Multiscale computational modeling of defects in uranium dioxide
Abstract
Manufacturing, extreme operation conditions, and storage introduce large variety of defects into uranium dioxide (UO2) nuclear fuel, which have diverse effects on fuel properties. In this study, a multiscale computational approach to model defect behavior in UO2 is presented, by passing information in stages from electronic structure and atomistic calculations to a stochastic kinetic method. Understanding the interaction of fission products with dislocations is critical to interpret their interaction with fuel microstructure. Atomistic simulations are employed to predict the segregation of ruthenium and cesium, both fission products, to edge and screw dislocations. Dislocation loops of different shapes and sizes are simulated to investigate their atomic structure. To understand the segregation behavior, comparisons are made between atomistic simulations and continuum-elastic based results. Segregation behavior is found to be directly related to the elastic strain field around the dislocation core and is affected by the orientation of dislocation and electrostatic interactions at the atomic defect site. A detailed mechanism of, and the effect of homogeneous strains on, the migration of uranium vacancies in UO2 is presented. Migration pathways and barriers are identified using density functional theory and the effect of strain fields are accounted for using a dipole tensor approach. This information is then passed to the kinetic Monte Carlo simulations, to compute the diffusivities in the presence of external strain fields. We report complex migration pathways for uranium vacancy and show under homogeneous strain, only the dipole tensor of the saddle with respect to the minimum is required to correctly predict the change in energy barrier between the strained and the unstrained case. Homogeneous strains as small as 2% have considerable effect on diffusivity of both single and di-vacancies, with the effect of strain more pronounced for single vacancies than di-vacancies. Further, strain lead to anisotropies in the mobility of vacancy and degree of anisotropy is sensitive to nature of applied strain field. Our results suggest that the influence of strain on vacancy diffusivity is greater when single vacancies dominate the defect structure, such as sintering, while the effect will be much less substantial under irradiation conditions where di-vacancies dominate. An early uncertainty quantification study is presented. The source as well as the nature of errors in the input parameters is investigated, and the sensitivity of each of these input parameters on the final computed diffusion rates is analyzed.
- Publication:
-
Ph.D. Thesis
- Pub Date:
- 2015
- Bibcode:
- 2015PhDT.......219G
- Keywords:
-
- Materials science;Nuclear engineering;Energy;Condensed matter physics