Estimates of peak areas and relative atomic amounts from wide‐scan XPS spectra

Abstract

AbstractIn an earlier investigation a polynomial method to estimate area and peak positions in wide‐scan XPS spectra was demonstrated with digital data (see N. H. Turner and A. M. Single, Surf. Interface Anal. 15, 215 (1990)). This method uses a density of points of about one per electron‐volt. It was found that in most of the cases studied the procedure gave area estimates that agreed reasonably well with determinations made from narrow‐region scans (which had a density of six to eight points per electron‐volt) with non‐linear least‐squares (NLLS) calculations. In addition, ratios between the areas of the various peaks were in reasonable accord with the NLLS determinations. However, these estimates were made with fairly intense peaks. In this study, area estimates were made by the polynomial procedure that included peaks with low intensity (i.e. surface contaminants or elements present in relatively small amounts). Also, approximations of the relative atomic amounts were made. These estimates then were compared to those from narrow‐region scan determinations that used the NLLS approach. Samples used included an insulator and conductors. In most instances the area ratios and the estimates of the relative atomic amounts from the polynomial procedure agreed to within a few per cent of values from the NLLS determinations. Poorer agreement was found in the estimates of adventitious carbon. This was due to both the comparatively small peak area and low relative intensity cross‐section for carbon. Also, the polynomial method did not work well when the adventitious carbon signal was comprised of several species. Much better agreement was obtained between the polynomial procedure and the NLLS approach with elements that have a relatively high photoelectron cross‐section.