Multiscale computational modeling of deformation mechanics and intergranular fracture in nanocrystalline copper

This study presents the development and validation of a two-scale numerical method aimed at predicting the mechanical behavior and the inter-granular fracture of nanocrystalline (NC) metals under deformation. The material description is based on two constitutive elements, the grains (or bulk crystals) and the grain-boundaries (GBs). Their behaviors are determined atomistically using the quasicontinuum (QC) method by simulating the plastic deformation of [ 1 1 ‾ 0 ] tilt crystalline interfaces undergoing simple shear, tension and nano-indentation. Unlike our previous work (Péron-Lührs et al., 2013) however, the GB thickness is here calibrated in the model, providing more accurate insight into the GB widths according to the interface misorientation angle. In this contribution, the new two-scale model is also validated against fully-atomistic NC simulation tests for two low-angle and high-angle textures and two grain sizes. A simplified strategy aimed at predicting the mechanical behavior of more general textures without the need to run more QC simulations is also proposed, demonstrating significant reductions in the computational cost compared to full atomistic simulations. Finally, by studying the response of dogbone samples made of NC copper, we show in this paper that such a two-scale model is able to quantitatively capture the differences in mechanical behavior of NC metals as a function of the texture and grain size, as well as to accurately predict the processes of inter-granular fracture for different GB character distributions. This two-scale method is found to be an effective alternative to other atomistic methods for the prediction of plasticity and fracture in NC materials with a substantial number of 2-D grains such as columnar-grained thin films for micro-scale electro-mechanical devices.