In our earlier work we have demonstrated the feasibility of performing multiscale simulations using a coupling method appropriate for metals, which combines a small region treated by quantum mechanics within DFT and a much larger region treated classically by EAM potentials. We have also demonstrated that a careful treatment of residual stress in the classical region needs to be made in order to ensure that there are no spurious effects in the multiscale calculation. With this theoretical basis, we were in a position to apply the method to the study of Ni embrittlement by S impurities at the grain boundaries, using a fully dynamical approach. This enabled us to address finite temperature effects and processes that may play a crucial role in the overall behavior of the system, such as amorphization and impurity segregation at the grain boundaries. Detailed studies of this nature constituted the next phase of the project, and provided new insight into the mechanisms of stress-corrosion cracking. These studies were a necessary complement to the simulations based on empirical potentials, such as the recently published work from the collaboration with the USC, and can serve as a validation of those results from a more fundamental perspective.
PDF
download
878987797
028-85220240
