## General

Using spin-polarized band structure calculations the magnetic properties of ferro-, antiferro- and ferrimagnetic materials are studied by the Linearized Augmented Plane Wave (LAPW) method. The Fixed Spin Moment (FSM) method (K.Schwarz and P.Mohn, J.Phys.F: Metal Phys, 14, L129, 1984)- originally implemented into the Augmented Spherical Wave (ASW) scheme - can be used in cases where the total energy as a function of the magnetic moment is needed for further analysis.

## INVAR Alloys

INVAR alloys (e.g. Fe65Ni35) have an almost vanishing thermal expansion coefficient in a certain temperature range. This behaviour is due to a strong magneto-volume coupling and can be explained on the basis of itinerant spin fluctuations.

## Non-collinear Magnetism

While for many magnetic materials the magnetic moments at all atoms point in the same direction (or have a simple antiferromagnetic up-dn alignment) there are other materials where the direction of moments varies from atom to atom. This is called non-collinear order of the magnetic momenta.
In order to calculate properties of such systems we have to extend our LAPW code (WIEN2k). The implementation of non-collinear magnetism utilises a rotated spinors basis set inside atomic spheres, and pure-spinors basis inside the interstitial. This allows calculations in both, the atomic-moment-approximation as well as in full non-collinear mode, and inclusion of spin-orbit coupling (for heavier elements) as well as LDA+U corrections (for "correlated" systems) are also possible. In the atomic-moment-approximation only the diagonal part of the spin-potential and the density matrixes inside the atomic spheres are taken into account, but in the interstitial region the full potential and density matrix is used. Often this is a very reasonable approximation since non-collinearity within an atom is small, but from on atom to the next the magnetisation rotates and this is taken into account properly. In the full mode also off-diagonal terms inside the spheres are considered and highest precission of the calculation is reached.