Extraction of mandelic acid with ionic liquids: parametric study, model and process optimization with L-SHADE

Mandelic acid (MA), an aromatic α-hydroxycarboxylic acid (C₆H₅CH(OH)COOH) derived from bitter almonds, is widely utilized in the pharmaceutical, cosmetic, and chemical industries. It serves as a key intermediate in the synthesis of antibiotics and other pharmaceuticals. It is valued in skincare applications for its gentle exfoliating properties, offering a less irritating alternative to glycolic acid in the treatment of acne, wrinkles, and photoaging. This study investigates an innovative method for extracting MA using two hydrophobic ionic liquids (ILs): trihexyl(tetradecyl)phosphonium phenylphosphinate ([P₆₆₆₁₄][Phos]) and trihexyl(tetradecyl)phosphonium decanoate ([P₆₆₆₁₄][Dec]) as extractants combined with heptane. Extraction experiments were conducted under different conditions to determine the main factors affecting extraction efficiency. Experimentally, the highest extraction yield of 96.36% was obtained at 25 °C using a solvent mixture of heptane and 160 g/L [P₆₆₆₁₄][Phos], with an aqueous phase pH of 2 to ensure that MA remains in its non-dissociated form. The organic phase was successfully regenerated at 45 °C with sodium carbonate, enabling efficient recovery of MA and reuse of the ILs extractant. An artificial neural network (NN) model was first developed to simulate and predict extraction performance based on experimental input variables, and the Linear Population Size Reduction Success History-based Adaptive Differential Evolution (L-SHADE) algorithm was applied to optimize the process conditions. The results show that the proposed ILs-based extraction system is a promising, environmentally friendly alternative for efficiently recovering MA, with great potential for use in both industrial and cosmetic manufacturing settings.