Abstract:In this study, a rapid method for detecting tetrahydrocannabinol (THC) content in beverages using a probability of detection (POD) model combined with surface-enhanced Raman spectroscopy (SERS) technology was established. Two pre-treatment schemes, direct dilution method and extraction method, were developed for different sample matricessuch as carbonated beverages, fruit and vegetable juices, and vitamin beverages. The experimental results showed that the Raman characteristic peaks at 542 cm-1 and 1 172 cm-1 were used as the peaks for identifying THC. In combination with the POD model, the detection limit (LOD) of tetrahydrocannabinol was determined to be 1 mg/kg in carbonated beverages and other matrices, and 2 mg/kg in fruit and vegetable juices. In the experiment, multiple-concentration-gradient positive samples (0~10 mg/kg) were designed for SERS detection, and a qualitative POD model was established to determine the change in detection probability with mass concentration. The sensitivity of this method was greater than or equal to 95%, with specificity greater than or equal to 90%, false negative rate lower than or equal to 5%, and false positive rate lower than or equal to 10%. Meanwhile, this rapid detection method and the reference method (the method by liquid chromatographytandem mass spectrometry) were analyzed through one-time evaluation. When the concentration of added THC was not lower than LOD, the POD values of this SERS method and the reference method were both higher than 95%, indicating that the two methods showed good consistency. This rapid detection method is easy to operate, fast and accurate, and requires only 30 minutes from sample pretreatment to measurement results. In combination with the POD model, the effectiveness of this method was verified. This method is expected to be applied for rapid on-site detection of THC in customs ports, market supervision and other fields.
