Study of the Efficiency of Visible-Light Photocatalytic Degradation of

Langmuir 2001, 17, 5025-5030

5025

Study of the Efficiency of Visible-Light Photocatalytic Degradation of Basic Blue Adsorbed on Pure and Doped Mesoporous Titania Films Elias Stathatos, Tatyana Petrova, and Panagiotis Lianos* Engineering Science Department, University of Patras, 26500 Patras, Greece Received March 8, 2001. In Final Form: May 14, 2001 Transparent mesoporous titania films have been deposited on glass slides by a sol-gel procedure in the presence of Triton X-100 reverse micelles in cyclohexane. Benzothiazolium, 2-[[4-[ethyl(2-hydroxyethyl)amino]phenyl]azo]-5-methoxy-3-methyl(T-4)-methoxysulfate (Basic Blue 41), has been adsorbed on these films from aqueous solutions, and the photodegradation of the dye by visible-light illumination has been monitored by absorption spectrophotometry. Film nanostructure has been optimized for maximum photodegradation efficiency by controlling the original reverse micellar composition, the ripening of the particles, and the thickness of the films. Films doped with silver ions, incorporated through the reversemicellar route, are more efficient photocatalysts than pure titanium films and become even more efficient when they are treated with UV radiation. Films doped with ruthenium ions are less efficient for photocatalysis but when they are treated with UV radiation, they also become more efficient photocatalysts than pure titania films.

1. Introduction TiO2 heterogeneous photocatalysis for treatment of wastewater and polluted air is one of the most extensively studied processes in recent years.1 Titania photocatalysts are usually employed as suspended powders. However, the manipulation of such powders and their removal are difficult; therefore, the most recent research is focused on the preparation of immobilized catalysts, e.g., in the form of thin films on solid supports. In any case, thin films are particularly interesting in applications to open-air purification panels.2,3 The efficiency of thin films is lower than that of suspended powders, owing to contamination by impurities and to the smaller quantity of active catalyst.4 For this reason, an intensive effort has been made to enrich titania with species that can enhance its photocatalytic capacity. Noble metals, silver in particular, have been used for this goal. Silver increases the photocatalytic capacity of titania by some different physical mechanisms that may act separately or simultaneously. Reduced silver particles of colloidal dimension (metallic silver) are considered to be electron traps. Thus by trapping electrons produced in electron-hole pairs after photons absorption, they prevent electron-hole recombination and liberate holes to participate in degradation reactions.5 Plasmon resonance effects in metallic silver nanoparticles are also held responsible for local enhancement of the electric field, facilitating electron-hole pair production and separation.4,6,7 Finally, silver is also reported to affect the * To whom correspondence should be addressed. Tel: 30-61997587. Fax: 30-61-997803. E-mail: [email protected]. (1) Fujishima, A.; Hashimoto, K.; Watanabe, T. TiO2 photocatalysis. Fundamental and Applications. BKC Inc.: Tokyo, 1999. (2) Negishi, N.; Takeuchi, K.; Ibusuki, T. J. Mater. Sci. 1998, 33, 5789. (3) Sopyan, I.; Watanabe, M.; Murasawa, S.; Hashimoto, K.; Fujishima, A. J. Photochem. Photobiol., A 1996, 98, 79. (4) Herrmann, J.-M.; Tahiri, H.; Ait-Ichou, Y.; Lassaletta, G.; Gonzales-Elipe, A.R.; Fernandez, A., Appl. Catal., B 1997, 13, 219. (5) Ilisz, I.; Dombi, A. Appl. Catal., A 1999, 180, 35. (6) Zhao, G.; Kozuka, H.; Yoko, T. Thin Solid Films 1996, 277, 147. (7) Lassaletta, G.; Gonzales-Elipe, A. R.; Justo, A.; Fernandez, A.; Ager, F. J.; Respaldiza, M. A.; Soares, J. G.; Da Silva, M. F. J. Mater. Sci. 1996, 31, 2325.

structure of titania, most likely leading to modification of its photocatalytic activity.8 In the present work, we study the photocatalytic efficiency of transparent mesoporous titania thin films, made by a sol-gel method involving reverse micellar templates.9-11 Films are structured in nanoparticles of controlled and uniform size, both in pure form or doped with silver or ruthenium. The metal ion is introduced in the films by initial solubilization in the reverse micellar solution. We have found that the presence of the impurity results in increasing photodegradation rate, and we thus search for ways to optimize it. We have tested our samples by monitoring rates of photodegradation of a model dye, Basic Blue 41 (BB) adsorbed on pure or doped-titania mesoporous films. This choice was dictated by the following motives. Dye photodegradation on a transparent film can be easily monitored by absorption spectrophotometry. A large portion of water-borne pollutants is synthetic textile dyes and industrial dyestuffs.12 Dye photodegradation can be carried out by visible light, through light absorption by the dye and subsequent electron separation by transfer to the semiconductor.12-14 Finally, TiO2 photodegradation reactions in thin film configuration is a process acquiring increasing industrial interest since applications of pure TiO2 can be envisaged by deposition on large surfaces for both indoor and outdoor air-purification systems.1-3,15 In that case, photodegradation is obtained by direct TiO2 absorption of near-UV radiation. (8) Tada, H.; Tiranishi, K.; Inubushi, Y.-I.; Ito, S. Langmuir 2000, 16, 3304. (9) Stathatos, E.; Tsiourvas, D.; Lianos, P. Colloids Surf., A 1999, 149, 49. (10) Stathatos, E.; Lianos, P.; Del Monte, F.; Levy, D.; Tsiourvas, D. Langmuir 1997, 13, 4295. (11) Stathatos, E.; Lianos, P.; Falaras, P.; Siokou, A. Langmuir 2000, 16, 2398. (12) Vinodgopal, K.; Wynkoop, D. E.; Kamat, P. V. Environ. Sci. Technol. 1996, 30, 1660. (13) Vinodgopal, K.; Kamat, P. V. Environ. Sci. Technol. 1995, 29, 841. (14) Nasr, C.; Vinodgopal, K.; Fisher, L.; Hotchandani, S.; Chattapadhyay, A. K.; Kamat, P. V. J. Phys. Chem. 1996, 100, 8436. (15) Negishi, N.; Iyoda, T.; Hashimoto, K.; Fujishima, A. Chem. Lett. 1995, 841.

10.1021/la0103620 CCC: $20.00 © 2001 American Chemical Society Published on Web 07/04/2001

5026

Langmuir, Vol. 17, No. 16, 2001

Stathatos et al.

Figure 1. Lamp emission spectrum, BB absorption spectrum, and BB chemical structure.

2. Experimental Section Materials. Titanium (IV) isopropoxide, polyoxyethylene(10) isooctylphenyl ether (Triton X-100), Ag(NO3), hydrated RuCl3, andbenzothiazolium,2-[[4-[ethyl(2-hydroxyethyl)-amino]phenyl]azo]-5-methoxy-3-methyl(T-4)-methoxysulfate (Basic Blue 41, BB, cf. chemical structure in Figure 1) were purchased from Aldrich and used as received. The rest of the reagents were from Merck. Millipore water was used in all experiments. Preparation of TiO2 Mesoporous Films Deposited on Glass Slides. Reverse micellar solutions of 0.2 M Triton X-100 and various water concentrations were prepared in cyclohexane. To each of these solutions we added 0.2 M titanium isopropoxide under vigorous stirring and at ambient conditions.9-11 Hydrolysis and condensation of titanium isopropoxide begin as soon as it is introduced in the reverse micellar solution, but it may take more than an hour before the solution becomes a visible gel. The thus prepared composite material can be deposited as thin film on a glass slide by dip-coating. The optical absorption onset for the glasses used in the present work was at 315 nm. The slide was previously sonicated for 20 min in ethanol and was, finally, copiously washed with Millipore water and dried in a stream of N2. The thus prepared slide was dipped into the gelling solution and was fast withdrawn at a speed of 2 cm/s. The composite organic-inorganic film was left to dry in air, and then it was slowly heated in air, up to 450 °C, at a rate of 3.5 °C/min. The film was finally sintered at 450 °C for about 15 min more. The procedure was repeated two additional times so that three depositions were finally made. Thicker films could be made by further successive coatings; however, thickness grows at the expense of transparency. Atomic force microscopy (AFM) and scanning electron microscopy (SEM) images of the films prepared by the above method, and published in previous publications,9-11 reveal a mesoporous structure that consists of TiO2 nanoparticles of practically monodisperse size with exceptional reproducibility. Mesoporous are structures with pore sizes lying in the range 2-50 nm, as in the present case. The size of the nanoparticles can be easily controlled by choosing the water/surfactant ratio in the original reverse micellar solution. Larger nanoparticles are made when more water exists in the solution. The advantage of the reverse-micellar route in making TiO2 particles, as compared with other methods, exactly lies on this capacity to control the mesoporous structure of the obtained films. The diameter of the nanoparticles employed in the present work, as estimated by using AFM images, was around 30 nm in the case of samples made with a water/surfactant ratio equal to 2 (0.2 M Triton X-100, 0.4 M water). The TiO2 films are too thin to give a detectable X-ray diffraction spectrum. We have freeze-dried the reverse micellar solution containing titanium isopropoxide about 1 h after mixing the components. The gel was first brought to liquid nitrogen temperature and then was continuously pumped while the vessel containing the solution was exposed to

Figure 2. Absorption spectrum of a typical titania film without (1) and with (2) adsorbed BB. ambient temperature. The solution remains frozen due to evaporation of the solvent, and it remains so until all solvent has been evaporated. The obtained material contains, of course, TiO2 and surfactant. It is then calcinated under the same conditions, as described above for films. The obtained powder was studied by X-ray diffraction, and it gave a diffraction pattern which corresponds to anatase. Given the conditions of the freeze-drying procedure, we believe that the above film also consists of anatase nanocrystallites. Incorporation of Metal Ions. Silver or ruthenium ions can be directly incorporated in titania film by solubilization in the original reverse micellar solution; i.e., instead of pure water, we used aqueous salt solution so that the final metal ion concentration is well-defined. The rest of the procedure remains the same as above. In the present work both ion concentrations were 0.02 M, i.e., at a 10% doping level with respect to titanium. Adsorption of the Dye on the Films. Both pure and doped titania films adsorb a substantial amount of BB when they are dipped in an aqueous solution of the dye, where BB is highly soluble. Adsorption is instantaneous, and the films are immediately deeply colored (cf. Figure 2). In typical preparations, dye concentration in solution was 10-2 M. Methods. Absorption measurements were made with a Cary 1E spectrophotometer. UV treatment of the films, aiming at reducing incorporated metal ions, has been obtained by illumination with a mercury-containing 400-W UV lamp. Illumination of the samples for dye photodegradation was achieved under ambient conditions with a commercial 250-W tungstenhalogen spot light satisfactorily simulating solar radiation (cf. Figure 1). The light intensity at the surface of the sample was 600 mW/cm2. Photodegradation was monitored by measuring the absorbance of the dye at its maximum. Measurements were made every 10 min of illumination.

3. Results and Discussion The data presented in this section are divided into two parts. First, we compared different titania films, formed at different preparation conditions, with respect to their photodegradation efficiency toward adsorbed dye. Then we have doped the most representative of these titania films with a dopant, i.e., ruthenium or silver, and we have compared photodegradation efficiency between doped and undoped titania. Comparison of Different Undoped Titania Films. Figure 1 shows the absorption spectrum of dilute aqueous Basic Blue 41 (BB) together with the emission spectrum of the lamp used for photodegradation. The chemical

Photodegradation of Dye

structure of BB is also inserted in Figure 1. Figure 2 shows the absorption spectrum of a representative titania film alone or with the adsorbed dye. The oscillating part in the spectrum of titania is due to interference fringes. Combination of the information in Figures 1 and 2 leads to the conclusion that the photodegradation of BB is realized by the photosensitization procedure;12-14 i.e., light is almost exclusively absorbed by the dye. Then the excited electron is transferred into the conduction band of the semiconductor leaving behind a hole localized on the dye. The thus separated electron is scavenged by atmospheric molecular oxygen so that e--h+ recombination is prevented and the dye is destroyed by the remaining hole. Photodegradation by direct excitation of titania is not excluded but it is very limited. Indeed use of filters cutting off the near-UV radiation of the lamp (400 nm cutoff filter) has a small effect on photodegradation rates (about 5%). The data presented from this point on are values of the maximum absorbance of the adsorbed dye at various steps of illumination. Decrease of absorbance is considered as photodegradation, i.e., mineralization of the dye. Eventhough we cannot measure CO2 emission during illumination, since the quantity of mineralized material in a thin film is too small to yield a detectable quantity of gas, the mineralization hypothesis is based on indirect evidence. There exists a controversy on whether decrease of dye absorbance in the presence of TiO2 is due to photodecomposition or to reversible photobleaching.16 Reversible photobleaching is excluded in our case for the following reasons. Reversible photobleaching is usually obtained in the absence of oxygen.16 It is the oposite in our case. No photodegradation of the dye was detected in the absence of oxygen. In addition, no recovery of the color of the bleached film was ever obtained under any conditions. Furthermore, bleached films can adsorb new dye by dipping, at the same level as the intact film. This can be repeated several times with the same results. Apparently, illumination leads to dye decomposition and the oxide layer becomes ready to adsorb new dye. This could not be possible if the dye remained on the film, losing only its color. With this in mind, we have first compared titania films which were all made by the reverse-micellar route, as described in section 2, by using different amounts of water/surfactant (w/s) ratio. We have found that hydrolysis of titanium isopropoxide and subsequent gelation by inorganic (i.e., -Ti-O-Ti-) polymerization, according to the sol-gel procedure, is very rapid for w/s g 3. Fast gelation leads to very rough, light-scattering films which are not appropriate for the present spectrophotometric studies. Thus the highest exploitable ratio was w/s ) 2.5 (0.2 M Triton X-100 + 0.5 M water). Titania films made by using different w/s ratios have been previously characterized by SEM, transmission electron microscopy (TEM), and AFM.9-11 As a general rule, larger water content leads to mesoporous titania films composed of larger nanoparticles. The total active surface of the mesoporous structure is larger in smaller particles, i.e., in the case of lower w/s values. Increase of w/s, i.e., increase of the nanoparticle size, makes rougher films with smaller active surface expansion. For this reason, we have found that films made with w/s ) 2.5 have smaller capacity of BB adsorption, less than 50% of the capacity of films corresponding to w/s ) 1. It is possible that no more than one BB molecule can be adsorbed on a single titania nanoparticle. As a consequence, smaller but more numerous nanoparticles can adsorb more dye than larger but less numerous nanoparticles. However, bigger particles (16) Mills, A.; Wang, J. J. Photochem. Photobiol. 1999, 127, 123.

Langmuir, Vol. 17, No. 16, 2001 5027

Figure 3. Variation of maximum absorbance of BB adsorbed on pure titania films as a function of illumination time with visible light. The three curves correspond to three films made with different water contents in reverse micellar solutions containing 0.2 M surfactant.

Figure 4. Absorption spectra of four titania films, (1) w/s ) 1, (2) w/s ) 2, (3) w/s ) 2.5, and (4) w/s ) 2, but dipping was made at a later stage of gelation.

are more effective in photodegrading BB than smaller particles, as can be seen in Figure 3, where it is obvious that photodegradation rates are much faster for w/s ) 2.5 than for w/s ) 2 or 1. Eventhough, more extensive studies are needed to explain this phenomenon, there may be two plausible explanations. It is known that, according to Brus effective mass model for nanosize semiconductors,17 absorption occurs at longer wavelength for bigger particles due to the size effects on the electronic properties of the semiconductor. Indeed, as seen in Figure 4, showing absorption spectra of films made with different values of w/s or different growth times, larger particles (higher w/s or longer growth) absorb light at longer wavelengths than smaller particles (lower w/s or short times). One then possible reason for higher rates with w/s ) 2.5 is the larger percentage of direct near-UV absorption of photons by the larger particles. However, this seems to us a rather (17) Brus, L. J. Phys. Chem. 1986, 90, 2555.

5028

Langmuir, Vol. 17, No. 16, 2001

Figure 5. Variation of maximum absorbance of BB adsorbed on pure titania films as a function of illumination time with visible light. The two curves correspond to two films made with the same reverse micellar solution but at different stages of gelation.

weak argument since the direct participation of titania in photon absorption is, as already said, very limited. Meanwhile, smaller particles have a higher energy bandgap. It is then possible that electron transfer from the dye to the higher-lying conduction band of the smaller-size semiconductor particles during photosensitization, is more difficult than in larger particles, which have a smaller band-gap. This can partly explain the smaller efficiency of smaller particles (cf. also ref 18). A second explanation is that in smaller particles it is more probable to have extended surface defects that will quench the excited species by providing electron-hole recombination sites and thus give lower photodegradation rates. As a conclusion, the above information tells us that more material is adsorbed on a film which has a finer mesoporous structure but it offers lower photodegradation rates. For this reason and in order to compensate the two opposing effects, we have opted for w/s ) 2 (0.2 M Triton X-100 and 0.4 M water) as the best combination for both satisfactory adsorption capacity and photodegradation efficiency. Most films are then made under this condition. As a verification of the fact that smaller particles give lower photodegradation rates but higher dye adsorption capacity, we have produced films made with w/s ) 2 at two different stages of gelation, i.e., dipping in the solution was done 10 and 180 min after component mixture. A larger waiting time means more extensive particle growth. In fact, approximately, the same effect is produced by using w/s ) 2 and waiting 3 h then by using w/s ) 2.5 and dipping a few minutes after preparation. Comparison of photodegradation rates between a film made at an early stage of gelation with a film made at a late stage of gelation is shown in Figure 5. Indeed, the latter film adsorbs less dye but degrades it faster. The quantity of the adsorbed dye is not shown in Figure 5. Maximum absorbances are normalized for the sake of comparison. Finally, by using films corresponding to w/s ) 2, we have also studied the effect of subsequent TiO2 layers deposition on the photocatalytic efficiency. It was then found that the dye is photodecomposed at exactly the same time, independent (18) Kavan, L.; Stoto, T.; Graetzel, M.; Fitzmomrice, D.; Shklover, V. J. Phys. Chem. 1993, 97, 9492.

Stathatos et al.

Figure 6. Variation of maximum absorbance of BB adsorbed on Ag-doped titania films as a function of illumination time with visible light.

of the number of layers. More layers make films that adsorb more dye. However, too many layers work against transparency. Thus we have optimized our system by working with films, consisting of three layers of TiO2, based on a reverse micellar solution with w/s ) 2. Titania Films with Dopants. Photocatalytic degradation of BB has also been studied with doped TiO2 films made by the above procedure, using a standard water/ surfactant ratio w/s ) 2. We have tried two dopants, Ag+ and Ru3+. Salts of these metals have been introduced at a concentration 0.02 M. i.e., at a 10% doping level with respect to titanium, by solubilizing them in the water used to make Triton X-100 reverse micelles. Formation of films by dipping and subsequent heating to 450 °C as well as BB adsorption for photodegradation experiments followed the same procedure as for pure titania films. We have examined two types of samples. In the first type, immediately after the doped film was taken out of the oven, the absorption spectrum of the film was registered and then BB was adsorbed by dipping. Finally, the film with adsorbed dye was illuminated with visible light and the photocatalytic degradation of the dye was monitored by absorption spectrophotometry. In the second type, after the films were taken out of the oven, they were first subjected to UV treatment for about 10 min, in the case of Ag-doped films, and for about 1 h, in the case of the Ru-doped films, then the dye was adsorbed and finally was photodecomposed by visible light illumination. UVtreated doped films are much more efficient for dye photodegradation than untreated films. This can be seen in Figures 6 and 7, where Ag-doped and Ru-doped TiO2 photocatalytic rates are, respectively, presented. Figure 6 shows that it takes about 2 h to photodegrade all adsorbed dye on pure TiO2 films. Photodegradation time diminishes by 50% with nontreated silver-doped films and it is limited to only 20 min when the films are previously treated with UV radiation. In the case of Ru-doping, a markedly different behavior was observed. Nontreated Ru-doped films decrease photodegradation efficiency, increasing photodegradation time from 2 h, for pure TiO2 films, to 5 h, for untreated doped films. On the contrary, Ru-doped films, previously treated with UV radiation, decreased photodegradation time to 50 min. As already discussed in the Introduction, silver doping of titania has

Photodegradation of Dye

Figure 7. Variation of maximum absorbance of BB adsorbed on Ru-doped titania films as a function of illumination time with visible light.

beneficial effects on its photocatalytic activity owing to the existence of three mechanisms that may function separately or jointly: metal particles created by silver ion reduction act as electron traps aiding electron-hole separation; dipole moment increase in the vicinity of metal particles also facilitates electron excitation and e--h+ separation; and silver dopants may structurally modify titania increasing its photocatalytic activity.4-8 We believe that enhancement of photocatalytic efficiency after UV irradiation is due to the creation of metal centers in the film, owing to reduction by electrons in the conduction band of the semiconductor. We thus adopt the above model accepted for silver impurities. Titania absorbs UV radiation, as can be seen in Figure 4. Absorption of UV photons excites a large number of electrons, which can reduce both Ag+ and Ru3+. Of course, metal colloidal particle formation by reduction of the cationic species is easier in the case of the monovalent silver than in the case of the trivalent ruthenium. For this reason only 10 min of UV illumination is necessary to transform silver-doped films, but it takes about 1 h in the case of ruthenium-doped films. It is most probable that a large portion of silver ions are reduced during heating of the film, even before UV illumination, thus the presence of silver always speeds up photodegradation. The presence of metal ion impurities in the TiO2 films had important effects on the spectroscopic characteristics of the films, which possibly mark important variations in the structure of the particles itself. Figure 8 shows the case of Ru3+. The presence of these ions is demonstrated by a marked absorption band in the visible. This band is due to ruthenium-localized transitions since it is also found in RuCl3 solutions. In addition, an important red-shift in the titania absorption onset suggests that ruthenium ions occupy titanium substitution sites in the titania lattice, introducing impurity states within the semiconductor energy band gap.19,20 It is possible that these states become recombination sites for electron-hole pairs, thus slowing down photodegradation rates,19 as observed with Ru-doped (19) Ohno, T.; Tanigawa, T. F.; Fujihara, K.; Izumi, S.; Matsumura, M. J. Photochem. Photobiol., A 1999, 127, 107. (20) Serpone, N.; Lawless, D.; Disdier, J.; Herrmann, J.-M. Langmuir 1994, 10, 643. (21) Brezova, V.; Blazkova, A.; Borosova, E.; Ceppan, M.; Fiala, R. J. Molecular Catal. A: Chem. 1995, 98, 109.

Langmuir, Vol. 17, No. 16, 2001 5029

Figure 8. Absorption spectrum of pure titania (made at w/s ) 2) (1), titania doped with 10% Ru3+ (2), and UV-treated film doped with 10% Ru3+ (3).

Figure 9. Absorption spectrum of pure titania (made at w/s ) 2) (1), titania doped with 10% Ag+ (2), and UV-treated film doped with 10% Ag+ (3).

samples before illumination with UV radiation. Of course, after exposure to strong UV radiation, metal centers are created which speed up photodegradation, as actually observed. The presence of silver ions marks a different behavior, as can be seen in Figure 9. Silver causes a blueshift in the film absorption onset suggesting a variation in the film nanostructure and not an introduction of impurity states. In Figures 8 and 9, it is noted that UV treatment of the film is not followed by the appearance of a new visible absorption band, indicating plasmon resonance absorption states. This may be due to the fact that the quantity of metal in the films is too small to give a detectable plasmon resonance absorption. In any case, these subjects are further studied in our laboratory. 4. Conclusion The above mesoporous films are efficient photocatalysts for the degradation of adsorbed Basic Blue 41. More dye is adsorbed on films made with smaller titania nanopar-

5030

Langmuir, Vol. 17, No. 16, 2001

ticles, but faster photodegradation rates are obtained with larger nanoparticles. More dye is also adsorbed on thicker films, but thickness grows at the expense of transparency. The reverse micellar route in combination with the solgel procedure can be successfully employed to optimize the efficiency of these films. Films doped with silver or ruthenium ions are very efficient photocatalysts if they are previously treated with UV radiation, i.e., when the

Stathatos et al.

ionic species are reduced, particularly in the case of silver dopants. Acknowledgment. We acknowledge financial aid from the program “Management of Industrial Wastes” of the Greek General Secretariat of Research and Technology. LA0103620