The Mpemba effect is when very hot liquids or mixtures (like ice cream) freeze faster than colder ones when they have the same amount and are in similar conditions. Scientists are still not sure if this effect can be repeated, exactly what it means, or why it happens. It is named after Erasto Mpemba, a Tanzanian teenager who studied it scientifically in the 1960s for the first time, along with Denis Osborne.

The Mpemba effect was first noticed in ice cream and water, and later in other mixtures. It has been studied a lot in water, but results have been mixed, with some experiments showing no clear effect. It has also been studied in magnetic alloys, tiny mechanical systems, and quantum systems.

Definition

The Mpemba effect is a scientific phenomenon that is not clearly defined in all studies, which makes it hard to compare different experiments. Monwhea Jeng suggested a definition for the effect in water: "There is a group of starting conditions and a pair of temperatures where two bodies of water are the same in these conditions but have different starting temperatures. In this case, the hotter water freezes faster than the colder one." This definition does not explain whether "freezing" means when ice first appears on the surface or when all the liquid turns to ice.

A more general definition is: "A hotter system cools faster than a colder one when both are cooled to the same low temperature." For water, this means when the liquid is completely frozen.

The effect is named after Erasto Mpemba, a student from Tanzania. In 1963, while in Form 3 at Magamba Secondary School in Tanganyika, he noticed that a hot ice cream mixture froze faster than a cold one during a cooking class. Later, he attended Mkwawa Secondary School in Iringa. The school headmaster invited Dr. Denis Osborne from University College in Dar es Salaam to give a physics lecture. After the lecture, Mpemba asked Osborne, "If two containers with the same amount of water are placed in a freezer—one at 35 °C (95 °F) and the other at 100 °C (212 °F)—which one freezes first?" Osborne was unsure at first but tested the idea later and found that Mpemba's observation was correct. In 1969, Osborne and Mpemba published their findings together while Mpemba was studying at the College of African Wildlife Management.

Mpemba and Osborne tested their idea by placing 70 ml (2.5 imp fl oz; 2.4 US fl oz) of water in 100 ml (3.5 imp fl oz; 3.4 US fl oz) containers inside a refrigerator's icebox on a polystyrene foam sheet. They found that water starting at 25 °C (77 °F) took the longest time to begin freezing, while water starting at 90 °C (194 °F) froze much faster. They concluded that evaporation and dissolved air were not major factors in the process. In their experiment, most heat was lost from the surface of the water.

History

Ancient scientists, such as Aristotle, studied how heat affects the freezing of water. Aristotle observed that water that is warmed first freezes more quickly because it cools faster. He explained this idea using a concept called antiperistasis, which suggests that a quality becomes stronger when surrounded by its opposite. For example, warming water may help it freeze faster because it is surrounded by heat.

Francis Bacon noted that slightly warm water freezes more easily than water that is extremely cold. René Descartes, in his book Discourse on the Method, linked this phenomenon to his vortex theory, which described how water particles move. He wrote that water heated for a long time freezes faster than unheated water because some of its particles that are less able to resist evaporation escape as vapor during heating.

In 1775, Scottish scientist Joseph Black studied a specific case of this effect by comparing water that had been boiled with water that had not been boiled. He found that boiled water froze more quickly, even when evaporation was controlled. He also observed that stirring unboiled water made it freeze at the same time as boiled water. Additionally, he noted that stirring very cold unboiled water caused it to freeze immediately. Joseph Black then discussed Daniel Gabriel Fahrenheit’s description of supercooling, which is when water is cooled below freezing without turning into ice. He argued that boiled water was less likely to be supercooled than unboiled water.

Modern experimental work

Modern studies using freezers with known properties have observed the Mpemba effect, where water cools below freezing without freezing first. Water that starts at a lower temperature often reaches a colder supercooled state before freezing. Some studies measure the time it takes for a sample to begin freezing (the start of recalescence, when heat from freezing first appears), the time to fully freeze, or the time between the start of recalescence and complete freezing. Others measure the time it takes for a sample to reach the freezing point of the fluid before freezing begins.

In 1995, David Auerbach studied glass beakers placed in a cooling bath. The water in these beakers cooled to −6 to −18 °C (21 to 0 °F) before freezing. In some cases, water that started hotter froze first. Auerbach noted that the time for freezing to begin varied randomly, and the Mpemba effect was more common when the surrounding temperature was between −6 and −12 °C (21 and 10 °F). James Brownridge later studied different starting conditions and containers, measuring the time to recalescence and found that hot samples sometimes froze first, but this also depended on the container’s properties.

Writing for New Scientist, Mick O'Hare suggested starting experiments with containers at 35 and 5 °C (95 and 41 °F) to maximize the effect.

In 2021, John Bechhoefer described a method to reliably reproduce the effect. In 2024, Argelia Ortega and others studied small drops (1 to 20 mL) in a Peltier cell using a thermographic camera. They found that hotter drops froze faster than colder ones, with a greater difference for larger drops. Hot drops finished freezing sooner after recalescence and had smaller temperature spikes during freezing.

Some researchers have criticized studies of the Mpemba effect for not considering dissolved solids, gases, and other factors. Even experiments that observe the effect under certain conditions often fail to see it under others.

In 2006, Philip Ball, a reviewer for Physics World, wrote that even if the Mpemba effect is real, it is unclear whether the explanation would be simple or complex. He noted that experiments must control many factors, such as water type, temperature, dissolved gases, container materials, cooling methods, and refrigerator temperatures. Ball suggested that the effect might be observed if warmer water melts frost on a cooling surface, improving heat transfer.

In 2016, Burridge and Linden studied the time it took water to reach 0 °C without freezing. They reviewed earlier work and found that large effects seen in early experiments were not repeated in later studies. They concluded that there is no strong evidence supporting meaningful observations of the Mpemba effect.

The original classroom observations of the Mpemba effect involved fresh ice cream, a colloid, freezing in a freezer.

A generalized version of the Mpemba effect is "when a hotter system reaches equilibrium faster than a colder one when both are cooled to the same low temperature." This has been modeled for simple systems, such as single particles moving randomly.

In 2019, Klich, Raz, Hirschberg, and Vucelja predicted a "strong Mpemba effect," where cooling could happen exponentially faster at specific initial temperatures. In 2020, this effect was demonstrated by Avinash Kumar and John Bechhoefer in a single-particle colloid. In 2022, the same group observed an "inverse Mpemba effect," where a colder system heated faster than a warmer one under certain conditions.

Since 2020, quantum researchers have studied Mpemba effects in quantum systems to show how initial conditions affect thermal changes. In 2024, a team at Trinity College described a quantum analysis of a system where "an initially hot system cools faster than an initially cooler one" when placed in a cold environment. They used computational models of spin systems and found that certain initial conditions can increase both thermalization speed and free energy.

In 2025, Zhang and others observed a quantum strong Mpemba effect in a single trapped ion. Chatterjee and others found the Mpemba effect naturally occurs during the cooling of nuclear spin states.

Theoretical explanations

The Mpemba effect is a phenomenon that scientists study, and its definition can vary in different research. Scientists have proposed several reasons for why it happens.

In 2017, two research teams discovered the Mpemba effect independently and at the same time. They also predicted a new type of effect called the "inverse Mpemba effect," where heating a cooled system that is far from equilibrium takes less time than heating a system that is closer to equilibrium. Zhiyue Lu and Oren Raz developed a general rule using a method called Markovian statistical mechanics. They predicted that the inverse Mpemba effect would appear in the Ising model and in diffusion processes. Antonio Lasanta and other researchers also predicted both the direct and inverse Mpemba effects in a system of granular gas starting from a far-from-equilibrium state. Lasanta’s study suggested that a common cause for both effects is a particle velocity distribution that differs greatly from the Maxwell–Boltzmann distribution.

In 2024, building on Kumar’s research, Isha Malhotra simulated single colloids placed in a double-well potential. She predicted that specific ranges of higher starting temperatures can lead to faster freezing, but not temperatures between those ranges.

Supercooling is a key part of many explanations for the Mpemba effect, especially the tendency of hot liquids cooled quickly to freeze at higher supercooling temperatures. Several molecular dynamics simulations have shown that changes in hydrogen bonding during supercooling influence the process. In 2017, Yunwen Tao and others suggested that the variety and unique behavior of hydrogen bonds might explain the effect. They noted that the number of strong hydrogen bonds increases with temperature, and that small clusters of strongly bonded molecules help form hexagonal ice when warm water cools rapidly. The researchers used vibrational spectroscopy and modeling with density functional theory-optimized water clusters.

Other explanations include the idea that freezing speeds vary greatly, and that the Mpemba effect may result from random, unpredictable factors, as stated by Andrei A. Klimov and Alexei V. Finkelstein.