Surface modification of LaFeO₃ by Co-Pi electrochemical deposition as an efficient photoanode under visible light

Abstract

LaFeO₃ is a promising visible light photocatalyst due to its favorable band gap and excellent stability in aqueous solution. The cathodic photocurrent for p-type characteristics reversing to anodic photocurrent on a LaFeO₃ photoanode is observed under visible light (>420 nm) and the anodic photoelectrochemical water oxidation performance is improved by electro-deposition of amorphous cobalt-phosphate (Co-Pi). It shows that the presence of Co-Pi down-shifts the onset potential by ∼560 mV for anodic photocurrent, and the improvement can be attributed to enhanced water oxidation due to the Co^II/Co^III-OH discharge in Co-Pi decorated layers by cyclic voltammetry test. The transition anodic photocurrent is improved by about six times after Co-Pi coating under visible light at 0.50 V vs. Ag/AgCl and the electrochemical impedance spectroscopy certifies the enhanced charge transfer, which contributes to the meliorative anodic photocurrent. The suppression of the photogenerated electron–hole recombination after Co-Pi coating on LaFeO₃ photoanode is directly demonstrated by the reduced photoluminescence spectrum. Combined with the accelerated carrier consumption on the surface and enhanced carrier separation on the photoanode, the incident photon-to-current conversion efficiencies of LaFeO₃ can be promoted from 1.37% to 2.14% at 400 nm owing to the presence of the Co-Pi layer.

---

Introduction

With the excessive consumption of fossil fuels and environmental deterioration, photocatalytic hydrogen generation from water is a promising way to explore renewable energy to maintain sustainable development.^1–3 Perovskite ABO₃ structures provide the basic stable framework to be modified in both crystal and electronic structures, and have found extensive applications in photocatalysis, solar cells and photoelectrochemical (PEC) water splitting due to their unique photo-response and photochemical activity.^4–9 Some perovskites such as SrTiO₃,⁶ NaTaO₃,⁷ and AgNbO₃,⁸ have be reported as photocatalysts for water splitting due to their high catalytic activity and good durability in aqueous solution. However, most perovskites show poor photon harvesting performance and are limited to utilize the ultraviolet (UV) light only, which is mainly attributed to their large band gap. Theoretical study suggests that charge densities of layered perovskites at the valence band maximum and conduction band minimum are located on nearly 180 degree B–O–B atom chains, which can provide transport channels for carriers. Moreover, they conclude that the partially substitution of B site atom is probable to achieve the visible light adsorption.¹⁰

Among most perovskites, LaFeO₃ is an intrinsically promising visible light photocatalyst for water splitting with suitable band gap of 2.07 eV.^11,12 Parida et al. found that LaFeO₃ powders could generate hydrogen and oxygen simultaneously under visible light and the apparent quantum efficiency reached to 8.07%.¹³ Sora et al. obtained the photocurrent density of 0.8 mA cm⁻² beyond 1 V vs. Ag/AgCl under simulated solar illumination when they used the LaFeO₃ films as a photoanode.¹⁴ Ultrathin LaFeO₃ films grown on Nb:SrTiO₃ exhibit thickness-dependence with sensitivity to less than 10 nm in both the charge transfer and the potential-dependent photo-response.¹⁵ Interestingly, Yu¹⁶ et al. prepared the LaFeO₃ photocathode by pulse laser deposition with p-type characteristic and achieved a stoichiometric evolution of H₂ and O₂ when it combined with α-Fe₂O₃ photoanode. However, it is revealed that the poor charge transport and slow reaction kinetics at the surface are the main obstacles restricting the photocatalytic performance of LaFeO₃ for water splitting.^13,14,17,19 In order to improve the charge transport and transfer, transition metals were doped into the LaFeO₃ photoanode and the Cu doping yielded the optimized PEC performance.^17,18 On the other way, the Pt decoration on the LaFeO₃ powder could promote the generation rate of hydrogen and boost to 3315 μmol g⁻¹ h⁻¹ under visible light.¹⁹ Other metal-oxides such as RuO₂,²⁰ IrO₂ (ref. 21) are also taken as a coating layer to reduce the oxygen evolution kinetics on the surface of photocatalysts. However, those materials are quite expensive. Low-cost and practical co-catalysts are desired to improve the surface chemical kinetics of LaFeO₃. An amorphous cobalt-phosphate (Co-Pi) based material, has been reported to exhibit low over potential for oxygen evolution reaction (OER) in the phosphate buffer solution at pH 7–8.²² The Co-Pi is thought to provide the oxygen–oxygen bond coupling which services as the active site to accelerate the carriers consumption in a chemical turnover-limiting process at the photoanode surface.²³ Hamann et al. deposited Co-Pi on porous α-Fe₂O₃ films and the onset potential for photocurrent negatively shifts due to the improvement of the surface chemical activity.²⁴ Other metal oxides such as ZnO²⁵ and WO₃,²⁶ also have achieved similar PEC performance promotions upon the Co-Pi coating.

Here we report that electro-deposition of Co-Pi on LaFeO₃ photoanode to get the improvement of the PEC performance. We observe an interesting phenomenon in pristine and coated LaFeO₃ photoanodes that the cathodic photocurrent with p-type characteristic appears in the negative potential regions and it reverses to anodic photocurrent in the positive regions with n-type characteristic. The onset potential for anodic photocurrent exhibits significantly shift from 0.23 V to −0.33 V vs. Ag/AgCl after Co-Pi coating due to the modification of the surface chemical activity. Six times of anodic photocurrent is achieved for Co-Pi coated LaFeO₃ at 0.50 V vs. Ag/AgCl under visible light and it should be attributed to the superior photogenerated electron–hole separation by electrochemical impedance spectrum (EIS) and photoluminescence (PL) spectrum. Our work displays both the cathodic and anodic photocurrent on LaFeO₃ and improves its solar oxygen evolution performance by a low-cost Co-Pi layer, which is instructive for the design of high efficient LaFeO₃-based PEC cells under visible light.

Experimental

Preparation of LaFeO₃ photoanode

LaFeO₃ nanoparticles were prepared by sol–gel method.¹³ All chemicals used were the analytical-grade reagents without any further purification. In detail, La(NO₃)₃·6H₂O and Fe(NO₃)₃·9H₂O with stoichiometric ratios were dissolved in ∼150 mL deionized water (18 MΩ) obtaining yellow transparent solution and kept stirring until the salts were dissolved completely. Citric acid (molar ratio of citric acid/metals = 2) was added into the solution and the resulting solution was concentrated with continuous stirring at 85 °C until a dark yellow sol was obtained. After that, the sol was transferred to the oven and kept at 120 °C overnight to obtain the dry ground gel, which was grinded completely and then calcined at 500 °C for 2 h to obtain the crystalline nanoparticles.

The slurry electrodes were fabricated by using doctor blade method. The crystalline LaFeO₃ powder was mixed with ethanol and then treated ultrasonically for several minutes until the powders dispersed completely. The slurry was obtained by continuous grinding the resulting mixtures after adding PEG 400 for 30 min. The fluorine doped tin oxide (FTO) glass was firstly ultrasonically cleaned in sequence of acetone, ethanol and deionized water for 10 min respectively before pasting. After pasted by doctor blade method, the slurry electrodes were transferred to the vacuum oven and kept at 120 °C overnight.

Electrochemical deposition of Co-Pi

Co-Pi was electrodeposited on LaFeO₃ photoelectrode by submerging into a 0.5 mM Co^II containing solution with 0.1 M potassium phosphate (KPi) buffer of pH ∼ 7. A Pt foil was worked as the counter electrode and Ag/AgCl was used as the reference electrode. The samples with Co-Pi deposited at a constant bias of +1.10 V vs. Ag/AgCl for 1, 3 and 5 hours are denoted as LaFeO₃/Co-Pi 1, LaFeO₃/Co-Pi 3, and LaFeO₃/Co-Pi 5, respectively.

Characterizations of materials

The X-ray diffraction (XRD) patterns were recorded on an analytical X'pert PRO diffractometer operating at 40 kV and 40 mA, using Cu Kα radiation with λ = 1.5406 Å in the 2θ ranging from 20° to 80°. Scanning electron microscope (SEM) images and Energy Dispersive X-ray analysis (EDX) were obtained by using a JEM 2100 microscope operated at 20 kV. X-ray Photoelectron Spectroscopy (XPS) studies were characterized on VG Multilab 2000 X-ray Photoelectron Spectroscope with an Al Kα radiation. The starting angle of the photoelectron was set at 90°. UV-vis absorbance spectra were characterized on LabRAM HR800 UV-vis spectrophotometer with scanning wavelength ranging from 200 nm to 900 nm using BaSO₄ as the reference. PL spectra were characterized on LabRAM HR800 Raman spectrometer with scanning wavelength ranging from 532 nm to 800 nm.

Electrochemical and PEC measurements

The pristine and Co-Pi coated LaFeO₃ photoanodes were fabricated by soldering a copper wire onto a bare part of the FTO substrate. The substrate edges and the metal contact region were sealed with insulating epoxy resin. Unless otherwise stated, all experiments were performed at ambient temperature (20 ± 1 °C) and with continues N₂ flowing. The photocurrent densities were evaluated by measuring the electrochemical performance of current density via liner sweep voltammetry (LSV) under chopped visible light illumination (>420 nm). The electrochemical signal was recorded using Autolab Electrochemical Workstation in a standard three-electrode electrochemical system. The slurry electrodes acted as the working electrode, Pt as the counter electrode, and saturated Ag/AgCl as the reference electrode. 0.1 M KPi solution (pH ∼ 8) was used as the electrolyte. The visible light was obtained by a full-spectra simulated sunlight from 500 W xenon lamps (CHF-XM500, Trusttech) with UV filters (>420 nm). The illumination intensity was calibrated to be 100 mW cm⁻² by Sanpometer SM206 solar power meter. The area of exposed FTO was set to be 1 cm² and the rest part was blocked by insulation tapes. The LSV of the current density under chopped visible light illumination ranged from −0.5 V to 0.4 V vs. Ag/AgCl with a scan rate of 5 mV s⁻¹. The cyclic voltammetry (CV) and EIS plots were noted under the same conditions. The incident photon-to-current conversion efficiencies (IPCE) were carried out on the CrownTech QTest Station500AD arranging from 400–650 nm in the standard three-electrode electrochemical system.

Results and discussion

SEM image of pristine LaFeO₃ slurry electrodes pasted on the FTO glass is shown in Fig. 1(a). A porous structure is obviously observed on LaFeO₃ film which should be attributed to the gas release during the calcination of the slurry electrode. The porous sizes are irregular and the surface morphology exhibits to be the coral island-shape consisted with the LaFeO₃ nanoparticles (Fig. S1(a) in ESI†). SEM images of Co-Pi coated LaFeO₃ electrodes with different deposition times are shown in Fig. 1(b)–(d) and the porous structures in the composites become inconspicuous. Upon more Co-Pi deposited, Fig. 1(d) shows us that the pore structure diminishes and the surface of LaFeO₃ is covered by Co-Pi (Fig. S1(c) and (d) in ESI†), similar with that in previous study.²²

Fig. 1 SEM images and corresponding EDX characterizations of (a) LaFeO₃, (b) LaFeO₃/Co-Pi 1, (c) LaFeO₃/Co-Pi 3 and (d) LaFeO₃/Co-Pi 5 photoelectrodes.

To confirm the existence of Co-Pi, EDX spectra (inset in Fig. 1) have been used to characterize the components of the corresponding samples. The stoichiometric of La/Fe/O is calculated about 1.00 : 0.85 : 2.97 from EDX spectrum as shown in the inset of Fig. 1(a). Two other peaks appear at 2.05 and 7.10 keV in the insets of Fig. 1(b)–(d), which correspond to the coated phosphorus and cobalt, confirming the existence of compound Co-Pi. It is noticed that the peaks of cobalt/phosphorus (Co/P) become stronger with the increase of electrochemical deposition time and the corresponding Co and P atomic concentrations are shown in Fig. S2 in ESI.† The ratios of Co : P change with deposition time rather than keep constant, indicating that the growth of Co and P is amorphous (Fig. S3 in ESI†) in the electrochemical deposition process. The Co : P ratios are about 1.26, 1.76 and 1.95 for these Co-Pi coated LaFeO₃ samples in the insets of Fig. 1(b)–(d), which are reasonable according to previous reported results.²²

XPS analysis is conducted to gain insight into the elemental composition and the binding states. The full XPS spectrum of LaFeO₃/Co-Pi 3 is shown in Fig. 2(a) and the chemical states of La, Fe, O, Co, P and K are well observed. The high-resolution Co 2p1/2 peak at 796.00 and Co 2p3/2 peak at 780.65 eV correspond to the typical Co^II or Co^III bound to oxygen.^22,25 Further, the P 2p peak at 133.20 eV in Fig. 2(c) is identified as phosphate in the electrodeposited Co-Pi, which confirms the existence of Co-Pi on the LaFeO₃ slurry electrode.

Fig. 2 XPS characterizations of LaFeO₃/Co-Pi 3 photoelectrode. (a) The full spectrum region and the corresponding peaks are identified in the figure, the spectra of (b) Co 2p and (c) P 2p.

LSV is measured to study the catalytic nature of electrochemical OER for the pristine and Co-Pi coated LaFeO₃ photoanodes in the dark (Fig. 3(a)). The onset potentials with an anodic current of 0.5 mA cm⁻² are 1.28, 1.21, 1.10 and 1.04 V for pristine and Co-Pi coated LaFeO₃ electrodes, respectively. It is clear that the onset potential for oxygen evolution exhibits negative shift by ∼240 mV, indicating the enhanced electrochemical catalytic activity introduced by Co-Pi. CV study is conducted to clarify the enhanced reaction kinetics in Fig. 3(b). Obvious anodic and cathodic peaks are observed for Co-Pi coated electrodes and the peaks at 0.47 V and −0.32 V correspond to the process of Co^II/Co^III redox, which agrees well with the previous study.^25,27,28 It is thought that the enhanced reaction kinetics in the water oxidation process should be attributed to the Co^II/Co^III-OH redox at the surface of Co-Pi.²⁹ We notice that the peak intensities at 0.47 V get increased with the increase of coated quantities of Co-Pi, in good coincidence with the order of the shift of the onset potentials.

Fig. 3 (a) LSV of pristine and Co-Pi coated LaFeO₃ photoelectrode measured in 0.1 M KPi buffer (pH ∼ 8) with a scan rate of 10 mV s⁻¹ in the dark. (b) CV of pristine and Co-Pi coated LaFeO₃ photoelectrode in 0.1 M KPi pH ∼ 8 ranged from −0.5 V to 0.8 V.

PEC performance of LaFeO₃ related photoanodes is investigated by comparing the photocurrent–potential characteristics of LSV measurement with chopped visible light illumination (>420 nm) (Fig. 4). An interesting cathodic photocurrent with p-type characteristic is observed in the negative potential regions (less than −0.33 V vs. Ag/AgCl) under visible light. Several investigations about the LaFeO₃ coating layers exhibit similar cathodic photocurrent,^16,30 which implies LaFeO₃ could be potential photocathode. The photocurrent turns to be anodic with n-type characteristic in the positive regions (larger than 0.23 V vs. Ag/AgCl). The corresponding onset potential of the anodic photocurrent for LaFeO₃/Co-Pi 1, LaFeO₃/Co-Pi 3 and LaFeO₃/Co-Pi 5 at the threshold photocurrent of 0.1 μA cm⁻² are 0.14 V, 0.06 V and −0.33 V, respectively. Compared with LaFeO₃ photoanode, the onset potential for the anodic photocurrent of LaFeO₃/Co-Pi 5 under visible light exhibits significantly negative shift of ∼560 mV, which is much larger than that in the dark (∼240 mV).

Fig. 4 LSV of (a) LaFeO₃, (b) LaFeO₃/Co-Pi 1, (c) LaFeO₃/Co-Pi 3 and (d) LaFeO₃/Co-Pi 5 photoelectrode measured in 0.1 M KPi buffer (pH ∼ 8) with a scan rate of 5 mV s⁻¹ under chopped visible light (λ > 420 nm, 100 mW cm⁻²) illumination.

To get more details of the photocurrent, the chronoamperometric measurements of the photoelectrodes at constant voltage of −0.50 V, −0.25 V, 0.25 V and 0.50 V with 30 s cyclic visible light illuminations are characterized as shown in Fig. 5(a)–(d), respectively. We find all the photoelectrodes exhibit cathodic photocurrent at −0.50 V potential and the cathodic densities tend to decrease with more Co-Pi deposited. At −0.25 V, similar reduced cathodic photocurrent are observed for Co-Pi coated ones. Especially for LaFeO₃/Co-Pi 5, the cathodic photocurrent are suppressed and turn to be anodic photocurrent at such a potential (Fig. 5(b)). Note that, LaFeO₃ could both generate H₂ and O₂ with a band gap ∼ 2.07 eV which implies its band structure staggered across the redox potential for water splitting. The cathodic or anodic photocurrent indicates the flow direction of the major carriers (electrons or holes) between the photoanode and electrolyte. The carriers flow is mostly determined by the band bending behavior at the photoanode/electrolyte surface with respect to the external bias (V_e). Flat band potential (V_fb) is the special V_e which makes the photoanode band equilibrium to the Fermi level of the bulk LaFeO₃. V_fb is calculated by Mott–Schottky plots for LaFeO₃ and Co-Pi coated electrode in Fig. 6 and all the plots exhibit the positive slopes which imply the n-type characteristics of the LaFeO₃ related photoanodes. Moreover, we also find negative slope for pristine LaFeO₃ photoanode with V_fb ∼ 0.43 V, indicating the possible p-type characteristic of LaFeO₃. It is found the V_fb of the LaFeO₃, LaFeO₃/Co-Pi 1, 3 and 5 to be 0.0 eV, −0.06 eV, −0.11 eV and −0.28 eV, respectively. To understand the cathodic reversing of LaFeO₃/Co-Pi samples, we take the Fig. 5(b) as an example where the V_e is kept constantly at −0.25 V. By comparison, only the V_fb of LaFeO₃/Co-Pi 5 is lower than V_e. If the value ΔV_fb = V_e − V_fb is defined as a descriptor to evaluate the direction and degree of band bending, only LaFeO₃/Co-Pi 5 gives a positive ΔV_fb compared to the other three ones (in Fig. 7) and shows a reverse band bending direction, which illustrates a reversing flow of the carriers. It is reasonable that LaFeO₃/Co-Pi 5 exhibits inverse anodic photocurrent. Since the cathodic photocurrent responses are associated with the generation of hydrogen from water involving the electron reaction in N₂ flow, holes of the LaFeO₃/Co-Pi 5 are involved in the OER reaction with an anodic photocurrent. As the V_e increases to positive (0.25 V and 0.50 V), the ΔV_fb of the four samples are always positive and only anodic photocurrent can be observed. Moreover, steady increases of the anodic photocurrent is shown in Fig. 5(c) and (d) with the coated quantity of Co-Pi. Especially, LaFeO₃/Co-Pi 5 exhibits six times enhancement of the photocurrent comparing to that of pristine LaFeO₃. Additionally, an interesting phenomenon that obvious decays and spikes of transient current are observed for cathodic photocurrent, while those decays and spikes disappear once the photocurrent turns to be anodic in Fig. 5. Co-Pi is known as a typical hole capturer that could trap and consume photogenerated holes of semiconductor which provides an intermediate of Co^IV and oxygen bridge (usually named as Co^IV-OXO) for the oxygen generation.^23,31,32 Since the cathodic photocurrent is linked to the hydrogen evolution, serious electron–hole recombination would occur on the holes-trapped photoelectrode when electrons are involved in the reaction, contributing to the sharp spikes. While the anodic photocurrent corresponds to the OER involving the holes reaction, the recombination of electron–hole on the photoelectrode can be suppressed and steady photocurrent is obtained due to presence of Co-Pi.

Fig. 5 Transient photocurrent response with 30 s cyclic visible light illumination at (a) −0.50 V, (b) −0.25 V, (c) 0.25 V and (d) 0.50 V of pristine and Co-Pi coated LaFeO₃ photoelectrode.

Fig. 6 Mott–Schottky plots of (a) LaFeO₃, (b) LaFeO₃/Co-Pi 1, (c) LaFeO₃/Co-Pi 3 and (d) LaFeO₃/Co-Pi 5 photoelectrodes with the frequency of 800 and 1000 Hz under the dark condition. The corresponding V_fb is fitted in each figures.

Fig. 7 Schematic of the band bending of LaFeO₃, LaFeO₃/Co-Pi 1, LaFeO₃/Co-Pi 3 and LaFeO₃/Co-Pi 5 at V_e of −0.25 V vs. Ag/AgCl (ΔV_fb = V_e − V_fb).

To investigate the charge transfer on the electrodes in the water oxidation process, EIS measurement is conducted with frequency ranging from 0.01 Hz to 100 kHz as shown in Fig. 8.^33,34 Based on the Nyquist plots of these samples in the dark (solid line) and under visible light (dash line), it is observed that the arc radius becomes smaller with the increase of the quantity of coated Co-Pi, demonstrating the reduction of the charge transfer resistance. On the other hand, the sample under visible light exhibits smaller arc radius than that in the dark, indicating a decrease of charge transfer under the illumination. To understand the carrier behaviours of pristine and Co-Pi coated photoanodes, PL spectrum is measured to study the photogenerated electron–hole recombination³⁵ and the result is shown in Fig. 9. The signal with the wavelength range from 550–800 nm could be assigned to the LaFeO₃ with a remarkable PL peak at about 700 nm as well as the Co-Pi coated ones. The peak intensities tend to be reduced with more Co-Pi coating, which indicates the suppressing of the photogenerated electron–hole recombination. This result is informative to clarify the role of Co-Pi coating on LaFeO₃ photoelectrode for improving the anodic PEC performance for water splitting under visible light.

Fig. 8 EIS Nyquist plots of pristine and Co-Pi coated LaFeO₃. The solid lines represent the samples measured in the dark condition and the dash lines are the responding samples measured under visible light illumination.

Fig. 9 PL spectra of LaFeO₃, LaFeO₃/Co-Pi 1, LaFeO₃/Co-Pi 3 and LaFeO₃/Co-Pi 5 photoelectrode.

IPCE of the pristine and Co-Pi coated LaFeO₃ photoelectrodes are measured at 0 V vs. Ag/AgCl in 0.1 M KPi of pH ∼ 8 in Fig. 10. The incident wavelength varies from 400 to 650 nm to avoid the interference effect of the UV light. The IPCE plot exhibits a good visible light response for both pristine and Co-Pi coated electrodes, which agrees well with the UV-vis for the pristine LaFeO₃ (1.37% at λ = 400 nm) in visible light range is similar with those previously reports.¹⁶ Upon Co-Pi coating, it is found the LaFeO₃/Co-Pi 1, 3 and 5 obtain improved IPCE of 1.42%, 1.85% and 2.14% at λ = 400 nm, respectively. The IPCE of LaFeO₃ coated by Co-Pi for a longer time is found to be poorer (Fig. S6 in ESI†) due to the introduced impurities in the coated layers and reduced light harvest with thicker coating layers. Our results imply that Co-Pi is a viable way to promote the photoconversion efficiency of LaFeO₃-based PEC cells under visible light.

Fig. 10 IPCE of LaFeO₃, LaFeO₃/Co-Pi 1, 3 and Co-Pi 5 measured at 0 V vs. Ag/AgCl in 0.1 M KPi of pH ∼8.

Conclusions

The photoelectrodes of Co-Pi coated LaFeO₃ are successfully fabricated by the electrochemical deposition method. It is found the onset potential of water oxidation negatively shift by about 240 mV in the dark at an anodic current of 0.5 mA cm⁻² in 0.1 M KPi with pH ∼ 8.0. The improved performance should be attributed to the enhanced water oxidation kinetics by the Co^II/Co^III-OH redox. The PEC performance has been measured under chopped visible light illumination (>420 nm) and the fabricated electrodes exhibit cathodic-to-anodic photocurrent reverse as the potential sweeps from negative to positive regions. The onset potential for anodic photocurrent under illumination shifts negatively by ∼560 mV for Co-Pi coated photoanode. Moreover, six times enhancement of the anodic photocurrent is observed on LaFeO₃/Co-Pi 5 photoelectrode. Based on the EIS and PL characterizations, the improved anodic photocurrent is attributed to the enhanced photogenerated electron–hole separation on Co-Pi coated photoelectrode. Furthermore, the IPCE of fabricated photoelectrode under visible light at 400 nm is promoted from 1.37% to 2.14% after Co-Pi coating. The understanding of the role of Co-Pi provides valuable guidance for the design of the LaFeO₃-base PEC cells for water splitting under visible light.

Acknowledgements

This work is supported by the National Basic Research Program of China (2013CB934800), National Natural Science Foundation of China (51302094, 51575217 and 51572097), the Hubei Province Funds for Distinguished Young Scientists (2014CFA018 and 2015CFA034), the State Key Laboratory of Digital Manufacturing Equipment and Technology Funding (DMET2015A01), as well as the Fundamental Research Funds for the Central Universities, HUST (2015QN009 and 2014TS037). Rong Chen acknowledges the Thousand Young Talents Plan, the Recruitment Program of Global Experts and Program for Changjiang Scholars and Innovative Research Team in University (No. IRT13017). We thank the technology support by the Analytic Testing Center of HUST for carrying out the XRD and UV-vis analysis and the equipment supports from AMETEK lab.

Notes and references

1. X. Lu, S. Xie, H. Yang, Y. Tong and H. Ji, Chem. Soc. Rev., 2014, 43, 7581–7593.
2. F. E. Osterloh, Chem. Soc. Rev., 2013, 42, 2294–2320.
3. Q. Huang, Z. Ye and X. Xiao, J. Mater. Chem. A, 2015, 3, 15824–15837.
4. Z. G. Yi and J. Ye, Appl. Phys. Lett., 2007, 91, 254108.
5. D. Wang, T. Kako and J. Ye, J. Am. Chem. Soc., 2008, 130, 2724–2725.
6. E. Guo and L. Yin, J. Mater. Chem. A, 2015, 3, 13390–13401.
7. X. Wu, S. Yin, B. Liu, M. Kobayashi, M. Kakihana and T. Sato, J. Mater. Chem. A, 2014, 2, 20832–20840.
8. G. Li, T. Kako, D. Wang, Z. Zou and J. Ye, Dalton Trans., 2009, 2423–2427.
9. I. E. Castelli, K. S. Thygesen and K. W. Jacobsen, J. Mater. Chem. A, 2015, 3, 12343–12349.
10. H. Chen and N. Umezawa, J. Appl. Phys., 2015, 117, 055106.
11. J. R. Hayes and A. P. Grosvenor, J. Phys.: Condens. Matter, 2011, 23, 465502.
12. I. R. Shein, K. I. Shein, V. L. Kozhevnikov and A. L. Ivanovskii, Phys. Solid State, 2005, 47, 2082.
13. K. M. Parida, K. H. Reddy, S. Martha, D. P. Das and N. Biswal, Int. J. Hydrogen Energy, 2010, 35, 12161–12168.
14. I. N. Sora, F. Fontana, R. Passalacqua, C. Ampelli, S. Perathoner, G. Centi, F. Parrino and L. Palmisano, Electrochim. Acta, 2013, 109, 710–715.
15. K. J. May, D. P. Fenning, T. Ming, W. T. Hong, D. Lee, K. A. Stoerzinger, M. D. Biegalski, A. M. Kolpak and Y. S. Horn, J. Phys. Chem. Lett., 2015, 6, 977–985.
16. Q. Yu, X. G. Meng, T. Wang, P. Li, L. Q. Liu, K. Chang, G. G. Liu and J. Ye, Chem. Commun., 2015, 51, 3630–3633.
17. Q. Peng, B. Shan, Y. Wen and R. Chen, Int. J. Hydrogen Energy, 2015, 40, 15423–15431.
18. J. Li, X. Pan, Y. Xu, L. Jia, X. Yi and W. Fang, Int. J. Hydrogen Energy, 2015, 40, 13918–13925.
19. S. N. Tijare, M. V. Joshi, P. S. Padole, P. A. Mangrulkar, S. S. Rayalu and N. K. Labhsetwar, Int. J. Hydrogen Energy, 2012, 37, 10451–10456.
20. E. Tsuji, A. Imanishi, K. Fukui and Y. Nakato, Electrochim. Acta, 2011, 56, 2009–2016.
21. S. Anantharaj, P. E. Karthikb and S. Kundu, J. Mater. Chem. A, 2015, 3, 24463–24478.
22. M. W. Kanan and D. G. Nocera, Science, 2008, 321, 1072–1075.
23. Y. Surendranath, M. W. Kanan and D. G. Nocera, J. Am. Chem. Soc., 2010, 132, 16501–16509.
24. B. Klahr, S. Gimenez, F. F. Santiago, J. Bisquert and T. W. Hamann, J. Am. Chem. Soc., 2012, 134, 16693–16700.
25. E. M. P. Steinmiller and K. S. Choi, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 20633–20636.
26. J. A. Seabold and K. S. Choi, Chem. Mater., 2011, 23, 1105–1112.
27. A. C. Tavares, M. A. M. Cartaxo, M. I. D. Pereira and F. M. Costa, J. Solid State Electrochem., 2001, 5, 57–67.
28. V. Devadoss, M. Noel, K. Jayaraman and C. A. Basha, J. Appl. Electrochem., 2003, 33, 319–323.
29. D. K. Zhong and D. R. Gamelin, J. Am. Chem. Soc., 2010, 132, 4202–4207.
30. V. Celorrio, K. Bradley, O. J. Weber, S. R. Hall and D. J. Fermín, ChemElectroChem, 2014, 1, 1667–1671.
31. J. G. McAlpin, Y. Surendranath, M. Dincă, T. A. Stich, S. A. Stoian, W. H. Casey, D. G. Nocera and R. D. Britt, J. Am. Chem. Soc., 2010, 132, 6882–6883.
32. M. W. Kanan, J. Yano, Y. Surendranath, M. Dincă, V. K. Yachandra and D. G. Nocera, J. Am. Chem. Soc., 2010, 132, 13692–13701.
33. T. Lopes, L. Andrade, F. Le Formal, M. Gratzel, K. Sivula and A. Mendes, Phys. Chem. Chem. Phys., 2014, 16, 16515–16523.
34. L. Bertoluzzi and J. Bisquert, J. Phys. Chem. Lett., 2012, 3, 2517–2522.
35. Y. K. Kim, S. H. Ahn, K. Chung, Y. S. Cho and C. J. Choi, J. Mater. Chem., 2012, 22, 1516–1520.

---

Footnote

† Electronic supplementary information (ESI) available: Characterization details about SEM images, XRD patterns, UV-vis of the samples. See DOI: 10.1039/c6ra01810f

---