How to Prevent Sinkholes in Candles

The ten-step pour, and the reason for each step

Every candle shrinks as it freezes. The only question is whether that shrinkage spreads evenly across the jar or concentrates in one place — and it concentrates when the outside sets before the middle. The full mechanism is on the sibling post, why does my candle have a sinkhole, and this page will not repeat it. What follows is the procedure that stops it happening.

Temperature control is the whole thing

If you take one thing from this page, take this: you cannot control cooling rate if you do not know starting temperature. Judging it by eye — "it looked ready", "it had gone cloudy at the edges" — introduces a swing of twenty degrees or more between batches, and a twenty-degree swing is the difference between a flat top and a crater.

Cooling environment, and how to preheat a jar

This is the point most makers have backwards. A candle that is hard in forty minutes feels like a success. It is not. Fast setting means a steep temperature gradient between the outside of the jar and the middle, and a steep gradient is exactly what drives the outside to freeze into a rigid shell while the core is still liquid and still shrinking. The shell cannot move inward, so the top surface gives instead. Slow is the objective. Everything below buys time.

Warming the vessel does two separate jobs. The first is sinkhole prevention. Cold glass under hot wax is a heat sink: the layer touching it sets into a rigid shell within seconds while the core still has minutes of shrinkage ahead of it, and that shell forces the shrinkage upward into a crater. Warm glass keeps the wax at the wall liquid longer, so the jar cools as one body and the shrinkage distributes.

The second job is glass adhesion. Wax that freezes instantly against cold glass then contracts away from it, leaving air gaps between candle and jar — the pale patches, or wet spots, that make a jar look faulty on a shelf even when it burns perfectly. They are cosmetic, but they cost sales. One step solves both problems.

The relief hole and the second pour

The ten steps reduce shrinkage and spread it evenly, but they do not abolish it. On some waxes and some jar shapes you will still get a shallow dip around the wick, and these two finishing operations deal with what is left.

The window is judged by touch, not by clock. You want the surface skinned firm enough that a needle leaves a clean hole rather than a smeared dent, and the core still soft enough that the hole reaches near-liquid wax. Press gently near the jar wall: if your finger sinks in, wait; if it is glassy and unyielding, you have missed it. Push a wicking needle or wooden skewer straight down two or three times, close to the wick but not touching it, most of the way to the base. The shrinking core must relieve its volume somewhere; given a channel it draws wax down and pulls air out instead of pulling the top into a crater. The holes close under the second pour.

Not every candle needs one. The working rule: if the top at twenty-four hours is visibly lower than the level you poured to, that wax and jar combination needs a second pour, and it will need one every time. Straight soy in a narrow container often needs very little. Pillar and mould work almost always does, which is why Soy Pillar Wax is formulated to shrink — it has to pull away from the mould to release, and the same shrinkage shows as a well around the wick.

On temperature, the top-up should go in a few degrees above your first-pour temperature — hot enough to fuse with the existing surface, not so hot that it liquefies the candle beneath and restarts the cooling problem. CSI publishes no second-pour figure for any wax in the range. That is a rule of thumb, not a manufacturer specification; for guidance on the wax you actually bought, ask us on WhatsApp rather than trusting this paragraph.

The pour log, and what changes at fifty jars

A Chennai flat in May and a Dehradun garage in January are not the same problem, and no article can tell you which of the ten steps you personally are getting wrong. A log can. Record five fields per batch, do it for two weeks, then read down the result column and see what the failing rows have in common. It takes about ninety seconds per pour.

After ten to fourteen batches the pattern is usually blunt: every failure was on the steel bench, or below a certain ambient, or on a day the fan was on, or in the last four jars of a large batch. That last one deserves its own section.

A five-jar pour is over in under a minute, so the whole batch gets the same pour and either all of it works or none of it does. A fifty-jar pour is a different operation. Wax in an open pitcher loses heat continuously, and faster as the volume drops, because a smaller mass with the same surface area cools more quickly. If your pour temperature was right at jar one, it is below the band by jar forty — and you get what the log predicts: the front of the tray flat, the back cratered. Makers scaling up call this "the wax has gone inconsistent". The wax has not changed. The process has.

One last note, because the question always follows. Sinkholes are almost never a wax quality problem. If you have ruled out the process — measured pour temperature, warmed jars, slow cooling on wood in still air, relief hole, top-up — and still have craters across a whole tray, check the material using the method in how to verify wax quality before you buy in India. Process first, though: it is the answer nineteen times out of twenty, and changing supplier without changing method just carries the problem across.

Frequently asked questions