Facile synthesis of iron-based compounds as high performance anode materials for Li-ion batteries

Abstract

Iron oxide@C composites and lithium ferrites were synthesized by a cotton-template method as anode materials for Li-ion batteries. α-Fe₂O₃@C composites with 3-D porous hollow secondary structures were prepared by directly burning the cotton containing the iron salt (FeCl₃ or Fe(NO₃)₃) in air, and the Fe₃O₄@C composites with similar structures were obtained by annealing α-Fe₂O₃@C under a N₂ atmosphere. α-LiFeO₂ and α-LiFe₅O₈ particles with sizes of 100–500 nm were also prepared using a similar method. Electrochemical measurements showed that all these samples demonstrated good electrochemical performances as Li-ion battery anodes, especially α-Fe₂O₃@C derived from Fe(NO₃)₃, which delivers a high reversible capacity of 990 mA h g⁻¹ at 100 mA g⁻¹ after 50 cycles. Both the porous hollow secondary structure and the suitable amount of amorphous carbon are significant for the electrochemical performances of the iron oxide@C composites. Such a method is simple, rapid and inexpensive and may facilitate the preparation of other high performance electrode materials with porous hollow structures.

---

1. Introduction

Li-ion batteries (LIBs) have attracted intense attention because advanced energy storage devices with higher energy and power density and longer lifetime are required for many applications, such as portable electronics, electric tools, electric vehicles and hybrid electric vehicles.^1–3 In the case of a battery, the electrode material plays a major role in determining the battery performance. The current LIBs predominantly use graphite as anode materials, however, it cannot satisfy the ever-growing demands of high energy density batteries due to the limited specific capacity (∼372 mA h g⁻¹). Metal oxides that undergo conversion- or displacement-type reactions (MO_x + xLi⁺ + xe⁻ → M + xLi₂O) are promising candidates for the next generation anode materials because of their high theoretical capacity (∼500–1000 mA h g⁻¹).^4–6 Iron oxide is one of the most intensively studied metal oxides for LIB anodes because of its high natural abundance, environmental benignity, low cost, and high theoretical capacity (928 mA h g⁻¹ for Fe₃O₄ and 1007 mA h g⁻¹ for α-Fe₂O₃).^7,8 However, like other transition metal oxides, when iron oxides used as anode materials for LIBs, they suffer from large volume expansion/contraction during the Li ion insertion and extraction process, leading to fast capacity fading and poor cycle stability.⁹

A strategy to solve this problem is the exploitation of hollow nanostructured materials. On one hand, hollow nanomaterials can enlarge the contact area between the material and electrolyte, and shorten the diffusion path of Li ions. On the other hand, the void space in hollow materials accommodates the structural strain and volume change during the Li ion insertion and extraction process.^10–12 There are two distinct strategies that are widely used to the preparation of iron oxide hollow nanostructures, namely template-free and template-based methods. For example, unique hollow Fe₃O₄ spheres were prepared by a template-free solvothermal method, which delivered a high reversible capacity of 870 mA h g⁻¹ at 100 mA g⁻¹ after 50 cycles.¹³ α-Fe₂O₃ hollow spheres with sheet-like subunits were synthesized through solvothermal method via a glycerol–water quasiemulsion-templating mechanism, and it delivered a reversible capacity of 710 mA h g⁻¹ at 200 mA g⁻¹ after 100 cycles.¹⁴ Kang et al. used microporous organic nanotubes as templates, which can be removed by burning in air, to prepare Fe₂O₃ nanotubes, which delivered a reversible capacity of 929 mA h g⁻¹ after 30 cycles.¹⁵ Another strategy to decrease the volume expansion and capacity fading of iron oxides is the fabrication of metal-oxide/carbon composites.^16–19 For example, nano-porous and mosaic structured Fe₃O₄@C spheres were prepared via a one-pot solvothermal route, which showed high specific capacity (∼1000 mA h g⁻¹) and excellent rate capability. The porous active carbon plays an important role in the improvement of electrochemical properties of Fe₃O₄, which not only acts as a host for the deposition of Fe₃O₄ particles, but also provides void spaces for active Fe₃O₄ to buffer the volume expansion.²⁰

Lithium ferrite is another important iron-based electrode material for LIBs, especially LiFeO₂ and LiFe₅O₈, which are usually used as cathode materials for LIBs.²¹ However, many problems still remain when lithium ferrite is used as cathode materials, for instance, low operating voltage, poor electrochemical activity and low capacity retention.²² Lithium ferrite also can be used as anode materials, however, there are few reports on lithium ferrite used as anode materials and almost all the reported results showed poor electrochemical performances.^23–25 Therefore, it is a challenge to develop a simple synthetic strategy for high performance lithium ferrite anode materials for LIBs.

Herein, we reported a simple cotton-templated method to synthesize iron oxide@C composites with 3-D porous hollow secondary structures and lithium ferrites with particle size in the range of 100–500 nm. Amorphous carbon coexists with the as-prepared iron oxides due to the incomplete burning of the cotton. All the iron-based compounds demonstrated good electrochemical performances when used as LIB anode materials. This method is simple, fast and inexpensive that may be applicable to the preparation of other porous hollow battery materials with excellent electrochemical performances.

2. Experimental

2.1 Materials preparation

Synthesis of iron oxides. All chemicals are of analytical grade (Kermel Co. Ltd Tianjing, China) and used without further purification. To synthesize α-Fe₂O₃@C, 2 g cotton was soaked in 10 mL 0.25 M FeCl₃ solution, and then dried at 80 °C for 12 h in an oven. After that, the cotton was burnt in air, and the as-obtained sample was named as α-Fe₂O₃–Cl. To obtain Fe₃O₄@C, α-Fe₂O₃–Cl was annealed in N₂ at 600 °C for 4 h, and the as-obtained sample was named as Fe₃O₄–Cl. When FeCl₃ solution was replaced by Fe(NO₃)₃ solution, and the other steps were the same as those above, the as-obtained samples were named as α-Fe₂O₃–N and Fe₃O₄–N, respectively.

Synthesis of lithium ferrites. To synthesize α-LiFe₅O₈, 2.5 mmol FeCl₂·4H₂O and 2.5 mmol ethylenediamine tetraacetate (EDTA) were dissolved in 10 mL deionized water, to which 2.5 mmol LiOH·H₂O was added subsequently, and then the solution was stirred at room temperature until it became clear. Next, 2 g cotton was soaked in the above solution, dried at 80 °C for 12 h in an oven, and then burnt in air. The as-obtained ash was pressed under 10 MPa, and then annealed in air at 750 °C for 12 h to obtain the final product. When synthesizing α-LiFeO₂, 2.5 mmol FeCl₂·4H₂O was dissolved in 10 mL deionized water, and 2 g cotton was soaked in the above solution. Afterwards, the cotton was dried at 80 °C for 12 h in an oven, followed by burning in air. The as-obtained ash was ground with 2.5 mmol LiOH·H₂O, then the mixture was pressed under 10 MPa, and finally annealed in air at 750 °C for 12 h.

2.2 Materials characterization

Powder X-ray diffraction (XRD) patterns were recorded on a RigakuSmartLab(9) diffractometer using CuKα X-rays at a scanning rate of 8° min⁻¹ between 10° and 80°. Thermal gravimetric analysis (TGA) was performed on a SDT Q600 (TA Instruments, U.S.A.) in air with a heating rate of 10 °C min⁻¹. Field emission scanning electron microscopy (FESEM) images and energy dispersive X-ray (EDX) spectrum were obtained using a FEI Nova Nano SEM 450 instrument at 10 kV.

2.3 Electrochemical measurements

The electrochemical performances of the iron-based anode materials were evaluated with standard CR2025 coin cells. The electroactive materials, acetylene black and polyvinylidene fluoride (PVDF) were mixed in a mass ratio of 60 : 20 : 20 and dispersed in N-methylpyrrolidone (NMP) to form a homogeneous slurry. Then the mixture was pressed onto a copper foil substrate and dried at 80 °C for 12 h. The resulting foil was cut into disc with a diameter of 12 mm, which was then pressed with a pressure of 10 MPa to enhance the contact between active materials and copper foil. Test cells were assembled in an argon-filled glovebox with metallic lithium foil as the counter electrode, 1 M LiPF₆ in ethylene carbonate–dimethyl carbonate–diethyl carbonate (EC–DMC–DEC, 1 : 1 : 1 vol%) as the electrolyte, and a polypropylene (PP) microporous film (Cellgard 2300) as the separator. The galvanostatic charge–discharge tests were conducted on a LAND battery program-control test system at different rates between 0.01 and 3.0 V.

3. Results and discussion

For the preparation of iron oxides, cotton was used as template, which can be removed by burning in air, and two different iron sources (FeCl₃ and Fe(NO₃)₃) were used as the reactants. It is well known that halogen compounds can be used as flame retardands.^26,27 When FeCl₃ was used as the reactant, the as-prepared iron oxides (α-Fe₂O₃–Cl and Fe₃O₄–Cl) may contain large amount of carbon due to the flame retardancy effect of Cl.²⁷ However, when Fe(NO₃)₃ was used as the reactant, the thermal decomposition of Fe(NO₃)₃ can be easily achieved, and O₂ generates during the process of decomposition,²⁸ so the cotton will combust much more sufficient, and there will be much smaller amount of carbon residue in the as-prepared iron oxides (α-Fe₂O₃–N and Fe₃O₄–N) than α-Fe₂O₃–Cl and Fe₃O₄–Cl. As shown in Fig. 1a and b, α-Fe₂O₃–Cl has 3-D porous hollow secondary structure with carbon in the inner and iron oxide nanoparticles coated on the surface of carbon. The hollow carbon inner wall was formed by the combustion of the cotton, and the porous iron oxide coating layer was formed by the decomposition of FeCl₃. The EDX spectrum of α-Fe₂O₃–Cl shown in Fig. 1c confirms the existences of Fe, O, C, Cl elements in α-Fe₂O₃–Cl, indicating that α-Fe₂O₃–Cl is composed of iron oxide, carbon and chloride. FESEM images of α-Fe₂O₃–N are shown in Fig. 1d and e. As can be seen, α-Fe₂O₃–N also has 3-D porous hollow secondary structure, and the porous wall of α-Fe₂O₃–N is mainly consisted by iron oxide nanoparticles without compact carbon wall in the inner side, which is quite different from α-Fe₂O₃–Cl. Fig. 1f shows the EDX spectrum of α-Fe₂O₃–N, which confirms that there are not only iron oxide but also carbon in α-Fe₂O₃–N, and the weak C peak indicates the small amount of carbon. The FESEM image of Fe₃O₄–Cl is similar to α-Fe₂O₃–Cl (Fig. 1g), while there are some differences between Fe₃O₄–N and α-Fe₂O₃–N, for example, the particle agglomeration can be clearly seen for Fe₃O₄–N, and some large octahedral particles can be also found (Fig. 1h).

Fig. 1 FESEM images (a and b) and EDX spectrum (c) of α-Fe₂O₃–Cl. FESEM images (d and e) and EDX spectrum (f) of α-Fe₂O₃–N. FESEM images of Fe₃O₄–Cl (g) and Fe₃O₄–N (h). (i) XRD patterns of iron oxide@C composites. The standard patterns of Fe₂O₃ (JCPDS 1-1053) and Fe₃O₄ (JCPDS 1-1111) are shown for reference.

XRD patterns of the as-prepared iron oxide@C composites are shown in Fig. 1i. It can be found that α-Fe₂O₃–Cl contains two phases of iron oxides, i.e., α-Fe₂O₃ (JCPDS 1-1053) and Fe₃O₄ (JCPDS 1-1111), and Fe₃O₄ is the minor phase. The existence of Fe₃O₄ can be attributed to the reduction of α-Fe₂O₃ by carbon during the cotton combustion process, which is so short that the proportion of Fe₃O₄ in α-Fe₂O₃–Cl is limited. Additionally, there are no peaks corresponding to carbon for all those samples, indicating that the carbons in all samples are amorphous. The diffraction peaks of iron oxide in α-Fe₂O₃–N can be assigned to α-Fe₂O₃ (JCPDS 1-1053), and no Fe₃O₄ was detected. When annealing the samples α-Fe₂O₃–Cl and α-Fe₂O₃–N in N₂ at 600 °C for 4 h, the XRD patterns of the as-obtained samples (Fe₃O₄–Cl and Fe₃O₄–N) match very well with that of Fe₃O₄ (JCPDS 1-1111), and the reduction of α-Fe₂O₃ to Fe₃O₄ can be attributed to the reduction effect of carbon at high temperature in inert atmosphere.²⁹

Fig. 2 shows the morphologies and crystal structures of lithium ferrites prepared by cotton-templated method. From Fig. 2a, it can be seen that α-LiFe₅O₈ is composed of agglomerated particles with particle size ranging from 100 to 500 nm, and some particles have polyhedral shapes with keen edges. The morphology of α-LiFeO₂ is similar to that of α-LiFe₅O₈ (Fig. 2b). It is worth noting that when the cotton was soaked in the solution containing both the iron and lithium sources, followed by drying, combustion, press and calcination, the as-prepared sample was proved to be α-LiFe₅O₈. When iron oxide was firstly synthesized using the cotton-templated method as precursor, and then ground with LiOH·H₂O followed by press and calcination, the as-prepared sample was proved to be α-LiFeO₂ (Fig. 2c). Such transformation has not been reported previously, and the formation mechanism is worth further studying.

Fig. 2 FESEM images (a and b) and XRD patterns (c) of lithium ferrites. The standard patterns of α-LiFe₅O₈ (JCPDS 17-115) and α-LiFeO₂ (JCPDS 17-938) are shown for reference.

To evaluate the electrochemical performances of the samples as anode materials for LIBs, the charge–discharge cycling was carried out at a current density of 100 mA g⁻¹ in the voltage range of 0.01–3 V. The charge–discharge voltage profiles of iron oxides and lithium ferrites for the first, second, fifth, tenth and thirtieth cycles are shown in Fig. 3. The first discharge capacity of α-Fe₂O₃–Cl is 1210 mA h g⁻¹, however, the discharge capacity decreases to 700 mA h g⁻¹ in the second cycle, and it is only 400 mA h g⁻¹ after the 30th cycle (Fig. 3a). Moreover, there is no distinct voltage plateau of the voltage versus capacity profiles for α-Fe₂O₃–Cl. While for α-Fe₂O₃–N, a distinct voltage plateau at approximately 0.75 V can be clearly identified during the initial discharge process, and it shifts to about 1.0 V and remains stable in the subsequent cycles (Fig. 3b). This charge–discharge voltage profiles correspond to the reduction of Fe₂O₃ to form Fe⁰ and Li₂O,¹⁵ which are in good agreement with previous report.³⁰ The first discharge and charge capacities of α-Fe₂O₃–N are 1316 and 947 mA h g⁻¹, respectively. The initial coulombic efficiency is 72%, which is higher than the recent reports about Fe₂O₃ anode materials.^15,31 The irreversible capacity loss in the first cycle can be attributed to the formation of a solid electrolyte interphase (SEI) film and electrolyte decomposition.³² The coulombic efficiency increases rapidly to 96% at the second cycle, reaches 98% at the 5th cycle, and maintains the higher values throughout the following cycle tests, indicating that the formed SEI film retains intact and good reversibility during the electrochemical reactions.³¹ The charge–discharge voltage profiles of Fe₃O₄–Cl are shown in Fig. 3c, the voltage plateau is much more clear, and the discharge capacity fades slower after the second cycle compared with that of α-Fe₂O₃–Cl. The charge–discharge voltage profiles of Fe₃O₄–N (Fig. 3d) are similar to that of α-Fe₂O₃–N, except that the initial coulombic efficiency is 67%, which is lower than that of α-Fe₂O₃–N. Fig. 3e shows that the first discharge capacity of α-LiFe₅O₈ is 1212 mA h g⁻¹, after the 30th cycle, the discharge capacity of α-LiFe₅O₈ slowly decreases to 694 mA h g⁻¹. The voltage plateau of α-LiFe₅O₈ is about 0.78 V during the initial discharge process, which shifts to about 1.0 V and remains stable in the subsequent cycles. The charge–discharge voltage profiles of α-LiFeO₂ (Fig. 3f) is similar to that of α-LiFe₅O₈, except that the first discharge capacity of α-LiFeO₂ is 1007 mA h g⁻¹, which is much lower than that of α-LiFe₅O₈, and the voltage plateau of α-LiFeO₂ is about 0.68 V during the initial discharge process, which is also lower than that of α-LiFe₅O₈.

Fig. 3 Charge–discharge voltage profiles of iron oxide@C composites (a–d) and lithium ferrites (e and f) at a current density of 100 mA g⁻¹.

The cycling performances of the samples are depicted in Fig. 4, at a constant current density of 100 mA g⁻¹ between 0.01 and 3.0 V. Both α-Fe₂O₃–Cl and Fe₃O₄–Cl display a low specific capacity and fast performance decline, only 322 and 355 mA h g⁻¹ are remained after 50 cycles, respectively. In contrast, both α-Fe₂O₃–N and Fe₃O₄–N demonstrate a higher specific capacity and better cycling stability, and the specific capacity reaches as high as 990 and 775 mA h g⁻¹ after 50 cycles for α-Fe₂O₃–N and Fe₃O₄–N, respectively (Fig. 4a). From Fig. 4b, both α-LiFe₅O₈ and α-LiFeO₂ demonstrate good cycling stability, which show a high specific capacity of 1212 and 1007 mA h g⁻¹ initially and maintain about 62% and 76% of the initial capacity after 50 cycles, respectively. Interestingly, the specific capacities of both α-LiFe₅O₈ and α-LiFeO₂ initially decrease until 30 cycles and then start to increase until around the 50th cycle. This capacity increase can be attributed to the existence of some activation process as previously reported.³³

Fig. 4 Cycling behaviour of iron oxide@C composites (a) and lithium ferrites (b) at a current density of 100 mA g⁻¹.

Since α-Fe₂O₃–N has highest specific capacity and best cycling stability among all the as-prepared iron-based compounds, the rate performance of α-Fe₂O₃–N was evaluated, and the results are shown in Fig. S1.† The electrode was measured at different rates from 100 mA g⁻¹ to 5000 mA g⁻¹, followed by a return to 100 mA g⁻¹. α-Fe₂O₃–N delivers a discharge capacity of about 960 mA h g⁻¹ when first cycled at 100 mA g⁻¹. Even at a high current density of 3000 mA g⁻¹, the reversible capacity still retains about 503 mA h g⁻¹, while it decreases to 397 mA h g⁻¹ at 5000 mA g⁻¹, which is still higher than the theoretical capacity of graphite (372 mA h g⁻¹). Moreover, when the current density returns to the initial 100 mA g⁻¹ after cycling at high rates, the reversible capacity still keeps about 940 mA h g⁻¹, which nearly recovers its original capacity. This results is superior to previous reports,^31,33 and suggest that the structure of α-Fe₂O₃–N remains stable even under high rate cycling.³¹

It is well known that the introduction of carbon into the metal oxide matrix can relieve the stress induced by volume expansion/contraction, and also increase the conductivity of metal oxides, leading to improved capacity retention and rate capacity.^9,20 In order to reveal the reasons for the excellent electrochemical performance of α-Fe₂O₃–N, the content of carbon in α-Fe₂O₃–Cl, α-Fe₂O₃–N and α-Fe₂O₃–N-anneal (the sample obtained by rapidly annealing α-Fe₂O₃–N in air at 600 °C for 10 min) were quantitatively measured. Fig. S2† shows TG curves of these three samples. By heating those samples from room temperature to 700 °C, a weight loss of 78 wt% can be found for α-Fe₂O₃–Cl, 5 wt% for α-Fe₂O₃–N and 1 wt% for α-Fe₂O₃–N-anneal. The proportion of Fe₂O₃ in α-Fe₂O₃–Cl is only 22 wt%, much lower than that in α-Fe₂O₃–N (95%). Since the theoretical capacity of carbon is much lower than that of α-Fe₂O₃ (372 mA h g⁻¹ and 1007 mA h g⁻¹, respectively), the reversible capacity of α-Fe₂O₃–Cl is much lower than that of α-Fe₂O₃–N (Fig. 4a). Fig. S3† shows the morphologies and the electrochemical performance of α-Fe₂O₃–N-anneal. As can be seen, α-Fe₂O₃–N-anneal has similar morphology to α-Fe₂O₃–N, but the electrochemical performance of α-Fe₂O₃–N-anneal is not as good as α-Fe₂O₃–N. The specific capacity of α-Fe₂O₃–N-anneal decreases from initial 1305 to 450 mA h g⁻¹, only maintaining about 34% of the initial capacity after 50 cycles. The reason for the relatively poor electrochemical performance of α-Fe₂O₃–N-anneal is that it contains too small amount of carbon (∼1%). Therefore, containing appropriate amount of carbon is an important reason for the excellent electrochemical performance of α-Fe₂O₃–N.

4. Conclusions

Electrochemically active α-Fe₂O₃@C, Fe₃O₄@C, α-LiFeO₂ and α-LiFe₅O₈ anode materials were synthesized by a simple cotton-templated method. All the iron oxides have 3-D porous hollow secondary structures with certain amount of amorphous carbon coexistence. Both porous hollow structures and proper amount of amorphous carbon are the key factors for the iron oxide@C composites to achieve excellent electrochemical performance. Especially, α-Fe₂O₃ not only has high reversible capacity (990 mA h g⁻¹ at 100 mA g⁻¹ after 50 cycles), but also has excellent rate performance (503 mA h g⁻¹ at 3000 mA g⁻¹, and 397 mA h g⁻¹ at 5000 mA g⁻¹). The as-prepared lithium ferrites also have excellent electrochemical performance. This method is simple, rapid and inexpensive, which may be used for the preparation of other battery materials with outstanding electrochemical properties.

Acknowledgements

The financial support from the National Natural Science Foundation of China (Grant no. 51125009), the Fundamental Research Funds for the Central Universities (Grant no. DUT13LK17), the Scientific Research Fund of Liaoning Provincial Education Department (Grant no. L2013030), the Open Project of State Key Laboratory of Rare Earth Resources Utilization (Grant no. RERU2013013), and the Hundred Talents Program of Chinese Academy of Sciences is greatly acknowledged.

References

1. J. M. Tarascon and M. Armand, Nature, 2001, 414, 359.
2. B. Kang and G. Ceder, Nature, 2009, 458, 190.
3. J. B. Goodenough and K. S. Park, J. Am. Chem. Soc., 2013, 135, 1167.
4. P. Poizot, S. Laruelle, S. Grugeon, L. Dupont and J. M. Tarascon, Nature, 2000, 407, 496.
5. M. Biswal, A. Suryawanshi, V. Thakare, S. Jouen, B. Hannoyer, V. Aravindan, S. Madhavicd and S. Ogale, J. Mater. Chem. A, 2013, 1, 13932.
6. J. Liu, H. Xia, L. Lu and D. Xue, J. Am. Chem. Soc., 2009, 131, 12086.
7. J. Lin, A. O. Raji, K. Nan, Z. Peng, Z. Yan, E. L. G. Samuel, D. Natelson and J. M. Tour, Adv. Funct. Mater., 2014, 24, 2044.
8. L. Su, Y. Zhong and Z. Zhou, J. Mater. Chem. A, 2013, 1, 15158.
9. J. E. Lee, S. H. Yu, D. J. Lee, D. C. Lee, S. I. Han, Y. E. Sung and T. Hyeon, Energy Environ. Sci., 2012, 5, 9528.
10. T. H. Kim and H. K. Song, Electrochim. Acta, 2013, 105, 47.
11. L. Zhang, H. B. Wu and X. W. Lou, Adv. Energy Mater., 2014, 4, 1300958.
12. J. Liu and D. Xue, Nanoscale Res. Lett., 2010, 5, 1525.
13. J. J. Zhang, Y. Yao, T. Huang and A. S. Yu, Electrochim. Acta, 2012, 78, 502.
14. B. Wang, J. S. Chen, H. B. Wu, Z. Y. Wang and X. W. Lou, J. Am. Chem. Soc., 2011, 133, 17146.
15. N. Kang, J. H. Park, J. Choi, J. Jin, J. Chun, I. G. Jung, J. Jeong, J. G. Park, S. M. Lee, H. J. Kim and S. U. Son, Angew. Chem., Int. Ed., 2012, 51, 6626.
16. X. Gu, L. Chen, S. Liu, H. Y. Xu, J. Yang and Y. T Qian, J. Mater. Chem. A, 2014, 2, 3439.
17. W. Xiao, Z. X. Wang, H. J. Guo, Y. H. Zhang, Q. Zhang and L. Gan, J. Alloys Compd., 2013, 560, 208.
18. B. Jang, M. Park, O. B. Chae, S. Park, Y. Kim, S. M. Oh, Y. Piao and T. Hyeon, J. Am. Chem. Soc., 2012, 134, 15010.
19. Y. Chen, B. H. Song, M. Li, L. Lu and J. M. Xue, Adv. Funct. Mater., 2014, 24, 319.
20. Z. P. Zeng, H. L. Zhao, J. Wang, P. P. Lv, T. H. Zhang and Q. Xia, J. Power Sources, 2014, 248, 15.
21. K. Li, H. Chen, F. Shua, K. Chen and D. Xue, Electrochim. Acta, 2014, 136, 10.
22. J. G. Li, J. J. Li, J. Luo, L. Wang and X. M. He, Int. J. Electrochem. Sci., 2011, 6, 1550.
23. M. M. Rahmana, J. Z. Wang, M. F. Hassana, Z. X. Chen and H. K. Liu, J. Alloys Compd., 2011, 509, 5408.
24. I. Uzunov, S. Uzunova, D. Kovacheva, S. Vasilev and B. Puresheva, J. Mater. Sci., 2007, 42, 3353.
25. S. Uzunova, I. Uzunova, D. Kovacheva, A. Momchilov and B. Puresheva, J. Appl. Electrochem., 2006, 36, 1333.
26. L. Costa, M. P. Luda and L. Trossarelli, Polym. Degrad. Stab., 2000, 68, 67.
27. G. E. Zaikov and S. M. Lomakin, Polym. Degrad. Stab., 1996, 54, 223.
28. B. J. Chen, L. Xu, W. Y. Li and X. L. Gou, Adv. Mater., 2005, 17, 582.
29. E. Kang, Y. S. Jung, A. S. Cavanagh, G. H. Kim, S. M. George, A. C. Dillon, J. K. Kim and J. Lee, Adv. Funct. Mater., 2011, 21, 2430.
30. L. Zhang, H. B. Wu, S. Madhavi, H. H. Hng and X. W. Lou, J. Am. Chem. Soc., 2012, 134, 17388.
31. F. Han, D. Li, W. C. Li, C. Lei, Q. Sun and A. H. Lu, Adv. Funct. Mater., 2013, 23, 1692.
32. H. Wu, N. Du, J. Z. Wang, H. Zhang and D. Yang, J. Power Sources, 2014, 246, 198.
33. S. H. Lee, S. H. Yu, J. E. Lee, A. Jin, D. J. Lee, N. Lee, H. Jo, K. Shin, T. Y. Ahn, Y. W. Kim, H. Choe, Y. E. Sung and T. Hyeon, Nano Lett., 2013, 13, 4249.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra06889k

---