Premium
Influence of the inelastic background on the accuracy of factor analysis of electron spectra
Author(s) -
Jansson C.,
Tougaard S.
Publication year - 1990
Publication title -
surface and interface analysis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.52
H-Index - 90
eISSN - 1096-9918
pISSN - 0142-2421
DOI - 10.1002/sia.740160134
Subject(s) - background subtraction , spectral line , physics , electron , computational physics , factor (programming language) , spectral analysis , optics , computer science , spectroscopy , nuclear physics , quantum mechanics , pixel , programming language
Abstract The role of the inelastic background in factor analysis of electron spectra is investigated. Ten sets of model spectra originating from different types of in‐depth profiles were considered. Since the background depends on the in‐depth concentration profile, it will normally vary strongly within a set of spectra. In spite of this, it is shown that the background does not give rise to any additional components in factor analysis because it does not vary as a linear spectral component. The inelastic background thus will be projected on the secondary factors in the same way as spectral noise. This and the fact that the background is a smooth function of energy explains why differentiation is an acceptable procedure for background removal when factor analysis is used to identify the relative components of spectral features. Background subtraction is found to be superior to spectral differentiation for those depth profiles where valid algorithms for background subtraction have been developed.