The Snap Popper: How a Bistable Dome Launches Itself

The setup: two shapes hiding in one shell

A snap popper is a shallow rubber or plastic spherical cap — a shell curved in two directions at once. Unlike a flat sheet, a doubly-curved shell can sit in two distinct low-energy shapes: the as-moulded convex dome, and an everted (turned-inside-out) dome. Both are equilibria; the shell is bistable. Plot its elastic energy against the apex deflection w and you get a double well:

U(w) ∝ (w² − w₀²)² → minima at w = ±w₀, an energy barrier at w = 0.

The natural shape sits in the deep well, the everted shape in a shallower, higher-energy well. Between them the stiffness turns negative (d²U/dw² < 0 for |w| < w₀/√3): push the apex far enough and it resists less the more you push. That negative-stiffness region is the seed of the snap.

Why it stays inverted: geometry beats material

What decides whether the inverted shape is a genuine second minimum — rather than springing straight back? Geometry, not material. The cap must be deep enough relative to its wall thickness. The clean benchmark is the Belleville (coned disc) spring, patented by Julien Belleville in 1867: its load–deflection curve develops a snap-capable, negative-stiffness region only when the cone height satisfies h₀/t > √2 ≈ 1.41. Below that ratio it is just a stiff, monotonic spring.

A popper follows the same rule: a dome several times deeper than its ~1–2 mm wall is bistable, while a shallow dome merely 'oil-cans' back. Theodore von Kármán and Hsue-Shen Tsien analysed exactly this snap-through of spherical shells in 1939, explaining why real shells buckle far below the classical linear prediction — the shell simply falls into a nearby lower-energy buckled state.

The snap: a limit-point catastrophe

The fast event is a limit-point (fold) instability, not a bifurcation. As the everted shell is nudged toward the barrier, its force–deflection path reaches a fold where dF/dw = 0; the quasi-static branch turns back on itself, but the shell cannot follow an unstable path, so it jumps dynamically across to the far stable branch. That leap is snap-through.

Why so fast? Past the fold there is no equilibrium left to hold it, so the release is limited only by the shell's own inertia. The timescale is the elastic-wave crossing time:

τ ≈ √(ρa²/E) ≈ √(1100·0.015²/3×10⁶) ≈ 0.3 ms

(with ρ ≈ 1100 kg/m³, a ≈ 15 mm, E ≈ 3 MPa). Real poppers snap in well under a millisecond — hence the audible pop as the shell slaps the surface and rings like a struck bell.

From snap to jump: launching itself

Snapping alone would not launch it — the geometry does. Sitting everted, the popper rests on its rim. When it snaps back to the convex shape, its centre of mass accelerates upward while the rim drives down against the table; the table's reaction delivers an impulse J = ∫F dt = m·v that throws the whole toy off the ground.

Worked example. Your hand does ~100 mJ of work inverting a m ≈ 3 g popper. If a fraction η ≈ 0.4 reaches bulk motion, KE ≈ 40 mJ → v = √(2·KE/m) ≈ 5 m/s → height h = v²/2g ≈ 1.3 m, dozens of times its own size. Releasing ~50 mJ in ~0.3 ms is a peak power near 150 W — a power amplification of ~10³ over the slow, ~0.1 W loading by hand. It is the same trick a click beetle or a Venus flytrap plays: load slowly, release in a flash.

The telltale pause: viscoelastic delay

Watch a popper closely and it does not fire the instant you set it down — it waits, sometimes a second or more, then leaps. That delay is a material effect, not energy 'charging up.' Rubber is viscoelastic: under the everted strain it slowly relaxes, and as stress relaxes the barrier pinning the metastable state shrinks. When the barrier vanishes the everted shape loses stability, and the inertial snap fires in ~1 ms.

This temporary stability has a name — pseudo-bistability — and the waiting time is governed by viscoelastic creep (relaxation time ~0.1–2 s, strongly temperature-dependent — warm the popper and it jumps sooner). Brinkmeyer and Santer identified this creep-driven delayed snap-back in 2012, and Gomez, Moulton and Vella later modelled its full dynamics. The snap itself, by contrast, is inertia-limited: Pandey, Moulton, Vella and Holmes showed in 2014 that its timescale is that of an elastic wave crossing the shell. Two clocks, orders of magnitude apart, living in one toy.

Where engineers use — and fight — dome snap-through

Dome snap-through appears wherever engineers want a fast, definite change of state. The snap-action disc thermostat (the 'Klixon,' Spencer Thermostat Co., 1930s) uses a bimetal dome that snaps at a set temperature to make or break contacts crisply — the sudden separation avoids a slow, arc-sustaining contact and is used the same way in circuit breakers. The identical physics gives keypads their tactile metal dome switches (~1–5 N click), gives Belleville stacks their over-centre latching, and drives bistable jumping robots and MEMS actuators.

It is also fought. Unwanted snap-through is the oil-canning of car roofs and appliance panels, and the high strain concentrated at a snapping dome's rim drives fatigue cracking — tactile domes are rated for a finite number of actuations (often several million) before the click fades and dies.

Snap-through (bistable dome) vs Euler buckling (column) — two elastic instabilities engineers must tell apart

Property	Snap-through (bistable dome)	Euler buckling (column)
Instability type	Limit point (fold) — path turns back on itself	Bifurcation (pitchfork) — a new path branches off
Governing relation	Double-well energy U ∝ (w²−w₀²)²	P_cr = π²EI/(KL)²
Stable equilibria	Two (natural + everted)	One below P_cr; symmetric pair after
Trigger	Deflection pushed past the fold / barrier	Compressive load reaching P_cr
Dynamics	Sudden dynamic jump to a distant state	Gradual growth of lateral deflection (ideal)
Reversible?	Yes — snaps back and forth, re-storing energy	Elastic if below yield, but often ends in collapse
Everyday example	Popper toy, snap-dome switch, oil-canning panel	Slender strut, plastic ruler pushed end-on