Glass Is Not a Slow Liquid: The Amorphous Solid Explained

Glass at room temperature has a viscosity of approximately 10²¹ Pa·s — liquid water is 10⁻³ Pa·s

The claim that glass is a slow-moving liquid — often illustrated by pointing to thicker panes at the bottom of old cathedral windows — is one of the most persistent myths in popular science. It appears in museum placards, physics lectures, and science communication articles worldwide. The quantitative reality makes the claim untenable: glass at room temperature is immovably rigid on any human or geological timescale. Its viscosity is approximately 10²¹ pascal-seconds, compared to 10⁻³ Pa·s for water and 10¹¹ Pa·s for glacier ice. Understanding what glass actually is — an amorphous solid, not a supercooled liquid in any practical sense — requires examining both its molecular structure and the physics of the glass transition.

Crystalline vs. amorphous structure

Most solids are crystalline: their constituent atoms or molecules are arranged in long-range repeating lattice structures. Metals, salts, ice, and most minerals are crystalline. The regular arrangement is thermodynamically favored — it represents the lowest energy configuration at equilibrium below the melting point.

Glass has no such long-range order. In silicate glass (the most common type — typically SiO₂ combined with Na₂O, CaO, and other additives), silicon atoms form tetrahedral units (SiO₄) bonded through oxygen atoms, but without the repeating 3D periodicity of a crystal. The structure is disordered — more similar in arrangement to a snapshot of a liquid than to a crystal. This disorder is frozen in during rapid cooling from the melt.

• The absence of long-range order means glass has no sharp melting point — it softens gradually over a temperature range
• Glass has no grain boundaries (unlike polycrystalline metals) — this is why it fractures as a smooth conchoidal surface rather than along crystalline planes
• X-ray diffraction of glass shows broad, diffuse rings rather than the sharp diffraction spots of a crystal — direct evidence of structural disorder
• Despite disorder, glass retains short-range order at the scale of a few angstroms — the SiO₄ tetrahedra themselves are regular

The supercooled liquid: technically correct, practically misleading

In thermodynamic terms, glass is rigorously classified as a supercooled liquid — a substance that has been cooled below its crystallization temperature without crystallizing. The glass transition is not a thermodynamic phase transition in the classical sense (there is no latent heat, no discontinuous jump in density or enthalpy at the transition); it is a kinetic phenomenon where molecular relaxation times exceed the timescale of the experiment.

Above the glass transition temperature (Tg), the material behaves as a viscous liquid — molecules can rearrange on laboratory timescales in response to stress. Below Tg, molecular rearrangement is so slow that the material appears — and for all practical purposes, is — solid. For common soda-lime glass (window glass), Tg is approximately 520–600°C. At room temperature (20°C), the glass is approximately 500°C below its glass transition.

The cathedral glass myth: debunked with measurements

The claim that old cathedral windows demonstrate glass flow by being thicker at the bottom has been thoroughly investigated and found false on multiple grounds. Robert Brill at the Corning Museum of Glass examined medieval European glass panes and found no systematic correlation between age and thickness variation. Edgar Zanotto at the Federal University of São Carlos calculated in 1998 that for glass to flow measurably at room temperature — even by just one atomic diameter — would require approximately 10³² years, roughly 10²² times the current age of the universe.

Why are old glass panes often thicker at the bottom? The answer lies in medieval glassmaking technique, not physics. The dominant method — the crown glass process — produced circular discs of spun glass with inherent thickness variation: the "crown" (bull's eye center) was thicker, as were the edges. Glaziers installing these panes typically placed the heavier, thicker edge downward for stability. The thickness variation is an artifact of manufacture and installation practice, not centuries of gravitational flow.

• Zanotto (1998) in the American Journal of Physics provided the definitive quantitative treatment of glass flow timescales
• Some medieval glass panes are actually thicker at the top — directly contradicting the flow-from-above narrative
• Plate glass and float glass (post-1950s) are manufactured to uniform thickness, eliminating any remaining ambiguity

Glass transition temperature and material design

The glass transition temperature (Tg) is a critical material parameter that defines the temperature range over which glass transitions between rigid solid behavior and viscous liquid behavior. Tg depends on chemical composition and is a key design variable in materials engineering.

Metallic glasses — alloys cooled rapidly enough to avoid crystallization — represent a growing class of engineering materials with superior strength and elasticity compared to their crystalline equivalents. Their Tg values make them processable at relatively accessible temperatures while remaining structurally superior to crystalline metals at room temperature.

Glass is a solid. It flows through geological time, if at all. Any physicist who says otherwise is speaking imprecisely about the thermodynamic classification, not predicting your windows will sag.