[SY-N6] Symposium N-6

[SY-N6] Atomistic Modelling of Pipe Diffusion: a Direct Comparison of MD, KMC, aKMC and DMD

Frederic Houlle¹, Jutta Rogal², ^○Erik Bitzek¹ (1.FAU Erlangen-Nuernberg, Germany, 2.ICAMS Ruhr University Bochum, Germany)

Understanding and influencing self-diffusion in complex microstructures is of fundamental importance for improving the high-temperature mechanical properties of materials. Modelling diffusion in a quantitatively predictive way requires information about the underlying atomic-scale processes and corresponding activation energies. While various atomic-scale simulation approaches are routinely employed to study diffusion in homogeneous environments, the presence of strain gradients and defects still poses significant challenges to predictive atomic-scale models. In particular, currently only few detailed studies that compare the application of different simulation methods to diffusion at extended defects exist.

Here we present our recent atomistic simulations results of vacancy-mediated diffusion in the vicinity of an edge dislocations in aluminum modelled by an EAM potential. The direct high-temperature molecular dynamics (MD), adaptive kinetic Monte Carlo (aKMC) and diffusive molecular dynamics (DMD) simulations all use the identical atomistic starting configuration. The activation energies of all nearest-neighbor vacancy jumps in this configuration were determined by the nudged elastic band (NEB) method and used to build the event catalogue for the kinetic Monte Carlo (KMC) simulations and to parametrize the DMD model. Significant differences between the direct MD simulations and the (a)KMC simulations are observed, which could be attributed to the coupling of the diffusive vacancy motion with the thermally-induced fluctuations of the dislocation.