Potential of high-throughput FIA-MS/MS and LC-MS/MS polyphenolic profiling to assess tea authenticity. Application to tea adulterations with chicory

• High-throughput FIA-MS/MS polyphenolic profiles to assess Tea Authenticity. • Excellent discriminations of different tea varieties against chicory by PLS-DA. • Discriminant polyphenols on each tea class and chicory were tentatively annotated. • Quantitation of adulteration in tea by FIA-MS/MS with prediction errors below 10.9% Tea can be found among the beverages more susceptible to fraudulent practices because of its high worldwide consumption and the increases on prices for some specific varieties due to climate change and geopolitical instability. Tea adulteration with other plants, such as chicory, is a common practice to gain an illicit profit. Polyphenols are abundant bioactive substances in tea, determining its quality and health function. In addition, they can be employed as markers to address authentication issues. The present contribution assesses the potential of polyphenolic profiling by high-throughput FIA-MS/MS and LC-MS/MS methodologies for tea authenticity. One hundred tea samples belonging to different varieties (green, black, red, oolong, and white teas) and 20 chicory samples were analyzed with both methodologies after a simple brewing process to profile fifty-five polyphenols belonging to different families. The resulting chemical descriptors were used to address tea classification and authentication by partial least-squares-discriminant analysis (PLS-DA). An excellent classification performance by PLS-DA was accomplished, with sensitivity and specificity values for FIA-MS/MS higher than 90% and 88.9%, respectively, and for LC-MS/MS higher than 85% and 86%, respectively. Good accuracy was also attained, with calibration errors below 10.5 and 14.5% for FIA-MS/MS and LC-MS/MS, respectively. Overall, FIA-MS/MS showed a better performance than LC-MS/MS, with the additional advantage of shorter analysis time as no chromatographic separation was required. The capability of phenolics to quantify tea adulterations with chicory was also assessed by partial least squares (PLS) regression, with prediction errors below 10.9 and 14.8% for FIA-MS/MS and LC-MS/MS, respectively, in the determination of adulterant levels. Thus, both methodologies demonstrated to be feasible for assessing tea authentication issues.