It was shown that for dislocations in crystalline solids, except for pure edge and screw dislocations, the local relative displacement of the two sides of the glide plane always deviated from the direction of the Burgers vector of the crystallographic Volterra dislocation. The generalized Peierls model was used; which made the simplifying assumption that the main atomic misfit was planar and was concentrated in the glide plane. Atomistic simulations showed that, for dislocations with an edge component, this assumption was generally satisfied. However, for some pure screw dislocations (as in body-centered cubic crystals), non-planar dissociation could occur. Division into an elastic energy for half-spaces, and a so-called atomic energy for glide planes, was useful but nevertheless somewhat artificial. The elastic energy diverged logarithmically in an infinite medium and, with a cut-off radius of some 1000b, was still a factor of 2 to 6 times larger than the atomic misfit energy; depending upon how the latter was defined. Hence, variations in the elastic energy, due to changes in shape, would have a strong effect upon the core configuration. It was expected that atomistic simulations of the dislocation core would also exhibit deviations when the cut-off radius for the region which was treated atomistically was sufficiently large.

Deviations from Volterra Dislocations. G.Schoeck: Philosophical Magazine Letters, 1998, 77[3], 141-6