Ethanol Detection Strategy with Multiple Digital Filtering of Passive FT-IR Interferograms

Digital filtering methods are evaluated for automated detection of ethanol using passive Fourier transform infrared (FT-IR) data collected during laboratory and open-air experiments. In applications where ethanol signals are overlapped by spectral interference signals (e.g., ammonia and acetone), the use of multiple digital filters is found to improve the sensitivity of the vapor analyte detection. The detection strategy applies bandpass digital filters to short interferogram segments that are acquired from the passive FT-IR spectrometer configuration. To implement the automated detection of the ethanol target analyte, the filtered interferogram segments are input into a piece-wise linear discriminant analysis. Through the use of a set of training data, discriminants are computed that are subsequently used for automated detection of ethanol vapor. A two-filter strategy with separate ethanol and ammonia filters is compared to a single ethanol filter approach. Bandpass parameters of the digital filters and the interferogram segment location are optimized with laboratory data. The laboratory data are generated for ethanol, ammonia, and acetone vapor mixtures in a gas cell, whose contents are viewed against various infrared background radiances. The optimized parameters, from the laboratory data, are subsequently tested with open-air remote sensing data. The open-air data consists of elevated temperature ethanol and ammonia plumes generated from a portable emission stack. The two-filter strategy outperforms the single-filter approach in laboratory and open-air scenarios, where the ammonia spectral interference dominates the ethanol spectral signature. Prepared in collaboration with Ohio University, Athens, OH.