Evaluation of a System for Generating Quantitatively Accurate Vapor-Phase Infrared Reference Spectra

A static delivery system for generating infrared vapor-phase reference spectra has been constructed and validated for liquid analytes with polarities up to that of methanol. Several important factors that determine the accuracy of these mixtures have been identified. These include thorough degassing of the liquid analyte, maintenance of water content inside the system below a predetermined value, isolation of the pure analyte vapor before inlet of the fill gas to avoid back-flow of vapor, forced mixing to overcome the effect of excessively slow diffusional mixing of the analyte vapor and fill gas, and measurement within the linear Beer's law range for the resolution and apodization employed. To maintain Beer's law linearity, it is vital to eliminate the effect of a nonlinear response of the mercury cadmium telluride detector. In this paper, the procedure developed to ensure quantitative accuracy from statically generated gas mixtures is described. The effect of detector response linearity on photometric accuracy is discussed, along with a method of empirically correcting for this nonlinear response. The accuracy of vapor-phase samples made by this approach has been validated by comparison to the spectra of certified standards and found to be within 2% of the certified value.