Void nucleation in heterogeneous materials induced by particle-matrix debonding in polycrystalline matrix

Abstract

Void nucleation represents the initial stage of void evolution and plays a pivotal role in subsequent damage progression in ductile materials. This study focuses specifically on void nucleation resulting from the debonding of spherical particles from a heterogeneous polycrystalline matrix. In contrast to most previous models that assume a homogeneous matrix surrounding the particle, this study introduces a more realistic representation by modeling the matrix as an assembly of grains with randomly assigned morphologies and crystallographic orientations. The crystal plasticity (CP) model and the cohesive zone model (CZM) are employed to simulate the plastic deformation of grains and the debonding behavior at the particle–matrix interfaces, respectively. Representative volume elements (RVEs), each containing multiple grains and a single spherical particle, are used in the simulations. Three particle volume fractions and two particle positions (intragranular and intergranular) are considered. For each combination, 200 RVEs with randomized grain morphologies and orientations are generated to investigate the influence of microstructural variability on the critical stress and equivalent strain required for void nucleation under various triaxial stress states. A comprehensive dataset was generated through extensive finite element simulations, followed by statistical analysis. Based on the results, two bivariate normal distribution models were developed, each employing a specific pair of variables—critical combination stress and equivalent strain—to characterize void nucleation behavior. Compared to previous homogenized void nucleation models, the proposed models account for grain morphology and orientation effects, thereby more accurately capturing the stochastic nature of void nucleation in heterogeneous polycrystalline materials under different triaxial stress conditions and particle volume fractions. These models have also been applied in macro-level finite element (FE) simulations to characterize the stochastic behavior of void nucleation in such materials.