Glass transition physics presents one of the most stubborn puzzles in condensed matter science. When you cool molten silica, something strange happens: the atoms slow down by a factor of about one quadrillion, yet they never arrange themselves into an orderly crystal.[s] The material becomes rigid while retaining the chaotic atomic arrangement of a liquid. Nobel laureate Philip Anderson called it “the deepest and most interesting unsolved problem in solid-state theory.”[s]
The Glass Transition Physics Problem
When pure water freezes at standard pressure, it releases heat and its molecules arrange into a crystalline lattice at 0°C. Glass formation works nothing like this. The temperature at which a liquid becomes glass depends on how fast you cool it. There is no latent heat release. The properties change smoothly rather than sharply.[s]
At the glass transition temperature, the structural relaxation time becomes so large that the liquid falls out of equilibrium on experimental timescales, transforming into a disordered amorphous solid.[s] The atoms are still trying to reach a more stable crystalline state, but they are moving so slowly that, as physicist Sriram Ramaswamy puts it, the material is “wandering desperately forever, not settling down to the state it actually wants to be in.”[s]
The Medieval Window Myth
You may have heard that old cathedral windows are thicker at the bottom because glass slowly flows over centuries. This claim belongs to a category of common physics misconceptions where the explanation most people learn turns out to be wrong. Scientists at Corning and Penn State definitively debunked this in 2017 using medieval glass from Westminster Abbey dating to 1268 AD.[s]
Their calculations revealed that medieval glass flows at most one nanometer over the course of one billion years.[s] The actual explanation for thickness variation is manufacturing: medieval windows were made using the crown process, in which glass was blown into a hollow globe, flattened, and spun into a disk. Panes cut from these disks were naturally thicker in the center, and glaziers installed them with the thicker end at the bottom for stability.[s]
Why Glass Breaks the Way It Does
Glass fractures in ways that seem excessive. Crack propagation in silica glass requires a fracture energy that far exceeds the thermodynamic cost of simply breaking bonds.[s] A 2026 molecular dynamics study using machine-learned interatomic potentials found that fracture energy rises by up to 33% even before cracks begin branching.[s]
A significant portion of the excess energy dissipates as heat, generating localized temperature spikes of several thousand Kelvin at the crack tip.[s] Fast-moving cracks create more than just additional surface area: they create a fundamentally different surface at the nanoscale, with altered energy density that would be invisible to standard post-mortem analysis.[s]
The Glass Transition Physics Debate
Physicists have argued for decades over whether glass formation is fundamentally a dynamic phenomenon or a thermodynamic one. The dynamic view holds that atoms simply slow down until the liquid appears solid, while its structure remains that of a liquid. The thermodynamic view insists some hidden structural change must drive the slowdown.[s]
Evidence supports both perspectives. In supercooled liquids, researchers observe dynamic heterogeneities: regions that suddenly relax next to regions that remain frozen. Does this correlate with structure? “What we find, what simulations find, what experiments find, time and again, is yes,” says physicist Rajesh Ganapathy.[s]
A 2026 study from Utrecht University added a new wrinkle: glass-like structures can exist in thermodynamic equilibrium, something many theories said should be impossible.[s] Using rod-shaped colloidal particles, researchers created a state where positions were disordered and frozen while particles could still rotate. “A glass and an equilibrium state exclude each other in many people’s minds,” notes researcher Thijs Besseling.[s]
What We Do Know
One structural account, configuron percolation theory, argues that glasses differ from liquids in their bond topology: thermally disrupted bonds in glasses constitute a small fraction of total chemical bonds, while liquids are overloaded with broken bonds.[s] In that framework, the formation of a percolation cluster made of these broken bonds results in the full loss of rigidity, which is the transformation of a solid into a liquid.[s]
Glass formers also fall into categories. Strong liquids like silica display nearly Arrhenius-like behavior with almost constant activation energy. Fragile liquids like o-terphenyl show a steep, super-Arrhenius increase with rapidly growing apparent activation energies.[s]
In conventional glasses, rigidity emerges without long-range order, and the materials are typically out of equilibrium.[s] The glass transition physics problem persists because these materials occupy a strange middle ground: solid enough to hold your drink, disordered enough to resist simple explanation.
Glass transition physics constitutes one of condensed matter science’s most persistent open problems. The phenomenology is dramatic: during vitrification, atomic mobility decreases by approximately 1015-fold while no long-range structural order emerges.[s] Philip Anderson characterized it as “the deepest and most interesting unsolved problem in solid-state theory.”[s] More than three decades after he predicted a breakthrough within a decade, the problem remains unsolved.
Glass Transition Physics Phenomenology
At the glass transition temperature Tg, the structural α-relaxation time becomes so large that the liquid falls out of equilibrium on experimental timescales, transforming into a disordered amorphous solid.[s] Unlike crystallization, vitrification exhibits no latent heat, no sharp temperature threshold, and Tg depends on cooling rate. The viscosity at calorimetric glass transition spans four orders of magnitude from 108.8 to 1013 Pa·s across different materials.[s]
The Angell fragility classification distinguishes glass formers by their deviation from Arrhenius dynamics. Strong liquids like SiO2 display nearly Arrhenius behavior with almost constant activation energy. Fragile liquids like o-terphenyl show steep, super-Arrhenius increases corresponding to rapidly growing apparent activation energies near Tg.[s]
The Structural Relaxation Myth Debunked
The claim that medieval cathedral windows exhibit thickness gradients from centuries of viscous flow represents a category of physics misconceptions where the explanation most people learn turns out to be wrong. Gulbiten and Mauro used Westminster Abbey glass from 1268 AD to calculate actual flow rates, finding medieval glass flows approximately 1 nm over 109 years.[s] The measured viscosity was 16 orders of magnitude lower than previous soda-lime silicate estimates, yet still far too high for observable flow on human timescales.
The thickness variation originates from the crown manufacturing process: glass blown into a hollow globe, flattened, and spun into a disk yields panes thicker at the center than edges.[s]
Dynamic Fracture Mechanics
Crack propagation in silica glass requires fracture energy far exceeding the thermodynamic cost of bond rupture.[s] First-principles molecular dynamics simulations using a machine-learned interatomic potential (Allegro MLIP trained on r2SCAN DFT data) demonstrate that structural fracture energy rises by up to 33% below the branching threshold at 0.72 of Rayleigh velocity.[s]
This increase partitions roughly equally between nanoscale roughening increasing real surface area and elevated intrinsic surface energy density. Dynamic fracture generates localized temperature spikes of several thousand Kelvin at the crack tip, driving fractoluminescence.[s] The critical stress intensity factor KIc measured computationally shows excellent agreement with experimental values.
Theoretical Frameworks in Glass Transition Physics
Mode-Coupling Theory (MCT) predicts atoms become increasingly caged by neighbors near a critical temperature, unable to escape or relax. MCT successfully captures early-stage dynamics but fails to explain relaxation beyond the predicted arrest temperature.[s] Dynamic Facilitation theory proposes cooperative movement enables relaxation even in deeply supercooled regimes, with the likelihood of cooperative regions decreasing at lower temperatures.[s]
Configuron Percolation Theory (CPT) offers a structural perspective: glasses differ from liquids in their bond topology. Thermally disrupted bonds in glasses constitute a small and often negligible fraction of total chemical bonds, while liquids are overloaded by broken bonds.[s] At Tg in this framework, a macroscopic percolation cluster of configurons (broken bonds with associated strain-releasing local adjustment) forms, causing complete loss of rigidity.[s]
The Hausdorff-Besicovitch dimensionality of the configuron set changes at Tg from 0 in glasses to approximately 2.5 in melts.[s]
Equilibrium Glass States
Utrecht University researchers demonstrated in 2026 that glass-like structures can exist in thermodynamic equilibrium, contradicting theoretical expectations.[s] Using long-ranged repulsive colloidal rods, they created a rotator glass phase where positional degrees of freedom are frozen while rotational freedom persists. When pushed toward crystallization via external electric field, the system relaxes back to the glassy state upon field removal.
Computer simulations corroborate the experimental findings. The results suggest rotational degrees of freedom may play an unappreciated role in glass transition physics, an element missing from many existing theories.[s]
Open Questions
The central debate persists: is vitrification purely kinetic or does a thermodynamic phase transition underlie it? Evidence for intermediate-range amorphous order growing in supercooled liquids supports the thermodynamic view. The theoretical Kauzmann temperature TK, where configurational entropy would vanish yielding an “ideal glass,” remains experimentally inaccessible because relaxation times diverge before reaching it.[s]
Glass transition physics may represent a problem that, as Ramaswamy suspects, “will never get resolved.” In conventional glasses, rigidity emerges without long-range order, and the materials are typically out of equilibrium.[s] Understanding this transition remains relevant for applications from ultrastable glasses in organic electronics to predicting behavior of biological systems exhibiting glass-like dynamics.[s][s]
