| Abstract Scope |
Zirconium hydride (ZrH2-x) is a promising moderator for microreactors due to its high hydrogen content. However, concerns remain over its thermal stability and potential hydrogen release at high temperatures. In this talk, we employ atomistic simulations to investigate how microstructures, especially grain boundaries and triple junctions, influence hydrogen diffusivity and retention in ZrH2-x. A machine learning interatomic potential, developed via an active learning framework, is used to model hydrogen behavior across different microstructural environments in ZrH2-x. Two stoichiometries, ZrH2 and ZrH1.75, are examined. Our results indicate that grain boundaries enhance hydrogen diffusivity in ZrH2, while have negligible effect in ZrH1.75, a trend attributed to local variations in the H/Zr ratio, which affect the diffusion mechanisms of hydrogen. Elevated temperatures reduce diffusivity difference across grain boundary types. Our calculations of the hydrogen out-diffusion rate and sink strength further reveal that grain boundaries do not act as effective hydrogen sinks. |