I can say that d8 separates nicely from d9-THC on this GC setup, though I admit it may be mixed with one of the other oddball peaks. Good call on the %d9…it did come in in the mid 50% ballpark by HPLC/DAD.
I agree also that here the main culprit is likely inadequate vacuum. If the pressure was always in the hundreds of millitorr, I’d bet the temperatures had to be pretty high to have the distillation go at all.
I don’t have wiped film material from this biomass, but I’m analyzed quite a number that I’ve distilled with WF, and kugelrohr (!), and they are all pretty close to the SBD result here. I should say that I’ve pretty much always had a diffusion pump operating when running cannabinoids, and I’m happy to report my usual pressures are in the 3-8x10-3 range. Distillation is pretty easy at 165° C for a first ‘post-terpene stripping’ run, and I can usually back off another 5-10° if I run a second pass (third distillation pass if you count the terpene stripping run),
I know nothing about analytical testing gear that you speak of.
However would it be possible to take a sample that tested clean and then divide it up a bit between a few test tubes and then elevate the temp in each test tube to different “mantle” temps from high to “oh my God the thing is smoking” hot? Then do the magic on your computer and test each test tube held at different temperatures and contrast. Perhaps on first run distillate anticipating and simulating temp conditions this way?
Shouldn’t rearrangement occur if temp is the issue and if so and tested in steps shouldn’t it be possible to predict when that happens this way?
It would be a good start, but there are more variables: time at at temp, packing material, pH, and absorptive media. Maybe a few more minor ones.
The real problem is not having a good assay. The method most labs are running for hplc/dad there is alot of co-eluting peaks, and I am sure the large d9 peak is masking these small isomers. I have seen some labs that don’t even test for d8. First step in my book is to devlope a hplc or gc assay. It maybe as simple as shooting a confirmatory GC injection to ensure your large peaks on hplc are pure. Drpaul, can you post the hplc trace the two samples you posted the GC for? I would like to see how they compare.
Time at temp is definitely worth exploring. We always run below .10 torr but we’ve had a couple batches with the same tell tale mystery peaks in the past. Making some adjustments in how we ramp up the mantle has prevented any further occurrences.
To shiny emulsion’s request, I didn’t run the HPLC on those oils, but I’ll see if I can get plots to post.
Experimental ‘production’ of these isomers is clearly a next step. Open test tubes would be pretty far from the conditions I were present when the problem run was done (lots of O2 present would be my first concern). Worth doing…there are so many spd systems out there now, it would be great to develop some guidelines to better prevent unwanted results.
The chromatograph you posted was fantastic. I have seen plenty of data for spinning band, but we don’t own or operate a short path so I haven’t seen that side by side before.
May I ask why GC was used here instead of HPLC? Is there an advantage for one over the other for cannabinoids? Also, do you know what the crude material tested like?
Just curious. Thanks
About how long per liter of crude in the boiling flask does a typical short path operator take to ramp the temperature up to the set point temperature for distillation?
I can’t speak for short path folks since we use a BR rig. However, it typically takes us about 45-60 minutes for a 1.5 liter run of crude to get to the point where the vapor temp starts to take off. BR systems use a Heat Rate vs. a temperature setting on the mantle as well so I couldn’t comment on what temp to use for short path. The way I like to explain the ramp up is using a kitchen analogy. When you boil water in the kitchen you can get there quicker by turning the heat on high until it starts to simmer then back off the heat to avoid it getting out of control. Using a similar approach on your mantle is a good idea.
I run a BR and have a consistent vac of 0.05tor. The oil looks great from a strictly qualitative stand point, but when I run it on my UPLC I get the same problematic “mystery” peak, between D9 and THCa, that many people are describing at about 8% w%/w.
I can confidently say that I ramp the temp of my mantle slow and there is no vac leak (aside from a potential leak at the cold trap, but it holds). What I cant confidently say is that my my filtering media does not make it into my boiling flask. I only use a small amount of DE, Silica, and a Celite bed, but I can almost always fine a small amount of fines in my roto flask - and very rarely in my boiling flask.
My two questions are this. Could any of these medias cause the metabolization of D9 to our THC imposter? and for the people using the Hocstrom filters and Summit filter papers, do you have any fines passing through? I think its important to note that my Hocstrom leaks like a bitch as well, I have to wrap the seal with electrical tape and seran wrap.
How slow is slow? We saw mystery peaks when we were ramping up slowly. Those went away when we sped up the ramp up time using the boiling water analogy I mentioned.
@nomadgt So our ramp takes about 3 hours to reach 180c on the pot probe, which is generally where we are at for our distillation. I have the heat rate at 27%.
Previously I was ramping at 40% and then slowing down, but I almost always was distilling at 200c at the pot probe.
Ill try to bring it to a boil quickly, and then ramp down. Do you have any recommendations on heat rate or temps on the BR?
@Soxhlet I will use that. Is this leaking a common issue with the hocstrom? I’m thinking of switching gears to a large scale filter press.
I can’t speak for the hochstrom, we don’t use filtration to remove our waxes. Ptfe is a more kosher material than electrical tape, which has adhesives.
That mystery peak is not calibrated? It sounds like Δ8, according to most common analytical chromatography for cannabinoids. You can buy Δ8 standard to calibrate the detector response factor, to give you accurate percentage.
Yes, the Hochstrom leaks unless you put a second thin but large O-ring around the base of the threads, so the top and bottom are then double-gasketed, which is standard practice in vacuum applications with dynamic meeting metal parts.
