Study on the Technology of Extracting Fucoxanthin from Hizikia fusiforme by Cyclic Ultrasound

The purpose of this study was to explore the extraction process of fucoxanthin from frozen Hizikia fusiforme by cyclic ultrasound. Based on single factor test, the response surface model and BP neural network model optimized by genetic algorithm were used to optimize the extraction parameters on the basis of single factor experiment. The results showed that the optimal extraction conditions, obtained by response surface model, were as follows: Extraction time 119 min, ultrasonic power 600 W, liquid-solid ratio 26.5:1 L/kg, circulating flow rate 8.6 L/min. Under these conditions, the actual extraction rate of fucoxanthin was 0.2195 g/kg, and the purity of the product was 3.42%. The optimal extraction conditions, obtained by BP neural network model, were as follows: Extraction time 118 min, ultrasonic power 680 W, liquid-solid ratio 23.5:1 L/kg, circulating flow rate 7.0 L/min. Under these conditions, the actual extraction rate of fucoxanthin was 0.2318 g/kg, and the purity of the product was 3.41%. The neural network model is feasible and slightly better than the response surface modelconcerning the extraction conditions, which has a certain reference value in large-scale algae extraction of fucoidin.