The observation sounds wrong on its face. A hotter object has more thermal energy. Cooling it to freezing should take longer than cooling an already-cold object to the same point. This is what intuition says, and intuition is not entirely wrong, but it is not the whole story.
In 1963, a Tanzanian schoolboy named Erasto Mpemba was making ice cream with his classmates at Mkwawa Secondary School. The recipe required boiling the milk and sugar mixture first, then cooling it before putting it in the freezer. Running out of time and freezer space, Mpemba put his hot mixture directly in the freezer rather than waiting for it to cool. It froze before his classmates’ pre-cooled mixtures. He asked his physics teacher about it. The teacher told him he must be mistaken.
Mpemba kept asking. A few years later, University of Dar es Salaam physicist Denis Osborne visited the school, and Mpemba asked him about it. Osborne took it seriously enough to test it, and the two published a paper in 1969 demonstrating the effect under controlled conditions. It has carried Mpemba’s name ever since.
The Mpemba effect is real but intermittent: it appears under some conditions and not others, and for decades nobody agreed on why. Proposed mechanisms proliferated. Hot water evaporates faster, losing mass, so there is less water to freeze. Hot water releases dissolved gases, and gas-free water freezes differently. Convection currents in hot water create a more efficient heat-exchange geometry. The freezer itself can be affected: the hot container melts frost on the shelf, creating better thermal contact with the freezer floor.
All of these are real effects. None of them fully explained the observations across all conditions.
A more complete account came in 2016, when a team at the University of Zaragoza published a theoretical analysis suggesting that the key was something called hydrogen bond stretching. Water molecules bond to each other via hydrogen bonds. In hot water, these bonds are stretched and store energy. As the water cools, those bonds relax and release stored energy, effectively cooling the water faster than expected. Cold water, with its bonds already at equilibrium, has no such stored energy to release.
This mechanism, if correct, would explain why the effect is conditional: it depends on the water’s temperature history, the dissolved content, the container, and the freezer setup. Some configurations produce a strong Mpemba effect. Others produce none.
The Zaragoza explanation is not universally accepted, and experimental work confirming or disproving it precisely is ongoing as of 2024. What the debate has established is that the physical chemistry of water near freezing is more complex than the elementary picture suggests.
The lesson Mpemba drew from his own story, which he recounted in interviews, was that he had been lucky to find someone willing to listen. The physics teacher’s dismissal was not unusual: a student claiming that a hot liquid froze faster than a cold one sounds like it must be a measurement error. The willingness to check, rather than assume the intuitive answer, is the thing that made the science happen.
The effect is named for a schoolboy who noticed something that experts told him wasn’t real. He noticed correctly.
