Simulation du piégeage d’une molécule d’iode I2 dans un défaut de Schottky, constitué d’une lacune d’uranium et deux lacunes oxygène, dans le dioxyde d’uranium UO2.
The evolution of nuclear fuel behavior in reactors is governed by the formation of irradiation damage and the evolution of the material's microstructure and properties under irradiation. Fuels, based on uranium and plutonium oxides, also see their chemical composition evolve during irradiation, with the creation of fission products and the modification of oxygen stoichiometry.
.
The electronic structure calculation, based on quantum mechanics and Density Functional Theory (DFT), is a primordial brick in the multiscale modeling scheme of nuclear fuel irradiation behavior.
It has enabled us to simulate properties such as the trapping of volatile fission products (I, Cs, Te) in the crystal structure of UO2 and (U,Pu)O2, the variation of the heat capacity of (U,Pu)O2 as a function of temperature and Pu content by ab initio molecular dynamics.
The determination of the electronic structure of actinide oxides is, however
complex due to the characteristics of actinide 5f electrons, which are highly localized around the nuclei and require recourse to approxima- tions beyond the computationally expensive standard DFT approximations (DFT+U).
A recent challenge is to develop interatomic potentials based on machine learning (Machine Lear- ning) from DFT calculations. These potentials enable molecular dynamics and finite-temperature property calculations to be carried out with reduced computation times compared with the ab initio approach, while maintaining the accuracy of the results. The first results obtained for UO2 in 2022 are very encouraging. The constitution of a DFT database is a prerequisite and requires numerous calculations.
The computing time resources required for these studies make the use of national computing centers indispensable.