On the impact of lattice parameter accuracy of atomistic simulations on the microstructure of Ni-Ti shape memory alloys

Ni-Ti is a key shape-memory alloy (SMA) system for applications, being cheap and having good mechanical properties. Recently, atomistic simulations of Ni-Ti SMAs have been used with the purpose of revealing the nano-scale mechanisms that control superelasticity and the shape-memory effect (SME), which is crucial to guide alloying or processing strategies to improve materials performance. These atomistic simulations are based on molecular dynamics (MD) modelling that relies on (empirical) interatomic potentials (IAPs). These simulations must reproduce accurately the mechanism of martensitic transformation and the microstructure that it originates, since this controls both superelasticity and the SME. As demonstrated by the energy minimization theory of martensitic transformations (Ball and James (1987 Arch. Ration. Mech. Anal. 100 13-52)), the microstructure of martensite depends on the lattice parameters of the austenite and the martensite phases. Here, we compute the bounds of possible microstructural variations based on the experimental variations/uncertainties in the lattice parameter measurements. We show that both density functional theory and MD lattice parameters are typically outside the experimental range, and that seemingly small deviations from this range induce large deviations from the experimental bounds of the microstructural predictions, with notable cases where unphysical microstructures are predicted to form. Therefore, our work points to a strategy for benchmarking and selecting IAPs for atomistic modelling of SMAs, which is crucial to modelling the development of martensitic microstructures and their impact on the SME.

Automated summary

Recent research has found that atomistic simulations used in studying shape-memory alloys often fail to accurately reproduce the mechanism and microstructure of martensitic transformation, leading to significant deviations in predicting the shape-memory effect. Accurate selection and assessment of interatomic potentials is crucial for atomistic modeling of shape-memory alloys.