Highlighting the Potential of Attenuated Total Reflectance - Fourier Transform Infrared (ATR-FTIR) Spectroscopy to Characterize Honey Samples with Principal Component Analysis (PCA)

Honey is a natural product of high nutritional value and of particular scientific and economic importance with unique sensory characteristics. Honey samples from different Greek regions and botanical origin were subjected to attenuated total reflectance - Fourier transform infrared spectroscopy (ATR-FTIR) to obtain the absorption fingerprint. Moreover, honey samples were analyzed spectrophotometrically to confirm and compare their in vitro antioxidant properties. Honeydew honeys presented significantly higher total phenolic content and antioxidant activity than the blossom varieties. Additionally, the FTIR spectra transformation revealed several bands, important for honey characterization based on geographical parameters (location, altitude), as well as for their botanical origin and individual sugar ratio. Honey classification according to certain FTIR bands was also evident utilizing a combination of statistical tools. Particularly, statistical analysis by means of the Python scipy.stats library, discriminant analysis by applying machine learning algorithms, and pairwise correlation were performed to FTIR band intensities for determining the optimum feature combinations to discriminate the geographical and botanical origins. Based upon the statistical analysis, carboxylic acids, amino acids, and the spectral region related to anomeric forms of carbohydrates or ring vibrations emerge as good indicators of honey geographical and botanical origins. The origins of the honey samples were predicted with high accuracy (93.55% and 96.77%, respectively) by applying principal component analysis (PCA). The results highlight the potential of FTIR coupled with spectral transformation and statistical analysis to evaluate honey origin and chemical profile.