Atomistic simulations allow for detailed analysis of the dynamics of dislocations. However, they are limited in time and length scales impeding the modeling of thermally activated processes. By representing the microstructure as an arrangement of discrete dislocation lines dislocation dynamics simulations achieve a substantial reduction in degrees of freedom. However, input from lower scales in the form of nucleation criteria is required.
In this talk, we present detailed comparisons between dislocation dynamics and MD simulations of nanoindentation of metals. We characterize and validate the plastic zones and nucleation rates for different initial microstructures. We use the method D2C which can be applied for computationally analyzing dislocation microstructure from different methods, e.g. molecular dynamics simulations, TEM microscopy or DD simulations.
We apply the methodology to atomistic simulations of nanoscratching in iron. By characterizing the curvature nucleation rate of dislocation loops versus scratching length we find that for large scratching lengths the nucleation of dislocations is accommodated by annihilation processes leading to an almost constant averaged curvature of the dislocations. Our data reveals a pronounced "length scale effect": With increasing scratching length the number of dislocations increases but the density of geometrically necessary dislocations remains constant resulting in decreasing shear stress.