Hi. I just had an extract tested with avazyme labs and it came back very different from the COA the seller was offering. From 5% to 1.9 CBG, 52% up to 67% CBD. I spoke with avazyme and they said that variations are common between labs and often between tests if the same sample due to unhomogenised samples. They also told me that round robin tests of the same extract sent to different labs sometimes had up to a 50% variability and a 5% variability within reliable samples was possible Where does this leave us in terms of reliable lab testing and knowing the cannabinoid content of extracts before we buy?
2 Likes
The lab is telling the truth, but also trying to play down their own incompetence. If the people running the analytical instrumentation know what they are doing and they get sent consistent/homogenized samples then everything should be pretty similar. However, many labs are either short of or entirely lacking competant analytical chemists and try to blame their customers rather than admit to their own faults. I’ve had extensive email threads with labs telling them where they might be going wrong (I’ve had over a year of post PhD analytical chemistry work so I am not as easy to be steamrolled) only to be first given some bullshit then silence
6 Likes
I’d slap myself if I got 5% variance of the same sample, however I’d accept that from another lab since I can’t hold some tech to the same standards (puns intended)
1 Like
Well my issue is whether the extract is 5% CBG or 1.9% or something else entirely and whether or not I can base a product line around thing CBG or I’d be another lying asshole in this industry.
Is there a lab that is beyond doubt in terms of integrity and quality? i hear good things about Steep Hill.
Do you work in a lab?
My product line is on hold until i can work out what the hell is actually in it.
@qga what test other than go or Hplc can be preformed to determine with accuracy cannabinoids content ?
Well if broken glassware is a phd analytical chemist he will be able to answer that better than myself being a knuckle dragging organic chemist phd lol.
The truth is, there is little regulation standards on laboratories in the US testing cannabinoids there are legit labs that are comparable to one another but it will take a few thousand to get this data to begin with.
I would take homogenous samples from a batch and send to 10 labs who are reputable and then compare which are the most similar results.
My goal of getting my batches third party testing is to be in the correct wheelhouse and be comparable to whichever lab my clients test with.
SOOOO the goal is not (unfortunately) not revealing the true cannabinoid content but rather getting the most accurate to match the rest of the labs throughout the United States for your clients. Many labs have similar results you just need to find them and not go bankrupt trying.
HPLC is the most accurate that comes to mind, GC could be used to confirm some results but main method HPLC should be most accurate @broken_glassware correct me if I am wrong.
5 Likes
Things for labs:
- Make sure they are ISO-17025 certified
- Make sure they participate in Proficiency Testing. The Emerald Test is a good inter-laboratory test that any reputable lab should participate in. Ask for their statistics from this test.
- Concentrate samples are usually weighed at very small amounts (10mg - 25mg), especially isolates, in order to be within the concentration range of their standard curve. Ask if they are taking a larger sample mass, and then diluting the solution before injecting into their instruments.
- Ask the lab how low their analytical balance/scale goes. They should be using something that at the very least has a resolution of 1/10th of a mg.
- Many labs won’t use any kind of confirmatory detection methods on their HPLC; a lot will typically utilize a single wavelength absorption detector. See if your lab has a DAD (diode array detector), or better yet a MS (mass spectrometer), for detection on the HPLC. This will help ensure that the proper molecule is being detected at the set retention time for the method, and not another molecule that absorbs at the same wavelength coming off the column at the same retention time.
6 Likes
If you are trying to measure Acid forms of the cannabinoids, then HPLC is the best way to go. In order to use GC to measure the acid forms, they must undergo some kind of derivative process to exploit the COOH functional group prior to injecting into the hot GC inlet port. Hot GC inlet = decarboxylation of your acid forms
1 Like
Yes conversion to the methyl ester via diazomethane to prevent decarboxation.
2 Likes
I spoke in detail with the owner of Avazyme, he mentioned many of the things you are discussing here and he was happy to talk at length for some time about things that went over my head. No hurry to get off the phone and not a hint of trying to defend or prove himself. I did make notes, I’ll have to dig them out.
The company who sent me the product want me to send 2ml back to them for testing at their lab again and to send some off for testing at a 3rd lab. He is as keen to work this out as I am and said he’d be happy to find a better lab of this shows his current one to be at fault. He’ll even pay for the new test and refund me the total cost of extract if the final outcome is off from is original claims. Can’t ask for more than that at this point.
Is there a lab in this industry that has a stellar reputation? If you wanted to know beyond a doubt what was in your extract what lab would you guys use? Steep Hill is a name that keeps coming up.
2 Likes
Perfectly said . Sad that is the case but it is so true .
1 Like
haha I only did analytical work kicking and screaming the whole way. I have a hypothesis about why there is a distinct lack of analytical chemists at west coast universities and canna companies. Analytical chem requires a extreme OCD for precision measurements and dilutions etc, however, west coasters are relaxed by nature, and cannabis related businesses also attract relaxed people, hence the terrible state of cannabinoid analytical chem services.
But your approach is fully sane. And I’d take an organic chemist to run distillation any day over any other chem discipline its an art so maybe the knuckle dragging pleases the distillation gods i dunno
3 Likes
I go to a school in the midwest and we have a top 10 analytical chem phd i tell people to start getting into this and it will be useful to them but they are all too lame i think not sure…
1 Like
yeah just have to be the kind of person to get off on teh >0.9999 R squared value on standard curve oh yeah
2 Likes
That would be me, but I’m not a chemist. I’m not an anythingist.
Maybe an antagonist on a good day.
3 Likes
Flora Research in Grants Pass, Oregon. Lab Director has 25 years as a forensic analytical chemist and is extremely generous with his time and knowledge. He has helped us twice with similar issues …
3 Likes
Comparing the accuracy of GC and HPLC tests is simple if you understand that with HPLC the THCA answer is multiplied by the theorectical conversion factor of .877 and then that number is added to the d9THC result to arrive at the " total smokable potency". The problem is that the real conversion rate from THCA to d9THC is about .68 because the reaction is never perfect. With GC, the answer is already the real “total smokable potency” number unless the sample is derivitized to prevent decarboxylation from occuring. In practice, if you submit a fresh flower sample to a HPLC lab you will get a number about 20% higher than the GC number. If you submit an already decarboxylated oil to the HPLC lab you get the same number as the GC test.
With GC, the derivitization step is much simpler than most people imagine. If adds about 2 minutes to the sample prep and about 10cents in cost. If you do the derivitization and then get an answer for both THCA and d9THC you are in the same situation as with HPLC. You have to then decide what conversion factor to use to arrive at the " total smokable potency.
With both GC and HPLC it is possible to get ± 5% precision, but this requires attention to detail. There is no fundamental reason why you can’t achieve comparable precision from either GC or HPLC.
The main difference is operating cost. The GC answer costs 15 cents, The HPLC answer costs anywhere from $5-15, a 100:1 difference. With GC you only have to calibrate on one molecule ( d9THC or CBD etc ). With HPLC every molecule requires two calibrations, one for the neutral and another for the acid form. The cost of the acid form calibration standards is 10 times the cost of the neutral form cal standards. And the acid form standards are unstable since they want to decarboxylate even if stored cold. I often suspect that a lot of the variation in lab results has to do with acid form cal standards which have partially decarboxylated and the operators don’t check for this. If an acid form THCA standard has decarboxylated only 5%, the calibration will be based on a lesser amount of THCA, so all the sample results will be 5% high.
7 Likes
Lab to lab variability is unfortunate but no amount of money or regulation is likely to help. Its just inherent at the current state of the technology. If you have your own GC or HPLC you can probably be accurate within ±10% on any given day with just one sample injected. If you inject 2-3 samples you can take the average and be closer to the truth, but that takes extra time the 3rd party lab can’t afford at the price they charge. If you have your own in-house GC or HPLC then you can decide to run 2-3 samples to be sure.
Hugh Goldsmith
SRI
5 Likes
Answers like this are why I come to Future4200.
I recently bought a kilo of crude. It’s COA said 62% CBD, my test at Avazyme ( a reliable lab that i use often) said 52%CBD. Both said 8% CBC. Funny thing is it is crystallising in the bottle at about 50% of the mass or more.
The guy I got it from made sure it was some of his best crude as an apology for a previous sample being wildly off the supplied COA. It’s great stuff, very smooth and potent, I just mixed up a batch at 2000mg/oz. It hits hard and clean.
How on earth is it crystallising if it’s only 50-60%?
think you nailed it on the head