2-Aminoimidazole borohydride as a hydrogen carrier

Abstract

2-Aminoimidazole borohydride (Im-NH₂BH₄) is synthesized via the reaction between 2-aminoimidazole hemisulfate ((Im-NH₂)₂SO₄) and sodium borohydride by a simple ball milling method. This compound holds theoretical hydrogen capacity of 8.1 wt%. It is designed based on the strategy of destabilizating BH₄⁻ using a large conjugated cation, Im-NH₂⁺ and maximizing the protonic-hydridic hydrogen interaction by balancing their numbers. Thermal dehydrogenation analyses demonstrate that it can release about 3.2 wt% hydrogen (Na₂SO₄ included) without gaseous impurities below 320 °C. The decomposition process has three exothermal steps and the onset dehydrogenation temperature is 50.5 °C. These results indicate that the above strategy is effective. The dehydrogenation process of this compound is discussed and the product at 320 °C is proposed to be C₃N₃H₃BH.

---

Introduction

Hydrogen is regarded as one of the most promising sources of energy for the future. However, effective storage of hydrogen is a giant obstacle for its wide application.^1–5 In the last two decades, a large number of hydrogen storage materials, such as metal hydrides,^6–11 complex hydrides,^12–15 chemical hydrides,^16,17 carbon materials,^18–20 metal–organic frameworks,^21,22 liquid organic hydrogen carriers,^23,24 Kubas-type materials^25,26 and their combinations, have been researched. Unfortunately, none of them can fulfil all the targets of the US Department of Energy (DOE).

Metal borohydrides are promising hydrogen storage materials owing to their very high gravimetric hydrogen capacities. However, they suffer from high dehydrogenation temperature and release of borane.²⁷ Many strategies are used to solve these issues, such as using additives,^28–30 borohydrides with mixed cations^31,32 and metal borohydride organic amine complexes.^33–35 Some of the above attempts have encouraging results.

Recently, we reported that the dehydrogenation temperature of LiBH₄ and NaBH₄ can be dramatically reduced by dissolving them in a imidazolium-type ionic liquids.³⁶ This originates from the destabilization of BH₄⁻ due to the more favourable charge transfer effect between BH₄⁻ and 1-n-butyl-3-methylimidazolium (bmim⁺), a large conjugated cation. However, the poor solubility of the metal borohydrides in the ionic liquid results in very low total hydrogen capacity. Nevertheless, it provides a new strategy to lower the dehydrogenation temperature of borohydrides.

Another strategy to promote the dehydrogenation of borohydrides is to utilize the protonic-hydridic hydrogen interaction. The borohydride–amine composites are largely based on this strategy. Moreover, guanidinium (Gua⁺, C(NH₂)₃⁺) type of borohydrides, such as methylguanidinium borodride,³⁷ guanidinium borodride (GuaBH₄),^38,39 guanidinium octahydrotriborate (GuaB₃H₈),⁴⁰ have been studied as hydrogen storage materials. Their dehydrogenation temperatures are also low mainly due to the interaction between the protonic H (H⁺) on N and the hydridic H (H⁻) on B atoms. However, dehydrogenation of GuaBH₄ releases some NH₃, due to the excess of the protonic H. This can be solved by adding proper amount of Ca(BH₄)₂ or LiBH₄ to balance the number of H from N–H and B–H.^41,42

Inspired by the above strategies, here we report 2-aminoimidazole borohydride (Im-NH₂BH₄, as shown in Fig. 1) as a novel hydrogen carrier. It consists of BH₄⁻ and a large conjugated cation, similar to the borohydrides dissolved in imidazolium-type ionic liquids. The amino group on the imidazole ring is to balance the number of protonic and hydridic hydrogen to maximize the H⁺–H⁻ interaction. The theoretic hydrogen capacity of Im-NH₂BH₄ reaches 8.1 wt% and it can release up to 68% hydrogen below 320 °C, making it a promising hydrogen carrier.

Fig. 1 Schematic diagram of the structure of 2-aminoimidazole borohydride (yellow, B; grey, C; blue, N; white, H).

Experimental section

Materials

(Im-NH₂)₂SO₄ was purchased from Adamas Reagents Co. Ltd. and NaBH₄ was purchased from Sigma-Aldrich Co. LLC. They were stored in an argon-filled glove box and used as-received without further purification.

Synthesis of Im-NH₂BH₄

Im-NH₂BH₄ was synthesized by ball milling a mixture of (Im-NH₂)₂SO₄ and NaBH₄ in 1 bar argon atmosphere using a planetary ball milling apparatus (Pulverisette 5) for 40 min. The molar ratio of NaBH₄ and (Im-NH₂)₂SO₄ was 2 : 1 and the mass ratio of balls to raw materials was 40 : 1. The effect of ball milling conditions was studied.

Characterization

The samples were characterized by powder X-ray diffraction (XRD, PANalytical X'Pert³ Powder, Cu Kα), Fourier transform infrared spectroscopy (FT-IR, Bruker Tensor 27 Fourier transform infrared spectrometer), solid-state nuclear magnetic resonance (SSNMR, AVANCE III 400 MHz WB solid-state NMR spectrometer, more details are available in the ESI†) and differential scanning calorimetry (DSC, NETZSCH DSC 204 HP calorimeter, heating rate of 5 K min⁻¹, argon flow).

The dehydrogenation process was characterized by temperature programmed desorption/mass spectrum (TPD/MS, Quantachrome Autosorb iQ automatic gas sorption analyser). The dehydrogenation kinetics was studied in a home-made Sievert type setup. The dehydrogenation capacity was quantified by a thermal conductivity detector (TCD) on the Quantachrome Autosorb iQ automatic gas sorption analyzer and also the Sievert's method as previously reported.³⁶ The TCD signal is calibrated using hydrogen as the external standard (Fig. S1†).

Results and discussion

The synthesis and purification of Im-NH₂BH₄

Fig. 2 shows the XRD patterns of (Im-NH₂)₂SO₄, NaBH₄ and the samples after ball milling with different rotating speeds. Most raw materials remain unchanged when the rotating speed is 100 rpm and 150 rpm. In the case of 200 rpm, most of the starting materials are transformed and Na₂SO₄ is observed. The peaks are ascribed to two different phases of Na₂SO₄, phase V and phase III. The former is the thermodynamically stable phase at room temperature, while the latter is a metastable phase.^43,44 The appearance of phase III may originate from the high energy process during the reaction under ball milling condition. There is some NaBH₄ remaining, as indicated by the peak at 28.9°. However, no other new phases are observed, which indicates that the generated Im-NH₂BH₄ might be amorphous. The sample with 200 rpm is denoted as Im-NH₂-RT.

Fig. 2 XRD patterns of (a) (Im-NH₂)₂SO₄, (b) NaBH₄ and (c–e) the samples after ball milling (Im-NH₂)₂SO₄ and NaBH₄ with different rotating speeds ((c) 100 rpm; (d) 150 rpm; (e) 200 rpm).

Im-NH₂-RT is further characterized by ¹¹B and ¹³C SSNMR spectra (Fig. 3). The main peak of ¹¹B SSNMR spectrum at −40.6 ppm and the main peaks of ¹³C SSNMR spectrum at 149.8 ppm, 146.2 ppm, 118.6 ppm and 109.6 ppm are assigned to BH₄⁻ and Im-NH₂⁺ respectively, which are in accordance with previous studies of NaBH₄ (ref. 31 and 45) and (Im-NH₂)₂SO₄.^46,47 The existence of the BH₄⁻ group is further evidenced by the FT-IR spectrum, in which characteristic B–H stretching bands at 2392 cm⁻¹, 2295 cm⁻¹ and 2229 cm⁻¹ are clearly observed (Fig. S2†).

Fig. 3 ¹¹B SSNMR spectrum (a) and ¹³C SSNMR spectrum (b) of Im-NH₂-RT.

All of these results suggest that Im-NH₂-RT contains BH₄⁻ and Im-NH₂⁺. Considering the fact that Na₂SO₄ appeared and most NaBH₄ and (Im-NH₂)₂SO₄ are consumed as suggested by XRD, it is quite reasonable to conclude that Im-NH₂BH₄ are formed from the metathesis of NaBH₄ and (Im-NH₂)₂SO₄ by ball milling:

2NaBH₄ + (Im-NH₂)₂SO₄ → 2Im-NH₂BH₄ + Na₂SO₄.

We also note some impurities in Im-NH₂-RT, as indicated by the peaks in the ¹¹B SSNMR spectrum at −23.1 ppm and the peaks of ¹³C SSNMR spectrum at 70.4 ppm, 39.2 ppm, 30.3 ppm. The peak of ¹¹B SSNMR spectrum at −23.1 ppm is especially strong, which may be ascribed to a N–BH₃ species according to the previous studies.^37,40 In the case of GuaBH₄, signals between −16 ppm and −24 ppm were assigned to the species with several guanidinium units containing terminal BH₃ groups with different chemical environments.³⁷ Hence, this N–BH₃ species is very likely to form from the partial dehydrogenation of Im-NH₂BH₄ to C₃N₃H₅BH₃ during the ball milling process. The broad shoulder peak in the left of the peak at −23.1 ppm may be assigned to some more complex species such as (C₃N₃H₅)₂BH₂ and (C₃N₃H₅)₃BH. The weak peaks of ¹³C SSNMR spectrum at 70.4 ppm, 39.2 ppm, 30.3 ppm imply that there may be sp³-C species in the Im-NH₂-RT. It is possible because of the strong reducibility of BH₄⁻. C–H bonds was also formed in the dehydrogenated products of GuaB₃H₈.⁴⁰ Furthermore, the shoulder peak from 3000 cm⁻¹ to 2890 cm⁻¹ in the FT-IR spectrum also indicates the existence of sp³ C–H bonds (Fig. S2†).

We spent considerable efforts to avoid the by-products by optimizing the conditions of ball milling. Unfortunately, no significant improvement was obtained. Separation Im-NH₂BH₄ from the Im-NH₂-RT was also attempted using anhydrous tetrahydrofuran (THF), acetonitrile and isopropanol. However, the extraction failed to yield sufficient product. Additionally, in the ¹H NMR spectra of the liquid extract, no BH₄⁻ signals were observed (Fig. S3†). Using liquid ammonia as solvent to synthesize Im-NH₂BH₄ also failed (more details are available in the ESI†). For now, we will proceed to study the dehydrogenation properties of Im-NH₂BH₄ using Im-NH₂-RT. Considering the high thermal stability of Na₂SO₄,^43,44 it is reasonable to consider that its effect on the dehydrogenation of Im-NH₂BH₄ may be little.

Dehydrogenation properties of Im-NH₂BH₄

Fig. 4 shows the dehydrogenation profile of Im-NH₂-RT. The TPD/MS curves (Fig. 4a) show that the onset dehydrogenation temperature is about 50.5 °C and there are three main hydrogenation peaks at 76.7 °C, 169.4 °C and 231.6 °C, respectively. No gaseous impurities were detected below 320 °C, which suggests that the strategy of balancing the number of H⁺ and H⁻ to suppress ammonia release might be effective. In contrast, ammonia is not negligible during the dehydrogenation of GuaBH₄. In addition, the onset dehydrogenation temperature and peak temperature of GuaBH₄ are 90 °C and 100 °C respectively, both of which are higher than the corresponding temperature of Im-NH₂BH₄.^38,41,48 This result suggests that combination of the charge transfer effect between BH₄⁻ and a large conjugated cation (Im-NH₂⁺) and the interaction between H⁺ and H⁻ indeed results in lower dehydrogenation temperature and also improved hydrogen purity.

Fig. 4 (a) TPD/MS curves of Im-NH₂-RT. The heating rate is 5 °C min⁻¹. The blue dotted line indicates the quantities of hydrogen released based on the total mass of sample, which calculated from the TCD integral area. (b) The dehydrogenation kinetics curve of Im-NH₂-RT below 320 °C.

The dehydrogenation capacity of each stage is 0.8 wt%, 0.9 wt% and 1.4 wt% respectively. The total quantities of hydrogen released is 3.1 wt% below 320 °C. Above 320 °C, some gaseous impurities are emitted which are most likely to derive from the decomposition of Im-NH₂⁺. Therefore, we only focused on the hydrogenation properties below 320 °C.

Study on the dehydrogenation kinetics up to 320 °C (Fig. 4b) manifests that Im-NH₂-RT can rapidly release 1.7 wt% hydrogen in the first 10 s and 3.0 wt% hydrogen in 20 min. At last, it can release 3.2 wt% hydrogen, which is in very good agreement with the TCD results. If the weight contribution of Na₂SO₄ is excluded, the hydrogen released will be 5.5 wt%. In other words, 2.8 equiv. H₂ was released from Im-NH₂BH₄ below 320 °C. Considering the partial decomposition during the ball milling process, it is suggested that Im-NH₂BH₄ can release 3 equiv. H₂ below 320 °C and approximately 1 equiv. H₂ in every stage.

The dehydrogenation products and process of Im-NH₂BH₄

After dehydrogenation at 320 °C, the sample turned black, which is denoted as Im-NH₂-320. Im-NH₂-320 was also characterized by XRD (Fig. 5), SSNMR (Fig. 6) and FT-IR (Fig. S2†). The XRD patterns shows that the peaks are mainly ascribed to phase III and phase II of Na₂SO₄.⁴⁴ The peak at 28.9° is assigned to NaBH₄, which also exists in Im-NH₂-RT. However, the assignment of the peak at 47.1° (marked by star) remains unsolved, which may be from the decomposition products of Im-NH₂BH₄.

Fig. 5 XRD pattern of Im-NH₂-RT.

Fig. 6 ¹¹B SSNMR spectrum (a) and ¹³C SSNMR spectrum (b) of Im-NH₂-320.

The main peak in ¹¹B SSNMR spectrum of Im-NH₂-320 at −40.6 ppm is indicative of BH₄⁻, which is originated from the residual NaBH₄ in Im-NH₂-RT. This is in agreement with the high thermal stability of NaBH₄, which remains stable up to 320 °C. The B–H stretching bands of BH4⁻ are also observed in FT-IR spectrum (Fig. S2†). The peaks at 13.5 ppm and 0.6 ppm may correspond to 3 fold coordinated B species such as –BN₃ and/or –BN₂H.^37,40 The peaks in ¹³C SSNMR spectrum at 161.9 ppm, 149.9 ppm and 15.6 ppm can be ascribed to the carbon atoms in the imidazole rings with different chemical environments compared to that of Im-NH₂BH₄. The peak at 42.7 ppm suggests formation of sp³ type carbon species. However, the characteristic signals of sp³-C–H disappear in the FT-IR spectrum of Im-NH₂-320 (Fig. S2†), indicating the formation of C–B or/and C–N bonds. The change of the imidazole ring after dehydrogenation is expected, as release 3 equiv. H₂ requires removing at least one H on the aromatic N.

Fig. 7 shows the DSC curve of Im-NH₂-320 below 320 °C, in which there are three exothermic peaks at 96 °C, 186 °C and 261 °C. This matches well with the three-stage hydrogenation of Im-NH₂BH₄ measured by TPD/MS. The enthalpies of the three stages calculated from the DSC curve are −30.2 kJ mol⁻¹, −28.3 kJ mol⁻¹ and −19.8 kJ mol⁻¹, respectively. Because of the endothermic effect of phase transformations of Na₂SO₄ from 200 °C to 280 °C,⁴⁴ the virtual enthalpies of the latter two stages should be more negative.

Fig. 7 DSC curve of Im-NH₂-RT. The heating rate is 5 °C min⁻¹.

There are four protonic H on N and four hydridic H on B atoms in Im-NH₂BH₄. The strong interaction between H⁺ and H⁻ results in the exothermic hydrogenation process at low temperature, which has been observed in other B–N–H hydrogen storage systems, such as GuaBH₄ and type compounds.^38,49 Compared to GuaBH₄, Im-NH₂BH₄ does not release ammonia during decomposition process as there are no excessive H⁺ in Im-NH₂BH₄. Furthermore, destabilization of BH₄⁻ by the large conjugated cation, Im-NH₂⁺ effectively reduces the dehydrogenation temperature of Im-NH₂BH₄.

Based on the measured hydrogen capacity and the structural characterization of the dehydrogenated products, the dehydrogenation process of Im-NH₂BH₄ could be expressed as:

Im-NH₂BH₄ → C₃N₃H₃BH + 3H₂.

Although the above conclusion is obtained using a mixture of Im-NH₂BH₄ and Na₂SO₄, we believe it is indicative of most of the dehydrogenation properties of Im-NH₂BH₄, as Na₂SO₄ should be inert during the dehydrogenation process. We will continue to purify Im-NH₂BH₄ to test this hypothesis. With an attainable hydrogen capacity of 5.5 wt% and high hydrogen purity, Im-NH₂BH₄ will be a highly attractive hydrogen carrier.

Conclusions

In summary, Im-NH₂BH₄ can be synthesized by the metathesis of NaBH₄ and (Im-NH₂)₂SO₄ using ball milling. The formation of Im-NH₂BH₄ is identified by XRD, SSNMR and FT-IT spectra, albeit with a low purity. Im-NH₂BH₄ starts to release hydrogen at 50 °C, and can release 3.2 wt% pure hydrogen (Na₂SO₄ included) by three exothermic steps below 320 °C. Im-NH₂BH₄ has a theoretical hydrogen capacity of 8.1 wt%, therefore it releases 68% of the stored hydrogen below 320 °C. We believe the interaction between BH₄⁻ and Im-NH₂⁺ and the interaction between H⁺ and H⁻ accounts for the low dehydrogenation temperature of Im-NH₂BH₄, which corresponds with the initial expectation.

Acknowledgements

The authors acknowledge the financial support from National Science Foundation of China (NSFC, No. U1201241, 11375020, 51431001 and 21321001).

Notes and references

1. L. Schlapbach and A. Zuttel, Nature, 2001, 414, 353–358.
2. S. I. Orimo, Y. Nakamori, J. R. Eliseo, A. Zuttel and C. M. Jensen, Chem. Rev., 2007, 107, 4111–4132.
3. J. Graetz, Chem. Soc. Rev., 2009, 38, 73–82.
4. Q. Lai, M. Paskevicius, D. A. Sheppard, C. E. Buckley, A. W. Thornton, M. R. Hill, Q. Gu, J. Mao, Z. Huang, H. K. Liu, Z. Guo, A. Banerjee, S. Chakraborty, R. Ahuja and K.-F. Aguey-Zinsou, ChemSusChem, 2015, 8, 2789–2825.
5. J. Yang, A. Sudik, C. Wolverton and D. J. Siegel, Chem. Soc. Rev., 2010, 39, 656–675.
6. C. J. Webb, J. Phys. Chem. Solids, 2015, 84, 96–106.
7. K. F. Aguey-Zinsou and J. R. Ares-Fernandez, Energy Environ. Sci., 2010, 3, 526–543.
8. R. Y. Yu, Q. A. Li and K. J. Li, Rare Met. Mater. Eng., 2007, 36, 1124–1128.
9. D. L. Zhao and Y. H. Zhang, Rare Met., 2014, 33, 499–510.
10. W. Liu, C. J. Webb and E. M. Gray, Int. J. Hydrogen Energy, 2016, 41, 3485–3507.
11. F. A. H. Yap, N. S. Mustafa and M. Ismail, RSC Adv., 2015, 5, 9255–9260.
12. J. F. Mao and D. H. Gregory, Energies, 2015, 8, 430–453.
13. E. Ronnebro, Curr. Opin. Solid State Mater. Sci., 2011, 15, 44–51.
14. X. Liu, C. Wu, F. Wu and Y. Bai, Progr. Chem., 2015, 27, 1167–1181.
15. W. Cai, H. Wang, D. Sun, Q. Zhang, X. Yao and M. Zhu, RSC Adv., 2014, 4, 3082–3089.
16. R. Moury and U. B. Demirci, Energies, 2015, 8, 3118–3141.
17. Z. H. Lu, Q. Yao, Z. J. Zhang, Y. W. Yang and X. S. Chen, J. Nanomater., 2014, 1, 11.
18. M. Sevilla and R. Mokaya, Energy Environ. Sci., 2014, 7, 1250–1280.
19. Y. H. Lu and Y. P. Feng, Nanoscale, 2011, 3, 2444–2453.
20. L. F. Wang and R. T. Yang, Catal. Rev.: Sci. Eng., 2010, 52, 411–461.
21. J. W. Ren, H. W. Langmi, B. C. North and M. Mathe, Int. J. Energy Res., 2015, 39, 607–620.
22. S. Proch, J. Herrmannsdorfer, R. Kempe, C. Kern, A. Jess, L. Seyfarth and J. Senker, Chem.–Eur. J., 2008, 14, 8204–8212.
23. Q. L. Zhu and Q. Xu, Energy Environ. Sci., 2015, 8, 478–512.
24. T. He, Q. J. Pei and P. Chen, J. Energy Chem., 2015, 24, 587–594.
25. C. Chung, J. Ihm and H. Lee, J. Korean Phys. Soc., 2015, 66, 1649–1655.
26. D. Bao, P. Gao, C. Li, G. Wu, Y. Wang, Y. Chen, H. Zhou and P. Yang, Eur. J. Inorg. Chem., 2016, 2016, 3371–3375.
27. H. W. Li, Y. G. Yan, S. Orimo, A. Zuttel and C. M. Jensen, Energies, 2011, 4, 185–214.
28. H. Kou, G. Sang, Y. Zhou, X. Wang, Z. Huang, W. Luo, L. Chen, X. Xiao, G. Yang and C. Hu, Int. J. Hydrogen Energy, 2014, 39, 11675–11682.
29. H.-W. Li, D. Matsumura, Y. Nishihata, E. Akiba and S.-I. Orimo, J. Jpn. Inst. Met., 2013, 77, 627–630.
30. G. Tu, X. Xiao, T. Qin, Y. Jiang, S. Li, H. Ge and L. Chen, RSC Adv., 2015, 5, 51110–51115.
31. D. Ravnsbaek, Y. Filinchuk, Y. Cerenius, H. J. Jakobsen, F. Besenbacher, J. Skibsted and T. R. Jensen, Angew. Chem., 2009, 48, 6659–6663.
32. T. Jaron, P. A. Orlowski, W. Wegner, K. J. Fijalkowski, P. J. Leszczynski and W. Grochala, Angew. Chem., Int. Ed., 2015, 54, 1236–1239.
33. Q. Gu, Z. Wang, Y. Filinchuk, J. A. Kimpton, H. E. A. Brand, Q. Li and X. Yu, J. Phys. Chem. C, 2016, 120, 10192–10198.
34. L. Liu, G. Wu, W. Chen, Z. Xiong, T. He and P. Chen, Int. J. Hydrogen Energy, 2015, 40, 429–434.
35. X. Zheng, X. Huang, Y. Song, X. Ma and Y. Guo, RSC Adv., 2015, 5, 105618–105621.
36. H. Fu, Y. Wu, J. Chen, X. Wang, J. Zheng and X. Li, Inorg. Chem. Front., 2016, 3, 1137–1145.
37. A. Doroodian, J. E. Dengler, A. Genest, N. Roesch and B. Rieger, Angew. Chem., Int. Ed., 2010, 49, 1871–1873.
38. T. J. Groshens and R. A. Hollins, Chem. Commun., 2009, 3089–3091,  10.1039/b900376b.
39. G. Qi, K. Wang, X. Li and B. Zou, J. Phys. Chem. C, 2016, 120, 13414–13420.
40. W. Chen, Z. Huang, G. Wu, T. He, Z. Li, J. Chen, Z. Guo, H. Liu and P. Chen, J. Mater. Chem. A, 2015, 3, 11411–11416.
41. Z. Tang, Y. Guo, S. Li and X. Yu, J. Phys. Chem. C, 2011, 115, 3188–3193.
42. G. Yanhui, G. Qinfen, G. Zaiping, M. Jianfeng, L. Huakun, D. Shixue and Y. Xuebin, J. Mater. Chem., 2011, 21, 7138–7144.
43. B.-K. Choi and D. J. Lockwood, J. Phys.: Condens. Matter, 2005, 17, 6095–6108.
44. S. E. Rasmussen, J. E. Jorgensen and B. Lundtoft, J. Appl. Crystallogr., 1996, 29, 42–47.
45. R. W. Schurko, I. Hung, S. Schauff, C. L. B. Macdonald and A. H. Cowley, J. Phys. Chem. A, 2002, 106, 10096–10107.
46. Y. Zhang, J. Labelled Compd. Radiopharm., 2002, 45, 1055–1064.
47. F. Wilde, C. Chamseddin, H. Lemmerhirt, P. J. Bednarski, T. Jira and A. Link, Arch. Pharm., 2014, 347, 153–160.
48. Y. Guo, Q. Gu, Z. Guo, J. Mao, H. Liu, S. Dou and X. Yu, J. Mater. Chem., 2011, 21, 7138–7144.
49. B. Peng and J. Chen, Energy Environ. Sci., 2008, 1, 479–483.

---

Footnote

† Electronic supplementary information (ESI) available: Additional figures and some experimental details. See DOI: 10.1039/c6ra21335a

---