Text from the interactive explanation

Effect of surface-active contaminants on bubble bursting yielding the smallest first-emitted droplet

Sergio Rodríguez-Aparicio, Alberto Cebrián-García, Emilio J. Vega, José María Montanero, María Guadalupe Cabezas

Bubble bursting · surfactants · J. Fluid Mech. 2026

A bubble pops at a liquid surface; its collapsing cavity fires a jet that breaks into droplets. Here one bubble is split down its axis. The left half bursts in clean liquid. The right half bursts in the same liquid with a trace of the surfactant Tween 80, so little that the surface tension barely changes. Watch its first droplet.

Red marks surfactant: molecules, a glow where they crowd, and arrows for the Marangoni pull. The circular loupes magnify 7×, following the cavity bottom, then the jet tip, then the first droplet.

Illustrative simulation. Cavity outlines are traced from the paper's high-speed frames (Fig. 5). The jet, the droplets and the surfactant motion follow the measured sizes, speeds and counts (Figs. 6, 9, 12, S2). Intermediate concentrations blend the two filmed cases. Readouts are the paper's measurements at the critical condition. It is not a flow solver.

01 — MEASUREDWhat the camera saw

These are the paper's own high-speed frames, shot at 12 800 frames per second against a back light. The clean interface is on the left; Tween 80 at c/ccmc = 0.0233 is on the right. The dark shape is the gas cavity left after the bubble's cap film breaks. For the first ~0.2 t0 (about 0.1 ms) both collapse identically. Then the contaminated cavity falls behind, and its bottom flips upward (the free-surface reversal) about 12 % later.

CLEAN

TWEEN 80

Time since the film ruptured, in units of the inertio-capillary time t0 = (ρRb3/σ)1/2 ≈ 0.50 ms. Orange: the frame after which each cavity has reversed. Frames from Figure 5(a) of the paper (La ≈ La*, Bo ≈ 0.01), reproduced under CC BY 4.0. We separated and re-oriented them from the published image strips. The hero's cavity outlines were traced from these frames.

02 — MEASUREDThe gauge sees nothing

A tensiometer reads the liquid's equilibrium surface tension σ. At the doses used here it barely moves: even at the highest Tween 80 concentration, σ falls only from 52.0 to about 51 mN/m (σ/σc = 0.98). The first droplet, photographed at 50 000 frames per second, tells a different story.

Make a guess. Surface tension drops by about 2 %. The first droplet's volume grows by…

—

Curve: the Langmuir equation of state the authors fitted to their Tween 80 measurements (Fig. 3a). Drag the marker or use the slider.

—

Figure 10 of the paper: the jet and its first-emitted droplet at La ≈ La* for four Tween 80 concentrations. Reproduced under CC BY 4.0; we cropped the panels and added the ring that marks the first droplet. At the highest dose the paper reports a radius 2.8 times larger and about twenty times the volume. The other ratios come from radii read off Fig. 9a.

03 — MECHANISMThe squeeze

Why does an invisible trace matter? Because the bubble concentrates it. As the cavity collapses, its surface flows toward the bottom while the bottom shrinks, so surfactant molecules pile up there. Surface tension depends on how crowded the surface is. That dependence stays almost flat until the surface is more than about half covered, then drops steeply. Squeeze a patch of surface and see.

—

The σ(Γ) equation of state for Tween 80 from the authors' fit (Fig. 3b): σ/σc = 1 + Ma ln(1 − Γ), with Ma = 0.076. Γ is surface coverage relative to a full monolayer. Shading marks Γ < 0.6, where σ hardly changes. The paper estimates ≈3× crowding at the cavity bottom (Γeq = 0.26) and ≈25× on the emitted droplets (Γeq = 0.033).

—

Cavity-bottom width w and upward bottom speed VB, read from Fig. 6. Zoomed frames at 420 000 frames per second from Fig. 5(b) (CC BY 4.0; separated and re-oriented), shown for the nearest filmed instant.

Sweep and shrink. Surface flow carries surfactant down the cavity walls while the cavity bottom contracts.

Crowding. Local coverage climbs into the steep part of σ(Γ). For Tween 80 at 0.0233 it passes 2.5 × Γeq.

Marangoni stress. The surface is pulled from crowded, low-σ regions toward cleaner, high-σ ones, which opposes the flow converging on the bottom.

A narrower, slower reversal. The bottom pinches narrower before it flips (w/Rb ≈ 0.04 vs 0.1) and rises more slowly (peak ≈ 23 vs 35 Rb/t0).

A wider, slower jet. The authors conclude that Marangoni stress is the main effect on the jet. It widens the jet and slows its surface, delaying pinch-off, and they propose that this lets more liquid flow into the first droplet.

04 — CONTEXTThe critical bubble

Not every bubble makes tiny droplets. The first droplet is smallest when the Laplace number La = ρRbσ/μ², which weighs surface tension against viscosity for a bubble of radius Rb, sits near La* ≈ 1100. There the collapsing capillary waves focus their energy best. In water that means bubbles of about 20 µm radius emitting droplets around 400 nm in radius at more than 115 m/s, which are hard to film. So the team used a more viscous 50/50 DMSO–water mixture, which reaches the same La* with bubbles a few hundred micrometres across.

Points: first-droplet radius measured at each Laplace number (read from Fig. 9a; clean points above La ≈ 2700 come from other DMSO–water mixtures). Line: our interpolation between them.

—

For this liquid (3.36 cSt, 1090 kg/m³, 52 mN/m), La ≈ 4.23 × Rb in µm. Adding Tween 80 lifts the whole curve but leaves the minimum near the same La.

05 — MEASUREDWhere the liquid goes

The jet does not stop at one droplet. At the critical condition the clean jet sends up seven. With more surfactant there are fewer, only two at c/ccmc = 0.023, but they are bigger. So the total liquid and kinetic energy thrown upward rise. At the lowest dose the total briefly dips, because fewer drops leave.

—

Values read from Fig. 12 (upward drops only); green tick = clean interface. The number of drops was highly reproducible. Total surface and volume varied a lot between repeats (large error bars in the paper). At 0.0233 the first droplet carries about 7× the momentum and 2.5× the kinetic energy of the clean one.

06 — METHODSThe rig

A 40 mm glass cube of liquid sits on a vibration-isolated optical table. Nitrogen is pushed through a needle in the floor of the tank. Each bubble detaches, rises and bursts within a few seconds, long enough for surfactant to settle on its surface. Two high-speed cameras, back-lit through a frosted diffuser, record the three acts at different speeds and magnifications.

Figure 2 of the paper (CC BY 4.0; cropped, tap the markers). The second camera is not drawn.

Tap a marker

Each letter in the sketch is one piece of the experiment.

Three acts, three camera settings

Photron FASTCAM SA5 and Mini UX50 cameras. The liquid was ultrapure water and DMSO, 50/50 by weight, at 22 ± 1 °C. Surfactant was stirred in for an hour and the solution used immediately. Bubbles were pure nitrogen, and every experiment was repeated five times.

Needle tips written with light

The needles that blow these bubbles have inner diameters of 10–40 µm. Their tips were 3D-printed in IP-S resin on a Nanoscribe Photonic Professional GT2 in its dip-in laser lithography (DiLL) configuration, a two-photon polymerization printer.

General background, not described in the paper: a tightly focused pulsed laser hardens the resin only in a tiny volume at its focus, so a part can be written point by point in three dimensions. In dip-in mode the objective lens is immersed directly in the liquid resin. The paper does not describe the tip geometry, so the shape drawn here is illustrative.

07 — IMPLICATIONSWhy it matters

Sea spray

Jet drops from tiny bubbles bursting at the ocean surface are a major source of aerosols around 0.1 µm in size. These particles can seed clouds and carry toxins, microplastics, bacteria and viruses into the air.

Clean water is an idealization

Natural water is never perfectly clean. Even the weak surfactant SDS, used here as a stand-in for impurities, clearly enlarged the first droplet. The authors note this fits seawater experiments in which bursting bubbles produced larger particles than clean-water predictions.

More mass, more energy

With surfactant, the jet handed more liquid and more kinetic energy to the spray. The authors suggest this may explain why surfactants increase the transfer of microplastics to the air, as other experiments have reported.

Across both surfactants the extra first-droplet size followed this fitted law (Fig. 11). It is very sensitive to the surfactant's strength Δσcmc (0.26 for Tween 80, 0.11 for SDS) and much less to how much is added. The authors offer a tentative explanation for the weak dependence on dose: Marangoni stress switches on only once crowding where the jet is born passes a threshold, and the local coverage reached there depends little on the starting concentration.

What the experiments cover. The tests used a DMSO–water mixture as a slower, larger stand-in for water. At the critical Laplace number its Bond number differs from water's, which is expected to play a secondary role. In seawater the surfactant layer may not reach equilibrium before a bubble bursts. Natural contaminants may also follow a different equation of state than Tween 80 or SDS.

Source

Sergio Rodríguez-Aparicio, Alberto Cebrián-García, Emilio J. Vega, José María Montanero and María Guadalupe Cabezas. Effect of surface-active contaminants on bubble bursting yielding the smallest first-emitted droplet. Journal of Fluid Mechanics 1032 , A49 (2026). doi:10.1017/jfm.2026.11412

Open Access under the Creative Commons Attribution licence (CC BY 4.0). Figures 2, 5 and 10 and the frames used in chapters 1–3 are reproduced from the article; we cropped them, separated and re-oriented the frame strips, re-arranged them and added markers. Values marked "read" are approximate readings from Figs. 3, 6, 9, 11, 12 and supplementary Fig. S2. The hero animation is an illustration built on those measurements and the traced Fig. 5 outlines; it is not the authors' simulation.