Evaluation of correction accuracy of several schemes for AES matrix effect corrections

Abstract A matrix effect correction is required to improve the accuracy of quantitative AES analysis. The correction includes terms involving the atomic density ( n ), electron back‐scattering factor ( R ) and electron escape depth ( L ). Many schemes have been proposed by various people for corrections of the R and L terms. However, up to now, there have been no systematic investigations of the correction accuracy of the proposed schemes. We have evaluated the correction accuracy, based on measured intensity data for AuCu alloys of different compositions. Comparison was made between the observed intensity ratio K (= I unk / I std ) and the calculated intensity, ratio K ′ (= C ( n unk / n std )( R unk / R std )( L unk / L std )), where C and I represent the concentration and intensity, respectively. The superscripts ‘unk’ and ‘std’ denote that the parameters are for unknown and standard specimens, here the pure elements. If the correction works well, the error Er (= K ′  K )/( K ) will become smaller. Evaluations were carried out on three schemes for the R correction and on seven schemes for the L correction using the Au 239 eV, Au 2024 eV and Cu 920 eV transitions. The root mean square (RMS) of the calculated errors showed several per cent for the best case and 20–30% for the worst case. The RMS error varied a few per cent between schemes for the R correction but it varied ∼30% for the L correction.