Maximizing Cannabinoid Separation for Potency Testing with LC

Researchers from the Department of Chemistry at Western Illinois University (Macomb, Illinois) conducted a study to optimize the separation of 18 cannabinoids for potency testing of hemp-based products, using liquid chromatography with a diode array detector (LC–DAD). The researchers found that differences in polarity and hydrophobicity among the columns significantly affected cannabinoid separation. LCGC International spoke to Liguo Song, the corresponding author of the paper stemming from this research, to discuss the study and its findings.

Your paper (1) outlines your research group’s search for the best separation of 18 cannabinoids, the maximum number of cannabinoids that have been quantified so far, for potency testing of hemp-based products using LC-DAD. Why did you feel this research was necessary?

The research was driven by the need for accurate and efficient analytical methods to quantify cannabinoids in hemp-based products. Precise potency testing of cannabinoids necessitates effective separation, which can be challenging because of their structural similarities. Although various methods exist, only two validated LC-DAD methods have achieved baseline separation of 18 cannabinoids, with the greatest challenge being the distinction between Δ⁹-tetrahydrocannabinol (Δ⁹-THC) and Δ⁸-THC. Conducting a thorough assessment of the most effective approaches is now both feasible and crucial to determining the optimal separation method.

Traditionally, what techniques were used to detect cannabinoids?

Gas chromatography (GC), paired with either flame ionization detection (FID) or mass spectrometry (MS), has been widely used for cannabinoid testing. However, acidic cannabinoids require derivatization before analysis to prevent decarboxylation under GC conditions.

What benefits does LC–DAD offer over previous techniques for this application?

Because of the thermal instability of underivatized acidic cannabinoids under GC conditions, LC, paired with either DAD or MS detection, has become a preferred alternative. Regarding LC detection, DAD is favored over electrospray ionization tandem mass spectrometry (ESI-MS/MS) because of its broader accessibility for crime laboratories, commercial suppliers, and farmers. Additionally, recent findings indicate that multiple reaction monitoring (MRM) cannot distinguish between Δ⁹-THC and Δ⁸-THC (2). However, the effectiveness of LC–DAD depends on achieving baseline separation of all cannabinoids, including unknown compounds that exceed the required limit of quantification (LOQ) in a sample. Therefore, LC–DAD method development must account for all cannabinoids previously quantified in hemp-based products. Furthermore, isocratic separation is preferred over gradient separation because changes in the mobile phase can cause baseline drift, as most neutral cannabinoids are detected at 230 nm or shorter wavelengths, leading to higher LOQs.

What were your main findings? Was there anything surprising or interesting?

The fastest potency testing of hemp-based products using LC-DAD was accomplished in just 15 min by employing two sequential columns (150 mm × 2.1 mm, 2.7 µm) with a diisobutyl-octadecyl (-OSi(iBu)₂C₁₈H₃₇) stationary phase. The mobile phase comprised 75.0% acetonitrile and 25.0% aqueous solution containing 0.03% formic acid and 0.5 mM ammonium formate at pH 2.97, with a flow rate of 0.5 mL/min. A slightly improved resolution of the 18 cannabinoids was obtained within 18.5 min using two sequential columns (150 mm × 2.1 mm, 2.7 µm) with a dimethyl-octadecyl (-OSi(CH₃)₂C₁₈H₃₇) stationary phase under similar conditions, except for a mobile phase containing 77.5% acetonitrile and a reduced flow rate of 0.45 mL/min because of back pressure higher than 600 bars. Furthermore, a rapid 7.0-min separation was achieved for potency testing of hemp-based products by LC-ESI-MS/MS using a column (150 mm × 2.1 mm, 2.7 mm) with a carbamate-embedded reverse-phase (RP) stationary phase (-OSi(CH₃)₂C₃H₆O(CO)NHC₁₂H₂₅). The mobile phase consisted of 70.0% acetonitrile and 30.0% aqueous solution of 0.01% formic acid and 1 mM ammonium formate at pH 3.38, at a flow rate of 0.5 mL/min.

The first two separations were less surprising than the final one. It was interesting that the two polar-embedded RP stationary phases (-OSi(CH₃)₂C₃H₆O(CO)NHC₁₂H₂₅and -OSi(CH₃)₂C₃H₆NH(CO)C₁₅H₃₁) demonstrated significantly better resolution of the D⁹-/D⁸-THC pair compared to the two traditional RP stationary phases (-OSi(CH₃)₂C₁₈H₃₇ and -OSi(iBu)₂C₁₈H₃₇).

What were the main challenges you encountered from an analytical perspective, and how did you overcome them?

Controlling the pH of the aqueous solvent is essential for achieving the desired separation. The pH was regulated using the HCO₂H/NH₄HCO₂ buffer system based on the Henderson–Hasselbalch equation, with a low total concentration of HCO₂H and NH₄HCO₂to achieve a low LOQ. However, the pH may vary depending on the LC-grade water used in its preparation, even when sourced from the same product but with different lot numbers. Consequently, slight adjustments to the relative concentrations of HCO₂H and NH₄HCO₂may be necessary to obtain the expected separation. Once the optimal ratio was established, the consistent use of the same products ensured robust pH control of the aqueous solvent.

Were there any cannabinoids that were easier to detect, or, conversely, harder to detect?

Most neutral cannabinoids, including cannabidivarin (CBDV), cannabidiol (CBD), cannabigerol (CBG), tetrahydrocannabivarin (THCV), D⁹-THC, D⁸-THC, cannabicyclol (CBL), and cannabicitran (CBT), had similar UV absorption spectra, with two apexes at approximate 210 and 230 nm. Cannabinol (CBN) had two UV absorption apexes at approximately 221 and 282 nm. Cannabichromene (CBC) had two UV absorption apexes at about 230 and 282 nm.

Most acidic cannabinoids, including cannabidivarinic acid (CBDVA), cannabidiolic acid (CBDA), cannabigerolic acid (CBGA), tetrahydrocannabivarinic acid (THCVA), D⁹- tetrahydrocannabinolic acid (D⁹-THCA), and cannabicyclolic acid (CBLA), had similar UV absorption spectra, with three apexes at approximate 224, 269, and 305 nm. Cannabichromenic acid (CBCA) and cannabinolic acid (CBNA) had only one major peak of UV absorption with their apexes at about 254 and 262 nm, respectively.

For cannabinoids with multiple UV absorption peaks, absorption intensity decreased as wavelength increased, while background absorption also declined. As a result, detection at 230 nm or shorter wavelengths was more prone to baseline fluctuations but remained crucial for most neutral cannabinoids. In contrast, detection at wavelengths longer than 230 nm was less affected by baseline fluctuations. However, the signal-to-noise ratio (S/N) was typically higher at absorption apexes at lower wavelengths.

How did you process the data to obtain the results you were looking for?

Key chromatographic parameters, including the retention factor, selectivity (separation factor), separation efficiency (theoretical plate number), and resolution, were calculated and compared.

Were there any limitations to this research that are important to note?

This research is limited to the quantification of cannabinoids naturally produced by hemp.

While your technique was used to separate 18 specific cannabinoids, you mention in your introduction that there are more than 150 cannabinoids that have been isolated and identified. How could the findings of this method be applied to develop ways to separate more of these cannabinoids?

Potency testing of hemp-based products is designed to quantify all major cannabinoids exceeding the required LOQ in a sample to evaluate its quality and safety. The findings of this research may aid in the more efficient separation of minor cannabinoids present below the required LOQ.

Moving forward, are you considering implementing any improvements to the method?

In addition to Δ⁹-THC, hemp naturally produces other THC isomers with psychotropic properties, for example, Δ⁸-, Δ¹⁰-, and Δ^6a,10a-THC, but in amounts too small to induce psychotropic effects. These THC isomers can also be synthetically produced in large quantities by converting CBD through non-specific isomerization reactions. As the increase in hemp production caused the price of CBD to plummet, the hemp industry started introducing synthetic THC isomers into the market. This shift was driven by arguments that the 2018 Farm Bill legalized their sale, as these compounds naturally occur in hemp and are derived from legally extracted CBD. Currently, Δ⁸-THC is the most popular synthetic THC isomer on the market, followed by Δ¹⁰-THC, while Δ^6a,10a-THC is also commercially available. However, these products typically contain not only the labeled THC isomer but also other synthetic THC isomers, because of the non-specific nature of the isomerization reactions used in their production. Moving forward, we aim to develop a method capable of simultaneously quantifying synthetic THC isomers.

Are there any next steps in this research?

Our next step is to develop a method that can simultaneously quantify synthetic THC isomers, along with the 18 cannabinoids described in this research.

Liguo Song is an associate professor of forensic chemistry and the B.S. forensic chemistry program director in the Department of Chemistry at Western Illinois University. His research is currently focused in the forensic analysis of abused drugs and toxicological specimens using gas chromatography, liquid chromatography, ambient ionization mass spectrometry, gas chromatography mass spectrometry, and liquid chromatography mass spectrometry. He has published numerous papers in analytical chemistry related to forensic analysis, mass spectrometry, chromatography, pharmaceutical analysis, biomedical analysis, capillary electrophoresis, and DNA analysis.

Reference

1. Owolabi, A.; Ogunsola, O.; Joens, E.; Kotler, M.; Song, L. A Systematic Study of Liquid Chromatography in Search of the Best Separation of Cannabinoids for Potency Testing of Hemp-Based Products. Molecules 2025, 30 (4), 758. DOI: 10.3390/molecules30040758

2. Meyer, G.; Adisa, M.; Dodson, Z.; Adejumo, E.; Jovanovich, E.; Song, L. A Liquid Chromatography Electrospray Ionization Tandem Mass Spectrometry Method for Quantification of up to Eighteen Cannabinoids in Hemp-derived Products. J. Pharm. Biomed. Anal. 2024, 238, 115847.https://doi.org/10.1016/j.jpba.2023.115847