Abstract
The metal K-edge absorption spectra of 3d transition metals and oxides are measured by x-ray-absorption spectroscopy. Experiments are performed on metal foils or fine powders dusted onto Scotch tape, at the Cornell High Energy Synchrotron Source. Comparisons to theory are made for both the metal and oxide data. We concentrate on the spectral nearedge structure (NES) and attempt to interpret its origins. For the metals, we find excellent agreement between the data and one-electron calculations. This agreement is much closer than previously found for the L23 edges in the same materials. The picture for the oxide near-edge structure is not so auspicious. Although symmetry-based molecular-orbital (MO) theory has been widely invoked, it is probably inadequate. Recent calculations indicate that while some of the near-edge peaks are indeed attributable to one-electron MO-type transitions whose energies are modified by the presence of the core hole, other features have no such simple origin. Finally we compare the oxide NES of the metal L23,K edges and the oxygen K edge altogether. These are found to agree with each other and to MO theory only for TiO2. The mismatch of the three NES's in the other oxides indicates the failure of one-electron theory for insulators. Instead, the near-edge structure must be dominated by core excitons, as previously shown for NiO. The presence of excitons explains the lack of agreement between the NES for core excitations of different atoms in the same solid.