Brittle intergranular failure in 2D microstructures: Experiments and computer simulations

Abstract

Brittle intergranular fracture (BIF) is a common mode of failure for monolithic ceramics and intermetallics, as well as for some refractory metals and metals exposed to environmental corrosion, stress corrosion cracking or high temperature creep. As interest in applications for these materials grows, research programs have been developed to characterize and predict their fracture behavior. In order to experimentally quantify the effects of microstructure on local BIF, systems which have a minimum number of variables which influence fracture must be used. Evaluation of materials with two dimensional (2D) microstructures can considerably reduce the complexity of the system. In addition, providing a biaxial stress state in the 2D microstructure ensures that all boundaries experience exclusively Mode I loading prior to failure. Biaxial elastic loading of this simplified microstructure allows the calculation of (a) local stress and strain fields (and their concentrations) prior to failure, as well as (b) prediction of grain boundary strength criteria, and (c) prediction of intergranular crack paths. This can be achieved by conducting computer simulations of the experimentally observed fracture phenomena in polycrystalline specimens having a given texture and microgeometry. These simulations use high resolution finite-difference grids below the crystal scale, and involve the derivation of a spring-network model for arbitrary in-plane crystal anisotropy. Since the grain boundary strength criterion is easily controllable in such simulations, it can be inferred by a comparison with actual experimental results. The latter is complemented by results on fracture of materials with very weak grain boundaries, thus providing a clear perspective on evolution of the failure process for varying degrees of embrittlement.