Fourier Transform near Infrared Spectroscopy Prediction of trans and cis Fats in Ground Cereal Products at Different Resolutions

Abstract

In this study, Fourier transform near infrared (FT-NIR) spectroscopy was investigated for use as a tool to determine the trans and cis fat contents in ground cereal products without the need for further oil extraction. To compare the calibration results obtained at different resolutions, the near infrared (NIR) spectra of samples were obtained using an FT-NIR spectrometer at resolutions of 4 cm−1, 8 cm−1 and 16 cm−1. Fat contents of samples were determined using a gas chromatography method. Generally, higher resolution provides better predictions for all types of fats. Each type of fat had its own optimum resolution: 4 cm−1 for trans and 8 cm−1 for cis fat models. At optimal resolution, the models predicted trans and cis fat contents with a SEP and r2 of 0.75% and 0.96, and 0.70% and 0.96, respectively. The results indicated that the trans and cis fat content of cereal products could be determined in minutes without the need for oil extraction within the accuracy required for sample screening ( RPD=4.3 or 4.8).