You've probably heard this one: the stained-glass windows of medieval cathedrals are thicker at the bottom because glass is secretly a liquid, slowly flowing downward over the centuries.
It's a lovely story. It's also half wrong — and the half that's wrong has been debunked with actual numbers for almost thirty years. Cathedral panes really are thicker at the bottom. But they didn't get that way by flowing. And a physicist has calculated exactly how long window glass would need to flow visibly at room temperature: a timescale that makes the age of the universe look like a rounding error.
So is glass a solid? Yes — but a strange one: a solid whose atoms are arranged exactly like a liquid's. How that state comes to exist is a problem a Nobel laureate called the deepest unsolved question in solid-state physics. Here's the whole story.
Left: medieval panes really are thicker at the bottom. Right: the actual reason — crown glass was spun into a disc of inherently uneven thickness, cut into panes, and glaziers tended to set the heavy edge down.
First, the truth about cathedral windows
Medieval window glass wasn't produced as flat, uniform sheets. In the crown method, a blob of molten glass was spun rapidly on a rod until centrifugal force flattened it into a disc — like spinning pizza dough — and that disc was inevitably thicker in some places than others. The alternative cylinder method (blow a cylinder, split it, flatten it by hand) was no more uniform.
When cutting those wobbly discs into panes and fitting them into frames, glaziers appear to have set the heavier, thicker edge down — the stable choice. (No medieval installation manual survives; this is inferred from the statistics of surviving windows.) So cathedral glass isn't thick at the bottom because it flowed. It was installed that way.
Three observations kill the flowing-glass story outright:
- Roman glassware, a thousand years older than any cathedral, shows no sagging at all in museums worldwide. Older glass should have flowed more.
- Antique telescope lenses and mirrors still deliver precise optics centuries later. Any flow would have wrecked their figure.
- Some cathedral panes are found thick side up or sideways. Apparently some glaziers just weren't fussy.
The physicist who actually did the math
"Fine, but couldn't it flow just a tiny bit?" In 1998, glass scientist Edgar Zanotto (Federal University of São Carlos, Brazil) answered that question head-on in the American Journal of Physics, in a paper bluntly titled "Do cathedral glasses flow?"
The logic: how fast a material flows is set by its viscosity, which for glass climbs explosively as temperature drops. Estimate the viscosity of medieval window glass at room temperature, and you can estimate how long visible sagging would take. His result:
Even GeO₂ glass — among the most flow-prone glasses at room temperature — would need 10³² years to sag visibly. Cathedral glass would take even longer.
A 1999 follow-up (Zanotto & Gupta) refined the estimate for window-glass compositions to about 10²³ years — a smaller number, but still ten trillion times the age of the universe (~10¹⁰ years). And in 2018, researchers measured the viscosity of a synthetic glass reproducing a medieval Westminster Abbey composition: relaxation turned out much faster than Zanotto's bound, yet the flow still amounts to about one nanometer per billion years. Every method lands on the same conclusion. Your windows are not flowing. Neither will your great-great-grandchildren's.
(Curious how a number like 10³² years is even calculated, given nobody can measure room-temperature glass viscosity? I walk through Zanotto's actual equations and parameters in a companion post.)
Why the myth felt so plausible
The myth survived because it contains a real piece of physics: glass genuinely has the atomic structure of a liquid.
An ordinary solid is a crystal — atoms lined up in a repeating lattice (salt, diamond, copper, ice). X-ray a piece of glass and you find no repeating pattern at all: the atoms are tangled exactly like a liquid's, but frozen in place. This is an amorphous solid, and the key to making one is cooling speed.
The same atoms in three states. Liquid: disordered and mobile. Crystal: cooled slowly, atoms find their ordered lattice positions. Glass: cooled fast, atoms freeze mid-disorder — a liquid caught in a snapshot.
Cool a liquid slowly and its atoms have time to line up into the low-energy crystalline arrangement. Cool it fast and viscosity skyrockets before the atoms can organize: they lock into place still disordered. That's glass — not a liquid flowing slowly, but a snapshot of a liquid that stopped moving. Picture a packed subway car when the power cuts out: everyone frozen exactly where they stood.
The temperature where this arrest happens is the glass transition temperature (Tg) — around 550 °C for window glass.
"The deepest unsolved problem in solid state theory"
Physicists still argue about what the glass transition fundamentally is. Nobel laureate Philip Anderson wrote in Science in 1995:
"The deepest and most interesting unsolved problem in solid state theory is probably the theory of the nature of glass and the glass transition."
Ordinary phase transitions are crisp: water freezes at 0 °C with a density jump and latent heat. The glass transition has no such sharp moment — viscosity just climbs until the material is de facto solid. Worse, the transition temperature itself depends on how fast you cool: quench quickly and the glass forms at a higher temperature than if you cool slowly. A "transition point" that moves with the experiment is deeply unsettling to theorists, and three decades after Anderson's remark there's still no consensus on whether it's a true phase transition at all.
So where does the "between solid and liquid" framing come from?
Pop-science videos love to call glass "neither solid nor liquid," and here's the twist: that framing traces back to Zanotto himself — the man who debunked the flowing-windows myth. In 2017, he and John Mauro (Penn State) proposed the modern definition of glass (Journal of Non-Crystalline Solids):
Glass is a nonequilibrium, non-crystalline state of matter that appears solid on a short time scale but continuously relaxes towards the liquid state; its ultimate fate, in the limit of infinite time, is to crystallize.
Thermodynamically, glass is not at equilibrium — it is creeping, infinitely slowly, toward lower-energy states. So "between solid and liquid" isn't baseless. But the framing becomes a myth the moment you drop the conditional clause: the "relaxation" in that definition is the same process Zanotto computed at 10²³-plus years. Formally, over infinite time: a relaxing nonequilibrium state. Practically, over any human timescale: a perfect solid. The same scientist wrote both papers, and both are true at once. Trouble starts only when someone compresses "glass would behave as a liquid if you waited trillions of times the age of the universe" into "glass actually flows." (The standard theory review of the field — Berthier & Biroli, Reviews of Modern Physics 2011 — likewise treats glass as a nonequilibrium disordered solid with liquid structure. Nobody in the field thinks your windows are moving.)
Amorphous solids are everywhere
| Object | Amorphous material | Why amorphous |
|---|---|---|
| Smartphone screens | Chemically strengthened aluminosilicate glass | Transparent; ion exchange hardens the surface |
| Obsidian | Natural volcanic glass (~70–75% SiO₂) | Lava cooled too fast to crystallize |
| Golf club heads (late 1990s), foldable hinges (in development) | Metallic glass (amorphous alloys) | No grain boundaries — strong, highly elastic |
| Some solar cells | Amorphous silicon | Cheap, thin, large-area deposition |
| Cotton candy | Amorphous sugar | Molten sugar spun into instantly-cooled fibers — sugar fiberglass |
Metals resist glass-forming strongly — classical routes needed cooling rates around a million degrees per second, though modern bulk metallic glasses achieve it far more gently through alloy design. The payoff is a metal with no grain boundaries: stronger and springier than its crystalline counterpart.
What stayed with me
I work in the semiconductor industry, and the amorphous state is quietly one of its heroes: the thermally grown SiO₂ that long served as the transistor's gate insulator is amorphous. A crystalline film would have grain boundaries, and grain boundaries are leakage paths; the disordered version is the uniform, reliable one. Order isn't always the desirable state — there are places where disorder is the performance. And while the glass transition remains an open problem, we've been running civilization on the unsolved material the whole time.
References
- E. D. Zanotto, "Do cathedral glasses flow?," Am. J. Phys. 66, 392 (1998). DOI: 10.1119/1.19026
- E. D. Zanotto & P. K. Gupta, "Do cathedral glasses flow? — Additional remarks," Am. J. Phys. 67, 260 (1999). DOI: 10.1119/1.19236
- O. Gulbiten, J. C. Mauro, X. Guo, O. N. Boratav, "Viscous flow of medieval cathedral glass," J. Am. Ceram. Soc. 101, 5 (2018). DOI: 10.1111/jace.15092
- P. W. Anderson, "Through the Glass Lightly," Science 267, 1615 (1995)
- E. D. Zanotto & J. C. Mauro, "The glassy state of matter: Its definition and ultimate fate," J. Non-Cryst. Solids 471, 490 (2017). DOI: 10.1016/j.jnoncrysol.2017.05.019
- L. Berthier & G. Biroli, "Theoretical perspective on the glass transition and amorphous materials," Rev. Mod. Phys. 83, 587 (2011)
Part of an ongoing series on solid-state physics in everyday life. The Korean version of this post is on my Naver blog.
댓글 없음:
댓글 쓰기