Speaker
Description
The martensitic phase of NiTi has long posed a challenge for first-principles modelling due to the discrepancy between the experimentally observed monoclinic B19′ structure and the B33 structure predicted by density functional theory as the low-temperature ground state. Recent molecular-dynamics simulations have shown that vibrational entropy can stabilize the B19′-like phase at finite temperature, providing a possible explanation for this apparent discrepancy. However, the predicted stabilization occurs at temperatures substantially higher than those observed experimentally.
In this work, we investigate the role of antisite defects in the finite-temperature stability of NiTi martensite using large-scale molecular-dynamics simulations driven by a newly developed machine-learning interatomic potential. We show that antisite disorder markedly shifts the stabilization of the B19′-like structure towards lower temperatures, bringing the predicted structural behaviour significantly closer to the experimentally relevant temperature range.