There is a great interest in finding alternatives and green solvents in extraction processes to replace petroleum based solvents. In order to investigate these possibilities, computational methods, as Hansen solubility parameters (HSP) and conductor-like screening model for real solvent (COSMO-RS), were used in this work to predict the solvation power of a series of solvents in salmon fish lipids. Additionally, experimental studies were used to evaluate the performance in lipids extraction using 2-methyltetrahydrofurane, cyclopentyl methyl ether, dimethyl carbonate, isopropanol, ethanol, ethyl acetate, p-cymene and d-limonene compared with hexane. Lipid classes of extracts were obtained by using high performance thin-layer chromatography (HPTLC), whereas gas chromatography with a flame ionization detector (GC/FID) technique was employed to obtain fatty acid profiles. Some differences between theoretical and experimental results were observed, especially regarding the behavior of p-cymene and d-limonene, which separate from the predicted capability. Results obtained from HPTLC indicated that p-cymene and d-limonene extract triglycerides (TAGs) and diglycerides (DAGs) at levels of 73 and 19%, respectively, whereas the other studied extracts contain between 75 and 76% of TAGs and between 16 and 17% of DAGs. Fatty acid profiles, obtained by using GC-FID, indicated that saturated fatty acids (SFAs) between 19.5 and 19.9% of extracted oil, monounsaturated fatty acids (MUFAs) in the range between 43.5 and 44.9%, and PUFAs between 31.2 and 34.6% were extracted. p-Cymene and limonene extracts contained lower percentages than the other studied solvents of some PUFAs due probably to the fact that these unsaturated fatty acids are more susceptible to oxidative degradation than MUFAs. Ethyl acetate has been found to be the best alternative solvent to hexane for the extraction of salmon oil lipids. Graphical Abstract ᅟ.

No abstract

No abstract

Near-infrared (NIR) spectroscopy was evaluated as a rapid method for the determination of oleic, palmitic, linoleic and linolenic acids as well as omega-3, omega-6, and to predict polyunsaturated, monounsaturated and saturated fatty acids, together with triacylglycerides, diglycerides, free fatty acids and ergosterol in salmon oil. To do it, Partial Least Squares (PLS) regression models were applied to correlate NIR spectra with aforementioned fatty acids and lipid classes. Results obtained were validated in front of reference procedures based on high performance thin layer and gas chromatography. PLS-NIR has a good predictive capability with relative root mean square error of prediction (RRMSEP) values below or equal to 1.8% and provides rapid analysis without the use of any chemicals making it an environmentally friendly methodology.