Conventional quantitative wavelength-dispersive spectrometers electron probe microanalysis (WDS-EPMA) requires the use of standard samples. However, standard samples are not always available for the heavy elements, especially for the actinides due to their high radiotoxicity which involves high complexity of fabrication and prohibitive costs. Standardless quantification can then be employed with the use of virtual standards. These standards require accurate knowledge of the absolute characteristic X-ray intensities emitted from samples when they are bombarded by an electron beam. X-ray intensities can be obtained by analytical calculations or by Monte Carlo (MC) simulations with the help of fundamental physical parameters. However, these parameters are not well known in the case of heavy elements and can lead to inaccurate results. Validation of the analytical calculations or MC predictions of the X-ray intensities by experimental measurements is required for standardless quantification purpose.
Absolute Mα and Mβ X-ray intensities were recorded for elements Pt, Au, Pb, U and Th with an electron microprobe using high-resolution wavelength-dispersive X-ray spectrometers Experimental X-ray intensities emitted from thick samples were measured for primary electron energies ranging from 6 keV to 38 keV. Recorded X-ray intensities were converted into absolute X-ray yields by careful evaluation of the detector efficiency.
Absolute Mα and Mβ X-ray yields simulated by the multi-purpose MC code PENELOPE [1] were compared with the experimental X-ray yields. Good agreement was found between experimental and simulated data, allowing future use of the simulated X-ray yields in standardless quantification methods. A procedure and a software program were developed to accurately obtain virtual standard values. Standardless quantifications of Pb, Th, U, Np, Pu and Am [2] were tested on several samples of known composition. Standardless quantification results were compared to analytical quantification results, when possible, and were shown in good agreements with the expected concentration value.