Explanation of Mpemba effect for freezing of water

[amal]

I recently came across an odd phenomenon known as Mpemba effect. It says that hot water freezes quicker than cool water provided that temperature difference is not too much. E.g. it can be, say, 30 degrees Celsius and 50 degrees Celsius but not something like 99 and 1. Unfortunately, the book did not say anything about why this happens. Can somebody tell me?

 

[blue spruce]

The Mpemba effect is caused by impurities in the water or on the surface of the containers holding the water. For experimental results that support this conclusion see:

“When does hot water freeze faster then cold water? A search for the Mpemba effect.” by James D. Brownridge in the American Journal of Physics # 79, (2011), 78-84.

Or

http://arxiv.org/abs/1003.3185


“Ordinary drinking water contain many harmless impurities,
which can act as ice nucleation sites, each of which have
an ice nucleation temperature, that is, the temperature at
which that site will imitate heterogeneous freezing. Still water
in a closed container will almost always supercool and
will therefore not begin freeze until its temperature falls to
the temperature of the ice nucleation site with the highest ice
nucleation temperature. Consequently, if a container of hot
water has ice nucleation sites with an ice nucleation temperature
-5.5 °C or higher than the ice nucleation site with the
highest ice nucleation temperature in the cold water, then the
hot water will freeze first. If this is not the case, the cooler
water will freeze first, and the Mpemba effect will not be
observed.”

 

[russ_watters]

In other words, the effect is not generally observed: it only happens under very specific conditions.

 

[blue spruce]

You that is correct.

 

[bahamagreen]

Wasn't this first observed by a student doing this experiment with a refrigerator/freezer and ice cube trays?

I wonder... if you time the freezing of cold water, then time the freezing of warm water... won't the warm water more likely activate the thermostat and engage the compressor for longer periods, thus chilling the warm water "more aggressively" than when using cold water?

The proper test would be to use a sub-zero environment that is not actively temperature controlled based on the measurement of temperature (no feedback control loop to confound the test) - a freezer that is cold and powered off.

 

[Drakkith]

Bahama, I could see that if hot water was right in the path of the freezing air, otherwise I don't believe that would make much of a difference.

 

[blue spruce]

Yes you are correct, the effect was re-discovered my Mr. Mpemba when he was a student.

A “proper test” would be to test the hot and cold water samples at the same time in the same freezer while maintaining all conditions except sample temperatures as identical as possible until both samples freeze. Sure the hot water will change the conditions inside the freezer; however, both samples will experience the change and under these conditions the cooler water will cool to 0 oC first. Note: still water will usually not begin freezing at 0oC, it will supercool. The temperature to which a water sample supercools will determine how long it will take to freeze. Heating water will often lower the temperature to which it will supercool.

I have maintained samples of water at -15oC for 422 days.

 

[russ_watters]

I did exactly that experiment for PF a few years ago. The thread is still around somewhere...

 

[ZapperZ]

I recently came across an odd phenomenon known as Mpemba effect. It says that hot water freezes quicker than cool water provided that temperature difference is not too much. E.g. it can be, say, 30 degrees Celsius and 50 degrees Celsius but not something like 99 and 1. Unfortunately, the book did not say anything about why this happens. Can somebody tell me?

I believe there have been several threads that contains a number of references for this effect. Here they are again:

http://arxiv.org/abs/1003.3185 (J.D. Brownridge Am. J. Phys. v.79, p.78 (2011)).
http://arxiv.org/abs/0704.1381 (Physica A: Statistical Mechanics and its Applications
Volume 387, Issue 4, 1 February 2008, Pages 757-763)
http://arxiv.org/abs/physics/0604224 (American Journal of Physics -- January 2009 -- Volume 77, Issue 1, pp. 27)
http://arxiv.org/abs/physics/0512262

Zz.