Nickel aluminum shape memory alloys via molecular dynamics

Shape memory materials are an important class of active materials with a wide range of applications in the aerospace, biomedical, and automobile industries. These materials exhibit the two unique properties of shape memory and superelasticity. Shape memory is the ability to recover its original shape by applying heat after undergoing large deformations. Superelasticity is the ability to undergo large, reversible deformations (up to 10%) that revert back when the load is removed. These special properties originate from a reversible, diffusionless solid-solid phase transformation that occurs between a high temperature austenite phase and a low temperature martensite phase. The development of the martensite microstructure is not well understood; this is especially true in regards to the role of size and mechanical constraints that dominate the properties in nanoscale samples. The goals of this research are to use molecular dynamics (MD) to (1) study the effects of simulation size on the martensite transformation to determine the ultimate limit of miniaturization, (2) to investigate the effects of mechanical constraints on the martensite transformation and resulting microstructure, and (3) to explore the effects of grain size in polycrystalline shape memory alloys. MD is well suited to study the transformation, as it shares a similar time scale with the extremely fast, diffusionless transformation. An extensive set of cooling and heating simulations were performed on Ni63Al37 disordered shape memory alloys (SMAs) to determine the effect of system size on the transformation. Simulation cell sizes in the range of 4.2 to 20 nm were studied. We discovered that decreasing system size only resulted in a slight increase of both transformation temperatures. However, the variability of the austenite transformation temperature increased considerably with decreasing simulation cell size, reaching 10% of the mean value for a system size of 10 nm. This variability can impose a fundamental limit on the miniaturization of this class of materials, as the reliability of device performance comes into question. Also, mechanical constraints were applied to force the cell angles to remain 90° in order to emulate the environment of a partially transformed polycrystal where grains are constricted by their neighbors. The mechanical constraints caused the austenite transformation temperature to decrease with decreasing size by up to 50%, and resulted in a two-domain microstructure for system sizes above 4.2 nm in order to accommodate the internal stresses. Finally, large scale MD simulations were done on polycrystalline samples with grain sizes ranging from 2.5 to 20 nm. We found that a critical grain size of 7.5 nm resulted in a minimum in the percent transformation to martensite. Below this critical size, martensite forms at the grain boundaries and the grains are able to rotate via grain boundary sliding to relieve internal stresses. In larger grains, martensite can nucleate and grow within the grains more easily. A uniaxial strain of up to 10% was applied to investigate the stress induced martensite transformation. Larger grains showed considerable work hardening when strained beyond about 2%. Plastic recovery was also calculated by unloading and relaxing at 4 and 10% strain. Samples strained to 10% were generally able to recover about 20-30% of the plastic strain, while samples strained to 4% showed varying amounts of recovery that peaked at 66% for a grain size of 7.5 nm