Detection of adulterants in dietary supplements with Ginkgo biloba extract by attenuated total reflectance Fourier transform infrared spectroscopy and multivariate methods PLS-DA and PCA

The infrared spectroscopy with attenuated total reflectance (ATR) sampling coupled with chemometric methods has been applied to non-destructive detection of adulterants in dietary supplements containing Ginkgo biloba extract. The sample set comprised the spectra of six drugs and sixteen dietary supplements with ginkgo leaf extract. Spectral data (900-1800 cm(-1)) were analyzed using multivariate partial least squares regression combined with a discriminant analysis (PLS-DA). The second derivative of spectra followed by mean centering was used as preprocessing method. Three models were constructed and validated for detection of potential adulterants: kaempferol, quercetin, and rutin. The iPLS-DA classification models achieved about 87.5%, 93,7%, and 87,5% of correct classification for adulteration with kaempferol, quercetin and rutin, respectively. The results obtained from classification models were verified by chromatographic fingerprints of unhydrolyzed sample extracts. Two-trace two-dimensional asynchronous correlation maps were constructed from pairs of spectra (each dietary supplement spectrum vs. averaged spectrum of drugs) and then analyzed by multiway PCA which revealed good discrimination between samples. (C) 2018 Elsevier B.V. All rights reserved.