Promoting visible light-driven hydrogen evolution over CdS nanorods using earth-abundant CoP as a cocatalyst

Abstract

Exploiting earth-abundant cocatalysts for photocatalytic hydrogen evolution is of great interest. Herein, earth-abundant CoP was used as a cocatalyst to fabricate a CdS nanorod based photocatalytic system. Meanwhile, visible light-driven H₂ evolution over the CoP decorated CdS nanorods was investigated. Moreover, the electrochemical behavior and photoelectrochemical behavior of the CdS/CoP photocatalytic system were explored to deeply demonstrate the essential role that CoP played in the photocatalytic process. Lastly, the photocatalytic mechanism was preliminarily discussed. The results indicated that the photocatalytic activity of the CdS nanorods could be remarkably enhanced due to the introduction of CoP. Under optimal conditions, a rate of H₂ evolution of approximately 2.12 mmol h⁻¹ was achieved and the apparent quantum yield was up to 27.1% at 435 nm, which was about 20 times higher than that over the CdS nanorods and 1.5 times higher than that over the Pt loaded CdS nanorods, respectively. This dramatic photocatalytic activity of the CoP modified CdS nanorods may be ascribed to the low H⁺ reduction overpotential on the CoP particles and the efficient electron transfer between CdS and CoP. The present work demonstrates that CoP is a promising alternative to Pt as an efficient cocatalyst for CdS.

---

1. Introduction

Light-driven H₂ evolution over semiconductors has been regarded as a promising strategy to convert solar energy into stored chemical energy.^1–3 Thus, semiconductor photocatalysts have been extensively investigated for hydrogen evolution in the past several decades.^4,5 Unfortunately, most of them suffer from low photocatalytic efficiency and poor stability, which severely limit their practical application.⁶ In order to enhance the photocatalytic activity and stability of semiconductors, a lot of effort has been devoted to the development of a cocatalyst.^7,8 Currently, platinum has been proven to be an efficient cocatalyst because of its high work function and low overpotential for hydrogen evolution.⁹ However, Pt is high-cost and rare, which is unfavorable to large scale applications. Therefore, it is desirable to find an inexpensive and earth-abundant alternative to Pt.

The roles of cocatalysts in a photocatalytic system include enhancing the separation of photogenerated charge carriers, providing active sites for catalytic hydrogen evolution, and suppressing photo-corrosion. Numerous noble-metal-free materials have been applied as cocatalysts in the past several decades, such as carbon-based materials,^10–14 transition metals,^15–17 transition metal oxides,^18,19 hydroxides,^20,21 sulfides^22,23 and so on. Recently, the experimental results show that transition-metal phosphides (TMPs), such as Ni₂P,²⁴ Cu₃P,²⁵ CoP^26–28 and FeP,²⁹ possess excellent electrocatalytic activity for the electrocatalytic hydrogen evolution. In the electrocatalytic processes above, they exhibit many attractive advantages, including high current density, low overpotential, good stability etc. These characteristics of phosphides imply that if phosphides are loaded on semiconductors, they could serve as an efficient cocatalyst in the photocatalytic H₂ evolution.¹⁵ Unfortunately, there are still limited reports on the applications of phosphides in the field of light-driven H₂ evolution although phosphides as cocatalysts have begun to attract interest. And the role which phosphides play in the photocatalytic process is not very clear. Therefore, it is essential to explore detailedly the photocatalytic behavior of the phosphide decorated semiconductor for light-driven H₂ evolution.

Among all the semiconductor photocatalysts, cadmium sulfide nanorods are promising candidates for visible light photocatalytic hydrogen evolution due to its suitable band position, large surface area and fast charge separation efficiencies.^30,31 Herein, we report a simple and low-cost synthesis route to load CoP as cocatalyst on CdS nanorods to build a highly effective photocatalytic system for hydrogen evolution, and the photocatalytic band structure of CoP/CdS system is proposed concretely. Under optimal loading condition, the CoP decorated CdS nanorods can achieve the photocatalytic activity of 2.12 mmol h⁻¹ under visible light irradiation, which is higher than the optimal value of CdS, Pt/CdS and the physical mixture of CoP and CdS. The band structure of the CdS/CoP photocatalytic system is also investigated with the characterization of UV-vis absorption and electrochemical analysis. The results indicate that CoP can extract photogenerated electrons from CdS and significantly improve the efficiency of subsequent hydrogen evolution reaction.

2. Experimental

2.1. Materials

All materials were of analytical grade and used as received without further purification. Cd(NO₃)₂·4H₂O, Co(NO₃)₂·6H₂O, thiourea, ethylenediamine, lactic acid and NaOH were obtained from Sinopharm Chemical Reagent Co., Ltd. (Beijing, China). NaH₂PO₂ were obtained from Aladdin Industrial Inc. (Shanghai, China). Pt/C (20 wt% Pt on Vulcan XC-72R), Nafion (5 wt%) and carbon black were obtained from Sigma-Aldrich (China). Ultrapure water with a resistivity of 18.2 MΩ cm, produced with a Milli-Q® Integral Water Purification System (Millipore), was used in all the experiments.

2.2. Synthesis of CdS nanorods, CoP decorated CdS nanorods and CoP particles

The CdS nanorods were synthesized using a modified solvothermal method.³¹ 20 mmol Cd(NO₃)₂·4H₂O and 40 mmol thiourea were dispersed in 60 mL ethylenediamine under stirring. After 30 min, the above solution was transferred into a 100 mL Teflon-lined autoclave and heated at 160 °C for 48 h. The resulting precipitate was collected, rinsed with water and ethanol alternately for several times, and dried under vacuum at 60 °C overnight.

A simple two-step reaction was used to synthesize the CoP decorated CdS nanorods. At first, 2 mmol CdS nanorods was dispersed in 30 mL water, and Co(NO₃)₂ solution was added according to designated molar ratio. Then 1 M NaOH solution was dropped into the above suspension with stirring for 4 h. The collected precipitate was washed with water and ethanol for several times and dried under vacuum at 60 °C overnight. Secondly, the precipitate obtained was placed next to the NaH₂PO₂ at a downstream side in a tube furnace. The furnace was heated to 250 °C in 30 min, and then slowly ramped to 350 °C with the rate of 1 °C per minutes and kept for 2 h in the atmosphere of Ar (50 sccm). As shown in Scheme 1, Co(OH)₂ was loaded on the surface of CdS nanorods using facile deposition–precipitation method, and then Co(OH)₂ was convert into cobalt phosphide through phosphidation treatment meanwhile maintaining the original morphology of CdS nanorods. The CoP decorated CdS nanorods are abbreviated to CoP/CdS-x%, in which x% is the molar percent of CoP cocatalyst.

Scheme 1 Formation process of the CoP decorated CdS nanorods.

The pure CoP particles were prepared in the same way as the CoP decorated CdS nanorods except for the presence of CdS nanorods.

2.3. Synthesis of the Pt loaded CdS nanorods

The Pt loaded CdS nanorods were prepared using photodeposition method. In a typical procedure, the as-prepared CdS nanorods were put into a quartz reactor containing water under stirring, and then various volume of H₂PtCl₆ aqueous solution were injected. Finally, the suspension above was irradiated for 30 min under stirring.

2.4. Characterization

X-ray diffraction (XRD) was carried out on a Rigaku D/max-7000 using filtered Cu Kα irradiation. Field-emission scanning electron microscopy (FESEM) image was taken on a JSM-6701F. Transmission electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM) images were obtained on a JEM-2100F operated at 200 kV. X-ray photoelectron spectroscopy (XPS) data were obtained with an ESCALab220i-XL electron spectrometer from VG Scientific using 300 W Al Kα radiation, the base pressure was about 3 × 10⁻⁹ mbar, the binding energies were referenced to the C1s line at 284.8 eV from adventitious carbon. The UV-visible absorption spectra were recorded with a UV-visible spectrophotometer UV-3900. Photoluminescence (PL) spectra were recorded using a HITACHI F-4600 fluorescence spectrophotometer with an excitation wavelength at 405 nm.

2.5. Electrochemical measurements

A typical three-electrode system was used and controlled by a CHI 760E electrochemical analyzer (Chen Hua, China). Ag|AgCl (in 3 M KCl solution) electrode and Pt wire were used as the reference and counter electrodes, respectively. The catalyst ink was prepared as follows: 5.0 mg of catalyst was added into a mixed solvent containing 980 μL of absolute ethanol and 20 μL of 5 wt% Nafion solution, and then the mixture was dispersed by sonication for about 30 min to form a homogeneous ink. HER electrode was prepared by loading 4 μL of catalyst ink onto a glass carbon electrode with 3 mm diameter. The linear sweep voltammetry with a scan rate of 5 mV s⁻¹ was conducted in 0.5 M H₂SO₄ solution saturating with N₂ for 30 min. The ITO glass electrode (effective area 0.1962 cm²) coated with the as-prepared inks 5 μL (5 mg mL⁻¹ in absolute ethanol solution) was used as the working electrode for photocurrent measurement, and a 10 v/v% lactic acid aqueous solution saturating with N₂ was used as the electrolyte. All of the potentials were referenced to a reversible hydrogen electrode (RHE) according to E(vs. RHE) = E(vs. Ag|AgCl) + E^θ(vs. Ag|AgCl) + 0.059pH.

2.6. Photocatalytic activity test

Photocatalytic H₂ evolution was performed in a Pyrex glass cell which had a flat, round upside window with an irradiation area of 38 cm² for light exposure. A 300 W xenon arc lamp with a 420 nm cut-off filter (PLS-SEX 300, Beijing Trusttech Technology Co., Ltd.) was used to simulate the visible light source. The spectrum of the lamp without filter was shown in Fig. S1 (ESI†). The illumination intensity was adjusted to 100 mW cm⁻². The H₂-solar system (Beijing Trusttech Technology Co., Ltd.) with a gas chromatogram (GC), equipped with a thermal conductivity detector (TCD), TDX-01 column and Ar carrier gas, was used to collect and on-line detect evolved H₂. 0.02 g of photocatalyst was suspended in a quartz glass cell with 80 mL of deionized water containing 10 v/v% lactic acid. The cell was kept at 5 °C by using a circulating water system. Before irradiation, the reaction system was pumped into vacuum. The hydrogen plant of photocatalytic experiment is shown in Fig. S2 (ESI†).

3. Results and discussion

3.1. Structure and property of CoP

The synthesis process for CoP is depicted in the Experimental section. Briefly, Co(OH)₂ was obtained by the way of deposition–precipitation method. Then, the Co(OH)₂ was treated with NaH₂PO₂ in Ar atmosphere to obtain CoP. Fig. 1a shows the XRD patterns of Co(OH)₂, CoP and the intermediate Co₂P/CoP composites. The XRD peaks of prepared Co(OH)₂ is in accord with the lattice diffraction of Co(OH)₂ (JCPDF 30-0433). After phosphidation, no peaks of Co(OH)₂ is observed. The main XRD peaks of CoP is clearly observed at 31.7°, 36.5°, 46.3°, 48.4° and 56.6°, which correspond to the (011), (111), (112), (211) and (301) lattice diffraction of CoP (JCPDF 29-0497), respectively. Meanwhile, we found that temperature was a vital factor to CoP synthesis. When the temperature of phosphidation was lower than 350 °C, a strong peak at 41° is observed, which corresponds to the (112) plane of Co₂P. When the temperature increases to 350 °C, the characteristic XRD peaks of Co₂P disappear. We speculate that Co₂P is an intermediate phase for the phosphidation of Co(OH)₂ to CoP, which agrees with the previously reported results.³² The HRTEM image shown in Fig. 1b shows the lattice fringes with spacing of 0.28 nm and 0.19 nm, which correspond to the (011) and (211) planes of CoP, respectively, matching well with the XRD data.

Fig. 1 (a) XRD patterns of samples with different annealing temperature, (b) TEM image of CoP, (c) LSV of CoP and 20% Pt/C in 0.5 M H₂SO₄ at a scan rate of 5 mV s⁻¹, (d) Tafel slope of CoP.

HER electrocatalytic performance of CoP is presented in Fig. 1c. The polarization curves of CoP demonstrate an onset overpotential of ∼40 mV versus RHE, and a lower overpotential of ∼123 mV at current density of 10 mA cm⁻². The Tafel slope, which determines the rate-controlling step of HER, is an intrinsic property of catalyst. For the CoP, we get a small Tafel slope of 50 mV dec⁻¹ showed in Fig. 1d, which suggesting that the rate-controlling step is electrochemical desorption.²⁸ Overall, these results suggest CoP is an excellent electrocatalyst for hydrogen evolution.

3.2. Structure and property of CoP/CdS

CoP/CdS was prepared using a two-step method (Scheme 1). The XRD patterns of the CdS nanorods and CoP/CdS are showed in Fig. 2a. The CdS nanorods and CoP/CdS show the same XRD patterns, which correspond to the lattice diffraction of hexagonal CdS (JCPDF 77-2306). No obvious peaks of CoP are observed because of the low deposition contents on the surface of CdS nanorods.

Fig. 2 (a) XRD patterns of CdS and CoP/CdS with various amounts of loading, (b) FESEM, (c) TEM and (d) HRTEM images of CoP/CdS.

To further characterize the morphology of CoP/CdS, FESEM, TEM, and HRTEM images were taken. Fig. 2b shows that CoP/CdS adopts a rod-like morphology with diameters of about 30–40 nm, and the loading of CoP has no effect on the morphology of CdS nanorods (Fig. S3, ESI†). The TEM image (Fig. 2c) shows that some CoP nanoparticles are dispersed on the surface of CdS nanorods. Fig. 2d shows the HRTEM image of CoP/CdS, and the distance of lattice fringe corresponds to the (011) facet of CoP and the (101) facet of CdS, respectively. CoP is closely interfaced with CdS lattice, which may be favorable to the charge transfer process. The chemical states of CoP/CdS were further analyzed with X-ray photoelectron spectroscopy (XPS). The XPS survey spectra (Fig. S4, ESI†) shows that Co, P, Cd, S elements are detected. The peaks in high resolution XPS spectra correspond to the binding energies of CoP and CdS, respectively (Fig. S5, ESI†). These results indicate that the CoP decorated CdS nanorods are successfully prepared in the process described above.

3.3. Evaluation of photocatalytic activity

The photocatalytic hydrogen evolution activity of CoP/CdS was measured using lactic acid as a sacrificial agent in a typical photocatalytic reaction system as described in Experiment section. Fig. 3a shows that the photocatalytic activity of the CdS nanorods can be dramatically enhanced due to the introduction of CoP. The effect of loading amount of CoP on the photocatalytic activity of the CdS nanorods was explored. The results show that the optimal loading amount of CoP is 3%. The average hydrogen evolution rate over CoP/CdS-3% is about 2.12 mmol h⁻¹ under visible light. In contrast, when the CdS nanorods are used as a photocatalyst, the hydrogen evolution rate is only 0.1 mmol h⁻¹. In addition, the photocatalytic activity of the Pt loaded CdS nanorods was also investigated (Fig. S6, ESI†). It can be found that the optimal hydrogen evolution rate of 0.86 mmol h⁻¹ is achieved in the presence of the Pt loaded CdS nanorods. In other words, the photocatalytic activity of CoP/CdS is about 20 times higher than that of the CdS nanorods, and outperforms the Pt loaded CdS nanorods.

Fig. 3 Photocatalytic hydrogen evolution rate in the presence of (a) CoP/CdS, (b) simple mixture CoP particles and CdS nanorods and (c) CdS nanorods loaded with various cocatalysts under visible-light irradiation (λ > 420 nm), (d) cycling runs of CoP/CdS-3% under visible-light irradiation.

It is noted that CoP itself shows no appreciable hydrogen evolution ability under the same condition, indicating that CoP serves as a cocatalyst in CoP/CdS. Meanwhile, the results shown in Fig. 3b indicate that the photocatalytic activity of the simple mixture of CoP particles and CdS nanorods is much weaker than that of CoP/CdS, suggesting that the good interface between CoP nanoparticles and CdS nanorods is necessary for fabrication of an efficient photocatalytic system. We also investigated the photocatalytic activities of the Co(OH)₂ modified CdS nanorods (Fig. 3c). Despite of the presence of cobalt oxidation (Fig. S5, ESI†), the photocatalytic hydrogen generation rate over CoP/CdS is much higher than that over the Co(OH)₂ modified CdS nanorods. Based on the results above, we can conclude that the photocatalytic activity enhancement of the CdS nanorods is mainly contributed to CoP.

The photocatalytic stability of CoP/CdS under visible light irradiation is shown in Fig. 3d. The reaction system was collected every one hour, and this process was repeated for several cycles. It shows that there is no significant decrease in hydrogen production through each cycle, indicating that the CoP/CdS composite possesses good stability for hydrogen evolution. For the third cycling, a slight decrease is observed. When 4 mL lactic acid was put into the decreased reaction system, the hydrogen generation increased and maintained for several hours. We can infer that the slight decrease might be mainly attributed to the lack of sacrificial agent.

The quantum yield of hydrogen (QE) is calculated according to the following equation:³³

where r_H2 stands for the average hydrogen rate (mol s⁻¹); N_A is Avogadro's constant 6.022 × 10²³ (mol⁻¹); I represents the light density of incident light (W cm⁻²); S means the irradiation area (cm²); t stands for the irradiation time (s); λ is usually used to express the wavelength of incident light (nm); h is Planck constant 6.626 × 10⁻³⁴ (J s) and c represents the speed of light 3 × 10⁸ m s⁻¹.

Using monochromatic light at λ = 435 nm, the average rate of CoP/CdS is 0.6732 mmol h⁻¹ with 20 mg catalyst, and the rate can reach 0.6081 mmol h⁻¹ after illumination for 24 h. The QE is calculated to be 27.08%, and the QE can still reach 24.47% after illumination for 24 h.

3.4. Band structure and mechanism of photocatalysis

UV-vis diffuse reflectance, photocurrent response, Mott–Schottky plot and photoluminescence are used to determine the band structure of CoP/CdS and understand the photocatalytic mechanism. UV-vis spectra of the CdS nanorods loaded with various molar percentages CoP are shown in Fig. 4a. The absorption intensity of CoP/CdS is obviously stronger in comparison with that of the CdS nanorods, especially in the visible-lights region. However, the absorption edge of CdS is almost not shifted after loading, implying that CoP is loaded on the surface rather than introduced in the crystal lattice of CdS. We also measured the UV-vis absorption spectrum of CoP (Fig. 4b). The band gap value of CoP is about 1.73 eV determined by Tauc's plot as shown in the Fig. 4b insert, which is consistent with the value reported previously.²⁶

Fig. 4 (a) UV-vis spectrum of CoP/CdS with the molar percent of CoP increasing from 0% to 7% (insert: Tauc's plots of the CdS nanorods and CoP/CdS-3%), (b) UV-vis spectra of CoP (insert: Tauc's plot of CoP), (c) photocurrent responses of CoP/CdS with various molar percent of CoP under visible-light irradiation at a bias-potential of 0 V, (d) PL spectrum of the CdS nanorods and CoP/CdS-3%, (e) Mott–Schottky plot of CoP, (f) schematic description of an artificial photocatalytic hydrogen evolution.

The photocurrent response of CoP/CdS was recorded on a home-built photoelectrochemical measurement system. Under irradiation, CoP and CdS can be excited to generate the electron–hole pairs. Next, the electrons and holes are oxidized and reduced on Pt wire and working electrode, respectively. As shown in Fig. 4c, the ITO electrodes coated with CoP/CdS demonstrate various photocurrent densities with the loading amounts of CoP changing. As the loading amount of CoP increases from 1% to 3%, a remarkable enhanced photocurrent is observed. However, the photocurrent significantly decreases when the dosage of CoP further increases. Thus, we can conclude that the appropriate loading of CoP on the CdS nanorods can efficiently suppress the recombination of the photogenerated charge carriers, while the excessive CoP may act as recombination center so as to decrease the electron lifetime. The result of photocurrent responds match well with the hydrogen evolution rate results. Fig. 4d shows the PL data with excitation wavelength at 405 nm. As can be seen from Fig. 4d, the CdS nanorods exhibit a broad emission peak at 540 nm. When CoP is loaded on the surface of the CdS nanorods, the emission is remarkably quenched. This phenomenon further confirms that there exists fast transfer of electrons from the CdS nanorods to the CoP nanoparticles, which can suppress the recombination of electrons and holes and improve the photocatalytic activity of the CdS nanorods.

The Mott–Schottky (MS) plot shows that CoP is n-type semiconductor, and the flat-band potential is −0.43 V vs. RHE (Fig. 4e and insert), which is more negative than the potential for hydrogen evolution (−0.059 V vs. NHE, pH 1), indicating that the CoP nanoparticles loaded can act as active sites for photocatalytic hydrogen evolution. Fig. 4f summarized the photocatalytic mechanism over CoP/CdS. The CoP nanoparticles and the CdS nanorods absorb light to generate electron–hole pairs, and then the carrier charges separate and migrate to the surface of semiconductors. The redox reaction occurs on the surface of the CdS nanorods in the absence of CoP. When the CoP nanoparticles are loaded on the surface of the CdS nanorods, the photo-generated electrons can transfer rapidly from the conduction band of CdS to the conduction band of CoP. Then, the electrons injected into the conduction band of CoP can conveniently react with H⁺ in the water to generate hydrogen due to the low H⁺ reduction overpotential on CoP particles. Here, the CoP nanoparticles loaded serve as the charge separation centres and active sites for photocatalytic hydrogen evolution. The enhancement of photocatalytic activity of the CdS nanorods may be mainly ascribed to decreasing of the H⁺ reduction overpotential and the more efficient separation of electron–hole pairs.

4. Conclusions

In conclusion, CoP is an efficient cocatalyst for the CdS nanorods. The photocatalytic activity can be obviously enhanced when the CoP nanoparticles are loaded on the CdS nanorods. In the case of CoP/CdS-3%, the hydrogen evolution rate and the apparent quantum yield at 435 nm can be up to 2.12 mmol h⁻¹ and 27.08%, respectively. This hydrogen evolution rate is about 21 times of that over the pure CdS nanorods or 2.5 times of that over the Pt loaded CdS nanorods. Here, the CoP nanoparticles loaded serve as the charge separation centres and active sites for photocatalytic hydrogen evolution. The dramatic photocatalytic activity of CoP/CdS ought to be ascribed to the low H⁺ reduction overpotential on CoP particles and the efficient electron transfer between CdS and CoP. The present work provides new possibility for designing an efficient Pt-free photocatalyst. Further efforts are currently being undertaken.

Acknowledgements

This work was financially supported by the Open Project of Bejing National Laboratory for Molecular Sciences (No. 20140163) and Shanghai Municipal Education Commission (Plateau Discipline Construction Program).

Notes and references

1. N. Armaroli and V. Balzani, Angew. Chem., Int. Ed., 2007, 46, 52–66.
2. J. Barber, Chem. Soc. Rev., 2009, 38, 185–196.
3. S. Y. Chen and L. W. Wang, Chem. Mater., 2012, 24, 3659–3666.
4. X. Chen, S. Shen, L. Guo and S. S. Mao, Chem. Rev., 2010, 110, 6503–6570.
5. T. Hisatomi, J. Kubota and K. Domen, Chem. Soc. Rev., 2014, 43, 7520–7535.
6. Y. Qu and X. Duan, Chem. Soc. Rev., 2013, 42, 2568–2580.
7. J. Ran, J. Zhang, J. Yu, M. Jaroniec and S. Z. Qiao, Chem. Soc. Rev., 2014, 43, 7787–7812.
8. J. Yang, D. Wang, H. Han and C. Li, Acc. Chem. Res., 2013, 46, 1900–1909.
9. H.-l. Liu, F. Nosheen and X. Wang, Chem. Soc. Rev., 2015, 44, 3056–3078.
10. Z. Chen, S. Liu, M.-Q. Yang and Y.-J. Xu, ACS Appl. Mater. Interfaces, 2013, 5, 4309–4319.
11. J. Zhang, Y. Wang, J. Jin, J. Zhang, Z. Lin, F. Huang and J. Yu, ACS Appl. Mater. Interfaces, 2013, 5, 10317–10324.
12. G. Xie, K. Zhang, B. Guo, Q. Liu, L. Fang and J. R. Gong, Adv. Mater., 2013, 25, 3820–3839.
13. X. Jiang, J. Nisar, B. Pathak, J. Zhao and R. Ahuja, J. Catal., 2013, 299, 204–209.
14. N. Zhang, Y. Zhang and Y. J. Xu, Nanoscale, 2012, 4, 5792–5813.
15. W. Y. Zhang, Y. X. Li, X. P. Zeng and S. Q. Peng, Sci. Rep., 2015, 5, 10589–10600.
16. B. Li, H. Cao, J. Yin, Y. A. Wu and J. H. Warner, J. Mater. Chem., 2012, 22, 1876–1883.
17. Z. Han, F. Qiu, R. Eisenberg, P. L. Holland and T. D. Krauss, Science, 2012, 338, 1321–1324.
18. L. J. Zhang, S. Li, B. K. Liu, D. J. Wang and T. F. Xie, ACS Catal., 2014, 4, 3724–3729.
19. C. Hu, K. Chu, Y. Zhao and W. Y. Teoh, ACS Appl. Mater. Interfaces, 2014, 6, 18558–18568.
20. L. J. Zhang, R. Zheng, S. Li, B. K. Liu, D. J. Wang, L. L. Wang and T. F. Xie, ACS Appl. Mater. Interfaces, 2014, 6, 13406–13412.
21. J. Ran, J. Yu and M. Jaroniec, Green Chem., 2011, 13, 2708–2713.
22. X. Zong, H. Yan, G. Wu, G. Ma, F. Wen, L. Wang and C. Li, J. Am. Chem. Soc., 2008, 130, 7176–7177.
23. Y. Li, H. Wang and S. Peng, J. Phys. Chem. C, 2014, 118, 19842–19848.
24. Z. Pu, Q. Liu, C. Tang, A. M. Asiri and X. Sun, Nanoscale, 2014, 6, 11031–11034.
25. J. Tian, Q. Liu, N. Cheng, A. M. Asiri and X. Sun, Angew. Chem., Int. Ed., 2014, 53, 9577–9581.
26. J. Tian, Q. Liu, A. M. Asiri and X. Sun, J. Am. Chem. Soc., 2014, 136, 7587–7590.
27. X. Yang, A.-Y. Lu, Y. Zhu, M. N. Hedhili, S. Min, K.-W. Huang, Y. Han and L.-J. Li, Nano Energy, 2015, 15, 634–641.
28. H. Yang, Y. Zhang, F. Hu and Q. Wang, Nano Lett., 2015, 15, 7616–7620.
29. J. Tian, Q. Liu, Y. Liang, Z. Xing, A. M. Asiri and X. Sun, ACS Appl. Mater. Interfaces, 2014, 6, 20579–20584.
30. Y. Li, L. Tang, S. Peng, Z. Li and G. Lu, CrystEngComm, 2012, 14, 6974–6982.
31. Y. Li, Y. Hu, S. Peng, G. Lu and S. Li, J. Phys. Chem. C, 2009, 113, 9352–9358.
32. D.-H. Ha, L. M. Moreau, C. R. Bealing, H. Zhang, R. G. Hennig and R. D. Robinson, J. Mater. Chem., 2011, 21, 11498–11510.
33. H. Xu, X. Li, S.-Z. Kang, L. Qin, G. Li and J. Mu, Int. J. Hydrogen Energy, 2014, 39, 11578–11582.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra04612f

---