Twist boundary energies for metals with long-ranged pairwise interatomic potentials

The recent calculation of Devlin (1982) predicting that the energy of rigid (001) twist boundaries in a central force pairwise interaction model with a long-ranged potential increases linearly with Sigma (for 5<or= Sigma <or=37) is considered. By use of a general argument based on the 'geometry of numbers' it is shown that this linear increase of the energy is too strong to be physically reasonable. Instead, the energy approaches a constant asymptotic value in this range, and any variations can be represented by a higher order term which varies as Sigma ^-34/.