A short walk through the strange five hours that pass between a cup of water and a cup of ice — including the case of a Tanzanian schoolboy who noticed something nobody believed.
Pour a liter of water into one big cup, or split it into two smaller ones — same volume, same freezer. The two cups will freeze noticeably faster. The reason is geometry.
When you split a volume V into n equal pieces of similar shape, the total surface area scales as n^(1/3). For two cups, that's about 1.26×, or +26%.
More surface means more channels for heat to escape. It also means the warmest molecules — the ones in the center — sit closer to the cold edge. Both effects shorten the time to ice.
Imagine 500g of water at room temperature (20°C) placed in a freezer at −18°C. We need to figure out how much energy must leave the water before it becomes solid — and roughly how long that takes.
First we cool the water from room temperature down to the freezing point. That uses the standard heat-capacity equation: Q = m·c·ΔT.
Then comes the harder part — the phase change at 0°C. Water doesn't get colder during this step; instead, energy goes into rearranging molecules into a crystal. That's the latent heat of fusion, 334 J per gram.
Out of the total 209 kJ, 80% goes into the phase change, not the cooling. Freezing isn't slow because water is hard to cool — it's slow because crystallization is energetically expensive.
Most of the wait isn't the cooling. It's the slow choreography of molecules locking into a crystal — at exactly 0°C, while nothing seems to be happening.
To turn energy into time, we use Newton's law of cooling: P = h · A · ΔT. Power equals a heat-transfer coefficient times surface area times the temperature gap between water and air.
For a 500ml cup with natural convection (h ≈ 15 W/m²·K) and a freezer at −18°C, the cooling phase takes about 40 minutes. The freezing phase — same surface, smaller gap — takes another four-plus hours.
Total: roughly five hours. Real freezers do better, around 2–4 hours, because they have forced convection and better thermal contact through metal shelves.
This is a back-of-the-envelope calculation. Real freezers have forced convection; cup material, shape, and air circulation matter a lot. Treat the number as a feel for the physics, not a stopwatch.
If you actually plotted the water's temperature over time, you'd see something strange: a long, flat plateau right at 0°C. The water stops cooling, even though it's still losing heat.
That heat is going into rearranging molecules into crystals — not into temperature. The thermometer reads zero for hours while the freezer is doing the bulk of its work, invisible.
This plateau is why a thermometer alone can't tell you how much ice has formed. The reading is the same at 5% frozen and 95% frozen.
A common confusion: "If water stays at 0°C while freezing, why do we use ΔT = 18°C in the heat-transfer formula?" The answer is that there are two different ΔTs in this problem, doing two different jobs.
Used in Q = m·c·ΔT. During freezing this is zero — water sits at 0°C the whole time. That's why we switch to latent heat for that phase.
Used in P = h·A·ΔT. Sets the speed of heat flow. Water at 0°C facing air at −18°C still gives an 18°C gap. Without that gap, no heat moves, and ice never forms.
The water can stay at one temperature while heat keeps streaming out — because the gap with the freezer air is what drives the flow, not the change in the water itself.
A microwave has a fixed wattage; load it with more food and each item gets a smaller share of energy. A freezer is similar — except instead of pumping energy in, it pumps energy out.
Most home freezers run on compressors rated around 100–250W. That's the upper limit on how fast heat can leave, no matter how many cups you stuff inside.
One or two cups? Barely matters. Five warm cups all at once? You'll feel it.
A thirteen-year-old boy in Tanzania was making ice cream in cooking class. The recipe said: boil milk and sugar, let it cool, then put it in the freezer. But on this day, the freezer was filling up fast. Mpemba skipped the cooling step and shoved his still-hot mixture in alongside everyone else's already-cooled ones.
Ninety minutes later, his was frozen. The cooled ones weren't.
He told his physics teacher. The teacher was unimpressed: "That's Mpemba's physics, not real physics." The classroom laughed. The boy didn't forget.
Years later, a visiting professor named Denis Osborne came to Mpemba's school. Mpemba asked him the question that had nagged him for years. Osborne, intrigued, ran the experiment back at his lab. The effect was real — under certain conditions, hot water did seem to freeze faster than cold. They published together in 1969.
The strangest part? Aristotle had noticed it. So had Francis Bacon and René Descartes. The phenomenon had been written about for two thousand years and quietly ignored.
"I have come, sir, to ask you a question about physics — but the boys are laughing at me."— MPEMBA TO DR. DENIS OSBORNE, 1969
Nobody knows for sure. Several mechanisms have been proposed, and they're probably all true in different situations.
So is it real? Sometimes, probably, under conditions that haven't been pinned down. It's the rare physics question where the right answer might still be: "It depends, and we're not sure on what."
The Mpemba effect is fascinating, but it's not a reliable trick. If you want fast ice, use the geometry tricks below — they work every time.
Pure water, cooled gently and undisturbed, doesn't freeze at 0°C. It can stay liquid down to roughly −38 to −42°C — the limit of homogeneous nucleation. Any small disturbance and the whole cup turns to slush in under a second. Search "supercooled water" on YouTube; it looks like a magic trick.
Sometimes an ice cube has a thin spike rising from its top, like a tiny obelisk. Water expands by about 9% when it freezes. If the surface freezes first and the inside expands, liquid water gets squeezed up through a small hole, freezing as it goes and building a tube.
Tap water has dissolved gases. As ice forms inward from the edges, those gases get pushed toward the center, ending up as tiny bubbles. That's why home ice is cloudy. Bartenders use directional freezing: insulate sides and bottom, freeze only from the top. The top half ends up glass-clear.