All reagents used during this project were of analytical grade or of higher purity. Reference standards of 1000 µg/ml CBD, CBDA, THC, and THCA in acetonitrile (ACN) were purchased from LCG Standards (London, UK). Water was obtained from a Rephile direct pure UP Ultrapure & RO Lab water system (Boston, MA, USA). HPLC-grade ACN, methanol, and formic acid were purchased from Sigma-Aldrich (Johannesburg, RSA).

For the calibration standards, a mixture solution containing CBD, CBDA, THC, and THCA was prepared with an initial concentration of 100 µg/ml in ACN which was then serially diluted to obtain seven concentrations of 50.00, 25.00, 12.50, 6.25, 3.13, 1.56, and 0.78 µg/ml. All calibration samples were prepared fresh before validation and stored at − 20 °C.

Ten commercially available CBD tea products were purchased and used to test the method application. The products evaluated had specific label claims regarding their CBD content with one product only stating that it contains C. sativa. Declared CBD amount per teabag was 2 mg for products A, E, and F and 5 mg for products B, D, and G–J. There was no stated amount of CBD in product C. The contents of the teabags were weighed and to obtain similar CBD concentrations in all extracted samples, products A, C, E, and F were extracted in 50 ml ACN and the remaining six products in 100 ml ACN. After the addition of extraction solvent, the samples were sonicated for 10 min, after which it was filtered with a 0.45-µm syringe filter into HPLC vials for analysis (Laanet et al. 2022). All samples were prepared in triplicate and each sample was injected three times to afford nine data points per product.

To determine the recovery and specificity of the analysis, four samples (A–D) were spiked with known amounts of CBD. A teabag of each product was selected, and the contents were split into four equal samples (200 mg). Fifty milliliters of ACN was added to each sample, followed by sonication for 10 min. All samples (0.5 ml) were filtered into HPLC vials after which two of the four replicates of each sample were spiked with 5 µl of a known concentration of CBD and the remaining two replicates were spiked with the same volume of ACN containing no CBD. The calculation that was used to determine recovery was:

$$\mathrm=\left(\frac-\mathrm}}\right)\times 100$$

In addition, tea infusions were prepared of all 10 samples according to the manufacturers’ recommendations. Deionized water was brought to boiling point after which a tea bag was added. The teabag was left to infuse for 5 min whereupon it was filtered (0.45 µm) into HPLC vials for analysis, and in order to test solution vs suspension, samples were also analyzed without being filtered (with only the tea bag material acting as a rough filter).

Validation and quantitative analysis were conducted on a Shimadzu iNexera LC-2040C system consisting of a quaternary pump, solvent degasser, autosampler, column oven, and photodiode array detector. LabSolutions software was used to control all instrument components and acquire, analyze, and store the data. Chromatographic separation was performed on a Zorbax Eclipse Plus C18 column (2.1 × 50 mm, 1.8 µm). Both MeOH and ACN in combination with water were tested, and it was found that the Zorbax Eclipse column using 0.1% formic acid (A) and ACN containing 0.1% formic acid (B) was optimal regarding peak shape and separation. To reduce the possibility of carryover between samples a gradient system was employed starting at 65% B for 2 min, increasing to 100% B at 2.5 min and kept at 100% until 4.5 min. The system was returned to 65% B until 7.5 min to allow for re-equilibration. Blank injections were performed after every 10 samples to test for and further reduce the possibility of carryover between samples. The flow rate used was 0.4 ml/min, and the injection volume for all samples was 2 µl. The column oven was set to 40 °C and the autosampler was set to 6 °C. The PDA detector stored UV data over the range of 190 – 600 nm and detection and quantification were conducted at 210 nm for CBD and THC and at 221 nm for CBDA and THCA (Zivovinovic et al. 2018; FDA 2020; Ryu et al. 2021).

The method was validated for linearity, precision (repeatability, interday, and intraday), limit of quantification, limit of detection, and accuracy according to the FDA Office of Regulatory Affairs (ORA-LAB.5.4.5) “methods, method verification and validation” (FDA 2020). Peak purity for all samples was conducted using Shimadzu’s peak purity algorithms, based on cosine vector comparison of spectra at five selected points across the peaks.