MIXTURE DESIGN OPTIMIZATION OF SOLVENT COMPOSITION FOR TOTAL SOLUBLE PHENOLIC COMPOUND EXTRACTION FROM FERRINHO BEANS

Ferrinho beans are a regional resource with potential for valorizing local agrobiodiversity, but reliable phenolic analysis depends on optimized extraction. This study optimized the solvent composition for phenolic extraction from Ferrinho beans and validated the predictive model. A four-component simplex-centroid mixture design was used with water, methanol, ethanol, and acetone. The design included 4 pure solvents, 6 binary, 4 ternary, 4 enriched formulations, and 4 replicates of the central point (22 trials). Response was absorbance at 750 nm by Folin-Ciocalteu assay. Modeling and desirability analysis identified the optimal composition (P < 0.05; R² > 0.85). Absorbance ranged from 0.134 to 0.456. Among pure solvents, water was highest (0.273), followed by methanol (0.249), acetone (0.140), and ethanol (0.134). The 1:1 water-acetone mixture yielded the highest experimental response (0.456), while water-methanol 1:1 decreased response to 0.174, showing that solvent combination does not always improve extraction. Ternary mixtures water-methanol-acetone (0.383) and water-ethanol-acetone (0.368) also performed well. A special cubic model predicted an optimum of 52.28% water, 5.24% methanol, 0.09% ethanol, and 42.39% acetone, with predicted absorbance of 0.457 and desirability 1.0. Validation showed <3% deviation between observed and predicted values. The validated solvent provides a robust basis for phenolic quantification and further characterization of Ferrinho beans.