How to Prevent Fragrance Oil Separation in Candle Wax

The five settings, as a specification

Most makers who lose batches to separation are not being reckless. They are working from a recipe — 200g of wax, 8% oil, add hot, stir, pour — and a recipe has no tolerances in it. It does not say how hot, how long, or what happens once the wax has cooled by jar twenty. A specification does: a value or a band for every variable that can move.

Three of the five settings below arrive with numbers already attached, published by the wax manufacturer and carried on the CSI product pages. The other two have no published number and never will, because they depend on your room in your city in your season. That does not make them optional. It makes them yours to establish once, write down and hold.

Notice what the specification does not contain: the oil. Swapping suppliers is the most expensive way to test a hypothesis you have not checked. A 92-oil catalogue does not become unstable overnight; your process does, one shortened stir at a time. The sibling page on why fragrance oil separates from wax owns the mechanism, and this page does not repeat it.

Two minutes, not forty seconds

Time yourself once, honestly. Almost everyone stops between thirty and fifty seconds — the wax looks uniform, the smell is everywhere, the arm is tired, and there is a jar waiting. It looks finished because dispersion looks exactly like dissolution to the eye.

It is not the same thing. Dispersion breaks the oil into droplets scattered through molten wax; dissolution puts it into the wax structure and holds it there. A dispersed batch is a suspension: while the wax moves, the droplets stay spread out. The moment it stops and begins to crystallise, they are pushed ahead of the advancing crystal fronts, they meet, they merge, and they arrive at the surface and against the glass. That is the film you find on day 7 on a batch that looked perfect on day 1.

The extra eighty seconds are the time it takes for shear from the stirrer to break the droplets down far enough, at a temperature high enough, for the oil to actually go in. Nothing visible happens during them. That is precisely why the step gets cut.

Two products make the two-minute stir a specification rather than an intention. Only one of them is a good idea at every batch size.

Timing and temperature, held together

Settings two and three are usually taught as separate instructions — add at this temperature, stir for this long — and treating them separately is what breaks them. They are one operation with a start condition and a duration, and the duration has to run inside the condition. The start condition is the 80–90°C band: Eco Soy CSI 400 specifies fragrance in at 80–90°C with two minutes of stirring, and CSI 464 melts and pours in the same band. Below it the wax is thickening and the oil sits as droplets no stirring will dissolve. Above it the volatile top of the fragrance leaves before the wax reaches the jar. A pen thermometer at ₹354 settles it in two seconds; molten soy looks identical at 75°C and 92°C.

The second half of the pairing is the cooling toward pour temperature. Wax thickens continuously on the way down and the window for moving oil into solution closes with it, so the stir is not a two-minute event followed by a pause: two minutes of committed work at the top of the band, then gentle continuous movement to the pour. Decide your pour temperature in advance and write it on the batch sheet rather than pouring whenever you happen to look up.

Cooling after the pour is setting four, the first with no published number. You are avoiding fast or uneven crystallisation, because oil is pushed out of solution at a crystal front and a violent front pushes hardest: stable ambient, no fan across the jars, no open window, no cold tile under warm glass, no moving the tray for several hours, never the fridge. Record the ambient you pour into and the hours the jars stay untouched, and treat them as fixed. The figures matter far less than their being the same every batch.

The twentieth jar of fifty

This failure mode catches makers exactly when they start selling properly, and it is invisible on a five-jar batch. You melt eight kilos, add the oil, stir your honest two minutes and start pouring. Jars one to eight are textbook. By jar twenty the wax has stood in the open for fifteen minutes, dropped well below the band and is being poured unstirred. By jar forty it is thick, cool and carrying whatever oil failed to dissolve.

The tell is worth knowing: if the late jars in a run sweat and the early jars do not, the fault is sequencing, not formulation. Same wax, same oil, same load — different outcome by pour order. Changing supplier will not fix that. Two things will.

Inspect at day 3, day 7 and day 14

Separation is a delayed fault. It is not visible when you pour, usually not the next morning, and very often not at forty-eight hours — which is exactly when most small makers photograph a batch, label it and box it. The oil migrates during cooling and through the first two weeks of cure, so the honest verification point is day 14, not pour day.

The cost of getting this wrong is not the wax. It is a customer opening a gift box to an oily jar with a smeared label. Three inspection points, three jars each, before release — sampled from the beginning, middle and end of the pour order, because the end of the run fails first.

Keep the log on paper at the bench, one sheet per batch, and keep the sheets — six months of them converts "I think the coconut soy batches were fine" into a fact you can check. Add the date and the room temperature at day 14, so a cluster of failures in April tells you something. If that cluster tracks the heat, it is the problem covered in can hot weather make candles sweat fragrance oil, and a reason to re-read why a candle sweats oil at all.

Hot months, and what this all costs

The ceiling is a formulation limit, not a weather forecast. A candle poured at 13% on Luxury Soy Wax Chunks is inside specification in January and still inside it in May — but in May it also sits in a warehouse, a courier van and a customer's kitchen at temperatures that give marginal oil every encouragement to leave. Working a point or two below the ceiling through the hot months buys margin at almost no cost.

The arithmetic, on a 200g candle. At 13% that is 26g of oil to 174g of wax; at 11%, 22g to 178g. On Sandalwood Fragrance Oil at 1kg pricing — ₹5,026.80, or ₹5.03 per gram — ₹130.70 of oil drops to ₹110.59, about ₹20 less per candle, or ₹10,053 across a 500-unit run. The 11% candle does not read as noticeably weaker to a customer; one with an oily film reads as defective to everyone.

Setting five, storage, follows the same logic and also has no published figure. State a maximum store temperature for finished stock and for unopened oil, keep both out of direct sun, and do not stack cartons against a west-facing wall through summer. If you cannot control the room, control the schedule: pour smaller, ship faster, hold less stock through the hottest weeks. Three days in the workshop and two in transit is a far smaller exposure than six weeks in a box waiting for a festival.

Set that against one lost run. Fifty 200g candles at 13% on Luxury Soy Wax Chunks is 1,300g of oil and 8,700g of wax: at Sandalwood 1kg pricing (₹5.03/g) and Chunks at ₹5,999 per 10kg (about ₹0.60/g), roughly ₹11,754 of raw materials before jars, wicks, labels or your time — and a separated batch is a write-off plus refunds, not a discount. The equipment that prevents it costs six percent of one bad batch.

A last note on wax choice, because it belongs to setting one. CSI 464 at ₹260 per 500g carries a stated ceiling of 10%; Luxury Soy Wax Chunks at ₹299 per 500g carry up to 13%. If your range needs the higher load, buy the wax formulated to hold it rather than pushing the one that is not. At 8% or 9% either is inside specification. Premium Coconut Soy CSI 468 at ₹325 per 500g and Eco Soy CSI 400 at ₹399 per 1kg are also in the range, and CSI 400 is the wax that publishes the 80–90°C add with two minutes of stirring — a good one to learn the specification on. A candle care sticker at ₹413 per 500 carries your storage line to the customer.

Frequently asked questions