Surface magnetism of Fe(001)

Results of all-electron self-consistent semirelativistic full-potential linearized augmented-plane-wave local-density and local-spin-density studies are reported for a seven-layer Fe(001) thin film. The calculated work function for the ferromagnetic state is found to be in excellent agreement with experiment, whereas that calculated for the paramagnetic state is significantly worse (namely, 0.5 eV too large), indicating the importance of spin polarization on this electrostatic property. For both states, partial densities of states (projected by layer and by orbital angular momentum), surface states, and charge (and spin) densities are presented and their differences employed to discuss the origin of surface magnetism. No Friedel oscillation is found in the layer-by-layer charge density. The surface-layer magnetic moment is found to have been increased by 0.73${\mathrm{\mu }}_{\mathit{B}}$ from the center layer to 2.98${\mathrm{\mu }}_{\mathit{B}}$/atom; a very small Friedel oscillation is obtained for the spin density, which indicates possible size effects in this seven-layer film. Layer-by-layer Fermi contact hyperfine fields are presented: While the core-polarization contributions are proportional to the magnetic moment, the conduction-electron contribution shows a pronounced Friedel oscillation in the central layer and, significantly, a change of sign and increase in the magnitude for the surface-layer contribution. The hyperfine field at the nucleus of the center-layer atoms is found to be in excellent agreement with experiment. The net result for the surface-layer atoms is a predicted decrease in magnitude of the total Fermi contact hyperfine field despite the large increase of their magnetic moments. The relevance of this prediction to experiment is discussed.