Gas Chromatographic/Infared Spectral Analysis Using a Computer Integrated Circular Variable Filter Spectrometer

Gas chromatography/infrared spectroscopy (GC/IR) is a powerful analytical technique for "on-the-fly" analysis of multi-component mixtures. The gas chromatograph separates the mixture into single components which are then analyzed by the spectrometer as they elute off the gc column and into the spectrometer. Standard instrumention usually involves a gas chomatograph, a fourier transform infrared spectrometer (FT-IR), and a computer system. The end products are a reconstructed gas chromatogram and a number of infrared spectra for each of the mixture components. These pieces of information often allow the identification of an unknown mixture. A CC/FT-IR/Computer System, however, is an elaborate and expensive laboratory instrument which requires extensive care, maintenance, and a well trained operator. It is not designed for field use and is seldom deployed outside of a laboratory. The objective of this research therefore was to develop a simple and inexpensive CC/IR/Computer System which would be rugged, easy to use, and readily adaptable for field use. To do this, we interfaced a fast scanning circular variable filter (CVF) infrared spectrometer to a mini-computer and then coupled the spectrometer/computer to a gas chromatograph. The system was configured such that the same end products generated by a CC/FT-IR/Computer System, i.e. reconstructed gas chromatograms and infrared spectra, are also generated by this GC/CVF-IR/ Computer System. A five component mixture of hazardous class type compounds with closely related boiling points was analyzed to demonstrate the feasibility of this new system.

were in the mixture. The individual spectra for each of the reconstructed g.c. peaks can then be displayed as a function of scan number. The spectra can also be signal averaged, smoothed, and subtracted. Figure 3 shows a typical sequence of events for this analysis.

Samples
Mixtures were prepared by mixing equal volumes of reagent grade acetone, chloroform, methanol, carbon tetrachloride, and benzene in carbon disulfide.

Software
Eight computer programs were written for this system. By using GCIRSUB: A subroutine used by GCIRAD to initiate data acquisition and store data on disk.
ADGCIR: An assembly language subroutine to sample the ADC.
GCIRCAL: Calibrates all the individually acquired IR spectra using the 0 VDC signal between filter elements one and three.
TWIRP: Plots out the reconstructed gas chromatograms.
8 GCIRSPEC: Plots out individual spectra corresponding to peaks from the recontructed gas chromatograms.
GCIRSUM: Signal averages any number of IR spectra.

RESULTS AND DISCUSSION
We want to show that this GC/IR system using a computer interfaced  can be easily seen. Of course the conpounds need to have IR active groups to cause a signal on the IR detector. In this initial case, after sixty scans, there is one 'IWIRP peak starting at scan #25. Figure 5 sho.vs the IR spectra associated with the scan numbers 22,24,25,26, and 28 plotted as a function of IR intensity vs wavenumber. Scan #22 is instrurrent background and since this is a single beam spectrometer the spectrum is due to ambient 002 and H20. At scan #24 ho.vever a distinctly different spectrum is seen and its intensity continues to increase at scan #25. At scan #28 the sample has conpletely eluted from the g.c.
column thru the cell and all that is left is ambient background.     Again as done previously the spectra associated with each reconstructed g.c. peak are plotted out as a function of % transmission vs wavelength ( figure 7). Scan #1 is the instrument background and the subsequent scans correspond to the five g.c. peaks. Figures 8, 9, 10, 11, 12, and 13 are the spectra displayed for more detailed analysis.   During this investigation we also conducted a comparision between the GC/IR TWIRP and a standard gc flame ionization detector (FID). The same five component mixture used previously was analyzed with the gc column reconnected to the FID as in normal gc operation. Figure 14 shows the comparision between the normal gc output and the TWIRP from While checking the reproducibility of the CVF scan and data acquisition we decided to try signal averaging the spectra to decrease the noise in a plotted spectrum. Figure 15 shows the original spectrum and an averaged spectrum where four spectra were added together and then signal averaged. As can be seen there is a noticable reduction in noise (the original spectra were made noisy on purpose by running at a higher than normal amplifier gain). Also a three point smooth   Purpose: This is a fortran subroutine used by GCIRAD.FR to initiate data acquisition and write the data onto disk.
Method: Based on the inputs from GCIRAD.FR specifying the file name and the length of time for data acquisition this program will run an assembly language subroutine, ADGCIR.SR, to sample the ADC and write the data out to a disk.
Output: Contiguous data blocks on disk. c