Electrochemically self-doped hierarchical TiO₂ nanotube arrays for enhanced visible-light photoelectrochemical performance: an experimental and computational study

Abstract

A two-step electrochemical anodization method was used to prepare typical hierarchical top-ring/bottom-tube TiO₂ nanotube arrays (TNTAs). Ti³⁺ self-doping into TiO₂ was achieved via electrochemical reduction at different negative potentials in the range from −1.0 V to −1.6 V. Compared with the pristine TNTAs, the TNTAs reduced at −1.4 V presented a dramatically enhanced photoelectrochemical performance, which showed a 2.4 times enhancement in photocurrent density under simulated AM 1.5G illumination and 2.3 times increase in visible-light photocurrent density. Approximately 100% improvement in photoelectrochemical catalytic efficiency was obtained in a phenol degradation experiment. First-principles calculations demonstrated that the new states induced by Ti³⁺ self-doping might act as a shallow donor level to promote the separation of photogenerated electron–hole pairs. Moreover, the light absorption improved by the hierarchical nanostructure and the excellent electron conductivity induced by Ti³⁺ doping also account for the enhancement in the photoelectrochemical performance. These results suggest a reasonable design of photoelectrodes for efficient photoelectrochemical applications in the future.

---

Introduction

The highly-ordered TiO₂ nanotube arrays (TNTAs) that were fabricated by the electrochemical anodization method on Ti foils have attracted tremendous interest due to their high definite surface area, vertically oriented charge transfer and transportation features,^1–3 which are suitable for extensive applications in solar cells,^4,5 water splitting,^6,7 and functional devices.⁸ However, the widespread usage of TNTAs is restricted by their wide band gap (3.2 eV) because the TNTAs merely respond to UV light. Moreover, the TNTAs usually show high recombination rate of photogenerated electron–hole pairs. Therefore, the development of effective measures to surmount these barriers has vital significance. In recent years, great efforts, such as doping with metal ions^9–12 and combining with small band gap semiconductors,^13–22 have been adopted to improve the photoelectrochemical (PEC) performance of TNTAs.

Hierarchical TNTAs with bottom-tubular and top-porous structures, which are obtained via a two-step anodization method, are selected as the photoanode since their photocatalytic activity has been proven better than conventional TNTAs due to their higher uniformity and enhanced light scattering activity.²³ In addition, doping with homospecies, Ti³⁺, has been considered as a hopeful approach to promote the PEC performance of the TNTAs photoelectrodes.^24–26 In comparison to the heteroelement doping, the self-doping method can reduce defect formation since it does not introduce significant structural distortion.^27–29 Ti³⁺ self-doped TNTAs photoelectrodes both improve the absorption of visible-light owing to the fact that Ti³⁺ generates interband states and electrical conductivity owing to the high donor density, thus leading to an improved PEC performance within not only the UV region but also the visible light region.³⁰

Numerous different TiO₂ self-doping methods have been developed thus far, such as hydrogenation,^31–33 laser irradiation,³⁴ heat treatment in reducing gas (CO or NO),⁷ microwave heating,³⁵ and high-energy-particle bombardment.³⁶ However, these self-doping methods are expensive and difficult to perform practically due to their harsh conditions and complicated processes. Recently, Zhang et al. prepared Ti³⁺ doped TiO₂ nanotubes by an electrochemical reduction method.³⁷ The electrochemical reduction approach is considered an uncomplicated, economical, and environmentally friendly solution to synthesize self-doping TiO₂. However, its underlying mechanism is still not clear, which further requires both experimental exploration and theoretical analysis.

In this study, Ti³⁺ self-doped TNTAs were successfully fabricated via the combination of the two-step electrochemical anodization method and the electrochemical reduction approach. Self-doped TNTAs have been widely researched both experimentally and theoretically. X-ray diffraction (XRD) and UV-visible spectrophotometry were used to measure the crystal structure and spectral properties of the prepared samples, respectively. PEC performance was evaluated using an electrochemical workstation. Photoelectrochemical catalytic activity driven by stimulated solar light was examined via the degradation of phenol. Finally, theoretical computation via density functional theory (DFT) was performed to understand the influence of self-doping on PEC performance.

Experimental section

Preparation of TNTAs

A TNTAs film on Ti substrate was prepared via a two-step electrochemical anodic oxidation process. Different solutions, such as acetone, ethanol and ultrapure water, were used to treat the Ti sheets (>99% purity, 35 mm × 15 mm) under ultrasonication in that order. Then, a mixed solution (HF : HNO₃ : H₂O = 1 : 4 : 5 in v/v) was employed to etch the treated Ti sheets for 1 min, followed by flushing with ultrapure water and drying in air. Ethylene glycol electrolyte solution (2 wt% H₂O and 0.3 wt% NH₄F) was used to anodize the Ti foils with a platinum sheet as the counter electrode. The distance between the anode and the cathode was 3 cm and the anodizing experiments were conducted at room temperature. The as-grown nanotube layers on Ti sheets, which were prepared in the first step of anodization at 60 V for 30 min, were removed in ultrapure water via ultrasonication. The second anodization was done at 30 V for 1 h on the same Ti sheet. The fabricated TNTAs samples were cleaned with ultrapure water and dried in air after the two-step anodization. The TNTAs obtained were calcined at 450 °C for 3 h in air ambient with a heating rate of 2 °C min⁻¹.

Preparation of electrochemically self-doped TNTAs

The two-step anodization process for the TNTAs preparation and electrochemical reduction process to produce Ti³⁺ self-doped TNTAs is presented in Scheme 1. The self-doping of the TNTAs was realized using the electrochemical reduction method on an electrochemical workstation (Zahner Zennium, Germany) in a three-electrode system at room temperature. The TNTAs, saturated calomel electrode (SCE) and platinum electrode were employed as the cathode, reference and anode electrodes, respectively. The self-doped TNTAs was prepared in a 0.5 M Na₂SO₄ aqueous solution via potentiostatic cathodic reduction at the different potentials of −1.0 V, −1.2 V, −1.4 V and −1.6 V (vs. SCE) for 10 min (denoted as −1.0 V TNTAs, −1.2 V TNTAs, −1.4 V TNTAs and −1.6 V TNTAs, respectively) at room temperature. The TNTAs were finally washed with ultrapure water and dried at 80 °C for 30 min in ambient atmosphere to give self-doped TNTAs.

Scheme 1 Two-step anodization process for the TNTAs preparation and electrochemical reduction process to produce Ti³⁺ self-doped TNTAs.

Characterization of the pristine and self-doped TNTAs

Field-emission scanning electron microscopy (FESEM; Hitachi S4800, Japan) was employed to characterize the morphologies of the samples. An X-ray diffractometer (XRD; Bruker, D8, Germany) was adopted to identify the crystal structure of the prepared samples, and a spectrophotometer (UV-vis; TU-1901, China), using BaSO₄ as the reference, was used to obtain UV-visible light absorption spectra. Raman spectroscopy was performed on a Laser Micro-Raman Spectrometer (Renishaw, inVia Reflex, UK) with a 514.5 nm excitation Ar⁺ laser. The oxygen vacancies and unpaired spins of Ti³⁺ 3d1 in the samples were investigated using electron paramagnetic resonance (EPR; Bruker, A320, Germany).

Photoelectrochemical measurement

1 M KOH (50 mL) electrolyte was used for the photoelectrochemical measurements in a quartz electrochemical cell with a standard three-electrode system, which was composed of the prepared sample, an Ag/AgCl electrode and a platinum electrode as the working, reference and counter electrodes, respectively. Photocurrent density was investigated on an electrochemical workstation (Zahner Zennium, Germany) to evaluate the photoconversion efficiency and charge transfer property of the samples. The light source was stimulated sunlight (AM 1.5, 100 mW cm⁻²) supplied by a CEL-S500 Xe lamp. Transient photocurrent response measurement was carried out under repeated on/off illumination cycles with visible-light (80 mW cm⁻², 565 ± 112 nm). Electrochemical impedance spectroscopy (EIS) was conducted using 10 mV voltage amplitude in a fixed frequency range (10⁻² to 10⁵ Hz). Mott–Schottky analysis was carried out at a frequency of 1 kHz to ascertain semiconductor properties. Incident-photon-to-current conversion efficiency (IPCE) measurements were performed under the monochromatic illumination of a PLS-SXE300C Xe lamp with different bandpass optical filters in the 350–550 nm range. Stability measurement was taken under continuous UV light (1 mW cm⁻², 365 ± 12 nm) illumination for over 2 h.

Computational details

Spin-polarized DFT calculations of all samples were carried out with projector augmented wave (PAW) pseudopotentials as implanted in the Vienna Ab initio Simulation Package (VASP).^38,39 The exchange and mutuality interactions were modeled using the generalized-gradient approximation (GGA) developed by Perdew, Burke, and Ernzerhof.⁴⁰ The cut-off energy for the plane wave expansion was set to 520 eV and a Monkhorst–Pack k-point mesh for the Brillouin-zone integration was generated with 4 × 4 × 4. For the doping structure, a relaxed 2 × 2 × 1 48-atoms anatase supercell (S.C.) was adopted, which is shown in Fig. S1.† In order to raise the predicted band gap, Hubbard U corrections were used for the Ti d states (U = 8.00 eV).

Evaluation of photoelectrochemical catalytic activity

To compare the photoelectrochemical catalytic activity of the pristine and self-doped TNTAs samples, photoelectrochemical degradation experiments were performed using phenol as an objective compound. The initial concentration of phenol was 15 mg L⁻¹ with supporting electrolyte (0.1 M Na₂SO₄), and the pH value was adjusted to 3 with 0.1 M HNO₃. A 500 W Xe lamp was employed to generate simulated sunlight (AM 1.5, 100 mW cm⁻²). An electrochemical workstation was used to conduct the photoelectrochemical degradation experiments using a standard three-electrode system, where the TNTAs electrode, a platinum electrode and an Ag/AgCl electrode served as the working photoanode, counter and reference electrodes, respectively. To maintain a uniform working condition, the potential was set at 0.3 V (vs. Ag/AgCl). To establish adsorption/desorption equilibrium, the electrode (1.5 cm × 1.5 cm) was soaked in 40 mL phenol solution in the dark for 30 min before photocatalytic degradation. The phenol concentration was then determined at the wavelength of 270 nm, every 30 min, using the spectrophotometer (UV-vis; TU-1901, China). A control experiment was also implemented by irradiating the same phenol solution without TNTAs photocatalyst to measure the self-photolysis of phenol.

Results and discussion

The wide-range and magnified top view SEM images of the −1.4 V TNTAs are presented in Fig. 1. There are no obvious changes in the morphology after electrochemical reduction of the TNTAs at −1.4 V (vs. SCE) compared to the pristine TNTAs shown in Fig. S2,† which indicates that the electrochemical reduction process did not lead to the destruction of the nanotubes. A hierarchical nanostructure of top-ring/bottom-tube was observed. The thickness of the top-layer is about 70 nm and length of bottom-tube is approximately 2.5 μm. The diameters of the top ring and bottom tube are in the range of 120–150 nm and 50–80 nm, respectively. Such a unique nanostructure with a top periodical layer and bottom vertical nanotubes facilitates both broadband optical confinement and unidirectional electron transport.³⁴

Fig. 1 (a) Wide-range and (b) magnified top view SEM images of the −1.4 V TNTAs. The inset in (a) is a cross-sectional SEM image.

Fig. 2a presents the XRD patterns of the pristine and −1.4 V TNTAs. The peak centered at ∼25.3° is the characteristic anatase phase peak (JCPDF 21-1272) with a preferential orientation of (101). There are no impurities observed in the XRD peaks after the electrochemical reduction.

Fig. 2 (a) XRD patterns and (b) EPR spectra of the pristine and −1.4 V TNTAs.

Electron paramagnetic resonance spectra could be applied to demonstrate the generation of Ti³⁺ and oxygen vacancy in the self-doped samples since the oxygen vacancy appears at g = 2.004 and the paramagnetic Ti³⁺ has a g-value of 1.94–1.99.⁴¹ As shown in Fig. 2b, the strong signals found at around g = 1.98 and g = 2.00 correspond to Ti³⁺ and oxygen vacancy, respectively. The intensity of the Ti³⁺ characteristic peak of the −1.4 V TNTAs is higher than that of the pristine TNTAs. No significant surface Ti³⁺ signature was observed at g = 2.02 due to surface oxidization. The results thus indicate that Ti³⁺ is present in the bulk rather than on the surface of the nanotubes.

The Raman spectra in Fig. S3† further confirm the formation of oxygen vacancies after the electrochemical reduction of the TNTAs. The peaks that are situated at 146, 398, 517, and 633 cm⁻¹ for the pristine TNTAs indicate the anatase phase.⁴² Compared with the pristine TNTAs, the decreased intensities of these four Raman peaks in the −1.4 V TNTAs is ascribed to the increased number of oxygen vacancies in the lattice structure.^43,44 Two extra oxygen deficiency electrons are transferred to the two Ti⁴⁺ atoms around them to form Ti³⁺. The bonding length of the Ti–O–Ti network and the change of atomic coordination in the Ti³⁺ self-doped TNTAs are caused by the presence of oxygen vacancies.⁴⁵ The high density of the Ti³⁺ dopant states and/or oxygen vacancies give rise to the approximate metallic behaviour of the material. Fig. S4† presents the obvious colour change from light gray to brown and even black after self-doping.^46–48

The UV-visible absorption spectra of the pristine TNTAs and the −1.4 V TNTAs are compared in Fig. 3. Compared with the pristine TNTAs, the absorption appears stronger in the wavelength range from 400 to 800 nm after self-doping, and the optical absorption edge of the −1.4 V TNTAs shifts toward the lower energy region. Self-doping contributed to the red shift as the band gap was narrowed. The strong absorption of the −1.4 V TNTAs in the visible region is one of the most important factors attributing to their visible-light driven photoelectrochemical activity.

Fig. 3 UV-visible absorption spectra of the pristine and −1.4 V TNTAs.

Different reduction potentials (−1.0, −1.2, −1.4 and −1.6 V) were applied to fabricate the self-doped TNTAs samples for 10 min to study the effects of reduction potentials on PEC performance. Linear sweep voltammetry was conducted under simulated sunlight illumination (AM 1.5G, 100 mW cm⁻²), and the result is presented in Fig. 4a. The photocurrent densities of the samples gradually grow with the decrease in reduction potential from −1.0 to −1.4 V. The photocurrent density of 0.51 mA cm⁻² obtained at 1.3 V versus the reversible hydrogen electrode (vs. RHE) for the −1.4 V TNTAs is maximal, whereas 0.15, 0.19, 0.23, and 0.28 mA cm⁻² are observed for the pristine TNTAs, −1.0 V TNTAs, −1.2 V TNTAs, and −1.6 V TNTAs, respectively. The photocurrent density of −1.4 V TNTAs is much higher than that of the pristine TNTAs, which represents a 2.4 times enhancement. This result indicates that Ti³⁺ doping into TNTAs at the correct level can effectively improve the photocurrent of TNTAs.

Fig. 4 (a) Linear-sweep voltammograms of the pristine and −1.4 V TNTAs under simulated sunlight with a scanning rate of 10 mV s⁻¹. (b) Photocurrent densities of the photoanodes under chopped (30 s on-off) illumination cycles with visible-light-illumination at 1.0 V vs. RHE. (c) IPCE spectra of the pristine and −1.4 V TNTAs obtained without external bias at an incident wavelength (350 to 550 nm) with the inset showing the magnified spectra (420 to 550 nm). (d) Time-dependent photocurrent densities of the pristine and −1.4 V TNTAs under continuous UV light illumination for over 2 h.

Transient photocurrent response was measured on the pristine and the self-doped TNTAs under repeated on/off illumination cycles under visible-light (80 mW cm⁻², 565 ± 112 nm) at 1.0 V vs. RHE (Fig. 4b). The photocurrent responses of the samples are fast and reproducible. The transient photocurrent density of −1.4 V TNTAs (3.3 μA cm⁻²) shows a 2.3 times increase as compared with the pristine TNTAs (1 μA cm⁻²).

The information of photo-activity at different wavelengths is as important as the photocurrent density gained under the full solar spectrum for photo-activity evaluation. IPCE measurements were adopted to characterize the photo-conversion efficiencies of the pristine and self-doped TNTAs. As shown in Fig. 4c, the IPCE results could be obtained according to the following equation:

where I is the photocurrent density, J_light is the calibrated and monochromated illumination power intensity, and λ is the incident-light wavelength. There is considerably enhanced photo-activity of the self-doped TNTAs compared with the pristine TNTAs over not only the UV but also the visible-light regions. The maximum IPCE obtained by the −1.4 V TNTAs is roughly 54% at 350 nm, whereas it is only 21% for the pristine TNTAs. The results indicate that the self-doped TNTAs could effectively respond to both UV light and visible light. The IPCE measurements show good agreement with the photocurrent performance in the full solar spectrum shown in Fig. 4a. Combined with the UV-visible absorption spectra (Fig. 3), the increased absorption within the visible-light for the −1.4 V TNTAs may be one of the significant factors leading to the PEC enhancement in the visible-light region.

Chemical and structural stability is another significant element to measure the potential of a PEC photoanode. Time-dependent measurement shows that the −1.4 V TNTAs maintain a more stable photocurrent density than that of the pristine TNTAs under continuous UV-light illumination (1 mW cm⁻², 365 ± 12 nm) for over 2 h (Fig. 4d). There is no noticeable structural degradation observed after the long-term measurement. This remarkable stability indicates that the −1.4 V TNTAs has potential application in long-term solar conversion.

EIS measurements were performed to scrutinize the interfacial properties between the photoanode and the electrolyte. The Nyquist plots are presented in Fig. 5a with the magnified spectra in the inset. The plot of the pristine TNTAs completely matches the transmission line model. The equivalent circuit of the TiO₂ electrode consists of charge-transfer resistance at the interface of the solid electrode and electrolyte and a distributed combination of electron-transport resistance through the solid electrode.^49–51 However, the plot of the −1.4 V TNTAs appears as an almost erect line at both high and low frequencies, which differs from the pristine TNTAs. The shrinking and disappearing of the semicircular arc for the −1.4 V TNTAs in the high-frequency region suggests the rapid charge transfer on the surface of electrode, thus corresponding to a low charge-transfer resistance.

Fig. 5 (a) EIS Nyquist plots of the pristine and −1.4 V TNTAs in the dark and the inset is the magnified plots at high frequencies. (b) Mott–Schottky plots of the samples measured at a fixed frequency of 1 kHz in the dark.

To characterize the semiconductor properties, flat-band voltage (E_fb), and carrier density (N_D) of the pristine and −1.4 V TNTAs, Mott–Schottky (M–S) plots were obtained at 1 kHz. The space charge capacitance (C_SC) of a semiconductor is presented as follows:⁵²

where C_SC is the space charge capacitance, N_D is the donor density, ε is the permittivity of the semiconductor electrode, E_fb is the flat-band potential ε₀ is the permittivity of the free space, E is the applied potential, k_B is the Boltzmann's constant, q is the elementary charge and T is the operation temperature. Fig. 5b shows the Mott–Schottky plots of the functional relationship between 1/C² and the applied potential. They show no changes for the semiconductor type of the TNTAs, which is n-type after self-doping of Ti³⁺ due to the positive slopes of the linear parts in the M−S plots. Furthermore, the plots were extrapolated to 1/C² = 0, and E_fb for the −1.4 V TNTAs was evaluated to be 0.10 V (vs. RHE), which has a negative shift compared with the pristine TNTAs (0.24 V vs. RHE). The charge separation at the interface between the electrolyte and semiconductor is promoted by the up shift of the Fermi level.⁵³

Furthermore, calculation of carrier density, N_D, can also be obtained from Fig. 5b using the following equation:⁵⁴

where ε₀ = 8.86 × 10⁻¹² F m⁻¹, e = 1.6 × 10⁻¹⁹ C, and ε = 48 for anatase TiO₂,⁵⁵ and the N_D values of the pristine and −1.4 V TNTAs were determined to be 5.78 × 10¹⁷ and 3.45 × 10¹⁹ cm⁻³, respectively. Therefore, the higher N_D achieved by self-doping is due to the positive shift of the Fermi level and enhanced bending degree of the band, and it signifies a faster carrier transfer. This causes the negative shift of the flat-band potential and facilitates charge separation at the semiconductor–electrolyte interface, thus an improved PEC performance.⁵⁶

The presence of oxygen vacancies influences the electronic properties of TNTAs, which was evaluated by a series of first-principles calculations, and the homologous electronic structures are given in Fig. 6. Fig. 6 shows the density of states of the anatase TiO₂ with a different number of oxygen vacancies. When one oxygen vacancy is introduced, new localized Ti³⁺ states are observed near the Fermi level (Fig. 6b). These new states may act as a shallow donor level, facilitating the separation of photogenerated electro–hole pairs. Furthermore, compared with the band structure of the pristine TiO₂, the Fermi level is shifted to the conduction band due to the presence of the oxygen vacancy. Therefore, an enhanced photoelectrochemical performance was observed in our experimental tests. As the number of oxygen vacancies increase to two, more Ti³⁺ states are formed near the Fermi level (Fig. 6c). These excess states may also serve as recombination sites or traps for charge carriers. Therefore, self-doped TiO₂ prepared at a more negative reduction potential (−1.6 V) shows a decreased photocurrent. It should be noted that as shown in Fig. 6d, the Ti³⁺ states disappear when we further increase the oxygen vacancy concentration. The theoretical calculations therefore demonstrate that there exists an optimal doping concentration, which is consistent with the experimental results.

Fig. 6 Density of states of (a) pristine TiO₂ and TiO₂ with a different number of oxygen vacancies: (b) one oxygen vacancy; (c) two oxygen vacancies; and (d) three oxygen vacancies. The Fermi level is defined to zero and marked by a vertical dash line.

Since phenol may not affect the light absorption of the photocatalysts due to its achromatic colour, it could be used as an objective compound to evaluate PEC catalytic activity. As shown in Fig. 7a, the direct photolysis of phenol was studied, and only 4% was decomposed after 4 h exposure. The phenol degradation rates by the pristine and −1.4 V TNTAs are 19% and 38%, respectively. 100% improvement was achieved in the phenol degradation after Ti³⁺ self-doping. The first-order dynamic equation is ln(C₀/C) = kt, where t is the irradiation time, k is the reaction rate constant, and C and C₀ are the reaction and original concentrations of the phenol aqueous solution, respectively. Fig. 7b shows the changes in ln(C₀/C) as a function of degradation time. The phenol degradation could be found to obey first-order reaction kinetics. There is a 1.45 times improvement of the reaction rate for the −1.4 V TNTAs (0.1238 h⁻¹) compared to that of the pristine TNTAs (0.0506 h⁻¹). Thus, this suggests that the electrochemical reduction self-doping method improves the photoelectrochemical catalytic activity actuated by stimulated solar light.

Fig. 7 Comparison of (a) degradation rates and (b) kinetic curves of the phenol degradation using the pristine and −1.4 V TNTAs under AM 1.5 G simulated sunlight.

Conclusions

Highly-ordered TNTAs could be self-doped by a simple electrochemical reduction method. The −1.4 V TNTAs show superior photoelectrochemical performance in comparison to the pristine TNTAs, which represents a 2.4 times enhancement in photocurrent density under simulated AM 1.5 G illumination. This shows approximately 100% improvement of photoelectrochemical catalytic efficiency in the phenol degradation experiment after self-doping. A combination of experimental and theoretical studies demonstrates that the enhanced PEC performance results from the unique hierarchical top-ring/bottom-tube nanostructure and the electrochemically reduced self-doping. The hierarchical nanostructure is beneficial for improving light absorption and unidirectional electron transport. Moreover, the TNTAs self-doping introduces new interband states, facilitates the separation of photogenerated electron–hole pairs, and increases the carrier density, leading to improved electrical conductivity and fast charge transfer at the semiconductor–electrolyte interface. These results provide an important guideline for further development and optimization of efficient photoelectrodes in PEC applications by modifying their morphological and electronic structures.

Acknowledgements

This study was financially supported by the National Natural Science Foundation of China (51462008), the National High Technology Research and Development Program of China (863 Program, Grant No. 2015AA034103), the Hainan Natural Science Foundation (20156220), the Tianjin University-Hainan University Collaborative Innovation Fund Project, the training program of outstanding dissertation for graduates in Hainan University, and the innovation research project for graduates in universities of Hainan Province.

References

1. A. Ghicov and P. Schmuki, Chem. Commun., 2009, 2791.
2. D. Kowalski and P. Schmuki, Chem. Commun., 2010, 46, 8585.
3. P. Roy, D. Kim, K. Lee, E. Spiecker and P. Schmuki, Nanoscale, 2010, 2, 45.
4. B. O'Regan and M. Grätzel, Nature, 1991, 353, 737.
5. J. J. Qiu, F. W. Zhuge, K. Lou, X. M. Li, X. D. Gao, X. Y. Gan, W. D. Yu, H. K. Kim and Y. H. Hwang, J. Mater. Chem., 2011, 21, 5062.
6. F. D. Hardcastle, H. Ishihara, R. Sharma and A. S. Biris, J. Mater. Chem., 2011, 21, 6337.
7. F. Zuo, L. Wang, T. Wu, Z. Y. Zhang, D. Borchardt and P. Y. Feng, J. Am. Chem. Soc., 2010, 132, 11856.
8. G. D. Du, Z. P. Guo, P. Zhang, Y. Li, M. B. Chen, D. Wexler and H. K. Liu, J. Mater. Chem., 2010, 20, 5689.
9. M. M. Momenin, Y. Ghayeb and Z. Ghonchegi, Ceram. Int., 2015, 41, 8735.
10. M. M. Momeni, M. Hakimian and A. Kazempour, Ceram. Int., 2015, 41, 13692.
11. M. M. Momeni and Y. Ghayeb, J. Alloys Compd., 2015, 637, 393.
12. M. M. Momeni and Y. Ghayeb, J. Electroanal. Chem., 2015, 751, 43.
13. M. M. Momeni, Appl. Surf. Sci., 2015, 351, 160.
14. M. M. Momeni and Y. Ghayeb, Ceram. Int., 2016, 42, 7014.
15. X. Zhang, M. Zeng, J. Zhang, A. Song and S. Lin, RSC Adv., 2016, 6, 8118.
16. M. M. Momeni and Y. Ghayeb, J. Mol. Catal. A: Chem., 2016, 417, 107.
17. M. M. Momeni, Y. Ghayeb and M. Davarzadeh, J. Electroanal. Chem., 2015, 739, 149.
18. M. M. Momeni and Y. Ghayeb, J. Solid State Electrochem., 2016, 20, 683.
19. M. M. Momeni and Y. Ghayeb, J. Appl. Electrochem., 2015, 45, 557.
20. M. M. Momeni and Y. Ghayeb, J. Mater. Sci.: Mater. Electron., 2016, 27, 1062.
21. M. M. Momeni and Y. Ghayeb, J. Mater. Sci.: Mater. Electron., 2016, 27, 3318.
22. M. M. Momeni and Y. Ghayeb, J. Iran. Chem. Soc., 2016, 13, 481.
23. Z. Zhang and P. Wang, Energy Environ. Sci., 2012, 5, 6506.
24. F. Zuo, K. Bozhilov, R. J. Dillon, L. Wang, P. Smith, X. Zhao, C. Bardeen and P. Feng, Angew. Chem., Int. Ed., 2012, 51, 6223.
25. Z. Zheng, B. Huang, X. Meng, J. Wang, S. Wang, Z. Lou, Z. Wang, X. Qin, X. Zhang and Y. Dai, Chem. Commun., 2013, 49, 868.
26. M. S. Hamdy, R. Amrollahi and G. Mul, ACS Catal., 2012, 2, 2641.
27. S. U. M. Khan, M. Al-Shahry and W. B. Ingler Jr, Science, 2002, 297, 2243.
28. S. Hoang, S. Guo, N. T. Hahn, A. J. Bard and C. B. Mullins, Nano Lett., 2012, 12, 26.
29. J. Zhang, C. Pan, P. Fang, J. Wei and R. Xiong, ACS Appl. Mater. Interfaces, 2010, 2, 1173.
30. B. H. Meekins and P. V. Kamat, ACS Nano, 2009, 3, 3437.
31. G. Wang, H. Wang, Y. Ling, Y. Tang, X. Yang, R. C. Fitzmorris, C. Wang, J. Z. Zhang and Y. Li, Nano Lett., 2011, 11, 3026.
32. X. Jiang, Y. Zhang, J. Jiang, Y. Rong, Y. Wang, Y. Wu and C. Pan, J. Phys. Chem. C, 2012, 116, 22619.
33. X. Chen, L. Liu, P. Y. Yu and S. S. Mao, Science, 2011, 331, 746.
34. N. Ohtsu, K. Kodama, K. Kitagawa and K. Wagatsuma, Appl. Surf. Sci., 2010, 256, 4522.
35. Z. H. Zhang, X. L. Yang, N. H. Mohamed, E. Ahmed, L. Shi and P. Wang, ACS Appl. Mater. Interfaces, 2014, 6, 691.
36. G. A. Kimmel and N. G. Petrik, Phys. Rev. Lett., 2008, 100, 196102.
37. Z. H. Zhang, N. H. Mohamed, H. B. Zhu and P. Wang, Phys. Chem. Chem. Phys., 2013, 15, 15637.
38. G. Kresse and J. Hafner, Phys. Rev. B: Condens. Matter Mater. Phys., 1993, 47, 558.
39. G. Kresse and J. Furthermuller, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169.
40. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865.
41. I. Nakamura, N. Negishi, S. Kutsuna, T. Ihara, S. Sugihara and K. Takeuchi, J. Mol. Catal. A: Chem., 2000, 161, 205.
42. X. Y. Pan and X. M. Ma, J. Solid State Chem., 2004, 177, 4098.
43. W. Wang, Y. R. Ni, C. H. Lu and Z. Z. Xu, RSC Adv., 2012, 2, 8286.
44. Z. K. Zheng, B. B. Huang, J. B. Lu, Z. Y. Wang, X. Y. Qin, X. Y. Zhang, Y. Dai and M. H. Whangbo, Chem. Commun., 2012, 48, 5733.
45. L. L. Si, Z. A. Huang, K. L. Lv, D. G. Tang and C. J. Yang, J. Alloys Compd., 2014, 601, 88.
46. A. Ghicov, S. P. Alba, J. M. Macak and P. Schmuki, Small, 2008, 4, 1063.
47. H. Tokudome and M. Miyauchi, Angew. Chem., Int. Ed., 2005, 44, 1974.
48. J. M. Macak, B. G. Gong, M. Hueppe and P. Schmuki, Adv. Mater., 2007, 19, 3027.
49. F. Fabregat-Santiago, G. Garcia-Belmonte, J. Bisquert, A. Zaban and P. Salvador, J. Phys. Chem. B, 2002, 106, 334.
50. F. Fabregat-Santiago, E. M. Barea, J. Bisquert, G. K. Mor, K. Shankar and C. A. Grimes, J. Am. Chem. Soc., 2008, 130, 11312.
51. J. Bisquert, J. Phys. Chem. B, 2002, 106, 325.
52. H. Tsuchiya, J. M. Macak, A. Ghicov, A. S. Räder, L. Taveira and P. Schmuki, Corros. Sci., 2007, 49, 203.
53. A. I. Kontos, V. Likodimos, T. Stergiopoulos, D. S. Tsoukleris, P. Falaras, I. Rabias, G. Papavassiliou, D. Kim, J. Kunze and P. Schmuki, Chem. Mater., 2009, 21, 662.
54. Z. Zhang, Y. Yu and P. Wang, ACS Appl. Mater. Interfaces, 2012, 4, 990.
55. P. K. Ghosh and M. E. Azimi, IEEE Trans. Dielectr. Electr. Insul., 1994, 1, 975.
56. Z. Zhang, M. N. Hedhili, H. Zhu and P. Wang, Phys. Chem. Chem. Phys., 2013, 15, 15637.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra05805a

---