The question of the long term alteration of nuclear glasses, which are matrixes currently used to store the long-lived high level nuclear wastes, is a key point in the perspective of a deep geological storage as planned in France. To cope with this question, experiments and simulations at different scales are combined to describe the behaviors from the real scale (the storage site) to the finest one (the atom).

In this context, this thesis aims to extend the development of a new Monte Carlo method used to simulate glass alteration at the atomic scale to complexify both the glass and the solution. This method is currently limited to SiO2-B2O3-Al2O3-Na2O glass compositions. Hence new elements with a significant impact on the glass alteration will be considered like Zr, Fe, Ca and Mg.

The work will be aimed at modifying the code as a first step to introduce these new elements then at determining the Monte Carlo parameters able to reproduce the experimental data measured in parallel. These results will be validated by complementary molecular calculations.

A consequent work will be also performed to connect this new Monte Carlo method to those currently developed at other scales.

The candidate will be based at CEA Marcoule in the « Laboratoire d’études du Comportement à long terme des Matériaux de Confinement » (DES/ISEC/DE2D/SEVT/LCLT) and will work in partnership with the « Laboratoire de Modélisation, de Thermodynamique et de Thermochimie » (DES/ISAS/DPC/SCCME/LM2T) at CEA Saclay. The candidate should have skills in computer science and chemistry of materials. At the end of this thesis, the experience gained will create numerous opportunities for the candidate in the field of physico – chemical modelisation.