Synchronous fluorescence spectroscopy for detecting adulteration and quantifying eugenol in blueberry spirit

This study investigates the feasibility of synchronous fluorescence (SF) spectroscopy as a rapid and direct method for the authentication and quality control of blueberry spirits. SF spectral data were analyzed using chemometric methods to both classify samples and quantify adulteration levels. For authentication purposes, unfolded total SF spectra were processed using principal component analysis combined with linear and quadratic discriminant analysis (PCA-LDA and PCA-QDA) to classify samples as authentic, ethanol-adulterated, or water-adulterated. Among the tested models, PCA-QDA achieved the best performance, with a balanced accuracy of 96.7%. Quantification of adulteration was carried out using partial least squares (PLS) regression applied to unfolded total and selected single-offset SF spectral data. Optimal PLS models provided accurate predictions of ethanol adulteration (coefficient of determination for prediction (R²P) = 0.96, root mean square error of prediction (RMSEP) = 0.6%, relative predictive deviation (RPD) = 3.0) and water adulteration (R²P = 0.97, RMSEP = 0.5%, RPD = 4.5). In addition, SF spectroscopy coupled with PLS regression was successfully applied to predict eugenol content in blueberry spirits, using reference values obtained by HPLC with fluorescence detection. The best model achieved R²P = 0.98, RMSEP = 0.1 mg/L, and RPD = 5.0. These results demonstrate the strong potential of SF spectroscopy as a fast, reliable, and cost-effective tool for the authentication and quality control of blueberry spirits, offering an attractive alternative to more complex and time-consuming analytical methods.