Synthesis of MoO₂ hierarchical peony-like microspheres without a template and their application in lithium ion batteries

Abstract

Three dimensional hierarchical structures constructed with low dimensional nanoscale building blocks have become a new research focus for lithium storage. Some hierarchical structured compounds, such as V₂O₅, Co₃O₄ and MoS₂ micro/nanoflowers, have been synthesized for energy storage applications. However, it still remains a challenge to fabricate well-designed hierarchical MoO₂ microflowers. This paper reports on a morphology-controlled synthesis of hierarchical peony-like MoO₂ microspheres as an anode material for lithium-ion batteries. This hierarchical structure is fabricated via a simple template-free solvothermal method. The as-derived hierarchical MoO₂ microspheres exhibit high specific capacity, high capacity retention, and remarkable cycling stability, when evaluated as an anode material for LIBs.

---

A large amount of effort has been made to achieve high energy conversion efficiency and capacity in the research of lithium-ion batteries (LIBs).^1–3 However, the development of anodes is largely inhibited by inferior long-term cyclic performance and fast capacity fading.^3,4 Well-designed micro/nano-structured materials give promise of solving this problem due to their size effects, which could not be achieved in bulk materials.^3,4 In particular, three dimensional hierarchical structures constructed with low dimensional nanoscale building blocks are valuable for improving the cycling performance and capacity retention in LIBs.^5–8 This can be explained in that nano-sized building blocks actually play key roles in decreasing the lithium/electron diffusion path and increasing the interfacial contact area with the electrolyte. Additionally, micro-sized secondary architectures provide good resistance to volume change and prevent self-aggregation and structural degradation during the lithium insertion/extraction process; meanwhile, the hierarchical structure can accommodate high volume expansion during intercalation, resulting in a longer lifespan and better cyclic stability. Up to now, various hierarchical micro/nano-structured metal oxides/sulfides, such as iron oxides,^9,10 V₂O₅,^11–13 Co₃O₄,¹⁴ CuO,¹⁵ MoS₂,^16,17 and SnO₂,^18,19 are already developed and applied as high-performance electrode materials.

Monoclinic MoO₂ is a promising alternative anode material owing to its high theoretical specific capacity (838 mA h g⁻¹), low electrical resistivity (8.8 × 10⁻⁵ Ω cm), high thermal and chemical stabilities, affordable cost, and high density (6.5 g cm⁻³).^3,20–24 Unfortunately, bulk MoO₂ material is not an ideal lithium storage host due to its low capacity retention and poor cyclic stability.^3,22,23 These problems mainly originate from the intrinsic slow kinetics and huge volumetric variations during lithium insertion/extraction processes, which creates high internal stress, leading to pulverization of the electrode material. Much effort has been devoted to the fabrication of hierarchical structured MoO₂ materials. Huang et al. have prepared hierarchical MoO₂ grown on cloth fibers and porous MoO₂ nanorod composites.²⁵ Mai et al. have obtained hierarchical nanostructured MoO₂/Co(OH)₂.²⁶

In this work, a morphology-controlled synthesis procedure has been designed to produce hierarchical peony-like MoO₂ microspheres. This hierarchical structure is fabricated via a simple template-free solvothermal method. The surface morphologies of the microflowers can be tailored by varying the reductants and the solvothermal reaction time. The as-derived hierarchical MoO₂ microflowers exhibit high specific capacity and remarkable cycling stability evaluated as an anode material for LIBs.

In a general synthesis process, commercial MoO₃ powders (1.008 g, 99.99%, Sigma Aldrich) and oxalic acid (2.646 g) were added into 40 mL de-ionized water under stirring for several hours at 80 °C. Then, 3 mL of the above solution and 30 mL isopropanol were mixed under stirring for 10 min at room temperature. Afterwards, the mixture was transferred into a 50 mL Teflon-lined stainless steel autoclave. The autoclave was kept in an oven at 200 °C for 12 h, and then cooled naturally and centrifuged with distilled water to remove impurity ions. After drying in a vacuum at 60 °C for 10 h, the black MoO₂ powders were fabricated without any further heat treatment.

The phase composition and morphology of the as-prepared products were characterized by X-ray powder diffraction (XRD, Rigaku D/max2500 XRD with Cu Kα radiation, λ = 1.54178 Å), and scanning electron microscopy (SEM, Quanta FEG 250, 10 kV).

The electrochemical tests were carried out using CR2016 two-electrode coin cells. In brief, the working electrodes were prepared by mixing the as-prepared MoO₂ powders, acetylene black, and polyvinylidene fluoride (PVDF) at a weight ratio of 8 : 1 : 1. The mixture were dispersed in N-methyl-2-pyrrolidone (NMP) to form a slurry. The slurry was pressed on a Cu foil and dried in a vacuum oven at 90 °C for 8 hours. The mass of the active materials is between 0.8 and 1.8 mg cm⁻². Li/MoO₂ coin cells were assembled in an argon-filled glove box (Mbraun, Germany) by stacking microporous Cellgard 2400 polypropylene membrane as the separator, lithium foil as the counter electrode, and 1 M LiPF₆ in ethylene carbonate (EC)/dimethyl carbonate (DMC) (at a volume ratio of 1 : 1) as the electrolyte. Cyclic voltammetry (CV) was tested on an electrochemical workstation (CHI604E, China) at a scan rate of 0.1 mV s⁻¹ in the voltage range 0.01–3.0 V (vs. Li/Li⁺). The galvanostatic charge/discharge measurements were conducted at various current densities in the voltage range 0.01–3.0 V (vs. Li/Li⁺) on a Land Battery Measurement System (Land CT 2001A, Wuhan, China). All the tests were performed at ambient temperature. The electrochemical impedance spectra measurement is performed with fresh cells by IM6ex electrochemical working station over the frequency range from 100 kHz to 0.1 kHz.

The SEM images of the as-generated products synthesized with different heat treatments are shown in Fig. 1. It can be concluded that the solvothermal times have a great influence on the final morphologies of the products. With the reactant heated for 4 h, the peony-like structure is still gestating. When the time is extended to 8 h, the flower with several thick petals can be observed clearly. As for the sample heated for 12 h, the peony-like structure is completely grown with thin petals; what make it more interesting is that the pistil is visible, thus making the flower more vivid. The petals of the flower will be pulverized by increasing the reaction time to 24 h, and will eventually be broken into small particles (Fig. 1e).

Fig. 1 SEM images of the as-prepared products synthesized for different times. (a and f) 4 h; (b and g) 8 h; (c and h) 12 h; (d and i) 24 h; (e and j) 48 h.

According to the self-assembly and traditional Ostwald ripening in some reported papers,^8,27–29 free-standing nanosheets are generated under a short reaction time. When the reaction time is prolonged, the free-standing nanosheets aggregate together into petal-building assemblies to minimize the total surface energy. Finally, obviously discrete and uniform flower-like structures constructed with nanosheet petal-building assemblies are observed. In general, most of the nanosheets remain unchanged in size and assemble into a petal-building flower-like structure, the chances of which can be compared to a thousand people lining up in a specific formation.

In our work, the formation of the hierarchical structure is different from the above. The peony-like MoO₂ microspheres develop as reaction time increases, which can be compared to a fertilized egg growing into a complete human being. It may be a consecutive diffusion growth instead of a self-assembly process. During the solvothermal stage, not only can the reaction time influence the morphology, but also the different reductants can change the eventual structure (see Fig. 2). As shown in Fig. 2a, when distilled water was used to replace the isopropanol while other parameters were kept constant, a unique polyhedron dominated in the full image. The symmetrical polyhedra are like carved stone. When ethylene glycol was substituted for isopropanol, the products exhibited uniform nanoparticles with sizes of around 100 nm (Fig. 2b). As the reductant was replaced by ethyl alcohol, the products possessed several kinds of particles with different sizes (Fig. 2c). Therefore, the selection of reductants is important in the fabrication of these hierarchical peony-like MoO₂ microspheres. The influence of solvents on the morphology of as-prepared products may be caused by different functional groups of solvents. They will change the surface of the microspheres. Different functional groups may lead to the distinctions of surface energy and surface charge. But the deep formation mechanism needs more investigation in the future. Hence, it was concluded that both the selection of reductants and the reaction time played critical roles in the fabrication of these hierarchical peony-like MoO₂ microspheres. The surface morphologies of the microflowers can be tailored by varying the solvothermal reaction time and the reductants.

Fig. 2 SEM images of the products prepared by the hydrothermal method at 200 °C at 48 h: (a) with water; (b) with ethylene glycol; (c) with ethyl alcohol; (d) with isopropanol.

As shown in Fig. 3, the as-generated sample consists of discrete and uniform peony-like microspheres with sizes of around 2 µm. The thickness of nanosheet petals in the flower is about 50 nm. The template-free formation of such a hierarchical architecture with several structural features, including the two-dimensional primary building blocks, is extraordinary. This kind of hierarchical architecture provides a good physical contact among the nanosheets, which is beneficial for ionic and electronic transference. Fig. S1a† shows the TEM observation of the hierarchical peony-like MoO₂ microspheres. The selected area electron diffraction (SAED) pattern (Fig. S1b†) shows a set of concentric diffractions, indicating the polycrystalline nature of the monoclinic MoO₂ phase. The diffraction rings could be indexed to the (011), (−211), (−212), and (−122) crystal planes of the monoclinic MoO₂ phase, respectively.

Fig. 3 SEM images of hierarchical peony-like MoO₂ microspheres by 12 h.

The XRD patterns of the as-generated products synthesized for different times are illustrated in Fig. 4. In Fig. 4a, the XRD pattern of the sample by 4 h only shows a little fluctuation; the intensity of the diffraction peaks is too weak to observe. When the time is extended to 8 h, some diffraction peaks of MoO₂ become obvious. As for the sample heated for 12 h (Fig. 4b), all the diffraction peaks can be readily indexed to the monoclinic MoO₂ phase (space group: P2₁/c (no. 14), JCPDS card no. 65-5787). Since the amount of as-prepared MoO₂ materials is very small, the intensity of all the diffraction peaks is weak. Meanwhile, the background is strong, caused by air scattering. The X-ray might irradiate the glass slide, and two faint broad peaks appearing at around 22° and 54° indicate the amorphous character of the glass slide. The upper part of the MoO₂ diffraction peak is sharp and gradually becomes broader, which reveals that the as-prepared MoO₂ has nanocrystalline characteristics.³⁰ If the as-prepared MoO₂ is poorly crystalline, it will just show broad diffraction peaks instead of a sharp and narrow crest. A similar phenomenon is found in Yang et al.’s report.⁷ The as-prepared MoO₂ microspheres had an average grain size of about 15 nm according to the Debye–Scherrer equation, based on the full width at half maximum (FWHM). As for the sample heated for 24 h and 48 h, the diffraction peaks can be readily indexed to the monoclinic MoO₂ phase, and the products are of high purity and high crystallinity. The intensity of diffraction peaks becomes stronger over time, indicating the formation of MoO₂ microspheres.

Fig. 4 (a) XRD patterns of MoO₂ microspheres obtained by a template-free hydrothermal method for different times. (b) XRD patterns of hierarchical peony-like MoO₂ microspheres obtained by 12 h.

Fig. 5 shows the CV profiles of hierarchical MoO₂ microspheres at a scan rate of 0.1 mV s⁻¹ in the voltage range 0.01–3.0 V. The profiles show three reversible cathodic peaks at 1.35 V, 0.5 V and 0.25 V, which are ascribed to the multi-step electrochemical reaction process. The peak at 1.35 V corresponds to the phase transition of Li_xMoO₂ during the lithium insertion process.^3,30,31 The cathodic voltage plateau decreases to 0.5 V, which can be associated with a complete conversion to metallic Mo and Li₂O.^32–35 Two anodic peaks at 1.75 V and 1.35 V are observed, which can be attributed to the deintercalation of the lithium ions. On one hand, there is only one cathodic peak corresponding to the lithium intercalation process; on the other hand, there are two anodic peaks corresponding to the lithium deintercalation process. This phenomenon is consistent with Zhang et al. and Lou et al.’s reports,^30,32–35 whereas it is different from Yoon et al.³⁶ and Madhavi et al.’s reports.³¹ This discrepancy may result from the lithiation behavior of the materials. The lithiation behavior of the MoO₂ electrode involves both the intercalation-type reaction and the conversion-type reaction.^31,37,38 The conversion-type reaction is a continuous process, and it is relative to the materials’ microstructure, surface properties and crystallinity, which may affect the intercalation of the lithium ions and lead to different cathodic peaks. Thus, during the discharge/charge process, the reactions of the MoO₂ electrode could be described as follows: MoO₂ ↔ Li_xMoO₂ ↔ Mo + Li₂O.^23,33 Additionally, the positions of redox peaks do not change in the subsequent cycles, indicating the high reversibility of phase transitions. The high reversibility observed is mainly attributed to the hierarchical microstructure of the electrode.

Fig. 5 Cyclic voltammograms of hierarchical MoO₂ nanoflowers for selected cycles at a scan rate of 0.1 mV s⁻¹.

As illustrated in Fig. 6a, the MoO₂ electrode shows the cycling performance at a current density of 45 mA g⁻¹ in the voltage range 0.01–3.0 V. Interestingly, the initial capacity of the MoO₂ electrode is low and fluctuant, taking five lithiation/delithiation cycles before being stabilized. This activated process for nanostructured MoO₂ materials has also been reported by other groups.³² This might originate from the fact that the material partially loses its crystallinity and transforms into an amorphous-like structure during the cycling, thus improving the Li diffusion kinetics so more Li can be reversibly intercalated/deintercalated. The MoO₂ electrode exhibits a relatively low capacity in the first cycle, and a maximum specific discharge capacity of 1003.7 mA h g⁻¹ can be reached at the 8th cycle. Afterwards, it gradually decreases to 925.5 mA h g⁻¹ after the 30^th cycle, with a capacity retention of 92.2% based on the maximum discharge capacity. The coulombic efficiency of the MoO₂ electrode is above 97%. In addition, it is noteworthy that the reversible capacity is higher than the theoretical specific capacity of MoO₂. The extra capacity can be ascribed to the interfacial and structural defects of the microspheres.^32,39,40 These defects might lead to better charge transfer kinetics and Li ion diffusion kinetics, and allow the insertion of extra Li ions to improve the specific capacity. Also, Mo can also accommodate extra Li ions.^40–43 When increasing the current density to 200 mA g⁻¹ (Fig. 6b), the MoO₂ electrode presents a specific capacity of 757 mA h g⁻¹ after 100 cycles, corresponding to a coulombic efficiency of 99%.

Fig. 6 (a) Cycling performance of the MoO₂ electrode at a current density of 45 mA g⁻¹ in the range 0.01–3.0 V vs. Li/Li⁺. (b) Cycling performance of the MoO₂ electrode at a current density of 200 mA g⁻¹ in the range 0.01–3.0 V vs. Li/Li⁺. (c) Rate capability of the MoO₂ electrode at the current densities of 100 mA g⁻¹ to 200 mA g⁻¹, 400 mA g⁻¹, 800 mA g⁻¹, 1200 mA g⁻¹ and 1600 mA g⁻¹, respectively.

Fig. 6c exhibits the rate capability of the MoO₂ electrode tested in the voltage window 0.01–3.0 V at different current densities. The specific discharge capacities show 963.4 mA h g⁻¹ at 100 mA g⁻¹, 730.8 mA h g⁻¹ at 200 mA g⁻¹, 586.6 mA h g⁻¹ at 400 mA g⁻¹, 495.5 mA h g⁻¹ at 800 mA g⁻¹, 429.9 mA h g⁻¹ at 1200 mA g⁻¹ and 398.9 mA h g⁻¹ at 1600 mA g⁻¹, respectively. As the cell was reset to work at the current density of 100 mA g⁻¹ after the rate performance measurements, the cell still maintained a high capacity of 707.7 mA h g⁻¹. The unique hierarchical structure may have an influence on the lithium insertion process, resulting in better electrochemical performance of the MoO₂ electrode, which can lead to several beneficial features: good physical contact between the active materials and the electrolyte, short lithium/electron diffusion paths, enhanced charge transfer kinetics and Li ion diffusion kinetics, and extra Li ion sites. It is an effective way to improve electrochemical properties by using well-designed hierarchical nanoscale structures.

To deeply investigate the enhanced electrochemical properties, electrochemical impedance spectra measurement is performed with fresh cells by IM6ex electrochemical working station in Fig. S2.† The charge transfer resistance (R_ct) of the MoO₂ electrode is about 200 Ω. This means better electrical conductivity, which can be attributed to short diffusion path in the hierarchical microstructures.

In summary, hierarchical MoO₂ peony-like microspheres with enhanced electrochemical performance have been successfully fabricated using a template-free hydrothermal method. As an anode material for LIBs, the unique hierarchical MoO₂ microspheres exhibit a high reversible specific capacity of 925.5 mA h g⁻¹ after 30 cycles (757 mA h g⁻¹ after 100 cycles for 200 mA g⁻¹), good cycling performance, high capacity retention and better rate performance. This enhancement in the electrochemical performance of the MoO₂ electrode can be attributed to the unique hierarchical architecture, which can effectively amplify the active mass–electrolyte contact area, minimize the lithium ion/electron transport distance, increase the ionic/electronic reactivity and provide more active inserted sites for lithium storage. The superior electrochemical performance suggests their promising applications as an anode material for rechargeable LIBs.

Acknowledgements

This work is financially supported by the National Natural Science Foundation of China (grant no. 51202297 and 51472271), Program for New Century Excellent Talents in University (NCET-12-0554), the National Basic Research Program of China (973 Program) grant no. 2013CB932901, and the Fundamental Research Funds for the Central Universities of Central South University (no. 2014zzts165 and no. 2014zzts168).

Notes and references

1. J. M. Tarascon and M. Armand, Nature, 2001, 414, 359.
2. S. Q. Liang, T. Chen, A. Q. Pan, D. W. Liu, Q. Y. Zhu and G. Z. Cao, ACS Appl. Mater. Interfaces, 2013, 5, 11913.
3. J. Liu, S. S. Tang, Y. K. Lu, G. M. Cai, S. L. Liang, W. J. Wang and X. L. Chen, Energy Environ. Sci., 2013, 6, 2691.
4. J. S. Chen, Y. L. Cheah, S. Madhavi and X. W. Lou, J. Phys. Chem. C, 2010, 114, 8675.
5. Q. Q. Xiong, J. P. Tu, S. J. Shi, X. Y. Liu, X. L. Wang and C. D. Gu, J. Power Sources, 2014, 256, 153.
6. N. Zhou, E. Uchaker, H. Y. Wang, M. Zhang, S. Q. Liu, Y. N. Liu, X. W. Wu, G. Z. Cao and H. Y. Li, RSC Adv., 2013, 3, 19366.
7. D. H. Teng, Y. H. Yu and X. P. Yang, RSC Adv., 2014, 4, 12309.
8. J. J. Zhang, Y. L. Chen, Y. F. Sun, T. Huang and A. S. Yu, RSC Adv., 2013, 3, 20639.
9. T. Yang, Y. G. Liu, Z. H. Huang, Q. Yang, Y. B. Chen, M. L. Hu, M. Guan and M. H. Fang, RSC Adv., 2014, 4, 41578.
10. S. L. Jin, H. G. Deng, D. H. Long, X. J. Liu, L. Zhan, X. Y. Liang, W. M. Qiao and L. C. Ling, J. Power Sources, 2011, 196, 3887.
11. A. Q. Pan, H. B. Wu, L. Zhang and X. W. Lou, Energy Environ. Sci., 2013, 6, 1476.
12. H. Yu, X. H. Rui, H. T. Tan, J. Chen, X. Huang, C. Xu, W. L. Liu, D. Y. W. Yu, H. H. Hng, H. E. Hoster and Q. Y. Yan, Nanoscale, 2013, 5, 4937.
13. L. Chen, X. Gu, X. L. Jiang, N. N. Wang, J. Yue, H. Y. Xu, J. Yang and Y. T. Qian, J. Power Sources, 2014, 272, 991.
14. L. Li, K. H. Seng, Z. X. Chen, Z. P. Guo and H. K. Liu, Nanoscale, 2013, 5, 1922.
15. C. Wang, D. Higgins, F. F. Wang, D. Y. Li, R. Q. Liu, G. F. Xia, N. Li, Q. Li, H. Xu and G. Wu, Nano Energy, 2014, 9, 334.
16. T. Yang, Y. J. Chen, B. H. Qu, L. Mei, D. N. Lei, H. N. Zhang, Q. H. Li and T. H. Wang, Electrochim. Acta, 2014, 115, 165.
17. S. Q. Liang, J. Zhou, J. Liu, A. Q. Pan, Y. Tang, T. Chen and G. Z. Fang, CrystEngComm, 2013, 15, 4998.
18. X. M. Yin, L. B. Chen, C. C. Li, Q. Y. Hao, S. Liu, Q. H. Li, E. D. Zhang and T. H. Wang, Electrochim. Acta, 2011, 56, 2358.
19. X. M. Yin, C. C. Li, M. Zhang, Q. Y. Hao, S. Liu, L. B. Chen and T. H. Wang, J. Phys. Chem. C, 2010, 114, 8084.
20. Q. Yang, Q. Liang, J. Liu, S. Q. Liang, S. S. Tang, P. J. Lu and Y. K. Lu, Mater. Lett., 2014, 127, 32.
21. L. X. Zeng, C. Zheng, C. L. Deng, X. K. Ding and M. D. Wei, ACS Appl. Mater. Interfaces, 2013, 5, 2182.
22. Y. M. Sun, X. L. Hu, J. C. Yu, Q. Li, W. Lu, L. X. Yuan, W. X. Zhang and Y. H. Huang, Energy Environ. Sci., 2011, 4, 2870.
23. B. K. Guo, X. P. Fang, B. Li, Y. F. Shi, C. Y. Ouyang, Y. S. Hu, Z. X. Wang, G. D. Stucky and L. Q. Chen, Chem. Mater., 2012, 24, 457.
24. Z. X. Huang, Y. Wang, Y. G. Zhu, Y. M. Shi, J. I. Wong and H. Y. Yang, Nanoscale, 2014, 6, 9839.
25. Y. M. Sun, X. L. Hu, W. Luo and Y. H. Huang, ACS Nano, 2011, 5, 7100.
26. K. M. Hercule, Q. L. Wei, A. M. Khan, Y. L. Zhao, X. C. Tian and L. Q. Mai, Nano Lett., 2013, 13, 5685.
27. Z. Liu, X. D. Wen, X. L. Wu, Y. J. Gao, H. T. Chen, J. Zhu and P. K. Chu, J. Am. Chem. Soc., 2009, 131, 9405.
28. B. Liu and H. C. Zeng, J. Am. Chem. Soc., 2004, 126, 8124.
29. X. F. Zhang, M. Z. Wu, S. Gao, Y. M. Xu, X. L. Cheng, H. Zhao and L. H. Huo, Mater. Res. Bull., 2014, 60, 659.
30. L. Zhou, H. B. Wu, Z. Y. Wang and X. W. Lou, ACS Appl. Mater. Interfaces, 2011, 3, 4853.
31. Z. Y. Wang, J. S. Chen, T. Zhu, S. Madhavi and X. W. Lou, Chem. Commun., 2010, 46, 6906.
32. Y. L. Liu, H. Zhang, P. Ouyang, W. H. Chen, Y. Wang and Z. C. Li, J. Mater. Chem. A, 2014, 2, 4714.
33. Y. L. Liu, H. Zhang, P. Ouyang and Z. C. Li, Electrochim. Acta, 2013, 102, 429.
34. W. H. Chen, H. Zhang, Y. Wang, Z. Y. Ma and Z. C. Li, Electrochim. Acta, 2014, 144, 369.
35. Y. L. Liu, H. Zhang, P. Ouyang, W. H. Chen and Z. C. Li, Mater. Res. Bull., 2014, 50, 95.
36. K. Palanisamy, Y. Kim, H. Kim, J. M. Kim and W. S. Yoon, J. Power Sources, 2015, 275, 351.
37. S. Yoon and A. Manthiram, J. Mater. Chem., 2011, 21, 4082.
38. J. H. Ku, Y. S. Jung, K. T. Lee, C. H. Kim and S. M. Oh, J. Electrochem. Soc., 2009, 156, A688.
39. X. F. Zhang, X. X. Song, S. Gao, Y. M. Xu, X. L. Cheng, H. Zhao and L. H. Huo, J. Mater. Chem. A, 2013, 1, 6858.
40. J. H. Ku, J. H. Ryu, S. H. Kim, O. H. Han and S. M. Oh, Adv. Funct. Mater., 2012, 22, 3658.
41. L. C. Yang, L. L. Liu, Y. S. Zhu, X. J. Wang and Y. P. Wu, J. Mater. Chem., 2012, 22, 13148.
42. Q. Wang and J. Li, J. Phys. Chem. C, 2007, 111, 1675.
43. J. Xiao, X. Wang, X. Q. Yang, S. Xun, G. Liu, P. K. Koech, J. Liu and J. P. Lemmon, Adv. Funct. Mater., 2011, 21, 2840.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra06062a

---