The EPMA, LA-ICP-MS and ICP-OES study of corrosion of structural materials for a nuclear reactor cooling circuit by molten fluoride salt treatment

Abstract

Electron probe microanalysis (EPMA), inductively coupled plasma atomic emission spectrometry (ICP-OES), and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) were applied to study the interaction of molten LiF-NaF salt mixtures with candidate structural materials (alloys) for a nuclear reactor-transmutor cooling circuit. The corrosion of test ampoules and inserted test specimens made of structural materials was brought about by the action of molten LiF-NaF coolant at 680 °C and its extent and character were examined in dependence on the time of exposure. The material corrosion changes were studied by mapping the sections of ampoule walls and inserted specimen surfaces with EPMA, whereas LA-ICP-MS was employed for linear scanning the salt/ampoule wall boundary. Corrosion-released structural material, dissolved in solidified molten salt, was analyzed by ICP-OES after the salt dissolution. The melt activity was proved to induce a surficial modification of a structural material up to the depth of 15–50 µm, which was associated with the coolant contamination. The X-ray maps by EPMA with its 1-µm lateral resolution revealed compositional changes in alloys, such as regular depletion of Cr to the depth of 10–25 µm. While the lateral resolution of LA-ICP-MS with the applied laser spot diameter of 25 µm was not exactly adequate to dimensions of the corroded regions and, consequently, yielded less information in comparison with EPMA, this technique was quite sufficient for the monitoring of the presence of alloy constituents in an adhered salt layer. It was concluded that: i) the EPMA study, involving semi-quantitative elemental mapping/content profiling and detailed spot quantitative analyses makes it possible to obtain quantitative assessment of the corrosion process; ii)LA-ICP-MS profiles can be converted from signal domain to elemental contents on a semi-quantitative level when applying signal normalization to the total sum of signals.