An imposing knock at the door, both impetuous and imperious.  I figured it for an Internet denizen.  “C’mon in, the door’s open.”

“You’re Moire?”

“I am.  And you are..?”

“The name’s Feder, Richard Feder, from Fort Lee, NJ.  I’m a stand-in for some of your commenters.”

“Ah, the post of business past.  You have a question?”

“Yeah.  How come hot water can freeze faster than cold water?”

“That’s really two questions. The first is, ‘Can hot water freeze faster than cold water?’ and the second is, ‘How come?‘  To the surprise of a lot of physicists, the experimental answer to the first question is, ‘Yes, sometimes.‘  But it’s only sometimes and even that depends on how you define freeze.”

“What’s to define?  Frozen is frozen.”

“Not so fast.  Are we talking surface ice formation, or complete solidification, or maybe just descent to freezing temperature?  Three very different processes.  There’s multiple reports of anomalous behavior for each one, but many of the reports have been contested by other researchers.  Lots of explanations, too.  The situation reminds me of Anne’s Elephant.”

“Why an elephant?  And who’s Anne?”

“Remember the old story about the blind men trying to figure out an elephant?  The guy touching its trunk said it’s a snake, the one at its side said it’s a wall, the dude at its leg said it’s a tree, and so on?  The descriptions differed because each observer had limited knowledge of something complicated.  This chilled-water issue is like that — irreproducible experiments because of uncontrolled unknown variables, mostly maybes on the theory side because we’re still far from a fundamental understanding.”

“Who’s Anne?”

“Anne is … an experience.  I showed her how the notion of Entropy depends on how you look at it.  Scientists have looked at this paradoxical cooling effect pretty much every way you can think of, trying to rule out various hypotheses.  Different teams have both found and not found the anomaly working with distilled water and with tap water, large amounts and small, in the open air and in sealed containers, in glass or metal containers, with and without stirring, with various pre-washing regimens or none, using a variety of initial and final temperatures.  They’ve clocked the first appearance of surface ice and complete opacity of the bulk.  They’ve tracked temperature’s trajectory in the middle of the container or near its wall… you name it.  My favorite observation was the 20th Century’s first-published one — in 1963 Erasto Mpemba noticed the effect while preparing ice cream.”

“What flavor?  Never mind.  Is there a verdict?”

“Vaguely.  Once you get approximately the right conditions, whether or not you see the effect seems to be a matter of chance.  The more sophisticated researchers have done trials in the hundreds and then reported percentages, rather than just ‘we see it’ or not.  Which in itself is interesting.”

“How’s that?”

“Well, to begin with, the percents aren’t zero.  That answers your first question — warm water sometimes does freeze faster than cold.  Better yet, the variability tells us that the answer to your second question is at the nanoscopic level.  Macroscopic processes, even chemical ones, have statistics that go the same way all the time.  Put a lit match to gasoline in air, you’ll always get a fire.  But if you set out 100 teaspoons of water under certain conditions and 37 of them freeze and the others don’t, something very unusual must be going on that starts with just a few molecules out of the 10²³ in those teaspoons.”

“Weird odds.”

“This experiment’s even more interesting.  You’ve got two bottles of water.  You heat up bottle A and let it cool to room temperature.  B‘s been at room temperature all along.  You put ’em both in the fridge and track their temperatures.  A cools quicker.”

“That’s where I came in.”

“Both start at the same temperature, finish at the same temperature, and their Joules-per-second energy-shedding rates should be the same.  A cools in less time so A releases less heat.  Entropy change is released heat energy divided by temperature.  Somehow, bottle A went into the fridge with less entropy than B had.  Why?  We don’t really know.”

~~ Rich Olcott

• – Thanks to Ilias Tirovolas, whose paper inspired this post.