Determination of peak positions and areas from wide‐scan XPS spectra

Abstract

AbstractIt has been found that by taking transforms of both Gaussian and Lorentzian line shapes into polynomial representations, it is possible in many instances to determine peak position, peak intensity and the full‐width at half‐maximum (FWHM) from wide‐scan XPS spectra that have a low number of data points per electron‐volt. This approach was tested with several samples, ranging from an insulator, Al2O3, to an excellent conductor, sputter‐cleaned gold. Several data sets were constructed from narrow‐scan spectra with points selected about 1 eV apart to simulate wide‐scan spectra. Then, comparisons were made between values calculated with a non‐linear least‐squares fit that used all of the data points from a given spectrum, and the polynomial procedure with the simulated wide‐scan data. It was found that the peak positions usually agreed to within 0.1 eV between the methods. The FWHM and intensities agreed to ∼20%. The area ratios usually were within ∼10%. Comparisons between actual and simulated wide‐scan data showed similar area ratio agreement. However, there appeared to be a systematic difference in the peak position of ∼0.5 eV, which was due to the spectrometer used in this study. Thus, in many cases, estimates of the relative atomic amounts can be made with those results from the polynomial procedure that would be comparable to those obtained from non‐linear least‐squares analysis of narrow‐scan data. The effects of such factors as baseline, overlapping peaks, computational techniques and line shape on the use of the polynomial method are discussed also.