Topological defects govern crack front motion and facet formation on broken surfaces

TL;DR

This paper reveals that topological defects in crack fronts, specifically steps, are key to understanding the formation of facets and patterns on fractured surfaces, linking 3D topology to 2D fracture dynamics.

Contribution

It demonstrates that steps are topological defects in crack fronts, explaining their stability, motion, and role in surface pattern formation during fracture.

Findings

Steps are topological defects that stabilize facet formation.

Crack front segments overlap, enabling step stability.

Step dynamics are coupled to local dissipation and crack front deformation.

Abstract

Patterns on broken surfaces are well-known from everyday experience, but surprisingly, how and why they form are very much open questions. Well-defined facets are commonly observed1-4 along fracture surfaces which are created by slow tensile cracks. As facets appear in amorphous materials5-7, their formation does not reflect microscopic order. Fracture mechanics, however, predict that slow crack fronts should be straight, creating mirror-like surfaces8-13. In contrast, facet-forming fronts propagate simultaneously within different planes separated by steps. It is therefore unclear why steps are stable, what determines their path and how they couple to crack front dynamics. Here we show, by integrating real-time imaging of propagating crack fronts with surface measurements, that steps are topological defects of crack fronts; crack front separation into discontinuous overlapping segments provides the condition for step stability. Steps drift at a constant angle to the local front propagation direction and the increased local dissipation due to step formation couples to the long-range deformation of the surrounding crack fronts. Slow crack front dynamics are enslaved to changes in step heights and positions. These observations show how 3D topology couples to 2D fracture dynamics to provide a fundamental picture of how patterned surfaces are generated.