Near-infrared spectroscopy as a rapid method for evaluation physicochemical changes of stored soybeans

Abstract Storage conditions of grains affect compositional characteristics. Determination of these characteristics is time-demanding and requires analytical methods. The objective of the study was to investigate the potential application of near-infrared (NIR) spectroscopy technique as a process analytical technology (PAT) for quality assessment of soybeans (BRS 284 cultivar) stored under different conditions (25 and 35 °C) for 9 months. In addition, NIR spectral information of BRS 284 was compared with other cultivar (BRS 232) in order to verify its potential to discriminate these samples. Chemical composition, pH, free fatty acid content and color were determined by traditional methods. NIR spectral information in the wavelength range of 400–2500 nm were obtained from the stored soybean. Principal component analysis (PCA) was applied as an exploratory tool of spectral data. Loadings of the first 3 principal components were used to identify most important wavelengths influencing the spectra. At 25 °C, soybeans exhibited little or no significant changes in physicochemical properties; on the other hand, at 35 °C, there was a reduction on color and pH and increase on acidity. PCA scores of first 2 principal components showed that none of the spectral pre-processing treatment (SNV, 1st and 2nd derivative) provided discrimination between soybeans cultivars and detection of physicochemical changes of stored soybean. Results have shown the potential application of NIR spectroscopy as a PAT for rapid assessment of quality changes of stored soybean.