Multiscale mechanism exploration and experimental optimization for rosmarinic acid extraction from Rosmarinus officinalis using natural deep eutectic solvents

This study aimed at efficient rosmarinic acid (RA) extraction from rosemary (Rosmarinus officinalis L.) using natural deep eutectic solvents (NADESs). Quantum chemical calculations and molecular dynamics simulation were used to dissect the microscopic mechanisms. Three choline chloride (ChCl) based NADESs and ethanol (traditional extractant as reference) were investigated. The σ-profiles of RA and each solvent were analyzed. The infinite dilution activity coefficient (γ∞) of RA in different solvents, the total non-bonded interaction energy and the average H-bond number were calculated, to predict the RA solubility, understand the bond type and strength in each system. Among the solvents studied, ChCl/LA (lactic acid) exhibited wider complementary σ-profile with that of RA, as well as lower lnγ∞, higher H-bond number and predominant electrostatic-interaction. Moreover, self-diffusion coefficients revealed that RA had better compatibility with ChCl/LA. Additionally, with different solvents as extractant, ChCl/LA outperformed the others by exhibiting the highest RA extraction yield. Based on results from single factor experiments, Box-Behnken design with response surface methodology was performed for further process optimization. The regressed model obtained was reliable from the ANOVA analysis of variances. RA extraction yield was 6.10% under optimal conditions of 69 °C, 5.6 h, liquid to solid ratio 34 and water content 13.5% (w/w). The combination of simulation and experiment in this study can shed light on the exploration of natural product extraction with high efficiency using green and novel NADESs.