Why Hot Water Sometimes Freezes Faster Than Cold Water

It sounds like a riddle with a wrong answer: two identical ice cube trays, one filled with hot water and one with cold, go into the freezer — and the hot one freezes first. Impossible, right? Hot water has further to go. Yet under the right conditions, this genuinely happens, and it has puzzled observers since Aristotle wrote about it more than two thousand years ago. In 1963, a Tanzanian student named Erasto Mpemba noticed his hot ice-cream mix froze before his classmates’ cold mix — and refused to let his teacher laugh it off. The phenomenon now bears his name. Here is why hot water sometimes freezes faster than cold water, and what the Mpemba effect really is.

What the Mpemba Effect Actually Claims

Precision matters, because the Mpemba effect is widely misunderstood. It does not claim hot water always freezes faster — that would violate common sense and thermodynamics. It claims that under certain conditions, initially hotter water can reach the frozen state before initially cooler water, despite starting further from the freezing point. The key phrase is “under certain conditions”: the containers, volumes, freezer environment, and measurement definitions all matter enormously. Scientists still debate the exact mechanisms, and some researchers argue the effect is finicky or even an artifact of inconsistent experimental setups. But careful experiments have reproduced it repeatedly — most convincingly with hot water around 90°C versus room-temperature water around 20°C in controlled freezers. It is real, it is conditional, and the “sometimes” in the question is doing heavy lifting. Understanding the candidate explanations is where the fun begins.

Theory 1: Evaporation Steals Mass and Heat

The oldest explanation is also the most intuitive. Hot water evaporates faster — significantly faster — than cold water. In an open container, the hot sample loses measurable mass as vapor, and evaporating water carries away enormous heat energy (evaporation is how sweating cools you). Less water left to freeze, plus heat already shed to the air, means less work for the freezer. Critics note that measured mass loss is often too small to fully account for observed time differences — evaporation contributes, but it may not be the whole story. Still, it is almost certainly part of the story, and it is the easiest mechanism to demonstrate at home: weigh the containers before and after, and the hot one’s weight loss is undeniable.

Theory 2: Convection Keeps the Hot Water Stirring

Here is the subtler mechanism. Hot water does not cool uniformly — it develops strong convection currents, with warm water rising and cooler water sinking in a continuous circulation. This internal stirring keeps the hot sample’s temperature more uniform and maintains efficient heat transfer to the container walls and the cold air. Cold water, by contrast, cools into stillness: the coldest, densest water settles, circulation weakens, and heat must escape slowly by conduction through stagnant layers. Think of it as the difference between a stirred pot and an unstirred one — the stirred pot always cools (or heats) faster. As the hot sample cools through the temperature range, its established circulation persists, giving it a heat-transfer advantage all the way down. Computer simulations of freezing water support this: the thermal history of the sample — how it got to each temperature — affects how fast it continues cooling. The water “remembers” being hot in the form of momentum.

Theory 3: Supercooling Gives Cold Water a False Start

This theory flips the perspective: maybe the mystery is not why hot water wins, but why cold water loses. Water does not always freeze at 0°C — very clean, still water can supercool several degrees below freezing without crystallizing, remaining liquid in a metastable state until something triggers ice formation. Initially cold water, cooling gently and undisturbed, is prone to supercooling: it sits below freezing, technically “colder,” without actually becoming ice. Hot water, with its vigorous convection and dissolved-gas changes, tends to nucleate ice promptly at the freezing point instead of supercooling. So the cold sample may reach a lower temperature first while remaining liquid, and the hot sample — freezing crisply at 0°C — wins the race to actual ice. The distinction between “coldest” and “frozen first” dissolves much of the paradox. In experiments, the cold sample often needs a nudge — a tap or a speck of dust — to finally crystallize, long after the hot sample has frozen solid.

Theory 4: Dissolved Gases and Frost Insulation

Two more contributors round out the picture. Heating water drives out dissolved gases (air, carbon dioxide), and degassed water freezes differently — some experiments suggest it nucleates ice more readily, though this effect’s size is debated. More practically, in frosty freezers, a hot container melts the insulating frost layer beneath it, settling into direct thermal contact with the cold shelf — dramatically improving heat conduction. The cold container sits atop its frosty insulation like a sled, cooling slowly. This is why the Mpemba effect shows up most reliably in frost-covered freezers and barely at all on bare wire racks: the “effect” is partly about the freezer, not just the water. Real-world conditions — frosty freezer, open containers, large temperature gap — stack multiple mechanisms in the hot sample’s favor simultaneously. Scientific American’s coverage of the Mpemba effect reviews how these theories have fared in recent controlled experiments.

Try It Yourself: The Kitchen Experiment

You can test this at home with genuine scientific rigor. You need: two identical containers (same size, shape, material), a freezer, and a thermometer. Fill one with hot tap water (not boiling — around 130–150°F is ideal and safer) and one with room-temperature water, using exactly equal volumes. Place them side by side on the same freezer shelf, start a timer, and check every 15 minutes without leaving the door open long. Record when each shows its first ice crystals and when each freezes solid. Run it twice, swapping positions to control for freezer hot spots. You may or may not see the hot sample win — and that is itself the lesson: the effect is conditional, and discovering which conditions matter is real experimental science. The Exploratorium’s freezing experiments offer structured protocols for young experimenters.

The Mpemba effect is real but conditional: initially hot water can freeze first through a combination of faster evaporation (less mass, less heat), persistent convection currents (better heat transfer), reduced supercooling (freezes promptly at 0°C instead of lingering liquid below it), and melted frost giving better shelf contact. No single theory explains every case — the freezer, the containers, and the temperature gap all matter.

Conclusion

Why does hot water freeze faster than cold water sometimes? Because freezing is not just about starting temperature — evaporation, convection, supercooling, dissolved gases, and frost contact all tilt the race, and under the right conditions they tilt it toward the hot sample. Aristotle noticed it, a Tanzanian schoolboy proved the skeptics wrong, and physicists are still refining the explanation. Try it in your own freezer — and whatever happens, you will have done real science. For more physics puzzles explained, visit our science section, meet our team, or send us your paradoxes.

Frequently Asked Questions