Microstructural analysis and molecular dynamics modeling of shape memory alloys

Abstract

Molecular dynamics (MD) is an effective tool for studying the microstructures in shape memory alloys at the atomic level. However, microstructural analysis of the results generated by MD is typically done by examining lattice constants or monoclinic angles, and only a limited set of the crystal variants can be identified. In the current work, a numerical method that can identify the crystal variants and multiple phases is proposed. By observing the pattern of the components in the transformation matrix of the lattice in the crystal, the corresponding variant of the lattice can be determined. The method is then applied to study the martensitic transformation induced by temperature and ultrahigh shear strain, and the resulting microstructures are verified by literature. Moreover, the microstructural evolution and the volume fraction variation of the crystal variants under ultrahigh shear strain are examined. The observed multiple stages of phase transitions imply that the current method is able to identify the austenitic, orthorhombic, trigonal (R phase), monoclinic, and body-centered orthorhombic crystal systems simultaneously. The proposed method also reveals that the nanometer-sized R phase is a transition zone that maintains the compatibility between the two phases in the different crystal systems. The proposed method is rapid, accurate, and sufficiently versatile to be applicable to other crystal systems and related materials.