Typically cannabis potency testing to determine the quantity of cannabinoids present in a sample is done with HPLC that has a UV detector. However, a slightly more accurate method of LC-MS/MS might also be utilized, especially for matrix heavy sustances such as with edibles or where trace analysis is useful.
HPLC separates compounds based on the selective distribution between a stationary phase and a mobile solvent phase. The stationary phase most often deployed for cannabinoid testing are silica particles bound with a C18 ligand. However, other stationary phases can also consist of other non-polar ligands such as biphenyl which is also commonly employed for cannabinoid separation. The mobile phase consists of a mixture of an organic phase (methanol or acetonitrile) with an aqueous phase (ultra-pure water) applied at a gradient of increasing organic phase. Without going into too much technical detail about hydrophobicity, steric interaction, hydrogen bonding capacities, etc… you can loosely say that molecules are attracted to the stationary phase based on polarity or the shape and charge distribution of each molecule. The more non-polar molecules will interact with the non-polar stationary phase more and elute towards the end of the gradient as the organic phase increases. Once again, I need to mention that this is a very simplified explanation as elution order can be influenced by the stronger dipole moment of an aprotic solvent such acetonitrile vs a protic solvent such as methanol. There are also a variety of solvent modifiers (ammonium formate, ammonium acetate, formic acid, TFA, etc) that can also influence chromatography.
Once these compounds get separated with HPLC (sometimes simply referred to as LC) they are typically detected with a UV detector. Using Beer’s Law, the concentration is directly proportional to absorabance of light that is obtained and compared to the absorbance of a previously formed calibration curve. A mathematic equation is then applied to the concentration of the diluted sample to compensate for the original dilution factor and exact sample mass. The result is the original concentration before dilution on a per mass basis.
A UV detector a cheap an effective way to measure cannabinoids since they are in such high concentration compared to other molecules in the plant. However, one drawback to UV detection is that this technique is not compound selective. Any light absorbing compound eluting at or sufficiently near the expected retention time will inaccurately quantitatively add to the actual value or potentially cause false positives in the absence of the intended compound of interest. With this said, LC-UV is regarded as the go to analytical technique for cannabinoids potency quantification in flower, concentrates, shatter, rosin, and wax since the concentration is typically high and any matrix effects can be diluted out to the point that they are insignificant.
Inaccuracies in data are more likely to arise when using LC-UV to quantify cannabinoids in edibles. Edibles have a much more complex matrix with a much lower concentration of THC and other cannabinoids per weight of total product. For these types are matrices, a more expensive mass spectrometry based detector that is selective to the compound of interest is best employed. Tandem mass spectrometry, also referred to as LC-MS/MS or triple quad mass spectrometry, adds an extra layer of separation with compounds of a particular mass to charge ratio and then fragments ions with that mass to charge ratio at specific electronic parameters and finally detects the concentration of those fragments to get a metaphorical fingerprint of the compound. To correctly identify the compound you need 4 sets of criteria: the retention time of the compound, the mass to charge ratio of the compound, the observance of 2 distinct fragments of that compound, and the ratio of those fragments at specified parameters programmed into the MS system. Whereas, LC-UV only has one form of identification based on the retention time on the LC column.