Soap Film Minimal Surface

A soap film is two layers of soap molecules with a thin sheet of water between them. Creating film costs energy, and that energy is strictly proportional to how much film there is: E = 2γA, where A is the area and γ is the surface tension of the solution (about 25 mN/m for soapy water, against 72 mN/m for pure water — the surfactant is what makes films stable enough to survive).

The factor of two is there because a free film has two surfaces, one on each side. Since the constant 2γ does not depend on shape, minimizing energy and minimizing area are the same problem. A film left alone slides downhill in energy until it cannot go any lower — so a soap film is an analogue computer that solves a hard geometry problem by simply relaxing into the answer.

This simulator does the same thing numerically. The film is a triangle mesh whose boundary vertices are pinned to the wire. Every step, each free vertex moves down the gradient of the total triangle area — the exact derivative ∂A/∂P, not an approximation — so the number under the canvas falls until the surface stops changing.

The Belgian physicist Joseph Plateau spent years dipping bent wire frames into soapy water and recording what came out, publishing his results in 1873 in Statique expérimentale et théorique des liquides soumis aux seules forces moléculaires. He had gone completely blind thirty years earlier, in 1843, and did essentially all of his soap-film work without being able to see it — measuring the angles in a foam through a sighted assistant who described each film to him, and dictating every word of the book's two volumes. The story that his blindness came from staring at the midday sun for 25 seconds in an 1829 afterimage experiment is often repeated, but modern medical opinion attributes it to chronic uveitis instead.

His experiments made the mathematical question concrete: given any closed curve in space, is there always a surface of least area spanning it? Physically the answer was obviously yes — you could just dip the wire. Proving it was another matter, and the question became known as Plateau's problem.

At every point a surface has two principal curvatures, κ₁ and κ₂ — the sharpest and gentlest ways it bends. The mean curvature is their average, H = (κ₁ + κ₂)/2.

The Young–Laplace law says the pressure difference across a film is Δp = 4γH. A film with air at the same pressure on both sides must therefore have H = 0 everywhere. That is the defining property of a minimal surface: it curves up in one direction exactly as much as it curves down in the perpendicular one, which is why every point looks like a saddle and why a minimal surface never has a bulge or a dome.

This is also the difference between a film and a bubble. A bubble encloses trapped air at higher pressure, so it has constant nonzero H and becomes a sphere. Only an open film, free to equalize, goes to zero.

Two coaxial rings give the one classical case you can work out by hand. The film is a catenoid, the surface swept by rotating a catenary about an axis:

Discovered by Leonhard Euler in 1744, it was the first non-trivial minimal surface ever found, and apart from the plane it is the only minimal surface of revolution. The constant c is the radius of the narrowest point — the neck.

Pull the rings apart and the neck thins. But c · cosh(h / c) = R only has a solution while the rings are close enough together. Past a critical separation the equation has no root at all — no catenoid exists — and the film has nowhere to go but break. It snaps to the Goldschmidt solution: two flat discs, one capping each ring.

Drag the separation slider up past 1.33 and the simulation reproduces this on its own — the neck pinches through and the film flashes over to two discs. The collapse is not scripted; it falls out of the same area descent that runs everywhere else.

The helicoid — the surface swept by a line rotating steadily as it advances along an axis, like a spiral staircase ramp — was shown to be minimal by Jean Baptiste Meusnier in 1776. It and the catenoid are conjugate surfaces: there is a continuous, isometric deformation carrying one into the other.

Isometric means every step of the bending is minimal, and nothing stretches or tears — distances measured within the surface never change. A catenoid can be cut along one line and unrolled into a helicoid as though the surface were inextensible cloth. For a long time these two were the only minimal surfaces anyone knew besides the plane.

When several films meet — a foam, or the film that forms inside a wire cube — they cannot join arbitrarily. Plateau observed strict rules, proved rigorously by Jean Taylor in 1976:

• Smooth sheets: films are smooth surfaces of constant mean curvature (zero, when the pressure is equal across them).
• Threes at 120°: films only ever meet three at a time, along a curve called a Plateau border, at equal angles of exactly 120°. Four sheets meeting is unstable and instantly resolves into two triple junctions.
• Fours at 109.47°: Plateau borders themselves meet four at a time, at a point, at the tetrahedral angle arccos(−1/3) ≈ 109.47° — the same angle as the bonds around a carbon atom.

Both angles are forced by the same logic: three equal surface tensions pulling on a line balance only at 120°, and four equal tensions pulling on a point balance only in the tetrahedral arrangement. Dip a wire cube and you get a small square film in the middle with twelve films running out to the edges — every junction obeying these numbers.

The simulator here relaxes a single connected film at a time, so it covers the smooth single-surface cases rather than multi-film networks like the cube.

The German architect Frei Otto turned this physics into a design method. At his Institute for Lightweight Structures in Stuttgart he built form-finding models: bend a wire into the plan of a roof, dip it, photograph the film, and measure the result. The soap film hands you a shape in which the tension is uniform in every direction — precisely what a fabric or cable-net roof needs if it is to carry load without wrinkling or slack.

The method produced the sweeping cable-net canopy over the Munich Olympic Stadium for the 1972 Games, built with the architect Günter Behnisch and the engineer Fritz Leonhardt, and following on from Otto's German pavilion at Expo 67 in Montreal. The Munich roof also sat right at the moment soap-film measurement ran out of precision: resolving the final cable geometry required some of the earliest large-scale computational form-finding. Otto received the Pritzker Prize in 2015, announced days after his death.

The idea outlived the soap. Minimal and constant-tension surfaces underpin modern tensile architecture, and the same mathematics shows up in the gyroid and other triply periodic minimal surfaces found in block copolymers, in butterfly wing scales that produce structural colour, and in the membranes of cell organelles.