Quantitative determination of physical and chemical measurands in honey by near-infrared spectrometry

Fourier transform near-infrared spectroscopy (FT-NIR) was evaluated to quantitatively determine 24 different measurands in honey. The reference values of 421 honey samples of different botanical origins were determined by classical physical and chemical methods. Partial least squares regression was used to develop the calibration models for the measurands studied. These calibrations were then validated using independent samples and proved satisfying accuracies for the determination of water (standard error of prediction: 0.3 g/100 g), glucose (1.3 g/100 g), fructose (1.6 g/100 g), sucrose (0.4 g/100 g), total monosaccharide content (2.6 g/100 g) as well as fructose/ glucose ratio (0.09) and glucose/water ratio (0.12). The prediction accuracy for hydroxymethylfurfural, proline, pH-value, electrical conductivity, free acidity and the minor sugars maltose, turanose, nigerose, erlose, trehalose, isomaltose, kojibiose, melezitose, raffinose, gentiobiose, melibiose, maltotriose was poor and unreliable. The results demonstrate that near-infrared spectrometry is a valuable, rapid and non-destructive tool for the quantitative analysis of some measurands related to the main components in honey.