COSMOtherm screening of deep eutectic solvents for ultrasound-assisted carotenoid recovery from salmon processing by-products

The selection of a suitable extraction medium is a key step in the recovery of carotenoids from complex food-processing matrices. In this study, COSMOtherm screening was combined with experimental solvent evaluation and ultrasound-assisted extraction (UAE) to recover carotenoids from freeze-dried Atlantic salmon (Salmo salar) skin-on trimmings. Ninety deep eutectic solvent (DES) combinations were screened according to the activity coefficient of astaxanthin at infinite dilution (lnγ), and five DES were selected for experimental testing. Menthol:octanoic acid (MeC8, 1:1) had the lowest predicted lnγ (−1.94) and also gave the highest total carotenoid content (TCC), 11.48 μg AE/g FDSM. This value was approximately 3.3 times higher than that obtained with water and 4.0 times higher than with 70% ethanol. MeC8 was therefore used for further UAE experiments. Preliminary tests showed that increasing extraction time beyond 15 min did not improve carotenoid recovery. Ultrasound power, extraction time, and sample mass were then investigated using a Box-Behnken design. A significant linear model was obtained (p = 0.0012), with sample mass being the only significant individual factor (p = 0.0002). Numerical optimization gave conditions of 20% ultra-sound power (40W), 15 min extraction time, and 0.50 g sample mass, with a predicted TCC of 30.68 μg AE/g FDSM. The experimental value obtained under these conditions was 31.74 ± 0.96 μg AE/g FDSM, corresponding to a relative deviation of 3.46%. Overall, the combined computational and experimental approach enabled the selection of MeC8 and the definition of suitable UAE conditions for carotenoid recovery from salmon processing by-products.