Dislocation propagation versus dislocation nucleation

Authors' response —

This issue needs to be analysed in the context from which it was quoted. In that paragraph of our text¹, we discuss the dislocations accumulated during sample preparation inside the grain volume between the grain boundaries. Such is the case of ref. 2 that we were commenting on. These grains, above 100 nm in size, are relatively large compared with those modelled in molecular dynamics simulations, and were produced by repeated rolling. As such, there are pre-existing dislocations in the interior of the grains^1,2. Some of these stored dislocations can move relatively easily on subsequent tensile testing with little hindrance from the high-angle grain boundaries and without the need to nucleate new dislocations. The plastic strains observed in the tensile tests were also partly attributed to the propagation of these ready-to-go dislocations². On the other hand, dislocation nucleation, which has been studied before for dislocation emission from the grain boundaries (hence the citation of the paper by Van Swygenhoven et al.³), requires high stresses at the grain boundaries and becomes necessary and dominant in small nanocrystalline grains with clean interiors, or after annealing wipes away those dislocations introduced into the grains during rolling².

As for the general importance of dislocation nucleation versus mobility, our article did not hold the opinion that nucleation must always be the rate-limiting process in plastic flow. In fact, we have mentioned this in a later paragraph, and previously discussed several mechanisms^4,5,6 for nanograins (especially those with grain size well below 100 nm). We proposed that dislocation-grain-boundary interactions, including dislocation emission from the grain boundaries as well as dislocation motion overcoming hindrance in the grain-boundary regions, can become the thermally activated, rate-limiting processes^4,5,6.