Quantitative analysis and geographical traceability of black tea using Fourier transform near-infrared spectroscopy (FT-NIRS)

Abstract

Near-infrared spectroscopy (NIRS) combined with chemometric tools was utilized as a rapid analysis method to assess quality and to differentiate geographical origins of black tea. A partial least squares (PLS) algorithm was employed for the calibration of models predicting the levels of caffeine, water extract, total polyphenols, and free amino acids, while a factorization method was proposed to trace black tea from different geographical origins. In the calibration set, the root mean squared error of cross validation (%) and the correlation coefficient (R) for caffeine, water extracts, total polyphenols and free amino acids were 0.102%, 0.654%, 0.552%, and 0.248% and 0.983, 0.977, 0.975, and 0.943, respectively. In the prediction set, the root mean squared error of prediction and R for the corresponding constituents were 0.160%, 0.685%, 0.594%, and 0.273% and 0.955, 0.962, 0.954, and 0.927, respectively. The identification accuracy for black tea from different geographical origins reached 94.3%. This study demonstrated that NIR spectroscopy can be successfully applied to rapidly determine the main chemical compositions and geographical origins of black tea.