518 Nuclear Instruments and Methods in Physics Research B13 (1986) 518-524 North-Holland, Amsterdam ANGULAR DISTRIBUTIONS OF IONS DESORBING FROM TiO, RICHARD L. KURTZ, R. STOCKBAUER and T.E. MADEY zyxwvutsrqponmlkjihgfedcbaZYXWVUTS Surface Science Division, National Bureau of Standards, Gaithersburg, MD 20899, USA The dependence of the electron- and photon-stimulated desorption (ESD, PSD) O’-ion yield on surface preparation from TiO, (110) and (001) surfaces has been studied. Angle-integrated electron-stimulated desorption yields have been measured versus annealing temperature from room temperature sputtered surfaces to 900°C annealed surfaces. Both the surface cation valence state and the surface geometry change as a function of annealing temperature, giving rise to a rich variety of ESD ion angular distribution (ESDIAD) patterns. These patterns are discussed in terms of possible models of local surface structure. 1. Introduction Electron- and photon-stimulated desorption (ESD, PSD) from TiO, surfaces are of importance because TiO, is generally regarded as a prototype system for the inter-atomic Auger decay model of ion desorption [l, 21. As part of an extensive study of the mechanisms and applications of ESD and PSD, we have been investigating the electronic excitation processes that lead to ion formation and desorption from single crystal surfaces of TiO, [3]. The details of the electronic excitation mechanism will be discussed later; the pre- sent paper is primarily concerned with the relationship between TiO, surface structure and the ESD ion angular distributions (ESDIAD). In ESD/PSD, the ion desorption directions are related to the orientation of the bond ruptured by the decay of the initial excitation and ESDIAD has been shown to be a useful tool for studying the structures of adsorbed molecules on metal surfaces [4]. The present study is the first application of ESDIAD to the surfaces of a bulk oxide. Information about the surface structure and bonding geometry of 0 ligands on both the (001) and (110) surfaces of TiO, (rutile) has been inferred from these measurements and compared with detailed models. TiO, rutile is one of the most studied transition- metal oxide systems and the preparation and charac- terization of its surfaces have been reported previously [5-91. Surfaces that have been bombarded by 5OOeV Ar’ have been shown to be oxygen deficient by Auger, X-ray and ultraviolet photoelectron spectroscopies: such surfaces have an excess of Ti, and the associated Ti 3d electron population is clearly evident in UPS. Annealing sputtered surfaces to temperatures above 400°C in ultrahigh vacuum (UHV) causes the surface to regain its stoichiometry via oxygen diffusion from the bulk [S-7]. The UPS spectra from these annealed surfaces show that the Ti 3d levels are indeed depopu- lated with the highest occupied level being the 0 2p, located between 3 and 9 eV below E, [5,8]. These spectra are identical to those obtained from UHV- fractured surfaces. The surface geometry is also affect- ed by annealing. The (110) surface has been reported to be stable to 900°C giving good (1 X 1) LEED patterns [7,8] although point defects have been re- ported to be introduced for T 2 600” [lo]. The (001) surface is not thermodynamically stable and it has been reported to facet; below 1000°C the facet planes are (011) (ref. [ll]). Recent synchrotron radiation studies have implied that ion desorption from TiO, surfaces exhibits strong yield dependences and angular anisotropies in the ion emission directions that depend on the specifics of surface preparation [3]. We have studied the yield behavior by performing angle-integrated electron- stimulated desorption (ESD) measurements and we have observed the angular distributions of emitted ions using ESDIAD. The results from these studies give specific information on the local surface structure and can assist in determining the probability that an O- ligand in a particular surface site will be involved in an inter-atomic Auger decay and desorb as an ion. Following this introduction, the details of the experi- mental technique are discussed in sect. 2. Results from the yield measurements and ESDIAD patterns are presented in sect. 3. Sect. 4 contains a discussion of the results which are interpreted in terms of plausible models of surface structure. 2. Experiment The TiO, single crystal used in these measurements was a rod 4 mm X 5 mm X 2 cm with two of the long sides aligned, cut, polished and etched to within 0.5” of the (110) and (perpendicular to this) the (001) crystal plane. This allowed a complete series of measurements