Atomistic Simulation of Ferroelectric Domain Walls

Abstract

Atomistic simulation of moving discommensurations is most useful when atomic details have a strong influence on the outcome. A classical example is the problem of dislocation core and movement in which the nature of atomic bonding has a direct effect on the configuration of the dislocation core, which in turn affects the manner the dislocation moves [1]. The final macroscopic characteristics of plastic deformation are intimately related to these details, of which a complete understanding can only come from the atomistic simulation of both the static and the (activated) non-equilibrium configurations of dislocations. The extension to the interface problem is encountered in martensitic transformations, involving either the parent/product interface or the product/product interface, which we call variant interface [2]. Variant interfaces are mobile and may be regarded as a group of dislocations, and like dislocations they move under a stress. The problem is relevant to shape memory alloys in which the stress and rate dependence of shape change is ultimately controlled by the atomic configurations at variant interfaces during their movement, even though the final (presumably equilibrium) multi-variant configurations are dictated by crystallography and elastic energy minimization. As a further extension to the interface problem, domain walls in ferroelectric crystals are like variant interfaces in that, crystallographically, both belong to the class of twin boundaries, but domain walls are complicated by electrostatic considerations [3, 4].