Résumé:

Virtually all real-life crystalline materials have defects. In particular, most metals and ceramics are aggregates of crystalline grains. Grain-boundaries (GBs), the two-dimensional lattice defects that separate the different grains of a crystal, control the mechanical properties of polycrystalline materials. Although GB motion is known to play important roles in plastic (i.e. irreversible) deformation, the microscopic origin of the plasticity of polycrystalline materials is still largely unknown, because of the limitations of available experimental tools to record, during deformation, the dynamics of the process with a nanometer resolution. To overcome these limitations, we use a colloidal analog of atomic polycrystals obtained by doping a copolymer micellar crystal with nanoparticles. Nanoparticles act as impurities, and as such, they segregate in the grain boundaries of the colloidal polycrystal, allowing their visualization by light and confocal microscopy and by scattering techniques. In addition, the microstructure of the polycrystal can be tuned by varying the nanoparticle volume fraction and the crystallization rate. To investigate the plasticity of colloidal polycrystals, we perform multispeckle time-resolved dynamic light scattering measurements on the samples submitted to cyclic shear deformations using a novel light scattering apparatus specifically designed to access the dynamics of the network of GBs. Plasticity is quantified by analyzing the correlation of the scattered intensity measured after a given number of deformation cycles. We demonstrate that the shear-induced GB dynamics at the origin of plasticity is aging in a length-scale dependence manner. We extract a characteristic length above which the GB dynamics is stationary, which is found to increase with the shear amplitude. Our data suggest a hierarchical organization of the GB network under shear and provides a novel framework to understand the plasticity of polycrystals and of disordered materials in general.