dapper - 5-12-2007 at 18:48
Now, I attempted to quantify the amount of alcohol present in samples of alcoholic beverages and did my data analysis today. All of my values were
about where I expected based on the published alcohol content of the particular bottles printed alcohol contents. Based on the label of a bottle of
'Jack Daniels' I believed the % alcohol to be "46% by volume". Some searching revealed that this actually means %w/w which was helpful, as well as the
knowledge that all bottles of jack daniels switched to 40% in 2003 without the label changing. This bottle then I am surmising began at 40%, it then
sat around the lab for approx 2years (during which anything could have happened to it - I wouldn't rule out someone taking a sip and filling it up
with water, which is insane because it has been sitting in a lab) until it was measured. I am trying to figure out a way to explain the loss of
alcohol because I'm reasonably sure that I didn't screw up too bad because all of my standards were carefully prepared and the calibration curve
turned out alright. The concentration of alcohol in the sample falls within the linear range.. My only explanation beyond sample tampering is that
ethanol would have evaporated out of the bottle and reduced the concentration which seems possible to me, but I have no idea how to calculate this
rate. I have the pertinent physical data on hand.... All my problems would be solved if I could obtain a new bottle of JD and test that but there
won't be an opportunity.
If I had more of the lab sample JD I would test it with a hydrometer for a rough estimate. Could I instead determine a boiling point or freezing point
test for the mixture from these tables? http://en.wikipedia.org/wiki/Ethanol_%28data_page%29#Thermod...
Thanks
Dapper
dapper - 11-12-2007 at 16:08
Two words for me: experimental design. I should have thought of ways to eliminate these problems before they started. I'll shoot for better results
next time.