Reexamination of a U-Zr diffusion couple experiment using quantitative phase-field modeling and sensitivity analysis

Interest in U-Zr metallic nuclear fuels has been on the rise, but sparsity in the available diffusion data continues to hinder efforts to model irradiation behaviors like constituent redistribution. In the current work, we develop a quantitative phase-field model for the influential β and γ phases of U-Zr and use it to reexamine data from a published diffusion couple experiment. We optimize the two phases’ kinetic parameters and use annealing simulations to show that the optimized model produces more accurate predictions than those obtained using the diffusion parameters currently employed by constituent redistribution models. We then demonstrate how to minimize the impact of user-defined model parameters such that the parameters based on experimental data dominate the model's response. Sensitivity analysis studies confirmed that the phases’ kinetic parameters are more influential than the interface energy for this type of problem, and γ phase kinetics were found to be more impactful than β phase kinetics. This observation is believed to be related to differences in phases’ solubility ranges and diffusion-limiting behaviors. These findings and the interpretive modeling technique employed in the current work will increase the efficiency of the data collection efforts necessary to reduce uncertainties in U-Zr diffusion parameters, expediting further fuel development.