.. ::::YS::CA' . 4 . V:: W CONDENSED MATTER THIRD SERIES, VOLUME 39, NUMBER 6 15 FEBRUARY 1989-II Shadow-cone-enhanced secondary-ion mass-spectrometry studies of Ag(110) Che-Chen Chang' and Nicholas Winograd Department of Chemistry, The Pennsylvania State University, 152 Davey Laboratory, University Park, Pennsylvania 16802 (Received 31 August 1988) The yield of particles ejected due to keV ion bombardment of Ag t 110( has been observed to de- pend strongly on the angle of incidence of the primary ion beam. The desorption angles of these particles are found to reflect both the crystallographic and chemical structure of the surface. Molecular dynamics calculations indicate that the desorption yield of all particles is significantly enhanced when the shadow cone of the incident beam of particles created by a surface atom inter- sects adjacent surface atoms. These angles may be accurately calculated by use of a full three- dimensional trajectory calculation of the scattering or by simple computation of the angle of this in- tersection using a two-body Moliere interaction potential function. With this approach it is possible to assign all of the angular anisotropies observed from secondary ions ejected from AgI 110I. More- over, a detailed analysis suggests that the topmost atomic layer is relaxed inward by (7. 8+2. 5)% and that the spacing between the first and third layer is contracted by (4. 1+2. 1)% relative to the bulk spacings. These results are in excellent agreement with recent Rutherford backscattering ex- periments. A number of new possible applications of this technique to the analysis of chemisorbed overlayers is also discussed. I. INTRODUCTION There has been considerable recent interest in employ- ing energetic ion beams in surface-structure characteriza- tions. Using MeV ion backscattering methods, for exam- ple, it has been possible to determine atomic locations with high precision and to unravel rather complex sur- face reconstructions. ' These determinations are possi- ble since the incident ion may be channeled or blocked by the ordered array of near-surface atoms such that atomic locations may be straightforwardly deduced from the an- gular dependence of the scattered-ion intensity. At lower energies, similar studies are possible by examining the in- tensity of ions backscattered through nearly 180. Here the scattered-ion intensity is sensitive to the incident con- dition where the shadow cone of a surface atom intersects an adjacent surface atom. This experimental config- uration, termed impact-collision ion-scattering spec- trometry (ICISS), has proven important in evaluating semiconductor surface rearrangements, in determining the structure of chemisorbed layers, and in measuring surface Debye-Wailer factors. Finally, by examining the angular distributions of ion-induced desorbed particles, the bonding geometry of chemisorbed atoms and mole- cules may be determined using rather straightforward channeling and blocking arguments. ' It is possible to combine a number of the above features to develop a new approach to ion-beam surface- \ structure studies. In our experiment, we take advantage of the fact that the shadow cone created by collision of the primary ion with a surface atom focuses Aux to specific surface coordinates. When the tail of this sha- dow cone strikes a surface or near-surface atom, the desorption yield should dramatically increase. With knowledge of the angle of incidence of the primary beam and the shadow-cone shape, it is then feasible to deter- mine surface geometries using a strategy similar to that developed for ICISS. In this work, we examine the basic concepts behind this unique scheme using the clean AgI 110I surface as a model. A number of theoretical approaches are exam- ined to check our proposed explanation of observed an- gular anisotropies. These range from full three- dimensional computer simulations of the ion-impact event to very simple approaches using only two-body in- teraction potentials to describe the shadow-cone shapes. The results of these calculations yield predictions which are in excellent agreement with previous Rutherford backscattering studies of the Ag I 110 I surface reconstruc- tion. ' Our studies suggest that the truncated top layer of the bulk crystal is relaxed inward by (7. 8+2. 5)% and that the spacing between the first and third layer is con- tracted by (4. 1+2. 1)% relative to bulk values. The im- plications of this new experimental scheme for determin- ing the structure of adsorbed layers are discussed. The possible advantages and disadvantages of this approach over ICISS are also considered. 39 3467 1989 The American Physical Society