Rancidity and oxidation in balms: the chemistry, the smell and the shelf life

How balm oils go rancid, why the smell is crayon or old nuts, which oils fail first, what actually extends shelf life, and how to date a handmade batch honestly.

A balm cannot go mouldy in the ordinary way, because there is no water in it. What it can do is oxidise. Knowing which part of the fatty acid the reaction attacks tells you which levers are worth pulling.

Rancidity is a chemical failure, not a biological one

Two processes are both called rancidity. Hydrolytic rancidity splits triglycerides into free fatty acids, needs water or an enzyme, and gives the soapy note of old coconut or palm fat. Oxidative rancidity, or autoxidation, reacts with oxygen and needs no water at all. In an anhydrous product the second is the one that matters, which is why an unpreserved balm can be safe from microbes and still unusable, and why balms do not need preservatives but do need antioxidants.

The bis-allylic hydrogen, not the double bond count

Almost everything written for makers ranks oils by how unsaturated they are. That is the wrong variable. Autoxidation begins by pulling a hydrogen off a fatty acid, and that abstraction is rate limiting, so what matters is how weakly the weakest hydrogen is held. A bis-allylic methylene is a CH2 flanked by two double bonds, the 1,4-diene motif. Lose a hydrogen there and the radical is delocalised over five carbons, far more stable, and therefore far easier to form, than the radical from a single double bond. The bond dissociation energies differ by about 10 kcal/mol, roughly 77 for a bis-allylic C-H against 87 for an ordinary allylic one.

Oleic acid is unsaturated and has no bis-allylic hydrogen at all, so on a balm's timescale it behaves almost like a saturated fat. That explains why high-oleic oils outlast their unsaturation, why jojoba oil and meadowfoam are exceptionally stable despite carrying double bonds, and why a few percent of rosehip sets the pace for a whole formula. The working index is percent linoleic plus twice percent linolenic, weighted across the blend.

How much faster? It depends what you measure

Published relative rates for the three acids disagree, and the disagreement is not sloppiness: different methods measure different parts of the same reaction.

Take the range, not one number. All agree on the shape: linoleic is one to two orders of magnitude faster than oleic, linolenic two to four times faster again. The ratio 1:100:1200:2500 that circulates in craft sources cannot be traced to a primary source and sits an order of magnitude away from every verified figure, so it is not used here. The ranking holds for bulk oils, which is what a balm is, and not for emulsions.

Autoxidation in three stages

The reaction is a free radical chain with the classic three-stage shape.

1. Initiation. Something abstracts a hydrogen atom, preferentially a bis-allylic one, leaving a carbon-centred radical. Heat, ultraviolet light and trace metal ions such as iron and copper all drive this step.
2. Propagation. The lipid radical takes up oxygen to give a peroxyl radical, which steals a hydrogen from a neighbouring chain, producing a hydroperoxide and a fresh lipid radical that repeats the cycle. One initiation event can consume hundreds of fatty acid molecules, which is why oxidation accelerates: it is autocatalytic.
3. Termination. The chain stops when two radicals combine, or when an antioxidant donates a hydrogen and forms a stable radical of its own. That is what tocopherol does, and it is why an antioxidant is a chain-breaker rather than an oxygen scavenger.

Hydroperoxides are almost odourless. The smell arrives when they decompose: hexanal from linoleic acid reads as cardboard and cut grass, nonanal and decenals from oleic acid as waxy or crayon-like, propanal and its relatives from linolenic acid as green or putty-like. Over a beeswax base, that combination is the old crayon smell people describe. There is usually a quiet induction period first, then a fast decline once the native antioxidants are spent.

Which oils fail first, and roughly when

Rank oils by the index and the order comes out right. Note what is missing below: the shelf lives in months that suppliers publish. Those are estimates with no stated method, temperature, packaging or endpoint, so they belong in a quotation, not a dataset.

Jojoba lasted about nine times longer than rosehip under identical stress, and the ranking tracked polyunsaturate content almost exactly. Jojoba has also been measured at 68.6 to 71.2 hours at 110 C, and 51 to 229 days at 60 C against 1 to 29 days for every other oil tested.

A balm inherits the shortest life in the blend, weighted by how much of it there is: ten percent rosehip in a stable base is not a two year product. The fuller dataset is in the oil comparison table, and the feel trade-offs in carrier oils compared.

The five accelerants, and what to do about each

Ranked by the difference each decision makes to a small batch maker:

Keep the melt under 85 C too, since heat both initiates oxidation and destroys native tocopherol. The order to hold in your head: bis-allylic content, storage temperature and light, headspace and packaging, native tocopherol, then trace metals. Asking a supplier for the pressing date or a certificate of analysis costs nothing and is covered in sourcing ingredients. Dosing decisions for the additives are in vitamin E and antioxidants.

The best storage evidence available is a 270-day real-time study of hemp seed oil (Bonazza et al., Molecules, 2024). In polypropylene at 25 C in diffused light it reached 40 meq/kg of peroxide by day 150, nearly three times the Codex limit for a cold pressed oil. The same oil in amber glass wrapped in foil, at 10 C in the dark, stayed under 10 meq/kg for the full 270 days. Packaging and temperature beat oil choice, which is the packaging guide argument in one experiment.

How to judge a suspect balm

Sensory assessment works if you do it the same way each time. Warm a little on the back of the hand, because volatiles need warmth to reach the nose, then sniff, wait and sniff again, comparing against a retained sample frozen at production.

Rancidity signals: a sharp nutty, putty or crayon note that grows on the second sniff, a bitter aftertaste in a lip product, a tacky feel the formula did not have when new (oxidation polymerises oils), and darkening at the exposed surface. Not rancidity: the honey note of yellow beeswax, the smell of unrefined shea, colour drift from an infusion, or a surface change that is sweating or bloom. Full diagnosis in rancid balm.

Peroxide value, and why one reading tells you nothing

Peroxide value is an iodometric titration in milliequivalents of peroxide oxygen per kilogram of oil (AOCS Cd 8b-90, ISO 3960). It is the number on a supplier certificate, and it is misread in two ways.

Legal is not fresh. Codex limits are 10 meq/kg for refined vegetable oils, 15 for cold pressed and virgin, under 20 for virgin olive and under 5 for refined olive. Genuinely fresh oil from a careful supplier tests at 0.1 to 1 meq/kg. A certificate showing 8 is fully compliant and describes an oil that has already spent much of its life before reaching your bench. Ask for the measured figure and its date, not a statement that the oil meets Codex.

The value is not monotonic. Hydroperoxides are intermediates: they form, accumulate, then decompose into the aldehydes you can smell. A badly rancid oil can therefore return a low or falling peroxide value. A single reading is close to uninterpretable, and means something only plotted over time and paired with a secondary measure such as p-anisidine value. Totox, twice the peroxide value plus the anisidine value, is quoted with thresholds of under 10 good and over 20 advanced, but those are trade convention, not a standard. Acid value, which tracks free fatty acids, rises throughout and is the better single number for an oil of unknown age.

The compounds that matter to skin are the secondary ones. Aldehydes are implicated in cytotoxicity at parts per million, which is why "it only smells slightly off" is a poor reason to keep using a lip product.

What rancid balm does, and does not do, to skin

Be proportionate. Rancid oil is not acutely toxic and nobody needs urgent care for using an old lip balm. What oxidised lipids reliably do is irritate: aldehydes and hydroperoxides are more reactive with skin proteins than the triglycerides they came from, so the balm stings on cracked lips and no longer does the barrier job you bought it for.

The substantive concern is sensitisation. Several fragrance terpenes are weak allergens when fresh and considerably stronger after air oxidation, limonene and linalool being the best-documented cases. That is an argument for keeping essential oils in balms low and not persisting with a scented product that smells stale, not for panic over a tin you used once.

Dating a handmade batch honestly

An expiry date on a handmade balm is an estimate and should be presented as one. No test reads a date off a finished product. What you can do is build the estimate defensibly.

1. Start from the highest-index oil in the blend, counting from its pressing date, not from the day you opened the bottle, and subtract the months it has already spent in your workshop. That subtraction is the correction most people skip.
2. Allow a little for antioxidant and good packaging, round down, and record the result with the lot number of every input.
3. Keep retained samples and smell them on a schedule. A year of that evidence beats any published table, including this one. The method is in shelf life testing.

Best before dates and the PAO symbol

In the EU and UK a cosmetic with a minimum durability of 30 months or less must carry that date, the hourglass symbol or "best before end of"; above 30 months it carries the open-jar period-after-opening symbol, which answers a different question about life once opened. Most handmade balms sit under 30 months, so a best-before date is the honest choice, and a 12M jar symbol on a product with an 18 month oxidative life is arguably misleading. The United States requires no expiration dating for cosmetics, but a voluntary best-before is what a buyer looks for when reading a balm label.

Practical stock rotation

Write two dates on everything: arrival date and lot on raw materials, batch code and best-before on finished stock. Buy the high-index oils (hemp, rosehip, grapeseed, standard sunflower) in the smallest pack you can tolerate, cold, and only when a batch is planned. A litre of rosehip bought cheaply in January and used in June has spent most of its life as inventory, and every argument on this page is lost before the balm is made.

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