Optimization and Validation of a Modified QuEChERS-UPLC-MS/MS Method for Multi-Residue Pesticide Analysis in Tea and Rosemary Leaves with Experimental Uncertainty Assessment

This study developed and validated a modified QuEChERS extraction procedure for the determination of 32 pesticide residues in tea (Camellia sinensis) found in Moroccan market originating from China and rosemary (Salvia rosmarinus) leaves using UPLC-MS/MS. A total of 202 pesticide residues were initially screened using the NF EN 15662 method, and 32 compounds showed satisfactory recovery and precision in tea. The optimized method, based on freezing before extraction and using PSA + BCG as purification sorbents, achieved recoveries of 70–120% for 97% of analytes. Calibration linearity ranged from 0.001 to 0.2 mg/kg (R2 ≥ 0.980), with spiking levels at 0.01–0.1 mg/kg analyzed in duplicate per level and per matrix. Results of LOQ and LOD are compared to multiple MRL databases of Tea (EU, USA, Canada, Codex Alimentarius) despite the scarcity of tea MRLs in most databases, providing an accurate representation of the complexity of the analysis. Uncertainty was evaluated through both validation-based and Horwitz-based approaches, allowing a robust comparison of theoretical and experimental variability. The optimized method was also applied to rosemary, where 31 of the 32 pesticides validated in tea showed satisfactory performance. Similar recoveries, precision, and LOQs were obtained for both matrices, demonstrating the robustness of the method across chemically diverse herbal substrates. Only carbofuran exhibited reduced performance in rosemary, requiring an increased functional LOQ because of matrix-induced ion suppression. This dual-matrix approach provides a reliable and reproducible method for pesticide monitoring in herbal products. This dual-matrix approach provides a reliable and reproducible method for pesticide monitoring in herbal products.