It is a well known fact that freezing point decreases on adding impurities. I feel that it should increase on adding impurities.

Reason: I have watched this video on supercooled water. It says that when ice forms, it needs some nucleation points. If water is pure and left undisturbed, then there would be no nucleation points and water won't freeze even below $0^o$C.

So, my reasoning is that when impurities are added, they provide additional nucleation points where water crystals can gather. This would cause water to freeze faster.

But it doesn't happen. Why?

  • 1
    $\begingroup$ It depends on the "impurity." If the impurity dissolves then that is one kind of impurity. If the impurity doesn't dissolve, then it could provide nucleation sites at the freezing point of water. $\endgroup$
    – MaxW
    Dec 15, 2015 at 17:57
  • $\begingroup$ The kinetics (nucleation sites) is very different from the thermodynamics (Gibbs free energy) of the situation. Whether nucleation occurs quickly or slowly does not impact whether a particular phase is thermodynamically stable. Thermodynamically, adding impurities will always reduce the melting point (enthalpy change is linear/polynomially influenced by impurity concentration, while entropy is logarithmically related). $\endgroup$
    – Jon Custer
    Dec 15, 2015 at 18:17
  • $\begingroup$ In a hypothetical case that the impurity is more soluble in solid ice than in liquid, then would the freezing point increase? $\endgroup$
    – Curt F.
    Dec 16, 2015 at 2:44

2 Answers 2


No, but good question. Freezing is a thermodynamic process. Nucleation is a kinetic process.

Kinetic Explanation: At the freezing point of water, water wants to become ice (thermodynamically driven). In order for freezing to initiate and ice to form a temperature dependent minimum number of water molecules must arrange them selves in a crystal or else the particle will dissolve and no freezing will occur. Impurities or containers with rough surfaces can help arrange and hold molecules which allows freezing to occur with a fewer number of molecules.

Thermodynamic Explanation: Freezing is a thermodynamic process dictated by $\Delta G_f = 0 = \Delta H_f -T\Delta S_f$. When water freezes it must dispel the impurities to form an ice crystal (reverse mixing). Since mixing increases entropy, this means that for water to freeze the negative change in entropy ($\Delta S_f$) is larger than that of pure water. This must be compensated by a lower temperature ($T$) neglecting and differences in ethalphy. The enthalpy ($\Delta H_f$) of mixing will contribute as well to the freezing temperature of water and must be accounted for. In the case of Sodium chloride though the enthalpy mixing is positive which would favor a higher melting point, but the entropy contribution is large and magnified by the temperature such that entropy dominates and reduces the temperature of freezing.

If you look at the phase diagram of water and salt you will see that the mixing of the salt and water depresses the melting temperature due to the increased change in entropy from mixing water and salt to make brine from ice and salt.

enter image description here

  • $\begingroup$ Nice answer. This makes sense. But is the kinetic part less dominant or is it completely irrelevant? $\endgroup$
    – Yashbhatt
    Dec 16, 2015 at 5:26
  • $\begingroup$ There is a molecular process for water to become ice and the kinetic part perimts the water to undergo that process. Even if the water were to be at -30C, without the kinetic process to form ice, the water can not freeze. $\endgroup$
    – A.K.
    Dec 16, 2015 at 15:59
  • $\begingroup$ So, if I were to add some salt to this water at $-30^oC$ it would freeze above that temperature, right? $\endgroup$
    – Yashbhatt
    Dec 17, 2015 at 4:59
  • $\begingroup$ no it would freeze immediately as the salt surface would provide nucleation sites for freezing. $\endgroup$
    – A.K.
    Dec 17, 2015 at 10:26
  • $\begingroup$ No. What I meant was that if I were to take that sample of water, heat it to normal temperature and then add salt, then would it freeze below or above $-30^0C$? $\endgroup$
    – Yashbhatt
    Dec 18, 2015 at 6:01

Think in terms of needing to remove the dissolved substance: the water that crystallizes is (fairly) pure $\ce{H2O}$, and the remainder is a concentrated solution (e.g. brine, if the solute is $\ce{NaCl}$. Work is done to separate the water from brine, so it takes more "effort" to freeze than does pure water.

There are many good explanation for the phenomenon on the web, e.g. Physics Animations.

BTW, some bacteria nucleate water freezing, such as Pseudomonas syringae. The protein from these bacteria is used in artificial snow-making. The water still has to be below 0 C to freeze, but supercools less.


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.