Optimization by artificial neural network and modeling of ultrasound-assisted flavonoid extraction from Phyllanthus emblica L. based on deep eutectic solvents

Phyllanthus emblica L. contains abundant flavonoids with multiple biological effects, drawing interest from the functional food industry. This research employed a deep eutectic solvent (DES) ultrasonic-assisted extraction method to extract flavonoids from Phyllanthus emblica L. To optimize the extraction process and assess total flavonoid yield, response surface methodology and artificial neural networks (ANNs) were employed to model the extraction process and predict the total flavonoids, yielding the optimal extraction conditions: 47 °C, 20% water content, 30 min of ultrasound time, and a liquid-solid ratio of 40. Under the optimal extraction condition, the extraction yield of total flavonoids reached 22.896 mg/100g. In addition, the extraction mechanism of different solvents was studied using molecular dynamics simulation. It was found that the use of choline chloride-ethylene glycol as the extraction solvent showed greater solvent accessibility surface area, more hydrogen bonds between choline chloride-ethylene glycol and the extract, and lower intermolecular interaction, resulting in higher extraction efficiency. Further research on the antioxidant and thermal stability of Phyllanthus emblica L. flavonoid extracts evaluated the application potential of total flavonoids. The DES ultrasonic-assisted extraction method effectively enhanced the flavonoid extraction yield from Phyllanthus emblica L. and reduced thermal decomposition, making it more suitable for application in the functional food industry.