The Glass That Breaks by a Single Law express
Glass is a material with attitude. It seems fragile, but powerful invisible forces can live inside it. In chemically strengthened glass, the surface is compressed like a stretched rubber band, while the interior is stretched. When a fast object hits such glass, that rubber band snaps in many places at once.
Scientists fired projectiles at glass and watched how it shattered. Then they built a computer model where the glass is represented as a network of springs, with internal stresses set as if 'frozen' deformation. The model faithfully reproduced the fracture pattern.
The big surprise: a weak impact produces large fragments, a strong one produces small ones, but if you plot 'how many fragments of each size,' all the curves collapse onto a single one. Just divide each fragment's area by the average area, and the same law holds for all impacts and all stress distributions.
At the microscale, something unusual emerged: ahead of the main crack, individual bonds snap, and for a brief moment the crack runs faster than sound in that glass. Then it stops, leaving 'tongues' and zones with tiny steps on the fracture surface: those very marks seen under a microscope.
So glass shattered in the lab tells us about how not just windows break, but also phone screens, car windshields, and protective panels.
Original abstract
as written by the authors arXiv:2602.20443
The dynamic fragmentation of residually stressed solids involves a complex interplay between stored elastic energy, stress wave propagation, and crack instabilities. In this work, we investigate the fracture mechanics of chemically toughened glass through high-velocity projectile impact experiments and a novel micromechanical network model. We rigorously incorporate residual stress into the discrete lattice framework via a prescribed inelastic strain (eigenstrain) distribution, formulated as equivalent body and surface forces to ensure mesh-independent fracture paths. Our experiments and simulations demonstrate that while the fracture topology shifts from coarse to fine with increasing impact energy, the cumulative fragment size distribution consistently follows an exponential decay. Crucially, we reveal a universal scaling law: fragment size distributions from diverse loading conditions and stress profiles collapse onto a single master curve when normalized by the mean fragment area. Furthermore, the model elucidates the determinants of fragmentation, showing that the resulting fragment size is governed not only by the magnitude of residual stress but also by the steepness of the stress gradient. At the microscopic scale, we identify a mechanism for dynamic instability where non-sequential bond breaking ahead of the crack tip leads to apparent local crack speeds exceeding the Rayleigh wave speed ($$c_r$$). These arrested micro-branches, analogous to the Burridge-Andrews mechanism, provide a physical explanation for the "tongue-like" features and hackle zones observed in post-mortem fractography.
The speed of sound in glass is about 5–6 kilometers per second. In the experiment, microcracks briefly outran it, even though that's impossible under normal conditions.