Why is ice slippery?

The textbook answer

In 1886, a young engineer named John Joly calculated that a skate presses on ice with 466 atmospheres, lowering its melting point to −3.5 °C¹. This idea, called pressure melting, stayed the main explanation for nearly a century¹.

Why the numbers don't add up

Pressure does lower the melting point of ice, but only by about 0.074 °C for every megapascal². To melt ice at −10 °C, you would need about 110 megapascals². Yet ice squeezed quickly from one direction already fails at around 10 megapascals². Most researchers doubt pressure melting for this reason, though the same review notes that local pressures under a blade can climb much higher². As physicist Daniel Bonn put it, you would need 10 elephants on a single skate to get a decent amount of melting¹¹.

To melt ice with a boot, you would have to squeeze all of your weight onto a spot about the size of a poppy seed¹⁰. Still, skating works at temperatures as low as −30 °C¹, and ice can be slippery even when you stand still on it¹.

Suspect 1: ice wears a wet skin

In 1859, Michael Faraday argued that ice carries a thin liquid layer near its melting point¹. Later microscope measurements put this liquid-like layer at no more than 12 nanometres thick at −24 °C, and 70 nanometres at −0.7 °C¹. In 2024, a team at Peking University imaged the ice surface atom by atom⁶. It slowly turned disordered above 120 kelvin, about −153 °C, starting at the boundaries between patches of differently stacked ice⁶.

Suspect 2: heat from rubbing

In 1939, Frank Bowden and T. P. Hughes suggested that friction makes heat that melts a thin film¹. A 2026 computer study from Oslo and Jülich backs a version of this: sliding warms the contact and can make the existing surface film about ten times thicker, without melting the ice⁹.

Suspect 3: rolling molecules and gooey water

In 2018, a team from Amsterdam and Mainz slid a steel ball over ice at temperatures from −100 °C to 0 °C⁴. Friction was high at −100 °C and lowest at about −7 °C⁴, and very low friction appeared only in the range typical for ice skating³. The team suggests that loose water molecules on top roll over the surface “like tiny spheres”, rather than a layer of liquid water doing the work⁴.

A 2019 study found a different picture: the film under a sliding bead was hundreds of nanometres thick and up to 100 times more viscous than water⁵.

Suspect 4: the shoe scrambles the ice

A 2025 computer simulation from Saarland University suggests something stranger. Sliding may scramble the top of the ice into a disordered, liquid-like layer without truly melting it⁷. The team points to electrical interactions between molecular dipoles in the ice and in the other surface, such as a shoe sole, not pressure or friction⁸. In their model, a film forms even near absolute zero, though there it would be gooier than honey⁸.