Does steam ever touch the spirit? Culinary steam, boiler chemicals and the water nobody specifies

Key takeaways

• In a column still, steam and wash are in direct contact. In the analyser column, “steam rising and wash descending through several levels” is the separation mechanism itself (Wikipedia, Column still) — the steam is not behind a coil wall.
• US law treats boiler chemicals as food additives. “Boiler water additives may be safely used in the preparation of steam that will contact food, under the following conditions,” with limits written per substance and per ppm in the steam — cyclohexylamine and morpholine at 10 ppm, diethylaminoethanol at 15 ppm, octadecylamine at 3 ppm (21 CFR 173.310).
• The EU writes the same rule as an outcome. “Steam used directly in contact with food is not to contain any substance that presents a hazard to health or is likely to contaminate the food” (Regulation (EC) 852/2004, Annex II, Ch. VII).
• India routes it through potability. FSSAI’s Schedule 4 requires ice and steam in direct contact with food to be made from potable water, with potability defined by IS 10500 : 2012 (FSSAI Licensing Regulations compendium).
• “Culinary steam” is an engineering specification, not an adjective. 3-A Accepted Practice 609-03 defines it as steam “free of entrained contaminants… suitable for use in direct contact with food products,” produced through a separator and a filter rated at 95% efficiency at 2 microns and larger, in 300-series stainless (3-A 609-03; Donaldson).

Yes, in more places than most specifications acknowledge. In continuous distillation the steam is injected into the column and meets the wash directly, and across every distillery steam is what sterilises the vessels, lines and hoses that afterwards hold new make, mature spirit and blending water. That makes the boiler a water treatment plant whose product is an ingredient — which is why three separate regulatory regimes have written rules about what may be dosed into it.

Where does steam actually touch the spirit?

Three places, in ascending order of how often they get written into a specification — which is to say, from rarely to never.

Inside a column still. The Coffey design works by countercurrent contact: “the first column (called the analyzer) in a column still has steam rising and wash descending through several levels, separating the wash into liquid water and alcohol vapour,” with “heat (usually in the form of steam) supplied to the base of the column” (Wikipedia). Where that heat is live steam rather than a reboiler, the steam and its condensate join the process stream. Anything volatile that rode in with the steam rises with the alcohol vapour into the rectifier.

In direct-injection heating. Mash cookers, wash heaters and hot-water sets are commonly heated by sparging steam into the liquid, because it is fast and thermally efficient. The condensate becomes part of the batch.

In cleaning and sterilisation. This is the universal case. Steam-in-place is how tanks, transfer lines, hoses and fillers are sanitised, and the condensate left behind sits on the product-contact surface until the next liquid displaces it — a blend, a reduced spirit, or the demineralised water about to be used for reduction. A producer can buy water to a part-per-billion specification and then wash the receiving vessel with something it never tested, which is the argument of our note on how purity drifts between the tanker and the tank.

Only cooling water is genuinely isolated from the product, and that is covered separately in our survey of where a distillery’s water actually goes.

Why is boiler water dosed with chemicals at all?

Because untreated feedwater destroys boilers, and the cure is deliberately volatile.

Two problems dominate. Hardness salts precipitate on the hottest surfaces as scale, wrecking heat transfer — which is why softening sits at the front of any boiler house, and why hardness units are a boiler engineer’s vocabulary before they are a blender’s. And carbon dioxide released from bicarbonate dissolves in returning condensate as carbonic acid, which eats the return lines from the inside.

The standard answer to the second problem is a neutralising amine — morpholine, cyclohexylamine, diethylaminoethanol — chosen precisely because it is volatile. It leaves the boiler with the steam, distributes itself through the system, and raises the pH of the condensate wherever it forms. The property that makes it work is the property that makes it a food-safety question: a corrosion inhibitor engineered to travel in steam will travel into anything that steam touches. Finland’s food authority states the operator’s duty plainly — such chemicals “must meet the requirements of chemical legislation,” and “chemicals that easily evaporate during a process are the most relevant with respect to foodstuffs” (Finnish Food Authority).

What do the rules actually say?

Three regimes, three drafting styles, one requirement.

The US list is the most useful document to read even outside the US, because it puts numbers on the question. Cyclohexylamine and morpholine are each capped at “10 parts per million in steam,” diethylaminoethanol at 15 ppm, octadecylamine — a filming amine — at 3 ppm, and trisodium nitrilotriacetate at “5 parts per million in boiler feedwater” (21 CFR 173.310). Note where the limit is measured: for the amines it is in the steam, not in the boiler drum, because the steam is the thing that reaches the food.

India’s route through IS 10500 is the one most likely to be under-read by a spirits producer, for the reason set out in our piece on FSSAI’s water standards for spirits: potable is a floor for safety, not a specification for flavour. A boiler fed with compliant municipal water is compliant, and may still carry a disinfectant residual into a vessel that will hold whisky.

How does boiler water get into steam in the first place?

By two routes that need separating, because they call for different controls.

Mechanical carryover is liquid boiler water entrained in the steam — droplets, not vapour. Spirax Sarco is blunt that “the most common cause of carryover… is a high Total Dissolved Solids (TDS) level,” because “as the boiler water TDS increases, the steam bubbles become more stable, and are more reluctant to burst and separate,” and the resulting foam “is carried over into the steam main” carrying “boiler water with a high level of dissolved and perhaps suspended solids.” The scale is worth pausing on: shell boilers run “in the range of 2 000 ppm for very small boilers, and up to 3 500 ppm for larger boilers” (Spirax Sarco) — set beside the single-digit µS/cm conductivity a demineralised blending water is bought to.

Vaporous carryover is the subtler mechanism: substances that travel in the steam phase rather than hitching a ride as droplets. The distinction is standard in boiler chemistry, where chemical carryover arises through foaming and “selective vaporous carryover… as a result of variances in the solvent properties of steam for different water phase impurities” (Veolia Water Handbook). The same source separates steam purity — “the amount of solid, liquid, or vaporous contamination in the steam” — from steam quality, which is only about how much liquid water the steam carries. A distillery buying a steam filter is buying quality; purity has to come from the feedwater and the dosing regime. Neutralising amines travel by the vaporous route by design, and a separator and a 2-micron filter will not touch them.

Could any of this actually reach a bottle of whisky?

Honestly: at a level nobody has published a measurement of. No survey of boiler-additive residues in spirits was available at the time of writing, and the mechanism is heavily diluted at every step — steam condensate is a small fraction of a mash, and a still is very good at leaving non-volatile material behind. A correctly run boiler house dosing an approved amine within its limit is not producing contaminated spirit.

What the mechanism deserves is a place on the risk register, for two reasons. First, sensory thresholds in spirits are brutally low for the wrong compound — the chlorophenol case, where taints are detectable from 0.1 µg/L, is the standing example, covered in our note on chlorine, chloramine and chlorophenol taint. Parts-per-million limits written for safety are not automatically tight enough for flavour. Second, the failure mode is an event rather than a drift: a foaming boiler, an overdosed condensate line, a steam cycle on a tank that then holds a blend. Those are far easier to rule out with a paper trail than to reconstruct after a batch tastes wrong.

What should a producer or buyer actually specify?

Six lines, none of them exotic, all of them answerable by an existing supplier.

1. Name the boiler treatment chemicals in use, by substance, with the safety data sheet and the supplier’s food-contact declaration. A product that cannot be matched to the 21 CFR 173.310 list or an equivalent food-grade approval does not belong on a plant where steam touches product.
2. Record the dose and the resulting concentration in steam, not just in the boiler. That is where the legal limit lives.
3. Map the direct-contact points: column still injection, direct steam heating, SIP on product vessels, hoses. Everything else is an engineering utility.
4. Apply 3-A 609-03 to those points — separator, filter at 95% efficiency at 2 µm and larger, sanitary 300-series stainless housings, differential pressure monitored (Donaldson).
5. Control feedwater TDS and blowdown, because carryover is a TDS problem before it is a filter problem (Spirax Sarco).
6. Displace and discard condensate after any steam cycle on a vessel that will hold spirit or blending water, and verify the flush with the same discipline as a bottling-line rinse.

A spirit’s water is not one stream but several, each with its own job and its own specification. Blending water goes into the bottle unchanged and therefore carries the tightest one. Steam is the stream nobody thinks of as water at all — and it is the only one that arrives at 150 °C, carrying whatever it was dosed with.

Glossary

• Culinary steam — steam suitable for direct contact with food; under 3-A 609-03, steam “free of entrained contaminants” and “relatively free of water in liquid form,” produced through a defined separator-and-filter arrangement.
• Plant steam (utility steam) — ordinary boiler steam for heating and power, carrying whatever the boiler house doses. Not qualified for direct food contact.
• Neutralising amine — a volatile alkaline compound (morpholine, cyclohexylamine, DEAE) dosed to raise condensate pH and prevent carbonic acid corrosion in return lines.
• Filming amine — a compound such as octadecylamine that forms a protective film on metal rather than neutralising acid. Limited to 3 ppm in food-contact steam.
• Carryover — boiler water leaving with the steam. Mechanical carryover is entrained droplets; vaporous carryover is material travelling in the steam phase.
• Steam purity — the amount of solid, liquid or vaporous contamination in steam. Distinct from steam quality, which measures only dryness.
• Blowdown — removal of concentrated boiler water to hold TDS below the level at which foaming and carryover begin.
• SIP (steam-in-place) — sanitising a closed vessel or line by admitting steam without dismantling it. Leaves condensate on the product-contact surface.

Frequently asked questions

Is boiler steam allowed to touch food at all? Yes, provided the water and its additives qualify. The US permits a named list of boiler water additives in “the preparation of steam that will contact food” with per-substance limits (21 CFR 173.310); the EU requires only that such steam contain nothing hazardous or contaminating (852/2004); India requires it to be made from potable water.

Does a pot-still distillery need to care? Less for the still itself — a pot still is heated by an external coil, pan or jacket, so the steam never meets the charge — but just as much for cleaning and sterilisation, which is where most direct contact happens in any distillery, regardless of still type.

Will a steam filter remove boiler treatment chemicals? No. A 3-A filter is rated for particulates — 95% efficiency at 2 microns and larger — and a separator removes liquid water. Volatile amines are in the vapour phase and pass straight through. They are controlled by choosing an approved additive and by dosing to the limit, not by filtration.

Should blending water be used as boiler feedwater? Generally no, and not for purity reasons. Demineralised water is aggressive toward boiler metallurgy without conditioning, and boiler feedwater has its own treatment logic. The two streams have different jobs — the point is that both need a written specification, not that they need the same one.

How would a contaminated-steam problem present in a spirit? As an event rather than a trend: a batch or vessel that tastes wrong after a boiler upset, a chemical changeover or an unusual steam cycle. The useful control is documentary — knowing which chemicals are on site and where the steam goes — backed by the sensory check on incoming water and on cleaned tanks described in our note on sensory testing of blending water.

Blenders International · Water Science series