Physicists built the first 'ideal glass' on a computer, and it behaves like a crystal

Physical Review Letters, 2026

Every window, phone screen, and drinking glass presents the same puzzle. Glass is rigid like a crystal, but at the molecular level, it looks like a frozen liquid. Its molecules are jammed together with no repeating pattern, no lattice, no order. How does something so disordered become so stable?

In 1948, Princeton chemist Walter Kauzmann predicted that if you could cool glass slowly enough, to cold enough temperatures, you would reach an ideal state where the amorphous molecules are packed as tightly as physically possible. At that point, the glass would behave like a crystal despite having no crystalline structure. For nearly eight decades, nobody could create this ideal glass, either physically or mathematically.

A team led by physicist Eric Corwin at the University of Oregon has now done it on a computer. Their results, published in Physical Review Letters, confirm Kauzmann's prediction and could eventually inform the design of new materials with enhanced properties.

Building density without structure

The challenge was to pack molecules as tightly as a crystal without introducing any repeating pattern. In a two-dimensional crystal, each disk-shaped molecule is surrounded by six neighbors, all touching, like cells in a honeycomb. The arrangement is perfectly regular and perfectly dense.

Corwin's team started from that crystalline template and developed a method to preserve the property of every disk being perfectly packed against its neighbors while entirely removing the repeating structure. The result was an arrangement that is as dense as a crystal but completely amorphous.

They confirmed the material's ideal glass status by testing how it responded to pressure, bending, melting, and other physical stresses. In every mechanical test, the modeled material behaved identically to a crystal. Corwin described the conclusion directly: their structure mechanically behaves identically to a crystal, even though it is completely amorphous.

Why this matters beyond theory

The definition of glass in physics is broader than the everyday meaning. Scientifically, anything solid made of amorphous molecular arrangements qualifies: many plastics, metallic glasses, and even some biological materials. Understanding the ideal glass state, the theoretical limit of how tightly amorphous molecules can pack, has implications for all of these materials.

A material approaching the ideal glass state would have enhanced properties: higher melting point, greater resistance to stress, potentially improved flexibility. These are properties that matter for manufacturing. Metallic glasses, for instance, are already valued for their strength and resistance to deformation, but they are difficult to produce because they must be cooled extremely rapidly from liquid to solid.

If researchers could develop a better understanding of the glass transition, the process by which a liquid becomes an amorphous solid, they might design metal alloys that can be cooled much more slowly while still forming a glassy structure. Slower cooling means simpler manufacturing.

From two dimensions to three

The current work is two-dimensional, using disk-shaped molecules on a flat plane. Real materials exist in three dimensions. Corwin and his team are working to extend their method to three-dimensional space, which is mathematically more challenging but necessary for practical application.

The team included doctoral student Viola Bolton-Lum and former doctoral students R. Cameron Dennis and Peter Morse, all from Corwin's lab. The research was supported by the Simons Foundation.

Limitations

The ideal glass exists only as a mathematical construction on a computer. It has not been created physically, and it remains unclear whether physical processes could ever produce a material approaching this ideal state. The gap between a perfect mathematical model and the messy reality of atoms cooling in a furnace is substantial.

The two-dimensional restriction is also significant. Many properties of materials change fundamentally when moving from two to three dimensions. Whether the mechanical equivalence between ideal glass and crystal holds in 3D is an open question.

Still, confirming a 78-year-old theoretical prediction is meaningful in its own right. It validates a conceptual framework that physicists have used for decades to think about glassy materials, and it provides a target for experimental work aimed at approaching the ideal glass state in real materials.