clock menu more-arrow no yes

Filed under:


Question: Why does hot water freeze faster than cold water in an ice cube tray placed in the freezer?

Answer: We have spent, in aggregate minutes, roughly half our life waiting for ice cubes to freeze. Most of that time was racked up back in the days when the ice cube compartment got so frosted there'd be only a tiny aperture remaining, a kind of ice cave. The ice cube trays would be embedded in frost, and when you inadvertently reached in with wet fingers to extract a tray your flesh would instantly fuse to the metal.

What we went through for a cold beverage.

So we care about this hot-water-freezing-faster issue. It's an ancient scientific debate. In fact, Aristotle mentioned the phenomenon once, we're told. It's also called the Mpempa Effect, after a Tanzanian schoolboy in the 1960s who made ice cream and discovered that hot milk seemed to freeze more quickly.

For many years, scientists said it was impossible. They argued that if you put hot water and cold water in the same freezer, at some point the hot water would have to cool to the temperature the cold water had been initially. Therefore by definition it would take the hot water longer to freeze.

But in 1977, Jearl Walker, a professor of physics at Cleveland State University, wrote an article in Scientific American explaining how hot water might actually freeze faster. His theory was that hot water evaporates more quickly, and loses energy more rapidly, than cold water. After a few minutes there would be less water in the "hot" tray. A smaller quantity of water would freeze more quickly.

We called Walker and he told us he's not so sure anymore. He referred us to a German scientist who has studied the issue more recently. His name is David Auerbach, and he's a physicist at the Max Planck Institute of Fluid Mechanics in Gottingen, Germany (a place that definitely sounds like a House O' Experts).

We called Auerbach and he gave us the skinny. In an old-fashioned refrigerator, he said, a hot ice cube tray will briefly melt the frost and ice beneath it. The liquid between the tray and the freezer surface will then re-freeze, and form an excellent contact for conducting heat away from the tray. The heat is sucked right out the bottom.

The cold water tray, meanwhile, doesn't have as good a contact with the freezer surface.

Auerbach has also found that even if you neutralize variables such as surface contact, hot water will sometimes form an almost imperceptible layer of ice around the edge before there's any sign of ice crystals in the cold water container. Why? Because the cold water container will often "supercool" more than the hot water container. That means cold water can get colder than the normal freezing point without actually turning to ice.

Mercifully this entire debate has been rendered moot by the invention of the ice machine.

The Mailbag:

Daniel F. of Oakland, Calif., writes, "Why is it that local Richter magnitudes are never given when earthquakes strike? For example, I would guesstimate that the 1989 Loma Prieta 7.1 magnitude earthquake was between 5.5 and 6.0 here in Oakland."

Dear Dan: We have grim news. According to an article in Science News by Richard Monastersky, the entire Richter scale is basically a hoax. Scientists do not actually use the Richter scale anymore. They just say they do, because they know that ordinary citizens like us are too ignorant to understand the scales they really use.

The Richter scale, invented by Charles Richter six decades ago, theoretically had no limit, but in practical reality only went up to about 7.0. Richter's scale was calibrated for the kind of crust found in Southern California. It measured how far a seismometer's needle moved during an earthquake.

The scale actually used now was developed by a successor of Richter's at Caltech, named Hiroo Kan-a-mori. His logarithmic scale is similar to Richter's, but what it measures is "moment magnitude," basically the total amount of energy released in an earthquake.

The magnitude of a quake, therefore, describes the total event, not just how it is felt in any given spot. Think of it this way: A 100-watt light bulb is a 100-watt light bulb no matter how far away you are standing.

There is, however, another way to measure an earthquake: Intensity. Intensity is what you are thinking of when you talk about the local effect of a quake. It's a subjective assessment of how much damage has occurred in a quake. No fancy instruments are used. Rather, an expert takes a clipboard and just looks at the damaged area and assigns a value to the intensity.

The intensity expert uses the Modified Mercalli Intensity Scale of 1931. The intensity scale, which doesn't get as much publicity as the magnitude scale, runs from 1 to 12 (actually it uses Roman numerals, so it goes from I to XII).

For example, this is an intensity level of I: " . . . sometimes birds, animals, reported uneasy or disturbed . . . doors may swing very slowly."

In other words, the same effect you get from the presence of a ghost.

Intensity V: "Awakened many, or most. Frightened few - slight excitement, a few ran outdoors. Buildings trembled throughout. Broke dishes, glassware, to some extent. . . . Pendulum clocks stopped, started or ran fast, or slow . . ."

(Which reminds us, we need to reset the pendulum clock that's attached to our VCR.)