Data-Mining of In-Situ TEM Experiments: Towards Understanding Nanoscale Fracture

Abstract

The lifetime and performance of any engineering component, from nanoscale sensors to macroscopic structures, are strongly influenced by fracture processes. Fracture itself is a highly localized event; originating at the atomic scale by bond breaking between individual atoms close to the crack tip. These processes, however, interact with defects such as dislocations or grain boundaries and influence phenomena on much larger length scales, ultimately giving rise to macroscopic behavior and engineering-scale fracture properties. This complex interplay is the fundamental reason why identifying the atomistic structural and energetic processes occurring at a crack tip remains a longstanding and still unsolved challenge. We develop a completely new approach combining quantitative in situ observations of nanoscale deformation processes at a crack tip with three-dimensional reconstruction of the dislocation structure and advanced computational analysis to address fracture initiation in a ductile metal. Our combinatorial approach reveals details of dislocation nucleation, their interaction process, and the local internal stress state, all of which were previously inaccessible to experiments. This enables us to describe fracture processes based on local crack driving forces on a dislocation level, a previously unseen high fidelity that paves the way to better understand and control local failure processes in materials.