Cluster models for predicting structural and chemical properties for adsorbates in metal and oxide surfaces

The molecular cluster model approach for the description of processes at surfaces is presented. It involves using finite clusters which contain surface atoms interacting with atomic or molecular adsorbates. Accurate $ab \ initio$ wavefunctions can be obtained for the clusters and thus makes it possible to describe the bonding in terms that are farniliar for molecular systems. The adsorbate-substrate interaction can be analyzed by means of new techniques developed to characterize the bonding as covalent or ionic. This is shown for F on Ag(lll) and O and CO on Cu(100) systems. The measures of adsorbate ionicity are (a) the expectation value of a projection opearator giving an indication of the number of electrons associated with the adsorbate, (b) the analysis of the dipole moment curve as function of the distance, and (c) the effect of an external electric field on the adsorbate geometry. Finally, the origin of the vibrational shifts for CO adsorbed on oxide surfaces will be briefiy discussed