MD study of plastic deformation by dislocations and phase transformations
Shape memory alloys (SMA) like Ni-Ti show a characteristic phase transformation between a martensite and an austenite phase depending on temperature and mechanical stress. The change in shape during transformation is exploited in various applications. After repeated use, this shape memory effect may suffer from functional fatigue, which is caused by changes in the microstructure. It is suggested that the dislocation density increases during repeated deformation and that dislocations inhibit the phase transformation. This project's aim is to study the interaction of dislocations and phase transformation on a atomistic level using classical molecular dynamics simulations.
Simulations of nanoindentation are performed with a Lennard-Jones style potential of Nickel-Titanium. Although such a simple potential cannot be expected to model all properties correctly, it exhibits a phase transformation between a cubic austenite phase and monoclinic martensite, induced either by stress or temperature. The analysis of dislocations and phase transformation requires suitable methods to identify higher-level structures from atomistic data, such as dislocations and the phase boundary. The methods developed are able to identify dislocations and their Burgers vectors in both austenite and martensite phase. Phase boundaries are detected as well and approximated by polyhedron meshes. This approach offers both qualitative results on how the entities interact, as well as quantitative values concerning dislocation densities and the volume fraction of each phase.
First results show a strong anisotropic behaviour. Depending on the crystal orientation the material responds either by plastic flow or almost purely pseudo-elastic. Regardless of the orientation, dislocations are found to stabilize regions of residual martensite as shown in the figure above.