Optimization of ultrasonic-assisted extraction of three main taxoids in the twigs of Taxus x media using multi-objective response surface methodology

A multi-objective response surface methodology (RSM) based on a desirability function analysis (DFA) was used to model and optimize the operational parameters of ultrasonic-assisted extraction of three main taxoids in the twigs of Taxus. media. Liquid-solid ratio, extraction temperature, extraction time, and ultrasonic power were chosen as four independent variables, with three objective variables considered: the extraction yields of 10-deacetylbaccatin III (10-DAB III), cephalomannine and paclitaxel. Under the predicted conditions with the highest desirability'', compared with the predicted value, the goodness of fit for the experiment extraction yields of 10-DAB III, cephalomannine and paclitaxel was 97.36, 100.81, and 97.46%, respectively. Analysis of variance (ANOVA), regression analysis, and verification test indicated a satisfactory correlation between the experimental data and predicted values. These results showed that RSM and DFA were good methods to solve the problem of multi-objective optimization.