Electrochemical performance of NASICON type carbon coated LiTi₂(PO₄)₃ with a spinel LiMn₂O₄ cathode

Abstract

NASICON type LiTi₂(PO₄)₃ particles are synthesized by a modified pechini type polymerizable complex method at 1000 °C in an air atmosphere. The synthesized LiTi₂(PO₄)₃ particles are ball milled and subsequently carbon coated from the carbonization of glucose (C–LiTi₂(PO₄)₃). Li-insertion properties are evaluated in half-cell configurations (Li/C–LiTi₂(PO₄)₃) and delivered an initial discharge capacity of 117 mAh g⁻¹ at a current density of 15 mA g⁻¹. Carbon coating alleviates the severe capacity fading of LiTi₂(PO₄)₃ during cycling. A full-cell with an operating potential of 1.5 V is constructed employing C–LiTi₂(PO₄)₃ as the anode with a spinel cathode, LiMn₂O₄, which delivered the first discharge capacity of 103 mAh g⁻¹ at current density of 150 mA g⁻¹. The LiMn₂O₄/C–LiTi₂(PO₄)₃ cell retains 72% of initial discharge capacity after 200 cycles and the results suggest that, the full-cell can be used for miniature applications by replacing other rechargeable systems like lead–acid, Ni–Cd and Ni–MH.

---

Introduction

Lithium-ion batteries (LIB) are one of the popular and prominent energy storage devices in this era for miniature applications like watches, toys, cameras etc. Lithium is the lightest metal (6.94 g mol⁻¹) with a theoretical capacity of 3862 Ah kg⁻¹ and a high electronegative potential (−3.05 V vs. normal hydrogen electrode) resulting in very high volumetric and gravimetric energy density of the cell which is attractive in portable and miniature applications. In addition, very little amount of self-discharge (<∼5% per month) and no memory effect when compared to other rechargeable power packs like lead–acid, Ni–Cd, and Ni–MH keeps LIB as a serious contender for high power applications such as electric vehicles (EV) and hybrid electric vehicles (HEV).^1–5 However, dendrite growth is the main problem for metallic lithium based rechargeable systems for practical applications.⁶ Hence, the concept of “host–guest” chemistry emerged and was successfully employed in commercial LIBs introduced by Sony Inc., utilizing carbonaceous anodes.⁷ Nevertheless, carbonaceous anodes endure problems in lithium plating and poor rate performances that led the way to search for new anode alternatives. The decomposition of conventional carbonate based electrolytes cannot be excluded for said anodes during insertion of Li-ions at lower potentials vs. Li.⁵ Several prospective insertion hosts such as anatase–TiO₂ (∼1.75 V vs. Li),^8,9 TiO₂–B (∼1.5 V vs. Li),^9,10 Li₄Ti₅O₁₂ (∼1.5 V vs. Li),^9,11 LiTi₂(PO₄)₃ (∼2.6 V vs. Li),⁸ TiP₂O₇ (∼2.6 V vs. Li),^8,12 VO₂ (B) (∼2.6 V vs. Li)¹³ and Li₃V₂(PO₄)₃ (∼1.7 V vs. Li)¹⁴ with compromising insertion potential and capacity when compared to graphitic anodes (∼0.1 V vs. Li) are proposed. Using LiTi₂(PO₄)₃ anodes provides several advantages like better thermal stability in both lithiated and de-lithiated states due to the presence of strong P–O covalent bond, three dimensional pathways for Li-ion transportation, flat operating potentials during Li-insertion/extraction, ease of synthesis and eco-friendliness. However, in LiTi₂(PO₄)₃, the T^4+/3+ redox couple is observed at ∼2.6 V vs. Li, which is too high for anodes and very low for cathodes in LIB. In addition, the separated arrangement of the TiO₆ octahedral unit supplies the inferior conducting properties of NASICON type LiTi₂(PO₄)₃. Therefore, less studies were carried out for NASICON type LiTi₂(PO₄)₃ as insertion host for LIBs. At the same time, LiTi₂(PO₄)₃ materials play a vital role in aqueous lithium-ion batteries as prospective anodes to bring down the working potential for the safe operation limit (<∼1.23 V), while employing high voltage cathodes such as LiMn₂O₄, LiMn_1.5Ni_0.5O₄, LiNiO₂, LiCoO₂etc..¹⁵ Luo et al.^16,17 reported the possibility of using LiTi₂(PO₄)₃ as insertion type electrodes in aqueous hybrid electrochemical capacitors. There has been no work reported on the utilization of anodes in full-cell configurations although the operating potential is low. Hence, an attempt has been made to synthesize NASICON type LiTi₂(PO₄)₃ phase by a modified pechini type polymerizable complex method. The carbon coating was conducted by carbonization of glucose under Ar atmosphere. The electrochemical properties were evaluated in both half-cell (Li/LiTi₂(PO₄)₃) and full-cell (LiMn₂O₄/LiTi₂(PO₄)₃) configurations at the constant current rate of 150 mA g⁻¹ along with powder characterizations and described in detail.

Materials and methods

NASICON type LiTi₂(PO₄)₃ was prepared by a modified pechini type polymerizable complex method as described by Mariappan et al.¹⁸ In the typical synthesis procedure, stoichiometric amounts of Ti metal (STREM, USA), LiOH·H₂O (Fisher, 99.9%) and NH₄H₂PO₄ (Fisher, 99.9%) were used as starting materials. In the first step, Ti metal (0.96 g) was dissolved in H₂O₂ (30%, Merck, Germany) and ammonia solution (25% Merck, Germany) solution (2 : 1 by volume). To form the polymerizable complex, citric acid to ethylene glycol (EG) was fixed to a 1 : 1 molar ratio and the same molar ratio of Ti metal present in the compound was also maintained. After the complete dissolution of Ti metal, citric acid (3.84 g, Aldrich, 99.5%), NH₄H₂PO₄ (3.45 g) and LiOH·H₂O (0.42 g) were added and continuously stirred until the solution becomes homogeneous. Then, EG was added drop-wise to the above solution and hot plate temperature was increased to 80 °C and maintained for 2 h. To complete the polyesterification process, the temperature of the hotplate was further increased to 140 °C to yield gel precursors. The obtained gel precursor was pyrolysed at 325 °C for 2 h in air to obtain black coloured intermediate powder. The powder was fired at 1000 °C for 24 h in air to obtain single phase LiTi₂(PO₄)₃ and subsequently ball milled (1 h) to reduce the aggregation of particles using SPEX 8000D, USA high energy ball miller. Carbon-coating was conducted after the ball-milling procedure by carbonizing the appropriate amount of glucose under Ar atmosphere at 800 °C for 2 h.

Powder X-ray diffraction measurements were conducted using Bruker AXS, D8 Advance with Cu–Kα radiation. Rietveld refinement was conducted for the obtained reflections using Topas V3 software. Surface morphological features and internal structure of the carbon coated LiTi₂(PO₄)₃ (hereafter-called C–LiTi₂(PO₄)₃) were analyzed using field emission scanning electron microscope (FE-SEM, JEOL JSM-7600F) and transmission electron microscope (TEM, JEOL 2100F), respectively. All the electrochemical studies were conducted using a standard two electrode coin-cell configuration (CR 2016). The composite electrodes were prepared by weighing the accurately weighed (20 mg) active material (C–LiTi₂(PO₄)₃ or LiMn₂O₄), 3 mg of super P carbon as conductive additive and 3 mg of teflonized acetylene black (TAB-2) as a binder. The resultant mixture was placed on a stainless steel mesh, which serves as a current collector for both electrodes. The composite electrodes were dried at 60 °C overnight before conducting cell assembly under an Ar filled glove box. Half-cells (Li/LiMn₂O₄ or Li/C–LiTi₂(PO₄)₃) were fabricated with metallic lithium as the anode and composite electrodes as the cathode, which was separated by a microporous glass fiber separator (Whatman, Cat. No. 1825 047, UK). 1 M LiPF₆ in ethylene carbonate (EC) and diethyl carbonate (DEC) (1 : 1 wt. %, DAN VEC) was used as the electrolyte. Cyclic voltammetric traces were recorded for Li/LiMn₂O₄, Li/C–LiTi₂(PO₄)₃ and LiMn₂O₄/C–LiTi₂(PO₄)₃ cells at scan rate of 0.1 mV s⁻¹ using Solartron, 1470E and SI 1255B impedance/gain-phase analyzer coupled with a potentiostat. Galvanostatic cycling profiles of the half-cells and full-cells were recorded using Arbin battery (2000) tester in ambient temperature conditions at constant current of 150 mA g⁻¹.

Results and discussion

Fig. 1a represents the Rietveld refined powder X-ray diffraction pattern (XRD) of C-LiTi₂(PO₄)₃ particles. The observed XRD patterns clearly indicate the formation of single phase LiTi₂(PO₄)₃ and absence of impurity phases like TiO₂ rutile and Li₃PO₄. The observed satellite reflections are indexed according to the rhombohedral–hexagonal structure with an R3̄c space group. The crystal structure of NASICON type materials is a three dimensional framework consisting of titanium in octahedral units (TiO₆) and phosphorous in tetrahedral position (PO₄), sharing oxygen vertices. Nevertheless, each PO₄ unit surrounded by four VO₆ unit forms a three dimensional chain, in which lithium is located within the three dimensional cavity. Three four-fold crystallographic configurations exist for the lithium atoms, which leads to twelve lithium positions within the unit-cell.^8,19 The lattice parameter values are calculated from the Rietveld refinement and found to be a = 8.515(8) Å and c = 20.912(7) Å which is consistent with literature values. The crystallite size calculated from Scherrer formula during refinement was 80 nm. A highly dispersed particulate morphology was observed for C–LiTi₂(PO₄)₃, as is clearly seen from the scanning electron microscopic image (Fig. 1b). This kind of morphology is due to the high energy ball-milling of LiTi₂(PO₄)₃ particles and subsequent coating with a thin layer of carbon. The presence of a carbon layer prevents the aggregation of particles during the carbonization of glucose, resulting in such dispersed morphology. In order to ensure the presence of uniform carbon coating over the LiTi₂(PO₄)₃ particles, transmission electron microscopy (TEM) was conducted and is presented in Fig. 1c and d. The presence of a filament like material is apparently noted in TEM picture, which is none other than amorphous carbon. Further, high resolution-TEM is recorded to estimate the thickness of the carbon layer, which was found to be ∼3 nm. The inter-planer d-spacing was measured at 5.98 Å and agrees well with hkl of (0 1 2) of the C–LiTi₂(PO₄)₃.

Fig. 1 (a) Rietveld refined X-ray diffraction pattern of carbon coated-LiTi₂(PO₄)₃ powders, (b) scanning electron microscopic images of carbon coated-LiTi₂(PO₄)₃ (c) transmission electron microscopic (TEM) pictures of carbon coated-LiTi₂(PO₄)₃ and (d) high resolution TEM images of carbon coated-LiTi₂(PO₄)₃.

The nature of the carbon covered LiTi₂(PO₄)₃ particulates during carbonization is very important for the improvement of electrochemical properties. Hence, Raman spectra were recorded and are presented in Fig. 2. From Fig. 2, it is apparent to see the presence of carbon with characteristic bands at 1548 and 1604 cm⁻¹. The observed frequencies correspond to the D and G bands of carbon respectively.²⁰ Further, appearance of frequencies at 968, 987, 1006 and 1093 cm⁻¹ are ascribed to the characteristic of LiTi₂(PO₄)₃ particles, in particular (PO₄)³⁻ anions.^21,22 Apart from the evidence of the carbon layer, in which the intensity ratio between D and G bands (I_D/I_G) are very crucial for the enhancement of conducting profiles. In the present case, the ratio is calculated to be 0.85, which indicates the carbonized layer predominantly contains sp² type carbon. The presence of such sp² type carbon enables better conducting profiles and hence, improved electrochemical properties are expected for C–LiTi₂(PO₄)₃ particulates.²³

Fig. 2 Raman spectrum of carbon coated-LiTi₂(PO₄)₃.

Fig. 3 shows the family of cyclic voltammograms (CV) of Li/C–LiTi₂(PO₄)₃ half-cells recorded between 2–3.4 V vs. Li at scan rate of 0.1 mV s⁻¹. The test cell showed an open circuit potential of ∼3.16 V vs. Li and was first discharged to a lower cut-off potential (2 V) to insert Li-ions into the NASICON type C–LiTi₂(PO₄)₃ lattice. During the cathodic sweep, the cell exhibits sharp peak potential of ∼2.29 V, which is an indication of reduction of Ti⁴⁺ in to Ti³⁺. In the anodic scan, the test cell showed an oxidation potential of ∼2.6 V, which corresponds to the Ti³⁺ to Ti⁴⁺ redox couple. The reduction and oxidation of the transition metal, Ti, reveals the insertion and extraction of Li-ions. However, in the second and subsequent cycles, the reduction peak potential shifted towards a higher voltage and this shifting of potential is mainly ascribed to the structural rearrangement in the C–LiTi₂(PO₄)₃ lattice during the first Li-insertion.²⁴ The very sharp oxidation/reduction peaks correspond to the two-phase reaction mechanism according to the following equilibrium LiTi₂(PO₄)₃ + 2Li⁺ + 2e⁻↔Li₃Ti₂(PO₄)₃

Fig. 3 Cyclic voltammogram of Li/carbon coated-LiTi₂(PO₄)₃ cell cycled between 2–3.4 V vs. Li at a scan rate of 0.1 mV s⁻¹, in which metallic lithium acts as both the counter and reference electrodes.

Galvanostatic charge–discharge studies of Li/C–LiTi₂(PO₄)₃ half-cells were cycled between 2–3.4 vs. Li at the current density of 15 and 150 mA g⁻¹ at room temperature. Fig. 4a shows the typical charge–discharge curves of the Li/C–LiTi₂(PO₄)₃ cell at a current density of 15 mA g⁻¹. The cell showed an initial discharge capacity of 117 mAh g⁻¹ (1.69 moles of lithium) with flat a potential of ∼2.43 V against a theoretical capacity of 138 mAh g⁻¹ (for the insertion of two moles of lithium). Reversible capacity of 114 mAh g⁻¹ is obtained in the first charge, which is equal to 1.65 moles of lithium. In the second discharge, the cell showed a discharge of 113 mAh g⁻¹ with a slightly higher insertion potential (∼2.45 V) when compared to the first discharge. The obtained charge–discharge curves are consistent with CV studies. A duplicate cell has been made to study the electrochemical Li-insertion behaviour at relatively high current of 150 mA g⁻¹ and is presented in Fig. 4b. The Li/C–LiTi₂(PO₄)₃ half-cell delivered the initial discharge and charge capacities of 114 and 109 mAh g⁻¹, respectively. An increasing polarization of the electrodes is also noted when increasing the current density from 15 to 150 mA g⁻¹. Cycling profiles of Li/C–LiTi₂(PO₄)₃ cells up to 200 cycles were tested at a current density of 150 mA g⁻¹ and are given in Fig. 4c. It is obvious to notice that the cell experienced capacity fading during cycling, for example discharge capacities of 87, 82, 80 and 78 mAh g⁻¹ were noted for the 50, 100, 150 and 200 cycles, respectively. However, severe capacity fading is observed only in the initial cycles (up to 10 cycles). The capacity fading is calculated from the 10^th cycle to the 200^th cycle and was found to be 0.09 mAh g⁻¹ per cycle. Nevertheless, capacity fading is obvious in the case of NASICON type LiTi₂(PO₄)₃, which is mainly attributed to the intrinsic nature of the compound and a similar kind of fading is already noted by previous researchers, for example Patoux and Masquelier⁸ reported a ∼0.63 mAh g⁻¹ capacity fade per cycle between 2–3.4 V vs. Li (∼120 and ∼95 mAh g⁻¹ for the 1^st and 40^th cycle, respectively at a low current rate of 0.1 C). Wessels et al.^25,26 reported the performance of carbon-coated LiTi₂(PO₄)₃ and it showed a capacity retention of 70% after 160 cycles at 1 C rate. Though, the oxygen deficient phase showed (LiTi₂(PO_4−δ)₃) two orders of magnitude increment in electronic conductivity and displayed the discharge capacity of 93 mAh g⁻¹ only compared to native phase (81 mAh g⁻¹) at 1 C rate.²⁷ The C–LiTi₂(PO₄)₃ prepared by pechni type polymerizable complex method and subsequent carbon coating using glucose yields better performing electrode as compared to previous reports.

Fig. 4 (a) Galvanostatic charge–discharge curves of Li/carbon coated-LiTi₂(PO₄)₃ cell cycled between 2–3.4 V vs. Li at current density of (a) 15 mA g⁻¹, (b) 150 mA g⁻¹ and (c) cycling performance of the above cell at a current density of 150 mA g⁻¹. The integers represent the cycle number.

The full-cell, LiMn₂O₄/C–LiTi₂(PO₄)₃ was constructed and the electrochemical properties were evaluated between 0.9–1.7 V at current density of 150 mA g⁻¹. In the case of full-cell arrangement, there is no Li-reservoir like that in metallic lithium half-cell configurations, the cathode, LiMn₂O₄, is the only available source. Hence, optimization of mass between two electrodes is necessary to utilize the full capacity of the both active materials. Based on the half-cell performance of both cells (Li/LiMn₂O₄ and Li/C–LiTi₂(PO₄)₃) under the same current rate (150 mA g⁻¹), and with the mass ratio fixed anode to cathode 1 : 1.04, the full-cell, LiMn₂O₄/C–LiTi₂(PO₄)₃, was potentiostatically cycled between 0.9–1.7 V at a scan rate of 0.1 mV s⁻¹ and is presented in Fig. 5. For the comparison purpose, electrochemical reactions in half-cell configuration of the individual components (LiMn₂O₄/C–LiTi₂(PO₄)₃) are also presented along with full-cell result in Fig. 5. The full-cell, showed an open circuit potential of ∼100 mV and charged first to extract lithium from the LiMn₂O₄ lattice, the extracted Li-ions are then simultaneously inserted in to LiTi₂(PO₄)₃ lattice. The sharp anodic peak ∼1.6 V corresponds to the oxidation of Mn³⁺ to Mn⁴⁺ or reduction of Ti⁴⁺ to Ti³⁺. Similarly, during the cathodic scan, the sharp peak ∼1.45 V is attributed to the oxidation of Ti³⁺ to Ti⁴⁺ or reduction of Mn⁴⁺ to Mn³⁺. In the electrochemical reaction, oxidation process corresponds to the removal of lithium and reduction reveals the insertion of lithium back and forth. The mirror like CV traces clearly indicate the reversibility of the cell during electrochemical cycling. A small variation in area under the curve is noted which indicates the capacity fade during cycling.

Fig. 5 Cyclic voltammogram (CV) of LiMn₂O₄/carbon coated-LiTi₂(PO₄)₃ cell cycled between 0.9–1.7 V at scan rate of 0.1 mV s⁻¹. For comparison, the CV traces of Li/LiMn₂O₄ and Li/carbon coated-LiTi₂(PO₄)₃ cells are also given.

The galvanostatic cycling performance of the LiMn₂O₄/C–LiTi₂(PO₄)₃ cells was conducted between 0.9–1.7 V at a current density of 150 mA g⁻¹ at room temperature. Typical charge–discharge curves of the LiMn₂O₄/C–LiTi₂(PO₄)₃ cell are given Fig. 6a and the capacity values are calculated based on the mass loading of the cathode. The cell displayed capacities of 121 and 103 mAh g⁻¹ for first charge and discharge, respectively. The discharge curve showed more or less a flat potential at ∼1.5 V and is consistent with the CV measurements. An irreversible capacity of about 18 mAh g⁻¹ is observed and the values are expected, since both half-cell configurations (Li/LiMn₂O₄ or Li/C–LiTi₂(PO₄)₃) are experiencing irreversible capacity loss in the first cycle. The cycling profiles of the LiMn₂O₄/C–LiTi₂(PO₄)₃ cells are illustrated up to 200 cycles in Fig. 6b. The full-cell experiences capacity fade during the cycles and this is evident from the cycling profile. The cycling behaviour is very similar to the half-cell, Li/C–LiTi₂(PO₄)₃. Further, the small capacity fading observed from the Li/LiMn₂O₄ cell cannot be excluded (given in the supplementary information†). In cycling, the LiMn₂O₄/C–LiTi₂(PO₄)₃ cell delivered discharge capacities of 87, 78 and 74 mAh g⁻¹ for 50, 100 and 200 cycles, respectively. Except for the first cycle, the cell showed a columbic efficiency of over 99%,which indicates the excellent reversibility during charge–discharge process and good agreement with CV studies. The operating potential (∼1.5 V) of the LiMn₂O₄/C–LiTi₂(PO₄)₃ full-cell is quite low when compared to the half-cell assemblies of either the anode (Li/C–LiTi₂(PO₄)₃) or the cathode (Li/LiMn₂O₄). At the same time, the use of metallic lithium in the practical rechargeable systems is not advisable in aprotic solvents, which may lead to dendritic growth and can result in short-circuiting of the cell. Meanwhile, the observed operating potential is higher than other conventional rechargeable battery systems like, lead–acid, Ni–MH, Ni–Cd and magnesium.^4,28 In the case of Ni–MH and Ni–Cd, the operating potential is highly restricted due to water splitting issues (∼1.23 V), use of environmentally hazardous materials and memory effect issues. Maintenance and cycleability is the main challenge for lead acid batteries. High temperature operation is also another important concern for those batteries, whereas the LiMn₂O₄/C–LiTi₂(PO₄)₃ system is overcomes the aforementioned setbacks noted in the other rechargeable systems by employing low cost, eco-friendly materials.⁴ Further, the operating potential of the C–LiTi₂(PO₄)₃ anode based system can be further improved by employing high voltage cathodes like LiNi_0.5Mn_1.5O₄, LiCoPO₄ depending on the requirements. Thus, the 1.5 V LiMn₂O₄/C–LiTi₂(PO₄)₃ cell can be used effectively for miniature applications like watches, toys, cameras, etc.

Fig. 6 (a) Galvanostatic charge–discharge curves of LiMn₂O₄/carbon coated-LiTi₂(PO₄)₃ cell cycled between 0.9–1.7 V at a current density of 150 mA g⁻¹ and (b) cycling performance of the above cell. The integers represent the cycle number.

Conclusions

NASICON type LiTi₂(PO₄)₃ was synthesized by a modified pechni type polymerizable complex method and subsequently ball milled to reduce the particle size. The resultant phase was carbon coated by the carbonization of glucose (C–LiTi₂(PO₄)₃). The carbon coating was confirmed by HR-TEM and Raman analysis and it was found ∼2.3 wt.% in the final compound. The Li-insertion properties were evaluated in half-cell configurations (Li/C–LiTi₂(PO₄)₃) and found to be 1.69 moles per formula unit. However, a noticeable amount of capacity fading was observed during cycling. The full-cell was constructed with LiMn₂O₄ cathode and the cell exhibited the initial discharge capacity of 103 mAh g⁻¹ at current density of 150 mA g⁻¹. The full-cell, LiMn₂O₄/C–LiTi₂(PO₄)₃ renders 72% of initial discharge capacity with over 99% columbic efficiency at a working potential of 1.5 V. The obtained results clearly indicate that the full-cell LiMn₂O₄/C–LiTi₂(PO₄)₃ can be effectively used for miniature applications and possibly replace lead–acid, Ni–Cd and Ni–MH rechargeable systems.

Acknowledgements

We thank the National Research foundation (NRF, Singapore) for financial support through the Competitive Research Programme (CRP) (Grant no. NRF-CRP4-2008-03) and the Clean Energy Research Project (CERP) (Grant no. NRF-2009-EWT-CERP001-036).

References

1. M. K. Song, S. Park, F. M. Alamgir, J. Cho and M. Liu, Mater. Sci. Eng., R, 2011, 72, 203–252.
2. O. K. Park, Y. Cho, S. Lee, H. C. Yoo, H. K. Song and J. Cho, Energy Environ. Sci., 2011, 4, 1621–1633.
3. V. Etacheri, R. Marom, R. Elazari, G. Salitra and D. Aurbach, Energy Environ. Sci., 2011, 4, 3243–3262.
4. J. O. Besenhard and M. Winter, ChemPhysChem, 2002, 3, 155–159.
5. V. Aravindan, J. Gnanaraj, S. Madhavi and H. K. Liu, Chem.–Eur. J., 2011, 17, 14326–14346.
6. K. Xu, Chem. Rev., 2004, 104, 4303–4418.
7. Y. Nishi, Chem. Rec., 2001, 1, 406–413.
8. S. Patoux and C. Masquelier, Chem. Mater., 2002, 14, 5057–5068.
9. Z. Yang, D. Choi, S. Kerisit, K. M. Rosso, D. Wang, J. Zhang, G. Graff and J. Liu, J. Power Sources, 2009, 192, 588–598.
10. A. R. Armstrong, G. Armstrong, J. Canales and P. G. Bruce, J. Power Sources, 2005, 146, 501–506.
11. T. F. Yi, L. J. Jiang, J. Shu, C. B. Yue, R. S. Zhu and H.-B. Qiao, J. Phys. Chem. Solids, 2010, 71, 1236–1242.
12. V. Aravindan, M. V. Reddy, S. Madhavi, S. G. Mhaisalkar, G. V. Subba Rao and B. V. R. Chowdari, J. Power Sources, 2011, 196, 8850–8854.
13. G. Armstrong, J. Canales, A. R. Armstrong and P. G. Bruce, J. Power Sources, 2008, 178, 723–728.
14. A. Aboulaich, R. Bouchet, G. Delaizir, V. Seznec, L. Tortet, M. Morcrette, P. Rozier, J. M. Tarascon, V. Viallet and M. Dollé, Adv. Energy Mater., 2011, 1, 179–183.
15. J.Y. Luo, W.J. Cui, P. He and Y.Y. Xia, Nat. Chem., 2010, 2, 760–765.
16. J. Y. Luo and Y. Y. Xia, J. Power Sources, 2009, 186, 224–227.
17. J. Y. Luo, J. L. Liu, P. He and Y. Y. Xia, Electrochim. Acta, 2008, 53, 8128–8133.
18. C. R. Mariappan, C. Galven, M. P. C. Lopez, F. L. Berre and O. Bohnke, J. Solid State Chem., 2006, 179, 450–456.
19. A. Aatiq, M. Menetrier, L. Croguennec, E. Suard and C. Delmas, J. Mater. Chem., 2002, 12, 2971–2978.
20. A. C. Ferrari and J. Robertson, Phys. Rev. B: Condens. Matter, 2000, 61, 14095.
21. I. de Fátima Gimenez, I. O. Mazali and O. L. Alves, J. Phys. Chem. Solids, 2001, 62, 1251–1255.
22. Y. Yue and W. Pang, J. Mater. Sci. Lett., 1990, 9, 1392–1392.
23. S. K. Martha, B. Markovsky, J. Grinblat, Y. Gofer, O. Haik, E. Zinigrad, D. Aurbach, T. Drezen, D. Wang, G. Deghenghi and I. Exnar, J. Electrochem. Soc., 2009, 156, A541–A552.
24. V. Aravindan, W. Chuiling, M. V. Reddy, G. V. S. Rao, B. V. R. Chowdari and S. Madhavi, Phys. Chem. Chem. Phys., 2012, 14, 5808–5814.
25. C. Wessells, R. A. Huggins and Y. Cui, J. Power Sources, 2011, 196, 2884–2888.
26. C. Wessells, F. La Mantia, H. Deshazer, R. A. Huggins and Y. Cui, J. Electrochem. Soc., 2011, 158, A352–A355.
27. J. Y. Luo, L. J. Chen, Y. J. Zhao, P. He and Y. Y. Xia, J. Power Sources, 2009, 194, 1075–1080.
28. G. Armstrong, A. R. Armstrong, P. G. Bruce, P. Reale and B. Scrosati, Adv. Mater., 2006, 18, 2597–2600.

---

Footnote

† Electronic Supplementary Information (ESI) available. See DOI: 10.1039/c2ra20826a/

---