Ice Is Slippery. Nobody Can Yet Agree Why
Everyone learns that ice is slippery because a thin skin of water coats it. Why that skin forms is a different matter, and the argument has run for nearly two centuries without a winner.
Step onto a frozen puddle and your foot can be gone before you've even tried to move. No run-up, no slide, no warning. The ice was slick the instant you touched it.
Ask almost anyone why and you'll get an answer straight away: there's a thin layer of water on top, and you slide on it. That part is close to right. Scientists broadly agree that a slick, liquid-like layer on the surface is what makes ice so hard to grip.
The trouble is the next question. Where does that layer come from? Physicists have been arguing about it for nearly two centuries, and the case is still open. Think of it as a whodunit with three long-standing suspects, a newcomer, and no conviction.
Suspect one: your own weight
This is the version most of us were taught. You stand on ice, your weight presses down, that pressure melts a sliver of the surface, and you glide on the meltwater.
It has a real alibi. Ice is unusual: squeeze it hard enough and its melting point drops, so it can turn to water a little below its normal freezing temperature. In the middle of the nineteenth century the engineer James Thomson worked out how the melting point falls as pressure rises, and his brother William, later Lord Kelvin, confirmed it by experiment. The link to skating came later, when the engineer John Joly pointed out that a skate blade touches the ice over such a tiny area that the pressure under its edge is enormous.
Then someone did the sums. A calculation from the 1930s found that a skier would need to weigh thousands of kilograms to shift ice's melting point by a meaningful amount. The colder the ice, the more pressure you'd need, yet people ski and skate on ice far colder than pressure melting can account for. Worse, at those low temperatures the pressure needed to melt a film would squeeze it so thin that it would drag, rather than help you glide.
Pressure is a real effect. It just can't carry the everyday case on its own.
Suspect two: the heat of the slide
In 1939, Frank Bowden and T. P. Hughes put forward friction instead. Rub two things together and they warm up. Slide over ice and that warmth could melt a thin film exactly where it's needed.
They tested the idea in an ice cave dug high in the Swiss Alps. Materials that carry heat away well, like metal, met more friction than poor conductors like wood. Their reading: a good conductor drains the heat from the contact, leaving less to melt the ice. Later experiments with temperature sensors fixed to skate blades found the ice under the blade warming as speed rose, which fits.
It's a sensible mechanism, and it still turns up in textbooks. But remember that frozen puddle. Friction heat needs motion, and ice is slick the moment a foot lands, before anything has slid. In one modern lab test, a spinning piece of metal stood in for a skate blade, and the slipperiness didn't change with speed, even though frictional heating should rise as things go faster. And for a ski moving at a realistic pace, measurements suggest the heat produced isn't enough to melt the surface at all.
Suspect three: it was wet all along
The third idea flips the question. Perhaps ice doesn't need you to make it wet. Perhaps it already is.
This one goes back to Michael Faraday. In a lecture in June 1850 he described experiments on regelation, the way two ice cubes freeze together when they touch. His explanation was that a single piece of ice carries a film of water on its surface, and when two pieces meet, the film trapped between them freezes and binds them. James Thomson disagreed and blamed pressure. Faraday came back with new experiments, but pressure melting dominated the textbooks for close to a century anyway.
Faraday's idea is now called premelting: a liquid-like layer that forms on a solid below the temperature at which the bulk of it melts. The reason, roughly, is that molecules at the surface have fewer neighbours to hold on to. Deep inside, every water molecule is locked into a regular crystal pattern. At the top, the molecules are freer to jiggle, so the outer skin behaves less like a solid and more like a very thin liquid.
Most scientists now accept that this layer exists, at least close to the melting point. It handles the frozen-puddle problem neatly. But it has its own weak spot: in extremely cold conditions the premelted layer disappears, and ice stays slippery regardless.
The newcomer: scrambled, not melted
That gap is where the latest suspect walks in. In 2025 a team of materials scientists published computer simulations arguing that the slippery layer doesn't need melting of any kind.
Their models slid ice surfaces past each other at temperatures too low for melting. Water molecules are lopsided in charge, slightly positive at one end and slightly negative at the other, so the two surfaces pulled at each other and formed tiny welds. As the sliding went on, the welds snapped and re-formed, over and over, tearing molecules out of their places in the crystal. What was left was a disordered layer, structureless rather than neatly patterned, and it grew thicker the longer the sliding continued.
The authors call this cold, displacement-driven amorphisation. In plain words: the surface goes liquid-like because sliding shoves its molecules out of order, not because anything gets warm enough to melt. They argue it explains the slipperiness best in very cold conditions, exactly where the older ideas struggle most. Similar thinking had already been used to explain why polishers find some faces of a diamond easier to work than others: sliding pulls surface atoms out of their bonds and leaves a disordered layer behind.
No conviction yet
So who did it? Possibly more than one of them.
A set of simulations published a few years before the newcomer arrived modelled a smooth slider moving over ice across a range of below-freezing temperatures, and found work for all three of the old suspects. The ice carried a premelted film just a few molecules thick that behaved much like very cold liquid water and acted as a lubricant. Pressing down made the film grow, as the pressure story predicts. And at the colder end, where the natural film thinned, friction from a slider that clung to the ice melted enough to rebuild the layer within a tiny fraction of a second. On that reading, slipperiness at the smallest scales is a team effort, with the balance shifting as conditions change.
The newcomer has picked up some support, along with disagreement about when it applies. One objection is that it may only kick in at high sliding speeds, which is hard to settle because simulating slow sliding takes a prohibitive amount of computing power. Part of the problem is language, too. Some researchers talk about how mobile the surface molecules already are; others talk about molecules being knocked out of place by sliding. Whether those are two names for one process or two different processes is itself disputed, and when similar effects get different labels, they can look like rivals even where they overlap.
The case stays open
Next time the footpath ices over and someone goes down, everyone watching will agree on what happened: they slipped on the wet skin on top. What nobody can yet say for certain is what made that skin. A film the ice grows for itself, a melt squeezed or rubbed out of it, or molecules shoved out of their crystal by the very foot that's about to go flying.
SOURCES
Paulina Rowińska, "Why Is Ice Slippery? A New Hypothesis Slides Into the Chat," Quanta Magazine, 8 Dec 2025. https://www.quantamagazine.org/why-is-ice-slippery-a-new-hypothesis-slides-into-the-chat-20251208/ (accessed 24 Oct 2026)
Robert Rosenberg, "Why Is Ice Slippery?," Physics Today, 1 Dec 2005 (DOI 10.1063/1.2169444). https://physicstoday.aip.org/features/why-is-ice-slippery (accessed 24 Oct 2026)
A. Atila, S. V. Sukhomlinov and M. H. Müser, "Cold Self-Lubrication of Sliding Ice," Physical Review Letters 135, 066204 (2025). https://doi.org/10.1103/1plj-7p4z (accessed 24 Oct 2026, via web.archive.org and https://arxiv.org/abs/2402.10843)
L. Baran, P. Llombart, W. Rzysko and L. G. MacDowell, "Ice friction at the nanoscale," Proceedings of the National Academy of Sciences 119, e2209545119 (2022). https://www.pnas.org/doi/10.1073/pnas.2209545119 (accessed 24 Oct 2026, via web.archive.org)
Comments