A joint experimental and theoretical study on the effect of manganese doping on the structural, electrochemical and physical properties of lithium iron phosphate

Abstract

Mn doped LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) cathode materials are synthesized by a carbothermal reduction method. The structure, morphology and electrochemical performance of the samples are characterized by X-ray diffraction (XRD), scanning electron microscopy (SEM), X-ray photoelectron spectroscopy (XPS), Raman, and charging–discharging tests. It is found that an appropriate amount of Mn doping did not affect the olivine structure and morphology of LiFePO₄, but greatly improves its electrochemical performance. Especially, the LiFe_0.97Mn_0.03PO₄ sample exhibits the best electrochemical performance, which shows the maximum discharge capacity of 148.2 mA h g⁻¹ at 0.1 C and a capacity retention ratio of 98% after 60 cycles at various rates. Based on first-principles density functional theory (DFT), the lithium ion migration channels, energy band structures, densities of states, and charge density diffusion of LiFePO₄ and LiFe_15/16Mn_1/16PO₄ are calculated to further investigate the influence of Mn doping on the LiFePO₄ lattice. The results prove that the higher lithium ion conduction and electronic conductivity of LiFe_15/16Mn_1/16PO₄ are attributed to Mn doping.

---

Introduction

As is well-known, olivine-type LiFePO₄ is one of the most promising cathode material candidates for lithium-ion power batteries, because of its prominent properties such as high theoretical capacity, low cost, high level of thermal stability and huge power generation.^1–3 However, one of the main obstacles for practical applications of LiFePO₄ is its poor rate capability, which can be attributed to poor electronic conductivity and slow rate of lithium-ion diffusion coefficient.^4,5 Up to the present, many effective approaches have been proposed to solve above problems. Carbon coating on the surface of LiFePO₄ particles by sintering carbon sources or carbonaceous materials is demonstrated to be an effective way to improve electrochemical performance.^6,7 However, a high amount of carbon addition could decrease the volumetric energy density of the composite materials. Another method is to dope metal ion, which has been considered as an efficient method of improving electron transfer and lithium ion diffusion within the LiFePO₄ lattice.^8–12 Chung et al. investigated different cation dopants to determine the effects of doping on the electronic conductivity of LiFePO₄, which sparked large amount of research looking at doping modification of LiFePO₄.⁸

Theoretical calculation has the advantage in supplementing the real experiments that one has full manipulate the relevant variables. The first-principles investigation has been shown to be a useful method to predict the properties of cathode materials in lithium-ion batteries.^13–16 For example, Ceder et al. calculated activation barriers to Li ion motion for LiMPO₄ (M = Mn, Fe, Co, Ni) olivine materials by using first-principles method, and demonstrated that Li ion diffuses through one-dimensional channels with high energy barriers to cross between the channels.¹³ In recent years, a lot of research works on metal ion doping of LiFePO₄ by first-principle have been reported.^17–21 Shi et al. investigated Mg doping LiFePO₄ at Li and Fe sites using first-principles density-functional theory, and elucidated the dependence of the electrochemical properties on the concentration of Mg dopant.¹⁷

It is essential to study systematically the micro-structural changes and electrochemical performance of metal ion doped LiFePO₄ for a better understanding on metal ion doping. Though some interesting experimental or theoretical results have been obtained, no joint experimental and theoretical investigations have been carried out to study the effect of metal ion doping on the physical and electrochemical characterization of LiFePO₄, and the intrinsic reasons of electrochemical properties enhancing of metal ion doped LiFePO₄ have not been well clarified.

In this paper, the Mn doped LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) were prepared via a simple carbothermal reduction method. The effect of metal ion Mn doping with different amounts on the structure, morphology and electrochemical performance of LiFePO₄ was studied in detail. In addition, by the first-principles, we calculated and compared the structural and electronic properties of LiFePO₄ and LiFe_15/16Mn_1/16PO₄. Combined with the experimental results, the mechanism that metal ion Mn doping improves the electrochemical performance of LiFePO₄ is proposed theoretically.

Experimental

Preparation of materials

The manganese doped LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) samples were synthesized using carbothermal reduction method. Stoichiometric amounts of analytical reagents Li₂CO₃, Fe₂O₃, MnCO₃, and NH₄H₂PO₄ were mixed in ethyl alcohol based on the desired amount of LiFe_1−xMn_xPO₄. Then proper acetylene black was added as reducing agent. The mixture was ball-milled for 10 h, and dried at 60 °C in the oven. After drying, the precursors were calcined under flowing nitrogen atmosphere at 350 °C for 6 h, followed by sintering at 700 °C for 10 h. After cooling to room temperature, the LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) samples were obtained.

Measurements

The structure of powders was performed with X-ray diffraction (XRD) on a SHIMADZU XRD-7000 X-ray diffractometer equipped with Cu Kα radiation. The morphology of sample was observed using a field-emission scanning electron microscope (FE-SEM) on a JEOL JSM-7600F FE-SEM system. An EASY ESCA X-ray photoelectron spectroscopy (XPS) equipment was used to determine the valence state of the samples with monochromatic Al X-ray radiation (1486.6 eV). Raman experiments were conducted on a Renishaw inVia Raman microscope equipped with 514.5 nm incident radiation in a quasi-backscattering geometry with parallel polarization incident light.

The electrochemical properties of LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) samples were measured by CR2032 button-type cell. The cathode electrodes were prepared by coating the slurry of a mixture composed of 80 wt% active material, 10 wt% acetylene black, and 10 wt% polyvinylidene fluoride (dissolved in N-methyl pyrrolidone) onto aluminum foil current-collector, and then dried at 120 °C for 10 h under vacuum. The electrolyte was 1 M LiPF₆ in 1 : 1 : 1 (v/v) ethylene carbonate, dimethyl carbonate and diethyl carbonate. A Celgard microporous polypropylene sheet was used as the separator, and a lithium metal foil was used as the anode electrode. The cells assembly was performed under an argon atmosphere in a glove box. The charge–discharge and cycling properties of the cells were evaluated between 2.2 and 4.2 V (vs. Li⁺/Li) using a EWARE BST-8 electrochemical test instrument at room temperature.

Theoretical calculations

In the present study, the theoretical calculations were carried out using first principles plane wave pseudo-potential with Cambridge Serial Total Energy Package (CASTEP) code, which was based on density functional theory (DFT) and molecular dynamics theory. The electronic exchange correlation function was realized using the Perdew Burke Ernzerhof (PBE) function within the generalized gradient approximation (GGA). Owing to the restriction of the computational resources, the experimental geometry of LiFe_15/16Mn_1/16PO₄ (2 × 2 × 1 supercell) was adopted to simulate Mn-doping model. For the geometry optimization and electronic property calculations of models, the Vanderbilt ultrasoft pseudopotential^22,23 was used with the cutoff energy of 300 eV, smearing was used with 0.2 eV and a Monkhorst Pack grid²⁴ with k-point for irreducible Brillouin zone sampling. The k-point sets 1 × 1 × 3 grids for the supercell. The energetic convergence threshold for self-consistent field (SCF) was 2 × 10⁻⁵ eV per atom, and the maximum ionic displacement was within 0.002 Å, and the maximum force tolerance was within 0.05 eV nm⁻¹, and the total stress tensor was reduced to 0.1 GPa.

Results and discussion

Fig. 1(a) shows the X-ray diffraction patterns of LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) samples. For all the samples, all dominant diffraction lines of samples can be indexed in ordered olivine structure, indicating a perfect crystallinity of the samples. Besides, no peaks of impurity phases could be detected in these patterns. It is obvious that a low amount of Mn doping do not affect the structure of the samples and formation of LiMPO₄ (M = Fe, Mn) solid solution reported by previous works.^25–27 The magnification part in the range of 24.8–26.4° of XRD patterns is shown in Fig. 1(b). There is a slight shift to lower angle of the diffraction peaks for the Mn-doped samples. These shifts indicate that the addition of manganese is successfully incorporated into LiFePO₄ lattice without altering olivine structure.

Fig. 1 XRD patterns of LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) samples (a). Magnification part in the range of 24.8–26.4° of XRD patterns (b).

The detailed lattice parameters of LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) samples are calculated and listed in Table 1. As being shown, the lattice parameters of Mn doped LiFePO₄ are larger than those of undoped sample. Moreover, with increased x, the increase in the unit cell volume is almost linear. It is in good agreement with A. Yamada's report.²⁶ That is because average value of the radius of Mn ion (0.84 Å) is larger than that of Fe ion (0.74 Å) in octahedral coordination.

Table 1 Lattice parameters of LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) samples

Sample	a (Å)	b (Å)	c (Å)	V (Å^3)
LiFePO4	6.0012	10.3176	4.6861	290.15
LiFe0.99Mn0.01PO4	6.0020	10.3189	4.6913	290.55
LiFe0.97Mn0.03PO4	6.0035	10.3213	4.6924	290.76
LiFe0.95Mn0.05PO4	6.0052	10.3216	4.6937	290.93
LiFe0.93Mn0.07PO4	6.0066	10.3288	4.6939	291.21

---

The morphological feature of the material is an extremely important factor for the electrochemical property of the cathode material, especially for its high-rate charge–discharge performance. The scanning electron microscopy images of LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) samples are shown in Fig. 2. It can be seen that the LiFe_1−xMn_xPO₄ is composed of irregular particles, which aggregate slightly with a broad particle size distribution of 100–600 nm. It indicates that a small amount of Mn doping does not have any influence on morphology of LiFe_1−xMn_xPO₄. The nano-sized LiFe_1−xMn_xPO₄ particles can shorten the diffusion path of lithium ions as well as providing rapid and easy Li⁺ transport.

Fig. 2 SEM images of LiFe_1−xMn_xPO₄ samples, x = 0, 0.01, 0.03, 0.05, 0.07 correspond to a–e.

In order to identify the oxidation state of Fe and Mn in the samples, X-ray photoelectron spectroscopy (XPS) characterization is carried out and reported in Fig. 3. The Fe 2p spectrum consists of two components (Fe 2p_3/2 and Fe 2p_1/2) because of spin–orbit coupling. The binding energy position of the main peak is related to the oxidation state of Fe. For the ferrous ion in LiFePO₄, the main peaks of Fe 2p_3/2 and Fe 2p_1/2 are emerged at ∼710 and ∼724 eV, respectively. For the ferric iron in FePO₄, the main peaks of Fe 2p_3/2 and Fe 2p_1/2 are centred at ∼713 and ∼726 eV, respectively. For the LiFe_0.97Mn_0.03PO₄ sample, the peak position (710.4 and 723.8 eV) in Fig. 3(a) is consistent with Fe²⁺, and there is no peak assigned to Fe³⁺. The Mn 2p spectrum shape for LiFe_0.97Mn_0.03PO₄ sample is similar with that of Fe 2p. It can be observed that a main peak centers at ∼641 eV which can be attributed to Mn 2p_3/2 in the typical form of Mn²⁺. From the analysis of XPS, we confirm that Mn²⁺ has been incorporated into the lattice and remained the valence as in the starting material. Therefore, the oxidation states of Fe and Mn in the LiFe_1−xMn_xPO₄ samples are +2.

Fig. 3 XPS spectra of Fe 2p (a) and Mn 2p (b) for LiFe_0.97Mn_0.03PO₄ sample.

Fig. 4 shows the Raman spectra of LiFe_0.97Mn_0.03PO₄ and LiFePO₄ samples. It can be seen that the peaks observed agree very well to the assignments of LiFePO₄ reported in the literatures.^28,29 The bands in the Raman spectra between 300 and 700 cm⁻¹ ν₂ and ν₄ comprise bending modes of a PO₄³⁻ anion. Three bands observed between 900 and 1200 cm⁻¹ belong to intramolecular symmetric and asymmetric stretching modes of the PO₄³⁻ anion (ν₁ and ν₃). The weak broad bands at ∼1350 and ∼1580 cm⁻¹ can be assigned to the D band and G band of carbonaceous materials, respectively. This proves that a small amount of residual carbon from acetylene black may be present in the LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) samples.

Fig. 4 Raman spectra of LiFe_0.97Mn_0.03PO₄ (a) and LiFePO₄ (b) samples.

Fig. 5(a) shows the first charge–discharge profile of LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) samples between 2.2 and 4.2 V at 0.1 C. All samples exhibit the typical electrochemical behavior of LiFePO₄, with a long charge and discharge voltage plateau at ∼3.5 and ∼3.4 V. It indicates that the well-known crystal LiFePO₄ is successfully synthesized and small amounts of Mn doping do not destroy the crystal structure of LiFePO₄. It is difficult to observe voltage plateau of Mn²⁺/Mn³⁺ at ∼4.0 V due to very small concentration of Mn in LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) samples. Pristine LiFePO₄ shows the lowest charge–discharge capacity of 135.2 and 131.9 mA h g⁻¹, followed by LiFe_0.93Mn_0.07PO₄ of 142.4 and 140.7 mA h g⁻¹, LiFe_0.99Mn_0.01PO₄ of 144.7 and 142.4 mA h g⁻¹, and then LiFe_0.95Mn_0.05PO₄ of 146.5 and 145.1 mA h g⁻¹. LiFe_0.97Mn_0.03PO₄ has the highest charge–discharge capacity of 151.4 and 148.2 mA h g⁻¹, which is nearly 90% of theoretical capacity. Moreover, the LiFe_0.97Mn_0.03PO₄ has an almost 100% capacity retention after 20 charge–discharge cycles at 0.1 C (Fig. 5(b)). This means that LiFe_0.97Mn_0.03PO₄ has a good electrochemical characteristic at low current rate.

Fig. 5 The first charge–discharge profile (a) and discharge cyclability (b) of LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) samples between 2.2 and 4.2 V at 0.1 C. Discharge curves (c) of LiFe_0.97Mn_0.03PO₄ at various rates. Cycling and rate performance (d) of LiFe_0.97Mn_0.03PO₄ and LiFePO₄.

To study the rate capability of LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07), the electrodes are discharged at various current rates. Fig. 5(c) presents the discharge capability of LiFe_0.97Mn_0.03PO₄ electrode at various current rates. The LiFe_0.97Mn_0.03PO₄ sample exhibits discharge capacity of 142.5, 130.5, 115.0, and 96.3 mA h g⁻¹ as it is discharged at 0.5, 1, 2, and 5 C, while that of LiFePO₄ is 126.1, 111.0, 94.2, and 71.8 mA h g⁻¹, respectively (Fig. 5(d)). Apparently, LiFe_0.97Mn_0.03PO₄ has an outstanding rate performance at more than 0.5 C. From Fig. 5(d), as the current rate is reduced again to 0.1 C, the capacity of the LiFe_0.97Mn_0.03PO₄ sample retains 98% after 60 cycles. It supports that the LiFe_0.97Mn_0.03PO₄ displays a better rate capability at various rates compared with other LiFe_1−xMn_xPO₄ electrodes, and enhances the rate capability and cycling performance effectively. Furthermore, it can be seen that all LiFe_1−xMn_xPO₄ electrodes exhibit excellent rate capabilities. This indicates that proper content Mn doping plays a significant role in enhancing the electrochemical performances of LiFePO₄.

In order to investigate the effect of Mn doping on electrochemical performance of LiFe_1−xMn_xPO₄, the first principles calculations have been performed to calculate and analyze the structure and electronic characterizations of LiFePO₄ and LiFe_15/16Mn_1/16PO₄. (Only a maximum of 112-atom LiFePO₄ supercell with 2 × 2 × 1 repetition can be built due to restriction of the computational resources.) Pure bulk olivine-type LiFePO₄ has an orthorhombic unitcell, which accommodates four LiFePO₄ formula-units with the space group of Pnma. Fig. 6 shows the three-dimensional framework of LiFe_15/16Mn_1/16PO₄ supercell, which contains 16 LiFePO₄ formula-units. It can be seen that the Mn doping model is built by replacing one Fe atom from the LiFePO₄ supercell with 2 × 2 × 1 repetition. Table 2 summarizes the average bond lengths of Fe–O, P–O, and Li–O adjacent to Mn atom in LiFe_15/16Mn_1/16PO₄ and LiFePO₄ supercell. As we know, the wide [010] direction lithium ion channels will promote Li ion migration. It is clear from Table 2 that the LiFe_15/16Mn_1/16PO₄ has shorter Fe–O bond lengths and longer Li–O distance, which implies broadening of the adjacent lithium ion channels. Therefore, it can be concluded that doping Mn atoms in LiFePO₄ supercell may improve the lithium ion conduction.

Fig. 6 The three-dimensional framework of LiFe_15/16Mn_1/16PO₄.

Table 2 The average bond lengths of Fe–O, P–O, and Li–O adjacent to Mn atom in LiFe_15/16Mn_1/16PO₄ and LiFePO₄ supercell

M–O bond length (average)	LiFe15/16Mn1/16PO4	LiFePO4
Fe1–O	2.109	2.119
Fe2–O	2.120	2.124
Fe3–O	2.108	2.124
P1–O	1.562	1.554
P2–O	1.568	1.557
P3–O	1.565	1.554
Li1–O	2.173	2.163
Li2–O	2.214	2.179

---

To further analyze the impact of Mn doping on the characterizations of LiFePO₄, we examine the electronic structures of LiFePO₄ with and without Mn doping. Fig. 7 shows calculated electronic band structures of LiFePO₄ and LiFe_15/16Mn_1/16PO₄ along the high symmetry point across the first Brillouin zone. The Fermi level is at 0 eV on the energy axis. The calculated band gap of LiFePO₄ displays an indirect band gap about 0.61 eV that is consistent with the previously reported ∼0.60 eV.^30,31 The calculation results are reasonable for a comparison, although it is underestimated comparing with the experimental value of 3.75 eV³² due to the well-known drawback of DFT. According to Fig. 7(b), band gap of LiFe_15/16Mn_1/16PO₄ is about 0.44 eV, which is much lower than that of LiFePO₄. Some papers^33–38 have demonstrated similar results by theoretical calculation. The narrower band gap indicates that the electronic conductivity of LiFePO₄ could be enhanced by doping Mn within LiFePO₄ lattice.

Fig. 7 Calculated electronic band structures of LiFePO₄ (a) and LiFe_15/16Mn_1/16PO₄ (b).

The density of states (DOS) of undoped and doped models are calculated to investigate how the band gap changes, and the results are shown in Fig. 8. Fig. 8(c) shows the partial density of states (PDOS) for Fe 3d states, P 3p states, O 2p states, and Li 2s and 2p states of LiFePO₄. It can be seen that two major valence bands include the high one located between −0.63 eV and the Fermi level and the low one located between −9.12 eV and −3.06 eV. The former is formed primarily by the Fe 3d orbit with minor contribution from O 2p orbit, while the latter is formed by a bonding state of the hybridized P 3p and O 2p orbit. The PDOS around the Fermi level is concerned due to the importance of the electron orbits near the Fermi level in the electron transport processes. The dominant electronic character of LiFePO₄ near the Fermi level is contributed by Fe 3d orbit. The DOS of Mn atom and its neighboring Fe, O, P, Li atoms are compared (Fig. 3(c)) to clarify the effect of Mn doping on the electronic structure of LiFePO₄ near the Fermi level. It is clear that the PDOS at the Fermi level mainly come from the Mn-3d orbit, which hybrids with its neighboring Fe-3d and O-2p orbits. The results demonstrate that the PDOS of Mn atoms play a key role in band gap reduction of LiFe_15/16Mn_1/16PO₄.

Fig. 8 The calculated TDOS and PDOS of LiFePO₄ (a and c) and LiFe_15/16Mn_1/16PO₄ (b and d) obtained from DFT calculations.

Charge density distribution of LiFePO₄ and LiFe_15/16Mn_1/16PO₄ are investigated and illustrated in Fig. 9 to give a clear description about bonding nature between Mn atoms and adjacent O atoms. It can be seen that no significant change in the charge density distribution is observed by introducing Mn doping. However, the electron clouds of Fe atoms overlapped with that of O atoms (Fig. 9(b)), while the corresponding covalent bond is weak in LiFePO₄. It indicates that there should be more electrons transfer from Fe atoms to O atoms in LiFe_15/16Mn_1/16PO₄. Furthermore, a covalent bond is also formed between Mn and its adjacent O atoms, although it is weaker than the corresponding Fe–O bond. These results demonstrate that a strong interaction of Fe–O bonding by introducing Mn doping in LiFe_15/16Mn_1/16PO₄ supercell induces the narrowed band gap.

Fig. 9 Charge density distribution of LiFePO₄ (a) and LiFe_15/16Mn_1/16PO₄ (b).

Conclusions

LiFe_1−xMn_xPO₄ (x = 0, 0.01, 0.03, 0.05, 0.07) samples were successfully prepared through carbothermal reduction method and investigated by XRD, SEM, XPS, Raman, and electrochemical tests. It was found that appropriate amounts of manganese doping did not affect the olivine structure and morphology of LiFePO₄, but greatly improved its electrochemical performance. The LiFe_0.97Mn_0.03PO₄ sample exhibited the best electrochemical performance, which shows the maximum discharge capacity of 148.2 mA h g⁻¹ and 96.3 mA h g⁻¹ at rates of 0.1 C and 5 C, respectively. Moreover, the capacity retention based on the first discharge capacity was 98% after 60 cycles. Furthermore, the doped configuration, structures and electronic properties of Mn-doped LiFePO₄ were investigated by the first-principles calculations. Combined with the experimental results, it can be found that the electronic conductivity and lithium ion migration rate of LiFePO₄ were enhanced from electronic band structures, DOS, and charge density distribution.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Grant no. 61271040), The Fundamental Research Funds for the Central Universities (Grant no. ZYGX2012J036).

Notes and references

1. S. I. Nishimura, G. Kobayashi, K. Ohoyama, R. Kanno, M. Yashima and A. Yamada, Nat. Mater., 2008, 7, 707.
2. P. G. Bruce, B. Scrosati and J. M. Tarascon, Angew. Chem., Int. Ed., 2008, 47, 2930.
3. J. K. Kim, J. Scheers, Y. J. Choi, J. H. Ahn and G. C. Hwang, RSC Adv., 2013, 3, 20836.
4. M. S. Islam, D. J. Driscoll, C. A. J. Fisher and P. R. Slater, Chem. Mater., 2005, 17, 5085.
5. C. A. J. Fishera and M. S. Islam, J. Mater. Chem., 2008, 18, 1209.
6. Z. X. Chi, W. Zhang, F. Q. Cheng, J. T. Chen, A. M. Cao and L. J. Wan, RSC Adv., 2014, 4, 7795.
7. Y. Wang, Z. S. Feng, C. Zhang, L. Yu, J. J. Chen, J. Hu and X. Z. Liu, Nanoscale, 2013, 5, 3704.
8. S. Y. Chung, J. T. Bloking and Y. M. Chiang, Nat. Mater., 2002, 1, 123.
9. T. Liu, J. Xu, B. Wu, Q. Xia and X. Wu, RSC Adv., 2013, 3, 13337.
10. B. Wang, B. Xu, T. Liu, P. Liu, C. Guo, S. Wang, Q. Wang, Z. Xiong, D. Wang and X. S. Zhao, Nanoscale, 2014, 6, 986.
11. Y. Wang, Z. S. Feng, J. J. Chen, C. Zhang, X. Jin and J. Hu, Solid State Commun., 2012, 152, 1577.
12. Y. Huang, Y. Xu and X. Yang, Electrochim. Acta, 2013, 113, 156.
13. D. Morgan, A. Van der Ven and G. Ceder, Electrochem. Solid-State Lett., 2004, 7, A30.
14. Z. Wang, X. Niu, J. Xiao, C. Wang, J. Liu and F. Gao, RSC Adv., 2013, 3, 16775.
15. M. Armanda and M. E. Arroyo y de Dompablo, J. Mater. Chem., 2011, 21, 10026.
16. S. Shi, C. Ouyang, M. Lei and W. Tang, J. Power Sources, 2007, 171, 908.
17. H. Zhang, Y. Tang, J. Shen, X. Xin, L. Cui, L. Chen, C. Ouyang, S. Shi and L. Chen, Appl. Phys. A, 2011, 104, 529.
18. Y. Wang, Z. S. Feng, J. Hu, L. Yu, J. J. Chen, L. L. Wang, X. J. Wang and H. L. Tang, Electrochim. Acta, 2014, 117, 431.
19. F. Zhou, K. Kang, T. Maxisch, G. Ceder and D. Morgan, Solid State Commun., 2004, 132, 181.
20. K. Hoang and M. Johannes, Chem. Mater., 2011, 23, 3003.
21. J. Lee, W. Zhou, J. C. Idrobo, S. J. Pennycook and S. T. Pantelides, Phys. Rev. Lett., 2011, 107, 085507.
22. S. P. Ong, L. Wang, B. Kang and G. Ceder, Chem. Mater., 2008, 20, 1798.
23. D. Vanderbilt, Phys. Rev. B: Condens. Matter Mater. Phys., 1990, 41, 7892.
24. H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Solid State, 1977, 16, 1748.
25. A. Yamada, M. Hosoya, S. C. Chung, Y. Kudo, K. Hinokuma, K. Y. Liu and Y. Nishi, J. Power Sources, 2003, 119–121, 232.
26. A. Yamada and S. C. Chung, J. Electrochem. Soc., 2001, 148, A960.
27. M. S. Whittingham, Chem. Rev., 2004, 104, 4271.
28. C. M. Burba and R. Frech, J. Electrochem. Soc., 2004, 151, A1032.
29. X. Lou and Y. Zhang, J. Mater. Chem., 2011, 21, 4156.
30. S. Shi, C. Ouyang, Z. Xiong, L. Liu, Z. Wang, H. Li, D. S. Wang, L. Chen and X. Huang, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 71, 144404.
31. J. Xu and G. Chen, Phys. Rev. B: Condens. Matter Mater. Phys., 2010, 405, 803.
32. A. Yamada, S. C. Chung and K. Hinokuma, J. Electrochem. Soc., 2001, 148, A224.
33. M. S. Kim, J. P. Jegal, K. C. Roh and K. B. Kim, J. Mater. Chem. A, 2014, 2, 10607.
34. A. Yamada, Y. Takei, H. Koizumi, N. Sonoyama, R. Kanno, K. Itoh, M. Yonemura and T. Kamiyama, Chem. Mater., 2006, 18, 804.
35. G. R. Gardiner and M. S. Islam, Chem. Mater., 2010, 22, 1242.
36. W. Huang, S. Tao, J. Zhou, C. Si, X. Chen, W. Huang, C. Jin, W. Chu, L. Song and Z. Wu, J. Phys. Chem. C, 2014, 118, 796.
37. D. B. Ravnsbæk, K. Xiang, W. Xing, O. J. Borkiewicz, K. M. Wiaderek, P. Gionet, K. W. Chapman, P. J. Chupas and Y. M. Chiang, Nano Lett., 2014, 14, 1484.
38. A. Paolella, G. Bertoni, E. Dilena, S. Marras, A. Ansaldo, L. Manna and C. George, Nano Lett., 2014, 14, 1477.

---

Footnote

† Current address: Guangzhou Tinci Materials Technology Co. Ltd., Guangzhou 510760, China.

---