Mpemba effect

[tusenfem]

Okay, the first time I start a thread here.

I was reading the newspaper this weekend, and in the book section there was this book about what we do not know about physics (unfortunately I cannot find the newspaper, but as it was in German it will not be useful for most BAUTers). One of the things that was mentioned was the Mpemba effect which states in popular language that a hot fluid freezes faster than a cold fluid. This was found out by the Tanzanian Mpemba as he put boiled milk in the freezer.

I always thought this was an urban legend, but wiki has a page on it and there are papers in the American Journal of Physics, which at least clears it up a bit. It is not that a glas of hot milk becomes icecream faster than a cold glas of milk, but the first layer of ice will occur on the hot fluid.

Unfortunately, I have no access to the AJP, so I was wondering if anyone has more info about experiments that have been made, or have done themselves. At least it is a very nice thermodynamical problem.

[antoniseb]

I seem to recall that hot water has fewer dissolved gases, and that this contributes to the effect. Personally, I have never tested it.

[mfumbesi]

I found an explaination attempt on this link http://www.school-for-champions.com/science/mpemba.htm

[Ken G]

It is not that a glas of hot milk becomes icecream faster than a cold glas of milk, but the first layer of ice will occur on the hot fluid.

Even that seems to be under debate, as the link from mfumbesi is indeed talking about the whole business freezing faster for warm milk than cold milk. But it might not be authoritative. There certainly is one very trivial example where very warm milk will freeze faster than very cold milk-- if the warm milk is actually so hot that it vaporizes and most of it evaporates. For more reasonable temperatures, it's not clear how big a role the reduction of the volume by evaporation plays. The dissolved gas issue seems more promising, but that should have been easy to test, one would think. A question that would be nice to answer is, does the temperature of the warm milk as a whole ever "pass" below that of the cold milk?

[George]

I think the African translation for this effect is known as "Mumbo Jumbo".

Yesh!

Supercooling. It is hypothesized that cold water, when placed in a freezing environment, supercools more than hot water in the same environment, thus solidifying the hot water faster.

Anyone care to help me with that one?

The effect of solutes such as calcium and magnesium carbonate.

What does this mean? Are we comparing the rate of freezing of two entirely different samples of water, or does warm calcium help water freeze faster. If so, what happens when the calcium cools to the temperature of the other sample? How would it behave differently? What has happened is time has elapsed allowing the cooler water to have frozen quicker, not slower.

Evaporation has some hope of helping, but it might have to be boiling before going into the freezer. This isn't the toughest type of experiment to do; where are they?

[01101001]

Straight Dope: Which freezes faster, hot water or cold water?

The results? The cold water froze about 10 or 15 minutes faster than the hot water, and there was no detectable difference between the boiled water and the other kind. Another old wives' tale thus emphatically bites the dust.
[...]
I happened to come across an article in Scientific American entitled "Hot Water Freezes Faster Than Cold Water. Why Does It Do So?"
[...]
Fascinating as all this no doubt is, all it basically proves is that very hot water freezes more slowly than very VERY hot water. The ordinary fumbler in the fridge, on the other hand, is dealing with temps more like the ones I was measuring, in which case cold freezes faster than hot. I rest my case.

This isn't the toughest type of experiment to do; where are they?

Physics FAQ: Can hot water freeze faster than cold water?

Yes -- a general explanation
Hot water can in fact freeze faster than cold water for a wide range of experimental conditions. This phenomenon is extremely counter- intuitive, and surprising even to most scientists, but it is in fact real. It has been seen and studied in numerous experiments.
[...]
EXPERIMENTS ON THE MPEMBA EFFECT
5. Mpemba and Osborne, "Cool", Physics Education vol. 4, pgs 172--5 (1969)
6. Ahtee, "Investigation into the Freezing of Liquids", Phys. Educ. vol. 4, pgs 379--80 (1969)
7. I. Firth, "Cooler?", Phys. Educ. vol. 6, pgs 32--41 (1979)
8. E. Deeson, "Cooler-lower down", Phys. Educ. vol. 6, pgs 42--44 (1971)
9. Osborne, "Mind on Ice", Phys. Educ. vol. 14, pgs 414--17 (1979)
10. M. Freeman, "Cooler Still", Phys. Educ. vol. 14, pgs 417--21 (1979)
11. G.S. Kell, "The Freezing of Hot and Cold Water", American Journal of Physics, vol. 37, #5, pgs 564--5, (May 1969)
12. D. Auerbach, "Supercooling and the Mpemba effect: When hot water freezes quicker than cold", American Journal of Physics, vol. 63, #10, pgs 882--5, (Oct 1995)
13. J. Walker, "The Amateur Scientist", Scientific American, vol. 237, #3, pgs 246--7, (Sept. 1971)
14. B. Wojciechowski, "Freezing of Aqueous Solutions Containing Gases", Cryst. Res. Technol., vol. 23, #7, pgs 843--8 (1988)

[Peter Wilson]

My high-school physics teacher explained that an ice-cube tray filled with warm/hot water melts, then re-freezes, a thin layer of ice where it touches the freezer, establishing a better thermal contact...like sticking your tongue to the freezer

I bought his explanation, and have not researched it any further. Of course, that was back-in-the-day when they taught that due to orbital changes, the next Ice-Age was just around the corner

[snarkophilus]

It looks like an open problem. Many mechanisms have been proposed, but I haven't seen anything in which the importance of the various processes has been compared. Lots of conjecture, little experimental work.

At first glance, no single effect ought to dominate over all temperature ranges. For instance, only so much gas can dissolve in a liquid... above or below certain temperatures, there won't be much difference between two liquids in this regard. At high temperatures, evaporation rates are probably important -- entropy is a powerful thing -- but at low temperature, there won't be much evaporation, so the effect will be minimal. (You'd also need to consider vapour pressure/humidity in the evaporation calculations.)

[George]

Physics FAQ: Can hot water freeze faster than cold water?

Thanks, perhaps there is some evidence, but we need to see the details (when time allows).

I am still dubious...

Why hasn't modern science answered this seemingly simple question about cooling water? The main problem is that the time it takes water to freeze is highly sensitive to a number of details in the experimental set- up, such as the shape and size of the container, the shape and size of the refrigeration unit, the gas and impurity content of the water, how the time of freezing is defined, and so on. Because of this sensitivity, while experiments have generally agreed that the Mpemba effect occurs, they disagree over the conditions under which it occurs, and thus about why it occurs.

I would still bet there are no normal general conditional experiments where "hot" water wins over ambient temp. water.

[tusenfem]

I just received some paper for Tim Thompson which I will read in the weekend and get back to y'all.

[Universal Curiosity]

Hmm... I never knew that this effect was so poorly understood. I thought that the explanation was that the particles in liquids at higher temps were moving about faster than lower temp liquids and therefore able to arrange themselves into the structure necessary for solidification faster. I'm not sure where I heard or read this or if I just made it up at some point in the past. It seems my explanation was wrong. I think I've told others that was the reason for the effect. Oops.

[Jerry]

Milk or Water?

Is the freezing from the surface down, or the container up? Homogenized, or not? How hot is hot? How cool is cold?

It would be very easy to set up conditions so that hot milk freezes faster than cold milk: Heat up unhomogenious milk, and the fat globuals will disperse in the fluid, while cold milk cooling from the surface under otherwise identical conditions will have an insulating layer of fat.

Another possibility, is boiling the milk will remove volatiles, increasing the freezing point.

Finally supercooling is more likely to occur when something is chilled slower - which would be more likely to occur in a cool liquid than a hot one.

All told, there are many competing physical phenomenon.

[Michael Noonan]

What about thermal gradient difference in the liquid itself?

It didn't seem to get a mention. Could there be some mechanism at play there perhaps, cheers.

[publius]

Another thing that contributes to this is most of the time, hot water pipes freeze and burst before cold water pipes. You know, it comes a "hard freeze" unexpectedly and you forgot to close up the vents under the house, or something. Hot water pipes, or so everyone seems to say, will burst first.

There is a lot of debate over just what causes that as well. The difference here is not one of temperature. Unless you've been running the hot water, and your pipes are well insulated, the water in the hot side will be about the same temperature as the cold side.

So we're talking about differences in water that has been heated (and in a typical water heater) and held there in the tank vs water than wasn't heated and didn't stay in the heater tank.

One theory is dissolved gases. You'll notice little bubbles in ice sometimes, and some argue that can absorb the expansion and prevent the pipe from bursting. Heating it and holding it at temperature drives the gases out.

Another is heating tends up to break up solid particle impurities in the water that will serve as nucleation centers. Here, the water in the cold side freezes first right on the edge, and the latent heat released keeps the center warmer. Pipe freezes from the outside in, and allows the water in the center to flow a bit, squeezing out of leaky faucets, etc fast enough to keep the pipe from bursting.

However, the heated water has smaller particles, and some supercooling happens. When it finally freezes, it goes quickly, and bursts the pipe.

Note here, this explanation depends on the unheated water actually freezing faster than the heated.

-Richard

[Jerry]

Another thing that contributes to this is most of the time, hot water pipes freeze and burst before cold water pipes. You know, it comes a "hard freeze" unexpectedly and you forgot to close up the vents under the house, or something. Hot water pipes, or so everyone seems to say, will burst first.

In my personal experience (living in a cold climate), it has always been the cold pipes that freeze first; but this may also be because they have outside bibs that extract heat. The other exception has been hot water pipes in trailors/campers where the heating system is shutoff.

Finally there is the case of an indoor system where the hot pipe froze but the cold did not: The toilet at the end of the cold pipe was always running.

[George]

Finally there is the case of an indoor system where the hot pipe froze but the cold did not: The toilet at the end of the cold pipe was always running.

There ya go! I think you just nailed it. Some toilet leaks are hardly perceptible. Also, there is their use as well.

[01101001]

OK so I might be rushing it a bit, but this is what I thought.

Zero point energy is 10^120 and so would have exploded in micro seconds.
That would still be an energy ball.

This is Q&A. I don't see any questions in your article so they must be answers. Now, I must ask: answers to what?

This topic is about:

I was wondering if anyone has more info about [Mpemba effect] experiments that have been made, or have done themselves.

What are your answers about?

If it's not about the topic, take it elsewhere. Please.

Thanks.

=======================

Edit: The Michael Noonan zero-point-energy article (was number 1038246) to which I responded seems to have disappeared. Perhaps someone else also didn't see the connection with the topic.

Edit: Spotted it, or at least some of it: in ATM.

[publius]

In my personal experience (living in a cold climate), it has always been the cold pipes that freeze first; but this may also be because they have outside bibs that extract heat. The other exception has been hot water pipes in trailors/campers where the heating system is shutoff.

Finally there is the case of an indoor system where the hot pipe froze but the cold did not: The toilet at the end of the cold pipe was always running.

I don't doubt that a bit -- things certainly vary. But take a poll of your local plumbers and see what they say. The ones around here all say that when they get a call about a bursted frozen pipe, odds are it will be a hot water pipe. If they were taking bets, they would come out ahead betting on hot. They'd loose sometimes, but win more more often than not. And because of that, plumbers are generally big believers in the "hot water freezes faster" effect. Indeed, it was from a plumber that I first heard this stuff as a young buck around 20 -- I was wiring houses in the summer, and so met up with a lot of plumbers and other trades.

One generally wants to insulate pipes under the crawl space. And it can help a bit to keep the water hot to insulate the hot water lines everywhere, but I don't know if it does much good. Anyway, I noticed a plumber *carefully* insulating his hot water lines all through the walls. Why, I asked. And that's when I first heard the hot water freezes faster. And I thought it was just the biggest load of bunk I ever heard, right up there with all that phase of the moon stuff.

And it wouldn't surprise me if this varies with the water supply -- water from different sources with different impurities may behave slightly differently.

-Richard

[George]

Indeed, it was from a plumber that I first heard this stuff as a young buck around 20 -- I was wiring houses in the summer, and so met up with a lot of plumbers and other trades.

I know many plumbers, if they say it is true, it likely is. Insulating hot water lines, however, is important in minimizing both the wait time for a hot shower and the wasted cold water expended while waiting. The electric bill also is less with these lines insulated.

I believe Jerry's explanation is adequate.

A 1/2 inch dia. water line will only hold about 2.5 gallons if it is 60 ft. long -- the length to the service meter, roughly. An average toilet flush will use this, or, in most cases, twice this volume. So, even if there are no leaking toilets or faucets, though these are quite common, then one flush gives the cold water line a big advantage. [The service lines to the house are larger than 1/2 inch, but this section doesn't count since it is both the hot and cold water feed line.]

The claim still seems mumbo jumbo to me. Will frozen water boil faster than tap?

[publius]

San Antonio may be different, but I'll bet you a rusty nickel they'll agree that hot water pipes do burst more than cold.

There's no harm in insulating hot water lines, none at all, and it will help a bit if you're using enough hot water. However, if you're like me, and maybe run the sink once in the morning, then take a shower at night, that water is going to cool down, unless you do some serious insulation indeed.

One solution that probably wastes more energy than it would save is a circulation system. You put a little pump in the hot water circuit and loop a return back to the heater. You need check valves and want to "wire" that carefully.

I knew a fellow who put one of those in. When he turned on the hot water, he'd get a few seconds of nice hot water, then it would turn cooler and finally cold. He had done something backwards, either put a check valve in backwards, or failed to put it in. Anyway, what was happenning was when he turned on the hot water, cold water at the bottom of the tank, which had just come in to replace the water flowing out, was flowing out through the circulation return line and back into the system.

Inline "instant heaters" are another way to get instant hot water. I don't know what the state of the art is, but if you want to heat a flow of water that rapidly it takes a pretty good wattage heater. It's hard to keep from wasting heat that way. They may have something good now that doesn't waste much, but I don't know.

ETA: It requires a subscription to get (from here), but someone wrote a paper on hot water pipes bursting more often:

http://prola.aps.org/abstract/PR/v8/i5/p500_1

-Richard

[George]

San Antonio may be different, but I'll bet you a rusty nickel they'll agree that hot water pipes do burst more than cold.

I bet you are right, but I will bet an ice cream that it is not because warm water freezes faster. We don't get much freezing down here. It has been stated that water freezes at 42F down here because "even the water ain't used to it!". [I guess that joke won't work in a BAUT crowd, shucks! ]

If plumbers would simply conduct a survey that ask the frozen hot water pipe owners their flush frequency, as well as, check for any slight leak in toilets or elsewhere, then we could confirm or tweak Jerry's toilet theory, err... Jerry's Hydrodynamic Hypothesis.

However, if you're like me, and maybe run the sink once in the morning, then take a shower at night, that water is going to cool down, unless you do some serious insulation indeed.

Yes, but with four kids, the time between persons can waste the heat in those lines where the last few bathers will suffer, voicing no small concern.

One solution that probably wastes more energy than it would save is a circulation system. You put a little pump in the hot water circuit and loop a return back to the heater. You need check valves and want to "wire" that carefully.

It would help heat the slab and eliminate cold feet, perhaps. Some northern motels have used this system for all their room heat.

[Jerry]

One generally wants to insulate pipes under the crawl space. And it can help a bit to keep the water hot to insulate the hot water lines everywhere, but I don't know if it does much good. Anyway, I noticed a plumber *carefully* insulating his hot water lines all through the walls. Why, I asked. And that's when I first heard the hot water freezes faster.

...And any thermal chemist who has worked with a calorimeter would tell you, all things being equal, that is nuts. The ASTM for using water to calibrate temperature calls out freshly distilled water; but if you pull it out of the still at 70C it will take more calories, and more time to complete the heat transfer, than if you cool the distillation down to room temperature for ~24 hours. In this case, desolved gases may depress the freezing point a few tenths of a degree, but it will still take ~49.7 more calories to freeze a cc of 70C water than a freshly distilled cc of water starting at 20C...unless the toilet is running.

[Cougar]

...a hot fluid freezes faster than a cold fluid. This was found out by the Tanzanian Mpemba as he put boiled milk in the freezer.

Well, there's a clue right there that this is a tall tale. Tanzanians don't have freezers.

[mugaliens]

Interesting - this appears to be similar to the all over the Internet "supporting evidence" that people were using to claim the coriolis effect couldn't possibly occur on human-sized bodies of water.

The only difference is that the incredulaty of the average reader here is much higher.

Wow. That opened a can of worms...

Ok, let's approach this systematically:

1. Closed system (stainless steel containers with hermetic lids)

2. Two identical masses of distilled water at the same room temp.

3. Heat one by 20 degrees.

4. Plunge both containers into a brine solution cooled to -10 degrees C.

The cooler one freezes solid before the warmer one.

That's a surprise?

As for the explained reasons behind the supposed Mpemba effect, all of them involve tampering with the systems (different gases/minerals in the hotter volume, less actually freezing with more off-gassing, etc.).

Of course if the conditions of the hotter sample are different than the conditions of the cooler sample then it's possible that the hotter sample will freeze before the cooler sample. Is that so surprising? Less mass, differing composition, additional means of cooling?

That's no "effect." It's simple physics 101, which also states that given identical masses, the cooler one freezes first, and this is borne out in experiment.

[snarkophilus]

As for the explained reasons behind the supposed Mpemba effect, all of them involve tampering with the systems (different gases/minerals in the hotter volume, less actually freezing with more off-gassing, etc.).

Yes, for a pure liquid system... but the original claim is based on the differences in initial conditions in a multi-phase system caused by the temperature difference. Even in your steel containers with hermetic lids, there will be different gas solubilities, and thus different pressures of water vapour, nitrogen vapour, et cetera. You can't just discount those explanations on the basis that it is a different experiment, because they are unavoidable differences. (It may well be that having a sealed container mostly offsets the solubility effects, with pressure build-up allowing only a limited amount of evaporation, but it doesn't have to. One would need to prove that.)

Same goes for Jerry's comment. If it was a pure ideal liquid in a closed container, then yes.

Something that has occurred to me while thinking about Jerry's distiller: by definition, these samples aren't starting at equilibrium, so we don't even have nicely defined starting points. There could be differences in freezing time depending on how long it took to heat up the sample. Something that might cause a difference (and something that might have easily affected Mpemba) is the amount of evaporation that took place during heating. If the hot sample mass is smaller than the cold (and even at equal volumes, this will be so!), then it might contain less energy.

[publius]

Let me clarify something I fear may be lost. There is a big difference between the Mpemba effect as stated, and hot water pipes bursting from freezing more than cold water pipes. You can have pipes that freeze without bursting, too if something can take up the expansion. I brought up the latter as a big reason by plumbers tend to be big believers in the Mpemba effect, but via the flawed reasoning that hot water pipes always contain water that is at a high temperature than the cold water pipes.

The hot water pipe effect depends on water *that has been heated in the past* having different properties than the water from the same source that has not been heated. These are two different possible things, here. Related, but different.

So, we heat a sample of tap water, hold it there for a while to simulate sitting in a water heater tank. Now, we cool that sample back down to the same temperature as the cold water supply. Can there be differences there? I don't think it takes any great stretch of imagination to accept there could be differences, and those differences are the reason for this purported behavior.

-Richard

[tusenfem]

Okay, I promised to come back to you after the weekend. It was busy, so I only managed to read one, but a very enlightening, paper by Monwhea Jeng, published in the American Journal of Physics, vol 74, pg. 515-522. Thanks to Tim Thompson for sending the pdf to me.

This paper described the notion of the Mpemba effect starting at the time of Aristotle, and when you read it you find that it is a highly difficult problem, to begin with, difficult in stating what one exactly means by "hot water freezes faster than cold water." First of al, he shows that not all experimenters did the same experiment or looked for the same effect. It would go too far to write all down what Jeng claims, and anyone who would like to have a copy of the paper (with numerous references to other papers) can get it from me.

Clearly it will not always function, if you take a liquid at 99.9 degrees celcius and one at 0.1 degrees celcius you will find that the cooler one will freeze sooner. Therefore, it is a parametric problem, such that "there is a temperature range within which warmer liquids will freeze faster than cooler."

Basically, there were a lot of explanations, convection in the fluid while freezing, the containers standing on frost which melts for the hot liquid giving a better conducting area at the bottom (but the lab experiments were performed with frost-less cool elements), degassing (but there are various experiments that did not show any difference), super-cooling (which basically gets to be the explanation for the bursting hot-water pipes, a related topic but not the same, as Publius says above), the size and shape of the container, etc. etc.

Basically, this is a very interesting and real phenomenon, but very difficult to investigate as it depends on lots of paramenters. A very interesting study was done by Walker with different containers and different thermocouples and different temperatures, and found that for some of them nothing "weird" happens and colder fluids freeze faster than warmer (Hurrah for thermodynamics!) but there are also parameter regions, in this case temperature regions where warmer fluids freezes faster (Hurrah for Mpemba).

Accofding to Jeng supercooling is no solution for the Mpemba effect, although a paper a year Earlier by Auerbach (Am. Journal of Physics, 1995, vol 63, pg. 882-885) seems (according to the abstract) to make a link between supercooling and the Mpemba effect.

So, interesting interesting, and I am not sure if we will soon find out the complete model for this effect. But it is a nice "sobering" phenomenon for physicists (like me) that "common sense does not always prevail" (as I often claim in the ATM area) and in this case "common physics sense" does not always prevail.

For copies, please contact me with an email address.

[Jerry]

Adding to the list of reasons hot water pipes burst first:

Hot water is stored in a chamber (water heater) that allows for expansion. Cold lines do not.

Plus - Cold lines are generally not insulated, allowing faster heat transfer from the center of a house to the edges, where pipes cool the quickest.

So the cold lines transfer heat from the center of the house to the edges, while the hot lines do not.

(This would likely only be a factor if highly conductive copper plumbing is used, very common in the US.)

[George]

Hot water is stored in a chamber (water heater) that allows for expansion. Cold lines do not.

That could be an important point, but how? Wouldn't the accumlator action of the water tank actually serve to reduce the number of hot water line ruptures?

[Jerry]

That could be an important point, but how? Wouldn't the accumlator action of the water tank actually serve to reduce the number of hot water line ruptures?

I was thinking about compression: When Ice starts to form in a line, it expands and increases the pressure in the line, which in turn, lowers the freezing point of ice. If the hot lines initially relieve pressure build-up, there is less freezing point depression and the ice forms more quickly. Then bulk effects come into play: the larger ice mass in the hot lines act like a big stopper, and the pressure in the coldest zone becomes critical.

Unfortunately, you can use almost the same reasoning to predict faster bulk effects in cold pipes, so it is an iffy argument.

Corrosion is not: Hot pipes corrode much more rapidly than cold pipes. (This is especially true if the heater introduces a galvanic action in the piping.) More corroded pipes break easier. Of all the possible explanations touched upon, I think this is the best.

Once upon a time I plumbed a hot tub up to an ordinary gas water heater. The heater was a floor below the tub, and I plumbed it so that passive circulation kept the tub warm. It worked great; cut my electric bill by 30%; for about two years - then the standard water heater simply fell apart, completely rusted out by the pool chemicals.