Synthetic antioxidants: what BHT actually buys, and where it is restricted

BHT at 0.01 to 0.1 percent, BHA, propyl gallate and TBHQ compared on induction time gains, EU cosmetic restrictions and where they beat tocopherol and rosemary.

Four synthetic phenols do almost all the antioxidant work in the food industry, and one of them, BHT, turns up in cosmetics constantly without anyone discussing it. They are cheap, colourless, odourless and effective at a few hundred parts per million, which is exactly what tocopherol and rosemary extract are not. They also carry a regulatory and reputational load that has grown steadily, and one of the four is effectively closed off in the EU. This page separates what the measurements show from what the marketing says.

• BHT: 0.01-0.1%
• BHT melts at 69-72 C
• E 321 in food
• Not permitted by natural standards

The four molecules and their doses

All four are hindered phenols, meaning a phenolic hydroxyl group with bulky substituents crowding it. That architecture is the whole reason they work, and it is shared with tocopherol. The practical differences between them are melting point, solubility and legal status rather than chemistry.

Those dose bands are not arbitrary. The US food additive regulations cap each of the four at 0.02 percent of the fat or oil content of the food, with a combined limit of 0.02 percent when more than one is used, and that convention has propagated into cosmetic practice. The upper end of the BHT band, 0.1 percent, is a cosmetic figure rather than a food one, and it is already well past the point where more delivers anything.

How a hindered phenol stops the chain

Rancidity is a radical chain reaction. A lipid radical takes up oxygen to give a peroxyl radical, that peroxyl radical abstracts a hydrogen from the next fatty acid, and the cycle repeats hundreds of times per initiation event. The mechanism in full is in rancidity and oxidation.

A chain-breaking antioxidant intercepts the propagating step by being an easier hydrogen donor than the fatty acid is. The phenol hands its hydroxyl hydrogen to the peroxyl radical, which becomes an inert hydroperoxide, and what is left is a phenoxyl radical. The trick is what that radical does next, which is nothing. Its unpaired electron is delocalised around the aromatic ring and the tert-butyl groups physically shield the reactive position, so it is too stable and too crowded to pull a hydrogen off an intact fatty acid. Instead it waits and couples with a second peroxyl radical, giving non-radical products. One molecule therefore terminates about two chains.

That stoichiometry, combined with a kinetic chain length in the hundreds, is why a few hundred parts per million is the right order of magnitude. You are not neutralising the oxygen in the jar or the fatty acids in the formula, you are mopping up the small number of radicals that actually get a chain going. Doubling the dose does not double a protection that was never proportional to concentration in the first place.

Potency and persistence are different things, and this is where the comparison with tocopherol gets interesting. In classic inhibition kinetics, alpha-tocopherol traps peroxyl radicals with a rate constant around 3 x 10⁶ per molar per second against roughly 1 x 10⁴ for BHT, so tocopherol is by far the faster scavenger. BHT nonetheless outperforms it over months of storage, because it is not consumed by side reactions, it survives heat, and it does not turn on the oil at ordinary use levels. Speed is not what a shelf life needs.

What the induction data actually shows

The measurable quantity is the induction period: how long an oil resists under forced oxidation before the reaction accelerates. Rancimat runs a stream of air through hot oil and detects the volatile acids that appear at the breakpoint.

Read that carefully before you conclude anything about natural beating synthetic. The doses were not matched: rosemary was run at twice the concentration of the synthetic blend, so the experiment shows what each option delivers at a plausible commercial loading, not which molecule is more potent. Normalise for cost instead and the picture inverts. BHT is one of the cheapest functional cosmetic ingredients there is, an order of magnitude below deodorised rosemary extract per unit of active, so at equal spend you could dose it several times over. What stops you is not money but the ceiling described in the next section. Past roughly 200 to 500 mg/kg there is nothing further to buy.

The overdose failure mode

Every chain-breaking phenol can turn prooxidant above its optimum, for the same reason it works below it. The phenoxyl radical is stable, not inert. When there are few of them they wait for a peroxyl radical to couple with. When there are many, some of them find an intact fatty acid or a hydroperoxide first and start a new chain, and the antioxidant becomes an initiator. This is documented for tocopherol at concentrations above about 100 ppm and it applies in the same way, at higher thresholds, to the synthetics.

In a balm this shows up three ways, none of which look like oxidation to the person seeing them.

• Yellowing. Spent BHT oxidises to quinone-type products that are yellow. A pale, unscented balm dosed at half a percent by an enthusiastic maker will discolour over months, and the antioxidant is the cause rather than the cure.
• White specks. BHT is a crystalline solid melting at 69 to 72 C. Added to a melt that is already cooling, or stirred into a batch below its melting point, it stays as grit that never dissolves. On a lip it is detectable. The general diagnosis is in undissolved wax specks.
• Surface bloom. BHT sublimes slowly at room temperature, so it can recrystallise as a fine white deposit on the surface of a balm or inside a lid, which reads to a customer as spoilage.

The discolouration route for propyl gallate is different and worth knowing even if you never use it. Gallates chelate iron and the complex is blue to black, so a gallate-containing product in contact with unlined steel, or made with an iron-rich unrefined oil, can go a dramatic colour. That alone rules it out for most small workshops.

Solubility, melt handling and when to add it

BHT is unusual among cosmetic additives in that heat helps rather than hurts. Tocopherol and rosemary extract degrade with prolonged warmth, so they go in at the end, below 50 to 60 C. BHT melts at 69 to 72 C, boils far above anything a balm sees, and is not damaged by a normal wax melt, so it goes in at the start of the oil phase where it also protects the oils during heating. That is a real advantage in a formula with a hard wax and a high melt temperature.

1. Add it to the oil phase, not the wax phase, at the beginning. Bring the oils to 70 C or above so the crystals melt rather than sit.
2. Or make a premix. Dissolve 1 g of BHT in 9 g of a stable liquid oil, warm, and dose that stock solution at ten times the target weight. This solves the weighing problem and the dispersion problem at once.
3. Stir until the melt is optically clear. Undissolved crystals are the commonest fault and they are invisible against white wax.
4. Do not add it after the pour has started. By then the batch is below the melting point and cooling further.

Weighing is the real obstacle at hobby scale. A 250 g batch at 0.02 percent needs 0.05 g, which a scale reading to 0.01 g can just about handle and a 1 g kitchen scale cannot handle at all. Either scale the batch up or use the premix. The general point about resolution and calibration is in weighing and calibration.

The legal position in the EU and the UK

This is where the four separate, and it is the part that changes. Verify the current entries before a launch rather than trusting any page, including this one.

Two points about that table matter more than the rest. First, the BHT figures come from a Scientific Committee on Consumer Safety opinion rather than from a numerical entry in an annex, which means there is no single legal limit to quote. In practice the opinion is the ceiling, because it is what a safety assessor will hold you to when signing a product report, and the UK position follows the same evidence. If you sell in the EU or Britain your assessor decides, not your supplier, which is the whole subject of safety assessment and the CPSR and part of the compliance route in selling balms in the UK and EU.

Second, Article 15 is the mechanism that quietly removes ingredients from the European market. A substance classified as a CMR under the CLP Regulation is prohibited in cosmetics automatically, unless it is specifically evaluated and found safe for that use. That is what has happened around BHA, and it is why the answer to "can I use BHA in the EU" is no even in the absence of a headline ban. BHT has been through the opposite process and come out with an opinion supporting its use.

The marketing pressure, and the evidence

BHT and BHA appear on almost every retailer restricted substances list and every "free from" screen, and no natural or organic certification permits them, so a maker who wants that positioning has no decision to make. That is a commercial fact, and it is a good enough reason on its own.

It is not the same as a safety finding, and it is worth being straight about the evidence. The strongest animal data concerns BHA, which produced tumours in the forestomach of rats and hamsters at high dietary doses. Humans have no forestomach, and the mechanism appears tied to that tissue, which is why IARC placed BHA in Group 2B rather than higher and why the food additive survives with an ADI. BHT sits in Group 3, the category for substances that cannot be classified either way. Both have been reviewed repeatedly as food additives and both retain acceptable daily intakes. Contact allergy to either is documented but uncommon in cosmetic patch test series.

The arithmetic for a cosmetic is the part that never makes it into the discussion.

A lip balm is the worst case, because it is the cosmetic most completely ingested. On that basis the exposure from a balm is roughly a thousandth of an intake that regulators consider acceptable every day for life. Deciding against BHT is a defensible commercial and precautionary choice. Presenting it as a toxicological one requires ignoring the numbers, which is the general problem with the vocabulary examined in chemical free and clean beauty claims.

Choosing between BHT, tocopherol and rosemary

Start by accepting that the antioxidant is the third lever. Choosing a lower bis-allylic oil does more for shelf life than any additive, and filling the container to the top does more than either, as the rankings in the oil comparison table make clear. Only once those are settled does this decision matter.

The naturals have their own page each, with doses and the pro-oxidant ceiling: vitamin E and antioxidants for the tocopherol family, rosemary extract for E 392, and vitamin E against rosemary extract for the choice between them.

What none of them will do

Three limits apply to every antioxidant on this page and none of them is negotiable. They do not preserve. An antioxidant has no effect on bacteria, yeast or mould, and if your product contains water, hydrosol, honey or fresh plant material it needs a preservative system instead, which is the subject of do balms need preservatives. They do not reverse anything: once an oil smells of crayon or old paint the aldehydes already exist, and adding BHT to a rancid batch produces a rancid batch with BHT in it, as set out in rancid balm. And they do not license a date. No dose of anything on this page justifies extending a best-before beyond what your oils and your testing support, which is the mistake behind most cases of a balm going off before its best-before.

The decision rule, then. If your product is pale, delicate, cheap to make and sold to people who read a label for function rather than philosophy, BHT at 0.02 to 0.05 percent is the technically better answer and costs almost nothing. If your positioning, your certification or your retailer says no synthetics, the naturals cost more and perform adequately, and you lose less than the debate implies. What you should not do is add a synthetic antioxidant to a formula built on grapeseed and rosehip and think the problem is handled. It is not. The substrate is the problem, and no phenol at 200 ppm fixes an oil phase that was chosen badly.

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