Ice presents a small everyday mystery. A cold pavement can be dangerously slippery even when the air temperature is far below freezing, while many other crystalline solids are perfectly easy to walk on. A steel skate can glide across a rink with remarkably little resistance. For generations, the standard explanation sounded simple: pressure from a person's weight lowers the melting point, producing a microscopic film of water that acts as a lubricant.

It is a satisfying story. It is also incomplete.

Modern experiments show that the surface of ice is an unusual molecular environment even before a boot or skate presses against it. Water molecules at the boundary are less rigidly organized than molecules deeper inside the crystal. Under many conditions, ice develops what physicists call a quasi-liquid layer through a phenomenon known as surface premelting.

But even that is not the whole answer. The slipperiness of ice emerges from an interplay of molecular mobility, temperature, pressure, frictional heating, surface roughness and the mechanical properties of the ice itself. There is no single magic film that explains every frozen sidewalk and skating rink.

The old pressure-melting explanation

The pressure theory has a respectable scientific history. Water is unusual because ordinary ice occupies more volume than liquid water. Increasing pressure can therefore lower ice's melting temperature. This is real physics and helps explain phenomena such as regelation, in which ice can melt under pressure and refreeze when that pressure is removed.

The problem is magnitude. The pressures generated by an ordinary person or even a skate cannot, by themselves, account for easy sliding over ice across the wide range of temperatures at which it occurs.

Modern friction measurements have made the limitation especially clear. Experiments reported in Physical Review X examined ice friction while independently varying temperature, sliding speed and contact pressure. The researchers found that increasing local pressure can actually hinder the mobility of surface molecules and eventually increase friction rather than simply creating a better lubricating water film. citeturn0search7

So pressure matters, but the familiar picture of body weight simply squeezing a layer of liquid water out of solid ice is not a sufficient explanation.

The surface of ice is not quite like its interior

To understand why, we have to zoom down to individual molecules.

Inside an ice crystal, water molecules participate in an ordered network held together by hydrogen bonds. A molecule buried within the solid is surrounded by other molecules in a relatively stable crystalline arrangement.

At the surface, that environment ends abruptly. Molecules no longer have identical neighbors in every direction. The hydrogen-bond network is disrupted, allowing surface molecules to adopt structures and motions different from those in the bulk crystal.

The result is a surface that can become partially disordered before the entire block of ice reaches its melting point. This is surface premelting.

A major review in Nature Reviews Chemistry describes a quasi-liquid layer at the surface of ice that can exist well below the bulk melting temperature. The review traces scientific interest in the phenomenon back to Michael Faraday in the nineteenth century and emphasizes that the precise structure, thickness and behavior of this layer remain active research questions. citeturn0search0

Calling it “liquid water” can therefore be too simplistic. Depending on temperature and conditions, the surface can contain molecular arrangements intermediate between an orderly crystal and ordinary bulk liquid water.

Are the molecules permanently liquid below zero?

This is where a popular description needs some care. It is tempting to say that ice is always covered by a permanent film of liquid water, no matter how cold it gets. The scientific picture is more nuanced.

Experiments and simulations show that molecular disorder at the surface changes strongly with temperature. As ice approaches its melting point, the quasi-liquid region generally becomes more pronounced. A 2024 study in Communications Chemistry, for example, describes quasi-liquid layers as an important feature of ice surfaces and notes experimental reports of their formation at temperatures tens of kelvins below the melting point. citeturn0search3

At still lower temperatures, the language of a continuous liquid film becomes less useful. Researchers studying friction down to extremely low temperatures have instead emphasized the thermally activated mobility of molecules in the outermost surface layers. citeturn0search7

The key idea survives: the outer surface of ice is dynamically different from the rigid crystal beneath it. But “permanently fluid water layer” should not be interpreted as a conventional puddle a few molecules deep existing unchanged at every subzero temperature.

Loose surface molecules can behave like tiny bearings

One particularly revealing set of experiments challenged the assumption that sliding must first melt the ice.

Researchers combined friction experiments with molecular simulations and concluded that the high mobility of molecules at the ice surface can itself produce exceptionally low friction. Nature summarized the result with a vivid analogy: skaters can effectively ride over highly mobile surface molecules rather than relying entirely on pressure-generated melting. citeturn0search5

In the 2021 Physical Review X study, the researchers found that at temperatures well below melting, friction followed a temperature dependence consistent with thermally activated diffusion of surface molecules. As temperature rises, those molecules can move more easily. citeturn0search7

This helps explain a familiar experience: very cold ice often feels less slippery than ice closer to 0°C. The surface molecular dynamics are temperature-dependent.

Ice is therefore not simply a hard solid with accidental water sitting on top. Its own surface possesses unusual mobility.

Sliding can create another lubricating layer

There is an apparent complication. If ice already has mobile surface molecules, why have other experiments found actual meltwater underneath moving objects?

Because several mechanisms can operate simultaneously.

When an object slides across ice, friction generates heat at the contact. Near the melting point, that heat can produce a thin layer of interfacial meltwater. Researchers led by Lydéric Bocquet and Daniel Bonn's collaborators probed this layer at nanometer scales and found something surprising: the lubricating material was not behaving like ordinary bulk water.

Their Physical Review X study reported a very thin interfacial film with complex viscoelastic behavior. It could be dramatically more viscous than ordinary water, partly because microscopic ice fragments were present in the layer. citeturn0search9

A commentary in Nature described its properties as intermediate between liquid water and ice. citeturn0search1

This is one reason the question “Why is ice slippery?” has resisted a one-sentence answer for more than a century. The lubricating interface under a moving skate is not necessarily identical to the equilibrium quasi-liquid surface that exists before the skate arrives.

Why temperature changes everything

Anyone who has skated outdoors knows that all ice is not equally fast. Ice near freezing can feel wet and soft. Extremely cold ice can feel dry and surprisingly resistant.

Laboratory measurements confirm that friction varies substantially with temperature. In experiments spanning a very wide range, researchers found that molecular mobility plays a major role at low temperatures, while near the melting point the mechanical softness of ice introduces another effect: the slider can deform or plough into the surface, increasing resistance. citeturn0search7turn0search10

This produces a counterintuitive result. Warmer does not simply mean more liquid and therefore endlessly lower friction. If the ice becomes soft enough for an object to dig into it, mechanical deformation can increase friction.

There can therefore be a temperature range in which ice is especially slippery, bounded by different physical mechanisms on either side.

Why skates need both smoothness and sharp edges

A skate blade beautifully exploits this complicated physics.

The broad, smooth running surface can maintain relatively low friction as it moves over ice. Yet the edges of the blade can bite into the surface when the skater tilts the skate, generating the grip required to accelerate, turn and stop.

The same 2021 friction study pointed out that this design makes physical sense: lower contact pressure on a smooth part of the blade favors easy sliding, while high local pressure at a sharp edge can penetrate the ice and produce the friction needed for control. citeturn0search7

A perfectly frictionless rink would actually be useless. A skater needs low resistance in one direction and controllable grip in another.

Michael Faraday suspected something strange in the 1800s

The idea that ice possesses an unusual surface is much older than modern molecular simulations. Michael Faraday investigated the behavior of ice in the nineteenth century and proposed that its surface could have liquid-like characteristics even below the ordinary melting point.

For more than 160 years, scientists have debated how this surface should be described and how much it contributes to ice friction. Modern spectroscopy, microscopy, computer simulations and nanometer-scale force measurements have turned Faraday's intuition into a much richer molecular picture. citeturn0search0

Yet some details remain unsettled. The thickness and structure of premelted ice depend on temperature, impurities, crystal orientation, humidity and defects. Even the phrase “quasi-liquid layer” covers surfaces that can be heterogeneous rather than uniform. Some regions may be more liquid-like than others. citeturn0search0turn0search2

That uncertainty is not a weakness in the explanation. It is evidence that a phenomenon we encounter every winter involves remarkably subtle surface physics.

So why do we slip?

The most accurate short answer is not “because pressure melts the ice.” It is that the interface between an object and ice has unusually low friction because water molecules at the surface are exceptionally mobile, while sliding can further modify that interface through frictional heating, melting and deformation.

Pressure is one participant, not the whole story. Temperature controls molecular mobility and hardness. Speed controls frictional heating. Surface roughness affects contact. A skate, boot sole and car tire therefore experience ice differently.

The strangest part is that the slippery behavior begins at a scale we cannot see. The surface that looks perfectly solid is not simply the last rigid layer of a crystal. Its outer molecules inhabit a less ordered, more mobile world, hovering between the behaviors we normally label solid and liquid.

That is why a frozen lake can support a person while its uppermost molecular layers help send that same person sliding across it. Ice is slippery not because it stops being a solid, but because at its surface the boundary between solid and liquid is far less absolute than it appears.