Atomistic breathing shell model calculations of dislocation core configurations in ionic crystals

Abstract

A study was undertaken to improve upon a previous rigid boundary atomistic calculation of the core properties of a/2〈110〉{110} edge dislocations in MgO. A flexible boundary method (Flex-II) based on linear elasticity theory and recently developed by Hoagland, Hirth and Gehlen (1976) was used to provide a more accurate determination of the boundary between the atomistic and the elasto-atomic regions. A central-force breathing shell model due to Sangster (1973) was used to take account of many-body effects in the interaction between ions. This marks the first application of the breathing shell model to the calculation of dislocation core configurations in ionic crystals. Within this model two potentials were used. One is due to Sangster (1973), the other is our previously developed Model 1 potential. The results obtained using the above potentials are analysed in terms of the stability of the dislocation symmetry type, the core displacement field, the volume of expansion and the total strain energy. It is found that, except for the strain energy, differences between the results for the two potentials are small. Thus for both potentials : (a) type I symmetry is the stable dislocation configuration while type II is on top of the Peierls energy barrier; (b) the strain field in regions a number of Burgers vectors beyond the core can be adequately expressed as the sum of the Volterra field and a linear elastic expansion field; (c) the dislocation causes a volume expansion of the entire crystal of 0·9 at. vol/plane. Non-linear effects and the core expansion field contribute to an increase in the atomistically determined strain energy factor, K _at, over the linear anisotropic elastic energy factor, K _el. The K _at results for Sangster's potential show that non-linear effects are stronger for this than the Model 1 potential. Calculations using Flex-II with the point ion and shell models also show that the results derived from the Model 1 potential are generally in better agreement with the breathing shell model than those derived from the previously developed Model 2 potential. Finally, a comparative study of the Flex-II method with the rigid boundary and other flexible boundary methods shows that, in terms of computational efficiency for a given accuracy, Flex-II is superior.