Quantitative analysis of AuPd alloys from the shape of XPS spectra by the fuzzy rule

摘要

The quantitative analysis of solid surfaces by XPS requires the determination of spectra intensities, related to atomic concentrations, recorded for constituents of investigated systems. The most reliable method of quantitative analysis is the multiline approach. The estimation of intensities requires satellites and background subtraction. The quantitative and qualitative information from intensity or binding energy shifts may be disturbed by overlapping spectra. To avoid these problems analyses based on the spectra lineshape were developed. These analyses use methods of statistical pattern recognition possible after selecting the set of standard spectra (reference set) consisting of a set of spectra recorded for different chemical states or atomic compositions (classes). The classification quality is characterized by the misclassification rate (error rate) estimated from the reference set. Fuzzy identification referring to the fuzzy membership results in quantitative estimation of chemical states and/or atomic compositions of constituents. The application of such algorithms to different lineshapes recorded at various kinetic energies results in quantification at various information depths. In the present work the fuzzy rule for analysing lineshapes of valence band, Au 4f, Pd 3d and Pd MNN Auger spectra recorded for AuPd alloys, Au and Pd, was found to be efficient and applicable for quantitative analysis of complex systems. Copyright (C) 2004 John Wiley Sons, Ltd.