Cold Enough, Helium Slips Through Cracks and Climbs Out of Its Cup

Cool helium far enough and it slips through gaps no ordinary liquid can pass, and creeps up and over the rim of its container. Here's what was actually measured, and the model that makes sense of it.

Most liquids are well behaved. Pour coffee into a cup and it stays in the cup. Stir it, come back later, and it's stopped swirling, because the liquid rubs against itself and the cup until the motion dies away.

Liquid helium, cooled far enough, stops playing by those rules. It can dribble through cracks so fine that an ordinary liquid can't get in at all. It can sit in an open dish and then, quite calmly, climb up the inside wall, over the rim and down the outside, until the dish is empty.

Some of that has been measured directly. Some of it is a way of picturing what was measured. It's worth keeping the two apart.

A sharp line in the cold

Helium is already odd before any of this starts. It stays liquid all the way down towards absolute zero, the coldest possible temperature, where most other substances would long since have frozen solid.

Keep cooling that liquid and, at about two degrees above absolute zero, something abrupt happens. Physicists call it the lambda transition. The name comes from a graph: plot how much heat the liquid soaks up as it cools, and you get a tall, narrow spike at that temperature, shaped like the Greek letter lambda.

Above the line, the liquid is called helium I, and it behaves more or less like any other liquid. Below it, it's helium II, the best-known example of a superfluid: a liquid whose odd behaviour is governed by quantum mechanics, the physics of the very small, showing up at a scale you can see.

You can watch the switch happen. As helium I is cooled towards the line, it boils violently. The moment it crosses, the boiling stops dead. Helium II carries heat more than a million times better than helium I does, better than any metal, so the hot spots that would make bubbles simply can't form.

No friction, measured

Viscosity is a liquid's internal friction, its thickness. Honey has a lot; water has a little. One standard way to measure it is to push the liquid through a very thin tube: the thicker the liquid, the wider the tube needs to be before anything gets through.

Helium II goes through tubes and channels far thinner than a thousandth of a millimetre, and in those experiments it behaves as though it has no viscosity at all. Force it through a very fine tube and there's no measurable drop in pressure from one end to the other. In an ordinary liquid, that drop is exactly what the friction costs you. Here it simply isn't there.

That flow was measured in 1938, by pushing the cold helium through narrow gaps and thin glass tubes, and it was so free of drag that the word "superfluid" was coined for it, by analogy with superconductors, which carry electric current without resistance.

Then the same liquid drags

Hang a disc on a thin wire, set it twisting back and forth in helium II, and watch the swings. They die away. The liquid is dragging on the disc. Work out the viscosity from how fast the swinging fades, and you get about one-tenth of the viscosity of air. Small, but definitely not zero.

A spinning cylinder tells the same story. Turn a cylinder inside the liquid and a small paddle near its centre will show whether the liquid is being dragged round. In helium II, it is, a little.

One measurement says no friction. Another, in the same liquid, says some. No ordinary liquid does that, and it was one of several puzzles that piled up in the late 1930s.

The way out is called the two-fluid model. It treats helium II as if it were two liquids flowing through each other. The superfluid part has no viscosity at all and can slip without loss through the narrowest channels. The normal part has ordinary viscosity and drags on things the way any liquid would. Push the helium through a fine tube and only the superfluid part gets through, so you see no friction. Twist a disc in the bulk liquid and the normal part grabs it, so you see some.

It's a model, not a recipe. Nobody can pour the superfluid part off into its own bottle. Helium II is a single quantum liquid, and fully describing it takes quantum mechanics. What the two-fluid picture does is predict what each experiment will show. It also tracks temperature: right at the lambda line the liquid behaves as if it's all normal, near absolute zero as if it's all superfluid, and in between the superfluid share grows as it gets colder. You can measure that share by checking how much of the liquid clings to a disc swinging back and forth in it.

The film that climbs

Now for the image everyone remembers.

Put helium II in a cup. Any liquid coats the sides of its container a little, because its atoms are slightly attracted to the surface. In an ordinary liquid, internal friction stops that coating from spreading far. In helium II there's no friction to stop it.

So a very thin film forms, around thirty millionths of a millimetre thick. This is the Rollin film, a creeping skin of superfluid. It spreads up the inside wall of the cup, over the rim and down the outside. At the bottom it gathers into a drop, which falls. Another drop forms, and another, until the cup is empty.

It works in reverse, too. Lower an empty test tube partway into a bath of helium II and it slowly fills, as a film flows up the outside of the tube and over the lip.

The film is looking for its own level. Set two containers side by side, joined by a surface it can creep along, and before long the levels in the two will match. Leave the outer container unsealed and the helium will creep out of that as well.

How fast? Just below the lambda line, the creep is slow. At lower temperatures it can reach about thirty-five centimetres a second.

None of this breaks physics. What does the lifting is a real pull between the helium and the wall, which in this thin film wins out over both gravity and friction. Ordinary liquids feel that pull too. Their own internal friction is what keeps them from following it very far.

The cup that empties itself

So picture it one more time: a cup of something that looks like any other clear liquid, sitting quite still on the bench. No one tips it. No one touches it. And drop by drop, from a skin too thin to see, it climbs out over the rim and leaves.

Every liquid feels that pull towards the wall. Helium, cold enough, is the one that has nothing holding it back.

SOURCES

1. Russell J. Donnelly, "The two-fluid theory and second sound in liquid helium," Physics Today (1 Oct 2009). https://physicstoday.aip.org/features/the-two-fluid-theory-and-second-sound-in-liquid-helium (accessed 24 Oct 2026)

2. "52.11: Superfluids," Physics LibreTexts (General Physics I: Classical Mechanics, Prince George's Community College). https://phys.libretexts.org/Courses/Prince_Georges_Community_College/General_Physics_I%3A_Classical_Mechanics/52%3A_Fluid_Dynamics/52.11%3A_Superfluids (accessed 24 Oct 2026)

3. "Rollin film," Wikipedia. https://en.wikipedia.org/wiki/Rollin_film (accessed 24 Oct 2026)

4. JR Minkel, "Strange but True: Superfluid Helium Can Climb Walls," Scientific American (20 Feb 2009). https://www.scientificamerican.com/article/superfluid-can-climb-walls/ (accessed 24 Oct 2026)