Optimization of pulsed ultrasonic‐assisted extraction of punicalagin from pomegranate ( Punica granatum ) peel: A comparison between response surface methodology and artificial neural network‐multiobjective genetic algorithm

Abstract

The present study focuses on the valorization of pomegranate peel waste by extracting bioactive compounds (mainly punicalagin) through pulsed ultrasonic-assisted extraction. Response surface methodology (RSM) and artificial neural network coupled with multiobjective genetic algorithm (ANN-MOGA) were used to determine the optimum extraction condition, namely, solvent volume, amplitude, time, and duty cycle. The responses for the optimization process include punicalagin content, ellagic acid, antioxidant activity, total phenolic content (TPC), total flavonoid content, and anthocyanin content. By applying ANN-MOGA it was found that at optimum condition of 35 ml of solvent, 35% amplitude, 23 min and 100% duty cycle, punicalagin, ellagic acid, antioxidant activity, and TPC was 6.63%, 18.59%, 52%, and 4.23% higher than RSM predicted values, respectively. Thus, the study proves that ANN-MOGA is an effective tool in predicting maximum punicalagin production from pomegranate peel. The optimization showed a close match between predicted and experimental data. Practical applications Pomegranate peel is a source of bioactive components. With the increase in demand for “clean label” products worldwide, valorization of pomegranate peel could be a potent alternative. This study investigates the ultrasonic-assisted extraction of major bioactive components from pomegranate peel in order to aid scaling up of the process. Also, the comparison between mathematical tools can give an insight into higher yielding process parameters.