PNNL

Abstract

A major reason for the departure of core level X-ray photoelectron spectra (XPS) of transition metal cations in oxides from the predictions of atomic models is shown to arise from ligand field splittings in the initial state of photoemission. This splitting often leads to a change in the spatial degeneracy of the initial state but the consequences of this for the XPS have not been identified in prior work. Further changes arise from ligand field splittings in the core-hole final states. Results are reported for non-empirical cluster model many body wavefunctions for the 3p XPS of Cr2O3. The agreement of the theoretical cluster model XPS with experiment is considerably improved over the pure atomic model. Furthermore, the treatment allows screening of the core hole through changes in the covalent character of the cluster orbitals. This is quite different from the usual description of screening in oxides within the framework of charge transfer configurations and it offers new insights into the role of charge transfer for satellite structure.