The Thickness Dependence of Energy Loss Spectra

Abstract

One factor limiting energy loss analysis is the effect of multiple scattering on core loss edge shapes. Multiple scattering distorts fine structures, leads to incorrect quantitative analyses and even affects analysis of extended fine structure (EXELFS). Two procedures for extracting the single scattering spectrum from a spectrum showing the effects of multiple scattering have been proposed. Johnson and Spence derive the single scattering profile by taking the logarithm of the fourier transformed spectrum. If all scattering angles are accepted by the spectrometer this is an exact procedure. Leapman and Swyt have had some success assuming that multiple scattering imposes low loss structure on the high loss part of the spectrum. It is of interest to know how stable these procedures are with thickness and whether the logarithmic deconvolution can be used in thicker specimens than the method described by Leapman and Swyt.