Honey in a balm: you are adding water to a product designed to have none

Honey has a water activity of 0.55 to 0.65 and 15 to 20 percent water. What that does to an anhydrous balm, plus honey powder, tack and crystallisation.

Honey appears in more homemade lip balm recipes than any other non-lipid ingredient, and almost none of them mention what it brings with it. Honey is between a sixth and a fifth water by weight. Putting it in a balm converts an anhydrous product, which needs no preservative because there is nothing for microbes to live in, into a water-containing one that does. This page sets out the numbers, the storage faults, the preservation problem and the three ways to get the honey idea without the honey.

• INCI: Mel (EU), Honey (US)
• Water 15-20%
• Water activity 0.55-0.65
• Insoluble in oil

What honey is made of

Honey is a supersaturated sugar solution, not a syrup with additives. Bees reduce nectar to under a fifth water and invert most of its sucrose, leaving a mixture that holds more dissolved sugar than water at that temperature should be able to carry. Every property that matters in a balm follows from that.

Four mechanisms keep honey from spoiling in its own jar, and it is worth naming them separately because they fail separately. Low water activity, which stops growth osmotically. Low pH. Hydrogen peroxide generated slowly by the enzyme glucose oxidase the bee added. And a very low nitrogen content, so there is little for an organism to build itself from. Only the first two survive being dispersed into a balm, and the first one only survives conditionally.

Why the self preservation does not transfer

Honey preserves itself by osmosis: the dissolved sugar holds the water so tightly that a cell placed in it loses its own water rather than gaining any. That works only while the sugar concentration stays where it is. Add water and the whole mechanism unwinds, which is why honey ferments in a damp store and why honey diluted for use in a kitchen spoils like any other sweet liquid.

Two things follow. Dispersing honey into a fat does not dilute it, so each droplet keeps roughly the water activity it had in the jar, around 0.55 to 0.65. That sounds reassuring until you look at what grows there. Osmophilic yeasts of the Zygosaccharomyces group grow down to about 0.61, which is inside that band, and honey with a moisture content near the 20 percent Codex ceiling is routinely described in the food trade as fermentation-prone for exactly this reason. Honey is not sterile and not self-sterilising: it is a material sitting just on the safe side of a line, and every route that adds water pushes it across.

The second point is the one that catches makers who reason from a single measurement. Water moves slowly through a fat phase, and materials drift toward equilibrium with the humidity around them: equilibrium relative humidity in percent is water activity times 100. A honey balm left open in a bathroom at 70 percent relative humidity has honey droplets drifting toward a water activity of 0.70, and at that point they will support growth. That is not a theoretical concern in a product applied with fingers, from an open tin, in the wettest room in the house. The wider argument is in do balms need preservatives.

The arithmetic at 1, 3 and 5 percent

Take honey at 17.5 percent water and 82 percent total sugars, the middle of the ranges above, and work out what each usage level puts into the batch.

The interesting comparison is with the measured behaviour of oil itself. Water activity measured on peanut oil reads 0.75 at 722 ppm of water and 0.93 at 923 ppm, so an oil holding less than a tenth of one percent of dissolved water is already at the ISO 29621 low-risk line. Honey at 1 percent puts in several times that. But the honey water is not dissolved in the oil, because a triglyceride can hold only a few hundred parts per million before water separates as its own phase. It sits as a dispersed syrup, and asking for the bulk water activity of the mixture is the wrong question. What matters is the water activity inside the droplets, and inside the droplets it is the water activity of honey.

Separation, beading and crystals

Honey has a density near 1.4 g/mL against roughly 0.92 for a molten balm, and it is completely immiscible with oil. Three faults follow, all of them predictable.

Sinking during the pour. Stirred into a melt and poured, honey droplets fall through the liquid until the wax network sets around them. Pour too warm or cool too slowly and you get a sweet, sticky layer at the base of the tin and a plain balm on top. This is the same physics as pigment sinking, and the same fixes apply: pour cooler, cool faster, and raise the yield stress of the melt. See balm separated in layers for the general case.

Beading. Over weeks, droplets migrate and coalesce, and the largest of them reach the surface as visible amber beads. Warmth accelerates this because it softens the wax network holding them apart. A balm that looked uniform on day one can bead by week six, which means a fault that appears after the sample has been approved.

Crystallisation. Honey crystallises on its own timescale, as glucose comes out of supersaturation as glucose monohydrate. The rate depends mostly on the glucose to water ratio and the fructose to glucose ratio: a high-glucose honey such as rape or clover granulates within weeks, while acacia and manuka stay liquid for a long time. In a balm the crystals form inside the droplets, and the result is a distinct gritty mouthfeel on a lip product. It is easy to mistake for shea grain, so check the diagnostic differences in grainy shea butter first: honey grit is soluble in the mouth and disappears, fat grain is not and does not.

Partial mitigations exist. High shear during cooling makes smaller droplets that migrate more slowly. A higher wax percentage and a faster set trap them better. Castor oil, being far more polar than a triglyceride, wets honey droplets better than the rest of the oil phase does and reduces coalescence. None of these is a fix, they are delays, and none of them addresses the microbiology.

Preserving a honey balm

A balm with honey in it is a water-in-oil dispersion with an unpreserved aqueous phase, and that is a harder preservation problem than a cream, not an easier one. The preservative has to be present in the water droplets, and most cosmetic preservatives partition strongly into the oil when there is far more oil than water. A dose calculated on the whole formula can end up almost entirely in the fat, where nothing is growing, and almost absent from the droplets, where something might.

What that leaves you with in practice:

• Workable

  A broad spectrum system at full dose

  Phenoxyethanol with ethylhexylglycerin, or a benzyl alcohol and dehydroacetic acid blend, both at the top of the supplier's range. Honey's own pH of 3.4 to 4.5 suits the acid-dependent systems.

• Constrained

  Sorbate and benzoate salts

  Effective only below about pH 5.5 and only in the water phase. They need the aqueous domains to be small and well distributed, and they are weak against moulds on their own.

• Not a preservative

  Tocopherol, rosemary, essential oils

  Antioxidants and aromatics. They slow oil going rancid and do nothing measurable to microbial growth at usage levels. In the EU only Annex V substances may be used as preservatives, and tocopherol is not among them.

The regulatory picture matters as much as the chemistry. In the EU and UK, preservatives must come from Annex V of Regulation (EC) No 1223/2009 at the listed concentrations, and a safety assessor will expect to see the choice justified. Whatever the market, the only evidence that a given system works in a given formula is a challenge test to ISO 11930 on the finished product in its finished packaging. Honey is precisely the case where guessing fails, because the water is unevenly distributed. The candidate materials and their limits are covered in preservatives for anhydrous products, and the contamination routes to close first are in workshop hygiene.

Honey powder, honey flavour and the other anhydrous routes

Most people who want honey in a balm want one of three things: the taste, the smell, or the word on the label. All three are available without the water.

Honey powder and dehydrated honey. Spray dried honey on a carrier, most often maltodextrin, typically 50 to 70 percent honey solids with a residual moisture of a few percent. It solves the water problem and inherits two others: it is a polar crystalline powder with no solubility in oil, so it behaves like sugar and feels gritty unless it is very finely milled and dispersed into a little warm oil first, and it is hygroscopic, so it pulls moisture out of humid air and can cake or turn tacky in an opened container. Usable at 0.5 to 2 percent, and no higher without accepting grit. Store the raw powder sealed with a desiccant.

Honey flavour. An oil-soluble flavour composition at 0.2 to 1 percent gives the taste and the smell with no water, no grit and no microbiology. It is the honest answer for a lip product. Remember that lip products are IFRA Category 1, the strictest there is, and that flavour and fragrance draw on the same headroom, which is worked through in lip balm flavour and sweeteners.

Honey extracts in glycerin or a glycol. Sold as honey extract or hydrolysed honey protein, these are honey in a humectant carrier. They swap a water problem for a humectant problem: glycerin is not miscible with oil either, so it beads, and it draws water to the surface of the product. The behaviour is the same one described in glycerin and humectants.

Yellow beeswax and propolis. The simplest route of all. Unbleached yellow beeswax carries a genuine honey and propolis note from the hive material that survives filtering, and at 10 to 15 percent of a lip balm it supplies the aroma for free. Propolis gives a more resinous version of the same idea. Both are hive products, so neither is vegan, and propolis is a recognised contact allergen, which is the subject of beeswax and propolis allergy.

Manuka, MGO ratings and what you can say

Manuka honey comes from Leptospermum scoparium. Its distinguishing feature is non-peroxide antibacterial activity from methylglyoxal, which forms slowly and non-enzymatically from dihydroxyacetone present in the nectar, over weeks to months of storage. The MGO number on a jar is a concentration in milligrams per kilogram: MGO 100, MGO 400, MGO 550. The UMF scale pairs an MGO figure with markers for authenticity, with UMF 10+ corresponding to roughly 260 mg/kg of methylglyoxal, UMF 15+ to roughly 510, and UMF 20+ to roughly 830. New Zealand's Ministry for Primary Industries separately defines monofloral and multifloral manuka for export using four chemical markers and a DNA marker, which is an authenticity test, not a potency one.

Now the dilution. A 3 percent addition of an MGO 400 honey delivers 12 mg of methylglyoxal per kilogram of finished balm, and it delivers it inside droplets rather than across the product. The antibacterial figures behind those ratings are measured on neat honey in an agar assay against a phenol standard. Nothing in that method survives being dispersed at 3 percent in fat, and the osmotic mechanism that does most of the work in neat honey has been abolished, because the surrounding matrix is oil rather than a wound bed.

Stickiness and the sensory ceiling

Sugar is hygroscopic, and honey droplets at the surface of a balm attract atmospheric water, which is what makes a honey lip balm feel tacky rather than glossy. The tack is humidity dependent: the same stick feels acceptable on a dry winter day and unpleasant in August. It also increases with time as droplets migrate to the surface.

Practical ceilings, from what makers consistently report rather than from a published sensory panel: about 1 percent in a lip balm before tack becomes the dominant impression, about 2 percent in an overnight product where cling is wanted rather than avoided, and lower again in anything applied to a large area, because sticky hands are noticed immediately. Sweetness is uneven at any level, since a droplet only tastes of anything once saliva reaches it. If a balm has gone tacky and honey is not the cause, the other mechanisms are in sticky or tacky balm.

Infant botulism and the under one rule

Honey is an established dietary source of Clostridium botulinum spores. In infants under about twelve months the gut flora is not yet established enough to suppress them, so ingested spores can germinate in the intestine and produce toxin there. This is infant botulism, and it is why public health authorities in the UK, the US and elsewhere advise that no honey be given to a child under one year. The advice is about ingestion of spores, not about the toxin, and no amount of heating a home cook can apply destroys spores.

Applying that correctly to topical products takes one distinction. Honey on intact skin is not a route to infant botulism, and honey-containing dressings are used clinically on adults. What matters is whether the product will be eaten. A lip balm is eaten in small quantities by definition. So is anything on the hands of a baby, and so is a nipple balm, which transfers directly into an infant's mouth at every feed. For those products the under one rule applies with full force, and honey should simply not be in the formula. The general position on products for the very young is in balms for babies, and the lip-specific version in lip balm for children.

Honey is not restricted as a cosmetic ingredient anywhere on this basis. The constraint is on the use case, not on the material, which is a distinction worth getting right before repeating either half of it as a rule.

The decision rule

Decide what you actually want from the honey, then take the cheapest route to it. If you want the smell, use yellow beeswax. If you want the taste, use an oil-soluble honey flavour. If you want the word on the ingredient list, use honey powder at 1 percent or less, mill it into warm oil, seal the container, and accept that it may go gritty. All three of those keep the product anhydrous and keep the preservative question closed.

If you want real honey in the balm, accept what you are making. It is a water-containing cosmetic that needs a preservative system, a challenge test, waterproof packaging, a shorter shelf life and a stability programme that watches for beading and crystals rather than only for rancidity, as covered in shelf life testing. That is a reasonable thing to do deliberately and an unreasonable thing to do by following a recipe that mentions none of it. For personal use in a tube you will finish in a month, at 1 percent, in a dry cupboard, the risk is low and the sensory penalty is the real cost. For anything sold, the honest position is that the honey has to earn a preservative, and most of the time it does not.

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