Electronic Structure Calculations: Materials with Weak and Strong Correlations

Doktorsavhandling, 1992

This thesis is devoted to studies of various electronic properties which can be extracted from photoemission spectroscopy. Both weakly and strongly correlated systems are discussed, ranging from free electron metal surfaces to strongly correlated copper oxide compounds. In order to study such a variety of systems, different calculational schemes have to be employed. This work is mainly based on band structure calculations using the local density approximation. However, for strongly correlated systems the need for model Hamiltonians is crucial. Subjects discussed are: * inverse photoemission spectra from the Cd(0001) surface obtained by a multiple scattering formalism. The results are in good agreement with recent experimental data. Especially, the one-step model is used to demonstrate image states excitations by secondary beams. * artificial structures, such as semiconductor heterojunctions and short-period superlattices, which are of vital importance for modern electronic devices. A method for tuning the band offset at semiconductor interfaces is proposed; by varying (n,m) for the GaAs/(AlAs)n(GaAs)m interface, the magnitude of the valence band offset can be designed. The metallic superlattice Mo/V was also studied, with emphasis on the application of hydrogen storage. * Schottky barrier heights. Final state screening effects can influence the Schottky barrier height, as deduced from measured core level shifts. According to our calculations, such effects may be of the order 0.1-0.2 eV. * chemical shifts in Cu2O, CuO and NaCuO2, where the Cu atom is formally mono-, di- and trivalent. We find similar chemical shifts between Cu atoms in YBa2Cu3O6.5 as between the Cu atoms in the model compounds. This suggests that different Cu atoms in YBa2Cu3O6.5 are formally mono-, di- and trivalent. The Cu 2p core level spectra were calculated using the Anderson impurity model, with parameters obtained from ab initio calculations. The results are related to the formal valence of Cu, and can be explained by a charge counting argument.