Specific features of matrix correction in the X-ray fluorescence analysis of samples of widely varied composition

The authors critically analyze the paper “Corrections for matrix effects in X-ray fluorescence analysis” [1], which was declared by R. Rousseau as a tutorial but contains some inaccuracies. Conclusions are confirmed by the results of experiments and theoretical calculations. It is shown that uniform linear calibration function I _i /I _i ^m = f(R _i ^th ) can hardly be obtained for test samples of any composition because of the difference in their microstructure and the variation of the background. At the same time, in the XRF analysis of multicomponent materials, such as steels, the Lucas-Tooth and Jongh equations provide a good alternative to the Rousseau algorithm.