Why Is Water Densest at 4°C?

Pure water reaches its maximum density at about 4 °C (39 °F), not at its freezing point. That's backwards from almost every other liquid, which simply gets denser and denser the colder it gets, right up until it freezes solid. Water cools, contracts, and gets denser too — but only down to 4 °C. Below that, it starts expanding again even though it's still losing heat, all the way down to the ice point. The cause is a straightforward tug-of-war between ordinary thermal contraction and the hydrogen bonds that give water its unusual, open molecular structure.

See the exact number Water Density by Temperature Calculator Open →

An anomaly among liquids

Most liquids follow a simple, boring rule: cooler means denser. Molecules move more slowly as heat leaves, they pack closer together, and density climbs in a fairly straight line until the liquid freezes. Mercury, alcohol, and oil all behave this way. Water follows that rule too — but only part of the way.

As liquid water cools from room temperature, it does get denser, exactly as expected. That trend continues down to 4 °C, where density peaks. Cool it any further and the trend reverses: density starts falling as the temperature keeps dropping, all the way to 0 °C. Then, at the freezing point, density drops sharply again as the liquid turns to solid ice. No other everyday liquid does this over such a familiar temperature range, which is why chemists call it the density anomaly of water.

Density of water by temperature

The change is small in absolute terms, but the shape of the curve is the whole story. Density rises to a peak at 4 °C, then falls on both sides of it:

Notice how narrow the peak is: density at 0 °C and at 10 °C are both lower than at 4 °C, confirming that 4 °C really is a maximum, not just a point on a slope. The precise peak sits at 3.98 °C — rounded to 4 °C in virtually every reference because the difference is meaningless outside a calibration lab.

Overall, density falls by only about 4% between the 4 °C peak and boiling. That's a small swing for such a consequential property.

The two competing effects

Density is simply mass packed into a given volume, so anything that changes how tightly water molecules pack together changes its density. Two separate effects act on that packing as temperature changes, and they pull in opposite directions:

Thermal contraction. This is the ordinary effect present in every liquid. As molecules lose kinetic energy, they vibrate less and drift closer together, which raises density. On its own, this effect would make water keep getting denser all the way down to its freezing point — exactly like mercury or oil.

Hydrogen-bond expansion. This effect is specific to water. Each water molecule can form up to four hydrogen bonds with its neighbors, and as the molecules slow down, those bonds have time to lock into a more open, partly ordered network — a loose preview of the rigid lattice that ice will form at 0 °C. That open network takes up more space per molecule than a disorganized liquid does, which lowers density.

Above 4 °C, thermal contraction is the stronger of the two effects, so density rises as the water cools. Below 4 °C, the hydrogen-bond network becomes the dominant effect, so density falls even though the water keeps losing heat. The 4 °C point is simply where the two effects are perfectly balanced — the single temperature at which water is packed as tightly as it ever gets in liquid form.

The role of hydrogen bonds

A water molecule is bent, with two hydrogen atoms bonded to one oxygen atom. Oxygen pulls electrons toward itself more strongly than hydrogen does, so each molecule ends up slightly negative near the oxygen and slightly positive near the hydrogens. That imbalance lets each molecule attract the oxygen or hydrogen atoms of its neighbors, forming a hydrogen bond — weaker than the bonds inside the molecule itself, but strong enough to matter a great deal in bulk.

At room temperature, these bonds are constantly breaking and reforming as molecules jostle around; the liquid stays disorganized and reasonably tightly packed. As the water cools, the molecules move more slowly and the bonds persist longer, nudging clusters of molecules toward the same tetrahedral geometry that ice adopts, where each molecule is hydrogen-bonded to four others in a fixed, roomy pattern. That geometry is inherently spacious — it's the same structure that makes solid ice about 9% less dense than liquid water, which is why ice floats instead of sinking. Below 4 °C, enough of this open, ice-like structure has already formed in the still-liquid water to outweigh the ordinary contraction from cooling.

Why the peak isn't at the freezing point

It's a fair question: if hydrogen bonding wins out completely at 0 °C, why does the density maximum sit a few degrees above freezing rather than right at it? The answer is that the two effects don't switch on and off — they both operate continuously across the whole range, and their relative strength shifts gradually as temperature drops. Thermal contraction is strongest, relatively speaking, at higher temperatures, while the hydrogen-bond network needs the molecules to be slow enough to actually organize, which only becomes significant as the liquid approaches freezing.

Between 4 °C and 0 °C, the hydrogen-bond effect is already large enough to overpower contraction, so density falls even as the water keeps cooling. By the time the water actually freezes, the hydrogen-bond network locks into its full rigid, hexagonal form, and density drops again in one further step, from about 999.85 kg/m³ as liquid water at 0 °C to roughly 917 kg/m³ as solid ice.

Why it matters in the real world

This isn't just a laboratory footnote — it governs how bodies of water behave through the seasons.

Lake "turnover" in spring and fall — when the whole water column briefly reaches a uniform temperature and wind can mix it top to bottom — happens precisely because the water is passing through this same density-versus-temperature relationship. Without the 4 °C anomaly, deep lakes would freeze from the bottom up, and most freshwater ecosystems as we know them simply couldn't exist.

What changes the number

Dissolved salt. Salt ions interfere with the hydrogen-bond network, weakening the anomaly and lowering the freezing point at the same time. At typical ocean salinity, seawater's theoretical density maximum falls below its freezing point, so seawater never actually reaches it — it just keeps getting denser all the way down until it freezes, more like an ordinary liquid.

Pressure. Higher pressure suppresses the hydrogen-bond network slightly and shifts the temperature of maximum density a little lower. The effect is small at everyday pressures and only becomes significant at depths found in the deep ocean.

Purity. Dissolved gases and minerals in ordinary tap water have a negligible effect on where the density peak sits — far smaller than the effect of salt at seawater concentrations. The 4 °C figure quoted throughout this guide assumes fresh, essentially pure water.

The bottom line

Water is densest at about 4 °C because that's the single temperature where ordinary thermal contraction and the expanding, ice-like hydrogen-bond network exactly balance each other. Above 4 °C, contraction wins and density rises as the water cools. Below it, the hydrogen-bond network wins and density falls even as the water keeps losing heat. That reversal — rare among liquids — is why ice floats, why lakes freeze from the surface down, and why deep water stays liquid and life-sustaining all winter long.