Expansion and Contraction of Wine in Containers

Nice experiment Joe. Bill Frazier Olathe, Kansas

Joe,

Great experiment. I am assuming that the cylinder you used was, well, cylindrical. Would not the volume of one inch of wine in the body of the carboy (say 10 inches in diameter) be significantly more than one inch of wine in the neck (say 1 inch in diameter)? Thus the change in height would be greatly multiplied in the neck of a properly topped carboy. All my carboys are filled at the moment. I will try some experiments in a gallon jug.

Sure, it's all relative (really bad pun). You are right about the relative changes and what you are thinking describes how glass thermometers and mercury barometers actually work. You have a large pocket of liquid attached to a relatively thin column on top. The relatively thin column shows the change in volume better and makes it easier to see.

The cylinder was cylindrical, it's made for general measurement of liquid volumes. There is a more accurate device called a volumetic flask that works on the principle you describe, but it's only calibrated at one point. I have one for 100 ml. That is too small to use here, the changes would be tough to see.

Joe

Not a bad experiment at all and not bat results considering the accuracy of the measurements and the range of temperature. Here is another approach:

Coefficient of expansion of Water @ 21 deg. C (70 deg F.) is 0.00021 Coefficient of expansion of Methanol (temp not given so STP assumed) is

0.00122

IF you can combine these properties linearly (a big if if someone wants to correct me) then the Coefficient of expansion of 14% wine will be 0.00034.

Now a 6 gal batch will be 22714 ml so the expansion with 1 deg C temp change would be 7.75 ml. or for 1 deg F temp change would be 4.31 ml. For a 5 deg change in your den this would be a whooping 21.5 ml. Now how much rise in fluid level that translates to in your carboy depends on how high up into the neck the carboy is filled and the inside diameter of the neck.neck.

Now these figures are for temperature changes around room temperature. Water coef. of expansion drops drastically at lower temperatures and methanol coef of expansion probably does as well. Note that the coef of expansion of water reverses itself below 39 deg F. So over the large change in temperature you used you would probably get a lower value than is observed at near room temperature. In fact you got 0.075 ml/l-deg F average over the range of 77 to 24 deg F. while it get 0.34 ml/l-deg F at temperatures near 70 deg F. Since most of us do not store our wine at temperatures that low, the numbers for 70 deg. F are probably more serviceable.

Does anyone know of any measurements on water methanol mixtures?

Ray

Hi Ray, You don't want to just add the values, you have 86% water and 14% alcohol in your sample wine, so factor that in.

I think you want to use ethyl, not methyl as your basis too. You want the CRC Handbook for those tables, any library will have one.

Glass changes with temperature too, that has to be factored in.

I'll run some numbers in Excel later and post, I have a copy of the CRC (Chemical Rubber Co. Handbook of Chemistry and Physics).

My wine did not change anywhere near that much as you suggest; a 5 gallon carboy dropping 50 F lost around 3 ounces, that's around 1.5 ml/F. But you bring up a good point, I went through 0C and water behaves pretty oddly there. That could explain your higher values, although I'm sure you need to touch them up a bit.

Water density is pretty odd, it actually reaches maximum density at ~4C, then turns around again. (Fish appreciate this, thankfully ice floats or they would have a bad day every winter...) It's the only substance that does this. I'll run the numbers over a range of temperatures, once I get the equations in it's plug and play... I have good temperature measurement equipment too, I may hook some up and start getting some empirical data too.

This is a little tricky as the relationships are not intuitive (to me at least).

Joe

I followed Ray's reasoning and tended to agree with him, so I did some calculations. It turns out the rate of change looks to be 5 times as much at the upper end of the temperatures wine is exposed to as at the lower end.

In short, forget about the 1.4ml/F idea, it's far from linear as Ray stated.

I would like to run some tests to prove these values out, but would be happy to email the spreadsheet this is based on to anyone interested, I tried to post an abbreviated version below. Maybe the tabs will be preserved, maybe not...

Joe

Here are some values:

Temp Ethyl Water Weight Glassware %ABV Density Dens F Density Density 1 US Gallon Correction @ ABV ml/l

32 0.80625 0.99987 3780.543 1.00048 10 0.98098 0.00 41 0.80207 0.99999 3781.090 1.00036 10 0.98055 0.43 50 0.79788 0.99973 3780.167 1.00024 10 0.97978 0.77 59 0.79367 0.99913 3777.962 1.00012 10 0.97870 1.08 68 0.78945 0.99823 3774.653 1.00000 10 0.97735 1.35 77 0.78522 0.99707 3770.340 0.99988 10 0.97577 1.58 86 0.78097 0.99567 3765.109 0.99975 10 0.97396 1.81 95 0.77671 0.99406 3759.050 0.99963 10 0.97197 1.99 32 0.80625 0.99987 3780.543 1.00048 11 0.97904 0.00 41 0.80207 0.99999 3781.090 1.00036 11 0.97857 0.47 50 0.79788 0.99973 3780.167 1.00024 11 0.97776 0.81 59 0.79367 0.99913 3777.962 1.00012 11 0.97665 1.11 68 0.78945 0.99823 3774.653 1.00000 11 0.97526 1.38 77 0.78522 0.99707 3770.340 0.99988 11 0.97365 1.61 86 0.78097 0.99567 3765.109 0.99975 11 0.97181 1.84 95 0.77671 0.99406 3759.050 0.99963 11 0.96979 2.02

32 0.80625 0.99987 3780.543 1.00048 12 0.97710 0.00

41 0.80207 0.99999 3781.090 1.00036 12 0.97659 0.51 50 0.79788 0.99973 3780.167 1.00024 12 0.97574 0.85 59 0.79367 0.99913 3777.962 1.00012 12 0.97459 1.15 68 0.78945 0.99823 3774.653 1.00000 12 0.97318 1.42 77 0.78522 0.99707 3770.340 0.99988 12 0.97153 1.64 86 0.78097 0.99567 3765.109 0.99975 12 0.96966 1.87 95 0.77671 0.99406 3759.050 0.99963 12 0.96762 2.04

32 0.80625 0.99987 3780.543 1.00048 13 0.97517 0.00

41 0.80207 0.99999 3781.090 1.00036 13 0.97461 0.56 50 0.79788 0.99973 3780.167 1.00024 13 0.97372 0.89 59 0.79367 0.99913 3777.962 1.00012 13 0.97254 1.19 68 0.78945 0.99823 3774.653 1.00000 13 0.97109 1.45 77 0.78522 0.99707 3770.340 0.99988 13 0.96941 1.68 86 0.78097 0.99567 3765.109 0.99975 13 0.96752 1.90 95 0.77671 0.99406 3759.050 0.99963 13 0.96545 2.07

32 0.80625 0.99987 3780.543 1.00048 14 0.97323 0.00

41 0.80207 0.99999 3781.090 1.00036 14 0.97263 0.60 50 0.79788 0.99973 3780.167 1.00024 14 0.97170 0.93 59 0.79367 0.99913 3777.962 1.00012 14 0.97048 1.22 68 0.78945 0.99823 3774.653 1.00000 14 0.96900 1.48 77 0.78522 0.99707 3770.340 0.99988 14 0.96729 1.71 86 0.78097 0.99567 3765.109 0.99975 14 0.96537 1.92 95 0.77671 0.99406 3759.050 0.99963 14 0.96327 2.10

Thanks Joe, I was doing what I did off the top of my head. I probably should have waited till I had access to my hand book at home.

Ray

I didn't think the format would be preserved, oh well.

The first column is temperature in F

2nd, density of ethyl alcohol at that temp 3rd, density of water at that temp 4th, weight of 1 US gallon at that temp 5th, coefficient of thermal expansion compensation for glass 6th, %ABV 7th, calculatated density of 1 liter at that temp ((%ABV * Alc Dens)+(remaining water * water dens)* glass expansion/l) 8th, change on ml from one step to the next due to changes on density and volume of the container; In other words, using the first group for an example, how much change you should expect to see in a 10% ABV solution for a 5 degree C (9F) change.

These are calculated values, I want to test them out too. They do not account for disolved solids, that is around 2% of the total on average from memeory.

Joe

Joe, as I said off-list, I would perform the same experiment with a

12% alcohol/water mixture in a graduated liter cylinder. The graduations were not fine enough to allow milliliter measures and so I "borrowed" a 250-mL cylinder from work. I began this before you posted your table.

My mixture was constituted at 78 degrees F and chilled to 50 degrees. Like you, my measurements were of contraction. I will not go into details here except to say they agree with your table in that the contraction is as predicted therein (to the extent I could approximate tenths of a milliliter), but they differ in that I began my 250 milliliters at 78 degrees.

Here is the problem I see with all of this and it is one I mentioned in my original post to you that prompted your initial experiment. The CRC tables all begin with the density of water being 1.000 at 4 degrees C (39.2 F), whereas we all constitute our must and the resulting wine at a much higher ambient temperature. Our mixtures have already expanded due to elevated temperature and so we begin with X liters of wine under airlock at, say, 68 degrees F and then the temperature rises. The volumetric expansion from 39.2 F, which your table predicts, is meaningless to us. We need the expansion from 68 F. It turns out one can calculate this from your table, but it isn't all that straightforward.

Any idea how to more easily work the problem we face in real life?

Now, having said all of that and asked my question, I still say the solution is not to be surprised by a volume increase due to rising temperature. Simply look at your carboys daily and when the wine approaches the airlock remove the airlock and then some of the wine. That is another use of a wine thief.

Jack Keller, The Winemaking Home Page

Hi Jack, I follow your logic, but the last column on that table was actually the difference between points of 5 C, not from maximum density. It's way too coarse at 5C steps so I started to make one in 1 C steps, then I decided I had better think this through.

I am actually wondering if any of this is valid since both the container and must are changing with temperature. I know the general coefficient of thermal expansion for glass, I may plug that in and run a few numbers after thinking it through.

Joe

... The

Got'cha, but just remember that the CRC numbers for glass are for borosilicate glassware (Pyrex. Also, there are major differences between CRC's F-3 and D-146 tables. The D-146 table looks better to me for our purposes, and the 1-degree C graduation is built in....

Ray, where are you on all of this?

Jack Keller, The Winemaking Home Page