Skip to main content

Which freezes faster - cold water or boiling water? Hailey, Sprighill, Florida

Usually cold water freezes before hot, as Newton theorized in 1701. But, hot water - even boiling water - can freeze faster than cold.

You Tube shows a video of university student Jim Sage on top of Mount Washington in heavy winter clothes boiling water on a small alcohol stove in the -34.8 F (-37.1 C) cold. Jim lifts the pot, and tosses boiling-hot water up then watches it snow down, like a blizzard, all around him.

Story continues below advertisement

<object width="600" height="480"><param name="movie" value=""></param><param name="allowFullScreen" value="true"></param><param name="allowscriptaccess" value="always"></param><embed src="" type="application/x-shockwave-flash" allowscriptaccess="always" allowfullscreen="true" width="600" height="480"></embed></object>

That's fast freezing, but faster than cold water?

It can happen, as Erasto B. Mpemba discovered in 1967 when he was a form-3 student (equivalent to North American 11th grade) making ice cream at Magamba Secondary School in Lushoto, a small town in Tanzania, Africa.

Erasto asked his science teacher to explain why his hot milk froze faster than a friend's cold milk. The teacher said, "You were confused; that cannot happen."

Erasto must have had some reservations about this reply because, over the next couple of years, he asked the same question of two more physics teachers, finally finding one with an open mind.

Erasto's day of discovery started like most school days. He woke up before dawn. Pulling on his school-uniform navy-blue shorts, he shivered in the cold. But the afternoon would be hot, this close to the equator, even though he was almost a mile high in the mountains. "I've time to make ice cream," he thought, and set out for milk.

He strode briskly along the dirt main street of Lushoto to the open-air market. Quickly he bought some milk, and headed back to school. He mixed sugar with milk, and put the pot on the stove to boil. It had begun to boil, when he noticed another boy not bothering to heat his mixture in order to seize space in the refrigerator's freezing compartment. Erasto grabbed the last available space, putting his boiling-hot liquid beside the cool one.

"The other boy and I went back in an hour and a half later, and found that my tray of milk had frozen into ice-cream while his was only a thick liquid, not yet frozen," Erasto wrote later. Erasto asked his physics professor why, and got only 'that cannot happen' answer.

Story continues below advertisement

Two years later, in Mkwawa High School, Erasto again asked why the hot liquid froze faster than the cold. His new teacher answered: You were confused.

But help was coming. The school's headmaster invited physicist Dennis Osborne of the University College in Dar es Salaam to lecture and answer questions. Erasto asked his question yet again.

Dr. Osborne smiled into Erasto's earnest face. "Is it true, have you done it?"


Everyday events are seldom as simple as they seem, mused Osborne to himself. "I do not know, but I promise to try this experiment."

And he did, eventually getting similar results. Osborne found cold water near 20 degrees C (70 F) began to freeze in about 100 minutes; whereas hot water about 80 degrees C (175 F) began to freeze much quicker - in 40 minutes.

Story continues below advertisement

Water, a commonplace substance is anything but simple. Certainly, as Mpemba and Osborne discovered, its freezing behaviour is not characterized by its average temperature.

No doubt, several mechanisms underlie the Mpemba effect. We have eliminated some.

For example, a vessel containing initially hot water would melt frost on the floor of a non-frost-free refrigerator. Then, the hot-water vessel would be in direct contact with the cooling element of the freezer, and therefore freeze faster than the vessel with the initially cold water, which would rest on frost. The frost would insulate the cold-water vessel.

But, this mechanism doesn't explain the entire effect, because, in one experiment, Osborne placed both vessels on good thermal insulators. The hot water still froze faster than the cold.

Osborn thought perhaps evaporation could account for the hot water freezing faster, since the more active hot-water molecules would escape the liquid, cooling the remaining liquid. But he found only small changes in volume due to evaporation, and concluded evaporation could account for no more than 30 per cent of the cooling.

Three mechanisms remain that could explain the effect.

Hot-top. Cooling occurs mainly from the top surface. But hot water freezes differently than cold - more heat escapes from the top for hot than cold water.

Cold water starts to freeze along its top surface, and along its sides. The top ice insulates the water from the colder air above, slowing freezing. The insulated upper water does not chill appreciably more and, therefore, does not become denser and drop to the bottom. Few convection currents form, so little convective heat transfer occurs.

Hot water also freezes along the sides and bottom like the cold water, but the top hot surface remains free of ice. Heat radiates from the liquid top at a much higher rate than from the top ice surface of the cold water. Moreover, many more hot water molecules zip out of the top hot water than from the cold ice. So, the temperature drops faster in the initially hot water.

Furthermore, in the hot vessel, as the hot surface water chills, contracts and becomes denser, it drops to the bottom of the container and forces warmer bottom water to the surface, further driving the heat transfer. These effects can combine so hot water freezes faster than cold.

Supercooling. Water has the property of remaining liquid below its freezing-point temperature - it can become supercooled. Both the initially hot and the initially cold water can supercool. But, if the cold water stays liquid to a lower temperature, then the supercooled hot water may, indeed, freeze first. Furthermore, supercooling is somewhat capricious, since it depends upon tiny bubbles to kick off ice crystal formation. So, which freezes faster can depend on how many bubbles are present in the respective containers.

Snap freezing. Hot water out of a hot water heater may have fewer of the tiny bubbles needed to form ice crystals. So, the hot water could supercool and suddenly snap freeze, thus beating the cold water to a frozen state.

But, as physicist Monwhea Jeng of Syracuse University writes, "There is no well-agreed explanation for how this phenomenon occurs." Mpemba's question remains unanswered, so far.

You, too, can check this phenomenon out in your very own kitchen.

If you do, please e-mail me your results. I'll post them with this article on the WonderQuest site.

Earlier, I sent the article to physicist Erik Ramberg of Fermilab for fact checks. He e-mailed me back: "I had heard these stories before about hot water freezing faster, but always dismissed them. But the essence of science is experimentation and observation, so I did a few simple experiments.

"I took two small containers of tap water (about 150 ml) and covered them with lids. I heated one in the microwave for 1 minute, yielding about 170 degrees F (77 C) water. The other was room temperature - about 70 degrees F (21 C). I fully expected the cool water to freeze much faster. To my surprise, although the cool water developed thicker ice on the top and side surfaces after an hour, the hotter water caught up after about 3 hours. I saw no significant difference between the two after 4 hours in the freezer."

I repeated the experiment twice more - once with only 60 ml of water. The conclusion was the same: the cooler water developed thicker ice to begin with, but for all practical purposes the hot and cool water freeze solid after the same amount of time. A wonderful surprise!"

I suspect the hot water radiates its heat quickly and that the ice developing on the surfaces insulates the water inside, causing the two to develop ice at a very similar rate.

"Your readers should be encouraged to become scientists and test this phenomenon themselves."

Erasto Mpemba is now Director of Game in Tanzania's Ministry of Natural Resources.

The description of Erasto's discovery of the 'Mpemba' effect while making ice cream is factual, based on Mpemba's story in the journal Physics Education.

Further Reading:

  • The Mpemba Effect (aka Cool?) by E.B. Mpemba and D.G. Osborne, published first in 1969 and reprinted in 1979: Phys. Educ. 1969 4 172-5, Phys. Educ. 1979 14).
  • Can hot water freeze faster than cold water? by Monwhea Jeng, University of California
  • Hot water freezing by Rod Nave, HyperPhysics, Georgia State University
  • Supercooling and the Mpemba effect , The Smithsonian/NASA Astrophysics Data System
  • Knight, Charles A., The Mpemba Effect: The Freezing Times of Hot and Cold Water, Letter in Am J Phys, Vol 64, May 1996, p524
  • Auerbach, David, Supercooling and the Mpemba Effect: When Hot Water Freezes Faster Than Cold, Am J Phys. 63, 882-885, (1995)
  • J. Walker, "The Amateur Scientist", Scientific American, vol. 2337, #3, pgs 246-7, Sept. 1971)

(Answered June 8, 2008)

April Holladay lives in Albuquerque, New Mexico. Her column, WonderQuest, appears every second Monday of the month on To read April's past columns, please visit her website . If you have a question for April, visit this information page .

You can also comment on this article at Google Groups .

Report an error
As of December 20, 2017, we have temporarily removed commenting from our articles. We hope to have this resolved by the end of January 2018. Thank you for your patience. If you are looking to give feedback on our new site, please send it along to If you want to write a letter to the editor, please forward to