J O U R N A L O F C H E M I S T R Y Materials Preparation of a TiO2 based multiple layer thin film photocatalyst Atsuo Yasumori,* Kenichi Ishizu, Shigeo Hayashi and Kiyoshi Okada Department of Inorganic Materials, Tokyo Institute of Technology, Tokyo 152-8552, Japan. E-mail: ayasumor@ceram.titech.ac.jp Received 30th April 1998, Accepted 20th August 1998 ATiO2 thin filmsemiconductor is expected to show high photocatalytic activity because of the short diVusion distances of photoexcited electrons and holes to the surface as well as the short spatial separation of reducing and oxidizing sites.Themultiple layer photocatalyst, reported herein, which consists of a TiO2 thin film, platinumelectrode and porous alumina substrate, was prepared by spin-coating a TiO2 sol and sputter-coating of a platinum electrode on a porous substrate.After adequate heat treatment, thismulti-layered photocatalyst showed a high eYciency forH2 generation from ethanol aqueous solution under UV illumination as compared with platinized TiO2 fine particles. Research into photocatalytic reactions mediated by metal oxide semiconductors received a further boost after the work of Fujishima and Honda on the photoelectrolysis of H2O using TiO2 electrodes.1 Since that study was reported, a considerable amount of work involving metal oxide semiconductor photocatalysts has been done.To date, no substance superior to TiO2, however, has been found. TiO2 is especially attractive because of its high photocatalytic activity and chemical stability in aqueous solution under light irradiation.Recently, two branches involving the study of the TiO2 photocatalyst have emerged especially for the purposes of solving environmental issues. One branch involves the use of highly dispersed fine particles in a porous material,2–4 and the other is concerned with TiO2 thin films.5–10 Most TiO2 thin films are prepared by coating a substrate with a TiO2 sol, thereby producing a thin film consisting of TiO2 particles.Generally, this methodology often results in a porous TiO2 film. The fine particles that make up the film have several advantages, for example, large specific surface areas, short diVusion distance to the surface for the photo-generated electrons and holes, and quantum size eVects for particles of <10 nm in diameter.11However, some disadvantages of fine particles include Fig. 1 Schematic illustration of a TiO2 based multiple layer thin film problems associated with the close proximity of the redox sites photocatalyst. due to small particle sizes, i.e., the progress of reverse reactions between the products (oxidants and reductants) on the surfaces of particles is hard to inhibit. Experimental Nonetheless, an attractive photocatalytic system for spatially separating the respective oxidation and reduction sites is an The TiO2 thin film was prepared by a sol–gel method.The unsupported thin film system consisting of a semiconductor flowchart for the preparation of the TiO2 sol precursor is and an appropriate metal electrode. This system is also shown in Fig. 2. This procedure was adapted from the method expected to minimize diVusion distances to the surface for the reported by Sakka and Kamiya.16 Reagent grade titanium photogenerated charge carriers.The fabrication of such a thin tetraisopropoxide (3.98 g, TTIP, Wako Pure Chemicals) was film system on the nanometer or mesoscopic scale is obviously stabilized with 0.33 ml of acetylacetone (AcAc, Wako) and very challenging.There are some reports on TiO2 thin film 130 ml of isopropanol (PriOH, Wako). Stabilization of TTIP photocatalysts supported on metal electrodes12–14 or on porous ensures that during hydrolysis only isopropoxyl groups are substrate.15 However, it is not clear whether these photocata- removed, and consequently this prevents the rapid growth of lytic systems were intended to spatially separate photo- the TiO2 particles.Stabilized TTIP was hydrolyzed with 0.98 g generated electrons and holes or not. of acetic acid (Wako), 0.76 ml of distilled water and 10 ml of In this work, a photocatalyst comprising a TiO2 thin film PriOH under an argon atmosphere resulting in a transparent and platinum electrode was fabricated on a silica gel coated yellowish TiO2 sol.The molar ratio of TTIP5AcAc5H2O was porous alumina substrate as schematically illustrated in Fig. 1. 150.2353. The particle sizes in the TiO2 sol were <10 nm in In this multiple layer system, UV light irradiation is expected diameter as observed by transmission electron microscopy. to induce the following redox reaction; oxidation on the A flowchart for the fabrication of the thin film photocatalyst surface of the TiO2 thin film and reduction on the metal is shown in Fig. 3. An alumina membrane filter (Whatman, electrode. For the reduction reaction, the oxidant is expected Anodisk 25) was used as the porous substrate. This filter is to migrate to the platinum electrode surface via the porous ca. 60 mm in thickness and has an asymmetric pore structure with pores of diameter of 0.02 mm and 0.2 mm.The surface of substrate and porous silica gel layer. J. Mater. Chem., 1998, 8(11), 2521–2524 2521the sample was heat-treated at a selected temperature in the range 300 to 700 °C. The surface and the cross-section of the thin film were observed by scanning electron microscopy (Hitachi S-2050 SEM and JEOL JCM-890S FE-SEM).In order to evaluate the thickness of the TiO2 layer, the TiO2 layer was distinguished from the platinum and silica layers by both secondary electron image (SEI) and back scattered electron image (BEI) observations of the sample cross section. Crystalline phases of the samples were identified by X-ray diVractometry (Rigaku Geigerflex System, Cu-Ka radiation). The photocatalytic activity of the sample was evaluated by measuring the photogeneration rate of hydrogen from an aqueous solution of ethanol.The simplified model of the photoredox system in an aqueous solution of ethanol mediated by the multilayered photocatalyst is also shown in Fig. 1. Ethanol is oxidized to acetaldehyde, acetic acid or CO2.17 Protons are reduced to hydrogen molecules on the platinum electrode.The photocatalyst was immersed in 500 ml of a 20 vol% aqueous ethanol solution under an Ar atmosphere. Illumination of the photocatalyst (TiO2 film side) was carried out using a 300 W Xe lamp (ILC Tech., LX300F). A water Fig. 2 Flowchart of the preparation of the TiO2 sol. cell was used as an infrared filter. The photogenerated hydrogen was detected by gas chromatography (Shimazu GC-6A with a thermal conductivity (TCD) cell and a 2 m long molecular sieve 5A column).Argon was used as the carrier gas. A platinized commercial TiO2 powder (Nihon Aerosil P- 25, anatase, particle diameter 10–50 nm) was used as a reference sample. Platinization of the powder was carried out by a photodeposition method in an H2PtCl6 aqueous ethanol solution.2 Results and discussion The XRD patterns of as-prepared samples showed only a halo indicating that the TiO2 layer was amorphous.Upon heattreatment at >300 °C for 4 h, a diVraction peak corresponding to the (101) plane of anatase appeared on the pattern and the crystalline phase of all heat-treated samples was predominantly anatase. The anatase crystallites were highly oriented along the (101) plane.The crystallite size was calculated by Sherrer’s equation assuming that there was no distortion in the crystal. As the temperature of heat treatment was increased, there was a parallel gradual increase in the peak intensities and crystallite size of anatase from ca. 10 nm (at 330 °C) to ca. 15 nm (at 500 °C). SEM images revealed that the surface (TiO2 film side) of the sample fired at 450 °C was devoid of any surface texture, i.e., it was very smooth and flat.Images of the cross section of this sample are shown in Fig. 4. The two SEI images show the pore structures of the alumina substrate and the overlayers. In the BEI image, the bright part was identified as the platinum layer. Clearly these images exhibit the expected structure of the multi-layered thin film system.We estimated the thickness Fig. 3 Flowchart for the fabrication of the TiO2 thin film of the TiO2 layer in all samples from the SEI+BEI image. photocatalyst. The change of TiO2 film thickness with the number of coating applications is shown in Fig. 5. It is apparent that film thickness increased in proportion to the number of coating applications.the porous substrate was coated with a silica sol (Nissan Chemical, Snowtex N) in order to make the surface smooth Each coating application produced a film about 10 nm thick. This film thickness corresponds to the crystallite size as thereby preventing the Pt and TiO2 fine particles from going through the substrate. This silica sol contains 20–21 wt.% of estimated from the XRD measurement.This shows that the TiO2 sol particles maintained their dispersive condition, with silica and the colloidal particles are 10–20 nm in diameter. The sol was diluted to 10 vol% with reagent grade ethanol the subsequent production of a dense and flat TiO2 thin film. The photocatalytic activity of the photocatalyst was and the coating was made just once a low speed spin coater (Thomas TM-701, ca. 300 rpm). After drying at 60 °C and evaluated by monitoring the rate of hydrogen evolution from an aqueous ethanol solution as already outlined in the calcining at 300 °C, platinum was sputter-coated on the silica gel layer. The platinum surface was subsequently coated with Experimental section. The hydrogen generation rate per unit surface area for the eight-coated sample as a function of firing the TiO2 sol, dried at 60 °C for 10 min and calcined at 300 °C for 10 min.This coating procedure for the TiO2 sol was temperature is shown in Fig. 6. The rate increased fairly steeply up to 400 °C and decreased gradually thereafter. The increase repeated until the desired film thickness was obtained. Finally, 2522 J. Mater. Chem., 1998, 8(11), 2521–2524activity above 450 °C are the eVects of the structural changes of the silica gel layer and TiO2 layer upon polycondensation or sintering among SiO2 or TiO2 fine particles.In the former case, the gel obtained from Snowtex N is known to sinter above 800 °C from its company’s catalogue. Further, in our previous photocatalytic work on the silica gel containing anatase fine articles, the photocatalytic activity also decreased above 500 °C, though the crystalline phase was unchanged at this temperature and the porous structure of the material was retained up to 800 °C.2 Considering the above information and our observations the porous structure of silica gel layer was also unchanged up to 800 °C, therefore, the silica gel layer is not responsible for the decrease of photocatalytic properties.A structural change of the TiO2 layer would result in grain growth and/or decrease of the concentration of surface OH groups and lattice defects such as Ti3+ and might aVect the photocatalytic activity.18,19 In order to investigate the change of fine structure of the thin film, precise observations of the surface and the cross-section of the thin film are required.However, at present, it is diYcult to observe such precise fine structure of the film by use of our SEM system. As mentioned already, there was no evidence of grain growth from the results of XRD measurement. In order to explain the eVect of firing Fig. 4 SEM images of the cross-section of the TiO2 thin film temperature on catalytic activity, further investigations on photocatalyst.properties of the thin films such as the number of OH groups on the TiO2 surface and the concentration and mobility of charge carriers, are therefore necessary. Fig. 7 shows the changes in hydrogen generation rate with film thickness for a sample heat-treated at 450 °C for 4 h. The hydrogen generation rate is presented as the rate per unit surface area as well as per unit mass of the TiO2 sample.The sample mass was evaluated assuming the density of an anatase crystal (3.54 g cm-3) based on the assumption that the film was dense. When the thickness of the film was less than ca. 60 nm, the rate was very low. This may be due to the formation of island textures of TiO2 on the platinum electrode in the early stages of spin coating.It is plausible that this texture induces the reverse reaction between photogenerated active oxidant and reductant on platinum, and consequently, hampers hydrogen evolution. Further coating applications pro- Fig. 5 Change of the TiO2 film thickness with the number of coating duced a uniform TiO2 film all over the platinum surface applications. resulting in a homogeneous spatial separation of reaction sites.As the film thickness was increased to >60 nm, the rate increased proportionally with the film thickness, while the rate per unit mass was almost constant and was also independent of film thickness. The photogeneration rate per unit mass was about eight times that of the reference sample of platinized TiO2 fine particles. These results strongly suggest that a similar number of photogenerated electrons per unit TiO2 mass diVuse to the surface of the platinum electrode regardless of the depth of the film.A potential gradient at the bottom level of the Fig. 6 Change in H2 generation rate with firing temperature. in H2 generation rate up to 400 °C is considered to be due to the crystallization of amorphous TiO2 into the photoactive anatase phase.However, it is not clear as to why the rate decreased above 450 °C considering that the crystalline phase of the thin film was not photoinactive rutile but was still anatase. Also, a sharp growth in anatase crystallite size was not observed. Changes in crystallite size aVect the band gap energy and the charge potential thereby influencing the overall photocatalytic properties.Fig. 7 Changes in H2 generation rate with thickness of the TiO2 thin film. Other possible causes of the decrease of photocatalytic J. Mater. Chem., 1998, 8(11), 2521–2524 2523conduction band and at the top level of the valence band is problems on the reduction site in this material are not so important as oxidative processes, such as the removal of probably formed across the TiO2 layer macroscopically, organic substances from polluted water and NOx from the because the front and reverse of the TiO2 (n-type semiconatmosphere by photo-oxidation, would be the main appli- ductor) layer are in contact with the solution and platinum cations of this catalyst. electrode, respectively.Therefore, photoexcited electrons Although poisoning or fouling of the surface by dense diVuse to the platinum electrode and holes diVuse to the pollutants is not completely avoidable in this material, our surface of TiO2 thin film along the potential gradients.Since fabricated system has potential for application to the removal the film was dense and consequently had few recombination of pollutants from the environment in comparison to finely sites such as surface defects, the separation of charge carriers divided particulate photocatalysts in porous media.occurred eYciently. Furthermore, the multiple layered structure on the porous substrate prohibited the reverse reaction between photoproduced oxidant and reductant thus achieving References a high photocatalytic activity as evidenced by the results 1 A.Fujishima and K. Honda, Bull. Chem. Soc. Jpn., 1971, 44, 1148. obtained in this study. 2 A. Yasumori, K. Yamazaki, S. Shibata and M. Yamane, J. Ceram. Soc. Jpn., 1994, 102, 702. 3 Y. Zhang, J. C. Crittenden, D. W. Hand and D. L. Perram, Conclusions Environ. Sci. Technol., 1994, 28, 435. 4 J. Harrmann and J. Mansot, J. Catal., 1990, 121, 340. A multilayered TiO2 thin film photocatalyst was prepared by 5 Y-M.Gao, H-S. Shen, K. Dwight and A.Wold, Mater. Res. Bull., sol–gel and sputtering methods. Compared with the reference 1992, 27, 1023. platinized TiO2 powder sample, the heat-treated photocatalyst 6 M-C. Lu, G-D. Roam, J-N. Chen and C. P. Huang, J. Photochem. Photobiol., 1993, 76, 103. showed a high eYciency for H2 generation from an aqueous 7 I. Sopyan, S.Murasawa, K. Hashimoto and A. Fujishima, Chem. solution of ethanol. The high catalytic activity of this thin film Lett., 1994, 723. system was achieved by a judicious arrangement of the TiO2 8 M. A. Aguado, M. A. Anderson and C. G. Hill, Jr., J. Mol. Catal., film and the platinum electrode on a porous substrate. This 1994, 89, 165. mesoscopic scale structure resulted in the reduction of the 9 A.Fernandez, G. Lassaletta, V. M. Jimenez, A. Justo, diVusion distances of charge carriers and the separation of A. R. Gonzalez-Elipe, J. M. Herrmann, H. Tahiri and Y. Ait- Ichou, Appl. Catal. B: Environ., 1995, 7, 49. redox reaction sites. 10 Y. Paz, Z. Luo, L. Rabenberg and A. Heller, J. Mater. Res., 1995, Most photocatalytic systems for environmental use will be 10, 2842.fabricated with TiO2 fine particles and porous support mate- 11 M. Anpo, T. Shima, S. Kodama and Y. Kubokawa, J. Phys. rials in commercial or industrial applications. In these systems, Chem., 1987, 91, 4305. micropores in the system act as adsorbents of pollutants; 12 D. Kim and M. Anderson, Environ. Sci. Technol., 1994, 28, 479. 13 S. Sato, H. Koshiba, H. Minakami, N.Kakuta and A. Ueno, however, poisoning or fouling easily occurs by the adsorption Catal. Lett., 1994, 26, 141. of not only the pollutant but also some components in the 14 C. Natarajan and G. Nogami, J. Electrochem. Soc., 1996, 143, reaction system. 1547. Because our fabricated material has a smooth surface TiO2 15 M. Takahashi, K. Mita and H. Toyuki, J. Mater. Sci., 1989, 24, layer (i.e. oxidation sites), the likelihood of fouling is relatively 243. low on this side. On the other hand, the platinum surface (i.e. 16 S. Sakka, K. Kamiya, K. Makita and Y. Yamamoto, J. Non- Cryst. Solids, 1984, 63, 223. reduction site) is located on the bottom of porous substrate. 17 K. Kato, Bull. Chem. Soc. Jpn., 1992, 65, 34. However, the thickness of the substrate is only ca. 60 mm, 18 Y. Oosawa and M. Gratzel, J. Chem. Soc., Faraday Trans. 1, 1988, therefore, the adsorbates are readily removed by washing or 84, 197. burning away. Further, since the redox reaction sites are 19 K. Kobayakawa, Y. Nakazawa, M. Ikeda, Y. Sato and spatially separated in our material, we need only to use one A. Fujishima, Phys. Chem., 1990, 94, 1439. reaction surface of the fabricated material for an adequate reaction system and good arrangement of apparatus. Any Paper 8/03265C 2524 J. Mater. Chem., 1998, 8(11), 2521–2524