2.4.3. UEAE Optimization by Experimental Design

A definitive screening design (DSD) [32] was applied to determine the best combination of UEAE variables to optimize simultaneously the recoveries of hydrophilic and lipophilic compounds from the raspberry, strawberry and blackberry pomaces. First, raspberry pomace was used as a model for the study, then the optimized and validated parameters were applied to the two other pomaces.

Six independent factors were considered as key factors: US amplitude (%), pH, E/S ratio (%, w/w), S/L ratio (%, w/v), time (min) and temperature (°C). Three levels were required per factor (minimum, centered, maximum). The number of experiments was defined by: N = 2 k + 1, where k represents the number of independent factors, which leads to 13 experiments, realized in triplicate. Three responses were considered: the total polyphenol content, antioxidant activity and oil extraction yields (assessed according to Section 2.3.1, Section 2.3.2 and Section 2.3.6). The yields of extraction were calculated as g per 100 g of milled DP.

To evaluate the results, a loss function was used. This tool allows one to give one score to an experiment, by assessing the yields and the corresponding deviations. A loss function tending to 0% is synonymous with the best extraction parameters, with the highest yields and the lowest standard deviations, while a loss function of 100% corresponds to the worst experiment.