Search

The selective surface modification by phosphonic acids of SiO2−TiO2 supports at the micrometer and molecular scale was investigated. Under aqueous conditions, phosphonic acids bind to TiO2 and not to SiO2 surfaces. A micropatterned support was prepared by electron beam microlithography and selectivity, of the surface modification was evidenced using scanning Auger electron spectroscopy (SAES). The second support was a mesoporous SiO2−TiO2 mixed oxide (10 mol % Ti) epoxidation catalyst prepared by sol−gel processing. Selectivity was deduced from the decrease of the catalytic activity upon modification and from chemical analysis; bonding modes to the surface were investigated using solid-state ²⁹Si and ³¹P MAS NMR. The possibility to introduce different organic groups by successive treatments with a phosphonic acid and a silylating agent was illustrated in the case of the mixed oxide.

¹⁷O-enriched phenylphosphonic acid (PhP*O3H2) was prepared from phosphonic acid dichloride and enriched water. This enriched phosphonic acid was used to prepare the titanium oxo-alkoxo-phosphonate Ti4(μ3-*O)(μ2-OPrⁱ)3(OPrⁱ)5(PhP*O3)3 (1) as a model compound, and titanium oxide/phenylphosphonate hybrids either by sol−gel processing with Ti(OPrⁱ)4 or by surface modification of a titania support. For comparison a gel sample was prepared using nonenriched PhPO3H2 and enriched water. The ¹⁷O MAS NMR spectra of these 5 solids have been recorded at various magnetic fields (7, 9.4, and 17.6 T) and the isotropic chemical shifts and quadrupolar constants have been determined for the different oxygen sites:  P&dbd;O, P−O−H, P−O−Ti, and OTix. These spectra give direct evidence for the presence of extensive Ti−O−P bonding in the hybrid materials.

Organosilsesquioxane−titanium oxide hybrids were prepared by nonhydrolytic condensation between CH3SiX3 and ³/4TiY4 (with X = Cl, Y = OⁱPr or X = OR, Y = Cl). The two routes led to hybrids in quantitative yield without formation of Si−O−Ti bonds. The lack of Si−O−Ti bonds may be due to greater rates of homocondensation of the mixture of methylchloroalkoxysilanes and titanium chloride alkoxide initially formed by redistribution reactions and/or to the instability of the Si−O−Ti bonds, which underwent redistribution reactions to form Si−O−Si and Ti−O−Ti bonds. The intermediate formation of Si−O−Ti bonds by redistribution was indirectly evidenced in solution. The heterogeneity of the hybrids has been related mainly to this instability.

Several silicon oxycarbide glasses with O/Si ratios 1.2 and 1.8 were prepared by pyrolysis below 1000 °C of organosilicon precursors. Their structure was investigated using29Si magic-angle spinning nuclear magnetic resonance. The spectra obtained are consistent with a purely random distribution of Si-O and Si-C bonds, which depends on the O/Si ratio of the glass only, regardless of the structure of the precursors. This result is interpreted by the occurrence, during the pyrolysis of entropically controlled redistribution reactions involving the exchange of Si-O and Si-C bonds.

The efficiency of the nonhydrolytic sol−gel process based on the condensation between metal chlorides and metal alkoxides has been studied in the preparation of aluminum titanate gels. Solutions of aluminum and titanium precursors in the ratio Al/Ti = 2 gave gels with the same metal ratio. The structural evolution of the samples is unusual; they directly crystallize at T > 600 °C into metastable pseudo-brookite β-Al2TiO5 and are reluctant to decompose into stable TiO2 (rutile) and α-Al2O3 at 1150 °C. This behavior must be connected to the high homogeneity of the samples originating from the nonhydrolytic sol−gel process.

The etherolysis/condensation of TiCl4 by diisopropyl ether (ⁱPr2O) as well as the direct condensation between TiCl4 and Ti(OⁱPr)4 are efficient nonhydrolytic sol−gel routes to titania. A comparison of the solution chemistry of the two routes was performed by means of NMR spectroscopy. In both cases the condensation is slow at room temperature; the true precursors are titanium chloroisopropoxides in equilibrium through fast ligand-exchange reactions. The condensation takes place at 100 °C after an induction period which drastically depends upon the OⁱPr/Ti ratio in the reaction mixture. This behavior is related to the composition of the starting chloroisopropoxides mixture. The key species in the induction process appears to be Ti(OⁱPr)Cl3, which catalyzes the first condensation reactions.