Rapid Classification and Identification of Benzodiazepine Drugs Utilizing Raman Spectroscopy Combined With PCA‐SVM Algorithms

ABSTRACT Rapid and accurate identification of benzodiazepine drugs is of great significance for pharmaceutical quality control and law enforcement supervision. In this study, a portable 785 nm Raman spectrometer and a handheld 1064 nm Raman spectrometer were first employed to collect the characteristic spectra of five benzodiazepine drugs, and the performance of the two spectrometers was compared. The results showed that compared with the portable 785 nm Raman spectrometer, the handheld 1064 nm Raman spectrometer exhibited a flatter background and clearer characteristic peaks. This method requires no sample pretreatment and enables rapid on‐site identification of benzodiazepine drugs in only 3–10 s. In addition, by combining principal component analysis (PCA) and support vector machine (SVM) algorithms, 100% accurate intelligent classification was achieved for five benzodiazepine drugs, as well as estazolam tablets and diazepam tablets from two different manufacturers. Through blind tests covering single drugs, interfering drugs, and diazepam–estazolam binary mixed drugs, the algorithm achieved an identification accuracy of 100% for both single drugs and interfering drugs and a classification accuracy of 93.3% for mixed drugs with different ratios—findings that further validated the feasibility of the PCA‐SVM algorithm. This method combines the advantages of rapidity, accuracy, nondestructiveness, and simplicity and is expected to provide technical support for the supervision, on‐site law enforcement, and quality monitoring of benzodiazepine drugs.