A Comparison of the Peak-to-Background Method and an Empirical Correction of the Results in the Energy-Dispersive Electron Probe Quantitative Analysis of Powder Materials

Abstract

Several algorithms for correcting the results of the quantitative electron probe elemental analysis of samples with rough surfaces and powder materials are compared. The effectiveness of the methods was estimated by the deviations of corrected weight fractions of elements from the results obtained for reference samples, i.e., polished plates of test materials. Using the most commonly used peak-to-background method, the intensity of continuous X-ray radiation was calculated by several methods. One method involved the analytical calculation of the bremsstrahlung generation function and the correction of the width and shape of the bremsstrahlung spectrum under the diagnostic lines of the elements based on the experimental spectra. The second, more rapid method was based on the direct simulation of the continuous radiation background by Monte Carlo methods in the NIST DTSA-II software environment. The last method of calculating the background of continuous radiation yielded smaller deviations of the results of quantitative analysis from the results obtained for reference samples. An empirical adjustment method was also tested. It was based on experimentally revealed patterns in the energy-dispersive spectra of powder samples. An analysis of the experimental data revealed an empirical dependence relating the parameters of characteristic photons to the value of accelerating voltage required to obtain the proper ratio of the weight concentrations of elements in the analysis of powdered materials. The proposed empirical method for correcting the results of analyses of powder samples based on the totality of the measurements performed is more effective.