Combined ANFIS and numerical methods to simulate ultrasound-assisted extraction of phenolics from chokeberry cultivated in China and analysis of phenolic composition

A 20-kHz ultrasound was utilized to extract phenolics from chokeberry cultivated in China using both distilled water and 50% aqueous ethanol as solvents. Sonication was performed at different acoustic energy density and temperature levels. Adaptive neuro-fuzzy inference system (ANFIS) was successful in correlating the operation parameters during extraction with total phenolic yield with the value of determination of coefficient higher than 0.97. In the constructed ANFIS model, “Gaussmf” was selected as the input membership function. Meanwhile, diffusion model was solved numerically to model the mass transfer process during sonication and identify the effective diffusivity of total phenolics. Generally, both AED and temperature had positive influences on the effective diffusivity of total phenolics. Furthermore, according to the analysis of high resolution mass spectrometry and high performance liquid chromatography, delphinidin-3- O -rutinoside, protocatechuic acid, chlorogenic acid and rutin had considerably high amounts in this Chinese chokeberry cultivar. In addition, through analyzing the phenolic composition in extracts, it was found that distilled water was an effective solvent for the extraction of some anthocyanins from chokeberry including cyanidin-3- O -glucoside, cyanidin-3- O -rutinoside and delphinidin-3- O -rutinoside, while sonication using 50% aqueous ethanol as solvent led to high recovery of phenolic acids and flavonols.