January 2010


I have waited to try this experiment for six months! Too bad when I first saw the video in YouTube I didn’t realize that one needs really cold temperatures for it too work. Despite the freezing days here in Europe in the past 2 weeks, it has not been cold enough for it to work in Edinburgh :(. Perhaps where you are it is cold enough (I hear it should be at least 0F, – 15ºC) so that you can try it and then send us a video of it!

For the time being let me just show you this YouTube video. The experiment is very simple: take a cup of boiling water, throw it up in very cold air and watch it freeze instantly and convert into snow!

The amazing thing is that, if you try it with cold water or just water that is not boiling, it won’t work. There seem to be two reasons for this. The first one is that boiling water may disperse faster in the air breaking into very small drops that freeze faster. The second one is the Mpemba Effect, which says that, under certain circumstances, boiling water freezes faster than cold water.

I find this effect fascinating, because it is so counter intuitive. Say we have two samples of the same amount of water inside 2 twin containers: one with boiling water and the other one with water at 25ºC. When put in the freezer, the boiling water has to first reach 25ºC and then freeze while the other recipient only has to do the second step. Clearly then, the second recipient will freeze faster than the one with boiling water.

This is so when both recipients are exactly in the same conditions but it turns out that they are not. First of all, the boiling water is loosing some water through evaporation at the beginning so, even if initially we put in the same amount of water, after a while it will have less water than its twin pot. Secondly the high temperature of boiling water could change the properties of the material that it’s in when heating it up: it may become a better conductor of heat and accelerate the cooling process. Also the convention currents in both pots will be different, presumably increasing the cooling speed in the boiling water one. Finally there are experiments that show that supercooling is favored in cold water versus hot water. Supercooling is an effect that occurs when water does not freeze at 0ºC but at a few degrees below 0.

Each of this effects are very small and separately could not cause an observable difference in the freezing time. However in some circumstances, the combination of all of them together seems to conspire in such a way that the boiling water freezes faster.

I found trying this experiment in my freezer quite hard because I think that opening and closing the door to check the samples influences it a lot. Then I tried doing it outside on a cold day, but it wasn’t cold enough and finally my Chemist friends tried doing it with dry ice and other fancy gadgets. Long story short: we have not been able to reproduce this effect.

It only happens for some very special initial conditions and, despite having played with the variables around, I haven’t been able to find the right set. In particular I would need to know how much water one needs to put in each container, what kind of material should one use and what should be the initial temperature of the non-boiling water. I found this video in You Tube where they manage to do it in their freezer:

I did try the exact same method and it didn’t work: perhaps my freezer does not reach the right temperature…

The challenge for you now is to try this at home and find the ideal set of initial conditions for it to work.

A colleague of mine has introduced me to this problem which combines mathematics and food. The challenge, as stated by George Hart, is to find three foods that do not go together well but every pair of them does.

In his statement of the problem, George Hart sais that they should not go well together “by any reasonable definition of foods going together”. But I think that this is so subjective that it would be almost impossible to agree on a reasonable definition for this. For instance, I have a bet with another friend in which I claim that if I like a given food alone then there is a way in which I can combine it with chocolate and still like it. However I realize that this wouldn’t mean that other people who are not chocoholics, would like the mix…

In fact, it is quite difficult for me to think of 3 foods that wouldn’t go together under any circumstance. So far, for every combination suggested in George Hart’s website,  I can find a way of cooking them or picking the right proportions such that the combination would taste good to me. But I guess some would say that I would just eat anything…

In that sense, it is even more difficult for me to think about the complementary problem: find 3 foods that go great together but such that every pair of them does not.

Even though I believe it is probably impossible to solve this because of difficulty in agreeing on what goes together well and what doesn’t, it is still a good conversation starter for a lazy afternoon and if you do come up with some ideas please post them in the comments of this post. I’d be happy to accept the challenge of finding a recipe that combines all 3 ingredients and eating it :).