Horst Hoffmann

Department of Physics, University of Regensburg, Germany

During the sixties uniaxial permalloy films were intensively investigated because of possible application as magnetic thin film memory. According to the Stoner-Wohlfarth single domain theory the magnetization of these films was expected to switch coherently at the anisotropy field HK from one stable state to the other with opposite direction of the magnetization. Hoffmann developed the ripple theory which included the influence of inhomogeneous local anisotropies, on the magnetic properties of the film. Due to local anisotropies, which are present event in the best films, coherent switching never can happen. Before the applied field approaches the Stoner-Wohlfarth astroid, the magnetization ripple blocks the magnetization reversal. The film breaks up into domains. The magnetization of the film finally is reversed by slow domain wall motion. The remnant state suffers of creep, which finally destroys the stored information. The magnetic thin film memory, as originally planned, was never produced, despite of year-long investigations.

Nanostructured arrays of ferromagnetic dots are now under discussion as magnetic storage systems. Will the former problems, which for this reason will be review in the first part, come up again? Even the best dots will include inhomogeneities. The situation seems to be even more serious than mentioned above, since we have now to include the external demagnetizing field from the edges of the nano dots. Dipolar coupling between the dots can be neglected by selecting the proper spacing. The saturation field of the dots is not given by their shape anisotropy (Stoner-Wohlfarth) but by the demagnetizing field. This field also dominates the magnetization reversal. The reversal starts with nucleation, which means local deviations of the magnetization from the direction at saturation. Curling as well as nucleation of reversed or partly reversed domains give such local deviation. The nucleation field and the stability at nucleation is most important for application. What about the creep problem, which has to be discussed?

The properties of multilayer films, for instance exchange coupling in giant magneto resistance films, depend on the structure of the interfaces. Several methods of structural investigations, partly on atomic scale will be discussed. Roughness is a very general terminus. One has to distinguish between mesoscopic and microscopic roughness as well as intermixing. The influence on the properties of the multilayer system is different for each part.