Detection of lethal fake liquors using digitally labelled gas-phase Fourier transform infrared spectroscopy

Abstract

Fake liquors have caused severe body injuries or even deaths worldwide, rapid detection of such lethal drinks is thus quite necessary. Methanol has been identified as a primary cause of the problem, so methanol monitoring is critical to the detection of fake liquors. The present work provides an effective strategy for rapid detection of different lethal fake liquors. Using gas-phase Fourier transform infrared spectroscopy in a digitally labeled approach, the spectral bands of methanol were extracted by the iterative discrete wavelet transform for classification, which is named as digitally labeled gas-phase Fourier transform infrared spectroscopy. In digitally labeled gas-phase Fourier transform infrared spectroscopy, principal component analysis and least square support vector machine were combined to discriminate problematic samples using the iterative discrete wavelet transform filtered signals. As a result, the method could cleverly extract spectral features of methanol from the alcoholic drinks in the presence of uncontrolled matrix effects. The recognition accuracy was higher than 97.0%, and each measurement was done within 3 min. The results illustrate that the digitally labeled gas-phase Fourier transform infrared spectroscopy method serves well to rapidly discriminate fake liquors as an efficient and promising tool, and could be well extended to detection of any other targeted volatile substance in complicated systems.