Modeling pseudo-elasticity in NiTi

Abstract

A comparison of the EAM-Finnis-Sinclair and the MEAM potential, two of the recently developed potentials to model NiTi, is carried out. The potentials are compared by studying the pseudo-elastic behavior in bulk NiTi for one specific crystallographic orientation. To this end we perform, for the first time, simulations where the transformation occurs not only under compressive but also under tensile loading along 〈100〉_B2 using both potentials. Results indicate that in both cases the MEAM potential captures the pseudo-elastic behavior more accurately. By using a lattice deformation model, it is demonstrated that the inaccurate transformation strains predicted by the EAM-Finnis-Sinclair potential are a direct consequence of its inability to predict experimental values of the lattice constants. Similarly, it is shown that the more precise values of the Young's modulus of the initial austenitic and the final martensitic phase estimated by the MEAM potential are the result of its ability to predict elastic constants more accurately than the EAM-Finnis-Sinclair potential. As a result, it is concluded that the MEAM potential is better suited to study the overall pseudo-elastic behavior in NiTi.