The stress tensor of an atomistic system

We prove that the stress tensor conservation equation expressing the local equilibrium condition of a body results from the invariance of its partition function under canonical point transformations. From this result the expression of the stress tensor of a general atomistic system (with short range interactions) in terms of its microscopic degrees of freedom can be obtained. The derivation, which can be extended to encompass the quantum mechanical case, works in the canonical as well as the micro-canonical ensemble and is valid for systems endowed with arbitrary boundary conditions. As an interesting by-product of our general approach, we are able to positively answer the old question concerning the uniqueness of the stress tensor expression.

A MATHEMATICAL MODEL FOR GRAIN GROWTH IN THIN METALLIC FILMS

We use geometrical arguments based on grain boundary symmetries to introduce crystalline interfacial energies for interfaces in polycrystalline thin films with a cubic lattice. These crystalline energies are incorporated into a multi-phase field model. Our aim is to apply the multi-phase field method to describe the evolution of faceted grain boundary triple junctions in epitaxially growing microstructures. In particular, we are interested in symmetry properties of triple junctions in tricrystalline thin films. Symmetries of triple junctions in tricrystalline films have been studied in experiments by Dahmen and Thangaraj.6,25 In accordance with their experiments, we find in numerical simulations that any two neighboring triple junctions belong to different symmetry classes. We introduce a local equilibrium condition at triple junctions which can be interpreted as a crystalline version of Young's law. The local equilibrium condition at triple junctions is purely determined by the grain boundary energies. In particular no triple junction energies are necessary to explain which triple junctions are possible. All triple junctions observed in the experiments as well as in the simulations fulfil the crystalline version of Young's law. Our approach is also capable of describing grain boundary motion in general polycrystalline thin films.