Abstract: Ultrasonic was used to help extract amygdalin from loquat seed with ethanol aqueous solution. To optimize the extraction process, the effects of operation parameters such as ethanol concentration, temperature, extraction time, liquid-to-solid ratio and ultrasonic power on extraction efficiency were investigated by single-factor experiments. Based on this, a two-level, five-factor experimental design was used to investigate the affecting significance of the operation parameters, and a quadratic model for extraction efficiency as a function of ethanol concentration, liquid-to-solid ratio and ultrasonic power that were determined to be main affecting variables was established using central composite design, and the optimization of the three operation parameters were performed using response surface methodology. Ethanol concentration of 75%, temperature of 60 ℃, extraction duration of 30 min, liquid-to-solid ratio of 12:1(mL/g) and ultrasonic power of 250 W were found optimum and the extraction efficient was observed to be 39.07 mg/g, with the relative error being 0.66% compared to the predicted value. Thus, the established model for extraction efficiency had high prediction accuracy.

Key words: loquat seed, amygdalin, ultrasonic-assisted extraction, optimization, response surface methodology