Extraction of pectin from watermelon rinds using sequential ultrasound-microwave technique: Optimization using RSM and ANN modeling and characterization

Abstract

This study aimed to optimize pectin extraction from watermelon (Citrullus lanatus) rind using sequential ultrasound-microwave assisted extraction (UMAE) with artificial neural network (ANN) and response surface methodology (RSM). The effects of pH, sonication time, microwave power, and irradiation time on pectin yield were evaluated. The ANN model showed higher precision in predicting yield compared to the RSM model. The optimal yield was 32.11 % under the conditions of pH 2.01, sonication time 54.23 min, microwave power 900 watts, and irradiation time 6.34 min. This study evaluated the effects of different extraction techniques, including ultrasound-assisted extraction (UAE), microwave-assisted extraction (MAE), and microwave-ultrasound-assisted extraction (MUAE), on pectin yield and their physicochemical properties under optimal UMAE conditions. The highest pectin yield was achieved with UMAE, followed by MUAE, MAE, and UAE. The Fourier transform infrared spectroscopy (FTIR) analysis confirmed the presence of homogalacturonan, galacturonic acid backbone, and methyl esters in the extracted pectin. The viscosity study revealed that the pectin solution showed pseudoplastic behavior at 0.2 % w/v. All extracted pectin in different methods had high-methoxy content ranging from 7.68 ± 0.56 % to 11.96 ± 1.29 % and a degree of esterification between 56.55 ± 0.68 % and 63.43 ± 1.54 %. However, UMAE showed significantly lower energy consumption and CO₂ emissions, suggesting it as a sustainable approach for pectin extraction from watermelon rind.

Keywords: Artificial neural network; Food waste; Optimization; Pectin extraction; Sequential ultrasound-microwave; Sonication.