Width-Enhanced Binary Representation for Library Searching of Vapor Phase Infrared Spectra

Abstract
A method is presented for the representation of vapor phase infrared spectra in a binary format which preserves both width and position information about each peak. Preselection of an optimally width-enhanced library of binary spectra, in advance of any actual library searching, is approached through the application of information theory and Grotch's method for prediction of matching histograms. A direct comparison of library searching results is conducted to complement the preselection efforts. The applicability of a channel-combination approach based on information theory is demonstrated. By this method, adequate library searching is achieved after a 25-fold reduction in the number of channels needed to describe each vapor phase infrared spectrum.