Follow along: IG @birectifier
I was recently gifted a large box of gin botanical samples to explore some analysis ideas. The main agenda is pitting birectifier analysis against Clevenger analysis (employing a new generation Clevenger design) even though we know soxhlet extraction with organic solvents is the truly superior method. However, this last method just isn’t pragmatic for many people and takes many thousands of dollars of equipment so we are going to see what mileage we get from the first two.
The first botanicals I chose to look at were two different cinnamon samples and I quickly ran into the trouble because cinnamon essential oil has a specific gravity denser than water so my particular Clevenger design would not work. You can get a specialty model for oils denser than water and the punch line here is that it may be well worth it.
I proceeded with birectifier analysis and fractioned 40 grams of each botanical with 250 ml of 40% ABV neutral spirit (100 ml of absolute alcohol) and collected six fractions like I have done with other botanicals.
The aim was to see what-aroma-was-where and if analysis yielded insights better than smelling the samples right out of the bag (I could not differentiate anything right out of the bag!). The next aim was to gain an actionable insight. If one sample had more essential oil than the other, could I quantify it to scale the essential oil to maintain consistency in my production? And what would that look like? Would I be scaling things by 10% or only dodging major outliers? The product of Clevenger apparatus steam distillation is an oil collected in a graduated section of glass that has a scale. The scale can be used to quantify the oil amount, but what would happen with the birectifier where mulitple fractions are collected in 25ml volumetric flasks?
[Oil denser than water sinks to the bottom of the volumetric flask.]
Most botanicals I’ve surveyed had significant information in each fraction, but cinnamon has proved to be an outlier. Fractions 1-4 basically tell you nothing in each sample. Fractions 5 and 6 proved very interesting because the essential oil was insoluble and essentially sank to the bottom of the volumetric flasks. One sample also readily appeared to have more essential oil than the other.
What was learned is that fractioning cinnamon with ethanol as opposed to steam distillation, added no insights, unlike many other botanicals, so Clevenger distillation is more practical for cinnamon. And, because I needed to transfer the essential oil at the bottom of the flasks to a different kind of volumetric glassware, a proper Clevenger design would again come in handy. My solution was to allow the flasks to settle, decant off half of the aqueous fraction, then transfer the remainder to two identical test tubes with conical bottoms which were the only practical vessel I owned. My transfer technique cut corners (not enough rinsing) and was clearly not a best practice but I’m still confident I could differentiate the two samples.
One sample is clearly inferior to the other and probably is not fit to replace the other in a formula. But if you had to, the usage rate of CB3 may be more than twice that of CB1 to equate the essential oils.
There are many frustrations transferring these sample of essential oil because they are so globby and can stick to the sides of everything. Essential oil do not readily form a single mass. Part of the modern Clevenger protocol is adding xylene to bind the essential oil into a single mass. I’m not sure if xylene still works for essential oils denser than water. Xylene is not used in old Clevenger protocols and I’ve had terrible luck without it. You end up with a bunch of separated globs of oil and cannot make a reliable reading. Xylene helps but does not solve all of the problems. Xylene, itself can stick to the sides of glassware.
The issue with soxhlet extraction is that you need hexane and dichloromethane which are hazardous and regulated organic solvents. Commercial operators face the burden of solvent recovery and proper disposal which requires costly equipment besides the analytic balance needed to weigh the essential oil. Typically only 2.0 gram samples are used so the oil yield must be weighed to 0.0001 grams. There may be ways to reduce the organic solvent to only 30 ml per test (as opposed to 100 ml), but I haven’t quite figured out if that is really doable. Hiram Walker was using both soxhlet extraction and Clevenger distillation as early as 1937.
San Fran Cinnamon Bark CB1 (I do not know more about the samples than the labels I inherited.)
Fraction 1: Fairly neutral, slight angular character
Fraction 2: Fairly neutral, but a faint zesty character
Fraction 3: Fairly neutral, but a faint character.
Fraction 4: Turbidity on visual inspection. No droplets floating above or below. Aroma almost like a translucent slice of cinnamon. Not filled in with bassier character.
Fraction 5: Small oil droplets but the majority of oil was separated. Beautiful cinnamon aroma. Cinnamon piquancy on the palate, but not overwhelming.
Fraction 6: Far less intense than fraction 5, possibly because there is no ethanol to hold onto any essential oil. Slight piquancy on the palate but far less than fraction 5.
BSG Cinnamon Bark CB3
Fraction 1: Fairly neutral, possibly less angular character, no real distinction between two samples.
Fraction 2: Fairly neutral, less zesty character than CB1?
Fraction 3: Fairly neutral, but a faint character representative of cinnamon.
Fraction 4: Turbidity on visual inspection. No droplets floating above or below. Aroma almost like a translucent slice of cinnamon. Not filled in with bassier character. Very challenging to evaluate if CB1 or more or less concentrated than CB3.
Fraction 5: Small oil droplets but the majority of oil was separated. Visually turbid unlike the other sample. Not as representative of cinnamon as the other sample? Very different sensations on the palate. Less intensity and another secondary character. Reminds me of bad bubble gum. Not as beautiful as the other sample. Not sure why. White stuff starting to precipitate unlike the other sample.
Fraction 6: Far less intense than fraction 5, possibly because there is no ethanol to hold onto any essential oil. Slight piquancy on the palate but far less than fraction 5. No turbidity unlike the fraction 5. No white precipitate. Less intense than the other sample.
[I don’t understand the source of the turbidity but it was observed in the inferior sample. It slowly precipitated at the bottom.]