My book mentions that water contracts on melting, but the book doesn't give any reason why it does so. It is mentioned that:

$1\,\mathrm g$ of ice of volume $1.091\,\mathrm{cm}^3$ at $0^\circ\mathrm C$ contracts on melting to become $1\,\mathrm g$ of water of volume $1\,\mathrm{cm}^3$ at $0^\circ\mathrm C$.

I searched on the internet but I failed to find any useful insight. Could someone please explain why water contracts on melting?

  • $\begingroup$ This may help - Examples of phase transition nuclei whose dynamics impede their own growth? $\endgroup$
    – mmesser314
    Jul 25, 2021 at 1:34
  • 2
    $\begingroup$ en.wikipedia.org/wiki/… Also includes antimony, bismuth, germanium, gallium, plutonium, and silicon. $\endgroup$
    – PM 2Ring
    Jul 25, 2021 at 14:06
  • $\begingroup$ After the first line did you mean "decreases with increase in pressure"? (instead of temperature) $\endgroup$ Jul 25, 2021 at 14:28
  • $\begingroup$ Type metal (an alloy) also slightly contracts when it melts. $\endgroup$ Jul 26, 2021 at 7:21
  • $\begingroup$ It's worth noting that liquid water is most dense at 4 deg C. It expands slightly as the temperature drops from 4 to 0. This is related to the hydrogen bonds that are present. $\endgroup$
    – Flydog57
    Jul 26, 2021 at 16:43

7 Answers 7


It's because of the crystal structure of the solids. When water freezes, the molecules form various structures of crystals which have empty gaps that cause the solid to be about 9% larger in volume than the liquid was. Metals usually form crystals when they freeze too, but they're often simpler crystals, if you will, and often don't have as much empty space in them as ice/snow does.


The melting phase transition transforms the long-range ordered-crystalline solid structure into the short-range-ordered average liquid structure. Looking at the process from the solid side, melting can be seen as the dramatic effect of a collective building-up of defects in the solid over a limited interval of temperature, eventually destroying the long-range order at the melting point.

The change of density accompanying the melting can be explained in terms of the kind of dominant defects, and these, in turn, depend on the solid structure. In particular, materials characterized by almost isotropic interactions between the molecules crystallize into compact three-dimensional structures, like the face-centered cubic (fcc) structure. In such a case, the dominant effect of the exponential growth of defects at the melting transition is a sudden decrease of density going from the solid to the liquid.

The situation is reversed in materials like water and elements like Bismuth or Silicon, where the molecular or atomic interactions are highly anisotropic. In the case of water, the shape of the molecule, and the important role of the hydrogen bond favor an open structure of the crystal in the same way as the anisotropic interaction between Silicon atoms favors the open diamond structure. At the melting point, the most frequent defects in such open structures induce a local and global collapse of the crystal structure, creating the conditions for a liquid phase at a higher density than the coexisting solid.

In summary, the reason for the increase of density at the melting point of some materials like water can be traced back to the presence of dominant anisotropic interactions favoring open crystalline structure in the solid phase.


You may try to find similarities between water and bismuth, both expand when solidify. Most other materials contract when solidify.

It mostly depends on crystalline structure. Water in form of ice happens to have crystalline structure that takes more space than liquid water.

water and ice


The reason is attributed to the hydrogen bonds that determine the structure of ice. The molecules of water in ice are arranged in a cage-like fashion, with rather hollow spaces in between them, thanks to the hydrogen bonds shown as dotted lines:
enter image description here

When ice melts, the energy supplied helps in breaking these hydrogen bonds, and the molecules of water come closer, hence increasing the density of water, or 'contracting it'. This process goes on till the water reaches about 4 degrees Celsius, at which water has the highest density. After this, water expands like any other liquid.
This property has an interesting, but equally important application in nature. As ice is less dense than water, it floats up to the surface of the water bodies and acts as a thermal insulator to the underwater world. This is a boon to aquatic animals and fish during harsh winters.
It also explains other properties of water like the unusually high latent heat of fusion and specific heat.


The structure of ice comes from hydrogen bonds. These occur because the electrons are more strongly attracted to the oxygen atom, so this atom is slightly negative, whereas the hydrogen atoms become slightly positive. You can think of the electrons as spending more time near the oxygen atom. When you put this together with the shallow V shape of a water molecule, the result is that the molecules come together with the hydrogen atoms on one molecule near the oxygen atoms on other molecules.

The resulting structure is not the smallest possible packing for the molecules. Water is denser than ice because in water the molecules can move into positions where there is closer packing as they move around. As the temperature of the water rises to approx 4°C the effectiveness of the hydrogen bonding decreases as the molecules move faster and so the water continues to contract. Above that temperature the speed of the molecules and collisions between them cause expansion, in the same way as for most substances.

Under very high pressures the hydrogen bond forces in ice can be overcome, and other, denser, forms of ice can exist.


It has something to do with water being a somewhat unusual compound that is denser and therefore, takes up less space as a liquid than a solid. Greater pressure pushes the material to take up less volume and liquid water takes up less volume than solid water.

If someone wants to give greater detail, feel free, but I think that's the gist of it.

  • $\begingroup$ This answer seems to only restate the question. $\endgroup$ Jul 28, 2021 at 14:47
  • $\begingroup$ @Chemomechanics High pressure creates a force that makes the material trend towards lower volume. Water is lower volume as a liquid than a solid at the temperature of the phase transition. That's an answer. It stood up better when it was the initial answer, which it was for a day or two. I invited people to give more detail which they have, so it doesn't stand up as well now. $\endgroup$
    – userLTK
    Jul 29, 2021 at 11:44

From the thermodynamic point of view, the Clausius–Clapeyron equation should be considered: dp/dT=lambda/[T(V2-V1)]

where lambda is the heat of transformation, p, T and V the usual meanings.

In the p-T phase diagram of pure H2O the equilibrium curve separating normal ice from water has a negative slope and hence dp/dT<0.

When you melt the ice, you must give heat and hence lambda is positive; obviously T>0.

Then (V2-V1) must be negative, that is V1 (ice volume) > V2 (water volume).

Generally speaking, when the equilibrium curve separating the liquid phase from the solid one is dp/dT<0, you have a larger volume in the solid; the reverse for dp/dT>0.

  • $\begingroup$ This answer seems to only restate the question. The negative slope is a consequence of expansion upon freezing, not a cause. $\endgroup$ Jul 28, 2021 at 14:46
  • $\begingroup$ I don't agree. In a phase diagram the coordinates are p and T; so the volume change depends on the latent heat sign. The negative slope has nothing to do with expansion, it is determined by the change of the melting temperature with pressure; volume is not considered. $\endgroup$
    – gryphys
    Jul 28, 2021 at 17:46
  • $\begingroup$ The phenomena are exactly equivalent. Every material whose phase-change temperature decreases with increasing pressure has a larger volume in the colder phase (as specified by the Clausius–Clapeyron equation, as you note). That says nothing about why water is this type of material. $\endgroup$ Jul 29, 2021 at 3:14
  • $\begingroup$ It seems to me the problem of the egg or the chicken dilemma. What I wrote is the explanation from the thermodynamic point of view. The previous answers face the problem from the structural point of view. This is the point of view of thermodynamic; in the C-C equation derivation, dG/dp=V is used. If you consider the problem from the structural point of view, you wil explain the volume change by considering how atoms are arrangd in the phases, from the thermodynamic point of view, you consider the thermodynamic quantities. It's the same, but viewed from different perspectives. $\endgroup$
    – gryphys
    Jul 29, 2021 at 8:23
  • $\begingroup$ it's always 'beatiful' to see anonymous downvoting without explanation or without evidencing (or being able to show) what is wrong. Not very scientific... $\endgroup$
    – gryphys
    Aug 27, 2021 at 16:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.