Why Is Fragrance Oil Separating From My Soy Wax?

What separation looks like, and three things that only look like it

Fragrance oil in a finished soy candle is meant to be held inside the set wax, invisible, until the melt pool releases it. Separation is any sign that some of it was never held, or has worked its way out. It shows up in four ways, and which one you have narrows the cause before you spend anything.

The tissue test: oil or air?

Soy wax in glass often shows patches where the wax has pulled away from the glass as it shrank. Through clear glass these look dark and wet, and makers read them as oil. They are gaps. The test takes a minute: on the surface, press a folded white tissue on the spot for ten seconds. A translucent, greasy mark that smells of your fragrance is oil. No mark means it is not oil at the surface. For a patch you can only see through the glass, look from above as well: if the top of the candle is dry and matte and the "wet" patch is only visible from the side, it is almost always pull-away, which is a cosmetic adhesion issue, not separation. Clear glass shows it most; that is a jar choice, not an oil problem.

Three look-alikes that are not separation

Frosting. A white, dry, crystalline bloom on the surface or against the glass. It is the wax itself, it is common in soy, and it is dry to the touch. The CSI 464 page describes "Low frosting, minimal blooming", and the CSI 468 page "minimal frosting"; minimal is not none. Glass pull-away (wet spots). Covered above: air, not oil. Surface cracks or sinkholes. Shrinkage as the wax sets, commonly from fast or uneven cooling. They can collect oil if the load is high, which is how two faults end up in one photo, but the crack is not caused by the oil.

The five causes, explained with what the wax pages say

We hold no lab data on why a given oil separates in a given wax, and nobody can give you a mechanism that is true for every oil. What follows is how the causes commonly behave, hedged where they vary, with the figures that CSI's own wax and oil pages state. Where a page is silent, we say so.

Cause 1 — load above the lower ceiling

Every wax holds only so much oil. Past that point, extra oil tends to have nowhere to go and appears as a pool, a film or beads; more oil does not raise throw in step with it. The rule is two ceilings: the lower of the wax page's stated maximum and the oil page's stated candle maximum governs. Luxury Soy Wax CSI 464 states "Fragrance Load Up to 10%". Luxury Soy Wax Chunks state "up to 13% fragrance load capacity" and Premium Coconut Soy Wax CSI 468 "Up to 13%". Eco Soy CSI 400, Natural Soy Pearl Wax and CSI's other soy and coconut waxes do not state a ceiling in the page text we hold; ask before you assume one. These are formulation limits, not regulatory ones, and the maximum is not the best load.

Worked example of the two-ceilings rule: Japanese Cherry Blossom states "6-13% for candles". In CSI 468 or the Chunks, 13% is permitted by both pages. In CSI 464 the wax page caps it at 10%. A maker who reads the oil page and pours 13% in CSI 464 puts 19.5 g into a 150 g candle where 15 g is the wax's stated limit: 4.5 g extra per candle, ₹21.79 per candle at the 1 kg price, ₹2,178.90 per 100 candles, for oil the wax page says it was not built to hold.

The arithmetic slip goes the other way. Load is a percentage of total candle weight. Adding 15 g of oil to 150 g of wax gives 9.1%, not 10%; adding 18 g gives 10.7%. Calculating on the wax weight makes you under-load slightly, so it rarely explains separation. What does: measuring oil by spoon or by millilitre (oils differ in density, so millilitres are not grams), "a little extra for luck", or using a second oil on top of the first in a blend and forgetting to count both. Weigh on a scale.

Cause 2 — oil added below the wax page's temperature

Too-cool addition tends to give poor incorporation: the oil does not disperse evenly and can reappear as mottling, a film or wet-looking patches. Too-hot addition, or long holding at high temperature, is a different problem (it can drive off lighter top notes and may darken some oils) and is not a common cause of separation. The numbers come from your wax page, not from us: CSI 464 states a pour temperature of 80°C–90°C; CSI 468 states a melting point of 50°C–53°C and a pour temperature of 58°C–60°C; Eco Soy CSI 400 says "For the best hot throw, add your fragrance or essential oils at 85°C (185°F), stir gently for 2 minutes, and pour directly", with a table entry of "Fragrance Add Temperature 80 to 90 degrees". The Chunks page gives no temperatures in the text we hold; ask on WhatsApp rather than borrow another wax's numbers. The full temperature question lives in Did I Add Fragrance Oil at the Wrong Temperature?, and the too-cold case in What Happens If Fragrance Oil Is Added to Candle Wax Too Cold?.

One practical trap: oil straight from a cold store, or a cold jug, drops the wax temperature the moment it goes in. In a 150 g test that is a small mass of wax losing heat quickly. Check the temperature again after the oil goes in, not only before.

Cause 3 — too little stirring

Fragrance oil and molten wax are both liquids, but they do not blend instantly. Poured in and swirled for a few seconds, oil tends to sit in streaks or settle, and those streaks can set as patches that weep later. Eco Soy CSI 400's page is the only CSI wax page in our captured text that gives a stirring time: "stir gently for 2 minutes". Treat that as a sensible default for other soy waxes too, labelled as our judgement. Gentle matters: whipping the jug adds air bubbles, which is a different surface fault. How long and how hard to stir is the whole subject of How Long Should I Mix Fragrance Oil Into Candle Wax?.

Cause 4 — cooling too fast or too slow

How a candle sets affects how evenly the wax holds its oil. Jars poured cold, set under an AC vent or a ceiling fan, or moved while setting tend to shrink unevenly: cracks, sinkholes, glass pull-away and, at higher loads, oil collecting in the gaps. At the other end, jars packed tightly together in a hot, closed room cool slowly and unevenly, and makers commonly report more surface sheen and mottling. We cannot give you a "correct" cooling time for every wax, and nobody else can either. What you can do is keep it consistent: warm the jars slightly if the room is cold, space them apart, keep air still, and do not move them until set. In an Indian summer (roughly April–June) a closed workroom can stay hot overnight; note the room temperature on your log.

Cause 5 — the oil and the wax do not agree

Some oils do sit less comfortably in some waxes. Formulations differ, and a given oil can behave well in one soy and less well in another. This is the one cause a new oil can fix, which is why makers reach for it first. It is also the one you can only prove by elimination: correct load, correct temperature, full stir, calm cooling, and a control oil that stays put beside yours. What the oil page tells you about soy compatibility, and how to run a three-candle compatibility check, is in How Do I Know If a Fragrance Oil Is Compatible With Soy Wax?.

The diagnostic order, and why the oil goes last

The order below is not a claim about frequencies we have measured; CSI has no such data. It is the order that gets you to an answer cheapest: arithmetic before instruments, instruments before test candles, test candles before a new oil. It is also the order in which a later step is meaningless if an earlier one has failed. There is no point testing an oil's compatibility at 13% in a 10% wax.

Is it actually the fragrance oil? Four questions

Answer these honestly before you order anything. One: is the load, in grams, on total candle weight, below the lower of the two ceilings? If you cannot say yes with a number, the oil is not yet a suspect. Two: did the oil go in within the wax page's stated temperature, and did you check it after adding? Three: did you stir for a timed two minutes? Four: did other oils, poured the same way in the same wax on the same day, stay put? If a different oil in the same batch run separated too, the oil is very unlikely to be the cause. Only when all four are yes, and this oil alone misbehaves, is it time for the control test in the next section. The fuller decision rule for replacing an oil that keeps separating is in Should I Change Fragrance Oil If It Keeps Separating From Candle Wax?.

The five-jar split test: one melt, one variable per jar

This test assumes step 1 has already passed on paper: every jar is a 150 g candle at 8%, which is 12 g of oil and 138 g of wax, weighed separately for each jar so the load is identical. 8% sits inside CSI 464's 10% and inside every oil page that states a soy maximum of 10% or more. Use your own wax, jar and wick; the question is about your candle, not ours.

Reading it: if A separates and B does not, it was temperature. If B separates and C does not, it was stirring. If C is clean and D is not, cooling is your problem. If C separates and E does not, your oil and that wax do not agree. If A is clean at 8%, the fault was the load you were using before, and the test has already answered you. If C and E both separate, the process still has a fault the test did not catch: re-check the thermometer against a second one, and ask on WhatsApp with photos.

With Sandalwood as the control instead, the full new spend is ₹743.40. All figures exclude labour, freight, packaging and GST. For comparison, a 1 kg bottle of Sandalwood bought on hope as a "more compatible" replacement is ₹5,026.80, 7.3 times the full diagnosis kit, and if the real cause was a cold add or a three-second stir, it would separate too. Test jars can be washed and reused once the verdict is in, so the consumed figure is an upper bound, not a bill.

What to buy: two tools, then a control oil

In the order the diagnosis uses them: a thermometer for causes 2 and 4, stirring sticks for cause 3, then a control oil and a woody alternative for cause 5. Oils were chosen on a stated soy maximum of 10% (so 8% sits inside both ceilings in CSI 464), stated notes, price per gram and stock at every size on 25 September 2026. CSI has no measured separation or stability data for any oil, and nothing below is ranked on it. Candle counts assume 150 g at 8%.

Where each separation problem is covered in depth

Once the split test has named the cause, go to the guide that owns it. Each one goes deeper than we can here, and each stays on its own question.

Fixing it before Diwali, 8 November 2026

Diwali falls on 8 November 2026, 44 days after these prices were checked. The split test gives a verdict in seven days, so a maker starting this week has time for one corrected production run with a full cure and a margin for dispatch. Confirm dispatch time to your city on WhatsApp before you promise a customer a date.

Worked example: 41 British Rose candles at 8% (500 g of oil, ₹1,593.00) with CSI 464 (₹2,870.87), 150 g clear jars and C1 wicks come to ₹8,576.17, or ₹209.17 per candle, excluding labour, freight, packaging and GST. Oil per candle at the 500 g price: ₹38.23 for British Rose, ₹60.32 for Sandalwood.

Frequently asked questions