Preparation and enhanced photoelectrochemical performance of a p–n heterojunction CuFe₂O₄/WO₃ nanocomposite film

Abstract

In this work, a p–n heterojunction film consisting of n-type WO₃ and p-type CuFe₂O₄ was synthesized via two steps. The n-type WO₃ film was deposited on the FTO substrate by a doctor-blade method and then modified with p-type CuFe₂O₄ nanoparticles by a deposition-annealing method. The composite film was characterized by X-ray powder diffraction, scanning electron microscopy, transmission electron microscopy, X-ray photoelectron spectroscopy and UV-vis diffuse reflectance spectroscopy, showing that the CuFe₂O₄ nanoparticles were deposited on the surface of WO₃ film. Meanwhile, photoelectrochemical measurements were used to investigate the photoelectrochemical properties. A photocurrent of 0.75 mA cm⁻² at 0.6 V (vs. Ag/AgCl) was achieved with CuFe₂O₄/WO₃, resulting in a 2.68 fold increase compared to pristine WO₃. The presence of a p–n heterojunction facilitates the separation of photoinduced electrons and holes, leading to more efficient charge transfer, resulting in a significant improvement in PEC performance.

---

1. Introduction

To solve the severe energy crisis and environmental pollution issues caused by the excessive consumption of traditional fossil fuels (coal, crude oil and natural gas) in the past decades, a lot of research has been reported. Among it, photoelectrochemical (PEC) water splitting, as an efficient, green and promising approach to generate renewable energy from sunlight, has triggered enormous research activities since Honda and Fujishima's report.¹ To date, various kinds of semiconductor materials have been synthesized and investigated.^2,3 Because of its photosensitivity, chemical stability and environmental friendliness, tungsten trioxide (WO₃) has been regarded as one of the ideal materials for solar-driven water splitting.⁴

Even though WO₃ has these promising characteristics, further improving the photoelectrochemical performance is necessary in order to satisfy practical application requirements. Besides loading WO₃ on the conductive substrates (FTO or ITO) is in favor of the reuse and recycle of photocatalyst, most of the researchers have focused on broadening the photoresponse region of semiconductor or improving the separation of the photo-generated electrons and holes.^5–7 The main methods for achieving the goal have included noble metal deposition,^8,9 metal or non-metal element doping,^5,6,10,11 surface modification,¹² or coupled with other semiconductors.^13–18 Among them, coupled with other semiconductors, which can promote the separation of the photoinduced electron–hole pairs via the different position of valence and conduction band from each other, is an effective method to improve the efficiency of water splitting.

MFe₂O₄ (M = Cu, Zn, Co etc.), which is a high thermal stable material with effective catalytic activity, has been used extensively in photocatalysis and solar water-splitting.^19–21 Since most of the ferrates are p-type semiconductors, some of them can couple with n-type semiconductors to form p–n heterojunction composite particles (CuFe₂O₄/CdS, CuFe₂O₄/SnO₂, CuFe₂O₄/TiO₂ and CaFe₂O₄/WO₃).^14,22–24 For example, Miyauchi et al. prepared CaFe₂O₄/WO₃ composite powders by mechanically mixing n-type WO₃ with p-type CaFe₂O₄ particles, and the composite photocatalyst showed better performance than WO₃ on the photocatalytic decomposition of acetaldehyde.¹⁴ The enhanced photocatalysis activity can be ascribed to the separation of the photo-generated electron–hole pairs by their potential difference at the p–n heterojunction. Even though the powdered composite photocatalyst is a great candidate in photocatalysis, its application is limited due to the low recyclability. To cover the shortage, loading p-type ferrate on the surface of n-type semiconductor film is expected to be an efficient route. Even though several works on the photocatalysis or photoelectrochemical property of n-type semi-conductor with p-type semiconductor (ferrate,^25,26 Cu₂O,^27–30 NiO,³¹ Si,³² CdTe³³ and BiOI³⁴ etc.) film have been reported, the n-type WO₃ film loaded with ferrate is lack of attention. To the best of our knowledge, this is the first report to study the photoelectrochemical performance of WO₃ film loaded with CuFe₂O₄ nanoparticles.

In this paper, we fabricated the CuFe₂O₄/WO₃ p–n heterojunction film by doctor-blade method and deposition-annealing method. Meanwhile, the photoelectrochemical properties had been systematically characterized by several methods, including linear sweep voltammetry, electrochemical impedance spectroscopy (EIS), Mott–Schottky and incident photon to current conversion efficiency (IPCE).

2. Experimental

2.1 Preparation of CuFe₂O₄/WO₃ heterojunction film

All chemicals were analytical grade and were used without further purification. In a typical synthesis, the WO₃ spherical nanopowders were prepared by a solution method.³⁵ Then, 0.175 g obtained WO₃ nanopowders were added to an agate jar with terpineol (0.8 mL) as a solvent. PEG20000 (0.05 g) and ethylcellulose (0.025 g) were added to the agate jar as additive. Besides, 0.2 mL acetylacetone and 0.2 mL triton were also added to the agate jar. The WO₃ paste was acquired after ball milling for 8 h. The WO₃ films were fabricated by the doctor blade method on the FTO glass substrates that had been cleaned up. Finally, the as-prepared WO₃ films were calcined at 450 °C for 0.5 h.

A deposition-annealing method was used to synthesize the CuFe₂O₄ nanoparticles on the as-prepared WO₃ film (1.5 cm × 1 cm). In detail, 0.060 g of Cu(CH₃COO)₂·H₂O was dissolved into 10 mL of de-ionized (DI) water to form a transparent solution firstly (solution A). Then, 0.196 g of K₃Fe(CN)₆ was dissolved into another 30 mL of DI water to form a transparent solution (solution B). After the two solutions were stirred for 10 min in the ice bath, respectively, the solution A was added into the solution B drop by drop under vigorously stirred in the ice bath. After further stirred in the ice bath for another 1 h, the as-prepared WO₃ film was put into the solution. The mixture was transferred into a refrigerator, keeping at 0–4 °C in a refrigerator for 2 h. Then the film was taken out and washed by dipping in the DI water for 30 seconds. After drying at 60 °C, the CuFe₂O₄/WO₃ film was calcined at 500 °C for 1 h. For comparison, the pristine WO₃ film and the mixed solution were also calcined to prepare pure WO₃ and bare CuFe₂O₄, respectively.

2.2 Characterization

The phase formations of the as prepared films were determined by an X-ray diffractometer (D/Max2250). The morphology of the materials was studied with a field emission scanning electron microscope (FESEM, NanoSEM 230) equipped with an Energy Dispersive X-ray (EDX) analyzer and a transmission electron microscope (TEM, TECNAI G2 F20). The surface elemental analysis was carried out by an X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi). UV-vis spectra were determined using a UV-vis spectrophotometer (UV-vis, PGeneral TU-1901) in diffuse reflectance mode.

2.3 Photoelectrochemical measurements

The PEC performance of films was characterized by a typical three-electrode electrochemical cell with an electrochemical analyzer (Zennium, Zahner, Germany), using an Ag/AgCl (saturated KCl) reference electrode and a platinum counter electrode. A 0.2 M Na₂SO₄ aqueous solution was used as electrolyte for all of photoelectrochemical tests. Photocurrent–potential plots were characterized by linear sweep voltammogram at 20 mV s⁻¹ under AM 1.5G illumination, and the scanned range was from 0 to 1.0 V (vs. Ag/AgCl). The electrochemical impedance spectra (EIS) were recorded at 0.3 V (vs. Ag/AgCl) with a 10 mV AC voltage perturbation, and the frequency were varied from 10 000 to 0.1 Hz. Additionally, incident photon to electron conversion efficiency (IPCE) of the film was obtained as a function of wavelength from 310 to 700 nm, and the absolute intensity of the monochromic incident light was measured by a silicon photodiode.

3. Results and discussion

The XRD patterns of WO₃ and CuFe₂O₄/WO₃ films are shown in Fig. 1. It can be easily seen that the XRD pattern of each sample matches well with the standard monoclinic phase WO₃ (JCPDS no. 83-0950). Moreover, there are also some peaks corresponding to tetragonal SnO₂ (JCPDS no.46-1088), meaning that the samples were loaded on the FTO substrate. The XRD pattern of the bare CuFe₂O₄ was also measured as shown in Fig. S1.† The peaks at 2θ values of 18.32, 30.18, 35.60 and 37.10° can be indexed to (111) (220) (311) and (222) crystal planes of CuFe₂O₄ (JCPDS no. 77-0010). However, typical peaks of CuFe₂O₄ or other materials are not observed in CuFe₂O₄/WO₃ (Fig. 1), which indicates that the concentration of CuFe₂O₄ is too low to be detected by XRD.

Fig. 1 XRD patterns of WO₃ and CuFe₂O₄/WO₃ films.

The SEM images of the WO₃ and CuFe₂O₄/WO₃ films are presented in Fig. 2. In Fig. 2a and b, the top-view of SEM images exhibit that both samples are nanoporous network composed of spherical particles, which are agglomerate and interconnected with each other. As can be seen in the cross sectional images (Fig. 2c and d), the thickness of both WO₃ and CuFe₂O₄/WO₃ films are about 3.2 μm however, the layer of CuFe₂O₄ can not be found. It means that CuFe₂O₄ may compactly deposit on the surface of WO₃ film. Furthermore, the energy dispersive X-ray (EDX) analyses were recorded as shown in the Fig. S2,† and Cu and Fe can be detected in the sample of CuFe₂O₄/WO₃ (Fig. S2b†).

Fig. 2 SEM images of the surface morphology of (a) WO₃ and (b) CuFe₂O₄/WO₃ films; cross-sectional micrographs of (c) WO₃ and (d) CuFe₂O₄/WO₃ films.

To get more detailed information on the crystalline structure of the composite, the TEM images of CuFe₂O₄/WO₃ at different magnifications are displayed in Fig. 3. The low resolution image (Fig. 3a) shows the particles are spheres that are highly agglomerated, and the size of the sphere is approximately 40 nm. It can also be seen that a smaller particle with irregular size adhere to the sphere. Fig. 3b displays the lattice fringes of the smaller particle and sphere. The interplanar distances observed in the smaller particle and the sphere are 0.254 and 0.365 nm, which correspond to the 311- and 200-planes of tetragonal CuFe₂O₄ and monoclinic WO₃, respectively. It is proved that CuFe₂O₄ nanoparticles are attaching to the WO₃ to form CuFe₂O₄/WO₃ composite.

Fig. 3 (a) Transmission electron micrographs and (b) high resolution TEM (HRTEM) image of CuFe₂O₄/WO₃.

The presence of CuFe₂O₄ on the surface of WO₃ film was further confirmed by the XPS spectra (Fig. 4). In Fig. 4a, the survey spectrum of WO₃ and CuFe₂O₄/WO₃ indicates peaks of elements W and O. The carbon peak is attributed to adventitious hydrocarbon from the XPS instrument itself, which is not indicated in the figure. Fig. 3b shows the XPS spectrum of the tungsten core level (W 4f). The two peaks at 35.47 and 37.62 eV in the spectrum can be assigned to W 4f_7/2 and W 4f_5/2, respectively, which are consistent with the reported values.³⁶ Fig. 4c and d show the high-resolution XPS spectra for Fe 2p and Cu 2p, respectively. From the Fe 2p spectrum, it was found that the peaks at 711.2 and 724.7 eV are from Fe 2p3/2 and Fe 2p1/2, and a accompanying satellite peak visible at binding energie of around 718.4 eV is indicative of the presence of Fe³⁺ cations.³⁷ Two distinct and intense peaks were appeared for the Cu 2p spectrum, which are Cu 2p3/2 = 932.6 eV and Cu 2p1/2 = 952.4 eV (ΔBE = 19.8 eV).³⁸ A satellite peak at about 942.0 eV can also be seen, which provides evidence for the presence of Cu²⁺.³⁷ The ICP analysis was employed to determine the concentration of various metals (Cu, Fe and W), and [Cu] : [Fe] : [W] value of 1.51 : 3.12 : 100 was estimated, so the ratio of p semiconductor to n semiconductor is about 1.51 : 100 in this system.

Fig. 4 XPS of CuFe₂O₄/WO₃ (a) full spectrum; (b) W 4f; (c) Fe 2p and (d) Cu 2p.

The UV-visible spectra of pure WO₃ and CuFe₂O₄/WO₃ films are presented in Fig. 5. The bare WO₃ film absorbs light with wavelengths smaller than 455 nm, corresponding to 2.72 eV of band gap energy.³⁹ The absorbance edge of CuFe₂O₄/WO₃ is 463 nm, corresponding to 2.68 eV of band gap energy. The slightly shift of absorbance edge is probably due to the higher absorption coefficient of CuFe₂O₄ for visible light (Fig. S3†). To investigate the flat band potential (V_fb) of samples, Mott–Schottky plots were employed with a frequency of 500 Hz as shown in Fig. 6. The value of V_fb is approximately equal to the value of x-intercept, which are 0.08 and 0.19 V (vs. Ag/AgCl) for WO₃ and CuFe₂O₄/WO₃, respectively. It can be found that a positive shift of flat band potential, and a shift of the Fermi level to positive potential. It is because that CuFe₂O₄ is a p-type semiconductor with a negative slope, whose flat band potential is about 0.5 V (vs. Ag/AgCl) (Fig. S4†).³¹ Moreover, the potential of Fermi levels of WO₃ and CuFe₂O₄ is in equilibrium after forming the p–n heterojunction.²⁴ So the CuFe₂O₄/WO₃ has a positive flat band potential compared with pure WO₃ film.

Fig. 5 UV-vis absorbance spectroscopy of the WO₃ and CuFe₂O₄/WO₃.

Fig. 6 Mott–Schottky plots of the WO₃ and CuFe₂O₄/WO₃.

To further study the effect of the valence band maximum (VBM) of the WO₃ and CuFe₂O₄/WO₃, the valence band X-ray photoelectron spectroscopy (VB XPS) was employed to determine the electronic structure. Fig. 7 indicates that the VBM of CuFe₂O₄/WO₃ moves to higher potential compared with that of WO₃. That is to say, the VBM of the WO₃ changes after coupling with CuFe₂O₄, in agreement with the changing of flat band potential.

Fig. 7 VB-XPS spectra of WO₃ and CuFe₂O₄/WO₃.

To investigate the photoresponsivity of the films, the photocurrent–potential was measured using a linear sweep voltammetry method as shown in Fig. 8. The current values of two samples are about zero under dark condition. In the range of 0.3–1.0 V (vs. Ag/AgCl), a significant improvement of photocurrent under irradiation indicates the enhanced PEC activity after loading of CuFe₂O₄. A photocurrent of 0.75 mA cm⁻² (0.6 V vs. Ag/AgCl) was achieved with CuFe₂O₄/WO₃, which is 1.68 times higher than that of bare WO₃ (0.28 mA cm⁻²), the conversion efficiency is about 0.48% (Fig. S5†), and it is higher than the results reported in the similar system.^35,40 For a semiconductor thin film photoanode, the magnitude of photocurrent is determined by the competition of photo-generated carriers migration and recombination.⁴¹ At the same time, most of the photogenerated electron/hole pairs recombine generating heat and/or light after excitation.⁴² In fact, p–n junction can greatly enhance the charge separation owing to the internal electric field.⁴³ Hence, the p–n heterojunction formed by n-type WO₃ and p-type CuFe₂O₄ has a beneficial effect on the separation of the photo-generated carriers and decreasing the recombination of electrons and holes. In addition, the onset potential of CuFe₂O₄/WO₃ (Fig. S6†) shows a positive shift compared with pure WO₃, which is consistent with the results of flat band potentials in Fig. 6a. To observe the effect of composition (ratio of p/n) on the photoelectrochemical performance, various composite films were fabricated by changing the concentration of Cu(CH₃COO)₂·H₂O and K₃Fe(CN)₆ (the details in the ESI†), and the photocurrent curves are shown in Fig. S7.† It can be concluded that the ratio of p/n ≈ 1.51 : 100 might be the optimized composition in our system. Transient photocurrent were also measured at 0.4 V (vs. Ag/AgCl) with chopped light, as shown in Fig. 6b. The photocurrent of both samples are about zero when the light is chopped, and the photocurrent show decay mode under irradiation, meaning that the recombination process occurs. The transient time constant (τ_tran) were calculated from the equation: exp(−t/τ_tran) = (I_t − I_f)/(I_i − I_f).^44,45 It is 3.5 s for WO₃ and 4.5 s for CuFe₂O₄/WO₃, respectively, indicating the p–n heterojunction accelerates the separation of photoinduced electrons and holes. The stability of WO₃ and CuFe₂O₄/WO₃ were also investigated, as shown in Fig. S8.† After irradiation of 1140 s, the photocurrents of WO₃ and CuFe₂O₄/WO₃ decrease with 64.1% (from 0.100 to 0.036 mA cm⁻²) and 56.2% (from 0.442 to 0.194 mA cm⁻²), respectively. It means that the CuFe₂O₄/WO₃ film shows a little better stability compared to the bare WO₃ film.

Fig. 8 (a) Linear sweep voltammograms of WO₃ and CuFe₂O₄/WO₃, (b) transient photocurrent of WO₃ and CuFe₂O₄/WO₃.

EIS measurement was employed to study the electron transfer resistance across the electrode–electrolyte interfaces under visible irradiation. Fig. 9a shows the Nyquist plots which can be used to characterize the charge transfer resistance and the separation efficiency of the photo-generated electrons and holes.⁴⁵ The impedance spectra, consisting of a semicircle, can be fitted with an appropriate equivalent circuit. As shown in Fig. 9b, the equivalent circuit includes a series resistance (R_s), a charge transfer resistance (R_ct) and a constant phase element (CPE). The semicircle of CuFe₂O₄/WO₃ is smaller than that of WO₃, and the fitting results of R_s are 2479 and 364 ohm for WO₃ and CuFe₂O₄/WO₃, respectively. It means that the p–n heterojunction decrease the charge transfer resistance at the interface of electrode and electrolyte, leading to more efficient charge separation. Correspondingly, the characteristic frequency peaks for WO₃ and CuFe₂O₄/WO₃ in Bode phase plots are shown in Fig. 9c. The peak of CuFe₂O₄/WO₃ shifts to low frequency from 53.5 Hz to 10.0 Hz compared with WO₃. Therefore, the lifetime of photoelectrons (τ) for the films can be estimated by the equation: τ = 1/2πf_max,⁴⁵ where f_max is the peak frequency, and the τ are 3.0 and 15.9 ms for WO₃ and CuFe₂O₄/WO₃, respectively. The presence of p–n heterojunction facilitates the separation of photoinduced electrons and holes, leading to prolongation of photoelectrons lifetime, resulting in a significant improvement in PEC performance.

Fig. 9 EIS plots of the WO₃ and CuFe₂O₄/WO₃: (a) Nyquist plots, (b) equivalent circuit and (c) Bode plots, the inset of (a) is the full plot of WO₃.

In order to further explore the quantitative correlation of light absorption on the films, IPCE measurements were performed at the bias of 1.0 V (Fig. 10). The light response region of WO₃ and CuFe₂O₄/WO₃ are both at 310–475 nm. The CuFe₂O₄/WO₃ film shows a higher IPCE value compared with WO₃ film in the whole photoresponse region. Moreover, the maximum IPCE value of CuFe₂O₄/WO₃ is 47.27% at 350 nm, which is 1.21 times higher than that of WO₃. As the number of electrons in the external circuit produced by an incident photon at a given wavelength divided by the number of incident photons, IPCE is determined by three microscopic processes: light harvesting (LH), separation of opposing charges (CS), and a collection of the charges (CC) at the electrodes. Considering the similar light absorbance of two samples as shown in Fig. 5, the different value of IPCE at the whole light response region indicates that the separation of opposing charges is the main factor. So the improvement of photoelectrochemical performance is ascribed to the presence of p–n heterojunction.

Fig. 10 Incident photo to current conversion efficiency of WO₃ and CuFe₂O₄/WO₃.

Based on the above results, a simple mechanism of the conduction and valence band positions for CuFe₂O₄/WO₃ is shown in Fig. 11. The space charge region caused by p–n heterojunction has a beneficial effect of improving the charges separation.²² In this way, photo-generated holes transfer to CuFe₂O₄, and more photo-generated electrons can transfer to conductive substrates without recombination. Therefore, it can be concluded that the p–n heterojunction promotes electron/hole separation, leading to an enhanced photoelectrochemical performance.

Fig. 11 Energy band diagram of hetero-system FTO/WO₃/CuFe₂O₄.

4. Conclusion

In summary, the CuFe₂O₄/WO₃ p–n heterojunction film was synthesized and deposited on the FTO substrate by a doctor-blade method and a deposition-annealing method. The results of TEM, SEM and XPS show that CuFe₂O₄ nanoparticles were deposited on the surface of WO₃ film. Mott–Schottky and electrochemical impedance spectroscopy analysis confirmed the formation of p–n heterojunction, and CuFe₂O₄/WO₃ heterojunction film shows an enhanced photoelectrochemical properties compared to WO₃ film. The presence of p–n heterojunction facilitates the separation of photoinduced electrons and holes, leading to more efficient charge transfer, resulting in a significant improvement in PEC performance.

Acknowledgements

This study was supported by the National Nature Science Foundation of China (No. 51304253), the Fundamental Research Funds for the Central Universities of Central South University (2015zzts021), and Y. L. acknowledges financial support from the China Scholarship Council (CSC File No. 201406370157).

References

1. A. Fujishima and K. Honda, Nature, 1972, 238, 37–38.
2. K. P. S. Parmar, H. J. Kang, A. Bist, P. Dua, J. S. Jang and J. S. Lee, ChemSusChem, 2012, 5, 1926–1934.
3. T. M. Breault and B. M. Bartlett, J. Phys. Chem. C, 2013, 117, 8611–8618.
4. Z.-F. Huang, J. Song, L. Pan, X. Zhang, L. Wang and J.-J. Zou, Adv. Mater., 2015, 27, 5309–5327.
5. T. Zhang, Z. Zhu, H. Chen, Y. Bai, S. Xiao, X. Zheng, Q. Xue and S. Yang, Nanoscale, 2015, 7, 2933–2940.
6. Y. Sun, C. J. Murphy, K. R. Reyes-Gil, E. A. Reyes-Garcia, J. M. Thornton, N. A. Morris and D. Raftery, Int. J. Hydrogen Energy, 2009, 34, 8476–8484.
7. Y. Liu, W. Li, J. Li, Y. Yang and Q. Chen, RSC Adv., 2014, 4, 3219–3225.
8. S. Ghosh, S. S. Acharyya, M. Kumar and R. Bal, RSC Adv., 2015, 5, 37610–37616.
9. A. Tanaka, K. Hashimoto and H. Kominami, J. Am. Chem. Soc., 2013, 136, 586–589.
10. B. Cole, B. Marsen, E. Miller, Y. Yan, B. To, K. Jones and M. Al-Jassim, J. Phys. Chem. C, 2008, 112, 5213–5220.
11. Y. Liu, J. Li, W. Li, Y. Yang, Y. Li and Q. Chen, J. Phys. Chem. C, 2015, 119, 14834–14842.
12. M. Khajeh Aminian and M. Hakimi, Catal. Sci. Technol., 2014, 4, 657–664.
13. L. Huang, H. Xu, Y. Li, H. Li, X. Cheng, J. Xia, Y. Xu and G. Cai, Dalton Trans., 2013, 42, 8606–8616.
14. Z. Liu, Z.-G. Zhao and M. Miyauchi, J. Phys. Chem. C, 2009, 113, 17132–17137.
15. J. Su, L. Guo, N. Bao and C. A. Grimes, Nano Lett., 2011, 11, 1928–1933.
16. Y. Liu, H. He, J. Li, W. Li, Y. Yang, Y. Li and Q. Chen, RSC Adv., 2015, 5, 46928–46934.
17. R. Saito, Y. Miseki and K. Sayama, Chem. Commun., 2012, 48, 3833–3835.
18. K. C. Leonard, K. M. Nam, H. C. Lee, S. H. Kang, H. S. Park and A. J. Bard, J. Phys. Chem. C, 2013, 117, 15901–15910.
19. H. Yang, J. Yan, Z. Lu, X. Cheng and Y. Tang, J. Alloys Compd., 2009, 476, 715–719.
20. R. Ding, L. Lv, L. Qi, M. Jia and H. Wang, RSC Adv., 2014, 4, 1754–1760.
21. J. H. Kim, J. H. Kim, J.-W. Jang, J. Y. Kim, S. H. Choi, G. Magesh, J. Lee and J. S. Lee, Adv. Energy Mater., 2015, 5(6) DOI:10.1002/aenm.201401933.
22. N. Nasrallah, M. Kebir, Z. Koudri and M. Trari, J. Hazard. Mater., 2011, 185, 1398–1404.
23. C. Karunakaran, S. SakthiRaadha, P. Gomathisankar and P. Vinayagamoorthy, RSC Adv., 2013, 3, 16728–16738.
24. A. Kezzim, N. Nasrallah, A. Abdi and M. Trari, Energy Convers. Manage., 2011, 52, 2800–2806.
25. M. Wang, L. Sun, J. Cai, P. Huang, Y. Su and C. Lin, J. Mater. Chem. A, 2013, 1, 12082–12087.
26. J. Cao, J. Xing, Y. Zhang, H. Tong, Y. Bi, T. Kako, M. Takeguchi and J. Ye, Langmuir, 2013, 29, 3116–3124.
27. L. J. Minggu, K. H. Ng, H. A. Kadir and M. B. Kassim, Ceram. Int., 2014, 40, 16015–16021.
28. J. Hou, C. Yang, H. Cheng, S. Jiao, O. Takeda and H. Zhu, Energy Environ. Sci., 2014, 7, 3758–3768.
29. M. Wang, L. Sun, Z. Lin, J. Cai, K. Xie and C. Lin, Energy Environ. Sci., 2013, 6, 1211–1220.
30. Z. Kang, X. Yan, Y. Wang, Z. Bai, Y. Liu, Z. Zhang, P. Lin, X. Zhang, H. Yuan, X. Zhang and Y. Zhang, Sci. Rep., 2015, 5, 7882.
31. R. Rajendran, Z. Yaakob, M. A. Mat Teridi, M. S. Abd Rahaman and K. Sopian, Mater. Lett., 2014, 133, 123–126.
32. A. Kargar, K. Sun, Y. Jing, C. Choi, H. Jeong, Y. Zhou, K. Madsen, P. Naughton, S. Jin, G. Y. Jung and D. Wang, Nano Lett., 2013, 13, 3017–3022.
33. J. A. Seabold, K. Shankar, R. H. T. Wilke, M. Paulose, O. K. Varghese, C. A. Grimes and K.-S. Choi, Chem. Mater., 2008, 20, 5266–5273.
34. G. Dai, J. Yu and G. Liu, J. Phys. Chem. C, 2011, 115, 7339–7346.
35. Y. Liu, J. Li, W. Li, Q. Liu, Y. Yang, Y. Li and Q. Chen, Int. J. Hydrogen Energy, 2015, 40, 8856–8863.
36. S. Cong, Y. Tian, Q. Li, Z. Zhao and F. Geng, Adv. Mater., 2014, 26, 4260–4267.
37. Y. Zhao, G. He, W. Dai and H. Chen, Ind. Eng. Chem. Res., 2014, 53, 12566–12574.
38. C. Reitz, C. Suchomski, J. Haetge, T. Leichtweiss, Z. Jaglicic, I. Djerdj and T. Brezesinski, Chem. Commun., 2012, 48, 4471–4473.
39. M. M. Momeni and Y. Ghayeb, J. Electroanal. Chem., 2015, 751, 43–48.
40. J. Zhu, W. Li, J. Li, Y. Li, H. Hu and Y. Yang, Electrochim. Acta, 2013, 112, 191–198.
41. Q. Liu, J. He, T. Yao, Z. Sun, W. Cheng, S. He, Y. Xie, Y. Peng, H. Cheng, Y. Sun, Y. Jiang, F. Hu, Z. Xie, W. Yan, Z. Pan, Z. Wu and S. Wei, Nat. Commun., 2014, 5, 5122.
42. J. Schneider, M. Matsuoka, M. Takeuchi, J. Zhang, Y. Horiuchi, M. Anpo and D. W. Bahnemann, Chem. Rev., 2014, 114, 9919–9986.
43. K. Zhang, X.-J. Shi, J. K. Kim and J. H. Park, Phys. Chem. Chem. Phys., 2012, 14, 11119–11124.
44. M. Ibadurrohman and K. Hellgardt, Int. J. Hydrogen Energy, 2014, 39, 18204–18215.
45. K. Chen, X. Feng, R. Hu, Y. Li, K. Xie, Y. Li and H. Gu, J. Alloys Compd., 2013, 554, 72–79.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra18279d

---