MadSci Network: Physics |
Where did you get the idea that melting point decreases in a vacuum? Boiling point certainly does, but melting point is something different. Melting happens when the temperature gets high enough that the motion of the molecules overcomes packing into a solid structure, so the liquid becomes more stable. The pressure has only a small effect on that. I can think of 4 ways a vacuum *might* seem to affect a melting point: 1) The pressure does have a *very small* effect on the solid/liquid equilibrium. For water, going from atmospheric pressure to near a vacuum *raises* the melting temperature by about 0.01 K. For most other substances (where the solid is more dense than the liquid) the effect is in the reverse direction (so that vacuum will lower the melting point), but the effect is still small, at least for the small pressure difference between atmospheric and vacuum. For the explanation of this effect, look in a physical chemistry textbook under "Clapeyron equation." 2) For some substances (especially metals), air can react with the solid and form a compound that will have a different melting point than the pure solid. This could affect the melting point in either direction, and doing it in a vacuum would get rid of this effect. Some high-temperature work with metals is done in a vacuum for this reason. 3) Solids can also "boil" since they have a vapor pressure (this is called sublimation). So, for example, if you make vacuum above some ice, it will eventually sublime and go away (like snow can evaporate on a dry sunny day even if the temperature is below freezing). But this is not melting. 4) For some compounds, if you heat them they will decompose before they melt. If you are under vacuum, the products of the decomposition will get carried away faster, so the substance will disappear faster. But this is not true melting either. About your other question, the prediction of melting points is in general extremely difficult, even more so than prediction of other properties. This is because it depends strongly on how the molecule can pack into a regular lattice arrangement, and that is very difficult to predict (except for simple molecules like argon). People have put in a lot of research to come up with schemes to predict the melting point based on molecular structure, but frankly none of these is very reliable (though it can work for a series of very similar compounds -- you can predict long alkanes based on shorter alkanes for example). If people have an important material they need to know the melting point of, measuring it experimentally is still the way to go. Measuring phase diagrams is still an important (and underappreciated) part of science.
Try the links in the MadSci Library for more information on Physics.