Highly porous N-doped TiO₂ hollow fibers with internal three-dimensional interconnected nanotubes for photocatalytic hydrogen production

Abstract

Fabrication of TiO₂ hollow fibers was conducted by atomic layer deposition (ALD) with polysulfone fibers (PSFs) as a template. After the ALD process, the PSFs were removed by heat treatment at 500 °C for 1 h to form anatase TiO₂ hollow fibers. They were then doped with nitrogen up to ∼5 at% by treatment in NH₃ at 500–600 °C. The N-doped TiO₂ hollow fibers remained porous and the wall was full of three-dimensional interconnected nanotubes. The X-ray photoelectron spectroscopic analysis indicated that nitrogen was substitutionally doped into oxygen sites of the TiO₂ lattice, resulting in band gap narrowing and more absorption in the visible light region. They exhibited significantly improved photocatalytic activity for water splitting due to a lower energy gap, higher reactive surface area (∼100 m² g⁻¹), multiple light reflection, and better charge separation efficiency. The fibers treated at 600 °C for 1 h, containing ∼2.5 at% nitrogen, generated 0.185 μmol g⁻¹ of hydrogen after 6 h of irradiation with a 150 W Xe lamp.

---

Introduction

Photoelectrochemical splitting of water by TiO₂ was first demonstrated in the early 1970s by Fujishima and Honda.¹ Titanium dioxide, as an inexpensive, nontoxic, chemically stable, and highly efficient photocatalyst, has been extensively applied for degradation of organic pollutants, air purification, and photocatalytic hydrogen evolution.^2,3 However, due to its wide band gap (3.2 eV), only about 4% of the solar energy on the earth's surface can be utilized. Therefore, it is important to develop novel anatase-based materials with an appropriate band gap to fit the visible spectral region. Impurity doping is one of the typical approaches to extend the spectral response of titanium dioxide to visible light. Some metal dopants such as Cr, Co, W, Mo, and V have been shown to enhance the photocatalytic activity of TiO₂,^4,5 nevertheless, metal doping may have disadvantages of thermal instability and carrier trapping. It is well known that the preparation of transition metal-doped TiO₂ by ion implantation technique is more expensive.⁶ Recently, it has been observed that TiO₂ doped with nonmetals, such as F, S, and N, would efficiently extend photoresponse from UV light region to visible light region.^7–9 N-doped TiO₂ was first reported to exhibit good visible light photocatalytic activity by Sato.¹⁰ Nitrogen doping seems to have caught more attention than other anionic elements due to small ionization energy and stability. The preparation method and condition greatly affect nitrogen doping. N-doped TiO₂ has been fabricated by various methods such as ion implantation,¹¹ pulsed laser deposition,¹² chemical vapor deposition,¹³ and sputtering deposition.¹⁴ In addition, nitrogen doping can also be obtained by heating TiO₂ in an ammonia atmosphere.

Inorganic hollow fibers have attracted considerable attention because of versatile potential applications. Several techniques have been employed to fabricate inorganic hollow fibers such as electrophoretic deposition,¹⁵ template,¹⁶ and electrospinning.¹⁷ Among these methods, the template method is a simple and easy way to control the dimensions of hollow fibers. Polysulfone is a popular membrane polymer, which can be used as a template to fabricate nanomaterials because of its thermal, mechanical, and chemical stability.¹⁸ Thin films of TiO₂ on nanostructured template have been produced by several techniques. The technique of ALD has advantages of conformal coating, precise thickness control, and good uniformity on large area for fabricating nanostructures.¹⁹ ALD is based on the sequential saturated surface reactions, where each precursor is alternately pulsed to the reaction chamber, and the reactions between the precursors and surface species are self-terminating. As a result, atomic level control of film growth can be achieved. Nowadays it has become a promising deposition method for fabricating thin film materials. In this study, we have developed a facile method to deposit TiO₂ by ALD on polysulfone hollow fiber template to form TiO₂ hollow fibers. Furthermore, N-doping was conducted by annealing the TiO₂ coated polysulfone hollow fibers in NH₃ flow at different temperatures for various lengths of time. The photocatalytic activity and hydrogen production efficiency of the hollow fibers were evaluated under visible-light irradiation. A possible mechanism for the enhanced visible-light photocatalytic activity is proposed.

Experimental

Hollow porous polysulfone fibers with the wall composed of interconnected nanofibers were used as the template material. Fig. 1 shows the schematic preparation steps of N-doped TiO₂ hollow fibers. The technique and apparatus for ALD have been reported previously.^20–22 TiO₂ thin film was coated on the polysulfone hollow fibers by ALD using titanium tetrachloride (TiCl₄) and H₂O as the precursors for Ti and O, respectively, with the substrate temperature set at 100 °C.²³ Each cycle consisted of a TiCl₄ precursor pulse of 0.24 s, a purge with N₂ for 5 s, followed by a H₂O pulse for 0.08 s, and a second purge with N₂ for 5 s to remove excess precursor and by-products. The as-deposited TiO₂ was amorphous, and therefore the samples were heated at 500 °C for 1 h in air to obtain anatase TiO₂ and to remove the polysulfone. The dimensions of the fibers have become smaller after removal of polysulfone and crystallization of TiO₂ by the heat treatment. Nitrogen doping was conducted by thermal treatment of TiO₂ hollow fibers in NH₃ flow (99.9%) at 500–600 °C for 1 h, or at 600 °C for a duration of 1 h to 4 h in a tube furnace.

Fig. 1 Schematic process of fabricating N-doped TiO₂ hollow fibers.

The crystal phases of the specimens were evaluated by X-ray diffraction (XRD, Shimadzu XRD6000) with Cu Kα radiation. The surface morphologies and cross sections of the specimens were examined by field-emission scanning electron microscopy (FESEM, JSE-6500F). X-ray photoelectron spectroscopy (XPS, Ulvac-PHI PHI 1600 ESCA) was performed to analyze the surface chemistry. Transmission electron microscopic (TEM, JEOL-3000F) examination was conducted at 300 kV. The nitrogen content was measured by a N–O analyzer (Horiba, EMGA-620W). Thermogravimetric (TG) analysis at a heating rate of 20 °C min⁻¹ was carried out on a Netzsch STA409 analyzer in air. The specific surface area was estimated by using the Brunauer–Emmett–Teller (BET) method on a TriStar 3000 gas adsorption analyzer with the degassing temperature of 100 °C and the relative pressure of N₂ (P/P₀) from 0.05 to 0.3. Optical transmittance was recorded on a UV-visible spectrometer (Hitachi 383). The optical property of the specimens was analyzed by photoluminescence (PL) spectroscopy using the 632.8 nm beam of He–Ne laser. The photocatalytic H₂ evolution reaction was carried out in a quartz cell using a 150 W Xe lamp as the light source. The visible light was passed through an optical filter, which allowed only wavelengths higher than 420 nm to illuminate the samples. The photocatalyst sample (0.02 g) was immersed into an aqueous methanol solution (5 ml methanol and 20 ml deionized water) under magnetic stirring. Before the irradiation, the reactor was purged with nitrogen for 1 h to remove residual air. The evolved H₂ gas was collected and analyzed by a gas chromatograph (Shimadzu GC-2014) at 110 °C using a 5 Å molecular sieve in the column and argon as the carrier gas.

Results and discussion

The shape and morphologies of an individual polysulfone hollow fiber are shown in Fig. 2. The average inner diameter, outer diameter, and thickness of the fiber are 450, 1000, and 275 μm, respectively. The wall is composed of numerous interconnected nanofibers. Fig. 3 shows the SEM images for the cross sections of near middle wall of the TiO₂ hollow fibers. Regardless of different treatment temperatures and durations in gaseous NH₃, they exhibit nearly identical surface morphologies. The region remains porous and is full of interconnected nanotubes which are formed due to the removal of original interconnected nanofibers within the polysulfone fiber. The structure of TiO₂ hollow fibers is also well preserved after treatment in NH₃. Therefore, polysulfone is a good framework template for preparation of oxide hollow fibers by ALD technique. A unique characteristic of ALD is that it enables formation of very uniform film with almost atomic precision not only on planar substrates but also on complex three-dimensional porous substrates. As shown by the TEM micrographs in Fig. 4(a) and (b), the interconnected nanotubes within the middle wall of the fiber have the wall thicknesses of 18.7 nm and 18.5 nm, for the samples heated at 600 °C for 1 h and 4 h, respectively. Since the samples were prepared by 400 cycles of ALD, the growth rate is calculated to be approximately 0.46 Å per cycle. From Fig. 4(c) and (d), the lattice fringes indicate that the N-doped TiO₂ hollow fibers have good crystallinity. The interplanar distances of 0.348 nm and 0.351 nm could be ascribed to the d-spacing of anatase TiO₂ (101), 0.352 nm (JCPDS-ICCD # 21-1272).

Fig. 2 SEM images of a polysulfone fiber. (a) Cross-section, (b) near the middle wall, and (c) higher magnification of (b).

Fig. 3 SEM images of the middle wall of TiO₂ hollow fibers. (a) Pure TiO₂ and those treated in NH₃ at (b) 500 °C for 1 h, (c) 550 °C for 1 h, (d) 600 °C for 1 h, (e) 600 °C for 2 h, and (f) 600 °C for 4 h.

Fig. 4 TEM images of nanotubes from N-doped TiO₂ hollow fibers. The TiO₂ nanotubes were treated in NH₃ at 600 °C for (a) 1 h and (b) 4 h. The corresponding HRTEM images in (c) and (d) are from the marked regions of (a) and (b), respectively.

Fig. 5 shows the XRD patterns of TiO₂ hollow fibers annealed in NH₃ at 500–600 °C for 1 h to 4 h. They all exhibit the anatase phase, and no any other diffraction peaks are observed. The XRD patterns of the (101) plane for the specimens are enlarged in the inset. Basically, it is hard to see shift in the peaks. The d-spacings are calculated and summarized in Table 1. It is seen that the values of d-spacing increase slightly after the nitridation, and they are close to those obtained from the HRTEM analysis. Note that the ionic radii of oxygen and nitrogen are close (1.38 Å vs. 1.46 Å), and the Ti–N bond length, 1.964 Å, is only slightly longer than that of Ti–O, 1.942 Å.²⁴ Therefore, the dimensional change due to N doping is relatively minor.

Fig. 5 XRD patterns of TiO₂ hollow fibers annealed in NH₃ under various conditions. The inset shows magnified (101) peak of the anatase phase.

Table 1 Values of d (101) for pure TiO₂ and N–TiO₂

Sample	d-spacing (Å)
TiO2	3.520
N–TiO2, 500 °C – 1 h	3.524
N–TiO2, 550 °C – 1 h	3.523
N–TiO2, 600 °C – 1 h	3.524
N–TiO2, 600 °C – 2 h	3.522
N–TiO2, 600 °C – 4 h	3.525

---

In order to evaluate the proper calcination temperature for removal of polysulfone, a sample was analyzed by TGA during heating to 500 °C (Fig. 6). The polysulfone after coating with TiO₂ reveals two stages of weight loss in one hour, as shown in Fig. 6(a). The weight losses from 15 min to 17 min (i.e., 330 °C to 370 °C) and from 17 min to 45 min (i.e., 370 °C to 500 °C) correspond to the partial degradation of sulfone groups and decomposition of polysulfone main chain respectively, leaving only oxide in the fibers. This result implies that polysulfone is completely decomposed at 500 °C after 1 h. Fig. 6(b) shows TG curves versus temperature for pure and TiO₂-coated polysulfone. The temperature was kept at 500 °C for 1 h. Both samples are rapidly decomposed at 500 °C.

Fig. 6 (a) TG curves versus time of the polysulfone fiber deposited with TiO₂ and (b) TG curves versus temperature for pure and TiO₂-coated polysulfone fibers.

Fig. 7(a) shows the absorption spectra of TiO₂ hollow fibers annealed in NH₃ under various conditions. The TiO₂ hollow fibers annealed in NH₃ at 500 °C and 550 °C do not absorb visible light. However, with annealing in NH₃ at 600 °C for 1 to 4 h, absorption in the visible light region increases substantially. The absorption band edge is shifted to the right and longer wavelength as the annealing time increases. The band gap energies can be calculated by the absorbance data using the equation: αhν = A(hν − E_g)^m, where α is absorption coefficient, h is Planck's constant, A is a constant related to the effective masses of electrons and holes, and m = 0.5 for allowed direct transition.²⁵ The band gap energies of pure and N-doped TiO₂ hollow fibers are determined to be 3.22 eV, 3.21 eV, 3.19 eV, 3.06 eV, 2.99 eV, and 2.75 eV for the six samples, as shown in Fig. 7(b). Recently, the origins of visible light absorption of C-, N-, and S-doped TiO₂ nanomaterials have been studied, and it is shown that additional electronic states above the valence band edge exist, which could define the red-shift absorption of the present materials and band-gap narrowing.²⁶ Other studies have indicated that oxygen sites in TiO₂ lattice substituted by nitrogen to form isolated impurity energy levels above the valence band accounts for the visible light response.²⁷ The light absorption of the N-doped TiO₂ in the visible light region is important to its practical application since it can be more easily activated by solar light.

Fig. 7 (a) UV-visible absorption spectra and (b) estimation of band gap energies of N-doped TiO₂ hollow fibers nitridated under various conditions.

The electronic structure and chemical environment of elements on the surface were investigated by XPS. The surface chemical composition and chemical states of the samples are shown in Fig. 8. For the sample nitridated at 500 °C, no core level N 1s peak is detected, Fig. 8(a). This implies that nitrogen is not incorporated into TiO₂. For the sample nitridated at 550 °C, a small peak at a binding energy (BE) of 400.2 eV corresponding to the N 1s core level is detected. Note that the peaks at 400 eV and 402 eV are usually assigned to molecularly chemisorbed γ-N₂,²⁸ and some studies have indicated that N 1s peak at 399–400 eV is due to NH₃ adsorbed on the surface.⁷ A peak with the binding energy of 396.4 eV is observed for the samples nitridated at 600 °C, and the peak intensity increases with the treatment time. The peak could be associated with the Ti–N bonding that suggests a substitution of O ions by N ions in the TiO₂ crystal lattice.²⁸

Fig. 8 XPS spectra of the N-doped TiO₂. (a) N 1s spectra for samples treated in NH₃, (b) Ti 2p and (c) O 1s are from the sample treated at 600 °C for 1 h.

Fig. 8(b) shows a Ti 2p XPS spectra of the sample treated at 600 °C for 1 h. Two peaks at 464.8 and 459.3 eV are observed, which can be assigned to Ti 2p_1/2 and Ti 2p_3/2, respectively. These values agree well with the reported XPS data for N-doped TiO₂,²⁹ and can be ascribed to the presence of Ti⁴⁺ in the anatase phase. For all samples, the oxygen 1s core level peak appears at the same position, indicating the same nature of oxygen in the structure. However, a small broadening of the O 1s peak is visible in the sample treated at 600 °C for 1 h, as shown in Fig. 8(c). This might be due to the presence of oxygen and nitrogen in the same lattice unit of TiO₂. The peak can be deconvoluted into two peaks, one at ∼530 eV and the other at ∼532 eV. The primary peak at 530 eV can be attributed to O–Ti–O type of oxygen coordination, whereas the smaller peak at 532 eV may arise from the presence of N–Ti–O bonds.²⁹

To understand and control the doping of nitrogen in TiO₂ hollow fibers, the concentrations of nitrogen in the samples were quantitatively determined by XPS and N–O analyzer, as listed in Table 2. The nitrogen concentration increases with the increase of temperature and annealing time. The values from the N–O analyzer agree well with those from the XPS analysis, except that the analyzer offers one more digit of resolution.

Table 2 Nitrogen concentrations (at%) of N-doped TiO₂ measured from XPS spectra and N–O analyzer

Sample	XPS	N–O analyzer
500 °C, 1 h	0	0.005
550 °C, 1 h	0.94	0.842
600 °C, 1 h	2.37	2.563
600 °C, 2 h	2.85	2.956
600 °C, 4 h	4.82	4.991

---

The PL emission has been commonly used for investigation of the efficiency of charge carrier trapping, immigration, and transfer behaviors of the photoexcited electron–hole pairs in semiconductors.³⁰ A comparison of the PL emission spectra of pure and N-doped TiO₂ is shown in Fig. 9. The PL intensities of the N-doped TiO₂ samples are consistently lower than that of pure TiO₂, and they are progressively lowered as the nitridation temperature increases. Since the PL emission comes from recombination of excited electrons and holes, the lower PL intensity indicates a lower recombination rate of electron–hole under the light irradiation,³¹ which may also imply a higher photocatalytic activity of the modified TiO₂. This efficient quenching of the PL can be ascribed to two possible pathways: (1) the excited electrons are trapped by the oxygen vacancies, while the holes are trapped by the doped nitrogen, thus the recombination rate of electron-holes is reduced; (2) the excited electrons can transfer from the valence band to the new defect levels introduced by nitrogen doping that exist near the minimum of conduction band.³²

Fig. 9 Photoluminescence spectra of pure and N-doped TiO₂.

Fig. 10 shows the amounts of hydrogen production by P25 and pure and N-doped TiO₂ fibers under illumination by a Xe lamp (150 W). The treatment at 500 °C basically does not influence the water splitting efficiency much. The activity gradually increases as the treatment temperature is above 550 °C, and the amount of H₂ increases with the irradiation time. After 6 h of irradiation, the amount of hydrogen production for the sample treated at 600 °C for 1 h is 0.185 μmol g⁻¹. The photocatalytic activity of N-doped TiO₂ is enhanced because the overpotential of H⁺/H and the recombination rate of electron–hole pairs are reduced. In addition, the isolated N 2p narrow band slightly above the O 2p valence band was in charge of the visible light response in the N-doped TiO₂.^27,33 The present result is a direct evidence that N doping narrows the effective band gap of TiO₂ and improves the light conversion efficiency in the visible region. It was observed, however, that the amount of H₂ decreased with the increase of annealing time at 600 °C. In other words, excess N doping results in decrease in photocatalytic activity. An optimum doping concentration can enhance the photocatalytic performance due to increased separation efficiency of photogenerated electron–hole pairs in the impurity energy levels with an adequate gap so as to inhibit their recombination. On the contrary, excess nitrogen will spread the impurity energy levels which would act as the recombination centers for electron and hole pairs and to decrease the photocatalysis efficiency.^27,34

Fig. 10 Hydrogen production of P25, pure and N-doped TiO₂ hollow fibers under visible light (λ > 420 nm) irradiation.

A proposed mechanism of water splitting with N-doped TiO₂ hollow fibers is illustrated schematically in Fig. 11. When the porous surface of the hollow fibers is irradiated with light, although some of the photons are not directly absorbed by the hollow fibers, they are trapped within the porous structure and then repeatedly reflected until being totally absorbed. Therefore, the porosity can improve the efficiency of photocatalysis.³⁵ Besides, the hollow fibers have a unique structure to entrap incident light inside the hollow structure.³⁶ Such a multiple reflection effect could further improve the collection efficiency of light and elevate photocatalytic performance. The principle of photocatalytic water decomposition of a semiconductor is based on the conversion of light energy into electrical energy on exposure to light. Incorporation of N into the crystalline lattice of TiO₂ would modify the electronic band structure of TiO₂, leading to formation of an N 2p band at above O 2p valance band. The electrons transfer from the N 2p level directly to the conduction band of TiO₂, resulting in higher photoabsorption efficiency. Noted that for hydrogen production electrons are responsible for reducing protons, and the ability of oxidation does not affect the performance because the upper level of valence band of TiO₂ is far more positive than the energy level for oxygen evolution. Furthermore, the conduction band remains almost unchanged after N-doping, being more negative than the energy level for hydrogen production.

Fig. 11 Schematic mechanism of photocatalytic water splitting for N-doped TiO₂ hollow fibers.

The present results indicate that nitrogen can be successfully doped into TiO₂ by thermal treatment in NH₃ that leads to considerable visible light response. The advantages of this nitridation over other methods include (i) use of inexpensive chemical precursors for the synthesis of N-doped TiO₂ and (ii) formation of highly porous N-doped TiO₂ nanostructure retains the shape and morphology of the polysulfone template. During the photocatalytic reaction, the process in the hollow fibers that possess a large microvoid structure and high fraction of pores should be more dominated by the Knudsen flow, as illustrated in Fig. 11. Knudsen diffusion occurs when the mean free path is relatively long compared to the pore size, so the molecules collide frequently with the pore surface.³⁷ The highly porous structure allows mass transfer of the reactants and products during the reaction. Consequently, Knudsen diffusion caused by the porous confinement enhances the efficiency of photocatalysis. In addition, the highly porous TiO₂ hollow fibers containing internal 3D interconnected nanotubes offer a larger reactive surface area (∼100 m² g⁻¹) for improved photocatalytic activity, compared to commercial P25 TiO₂ (50 m² g⁻¹). It is believed that the main contribution to higher photocatalytic activity of the N-doped TiO₂ is band-gap narrowing because of enhanced light absorption in the visible light region. The larger surface area and better light harvesting also enhance the efficiency of photocatalysis, but to a less degree.

Conclusion

Highly porous TiO₂ hollow fibers with internal interconnected nanotubes were fabricated by ALD method using polysulfone fibers as a template. Nitrogen doping was achieved by heat treatment in NH₃ at above 550 °C. The amount of nitrogen increased with the temperature and duration. Nitrogen was inserted into the anatase TiO₂ lattice to form impurity levels at above the valence band that narrowed the band gap to enhance visible light absorption. The highly porous structure of TiO₂ hollow fibers possessed a large reactive surface area and better light harvesting that also enhanced the efficiency of photocatalysis. The sample treated at 600 °C for 1 h in NH₃ exhibited the highest water splitting efficiency.

Acknowledgements

The authors are grateful to Ministry of Science and Technology of Taiwan for supporting this research under the Contract No. NSC 99-2221-E-007-066-MY3 and NSC 101-2120-M-007-013. The polysulfone fibers supplied by Dr Jeng-Liang Guo in Material and Chemical Research Laboratories, Industrial Technology Research Institute, Hsinchu, Taiwan is highly appreciated.

References

1. A. Fujishima and K. Honda, Nature, 1972, 238, 37.
2. A. L. Linsebigler, G. Lu and J. T. Yates, Chem. Rev., 1995, 95, 735.
3. M. R. Hoffmann, S. T. Martin, W. Choi and D. W. Bahnemann, Chem. Rev., 1995, 95, 69.
4. A. Di Paola, G. Marci, L. Palmisanok, M. Schiavello, K. Uosaki, S. Ikeda and B. Ohtani, J. Phys. Chem. B, 2002, 106, 637.
5. J. Zhu, Z. Deng, F. Chen, J. Zhang, H. Chen, M. Anpo, J. Huang and L. Zhang, Appl. Catal., B, 2006, 62, 329.
6. H. Yamashita, M. Honda, M. Harada, Y. Ichihashi, M. Anpo, T. Hirao, N. Itoh and N. Iwamoto, J. Phys. Chem. B, 1998, 102, 10707.
7. R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki and Y. Taga, Science, 2001, 293, 269.
8. T. Umebayashi, T. Yamaki, H. Itoh and K. Asai, Appl. Phys. Lett., 2002, 81, 454.
9. W. Ho, J. C. Yu and S. Lee, Chem. Commun., 2006, 10, 1115.
10. S. Sato, Chem. Phys. Lett., 1986, 123, 126.
11. A. Ghicov, J. M. Macak, H. Tsuchiya, J. Kunze, V. Haeublein, S. Kleber and P. Schmuki, Chem. Phys. Lett., 2006, 419, 426.
12. Y. Suda, H. Kawasaki, T. Ueda and T. Ohshima, Thin Solid Films, 2004, 453–454, 162.
13. O. Diwald, T. L. Thompson, T. Zubkov, E. G. Goralski, S. D. Walck and J. T. Yates, J. Phys. Chem. B, 2004, 108, 6004.
14. J. Premkumar, Chem. Mater., 2004, 16, 3980.
15. R. W. Sicka, S. H. Rose and T. M. Harkulich, US Pat., 3445361, 1969.
16. H. Imai, Y. Iwaya, K. Shimizu and H. Hirashima, Chem. Lett., 2000, 29, 906.
17. K. H. Lee and Y. M. Kim, Key Eng. Mater., 1991, 61–62, 17.
18. K. I. Liu, Y. C. Hsueh, H. S. Chen and T. P. Perng, J. Mater. Chem. A, 2014, 2, 5387.
19. D. B. Farmer and R. G. Gordon, Nano Lett., 2006, 6, 699.
20. C. C. Wang, C. C. Kei, Y. W. Yu and T. P. Perng, Nano Lett., 2007, 7, 1566.
21. C. C. Wang, C. C. Kei and T. P. Perng, Nanotechnology, 2011, 22, 365702.
22. W. T. Chang, Y. C. Hsueh, S. H. Huang, K. I. Liu, C. C. Kei and T. P. Perng, J. Mater. Chem. A, 2013, 1, 1987.
23. H. S. Chen, P. H. Chen, S. H. Huang and T. P. Perng, Chem. Commun., 2014, 50, 4379.
24. C. D. Valentin, G. Pacchioni and A. Selloni, Phys. Rev. B: Condens. Matter Mater. Phys., 2004, 70, 085116.
25. J. Tauc, R. Grigorovici and A. Vancu, Phys. Status Solidi, 1966, 15, 627.
26. X. Chen and C. Burda, J. Am. Chem. Soc., 2008, 130, 5018.
27. H. Irie, Y. Watanabe and K. Hashimoto, J. Phys. Chem. B, 2003, 107, 5483.
28. M. C. Saha and H. G. Tompkins, J. Appl. Phys., 1999, 72, 3072.
29. C. S. Gopinath, J. Phys. Chem. B, 2006, 110, 7079.
30. Y. Yu, J. C. Yu, C. Y. Chan, Y. K. Che, J. C. Zhao, L. Ding, W. K. Ge and P. K. Wong, Appl. Catal., B, 2005, 61, 1.
31. Y. M. Wu, J. L. Zhang, L. Xiao and F. Chen, Appl. Surf. Sci., 2010, 256, 4260.
32. H. Tang, H. Berger, P. E. Schmid and F. Levy, Solid State Commun., 1993, 87, 847.
33. C. Di Valentin, G. Pacchioni, A. Selloni, S. Livraghi and E. Giamello, J. Phys. Chem. B, 2005, 109, 11414.
34. H. H. Chen, F. Lei, J. T. Zhao, Y. Shi and J. J. Xie, J. Mater. Res., 2013, 28, 468.
35. B. Lu, X. Li, T. Wang, E. Xie and Z. Xu, J. Mater. Chem. A, 2013, 1, 3900.
36. M. Rahman, F. Tajabadi, L. Shooshtari and N. Taghavinia, ChemPhysChem, 2011, 12, 966.
37. S. M. Mak, B. T. Tey, K. Y. Cheah, W. L. Siew and K. K. Tan, Adsorption, 2009, 15, 507.

---