Investigating damage evolution at the nanoscale: Molecular dynamics simulations of nanovoid growth in single-crystal aluminum

Nanovoid growth was investigated using molecular dynamics to reveal its dependence on void size, strain rate, crystallographic loading orientation, initial nanovoid volume fraction, and simulation cell size. A spherical nanovoid was embedded into a periodic face-centered cubic (fcc) Al lattice, and a remote uniaxial load was applied to elucidate dislocation nucleation and shear loop formation from the void surface as well as the subsequent void growth mechanisms. The nucleation stresses and void growth mechanisms were compared for four different strain rates (10⁷ to 10¹⁰ seconds ^-1), five different simulation cell sizes (4-nm to 28-nm lengths), four different initial nanovoid volume fractions, and seven different tensile loading orientations representative of the variability within the stereographic triangle. The simulation results show an effect of the size scale, crystallographic loading orientation, initial void volume fraction, and strain rate on the incipient yield stress for simulations without a void (single-crystal bulk material). For instance, the crystallographic orientation dependence on yield stress was less pronounced for simulations containing a void. As expected, dislocations and shear loops nucleated on various slip systems for the different loading orientations, which included orientations favored for both single slip and multiple slip. The evolution of the nanovoid volume fraction with increasing strain is relatively insensitive to loading orientations, which suggests that the nanoscale plastic anisotropy caused by the initial lattice orientation has only a minor role in influencing the nanovoid growth rate. In contrast, a significant influence of the initial nanovoid volume fractions was observed on the yield stress, i.e., a ∼35 pct decrease in yield stress was caused by introducing a 0.4 pct nanovoid volume fraction. Furthermore, a continuum-scale bridging parameter m - which is a material rate sensitivity parameter in continuum damage mechanics - was calculated and found to be close to 1. Consequently, atomistic simulations of this type can indeed inform continuum void growth models for application in multiscale models.