HPLC SOP for cannabinoid assay on distillate

Testing and Analytics
1

Hey there,

I’m wondering if anyone has an SOP to share for cannabinoid assays on distillate utilising HPLC.
Our solvents are methanol/TFA and water on a gradient which produces fine results for flower.
Currently tinkering on a method for distillate, mostly interested in THC/THCA/CBD/CBDA. Obviously there will be hardly any acids due to distillation, and mostly I’m interested in THC.
We have standards for all, of course.

I’ve reached a point of 140mg of sample in a 1000ml dilution, but results have been a bit fucky and thought I’d reach out to see if anyone would be generous enough to either troubleshoot or provide an SOP.

For example, we recieved results on distillate at 72% total THC. I did a second pass on short path and utilising the above method it was shown to be 70.5% or 76% total THC. Probably sampling error, but I would have assumed to to be at least a bit higher.

3 Likes
2

Your column vendor may have a method published, I’d start there.

There is zero difference in the HPLC method for us between distillate, flower, isolate, and whatthefuckisthis.

Sample prep is another story. We’ve developed a gravimetric two-dilution prep that uses different quantities depending on the expected cannabinoid content of the sample. It’s rather simple now, though getting there wasn’t.

7 Likes
3

I hope you are not really dissolving 140 mg in 1000 mL… are you aiming for 140mg/L in a stepwise manner? How do you usually proceed?
Sample prep is the first critical step.

You should look into series of open access papers focused on cannabinoid HPLC analysis published in the last 10 years, which in general provide more or less detailed SOPs. Restek’s blog should also be of great help.

If you aim for THC and CBD, you should still include at least delta8-THC, CBN, CBG and CBC, which are also generally there.

3 Likes
4

Why are you using EtOH for mobile phase? I’d recommend ACN or MeOH due to costs alone.

Note that elution order will change when swapping from ACN to MeOH

2 Likes
5

Good morning everyone. Thank you for your responses, I’ll address them step-wise to keep the thread tidy.

A simple yet obvious solution, I will reach out to them. As far as I know they only have flower methods. There is currently no Australian pharmacopoeiac standard for cannabinoids although I believe the regulatory bodies are working on it.

No, we do dilutions.

Wonderful, thank you for the resource.

As I said, we only have standards for THC/CBD(+A) and at the moment our HPLC is utilised as a general indicator of potency for flower. We have to send them off to an external lab for validation before we send them to market (pricey, as you can imagine). We’ve hit a point where our flower results are within a pretty tight range of what our external lab sends back to us.

Developing this distillate method will save the money and turnaround time it takes for a cannabinoid analysis of our distillate and help me develop the short path method itself. At the current time, the minor cannabinoids hold no real interest for us other than curiousity, although that may change as our market develops. Our column provider has already offered (at a price) a more robust HPLC method to include the minors, but for now we’re sticking with this. Lots of spinning plates here and it isn’t urgent, but thank you for the suggestion.

My apologies, we’re using methanol - not ethanol. I have corrected that in the initial post.

1 Like
6

you should take the advice to include at minimum these additional cannas, as these are critical separations that can save you tons of time and money.

you already mentioned buying distillate that you did a second pass on . you definitely want to test any product before you buy it. Converted CBD will generally have varying amounts of d8thc , if your method does not separate these isomers you wouldn’t be able to tell. Which could be a part cause of why your disti is testing lower than you expect.

I would not consider CBN as a minor, really one of the top 3 (other two being THC and CBD)

CBG/CBD and CBC/d9-THCa are other critical separations that you’ll see regularly. you should adjust your method until these separations are mostly resolved, you dont necessarily need to maintain calibrations for every compound if you dont care to quantify those as accurately

you should be testing your flower input as well so I would at least add d9-THCa to the list as well, you already have the HPLC. If you were only planning to test for neutrals you should have bought a GC-FID (cheaper and easier to run/maintain)

look at restek and cayman for cannabinoid mixtures for standards. they will save you money and time to buy the mixture vs a bunch of isolates and mixing them together at a lower concentration

feel free to pm me if you have more questions

1 Like
7

I didn’t explain well. We do distillation on ethanol extracted crude oil here via short path. The distillate we sell is marketed for THC only. Australian regulations require any cannabinoids over 2% w/w be on the label, and so you and the previous poster were correct that a more robust method of assay is helpful, it is certainly something we are working towards, but for now I’m just trying to get a sample prep method together for the distillate using what we have on hand. It will be easier to ask for more analytical budget once that is squared away.

Typically, due to feedstock, our CBD in the distillate is extremely low, however we have had issues with CBG which is also an argument for a more robust method. Again, the eternal dance between budget and science.

The flower we extract from is typically cannabis grown for THC rather than hemp, so I feel like there’s a reasonable expectation most of the THC is delta-9. Regardless, d8 has never appeared on any of our lab results to a degree worth worrying about, but you raise a good point.

Agreed, poor phrasing on my part.

The HPLC was purchased before my time and it isn’t in the budget at the current time to outfit a GC in the lab. We’ve done cost analysis on a GC outfit, particularly as there is some interest in terpene analysis, but as it stands that’s on the backburner.

I believe our supplier (Agilent) now has mixed standards, we will probably move to that once our stock of seperate standards have run their course.

Thank you for the response, appreciate the PM offer.

1 Like
8

Understandable, you gotta work with what is available

9

I want to reiterate what @Lincoln20XX states above,

in response to:

If your method is reliable for flower, you can use the same solvent system and method with very minor changes to your prep. Basically you need to estimate starting cannabinoid content (use just the major compound to make it simple), and do the math for diluting your sample so that your compounds of interest fall within your validated calibration range.

So if, for example, your normal prep is to take 100mg flower & sonicate in 10mL MeOH, decant 100mL of this solution into the HPLC vial and add 900mL 50%MeOH, then you would need to figure out the dilution for a more concentrated sample to end up in this same range. E.G. 20% flower diluted 1000x = 200ppm cannabinoids in solution, so diluting 85% distillate to that range should be trivial. No added expense.

3 Likes
10

Same boat! Purchased an Agilent 1200 and I’m tasked to make it work. Humbly I ask if anyone wants to share their distillate / flower prep SOPs, preferred standard vendors, and general tips.

I’ve ran a curve and some “known samples” that we’d tested with our trusted local regulated lab and the results were regularly 10-20% lower thc than expected and I’m trying to understand what I’m doing wrong.

So far for me:

Purchased 8 cannabinoid standards in 1ml vials (from Cayman chemical). Prepared a 6 point calibration curve by adding all standards together and then making 1ug/ml, 5ug/ml, 25ug/ml, 50ug/ml, 125ug/ml, 250ug/ml, and 500ug/ml samples. (The 1u didn’t look great so we omitted).

But that’s how we made the curve.

First time running machine really but have been ~lightly~ consulted.

Mobile phases I believe are 1) 24% water / 76% methanol + some formic acid and 2) Acetonitrile.

Our prep method for Distillate is:

Weigh out 50ug sample. Dilute in 5ml of methanol and vortex or mix for 2 minutes. Add 1.6ml of water and mix for 2 minutes. Take sample and filter it into 2ml beaker. This is our high dilution sample and we make a lower dilution of each sample (I think we did 10x but I’ll check notes).

Prep for flower is pretty much the same but with a 250ug sample.

Our consultant gave me a few Excel files I’m still studying. One is plotting the 6 point curve of the 8 cannabinoids on a graph and giving me a slope and intercept. Another Excel file has to do with entering the initial weight of the sample, as well as the area under the peaks of interest and doing the calculations to give us percentages.

Would love input, Wouldn’t mind constructive criticism. Will hopefully update standard vender and better details by end of week.

(Also before I get heat for it my boss purchased the HPLC and it’s my job to maintenance all equipment and keep track of their METRC work. So it wasn’t my call to hire the consult, or buy the machine in the first place. But I am excited to get it into my tool belt and do a good job at my job🙏)

11

These are all VERY small amounts to measure. We use between 0.1g and 1.0g for concentrate samples, and ~5g for biomass. The system we built ensures that every measurement we make or take is no less than three orders of magnitude away from the variance inherent in the measuring tool in use. Usually five or six.

As a result, we also catch when our third party lab makes a mistake in their calibrations.

Are your scales calibrated? Are your balances stable and level and grounded? Are you measuring your liquids gravimetrically or volumetrically? If you’re working in micrograms or milligrams the tiniest variance can skew your results by a shocking amount. How consistent are your techs?

Are you recording every mass at every step and taking that into account for your final results or are you assuming it’s done correctly and per your SOP? If you’re assuming, you do not actually know anything. Go back to square one.

Follow this pair of questions as deeply as you need to:

  1. How do we know what we know?
  2. How do we know that we know what we know?
3 Likes
12

Nonsense!!

…and if true, the entire reason you’re failing.

50ug +/- how many?!?

You can purchase a balance that will get you there, but I’m certain you don’t have one…Microbalances and Micro Balance Scales | Mettler Toledo

I believe you mean 50mg. And I suspect this may STILL be your problem.

Is that 50+/- 1mg. or 0.1mg what is the magnitude of the error you’re introducing here? How about those other measurements? 5ml +/- what? 1.6ml +/- what?

You’re much better off weighing your liquids (gravimetric vs volumetric) if you need accuracy (and you do).

Edit: and I see @Lincoln20XX has already gone here. In a kinder gentler fashion.

2 Likes
13

AOAC 2018.10.pdf (584.2 KB)

I’d start here for anyone trying to setup a method.

3 Likes
14

Disclosure: my company hired a consult to help. He has educated me and I am thankful for that but I do think there are many places to tidy up. Not that its an exclude I still need to fully understand what I’m doing and that’s the goal

I agree with 50ug being silly. but yes I do mean mean 50ug ± 10ug. (not mg) and I totally see how that would make for large error.

We have a scale that is certified up to .001g so you are absolutory right.

These are done with pipets. So ill weigh water on a scale to make sure the 5ml is = 5g and then I use the pipet and assumed there should be negligible error there, and same with the 1.6ml.

Thank you for your input!

15

This is key thank you! Honestly I needed to hear that the 50ug suggestion was out of pocket.

that is beautiful

I do know our scales are calibrated and do double check them in the morning (but with the 50ug I was only able to be accurate ± 10ug so surely that data is wrong. I’m still wrapping my head around making our curves and getting all my dilutions correct but I will get more exposure and experience. The guy my company hired to train me made this SOP based off of his CBD lab so I think there may be inconsistencies (and that’s not to deflect blame either, I just want to know how to get consistent and correct results).

No techs so far, I manage the production lab and I need to fully understand the HPLC and its process before trying to train any of my techs on it. Very ground level at the moment. I do record every weight when its samples. I gravimetrically check my pipets and then volumetrically for actually adding solvents and water.

Thank you! Ill do some thinking and research on other SOPs and hopefully come back a little more educated on the topic soon.

1 Like
16

0.001 gram is 1mg. There are 1000mg/g

1ug is 0.000001g. There are 1000000ug/g

So you are measuring 50+/- 1mg

You want 50+/- 0.1 at a minimum

And 5000 +/- 0.1mg would allow greater accuracy

Weighing 50ug to the nearest 10ug would also be problematic, but if you’re describing the scale as “certified to 0.001g” then you’re not weighing out 50ug.

4 Likes
17

The smallest measurement we take in our system is a bit over 100mg, most are over 1000mg. Our scale is accurate to +/- 0.1 mg.

Go get yourself a proper analytical balance. It doesn’t need to be a NIST traceable one, but with your level of uncertainty your calibration weights probably should be.

A consult from someone who understands math, metrology, and basic chemistry would probably be a good thing as well. I suggest you don’t let your bosses decide who to hire.

3 Likes
18

Big facts. I appreciate you guys highlighting where we can get better.

In a few weeks Ill reply with an update yall with an updated sop and better results to share!

19

Good to know

Good Move

This one is fair. I feel like with the insight you guys gave me, the machines manual, and things you guys just suggested I’m going to give it a go without a consult for now. The fellow is nice really I was frustrated with myself for not knowing it all the first go.

2 Likes