Join 3,498 readers in helping fund MetaFilter (Hide)


From 100 to 0
June 27, 2012 9:19 AM   Subscribe

The Royal Society of Chemistry is offering £1000 to the person or team producing the best and most creative explanation of the phenomenon, known today as The Mpemba Effect

More on Mpemba, paper in full here.


(Previously)
posted by Kiwi (95 comments total) 17 users marked this as a favorite

 
Aren't there multiple possible mechanisms, such as 1) faster evaporation and mass reduction of warmer water, or 2) thawing of surface frost on the freezing surface to give better heat transfer, and so on, any of which might dominate under certain conditions?
posted by Mapes at 9:29 AM on June 27, 2012 [1 favorite]


Water is weird.
posted by tommasz at 9:30 AM on June 27, 2012 [8 favorites]


Best, or most creative... I'm not sure you can have both.

Anyway, it's not scientific, but I've always expected that it was a result of the molecules themselves transporting external temperatures. Hot molecules move faster, thus move the cold temperatures to the middle faster. Of course, this theory no longer works once the water becomes not-hot because at that point the molecules are no longer moving fast.

The counter-to-the-counter-point to that would be something along the lines of molecules changing speed at a slower rate than the temperature change, which I'm sure would also violate some law somewhere, and I can't really be bothered to look it up.

If somebody wins a grand off that theory, buy me lunch and I'll consider us even.
posted by Blue_Villain at 9:35 AM on June 27, 2012


This thread is going to be very frustrating for those that clicked the links.
posted by DU at 9:35 AM on June 27, 2012 [16 favorites]


If the effect was observed with ice cream then it can't be (primarily) convection within the material to be frozen. Presuming the effect is real, I suspect that the hot material may set up persistent convective air currents within the freezer itself, and that these air currents account for the faster cooling.
posted by Joe in Australia at 9:41 AM on June 27, 2012 [1 favorite]


How about this effect: Popular media announces something (e.g., an award for a solution) related to a problem that has stumped experts in the field, summarizing the problem for a lay audience. Without fail, someone who misunderstands or skims over the problem will think it's a simple problem and claim victory.
posted by supercres at 9:43 AM on June 27, 2012 [2 favorites]


Er, think it's an easy problem with a simple solution and claim victory.
posted by supercres at 9:43 AM on June 27, 2012


That's really a very neat phenomena. Does anyone know how to pronounce Mpemba?
posted by annsunny at 9:45 AM on June 27, 2012 [1 favorite]


Here's the video version.
posted by Obscure Reference at 9:45 AM on June 27, 2012


Haven't ice rink folks via the zamboni/resurfacer know this for years ? (Yes, there's also the "hot water helps even our chips/grooves/etc in the ice" side-effect, but that it froze faster was the main benefit)
posted by k5.user at 9:47 AM on June 27, 2012 [1 favorite]


I like Joe's idea. It wasn't mentioned on the wiki page at least, and the convection or air could be carrying the coldest draft straight to the hotter container, and the current could persist for a while even after the water cooled. To test it and check some other things like evaporation, try it with identical sealed containers (seal in identical water before heating the one.) Test for freezing either with a sealed in thermometer or by watching for the correct amount of expansion. If you change the material of the identical containers to be more insulating, the slower heat loss would mean less convection.
posted by TreeRooster at 9:48 AM on June 27, 2012


It's like when your wife is 8 months pregnant and 1000 degrees and 3 minutes later she's cranking the thermostat that you always have set at 19 because JESUS FUCK DON'T ANGER IT.
posted by jimmythefish at 9:48 AM on June 27, 2012 [1 favorite]


Good stuff in the third link:
Regardless of the mystery, Mpemba’s discovery is a wonderful illustration of many important lessons in science: the significance of experiment over theory, the danger of clinging to preconceived notions, the difficulty in evaluating even seemingly simple real-world physics problems, and the importance of perseverance in the face of unreasoning denial.
posted by exogenous at 9:48 AM on June 27, 2012 [2 favorites]


Does anyone know how to pronounce Mpemba?

I think it's I MmmPemba. Like "Mmm... cookies!"
posted by Wolfdog at 9:48 AM on June 27, 2012 [4 favorites]


Haven't ice rink folks via the zamboni/resurfacer know this for years ?

The effect is named after Tanzanian Erasto Mpemba. He first encountered the phenomenon in 1963...
posted by DU at 9:49 AM on June 27, 2012 [1 favorite]


I also like the "hot material melts a thin layer of ice on which the container is sitting, which makes a better thermal bond between the container and the freezer" theory.
posted by Joe in Australia at 9:56 AM on June 27, 2012


Here is young Erasto Mpemba in his own words, and it is a story quite well told. (May be behind paywall if you are not at a university.)
posted by TreeRooster at 9:57 AM on June 27, 2012 [1 favorite]


Maybe the hot water molecules are quicker to line up all laticey-like than the colder, slower water molecules. Once they have a head start, the freezing happens quicker.
posted by DaddyNewt at 9:58 AM on June 27, 2012 [4 favorites]


When there is a bigger the difference in temperature between two objects, the energy in the form of heat flows between them faster, right? So starting from a higher difference in temperature, there is higher potential energy. So maybe energy has momentum. That would be pretty weird.
posted by 2bucksplus at 9:59 AM on June 27, 2012 [2 favorites]


He first encountered the phenomenon in 1963...
And Zambonis are from 1949.
posted by k5.user at 9:59 AM on June 27, 2012 [2 favorites]


most creative explanation

Hot water has a higher concentration of fearons than cold water, which in turn contains higher concentrations of DsBLeeF. Hence, when water is introduced to freezing temperatures, the solution is clear: the cold water suspends DsBLeeF and the hot water freezes in fear. Also known as the O+M effect.
posted by It's Raining Florence Henderson at 10:01 AM on June 27, 2012 [1 favorite]


It's because the warmth makes the molecules stretch, and then the cold makes them snap back quickly like a rubber band. You're welcome.
posted by Burhanistan at 10:01 AM on June 27, 2012 [1 favorite]


Presuming the effect is real, I suspect that the hot material may set up persistent convective air currents within the freezer itself, and that these air currents account for the faster cooling.

The effect is real, and it doesn't just happen in freezers. The effect has long been known (surely much earlier than Mpemba "discovered" it) in places that are cold. Zambonis depend on it. Hell, there's even a scene in Home Alone that utilizes it.
posted by Sys Rq at 10:03 AM on June 27, 2012 [1 favorite]


Do other liquids behave like this... does room temperature gallium resolidify more slowly than liquid metal at twice its temperature?
posted by Slap*Happy at 10:05 AM on June 27, 2012 [1 favorite]


Haven't ice rink folks via the zamboni/resurfacer know this for years ?

I thought so, too. Still, it should be an easy £1000 for the first Zamboni driver to pick up the phone.
posted by Blazecock Pileon at 10:06 AM on June 27, 2012


Does anyone know how to pronounce Mpemba?

You could ask him yourself....
posted by Kiwi at 10:08 AM on June 27, 2012


The obvious answer is that science. Also molecules.
posted by shakespeherian at 10:13 AM on June 27, 2012 [4 favorites]


We need to get someone on a space station to try freezing zero-g blobs of water so we can rule out the effects of containers.
posted by Jpfed at 10:14 AM on June 27, 2012


The thing that is so counter-intuitive is that the warm water at some point will cool to the temperature of the cool water. So at that point, how is it indistinguishable from the cool water when it was put in the freezer? Meanwhile the cool water must be cooler than when it was put in the freezer. Then the cooler warm water must cool to the cooler temperature of the cool water, and so on. How could the warmer water overtake, since every time it catches up to the cool, it must play by the same rules?
posted by weapons-grade pandemonium at 10:17 AM on June 27, 2012 [9 favorites]


Resurfacers have little in common w this effect. They're resurfacers, not icemakers
posted by MangyCarface at 10:18 AM on June 27, 2012


Resurfacers have little in common w this effect. They're resurfacers, not icemakers

They spread hot water on an ice rink. The hot water then freezes, rather than, you know, melting the rink.
posted by Sys Rq at 10:19 AM on June 27, 2012


Joe in Australia: If the effect was observed with ice cream then it can't be (primarily) convection within the material to be frozen.
Why not? Do you believe that unfrozen ice cream liquid doesn't have convection?
posted by IAmBroom at 10:23 AM on June 27, 2012


Yeah, but they don't rely on hot water freezing more quickly than cold water. They rely on hot water melting the ice beneath it, which then refreezes along with the newly-sprayed water, so as to create a thicker layer of fresh ice.
posted by Holy Zarquon's Singing Fish at 10:24 AM on June 27, 2012 [1 favorite]


Sys Rq: Resurfacers have little in common w this effect. They're resurfacers, not icemakers

They spread hot water on an ice rink. The hot water then freezes, rather than, you know, melting the rink.
That's a matter of the ratio of added thermal energy (a thin spray of hot water) to the thermal mass of the rink (a relatively thick layer of ice). Still has nothing to do with this problem.
posted by IAmBroom at 10:26 AM on June 27, 2012


Slap*Happy: Do other liquids behave like this... does room temperature gallium resolidify more slowly than liquid metal at twice its temperature?
Now THAT is a good question. If true, it would wipe out all the polar-liquid hypotheses in a single go.
posted by IAmBroom at 10:27 AM on June 27, 2012


How could the warmer water overtake...?

I think the answer is that while it was cooling down, the hotter water was evaporating faster than the cooler water, losing mass. When it finally overtook the cooler water, it had less mass to cool and freeze.
posted by exphysicist345 at 10:28 AM on June 27, 2012


£1000 is a pretty pathetic prize to offer in return for explaining something mysterious about the most common chemical compound in the universe. Put up £100000 or a million, you cheap bastards. Make it worth someone's while.
posted by anigbrowl at 10:30 AM on June 27, 2012 [2 favorites]


That's a matter of the ratio of added thermal energy (a thin spray of hot water) to the thermal mass of the rink (a relatively thick layer of ice). Still has nothing to do with this problem.

Right, but the ratio is more even with hot water than cold water, and yet it freezes faster.
posted by Sys Rq at 10:34 AM on June 27, 2012


It's because cold is actually a malevolent, conscious force that actively despises heat. So whenever cold sees heat getting too uppity (as when heat performs one of its more useful and impressive tricks like boiling water), cold overreacts and the knives really come out.
posted by saulgoodman at 10:35 AM on June 27, 2012


Meanwhile the cool water must be cooler than when it was put in the freezer. Then the cooler warm water must cool to the cooler temperature of the cool water, and so on. How could the warmer water overtake, since every time it catches up to the cool, it must play by the same rules?

Don't get in a footrace with a tortoise, is my advice to you.
posted by Wolfdog at 10:40 AM on June 27, 2012 [6 favorites]


Doesnt it have something to do with the difference thermal diffusion rates for ice, water, and water vapor? Btw, I had heard this before, but didn't realize that it had been demonstrated experimentally. So, I should build my water for ice cube making? Do commercial ice machines make use of this?
posted by borges at 10:42 AM on June 27, 2012


Does this effect still hold when the water is stirred to keep the temperature homogeneous throughout the entire container? If so, then something strange is definitely going on because hot water would have to become cold water on its way to freezing.
If not, then it must have something to do with the temperature distribution between the top, middle, and bottom layers (not really layers, since it would be continuous).
posted by rocket88 at 10:42 AM on June 27, 2012 [1 favorite]


Water is weird...this page provides a possible explanation for the effect : http://www.lsbu.ac.uk/water/explan.html
posted by borges at 10:50 AM on June 27, 2012


Also, degassing! As mentioned in the third link, I think.

Don't all the fancy bartending websites tell you to boil your water before making ice cubes to get the best ice cubes? They should ask fancy bartenders about this.
posted by newg at 10:54 AM on June 27, 2012


I haven't seen this mentioned before, but does the difference in freezing rates increase when a colder freezer is used?
posted by sensate at 10:55 AM on June 27, 2012


The full paper is very readable and addresses a lot of the initial things that occur upon thinking about this problem.
posted by lazaruslong at 10:57 AM on June 27, 2012


k5.user: And Zambonis are from 1949.

And Aristotle wrote about it around 350BC. The post isn't about who knew about it first, it's who can best explain it.
posted by thebestsophist at 11:17 AM on June 27, 2012


Is this like (people who believe that) wet sponges soak up more water than dry ones?

Because they don't (when all other variables are held constant).
posted by achrise at 11:24 AM on June 27, 2012


you can supposedly boil your water to get clearer ice cubes, as it (supposedly) drives out the dissolved gasses, but I have not found this to be true in practice.

you should not use hot water from the tap, unless you like drinking dissolved contaminants from your boiler and pipes.
posted by danny the boy at 11:34 AM on June 27, 2012 [1 favorite]


Dissolved contaminants from your boiler and pipes should be steeped for exactly 3 minutes, and not one second longer. Add one sugar cube, and a splash of cream. Stir 3 times. Perhaps a dash of arsenic.
posted by It's Raining Florence Henderson at 11:37 AM on June 27, 2012 [1 favorite]



I'll see if I can remember the explanation offered to me by one of my physics grad students working in the lab :

It has to do with the energy requirements for a phase change from liquid to solid. Warm water has more energy available, which makes it easier for molecules near the edge to make the phase change. This then accelerates other molecules cooling down. And so on.


As for the zamboni - they need the water laid down to freeze quickly. A layer of hot water refreezes faster than a layer of cold water. It has the added benefit of smoothing out ridges, too.
posted by Pogo_Fuzzybutt at 11:41 AM on June 27, 2012


Water can't actually freeze at all, since corollary to Zeno's Paradox it can never actually reach 32°F from any other temperature. Ice is a myth. I sent this answer in and they agreed, and are sending me the prize money, though they did warn me that it cannot logically ever arrive.
posted by George_Spiggott at 11:45 AM on June 27, 2012 [11 favorites]


I'm glad they were willing to meet you halfway.
posted by Holy Zarquon's Singing Fish at 11:49 AM on June 27, 2012 [3 favorites]


I think the answer is that while it was cooling down, the hotter water was evaporating faster than the cooler water, losing mass. When it finally overtook the cooler water, it had less mass to cool and freeze.

If that is the case, then warm water does not freeze faster than cool water. Small amounts of water freeze faster than large amounts of water, and we knew that already. You can't go off the baseline to avoid the tag.
posted by weapons-grade pandemonium at 11:54 AM on June 27, 2012


Zeno's Paradox

There's a reason why archery teachers tell their students to imagine the arrow going through the target, instead of only just reaching the target.
posted by explosion at 11:57 AM on June 27, 2012


(Stupid though that was, it's worth questioning whether a thing must traverse the intervening temperatures beween one measured temperature and another. On a macro level I suppose it does as a matter of overwhelming statistical likelihood, but I suspect not as a strict thermodynamic requirement.)
posted by George_Spiggott at 12:06 PM on June 27, 2012


What I want to see is identical little water balloons, in a freezer with well mixed air, in free fall. I wonder if £1000 is enough for Branson to include this on a flight of his next spaceship?
posted by TreeRooster at 12:19 PM on June 27, 2012


My wife just boiled some water for tea as I discovered this thread, so I am conducting an ad hoc experiment. Two ice cube trays filled with the same amount of water side by side in the freezer, one with fresh heated boiling water from the tea pot, the other with body temperature warm water from the tap. The clock is operating, the race is on.
posted by vibrotronica at 12:39 PM on June 27, 2012


Heat is like unto Hell, but water is innocent (or how could you baptise someone with it?), therefore the answer is that the forgiving power of God and the redemption of Jesus floweth most readily unto that which most needeth it.

Well, they did say creative
posted by George_Spiggott at 12:48 PM on June 27, 2012 [3 favorites]


"If you take two similar containers with equal volumes of water, one at 35 °C (95 °F) and the other at 100 °C (212 °F), and put them into a freezer, the one that started at 100 °C (212 °F) freezes first. Why?"

I'm not going to say it's not true, because apparently it is tested and true.

BUT

If it is true, how can the boiling (100 °C) water get to freezing (0 °C) without temporarily being all temperatures in between, including 35 °C, which is the temperature of the other water at the start?

And if the formerly boiling water does hit 35 °C along the way, why does it not then freeze as slowly (from that moment) as the other water froze from the start?

What is the difference between the two containers of 35 °C water?
posted by pracowity at 12:53 PM on June 27, 2012


Momentum?
posted by shakespeherian at 12:57 PM on June 27, 2012 [2 favorites]


IAmBrooom asked: Do you believe that unfrozen ice cream liquid doesn't have convection?

I believe the children are our future. But no, I don't believe that the relatively-viscous liquid ice cream has convection that is anywhere near as rapid as that of water.
posted by Joe in Australia at 1:01 PM on June 27, 2012


At 35 minutes, ice is forming on the surface of the (formerly) boiling water, no visible ice on the body temp water. I'll be damned.
posted by vibrotronica at 1:11 PM on June 27, 2012 [3 favorites]


We need to get someone on a space station to try freezing zero-g blobs of water so we can rule out the effects of containers.
posted by Jpfed at 10:14 AM on June 27 [+] [!]


I thought that last word was 'centaurs' for a second, and pictured in my mind the most fantastical expirment possible... until the space-centaurs come and wreck it all.
posted by FatherDagon at 1:14 PM on June 27, 2012


Momentum?

Temperature does not have inertia. It never overshoots, for example.
posted by Mapes at 1:16 PM on June 27, 2012


(Let me be more specific before someone brings up thermostats and hot plates and PID controllers: by lack of overshoot I mean that if you bring a hotter/colder material in contact with a constant-temperature reservoir, the object's temperature will never overshoot that of the reservoir.)
posted by Mapes at 1:20 PM on June 27, 2012


If it is true, how can the boiling (100 °C) water get to freezing (0 °C) without temporarily being all temperatures in between, including 35 °C, which is the temperature of the other water at the start?

You are assuming homogeneous distribution of temperatures. Water at the edges of the container will cool before water in the center, and it is likely the precise distributions from 35C to 0C will never look like the distribution from 100C to 0C. One distribution freezes easier than the other, apparently.
posted by effugas at 1:26 PM on June 27, 2012


Temperature does not have inertia.

I think that was just a joke.
posted by pracowity at 1:28 PM on June 27, 2012 [2 favorites]


vibrotronica - do you have the means to measure the current temperature of the two samples of water? I can't seem to find any mention of measurements of the actual temperatures involved. I'd like to know for certain whether the (formerly) hot water sample is crystalizing (freezing) at the same temperature as the (formerly) cold water sample.
posted by It's Raining Florence Henderson at 1:28 PM on June 27, 2012


Sorry, it's Raining Florence Hendersons, all I have is a freezer and a stopwatch.

At 1 hour 5 minutes, the no-longer-boiling tray is completely iced over but still liquid inside the cubes, while the warm water tray is only about 70% iced with liquid water still visible on the surface.
posted by vibrotronica at 1:41 PM on June 27, 2012


Heat is like unto Hell, but water is innocent (or how could you baptise someone with it?)

Dante would like to have a word with you.
posted by mykescipark at 1:46 PM on June 27, 2012


I present the Pop-Tarts theory:

1. So hot it's cool.
2. So cool it's hot.
3. Therefore the hot freezes faster.
posted by Kabanos at 1:54 PM on June 27, 2012 [2 favorites]


Well, the problem here is that the heating of the water relies heavily on the need to boil it in a pot first. However, to empirically observe this phenomenon one must watch the pot, which negates the ability of the pot to poil. Perhaps you could try to observe the phenomenon backwards, since it will be much clearer in hindsight.

While a description of the actions you took to replicate this phenomenon might be heard by more people, most journals will want a full write up. You could also try to photograph the phenomenon without watching the pot, but most of the scientific literature has much longer length requirements than 1000 words.

One possible solution involves using a kettle instead of a pot, but since the pots in most kitchens are black, you aren't really proposing anything different. And to be honest, since this experiment involves so much heat, it probably shouldn't be done in a kitchen anyway.
posted by This_Will_Be_Good at 1:55 PM on June 27, 2012


Something DaddyNewt mentioned above is similar to what I was thinking.

Solid water forms a crystal lattice with a specific configuration. Water in the liquid state, while fluid, also has a fleeting microstructure to it. Specifically, water molecules align to form "cages" of water molecules with space in between them. They do this because their dipoles align with partial negative charges and partial positive charges in proximity.

In order for liquid water to freeze, you can consider it the sum of two processes. The first process requires the liquid water to break apart these microcages. Then the molecules can align themselves in the solid crystal lattice configuration. Warmer water, because of the dispersion of the molecules due to higher temperature, does not have to overcome the microcage interactions of liquid water as much. That is, the microcages are already disrupted.

Hence, the warmer water molecules can go directly to aligning in solid crystal lattice configurations They don't first have to undergo an entropically favored (but enthalpically disfavored) process by which they become more disordered first (the non-caged arrangement), before then being able to undergo the transition to the entropically disfavored, but enthalpically favored solid lattice arrangement.

TL, DR version: I suspect it's all about the transition barrier of the entropic process. For those of you who have studied Gibb's Free Energy, you know that i spontaneous processes, sometimes entropy issues can be large enough to overwhelm enthalpic ones when the entropy change is significant and temperature is higher.

I'd have to run some calculations and controlled experiments to figure if this were the case, but it's a theory, anyway. :)

Also, love the suggestion above that other liquids should be tested. If the effect is substantially restricted to the behavior of water, then I'd suspect it's almost certainly entropic issues related to molecular arrangements in liquid vs solid state.
posted by darkstar at 2:02 PM on June 27, 2012 [4 favorites]


As a minor addition to darkstar (and DaddyNewt)'s thoughts: isn't water unusual in that it expands on freezing? It seems like to arrange itself this way, which in some sense works against gravity given a rigid vessel, a shortage of random, energetic motion would tend to delay this process.
posted by George_Spiggott at 2:09 PM on June 27, 2012 [1 favorite]


darkstar, it seems like you're suggesting either that the warmer water, once it reaches 0°C, retains a "memory" of being warm that the colder water doesn't; or that the warmer water freezes at a different temperature than the colder water. Can you clarify?
posted by Mapes at 2:11 PM on June 27, 2012


1 hour 40 min: boiling water is now completely frozen into cocktail-ready cubes. Non-boiling still liquid in the middle. Terminating experiment on account of I have Flaming Lips tickets.
posted by vibrotronica at 2:16 PM on June 27, 2012 [7 favorites]


Experiment suggests your flaming lips will turn to ice before my 98.6 degree lips. Enjoy!
posted by It's Raining Florence Henderson at 2:31 PM on June 27, 2012 [3 favorites]


Mapes, I'm not sure, but it's not my intention to imply those things. Basically, both the warmer water and the cooler water are identical when they are making the final transition to freezing. I think what I'm suggesting is that they had to take different pathways to get there, though.

That is, the cooler water had to go through an entropic detour first to reach its destination, while the warmer water didn't. So that, although the warmer water started farther away from its destination (in a sense), it gets there first.

Of course, this is spit-balling, and I'd really need to do some reading up on the entropic energy of the various microstates we're talking about, etc. But entropy is my primary suspect, at this point. :)
posted by darkstar at 2:33 PM on June 27, 2012


There's a decent chart of specific entropy of water by temperature, pressure, density, etc., here.
posted by It's Raining Florence Henderson at 2:38 PM on June 27, 2012


It'd be awesome to have precise measurements of both vessels over time, and particularly at point at which each is completely frozen. I believe it's well understood that water can be liquid or frozen at the same temperature (you readily see this in flowing rivers). If more-energetic water can (to clumsily paraphrase darkstar's much better language) more easily overcome the barriers to forming a lattice than less-energetic water can, then the cooler water may very well be freezing at a lower temperature, because it has less energy to spend on freezing, as it were.

Does that make sense? I thought not.
posted by George_Spiggott at 2:42 PM on June 27, 2012


The higher the temperature, the longer the run-up and more speed you get, obviously. Like a cheetah running 100m racing a man with a 60m headstart and eventually overtaking him. Next problem please!

I'll have it in used bills please.

Actually, this effect always annoyed the crap out of me that I never understood it in physics. I didn't realise that no one else did either.
posted by NailsTheCat at 2:50 PM on June 27, 2012


Wow, amazing thread. I had heard about this. Is it also true that hot water has less oxygen and could this play a role?
posted by marienbad at 3:10 PM on June 27, 2012


Well, I'm not sure whether the "entropy" argument holds water (heh, heh). But entropy can explain a lot about spontaneous processes that aren't enthalpically favored.

So it's kind of like when I teach organic chemistry: if you don't know the answer for why something behaves as it does in an organic reaction, the best thing to suggest is almost always "resonance stabilization". At the very least, it shows you've learned some of the jargon! :)
posted by darkstar at 3:21 PM on June 27, 2012


It was very nice to see that brief allusion to the long history of the issue in Kiwi's RSC link, because I've never known how seriously to take reports of the existence of the phenomenon in the first place:
Why does hot water freeze faster than cold water?

It seems a simple enough question - yet it has baffled the best brains for at least 2,300 years.

* Aristotle agonized over it fruitlessly in the fourth century BC
* Roger Bacon in the 13th century used it to advocate the scientific method in his book Opus Majus
* Another Bacon, Francis, wrote in his 1620 Novum Organum, that "slightly tepid water freezes more easily than that which is utterly cold" but could not explain why
* Descartes was defeated by it in the 17th century AD
* Even perplexed 20th and 21st century scientists and intellectuals have swarmed over it without result
My guess is that this is due to the way water expands when it freezes, unlike the great majority of other liquids (liquid gallium happens to be another one of the very rare exceptions).

Water reaches its maximum density (0.999973 g/ml) around 4°C (~39°F). Water at 0°C is 0.999841 g/ml, which is slightly less dense than water at 8°C (0.999849 g/ml).

Imagine putting a container of water into the freezer which is 8°C or less.

The floor of the freezer is below 0, so the water right next to the bottom of the container will get denser and denser as it cools toward 4°C, and since denser things sink, it will stay right there.

As it cools further, however, it becomes less dense, but it will still sit tight because it's still denser than the 8°C water above it, yet as it approaches 0°C, instead of sitting there and freezing, it's suddenly less dense than all the water above it, and it will (neglecting heat exchange as it rises) go all the way to the top of the container.

The new parcel of water which replaces it at the bottom goes through the same process, which will continue until all the water is very close to 0°C, and then ice will form at the bottom and freezing will proceed.

Freezing will only really take place at the bottom and sides because ice gives off a tremendous amount of heat as it freezes-- heat which is quite distinct from the heat it gave off as it cooled down-- and that heat has to go somewhere other than the water itself.

By contrast, the same container with the same amount of water at, say, 12°C will behave very differently.

In that case, the water as it approaches 0°C is less dense only than the band of water just above it which has been cooled to a temperature just above 8°C. That parcel of 0°C water will rise only to the boundary layer with the 8+-12°C water and stop. A very narrow band of very cold water will form at the bottom of the container, and freezing will take place there while the great bulk of the water is still at an elevated temperature and much sooner than than any freezing took place for the container of 8°C or less water.

I think the earlier appearance of the first ice (in this scenario I'm suggesting) is the key to the mystery.

A quantity of water at 0°C gives off about the same amount of heat when it freezes as it takes to raise that identical quantity of water to almost 80°C (if I did that estimate correctly), making it the bulk of the heat that is removed when water is frozen unless the water starts out piping hot.

That heat has to flow into the heat sink and be removed, and because the temperature of the water as it freezes is 0°C, and the temperature of the heat sink (the floor of the freezer) isn't much less, the rate of that heat flow is low, and takes a long time because the rate of heat flow is proportional to the difference in temperature between source and sink.

In short, I'd say warm water freezes faster than really cold water because a rate-limiting step of the process (removal of the latent heat of fusion of the freezing water) gets started sooner for warm water, and that the point at which it really speeds up should be right around 8 to 12°C
posted by jamjam at 4:25 PM on June 27, 2012


But does that explain why 95C water freezes slower than 100C water, jamjam, the specific example given? You're thinking about much colder temperatures, but the specific example we're given uses 100C and 95C.
posted by Malor at 4:38 PM on June 27, 2012


I think it relates to what I said early on. It has to do with the physical properties of water molecules. It's more efficient to form a lattice with very short strings of molecules than it is to form a lattice with long strings of molecules.
Mind you, I graduated from college 30 years ago and I never had a physics class in my life, but my last name is Newton, so come at me physics bros.
posted by DaddyNewt at 4:50 PM on June 27, 2012


This reminds me of my Recipe for Ice Cubes. To make really nice, really clear ice cubes, use the hottest water you can easily manage.

It's true! Trying filling an ice cube tray with really hot water and another one with cold water. After you freeze them, compare the clarity of the cubes from each tray. You'll see what I mean.

After I learned about this, I became an absolute asshole about ice cubes. I'd go to someone's house, they'd give me a drink, and I'd think "you have cloudy ice cubes." It would totally make me lose respect for whoever made the cubes. If you don't care about your ice cubes, what else don't you care about?

That's why I'm sharing the Recipe with you, so assholes like me won't think you're an idiot.
posted by twoleftfeet at 6:59 PM on June 27, 2012 [1 favorite]


I had the idea that this whole thing had been debunked: that it didn't really happen. I even thought there was an askme about it a while back. Is the effect easily reproduced? Or is it more like cold fusion: some can see it and some can't.
posted by DarkForest at 7:13 PM on June 27, 2012


Perhaps you might like to explain this phenomenon; at the exact time I began reading this thread, it was being discussed on the radio in the next room...

Do radio discussions cool faster than internet threads????
posted by arzakh at 8:51 PM on June 27, 2012


The answer is easy: Velocity.
posted by Neale at 11:00 PM on June 27, 2012


The heat agitates the phlogiston, which in turn elicits a sympathetic response from the cooling humours. The effect is most pronounced between Lady Day and Michaelmas with the moon gibbous.
posted by Abiezer at 8:29 AM on June 28, 2012 [1 favorite]


For years my brother and I had this discussion (very tongue in cheek), not knowing it was a thing. We even named it the Law of Phase Change Momentum. If pressed we would discuss at length how it applied not only to freezing boiling water, but also to boiling ice - colder ice boils faster
posted by ElGuapo at 11:30 AM on June 28, 2012


Mapes: "65Momentum?

Temperature does not have inertia. It never overshoots, for example.
"

But what about the process of lattice-making?* Couldn't that have momentum? Or perhaps not momentum per se, but a sort of organisational efficiency, prohibiting "noise" in the process of freezing. Like how a multi-lane highway with lots of merging/crossover lanes will just be an utter mess, especially at slower speeds...but efficiently moving, well-designed multi-lane highways will facilitate traffic, at any speed. And the hot water molecules self-organize much better than the cold ones, no? So they set up nice clear trajectories for themselves as they get colder, making the whole operation go more smoothly. As evidenced by less cloudy cubes. Same is true for ice/snow in a cold snap, right?

*Sounds like a title track from a first album by a band of science geeks. Or carpenters.
posted by iamkimiam at 10:58 AM on June 29, 2012


I… is this really even a thing? When we were kids mu brother and I, in disbelief of our mom's insistence that hot water froze faster, made ice cubes with one tray of hot water and one tray of cold. The cold water ones froze faster, which didn't surprise us at all.

Oh, and someone upthread mentioned damp vs. dry sponges — a slightly damp sponge soaks up water much more quickly and effectively for reasons related to why water has surface tension. It doesn't necessarily hold MORE water than a dry sponge.
posted by DoctorFedora at 3:13 PM on June 29, 2012


« Older “[...] it took more than a dozen calls to work out...  |  A German court has ruled that ... Newer »


This thread has been archived and is closed to new comments