Multiscale modelling of radiation-enhanced diffusion phenomena in metals

Abstract: A multiscale modelling framework and an experiment campaign are used to study void swelling and Cu precipitation under irradiation. Several aspects regarding defect and solute diffusion under irradiation have been studied in this thesis.First, a self-diffusion model in bcc Fe has been constructed in order to describe the non-linear effects, especially the magnetic transition, around the Curie temperature. First principles calculations are applied to obtain the parameters in the model. The paramagnetic state is simulated by statistical sampling of randomly arranged spin states on each atom. The model fits well with the experimental observations.Then, a combination of atomistic calculations and the finite element method (FEM) is developed in order to solve the diffusion equations of point defects, which are under the influence of a dislocation strain field. The dislocation bias, a key parameter in void swelling models, is hence obtained numerically. The method has been applied in different structural lattices. In the bcc materials, anomalous bias factors have been found for both edge- and screw dislocations. For the edge dislocations, the traditional assumption that the dislocation bias value is proportional to the Burgers vector has been proven not appropriate. For the screw dislocation, a negative bias value is obtained. This implies that vacancies, instead of self-interstitials, are preferentially absorbed into the screw dislocations. Thus a possible complementary mechanism is here introduced for explaining the long swelling incubation time before the steady swelling in bcc materials compared to that in fcc materials.Edge dislocations in fcc materials split into partial dislocations due to their  relatively low stacking fault energy. This feature complicates the analytical derivation of the dislocation bias. However, by transforming the analytical dislocation-point defect interaction energies to discrete interaction maps numerically applied in the FEM method, it is possible to perform a systematic study on typical fcc materials, i.e. Cu, Ni and Al. The impacts on the dislocation bias from elastic constants and stacking fault energy have been studied. It is found that the partial splitting distance is the dominating factor that determines the dislocation bias. A prediction method has been hence developed to obtain the dislocation bias of the austenitic alloys, for which it is difficult to use an atomistic description of the interaction maps. A prediction of about 8% dislocation bias of a typical austenitic 316 alloy has been made without performing specific atomistic calculations in the austenitic alloys.Finally, Cu precipitation under irradiation has been studied using both experiment and simulations. Cast iron and FeCu alloy samples were irradiated for a week with 2 MeV electrons. The resistivity of the samples was measured in situ. The microstructure of the samples was then examined by atom probe tomography. No Cu precipitation was found in the cast iron sample while small Cu clusters are observed in the FeCu model alloy. To simulate the clustering process, Kinetic Monte Carlo (KMC) and rate theory methods are used. Both the KMC and rate theory simulations show clearly the Cu clustering process in the FeCu alloy but not in cast iron within the irradiation dose.