Texture and grain neighborhood effects on Ni-Ti shape memory alloy performance

This work demonstrates how the statistical pseudoelastic performance of individual grains is affected by the local grain neighborhood in polycrystalline shape memory alloys (SMAs). This is achieved using a microstructural finite element (FE) model calibrated to homogenizedTi-50.9 at% Ni SMA. The results show a three-fold variation in the grain level axial transformation strain pT in randomly textured polycrystals, and a similar to 20-30% reduction in average pT if plastically predeformed. A key outcome is a performance function to predict pT of a grain, based on the orientations of the grain and its neighbors. Two key strategies to improve polycrystalline SMA performance are identified. The first is to minimize the number of grain boundaries between high-and low-performing grains: plate and bamboo geometries achieve this. The second is to employ high-symmetry orientation relationships between these grains. The results draw on recent experimental studies of grain level performance and provide a theoretical framework to interpret future diffraction tomography studies.