Time-series characterization of various honey types under different storage conditions based on total polyphenol content and fluorescence properties

Abstract

This study investigated fluorescence spectroscopy for evaluating honey quality changes during storage. Twenty-two honey varieties were stored at high (35 °C) and low (4 °C) temperatures for six months, with excitation-emission matrix (EEM) and total polyphenol content (TPC) measured every two months. Under high-temperature storage, TPC increased significantly while remaining stable at low-temperature. EEM measurements revealed five characteristic fluorescence peaks attributed to various compounds including amino acids, flavonoids, phenolic acids and Maillard reaction products. Using principal component scores obtained from principal component analysis (PCA) dimensionality reduction, support vector machine (SVM) classification achieved 81.82 % accuracy in distinguishing between early storage periods and late storage periods for high-temperature samples, while maintaining 59.09 % accuracy for low-temperature samples. Partial least squares regression (PLSR) models constructed using EEM data demonstrated robust TPC prediction capability with R²cv of 0.92, root mean square error cross validation (RMSECV) of 40.66 μg gallic acid equivalent/g and residual prediction deviation (RPD) of 3.61. Variable importance in projection (VIP) analysis indicated that fluorescence regions associated with flavonoids, phenolic acids and Maillard reaction products significantly contributed to TPC prediction. These findings demonstrate the potential of fluorescence spectroscopy as a non-destructive method for evaluating honey quality changes during storage, particularly under high-temperature conditions.