Is this a reliable way to determine the alcohol %?

[LennoxLewis]

Alcohol (ethanol) has a freezing point of -114 degrees Celcius. Can this be used to determine the alcohol percentage in a drink?

Last week i put a glass of whiskey mixed with some water (about 1:1) and the surface was just frozen after i left it a night in the freezer, which was somewhere between -22 and -17 degrees Celcius (i didn't take an accurate reading because i just wanted it to be really cold).

So, given that whiskey has about 40% alcohol, the total theoretical percentage is about 20%, which means the freezing point of this "cocktail" is (0.2 * -114 + 0.8 * 0) = -23 degrees Celcius. Although not very practical because you're bound to a certain temperature range and it's hard to decrease the temperatur step-by-step, is this reasoning correct? The small experiment indicates so..





By the way, I've placed this one in the physics and not chemistry forum because it only deals with freezing points and volume ratios. Admin, feel free to move it if necessary.

 

[0xDEADBEEF]

Sorry, but this seems to be wrong. Binary mixes behave very "non linear" when it comes to phase transitions see http://en.wikipedia.org/wiki/Eutectic What you have discovered seems to be more of a phase separation, where some water freezes out leaving the rest of the liquid with a higher alcohol concentration.
Water alcohol mixes might be some special case though, where the "hanging part" of the transition line is very flat, and your idea could work, but then I would expect no phase separation but the whole mix to freeze solid.

 

[LennoxLewis]

I see. Is there an other physical (not chemical) way to determine the alcohol level? Counting the number of glasses that gets me smashed doesn't count!

 

[DaleSwanson]

The first thing that came to my mind was weighing it and using the known densities. Ethanol is 0.789 g/mL. This would be a problem though if there were other things besides ethanol and water. But I suppose you could find the density of pure whiskey yourself first, then use that to figure out how much of the mixture is added water.