Quantitative Evaluation of Chromatographic Data from Nonlinear Detectors and the Sulfur Flame-Photometric Detector

Abstract

The most reliable quantitative gas chromatography uses an internal standard and area calibration. This method is very accurate when the detector is linear; however, when the detector is nonlinear, severe errors may result. Theoretical and experimental study of the potential error due to detector nonlinearity was carried out, and the effect of fluctuations in the experimental conditions was evaluated for calibrations based on peak area (A), peak height (R_m), and a two parameter calibration (AR) which uses the combination A/R_m^((n−1)/n) as the variable where n is an empirical constant. In the theoretical development it was assumed that the shape of the peak of the concentration-versus-time curve is Gaussian and that the detector response is proportional to the nth power of the instantaneous concentration of sample in it. The model shows that calibration based on peak height will always depend on the experimental conditions. Calibration based on peak area will be independent of the experimental conditions if the detector is linear. Calibration based on A/R_m^((n−1)/n) will be always independent of the experimental conditions. The sulfur flame-photometric detector (FPD), which is inherently nonlinear, was used to derive sets of data on mixtures of organic sulfides and thiophenes. The dimensionless standard deviation (DSD) of the analysis of the same sets of data was compared for the three calibration methods. The two-parameter calibration yielded the smallest DSD when the same solution was analyzed in various conditions. Changes in the size of the sample, in the temperature, in the rate of temperature programming, and in the rate of flow of carrier least affected the DSD of the two-parameter method.