J . Chenz. SOC., Faraduy Trans. I, 1989, 85(5), 1043-1048 Preparation and Characterization of Ti0,-SO, Aerosil Colloidal Mixed Dispersions Colin Morrison and John Kiwi* Institut de Chimie Physique, Ecole Polytechnique Fe'de'rale, CH-1015 Lausanne, Sw itzerland TiO, sol adsorption on Si0,-Aerosil has been carried out by three different methods to produce Ti0,-SiO, Aerosil sols having different mole ratios. Characterization of these particles by electrophoresis, spectroscopy and electron microscopy is presented. Corrosion of Si0,-Aerosil was not observed under photolytic conditions at pH 3 when loaded with 10 atom YO TiO,. The purpose of this study is to elucidate the role of SiO, as a substrate when TiO, colloid is dispersed on its surface via different techniques and at different loading levels.TiO, is the electron-donating species in a colloidal or a particulate semiconductor-SiO, system.' Ti0,-SiO, obtained by the precipitation method has been reported to be active in isomerization of butanes and in amination reactions., The properties of these coprecipitated silica-titania gels have also been Ti0,-SiO, glasses have recently been a subject of much attention in the ceramic field.6i ' The reverse situation to ours, namely the deposition of SiO, on TiO, surfaces, has been a subject of wide interest in material science.'. Photocatalytic studies on organic reactions1'. l1 and water-based systems over titania-silica oxide catalysts have been more recently described. l3 Experimental Hydrolysis of TiCl, was carried out by methods that have been well documented in the 1iterat~re.l~ Freeze-drying of Ti0,-SiO, systems was carried out according to work reported previously.l5 U.v.-visible and diffuse reflectance spectra were recorded on a Hitachi-Perkin-Elmer 340 recording spectrophotometer, equipped with an integrating sphere. The electro- phoretic mobility was measured with a commercial mark I1 microelectrophoresis apparatus (Rank Bros, Cambridge). The average velocity was calculated by dividing the velocity by the field strength (potential/distance between the electrodes). Electron microscopy (t.e.m.) was carried out wit! a Philips 300s instrument. The limit for resolution for such an instrument is 3 A. Continuous photolysis experiments were carried out with a Rofin Xe lamp (100 mW crn-,). Prior to being photolysed, the samples were flushed with 48Ar for the removal of oxygen.Hydrogen was analysed by gas chromatography using a Carbosieve 5 A column at 40 "C and a Gow-Mac conductivity detector. The following methods were used to deposit TiO, on Aerosils. Method A TiO, was prepared by hydrolysis of TiC1, in 1 : 1 HCl pH 0 (5 g dmP3 TiO, equiv.) at 0 "C and was subsequently dialysed to pH 3. SiO, (Degussa, Aerosil) suspension was prepared by addition of 1 g of the solid to 100 cm3 water at pH 3 and sonicating. An appropriate amount of TiO, stock was added to water at pH 3 and diluted to 100 cm3 10431044 Ti0,-SiO, Aerosil Colloidal Mixed Dispersions with water at pH 3 (adjusted with HCl). This was rapidly added to the sonicating SiO, suspension and then allowed to sonicate for a further 5 min.The suspensions were stored in plastic bottles and were periodically shaken. Method B TiO, was prepared at pH 0 and 0 "C as above and an appropriate amount was added to water at pH 0 and adjusted to 100 cm3. A suspension of 1 g SiO, in water at pH 0 was sonicated while the TiO, was added rapidly. Sonication was continued for 5 min. Later, this solution was dialysed to pH 3. Method C Method A was followed and the freshly prepared suspensions were rapidly frozen in liquid N, followed by freeze-drying. Results and Discussion SiO, of different surface areas and particle sizes were covered with TiO, (prepared from TiC1, and used in sol form) in order to optimize surface coverage for u.v.-photoinduced H, production. As shown above we have chosen three methods: (A) TiO, sol dialysed to pH 3 was mixed with a suspension of SiO, at pH 3.This first method was chosen because at pH 3 SiO, and TiO, are oppositely charged and should attract one another. Uniform deposition on the SiO, particles would be very dependent6-' on how the two components were mixed. To overcome the mixing problem, method B was devised. At pH 0, the TiO, would exist as Ti02+ ions8 and would use the SiO, substrate' for nucleation and growth. Taking into account this observation, undialysed TiO, sol at pH 0 was mixed with a suspension of SiO, at pH 0 followed by dialysis to pH 3. A third approach was also tried (C), in which the suspension formed in method A was freeze-dried to a powder. Freeze-drying was performed in order to deposit the TiO, onto the SiO, surface and to form a powder which could be resuspended in water.In such a powder it is less likely that particles would agglomerate in suspension on storage. It would also be more convenient to heat-treat. Aerosil 90 (A 90) has been reported to have a mean size of 20 nm with particle sizes varying from 12 to 40 nm. The B.E.T. surface area has been reported to be 90 m2 g-'. Aerosil 200 and Aerosil 380 had mean sizes of 12 and 7 nm with 200 and 380 m2 8-l B.E.T. surface areas, respectively. This information was kindly supplied by Degussa AG (Frankfurt). Some samples were heated in air at 400 "C for 24 h.? Solvent evaporation gives rise to surface modifications of catalyst prepared by method B on heating. This is not the case for the heating of the freeze-dried samples as prepared by method C.15 Fig.1 compares the u.v.-visible absorption characteristics of TiO, sol alone (3.5 cm3 of 5 g dm-3 in 25 cm3 at pH 3) with that of a solution phase of 10 atom O/O Ti-SiO, Aerosil 380 sols prepared by methods A and B. The solution phase of the suspensions shows only slight absorption below 300 nm, whereas the TiO, sol has a sharp absorption above 300nm. This observation confirms that the TiO, has been adsorbed onto the surface of Aerosil 380. This was found to be the case for all the samples prepared. At this point we would like to estimate the monolayer coverage for the three SiO, Aerosils used assuming that (a) the particles are spherical and (b) the TiO, is close- packed with a cross-sectional area of m2 molecule-'.SiO, A 90 (Aerosil with t The Degussa literature indicates that the surface area remains constant up to 500 "C before becoming smaller.C. Morrison and J . Kiwi 1.0 0.90 0.80 0.70 1045 1 I I 1 - - - - 0.0 1 I 20 0 300 4 00 500 600 700 800 Alnm Fig. 1. Absorption characteristics of TiO, sol alone (a) (for details see text) and of 10 atom % Ti Aerosil 380 (b). 90 m2 g-') having a mean particle size of 20 nm (ref. 16) will have a surface particle of 12.56 lo-'' m2. This renders 7.16 x 10l6 particles SiO, A 90 per gram and 1256 molecules per SiO, particle. Therefore, 1 g SiO, A 90 can hold 0.9 x 10,' molecules of TiO, (or 0.15 mmol). This is equivalent to a 0.90 atom % TiO, coverage for one monolayer. In the same way, for SiO, A 200 and SiO, A 380 monolayer coverage is estimated to be 2.0 and 3.8 atom YO TiO,.It is clear that TiO, loadings (5 and 10 atom YO) were chosen to give more than a monolayer coverage, although in one case, SiO, A 380 at 2 atom YO TiO, only a submonolayer was available. The mechanism of Ti0,-SiO, sol formation during the preparation of the catalyst involves terminal hydroxyl groups on the colloidal silica" interacting with titania hydrates so that they can be uniformly distributed during sol preparation. More important is that hydroxyl groups provide a mechanism for Ti-OH hydrates to attach to the colloidal silica particle via hydrogen bonding and condensation with surface silanol groups." Therefore, the pore spaces of the Aerosils become filled with the added titania.The heated samples (500 "C) show the tendency of these gels to undergo crystallization during sintering. l9 The presence of internal surfaces on the colloidal microstructure is a contributing factor when heat-treating. Following the absorption spectrum of heteropolyblue for silicon1' we tested SiO, dissolution after 24 h photolysis at pH 3 in Ti0,-loaded and Ti0,-unloaded materials under the conditions described in the Experimental section. This pH has been selected for our studies since corrosion to silicates and hydroxysilicates sets in only at ca. pH 10.7. Good stability for Aerosil materials in the presence of electrolytes up to pH 9 has been reported." There was no difference between the SiO, dissolution in the presence or absence of 10 atm O h TiO, at pH 3, the overall effect being rather small.Therefore, no photosensitization of SiO, induced by TiO, took place. Fig. 2 shows the diffuse reflectance spectra (d.r.s.) for the titania-silica oxide catalysts.1046 Ti0,-SiO, Aerosil Colloidal Mixed Dispersions 0.8 0.6 8 5 -e 0, 4 0.4 0.2 I I I I I I I 2 00 3 00 400 500 Ahm Fig. 2. U.v.-visible diffuse reflectance spectra of (a) SiO, Aerosil 380, (b) 2 atom % Ti-Aerosil 380, ( c ) 5 atom YO Ti-Aerosol 380, ( d ) 10 atom YO Ti-Aerosil 380, ( e ) TiO, sol. All materials were freeze-dried. Spectra were recorded at 293 K. MgCO, was used as a reference. It is clearly seen that with the increasing SiO, A 380 content the absorption band of these catalysts shifts toward shorter wavelengths. Fig. 2 therefore shows the effects of titania interactions with the silica surfaces.The hypsochromic shift increases with TiO, loadings up to a Ti0,-SiO, A 380 content of 10 atom YO. The inflection observed in the spectra around 400 nm (3.2 eV) for TiO, sol [fig. 2(e)] is due to the band gap of this material. Fig. 2 shows that higher-energy transitions take place as the silica content of the catalyst increases. Therefore, in a highly dispersed state the TiO, particles coat the SiO, colloidal substrate incompletely and the spectra reflect the dual character of these powders. TiO, anchored on Vycor glass2’ also has been reported to induce a hypsochromic shift in Ti0,-SO, catalysts analogous to results obtained in fig. 2. Fig. 3 ( a ) (trace 1) shows the electrophoretic mobility and an isoelectric point (IEP) for TiO, Degussa P-25 of pH 6.4.This value agrees well with the value of pH 6.6 reported by the manufacturer.,l Trace 2 shows the electrophoretic mobility of TiO, prepared by hydrolysis of TiCl, (see Experimental). The value found for the IEP is shifted considerably towards more acid values, showing that, despite the dialysis carried out (pH 3), the low residual chloride content plays a significant role on the TiO, surface. The mobility of SO, Aerosil 380 is shown in trace 3 as function of pH. This curve is similar to the values found for Cab-0-Sil.”. 22 An IEP of pH 2.3 is found in our studies. OtherJ. Chem. SOC., Faraday Trans. I , Vol. 85, part 5 Plate 1. For legend see over. C. Morrison and J. Kiwi Plate I (Facing p . 1046)J. Chem.SOC., Faraday Trans. I , Vol. 85, part 5 Plate 1 Plate 1. Electron micrographs of samples with a magnification of 194000~ : (a) Aerosil 380, (6) TiO, sol dried at 130 "C, ( c ) 10 atom % Ti-Aerosil 380 dialysed together (method B). C. Morrison and J. KiwiC. Morrison and J . Kiwi 1047 c -0.9 Fig. 3. (a) Electrophoretic mobility as a function of pH for mol dm-3 (NaOH + HC1) solutions: (1) TiO, Degussa P-25, (2) TiO, sol freeze-dried, (3) Aerosil 380. (b) Same conditions as in (a): (1) 10 atom YO Ti-Aerosil380, (2) 5 atom YO Ti-Aerosil 380, (3) 2 atom YO Ti-Aerosil380. Aerosils, e.g. A 90 and A 380, were found to have IEP values between pH 2.5 and 3.2. The production of these surface species depends, therefore, on the manner of preparation and thermal pretreatment of the samples. Such IEP values indicate fairly acidic oxides.It was difficult to observe the positive mobilities in the low-pH region for pure Aerosil samples, since these values were in general quite small. To obtain these data the dispersions were left overnight at the lowest pH. Equilibration times at each subsequent pH were at least 1 h and varied depending on the time required for the pH to stabilize. In this way the electrophoretic mobilities found will reflect precisely the adsorption capacity for each catalyst at a given pH. Fig. 3 (b) shows the effect of TiO, deposition on the electrophoretic mobility of Aerosil 380 as a function of pH. The IEP increases to higher pH values as more titania is added to Aerosil 380. This behaviour indicates that interaction is taking place between titania and the silica surface.The electrophoretic behaviour of both components are pH-dependent and the shape of traces 1-3 in fig. 3(b) reflect some complex phenomena at the catalyst surface. A possible explanation for this behaviour is that the Aerosil 380, when loaded with titania, is affected in its average surface charge/pH b e h a ~ i o u r . ~ ~ This is shown by the evolution of the polarization curves as a function of pH in traces 1-3. The binding links (Ti-0-Si) are readily accessible to protons at pH values between 2 and 6 as seen in fig. 3(b). The very low positive mobilities observed in fig. 3(b) seem to be due to an opening of the silica structure at low pH values and probable disruption of the (Ti-0-Si) links.* Plate 1 shows the results of electron microscopy studies.Work was carried out at 100 kV. Plate 1 (a) presents Aerosil380 powders with a mean size for the primary particle1048 Ti0,-SiO, Aerosil Colloidal Mixed Dispersions (aggregate size of the crystal) of 170 A and particle sizes between 70 aFd 400 A. Plate 1 ( b ) shows dried TiO, sol (1 30 "C), with a primary particle size of 170 A. A 10 atom % Ti/A 380 sample prepared according to method B is shown in plate 1 (c). As seen from this electron micrograph, as the samples of titania and silica mix together, the primary particles conserve their original sizes. These samples were dried at 130 "C for observation purposes. A 10 atom %/A 3800freeze-dried sample (method C ) showed particles with a primary particle size of ca.60 A. The smaller size found in this preparation agrees well with other materials prepared by this method and reported recently in our 1ab0ratory.l~ In conclusion, a few methods have been developed wherein colloidal silica is loaded with titania in aqueous solution. The representative system studied consisted of TiO, particles immobilized on negatively charged SiO, colloids (pH 3). Particle-particle interactions in aqueous dispersions containing SiO, and TiO, have been characterized by diffuse reflectance spectroscopy, electron microscopy and microelectrophoresis. These methods have been useful in reporting the loading and degree of stoichiometry of the surfaces formed. Photoinduced hydrogen evolution at pH 3 proceeds without silica corrosion. References 1 H.Hattori, M. Itoh and K. Tanabe, J. Catal., 1975, 38, 172. 2 M. Itoh, H. Hattori and K. Tanabe, J. Catal., 1974, 35, 225. 3 S. Kaneko and K. Tsukamoto, Chem. Lett., 1983, 1425. 4 H. Morikawa, T. Osuka, F. Marumo, A. Yasumori and M. Yamane, J. Non-cryst. Solids, 1986, 82, 5 R. Butz and H. Wagner, Phys. Stat. Sol., 1986, 94, 71. 6 W. Beier, A. Goktas and G. Frischat, J. Am. Ceram. Soc., 1986, 69, C148. 7 C. Sherer and C. Pantano, J. Non-cryst. solids, 1986, 82, 246. 8 D. Furlong, K. Sing and G. Parfitt, J. Colloid Interface Sci., 1979, 69, 409. 9 R. Iler, US Patent 2885366, 1959. 97. 10 S. Kodama, H. Nakaya, M. Anpo and Y. Kubokaba, Bull. Chem. SOC. Jpn, 1985, 58, 3645. 11 V. Nikisha, B. Shelimov and V. Kazanskii, Kinet. Catal., 599, 15, 676. 12 M. Anpo, H. Nakaya, S. Kodama, Y. Kubokawa, K. Domen and T. Onishi, J. Phys. Chem., 1986,90, 13 A. Frank, I. Willner, Z. Goren and Y. Degani, J. Am. Chem. SOC., 1987, 109, 3568. 14 J. Kiwi, Chem. Phys. Lett., 1981, 83, 594. 15 K. R. Thampi, M. Subba Rao, W. Schwarz, M. Gratzel and J. Kiwi, J. Chem. Soc., Faraday Trans. 1, 16 D. Boltz and J. Howell, Colorimetric Determination of Non-metals (John Wiley, New York, 1978), p. 17 Aerosil Pigments, Descriptive no. 1 I, 23, Degussa AG, Frankfurt, Federal Republic of Germany. 18 Colloid Science, ed. H. Kruyt (Elsevier, Amsterdam, 1952). 19 G. Parfitt, Progr. Surf. Membr. Sci., 1976, 11, 181. 20 M. Anpo, N. Aikawa, Y. Kubokawa, M. Che, C. Louis and E. Giamello, J , Phys. Chem., 1985, 89, 21 Degussa Pigments, Aerosol Process, Hanau 1, Federal Republic of Germany, 1977. 22 R. Harding, J. Colloid Interface Sci., 1972, 40, 164. 23 R. James and G. Parks, in Surface and Colloid Science, ed. E. Matijevic (Plenum Press, New York, 1633. 1988,84, 1703. 430. 5689. 1983), vol. 12, p. 119. Paper 8/01390J; Received 1 lth April, 1988