Optimised photocatalytic activity of grid-like mesoporous TiO₂ films: effect of crystallinity, pore size distribution, and pore accessibility

Abstract

Photocatalytic activity of anatase cubic-based ordered mesoporous thin films was related to the morphology of the crystalline porous network obtained upon thermal treatment with a specific incremented sequence. The porosity and pore size distribution of these thin films were investigated with a novel Environmental Ellipsometric Porosimetry (EEP) technique. Network crystallinity was assessed by XRD. In parallel, the evolution of the photocatalytic activity was studied through UV-induced photodegradation of methylene blue and lauric acid within the films at the various steps of the temperature treatment. The photoactivity was linked to the porous characteristics of the films and we concluded that the activity is optimal when the porosity is high and completely accessible, and when the nanoparticle and pore size have dimensions of 7.5 and 5.5 nm respectively. Such an optimal system was obtained after a sequential thermal treatment ending with 10 min at 600 °C in air, for which the films adopts an ordered bidirectional grid-like structure.

---

Introduction

Crystalline TiO₂ thin films are very attractive materials because of the photoinduced electronic transfer properties associated with the anatase metastable phase, its high chemical stability, its safety, and low cost. Main potential applications include superhydrophilic surface,¹ self cleaning coatings,² photovoltaic cells,³ and photocatalysed depolluting layers.⁴ While the first two require smooth and dense coatings, the latter two applications need high surface area and high accessibility in order to facilitate diffusion within the active network. Photocatalysts have been studied for application in water and air purification as well as for decomposition of water.^5,6 High photocatalytic activity is associated with high crystallinity and high surface area^7,8 since they respectively prevent the electron–hole recombination and the increase of reactant quantity onto the photocatalyst surface. A high surface area also favours the production of active oxygen species. Mesoporous TiO₂,^9–11 Ta₂O₅,¹² and MgTa₂O₆^13,14 have been reported to be efficient for the photocatalytic decomposition of dye, acetone, water, and so on. Concerning photovoltaic applications, the mesoporosity must be as open as possible to enhance dye impregnation and dielectric incorporation. Ordered mesoporosity is obtained through the Evaporation Induced Self-Assembly (EISA) templating method.^15,16 Our recent works on mesostructured thin films showed that the preparation of highly ordered and porous anatase coatings requires perfect tuning between the chemicals, the deposition process and the treatment critical parameters.^17–19 Indeed, a typical mesoorder transition from the contracted bcc Im3m structure to the vertical grid-like structure was obtained upon nucleation growth of metastable anatase nanocrystalline particles within the inorganic framework.²⁰ This transition is characterised by the transformation of discrete ellipsoidal cavities forming a closed porous network (amorphous state) into an open interconnected porosity (nanocrystalline state) by the combined effect of spatially quenched growth and diffused sintering of particles.

Recent studies showed that the photocatalytic activities of this sort of material are greatly influenced by the particle size and crystallinity. Photocatalytic activities are deduced by assessing the kinetic of decomposition of organic species (typically methylene blue (MB) dye) upon UV light irradiation.^21–24 Anatase shows the highest photocatalytic activity in the crystal phase of TiO₂. UV light irradiation leads to the generation of excited electrons and holes. Their generation is related to the reduction of Ti⁴⁺ and the oxidation of O²⁻, which is associated with the formation of reactive OH˙ and O˙ radical species, that are responsible for the organic decomposition into CO₂ and water. On the other hand, since this phenomenon takes place at the material surface, its accessibility to organics must be high and must be directly linked to the porosity. However no precise investigation concerning the effect of the pore size combined with nanocrystallite dimension on such photoactivity has ever been reported to the best of our knowledge.

In this work, we compare the kinetics of UV induced MB and lauric acid (LA) decomposition when the latter is impregnated within mesoporous TiO₂ thin films having various mesostructures, crystallinity, crystallite dimensions, porosity, pore size and surface accessibility. These films were prepared by varying the thermal treatment conditions applied to an initial bcc mesoporous titania film templated by Pluronic F127. The photodecomposition of MB and LA was assessed by UV-visible spectroscopy and water contact angle measurement, respectively, and was related to the crystallinity of the walls deduced from XRD and to the pore size distribution investigated by a novel ellipsometric porosimetry technique: Environmental Ellipsometric Porosimetry.²⁵ We show that the highest activity arises when the porosity is constituted of an open grid-like network integrating pore size and crystallite dimensions of 5.5 and 7.5 nm respectively. This optimised porous morphology was achieved by applying a sequential thermal treatment ending with 10 min at 600 °C in air.

Experimental

Orthorhombic mesoporous thin TiO₂ films were prepared as previously reported. In a typical preparation,^18,20 TiCl₄ precursor was dissolved in a solution containing EtOH, H₂O and F127 (Pluronic block copolymer EO₁₀₆-PO₇₀-EO₁₀₆) (molar ratio: 1 Ti, 10 H₂O, 40 EtOH, 0.005 F127). Thin films were deposited by dip coating on (001) Si wafers under constant humidity of 60%. They were then allowed to age for 24 h at 30% humidity before been thermally treated up to 300 °C for several hours. Further thermal treatments in air were carried out to permit elimination of organics, crystallisation of the network, and opening of the porosity. The following typical sequence of treatment was applied after stabilisation at 300 °C: 20 min at 400 °C + 15 min at 500 °C + 10 min at 600 °C + 5 min at 700 °C + 2 min at 800 °C (all segments are essential). Decreasing period with increasing temperature was chosen so as to prevent extensive particle growth through diffuse sintering favoured at high temperatures. Evolution of the mesostructure with such treatments was previously investigated by in situ SAXS in our groups,²⁰ but is once more reported in this paper for clarity. Additionally, the crystallinity, and porosity of such films were investigated by Wide Angle XRD and EEP (Environmental Ellipsometric Porosimetry) respectively at each of these latter temperatures. Finally, photocatalytic activity was assessed by MB and LA photodegradation upon UV light irradiation. Many researchers have used a commercial TiO₂ powder, P-25 (Degussa), as a comparison photocatalyst.^22–24 However, it is very difficult to compare the photocatalytic activity of the thin film with that of the powder. Thus, in this study, a TiO₂ thin film without an organized mesoporous structure was used as the comparison film. This film was prepared by the same method as the 500 °C calcined mesoporous TiO₂ one without using F127, and had anatase crystalline phase and ca. 3 vol% porosity measured by EEP.

XRD investigations

2D Diffraction patterns revealing the mesoorder were obtained at the Austrian SAXS beam line of Elettra synchrotron. Details concerning the set up and the conditions of analysis are reported elsewhere.²⁰ Wide angle diffraction patterns were collected on a Philips D8 apparatus in Detector Scan mode (λ_CuKα, θ fixed between 0.3 and 0.5°, overnight acquisition time).

Environmental ellipsometric porosimetry analysis

EEP investigations were conducted on a Wollam VASE M2000U ellipsometer equipment modified for water adsorption/desorption analysis. This method consists of allowing the film to stand in an atmosphere where the relative humidity is controlled. The refractive index and the thickness of the films are measured in the visible wavelength range during increasing and decreasing relative humidity. Depending on the pore size, water is adsorbed at a specific relative humidity. These data are then transformed into adsorbed volume fraction of water versus relative pressure of water using the Bruggeman Effective Medium Approximation model and classical laws of the thermodynamics. The pore size distributions are evaluated using the Kelvin equation (i.e.RT ln(P/P₀) = − γ_l/v cosθ_l/s (dS/dV)), characterising the liquid/vapour interface geometry when capillary condensation of water occurs inside the cavities. The geometrical term dS/dV was calculated by taking into account the ellipsoidal pore geometry. In the present case, an aspect ratio of ρ = a/b = 2.3 (from 300 °C) is used after deducing the pore anisotropy from the lattice degree of contraction and the TEM pictures.²⁵ The surface tension γ_l/v is corrected by the effect of a high liquid surface curvature met in close confinement (mesopores) known as Tolman's correction. The contact angle θ_l/s is directly measured by depositing a water drop directly on the samples and was found to be less than 10° for every temperature. The real radius r_pore is deduced by adding the layer thickness of adsorbed water molecules t, directly evaluated by ellipsometry measurement on a flat surface of non-porous TiO₂ at various P/P₀, to Kelvin's radius r deduced from the isotherms. Details of such calculation are available in another article.²⁵

Photocatalytic properties

Photodegradation of MB in aqueous solution. TiO₂ films were irradiated with UV light (24 W, λ = 365 nm) overnight to decompose pollutants adsorbed on its films. Then, the TiO₂ films were placed in a Petri dish filled with 0.01 mM MB aqueous solution (10 mL) and MB was adsorbed onto the film surface in the dark for 1 hour. The photodegradation of MB was initiated and conducted by UV light irradiation with stirring. We monitored the absorbance of the MB solution corresponding to λ_max around 650–665 nm by UV-visible spectroscopy versus irradiation time (2 hour intervals). The amount of MB decomposition was determined by a linear relationship between the concentration and the absorption of MB.

Photodegradation of LA coated on TiO₂ films. TiO₂ films were irradiated with UV light overnight to decompose pollutants adsorbed on its films. LA was coated on the TiO₂ films by dip-coating with 2 wt% LA ethanol solution. Then, the obtained films were dried at 60 °C for 30 minutes. The photodegradation of LA was conducted by irradiating with UV light, monitoring the water contact angle of the TiO₂ film surface by using a charge coupled device (CCD) camera versus irradiation time (10 minutes–1 hour intervals).

Results and discussion

Mesostructure

Mesostructure evolution with increasing temperature is illustrated by SAXS patterns and TEM images in Fig. 1, reproduced from reference 20. The initial structure is a bcc arrangement of spherical micelles (lattice parameter a = 18 nm) that underwent a uniaxial contraction upon calcination up to 400 °C. At this stage, the structure is orthorhombic and transforms into a grid-like structure, with interconnected pore network planes separated by 13 nm, upon further heating.

Fig. 1 TEM and 2D-SAXS analyses of Im3m contracted mesoporous TiO₂ thin films treated at 300 °C ((a) and (c)) and 600 °C ((b) and (d)) reproduced with permission from Chem. Mater., 2003, 15, 4562. Copyright 2003 Am. Chem. Soc. TEM: (a) [110] zone axis showing the typical cubic type arrangement of 7 nm wide discrete pores; (b) the typical grid-like structure obtained after pore merging in the [111] direction (scale bar = 50 nm). SAXS: c) typical diffraction pattern of a contracted Im3m structure with initial a = 18 nm; (d) typical signature of a grid-like structure. The characteristic off-plane diffraction peaks are much more intense and at the same distance than the formal (1−10) one; (d₍₁₋₁₀₎ = d_(grid-like) = 13 nm).

This was attributed to the crystallisation into anatase of the inorganic walls. More precisely, this structural modification is related to the diffuse sintering of nanocrystallites that follows the thermal induced nucleation-growth of anatase. According to our previous study, the intensity of the diffraction peak corresponding to the grid-like structure (formal (1−10)) is maximal between 500 °C and 600 °C, suggesting that the ordered mesoporosity is retained up to this latter temperature. If the system is maintained for more than 10 min at 700 °C, or heated above this critical temperature, the ordered mesoporosity is progressively lost.

Microstructure

Fig. 2 represents the evolution of XRD patterns collected on films treated at temperatures given in the Experimental section. They show that anatase nanocrystals are detected from 500 °C and no transformation into the thermodynamically stable rutile phase occurs even at higher temperatures in the present conditions, as is systematically observed with mesosporous TiO₂.

Fig. 2 XRD diagrams of TiO₂ mesoporous films calcined at various temperatures (* labels represent the Si wafer diffraction peak). Inset: evolution of the crystallite dimension deduced from Scherrer's equation applied on the (101) diffraction peaks.

Using the Scherrer equation, we observed that nanocrystals grow up to 7.5 and 15 nm in size after 600 °C and 700 °C treatment, respectively. Above 700 °C, extensive sintering takes place and larger particles are detected. This is in agreement with the SAXS data that reveal a mesoorder collapsing at this temperature. Indeed, it is most probable that the extensive increase of crystallite dimension through such thermal solicitation is responsible for the collapsing of the mesostructure.²⁰ This phenomenon has been well described using in-situ SAXS-WAXS investigations and confirmed by kinetic studies of both crystallization and degradation.²⁶

Porosity

The Environmental Ellipsometric Porosimetry (EEP) investigations data are presented under isotherms of water adsorption–desorption form for each of the TiO₂-based thin films prepared at different temperatures in Fig. 3. These isotherms are reported in terms of volume percent of water adsorbed versus relative pressure of water applied at the film atmosphere. Much information can be withdrawn from such plots and the first evident one is the total porosity of each film that corresponds to the maximal volume of water adsorbed at high pressure. This information can also be deduced at low pressure when pores are emptied and match the optical characteristic of air. In these conditions a model such as the Bruggeman Effective Medium Approximation (BEMA) is ideal and always corroborates the value deduced from maximal water adsorption, providing all pores are accessible and filled with water. The maximal porosity is measured to be 43 vol% at 400 °C while at 300 °C only 27% of the total volume is accessible to water, suggesting that the remaining 16% of the total volume is composed of void. In other words, only 60% of the pore volume is accessible to water at 300 °C. This is not surprising when considering that at this temperature, pores are discrete entities which may not necessary be interconnected by intra-wall microporosity as a result of incomplete decomposition of the organic residue. It has been shown that PEO-PPO block copolymers decompose between 150 and 300 °C with a preferential elimination of the PPO core at low temperature and a more difficult decomposition of the PEO chains located inside the walls at high temperature.²⁷ It is likely that once the PPO combustion products departed the system, creating the porosity and the unidirectional shrinkage of the mesostructure, part of the pores healed up making them inaccessible to water. Such partial wounding is envisaged because of the amorphous nature of the organic-rich (i.e. complexing carbonate species) network within which oxo-polyhedra still have a certain local degree of movement. In other words, only part of the pores are accessible to water at 300 °C, while the totality is accessible at 400 °C. This sudden porosity opening can only be explained by a local migration of mater that is attributed to particle diffusive sintering at high temperature. Comparing the actual EEP and XRD results, one observes that since no diffraction peak could be detected by XRD at 400 °C, one expects that only very small particles are present and that nucleation just took place, initiating at the same time as the opening of pores. Upon further temperature increase, crystallites progressively grow and reach a plateau at 7.5 nm for 600 °C. This average particle size is in the range expected for the actual broadness of the inorganic wall separating two distinct pores. This growth is accompanied by a slight decrease of the porosity to 40% associated with the densification of the mineral phase combined with a slight additional film contraction. At 800 °C, only 30% pore volume remains, revealing a progressive thermal densification.

Fig. 3 Adsorption–desorption isotherms plotted for TiO₂ mesoporous thin films calcined at various temperatures (XXX A and XXX D represent adsorption and desorption curves respectively preceded by the temperature of treatment).

The second piece of very useful information that one can extract from such isotherms concerns the pore size distribution and the interconnection between them. All plotted isotherms show stiff adsorption and desorption slops at two different relative pressures. These hystereses are characteristic of mesoscale porosity with well define pore dimensions and pore interconnection through windows of smaller dimension (bottle necks or restrictions). Upon calcination, the hysteresis is progressively shifted towards higher relative pressures, which suggests (even after surface energy correction) a significant pore enlargement. Pore size distribution can be deduced from Kelvin's equation that characterises the liquid/solid interface curvature at the condensation. Several models can be applied that depend on the pore morphology. For the present example, we expected the pores to exhibit ellipsoidal morphology as a result of the contraction of the initial spherical pore. The anisotropy factor at each temperature was deduced from the film (or lattice parameter) corresponding to contraction as previously described. With such geometry we supposed that windows also have this ellipsoidal shape. Pore size distributions deduced from these isotherms are plotted for adsorption and desorption branches in Fig. 4.

Fig. 4 (a) Pore size distributions deduced from the adsorption branches of isotherms given in Fig. 3 and corresponding plots of the average pore size distributions deduced from the adsorption and desorption branches versus temperature of treatment.

They revealed that pores do grow in diameter from 3 to 10 nm, and become less monodimensional with increasing temperature. The evolution of the average pore dimension with increasing temperature is illustrated in Fig. 4. One observes that no significant variation occurs between 500 °C and 600 °C, suggesting a domain of higher stability of the nanocrystallites corresponding with that of the pore size (plateau regime). This stability period corresponds to the stability range of the grid-like mesostructure characterised by the constant maximal intensity of the corresponding diffraction peaks (see Fig. 1). Since in this period, pores are around 5.5 nm in diameter and particles are 7.5 nm, one confirms in a first approximation that particles have the size of the wall dimension that is deduced from the grid periodicity minus the pore dimension (13 − 5.5 = 7.5 nm). The pore interconnection is undertaken by smaller windows with averaged dimension of 3.5 nm. Larger pores are formed above this period in association with further sintering of the nanoparticles. The Kelvin's equation model is not relevant for micropores and can thus not be used to precisely characterise the porosity by analysing the adsorption/desorption branches at low relative pressures. However, one observes that close to 10% volume of water was adsorbed at these low pressures for 300 °C and 400 °C against 3% for 500 °C and 600 °C, and 1% for 700 °C and 800 °C. These values represent only 25, 8 and 2% of the total volume of pores respectively in the error of the experimental analysis. Amorphous thin films contain significant volumes of micropores, while only the grid-like nanocrystalline frameworks exhibit microporosity that corresponds to nanoparticle interstices. This becomes almost insignificant when particles grow by sintering above 600 °C.

Photoinduced catalysis investigations

Mesoporous TiO₂ reported previously did not have the structural stability to withstand higher than 600 °C calcination.⁹ The photocatalytic activity of our films was determined by the decomposition rate of methylene blue. Fig. 5 shows the absorption spectra of MB solution under UV light irradiation for a film calcined at 600 °C. The amount of MB decreased, while a blue shift of the maximum peak around 650–665 nm was observed, as the photocatalytic reaction proceeded, and this peak disappeared completely after overnight irradiation. It was reported that the blue shift was caused by photooxidative N-demethylation of MB resulting from the production of intermediate species.²¹ The amount of MB decomposition over mesoporous TiO₂ is deduced from the corresponding absorption band decrease as illustrated in Fig. 5, which revealed that the normalised activity is optimal for 600 °C treatment (see Fig. 6). Such a limit was reported at 500 °C by other groups.⁹ Slight photodegradation of MB was observed in spite of not using the TiO₂ film. Fig. 6 was obtained by normalizing with result of the blank test. At 300 °C, mesoporous TiO₂ hardly exhibits photocatalytic activity, which is due to its amorphous structure and organic residues. The amorphous structure leads to the recombination of the photo-generated electrons and holes. At 700 °C, the photocatalytic activity decreases as a result of the enlargement of the anatase nanocrystals and the corresponding reduction of the porosity. Yu et al. pointed out that the growth of the crystallite size from 8.5 nm to 15.4 nm in mesoporous TiO₂ led to reduction of its bandgap from 3.27 eV to 3.20 eV, which caused a decrease in the photocatalytic activity.⁹ As shown in Fig. 2, the crystallite size was increased dramatically from 7.5 nm to 15 nm by calcining at 700 °C. Presumably, the bandgap reduction was caused by the crystalline growth of the mesoporous TiO₂, and this reduction of the bandgap caused the decrease of the photocatalytic activity. The 800 °C calcined films would exhibit lower photocatalytic activity because of the same reasons as the 700 °C calcined one.

Fig. 5 Variation of absorption spectra of MB-TiO₂ films calcined at 600 °C upon UV irradiation.

Fig. 6 Photodegradation of MB over TiO₂ films calcined at various temperatures under UV light irradiation.

The photon irradiation having higher energy than the bandgap of the photocatalyst provides excited electrons in the conduction band and holes in the valence band. The photo-generated electrons and holes react with oxygen and water from the environment to produce active oxygen and radical species, which decompose organics into CO₂ and H₂O. The catalytic activity is thus increased for high concentrations of these reactive species in agreement with high surface area and high accessibility to O₂ and H₂O. In Fig. 6, one observes that the non-porous control TiO₂ film, which was prepared by the same method as that for mesoporous TiO₂ but without adding F127, showed lower activity than mesoporous TiO₂, confirming the benefit of the mesoporosity.

TiO₂ thin film exhibits photo-induced hydrophilic property as well as the photodegradation of organic compounds by irradiating with UV light. Sakai et al. proposed the mechanism of the photo-induced hydrophilic property.²⁸ Most of the photo-generated electrons and holes are consumed by producing active oxygen and radical species, and moreover, some of the holes are trapped at lattice oxygens in TiO₂. These trapped holes make Ti and the lattice oxygen bond unstable, and then water molecules adsorb at these bond, finally forming new hydroxyl groups. The increasing number of hydroxyl groups were ascribed to the hydrophilic property of the TiO₂ film surface. In Fig. 7, the photodegradation of LA was conducted by monitoring the water contact angle of the TiO₂ thin film surface. All films were irradiated by UV light overnight before coating with LA. Except for the 300 °C calcined film, the water contact angle of the TiO₂ films became less than 5°, indicating these films showed photo-induced hydrophilic conversion (not shown). The amorphous structure and the organic residues of the 300 °C calcined film resulted in the lack of photo-induced hydrophilic conversion.²⁹

Fig. 7 Photodegradation of LA over TiO₂ films calcined at various temperatures under UV light irradiation.

After coating with LA, the water contact angle of TiO₂ films was increased to 65–85 degrees because LA has hydrophobic properties. In the mesoporous TiO₂ films, the 600 °C calcined film exhibited the highest decrease rate of the water contact angle, indicating this film showed the highest photocatalytic activity in the other mesoporous TiO₂ films similar to the photodegradation of MB as displayed in Fig. 6. The small particle size as well as high crystallinity of this film improves the photodegradation of LA because of decreasing the recombination of the photo-generated electrons and holes. However, the photocatalytic activity of this film was inferior to that of the control film. The mesoporous structure enables a large amount of LA to be adsorbed in the TiO₂ films. During the photodegradation reaction, LA and its decomposed intermediates were diffused from the mesopores to the TiO₂ film surface, resulting in the lower decrease rate of the water contact angle over the mesoporous TiO₂ films than that over the control film.

The results of the photodegradation of MB and LA indicate that the optimal conditions are achieved by the network morphology of the grid-like mesostructure described above and stabilised at 600 °C, for which the crystalline particles remain as small as those composing the network at 500 °C.

Conclusion

The present work shows, through the very powerful environmental ellipsometric porosimetry technique, that mesostructured TiO₂ based films open their porosity to accessibility upon thermal induced diffusive sintering. It also permits the accurate definition of the pore size distribution and the interconnection between them. In the present case, it revealed that the mesoporosity is still ordered and well defined up to 600 °C. According to our photocatalytic measurements, we confirm that the activity is more efficient when the porosity is high and completely accessible to water, oxygen, and adsorbate species and when nanoparticles and pores have dimensions of 7.5 and 5.5 nm respectively. Such an optimal system was obtained after a thermal treatment of 10 min at 600 °C in air, for which the structure is grid-like.

References

1. R. Wang, K. Hashimoto, A. Fujishima, M. Chikuni, E. Kojima, A. Kitamura, M. Shimohigoshi and T. Watanabe, Nature, 1997, 388, 431.
2. R. Blossey, Nat. Mater., 2003, 2, 301.
3. B. O'Regan and M. Grätzel, Nature, 1995, 95, 45.
4. M. Yamagishi, S. Kuriki, P. K. Song and Y. Shigesato, Thin Solid Films, 2003, 442, 227.
5. A. Fujishima and K. Honda, Nature, 1972, 238, 37.
6. O. Carp, C. L. Huisman and A. Reller, Prog. Solid State Chem., 2004, 32, 33.
7. T. Torimoto, R. J. Fox III and M. A. Fox, J. Electrochem. Soc., 1996, 143, 3712.
8. I. Sopyan, M. Watanabe, S. Murasawa, K. Hashimoto and A. Fujishima, J. Electroanal. Chem., 1996, 415, 183.
9. J. C. Yu, X. Wang and X. Fu, Chem. Mater., 2004, 16, 1523.
10. J. Tang, Y. Wu, E. W. McFarland and G. D. Stucky, Chem. Commun., 2004, 1670.
11. H. Wang, J. J. Miao, J. M. Zhu, H. M. Ma, J. J. Zhu and H. Y. Chen, Langmuir, 2004, 20, 11738.
12. Y. Takahara, J. N. Kondo, T. Takata, D. Lu and K. Domen, Chem. Mater., 2001, 13, 1194.
13. M. Uchida, J. N. Kondo, D. Lu and K. Domen, Chem. Lett., 2002, 5, 498.
14. D. Grosso, C. Boissière, B. Smarsly, T. Brezesinski, N. Pinna, P. A. Albouy, H. Amenitsch, M. Antonietti and C. Sanchez, Nat. Mater., 2004, 3, 787.
15. Y. Lu, H. Fan, A. Stump, T. L. Ward, T. Rieker and C. J. Brinker, Nature, 1999, 398, 223.
16. D. Grosso, F. Cagnol, G. J. de A. A. Soler-Illia, E. L. Crepaldi, P. A. Albouy, H. Amenitch, A. Brunet-Bruneau and C. Sanchez, Adv. Funct. Mater., 2004, 14, 309.
17. D. Grosso, G. J. de A. A. Soler-Illia, F. Babonneau, C. Sanchez, P. A. Albouy, A. Brunet-Bruneau and R. Balkenende, Adv. Mater., 2001, 13, 1085.
18. E. L. Crepaldi, G. J. A. A. Soler-Illia, D. Grosso, F. Cagnol, F. Ribot and C. Sanchez, J. Am. Chem. Soc., 2003, 125, 9770.
19. P. C. A. Alberius, K. L. Frindell, R. C. Hayward, I. J. Kramer, G. D. Stucky and B. F. Chmelka, Chem. Mater., 2002, 14, 3284.
20. D. Grosso, G. J. de A. A. Soler-Illia, E. L. Crepaldi, F. Cagnol, C. Sinturel, A. Bourgeois, A. Brunet-Bruneau, H. Amenitsch, P. A. Albouy and C. Sanchez, Chem. Mater., 2003, 15, 4562.
21. T. Zhang, T. Oyama, A. Aoshima, H. Hidaka, J. Zhao and N. Serpone, J. Photochem. Photobiol. A, 2001, 140, 163.
22. C. Randorn, S. Wongnawa and P. Boonsin, ScienceAsia, 2004, 30, 149.
23. S. Yang and L. Gao, J. Am. Ceram. Soc., 2005, 88, 968.
24. H. Chang, C. Su, C. H. Lo, L. C. Chen, T. T. Tsung and C. S. Jwo, Mater. Trans., 2004, 45, 3334.
25. C. Boissière, S. Lepoutre, D. Grosso, L. Nicole, A. Brunnet-Bruneau and C. Sanchez, Langmuir, 2005 , in press.
26. B. L. Kirsch, E. K. Richman, A. E. Riley and S. H. Tolbert, J. Phys. Chem. B, 2004, 108, 12698.
27. F. Kleitz, W. Schmidt and F. Schuth, Microporous Mesoporous Mater., 2003, 65, 1.
28. N. Sakai, A. Fujishima, T. Watanabe and K. Hashimoto, J. Phys. Chem. B, 2003, 107, 1028.
29. T. S. Yang, C. B. Shiu and M. S. Wong, Surf. Sci., 2004, 548, 75.

---