A Discrete Model for Diffusion-Induced Grain Boundary Deterioration

Abstract

Polycrystalline materials may suffer internal damage due to diffusion of chemically aggressive species during service. Diffusion rates are greatly enhanced on grain boundaries (GB). This can be modelled with discrete networks, where the GB structure is represented by links with local diffusivities. We present a site-bond model for concentration-driven diffusion that can be used to study the accumulation of chemical species at GB, leading to deterioration and eventual cracking. We employ realistic distributions of GB energies and corresponding diffusivities from published works. We show how the model can be used to predict macroscopic diffusivities with little experimentation. We demonstrate how the grain boundary structure controls the extent of internal damage resulting from the diffusion of chemical species.