The XP spectrum consists of peaks on a characteristic stepped background. The simplistic interpretation of this, used in most data systems, assumes that the step derives entirely from extrinsic losses and can be removed by a "Shirley" type of algorithm. Tougaard has demonstrated the weakness of this assumption, showing how removal of the background by the simple Shirley method will subtract, also, important components of the peak intensity itself. The reason for this is that both peak asymmetries and other intrinsic losses contribute intensity in the same form as that resulting from extrinsic losses. The Tougaard background was developed to account for extrinsic losses alone and has been shown to merge with the general loss background at a point some 50 eV distant from the peak position. The use of the Tougaard universal background for the transition elements removes very little intensity from the vicinity of the main peak and this poses a problem for quantitative analysis since use of a window wide enough to yield a peak intensity which is independent of the integration limits adds in a great deal of intensity, which has probably not been included in the standards used for quantification, whereas the reduction of the range to that of a conventional narrow window leaves, after subtraction of the Tougaard background, an asymmetric peak still retaining a step-like background and an integrated area which is not closed but depends on the choice of window size. In this work we show that the asymmetry associated with intrinsic, final state, losses in the photoelectron emission can be defined by a single shape parameter. This parameter, kappa, is shown to be independent of the kinetic energy of the peak, is similar for different peaks of a given element and does not depends on the spectrometer type or operating parameters. It varies uniformly with atomic number in a manner which is similar to that of the melting point of the transition elements, i.e. reflects the internal energy in those cases where this is defined by the density of overlapping valence levels. The use of this parameter in curve fitting provides a means of characterising the intrinsic component of the losses. We have handled this by a Shirley type routine after subtraction of the Tougaard background. The advantage in quantification is that a closed area under the peak is obtained, i.e. the advantage of the Shirley background is retained but in a defined consistent manner. (C) 1998 Elsevier Science B.V. All rights reserved.
