Head to Tail Drinking: Breaking up Distillations in the Rotovap

posted by Dave Arnold

A Thai basil distillation broken up into 27 pieces based on volatility

We’ve been working hard over the past four years to make the best possible hooch. I thought we were doing a top-notch job, but yesterday I discovered I was wrong.  That’s because yesterday we split a distillation run into 27 distinct parts based on volatility, nosed and tasted them all, and blended the best of them back together:  super-hooch.  This test was done with home-made Thai basil eau de vie, and it was truly excellent.

As many of our readers know, the rotary evaporator (rotovap) is perhaps my favorite piece of equipment.  It allows us to distill at very low temperatures, with very complete flavor recovery.  For example, the aforementioned eau de vie never went above room temperature.  For more on how a rotovap accomplishes these miracles, see the rotovap primer.

Our distillations are more akin to gin than to traditional distillation –they are re-distillations of already purified spirits, with the intention of capturing flavor. Traditional distillation’s primary purpose is to concentrate ethanol.  You start with a lower proof alcohol (such as whiskey mash, or fermented potatoes) with some impurities. When you distill, you concentrate these impurities as well as the ethanol. The first liquid to leave the still is referred to as the heads, and it is full of nasty stuff.  After the heads come the heart –the good stuff you want to drink.  Lastly, the tails – no good.  Too much heads or tails makes for popskull moonshine that can make you sick.  Distillation enthusiasts often ask me if I cut the heads and tails out of our rotovap distillate.  With re-distillation there is no need to remove heads and tails  for safety, though it is typical to remove them for taste.  At the school, we usually don’t separate the heads from the heart — I like the heads. But we do cut off the tails –we just stop the distillation when the flavor tapers off. Then  it occurred to me: What if some phases of the distillation process produce better flavors than others?  In  alcohol-concentration distillation the breakdown into heads heart and tails makes sense, but in flavor-based distillation there isn’t any reason to assume that the best flavors should be concentrated in the center of the run, or that undesirable flavors will only occur in the beginning or end. I pride myself on getting the maximum flavor out of my distillations, but what if some of the flavors I’ve been collecting are not so good?

Because different aroma and flavor compounds have different volatilities, they distill at different times — and so the flavor of the distillate changes  over the course of a run.  What if  I separated the entire distillation into small samples of slightly different volatility and checked their flavor individually? Most people with rotovaps can’t taste the distillate as they go.  Rotovaps are sealed systems under vacuum –that’s how they work. To taste the distillate, you have to stop the distillation, break the vacuum, pour off the distillate, and restart. Years ago, I addressed this limitation by building a pump system that connects to the output of my rotovap (see  picture below).  To this pump system I added a bunch of luer-lock stopcocks that allow me to divert the flow of distillate into any one of four small graduated cylinders, or into a larger collection beaker.

How we separate the distillation samples: The peristaltic pump pumps the distillate out of the vacuum into the manifold. The pump’s other –arguably more important-- function is to save flavor that would otherwise be lost –see the primer
Closeup of the manifold

Thai Basil Eau De Vie

We blended 750 mls of vodka with 82.5 grams of Thai basil leaves, put them into the rotovap and distilled with a bath temperature of 40°C and a condenser temperature of -17°C. In a rotovap, the product you’re distilling is a good 15-20°C cooler than the water bath, so the Thai basil was never heated above room temperature.  With the new stopcock system we easily separated the distillation into 20ml samples. We collected every 20 mls that came off the rotovap, then capped and labeled them.  Some samples were a little more, or a little less than 20 mls, so we recorded the actual number of mls of each sample, accurate to 0.5 ml. After we collected 542 mls of distillate (27 samples), we shut down the rotovap and got to tasting.

Tasting and mixing.

We checked every sample for alcohol content with a refractometer.  The first two samples read lower in alcohol than the second two did.  I don’t know if they were actually lower in alcohol, or if impurities made them read lower.   We smelled each sample in order, then tasted them.  We organized the samples into four groups:

  • Green: everybody liked these
  • Yellow: everyone thought these were OK
  • Orange: some people hated these, some thought they were OK
  • Red: everyone hated these

The samples were tasted by 6 people –we will do tastings with larger groups very soon. Here’s the tasting chart, with abbreviated notes:

As I had hypothesized, there were samples right in the heart of the distillation that we all hated –I had hoped there wouldn’t be; because diverting the distillate at precise times isn’t the easiest thing to do.

Blending spirits is a complex problem.  Sometimes, small additions of flavors you don’t like can actually make the drink as a whole taste better –counter-intuitive, but true.  We had to test whether removing the parts of the distillation run that we didn’t like improved the taste of the whole distillate. We did some blending and made four different eau de vies:

  • 3 mls from every vial in the green group: “Green”
  • 3 mls from every vial in the green group AND the yellow group: “Yellow”
  • 3 mls from every vial in the green group, the yellow group and the orange group: “Orange”
  • 3 mls from every vial: “Red”

I thought Green would be the hands-down winner, but three of the six people at the tasting preferred Yellow.  Everyone chose Green and Yellow as their favorite two liquors.  The people who chose Green had a strong preference for Green, while Yellow lovers tended to like both. Half the tasters preferred the Red liquor to the Orange one.  I was surprised that the Orange liquor would score lower on some people’s ratings than the red because the red samples were unanimously disliked on their own –goes to show the whole isn’t just the sum of its parts.  One fact is certain: selectively removing portions of the distillate that occur in the “heart” of the run can improve the flavor of the hooch you make.

We ran two more Thai basil distillations using the same amount of liquor and basil as the first test: one we did the old way, and one we did by diverting the distillate according to the recipe for “Yellow” above.  The one made like Yellow was better than the one made the old way, showing that the same flavors come out of the distillation at the same point in the run from batch to batch.

Now we have a problem:  how do we distill this way on a regular basis?  How are we going to get our yield up? Is there a way to re-use the red and orange part of the distillation, the way a pro distiller does?

18 thoughts on “Head to Tail Drinking: Breaking up Distillations in the Rotovap

  1. Interesting research. I’m curious about the transition from a “green” in sample 10 all the way to a “red” in sample 11. Any idea what’s going on there? Have you considered trying to refine your sample sizes in that region to see where the actual transition from good to bad is happening? Of course, reproducing it with enough fidelity for it to be a useful data point may be a bit touchy…

  2. The easiest way to increase yield would be to sample in smaller increments when you are approaching a red area. This will help you to isolate the compounds you don’t like.

    The easiest way to increase yield would be to build a vacuum distillation tower. But depending on the relative volatilities of the compounds your are trying to separate you may need a rather tall column. An actual column will also let you reflux and isolate each red/yellow zone to a specific temperature and pressure range. In the end this can provide a product of higher purity and much higher yields.

  3. Dave,
    could you please go and find an anlytical chemist to help you?
    Should be no big issue as most chemists are easy to bribe with booze (like most chefs LOL).

    You definitely need someone to run GCs on all those samples to see if and what kind a jumps in flavor contents there are, sounds extremely interesting to me.
    It would be even more interesting to also run a sniffing GC on those samples to find out if peaks that lets say suddenly appear in a red vial are actually foul in smell on their own or not….. (like pro wine people sometimes do see: http://www3.interscience.wiley.com/journal/113321779/abstract?CRETRY=1&SRETRY=0)
    Very interesting indeed….

    As a side note I think your fraction collector is severely over engineered: one tube that you move from flask to flask will do….. (Maybe with a syringe needle at its end to make it a bit heavier not to fall out of you flasks )
    If your flow rate is constant enough you could use a time trigger to change fractions either manually (by some sad intern) or with some smart mechanical device We have lots of fraction collectors that work like this in chemistry labs. Some move the needle some a train of flaks under a fixed needle.

    On the recycling of EtOH: Here you would want to do the opposite of what you do in the rotavap (in your case a dirty distillation “on purpose” similar to the way large pot stills for whiskey are very “imperfect” from an engineering point of view for purifying EtOH.)
    What you need now is a proper distillation column as long as you can get. This is to still the EtOH and leave everything else behind. Theoretically you can get 192 proof EtOH this way with some residual water but no taste to it. Re-dilute this with some quality water and you have something very close to Vodka. I know Russian chemists that do this all the time in their labs. (At Lomonossow University in Moscow chemistry is still fun, they smoke, eat and dink in the lab. LOL)

  4. A very interesting article. Especially since I had not read your rotavap primer. I spent a lot of time in a lab waiting for stuff to dry out in rotovaps: if I’d known I could have been handcrafting alcohol on the machine next to it while I waited (and legally, since I’m in New Zealand), it would have been more fun.

    As to today’s article, a couple of thoughts spring to mind. Firstly, the lower the temperature, the bigger the spread of time between different components. That might make it easier to separate what you want from what you don’t. Maybe don’t turn the bath on at all, or even cool it? Would slow the process down of course.

    Another question is: have you ever considered column chromatography? It’s used in the lab to separate similar compound (such as desired products and side-products or impurities). Basically you put everything in at the top and it travels through at different speeds depending on polarity. This might help you separate out the flavours you want from those you don’t. However, it is very slow, typically small scale, and in my experience uses unpleasant solvents (although there may be alternatives).

    1. Hi Alexander,
      There is an outfit called Cherry Instruments that does liquid-liquid chromatography with stuff like ethanol. They have approached us to do some stuff but we have been bad following up. Does sound interesting.

      1. Hi Alexander and Dave,

        column chromatography really doesn´t scale very well for productive use in a kitchen…
        The idea of using counter current chromatography is more charming, though. This is scalable (though prolly quite costly).
        Cherry is one of the experts in this filed.
        If I was in your position I´d definitely follow this invitaiton to do stuff with them. Maybe it is just a great learning experience, mabe even some productive stuff comes out of it. Good fun….

        1. Dave, I must say I am very excited to see this work.

          When we spoke on the phone the first time, I remember discussing the idea of breaking complex flavors into smaller pieces that can be recombined to create new and unique flavors. Very well done! It seems you have accomplished this!

          I believe our next go around with countercurrent chromatography will provide even more impressive results. I am looking forward to speaking with you in regards to this. I have some new tricks up my sleeve that should improve our method. And, thus far we’ve only touched the surface of it’s potential for this type of application.

          1. Hi Luke,
            Great to hear from you. Let’s talk soon. I’d like to push the application of countercurrent chromatography. Do you still think it is possible to make a kitchen scale/price version of the equipment?

  5. Hi Dave-

    I loved reading about this experiment; it has a lot of similarities to projects I’d like to do but haven’t had funding/foodsafe equipment for. I think you are exactly on the nose re: varying locations of tasty volatiles in a distillation for more than ethanol. I think that as you are taking cuts, you are getting not only different volatiles present but also different proportions between volatiles as they peak or recede in the distillation run, based on their affinity for ethanol, which is concentrated in the vapor. So flavor-wise, probably some of what you are getting in your cuts is perceptual differences based on aroma synergy, masking, etc. resulting from proportional changes in volatile composition. If you were using, for example, a gas chromatography column or a column still with many, many side draws and lots of reflux, you could isolate different compounds or groups of compounds and recombine them in whatever compositions you want to get your desired flavor effects…

    I was at Hangar One this weekend and one of the guys there mentioned doing some tastings with you of conventional vs. vacuum distilled products and finding that sometimes, the flavor in vacuum-distilled products tended to dissipate. Have you/do you find this? I’m in grad school at UC Davis looking at physical and analytical chemistry and sensory analysis of foods and beverages and am particularly interested in distilling so would love to talk more about it with you. My lab here does a lot of GC work which could be interesting to both of us. Every time I read one of your posts I am totally fascinated at the physiochemical mechanisms that might be behind what you find!

    1. Hi Arielle,
      I missed your comment. It was redirected to spam. I would love to do some work with you guys.

  6. This gives me occasion to wonder whether we’ll get another update on your DIY rotovap soon!

      1. But Wipop can. Just found out prototyping around here is dirt cheap. Anyway, don’t you have a shiny new rotovap now el Dave?

        1. Wipop!!!! We miss you! Dave DOES have a new shiny rotovap! Come back so you can help boil it over.

  7. Consider using a microcontroller paired with a flow sensor and actuation for the stopcock. preprogram it to separate your distillation. At its simplest it could sort the distillation into good and bad as long as you know that the first 20ml are good, the 2nd 20 are bad, etc.

  8. I would like to discuss distillation with you further if you don’t mind. I have a business where we are developing a new spirit and would like to begin a dialogue with you. Please let me know if you are interested. Thanks.

Comments are closed.