Computational Approaches for Studying the Nucleation of Voids at the Nanoscale.

Solid state precipitation and growth of intermetallic phases from metal solid solutions is important for the design and manufacturing of alloys. Intermetallic precipitates, particularly on the nanoscale, are one of the most effective methods to strengthen metals. It has been demonstrated that in Mg alloys this microstructure can result from the defects generated by mechanical deformation. Vacancies and voids are commonly found among these defects and may play a crucial role in the nucleation process, as they change the composition in their vicinity by attracting or repelling solute.



The formation of voids is itself a nucleation process. Current models assume that voids nucleate from a uniform concentration of vacancies. However, in direct molecular dynamics simulations we have observed evidence that agglomerations of vacancies may constitute a meta-stable intermediate state not considered in existing models for the nucleation of voids.



Nucleation is a “rare event,” making extracting kinetic information from simulations challenging. By means of Replica Exchange Transition Interface Sampling (RETIS) we quantify the void nucleation process by sampling independent phase-space paths. Using this data, we aim to construct non-classical void nucleation theories suitable for far-from-equilibrium metals processing.