Dispersive infrared spectroscopy measurements of atmospheric CO₂ using a Fabry-Pérot interferometer sensor

In this paper, we present the first dispersive infrared spectroscopic (DIRS) measurement of atmospheric carbon dioxide (CO2) using a new scanning Fabry-Pérot interferometer (FPI) sensor. The sensor measures the optical spectra in the mid infrared (3,900 nm to 5,220 nm) wavelength range with full width half maximum (FWHM) spectral resolution of 78.8 nm at the CO2 absorption band (~4,280 nm) and sampling resolution of 20 nm. The CO2 concentration is determined from the measured optical absorption spectra by fitting it to the CO2 reference spectrum. Interference from other major absorbers in the same wavelength range, e.g., carbon monoxide (CO) and water vapor (H2O), was taken out by including their reference spectra in the fit as well. The detailed descriptions of the instrumental setup, the retrieval procedure, a modeling study for error analysis as well as laboratory validation using standard gas concentrations are presented. An iterative algorithm to account for the non-linear response of the fit function to the absorption cross sections due to the broad instrument function was developed and tested. A modeling study of the retrieval algorithm showed that errors due to instrument noise can be considerably reduced by using the dispersive spectral information in the retrieval. The mean measurement error of the prototype DIRS CO2 measurement for 1 minute averaged data is about ±2.5 ppmv, and down to ± 0.8ppmv for 10 minute averaged data. A field test of atmospheric CO2 measurements were carried out in an urban site in Hong Kong for a month and compared to a commercial non-dispersive infrared (NDIR) CO2 analyzer. 10 minute averaged data shows good agreement between the DIRS and NDIR measurements with Pearson correlation coefficient (R) of 0.99. This new method offers an alternative approach of atmospheric CO2 measurement featuring high accuracy, correction of non-linear absorption and interference of water vapor.