The bottle you opened is no longer the same material: oxidation, storage, and whether refrigerating and decanting actually help

A bottle of lemon oil opened two years ago still carries the same label, but its contents have changed. The smell may dull or grow unfamiliar edges, and some oxidation products sensitise skin more readily than the fresh material.

Pure linalool does not sensitise. Oxidised linalool does.

In 2004 Sköld, Börje, Harambasic and Karlberg published work in Chemical Research in Toxicology that joins the whole causal chain together.

Linalool is unsaturated and autoxidises on exposure to air. They isolated and identified the oxidation products (two hydroperoxides plus several secondary products) and quantified them over time by GC and HPLC. One hydroperoxide reached 15% of an oxidised sample.

They then measured sensitising potential with the local lymph node assay:

Pure linalool showed no sensitizing potential. The air-exposed samples of linalool produced clearly positive responses, and the hydroperoxides were the strongest allergens of the tested oxidation products.

Their closing sentence is the one to remember:

The sensitizing potential differs with the composition of the oxidation mixture and thus with the air exposure time.

Linalool is not the allergen. Your old bottle is. And how much of an allergen it is depends on how long it has been open.

This is not a laboratory curiosity

In 2023 Schubert and colleagues, on behalf of the IVDK, published patch test results from consecutive patients in Contact Dermatitis. Between July 2018 and December 2020, 5,511 consecutive patients were patch tested with limonene hydroperoxides (0.3% in petrolatum) and linalool hydroperoxides (1% in petrolatum).

	Positive reactions
Limonene hydroperoxides	170 (3.1%)
Linalool hydroperoxides	483 (8.8%)

Be clear about the population. These are dermatology patients undergoing patch testing, not the general public — being referred for a patch test is itself a selection. The IFRA article cited Diepgen and colleagues, who found roughly 1.8% and 1.9% for Fragrance Mix I and II in the general population; the two sets of figures cannot be placed side by side.

Even allowing for that bias, 8.8% is high. And the crucial point is what they react to: not linalool, but oxidised linalool. What reached their skin was product that had been sitting around.

The IFRA article made the point that sensitisation is cumulative and irreversible. Here is the part that lands on you: nobody handles these materials more often than you do, and a three-year-old bottle of linalool is more dangerous than a fresh one.

Which materials degrade fastest, in our data

The database carries two relevant fields: storage (a storage recommendation) and shelf_life. The first has only four distinct values, the strictest being "store under nitrogen".

Which 40 materials call for nitrogen? The answer reads like a textbook of autoxidation:

Class	Examples
Fatty aldehydes	octanal (C-8), nonanal (C-9), decanal (C-10), dodecanal (C-12), 2-methyl undecanal, 9-decenal
Unsaturated aldehydes	(E,Z)-2,6-nonadienal, (Z)-6-nonenal, cinnamaldehyde, α-amyl and α-hexyl cinnamaldehyde, phenyl acetaldehyde
Citrus oils	lemon oil (four variants), sweet orange peel oil (two), mandarin oil (three), grapefruit isolate
Monoterpenes	dextro-limonene, dipentene, delta-3-carene
Others	linalool and laevo-linalool, benzyl alcohol, cinnamyl alcohol, ho leaf oil, dihydro-β-ionone

Aldehydes and terpenes. An aldehyde has an abstractable hydrogen next to the carbonyl; a terpene has double bonds. Both are textbook starting points for autoxidation. And sitting right there in the list is linalool itself, the subject of both studies above.

The storage recommendations, on their own, finger the same molecules.

The shelf life figures

Of the 21,494 materials with supply information, only 1,251 (5.8%) carry a shelf_life value. By family:

Odour family	n	Median recommended shelf life	Share at 12 months or less
Citrus	96	12 months	52%
Fruity	77	12 months	51%
Spicy	63	24 months	43%
Floral	183	24 months	32%
Woody	73	24 months	29%
Herbal	77	24 months	27%
Green	57	24 months	26%
Balsamic	40	24 months	8%

Citrus and fruity carry half the shelf life of the other families, and almost nothing balsamic is short-lived. That matches the chemistry above: citrus oils are dominated by monoterpenes, balsamics by resins and larger molecules.

Three things about how to read this field.

1. It is a floor, not an expiry. Every raw value reads "X month(s) or longer if stored properly". It says "at least this long", not "unusable after".

2. The values are a convention, not a measurement. 96% of the 1,251 values sit on four round numbers — 6, 12, 24 and 36. This is a graded recommendation, not a per-material measurement — the same phenomenon as the rounding in the substantivity field from the evaporation curve article.

3. Coverage is only 5.8%. Every row above carries its denominator; the smallest, balsamic, is 40 materials. This is directional evidence.

We also excluded the group with no primary odour family (n=277, median 12 months) — mostly non-fragrance chemicals, which would distort the comparison.

Do refrigerating and decanting actually help?

Yes, and you can derive it from the mechanism rather than guessing.

Autoxidation needs four things: oxygen, time, heat and light (plus metal-ion catalysis). Every countermeasure removes one of them:

Practice	Removes	Effect
Decant into a small bottle filled to the top	Oxygen	The most effective, and free. A 30 mL bottle with a fifth left holds 24 mL of air
Close it tightly, open it rarely	Oxygen	Every opening exchanges the headspace
Refrigerate	Heat	Effective, but do not freeze, and let it reach room temperature before opening or condensation gets in
Dark glass, kept out of light	Light	Light initiates radicals. Plastic has leaching problems of its own — use glass
Nitrogen blanket	Oxygen	The professional version, and exactly what the database recommends for its 40 most fragile materials

Decanting is the highest-return habit here. The volume of headspace air and the exposed surface area directly set the oxidation rate, and decanting shrinks both at once. Split an opened bottle into three small ones and two of them return to near-unopened condition. I decant the day a bottle arrives now; the ones I kept putting off were always the first to turn.

As for refrigeration: it slows the reaction, it does not stop it. A refrigerated three-year-old bottle is still a three-year-old bottle.

How long does an opened bottle last?

Months are the wrong unit for this question. The useful answer is an action: put the bottle next to your retain.

The batch variation article asked you to keep 1 to 2 mL of every batch, sealed and refrigerated. This is where that pays off: smell the working bottle beside the retain. It is the only way to answer "has it gone off". Oxidation is gradual, and if you smell a fraction of the change every day, you will never notice it.

Typical signs of oxidation:

Citrus is the easiest to catch because the contrast is so strong: the bright top leaves first, then turpentine-like, waxy, stale notes move in.

Aldehydes go rancid and fatty. C-8 through C-12 are bright when fresh and smell like stale nuts or old oil when they are not.

The most insidious class is the terpene alcohols such as linalool. The fresh floral top fades, a woody, camphoraceous edge creeps in, and at no point does the bottle smell bad. It just smells ordinary.

Your eyes help too: darkening, cloudiness, thickening. Citrus oils show this most clearly.

There is a trap here, the one from the individual differences article. What you are judging is "has this changed", a question one person can measure reliably: same person, side by side, you are your own control. "Is it still usable" is a different question, and the answer depends on what you want it for.

Two more fields for the record

The evaluation note, the purchase record, and now two more:

Field	Why
Date opened	Shelf life runs from opening, not from delivery. If it is not on the bottle, in six months you are guessing
Date and result of the last comparison against the retain	Without it you will not know how long it has been since you checked

In practice: decant, date and set aside a retain the day it arrives. All three together take under five minutes, and they decide whether that bottle is a trustworthy piece of data or a question mark you will be afraid to use in six months.

---

Materials keep changing after you buy them. The batch article said the same name stops being the same material, because batches move. This one says the same bottle stops being the same material, because time moves. Both have the same remedy, and it is the one you already know: keep a retain, write the date, compare side by side.

References

M. Sköld, A. Börje, E. Harambasic & A.-T. Karlberg, Contact allergens formed on air exposure of linalool, Chemical Research in Toxicology, 17(12), 1697–1705 (2004). PMID 15606147

S. Schubert et al. (IVDK), Patch testing hydroperoxides of limonene and linalool in consecutive patients, Contact Dermatitis, 89(2), 85–94 (2023). PMID 37177844

---

Next comes the close of this series: joining the methods from all fifteen articles into a single working process, from receiving a new material to a formula that can be reproduced, modified, and handed to someone else.