Chemosphere 44 (2001) 1087±1092
Photocatalytic activity of TiO2 ®lms grown on dierent substrates Ying Ma, Jian-bin Qiu 1, Ya-an Cao, Zi-shen Guan, Jian-nian Yao
Center for Molecular Science, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100080, People's Republic of China Received 3 May 2000; accepted 24 August 2000
Abstract Titanium dioxide ®lms were prepared on glass, indium±tin oxide (ITO) glass and p-type monocrystalline silicon and studied for the photocatalytic degradation of rhodamine B in an aqueous medium. Raman, AFM, and XPS spectroscopic investigations of these ®lms indicated that microstructure of titanium oxide ®lms were greatly aected by the substrate materials. Rutile was con®rmed to be easily formed on the surface of ITO glass, and TiO2 tended to grow as closely packed particles that were elongated strips with an average size of 20 nm, and had lovely contrast with the perfectly round particles grown on p-type monocrystalline silicon. Charge transfer between the ®lm and silicon substrate was veri®ed by surface photovoltage spectra. This may be the real reason why the ®lms grown on ITO glass and silicon substrates exhibit higher photocatalytic reactivity than the ®lm on glass substrate. Moreover, the dierent surface properties also seem to be responsible for the dierent activity. Ó 2001 Elsevier Science Ltd. All rights reserved. Keywords: Photocatalysis; Rhodamine B; Surface properties; Surface photovoltage spectra; Charge transfer
1. Introduction Heterogeneous semiconductor photocatalysis, as a low-temperature nonenergy-intensive approach for chemical waste destruction, represents an emerging area of environmental catalysis (Fox and Dulay, 1993; Homann et al., 1995; Davis and Green, 1999). Many works (Peral et al., 1988; Yu et al., 1998; Paola et al., 1999; Yumoto et al., 1999) on photocatalytic processes using a variety of semiconductor materials have been *
Corresponding author. Tel.: +86-10-6488-8154; fax: +8610-6487-9375. E-mail address: [email protected]
(J.-n. Yao). 1 Graduate student from Department of Chemistry, Fujian Normal University. 2 Present address: Institute of Photographic Chemistry, Center for Molecular Science, Academica Sinica, Chinese Academy of Sciences, Beijing 100101, Peoples's Republic of China.
developed since photochemical conversion and light harvesting of TiO2 materials were ®rst introduced in the 1970s (Fujishima and Honda, 1972). But it was not until 1983 that organic pollutant oxidative mineralization sensitized by semiconductor was clearly recognized for the ®rst time (Pruder and Ollis, 1983). Due to its stability, nontoxicity, and broadly de®ned goal of eciently detoxifying hazardous organic pollutants, TiO2 has been studied extensively over the last two decades, and the mechanism of TiO2 photocatalysis has also been studied in detail by many in situ experimental methods recently (Hwang et al., 1998; Nosaka et al., 1998; Maeda et al., 1999). But the application of heterogeneous photocatalysis for large-scale wastewater treatment has not advanced rapidly and low rates of oxidation and lack of pilot plant studies to validate models for the scaleup, design, and optimization of photocatalytic reactors are the main reasons (Feitz et al., 1999). To decrease the cost of the reactor the ®ltration and suspension of small particle photocatalysts are needed to avoid.
0045-6535/01/$ - see front matter Ó 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 5 - 6 5 3 5 ( 0 0 ) 0 0 3 6 0 - X
Y. Ma et al. / Chemosphere 44 (2001) 1087±1092
Furthermore, there is a growing interest in the use of thin, transparent ®lm of TiO2 for generating a selfcleaning surface presently (Fujishima et al., 1999a), so the design of highly ecient and transparent titanium oxide ®lm photocatalyst is strongly desired. In order to obtain a high reaction rate, the recombination rate of photogenerated electrons and holes must be kept low. An ecient charge separation can be obtained by coupling two semiconductor particles (Spanhel et al., 1987; Gopidas et al., 1990; Paola et al., 1999), coupled semiconductor systems were usually acquired by mixing semiconductor powders (Serpone et al., 1995) or by deposition of semiconductor ®lms on substrates (Liu et al., 1993). In the present work, we prepared TiO2 ®lms on indium±tin oxide (ITO) glass and p-type monocrystalline silicon to study if these coupled semiconductor systems composed of substrates and TiO2 ®lms grown on them could enhance the reaction eciency. It is also well known that many other variables such as crystal structure, particle size, surface area and porosity greatly in¯uence the photocatalytic eciency of TiO2 . So, some surface properties of the ®lms were investigated in detail and the comparison was made. Dyestus constitute a large group of toxic pollutants in the world, and early studies (Vinodgopal et al., 1996a; Zhao et al., 1998) con®rmed that photodegradation of dyes under visible light in TiO2 suspension makes it possible to remedy dyestu euents using photocatalytic method. So rhodamine B, one of xanthene dyes is used as the target compound to carry out photocatalytic experiments in this work. 2. Materials and methods
2.2. Catalyst characterization The Raman spectra of the resulting TiO2 ®lms were recorded on a Raman spectrometer (Renishaw model 2000). The exciting light used was 514.5 nm supported by Ar laser with the power on samples less than 5 mW. The XPS spectra were recorded by an XSAM 300 multifunctional surface analysis spectrometer (KRATO, UK) using a standard Al Ka source. Detailed spectra were recorded for the following regions: C1s, O1s and Ti2p, and the C1s peak of hydrocarbon contamination set at 284.8 eV was used. Surface morphology of the catalyst was observed on atomic force microscopy (AFM, Park Scienti®c Apparatus). A home-built surface photovoltage spectrometer (SPS) (Wang et al., 1989a) was used to acquire the SPS spectra of the ®lms on dierent substrates. 2.3. Photoreactivity experiments Photodegradation of rhodamine B in an aqueous medium was used as a ``probe'' reaction to assess the photocatalytic activity of all the ®lms. Ten ml of rhodamine B aqueous solution was placed in a cylindrical glass reactor to carry out the experiments, and the starting concentration of the solution was 1 10 5 mol l 1 . A 500 W high pressure mercury lamp cooled with a water jacket was used as the irradiation source and placed 12 cm apart from the reactor. A 2:5 cm 2 cm TiO2 ®lm was applied as photocatalyst in all experiments. Under irradiation, the solution was stirred continually and open to air.
2.1. Catalyst preparation
3. Results and discussion
TiO2 thin ®lms on dierent substrates were prepared by dip-coating method similar to our previous work (Ma and Yao, 1998). Some molar acetylacetone was added in tetrabutyl titanate ethanolic solution and followed by magnetic stirring for 3 h to complete the chelation. The mixture was further stirred for 1 h after polyethylene glycol (600 of mol. wt.) and water were introduced. Stable deep-yellow coating solution was then obtained after ®ltration. Clean glass, ITO glass and p-type monocrystalline silicon (p-Si, (1 1 1) surface) were impregnated in the above coating solution and followed by pulling up at 1.5 mm s 1 . Dried precursor ®lms on different substrates were ®rst heated at 250°C in air to remove organic counterparts. The process of impregnation, drying and presintering was repeated for several times to acquire relative thick amorphous ®lms, then the ®lms were sintered at 450°C in the air for 45 min to obtain the ®nal catalysts.
3.1. Structural properties The Raman spectra of the ®lms supported on different substrates are shown in Fig. 1. For TiO2 /G and TiO2 /p-Si (line a and b), ®ve peaks were observed: 142, 196, 396, 518, and 638 cm 1 . All of these peaks can be attributed to anatase (Kelly et al., 1997), and no other peaks can be observed. Whereas, not only did all of the above peaks shift and broaden in Fig. 1(c), but also a new band around 440 cm 1 attributable to rutile (Kelly et al., 1997) was observed. That is to say TiO2 ®lm grown on ITO glass substrate consists of both anatase and rutile. Now that TiO2 /G and TiO2 /p-Si prepared under the same conditions consist of only anatase, the results indicate that ITO possibly promotes the anatase± rutile transformation in some sense. In addition, the broadening and shift of the peaks as shown in Fig. 1(c) are also possibly due to the coexistence of both crys-
Y. Ma et al. / Chemosphere 44 (2001) 1087±1092
sorbed on the surfaces. No dierence was observed in Ti2p XPS spectra, which are not shown here. The above results show that TiO2 surfaces grown on ITO and p-Si are more close to defect-free than that grown on glass substrate. 3.2. AFM results
Fig. 1. The Raman spectra of dierent ®lms: (a) TiO2 =G; (b) TiO2 =p-Si; (c) TiO2 =ITO.
talline phases, and small grain size is another possible reason (Kelly et al., 1997). Fig. 2 presents O1s XPS spectra for TiO2 ®lms. For TiO2 /ITO and TiO2 /p-Si, we observed almost symmetric O1s peaks, while an obvious broadening at a higher binding energy was observed for TiO2 /G. Previous studies (Wang et al., 1999b) have assumed the O1s peak broadening to be due to the hydroxyl species (OH) ad-
Fig. 2. The O1s XPS spectra for TiO2 =G, TiO2 =p-Si, and TiO2 =ITO.
The AFM images of the three kinds of ®lms are shown in Fig. 3. As shown in Fig. 3(a), nearly round particles of about 80 nm arrayed uniformly on the TiO2 / G surface. For TiO2 /p-Si, perfectly round particles ranging from 50 to 150 nm can be observed and the surface roughness is larger than the other ones. As for TiO2 /ITO surface, elongated particles are closely packed and it is dicult to discern the grain boundary. The grain size is about 20 nm and the surface roughness is small as seen in Fig. 3(c). Therefore, not only the crystalline phase was aected by the substrate materials, but also the surface morphology including the particle shapes and arrangement greatly depended on the supporting materials. 3.3. Photocatalytic studies Fig. 4 illustrates the photocatalyzed disappearance of rhodamine B in the presence of dierent samples. The results indicate that all of the reactions follow pseudo®rst-order kinetics. As shown in Fig. 4, TiO2 /G exhibits the lowest photoactivity under, otherwise, identical conditions, TiO2 /p-Si and TiO2 /ITO are much more ef®cient. Detailed results are shown in Table 1. It is shown that TiO2 /p-Si photocatalyses the oxidative conversion of rhodamine B at a twofold faster rate than TiO2 /G, and TiO2 /ITO is a more ecient catalyst than TiO2 /G by a factor of about 3. As shown above, the greatly dierent photoactivities for the three samples are possibly due to the dierence in their microstructures. A recent study has shown that the anatase/rutile phase interactions should not play a signi®cant role in promoting the photoreactivity for P25 powders (the widely used commercial TiO2 photocata-
Fig. 3. AFM micrographs of (a) TiO2 =G, (b) TiO2 =p-Si, and (c) TiO2 =ITO.
Y. Ma et al. / Chemosphere 44 (2001) 1087±1092
Fig. 4. Photocatalyzed conversion of rhodamine B: (a) TiO2 =G; (b) TiO2 =p-Si; (c) TiO2 =ITO.
lyst) since the majority of the anatase grains are not in contact with the rutile ones (Zhang et al., 1998). But in our work, rutile and anatase coexisted in TiO2 grains and no separation was observed in TiO2 /ITO, so the interaction between the two phases probably made a contribution to its high photoreactivity. In addition, it can be seen from surface morphologies of the samples that TiO2 /p-Si presents the highest surface roughness, which may enhance its photoactivity. On the other hand, however, surface adsorbed hydroxyl (OH) is believed to be one of the main oxidizing species and bene®t ecient photocatalytic reactions (Fox and Dulay, 1993; Homann et al., 1995). From the view of this point, TiO2 /G should be a better photocatalyst. Most of all, although there is so much dierence between TiO2 /p-Si and TiO2 / ITO, their activities to assist the decomposition of rhodamine B are relatively close, and both of them are much higher than TiO2 /G. So the enhanced rate of disappearance of rhodamine B cannot be justi®ed by a simple improved surface roughness or a mixed crystalline phase. It is well known that a vectorial transfer of electrons and holes from one semiconductor to another can result in the improvement of photocatalytic eciency in a coupled semiconductor system with favorable energetics (Hotchandani and Kamat, 1992). SnO2 /TiO2 composite ®lms have been demonstrated to enhance the photocat-
alytic degradation rate greatly in a previous study (Vinodgopal et al., 1996b). Similar charge transfer could have also happened at the interface of TiO2 /ITO since ITO possesses a band gap of 3.5 eV and similar band edge position to SnO2 . As shown in Fig. 5, lines a and b are the SPS spectrum and positive electric ®eld induced SPS (EFISPS) spectrum for TiO2 /ITO, respectively. It is evident that the response signal was enhanced by applied positive electric ®eld, downward band bending of the conduction and valence band at the interface between the ®lm and ITO substrate was therefore con®rmed. The accompanying peak noticeable in Fig. 5(b) is ascribable to some surface state energy levels. Band gap of 3.40 eV can be calculated from the threshold response of 365 nm in Fig. 5(a) (where the extrapolation of line a and the transverse are crossed), then the electron energy diagram as shown in Fig. 6 was acquired. Clearly, the electrons excited to conduction band from valence band of TiO2 by near UV light can ¯ow along the band bending to the conduction band of ITO substrate. Thus the photogenerated charges separate eciently, the holes on TiO2 surface then oxidize the adsorbents followed by the decomposition of rhodamine B. It should be also noted that it is dicult for the electrons in the ITO conduction band to be excited under our experimental conditions. As for TiO2 /p-Si, charge transfer can also happen according to the energy level correlation between
Fig. 5. The SPS and EFISPS spectra of TiO2 =ITO.
Table 1 Results of disappearance of rhodamine B for the three ®lms Sample
Film weight (mg)
Degradation ratioa , DC=C0
Rate constant, k
TiO2 =G TiO2 =p-Si TiO2 =ITO
0:58 0:02 0:52 0:02 0:50 0:02
0:256 0:011 0:528 0:020 0:634 0:019
4:93 0:10 10
1:25 0:04 10
1:68 0:07 10
3 2 2
Photodegradation ratio measured after reaction for 1 h. Degraded phenol measured after reaction for 1 h catalyzed by unit photocatalyst.
Half-time, t1=2 (min)
Speci®c photocatalytic activityb
mol g 1 h 1
1483 63 2 46 2
2:21 0:11 10
5:08 0:20 10
6:34 0:20 10
5 5 5
Y. Ma et al. / Chemosphere 44 (2001) 1087±1092
Fig. 6. Schematic diagram representing the charge transfer process in a TiO2 =ITO system.
monocrystalline Si and TiO2 (Fujishima et al., 1999b). The SPS spectrum of TiO2 /p-Si provided evidence of the charge transfer. The comparison between the SPS spectra for p-Si and TiO2 /p-Si is illustrated in Fig. 7. All the response signals in Fig. 6(a) stem from the indirect transition in p-Si, since its band gap is 1.11 eV. Except for the signal around 350 nm due to TiO2 , Fig. 7(b) resembles Fig. 7(a). It means that both photovoltaic responses for the ®lm and p-Si substrate were recorded on the TiO2 surface. It is known from the principles that the photovoltage for p-Si could not be observed on TiO2 surface if it were not for the charge transfer between TiO2 and p-Si substrate. Whereas, charge transfer from TiO2 to ITO cannot be observed due to the SPS experimental limitation. As shown in Fig. 8, photogenerated electrons ¯ow to TiO2 conduction band, while photogenerated holes tend to aggregate at the valence band of p-Si. The increase of the photoelectrons seems to make a signi®cant contribution to the improvement of photoreactivity for TiO2 /p-Si in consideration of the nonreactivity of the holes of p-Si. On the other hand, the majority of the valence band holes of TiO2 seems to be
Fig. 8. Schematic diagram representing the charge transfer process in a TiO2 =p-Si system.
trapped by the surface adsorbents to initiate the photooxidation reactions instead of aggregation in p-Si, since TiO2 /p-Si exhibits high photoreactivity. Further studies are needed to make sure the exact details.
4. Conclusion Substrate materials greatly aected the microstructure of the TiO2 ®lms supported on them, including the crystalline phase, particle shapes and the way the particles array. Anatase was the main phase in both TiO2 /G and TiO2 /p-Si samples, while mixed phase of anatase and rutile formed in TiO2 /ITO. Particles on TiO2 /G and TiO2 /p-Si surfaces are nearly and perfectly round, respectively, for TiO2 /ITO surface, however, elongated particles were observed. The photoreactivity is very dierent for the three samples, comparison of them gives the following decreasing order of activity: TiO2 =ITO > TiO2 =p-Si > TiO2 =G;
Fig. 7. The SPS spectra of (a) p-Si and (b) TiO2 =p-Si.
such sequence may owe to the following reasons. Charge transfer between the ®lm and the substrate seems to be the main reason why TiO2 /p-Si and TiO2 / ITO present high photoactivities, since the charge separation is more ecient in such systems. On the other hand, anatase/rutile interaction may also enhance the photoactivity of TiO2 /ITO, and the high photoactivity of TiO2 /p-Si can also be attributed to its high surface roughness. Above all, supporting materials were con®rmed to be an important role to aect the photoreactivity of the TiO2 ®lms. This will be helpful to develop new, highly ecient immobilized photocatalysts in further studies.
Y. Ma et al. / Chemosphere 44 (2001) 1087±1092
Acknowledgements This work was sponsored by the Chinese Academy of Sciences, National Science Foundation of China and `National Research Fund for Fundamental Key Projects No. 973 (G19990330)'. The authors would also like to thank Dr. Zhang Xintong for help with the AFM and SPS experiments. References Davis, A.P., Green, D.L., 1999. Photocatalytic oxidation of cadmium-EDTA with titanium dioxide. Environ. Sci. Technol. 33, 609±617. Feitz, A.J., Waite, T.D., Jones, G.J., Boyden, B.H., Orr, P.T., 1999. Photocatalytic degradation of the blue green algal toxin microcystin-LR in a natural organic-aqueous matrix. Environ. Sci. Technol. 33, 243±249. Fox, M.A., Dulay, M.T., 1993. Heterogeneous photocatalysis. Chem. Rev. 93, 341±357. Fujishima, A., Honda, K., 1972. Electrochemical photolysis of water at a semiconductor electrode. Nature 238, 37±38. Fujishima, A., Hashimoto, K., Watanabe, T., 1999a. TiO2 Photocatalysis Fundamentals and Applications, ®rst ed. BKC, Tokyo (Chapter 3). Fujishima, A., Hashimoto, K., Watanabe, T., 1999b. TiO2 Photocatalysis Fundamentals and Applications, ®rst ed. BKC, Tokyo, p. 128. Gopidas, K.R., Bohorquez, M., Kamat, P.V., 1990. Photophysical and photochemical aspects of coupled semiconductors: charge transfer processes in colloidal cadmium sul®de-titania and cadmium sul®de-silver (I) iodide systems. J. Phys. Chem. 94, 6435±6440. Homann, M.R., Martin, S.T., Choi, W., et al., 1995. Environmental applications of semiconductor photocatalysis. Chem. Rev. 95, 69±96. Hotchandani, S., Kamat, P.V., 1992. Modi®cation of electrode surface with semiconductor colloids and its sensitization with chlorophyll a. Chem. Phys. Lett. 191, 320±326. Hwang, S.-J., Petucci, C., Raftery, D., 1998. In situ solid-state NMR studies of trichloroethylene photocatalysis. Formation and characterization of surface-bound intermediates. J. Am. Chem. Soc. 120, 4388±4397. Kelly, S., Pollak, F.H., Tomkiewicz, M., 1997. Raman spectroscopy as a morphological probe for TiO2 aerogels. J. Phys. Chem. B 101, 2730±2734. Liu, D., Kamat, P.V., 1993. Photoelectrochemical behavior of thin cadmium selenide and coupled titania/cadmium selenide semiconductor ®lms. J. Phys. Chem. 97, 10769±10773. Ma, Y., Yao, J.N., 1998. Photodegradation of rhodamine B catalyzed by TiO2 thin ®lms. J. Photochem. Photobiol. A 116, 167±170. Maeda, H., Ikeda, K., Hashimoto, K., Ajito, K., Morita, M., Fujishima, A., 1999. Microscopic observation of TiO2
photocatalysis using scanning elecrochemical microscopy. J. Phys. Chem. B 103, 3213±3217. Nosaka, Y., Kishimoto, M., Nishino, J., 1998. Factors governing the initial process of TiO2 photocatalysis studied by means of in-situ electron spin resonance measurements. J. Phys. Chem. B 102, 10279±10283. Paola, A.D., Palmisano, L., Venezia, A.M., Augugliaro, V., 1999. Coupled semiconductor systems for photocatalysis. Preparation and characterization of polycrystalline mixed WO3 =WS2 powders. J. Phys. Chem. B 103, 8236±8244. Peral, J., Casado, J., Domenech, J., 1988. Light-induced oxidation of phenol over zinc oxide powder. J. Photochem. Photobiol. A 44, 209±217. Pruder, L., Ollis, D.F., 1983. Photoassisted heterogeneous catalysis: the degradation of trichloroethylene in water. J. Catal. 82, 404±417. Serpone, N., Maruthamuthu, P., Pichat, P., Pelizzetti, E., Hidaka, H., 1995. Exploiting the interparticle electron transfer process in the photocatalyzed oxidation of phenol, 5-chlorophenol and pentachlorophenol: chemical evidence for electron and hole transfer between coupled semiconductors. J. photochem. Photobiol. A 85, 247±255. Spanhel, L., Weller, H., Henglein, A., 1987. Photochemistry of semiconductor colloids. 22. Electron ejection from illuminated cadmium sul®de into attached titanium and zinc oxide particles. J. Am. Chem. Soc. 109, 6632±6635. Vinodgopal, K., Wynkoop, D.E., Kamat, P.V., 1996a. Environmental photochemistry on semiconductor surfaces: photosensitized degradation of a textile azo dye, acid orange 7, on TiO2 particles using visible light. Environ. Sci. Technol. 30, 1660±1666. Vinodgopal, K., Bedja, I., Kamat, P.V., 1996b. Nanostructured semiconductor ®lms for photocatalysis systems and its role in photocatalytic degradation of a textile azo dye. Chem. Mater. 8, 2180±2187. Wang, D.J., Liu, W., Xiao, L.Z., 1989a. Application of surface photovoltage spectroscopy in chemistry. Huaxue Tongbao 10, 32±37. Wang, L.-Q., Ferris, K.F., Skiba, P.X, et al., 1999b. Interactions of liquid and vapor water with stoichiometric and defective TiO2
110 surfaces. Surf. Sci. 440, 60±68. Yu, J.C., Lin, J., Kwok, R.W.M., 1998. Ti1 x Zrx O2 solid solutions for the photocatalytic degradation of acetone in air. J. Phys. Chem. B 102, 5094±5098. Yumoto, H., Inoue, T., Li, S.J., Sako, T., Nishiyama, K., 1999. Application of ITO ®lms to photocatalysis. Thin Solid Films 345, 38±41. Zhang, Z., Wang, C.-C., Zakaria, R., Ying, J.Y., 1998. Role of particle size in nanocrystalline TiO2 based photocatalysts. J. Phys. Chem. B 102, 10871±10878. Zhao, J., Wu, T., Wu, K., Oikawa, K., Hidaka, H., Serpone, N., 1998. Photoassisted degradation of dye pollutants. 3. Degradation of the cationic dye rhodamine B in aqueous anionic surfactant TiO2 dispersions under visible light irradiation: evidence for the need of substrate adsorption on TiO2 particles. Environ. Sci. Technol. 32, 2394±2400.