Bubbles, foam and pinholes trapped by a melt that thickened before the air escaped
Foam is air whipped in and frozen in place as the melt thickens. Degassing holds, pour technique and a surface reflow that rescues it without a full remelt.
The air in a bubbly balm went in while you were stirring, and it stayed in because the melt thickened faster than the bubbles could climb out of it. Nothing about the recipe caused it. It is a race between a rising bubble and a falling temperature, and the numbers on both sides are close enough that a few degrees at the pour decides the outcome.
Entrained air only escapes while the melt is thin. A 0.2 mm bubble rises roughly 1 mm per second at 75 C and clears 10 mm of melt in about 11 seconds; the same bubble takes half an hour at the first sign of clouding, and once crystals give the melt a yield stress it never rises at all. Rest the melt 5 to 10 minutes at 75 C, then pour above 60 C.
Where the air comes from
Molten balm is a thin, low surface tension liquid that wets almost anything, which makes it very good at swallowing air and very bad at giving it back. There are five routes in, and they leave different marks.
Vortex entrainment is the big one. Stir fast enough to pull a visible dimple in the surface and the dimple is a funnel: air is dragged down the axis and sheared off into a cloud of fine bubbles. Nobody sets out to whisk a balm, but a hand stirring "briskly" to dissolve the last of the wax is doing exactly this. Splash entrainment happens at the pour, when a stream falling more than a centimetre or two punches its own air pocket into the surface it lands on. Adsorbed air on powders arrives with every dry material you add: mica, zinc oxide, clay, starch and pigment all carry air on their particle surfaces, which is why a pigmented balm often foams the moment the colour goes in. Cold additive shock chills the melt locally, so a cold slug of fragrance or a cold spatula creates a thickened, bubble-holding pocket at exactly the moment you are stirring it. And wall air is the one people miss: a cold, dry container has a boundary layer of air that the arriving melt does not always displace cleanly, leaving a ring or ladder of bubbles against the wall.
All of these are reversible in principle. The melt is fluid, the bubbles are buoyant, and given time they will surface and burst. Time is the thing you do not have.
Why the bubbles cannot get out again
A bubble rising in a viscous liquid reaches a terminal velocity almost immediately, and for small bubbles that velocity follows Stokes' law: speed is proportional to the square of the bubble radius and inversely proportional to the viscosity of the liquid. Two consequences fall straight out of that. Halving the bubble diameter quarters the rise speed, so the fine foam is exactly the fraction that survives while the big obvious bubbles clear themselves. And viscosity climbs steeply as a balm melt cools, so the escape window closes far faster than the temperature falls.
| Melt condition | Viscosity | Rise speed, 0.2 mm bubble | Clears 10 mm | Clears 10 mm, 0.05 mm bubble |
|---|---|---|---|---|
| Thin melt, 80-85 C, low wax | 10 mPa s | 1.9 mm/s | 5 s | 1.4 min |
| Working melt, about 75 C | 20 mPa s | 0.93 mm/s | 11 s | 3 min |
| Cooled to about 65 C | 50 mPa s | 0.37 mm/s | 27 s | 7 min |
| First cloud, about 58 C | 200 mPa s | 0.09 mm/s | 1.8 min | 29 min |
| Visibly thickening, about 55 C | 1000 mPa s | 0.019 mm/s | 9 min | 2.4 h |
| Below the cloud point, network formed | yield stress | 0 | never | never |
That table is the whole page in one block. A bubble you could have lost in a few seconds at 75 C needs half an hour at 58 C, and a maker who stirs a fragrance in at 55 C and pours immediately has given the air no chance whatsoever. The temperature at which you stop touching the melt matters more than any other number here.
The yield stress cut-off, which is where the window really shuts
The last row of the table is not an extrapolation of the others. Once wax and hard butter start crystallising, the melt stops being a liquid with a high viscosity and becomes a soft solid with a yield stress: it does not flow at all until the stress applied to it passes a threshold. A bubble can only move if its own buoyancy yields the material around it, and the buoyancy stress available scales with the bubble radius. For a bubble 0.2 mm across in a melt of density 870 kg per cubic metre, that stress is under 1 Pa.
Set balm has a yield stress in the hundreds to thousands of pascals, and even the first sparse wax network to appear on cooling is orders of magnitude beyond what a bubble can overcome. This is why waiting does not work. There is no slow leak of air out of a setting balm; the population of bubbles present at the moment the network forms is the population you have for the life of the product, and every one of them is permanent. Static bubbles in yield stress fluids are a well studied case in the non-Newtonian flow literature (Chhabra, 2006), and the practical criterion is blunt: at balm bubble sizes, nothing escapes.
This also explains why bubbles are worse in hard formulas than soft ones. A 25% beeswax lip balm develops a network 8 to 10 C higher than a 12% wax body balm does, so it shuts the escape window earlier in the cool and takes bubbles from a hotter, thinner state that felt safe to stir.
Read the pattern before you reprocess
Where the bubbles sit tells you which of the five entry routes you used, and that decides whether a surface reflow will do or the batch has to come back to the pot.
| What you see | Cause | Confirm by | Fix |
|---|---|---|---|
| White froth or a foam crust on top only | Surface foam floated up and set there, from vortex stirring or splash | Scrape 2 mm off: clean balm underneath | Reflow the top, no remelt needed |
| Fine pinholes evenly through the body | Air entrained into the bulk and locked by the network, poured too cool | Cut a tin in half or split a tube lengthways | Full remelt, rest at 75 C, pour hotter |
| A ring or ladder of bubbles against the wall | Cold or dusty container, melt did not displace the wall air | Bubbles are flattened against the metal or plastic, not spherical | Warm containers to about 40 C, pour down the wall |
| Bubbles concentrated in the bottom third of a tube | Melt bridged and set at the neck before the base filled | Tube is light against its stated fill weight | Raise pour temperature, fill in one continuous stream |
| Foam that appeared only when the colour went in | Adsorbed air carried on pigment particles | Run the same batch uncoloured | Pre-wet pigment as an oil slurry, add hot |
| Bubbles plus spitting and a cloudy melt | Water, not air: steam bubbles from a wet ingredient or a wet vessel | The melt crackles and the surface looks hazy rather than clear | Stop. See water contamination |
| One large central void, smooth walled | Not a bubble. Solidification shrinkage | The cavity is a single cavity in the thermal centre | See tunnelling and dips |
Rescue the batch in front of you
Surface foam is a five minute job. Bubbles through the body are a remelt. Decide which you have first, because reflowing a batch that is bubbly all the way down wastes the effort and leaves a glossy lid on a faulty product.
Surface reflow. Work a heat gun on its lowest setting 15 to 20 cm above the open containers, moving constantly. You are aiming to hold the top 2 mm at about 70 C for a few seconds, no more. The foam collapses and the surface levels itself. Stop the moment it goes glossy; going deeper melts the column under it and you will get a dip as it shrinks again. A tray of tins takes under a minute. This works on tins and jars and does not work on tubes, where you cannot reach the surface without deforming the plastic.
- Scrape the whole batch back into one vessel. Include the tin scrapings. Reprocessing container by container guarantees an inconsistent result across the batch.
- Melt to 75 C without stirring more than you must. A water bath heats evenly and does not need the agitation that direct heat does. See melting methods for why the bath wins here.
- Rest it undisturbed for 5 to 10 minutes at 75 C. This is the degassing hold, and it is the step people skip. At that temperature the table above says a 0.05 mm bubble clears 10 mm in about three minutes, so a 30 mm depth of melt is genuinely clear after ten.
- Skim, do not stir. Any residual foam sits as a thin ring at the edge of the surface. Lift it off with the edge of a spatula and discard it.
- Add heat sensitive materials with a slow fold. Fragrance, flavour and tocopherol go in below 60 C, moved with a spatula held under the surface. Never bring the blade up through the meniscus.
- Pour at 60 C or above. Tins 60 to 68 C, tubes 65 to 72 C, as set out in pour temperatures. Below 60 C you are pouring into the trapping regime.
- Two firm taps on the tray. Set the filled tray down flat and tap it twice against the bench. Small bubbles coalesce into larger ones, which rise fast enough to matter.
Every remelt costs shelf life. Holding an oil blend at 75 C consumes antioxidant and starts the oxidation clock again, and a batch rescued twice should get a shortened best-before rather than the original one. Judge by smell before you reprocess: reheating a batch that has already gone off gives you smooth, bubble-free rancid balm.
Pouring without putting the air back in
Most bubble problems are created in the last thirty seconds. The melt was clean in the pot and the pour ruined it.
Pour down the wall, not into the middle. Tilt a tin 15 to 20 degrees, land the stream on the inside wall a few millimetres above the rising surface, and let it run down and fill from below. The stream should be continuous and slightly fat rather than a thin dribble, because a thin stream breaks into drops and every drop punches its own air pocket. Keep the spout under about 10 mm from the surface throughout, which means chasing the level down as the container fills rather than holding the jug at a fixed height.
Warm the containers. A tray held at 35 to 45 C stops the melt skinning against the wall and cuts wall bubbles almost completely; it also buys a few extra seconds of thin melt at the surface, which is where the last bubbles have to go. This matters most for tubes, where the wall to volume ratio is high and the fill is deep and narrow. Full method for those is in filling lip balm tubes.
And keep the pour above 60 C. This is the one rule that carries the most weight, and it sits in tension with the advice to pour cool for shrinkage control. When those two conflict, hold the melt hot in the pot, cool it in the jug, and pour at the top of the window with a two stage top up rather than pouring a single cold fill.
Preventing it at the bench
- Degassing hold: 5 to 10 minutes at 75 C, undisturbed.
- Pour floor: 60 C. Below that the melt traps what it holds.
- Pour height: under 10 mm from spout to surface.
- Container preheat: 35 to 45 C.
- Rise speed falls as the square of bubble diameter: fine foam is the fraction that survives.
- Reflow depth for a surface fix: about 2 mm at 70 C.
Tool choice does more than technique. A flat silicone spatula moved in a slow figure of eight, blade kept below the surface, dissolves wax perfectly well and entrains almost nothing. A balloon whisk is an air entrainment device and has no place in an anhydrous melt. Neither does a stick blender: it is designed to shear air and water into an emulsion, and in a single phase melt it produces a fine, uniform foam that no amount of resting will clear. If a formula seems to need one, the real problem is usually undissolved solids, and the answer is a longer hold at temperature rather than more shear. Our notes on workshop equipment cover what is worth owning.
Dry powders need pre-wetting. Slurry mica, oxides or starches and silica into a small amount of warm oil, work out the lumps, and stir the slurry into the hot melt rather than dusting dry powder onto the surface. The same discipline reduces the streaking and settling covered under pigment sinking. Materials that carry their own water, honey being the obvious one, bring a different problem entirely, and the bubbles they cause are steam rather than air.
Bubbles that are not bubbles
Three faults get filed under bubbles and none of them respond to degassing. A single smooth walled cavity in the middle of a tin or down the axis of a tube is solidification shrinkage, and the fix is a two stage pour. Hard specks that catch the light and feel gritty are undissolved wax, poured before it had fully dissolved. A haze of very fine points that dulls the surface without any individual bubble being visible is usually crystal texture rather than air, and belongs with matte instead of glossy balm, though heavy micro-foam produces the same optical result by scattering light off thousands of tiny gas interfaces.
The one to take seriously is steam. If the melt crackles, spits or looks cloudy while hot, there is water in it, from a damp vessel, a wet botanical, or an ingredient that was never anhydrous. That is not a cosmetic fault. Water in a nominally waterless product creates the conditions for microbial growth, which is the whole argument in do balms need preservatives and the reason moisture control is a discipline rather than a nicety.
Whipped products, where the air is the product
Everything above inverts for whipped body butter. There the goal is to beat in as much gas as the fat network can hold, typically 20 to 60% by volume, and the same yield stress that traps unwanted bubbles is what stops the wanted ones from coalescing and escaping. The technique reverses accordingly: you whip cold rather than hot, at the point where enough crystals exist to stabilise the bubble walls but the mass is still soft enough to fold, and a stick blender or a stand mixer is the correct tool rather than a mistake.
The failure mode reverses too. A whipped butter that collapses in the jar has lost its bubbles because the network softened and let them coalesce, which is a different problem with a different fix, covered in whipped body butter collapsed. If you are choosing between the two formats, the trade in density, yield and feel is set out in body balm versus body butter. A poured body balm should have no air in it at all.
What the bubbles actually cost you
Decide with the fill weight, not the appearance. Set balm has a working density near 0.92 g/mL, so a 30 mL tin holds about 27.6 g full. If the label says 25 g net, your entire margin is 2.6 g, about 9% of the volume. Ten percent entrained air eats all of it and puts you under the declared weight, which is a labelling problem rather than an aesthetic one, and a good reason to check filled units against a calibrated balance as described in weighing and calibration.
Below that threshold it is a judgement call. Surface foam on a tin is a five minute reflow and there is no reason to sell it unreflowed. Fine pinholes through a body balm are invisible in use and cost nothing but appearance, and remelting a whole batch for them spends more shelf life than it recovers. Bubbles in a twist-up stick are the exception worth being strict about, because a plane of voids is a weak plane, and a stick that snaps mid-column becomes a tube that will not twist up in a customer's pocket. Foam is a process fault, not a safety one, so the honest position is to fix the pour, use the bubbly stock yourself, and stop calling a reflow a rescue when the batch never needed one.
Frequently asked questions
Are air bubbles in lip balm dangerous?
No. The bubbles contain ordinary air, and a balm full of them is chemically identical to a balm with none. The only real consequences are appearance, a short fill weight if there is enough air to matter, and mechanical weakness in a twist-up stick, where a plane of voids can let the column snap. Steam bubbles from water contamination are a different matter and do need investigating.
Why do bubbles rise out of the melt but not out of the balm?
Because a cooling balm stops being a liquid. While the melt is thin, a bubble rises under buoyancy at a speed set by Stokes' law, roughly a millimetre a second for a 0.2 mm bubble at 75 C. Once wax crystals form a network, the material has a yield stress, and the buoyancy of a bubble that size generates under 1 Pa of stress against a network worth hundreds. It cannot move at all.
Can I get bubbles out without remelting the whole batch?
Only if they are on the surface. A heat gun on its lowest setting, held 15 to 20 cm away and moving constantly, reflows the top 2 mm and collapses surface foam in seconds. Stop as soon as the surface goes glossy. If the bubbles run through the body of the balm, reflow just puts a clean lid on a faulty product and you need a full remelt.
What temperature should I pour at to avoid air bubbles?
Keep the pour above 60 C, with tins at 60 to 68 C and lip balm tubes at 65 to 72 C. Below 60 C most formulas are already clouding, viscosity has climbed enough that entrained air needs many minutes to clear, and the crystal network that traps it permanently is minutes away. Pour hot, then let the filled containers cool undisturbed.
Should I use a stick blender to mix a balm?
No. A stick blender exists to shear two phases together and force air into an emulsion. In a single phase anhydrous melt it produces a fine, uniform foam that resting cannot clear, because the bubbles it makes are small enough to rise at a fraction of a millimetre a minute. Use a flat spatula in a slow figure of eight, kept below the surface. Whipped butters are the deliberate exception.
Why does my coloured balm foam when the plain one does not?
Dry pigment particles carry air adsorbed on their surfaces, and dusting powder onto a melt introduces all of it at once. Mica, iron oxides, zinc oxide, clays and starches all behave this way. Make a slurry with a little warm oil first, work out the lumps, then stir the slurry into the hot melt and hold it at temperature for a few minutes before pouring.
Does tapping the tray really help?
Two firm taps on a flat bench help more than they should, because tapping makes small bubbles coalesce into larger ones and rise speed scales with the square of the diameter. Merging four 0.1 mm bubbles into one roughly 0.16 mm bubble more than doubles its speed. It only works in the seconds before the surface skins, so tap immediately after filling.
Sources and further reading
- Bird, R. B., Stewart, W. E. and Lightfoot, E. N., Transport Phenomena, 2nd edition, Wiley, 2002 (creeping flow past a sphere and the Stokes drag result).
- Clift, R., Grace, J. R. and Weber, M. E., Bubbles, Drops and Particles, Academic Press, 1978, reprinted by Dover, 2005.
- Chhabra, R. P., Bubbles, Drops and Particles in Non-Newtonian Fluids, 2nd edition, CRC Press, 2006 (bubble motion and entrapment in yield stress fluids).
- Noureddini, H., Teoh, B. C. and Clements, L. D., Viscosities of vegetable oils and fatty acids, Journal of the American Oil Chemists' Society, 69(12):1189-1191, 1992.
- ASTM International, ASTM D445 Standard Test Method for Kinematic Viscosity of Transparent and Opaque Liquids, West Conshohocken.
- The Weights and Measures (Packaged Goods) Regulations 2006, SI 2006/659, United Kingdom (average quantity rules for declared net contents).
Reviewed and updated 6 September 2026. Spotted an error? Tell us and we will fix and log it.