Qualitative analysis using Raman spectroscopy and chemometrics: a comprehensive model system for narcotics analysis.

摘要: The rapid, on-site identification of illicit narcotics, such as cocaine, is hindered by the diverse nature samples, which can contain a large variety materials in wide concentration range. This sample variance has very strong influence on analytical methodologies that be utilized and general prevents widespread use quantitative analysis narcotics routine basis. Raman spectroscopy, coupled with chemometric methods, used for situ qualitative narcotics; however, careful consideration must given to dealing extensive types. To assess efficacy combining spectroscopy chemometrics target analyte under real-world conditions, large-scale model system (633 samples) using (acetaminophen) mixed excipients was created. Materials exhibit problematic factors fluorescence, variable scattering intensities, peak overlap were included challenge data preprocessing classification methods. In contrast spectral matching approaches, we have taken model-based approach account variances data. first derivative spectra from fingerprint region (750-1900 cm(-1)) yielded best classifications. Using robust segmented cross-validation method, correct rates better than ∼90% could attained regression-based classification, compared ∼35% SIMCA. study demonstrates even high degrees variance, evidenced dramatic changes spectra, it possible obtain reasonably reliable combination chemometrics. set now validate more advanced or machine learning algorithms being developed analytes narcotics.