Tunnelling, dips and sink holes in balm: why they form and how to pour them out

Balm shrinks as it solidifies and the last liquid to freeze leaves a cavity. Pour temperature table, a two-stage pour, reflow method and the hidden tube hollow.

Every balm shrinks as it sets, because solid fat is denser than the melt it came from. That fact is unavoidable. What you control is where the missing volume ends up: spread invisibly across the surface, concentrated in an ugly crater, or hidden as a void down the middle of a tube where nobody sees it until the customer does.

The physics, in one paragraph

Use 0.92 g/mL as the working density of finished balm. The melt is less dense, because thermal expansion and the disordered liquid state both take up more room than a packed crystal lattice does. Take an illustrative molten density of 0.87 g/mL: a 27 g fill occupies about 31.0 mL hot and about 29.3 mL cold, so roughly 1.7 mL has to disappear during setting. In a 1 oz tin about 55 mm across that is close to a millimetre and a half of depth, all of it concentrated in the last region to freeze. Nothing about that is a fault. The fault is letting all of it appear in one place.

Solidification runs inward from every cooled surface, which for a tin means the base and the walls, and for a tube means the whole cylindrical wall. The freezing front advances until the remaining liquid is a shrinking pocket in the thermal centre. If that pocket is open to the surface it simply sinks and you get a dip. If a skin has already closed over it, the contraction pulls a vacuum instead and you get a sealed cavity: a tunnel. Pouring hot makes both worse, because a hotter melt has more contraction still to come after the shell has formed.

Diagnose it

Pour temperatures that avoid it

The general rule is to pour just above the temperature at which the melt begins to cloud and thicken, which is formula-specific and worth measuring once for each recipe you use. Full context for those windows is in pour temperatures. If the formula contains shea, remember that the melting and cooling requirements of the butter come first, because dropping the pour temperature to control shrinkage must not mean pouring below the point that avoids shea grain. In that conflict, hold the melt hot in the pot, cool it in the jug to the pour window, then pour.

The two-pour procedure

1. Hold back 10% of the batch. Keep it in the melting vessel at 70 to 75 C, covered, while the first pour sets. This is separate from the 8 to 15% overage you allow for pot and spatula losses.
2. Fill to 90 to 95% of the container. For tubes, stop a few millimetres below the rim, not level with it.
3. Wait for a skin and the start of a sink. Typically 3 to 8 minutes at room temperature. You want the outer shell set and the centre still soft, not a fully solid puck.
4. Top up warm. Add melt at roughly the same pour temperature, filling the depression and slightly proud of the rim. It will fuse with the first pour rather than sit on it as a separate layer.
5. Do not chill. Let the topped-up containers set at room temperature on a wooden board or a folded towel, away from a draught. A cold metal bench pulls heat out of the base and drives the shrinkage downward and inward.
6. Weigh and section one unit per batch. Fill weight plus a lengthways cut on one tube is the only reliable evidence there is no hidden void. Method detail is in filling lip balm tubes.

Tubes versus tins: the expensive version of the fault

In a tin the fault is visible and therefore cheap. You see the dip, you reflow it, nobody is harmed. In a twist tube the same physics produces a slim void running down the axis, invisible from the top because the meniscus looks fine and invisible on the scales because the fill weight was taken as liquid, which is why the filling sequence in how to make lip balm ends with a set-and-check step. The customer winds the column up, the wall of the column proves to be a shell around a hole, and the top two centimetres break off in the first week.

Tubes are more prone to it than tins for two reasons: the cylindrical geometry cools from all sides toward a single axis, and tubes are usually poured hotter to get a clean fill through a narrow aperture. The countermeasures are to warm the empty tubes to around 30 C before filling, to fill in two stages, and to let the rack set at room temperature rather than in a fridge. Container weights to check against are in the packaging guide, and the 4.2 to 4.9 g fill of a standard oval tube leaves very little margin for a void before the customer notices.

Reflow: fixing what you already poured

Surface dips are repairable in the container. Hold a heat gun or a hair dryer at low setting (see workshop equipment) 10 to 15 cm above the surface and keep it moving until the top 2 to 3 mm liquefies and levels, then set the container down and leave it. Do not tilt it while liquid, and do not lid it warm, because condensation and slow cooling are both worse than the dip you started with. A soldering-iron style wax pen also works for a small number of tins. What reflow cannot do is fill a sealed internal cavity: heat reaches the surface first, the top melts and runs into the hole, and you finish with a lower surface and the same void partly filled. For deep tunnels, pierce the centre first with a warm skewer, then top up with melt, then reflow the surface.

Anything beyond that is a remelt, and a remelt is a full reprocess: scrape everything back into one vessel, melt to 70 to 80 C, cool with stirring, pour again in the correct window. Reprocessing inside the container is the mistake to avoid, because a tin sitting in a warm oven melts and then cools slowly in place, which produces a coarse crystal network, poorer oil binding and a real chance of weeping later.

Preventing it in the formula, not just the process

Process fixes come first, but some formulas are simply more prone to shrinkage faults. High total wax means a higher solid fraction and more contraction, and fast-crystallising hard waxes concentrate that contraction into a short temperature window. A formula that dips badly at 20% wax may behave itself at 15% with a hard butter making up the difference, a trade explained in balm texture science. Very thin, fast-absorbing oils also drain into the shrinkage cavity and can leave a slick in the crater, which looks like two faults at once.

When it is not worth saving

Reflow a tin, always. Reprocess a batch of tubes only if the void is confirmed and the batch is large enough to justify the melt, because emptying wound tubes is slow, messy work and the recovered balm is often contaminated with plastic shavings. If it is a handful of tubes, write them off. And if the dip comes with a slick of exuded oil or a gritty texture, stop treating it as a pouring fault: those are separate mechanisms, and a beautifully levelled surface on a balm that weeps or grains is still a batch you cannot sell. Fix the underlying formula first, then worry about the finish.

Reviewed and updated 6 September 2026. Spotted an error? Tell us and we will fix and log it.