Hemp seed oil: green, fragile, and nothing to do with CBD

Hemp seed oil runs 50 to 60 percent linoleic and 15 to 20 percent alpha linolenic with an iodine value of 140 to 170. Use at 2 to 15 percent and a 6 to 9 month life.

Hemp seed oil has the most attractive fatty acid profile of any cheap cosmetic oil and one of the shortest usable lives, and the second fact is a direct consequence of the first. It is also the ingredient most often confused with a completely different material sold at fifty times the price. This page separates the seed oil from the extract, puts numbers on the stability problem, and says where a green, grassy, fast ageing oil still earns a place in a balm.

• INCI: Cannabis Sativa Seed Oil
• Bis-allylic index 89
• Iodine value 140-170
• 6-9 months
• No meaningful CBD

What is actually in the oil

Hemp seed oil is cold pressed from the seed of Cannabis sativa L. It is a conventional triglyceride oil, unusual only in how much of its fatty acid content is polyunsaturated: around three quarters of the chains carry two or more double bonds, and it is one of the few common cosmetic oils to contain meaningful gamma-linolenic acid alongside a high alpha-linolenic content.

A commonly used single set of figures, and the one carried in the oil comparison table, is 13.1 percent oleic, 56.8 linoleic and 16.2 alpha-linolenic. The tocopherol content is worth noting: 500 to 1000 mg/kg is a respectable native antioxidant load, and roughly 85 percent of it is gamma-tocopherol, which is a better chain breaking antioxidant in a bulk oil than the alpha form. Hemp seed oil is not unstable because it is poorly protected. It is unstable because of what it is made of.

Iodine value, bis-allylic count, and what each one predicts

Two numbers circulate for hemp stability and they measure different things. The iodine value, grams of iodine reacting with 100 g of fat by the Wijs method (ISO 3961 or AOCS Cd 1d-92), counts double bonds. It is quoted for hemp at 140 to 170. That figure comes from trade specifications and published analyses, not from a compositional standard, because hemp is not one of the oils named in Codex CXS 210-1999. Ask a supplier for a measured value with the method stated if you need one for a specification.

The number that predicts rancidity is the bis-allylic count, the CH2 groups sitting between two double bonds, because that is where autoxidation abstracts its first hydrogen. The house index is linoleic percent plus twice linolenic percent, and hemp scores about 89, which puts it in the shortest lived band of common oils. The reasoning behind the index is in rancidity and oxidation. Hemp is one of the cases where that shorthand undercounts: gamma-linolenic contributes two sites per chain and stearidonic three, and neither is in the formula. Counted properly, 100 hemp fatty acid chains carry roughly 98 bis-allylic positions, close to three per triglyceride molecule.

Meadowfoam is the row that shows why iodine value alone misleads. It carries plenty of double bonds and effectively no bis-allylic hydrogens, because its dienes are not methylene interrupted, and it is one of the most stable oils in cosmetic use. Use the iodine value as an identity check and the bis-allylic count as the stability prediction.

Hemp seed oil contains no meaningful cannabinoids

Cannabinoids are produced in glandular trichomes on the flowering tops and bracts of the plant. The seed does not make them. Whatever turns up in cold pressed seed oil arrives as surface contamination from resin transferred onto the hull during harvest and threshing, and it is measured in milligrams per kilogram, not percent. Regulators have set limits at that scale: in the European Union, Commission Regulation 2022/1393 fixes a maximum level of 7.5 mg/kg of delta-9-THC plus its acid form in hemp seed oil sold as food. A limit expressed in single figures of milligrams per kilogram is not a source of pharmacological activity, and it says nothing at all about cannabidiol, which is not the regulated analyte.

In the United States, the 2018 Farm Bill defines hemp as Cannabis sativa with no more than 0.3 percent delta-9-THC on a dry weight basis, and the Food and Drug Administration issued no questions letters in 2018 for GRAS notices covering hulled hemp seed, hemp seed protein powder and hemp seed oil as foods. Cosmetic status is separate again: Cannabis Sativa Seed Oil is a long established cosmetic ingredient with its own entry in the European Commission's CosIng database, while extracts of the flowering parts and isolated cannabidiol sit under an entirely different and much less settled set of rules.

Chlorophyll, the green cast, and why it fades

Unrefined hemp seed oil is a deep bottle green shading to olive, and the colour is chlorophyll and its magnesium free breakdown products, the pheophytins, carried over from the seed coat. Cold pressed oils typically carry tens of milligrams per kilogram of total chlorophyll pigments, enough to dominate the appearance of anything they go into.

The colour is not cosmetically neutral, for two reasons. The first is appearance. At 10 percent in a pale shea and beeswax balm the result reads olive rather than cream, and over yellow beeswax it turns khaki. In a tinted product it shifts every pigment toward the cool side, which matters when you are matching shades. The second reason is chemistry: chlorophyll is a photosensitiser. Under light it promotes ordinary triplet oxygen to singlet oxygen, which attacks double bonds directly without needing a radical initiator or an induction period. A green oil in a clear jar on a windowsill is running two oxidation mechanisms at once, and only one of them is slowed by a tocopherol.

The green fades over weeks to months in the light, to a dull yellow brown. That fading is not the product improving. It is the pigment being bleached by the same process that is consuming the oil around it, so a balm that has lost its green cast is a balm that has been photo-oxidised, and colour loss in a hemp containing formula should be read the same way as the faults in tinted balm colour fading. Refined hemp oil is pale gold, has had most of the chlorophyll removed with it, and is the sensible choice for lip and pale tinted products. It gives up part of the tocopherol load and the unrefined description in exchange.

Use levels and the blend arithmetic

Two to 15 percent is the working window, and where you sit in it is decided by smell and colour rather than by skin performance. In a lip product, 2 to 5 percent of a refined grade, because no flavour covers grassy at higher levels. In a face or body balm, 5 to 15 percent, where the green reads as a feature. In a stock line that sits in a market box for six months, none, whatever the profile promises.

The useful discipline is to work out what hemp does to the whole blend rather than to argue about the oil in isolation. Weight the index by percentage and read the result.

That table contains the whole strategy. Ten percent hemp in a stable base lands the blend at 18, which is a perfectly ordinary balm, while the same 10 percent in standard sunflower barely registers against a base that was already fragile. Pair hemp with meadowfoam seed oil or a high oleic oil and you buy months. Pair it with grapeseed or rosehip and you have simply added a second fragile oil. The arithmetic for any blend you like is in the oil blend calculator.

The smell, and the timeline it follows

Fresh cold pressed hemp seed oil smells grassy and green, with a nutty, faintly bitter finish that some people like a great deal and some cannot get past. That smell is the baseline, and its deterioration is a better shelf life indicator than any number a home maker can measure.

In a part full clear bottle at room temperature the sequence runs roughly like this. For the first two to three months the green top note stays. Between three and four months it flattens: not unpleasant, just dull, and this is the stage most people miss because nothing smells wrong. By five to six months a putty or linseed paint note appears, the characteristic signature of alpha-linolenic breakdown, and by seven to nine months it goes cardboard and faintly fishy. Cold and dark, in a full sealed amber bottle, the same oil is still at the flat stage at nine months.

The chemistry behind the sequence is that hydroperoxides, the primary oxidation products a peroxide value measures, are odourless. The smell arrives only when they decompose into aldehydes: hexanal from the linoleic fraction, propanal and 2,4-heptadienal from the linolenic. So a peroxide value taken at month three can look acceptable while the oil is already committed, and the smell can arrive suddenly. What that looks like in a finished product is in rancid balm.

Storage, nitrogen, and the cold storage arithmetic

Headspace does more damage than most makers expect, and it can be quantified. The air above the oil in a half empty one litre bottle is about 500 mL, containing roughly 105 mL of oxygen, which is about 4.3 millimoles. If all of that ends up as hydroperoxide in the 500 g of oil below it, that alone is around 17 milliequivalents per kilogram of peroxide value, above the Codex limit of 15 for a cold pressed oil, from the air in the bottle and nothing else. Decanting into small bottles filled to the neck is therefore not fussiness, it is the largest single lever you have.

Nitrogen flushed packaging does the same job at the supplier's end by displacing that oxygen before sealing. It is worth paying for on hemp, rosehip and the other high index oils, and worth nothing if you then leave the bottle open on the bench. Ask for the pressing date rather than the best before date, and for a peroxide value with the method named, as set out in sourcing ingredients.

Temperature is the second lever, and the conventional planning assumption is that reaction rate roughly doubles for each 10 C rise, the Q10 assumption behind ASTM F1980 and behind most oven ageing protocols. On that basis a fridge at 5 C rather than a 21 C cupboard is a 16 C drop, about a threefold extension. A freezer looks like fifteenfold on the same arithmetic, which is where the assumption should be treated with suspicion: extrapolating a single Q10 across 40 degrees, through the point where fats begin to crystallise, is not a measurement. The limits of that method are the subject of accelerated ageing.

Put the levers together with the multipliers in the shelf life calculator, which starts a fragile oil at 2.5 months. Amber glass filled to the neck, a cool store at 8 to 12 C, tocopherols at a sensible level, fresh stock and small containers gives 2.5 multiplied out to about 8 months. The same oil in clear plastic in a warm room, half empty and without antioxidant, comes out near 1.2 months. Same oil, one order of magnitude apart, which is the finding of the 270 day storage study that anchors those factors: hemp seed oil in polypropylene at 25 C in diffused light passed 40 meq/kg by day 150, while the same oil in amber glass with foil at 10 C stayed under 10 meq/kg for the full nine months. A best before date on a hemp containing balm should be set with a real protocol behind it, as in shelf life testing, and stated conservatively.

Cost, and whether the profile earns its place

Hemp seed oil sits in the middle of the cost band: usually a little above sunflower, soybean and sweet almond per kilogram, well below argan, rosehip, tamanu and sea buckthorn. At a 10 percent inclusion the difference between hemp and a commodity oil is a rounding error in the cost of a tin, so price is not the argument either way. The real cost is the shortened best before date it imposes on every batch that contains it, and the write offs when that date is set optimistically.

The case for hemp is genuinely good on paper. Linoleic acid is the precursor of the acylceramides that hold the barrier's lipid lamellae together, low linoleate content is associated with barrier disturbance, and hemp delivers linoleic acid together with alpha-linolenic in a ratio close to three to one. The caveats are equally real. Topical delivery of a fatty acid into the barrier from a wax stick is not the same as dietary supply, and the evidence is thinner than the marketing suggests: what is known about depth and mechanism is in lipid penetration into skin and ceramides and barrier lipids. Higher linoleic content is available from safflower and standard sunflower at lower cost, without the green colour and with less alpha-linolenic to oxidise. And the gamma-linolenic content that makes hemp distinctive is 1 to 4 percent of an oil used at 10 percent, which is 0.1 to 0.4 percent of the formula. If GLA is the objective, evening primrose and borage oil carry five to ten times as much of it.

The verdict

Use hemp seed oil when three things are true at once: you sell or use the product within a few months of making it, you can store the oil cold and dark and buy it in a pack size you will finish, and the green colour and grassy smell suit what you are making. A face balm, a hand salve or a beard product made in small batches for a customer who will use it in eight weeks is a good home for it. Under those conditions its fatty acid profile is a real advantage and its fragility never gets tested.

Leave it out when any of the three fails. A market stall line with a twelve month best before, a pale tinted lip product, a fragrance forward formula that the grass will fight, or a household that keeps its balm in a bathroom are all cases where hemp costs more than it gives. The substitution is easy and unglamorous: a high oleic oil for the base, and if you want the linoleic content, a high linoleic oil chosen with the index in mind from carrier oils compared. There is no formulating trick that makes a bis-allylic count of 98 per hundred chains behave like 10.

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