RSC - Advancing the Chemical Sciences


The Mpemba Effect: A brief history

Ice cream


How did the phenomenon of hot water freezing faster 

than cold water come to be called The Mpemba Effect?

From Aristotle to Mpemba

The phenomenon by which, under certain conditions, hot water freezes faster than cold water has been observed by some of the world's finest minds, and has been used in everyday life for centuries.

In the 4th Century BC, Aristotle observed that "The fact that the water has previously been warmed contributes to it's freezing quickly: for so it cools sooner. Hence many people, when they want to cool water quickly, begin by putting it in the sun".

Roger Bacon in the 13th Century  used the effect to advocate the scientific method in his Opus Majus, while in his 1620 Novum Organum, Francis Bacon wrote that "slightly tepid water freezes more easily than that which is utterly cold". Rene Descartes also tried to solve the problem in 1637 and through the years, many scientists have attempted to explain it with little success.

In 1963, the question was catapulted back into the public eye when a young Tanzanian boy called Erasto Mpemba noticed that if he put hot ice-cream mix into the freezer, it froze more quickly than the cooled ice-cream mix of his fellow classmates. 

Mpemba's tenacity in the face of his teacher's initial dismissal of his observations and the ridicule of his classmates prompted him to repeat the experiment with hot and cold water, and to stand up and ask visiting physics lecturer Denis Osborne about the phenomenon.

In 1969, Mpemba and Osborne published a paper together (See "Cool?" in the box below), and the effect became known as The Mpemba Effect. In August 2012 Osborne had the following to say about his work with Mpemba:

"In line with his question made in front of his school staff and peers, we tested and found that hot water in Pyrex beakers on polystyrene foam in a domestic freezer froze before cooler samples. We attributed this to convection creating a continuing hot top, noting that:

  1. If two systems are cooled, the water that starts hotter may freeze first, but we did not look for ice and measured the time as that until a thermocouple in the water read 0C.
  2. A graph of 'time to start freezing' against initial temperature showed that the water starting at about 26C took longest to freeze (water starting at 60C took twice as long as water starting at 90C).
  3. Thermocouples near top and bottom showed a temperature gradient in the water. A hot starter kept a hot top while its lower levels were cooler than for the cool starter.
  4. An oil film on the water surface delayed freezing for several hours, suggesting that without this film, most of the heat escaped from the top surface.
  5. Changes in volume due to evaporation were small; the latent heat of vaporization for all the water to cool to 0C and start freezing accounted for less than 30% of the cooling.
  6. We used recently boiled water for all the trials, making dissolved air an unlikely factor.

We failed to check and report the ambient temperature in the freezer or its consistency during cooling. Lower ambient air temperatures might increase heat loss rates from the top surface, cause more rapid convection and increase the difference in freezing times.

Different mechanisms may assume more importance in different situations. We gave one example, with Mpemba's initial discovery in mind, and we wrote: 'rapid cooling of a system that starts hot may be accelerated if it establishes thermal contact with the case of the freezer cabinet through melting the layer of ice and frost on which it rests'."

Critical analysis of selected publications

Critical analysis of selected publications on the Mpemba effect

Downloadable Files

Cool? A paper by Erasto Mpemba and Denis Osborne, 1969
A Journal published then by The Institute of Physics and The Physical Society
PDF iconPDF (110k)  

PDF files require Link icon Adobe Acrobat Reader