Fourier transform infrared spectroscopy and chemometrics for chemical property prediction of chemically interesterified lipids with butterfat and vegetable oils during storage

The storage stability of structured lipids has a great importance for food industry and should be defined clearly. The aim of the study is to manufacture structured lipids by chemical interesterification of butterfat with different vegetable oils, to determine some chemical properties of interesterified lipids throughout storage and to predict free fatty acid, peroxide value, malondialdehyde and mono-di-triacylglycerol contents of these lipids by using middle infrared spectroscopy (FT-MIR). Prediction models were constructed by using partial least square (PLS) regression with an external cross-validation. The PLS model for TAG with FT-MIR data showed an excellent predictive potential with higher R2 cal=0.98, R2 cv=0.99 and lower RMSE values. The model for monoacylglycerol content (MAG) showed a good predictive ability with higher R2 cal=0.88, R2 cv=0.90 and lower RMSE values. The PLS model constructed with FT-IR spectra for diacylglycerol (DAG) content have good capability of prediction due to higher R2 and lower RMSE values. For the prediction of malondialdehyde (MAD) of interesterified lipids by FT-MIR spectra, regression coefficient of calibration set was found as 0.84. For the prediction of peroxide value (PV) of interesterified lipids with FT-MIR spectra, R2 CV=0.76 and RMSEC=1.52 and RMSECV=1.15. Infrared spectroscopy technique could be used for analyzing the chemical changes of fats during storage and also suggests a rapid and non-destructive techniques as good alternatives to the traditional analytical methods.