JOURNAL OF CATALYSIS131, 260--275 (1991) The Formation of Titanium Oxide Monolayer Coatings on Silica Surfaces S. SRINIVASAN,* A. K. DATYE,* M. HAMPDEN-SMITH,? I. E. WACHS,:~ G. DEO,$ J. M. JEHNG,:~ A. M. TUREK,:~ AND C. H. F. PEDEN§ Center for Microengineered Ceramics and Departments of*Chemical and Nuclear Engineering, and tChemistry, University of New Mexico, Albuquerque, New Mexico 87131; ~Zettlemoyer Center for Surface Studies and Department of Chemical Engineering, Lehigh University, Bethlehem, Pennsylvania 18015; §Sandia National Laboratories Division 1846, P.O. Box 5800, Albuquerque, New Mexico 87185 Received February 25, 1991; revised April 19, 1991 The formation of a dispersed titanium oxide layer on Cabosil-fumed silica and on nonporous silica spheres was studied by infrared and Raman spectroscopies and by transmission electron microscopy (TEM). The procedure for obtaining the titania coatings involved reacting the silanol groups on the silica surface with titanium alkoxides under a N 2 atmosphere. This self-limiting reaction led to a coating of dispersed titania on the silica spheres with a weight loading between 0.5 and 1.4 x 10 -3 g/m 2. The dispersed titanium oxide on the silica spheres was visible as a surface texturing of the silica in TEM images, and led to over two orders of magnitude increase in the reactivity of the silica spheres for l-propanol dehydration. Raman spectroscopy and TEM confirmed that the dispersed titania was stable to calcination in dry air at 973 K or to heating under a vacuum of 2 x 10-7 Tort up to 1058 K. However, under alcohol dehydration reaction conditions, the dispersed titania transformed into crystals of anatase, 3 nm in diameter. On Cabosil-fumed silica, on the other hand, a similar preparation resulted in a titania loading (per square meter) that was only 7% of that seen on the silica spheres. Higher loadings caused the appearance of bands due to crystalline TiO2 (anatase) in the Raman spectra. The lower monolayer capacity on Cabosil silica can be correlated with the presence of singly bound hydroxyls as seen by IR. The Stober spheres on the other hand show hydroxyl bands that show significant hydrogen bonding. © 1991Academic Press,Inc. INTRODUCTION There is growing interest in synthesis of dispersed transition metal oxides as novel catalytic materials (1). The catalytic behav- ior of a dispersed oxide is often remarkably different from that of the bulk oxide. Exam- ples of such behavior are the observed selec- tivity of dispersed vanadia on titania in se- lective catalytic reduction of nitric oxide with ammonia (2) or in the generation of strong acid sites in systems such as WO3/ AI203 (3). The structures of these dispersed oxides have been characterized by tech- niques such as Raman spectroscopy (4), EXAFS (5), and selective chemisorption (6) and the reactivity of the active sites has been probed by the use of model catalytic reac- tions. In studies of dispersed oxides, it is often important to know if the oxide is pres- 0021-9517/91 $3.00 Copyright© 1991by Academic Press,Inc. All rightsof reproduction in any formreserved. ent as a dispersed phase or whether it has aggregated into crystals. At low loadings, the small size of these crystals causes sig- nificant line broadening in X-ray diffraction (XRD) powder patterns. Hence, the ab- sence of lines from the crystalline phase in an XRD pattern does not necessarily rule out the presence of small particles of the crystalline oxide. While high-resolution transmission electron microscopy (TEM) can be used for the detection of particles smaller than 1 nm, the contrast from the support often makes unambiguous detection difficult (7). The problem becomes espe- cially acute when the dispersed phase is present as a two-dimensional overlayer and the contrast is insufficient to distinguish the dispersed phase from the supporting oxide. Hence, in previous studies of dispersed ox- ides, TEM has been used most often only 260