Titanium dioxide macroporous materials doped with iron: synthesis and photo-catalytic properties

Abstract

A series of iron doped three-dimensional (3D) ordered macroporous TiO₂ hybrid materials (Fe–Ti–Oxide-M) have been successfully prepared by using polymethylmethacrylate nanospheres (PMMAs) as templates. The crystal structure of Fe–Ti–Oxide-M (50%) could be assigned to the Fe₃Ti₃O₁₀ phase by XRD characterization. Synthetic samples were also studied by TEM and UV-vis absorption spectra in detail. All obtained samples possessed 3D ordered macroporous structures, and the absorption spectra of Fe–Ti–Oxide-M gradually extended to the visible light region with increased doping quantities of Fe. Furthermore, the photo-catalytic degradation of tetracycline (TC) was investigated, and the observed photo-catalytic performances of Fe–Ti–Oxide-M (1%, 2%, 5%, 10%, 20%, and 50%) were associated with their composition, which further indicates that the introduction of Fe has a remarkable influence on the photo-activity of 3D ordered macroporous TiO₂ materials.

---

Introduction

Recently, much attention has been focused on strategies for the photo-catalytic purification of polluted air and wastewater.^1–3 It is worth noting that attempts to tune the chemical composition and the morphology of the matrices could provide a flexible means to optimize the catalytic performance of catalysts.^4,5 Therefore, it remains a fascinating challenge to design and synthesize novel photo-catalysts with suitable chemical modifications and microstructure for the photo-degradation of organic pollutants using solar energy.^6–8

Yablonovitch and John first proposed the concept of photonic crystals.^9,10 In photonic crystals, the 3D ordered arrangement of two different refractive index materials generates a periodic potential field, which will lead to the emergence of the energy band gap. Therefore, the band-edge effect of photonic crystals plays a role in enhancing the light absorption of substances, namely the photonic crystal slow photon effect, which extends the interaction time of the light and gain materials.¹¹ Recently, the use of opaline lattices as templates to assemble 3D ordered macroporous materials, as photonic crystals,^12–14 has attracted intense attention, because of their applications in catalysis,¹⁵ magnetics,^16,17 and photoluminescence.^18–20

The photonic crystal structure with the regulation of light (especially the slow photon effect) will help to promote the activity of material, which endows 3D ordered macroporous materials with potential applications in the field of photo-catalysis.²¹ TiO₂, due to its nature of chemical stability, environmental-friendliness, and high activity, has been widely used as a highly-efficient photo-catalyst. So, combining a 3D ordered macroporous structure and a TiO₂ matrix into a single material has become a challenge to explore novel photo-catalysts. For example, Ozin’s group²² found that the photo-activity of TiO₂ fashioned as inverse opals can be dramatically enhanced by utilizing the slow photon effect with energies close to the electronic band gap of the semiconductor. Moreover, Song et al.²³ reported that the absorption of TiO₂ in the ultraviolet region could be enhanced based on the photonic crystal slow photon effect by optimizing the photonic crystal band gap, which could increase the catalytic efficiency up to 2.3 times. Therefore, it can be considered that the catalytic properties of TiO₂ can be further improved if it is endowed with 3D ordered macroporous materials.

However, as is well known, TiO₂, due to its wide band gap (E_g = 3.2 eV), can only respond to ultraviolet light (less than 387 nm), which means a low utilization rate of solar energy.²⁴ So, although the 3D ordered macroporous structure can further enhance the photo-activity of a TiO₂ matrix, the wide band gap still limits its practical application. Despite this, it is still an interesting task to enhance the efficiency of TiO₂ matrices through visible light sensitization.²⁵ Recently, some studies have been conducted with a view to improving the visible light sensitivity of catalysts based on a TiO₂ matrix.^26,27 One of the major strategies for inducing a TiO₂ matrix to possess a visible light response is chemical doping with transition metals containing partially filled d-orbitals, such as V, Ni, Cr, Mo, Fe, Sn, Mn, and so on.^28–32 Moreover, it has been hypothesized that the addition of transition metals to a TiO₂ matrix could also enhance the rate of photo-catalytic oxidation, due to electron scavenging by the metal ions at the semiconductor surface.³³ Therefore, chemical doping of TiO₂ matrices might be an alternative way to further design novel TiO₂ catalysts with 3D ordered macroporous structures.

Inspired by the aforementioned points, in this work, we report on the preparation of a series of 3D ordered macroporous TiO₂ hybrid materials with different doping quantities of Fe (Fe–Ti–Oxide-M; Fe/(Fe + Ti) = 1%, 2%, 5%, 10%, 20%, and 50%) by using polymethylmethacrylate nanospheres (PMMAs) as templates. The morphologies and absorption spectra of Fe–Ti–Oxide-M samples were systematically investigated. More importantly, the series of Fe–Ti–Oxide-M samples exhibited different photo-sensitized degradation of tetracycline (TC) dye under the illumination of visible light, and these properties were also found to be quantity of Fe dependent.

Experimental

Materials

All the reagents were used without further purification.

Syntheses

Synthesis of monodisperse PMMA spheres. The methyl methacrylate was previously washed three times with 0.1 M NaOH aqueous solution and deionized water, respectively. In a round-bottomed flask, the persulfate solution (0.16 g potassium persulfate) was added to the reaction flask containing a mixture of deionised water (80 mL) and washed methyl methacrylate (5 mL) at 70 °C, and then the resulting mixture was stirred for 2 h. N₂ gas was flowed during the whole process. The PMMAs were collected by centrifugation. Finally, the obtained PMMAs were dried in a vacuum.

Synthesis of Ti/Fe mixture sol. First, x mol ferric nitrate (Fe(NO₃)₃·6H₂O) and 0.34 mL acetic acid were dissolved in 5.75 mL ethanol, then y mol isopropyl titanate was added to the solution. The total moles of ferric nitrate and isopropyl titanate were 0.01 mol, and the molar ratios of x/(x + y) were 1%, 2%, 5%, 10%, 20%, and 50% respectively. Finally, the mixture was stirred for 10 min.

Synthesis of Fe–Ti–Oxide-M. PMMA blocks were carefully immersed in the above solution for a few minutes, then the PMMA blocks were taken out of the solution and dried at room temperature. The PMMAs were finally removed by calcination at 550 °C for 2 h. The final product was Fe–Ti–Oxide-M (1%, 2%, 5%, 10%, 20%, and 50%).

Characterization

Powder X-ray diffraction (XRD) patterns were obtained on a D/MAX-2500 diffract meter (Rigaku, Japan) using a Cu Kα radiation source (λ = 1.54056 Å). Transmission electron microscopy (TEM) images were collected on an F20 S-TWIN electron microscope (Tecnai G2, FEI Co.), using a 200 kV accelerating voltage. UV-vis diffuse reflectance spectra of the samples were obtained from a UV-vis spectrophotometer (UV2550, Shimadzu, Japan). BaSO₄ was used as a reflectance standard. Specific surface area was measured by using a NOVA2000e analytical system made by Quantachrome Corporation (USA), the BET surface areas were obtained from nitrogen adsorption at −196 °C, and the samples were previously outgassed overnight at 120 °C under vacuum conditions. Total organic carbon (TOC) analyses were conducted on a multi N/C 2100 (Analytik Jena AG, Germany) TOC analyzer.

Photo-catalytic activity measurement

The photodegradative reaction for TC was carried out at room temperature under visible-light. The photochemical reactor contained 0.1 g sample and 100 mL TC solution (20 mg L⁻¹). To determine the initial absorbency of samples, the reactor was kept in darkness for 30 min to reach absorption equilibrium. The photochemical reactor was irradiated with a 150 W Xenon lamp which was located with a distance of 8 cm at one side of the containing solution. UV lights with wavelengths less than 420 nm were removed by a UV-cutoff filter. The sampling analysis was conducted at 10 min intervals. The photocatalytic degradation ratio (DR) was calculated by the following formula:
A₀ is the initial absorbency of TC that reached absorption equilibrium, while A_i is the absorbency after the sampling analysis.

The TC concentration was measured by a UV-vis spectrophotometer with the maximum absorption wavelength at 357 nm.

Results and discussion

Structure and microscopy characterization of Fe–Ti–Oxide-M

The synthesis steps of the macroporous materials are described in Scheme 1. The Fe–Ti–Oxide-M was prepared by using a Ti–Fe mixed solution as a precursor. In the mixed solutions, the quantities of acetic acid and ethanol took an important role, since the acetic acid as a chelating agent could effectively buffer the hydrolytic rate of isopropyl titanate, and the ethanol was used to adjust the viscosity of the solution. Then, the mixed solution can be filled into the interspace of assembled PMMAs through capillary infiltration, and solidified under mild conditions. The template PMMAs were finally removed by calcination, and the obtained products are Fe doped 3D macroporous materials based on a TiO₂ matrix.

Scheme 1 Synthetic procedure for Fe–Ti–Oxide-M.

In order to study the structures of the Fe–Ti–Oxide-M, X-ray diffraction patterns were characterized as shown in Fig. 1. The XRD pattern of Fe–Ti–Oxide-M (1%) in Fig. 1a was comparable with pure TiO₂ (JCPDS ID: 21-1272). Moreover, no traces of impure reflections could be found when the quantity of Fe was increased even up to 5% (Fig. 1c), which indicated that the Fe might be doped into the TiO₂ crystal lattice if the quantity of Fe is below 5%. Nevertheless, a new diffraction reflection at ca. 26.9°, which could be attributed to the Fe₃Ti₃O₁₀ phase (JCPDS ID: 43-1011),³⁴ appeared when increasing the quantity of Fe to 10% (Fig. 1d), while the relative intensity of reflections from TiO₂ became weaker in contrast. This confirmed that the Fe₃Ti₃O₁₀ phase began to form in the Fe–Ti–Oxide-M (10%) sample. Furthermore, when the quantity of Fe was equal to Ti (Fig. 1f), the final sample exhibited the pure Fe₃Ti₃O₁₀ phase according to the simulated XRD pattern of Fe₃Ti₃O₁₀ (Fig. 1g). Fe₃Ti₃O₁₀ is a kind of iron titanium oxide, and it contains stoichiometric iron and titanium atoms. To the best of our knowledge, investigations on macroporous materials composed of pure Fe₃Ti₃O₁₀ have been lacking. Therefore, the formation of the Fe₃Ti₃O₁₀ phase in the product would offer alternative novel macroporous materials in the field of photo-catalysis. Moreover, the crystallite sizes play an important role in the stability of the macroporous skeleton, because crystallite sizes that are too large might make the skeleton prone to collapse. So, we further calculated the crystallite sizes using the Scherrer equation, and the crystallite sizes of all samples range from 16.0 to 24.0 nm. The small size could ensure the stability of the macroporous structure.

Fig. 1 XRD patterns of Fe–Ti–Oxide-M (1%) (a), Fe–Ti–Oxide-M (2%) (b), Fe–Ti–Oxide-M (5%) (c), Fe–Ti–Oxide-M (10%) (d), Fe–Ti–Oxide-M (20%) (e), Fe–Ti–Oxide-M (50%) (f), and the simulated XRD pattern of Fe₃Ti₃O₁₀ (g).

Transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) provided insight into the morphology of Fe–Ti–Oxide-M. As depicted in Fig. 2a, the Fe–Ti–Oxide-M (1%) block is constructed with a uniform macroporous structure, and the block has a large size with a continuous macroporous structure. We can see from the magnified view of Fe–Ti–Oxide-M (1%) (Fig. 2b) that the sample consists of ca. 280 nm diameter spherical pores which are interconnected through ca. 85 nm wide apertures and surrounded by a continuous gel framework, indicating the presence of close-packed spherical pores inside the samples. Additionally, the TEM images of Fe–Ti–Oxide-M (2%, 5%, 10%, 20%, and 50%) all exhibit close-packed macroporous structures (Fig. 2c–g), indicating that the quantity of Fe inside the TiO₂ matrix did not observably influence the macroporous structure of the samples. In Fig. 2h, the HRTEM image of Fe–Ti–Oxide-M (50%) reveals that Fe–Ti–Oxide-M is composed of the Fe₃Ti₃O₁₀ phase with an interplanar spacing of about 3.30 Ǻ, which corresponds to the (111) crystalline plane.

Fig. 2 TEM images for Fe–Ti–Oxide-M (1%) (a), magnified view of Fe–Ti–Oxide-M (1%) (b), Fe–Ti–Oxide-M (2%) (c), Fe–Ti–Oxide-M (5%) (d), Fe–Ti–Oxide-M (10%) (e), Fe–Ti–Oxide-M (20%) (f), Fe–Ti–Oxide-M (50%) (g), and HRTEM image of Fe–Ti–Oxide-M (50%) (h).

The macroporous structure of Fe–Ti–Oxide-M was also characterized by nitrogen adsorption–desorption isotherms. The N₂ adsorption–desorption isotherms (Fig. 3) exhibit a hysteresis loop range of 0.8–1.0 P/P₀. The BET surface areas of all samples are around 50 m² g⁻¹, due to their 3D ordered macroporous structure. This result is a little higher than the reported SiO₂³⁵ and TiO₂³⁶ materials with 3D ordered macroporous structures. Commonly, the measured BET surface areas of macroporous materials are related to the small particle sizes of the crystallites, so the BET surface area data also reveal that the crystallite sizes of Fe–Ti–Oxide-M are relatively small, in agreement with their XRD results calculated by the Scherrer equation. Moreover, the relative large surface areas of Fe–Ti–Oxide-M can facilitate the diffusion of solvent in the system of photo-degradation, which might make a key improvement on the catalytic efficiency.³⁷

Fig. 3 N₂ adsorption–desorption isotherms of Fe–Ti–Oxide-M (1–50%).

Generally, chemical doping with metal ions can alter the semiconductor properties of a TiO₂ matrix. So, the UV-vis absorption spectra of Fe–Ti–Oxide-M were further characterized as shown in Fig. 4A. The Fe–Ti–Oxide-M (1%) exhibited a broad band from 400 to 200 nm in the UV region, and the maximal absorption was achieved at about 300 nm. In comparison, the Fe–Ti–Oxide-M (2–50%) demonstrated successive red shifts responding to the introduction of Fe, and the absorption edge of Fe–Ti–Oxide-M (50%) extended to about 700 nm. This phenomenon is in agreement with the previous reports.^38–40 Furthermore, we calculated both the indirect and direct band gaps (E_g) of samples based on their UV-vis absorption spectra in order to fully investigate the photophysical properties of the materials. As shown in Fig. 4B and C, the E_g of the samples became gradually smaller with increasing quantities of Fe, which suggested that the ability to absorb visible light can be controlled by altering the amount of Fe inside the Fe–Ti–Oxide-M. Moreover, in regard to the XRD data, the absorption of Fe–Ti–Oxide-M (20–50%) should be attributed to the Fe₃Ti₃O₁₀ phase. Therefore, we can speculate that two things may contribute to the improved red shift of the TiO₂ matrix. One is that the Fe atoms were doped into the TiO₂ crystal lattice, which embedded new energy levels into the band gap of TiO₂, and a similar phenomenon has been reported by Miyauchi’s group recently.⁴¹ The additional reason is that the Fe₃Ti₃O₁₀ phase with a narrow band gap had formed in the samples, which further induced the red shift of the absorption spectra of Fe–Ti–Oxide-M (20–50%).

Fig. 4 UV-vis absorption spectra (A) of Fe–Ti–Oxide-M (1–50%). The (B) and (C) spectra show the corresponding indirect and direct band gaps, respectively.

Photocatalytic properties

Organic compounds are used extensively in industrial products in the printing, textile, medication, and photographic industries. Many organic compounds are nontoxic themselves, but they could form highly toxic complexes with some heavy metal ions in wastewater, and cause the pollution of water resources.^42,43 TC, a popular broad-spectrum antibacterial agent, has been widely applied in bacterial infection treatment. However, there are serious threats to the ecosystem and human health if TC is released into the environment. So, the degradation of TC has been established to be one of the most promising technologies for environmental remediation.⁴⁴ Recent suggestions for the degradation of TC have mainly focused on semiconductor-based photo-catalysts using solar energy,⁴⁵ and the degradation efficiency directly determines the application of the photo-catalysts.

Based on the interesting absorption spectra and crystal structure, the photo-catalytic activity of the Fe–Ti–Oxide-M was studied on degradation of TC in an aqueous solution (20 mg L⁻¹). During the process of photo-degradation, the major absorbance (ca. 357 nm) from TC in the solution gradually decreased under visible light (λ ≥ 420 nm) for 1 hour (Fig. S1†), which suggested significant photo-degradation of TC by the Fe–Ti–Oxide-M.

As shown in Fig. 5, the photocatalytic degradation ratios of Fe–Ti–Oxide-M (1%, 2%, 5%, 10%, 20%, and 50%) are 34.1%, 38.0%, 65.0%, 78.3%, 53.2%, and 51.0% after irradiation for 1 h, respectively. Furthermore, the blank experiments were also investigated under different conditions: first, the bare TC solution was exposed to visible light for 1 h; second, the Fe–Ti–Oxide-M and TC mixed solution was reacted for 1 h in the dark. The photocatalytic degradation ratios of both blank experiments are very low, which indicated that the TC could hardly be degraded without the samples under visible light irradiation, while the Fe–Ti–Oxide-M shows much higher photo-activity.

Fig. 5 Photo-catalytic degradation ratios (detection wavelength is 357 nm) of Fe–Ti–Oxide-M (1–50%) under visible light.

To ensure the accuracy of the photodegradation ratios in consideration of the influence of physical adsorption, control experiments of physical adsorption in the samples over 30 min were further carried out. As shown in Fig. 6, all samples reached adsorption equilibrium in 20 min, which revealed that the photo-activities of samples take the main role in the photodegradation of TC with little relation to physical adsorption. Moreover, the total organic carbon (TOC) under visible-light for the Fe–Ti–Oxide-M (10%) (Fig. S2†) has been further studied to determine the extent of mineralization during photodegradation. The TOC removal percentage was calculated, and 86% of the TOC was removed after 420 min under visible light irradiation, which indicated the obvious mineralization of TC.

Fig. 6 Adsorption equilibrium curves of different samples over 30 min.

In our work, the doping quantity of Fe could efficiently extend the absorption of the TiO₂ macroporous materials to the visible region, and the Fe–Ti–Oxide-M (50%) possessed the largest absorption area in the visible region, so the Fe–Ti–Oxide-M (50%) should exhibit the best photo-catalytic performance in theory. However, Fe–Ti–Oxide-M (10%) demonstrated the highest photo-catalytic degradation ratio compared with other samples (Fig. 7). Some reasons are proposed to account for the high photocatalytic activity of the hierarchical Fe–Ti–Oxide-M (10%): as previously reported, metal ions can introduce new energy levels of the transition metal ions into the band gap of TiO₂ without changing its valence band edge, which results in the absorption of visible light for metal doped TiO₂.⁴⁶ On the other hand, the solubility of Fe in the TiO₂ matrix is commonly very low (about 1 at% for anatase), so the redundant Fe should be left as dopants on the surface of TiO₂, which would depress the photo-activity of TiO₂ by contrast.⁴⁷ With the combination of the above two factors, there should be an optimal value of doping quantity of Fe in the TiO₂ matrix, and deviation from the optimal value will work against the purpose of increasing the photo-activity of TiO₂.

Fig. 7 Histogram of photo-catalytic degradation ratios vs. ratio of Fe/(Fe + Ti).

Therefore, the optimal value of doping quantity of Fe is 10% in the 3D macroporous Fe–Ti–Oxide-M system. Because the Fe–Ti–Oxide-M (50%) was composed of a Fe₃Ti₃O₁₀ phase, the results further suggest that TiO₂ doped with a lower concentration of Fe has a better photo-catalytic degradation ratio of TC than pure Fe₃Ti₃O₁₀ under visible light.

Conclusion

Fe–Ti–Oxide-M (1%, 2%, 5%, 10%, 20%, and 50%) with a 3D macroporous structure has been successfully synthesized. Tuning the doping quantity of Fe can significantly affect the crystal structure of the resulting Fe–Ti–Oxide-M. The obtained samples exhibited the Fe₃Ti₃O₁₀ phase when the quantity of Fe was equal to Ti. Moreover, the Fe–Ti–Oxide-M showed a significant red shift in response to the doping of Fe. The photo-catalytic properties of Fe–Ti–Oxide-M were studied under visible light in-depth. Taking into account the crystal structure and absorption spectra, the Fe–Ti–Oxide-M (10%) showed the optimal catalytic activity among the samples with different doping quantities of Fe in this system.

Acknowledgements

The authors are grateful to the National Natural Science Foundation of China (21201085 and 21276116), Start-Up Foundation of Jiangsu University (11JDG105, and 11JDG153), and Natural Science Foundation of Jiangsu Province (BK2012294 and BK2011528).

References

1. Y. Q. Lei, S. Y. Song, W. Q. Fan, Y. Xing and H. J. Zhang, J. Phys. Chem. C, 2009, 113, 1280.
2. A. Testino, I. R. Bellobono, V. Buscaglia, C. Canevali, M. D. Arienzo, S. Polizzi, R. Scotti and F. Morazzoni, J. Am. Chem. Soc., 2007, 129, 3564.
3. Y. B. Zhao, W. H. Ma, Y. Li, H. W. Ji, C. C. Chen, H. Y. Zhu and J. C. Zhao, Angew. Chem., Int. Ed., 2012, 51, 3188.
4. B. Mandlmeier, J. M. Szeifert, D. Fattakhova-Rohlfing, H. Amenitsch and T. Bein, J. Am. Chem. Soc., 2011, 133, 17274.
5. H. Tong, S. X. Ouyang, Y. P. Bi, N. Umezawa, M. Oshikiri and J. H. Ye, Adv. Mater., 2012, 24, 229.
6. Y. Q. Lei, G. H. Wang, S. Y. Song, W. Q. Fan, M. Pang, J. K. Tang and H. J. Zhang, Dalton Trans., 2010, 39, 3273.
7. J. C. Zhao, C. C. Chen and W. H. Ma, Top. Catal., 2005, 35, 269.
8. Y. P. Huang, W. H. Ma, J. Li, M. M. Cheng, J. C. Zhao, L. J. Wan and J. C. Yu, J. Phys. Chem. B, 2003, 107, 9409.
9. E. Yablonovitch, Phys. Rev. Lett., 1987, 58, 2059.
10. S. John, Phys. Rev. Lett., 1987, 58, 2486.
11. J. I. L. Chen, G. Von Freymann, S. Y. Choi, V. Kitaev and G. A. Ozin, J. Mater. Chem., 2008, 18, 369.
12. I. Soten, H. Miguez, S. M. Yang, S. Petrov, N. Coombs, N. Tetreault, N. Matsuura, H. E. Ruda and G. A. Ozin, Adv. Funct. Mater., 2002, 12, 71.
13. P. D. Yang, A. H. Rizvi, B. Messer, B. F. Chmelka, G. M. Whitesides and G. D. Stucky, Adv. Mater., 2001, 13, 427.
14. Y. A. Vlasov, N. Yao and D. J. Norris, Adv. Mater., 1999, 11, 165.
15. C. Wang, A. Geng, Y. Guo, S. Jiang and X. Qu, Mater. Lett., 2006, 60, 2711.
16. Y. N. Kim, S. J. Kim, E. K. Lee, E. O. Chi, N. H. Hur and C. S. Hong, J. Mater. Chem., 2004, 14, 1774.
17. J. Gallery, M. Ginzburg, H. Miguez, S. M. Yang, N. Coombs, A. Safa-Sefat, J. E. Greedan, I. Manners and G. A. Ozin, Adv. Funct. Mater., 2002, 12, 382.
18. Y. Liu and S. Wang, Colloids Surf., B, 2007, 58, 8.
19. R. Withnall, T. G. Ireland, M. I. Martinez-Rubio, G. R. Fern and J. Silver, J. Mod. Opt., 2002, 49, 965.
20. A. Ghadimi, L. Cademartiri, U. Kamp and G. A. Ozin, Nano Lett., 2007, 7, 3864.
21. B. L. Su, Q. J. Zhang, D. Bonifazi and J. L. Li, ChemSusChem, 2011, 4, 1327.
22. J. I. L. Chen, G. Von Freymann, S. Y. Choi, V. Kitaev and G. A. Ozin, Adv. Mater., 2006, 18, 1915.
23. J. Liu, G. L. Liu, M. Z. Li, W. Z. Shen, Z. Y. Liu, J. X. Wang, J. C. Zhao, L. Jiang and Y. L. Song, Energy Environ. Sci., 2010, 3, 1503.
24. Q. Wang, C. C. Chen, W. H. Ma, H. Y. Zhu and J. C. Zhao, Chem.–Eur. J., 2009, 15, 4765.
25. A. J. Cowan, C. J. Barnett, S. R. Pendlebury, M. Barroso, K. Sivula, M. Grätzel, J. R. Durrant and D. R. Klug, J. Am. Chem. Soc., 2011, 133, 10134.
26. R. M. Navarro Yerga, M. C. Alvarez Galvan, F. del Valle, J. A. Villoria de la Mano and J. L. Fierro, ChemSusChem, 2009, 2, 471.
27. X. B. Chen, S. H. Shen, L. J. Guo and S. S. Mao, Chem. Rev., 2010, 110, 6503.
28. X. Chen and S. S. Mao, Chem. Rev., 2007, 107, 2891.
29. D. H. Kim, D. K. Choi, S. J. Kim and K. S. Lee, Catal. Commun., 2008, 9, 654.
30. D. W. Hwang, H. G. Kim, J. S. Lee, J. Kim, W. Li and S. H. Oh, J. Phys. Chem. B, 2005, 109, 2093.
31. E. Borgarello, J. Kiwi, M. Grätzel, E. Pelizzetti and M. Visca, J. Am. Chem. Soc., 1982, 104, 2996.
32. R. Konta, T. Ishii, H. Kato and A. Kudo, J. Phys. Chem. B, 2004, 108, 8992.
33. E. C. Butler and A. P. Davis, J. Photochem. Photobiol., A, 1993, 70, 273.
34. G. Chen, J. Chen, J. H. Peng and C. Srinivasakannan, J. Alloys Compd., 2011, 509, L244.
35. W. Q. Fan, J. Feng, S. Y. Song, Y. Q. Lei, S. Dang and H. J. Zhang, Opt. Mater., 2011, 33, 582.
36. B. T. Holland, C. F. Blanford and A. Stein, Science, 1998, 281, 538.
37. L. S. Zhong, J. S. Hu, L. J. Wan and W. G. Song, Chem. Commun., 2008, 1184.
38. X. H. Wang, J.-G. Li, H. Kamiyama, M. Katada, N. Ohashi, Y. Moriyoshi and T. Ishigaki, J. Am. Chem. Soc., 2005, 127, 10982.
39. Y. Q. Wang, H. M. Cheng, Y. Z. Hao, J. M. Ma, W. H. Li and S. M. Cai, J. Mater. Sci., 1999, 34, 3721.
40. F. Gracia, J. P. Holgado, A. Caballero and A. R. Gonzalez-Elipe, J. Phys. Chem. B, 2004, 108, 17466.
41. M. Liu, X. Q. Qiu, M. Miyauchi and K. Hashimoto, J. Am. Chem. Soc., 2013, 135, 10064.
42. X. M. Zhou, H. C. Yang, C. X. Wang, X. B. Mao, Y. S. Wang, Y. L. Yang and G. Liu, J. Phys. Chem. C, 2010, 114, 17051.
43. F. Cao, W. D. Shi, L. J. Zhao, S. Y. Song, J. H. Yang, Y. Q. Lei and H. J. Zhang, J. Phys. Chem. C, 2008, 112, 17095.
44. X. Xiao, R. P. Hu, C. Liu, C. L. Xing, X. X. Zuo, J. M. Nan and L. Wang, Chem. Eng. J., 2013, 225, 790.
45. G. H. Zhang, X. Zhang, Y. F. Wu, W. D. Shi and W. S. Guan, Micro Nano Lett., 2013, 8, 177.
46. K. Nagaveni, M. S. Hegde and G. Madras, J. Phys. Chem. B, 2004, 108, 20204.
47. W. Choi, A. Termin and M. R. Hoffmann, J. Phys. Chem., 1994, 98, 13669.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3ce41627e

---