建立了可用于甲基苯丙胺、 氯胺酮、 海洛因、 可卡因快速定性鉴定分析的衰减全反射-傅里叶变换红外光谱(ATR-FTIR)方法, 采用特征吸收峰作为定性判别依据。 长期以来, 由于缺乏合适的定性判别依据, 红外光谱法仅被用于毒品的快速定性筛查分析。 为扩大红外光谱法在禁毒领域的应用, 收集并分析了1 380份各类缴获毒品样品, 对特征吸收峰法和簇类独立软模式法(SIMCA)两种定性判别方法进行了考察和比较。 在标准品化合物红外光谱图的2 500~650 cm-1范围内挑选5~8个相对吸收强度较高且不受常见掺杂物干扰的吸收峰作为特征吸收峰。 采用特征吸收峰法(全部特征吸收峰均检出)作为定性分析依据时, 646份验证样品的阳性检出率为98.1%。 采用516份建模样品建立了不同盐型毒品的SIMCA定性判别模型, 并对646份验证样品进行了定性分类, 总识别率为95.4%, 拒绝率为100%。 结果表明, 特征吸收峰法和SIMCA法均具有专属性强、 定性结果准确可靠的特点; 但特征吸收峰法操作简单, 即使是没有红外光谱知识背景的一线民警在经过简单培训后也可掌握, 利于推广和普及; 而SIMCA模型的建立需要大量有代表性的建模样品和专业的数理统计软件, 推广和普及难度较大。 基于特征吸收峰的ATR-FTIR法将极大提高我国毒品样品定性检验的鉴定效率、 降低检验鉴定成本。

For the first time, this study has established an attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR) method for fast qualitative analysis of methamphetamine, ketamine, heroin, and cocaine. Characteristic peak method was chosen as the qualitative analysis criterion of this method. Due to the lack of proper qualitative identification criterion, FTIR method has long just been used for fast qualitative screening in drug analysis. In order to expand its application in forensic sciences, two qualitative identification criteria of characteristic peak method and soft independent modeling by class analogy (SIMCA) classification method were investigated and compared. A total of 516 calibration and 864 prediction samples were used for method development and validation. For characteristic peak method, five to eight peaks in the range of 2 500~650 cm-1 with relative high intensity and interference-free from common cutting agents were selected as the characteristic peaks. When characteristic peak method (all of the characteristic peaks should be detected) was used for the qualitative analysis of 646 validation samples, the positive detection rate was 98.1%. A SIMCA classification model was constructed based on 516 calibration samples, consisting of 7 PCA models for methamphetamine HCl, ketamine HCl, ketamine base, heroin HCl, heroin base, cocaine HCl, and cocaine base. The classification results of 646 validation samples showed a recognition rate of 95.4%, and a false rejection rate of 100%. Therefore, the results of both the characteristic peak method and the SIMCA method were reliable and accurate. It is easy for the characteristic peak method to be used by local police officers after simple training, thus suitable for popularizing. In comparison, as the building of SIMCA model needs the collection of large number of representative calibration samples and the use of professional mathematical software, the SIMCA method is not suitable for popularizing. Therefore, the characteristic peak method was chosen as the criterion method for fast qualitative identification analysis. The ATR-FTIR method that is based on characteristic peak method will greatly increase the efficiency and reduce the cost of drug qualitative identification analysis.