Effects of Fe²⁺ ion doping on LiMnPO₄ nanomaterial for lithium ion batteries

Abstract

To improve the electrochemical performance of LiMnPO₄, a modified polyol reaction has been successfully developed for the preparation of LiMn_xFe_1−xPO₄/C samples (x = 0.2, 0.5, 0.8). The secondary particles of the acquired LiMn_xFe_1−xPO₄/C samples are spherical or annular, and the primary particles are nano-sized (20–80 nm). The LiMn_0.8Fe_0.2PO₄/C has a higher energy density of 637 W h kg⁻¹ compared with LiMn_0.2Fe_0.8PO₄/C (575 W h kg⁻¹) and LiMn_0.5Fe_0.5PO₄/C (584 W h kg⁻¹). The TEM image shows that the primary particles of LiMn_0.8Fe_0.2PO₄/C are uniformly covered by a 3 nm carbon layer, and it can deliver a high discharge capacity of 161 mA h g⁻¹ at 0.05C. At a 0.5C discharge rate, the LiMn_0.8Fe_0.2PO₄/C can maintain 80.4% of its initial capacity after 900 cycles, and it also maintains 90% and 83% of its initial capacity at 45 °C and 55 °C respectively after 100 cycles. The results demonstrate that the modified polyol process is feasible for Fe-doping and carbon-coating to enhance the electrochemical performance of LiMnPO₄.

---

1. Introduction

Lithium-ion batteries (LIBs) are widely applied in consumer electronics. They are one of the most popular types of rechargeable batteries for portable electronics due to their many advantages, such as high energy density, no memory effects, etc.^1,2 Besides, LIBs are also growing in popularity for military, electric vehicle, and aerospace applications due to their excellent properties, which originate from a host of aspects including the cathode, anode, electrolyte, etc.³ As for cathode materials, there are many kinds, including layered oxide LiCoO₂, spinel LiMn₂O₄ and olivine LiFePO₄, which have been widely studied and commercially applied.⁴ As for the potential anode materials, transition metal oxides (including WO₃,⁵ MnO,⁶ MoO₂,⁷ etc.) have shown higher capacities than that of the commercially employed graphite anode (specific capacity of 372 mA h g⁻¹).

In term of redox energy levels, these cathode materials can only be charged and discharged at around 3.7 V, which limits their energy density. Therefore, the development of cathode materials, which can provide high energy density at fast discharge rates, becomes imperative for lithium-ion batteries.^8,9 For these reasons, LiMnPO₄ has become a hotspot of research recently. The most remarkable property of LiMnPO₄ is that the discharge voltage plateau is at around 4.1 V, which enables LiMnPO₄ to deliver a larger power density compared with other cathode materials at the same current density. Unfortunately, under high current density, the accessible capacity and cyclic performance of LiMnPO₄ are still undesirable.¹⁰

To solve the problems, many groups have conducted many research works: (1) One method is reducing the particle size by controlling the morphology and texture of the material.^11–13 For example, Bakenov and Taniguchi¹² demonstrated that the nanoparticles of LiMnPO₄ prepared at a relatively low temperature of 500 °C could deliver a high capacity of about 150 mA h g⁻¹. (2) Another strategy to improve the performance of electrode materials is through doping with metallic ions (such as Mg,¹⁴ Fe,¹⁵ Ni,¹⁶ Cu,¹⁷ Zn¹⁸ etc.). Cation-doping is an effective way to modify the intrinsic properties of electrode materials. Cation-doping in olivine LiMnPO₄ can also be used to further improve its electrical conductivity, favoring fast charge–discharge rates.¹⁵ (3) The surface modification of battery materials is also a rewarding way to improve electronic conductivity, and the most common way is coating with carbon.^6,19,20 A perfect carbon coating can not only reduce the interface contact resistance of LiMnPO₄, but also weaken the dissolution of manganese in the electrolyte.¹³

Based on the above principles, an olivine cathode material for LIBs has been designed and synthesized by a modified polyol reaction. Although the polyol method has been used for the preparation of nano-sized LiMnPO₄ by Wang et al.,²¹ the high specific capacity can only be released at high temperature (50 °C, 159 mA h g⁻¹, 0.1 C). While discharging at high rates (such as 1 C), it can only release 113 mA h g⁻¹ at room temperature for due to its low ion conductivity. So we demonstrate a modified polyol process in which metallic cation ions can be doped into the lattice of LiMnPO₄ by adding an annealing process after the liquid phase synthesis. The ion conductivity has been greatly improved. The acquired LiMn_0.8Fe_0.2PO₄/C demonstrates excellent rate capability and cyclic performance.

2. Experimental

Materials

Manganese acetate tetrahydrate (Mn(CH₃COO)₂·4H₂O, 99.0%) and ferrous sulfate heptahydrate (FeSO₄·7H₂O, 99.0%) were purchased from Alfa Aesar (China) Chemical Co., Ltd. Diethylene glycol (DEG, 99.5%) and lithium dihydrogen phosphate (LiH₂PO₄, 99.9%) were purchased from Shanghai Aladdin Bio-Chem Technology Co., Ltd. Polyvinylidene difluoride (PVDF) and N-methyl-2-pyrrolidinone (NMP) were purchased from Shanghai Kureha Chemical Ltd. Carbon black, Li foil and Celgard 2300 were purchased from Hefei Kejing Material Technology Co., Ltd, China. LiPF₆ (dissolved in ethylene carbonate, ethylene methyl carbonate, and dimethyl carbonate with a volume ratio of 1 : 1 : 1) was purchased from Tianjin Jinniu Power Sources Material Co., Ltd, China. All the chemicals were of analytical grade and were used as-received without further purification.

Synthesis of LiMn_xFe_1−xPO₄/C

The LiMn_xFe_1−xPO₄/C were prepared using the following polyol synthesis. The manganese acetate tetrahydrate and ferrous sulfate heptahydrate were dissolved into deionized water with the ratio of x : (1 − x) (x = 0.2, 0.5, 0.8) and the mixed solution was poured into diethylene glycol in a reactor. This DEG–H₂O solution was vigorously stirred and heated to over 140 °C, keeping it at 140 °C for 1 h, and then the lithium dihydrogen phosphate aqueous solution was dropped into this system. Finally the DEG suspension was kept at this temperature for another 4 h. After cooling down to room temperature, the precursor was separated by centrifugation and washed three times with ethanol in order to remove the residual DEG and organic remnants. Then the precursor was poured into a sucrose solution and dried at 110 °C to obtain a sucrose coating, and then the mixture was annealed for 4 h at 680 °C in a furnace purged with an Ar/H₂ (98 : 2 by vol%) mixture to obtain the samples of LiMn_xFe_1−xPO₄/C (x = 0.2, 0.5, 0.8).

Characterization

The microstructure of the products was examined using a Rigaku D/MAX-2500 X-ray diffractometer with Cu Kα radiation. The XRD patterns were recorded in the 2θ range from 10–90° at room temperature with a tube voltage of 35 kV, a tube current of 30 mA, a rate of 5° min⁻¹ and a step size of 0.02°. The morphology of the products was examined using a Hitachi S4800 field emission scanning microscope and a JEM-2010FEF field emission transmission electron microscope. X-ray photoelectron spectroscopy (XPS, ESCALAB 250) was used to confirm the oxidation state of iron.

Electrochemical measurements

Electrochemical experiments were conducted using a Land CT2001A battery test system. Cathodes for the electrochemical tests were first prepared by mixing the active material with 5 wt% carbon black and 5 wt% PVDF in NMP to form a slurry, and then the cathode slurry was pasted on an Al foil and dried in a vacuum oven at 120 °C over 5 h. The CR2032 button cells were assembled in an argon filled glove box with Li metal as a counter electrode. The electrolyte was 1 M LiPF₆ in ethylene carbonate (EC) : ethylene methyl carbonate (EMC) : dimethyl carbonate (DMC) (1 : 1 : 1 in volume). Galvanostatic charge–discharge measurements were conducted in the voltage range between 2.2 and 4.5 V (vs. Li⁺/Li).

3. Results and discussion

Characterization of LiMn_xFe_1−xPO₄/C (x = 1, 0.8, 0.5, 0.2, 0)

Fig. 1 presents the X-ray diffraction (XRD) patterns of the LiMn_xFe_1−xPO₄/C (x = 1, 0.8, 0.5, 0.2, 0) nanomaterials and LiMnPO₄ nanoparticles (prepared for this study as reported in ref. 22). It is clear that all samples are pure ordered olivine compounds with an orthorhombic structure (space group Pnma). It can be confirmed that the LiMn_xFe_1−xPO₄/C samples (when x = 0.8, 0.5, 0.2) are a solid solution of LiMnPO₄ and LiFePO₄. All the XRD reflection peaks of LiMn_xFe_1−xPO₄/C (x = 1, 0.8, 0.5, 0.2) shift slightly to a lower 2θ value compared with LiFePO₄ (Fig. 2). The lattice volumes calculated from the XRD data are summarized in Table 1. After substituting Mn with Fe, the pure olivine phase is obtained, and the lattice volume decreases from 301.5 ų in LiMnPO₄/C to 299.4 ų in LiMn_0.8Fe_0.2PO₄/C, 297.1 ų in LiMn_0.5Fe_0.5PO₄/C, and 294.3 ų in LiMn_0.2Fe_0.8PO₄/C. These results indicate that the doping has been achieved. Obviously, the crystal volume shrinks by doping with Fe.

Fig. 1 XRD patterns of the samples: LiMnPO₄/C (a), LiMn_0.8Fe_0.2PO₄/C (b), LiMn_0.5Fe_0.5PO₄/C (c), LiMn_0.2Fe_0.8PO₄/C (d), LiFePO₄/C (e).

Fig. 2 Detailed XRD patterns of the samples: (a) LiMnPO₄/C, (b) LiMn_0.8Fe_0.2PO₄/C, (c) LiMn_0.5Fe_0.5PO₄/C, (d) LiMn_0.2Fe_0.8PO₄/C, (e) LiFePO₄/C.

Table 1 Lattice volume of the samples

x	Lattice volume (Å^3)
1	301.5
0.8	299.4
0.5	297.1
0.2	294.3
0	290.1

---

Scanning electron microscopy (SEM) images of LiMn_0.2Fe_0.8PO₄/C (a), LiMn_0.5Fe_0.5PO₄/C (b) and LiMn_0.8Fe_0.2PO₄/C (c) are shown in Fig. 3a–c, respectively. The SEM images show that most of the particles aggregate into secondary particles. The average primary particle sizes of LiMn_0.8Fe_0.2PO₄/C (c) are about 60–80 nm, and the average primary particle sizes of LiMn_0.2Fe_0.8PO₄/C (a) and LiMn_0.5Fe_0.5PO₄/C (b) are about 30–50 nm. The secondary particles of LiMn_0.2Fe_0.8PO₄/C (a) and LiMn_0.5Fe_0.5PO₄/C (b) are spherical or annular. The nanoscale sizes of the samples are desirable owing to decreasing the transport length for Li ions and electrons.^23,24

Fig. 3 Electron microscopy images of the samples: (a) LiMn_0.2Fe_0.8PO₄/C, (b) LiMn_0.5Fe_0.5PO₄/C and (c) LiMn_0.8Fe_0.2PO₄/C, and (d) TEM image and EDS spectrum of LiMn_0.8Fe_0.2PO₄/C.

Fig. 3d shows a TEM image of a LiMn_0.8Fe_0.2PO₄/C particle covered by a 3 nm carbon film, and the atomic layers can also be observed clearly in the image. Obviously, the uniformly coated carbon layer limits the growth and solid-state sintering of primary particles so to maintain the same size. It is indeed the carbon coating layer that assures this favorable configuration since, without its occurrence, an extensive densification of the primary particles would have taken place during the calcination step.²⁵

The energy-dispersive X-ray spectroscopy (EDS) test of LiMn_0.8Fe_0.2PO₄/C was performed as shown in Fig. 3d, and the characteristic peaks of Mn, Fe, P and O elements are found in the EDS spectrum as expected, in which the molar ratio of the Mn : Fe : P is about 0.8 : 0.2 : 1. The results indicate the fairly uniform distribution of the elements in LiMn_0.8Fe_0.2PO₄/C.

XPS was used to analyze the chemical state of iron in the LiMn_0.8Fe_0.2PO₄/C sample. As shown in Fig. 4, the Fe 2p spectrum is split in two parts due to spin–orbit coupling, namely Fe 2p3/2 and Fe 2p1/2. The binding energy positions at 710.6 eV and 724.6 eV are ascribed to Fe 2p3/2 and Fe 2p1/2, respectively,^19,26 which perfectly match with previously reported spectra of LiFePO₄ and are characteristic of Fe²⁺ cations.²⁷

Fig. 4 XPS spectrum of Fe 2p in LiMn_0.8Fe_0.2PO₄/C.

Electrochemical performance of LiMn_xFe_1−xPO₄/C (x = 1, 0.8, 0.5, 0.2, 0)

The samples with carbon contents of 1%, 3%, 5%, 7% and 10% (wt) were also prepared from LiMn_0.8Fe_0.2PO₄/C. The influence of carbon content on the cycle performance of the samples is discussed in Fig. 5 and Table 2. It is obvious that with increasing the carbon content, the discharge capacity of the samples increases at first, and then decreases. This phenomenon is mainly because the electronic conductivity of LiMn_0.8Fe_0.2PO₄/C increases with increasing the carbon content before reaching the optimal content. So the polarization in the electrode will be weakened, which leads to the higher specific capacity. However, after the optimal carbon content, the ratio of LiMn_0.8Fe_0.2PO₄ in the electrode active material will be gradually reduced and there is no obvious increase of the conductivity, which causes the decrease of the specific capacity. When the carbon content reaches 5%, the sample has the highest specific capacity (158 mA h g⁻¹ at 0.5C), capacity retention (100% after 100 cycles), and coulombic efficiency for the first cycle (98.1%).

Fig. 5 The influence of carbon content on the cycle performance of LiMn_0.8Fe_0.2PO₄/C for 1%, 3%, 5%, 7% and 10% (wt).

Table 2 Influence of carbon content on electrochemical performance

Carbon content (wt%)	1	3	5	7	10
Capacity in first cycle (mA h g^−1)	121	128	158	144	139
Capacity retention after 100 cycles (%)	94	98.4	100	100	100
Columbic efficiency (%)	82.4	88.2	98.1	95.3	96.5

---

Additionally, the typical charge–discharge curves of the samples at 0.05C are shown in Fig. 6a. When discharged at a slow rate of 0.05C, all the Fe-doped materials can deliver a discharge capacity of about 160 mA h g⁻¹, and the coulombic efficiencies in the first cycle are all above 90%. Two voltage plateaus can be observed in all the discharge curves at 0.05C. One plateau is located at around 4.0 V and the other plateau is located at around 3.4 V. The capacity releasing around 4.0 V can be attributed to the redox reaction of Mn²⁺/Mn³⁺, and the capacity releasing around 3.4 V can be attributed to that of Fe²⁺/Fe³⁺ for LiMn_xFe_1−xPO₄/C (x = 0.8, 0.5, 0.2).^28,29 This result also indicates the existence of Fe²⁺ and it is consistent with the XPS result above. It is apparent that the plateaus of LiMn_0.5Fe_0.5PO₄/C and LiMn_0.2Fe_0.8PO₄/C at 3.4 V contribute much more capacity than those of LiMn_0.8Fe_0.2PO₄/C, and these results indicate more Fe²⁺ ions in LiMn_0.5Fe_0.5PO₄/C and LiMn_0.2Fe_0.8PO₄/C. The ionic radius of Mn²⁺ is larger than that of Fe²⁺, thus the lattice volume of LiMn_0.5Fe_0.5PO₄/C and LiMn_0.2Fe_0.8PO₄/C should be smaller than that of LiMn_0.8Fe_0.2PO₄/C. This result is consistent with the XRD results.

Fig. 6 Discharge curves of the different cathodes at different rates: (a) the typical charge–discharge curves of the four samples at 0.05C, (b) LiMn_0.8Fe_0.2PO₄/C, (c) LiMn_0.5Fe_0.5PO₄/C, (d) LiMn_0.2Fe_0.8PO₄/C.

From Fig. 6a, the reversible voltage profiles can be observed, with the samples showing reversible discharge capacities of 163 mA h g⁻¹ (LiMn_0.2Fe_0.8PO₄/C), 162 mA h g⁻¹ (LiMn_0.5Fe_0.5PO₄/C) and 161 mA h g⁻¹ (LiMn_0.8Fe_0.2PO₄/C), and measured energy densities of 575 W h kg⁻¹ (LiMn_0.2Fe_0.8PO₄/C), 584 W h kg⁻¹ (LiMn_0.5Fe_0.5PO₄/C) and 637 W h kg⁻¹ (LiMn_0.8Fe_0.2PO₄/C). The energy density of LiMn_0.8Fe_0.2PO₄/C is a significant improvement over LiMn_0.2Fe_0.8PO₄/C and LiMn_0.5Fe_0.5PO₄/C.

The discharge capacity and capacity retention of the different cathodes at different rates are shown in Fig. 6b–d and 7. When the LiMn_0.8Fe_0.2PO₄/C electrode is discharged at 10C, a capacity of about 93 mA h g⁻¹ can be released. Meanwhile, the LiMn_0.5Fe_0.5PO₄/C and LiMn_0.2Fe_0.8PO₄/C samples can deliver discharge capacities of 89 mA h g⁻¹ and 91 mA h g⁻¹ even at a discharge rate of 10C.

Fig. 7 Discharge capacity retention of the materials at different rates: (a) LiMn_0.8Fe_0.2PO₄/C, (b) LiMn_0.5Fe_0.5PO₄/C, (c) LiMn_0.2Fe_0.8PO₄/C.

Fig. 7 shows the capacity retention of LiMn_0.8Fe_0.2PO₄/C, LiMn_0.5Fe_0.5PO₄/C and LiMn_0.2Fe_0.8PO₄/C at different rates. All of them retain about 95% of the initial discharge capacity at a discharge rate of 0.5C. It is interesting to note that LiMn_0.8Fe_0.2PO₄/C can retain nearly 97% of the initial discharge capacity when discharged at 0.5C. At 10C, LiMn_0.8Fe_0.2PO₄/C mainly discharges at about 3.2 V, while LiMn_0.5Fe_0.5PO₄/C and LiMn_0.2Fe_0.8PO₄/C mainly discharge at around 2.8 V and 2.5 V. A larger amount of Mn can supply a higher discharge plateau potential which can lead to a higher energy density.

Polarization is caused by limited electronic conductivity and a slow Li ion transfer rate, leading to deterioration in the performance of the cathode electrode at high current density. The polarization can be decreased by using nano-crystallization and metal element doping methods. Therefore, it is reasonable to deduce that the transfer rate of Li⁺ has been greatly enhanced by Fe-doping.

The cyclic performance of LiMn_0.8Fe_0.2PO₄/C under a discharge rate of 0.5C is shown in Fig. 8a. It is apparent that the LiMn_0.8Fe_0.2PO₄/C electrode can deliver a capacity of 158 mA h g⁻¹ at the initial cycles and release 80.4% of the initial capacity after 900 cycles. The capacity fading can be ascribed to structural instability and an increase in the impedance of the cell which would result in the polarization.³⁰ The high capacity of LiMn_0.8Fe_0.2PO₄/C is mainly due to minimized polarization and improved electrical conductivity as described above, and the good capacity retention should result from enhanced structure stability.

Galvanostatic charge–discharge measurements of LiMn_0.8Fe_0.2PO₄/C at a 0.5C rate (1 C = 170 mA g⁻¹) in the voltage range of 2.2–4.5 V are shown in Fig. 8b. The charge–discharge capacities in the range of 2.2–4.5 V in the 1st, 5th, 10th, 100th, 500th and 900th cycles are 161/158 mA h g⁻¹, 159/157 mA h g⁻¹, 159/158 mA h g⁻¹, 156/156 mA h g⁻¹, 146/146 mA h g⁻¹ and 126/126 mA h g⁻¹ respectively. The corresponding coulombic efficiencies are 98.1% for the 1st cycle, 98.7% for the 5th cycle, 99.3% for the 10th cycle, 100% for the 100th cycle, 100% for the 500th cycle and 100% for the 900th cycle.

Fig. 8 Cyclic performance of LiMn_0.8Fe_0.2PO₄/C: (a) 900 cycles of LiMn_0.8Fe_0.2PO₄/C, (b) charge–discharge curves of different cycles.

It is known that there are two main reasons which lead to the deterioration of the cycle life of LiMnPO₄: one is the Jahn–Teller effect caused by Mn³⁺, the other is the dissolution of manganese in the electrolyte which is more serious at high temperature. To verify the cycle performance of the materials at high temperature, the high temperature cycle performance of LiMn_0.8Fe_0.2PO₄/C is compared with LiMn_0.8Fe_0.2PO₄ (without a carbon coating). The cycle performance of the materials at high temperature was measured at 45 °C and 55 °C. From the results of Fig. 9, the LiMn_0.8Fe_0.2PO₄/C electrode delivers 90% of the initial capacity after 100 cycles at 45 °C and 83% of the initial capacity at 55 °C, while LiMn_0.8Fe_0.2PO₄ only releases 61% of the initial capacity at 45 °C and 34% at 55 °C. So it can be concluded that carbon coating is effective in restraining the dissolution of manganese, and good cyclic performance at high temperature can be promoted.

Fig. 9 High temperature cyclic performance of LiMn_0.8Fe_0.2PO₄/C and LiMn_0.8Fe_0.2PO₄: (a) LiMn_0.8Fe_0.2PO₄/C at 45 °C, (b)LiMn_0.8Fe_0.2PO₄/C at 55 °C, (c) LiMn_0.8Fe_0.2PO₄ at 45 °C, (d) LiMn_0.8Fe_0.2PO₄ at 55 °C.

4. Conclusion

In summary, to improve the rate capability and cyclic performance of olivine LiMnPO₄, high performance Fe-doped LiMn_xFe_1−xPO₄/C was prepared by a modified polyol method. The resulting olivine LiMn_0.8Fe_0.2PO₄/C sample not only shows excellent rate capability (93 mA h g⁻¹ at 10C) and high energy density (637 W h kg⁻¹), but also exhibits a great improvement in cyclic performance (retaining 80.4% after 900 cycles, and 90% at 45 °C and 83% at 55 °C after 100 cycles) at the charge–discharge rate of 0.5C. It is believed that LiMn_0.8Fe_0.2PO₄/C is a promising cathode material for high power applications in the future.

References

1. M. Winter and R. Brodd, Chem. Rev., 2004, 104, 4245–4270.
2. M. Yoshio, Y. Xia and N. Kumada, et al., J. Power Sources, 2001, 101, 79–85.
3. L. Liu, Y. Tian, Y. Zhai and C. Xu, Trans. Nonferrous Met. Soc. China, 2007, 17, 110–115.
4. H. Xu, X.-J. Liu, Q.-W. Peng and Z.-W. Lu, Mater. Express, 2014, 4, 72–78.
5. H.-G. Wei, D.-W. Ding, X.-R. Yan, J. Guo, L. Shao, H.-R. Chen, L.-Y. Sun, H. A. Colorado, S.-Y. Wei and Z.-H. Guo, Electrochim. Acta, 2014, 132, 58–66.
6. H.-L. Cao, X.-Z. Wang, H.-B. Gu, J.-R. Liu, L.-Q. Luan, W. Liu, Y.-R. Wang and Z.-H. Guo, RSC Adv., 2015, 5, 34566–34571.
7. S. Qiu, G.-X. Lu, J.-R. Liu, H.-L. Lyu, C.-X. Hu, B. Li, X.-R. Yan, J. Guo and Z.-H. Guo, RSC Adv., 2015, 5, 87286–87294.
8. K. Kang, Y.-S. Meng, J. Bréger, C. P. Grey and G. Ceder, Science, 2006, 311, 977–980.
9. J. Molenda and J. Marzec, Funct. Mater. Lett., 2008, 1, 91–95.
10. K. Saravanan, V. Ramar, P. Balaya and J. J. Vittal, J. Mater. Chem., 2011, 21, 14925–14935.
11. N.-H. Kwon, T. Drezen, I. Exnar, I. Teerlinck, M. Isono and M. Graetzel, Electrochem. Solid-State Lett., 2006, 9, 277–280.
12. Z. Bakenov and I. Taniguchi, Electrochem. Commun., 2010, 12, 75–78.
13. S. K. Martha, B. Markovsky, J. Grinblat, Y. Gofer, O. Haik, E. Zinigrad, D. Aurbacha, T. Drezen, D. Wang, G. Deghenghi and I. Exnar, J. Electrochem. Soc., 2009, 156, 541–552.
14. T. Shiratsuchi, S. Okada, T. Doi and J. Yamaki, Electrochim. Acta, 2009, 54, 3145–3151.
15. G. Yang, H. Ni, H.-D. Liu, P. Gao, H.-M. Ji, S. Roy, J. Pinto and X.-F. Jiang, J. Power Sources, 2011, 196, 4747–4755.
16. C. A. J. Fisher, V. M. Hart Prieto and M. Saiful Islam, Chem. Mater., 2008, 20, 5907–5915.
17. J. Ni and L. Gao, J. Power Sources, 2011, 196, 6498–6501.
18. C.-X. Hu, S.-M. Guo, G.-X. Lu, Y. Fu, J.-R. Liu, H.-G. Wei, X.-R. Yan, Y.-R. Wang and Z.-H. Guo, Electrochim. Acta, 2014, 148, 118–126.
19. S.-M. Guo, J.-R. Liu, S. Qiu, Y.-R. Wang, X.-R. Yan, N.-N. Wu, S.-Y. Wang and Z.-H. Guo, Electrochim. Acta, 2016, 190, 556–565.
20. S.-M. Guo, J.-R. Liu, S. Qiu, W. Liu, Y.-R. Wang, N.-N. Wu, J. Guo and Z.-H. Guo, J. Mater. Chem. A, 2015, 3, 23895–23904.
21. D.-Y. Wang, H. Buqa, M. Crouzet, G. Deghenghi, T. Drezen, I. Exnar, N.-H. Kwon, J. H. Miners, L. Poletto and M. Grätzel, J. Power Sources, 2009, 189, 624–628.
22. J. Zong and X.-J. Liu, Electrochim. Acta, 2014, 116, 9–18.
23. C. Delacourt, P. Poizot, S. Levasseur and C. Masquelier, Electrochem. Solid-State Lett., 2006, 9, A352–A355.
24. D.-H. Kim and J. Kim, Electrochem. Solid-State Lett., 2006, 9, A439–A442.
25. H.-G. Jung, S.-T. Myung, C. S. Yoon, S.-B. Son, K. H. Oh, K. Amine, B. Scrosati and Y.-K. Sun, Energy Environ. Sci., 2011, 4, 1345–1351.
26. Y.-R. Wang, Q.-L. He, J. Guo, J.-M. Wang, Z.-P. Luo, T. D. Shen, K.-Q. Ding, A. Khasanov, S.-Y. Wei and Z.-H. Guo, ACS Appl. Mater. Interfaces, 2015, 7, 23920–23931.
27. M. S. Bhuvaneswari, N. N. Bramnik, D. Ensling, H. Ehrenberg and W. Jaegermann, J. Power Sources, 2008, 180, 553.
28. X.-Y. Chang, Z.-X. Wang, X.-H. Li, L. Zhang, H.-J. Guo and W.-J. Peng, Mater. Res. Bull., 2005, 40, 1513–1520.
29. C.-H. Mi, X.-G. Zhang, X.-B. Zhao and H.-L. Li, Mater. Sci. Eng., B, 2006, 129, 8–13.
30. J. Zhou and P. H. L. Notten, J. Power Sources, 2008, 177, 553–560.

---