IFF
Scientific Report 2000/2001

The institute Theory III investigates the mechanisms of the formation of structures and their consequences in condensed matter. The research starts from electronic properties which define the shortest length and time scales, but it also encompasses the macroscopic consequences. The analytical and numerical investigations are in many ways closely connected with experimental studies performed in other groups of the IFF, but also with activities in other institutes of the Research Center Jülich.

Central points of interest for the research in Theory III are in the field of electronic structure of solids (F&E-Nr. 23.20.0). Materials classes under consideration are metals and semiconductors, specifically with respect to their importance for information technology (F&E-Nr. 23.42.0). A second mainstream is formed by cooperative phenomena in condensed matter (F&E-Nr. 23.15.0). Questions here aim at the dynamics of structure and pattern formation and the statistical mechanics of order and disorder processes. Specific activities in the field of complex fluids (F&E-Nr. 23.30.0) are concerned with structure and dynamics of soft matter. The research of Theory III employs all analytical and numerical techniques applicable to many-body problems in condensed matter. In addition the development of new methodological concepts and numerical procedures is part of our research interest. The development of parallel program codes adapted to massively parallel computers has received special attention in recent years.

The explanation of the microstructure and dynamics of real solids requires the understanding of the electronic properties. One of the most important methods for the calculation of the electronic structure of real solids is the density functional theory in connection with appropriate numerical procedures. While in recent years bulk properties of metals and semiconductors have been at the center of our research a main concern now is directed towards the understanding of surface and interface properties, with particular emphasis on magnetism.

The technological success of the GMR-effect has also triggered new theoretical efforts in layered structures. A particular related problem is the tunneling of electrons through epitaxial multilayers of the type ferromagnetic-insulating-ferromagnetic. It was recently demonstrated by us that the tunneling process in this case can be understood in terms of the complex band structure of the bulk insulator, specifically the metal-induced gap-states in the energy-region of the gap. The decay of the wavefunction into the insulating region is described by an imaginary part of the Bloch vector. The spectral distribution of this part of the wave-vector has been calculated for various semiconductors like Si, GaAs, etc. One result is that in most cases and for large enough film thickness the tunneling is dominated by states of normal incidence on the interface. Based on these results we can discuss the spin-dependent tunneling in those junctions.

The growth of such layered structures is achieved through hetero- or homo-epitaxy, where the layers grow on some substrate of different or of the same material. For hetero-epitaxy with large lattice mismatch this growth-process does not give homogeneously growing layers but results in isolated islands growing in height, but not in width. To understand the atomistic reasons for different growth processes, pseudopotential-calculations based on local density-functional concepts were performed. In particular it was tried to understand the influence of a surfactants like As, Sb which can change the Stranski-Krastanov island growth into layer growth. These calculations give detailed explanations for the different observed growth modes and they are in full agreement with experiments for example for homo-epitaxial growth of As- or Sb-covered Si(111)-surfaces.

On larger scales the lattice-misfit between the substrate and a cluster adsorbed during the hetero-epitaxy-process leads to repulsive elastic forces between the adsorbed atoms of the cluster. This is a collective effect which becomes increasingly important with the size of the growing adsorbate cluster. We have studied the influence of the resulting elastic stress onto the growth-modes of the clusters. Large-scale computer-simulations employing a recently developed multigrid-scheme gave quantitative results for the different growth patterns. A scaling-theory was able to explain the large scale behavior. New laws for the growth-rates depending on a new elastically controlled length scale are obtained.

The investigation of glasses and supercooled liquids as disordered materials was further extended by molecular dynamics methods. The spatial heterogeneity is a major reason for the non-standard dynamics of the glass. In particular it is thought that this heterogeneity is the reason for non-exponential relaxation observed in experiments. To understand this effect, Lennard-Jones models and several other realistic microscopic models for glasses were investigated. Large-scale simulations were performed concerning the diffusion and relaxation properties and their dependence on temperature and time. A non-Gaussian distribution of fluctuations is observed and traced back to collective hopping of groups of atoms. Based on our simulations we are lead to assume that this mechanism is common to all glass-forming systems. In consequence this heterogeneity will always increase in the intermediate time-domain, the beta-relaxation region, at all temperatures and in many types of materials.

Microstructural features evolve in crystalline solids by reaction-diffusion kinetics of defects. The results depend crucially upon the spatial dimension of the underlying diffusion process – as is well known for example in the process of Ostwald-ripening of inclusions – which here is commonly assumed to be three-dimensional. During irradiation of metals, however, clusters of interstitials will rather give rise to one-dimensional diffusion. It can even be argued, that additional transversal diffusion can lead to reaction-diffusion kinetics somewhere intermediate between one and three dimensions. An analytical theory for this effect is fully confirmed by Monte-Carlo simulations.

A non-scientific but nevertheless seemingly important activity during this past year 2000 was the organisation of the events for "2000: Jahr der Physik", a joint effort by the German Physical Society (DPG) and the sponsoring Bundesministerium für Bildung und Forschung (BMBF), an effort towards the "public understanding of science". Some 450 physicists have been involved in the presentation of exhibitions and lectures during five major events in Berlin and Bonn, the number increasing to about 2000 when one considers furthermore the around 200 satellite-events throughout the country. The reports about resonance in the public and in the press have been rather encouraging. As the coordinator for the German Physical Society I would like to thank the board of directors of the Research Center Jülich, J. Treusch, R. Wagner and H. Grübel, for their emphasis and continuous support during this year. I thank my colleagues at IFF for their help and patience in many occasions. My very special thanks go to the members of this institute Theory-3 for their understanding, support and encouragement, and for their generous tolerance in this special year 2000.

H. Müller-Krumbhaar

Complex Band Structure and Tunneling through Insulators

Ph. Mavropoulos, N. Papanikolaou, M. Freyss, R. Zeller, and P. H. Dederichs
PDF-Datei, ca. 0.5MB