Dislocations are linear topological defects which divide slipped and unslipped regions of a crystal lattice, which play an important role in determining the critical stress for shear deformation of crystalline materials. They also influence the chemical properties of a crystal, both because they enhance material transport by serving as fast diffusion pathways, and because the large strain fields in the vicinity of a dislocation core provide a substantial sink for point defects, including vacancies and chemical impurities. Indeed, calculations show that, in olivine, the concentration of incompatible elements such as Sr2+, Ba2+, and U4+ may be orders of magnitude greater near a dislocation than in the bulk crystal. In return, point defects can modify the intrinsic properties of a dislocation, whether by pinning it place, easing its motion through the crystal lattice, or even changing the atomic structure of its core.
The unifying feature of all aspects of point defect-dislocation interactions is that they are governed by atomic-scale processes, which can be difficult – even impossible – to study using existing experimental techniques. However, Moore’s Law increases in processing power combined with algorithmic improvements mean that it is now practical to simulate atomic clusters of sufficient size to accurately capture the properties of dislocation cores in complex crystalline materials, and to model their interactions with other defects. In this talk, I will describe a suite of computational tools which have been developed here to facilitate automated, high-throughput modelling of dislocations, and the ways in which they both influence, and are influenced by, point defects. These tools can used to better understand geological phenomena, ranging from changes in olivine deformation fabrics above subduction zones, to the observed short-range heterogeneity of incompatible element concentrations in highly deformed silicates.