3D Cola and Mentos Eruption: Bubble Growth and Foam Jet

Drop Mentos into a bottle of cola: bubbles nucleate on the candy, grow by diffusion, fill the bottle with foam and force it out of the neck.

Switch to 2D

Presets

Playback

Parameters

Camera

Layers

Quality

Model check

The science

What is modelled

Cola holds CO2 dissolved far above what the open air can keep. Nothing happens until gas has somewhere to form. A Mentos candy has a rough, wettable surface covered in tiny pits, so bubbles nucleate on it in huge numbers. Each bubble then grows by pulling CO2 out of the surrounding liquid by diffusion. The page adds up the gas made by every bubble in the foam, lets the foam fill the bottle, and forces it out of the neck as a jet.

Equations

  • Dissolved gas: c0 = Vc / 22.414 mol/L, where Vc is the carbonation in "volumes" (gas volume at 0 C and 1 atm per volume of drink). Open to air with CO2 at the bubble surface, the liquid can keep only ceq = kH(T) p with kH = 0.034 exp[2400 (1/T − 1/298.15)] mol/(L atm). The supersaturation Δc = c − ceq is the driving force and is larger in a cold drink... but a cold drink also diffuses slowly.
  • Bubble growth (Epstein-Plesset, surface tension and the early transient neglected): dR/dt = D Δc / (ρg R), so R2 = R02 + 2 D Δc t / ρg, with ρg = p/(Rgas T). The diffusion coefficient scales as D ∝ T / η(T) from 1.91×10-9 m2/s at 20 C.
  • Gas from one bubble: dn/dt = 4 π R D Δc. New bubbles appear at dN/dt = Jc · (candies in the liquid) · (surface factor) · (Δc/Δcref) e−age/τ and leave the foam after 0.9 s.
  • The bottle holds 2.2 L. Gas first takes up the free space; once liquid plus gas fill the bottle, the foam must leave through the neck: v = Qgas / (α Aneck) with a foam gas fraction α = 0.6.
  • Droplets rise against gravity and quadratic drag: dv/dt = −g − KD |v| v, apex h = ln(1 + KD v2/g) / (2 KD).

Checks run in the page

  • A bubble is grown with a small time step and compared with the closed-form radius.
  • The CO2 account (dissolved, released as gas, carried out in the liquid) is closed over the whole run.
  • The gas volume of a complete release is computed from the ideal-gas law and compared with the "volumes of CO2" definition; Henry constants at 25 C and 4 C are shown.
  • A foam drop leaving the neck at the peak speed is integrated with RK4 and its apex compared with the closed form.

Honest simplifications

  • The number of bubbles a candy produces per second, Jc, is not derived. It is the one fitted constant, chosen so that four candies in 1.7 L of diet cola at 20 C give a jet of about 2 m, in the range people report. The relative effects (temperature, number of candies, neck width, carbonation) come from the physics; the absolute heights are only as good as that calibration.
  • Surface factors for diet, sugared and plain soda water are qualitative: lower surface tension and additives such as sweeteners and benzoate make nucleation easier, but the numbers here are illustrative.
  • Bubbles are treated as non-interacting spheres in a well-mixed liquid; coalescence, foam drainage, the liquid column and turbulence in the neck are ignored. The foam gas fraction and droplet drag constant are single typical values. Carbonation is taken as an input at the chosen temperature, not recomputed from bottle pressure.
  • This is a teaching model of why the eruption happens, not a prediction. Do not try it with anything but a bottle of cola outdoors: the jet is messy and the bottle must be open and pointing up.

What to try

  • Compare ice-cold and warm cola: the temperature graph shows how strongly the peak height changes.
  • Use one candy, then ten; widen the neck and see why a narrow opening makes the jet.
  • Switch to plain soda water or flat cola.
Loading model...
Loading Blender model...
Jet height against time
Dissolved and released CO2
Peak height against drink temperature
Made with 🟠 Blender · Three.js