126.
30. G. Kucinskis, G. Bajars and J. Kleperis, J. Power Sources, 2013, 240, 66.
31. I. Belharouak, C. Johnson and K. Amine, Electrochem. Commun., 2005, 7, 983.
32. Y. L. Cao, L. H. Yu, T. Li, X. P. Ai and H. X. Yang, J. Power Sources, 2007, 172, 913.
33. X. Zhou, F. Wang, Y. Zhu and Z. Liu, J. Mater. Chem., 2011, 21, 3353.
34. J. Ha, S. K. Park, S. H. Yu, A. Jin, B. Jang, S. Bong, I. Kim, Y. E. Sung and Y. Piao, Nanoscale, 2013, 5, 8647.
35. Y. Tang, X. Rui, Y. Zhang, T. M. Lim, Z. Dong and H. H. Hng, J. Mater. Chem. A, 2013, 1, 82.
36. Y. Zhang, Y. Tang, S. Yin, Z. Zeng, H. Zhang, C. M. Li, Z. Dong, Z. Chen and X. D. Chen, Nanoscale, 2011, 3, 4074.
37. S. Yin, Y. Zhang, J. Kong, C. Zou, C. M. Li, X. Lu, J. Ma, F. Y. C. Boey and X. D. Chen, ACS Nano, 2011, 5, 3831.

---