Determination of nonvolatile components of heated soybean oils separated with high-efficiency mixed-bed polystyrene/divinylbenzene columns

Author： Abidi S. Kim I. Rennick K.

Publisher： Springer Publishing Company

ISSN： 0003-021X

Source： Journal of the American Oil Chemists' Society, Vol.76, Iss.8, 1999-08, pp. : 939-944

Disclaimer: Any content in publications that violate the sovereignty, the constitution or regulations of the PRC is not accepted or approved by CNPIEC.

Previous Menu Next

Abstract

Whole heated soybean oils and their polar fractions were analyzed for nonvolatile components by high-performance size-exclusion chromatography (HPSEC) with evaporative light scattering detection (ELSD). High molecular-weight (MW) polymer compounds with MW ≥ trimer were efficiently separated with new 3-µm mixed-bed styrene/divinylbenzene copolymer columns. Peaks of high MW polymer components in the new column system appeared to be sharper and more symmetrical than those obtained with other columns. In the model systems studied, continuous addition of water to partially simulate frying conditions resulted in a significant increase (up to 30%) in the polar lipid content of the heated oils evaluated. Due to relatively high concentrations of monomeric triglycerides (84.6–93.5%) present in the whole unfractionated oils, small but erratic variations in the compositional distribution of components were observed in oils containing different amounts of added water. On the other hand, HPSEC-ELSD analyses of the polar fractions (monomeric triglycerides, 25.4–62.6%) showed significant changes in the content and composition of nonvolatile components with the amount of water added. In general, prolonged heating with increasing amounts of water accelerated hydrolysis and polymerization of heated soybean oils. Discrepancies in total polymerization of heated soybean oils. Discrepancies in total polymeric materials obtained from HPSEC composition data for whole oils and polar fractions are discussed in terms of nonuniformity in sample matrices, detection limitations for minor components, and a nonlinear ELSD response rationale.