Capability and Uncertainty of Standardless Procedures for Quantitative Electron Probe X-ray Microanalysis

Abstract

Electron probe microanalysis (EPMA) has become a well established technique for determining compositions of bulk and multilayer samples. Quantification is traditionally based on the use of standards and many uncertainties, resulting of the experimental conditions, the atomic data and the physical model describing the X-ray emission, can be canceled. Depending of the instrument, the operator experience and the nature of the sample, the uncertainty in the results by the use of standards can be expected to be less than 2%, even when mass absorption coefficient is very large. Unfortunately, measurements on many standards are time consuming and require the appropriate standard for the corresponding element. As a solution, standardless procedure [1-5] is an elegant way to overcome these obstacles. This method attempts to remove the need of standardization through calculation of pure element intensities. Two standardless procedures are largely used. The first one, based on absolute calculations, takes into account all aspects of X-ray generation, propagation, and detection, and the second one, based on mathematical fits, makes use of measured intensities from a limited set of pure standards. Since standardless procedures imply a decrease in accuracy, the aim of this paper is to evaluate the capability and the uncertainty of standardless analysis based on absolute quantification procedure, putting the emphasis on fundamental aspects.

Standardless quantification methods in electron probe microanalysis

“…The relative uncertainties are about 5% for major and minor elements, i.e., with concentrations greater than 10% and between 1 and 10%, respectively, and somewhat greater for trace elements (concentrations lower than 1%) [6]. Mineral samples constitute a special case for which, provided the adequate standards are available, the relative errors of the elemental concentrations are lower than 2% in most of the typical situations [7].…”

Standardless quantification by parameter optimization in electron probe microanalysis

“…When a wavelength dispersive spectrometer (WDS) is used, the relative uncertainties are about 5% for major and minor elements, i.e., with concentrations greater than 10% and between 1 and 10%, respectively, and somewhat greater for trace elements (concentrations lower than 1%) [2]. In the special case of mineral samples, the relative errors of the elemental concentrations become lower than 2% in most of the situations [3].…”