Crossover from sluggish to accelerated diffusion in multi-principal element alloys induced by lattice distortion

Sluggish diffusion and lattice distortion are core factors in multi-principal element alloys (MPEAs) that are responsible for their unique properties. Nonetheless, the existence and mechanisms of sluggish diffusion remain controversial. In this study, we reveal the intricate interplay between diffusion and lattice distortion using equilibrium Monte Carlo (MC) and kinetic Monte Carlo (kMC) simulations driven by a Graph Neural Network (GNN) model that captures local atomic environment-dependent energy states and path-dependent transition rates. Our results demonstrate that lattice distortion significantly elevates equilibrium vacancy concentrations in MPEAs. Integrating insights from vacancy concentration and kMC-derived diffusivity, we identify a crossover in the self-diffusion coefficient of MPEAs as a function of the degree of lattice distortion. Sluggish diffusion is observed in systems with moderate distortion, where increased vacancy concentration, accelerated vacancy jump frequency, and stronger correlation effects compete. Unexpectedly, a higher degree of lattice distortion leads to accelerated instead of sluggish diffusion due to the weak correlation effect. Our findings establish the critical connection between structural features and kinetic properties in MPEAs, offering a foundation for designing high-performance alloys with tailored kinetics by strategically tuning structural characteristics within the vast compositional space.