It’s interesting you are using GC/MS. Most analytical testing labs utilize HPLC due to cost, sample prep, replicability without a mass spec, a couple of other reasons. On a side note, you may find you have a significant lab to lab variance because you are using a different instrument compared to the industry standard. Keep that in mind if you find clients are going to other labs and getting consistently different results. I found that even when using LC/MS/MS rather than HPLC sometimes yielded significant (more than 1-3%) lab to lab variance with potency.
Are you seeing the same peaks on an LC/MS/MS? If so, that will tell you a lot. Even if you aren’t seeing the same peaks due to the nature of mass spec MRM scanning, because you strongly believe these are cannabinoids if you can focus on the parent ion, and systematically eliminate the compounds you can identify easily, you may be able to work backward and figure out what compound the unknowns are. And if you can do that, the daughter fragments will further confirm your results. That may be tough to do since cannabinoids have such similar masses, but you may get lucky if you work through it systematically! Best of luck!
The original test is always HPLC-PDA, which is the chromatogram above. There is some coelution on there because we test 20 cannabinoids plus internal marker. The mass spec is used to achieve baseline resolution and confirm.
I was just told the CBG isomers are based on LCMS data and that the GCMS hasn’t been run yet.
Is this a concentrate sample? Looks like degradation products based on the low potency values. If it is a concentrate, then what does the sample smell like?
Ok they just may have gained a customer. That is awesome that they went that far to look into it. Shows passion in their work.
The rxn’s do some strange, funky stuff. I have to compile all my tests, but I pulled enough samples from a rxn yesterday to get a kind of clearer understanding what happens. Two big things I have noticed is that CBD morphs into some strange who knows what that fronts early to the point the peak will elude over CBC (on gc-FID CBC is first then CBD on our rig) which sucks because it skews the conversion efficiency percentage. Then the other observation is that d8 will increase and decrease over time. If I have the opportunity I would like to get samples every 5 to 10 minutes and run them, to do kind of a “stop motion” video to get some sort of visual of whats happening during the process. But like stated without knowing what the peaks are, it’ll just a pretty squiggly line that moves around.
And this is why I like testing with them. So many sellers get a “clean” COA they can sell with and refuse to send off to other labs. Look at how much detail they just explained in this. Idk any lab on the west coast that would do this.
I think I spoke with some of your colleagues on the phone recently.
We’re trying our best to help identify these peaks and give people the data they need to make their processes better.
Too many new clients use our services, see some THC they didn’t know they had, or see contaminants they’d rather not think were in there and then proceed to tell us that we’re wrong because some lab in CO has a ND for d9-THC with an LOD of 0.15%.
The clients that truly know how to use the analytics from our lab use the numbers to get better. Our clients make higher quality material, because they have better information.
As others have said, identifying from an HPLC chromatogram, without a standard, is very difficult if not impossible if they are unknowns.
I think the best way to handle this would be to run both LC/MS and GC/MS to get a full list of identified cannabinoids and molecules. Keep in mind that a lot of mass spec identification is done through pulling mass/ion fragments from various databases, so if these unidentified compounds are not in any database it’s going to give you an estimate based on the next closest match. Once you have the identity of what is in the actual sample then you can take a look at the actual chemical structure of all of the compounds and then based on your HPLC parameters you can start to label the unknown peaks based on chemical structures and theoretical retention times based on interactions between each molecular structure and the mobile phase/stationary phase.
As you can see this still includes a bit of guess work. If you wanted to have an exact identification then I would suggest the following:
Run the sample on HPLC and collect every peak as its own fraction.
Submit each individual peak fraction for LC/MS and GC/MS analysis
Confirm mass spec identity through a secondary source, usually NMR is the next tool in the identification arsenal when doing unknown chemical identification
From here you should have sufficient data from each individual peak that you can then label going forward so long as your process is consistent and is analyzed via the same methodology.
The chances of there being library spectra for compounds that don’t have a standard reference material available is pretty slim…
Why would you collect from an HPLC and run the fraction on an LC/MS? That is going to be a tiny amount of analyte (mainly mobile phase and trace amount of compound of interest). Wouldn’t it be a lot easier to use the same column and gradient you use on the HPLC on the LC/MS to get the same elution order and fairly similar retention times?
If you use mass spec coupled with NMR you would be able to get an exact structure of the compound…that’s the point of combining mass spec with NMR. To figure out yourself what the exact structure is.
Mass spec will give you all of the fragmentation ions and NMR (both hydrogen and carbon) will help you determine the exact molecular structure and you can combine that knowledge with the fragmentation patterns to build the structure.
