Extraction of phenolic compounds from cocoa shell: Modeling using response surface methodology and artificial neural networks

Abstract

This work’s objective was to model and optimize a green extraction method of phenolic compounds from the cocoa shell as a strategy to revalorize this by-product, obtaining novel high-value products. According to a Box-Behnken design, 27 extractions were carried out at different conditions of temperature, time, acidity, and solid-to-liquid ratio. Total phenolic compounds, flavonoids, flavanols, proanthocyanidins, phenolic acids, o-diphenols, and in vitro antioxidant capacity were assessed in each extract. Response surface methodology (RSM) and artificial neural networks (ANN) were used to model the effect of the different parameters on the green aqueous extraction of phenolic compounds from the cocoa shell. The obtained mathematical models fitted well for all the responses. RSM and ANN exhibited high estimation capabilities. The main factors affecting phenolic extraction were temperature, followed by solid-to-liquid ratio, and acidity. The optimal extraction conditions were 100 °C, 90 min, 0% citric acid, and 0.02 g cocoa shell mL−1 water. Under these conditions, experimental values for the response variables matched those predicted, therefore, validating the model. UPLC-ESI-MS/MS revealed the presence of 15 phenolic compounds, being protocatechuic acid, procyanidin B2, (−)-epicatechin, and (+)-catechin, the major ones. Spectrophotometric results showed a significant correlation with the UPLC results, confirming their potential use for screening and optimization purposes. Aqueous phenolic extracts from the cocoa shell would have potential use as sustainable food-grade ingredients and nutraceutical products.