Controllable synthesis of uniform BiOF nanosheets and their improved photocatalytic activity by an exposed high-energy (002) facet and internal electric field

Abstract

To date, it still remains a big challenge to develop a new photocatalyst for photocatalysis technologies. Herein, a BiOF photocatalyst with a regular nanosheet shape has been, for the first time, prepared by a simple hydrothermal method. The samples are characterized by X-ray diffraction (XRD), scanning electron microscopy (SEM), high-resolution transmission electron microscopy (HRTEM), ultraviolet-visible diffuse reflectance spectroscopy (UV-DRS), electrochemistry impede spectroscopy (EIS) and nitrogen sorption isotherms. Also, ab initio density functional theory (DFT) calculations have been carried out to give insight into the energy band and electronic structures of BiOF. Furthermore, rhodamine B (RhB) is chosen as the representative dye pollutant to evaluate the photocatalytic activity of BiOF. The results show that the uniform BiOF nanosheets grow preferentially along the [110] and [100] orientation, and 75.4% of the (002) facets are exposed. After 60 min of ultraviolet light irradiation (<420 nm), 79.3% of RhB is degraded by BiOF, while only 33.7% of RhB is degraded by the commercial rutile TiO₂. The apparent kinetic rate constant (0.02534 min⁻¹) of BiOF is 3.88 times higher than (0.00652 min⁻¹) rutile TiO₂. Moreover, the calculation results demonstrate that the high-energy (002) facets are more active than the low-energy (020) and (200) facets. For the layered BiOF there is an internal electric field (IEF) perpendicular to the [Bi₂O₂]²⁺ slabs and fluorine anionic slabs, which is favorable for the efficient separation of the photogenerated electrons and holes. It is the synergetic effect of the surface structure and bulk IEF that greatly improves the activity of the BiOF nanosheets. We expect that bulk IEF adjustment is another new strategy to develop new, efficient photocatalysts for layered materials.

---

1. Introduction

Since the first report by Fujishima and Honda,¹ photocatalytic technology has been intensively studied and various photocatalysts have been developed. Besides TiO₂,² other efficient photocatalysts have also been developed, such as oxides,³ sulfides,⁴ oxysalts,⁵ polymers,⁶ and so on. Recently, bismuth-based compounds, as efficient photocatalysts, have drawn a lot of attention. Usually, bismuth-based compounds possess hybridized band structures due to the lone pair of electrons of Bi³⁺.⁷ Among them, bismuth oxyhalides, as a new class of photocatalysts, comprise a layer of [Bi₂O₂] slabs interleaved with double slabs of halogen.⁸ Due to the unique layered structure, there is a strong internal electrostatic field (IEF) between the slab layers, which is perpendicular to the slab layers.⁹ As a result, the separation of the photogenerated electron and hole can be greatly improved by the IEF, and a high photocatalytic activity can be achieved for bismuth oxyhalides.⁹ On the basis of the electronic structures calculated by density functional theory (DFT), BiOF has a direct band gap, whereas the other three BiOX (Cl, Br, I) materials have an indirect band gap.¹⁰ To the best of our knowledge, the BiOX (Cl, Br, I) photocatalyst has been studied extensively. In particular, the crystal orientations of these three materials have been well controlled.^11,12 However, BiOF has been scarcely reported as a photocatalyst.¹³ Furthermore, it is well known that the chemical properties of fluorine are significantly different from those of Cl, Br and I. Because fluorine has the highest electronegativity,^14,15 it can strongly adsorb or trap electrons, thus effectively affecting the electron distribution. It has been widely reported that the activity of photocatalysts can be effectively improved by fluorine.^16–21 The current question for us is whether BiOF has the same crystal facet effect as BiOX (X = Cl, Br, I) or not, which needs to be investigated extensively and answered clearly. Moreover, most researchers have focused on the surface properties of the catalysts; it is desirable to develop catalytic materials by exploring and employing a new strategy. We expect that bulk IEF adjustment is another new strategy to develop new, efficient photocatalysts for layered materials.

Herein, uniform BiOF nanosheets are, for the first time, achieved by a simple hydrothermal method. We have mainly investigated the influence of pH on the samples because bismuth-containing inorganic salt precursors are fairly sensitive to the pH value.²² Besides, we have investigated the photocatalytic activity of BiOF for the degradation of RhB dye. Finally, we have calculated the surface energies of the (002), (020) and (200) facets and the electronic structure of BiOF. Moreover, we have revealed the correlation of the photocatalytic properties with the active facets and IEF of BiOF.

2. Experimental

2.1 Sample preparation

All reagents were of analytical grade, purchased from Beijing Chemical Reagents Industrial Company of China, and were used without further purification.

In a typical synthesis, 1 mmol Bi(NO₃)₃·5H₂O was dissolved in a 15 mL mixture of C₂H₅OH–H₂O (the volume ratio is 14 : 1) under magnetic stirring at room temperature. Then, 6 mmol NH₄F was added into the above solution under stirring. After stirring for 1 h, 15 mL of 2 mol L⁻¹ NH₃·H₂O was added into the above solution. The resulting suspension was transferred into a Teflon-lined stainless steel autoclave and maintained at 160 °C for 12 h. The resultant precipitate was washed with distilled water and absolute ethanol several times, respectively. Finally the sample was dried at 60 °C for 6 h.

2.2 Characterization

The crystal phases of the samples were characterized by X-ray diffraction (XRD, Rigaku D/max-2550VB), using graphite monochromatized Cu K_α radiation (λ = 0.154 nm) in the range of 20–80° (2θ) at a scanning rate of 5° min⁻¹. The acceleration voltage and applied current were 40 kV and 50 mA. The samples were characterized on a scanning electron microscope (SEM, Hitachi SU-1510) with an acceleration voltage of 15 keV. The samples were coated with a 5 nm-thick gold layer before observations. The fine surface structures of the samples were determined using a high-resolution transmission electron microscope (HRTEM, JEOL JEM-2100F) equipped with an electron diffraction (ED) attachment with an acceleration voltage of 200 kV. UV-vis diffused reflectance spectra of the samples were obtained using a UV-vis spectrophotometer (UV-2550, Shimadzu, Japan). BaSO₄ was used as a reflectance standard in a UV-vis diffuse reflectance experiment. Nitrogen sorption isotherms were performed at 77 K and <10⁻⁴ bar on a Micromeritics ASAP2010 gas adsorption analyzer. Each sample was degassed at 150 °C for 5 h before measurements. The surface area and the pore size distribution were calculated by the Brunauer–Emmett–Teller (BET) methods.

2.3 Electrochemical impedance spectroscopy (EIS) measurement

Electrochemical impedance spectroscopy (EIS) was performed from 0.1 Hz to 100 kHz at an open circuit potential of 0.3 V, and an alternating current (AC) voltage amplitude of 5 mV in 1 M KNO₃ aqueous solution. The data were analyzed by ZSimWin software.

2.4 Evaluation of the photocatalytic activity

Under UV light irradiation, the photocatalytic activities of the samples were evaluated using RhB as the probing molecule. Typically, 0.1 g of BiOF was added into the dye solution (200 mL, 10 mg L⁻¹). Under darkness, the suspension was stirred for 30 min to reach an adsorption–desorption equilibrium of dye molecules on the photocatalyst, then it was irradiated with a 500 W Xe arc lamp equipped with ultraviolet light (λ ≤ 420 nm). During photoreaction, 4 mL of the suspension was collected at a given interval time and centrifuged to remove solids. The concentration of dye was determined by using a UV-vis spectrophotometer.

2.5 Theory calculation

All calculations were carried out using density functional theory (DFT) with the exchange–correlation function described by GGA-PBE.²³ A BiOF (1 × 1) cell was used and calculations were enabled by a VASP code, in which the projector augmented wave (PAW) method represented the electron–ion interaction with a kinetic energy cutoff of 500 eV after the convergence calculation.^24,25 During optimizations, the energy and force converged to 10⁻⁴ eV per atom and 10⁻³ eV Å⁻¹, respectively. The k-points were 7 × 7 × 3, 7 × 7 × 1, 3 × 3 × 1 and 3 × 3 × 1 for the BiOF-bulk, BiOF-002 (1 × 1 × 4), BiOF-020 (1 × 2 × 4) and BiOF-200 (2 × 1 × 4) super cells, respectively. For the energy, the k-points increase to 9 × 9 × 5, 9 × 9 × 3, 5 × 5 × 3 and 5 × 5 × 3 for the BiOF-bulk, BiOF-002 (1 × 1 × 2), BiOF 3-020 (1 × 1 × 2) and BiOF-200 (1 × 1 × 2) super cells, respectively. For DOS calculations, the k-points increase to 9 × 9 × 5 for the BiOF-bulk. To generate 20 k-points by line-mode, between each specific connection point of the Brillouin zone for the energy band calculations, the vacuum in all the models was kept at 12 Å. The surface energy is represented by formulae (1) as follows:²⁶where E^slab is the total energy of the slab, E^bulk is the energy per unit of BiOF, N is the total number of units of BiOF contained in the super cell, and A is the surface area of the slab.

3. Results and discussion

3.1 Effect of the pH value on the samples

Fig. 1 shows the XRD patterns of the samples obtained at different pH values, while the other reaction conditions were kept unchanged. When 6 and 15 mL of NH₃·H₂O (2 mol L⁻¹) were added, the pH values of the system were 8 and 9, respectively. The as-obtained samples are pure-phase BiOF. All the diffraction peaks of the sample prepared at pH = 9 are in good agreement with the standard card (JCPDS: 73-1595). No impurity peaks, e.g., Bi₂O₃, BiF₃ and Bi, can be detected, indicating the formation of phase-pure BiOF. The sharp and strong peaks indicate a good crystallinity of the sample. The SEM images show that the samples prepared at pH = 8 and 9 show similar particle shapes (Fig. S1, see ESI†). When 3 mL of 1 mol L⁻¹ HNO₃ is added (pH = 2.6), the crystalline phase cannot be determined accurately (Fig. 1). When 7 mL of 1 mol L⁻¹ HNO₃ is added (pH = 2.2), the sample is mainly composed of the Bi₇F₁₁O₅ phase, along with some impurity phases that cannot be defined (Fig. 1). Without adding HNO₃ or NH₃·H₂O, the as-prepared sample is phase-pure BiF₃, with an irregular particle shape (Fig. S2, see ESI†).

Fig. 1 X-ray diffraction patterns (XRD) of the samples prepared at different pH values: BiOF; JCPDS: 73-1595; Bi₇F₁₁O₅: marked with red facet indexes.

It is obvious that the pH value is crucial in the formation of phase-pure BiOF. It has been reported that alcohols can coordinate to Bi(III) to generate alkoxide complexes.²² We could assume that a Bi(III)–C₂H₅OH complex may form. The alkoxides would slowly release the Bi(III) ions to form BiOF. The whole reaction involved could be described as follows:

Bi³⁺ + C₂H₅OH ↔ [Bi(C₂H₅OH)]³⁺ (2)

Bi³⁺ + H₂O ↔ [BiO]⁺ + 2H⁺ (3)

[BiO]⁺ + F⁻ → BiOF (4)

Herein, the added NH₃·H₂O greatly contributes to the formation of [BiO]⁺, promoting the hydrolysis of Bi(III) ions to form BiOF. It is interesting that C₂H₅OH can be used as the complexing agent to control the release of Bi(III), as well as the reaction solvent.

Typically, the sample prepared at pH = 9 is characterized by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Uniform, well-defined nanosheets are obtained (Fig. 2a). Fig. 2b shows that the nanosheets are 3–4 μm × 3–4 μm large. From high-resolution transmission electron microscopy (HRTEM) (Fig. 2c), two sets of lattice fringe spacings are determined to be 0.187 nm and 0.264 nm, which can be indexed to the (200) and (110) planes of BiOF nanosheets, respectively. This indicates that the nanosheets grow preferentially along the [100] and [110] directions. The selected-area electron diffraction (SAED) pattern shows the sharp diffraction spots along the [001] zone axis. The clear diffraction spots can be identified as the (110), (200) and (1–10) planes of BiOF.

Fig. 2 (a) SEM and (b) HRTEM images, (c) lattice fringe image (the inset shows the ED pattern) of the BiOF nanosheets prepared at pH = 9 and (d) calculation of the geometric model of a single nanosheet.

Moreover, from the XRD patterns in Fig. 1, the intensity ratio of the (002)/(101) peaks of BiOF nanosheets is calculated to be 49/100, which is significantly smaller than that (96/100) of the bulk phase (standard card), indicating that the growth of the (002) facets may be suppressed. As reported in ref. 27, more oxygen atoms are positioned on the (001) facets, and the glycerol molecule tends to bind on the (001) facet via an O_surface–H–O–C interaction, which would suppress the growth of the (001) facet along the [001] direction. Our calculation results (Section 3.3 in the main text) also show that the (002) facet contains more oxygen atoms, i.e. a high oxygen atom density on the (002) facets. Hence, ethanol molecules may tend to bind on the (002) facet via an O_surface–H–O–C interaction. As a result, the growth of (002) would be suppressed along the [001] direction, and the growth rates of the crystal in the [110] and [100] directions are much higher than that in the [001] direction, leading to the formation of nanosheets, as shown in Fig. 2d. On the basis of the analyses above, the (002) facets are mainly exposed for the BiOF nanosheets.

We have obtained the facet exposure ratio by geometric calculation.²⁷ The calculation method has been reported by Wang et al.²⁷ As shown in Fig. 2d, the length, width and thickness of the BiOF nanosheet are measured to be 4 μm, 4 μm and 0.6 μm, respectively. As a result, the exposure percentage of the (002) facet is calculated to be 75.4%.

3.2 Photocatalytic activities of BiOF

Fig. 3a and b show the photoactivity curves and the apparent reaction kinetic curves of the samples for the RhB degradation under ultraviolet light (λ ≤ 420 nm). The blank experiment without adding any photocatalyst demonstrates that after 60 min irradiation, 20% of RhB is degraded. After 60 min, 79.3% of RhB is degraded by BiOF, while only 33.7% of RhB is degraded by commercial rutile TiO₂. In addition, after 40 min, 79.5% of RhB is degraded by commercial P25. The apparent rate constant of the commercial P25, BiOF and rutile TiO₂ is 0.04196 min⁻¹, 0.0253 min⁻¹ and 0.065 min⁻¹, respectively. It is obvious that the BiOF sample exhibits a higher photocatalytic activity than rutile TiO₂, but exhibits a lower photocatalytic activity than the commercial P25. The BET area of BiOF is 2.1 m² g⁻¹, which is only 4.2% (50.0 m² g⁻¹) of rutile TiO₂. Compared with rutile TiO₂, the higher activity of BiOF nanosheets may be decided by other factors, instead of the BET area. Herein, we mainly considered the role of the internal electric field (IEF) of BiOF with a typical layer structure. As reported in ref. 29, IEF plays a key role in improving the activity for layered BiOX photocatalysts. Fig. 3c presents the cycling experiments of BiOF for RhB degradation. The photocatalytic activity of BiOF does not decrease obviously after three cycles, indicating a good stability of BiOF.

Fig. 3 (a) Degradation curves and (b) reaction kinetic curves of RhB in the samples under UV-light irradiation (λ < 420 nm): 200 mL of 50 mg mL⁻¹ RhB; 100 mg of powders. (c) Cycle curves and (d) Nyquist plots of electrodes in the frequency range of 0.1–10 kHz at an open circuit potential of 0.30 V in 1 M KNO₃ aqueous solution.

To understand the high activity of BiOF, electrochemical impedance spectroscopy (EIS) has been performed to investigate the electron transfer ability (Fig. 3d). It is well known that a smaller semicircle radius in a Nyquist plot means a smaller electrical resistance of the electrode.²⁸ The semicircle radius of rutile TiO₂ is larger than that of BiOF, indicating a smaller electrical resistance for BiOF. It seems that the featured layer structure of BiOF favors the transfer and separation of electron–hole pairs, leading to improved photocatalytic properties.

For the BiOF layered structure, there are strong covalent bonds in the [Bi₂O₂] intralayer and weak van der Waals forces in the interlayer. As a result, the non-uniform charge distribution will polarize the relative atoms and atomic orbitals to induce an internal electric field (IEF).²⁹ Fig. 4 shows a schematic diagram of the IEF, which is perpendicular to the [BiO]⁺ slabs and F⁻ anion slabs, namely, along the [001] direction. This IEF could facilitate the separation and transfer of photogenerated charges. In other words, the unique layered structure of BiOF could be responsible for the formation of the IEF. In summary, we highlighted that the photocatalytic activity of BiOF can be significantly improved by the IEF of BiOF, instead of the BET area.

Fig. 4 Diagram of the internal electric field (IEF) within BiOF.

Moreover, the inset of Fig. 5a shows the ultraviolet-visible diffuse reflectance spectrum (UV-DRS) of BiOF. The absorption edge of BiOF occurs at 384 nm and the band gap is determined to be 3.93 eV (Fig. 5a). It is obvious that BiOF has a wider band gap than that of rutile TiO₂ and Degussa P25 TiO₂ (Fig. 5a). For BiOX, the valance band maximum is mainly comprised of O_2p and X_np states (i.e. n = 2, 3, 4 and 5 for X = F, Cl, Br and I, respectively), while the conduction band minimum is mainly composed of Bi 6p states.²² It has been reported that with an increase of the atomic number of X, the contribution of X_np states of the halogen atom (n = 2, 3, 4 and 5 for X = F, Cl, Br and I, respectively) increases remarkably, and the dispersive characteristic of the energy band becomes more obvious, thus narrowing the band gap.²² It is a fact that the band gaps of BiOF, BiOCl, BiOBr and BiOI are 3.93, 3.4, 2.8 and 1.8 eV, respectively.²² Among them, BiOF has the widest band gap.

Fig. 5 (a) Tauc plot (absorption² vs. energy) of BiOF: the inset shows the UV-visible diffuse reflectance spectrum (UV-DRS) and (b) the ultraviolet-visible diffuse reflectance spectra (UV-DRS) of BiOF, rutile TiO₂ and Degussa P25 TiO₂.

As observed from Fig. 5b, the absorption edges of BiOF, rutile TiO₂ and Degussa P25 TiO₂ are about 384 nm, 400 nm and 420 nm, indicating that the three catalysts can be excited by ultraviolet light (λ ≤ 420 nm). We conclude that the absorbance order is rutile TiO₂ > Degussa P25 TiO₂ > BiOF by calculating the integral area of the curve. It is obvious that BiOF exhibits a lower absorbance than rutile TiO₂. However, BiOF exhibits a higher photocatalytic activity than rutile TiO₂. We assume that the result may not be decided by the absorbance. Furthermore, our calculation results (in Section 3.3) also demonstrate that the exposed (002) facet is a high-energy facet, which plays a crucial role in improving the activity of the catalyst. On the other hand, due to Degussa P25 TiO₂ having a wilder absorbance edge than BiOF, Degussa P25 TiO₂ exhibits a higher photocatalytic activity than BiOF.

Furthermore, trapping experiments have been performed to explore the photocatalytic degradation process, in which ammonium oxalate and dimethylsulfoxide (DMSO) are used as the hole and radical scavengers, respectively (Fig. 6). When adding 10 mL DMSO, the activity of BiOF decreases slightly, while the degradation reaction is obviously inhibited when adding 1 mmol ammonium oxalate. This indicates that the holes could be the major oxidative species for RhB over BiOF.

Fig. 6 Effect of scavengers on the degradation activity of RhB over BiOF.

3.3 Theory calculation

We have calculated the surface energies of the typical (002), (020) and (200) facets by using the VASP package. The surface energies of the (002), (020) and (200) facets are calculated to be 3.11, 1.02 and 1.02 J m⁻² (Table 1). It has been demonstrated that the high-energy facets exposed play a crucial role in improving the activity of a catalyst.^30–32 Thus the high-energy (002) facet is expected to be more active than the others.²⁸ Further observed from Fig. 7 (b)–(d), the numbers of oxygen atoms are determined as 8, 4 and 4 on the (002), (200) and (020) facets, respectively. We could expect that more oxygen atoms on the (002) facets exposed can catch the photogenerated electrons, which favors the charge separation. As a result, the high-percentage of (002) facets is beneficial for improving the activity of BiOF.

Table 1 Surface energies of different facets of BiOF calculated by the VASP package

Model	Atomic layers	Total energy (eV)	Surface energy (J m^−2)
Bulk	1	−32.527217	—
002	2	−59.596692	3.11421906
020	4	−124.651986	1.020443973
200	4	−124.651986	1.020443973

---

Fig. 7 Atomic configurations of the monoclinic BiOF by VASP: (a) super cell; (b) (002); (c) (200); (d) (020).

Moreover, ab initio density functional theory (DFT) calculations have been carried out to provide insight into the energy band and electronic structures of BiOF, because the electronic structure can provide an important insight into the physicochemical behavior of a material. Fig. 8 shows the energy band and electronic density of states (DOS) of BiOF, which are calculated by the CASTEP package. In Fig. 8a, the valence band maximum (VBM) and conduction band minimum (CBM) between the Z and R points confirm the direct band gap property of BiOF, and are in good agreement with the result of the measured absorption spectrum in Fig. 6a. The direct band gap between VBM and CBM is calculated to be 3.40 eV, with an expected diminution which is a common feature of DFT calculations.³³ For Bi-based semiconductors, it is usually found that the Bi 6s and O 2p orbitals can form a preferable hybridized valence band (VB).³⁴ Furthermore, both the total density of states (DOS) and the partial DOS are employed to understand the electronic properties of BiOF. In Fig. 8b, the VB apex of BiOF is found to be not only composed of F 2p, O 2p and Bi 6p orbitals, but also a little of the Bi 6s orbital. In addition, the CB bottom is mainly composed of the O 2p and Bi 6p orbitals. For BiOF, the charge transfer upon photoexcitation occurs from the hybrid orbitals of both F 2p and O 2p to the hybrid orbitals of both O 2p and Bi 6p. In summary, the hybridized orbitals are favorable for the separation efficiency of the electrons and holes.^18,19 As we know, most researchers have focused on the surface properties of catalysts in recent years, it is therefore desirable to develop catalytic materials by exploring and employing new strategies. We expect that bulk IEF adjustment is another new strategy to develop new, efficient photocatalysts for layered materials.

Fig. 8 Energy bands structure (a) and total and partial density of states (DOS) (b) of BiOF.

4. Conclusions

To conclude, we have developed an innovative method to acquire uniform BiOF nanosheets, of which 75.4% of the high-energy (002) facets are exposed. The BiOF nanosheets show a higher activity than commercial rutile TiO₂, which is mainly due to the synergetic effect of the high-energy (002) facet exposure and the bulk IEF. This study may show that bulk IEF adjustment is a new strategy to develop new, efficient photocatalysts for layered materials.

Acknowledgements

This work is financially supported by National Science Foundation of China (21377060, 20943004), the Project Funded by the Science and Technology Infrastructure Program of Jiangsu (BM201380277), Jiangsu Science Foundation of China (BK2012862), Six Talent Climax Foundation of Jiangsu (20100292), Jiangsu Province of Academic Scientific Research Industrialization Projects (JHB2012-10), the Key Project of Environmental Protection Program of Jiangsu (2013005), A Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD) and Jiangsu Province Innovation Platform for Superiority Subject of Environmental Science and Engineering sponsored by SRF for ROCS, SEM (2013S002) and “333” Outstanding Youth Scientist Foundation of Jiangsu (2011–2015).

References

1. A. Fujishima and K. Honda, Nature, 1972, 37–38.
2. R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki and Y. Taga, Science, 2001, 293, 269–271.
3. M. Muruganandham, R. Amutha, G.-J. Lee, S.-H. Hsieh, J. J. Wu and M. Sillanpää, J. Phys. Chem. C, 2012, 116, 12906–12915.
4. L. Huang, X. Wang, J. Yang, G. Liu, J. Han and C. Li, J. Phys. Chem. C, 2013, 117, 11584–11591.
5. S. Li, S. Liu, S. Liu, Y. Liu, Q. Tang, Z. Shi, S. Ouyang and J. Ye, J. Am. Chem. Soc., 2012, 134, 19716–19721.
6. S. J. Li, S. M. Liu, S. X. Liu, Y. W. Liu, Q. Tang, Z. Shi, S. X. Ouyang and J. H. Ye, J. Am. Chem. Soc., 2012, 134, 19716–19721.
7. T. Wen, D.-X. Zhang and J. Zhang, Inorg. Chem., 2012, 52, 12–14.
8. H. Huang, Y. He, Z. Lin, L. Kang and Y. Zhang, J. Phys. Chem. C, 2013, 117, 22986–22994.
9. L. Zhang, W. Wang, J. Yang, Z. Chen, W. Zhang, L. Zhou and S. Liu, Appl. Catal., A, 2006, 308, 105–110.
10. X. Xiao, R. Hao, X. Zuo, J. Nan, L. Li and W. Zhang, Chem. Eng. J., 2012, 209, 293–300.
11. Y. Xu and M. A. Schoonen, Am. Mineral., 2000, 85, 543–556.
12. N. Li, X. Hua, K. Wang, Y. Jin, J. Xu, M. Chen and F. Teng, J. Mol. Catal. A: Chem., 2014, 395(12), 428–433.
13. N. Li, X. Hua, K. Wang, Y. Jin, J. Xu, M. Chen and F. Teng, Dalton Trans., 2014, 43, 13742–13750.
14. W. Su, J. Wang, Y. Huang, W. Wang, L. Wu, X. Wang and P. Liu, Scr. Mater., 2010, 62, 345–348.
15. S. Liu, J. Yu, B. Cheng and M. Jaroniec, Adv. Colloid Interface Sci., 2012, 173, 35–53.
16. J. C. Yu, W. Ho, J. Yu, S. Hark and K. Iu, Langmuir, 2003, 19, 3889–3896.
17. H. Park and W. Choi, J. Phys. Chem. B, 2004, 108, 4086–4093.
18. J. C. Yu, J. Yu, W. Ho, Z. Jiang and L. Zhang, Chem. Mater., 2002, 14, 3808–3816.
19. D. Li, H. Haneda, N. K. Labhsetwar, S. Hishita and N. Ohashi, Chem. Phys. Lett., 2005, 401, 579–584.
20. Q. Xiang, K. Lv and J. Yu, Appl. Catal., B, 2010, 96, 557–564.
21. K. Lv, B. Cheng, J. Yu and G. Liu, Phys. Chem. Chem. Phys., 2012, 14, 5349–5362.
22. J. Yu, Q. Li, S. Liu and M. Jaroniec, Chem.–Eur. J., 2013, 19, 2433–2441.
23. J. Li, Y. Yu and L. Zhang, Nanoscale, 2014, 6, 8473–8488.
24. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865.
25. G. Kresse and J. Furthmüller, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169.
26. G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 1758.
27. D. Wang, G. Gao and Y. Zhang, Nanoscale, 2012, 4, 7780.
28. Q. Li and X. Hu, Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 74, 035414.
29. Y. Lv, Y. Liu, Y. Zhu and Y. Zhu, J. Mater. Chem. A, 2014, 2, 1174–1182.
30. J. Jiang, K. Zhao, X. Xiao and L. Zhang, J. Am. Chem. Soc., 2012, 134, 4473–4476.
31. B. S. Naidu, B. Vishwanadh, V. Sudarsan and R. K. Vatsa, Dalton Trans., 2012, 41, 3194–3203.
32. J. Wang, F. Teng, M. Chen, J. Xu, Y. Song and X. Zhou, CrystEngComm, 2012, 15, 39–42.
33. J. Shen, J. Zai, Y. Yuan and X. Qian, Int. J. Hydrogen Energy, 2012, 37, 16986–16993.
34. A. Kudo, H. Kato and S. Nakagawa, J. Phys. Chem. B, 2000, 104, 571–575.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra14769g

---