Simultaneous optimization of ultrasonic-assisted extraction of Danshen for maximal tanshinone IIA and salvianolic acid B yields and antioxidant activity: A comparative study of the response surface methodology and artificial neural network
Zhe-Xuan Yu,Yang-Yang Zhang,Xi-Xi Zhao,Li Yu,Xiao-Bo Chen,Hai-Tong Wan,Yu He,Wei-Feng Jin
DOI: https://doi.org/10.1016/j.indcrop.2020.113199
IF: 6.449
2021-03-01
Industrial Crops and Products
Abstract:<p>Dried roots and rhizomes of <em>Salvia miltiorrhiza</em> Bge. (Danshen in Chinese) are widely used in Chinese herbal medicine. These components from Danshen have significant anti-oxidation properties owing to high levels of tanshinone IIA (TIIA) and salvianolic acid B (Sal B). To make the best use of this natural resource, response surface methodology (RSM) and artificial neural network (ANN) were used for the modeling and optimization of ultrasound-assisted extraction (UAE) of TIIA and Sal B together to determine the antioxidant activity of the extracts obtained from Danshen. In this study, the Box-Behnken design (BBD) was used to improve extraction time (<em>X<sub>1</sub></em>), solvent-to-material ratio (<em>X<sub>2</sub></em>), extraction temperature (<em>X<sub>3</sub></em>), and ethanol concentration (<em>X<sub>4</sub>)</em> for the optimal combination of the comprehensive yield of TIIA and Sal B (<em>Y<sub>1</sub></em>) and the antioxidant activity (<em>Y<sub>2</sub></em>). The optimal process parameters were determined to be as follows: extraction time, 73 min; solvent-to-material ratio, 11 mL/g; extraction temperature, 76℃; and ethanol concentration, 80 %. Using these conditions, the predictive optimal combination revealed a comprehensive evaluation value of 16.2281 and an antioxidant activity of 1.1453 mM FeSO<sub>4</sub>/5 g, while the experimental average values for these parameters were determined to be 16.1826 and 1.1415 mM FeSO<sub>4</sub>/5 g, respectively. It was clear that the ANN model had higher accuracy in predictive and optimization capabilities, with higher <em>R<sup>2</sup></em> and lower RMSE, MAE, and relative deviations values, than did RSM. Hence, the ANN model proved to be more effective for the analysis and improvement of the extraction process.</p>
agronomy,agricultural engineering