Abstract:
This dissertation contains analyses of new methods for determining the concentration and temperature of atmospheric gases from the data generated by an open-path Fourier transform (OP-FTIR) spectrometer. The concept of FTIR and the subset OP-FTIR are explained in terms of the physical instrumentation and the traditional Beer-Lambert Law based absorbance quantification. The important problems of background collection and water vapor interference with target gas features are also introduced. The process of synthetic spectrum generation is the foundation for this work and is described in detail. The inputs that are required to model the physics of the absorption of infrared radiation by small molecules are explained. The effects that each input has on the final spectrum as recorded by the OP-FTIR are also discussed at length. Also described is the modeling of the optics of the OP-FTIR instrument. Particular attention is paid to the temperature effects on the spectrum of the most important atmospheric infrared absorber, water vapor. A method is explained that is successful at determining the atmospheric temperature along the beam by using two water vapor absorption lines (3281 cm-1 and 3283 cm-1) in the single beam spectrum that have opposite and strong temperature dependencies. The regression model is based on synthetic data created with the HI-TRAN database and shows good agreement with field data. Lastly a new way to quantify gases from the single beam spectrum of the OP-FTIR is introduced and tested. This method contrasts with traditional absorbance based methods and avoids the pitfalls associated with the background spectrum. The input spectra are divided into two arrays. One of these arrays is associated with the points in wave-number space where the target gas has less absorbance and the second array contains information about the points where the target gas absorbs most. A series of reference transmittance spectra are divided from the input spectra and the arrays converge at the correct concentration. The data suggests that this method is very robust to noise, sinusoidal interference, and instrumental changes. The method also shows the ability to quantify two overlapping gasses through iteration with few cycles.