We are working on developing simple GC methods for measuring cannabinoids in oils, butters etc.
Our current focus is to just do a simple extraction like we do for flower and concentrates and then modify the GC hardware to catch the oils on a pre-column which is then backflushed to reject the oils while allowing the cannabinoids to proceed onto the analytical column.
I would appreciate any guidance from others who have solved this problem in other perhaps better ways.
Hugh
SRI
I think you’re right on target.
my current procedures is to extract with etoh and double the run time. Pretty sure that’s @Dr_Jebril’s approach as well.
Back-flushing the big/slow stuff makes way more sense, and has been available on more expensive systems for some time now.
seems like a game changer to me (not a chemist).
Even doubling the run time is not enough to get the oil oil out of the column. With the .25 micron film column we formerly used you really had to get to 350C to get it to elute. With the new 1 micron film column we now have to use, it can’t really go hotter than 300C and the olive oil never elutes., hence our interest in finding a backflush method.
If you used something like 90-95% methanol as your solvent you could reject most of the fats out of the solution and just be testing primarily the cannabinoids. Might be a good idea to cut down some work off the bat.
Maybe make a 310c++ with a cleaning cycle built in with a small cleaning solvent reservoir 
push some heptane, or some kind of proprietary mix through the column to clean it after an edibles/tincture test.
fixed it for you…
In fact in my case, no. I simply dissolve the whole sample with acetone, filtered to 0.2 microns, and it is all injected then. It promote some parasites peaks, but I’m fine with it. I use various machines, one is the workhorse for any samples, and other are more devoted to “Clean analysis” were all the chromatograms matter…
Using ethanol or methanol makes two phases. A little portion of cannabinoids stay in the oil phase. There are various ways to correct this, but at then end I don’t find this very practical.
Total dissolution is more straightforward, but indeed can promote some mess in the GC…
The method I use for oils ends at higher temps than with flowers/extracts (320°C instead of 280°C) and stays there for five more minutes.
In case of MCT oil, is is enough to get ill all out. This oils can be clearly seen with the FID.
If it does not get all out, then it promotes visible mess in following injection, usually right were the cannabinoids show up…
In case of other oils such as olive, hemp, safflower, rapeseed, grapesseed… then what actually happens to most of the oil is still a bit elusive to me… I tried some long runs, but very few clear peaks are usually seen. A portion appears as a long tail in the middle of my method (between terps and cannabinoids), as a long tail. Some seems to promote a constant small background. Some will show up after the machine cools down, upon next start up…
The background and most of parasites peaks disappear after a few injections. If one want to get rid if them faster, injecting larger blank solvent volumes and increasing the injector temperature helps. (using various solvents for cleaning also helps, even using some MCT In some cases I believe)
Regarding analysis of tincture, after 1000s test, I found that this approach was convenient.
It does not really impairs routine tests. What I found more annoying, is how certain fresh oils seems to promote some artifacts (slight isomerisation ?) when preparing dilute tinctures with hemp extracts… 
Creating a solution in methanol and then introducing to a deep freezer <-20°C for ~2 hours, decanting the solution, and filtering down to 0.2µm gives relatively clean samples with no artifacts. (We dry down a sample of the MeOH solution and reconstitute in ethyl acetate before injection)
The issue with this method is validating the extraction efficiency and it may throw a wrench in routine testing due to sample prep time.
This paper has some good information for various sample preps and analysis for different matrices used in cannabis edibles. Tends to lean GC for many preps so it may be helpful.
Analysis of “Marijuana Edibles” – Food Products Containing Marijuana.pdf (196.8 KB)
That kind of cleaning works with HPLC but not with GC. In GC the cleaning solvent just floats by the contaminants on its way through the column. It does not wash it through. Are you sure methanol does not dissolve olive oil, especially when working with already low concentrations of olive oil. Typically we dissolve 100mg of olive oil ( containing cannabinoids ) in 40 ml of solvent ( could be methanol but we use acetone normally ).
Even if the methanol was less efficient at dissolving the olive oil, after repeated injections the olive oil ( mostly triolein I think ) would accumulate and require a new column ( expensive ) or some kind of service. We are trying to get a “bulletproof” method.
Good idea and we do that in extreme circumstances, but that requires removal of the column from the GC. It works about 50% of the time based on our experience. But we are looking for “bulletproof” and “drop dead easy” because most of our customers can’t remove and replace columns without causing other problems.
Acetonitrile might be an even better solvent for wax & fat rejection.
Otherwise water in your methanol also pushes out non-polar contaminants pretty readily
that Swagelok tee between the guard and primary column. shut off main column and use carrier to feed 50-100ml through guard column backwards into waste.
requires reservoir and a manual valve to select between liquid and gas backlash
?!?
I can see how that would work with HPLC, but with GC the residual solvent vapors would take days to dissipate out of the connecting tubing. So you would have a high but slowly decreasing ( drifting ) background. Think its better just to use the carrier gas to backflush the oils as part of the analysis.
Hi Hugh. From this post, it looks like you implemented a solution:
Do you have any information on that on your web site?
Hopefully there will be a document soon.
The additional hardware required to backflush a precolumn, plus the precolumn will add about $1500 to the current GC configuration cost. We hope to have a DIY kit, but it will require soldering so the GC might have to come back to the factory if you are not into soldering.