CBD and hemp extract in balms: the arithmetic, the law and the evidence
How to convert a milligram claim into a percentage, the UK novel food and THC limits, degradation by heat and light, and what topical CBD evidence actually supports.
A CBD balm is an ordinary anhydrous balm with a milligram number on the lid, and almost every difficulty it creates comes from that number rather than from the formulation. The arithmetic is trivial and people still get it wrong. The law is not trivial, and the limit that catches UK makers is not the one they have read about. And the evidence for what the ingredient does on skin is thinner than the price suggests.
Weight percent equals milligrams divided by the fill weight in grams, divided by ten: 300 mg in a 30 g tin is 1.0 percent. In the UK a finished cosmetic must contain under 1 mg of THC per unit, which is 33 ppm in a 30 g tin, far tighter than the 0.2 percent crop figure. Add CBD below 70 C, package opaque, and make no therapeutic claim.
The arithmetic: turning a milligram claim into a percentage
CBD is sold on milligrams per container and formulas are written in weight percent, so the first job is the conversion. One percent is 10 mg per gram of finished product, so:
weight percent = milligrams per unit divided by (fill weight in grams times 10)
| Claim on the lid | In a 5 g lip pot | In a 15 g tin | In a 30 g tin | In a 50 g jar |
|---|---|---|---|---|
| 100 mg | 2.00% | 0.67% | 0.33% | 0.20% |
| 250 mg | 5.00% | 1.67% | 0.83% | 0.50% |
| 300 mg | 6.00% | 2.00% | 1.00% | 0.60% |
| 500 mg | 10.0% | 3.33% | 1.67% | 1.00% |
| 1000 mg | 20.0% | 6.67% | 3.33% | 2.00% |
The table already tells you something. A 1000 mg claim on a small container is not a stronger product so much as a different one: at 20 percent of the formula CBD is a major phase, not an additive, and it changes how the balm sets and feels. Most commercial topical CBD sits between 0.3 and 3 percent.
The second step is the one people skip. You weigh out an extract, not pure cannabidiol, so divide by the potency on the certificate of analysis. For a 1.00 percent CBD target, a 99 percent isolate needs 1.01 percent of the formula, an 80 percent broad spectrum distillate 1.25 percent, and a 65 percent full spectrum distillate 1.54 percent. Skip it and you undershoot a 300 mg claim by a third.
Worked example. Twenty 30 g tins is a 600 g batch, and 300 mg per tin needs 6.0 g of CBD in it: 6.06 g of a 99 percent isolate, or 7.5 g of an 80 percent distillate. Batch overage does not dilute the claim, because the dose is set by percentage. Make 660 g to allow for what stays in the jug, scale every ingredient including the CBD, and each tin still carries 300 mg. The batch calculator handles the scaling; it cannot check your extract potency.
Fill tolerance moves the claim. A tin filled to 30 g plus or minus 3 percent carries 291 to 309 mg at a nominal 300. If you are declaring a number, declare it as a minimum and know your filling variation, because the label figure has to be true for the tin the customer opens, not for the average of the batch.
Isolate, broad spectrum and full spectrum
The three grades differ mostly in what comes along with the cannabidiol.
| Grade | CBD content | Form and colour | THC | What it does to the balm |
|---|---|---|---|---|
| Isolate | 98-99.9% | White crystalline powder or shards, odourless | None detectable | Nothing visible. Dissolves clear in warm oil, no colour, no smell, no taste |
| Broad spectrum distillate | 70-90% | Viscous amber to dark amber resin | Removed, usually to below the limit of quantification | Adds a honey to brown tint and a faint herbal note. Minor cannabinoids and some terpenes retained |
| Full spectrum distillate or extract | 50-80% | Dark amber to near black, thick and sticky | Present, quantity stated on the certificate | Visible colour, distinct hemp aroma and taste. The most restricted grade in the UK |
Handling differs too. Isolate is a powder that dissolves readily in warm oil and weighs accurately on a gram scale. Distillates are thick enough at room temperature to be awkward: warm the container to 40 to 50 C before weighing, accept that some stays in the pot, and weigh straight into the batch vessel. Above roughly 8 to 10 percent, isolate can exceed its solubility in the oil phase and recrystallise as the balm cools, giving fine white crystals and a gritty feel. That is a solubility failure rather than contamination, and another reason the 1000 mg small tin is a bad idea.
The UK limit that actually binds: 1 mg per unit
The figure most people quote is 0.2 percent THC, or 0.3 percent from US sources. Neither is the limit on your finished product. The first is the UK ceiling for a growing crop under an industrial hemp cultivation licence; the second is the hemp definition in the 2018 US Farm Bill, delta-9 THC on a dry weight basis. Both describe plants. Your balm is not a plant.
In the UK, delta-9 THC and its stereochemical variants are controlled drugs under the Misuse of Drugs Act 1971, as are cannabinol, CBN, and cannabinol derivatives. A consumer product containing them is lawful only if it meets the exempt product definition in regulation 2 of the Misuse of Drugs Regulations 2001, which needs all three of: the product is not designed for administration of the controlled drug; the drug cannot be recovered by readily applicable means, or in a yield that constitutes a risk to health; and no one component part contains more than one milligram of it.
That third condition is a fixed mass per unit, not a concentration, so the bigger your container the tighter the concentration limit becomes.
| Container | Fill weight | Max THC, percent | Max THC, ppm |
|---|---|---|---|
| Lip pot | 5 g | 0.020% | 200 |
| Small tin | 15 g | 0.0067% | 67 |
| Standard tin | 30 g | 0.0033% | 33 |
| Large jar | 50 g | 0.0020% | 20 |
| Body balm jar | 100 g | 0.0010% | 10 |
Here is the trap, in numbers. A full spectrum distillate whose certificate declares 0.2 percent THC, used at 3 percent of the formula, puts 0.006 percent THC in the balm. That is 60 ppm, or 1.8 mg in a 30 g tin: nearly double the exemption limit, from an input the supplier described as compliant. To stay inside 1 mg per 30 g tin at a 3 percent inclusion, the distillate itself must be below about 0.11 percent THC. Do this multiplication for every lot, and remember CBN counts separately.
There is a second UK and EU layer above the drug law. The Cosmetic Regulation prohibits narcotics listed in the 1961 Single Convention, which covers cannabis, cannabis resin, and extracts and tinctures of cannabis. Following the Court of Justice ruling in the Kanavape case in November 2020, the Commission updated the CosIng entry so that cannabidiol produced synthetically, or obtained from parts of the plant that the Convention excludes, is not caught by that prohibition, while CBD derived from an extract, tincture or resin of cannabis remains within scope. Which side of that line your input sits on depends on the plant part and the process, which means it depends on documents from your supplier. This is a question for your safety assessor and, in Great Britain, for the responsible person, and it is covered from the compliance end in selling balms in the UK and EU and safety assessment and the CPSR. No website can settle it for your specific material.
Novel food, and why a lip balm sits awkwardly
Novel food is a food regime, not a cosmetics one, and the distinction matters. In Great Britain, CBD extracts intended for ingestion are novel foods requiring pre-market authorisation from the Food Standards Agency, which required applications to be validated by 31 March 2021 and maintains a public list of products permitted to remain on sale while assessment continues. The FSA also publishes provisional daily intake advice for healthy adults, revised downwards in February 2024 to 10 mg per day from the 70 mg it had advised since 2020.
A balm applied to arms and legs is unambiguously a cosmetic and sits outside all of that. A lip balm is where it gets uncomfortable. Legally it is still a cosmetic, but a meaningful fraction is swallowed, safety assessment for lip products already has to consider an oral exposure route, and marketing that mentions a dose, an intake or a wellbeing benefit invites a regulator to read the product as a food or a medicine instead. Against the FSA's 10 mg a day, a 5 g lip pot at 6 percent carrying 300 mg is hard to defend even though it was never sold as food.
The practical conclusion is unglamorous: if you want to make a CBD product, make a body balm or a rub. A CBD lip balm carries the most regulatory exposure of any format and the least plausible benefit.
Heat: decarboxylation, degradation, and when to add it
Raw cannabis produces cannabidiolic acid, CBDA, which loses carbon dioxide on heating to give CBD. The reaction proceeds usefully from around 110 C and is typically run at 110 to 140 C for 30 to 60 minutes. You will never do it: isolates and distillates are sold already decarboxylated, and a certificate showing a high CBDA figure means you bought the wrong material, not that you need to bake it.
What matters in a balm pot is the opposite reaction. CBD has no cliff edge above which it dies; loss accumulates as a function of time at temperature, with oxidation the main route. Crystalline CBD melts at 66 to 67 C, which is convenient, because an isolate then dissolves easily in an oil phase held anywhere in the normal pouring range. There is no reason to expose it to wax melting temperatures at all.
- Melt the wax phase as usual. Waxes and butters at whatever the formula needs, up to 90 C if a hard wax requires it.
- Cool to 65 to 70 C. Off the heat, stirring, checked with a thermometer rather than by eye.
- Stir in the CBD until the melt is clear. Isolate takes a minute or two. Distillate needs longer and should be pre-warmed.
- Add volatile actives last. Menthol, camphor and essential oils below 60 C, after the CBD has dissolved.
- Pour and set at room temperature. Never reheat a CBD batch to fix a fault in the pour.
The process to avoid is adding CBD early and then holding the batch hot while you fiddle: twenty minutes at 90 C costs potency you have paid for and cannot see. Keep acids out too, since cannabidiol cyclises to THC under strongly acidic laboratory conditions. That is not a realistic pathway in a neutral anhydrous balm, but it is a good reason not to invent one.
Light, oxygen, packaging and honest potency dating
Cannabinoids degrade by oxidation, and storage studies on cannabis preparations going back to the 1970s consistently identify light as the largest single factor, ahead of oxygen and ahead of moderate temperature. Preparations kept dark held their content for a year or more; the same material in the light did not. Nothing about a balm changes that chemistry.
So the packaging decision is not decorative. Use an opaque container, an aluminium tin or amber glass in a carton, fill close to the top so there is little headspace oxygen, and store and ship cool. A clear glass or PET jar on a lit shop shelf is the worst case available, and no antioxidant will rescue it. Container and closure choices are in the packaging guide.
Dating a potency claim is harder than dating a balm, because two clocks run at once. The base ages by oxidation of its oils, the ordinary problem covered in rancidity and oxidation, and faster if the carrier is fragile: a balm built on hemp seed oil because the marketing wanted the word hemp goes rancid long before it goes weak. Separately, the cannabinoid content falls. A label saying 300 mg has to be true at the end of the shelf life, not on the day you poured it, which is why some producers formulate 10 to 20 percent above the claim. That is defensible only with data showing the loss you are compensating for. Accelerated ageing at 40 C predicts it badly, because the dominant route is light and oxygen driven rather than thermal, so real time testing of retained samples is the only honest basis. See accelerated ageing and shelf life testing.
Reading a certificate of analysis
A CBD input without a current, lot specific certificate from an independent laboratory is unusable, because every number you print descends from it. Check that the lot number matches the drum, that the analysis date is recent, and that the laboratory is named and accredited to ISO/IEC 17025.
| Panel | What it reports | What to check |
|---|---|---|
| Cannabinoid potency | CBD, CBDA, delta-9 THC, THCA, CBN, CBG and others by HPLC | Total CBD is CBD plus 0.877 times CBDA. Use the total, and use it to set your inclusion rate |
| THC and CBN | Quantified separately, with a limit of quantification | A non-detect means nothing without a stated LOQ. Multiply through to milligrams per unit for your container |
| Residual solvents | Ethanol, hexane, butane, propane, heptane | Relevant to hydrocarbon and ethanol extractions. Supercritical CO2 material should be clean |
| Pesticides | Multi residue screen | Confirm the screen covers the compounds used on hemp in the country of origin, not a generic short list |
| Heavy metals | Lead, cadmium, arsenic, mercury | Hemp is used for phytoremediation precisely because it takes metals up from soil. This panel is not optional |
| Microbiology and mycotoxins | Total counts, yeast and mould, aflatoxins, ochratoxin A | Most relevant to crude extracts and plant material rather than to purified isolate |
Keep the certificate filed against every batch it was used in. If anyone ever asks what was in a tin you sold, that paper trail is the only answer you will have. The supplier questions to ask before the drum arrives are in sourcing ingredients.
What the topical evidence supports, and what it does not
Start with the physics, because it constrains everything else. Cannabidiol is extremely lipophilic, with a partition coefficient around six to seven on the log scale. Molecules like that partition happily into the lipids of the stratum corneum and then stay there, because moving on into the watery viable epidermis is thermodynamically unattractive. From a simple anhydrous balm with no penetration enhancement, most of the CBD you apply sits in the outermost dead layer of skin. What does and does not get further is covered in lipid penetration into skin.
What exists in the literature is a scatter of small studies: open label work in peripheral neuropathy, case series in blistering skin conditions, and small trials of CBD containing emollients in inflammatory skin disease. They are frequently uncontrolled or run without a vehicle only comparator, and often funded by parties with an interest. That is preliminary evidence, not a demonstration. Purified CBD is a licensed prescription medicine for specific rare epilepsies, taken orally at hundreds of milligrams a day, and that licence tells you nothing about a balm.
Two confounders swamp the ingredient in ordinary use. The first is the vehicle: an occlusive anhydrous balm slows water loss and improves the feel of dry skin by itself, which is what most people are noticing, and the mechanism is in occlusive, emollient and humectant. The second is the company it keeps. Most CBD muscle rubs also carry menthol and camphor, which produce an unmistakable cooling and warming sensation well below their safety limits. Those are the materials doing the perceptible work: see menthol, camphor and actives, the ceilings in menthol and camphor safety limits, and a worked formula in muscle rub balm.
Stated plainly: no well powered, vehicle controlled trial shows that a topical CBD balm relieves pain, reduces inflammation or treats a skin condition better than the identical balm without it. Tolerability is easier. Cosmetic level CBD is generally well tolerated on skin, and where a full spectrum product provokes a reaction, the terpenes and residual plant material are likelier culprits than the cannabidiol.
What you may say on the label
A cosmetic cleanses, perfumes, changes appearance, protects, keeps in good condition or corrects body odour. Pain, inflammation, arthritis, eczema, healing and anything presenting the endocannabinoid system as a therapeutic target are drug claims, and one of them turns your cosmetic into an unlicensed medicine in the UK and an unapproved new drug in the US. The wording that stays on the right side of the line is in cosmetic versus drug claims.
The milligram figure is a quantity statement rather than an efficacy claim, and is permitted provided it is accurate. Ingredient naming follows the usual rules: Cannabidiol for isolate, Cannabis Sativa Seed Oil for hemp seed oil, which contains no meaningful CBD at all, and an extract name reflecting the actual plant part for a distillate. That last one is part of the Annex II question above rather than a formality, and the conventions are in INCI names explained. US sellers face the format rules and the facility registration duties set out in labelling cosmetics for the US.
The honest position
CBD is the most expensive gram in a balm and the least demonstrated. Add a certificate for every lot, potency verification of the finished product if you print a number, and a safety assessment that has to reason about an ingredient with a live regulatory question attached, and the compliance burden is large relative to a 1 percent addition whose topical effect nobody has shown.
The decision rule, if you still want to make one: include CBD only if all four of these are true. You will obtain and keep a lot certificate and do the THC per unit multiplication yourself. You will package opaque and fill full. You will date the potency claim against real data rather than optimism. And you will describe the product without a single therapeutic word. If any one of those will not happen, leave it out and sell the balm on the balm, which is what the customer is feeling anyway. Nothing here is legal advice or a safety assessment, and the limits are set out in the disclaimer.
Frequently asked questions
How many mg of CBD is 1 percent in a balm?
One percent is 10 mg per gram of finished product. So a 30 g tin at 1 percent contains 300 mg, a 15 g tin at 1 percent contains 150 mg, and a 5 g lip pot at 1 percent contains 50 mg. To go the other way, divide the milligram claim by ten times the fill weight in grams to get the percentage.
Is CBD legal in a cosmetic in the UK?
CBD itself is not a controlled drug, but THC and CBN are, and a finished product must contain under 1 milligram of each per unit to meet the exempt product definition in the Misuse of Drugs Regulations 2001. Separately, cosmetics law restricts CBD obtained from cannabis extracts, tinctures and resin. The answer depends on your specific input and its documentation.
Does the 0.2 percent THC limit apply to my balm?
No. The 0.2 percent figure is the ceiling for the growing crop under a UK industrial hemp licence, and the US 0.3 percent figure is the hemp definition in the 2018 Farm Bill. Both describe plants. A finished UK consumer product is governed by the 1 milligram per unit exemption, which in a 30 g tin works out at 33 parts per million.
At what temperature should I add CBD to a balm?
Between 65 and 70 C, after the waxes have melted and the batch has cooled. Crystalline CBD melts at 66 to 67 C and dissolves readily at that temperature. Degradation is cumulative time at temperature rather than a sudden threshold, so avoid holding a CBD containing batch hot, and never reheat one to fix a pour fault.
Does CBD balm work for pain?
There is no well powered, vehicle controlled trial showing that a topical CBD balm relieves pain better than the same balm without it. CBD is very lipophilic and largely stays in the outer dead layer of skin. Most CBD muscle rubs also contain menthol and camphor, which produce the sensation people attribute to the CBD.
What is the difference between hemp seed oil and CBD?
Hemp seed oil is pressed from the seed and contains no meaningful cannabidiol. It is a fast drying, linoleic rich carrier oil with a short shelf life. CBD is an extract of the flowering material, or a synthesised or purified isolate, and is a separate ingredient with a separate price and separate regulatory treatment. Products marketed as hemp oil are frequently the former.
How long does a CBD balm keep its potency?
Long enough is a testing question, not a guess. Cannabinoids degrade mainly through light and oxygen, so an opaque, well filled tin stored cool holds potency far longer than a clear jar on a lit shelf. Accelerated ageing at 40 C predicts this badly, so the honest route is real time testing of retained samples from your own batches.
Sources and further reading
- UK Government, Misuse of Drugs Regulations 2001, regulation 2: interpretation and the definition of exempt product, legislation.gov.uk.
- UK Government, Misuse of Drugs Act 1971, Schedule 2: controlled drugs, including cannabinol and tetrahydrocannabinols, legislation.gov.uk.
- Food Standards Agency, Cannabidiol (CBD) guidance for business, the public list and the provisional daily intake advice, FSA.
- European Commission, CosIng cosmetic ingredient database, entry for Cannabidiol, and Annex II entry 306 of Regulation (EC) 1223/2009.
- Court of Justice of the European Union, Judgment in Case C-663/18, B S and C A (Kanavape), 19 November 2020.
- US Food and Drug Administration, FDA regulation of cannabis and cannabis derived products, including cannabidiol.
- Fairbairn, Liautaud and Rowan, The stability of cannabis and its preparations on storage, Journal of Pharmacy and Pharmacology, 1976.
Reviewed and updated 6 September 2026. Spotted an error? Tell us and we will fix and log it.