|MadSci Network: Physics|
Good question indeed. Although there seem to exist several opinions about this, I stick to the following: As the water gets heated, there are places near the heating elements (i.e. at the bottom of the pot or kettle) where the temperature rises above 100 degrees for an instant without the water starting to boil immediately. So there is a local boiling retardation (Siedeverzug) which gets `released' by production of tiny bubbles of water vapour that rise and almost instantly disappear because the temperature of the surrounding water is less than 100 degrees. If you have ever attended a chemistry class the teacher should have mentioned that when heating pure liquids (i.e. with no solid components) above a flame you should always have something in there like a glass stick or tiny glass balls that improve convection in order to prevent boiling retardation. The `release' of the retardation can be almost explosion-like, and that's what happens on a small scale when heating water in a pot or kettle.
It should be interesting to investigate what happens if one uses demineralized, pre-boiled water. I think the effect should be more pronounced there. On the other hand, slightly polluted water might give off less noise as boiling retardation is reduced.
Hope that helps, and viele Grüße von Erlangen nach
Try the links in the MadSci Library for more information on Physics.