[David Cooper (OP)]

https://medium.com/the-physics-arxiv-blog/d8a2f611e853

The above is a blog entry linked to by the article below. The article below is lighter on detail, but contains a good graph which the blog entry lacks.

http://www.dailymail.co.uk/sciencetech/article-2483383/Mystery-hot-water-freezes-faster-cold-solved--strange-behaviour-atom-bonds.html

[David Cooper (OP)]

If you look at the graph on the Mail Online article, you can see that the heated water, after it has returned to the initial temperature, cools more slowly from there than the unheated water did, and it stays warmer too once it reaches zero, so that elevated temperature should allow faster loss of energy to the freezer and account for the faster freezing time. This change of energy in the covalent bonds must take a long time to recover to the normal level for the starting temperature when it gets back there (though in this experiment it isn't allowed to stop there at all), but the same lag in change of the bonds must prevent the water that started cold from giving up that energy once it's below zero, so it starts to supercool and is unable to maintain its temperature at around zero - the bonds aren't able to release the energy fast enough. With the heated water, heating them allowed these bonds to give up that stored energy more rapidly while they were hot, and they then cooled too quickly to take the energy back in.

[lightarrow]

David, I am not addressing you here, but all of those people who claims that "hot water freezes faster than colder one".
Whatever the "explanation" invoked, hot water can *never* freeze faster than colder one: when hot water starts to lower its temperature, and this requires some time t₁, it becomes "cold water" and this require a time interval t₂ to freeze.
t₁ + t₂ > t₂
unless they can go back in the past...

So the "effect" have, at least, to be called in a different way than "hot water freezes faster than colder one".
Or do we want to throw away centuries of physics just to make someone happy with its "discover"?

--
lightarrow

[David Cooper (OP)]

It's usually stated rather badly (as in the title of this thread where there's little room to play with), but clearly the hot water must return to being cold water before it freezes, so it's really a case of two different kinds of cold water freezing differently as a result of one of them being boiled/heated and cooled rapidly first (although if you look at the graph on the Mail Online page you'll see that one lot is put straight in the other freezer while the other lot is only just being put on to boil, but the boiled one will still freeze first).

I'd like to see more graphs to show various cases with both lots of water being boiled first but at different times to ensure that they've both had the same amount of disolved gas released from them, but I assume this has all been done and the difference in freezing speed remains. If this new theory is correct, what I find most interesting about it is how long it must take for the covalent bonds to become stretched and to take in energy, because they don't appear to have time to take energy in on the way back down from boiling once the water's put in the freezer, even though it takes many minutes.

[alancalverd]

It's a very well-known phenomenon but quite difficult to replicate in a domestic freezer. Erasto Mpemba gets recent credit for studying it in detail as a schoolboy and later pursuing a career in physics, but Charles Darwin (possibly a more reliable observer than the Daily Mail) reported the experiment a long time before Mpemba was born.

Darwin's classic trial was between two metal buckets of water placed outdoors on a cold, windy night. The one just taken from the fire, froze before the one previously at room temperature.

If you put a tiny thermistor into a vessel of cool water in a freezer, you can detect the onset of anomalous convection as it chills below 4 deg C. Colder water in contact with the surface of the vessel rises instead of sinking, so the rate of heat loss at the edge decreases (Newton's Law of Cooling) and the formation of an ice crust at the top and eventually round the sides reduces the rate of heat exchange between the hotter liquid molecules and the ambient air.  However if the starting temperature is somewhere in the region of 30 - 60 deg C  and the rate of cooling is sufficient (you really do need a metal container and a blast chiller to get a repeatable result) the normal convection flow has sufficient inertia to continue dragging the topmost layer down the side of the vessel below 4 deg C, so the entire mass of water can reach 0 deg C more quickly.     

[lightarrow]

Darwin's classic trial was between two metal buckets of water placed outdoors on a cold, windy night. The one just taken from the fire, froze before the one previously at room temperature.

If you put a tiny thermistor into a vessel of cool water in a freezer, you can detect the onset of anomalous convection as it chills below 4 deg C. Colder water in contact with the surface of the vessel rises instead of sinking, so the rate of heat loss at the edge decreases (Newton's Law of Cooling) and the formation of an ice crust at the top and eventually round the sides reduces the rate of heat exchange between the hotter liquid molecules and the ambient air.  However if the starting temperature is somewhere in the region of 30 - 60 deg C  and the rate of cooling is sufficient (you really do need a metal container and a blast chiller to get a repeatable result) the normal convection flow has sufficient inertia to continue dragging the topmost layer down the side of the vessel below 4 deg C, so the entire mass of water can reach 0 deg C more quickly.     

Ok but these are not the only effects. The hot water in the first bucket has not the same mass as the one in the other, it has less because it's at lower density, so he was actually comparing two different amounts of water. Furthermore, the hot water vaporizes quicker and when it has the same T of the other at the beginning, it has already lost some of it.

--
lightarrow

[alancalverd]

You only need to vaporise a tiny amount of water to shift a lot of heat. Latent heat of vaporisation = 760 cal/gram, so you could cool a litre of water by 0.76 degree for the loss of 1 cc. The density difference is also negligible compared with the difference in freezing rates.