Ceramics-Silikáty 43, (2) 41 - 47 (1999)

Modeling of physical phenomena taking place in microanalysis can show limitations as well as possibilities of the method. Experimental microanalytical data, particularly k-ratios, have been compared with calculated ones obtained by mathematical simulation. The model uses the single scattering Monte Carlo algorithm, where "Bethe cross section" is introduced for the description of inelastic events. Various ionization cross sections have been tested to show their influence on the distribution functions. Accuracy of the simulation has been tested at the boride and carbide data sets, comparing calculated k-ratios with experimental data. They have revealed the uncertainty in the mass absorption coefficients and their limitation to the quantitative analysis of very light elements. Layered samples present the most crucial test for the model. Monte Carlo calculation results are compared with experimental data and commercial analytical programs. The paper has shown that the correction program based on first principles gives comparable accuracy with analytical approaches for all cases. The progress in the quantitative analysis is conditioned by the progress in the theory as well as by the progress in the accurate experimental determination of some physical parameters.