Use of spectroscopy for studying atomic and magnetic structure

Ondřej Šipr, Pavel Machek

Our focus is on qualitative and quantitative understanding of relation between geometric and spectroscopic properties of materials, with special emphasis on clusters and nanostructures. Studying these materials is interesting both for fundamental reasons (they form a link between atoms and solids) and for practical reasons (because of potential use of these materials in information and other technologies).

X-ray spectroscopy of solids makes it possible to gain information about electronic structure of materials, with special emphasis on its local aspects - it is chemically selective, meaning that by tuning the energy of the x-rays very can choose atoms of which chemical species will be probed. Because of the high intensity of x-ray radiation produced by current sources (synchrotrons), one can study local electronic and magnetic structure related even to those atoms which are present in very low concentrations. Therefore x-ray spectroscopy is especially well-suited for studying multicomponent materials such as those used in spintronics (layered nanostructured, diluted magnetic semiconductors). In many cases, x-ray spectroscopy is the only technique to study local magnetic moments (and consequently local aspects of magnetic anisotropy).

By analyzing x-ray spectra and x-ray magnetic circular dichroism (XMCD), the information about magnetic moments can be obtained in an indirect way only, via the XMCD sum rules. A reliable application of these rules requires a solid theoretical background, otherwise the results are ambiguous or questionable. Our theoretical approach is based on the Green's function or multiple-scattering formalism in a real space, which naturally emphasises the link between the geometric and electronic structure.

An example of a study of the interrelation between magnetism and structure: Local magnetic moments in Fe and Co clusters of 2-9 atoms supported by the Au(111) surface decrease approximately linearly with the number of neighbors (upper graphs). On the other hand, for free clusters fo the same size and shape no such correlatino appears. Different symbols in the graph correspond to different cluster sizes, as indicated in the legend.