RESUMEN
The separation of cannabinoids from hemp materials is nowadays one of the most promising industrial applications of liquid-liquid chromatography (LLC). Despite various experimental research efforts to purify cannabinoids, there are currently few works on process modeling. Thus, this study aimed to explore a straightforward approach to model the LLC separation of cannabinoids from two hemp extracts with different compositions. The feed materials were simplified to mixtures of preselected key components (i.e., cannabidiol, tetrahydrocannabinol, cannabigerol, and cannabinol). The elution profiles of cannabinoids were simulated using the equilibrium-cell model with an empirical nonlinear correlation. The model parameters were derived from the elution profiles of single-solute pulse injections. For the validation of the proposed approach, LLC separations with the two hemp extracts were performed in descending mode with the solvent system composed of hexane/methanol/water 10/8/2 (v/v/v). The injected sample concentrations were gradually increased from 5 to 100 mg/mL. The results showed that the approach could describe reasonably well the elution behavior of the cannabinoids, with deviations of only 1-2 min between simulated and experimental elution times. However, to improve the prediction accuracy, the model parameters can be refitted to the elution profiles of 3-4 systematically selected pulse injections with specific hemp extracts.