Predicting the martensitic transition temperatures in ternary shape memory alloys Ni0.5Ti[formula omitted]Hf[formula omitted] from first principles

Abstract

Martensitic transition temperatures (MTTs) can be tuned by alloying binaries with other elements to create multi-component shape memory alloys (SMAs). However, it is inefficient to use the trial-and-error approach to find compositions with desirable operating temperatures because of the large number of combinations of metals possible to form ternaries, quaternaries, etc. Thus it is crucial to develop the theoretical capability of accurately predicting MTTs as a function of composition, in order to provide experimentalists with necessary and reliable guidance. Previous work has focused on developing and applying first-principles methods to compute phase transitions and MTTs in binary SMAs such as NiTi (nitinol), but certain technical problems associated with the multi-component SMAs remain unsolved. In this work, we employed ab initio molecular dynamics (MD) and thermodynamics integration to study the NiTiHf-based high-temperature ternary SMAs. We overcome the technical challenges to accurately obtain the Gibbs free energy in cubic ternaries where the reference structures are unknown. Specifically, we examined the cubic, monoclinic and orthorhombic structures of Ni0.5Ti0.5−xHfx for x∈[0,0.5], and our results suggest that the cubic-to-monoclinic martensitic transition occurs when x<0.08, for x>0.17 the martensitic transition is between the cubic and orthorhombic phases, whereas in between our calculations cannot distinguish these two martensite structures near the MTT. The computed MTTs vs Hf content x are in good agreement with measured data. Thus our current work paves the way for computational design of multi-component SMAs with desired properties.