Can water be too pure for whisky? The case against remineralising blending water

Key takeaways

• The “too pure” worry is real — for drinking water. The WHO review of the topic opens by noting that demineralised water “is highly aggressive and if untreated, its distribution through pipes and storage tanks would not be possible,” and that “distilled water has poor taste characteristics” (Kozisek, WHO).
• Drinking-water guidance therefore sets floors, not ceilings. The 1980 WHO report recommended “the minimum TDS in drinking water should be 100 mg/L” with “an optimum TDS … about 200-400 mg/L,” a minimum calcium of 30 mg/L and bicarbonate of 30 mg/L; European law went further, requiring a minimum hardness for softened or desalinated supplies (Kozisek, WHO).
• None of that logic survives contact with a spirit. Those recommendations are built on drinking roughly two litres a day. Reduction water at WHO’s optimum 400 mg/L would put under 200 mg of dissolved solids into a 700 mL bottle — and a fraction of a milligram-scale contribution into a dram.
• Minerals in reduction water are a defect risk, not a nutrient. Spirits chemist Gary Spedding puts it bluntly: “Limestone water is great for mashing, but purified water is the standard for reduction or dilution of a spirit before bottling,” and “as little as two ppm of calcium in the final product can cause precipitation” (Artisan Spirit).
• The pH scare is a measurement artefact. Fresh ultrapure water “can read anywhere between pH 5.0 and 8.0” because of atmospheric CO₂, and its buffer capacity is “about 1000 times less” than an acetate buffer — far too little to move a spirit already sitting between roughly pH 3.4 and 4.7 (Thermo Fisher; Chemosensors).

No — not for the job blending water is asked to do. Water can absolutely be too pure to drink as a sole supply, and the public-health literature on that is substantial and settled. But reduction water is a food ingredient added to a finished spirit, not a daily water supply, and the two are governed by opposite design goals: drinking water is specified upward from a health floor, while blending water is specified downward toward zero, because every mineral it carries is a potential haze, scale or flavour defect in the bottle.

Why does demineralised water have a bad reputation at all?

Because for its original application — supplying whole populations with desalinated seawater — it genuinely was a problem, and the people who found that out documented it thoroughly.

Frantisek Kozisek’s chapter for the World Health Organization, Health risks from drinking demineralised water, defines the material precisely: “water almost or completely free of dissolved minerals as a result of distillation, deionization, membrane filtration (reverse osmosis or nanofiltration), electrodialysis or other technology,” in which “TDS could be as low as 1 mg/L” (Kozisek, WHO). That is, near enough, the same material a distillery reduces with.

The chapter then lists why early desalination programmes concluded such water was not “fully appropriate for consumption”:

• It attacks its own container. “Demineralised water is highly aggressive and if untreated, its distribution through pipes and storage tanks would not be possible. The aggressive water attacks the water distribution piping and leaches metals and other materials from the pipes and associated plumbing materials.”
• It tastes of nothing. “Distilled water has poor taste characteristics.” Elsewhere the chapter notes that water below roughly 50 mg/L TDS “can have negative taste characteristics” and is “reported to be less thirst quenching.”
• It removes minerals from the diet. Low-mineral water “markedly … increased diuresis (almost by 20%, on average)” and increased urinary loss of sodium, potassium, calcium and magnesium.

Those are three serious objections — and all three are objections to drinking it in quantity, continuously, as your only water. Hold that thought.

What minimum mineral levels do the drinking-water rules actually set?

Real ones, and they are worth quoting because they set the benchmark that “too pure” is measured against.

Working from experimental studies in rats, dogs and human volunteers, the 1980 WHO report “recommended that the minimum TDS in drinking water should be 100 mg/L,” with “the optimum TDS … about 200-400 mg/L for chloride-sulphate waters and 250-500 mg/L for bicarbonate waters.” It also “recommended that the minimum calcium content of desalinated drinking water should be 30 mg/L” and “a bicarbonate ion content of 30 mg/L as a minimum essential level needed to achieve acceptable organoleptic characteristics, reduced corrosivity, and an equilibrium concentration for the recommended minimum level of calcium” (Kozisek, WHO).

Later work refined the figures downward for calcium and upward in specificity for magnesium — suggesting “the minimum magnesium content of drinking water should be 10 mg/L and the minimum calcium content should be 20 mg/L,” with optima “of about 20-30 mg/L” magnesium and “about 50 (40-80) mg/L” calcium (ibid.).

And this made it into law. The chapter records that “the first European Directive established a requirement for minimum hardness for softened or desalinated water (≥ 60 mg/L as calcium or equivalent cations)” (ibid.). Which is the cleanest possible statement of the point: in drinking-water regulation, purifying water too far is a compliance failure you must correct by adding minerals back.

Drinking-water figures from Kozisek, WHO.

Does any of that carry over to a bottle of whisky?

Almost none of it, and the reason is dose.

Every health recommendation above is calibrated to a person drinking on the order of two litres of water a day, every day, for years. Reduction water reaches a drinker through a very different route. Proofing water “can make up as much as 50–60% (or more, in some cases) of the final bottled product” (Caley Shoemaker, Distiller magazine) — a large share of the bottle, and a trivial share of anyone’s water intake.

Run the arithmetic at 40% ABV. Roughly 60% of a 700 mL bottle is water, so about 420 mL. Suppose you insisted on remineralising that water all the way to WHO’s optimum drinking figure of 400 mg/L:

• Per bottle: about 168 mg of added dissolved solids.
• Per 30 mL dram: about 7 mg.
• Against a daily reference intake of two litres of 400 mg/L water — 800 mg per day — a dram supplies under 1%.

So the nutritional case for remineralisation collapses on contact with a spirit: you would be adding a nutritionally irrelevant quantity of minerals to a product nobody drinks by the litre. What you would not be adding is nothing — because in a hydro-alcoholic matrix, a few milligrams of the wrong ion is a defect.

What actually happens if you put minerals into reduction water?

You buy yourself the haze problem that purification exists to prevent.

The mechanism is that ethanol is a far worse solvent for inorganic salts than water is. Sodium chloride dissolves to 36.0 g per 100 g of water at 25 °C but only 0.065 g per 100 g of ethanol — a difference of more than 500-fold (PubChem, citing the CRC Handbook and Kirk-Othmer). A 40% ABV spirit sits between those two solvents, so ions that were comfortably dissolved in your remineralised water arrive in a liquid that holds them far less willingly.

Calcium is the specific offender, because the spirit brings its own reaction partner. Oxalic acid is present in grain and extracted from oak, and it meets calcium in the reduction water to form calcium oxalate — the “floc” or “whisky diamonds” that appear in bottles months or years after filling. The trade guidance is unambiguous. Gary Spedding, writing in Artisan Spirit, states that “limestone water is great for mashing, but purified water is the standard for reduction or dilution of a spirit before bottling,” that “water must be purified, either by reverse osmosis (RO) or deionization,” and that “the introduction of calcium would cause calcium oxalate precipitation in the bottle.” On the threshold, he relays that “as little as two ppm of calcium in the final product can cause precipitation” (Artisan Spirit).

Two parts per million. WHO’s minimum recommended calcium for drinking water is 20,000 parts per billion — ten times the concentration that a spirits chemist identifies as a precipitation risk. The two specifications are not in tension; they are answering different questions. Our note on why whisky goes cloudy when you add water separates this kind of mineral floc, which is a fault, from the harmless louching that any non-chill-filtered whisky shows on dilution.

But doesn’t demineralised water taste flat — and won’t that flatten the whisky?

The first half is true. The second does not follow, and this is where the honest disagreement in the industry sits.

Demineralised water does taste flat, for the reason WHO gives: taste in water comes largely from its dissolved minerals, and removing them removes the taste. But a whisky is not carrying water as its flavour; it is carrying congeners, cask extractives and ethanol, and the water’s job is to be the one component that adds nothing of its own — a position we set out in does the type of water you add to whisky actually change the taste?.

That said, not every distiller agrees, and it would be dishonest to pretend otherwise. Distiller magazine frames the split fairly: “water sources such as springs, wells and aquifers tend to have higher mineral content and unique flavor profiles that can add a sense of place to bottled spirits,” while “some distillers prefer to let their other ingredients tell the story of their spirit or brand, and in those cases, a pure clean flavor profile is desirable in your water.” RO water, the same piece notes, “is very consistent in flavor and provides a clean, characterless additive to spirits,” and deionised water “is very consistent in flavor, but it lacks uniqueness or character” (Shoemaker, Distiller).

Read that carefully and the disagreement turns out to be narrower than it looks. Nobody in that account is arguing for remineralising purified water. The argument is between using a mineral-bearing source in the first place — the spring-water terroir tradition — and purifying. A distillery that has already installed RO and then doses minerals back in has taken on the haze risk of the first approach and the cost of the second, with the batch-to-batch variability of neither properly controlled.

Is demineralised water acidic, and will it drag a spirit’s pH down?

No, and this is the most common technical objection worth defusing, because it rests on a measurement artefact.

Purified water does read low on a pH meter. Thermo Fisher answers the question directly: asked whether dispensed ultrapure water should be neutral, the reply is “No. This is a common myth of ultrapure water … Fresh ultrapure water can read anywhere between pH 5.0 and 8.0. As time passes, the CO₂ absorption in the water creates carbonic acid, which causes the pH level to drop” (Thermo Fisher).

Two things follow. First, the reading itself is unreliable: “pH measurements of ultrapure water tend to be unstable and can be inaccurate due to its low ionic strength and low buffering capacity; therefore, pH is not considered a measurement of ultrapure water quality” — indeed ASTM D1193, the reagent-water specification, “does not list pH readings for Type I, II or III water” (ibid.). Second, and decisively, a low pH reading is not the same as acid content: “the acidity and buffer capacity of low pH pure water is minimal … the pH of ultrapure water and an acetate buffer are very similar, but the acidity and buffer capacity of the pure water is about 1000 times less” (ibid.).

Set that against the spirit. Whisky sits between roughly pH 3.4 and 4.7 (Chemosensors), buffered by its own acetic, lactic and wood-derived acids. A liquid with a thousandth of the buffer capacity does not move it. This is also why judging blending water by its pH figure alone is a mistake — see pH and TDS explained for how to read a spec sheet properly.

So is the “aggressive water” warning irrelevant?

No — this is the one drinking-water concern that transfers, and it matters. WHO’s point that “low-mineralized water is unstable and therefore, highly aggressive to materials with which it comes into contact” and “more readily dissolves metals and some organic substances” applies just as much in a distillery as in a municipal main (Kozisek, WHO).

The answer, though, is not to remineralise the water. It is to specify the pipework, tanks and fittings that touch it — stainless steel of the right grade, appropriate polymers, no galvanised or copper runs, no long dead legs. Purified water is a demanding fluid to store and move, which is the subject of our note on what storage and pipework do to a blending water spec. Solving a materials problem by degrading your product specification is the wrong trade.

What should a buyer take from this?

1. Do not benchmark blending water against drinking-water minima. IS 10500 and its equivalents are a safety floor for the water; they are not a target for a reduction spec, as we set out in FSSAI, IS 10500 and the industry floor.
2. Treat any offer of “remineralised” or “mineral-enhanced” proofing water as a red flag. It is a drinking-water product being sold into a spirits application.
3. Ask for calcium specifically, not just TDS. At a two-ppm threshold for precipitation risk, a bulk TDS figure can hide the ion that actually matters.
4. Ask how the water is stored and delivered. The aggressiveness of purified water is real; the control is materials and hygiene, not minerals.
5. If you want mineral character, source it, don’t dose it. That is a legitimate stylistic choice — but it belongs at the water source, with the haze risk understood and managed, not at the end of a purification train.

Glossary

• Demineralised water — water almost or completely free of dissolved minerals, produced by distillation, deionisation, membrane filtration, electrodialysis or similar; TDS can be as low as 1 mg/L.
• Remineralisation — deliberately adding calcium, magnesium or bicarbonate back to purified water, standard practice in desalination for drinking supply and inappropriate for spirit reduction.
• Aggressive (or corrosive) water — low-mineral water’s tendency to dissolve metals and other materials from the surfaces it contacts, because it is chemically unsaturated.
• Buffer capacity — a solution’s ability to resist a change in pH when acid or base is added. Purified water has almost none, which is why its pH reading is both unstable and inconsequential.
• Calcium oxalate (“floc”, “whisky diamonds”) — a crystalline precipitate formed when calcium from reduction water meets oxalic acid from grain and oak; insoluble in both water and alcohol.
• Salting out — the reduced solubility of inorganic salts in a water-ethanol mixture compared with water alone, a consequence of ethanol’s lower dielectric constant.
• TDS (total dissolved solids) — the total mass of dissolved material per litre; a floor parameter in drinking-water guidance and a ceiling parameter in blending-water specification.

Frequently asked questions

Is demineralised water safe to put in a spirit? Yes. The health concerns in the WHO literature relate to demineralised water used as a population’s principal drinking supply, at roughly two litres a day. As a diluent inside a bottled spirit it contributes a nutritionally negligible fraction of anyone’s mineral intake, while removing a defect risk from the product.

Should I ask my supplier to remineralise blending water to meet IS 14543 or a similar packaged-water standard? No — those standards govern water sold as a drink. Blending water is an ingredient in an alcoholic beverage; India’s Alcoholic Beverages Regulations point at the drinking-water standard IS 10500 as a floor for safety, and a proofing specification is written far below that floor deliberately.

Won’t purified water make my whisky taste thin? Thinness on dilution is a function of how far you have reduced the spirit and how it was filtered, not of the water’s mineral content. Purified water is chosen precisely because it contributes no taste of its own — the reasoning behind why water opens whisky.

My blending water certificate shows pH 5.6. Should I reject it? Not on that basis alone. Fresh ultrapure water reads anywhere from pH 5.0 to 8.0 depending on how much atmospheric CO₂ it has absorbed, and pH “is not considered a measurement of ultrapure water quality” (Thermo Fisher). Conductivity, TOC and the ion panel tell you far more.

Some distilleries make a feature of their mineral-rich spring water. Are they wrong? Not wrong — making a different choice, and generally at the source rather than by dosing. Mineral-bearing reduction water can carry a sense of place, and it also carries a real precipitation risk that has to be managed, often by chill filtration. What almost nobody does is purify water and then add minerals back into it.

Blenders International · Water Science series