Principal component analysis to enhance enantioselective Raman spectroscopy

Abstract

Enantioselective Raman (esR) spectroscopy is an innovative technique with a high potential for online process monitoring in chiral media, e.g. in the pharmaceutical industry. A prerequisite for an effective application is to combine the experimental approach with suitable concepts for data analysis. In this work, we present a chemometric approach to analyze the esR spectra recorded in an automatized polarization-resolved Raman set-up. It is demonstrated that the proposed method is capable of distinguishing between the enantiomers of the chiral alcohol 4-methylpentan-2-ol in a fully unsupervised fashion. Furthermore, it is shown that the difficulty of facing only small intensity differences between the esR spectra of the enantiomers can be overcome by feeding difference spectra between the pure enantiomers and the racemate into the principal component analysis (PCA) algorithm. The enantiomers are clearly discriminable along the first principal component.