The classical modelling of dislocation climb based upon a continuous description of vacancy diffusion was compared to recent atomistic simulations of dislocation climb in body-centered cubic iron under vacancy supersaturation. Quantitative agreement was obtained, showing the ability of the classical approach to describe dislocation climb. The analytical model was then used to extrapolate dislocation climb velocities to lower dislocation densities, in the range corresponding to experiment. This permitted testing of the validity of the pure climb creep model proposed by Kabir et al. (2010). The stress entered into the creep model of Kabir et al. only through control of the dislocation density and the vacancy supersaturation. As the climb velocity did not then depended upon the dislocation orientation all dislocations, whatever their orientation, were climbing at the same velocity. If there was no specific climb direction being imposed by the stress, no average macroscopic strain could develop and there would be no creep. The overall comparison with atomistic simulations showed that a classical approach, at the mesoscopic scale, quantitatively described dislocation climb. Such an approach not only permitted rationalization of the results of atomistic simulations, but could also extrapolate them into the range of dislocation densities corresponding to experiment. Thanks to such an extrapolation, based upon a physically sound model, a fair test could be made of the validity of the creep model proposed by Kabir et al.; thus showing its inability to reproduce experimental data on the creep of iron.

Predicting Dislocation Climb: Classical Modeling Versus Atomistic Simulations. E.Clouet: Physical Review B, 2011, 84[9], 092106