Quantitative elemental concentrations by energy-filtered imaging

Abstract

There is widespread interest in elemental distribution maps produced from energy-filtered core-loss images obtained wim commercial imaging energy-filters and slow-scan charge-coupled device (CCD) cameras on transmission electron microscopes (TEMs). Earlier work on the feasibility of mapping solute segregation in stainless steels by energy-filtered imaging confirmed the utility of jump-ratio images (created by division of a post-edge image by a pre-edge one) for rapid assessments of elemental distributions. The effects of diffraction contrast and thickness variations are largely corrected for in such images. However, quantitative compositional information requires the use of net core-loss intensities following subtraction of an extrapolated background. Such core-loss intensities are influenced by diffraction contrast and thickness variations; corrections for these effects may be necessary for a quantitative interpretation. In the present work, energy-filtered images are treated similarly to quantitative electron energy-loss spectrometry (EELS) data. An image showing number of atoms per unit area, nx, is obtained by dividing the core-loss intensity image, Sx(Δ,β), by the low-loss image, J1(Δ,β), obtained with identical energy window Δ and collection half-angle β, and by the partial ionization cross-section, σx(Δ,β). Further normalization by specimen thickness, t, yields an image showing elemental concentration in atoms per unit volume: