Water contamination control: where moisture gets in and how to keep it out

Every route water takes into an anhydrous balm, with workshop humidity targets, drying protocols for botanicals and tallow, condensation rules and in process checks.

Anhydrous is a description of your process, not a property of your ingredient list. A formula of nothing but wax, butter and oil becomes a wet product the moment a lid drips, a herb goes in leathery rather than brittle, or a tin comes out of a cold van into a warm room. This page maps the routes water takes in, gives the number each control is set to, and shows the two checks that catch a wet batch before it is filled.

• Target aw: below 0.75
• Workshop: 40 to 60 percent RH
• Botanicals: below 10 percent moisture
• Tallow dry-off: 100 to 105 C
• Dry down after wet cleaning: 30 min

What "dry enough" means as a number

The measure is water activity, aw, the ratio of the vapour pressure of water in the material to that of pure water at the same temperature, on a 0 to 1 scale. It is not water content, which is the whole reason honey keeps and diluted honey does not. Organism-by-organism minima are on do balms need preservatives.

Two numbers bound the problem. About 0.60 is the approximate floor for growth of any organism, cited as such in ISO 29621: below it, cell division stops rather than slows. 0.75 is the line ISO 29621 draws for a product to count as microbiologically low risk on water activity alone. The gap between them belongs to the moulds and the osmophilic yeasts. Most moulds need about 0.80, but xerophilic species multiply down to around 0.65, which is why 0.75 is a pragmatic industry limit rather than a biological threshold.

Here is the part that changes how you work. Water activity in a balm is a local property, not a batch average. Oil and water do not mix, so introduced water never dilutes itself through the melt. It sits as discrete droplets, and inside a droplet the water activity is close to 1.0 whatever the bulk figure says. A 50 mg bead of condensate in a 15 g tin is 0.3 percent of the mass and would barely register on a bulk measurement, yet it is a hospitable microhabitat walled off by lipid, which no antioxidant and no oil-soluble preservative can reach. That is the mechanism behind almost every case of mould in balm: not a damp product, but a dry product with one wet spot in it.

The routes water takes in

There are six, and they differ by orders of magnitude. Rank them by how much water each can deliver rather than treating them as equally urgent.

The top three are preparation decisions made once and then repeated. The bottom three are habits, and they produce the batch that was fine last month and cloudy this month. Cleaning and personal practice are on workshop hygiene.

Workshop humidity: 40 to 60 percent, on a hygrometer you have checked

Hold the making space at 40 to 60 percent RH. Both ends of the band are set for practical reasons rather than microbiological ones. Below 40 percent, powders and pigments fly, static makes weighing unreliable, and dried botanicals crumble to dust in the hand. Above 60 percent, the dew point climbs to within a few degrees of room temperature so cold surfaces start to wet, hygroscopic materials such as clays, starches and glycerin gain measurable weight on the shelf, and surface mould becomes possible on damp walls and cardboard stock.

A cheap digital hygrometer is often 5 to 10 percentage points out of calibration. Check it before you trust it, using the saturated salt slurry method of ASTM E104: half fill a small pot with the salt, wet it to a slush with distilled water leaving undissolved crystals visible, seal the pot and the hygrometer together in a jar, and leave it 8 to 12 hours at a stable temperature.

Two points beat one, because a meter can read correctly at 75 percent and badly at 33 percent. Write the offset on the instrument and apply it, or replace the meter if it is more than 5 points out; the same discipline covered on weighing and calibration applies here. If the space runs high, a dehumidifier with a humidistat is the cheapest fix. A rolling water bath in the same room as open stock is a humidifier pointed at your packaging.

Condensation, worked

Condensation is the underrated route, because nothing looks wet. Any surface below the dew point of the surrounding air pulls water out of it, invisibly at first, as a film rather than beads.

Work a case. A box of tins has come in from an unheated garage and is at 20 C. The workshop is at 24 C and 60 percent RH, so the dew point is 15.8 C and the tins are 4.2 C above it. They stay dry, and you can fill them. Now change one variable. Let the room drift to 75 percent RH on a humid afternoon and the dew point rises to about 19.3 C, leaving 0.7 C of margin, which is inside the error of a cheap hygrometer. At 80 percent RH the dew point is 20.3 C, the tins are below it, and every one collects an invisible film before the balm arrives.

The rule that follows: keep at least 3 C between your coldest packaging surface and the room dew point, and bring packaging into the workshop the day before rather than the hour before. A cardboard case takes hours to equilibrate and the tins in the middle take longest.

Drying botanicals below 10 percent moisture

Fresh leaves and flowers are 70 to 85 percent water by weight. Herb material dried for storage conventionally lands at 8 to 12 percent residual moisture, and for oil infusion you want the bottom of that: below 10 percent, preferably 6 to 8. Anything wetter releases water into the oil during infusion, where it settles as a cloudy layer at the bottom of the jar and reappears as haze in the balm.

1. Dry at the right temperature for the part. Leaves and flowers hold their volatiles best air dried in the dark or in a dehydrator at 35 to 40 C. Roots, barks and resins tolerate 45 to 50 C. Above about 55 C you trade aroma and colour for speed.
2. Dry to brittleness, not to leathery. A leaf that folds is still wet. A leaf that snaps and a stem that breaks rather than bends is the field test, and it is a reliable proxy for the 10 percent line.
3. Confirm by weight if the batch matters. Weigh 5 to 10 g of dried material, hold it at 100 to 105 C for two hours, cool it sealed and weigh again. Loss on drying is the mass lost as a percentage of the start. It is the pharmacopoeial method in miniature and accurate enough to settle an argument.
4. Store sealed and full. A half empty jar in a humid room is a moisture reservoir that the dried material will equilibrate with.
5. Settle and decant after infusion. Let strained oil stand 24 to 48 hours. Water separates to the bottom as a distinct cloudy layer. Decant the clear oil and discard the last few millilitres rather than filtering them.

Method matters as well as material. A heat infusion at 40 to 60 C drives some residual moisture off; a sealed cold or solar infusion traps it. Full ratios and yields are on herbal infused oils, with the arithmetic on the infusion calculator. A wet infusion is the commonest failure in the category, which is why calendula salve specifies dried florets and nothing else.

Rendering and drying tallow at 100 to 105 C

Home-rendered fat arrives wet: raw suet carries water in the connective tissue and wet rendering adds more deliberately. The physics gives a free indicator. While free water remains, the fat cannot rise much above 100 C, because the energy going in is spent boiling water rather than heating fat. The temperature sits on a plateau, the surface bubbles, and a lifted spatula crackles.

The protocol is therefore to keep going. Hold the fat at 100 to 105 C until the bubbling and crackling stop and it goes visibly clear rather than cloudy, usually 20 to 45 minutes past the point where it looks finished. Strain hot through fine muslin, let it set, then check the underside of the block: any water that remained separates as a wet greyish layer to be scraped off and discarded.

Properly dried tallow keeps for months at room temperature and years frozen. Under-dried tallow goes off early and takes the balm with it, which is why tallow has an undeserved reputation for instability and why tallow balm is specific about starting material. Beeswax cleaned by melting under water needs the same treatment: separate it fully, then re-melt it dry.

Honey, aloe, hydrosol and glycerin

Honey is the interesting case, because it is genuinely self-preserving at 0.55 to 0.65 and people reason from that to safety in a balm. Droplets of it sit close to the 0.60 floor with no margin, so moisture from humid air during filling, or from a finger, moves them into osmophilic yeast territory. Honey also carries Clostridium botulinum spores, which is why it stays out of anything for infants however dry the product is; see balms for babies. For the character without the water, use a flavour or a dried powder, as on honey in balms.

Glycerin is a trap of a different shape. It brings no meaningful water but is strongly hygroscopic, so a balm containing it drifts up toward ambient humidity over months in a bathroom rather than staying where you filled it. Water activity is solute dependent too: FDA's compilation notes an organism inhibited at aw 0.975 in ethanol that grows at 0.89 in glycerol, so depressing aw with glycerol buys less than the number suggests. It will not stay dispersed in an anhydrous base either, so it beads out and reads as balm sweating. Limits are on glycerin and humectants.

Wet cleaning and the 30 minute dry down

Cleaning is necessary and it is also the largest scheduled introduction of water into your workshop. The fix is not to clean less, it is to separate wet work from dry work in time. Run the day in one direction: wet jobs first, then a dry down, then weighing, melting and filling, then wet jobs again at the end. Wash with hot water and detergent, sanitise the clean dry surface with 70 percent alcohol, and allow at least 30 minutes before the first ingredient is weighed. Thirty minutes is a working minimum for a wiped bench and air-dried stainless in a 40 to 60 percent room. A narrow-necked jug, a spouted pan or anything with a thread or a crimp needs longer, or an alcohol rinse to displace the water and flash off.

Three habits do the rest. Never wash tins, tubes or jars, because a washed container cannot be dried inside a rolled edge; buy them clean, store them sealed and inverted, and sanitise with alcohol only if you must. Never put a flat lid on a water bath, because condensate collects underneath and drips into the melt; use a domed lid, tilt it, or work uncovered. And cool filled containers uncapped, since capping warm traps vapour that condenses on the lid and falls back onto the product. See equipment.

In-process checks that cost nothing

Two checks catch a wet batch before it reaches packaging, and both take under a minute. Two more confirm it afterwards.

1. Melt clarity. A fully melted, fully dry oil and wax phase is transparent: look through it at a printed line and you should be able to read it. Water droplets scatter light, so a wet melt reads as a milky cast that does not clear with more heat or more time. Undissolved wax shows as discrete specks and does clear on holding; air rises out. If haze survives 15 minutes at 75 C, treat it as water.
2. The hot spatula test. Heat a clean stainless spatula until a flicked droplet of water flashes off instantly, take it off the heat, and place one drop of the melt on it. Dry balm spreads silently. Water gives an audible spit, a crackle, or a fine spray of fat. It is qualitative, and it finds tenths of a percent rather than parts per million, but that is the range that matters.
3. Weigh the infused oil. If you dry an oil deliberately, weigh the jug before and after: loss on a gentle 60 to 70 C hold is water leaving, and when the mass stops falling the free water has gone.
4. Keep a retained sample. One filled unit per batch at room temperature, inspected at 4, 8 and 12 weeks. Water declares itself as haze, a wet ring, off odours or visible colonies well before the best-before date. Log it against the batch code, as on batch records.

None of these detect the parts-per-million water oils carry as supplied, typically 300 to 1000 ppm. That is normal rather than a defect, and quantifying it needs Karl Fischer titration to ISO 8534 in a laboratory. It matters for oxidation, not microbiology; see rancidity and oxidation. Trace water plus free fatty acids is also what attacks a tinplate seam from the inside, behind rust spots on balm tins.

What the rules require, and what you have to decide yourself

In the EU and the UK, Regulation (EC) No 1223/2009 Article 8 requires manufacture in accordance with good manufacturing practice, with ISO 22716 as the recognised reference. Neither names a humidity figure or a drying temperature; both require defined procedures, followed and recorded. The 40 to 60 percent band, the 3 C dew point margin and the 30 minute dry down here are defensible working specifications, not legal limits, and once written into your procedures you are expected to show you meet them. In the United States, MoCRA directs FDA to make cosmetic GMP regulations and that rulemaking is still in progress, so the current expectation is the FD&C Act prohibition on adulterated products rather than a humidity rule.

Whether your product qualifies as microbiologically low risk on water activity, and so whether an ISO 11930 challenge test can be waived, is a judgement for the safety assessor who signs your Cosmetic Product Safety Report Part B, not for a website and not for the category name. Enumeration testing to ISO 17516 is a separate question that low water activity does not answer, because a dry product can still carry a high count. See safety assessment and CPSR.

The decision rule is narrow and it is the honest one. This site cannot tell you the water activity of your balm. No aw values have been published for shea, cocoa butter, beeswax or candelilla, so there is no table to look it up in, and measuring it needs an instrument costing more than most small makers' entire kit. What you can do is eliminate the routes, which is cheap and verifiable, then treat any haze, any spit and any bead of liquid under the surface as a wet batch. Remelting does not rescue one, because the water is still there once the visible boiling stops; the disposal decision is on water contamination in balm. If your product genuinely needs water in it, you are making an emulsion, and the answer is to formulate and preserve it as one, starting from what preservatives for anhydrous products can and cannot do.

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