Modeling of solid fat content of chemically interesterified fully hydrogenated soybean oil and canola oil blends as a function of temperature and saturated fatty acids

Solid fat content (SFC) curve is one the most important physical properties of fats determining their plasticity and melting behavior. It is measured using pulsed nuclear magnetic resonance spectroscope or differential scanning calorimeter. However, the instruments may not be readily available at all laboratories. As a case study, in this research, application of empirical modeling for prediction of the SFC curve of chemically interesterified binary blends of fully hydrogenated soybean (FHSBO) and canola oil (CO) was evaluated. The SFC curve of interesterified fats as a function of temperature (SFCf(T)) or saturated fatty acids (SFA, SFCf(SFA)) were S-shaped and could be described using the sigmoidal Gompertz model. The SFCf(T) models fitted the experimental data with R(2)s greater than 0.95 and mean absolute errors (MAE) less than 1.61 %. The SFCf(SFA) curves could also be described using the Gompertz model with R(2)s greater than 0.97 and MAEs less than 1.18 %. Finally, a two-variable Gompertz model considering both temperature and SFA as variables was developed which fitted the experimental data with R-2 of 0.98 and MAE of 1.00 %. To validate the final model, the SFC curve of different blends including interesterified FHSBO/CO/sunflower oil, palm olein/CO or palm olein/CO/sunflower oil blends was predicted. The model could predict the SFC of interesterified FHSBO/CO/ sunflower oil blends with correlation coefficient of 0.98 and MAE of 1.02 %. When applied for palm olein-based blends, it predicted the SFC curves with correlation coefficient and MAE of 0.87-0.88 and 2.6-3.5 %, respectively.