A customer photo, and the story in it. Up front the candle looks fine: level, clean wax surface, no frosting, centered wick. But along the rim runs a 3-millimeter channel of bare glass - a wall of wax that has pulled away from the inside of the jar. The question in the message is always the same: "Does this mean the batch is bad? Why doesn't my candle touch the glass?" That gap - "wet spots" or "pull-away" in trade language - is the most common "defect" in candle retail, and the most common one to misdiagnose. It is the default outcome of pouring a shrinking material into a rigid one, and until that is understood, every fix that gets applied treats the symptom.

Why the Gap Forms - It Is a Two-Material Failure, Not a Pouring Error

Liquid wax contracts as it sets. In volume terms: typical soy and coconut blends shrink 1 to 3 percent between the pour and full solidification. Put that against a glass wall that does not move, and you have two systems in conflict: one shrinking, one rigid. The interface between them must have a loser, and it is always the bond. Glass is a heat sink - the outermost layer of wax, the one in contact with the wall, cools and solidifies first. While the core is still liquid, that skin is already contracting against the glass. The remaining contraction pulls the interface apart and it lets go of the wall: a channel running the full height of the candle. The sequence is non-negotiable. It is not a fault of the batch. It is arithmetic.

Three modifiers decide how visible it is. First, the wax: soy and coconut blends contract more visibly than paraffin and mineral blends - which is exactly why the "natural" candles show the gap and the cheap ones do not. Second, the shape: tall narrow jars carry more surface for their volume, so a larger share of the wax solidifies against the wall, and the gap runs longer and reads worse. Third, the cooling environment. A cold room and a draft solidify the wax faster, and it shrinks harder and faster against the wall. A supplier's own troubleshooting page sets the bar at room temperature above 20 degrees Celsius during and after the pour, and notes that wax shrinkage "is an inherent property of the wax and cannot be fully avoided." That last sentence is the position of the entire industry, and almost nobody has read it. The gap cannot be poured away. It can only be minimized, managed, and sold for what it is.

The Five Spec Calls That Control It

None of these is a trick. Each one moves a number.

  • Strip the glass. Adhesion is surface energy, and surface energy dies where there is oil. A factory film, a finger smudge, dish-soap residue - each becomes the point where the wax lets go first. Wash, dry completely. This is the free fix, and the one that fails the quietest.
  • Warm the jar before the pour. A preheated wall narrows the first cooling gradient, so the first skin of wax solidifies slower and bonds better in that first minute. But lab testing by the suppliers carries the same caveat: preheating buys temporary adhesion, and the pull-away continues over hours and days. It is a starting bonus, not a cure.
  • Pour in the upper half of the wax's spec window, not above it. Too low and the near-wall layer crystallizes before it can bond - instant gap. Too hot and the whole batch shrinks faster against the wall. Every wax spec sheet has a window; the middle of the window is the target, not the ceiling.
  • Cool at room temperature, away from drafts, and do not seal early. Above 20 degrees Celsius, no window air, no AC downdraft, for 24 to 48 hours. Boxing candles in early to "cure" them fast traps heat and moisture and converts an aesthetic gap into frosting and a mushy top. Slow is the spec.
  • Leave 12 millimeters (a half inch) of headspace below the rim. This is the most under-weighted number in the entire vessel specification. When the melt pool grows to the rim, the thinnest, most stressed part of the jar sits over an 800 to 1,200 degree Fahrenheit flame field while a 65 to 93 degree melt pool soaks the rim from below. A jar filled to the brim is a jar on a countdown to a crack. Brim fills are how most "the glass shattered in the customer's home" claims start.

And when the gap survives all of it - because on some candles it will - there is a finishing move: a heat gun, a few seconds on the cold channel, re-melts the outer skin, and it re-bonds to the wall. It is a known production fix. It is ten seconds of labor per affected jar. A QC line that rakes light across the face catches the channel, grades it, and routes it to touch-up instead of the trash. The gap is cosmetic. It moves the throw and it shortens nothing. A line that rejects a candle for a two-millimeter channel is paying scrap cost for an aesthetic standard the market does not ask for.

The Second Glass: the One That Cracked

Pull-away is the soft failure. The hard failure is the jar itself, and it runs on the same physics in reverse. The glass took a thermal shock. The usual suspects, in order of frequency: a hot jar set on a cold, wet, or marble surface; a jar moved while hot; a dishwasher cycle (thermal shock plus soap etching, attacking the glass from the inside). Third, the brim-fill rim heating described above. Fourth, the wrong glass: a cheap soda-lime jar under a large multi-wick flame, instead of borosilicate, which is made for thermal swings. The flame at the wick runs 800 to 1,200 degrees Fahrenheit. The glass does not see the flame, but it sees the melt pool - and a rim soaked at 65 to 93 degrees, then struck by a cold table, is doing arithmetic the glass cannot win. When a customer's jar has cracked at home, the first question is not "what did they do wrong." It is "what did we spec - which glass, what fill level, how big a flame?" The answer is usually in the spec, and the spec is where the fix has to be written.

That is the entire container design problem in one line: the wax wants to shrink, the glass wants to stay, and the interface breaks. You cannot pour away the shrink, and you cannot argue with the glass, but you can control the temperatures, the cleaning, the fill level, and the glass itself. Do those four things and the gap becomes a two-millimeter feature you describe, instead of a defect you apologize for. Candles poured and cooled to that spec are at YSCandle.com.