The growth of new near-perfect grains during recrystallization of deformed metals is governed by the migration of the grain boundaries surrounding the new grains. The grain boundaries migrate through the deformed metal driven by the excess energy of the dislocation structures created during deformation. Recently, it has been found that recrystallization is far more inhomogeneous than previously thought. The purpose of this PhD-project is to study recrystallization by computer simulations with special focus on inhomogeneous growth. Two types of simulations have been employed: geometric- and molecular dynamics simulations (MD). The geometric simulation method developed and used in this study is a generalization of an existing method. The geometric simulations have been used to investigate simplistic situations where one type of growth-inhomogeneity exists in an otherwise very basic model. Simulations of grains possessing distributions of growth rates and simulations of anisotropic growing grains, where grains have individual preferred growth-directions, have been performed. The MD simulations have been used to study grain boundary motion driven by dislocation structures at the atomic level. This is the first time that dislocation structures have been used to drive grain boundary migration in MD simulations. Different types of dislocation structures and grain boundaries have been simulated using different inter-atomic potentials. The geometric simulations show that the introduction of growth-inhomogeneities into a simple recrystallization-model can affect the recrystallization kinetics and microstructure significantly, which makes it very important to understand the origin of such inhomogeneities. The MD simulations show that grain boundary migration during recrystallization is strongly affected by the dislocation structures in the deformed metal due to local effects: Inhomogeneous boundary morphologies and dislocation-structure-dependent migration rates are observed. The effects that the dislocation structures have must be taken into account in order to create realistic recrystallization models, and through that improve the processing and properties of metals.
Nanobioteknologi og medikomaterialer; Risø-PhD-36; Risø-PhD-36(EN); Risø-PhD-0036