Why alcohol plus water is less than the sum of its parts
Half a litre of ethanol and half a litre of water make about 965 mL. Nothing evaporates, nothing leaks — and if you blend spirits, this is the number that decides whether you hit your label strength.
It sounds like a trick question. Measure 500 mL of pure ethanol, measure 500 mL of water, pour them together, and the mixture comes to roughly 965 mL. Nothing has escaped. Weigh it and the mass is exactly what you put in. The volume simply isn't conserved.
What's actually happening
Volume, unlike mass, is not an additive property of a mixture. The volume a liquid occupies depends on how its molecules pack together, and when you mix two different liquids, they repack.
Water is unusual to start with: its hydrogen bonds hold it in a relatively open, structured arrangement. Ethanol molecules are a small hydroxyl group attached to a bulkier ethyl tail. The hydroxyl end hydrogen-bonds happily with water, and the ethyl tail is small enough to tuck into the gaps in water's open structure. Molecules that were previously sitting in each other's way end up nesting, and the total space needed falls.
The formal name for this is excess molar volume — the difference between the real volume of a mixture and the volume you'd predict by adding the pure components. For ethanol and water it's negative across the whole composition range, and it's strongly negative in the middle.
How much, and where
The contraction is not a fixed percentage. It's near zero at both ends — pure water and pure ethanol obviously don't contract — and peaks somewhere around 45–55% ethanol by volume, at about 3.6% of the total.
| Mixing (by volume) | Predicted | Actual (approx.) | Contraction |
|---|---|---|---|
| 100 mL ethanol + 900 mL water | 1000 mL | ~992 mL | ~0.8% |
| 250 mL + 750 mL | 1000 mL | ~978 mL | ~2.2% |
| 500 mL + 500 mL | 1000 mL | ~964 mL | ~3.6% |
| 750 mL + 250 mL | 1000 mL | ~972 mL | ~2.8% |
| 900 mL + 100 mL | 1000 mL | ~985 mL | ~1.5% |
Values are at 20 °C and rounded — the exact figures come from measured density tables, which is the only honest way to get them. There is no tidy closed-form equation for ethanol-water density; the relationship is empirical, tabulated by national metrology institutes, and interpolated between measured points.
That 50/50 peak is not a coincidence of the volume scale. It's near where the molar ratio gives the most opportunity for the two molecular shapes to interlock. Push toward either pure component and there's progressively less of the other to nest with.
What this means if you're blending
Suppose you have 40 litres of 70% ABV spirit and you want to bring it down to 40% ABV for bottling.
The naive route: alcohol content is 40 × 0.70 = 28 L of ethanol; to make that 40% you need 28 ÷ 0.40 = 70 L of final volume; so add 30 L of water. Do that and you will not land on 40%. The mixture contracts, so the final volume comes out under 70 L — and the same 28 L of ethanol in less than 70 L is stronger than 40%.
The two ways to get it right:
- Blend to volume, not by volume. Add water gradually until the batch reaches the 70 L mark in a calibrated vessel. Contraction takes care of itself because you're measuring the thing you actually care about.
- Work in mass. Mass is conserved, always. Convert the strengths to mass fractions using density tables, do the blend arithmetic on kilograms, and weigh the water in. This is what production blending software does, and it's why distillery blend sheets are usually in kilograms rather than litres.
Temperature, and why ABV always carries a temperature
Ethanol expands with temperature about four times as much as water does. So an alcohol-water mixture's volume strength shifts as it warms or cools, even though its mass composition is untouched. A spirit measured at 25 °C and again at 15 °C will give you two different ABV readings off the same hydrometer.
This is why alcohol strength is legally defined at a reference temperature — 20 °C in most of the world, 60 °F (15.56 °C) in the US — and why hydrometer readings get temperature-corrected before they mean anything. An uncorrected warm reading runs low; the spirit is stronger than the glass says.
Not just alcohol
Ethanol-water is the famous case because it's large, well-measured and commercially important. But volume non-additivity is general:
- Sulfuric acid and water contract, and release enough heat that the mixing order is a safety rule, not a preference.
- Methanol, isopropanol and glycols in water all contract, each with its own curve — you cannot substitute the ethanol table for isopropanol and expect the right answer.
- Some mixtures expand. Certain organic pairs have a positive excess volume, where the molecules pack together worse than they did apart.
The general lesson is the useful one. Mass is conserved and volume is not. Whenever accuracy matters in a liquid blend, do the arithmetic in mass and let a measured density table carry you back to volume at the end — rather than adding litres to litres and hoping the physics cooperates.