a Multivariate Statistical Approach to Identifying Organic Compounds Using AN Oscillating Plasma Glow Discharge Detector for Gas Chromatography.
An oscillating plasma glow discharge detector for gas chromatography is used to obtain fingerprint information about an analyte by combining both the average cell current and oscillation frequency signals. Five homologs each of the n-alkanes, 1-alkenes, 1-alkynes, 2-ketones and aldehydes are studied. Although triplicate determinations had some scatter due to noise, they showed clustering that allows several of these compounds to be distinguished from the others by using a two-dimensional plot of the ratios of frequency peak area to current peak area and frequency peak height to current peak height. Fingerprint identification information is improved by changing the cell pressure, applied voltage and electrode spacing. Changes in the discharge operating conditions produce changes in the analyte peak responses. The relative magnitudes of the analyte current and frequency peak responses also change with respect to each other under different discharge conditions. Unique fingerprints or patterns of responses are created for each analyte by changing the discharge operating conditions. The detector responses toward 10 organic compounds, representing seven different functional groups, are recorded under 56 different combinations of discharge conditions. The ratios of the frequency to current peak responses (heights and areas) for three of the 56 sets of conditions investigated provide enough information to distinguish between nine compounds. Principal component analysis and hierarchical cluster analysis, multivariate exploratory techniques, are used to observe natural clustering in the data.
- Pub Date:
- Chemistry: Analytical; Physics: Fluid and Plasma; Computer Science