Hierarchical cluster analysis applied for the automated recognition of psychoactive substances and of their main precursors

Abstract

Ephedrine is a psychoactive substance abused for its stimulant effect. In addition, ephedrine, as well as its isomers and analogues are controlled substances as they are the main precursors of more potent stimulants, i.e. amphetamines. We are presenting a Hierarchical Cluster Analysis (HCA) application built for the automated recognition of ephedrines based on their spectra recorded with a portable infrared laser spectrometer operating in the 1405–1150 cm−1. The spectra have been preprocessed with a feature weight wE, which emphasizes the most relevant absorptions. The class identity of each compound was then assigned by using HCA, based on the scores obtained by subjecting the preprocessed spectra to Principal Component Analysis (PCA). Despite the narrowness of the spectral window in which the spectra are recorded, the system is very sensitive, as it detects all the controlled ephedrines. It is also quite selective, as it distinguishes ephedrines from other illicit phenethylamines, such as stimulant and hallucinogenic amphetamines.