Abstract
Electron probe microanalysis (EPMA) is used to quantify with a high accuracy the amount of different elements present on a sample of unknown composition. EPMA is largely used to quantify the amount of actinides present in fresh and irradiated fuels, to manage waste disposal and to date rocks. However, quantitative EPMA is not always possible to achieve for these materials due to the lack of suitable reference standards for the radionuclides. To overcome this difficulty, standardless methods of analysis are employed with mean of virtual calculated standards. These calculated standards are generally obtained from empirical formulae based on experimental extrapolations or from theoretical calculations that require physical parameters which are poorly known as it is the case for the X-ray production cross section.The accurate knowledge of these cross sections is required in many applications such as in particle transport code and in Monte Carlo simulations. The computer simulations are widely used in the medical field and particularly in medical imaging and in electron beam therapy. In the field of astronomy, these data are used to perform simulations that predict the compositions of stars and galactic clouds, and the formation of planetary systems.In the present work, L- and M-shell absolute x-ray production cross sections were determined experimentally for elements lead, thorium and uranium by electron impact using ultrathin self-supporting targets with thickness varying from 0.2 to 8 nm. The measured cross sections have been compared, with the distorted-wave Born approximation (DWBA) calculated by Bote et al. and with the predictions of analytical formulae widely used in practical applications. For the conversion of inner-shell ionization cross sections into x-ray production cross sections, atomic relaxation parameters were extracted from the literature. The predictions of the DWBA calculations are in excellent agreement with our measured x-ray production cross sections. This confirms the predictive results of this model and its usefulness for the calculation of virtual standards.The DWBA calculations were used into the Monte Carlo simulation code PENELOPE to calculate the X-ray intensity produced by pure actinide standards. The X-ray intensities were calculated for elements with atomic number 89 ≤ Z ≤ 99 and for accelerating voltage ranging from the ionization threshold up to 40 kV with a step of 0.5 kV. For a practical use, the calculated intensities for the most intense L and M lines were stored in a database.The predictions of our calculated standards have been compared with the x-ray intensity of known composition actinide samples (such as U, UO2, ThO2, ThF4, PuO2…) and with the data acquired during previous measurement projects. Actinide quantifications performed by virtual standards were found to be in fair agreement with the expected results. This confirms the reliability of the developed virtual standards and demonstrates that actinide quantification by EPMA can now be possible to perform without material actinide standards and with a good level of accuracy.