Why Ice Is Slippery: The Quasi-Liquid Layer Explained

The most widely taught explanation for why ice is slippery is quantitatively wrong by a factor of 100

The pressure-melting hypothesis — the idea that the weight of a skater pressing down on a thin blade melts a thin layer of ice, creating a lubricating water film — appears in physics textbooks, classroom demonstrations, and popular science writing. It is also, when actually calculated, far too small an effect to account for ice slipperiness at temperatures below -1°C or -2°C. The pressure exerted by a typical skate blade is approximately 5–10 MPa. The freezing point of ice decreases by only 0.0075°C per atmosphere (0.1 MPa). A 10 MPa pressure reduces the melting point by approximately 0.8°C. At -5°C — a normal skating rink temperature — pressure melting is nearly irrelevant. Ice is already slippery before any weight is applied. The actual explanation involves a phenomenon first observed by Michael Faraday in 1842 and confirmed by molecular-level measurements only in the 21st century.

Faraday's 1842 discovery: surface melting

Michael Faraday first reported evidence of a liquid-like layer on the surface of ice at temperatures below the bulk melting point in 1842. He demonstrated that two pieces of ice held in contact would freeze together (regelation), even when pressed together at temperatures well below 0°C — an observation he attributed to a liquid water film on each surface that froze when the two surfaces met and merged. He called this surface condition "quasi-fluid."

The scientific community debated the phenomenon for more than a century without consensus. Direct experimental confirmation proved technically challenging because the layer is extremely thin — nanometers to tens of nanometers — and ice surfaces are difficult to study in situ without disturbing them.

• James Thomson (Lord Kelvin's brother) proposed the pressure-melting alternative in 1850 as a competing explanation for ice regelation
• The pressure-melting explanation gained dominance in textbooks partly because it was mathematically tractable and physically intuitive
• Faraday's surface-melting concept, later formalized as "premelting" or the quasi-liquid layer (QLL), was revived by experimental evidence in the 1980s–2000s

The quasi-liquid layer (QLL): what it is and why it exists

Surface molecules of a crystalline solid, including ice, are in a fundamentally different physical environment than interior molecules — they have fewer neighboring molecules bonding to them, leaving unsatisfied bonding capacity. For most solids, this surface energy asymmetry causes surface molecules to remain in the crystalline structure because the energy cost of leaving the crystal lattice exceeds thermal energy available at temperatures below melting.

For ice, however, the hydrogen bonding network of water is uniquely flexible. Surface water molecules at the ice-air interface can partially satisfy their bonding requirements by rotating and rearranging into a disordered, liquid-like configuration at temperatures significantly below 0°C. This disordered layer — the QLL — sits atop the crystalline ice surface and is neither bulk ice nor bulk liquid water, but a structurally intermediate phase.

Modern neutron and X-ray studies

Direct measurement of the QLL required techniques capable of probing a nanometer-scale surface layer at subzero temperatures without disturbing it. Several approaches have succeeded:

Sum frequency generation (SFG) spectroscopy: An interface-specific laser technique sensitive to the molecular ordering at the ice surface. Studies by Shen (UC Berkeley) and collaborators confirmed disordered OH stretching modes at the ice surface at temperatures down to -35°C, consistent with the QLL model.

Neutron reflectometry: Using specular neutron reflection, Beaglehole and Nason (1980) and later Suter et al. (2006) demonstrated a density gradient at the ice surface consistent with a disordered near-surface layer.

X-ray diffraction: Synchrotron-based grazing-incidence X-ray diffraction studies by Dosch, Lied, and Bilgram in the 1990s directly measured the surface structure of ice, showing disordered layers at the surface at temperatures below 0°C extending to ~30 nm thickness at -1°C.

• The QLL thickness diverges as temperature approaches 0°C — it grows thinner at colder temperatures, consistent with thermodynamic surface melting models
• The QLL is present at ice-air, ice-solid, and ice-liquid interfaces with different thicknesses depending on surface chemistry
• The QLL is responsible for regelation (ice refreezing behind a wire pulled through ice under tension) — not pressure melting

Why pressure melting still fails at temperatures relevant to sports

Quantitative analysis of pressure-melting demonstrates its insufficiency as a primary explanation for ice friction:

• A 75 kg skater on a blade 30 cm × 3 mm produces approximately 8.3 MPa contact pressure
• This reduces the melting point by approximately 0.6°C — negligible at -5°C (skating rinks) or -20°C (outdoor winter)
• Ice is slippery even for a stationary object resting on it with zero applied pressure differential — pure pressure melting cannot explain static slipperiness
• Ice at -30°C is noticeably less slippery than at -5°C despite any frictional heat, consistent with QLL thickness reduction at lower temperatures

Frictional heating from blade-ice contact does contribute to lubrication during active skating — but this is a dynamic, speed-dependent effect, distinct from the static slipperiness of an ice surface. The QLL explains static slipperiness; frictional heating amplifies dynamic lubrication during motion. Both are real. Only one appears in textbooks.