Abstract
Photoemission, Auger, and x-ray absorption spectroscopy have been used in a systematic study of small metal clusters on a variety of supports. The degree of cluster-support interaction for clusters of group-VIII and noble metals can be divided into two categories: supports with localized p or d orbitals with binding energies overlapping those of the cluster d orbitals and supports without such orbitals. The first type is considered strongly interacting, whereas the latter type is only weakly interacting. For weakly interacting substrates such as carbon, the energy shifts in photoemission, Auger, and x-ray absorption edges, as well as changes in x-ray edge intensities, photoemission valence-orbital intensities and splittings, and photoemission and Auger linewidths all show that initial-state properties are much more sensitive to cluster size than are the final-state properties. The photoemission spectra of small clusters and those of alloys and intermetallic compounds are quantitatively compared. For weakly interacting substrates and host metals, the photoemission spectra of clusters and alloys are virtually identical, depending only on the average coordination number n¯. In these systems the net interatomic charge transfer to the substrate or host atoms is very small. However, there is a significant intra-atomic charge transfer, which increases the d-electron count with increasing cluster size or alloy concentration. For strongly interacting supports, the cluster binding energy is usually shifted to lower binding energy. This shift can be understood from simple molecular-orbital arguments. The experimental conclusions are supported by calculations with the use of the thermodynamic model of Johansson and Mårtensson. Their model accurately predicts the observed binding-energy shifts and shows that initial-state effects dominate for weakly interacting systems and that final-state processes are relatively more important for the reactive substrates.