SiO2 nanocomposites—Phase transformations and photocatalytic studies

SiO2 nanocomposites—Phase transformations and photocatalytic studies

Colloids and Surfaces A: Physicochem. Eng. Aspects 361 (2010) 25–30 Contents lists available at ScienceDirect Colloids and Surfaces A: Physicochemic...

949KB Sizes 1 Downloads 45 Views

Colloids and Surfaces A: Physicochem. Eng. Aspects 361 (2010) 25–30

Contents lists available at ScienceDirect

Colloids and Surfaces A: Physicochemical and Engineering Aspects journal homepage: www.elsevier.com/locate/colsurfa

New TiO2 /SiO2 nanocomposites—Phase transformations and photocatalytic studies A. Nilchi a,∗ , S. Janitabar-Darzi a , A.R. Mahjoub b , S. Rasouli-Garmarodi a a b

Nuclear Science and Technology Research Institute, P.O. Box 11365-8486, Tehran, Iran Department of Chemistry, Faculty of Sciences, Tarbiat Modares University, 14115-175 Tehran, Iran

a r t i c l e

i n f o

Article history: Received 23 October 2009 Received in revised form 3 March 2010 Accepted 4 March 2010 Available online 11 March 2010 Keywords: TiO2 /SiO2 nanocomposites Sol–gel process Photocatalyst Congo Red Photodegradation

a b s t r a c t A series of TiO2 /SiO2 nanocomposites were synthesized via a novel sol–gel method. A phase transformation study of as-synthesized nanocomposite due to thermal treatment up to 1100 ◦ C has showed the existence of anatase phase in all the tested temperatures. At thermal treatment, when the temperature exceeded 400 ◦ C, brookite phase was formed beside anatase phase. At 950 ◦ C, the amorphous silica matrix was transformed to crystobalite and the brookite phase disappeared. Finally, the small peaks of rutile phase were detectable at 1100 ◦ C. The presence of tetrahedral coordination of TiO2 in SiO2 matrix was confirmed by UV–vis study. The produced nanocomposites have good photocatalytic properties due to its anatase phase, tetrahedral coordination of TiO2 in SiO2 matrix and very large surface area. The photocatalytic properties of the composites were compared for the degradation of Congo Red (CR) azo dye. Further studies were also devised to compare the catalytic efficiency of the composite with the synthesized pure TiO2 . The results revealed that the as-prepared composite is the most effective. © 2010 Elsevier B.V. All rights reserved.

1. Introduction TiO2 /SiO2 composites may be applied as efficient catalysts in various chemical reactions. Hence, a number of researches on this composite have been performed [1,2]. In the case of TiO2 /SiO2 , the addition of SiO2 not only reduces the particles size and increases the specific surface area but also enhances the thermal stability of TiO2 particles against anatase to rutile phase transformation [3–5]. It has also been reported that the photocatalytic reactivity of TiO2 /SiO2 nanocomposites is highly dependent on the Ti/Si ratios [6]. By considering both the photocatalytic activity and the mechanical stability, the addition of about 50% SiO2 seems to be the best choice [7,8]. TiO2 has three kinds of crystal phase: anatase, rutile and brookite. Anatase-type titania exhibits high photocatalytic activity and thus has recently attracted a great deal of attention in the field of photocatalysts for decomposition of environmental pollutants and antibacterial applications [9]. The rutile phase has been found to be rarely active for the photodegradation of organic species in aqueous solutions [10] and only few studies have examined the photocatalytic activity of brookite TiO2 [11]. However, anatase nanocrystals exhibit high photocatalytic activity, compared to the bulk crystals, as they facilitate diffusion of excited electrons and holes toward the surface before their recombination.

∗ Corresponding author. Tel.: +98 21 88003315; fax: +98 21 88003793. E-mail address: [email protected] (A. Nilchi). 0927-7757/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfa.2010.03.006

The sol–gel method is one of the most promising techniques to prepare ceramics. However, as-prepared ceramics by the sol–gel method are usually amorphous, and thus the heat-treatment generally needed to densify and crystallize them [12,13]. As to sol–gel derived TiO2 and its composites, a high temperature process over 400 ◦ C is required to form anatase nanocrystals. In the present study, nanosized pure anatase TiO2 and a series of TiO2 /SiO2 nanocomposites were synthesized via sol–gel method and preparation of anatase/SiO2 nanocomposite under atmospheric pressure at room temperature is achievable. These TiO2 /SiO2 nanocomposites have also been successfully applied as photocatalysts for the liquid-phase degradation of CR. The dependence of photocatalytic activity of TiO2 /SiO2 nanocomposites on some factors such as the type of TiO2 phase present into the composites, particles size, and the effect of BET surface area, are discussed in detail in order to provide vital information for the design and application of such highly efficient photocatalytic systems in the degradation of toxic compounds diluted in a liquid phase.

2. Experimental 2.1. Chemicals The chemicals used in this study were titanium tetrachloride (TiCl4 , 99.9%), Fluka, as a titanium precursor, tetraethylorthosilicate (TEOS, 98%), as silica source, Congo Red (C32 H22 N6 Na2 O6 S2 ), HNO3 (70 wt%, d = 1.42 g cm−1 ), NH4 OH (25 wt%), and anhydrous

26

A. Nilchi et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 361 (2010) 25–30

ethanol (C2 H5 OH) from Merck. All reactant were used without further purification. 2.2. TiO2 and TiO2 /SiO2 composite preparation TiO2 and TiO2 /SiO2 nanocomposites were prepared by the sol–gel method. For TiO2 preparation, TiCl4 was added drop wise to distilled water under vigorous stirring in an ice water bath. The hydrolysis reaction was highly exothermic and released HCl. The produced dispersion was treated by NH4 OH and pH adjusted to 7 (suspension T). The resulting solid was collected by filtration, washed with distilled water in order to remove the NH4 Cl side product, and dried at room temperature. The powder produced was denoted as TR. The TR powder was calcined for 1 h at 400, 600 and 900 ◦ C. The obtained samples were denoted as T400, T600 and T900. For composite preparation, suspension T was prepared again and filtered. It was washed with distilled water for several times. The precipitation was dispersed in 200 mL of 0.3 M HNO3 . The mixture was refluxed under vigorous stirring at 70 ◦ C for 16 h as titania sol was prepared. 25 mL of tetraethylorthosilicate was added drop wise to the above sol and stirred at 70 ◦ C. The resulting powder was filtered and washed with distilled water and then dried at room temperature. The prepared composite was denoted as TSR. The composite powder calcined for 1 h at 400, 600, 800, 9500 and 1100 ◦ C and the obtained samples were denoted as TS400, TS600, TS800, TS950 and TS1100. 2.3. Characterization Phase identification and crystallite size of the products were characterized by X-ray diffraction (XRD) obtained on Philips Xpert diffractometer using a scan rate of 2◦ /min and Cu K␣ line ( = 1.54056 Å) radiation with working voltage and current of 40 kV and 40 mA, respectively. The crystallite size of the powders was determined by Scherrer equation [14]. Percent of each component of composite was determined by X-ray fluorescence (XRF) spectroscopy using Oxford ED 2000. Spectroscopic analysis of the ceramics was performed using a Fourier transform infrared spectroscopy (FT-IR) spectrometer (Perkin–Elmer 843) and UV–vis spectrophotometer (Shimadzu UV 2100). Brunauer–Emmett–Teller (BET) specific surface area and Barret–Joyner–Halenda (BJH) pore size distribution of the sample was determined through nitrogen adsorption (Quantachrome NOVA 2200 e). Thermogravimetry–differential scanning calorimetry (TG–DSC) was carried out using STA 150 Rhenometric Scientific Unit. Measurements were taken with a heating rate of 10 ◦ C/min from 25 to 800 ◦ C in an argon atmosphere. The morphology of the products was studied by scanning electron microscopy (SEM, Philips XL30) and transmission electron microscopy (TEM, Philips-EM208S). 2.4. Photocatalytic reaction For all photocatalytic experiments, a cylindrical glass was used as the reactor, which was filled with 300 mL of aqueous suspension of CR (5 ppm) containing 1 g/L of photocatalysts. It was an open Pyrex glass tube with double walls, so that a jacket of water was cooling the reactor to constant temperature of 25 ± 1 ◦ C. A lowpressure mercury lamp (18 W) with main wavelength at 254 nm (Philips) was used as the UV source. It was placed in a quartz vessel and immersed in center of photo reactor. Air was continuously bubbled into the solutions by an aquarium pump in order to provide a constant source of dissolved oxygen. Prior to irradiation, the suspension was magnetically stirred in the dark for approximately

Fig. 1. (a) XRD spectra of as-prepared TiO2 and TiO2 samples heat-treated at different temperatures. () Peaks due to anatase and () peaks due to rutile. (b) XRD spectra of as-prepared TiO2 /SiO2 composite and samples heat-treated at different temperatures. () Peaks due to anatase, () peaks due to rutile, () peaks due to brookite, and (䊉) peaks due to SiO2 (crystobalite).

30 min to ensure establishment of an adsorption/desorption equilibrium among the photocatalyst particles, CR, and atmospheric oxygen. During the course of UV irradiation, a suspension of about 5 mL was taken out after regular intervals, centrifuged, and then filtered through a millipore filter. The filtrates were then studied by UV–vis spectroscopy. 3. Results and discussion The crystalline structures of synthesized samples which were prepared at different annealing treatments were investigated by XRD measurements. The XRD patterns obtained for TiO2 samples and TiO2 /SiO2 nanocomposites are shown in Fig. 1a and b, respectively. According to Fig. 1a, as-prepared TiO2 powder is initially amorphous (TR) and converts to a crystalline anatase at higher temperatures. The samples T400 and T600 show only anatase phase to be present; whereas the sample T900 show all of anatase phase transforming to rutile one, as the annealing temperature increase. The XRD patterns of the synthesized composites (Fig. 1b) show that as-synthesized TiO2 /SiO2 nanocomposite (TSR) without calcinations, has crystalline anatase phase in amorphous silica matrix. Amorphous silica transforms to crystobalite at 950 ◦ C but anatase phase is present in all of the synthesized composite samples up to 1100 ◦ C. The XRD patterns related to the TS400, TS600 and TS800 exhibit diffraction lines attributed to brookite phase beside the anatase. On the other hand, heat treatment at temperature above 400 ◦ C makes small portion of anatase transform to brookite. The existence of brookite in XRD patterns is clearly evidenced from the presence of the (1 2 1) peak at 2 = 30.81◦ . In order to interpret the diffractograms, it is necessary to take into account that the

A. Nilchi et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 361 (2010) 25–30

27

Table 1 Anatase grain size of prepared samples and surface area of prepared composites versus calcination temperature. Heat treatment temperature (◦ C) Anatase grain size in composites (nm) Anatase grain size in TiO2 samples (nm) Surface area of composites (m2 /g)

Room temperature 5 – 707.6

main (1 0 1) diffraction peak of anatase at 2 = 25.28◦ overlaps with the (1 2 0) and (1 1 1) peaks of brookite at 2 = 25.34◦ and 25.69◦ , respectively [15]. In samples prepared at higher calcination temperature, the brookite phase disappears. Both the interactions Si–O–Ti and the high dispersion, prevent the crystalline transition to rutile phase [16] and only at very high temperature (1100 ◦ C), small peaks of rutile phase appears in the nanocomposite. The rutile phase is the stable form and can be produced by the heat treatment of metastable anatase or brookite. The sizes of the anatase crystallites in the prepared TiO2 and TiO2 /SiO2 nanocomposite samples measured by peak half-width according to the Scherrer equation, and are shown in Table 1. Comparing T400 and TS400, samples prepared at 400 ◦ C, as the molar ratio of SiO2 to TiO2 increased from 0 to about 1/1, the crystallite size of anatase TiO2 decreased from 9 to 5.09 nm. This result suggested that the doping of SiO2 into TiO2 could effectively retard the growth of nanoparticles and thus reduce the particle size. This observation may have resulted from the formation of the Ti–O–Si bond and due to the presence of amorphous SiO2 around TiO2 , which would prevent the growth of TiO2 particles [17]. The XRF analysis shows that the composite consists of 50% TiO2 and 46% SiO2 . Fig. 2 shows the FT-IR spectra of prepared TiO2 sample T400, and TiO2 /SiO2 nanocomposites (TSR, TS400, TS600, TS800 and TS1100), respectively. In the composite samples, three characteristic bands can be observed at around 1100, 950, and 490 cm−1 . The band at around 1100 cm−1 can be assigned to the stretching of the Si–O–Si bond of the SiO2 matrixes. TiO2 and TiO2 /SiO2 nanocomposites exhibit a band at around 490 cm−1 which is representative of TiO2 matrixes, while the composites exhibit an additional band at around 950 cm−1 . This band has been assigned to the stretching of the Si–O¯ species of Si–O–Ti or Si–O defect sites which are formed by the inclusion of Ti4+ ions into the SiO2 matrixes [18,19]. Thus, the appearance of the band at around 950 cm−1 indicates that the

Fig. 2. FT-IR spectra of pure TiO2 (T400) and TiO2 /SiO2 composite samples (TSR, TS400, TS600, TS800, and TS1100).

400 5.1 9 405.3

600 5.6 11.6 373.6

800 7.8 – 206.5

950 15.1 – –

1100 26.7 – 14.8

titanium oxide species are embedded into SiO2 matrixes within the TiO2 /SiO2 nanocomposites. N2 adsorption–desorption isotherms and pore size distribution analysis of the pure TiO2 and the composite materials prepared at different calcination temperatures are illustrated in Fig. 3. Isotherms of T400 (TiO2 ), TS400, TS600, TS800 and TS1100 composites are of type IV which indicates the presence of mesopore. However in the case of TS400, TS600 and TS800, the hysteresis loop is of a H2 type with a triangular shape and a steep desorption branch. Such behavior was observed for many porous inorganic oxides and was attributed to the pore connectivity effects [20]. Indeed, H2 hysteresis loops were observed for materials with relatively uniform channel-like pores, when the desorption branch happened to be located at relative pressures in the proximity of a lower pressure limit of adsorption–desorption hysteresis [21]. The shape of the hysteresis loop for these nanocomposites clearly indicates some pore blocking, as the almost horizontal plateau of adsorption ends. Therefore, it can be concluded that the porosity of the samples is characterized by some cavities which are connected with each other and with the external surface via narrow pores, the so-called ink-bottle type of porosity. The cavities are being created by the partial decomposition of the organosiliceous component of the material. The slightly lower surface area and pore volume of the samples calcined at the higher temperature of 400 ◦ C compare to TS400 is due to some additional thermal sintering or destruction. For calcined composite at 1100 ◦ C (TS1100), hysteresis loop is an H4 type. Type H4 loops feature parallel and almost horizontal branches and their occurrence has been attributed to adsorption–desorption in narrow slit-like pores [22]. The surface area of TiO2 /SiO2 composite samples calculated from BET is listed in Table 1. The surface area decreases from 707.6 to 14.8 m2 /g with increasing calcination temperature up to 1100 ◦ C. The decrease of the surface area upon calcination is due to the crystallization of the walls separating mesopores. For comparison purposes, the surface area of pure TiO2 was also calculated to be 90.4 m2 /g. It can be seen that the surface area of TS400 is much larger than that of T400 (pure TiO2 ) prepared at the same temperature. Pore diameter of T400 (pure TiO2 ) is 3.67 nm according to its BJH plot. TS400, TS600, and TS800 composites, possess a similar pore diameter, which by means of BJH plots were found to be 4.65 nm. However, TS1100 have a smaller pore diameter in comparison to the other prepared composites, which could be due to agglomerated particles and generation of crystobalite SiO2 and rutile TiO2 in the composite. The BJH plot of TS1100 also shows that calcination at very high temperature causes the broadening of the distribution. Differential scanning calorimetry and thermogravimetric curves of as-synthesized TiO2 /SiO2 composite (TSR) are shown in Fig. 4. The decrease in weight up to 150 ◦ C is attributed to desorption of physisorbed water and organic residue (confirmed by an endothermic peak on the DSC curve at about 100 ◦ C). A diffuse exotherm at around 400 ◦ C is the result of the small portion of anatase crystallizing to brookite. Above 400 ◦ C, the change in weight is very small. Fig. 5 shows the typical TEM image of as-prepared TiO2 /SiO2 nanocomposite (TSR) and SEM images of calcined nanocomposites. It can be seen from TEM image that granular TiO2 nanocrystal-

28

A. Nilchi et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 361 (2010) 25–30

Fig. 3. N2 -adsorption–desorption isotherms and BJH pore size distribution curve of T400, TS400, TS600, TS800 and TS1100.

lites (deep dark spots) are dispersed in the amorphous SiO2 matrix. The SEM images show that all the composites have agglomerated grainy structure and the particles size of the samples increased with increasing calcination temperature. Fig. 6 shows the absorption spectra (in the range 200–800 nm) of the as-prepared samples dispersed in ethanol. The band gap of the samples was calculated from the straight part of the optical absorption spectra [23]. By this method, the band gap of the pure anatase, T400, was calculated to be 3.24 eV. Comparing band gap of the pure TiO2 and composites revealed that by introducing SiO2 into the TiO2 matrix, a remarkable shift in the absorption band toward shorter wavelength regions is obtained. Such a large shift toward shorter wavelengths for the TiO2 /SiO2

nanocomposites is attributed to the size quantization effect arising from the presence of extremely small titanium oxide particles and/or the presence of highly dispersed titanium oxide species having a low coordination number [24]. A clear red shift in the absorption edges of the composites is observed by increasing of annealing temperature as shown in Fig. 6. Furthermore, the inset shows that the optical band gap of nanocomposites decreases (4.25–3.82 eV) by increasing temperature to 1100 ◦ C. The shift can be ascribed to the difference in grain size of the samples. Zribi et al. have reported a similar trend with regard to optical band gap and temperature and suggested that the variation of density and the structural modifications may have caused the changes in the shape of the fundamental absorption edge [25].

A. Nilchi et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 361 (2010) 25–30

29

Fig. 4. TG–DSC curves for the as-prepared TiO2 /SiO2 nanocomposite (TSR).

This spectral technique has also been employed to characterize the chemical nature and the coordination sphere of Ti ions incorporated into the silica matrix. A peak corresponding to isolated Ti species which have absorption maxima at 210–225 nm is observed in composite spectra. The absorption peak at 200–260 nm can be attributed to the charge transfer absorption process involving an

Fig. 6. UV–vis absorption spectra of the pure TiO2 (T400) and nanocomposites (TSR, TS400, TS600 and TS1100) dispersed in ethanol. Inset: changes of optical band gap of the composites with calcination temperature.

Fig. 5. TEM image of as-prepared composite (TSR) and SEM images of calcined composite TS400, TS600, TS800, TS950 and TS1100.

30

A. Nilchi et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 361 (2010) 25–30

the photocatalytic activity. The brookite phase exists in the TS400 and TS800 composites; however the photocatalytic properties are not affected. 4. Conclusion

Fig. 7. Variation of photocatalytic activity of TiO2 and TiO2 /SiO2 composite heattreated at different temperatures.

electron transfer from the O2− to Ti4+ ions of the highly dispersed tetrahedrally coordinated TiO4 unit of the catalysts [26,27]. Anpo and co-workers have reported that titanium oxides having a tetrahedral coordination can be chemically supported onto silica matrix and have shown that such catalysts exhibit highly photocatalytic reactivities [27]. 3.1. Photocatalytic activity Photocatalytic degradation of Congo Red, cause its absorption bands to decrease with time, indicating the destruction of its structure. Fig. 7 shows photocatalytic efficiency of the composites and pure TiO2 (T400) for removal of CR from aqueous solution as a function of time at  = 497 nm. The efficiency or degree of photodegradation (X) is given by: X = (C0 − C)/C0 , where C0 is the initial concentration of dye, and C the concentration of dye at time T. Fig. 7 reveals that the as-prepared composite is the most effective photocatalyst. CR molecules are strongly adsorbed and degraded on the surface of TSR so that more than 99% of dye decolorization is performed in 60 min. The order of efficiency in applied photocatalyst is TSR > TS400 > T400 > TS800 > TS1100. These results are in a good conformity to the BET surface area of the composite samples. As a result of thermal treatment at 1100 ◦ C, the activity of composite significantly decreases due to the noticeable reduction of its specific surface area. In the composites, the decrease of anatase particle size leads to a larger surface area, and consequently the number of available surface active sizes increase. A reduction in particle size should also lead to a high photonic efficiency favoring a higher interfacial charge carrier transfer rate. The samples calcined at 800 and 1100 ◦ C are weaker photocatalyst than the pure synthesized TiO2 . It can be deduced from the results obtained that the high content (100%) of anatase phase of TiO2 in T400 would be the reason of its photocatalysis superiority in comparison with TS800 and TS1100, although the sample calcined at 800 ◦ C has a larger surface area. Photocatalytic activity of the TiO2 catalysts not only depends upon BET specific surface area, but also on several factors including, crystallinity, crystallite size and crystal structure. The contemporaneous presence of different phases of the same semiconductor is usually beneficial to enhance

Nanosized TiO2 and a series of TiO2 /SiO2 nanocomposites with spherical shape particles were synthesized via sol–gel process and used as photocatalysts. In the TiO2 /SiO2 composites, TiO2 nanocrystals are highly dispersed in the amorphous SiO2 matrix and the formation of Ti–O–Si bond and amorphous SiO2 in TiO2 /SiO2 could effectively increase the stability of anatase TiO2 , limit the growth of crystallites, and as a result significantly increase the surface area. Such an increase in the surface area is expected to improve the photocatalytic activities of the TiO2 /SiO2 ceramic. In the prepared nanocomposites, the isolated titanium oxide species in tetrahedral coordination are present as a separate entity in the SiO2 matrix which can be another reason for its photocatalytic properties. Finally, it can be deduced from the results that the calcination temperature govern on morphology, size and structure of the TiO2 nanoparticles dispersed in SiO2 matrix and also affects the photochemical properties of the nanocomposites. References [1] Y. Kotani, A. Matsuda, T. Kogure, M. Tatsumisago, T. Minami, Chem. Mater. 13 (2001) 2144–2149. [2] H. Chun, W. Yizhong, T. Hohgxiao, Appl. Catal. B: Environ. 30 (3–4) (2001) 277–285. [3] D.J. Reidy, J.D. Holms, M.A. Morris, Ceram. Int. 32 (3) (2006) 253. [4] K. Wang, Y.X. Chen, F.X. Ye, Chin. J. Catal. 24 (12) (2004) 931. [5] Z. Xu, J. Liu, E. Wang, C. Qin, Q. Wu, Q. Shi, J. Mol. Struct. 873 (2008) 41. [6] S. Dohshi, M. Takeuchi, M. Anpo, Catal. Today 85 (2003) 199–206. [7] C.H. Kwon, J.H. Kim, I.S. Jung, H. Shin, K.H. Yoon, Ceram. Int. 29 (2003) 851–856. [8] L. Zhou, S. Yan, B. Tian, J. Zhang, M. Anpo, Mater. Lett. 60 (2006) 396–399. [9] A.L. Linsebigler, G. Lu, J.T. Yates, J. Chem. Rev. 95 (1995) 735–758. [10] M. Addamo, M. del Arco, M. Bellardita, D. Carriazo, A. Di Paola, E. García-López, G. Marcì, C. Martín, L. Palmisano, V. Rives, Res. Chem. Intermed. 33 (2007) 46–479. [11] A.D. Paola, G. Cufalo, M. Addamo, M. Bellardita, R. Campostrini, M. Ischia, R. Ceccato, L. Palmisano, Colloids Surf. A: Physicochem. Eng. Asp. 317 (2008) 366–376. [12] Y. Abe, N. Sugimoto, Y. Nagao, T. Misono, J. Non-Cryst. Solids 104 (1988) 164. [13] R.N. Viswanath, S. Ramasamy, Colloids Surf. A: Physicochem. Eng. Asp. 133 (1998) 49. [14] P. Peshev, I. Stambolova, S. Vassilev, P. Stefanov, V. Blaskov, K. Starbova, N. Starbov, Mater. Sci. Eng. B 97 (2003) 106–110. [15] D.P. Agatino, C. Giovanni, A. Maurizio, B. Marianna, C. Renzo, I. Marco, C. Riccardo, P. Leonardo, Colloid Surf. A: Physicochem. Eng. Asp. 317 (2008) 366–376. [16] J. Aguado, R. Grieken, M. Lopez-Munoz, J. Marugan, App. Catal. A: Gen. 312 (2006) 202–212. [17] M. Machida, K. Norimoto, T. Watanabe, K. Hashimoto, A. Fujishima, J. Mater. Sci. 34 (11) (1999) 2569. [18] R.J. Davis, Z. Liu, Chem. Mater. 9 (1997) 2311. [19] M.A. Camblor, A. Corma, J. Pérez-Pariente, J. Chem. Soc., Chem. Commun. (1993) 557. [20] H. Liu, L. Zhang, N.A. Seaton, J. Colloid Interf. Sci. 156 (1993) 285. [21] M. Kruk, M. Jaroniec, A. Sayari, A. Langmuir. 13 (1997) 6267. [22] M. Kruk, M. Jaroniec, Chem. Mater. 13 (2001) 3169–3183. [23] M.N. Kamalasanan, S. Chandra, Thin Solid Films 288 (1996) 112–115. [24] H. Yamashita, S. Kawasaki, Y. Ichihashi, M. Harada, M. Takeuchi, M. Anpo, J. Phys. Chem. B 102 (1998) 5870–5875. [25] M. Zribi, M. Kanzari, B. Rezig, Thin Solid Films 516 (2008) 1476–1479. [26] X. Yan, J. He, D.G. Evans, Y. Zhu, X. Duan, J. Poro. Mat. 11 (2004) 131–139. [27] H. Yamashita, Y. Ichihashi, M. Harada, G. Stewart, M.A. Fox, M. Anpo, J. Catal. 158 (1996) 97.