Author:
O’Keefe Michael A.,Sattler Margaret L.
Abstract
Image simulation has become one of the preferred techniques for analysis of high-resolution transmission electron micrographs, in both bright-field and dark-field modes. This is especially true of microscope images used in stuctural studies, both for perfect crystal structures, and for defects within periodic structures. In using image simulation for structural analysis, comparison is made point-by-point (pixel by pixel) between the experimental image and one simulated under identical imaging conditions for a model structure. Comparison with a matching simulated image enables features in the experimental image to be identified as belonging to structural features in the specimen, such as groups of atoms, or individual atoms. In the case of amorphous structures, however, no such one-to-one correspondence between simulations and experimental high-resolution images can be expected. It is thus much more difficult to determine whether the model from which one is simulating images really does describe the appropriate amorphous structure. Amorphous structures are characterized not in terms of atom positions within a well-defined unit cell, but interms of a “radial distribution function” (RDF), a function that gives the average number of atoms lying at any given distance from an average atom. The RDF is thus a non-periodic Patterson function, and a single RDF can arise from many different arrangements of atoms, provided only that atomic positions within the structure have the “right” statistical distribution.
Publisher
Cambridge University Press (CUP)