A bond-order potential for hexagonal close-packed Ti, and a central-force Finnis-Sinclair potential, were used to study the core structure of the 1/3<1¯2▪0> screw dislocation atomistically. It was noted that the qualitative features of the core structures were similar in the 2 cases. The dislocation could either dissociate into Shockley partials on the basal plane, or spread in a continuous manner into the prism plane. However, spreading of the dislocation core into the prism plane was always energetically favored over splitting into the basal plane, in the case of the bond-order potential. The opposite was found to be true in the case of the central-force Finnis-Sinclair potential. The results which were obtained by using the bond-order potential thus explained the strong preference for prism slip over basal slip in Ti. The most important overall parameter was the energy of the intrinsic stacking fault on the basal plane. This was so high in the case of the bond-order potential that splitting into Shockley partials was not energetically favorable. The reason for such a high stacking-fault energy was the non-central character of atomic interactions which arose from the significant contribution that d-electrons made to the bonding. This was correctly modelled by the bond-order potential.

Atomistic Simulation of Titanium: Structure of 1/3<¯12▪0> Screw Dislocations and Slip Systems in Titanium A.Girshick, D.G.Pettifor, V.Vitek: Philosophical Magazine A, 1998, 77[4], 999-1012