Quantitative treatment of coarsely binned low-resolution recordings in molecular absorption spectroscopy

Optical multichannel detectors like photodiode arrays or CCD cameras combined with grating spectrometers are commonly used as detection systems in quantitative absorption spectroscopy. As a trade-off to broad spectral coverage, banded spectral features are sometimes recorded with insufficient spectral resolution and/or insufficiently fine detector binning. This renders the true physical spectrum of recorded intensities changed by instrumental and spectrum specific artefacts thus impeding comparability between results from different set-ups. In this work, it is demonstrated that in the case of a "well-behaved"--i.e. free of ro-vibronic structure--absorption band like the iodine monoxide IO(4<--0) transition, these effects can easily change the apparent peak absorption by up to 50%. Also deviations from the strict linearity (Beer-Lambert's law) between absorber concentration and apparent, i.e. pixelwise optical density occur. This can be critical in studies of chemical kinetics. It is shown that the observed non-linearity can cause errors of up to 50% in the determination of a second order rate coefficient for the IO self reaction. To overcome the problem, a consistent and rigorous integral approach for the treatment of intensity recordings is developed. Linearity between optical density and absorber concentration thereby is re-established. The method is validated using artificial test data as well as experimental data of the IO(4<--0) absorption transition, obtained in the context of I2/O3 photochemistry studies. The agreement is accurate to within +/-2% (test data) and +/-3% (experimental data) supporting the validity of the approach. Possible consequences for other spectroscopic work are indicated.