Molecular dynamics (MD) simulations of fused silica glass deforming in pressure-shear, while revealing useful insights into processes unfolding at the atomic level, fail spectacularly in that they grossly overestimate the magnitude of the stresses relative to those observed, e. g., in plate-impact experiments. We interpret this gap as evidence of relaxation mechanisms that operate at mesoscopic lengthscales and which, therefore, are not taken into account in atomic-level calculations. We specifically hypothesize that the dominant mesoscopic relaxation mechanism is shear banding. We evaluate this hypothesis by first generating MD data over the relevant range of temperature and strain rate and then carrying out continuum shear-banding calculations in a plate-impact configuration using a critical-state plasticity model fitted to the MD data. The main outcome of the analysis is a knock-down factor due to shear banding that effectively brings the predicted level of stress into alignment with experimental observation, thus resolving the predictive gap of MD calculations.

熔融石英玻璃在压力剪切作用下变形的分子动力学（MD）模拟，尽管揭示了原子级展开过程的有用见解，但由于相对于观察到的应力而言高估了应力的大小，因此失败了。例如，在板撞击实验中。我们将这种间隙解释为弛豫机制的证据，该机制在介观长度尺度上起作用，因此在原子级计算中未考虑在内。我们专门假设主要的介观弛豫机制是剪切带。我们首先通过在相关温度和应变率范围内生成MD数据，然后使用适合于MD数据的临界状态可塑性模型，在板碰撞配置中进行连续剪切带计算，从而评估该假设。