Friday, January 09, 2009

Evaporation Rates of Ethanol Solutions

A few years ago, a crazy post-doc posted the following sign at the bacterial bench:

"Do not use 70% ethanol to sterilize the wand, it does not burn completely. Better yet, 70% becomes 50%, becomes 30%... WATER DOESN'T BURN!!"
More recently in a comment, Kevin Z muses, "I don't really see how it matters to make fresh ethanol dilution or keep a 85% stock around." Being graduate students, and having totally forgotten basic high school science, this became a question: Is this true, does ethanol in solution evaporate preferentially over time? (Of course, being graduate students, it took 2 years to come up with and test the question).

There were two schools of thought on the possible answer:
1) A mixture of water and ethanol will form a perfect solution with its own unique properties (different percentages will have different evaporation rates but the ethanol percentage will remain constant over time)
2) Ethanol, being more volatile than water, will preferentially leave solution resulting in a decrease in ethanol content over time.

To answer the question, a series of ethanol dilutions ranging from 0 to 100% were prepared and left open on the benchtop in 15 mL tubes (10mL of each dilution). The volume was recorded every day for a period of 1 week and plotted.
















Figure 1: Change in volume of open ethanol solutions over a period of 1 week. Slopes represent evaporation rate (mL/hour).

As Figure 1 demonstrates, each solution has a unique evaporation rate and slopes appear roughly linear. However, closer examination reveals that the graphs are not linear but rather the slope of each curve changes subtly over time. That is, for example, the calculated slope of the line over the first 3 days differs from that over the last 3 days (data not shown). This change in evaporation rate over time suggests that the ethanol content of each solution is changing, in line with view (2).

Because the changes in slope were subtle, a second metric was chosen to verify that the ethanol content of each solution was changing over time. For this, the mass of one mL of each solution was measured at the end of the week and compared to the density of solutions of known composition.














Figure 2: Density of various ethanol solutions. Diamonds indicate the density of freshly prepared solutions at the indicated ethanol percentages. Bars indicate the measured density after one week on the benchtop.

As can be seen in Figure 2, the density of ethanol solutions does, in fact, change over time - indicated by the difference between the points and the height of each bar. The difference is most noticeable at mid-range mixtures (~30 - 50% EtOH). Final ethanol percentages after one week were calculated based on the known standards. (I realize the x-axis doesn't scale correctly [eg. the space between 0 and 10 is the same as between 10 and 30], this was a quirk of Excel that I couldn't figure out in a short amount of time. The standard curve and R2 were determined on a separate graph with proper scaling [not shown] using the same data) The summary can be seen in the table below.

Table 1: Summary of Ethanol Densities
% EtOH in solutionDensity (g/mL)Density (after 1 week)Estimated final % EtOH
01.0010.99514.304
100.97190.980210.78
300.94070.969515.43
500.9010.937429.39
700.8560.874556.73
900.79340.80586.95
1000.7610.766103.9


Long story short, the postdoc in question was correct. After a week on the bench, 70% ethanol used to sterilize bacteria spreading wands will be a shadow of its former self. Regarding Kevin Z's comment, if you're in the habit of leaving 85% EtOH open on the bench, then by all means you should be making it fresh - but I suspect that's not the case. This carries over to other areas in the lab as well. Oft times, the people filling TC spray bottles with 70% ethanol for sterilizing work surfaces (maybe this is where the confusion at the bacterial bench comes from?) will prepare too much and leave the excess in a graduated cylindar on the bench - sometimes covered, sometimes not - to fill bottles at a later time. This, again, will lead to evaporation of ethanol and spray bottles filled with less than 70%.

This could have been figured out in a much simpler way. Water and ethanol form an azeotrope - a mixture whose composition can't be changed by distillation - at a ratio of 95.6 : 4.4 ethanol : water. In this mixture, the vapour has the same constituent ratio as the solution. Other ratios will have one component evaporating off at a different rate than the other. Which is why, of course, we're able to distill spirits to increase their alcohol content. Check out the above link to read more about azeotropes, phase diagrams and Raoult's Law.

The real lesson, though, is to put the cap back on your Jameson's when you're not drinking it.


13 comments:

Anonymous Coward said...

has hell frozen over?

rob said...

While this reviewer finds the work worthy of publication on the Bayblab, my main concern is with the lack of statistical analysis and low n. I am also concerned about originality of this work as I think I learned this in undergrad.
Mostly, however, I'm just disappointed I can't be an annoying know-it-all to the post-doc in question.
Good work on finally getting this work out. I know that the scientific community needs to know that these sorts of ground breaking experiments are taking place.
BAYBLAB EXCLUSIVE.
If anyone else wants to publish original work on the bayblab, let us know. We also take negative results [obviously]

Bayman said...

I always follow my nose when it comes to lab chemicals. Never trust an alcohol source until you sniff it first. Mine is good to +/- 5%. Have to say I prefer the smell of isopropanol...actually come to think of it, butnaol is even better.

bets said...

Hey, whoever did this experiment-
Table 1 also means that 100% ethanol becomes 103% after a week and plain water (0%) develops ethanol after a week (4%). That means we can make ethanol from water! Tell the biofuel guys to stop research and concentrate on this.

Where are the grad students getting to? I am also a grad student!

Kamel said...

I ran the experiment, and I am aware of that 'quirk' of the data. (I typed up the table after all!)

I decided not to force the numbers to 0% and 100% even though there was no ethanol (to my knowledge) added since those were the actual measured numbers. Chalk it up to pipetting error, error with the scale, environmental differences (temp/pressure) a week later or what have you. Of course the numbers are there so you can rejig the graph if you want to force it through 0 and 100. And if anybody wants all the raw data, I'm happy to send it.

simonex2609 said...

A Good piece of information.

Antabuse Generic said...

very welldone

Rodney Benner said...

Wow. I left my beer overnight in the fridge, uncovered, and I wanted to know if I could still get a buzz from it in the morning.

It's news to me that people use ethyl alcohol for more than just imbibing. What'll they think of next... like using DNA molecules to treat my cirrhosis?

Thanks science!

Anonymous said...

Does capping the solution completely stop the ethanol from evaporating or just slow the process down to a negligible degree?

If evaporation still occurs, how long would a capped 70% solution take to turn into water? A few decades?

Tuulia said...

To get quite absolute EtOH-% in each solution you could have done that easily with NMR. I'm sure you have contacts to some NMR groups (or at least you should have if you work with chemistry related field).

Anonymous said...

This experiment is interesting, but without replicates from which to calculate standard deviation/error how can you say your results are significant? I think the variation you're seeing could easily be due to your experimental error, especially given the fact that your largest difference is only around 0.04 g/ml. Without error calculation, it's difficult to say that these data mean anything at all. Just sayin....

Rob said...

@ Anon 10:05
A very legitimate criticism of the experiment. I am reasonable confident that the variation was not due to experimental error but this hasn't been demonstrated at all. This was a pretty rough experiment but I think the conclusions are justified. Thanks for reading and your comment!

drug possession attorney Miami said...

I recently discovered your blog and still have high hope for this to carry on. Keep up the good perform, their nearly impossible to find beneficial ones. We have added to our faves. As well as the concept of Many thanks. For more info visit on drug possession attorney Miami