Spin-orbit coupling induced phenomena at metal surfaces and interfaces; first principles investigation of magnetocrystalline anisotropy and magnetooptics

The most prominent spin-orbit coupling (SOC) induced phenomena in surface magnetism, namely, the magnetocrystalline anisotropy (MCA) and the surface magneto-optical Kerr effect (MOKE) are investigated with first-principles calculations by means of the all-electron full-potential linearized augmented plane-wave (FLAPW) method within the local spin density approximation. Accuracy, as well as computational reliability, is achieved by adopting the second variational scheme for SOC inclusion and the state tracking and torque methods for MCA calculations. The results for (001), (110) and (111) oriented Co/Cu _n superlattices (n $\le$ 5) show a marked dependence of the interface MCA on both the orientation and the thickness of the Cu layer due to the hybridization strength of Co and Cu d bands that is strongly affected by the different geometry. For the (110) surface, we calculate the in-plane MCA employing fcc Co (110) as a free standing monolayer or as an overlayer on a Cu substrate. We find an in-plane MCA of the same order of magnitude as the perpendicular MCA, and exhibiting a significant two-fold anisotropy—as expected from the phenomenological approach. Introducing the in-plane strain and the non-magnetic Cu substrate results in a large enhancement of the in-plane MCA, which is attributed to the different influence of the strain induced change of band structure for different &phis;, where the coupling between the out of plane d states plays a crucial role.; We develop a first principles approach to calculate the MOKE from the optical conductivity determined by linear response theory in the long wavelength limit applied to the single-particle eigenvalues and wavefunctions obtained by the FLAPW method. We apply this approach to bulk bcc Fe and fcc Co and obtain the conductivity and Kerr spectra that are comparable to experiment and with negligible anisotropy about the magnetization direction. From applications to the CoPt_n (n = 1 and 3) alloys and multilayers, we find that the Kerr spectra is very sensitive to structural changes and composition, which is explained by the change of band structures in the different geometries rather than by the correlation between the MOKE and SOC strength. The strong spectra of these systems is mostly attributed to the large SOC strength of Pt. Finally, we carry out MOKE calculations for Fe and Co free standing thin films and find that the surface effects cause a red-shift for the peak at the mid-visible light energies and a reduction of the infrared spectra mainly due to band narrowing and enhanced magnetization, both of which imply better agreement with experiment.