Why Is Fragrance Oil Separating From My Soy Wax?

What separation actually is — and what it is not

A properly made soy candle is a single material: fragrance oil dissolved into molten wax and then locked into the wax as it sets. Separation is the failure of that union. The oil is present in the jar, but it is sitting somewhere on its own — floating on the melt, pooled in a layer on the set top, collected in soft translucent patches, or slicking the inside of the pitcher. It is not evenly distributed and it is not held.

Three different faults get called “separation” in Indian candle-making groups, and they need different fixes. Separation in the pot is oil that never dissolved: you see it before you pour. Separation on setting is oil that dissolved but could not be held as the wax crystallised: it appears within hours of pouring, usually as a glassy or oily film on the top. Sweating is something else entirely — fine beads that appear on a candle that set soundly and has since met heat, and it belongs to storage and load margin rather than to mixing (why your candle is sweating fragrance oil).

A fourth thing, oil escaping through a gap where the wax has pulled away from the glass, is a leak rather than a separation, and it usually points at adhesion and transit heat. And two common soy faults that look alarming are not oil at all: frosting is a white crystalline bloom of the wax itself (why soy candles frost), and wet spots are patches where the wax has let go of the glass (wet spots and sinkholes). Neither is fixed by changing how you mix.

Why wax holds oil at all — and the two ways it fails

Sugar in tea is the closest kitchen picture. In hot tea, sugar dissolves and disappears; in cold tea, the same spoonful sits at the bottom. Add too much even to hot tea and the excess never goes in. Molten soy wax behaves like the hot tea: it dissolves fragrance oil into a single clear liquid. As the wax cools it crystallises into a network of tiny plates, and the oil has to be trapped in the spaces between them. That network is the cage, and it has a capacity.

So there are exactly two ways this goes wrong. Either the oil never dissolved — the wax was too cool to take it, or it was not stirred long enough to disperse before the wax started to set — or there was more oil than the cage can hold, in which case the excess is pushed out as the wax crystallises and collects wherever it can. Everything in the list of causes below is one of those two failures with a different origin.

The capacity of the cage is what a wax page means by its fragrance load. Luxury Soy Wax Chunks state up to 13% of total wax weight, with 8% as a starting point and 10% recommended, and the page describes the wax as able to hold a 10–13% load without bleeding or seeping. Premium Coconut Soy CSI 468 states up to 13%. Luxury Soy Wax CSI 464 states up to 10%. Eco Soy CSI 400 does not publish a ceiling — ask on WhatsApp before planning a load around it. Those are formulation maximums from the wax manufacturers, not targets.

Timing matters as much as temperature, because the cage does not appear all at once. Soy begins to cloud and thicken well before it is solid, and once that starts, any oil not already dispersed has nowhere even to try to go. This is why a slow, thorough stir immediately after adding the oil does more than a vigorous one later, and why topping a thickening melt up with cold flakes just before pouring is such a reliable way to produce a marbled, oily top.

The six causes, in the order to check them

Two things are worth saying plainly about this list. First, causes one and two are process, not product: they cost nothing to fix and they are where the great majority of separated soy batches come from. Second, none of these is a reason to distrust the wax. A verified CSI buyer describes Eco Soy CSI 400 simply as “good quality wax”, and the same grades that separate in a rushed 65°C pour behave perfectly in the same kitchen a week later with a thermometer on the bench.

Cause one: the oil went in too cool

CSI’s soy pages publish the window. Luxury Soy Wax Chunks: melt to approximately 80–85°C, add dye at full melt, add fragrance at 80–90°C and stir slowly for 2 full minutes, then pour at 75–80°C. Eco Soy CSI 400: add fragrance at 85°C, within 80–90°C, stirring gently for 2 minutes, then pour directly. Luxury Soy Wax CSI 464: the product page gives 80–90°C as the working range for melting and pouring. Premium Coconut Soy CSI 468 is the exception on this bench — its page gives a melting point of 50–53°C and a pour of 58–60°C but does not publish an add temperature, so message CSI rather than borrowing the soy figure.

Two habits cause a cool add without anybody noticing. The first is melting on a stove or in a double boiler, turning the heat off, going to fetch the oil and the scale, and returning four minutes later to wax that has dropped fifteen degrees. The second is an Indian winter or a hill-station kitchen, where a steel pitcher loses heat far faster than the same pitcher in a Chennai May. A CSI customer describes precisely this problem before switching to a melter: the temperature “would drop within seconds”, and “the wax stays warm for a longer time in the melter”. A Mini Electric Wax Melter at ₹2,360.00 holds temperature; a Pen Thermometer at ₹354.00 at least tells you when it has gone.

The opposite error is rarer but real. Wax held far above the window for a long time, or heated repeatedly, is not a way to dissolve oil better — it is a way to lose the light top notes of the fragrance before the candle is even poured. Flashpoint is not the relevant figure here: it is a transport and handling property, not the temperature at which an oil “burns off”. Sandalwood lists a flashpoint of 203°C and Blue Ocean 143°C — both comfortably above an 80–90°C add. The reason not to overheat is scent quality and wax colour, not ignition (best temperature for adding fragrance oil to soy wax; at what temperature to add fragrance).

Causes two and three: the stir and the dose

The stir is not a formality. CSI’s soy pages specify slow stirring for two full minutes, and the word slow is deliberate: fast whipping drags air into the wax and produces bubbles and a pitted top, while a slow, deliberate figure-of-eight that reaches the base and the sides of the pitcher is what actually disperses the oil. Time it. Almost everybody who thinks they stirred for two minutes stirred for about forty seconds. A Bamboo Stirring Stick (₹118.00 for 5, ₹1,062.00 for 50) is fine for a jar-sized melt — a customer calls it “quite helpful In mixing the colours and stirring the wax” — and a CSI Pro Mixer at ₹349.00 is worth having for larger pitchers, used briefly and at low speed.

Scraping matters as much as stirring. Fragrance oil is less dense than wax and will sit at the top if it is not carried down; wax cooling against the metal sides sets first and takes dispersed oil with it. Run a Silicone Spatula (₹175.82) or a Steel Scraper (₹236.00) around the sides and across the base once during the two minutes. In a large melter, that is the difference between an even batch and the last three jars of a run coming out streaky.

The dose is arithmetic, and it is the cause that produces the most uniform-looking failure: every jar in the batch shows the same oily layer at the same time. Recalculate before you blame anything else. Loads are percentages of wax weight on CSI pages, so 200 g of wax at 10% takes 20 g of oil, and a 200 g pour containing 26 g of oil is at 13% — the ceiling for the Chunks and above the ceiling for CSI 464. Two ways of measuring cause most over-doses: calculating the percentage on the finished candle rather than the wax, and measuring oil in millilitres rather than grams.

The oil page sets a limit as well, and where the two differ the lower figure governs. Sandalwood states up to 10% for soy, Dark Vanilla Oud gives 6–10% for candles, Lavender 6–10%, and Fresh Strawberries stops at 6–9%. A wax rated to 13% does not lift any of those (how to make candles smell strong without exceeding the safe fragrance load).

Causes four to six: the oil, the heat history and the order of addition

Not every fragrant liquid is a candle fragrance oil. Reed diffuser bases are formulated to stay fluid and to wick up rattan, attars and alcohol-based perfumes are built for skin, and cosmetic or massage oils are carrier oils with fragrance in them. None of those is formulated to dissolve into wax and be held there, and they separate at loads a candle oil handles comfortably. The test is the product page: a CSI fragrance oil page states a candle usage figure — 6–10%, or up to 10% for soy — and that figure is the sign it has been assessed for wax.

Essential oils are a separate conversation. They are volatile, variable between batches and generally behave differently in wax from a formulated fragrance oil; if you work with them, treat any load figure as something to establish for yourself in a controlled batch rather than borrowing a fragrance-oil number. Blending two fragrance oils does not itself cause separation, but it does create an untested mixture — keep the combined weight inside the lower of the two pages’ ranges and test one blend at a time.

Heat history is cause five, and it catches production makers more than hobbyists. Wax that sits in a melter at working temperature all afternoon, is topped up and reheated twice, and finally poured at four o’clock is a different material from wax melted and poured within twenty minutes. So is a batch poured into a 34°C Ahmedabad kitchen with no air movement, where the wax stays liquid for a long time and the oil has hours in which to drift upwards before the crystal network forms. Neither is a mixing failure; both look exactly like one (how long soy wax should cool before pouring).

Cause six is the order of addition, which the Chunks page settles for you: dye at full melt, fragrance at 80–90°C, stir slowly for two minutes, pour at 75–80°C. Liquid dye added late, or into wax that has begun to cloud, streaks and can carry oil with it into patches — a Candle Dye Liquid is ₹236.00 for 15 gm and a very small quantity is enough. If you use Vybar Additive (₹156.00 for 50 g, ₹3.12 per gram), its page puts it in at 0.5–2% of wax weight — the lower end, 0.5–1%, for soy — added to the melted wax before the fragrance and colour, which is ₹6.24 at 1% of a 200 g batch. Treat it as an option to test in a controlled batch, not as a repair: it will not rescue a pour that went in at 65°C, and single-pour grades like Eco Soy CSI 400 are sold as needing no additives (does Vybar improve hot throw).

Rescuing a separated batch

The economics are usually comfortable. Bringing a 200 g batch from 12% back to 10% costs ₹23.96 of wax; from 15% to 10%, ₹59.90. A full kilogram of CSI 464 poured at 13% needs 300 g of plain wax to come back to its stated 10% — ₹152.22 — against 1 kg of wax and 130 g of oil already in the pot. The rescue almost always costs less than the fragrance you would otherwise throw away, and it produces more candles than you started with, which is worth remembering before you quote a replacement batch to a customer.

Two honest limits. First, a rescue dilutes the scent — a batch brought from 15% to 10% now smells like a 10% candle, which is the point, but tell yourself that before you burn-test it and feel disappointed. Second, if the problem is an oil that was never formulated for wax, no amount of remelting will hold it; that batch becomes wax melts, or unscented wax for future rescues, and the lesson is on the product page you did not read. Repeated remelting is also not free: every cycle costs heat, time and a little fragrance.

The pour that does not separate

Write the batch down before you start: wax grade and pack, wax grams, oil and its page range, oil grams, the load those two produce, the add temperature, the pour temperature, the stir time, the room temperature and the time. Then fill in what actually happened. A batch card takes thirty seconds, and it is the only thing that lets you say in November why a batch poured in September behaved differently from the one before it (how to make soy candles step by step; using CSI soy wax for smooth, crack-free candles).

Indian conditions deserve two specific adjustments. In a hot kitchen in April or May — Delhi, Ahmedabad, Nagpur — a poured jar can stay liquid for a long time, so keep jars out of direct sun and away from the melter’s own warmth, and do not stack or box candles until they are fully firm. In the monsoon, in Mumbai, Kolkata or Kochi, the enemy is moisture rather than heat: keep wax sealed, keep jars dry before pouring, and do not leave an open bag of flakes near a window. Neither condition causes separation on its own, but both lengthen the window in which a marginal batch can go wrong.

Finally, buy the way you pour. A Pen Thermometer at ₹354.00 is the cheapest insurance in this entire article against a batch worth ₹280.66 in materials alone. A Steel Pouring Pitcher 600 ml at ₹708.00 pours without dribbling down the glass, which is a separate problem that looks like a leak. And if your wax cools faster than you can work — a Himachal winter, an air-conditioned studio — a Mini Electric Wax Melter at ₹2,360.00 holds the window for you rather than against you.

Frequently asked questions