Inelastic Background and Peak Area Determination in X-Ray Photoelectron Spectroscopy (ESCA)

Author： Proctor Andrew Hercules David M.

Publisher： Society for Applied Spectroscopy

ISSN： 0003-7028

Source： Applied Spectroscopy, Vol.38, Iss.4, 1984-07, pp. : 505-518

Disclaimer: Any content in publications that violate the sovereignty, the constitution or regulations of the PRC is not accepted or approved by CNPIEC.

Previous Menu Next

Abstract

Analysis of the Au 4f ESCA spectra of thin gold films deposited over a graphite base shows that as the separation from the main photoelectron peak, ΔE, increases the background decreases. Conversely, the effect of a contaminant overlayer on solid gold gives rise to a background which increases as ΔE increases. There is, therefore, a considerable variety of background profiles which can be encountered in real samples, which depends markedly on the element, its concentration, and its distribution in the sample. Over the region of the photoelectron line signal itself, the simple integral description which assumes that the background is proportional to the number of electrons of higher kinetic energy is reasonably valid and is suggested for peak area measurement when a nonlinear profile is needed.Direct integration and curve fitting techniques constitute the two main methods of peak area measurement. Simulated Ni 2p spectra are used to compare the effectiveness of the simple integral and linear background profiles for producing the correct peak areas. Although the simple integral profile shows more consistency and is to be preferred in general, certain fortuitous cancellation effects, particularly with nonmonochromatized spectra, make areas using the linear background fairly accurate in some instances. Curve fitting can overcome many of the problems of direct integration resulting from severe peak overlap. Two types of Gaussian/Lorentzian functions with a suitable exponential tailing function, which has a finite area over reasonable integration limits, are discussed with regard to their use in non-linear least-squares curve fitting methods. Also discussed is the usefulness of composite and difference spectra as simpler methods of determining the relative areas of overlapping peaks.