Terpene oxidation products and what they mean

Do you know this? A harvest that smelled fresh, of bright citrus and pine needles, smells rather flat, hay-like, or slightly harsh a few months later. This is rarely just imagination – it's chemistry.

Terpenes are small molecules with one or more carbon double bonds (C=C). These double bonds are exactly what's chemically vulnerable: oxygen from the air, UV light, and heat react with them and convert them into other molecules – alcohols, ketones, aldehydes, epoxides, and hydroperoxides. In the scientific literature, these breakdown products are collectively referred to as terpene oxides.

This reaction is called autoxidation, because it happens by itself as soon as terpenes come into contact with atmospheric oxygen – with no bacteria or mold involved at all. That's exactly why a harvest can smell 'bad' without anything being contaminated or moldy.

The most detailed study on this to date comes from Raeber and colleagues (2025, Phytochemical Analysis, ETH Zürich). The team exposed 29 individual terpenes to UV light and heat over 28 days and regularly measured how much of each remained.

The most striking result: even chemically related terpenes break down at completely different rates. α-Terpinene was completely gone within 72 hours under UV exposure. Myrcene took, depending on the condition, between roughly 96 hours (under UV) and about two weeks (in whole, dried bud). β-Caryophyllene was nearly completely broken down after around 17 days under UV.

At the other end of the scale was α-pinene: after 18 days, 94% of the starting amount was still detectable – strikingly stable. Limonene also showed above-average stability, though not quite at the level of pinene.

One detail that's practically relevant for you: whole, dried bud protected the terpenes noticeably better than isolated, dissolved terpenes in the lab test. The plant matrix – trichomes, cell structure, surrounding compounds – apparently slows down the reaction. Pure terpene solutions (as found in some extracts or isolates) are therefore tendentially even more susceptible than intact bud.

One molecule shows up across several breakdown pathways: p-cymene. It doesn't primarily form in the plant, but as a shared end product of various breakdown pathways – among other things, the study was able to confirm that the breakdown of myrcene produces not only p-cymene but also α-pinene and β-pinene.

Because p-cymene is formed this way from multiple starting substances, its appearance in a terpene profile serves as a kind of aging signal: the more p-cymene is detectable relative to the original terpenes, the more oxidation has presumably already taken place.

For some of the best-known cannabis terpenes, the most important breakdown products are now well documented:

Limonene oxidizes into limonene oxide, carvone, and limonene hydroperoxide – this conversion is well established, especially from fragrance and allergy research, since limonene also occurs in many perfumes and cleaning products.

β-Caryophyllene turns into caryophyllene oxide – the molecule that, incidentally, drug-detection dogs also pick up on when searching for cannabis, because unlike THC it isn't so volatile that it quickly becomes undetectable.

α-Humulene forms, among other things, humulene epoxide II.

For pinene, general terpene chemistry describes products such as verbenone, pinene oxide, or camphor-like compounds – though the data situation specifically for cannabis is thinner here, which is why this attribution should be read a bit more cautiously than for limonene or caryophyllene.

Aromatically, oxidation first becomes noticeable in the light, volatile monoterpenes – exactly the molecules responsible for bright citrus and pine notes. They often break down first.

What remains is a profile in which the heavier sesquiterpenes and the newly formed oxidation products carry relatively more weight. This explains why aged product often smells flatter, heavier, or slightly hay-like instead of fresh and lively – the 'bright' top notes have simply disappeared first.

Accuracy is worth more than alarmism here. At least one published clinical case report exists of a perioral contact dermatitis (skin irritation around the mouth) in a person who vaped cannabis – traced back to limonene hydroperoxide, an oxidation product of limonene, not to limonene itself.

This is a documented single case, not evidence of a widespread risk. The distinction matters: fresh, unoxidized caryophyllene is not considered an allergen. Dermatological literature also describes caryophyllene oxide itself as a comparatively weak and rare sensitizer. What acts as a sensitizer is primarily oxidized or aged forms and especially short-lived hydroperoxide intermediates – not the fresh terpene in the bud.

In practice, this means one more, understandable reason not to store product open or exposed to light and heat for longer than necessary – without this becoming a known, common health risk.

A methodological point from the study is also relevant to you as a reader of certificates of analysis (COAs): Raeber et al. explicitly recommend stating terpene values as quantitative concentration (e.g. mg per gram) instead of as a normalized percentage.

The reason: when one terpene degrades, the percentage shares of all other terpenes automatically shift upward – even if their absolute amount stayed the same. A pure percentage value can visually mask real terpene losses and look like a mere 'profile shift', even though substance was actually lost. So this is a known pitfall even in professional analytics.

All the chemistry described here is essentially the explanation for why the usual storage recommendations work: cool, dark, airtight, and with as little air space in the container as possible slow down exactly the three factors that drive oxidation – heat, light, and oxygen contact. You can find concrete storage tips in the article Storage and avoiding terpene loss.