Consistent, combined quantitative Auger electron spectroscopy and x-ray photoelectron spectroscopy digital databases: Convergence of theory and experiment

Auger Electron Spectroscopy (AES) and x-ray photoelectron spectroscopy (XPS ) have many aspects in common. Therefore, tests of aspects for one spectros copy, applicable to the other, should be validated for both. Digital databa ses of elemental spectra for both AES and XPS have thus been measured using an electron spectrometer that has been fully calibrated for its intensity and energy axes. These databases of true spectra give absolute Auger electr on yields and relative photoelectron yields. The AES data are measured for both 5 and 10 keV electron beam energies, whereas the XPS data are measured for both Al and Mg unmonochromated x rays at the magic angle. In addition to these we have measured reflected electron energy loss spectroscopy (REEL S) data to provide a third database. The combination of these databases all ows a refinement of the theories to obtain an overall convergence between t heory and experiment. Improvements to the theory have been obtained by iden tifying three classes of parameter, those (i) for both AES and XPS, such as electron transport, the methodology of evaluation of peak areas, and the s pectrometer response function, (ii) for AES only, such as the electron ioni zation cross section, backscattering, and specific electron backgrounds, an d (iii) for XPS only, such as the photon-ionization cross section. Addition ally, the use of REELS data for background subtraction seems significantly better than the use of any one single Tougaard universal cross section in t he determination of the AES and XPS intensities in the databases. The forma lism for quantitative analysis in AES and XPS, using relative sensitivity f actors, has been revised to develop an accurate matrix-less formalism that is very simple for use by the analyst. This formalism, involving "average m atrix sensitivity factors" rather than "pure element sensitivity factors," has the same accuracy as the matrix formalism but its simplicity permits re ady extension to systems beyond binary. [S0734-2101(00)08704-2].