以“软度”衡量非晶态固体塑性的结构起源:一种基于机器学习的定量理解（英文）

“Softness” as the structural origin of plasticity in disordered solids: a quantitative insight from machine learning

非晶态固体没有长程的平移对称性,因而缺少像晶体中那样具有明确定义的缺陷及其运动来解释塑性的产生.长期以来,人们推测非晶态固体的塑性可能起源于物理意义上的局部软区.与此相比,在Cubuk等人最近发表在《科学》杂志上的论文中,作者基于机器学习技术定义了一个微观结构量"软度",并提出可以通过"软度"来衡量多种不同非晶态固体(分子玻璃、胶体玻璃、金属玻璃等)的塑性的结构起源.尽管基于机器学习而得到的"软度"的物理意义仍值得进一步研究,但是由此得到的"软度"区域确实显示出与局部重排区域非常好的相关性.这个发现为探究多种非晶态固体塑性的结构起源提供了一个定量的理解.

Disordered solids,including various types of glasses,such as metallic glasses (MGs),ceramic glasses and polymeric glasses,possess an amorphous or aperiodic arrangement of their building blocks,such as atoms,ions and molecules.At first sight,this is similar to the structure of their corresponding liquids,lacking any long-range translational symmetry as in crystalline solids.As a result,disordered solids usually exhibit a unique combination of properties,and are an indispensable class of engineering materials.For instance,MGs exhibit superb strength, high elastic limit,and outstanding corrosion-and wear- resistance,which is ideal for making sporting and medical apparatus.However,disordered solids,particularly structural glasses,could be very brittle under an ambient condition.Catastrophic failures can take place in them without noticeable plasticity [1,2].In practice,this severely hinders the applications of disordered solids as a reliable structural material;therefore,the origin of plas- ticity in disordered solids,if any,has been remaining in decades as one of the most active topics in materials science and engineering.