Standardizing Botanicals: Me and My Soxhlet Extractor

Follow @b_apothecary

[2/20/16 I have made a ton of progress here, but I haven’t shared it with anybody!]

[This is just one post in hopefully a series about learning to standardize botanical charges for distillations most particularly gin, also aromatized wines, and bitters.]

Long ago I linked to a great paper called Controlling Gin Flavor from Herman Wilkie’s team at Hiram Walker in 1937. Wilkie is a very important distilling figure and it should be known is the true father of vacuum distilled alcoholic beverages. In the paper, back in 1937, Wilkie mentions a new era they had just entered where the botanical charge of a gin was scaled for oil yield. This acknowledges that the oil yield is inconsistent and if you just weigh your botanicals, you will end up with a less than consistent product. And sadly I suspect we have returned to the pre-Wilkie era which in my opinion is less than craft.

Gin production in the past has been characterized by lack of control over many of the important variables such as quality of spirits, quantity and quality of flavor in the various botanicals used, variable types and methods of operating the still, etc. Critical study of these variables disclosed valuable information which led to standardization of spirits and operations which, with proper selection of botanicals and regulation of the quantity of each ingredient used in the formula in accordance with its flavor value, now permits the production of gin under technical control which guarantees uniformity and quality of final product.

Controlling Gin Flavor

Wilkie notes that some distillery labs use the Clevenger Method of finding the oil yield which simply employs steam distillation while Hiram Walker uses a method, likely a Soxhlet extractor, with an ether as the solvent. The oil extracted is simply weighed then converted to a percent oil yield. What the paper doesn’t mention is how large their sample size is which is very important for what I aim to do.

No small producers to my knowledge are performing any of this analysis and these days it should be easier than ever with teaching resources like youtube, equipment procurement resources like ebay, and already purified chemicals affordably available from the likes of Fisher Scientific (but you need a commercial account and clearance to ship).

photo 5

To explore this type of analysis I bought a 500 mL Soxhlet extractor from ebay and already made some miss steps. Many Soxhlet extractors use a thimbleĀ to hold the botanicals and I bought one for $40 that I probably didn’t need. According to some youtube soxhlet demos, the bottom of the extractor can be lined with a simple bleached cotton pad and the botanicals simply tucked into a coffee filter. Its a much cheaper solution and even increases the volume the extractor can hold.

photo 2photo 1
The soxhlet extractor works by condensed solvent filling a chamber holding the botanicals until it reaches the level of a siphon tube eventually drains the chamber similar to flushing a toil. the drained solvent eventually evaporates refilling the chamber with fresh warm solvent. This means that the duration for running the apparatus can be considered in terms of flushes. Great advice is taken from here.

The amount of powder depends on the weight of the drug. If the powder is from roots or stem parts, it will be comparatively heavier than leaf powder. So heavier powder will be needed more as it will settle well in the extractor. What I mean to say is that the weight of the material is not a problem. It depends upon the size of the extractor you are using. Only thing is that it should be filled in extractor at least 1 inch below the siphon tube to avoid its entry there and finally in the flask.


So do not over fill the cavity.

Solvent should be filled from the top and not directly in the flask. Once you start filling the solvent you can see the drug getting wet and finally you will add it till the first cycle runs. Now you should add solvent which is sufficient to run at least two to three more cycles (from the top only to get initial efficient extraction). This way you will find that the drug is entrapping solvent for one cycle and flask is having sufficient solvent to run two to three more cycles. This is the normal practice. Regarding time for extraction, it is normally 24 hours or 72 cycles. But you can check for the completion of extraction when you see that the solvent coming through the siphon into the flask has become free of extracted material. For that you can use a watchglass. Just when the cycle is about to run, you need to take little (1-2ml ) of the solvent from the cycle in a watchglass and allow it to evaporate at room temperature. If you find a deposition in the watchglass, then it needs further extraction and vice versa.
Hope it will help you.


72 cycles (or flushes) seem like a long time but you can also refine your process by observing when the solvent starts to run clear. I think the 1-2 mL sample can be thiefed out of the extractor by reaching a pippette down through the condensor which is open (though you can’t really see it in my picks) then evaporating it. A microscope might aid in observing the residue. Once the amount of cycles are standardized, the time per cycle can be calculated and the total time taken from that.

As far as I am concerned, we use 10gm of power of plant materials for each 100 ml of solvent. For example, the solvent container that you used has a 500ml capacity means, we can pour 300ml and process 30gm of plant power (10gm per 100ml of solvent). In our lab, we will continue the extraction process up to the point, where the solvent color in the thimble becomes colorless as water.


So here is a best bet.

If we come to the point of solvent type, there is a custom to use three types of solvent, i.e. high polar, mid-polar and non-polar solvents. Some researcher uses any one solvent for each of the categories, however most of others, can decide a particular solvent, especially either from non-polar (such as hexane) or high polar (methanol, ethanol).

This something I haven’t completely figured out. I used hexane because its what I had. It is also less toxic than dichloromethane and waste disposal does become a consideration. In the end I lost about 50 grams of hexane (33.2 mL) which were stuck to the botanicals when I removed them from the extraction chamber.

photo 10photo 6These measurements with the scale are about 20 minutes apart.

I then tried to recover the hexane from the flat bottomed boiling flask.

photo 1


This is actually an early photo after probably one flush. Most often the low boiling point solvent is recovered with a rotovap which is known for speed and efficiency but I only had a high school quality vacuum distilling rig.

photo 9


Yet it was able to collect the hexane.

photo 8


Some how I only recovered 150 mL of my initial 300 mL of hexane, but I do know 33.2 mL was stuck to the botanicals and was lost to the atmosphere. Better systems could likely dramatically decrease the loss and inefficiency. Glycol instead of water to condense both rigs might be a good place to start.

photo 7

But what did I get? Pretty much nothing. My first test was run with wormwood which was likely a bad idea because the typical oil yield is so low (0.35%) where if experimenting with cloves they might have yielded over 10% and given a better feel for the process.

And what exactly is all that stuff and can it be thought of as oil? Should the contents of the flask have been filtered before it was vacuum distilled? We think of oil as volatile, so when we examine botanicals with very low oil yields but very high amounts of soluble non-volatile stuff like bitter alkaloids (not sure if I picked those up actually), should a different method be used like steam distillation?

It looks like somethings precipitated but are they still figured in the oil weight?

I also suspect a big problem I’m having is that I’m using old and tired botanicals who’s oil yields are not anywhere they should be and thus have no place in a gin. So I think I’m experimenting with some failures but there should be some value in there somewhere.

The next step is to try out my new glass steam distillation rig with clevenger oil separator. More to come. [2/20/16 I’ve also made a ton of progress with the clevenger apparatus and learned to differentiate its capabilities with the soxhlet. maybe I’ll sahre it later in the year.]

Follow @b_apothecary