Nanostructure and surface effects on yield in Cu nanowires

The yield strengths of nanomaterials are highly sensitive to their internal and surface structures. However, it is difficult to identify a priori which structural feature will govern plastic yield. We employ very large scale molecular dynamics simulations to explicitly identify the relevant yield mechanisms for Cu nanowires with four distinct, experimentally realizable nanostructures: single crystal (SC), nanotwinned single crystal (NTSC), nanocrystal (NC) and nanotwinned nanocrystal (NTNC). By characterizing the deformation at the yield point on the atomic scale, our simulations elucidate the effects of surface defects, nanotwins and grain boundaries on the commencement of yield and reveal several critically important features of the yielding process. First, the initial yields in all nanowires occur via dislocation nucleation at different characteristic nanostructural features. SC and NTSC nanowires yield via dislocation nucleation from surfaces or surface defects, while NC and NTNC nanowires yield via dislocation nucleation from grain boundary triple junctions. Second, our simulations highlight the relative potency of stress concentrators arising from different imperfections in modulating the yield strength of nanowires. Grain boundary triple junctions are as effective as surface defects at acting as stress concentrators. However, the higher density of triple junctions in NC and NTNC nanowires renders these structures considerably weaker than their SC and NTSC counterparts. Third, the presence of nanotwins only marginally enhances the yield strength of nanocrystalline Cu nanowires, which is in line with experimental observation in NTNC Cu nanowires but contrary to that in bulk ultrafine-grain nanotwinned Cu. The reason for this divergent behavior is that in nanowires yield strength is governed by dislocation nucleation from triple junctions in contrast to dislocation propagation in the bulk. Finally, excellent agreement is obtained between the relative yield strengths, stress–strain behavior and dislocation nucleation conditions of nanowires in our simulations and existing experimental data. This suggests that our predicted atomistic processes controlling yield in our simulations may also control yield in experiments.