Can NMR spectroscopy discriminate between NPS amphetamines and cathinones? An evaluation by in silico studies and chemometrics

The use of new psychoactive substances (NPS) has become a problem that affects many countries. Because new chemical structures appear frequently and reliable reference data do not always exist, identifying these compounds is an operational problem. Theoretical chemistry associated with statistical methods has scarcely been used to tackle such problems and can provide an unvaluable aid. In this study, we have employed quantum chemistry tools to simulate NMR spectra and to evaluate, as a paradigm, the main differences between amphetamines and cathinones. We have generated and examined the 1H and 13C chemical shifts of 21 homologous amphetamine and cathinone structures. In our computational approach, we used a test set and a proof set to generate the shifted factors on the basis of experimental observations. We employed three different DFTs (B3LYP, M062X, and PBE0) with a TZVP basis set and different solvents (water, ethanol, chloroform, and gas-phase). To evaluate the data, we applied statistical approaches, which allowed us to assess the methods and to distinguish between the studied NPS. This study provided reliable information that helps to explain how NMR evaluation differs in assessing these molecules and promoting their discrimination in forensic analysis.