Deformation and failure of glassy materials

Elastoplastic deformation of disordered solids and the formation of polymer crazes in amorphous polymer glasses are studied using large-scale molecular dynamics simulations. It is shown that the pressure-modified von Mises criterion accurately describes the maximum shear yield stress under general loading conditions. The pressure coefficient is insensitive to most model parameters, but is related to the bead geometry in analogy to friction coefficients. The yield stress decreases linearly with rising temperature and the strain rate dependence can be described by a power-law, or in a limited range, by a logarithm. The rate dependence does not vary with temperature, which is inconsistent with simple rate-state models of thermal activation such as the Eyring model. An analysis of the dynamics of the local stress distribution as well as modern phenomenological theories of rheology of glassy materials are discussed in light of these findings.; We then present a comprehensive investigation of the deformation of glassy polymeric systems into a dense load-bearing network of fibrils and voids called a craze at large strains. This expansion takes place in the form of a drawing process, where the strain rate is strongly localized in a narrow interface region between dense polymer and craze. The expansion is controlled by some polymer chain segments between entanglements that are stretched taut during crazing. We also find that the distribution of tension in the craze develops an exponential force tail in close analogy to compressed jammed systems such as granular media. This highly anisotropic stress distribution and the localization of large forces on relatively few chains indicate that earlier models of the crazing process that treat the polymer as a viscous fluid with hydrodynamic interactions are incorrect. Simulations and simple scaling arguments are presented that describe craze breakdown through disentanglement or chain scission.; Glassy polymers exhibit an unusually high fracture energy, which is 3–4 orders of magnitude larger than the interfacial free energy change. We conclude by calculating this enhancement in a multiscale approach that combines molecular level results of craze failure with an analytic argument for the onset of crack propagation on the continuum scale.