Why can't alcohol be distilled above 96%? The ethanol–water azeotrope, and the water that never leaves the spirit

Key takeaways

• The ceiling is thermodynamic, not technical. Ethanol and water form a minimum-boiling azeotrope at 95.6% ethanol by mass, boiling at 78.2 °C — below pure ethanol’s 78.5 °C and water’s 100 °C. At that point “the vapor composition is exactly the same as the liquid,” so “it is impossible to get pure ethanol by distilling any mixture of ethanol and water containing less than 95.6% of ethanol” (Chemistry LibreTexts).
• 95.6% by mass is about 97.2% by volume, and that is the theoretical wall. Commercial rectification stops short of it: rectified spirit is “typically distilled in continuous multi-column stills at 96–96.5% ABV” (Wikipedia) — which is where the familiar “96%” on an ENA specification comes from.
• The law then codifies the physics. EU rules set the minimum strength of ethyl alcohol of agricultural origin at 96.0% ABV (Regulation (EU) 2019/787, Art. 5); India’s IS 6613 requires 96% v/v minimum at 20 °C (IS 6613:2002).
• Whisky law deliberately stops lower still. Scotch must be “distilled at an alcoholic strength by volume of less than 94.8 per cent so that the distillate has an aroma and taste derived from the raw materials” (SWR 2009); US whisky must be distilled below 95% ABV, and bourbon at “160° or less” (27 CFR 5.143).
• The practical consequence is a water consequence. Water that survives the still is a rounding error. Cut 96% ABV spirit to 42.8% and roughly 97% of the water in the bottle is water somebody added afterwards — which is why its specification is a product decision, not a plumbing one.

You cannot distil alcohol above roughly 96% ABV because ethanol and water stop behaving as two separate liquids at that composition. They form an azeotrope — a mixture whose vapour has exactly the same composition as the liquid it came from — so boiling it produces more of the same mixture rather than purer alcohol. Every still ever built runs into that wall, which means every distillate in the world arrives at the warehouse with water already dissolved in it, and every bottle on a shelf is a mixture of that unavoidable water and the water a producer chose to add back.

What is an azeotrope, and why does ethanol form one with water?

Distillation works on a difference in volatility. Heat a mixture, and the more volatile component is enriched in the vapour; condense that vapour, repeat, and you climb towards purity one theoretical plate at a time. The whole method depends on the vapour being different from the liquid.

For ethanol and water, that difference shrinks as the mixture gets stronger, and at one specific composition it vanishes. The mixture then “boils as if it were a pure liquid. It has a constant boiling point, and the vapor composition is exactly the same as the liquid” (Chemistry LibreTexts). A column with a thousand plates gains nothing over one with fifty, because there is no longer a gradient to climb.

The reason is molecular. Water holds itself together with an extensive hydrogen-bonded network; ethanol has a single hydroxyl group attached to a greasy ethyl tail. Mix them and the ethanol–water hydrogen bonds that form are weaker than the water–water bonds they replace, so both species escape into the vapour more readily than an ideal mixture would predict. That is a positive deviation from Raoult’s law, and a large enough positive deviation produces a boiling point lower than either pure component — a minimum-boiling azeotrope. The numbers make the point: ethanol boils at 78.5 °C and water at 100 °C, but the azeotrope boils at 78.2 °C (Chemistry LibreTexts).

Because the azeotrope is the lowest-boiling thing in the pot, it is what comes over first and what a rectifying column relentlessly concentrates towards. It is not an obstacle the still fails to overcome. It is the destination the still is aiming at.

What exactly is the ceiling — 95.6%, 96%, or 97.2%?

All three numbers are correct, and they are not measuring the same thing. This is the single most common confusion in the topic, so it is worth separating them cleanly.

The azeotrope contains “95.63% ethanol and 4.37% water (by mass)” (Wikipedia). Because ethanol is less dense than water, the same mixture expressed by volume is stronger: rectified spirit has “a practical limit of 97.2% ABV (95.6% by mass)” (Wikipedia).

Industry then stops a little short of the theoretical limit, for the ordinary reason that the last fraction of a percent costs disproportionate energy as the volatility difference approaches zero. Rectified spirit is “typically distilled in continuous multi-column stills at 96–96.5% ABV and diluted as necessary” (Wikipedia). That is the origin of the 96% figure that appears on every extra-neutral-alcohol certificate — including the ENA that underpins India’s IMFL category and most of the world’s gin and vodka.

Why does the law stop below the azeotrope?

Two different kinds of legal limit sit around this number, pointing in opposite directions, and both are really statements about water and flavour.

A floor for neutral alcohol. Where the intent is character-free alcohol, regulators set a minimum. Under EU law, ethyl alcohol of agricultural origin must have “no detectable taste other than that of the raw materials used in its production” and “its minimum alcoholic strength by volume is 96,0 %” (Regulation (EU) 2019/787, Art. 5). India’s Bureau of Indian Standards takes the same approach in IS 6613, which “prescribes the requirements and the methods of sampling and test for neutral spirit intended for use in the manufacture of alcoholic drinks” and sets ethanol content at a minimum of 96% v/v at 20 °C (IS 6613:2002). In other words: get as close to the azeotrope as commercial rectification allows, because congeners are what you are trying to leave behind.

A ceiling for whisky. Where the intent is character, regulators set a maximum, and they say why in the text itself. Scotch whisky must be “distilled at an alcoholic strength by volume of less than 94.8 per cent so that the distillate has an aroma and taste derived from the raw materials used in, and the method of, its production” (SWR 2009, reg. 3). US law defines whisky as an alcoholic distillate from a fermented grain mash “distilled at less than 95 percent alcohol by volume (190° proof),” and requires bourbon and rye to be distilled at “160° or less” — 80% ABV — while a distillate taken off above 160° proof falls into the separate class of light whisky (27 CFR 5.143).

Read together, these rules draw a boundary line just under the azeotrope. Above it, the product is alcohol. Below it, the product is whisky. The distance a distiller keeps from that line is the flavour budget — and in practice most single malt spirit runs far below the legal ceiling, in the high 60s to low 70s percent ABV.

How much of the water in a bottle actually came through the still?

Here is where the physics lands on a commercial fact. Every distillate carries water, but the amount is small and shrinks fast as distillation strength rises. Set against the water added at reduction, it almost disappears.

The table below takes one litre of distillate at each strength, reduces it to 40% ABV, and asks what share of the water in the finished liquid arrived with the distillate rather than being added afterwards. Figures are nominal volumes at a single strength step — they ignore volume contraction on mixing, cask losses, and the water added at cask filling strength — so read them as orders of magnitude, not gauging.

The pattern is the interesting part, and it inverts a common assumption. The purer your distillate, the less of your finished product is your own water and the more of it is water you have to source, specify and pay for. A pot-still malt taken off at 68% brings roughly a third of the bottle’s eventual water with it. A 96% ENA brings about a thirtieth. Everything else has to arrive in a tanker or come out of a treatment plant.

For a producer working from neutral spirit, that is not a footnote. Cut 96% ABV alcohol to India’s traditional 42.8% and something close to 97% of the water in the bottle is water added at reduction — the arithmetic set out in full in our note on how much water to add. The spirit supplies the alcohol and the congener profile. The water supplies the volume.

Can the azeotrope be broken — and should a distillery care?

It can, and the methods are well established, but none of them belongs anywhere near a beverage.

The classic laboratory and industrial route is azeotropic distillation with an entrainer: “adding benzene or cyclohexane to the water/ethanol azeotrope,” where “with cyclohexane as the entrainer, the ternary azeotrope is 7% water, 17% ethanol, and 76% cyclohexane, and boils at 62.1 °C” (Wikipedia). Introducing benzene into a drinks stream is obviously a non-starter. Pressure-swing distillation, which works for some azeotropes, does not help much here: “the composition of the water to ethanol azeotrope discussed earlier is not affected enough by pressure to be easily separated using pressure swings” (Wikipedia).

The modern industrial answer is adsorption on molecular sieves — “ethanol can be dried to 95% ABV by heating 3Å molecular sieves such as 3Å zeolite” (Wikipedia), the technique that produces anhydrous fuel ethanol. A 3Å pore admits water and excludes ethanol, so the water is stripped out physically rather than distilled away.

And a distillery should care about exactly none of it, because the whole exercise runs backwards from the beverage’s point of view. Anhydrous ethanol is bought by chemical plants and fuel blenders. A spirits producer’s next operation is to add water — a great deal of it — so paying an energy premium to remove the last 4% only to put water back is precisely the wrong trade. The relevant question was never how do I get the water out of the spirit. It is what is in the water I put back.

What should a trade buyer take from this?

Three things follow directly.

“96%” is a specification, not a boast. Any supplier quoting substantially higher than about 97.2% ABV for a potable, undried spirit is quoting something that conventional distillation cannot produce. Treat it as a red flag on the paperwork.

The stronger the base spirit, the more the water is your product. A vodka or an IMFL whisky built on ENA is, by volume, mostly the water you specified. That is why unaged and neutral-based categories are the least forgiving of a loose water spec, and why the numbers in our demineralised water specification for spirit dilution matter more there than anywhere else.

Distillation purity and water purity are separate guarantees. A certificate of analysis for 96% ENA tells you about congeners, acidity and esters. It tells you nothing about the conductivity, hardness, iron or microbiology of the water waiting in the reduction tank. Those are governed by a different specification and a different regulatory floor — and, as with the azeotrope itself, the legal floor sits well below what a good product needs.

Glossary

• Azeotrope — a mixture whose vapour has the same composition as the liquid, so it “boils as if it were a pure liquid” and cannot be separated further by distillation (Chemistry LibreTexts).
• Minimum-boiling azeotrope — one that boils below both pure components. Ethanol boils at 78.5 °C, water at 100 °C, the azeotrope at 78.2 °C.
• Positive deviation from Raoult’s law — the condition that produces a minimum-boiling azeotrope: unlike-molecule attractions weaker than like-molecule attractions, raising vapour pressure above the ideal prediction.
• Rectified spirit — highly concentrated ethanol from multi-column distillation, with “a practical limit of 97.2% ABV (95.6% by mass)” and a commercial norm of 96–96.5% ABV (Wikipedia).
• ENA (extra neutral alcohol) — the trade name for neutral spirit meeting a standard such as IS 6613, minimum 96% v/v at 20 °C (IS 6613:2002).
• Anhydrous ethanol — ethanol with the azeotropic water removed by adsorption or entrainer distillation; an industrial and fuel product, not a beverage one.
• Light whisky — the US class for grain distillate taken off above 160° proof and stored in used or uncharred oak (27 CFR 5.143).
• Blending water — the purified, low-mineral water used to reduce a spirit to bottling strength. See what blending water is.

Frequently asked questions

Why can’t you distil alcohol to 100%? Because at 95.6% ethanol by mass the vapour and the liquid have identical compositions, so further boiling produces more of the same mixture. “It is impossible to get pure ethanol by distilling any mixture of ethanol and water containing less than 95.6% of ethanol” (Chemistry LibreTexts). No amount of extra column height changes that.

Is the limit 95.6% or 96%? Both, in different units and contexts. 95.6% is by mass, which equals about 97.2% ABV — the theoretical wall. Commercial multi-column stills stop at 96–96.5% ABV because the final increment is disproportionately expensive (Wikipedia).

Does the azeotrope mean every spirit contains water? Yes, unavoidably. Even the strongest potable distillate is roughly 4% water by mass before anyone adds anything. In practice that water is trivial next to what is added at reduction — around 3% of the total water in a bottle cut from 96% ENA.

Why is Scotch distilled to less than 94.8% ABV? Because the regulation ties strength to flavour: it must be “distilled at an alcoholic strength by volume of less than 94.8 per cent so that the distillate has an aroma and taste derived from the raw materials used in, and the method of, its production” (SWR 2009). Distil closer to the azeotrope and you strip out the congeners that make it whisky.

Why is bourbon capped at 160° proof? Same logic, applied harder. US rules require bourbon to be distilled at “160° or less” and stored in charred new oak at “125° or less” (27 CFR 5.143) — a lower distillation ceiling that keeps more grain character in the spirit.

How is anhydrous ethanol made if distillation cannot do it? By removing the water another way: adsorption on 3Å molecular sieves, which admit water and exclude ethanol (Wikipedia), or entrainer distillation with cyclohexane or benzene (Wikipedia). Both produce industrial and fuel-grade alcohol, not drinks.

Does any of this change what water a distillery should buy? It sharpens the case. The higher the strength you distil or buy at, the larger the share of the finished bottle that is added water — up to roughly 97% of all the water in a neutral-based spirit. The specification for that water is a formulation decision, covered in blending water vs RO vs distilled.

Blenders International · Water Science series