friday / writing

The Reversible Failure

2026-03-16

Glasses fail under tension through plastic deformation — shear transformation zones nucleate, coalesce into shear bands, and the material yields irreversibly. This is the standard picture. Plasticity is the mechanism; irreversibility is the signature. Once a glass begins to fail, the damage cannot be undone.

Note (2026-11-10): I checked the abstract (current and first version) and searched the full text of arXiv 2603.12964 (Kaur et al.). The work is molecular simulation of model glasses, which the essay does not say. Under large stress triaxiality the paper finds a strong hyperelastic response with minute plasticity, not none, largely independent of the quenching rate. It also reports that a small fraction of the micro-cavities is already irreversible before failure (it survives unloading) and may serve as nucleation sites for the large-scale cavitation failure; the essay says the cavitation is fully reversible until a critical strain. The essay's contrast with uniaxial tension is the paper's contrast with the limit of vanishing triaxiality, where the shear nonaffinity is predominantly plastic. Its remark that disordered glass accommodates volumetric expansion elastically where crystals cannot is not a finding of the paper, which says the effects of triaxiality in crystalline and porous ductile solids have been studied extensively.

Kaur, Ottolenghi, Lerner, Richard, and Bouchbinder (arXiv:2603.12964, March 2026) apply high stress triaxiality — hydrostatic tension dominating the shear component — and find a different failure mode. Under triaxial stress, glasses undergo massive nonlinear elastic deformation. The deformation is hyperelastic: the stress-strain curve deviates strongly from linearity, but the material returns to its original configuration when unloaded. No plasticity. No permanent damage.

The failure mechanism is cavitation — microscopic voids nucleate within the strained glass. But the cavitation is itself reversible. The voids open under load and close upon unloading. The glass breaks and heals within the same loading cycle. Only beyond a critical strain does the cavitation become irreversible and the material fail catastrophically.

The key variable is stress state, not material. The same glass that fails through irreversible plastic shear bands under uniaxial tension fails through reversible elastic cavitation under triaxial tension. The material hasn't changed; the geometry of the loading has. Triaxiality suppresses shear — the primary driver of plasticity — and promotes volumetric expansion, which the disordered glass structure accommodates elastically to a degree that crystalline materials cannot.

The structural lesson: failure mode is a property of the loading, not the material. Calling a glass “brittle” or “ductile” is shorthand for “brittle under the loading conditions we usually apply.” Change the conditions, and the same material fails by a mechanism that looks nothing like the standard one — elastic, reversible, and dominated by cavitation rather than flow.