Solar photocatalytic degradation of methylene blue using doped TiO2 nanoparticles

Solar photocatalytic degradation of methylene blue using doped TiO2 nanoparticles

Available online at www.sciencedirect.com ScienceDirect Solar Energy 103 (2014) 473–479 www.elsevier.com/locate/solener Solar photocatalytic degrada...

817KB Sizes 3 Downloads 156 Views

Available online at www.sciencedirect.com

ScienceDirect Solar Energy 103 (2014) 473–479 www.elsevier.com/locate/solener

Solar photocatalytic degradation of methylene blue using doped TiO2 nanoparticles R.R. Bhosale a, S.R. Pujari a, G.G. Muley b, S.H. Patil c, K.R. Patil c, M.F. Shaikh d, A.B. Gambhire d,⇑ a D.B.F. Dayanad College of Arts and Science, Solapur, 413002 Maharashtra, India Department of Physics, Sant Gadge Baba Amravati Univeristy, Amravati, 444602 Maharashtra, India c Catalysis Division, National Chemical Laboratory, Pashan Road, Pune 411008, India d Department of Chemistry, Shri Anand College of Science, Pathardi, Ahmednagar, 414102 Maharashtra, India b

Received 25 December 2013; received in revised form 1 February 2014; accepted 26 February 2014 Available online 20 March 2014 Communicated by: Associate Editor Gion Calzaferri

Abstract Doped-TiO2 nanoparticles (M:TiO2: Fe, Zn, Zr, Sb, Ce and nM:TiO2: B, C, N, P, S) with anatase structure were prepared by sol–gel method and characterized by X-ray diffraction (XRD), Transmission electron microscopy (TEM), X-ray photoelectron spectra (XPS), Brunauer–Teller method (BET), UV–Vis diffuses reflectance spectroscopy (DRS). Results revealed that the anatase structure is highly stable for all doped TiO2 prepared compounds with enhancement in the surface area. UV–Vis diffuse reflectance spectra showed that these dopants were responsible for narrowing the band gap of TiO2 and shifting its optical response from ultraviolet to visible-light region. The photocatalytic activities of these multi-doped TiO2 catalysts were investigated by degradation methylene blue in aqueous solution under solar-light illumination. The results showed an appreciable enhancement in the photoactivity of the C-doped TiO2 as compared to other multi-doped TiO2 because of the formation of Ti+3 species which prevent the recombination of electron–hole pairs in C-doped TiO2. Ó 2014 Elsevier Ltd. All rights reserved.

Keywords: Carbon; Iron; Solar photocatalysis; Methylene blue

1. Introduction Heterogeneous photocatalysis has received great attention as an advanced oxidation process for the removal of toxic organic and inorganic contaminants from water (Sharma et al., 2012; Wang et al., 2010; Gernjak et al., 2004; Gambhire et al., 2011). However, the development of a practical photocatalytic system focused on the cost ⇑ Corresponding author. Tel.: +91 02428 222736, fax: +91 02428 223033. E-mail address: [email protected] (A.B. Gambhire).

http://dx.doi.org/10.1016/j.solener.2014.02.043 0038-092X/Ó 2014 Elsevier Ltd. All rights reserved.

effectiveness by the use of renewable solar energy source. Photocatalytic degradation of organic contaminants using solar irradiation could be highly economical compared with the processes using artificial UV–Vis irradiation which required substantial electrical power input. Hence, development of solar light active photocatalytic materials is a subject of extensive current research in this field. Doping TiO2 with transition metals having electronic coupling capability (Naseri et al., 2011; Choi et al., 1994) or non-metals such as Boron (Begum et al., 2008), Carbon (Xiao et al., 2008; Khan et al., 2002), Nitrogen (Asahi et al., 2001; Gole et al., 2004), Sulfur (Wang et al., 2007;

474

R.R. Bhosale et al. / Solar Energy 103 (2014) 473–479

Umebayashi et al., 2002) is known to enhance their photocatalytic response in the visible light region. Choi et al. (1994) studied the effects of 21 different dopants on the photocatalytic activity of TiO2, Fe dopant proved to be better than Ru, V, Mo, Os and Re. Asahi et al. (2001) reported that N-doping shifted the absorption edge of TiO2 to visible light region, thereby exhibiting photocatalytic degradation of MB solution and gaseous acetaldehyde under visible irradiation. Wang et al. (2007) reported the shift of photo response of TiO2 from UV to the visible region by a C-dopant. Irie et al. (2003) prepared carbondoped anatase TiO2 nanoparticles by oxidative annealing of TiC under O2 flow at 600 °C. The modification of TiO2 by co-doping with metal and non-metal and the cooperative actions of co-doping were also investigated to improve the photocatalytic activity (Rengifo-Herrera and Pulgarin, 2010; Zhao et al., 2004; Sakatani et al., 2003; Xiao et al., 2008). At present, the doping of one kind of atom into TiO2 has gained much attention due to superior control on the concentration of dopant and fabrication of efficient, cost-effective photocatalysts in order to ease global environmental issues. However, there are few publications reporting the comparative study on photocatalytic degradation of MB using multi-doped TiO2. Furthermore, most of the research work has been carried out by irradiating catalyst suspension with artificial visible light (Lv et al., 2013) or by using UV light irradiation (Liu et al., 2011) and it is not feasible and economical for the treatment of huge quantity of industrial effluents. The present study focuses on the efficient use of sunlight and the ability of prepared photocatalyst to destroy MB under solar light irradiation.

and thiourea were used for the preparation of the nM:TiO2 samples, respectively. In a typical experiment, 0.1 mol of titanium butoxide was dissolved in 100 ml anhydrous ethanol to form solution. A certain amount of boric acid, ethylene glycol and citric acid, ammonia, ortho-phosphoric acid and thiourea were dissolved in a mixture of 50 ml deionized water containing 2 ml nitric acid and 50 ml of ethanol separately. To this, TiO2 solution was added drop-wise under vigorous stirring to form the precipitate by simultaneous addition of ammonium hydroxide pH at 7 (excluding N-doped TiO2 solution). After keeping the precipitate for aging (5 days), it was concentrated and dried. The samples, after overnight drying at 110 °C, were calcined for 2 h at 500 °C. 2.3. Characterization X-ray powder diffraction (XRD) patterns have been recorded on a model D8 Bruker AXS with monochromatic Cu radiation (40 kV and 30 mA), over the 2h collection range of 20–80°. The particle size of anatase was calculated from XRD measurement. Anatase to rutile ratio was estimated from integrated intensities of the reflection of 1 0 1 and 1 1 0 respective phases. BET surface area measurements were carried out using a Quantachrome NOVA 1200 instrument. The microscopic nanostructures were observed by transmission electron microscopy (TEM; FEI, Tecnai F30, HRTEM, FEG operated at 300 kV). FT-IR spectra were recorded on a Shimadzu-8400 spectrometer in the

2. Experimental details

Anatase

2.1. Preparation of transition metals doped TiO2 (M:TiO2)

Rutile

j i h

g f

e

Intensity (a.u.)

Fe(III), Zn(II), Zr(IV), Sb(III), and Ce(IV)-loaded (3 wt.%) TiO2 nanomaterials were prepared by sol–gel process. Titanium butoxide (98%, Aldrich) was used as the precursor of TiO2. In a typical procedure, 25 ml of titanium butoxide was hydrolyzed in 300 ml water containing 1.5 ml nitric acid. The cationic surfactant cetyltrimethylammonium bromide (CTAB), 20% (10 ml) in ethanol was dropped into the above solution. Gel formed was stirred continuously at room temperature to form a highly dispersed sol. To this, Fe, Zn, Zr, Sb, and Ce solutions (3 wt.%) were added separately and stirred again for about 5 h. After keeping the sol for aging (5 days), it was concentrated and dried at 80 °C. The samples, after overnight drying at 110 °C, were calcined for 2 h at 500 °C.

k

d c b a

2.2. Preparation of non-metals doped TiO2 (nM:TiO2) 20

B, C, N, P and S doped TiO2 samples were synthesized using the controlled hydrolysis of titanium butoxide. The dopant starting materials boric acid, mixture of ethylene glycol and citric acid, ammonia, ortho-phosphoric acid

30

40

50 2 (θ 0 )

60

70

80

Fig. 1. XRD profiles of (a) pure TiO2, (b) Fe–TiO2, (c) Zn–TiO2, (d) Zr– TiO2, (e) Sb–TiO2, (f) Ce–TiO2, (g) B–TiO2, (h) C–TiO2, (i) N–TiO2, (j) P– TiO2, and (k) S–TiO2, calcined at 500 °C.

R.R. Bhosale et al. / Solar Energy 103 (2014) 473–479

475

Table 1 BET surface area, particle size, and band gap energy prepared samples. Samples

Average crystalline size (nm)

BET (m2/g)

MB removal after 60 min (%)

Energy of the band gap (eV)

Pure TiO2 Fe–TiO2 Zn–TiO2 Zr–TiO2 Sb–TiO2 Ce–TiO2 B–TiO2 C–TiO2 N–TiO2 P–TiO2 S–TiO2

12.61 10.20 25.27 14.47 11.04 13.50 23.77 9.88 11.27 17.36 14.50

34 149 80 104 138 115 88 173 147 67 63

40 93.1 84 80 85 80 76 98.7 91.3 75 37

3.2 2.64 2.7 2.8 2.25 2.90 2.85 2.44 2.31 2.86 2.94

Fig. 2. The TEM images of (a) Fe–TiO2, (b) N–TiO2 and (c) C–TiO2.

range of 4000–500 cm 1. X-ray photoelectron spectroscopy (XPS; ESCA-3000, VG Microtech, Uckfield, UK) was used to study the chemical composition of the samples. Non-chromatic X-ray beams of Al Ka (hm = 1486.6 eV) and Mg Ka (hm = 1253.6 eV) radiation were used as the excitation source. A hemispherical sector analyzer and multichannel detectors were used to detect the ejected photoelectrons as a function of their kinetic energies. XPS spectra were recorded at a pass-energy of 50 eV, 5-mm slit width and a take-off angle of 55°. The spectrometer was

calibrated by determining the binding energy values of the Au4f7/4 (84.0 eV), Ag3d5/2 (368.4 eV) and Cu2P3/2 (932.6 eV) levels using spectroscopically pure materials. The instrumental resolution under these conditions was 1.6 eV full-width at half-maximum (FWHM) for Au4f7/4 level. The Cls (285 eV) and Au4f7/4 (84.0 eV) were used as internal standards when needed. UV–Vis diffuse reflectance spectra (UV–Vis-DRS) were recorded in an air at room temperature in the wavelength range of 200–800 nm using a PE LAMBDA35 spectrophotometer.

476

R.R. Bhosale et al. / Solar Energy 103 (2014) 473–479

2.4. Photocatalytic activity

b j Absorbance (a.u.)

All the solar photocatalytic experiments were carried out at the same conditions on March 2012 from 12.00 p.m. to 1.00 p.m. Solar light was used as the irradiation source, and the average insolation of the solar irradiation was 25.28 W/m2 measured by an UV irradiance meter at range of 375–475 nm. Photocatalytic experiments were carried out by adding 20 mg of photocatalyst to 50 ml (30 ppm) solution of MB in closed cylindrical Pyrex bottles (100 ml). The whole set up was then placed in sunlight between 12.00 p.m. and 1.00 p.m. in the month of March. Prior to irradiation, the suspensions were magnetically stirred in the dark for 30 min to ensure establishment of an adsorption–desorption equilibrium among the photocatalyst, MB and atmospheric oxygen. At a given irradiation time interval, 10 ml of the suspension was collected, and then filtered through a Millipore filter to separate the photocatalyst. The changes in MB concentration were analyzed by a UV– Visible spectrophotometer and the absorption peak at 650 nm was recorded.

g i h f e

3. Results and discussion

d a

3.1. XRD analysis To understand phase symmetry in the calcined samples, a systematic X-ray diffraction study was undertaken. Fig. 1(a–k) shows the XRD patterns of the pure TiO2, M:TiO2 and nM:TiO2 samples, respectively. Pure TiO2 (Fig. 1(a)) shows two main peaks at 2h = 25.4 and 27.5, corresponding to (1 0 1) phase of anatase and (1 1 0) phase of rutile, respectively. In the case of multi-doped TiO2 (Fig. 1(b–k)), the rutile phase is <3%, which means multidoping retards the transformation from anatase to rutile phase. Multi-doping of the TiO2 stabilizes a well-crystallized pure anatase upon calcination at 500 °C, in contrast with the simultaneous growth of the rutile phase observed for the pure TiO2. The results are in good agreement with previous reports (Castro et al., 2009; Yang et al., 2009). Further, all peaks measured by XRD analysis could be assigned to those of TiO2 crystal. No peaks corresponding to the metal and non-metal oxide is detected, suggesting that it exist as the amorphous phase without getting incorporated into the TiO2 phase; that is, they are in a highly dispersed form on the surface. The peaks of TiO2 have been slightly shifted due to solid solution of metal and nonmetal ion with TiO2. The average particle sizes of the samples were calculated using Debye–Scherrer formula based on the XRD peak broadening analysis at 101 peaks, listed in Table 1. The particle size calculated from XRD data is as large as 15–25 nm for Zn, Zr, Sb, Ce, B, P and S doped TiO2 and as small as 10–13 nm for the Fe, C and N doped TiO2. This apparent fall in the particle size (higher specific surface area) will ensure high photocatalytic activity for the

k c

200

300

400

500

600

700

800

wavelength (nm) Fig. 3. (a–k) UV–Vis-DR spectra of (a) pure TiO2, (b) Fe–TiO2, (c) Zn– TiO2, (d) Zr–TiO2, (e) Sb–TiO2, (f) Ce–TiO2, (g) B–TiO2, (h) C–TiO2, (i) N–TiO2, (j) P–TiO2, and (k) S–TiO2, calcined at 500 °C.

Fe, C and N doped TiO2. when it is used for photocatalytic applications. 3.2. TEM analysis In order to further confirm the effect of metal and nonmetal doped samples on particle size and hence higher specific surface area of composite powder, the particle size of Fe–TiO2, N–TiO2 and C–TiO2 were observed using TEM. Fig. 2(a–c) shows TEM images of Fe–TiO2, N–TiO2 and C–TiO2, respectively and its corresponding Fourier transfer patterns (FTT) are also presented in the inset of figures. It can be seen that the particle size of composite powders are about 10–15 nm.

R.R. Bhosale et al. / Solar Energy 103 (2014) 473–479

477

3.3. UV–Vis DR spectral analysis

3.4. XPS studies

UV–Vis diffuse reflectance spectroscopy (Fig. 3(a–k)) permits the detection of framework of TiO2 in the samples. In all the samples, characteristic band for tetrahedrally coordinated titanium appears at about 350 nm. A progressive red-shift in the band-gap absorption is noticed with metal and non-metal loading than that of pure TiO2 (Fig. 3(a)). However, the edges of the absorption of the metal and non-metal ion doped samples were shifted to approximately 500 nm, corresponding to band-gap energy of 2.11 eV. All doped samples show enhanced absorption in the range 400–600 nm, with higher increase for samples containing iron and antimony. The absorption onsets were determined by linear extrapolation from the inflection point of the curve to the baseline.

To investigate the O, Ti, C, N and Fe in the samples, we measured the O 1s, Ti 2p, N 1s, C 1s and Fe 2p XPS spectra. The O 1s and Ti 2p XPS spectra of the C–TiO2, N–TiO2 and Fe–TiO2 samples are shown in Fig. 4(a–f). Fig. 4(a–c) exhibits a main peak at around 530 eV, which can be assigned to bulk O2 from TiO2 (Wu et al., 2006). The O 1s peak of the C–TiO2 shows broadening and asymmetry towards the higher binding energy side. These peaks are resolved into three components with binding energies values of 530.4, 531.7 and 533.5 eV for the first, second and third peak, respectively. The binding energy of the first, second and the third peak is well matching with the binding energy of TiO2 and Ti2O3 lattice oxygen, respectively, while the third peak is either adsorbed oxygen or hydroxyl species on the surface. Fig. 4(d) shows Ti2p XPS spectra of the

(a) C-doped TiO2

(b) N-doped TiO2

O 1s

(c) Fe-doped TiO2

O 1s

O 1s

O1s(Ti2O3) O1s(OH)

Intensity (a.u.)

Intensity (a.u.)

Intensity (a.u.)

O1s(TiO2)

529.6

530.6

(a)

530

535

540

524

526

(e)

Ti 2p3/2

532

534

536

526

538

I

456

458

460

462

464.24

465.19

459.48

452 454 456 458 460 462 464 466 468 470 472

Binding Energy (eV)

(h)

N 1s

(i)

Fe 2p

399.1 397.7

Fe 2p1/2

Fe 2p3/2

(f)

(e)

270

538

(c)

Intensity (a.u.)

Intensity(a.u.)

287.2

536

Binding Energy (eV)

C 1s

284.4

534

450 452 454 456 458 460 462 464 466 468 470 472

Binding energy (eV)

(g)

532

Ti 2p3/2

Intensity (a.u.)

454

530

(f)

(d)

452

528

Binding Energy (eV)

Ti 2p3/2

458.53

Intensity (a.u.)

II

Intensity (a.u.)

530

Binding Energy (eV)

Binding energy (eV)

(d)

528

Intensity (a.u.)

525

275

280

285

Binding Energy(eV)

290

390

392

394

396

398

400

402

Binding Energy (eV)

404

406

700

705

710

715

720

Binding energy (eV)

Fig. 4. (a–i) High resolution XPS spectra of the O 1s, Ti2p, C 1s, and N 1s regions taken on (a) C–TiO2 and (b) N–TiO2.

725

730

R.R. Bhosale et al. / Solar Energy 103 (2014) 473–479

C–TiO2. The Ti2p level shows asymmetry toward the lower binding energy side. The peaks are resolved into two components, with binding energy values of 457.5 and 458.6 eV for the first and second components, respectively. These binding energies are well matching with the binding energy values of Ti2O3 and TiO2 (Yu and Zhao, 2001). The binding energies and spin orbital splitting (difference between binding energy of Ti2p3/2 and Ti2p1/2) is well matching with reported values of Ti2O3 and TiO2. The formation of Ti2O3 might be due to the reduction in residual carbon in the layer from organic radicals. The carbon produced in decomposition of organic radicals during thermal treatment at 500 °C draws oxygen from the surrounding atmosphere, which causes a reduction in some Ti (IV) to Ti (III) species. Fig. 4(e–f) reveals the characteristic doublet Ti 2p3/2 and Ti 2p1/2 at around 459 and 465 eV, respectively, which indicates that Ti exists in the Ti4+ form on the surface of the N–TiO2 and Fe–TiO2 sample (Reddy et al., 2006). In Fig. 4(g), the C 1s XPS spectra of the C–TiO2 sample are shown. The C 1s spectra of the sample shared two peaks at around 284.4 and 287.2 eV binding energy. The 284.4 eV peak is due to carbon-containing species adsorbed on the surface, and the 287.2 eV peak indicates the presence of C–O bonds (Ren et al., 2007). The N 1s XPS spectra are shown in Fig. 4(h). The peak at around 400 eV and 397 eV in the N–TiO2 are attributed to N 1s electron within different Ti-N bonding environments (Chen et al., 2005). Fig. 4(i) shows Fe2p high resolution XPS spectra of Fe-TiO2 sample. The spectra shows peak position at binding energy of 710.8 eV which corresponds to element Fe+3. 3.5. Photocatalytic activity The photocatalytic activities of pure TiO2, M:TiO2 and nM:TiO2 samples are shown in Fig. 5. In the presence of pure TiO2, decomposition of MB was not observed. However, in the presence of the metal and non-metal doped TiO2 samples, the decomposition of MB obviously increased. Among the different metal and non-metal incorporated samples, C–TiO2 sample exhibited the highest photocatalytic activity under solar light irradiation, only 5–7% of MB remained, and in the case of Fe–TiO2, N–TiO2 10% of MB remained after exposure to solar light for 60 min. While as high as 20–25 % remained in the case of Zn– TiO2, Zr–TiO2, Sb–TiO2, Ce–TiO2, B–TiO2, P–TiO2, and S–TiO2. From the observed results it was found that the C–TiO2 is found to be composite photocatalyst, in this case once optical excitation occurs, the photogenerated electrons can be transferred to the lower-lying conduction bands of carbon while the holes will accumulate in the valence band of TiO2, and effectively scavenged by the oxidation of MB, where as the photogenerated electrons can be transferred into the surface of carbon rather than undergoing bulk recombination. C–TiO2 sample calcined at 500 °C temperature present a higher surface area. The high surface area of metal and non-metal ion doped TiO2 effectively

C-TiO2 Fe-TiO2

30

N-TiO2 Zn-TiO2 Sb-TiO2 Zr-TiO2

25

Ce-TiO2

Degree of decomposition (%)

478

S-TiO2 P-TiO2

20

B-TiO2 Pure TiO2

15

10

5

0 0

10

20

30

40

50

60

Irradiation time (min) Fig. 5. Rate of decomposition of MB by using (a) pure TiO2, (b) Fe–TiO2, (c) Zn–TiO2, (d) Zr–TiO2, (e) Sb–TiO2, (f) Ce–TiO2, (g) B–TiO2, (h) C– TiO2, (i) N–TiO2, (j) P–TiO2, and (k) S–TiO2, calcined at 500 °C.

concentrates MB around the loaded photocatalysts and produces high concentrations of organic compounds for the TiO2 photocatalyst. The carbon in C–TiO2 reduces TiO2 to form Ti+3 ions. Ti+3 can trap photogenerated electrons in the conduction band and prevent the recombination of electron–hole pairs under visible light radiation. Therefore, the increase in Ti+3 content enhances photocatalytic activity (Yu and Zhao, 2001), which is evident from XPS spectra showing the formation of Ti+3 species. Moreover, M:TiO2 and nM:TiO2 samples shows redshift in the absorption range compared with pure TiO2. The existence of oxygen deficiencies, probably located at the anatase–rutile boundary, leads to localized electronic states between the valence and conduction band, showing certain absorption in the visible range. This seems to indicate that the absorption could also enhance the efficiency of the photocatalytic reaction since the number of photons participating in the photocatalytic reaction is larger. Furthermore, the higher specific surface area of C–TiO2 sample implies higher adsorption capacity. The conjunction of all this features can lead to the improvement of the photocatalytic efficiency of the C–TiO2 composite material. 4. Conclusions Multi-doped TiO2 nanoparticles were successfully prepared by a sol–gel method. The prepared samples were characterized by XRD, TEM, FT-IR, XPS and UV–Vis DRS. It was found that the prepared photocatalyst exhibited smaller

R.R. Bhosale et al. / Solar Energy 103 (2014) 473–479

shape particles and higher specific surface area. In addition to oxides of Ti+4, there was a certain amount of Ti+3 oxides existed. Furthermore, the absorption edge of metal and nonmetal doped TiO2 exhibited significant red shift to visible region, which may be caused by the formation of new band gap level. The high photocatalytic activity of the C–TiO2 under visible irradiation can be attributed to small particle size, high specific surface are, optical absorption and displayed photocatalytic activity in the visible region. Acknowledgement This work is financed by University Grants Commission, New Delhi, India (Grant No. 47–2028/11). References Asahi, R., Morikawa, T., Ohwaki, T., Aoki, K., Taga, Y., 2001. Science 293, 269. Begum, N.S., Ahemad, H.M.F., Hussain, O.M., 2008. Bull. Mater. Sci. 31, 741. Castro, A.L., Nunes, M.R., Carvalho, M.D., Ferreira, L.P., Jumas, J.-C., Costa, F.M., Floreˇncio, M.H., 2009. J. Solid State Chem. 182, 1838. Chen, X., Lou, Y., Samia, A., Burda, C., Gole, J.L., 2005. Adv. Funct. Mater. 15, 41. Choi, W., Termin, A., Hoffmann, M.R., 1994. J. Phys. Chem. 98, 13669. Gambhire, A.B., Lande, M.K., Arbad, B.R., Rathod, S.B., Gholap, R.S., Patil, K.R., 2011. Mater. Chem. Phys. 125, 807.

479

Gernjak, W., Maldonado, M.I., Malato, S., Ca´ceres, J., Krutzler, T., Glaser, A., Bauer, R., 2004. Sol. Energy 77, 567. Gole, J.L., Stout, J.D., Burda, C., Lou, Y., Chen, X., 2004. J. Phys. Chem. B 108, 1230. Irie, H., Watanabe, Y., Hashimoto, K., 2003. Chem. Lett. 32, 772. Khan, S.U.M., Al-shahry, M., Ingler Jr., W.B., 2002. Science 297, 2243. Liu, H., Zhou, Y., Huang, H., Feng, Y., 2011. Desalination 278, 434. Lv, J., Sheng, T., Su, L., Xu, G., Wang, D., Zheng, Z., Wu, Y., 2013. Appl. Surf. Sci. 284, 229. Naseri, N., Yousefi, M., Moshfegh, A.Z., 2011. Sol. Energy 85, 1972. Reddy, B.M., Rao, K.N., Reddy, G.K., Bharali, P., 2006. J. Mol. Catal. A. Chem. 253, 44. Ren, W., Ai, Z., Jia, F., Zhang, L., Fan, X., Zou, Z., 2007. Appl. Catal. B: Environ. 69, 138. Rengifo-Herrera, J.A., Pulgarin, C., 2010. Sol. Energy 84, 37. Sakatani, Y., Nunoshige, J., Ando, H., Okusako, K., Koike, H., Takata, T., Kondo, J.N., Hara, M., Domen, K., 2003. Chem. Lett. 32, 1156. Sharma, M., Jain, T., Singh, S., Pandey, O.P., 2012. Sol. Energy 86, 626. Umebayashi, T., Yamaki, T., Itoh, H., Asai, K., 2002. Appl. Phys. Lett. 81, 454. Wang, X., Meng, S., Zhang, X., Wang, H., Zhang, W., Du, Q., 2007. Chem. Phys. Lett. 444, 292. Wang, C., Ao, Y., Wang, P., Zhang, S., Qian, J., Hou, J., 2010. Appl. Surf. Sci. 256, 4125. Wu, H.X., Wang, T.J., Jin, Y., 2006. Ind. Eng. Chem. Res. 45, 1337. Xiao, Q., Zhang, J., Xiao, C., Si, Z., Tan, X., 2008. Sol. Energy 82, 706. Yang, X., Cao, C., Erickson, L., Hohn, K., Maghirang, R., Klabunde, K., 2009. Appl. Catal. B: Environ. 91, 657. Yu, J., Zhao, X., 2001. Mater. Res. Bull. 36, 97. Zhao, W., Ma, W.H., Chen, C.C., Zhao, J.C., Shuai, Z.G., 2004. J. Am. Chem. Soc. 126, 4782.