Synthesis and performance of hollow LiNi_0.5Mn_1.5O₄ with different particle sizes for lithium-ion batteries

Abstract

High voltage spinel LiNi_0.5Mn_1.5O₄ is a promising cathode material for next generation lithium ion batteries. A simple method of synthesizing hollow LiNi_0.5Mn_1.5O₄ spinel using MnCO₃ as the manganese resource is presented. The hollow structure forms during the calcination process at 850 °C. The transformation from MnCO₃ to manganese oxide and inter-diffusion of Mn and Ni atoms are excluded as reasons for the formation of the hollow structure. Four hollow LiNi_0.5Mn_1.5O₄ samples with different particle sizes were synthesized by controlling the reactant concentration. The effects of particle size on the electrochemical performance of hollow LiNi_0.5Mn_1.5O₄ have been investigated in detail. The hollow LiNi_0.5Mn_1.5O₄ samples with particle size less than 1 μm and some small broken particles of about 200 nm show poor rate capability and cycling performance due to their poor contact with conductive additive and high interface resistance. The hollow LiNi_0.5Mn_1.5O₄ samples with diameters of 2 or 6 μm exhibit better rate capability and cycling performance. This is because most of the micro-sized particles can make direct contact with the conductive additive and have low interface resistance; moreover, the hollow structure also decreases the Li⁺ ion and electron diffusion distance.

---

Introduction

As a result of the escalating energy crisis and greenhouse gas emission issues, green energy sources and electric vehicles have received much attention. High energy and high power rechargeable lithium-ion batteries have become dominant energy storage systems. One strategy for increasing the energy density of lithium ion batteries is to increase its operation voltage by utilizing high-voltage positive electrodes. Among cathode materials, LiNi_0.5Mn_1.5O₄ has received much attention for its excellent electrochemical properties with a potential plateau at around 4.7 V and low cost.^1–5 In addition, LiNi_0.5Mn_1.5O₄ has high power density due to its facile three dimensional lithium-ion diffusion.^6,7 The nominal spinel LiNi_0.5Mn_1.5O₄ has two different space groups: Fd3̄m with disordered Ni and Mn on the octahedral sites and P4₃32 with ordered Ni and Mn.^8,9 LiNi_0.5Mn_1.5O₄ with disordered Fd3̄m has a higher Li⁺ ion diffusion coefficient and greater electronic conductivity afforded by charge transfer between Ni and Mn.^10,11

To improve the rate capability of electrode material, nanosized materials have attracted considerable attention. Nanosized LiNi_0.5Mn_1.5O₄ was synthesized with polyethyleneglycol^12,13 or polyvinylpyrrolidone additives.^14,15 These additives are used as templates or dispersants. The nanosized LiNi_0.5Mn_1.5O₄ material can decrease the lithium ion and electron diffusion length. However, nanosized materials have some disadvantages such as low volumetric density and high interface resistance.^16,17 Moreover, the large specific surface areas of nanosized materials will cause serious side reactions between electrode and electrolyte, which is harmful to their cycling performance.

Another way to improve rate capability is to use hollow and hierarchical materials. LiNi_0.5Mn_1.5O₄ materials with hollow structures were synthesized using MnO₂ (ref. 18 and 19) or Mn₂O₃ (ref. 20) as the manganese resource. MnO₂ and Mn₂O₃ are obtained by calcining MnCO₃ at 400 °C and 600 °C, respectively. The shell provides short distances for Li⁺ and electron diffusion, leading to better rate capability. The interior void space can buffer the structural strain and volume change associated with the repeated Li⁺ insertion/extraction processes, thus improving the cycling stability.¹⁸ However, the effects of practical size on performance for hollow materials have not been studied systematically.

Herein, spinel LiNi_0.5Mn_1.5O₄ with a hollow structure was synthesized by a simple method using MnCO₃ as the manganese resource. The formation process of hollow structures was discussed in detail. Moreover, four hollow LiNi_0.5Mn_1.5O₄ samples with different particle sizes were synthesized by changing the size of MnCO₃. The effects of particle size on the electrochemical performance of hollow LiNi_0.5Mn_1.5O₄ were systematically investigated.

Experimental

Preparation of MnCO₃

MnCO₃ microspheres were prepared by a precipitation method described as outlined below.²¹ 200 mL ethanol was added to 1 L of MnSO₄ solution under stirring. Subsequently, 1 L of NH₄HCO₃ solution was quickly added to the abovementioned mixture and further stirred for 5 h. The mixture was then centrifuged and washed with water and ethanol several times. The as-obtained MnCO₃ microspheres were dried at 60 °C. The concentration of the NH₄HCO₃ solution was ten times that of MnSO₄ solution. The MnSO₄ solution concentrations were 0.01, 0.02, 0.04 and 0.08 mol L⁻¹, respectively, and the prepared MnCO₃ microspheres were denoted as MnCO₃-1, MnCO₃-2, MnCO₃-4 and MnCO₃-8.

Preparation of LiNi_0.5Mn_1.5O₄

Spinel LiNi_0.5Mn_1.5O₄ samples were prepared using the abovementioned MnCO₃ as the manganese resource and denoted as LNMO-1, LNMO-2, LNMO-4 and LNMO-8. Stoichiometric amounts of LiOH·H₂O, Ni(NO₃)₂·6H₂O and MnCO₃ taken in the ratio Li : Ni : Mn = 1.05 : 0.5 : 1.5 were dispersed in ethanol. Ethanol was slowly evaporated at room temperature under stirring. The mixture obtained after evaporation was calcined at 400 °C for 4 hours and at 850 °C for 10 hours and then cooled at a rate of 0.5 °C min⁻¹ in air to obtain the product.

Characterization

The obtained materials were characterized by scanning electron microscopy (SEM) and powder X-ray diffraction (XRD). XRD characterization was carried out with a D/max-RB diffractometer using a Cu Kα source and recorded with a step of 0.05°.

Electrochemical tests of the homemade samples were carried out using coin-type cells (2025). The cathode materials of the cells were made from a slurry containing 80 wt% active material, 10 wt% conductive acetylene black as a conductive agent, and 10 wt% polyvinylidene fluoride (PVDF) as a binder dissolved in n-methyl pyrrolidinone. The slurry was evenly coated onto an aluminum foil and then dried in a vacuum oven at 120 °C overnight. The foil was then punched into a circular electrode (1.4 cm in diameter). The loading weight of the active materials on the electrode was about 2 mg cm⁻². Cells with lithium metal as the counter electrode were assembled in an argon-filled glove box. The electrolyte was 1 mol L⁻¹ LiPF₆ in a mixture of ethylene carbonate and dimethyl carbonate with a ratio of 1 : 1 by weight. Charge–discharge tests were carried out on a NEW WARE battery tester. When the current densities were higher than 0.5C, the cells were first galvanostatically charged to 4.95 V. Cell voltage was kept at 4.95 V until the current decreased to 0.1C. The cells were then discharged to 3.5 V at different rates. Cyclic voltammetry (CV: 3.5–5.1 V, 0.1 mV s⁻¹) and electrochemical impendence spectroscopy (EIS) were carried out on a CHI650D electrochemical workstation. EIS measurements were conducted with an AC amplitude of 5 mV at 4.73 V in the frequency range from 10⁵ Hz to 0.01 Hz.

Results and discussion

The XRD patterns of MnCO₃-1, MnCO₃-2, MnCO₃-4 and MnCO₃-8 are shown in Fig. 1(a). The XRD patterns are indexed to the rhombohedral structure of MnCO₃ (JCPDS card no.: 83-1763). The diffraction peak intensity of MnCO₃-1 is stronger than those of the others, indicating its larger single-crystal size.

Fig. 1 XRD patterns of four MnCO₃ samples (a) and four LiNi_0.5Mn_1.5O₄ samples: full range (b) and enlarged region (c).

The XRD patterns of the four LNMO samples are shown in Fig. 1(b and c). All patterns can be assigned to cubic spinel LiNi_0.5Mn_1.5O₄ (JCPDS card no.: 80-2162). These LNMO samples have the Fd3̄m space group, as revealed by the absence of superstructure reflections at 15.3°, 24.3°, 34.7° and 41.3° in Fig. 1(c), which originates from the ordering of Mn⁴⁺ and Ni²⁺ ions in the 16d octahedral sites.^22–24 Moreover, there are some minor residual peaks at 37.5°, 43.6°, and 63.4° that can be attributed to Li_xNi_1−xO impurity.^25,26 The impurity contents of these samples are low and similar, as shown in Fig. 1(b and c). The low impurity content is beneficial for the electrochemical properties because impurities can block Li⁺ mobility in the material and lead to poor performance.²⁵

SEM images of the MnCO₃ samples prepared with different concentrations of MnSO₄ solution are shown in Fig. 2(a–d). All the particles have spherical morphology. MnCO₃ particle size gradually decreases with increasing MnSO₄ solution concentration due to the increasing nucleation rate.

Fig. 2 SEM micrographs of four MnCO₃ samples (a–d) and four LiNi_0.5Mn_1.5O₄ samples (e–l).

SEM images of the four LNMO samples are shown in Fig. 2(e–l). All four samples retain the spherical MnCO₃ shape. The average particle sizes of LNMO-1, LNMO-2, LNMO-4 and LNMO-8 are around 5.8, 2.1, 1.1 and 0.9 μm in diameter, respectively. That is to say, with increasing MnSO₄ solution concentration, LNMO particle size decreases gradually in a similar manner to MnCO₃. Therefore, the particle size of LiNi_0.5Mn_1.5O₄ can be controlled by controlling precursor size, which is determined by reactant concentration.

It is interesting that all LNMO samples have hollow structures, as shown in Fig. 2(i–l). The shell thickness of LNMO-1, LNMO-2, LNMO-4 and LNMO-8 is around 1.3, 0.5, 0.2 and 0.2 μm, respectively. It can be noted that the LNMO samples' shells have hierarchical structures and are composed of interconnected primary particles. The primary particles of LNMO-2, LNMO-4 and LNMO-8 have similar sizes of about 200 nm. The primary particles of LNMO-1 are larger. Moreover, the shell of partial LNMO-4 and LNMO-8 breaks into small particles of about 200 nm, especially in LNMO-8.

To understand the LNMO hollow structure formation process, LNMO-4 is taken as an example. The mixture before calcination, when the temperature reached 850 °C, and after calcination at 850 °C for 1, 2, 4, 6, and 10 hours was characterized by SEM, as shown in Fig. S1.† It can be seen that the hollow structure formed during the calcination process at 850 °C. As shown in Fig. S1(b),† no hollow structure was formed when the temperature reached 850 °C, at which MnCO₃ has transformed to manganese oxide.^27,28 This indicates that the transformation from MnCO₃ to manganese oxide is not the reason for the formation of hollow LNMO.

During the synthesis process, MnCO₃, Ni(NO₃)₂ and LiOH were dispersed into ethanol to mix these reactants, in which Ni(NO₃)₂ dissolved in ethanol. Ethanol was then evaporated under stirring. After the evaporation of ethanol, dissolved Ni(NO₃)₂ will deposit on the surface of solid MnCO₃ microspheres, as shown in Fig. S1(a).† The inter-diffusion of Mn and Ni atoms during calcination may be the reason for the formation of the hollow LNMO. To validate this assumption, stoichiometric amounts of LiOH and MnCO₃-4 taken in the ratio Li : Mn = 1 : 2 were mixed without Ni(NO₃)₂, and calcined under the same conditions as LNMO. SEM images of the product are shown in Fig. S2.† As seen, the obtained product still has a hollow structure without inter-diffusion of Mn and Ni atoms. Therefore, the hollow structure of LNMO is not caused by the inter-diffusion of Mn and Ni atoms. After exclusion of the abovementioned causes, the formation of the hollow structure may be caused by fast outward diffusion of Mn and Ni atoms and the slow inward diffusion of O atoms during the calcination process at 850 °C.¹⁸

In order to demonstrate the effects of the microstructure on the electrochemical performance of the spinel products, cells using the spinel products as positive electrodes were cycled at various rates. The rate capability and cycling performance of these cells were compared.

Fig. 3 displays the charging and discharging curves of the four LiNi_0.5Mn_1.5O₄ samples. The discharging curves at 0.2C show a dominant plateau at around 4.7 V, which is attributed to a Ni²⁺/Ni⁴⁺ redox couple. A minor plateau in the 4 V region is also observed and associated with the Mn³⁺/Mn⁴⁺ redox couple. The length of the 4 V plateau in the discharge curves can be used to evaluate the relative Mn³⁺ content of the spinel.^29,30 A longer plateau indicates a higher Mn³⁺ content. The discharge capacity proportion in the 4 V region at 0.2C for LNMO-1, LNMO-2, LNMO-4 and LNMO-8 is 21.7%, 9.8%, 14.6% and 7.0%, respectively. This means that LNMO-1 has the highest Mn³⁺ content. The high Mn³⁺ content may be caused by the inhomogeneous Ni/Mn distribution in LNMO-1.³¹ With the largest MnCO₃ particle sizes, Ni and Mn atoms have to migrate over a distance as long as several microns to form LNMO-1. This long distance leads to inhomogeneous distribution of the elements, even after calcination.

Fig. 3 Charge and discharge capacity curves of four LiNi_0.5Mn_1.5O₄ samples.

As shown in Fig. 3, with increasing current density, the voltage difference between the charging and discharging plateau increases, meaning increasing polarization. A smaller voltage difference between the charging and discharging plateau indicates a better rate capability. Moreover, the rate capability can be also evaluated by the proportion of constant current charge step in the total charge capacity at high currents.¹⁸ The constant current charge step contributes 55.0%, 74.4%, 49.3% and 0% to the total charge capacity at 5C in LNMO-1, LNMO-2, LNMO-4 and LNMO-8, respectively. The results indicate that LNMO-2 has the best rate capability and LNMO-8 has the worst.

To further evaluate the rate capability, the LNMO samples were cycled at various discharge rates ranging from 0.2 to 20C, as shown in Fig. 4(a). For ease of comparison, the discharge capacity values at various C rates are normalized to the discharge capacity value at the 0.2C rate and plotted in Fig. 4(b).

Fig. 4 Discharge capacity and rate capability retentions of four LiNi_0.5Mn_1.5O₄ samples at different discharge rates.

Because the constant current charge step is followed by an additional constant voltage charge step until the current drops to 0.1C when the rate ≥ 1C, the discharge capacity at 1C is higher than that at 0.5C. As the current density increases from 0.2C to 20C, the discharge capacities of LNMO-1 and LNMO-2 decrease gradually and linearly. However, the discharge capacity of LNMO-4 and LNMO-8 decreases rapidly when the current density is greater than 10C and 5C, respectively. That is to say, the rate capability of LNMO-4 and LNMO-8 is worse than that of LNMO-1 and LNMO-2, especially at high current density. LNMO-4 and LNMO-8 can hardly deliver capacity at 20C. However, LNMO-1 and LNMO-2 are still able to deliver a capacity as high as 51.5% and 60.9% of discharge capacity at 0.2C.

Mn³⁺ in spinel can enhance conductivity, which is beneficial to high rate performance.^32,33 However, as discussed above, the Mn³⁺ content of LNMO-1 and LNMO-4 is greater than that of LNMO-2 and LNMO-8. Therefore, the rate capability differences of the four samples are more related to differences in their microstructures. Different microstructures lead to different conductive structures. The SEM images of the four LNMO electrodes are shown in Fig. 5. The very small particles of about 50 nm are conductive additive (super-P). As seen in Fig. 5, due to their small particle sizes, only a small number of LNMO-4 and LNMO-8 particles can directly contact the conductive additive and many interfaces exist between the particles. Only conductive additive can conduct large electric current, so the discharge capacities of LNMO-4 and LNMO-8 decrease rapidly when the current density is higher than 10C and 5C, respectively. LNMO-8 has more small broken particles of about 200 nm, resulting in its having the worst rate capability. For micro-sized LNMO-1 and LNMO-2, most of the particles can directly contact the conductive additive and few interfaces exist between the micro-sized particles due to their large particle size. In addition, the hollow structure can decrease the Li⁺ ion and electron diffusion distance. Therefore, LNMO-1 and LNMO-2 exhibit excellent rate capability. The shell of LNMO-2 is thinner than that of LNMO-1. Therefore, LNMO-2 exhibits better rate capability. It can be seen from the abovementioned results that large decreases in particle size are unfavorable for enhancing rate capability. In addition, a micro-sized particle with a hollow structure is optimal for rate capability because it has low interface resistance while decreasing the Li⁺ ion and electron diffusion distance. The phenomena that hollow LNMO samples with large particle sizes exhibit better rate capability¹⁸ and hollow LNMO samples with small particle sizes exhibit poor rate capability³⁴ were also reported by other groups.

Fig. 5 SEM micrographs of the four LiNi_0.5Mn_1.5O₄ electrodes.

To further verify the abovementioned discussion, cyclic voltammetry and electrochemical impendence spectroscopy measurements were performed on the cells. CV curves of the LNMO samples are compared and shown in Fig. 6(a). The main peaks around 4.7 V are attributed to a Ni²⁺/Ni⁴⁺ redox couple and the weak peaks at 4.0 V are attributed to a Mn³⁺/Mn⁴⁺ redox couple. The peaks at 4.0 V are related to the Mn³⁺ content of spinel.^35,36 A comparison of the peaks at 4.0 V shows that the Mn³⁺ content of LNMO-1 and LNMO-4 is greater than that of LNMO-2 and LNMO-8, which is consistent with the estimated results from the discharging curves. The gap between the oxidation and reduction peaks of LNMO-8 is much larger than those of the others, which suggests that its poor electrochemical reversibility is attributed to poor conductive structures and low Mn³⁺ content.

Fig. 6 Cyclic voltammetry curves (a) and electrochemical impendence spectroscope (b) of four LiNi_0.5Mn_1.5O₄ samples.

The Nyquist plots of EIS are shown in Fig. 6(b). The EIS was fitted simply by the equivalent circuit inserted in Fig. 6(b). As shown, each plot consists of a depressed semicircle in the high frequency region and a sloping line at the low frequency range. The depressed semicircle reflects the interface impendence (R), including the interfacial layer and charge transfer reaction.^37,38 The values of R are 85.7, 81.2, 83.1 and 453.8 Ω for LNMO-1,2,4,8, respectively. The interface impendence of LNMO-8 is greater than that of the others, which is due to many interfaces between the particles and the lowest Mn³⁺ content in LNMO-8.

In addition to rate capability, the LNMO sample cycling performance at charging/discharging rates of 1C, 2C and 5C are compared in Fig. S3† and the discharge capacity retentions are plotted in Fig. 7. LNMO-1 and LNMO-2 show better capacity retention than LNMO-4 and LNMO-8. The capacity retention values of LNMO-2 are all over 90% after 300 cycles at 1C, 500 cycles at 2C and 500 cycles at 5C. The hollow structure can buffer the structural strain and volume change during cycling, thus improving the cycling stability.¹⁸ However, the hollow structure of LNMO-4 and LNMO-8 is not stable due to its thin shell, as shown in Fig. 2(g and h). The poor cycling performance seen in LNMO-4 and LNMO-8 is due to their large specific surface areas, resulting in serious side reactions such as decomposition of the electrolyte and dissolution of Mn. With larger particle sizes, LNMO-1 and LNMO-2 have smaller specific surface areas.

Fig. 7 Capacity retention of four LiNi_0.5Mn_1.5O₄ samples.

Conclusions

Hollow spinel LiNi_0.5Mn_1.5O₄ was synthesized by a simple method using MnCO₃ as the manganese resource. The hollow structure forms during the calcination process at 850 °C and is not caused by transformation from MnCO₃ to manganese oxide or inter-diffusion of Mn and Ni atoms. The particle size of hollow LiNi_0.5Mn_1.5O₄ is controlled by changing precursor size, which is controlled by reactant concentration. An excessive decrease in particle size is unfavorable for enhancing rate capability because only few particles can directly contact the conductive additive and many interfaces exist between the particles when they are too small. The LiNi_0.5Mn_1.5O₄ sample, which is 2 μm in diameter, with a 500 nm thick shell exhibits the best rate capability and cycling performance. This is attributed to the low interface resistance and good contact between micro-sized particles and the conductive additive. Moreover, the 500 nm thick shell decreases the Li⁺ ion and electron diffusion distance.

Acknowledgements

We thank the National Natural Science Foundation of China (Grant No. 21273058), the China Postdoctoral Science Foundation (Grant No. 2012M520731 and No. 2014M70350) and the Heilongjiang Postdoctoral Financial Assistance (LBH-Z12089) for their financial support.

Notes and references

1. J. Hassoun, K. S. Lee, Y. K. Sun and B. Scrosati, J. Am. Chem. Soc., 2011, 133, 3139.
2. H. G. Jung, M. W. Jang, J. Hassoun, Y. K. Sun and B. Scrosati, Nat. Commun., 2011, 2, 516–520.
3. J. B. Goodenough and K. S. Park, J. Am. Chem. Soc., 2013, 135, 1167.
4. J. B. Goodenough, Energy Environ. Sci., 2014, 7, 14.
5. O. K. Park, Y. Cho, S. Lee, H. C. Yoo, H. K. Song and J. Cho, Energy Environ. Sci., 2011, 4, 1621.
6. P. Strobel, A. I. Palos, M. Anne and F. L. Cras, J. Mater. Chem., 2000, 10, 429.
7. C. M. Julien and A. Mauger, Ionics, 2013, 19, 951.
8. J. H. Kim, S. T. Myung, C. S. Yoon, S. G. Kang and Y. K. Sun, Chem. Mater., 2004, 16, 906.
9. A. Manthiram, K. Chemelewski and E. S. Lee, Energy Environ. Sci., 2014, 7, 1339.
10. J. H. Kim, S. T. Myung, C. S. Yoon, S. G. Kang and Y. K. Sun, Chem. Mater., 2004, 16, 906.
11. M. Kunduraci and G. G. Amatucci, Electrochim. Acta, 2008, 53, 4193.
12. J. C. Arrebola, A. Caballero, M. Cruz, L. Hernán, J. Morales and E. R. Castellón, Adv. Funct. Mater., 2006, 16, 1904.
13. X. Zhang, F. Cheng, K. Zhang, Y. Liang, S. Yang, J. Liang and J. Chen, RSC Adv., 2012, 2, 5669.
14. H. B. Lin, Y. M. Zhang, J. N. Hu, Y. T. Wang, L. D. Xing, M. Q. Xu and W. S. Li, J. Power Sources, 2014, 257, 37.
15. Y. Wang, H. Li, P. He, E. Hosono and H. Zhou, Nanoscale, 2010, 2, 1294.
16. N. Arun, A. Jain, V. Aravindan, S. Jayaraman, W. C. Ling, M. P. Srinivasanc and S. Madhavi, Nano Energy, 2015, 12, 69.
17. S. Lee, Y. Cho, H. K. Song, K. T. Lee and J. Cho, Angew. Chem., Int. Ed., 2012, 51, 8748.
18. L. Zhou, D. Zhao and X. W. Lou, Angew. Chem., 2012, 124, 243–245.
19. W. W. Wu, H. F. Xiang, G. B. Zhong, W. Su, W. Tang, Y. Zhang and C. H. Chen, Electrochim. Acta, 2014, 119, 206.
20. C. Zhu and T. Akiyama, RSC Adv., 2014, 4, 10151.
21. J. B. Fei, Y. Cui, X. H. Yan, W. Qi, Y. Yang, K. W. Wang and J. B. Li, Adv. Mater., 2008, 20, 452.
22. E. S. Lee, K. W. Nam and E. Hu, Chem. Mater., 2012, 24, 3610.
23. K. R. Chemelewski and A. Manthiram, J. Phys. Chem. C, 2013, 117, 12465.
24. D. W. Shin, C. A. Bridges, A. Huq and M. P. Paranthaman, Chem. Mater., 2012, 24, 3720.
25. J. Song, D. W. Shin, Y. Lu, C. D. Amos, A. Manthiram and J. B. Goodenough, Chem. Mater., 2012, 24, 3101.
26. G. Liu, K. S. Park, J. Song and J. B. Goodenough, J. Power Sources, 2013, 243, 260.
27. L. Wang, F. Tang, K. Ozawa, Z. G. Chen, A. Mukherj, Y. Zhu and G. Q. Lu, Angew. Chem., 2009, 121, 7182.
28. Y. Wang, X. Shao, H. Xu, M. Xie, S. Deng, H. Wang and H. Yan, J. Power Sources, 2013, 226, 140.
29. J. Xiao, X. Chen, P. V. Sushko, M. L. Sushko, L. Kovarik, J. Feng and M. S. Whittingham, Adv. Mater., 2012, 24, 2109.
30. Z. Moorhead-Rosenberg, D. W. Shin, K. R. Chemelewski, J. B. Goodenough and A. Manthiram, Appl. Phys. Lett., 2012, 100, 213909.
31. Y. Xue, Z. Wang, L. Zheng, F. Yu, B. Liu, Y. Zhang and K. Ke, Sci. Rep., 2015, 5, 13299.
32. M. Kunduraci and G. G. Amatucci, J. Electrochem. Soc., 2006, 153, A1345.
33. M. Kunduraci and G. G. Amatucci, Electrochim. Acta, 2008, 53, 4193.
34. H. Luo, P. Nie, L. Shen, H. Li, H. Deng, Y. Zhu and X. Zhang, ChemElectroChem, 2015, 2, 127–133.
35. Y. Xue, Z. Wang, F. Yu, Y. Zhang and G. Yin, J. Mater. Chem. A, 2014, 2, 4185.
36. B. Hai, A. K. Shukla, H. Duncan and G. Chen, J. Mater. Chem. A, 2013, 1, 759.
37. L. J. Xi, H. E. Wang, Z. G. Lu, S. L. Yang, R. G. Ma, J. Q. Deng and C. Y. Chung, J. Power Sources, 2012, 198, 251.
38. H. Fang, Z. Wang, B. Zhang, X. Li and G. Li, Electrochem. Commun., 2007, 9, 1077.

---

Footnote

† Electronic supplementary information (ESI) available: SEM micrographs and cycling performance. See DOI: 10.1039/c5ra17933e

---