|
Peak positions in experimental photoemission spectra are often compared with band structures or spectral densities, the latter preferably obtained by first-principles calculations. However, valuable information is contained also in the peak heights. To investigate these, a direct link between ab initio calculations for the electronic structure of the system and experimental spectra is inevitable. It is provided by the one-step model of spin- and angle-resolved photoelectron spectroscopy (SPARPES) which accounts for the geometric, electronic, and magnetic structure of the system and for the appropriate boundary conditions. Since the transition matrix elements of initial and final states are computed, direct comparison of experimental and calculated intensities allows to conclude on ground-state properties.
In the first part of the presentation I shall briefly introduce the one-step model for SPARPES in its formulation within relativistic multiple scattering theory (i.e. spin-polarized relativistic layer Korringa-Kohn-Rostoker, SPRLKKR). Based on the Dirac equation - rather than on the Schrödinger equation - spin-orbit coupling and magnetism are treated on equal footing. In the second part, effects due to spin-orbit coupling, matrix elements and scattering in SPARPES will be presented. The importance of spin-orbit coupling is elucidated by the spin-polarization of photoelectrons emitted from nonmagnetic surfaces. That a naive conclusion from the photoelectron's spin on that of the initial state appears questionable will be shown for Au(111). The importance of matrix elements will be demonstrated for Ni(111) and by magnetic linear dichroism of Fe(001). Eventually, spin-dependent scattering within the final state will be discussed for photoelectron diffraction in core-level photoemission and for spin motion in ultrathin magnetic films. |
|