October 9 - 13, 2005, FACSS 2005, Quebec City, PQ, Canada

Applying Self-Modeling Mixture Analysis to Simplify Vibrational Interpretations

Michel Hachey, Michael Boruta

Abstract

With the rising popularity of in-situ real time reaction monitoring equipment, it is becoming more common for analysts to make vibrational interpretations of mixtures. While analysts will usually do everything in their power to avoid the infrared spectral interpretation of mixtures, it is not practical in this instance to depend on sample purification or chromatographic separations. And while spectral subtraction approaches are quick and simple, they require prior knowledge of the interfering components which can be missing for novel reactions. In absence of such knowledge, deconvolution and second derivatives are often used as a last resort to help overcome the overlap band problem. However, these latter approaches do not really help correlate spectral features to specific components. In this talk, we will discuss how multivariate mixture analysis can be applied to simplify vibrational interpretation of mixtures by extracting pure component spectral profiles from evolving mixtures.

This describes an alternative approach to simplifying vibrational interpretation of complex sample mixtures via curve resolution techniques.

Download the presentation in MS PowerPoint (1.62 Mb ZIP file) or Adobe Acrobat format (813 Kb PDF file).

Relevant Product Line: Optical Spectroscopy