The adiabatic theory of spin-density waves is developed on the basis of spin-density-functional theory. The wave-number-dependent exchange constant matrix is obtained from spin-density-functional calculations with constrained moment directions. The central assumption considers a fast electronic and a slow magnetic time scale, and postulates negligible correlation of the fast motion between different ionic sites. The parameter-free calculated magnon spectra for Fe, Co, and Ni are in excellent agreement with available experimental data. In the case of Fe, they show strong Kohn anomalies. Using Planck statistics at low temperature, the temperature dependence of the magnetization is well described up to half the Curie temperature. It is conjectured that correlated local-moment clusters survive the Curie transition. On this basis, calculated Curie temperatures are obtained within 10% deviation from experiment for Fe and Co, but 30% to low for Ni. ͓S0163-1829͑98͒04425-7͔

No abstract

It is shown that the magnon spectrum of magnetically ordered crystals can be calculated using a frozen magnon scheme and spin density functional theory in the local approximation. Ab initio calculated magnon spectra for Fe and Ni are presented. Kohn anomalies are predicted for the magnon spectrum of Fe. Considering the magnons as true low-lying thermal excitations and using mean-field semi-classical statistics at elevated temperatures, the finite-temperature magnetization and Curie temperature TC are calculated, which compare well with experiment for Fe, Co, and Ni.

No abstract

For cubic (001), ( 110) and ( 111) surface systems with in-plane or perpendicular magnetization, valence-band photoemission along the surface normal is studied analytically by evaluating electric dipole transition matrix elements between half-space initial and final states of the appropriate double-group symmetry. Explicit expressions are obtained for the spin-polarization vector of the photoelectrons, and the spin-averaged intensity and its change upon reversal of the magnetization direction, i.e. magnetic dichroism, for circularly and linearly polarized incident light. These results firstly elucidate the origin of spin polarization and dichroism in terms of an interplay between spin-orbit coupling and exchange, and secondly provide a systematic overview of possible effects. In particular, we predict new types of magnetic linear dichroism for s-polarized light in the case of magnetization perpendicular to surfaces with a twofold rotation axis and in the case of in-plane magnetization of fcc (111) or hcp (0001) surfaces.