Extraction of Ursolic Acid from Apple Peel with Hydrophobic Deep Eutectic Solvents: Comparison between Response Surface Methodology and Artificial Neural Networks

Abstract

Extracting ursolic acid (UA) from plant resources using organic solvents is incompatible with food applications. To address this, in this study, 15 edible hydrophobic deep eutectic solvents (HDESs) were prepared to extract UA from apple peel, the extraction conditions were optimized, and the optimization strategies were compared. It was found that the solubility of UA in the HDESs can be 9 times higher than the traditional solvent such as ethanol. The response surface optimization concluded that temperature had the greatest effect on the extraction and the optimized test conditions obtained as follows: temperature of 49 °C, time of 32 min, solid–liquid ratio of 1:16.5 g/mL, respectively. Comparing the response surface methodology (RSM) and artificial neural networks (ANN), it was concluded that ANN has more accurate prediction ability than RSM. Overall, the HDESs are more effective and environmentally friendly than conventional organic solvents to extract UA. The results of this study will facilitate the further exploration of HDES in various food and pharmaceutical applications.

Keywords

• apple peel
• ultrasonic-assisted extraction
• response surface methodology
• artificial neural network