Reconstruction of tritium amount and depth profile in PFMs for fusion reactor using MLEM and BIXS spectra

Abstract The maximum likelihood expectation maximization (MLEM) algorithm has been introduced as a robust solution to address the challenges associated with the high condition number of response matrix encountered in the beta ray induced x-ray spectra (BIXS) technique, specifically for the quantitative reconstruction of tritium amount and depth profiles within plasma facing materials (PFMs) in a fusion reactor. Results suggest that the MLEM algorithm demonstrates rapid convergence properties in the reconstruction process. The reconstruction error for tritium amount in tungsten with typical tritium distributions can be as low as 0.1% after 10 5 -iterations, and the maximum reconstruction error observed in the depth distribution for a single layer consistently remains below 10%. While the reconstruction error for each layer after including the tritium activity as weighted factor is consistently less than 0.5%. Consequently, utilizing MLEM and BIXS techniques, the quantitative determination of tritium within PFMs can be achieved non-destructively, and it is also independent of prior knowledge regarding the tritium amount within the materials. This approach holds significant promise for the assessment of tritium retention, diffusion behavior and the development of tritium removal strategies.