thanks for the info! I must admit though, I just bought the GC and H2 generator. I’m 99% looking for d9THC and CBD concentrations. Tomorrow will order standards from Restek 3-in-1 of CBC, CBD and THC I believe.
I know this is not as accurate as HPLC, i have worked in pharma research with some beautiful units. FTIR will be dirty as all getup with the mix of cannabinoids. I looked at all the near IR stuff and the Photonics. Crazy expensive and it’s been a battle to squeeze out some money for the GC. You know the way it goes… they want a lab, but are shocked it costs money. I tease I work with good guys, just on a tight budget.
I don’t write off possible upgrades in the future though. This is a quick and dirty. This is a sanity check vs testing labs. Also this system will do terpenes with the standard kit from Restek and another column, also can do residual solvent with a third. All run through FID.
Now we get to my issue, the instrument I purchased did not have an FID detector on-board. I need to buy one of those, got anywhere cheap for that? I know I should have just got one with the FID, although a new one isn’t a terrible idea, but I did get the unit for $600 delivered
I’ve had quite a bit of experience with GC and GC/MS systems over the years, and had three of the SRI GCs, so I have a couple of comments.
First, HPLC with UV or UV/DAD will not be a great platform for terpene analysis, as these compounds are nearly transparent to UV, so no or very low signal. GC/FID is a much better candidate, maybe even better than GC/MS, in that the relative response to many terpenes will be similar (FID is sometimes referred to as a “carbon counting” detector, so a mole of a 5-carbon compound would have roughly half the response of a mole of a 10-carbon compound). Separating complex mixtures of terpenes is still a bit of a headache, but a long (>45 m x 0.32 mm ID?) 624 column could probably do it, and double as a column for residual solvents.
GC/MS has great identification power in general, but since its absolute response depends on ionization of compounds (and they differ in ease of ionization!), response to structurally similar compounds may yield quite different absolute signals.
Additionally, the mass spectra of several of the cannabis terpenes are so close that using library matching may not provide solid identification of peaks! Back to the collection of individual compounds for retention times…
Finally, I agree totally that an effort to identify the residuals present in high purity distillates is really needed. High potency distillates (>90% THC), including “water clear” preparations can still have several percent of minor cannabinoids and “other” compounds. Without better methods to identify these ballast materials, we’ll be stuck with guessing about the processes contributing to discoloring, odor problems, etc. Right now I’m collecting application notes on phospholipid, FAME (fatty acid methyl ester) analytical methods, and tuning the bejeezus out of my GC/MS to try some new techniques!
Here’s a great example of two samples, one exhibiting the red ring of death. The one on the left is over a year old (78.39% THC, 82.08% total cannabinoids) with no discoloration. On the right is a sample that discolored within hours after distillation (94.57% THC, 99.01% total cannabinoids; analyses by CW Analytical, Oakland, CA).
Awesome! I’m sure SRI has refurbished FID units. One tip though: It makes a big difference where the FID electrode is clipped to the outlet tube, so once you calibrate it, do not move it! You just need some terpenes to standardize the machine, and you will be ready to rock!
I agree that GC is more sensitive than HPLC for UV light-transparent volatiles like terpenes. Whatever they decompose into, it should never really change the FID response (unless terpene mixtures make a difference)… so yes, the response should be absolute, as you say. The 3rd-party analytical labs standard for detection of cannabis terpenes is mass spec from HPLC separation, which is also a type of ion detection, just a lot more specific than FID, since the ion spectra are used to both identify and also to quantify the compounds present. Just for cannabinoids, they mainly use UV DAD calibrated for a single wavelength for the chromatograms, but simulataneous detection of a fairly wide range of UV wavelengths occurs in the background, allowing each peak to be further analyzed with regard to its UV absorbance “fingerprint”. I know you already know all of this @drPaul , I am just stating it for other readers to follow the conversation. I sincerely thank you for coming in to clear up my easily-misconstrued statements for us!
Also, I did not know about FID’s carbon counting ability! That is a really interesting fact to keep in mind! Thank you!
Does that only follow for C-C bonds of the same type, and/or the same arrangement (e.g. rings vs chains)? For example, would a C=C=C=C chain give double the response for a single bond (C-C) and a triple bond (C=-C)? Also, would a C4 chain give about the same response as a C4 ring?
I think its based on voltage change when it burns. This is a huge assumption, but you would think that programming would account for bond energies. Realistically it’s just a retention time and the juice it gets from burning the bonds.
I hadn’t thought about a refurb unit. I was definitely on-board with a new unit but if the refurb is cheap enough I will grab it.
my assumption was that by using a very hot reducing (H2) flame, one could somehow steal an electron for every CO2 + H2O created in said flame. ie every Carbon. I admit that I have made no attempt to look at that assumption critically. it burns, and we measure electrons. it’s amazingly linear over a couple or 3 orders of magnitude. Calibrating with CBD works pretty well to quantitate THC. “carbon counting” sounds like a great description to me.
steal an electron? That would be a free radical. That would probably be Mass Spec-ish. It burns, we do not count electrons. We measure the temperature over time and the temperature causes a voltage change that then is interpreted as a curve and we integrate the area under the curve for concentration vs the calibration curve.
ions, not radicals. It’s guess it’s not a temp measurement it’s “Ions are detected using a metal collector which is biased with a high DC voltage. The current across this collector is thus proportional to the rate of ionisation which in turn depends upon the concentration of HC in the sample gas.” and “The measurement of ion per unit time make this a mass sensitive instrument.”
I learn or re-learn things as often as I forget other things.
It is a voltage change measurement though. Yes, I took some apart but not in the last decade
Well, a new one is great if you can afford it! I imagine a refurb would be fine also, since the mechanism is not overly complex and thereby not so sensitive to outright damage. The most common mode of failure (the main way people screw it up) is touching, abrasive cleaning, changing position of the clip or otherwise fucking with the metal tube ion “exhaust” electrode… and generally all these things do are the same thing: they throw off the calibration. So the fix is to set it up and clean it if you think it needs it (usually it does not, and it must be cleaned gently), and recalibrate. Then leave it alone!
“The simplest explanation of these observations is that every hydrocarbon is de-
graded to the same distribution of single-carbon radicals before ionization takes place. The visible light from an FID flame is due to emission at 431.5 nm from excited (A2A)CH radicals, and this emission also has the equal-per-carbon and linear response
of the ion production.”
You were and are correct, Sir! I was not in disagreement with anything you had said! Free ions are also known as free radicals. A lone free electron can also be considered a radical, I believe. “Ion” is a description of electrically charged matter, and to be more precise, cations are positively charged and anions are negatively charged. “Radical” is more a description of charged matter’s activity in relation to other matter. They are two sides of the same coin.
Ahh, this quote says much more clearly than my father’s day booze and dabs head is capable of:
A radical is a molecule or atom that has at least one unpaired electron (one electron is in an orbit by itself), but this moelcule or atom does not carry a charge like an ion because the number of orbiting electrons still matches the number of protons in the nucleus.; however it is very reactive.
But according to my read from earlier the GC first makes a radical and I think it was saying the conversion to ion occurs after. I’ve been diving down the rabbit hole all weekend on theory. Tomorrow it’s back to the salt mines and actual application of theory. It’s good, my brain feels full.
Damn skippy brother! 97.5% of my interactions in this site have been positive. For the internet that may be a world record
edit: We are not stealing, that would be wrong! We are re-arranging the molecules through flame ionization. Here is the simplest I could find:
CH + 0–> HCO*
The star indicates the radical. I’m not a organic chemist so at this point we are running up against the end of my knowledge. Hopefully someone else can drive the next question
One note on FID maintenance: if you’re running silylated derivatives, there is the possibility of silicon ‘plating’ out on the FID collector, forcing periodic cleaning. Just the price you pay (clean, recalibrate, drink coffee, run actual samples, clean…repeat) for trying to get the most out of your GC!
Actually, that’s probably good practice to change out the electrode after a bit if you run anything that can remotely corrode your column (like strong base, for example), because it is LINED with siloxane derivatives that can break down and do that same thing!
