The fragrance was there on the day of the pour and gone by the third week

Top notes evaporate from a hot melt and the wax absorbs the rest. Adding temperatures, fixative levels and the packaging that keeps a balm smelling of something.

A balm that smelled of grapefruit on the day it was poured and of warm wax three weeks later has not spoiled. Nothing has oxidised, nothing has separated, and the base oils are exactly what they were. You have lost the lightest fraction of the fragrance, and a good part of it left before the balm ever set.

• Add fragrance below 45-50 C
• Citrus tops: days to three weeks
• Fixatives: 5-15% of the load
• Lip products are IFRA Category 1

Three faults get called "the smell went"

Before changing anything, decide which of three you actually have, because two of them have nothing to do with each other.

It faded. The balm smells of less and less of what it smelled of, and eventually of the base: wax, butter, a faint nuttiness. Nothing new has appeared. That is volatile loss, and it is this page.

It turned. The balm smells of something else: turpentine, paint, varnish, old cardboard. The intensity may be unchanged. That is oxidation of the terpenes, it raises the sensitising potential of the product, and it is dealt with under scent has turned. The two overlap because the molecules that evaporate first are largely the molecules that oxidise first, so a nine month old citrus balm can be both faded and turned at once.

It was never right. The balm has smelled thin since day one, which is a dosing question rather than a stability one. The mirror image, a batch that came out far louder than intended, is under balm smells too strong.

Volatility classes, and the timetable they set

Perfumery sorts materials into top, heart and base notes by how long they last on a blotter. Underneath that convention is a physical property: vapour pressure, the tendency of a molecule to leave a liquid for the air above it. It is tabulated far less often than boiling point, but within a chemical class the two track each other closely enough for boiling point to work as a proxy.

Two consequences fall straight out of the table. A single-oil sweet orange balm has nothing in it but the first row, so there is no version of that product that still smells of orange at six months. And a blend does not fade, it distorts: the tops leave, the base stays, and the balm progressively becomes the least interesting part of the accord you designed. A lavender balm at four months is not a weaker lavender balm, it is a camphoraceous, slightly woody balm with the sweetness gone.

What you lose at the pot, before anything is poured

This is the single largest and most avoidable loss, and almost every complaint about weak scent starts here. Vapour pressure rises steeply with temperature: limonene sits at roughly 1.5 to 2 mmHg at 25 C, and by 70 C it is higher by something close to an order of magnitude. Stir an open jug of 70 C fat with a citrus oil freshly poured into it and you are running a small, inefficient distillation.

The loss is not measured anywhere for craft batches, and it depends on the temperature, the open surface area, how long the jug stands uncovered, and how vigorously you stir. As a working estimate, expect somewhere between a fifth and a half of the most volatile fraction of a citrus addition to be gone before the balm sets. You can bound it yourself in one afternoon: weigh the covered jug immediately after the addition, weigh it again at the moment of pouring, and the difference is almost entirely fragrance, because nothing else in an anhydrous balm is volatile at that temperature.

The arithmetic matters because the label does not change. A 100 g batch at 1.0 percent carries 1.0 g of oil. If the addition is a sweet orange oil that is roughly 90 percent limonene and you lose a third of that fraction to a hot open melt, the wearer receives about 0.7 g. The declaration, the allergen calculation and your allergen labelling all still describe 1.0 g, because that is what went in. Under-dosing by evaporation does not buy you regulatory headroom, it only buys you a weak balm.

The fix is a temperature, not a technique. Melt the fats properly at 70 to 80 C as set out in pour temperatures, then take the jug off the heat and let it come down to 45 to 50 C (113 to 122 F) before the fragrance goes anywhere near it. Stir for fifteen seconds, keep the jug covered between stirs, and pour. For most makers this one change is the whole answer, and it costs nothing.

Why the wax holds on to half of what survives

An aromatic molecule in a balm is not sitting on the surface waiting to leave. It is dissolved in fat, and to reach a nose it has to escape a lipophilic solvent first. How strongly it is held depends on how fat-loving it is, which is exactly the property that also makes it a base note. Sesquiterpene alcohols, resinoids and musks partition heavily into the lipid phase and release slowly. Small oxygenated monoterpenes are far less comfortable there and leave readily. Monoterpene hydrocarbons barely notice the fat at all.

The practical result is that a wax base is not a neutral container. It systematically favours the heavy end of your blend, which is why the same accord smells balanced in alcohol and base-heavy in a tin. The solid perfume page treats this as a design problem rather than a fault, and the correction is the same in a balm: over-weight the top notes relative to what you would use in a spray, typically pushing the top fraction to 30 percent or more of the blend rather than the conventional 20 to 25.

The same partition argument explains a difference makers notice and rarely explain. A high-castor lip balm holds flavour and fragrance longer than a squalane-heavy one, because a viscous, polar oil phase releases volatiles more slowly than a thin, mobile one. That is also why the two questions in lip balm flavour and sweeteners and this page keep meeting: on a lip product, perceived flavour and perceived scent fade on the same clock and for the same reason. Background on what dissolves in what is under solubility in anhydrous formulas.

Diagnose it in an hour with two tins

The only reliable measurement is a comparison, because nobody can remember a smell accurately for three weeks. This is why retained samples exist.

1. Open the sealed retain from the same batch. Same pour, same fragrance, never opened, stored cool and dark. If it smells right and the shelf sample does not, the fragrance was dosed correctly and the loss happened in the container. Go to packaging.
2. Smell them side by side, warm. Rub a smear of each on a different hand, wait thirty seconds, and compare. Volatiles read very differently at 32 C skin temperature than at 20 C in a tin.
3. If the retain is weak too, the loss happened at the pot. The fragrance either never went in at the level you think, or it left during the melt. Check the addition temperature in your batch records. If you did not record one, that is the finding.
4. Blotter the neat oil. One drop on paper, wait a minute, smell. If it reads sharp, flat or turpentine-like rather than simply weak, you have a turned oil rather than a fading balm and the whole diagnosis changes.
5. Get a second nose. Olfactory adaptation is fast and complete: a few minutes in a room that smells of bergamot and you stop smelling bergamot. A maker who has spent a morning with the product is the worst available judge of it. Protocols are in sensory testing.

Fixatives: what they do and what they do not

A fixative is a low-volatility material added to make a blend last. Two mechanisms are commonly claimed for them and only one is worth much. The theoretical one is that a heavy, poorly volatile component lowers the mole fraction and therefore the partial pressure of the volatiles around it, which is true and small. The one that matters in practice is simpler: the fixative is still there in week six when the tops have gone, so the balm still smells of something deliberate rather than of wax. It changes what the dry-down is, more than it changes when the dry-down starts.

Two honest limits. First, a fixative cannot hold a molecule that has already left, so it does nothing for the first three weeks of a citrus balm's life. Second, none of these materials is safety-neutral. Benzoin in particular carries several of the fragrance allergens that have to be declared above 0.001 percent in a leave-on product, so a resin added to solve a sensory problem creates a labelling one. The whole picture is in essential oils in balms.

The other lever, and the one professionals reach for first, is to pick a longer-lasting material with a similar smell rather than to prop up a short one. Litsea cubeba or lemon myrtle read lemony and last longer than expressed lemon, though both are citral-rich and capped at 0.11 percent in lip products. A folded or terpeneless citrus oil has had much of the limonene stripped out, which makes it stronger per gram and slower to disappear. And a compounded fragrance oil is built by a perfumer to survive a matrix, which is one of the real arguments in essential oils versus fragrance oils.

Headspace and closures decide the rest

Once the balm is set, everything that leaves it goes through the gap around the lid. A friction-fit slip lid on a slider tin is not a seal, it is a press fit with a continuous path to the air, and a shallow tin that is half used has a large open surface behind it. A screw closure with a liner, or a capped twist-up tube presenting one small face, loses a fraction of that.

Fill level is part of the same variable. A 15 mL jar filled to 9 mL has 40 percent of its volume in air from the day it was capped, and that air saturates with whatever is most volatile in the product. Choose a container close to the fill you intend rather than a generous one, which is the same argument made from the oxidation side in the packaging guide. The trade-offs between the two most common lip formats are set out in lip balm tube versus tin, and how a customer stores the thing afterwards is covered in storing balms at home.

Rescuing the batch in front of you

Remelting is safe here. The product is anhydrous, so there is no microbiological consequence to a second heat cycle, and nothing in the fat phase is damaged by a short hold at 50 to 60 C. The only real cost is that a second melt drives off more of whatever volatiles are left, so redose rather than simply reheat and hope.

1. Work out the current dose first. Take the original percentage, not what you think is in there now. You are topping up an unknown, so the total on the label has to be the sum of both additions.
2. Check the ceiling before you weigh anything. Run the new total through the dilution calculator against the IFRA Standard for each constituent in your product category. Lip products are Category 1, and the summation rule means two materials each at 60 percent of their own limit is already a failure.
3. Melt the whole batch at the lowest temperature that liquefies it. Typically 50 to 60 C for a tin balm, in a gentle water bath, covered. Do not return to 75 C, because that undoes the point of the exercise.
4. Cool to 45 to 50 C and add the top-up. Weigh it to 0.01 g. Never count drops: a drop runs from about 20 to 40 mg depending on the oil and the reducer.
5. Stir fifteen seconds, covered, and pour immediately. Every extra minute the open jug stands is dose leaving.
6. Cap as soon as it is solid. The surface loses volatiles for as long as it is warm.
7. Retain two tins and re-judge at four weeks. The point of the rescue is a product that still smells right in a month, not one that smells right on Thursday.

Two cases where the rescue is not available. If the batch has already been sold, redosing is not an option and neither is recalling on a sensory grumble; note it against the batch and change the process for the next one. And if the formula relies on a hard wax to hold its shape, a 50 C melt will not liquefy it and a 75 C melt will cost you most of the addition, which is the conflict flagged in the warning above.

Preventing it in the next batch

One formulation note that sits outside the fragrance itself. A carrier that stays odourless for the life of the product is doing quiet work here, because a base that develops its own crayon and cardboard notes competes with whatever scent survives. Jojoba is the usual choice for exactly this reason: the most stable scent in a balm is often the one carried by the most stable base. Confirm your own numbers with retained samples rather than assumption, using the method in shelf life testing.

The limit nobody can formulate around

There is an honest ceiling on this, and it is worth saying rather than selling around. A natural citrus fragrance in an anhydrous balm cannot be made to last twelve months at the intensity it had on day one. The molecules that smell of citrus are the molecules that leave, and every lever above changes the slope of the curve rather than removing it: cooler dosing recovers what you were throwing away, better closures slow the exit, fixatives replace the ending with a different one. None of them makes limonene less volatile.

So this is a design decision, taken before the first melt. Accept a short scent life and say so, with a period-after-opening that reflects it. Or build the accord from materials that survive a wax matrix, in which case it will read woody, resinous or vanillic rather than fresh. Or use a compounded fragrance oil with an IFRA certificate, which is a perfumer solving the same problem with a wider palette than you have. A fourth option is genuinely respectable: leave the fragrance out, use an unrefined butter for its own scent, and remove an entire category of complaint, cost and allergen declaration from the product.

Reviewed and updated 6 September 2026. Spotted an error? Tell us and we will fix and log it.