The interpretation of near edge structure

Abstract

2014 The near edge structure up to 30 eV above the ionisation threshold of inner shell edges can be interpreted either as the result of interference of waves scattered from neighbouring atoms or in terms of the local projected conduction band density of states. Methods of calculating near edge structure based on both these viewpoints will be discussed and applications to the near edge structures observed in semiconductors and insulators reviewed. Particular emphasis will be given to those cases where the calculations lead to a simple interpretation of near edge structure. The characteristic inner shell edges observed in electron energy loss spectroscopy have traditionally been used for elemental analysis. Now that spectrometers capable of giving about 1 to 1.5 eV resolution on conventional TEMs and 0.3 to 0.5 eV resolution on cold field emission STEMs have become available the observation of fine structure oscillations has become routine. In particular the large modulations within the first 20 to 30 eV above threshold have become easy to observe and can be detected even when the element is present in low concentrations (below about 10% relative to the majority constituent). It would be advantageous if the fine structures could give some new information on the microstructure of the material, especially if that information could not be obtained in other ways. The materials scientist would like to know whether the near edge structure gives information about elemental charge state, local bonding or the local nature of the conduction band. Alternatively the fine structure might be sensitive to environment or coordination number. Another possibility is that it only reflects interatomic bond distances, but in a more convoluted way than the much weaker extended fine structure which starts about 30eV above threshold. One approach to understanding near edge structure is to collect "fingerprints" from elements in similar environments and try to identify common features (see the article by Brydson in this volume). The other approach, which is adopted in this paper, is to calculate the near edge structure using appropriate theory and identify those features which show sensitivity to a property of interest. In practice both methods can be valuable and they readily complement each other. Calculations have traditionally relied on two theoretical approaches. The most common method, which has been extensively used by the X-ray absorption community, is the cluster calculation or XANES method of Durham et al. (1982) [1]. The key idea is that the fine structure Article available at