The separation of surface and bulk contributions in ELNES spectra

A “Thickness Series”: Weak Signal Extraction of ELNES in EELS Spectra From Surfaces

“…Acquisition of EEL spectra in transmission mode through the sample over different terraces provides a set of measurements with constant contribution of surface signals but systematically varying bulk component. It should be noted that the whole data set should be restricted to within the single scattering regime in order to ensure a linear integration of all components over thickness without further data processing and deconvolution (Schattschneider et al, 2002). The thickness in the thickest area (point P7) is 60 nm in the present data set, which is considerably smaller than the total inelastic mean free path (>100 nm).…”

“…The approach of separating ELNES into two components in a TEM foil, in this case the bulk and the oxidized surface of a thin film, having comparable thickness, has been reported (Stöger et al, 2002) from two scaled ELNES spectra containing different component ratios. Although effective for relatively thick layers, the simple subtraction of two-scaled spectra introduces a large propagation error of ~15%, resulting from the uncertainty in estimated contributions of both layers, noise in the spectra, and uncertainty in quantification as directly stated in Schattschneider et al (2002). Therefore, an improved method needs to be developed to extract weak signals with higher precision.…”