The energy cusp that appeared around the [001]  = 5 twist grain boundary in this compound was studied by using molecular dynamics techniques. A definite cusp was found at  = 5. However, it was much shallower than those predicted by previous simulations of Cu or Au [001] twist boundaries. A proposed Read-Shockley type formula did not fit the energy data well when the Burgers vector of the expected secondary dislocations was used. Also, the corresponding structural unit description of near  = 5 boundaries was found here to be ambiguous. Dissociation of the secondary screw dislocation grid, into a complicated network of partial dislocations, was suggested to be a possible explanation for these discrepancies. It was also suggested to account for the very shallow cum indiscernible cusps which existed around [001]  = 5 twists in certain materials. It was noted that all of the twist boundaries within the misorientation range of  = 17 to  = 5 could have energy states that fell well below the energy versus misorientation curve. These corresponded to structures which contained well-formed single crystal structural units.

I.P.Antoniades, A.J.Patrinos, G.L.Bleris: Materials Science Forum, 1996, 207-209, 833-6