Evaluating Open-Path FTIR Spectrometer Data Using Different Quantification Methods, Libraries, and Background Spectra Obtained under Varying Environmental Conditions

Studies were performed to evaluate the accuracy of open-path Fourier Transform Infrared (OP-FTIR)spectrometers using a 35 foot outdoor exposure chamber in Pittsboro, North Carolina. Results obtained with the OP-FTIR spectrometer were compared to results obtained with a reference method (a gas chromatograph equipped with a flame ionization detector, GC-FID). Concentration results were evaluated in terms of the mathematical methods and spectral libraries used for quantification. In addition, the research investigated the effect on quantification of using different backgrounds obtained at various times during the day. The chemicals used in this study were toluene, cyclohexane, and methanol; and these were evaluated over the concentration range of 5-30ppm. Two different quantification methods were used in the analyses to calculate concentrations and were performed manually using Grams 386 software (Galactic, Inc.): the traditional peak area method (which involved calculating the area of the peaks utilizing the trapezoidal rule of integration) and an interactive subtraction method. The different spectral libraries used for quantification included: the Hanst library(137ppm-m toluene, 388 ppm-m cyclohexane, and 27 ppm-m methanol), and the EPA library (249 ppm-m toluene). An EPA library spectrum for cyclohexane and methanol was not available. To investigate the effect on quantification of using different backgrounds, concentration data files were analyzed using the backgrounds taken at the start of the experiments, and subsequent backgrounds obtained later the same day. The backgrounds contained varying amounts of water vapor and other atmospheric constituents. Using the Hanst library for quantifying toluene, there was no significant difference between open-path and GC-FDD concentrations, regardless of the method of quantification, for concentrations of 10 ppm and below. However, for concentrations between 10 ppm and 30 ppm, there was a significant difference between open-path and GC-FBD concentrations. Using the EPA library for quantifying toluene, there was no significant difference between open-path and GC-FID measurements, regardless of the method of quantification, for concentrations of 15 ppm and below. However, for concentrations between 15 ppm and 30 ppm, there was a significant difference between open-path measurements analyzed by the subtraction method and GC-FED concentrations. There was no significant difference between GC-FID concentrations and open-path measurements quantified by the peak area method for all concentrations studied. For cyclohexane, all concentrations were significantly higher than the reference method, regardless of the quantification method. There was no significant difference between concentrations obtained using the two quantification methods for concentrations of 5 ppm and below. A regression through the origin was performed on the cyclohexane peak area and subtraction curves and each respective slope was compared to the slope of the GC-FID curve. When a ratio factor for the slope of the cyclohexane peak area and subtraction curves was applied to the open-path measurements, there was no significant difference between the GC-FID measurements and open-path FTIR concentrations. Water vapor was used as a marker to characterize the backgrounds obtained each day. Each data file was analyzed with the original and subsequent backgrounds from the same day to study the effect on quantification. The magnitude of the effect varied among chemicals as follows: (1) In general, wetter and dryer surrogate toluene backgrounds significantly increased concentrations. (2) For cyclohexane and methanol, wetter surrogate backgrounds significantly increased concentrations and dryer surrogates significantly decreased concentrations. (3) Differences in water vapor partial pressures of less than 5% resulted in quantification results being significantly different than the results obtained with the original background for all three chemicals.Master of Science in Public Healt