Comparative study on optimization of NADES extraction process by dual models and antioxidant activity of optimum extraction from Chuanxiong-Honghua

The natural deep eutectic solvent (NADES) was used as the extraction solvent to extract the four ingredients from Chuanxiong-Honghua. NADES-7 composed of l-proline and lactic acid (1:2) has a better extraction effect compared with traditional solvents. The extraction process was optimized using response surface methodology (RSM) and ant colony algorithm-enabled backpropagation neural network (ACO-BPNN), which were compared with each other. ACO-BPNN has a higher R2 and predictive value than RSM. The optimal process parameters predicted by the ACO-BPNN were determined to be as follows: solvent-to-material ratio, 8 mL/g; extraction time, 67 min; water concentration, 48%; and extraction temperature, 50 °C. The best predicted comprehensive evaluation value was 4.2634, and the mean value obtained from the experiment was 3.8678. Compared with traditional solvents, the antioxidant activity of NADES-7 extract was better, and the mean value of DPPH, FRAP (ferric reducing antioxidant power), and ABTS assays were 0.1555, 0.1132, and 0.0894 mmol Trolox/g dry weight, respectively. Experimental results show that ACO-BPNN can better optimize the extraction process of many different polar compounds in plant combinations.