The Mpemba Effect: Why Hot Water Sometimes Freezes Faster Than Cold

A Tanzanian Student Challenged His Physics Teacher

In 1963, Erasto Mpemba was a secondary school student in Tanganyika (now Tanzania) making ice cream with his classmates. The protocol called for boiling a milk mixture, cooling it, then placing it in the freezer. Mpemba, rushing to secure a spot in the freezer, skipped the cooling step and placed his hot mixture directly inside. It froze before his classmates' cooled mixtures. His teacher dismissed the observation. Mpemba persisted for years until physicist Denis Osborne of University College Dar es Salaam took the claim seriously, and the two published a joint paper in 1969.

The phenomenon had been noted before — by Aristotle, Francis Bacon, and Descartes — but Mpemba and Osborne's paper brought it into modern scientific discourse. It has remained controversial ever since.

Defining the Effect Precisely

The Mpemba effect refers to the observation that, under certain conditions, a body of hot water can reach a frozen state faster than an otherwise identical body of cold water when both are placed in the same sub-zero environment. The definition matters. "Freezing" can mean reaching 0°C, beginning to form ice, or becoming fully solid. Different definitions produce different experimental outcomes.

Much of the scientific disagreement about the Mpemba effect stems from inconsistent definitions. An experiment that confirms the effect under one definition may fail under another.

Proposed Explanations

Multiple mechanisms have been proposed. None has achieved consensus, and the effect likely results from a combination of factors that vary with experimental conditions.

Evaporation

Hot water evaporates faster. Evaporation removes mass and carries away latent heat, cooling the remaining water more rapidly. A container of initially hot water may lose 10–15% of its mass before reaching the temperature of the initially cold sample. Less water means less thermal energy to remove. The math supports this as a partial explanation.

Convection Currents

Hot water develops stronger convection currents due to larger temperature gradients between the water and its container walls. These currents transport warm water to the surface and cold water to the bottom more efficiently, potentially accelerating heat loss.

• Cold water may develop a stable stratification layer, with the warmest water insulated in the center
• Hot water's vigorous convection maintains contact between the warmest water and the cooling surfaces
• The effect depends heavily on container geometry — wide, shallow containers amplify convection effects
• Some experiments using stirred samples (eliminating convection differences) fail to reproduce the Mpemba effect

Dissolved Gases

Hot water holds less dissolved gas than cold water. Boiling drives out dissolved oxygen and nitrogen. Some researchers hypothesize that dissolved gases alter water's freezing behavior — possibly by affecting nucleation sites or changing thermal conductivity slightly. The evidence for this mechanism is weak, and controlled experiments have produced mixed results.

Supercooling

Water does not always freeze at 0°C. It can remain liquid well below freezing — a state called supercooling — until nucleation triggers crystallization. Previously heated water may supercool less than water that was never heated, possibly because heating alters the distribution of dissolved impurities or micro-bubbles that serve as nucleation sites.

The Experimental Challenge

Reproducing the Mpemba effect reliably is notoriously difficult. Small differences in experimental setup produce different outcomes.

• Container material (glass, metal, plastic) affects heat transfer rates
• Container shape and volume alter convection patterns and surface-to-volume ratio
• Placement in the freezer (on a shelf, on frost, in contact with a metal surface) changes conductive cooling
• Water purity, dissolved mineral content, and dissolved gas levels all influence freezing behavior
• Ambient humidity in the freezer affects evaporative cooling rate

A 2016 study by Burridge and Linden at the University of Cambridge demonstrated the effect under carefully controlled conditions but emphasized that results depended on which definition of "freezing" was used. A 2020 paper by Lu and Raz provided a theoretical framework suggesting the effect is real but requires specific initial conditions — it is not a universal phenomenon.

Skeptical Perspectives

Some physicists argue the effect is an experimental artifact. If evaporation, convection, and container contact effects are all carefully controlled, the thermodynamic expectation is straightforward: cold water should freeze first because it starts with less thermal energy to remove. The Mpemba effect, by this view, reflects uncontrolled variables rather than anomalous physics.

Beyond Water: Mpemba-Like Effects in Other Systems

In 2020, researchers demonstrated Mpemba-like effects in non-water systems — colloidal particles in solution and theoretical models of magnetic spin systems. These findings suggest the effect may reflect a general property of out-of-equilibrium systems rather than something unique to water. If confirmed, this would place the Mpemba effect in a broader context of non-equilibrium statistical mechanics.

More than 60 years after a Tanzanian student challenged his teacher, the question remains open. Hot water can freeze faster than cold water under specific conditions. Whether those conditions reveal deep physics or merely uncontrolled experimental artifacts is a debate that the scientific community has not yet settled.