Structural materials used in nuclear reactors are subjected to the simultaneous action of mechanical stress and irradiation. Under these combined effects, the components are irreversibly deformed by a phenomenon known as irradiation creep. To guarantee reactor performance and safety, it is important to understand, control and predict these deformations. However, although numerous mechanisms have been proposed to explain this phenomenon, their relevance remains largely an open question.
Under irradiation, atomic defects (vacancies, interstitial atoms) are created and diffuse into the material. They may agglomerate, forming defect clusters (cavities, dislocation loops), or be absorbed at different sites in the microstructure (e.g. dislocations). The anisotropic formation of point defect clusters can induce deformation. Absorption of point defects by dislocations, leading to their climb and thus facilitating glide, can also induce deformation.
In order to assess the relevance of the various mechanisms proposed, a coupled approach based on both experiments and numerical simulations is planned. This approach is based on the use of a model material: aluminum.
- From an experimental point of view, we will study the evolution of nanometric defects (dislocations, dislocation loops) under irradiation and stress. Observations will be made using a transmission electron microscope (TEM), with electrons used both to create defects and to observe defect clusters and dislocations. During the experiment, the material will be subjected to stress, and the deformation mechanisms will be observed at the nanometric scale in real time. Experiments involving in situ ion beam irradiation in a TEM will also be carried out.
- The aim of the simulations will be not only to validate certain classes of mechanisms, in close connection with experimental observations, but also to identify the most relevant mechanisms, at the atomic scale. To this end, a multi-scale simulation is envisaged. Numerical tools will be used to investigate mechanisms at the atomic scale (atomistic calculations with a machine learning potential), and to simulate defect diffusion and dislocation glide at experimental sizes and times (object kinetic Monte Carlo and dislocation dynamics).