Dear  Xi Zhang, Yongli Huang, Zengsheng Ma and Chang Q Sun
Rupert Harper [info@harperfamily.org.uk]

Congratulations on establishing a MUCH more plausible explanation of the
Mpemba effect than any previously published. I entered the Royal Society
of Chemisty's competion and was very dissapointed with the winners
explanation which didn't address the peculiar properties of water as you
have done. Well done!!

Your ideas and mine overlap although you are, obviously, much better
qualified than I. Here is my explanation which might be of value to you.
You are welcome to pursue my ideas, especially experimentally, but a
small acknowledgement would be welcome if you publish! You might have a
go at the simple experiment I propose in the last sentence.

It is well known that water is a polar molecule and that it is a liquid
at room temperature because of the hydrogen bonds it forms. Because of
its liquid nature it is reasonable to assume that, at any one time, less
than the maximum number of hydrogen bonds are extant. In simple terms,
if you stir a glass of water a very large number of the bonds must be
broken. My hypothesis, although it would be impossible to prove, is that
water at rest is far more viscous than water that has been agitated. Of
course, any attempt to measure the viscosity must agitate the water
hence the impossibility of proof!

However, water that has been heated to a higher temperature would be
less viscous on account of the kinetic energy of the water molecules.

My further hypothesis is thus. Water at 100 degrees is less viscous but
as soon as it is placed at rest a very large number of hydrogen bonds
can form very quickly on account of the low viscosity. In other words,
in a few seconds, it forms the necessary hexagonal crystalline lattice
arrangement of the molecules without actually freezing. I can only
assume that the kinetic energy of the molecules is converted into pure
heat. Due to the elevated temperature this heat is quickly dissipated
either by radiation or conduction, both of these processes being
accelerated by higher temperature differences and gradients respectively.

However water at 35 degrees is more viscous and so, when placed at rest,
the molecules find it much harder to arrange themselves into the right
hexagonal lattice arrangement. Furthermore, as the water cools further
viscosity increases and the molecules take longer and longer to arrange
themselves; particularly when ice actually begins to form which retards
the process further. So even when the water is at 0 degrees the
molecules have residual kinetic energy and, as they form hydrogen bonds,
they will release heat which slows the cooling process further. Contrast
this with the hot water which has already arranged itself into the right
structure and dissipated its surplus molecular kinetic energy as heat
very rapidly early in the cooling process.

My hypothesis could be disproved by showing that the Mpemba effect
happens if the water is vigourously stirred whilst being cooled.
Likewise, if stirring mitigates the effect I can upgrade my hypothesis
to a theory! Unfortunately I don't have laboratory facilities at my
disposal to perform this experiment.

--
Rupert Harper,
B.Sc.(Eng), C.Eng., C.I.T.P., F.B.C.S., A.C.G.I.
+44 1903 871401

If you can see the answer then you are a visionary.
If you can draw it then you are an artist.
If you can build it then you are a sculptor.
If you can explain how it works then you are a scientist.
If you can do the calculations then you are a mathematician.
If you can see it, draw it, build it, explain it, do the calculations and make it work then you are an Engineer.


(c) Rupert Harper, 2005, 2006 - may be freely distributed on condition that original authorship and copyright are acknowledged