Rapid Identification of Aliphatic Ketones from Vapor-Phase Infrared Spectra

The need for the identification of compounds present in complex gaseous mixture or after separation by gas chromatography (GC) using vapor-phase infrared (VPIR) spectrometry is becoming increasingly important. This is primarily due to advances in Fourier transform infrared (FT-IR) spectrometry allowing the acquisition of complete spectra in a few tenths of a second with detection limits of less than 1 ppb in long path infrared gas cells and below 10 nanogarms in the GC/FT-IR interface. With a GC/FT-IR system hundreds of spectra can be generated per chromato-gram. Clearly, the limiting processes are the spectral searching and, in cases where the search results are ambiguous, spectral interpretation steps. A powerful digital computer is needed to control the interferometer and perform the Fourier transformation from the time domain to the frequency domain in order to obtain the spectrum. This makes the incorporation of automatic spectral interpretation in the same computer an obvious and sensible capability. It should be kept in mind that the ultimate goal of the chemist is not simply to produce a collection of spectra but rather to determine what chemical compounds are present in the unknown samples and what their significance is to the problem at hand. The laborious task of spectral identification has been greatly simplified by the use of computer-based library searching routines. Two general classes of algorithms has been developed: procedures that seek to make an exact identification of an unknown by direct comparison with spectra of known compounds (which require the spectrum of the unknown to be in the reference databases), and more general approaches striving to identify all the functional groups in the unknown. A wide range of encoding schemes has been proposed 1 to represent spectra in a digital form suitable for library searching and interpretation. The ideal storage format retains the minimum amount of information necessary for correct identification of unknown compounds in a form permitting rapid numerical comparison. The optimum instrumental parameters for acquiring spectra for spectral searching and identification are a function of the sample, so that practical considerations require some degree of standardization in sampling and measurement. In addition, specific computer characteristics such as word size, amount of semiconductor memory, and types of mass storage devices must also be taken into account. Buechi et al. 2 have discussed the compilation of spectral libraries in terms of five operations: selection, digitization, completion, formatting, and verification. For infrared spectral reference data …