Determination of the activation energy by stochastic analyses of molecular dynamics simulations of dislocation processes

Abstract

An investigation is reported of the probability and the probability density of thermal activation of stress-driven dislocation processes, as simulated using molecular dynamics (MD). Stochastic analyses of the survival probability are found to lead to simple relationships between the loading history and the distribution of the interaction time and strength. It is shown that the determination of the activation energy associated to a thermally activated event can be achieved by a reduction of the stochastic process to a process obeying the Poisson's distribution, preserving the activation probability at the survival time. The method is applied to the kink-pair mechanism for screw dislocations in iron. Predictions are compared with experimental results and with other methods reported in the literature, which allows the difference in the approximations and in the assumptions considered in these models to be underlined.