The Mpemba Effect

Is this a paradox at all? The Mpemba effect is not logically paradoxical — there is no formal contradiction in the premises. It is counterintuitive: it violates the expectation that less energy means faster cooling. But counterintuitive results are common in thermodynamics. The effect is "paradoxical" in the loose sense that it surprises people who apply naive arguments to it.

Why has it not been settled? The effect is exquisitely sensitive to conditions: the starting temperature, the shape of the container, the ambient humidity, the degree of agitation, whether the water has been boiled before. A mechanism that operates strongly in one setup may be irrelevant in another. This makes reproducibility genuinely difficult and means that experiments confirming the effect and experiments failing to find it may both be correct — they may simply be measuring different systems.

The historical record. Aristotle noted in the Meteorologica that "hot water freezes more quickly." Francis Bacon made the same observation. Neither pursued it as a scientific question. Mpemba's contribution was not the observation but the insistence, against dismissal from his teacher, that it was a real effect worth investigating — and the fortunate timing of Osborne's visit.

Aristotle stated in the Meteorologica (c. 340 BCE) that "the fact that water has previously been warmed contributes to its freezing quickly: for so it cools sooner." Francis Bacon included the same observation in his Novum Organum (1620). Both treated it as a matter of observation rather than a problem requiring explanation.

The effect was essentially forgotten until 1963, when Mpemba noticed it while making ice cream in Tanzania. His physics teacher dismissed it; the class mocked him. When Denis Osborne visited the school in 1966, Mpemba asked him directly. Osborne's willingness to take the question seriously — and his experimental confirmation — are credited with reviving scientific interest.

The 1969 paper generated a long tail of experimental and theoretical work. The question became something of a proving ground for proposed mechanisms of water behaviour. It remains unresolved.

It is a genuine physical effect without a settled explanation. Most well-documented effects in everyday physics have agreed mechanisms. The Mpemba effect, after more than sixty years of research and hundreds of papers, does not. It is a rare example of something that can be shown in a kitchen being scientifically contested at the mechanistic level.

It illustrates how experimental conditions interact with physical mechanisms. The sensitivity of the Mpemba effect to container geometry, water purity, and freezer type makes it a case study in the way that real-world systems can exhibit behaviour that simple models miss — because simple models hold all these variables constant, and reality does not.

It is a lesson in scientific credibility. Mpemba's observation was dismissed by his teacher. It was confirmed by a physicist who was willing to test rather than theorise. The episode is frequently cited in discussions of how scientific questions should be evaluated and by whom.

Industrial cooling. Many industrial processes involve cooling hot liquids rapidly. If the Mpemba effect operates reliably under specific conditions, it could have engineering implications for the design of cooling systems.

Cryopreservation. Biological samples (cells, embryos, organs) are preserved by cooling. The rate and dynamics of cooling affect ice crystal formation and hence cell survival. Understanding precisely how temperature history affects freezing dynamics — the Mpemba question, in essence — is directly relevant.

Climate science. Ocean temperatures and their evolution over time affect ice formation at the poles. The Mpemba effect is a reminder that simple thermal models of cooling can miss effects that depend on the prior history of the water.

When the effect occurs

The Mpemba effect requires a mechanism that gives the hotter sample an advantage that more than compensates for its higher starting temperature. This typically requires at least one of:

1. Mass loss through evaporation: significant if the container is open and the temperature difference is large.

2. Differential supercooling: the hot sample nucleates ice at 0°C; the cold sample supercools past 0°C, delaying freezing. Supercooling can delay freezing by minutes in pure water.

3. Altered thermal conductivity: through dissolved gas concentration or temperature-dependent water structure.

The effect is least likely in conditions of:

• Sealed containers (eliminating evaporation)
• Very pure water (reducing nucleation from dissolved gases)
• Uniform freezer temperature (eliminating convection effects)
• Small temperature difference between samples

This explains why laboratory attempts to reproduce it under "ideal" conditions often fail.

The Markovian cooling model

In a simple Newton's-law model, let $T(t)$ be the temperature at time $t$ and $T_F$ the freezer temperature. Then:

$$\frac{dT}{dt} = -k(T - T_F)$$

This gives exponential cooling:

$$T(t) = T_F + (T_0 - T_F)e^{-kt}$$

For two samples with starting temperatures $T_h > T_c > T_F$, both reach any given target temperature $T^*$ at times:

$$t_h = \frac{1}{k}\ln\frac{T_h - T_F}{T^* - T_F}, \quad t_c = \frac{1}{k}\ln\frac{T_c - T_F}{T^* - T_F}$$

Since $T_h > T_c$, we get $t_h > t_c$: the hot sample always takes longer. The Mpemba effect cannot occur in this model. It requires a mechanism that breaks the simple Markovian assumption — some property of the hot sample that changes $k$, reduces the effective mass, or delays freezing in the cold sample.